2009: Proceedings of the 15th International Congress of Speleology, Kerrville, Texas, July 19-26, 2009, Volume 2, Symposia, Part 2

2009: Proceedings of the 15th International Congress of Speleology, Kerrville, Texas, July 19-26, 2009, Volume 2, Symposia, Part 2

Material Information

2009: Proceedings of the 15th International Congress of Speleology, Kerrville, Texas, July 19-26, 2009, Volume 2, Symposia, Part 2
Alternate Title:
International Congress of Speleology
Alternate Title:
ICS Proceedings
White, William Blaine, 1934---
Greyhound Press
Publication Date:


Conference Proceeding
serial ( sobekcm )


General Note:
Note: an uncompressed version is available by selecting the "Additional Digital Versions" URL below. From the Preface: "Five Hundred papers were presented at the Fifteenth International Speleological Congress, Kerrville, Texas, USA on July, 19-26, 2009 by speleologists from all over the world. These volumes contain the written record for those papers. Authors who chose to do so were invited to prepare full papers of up to six pages. Authors who preferred a more limited text contributed abstracts of their papers for the Proceedings. The papers fall into two categories: those that were incorporated into the 13 symposia; 300 papers and those that were contributed to topical sessions; 200 papers. Written accounts appear for both oral presentations and papers that were presented as posters. The papers are arranged alphabetically by first author in the sequence Plenary lectures, Symposia papers, and Contributed papers. Both abstracts and papers received comprehensive technical reviews by the Science Committee. The authors had the opportunity to revise their papers in response to reviewer's comments. It is hoped that the review process has improved the clarity of the papers so that information transfer is enhanced."
Open Access - Permission by Publisher
General Note:
See Extended description for more information.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
K26-00111 ( USFLDC DOI )
k26.111 ( USFLDC Handle )
8371 ( karstportal - original NodeID )
9781879961340 ( ISBN )

USFLDC Membership

Added automatically
Karst Information Portal

Postcard Information



This item has the following downloads:

Full Text


Cover design: Beth Fratesi Layout and design: Greyhound Press Editor: William B. White15th International Congress of SpeleologyKerrville, Texas United States of America July 19, 2009 Volume 2, Symposia, Part 2Proceedings


PROcee CEE Di I NGs S 15t T H INte TE RNati ATI ONal AL CONGRess ESS OF Spele PELE Ol L OGYKerrville, Texas United States of America July 19, 2009Produced by the organizing Committee of the 15th Internatioal Congress of Speleology Published by the International Union of Speleology 2009 National Speleological Society, Inc. Individual authors retain their copyrights as indicated in the text. All rights reserved. No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any data storage or retrieval system without the express written permission of the copyright owner. All drawings and maps are used with permission of the artists. Unauthorized use is strictly prohibited. Printed in the United States of America. Library of Congress Control Number 2009930608 ISBN 978-1-879961-34-0VOlu LU Me E 2 SYMp P Osia SIA Pa A Rt T 2


Proceedings of the 15th International Congress of Speleology Errata and Omissions The Proceedings of the 15th International Congress of Speleology contain either abstracts or full papers of the 500 contributions presented at the Congress. The three volumes of the Proceedings total 2130 pages. The pathway to this mass of material was as follows: Prospective authors submitted an initial abstract to the ICS Science Committee. These abstracts were reviewed by the Committee to ascertain that the subject matter was appropriate for the Congress. The abstracts were then returned to the authors with suggestions and an invitation to prepare a full paper limited to six printed pages. Few papers were rejected, but some were withdrawn so that of 540 initial submissions, 500 were presented at the Congress. The draft papers were sent to the Science Committee who distributed them for review after which they were returned to the authors for such adjustments as the reviewers deemed necessary The final papers were received by the Science Committee for formal acceptance and were forwarded to the editor. The edited papers were then transmitted to Production Manager for page layout and preparation for the printer. All of this movement of abstracts and manuscripts was done electronically. In the process of transmittals, various reviews, and editorial handling, a few errors and omissions were created. The lists that follow contain the additions and corrections that have been brought to our attention. We have limited the corrections to matters of fact; small errors in spelling, punctuation, and formatting are not addressed. We apologize to the authors whose papers were mishandled in some manner. The Editorial Team Errata Volume 1, Page 541 Cave Sediments Related to Cretaceous Tertiary Paleokarst Developed in Eogenetic Carbonate Rocks: Examples from SW Slovenia and NW Croatia by Bojan Otoni The abstract was truncated in printing with only the first few lines appearing in the Proceedings. The full abstract follows. In the SW Slovenia and NW Croatia a regional paleokarstic surface separates the passive margin shallow marine carbonate successions of different Cretaceous formations from the Upper Cretaceous to Eocene palustrine and shallow marine limestones of the synorogenic carbonate platform. Thus, the paleokarst corresponds to an uplifted peripheral foreland bulge, when diagenetically immature eugenetic carbonates were subaerially exposed and karstified


Among the subsurface paleok a rstic features vadose and phreatic forms are recognized. For the epikarst, pedogenic features and enlarged root related channels are characteristic. Vado se channels, shafts and pits penetrate up to a few tens of meters bellow the paleokarstic surface, where they may merge with originally horizontally oriented phreatic cavities. The latter comprise characteristics of caves forming in fresh/brackish water le nses. T he phreatic cavities were found in different positions regarding to the paleokarstic surface, the lowest one being some 75 meters below it. Usually only one distinct paleocave level occurs per location, although indistinct levels of spongy porosity and/or irregularly dispersed cavities of different sizes have been noticed locally. The cavities had been subsequently partly reshaped and entirely filled with detrital sediments and flowstones in the upper part of the phreatic, epiphreatic and vadose zones. The internal cave sediments and flowstones may also occur as clasts in deposits (mostly breccias) that fill subsurface paleokarstic cavities and cover the paleokarstic surface. In general, the variety of cave infilling deposits and the amount of surface derived material decrease with the distance from the paleokarstic surface. Below 1318O values of cavity deposits usually exhibit good correlation with trend sign ificant for meteoric diagenesis. Relatively small p hreatic cavities of the lowermost part of the paleokarstic profile s are commonly geopet ally infilled with laminated mudstone derived from incomplete dissolution of the hostrock overlain by coarse grained blocky calcite of meteoric or mixing meteoric/ marine origin. S omewhat larger phreatic caves located shallower below the paleokarstic surface usually exhibit more complicated stratigraphy. Although the lower parts of the caves are still mainly infilled with reddish stained micritic carbonate sediment different types of flowstone, especially calcite rafts become more prominent higher in the cave profiles. Gradually in the upper parts of the cav es, sediments derived from the paleokarstic surface prevail over autochthonous deposits. Especially channels of the epikarst zone are almost entirely infilled with pedogenically modified material derived directly from the paleokarstic surface. Regardless of their origin, cave deposits had been often intensively modified by pedogenic processes while they were exposed to the paleokarstic surface by denudation. Just prior to marine transgression over the paleokarstic surface some cavities or their parts had been infilled by marine derived microturbidites. It will be shown that especially deposits related to denuded phreatic caves may be of great importance for the study of speleogenetic, geomorphologic and hydrogeologic evolution of a specific karst region. Volume 2, page 650 Medical and Governmental Considerations of CO2 and O2 in Volcanic Caves by William R. Halliday The final sentence of the first paragraph on page 652 contains incorrect wording. The sentence should read: The issue resurfaced when U.S. Geological Survey and National Park Service personnel applied OSHA standards to volunteers in volcanic caves with nontoxic levels of O2 and CO2. Volume 2, page 662 Unusual Rheogenic Caves of the 1919 Postal Rift Lava Flow, Kilauea Caldera, Hawaii by William R. Halliday The first paragraph on page 664 contains several errors and misstatements. The corrected paragraph should read:


Noxious gas (probably HCl) was encountered only in one tiny cave on the edge of Halemaumau Crater. Presumed sulfate fumes were encountered in numerous caves but were found to be essentially non toxic. Eye irritation rarely was encountered ( Halliday, 2000b). Two types of CO2 monitors previously untested in volcanic caves were required for the last five field trips. They were found to be useless in hyperthermal caves and no sig nificant elevation of CO2 was identi fied in normothermic examples (Halliday, 2007). In no cave was significantly elevated CO2 identified by changes in normal bre athing (Halliday, this volume ). Volume 2, Page 785 Symposium #11, Speleogenesis in Regional Geological Evolution and Its Role in Karst Hydrogeology and Geomorphology was arranged by Alexander Klimchouk and Arthur N. Palmer (no t by John Mylroie and Angel Gins as listed on the title page of the symposium in the Proceedings). Volume 2, Page 1033 Uranium Mapping in Speleothems: Occurrence of Diagenesis, Det rital Contamination and Geochemical Consequences The correct authors for this paper are: Richard Maire, Guillaume Deves, Ann -Sophie Perroux, Bassam Ghaleb, Benjamin Lans, Thomas Bacquart, Cyril Plaisir, Yves Quinif and Richard Ortega. The names of Bassam Ghaleb and Yves Quinif were omitted in the Proceedings Volume. Volume 3, Page 1307 Species Limits, Phylogenetics, and Conservation of Neoleptoneta Spiders in Texas Caves by Joel Ledford, Pierre Paquin, and Charles Griswold James Cokendolpher, Museum of Texas, Texas Tech University, Lubbock, Texas was also a co author for this paper. Omissions The Fossil Bears of Southeast Alaska by Timothy H. Heaton and Frederick Grady was inadvertently omitted in the final stages of page layout. The reviewed and edited paper follows:


THE FOSSIL BEARS OF SOUTHEAST ALASKA TIMOTHY H. HEATON1, FREDERICK GRADY2 1Department of Earth Sciences, University of South Dakota, Vermillion, SD, 57069, USA 2Department of Paleobiology, Smithsonian Institution, Washington, DC, 20560, USA Southeast Alaska is home to brown bears ( Ursus arctos) and black bears ( U. americanus ) with an unusual distribution. Both species inhabit the mainland while on ly black bears inhabit the i slands south of Frederick Sound and o nly brown bears inhabit the islands north of Frederick Sound. Brown bears of the northern islands belong to a distinct lineage and are genetically more similar to polar bears than their mainland counterparts. Bears are among the most common fossils found in caves in the regi on, and they indicate that both species made greater use of caves as dens when the climate was colder. B ut no bear fossils are known from the Last Glacial Maximum (LGM), even at On Your Knees Cave where foxes and marine mammals have been recovered across m ost of this interval. This begs the question of whether bears survived the LGM on coastal refugia or recolonized the islands after the ice retreated. No evidence has been found to settle the question for black bears. Black bears are far more common than b r own bears in On Your Knees Cave for the period before the LGM but they were slower than brown bears in expanding their range across the islands after the ice melted The evidence for survival in a local refugium is much stronger for brown bears. While the y are less common before the LGM, they had a greater distribution than black bears immediately following the LGM, including some of the outermost islands of the archipelago. The lack of brown bear fossils from mainland sites during early postglacial times may indicate that the mainland was not the source of this population The distinct genetic character of modern island brown bears also suggests that they did not derive from the mainland Two fossil brown bears from caves of Prince of Wales Island have had successful DNA extractions and match the distinct lineage that now lives only on the northern islands of Southeast Alaska. A refugium for brown bears may have been offshore on the continental shelf which was exposed during the LGM but was flooded by rising sea level in the early postglacial period. 1. Introduction Our research in southeast Alaska began in 1991 after several bear skeletons were found in El Capitan Cave on Prince of Wales Island by a caving expedition (HEATON and G RADY, 1992, 1993). El Capi tan Cave is Alaskas largest known cave and has passages that flood during storms, but the fossils were found in a quiet upper passage near the surface. One skeleton was complete and undisturbed, suggesting that the bears were denning in the cave, so caver s called this passage the Hibernaculum. It was apparent that the bears accessed the cave by an entrance that had become sealed with soil and logs, and we were able to reopen this entrance to conduct an excavation of the site. Soon cavers discovered skeleto ns in other caves of the region with similar dimensions, namely horizontal passages 1.5 2.5 meters in diameter. Several natural trap caves with bear fossils were also discovered. Although our research has expanded to include a variety of mammals, birds, an d fishes (H EATON and G RADY, 2003), bears have remained a major focus, and our fossil discoveries have contributed to solving the question of whether animals survived the Ice Age in Southeast Alaska. Most islands of Southeast Alaska are home to bears, but currently there is no more than one species per island. Black bears ( U rsus americanus ) inhabit Prince of Wales Island and most other islands south of Frederick Sound, while brown bears ( Ursus arctos) inhabit the islands north of Frederick Sound, namel y Adm iralty, Baranof, and Chicha gof (ABC) islands. Both species inhabit the nearby mainland (MACDONALD and COOK, 2007). Prior to the discovery of a fossil record KLEIN (1965) proposed that this island distribution resulted from a postglacial colonization histo ry: brown bears arriving from the north and black bears from the south. This hypothesis was based on the prevailing assumption that no land animals survived the Last Glacial Maximum (LGM, 24,00013,000 radiocarbon years B.P.) in Southeast Alaska because of complete ice cover. Although the islands of Southeast Alaska exhibit a nested mammalian fauna suggestive of recent colonization (CONROY et al., 2000), fossil and genetic studies of bears have revealed a much more complex history in the region. The complete skeleton from El Capitan Cave, as well as portions of several others, were of black bears, distinguished from the living bears on the island only by their large size. Their size seemed especially significant since they appeared to be females based on the lack of bacula and the gracile structure of their skulls. Even more significant was the discovery of even larger bear remains that we identified as brown bear. Finding that Prince of


Wales Island had been home to additional species in early postglacial t ime conflicted with the simple postglacial colonization model held by K LEIN (1965) and other biologists. In addition to brown bears we also discovered fossil remains of Arctic fox ( Alopex lagopus ), red fox (Vulpes vulpes ), and caribou ( Ranifer tarandus ) t hat no longer inhabit the island. Rather than lacking a fauna at the end of the Ice Age, Prince of Wales Island simply had a different fauna that was adapted to the colder and less forested habitat Following this initial discovery we set out to expand ou r dataset both geographically and chronologically by searching for caves with fossil deposits on different islands and the mainland, in diverse habitats, and of greater antiquity. During the 1990s fossil sites were brought to our attention by cavers explor ing the region, often working with the support of Tongass National Forest and guided by forest agendas. After 2000 we began coordinating searches for caves specifically to fill in gaps in our dataset. In spite of limits imposed by limestone distribution an d the difficulty of finding sites over 12,000 years old, a long history for both brown and black bears has emerged. During this same period geneticists began DNA studies on living bear populations in Southeast Alaska that complemented our work (H EATON et a l. 1996), and we have worked in conjunction with ancient DNA researchers to trace bear lineages back in time. What has emerged is a greatly expanded, but not entirely complete, picture of bear history in Southeast Alaska. 2. Postglacial History The postg lacial record of bears in Southeast Alaska is spectacular. Following the discovery of black and brown bears in El Capitan Cave (130 m elevation), additional brown bear skeletons were found in two high elevation caves (over 500 m) on northern Prince of Wale s Island: two juveniles in a natural trap called Blowing in the Wind Cave, and parts of 12 individuals in a horizontal tube called Bumper Cave, including skeletons of what appeared to be a mother and her two cubs (Table 1). By contrast, lower elevation cav es (below 200 m) on the island, such as Kushtaka and On Your Knees caves (den sites) and Tlacatzinacantli Cave (a natural trap) contained only black bears from the postglacial interval (Table 2). This apparent partitioning of den sites by the two species m ust be kept in mind when considering other parts of Southeast Alaska where samples from diverse elevations are not available. This does not mean that brown bears were restricted to high elevations because their isotopic signature indicates a stronger marin e diet than black bears (H EATON 1995; H EATON and G RADY, 2003). Table 1. List of radiocarbon dated brown bear ( Ursus arctos ) fossils from caves of Southeast Alaska in order of age. Laboratory # Age (years B.P. ) 13 C Site Island Sample AA15224 7,205 65 17.9 Bumper Cave POW Dentary AA56996 9,590 95 20.5 Deer Bone Cave Coronation Radius AA 07794 9,760 75 18.0 El Capitan Cave POW Humerus AA 10451 9,995 95 18.5 Blowing in the Wind Cave POW Ribs AA-5 2223 10,700 100 17.1 Enigma Cave Dall Humerus AA15225 10,970 85 19.5 Bumper Cave POW Molar AA15223 11,225 110 16.8 Bumper Cave POW Humerus AA 52221 11,600 100 14.6 Enigma Cave Dall Dentary AA44450 11,630 120 18.2 Colander Cave Coronat ion Humerus AA15222 11,640 80 17.8 Bumper Cave POW Rib AA 15226 11,715 120 16.0 Enigma Cave Dall Humerus AA 32122 11,910 140 18.1 El Capitan Cave POW Rib2 AA52222 11,930 120 14.6 Enigma Cave Dall Skull AA10445 12,295 120 18.3 El Capi tan Cave POW Pelvis AA 33783 26,820 700 16.3 On Your Knees Cave POW Astragalus AA 52219 29,040 600 16.3 On Your Knees Cave POW Rib AA52220 29,590 980 17.7 On Your Knees Cave POW M2/ AA33792 31,700 1900 16.2 On Your Knees Cave POW Molar AA52218 31,900 1,300 19.6 On Your Knees Cave POW Claw AA 52207 33,300 1,500 17.0 On Your Knees Cave POW Phalanx 1 AA15227 35,365 800 15.9 On Your Knees Cave POW Femur AA52215 38,800 3,000 10.0 On Your Knees Cave POW Phalanx 2


AA 33791 39, 400 3100 17.1 On Your Knees Cave POW Tooth AA 52216 34,000 + 17.4 On Your Knees Cave POW M/1 AA52201 40,900 + 16.8 On Your Knees Cave POW P4/ AA52217 41,100 + 15.4 On Your Knees Cave POW Vertebra Table 2. List of radiocarbon dated black bear ( Ursus americanus) fossils from caves of Southeast Alaska in order of age. Laboratory # Age (years B.P. ) 13 C Site Island Sample CAMS 27263 2,790 60 23.2 Kushtaka Cave POW Artifact AA57000 3,425 50 12.5 Lawyers Cave Mainland Humerus CAMS 31068 3,960 50 20.7 On Your Knees Cave POW Dentary AA 36637 4,847 58 21.2 Hole 52 Cave Mainland Skull SR 5265 6,290 50 Lawyers Cave Mainland Phalanx AA10447 6,415 130 22.1 El Capitan Cave POW Skull CAMS 24967 8,630 60 21.4 Kushtaka Cave POW Rib AA 18451R 9,330 155 23.9 Kushtaka Cave POW Femur AA 32118 10,020 110 22.1 Tlacatzinacantli Cave POW Femur AA36641 10,080 120 21.6 Hole 52 Cave Mainland Phalanx AA 33780 10,090 160 21.2 On Your Knees Cave POW Phalanx CAMS 42381 10,300 50 20.7 On Your Knees Cave POW Artifact AA 36636 10,350 100 18.9 Hole 52 Cave Mainland Skull AA36640 10,420 110 21.6 Hole 52 Cave Mainland Skull AA 07793 10,745 75 21.1 El Capitan Cave POW Humerus AA 32120 10,860 120 21.8 Tlacatzinacantli Cave POW Skull AA 32117 10,870 120 21.8 Tlacatzinacantli Cave POW Ulna AA36638 10,930 140 19.8 H ole 52 Cave Mainland Skull AA 32119 10,970 120 22.4 Tlacatzinacantli Cave POW Fragment AA 33202 11,460 130 19.9 Hole 52 Cave Mainland Canine AA 10446 11,540 110 20.0 El Capitan Cave POW Skull AA10448 11,565 115 18.7 El Capitan Cave POW Sku ll AA 21569 28,695 360 20.7 On Your Knees Cave POW Calcaneum AA 21570 29,820 400 20.8 On Your Knees Cave POW Vertebra AA 33781 36,770 2300 18.6 On Your Knees Cave POW Femur AA33194 38,400 3000 18.4 On Your Knees Cave POW Humerus AA 33198 39,000 3100 19.5 On Your Knees Cave POW Rib AA 16831 41,600 1500 20.7 On Your Knees Cave POW Tibia AA 36653 25,000 + 22.0 On Your Knees Cave POW Premolar AA36655 27,000 + 18.2 On Your Knees Cave POW Baculum AA 33196 38,500 + 19.4 On Your Knee s Cave POW Scapula AA 52206 38,500 + 20.8 On Your Knees Cave POW Metapodial AA 52204 39,100 + 20.2 On Your Knees Cave POW Canine AA33200 39,400 + 19.3 On Your Knees Cave POW Canine AA 33195 40,100 + 18.4 On Your Knees Cave POW Humerus AA 33199 40 ,200 + 19.9 On Your Knees Cave POW Canine AA 44448 41,000 + 21.7 On Your Knees Cave POW Molar SR 5110 43,050 + On Your Knees Cave POW Vertebra SR 5111 44,940 + On Your Knees Cave POW Skull


Several postglacial deposits have also been found on the mainland near the town of Wrangell and on two of the outermost islands of the Archipelago: Coronation and Dall Islands (HEATON and GRADY, 2003) Today only black bears inhabit Dall Island while no bears inhabit Coronation Island (MACDONALD and COOK, 2007) Three early postglacial cave deposits have turned up six individuals, all of which match brown bear (Table 1). Deer Bone C ave is a den cave while Colander Cave is a natural trap, and Enigma Cave is larger and more complex with bear skeletons both in horiz ontal den passages and at the bottom of pits. All these caves are at 200 m elevation or lower. By contrast, two postglacial cave deposit s on the mainland, a den site called Lawyers Cave and a complex cave with horizontal passages and pits called Hole 52, c ontain only black bear remains (Table 2). Brown bears may have denned at higher elevation, but no such sites are known. The remarkable conclusion from these sites is that the two bear species had nearly the opposite distribution in the early postglacial period than they do today. Currently both species inhabit the mainland while only black bears inhabit the southern islands of Southeast Alaska. Shortly a fter the Ice Age only brown bears inhabited the outer islands, both species occupied the large Prince of Wales Island, and only black bears are documented from the mainland. Discovering the postglacial history of bears in the northern islands of Southeast Alaska, where only brown bears live today, has been hampered by a paucity of limestone and a lack of any fossil discovery. Since brown bears thrived in the southern islands in early postglacial times, there is no reason to doubt their presence farther north. Whether black bears ever colonized the northern islands remains a mystery. To the south of Alaska a p attern similar to Prince of Wales Island has been documented by Canadian investigators Haida Gwaii (Queen Charlotte Islands) and Vancouver Island are currently home only to black bears. Fossil black bears have been found dating back to 10,000 years B.P. o n Haida Gwaii (R AMSEY et al. 2004; F EDJE et al. 2004) and from about 9,800 to 12,000 years B.P. on Vancouver Island (N AGORSEN et al. 1995; N AGORSEN and K EDDIE, 2000). Brown bears from Haida Gwaii have been found dating from 10,000 to 14,500 years B.P., showing that they once were widespread on coastal islands. Another remarkable pattern visible in Tables 1 and 2 is the sheer number of early postglacial bears. With the exception of the sealed hibernaculum of El Capitan Cave, all of these sites remain open for potential denning today. Yet far more specimens of both black and brown bears date between 9,000 and 12,000 years B.P. than date to the 9,000 years since then. Most of these remains were exposed on the cave floors (not fully buried) so were not selected for dating based on their potential antiquity. Either bears were more numerous in early postglacial times or they were denning in caves much more regularly. The fact that natural trap caves (at least a third of the sites) show this same pattern suggests a high bear population. None of the other species we have studied show this distinct chronological pattern. Perhaps the early successional stages of forest development following the melting of the glaciers provided a high density of berries and other edible foods preferred by bears for the herbivorous part of their diet. Since climax forest s are lacking in such foods, modern bears are attracted to forest clear cuts, shorelines, and other disturbed areas where such plants grow. 3. Ice Age History The single site in Southeast Alaska that has produced an extensive Ice Age record (prior to 13,000 radiocarbon years B.P.) is On Your Knees Cave. It is a small cave on the northern tip of Prince of Wales Island discovered during a logging survey and had only a few bones initially exposed. The significance of the site was only recognized when a partial brown bear femur was radiocarbon dated to 35,365 years B.P. (Table 1). Excavation began in 1996 and continued until 2004. An extensive record of mammals, birds, and f ish was discovered covering at least the last 45,000 years (H EATON and G RADY, 2003) plus an extensive archaeological record including the oldest human remains from Alaska or Canada (D IXON et al. 1997). Devils Canopy Cave on Prince of Wales Island is the only other site where we obtained an Ice Age radiocarbon date (on marmot), but extensive excavation produced only a few rodent and insectivore remain s. Our extensive efforts to find an Ice Age site on the outer islands of Southeast Alaska have so far been unsuccessful. For a single site, On Your Knees Cave provides a superb record of animals during the LGM and the preceding interstadial. As can be seen in Tables 1 and 2 many bone dates are beyond the radiocarbon limit, but uranium dates on speleothem fragments date back to 185,800 2,800 years B.P. (D ORALE et al. 2003). Both black and brown bears were present and probably used the cave as a den from at least 41,000 years B.P. until the approach of the LGM (Tables 1 and 2). We have not dated enough samples to be certain exactly when their use of the cave ceased, but no bear remains have been dated to the glacial maximum itself. A sample of 25 ringed seal ( Phoca hispida) specimens were radiocarbon dated from 24,150 490 to 13,690 240 years B.P., which is the very interval that the


bears (and caribou) are missing. Arctic and red foxes, other marine mammals, and sea birds also date to the LGM so the cave was available and used as a den (by foxes) during that interval. One ringed seal humerus has bite marks that match bear canines, but it could be a polar bear ( Ursus maritimus) kill that was scavenged by foxes. Black bear fossil s outnumber brown bear fossil s in On Your Knees Cave by a ratio of about 10:1. This is not evident in Tables 1 and 2 because we selected specimens of both species for dating. This difference could represent a greater abundance of black bears or a partitioning of den sites by elevation like we see during the postglacial period. Other elements of the fauna suggest that conditions during the interstadial were similar to the early postglacial interval before a climax forest was established. 4. Genetics T ALBOT and S HIELDS (1996) found that brown bears of Admiralty, Baranof, and Chicha gof (ABC) islands (Southeast Alaskan islands north of Frederick Sound) a re distinct from all other populations based on mitochondrial DNA and are more closely related to polar bears than to their mainland counterparts Using nuclear microsatellite variations P AETKAU et al. (1998) confirmed this result for female s but detected some exchange of males with the local mainland population. L EONARD et al. (2000) discovered a fossil from Yukon Territory match ing the ABC bears and dating to 36,500 1,150 years B.P. so this clade had a wider distribution before the LGM Nevertheless, the current restricted range of this clade suggests that the islands of Southeast Alaska acted as a refugium for this population during the glacial maximum (HEATON et al., 1996) Further support for this hypothesis comes from early postglacial fossils of Prince of Wales Island and Haida Gwaii. After several failed attempts at extracting ancient DNA, B ARNES et al. (2002) reported that a brown bear fossil from Blowing in the Wind Cave ( AA10451 on Table 1) belong s to the ABC clade. Further work by S arah B ray (personal communication) also linked a bear from Bumper Cave ( AA16553 on Table 1) and ones from Haida Gwaii to the ABC clade. STONE and COOK (2000) found that black bears from the southern islands of Southeast Alaska belong to a mitochondri al lineage that is also found on the islands and coastal mainland of British Columbia and down the coast to northern California. Several other mammal species have distinct coastal lineages with a similar range, but it remains unclear whether the source of these lineages was south of Cordilleran glaciers or on coastal refugia, possibly in Southeast Alaska (COOK et al., 2001, 2006). 5. Conclusions The absence of a fossil record of bears from the LGM leaves open the question of whether they survived the glaci al expansion in Southeast Alaska on coastal refugia or recolonized afterward. Cave faunas document that both brown and black bears were present during the preceding interstadial and reappeared in great numbers soon after the ice melted. Genetic evidence fo r a distinct coastal lineage, where refugial isolation is the simplest explanation, is strong for brown bears but more equivocal for black bears. Both bears are refugial species in the sense that they were adversely affected by glaciation and struggled to survive under unfavorable climatic conditions. By contrast, other carnivores such as ringed seals, Arctic foxes, and likely polar bears flourished and expanded their ranges during the LGM. The extent to which the Arctic and refugium faunas competed with one another is unknown, but their interactions could have been a factor in the temporary loss of black and brown bears from On Your Knees Cave. What we learn from postglacial bears is that the species were able to move about freely and colonize territory th at was favorable for them, rather than being restricted by barriers and competition. Solving the full puzzle of bear history in Southeast Alaska will require finding additional faunas of similar antiquity to On Your Knees Cave, as a single site cannot document the movements of species. During the LGM the expanding glaciers pushed mammal populations westward, while falling sea level opened up new habitat to the west and changed the configuration of the coastal corridor. The possibility that populations of be ars and other mammals found suitable refugia to survive the LGM in Southeast Alaska is very possible. References BARNES, I., P. MATHEUS, B. SHAPIRO, D. JENSEN, and A. COOPER (2002) Dynamics of Pleistocene population extinctions in Beringian brown bears. S cience 295:2267 2270. CONROY, C. J., J. R. DEMBOSKI, and J. A. COOK (1999) Mammalian biogeography of the Al exander Archipelago of Alaska: a north temperate nested fauna. Journal of Biogeography 26:343 352.


COOK, J. A., A. L. BIDLACK, C. J. CONROY, J. R. DEMBOSKI, M. A. FLEMING, A. M. RUNCK, K. D. STONE, and S. O. MACDONALD (2001) A phylogeographic perspective on endemism in the Alexander Archipelago of the North Pacific. Biological Conservation 97:215 227. COOK, J. A., N. G. DAWSON, and S. O. MACDONALD (200 6) Conservation of highly fragmented systems: The north temperate Alexander Archipelago. Biological Conservation 133:1 15. DIXON, E. J., T. H. HEATON, T. E. FIFIELD, T. D. HAMILTON, D. E. PUTNAM, and F. GRADY (1997) Late Quaternary regional geoarchaeology of Southeast Alaska karst: A progress report. Geoarchaeology 12(6):689 712. DORALE, J. A., T. H. HEATON, and R. L. EDWARDS (2003) U Th dating of fossil associated cave calcites from southeastern Alaska. Geological Society of America Abstracts with Programs vol. 35, no. 6, p. 334. FEDJE, D. W., Q. MACKIE, E. J. DIXON, and T. H. HEATON (2004) Late Wisconsin environments and archaeological visibility on the northern Northwest Coast. In Entering America: Northeast Asia and Beringia before the Last Glacial Maxi mum D. B. Madsen (ed.), University of Utah Press, p. 97 138 HEATON, T. H. 13C values from vertebrate remains of the Alexander Archipelago, southeast Alaska. Current Research in the Pleistocene, vol. 12, pp. 95 97. HEATON, T. H., and F. GRADY (1992) Preliminary report on the fossil bears of El Capitan Cave, Prince of Wales Island, Alaska. Current Research in the Pleistocene 9:97 99. HEATON, T. H., and F. GRADY (1993) Fossil grizzly bears ( Ursus arctos ) from Prince of Wales Island, Alaska, offer new insights into animal dispersal, interspecific competition, and age of deglaciation. Current Research in the Pleistocene 10:98 100. HEATON, T. H., and F. GRADY (2003) The Late Wisconsin vertebrate history of Prince of Wales Island, Southe ast Alaska. I n Ice Age Cave Faunas of North America, B. W. Schubert, J. I. Mead, and R. W. Graham (eds.), Indiana University Press, p. 17 53. HEATON, T. H., S. L. TALBOT, and G. F. SHIELDS (1996) An Ice Age Refugium for Large Mammals in the Alexander Archipelago, Southeastern Alaska. Quaternary Research, 46(2):186 192. KLEIN, D. R (1965) Postglacial distribution patterns of mammals in the southern coastal regions of Alaska. Arctic 18:7 20. LEONARD, J. A., R. K. WAYNE, and A. COOPER (2000) Population genetics of Ice Age brown bears. Proceedings of the National Academy of Sciences 97(4):1651 1654. MACDONALD, S. O., and J. A. COOK (2007) Mammals and amphibians of Southeast Alaska. Museum of Sou thwestern Biology Special Publication 8:1 191. NAGORSEN, D. W. and G. KEDDIE (2000) Late Pleistocene Mountain Goats ( Oreamnos Americanus ) From Vancouver Island: biogeographic Implications Journal of Mammalogy 81(3): 666 675. NAGORSEN, D. W. G. KEDDIE, a nd R. J. HEBDA (1995) E arly Holocene black bears, Ursus americanus from Vancouver Island Canadian Field Nat uralist 109(1):11 18. PAETKAU, D., G. F. SHIELDS, and C. STROBECK (1998) Gene flow between insular, coastal and interior populations of brown bears in Alaska. Molecular Ecology 7:1283 1292. RAMSEY, C. L., P. A. GRIFFITHS, D. W. FEDJE, R. J. WIGEN, and Q. MACKIE (2004) Preliminary investigation of a late Wisconsinan fauna from K1 cave, Queen Charlotte Islands (Haida Gwaii), Canada. Quaternary Research 62(1):105 109. STONE. K. D., and J. A. COOK (2000) Phylogeography of black bears (Ursus americanus) of the Pacific Northwest. Canadian Journal of Zoology 78:1218 1223. TALBOT, S. L., and G. F. SHIELDS (1996) Phylogeography of brown bears ( Ursus arctos ) o f Alaska and paraphyly within the Ursidae. Molecular Phylogenetics and Evolution 5:477 494.


Symposium 9 LAVA CAVESArranged by: Stephan Kempe William R. Halliday


15th International Congress of Speleology Lava Caves 621 2009 ICS Proceedings e subject of our research is an articial cavity located not far from the town of Orvieto, at the edge of a volcanic ash plateau, belonging to the Volsinian volcanoes, a system active in the uaternary, between 0.6 and 0. Ma. is was a polycentric system, whose eruptions, mainly of explosive type, covered an area of about 2,000 hectares, at the centre of which is the lake of Bolsena, the largest European volcanic lake. During its history, this system has seen four dierent phases (some simultaneously): (a) Paleo-volsinian phase, the most ancient, about 0.6.5 Ma, (b) Bolsena-Orvieto phase, in whose bedding plane is located the tunnel we studied, (c) Latera caldera phase, (d) phreato-magmatic phase (Monteascone phase). e tectonic events which determined the rise of Volsinian volcanism and, broadly speaking, all the Tyrrhenian uaternary volcanism, are the same as the origin of a highly seismic belt along the Apennine mountain chain, responsible of the dramatic earthquake that ravaged the town of Assisi in 1997. e Apennine belt is on a seismically active plane corresponding to a subduction line of the Adriatic plate, thrust by Tyrrhenian events. Responsible for such dynamic is the pushing force of the African plate on the Eurasian continent which is shiing the Italian peninsula in a counterclockwise movement, widening the Tyrrhenian Sea and contracting the Adriatic. Two seismic belts run N-S through the peninsula. One is aligned with the Apennines as far south as the Calabrian arch. It is known for high magnitude earthquakes. e other is along the Tyrrhenian area and has lower intensity earthquakes. e depth of the Moho is about 25 km under the Tyrrhenian belt, and considerably deeper under the Apennines. e subduction zone dips about 80 degrees. Its origin was in the Cretaceous, and it is still active. Crustal thinning caused upwelling of huge amounts of sub-crustal magma, which, on reaching the surface, gave origin to wide volcanic topography developed on NW-SE fractures (Fig. 1). e converging plates (Adriatic subducting, Tyrrhenian thrusting) have produced wide volcanism, from Tuscany to the Somma-Vesuvio system. e process was in two distinct uaternary phases which generated chemically distinct products: the ANATECTIC TUSCAN PROVINCE (PAT), and the later ROMAN MAGMATIC PROVINCE (PCR; Washington, 1906). Extension was particularly THE ANCIENT SPRING: WATER IN THE LAND OF FIREEDOARDO BELLO O CCHI Club Speleologico Proteo, Vicenza, Italy, senalpha@gmail.com A group of speleologists and archaeologists have studied an articial cavity in Central Italy which dates back to the h century b B .c C It is a former Etrurian aqueduct, bearing witness to the ancient civilization of the Etruscans, pre-dating Roman times in Tyrrhenian Italy. e tunnel system was constructed in alkaline uaternary volcanic deposits, rich in leucite, within the Roman Magmatic Province. ey are characterized by a high potassium content (and therefore radiogenic K-40 as well) and uoride. is kind of magma is very rich in incompatible elements erupting to the surface through fractures caused by crustal thinning, a consequence of the complex tectonic movements of the Italian peninsula that is still active. e eruptions date back 0.5 Ma and caused the formation of a landscape typical for Etruria. is research revealed interesting archaeological details. e tunnel system is well constructed, complex, and compelled the builders to overcome diculties by adopting hydraulic engineering solutions that are remarkable for this time. is is conrmed by the fact that the spring intercepted by the tunnels has supplied water to a village without interruption until the 1960s. e chemical and radiological analyses revealed additional unexpected features. In the dry season, the spring discharges water exclusively of hypogenic origin, which exhibits a very high level of alpha radioactivity. It suggests that decomposition of the U-238 family is not the only source responsible for the high radioactivity, but that -232 is also an important source adding to the alpha activity.


Lava Caves 622 2009 ICS P roceedings 15th International Congress of Speleology active during the Tortonian-Messinian period. Main features are magmatic acidic eruptions, silica saturated (SiO2 > 60%) for the PAT, with domes at the surface and hypabyssal formations, due to high viscosity of lavas (Marinelli, 1961). Its age is between 4 and 1.3 Ma. e PCR was characterized by explosive features with rare emissions of lava, and the products are typically of the under-saturated, ultrapotassic, alkaline series, whose main mineral is leucite. Its main centers are the Volsinian, Vican and Sabatin at N of Rome, the Alban Hills by the capital, the Roccamonna system at Caserta, the Campi Flegrei and the Somma-Vesuvio to the south. Almost all of these have been active beginning 0.6 Ma. us they were co-existant with uaternary glaciations. While such magmatic regions are widely distributed, there are very few places on Earth with this particular type of alkaline magmatic bedding found in the PCR and the Leucyte Hills, in Wyoming. e cavity which we studied belongs to the Volsinian system, part of the PCR. Our researches revealed a very high emission of alpha radioactivity. is particular kind of volcanism is notoriously rich in incompatible elements, coming from immediately below the aesthenosphere. Its origin can be attributed to partial fusion of the mantle caused by the ascent of uids coming from the deep mantle, very rich in volatiles acting as carrier gases of incompatibles, including uranium and thorium. is particular process, called mantle degasation, causes the fusion of material from upper parts of the mantle. It has been suggested that this mechanism happens at a depth of about 80 km. Ascending magma interacts with the embedding rocks with processes of metasomatism, thus changing considerably its composition. Such processes happen in the upper mantle or the lower crust and the result is the formation of so-called glimmeritic materials, sometimes referred to as MARID, standing for Mica (phlogopite), Amphibole (K-Richterite), Rutile Ilmenite,Diopside. is kind of rock crystallizes from magmas strongly enriched with volatiles. So the volcanic products characteristic of the PCR are particularly rich in uranium and thorium and also potassium-40. Uranium and thorium are present together in hypogenic environments but in supercial environments follow dierent courses. Uranium has two oxidation states, while thorium can be found only in the tetravalent state. When they enter the supercial environment, uranium shis to the major oxidation state, and is soluble as a complex. orium is not. us we say that uranium is mobile, referring to this behaviour. Both are the progenitors of distinct radioactive families consisting mainly of alpha emitters, among which radon is gaseous and can be released without interacting signicantly with the chemical environment. Radon is localized half-way along the U-Pb decay series, and is in turn the father nucleus of a narrow series of alpha and beta emitters characterized by short half lives. Figure 1: Geological map of the Volsinean system. e tunnel is located in the circled area. Figure 2: Radon concentration in air. Figure 3: Average Rn concentration in Bq/day.


15th International Congress of Speleology Lava Caves 623 2009 ICS Proceedings e histograms in Figures 2 and 3 show seasonal variation of radioactivity. Such seasonal nature is also present in the karstic environment at the Buso della Rana Cave at Vicenza. Here we have a completely dierent environment, and the main emission source is water, specically its hypogenic component. In the hot season from May to October we have a long period of drought, and the tu beds drain no surface waters. During such periods, the spring drains waters of exclusively hypogenic origin (cold thermalism: 13.1 C). ese, before reaching the surface, undergo a deep water/rock interaction, absorbing radioactive elements from below. We performed radioactivity measurements using LR115 type SSNTDs (Solid State Nuclear Track Detector), closed in a container (Goblet A type) specically designed to let only air ow inside, excluding solid matter, thus having measures of alpha particles coming exclusively from radon decay e result, as developed and analysed in the laboratory of GT-Analytik of Innsbruck, is the measure of radon concentration. is suggests a seasonal nature of radon concentration in air, and what is most important, such variation follows the pattern of hypogenic origin in water: in the hot season, waters have almost exclusively a deep origin, and the radon concentration is higher than in the rainy season, in which we have mixed waters, hypogenic and meteoric. Further research will include a more detailed program, comparing SSNTD measures coming from dierent geological environments. e object of our research is an articial gallery at various angles. It follows the contact between an upper leucytic tephra level and a lower pyroclastic series. is articial cave structure is well constructed and complex, and is divided in two sections: the main intercepts a perennial spring, the other gives access to a tank, nowadays inactive. e section is typically long and narrow, with an ogival vault, big enough to let an operator stand comfortably. e aim of the engineers was not to collect water dripping from the vault, but to intercept a stream. e stream is perennial and consistent. In the absence of a collection work, water found its way to the surface following two directions: vertically through fractures in the thick tephra bank, and horizontally between the tephra bank and the under-bedding layer of loose ash. In such conditions the spring would have split into a swampy surface along the vertical line, and more streams making a buried spring at the inter-bedding banks, along the horizontal line. is made rational utilization impossible, since scattered uncontrolled springs turned a wood into a marshland. e ancients made several other attempts to dig tunnels nearby (the longest being about ten meters). e successful tunnel is characterized by a long corridor parallel to the external lining, and by adjustments and changes of direction. Its remarkable that planimetry changes by right angles is the technique attested in southern Etruria, thus making the tunnel develop by right-angled lines. e only curve corresponds to a by-pass of probable Roman age, made to pass beyond a collapsed part of the tunnel corresponding to a change of direction. e chamber of the spring is a narrow space half lled by three settling tanks built less than a century ago, the rst of which collects water directly from a hole in the rock, and the third drains to a small gutter on a side of the corridor by the exit, where a pipe conveys it to a fountain with a wash-basin. Deep inside the tunnel the walls are made of a sequence of unlithied ash. ey are beds of pyroclastic surges, hard enough to sustain the weight of the overlying deposits formed by Pelan clouds and by the collapse of Plinian columns. e tunnel alternates between layers of ne ash and sharp cornered breccia. e lateral branch is aected by structural problems of instability making exploration dangerous because of high risk of slides. As soon as one passes the narrow and partly collapsed access tunnel, the explorer nds himself in an irregular shaped passage showing visible evidence of slides. e ceiling is the oor of thick leucytic tephra that separates the tunnel from the ground surface, making exploration of this branch somewhat dangerous, but it can give information about the sedimentation surface. e deposition of tephra appears to have been uniform, suggesting a single eruptive event, while the lower beddings were in a sequence of very violent events, like Pelan clouds or Plinian columns. e walls of this latter part of the tunnel are made of clayish incoherent material incorporating sharp edged rocks, presumably not consisting of bombs, but of volcanic tuaceous rocks included in the plastic bulk turned later into clay by the hydrolizing action of thermal waters on silicates Figure 4; Sketch of the Etruscan tunnel.


Lava Caves 624 2009 ICS P roceedings 15th International Congress of Speleology While the Romans are known as surface builders, Etruscans could reasonably be considered a population of mining engineers. Both civilizations were involved in hydraulics, but the latter le us a huge number of hypogean works for collecting and conveying water, both spring and meteoric. In northern Latium area are numerous tunnels dug by the Etruscans for the control of waters, even in areas where the actual climatic conditions caused long periods of drought; indeed, most of those tunnels no longer drain water. Some of them had the function of draining meteoric waters, some others of collecting waters, others of tapping perennial springs as did the tunnel we studied. Tunnels made for collecting and conveying surface waters into basins far from the resurgence are just below the surface, between earthy and the tuaceous horizons. Turf, although permeable, drains water much less than the earthy surface, so these pipes drained water dropping from the vaults and transferred it to storage basins at the foot of the hills. is is the classic type very well represented in the Campagna Romana (Roman Plains) and in southern Etruria, whose landscape is characterized by mild slopes. North, at the borders of the Vulsinian system, are harsh morphologies with hills abruptly interrupted by steep slopes, made of accumulations of volcanic deposits not always lithied and subject to slides, covering clay soils originated from Tyrrhenian Pliocenic sea oors. is is the typical landscape of the upper Viterbo region, giving to it an original characteristic of wilderness where mankind has changed very little of the natural landscape. e presence of many sliding surfaces, enhanced by many seismic events and gas emissions, has dramatically inuenced settlement in the last two thousand years of recorded history. Striking examples are the clis of Orvieto and Civita by Bagnoregio: thick Pleistocene tu on a stable layer of Pliocene clays. Why did the ancient Etruscans dig such well-constructed works of hydraulic engineering where nowadays runs no water? is was because of conditions of paleometeorology. During the rst millenium B.C., conditions were dierent from the present. It was cooler and a bit rainier. We call this sub-Atlantic weather. We have evidence of this from the study of ancient pollen, and we know that two thousand years ago forests were more widespread than today, mainly of oaks but of beech trees as well. Nowadays we have beech-woods only on the slopes of Mount Venere (Venus Mountain), a volcano by Vico caldera lake, and some other beech trees above Mezzano maar lake. It is likely that, two thousand years ago, beech woods were at the border between warm and cold climates, and in the last millennium the warm zone prevailed over the cold one. e Mediterranean weather of today with cool winters and droughty torrid summers, developed recently, in terms of centuries, and the climate in the Etruscan age was characterized by cooler summers and more rainfall than today, enough to justify the excavation of tunnels most of which nowadays are dry. Acknowledgements We wish to thank individuals and institutions that have collaborated in the outcome of this work: Professor Paoloetto, of the University of Padua, founder of our group;GTAnalytic, in the person of doctor Jochen Gschnaller, from Innsbruck, for dosimeters LR 115; Helmut Singer Elektronik, from Aachen (Germany) for the Geiger counter FAG Kugelsher; Doctor Olivia Iacoangeli, geologist; and all the members of Grupo Speleological Proteo, particularly our president, Robert Farinati who supported us, and all the people of Castelgiorgio (Terni) and Benano Valley. Reference RAVELLI Franco, Etruscan cunicoli:tunnels for the collection of pure water, Twelh International Congress on Irrigation and Drainage, Fort Collins, USA, 1984


15th International Congress of Speleology Lava Caves 625 2009 ICS Proceedings PAHOEHOE AND LAVA TUBES IN PAYUNIA, NORTH PATAGONIA, ARGENTINACARl L Os S Be E Ne E De E TTO Federacin AA rgentina de Espeleologa, carlos_benedetto@inae.org.ar Payunia is a geographical region located in North Patagonia that includes parts of the Argentine provinces of Mendoza, Neuqun, and La Pampa. It owes its name to the Payn Matru volcano, located 280 km south of the city of Malarge, Mendoza. e base of the volcano is 28 km in diameter and its elevation is 3800 m above sea level. Volcanologists describe the region as having the largest basaltic ows in the world, at nearly 180 km in length, which contain many unexplored lava tubes. Almost 800 volcanoes occur in the region that produced numerous pahoehoe ows, some extremely uid, in which Argentine speleologists, in cooperation with British colleagues, have described several caves, the largest of which is 840 m long. One of these caves contains the richest suite of subterranean fauna in the region, made up largely of species not yet described. e main lava tubes explored to date are Cueva del Tigre, Cueva del Borne, Hoyo Dolo, Alero del Manzano, Cueva Doa Otilia (Mendoza province), Cueva de la Salamanca (Neuqun province), Caverna Halada (La Pampa province). Chacras de Aguado (Malarge, Mendoza) is formed in a very low viscosity pahoehoe. Doa Otilia and Manzano are mineralogically and biologically important. Italian volcanologists have geophysically detected potentially longer lava tubes not explored yet in the 180-km-long lava ow in South Mendoza.1. Introduction Payunia is a North Patagonic region with more than 600 inactive volcanoes that were active in the Pleistocene. It is named for the volcano Payn Matr, the largest of the region. All of the volcanoes are located in the Province of Mendoza, but the lava ows spread to the provinces of La Pampa and Neuquen. Known pahoehoe formations are located in the three provinces (Brojan et al, 1998; Benedetto, 1999, 2008; Brojan 2000), but the most important ones are located in Mendoza. Llambas (2003) identied basaltic uaternary eusions that have been sites of speleological explorations. e list of caves surveyed thus far are listed in Table 1 and demonstrate that the largest number of caves are concentrated near the volcanoes that gave them origin. Based on these data, maps were updated e region is arid (precipitation of about 300 mm per year) and biogeographically is a transition from PatagoniaCordillera-Monte. e average surface temperature is between 10oC and 11oC, which coincides with the average temperature hypogeum measured in all visited caves. 2. Doa Otilia Cavee Association CAE (Centro Argentino de Espeleologa) mapped 838 m of galleries in an horizontal development in Doa Otilia Cave. is is the longest known lava tube in Argentina, located 70 km to the southeast of Malarge city, and it is under the close supervision of the Family Zagal, the farmers not owners of the place. Speleologist can access the cave through a small hole in the oor, not easy to distinguish in epigean explorations, like almost all the basaltic caves of the region. In the inside, there are no major diculties for exploration. e cavity ends in a chaos of blocks, but the Name ProvinceDepartmentElev. (m) Borne MendozaMalarge Campana MahuidaNeuqunLoncopu Cerro GuachoNeuqunLoncopu Cueva del ZorroNeuqunPehuenches Doa Otilia MendozaMalarge1926 El Jagel NeuqunPehuenches982 El Manzano MendozaMalarge1350Table 1: List of caves explored. Halada La PampaPueln759 Hoyo Dolo MendozaMalarge1650 La Escondida NeuqunAelo La Salamanca NeuqunPehuenches Los Gatos NeuqunPehuenches880 Pozo del CampamentoMendozaMalarge1264 Puesto La BardaMendozaSan Rafael592 Tigre MendozaMalarge1474 Zagal MendozaMalarge1800


Lava Caves 626 2009 ICS P roceedings 15th International Congress of Speleology oor is comprises mostly sand and gravel. e penetration of roots from surface plants, which search to absorb the high humidity inside in the oor of the cave. e fauna noted to date have been grossly identied by Peralta and Benedetto (2007) as follows: 1. Animals associated with plants: -Order Collembola; -Class Oligochaeta; -Order Acariformes; Nematoda-Phyllum 2. Class Insecta: O.Diptera: Tipulidae (adult), other dipterans s / d; O. Blattaria, O. Coleoptera Part of the exoskeleton (accidental) 3. Class Arachnida: O. Araneae, O. Opiliones (troglobitic?) 4. Class Chilopoda 5. Reptile bones and mammals (accidental?),3. Alero El ManzanoAlero El Manzano is located on a basaltic wall in front of National Route 40, north of El Manzano, about 100 km south of Malarge city. is is a small cavity with no more than 10 m of development, but where Paolo Forti (Benedetto et al, 1998) identied its importance as having the highest concentration of phosphate cave minerals in the world. ere is a small colony of bats whose guano was, perhaps, the cause of these mineral formations4. Hoyo Dolo CaveHoyo Dolo Cave, close of the farm (puesto) Del Pozo, near the Dolo Vulcano, and on the southern tip of the province of Mendoza is a unique basaltic caves located within a protected natural area (Payunia). e CAE Association explored it and documented its 350 m length. e entrance is an opening of 20 m with a depth of 12 to 15 m. Inside, there are large blocks of collapses and toward the northeast and southwest are two large galleries where sunlight does not penetrate. e smaller gallery to the northeast is 20 m long with a height ranging from 5 15 m. e other gallery is 250 m long. e caved is a product of lava ows from Dolo Vulcano, and satellite-derived observations shows that the true development of galleries as almost 2.6 km. Speleologists could explore only 350 m inside, but the original extent was longer (Milillo, 1988).5. Del Tigre CaveDel Tigre Cave is the basaltic cave most visited by tourists, but it is not ocially sanctioned to such activities. It is located 58 km southeast of Malarge city. From the place of descent into the cave, the passage is divided into two branches: one goes to the north and the other one to the south. e second one ramies towards the end and has numerous landslides. Most of the cave oor is bare sand. e rst exploration in this cave was made by the CENTRO ARGENTINO DE ESPELEOLOGA (1973), but more explorations were done by the local Association INAE and the U.K.s Mendip Caving Group. Joint explorations of Argentinian and British speleologists that sought to open the nal collapses of the North gallery by following air ow were unsuccessful. Dr. Franco Urbani made an inventory of the minerals in the walls and ceiling of the cavity (Urbani & Benedetto, 1997) and, despite its extreme dryness, a biological survey in 1991 (Trajano, 1991) provided the following results: F. Chordata: cl. Mammalia: O. Chiroptera: Vespertilionidae. Lasiurus sp: observed two individuals O. Edentata: Tolypeutes matacus. O. Rodentia: Lagidium sp. O. Carnivora: Canis familiaris, Felis concolor; Lyncodon patagonicus. O. Bovidae: Capra hircus F. Arthropoda: Cl Hexapoda: O. Diptera: Muscidae: puperios O. Lepidoptera: Tineidae Cl Arachnida: O. Araneae: Pholcidae In 2008, during a visit by the Argentinian Federation, speleologists together with biologists of PCMA Project (Programa para la Conservacin de los Murcilagos de Argentina) collected some bats in this cave, but not of the kind mentioned by Trajano, which suggests a greater biological importance (Daniela Rodrguez & Vernica Chillo, com. Pers).6. Zagal CaveZagal Cave is located near Doa Otilia and Cerro Patahuilloso. Its name derives from the family that inhabits the region, the same family who takes care of Doa Otilia Cave. e 326 m long lava tube was explored and surveyed by the CAE. e access entrance is at the northern end of a circular depression that slightly slopes towards it. is entrance is narrow and upright for about ve meters, reaching a room with a block collapse, then descends gradually and narrows. e oor is composed of sandy sediment, with evidence of water circulation.


15th International Congress of Speleology Lava Caves 627 2009 ICS Proceedings 7. Del Borne Cave In 2007, the Instituto Argentino de Investigaciones Espeleolgicas (IN.A.E.) explored a new cave near Cueva del Tigre that they called Cueva del Borne. Its entrances is a large mouth, but the inner length is not remarkable. is cave was originally a lava tube but collapse has converted it to a clastic cave. e central structure (lava tube) confused with clasts. In a Project PCMA visit in late 2008, the entrance was detected by an overwhelming presence of fauna, especially chiroptera, birds, and mammals. In the rst sampling, the cave revealed a kind of natural museum with a great variety of animal species that have been predated by older animals. e cavity appears to be a kind of restaurant of a puma (Felis concolor). e cave appears to be a natural repository information about past conditions on the surface of the region (Daniela Rodrguez & Vernica Chillo com. pers.).8. Pozo del Campamento and ProjectionsIn 2005, INAE participated with Giorgio Pasquar (University of Miln Italy) in the partial exploration of Pozo del Campamento, an articial, 50 m deep, vertical pit in which stratigraphic observations were made in association with an almost 180 km long, the longest lava ow in the world (Pasquar et al, 2008). e ow reaches Ro Salado Valley (La Pampa province), maintaining a straight and narrow without following topography control. Pozo del Campamento is located in Salinillas, Malarge, Mendoza, near the border with La Pampa province. Lava tubes were detected that lacked previous access from the surface. ese investigations revealed extremely long lava ows that originated in the Andean retroarc, forming in the late uaternary. Pasquar et al. (2008) describes the rock that forms the casting as a hawaita with a low content of olivine and plagioclase phenocrysts. During an exploration, the team of speleologists descended this articial pit, which allowed the investigation of the strata in order to project information to all the region. 9. La Halada caveIn La Pampa province, near the ow described by Pasquar et al. (2008), the association CAE explored La Halada Cave (also called La Alada) (development 370 m, 10 m height dierence) thirty years ago, but subsequently the GEA association made a second survey. Like the other basaltic cavities, its entrance is visible only from very close as it is along ground level. e entrance is circular, only one meter diameter, with a vertical, pipe-shaped tube about 2 meters long that reaches the oor of cavity. Passage development is horizontal with a slight slope inwards. e galleries are broad (width maximum is 22.90 m) with an average height of 1.60 m. It has a main duct that branches into three galleries Cavern fauna were not observed but a lot of skeletons belonging to epigean fauna: birds, vizcachas, and small mammals. Early geologic reports remark that the cave is formed in olivine basalts from Cenozoic casting corresponding to 8 to 12 meters thick. In the region, cast of basalts of various ages can be distinguished (Grupo Espeleolgico Argentino, 2002, Martnez 1998).10. Lava Tubes in North NeuqunIn 1983, the CAE explored Salamanca Cave which is few kilometers from Mendoza border and a few meters from National Route 40 to the East, in the province of Neuqun near Buta Ranquil city. e caves entrance is similar to those described above. It formed in Cenozoic basalts, with a unilineal development of 204 meters. Subsequently, GEA explored the Jagel Cave (development of 324 meters, -25 meters) and Los Gatos (312 m of development, 36 meters gap) near Rincn de los Sauces city. ere are other data that have shown associations in Neuquen (Mahuida Campana, Cerro Guacho, Cueva del Zorro and La Escondida), which have the coordinates but not the full cadastral information at disposal. Speleology is forbidden now in Neuqun due to political problems, so we dont know if cavers had explored other lava tubes in that area. e information of this communication was produced before the prohibition.ReferencesBENEDETTO, C., P. FORTI, E. GALLI and A. ROSSI. 1998. Chemical deposits in volcanic caves of Argentina. En: Proceeding of the 8th International Symposium on Vulcanospeleology, Bucarest, Rumania. International Journal of Speleology (Edizione Italiana) (1998), 27B(1):155 BENEDETTO, C, 1999. Volcanic Caves in Argentina. En: IX International Symposium on Vulcanospeleology, Sicilia, Italia, pp 219 BENEDETTO, C, 2005. Vulcanoespeleologa en Payunia. En: Boletn SPELAION (INAE) 10 (30), pg,. 7. Malarge BENEDETTO, C, 2008. Estado actual del conocimiento de los tubos lvicos en la regin de payunia (Mendoza,La Pampa, Neuqun Argentina) En:


Lava Caves 628 2009 ICS P roceedings 15th International Congress of Speleology III Congreso Argentino de Espeleologa MalargeMendoza, febrero de 2008 pp 147 BENEDETTO, C., P. FORTI, E. GALLI, and A. ROSSI. 1998. Chemical deposits in volcanic caves of Argentina. En: Proceeding of the 8th international symposium on Vulcanospeleology, Bucarest, Rumania. International Journal of Speleology (Edizione Italiana) (1998), 27B(1):155 BENEDETTO, C. and M. PERALTA, 2007. Observaciones sobre la ecologa de la cueva Doa Otilia (Malarge, Mendoza, Argentina). Actas del V Congreso de Espeleologa-Federacin Espeleolgica de Amrica Latina y del Caribe (FEALC) y I Congreso de FEPUR, Aguadilla, Puerto Rico. BROJAN M., A. CASTRO and G. CASTRO 1998. Cavidades en lava y caliza al Sureste de Malarge, Mendoza, Argentina. Memorias del III Congreso Espeleolgico de Amrica Latina y el Caribe (III C.E.A.L.C.), Boletn El Gucharo N 43, Sociedad Venezolana de Espeleologa. Caracas. BROJAN M. 2000. Biologa en Cueva Doa Otilia. Malarge, Mendoza, Argentina. Actas del Ier. Congreso Nacional Argentino de Espeleologa, Revista Spelaion N 7, pp 55, Instituto Argentino de Investigaciones Espeleolgicas. Malarge CENTRO ARGENTINO DE ESPELEOLOGA. 1973. Algunas cuevas de la provincia de Mendoza y La Pampa, Repblica Argentina. Boletn Sociedad venezolana de Espeleologa, 4 (2): 141 CNCN Catastro Espeleolgico Argentino. Archivo de la Federacin Argentina de Espeleologa. 2007 GRUPO ESPELEOLGICO ARGENTINO. 1981. Informe sobre la caverna Halada, provincia de La Pampa, (L-1). Grupo Espeleolgico Argentino GEA, Buenos Aires, pp 1. LLAMBIAS, E., 2003. Geologa de los Cuerpos Igneos. Asociacin Geolgica Argentina Serie B Didctica y Complementaria Nro. 27. Buenos Aires MARTINEZ, O.N. 1998. Descripcin topogrca y geolgica de la caverna Halada y su entorno, Provincia De La Pampa. Revista Salamanca N, Grupo Espeleolgico Argentino, Buenos Aires, MILILLO Juan C. 1988. Expedicin a la Caverna Hoyo Dolo. Las Brujas N 1: 33. Anales 1983. Centro Argentino de Espeleologa. Buenos Aires. PASUAR, G., A. BISTACCHI, L. FRANCALANCI, G. BERTOTTO, E. BOARI, M. MASSIRONO, Na ROSSOTTI, 2008. Very long pahoehoe inated basaltic lava ows in the Payenia Volcanic province (Mendoza and La Pampa, Argentina). En: Revista de la Asociacin Geolgica Argentina 63 (1): 131. Buenos Aires TRAJANO, E., 1991. Notas biolgicas sobre cavernas argentinas. Resultados de la primera expedicin espeleolgica argentino-brasilea, NeuqunMendoza. Spelaion 2 (2): 3. URBANI, F and C. BENEDETTO, 1996. Mineraloga de algunas muestras de espeleotemas de cavidades de la Municipalidad de Malarge Mendoza, Argentina


15th International Congress of Speleology Lava Caves 629 2009 ICS Proceedings EXPLORATION OF MANU NUI LAVA TUBE, Hawaii AWAII USAANNANN BO O ST T ED D PET T ER R BO O ST T ED D 630 Valley Forge DDrive, NN ewport NN ews, VA A 23602, USA A We have been exploring and surveying Manu Nui Lava Tube System since 2003, and so far have discovered over 3.5 km of passage, with a vertical extent of 347 meters. e cave, located in the rain forest above Kona on the Big Island of Hawaii, is unusual for several reasons: (a) it is profusely decorated with lava stalactites up to 1 m long, oen hanging at signicant angles from the vertical; (b) areas of vividly colored lava, ranging from black and dark brown through red and orange to ochre; (c) areas of extensive fountaining, where lava splashed onto the walls and ceiling; (d) the slope of the passages is very steep for a lava tube system; and e) it contains bones from a now-extinct Hawaiian goose. In this talk, we will illustrate all of these features, and engage the audience in discussions of their origin. 1. Introductione Big Island of Hawaii is well known for its abundance of large and long lava tubes. Manu Nui is a lava tube system on Hualalai, one of the ve volcanoes on the Big Island of Hawaii. e biggest volcano on this island is Mauna Loa, which is also considered to be the biggest volcano in the world. It rises to 4 km above sea level or about 9 km above the seaoor. Hualalai is the third youngest and third-most historically active volcano on the Island of Hawaii. Some 300,000 years ago, the summit of Hualalai was at sea level, and today it is at 2,521 m above sea level, making it the third tallest volcano on the island. Detailed mapping has shown that 95% of the surface area of this basaltic shield volcano is less than 10,000 years old. 80% is less than 5,000 years old, more than half is less than 3,000 years old, and 25% is covered by ows less than 1,000 years old. e most recent ow, which occurred in 1801, owed over part of Manu Nui, so we are unable to know how far the tube system once extended. Hualalai is considered to be dormant, and, according to the USGS Web site (USGS 2009), is likely to erupt again in the next 100 years. Manu Nui is a 3.5 km system on the western ank of Hualalai, noted for its steepness and colorful speleothems. Manu Nui is part of a ow that occurred about 2,200 to 2,300 years ago. e geologic setting is shown in Figure 1, where it can be seen that the cave starts just below the Kaupulehhu crater (elevation 1,870 m). e cave has two branches: the northern branch ends where it meets the 1801 ow. e southern branch is still being explored, and may continue down the main axis of the ow. e most striking feature of the system is its very steep grade, which averages 27 percent (or 16 degrees). is is considerably higher than the average 4 percent gradient of the famous Kazamura Cave. 2. Exploration Historye rst recorded exploration of the lower part of the cave started in the late 1990s. e late Kathy Marcelius bought the lot that contains the lowest 500 m of the system, initially with the goal to preserve the natural ora and fauna of the tract. She discovered many pukas (collapse entrances), and named them aer the various native plants that grew in or near these entrances (for example, the Alani, Koa, Palani, and Clermonita entrances). She explored much of the main passage, and was amazed by the many stalactites and colorful lava ows. Realizing that this cave might be quite exceptional, she contacted William Halliday of the Hawaii Speleological Survey (HSS). A visit by members of the HSS conrmed the interesting nature of the cave. In February 2003, Kathy guided Don Coons, Ric Elhard, Rose Herera, and the authors to the system, and the main passage between the Maile and Koa Tree and Clermontia entrances was surveyed. Many side passages were noted, and photos taken. Don, Ric, and Barb Capocy came back the next month and surveyed the main passage from the Maile Figure 1: Google Earth view of Kaupulehhu crater area, with geologic overlay. Manu Nui cave system is shown as the dark black lines.


Lava Caves 630 2009 ICS P roceedings 15th International Congress of Speleology down to the Refrigerator entrance, again noting many side passages. Since the cave owed from an upper lot onto her lot, Kathy bought the adjacent lot to prevent logging and development. Over the next six years, the authors, together with Kathy and various friends, continued the cartography and surveyed most of the upper section of the main cave, and most of the small parallel tubes in both the upper and lower sections in a series of eighteen survey trips. On our most mauka (up-ow) trip, we came across some agging tied around a tree. Aer some discussion with Doug Medville, it became apparent that he had been surveying the uppermost part of the cave, which had originally been explored by Rob Pacheco and friends. Doug joined us on a trip and helped us resurvey much of what he had done before noticing that he was duplicating his eort in the upper portion of the cave. In the process, Peter found an obscure route through a breakdown pile in one of the pukas that turned out to be the beginning of the parallel branch to the south. On the next trip, the authors and Kathy explored makai (down-ow) in the longest section of the system yet found with no entrances: e Energizer Bunny Passage Sadly, this was to be her last trip, as she passed away less than a year later, a victim of cancer. e authors continued exploring and surveying the system makai in the southern branch in 2008 with her son, Donovan. As of this writing, several leads remain. e southern branch continues makai and there are still some small side passages and parallel tubes to be surveyed in the northern branch. 3. Genesis of Some of the Lava Stalactites in Manu Nuie western slope of Hualalai is notable for its steepness, which means that there are many lava falls (or cascades) in Manu Nui, and at the base of many of them there are, or were, lava plunge pools similar to those in the steepest parts of Kazamura, only smaller as the Manu Nui tube is considerably smaller than that of Kazamura. e western slope of Hualalai is also known for its high rainfall (60 to 70, or 150 cm to 180 cm, per year). Almost every aernoon, this slope is shrouded in mist and clouds when the rest of the island is clear and sunny. We speculate that pools of rainwater collected in the tubes dessicated plunge pools prior to the later ows of lava through the Manu Nui tube system. When very hot lava owed through the tube, it would have mixed with the cold water, with spectacular results. At present, the youngest of Hawaiis volcanoes, Kilauea, has actively owing lava, which reaches the ocean. us, we can readily observe, today, how when owing lava meets the ocean, we get spectacular fountaining of lava and huge plumes of steam. is is quite spectacular, especially when viewed at night, as not only is the lava thrown up into the air, but the sudden mixing of hot lava and cool water results in huge steam clouds. On the USGS website (USGS 2009), we read this explanation: When lava pours into the ocean at high rates from a lava-tube entry, beautiful and spectacular explosions called tephra jets commonly occur. With temperatures higher than 1,100 degrees Celsius, lava can instantly transform seawater to steam, causing explosions that blast hot rocks, water, and molten lava fragments into the air. In general, the more intense the incoming waves, the more energetic the tephra jets. is source goes on to discuss four types of explosions when lava meets ocean water while a lava delta is being formed, namely tephra jests, blasts, bubble bursts and littoral lava fountains. Of interest to us is the statement: When seawater and lava mix within the connes of a lava tube, pressure may build to cause explosions that blast a hole through the roof of the tube. Two types of steam-driven explosions may be generated in such a conned environment: littoral lava fountains and bubble bursts. Litoral lava fountains are described as: Spectacular and rare, this type of lava-seawater explosion produces fountains of molten lava and steam that reach heights of more than 100 m. e explosions of molten spatter, bombs, and smaller tephra fragments quickly build a circular cone on the subsided lava delta, sometimes in a matter of minutes. Originating from deeper within the subsided delta and closer to the shoreline than bubble bursts, littoral lava fountains are much more energetic and dangerous. (USGS 2009) Bubble bursts are characterized as: Sporadic bursts of molten, dome-shaped lava sheets emanating from a circular rupture in the roof of a tube a few meters inland from the shoreline. Individual bubbles can reach diameters of 10 m in less than 2 seconds before they burst. e bubble fragments continue on their radial trajectories for up to 10 m more before falling to the ground. At the end of a burst, a pool of lava that remains in the roof of the lava tube gradually drains away. ese bursts are frequently accompanied by a loud boom that shakes the entire delta. (USGS 2009) If we apply what the USGS vulcanologists have observed and documented when lava meets ocean water, we can begin to understand what might happen if hot lava met a pool of rainwater in a conned space, such as a lava tube, away from an ocean, and the meeting did NOT result in the blasting of a hole through the roof of the tube. We can imagine fountains of molten lava and great clouds of steam, but not


15th International Congress of Speleology Lava Caves 631 2009 ICS Proceedings enough to damage the tube. Small and large fragments of molten lava would be thrown up randomly and splatter onto the tubes ceiling and walls, as illustrated in Figure 2. Parts of the larger fragments of molten lava would become elongated and start to resemble small stalactites. e pressure from the steam would instantly cause the fragments of molten lava to outgas, shrink, solidify and become permanent. If another fragment was to be projected onto a newly-formed stalactite, the process of sticking, elongating, outgassing, shrinking and solidifying would be repeated, and the stalactite would grow in length and volume. Since the fragments of liquid lava fountained up would be of varied sizes and trajectories, the growth of these stalactites would be very random some would grow longer and thicker than others, and the shapes would vary greatly. However, the side of the stalactite facing the fountain would receive all additional fragments of molten lava, and the stalactite would grow in width in that direction, although gravity would ensure that it also grew in length. As a stalactite grows thicker and longer from successive fragments, it would also shield some stalactites from receiving more fragments. It would be like a large tree in a forest blocking the sun from smaller trees. In other words, a large stalactite would throw a shadow over the stalactites no longer in the range of the lava fountain. Some of the fragments of molten lava may not reach the top of the stalactite, but instead stick to the bottom, or tip of the stalactite. In such cases, the stalactites tip would grow towards the fountain, sometimes with a slight curl, which would get more predominant with successive fragments hitting and sticking to the tip of the stalactite. is process would continue until the water in the pool had all evaporated into steam, or the ow of the lava was interrupted. Today, when we visit Manu Nui, we observe that parts of the cave have a huge profusion of thick, oddly-shaped stalactites hanging away from the vertical (see Fig. 3). We nd stalactites growing like thick blades away from the walls in the shape of sloping benches, and oen reducing the width of the passage considerably. We nd other areas that are covered with small fragments of dierent colored lava. e lava ranges from black through dark and light brown, red, butterscotch, dark and light orange to ochre. Some grayish and greenish tints are also seen. is has the eect of color coding the profusion of fragments so that we can better observe how some of them would have landed. For Figure 2: A series of diagrams, not to scale, illustrating the cross section of a lava tube, showing what could happen when lava fountains in a conned space: a) shows a pool of water trapped underneath owing lava, turning to steam, and forcing random agments of lava to hit the walls and ceilings in random trajectories; b) shows more agments hitting walls and other agments of lava, resulting in stalactite-like speleothems; c) e fountain is losing steam and trajectories are not as high. Stalactites grow towards the source of the fountain, and some are long enough to shadow the areas behind them. Figure 3: Lava stalactites up to 1 m long, hanging at signicant angles om the vertical.


Lava Caves 632 2009 ICS P roceedings 15th International Congress of Speleology example, we observe that one elongated ochre fragment, in the shape of a streamer, was about 25 cm long when it reached a wall some 4 m above the oor. e fragment was lying horizontally when it hit. About 8 cm landed against one stalactite, the middle 10 cm hit the wall, and the nal 8 cm landed against another stalactite. We can easily observe the shadow zone caused by the two stalactites on the wall behind them. ere are many hundreds of such fragments observable at one time. And because of the wide variety of size and color, we can observe a denite pattern in the direction of the blobs trajectories (see Figs. 4 and 5). Space does not permit us to give a detailed description of each of the several places where fountaining in the cave is evidenced by a profusion of splatters, curved stalactites, or multi-colored blobs, but an observer with attention to detail and time on his hands could painstakingly document the occurrences where one color overlaps the other and perhaps reconstruct which colors owed earliest, and which colors owed last. Instead, we would prefer to seek answers to other questions, like How can so many colors originate from one spot? When we see lava owing it is red, but it generally cools to black. Can Manu Nui be like Mary Poppins jar of medicine, which poured a teaspoon of red medicine then, almost immediately, a teaspoon of blue? Disney jokes aside, could the answer lie in chemistry, heat, water, cooling time, steepness, or a combination of circumstances? We know that the chemical composition of lava varies, not only among igneous rocks, but among basaltic rocks, and indeed among ows of lava from the same vent. e chemical composition of dierent ows has been likened to a ngerprint. e chemical composition of lava is, according to Wikipedia, composed by weight, 45% 55% silicon dioxide, 5% 14% iron oxide, 14% alumina, 5% 12% magnesium oxide, 10% calcium oxide or lime, 2% 6% alkalis, and 0.5 2% titanium dioxide. Iron oxide and magnesium oxide seem to be the chemicals with both a signicant component and a signicant variation, but can they account for such a wide variety of colors from black to ochre, via red and olive green? It is our understanding that lava has not been chemically analyzed to determine what causes the dierent colors. It is popularly held that magnesium causes the black lava and that iron causes the red color when it oxidizes, but can these two facts explain the kaleidoscope of colors in Manu Nui? It is also held that rapid cooling on the outside causes a veneer of black lava around an inner core of lighter-colored lava. So the color Figure 4: Splatters of vividly colored lava, ranging om black and dark brown through red and orange to ochre. Figure 5: Sample spatters that indicate extensive fountaining, where lava splashed onto the walls and ceiling.


15th International Congress of Speleology Lava Caves 633 2009 ICS Proceedings black could indicate chemistry or genesis. Dr. William Halliday told us that the chemical composition of lava is very similar to that of clay, and he had observed that 90-year-old lava had already decomposed to clay. We are not potters, but a cursory survey of ceramics websites informed us that the color of terra cotta is controlled by not only the chemical composition of the clays, but also the ring temperature, the composition of gasses in the kiln, and the water content of the clay. We read (CERAMICS 2009): e iron-content of the clay used for earthenware gives a color which ranges om bu to dark red, or even cream, grey or black, according to the amount present and the atmosphere (notably the oxygen content) in the kiln during ring. Clays with low iron content can result in paler colors on ring, ranging om white to yellow.4. Description of Other Interesting FeaturesOne unusual feature, which the authors have not seen in any other lava tube, is illustrated in Figure 6. It is a mini volcano, about 15 cm tall, with a 5 cm wide opening at the top. e mini-volcano seems to have formed above a small opening to a lower passage that crosses underneath, whose ceiling is less than a meter below the oor of the passage with the volcano. Apparently, there was a small opening during one of the last ows through the system, and lava welled up from below into the upper passage, and the strong wind kept the opening intact. e connection is still open today, as evidenced by a breeze blowing through the opening. We noticed the feature when a team surveying the upper passage was able to clearly hear another team in the lower tube. e initial explorations of the cave system revealed a variety of bones. While most of these are common (goats, pigs, cows, and small birds), one set of bones seemed quite unusual: see Figure 7. Don Coons initially identied these bones as having belonged to a large extinct goose, but this has yet to be conrmed. Kathy, an ardent lover of nature, named the cave for this avian friend who perished in her cave Manu Nui means Big Bird in Hawaiian.4. ConclusionsMore work remains to fully survey and document this fascinating cave system. We hope that it can be preserved for the future as Kathy would have wanted it to be. We gratefully acknowledge the leadership and generous support of Kathy Marcelius and her family. References:USGS 2009: http://hvo.wr.usgs.gov/hazards/oceanentry/ delta explosions/and references therein. CERAMICS 2009: http://www.visual-arts-cork.com/ ceramics.htm Figure 6: Mini-volcano feature, which is about 15 cm tall and connects overlain passages. Figure 7: Bones om a large extinct Hawaiian goose.


Lava Caves 634 2009 ICS P roceedings 15th International Congress of Speleology SOME ASPECTS OF SPELEOGENESIS IN EXTRATERRESTRIAL ENVIRONMENTS: 1MERCURY AND MARSVITTOR TTOR IO O CA A ST T ELLAN AN I1, ARR ARR IGO O A A CIGNA NA2 1Lab. AA strosica Spaziale, Frascati (RRoma) Italy2Fraz. TT uo, Str. Bottino 2, I-14023 Cocconato (ATAT), Italy In this paper a general view of the aspects and problems of the extraterrestrial speleology is reported on the basis of the documents available from the Laboratory of Space Astrophysics (Italian Council of Research, CNR) in co-operation with NASA. In order to show the dierent aspects of the problem, some structures referred to speleogenetic processes observed on the surface of Mercury and Mars are here described. In the rst case structures due to the eusion of volcanic lava are recorded; in the case of Mars the situation is more complicated because the evidence of erosion caused by uid masses is associated to the volcanic phenomena. e fusion of permafrost in the underground of Mars could result in a pseudokarstic mechanism also able to produce extended caves.1. Introduction is paper is a revised and updated version of a paper delivered in 1978 at the 13th National Congress of Speleology held in Perugia, Italy, whose proceedings have never been published by the organizers, following a simple preprint without any illustration. Since Professor Vittorio Castellani, astrophysicist and speleologist, member of the Italian Accademia Nazionale dei Lincei, died on May 20, 2006, this paper gratefully is dedicated to his memory.2. Generalities Speleology, as study of the dierent kinds of cavities found in the Earth, has gathered much information not only on the karst phenomena sensu strictu, but also on examples of parakarst and pseudokarst phenomena. Space explorations have shown the existence of similar examples on the surface of other planets of the solar system. erefore, a new reason to investigate more deeply such phenomena on the Earth aroused in order to obtain a more detailed knowledge on the mechanism of their evolution with special regard to the environmental conditions. In this paper a general view of the aspects and problems of the extraterrestrial speleology is reported on the basis of the documents available within the Laboratory of Space Astrophysics (Italian Council of Researches, CNR) in the framework of the co-operation with the NASA. In order to show the dierent aspects of the problem, in Table 1 the main characteristics of the so called inner planets, which are similar to the Earth on account of their solid rock surface, are reported. Venus, as well as the Earth, is the only planet with a real atmosphere. For this reason heavy clouds cover its surface and consequently the information on the surface morphology was obtained mainly by radar images. Pictures of the surface were taken by the Venera missions, which landed on Venus and returned pictures of area around the landing place to show a surface with basalt rocks of dierent size. A study on the development of lava tubes in extraterrestrial conditions (Badino, 2008) is based on the heat released by Order from the Sun Name Masse (Earth= 1) Gravity (Earth= 1) Atmospheric. pressure (millibar) Temperature (K)CratersLava Volcanic structures 1Mercury0.0550.380 (He trace)440YesYesYes 2Venus0.8160.829.105 (CO2)748Yes YesYes 3Earth1.0001.0001000 (N2+O2)293(Yes)YesYes 3aMoon0.01230.165 0 245YesYesYes 4Mars0.110.386 (CO2) 210YesYesYes 4aPhotos-NegligibleYesNoNo 4bDeimos-NegligibleYesNoNoTable 1: Characteristics of the planets similar to the Earth.


15th International Congress of Speleology Lava Caves 635 2009 ICS Proceedings a lava ow. In case of the existence of an atmosphere the heat exchange is due to both convection and radiation, otherwise the release of heat is by irradiation only. On Mars and Venus the atmosphere is composed by CO2 while on the Moon and Mercury the atmosphere is absent. In Table 2 the heat loss of lava in dierent planets is reported. e convection was evaluated for two wind velocities (w) of 10 and 100 ms-1 respectively and radiation for lava temperature (T) at 1000 C. According these data a rather rough conclusion could be that the development of lava tubes in the dierent planets should not be much dierent from the Earth with the exception of Venus on account of its dense atmosphere. But, development of lava tubes depends on many other factors as the chemical composition of lava, the slope, the gravity, etc. An exhaustive evaluation of there factors is complicated and outside the scope of this paper; it will be further developed in the nal version of Badino (2008). A global evaluation of the phenomenon can be obtained by a numerical calculation only since the radiation is proportional to T4 while the convection is directly proportional to T. erefore the lava ux must be analyzed along its whole pathway. If convection would prevail the lava tubes on Mars would be dramatically longer than on the Earth. e surface of other planets is presently well known thanks to the space missions. eir surfaces have many craters due to the impact of meteorites. is fact is due to the absence of atmospheric eects which, on the Earth, acts as a shield for most of the meteorites and smoothes rather rapidly the craters formed on our planet. With the exclusion of Mars satellites, which are nothing more than a couple of big boulders moving in the space, relevant volcanic phenomena seem to be widespread. Basalt rocks and detritus probably form planets surface.3. MercuryMercury, the least massive among the inner planets, shows many craters, which may be attributed to meteoric impacts. Among the most relevant examples, the so-called Caloris basin, a huge circular sea is the result of a catastrophic impact, which modied a wide part of the planet surface. It must be emphasized here that such events may produce a local fusion (or re-fusion) of the crust, so that meteoric impacts may be the origin of secondary eects of a nonendogen volcanism. If evident volcanic craters are absent, there are wide covers of the original soil, which are generally supposed to be due to lava layers. In other words, an important eusive volcanism, with the lava, emerging from crust fractures, could have covered large areas of the planet surface. In Figure 1, a region of Mercurys surface is reported, where many indications support this hypothesis: a. e area in the top, where there are practically no craters, can be assumed to have been successively covered by lava ow, since it cannot be assumed that this area was not originally hit by meteorites as well as the rest of the planet. b. Typical ridges, laying north-south, are evident on such a cover, with the same appearance of the ow of a viscous uid. c. e existence of many craters lled and sometimes nearly smoothed by a uid. Sometimes the ridges rise above overlap the smoothed craters supporting the hypothesis that the covering layer is Planet Convection w=10 ms-1Convection w=100 ms-1IrradiationTotal Earth (ground level) 7 30 100107-130 Moon 0 0 100 100 Mercury 0 0 100 100 Venus (ground level) 400 1600 80480-1680 Mars (at 20 km) 1 4 100101-104Table 2: Heat loss of lava [kW m-2]. Figure 1: Lava waves on Mercury. Notice the crater coered in the upper le corner.


Lava Caves 636 2009 ICS P roceedings 15th International Congress of Speleology successive to the formation of the crater. If the prevailing conditions of the planet are these, then tectonic caves and/ or lava tubes could exist. e presence of widespread fractures of the planet crust is evident in Figure 2 where a computerized photomontage of the area to the east of the Caloris basin is reported. A series of deep rills is seen on the le of the picture, as an evident consequence of the impact that originated the same basin. It is interesting to observe that the fracture, in the middle of the picture up to the basis of the typical group of four craters (Mickey Mouse), shows in the center top a series of interruptions suggesting the presence of tectonic caves. A support to this hypothesis is found in Figure 3 with a particular rill crossing the lower edge of the central crater of Mickey Mouse. In the lower le corner of the picture, a channel between a substantially homogeneous plain and the center of the crater can be seen. It is interesting to observe that here the rill is formed by the union of a series of elliptical hollows as it happens when a lava tube collapses. is is an important criterion to distinguish surface collapses from a succession of micro-craters due to secondary ejecta from an impact crater. Research of lava tubes is carried out on the same principles, on the basis as observations of surfaces with partially or totally collapsed lava tubes. Sometimes the winding of a series of micro-craters may give a rather precise indication of the presence of a lava tube below (Greeley, 1977: Fig. 3). It is dicult to obtain a clear indication of the existence of volcanic caves by observing the surface of Mercury. Frequently the presence of some indication never reaches the level of a clear proof. Also in Figure 3, on the right of the channel mentioned above, a series of collapses can be seen as the result of a lava tube connected to the central crater. But perhaps it could be also a branch of the system of fractures already observed. e situation is much more uncertain because on Mercury there is a strict association between some craters and eusive phenomena. Figure 4 show a typical case: the craters in the picture appear to be lled by a homogeneous uid but, nevertheless, the nearby soil and heights appear to be substantially untouched. e area appears to have been subject to lava ows, which locally lled hollows. In the center of the picture, two craters seam to be connected by a channel; as well lava channels could be the rills starting from the western ridge of the central crater and from the huge crater at the north-western border of the picture. At the moment the existence of lava tube cannot be denitively proven. Perhaps, in the future, more detailed pictures could show important details. It is also possible that the nature itself of the volcanic phenomena on Mercury precludes to existence (or the observation) of lava tubes in this planet. Figure 2: Computerized photo mosaic of the E edge of basin Caloris. A series of deep rilles are seen on the le of the picture. Figure 3: Particular of a group of craters (Mickey Mouse). In the lower le corner, channels can seem to be a series of elliptical hollows as it happens when a lava tube collapses. Figure 4 : In the middle, two craters connected by a channel. Other channels, sometimes coered, are seen at the top le.


15th International Congress of Speleology Lava Caves 637 2009 ICS Proceedings 4. Marse study of the surface structures on Mars is somewhat more dicult than the previous ones. In fact, in addition to the tectonic and volcanic structures, uvial molding processes are evident supporting the hypothesis that liquid water exists on the surface of the planet (Leovy, 1977; Jakowsky & Haberle, 1992). Figure 5 shows a typical and striking example, where a uvial mold on a surface with many craters can be observed. Note the strange domeshaped structure inside the principal craters. Tectonic structures are frequent both as a consequence of strong meteoric impacts and as the result of a real planetary tectonics. e most striking example of such tectonics is given by the Vallis Marineris, a more than 1000 km long deep canyon in the Mars surface. Beyond these tectonic structures Mars shows the existence of strong volcanism with the largest volcanoes in the solar system, known up to now. e most imposing volcano (Olympus Mons) is more than 600 km wide and has a height from the planet surface of 26 km (Greeley, 1977: Fig. 17). e presence of wide areas without craters suggest the existence of a covering layer, which, in the case of Mars, could also be due to a water ow or to sand storms known to occur on this planet. Close to volcanoes there are apparent lava tubes and volcanic caves. In Figure 6, the Arsia Mons, a volcano 19 km high and wider than 100 km is shown. In addition to a series of fractures in the upper right corner of the picture, three lava tubes appear to originate at the volcanic shield of the crater. Upstream, the central rill can be observed. It could be explained as the partial collapse of the lava tube that feeds the terminal channel. A more detailed study of the surface morphology suggests the existence of caves more exactly termed parakarstic. e landslides along the slopes of the Vallis Marineris (Fig. 7) have been interpreted as due to permafrost, i.e. to the existence of underground frozen deposits that under proper conditions (e.g., volcanic heat) can change into a liquid phase and result in a uid ow under the thrust of the overlying layers (Coradini & Flamini, 1979). Good evidence of the existence of these phenomena is reported in Figure 8 where a uid mass originating from a collapsed structure has covered the plain on the le of the picture. e fusion of permafrost in the underground of Mars could result in a pseudokarstic mechanism able to produce extended caves. Up to now, only lava and tectonic caves (i.e., pseudokarst) had been identied or, at the most, some parakarst phenomena. Grith and Schock (1995) proposed a geochemical model to explain the formation of hydrothermal carbonates on Mars from the alteration of Figure 5: Photo mosaic of Regio Crisis om 1,600 km; the structures around the craters are due to the ow of a uid (water?). Figure 6: One of the Mars highest olcanoes, the Arsia Mons (diamater 100 km, and height 19 km). Notice the channels om the crater towards the center of the picture. Figure 7: Photo mosaic of the area to W of Vallis Marineris om 4300 km. Notice the landslides along the slope or the valley and the channel due to the collapse of caves.


Lava Caves 638 2009 ICS P roceedings 15th International Congress of Speleology basalt. is model agrees perfectly with the observations when applied to Iceland rocks. In Mars conditions, the alteration of basalt in presence of water and CO2, e.g., at a temperature of 250 C, could produce 7% carbonate, which could be found as vein-lling materials and mineral replacements. According to the above mentioned authors, if the hydrothermal processes developed, in areas where the water table intersect the surface, warm, CO2-laden uids should have degassed to produce locally extensive travertine. In this case, classic karst processes could have produced caves. Also, if the existence of hydrothermal systems on Mars seems to be assessed (Farmer, 1996), up to now no carbonates have been detected. Recently, Halevy, Schrag and Zuber (2007) speculated on the possibility of an evolving Mars environment base on the sulfur cycle instead of the carbon cycle, as supported by the abundance of sulfur minerals on the Martian surface. In this case, caves could have developed in gypsum instead of limestone. In Figure 9, the gullies eroded into the wall of meteor impact crater covered by soil (loess?) in Noachis Terra could start from the entrance of a cave. An image taken by NASAs Mars Reconnaissance Orbiter on 8 August 2007 shows the entrance of a pothole with a wall on this dark feature, suggesting it is a pit at least 78 meters deep (Shiga, 2007) (Fig. 10). e evidence of a surface layer with a round void (about 150 m in diameter) leading to an underground passage suggests the possibility of a skylight produced by the collapse of the roof of a lava tube or a wide passage. In any case, this feature is not an impact or a volcanic crater since there are no ejecta around it. Karst phenomena strictu sensu are probably absent due to the lack of limestone. Based on these assumptions, the above noted morphological characteristics of the planets surface might hold speleological interest. Mars satellites are not taken into account due to the absence of any mechanism of development. Along the Vallis Marineris there is some evidence of features that could be interpreted as long lava tubes (Fig. 11). In this case they could be the longest lava tubes in the solar system. Also without a direct exploration, relevant indications support such a hypothesis. On the highlands, close to the slopes of the Vallis Marineris, and in addition to the landslides quoted above, there are typical pseudo-uvial Figure 8: e landslides at the le of the valley originated the uid mass that coered the plain on the bottom of the picture. Figure 9: Gullies eroded into the wall of meteor impact crater in Noachis Terra could start om the entrance of a cave. (NASA/JPL/Malin Space Science Systems). Figure 10: A pit at least 78 meters deep e evidence of a surface layer with a round oid (about 150 m in diameter) leading to an underground passage suggests the possibility of a skylight produced by the collapse of the roof of a lava tube or a wide passage (Image: NASA/JPL/University of Arizona).


15th International Congress of Speleology Lava Caves 639 2009 ICS Proceedings valleys, which at rst seem to be unjustied on account of the absence of any trace of the corresponding rivers. If the existence of permafrost in the Martian subsoil is accepted, then such valleys can be attributed to the action of underground drainage channels along fusion lines of permafrost. Above these channels and close to the surface, collapse structures would have the shapes here observed (Fig. 12). In such a case, the Marstian subsoil could host pseudokarstic phenomena, although it is not possible to dene their real importance.ReferencesBadino G. (2008) First approach to the lava tubes thermal exchanges in extraterrestrial environment (in progress) Coradini M. and Flamini E. (1979) A thermodynamical study of the Martian permafrost. Journal of G eophysical RResearch, 84, pp 8115. Greeley R., (1977) Lava Tubes in Other Planets. AA t ti Seminario Grotte Laviche, Catania 1975, Gruppo Grotte Catania CAIEtna, pp 181. Grith L.L. & Shock E.L. (1995) A geochemical model for the formation of hydrothermal carbonates on Mars. N N a ture, 377, p. 406-408. Jakowsky B.M. and R.M. Haberle (1992) e Seasonal Behavior of Water on Mars. In Kieer, H.H. et al., (Eds.) Mars, University of Arizona Press, Tuscon & London, p. 969-1016. Leovy L.J. (1977) Latmosfera su Marte. Le Scienze, 111, pp 32. Figure 11: Another view of the Vallis Marineris; various structures, probably lava tubes, run parallel to the valley. Image om NASA (courtesy of Windows to the Universe, http://www.windows.ucar.edu). Figure 12: Caldera collapse pit at the summit of one of the large olcanoes on the arsis plateau. Image om NASA (Courtesy of Windows to the Universe, http://www.windows.ucar.edu).


Lava Caves 640 2009 ICS P roceedings 15th International Congress of Speleology SOME ASPECTS OF SPELEOGENESIS IN EXTRATERRESTRIAL ENVIRONMENT: 2MOON AND VENUSVITTOR TTOR IO O CA A ST T ELLAN AN I1, ARR ARR IGO O A A CIGNA NA2 1Lab. AA strosica Spaziale, Frascati (RRoma) Italy2Fraz. TT uo, Str. Bottino 2, I-14023 Cocconato (ATAT), Italy Some structures observed on the surface of the Moon are interpreted as resulting from speleogenetic processes. e volcanic structures are particularly prominent speleological features. Both lava tubes and tectonic caves have been identied. In particular, the lower gravity and the lack of an atmosphere relatively dense (and therefore able to drain negligible amounts of heat in short time intervals) allowed the formations of extremely long channels in these extraterrestrial environments. Observations of Venus surface, showed the existence of volcanic activity on this planet, and there are some indications about the possibility of the existence of volcanic caves as lava tubes. Nevertheless it must be taken into account the characteristics of this planet: if the high density of its atmosphere (9*105 millibar) would facilitate an exchange of heat, the high temperature (748 K) would reduce such an exchange. erefore, the lava tubes could perhaps be larger than those on Earth.1. MoonOn the Moon the volcanic structures are quite evident, as cupolas emerging above the soil surface or real volcanoes. Tectonic phenomena are also observed in the vicinity and inside some large impact craters. Figure 1 shows Maris Humorum, with a reported diameter of about 350 km. On the right side there are some ridges similar to those observed on Mercury, and on the le side there is a series of probable tectonic fractures. It is interesting to observe that along the southwest edge of the principal crater there are some relics of craters lled by lava. en, it seems that aer the original formation of craters, further lava ows perhaps went o some fractures produced by the original event. On the west edge, in the inner zone, a series of rills with the very characteristics of lava tubes and covering relics of craters, support the hypothesis of a succession of lava ows. Somewhat above, the disappearance of half a crater, suggests that a lava ow from southeast stopped at a tectonic fracture. Examples of clear lava tubes coming from wide craters are evident in Figures 2 and 3. Figure 3 show a general view of this area: from the structure reported above, a rill with many covers starts and extends till the near by sea. e morphology of this rill leaves few doubts about its origin as a lava tube. e Circus of Plato is a crater lled by a sea about 100 km of diameter. e material inside the sea is also outside, without clear surface connections. For this reason already in 1970 an underground connection between the two areas was considered likely (Kosofsky & El Baz, 1970). Below, to the le, a large rill (lava tube?) seems to start from the slope of the principal crater. Above, to the right, there are a small crater and a stretched hollow that have evaded a clear interpretation. Below, to the right, a peculiar formation (Vallis Alpinis), whose origin is clearly connected to a uid ow, is still under discussion. In Figure 4, a view of the Vallis Alpinis (about 150 km long and 8 km wide) is reported with its evident connection to the higher Mare Imbrium. It could Figure 1: Mare Humorum (diameter about 350 km) in the middle; right: crater Gassendi, le: crater Vitello (23S, 38W) on the right some ridges are visible, on the le some rilles. e width of each path of the photo mosaic is 86 km.


15th International Congress of Speleology Lava Caves 641 2009 ICS Proceedings be assumed that the valley acted to channel lava ow along a tectonic fracture to produce a central lava tube. On the right side of the picture and inside the same Mare Imbrium, there are other rills, which can be interpreted as lava tubes. Typical characteristics of the lunar volcanism is the frequent origin of lava tubes from hollows rather than from volcanoes. It is possible that this feature is a general phenomenon based on previous observations. For instance, one of the most prominent lunar rill (Rima Hadlet) appears to start from a deep hollow (Fig. 5). Another quite similar structure is seen in Figure 6. In this case, instead of the rill, there is a series of collapse structures attributable to a underlying cavity. It is very probable that it consists in lava tubes (notice the branch in the lower right part of the picture) due to a mechanism of lava eusion quite similar to the one observed on Mercury. An interesting and typical example is given by the so called Schroter Valley, a channel originating from a deep hollow at the base of a hill (Cobras Head) about 1500 m high. e crater Aristarcus (about 40 km of diameter) is on the le and the crater Erodotus is on the right. e valley (Fig. 7) appears to be a long winding hollow with a at oor, about 1,300 m deep. e whole area was evidently covered by a huge lava ow as supported by the presence of the relic of a crater in the lower le part and by the many rills. An enlarged view of the upper part of the valley (Fig. 8) Figure 2: Crater Plato (diameter about 100 km). e material inside the sea is is found also outside, without clear surface connections. Perhaps there was an underground source of this material inside the crater or an underground connection connected the two areas. Below, to the right, a large rille (lava tube?). e width of each path of the photo mosaic is 12 km. Figure 3: A long channel, sometimes coered, connects crater Plato (bottom lefy) to Mare Imbrium (centre). Vallis Alpinis is in the bottom right. e width of each path of the photo mosaic is 1 km. Figure 4: Mare Imbrium (le) and Vallis Alpinis (about 150 km long and 8 km wide) whose origin is much discussed. A channel with many meanders is along the valley; other channels are seen in the Mare and on the bottom. Figure 5: One of the most relevant Lunar channels (Rima Hadley) seems t start om a deep hollow. e width of each path of the photo mosaic is 4 km.


Lava Caves 642 2009 ICS P roceedings 15th International Congress of Speleology shows that in its at bottom there is a deep winding channel with the typical size of the rills to be noticed in the same zone (see the upper right side). e interpretation of this structure is not easy. e neat meanders of the inner channel seem to be due to an erosion by an absolutely non-viscous uid. Such erosion may be connected to the genesis of the whole structure. It is also rather uncertain if the Cobras Head is the source vent of such a uid. Some characteristics could suggest an opposite situation, so that the hollow in the upper part of the valley could be interpreted as a real lava pothole. In any case, it is evident that on the Moon there is an extended number of channels and lava tubes as well as tectonic caves to be still studied and explained in detail (Anelli, 1973).2. Venuse rst radar images of Venus (Saunders & Malin 1976) suggested the existence of craters. Successively detailed images conrmed this hypothesis. Because Magellans radar viewed the Venusian surface from varying angles, 3dimensional images of the planets terrain were also possible in the early 1990s. e Venera missions analyzed Venus atmosphere and found that it is made up of about 96% carbon dioxide, with little oxygen. e 7th through the 14th Venera missions all successfully landed on Venus, each spacecra spending a longer time on its surface than the previous one. Venera 10 sent back the rst black and white photographs of Venus terrain, while Venera 13 sent back the rst color photos. e Venera missions also measured a surface temperature of 475 C, detected lightning, sampled the soil where it landed, and found a type of basalt that is common on Earth In Figure 9 an image of a crater with multiple rings above the crater oor shows some channels in the upper le corner. In Figure 10, a very detailed image of the Lada Regio has many channels, particularly in the middle center at area, that could be interpreted as lava tubes. Nevertheless it must be taken into account the characteristics of this planet: if the high density of its atmosphere (9*105 millibar) would facilitate an exchange of large amounts pf heat, the high temperature (748 K) would reduce such an exchange. erefore, lava tubes could perhaps be larger than on the Earth. Figure 6: Series of hollows towards the hedge of the Oceanus Procellarum; they have the appearance of collapse structures attributable to a underlying cavity. e width of each path of the photo mosaic is 5.4 km. Figure 7: Crater Aristarcus (middle) (diameter about 40 km) and crater Erodotus to the right (diameter about 35 km) with a at oor. e Valley of Schroeter (250 km long) starts om the Cobras Head. e width of each path of the photo mosaic is 11 km. Figure 8: Particular of the Schroeter Valley about 1,300 m deep. e winding channel sometimes is coered by the slope of the valley. e width of each path of the photo mosaic is 4.3 km.


15th International Congress of Speleology Lava Caves 643 2009 ICS Proceedings 3. Conclusionse framework here reported shows that earth-like speleogenesis may develop at the surface of many planets and satellites. A comparative study of these phenomena can explain the role played by the environmental conditions where the phenomena develop. e relevant parameters for this comparison are reported by Castellani and Cigna (these proceedings, Table 1). Many preliminary assumptions and working hypotheses may be easily listed. As an example, the large dierence between Mercury and Mars must be emphasised notwithstanding the same gravity at their surface. en it must be stressed the wide diusion of volcanic phenomena also with dierent aspects. e genesis of lava channels and tubes is characterised by various factors as it was reported by Wood (1977): Flow rate Flow speed: Viscosity Slope Gravity Cooling velocity Since some of these factors may vary largely in the dierent planets here considered, also the sizes of the structures described here may be very dierent from those observed on the Earth. In particular, the lower gravity and the lack of a relatively dense atmosphere (and therefore able to drain signicant amounts of heat in short time intervals) allowed the formations of extremely long channels in the extraterrestrial environments here examined (Greely, 1976, 1977). As it is not possible now to improve the knowledge of the phenomena by a direct observation, it is necessary to extend the study by developing models of the structures here described in function of known and measurable quantities. Acknowledgments e kind assistance and instrumental cooperation in the preparation of the original version of the paper is gratefully acknowledged to prof. M. Fulchignoni, dr. M. Coradini and dr. M. Poscolieri. e contributions of Dr. V. Caloi who provided references, and prof. G. Badino who supplied a text in progress are here acknowledgedReferencesAnelli F. (1973) Il geografo sulla luna. Realt Nuova, 4, Milano pp 1 Greely R., (1976) Terrestrial Analog to Lunar Sinuous Rilles. Proceedings International Symposium V ulcanospeleology and Extraterrestrial AA pplication. White Salmon, Washington 1972, Western Speleological Survey National Speleological Society, Seattle, p 85. Greely R., (1977) Lava Tubes in Other Planets. AA t ti Seminario Grotte Laviche, Catania 24-30 Agosto 1975, Gruppo Grotte Catania CAI Etna, pp 181. Figure 9: A crater on Venus with multiple rings that rise aboe the crater oor and some channels in the upper le corner. Image om NASA (Courtesy of Windows to the Universe, http://www.windows.ucar.edu). Figure 10: A very detailed image of Lada Regio with many channels, particularly in the middle centre at area, which could be interpreted as lava tubes. Image om NASA (Courtesy of Windows to the Universe, http://www.windows.ucar.edu).


Lava Caves 644 2009 ICS P roceedings 15th International Congress of Speleology Harter J.W. III (1976) Mathematical Analysis of Some Lava Tube Mechanics. Proceedings International Symposium Vulcanospeleology and Extraterrestrial A A p plication. White Salmon, Washington 1972, Western Speleological Survey National Speleological Society, Seattle, p. 70-73. Kosofsky L.J. and F. El Baz (1970) e Moon as Viewed by L unar OOrbiter. NASA, Washington, DC, 152 pp. Saunders R.S. and M.C. Malin (1976) Venus: Geological Analysis of Radar Images. Proceedings International C olloquium Planetary Geology, RRome 1975; Geologica Romana, 15, pp 507 Wood C. (1977) Factors Contributing to the Genesis of Caves in Lava. AA t ti Seminario Grotte Laviche Catania 24 Agosto 1975, Gruppo Grotte Catania CAI Etna, pp 181


15th International Congress of Speleology Lava Caves 645 2009 ICS Proceedings HIGH PRECISION U/TH DATING OF RECENT LAVA FLOWS USING SYNGENETIC NONSILICATE CAVE MINERALSJULIAN AN D D ILLON ON VICTORTOR PO O LYA A K, YEMAN AN E A A SMERO RO MD Department of Earth and Planetary Sciences, University of NN ew Mexico, AA lbuquerque, NN M 87131 Precise radiometric dating of recent lava ows has been a problem due to the inability to resolve inherited geochemical isotopic signatures complicated by complex magma chamber residence histories. As a result, direct U/ or K/Ar measurements of young basalts oen yield ages with large uncertainties. Other techniques such as cosmogenic exposure ages require physical assumptions of erosion rates and production rates that are not always constrained, but are oen the best or only techniques available. Syngenetic minerals, such as gypsum, occur in many lava tubes and may potentially provide reliable ages for young basalt ows. Our rst results come from Big Skylight and Four Windows caves, which are located in the Bandera basalt ow in the Zuni-Bandera volcanic eld, New Mexico. In these caves thick gypsum crust (~0.5 1 cm) occur on primary tube surfaces. e Bandera lava ow has been well-dated using 14C from charcoal underneath the ow (12.5 Ka) as well as 36Cl and 3He (11.2 Ka) Initial results on the gypsum crusts show that U and Sr isotopic signatures are similar to those of the basalt. 234U values range from -6 to +30 which are moderately more positive than their basalt hosts, which is always 0 for unaltered basalts. e 87Sr/86Sr isotopic ratio for the gypsum crust average 0.7039149 + 0.0003, which closely mirror published basalt value of 87Sr/86Sr = 0.703667 and which are lower than the 87Sr/86Sr ratio of 0.7075.071226 for local soil. Slight dierences of 234U and 87Sr/86Sr between the gypsum crust and basalt indicate some limited contribution from a non-basaltic source. e oldest gypsum U-series ages range from 9 Ka, but some samples within these crusts also yield ages between 5-7 ka. Our initial results indicate that U-series ages of these gypsum crusts could potentially provide a new method for dating young basalt ows, if we are able to delineate pristine samples from those that experienced cryptic alteration. 1. IntroductionRadiometric dating of uaternary (2.0 Ma to present) lava ows is important for cosmogenic exposure age techniques, which are used to constrain erosion rates, as well as neotectonics, hazard assessment, evolution of magma chambers, and magma recurrence intervals. K/Ar (Ar/Ar) method, cosmogenic nuclide dating of surfaces, and in cases where eruption related charcoal can be found the 14C method have been used to date volcanic rocks. e K/Ar method suers from problems due to the low abundance of K in the basalts, as well as ideal conditions for the entrapment and isolation of radiogenic argon from atmospheric argon, particularly in basalts that are aphinitic(Sims et al, 2007). Cosmogenic nuclides have been successfully applied to many young basalts where erosion rates and cosmogenic production rates have been independently determined, and exposure (or lack thereof) to cosmogenic rays is well known. For basalts that overly charcoal deposits, a lower limit on eruption age can be determined from 14C measurements, but only up to the upper limit of 14C dating (50 ka). Direct uranium series dating of young lava is problematic due to the unfavorable U/ ratios in volcanic rocks. Daughter nuclide disequilibria of 238U and 230 have been successfully used to date carbonate rocks up to 500,000 years (Cheng et al., 2000), but lava ows require isochron techniques because of their high /U ratios and have large errors (i.e., %, Sims et al, 2007). Moreover such the ages do not dierentiate between timing of magma chamber residence, transfer to the surface and ow emplacement. Syngenetic non-silicate minerals potentially can overcome these limits to the extent that many minerals, such as gypsum have favorable U/ ratios and they do form at the time of eruption. Primary igneous sulfates have been directly observed as anhydrite in a number of localities and settings (Luhr, 2008). To determine if our samples are pristine and from


Lava Caves 646 2009 ICS P roceedings 15th International Congress of Speleology a magmatic source, we focus on U, Sr and S isotopic measurements on the well-dated Bandera ow from the Zuni-Bandera volcanic eld in northwestern New Mexico, USA. Data on U and Sr from the Bandera basalt are known (Asmerom, 1999) and should be reected in eruption products. e S isotopic composition should near 0 for a high temperature source (Sakai et al, 1982). e Bandera ow has 3He, 36Cl, 14C, as well as an aberrantly older K/Ar ages. 2. Sample Descriptione Bandera ow is branched and segmented, with all samples collected from the Big Tubes recreation area in El Malpais National Monument, with a majority of samples collected and analyzed from the largest cave, Big Skylight cave (Fig. 1). Other samples were collected from nearby Four Windows Cave and Giant Ice Cave. Gypsum crusts were not observed on any broken surfaces on the cave walls or on breakdown blocks. e crusts were located near entrances and skylights, where preservation of the gypsum is presumed to be highest because of low humidity as opposed to deeper into the cave systems where higher relative humidity might form condensates that remove these easily dissolvable crusts. Samples were collected in-place, as well as broken pieces along lava shelves. When multiple lava benches were present, samples were collected at multiple levels. Gypsum crusts were also present on the primary ceilings of the caves, but not accessible. e gypsum crusts were up to 1 cm thick in places, and as thin as 5 mm. Crusts were locally continuous, and showed a rosette texture of crystal terminations on surfaces. Some samples, aer collection, showed voids between the basalt surface and gypsum surface. A sample of soil was collected from directly above the entrance of Big Skylight Cave for 87Sr/86Sr comparisons.3. MethodsGypsum crust samples were lightly powdered and dissolved in 1N HNO3. is process took 2448 hours. e gypsum solutions were spiked with a mixed 233U-236U-229 spike for the Udating and 84Sr spike for Sr concentration determination. U and were coprecipitated from the spiked solutions for Useries analyses. All separates were analyzed using a ermo-Neptune multicollector ICPMS. Figure 1: (A) Big Skylight Cave entrance. (B) Lava Grooes and sheles (ow lines) illustrating well-preserved passage walls in Big Skylight Cave. e white arrow points to gypsum crust (site 2). (C,D) Gypsum crust on cave walls in Big Skylight Cave (site 1).


15th International Congress of Speleology Lava Caves 647 2009 ICS Proceedings 4. ResultsOur apparent ages from cave speleothems show a bimodal distribution, with an upper group of 9-10 ka, and a lower group of 5-7 ka (Table 1). For the initial Big Skylight samples with sucient thickness, top and bottom age proles of the crusts were also analyzed yielding the same ages within error. 234U values range between +9 and +30. 87Sr/86Sr isotope ratios are tightly clustered between 0.703659 0.7039974 with Sr concentrations ranging from 300 to 1400 ppm, with the exception of sample FW-1, which was has anomalously low Sr abundance. Our soil 87Sr/86Sr value is 0.712255 .00076% (1). Cosmogenic ages (Table 2) are from Dunbar and Phillips (2004), and McLaughlin (1994). 40Ar/39Ar ages are from McIntosh (1994). 5. Discussione age of the Bandera basalt ow is established by multiple methods to be 9-12 ka (Dunbar and Phillips (2004). Given that our gypsum crusts are speleogenetic (having formed as a product of the caves origin), then they should yield U-series ages within the same age range as the other methods. Our results show a bimodal distribution of ages with the younger being about half the reported age and the older representing an age close to the expected age. e bimodal distribution of Table 1: Uranium series data: lava tube gypsum crusts.


Lava Caves 648 2009 ICS P roceedings 15th International Congress of Speleology ages forces the question of are these indeed speleogenetic or are they speleothemic crusts? Field observations and isotopic analyses provide additional insights. Speleothems in the lava tube caves of El Malpais National Monument are opalline and calcitic coralloids, crusts, and moonmilk. With the exception of some occurrences of calcite moonmilk, all speleothems in these caves seem to be signicantly thinner and smaller than the gypsum crusts. e gypsum crusts are relatively thick in comparison and seem to have no obvious secondary origin. For instance, coralloids form readily from condensates and sprays which are observed in these caves today. e calcite moonmilk deposits are also located in areas where condensates occur. But the gypsum crusts seem to be located in areas where secondary deposits are absent or rare. ey are also located on original tube surfaces, and not found on breakdown. e gypsum crusts, if formed by nal volcanic activity related to the lava tube, should have inherited isotopic signatures similar to the basalt/volcanic gases. 234U values of the crust are low (9-30 ), and while higher than the basalt (0 ), they are still reasonably low enough to have formed from the volcanic gases associated with the origin of the lava tube. e 234U values for coralloids and moonmilk are respectably higher (~133-226 LaPointe personal communication). 87Sr/86Sr isotope measurements for the gypsum crust (0.703659 0.7039974) also match the published Bandera basalt values of 0.703667 (Asmerom, 1999), and dier signicantly from the soil (0.712255, our value for soil above Big Skylight Cave; 0.710660.7122, Van der Hoven and uade (2002)). e U isotope ratios and Sr isotope ratios of the gypsum and Bandera basalt are similar enough to explain their origin from a magmatic source rather than from surrounding pedogenic soils. Slight dierences may be due to inheritance of groundwater signatures of the surrounding basalt, but given the young age of the lava ow and the thickness of the sample sizes as well as the consistency in ages from the upper and lower samples, we nd it unlikely that secondary deposition would be possible to deposit a suciently thick gypsum crust in the allotted time, especially since top and bottom subsamples of the crusts are the same age. Furthermore, no gypsum crusts have been observed on broken lava ceiling surfaces, suggesting one depositional event. e variation in ages is attributed to secondary alteration, which gypsum is highly susceptible to, or to later anhydrite-to-gypsum transition. Current research is on identication of alteration signatures, with emphasis on the stable isotopes of S. Other possible explanations for the bimodal cluster of ages include a subsequent pulse of volcanic activity around 5 ka, consistent with the McCartys ow in the Zuni-Bandera volcanic eld. However, the McCartys ow uses a dierent vent system entirely, and the Bandera ow tube system is heavily segmented with no indication of subsequent pulses of volcanism. Also, preservation of two stages of gypsum crusts is highly unlikely. Our eld observations and geochemical results show that the gypsum is not a secondary deposit (speleothem) and support that it is speleogenetic. We interpret our highest ages to represent the age of the gypsum crust emplacement and timing of the end of Bandera volcanic activity. Table 2. Bandera ow ages.


15th International Congress of Speleology Lava Caves 649 2009 ICS Proceedings 6. Conclusione recent recognition of deposition of primary igneous sulfates has provided an opportunity to use U/ disequilibrium techniques as a way to determine meaningful ages of quaternary volcanism. Isotopic signatures of the gypsum crust samples from the Bandera ow indicate a magmatic source, while apparent ages agree with previously published cosmogenic exposure ages as well as underlying charcoal 14C. Due to the unstable nature of gypsum or early anhydrite-to-gypsum transitions, analyses of multiple samples increase reliability of apparent ages. Application of gypsum crust dating may be applied to other lava tube systems once the system is better understood. AcknowledgmentsFunding for this research was provided by NSF grant EAR-0841426. We like to thank the El Malpais National Monument for allowing us access and samples from the Bandera ow. ReferencesASMEROM, YEMANE (1999) -U fractionation and Mantle Structure. Earth and Planetary Science Letters 166, 163. CHENG, H., R.L. EDWARDS, J. HOFF, C.D. GALLUP, D.A. RICHARDS, Y. ASMEROM, (2000) e half-lives of uranium-234 and thorium-230. Chemical Geology 169, 17. D UNBAR, NEILA w W ., FRED m M Phillips HILLIPS c C osmogenic 36Cl ages of lava ows in the ZuniBandera volcanic eld, north-central New Mexico, U.S.A. NN ew M exico Bureau of Geology and Mineral R Resources Bulletin. 160. 309. LAUGHLIN, A.W., J. POTHS, H.A. HEALEY, S. RENEAU, and G. WOLDEGABRIEL, (1994) Dating of uaternary basalts using the cosmogenic 3He and 14C methods with implications for excess 40Ar. Geology 22, 135. Luhr U HR J. (2008) Primary igneous anhydrite: Progress since its recognition in the 1982 El Chicon trachyandesite. Journal of Volcanology and G eothermal RResearch. 175, 394. MCINTOSH, W.C. (1994) 40Ar/39Ar geochronology of late Miocene to Pleistocene basalts of the ZuniBandera, Red Hill-uemado, and Potrillo volcanic elds, New Mexico(abs). NN ew M exico Geology. 16, 60. Sakai A KAI H., T.J. Casadevall ASADEVALL J.G. Moore OORE (1982) Chemistry and Isotope ratios of sulfur in basalts and volcanic gases at Kilauea Volcano, Hawaii. G eochimica et Cosmochimic AA cta. 46, 729. Sims I MS K. W., Ackert CKERT R.P. Jr R ,, Ramos AMOS F. C., Sohn OHN R. A., Murrel URREL M. T., De E Paolo AOLO D. J. 2007) Determining eruption ages and erosion rates of uaternary basaltic volcanism from combined Useries disequilibria and cosmogenic exposure ages. Geology. 35, 471. Van A N der DER Hoven OVEN S.J., and J. uade UADE (2002) Tracing spatial and temporal variations in the sources of calcium in pedogenic carbonates in a semiarid environment. Geoderma. 108, 259.


Lava Caves 650 2009 ICS P roceedings 15th International Congress of Speleology MEDICAL AND GOVERNMENTAL CONSIDERATIONS OF CO2 AND O2 IN VOLCANIC CAVESWILLIA A M R R HA A LLIDA DA Y, M.D D ., F.C.C.P. Past Medical DD irector, Washington State DDepartment of Labor and Industries, 6530 C ornwall Court, NN ashville, TNTN 37205 USA A wrhbna@bellsouth.net Occurrences of hypercarbia in carbonate caves are well comprehended in most of the global speleological community. In some volcanic caves, a dierent scenario also must be considered: simple addition of magmatic CO2 to ambient air. In this scenario, hypoxia plays a minimal role, if any. In others, simple percent for percent replacement of O2 by CO2 may occur, just as in carbonate caves. In addition, other toxic gases of magmatic origin may exist in volcanic caves, but their toxic manifestations are easily detected clinically. Basic investigations of separate eects of various levels of hypoxia and of hypercarbia were conducted so lo ng ago that they  cannot  readily be retrieved electronically. Medical textbooks of the mid-20th Century summarized such studies in volunteer healthy adult males. At approximately 4% atmospheric CO2. breathing is alarmingly increased, with the limit of tolerance in highly motivated test subjects between 7 and 9%. Eventual unconsciousness occurs at 10%. No such early warning sign exists for hypoxia; even reduction of O2 to 10% (which is lethal eventually) causes only a slight increase in breathing. Such textbook summaries, however, fail to emphasize the wide individual human variations in response to hypoxia and to hypercarbia clearly specied in the original reports. Congress assigned a politically charming but technically impossible task to the Occupational Safety and Health Administration (OSHA): setting standards which would provide safe work places for all American workers young and old alike, and both healthy and terminally ill cigarette smokers. OSHA chose t o set standards  that would  protect most workers in most workplaces, with selective enforcement. is allows volcanologists to run great risks on volcanoes, for example.  Inevitably,  its standard for CO2 responded to anxiety produced by breathing a mere 4% CO2 and its impact on workers, their union shop stewards, the general public and even concerned employers. A wide safety margin was included to protect aged workers and those with breathing impairments. is likely included some with a nancial stake in maximizing their symptoms. Further, OSHA determinations were developed to apply to closed work spaces much like the scenario seen in carbonate caves, where hypercarbia gives the alarm but hypoxia kills. us, OSHA standards probably bear little resemblance to medical principles in healthy, highly motivated volunteers in volcanic caves. While the promulgated gures vary slightly from country to country, this pattern exists broadly across developed, socially conscious nations. is issue surfaced in 1990 when Howarth and Stone found a notable ecosystem in a cave with 6% CO2 (60,000 ppm) and 15% O2, and recommended that fellow biospeleologists seek similar caves for comparative studies.  However, working in such environments runs contrary to law and OSHA regulations. e issue resurfaced when the U.S. Geological Survey and National Park Service administrators of two national parks applied OSHA standards to volunteers in volcanic caves with nontoxic levels of O2 and CO2. However, medical research suggests that establishing minimum blood saturation limits and using portable oximeters would provide more eective protection than applying OSHA standards and requiring CO2 detectors while working in volcanic caves. 1. Hypercarbia and Hypoxia in Closed SpacesOccurrences of hypercarbia in carbonate caves are well comprehended in most of the global speleological community. In some parts of Australia, cavers encounter 8% CO2 and sometimes more. In such caves, simple percent for percent replacement of O2 by CO2 occurs. e same


15th International Congress of Speleology Lava Caves 651 2009 ICS Proceedings eect exists in closed work spaces, and OSHA has set exposure limits to protect American workers. With some minimal variations, similar standards exist broadly across developed, socially conscious nations. Two other American regulatory agencies also have promulgated such standards: the Environmental Protection Agency (EPA) and National Institute for Occupational Safety and Health (NIOSH). In the past, the OSHA standards were dicult to access electronically, and statements by manufacturers of gas monitors were commonly consulted. Some manufacturers have restated the OSHA standards to conform to the inherent capabilities of their particular devices but some others (e.g., InspectAPedia) provide independent information. OSHA currently permits a maximum concentration of 1% CO2 (10,000 ppm) for a ten hour work shi and a maximum of 3% (30,000 ppm) for any ten minute period (OSHA 2008). Also, it promulgated a minimum O2 level of 19.5%, which would mean a maximum CO2 of 1.4%. In contrast, EPA recommends a maximum concentration of 0.1% CO2 for continuous exposure (InspectAPedia, 2009). All these maxima and minima correlate poorly with basic medical investigations of various levels of hypoxia and hypercarbia. Basic research on separate eects of various levels of hypoxia and hypercarbia was conducted so long ago that it cannot be readily retrieved electronically. At approximately 4% atmospheric CO2, breathing becomes alarmingly increased, but the limit of tolerance in highly motivated, healthy young test subjects was found between 7 and 9% (Sollman, 1942). Eventual unconsciousness is reached at 10 to 12%. No such early warning sign exists for hypoxia. Even reduction of O2 to 10% (which is lethal eventually) causes only a slight increase in breathing. Early basic studies also noted wide individual human variations in responses. is is sometimes ignored in later restatements.2. Hypercarbia and Hypoxia in Volcanic CavesA more benign scenario is encountered in some (but not all) volcanic caves with hypercarbia. Some lava tube caves (e.g., the Undara system, ueensland, Australia) are overtopped with such impermeable layers of rock or soil that they function as closed spaces like those commonly encountered in karstic terrains. At the other extreme, many caves in the ows of Kilauea volcano, Hawaii, USA, are within ows with plentiful connections to extratubal spaces and to the surface. Winds blow through even single-entrance caves in such ows, and reverse abruptly with little lag time from the surface. Yet such caves are permeable to magmatic CO2. is may produce a signicantly dierent scenario of additive hypercarbia, quite dierent from the replacement hypercarbia characteristic of closed work spaces and karstic caverns (Table 1). For example, cave air containing 6% CO2 also contains 19.5% O2, which is an insignicant decrease. Anyone slowly dying from a CO2 concentration of 10% in such caves still has plenty of oxygen. is issue surfaced in 1990 when Howarth and Stone (1990) found a notable ecosystem in a volcanic cave with Table 1: Additive hypercarbia and hypoxia in the olcanic cave atmosphere.


Lava Caves 652 2009 ICS P roceedings 15th International Congress of Speleology 6% CO2 and 15% O2 and recommended that fellow biospeleologists seek similar caves for comparative studies. Medical principles indicate that, for healthy volunteers not subject to OSHA constraints, this would be an unpleasant experience but within functional limits. For persons conducting such studies in the course of employment, however, it would be unlawful. e issue resurfaced when US Geological Survey and National Park Service personnel applied OSHA standards to volunteers in volcanic caves with non-standard levels of O2 and CO2. 3. Application of OSHA and EPA Standards to Investigators in Volcanic Cavese intention of the U.S. Congress in creating OSHA was to provide safe work places for all American workers; young and old, healthy and terminally ill alike. is was a politically charming but technically impossible task. OSHA chose to set standards that would protect most workers in most work places, most of the time. Selective enforcement is inherent in the process. Without exibility, volcanologists would be barred from risky eld studies on active volcanoes, for example. Inevitably, OSHAs standard for CO2 responded to the anxiety produced by breathing a mere 4% CO2 and its impact on frightened workers, their union shop stewards, the general public and even concerned employers. A wide safety margin was included to protect aged workers and those with breathing impairments. is likely included some with a nancial stake in maximizing their symptoms. us, these standards have little relevance to speleologists and other investigators especially volunteer investigators. Basic medical data cited above suggest that establishing minimum blood saturation limits and using portable oximeters would provide more eective protection than blindly applying OSHA standards and requiring CO2 detectors in well-aerated volcanic caves, especially in caves where investigators undergo no signicant increase in respiration.AcknowledgmentAustralian caver Garry K. Smith is the originator of the concept of two scenarios for increased CO2 in caves.ReferencesHowarth, F.G. and F. Stone, 1990. Elevated Carbon Dioxide Levels in Bayless Cave, Australia: Implications for the Evolution of Obligate Cave Species, Pacic Science, Volume 44, no. 3, pp. 207. InspectAPedia. 2009. Toxicity of Carbon Dioxide Gas Exposure, CO2 Poisoning Symptoms, Carbon Dioxide Exposure Limits, and Links to Toxic Gas Testing. http://www.inspect-ny.com/hazmat/ CO2gashaz.htm. 6 pages. OSHA. 2008. Carbon Dioxide in Workplace Atmospheres. http:://www.osha.gov/dts/sltc/methods/inorganic/ id172/id172.html. 10 pages. Sollman, T. 1942. A Manual of Pharmacology, and its Application to erapeutics and Toxicology. Philadelphia and London, W.B. Sanders Co., pp 806


15th International Congress of Speleology Lava Caves 653 2009 ICS Proceedings A STEEPLY INCLINED PLIOCENE ? LAVA TUBE CAVE IN DEATH VALLEY NATIONAL PARK, CALIFORNIAWilli ILLI Am M R R HAlli LLI DAy Y1, D D AviVI D Ek K2 1Honorary President, IUS Commission on Volcanic Caves, 6530 Cornwall Court, NN ashville, TNTN 37205 USA A wrhbna@bellsouth.net2AA ssistant Chief, RResource Management, DDeath Valley NN ational Park, CA A 94062 USA A karst@pahrump.net Gneiss Cave is a misnamed ancient lava tube cave which opens on a cli in the northwest end of the Mormon Point Turtleback of Death Valley National Park. While small, it is especially signicant because of it s location in a fault-bounded  block of basalt of uncertain but clearly pre-Pleistocene age. Several geologists have mapped this area, with notably dierent interpretations. One of these maps shows the cave area as gneiss, hence the name of the cave. It is entered through a comparatively spacious chamber which opens widely on the cli face. From it, a tubular passage about 1 m in diameter extends upward at about 45o for a total length of 11.7 m. It is notable for a dusty sheet of banded brown owstone about 2.5 cm thick which does not react to acid. A pathognomonic central ridge of cauliower lava extends along t he oor of the tubular section of the cave and the upper  part of the entrance chamber. Other documented lava tube caves of pre-Pleistocene age are (1) Pahihi Gulch Cave, Maui Island, Hawaii, USA, and (2) an apparently unnamed lava tube cave on the Pacic Island of Truk, reported by Rogers and L egge.  One or more ill-dened caves in Pliocene basalt in  Colorado (USA) and some small lava tube caves opening on a cli in Jalisco, Mexico, are under study. ree well-known, heavily marine-eroded horizontal caves near sea level on the Hawaiian island of Kauai also may qualify.1. Location and History of Gneiss Cavee entrance room of mis-named Gneiss Cave is easily seen by visitors driving on California State Highway 178 just south of Mormon Point in Death Valley National Park, about 30 m above the highway (Figs. 1, 2). Because the scramble to it is steep and rubbly, it has had few visitors and is free of grati and trash. It was named Gneiss Cave because a geological map showed the bedrock surrounding the cave to be gneiss. When the junior author was transferred to Death Valley National Park in 2005, he became condent that it is not in gneiss. But, he was unable to classify the surrounding bedrock or the cave itself. In 2007, the senior author joined him in a further study of the cave (Halliday, 2008) (Fig. 3). Because of the presence of a central longitudinal ridge of cauliower lava (Larson, 1993) extending nearly the entirety of the length of its oor (Fig. 4), it now is considered a lava tube cave, the only example identied in this national park. An accreted tube lining Figure 1: About 30 m aboe a state highway and salt pan, the entrance of Gneiss Cave is in a fault-demarcated block of mac rock, presumably basalt or metabasalt. With a hand lens, an accreted tube lining is easily seen. Figure 2: Looking back down om the entrance to the salt plan on the oor of Death Valley.


Lava Caves 654 2009 ICS P roceedings 15th International Congress of Speleology several centimeters thick also is present.2. Features of Gneiss Caveis is not a large cave. e comparatively spacious entrance room (Fig. 5) is no more than 7 m wide and about 3 m long, with a ceiling height of about 2 m. From it, a tube-shaped passage extends upward at an angle of 45o to a tapered end 11.7 m from the drip line. e upper section contains a sheet of banded brown owstone (Fig. 6) which does not eervesce with acid. It is 1 to 2 cm thick and partially covers the longitudinal ridge and a now-eroded space alongside the ridge. It closely resembles banded siliceous owstone and dripstone seen in some lava tube caves of southwestern Washington State. Plentiful rodent and bird droppings are present, and a little amberat.3. Geology of the Area of Gneiss Cavee geology of the cave area is unusually complex, and the age of the cave will be dicult to determine conclusively. Several geologists have mapped the area, with notably dierent interpretations (e.g., Otton, 1974; Wright et al., 1974, Gregory and Baldwin, 1988). Everyone agrees that it is on the northwest end of the Mormon Point Turtleback, but even the denition of turtleback is less than fully agreed upon. e AGI Glossary of Geology (Jackson, 1997: 684) denes the term as an extensive smooth curved topographic surface, apparently unique to the Death Valley (Calif.) region that resembles the carapace of a turtle or a large elongate dome with an amplitude up to a few thousand meters. Wright et al. (1979) consider them to be colossal fault mullions resulting from severe crustal extension localized along undulating, northwestplunging zones of pre-existing weakness coinciding with anticlinal folds in Precambrian metasediments. From the western foot of the turtleback, the cave appears to be in an unmapped gully-lling basalt ow plunging steeply down Figure 3:. Map of Gneiss Cave. Figure 4: A central longitudinal ridge of cauliower lava extends down the tubular passage and into the entrance room. Figure 5: Looking diagonally across the entrance room to the lower end of the tubular passage.


15th International Congress of Speleology Lava Caves 655 2009 ICS Proceedings the Black Mountains, but this is something of an optical illusion. From a greater distance, it is seen to open on the near-vertical side of the turtleback in the lowest of several fault-demarcated blocks of basalt or metabasalt (Fig. 7). e intricate faulting in this vicinity is especially well-described in Wright et al. (1974). e basalt or metabasalt may be part of the mid-Pliocene Furnace Creek Formation (McAllister, 1970) or a local basalt member of the Pliocene Furnace Formation exposed a few kilometers farther north (Gregory and Baldwin, 1988). e possibility of much greater age cannot be ruled out. Its lithology and relationship to other bedrock diers markedly from that of Cenozoic basalt seen on Recent surfaces nearby and it is clearly pre-Pleistocene.4. Other Pre-Pleistocene Lava Tube CavesGneiss Cave is a member of a very small group of recorded pre-Pleistocene lava tube caves. ese include (1) Pahihi Gulch Cave on the island of Maui, Hawaii, shown on the Stearns and Macdonald (1942) map of Maui as being within Tertiary lava, an apparently unnamed lava tube cave on the Pacic island of Truk (Rogers and Legge, 1985) and perhaps three well-known horizontal caves near sea level near Haena, Kauai Island, Hawaii, which are heavily marine-eroded (Halliday 1981). Recently, several ill-dened caves have been reported in Pliocene basalt in Colorado, USA (Medville and M edville 2008). At least one probably is a  remnant of lava tube. In Jalisco, Mexico, several small lava tube caves  open in a blu of uncertain age  (John Pint, e-mail communications, 2008). If Gneiss Cave is in metabasalt, it is the rst lava tube cave to be identied in this rock. Despite reports to the contrary, (e.g. Alexander, 1980: 78), no lava tube cave has been found in 1.1Ga metabasalt that are notable for other, remarkably preserved pahoehoe features on the north shore of Lake Superior (John C. Green, written communication, 2008). Consequently, this small cave merits especially intensive additional studies.ReferencesAlexander, E.C. Jr. (Ed.) 1980, An Introduction to Caves of Minnesota, Iowa and Wisconsin. Guidebook for the 1980 National Speleological Society Convention. p 78. Gregory, J. and E.J. Baldwin (Eds.), 1988, Geology of the Death Valley Region. Guide for the South Coast Geological Society Field Trip: October 12, 1988. Halliday, W.R. (Ed.), 1981, Haena Caves. In: Introduction to Hawaiian Caves: Field Guide for the 6th International Symposium on Vulcanospeleology, Figure 6: Sheet of banded brown owstone which does not react to acid, located near the top of the tubular passage. e length of the pocket comb used for scale is 115 centimeters. Figure 7: e northwest end of Death Valleys Mormon Point Turtleback, with the Black Mountains in the background. Large and small blocks of mac rock are clearly seen. Gneiss Cave is just aboe the valley oor near the center of the photograph.


Lava Caves 656 2009 ICS P roceedings 15th International Congress of Speleology Hilo, Hawaii. August 1981. McAllister, J.F. 1970. Geology of the Furnace Creek Borate Area, Death Valley, Inyo County, California. California Division of Mines and Geology Map Sheet 14, 1970. Reprinted in B.W. Troxel, editor. 1974. Guidebook, Death Valley Region, California and Nevada. Prepared for the 70th Annual Meeting of the Cordilleran Section, Geological Society of America Field Trip Number 1. pp 84. Shoshone, California, Death Valley Publishing Company. Medville, Douglas and Hazel Medville. 2008. Cave Development in Volcanic Rocks in Colorado. Rocky Mountain Caving, 25, number 4, Autumn, pp 30. Otton, J.K. 1974. Geologic Features of the Central Black Mountains, Death Valley, California. Pages 85. In: B.W. Troxel, 1974. Guidebook: Death Valley Region, California and Nevada. Prepared; for the 70th Annual Meeting of the Cordilleran Section, Geological Society of America Field Trip Number 1. Shoshone, California, Death Valley Publishing Company. Rogers, B.W. and C.-J. Legge. 1985. e 1984 Micronesian Karst Reconnaissance (abstract). National Speleological Society Bulletin 47, number 1. N.p. (page 62). October 1985. Stearns, H.T. and G.A. Macdonald. 1942. Geology and Ground Water Resources of the Island of Maui, Hawaii. 1942. Hawaii Territorial Division of Hydrography Bulletin 7, Plate I. Troxel, B.W. 1974, Geologic Guide to the Death Valley Region, California and Nevada. In: Troxel, B.W. 1974. Guidebook: Death Valley Region, California and Nevada Prepared; for the 70th Annual Meeting of the Cordilleran Section, Geological Society of America Field Trip Number 1. pp 3. Wright, L.A., Otton, J.K. and B.W. Troxel, 1974. Turtleback Surfaces of Death Valley Viewed as Phenomena of Extensional Tectonics. Geology, Volume 2, pp 53 54. Reprinted as pages 79 in B.W. Troxel, editor. 1974. Guidebook: Death Valley Region, California and Nevada. Prepared for the 70th Annual Meeting of the Cordilleran Section of the Geological Society of America Field Trip #1. Shoshone, California, Death Valley Publishing Company.


15th International Congress of Speleology Lava Caves 657 2009 ICS Proceedings MINERALLINED THERMAL EROSION CHANNELS IN A HOLLOW TUMULUS COMPLEX, KILAUEA CALDERA, HAWAII, USAWilli ILLI Am M R R HAlli LLI DAy Y Hawaii Speleological Survey of the NN ational Speleological Society, 6530 Cornwall Court, NN ashville,TN TN 37205 USA A wrhbna@bellsouth.net Sharply incised dendritic channels up to several centimeters in depth and width were found in the sloping o or of a chamber of Fractured Tumulus Cave, a cavernous hollow tumulus complex in Kilauea  Caldera, Hawaii Island, USA. Much of these channels is lined with a silvery, iridescent optically active mineral or minerals. While the occurrence of thermal erosion now is widely accepted, these may be the rst recorded examples of small-scale eroded channels in a cave. Very low concentrations of Na2O and K2O indicate that their linings and thence the lava in which they are incised are early eruption products. Backscatter a nd  other commercial electron microscopy revealed six groups of oxides in one of the linings. Group 1 may be a magnesiocuprous ferrite. e oxide in  Group 2 is primarily Fe2O3. Group 3 apparently is a complex oxide such as TiFe2O5. Group 4 appears to be ilmenite. Group 5 may be a titanomagnetite and Group 6 is a pyroxene. 1. IntroductionHollow tumulus caves are a comparatively uncommon volcanic landform but several examples exist in the 1919 Postal Ri lava ow in Kilauea Caldera (Fig. 1). ey perhaps were rst described by James Dana (1891). In the United States, other examples exist in Idaho (e.g., Abo Dome) and presumably elsewhere. e Postal Ri lava ow is located directly below the Hawaiian Volcano Observatory of the U.S. Geological Survey and probably is the moststudied part of the oor of Kilauea Caldera. As discussed in a companion paper (Halliday, this volume), the Hawaii Speleological Survey has identied a total of about 250 rheogenic caves in this ow. Some are complexes containing more than one hollow tumulus. Fractured Tumulus Cave is a comparatively large example of hollow tumulus complex. It consists of two hollow tumuli with a subterranean crawlway connection through stable breakdown (Fig. 2). Its entrance room is unusually large for this type of cave (Figs. 2, 3) but the notable features in this study are in the smaller inner chamber. Figure 1: Aerial view of Kilauea Caldera looking north. e 1919 Postal Ri lava ow is on the le (west) beyond the large inner pit. Figure 2: e entrance room of Fractured Tumulus Cave is nearly 30 meters wide. e inner room is much narrower, and its oor slopes at about 20 degrees.


Lava Caves 658 2009 ICS P roceedings 15th International Congress of Speleology 2. ermal Erosion Channels in Fractured Tumulus CaveTo date, Fractured Tumulus Cave is unique in Kilauea Caldera in that it contains dendritic thermal erosion channels up to several centimeters in depth and width (Figures 2, 4). ese channels are sharply incised into the sloping oor of the chamber within the smaller of the two tumuli. While the general concept of thermal erosion now is widely accepted, this may be the rst recorded example of small-scale thermally eroded channels in a cave. ey are especially notable for linings of silvery, iridescent and optically active mineral or minerals (Fig. 5). is silvery material also formed owstone and crack llings, and small ponds. It resembles fragments up to 1 cm thick observed as oat in a breakdown area of Jonathans Cave, Puna District, Hawaii, not yet studied Very low concentrations of Na2O and K2O indicate that these linings and thence the lava in which they are incised are early eruption products (Scott Cornelius, e-mail communication, 2007). 3. Procedures and FindingsUnder a National Park Service research permit which required on-site supervision by Don Swanson of the US Geological Survey, small examples were collected for study. Using a binocular microscope, Swanson subsequently identied platy surface minerals as magnesioferrite and titanomagnetite (Don Swanson, oral communication, 2007). Commercial electron backscatter (BSE) and reected light microscopy at Washington State University identied six groups of oxides in the channel lining (Scott Cornelius, e-mail communication, 2007): Group 1. e mineral contains major Cu, and may be a magnesio-cuprous ferrite. It should be noted that small quantities of copper minerals have been found as dripstone in other caves in Kilauea Caldera and elsewhere in Hawaii. Group 2. e oxide is primarily Fe2O3. Group 3. For the formula, ve oxygens seemed to work best, as in TiFe2O3. Group 4: ..appears to be ilmenite. Figure 3: Map of Hollow Tumulus Cave. e inner room containing the incised channels is reached by crawling through stable breakdown.


15th International Congress of Speleology Lava Caves 659 2009 ICS Proceedings Group 5: ..may be a titanomagnetite. Group 6: ..a pyroxene, the location of which is o the photos. Numerical data substantiating these conclusions are presented in Tables 1 and 2, and the tested loci are shown in Figures 6 to 10. Color renditions of these gures will be supplied by e-mail upon request. Figure 4: Looking down one of the incised channels of the inner room. e brown object on the lower right is my boot. Note that with illumination at this angle, the silery, optically active lining of the channel appears tan. In less oblique light it appears nearly indigo. Table 1: WWU ion analysis of selected study points. Figure 5: e specimen analyzed is marked by an arrow. In direct light it appears silery and iridescent.


Lava Caves 660 2009 ICS P roceedings 15th International Congress of Speleology Cornelius noted that for these oxides, reected light microscopy in some ways seems better than BSE. He explained that the contrast in BSE is due to dierences in the average atomic number of the phases present. e data are on a thermal scale, with brighter colors (yellow to white) representing phases with the highest average atomic number. e light brown are mostly or all pyroxenes, olivines or other ferro silicates. e dark brown are undierentiated feldspar and glass. e black is epoxy mounting medium. 4. Optical Propertiese unusual optical properties of this lining were not studied. Depending on the angle of illumination, it appears silvery, tan, or indigo.5. ConclusionsData presented here demonstrate that Fractured Tumulus Cave contains mineralogical and other features not found in other caves in the Postal Ri lava ow. Further study is strongly indicated. Table 2: WWU oxide analysis of selected study points. Figure 6: Back scatter electron microphotograph (BSE) of lining shown in Figure 5. Numbers refer to analysis sites listed in Tables 1 and 2. Electron microscopy by Scott Cornelius, Washington State University (WWU). Figure 7: Same as Figure 6, by reected light.


15th International Congress of Speleology Lava Caves 661 2009 ICS Proceedings AcknowledgmentsMy profound thanks to Don Swanson of the U.S. Geological Survey for invaluable eld assistance and laboratory studies, to Harry Shick for notable eld assistance, to Scott Cornelius of Washington State University for going beyond the call of duty in analyzing the specimen, and to Jim Martin, former Superintendent of Hawaii Volcanoes National Park for encouraging the Hawaii Speleological Survey to conduct studies in Kilauea Caldera. All work described here was self-funded.ReferencesDana, James D. 1891. Characteristics of Volcanoes. New York, Dodd, Mead and Co., pp 117. Halliday, William R. (this volume) Unusual rheogenic caves of the 1919 Postal Ri lava ow, Kilauea Caldera, Hawaii. Figure 8: BSE of another section of lining shown in Figure 5. Figure 9:. Same as Figure 8, by reected light. Figure 10: Enlargement of section of Figure 8.


Lava Caves 662 2009 ICS P roceedings 15th International Congress of Speleology UNUSUAL RHEOGENIC CAVES OF THE 1919 POSTAL RIFT LAVA FLOW, KILAUEA CALDERA, HAWAIIWILLIA A M R R HA A LLIDA DA Y Hawaii Speleological Survey of the NN ational Speleological Society, Commission on Volcanic Caves of the UIS, 6530 C ornwall Court, NN ashville, TNTN 37205 USA A wrhbna@bellsouth.net In follow up of a 1991 study by George P. L. Walker of the surface of a small part of the 1919 Postal Ri lava ow in Kilauea Caldera, Hawaii Speleological Survey teams studied the entire 1919 ow. Walker described and discussed tumuli, lava rises, lava rise pits, and lava ination cles. He described a nd depicted one hollow tumulus and noted the presence of one subcrustal grotto in a lava  rise. In  our intermittent 12-year study of the entire ow, Hawaii Speleological Survey teams identied and studied approximately 250 rheogenic caves and a few crevice caves. Most of the rheogenic caves are shallow cavities drained aer subcrustal injection of lava rather than classical lava tube conduit caves: ow lobe caves, lava rise caves, hollow tumuli and the like. Only two lengthy lava tube caves were identied, and one of these shares morphological characteristics with some hollow tumuli. In and near Walkers study area, we identied and mapped (1) a second hollow tumulus with a sizeable melthole connection to t he one  identied and mapped  by Walker, (2) a complex circumferential cave in the perimeter ridge of a lava rise  whose surface features  he identied and mapped, (3) a penetrable crevice cave which drained a lava pond in a second lava rise which he identied, and we mapped and inventoried several other hollow t umuli and other lava rise caves just outside  his study area. e internal morphology of the lava rise caves diers signicantly from that of the hollow tumuli, thus substantiating Walkers dierentiation of lava r ises from tumuli  and providing additional insight into the emplacement of pahoehoe ow elds.  1. History and Techniques of the StudyIn 1991, George P.L. Walker published a notable paper on ination features of a small part of the 1919 Postal Ri lava ow in Kilauea Caldera, Hawaii Island, Hawaii (Walker, 1991). He described and depicted one hollow tumulus and noted the presence of one sub-crustal grotto in a lava rise. In an intermittent 12-year study from mid-1994 to Spring 2006, teams of the Hawaii Speleological Survey continued his work and expanded it to the entire lava ow (Figs. 1-5). We undertook 180 ten to eleven hour day trips in the course of 23 eld seasons. Approximately 251 caves were identied and 83% of these were investigated with forms reported here. Of the others, eleven (4%) were too tight, eleven (4%) too hot, ten (4%) were both too hot and too tight. ree (1%) were bypassed because of fragile features and three Figure 1: Looking southeast across the 1919 Postal Ri lava ow om the rim of Kilauea Caldera. e south (uphill) end of Walkers study area is in the immediate foreground. Shown are Lava Rise Cave E-3 (E-3), Almost Too Hot Cave (ATH),.Red Slope Cave (RS), and Sleeping Ohia Cave (SO). Sleepings Sister Cave is just to the le of Sleeping Ohia Cave, out of the photograph. Figure 2: Looking east across the 1919 Postal Ri lava ow om the caldera rim. e north end of Walkers study area is in the immediate foreground. Shown are Lava Rise E-3 (E-3), Lava Rise E-5 Caves (E-5), Tumulus E-4 (E-4), Tumulus E-6 (E-6), Tumulus E-1 (E-1) and Lava Rise E-x Caves (E-x).


15th International Congress of Speleology Lava Caves 663 2009 ICS Proceedings (1%) were found to be tectonic rather than rheogenic. Two (1%) were lost or forgotten, and permission was denied for entry into approximately four caves (2%). Many of the caves are hyperthermal (up to 77o C), with 100% relative humidity, thermostratication, and/or changing underground wind currents. is required development of new exploration techniques including identication and use of relatively cool layers of air for escape routes of one minute or less. Core temperatures were monitored. Sublingual temperatures of 38o C were found to cause serious impairment of critical functions requiring emergency replacement of uids and electrolytes as well as cooling. Symptoms from core temperatures of 37.5o generally beneted from these replacements. Some individual variations were found in the exposure needed to cause such hyperthermia. Tables of broadly tolerable exposure to various temperatures were developed empirically and published (Halliday, 2000a) but occasionally were exceeded briey by participants with especially high heat tolerance. Light weight, hard weave cotton clothing was used at all times, together with a variety of face masks in hyperthermal caves. Figure 3: Selected maps om Walker (1991) showing his study area and features. No cavernous features have been found in his southern group (E-2, E-7E-10). Figure 4: View of Tumulus E-1 om Lava Rise E-x. e latter is outside Walkers study area. Tumulus E-6 also is seen on the le.


Lava Caves 664 2009 ICS P roceedings 15th International Congress of Speleology Noxious gas (probably HCl) was encountered only in one tiny cave on the edge of Halemaumau Crater. Presumed suldic fumes were encountered in numerous cave but were found to be essentially non-toxic. Eye irritation rarely was encountered (Halliday, 2000b). Use of two types of CO2 monitors was required in previously untested volcanic caves for the last ve eld trips. ey were found to be useless in hyperthermal caves and no signicant elevation of CO2 was identied in normothermic (normal body temperature) examples (Halliday, 2007). In no cave was signicantly elevated CO2 identied by changes in normal breathing (Halliday, this volume).2. Types of Rheogenic CavesOne major lava conduit system consisting of three individual caves was identied, inventoried and mapped: the Postal Ri System (Fig. 2). With a total length of 1,081 m, it is the master conduit for the 1919 ow in the caldera but extends for less than a third of the length of the ow. Directly beneath the US Geological Surveys Hawaiian Volcano Observatory, its 1919 origin from overow of Halemaumau pit crater was documented in detail in early serial publications of that institution. Nearly all the other non-tectonic caves in this ow were formed by drainage of sub-crustal injection and lava breakout. ese form a continuum between rough-shaped hollow breakouts a few meters in diameter and approximately a meter in height, hollow tumuli of several types, and drained injection spaces, the largest of which have some characteristics of a partially developed conduit tube. Drained lava rises, drained ow lobes and drained lava tongues are common. Meltdown of short-lived lava between adjoining cavernous landforms has produced complex oor plans and limited vertical development. No hollow pressure ridges comparable to those in the Myvatn region of Iceland were found.3. Forms of Rheogenic CavesSeveral very small tumuli and ow lobe caves were found with maximum ceiling heights of 1 to 1.5 m and diameters of a very few meters. e Puka X group of caves are little more than hollow breakouts. Lehua Cave is an example of a larger, igloo-shaped hollow tumulus. More commonly, such tumuli are elongated in the direction of ow. New Cave is an example almost completely buried by ows subsequent to its development. Some larger hollow tumuli are oblong and are oriented transversely to the ow. Black Hole Cave is an unusual example. Only its outer shell is essentially intact. Beneath it, breakdown is so extensive that little of the original domed space remains. Hollow whale back tumuli are larger and even more elongate. Sleeping Sister Cave is a prototype. Its upper end is a very small sub-crustal tube. It curves and soon enlarges to a width of more than 15 m within a whaleback ridge. e outer chamber of Tumulus E-1 Cave was studied by Walker (Fig. 3). Our study found the orice he labeled drainage tube to be a melt hole of lava originally separating cavities in Tumuli E-1 and E-4 (Fig. 5). Figure 5: Plan of Tumulus E-1 Cave showing melt hole connection to chamber in Tumulus E-4. Ceiling heights are in meters. Figure 6: Plan of Almost Too Hot Cave. Ceiling heights are in meters.


15th International Congress of Speleology Lava Caves 665 2009 ICS Proceedings Hollow sinuous tumuli and feeder and drainage tubes are common. Almost Too Hot Cave also has a wide terminal room but is an illustrative example of hollow sinuous tumulus (Fig. 6). About half of its height rises above the local lava surface. Much of Red Slope Cave has the same form, but it is more complex. Much of it is the product of deation of a seemingly ordinary lava tube conduit cave, with a somewhat tubular residual passage along one edge of the original tube. Locally, the original roof slumped irregularly, producing up to three sub-parallel passages. e original cave was immature, and brief periods of overload produced local upper levels, breakouts, and small tumuli. Figure 7: Plan of Lava Rise E-3 Cave. Ceiling heights are in feet.


Lava Caves 666 2009 ICS P roceedings 15th International Congress of Speleology Figure 8: Plan of Lava Rise E-5 Cave. A small melt hole connects it to an originally separate cave. Christmas Cave is even more complex. Its main entrance is at the lower end of a lava trench formed by similar slumping of the roof of an ordinary lava tube cave, and its long southern extension is in the preserved margin of that cave. is cave is of special interest because of its nesting of


15th International Congress of Speleology Lava Caves 667 2009 ICS Proceedings drained ow lobes (pancake rooms), as theorized by Hon et al. (1994) as an important factor in the homogenization of pahoehoe ow elds. Other parts of this complex cave include a hollow hornito and hollow ow lobes connected to the its northeastern section through a melt hole at the eastern extreme of the mapped portion. Merry Go Round Cave also has a hornito-like structure protruding upward from its main passage. e original connection solidied, and feathery lava extrusions are present on the cave ceiling. Flow lobe caves tend to reect local topography during the emplacement of the parent ow. Plywood Cave extends down the side of a solid tumulus, retrograde to the direction of general ow. New Entrance Cave is a ow lobe cave with budding of a secondary chamber or passage. A slumped area draining into the budded passage demonstrates the speleogenetic sequence. Big Ell Cave is within a hollow tumulus with a deep surface indentation above the point of budding of a large side passage formed by this process. Natural Bridge Cave and nearby South Twin Puka Cave uniquely consist of budded crawlways reecting large terminal toes of the ow. Lava rise caves appear to have originated through a mechanism also seen in Red Slope and Christmas caves. ese are largely perimeter caves, circumferential to large deated areas (Figs. 7, 8), characteristically containing numerous lavaballs. Lava Rise E-3 Cave is especially notable for its lengthy peripheral and short central drain passages. Lava Rise E-5 Cave has a central cavernous crevice from which very uid pahoehoe was forced late in the speleogenetic process. e base of Tumulus E-6 can be inspected in the north cavern of Lava Rise E-5 Cave.AcknowledgmentsMy heartfelt thanks to Jim Martin, former Superintendent of Hawaii Volcanoes National Park, for long-term encouragement and actual participation in this study, and to Bobby Camara, Cave Specialist of that park, who assisted notably. Don Swanson of the U.S. Geological Survey and Ken Hon of the University of Hawaii, Hilo Branch, made valuable suggestions in the eld. Too many members and cooperators of the Hawaii Speleological Survey participated in eld studies to name them all here, but without their generous assistance, the project would not have been possible. It was self-funded by participants. ReferencesHalliday, W.R. 2000a. Mapping Caves in Kilauea Caldera: Proposed Standards and Guidelines. Hawaii Speleological Survey Newsletter #7, June 2000, pp 33. Halliday, W.R. 2000b. Kilauea Caldera 2000. Hawaii Speleological Survey Newsletter #8, Dec. 2000. p 21. Halliday, W.R. 2007. e end of the project. February 2006 in Kilauea Caldera. Hawaii Speleological Survey Newsletter #21. pp 10. Hon, K. et al. 1994. Emplacement and ination of pahoehoe sheet ows: observations and measurements of active lava ows on Kilauea Volcano, Hawaii. Geological Society of America Bulleetin 106:351. Walker, G.P.L. 1991. Structure, and origin by injection under surface crust, of tumuli, lava rises, lava rise pits, and lava ination cles in Hawaii. Bulletin of Volcanology, 53:546.


Lava Caves 668 2009 ICS P roceedings 15th International Congress of Speleology Principles RINCIPLES of OF pyroduct PYRODUCT lava LAVA tunnel TUNNEL formation FORMATIONSTeph EPH AN Kempe EMPE Institute of AA pplied Geosciences, University of TT echnology DDarmstadt, Schnittspahnstr. 9, DD -64287 D D armstadt, Germany, Kempe@geo.tu-darmstadt.de Shield volcanoes owe their shape to the fact that low-viscosity, high-temperature lavas form internal tunnels in which the lava can be transported for tens of kilometers. Originally described as tunnels and termed pyroducts (in analogy to aqueducts) they are integral features of phoehoe lava ows. e term lava tube, implying that lava is simply piped downhill, should be avoided. Contrary to the popular idea that lava tunnels form by the crusting over of lava channels, they form at the tip of the lava ow by a repeated process of advance and ination. Within the stack of lava sheets formed initially, the hottest conduit will attract most of the ow. Soon aer the lava will start eroding downward, thus creating an underground canyon-like tunnel with a river of low viscosity lava at the bottom. Back-cutting lavafalls can quickly enlarge this canyon uphill. Collapse of the primary roof can open skylights (pukas) that allow convective cooling of the lava. It reacts by freezing over, forming a secondary roof, under which the owing river is once more protected from heat loss. Field investigations suggest categorization into single-trunked, double(or multiple)-trunked and superimposed-trunked systems. First category examples are some of the very long caves like Kazumura, Ke`ala and Ainahou Caves (all Kilauea, Hawai`i). Doubletrunked systems operate contemporarily side-by-side inuencing each other. One known example is the interaction between the upper part of the Huehue Flow and the Mystery Flow (Hualalai, Hawai`i). e nal category includes superimposed tunnels all active at the same time. Such systems can come about by an increase in lava volume during an eruption causing new tunnels to form on top of the old ones that also stay active but cannot swallow the increased ow. Such a system most likely is represented by Kulakai Cave (Mauna Loa, Hawai`i).1. IntroductionTransport of lava through interior tunnels of phoehoe lava ows is an important volcanic process. It allows lava to ow over long distances. Within the tunnel, heat is lost only conductively while surface lava looses heat convectively at the upper and conductively at the lower interface. Lava transport through tunnels is therefore the reason why basaltic shield volcanoes attain very low slopes (oen <2, Cave Total length, km Main trunk length, km End-to-end bee-line, km SinuosityVertical distance m SlopeVolcano Kazumura Cave 65.5041.8632.11.301101.81.51K, A Keala Cave 8.607.075.591.251861.51K, A J. Martin/Pukalani System6.26 K, A Epperson`s Cave 1.931.130.801.41--K, A Thurston Lava Tube 0.4900.4904321.1320.12.4K, A Ainahou Ranch System 7.114.82*4.271.133233.83K, A? Keauhou Trail System, A? Charcoal System 1.5 1.4 602.6K, Earthquake System 0.34 334.7 Huehue Tube, HH (Clagues Cave**) 2.731.391.181.15157.16.49H, HH) Paauhau Civil Defense C. 1.000.580.501.14494.87MK*horizontal; ** upper part of Huehue Table 1: Comparison of some morphological indices of some of the Hawaiian lava tunnels (for sources of data see KEMPE, 2002). (K,A: Kilauea, Ai-la`au; H, HH Huallai, Huehue ow of 1801; MK: Mauna Kea).


15th International Congress of Speleology Lava Caves 669 2009 ICS Proceedings compare Table 1) and why large continental lava plateaus fed by multiple eruptions can form. ese facts are known, even though many volcanology textbooks do not mention them or deal with them only marginally. Overall, a phoehoe lava ow is stationary, growing only at its tip One can walk across its active tunnel without even noticing it. A`a ows on the other hand move like glaciers, shoving down-hill in their entire width. ese ows do not develop a tunnel. e rst to describe a lava tunnel aer having inspected it by himself was apparently Olafsen L AFSEN (1774) who visited Island 1752 to 1757 (Kempe E MPE 2008). In his description of Surtshellir ( 358, p130; later almost verbally copied by Rosenm O SENM ller LLER & Til IL Lesius ESIUS 1799) we read: DD er ieende Hraun ist wie ein Strom durch diesen Canal geossen; (the running lava owed through this channel like a river). Troil R OIL (1779; p 225) who visited Island with Joseph Banks and Daniel Solander in 1772 wrote: DD ie obere RR inde wird bisweilen kalt und fest, obgleich die geschmolzene Materie noch unter derselben weglu, dadurch entstehen groe Hhlen, deren Wnde, Betten und D D ach aus Lava besteht, und wo man eine Menge TT ropfstein aus Lava ndet. (e upper crust sometimes cools and solidies, even though the molten matter keeps running underneath; in this way large caves form, the walls, oors and ceiling of which are composed of lava and where a lot of dripstones of lava occur). e rst one to report seeing an active tunnel was Coan O AN (1844) who in 1843 ascended Mauna Loa: But we soon had ocular demonstration of what was the state beneath us; for in passing along we came to an opening in the superincumbent stratum, of twenty yards long and ten wide, through which we looked, and at the depth of y feet, we saw a vast tunnel or subterranean canal, lined with smooth vitried matter, and forming the channel of a river of re, which swept down the steep side of the mountain with amazing velocity. e sight of this coered aquaduct or, if I may be allowed to coin a word, this pyroduct tiled with mineral fusion, and owing under our feet at the rate of twenty miles an hour, was truly startling. us Coan used the term tunnel, but he also coining a new one: pyroduct. is is a specic term for a very specic natural phenomenon and should take precedence over younger terms. On the other hand famous geologists like James Dana continued to use tunnel, while J.W. Powell introduced volcanic pipes and Tom Jaggar used tunnel as well as tube; only aer 1940 the term lava tube became standard (pers. com. J. Lockwood OCKWOOD ). In this contribution I will go back to the old term tunnel and, for reasons of rules of scientic nomenclature, pyroduct as well. ere is also good reason why to avoid the term tube, because it invokes the picture of pipes in which lava can ow up and down under pressure like in a plumbing system. Furthermore, tube implies a circular cross-section that is only rarely found. e discovery of long lava ows on Moon, Venus and Mars and even active volcanoes on Io has increased the interest in pyroducts in the last few decades. On Earth the longest surveyed lava tunnel is Kazumura Cave (65.5 km) (Hawai`i, Kilauea Volcano) (Allred L LRED et al., 1997) and the longest terrestrial tunnel-fed ow is that of Undara/Australia (Atkinson TKINSON 1993). e authors group explored and surveyed many other caves on Hawai`i (Kempe E MPE 2002) and in Jordan (Kempe E MPE et al., 2006a) that give opportunity to study formation and evolution of pyroducts from the inside.2. Formation of PyroductsIn many text books lava tunnels are described as having formed by crusting over channels (e.g., Francis, 1993). Such caves do exist, but they are clearly in the minority. Mostly they form short, roofed sections of open-surface lava rivers contained in levees. In cross-section such cave roofs show accreted layers growing from the sides inward and having a central vertical parting where the growing lateral shelves met. e long lava tunnels form, however, by ination (Hon O N et al., 1994). It is a process that is incremental and it starts at the distal tips of the phoehoe ows where hot lava rapidly covers the ground in thin sheets. e sheet will cool quickly, causing the dissolved gases to form bubbles, decreasing the overall density of the rock. e next pulse of advancing lava will li this sheet up (ination by buoyancy) before forming the next distal surface sheet. Multiple advances can occur, forming a primary roof with several sheets, separated by sheer interfaces (only the rst or top sheet will have subaerially formed ropy structure) (Fig. 1). e oldest lava sheet is therefore on top of the stack in contrast to normal stratigraphic conditions. Below the primary roof the lava can stay hot and can keep owing. is is the initial tunnel. us these caves are characterized by roofs build of one or several, sometimes more than ten continuous sheets of lava. is roof structure can be studied at roof collapses, called pukas in Hawai`i.2. Internal developmentIf the area to be covered by the rst advance is rather at, many small, parallel conduits can develop Each of them can start to erode down soon aer. One of the threads will, however, erode fastest and attract the largest ow volume. It will then drain the other parallel ducts one by one of their lava, oen leaving them as small-scale labyrinths high above the nal oor (Fig. 2). Since these mazes are drained from


Lava Caves 670 2009 ICS P roceedings 15th International Congress of Speleology their lava when the whole system is still very hot, their oors are mostly smoother than that of the later main tunnel. As the erosion continues, the lava runs with an open surface in a self-generated underground canyon. is canyon is cut into older rocks, not associated with the current eruption. is fact can be studied at places, where the thin lining of the side walls has fallen away (e.g., Greeley R EELEY et al., 1998; Kempe E MPE 2002). Oen we nd a`a blocks behind or even ash layers, both certainly not integral parts of phoehoe ows. Downcutting is facilitated by a variety of processes. One of the more spectacular processes is canyon formation by backcutting lavafalls (Kempe E MPE 1997; Allred LLRED & Allred LLRED 1997) (Fig. 3). ese are quite common in the long Hawaiian caves, but none are yet found in Jordan. e falling lava hammers out the rubble from the oor. It consists of less dense rocks that oat up and are transported on the surface of the river. us, other than in a water river, the bed is not protected by bedload and therefore prone to continued erosion. e mobilized blocks are cool and receive a coating of lava forming lavaballs. Some of the lavafalls seem to be stationary forming large plunge-pools and chambers (Allred LLRED & Allred LLRED 1997). us the tunnel will grow in depth and width in an uphill direction. e passage above the lava falls is quite small in contrast. As one proceeds uphill, the canyon will become larger and larger until one enters the next plunge pool chamber. It is not quite understood how much mechanical erosion and how much melting of the river bed occurs (e.g. Greeley R EELEY et al., 1998; and citations in Kempe E MPE 2002). Other enlarging processes may occur, such as small phreathic explosions, blowing out sections of the wall or oor as groundwater is vaporized. As the downcutting continues the river meanders, undercutting walls and destabilizing the roof. Breakdown falling into the owing lava is also carried away. If the primary ceiling collapses entirely, a skylight or puka opens up If the ow is still active, the rubble can be carried away and we speak of a hot puka. If the collapse occurs aer the termination of the ow, there will be a breakdown pile, sometimes giving easy access to the cave below, sometimes sealing it completely this is termed a cold puka. If a puka opens up during activity then hot gases can escape from the tunnel and heat exchange can occur convectively. e heat exchange is specically ecient if two pukas open up en the upper one will serve as an exit or chimney of hot gases, while cold external air is drawn into the cave at the lower puka. is intrusion of cold air causes the surface of the lava river to freeze over, forming a secondary ceiling in the canyon (Fig. 3). e secondary ceiling will split the passage into two levels, one on top of each other extending between the two pukas but not very much further. is process can occur several Figure 1: Sketch illustrating pyroduct (lava tunnel) formation. At the tip of a phoehoe ow lava advances quickly in form of a delta of thin, ropy lava. e next pulse of lava lis the rst sheet up (ination). is process is repeated until a stack of lava sheets (the primary roof) is formed, below which the hottest ow thread becomes the later tunnel.


15th International Congress of Speleology Lava Caves 671 2009 ICS Proceedings times consecutively, forming e.g. a tertiary ceiling. Later spills from below through breakdown holes or from the upstream end of the secondary ceiling can reinforce it from above. In Ke`ala Cave, Hawai`i, one section of secondary ceiling is over 1 km long. Very oen the upstream end of the secondary ceiling is sealed. is is caused by lavaballs oating on the lava river that are too buoyant to be dragged below the secondary ceiling. Instead they strand on the upper edge of the secondary ceiling. e accumulated blocks are then welded together by splashed-up lava. Floating blocks can be very large, in Waipouli Makai Cave there is a block about 12 m wide, 8 m long and 5 m thick welded into the ceiling of the cave (Kempe E MPE et al., 2006b). e cold, oxygen containing external air that is drawn into the Figure 2: Detail of the ground plan of the Huehue Cave illustrating how a primary maze of parallel lava threads (A) was drained until the master trunk that cut down fastest remained (E). (A) Initial pattern of lava tunnels, ow was om right to le in up to ve parallel conduits. (B) e southern-most conduits were drained rst. (C) Further downcutting reduced the number of active conduits to three. At the same time, lava om a parallel ow (Mystery Flow) coered the area. (D) Only one tunnel remained (following the thick line in B and C), its bed 2 m below the original surface. e added overburden caused collapse of pukas (labeled 8, 6 and 4) the breakdown of which was remoed by the lava river. Now external air began owing uphill om Puka 8 to 4. As a consequence a secondary ceiling oze out aboe Puka 8 and below Puka 8. (E) Various spill-events reinforced the secondary roof, even spilling into the already drained upper conduits, closing the northern-most one. ese spilled lavas were oxidized by the passing surface air and attained various shades of red caused by the crystallization of very ne-grained hematite. e last event was the collapse of puka 5 while the tube cooled. Its material is still in place.


Lava Caves 672 2009 ICS P roceedings 15th International Congress of Speleology cave can oxidize lava surfaces that are still hot. e iron, contained in the volcanic glass, is oxidized to ne-grained hematite, tinting the surfaces of secondary ceilings in various hues of red. us, red lava is not any dierent from the lava in the tunnel, it just was exposed dierently to oxygen during cooling. Hot pukas can also serve as temporary rootless vents when the tunnel below is obstructed or even closed entirely. Lava can then ow out of them, forming relatively rapidly cooling, thin, ropy phoehoe. e Puka 17 Flow out of the lower part of the Huehue tunnel (Fig. 4) is an example. Pukas, cold or hot, can also serve as entrances for lava of later ows. e upper and lower ends of Ke`ala Cave were plugged by later lava invading into pukas. All these processes act to form caves of complex pattern, morphologically not representing tubes at all. Also the total length of the caves is usually much larger than the simple distance along the main tunnel. Table 1 gives some basic morphometric data, such as sinuosity and slope for some tunnel systems. 3. General Types of PyroductsOverall, we can dierentiate several general types of lava tunnels. Here I would like to introduce three new terms to describe the general functioning of the lava tunnels. ese are: (a) single-trunked systems, (b) double(or multiple)trunked systems and (c) superimposed-trunked systems Most of the lava tunnels yet documented in enough detail appear to belong to the single-trunked category. ey are fed by one eruption vent and the meso-morphological internal structure can be explained by the processes discussed above. e tunnel size depends on the lava discharge rate and on the length of activity (days, weeks, months and possibly even years), i.e., on the time available for erosion. If the eruption stops or the tunnel collapses or is blocked, the tunnel will cool. e next or even in case of a blocked tunnel the same eruption will then create a new pyroduct. Normally, it will be situated to either side of the previous ow because it now forms a topographic ridge (ow lobe). Typical examples of single-trunked systems are Kazumura, Ainahou Ranch, Ke`ala and others of the long Hawaiian caves. If lava from the new tunnel should spill through a puka, or break (because of its overburden) into any older, underlying tunnel, then the older tunnel will be lled by lava cooling in the same pattern as at the surface. Double-trunked systems are comprised of two lava tunnels, active side by side at the same time and fed by two separate eruption points. Such tunnels can interact and cause more complex morphologies than described above. One example is the interaction between the Huehue Flow and the Mystery Flow (Fig. 4; Kempe E MPE 2002). In this case the Huehue ow established its tunnel rst. en a second eruption point (the very inconspicuous, low Mystery Shield) erupted lava, establishing a small tunnel system to the side of the Huehue ow lobe. Part of the Mystery lava formed surface lavas that quickly cooled forming a`a ows. ese superseded the upper part of the Huehue tunnel to the le. Figure 3: Longitudinal section of an evoling lava tunnel. Top: Erosional enlargement of the underground canyon by backcutting lavafalls. Bottom: Upon static failure and partial collapse of the primary roof a skylight (puka) opens up, allowing cold air to enter the tunnel, eezing an internal secondary roof below which erosional enlargement can continue. Spills om uphill or through holes reinforce the secondary roof.


15th International Congress of Speleology Lava Caves 673 2009 ICS Proceedings Once thick enough, the primary, sheeted roof of Huehue collapsed and le a roof composed of Mystery a`a lava. e resulting breakdown was removed with the active lava river. Due to the large, hall-like cavity that formed, a secondary roof froze out over the active ow of Huehue. Later rockfall covering the newly formed false oor gives the upper passage the appearance as if the tunnel was formed in a`a, an impossibility near to a vent issuing very hot basaltic lava. e least understood and documented category is the superimposed-trunked system. It is dened as a set of lava tunnels superimposing and crossing each other, all being all active at the same time. e upper tunnels stop their activity rst, so that the lower ones carry on for some time before they also stop operating and become emptied. ere may even be connecting openings between the levels exchanging lava between cross-overs. Such systems could arise when a volcanic vent increases its output volume during an ongoing eruption. en the already established pyroducts cannot accommodate the increased ow volume and a new story of independently operating tunnels is build on top of the already active one. To my knowledge the Kulakai System on Hawai`i is an example of such a superimposed-trunked system.4. ConclusionsIn spite of the tremendous progress made in lava cave exploration, we still are far from understanding all the features and processes that interact during lava tunnel formation. It is clear, that the concept of a tube, simply piping lava downhill, is far too simple to explain the observed morphologies. Furthermore, many published lava cave maps are rather useless because they are not linked to a geological map of the ow; many of them do not even oer morphological details and cross-sections (if they are made at all) do not show the structure of the lava ow itself. us, much more process-oriented analysis is needed in order to advance lava cave research.References:Allred LLRED K. and C. Allred LLRED (1997) Development and morphology of Kazumura Cave, Hawai`i. Journal of Cave Karst Studies, 59(2): 67. Allred L LRED K., C. Allred LLRED and R. Richards ICHARDS (1997) Kazumura Cave Atlas, Island of Hawai`i. Special Publication Hawaii Speleological Survey: 81 pp. Figure 4: Geological map of the Huehue ow (Hualalai Volcano, Hawai`i) of 1801 according to surveys of the authors group Flow was om right to le, vertical distance 500 m. Oldest are the lavas (light grey) of the Puhia Pele vent (a series of spectacular spatter cones) that was gas-rich and formed an open channel system (with a few roofed-over sections). Aer its termination the Huehue Flow erupted (numbers label pukas on the Huehue Cave, thick line; lava light middle grey le and at Puka 1) accompanied by the Mystery Shield eruption that was active for only a short time (upper right) and coered much of the upper part of the Huehue Flow (dark middle grey). Zoes Puka is a tunnel belonging to the Mystery ow. From Puka 17 a surface ow issued (dark grey) and the terminal lava om the mystery shield formed several short a`a ows (dark grey at the right).


Lava Caves 674 2009 ICS P roceedings 15th International Congress of Speleology Atkinson TKINSON A. (1993) e Undara Lava Tube System, North ueensland, Australia: updated data and notes on mode of formation and possible lunar analogue. Proc. 6th Intern. Symp Volcanospeleol., Hilo, Aug 1991: 95. Coan O AN T. (1844) Letter of March 15, 1843, describing the Mauna Loa eruption of 1843. Missionary Herald, 1844. Francis R ANCIS P (1993) Volcanoes, a Planetary Perspective. Oxford University Press: 443 pp. Greeley R EELEY R., S.A. Fagents AGENTS R.S. Harris ARRIS S.D. Kadel ADEL and D.A. Williams ILLIAMS (1998) Erosion by owing lava, eld evidence. J. Geophys. Res., 103 (B11): 27,325,345. Hon O N K., J. Kauahikaua AUAHIKAUA R. Denlinger ENLINGER and K. Mackay ACKAY (1994) Emplacement and ination of phoehoe sheet ows: observations and measurements of active lava ows on Kilauea Volcano, Hawai`i. Geol. Soc. Amer. Bull., 106: 351. Kempe E MPE S. (1997) Lavafalls: a major factor for the enlargement of lava tubes of the Ai-la`au Shield phase, Kilauea, Hawaii. Proceedings 12th International Congress of Speleology, La Chaux-deFonds, Switzerland, 1: 445. Kempe E MPE S. (2002) Lavarhren (Pyroducts) auf Hawai`i und ihre Genese. In: Angewandte Geowissenschaen in Darmstadt, Rosendahl, W. & Hoppe, A. (Eds.), Schrienreihe der deutschen Geologischen Gesellscha, 15: 109. Kempe E MPE S. (2008) Immanuel Kants remark on lava cave formation in 1803 and his possible sources. Proceedings 13th Intern. Sympos. on Volcanospeleology, Jeju Island, Korea, 1.-5. Sept. 2008: 35-37. Kempe E MPE S., A. Al L Malabeh ALABEH M. Frehat REHAT and H.V. Henschel ENSCHEL (2006a) State of lava cave research in Jordan. Proc. 12th Intern. Symp on Vulcanospeleology, Tepotzln, Mexico, July 2, 2006, Assoc. for Mexican Cave Studies, Bull. 19 and Socieded Mexicana de Exploraciones Subterrneas Bol., 7: 209. Kempe E MPE S., H.V. Henschel ENSCHEL H. Shick HICK and F. Trusdell R USDELL 2006b) Geology and genesis of the Kamakalepo Cave System in Mauna Loa picritic lavas, Na`alehu, Hawai`i. Proc. 12th Intern. Symp on Vulcanospeleolpgy, Tepotzln, Mexico, 2-7 July, 2006, Assoc. for Mexican Cave Studies, Bull. 19 and Socieded Mexicana de Exploraciones Subterrneas Bol., 7: 229. Olafsen L AFSEN E. (1774, 75) Des Vice-Lavmands Eggert Olafsens und des Landphysici Bianre Povelsens Reise durch Island, veranstaltet von der Kniglichen Societt der Wissenschaen in Kopenhagen (from the Danish 1st edn. 1772) 2 vol., Heinecke und Faber, Kopenhagen und Leipzig, 1 map, 51 copperpl., 328+xvi+244 pp. Rosenm O SENM ller LLER J.C. and W. G. Tillesius ILLESIUS (1799) Beschreibung merkwrdiger Hhlen, ein Beitrag zur physikalischen Beschreibung der Erde. Breitkopf und Hrtel, Leipzig, 10 copperpl. 294 pp Troil R OIL U.v V Briefe welche eine von Herrn Dr. Uno von Troil im Jahr 1772 nach Island angestellte Reise betreen. Magnus Swederus, Upsala und Leipzig, 342 pp.


15th International Congress of Speleology Lava Caves 675 2009 ICS Proceedings Interpreting NTERPRETING the THE genesis GENESIS of OF Thurston HURSTON Lava AVA Cave AVE Kilauea I LAUEA Hawai AWAI i ISTeph EPH AN Kempe EMPE1, HORs S TVOlke LKE R He E Nschel SCHEL2 1Institute of AA pplied Geosciences, University of TT echnology DDarmstadt, Schnittspahnstr. 9, DD -64287 DDarmstadt, Germany, e-mail Kempe@geo.tu-darmstadt.de2Henschel & RRopertz, AA m Markt 2, DD -64287 DDarmstadt, Germany, e-mail: h-v.henschel@henschel-ropertz.de urston Lava Tube (alias Keanakakina), discovered east of the Kilauea Iki Crater on Hawai`i in 1913, soon became an attraction in the Hawai`i Volcanoes National Park. It is visited daily by many tourists who get here their only chance to experience a true pyroduct. Nevertheless, not much is known about urstons speleogenesis and previously published maps are not very detailed. We resurveyed the cave in 1996 with high precision, establishing its morphometry (horizontal length 490 m, direct distance 432.5 m between ends, sinuosity 1.13, vertical drop 20 m, average slope 2.4, width 10.5 to 3.5 m, height 11.5 to 1.6 m). Two lava falls with a total drop of 1.8 m are encountered. Volcanologically, the cave is important because it is situated near the vent of the Ai-la`au Shield, which at 1,195 m a.s.l. that yielded the last massive summit eruption of Kilauea 350 years ago and produced Kazumura Cave. e cave appears strangely dull; the typical smooth, continuous glazing normally found is missing throughout. e cave ends at a lava sump, posing a puzzle because lava seems to have upwelled from below. No other cave downhill can be linked with urston. e lava falls show that urston is not just the upper part of a larger canyon, separated by an internal, secondary roof. However, it is possibly an independent tunnel created on top of the original lava tunnel during a burst of activity of the Ai-la`au Shield and that the upwelled lava came from this lower tunnel.1. IntroductionOn Hawai`i, only a few caves are accessible to the general public. ere are two show caves (sections of Kula Kai and Kazumura caves) that oer ecologically and conservatively oriented tours for small groups and one cave in Kaumana State Park that people can visit at their own risk. In addition, Hawai`i Volcanoes National Park oers guided educational tours to a few selected caves. Otherwise, the visitor to the Hawaiian volcanoes can go underground only in urston Lava Tube (or Keanakakina, i.e. Tunnel of urston, keana meaning the cave and kakina being the Hawaiian Name of urston). It is a popular tourist attraction in the Hawaii Volcanoes National Park and visited daily by hundreds, if not over a thousand tourists. In spite of its many literature references, not much is known about its speleogenesis and previously published maps lack detail (Powers O WERS 1920; Wood OOD 1981; Halliday ALLIDAY 1982). We therefore resurveyed the cave on March 9th, 1996, in high precision by digital compass and level mounted on antimagnetic tripods to keep instruments at a xed distance from the rock (Figs. 1, 2). Using forward and backward shots served to eliminate any magnetic inuence of the rock (which is small anyway according to our survey experience in Hawaiian caves). Width and heights were recorded every 5 m. Table 1 gives the most important results. e cave was discovered in 1913. A report signed by Wade Warren ayer in the visitors book of the Volcano House says: O O n AA ug.2nd a large party headed by L.A A. urston explored the lava tube in the twin Craters recently discoered by Lo rrin urston, Jr. TT wo ladders lashed together gave comparatively easy access to the tube and the whole party, in cluding several ladies, climbed up. NN o other human beings had been in the tube, as was evidenced by the perfect condition of t he numerous stalactites and stalagmites. DDr. Jaggar estimated the length of the tube as slightly over 1,900 feet. It runs northeasterly om the crater and at the end pinches down until the oor and roof come together Halliday ALLIDAY (1997). e primary entrance to the cave today is reached across a bridge. It opens in the wall of an elongated depression called Kaluaiki, possibly a pit crater collapsed over the original magma chamber that delivered the lava producing the cave. e present exit is a roof collapse where the tourists leave the cave via stairs. e tourist section has yellow electric lighting to minimize lampenora and the oor is covered with gravel. Beyond the stairs, an open gate is found with a sign advising to carry proper lighting when visiting this 343 m long back section (more correctly 357 m).2. Volcanological importanceVolcanologically, the cave is important since it is situated


Lava Caves 676 2009 ICS P roceedings 15th International Congress of Speleology Figure 1: Map of urston Lava Tube. very near to the original vent of the Ai-la`au Shield, the site of the last massive summit eruption of Kilauea (Holcomb OLCOMB 1987) that lasted from about 500 to 350 aBP. e Ai-la`au lavas cover a very large area east of Kilauea Caldera all the way to the ocean near Hilo. ese tube-fed pahoehoe lavas contain not only the longest lava cave known (Kazumura Cave) but also a number of other very long lava tunnels (Keala Cave, John Martin Cave, Pahoa Cave). Since urston runs underneath the highest point of the Ai-la`au shield (the 3,840 foot contour), it appears to be the tube that sustained the last active ow, possibly producing the lava which reportedly invaded Kazumura (Allred L LRED and Allred L LRED 1997). urston is heading 45 N, ending just inside the park boundary. It is aiming at a prominent ow


15th International Congress of Speleology Lava Caves 677 2009 ICS Proceedings Figure 2: Longitudinal section of urston Lava Tube. Length (from the beginning of cave roof which is 13.5 m upslope of Station 18 above entrance bridge to lava sump end at Station 0) Inclined Horizontal Total cave (m) 490.84490.08 (Station 0 to Station 18= 476.576 m) Wild section (m) 357.43356.76 Tourist section (m) 133.41133.32 Total survey length (m) 531.75(total of 19 stations) Linear extent (m) 432.5 Sinuosity (490.076/432.5) 1.133 Vertical extent (m) (Station 0 at lava sump to oor at Station 18 at downslope end of bridge) -20.08 Width (m) max. 10.5min. 3.5 Height (m) max. 11.5min. 1.6 Total lava fall height (m) 1.88.96% of total vertical extent Slope () (tan-1 (20.08/476.576))* 2.413 Entrance: side of collapsed crater at ca.1195 m (3920 ) elevation End: lava sump* Because the cave roof starts earlier than the cave oor, we can use only the cave oor length, which is shorter than the total cave length, in order to calculate slope. [for comparison, length by Powers O WERS (1920): 1494 feet total (455 m), straight: 1360 feet (425m); slope 2.5]. Table 1: Survey data for urston Lava Tube.


Lava Caves 678 2009 ICS P roceedings 15th International Congress of Speleology bulge at the NE of the Shield. e upper end of Kazumura runs in parallel, slightly less than a kilometer further to the north near the highway (Allred L LRED et al., 1997). is makes is unlikely that both caves belong to the same lava ow, unless the northward turn of urston shortly before its end indicates a sharp bend in the tunnel system (Fig. 3). When comparing the sinuosity and slope of the cave with those of others in the ow eld (at least for those for which we have data) urston shows similar characteristics (Table 2). When inspecting the cave, a series of questions arise. For the casual observer the cave appears strangely dull, without many detailed features. Also the typical smooth, continuous glazing found in lava tubes is mostly missing. And nally the cave ends at a kind of lava sump, which poses quite a puzzle (Figs. 1 and 4). Figure 3: Modied clip of USGS Kilauea geological map (NEAL and LOCKWOOD, 2004). Lava Cave Total length, km Main passage length, km End-to-end length, km Sinuosity Vertical extent, m Slope deg. Volcano Kazumura Cave161.041.8632.11.301101.81.51K, A Keala Cave28.607.075.591.251861.51K, A John Martin/Pukalani System36.26 K, A Eppersons Cave41.931.130.801.41--K, A urston Lava Tube50.4900.490 0.4321.1320.12.4K, A Ainahou Ranch System67.114.82*4.271.133233.83K, A? Keauhou Trail System73.002.271.991.13213.35.36K, A?Table 2. Topographic data of some of the tubes om the Ai-la`au ows. 1: ALLRED et al., 1997; 2: KEMPE 1997; 3 4; 5: unpublished data; 6: WOOD, 1980; 7: KEMPE et al., 1997. Volcano: K, A: Kilauea, Ai-la`au. Figure 4: urston Lava Tube ends in a Chamber where the ceiling sinks below the oor that appears to consist of material up-welled om below forming a low bulge.


15th International Congress of Speleology Lava Caves 679 2009 ICS Proceedings 3. Features of the caveNevertheless that cave shows several interesting details like two lava falls (Fig. 5), below which the cave is wider and higher than above. Looking mauka (uphill) the undercutting of the former bottom sheet and of the wall linings is noticed. Ledges bending downward at the lip of the lava fall can be followed for some distance upstream, indicating that the nal ow in the cave did not ll it entirely. e cave also features ceiling cupolas of dierent sizes. Powers O WERS (1920) noted that the cupolas become larger and wider along the tube. For those nearer to the entrance he suggested that they resulted from a blow torch eect, i.e. from the melting of the primary ceiling by hot gas jets escaping from the owing lava beneath. However, the blow torches should have been moving makai (downhill) with the ow and elongated cupolas or ceiling notches should have been formed. Some of the cupolas are elongated, others not. For the cupolas further down, Powers suggested breakdown as their cause, the blocks of which have been carried out of the tube during its activity. Most of the cupolas have received a new lining and some have horizontal rims, indicating former lava stands. We found (Figs. 1, 2) seven cupolas in the rst two thirds of the tourist section and eight in the beginning of the wild section. None occur further in and they do not become wider. ere are smaller and more cylindrical and larger and more elongated cupolas in both sections. All of them occur in the center of the passage. is, and their forms, speak (at least for the cylindrical) against their origin as breakout cupolas. We suggest that they are former hornitos, vents in the primary ceiling that allowed hot gases and spatter to escape. in secondary overow, reinforcing the roof may have buried and closed them in the nal phase of the eruption. Powers O WERS (1920) suggested that the Great Hall (Figs. 1, 2), shortly before the end of the cave, is actually a window caused by breakdown of the intervening ceiling in between and up into another tube above urston, again an observation that we could not corroborate. e oor is almost devoid of ow lobes, indicative of very hot conditions when the ow stopped, not allowing sucient cooling of the surfaces skin to be rippled. Many cooling cracks are noticed in the makai section extending into the oor deeper than the thickness of the bottom sheet of the cave (which is just a few cm thick), again indicative of very hot conditions far beyond the bottom sheet of the cave. ere are also a signicant number of squeeze-ups (termed volcanoes by Powers O WERS 1920) (Fig. 6), partly related to the cracks, forming very at, glazed mounds, again indicating very hot conditions when they where extruded from the underlying lava by the expanding gas during solidication. On the walls many runners occur, partly bleeding in series out of horizontal partings in the wall. Overall, ceiling, walls and oor are irregular on the centimeter scale. e millimeter-thick, continuous, and shining glazing, so typical for most lava caves, is missing, possibly being destroyed by the ongoing degassing of the lava surrounding the cave aer the evacuation of the cave, again speaking for sustained and very hot conditions. Also the typical cylindrical lava stalactites are missing, save for short stumps. ey may, however, have been removed over the years by visitors since the initial description of the cave talks of a rich decoration (see above).4. Discussion of speleogenesisA lava sump seals the cave makai (Fig. 4) and it appears as if lava welled up from underneath (Powers O WERS 1920, Figure 5: e rst lava fall viewed mauka. e undercutting of the bottom sheet is clearly visible. Figure 6: One of many low, dome-and-cone-shaped mounds on the oor that seem to be squeeze-ups om below.


Lava Caves 680 2009 ICS P roceedings 15th International Congress of Speleology convex surface). Flow lobes or ropy textures are missing which would indicate that the ow in the cave just lled it to the roof at a low spot. us it is conceivable that urston represents an upper level of a much larger conduit system, as suggested by Halliday A LLIDAY (1982), stating that the cave is part of a Jameo System, i.e. a multi-level lava conduit. If this is so, then the two caves were above each other and not created by down-cutting and consecutive formation of a secondary ceiling separating a canyon-like tunnel. Such separations are clearly later additions and can be recognized at cross-sections (Kempe E MPE 2002). Inspection of the lava below the entrance of the cave shows that oor is not a secondary ceiling. If urston belongs to a multi-storied cave system, then it must have formed during an increase in eruption volume, exceeding the capacity of the lower tube and establishing a contemporaneous upper conduit above it, which, when lava supply subsided, fell dry and was sealed at the end by lava up-welled from the lower conduit. Another feature speaks also against the hypothesis that the oor of urston is a secondary ceiling and that is the presence of the two lava falls. ey indicate that the oor formed by active ow because these falls show clear signs of their back-cutting (Fig. 5). us, if we assume the presence of multi-storied conduits, then they must have been established by consecutive overow events, creating several caves on top of each other. Such situations have rarely been documented. Parts of Kulakai Cavern could represent such a cave type, based on the geological mapping of its surface by our group. On the far side of the collapse crater another section of cave was found, as reported by W.R. Halliday and J. Martin (Halliday ALLIDAY 1992) (Fig. 3). It has a NWSE direction, at a 90 angle from urston. Its relation to urston and to a presumed multi-story tube system remains unclear from the available map (Halliday A LLIDAY and Martin ARTIN unpublished). e correct interpretation of the nature of urston lava tube is intimately associated with the question of where the Ai-la`au vent exactly was. Holcomb O LCOMB (1987) suggests it was at the eastern notch of the Kilauea Iki collapse structure. ere vertical lava sheets are preserved. However, the topographic high is to the east of it, above urston Lava Tube (Fig. 3). erefore it is conceivable, that Kilauea Iki served as a gas vent, while a second vent produced the nal lava ows. It could have been below the Kaluaiki collapse crater. Otherwise one would need to explain how the topographic high came about. is question and some of the others posed in this paper, suggest that we do not understand the speleogenesis of urston Lava Tube very well, in spite of the fact that it may be the most visited and the most oen mentioned lava tube world-wide.References:Allred LLRED K., and C. Allred LLRED (1997) Development and morphology of Kazumura Cave, Hawaii. Journal of Cave Karst Studies, 59(2): 67. Allred L LRED K., C. Allred LLRED and R. Richards ICHARDS (1997) K azumura Cave AA tlas, Island of Hawaii Special Publications of Hawaii Speleological Survey: 81pp. Halliday A LLIDAY W.R. (1982) Kaluaiki and urston Lava Tube: An unrecognized jameo system? Proceedings of the 3rd International Symposium on Vulcanospeleology, Bend, Oregon, July 30Aug. 1: 52. Halliday A LLIDAY W.R. (1992) Mauka urston and Ash caves, Kau District, Hawaii County, Hawai`i. Cascade Caver, May-June: 37. HSS Chairmans Letter. Hawaiian Speleological Survey: 7. Halliday A LLIDAY W.R. (1997) omas A. Jaggar JR. Speleologist and Caver. Geo2, 24 (3): 87. Holcomb O LCOMB R.T. (1987) Eruptive history and longterm behavior of Kilauea Volcano. In Volcanism in Hawaii, US Geological Survey Professional Papers, 1350(1): 261. Kempe E MPE S. (1997) Lavafalls: a major factor for the enlargement of lava tubes of the Ai-la`au Shield phase, Kilauea, Hawaii. Proceedings 12th International Congress of Speleology, La Chaux-deFonds, Switzerland, 1: 445. Kempe E MPE S. (2002) Lavarhren (Pyroducts) auf Hawai`i und ihre Genese. In AA n gewandte Geowissenschaen in DDarmstadt, W. Rosendahl & A. Hoppe (Eds.), Schrienreihe der deutschen Geologischen Gesellscha, 15: 109-127. Kempe E MPE S., H. Buchas UCHAS J. Hartmann ARTMANN M. Oberwinder BERWINDER J. Strassenburg TRASSENBURG and K. Wolniewicz O LNIEWICZ ) Mapping lava ows by following their tubes: the Keauhou Trail/Ainahou Ranch Flow Field, Kilauea, Hawaii. Proceedings 12th International Congress of Speleology, La Chaux-deFonds, Switzerland, 1: 453.


15th International Congress of Speleology Lava Caves 681 2009 ICS Proceedings Powers OWERS S. (1920) A lava tube at Kilauea. Bulletin H awaiian Volcanological OObservatory, March 1920: 46. Wood O OD C. (1980) Caves of the Hawaiian volcanoes. Caving International Magazine, 6&7: 4. Wood O OD C. (1981) Caves of glass, lava tube caves of Kilauea Volcano, Hawaii. Descriptive broadsheet: photos, diagrams and full cave surveys (unpubl. rep.)


Lava Caves 682 2009 ICS P roceedings 15th International Congress of Speleology Archaeology RCHAEOLOGY and AND 14C dates DATES of OF the THE Kamakalepo AMAKALEPO /Waipouli AIPOULI / Stonehenge TONEHENGE Area REA Nahalehu AHALEHU Hawai AWAI i IST T EPHAN AN KEMPE1, HOR OR ST T VO O LKER R HEN N SCHEL2, HARRy Y Schick CHICK3, BAsil SIL HANse SE N4 1Institute of AA pplied Geosciences, University of TT echnology DDarmstadt, Schnittspahnstr. 9, DD -64287 DDarmstadt, Germany, email Kempe@geo.tu-darmstadt.de2Henschel & RRopertz, AA m Markt 2, DD -64287 DDarmstadt, Germany, e-mail: h-v.henschel@henschel-ropertz.de3General DDelivery Keaau 96749 Hawai`i, USA A4PO O Box 759 NN a`alehu, 96772 Hawai`i, USA A e archaeology of Hawaiian lava caves is poorly documented. Here we report about a small area south of Na`alehu, Hawai`i, near the coast. On aerial photos, a small outcrop of ash is clearly visible. It belongs to the agriculturally valuable Pahala Ash sites that sustained early Hawaiian populations. e area called Kamakalepo is just east of South Point, where similar soils provided for some of the earliest settlements on Hawai`i. It contains unique archaeological features both above and below ground and was studied by the authors over the last several years. We now have two 14C carbon dates setting time constrains on the time of settlement in the Kamakalepo area. A large cave system consisting of four sections of a once much longer tunnel in Mauna Loa lavas was used extensively by the native Hawaiians. e system is entered through two pukas (ceiling collapses): Lua Nunu o Kamakalepo (Pigeon Hole of the Common People) and Waipouli (Dark Waters). Both pukas give access to uphill (mauka) and downhill (makai) caves, totalling 1 km in length. Two further pukas belong to the system, Pork Pen Puka (mauka of Lua Nunu) and Stonehenge Puka (makai of Waipouli) for which no local names are known. Pork Pen Puka is a depression set into the roof of Lua Nunu Mauka Cave, the bottom of which is a secondary ceiling to the cave below. Stonehenge Puka is a large root-less vent with raed blocks around its perimeter, 60 m x 40 m wide and up to 20 m deep. Underground, the Lua Nunu caves are the ones used primarily. e main features are two large defensive parapets across the cave erected from breakdown blocks. e wall in the Makai Cave, 40 m inside the entrance, collapsed mostly but the one in the Mauka Cave, approximately 60 m into the cave, is well preserved. It is approximately 2 m high and up to 1 m thick; because it was erected on breakdown it reaches 3.7.5 m above the oor. It has a length of 25 m from wall to wall with a doorway slightly o the middle. e wall could be defended from a platform behind by throwing sling stones (polished beach pebbles) and spears. Further in, Bonk counted 102 sleeping platforms, extending well beyond the zone of light. Charcoal, seafood shells and some sh bones can be found, suggesting that the place has in fact served its purpose. In the far back of the cave, we opened a crawl, giving access to more than 100 m of additional cave. Even here we found charcoal bits on the oor, suggesting that the Hawaiians had already explored this section, albeit by a now collapsed crawl. A piece of charcoal was dated to 282 years. is places the exploration of this section of the cave at around .. 1600 (1 cal .. 1523, 2 cal .. 1495), i.e., into the time of the highest population density of the pre-discovery Hawai`i. Both of the Waipouli Caves show little signs of Hawaiian presence. Mauka, just a few bits of charcoal and seafood shells are found. e oor is too rough to be of any use. e makai part is lled by brackish water that is caped by freshwater at times of high groundwater ow. Underwater, at a depth of 10 m we recovered a whale vertebra, now dated to 524 years (1 cal .. 1327, 2 cal .. 1303). Since the bone may have been washing in from the beach, it would mark a time prior to which the last ow in the area, a black pahoehoe lava, covered the area and closed Waipouli Makai by intruding it from the downhill Stonehenge Puka.


15th International Congress of Speleology Lava Caves 683 2009 ICS Proceedings 1. IntroductionEast of Southpoint, Hawaii, and south of Na`alehu is an outcrop of agriculturally valuable Pahala Ash forming a lemon-shaped area on aerial pictures. It is one of the sites that sustained early Hawaiian populations (Kirch, 1985). e area under investigation (Fig. 1), called Kamakalepo, contains unique archaeological features both above and below ground (Bonk, 1967; Kempe, 1999; Kempe et al., 2006a). Figure 1: Situation of Kamakalepo area with important landmarks, caves, and their GPS xes (dots). Kind Male Female undecided Simple stick man, hands down 3A2;2A3;2A4; 5A5; 3A7; =15 2A2;2A3;1A4; 4A5; 1A6; 2A7; 2A10; =14 1A3, 3A3; 1A6; 4A9; 4A10; = 13 Simple stick man, one hand up 1A2; 2A7;1A10; = 41A5 Stick man, legs spread 1A1; 2 A7 3A7; 3A10 Triangular or square bodies 1A1; 2A2;1A3; 1A6; 1A8; = 6 Full head, double line body 2A8 3A4; 1A7; 4A8 1A6; 8A8 Filled frontal 1A6 Lateral views with outlines 1A4? Lateral views lled bodies 1A7 1A7 Monkeys 1A6 1A8 Rectangular basins 2A1 Others (many lines but unclear) 3A7Table 1: Classication of petroglyphs (Petroglyph Valley, Kamakalepo) (read: 3A2 = 3 specimens in Area A2).


Lava Caves 684 2009 ICS P roceedings 15th International Congress of Speleology A large cave system of a once much longer tunnel in Mauna Loa lavas (Kempe et al., 2006b) was used extensively by the native Hawaiians. It consists of four sections, that can be entered through two ceiling collapses (pukas): Lua Nunu o Kamakalepo (Pigeon Hole of the Common People) now overgrown by acacia shrubs and Waipouli (Dark Waters). Both of these pukas give accesses to uphill (mauka) and downhill (makai) caves, all-in-all 1 km in length (Table 1 and Figs. 5 to 8 in Kempe et al. 2006b) Two further pukas belong to the system, Pork Pen Puka (mauka of Lua Nunu) and Stonehenge Puka (makai of Waipouli) for which no local names are known. Pork Pen Puka is a depression set into the roof of Lua Nunu Mauka Cave, the bottom of which is a secondary ceiling to the cave below. Stonehenge Puka is a large root-less vent with raed blocks around its perimeter, 60*40 m wide and up to 20 m deep (Kempe et al., 2006b).2. Usage of cavesUnderground, the caves of the Lua Nunu are the ones used primarily (Figs. 2 and 3). An old, now mostly obliterated path led down from the northeast rim with the other sides overhanging. Within the puka small outcrops of Pahala Ash exist, possible former eld plots or agropits. Retaining walls are found at both entrances providing for level ground, on which foundations of huts are found. In both caves large defense walls were erected by stacking breakdown blocks. e wall in the Makai Cave, 40 m from the entrance, collapsed mostly but the one in the Mauka Cave, Figure 2: Map of Lua Nunu o Kamakalepo Mauka Cave. Note archeological details.


15th International Congress of Speleology Lava Caves 685 2009 ICS Proceedings approximately 60 m into the cave, is well preserved. It has all the characteristics of a medieval defense wall: It is about 2 m high and up to 1 m thick. Because it was erected on breakdown, it reaches 3.7 and 5.5 m above the oor (Fig. 4). It stretches from wall to wall and, due to its convex-mauka curvature, reaches a length of approximately 25 m (the cave being 23 m wide and 14 m high in its center). A doorway, slightly o center, admits access and platforms behind the wall permit the defenders to throw sling stones and spears at any attacker. Sling stones (wave-worn pebbles) are still found on the oor. While the defenders would stand in the dark, the attackers would be outlined by daylight providing good aims. Behind the wall, Bonk (1967) counted 102 sleeping platforms that extend well into the zone of complete darkness. Charcoal, seafood shells and a few sh bones can be found everywhere, suggesting that the place has in fact served its purpose. Artifacts were collected in 1908 by Meineke and 1967 by Bonk. In the far back of the cave, we opened a crawl, giving access to more than 100 m of additional cave. Even here we found a few charcoal bits on the oor, suggesting that the Hawaiians had already explored this section, albeit by a now collapsed crawl. A piece of charcoal was dated to 283 a BP. is places the exploration of this section of the cave at around .. 1600 (1 cal .. 1523, 2 cal .. 1495), i.e. into the time of the largest population density of pre-discovery Hawaii. Similar dates have now been obtained for charcoal from the Pa`auhau Civil Defence (Kempe et al., 2003): Hd 26237: 270 aBP; 1 cal AD 1524-1792, 2 cal .. 1495-1951; and Hd-26276: 211 a BP; 1 cal .. 1650, 2 cal .. 1643) also indicating dates centered to late preor early post-contact occupation times. Underground fortications have been described from other caves on Hawaii (e.g., Kennedy & Brady, 1997). An elaborate example is the Cave of Refuge on the Hakuma Horst (Kalapana, Puna District). ere the defense function was obtained by narrowing the entrance to the cave to a crawlway that could be entered by attackers only one at a time (Kempe et al., 1993). La Plante (1993) reported about fortications (defense walls, fortied crawlways) from the Puna District (most probably Pahoa Cave) without giving details about locations or constructional dimensions. Small Figure 3: Map of Lua Nunu o Kamakalepo Makai Cave. Note archeological details. Figure 4: View mauka of the 25 m long defense wall in the Lua Nunu o Kamakalepo Mauka Cave. Note persons for scale and gate at the center of the wall.


Lava Caves 686 2009 ICS P roceedings 15th International Congress of Speleology defense walls, now crumbled seem to have protected the cave passages below Keala Pit as well (Kempe & KetzKempe, 1997). More data probably exist in internal reports of various agencies without ever having been published. e Lua Nunu o Kamakalepo Makai Cave has also been fully explored by the Hawaiians. Platforms and re places extend almost 100 m into the cave. At the makai end, the black phoehoe lava that intruded the puka secondarily (see Kempe et al., 2006b) forms a separate, less than a meter high cave. It was also entered by the Hawaiians as bits of charcoal on the oor reveal. We found its entrance closed articially by rocks, probably to hide the entrance of this chamber of last refuge (see Secrete Hall on map, Fig. 3). Both of the Waipouli Caves show little signs of Hawaiian presence. In the mauka sections, just some charcoal is found and a few bits of seafood remains. e oor is too rough to be of use. e makai part is lled by a brackish water lake that is caped by freshwater at times of high groundwater ow. We found one large beach stone on the steep entrance slope and a whale vertebra mid-lake at a water depth of about 8 m Both of the Waipouli Caves show little signs of Hawaiian presence. Mauka, just a few bits of charcoal and seafood shells are found. e oor is too rough to be of any use. e makai part is lled by brackish water that is caped by freshwater at times of high groundwater ow. Underwater, at a depth of 10 m we recovered a whale vertebra (Fig. 5), now 14C dated to 524 a BP (1 cal .. 1327, 2 cal .. 1303) (data courtesy of Dr. Bernd Kromer, Heidelberg). Since the bone may have been washed in from the beach, it would mark a time prior to which the last ow in the area, a black phoehoe lava, covered the area and closed Waipouli Makai by intruding it from the downhill Stonehenge Puka (Kempe et al., 2006b). e water was pumped up in the 20th Century for cattle. A concrete platform at the entrance is all that is le. Also, further makai an over 20 m deep well was dug through the cave roof and the water was pumped up by a winddriven pump. Part of its collapsed trestle fell into the well and landed in the water, where it now forms interesting rusticles. Stonehenge Puka was also used by Hawaiians: Its southern wall is overhanging and providing a natural shelter. Here a few very small platforms were erected.3. Above-ground usageAbove ground, the area shows many signs of usage as well. First of all there is a beach stone paved path, giving access to the area from the west (mauka). e area south of Lua Nunu is covered by ash and could have been used for agriculture, explaining the presence of the underground settlement. At the western rim of the ash plain, on the overlaying bare lava, we found two small heiaus, compact stone platforms used either for dwelling huts or religious purposes (Fig. 1 for location). e Pork Pen Puka has stone walls along its perimeter and throughout its centre, suggesting that it was used to keep pigs. At the eastern side of the ash outcrop, there is a rectangular structure build from phoehoe plates which probably also was a pen either for pigs, or for goats and cows if erected aer contact. Nearby, a shallow cave was found, showing also signs of occupation. Paths connected the Lua Nunu with Waipouli (mostly overgrown now) and led towards the coast from Waipouli eastward. At the end of the path a large carbonate-cemented beachrock was placed, obviously a well-visible signal to guide the traveler to the beginning of the path across the Waipouli a`a.4. Petroglyphs Within the studied area, three petroglyph sites were found. e one furthest south has mostly animal gures. e second one, north of Stonehenge, is composed of postcontact petroglyphs: It displays a pentagram, a large cross made from ve squares each of it inscribed with a + and a X, and a saber with a two line inscription reading: KA IEIE PALA and IKA UA NOE (the Mellow IeIe, a plant, and Strong Misty Rain; possibly the names of two lovebirds). To the north, at the seaward end of a shallow valley, ten groups with almost a hundred petroglyphs occur within an Figure 5: Weathered whale vertebra om the deeper part of the lake in Waipouli Makai Cave.


15th International Congress of Speleology Lava Caves 687 2009 ICS Proceedings area of 50*50 m. ere simple stickman occur next to more complicated full body pictures, both in frontal as in lateral views. Two of the larger gures have long tails, possibly pictures of monkeys (one of them clearly a male specimen) (Fig. 6), thus placing the petroglyphs into the early postcontact time. Marks made by sharpening tools occur as well as many pound marks, some almost obliterating some of the glyphs. A total of 92 glyphs were identied that distribute among several types (Table 1). Many dierent styles are present: e group of simple stickmen with arms and legs bend at right angles dominates; male and female glyphs occur with a similar frequency (Fig. 7; Table 1 rst line). One of the male stickmen has two lines extending down its head, like indicating long hair. Five stickmen have one hand raised as if in greeting. A few stickmen have simple spread legs like in an inverted Y. e triangular-bodied gures appear all without a penis and could therefore reasonably be labeled as female. e gures with open circle heads and a double-lined body have a variety of hands, mostly with three ngers, but one even has ve ngers and toes. Interesting are gures shown in side-view, among them a large gure in Area 8. e two ape-like glyphs are among the largest. One, with a penis, is shown laterally (A6), the other (A8) in frontal view with a long thin tail between its legs. Otherwise, no other animal pictures are seen, except a possible goat head (triangle and two curved lines). Overall, the site seems to be restricted (with the exception of the two monkeys and the goat) to glyphs of humans, both female and males. Circular depressions and rings are missing, so prominent in other Hawaiian petroglyph sites, and in spite of the proximity to the sea, no marine animals are depicted (Cox & Stasack, 1977). Area 9 features a vertical slab which is pounded upon forming a spot about 1 m in diameter (Fig. 8); the surfaces Figure 6: e picture of a male primate being with a long tail, possibly a monkey brought by sailors to the island in the early post-contact period. Figure 7: Examples of simple stickmen petroplyphs, right with a penis (male) le without (female).


Lava Caves 688 2009 ICS P roceedings 15th International Congress of Speleology of the inclined slabs below are also heavily abraded. Both slabs contain traces of almost erased stickmen. We interpret this area as a sling-stone practice target. Behind the slabs, an approximately 5 m long cave extends, which contains four bamboo poles of unknown age. e valley is also heavily impacted by Hawaiian quarrying: all along the rims of the valley the upper lava layers have been dug up, partly down to 2 to 3 m, and piles of broken rocks litter the perimeter of the quarries. uarrying has been going on also in the area between the Petroglyph Valley and Lua Nunu. Many of the sites display longitudinal grooves caused by grinding. What exactly the rock was quarried for remains unknown since no intermediate products were noticed.5. Conclusions e archeological evidence specically the number of sleeping platforms behind the defense walls suggests that the Kamakalepo area sustained a sizeable population. At peak times it may have counted hundred or more people. Clearly the area was still settled in early postcontact times as illustrated by petroglyphs of monkeys, a saber, a Christian (?) cross and an inscription. Writing was introduced to the islands only aer 1820. e only directly accessible water in the area is the lake in Waipouli. Paths leading towards it suggest that it was used by the Hawaiians intensively, in spite of the fact that not much archeological evidence is found inside. Any stairways or walls may have been obliterated either by later rock fall or by the farmers in the early 20th century. is water supply is, however, treacherous and in times of drought the water turns brackish, salty enough to make it even unt for cattle. In times of drought drip water in the caves ceases also, which is, in other areas of the island, a major source of water (compare Martin, 1993; Kempe & KetzKempe, 1997). erefore the Kamakalepo settlement may have been sustainable under a dierent climate condition, such as during the Little Ice Age in the 17th and 18th century, when more groundwater may have been available. e dated charcoal (1 cal .. 1523-1656, 2 cal .. 14951793) from one of the caves substantiates this interpretation. Because the dated whale bone was found deep inside the lake of Waipouli it may not be an artifact at all. Its age (1 cal .. 13271439, 2 cal .. 13031451) may mark the time when it was washed into the Waipouli tunnel from the sea. us it indirectly gives a date aer which the tunnel was closed by the black phoehoe ow that invaded and closed the seaward section of the Waipouli tunnel via Stonehenge Puka.References:Bonk, W.J. (1967) Lua Nunu o Kamakalepo: A cave of refuge in Ka`u, Hawai`i. Internal report, unpublished: 75. Cox, J.H. and E. Stasack, (1977) Hawaiian Petroglyphs. Bernice P. Bishop Museum Special Publications 60, 100 pp. Kempe, S. (1999) Waipouli and Kamakalepo, two sections of a large and old Mauna Loa Tube on Figure 8: is near-vertical slab features multiple pound-marks, possibly a training target for shooting sling-stones.


15th International Congress of Speleology Lava Caves 689 2009 ICS Proceedings Hawai`i. Abstract, NSS Convention 1999, Vulcanospeleological session. And: J. Cave Karst Stud. Nat. Speleol. Soc. 62 (April 2000) (1): 43. Kempe, S., C. Ketz-Kempe, W.R. Halliday, and M.S. Werner, (1993) e Cave of Refuge, Hakuma Horst, Kalapana, Puna District, Hawai`i. Pacic Stud. 16(2): 133. Kempe, S. and C. Ketz-Kempe, (1997) Archaeological observations in lava tubes on Hawai`i. Proc. 12. Intern. Congr. Speleol. Aug 10, 1997, La Chauxde-Fonds, Switzerland, Vol. 3: 13-16. Kempe, S., I. Bauer, and H.-V. Henschel, (2003) e Pa`auhau Civil Defence Cave on Mauna Kea, Hawaii, a lava tube modied by water erosion. J. of Cave and Karst Studies 65(1): 76. Kempe, S., H.-V. Henschel, H. Shick, and B. Hansen, (2006a) Archaeology of the Kamakalepo / Waipouli / Stonehenge area, underground fortresses, living quarters and petroglyph elds. Proc. 12th Intern. Symp. on Vulcanospeleology, Tepotzln, Mexico, July 2, 2006, Assoc. for Mexican Cave Studies, Bull. 19 and Socieded Mexicana de Exploraciones Subterrneas Bol. 7: 243. Kempe, S., H.-V. Henschel, H. Shick, and F. Trusdell, (2006b) Geology and genesis of the Kamakalepo Cave System in Mauna Loa picritic lavas, Na`alehu, Hawaii. Proc. 12th Intern. Symp. on Vulcanospeleolpgy, Tepotzln, Mexico, July 2-7, 2006, Assoc. for Mexican Cave Studies, Bull. 19 and Socieded Mexicana de Exploraciones Subterrneas Bol. 7: 229. Kennedy, J. and J.E. Brady, (1997) Into the netherworld of island earth: a reevaluation of refuge caves in ancient Hawaiian society. Geoarchaeology 12(3): 641. Kirch, P.V. (1985) Feathered Gods and Fishhooks, an Introduction to Hawaiian Archaeology and Prehistory. University of Hawai`i Press, Honolulu, 349 pp. La Plante, M. (1993) Recently discovered Hawaiian religious and burial caves. Proc. 6th Intern. Symp. Volcanospeleol., Hilo, 1991: 7. Martin, J. (1993) Native Hawaiian water collection systems in lava tubes (caves) and fault cracks. Proc. 6th Intern. Symp. Volcanospeleol., Hilo, 1991: 10.


Lava Caves 690 2009 ICS P roceedings 15th International Congress of Speleology Jordanian ORDANIAN lava LAVA caves CAVES and AND their THEIR importance IMPORTANCE to TO understand U NDERSTAND lava LAVA plateaus PLATEAUSStephan Kempe1, AA hmad AA l-Malabeh2 & Horst-Volker Henschel3 1Institute of AA pplied Geosciences, University of TT echnology DDarmstadt, Schnittspahnstr. 9, DD -64287 DDarmstadt, Germany, Kempe@geo.tu-darmstadt.de2Hashemite University, DDepartment of Earth and Enironmental Sciences, P.O O Box 150459, Zarka 13115, Jordan, a_ malabeh@yahoo.com;3Henschel & RRopertz, AA m Markt 2, DD -64287 DDarmstadt, Germany, h-v.henschel@henschel-ropertz.de e Arabian plate is covered by large Cenozoic (Oligocene-uaternary) basalt elds, the Harrats, over a north-south distance of about 3000 km from Jordan and Syria through Saudi Arabia to Yemen with an estimated volume between 103 and 105 km3. ese are among the largest basalt plateaus worldwide. We study the Jordanian Harrat Al-Shaam, the most northern of these plateaus that covers about 45,000 km2. It forms a gently undulating plateau dotted with tephra cones, shield volcanoes, pressure ridges and crossed by a few, up to 80 km long dikes. e, with < 4, northwest-southeast dipping plateau drops from approximately 1100 m to 700 m at Al-Mafraq and to 550 m asl in the Al-Azraq area. It forms a succession of ow sheets, the youngest of these is over 400 ka old (Al-Fahda area). e Harrat is covered by a 1-2 m thick loess layer that has been washed into the depressions forming playas (locally known as Qa). In these lavas we explored, surveyed, and studied a total of 17 lava caves since September 2003. 2824 m of passages were surveyed as of spring 2008. Nine of these are lava tunnels, six are pressure ridge cavities and two are of doubtful origin. e discovery of these lava tunnels is surprising considering their old age and the fact that the loess is easily washed into caves lling them eventually. e presence of the lava tunnels underscores the fact that the Harrat consists to a large part of tube-fed pahoehoe, thus explaining its overall low slope.1. Introductione Arabian plate is covered by seven larger and several smaller Cenozoic (Oligocene-uaternary) basalt elds, the Harrats. ey stretch over a north-south distance of about 3000 km from Jordan and Syria through Saudi Arabia to Yemen. e estimated volume of eruptive material equals to between 103 and 105 km3. ese wide-spread, poorly studied basalt elds are considered to be among the largest of predominately alkali-olivine basalt plateaus in the world (e.g., Al L Malabeh ALABEH 1994). Our group studies the lava caves contained in the Jordanian section of the Harrat Al-Shaam, the 700 km long, most northern of these plateaus that covers about 45,000 km2 (approximately 25% of the Arabian Harrats) (Fig. 1). is Harrat is in Jordan approximately 220 km wide in the N and 30 km in the south. Geomorphologically it forms a gently undulating lava plateau dotted with prominent tephra cones, low shield volcanoes, numerous pressure ridges and crossed by a few, up to 80 km long eruptive ssures. e plateau generally dips to the south and southeast, starting at an elevation of approximately 1100 m (asl) along the Syrian border and dropping down to 700 m at Al-Mafraq and to 550 m in the Al-Azraq area. e overall slope is at most 4. e structure of the basalt plateau is a succession of ow sheets which form stepped clis along wadi walls or faults. e youngest of these ows are over 400 ka old (Al-Fahda area) (Tarawneh ARAWNEH et al., 2000). It, and the other younger lava elds, does not show wadi incision yet, while the older ow series are heavily incised. e Harrat is covered by a 1 m thick loess layer that has been washed into the depressions forming playas (locally known as Qa) giving the less incised areas a mottled appearance.2. Lava CavesIn these lavas we explored, surveyed and studied a total of 17 lava caves since September 2003. 2824 m of passages were surveyed as of spring 2008 (KEMPE et al., 2008) (Table 1). Of the total 1,486 m, or close to 53%, was surveyed in September 2005, among them the 923.5 m long Al-Fahda Cave, currently the longest cave in Jordan. Eight of the lava caves are lava tunnels. One cave (Treasure Pit) is pit dug by treasure hunters that probably leads into a sediment-lled lava tunnel. Six caves are pressure ridge cavities and two caves (Beer Al-Wisad and Uwaiyed) are of unusual origin. 2.1 Lava tunnels


15th International Congress of Speleology Lava Caves 691 2009 ICS Proceedings Al-Fahda (the lioness) Cave (Al L Malabeh ALABEH et al., 2006) was named aer the local name for one of the youngest lava elds (KAr age 0.46    0.01  Ma sample HAS-7; Tarawneh ARAWNEH et al., 2000) in the Harrat. It was rst mentioned without any speleological details by Helms E LMS (1981, p.138) as El-Mughara in connection with the investigation of the famous Bronze Age desert city Jawa. Helms described a channel that leads to the entrance of the cave, apparently dug in an attempt to store water in times of plenty in the cave for times of need. is channel led to the rediscovery of the cave by the second author, who followed it from Wadi Rajil (830 m asl) Figure 1: Map of the Harat As Shaam Lava eld in the north of the Arabian plate. Name of Cave Type Length mDepth mHyena presence Al-Fahda Cave Lava Tunnel 923.5 6.7 +++ Al-Badia Cave Lava Tunnel 445.0 17.2 ++ Hashemite University CaveLava Tunnel 231.1 10.0 Al-Ameed Cave Pressure Ridge 208.0 4.0 ++ Dabi Cave Lava Tunnel 193.6 1.8 +++ Abu Al-Kursi East Lava Tunnel 153.7 12.2 ++ Kempe Cave Lava Tunnel 139.4 11.3 +++ Al-Howa Lava Tunnel 97.1 10.8 Obada Cave Pressure Ridge 90 3.4 ++ Al-Hayya Cave Pressure Ridge 81.3 4.2 +++ Abu Al-Kursi West Lava Tunnel 77.1 8.1 ++ Azzam Cave Pressure Ridge 44.1 4.2 Al-Raye Cave Pressure Ridge 42.0 3.5 Dahdal Cave Pressure Ridge 28.9 0.0 + Beer Al-Wisad Pit (unknown) 11.4 11.5 Uwaiyed Cave Upward stooping (?) 11 3.1 +++ Treasure Pit Tunnel ? 7.2 5.8 Total 2,824Table 1: List of presently (spring 2008) surveyed lava caves in Jordan (altered aer KEMPE et al., 2006b).


Lava Caves 692 2009 ICS P roceedings 15th International Congress of Speleology in the north downslope to the main entrance (730 m asl) (Al L Malabeh ALABEH et al., 2006) and surveyed by us in 2005 (Table 2; Fig. 2). e cave is also known under the name of Khsheifa Cave and was surveyed in parallel by Frumkin R UMKIN et al. (2008) yielding astonishingly similar results (their length 920 m). e cave has a very low slope, according to our survey of about 0.7. Such a low slope is typical for tube-fed pahoehoe ows. Al-Fahda Cave is unusually wide but very low and has Stations Horizontal Length mStations m 2-54a Main survey downslope488.60 End-to-end (as the crow ies) 684.0 8-11 Back of entrance 18.68 Sinuosity (771.03/684) 1.13 19-22To second entrance14.462-54aVertical (entrance to deepest point)-6.74 50-51aW-passage of terminal split28.4571-54aVertical extent of Main Passage-8.41 4-5 Connection 6.0571-54a Horizontal length 755.12 5-71 Upslope passage 266.21 Slope 1 slope () (tan-1 (8.41/755.12)* 0.64 67a-79Mahmouds Test Passage.101.07 Slope 2 slope () (tan-1 (8.41/684)0.70 Total 923.52Width Maximal at St. 8 17.5 Main Passage length Minimal at St. 64 3.55 4-54aDownslope passage482.86 Mean of main passage (39 stations)7.51 4-67a Upslope passage 187.10 Height Maximal St. 14 4.67 67a-79Mahmouds Test Passage101.07 Mean of main passage (39 stations) 1.21 Total 771.03Table 2: Survey results of Al-Fahda Cave (AL-MALABEH et al., 2006). Figure 2: Map of Al-Fahda Lava Cave, longest lava cave in Jordan.


15th International Congress of Speleology Lava Caves 693 2009 ICS Proceedings a very at oor (at least were the rock oor is visible) (Fig. 3). It appears that this was caused by a later invasive ow lling the lower part of the tunnel. It shows a blocky surface and ends in two ow lobes shortly before the cave itself ends. It remains unclear if this ll is autochthonous, i.e, generated within the tunnel as a terminal slump of a higher viscosity, or allochthonous, i.e. caused by an invasion of a later ow of the Al-Fahda ow eld through a ceiling hole (above the current accessible section of the cave). e main entrance to the cave today is through a late central ceiling collapse, exposing a cross section through the primary roof. It is composed of two relatively thick layers, in contrast to other caves that have up to 12 sheets in the primary ceiling (in case of Abu AlKursi) (for a more detailed study of the cave see Al L Malabeh A LABEH et al., 2006). Hashemite University Cave is speleologically interesting also; it is reached through a collapse hole at the crest of a ridge. ere the primary, 7 m thick roof is exposed consisting of only three pahoehoe layers. e uphill passage running NW is blocked by breakdown but from the north another low passage lled with sediment joins. e downhill tunnel is 180 m long before it opens up to a nearly circular room approximately 20 m in diameter. ere, the cave ends in a lava sump. In a way, this is similar to the terminal lava sump of urston Lava Tube (see Kempe E MPE et al., this volume). It poses a structural riddle since one would expect that the back-up of the residual ow in the tunnel should close the cave at a narrow point but not at a wide passage. One possible solution could be assuming that the oor is a secondary ceiling (Kempe E MPE 2002). A blowhole, situated near station 26, indicates that there could be an open passage underneath, giving some credibility to this hypothesis. In case of Kempe Cave, we can identify the source volcano for the rst time in Jordan (Fig. 4) (Kempe E MPE et al., 2008). It is a low shield volcano in between larger stratovolcanoes. e crater is 120 m across and the rim is very even, suggesting that it once held an overowing lava lake. e slope between crater (976 m asl) and cave (936 m) is only 1.2. e cave itself (Fig. 5) is very low and curves around in half-circle; it is the most sinuous among the caves yet explored. Due to the fact that it is the cave furthest east and therefore the driest it also contains unusual speleothems, among them curvy gypsum owers.2.2 Pressure ridge cavesA group of caves not showing any clear direction of slope nor any signs of horizontal ow, is grouped as pressure Figure 3: View up-slope into Al-Fahda Cave om the entrance. Note at, sediment coered oor and low and wide character of the tunnel (persons for scale). Figure 4: Geological map of the Kempe Volcano (Map by Al-Malabeh and Kempe based on Field observations and Google Earth images).


Lava Caves 694 2009 ICS P roceedings 15th International Congress of Speleology ridge caves. ey can be quite long (Fig. 6, map of AlAmeed Cave), are very wide and low in general and can have several branches, petering out at their ends. Similar caves are kwon from Hawaii, but are not well documented. Pressure ridge caves apparently form when half-solidied surface sheets possibly yield to the shoving of the hotter lava below by doming upward, oen with axes perpendicular to the direction of pressure. e caves are however, not bound to pronounced tumuli put can occur under low, dome-like rises.2.3 Other lava cavesUwaiyed Cave is a circular 10 m wide chamber in highly weathered old basalt that may be caused by upward stooping of a hypogene, collapsed limestone cave at depth (Kempe EMPE et al., 2009) (Fig. 7). Another (Beer AlWisad) one is an 11 m deep pit, also in very old lava, of unknown origin. 3. Conclusionse discovery of so many lava tunnels in the Harrat is surprising considering their old age and the fact that the loess is easily washed into caves lling them eventually. Al-Fahda, Hashemite University, Dabie, Kempe and the two Abu Al-Kursi Caves are all closed by sediments. Only Al-Howa Cave is terminated on both ends by roof collapse due to the loading of a later a`a lava ow. Al-Fahda, Al-Badia (Beer Al-Hamam), and the two Abu Al-Kursi Caves are rather wide, while Al-Howa, Hashemite University, Kempe and Dabie Caves are of smaller dimensions. All have very low gradients. Lava falls and plunge pools, so oen encountered in Hawaii (Kempe EMPE 1997; Kempe EMPE this volume), were not found in these caves. A secondary ceiling is possibly present only in Hashemite University Cave. Benches and shelves marking older ow levels occur in Dabie Cave, Al-Fahda and in one place in Hashemite University Cave. Branching is rare, apart from Al-Fahda Cave only Hashemite University and possibly Kempe Cave display branching. e presence of the lava tunnels underscores the fact that the Harrat consists to a large part of tube-fed pahoehoe, thus explaining its overall low slope. Compared to Hawaiian tunnels (see data in Kempe E MPE 2002; Kazumura, Keala and Huehue, some of the longest caves on Hawaii have sinuosities of 1.30, 1.25 and 1.2), most caves show a rather Figure 5: Map of Kempe Cave.


15th International Congress of Speleology Lava Caves 695 2009 ICS Proceedings low sinuosity (Al-Fahda: 1.13), in spite of the fact that it has a lower slope than the mentioned Hawaiian caves (1.51, 1.51, 4.58 respectively). e hypothesis that there should be a reverse relation between slope and sinuosity can, therefore, not be substantiated. e winding of the cave should have provided for a alweg, i.e. a path along which the lava ow was maximal with slip-o and undercut slopes to the sides depending on curvature. e high proportion of pressure ridge caves and their length are another interesting nding. One of the reasons for this high proportion of caves not formed by underground linear ow of lava may be the low slope of the terrain being in places even below 1. Many of the caves (compare Table 1) have been used by hyenas, wolves, foxes and porcupines. Specically hyenas le many bones of their prey, abundant coprolites, dens dug into sediment and scent marks (Kempe E MPE et al., 2006a). e caves therefore are also of high paleontological and taphonomic importance.References:Al L Malabeh ALABEH A. (1994) Geochemistry of two volcanic cones from the intra-continental plateau basalt of Harrat El-Jabban, NE-Jordan. Geochemical Journal 28: 517. Al L Malabeh ALABEH A., M. Frehat REHAT H.V. Henschel ENSCHEL S. Kempe EMPE (2006) Al-Fahda Cave (Jordan): the longest lava cave yet reported from the Arabian Plate. Proc. 12th Intern. Symp. on Figure 6: Map of Al-Ameed Cave.


Lava Caves 696 2009 ICS P roceedings 15th International Congress of Speleology Figure 7: Map of Beer Al-Wisad. Vulcanospeleology, Tepotzln, Mexico, July 2, 2006, Assoc. for Mexican Cave Studies, 19 and Socieded Mexicana de Exploraciones Subterrneas 7: 201. HELMS, S.W. (1981) Jawa, Lost City of the Black Desert. Methuen, London, 270 pp. Ibrahim B RAHIM K., and A. Al L Malabeh ALABEH (2006) Geochemistry and volcanic features of Harrat El Fahda, a young volcanic eld in northwest Arabia, Jordan. Journal of Asian Science, 127(2): 127. Frumkin R UMKIN A., M. Bar AR Matthews ATTHEWS A. Vaks AKS Paleoenvironment of Jawa basalt plateau, Jordan, inferred from calcite speleothems from a lava tube. uaternary Research 70: 357. Kempe E MPE S. (1997) Lavafalls: a major factor for the enlargement of lava tubes of the Ai-laau Shield


15th International Congress of Speleology Lava Caves 697 2009 ICS Proceedings phase, Kilauea, Hawaii. Proc. 12th Intern. Congr. of Speleology, La Chaux-de-Fonds, Switzerland 1: 445. Kempe E MPE S. (2002) Lavarhren (Pyroducts) auf Hawaii und ihre Genese. In: Angewandte Geowissenschaen in Darmstadt, W. Rosendahl & A. Hoppe (Eds.), Schrienreihe der deutschen Geologischen Gesellscha 15: 109. Kempe E MPE S., A. Al L Malabeh ALABEH D. Dppes PPES M. Frehat REHAT H.V. Henschel ENSCHEL W. Rosendahl O SENDAHL (2006a) Hyena Caves in Jordan. Proc. 12th Intern. Cave Bear Symposium in essaloniki/Loutra, Scitic Annals, School of Geology, Aristotle Univ. of essaloniki, Spec. Vol. 98: 201. Kempe E MPE S., A. Al L Malabeh ALABEH M. Frehat REHAT H. V. Henschel ENSCHEL (2006b) State of lava cave research in Jordan. Proc. 12th Intern. Symp. on Vulcanospeleology, Tepotzln, Mexico July 2-7, 2006, Assoc. for Mexican Cave Studies, Bull. 19 and Socieded Mexicana de Exploraciones Subterrneas Bol. 7: 209. Kempe E MPE S., A. Al L Malabeh ALABEH H.V. Henschel ENSCHEL (2008) Kempe Cave: an unusual, meandering lava tunnel cave in NE-Jordan. Proceedings 13th Intern. Symposium on Volcanospeleology, Jeju Island, Korea, Sept 1-5, 2008: 38. Kempe E MPE S., A. Al L Malabeh ALABEH M. Frehat REHAT H.V. Henschel E NSCHEL (2009): Point karstication in the southern Jordanian desert: A sign of oileld degassing? Submitted to Acta Carsologica. Tarawneh A RAWNEH K., S. Ilani LANI I. Rabba ABBA Y. Harlavan ARLAVAN S. Peltz ELTZ K. Ibrahim BRAHIM R. Weinberger EINBERGER G. Steinitz TEINITZ (2000) Dating of the Harrat Ash Shaam Basalts Northeast Jordan (Phase 1). National Reserve Authority; Geological Survey Israel.


Lava Caves 698 2009 ICS P roceedings 15th International Congress of Speleology POSSIBLE LAVA TUBE DEVELOPMENT IN THE SAN LUIS VALLEY AND SOUTHEASTERN SAN JUAN MOUNTAINS, COLORADODODO UGLA A S MED D VILLE, HA A ZEL MED D VILLE 11762 Indian RR idge RR d., RReston VA A 20191, medville@verizon.net Entrances to what appear to be lava tube caves are observed on two Pliocene shields in south-central Colorado: the Culebra Volcano on the Taos Plateau volcanic eld in the San Luis Basin and Los Mogotes; a shield in the southeastern San Juan Mountains, 42 km to the west. e Culebra volcano is a 2 km2 isolated shield on the east side of the Rio Grande River. Entrances to ten short lava tube segments up to 20 m in length occur along the base of two 8 to 10 m high escarpments on the west side of the shield. e chemical composition of four groundmass samples taken from inside the caves indicates that the tubes are in a silicic alkaline basalt (SiO2: = 52.9%, Na2O + K2O: = 5.2%). Although estimated to be age-equivalent to the 3.4-3.9 Ma Servilleta basalt to the south and east, the age of the Culebra Volcano has not been determined. Based on their position with respect to the Rio Grande River and the exposure of the host rock on the opposite bank of the river, the tubes appear to predate the draining of historic Lake Alamosa (440 ka) and to have been exposed following incision of the modern Rio Grande into the ow. Entrance to possible lava tube segments up to 10 m in length are also observed in basaltic rock on the anks of a 270 km2 shield volcano (Los Mogotes) in the southeastern San Juan Mountains. e entrances are exposed in a sequence of up to 12 escarpments in 4.44.75 Ma Hinsdale trachybasalts on the south side of the shield where the Conejos River has exposed individual ow units. Deformed rock around entrances, oval shaped passages, and the presence of a feature resembling a levee in one of the cave segments may be evidence of genesis by owing lava. A K/Ar date of 4.4 Ma was obtained by USGS for an exposure 1.25 km west of the entrances and at the same elevation. If this date is representative of the age of the ow episode containing the entrances, the observed lava tube remnants could be among the oldest in the U.S.1. Introductione Taos Plateau volcanic eld in the southern San Luis Valley and the adjacent southeastern San Juan Mountains, both in south-central Colorado, contain several late Tertiary (Pliocene) to early uaternary (Pleistocene) shield volcanoes composed of silicic alkalic basalts, andesitic basalts, and trachybasalts. Entrances to what appear to be lava tube caves are observed on two of these shields; the Culebra Volcano on the Taos Plateau and Los Mogotes; a shield 42 km to the west in the southeastern San Juan Mountains. e locations of these features are shown in Figure 1 with the town of Alamosa in the central San Luis Valley shown at the top of the gure.2. Culebra Volcanoe Culebra Volcano is a 2 km2 isolated shield on the east side of the Rio Grande River, 6 km south of the San Luis Hills and on the Costilla Plain of the Taos Plateau. is low shield was named by BURROUGHS (1971) and is labeled as a Tertiary olivine andesite on the map by THOMPSON and MACHETTE (1989). e summit is at elevation 2,357 m and rises only 70 m above the Costilla Plain. e age of the Culebra Volcano is not known but is postulated in THOMPSON et al. (2007) to be age equivalent to the Pliocene Servilleta basalt (3.54.2 Ma). Flows from this small shield are exposed in two escarpments, each up to 10 m in height along the west side of the volcano, where the Rio Grande River has incised into the ow, exposing it on both sides of the river in a two kilometer long and 20 m deep canyon labeled e Box on topographic maps. e eastern escarpment contains entrances to ten short cave segments up to 18 m in length. e entrances to the largest of these segments are 45 m wide and 34 m high (Fig. 2) and are at the base of the lowest cli, only 12 m above river level. e passages in these caves parallel each other, are perpendicular to the river, and slope upward toward the top of the shield, where within only 20 m they become choked with rock fall and animal midden. A third short cave, shown in Figure 3, is 7 m in length and is found at the base of an upper escarpment, directly above the two entrances shown in Figure 2. e directional orientation of this short segment is the same as the lower caves.


15th International Congress of Speleology Lava Caves 699 2009 ICS Proceedings In Figure 2, deformation of the host rock around entrances can be observed. Liquid lava owing through a molten core may have exerted lateral pressure on the surrounding viscous lava, resulting in plastic deformation of this slowly cooling material and ination of the malleable crust immediately above the tube as noted in KAUIHIKAUA et al. (1998). is may explain the curvilinear texture of the ow cross section seen around several of the cave entrances as illustrated in Figure 4. An alternative interpretation (Kauahikaua, pers. comm.) is that lava owed around preexisting material which was subsequently removed following exposure of the entrances by the downcutting Rio Grande River, resulting in a mold having the form of the preexisting object. e caves do not exhibit many of the ow related features usually observed in lava tubes; perhaps as a result of weathering and spalling of original wall rock near the entrances and passage modication resulting from back ooding of the Rio Grande into the caves. Away from the river and toward the accessible end of the largest two caves, the passage morphology becomes more tube-like with smooth curved ceilings (Fig. 5). As a result of midden ll, soil blown into the caves, and deposition of sediments by the Rio Grande River, the original bedrock oors of the caves cannot be observed, hindering interpretation.2.1. Groundmass compositione composition of the Culebra Volcano ows have been variously described as an olivine andesite in Figure 1: Map of south-central Colorado showing Los Mogotes and the Culebra Volcano. Figure 2: Possible lava tube entrances on Rio Grande River below Culebra shield olcano. Figure 3: Low entrance to cave at base of upper ow unit on east side of e Box.


Lava Caves 700 2009 ICS P roceedings 15th International Congress of Speleology THOMPSON and MACHETTE (1989), and andesitic in BURROUGHS (1971). To determine the chemical composition of the rock, four groundmass samples were taken, two from the interior of caves in the lower escarpment and the other two from the interiors of caves in the upper escarpment. LUISZER (2008) used X ray uorescent spectroscopy to identify the constituents of each sample. Results obtained indicate little variation between upper and lower escarpment samples. e alkalinity (Na2O + K2O) and silica content (SiO2) for each sample indicate that the tubes are in a silicic alkalic basalt/trachybasalt with mean SiO2 percent weight of 52.9% and mean alkalinity (Na2O + K2O) of 5.2%. is material is more viscous than Hawaiian basalts.2.2. Cave ageTo determine the age of the ow in which the tubes are found, 40Ar/39Ar radiometric dating was carried out on a wall sample taken from the largest of the tubes. As of the preparation of this paper, results were not available, but are expected to indicate an age that is comparable to that of the nearby and more widespread 3.4 Ma to 3.9 Ma Servilleta basalt as suggested in THOMPSON et al. (2007). If correct, the Culebra Volcano would be considerably older than the nearby Mesita Hill shield, 5 km to the southeast, where a 40Ar/39Ar age of 1.03 +/0.01 Ma was obtained by APPELT (1998). e distal end of the ow in which the caves are found is buried beneath uaternary deposits on the west side of the Rio Grande River, where a 12 m escarpment and 50 m wide surface exposure above it are seen. e ow predates the modern Rio Grande River, which occupied its current course as a result of the draining of a 105 km long, and 48 km wide Pleistocene paleolake (Lake Alamosa), dammed by the San Luis Hills, 6 km to the north of the Culebra Volcano as described in MACHETTE et al. (2007). e elevation of shoreline deposits and calcic soils on barrier bars indicate a maximum elevation of 2330 m for Lake Alamosa and overow through a gap in the San Luis Hills and onto the Costilla Plain around 440 ka, well aer the emplacement of the Culebra Volcano ows and its lava tubes. Bisection of the ow by the modern (post440 ka) Rio Grande exposed the observed lava tubes and beheaded them at their lower ends since there is no evidence of continuations of any of the tubes on the west side of the river. 3. Los MogotesLos Mogotes is an east-dipping 130 km2 shield volcano in the southeastern San Juan Mountains, 42 km. SW of the Culebra Volcano. Basaltic rocks on the Los Mogotes summit (elev. 3010 m asl) have been dated by APPELT (1998) with 40Ar/39Ar dates of 4.75 +/0.28 Ma obtained. e eastward owing Conejos River, 490 m below the summit, has cut across the south ank of the shield, exposing a sequence of at least 12 ows as a series of 6to10-m high clis (Fig. 6) in late Cenozoic Hinsdale Formation silicic alkalic basalts and basaltic andesites, described by LIPMAN and MEHNERT (1975). Figure 4: Curved/deformed beds around entrance below Culebra shield olcano. Figure 5: Curved ceiling at back of cave below Culebra shield olcano. Figure 6: Exposed beds of basaltic rock on south side of Los Mogotes.


15th International Congress of Speleology Lava Caves 701 2009 ICS Proceedings Numerous openings are observed in these clis with typical cross sections being circular or oval shaped. Passages can only be followed for a few meters before being choked by cemented rounded blocky material, possibly ll from younger ow episodes at higher elevations on the shield. ese openings may be remnants of lava tubes dating back to emplacement of the Hinsdale Formation basaltic rocks. e largest of the cavesLos Mogotes Cavehas an entrance that is 5 m wide and 4 m high and can be followed for 10 m to a rock choke (Fig. 7). A laterally protruding n of rock that can be seen on the right wall of this short cave segment resembles a levee as described and illustrated in LARSON (1993): a free standing lateral remnant of a lava tongue or ow caused by cooling along the edges and subsequent evacuation. If this is the case, it provides evidence for genesis of the cave by owing lava. A second cave entrance is in a bed 8 m below the one containing the Los Mogotes Cave and is few meters oset from the former. A 1.5 m high and 2 m wide entrance leads to an alcove and a curved passage segment 12 m long, ending at a choke in blocky, ashy material. e bedrock around several of the entrances on the south side of Los Mogotes has the same appearance as seen around the Culebra Volcano entrances: deformation of the host rock around the entrances. 3.1. Groundmass compositionLos Mogotes basalts are both silicic and alkalic. e composition of eleven samples taken from this shield is plotted in Figure 8 of LIPMAN and MEHNERT (1975). ese have a mean SiO2 content of 51.6% and a mean Na2O + K2O content of 4.9%. For this study, a sample was taken from the wall of Los Mogotes Cave and another from the wall of a 12-m long cave in the ow unit immediately below the one containing the Los Mogotes Cave. X ray uorescent spectroscopy analysis of the composition of the two samples (LUISZER, 2008) indicates that the host rock is a trachybasalt having a mean SiO2 content of 49.97% dry weight and a mean Na2O + K2O content of 6.23% dry weight; somewhat more alkalic and less silicic than the samples analyzed in LIPMAN and MEHNERT (1975). 3.2 Cave ageAlthough 40Ar/39Ar dating was not carried out for this study, LIPMAN (1975) provided a K/Ar date of 4.4 Ma from an exposure 1.25 km west of the entrances and at about the same elevation as the Los Mogotes cave entrances. If the entrances and short passage segments seen on the south side of Los Mogotes are remnants of lava tubes and if, as appears to be the case, some of them are in the same bed as the dated sample, these features could be among the oldest lava tubes in the United States.4. ConclusionsOpenings to what appear to be short and truncated lava tube segments in late Pliocene basaltic rocks (3.5 Ma to 4.4 Ma) are observed on the anks of two shield volcanoes in south Central Colorado. ese are the rst lava tubes to have been documented in this State and appear to be among the oldest in the United States.ReferencesAppelt PPELT R.M. (1998) 40Ar/39Ar geochronology and volcanic evolution of the Taos Plateau volcanic eld, northern New Mexico and southern Colorado: Socorro, New Mexico Institute of Mining and Technology, M.S. thesis, 58 pp. BURROUGHS, R.L. (1971) Geology of the San Luis Hills, South-Central Colorado in NN ew M exico Geological SocietyTT wenty Second Field Conference, Guidebook of the San Luis Basin, Colorado, 277. KAUAHIKAUA, J., K. Cashman, T. Mattox, C. Heliker, K. Hon, M. Mangan, and C. ornber (1998) Observations on basaltic lava streams in tubes from Kilauea Volcano, island of Hawai`i. Journal of G eophysical RResearch 103, No. B11, 27,303,323. Larson A RSON C.V. (1993) An Illustrated Glossary of Lava Tube Features, Western Speleological Survey Bulletin No. 87, 56 pp. Lipman I PMAN P.W. (1975) Geologic Map of the Lower Conejos River Canyon Area, Southeastern San Juan Mountains, Colorado. U.S. Geological Survey Figure 7. Five meter wide, four meter high entrance to Los Mogotes Cave with deformed beds around the entrance.


Lava Caves 702 2009 ICS P roceedings 15th International Congress of Speleology Miscellaneous Investigation Series Map I-1901. LIPMAN, P.W. and H.H. MEHNERT (1975) Late Cenozoic Basaltic Volcanism and Development of the Rio Grande Depression in the Southern Rocky Mountains. G eological Society of AA merica Memoir 144, 119. Luiszer U ISZER F.G., University of Colorado, Boulder, Department of Geological Sciences (2008) Results of X Ray Fluorescent Spectroscopy on Submitted Basalt Samples, e-mail communication. Machette A CHETTE M.N., D.W. Marchetti ARCHETTI and R.A. Thompson H OMPSON (2007) Ancient Lake Alamosa and the Pliocene to Middle Pleistocene Evolution of the Rio Grande. Chapter G in 2007 R Rocky Mountain Section Friends of the Pleistocene Field TT rip uaternary Geology of the San Luis Basin of Colorado and N N ew Mexico, 157. Thompson H OMPSON R.A., and M.N. Machette A CHETTE (1989) Geologic Map of the San Luis Hills Area, Conejos and Costilla Counties, Colorado. U.S. Geological Survey Miscellaneous Investigations Series Map I1906. Thompson H OMPSON R., M. Machette ACHETTE R. Shroba HROBA and C. Ruleman ULEMAN (2007) Geology of Mesita VolcanoEruptive History and Implications for Basin Sedimentation during the uaternary. Chapter H in 2007 R Rocky Mountain Section Friends of the Pleistocene Field TT rip uaternary Geology of the San Luis Basin of Colorado and NN ew Mexico, 169.


15th International Congress of Speleology Lava Caves 703 2009 ICS Proceedings SEVENTEEN YEARS BENEATH HUALALAI: A SUMMARY OF CAVING EXPERIENCEDODO UGLA A S and HA A ZEL MED D VILLE, PET T ER R and ANN ANN BO O ST T ED D DON DON COON OON S, DA DAVE BUNN NN ELL, N N EVIN N and J UD D Y DA DAVIS, BO O B R R ICHARD ARD S, JO O HN N RO RO SEN N FELD D ST T EVE SMIT T H, B ERNARD RNARD and SAND AND Y SZUKA A LSKI Hawaii Speleological Survey 11762 I ndian RR idge RR d., RReston, VA A 20191 Hualalai, a shield volcano on the western side of the island of Hawai`i, rises to an elevation of 2521 m and has a surface area of about 751 km2. In the seventeen year period 1992-2009, 485 lava tube caves have been recorded beneath 30 of the 173 Hualalai lava ows and over 111 km of passages have been surveyed in 203 of these caves. e Hualalai lava tubes are found in ows ranging from just over 200 years to over 10,000 years in age and vary greatly in pattern from simple linear conduits to multi-level braided and branching complexes that present exploration and survey challenges. Hualalai contains some of the worlds longest and most vertically extensive lava tubes, including the deepest single pit in the United States (Na One, 263 m deep), the 24 km Hualalai Ranch tube complex with 452 m of relief, the 10.8 km Hu`ehu`e Cave in the historic 1801 ow with 498 m of relief, Pueo Cave, a highly braided lava tube containing over 6 km of passages, and the 10.7 km Lama LuaKa`upulehu cave complex extending over a linear distance of 5.4 km and having a relief of 370 m. e surveyed distance from the Lama Lua entrance to the end of one of its passages is 2.65 km and is entirely in darkness; this is perhaps the greatest traversable distance in total darkness for any lava tube on earth. Above 1200 m elevation, the gradients of the ows increase to up to 15 degrees and several long and vertically extensive lava tubes have been surveyed in upland dry forests. Notable caves include: (a) Umi`i Manu, the second most vertically extensive lava tube on Hawaii, extending for a linear distance of 3.4 km and having a vertical relief of 570 m, (b) Ambigua Cave, having a linear extent of 1.3 km and a vertical relief of 306 m, and (c) Manu Nui, a steep and decorated lava tube having a surveyed length of 3.7 km and a vertical relief of 352 m. 1. IntroductionHualalai, a shield volcano on the western side of the island of Hawai`i (Figs. 1-A, 1-B) reaches an elevation 2521 m (8271 feet) and covers an area of 751 km2 (290 mi2); about 7.2% of area of the island. A majority of the lava ows on Hualalai are Holocene in age (10,000 years bp or younger), a few are Pleistocene in age (>10,000 years bp) and one is historic. e historic ow occurred in ca. 1800 and produced two ow elds; the Hu`ehu`e ow eld seen one km. north of the Kona International airport and the somewhat older Ka`upulehu ow eld 8 km. to the north. e Hu`ehu`e eld in turn, consists of two historic ows: the primarily pahoehoe Manini`owali ow and the primarily a`a Puhi-a-Pele ow. Of the 173 Hualalai ows that contain some pahoehoe, only 30 have been examined for caves. However, beneath these 30 ows, 485 caves have been found of which 203 have been surveyed. e combined survey length of these caves is 111 km (69 miles). e caves are found at all elevations on Hualalai and cave temperatures range from 30 C on the Pacic coast to 20 C on the upper slopes of Hualalai.2. Signicant Lava Tubes on Hualalaie lava tubes found on the slopes of Hualalai are among the longest and most vertically extensive on earth with ve of these caves having a vertical extent of over 300 meters. Several of the more notable caves on Hualalai are described below, starting with caves at and just above sea level to those found at elevations of up to 2000 meters above sea level on Hualalais upper slopes.2.1 Under the Wall CaveLocated a few km south of the Kona International airport, this ne unitary tube extends for a linear distance of 2.5 km and has a vertical range of 125 meters. e upper part of the cave consists of a 10 meter wide and 6 meter high tunnel containing a variety of man-made features including walls,


Lava Caves 704 2009 ICS P roceedings 15th International Congress of Speleology ramps, and platforms. A description and map of the cave are provided in MEDVILLE (2002).2.2 e Hu`ehu`e (Manini`owali) Cavee Huehue Cave, found in the historic 1801 ow 2 km north of the Kona International airport, was surveyed in the 1992-96 time frame by teams of American and German caver/speleologists led by D. Medville and Prof. S. Kempe, respectively. e cave and its geology are documented in OBERWINDER (1996), MEDVILLE and MEDVILLE (1997), and KEMPE and OBERWINDER (1997). Although braided in places, the cave usually consists of a single large conduit. e cave originally consisted of four separate segments but in a series of three trips made in March 1997 by S. Kempe, M. Oberwinder, D. and H. Medville, and W. Storage, the segments were connected, resulting in a single tube with 10.8 km. of surveyed passage and 498 meters of vertical relief extending over a linear distance of 6.1 km. e cave is about 8 meters beneath the surface and contains passages that are generally 5-6 meters in width and 4-5 meters in height (Fig. 2). 2.3 Hualalai Ranch Cave Systemis is a huge distributary system that contains over 24 km of passages in a 3,000-5,000 year old ow and as such, is the largest lava tube known to exist on Hualalai. Although the source of the ow containing this cave is a ventPuu Alauawa at elevation 660 meters asl, the cave is terminated by a local highway (Route 190) at its upper end at elevation 634 meters. Here, a single passage can be followed downhill for over 450 vertical meters, branching into a multi-level complex of parallel passages. At its mid-section, the cave contains numerous parallel passages up to 20 meters beneath the ow and extending laterally for 600 meters across the ow. e cave has a linear extent of over 4 km. and was explored and surveyed in the 1997-2007 time period by J. Rosenfeld, N. and J. Davis, P. Carter, J. Wilson, R. Pacheco, B. Liebman, D. Coons and many others. e project is described in ROSENFELD (2001) and DAVIS (2003).2.4 e Ka`upulehu/Lama Lua ComplexAdjacent to the historic (1800-1801) Ka`upulehu a`a ow and on its eastern margin, a 1,500-3,000 year old ow contains another complex of large caves. Although 19.2 km of passages have been surveyed in 58 caves, half of the total (10.5 km.) is found in two long, aligned caves extending over a linear distance of nearly 6 km. e upper of the caves, Lama Lua, has 5.5 km. of passage and four entrances, three of which require vertical equipment to enter. e cave was Figure 1-A: Map of Hawai`i showing location of Hualalai Volcano (credit: USGS). Figure 1-B: Hualalai Volcano, Hawai`i (credit: USGS). Figure 2. Passage in Hu`ehu`e Cave (Photo by Dave Bunnell).

PAGE 100

15th International Congress of Speleology Lava Caves 705 2009 ICS Proceedings explored and surveyed in the 1997-2004 time frame by N. Davis, J. Davis, J. Rosenfeld, M. Kramer, P. Carter, D. Medville, H. Medville, G. Clemmer, D. Bunnell, R. Pacheco, and T. Lappin. A description and map of the cave are provided in MEDVILLE and DAVIS (2007). e highest of the entrances is a collapse pit, 10 meters in diameter and 10 meters deep. is leads to a 15 meter wide and 10 meter high passage nearly 20 meters beneath the surface. is passage can be followed down the ow for nearly 3 km to lower entrances. A parallel passage 700 meters below the Lama Lua entrance (Fig. 3) extends to the ENE for over 2 km before ending at a lava seal. e surveyed distance to this point from the Lama Lua entrance is 2.65 km., entirely in darkness and possibly the greatest known distance in darkness from an entrance for any lava tube on earth. e two parallel passages in Lama Lua are 550600 meters apart, comparable to the lateral extent of the Hualalai Ranch complex. e lower of the two caves; the Ka`upulehu System, has nearly as much surveyed passage (5 km.) and is a continuation of the Lama Lua Cave, separated from it by two shattered rock rings 275 meters apart. e shatter rings are a result of lling of the tube beneath and a breakout to the surface with subsequent draining of lava back into the tube as described in KAUAHIKAUA et. al. (1998). is cave extends from the lower of the two shatter rings to within 200 meters of the coastal highway over a linear distance of 2.2 km. e cave was explored and surveyed by D. and H. Medville in 1993-97.2.5 Kiholo Bay CavesA 3,000-5,000 year old ow crosses the coastal highway (Route 19) at the Kiholo Bay scenic viewpoint at mile post 82. is ow contains a concentration of 105 caves in an 18.5 km2 area. During the time period 1992-2009, 43 of these caves were surveyed, the longest of which has a length of 1.3 km. Unlike the ows described above, this ow does not contain a single massive conduit but rather numerous shallow (4-6 meters beneath the surface) and shorter caves that parallel each other with no apparent overall pattern. e caves are found between sea level and an elevation of 250 meters with several of the sea level caves containing pools of brackish water. e most notable of these is informally called Peles Water Cave and contains a sea level pool extending for mm meters (Fig. 4). Another such cave, the Keanalele Water Hole, is found on the shore of the Pacic Ocean and is a local tourist attraction. 2.6 Pueo Cave In the vicinity of Pu`u Wa`awa`a and at an elevation of 500 meters asl, a 3,000-5,000 year old ow extends for nearly 9 km. toward the ocean. e caves in this ow are shallow: only 3-5 meters beneath the surface and contain numerous small entrances. ese caves however, are highly braided and complex, perhaps indicating lava owing in multiple diverging and recombining lobes before the molten cores of these lobes were evacuated with a resulting braided tube remaining. e largest of the tubes in this ow, Pueo Cave, contains over 6.5 km. of passages in an area less than 0.2 Figure 3: Passage in Lama Lua (Photo by Nevin Davis). Figure 4: Water Cave aboe Kiholo Bay (Photo by Dave Bunnell).

PAGE 101

Lava Caves 706 2009 ICS P roceedings 15th International Congress of Speleology km2. e cave contains over 40 loops and is one of the most complex of the Hualalai caves. It was surveyed in 2004-2007 by D. and H. Medville, N. and J. Davis, and S. Smith and is described in MEDVILLE (2008). A similar cave (Two Owl Cave) is 150 meters below Pueo Cave but unconnected to it and contains over 2 km of surveyed passage.2.7 Mauka Hualalai: Umi`i Manu and Ambigua CavesHigher on Hualalai the gradients of the ows increase and the tubes tend to be vertically extensive. One example of such a tube is Umi`i Manu, literally Bird Trap, named aer the numerous skeletal remains of the extinct Hawiian ightless goose, collected in this cave by ornithologists from the University of Hawai`i and the Smithsonian Institution. is goose was three to four feet in height and the bones have been radiocarbon dated to 500-900 years bp. e upper end of the cave is at an elevation of 1890 meters and is only 300 meters below the source vent for the ow containing the cave. At this elevation the average gradient on Hualalai is 13 degrees. e cave extends for a linear distance of nearly 3.4 km and has a vertical extent of 570 meters, arguably the second most vertically extensive lava tube known, aer Kazumura Cave. Umi`i Manu consists of a single large conduit, generally 5 meters wide and high and follows the steepest gradient of the ow for its entire length. In addition to the goose bones, the cave contains the skeletal remains of other birds, including rails, petrels, and nenes. e cave is a good example of a high gradient conduit with very little meandering or braiding of passages. An outline map and summary are provided in MEDVILLE (2003). Another example a few km to the north is Ambigua Cave where the local gradient is about ten degrees. As is the case with Umi`i Manu, Ambigua Cave is generally a single large conduit; it has a vertical extent of 306 meters and a surveyed length of 1.96 km. A map and description are provided in SZUKALSKI (2008).2.8 Below Kaupulehu Crater: Manu Nui and Bealls CaveAlso on the steep upper anks of Hualalai and in an area receiving over 90 cm. of rainfall per year are two other extensive and complex cave systems: Manu Nui and Bealls Cave. Manu Nui, surveyed by P. and A. Bosted, K. Marcellius (the owner), D. Coons, and others, is an extremely colorful, complicated, and steep tube system containing over 3.5 km of passage and having a vertical extent of 347 meters. e cave is braided and contains long, dagger-like stalactites (Fig. 5). A discussion of the caves exploration and the genesis of its lava stalactites is given in BOSTED, A and P. BOSTED (2009). Somewhat lower in elevation (800 meters asl) and in an adjacent prehistoric ow, Bealls Cave contains over 2.1 km of passage extending over a vertical range of 105 meters. As a result of the humid tropical climate at this elevation, Bealls Cave is wet and contains numerous colorful slimes. e cave was surveyed in the period 2004-2006 by D. and H. Medville, N. and J. Davis, and others.3. Pit Craters and Open Vertical Volcanic ConduitsIn addition to lava tubes, Hualalai is known for the numerous pit craters and open vertical volcanic conduits along its ri zones and on its summit area. Some of these are described below. e historic Ka`upulehu ow eld contains ultramac xenolith nodules, the primary constituents of which are clinopyroxene-olivine and olivine. ese nodules are embedded in the walls of several pits having vertical extents of up to 30 meters. ese pits, at an elevation of 1000 meters Figure 5: Passage in Manu Nui (Photo by Peter and Ann Bosted).

PAGE 102

15th International Congress of Speleology Lava Caves 707 2009 ICS Proceedings asl, were explored and surveyed by W.R. Halliday and others in the mid-1990s and are interpreted as breakouts from a lava tube that transferred the nodules from near-vent regions. e source of the 1800-01 Hu`ehu`e ow, Puhi-a-Pele consists of an aligned series of hollow spatter cones/spatter vents containing vertical openings, the deepest of which has been explored by J. Rosenfeld, N. Davis, P. Carter and others to a depth of 130 meters beneath the local surface via a series of elongated pits. is vertical cave has not been followed to an ultimate bottom due to the decreasing width of the passage, a lack of rigging points (many rebelays and redirections are required), and generally incompetent rock. At an elevation of 1870 meters asl on the NW ri zone of Hualalai, the Ka`upulehu crater is a 70 meter diameter vent within which a 5 meter diameter inner pit descends to a depth of 112 meters. e pits base consists of a chamber, 15 meters wide and 30 meters long with a downsloping rubble covered oor. e pit was descended and surveyed by P. Carter and N. Davis in February 2004. e deepest of the Hualalai pits, Na One, is on the volcanos southeastern ri at an elevation of 1850 meters. A combination pit crater and open vertical volcanic conduit, Na One has a vertical extent of 263 meters. e entrance is a 180 meter wide pit crater, roughly 100 meters deep. An opening in the oor of this pit leads to the top of a 160 meter deep sha. e debris oor of the chamber at the bottom of this pit is 30 meters wide and 60 meters long. When rigged with a Tyrolean traverse across the top, Na One can be descended in a single free drop, the deepest pit known in the United States and among the deepest volcanic pits on earth (Figure 6). ere has only been one descent of Na One, in February 1994 by K. Allred and D. Coons.4. SummaryHualalai Volcano in West Hawai`i, contains one of the worlds greatest concentrations of long and vertically extensive lava tubes and volcanic pits. Although over 110 km of passages have been surveyed in over 200 caves, including ve having a vertical extent of over 305 meters (1,000 feet), a large majority of this shield volcano has yet to be investigated. New discoveries, exploration, and surveys are expected for many years to come. ReferencesBOSTED, A. and P. BOSTED (2009) Exploration of Manu Nui Lava Tube, Hawai`i, USA Proc. 13th Int. Congress of Speleology, Kerrville TX, July 2009. DAVIS, N. (2003) e Maturing of the Hualalai Ranch Caves Survey Project, Hawai`i Speleological Survey Newsletter No. 13, Spring 2003, page 11. KAUAHIKAUA, J. K., K. Cashman, T. Mattox, C. Heliker, K. Hon, M. Mangan and C. ornber (1998). Observations on basaltic lava streams in tubes from Kilauea Volcano, island of Hawai`i. Journal of G eophysical RResearch, 103, No. B11, 27,303-27,323. KEMPE, S. and M. OBERWINDER (1997) e Upper Huehue Flow (1801 eruption, Hualalai, Hawaii): An example of interacting lava ows yielding complex lava tube morphologies. Proc. 10th Intern. Congr. Speleol. 10-17 Aug. 1997. MEDVILLE, D. (2002). Under the Wall Cave. Hawai`i Speleological Survey Newsletter, No. 11, Spring 2002, pp. 3-12. MEDVILLE, D. (2003) Umi`i Manu, Hawai`i Speleological Survey Newsletter No. 13, Spring 2003, pp 34-35. MEDVILLE, D. (2008) e Survey of Pueo Cave, Pu`u Wa`awa`a Ahupua`a, Hawai`i Speleological Survey Newsletter No. 23, Spring 2008, pp. 16-24. MEDVILLE, D. and N. DAVIS (2007) e Exploration and Survey of the Lama Lua SystemNorth Kona, Hawai`i. NSS News Vol. 65, No. 8, Aug. 2007, pp.10-18. MEDVILLE, D. and H. MEDVILLE (1999). e Exploration and Survey of Hu`ehu`e Cave, NSS News Vol. 57 No. 2, Feb. 1999. Figure 6: Na One on Hualalai (Photo by Dave Bunnell).

PAGE 103

Lava Caves 708 2009 ICS P roceedings 15th International Congress of Speleology OBERWINDER, M. (1996) Genese und interne Struktur des oberen Teiles des Lavastromes von 1801. MS thesis, University of Keil. ROSENFELD, J. (2001) January-February 2000 Hualalai Ranch Cave Expedition, Hawai`i Speleological Survey Newsletter No. 9, June 2001, pp. 29-30. SZUKALSKI, B. (2008). Ambigua Cave, Hawai`i Speleological Survey Newsletter No. 24, Fall 2008, pp. 3-14.

PAGE 104

15th International Congress of Speleology Lava Caves 709 2009 ICS Proceedings Composition OMPOSITION of OF bacterial BACTERIAL mats MATS in IN El L Malpais ALPAIS National ATIONAL Monument ONUMENT New EW Mexico EXICO USA: comparison COMPARISON and AND contrasts CONTRASTS with WITH bacterial BACTERIAL communities COMMUNITIES in IN Hawai AWAI i I lava LAVA tubes TUBESMONic IC A MOy Y A1, MATThew HEW G. GARci CI A1, Mich ICH Ael EL N N Spil PIL De E2, and Di DI ANA E. N N ORThup HUP1 1DDepartment of Biology MSC03 2020, University of NN ew Mexico, AA lbuquerque, NN M 87131 USA A2Institute of Meteoritics, MSC03 2050, University of NN ew Mexico, AA lbuquerque, NN M 87131 USA A Cave bacterial mats cover the walls of lava tubes around the world, including the lava tubes in New Mexico, yet little is known about their bacterial composition and role in the ecosystem. We undertook a study of the dierently colored bacterial mats found in Pahoehoe, Four Windows, and Roots Galore Caves in El Malpais National Monument (ELMA), located southwest of Grants, New Mexico. To determine the bacterial community composition and the phylogenetic relationships of the bacterial mats found in these three caves, we aseptically sampled bacterial mats found in the twilight and dark zones of each cave. Bacterial DNA was extracted and puried, the 16S rRNA gene was amplied using polymerase chain reaction (PCR), and approximately 1400 bases were sequenced from clone libraries. Bacterial identities of the closest relatives were found using Ribosomal Database Project II and BLAST, while a maximum parsimony phylogenetic tree was constructed using PAUP. Our results reveal the presence of a diverse bacterial community comprising the dierently colored lava tube mats that includes members related to the AA ct inobacteria, Gammaproteobacteria, AA lphaproteobacteria A A cidobacteria, Firmicutes and Chloroexi.. ere exists common overlap in bacterial communities across the three cave sites, but most notably within the Actinobacteria, the bacterial group that produces many of the antibiotics in use today. In comparison with a parallel study in Hawaiian lava tubes, ELMA microbial mats were less diverse, but overlapped in some phyla present. Our studies show that there is less diversity in yellow bacterial mats than white bacterial mats in both the New Mexican and Hawaiian lava tubes. Putative AA ct inomyces were found among the AA ct inobacteria, which suggests that heterotrophy occurs in these lava tubes. In addition, putative N N i trosococcus were found among the Gammaproteobacteria, suggesting that ammonia oxidation may also occur. Our studies are shedding light on the nature of these communities and their possible roles in the ecosystem. 1. Introductione colorful mats that exist in caves and lava tubes all over the world are known to be microbial thanks to culture studies, scanning electron microscopy, and cultureindependent molecular phylogenetic techniques. Until recently scientists used only culture-based techniques to study microorganisms in environments such as caves. Researchers have assumed that the microbial mats in lava tubes are primarily composed of actinomycetes. However, our preliminary studies have revealed many new microbial species waiting to be identied in these mats. ese unidentied species could have some useful medicinal value as has been shown in actinomycetes. Some types of actinomycetes are medicinally and culturally signicant because they excrete antibiotic products to repel invaders (Lazzarini et al. 2000). e antibiotic properties of many bacteria species make them promising biotechnology targets. Humid lava tube caves contain highly visible mats of bacteria and other microorganisms, nicknamed lava wall slime, but they have been studied even less than limestone caves (Northup and Welbourn 1997; Northup et al. 2008). Howarth (1981) has suggested that nutrient recycling (e.g. nitrogen) occurs in the microbial mats. Ashmole et al. (1992) have found microbial mats present in humid caves in the Canary and Azores Islands, but never in caves lacking moisture. Staley and Crawford 1975 have found microbial mats consisting of dierent species of bacteria, including actinomycetes in the genus Streptomyces in research done in lava tubes in Washington We have had limited success in culturing microorganisms from the environment, including caves, using standard microbiological media (Northup et al. 1994; Amann et al. 1995; Hugenholtz et al. 1998). Molecular phylogeny, using the 16S ribosomal rRNA gene, has revolutionized our understanding of the great diversity and distribution of life present in the environment (Pace 1997). Many

PAGE 105

Lava Caves 710 2009 ICS P roceedings 15th International Congress of Speleology novel bacterial and archaeal species have been detected as a result of this new technology in a variety of environments. Bacteria have been found in some of the most extreme areas including deep-sea hydrothermal vents, kilometers below the surface of the Earth in rock, and in caves. ese microorganisms are important participants in the precipitation and dissolution of minerals in caves (Northup and Lavoie 2001; Barton and Northup 2007) and in a variety of surface settings (Ehrlich 1999). However, researchers have barely begun to characterize the microbial diversity of caves and the roles of microorganisms in the subsurface. Additionally, we know little of what abiotic factors control the lava tube microbial diversity. We feel that investigating these mats, using culture-independent techniques, will provide valuable insights into the nature of these communities and what determines their diversity. is study intends to identify many novel bacterial species that inhabit the walls and ceiling of several El Malpais lava tubes and compare them to the parallel study of the bacterial communities in Hawai`i lava tubes (Garcia et al., this volume). is study will advance our knowledge of the dierences and similarities found among lava tubes with varied age ow, surface conditions and dierent colored bacterial mats. 2. Methodse three lava tubes sampled at El Malpais National Monument, Four Windows, Pahoehoe, and Roots Galore, occur in the Bandera lava ow, which is approximately 10,000 years old. El Malpais National Monument is located southwest of Grants, New Mexico, USA. Age of ow and average area rainfall data for Hawaiian and El Malpais lava tubes were ascertained using a variety of online resources and from previous investigations (Laughlin and Woldegabriel, 1997). We recorded entrance elevation and GPS coordinates, and cave temperature and humidity were measured using an IMC Digital ermometer probe. Small samples of wall rock covered with bacterial mats were collected from the three El Malpais Lava tubes under a National Park Service collecting permit. Samples were covered with sucrose lysis buer to preserve the DNA and transported to the lab where they were stored in a -80C freezer until DNA extraction. Yellow and white microbial mats were sampled from Pahoehoe and Roots Galore Caves, and white and gold microbial mats from Four Windows Cave, in the El Malpais National Monument. DNA was extracted and puried using the MoBio PowerSoil DNA Isolation Kit using the manufacturers protocol (MoBio, Carlsbad, CA). Extracted DNA was amplied with universal bacterial primers 46 forward (5GCYTAAYACATGCAAGTCG-3) and 1409 reverse (5-GTGACGGGCRGTGTGTRCAA-3)(Vesbach, personal communication). Amplicons were cleaned and Figure 1: Parsimony tree of bacterial clone sequences om a yellow microbial mat om Pahoehoe Cave in El Malpais National Monument. Numbers on the branches indicate bootstrap values om 1000 re-samplings and indicate the degree of support for this tree topology.

PAGE 106

15th International Congress of Speleology Lava Caves 711 2009 ICS Proceedings puried using the Qiagen PCR cleanup kit (Qiagen, Germantown, Maryland) and were cloned using the TOPO TA Cloning kit (Invitrogen, Carlsbad, CA) and sent to Washington University Genome Sequencing Facility for sequencing with primers M13F and M13R. Once received, sequences were edited and contiged with Sequencher 4.8. (Gene Codes, Ann Arbor, Michigan). To check the orientation of our sequences and to convert from antisense to sense, OrientationChecker (www.cardi.ac.uk/biosi/ research/bioso/) was used. Chimeras were detected using the Mallard soware (http://www. bioinformatics-toolkit. org/Mallard/). Rarefaction curves were generated using Dotur (http://schloss. micro.umass.edu/soware/ dotur.html) to ascertain whether sequencing had detected a comprehensive set of community members. Sequences were analyzed using BLAST (NCBI; Altschul et al. 1997) to identify closest relatives. Initial alignment was completed with Greengenes (greengenes.lbl.gov/) and manually corrected using BioEdit editor (http://www.mbio. ncsu.edu/BioEdit/bioedit.html), guided by 16S primary and secondary structure considerations. Parsimony analysis was performed using PAUP (version 4.0b10, distributed by Sinauer; http://paup.csit.fsu.edu/) and bootstrap analyses were conducted on 1000 re-sampled datasets using PAUP. Samples of the lava tube wall rock covered with microbial colonies were examined on a JEOL 5800 scanning electron microscope (SEM) equipped with an Oxford (Link) Isis energy dispersive x-ray analyzer (EDX). Rock samples with adherent bacterial colonies were mounted directly on an SEM sample stub while in the cave and then coated by evaporation with Au-Pd in the lab prior to imaging.3. Results and Discussion e phylogenetic analysis of yellow microbial mats from Pahoehoe and Roots Galore Caves revealed that the sequences group within six phyla: AA ct inobacteria, Gammaproteobacteria, AA lphaproteobacteria AA cidobacteria, Firmicutes and Chloroexi. Pahoehoe Cave yellow bacterial mats are more diverse than the Roots Galore mats, which only contacted sequences from the A A ct inobacteria, Gammaproteobacteria, and AA cidobacteria phyla,. Interestingly the Pahoehoe phylogenetic tree (Fig. 1) and the Roots Galore tree (Fig. 2) overlap in the G ammaproteobacteria, AA cidobacteria and AA ctinobacteria. Some of the close relatives that group with our sequences were isolated from other caves around the world, such as Frasassi cave (AA ct inobacteria) and Oregon caves (Gammaproteobacteria). Other closest relatives were environmental isolates from a variety of soils. Figure 2: Parsimony tree of bacterial clone sequences om a yellow microbial mat in Roots Galore Cave in El Malpais National Monument. Numbers on the branches indicate bootstrap values om 1000 re-samplings and indicate the degree of support for this tree topology.

PAGE 107

Lava Caves 712 2009 ICS P roceedings 15th International Congress of Speleology In Figure 2 the AA ctinobacteria clone sequences from our study have mainly cultured closest relatives, something that is rarely encountered in our environmental studies where novel species are common. Also, there are close relatives that came from Oregon caves and Hawaiian lava tubes. In addition, there is a close relative that came from an FeMn nodule, suggesting the presence of possible iron and manganese oxidizing microbes in this lava tube. Because of the iron in the basalt of the El Malpais lava tubes, this is not a surprising nding and reinforces the idea of using various iron media to isolate iron bacteria for further physiological and biochemical study. In comparing results from this study and a parallel study by Garcia et al. (this volume) of microbial mats in lava tubes on Hawai`i, we see a combined diversity that spans 13 phyla, including four of the subdivisions of the Proteobacteria. e greatest overlap amongst New Mexican and Hawaiian lava tubes occurred among the AA ct inobacteria and AA cidobacteria phyla (Table 1), with all but one lava tube with closest relatives found within those phyla. e second most abundant group found was Gammaproteobacteria, which had eight lava tubes with closest relatives among this phylum. ere is a slight trend for yellow microbial mats to be more diverse than white mats, but more sequencing and analyses are needed before a denitive assessment can be made. Scanning electron micrographs revealed the presence of many dierent possible microbial cell shapes, including cells in the shape of rods (e.g. Fig. 3), laments, and spores. e fuzzy rods seen in Figure 3 are one of the more common morphologies seen and may indicate the widespread presence of AA ct inobacteria in these mats.4. ConclusionsWe conclude that the Hawai`i lava tube microbial mats are quite diverse, overall containing 13 phyla of bacterial life as opposed to seven phyla for the New Mexico lava tubes. Hawai`i receives considerable more moisture, especially on the eastern side of the island, than does the region of New Mexico where El Malpais National Monument is located. Additional moisture will inltrate the lava tubes and may bring additional nutrients to fuel some microbial metabolic lifestyles. e Hawai`i lava tubes samples are several thousand years younger than are those of El Malpais, but whether this is a factor in the decreased diversity Phyla ActnPPPPAcdClfxCyanNitVerGemPlancBactDeinOP11Firm Cave 4W Pahoe##### # RG### Bealls#/#/#/#/#/#/#/#/#/#/#/ #/ BP Epper><><>< ><><>< >< Kau/\ /\/\/\ Kula# ## # Maels ***** ** *Table 1: Comparison and contrast of Hawai`i and El Malpais caves by bacterial phyla. ActnActinobacteria, PAlphaproteobacteria, PBetaproteobacteria, PGammaproteobacteria, PDeltaproteobacteria, AcdAcidobacteria, ClfxChloroexi, CyanCyanobacteria, NitNitrospirae, VerVerrucomicrobia, GemGemmatimonadetes, PlancPlanctomycetes, BactBacteroidetes/Chlorobi Group, DeinDeinococcus-ermus, OP11 and FirmFirmicutes. e symbols represent a dierent color of microbial mat as follows: *blue/green ooze, #yellow, white #/yellow and white, >
PAGE 108

15th International Congress of Speleology Lava Caves 713 2009 ICS Proceedings will need to be determined by additional sampling and sequencing of older lava tubes on Hawaii. Many novel species were detected in the clone libraries and many of the sequences from this study are from other volcanic or cave environments. Our knowledge of lava tube microbial mats is increasing as a result of these studies and we know know that the microbial mats can be quite diverse in their phylogenetic makeup.AcknowledgementsWe would like to thank Dr. Penny Boston, Dr. Fred D. Stone, Doug Medville, Hazel Medville, and Don Coons for assistance with eld-work and sampling. For permits and access to the lava tubes we thank the Andy Bundshuh of El Malpais National Monument. John Craig provided lab assistance. We thank Kenneth Ingham for providing amazing photographs of our sample sites. is work would not have been possible without the funding provided by the Cave Conservancy of the Virginias.References Altschul S.F., T.L. Madden A.A. Scher J. Zhang Z. Zhang W. Miller D.J. Lipman (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. NN uc leic AA cids RResearch 25, 3389 3402. Amann, R.I., W. Ludwig, and K. Schleifer. (1995) Phylogenetic Identication and In Situ Detection of Individual Microbial Cells without Cultivation. M icrobiological RReviews 59, 143. Ashmole, N.P., P. Orom M. J. Ashmole, and J.L. Martin. (1992) Primary faunal succession in volcanic terrain: Lava and cave studies on the Canary Islands. Biological Journal of the Linnean Society 46, 207 234. Barton, H.A. and Northup, D.E. (2007) Geomicrobiology in Cave Environments: past, Current and Future Perspectives. Journal of Cave and Karst Studies. 6,:163-178. Ehrlich, H.L. (1999) Microbes as geologic agents: eir role in mineral formation. Geomicrobiology Journal 16(2), 135. Howarth, F.G. (1981) Community structure and niche dierentiation in Hawaiian lava tubes. Island Ec osystems: Biological OOrganization in Selected Hawaiian Communities 15, 318. Hugenholtz, P., B.M. Goebel, and N.R. Pace, 1998. Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity. Journal of Bacteriology 180, 4765. Laughlin, A.W. and G. Woldegabriel (1997) Dating the Zuni-Bandera volcanic eld. Bu lletin: NN ew Mexico Bureau of Geology & Mineral RResources 156, 25. Lazzarini, A., L. Cavaletti, G. Toppo, and F. Marinelli. (2000) Rare genera of actinomycetes as potential producers of new antibiotics. AA n tonie van Leeuwenhoek International Journal of General and Molecular Microbiology 78, 99. Melim, L.A., D.E. Northup, M.N. Spilde, B. Jones, P.J. Boston, and R.J. Bixby. (2008) Reticulated Filaments in Cave Pool Speleothems: Microbe or Mineral? Journal of Cave and Karst Studies 70, 135. Northup, D.E., C.A. Connolly, A. Trent, V.M. Peck, M.N. Spilde, W.C. Welbourn, and D.O. Natvig. (2008) e Nature of bacterial communities in Four Windows Cave, El Malpais National Monument, New Mexico, USA. AA M CS Bulletin 19, 119. Northup, D.E. and K.H. Lavoie, (2001) Geomicrobiology of caves. Geomicrobiology Journal 18(3), 199. Northup, D.E. and W.C. Welbourn. (1997) Life in the twilight zone: Lava tube ecology. NN ew M exico Bureau of Mines & Mineral RResources Bulletin 156, 69. Northup, D.E., D.L. Carr, M.T. Crocker, K.I. Cunningham, L.K. Hawkins, P. Leonard, and W.C. Welbourn, (1994) Biological Investigations in Lechuguilla Cave. NN at. Spel. Soc. Bull. 56, 54. Pace, N.R. (1997) A molecular view of microbial diversity and the biosphere. Science 276 (5313), 734. Staley, J.T., and R. Crawford. (1975) e biologists chamber: lava tube slime. Cascade Caver 14, 20.

PAGE 109

Lava Caves 714 2009 ICS P roceedings 15th International Congress of Speleology UMM JIRSAN: ARABIAS LONGEST LAVATUBE SYSTEMJO O HN N J. PINT NT UIS Commission on Volcanic Caves Ceibas 172, Pinar de la Venta, CP 45221, Zapopan, Jalisco, Mexico, RR an chopint@Yahoo.com is system is located in Harrat Khaybar Lava Field, 130 km north of Medina in the Kingdom of Saudi Arabia. e system consists of three lava-tube passages separated by two collapses and measures 1481.2 m in length with a typical passage height of 8 m and a maximum passage width of 45 m. Sediment covering the cave oor was measured at 1.17 m deep. Wolves, foxes, swis and snakes inhabit or use the cave. Caches of human and animal bones are found in many places, lying on the surface of the oor sediment. Carbon dating revealed that various human skull parts are from 150 to 4,040 years old and the oldest animal bone dates 2,285 years BP. Many basalt fragments of a size and shape useful for gouging or scraping were found inside the longest cave passage, about 180 m from the closest entrance. It is conjectured that older bones and tools might lie beneath the sediment and digging under the guidance of an archaeologist is recommended. Umm Jirsan is one of at least 40 strings of collapses appearing on the most accurate geological map of Harrat Khaybar. Some of these strings are over 15 km long, suggesting that other, much longer lava tubes may be found in this area.1. IntroductionAl-Malabeh et al. (2006) reported that Al-Fahda Cave in Jordan had been surveyed with a total passage length of 923.5 meters. is was, at the time, the longest known surveyed cave on the Arabian Peninsula. In 2007, Umm Jirsan Lava-Tube System in Saudi Arabia was measured to be 1,481.2 meters long with features of possible archaeological signicance. Mapping and limited studies of Umm Jirsan System were carried out by members of the Saudi Geological Survey and the author during ve days.2. General DescriptionUmm Jirsan System is located near the center of Harrat Khaybar Lava Field (Fig. 1), which lies due north of Al Madinah (Medina) in western Saudi Arabia. ese lavas have an area of approximately 12,000 km2 and are mildly alkaline with low Na and K content and include alkali olivine basalt (AOB), hawaiite, mugearite, benmoreite, trachyte, and comendite. e age of Khaybar lavas range from ~5 Ma to historic (Roobol and Camp, 1991). e age of the lava ow in which Umm Jirsan is found has not been determined, but volcanologist M.J. Roobol suggested it may be three million years BP (Roobol, 2007). Figure 1: Location of Harrat Khaybar Lava Field in Saudi Arabia.

PAGE 110

15th International Congress of Speleology Lava Caves 715 2009 ICS Proceedings e main passages of the Umm Jirsan System system extend east and west of Collapse 1, which measures 89 m long by 55 m wide with a depth of 13 m and which is shown in Figure 2. A breakdown slope on the south side of this collapse supports a narrow path from the surface to the oor. In many places along this path, the basalt has been polished, perhaps by the feet of human visitors. e entrance to the east passage (shown in Figure 2) measures 10 m high and 35 m wide. A shallow water channel can be seen along the north wall of the entrance room and mounds of rock-dove guano along the opposite side. Sediment covers the original oor of the cave. is was found to be 1.17 m deep at station 3, with the measurement taken in the center of the passage. North of station 4, much of the surface of the passage oor takes the form of mounds roughly 15 cm in diameter and varying in height. XRD analysis showed one of them to be composed principally of quartz, albite and kaolinite as well as nontronite, biotite, microcline and augite with traces of saponite, montmorillonite and hematite. A number of these mounds are shown in Figure 3. Lava stalactites as well as gypsum and calcite speleothems are found in several parts of the passage. e maximum height of the passage is 12 m and the maximum width 45 m. e length of the passage is 948.6 m. An air temperature of 24 C was recorded at station 5 on May 18, 2007. e west passage has only one entrance (Fig. 4) and measures 341 m in length with a maximum passage width of 45 m and maximum height of 12 m. Lava levees are prominent in this section of the system and the sediment mounds are notably lacking. Moist spots are found in both of the caves long passages, with evidence of water ow particularly noticeable in the east passage. A third passage, 34 m long with a maximum of 20 m wide and 4 m high connects Collapse 2 to Collapse 3.3. Bones and CoprolitesCoprolites and guano indicate that wolves, foxes, hyenas, rock doves, bats, sheep or goats, and swis have inhabited the cave at some point in its history. Swis (possibly Apus pallidus) and bats (not identied) were seen in the west passage of the cave and a swi nest measuring 9 cm in diameter was found on the passage oor. Animal sounds thought to be wolf growls were noted in this same passage. Recently made fox and snake tracks were seen in the east passage. Figure 2: Collapse 1, with a view of the entrance to the west passage. Figure 3: Mounds on the oor of the east passage. Figure 4: A human gure proides scale in the entrance to the west passage. Figure 5: Fragments of two human skull caps found in the Wolf Den in the west passage.

PAGE 111

Lava Caves 716 2009 ICS P roceedings 15th International Congress of Speleology Bones, presumably carried in by predators, were found throughout the cave but were particularly concentrated at the extreme western end of the system in the same place (the Wolf Den) where growls were heard. From among many bones lying on the surface, a human skull, two human skullcap fragments (Fig. 5), one particularly large animal bone and one springy, curved stick were collected, removed from the cave and radio-carbon-dated. e results are shown in Table 1. One of the human skull fragments was dated at 4040 years BP and the still unidentied animal bone (Fig. 6) was found to be 2285 years old. Since these items were found lying on the surface of the sediment layer covering the original oor, it is speculated that still older human and animal remains might be found by excavations carried out by competent investigators.4. Tool-Shaped Basalt FragmentsAt station 5 in the east passage, 180 m east of Entrance Collapse 1, up to 20 fragments of basalt were found lying on the surface within one meter of one another. ese items either had a point at one end or a sharp edge on at least one side. Most of them ranged in length from 7 to 13 cm and were of a shape that ts comfortably in the human hand. So far, no sign of chipping has been detected in these items, but the concentration of so many fragments usable as tools in one small area, raises the question of whether primitive people without tool-chipping skills may have gathered usefully shaped fragments of basalt (the most common rock in the area) for use as simple tools, perhaps for removing Figure 6: is bone, over 2000 years old, is still awaiting identication.Sample record index No.: 2334 Job number: NB-186?RMF-1/2007 No.Sample nameLab. No. Age 14C (BP) Calibrated age range 68.2%Calibrated age range 94.5% 1OO.STK.1GdA-1155300 1522AD (50.0%) 1574AD 1627AD (18.2%) 1646AD 1495AD (69.9%) 1602AD 1616AD (25.5%) 1651AD 2OO.SKU.1GdA-1156150 1670AD (12.8%) 1695AD 1727AD (29.3%) 1779AD 1799AD ( 7.9%) 1813AD 1854AD ( 4.3%) 1867AD 1918AD (13.9%) 1943AD 1667AD (16.3%) 1709AD 1718AD (31.5%) 1784AD 1796AD (30.3%) 1890AD 1910AD (17.3%) 1951AD 3OO.SKU.2GdA-11574040 2618BC ( 5.6%) 2609BC 2598BC ( 1.4%) 2595BC 2583BC (19.3%) 2560BC 2537BC (41.8%) 2491BC 2832BC ( 2.2%) 2821BC 2631BC (93.2%) 2474BC 4OO.SKU.3GdA-115834101749BC (68.2%) 1668BC 1867BC ( 2.8%) 1848BC 1774BC (92.6%) 1624BC 5OO.BON.1GdA-11592285 399BC (56.9%) 360BC 274BC (11.3%) 260BC 404BC (60.7%) 352BC 295BC (33.0%) 228BC 221BC ( 1.7%) 211BCTable 1: Radio-carbon dates for items collected in the Wolf Den, courtesy of Radiocarbon Laboratory, Gliwice, Poland.

PAGE 112

15th International Congress of Speleology Lava Caves 717 2009 ICS Proceedings meat from bones. It should be noted that thin layers of basalt naturally spall o the cave walls in Umm Jirsan. Breaking these plates may have provided cave visitors with a source of fragments from which to select simple tools. No search for similar sites has yet been undertaken in the Umm Jirsan System, nor has digging been undertaken.5. Stone Wallse remains of a stone wall bisect the short passage between Collapses 2 and 3 at its narrowest point. What appear to be the foundations of a rectangular stone building were found at the east end of the east passage. It should also be noted that the breakdown slope on the south side of Collapse 1 may have been reshaped by human hands in order to accommodate the footpath. None of these things have been investigated by experts.6. Conclusions and SuggestionsUmm Jirsan appears to be signicant not only because of its size but also for its archaeological potential. Although the cave was mapped and a few studies were carried out, much remains to be done: the age of the cave has yet to be determined; archaeologists have yet to visit the cave; investigations of what lies beneath the surface of the sediment should be carried out; biologists, micro-biologists and mineralogists ought to do a preliminary study. e few such studies undertaken in other lava caves in Saudi Arabia have resulted in the discovery of artifacts, bones and a human skull (Pint, 2006) and cave minerals rare enough for Saudi Arabias Ghar Al Hibashi to be included among the worlds top ten volcanic caves for hosted minerals. (Forti et al., 2004). In addition, a search for other lava caves in the area should be undertaken. Roobol and Camp (1991) show at least 40 strings of collapses on Harrat Khaybar, some of them up to 15 km long. e existence of intact lava tubes between the Umm Jirsan collapses suggests that other, longer volcanic cave systems may be found in the same lava eld. At present, no Saudi organization has plans to continue vulcanospeleological studies in Harrat Khaybar, but proposals from non-Saudi entities for organizing and nancing such studies might be accepted by Saudi universities or the Saudi Geological Survey.Acknowledgementse author wishes to thank Dr. Zohair Nawab, president of the Saudi Geological Survey for supporting eld trips to Umm Jirsan Cave. anks are also due to geologist Mahmoud Al-Shanti who organized and led these eld trips as well as the survey team. Additional thanks go to Mohammed Al-Moheisen, Saad Aslimi, Hamadi Al-Harbi and Obaidallah Al-Mutairi for putting their lives on the line in the lonely wastes of Harrat Khaybar.ReferencesAL-MALABEH, A., M. FREHAT, H. HENSCHEL, S. KEMPE, (2008): Al-Fahda Cave (Jordan): e Longest Lava Cave Yet Reported from the Arabian Plate. Proceedings of the X, XI, and XII International Symposia on Vulcanospeleology, AMCS, pp 201. FORTI, P., E. GALLI, A. ROSSI, J. PINT, S PINT (2004): Ghar Al Hibashi Lava Tube: e Richest Site in Saudi Arabia for Cave Minerals. AA ct a Carsologica 33, 190. PINT, J. (2006): Vulcanospeleology in Saudi Arabia. A A ct a Carsologica 35, 107 ROOBOL, M.J. (2007): personal communication to author. ROOBOL, M.J. and V.E. CAMP, (1991): Geologic map of the Cenozoic lava eld of Harrats Khaybar, Ithnayn, and K ura, Kingdom of Saudi A A rabia. Saudi Directorate General of Mineral Resources Geoscience Map GM131, with explanatory text, 60 pp. Figure 7: One of many pointed or sharp-edged basalt agments found near station 5.

PAGE 113

Lava Caves 718 2009 ICS P roceedings 15th International Congress of Speleology Identification DENTIFICATION of OF the THE Microbial ICROBIAL Communities OMMUNITIES Associated SSOCIATED with WITH Roots OOTS in IN Lava AVA Tubes UBES in IN new NEW mexico MEXICO and AND hawai HAWAI i IJessic ESSIC A R R SNi I De E R1, MONic IC A MOy Y A1, MATThew HEW G. GARci CI A1, Mich ICH Ael EL N N Spil PIL De E2, Di D I ANA E. N N ORThup HUP1 1DDepartment of Biology, MSC03 2020, University of NN ew Mexico, AA lbuquerque, NN M 87131, USA A 2Institute of Meteoritics, MSC03 2050, University of NN ew Mexico, AA lbuquerque, NN M 87131, USA A Although roots have been found to be an essential energy source in lava tubes around the world, the role of the roots and their microbial communities in the cave environment is largely unknown. We investigated bacterial communities found on roots and walls in two lava tubes in the El Malpais National Monument, New Mexico, USA, and two lava tubes in Hawai`i, USA, using culture-independent methods. Root, wall, soil and water samples were collected to determine carbon levels and for DNA extraction. Root and wall samples were collected for scanning electron microscopy (SEM) to look for presence of microorganisms. All samples were collected with permits or permission of landowners. Samples of these communities were taken aseptically and stored on site in sucrose lysis buer to lyse the cells and for preserve the DNA. DNA was extracted and puried using the MoBio Power Soil DNA Extraction Kit, amplied using polymerase chain reaction (PCR), cloned using Topo TA Cloning and sequenced using Big Dye Terminator v1.1 sequencing. Closest relatives were identied through searches of the NCBI BLAST database. Alignment was done using Greengenes and neighbor joining phylogenetic trees with 100 bootstrap replicates were constructed using Paup version 4.0b10. Bacterial communities of the roots and walls were compared using presence/absence charts. Preliminary results show that the water drips collected from the roots had three times the amount of dissolved organic carbon as drips collected from the walls, suggesting that the roots are an area of increased nutrients in the lava tube. SEM analysis found evidence of bacteria and fungus associated with the roots while only bacteria were noticed on the samples of wall using SEM. Both root and wall samples from the New Mexico lava tubes had closest relatives within the AA cid obacteria, AA lphaproteobacteria and AA ctinobacteria. However, only the wall bacterial communities had closest relatives in the Gammaproteobacteria and Firmicutes, while only the roots had closest relatives in DD eltaproteobacteria, Bacteroidetes and Betaproteobacteria. is study suggests that the roots support a diverse microbial community in the lava tubes and is one of the rst projects to look at root-associated microorganisms in cave environments. 1. IntroductionFew studies have examined the role of roots growing into caves, aside from their use as the sole food source of the troglobitic planthopper O O liarus polyphemus in Hawaiian lava tubes (Howarth, 1972), and no studies have investigated the microbial communities associated with the roots growing into caves. Traditionally cited sources of energy in caves include organic debris entering the cave via sinking streams, gravity or oods (Poulson, 2005), drip water percolating down into the cave, or remains of any type le by trogloxenes visiting the cave (Gillieson, 1996). In addition, chemoautotrophs, use elements from the cave walls or soils as an energy source (Barton and Northup, 2007). However, it is unlikely that roots growing into caves have no eect or bearing on the cave environment and food web. Indeed, we believe roots are another signicant source of food for the cave environment, especially in oligotrophic, or nutrient poor, lava tubes. While plant roots withdraw and store essential nutrients from the surrounding soil matrix, they can also aect the local area around them. Surrounding all plants roots is an area called the rhizosphere, a layer of soil up to 20 mm thick that is aected biologically, chemically and physically by the presence of the roots and is rich in microorganisms directly or indirectly associated with the plant root. Roots excrete numerous exudates and produce dead material in the form of ne root turnover, and this nutrient input results in higher levels of microbial diversity and activity in the rhizosphere compared to that of root/rhizosphere free soils (Madigan et al., 2008). Roots growing into a lava tube cave bring in their at least part of their rhizosphere, including its load of carbon, nutrients and numerous new microbes, into the cave environment. While the presence of roots aects the cave environment, growing into the cave environment also aects the rhizosphere of the roots. As the root grows out of the soil of the epikarst and into open

PAGE 114

15th International Congress of Speleology Lava Caves 719 2009 ICS Proceedings air of the cave atmosphere the rhizosphere loses much of it accompanying soils. It can be assumed that the microbial community associated with the root is altered by this change in the rhizosphere; however, no research has looked at how the rhizosphere changes as it grows into the cave. e roots (and rhizospheres) incursion into the lava tube cave may seed the cave with novel microorganisms. In order to shed more light on the role of roots growing into the lava tube cave environment, we plan to address the following questions: 1. What is the composition of the microbial communities associated with roots growing into lava tubes in New Mexico and Hawaii? 2. Are the roots an area of increased nutrient (e.g. carbon, nitrogen and phosphorus) levels in the lava tube environment? 3. How related are the wall and root microbial communities? Which groups are similar or dierent in the two cave habitats?2. Materials and Methods 2.1 Cave description and sample collectionTwo caves in New Mexico and two caves in Hawai`i, all with active root growth, were selected as collection sites. Roots Galore Cave and Pahoehoe Cave are located in the El Malpais National Monument in northwestern New Mexico. urston Lava Tube and Kula Kai Caverns are located on the Hawaiian Islands. Root samples from each cave were collected aseptically from active root growth near the root apex. Small samples of rocks from the walls and oors at least 2 meters from any noticeable root growth were collected from each cave. All samples were collected aseptically, under an ocial collecting permit or landowner permission and were stored on site in sucrose lysis buer (SLB) to ensure preservation of the DNA. Samples were then transported back to the lab and stored at -80 C until DNA extraction. Additional samples were collected aseptically and stored on site in dry tubes for scanning electron microscopy and nutrient analysis. 2.2 SEM imaging and nutrient analysesSamples for SEM imaging were coated with Au/Pd and viewed using a JEOL 5800 LV SEM at the University of New Mexico. Dry samples collected for nutrient analysis were desiccated, ground, and inorganic carbon was removed by HCl fumigation (Harris et al, 2001). Percent nitrogen and percent carbon was determined by high temperature combustion; the resulting gases were eluted on a gas chromatography column, detected by thermal conductivity and integrated to yield carbon and nitrogen content. Analysis was performed on a ermouest CE Instruments NC2100 Elemental Analyzer (ermouest Italia Sp.A., Rodano, Italy (Pella, 1990)). Soil extractable nitrogen was determined by extraction with 2N KCl followed by analysis for NH4-N using method 98W(1a), 4500-NH3-G(2) and NO3-N using method 100W (1b), 4500-NO3-F (2), on a Techincon AutoAnalyzer II (Mulvanery, 1996). Total phosphorus was determined by combustion at 500 C for one hour, followed by addition of 1N HCl and incubation at 80 C for 30 minutes. Aer dilution the aliquots were analyzed for PO4 using method 94-70W (1c), 4500-P-F (2), on a Technicon AutoAnalyzer II. Total organic carbon (TOC) samples were analyzed using the persulfate digestion method (APHA, 1998) method on a Shimadzu TOC-5050A instrument. All analyses were completed at the UNM Biology Annex Labs.2.3 Phylogenetic studies of microbial communities associated with walls and rootsDNA was extracted and puried using the MoBio Power Soil DNA Extraction Kit (MoBio Laboratories, Inc., California). e 16S rRNA gene was amplied using universal bacterial primers 46 forward (5-GCYTAAYACATGCAAGTCG-3) and 1409 reverse (5-GTGACGGGCRGTGTGTRCAA3) with an amplication reaction mixture that contained 30mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 5 mg bovine serum albumin (BoehringerMannheim), 200 mM (each) deoxynucleoside triphosphates, 100 pmol of each primer and 0.5 U of T T a q polymerase (AmpliTaq LD; Perkin-Elmer) in a nal reaction volume of 25l. Amplicons were cleaned and puried using the Qiagen PCR cleanup kit (Qiagen, Germantown, MD) and were cloned using the TOPO TA Cloning kit (Invitrogen, Carlsbad, Calif). RG71 samples were sequenced using ABI PRISM Big Dye Terminator v1.1 sequencing kit (Perkin-Elmer, Foster City, Calif), while RG88 and PH1 and TH10 samples were sent to Washington University Genome Sequencing Facility for sequencing with primers M13F and M13R. Orientation of all sequences was checked using Orientation Checker (http://www.bioinformatics-toolkit.org) and sequences were screened for possible chimeric artifacts using Mallard (http://www.bioinformatics-toolkit.org). Closest relatives of the genetic sequences were selected using NCBI Blast. Alignment using 650bp was developed using GreenGenes (http://greengenes.lbl.gov/cgi-bin/nph-index.cgi) and manually rened using BioEdit multiple sequence editor (http://www.mbio.ncsu.edu/BioEdit/ BioEdit.html). Neighbor joining and unweighted maximum parsimony phylogenetic analysis was performed using PAUP version 4.0b10. Bootstrap analyses were conducted on 100 resample datasets.

PAGE 115

Lava Caves 720 2009 ICS P roceedings 15th International Congress of Speleology 3. Results and Discussion 3.1 SEM imaging and nutrient analysesere were signicant dierences in root length and lushness of associated fungal and microbial mats between New Mexican and Hawaiian lava tube caves in which we sampled. In Roots Galore Cave, roots grew approximately 20 cm into the cave and were covered with white fungal and bacterial mats (Fig. 1). Roots growing into Pahoehoe Cave were shorter, with only 5 to 8 cm of growth and showed minimal white microbial mats. e roots in the Hawaiian lava tubes typically were between one and four meters, some even growing through the cave and back into the oor of the cave. However, they did not show the thick white fungal mats found on the roots growing into the New Mexico lava tubes, as seen in Figure 2. Samples from Roots Galore Cave viewed by SEM showed fungal growth associated with the root, included one fungal mass appearing to grow into the root (Fig. 3). e root also appeared to have a thick mat of microorganisms growing over most to of the root outer surface (Fig. 4). Figure 1: Macroscopic photographs of the roots in Roots Galore Cave, New Mexico. e roots have about 20 cm of growth and the thick white microbial maps are visible in the photograph. Photos copyright 2007 and 2008, Kenneth Ingham. Figure 2: Macroscopic photographs of the roots in Kula Kai Caverns, Hawai`i. Unlike the roots in Figure 1, the Hawaii roots are 2 meters long and do not appear to have microbial mats associated with them. Photos copyright 2007 and 2008, Kenneth Ingham. Figure 3: SEM images of root samples om Roots Galore. e scale bar in the image is 1mm long. Notice how the fungus appears to be growing into the root at the bottom of the image while other strands anchor the mass to the roots. Figure 4: SEM images of root samples om Roots Galore. e scale bar is 200m long. Microbial mats, including both fungus and bacteria, coer the surface the root and help increase the surface area exposed to soil of the root.

PAGE 116

15th International Congress of Speleology Lava Caves 721 2009 ICS Proceedings Preliminary TOC levels were determined for drip water in Roots Galore Cave. Our results showed that the TOC levels in the water dripping from the roots to be three times greater than that dripping down from the ceiling where no roots were growing. ese preliminary results suggest that the roots are providing additional carbon to the cave.3.2 Phylogenetic studies of microbial communities associated with walls and rootse microbial communities on roots in the New Mexico lava tubes were associated with Pinus ponderosa roots growing into the lava tubes. Closest relatives to the microbial communities associated with the roots grouped within ve phyla: Bacteroidetes, A A ct inobacteria, A A cidobacteria, A A lphaproteobacteria, and Betaproteobacteria. Figure 5 shows a phylogenetic tree of the Roots Galore root microbial mats. 44% of the 18 unique clone sequences grouped within the AA cidobacteria, while 33% grouped within the A A l phaproteobacteria. A number of the clones had closest relatives that were known rhizosphere bacterial species. For example, RG88B09, RG88B02 and RG71A10 all had closest relatives that have been associated with trembling Aspen roots. Other sequences were more closely related to soil microbes, such as RG71B12, which had a closest relative from Holocene lake sediment. Surprisingly, some microbes found on the roots had closest relatives that were common in other caves with no root samples. RG88B11 has a closest relative found in a Hawaiian lava tube bacterial mat and RG71C09 grouped with bacteria found in Roman catacombs. Closest relatives of clones from the Pahoehoe Cave wall sample grouped in six phyla: AA ct inobacteria, AA lphaproteobacteria, Gammaproteobacteria, AA cidobacteria, Firmicutes and Chloroexi (Fig. 6). e closest relatives of the clones were fairly evenly distributed among the six phyla. Most of the clones had closest relatives that resided in other lava tube caves or soils, such as one clone with a closest relative found associated with trembling Aspen. Bacteria found on roots in Hawai`i lava tubes had closest relatives that grouped in AA l phaproteobacteria, Betaproteobacteria and Cyanobacteria. A majority of the clone sequences had closest relatives that live associated Figure 5: Neighbor joining tree of bacteria associated with roots in Roots Galore Cave. Note that while most of the clones have closest relatives found in soils, two of the clones have closest relatives that reside in other cave enironments.

PAGE 117

Lava Caves 722 2009 ICS P roceedings 15th International Congress of Speleology with dolomite rocks, soils and Hawaiian volcanic deposits. Closest relatives of clones from the roots in Kula Kai Caverns in Hawai`i grouped in AA ct inobacteria, A A cidobacteria, Bacteriodetes and Gammaproteobacteria, with 75% of the clones having closest relatives in the A A ct inobacteria. Most of the clones from Kula Kai Caverns have closest relatives that were associated with dolomite deposits in the Alps or with other soils. Figure 6 shows a chart of all phyla found in the roots and walls of New Mexico and Hawai`i lava tubes.4. ConclusionsComparisons of the phylogenetic trees from roots and walls in lava tubes show that while most of the bacterial sequences on the roots are typical root-associated bacteria, some bacteria more commonly found in bacterial mats on the walls of caves were found on the roots. In addition, evidence that some of the soil bacteria more commonly found associated with soil and plant roots have been also been found on the walls of the lava tubes. For example, RG88B11 and RG71C09 clones both group with bacteria mats found in Hawaiian lava tubes and Roman catacombs, respectively. Clones with these similar closest relatives have also been found on the walls of lava tubes in New Mexico. Such results may lend support to the suggestion that there are organisms that are indigenous to subsurface environments, such as caves and catacombs. is suggested that the roots could be picking up the bacteria as they grew through the ceiling of the cave, that the roots are introducing, or seeding, bacteria into the cave environment or that both the roots and the cave walls are acquiring bacteria from the soil overlaying the cave. ese preliminary results suggest that the microbial communities on the roots and the walls are related and may be acquiring microbes from each other or another source. In addition, preliminary results of the nutrient analysis suggest that the roots represent an area of carbon enrichment in the cave and water dripping o of the roots is also carbon enriched. Our analyses suggest a diverse community of microorganisms inhabits both the root masses entering the lava tubes and the walls of the lava tubes.AcknowledgementsWe thank the University of New Mexico SRAC grant, the UNM Biology Department GRAC grant and Springeld Fellowship, the Southwest Region of the NSS Cave Conservation Grant, the Snider family and Kenneth Ingham for funding this project. Herschel Schultz and Andy Bundshuh of the El Malpais National Monument, Kula Kai Caverns sta, the Hawaii Cave Conservancy, and Hawaii Volcanoes National Park provided permits or permission to sample in the caves. Fred Stone provided invaluable help in identifying appropriate lava tubes in Hawai`i, obtaining land owner permissions, and helping with sampling and photography. Finally we would like to thank Don Coons and Penny Boston for sampling help, Kenneth Ingham for his photography, John Craig for all his help in the lab and the nutrient analysis and Armand Dichosa for helping oversee two high school students working on this project. Phylum NM cave rootsNM cave walls Hawai`i cave roots Hawai`i cave wallsActinobacteria X X  X Planctomyocetes X    X Firmicutes X X  X Bacteroidetes X    Alphaproteobacteria X X X X Betaproteobacteria X  X X Deltaproteobacteria X  X Gammaproteobacteria X X  X Acidobacteria X X  X TM7 X      Verrucomicrobia X    X Gemmatimonadetes X      Chloroexi  X    Cyanobacteria    X  Nitrospira      XFigure 6: Microbial populations by phyla om New Mexico and Hawaii cave roots and walls. e New Mexico cave roots and Hawai`i cave walls appear to have the most diverse communities of microorganisms.

PAGE 118

15th International Congress of Speleology Lava Caves 723 2009 ICS Proceedings ReferencesAmerican Public Health Association (APHA). PHA (1998) Standard Methods for the Examination of Water and Wastewater, 20th Edition. Barton H.A. and D.E. Northup (2007) Geomicrobiology in Cave Environments: past, current and future perspective. Journal of Cave and Karst Studies 69(1), 163. Gillieson, D. (1996) C aves: Processes, DDevelopment and Management. Blackwell Publishers, Oxford, UK, 324 pp. Howarth, F.G. (1972) Cavernoiles in lava tubes on the island of Hawaii. Science 175, 325. Harris, D., W.R. Horwath, and C. Van Kessel (2001) Acid Fumigation of Soils to Remove Carbonates Prior to Total Organic Carbon or Carbon-13 Isotopic Analysis. Soil Science Society of America Journal 65,1853. Madigan, M.T., Martinko, J.M., Dunlap, P.V., and D.P. Clark (2008) Brock Biology of Microorganisms 12th edition. Pearson Prentice Hall, Inc, New York, NY, 1168p. Mulvaney, R.L. (1996) Nitrogeninorganic forms. In Methods of soil analysis. Part 3. 3rd ed. SSSA A Book Ser. 5, D.L. Sparks et al. (Eds.) ASA and SSSA Press, Madison, WI, pp 1123. Poulson, T.L. (2005) Food Sources. In Encyclopedia of Caves, D. C. Culver and W.B. White (Eds.), Elsevier Academic Press, New York, pp 255. Pella, E. (1990) Elemental organic analysis: 1. Historical developments. AA m. L ab., 22(3), 116.

PAGE 119

Lava Caves 724 2009 ICS P roceedings 15th International Congress of Speleology ASSUWAYDA LAVA CAVES SOUTHERN SYRIA: SPELEOLOGICAL STUDY COMBINING GEOLOGY AND HISTORYJO O HNN NN Y W. TA TA WK, FAD AD I H. NAD NAD ER R SA A MI KAR AR KA A BI, WA A LEED D JAD AD Splo-Club du Liban, P.O O B ox: 70-923 AA ntelias, Lebanon A very limited number of written documents and maps of Syrian volcanic caves is currently available. In January 2008, an expedition was organized by cavers from the Splo-Club du Liban to As-Suwayda Province in southern Syria in order to explore lava caves. e rst target was set on the known Aariqa cave, which is in the center of the Aariqa town. In December 2008, another expedition took place to nalize the survey of Aariqa cave, which was extended to 562 m. Another cave was also explored and surveyed near Umm ar Rumman town. is cave resulted in 1615 m of underground tunnels, now the longest known development of lava tubes in Middle-East, passing Umm Jirsan in Saudi Arabia (total development: 1481 m). is paper discusses both Umm ar Rumman and Aariqa lava caves and provides speleological documentation combined to geologic and historical investigations. Umm ar Rumman cave is south of As-Suwayda near the border with Jordan and about 20 km southeast of the city of Bosra. is lava tube formed in the earliest uaternary lava sheets (the pahoehoe lavas of Q1). e cave entrance (14 m deep and 20 m wide), which is situated in a at agricultural area, may have been formed by roof-collapse. Almost all features found in volcanic caves are also found in Umm ar Rumman cave; e.g.: levees and gutters, ow ledges, splash stalactites, lava columns and stalagmites, as well as ras. In addition, beautiful calcite speleothems decorate this cave. e Aariqa lava cave was also called Aahir and it is within the Recent lava sheets (Q4; dated to 4000 years BP). e cave was used since the Nabatean period (64 r.f. / .. 106) until the Arab rebellion against the French mandate (1920). e entrance is an impressive, open-collapse with constructed structures that function as facilities for operational personnel and visiting tourists. In addition to the famous basalt door (probably Nabatean in age), remains of stone walls, bones, and pottery were found. e cave, characterized by a at oor, hosts remnants of human construction (housing). Some human activities are suggested by pottery, bones, re places, and housing traces. A small fragment of pottery was dated to the Arab Period in southern Syria (.. 634.. 643). Inscriptions are also seen on the passage walls. Near the end of the cave a narrow passage leads to a second entrance.1. IntroductionAs-Suwayda, Daraa and Hawran (Golan) provinces form the southwestern portion of the Syrian Arab Republic, bordering Jordan to the south and Palestine/Israel to the west (Fig. 1). In the center of this area lies Jabal Ad-Drouz volcanic range, which trends northwestsoutheast and has a maximum elevation of 1785 m. This range has numerous volcanic cones, often organized in ridges whose slopes are covered with many lava flows (DUBERTRET, 1933). This volcanic terrain (called AlHarra) stretches southwardly crossing Jordan and partly northern Saudi Arabia. In Jabal Ad-Drouz, the annual precipitation ranges between 200 and 350mm, while in the nearby Al-Harra plain it does not exceed 100 mm. The average annual temperature is between 15 and 19C (PONIKAROV, 1967). Although the geographic and geologic aspects of this volcanic region were properly studied previously during mapping surveys, no signicant exploration and surveying of the lava caves have been published to date. is contribution discusses two major caves in As-Suwayda province: Umm ar Rumman and Aariqa (Fig. 1). e rst one (Aariqa) has historical signicance as it used to be apparently used for housing in historical and probably pre-historical times. Whereas Umm ar Rumman cave is a fantastic lava cave with beautiful speleothem decorations (volcanic and calcite) and features typical of lava tubes. Here, it must be mentioned that the whole area features historical Nabatean and Byzantine settlements before the Arabian period. A typical example of this rich historical area is the nearby town of Bosra, which hosts a huge amphitheater made-up almost exclusively of carved basalt stones.

PAGE 120

15th International Congress of Speleology Lava Caves 725 2009 ICS Proceedings 2. Geologic SettingsMultiple, overlying, and overlapping volcanic sheets characterize the As-Suwayda region (Fig. 2). Neogene deposits, exposed in As-Suwayda, are mostly eusive basalt rocks of Pliocene or possibly Miocene age. Towards the west (in Hawran), these deposits pass into lacustrine, continental, and marine facies. PONIKAROV (1966) argued that Miocene undivided basalts (N1) must underlie the thick sequence of Pliocene basalts in the central part of the Jabal Ad-Drouz Range (with an estimated thickness of ~750 m). e Miocene basalts do crop out south of Damascus (capital city of Syria) and consists of up to a 500 m thickness of dolerites, ankaramites, and plagio-dolerites. PONIKAROV (1967) subdivided the Pliocene sediments into two parts: a lower part composed of terrigenous rocks and an upper one consisting of basalts. e Pliocene basalts (N2) are widespread in the As-Suwayda region, covering some 4000 km2 and including many chains of volcanic bodies. e greatest thickness of the Pliocene eusives is observed in the massive of Jabal Ad-Drouz where it presumably reaches 700 m. N2 comprises all the desert district of AlHarra (the northern part of which is located in Syria while the southern part mainly in Jordan). Figure 1. Simplied map of the Syrian Arab Republic, showing the proinces and the locations of Umm ar Rumman and Aariqa caves in As-Suwayda Proince.

PAGE 121

Lava Caves 726 2009 ICS P roceedings 15th International Congress of Speleology e uaternary system, whose deposits are dominated by basalts in the study area, has been organized by Lower, Middle, Upper and recent series of deposits by PONIKAROV (1966). e lower series (Q1) series of eusives is again classied into two age groups: 1) the sheet of old pahoehoe lavas (Q1), which occupies the foothills of Jabal Ad-Drouz Range, together with the vast distribution in the Hawran Province, the total area of this sheet reaches 3,600km2; and 2) the overlying sheet of dermolithic lavas (Q1). e outcrops of Q1 are either arched parts of lava swells or blown up lava bubbles with a system of radials patterned ssures. e relative elevation of these positive surface elements does not exceed 8-10 m. e general orientation of lava ows is suggested to be westwards (to the Hawran valley). Probably the maximum thickness of separate lava sheets does not exceed 12-15 m. In these sheets, the Umm ar Rumman cave (about 1600 m) was found (see below). e Middle Series (Q2) of the uaternary basalts is of limited distribution in the study area (total area is around 100 km2). ese consist mainly of a pahoehoe lava ow stretching in a narrow strip with a westward direction. In Jabal Ad-Drouz Range, the Q2 eusives include scoria, scoriaeceous agglomerates, and incoherent pyroclastics (lapilli and volcanic bombs). e Upper uaternary epoch (Q3) appears not to have witnessed volcanic activity. PONIKAROV (1966) recognized seven alternating lava sheets of Recent age in Syria. e rst and oldest basalt sheet (Q4) is limited in surface-extent (70 km2) and located in the northeastern part of the As-Suwayda province. is sheet was formed by 2 or 3 eruptions, the thickness of the each erupted lava sheet reached some 10 m. ese predate ows and sheets of basalts (Q4) of a more extensive nature occupying an area of 170 km2 and underlying a third group of younger lavas (Q4). Another basalt sheet (Q4) consists of pahoehoe lavas and believed to be part of the volcanic plateau of Al-Laja. e sources of eruptions of these lavas are the volcanoes located in the region of the village Majadel (DUBERTRET, 1933). e average thickness of each sheet is considered to be 8 to 10 m. Another distinctive group of basalt ows (Q4) is represented by lava ows from volcanoes of Tell Shihan and Ard El-Kra. Aariqa cave is found within these basalts (see below). e age of bone fragments of mammals, goats, and gazelles, was found to be 4,000 years (by carbon dating method; Ponikarov, 1966). e length of the pahoehoe lava ow from the eruption source (Tell Shihan) to the edge tongues reaches 45 km. Its width in the east, near the volcano, does not exceed 6 to 8 km, gradually increasing to the west reaching 10 km. e ow surface of the pahoehoe lavas features ropy lavas indicating a westward direction of lava ows. ese basalt ows are overlain by the youngest recent eusives (Q4) observed in the study area; they are believed to have occurred in historic time within the last 4000 years. ey include block aa-lavas of the Shahba volcanoes and the southern portion of the Tell AsSafa Massif represented by pahoehoe lavas (PONIKAROV, 1966). 3. Lava CavesOur speleological investigations led us to the discovery and exploration of two lava cave systems in dierent lava sheets (with respect to the age of the lavas). 3.2 Umm ar Rumman Cave (Q1) Umm ar Rumman cave is located south of As-Suwayda near the border with Jordan, and about 20 km southeast of Bosra city. is lava tube is located within the earliest uaternary Figure 2. Schematic block-diagram showing the multiple uaternary and Recent lava sheets in As-Suwayda region and the conceptual positions of Umm ar Rumman and Aariqa lava caves. e block diagram is modied om PONIKAROV (1966).

PAGE 122

15th International Congress of Speleology Lava Caves 727 2009 ICS Proceedings sheets (the pahoehoe lavas of 1Q1, described above) in a at agricultural area. It is characterized by an entrance (14 m deep and 20 m wide) that may have been formed by roof-collapse. e entrance is cluttered with fallen rocks, a big opening leads through an inclined gallery 10 m deep, to reach a linear gallery characterized by a well-traced trail. e total development of Umm ar Rumman cave is 1615 m (Fig. 3), and it contains braided galleries (Fig. 4). As the longest reported lava tube in Arabia was the Umm Jirsan cave in Saudi Arabia with a development reaching 1,481 m (PINT, 2008), Umm ar Rumman becomes now the longest surveyed lava tube development in the Middle East. Umm ar Rumman is a typical lava cave hosting almost all features found in volcanic caves: levees and gutters, ow ledges, splash stalactites, lava columns and stalagmites, as well as ras. In addition, beautiful calcite speleothems decorate this cave. e average diameter of the tube is 7.5 m with a height of 8 m. A huge collapse is 190 m from the entrance. A splash stalagmite is found, about 1 m high, near a molded tree. e collapse ended with a braided maize. Aer a small crawl, a second part continues. Calcite gours cover the oor where we found many fragments of pottery that, aer examination, appeared to belong to the Islamic period (Ayyoubide or Mamlouk, ref. Dr. Leila Badr, conservator of AUB Museum, American University of Beirut). In this part of the cave, many collapses change the homogeneity of the cave prole (Figure 3). At some places, the roof reaches the height of 14 m. At 800 m from the entrance, a second braided maze (Fig. 5) has calcite speleothems (i.e., popcorn, stalagmites, helictites, and others). At the end of the right sided tunnel, the caves oor and walls are reddish with a large amount of fallen rocks. Umm ar Rumman cave ends with a narrow 10 m long tunnel. 3.2 Aariqa Cave (Q4) e Aariqa cave is situated in the center of the Aariqa Figure 3. Map of Umm ar Rumman Cave (surveyed by members of the Splo-Club du Liban and drawn by Johnny Tawk). Figure 4. Photo inside Umm ar Romman Cave, by Johnny Tawk. Figure 5. Photo of a Braided Maze inside Umm ar Romman Cave, by Johnny Tawk.

PAGE 123

Lava Caves 728 2009 ICS P roceedings 15th International Congress of Speleology village. It was also called Ahir cave and was used during dierent historical periods in Syria. is cave is northward of the Umm ar Rumman and within the relatively younger Q4 recent pahoehoe lavas which have been dated at about 4,000 years (see above). At the end of the cave, the transition from pahoehoe basalt of the Q4 and aa of the overlying (younger) Q4 sheet has been observed. e entrance is an impressive open collapse seen from the main road with an average of 14 m wide and 16.2 m depth (Fig. 6). At -14 m from the road, and at the le side, a basaltic stair under two arches goes down 5 m towards Aariqa spring, which is used for domestic purpose in the city. e total development of the cave is 562 m (Fig. 7). e entrance of the cave is protected by a carved monolithic basaltic door from Nabatean or Roman era (64 r.f. to .. 391) about 90 cm wide and 110 cm high, no inscriptions are observed. Aer three steps, you could reach the rst part of the cave which is an east-west 165m long tube (Fig. 8). is part is developed as a show cave, electrical cables and projectors are seen on both sides. It is 16 m large and 9 m high with a at muddy clay oor caused by dripping water from lateral sides. Scarce calcite stalactites are apparent. At the end of the tube, a large chunk of wall is fallen creating lining, the wall is glazed. Beyond this tube, the morphology of the cave takes dierent aspect followed with four smaller tubes linked by tight and low passages. e rst tube is distinguished by an important rock collapse, on which we could map seven enclosures separated by non-carved stone walls not more than 30 cm high. e interiors of these enclosures reveal a replace, animal bones and fragments of pottery (Arab period, aer .. 634, pres. Communication Dr. Leila Badr, conservator of AUB Museum, American University of Beirut), documenting a past human occupation. ough some fragments are recent, a detailed study must be carried out in situ. e second tube is 72 m long, 13 m wide, and 5 m high (Fig. 7). Here also, non-carved stone structures are located on both sides of the oor, animal bones and pottery. is tube is at -12 m from the touristic area. e third tube is 40 m long, 10 m wide, and 5 m high. is part represents the continuation of the second tube, which is separated by a ceiling collapse. At the end of the third tube, a side passage ascends 20 m high and reaches a second entrance to the cave in the garden of a private house. At the end of the cave, a rounded construction, bones, and pottery were observed. Also, dripping water is noted. Also, two cupolas are seen in the ceiling (Fig. 7). In general, the temperature of the cave is 18 C (December 2008) and some volcanic formations are spotted in this cave, including linings, splash stalactites, breakdown areas, and contraction cracks. Figure 6. Entrance of Aariqa Cave, by Johnny Tawk. Figure 7. Map of Aariqa Cave (surveyed by members of the Splo-Club du Liban and drawn by Johnny Tawk). Figure 8. Photo of the touristic part inside Aariqa Cave.

PAGE 124

15th International Congress of Speleology Lava Caves 729 2009 ICS Proceedings 4. ConclusionsTwo expeditions led to the exploration and surveying of two signicant lava caves in Syria and the Middle East. e rst, Umm ar Rumman cave with 1,615 m of underground development can be considered the longest of its kind in Arabia and the Middle East. is cave, probably older than a million year (in the uaternary lava sheet 1Q1) is also decorated with the full range of volcanic and breathtaking calcite speleothems. e second, Aariqa Cave, has a development reaching 562 m, still it holds an important historical aspect as housing remains were found in its galleries. Aariqa Cave must be younger than 4,000 years old as it is formed in the recent lava sheet 5Q4.ReferencesDUBERTRET, L. (1933) La Carte Gologique au millionime de la Syrie et du Liban. RR evue de Gographie Physique et Golologie DD ynamique, 6 (4), 269. PINT, J. and S. PINT (2008) Umm Jirsan: Arabias longest lava cave. AA rticle on a Web site: http://www. saudicaves.com/jirsan/index.htm. PONIKAROV, V.P. (1966) e geology of Syria: E xplanatory NN otes on the Geological Map of Syria, scale 1:200 000. Ministry of Industry, Syrian Arab Republic. PONIKAROV, V.P. (1967) e geology of Syria: E xplanatory NN otes on the Geological Map of Syria, scale 1:500 000. Part I: Stratigraphy, Igneous RRocks and TT ectonics. Ministry of Industry, Syrian Arab Republic.

PAGE 125

Lava Caves 730 2009 ICS P roceedings 15th International Congress of Speleology

PAGE 126


PAGE 128

15th International Congress of Speleology Rare Species 733 2009 ICS Proceedings 1. IntroductionCarlsbad Caverns, Guadalupe Mountains, New Mexico (USA) is famous for its large, highly decorated chambers but also popular with visitors for the summer evening exodus of almost 500,000 Mexican free-tailed bats (Betke E TKE et al., 2008). However, this region has numerous smaller, less spectacular caves that provide important resources for other cavernicolous bat species. ese caves oer appropriate microclimate for roosting bats species-specic temperatures and humidity, which insure reproductive success and survival (Tuttle U TTLE 1979; Gore ORE and AND Hovis OVIS 1998). Because caves provide critical habitat for bats, exclusion from these sites can negatively impact populations. Unfortunately, bats are highly intolerant of human intrusion into their roosts (Mann A NN et al., 2002; Buecher U ECHER 2006; Elliot LLIOT T, 2006) and will abandon important sites if disturbed too frequently (Mohr O HR 1972; Tuttle UTTLE 1977; Tuttle UTTLE 1979). Given that cave resources are fragile and bats need protection from human disturbance, land managers oen install gates at popular caves to protect both biotic and abiotic cave resources. Unfortunately, improperly designed cave gates intended to protect bats can oen do more harm than good. Tuttle U TTLE (1977) discovered that at a number of caves, specically gated to protect bats, the animals abandoned the sites within two years. is is due, in part, to alteration and reduction of the bats yway but can also result from changes in cave microclimate by modifying air exchange. Additionally, ill-conceived gates can provide a location for predators to grab bats when they are forced to y slower to negotiate a gate (Tuttle U TTLE 1977). Cave gates can also increase circling behavior by bats and cause bats to retreat more oen into the cave prior to emerging from a site. is behavior is more pronounced in larger colonies whose gates are constructed in small passages (i.e. less available ight space per bat), impeding smooth emergence by the animals (Spanjer PANJER and AND Fenton ENTON 2005). Any bottleneck that gates create for bats emerging daily from a cave increases ight time and energy costs, which can reduce reproductive success and survival.2. History of a Cave GateYellow Jacket Cave overlooks Dark Canyon in the Guadalupe Mountains and is managed by Bureau of Land Management (BLM). e cave is approximately 10 miles from Carlsbad New Mexico and has over 1.5 miles of passage. In the spring of 1979, cavers from the Southwest Region (SWR) of the National Speleological Society (NSS) began mapping Yellow Jacket and discovered evidence of a major bat roost. In the spring of 1980 they reported to BLM ... active bat roosts everywhere. It was suspected GATING A CAVE PROTECTS A BAT COLONY ... EVENTUALLYDD EBBIE C. BUECHER R1, ANDR ANDR EA A K. GOOD OOD BAR AR2 1Buecher Biological Consulting, 7050 E. Katchina Court, TT ucson, AA Z 85715 USA A2Volunteer for Bureau of Land Management, 3027 E, DDerrick, Carlsbad, NN M 88220 USA A e Guadalupe Mountains of New Mexico (USA) are best known for the highly decorated Carlsbad Caverns, which is home to a large maternity colony of Mexican free-tailed bats (T T ad arida brasiliensis). e Guads also have numerous smaller, less spectacular caves that are critical habitat for other bat species. Because cave resources can be damaged with unlimited human visitation, land managers oen protect caves by installing cave gates. However, not all bat species readily accept gates and some bats have abandoned historic roosts once the sites are gated. Unfortunately these animals originally chose the caves because they had the appropriate species-specic temperature and humidity for their survival. We describe eorts to protect a maternity colony of cave myotis (Myotis velifer) in a Bureau of Land Management (BLM) cave near Carlsbad, New Mexico. A long-term (>15 yr.) monitoring program documents the near-demise of a signicant bat colony, resulting both from human vandalism and the consequences of a cave gate. We describe how the size and placement of a gate, designed to protect the bats, resulted in cave myotis abandoning the cave. Fortunately BLM management appropriately addressed the problem and altered the gate design and placement so that bats have reoccupied the cave in historic numbers. We discuss possible consequences of gating caves without fully understanding how the bat species using the site might accept a gate. Because of before and aer documentation of how this particular bat species responded to a gate, we strongly recommend long-term monitoring aer installation of any cave gate designed to protect a bat colony.

PAGE 129

Rare Species 734 2009 ICS P roceedings 15th International Congress of Speleology that the cave was used as a nursery. is was conrmed. Because of the nursery, we delayed further work. When SWR completed their map (Belski E LSKI et al., 1986), they had documented 10 distinct areas in the cave where bats day-roost. Unfortunately, the proximity of Yellow Jacket to Carlsbad New Mexico resulted in its use as a party cave by locals and there was oen evidence of vandalism, including re rings, spent recrackers, and grati. Because of these threats to the resource, BLM developed a cave management action plan in 1986 for the protection of Yellow Jacket Cave. Although the number of bats using the cave and the specic species was unknown, it was recommended that a cave gate be installed to protect the colony. Prior to the installation of this gate, BLM conducted 5 bat counts between June and August. ese counts were conducted at the entrance of the cave as bats emerged to forage. During summer 1986 the average number of bats was 2,158 with the highest bat count at 2,742 in early July. BLM and SWR installed a steel gate in December 1986 in a small passage (1 meter wide x 0.5 m high) approximately 25 m from the entrance. During summer 1987, BLM volunteers conducted 12 post-gate bat counts to monitor gate acceptance by the bats. e average number of bats emerging from the cave that summer was 1,145, with the largest number of bats at 3,246 in early August. Although the numbers looked reasonable, it was discovered that bats were oen roosting in passages on the outside of the gate, avoiding the gate entirely and exposing the animals to continued harassment. In 1988 bats still avoided the gate and were still roosting in a side passage between the gate and the cave entrance. In an attempt to mitigate the situation, BLM le the gate open throughout the summer and continued to conduct evening emergence counts. Despite the open gate, bats still roosted outside the gate. Also during this year, garbage was repeatedly removed from the entrance area and BLM documented new re rings and fresh grati. Due to budget cuts, no bat counts were conducted between 1988 and 1997, however cavers reported fewer bats using Yellow Jacket. Although the reduced yway, created by the gate, no doubt impacted the bats, the low numbers were most likely due to harassment by humans. Evidence indicated that the cave continued to be used by local youths for parties, with beer bottles and re rings oen found in the entrance rooms. It was also reported that people harassed bats using tennis rackets and shotguns. is information conrmed the need to re-evaluate the situation and determine if a more bat-friendly cave gate might better protect the bats. In 1997 BLM contracted a bat biologist to evaluate Yellow Jacket Cave, determine which bat species used the site and propose possible ways to protect the colony and restore the historic population. is study included seven emergence counts at the entrance and installation of a passive infrared bat counter to monitor bat activity throughout the night. It was determined that Yellow Jacket had a maternity colony of cave myotis (Myotis velifer). Cave myotis are the largest myotis species in North America with a wingspan of 28 cm and weighing 12 g. Compared to other myotis, they are recognized by a longer forearm (40 mm), lack of a keeled calcar and a sparsely furred area between the scapula (Fitch ITCH et al., 1981; Adams D AMS 2003). During summer 1997 the cave myotis used the cave in low numbers (average = 290) but the population increased in the fall (799 in early September). is increase in bat numbers was observed each year and suggested that Yellow Jacket was important as a swarming site, where solitary bachelor males rejoin females to mate in the fall prior to entering hibernation (Fenton E NTON 1969; Spanjer PANJER and AND Fenton ENTON 2005). At the end of the study, Buecher U ECHER et al. (1997) recommended that the existing gate be removed and a more bat-friendly gate be installed closer to the entrance in passage with a larger crosssection. is would oer an area for bats to circle on either side of the gate, provide more choices for ying through the bars and would reduce possible predation at the gate. In March 2000, the original gate was removed by SWR and a new steel bat-friendly gate (Powers O WERS 2004) was built in the large entrance passage. is location provided protection of all historic bat roost sites documented by SWR (Belski E LSKI et al., 1986). Post-construction emergence counts at Yellow Jacket by BLM volunteers over foue years (2001) documented dramatic increases in bat-use (Table 1) by cave myotis. e average number of bats now using Yellow Jacket Cave increased each year and the highest number of bats using the cave in 2004 was 6,155 bats. By all appearances, the bats denitely approve of the new gate design and its location in the larger cave passage. YearAverage No.Max. Number 2001 1663 4000 2002 2111 3845 2003 2175 4205 2004 3191 6155Table 1: Average number and maximum number of bats documented in Yellow Jacket Cave each summer between 2001-2004.3. DiscussionPre-gate emergence counts indicated that the numbers of bats using Yellow Jacket were much higher before the cave was gated. In 1988 a gate was constructed in a small

PAGE 130

15th International Congress of Speleology Rare Species 735 2009 ICS Proceedings cave passage because it was cheaper to build and easier to secure against vandals. Unfortunately, the placement of horizontal and vertical steel bars decreased the bats ight window by almost half. e gate may also have provided a spot for predators to take bats from the air as they slowed to negotiate the gate. Aer installing the rst gate (Fig. 1a), the fact that bats avoided it suggested that the gate was detrimental to the bat colony. Fortunately BLM management discovered this through post-gate emergence counts. ey removed the original gate and restored the passage to historic dimensions. ey then constructed a new gate in a larger passage using steel bars with wider horizontal spacing and eliminated many of the vertical bars (Fig. 1b). is bat-friendly design resulted in bats reoccupying the cave in historic numbers. is emphasizes the importance of conducting both preand post-gate emergence counts to fully understand how bats behave prior to a gate and, once a gate is installed, how well they accept the new obstacle in their yway. In 1986 BLM attempted to protect the resources of Yellow Jacket Cave using the only method at their disposal, a cave gate. Although their eorts almost caused the demise of a maternity colony of cave myotis, it is important to understand that at the time, little was known regarding how bats responded to cave gates and why some bat species accepted cave gates, while others did not. It was, unfortunately, as a result of abandonment of known bat roosts across the U.S., due to a variety of ill-conceived gate designs, that we have learned which gates are best for bats (Powers OWERS 2004), which bat species may or may not accept gates (Sherwin H ERWIN et al., 2004) and what gates do not negatively impact cave microclimate (Richter I CHTER et al., 1993; Kennedy ENNEDY 2004). Gating caves has become a common management practice to protect fragile cave resources, both biotic and abiotic (Tuttle U TTLE and AND Taylor AYLOR 1998; KERBO, 2004). Because of our before and aer documentation of how this particular bat population responded to a gate, we strongly recommend long-term monitoring aer installation of any cave gate designed to protect a bat colony. Evidence of additional circling at bat gates (Spanjer P ANJER and AND Fenton ENTON 2005) requires more energy that can negatively impact bats. is may be more problematic at maternity roosts where pregnant females are less maneuverable and where newly volant young are less experienced yers. Gating policies should consider all cavernicolous bat species in an area because caves are oen used by multiple species (Tuttle U TTLE 1977; Ludlow UDLOW and AND Gore ORE 2000; Spanjer PANJER and AND Fenton ENTON 2005). Without fully understanding which bat species are using a site and whether they accept cave gates may result in negative consequences for populations. Given the sensitivity of bats to human disturbance but also the resistance of some species to cave gates, we must approach this issue with care in order that we do not exclude the very species we desire to protect.Acknowledgmentsis study was conducted with nancial support from the BLM Carlsbad Region oce. Bat counts were conducted through the years by many BLM volunteers including, but not limited to: Andrea Kurman, Sandy Major, Bob Brink, Darcy Clements, Bob Buecher, Debbie Buecher, Harry Burgess, Jackie Burgess, and Jim Goodbar. Dave and Carol Belski conducted all bat counts in 2001, making this comparative analysis possible. We especially thank Jim Goodbar, BLM Cave Specialist for his diligence in seeking a solution to restore this cave myotis maternity colony. We thank the editors (VORIES et al., 2004) of the Proceedings on Bat Gate Design: A Technical Interactive Forum for a concise review of various bat-friendly cave gates, including supply lists and detailed plans. is resource is invaluable for land managers. Figure 1: Top photo is of the 1986 cave gate. Note how much the bars of the gate reduce the ight window for bats. Bottom photo is the 2000 bat-iendly gate installed by BLM and SWR. Note the larger ight window and number of choices for bats exiting or entering the cave.

PAGE 131

Rare Species 736 2009 ICS P roceedings 15th International Congress of Speleology ReferencesAdams DAMS R.A. (2003) B ats of the RRocky Mountain West: N N atural history, ecology and conservation. University Press of Colorado, Boulder, CO, 289 p. Belski E LSKI D., C. Belski ELSKI and F. Hardy ARDY (1986) NSS Convention Guidebook No. 25. Southwestern Region of the National Speleological Society, 25, 66 p. BETKE, M., D.E. HIRSH, N.C. MAKRIS, G.F. MCCRACKEN, M. PROCOPIO, N.I. HRISTOV, S. TANG, A. BAGCHI, J.D. REICHARD, J.W. HORN, S. CRAMPTON, C.J. CLEVELAND and T.H. KUNZ. (2008) ermal imaging reveals signicantly smaller Brazilian freetailed bat colonies than previously estimated. Journal of Mammalogy 89, 18. Buecher U ECHER D.C., R.H. Buecher U ECHER H. Burgess URGESS and J. Burgess URGESS (1997) Results of 1997 summer bat monitoring at McKittrick Cave, Endless Cave, Yellow Jacket Cave and Lair Cave. Report for Bureau of Land Management, Carlsbad New Mexico, 155 p. Buecher U ECHER D.C. (2006) Do not disturb hibernating bats or nursery colonies. Cave Conservation and R R estoration. National Speleological Society, AL, p. 43. Elliott L LIOTT W.R. (2006) Biological dos and donts for cave conservation and restoration. Cave Conservation and R R estoration. National Speleological Society, AL, p. 33. FENTON, M.B. (1969) Summer activity of Myotis lucifugus (Chiroptera: Vespertilionidae) at hibernacula in Ontario and uebec. Canadian Journal of Zoology 47, 597. FITCH, J.H., K.A. SHUMP, Jr., and A.U. SHUMP (1981) Myotis velifer. Mammalian Species, AA m erican Society of Mammalogists 149, 1. GORE, J.A. and J.A. HOVIS (1998) Status and conservation of southeastern myotis maternity colonies in Florida caves. Florida Scientist 61,160 170. KENNEDY, J. (2004) Preand post-gate microclimate monitoring. P roceedings of Bat Gate DDesign: AA T T echnical Interactive Forum, Austin, TX, p. 353 357. KERBO, R. (2004) Cave and karst resources. Proceedings of B at Gate DDesign: AA TT echnical Interactive Forum, Austin, TX, p. 3. LUDLOW, M.E. and J.A. GORE (2000) Eects of a cave gate on emergence patterns of colonial bats. Wildlife Society Bulletin 28, 191. MANN, S.L., R.J. STEIDL and V.M. DALTON (2002) Eects of cave tours on breeding Myotis velifer. Journal of Wildlife Management 66, 618. MOHR, C.E. (1972) e status of threatened species of cave-dwelling bats. NN a tional Speleological Society Bulletin 34, 33. POWERS, R.D. (2004) e angle iron bat gate. Proceedings of B at Gate DDesign: AA TT echnical Interactive Forum, Austin, TX, p. 159. RICHTER, A.R., S.R. HUMPHREY, J.B. COPE and V. BRACK, JR. (1993) Modied cave entrances: thermal eect on body mass and resulting decline of endangered Indiana bats (Myotis sodalis). Conservation Biology 7, 407. SPANJER, G.R. and M.B. FENTON (2005) Behavioral responses of bats to gates at caves and mines. Wildlife Society Bulletin 33, 1101. SHERWIN, R.E., J.S. ALTENBACH and S. HAYMOND (2004) e responses of bats to gates. Proceedings of B at Gate DDesign: AA TT echnical Interactive Forum, Austin, TX, p. 333. TUTTLE, M.D. (1977) Gating as a means of protecting cave dwelling bats. P roceedings of the 1976 NN ational Cave Management Symposium, Aley and Rhodes, p. 77. TUTTLE, M.D. (1979) Status, causes of decline and management of endangered gray bats. Journal of Wildlife Management 43, 1.

PAGE 132

15th International Congress of Speleology Rare Species 737 2009 ICS Proceedings Tuttle, M.D. and D.A.R. Taylor (1998) Bats and mines. Bat Conservation International, Resource Publication No. 3, Austin, Texas, 50 p. VORIES, K.C., D. THROGMORTON and A. HARRINGTON. (2004) Proceedings of Bat Gate D D esign: AA TT echnical Interactive Forum held in AA ustin, T T exas March 4-6, 2002. Carbondale (IL): USDOI Oce of Surface Mining Alton, Illinois and the Coal Research Center, Southern Illinois University at Carbondale. 434 p. Proceedings available on the Internet
PAGE 133

Rare Species 738 2009 ICS P roceedings 15th International Congress of Speleology REGIONAL HABITAT CONSERVATION PLANNING IN KARST TERRAIN, WILLIAMSON COUNTY, TEXAS, USAS.W. CAROThe HE Rs S1, K. Whi HI Te E1, G.L. GAlb LB RAi I Th H1 and G.D D BOy Y D2 1SWCA A Enironmental Consultants, 4407 Monterey OOaks Bld., Bldg. 1, Suite 110, AA ustin, TT X 787492Williamson County Conservation Foundation, 400 West Main St., Suite 216, RRound RRock, TT exas 78664 Abstract Regional Habitat Conservation Plans (RHCPs) are an increasingly popular option for local governments to balance economic growth with conservation of habitat for rare and endangered species. For most of this decade, Williamson County, Texas has remained one of the fastest growing areas of the United States with much of that growth occurring on the Edwards aquifer karst. Nearly ten percent of the known caves in Texas lie within the path of growth west of the cities of Round Rock and Georgetown. Karst endemic species include three endangered karst invertebrates, more than a dozen other rare karst invertebrates, and several salamanders. Concerned community leaders, including several members of the Williamson County Commissioners Court formed the Williamson County Conservation Foundation in 2002 to develop the Countys regional planning approach. Approved by the U.S. Fish and Wildlife Service in October of 2008, the Williamson County RHCP provides expedited compliance with the Endangered Species Act while recovering two endangered species and precluding the need to list others. Mitigating species impacts on a regional scale allows high quality conservation areas to be selected on the basis of biological diversity and preserve integrity. A minimum of nine high quality karst preserves have been or will be established within the 100,000 acres of undeveloped karst by 2038.

PAGE 134

15th International Congress of Speleology Rare Species 739 2009 ICS Proceedings Inciting NCITING public PUBLIC interest INTEREST and AND p P ROFESSIONAL Partnerships ARTNERSHIPS in IN the THE inventorying INVENTORYING and AND monitoring MONITORING of OF cave CAVE invertebrates INVERTEBRATES Kartchner ARTCHNER caverns CAVERNS State TATE park PARKRORO BERT RT CA A SA AVANT ANT 1, 2, ST T EVE WILLSEY 1, ST T EVE CA A SPER R 3, SUPA A PAN AN SERA RA PHIN N 3 1 AA rizona State Parks, 1300 W. Washington St., Phoenix, AA Z 850072 DDept. Mining & Geological Engineering, University of AA rizona, TT ucson, AA Z 857212 DDept. Material Science& Engineering, University of AA rizona, TT ucson, AA Z 85721 Abstract Kartchner Caverns State Park (KCSP) is a commercial cave operated by Arizona State Parks. Opened in 1999 it was one of the most environmentally sensitive developments of a show cave anywhere in the world. When our visitors explore the large chambers and varied formations at Kartchner Caverns, they cannot fully realize or appreciate the myriad of tiny eyes, antennae and other sensory organs that our oligotrophic invertebrates are employing to detect their passage. Most of our invertebrates are associated with the seasonal inux of fresh bat guano. Some are parasites on the Myotis velifer bats, while others nd a home in the cracks of the breakdown rock throughout the cave. Like the microorganisms that ourish in our cavern complex, cave invertebrates ourish in environments that are too small to be exploited by macroorganisms, and observed during cave tours. Before construction was begun in the cave a 4-year study of the undisturbed cave environment was undertaken. One study included an inventory of the invertebrates found in the cave. irty-eight species were documented, several of which are thought to be new species. One goal of the rst study was to establish a baseline for future comparative studies. e Cave Science Unit at KCSP invites qualied invertebrate researchers to assist in ongoing habitat studies and inventory of known and dierent species. In 2007 a graduate student from the Materials Science and Engineering Department at the University of Arizona performed high-resolution imaging and measurement of various invertebrate specimens that had been previously collected in the caves. Inspiring images of various specimens were produced by a scanning electron microscope (SEM) for the study. ese were shared in presentations by sta and the researcher in a number of public forums. Preliminary ndings from two species of beetles found aer the predevelopment studies indicate they could also be new species. Both of these beetles were imaged by the SEM, which provided high-resolution details of their body morphologies for future study. SEM images of cave spiders showed no bacterial growth on their bodies, whereas other invertebrates where the spiders were found, exhibited bacterial components in abundance. Images like these pose intriguing questions about ecosystem dynamics. High-resolution imaging of our invertebrate world continues to inspire and awe park visitors. rough technology, shared goals, and partnerships, we hope to entice and develop a new generation of cave resource stewards. Future studies hope to address questions like, Why is the surface of the spiders carapace so crenulated? Does it reect or refract light, provide greater absorption of air born molecules, or help the spider sense subtle changes in air pressure that signal prey is near? How do our tiny springtails sense food sources like bacteria and fungi from great distances across the cave? As a small part of the interpretive program at Kartchner, we found a very interested public fascinated by the SEM images and interested in protecting its invertebrate world. Cave invertebrates serve as important indices for measuring the balance of Kartchners cave ecosystem. us, we look forward to partnering with others on a cost-eective and sustainable monitoring program that accurately assesses and conserve the invertebrates of Kartchner Caverns and other caves in the area.

PAGE 135

Rare Species 740 2009 ICS P roceedings 15th International Congress of Speleology The HE Role OLE of OF Karst ARST Fauna AUNA Areas REAS in IN the THE Advancement DVANCEMENT of OF Karst ARST Conservation ONSERVATION in IN Central ENTRAL Texas EXASC. Cl L Ove VE R, P.G. Cl L Am M ONs S1, Pie IE RRe E PAQui UI N2 1Karst and Endangered Inertebrate R Research Laboratory, SWCA A Enironmental Consultants, 4407 Monterey O Oaks Boulevard, Building 1, Suite 110, AA ustin, TT exas, 78149, U.S.A A.1cclamons@swca.com2ppaquin@swca.com Abstract Urban development in central Texas is a threat to the integrity of natural habitats, especially caves and karst. Sixteen cave-restricted invertebrate species are protected by the U.S. Endangered Species Act (ESA) in Central Texas, while many others are considered species of concern. In order to prevent extinction of listed cave invertebrate species, to favor conservation, and ultimately to achieve species recovery, the U.S. Fish and Wildlife Service (USFWS) proposed guidelines to dene areas that could be designated as units for conservation. e USFWS commissioned a study that dened and delineated Karst Fauna Regions (KFR) based on potential geologic and geographic barriers to karst invertebrate dispersal, and limits of species distribution. Since 1992, studies have shown that the boundaries proposed for the KFRs do not always correspond to the known species boundaries. In addition, some species previously thought to be restricted to one KFR have been collected over several contiguous KFRs. e concept of a Karst Faunal Area (KFA) was rst proposed in the RR ecoery Plan for Endangered Karst Inertebrates in TT ravis and Williamson Counties, TT exas. Although titled as a recovery plan, the document objective is limited to downlisting the species, and recommends criteria and specic actions required to achieve that objective. ese criteria include the following phrase: three karst fauna areas within each karst fauna region in each species range should be protected in perpetuity. In the Recovery Plan, a KFA was envisioned as a separate, protected unit of occupied habitat that would be ocially designated and preserved for the benet of listed troglobitic species recovery. Since the KFA concept was introduced, additional scientic knowledge of karst invertebrates has been gained, and many karst preserve areas have been established. However, only one of these preserves is considered a KFA by USFWS. is unit was established in November 2008, more than 14 years aer the publication of the Recovery Plan. Some of the criteria proposed to dene KFAs are rather vague and subjective, particularly the requirement that caves within a given KFA must be independent from other caves. In the only currently designated KFA, the location on an isolated hilltop clearly suggests an isolated cluster of caves, but independence will be dicult (or impossible) to determine where the surface topography diers. In order to circumvent this problem and use a quantiable scientic baseline to establish the independence among clusters of caves, we propose to use genetic data to assess the degree of connections between caves. We provide an example using mtDNA data of a troglobitic Cicurina to determine the structure of the genetic diversity of a small karst area and identify Signicant Evolutionary Units (ESUs). ESUs are more appropriate units to guide conservation strategies because they are measurable, internationally recognized and scientically sound.

PAGE 136

15th International Congress of Speleology Rare Species 741 2009 ICS Proceedings SPELEOTHEMS AND LIVING ORGANISMS: WHAT KIND OF RELATIONSHIP?PAO AO LO O FORTi I Italian Institute of Speleology, University of Bologna, Italy, paolo.forti@unibo.it e idea that speleothems may be somehow inuenced by living organisms is rather old, but specic studies have only started in the last few decades and presently there are only few systematic papers on this topic. e role of micro-organisms is perhaps the best investigated though not fully understood, while studies on upper organisms and speleothems are scarce and details on the involved genetic mechanisms are not always given. e aim of the present paper is to enhance the interest of the scientic community on this peculiar topic. In fact the complex biochemical reactions involved in the development of the dierent cave deposits clearly have an importance far exceeding the simple speleogenetic interest.1Introductione idea that the development of the secondary chemical deposits in caves may be somehow inuenced by living organisms is rather old (Shaw H AW 1997): the shape and the internal structure of some speleothems (stalactites, stalagmites, coralloids) suggested to the early visitors of caves (ALDROVANDI, 1648) the possibility that they grow as plants even though the current idea was that minerals (and therefore also the speleothems) were living organisms but at a lower level with respect to plants or animals until the second half of XVIIth century. In the second half of the XVIIth century a scientist put forth the idea that speleothems are true rock plants (Beaumont EAUMONT 1676), this theory was later perfected by J. P. Tourneford OURNEFORD who wrote: at certain rocks nourish themseles in the same way as plants. Perhaps they reproduce also in the same way.. that there are seeds which gradually swell up and develop the regular structure which is perhaps hidden beneath their surface us the congelations grow up om seeds. us, in the XVIIIth century, some of the most common types of speleothems were oen represented just as part of a tree: with stalactites as roots, stalagmites and columns as trunks, helictites as leaves or owers. Since the second part of XVIIIth century the progress in chemical studies allowed the detection of the main mechanisms by which calcite and other minerals deposit in caves and consequently any possible biogenic interaction in speleothem evolution was rejected for over one century. But the increase of scientic observation inside caves which characterised the XXth century allowed the opportunity to reconsider the whole matter. Today the fact that living organisms may inuence the external shape and/or the chemical composition of speleothems is generally accepted. Nevertheless systematic studies on this topic have never been done, except for a few dealing with microbiology (Sasowsky ASOWSKY AND Palmer ALMER 1994; Northup ORTHUP et al., 1997; FORTI, 2002). Research in this specic eld started only y years ago, but their development progressively highlighted the role played by micro-organisms in the genesis and the evolution of secondary cave minerals. erefore at present, some even doubt that caves may host speleothems developed without the active and/or passive control of living organisms. e role of micro-organisms is perhaps the best investigated, even if it is far from being fully understood. Studies on upper organisms and speleothems in caves are scarce and normally refer only to the occurrence of biologically controlled chemical deposits saying nothing on the involved genetic mechanisms. 2The Role of Microorganisms Presently, it is well established that microorganisms can directly cause biomineralization through enzymes, or can produce substances that lead to the precipitation of minerals (e.g. by changing the pH in their surrounding) or they may become the privileged support for nucleation. e microbial processes in caves oen involve redox reactions. e microbial players are varied: aerobic (chemiolithotroph) microorganisms, which obtain energy directly from the oxidation of inorganic compounds, but also anaerobic (heterotroph) organisms which obtain energy from the oxidation of organic matter and reduce inorganic compounds. 2.1 e sulfur cyclee microbial reactions of the sulfur cycle are perhaps the best studied and have been proved to cause the development of a lot of cave minerals: native sulfur, gypsum and iron oxides-hydroxides are the most common speleothems developed by them, but plenty of others have been reported in literature (Hill I LL and AND Forti ORTI 1997; SHOPOV,

PAGE 137

Rare Species 742 2009 ICS P roceedings 15th International Congress of Speleology 2004; Forti ORTI et al., 2009). Sometimes the large amount of organic matter produced in the sulfur cycle allow for the evolution of speleothems (pseudo-stalactites) consisting of single organic mat (mucus), which are normally called mucolites. Apart from those related to the sulfur cycle, many other kinds of biomineralization can occur in the cave environment: the most important of which are: 1the salpeter evolution; 2the phosphate deposition; and 3-the guano digestion.2.2 e salpeter evolutionActually it is well known that the deposition of saltpeter (nitrocalcite) and all the other cave nitrates is driven by nitrifying bacteria, but in the early times of cave science, there were several supposed origins for nitrates in caves. e most curious among them referred to the upliing of nitric gas from deep inside the earth (Zimmermann I MMERMANN 1788), or postulated the deposition by electric currents (Giovene IOVENE 1819). Only in 1839 the Danish LUND put forth the hypothesis that: Salpeter earth (in caves) derives om surface organic matter However, this theory took over 150 years to be demonstrated: in fact only in 1981 C. Hill I LL established that: nitrates leached om the surface soils, transported into caves by seeping waters, are deposited by the aid of the nitrogen bacteria Nitrobacter Moreover, it is presently proved that the same mechanism may also leach nitrates from guano, rat droppings, and urine and deposit them into cave earth. erefore it is only in very peculiar cave environments (like the volcanic caves) that salpeter may have an inorganic origin related to weathering of basaltic rock (Hill ILL and AND Eller LLER 1977).2.3 e phosphate depositione sources for PO4 -3 ions to produce minerals in caves are normally represented by bones and/or guano deposits inside the cave. e reaction between phosphoric acid, cave walls, clay and sand in the oor and/or other minerals dispersed in the hosting rock is absolutely an inorganic process, but the transformation of organic phosphorous into PO4 -3 ions seems to be always driven by the micro-organisms ruling the complex mineralization (mainly oxidation) processes of the organic mat inside a cave. erefore probably almost all of the known cave phosphates are at least partially biogenic products. No specic study on this topic has been done until now.2.4 e guano digestionMineralization of guano is a complex mix of dierent reactions, many of which are surely biologically driven. e previously described, related to the sulfur cycle, the saltpeter and phosphate evolution are among them and surely the most important, all of them occurring inside guano deposits. Micro-organisms may reasonably control many other processes, like those causing the deposition of halite, gypsum, iron and manganese oxides-hydroxides. Until present no specic study on the eventual biologically driven guano reactions has been done. Recently, in a dierent context (without guano), it was possible to demonstrate that the iron oxides-hydroxides depositing inside Odyssey Cave, Bugonia, Australia are surely a biogenic mineralization: here the bacteria Lepothrics spp and Gallionella sp were proved to precipitate ferrihydrite with characteristic morphologies (Contos ONTOS 2001; Contos ONTOS et al., 2001).2.5 Biogenic speleothems in silica-rich cave environmentse presence of high silica content in the cave wall and/or sediments may allow the development of peculiar micro-organisms which may in turn give rise to biogenic mineralizations. In some volcanic caves of Japan (Kashima ASHIMA et al., 1989) the development of several silica coralloids and helictites has been found to be strictly related to the presence of colonies of diatoms (genus Melosira). In fact these speleothems consist mainly of skeletons of such organisms that are cemented by small amount of silica. eir presence is strictly conned to the rst part of the caves where a little of the external light can still reach the colonies of Melosira, because they need the energy supplied by the light to live. e light control is evident not only by the fact that these speleothems develop only in the threshold zone but also by their shape, which is always pointing towards the cave entrance. Beside this proven occurrence of biologically controlled speleothems in silica rich cave environment, there are several other cases in which a biogenic origin seems to be highly probable. Filamentous organic structures have been reported in opal coralloids from quartzite caves in Venezuela (Onac NAC et al., 2001) and in opal-sulfur speleothems in a gypsum cave of Sicily (FORTI AND ROSSI, 1987). Moreover in many of the lava tubes weathering of basaltic rock caused the evolution of a widespread amorphous silica moonmilk, extremely rich in organic matter (FORTI, 2002; CALAFORRA et al., 2008), thus suggesting that the weathering process is probably driven by micro-organisms. In Pico Island (Azores) there are gigantic opal owstones (up to 5 or 6 meters (m) long and over 1 m thick) inside the Argar do Carbalo volcanic cave, which seem to derive from the diagenesis of the previously described silica moonmilk and, therefore, should be considered biogenic speleothems.2.6 Biogenic carbonate speleothemsIt is presently well accepted that by far the large majority of carbonate speleothems developing in a cave environment are

PAGE 138

15th International Congress of Speleology Rare Species 743 2009 ICS Proceedings absolutely inorganic, however at least some of these deposits undergo biological control. Fungi, algae and bacteria have all been implicated in the precipitation of carbonate dripstones but until recently it was not proved if these micro-organisms helped actively (by driving the chemical process) or passively (serving as crystallization nucleus) in the deposition, or if they were just accidentally buried there, which seems rather to be the more common occurrence. Perhaps the best example of these processes are represented by the gigantic stromatolitic stalactites developing worldwide just at the entrance of many caves mainly in the tropical area. e algae colonies responsible for their development need strong light to survive and therefore such formations are present only up to the threshold zone. Over a score of several tens of years, an extensive documentation of the microbial precipitation of calcium carbonate exists in non-cave carbonate/travertine literature (Ehrlich H RLICH 1996), their direct involvement in cave carbonate deposition have only been demonstrated very recently (Contos O NTOS et al., 2001; Contos ONTOS 2001). In her experiments CONTOS proved the active role played by bacteria in precipitating calcite crystals, and also experimentally evidenced that micro-organisms in Nullarbor caves (Australia) control also the shape of the generated calcite crystals, which are very dierent from those of the inorganic precipitated ones. Deposition of dolomite within cave environment was recently strictly related to the presence of bacteria (VASCONCELOS, 1995; VAN LITH, 2001; PANIERI et al., 2008). Fungal hyphae may act as nuclei for crystallization and a site for attachment for crystals. Algae can trigger the precipitation of calcium carbonate from solution, and may subsequently trap and bind the particles to carbonate speleothems. is may occur as the algae change the microclimate by respiring carbon dioxide (CO2) and consequently causing the CaCO3 to precipitate. Anyway due to their photosynthetic nature, except in show-cave environments, algae will only contribute to carbonate deposition at the entrance and twilight region of the caves, being responsible for the evolution of the previously cited stromatolithic stalactites, which may sometimes reach even gigantic dimension in the tropical areas. Finally bacteria which utilize CO2 (like othrix in the sulfur cycle) have been proved to cause accelerated carbonate speleothem growth. Microbiological reactions seem to be frequently responsible for moonmilk deposition. In fact the two most common mechanisms for the evolution of moonmilk (Hill I LL and AND Forti ORTI 1997) are: 1biochemical corrosion of the bedrock by organic acid produced by microorganisms (AA rt hrobacter, Flavobacterium, Pseudomonas); and 2active precipitation of moonmilk by bacteria (Macromonas bipunctata). Finally it is important to note that microorganisms should be fundamental also in the deposition of moonmilk made by dierent minerals: not only in the case of the amorphous silica already cited, but also when the moonmilk is made of gypsum, amorphous silica, etc. (Forti O RTI 2000) even if no specic studies have been done on this topic at present.3 e Role of Plants and Animals in the Evolution of SpeleothemsIf the actual knowledge on the role played by microorganisms in speleothem evolution is surely not exhaustive but fairly good, the situation is far worse when considering the eect of plants and animals over the secondary chemical deposits in caves. In fact, in these two elds the research is extremely scarce and the available few papers always deal with spot observations without any attempt to consider the topic from a general point of view. Presently its well accepted that plants (mainly roots) and remnants of animals (spider nets, bones, etc) may passively improve the development of speleothems by enhancing capillary migration of waters to places where evaporation may occur. Further other cave formations exist, the genesis of which are directly related to plants and/or animal. Plants may induce the genesis of the low pulp density peat stalactites, which are generated by evaporation of suspensions coming from the surface through porous cap rock (BURKE, 1967). Animals with their dejections are directly responsible for the development of organic formations without any xed chemical composition like amberat speleothems, related to the dung and urine of cave rats (SHOPOV, 2004), and mumijo, which is specic of the caves in the high mountains of central Asia where it is secreted from the excrements of animals like rabbits or mice which eat the peculiar grass and shrubs of these regions (HILL AND FORTI, 1997). Moreover animals, like bats, may deposit huge amounts of excreta in caves, which later undergo bio-mineralizations driven by micro-organisms. Anyway, in most cases, an active involvement of plants and animals in the evolution of speleothems is still speculative.3.1 e inuence of roots on speleothem growthe root apparatus is the single portion of a tree which may somehow interact with speleothem evolution. e interaction may aect both the morphology of the speleothem (passive eect) and its chemical composition (active eect). However, plants normally cannot directly control the mechanisms of chemical deposition deep into the caves, because they need light to survive and their roots can hardly reach depths of several tens of meters. On the contrary in the show caves, where light is articially supplied, plants oen become not only an element of

PAGE 139

Rare Species 744 2009 ICS P roceedings 15th International Congress of Speleology disgurement, but also may lead to a halt in the calcium carbonate deposition or even cause the corrosion of speleothems, due to the acid secretion of their roots. A peculiar lithogenetic eect induced by plants was described in the gypsum area of Bologna (Forti O RTI 1983): with the life activity of plants causing a concentrated local increase of CO2 close to the root apparatus that allows the development of the incongruent dissolution of gypsum which leads to the evolution of carbonate speleothems to form layers inside gypsum owers or even thin, pure calcite owstones. e morphological (passive) eects induced by roots on speleothems are much more frequent and evident: in fact when roots enter cave voids their surface may become a preferential area for the ow of the seeping water and, if the suitable environmental conditions (diusion of CO2 and/or evaporation) exist, for calcium carbonate deposition. is mechanism causes the evolution of peculiar stalactites and columns, with a tilted oen anastomized shape, over which several pseudo-helictites grow: these speleothems, growing over roots, have been observed with the same characteristics all over the world and are normally called rootsicles (Hill I LL and AND Forti ORTI 1997). Finally, in the wet tropical environment, the root apparatus of large trees may become the main driving factor for the evolution of peculiar speleothems, called Showerhead, which were rst described inside Brazilian caves (Lino I NO 1989) and then observed in many other tropical areas. Showerheads are cone-shaped, stalactitic speleothems from which a steady or intermittent shower of water can emerge: in many cases they develop along large ssures widened by the presence of roots which in turns are partially or totally transformed into rootsicles. 3.2 e inuence of animals on speleothem growthWhen considering the inuence of animals on chemical cave deposits, it must be clear that lithogenetic phenomena like corals or others biogenic structures growing inside sea caves cannot be considered speleothems, because a cavern environment is normally not fundamental for their genesis and/or evolution. erefore on the basis of the existing literature it should be stated that the inuence of animals on speleothems is extremely scarce. In reality it is highly probable that upper living organisms should be able to inuence specic speleothems inside peculiar cave environments, therefore the lack of papers on this topic would be the consequence of the scarcity of specic observations instead of the rarity of the event. Most of the existing papers deal with marine caves in which the underwater biogenic calcite overgrowth on preexisting continental (inorganic) speleothems is induced by serpulids (Antonioli NTONIOLI et al., 2001). is kind of overgrowth rarely gives rise to specic new speleothems, normally just enlarging the pre-existing speleothem, the morphology of which remains reasonably unaected. Sometimes, if the environmental conditions are suitable, the biogenic deposition induced by serpulids may allow for the development of single biogenic formations, the most characteristic of which are presently the biogenic trays observed inside the submerged Lu Lampiune cave in Apulia (Onorato NORATO et al., 2003). e speleothems inside this cave consist of big (up to 2 m long and 40 centimeter (cm) in diameter) clearly deected stalactites: in fact their tip always points towards the dominant water ow inside the cave: the resulting growth direction changes from sub horizontal close to the cave entrance to almost vertical in the cave bottom. e shape of these speleothems is not conical but attened with the major axis being 3 to 4 times greater than the smaller one and with the tip larger than the base. e analysis of the internal structure evidenced the absence of even a small pre-existing continental stalactite, being totally developed due to a biogenic deposition in the marine environment. Morphologically they are extremely similar to the gypsum trays described inside the caves of New Mexico (Calaforra ALAFORRA and AND Forti ORTI 1994). In both these occurrences the deposition is controlled by the same agent (they develop against the ow direction), the uid being the only dierence (water in the Lu Lampiune cave and air in the New Mexico caves). Practically the direction of evolution of these biogenic stalactites is controlled by the serpulids, which must lter the water to obtain the necessary trophic support, and therefore tend to direct only upstream the water ow. e apical part is attened and enlarged with respect to the basal section of the stalactites for the same reason: in fact the trophic support is highest at the apex rapidly decreasing toward the base, thus allowing the enlargement of the tip and avoiding the radial growth of the other part of the trays. Until now only one type of large biogenic owstone made by upper living organisms has been described from a continental karst system in the southern part of Italy (Poluzzi O LUZZI and AND Minguzzi INGUZZI 1998). e Vallone Cufalo gypsum cave (Verzino, Italy) is an active sinkhole with a river owing inside. It hosts a large owstone consisting of a gently terraced calcite crust up to 50 cm thick, 4 to 5 m wide, and some tens of meters long, covering the cave oor along the subterranean stream. Its complex genesis has been attributed to the large community of larvae of a troglobitic insect (TT ric optera wormaldia), living inside the cave over a large deposit of anthropogenic organic matter (olive oil factory waste discharged every year into the cave). eir life processes cause the production of large amounts of CO2 which in turn react with the

PAGE 140

15th International Congress of Speleology Rare Species 745 2009 ICS Proceedings water saturated in gypsum thus causing the deposition of calcite just around the worms. e nal morphology of the owstone was the result of a combined action of larvae living activities and kinetic energy of the owing water: in fact while the larvae were alive the tubes were bent upstream to catch as much fresh water as possible, but aer their death and/or evolution the further deposition of CaCO3 was controlled by the kinetic energy of the water transforming the upper part in a normal owstone evolving downstream. It is hard to believe that the Vallone Cufalo would be the only cave in the world in which the environmental conditions are suitable for an active upper living organism control over speleothems; therefore it seems reasonable that chemical deposits might develop due to the presence of animals in many other caves. 3.3 e inuence of humans on speleothem growthNormally humans are responsible for speleothem destruction and not the reverse: however, some examples exist of chemical cave deposits induced by the presence of humans. e early human frequentation of caves sometimes allowed for the development of peculiar speleothems: like those in the Romanelli cave (Apulia) where the organic compound mellite, originated by the reaction between coal, food remains and terra rossa within an ancient hearth, gave rise to euhedral honey shining large crystals and owstones (GARAVELLA AND UAGLIARELLA, 1974). In the second half of the XX century cavers were responsible of the development of a peculiar kind of speleothem in many natural cavities: the carbidimites (TUCKER, 1985). ey are composed of vaterite self-transforming into aragonite and they developed over and/or close to huge deposits of exhausted carbide thanks to the reaction between the calcium hydroxide present in the deposits and the CO2 of the cave atmosphere. ey oen gave rise to peculiar speleothems similar to small ice cream cones (IOWA GROTTO, 1959). Luckily these speleothems are no longer under development because cavers learned to take exhausted carbide out of the cave and more recently the light emitting diode (LED) technology solved denitively the problem.4 Final Remarksis short, and surely not exhaustive, overview on the biotic inuence over the genesis and the evolution of speleothems clearly puts in evidence the very important role played by living organisms, mainly micro-organisms, over lithogenetic cave processes (speleothems and cave minerals evolution). e complex biochemical reactions involved in the development of the dierent deposits, though still not completely understood, clearly have greater interest and importance far exceeding simple speleogenesis. Normally these are low-enthalpy reactions (which involve low to very low quantity of energy) and their knowledge is fundamental to improve our understanding of the natural mechanisms by which even ore bodies of economic interest are formed and then mobilized. However the study of biologically driven speleothems is also fundamental to enhance the knowledge on peculiar environments like the chemioautotrophic ones, which are presently not well known despite their scientic interest. It is therefore reasonable to expect an increase in the cooperation between biologists and geochemists in the near future in order to obtain a fast improvement in the study of these phenomena. In conclusion it can now be stated that: Speleothems denitely do not develop via vegetative growth But without biogenic control caves would be very poorly decorated.ReferencesALDROVANDI U. (1648) Musaeum Metallicum Bologna, Ferronius, 979 p. Antonioli N TONIOLI F., Silenzi ILENZI S., Frisia RISIA S. (2001) Tyrrhenian holocene palaeoclimate trends from spelean serpulids. ua ternary Science RReviews, 20(15), 1661. Beaumont E AUMONT J. (1676) Two letters concerning rockplants and their growth. P hilosophical TT ransactions OO f the R Royal Society, 11 (129), 732. BURKE, A.R. Deposition of peat from aqueous suspension: natural occurrence of stalactitic and related forms. NN a ture 214, 5087, 532. Calaforra A LAFORRA J.M., Forti ORTI P. (1994) Two new types of gypsum speleothems from New Mexico: Gypsum trays and Gypsum dust. NN a tional Speleological Society Bulletin 56, 32. CALAFORRA, J.M., FORTI, P., LES, J. (2008) Lava tubes of the Rohio lava eld (Rapa Nui, Chile: exploration and scientic interests Proceedings of the 14th International Symposium on Vulcanospeleology, 1-5 September 2008 Jeju Island Korea, p.46. CONTOS, A. (2001) Biomineralisation in Caves PhD esis, University of Sidney, 221 p.

PAGE 141

Rare Species 746 2009 ICS P roceedings 15th International Congress of Speleology Contos ONTOS A.K., James AMES J.M., Heywood EYWOOD B.R., Pitt ITT K., Rogers OGERS P.A.W. (2001) Morphoanalyses of bacterially precipitated subacqueous calcium carbonate from Weebubbie Cave, Australia. Geomicrobiology Journal. 8(3), 331. EHRLICH (1996) Geomicrobiology. M. Dekker Ed., New York, 716 p. Forti O RTI P. (1983) Un caso di biocarsismo nei gessi: le i noo rescenze sopra i massi aoranti. Sottoterra 66, 21. Forti O RTI P. (2000) I depositi chimici presenti nella grotta Serano Calindri. Sottoterra 110, 31. FORTI, P. Biogenic speleothems: an overview. International Journal of Speleology 30A /4, 39n56. Forti O RTI P., Rossi OSSI A. (1987) Le concrezioni poliminerali della Grotta di Santa Ninfa: un esempio evidente dellinuenza degli equilibri solfuri-solfati sulla minerogenesi carsica AA t ti e Memorie Commmissione Grotte EugenioBoegan 26, 47. FORTI, P., GALLI, E., ROSSI, A. (2009) Minerogenesis in the Naica Caves (Chihuahua, Mexico) XV International. Spepeleological. Congress, Kerville, Texas. GARAVELLA, C.L., UAGLIARELLA, A.F. Mellite di Grotta Romanelli (Otranto) Periodico di Mineralogia, 43(1), 39. Giovene I OVENE G.M. (1819) Del nitro e degli altri sali che laccompagnano. AA t ti AA c. Soc. Sci. Modena, 18(2), 1-7. Hill I LL C.A. (1981) Origin of cave salpeter. NN SS Bull 43(4), 110. Hill I LL C.A., ELLER, P.G. (1977) Soda-niter in earth cracks of Wupatki National Monument. NN SS Bulletin 39(4), 113. HILL, C.A., FORTI, P. (1997) Cave minerals of the World, N.S.S. Huntsville, 464 p. IOWA GROTTO, 1959 Haines Cave, Wisconsin Iowa Cave Book, 3, art.44, 1. Kashima A SHIMA N., Ogawa GAWA T., Hong ONG S.H. (1989) Volcanogenic speleo-minerals in Cheju Island, Korea. J. Spel. Soc. Jap., 14, 32. LINO, C.F. (1989) Cavernas: o fascinante Brazil subterraneo Rios, Sao Paulo Ed., 280 p. Lund U ND P.W. (1839) Extract from a letter on the fossil mammifera discovered by him in Brasil. AA n n. NN at. Hist., London, 3(17), 235. NORTHUP, D.E., REYSENBACH, A.L., PACE, N.R. (1997) Microorganisms and speleothems in Cave Minerals of the World, Hill C.A., Forti P. (Eds.), NSS, Huntsville, 261. Onac N AC B.P., Veni ENI G., White HITE W.B. (2001) Depositional environment for metatyuyamunite and related minerals from Caverns of Sonora. European J. of Mineralogy 24(1), 135. ONORATO, R., FORTI, P., BELMONTE, G., POTO, M., COSTANTINI, A. (2003) Grotta sottomarina di Lu Lampiune, novit esplorative e prime indagini ecologiche alassia Salentina 26, 55. PANIERI, G., FORTI, P., GASPAROTTO, G., SOLIANI, L. (2008) Studio delle inclusioni solide presenti nei cristalli di gesso della Grotta delle Spade (Naica, Messico) Mem. Ist. It. Spel. s. II, 19, 335. Poluzzi O LUZZI A., Minguzzi INGUZZI V. (1998) Un caso di biocostruzione in un ambiente di grotta. Mem. Ist. It. Spel. s. II, 10, 93. Sasowsky A SOWSKY I.D., Palmer ALMER M.V. (1994) Breakthroughs in K arst Geomicrobiology and RRedox Geochemistry. Symposium Abstract, Karst Waters Institute, 112 p. SHAW, T.R. (1997) Historical introduction in Cave Minerals of the World HILL C.A., FORTI P. (Eds.), National Speleological Society, Huntsville, 28. SHOPOV, Y. Sediments: Biogenic in Encyclopedia of Caves and Cave Sciences, J.Gunn (ed.), 636. Tourneford O URNEFORD J.P. (1704) Description du labyrinthe du Candie, avec quelques observations sur laccroissement et sur la gnration des pierres. Mem.R R .A A cad. Paris, 406.

PAGE 142

15th International Congress of Speleology Rare Species 747 2009 ICS Proceedings TUCKER, T.E. Moonmilk and related minerals in Missouri caves. NN S S Bulletin 47, 66. VAN LITH, Y. (2001) e role of sulphate reducing bacteria in dolomite formation PhD thesis Swiss Federal Institute of Technology, Zurich, 186 p. VASCONCELOS, C., MCKENZIE, J.A., BERNASCONI, S., CRUJIC, D., TIENS, A. (1995) Microbial mediation as a possible mechanism for natural dolomite formation at low temperature NN a ture 377, 220. ZIMMERMANN, E.A. (1788) Voyage la nitrire naturelle que se truoe Molfetta dans la terre de Bari en A A pu lia. Venezia, Storti, 44p.

PAGE 143

Rare Species 748 2009 ICS P roceedings 15th International Congress of Speleology CONSERVATION AND MANAGEMENT OF THE BRAZILIAN FREETAILED BAT: COLONY SIZE AND ACTIVITY PATTERNS USING THERMAL IMAGINGNN ICKO O LA A Y I. HR R ISTO TO V1, MAR AR GR R IT T BET T KE1,2, T T HO O MA A S H. KUN N Z1 1Center for Ecology and Conservation Biology, Boston University, Boston, MA A 022152 DDepartment of Computer Science, Boston University, Boston, MA A 02215 Abstract e Brazilian free-tailed bat (TT ad arida brasiliensis) is one of the most abundant and conspicuous species in North America, where it provides one of the most important but underappreciated agroecosystem services to mankind. During peak lactation individual bats consume up to two-thirds of their body mass in insects each night. When multiplied by the hundreds of thousands of bats roosting in caves, bridges and other sites, enormous quantities of some of the most damaging pests to agriculture are consumed nightly. Historical records, however, indicate that there has been a severe decline in numbers during the last century. An accurate assessment of the agricultural and ecological value as well as the need for management and conservation of this species, require knowledge of nightly and seasonal patterns in colony size, and foraging activity. We present results from the application of a new method for censusing and quantifying the nightly and seasonal patterns of activity in Brazilian free-tailed bats using thermal infrared imaging and computer vision processing. ermal infrared cameras were positioned both outside inside Carlsbad Cavern and the colony size and nightly activity of the colony was recorded in high temporal resolution. Automatic computer vision algorithms were used to analyze the thermal data producing graphical representations of the activity pattern for the entire colony. Results indicate high variation in the size and activity of the colony as a function of seasonal reproductive behavior and local weather conditions. Our results provide an accurate estimate of colony size and the rst quantitative description of nightly and seasonal time budgets for this species. Such information is essential for modeling the impact of Brazilian free-tailed bats on agriculture and for understanding colony and population dynamics in response to changes in weather patterns and food availability.

PAGE 144

15th International Congress of Speleology Rare Species 749 2009 ICS Proceedings BIOGEOGRAPHICAL DISTRIBUTION OF SUBTERRANEAN FAUNA IN APULIA ITALY, IN THE CONTEXT OF THE PALEOGEOGRAPHICAL EVOLUTION OF THE AREASAlv LV ATORe E INgusci GUSCI O1, Em M ANuel UEL A R R Ossi SSI1, MARi I O PARise ISE2 1 Biospeleology Laboratory S. RR uo, NN ard LE, Italy (laboratorio@avanguardie.net)2 NN ational RResearch Council of Italy, IRR PI, Bari, Italy (m.parise@ba.irpi.cnr.it) Apulia (South-East Italy) is among the most remarkable areas as regards biospeleology in Italy, being the only continental region of the country where the biogeographical region coincides with the administrative one. Based upon analysis and distribution of the subterranean fauna, and the paleogeographic reconstruction of the Mediterranean Basin, Italy is subdivided into seven biogeographical regions, each one being characterized by a certain degree of fauna homogeneity. e largest biogeographical region is the Apenninic, which covers a large part of the Italian peninsula. Apulian troglobites represent a completely distinct set from the Apenninic subterranean fauna; the two regions have in common very few troglobites, whilst a total lack in Apulia of typical forms of the Apennines has to be registered. On the other hand, several troglophiles are present in Apulia, probably entered in the region during the uaternary time. e rst studies about Apulian subterranean fauna, which illustrated a distinction among the three karst sub-regions in Apulia (from North to South, Gargano, Murge, and Salento), are discussed in this contribution. en, a quantitative analysis of troglobite presence and distribution in the Apulian caves is presented, aimed at highlighting the real dierences between the three sub-regions, and the reasons for them as well. A particular focus is given to Salento, the southernmost portion of Apulia, that is the area where the highest number of troglobites and endemic species is recorded.1. Introductione distribution of biota is controlled by a combination of factors, that include, but are not limited to, the geological history, the chemical characters of the aquifer, the connections between near regions, and so on. Central-southern Europe, with particular regard to the Mediterranean area, presents a taxonomic diversity and specie richness higher than other parts of Europe. is derives from a number of reasons, including the presence of extensive, temperate to Mediterranean climate, karst areas, extensive shallow embayments in the Tertiary, the occurrence of a salinity crisis in Miocene, the absence of Pleistocene glaciation, the common presence of anchihaline habitats in the coastal areas, and abundant freshwater subterranean habitats (HOLSINGER, 1993). e Mediterranean Sea had therefore a very complex history that includes the event of drying up about 6 million years ago, during the Messinian crisis. is event likely resulted in a greater invasion rate of the subterranean realm from marine waters, and probably contributed to the great subterranean diversity observed in many Mediterranean countries (GIBERT and AND CULVER, 2005). In Italy, a total of 265 stygobionts, 321 troglobionts, and an additional 317 troglobionts from the undergound shallow medium (MSS or milieu souterrain superciel; JUBERTHIE et al., 1980) have been described (RUFFO and AND STOCH, 2005). Based upon analysis and distribution of the subterranean fauna, and the paleogeographic reconstruction of the Mediterranean Basin, Italy is generally subdivided into seven biogeographical regions, each one being characterized by a certain degree of fauna homogeneity (Fig. 1). Most of them include dierent administrative regions, encompassing the legal boundaries. e largest biogeographical region is the Apenninic, which covers a large part of the Italian peninsula. ere are only three exceptions where the biogeographical region coincides with the administrative one: two out of three are the major islands, Sardinia and Sicily. e third is Apulia, the heel of the Italian boot, located in the southeastern part of the country. is points out to the remarkable biogeographical importance of Apulia region, which in turn stems from its geological history. In addition, Apulia hosts a great variety of endemic species.2. PaleogeographyIn the last years, a large amount of works have been produced about the paleogeographical evolution of the central Mediterranean, mainly in the aermath of discoveries of dinosaur footprints that pushed the geologists to adapt the previously established reconstructions in the new light provided by these additional data. e most

PAGE 145

Rare Species 750 2009 ICS P roceedings 15th International Congress of Speleology signicant element of central Mediterranean is Adria, a structural element involving the crust and lithosphere, whether considered an independent microplate or a promontory of the African plate. From the paleogeographic point of view, a variety of environments including dry land, at carbonate islands, tidal ats, marshes and shallow lagoons, deep sounds and basins may have coexisted on Adria. As pointed out by BOSELLINI (2002), many authors used the name Apulia for the same tectonic element, thereby generating considerable confusion. e easternmost platform of central-southern Italy, the Apulia Platform, is a JurassicCretaceous shallow-water carbonate bank bounded by deep-water basins and one of the so-called peri-Adriatic carbonate platforms (DARGENIO, 1976; ZAPPATERRA, 1994; BOSELLINI, 2002). It belongs to Adria, like other JurassicCretaceous platforms and deeper basins. e Apulia carbonate platform is essentially a Mesozoic paleogeographic element, which, in large part, acted as a rigid block during Alpine (Tertiary) orogenesis. During the Mesozoic, it was an isolated carbonate bank situated along the southern margin of the Mesozoic Tethys Ocean and was created during Early Jurassic riing of the margin. Today, it is partly buried under the Apennine thrust sheets and partly constitutes the weakly deformed foreland of the Apennine and Dinaric chains (CHANNELL et al., 1979; UNDERHILL, 1989; BOSELLINI et al., 1999a, 1999b). e dinosaur footprints and tracks recently found in the Apulia carbonate platform (GIANOLLA et al., 2000; NICOSIA et al., 2000) put strong constraints to plate tectonic and paleogeographic reconstructions of the eastern Mediterranean area: according to the new constraints, Adria was an African Promontory and the Apulia Platform was not an isolated Bahamian bank, but rather a sort of Florida Peninsula, a carbonate peninsula directly attached to the Cyrenaica spur of Africa during the Jurassic and Early Creatceous, thus subdividing the Mesozoic Mediterranean into a western Ionian basin and an eastern Levantine basin (BOSELLINI, 2002).3. Distribution of hypogeous fauna in Apuliauantitative analysis carried out in caves and wells of Apulia in the last years showed the presence in the region of a total number of 42 troglobites (Table 1), 29 of which are aquatic, whilst the remaining 13 are terrestrial. In Table 2, the details of the distribution in the three karst subregions are given, also as regards the endemic species. ese numbers indicate that the majority of Apulian troglobites (69%) consist of aquatic animals, as a likely consequence of the many trasgressive phases that the region experienced during its geological history, and that were one of the main reasons for the limited distribution of the terrestrial fauna. Colonization of subterranean waters may have occurred by adaptation of some organisms coming from surface, fresh or marine, waters through a passage in anchialine or interstitial waters, and eventually to subterranean waters. Endemics in Apulia are 15 (9 aquatics and 6 terrestrials), thus representing 52% of the whole Apulian hypogeous fauna. Some of them are exclusive of this region, even as regards genus and family. Salento is the sub-region that presents the highest number of both troglobites and endemic organisms. e presence in Apulia of a specialized fauna, together with numerous endemic organisms, has to be related to the widespread, multi-phase, karstic phases occurred in the region, as well as to the paleogeographic history of Apulia, above recalled. Apulian troglobites represent a completely distinct set from the Apenninic subterranean fauna; the two regions have in common very few troglobites, whilst a total lack in Apulia of typical forms of the Apennines has to be registered. On the other hand, several troglophiles are present in Apulia, probably entered in the region during the uaternary time. e rst studies about Apulian subterranean fauna highlighted a distinction among the three karst sub-regions in Apulia: from North to South, Gargano, Murge, and Figure 1: e Italian stygofaunistic proinces (aer STOCH, 2008).

PAGE 146

15th International Congress of Speleology Rare Species 751 2009 ICS Proceedings OOrder Family SpeciesA A /TTD D istributionEndemic Hachondrida Desmoxyidae Higginsia ciccaresei AS Y Acarina Halacaridae Lohmannella stammeri AS Y Acarina Ixodidae Ixodes vespirtilionis TGM N Araneae Anapidae Zangherella apuliae TS N Pseudoscorpiones Chthoniidae Chthonius ruoi TS Y Pseudoscorpiones Syarinidae Hadoblothrus gigas TMS Y Cyclopoida CyclopidaeD D iac yclops antricola AGS N Cyclopoida CyclopidaeD D iac yclops clandestinus AG N Cyclopoida Cyclopidae Halicyclops troglodytes AS N Cyclopoida Cyclopidae Metacyclops stammeri AS N Cyclopoida Cyclopidae Metacyclops subdolus AS N Harpacticoida AmeiridaeN N i tocrella stammeri AGMSN Harpacticoida Ameiridae Parapseudoleptomesochra italica AGMSN Harpacticoida Ameiridae Psyllocamptus monachus AS Y Harpacticoida Parastenocarididae Parastenocaris proserpina AG N Podocopida Loxoconchidae Pseudolimnocythere hypogea AGMSY Podocopida CandonidaeT T r apezicandona italica AG Y Amphipoda bogidiellidae Bogidiella chappuisi AS N Amphipoda Hadziidae Hadzia adriatica AGM Y Amphipoda Hadziidae Hadzia minuta AS Y Amphipoda Metaingolellidae Metaingolella mirabilis AS Y Amphipoda NiphargidaeN N i phargus gr. orcinus AGS N Amphipoda NiphargidaeN N i phargus longicaudatus AGS N Amphipoda NiphargidaeN N i phargus parenzani AS N Amphipoda Pseudoniphargidae Pseudoniphargus adriaticus AGM N Amphipoda NiphargidaeN N i phargus aquilex AGM N Amphipoda Salentinellidae Salentinella angelieri AG N Amphipoda Salentinellidae Salentinella gracillima AS Y Decapoda PalaemonidaeT T yp hlocaris salentina AGMSY Isopoda TrichoniscidaeA A e gonethes cervinus TG N Isopoda Trichoniscidae Castellanethes sanlippoi TM Y Isopoda Microparasellidae Microcharon arganoi AG Y Isopoda Trichoniscidae Murgeoniscus anellii TM Y Isopoda TrichoniscidaeT T ric honiscus ruoi TS Y Mysidacea Lepidomysidae Spelaeomysis bottazzii AGMSY Mysidacea Stygiomysidae Stygiomysis hydruntina AS Y ermosbaenacea Monodaellidae Monodella stygicola AGMSY Glomerida Glomeridae Glomeris stammeri TS Y Collembola Hypogastruridae Bonetogastrura cavicola TS N Collembola ParonellidaeT T r oglopedetes ruoi TMS N Diplura Campodeidae Plusiocampa dolicophoda inguscioi TG Y Coleoptera Carabidae Italodytes stammeri TMS YTable 1: Hypogeous species of Apulia. e third column (A/T) indicates whether the species is aquatic (A) or terrestrial (T). Key to the fourth column (Distribution): G = Gargano; M = Murge; S = Salento. Salento. RUFFO (1955), for example, identies among the Apulian troglobites: i) balcanic species, with trans-Adriatic distribution; ii) species with a southern paleogenic transJonian distribution; iii) species with a paleo-mediterranean distribution. A clear distinction was, in particular, done between the hypogeous fauna in Gargano and that in Murge-Salento (RUFFO, 1955; ARIANI, 1982). According to these studies, Gargano was characterized by few evolved troglobites, a fact that was considered as due to the unfavourable ecological characters of the deep

PAGE 147

Rare Species 752 2009 ICS P roceedings 15th International Congress of Speleology Troglobites Troglobites present only in one sub-region Endemic troglobites aquaticterrestrialtotalaquaticterrestrialtotalaquaticterrestrialtotal Gargano163195 2 7213 Murge86141 2 3-22 Salento2182911 6 177310Table 2: Troglobite distribution in the three Apulian karst sub-regions. subterranean waters, and to isolation of the deep water table from that shallow. ese considerations brought the above authors to conclude that there was no connection between Gargano and the rest of Apulia during Middle Miocene. Recent researches, however, have documented in coastal areas of Gargano the presence of troglobites of ancient origin, as Spelaeomysis bottazzii and T T yphlocaria salentina, thus delineating a dierent paleogeographical framework (ROSSI and AND INGUSCIO, 2001, 2003). Figure 2 illustrates the distribution of 6 among the most signicant hypogeous species in Apulia. Hadoblothrus gigas is the Apulian terrestrial troglobite which shows the most advanced adaptations. e genus Italodytes (Fig. 2B) is present in Apulia with two sub-species (Italodytes stammeri stammeri and Italodytes stammeri antoniettae). e genus Monodella, considered a Tethyan relict, has two species (Fig. 2C): one (Monodella stygicola) is in Apulia (RUFFO, 1949), whilst the other (Monodella argentarii) has been documented in Tuscany, and precisely in a cave in the Argentario Promontory. It has to be noted that about 50 million of years ago this Promontory, and part of Apulia were among the few emerged territories of the Italian peninsula. As for the genus Salentinella, Apulia hosts two distinct species (Fig. 2D): Salentinella angelieri, and Salentinella gracillima. e rst, highly euryhaline, lives almost exclusively in coastal areas. Salentinella gracillima, according to RUFFO (1982) has a paleo-mediterranean marine origin, and colonized subterranean waters during Middle Miocene; other authors, however, disagree with this hypothesis, retaining that the species followed the movements of the Mediterranean coastlines during Pliocene time (PESCE and AND PAGLIANI, 1999). Outside Apulia, the genus TT yp hlocaris (Fig. 2E) presents only two other species, respectively in Cyrenaica and in Palestine. It could represent species with paleomediterranean distribution that today are relict of a fauna which lived in subtropical climate, and later on survived to the post-Pliocene climatic changes only at a few sites. Spelaeomysis bottazzii (Fig.23F) is the more widespread stygobiont (PESCE, 1975a, 1975b), and the only one for Figure 2: Distribution of the most signicant species in Apulia: (A) Hadoblothrus gigas; (B) Italodytes stammeri stammeri (dots), and Italodytes stammeri antoniettae (squares); (C) Monodella stygicola; (D) Salentinella angelieri (squares), and Salentinella gracillima (dots); (E) Typhlocaris salentina; (F) Spelaeomysis bottazzii.

PAGE 148

15th International Congress of Speleology Rare Species 753 2009 ICS Proceedings which data are available about its resistance to biological pollution. Worldwide distribution of the genus let suppose the past existence of correlations among the American, Mediterranean and Indopacic areas. is species feeds out of organic substances present in so carbonate rocks; this characteristic may had contributed to colonization of subterranean waters, following the coastline movements during the Pliocene. e species-area relationship can be considered as the most important factor in predicting species extinction (HOLT et al., 1999; HOLSINGER, 2005). Many hypogeous species are endangered by widespread urbanization, and the eects caused by a number of anthropogenic activities that cause pollution and environmental degradation to the karst (SKET, 1999; PARISE and AND GUNN, 2007). THOMAS and co-workers (2004) predicted the extinction of 15-37% of total species (over 1 million) by year 2050. e outcomes from a study in the Nard district (Salento sub-region) showed that there is a possible trend number of species resources relationship which could lead to the extinction of stygobionts in the near future (MASCIOPINTO et al., 2006), based upon the observed decline in hypogeous aquatic fauna biodiversity. Aquatic hypogeous species, being highly sensitive to changes in the hydrological cycle and to habitat modications caused by anthropogenic activities, have been proposed as indicators of environmental quality in Salento (MASCIOPINTO et al., 2006). To provide an example, Spelaeomysis bottazzii is a stygobiont omnivore that has largely increased in number and size in Salento with respect to other, smaller, species that are less resistant to water quality modications. In fact, Spelaeomysis bottazzii is eurytherm and tolerant to a wide range of salinity.4. Conclusione data presented here about distribution of the subterranean fauna in Apulia point out to the remarkable relevance of this southern Italian region. Among the main factors at the origin of such an importance, the geological history has to be considered. Until the Pleistocene, the three sub-karst regions (Gargano, Murge, and Salento) had insular character, hosting stygobionts of marine origin that are considered to be Tertiary relicts from Tethys, and show anities with Caribbean and Indopacic species. e successive paleogeographical setting still needs further research in order to fully explain the observed specie distribution that does not appear simply to be distinct in Gargano and in Murge-Salento, as so far has been considered.ReferencesARIANI, A.P. (1982) Osservazioni e ricerche su TT yphlocaris salentina e Spelaeomysis bottazzii. Approccio idrogeologico e biologico sperimentale allo studio del popolamento acquatico ipogeo della Puglia. A A n nali dellIstituto di Zoologia dellUniversit di N N apoli 1979-80, 203. BOSELLINI, A. (2002) Dinosaurs re-write the geodynamics of the eastern Mediterranean and the paleogeography of the Apulia platform. Earth S cience RReviews 59, 211. BOSELLINI, A., M. MORSILLI, C. NERI (1999a) Longterm event stratigraphy of the Apulia Platform margin (Upper Jurassic to Eocene Gargano, southern Italy). J ournal of Sedimentary RResearch 69, 1241. BOSELLINI, A., F.R. BOSELLINI, M.L. COLALONGO, M. PARENTE, A. RUSSO, A. VESCOGNI (1999b) Stratigraphic architecture of the Salento coast from Capo dOtranto to S. Maria di Leuca (Apulia, southern Italy). RR i vista Italiana di Paleontologia e Stratigraa 105, 397. CHANNELL, J.E.T., B. DARGENIO, F. HORVATH (1979) Adria, the African Promontory, in Mesozoic Mediterranean palaeogeography. Earth-Science R R eviews 15, 213. DARGENIO, B. (1976) Le piattaforme carbonatiche periadriatiche. Una rassegna di problemi nel quadro geodinamico Mesozoico dellarea Mediterranea. Memorie della Societ Geologica Italiana 13, 1. GIANOLLA, P., M. MORSILLI, A. BOSELLINI (2000) Dinosaur footprints. In e Eastern Margin of the A A pu lia Platform, A. BOSELLINI, M. MORSILLI, C. NERI (Eds.), Guide Book, WG4 Meeting, Vieste, Gargano, p. 34. Gibert, J. and D.C. Culver (2005) Diversity patterns in Europe. In Encyclopedia of caves, D.C. CULVER and B.W. WHITE (Eds.), Elsevier Academic Press, p. 196201. HOLSINGER, J.R. (1993) Biodiversity of subterranean amphipod crustaceans: global patterns and zoogeographical implications. J ournal of NN atural History 27, 821.

PAGE 149

Rare Species 754 2009 ICS P roceedings 15th International Congress of Speleology Holsinger, J.R. (2005) Vicariance and dispersalist biogeography. In Encyclopedia of caves, D.C. CULVER and B.W. WHITE (Eds.), Elsevier Academic Press, p. 591. HOLT, R.D., J.H. LAWTON, G.A. POLIS and N.D. MARTINEZ (1999) rophic rank and the speciesarea relationship. Ecology 80, 1495. JUBERTHIE, C., B. DELAY and M. BOUILLON (1980) Extension du milieu souterrain en zone non calcaire: description dun nouveau milieu et de son peuplement par les Coloptres troglobies. Memoires Biospeleologie 7, 19. MASCIOPINTO, C., F. SEMERARO, R. LA MANTIA, S. INGUSCIO and E. ROSSI (2006) Stygofauna abundance and distribution in the ssures and caves of the Nard (Southern Italy) fractured aquifer subject to reclame water injections. Geomicrobiology Journal 23, 267. NICOSIA, U., M. MARINO, N. MARIOTTI, C. MURARO, S. PANIGUTTI, F.B. PETTI, E. SACCHI, (2000) e Late Cretaceous dinosaur tracksite near Altamura (Bari, southern Italy). G eologica RRomana 35, 231. PARISE, M. and J. GUNN (Eds.) (2007) NN a tural and A A nthropogenic Hazards in Karst Enionments: R Recognition, AA nalysis, and Mitigation. Geological Society of London spec. publ. 279, 202 p. PESCE, G.L. (1975a) On a Stygiomysis (Crustacea, Mysidacea) from the southern Italy. Bollettino d el Museo Civico di Storia NN aturale di Verona 2, 439. PESCE, G.L. (1975b) A new locality for Spelaeomysis Bottazzii with redescription of the species (Crustacea, Mysidacea). Bollettino del Museo Civico d i Storia NN aturale di Verona 2, 345. Pesce, G.L. and T.A. Pagliani (1999) Gli ambienti anchialini della Puglia e la loro fauna. In Il carsismo dellarea Mediterranea, G. BELMONTE, G. CICCARESE and L. RUGGIERO (Eds.), Ed. Del Grifo, p. 89102. Rossi, E. and S. Inguscio (2001) AA n imalia tenebrarum. Biospeleologia pugliese. Ideemutimediali, Nard, 96 p. ROSSI, E. and S. INGUSCIO (2003) Fauna ipogea pugliese: nuovi dati e osservazioni. alassia Salentina suppl. 26, 219. RUFFO, S. (1949) Monodella stygicola n. g. n. sp. nuovo crostaceo Termosbenaceo delle acque sotterranee della Penisola Salentina. AA r chivio Zoologico Italiano 34, 31. RUFFO, S. (1955) Le attuali conoscenze sulla fauna cavernicola della regione pugliese. Memorie di B iogeograa AAdriatica 3, 1-143. RUFFO, S. (1982) Gli antipodi delle acque sotterranee italiane. Lavori della Societ Italiana di Biogeograa, n.s. 7: 139 RUFFO, S. and F. STOCH (Eds.) (2005) Checklist e distribuzione della fauna italiana. Mem. Museo Civico Storia Nat., Verona, Sez. Scienze della Vita 16, 307 p. SKET, B. (1999) e nature of biodiversity in hypogean water and how it is endangered. Biodiversity Conservation 8 (10), 1319. STOCH, F. (Ed.) (2008) Le acque sotterranee. uaderni Habitat, Ministero dellAmbiente 20, 156 p. THOMAS, C.D., A. CAMERON, R.E. GREEN, M. BAKKENES, L.J. BEAUMONT, Y.C. COLLINGHAM, B.F.M. ERASMUS, M.F. DE SEUEIRA, A. GRAINGER, L. HANNAH, L. HUGHES, B. HUNTLEY, A.S. VAN JAARSVELD, G.F. MIDGLEY, L. MILES, M.A. ORTEGA-HUERTA, A.T. PETERSON, O.L. PHILLIPS and S.E. WILLIAMS (2004) Extinction risk from climate change. NN a ture 427, 145. UNDERHILL, J.R. (1989) Late Cenozoic deformation of the Hellenide foreland, western Greece. Bulletin of t he Geological Society of AA merica 101, 613. ZAPPATERRA, E. (1994) Source-rock distribution model of the Periadriatic Region. AA m erican AA ssociation of Petroleum Geologists Bulletin 78, 333.

PAGE 150

15th International Congress of Speleology Rare Species 755 2009 ICS Proceedings DETECTION PROBABILITIES OF KARST INVERTEBRATES IN CENTRAL TEXASJEAN AN K. KR R EJCA A1, BUT T CH WECKER R LY2 1Zara Enironmental LLC, 118 W. Goforth RR d., Buda, TT X 78610 USA A2DDepartment of Biology, TT exas State University, 601 University DDrive, San Marcos, TT X 78666 USA A Protection of federally listed endangered troglobites in central Texas focuses on caves that are occupied by the species. e determination of occupancy is based on presence/absence surveys for those taxa. Under current U.S. Fish and Wildlife Service recommendations, three surveys are used as a standard to determine presence or absence, and certain environmental and seasonal conditions must be met. We used survey data from 23 caves on Camp Bullis Military Reservation, Bexar County, Texas to test the validity of the survey protocols. Presence/absence matrices were created for three cave species, two beetles Batrisodes uncicornis (Coleoptera: Pselaphidae), RR h adine exilis (Coleoptera: Carabidae), and a harvestman Chinquipellobunus madlae (Opiliones: Stygnopsidae). Eleven environmental and seasonal covariates that have been suggested to aect detection probability were tested for t to the detection data. B. uncicornis and RR exilis were determined to have constant detection probabilities of 0.1226 and 0.1875. C. madlae was found to have a survey specic detection probability (average p = 0.2424), and in no case was detectability tied to any of the measured covariates. Parametric bootstrapping was used to simulate the number of surveys needed to have a 5% chance of not detecting the species if they were present at the site. e number of surveys needed ranged from 10 to 22. ese results indicate that more surveys should be performed before determining absence from a site. e results also indicate that most of the time cave species are not available to be surveyed, and we hypothesize that they retreat into humanly inaccessible cracks connected to the cave.1. IntroductionDetection probability (p), or detectability, is the chance that a karst invertebrate will be observed if the cave is occupied by that species. In order for a species to be observed it must be both available (e.g. not hiding in a humanly inaccessible crack) and seen by the researcher. Occupancy (t) is the proportion of sites that are occupied, or the proportion of areas where the species is present. Failure to take into account detection probabilities when using species counts can lead to underestimating cave occupancy, since nondetections in survey data do not necessarily mean that a species is absent unless the probability of detection is one (Mac AC Kenzie ENZIE et al., 2002; Bailey AILEY et al., 2004). If the probability of detection is less than one, then surveys should be designed to account for imperfect detection. Cave organisms are small and live in an environment that is dicult to sample because of constricted crawlways, vertical drops, low oxygen levels, and an abundance of mesocaverns, or tiny cracks and voids connected to the cave but inaccessible to humans. For the sixteen species of federally listed terrestrial karst invertebrates in central Texas, recovery is based on protecting habitat around caves known to contain the species, therefore estimating occupancy of caves is of paramount importance. Monitoring the populations in these caves and conducting surveys in new caves are listed as key components to the recovery strategy (USFWS 1994). e U.S. Fish and Wildlife Service (2006) provides survey recommendations for these taxa and detail that permitted surveyors must have several years of experience with these or similar species under a permit holder. During the three surveys required to ascertain presence or absence of a species in a cave, certain environmental and seasonal conditions must be met. us far these conditions (number of visits, season, temperature, recent rain) have been determined based on non-quantied observations by researchers balanced with an estimation of observer impact on the environment (James Reddell and USFWS Bexar County Karst Invertebrate Recovery Team, pers. comm.). Since newly found caves are rapidly being impacted by development, and the data from early counts of karst invertebrates are being relied upon for guidance of preserve designs, it is imminently important to estimate the utility of the recommended survey protocol with condence. e focus

PAGE 151

Rare Species 756 2009 ICS P roceedings 15th International Congress of Speleology of this study is to determine the detection probabilities for several terrestrial karst invertebrates, to assess whether certain environmental parameters aect detectability, and to use detectability to determine the number of surveys required to be condent in a determination of absence from a site.2. Methods 2.1 Study sitesCaves on Camp Bullis Military Reservation, Bexar County, Texas were used for this study, and the raw dataset along with detailed information about each site is reported in George E ORGE Veni ENI and AND Associates SSOCIATES (2006). Cave sites were subdivided into zones, and these individual zones are the survey units. Surveys were conducted three times per year, during the spring (May), summer (July 15 August 15), and fall (October). ese started in the fall of 2003 and included spring 2007, for a total of eleven sample events. Prior studies have used this method (Elliott L LIOTT 1994) and it is consistent with U.S. Fish and Wildlife Service endangered species survey recommendations (2006).2.2 Detection probabilities, occupancy, and number of surveyse program PRESENCE (Proteus Wildlife Research Consultants, Dunedin, New Zealand) includes markrecapture models modied by Mac A C Kenzie ENZIE et al. (2002) for use with presence-absence data. It was used to analyze the t of several models to the dataset. e rst test was to determine whether our dataset that included multiple years and seasons could be considered closed during the period of the surveys, fall 2003 to spring 2007. Closure means the cave zone did not experience a change in occupancy by the species during the time interval of surveys and is an assumption of the occupancy models (Mac A C Kenzie ENZIE et al,. 2002). To determine closure three models were compared. e rst model considered the detection probability as specic to each survey event, the second as specic to each season, and the third as constant across all survey events. e models were compared using Akaikes Information Criterion (AIC) and AIC weights (Burnham U RNHAM and AND Anderson NDERSON 2002). Once the assumption of closure was validated, detection probabilities were modeled as either constant among surveys, specic to individual surveys, or inuenced by one of eleven covariates discussed below. Aer model selection analysis, we determined the number of surveys needed to have a 5% chance of not detecting the species at sites where they are present, based on estimated probabilities of detection. For Chinquipellobunus madlae, we found that detectability varied with each survey. erefore, we conducted a parametric bootstrapping simulation obtaining 1000 pseudo samples (Manly A NLY 1997). We used the formula s ip11 where p is the detection probability on survey i and s is the number of surveys (Jackson A CKSON et al., 2006). For Batrisodes uncicornis and RR hadine exilis, whose detectability was constant across surveys, the calculations were based on the simpler formula sp 11 where p is the detection probability and s is the number of surveys performed (formula 6.1 in Mac A C Kenzie ENZIE et al., 2006). Simulations for each dierent number of surveys (2, 4, 6, etc.) were performed using the statistical soware R, and consisted of 1,000 bootstrapped samples produced with a parametric and not a nonparametric bootstrapping algorithm. en for each dierent number of surveys, the mean probability of failing to detect the species was calculated. For Chinquipellobunus madlae, the varying values of p allowed us to create 95% condence intervals (Fig. 1). 2.3 Covariates Detection probabilities were modeled as either constant among surveys, specic to individual surveys, or inuenced by one of eleven covariates. Of these eleven covariates, four were unique to each cave site and seven were unique to each sample event. ey were chosen based on personal observation, interviews with local cave biologists (James Figure 1: Simulations using the survey specic detection probabilties measured for Chinquipellobunus madlae show that more surveys decrease the probability that this species will not be detected at sites where they are present. Upper and lower 95% condence intervals are shown as dashed lines. ese ndings suggest that 10-12 surveys are needed to be 95% condent that C. madlae are absent om a surveyed site.

PAGE 152

15th International Congress of Speleology Rare Species 757 2009 ICS Proceedings Reddell, Peter Sprouse), USFWS recommendations (2006), and other research (Schneider C HNEIDER and AND CulverULVER 2004). Site covariates included cave length, cave depth, size of oor search area, and size of wall search area. Seven sample covariates changed with each event and included four continuous variables: search time, in-cave temperature, in-cave relative humidity and surface air temperature. e remainder corresponded with USFWS survey recommendations and consisted of a yes/no determination for falling within the recommended surface temperature range, recommended sampling season, and a recent rain event.2.4 SpeciesBatrisodes uncicornis is an eyed troglophilic (not restricted to caves, but can spend entire life cycle in a cave) pselaphid beetle (Fig. 2) that occurs in caves throughout central Texas. is species is not endangered, but it is closely related to endangered TT e xamaurops reddelli and Batrisodes texanus. It is known to occur in 9 caves containing 21 zones that were sampled 11 times. Chinquipellobunus madlae is a troglobitic (restricted to caves) harvestman (Fig. 3) that occurs in caves throughout central Texas. is species is not endangered, but it is related to endangered TT e xella cokendolpheri, TT exella reyesi, and T T e xella reddelli harvestmen. Chinquipellobunus madlae is known to occur in 22 caves containing 61 zones that were sampled 11 times. R R h adine exilis is a federally listed carabid beetle (Fig. 4) restricted to Bexar County, Texas. Survey results were used from 23 caves subdivided into 65 zones with 11 sample events.3. Results e assumption of closure was met for all taxa, indicating that species do not colonize a site or become extinct from a site within the study period. Lower AIC values indicated the data for Batrisodes uncicornis were most consistent with constant detection probabilities and the data for the other two species varied by survey rather than being seasonal or constant. Aer closure was met, data from all years were used to test whether detection probabilities were either constant among surveys, specic to individual surveys, or inuenced by one of eleven covariates. Of the three species, Figure 2: Batrisodes uncicornis, a tiny (2 mm) troglophilic beetle, om B-52 Cave, Bexar County, Texas. Figure 3: Chinquipellobunus madlae, a troglobitic harvestman (2-3 cm), om Flying Buzzworm Cave, Bexar County, Texas. Figure 4: Rhadine exilis, an endangered troglobitic ground beetle (1-1.5 cm), om Banzai Mud Dauber Cave, Bexar County, Texas.

PAGE 153

Rare Species 758 2009 ICS P roceedings 15th International Congress of Speleology Chinquipellobunus madlae was the only dataset found to have a clear best model, which was that the detectability was dierent for every survey. Detection probabilities ranged from 0.0595 to 0.3769, with a mean of 0.2424, standard error of 0.0943 and coecient of variation of 0.3887. e proportion of sites occupied (t) was 0.85 with a standard error of 0.06. e other two species had several models that rose above the rest but were not distinct enough to choose between, and in those cases the most parsimonious of the higher ranking models, constant probability of detection, was chosen. In the case of Batrisodes uncicornis, the constant detection probability was 0.1226, the proportion of sites occupied (t) was 0.45 with a standard error of 0.16. In the case of RR hadine exilis, the constant detection probability was 0.1875 the proportion of sites occupied (t) was 0.71 with a standard error of 0.07. Parametric bootstrapping yielded the following recommended number of surveys for Batrisodes uncicornis, Chinquipellobunus madlae, and RR hadine exilis: 22, 10-12 and 14. ese are the recommended number of surveys to conduct to reduce the probability of non-detection, given presence, to 5%.4. DiscussionMany caves are surveyed to determine whether they are occupied by rare and endangered troglobites, and several researchers have examined accumulation curves and patterns of species richness in karst areas of West Virginia and Slovenia (Culver U LVER et al., 2004; Schneider CHNEIDER and AND Culver ULVER 2004). ese studies focused on determining the number of cave species in a region and how many caves would need to be sampled to obtain an accurate estimate of species richness for the area rather than for a single cave. Results included a lack of asymptotes or plateaus in species accumulation curves, with one explanation being that repeated visits are oen necessary to collect all of the species found in a single cave (Schneider C HNEIDER and AND Culver ULVER 2004). Culver U LVER et al. (2004) give an example of a new taxon being found aer 6 visits, and two examples of new taxa being found aer over a hundred visits to a cave. In the instance of Lakeline Cave, Williamson County, Texas, at least 45 biological surveys have been performed by experienced cave biologists of the entire cave (approximately 23 m long), and on approximately the 40th visit a new species of troglobitic pseudoscorpion was found. Clearly some species are commonly not available or not detected, however prior to this work no researchers have attempted to calculate detection probabilities or estimate the number of visits required to a single cave to nd a troglobite. e detection probabilities calculated herein suggest that modications should be made to recommended survey techniques to condently estimate occupancy. Even in taxa that are large and easy to see ( Chinquipellobunus madlae Figure 3), in our analysis of caves where they are known to occur, the proportion of sites occupied was 0.85 and the detection probability averaged only 0.24. With 10-12 visits recommended to condently determining absence for this taxon, many more should be required of smaller, slower moving and more inconspicuous troglobites such as T T e xella species. Suggestions about appropriate sampling conditions for cave fauna come from qualitative observations by cave biologists, and in Texas have generally included seasonal and weather conditions that are thought to make the interior of these shallow caves more favorable for nding cave species. In our lengthy list of possible covariates, however, none clearly demonstrated an association with detectability of these species. For one of three taxa, detectability denitively varied with each survey event, indierent of all the covariates tested. For the other two taxa, the distinction was less clear and confounded by a small number of detections in the matrix of observation events. Patterns of species detections appear irregular, and more work needs to be done both on the environment and experimentally on the species to determine if the environmental variables we measure during these studies are actually related to detection probability. For example, dataloggers in caves can demonstrate if seasonal, temperature, or rainfall variation on the surface is reected in the cave environment at dierent endangered species localities. e other critical component is to use experimental manipulation of the taxa to determine if they respond to the magnitude of changes that actually occur within the cave. When the species analyzed herein are not available, the most obvious hypothesis is that they retreat into inaccessible cracks that are connected to the cave. ese spaces, called mesocaverns (or sometimes called epikarst, voids, or unenterable caves), should then be considered a priority for conservation. Presently management focuses on caves and surface habitat immediately surrounding caves. Cave entrances and the surrounding surface area are important because they provide a nutrient source for cave ecosystems, but this suggests that a greater area of karst that is connected to caves may be where the species oen reside. Knapp N APP and AND Fong ONG (1999) also concluded that the stygobites they studied occur primarily in a larger area of epikarst that is connected to the cave pools they could access, and considered the pools a small window into that habitat. AcknowledgementsJackie Schlatter (Chief of Natural and Cultural Resources, Ft. Sam Houston, Texas) provided permission to use the dataset

PAGE 154

15th International Congress of Speleology Rare Species 759 2009 ICS Proceedings from Camp Bullis Military Reservation and coordination of the funding to collect the data. George Veni (George Veni & Associates) was the principal investigator and coordinator for all the research at Camp Bullis during the time these data were collected and always encouraged analyses to further the knowledge of these little studied systems. Cyndee Watson (USFWS) was involved in the project conception and provided many useful suggestions throughout the duration of the project. Rob Myers assisted with review and revisions of a dra version. James Reddell (Texas Memorial Museum) and Peter Sprouse (Zara Environmental LLC) provided suggestions of covariates to test. Kellie Cowan (Zara Environmental LLC) spent countless hours extracting and arranging the data for input into PRESENCE. Bill Elliott started monitoring cave species in central Texas and provided the framework for future research. Many people were involved in the in-cave data collection, including USFWS karst invertebrate permit holders Andy Gluesenkamp, Rob Myers, James Reddell, Marcelino Reyes, Peter Sprouse, Bev Shade and George Veni, and their observations and dedication created the large dataset used in this project and that will certainly be used for years to come. Funding for the analysis and this report was provided by Texas Parks and Wildlife through the Endangered Species Act Section 6 grant program of the U.S. Fish and Wildlife Service. In particular, Craig Farquhar (TPWD) did an outstanding job taking care of the administration of these funds. Matching funds and general support of the author were provided by Zara Environmental LLC and Texas State University.ReferencesBAILEY, L.L., T.R. SIMONS, and K.H. POLLOCK (2004) Estimating site occupancy and species detection probability parameters for terrestrial salamanders. Ec ological AA pplications 14(3), 692. Brunham, K.P., and D.R. Anderson (2002) Model selection and multi-model inference: a practical informationtheoretic approach. Springer-Verlag, New York, New York, USA. CULVER, D.C., M.C. CHRISTMAN, B. SKET, and P. TRONTELJ (2004) Sampling adequacy in an extreme environment: species richness patterns in Slovenian caves. Biodiversity and Conservation 13 12091229. Elliott, W.R. (1994) Community ecology of three caves in Williamson County, Texas: a three year summary. Annual report submitted to Simon Development Co., Inc., U.S. Fish and Wildlife Service, and Texas Parks and Wildlife Department, 46 pp. JACKSON, J.T., F.W. WECKERLY, T.M. SWANNACK, M.R.J. FORSTNER (2006) Inferring absence of Houston Toads given imperfect detection probabilities. Journal of Wildlife Management 70(5) 14611463. KNAPP, S.M., and D.W. FONG (1999) Estimates of Population Size of Stygobromus emarginatus (Amphipoda: Crangonyctidae) in a Headwater Stream in Organ Cave, West Virginia. Journal of Cave and Karst Studies 61(1) 3. MACKENZIE, D.I., J.D. NICHOLS, G.B. LACKMAN, S. DROEGE, J.A. ROYLE, and C.A. LANGTIMM (2002) Estimating site occupancy rates when detection probabilities are less than one. Ecology 83(8) 2248-2255. MacKenzie, D.I., J.D. Nichols, J.A. Rolye, K.H. Pollock, L.L. Bailey, and J.E. Hines (2006) OO ccupancy Estimation and Modeling: inferring patterns and dynamics of species occurrence. Elsevier Inc., Amsterdam. Manly, B.F.J. (1997) RR and omization, bootstrap and monte carlo methods in biology. 2nd ed., Chapman & Hall, New York. SCHNEIDER, K. and D.C. CULVER (2004) Estimating subterranean species richness using intensive sampling and rarefaction curves in a high density cave region in West Virginia. Journal of Cave and Karst Studies 66(2) 39. U.S. Fish and Wildlife Service (1994) Recovery Plan for Endangered Karst Invertebrates in Travis and Williamson Counties, Texas. Albuquerque, New Mexico. 154 pp. U.S. Fish and Wildlife Service (2006) USFWS Section 10(a) (1) (A) Scientic Permit Requirements for Conducting Presence/Absence Surveys for Endangered Karst Invertebrates in Central Texas. Fish and Wildlife Service, 10711 Burnet Road, Suite 200, Austin, Texas 78758. March 8, 2006. 21 pp. Veni, George, and Associates (2006) Hydrogeological, Biological, Archeological, and Paleontological karst investigations, Camp Bullis, Texas, 1993-2006. Report for Natural and Cultural Resources, Ft. Sam Houston, by George Veni and Associates, San Antonio, Texas, 2,421 pp.

PAGE 155

Rare Species 760 2009 ICS P roceedings 15th International Congress of Speleology BATS: GOING . GOING . GONE WITH THE WINDTT HO O MA A S H. KUN N Z1 AND AND ED D WARD ARD R R ARN ARN ETT TT2 1Center for Ecology and Conservation Biology, DDepartment of Biology, Boston University, Boston, MA A 02215, USA A2Bat Conservation International, 500 NN orth Capital of TT exas Highway, AA ustin, TT X 78746, USA A Abstract roughout much of the world, wind energy has become an increasingly important sector of the renewable energy industry. Environmental benets of wind energy accrue from the replacement of energy generated by fossil and nuclear fuels, thus reducing some adverse environmental eects from these industries. However, development of the utility-scale wind energy industry has led to some unexpected environmental costs. In particular, large numbers of bat and bird fatalities have been reported at utilityscale wind energy facilities in both forested and agricultural landscapes. Less is known about adverse eects of wind energy developments in arid regions and at oshore sites where wind energy capacity is oen high. As the world demand for renewable energy increases, we are witnessing a near exponential growth of wind energy facilities. Given the current unregulated development of this industry in many parts of the worldincluding the United Statescontinued large numbers of bat and bird fatalities can be expected. To date, most of the bat fatalities have been reported from onshore localities in Europe and North America, where several local eorts are being made to monitor operational wind-energy facilities for turbine-related fatalities. Unfortunately, few if any assessments of bat or bird fatalities are being made in most other regions of the world. e wind energy industry is developing at such a rapid rate that the projected cumulative impacts of fatalities are staggering, aecting both cave-dwelling and tree-roosting species. Recent research has shown that adverse impacts of existing wind energy facilities on bats can be mitigated by operationally feathering turbine rotors at low wind speeds. e term feathering means that the turbine blades are pitched parallel to the wind and thus hardly move. is can be done at low wind speeds when bats are more oen killed, and during high seasonal periods of bat activity (e.g., during fall migration). Preliminary research suggests that this type of mitigation can reduce fatalities by up to 92%. us, the wind energy industryto retain its green imageshould be required to implement such measures during designated times of the year and at low wind speeds to reduce adverse impacts on bat populations. Such considerations should be part of the permitting process and required for the authorization of government-funded tax credits, Additional research is needed to identify geographic regions that are more appropriate for placement of new wind energy facilities, preferably where little or no bat activity has been documented. Finally, research is needed to assess the potential adverse impacts of communityscale wind energy developments on bats and birds, where such development could be the next cat in the backyard.

PAGE 156

15th International Congress of Speleology Rare Species 761 2009 ICS Proceedings Cave AVE animals ANIMALS in IN show SHOW caves CAVES in IN Slovenia LOVENIAANDRANDR EJ MIHEVC Karst RResearch Institute ZR R C SA A ZU, TT ito trg 2, 6230 Postojna, Sloenia Abstract Since the discovery of the cave salamander Proteus anguinus in Postojnska jama cave in 1796 and the rst cave beetle Leptodirus hohenwarti in 1831, cave animals were important attraction for visitors in show caves in Slovenia. Proteus also became one of the symbols of the Postojnska jama. Animals were later, specially the proteus, on the display to visitors. At rst they were displayed in pools or aquarium in the cave. Now a part of the cave, former speleobiological station with special entrance is arranged for visitors displaying proteuses and several other smaller cave animals like beetles and crustaceans in special aquariums or terrariums. Visitors can see the proteus also in two other managed caves in Slovenia. Besides them black subspecies, Proteus anguinus parkelj can be seen in the natural environment in one karst spring. In Postojnska jama cave tourism started in 1818 and there is less cave animals in some managed passages. In other parts of the cave impact of tourism is less pronounced. ere is no impact of tourism on the proteus population which lives there. ere is also no impact reported on proteus in other managed caves.

PAGE 157

Rare Species 762 2009 ICS P roceedings 15th International Congress of Speleology ON RARITY AND THE VULNERABILITY OF SUBTERRANEAN FAUNAOANAOANA T T EODORA ODORA MO O LDO DO VAN AN Emil RR acoi Institute of Speleology, Clinicilor 5, 400006 Cluj-N N apoca, RRomania, oanamol@hasdeu.ubbcluj.ro Abstract Rare species have small numbers of individuals or small ranges and are thus considered to be the most threatened. Many subterranean species can be considered rare, but little is known about how threatened they actually are. According to the IUCN there are three categories of threatened species: vulnerable, endangered, and critically endangered. We oen declare subterranean species as vulnerable, if not endangered, but concise arguments for this axiom are rarely presented. Our research over the past decade on terrestrial and aquatic invertebrate fauna in caves or other underground habitats, in both show caves and protected caves will be presented as either arguments or counter-arguments for the vulnerability of subterranean fauna. We also discuss the dierent human direct or indirect impacts and propose some management rules for protection of subterranean fauna.

PAGE 158

15th International Congress of Speleology Rare Species 763 2009 ICS Proceedings CAVE MICROBIAL COMMUNITIES: IS PROTECTION NECESSARY AND POSSIBLE?DD IANA ANA E. NORT NORT HUPD Department of Biology, MSC03 2020, University of NN ew Mexico, AA lbuquerque, NN M 87131, USA A Microbial communities in caves vary from the striking microbial mats observed in many lava tubes worldwide, to occasional colonies on the wall, to invisible biolms on rock walls and ceilings of caves, to microbial end products, such as iron oxides. e investigations of the last decade, using cultureindependent techniques in which we extract DNA from environmental samples and sequence clones to identify organisms present, have revealed a wealth of microbial species never described. Some of those organisms have been implicated in the dissolution and precipitation of cave rock and secondary minerals. Some may be critical in cycling nitrogen and carbon from the surface to the subsurface. Others have been shown to have the ability to kill cancer cells or produce antibiotics that might replace some of those antibiotics that are failing due to antibiotic resistance. us, we know that microorganisms found in caves are valuable, both for their intrinsic nature, and because they may prove useful to humans. Are they valuable enough to us to consider protecting them from we humans that explore their native cave habitat? To what degree do we, as cave visitors, impact these communities when we visit caves (Northup and Welbourn, 1995)? e degree to which we impact cave microbial communities depends on the nature of the cave. Mammoth Cave in Kentucky, USA, and other caves like it, have rivers or streams running through major portions of the cave. Such caves are likely impacted by human visitation much less than arid-land caves, such as Lechuguilla Cave in New Mexico, USA. ese drier and warmer caves contain signicant microbial communities that can fall prey to a variety of impacts. When we explore caves, we leave behind pieces of ourselves: skin cells, bacteria and fungi from our hair and skin, hair, and occasionally worse things such as feces, urine, or mud and dirt from other caves, which carry their own microbial passengers. One of the major impacts that we can have on low-nutrient caves is the enrichment of the organic carbon present in those caves. We also compact any soil present and might leave other pollutants that aect microbial communities. How can we protect these microbial communities from such threats? Several strategies can lower the impact that we have on cave microbial communities. Cleaning our gear, clothes, boots, and bodies between cave trips can limit the amount of cross-contamination that occurs among caves and lower the amount of organic carbon enrichment that occurs. Establish trails and camp areas (if camping is an issue) to conne human impact to a limited area of the cave. Encourage everyone to eat over bags to catch crumbs. A much more controversial strategy is the establishment of microbial preserves to preserve areas of unusually promising microbial potential by limiting the amount of human impact on the area. Cave microbial communities can represent an extremely valuable resource that is worth protecting by modifying our behavior in visiting caves. e payo may be an antibiotic that saves your life someday!1. IntroductionMicroorganisms in caves range from completely invisible to highly colorful microbial mats (Fig. 1) that line the walls of lava tubes to microbial waste products, such as iron oxides. It is hard to appreciate and value something that you cannot even see, but there are many compelling reasons to protect and conserve cave microorganisms and their habitat. In the last two decades we have seen a major increase in research into the role that microorganisms play in the dissolution of bedrock and other surfaces and the precipitation of secondary mineral deposits (Barton A RTON and AND Northup ORTHUP 2007; Northup ORTHUP and AND Lavoie AVOIE 2001). New discoveries of sulfur oxidizing bacteria in caves are revealing microbial roles in cycling sulfur in caves and enlargement of cave passages through sulfuric acid

PAGE 159

Rare Species 764 2009 ICS P roceedings 15th International Congress of Speleology dissolution, as well as a possible role in sulfuric acid driven speleogenesis (Engel N GEL et al., 2004; Hose O SE et al., 2000). Spilde P ILDE et al. (2005) have shown a microbial role in the production of ferromanganese deposits in arid-land caves. ese and other studies are showing key geomicrobiological roles for microorganisms in caves. Although much remains to be learned about the microbial role in energy transfer and elemental cycling, some evidence suggests that microorganisms facilitate the transfer of energy between cave life and organic carbon and serve as food for the cave life (Simon IMON et al., 2007). Perhaps most exciting is the amount of novel biodiversity that culture-independent molecular studies are revealing in caves (e.g. Barton A RTON et al., 2004; Gonzalez O NZALEZ et al., 2006; Northup O RTHUP et al., 2003). Some of these novel (and not so novel) species may produce chemical compounds that are very useful to humans, such as new antibiotics to replace those to which bacteria are now resistant (Dapkevicius, Terrazas and Northup, unpub. data). e geomicrobiological studies and those of novel microbial biodiversity also serve to aid our understanding of how to detect life on other planets, such as Mars, where life is likely to shelter from harsh surface conditions in the subsurface (Boston O STON et al., 2001). us, our research is emphasizing the critical nature of cave microbial communities and suggests that their conservation is important.2. reats to Microbial PopulationsCave microorganisms are susceptible to a variety of threats, including human visitation, soil compaction, pollutant spills, cave restoration, and organic carbon enrichment. Whether microorganisms reside in arid-land caves, or those in areas with more rainfall, aects the degree to which these threats are an issue for microbial populations. Rivers and streams running through caves can carry away pollutants and dampen the eects of various threats; however, rivers and streams can also be the vehicle for introducing pollutants. Arid-land and caves with little inow of moisture, in particular, are much more subject to the eects of organic carbon enrichment and other impacts that result from human visitation of caves. When we visit caves, we shed tens of thousands of skin cells, many of which are life ras for our own microbial inhabitants, as well as hair and bers and mud from our clothing (Fig. 2). If we are sick and vomit in the cave, we greatly enrich organic carbon in the habitat. Longer cave trips may bring the issues of urine and feces deposition (Fig. 3). While cricket and beetle feces are a natural part of the ecosystem, human feces are not. ere is the matter of scale and the fact that human feces are almost 50% microorganisms. Urine can lead to the buildup of harmful compounds that change the microbial ecosystem (Lavoie AVOIE 1995). As we walk through areas of the cave with soil or detrital material, we cause compaction of the soil, which decreases the available oxygen. Some visitors draw their names and dates in microbial mats (Fig. 4). Our studies (Lavoie and Northup, unpub. data.) suggest that human associated bacteria (e.g. Staphylococcus aureus) and fungi are preferentially found in areas with more human impact. If the cave is given time to rest (i.e. no human visitation) and we limit the amount of organic carbon buildup, these exotic populations generally die o. However, some exotic populations can persist and damage cultural artworks, such as those found in the caves of France and Spain (Jurado U RADO et al., 2008). Figure 1: White, yellow, gold, and, pink microbial mats commonly line the walls and ceilings of moist lava tubes worldwide. Photos copyright 2008 Kenneth Ingham.

PAGE 160

15th International Congress of Speleology Rare Species 765 2009 ICS Proceedings Also, our well intentioned eorts to restore and clean caves can lead to many problems for microbial communities as detailed in Boston O STON et al. (2005). By trying to protect pristine pools in Lechuguilla Cave, we unintentionally enriched the amount of organic carbon in the pools from plasticizers in the tubing used to obtain drinking water. is led to a population explosion of a native bacterial population that appears to have then supported introduced E. coli in the pool (Hunter U NTER et al., 2004). rough a variety of ways, we provide challenges to subterranean microbial populations. Where native microbial populations reside in oligotrophic habitats within caves, we will see the most profound eects. Oligotrophic microorganisms dont simply get fatter when you feed them more; they oen die o, allowing more surface-adapted microorganisms to take their place (Koch O CH 1997). us, human visitation can introduce new organic matter and exotic microorganisms into caves, which may harm native microbial populations.3. Is Protection Possible During Active Exploration and Scientic Investigation?ere are some relatively simple things that can be done to protect microbial communities in caves, but the success of these recommendations rests in the acceptance of the value of these microbial populations by cavers, scientists, and other visitors to caves. It is hard to think about and protect things that you cannot even see. If you believe that cave microorganisms are a key component of the cave ecosystem or that an eective treatment for cancer or a new antibiotic that could save your life someday could come from a cave microorganism, then you are likely to be willing to go the extra mile for the microbes. Getting people to this point will take more research into what harms and what protects microbial communities in caves and using that information to educate cave visitors about the problems and solutions. One of the payos is that educational programs about cave microorganisms oen excite and engage cave visitorsthe microbes have wonderful stories to tell.4. RecommendationsTo know how to protect something, you need to understand it. Our knowledge of cave microbial communities is rudimentary, limiting our ability to know precisely what eorts will protect microbial communities in caves. Several laboratories around the world are conducting outstanding culture-independent molecular studies of cave microbial communities to identify novel biodiversity, while others are culturing cave microorganisms to shed light on their Figure 2: Humans leave detritus when they visit caves, such as hair, skin cells, food, and mud om other caves. Illustration by N. Robin Wilson, copyright 1993. Figure 3: Long caving trips raise the issue of human feces disposal and possible impact as feces olatilize as they leave the body. Illustration by N. Robin Wilson, copyright 1993. Figure 4: Cave visitors oen draw their names in microbial mats that line the walls of lava tubes. Photo copyright 2008 Kenneth Ingham.

PAGE 161

Rare Species 766 2009 ICS P roceedings 15th International Congress of Speleology physiology and biochemistry, but we need more scientists involved. e rst molecular study of microbial diversity was published in 1997, and while many others have followed it, much remains to be learned. Microbial inventories across gradients of depth, nutrient richness, distance from entrances, human impact, etc. are needed to compile a more complete picture of cave microbial communities. Many interesting ecological and evolutionary questions about cave microorganisms await researchers (e.g., Snider N IDER et al., 2009). us, research and inventory are key steps in the journey to protect cave microorganisms. e following recommendations are based on our preliminary investigations and insights, but their eectiveness remains to be tested. As cavers and visitors to caves, we can take several actions to conserve microbial habitat and microorganisms: Establish trails for movement through the cave. When you establish trails, use inert markers that do not enrich organic carbon in the cave and do not degrade. If there are no marked trails, always walk where the elephant tracks are. For caves in which camping is necessary for exploration, establish camps to concentrate human impact. Eat over bags to catch all crumbs. Whats a crumb to you is a supermarket to a microorganism. Clean your clothes and boots between cave trips to prevent cross contamination between caves. Brush your hair to remove loose hairs before going caving. Find ways around pristine pools and avoid dipping anything, including yourself, in the pool. Establish a clean pitcher for obtaining water. Educate new cavers in the ways to preserve and protect microbial communities. Scientists, cavers, and cave managers who nd unusual deposits that may be microbial should consider establishing a microbial preserve to allow investigation before visitation occurs to any extent. If you see something really intriguing, send a photo to one of the microbiologists around the world who studies these communities in caves. Scientists oen study a few areas very intensively and may miss key discoveries. Cavers and scientists should collaborate on microbial discoveries for mutual benet. Scientists can excite cavers and visitors by providing engaging information about their ndings through public talks, articles, and other media that bring the science to the public.5. ConclusionsOur knowledge of microbial diversity in caves is growing rapidly and revealing a wonderland of microorganisms (Fig. 5) that participate in precipitation and dissolution of cave mineral deposits that have roles in nutrient cycling within the cave ecosystem, that may produce chemical substances of great use to humans, and that serve as an analog for possible life on other planets. ese important communities are, however, threatened by some of our actions when we visit or live and work above caves. By being conscious of the ways in which we may enrich organic carbon in caves, we can do much to protect microbial habitats and microorganisms in caves. Are cave microorganisms threatened? In 1997, Jim Staley wrote the following concerning microorganisms in general: Our knowledge of microbial diversity, particularly bacterial diversity, is so meager that we do not yet know if and when most species are threatened. is is particularly true of cave microorganisms and enhanced eorts to study and understand cave microbial communities are essential to our being able to truly answer the question of whether these populations are threatened.AcknowledgementsMany cavers over the years have provided immeasurable help in carrying out the various research projects that led to observations that formed the basis for the ideas contained in this manuscript, and in providing leads to new microbial habitats. ey include, but are not limited to: Kenneth Ingham for all his great microbe photography; Figure 5: Cave microorganisms represent a wonderland of organisms as seen in these scanning electron micrographs om caves in New Mexico, Arizona, Mexico, and the Cape Verde Islands. Photomicrographs courtesy of Michael N. Spilde.

PAGE 162

15th International Congress of Speleology Rare Species 767 2009 ICS Proceedings Val Hildreth-Werker and Jim Werker for photography, engineering, and lots of great research; Andi Hunter, for her invaluable research into the contamination of pools in Lechuguilla Cave; and Penny Boston and Mike Spilde, with whom I have had many stimulating conversations about microbes. e Charles A. and Anne Morrow Lindbergh Foundation, Mammoth Cave National Park, and T & E, Inc. provided nancial support for the human impact studies that were carried out in collaboration with Kathy Lavoie who contributed substantially to the ideas on human impact. anks go to Leslie Melim and Kenneth Ingham for insightful comments on the manuscript.ReferencesBoston OSTON P.J., Northup ORTHUP D.E., and K.H. Lavoie A VOIE (2005) Protecting microbial habitats: Preserving the unseen. In C ave Conservation and RRestoration, Hildreth-Werker V., J.C. Werker (Eds), National Speleological Society, Huntsville (AL), p. 61. Boston O STON P.J., Spilde PILDE M.N., Northup ORTHUP D.E., Melim ELIM L.A., Soroka OROKA D.A., Kleina LEINA L.G., Lavoie AVOIE K.H., Hose OSE L.D., Mallory ALLORY L.M., Dahm AHM C.N., Crossey ROSSEY L.J., and R.T. Scheble C HEBLE (2001) Cave biosignature suites: Microbes, minerals and Mars. AA s trobiology Journal 1, 25. Engel N GEL A.S., Stern TERN L.A., and P.C. Bennett ENNETT (2004) Microbial contributions to cave formation: new insights into sulfuric acid speleogenesis. Geology 32 369372. Gonzalez ONZALEZ J.M., Portilo ORTILO M.C., and C. Saiz A IZ Jimenez IMENEZ (2006) Metabolically active Crenarchaeota in Altamira Cave. N N a turwissenschaen 93, 42. Hose OSE L.D., Palmer ALMER A.N., Palmer ALMER M.V., Northup ORTHUP D.E., Boston OSTON P.J. and H.R. Duchene U CHENE (2000) Microbiology and geochemistry in a hydrogen sulphide-rich karst environment. Chemical Geology 169, 399. Hunter U NTER A.J., Northup ORTHUP D.E., Dahm AHM C.N., and P.J. Boston OSTON (2004) Persistent coliform contamination in Lechuguilla Cave pools. Journal of Cave and Karst Studies 66, 102. Jurado U RADO V., Sanchez ANCHEZ Moral ORAL S., and C. Saiz AIZ Jimenez IMENEZ (2008) Entomogenous fungi and the conservation of the cultural heritage: A review. International Biodeterioration & Biodegradation. 62, 325. Koch O CH A.L. (1997) Microbial physiology and ecology of slow growth. Microbiology and Molecular Biology R R eviews 61, 305. Lavoie A VOIE K.H. (1995) e eects of urine deposition on microbes in cave soils. In Proceedings of the 1993 N N a tional Cave Management Symposium: held in Carlsbad, NN ew Mexico, OOctober 27-30, 1993, Pate, D.L. (Ed.), National Cave Management Symposium Steering Committee, Huntsville (AL), p. 302. Northup O RTHUP D.E., and K.H. Lavoie AVOIE (2001) Geomicrobiology of caves: A review. Geomicrobiology Journal 18, 199. Northup O RTHUP D.E., and W.C. Welbourn ELBOURN (1995) Conservation of invertebrates and microorganisms in the cave environment. In Proceedings of the 1993 N N a tional Cave Management Symposium: held in Carlsbad, NN ew Mexico, OOctober 27-30, 1993, Pate, D.L. (Ed.), National Cave Management Symposium Steering Committee, Huntsville (AL), p. 292. Simon I MON K.S., Pipan IPAN T., and D.C. Culver ULVER (2007) Conceptual model of the ow and distribution of organic carbon in caves. Journal of Cave and Karst Studies 69, 279. Snider N IDER J.R., Goin OIN C., Miller ILLER R.V., Boston OSTON P.J., and D.E. Northup ORTHUP (2009) Ultraviolet radiation sensitivity in cave bacteria: Evidence of adaptation to the subsurface? International Journal of Speleology 38 112. Spilde P ILDE M.N., Northup ORTHUP D.E., Boston OSTON P.J., Schelble CHELBLE R.T., Dano ANO K.E., Crossey ROSSEY L.J., and C.N. Dahm AHM (2005) Geomicrobiology of cave ferromanganese deposits: A eld and laboratory investigation. Geomicrobiology Journal B, 99. Staley T ALEY J.T. (1997) Biodiversity: Are microbial species threatened? C urrent OO pinion in Biotechnology 8, 340.

PAGE 163

Rare Species 768 2009 ICS P roceedings 15th International Congress of Speleology MANAGEMENT OF ENDANGERED KARST INVERTEBRATES ON THE BALCONES CANYONLANDS PRESERVE, AUSTIN, Te E Xas ASKAT AT HLEEN N M. O O CONNOR ONNOR1, MAR AR K SAND AND ER R S2, PA A UL FUSHILLE1 1TT ravis County NN atural RResources and Enironmental uality DD ivision 1010 Lavaca AA ustin, TT X 787462City of AA ustin Wildland Conservation DD ivision PO O Box 1088, AA ustin, TT X 78767 Abstract In 1996, the USFWS issued a 10(a)1b permit to Travis County and the City of Austin to mitigate for the loss of endangered species habitat due to urban development activities and to facilitate the recovery of eight federally listed endangered species, six of which are troglobitic invertebrates. A minimum of 30,428 acres in western Travis County, Texas will be set aside as preserve land to fulll mitigation requirements. ere are 62 karst features listed under the Balcones Canyonlands Preserve (BCP), including 35 caves containing at least one of the listed troglobitic species. Currently, Travis County Natural Resources and City of Austin Wildland Division sta manages and protects 32 BCP karst features, including a cave cluster which supports one of the most diverse, terrestrial cave-adapted assemblages in the southwestern U. S. Management practices on the BCP include conducting annual faunal surveys within caves, seasonal cricket counts at cave entrances, and red imported reant (Solenopsis inicta) control. Caves located near urbanized areas provide a unique challenge for karst managers but also provide an opportunity for public education. In the spring of 2007, the City of Austin and Travis County began conducting quarterly surveys on eight selected caves in order to examine seasonal changes in species richness and abundance. Data from three seasons (spring, fall and summer) in both 2007 and 2008 was used for analyses. e relationships between the cave fauna community and three variables (site, year of sampling, and season of sampling) were tested using Mantel tests (Mantel 1967) between two similarity matrices calculated with two indices (S8: presence/absence and S17 which also accounts for abundances). No signicant relationships were found between the fauna, year and season with both indices, but a highly signicant relationship with the sites suggests that the fauna diers from cave to cave. A partial Mantel test (Smouse et al. 1986) found a signicant relationship between TT e xella reyesi and season when controlling for the variability due to the sites. Although still preliminary, these results suggest that a given cave may not necessarily be a biological replicate of another cave despite geographical proximity, similar sampling intensity, and common geological history. Data collection in selected caves will continue seasonally for further subsequent analyses.

PAGE 164

15th International Congress of Speleology Rare Species 769 2009 ICS Proceedings Conservation ONSERVATION of OF species SPECIES on ON the THE wrong WRONG track TRACK ? Is S conservation C ONSERVATION of OF species SPECIES losing LOSING its ITS way WAY ?BORis IS Ske KE T1 and FRANci CI GAb B ROv V ek EK2 1OO ddelek za biologijo, Biotehnika fakulteta, Univerza v Ljubljani, Ljubljana, Sloenia2Intitut za raziskoanje krasa ZR R C SA A ZU, TT ito trg 2, Postojna, Sloenia Abstract To eciently protect and conserve the subterranean fauna, knowledge of its composition, ecology, and the distribution of its taxa are necessary. Detailed investigations with new methods (mainly DNA analyses) show that the number of species is much greater, the distribution areas of species are much lower, and their endemicity higher than suspected. To complete the pictures of the real biotic richness of countries, intensive re-investigations are necessary. e researchers have to overtake the rapid destruction of the nature, destruction of access points and even rarefaction or extinction of subterranean species. erefore the rst action of national agencies for environmental management, for nature conservation, departments for nature resources, and similar authorities, should be the legal protection of the subterranean environment which is in full accord also with more practical needs of the local population. It either results or even originates in protection of ground water resources. As the second, an obligation should be felt to encourage and support researchers to do the above mentioned investigations. Specialists are scarce and busy. ey require support if they are at all availableglobalization is necessary in this context. e additional training of local students and scientists, as well as registering guest scientists is useful, but it should not hamper their research. is is also in accord with the Convention on Biological Diversity (Article 7. Identication and Monitoring; Article 12. Research and Training), signed by presidents of many countries. Only as the third, a supplementary is development of protected species lists that are forbid collecting. Such lists are useful only for the commercially interesting species. Numerous cases illustrate how rapidly the access localities for subterranean fauna (including the type localities, of prime importance for science) are being destroyed. Calculations also show that collecting subterranean animals mostly can not endanger their populations. By means of fractal analysis we roughly estimated that less than 10% of the underground voids volume and less than 1 of the underground surface inhabitable by invertebrates is accessible to man. On the other hand, the whole system is accessible to pollution and other obstructions from the surface. uswith few exceptionsspecies protection policies protect animals only from researchers; beside that, it can mask the absence of an eective habitat protection measure. Similarly counter-productive is the exaggeration of provisions on access to genetic resources, triggered by a misleading interpretation of the Convention on Biological Diversity. If these obstacles to the research are not soon removed, extinction will move forward much faster than research and protection of subterranean fauna. Both problems, threats to the subterranean environment, and administrative obstacles to investigations, are common to both the developed and the developing world.

PAGE 165

Rare Species 770 2009 ICS P roceedings 15th International Congress of Speleology KARST INVERTEBRATE HABITAT AND THE ROLE OF EXCAVATIONPe E Te E R Sp P ROuse USE and Je E AN KRejc EJC A, Ph H .D. Zara Enironmental LLC, 118 W. Goforth RR d., Buda, TT X 78610 USA A Abstract Two federally listed endangered troglobite species, the ground beetle R R h adine persephone and the harvestman TT e xella reyesi, were found in excavated karst features in Travis County, Texas within 2 m of the feature entrance during species surveys for placement of a water treatment plant. ese may be the shallowest records for these cave species, with implications for the conceptual denition of habitat and practical determination of habitat and search eort. In the case of RR pe rsephone a feature was excavated to a depth of 2 m, where an articially-enlarged space 1.5 m in diameter was bounded by a bedding plane of vuggy limestone with small voids extending in all directions in un-enterable mesocaverns. Un-baited sticky traps were placed and the feature was covered with plywood and black plastic sheeting to induce darkness, retain humidity, and buer temperatures in the near surface zone, all with the goal of increasing the probability of detecting karst invertebrates that may occupy adjacent mesocavernous voids. Biologists were surprised to nd 10 RR pe rsephone on these traps. In two other features of similar depth extensive bedrock excavation revealed specimens of T T r eyes during digging, but were not caught in traps or during visual searches following the covering of features with plastic sheeting and plywood.

PAGE 166

15th International Congress of Speleology Rare Species 771 2009 ICS Proceedings KARST LANDSCAPE EVOLUTION: IMPACTS ON SPECIATION, BIOGEOGRAPHY, AND PROTECTION OF RARE AND ENDANGERED SPECIES GEORGE VENIN N a tional Cave and Karst RResearch Institute, 1400 Commerce DDrive, Carlsbad, NN ew Mexico 88220, USA A gveni@nckri.org Karst areas provide habitat to a diverse array of species, oen adapted to spending their entire lives underground. Many of these species are rare and some are endangered. e evolution and biogeography of cave-dwelling species is strongly inuenced by the evolution of karst landscapes relative to the following factors: lithology, geologic structure, burial, hydrology, and climate. e resulting karst provides potential habitat for cave-dwelling species, connectivity between populations, and restrictions and barriers to gene ow. Speciation oen results when populations become isolated. Genetic isolation can be complete, a barrier, or partial, a restriction to a species distribution. In the case of a restriction, gene ow occurs through a relatively small area and/or an area that is traversable only for relatively short periods of time. e ecological and genetic eect of species moving through restrictions thus is oen limited or diluted. Rare karst species most commonly occur where barriers and restrictions produce small habitats, and are more easily endangered where unsustainably impacted by human activities. Evaluating the origin and evolution of karst landscapes and caves also provides a valuable means of protecting and managing cave ecosystems. Key management tools include delineation and establishment of: Karst Fauna Regions, dened by hydrogeological barriers and/or restrictions to the migration of troglobites over evolutionary time, which result in speciation between regions and the creation of similar groups of troglobites within the caves of a particular region; K arst Fauna Areas locations, protected in perpetuity, known to support one or more sites of rare or endangered species and are distinct by acting as individual systems separated from other karst fauna areas by geologic and hydrologic features and/or processes that create barriers to the movement of water, contaminants, and troglobitic fauna; Rare and Endangered Karst Species Zones, which use biological and geological factors to estimate the likely boundaries of habitat for such species and areas of probable and improbable habitat; Karst Preserves, locales protected from non-natural activities to sustain the surface and subsurface components of rare karst ecosystems known to occur within certain caves. Examples are provided from caves with rare and endangered species in central Texas, U.S.A.1. Introductione origin and evolution of cave-dwelling animals is dependent on the occurrence and evolution of caves and conditions that would cause surface-dwelling creatures to retreat underground. Speciation occurs as cave habitat becomes available or attractive, and as incipient cave dwellers begin to diverge genetically from their epigean ancestors. As species become increasingly cave-adapted, their ability to survive on the surface decreases until they evolve into obligatory cave dwellers, troglobites. Speciation continues as caves and karst areas become fragmented by geologic processes and cavernicole populations become isolated, unable to cross the intervening non-cavernous areas. Several such isolated (endemic) species have been listed as endangered in the United States (U.S.). A clear understanding of these species origin, distribution, threats, and management requires an analysis of their cavernous habitat and its geologic evolution. e primary factors that determine the presence, size, shape and extent of karst caves are p redominantly soluble rock; fractures or other permeable zones within the rock; water that is chemically undersaturated with respect to the primary soluble minerals present; s ucient hydraulic gradient to promote ecient groundwater circulation; and t ime.

PAGE 167

Rare Species 772 2009 ICS P roceedings 15th International Congress of Speleology Generally, caves become larger, longer, deeper, and more interconnected when any of the above variables increase, up to the point where karst denudation results in a net decrease in cave size. ese and related variables can therefore be examined to delineate areas where caves, and related humanly inaccessible mesocavernous voids, occur to provide potential habitat to cavernicolous species. Cavernicoles generally evolve as a result of past climatic changes resulting in at least the local extinction of ancestor populations, prompting them to enter caves for shelter, moisture, or other needs. If the cave environment is suciently favorable, adaptation continues until the species become obligate cave dwellers, troglobites. Variations in geological and environmental conditions are key factors aecting the distribution, diversity, and conservation of cavernicoles and especially troglobites, the focus of this paper. Examples below are from the Austin and San Antonio areas of central Texas, which has some of the highest subterranean biodiversity in the United States (e.g. Hobbs, 2005).2. Troglobite Biogeography and Karst Landscape EvolutionSpeciation on islands, where a single ancestor may give rise to dierent species isolated on dierent islands, is similar in some respects to troglobite speciation in karst. By denition, troglobites are restricted to cave environments and thus to the geologic outcrops in which caves occur. A common surface ancestor can have several troglobite descendant species within separate karstied outcrops, assuming environmental and ecological conditions within the outcrops vary enough to promote dierent specialized adaptations. As karst landscapes evolve, such as by dissection into separate karstied areas and changing of groundwater levels, troglobite populations may become further isolated and speciated. As an example, the following discussion presents an abbreviated hydrogeologic history of karst development in the Austin and San Antonio areas (primarily summarized from Veni and Associates, 1992, 1994). Karst development in the Austin and San Antonio regions of Texas began with the deposition of the Edwards and associated formations during Cretaceous time (for brevity, only the Edwards Limestone will be discussed in this paper; similar processes and histories occurred for the other associated units). e rst episode of karstication occurred during the late Early Cretaceous when the San Marcos Platform was uplied and subaerially exposed. By the Late Cretaceous, sea levels rose to bury the Edwards under a thick sequence of carbonate and ne-grained clastic sediments. During the very Late Cretaceous or Early Tertiary, the Edwards Plateau and Edwards Limestone were lied above sea level. Balcones faulting in the Early Miocene accelerated stream incision. By the Middle Miocene, enough of the limestone was exposed to create a sucient hydraulic gradient to initiate groundwater ow and conduit development, which accelerated with continued downcutting and establishment of springs at lower elevations. Erosion of the southern and eastern margins of the Edwards Plateau exposed the Edwards Limestone in bands, isolated from the limestone on the plateau in downthrown blocks of the Balcones Fault Zone. e absolute ages of their caves have not been precisely determined but some have been roughly estimated based on stream incision rates. Caves northwest of San Antonio are estimated as old as 3.58 to 6 Ma, while some in the Jollyville Plateau of the Austin area are potentially as old as 12.5 Ma. Troglobites, and/or their ancestral species, likely began to occupy caves in the sequence that caves developed and became habitable. White (2006) described how relay ramps (sloping fault blocks) sequentially exposed limestone sections to karst development, determining the timing of occupation. Continued stream incision has divided some of the fault blocks into segments connected by only narrow sections of cavernous limestone, further restricting the distribution of their troglobites. Fine-scale distribution of species through the karst was determined by localized zones of preferential cave development and the creation of suitable conditions for troglobites. Veni (2005) identied distinct preferential modes of conduit development within the eight stratigraphic members of the Edwards Limestone and used them to evaluate troglobite habitat and management strategies. Such interpretation is crucial to predicting the extent and locations of mesocaverns, small, humanly impassable, solutionally enlarged voids that provide potential habitat for cavedwelling species in the areas between caves. Mesocaverns include solutionally widened bedding planes and fractures, anastomosed bedding planes and fractures, honeycomb solution zones, non-cemented collapse or faultbrecciated areas, porous cave sediments, and caves that have been near-completely lled with sediment. Mesocaverns may not hydrologically connect certain caves, but could provide avenues of movement between those caves for troglobites. e minimum width of mesocaverns for a signicant troglobite fauna is probably 5-10 mm; this width corresponds to the threshold of turbulent groundwater ow that could carry particles of organic nutrients to the species. Although some species can traverse smaller openings, the lack of food probably restricts their migration. is hypothesis is supported by the

PAGE 168

15th International Congress of Speleology Rare Species 773 2009 ICS Proceedings absence of troglobites in the Georgetown Limestone, which overlies the Edwards Limestone and is reported with fracture and bedding plane widths of generally <1 mm (Collins, 1989) and studies in Europe which show cave fauna generally inhabit voids greater >1 mm in width (Juberthei and Delay, 1981). 3. Troglobite Biogeography and ConservationIn 1988, ve invertebrate troglobites in the Austin, Texas, area were federally listed as endangered by the U.S. Fish and Wildlife Service (USFWS) and another seven troglobites were listed as endangered in 2000 in the San Antonio area, about 100 km to the southeast (USFWS, 1988, 2000). Listing was in response to the threats and adverse impacts posed by urban growth into the species geographically small habitat. ose threats include, but are not limited to, the destruction and adverse modication of habitat, degradation of water quality, introduction of non-native species, especially predatory red imported re ants ( Solenopsis inicta ), and the inadequacy of existing regulations to protect the species. To establish eective protection and management strategies for the listed species, USFWS funded studies and programs (e.g. Veni and Associates, 1992) that biogeographically classied the species into four categories: Karst Fauna Regions, Karst Fauna Areas, Karst Zones, and karst preserves.3.1 Karst fauna regions (KFR)KFRs are the largest biogeographic unit. eir conservation purpose is to identify distinctive biographic regions, which is crucial to understanding and maintaining genetic diversity in the ecosystems that support rare and endangered species. ey are dened by hydrogeological barriers and/or restrictions to the migration of troglobites over evolutionary time, which result in speciation between regions and the creation of similar groups of troglobites within the karst of a particular KFR. Geologic barriers are stratigraphic, structural, or hydrologic. e primary stratigraphic barrier is the simple lack of cavernous rock, but others include impermeable layers within an otherwise cavernous sequence. Structural barriers are usually coupled with stratigraphic barriers through fault juxtaposition of cavernous and noncavernous units. Hydrologic barriers vary according to the needs of the species in question; terrestrial species have a downward limit at the water table, which serves as the upper limit for aquatic species. Conditions that decrease the input of moisture or nutrients into a cave or mesocavernous voids beyond the organisms ability to survive are also barriers. KFRs are initially developed a priori, by dening the geologic features and conditions that may isolate or restrict species, and comparing those regions to the distribution of advanced troglobites. Tentative KFRs that do not contain a distinctive troglobite fauna are combined with one or more adjacent KFRs that possess the same suite of species to create biologically distinctive KFRs (e.g. combining of the initially proposed McNeil and Round Rock KFRs in the Austin area; Veni and Martinez, 2006). Distinctiveness does not mean all troglobites are endemic to each KFR. Recently adapted troglobite species are widely distributed and less likely found in only one KFR, while advanced troglobites are more likely to be suciently speciated and restricted to certain regions. Additionally, KFRs which are not completely separated geologically, or only recently separated, will share species due to insucient time for speciation or sucient gene ow through the restricted but connecting area. KFR barriers and restrictions may not necessarily be from absent or narrow cavernous outcrops but locations where the outcrops extend under valleys with owing streams and are thus below the water table. In such areas, restrictions occur where the water table periodically drops suciently below the stream bed and for suciently long periods to allow potential occupation of caves and mesocavernous voids by troglobites from each side of the valley (this paper only addresses terrestrial troglobites, not aquatic troglobites, stygobites, although many of the same principles can be applied to stygobite biogeography and management). e best example occurs in the Austin area. Ten troglobites were assessed in KFRs south side of the Colorado River and 28 in KFRs to north, only but two species were found common to KFRs on both sides of the river (Veni and Associates, 1992). Species distribution among KFRs is not simply controlled by hydrogeological factors. Species which spend much of their time actively searching for prey or forage tend to be more widely distributed. In the San Antonio area, the carabid beetle RR h adine exilis occurs in four of the areas six KFRs, absent only from the Alamo Heights and Culebra Anticline KFRs which are geologically separated from the others (USFWS, 2000). In contrast, Cicurina spiders, with 57 recognized troglobites in Texas alone (Paquin and Duprr, 2009), generally do not move far from their webs and are thus more likely to speciate. Considerable biological information on troglobite diversity and distribution has been collected since the KFRs of central Texas were rst proposed based on statistical analyses in the mid-1990s. A statistical reevaluation of the KFRs geological, topographical, and biological factors is needed, including results of DNA research that provides insight into the biogeography and evolution of the species. Such analyses should include the distribution of rare, nonlisted troglobites for insight into region boundaries not

PAGE 169

Rare Species 774 2009 ICS P roceedings 15th International Congress of Speleology evident by data from the listed species alone, and for use in management to potentially preclude the need to list those species (Veni and Martinez, 2006).3.2 Karst fauna areas (KFA)e USFWS dened a KFA as an area known to support one or more locations of a listed species and is distinct in that it acts as a system that is separated from other karst fauna areas by geologic and hydrologic features and/or processes that create barriers to the movement of water, contaminants, and troglobitic fauna. Karst fauna areas should be far enough apart so that if a catastrophic event (for example, contamination of the water supply, ooding, disease) were to destroy one of the areas and/or the species in it, that event would not likely destroy any other area occupied by that species (ODonnell et al., 1994). Conservation and recovery of endangered troglobites is accomplished by preserving a sucient number of KFAs for each listed species, based on the number of KFRs in which a species occurs (e.g. ODonnell et al., 1994). KFAs that contain more than one of the listed species, as well as non-listed troglobites, are especially sought to preserve maximum biodiversity. Key factors used to determine the conguration of a KFA includes, but are not limited to KFA and cave size and shape, location of the cave entrance, drainage into the cave, vegetative buer, connectivity with other KFAs or preserved natural areas, and environmental quality. For a KFA to meet recovery criteria, the USFWS must concur that the area has accomplished all of the above criteria and criteria in ODonnell, 1994, and that it is protected in perpetuity. Additionally, while not directly mentioned to avoid confusion in nomenclature, critical habitats designated for the endangered troglobites in the San Antonio area were largely based on the principles that dene KFAs (U.S. Fish and Wildlife Service, 2003). 3.3 Karst ZonesKarst Zones identify the potential occurrence of endangered karst troglobites within KRFs and adjacent areas. Like KFRs, they are based on biological and geological factors. Four zones are generally established: Zone 1: areas known to contain endangered cave fauna; Zone 2: areas having a high probability of suitable habitat for endangered cave fauna; Zone 3: areas that probably do not contain endangered cave fauna; and Zone 4: areas which do not contain endangered cave fauna. Similar zones could also be dened for rare, common, and specic species. Based on the denition of KFRs, it could be assumed that if listed troglobites occur in one cave of a KFR, they occur throughout the KFR, as strongly indicated by revision of Karst Zone mapping as new localities for the endangered species are discovered (e.g. Veni and Martinez, 2006). However, that cannot be used as the basis for eective management of human activities that may adversely impact the species. e USFWS has used Karst Zones in several ways, but primarily as management zones, determining what level of action and research is needed in the protection and study of areas within these zones (e.g. USFWS, 2001). Karst Zones are also used by businesses seeking to avoid environmentally sensitive areas, which would likely be more expensive to develop, leaving them more available for possible purchase for protection. e City of San Antonio incorporated Karst Zones into its geographic information system model for identifying environmentally valuable land for acquisition and conservation (Stone and Schindel, 2002). Zone 1 is the most critical Karst Zone. It answers the question How far can I be from a cave with endangered species before I am reasonably certain that the species may not be present? Contacts between geologic units where caves are common versus units where caves are rare or absent are the most reliable factors in delimiting Zone 1 boundaries, but many Zone 1 boundaries are not that simple to dene. Where no known discontinuity occurs in the cavernous limestone and for lack of other possible options, Zone 1 boundaries might be drawn along creek beds and the locally narrowest or lowest drainage divide. ese locations are where the limestone is thinnest and may pose some restrictions on species distribution. Faults with cavernous rock on either side do not seem to restrict species distribution, but they may be selected as a Zone 1 boundary if other possibilities are exhausted. While some caves form along faults, fault planes lled with calcite or gouge are unlikely sites for cave development. Other factors considered in the delineation of Zone 1 boundaries include: Comparing the lowest known cave elevation with the lowest topographic elevation to be sure at least the known cavernous zone in the rock is encompassed. Examining the distribution of troglobites in dierent caves. If the same troglobites, and especially the same endangered troglobites, occur in dierent caves, those caves may warrant grouping as a single Zone 1 area. e quality of the collections should also be weighed. Collections conducted only once, under poor conditions, cursorily, and/or by non-specialists in the collection of cave species, should be given greater

PAGE 170

15th International Congress of Speleology Rare Species 775 2009 ICS Proceedings weight for similarity of species, since more detailed studies would likely yield more similarities. Evaluating the type and extent of cave development in the area will help determine how realistic it may be for cavernous voids to occur in locations considered as zone boundaries. Assessing other caves in the area, especially if they occur between caves with listed species. is demonstrates the presence of potential habitat for the species, unless those caves have been carefully biologically surveyed and the species were not found. ese factors are not always consistent. For example, the geology may suggest a restriction, but the biology may indicate the opposite. All available factors and information should be considered to determine which features and locations are the mostly likely potential boundaries. Zone 2 is oen a potential Zone 1, where no reason is known to preclude the presence of the listed species, but where the listed species have not been found. is is primarily true where Zone 2 is adjacent to Zone 1. In most such cases, Zone 2 areas are where caves have not yet been discovered and/or biological surveys in the caves have not been conducted. Major exceptions to these types of Zone 2 are: Nearby but hydrogeologically isolated areas where troglobite populations may not have yet diverged. Karst that extends contiguously to where a dierent suite of species lls the listed species ecological niche and the interface between the groups has not been delineated. Areas that would generally be considered Zone 1, except for being buried under thick alluvium or poorly permeable strata that limit nutrient and possibly moisture input to the underlying rocks, making any underlying voids less suitable habitat. Zone 3 is typically poorly karsted limestone near or adjacent to Zone 1 or 2, where troglobites are rare or not known. Zone 4 is either a non-karst area, or a biologically wellstudied karst area proven to have a dierent suite of species.3.4 Karst preservesKarst preserves are locales protected from non-natural activities to sustain the surface and subsurface components of rare karst ecosystems known to occur within certain caves. While oen called cave preserves, their eectiveness depends on protecting not just the known portions of caves but the parts of the surrounding karst that biologically and hydrologically aects the caves. Key considerations include maintaining for caves and their mesocaverns areas: a high quality of water entering a preserve by surface or subsurface routes, natural water quantity, stable microclimatic conditions, natural quantities and quality of nutrients entering the subsurface, healthy surface and subsurface ecological communities, and a minimum of contaminants, non-native plants and animals, and non-natural disturbance. e intent of the karst preserves is to provide refuges that insure the survival of the endangered troglobites, including worst case scenarios where all surrounding land would be fully impacted by urban or other land use detrimental to the species. Karst preserves are developed in a similar manner to KFAs and could rise to the level of a KFA if they are protected in perpetuity and the USFWS concurs that the preserve meets recovery criteria. e USFWS (2006) provides the most current and comprehensive discussion of eective karst preserve design and maintenance, even though the report is in dra form for public comment (the nal dra should be completed by the end of 2009). 4. Conclusionse study, conservation, and recovery of rare and endangered karst troglobites is complex and vexed with uncertainties, since the species can only be observed for short periods and most of the habitat is physically inaccessible to humans. While many strictly biological and environmental factors dramatically aect the species, understanding the karst landscapes hydrogeological evolution is critically important in evaluating their biogeography and establishing eective management practices.ReferencesCollins, A.D. (1989) Geochemistry and ow characteristics of Edwards Aquifer springs: Washita Prairie, central Texas. Baylor Geological Studies Bulletin, 48, p. 10. Hobbs, H.H., III (2005) Diversity patterns in the United States. In Encyclopedia of Caves, David C. Culver and William B. White (Eds.), Elsevier Academic Press, p. 170. Juberthei, C., and B. Delay (1981) Ecological and biological implications of the existence of a supercial underground compartment. In Proceedings of the Eighth International Congress of Speleology, Bowling Green, Kentucky, p. 203.

PAGE 171

Rare Species 776 2009 ICS P roceedings 15th International Congress of Speleology ODonnell, L., W.R. Elliott, and R.A. Stanford (1994) Recovery plan for endangered karst invertebrates in Travis and Williamson counties, Texas. U.S. Fish and Wildlife Service, Region 2, Albuquerque, New Mexico, 154 p. Paquin, P., and Duprr (2009) A rst step towards the revision of Cicurina: redescription of type specimens of 60 troglobitic species of the subgenus Cicurella (Araneae: Dictynidae), and a rst visual assessment of their distribution. In Zootaxa 2002, Magnolia Press, p. 1. Stone, D., and G.M. Schindel (2002) e application of GIS in support of land acquisition for the protection of sensitive groundwater recharge properties in the Edwards Aquifer of south-central Texas. Journal of Cave and Karst Studies, 64(1), 38. U.S. Fish and Wildlife Service (1988) Endangered and threatened wildlife and plants; nal rule to determine ve Texas cave invertebrates to be endangered species. F ederal RRegister, 53, 36,02936,033. U.S. Fish and Wildlife Service (2000) Endangered and threatened wildlife and plants; nal rule to list nine Bexar County, Texas invertebrate species as endangered. F ederal R R egister 63(248) 81,419-81,433. U.S. Fish and Wildlife Service (2001) Karst feature survey protocols, May 23, 2001 version. Austin Field Oce, U.S. Fish and Wildlife Service, 10 p. U.S. Fish and Wildlife Service (2003) Endangered and threatened wildlife and plants; Designation of critical habitat for seven Bexar County, TX, invertebrates species. Federal Register, 68(67), 17,156-17,231. U.S. Fish and Wildlife Service (2006) Recovery plan for nine endangered karst invertebrates in Bexar County, Texas. U.S. Fish and Wildlife Service, 84 p. Veni, G. (2005) Lithology as a predictive tool of conduit morphology and hydrology in environmental impact assessments. In Sinkholes and the Engineering and Enironmental Impacts of Karst, Geotechnical Special Publication No. 144, American Society of Civil Engineers, p. 46. Veni, G., and Associates (1992) Geologic controls on cave development and the distribution of cave fauna in the Austin, Texas, region. Report for the U.S. Fish and Wildlife Service, 77 p. Veni, G., and Associates (1994) Geologic controls on cave development and the distribution of endemic cave fauna in the San Antonio, Texas, region. Report for Texas Parks and Wildlife Department and U.S. Fish and Wildlife Service, 99 p. Veni, G., and C. Martinez (2006) Revision of karst species zones for the Austin, Texas, area. George Veni and Associates report for the U.S. Fish and Wildlife Service, 48 p. White, K. (2006) Paleohydrology of the Edwards Aquifer karst and the evolution of rare and endangered Cicurina cave spiders, south-central Texas. Ph.D. Dissertation, e University of Mississippi, 125 p.

PAGE 172

15th International Congress of Speleology Rare Species 777 2009 ICS Proceedings PRESERVE DESIGN FOR ENDANGERED KARST INVERTEBRATES IN BEXAR COUNTY, TEXASCYNT NT HIA A A A WAT AT SON ON United States Fish and Wildlife Service, 10711 Bu rnet RRoad, Suite 200, AA ustin, TT exas 78758 USA A Nine federally endangered karst invertebrates inhabit caves and mesocaverns in Bexar County, Texas. ese animals depend on high humidity, stable temperatures, and nutrients derived from the surface. reats to these species include habitat loss and degradation associated with development. e recovery strategy (Service 2008) to protect these species includes the perpetual preservation and management of an adequate quantity and quality of habitat that spans each species range. Multiple preserves (quantity) across each species range are desirable to 1) reduce the risk that a catastrophic event would extirpate the species 2) to protect the genetic diversity and 3) allow possible migration or population dynamics necessary for long-term viability. uality of habitat refers to the condition and orientation of preserve land with respect to the known localities for the species. Preserving habitat, being conservative in terms of allowing enough acreage for adaptive management, monitoring, and conducting research to rene our understanding of the species are key components of recovery and ensuring the establishment of karst preserves that will protect these species in perpetuity. e reasoning and scientic support behind the recommended quantity and quality of habitat necessary to protect these species will be presented in a poster presentation at the International Congress on Speleology. 1. IntroductionOn 26 December 2000, nine karst invertebrates were listed as endangered species in Bexar County, Texas. ese species are troglobites and are restricted to the subterranean environment. ey have pale coloration and small or absent eyes. eir habitat includes caves and mesocavernous voids in karst (terrain characterized by sinkholes and caves, produced by solution of bedrock). ese animals depend on high humidity, stable temperatures, and nutrients derived from the surface. Examples of nutrient sources include leaf litter fallen or washed in, animal droppings, and animal carcasses. It is imperative to consider that while these species spend their entire lives underground; their ecosystem is dependent on the epigean (surface) habitat. Herein, information from Appendix B, titled Preserve Design of the dra Bexar County Karst Invertebrate Recovery Plan, is presented explaining the reasoning and scientic support for the recommended quantity (size) and quality of karst preserves necessary to recovery these species. 2. Preserve Design Principlese objective of designing a karst preserve is to protect the surface and subsurface drainage basins of an occupied karst feature and adequate surface habitat to maintain native plant and animal communities around the feature. Details of the area needed to protect the feature are dicult to dene due to limited information on the dynamics of the species and ecosystem processes. Furthermore, population trends of all the listed invertebrates are dicult to obtain due to small sample sizes. is means that the only way to determine with certainty that a preserve is inadequate to support karst invertebrates is to document the extinction of a population by observing no specimens for many years. Because it is unknown if these species can be reintroduced or migrate (except over evolutionary time) into habitat, this is not acceptable. In addition, if a preserve is later found to be inadequate to support the species due to developments being too close or dense, the potential for preserving that land or for adaptive management is lost. Because these species have relatively long life-spans and low requirements for food, a decline in population size or even the complete extinction of the population may take decades. Observations of a listed species over many years on a preserve that is too small for species preservation may not reveal declines. If these observations are used as evidence that a preserve size was adequate, then the potential for long-term preservation of that species may become lost due to irreversible development surrounding the preserve. So, due to the unique considerations of population viability and habitat requirements for these species, preserve design should be based on estimates and assumptions that favor a high probability for species conservation.

PAGE 173

Rare Species 778 2009 ICS P roceedings 15th International Congress of Speleology e concept of how much is enough should be answered in the context of the surrounding conditions (Harris l984). ree critical elements identied for maintaining habitat islands are the actual habitat size, the distance from similar habitat, and the degree of dierence in the intervening matrix. Lord and Norton (1990) also cite ecosystem vulnerability to extrinsic disturbances. Because karst ecosystems cannot be recreated once destroyed, preserves should be designed and congured conservatively and incorporate the suite of biotic and abiotic factors needed to promote the integrity of fully-functioning ecosystems. To promote long-term, conservation of karst species and ecosystems, preserves should be designed to rely on minimal management to control threats. Size and Shape of Preserves Based on existing literature on habitat patch size, fragmentation, isolation, edge eects, corridors, and other factors considered in minimizing threats to ecosystem stability, a karst preserve should be at least 28 to 40 hectare (ha), including a core and buer area, to protect the integrity of the biotic communities that support the karst ecosystem. In determining the actual size and conguration of a karst preserve, all of the factors listed below should be incorporated in preserve designs. Protection of Water uality and uantity Karst hydrology is more dicult to predict than that of surface water or of porous media groundwater movements. A detailed hydrogeologic investigation should be conducted to determine these drainage basins, local recharge areas, and direction of groundwater movement. It is oen challenging to map these basins. For example, Flint Ridge Cave in Travis County, Texas was initially mapped as having a 0.30 ha drainage basin (State Department of Highways and Transportation 1989), later mapped as 15.8 ha (Veni 2000), and most recently found to be 22 ha (Hauwert et al. 2005). For information on how to determine subsurface drainage basins see Veni 2003, Veni 2004, and Veni and Associates 2002.3. Protection of Habitat Area Needed to Sustain Viable Native Plant CommunitiesA minimum of 28 to 40 ha is likely needed to support a self-sustaining woodland-grassland mosaic community (Service 2003). is includes a core area of 24 to 36 ha and a minimum 20 m buer to protect this core plant community from detrimental edge eects. ese gures are the minimum size needed for an isolated preserve. Preserves that are adjacent to and share a large perimeter with another large preserve, or that are surrounded by low levels of development and native vegetation, in perpetuity, may be smaller. A preserve should be larger the more isolated it is from similar plant communities, or where it may become isolated in the future due to development. Long, narrow corridors that have some advantages to the vertebrate community are not likely to be eective in maintaining the native plant community because this conguration may be more vulnerable to edge eects and exotic species invasion (Saunders et al. 1990, Kotanen et al. 1998, Suarez et al. 1998, Meiners and Steward 1999).4. Protection of Habitat Area Needed to Sustain Viable Native Animal CommunitiesCave Crickets e native animal community important for sustaining and providing nutrient input for karst ecosystems includes cave crickets (Ceuthophilus spp.) as well as other surface fauna. e cave cricket is a particularly important nutrient component (Barr 1968, Reddell 1993) and found in most Texas caves (Reddell 1966). Cave crickets forage on the surface at night up to 105 m from a cave entrance (Taylor et al. 2005). Also, Taylor et al. (2007), compared caves in urban areas to those in natural areas, and found signicant dierences in isotope ratios of cavernicoles between these two levels of impact demonstrating that nutrient ows are dierent in urban and rural areas. ey also found the number of cave crickets is strongly correlated to the number of other cave taxa. erefore, the foraging area of cave crickets and a protective buer should be encompassed in the boundaries of the preserve. Terrestrial Vertebrates Species that occasionally use caves such as raccoons (Procyon lotor), white throated salamander (Plethodon albagula), cli frog (Eleutherodactylus marnocki), and snakes and mice, may play an important role in the ecology of cave systems. Mammals may use caves for shelter from surface temperature extremes and aridity. ough we know of no studies delineating the exact role of mammals in central Texas cave ecology, the presence of a large amount of mammal-derived energy indicates their importance. is energy is in the form of scat, nesting material, and carcasses. Cave collembolan or springtails (a food source for endangered karst invertebrate predators), are frequently seen feeding on the scat (and associated fungus and microorganisms) and mammal carcasses. 5. Continuity of Habitat and Edge EectsEdge eects are changes to the oral and faunal communities where dierent habitats meet. Preferably, preserves will be in an approximately circular or square conguration, to minimize edge eects. e more edge a habitat fragment or patch has, the larger the patch or fragment size should be to protect the core area from the

PAGE 174

15th International Congress of Speleology Rare Species 779 2009 ICS Proceedings deleterious edge eects (Ranny et al. 1981, Lovejoy et al. 1986, Yahner 1988, Laurance 1991, Laurance and Yensen 1991, Kelly and Rotenberry 1993, Holmes et al. 1994, Reed et al. 1996, Turner 1996, Suarez et al. 1998). Minimizing edge eects in a preserve design means keeping the edge-toarea ratio low through increasing the patch size (Holmes et al. 1994) and/or using optimal preserve shapes. e length and width of the edge, as well as the contrast between the vegetational communities, all contribute to the amount of impacts (Smith 1990, Harris 1984). Some types of edge eects include increases in solar radiation, changes in soil moisture due to elevated levels of evapotranspiration, wind bueting (Ranny et al. 1981), changes in nutrient cycling and the hydrological cycle (Saunders et al. 1990), and changes in the rate of leaf litter decomposition (Didham 1998). ese edge eects alter plant communities, which in turn impact the associated animal species. Edge eects can also aect animal species directly. Vegetation 2 m from an edge can be visibly aected within days (Lovejoy et al. 1986). Hard edges can act as a barrier to distribution and dispersal patterns of birds and mammals (Hansson 1998, Yahner 1988). Invertebrate species are also aected by edges. Mader et al. (1990) found that carabid beetles and lycosid spiders avoided crossing unpaved roads that were less than 3 m wide. Roads can also constitute a hindrance to movement in forest-inhabiting mice and other small mammals (Mader et al. 1990). Increases in predation (Andren 1995, Bowers et al. 1996, Suarez et al. 1998) and competition for food sources (Hanski 1995) and den sites (Rosatte et al. 1991) also occur near edges. Saunders et al. (1990) suggest that as little as 100 m of agricultural elds may be a complete barrier to dispersal for small organisms such as invertebrates. Edges oen allow just enough disruption for invasive or exotic species to gain a foothold where the native vegetation had previously prevented their spread (Saunders et al. 1990, Kotanen et al. 1998, Suarez et al. 1998, Meiners and Steward 1999). e invasion of red-imported re ant (Solenopsis inicta) (RIFA), an aggressive predator and threat to the karst invertebrates (Elliott 1994, Service 1994), is known to be aided by any disturbance that clears a site of heavy vegetation and disrupts the native ant community (Porter et al. 1988, Taylor et al, 2007). In southern California, Suarez et al. (1998) found that densities of another exotic ant species, the Argentine ant (Linepithema humile), (similar to RIFA), are highest within 100 m and rare or absent less than 200 m of an urban edge. Native ant communities tended to be more abundant in native vegetation and less abundant in areas with exotic vegetation. Edge eects on various habitats and taxa vary from as little as 15 m to as much as 5 km (Laurance and Yensen 1991). e eects of edge on fauna generally exceed the eects on vegetation. A rule of thumb for the protection of a forest from a clearcut edge is the three tree height rule (Harris 1984). Tree heights for the Edwards woodland association in Texas are 3 to 9 m (Van Auken et al. 1979). An average tree height of 6.6 m was used, and therefore an edge eect of approximately 20 m is estimated. e three tree height approach described by Harris (1984) was based on the distance that eects of storm events (wind-throw) from a surrounding clear-cut edge will penetrate into an oldgrowth forest stand. Since the eects of edge on woodland/ grass land mosaic communities have not been well studied, the three tree height recommendation is considered to be the best available peer-reviewed science to protect woodland areas from edge eects (Dr. Kathryn Kennedy, Center for Plant Conservation, pers. comm. 2003). Other studies, found that invasive species were within 16 to 137 m and 20 to 30 m from an edge; hence, we may be underestimating the area needed to buer against invasive species. For animal communities, reported edge eects are typically 50 to 100 m or greater (Lovejoy et al. 1986, Wilcove et al. 1986, Laurance 1991, Laurance and Yensen 1991, Kapos et al. 1993, Andren 1995, Reed et al. 1996, Burke and Nol 1998, Didham 1998, Suarez et al. 1998 Avoiding Internal Roads and Habitat Fragmentation Because roads may hinder movement of fauna, no internal roads or other permanent habitat fragmentation should occur within a karst preserve. Where human access is critical, a bridge could be installed in lieu of a road, provided it does not alter a critical component of the karst ecosystem, such as the quality and quantity of water entering the subsurface. Internal clearing activities and other disturbances of soil and native vegetation should also be avoided to help minimize RIFA infestations. Urban runo should be diverted away from the karst ecosystem to avoid contamination and increased RIFA activity. Preserve Non-cave Karst Areas Between Known Cave Localities Good connectivity with mesocaverns for population dynamics of troglobites should be maintained. Restrictions on impervious cover and the use of Best Management Practices in the karst areas extending to the entire range of the listed species would provide some landscape scale consideration to the species that may

PAGE 175

Rare Species 780 2009 ICS P roceedings 15th International Congress of Speleology otherwise be susceptible to problems caused by isolation. ese conservation actions will not only help maintain mesocaverns, they will also potentially supply corridors for migration of troglobites, provide surface corridors for trogloxenes, provide genetic diversity for maintaining native ora and fauna, and buer water quality and quantity entering the subsurface. 6. DiscussionTo summarize, it is important to design karst preserve that 1) are 28 to 40 ha; 2) in an approximately circular or square conguration to reduce edge eects; 3) protect the surface and subsurface drainage basins; 4) protect the native animal community including the cave cricket foraging area and a buer; 5) protect the native plant community; and 6) protect mesocavernous areas between occupied caves. e current science indicates that the information above should protect an adequate quantity and quality of habitat to ensure that the endangered karst invertebrates of central Texas can survive in perpetuity; however, further research is needed to rene our understanding of these species. For a comprehensive discussion on preserve design see the dra Bexar County Karst Invertebrate Recovery Plan (Service 2008) Appendix B, titled Preserve Design. Much of the information compiled above was due to the diligent eorts of sta at the US Fish and Wildlife Services Austin Ecological Services Oce and the Karst Invertebrate Recovery Team.7. ReferencesAndren NDREN H. (1995) Eects of landscape composition on predation rates at habitat edges. In Mosaic Landscapes and Ecological Processes, L. Hansson, L. Fahrig and G. Merriam, (Eds.) Chapman and Hall, London. p. 225. BARR, T.C. Jr. (1968) Cave ecology and the evolution of troglobites. Evolutionary Biology 2, 35. BOWERS, M.A., K. GREARIO, C.J. BRAME, S.F. MATTER, and J.L. DOOLEY, Jr. (1996) Use of space and habitats by meadow voles at the home range, patch and landscape scales. OO ecologica 105, 107-115. BURKE, D.M. and E. NOL (1998) Inuence of food abundance, nest-site habitat, and forest fragmentation on breeding ovenbirds. e AA uk 115(1): 96. DIDHAM, R. (1998) Altered leaf-litter decomposition rates in tropical forest fragments. OO ecologia 116, 397. Elliott L LIOTT W.R. (1994) Conservation of Texas caves and karst. In e Caves and Karst of TT exas, 1994 N N SS Conention guidebook. W.R. Elliott and G. Veni, (Eds.), National Speleological Society, Huntsville, Alabama. p. 85. Hanski A NSKI I. (1995) Eects of landscape pattern on competitive interactions. In Mosaic Landscapes and Ecological Processes, L. Hansson, L. Fahrig, and G. Merriam, (Eds.) Chapman and Hall, London. p. 203. HANSSON, L. (1998) Local hot spots and their edge eects: small mammals in oak-hazel woodland. OO i kos 81, 55-62. Harris A RRIS L. (1984) e agmented forest: island biogeography theory and the preservation of biotic diversity. University of Chicago Press. Chicago, Illinois. Hauwert A UWERT N., M. Litvak ITVAK and J. Sharp HARP Jr. (2005) Characterization and water balance of internal drainage sinkholes. In TT e nth Multidisciplinary Conference on Sinkholes and the Engineering and Enironmental Impacts on Karst. San Antonio, Texas. HOLMES, E.E., M.A. LEWIS, J.E. BANKS, and R.R. VEIT (1994) Partial dierential equations in ecology: spatial interactions and population dynamics. Ecology 75, 17. KAPOS, V., G. GANADE, E. MATUSUI, and R.L. VICTORIA (1993) Carbon 13 isotope as an indicator of edge eects in tropical rainforest reserves. Ecology 81, 425. Kelly E LLY P.A. and J.T. Rotenberry O TENBERRY 1993. Buer zones for ecological reserves in California: replacing the guesswork with science. In Interface Between Ecology and L and DDevelopment in California, J.E. Keeley (Eds.) Southern California Academy of Sciences, Los Angeles. p 85. KOTANEN, P.M., J. BERGELSON, and D.L. HAZLETT (1998) Habitats of native and exotic plants in

PAGE 176

15th International Congress of Speleology Rare Species 781 2009 ICS Proceedings Colorado shortgrass steppe: a comparative approach. Canadian Journal of Botany 76, 664. LAURENCE, W.F. (1991) Edge eects in tropical forest fragments: application of a model for the design of nature reserves. Biological Conservation 57, 205. LAURENCE, W.F. and E. YENSEN (1991) Predicting the impacts of edge eects in fragmented habitats. Biological Conservation 55, 77. LORD, J. and D. NORTON (1990) Scale and the spatial concept of fragmentation. Conservation Biology 4, 197. Lovejoy O VEJOY T.E., R.O. Bierregaard I ERREGAARD A.B. Rylands Y LANDS J.R. Malcolm ALCOLM C.E. uintela UINTELA L.H. Harper, K.S. Brown, A.H. Powell, G.V.N. Powell, H.O.R Schubert, and M.J. Hays (1986) Edge and other eects of isolation on Amazon forest fragments. In Conservation Biology: e Science and S carcity of DD iversity. M. Soul (Eds.) Sunderland, Massachusetts. p. 7. MADER, H.J., C. SCHELL, and P. KORNACKER (1990) Linear barriers to arthropod movements in the landscape. Biological Conservation 54, 209. MEINERS, S. and P. STEWARD (1999) Changes in community and population responses across a forest to eld gradient. Ecography 22, 261. PORTER, S.D., B. VAN EIMEREN, and L.E. GILBERT (1988) Invasion of red imported re ants (Hymenoptera: Formicidae): microgeography of competitive replacement. AA n nals of the Entomological Society of AA merica 81, 913. Ranny A NNY J.W., M.C. Bruner R UNER and J.B. Levenson EVENSON (1981) e importance of edge in the structure and dynamics of forest islands. in F orest Island DD ynamics in Man-D Dominated Landscapes. R.L. Burgess and D.M. Sharpe, (Eds.) Springer Verlag, New York. p. 67. REDDELL, J.R. (1966) A checklist of the cave fauna of Texas. II. Insecta. e TT exas Journal of Science 18, 25. Reddell E DDELL J.R. (1993) Response to the petition to delist seven endangered karst invertebrates. Letter to U.S. Fish and Wildlife Service, Austin, Texas. July 10, 1993. REED, R., J. JOHNSON-BARNARD, and W.L. BAKER (1996) Fragmentation of a forested rocky mountain landscape, 1950. Biological Conservation 75, 267. Rosatte O SATTE R.C., M.J. Powers O WERS and C.D. Mac A C Innes NNES (1991) Ecology of urban skunks, raccoons, and foxes in metropolitan Toronto. in Wildlife Conservation in M etropolitan Enironments: NN ational Institute for Urban Wildlife Symposium Series 2. L.W. Adams and D.L. Leedy, (Eds.) National Institute for Urban Wildlife. Columbia, MD. p. 31. Smith M ITH R.L (1990) Ecology and Field Biology, 4th edition. Harper Collins Publishers, Inc., New York. SAUNDERS, D.A., R.J. HOBBS, and C.R. MARGULES (1990) Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5, 18. SUAREZ, A.V., D.T. BOLGER, and T.J. CASE (1998) Eects of fragmentation and invasions on native ant communities in coastal Southern California. Ecology 79, 2041. Service (U.S. Fish and Wildlife Service). 1994. Recovery plan for endangered karst invertebrates in Travis and Williamson Counties. Austin, Texas. 154 pp. Service (U.S. Fish and Wildlife Service). 2003. Designation of critical habitat for seven Bexar County, TX, invertebrate species, nal rule. Federal Register 68: 55063. Service (U.S. Fish and Wildlife Service). 2008. Dra Karst Invertebrate Recovery Plan. 125 pp. State Department of Highways and Transportation. 1989. Final Environmental Impact statement, Austin Outer Parkway SH45 Segment 3. TAYLOR, S.J., J. KREJCA, and M.L. DENIGHT (2005) Foraging range and habitat use of Ceuthophilus secretus (Orthoptera: Rhaphidophoridae), a key trogloxene in central Texas cave communities. A A m erican Midland NN aturalist 154, 97.

PAGE 177

Rare Species 782 2009 ICS P roceedings 15th International Congress of Speleology Taylor AYLOR S.J., J. K. Krejca R EJCA and K. Hackley ACKLEY (2007) Examining possible foraging dierences in urban and rural cave cricket populations: carbon and nitrogen isotope ratios ( 13 C, 15 N) as indicators of trophic level. Illinois Natural History Survey Technical Report 2007(59), 97 pp. TURNER, I. (1996) Species loss in fragments of tropical rain forest: a review of the evidence. Journal of A A p plied Ecology 33, 200. VAN AUKEN, O.W., A.L. FORD, and A. STEIN (1979) A comparison of some woody upland and riparian plant communities of the Southern Edwards Plateau. S outhwestern NN aturalist 24, 165. Veni E NI G. (2000) Hydrogeologic assessment of Flint Ridge Cave, Travis County, Texas. Consulting report submitted to the City of Austin. Veni E NI G. (2003) Delineation of hydrogeologic areas and zones for the management and recovery of endangered karst invertebrate species in Bexar County, Texas. Report for U.S. Fish and Wildlife Service, Austin, Texas, prepared by George Veni and Associates, San Antonio, Texas. Dated 23 December 2002 with minor revisions submitted April 12, 2003. (pgs) Veni E NI G. (2004) Environmental impacts assessments. In Encyclopedia of Cave and Karst Science, John Gunn, (Ed.), Fitzroy Dearborn Publishers, London, p. 319. Veni E NI G. and Associates (2002) Hydrogeologic and biological assessment of caves and karst features along proposed state highway 45, Williamson County, Texas. A report for Hicks and Company. Wilcove I LCOVE D., C. Mc C Lellan E LLAN and A. Dobson OBSON (1986) Habitat fragmentation in the temperate zone. In Conservation Biology: e Science of Scarcity and D D iversity. M. Souleb (Ed.), Sinauer Associates, Inc. Sunderland, Massachusetts. p. 217 YAHNER, R.H. (1988) Changes in wildlife communities near edges. Conservation Biology 2, 333.

PAGE 178

15th International Congress of Speleology Rare Species 783 2009 ICS Proceedings PHYLOGEOGRAPHIC MODELING OF THE EDWARDS AUIFER KARST AS A MANAGEMENT TOOL FOR RARE AND ENDANGERED SPECIES IN CENTRAL TEXAS; A CASE STUDY USING TROGLOBITIC CICURINA SPIDERSK. Whi HI Te E and P. PAQui UI N Cave and Endangered Inertebrate RResearch Laboratory, SWCA A Enironmental Consultants 4407 Monterey OOaks Bld., Bld g. 1, Suite 110, AA ustin, TT X 78749 Abstract Four species of troglobitic Cicurina cave spiders known from the Balcones Escarpment in and around San Antonio, Texas are listed as endangered under the Federal Endangered Species Act (ESA). Many other troglobitic Cicurina from central Texas are petitioned for listing. A compelling argument for listing is that economic development is degrading and fragmenting habitat faster than it can be surveyed for biota. Among thousands of known caves in central Texas, only a small percentage have been surveyed due to the limited number of researchers and restrictions on access to private land. Based on mtDNA data, diversity among troglobitic Cicurina spiders is the product of the progressive availability of vadose zone habitat as discrete recharge areas developed within the broader Edwards aquifer system. Older genetic lineages occur in structurally high, mature karst terrains while the younger lineages occur in structurally low, emergent karst terrains. Since these areas are geographically concordant, non-overlapping and strongly correlated with geologic structure, genetic mapping of the vadose zone provides the rst phylogeographic model for predicting the distribution of listed species and for predicting which un-sampled areas are likely to be rich in genetically distinct populations or species. As a short cut to traditional biota survey strategies, phylogenetic modeling provides a robust basis for regional conservation planning and more ecient allocation of limited karst management resources.

PAGE 179

Rare Species 784 2009 ICS P roceedings 15th International Congress of Speleology

PAGE 180


PAGE 182

15th International Congress of Speleology Speleogenesis 787 2009 ICS Proceedings originORIGIN and AND development DEVELOPMENT of OF the THE dam DAM valley VALLEY lakes LAKES and AND related RELATED karst K ARST hydrogeologic HYDROGEOLOGIC systems SYSTEMS : Amtkeli MTKELI river RIVER Western ESTERN Caucasus AUCASUS G.N N Amelichev AMELICHEV B.A A VAkh KH Rushev USHEV and v V .N. Dubly DUBLY ANskiy SKIY Ukrainian Institute of Speleology and Karstology, 4 Vernadsky Prospect, Simferopol 95007, Ukraine, Abstract Dammed lakes in river valleys of mountain karst regions can be formed by rockslides. Ancient relict cave systems, which had already been abandoned by their ow, can have their hydrologic function reactivated as a result. Interest in their study is due to the fact that they represent a natural model of an engineering situation that arises during dam construction in karst. e Amtkel karst area is situated on the south slope of the Abkhazsky mountain range, in the western Caucasus, where the River Amtkel crosses the belt of Cretaceous limestone. Aer an earthquake-induced rockslide in 1891, a dammed lake was formed in the river. Relict caves in the slopes of the valley were ooded and began to function as water intakes. e valley downstream from the rockslide dam was drained, which made former intakes at the river bed open and accessible. In order to study cave hydrogeologic systems of the area, topographic, geologic, bathymetric, hydrochemical, thermal, and speleological investigations have been conducted. e large karst hydrogeologic system has been revealed, where ancient components received intense recharge from lake waters. Areas of recharge, transit, and discharge have been identied in the system, with discharge occurring to a karst spring located 9 km away from the dammed lake. ree hydrochemical facies were identied in waters of the area: non-karstic groundwaters in the Paleogene sediments, Amtkeli Lake waters with two temperature sub-facies (shallow and bottom), and river waters. Chemical processes in the system have been modeled through the use of mixing equations for both low-ow and high-ow regimes. In this way the speleogenetic evolution of the karst system of the Amtkeli Lake and river during the Pliocene-uaternary has been reconstructed.

PAGE 183

Speleogenesis 788 2009 ICS Proceedings 15th International Congress of Speleology Base ASE level LEVEL rise RISE and AND PEr R Asc SC Ensu NSU M Model ODEL of OF Speleogenesis PELEOGENESIS PAMS. Interpretation N TERPRETATION of OF deep DEEP phreatic PHREATIC karsts KARSTS vauclusian VAUCLUSIAN springs SPRINGS and AND chimney CHIMNEY shafts SHAFTS .Philippe HILIPPE A A UDRA DRA1 Lu U DOvic VIC MO O CO O CHA A IN N2, and Je E ANYves VES BIGOT OT3 1PolytechN N ice-Sophia, Engineering School of NN ice Sophia AA ntipolis University, 1645 route des Lucioles, 06410 BIOT OT France2University of AA ix-Marseille, CER R EGE, Europle de lA A rbois, BP 80, 13545 AA ix-en-Proence, Cedex 4, France & Centre de Sdimentologie Palontologie Gologie des systmes carbonats  13331 Marseille, Cedex 03, France3French AA ssociation of Karstology In Mediterranean karsts, the Messinian Salinity Crisis induced rst a deepening of the karst systems, then a ooding aer the Pliocene transgression, and nally a reorganization of the drains aer this base level rise. is reorganization mainly corresponds to the development of phreatic lis: the chimney-shas and the vauclusian springs. Such a per ascensum speleogenesis appears with a base level rise, which is caused by eustatism, by uvial aggradation or valley inlling, or by continental subsidence. Consequently, we explain the origin of most of the deep phreatic cave systems (which are not hypogenic) by a base level rise which ooded the deep karst, producing phreatic lis connected to vauclusian springs.1. IntroductionWhere no impervious aquiclude is present, cave levels can be correlated to base level (G et al. 2001; A & G 2004; H et al. 2007). Authors generally explain them as the result of descending base level caused by valley incision. Cave levels are implicitly associated with a per descensum evolution. Since the ages of cave levels are correlated to successive stages of valley entrenchment, the lowest levels are considered the youngest, and conversely (P 1987). And when a base-level rise is taken into account, its role is generally limited to the ooding and lling of cave systems, without noticeable speleogenesis. Studies of speleogenesis associated with the MessinianPliocene eustatic cycle, i.e., the succession of Messinian Salinity Crisis (MSC) and Pliocene High Stand (PHS), demonstrate the speleogenetic role of base-level rise as a per ascensum process, by the formation of phreatic lis, or chimney-shas (Mff et al. 2006). By extension, other contexts of base level rise, mainly caused by uvial aggradation, produce a similar speleogenesis, making it possible to extend the P er AA scensum Model of Speleogenesis (PAMS). is paper presents our results, carried out rst in the French Mediterranean area and associated with the impacts of the MSC. Second, we extrapolate to other contexts of base-level rise that also show a PAMS. eir origin could be eustatic, climatic (transgression or uvial aggradation), or tectonic (regional subsidence).2. e PAMS Associated with the MessinoPliocene Cycle in the Mediterraneane French Mediterranean periphery displays a cluster of deep phreatic cave systems (Fig. 1). Many authors once interpreted it by the Four State Model (F 1977), assigning a bathyphreatic origin with a speleogenesis not inuenced by the base-level position. From the 1980s onward, according to concepts developed by Clauzon et al. (1997), the origin of such a deep-phreatic speleogenesis gradually shied to the MSC. is revision provides conceptual tools based on the inuence of large-scale base level changes on deep phreatic cave systems. Recent studies have identied several types of ooded cave systems (Mff 2007; A 2007). is typology is built not only on morphological criteria, but also on the elevation of the caves according to the current base-level position. It is possible to distinguish ooded cave systems located mainly below the base level (marine or uvial), from those currently in the vadose zone but having a phreatic origin.2.1 Flooded coastal karst, the Port-Miou submarine springe Port-Miou submarine spring, near Marseille, is fed by part of the Provence karst (Fig. 1). It has been explored for more than 2 km and down to 179 m depth (Fig. 2). e oshore bathymetry reveals submerged karst features (dolines, poljes, and canyons) (Bf & M 1988; CG 1996). e Cassidaigne Canyon is interpreted as an old pocket valley developed during the Messinian low

PAGE 184

15th International Congress of Speleology Speleogenesis 789 2009 ICS Proceedings Figure 1: Deep-phreatic cave systems in Mediterranean France. All cave systems are connected to the Mediterranean or to the Pliocene rias (ooded valley) (map aer Clauzon et al. 1997; Camus 2003; caves updated om Audra 1997). sea level. Since the Pliocene transgression, the deep karst has been ooded. Sea water enters several kilometers into the aquifers through the old Messinian drains. is intrusion is responsible for the salinity of the spring (G 2001; B et al. 2004; C 2007).2.2 Flooded continental karst: the Fontaine de Vauclusee Fontaine de Vaucluse drains the largest karst area of France (1130 km2, Q = 23 m3/s; Fig. 1). It is famous for its considerable depth of 308 m, i.e., 224 m below current sea level (B & G 1987). Wall karren are developing down to 170m below sea level. ey testify to past epiphreatic conditions by successive ooding and draining. e Fontaine de Vaucluse appeared during the MSC (G & A 2004). Seismic investigations reveal Figure 2: Port-Miou submarine spring. In the Messinian drain, opening oshore at the head of the Cassidaigne Canyon, the water is blocked by the density of the salt water. e underground ow follows a phreatic li connected to the submarine spring (Blavoux et al. 2004; survey aer Douchet & Fage 1993).

PAGE 185

Speleogenesis 790 2009 ICS Proceedings 15th International Congress of Speleology a Messinian canyon lled with sediments, located 20 km to the west and originating from the Fontaine de Vaucluse (Sf et al. 1997). is pocket valley has been lled during the Pliocene (Fig. 3). e ll blocked the canyon at depth and forced the ow upward and to use the past overow route as a perennial spring. e lowest part of the karst is ooded to a great depth. A similar evolution occurred in Ardche, where the Goul du Pont and Goul de la Tannerie springs have been explored by scuba divers down to -220 m (Fig. 1).2.3 Drained karst: the Ardche e canyons were deeply entrenched during the Messinian and then lled with sediments during the Pliocene, causing a base-level rise of similar amplitude. is rise rst occurred by ooding of the valleys during the Pliocene transgression, then by uvial aggradation through to the end of the Pliocene (Fig. 4). Cave levels are correlated with the successive positions of the base level during the MessinianPliocene cycle. Foussoubie is a 25-km long cave system with a main drain displaying a regular gradient (2.5%) between the sinkhole and the resurgence in the Ardche Gorge (Fig. 5). Above the resurgence are vertical series with phreatic features that Figure 3: Speleogenetic model of the Vaucluse karst. e Messinian canyon of the Rhne River is lled with Pliocene deposits. Its bottom is at 900 m below current sea level. e Messinian cave system, which was probably connected to this canyon, has been ooded and lled with sediment during the Pliocene. Since that time, the underground ow has used a chimney-sha opening to the current base level at the Fontaine de Vaucluse. Figure 4: PAMS during the Messinian-Pliocene cycle. Le: Messinian canyon entrenchment caused the deepening of karst drainage. Center: Pliocene base level rise occurred in two steps by marine ingression as ria (dark gray), then by uvial aggradation (light gray). Deep drainage uses phreatic lis to emerge as vauclusian springs, recording successive positions of the base level. If the Messinian canyon is located below the current base level, it remains fossil; the karst remains ooded and discharges by a vauclusian spring (fontaine de Vaucluse type). Right: if the Messinian canyon is located aboe the current base level, the canyon is exhumed and the karst is drained. e current drainage uses the deep Messinian drain; the Pliocene phreatic lis are abandoned as fossil chimney-shas. Figure 5: PAMS in the Foussoubie Cave System, Ardche. e phreatic lis resurge at vauclusian springs connected to the higher base-level positions corresponding to the Pliocene lling. During the Pleistocene, the Messinian canyon was exhumed and cleared of its Pliocene ll down to its bottom. Current drainage reuses the deep Messinian drain. e Pliocene phreatic lis are abandoned as fossil chimney-shas (Bigot 2002; Mocochain 2007; Mocochain et al. 2006, 2008).

PAGE 186

15th International Congress of Speleology Speleogenesis 791 2009 ICS Proceedings clearly show a rising ow. e rectilinear long prole shows that the main drain developed during the Messinian, at a base level at the bottom of the Ardche Canyon (B 2002; Mff 2007). Filling of the canyon causes the development of resurgences as phreatic lis, or chimneyshas. e elevations of resurgences record the stages of base-level rise due to Pliocene uvial aggradation. During the Pleistocene, the Messinian canyon of the Ardche was exhumed by clearing away of the Pliocene lling: Foussoubie chimney-shas became fossil, and the Messinian drain returned to a vadose ow (Figs. 4, 5). In partly exhumed canyons, the lower part of the karst has remained ooded since the beginning of the Pliocene, and they discharge as vauclusian springs (Fontaine de Vaucluse type). In the entirely exhumed canyons, the karst is drained and the chimney-shas are fossil (Foussoubie type). In turn, the chimney-shas, which are systematically associated with paragenesis, are interpreted as a record of the PAMS, which originates from a base-level rise. Consequently, a baselevel rise is interpreted to be a founder speleogenetic event. Besides the Messinian-Pliocene cycle, other causes of baselevel rise also produce per ascensum speleogenesis and the development of chimney-shas. 3. Extrapolation of the PAMS to Other Causes of Base-Level Risee speleogenetic role of the Messinian-Pliocene cycle could be attributed to a dramatic base-level drop that allowed a deepening of karst drainage, followed by a base-level rise of similar magnitude. is base-level rise ooded the deep drainage and developed chimney-shas, sometimes associated with new horizontal cave levels, as in Saint-Marcel Cave, Ardche (Mff et al. 2006). e occurrence of deep phreatic karsts, vauclusian springs, and chimney shas all around the Mediterranean is a consequence of speleogenesis during the Messinian-Pliocene cycle (Figs. 1, 7). Besides the Messino-Pliocene cycle, the PAMS applies to every kind of base-level rise (following a low base-level position). A base-level rise is shown by lling of the lowest parts of valleys by water, ice, or sediment. e driving force could be eustatic (transgression), tectonic (subsidence), climatic (clearing of slopes soils, glacial advance), or even anthropic (e.g., man-made dams).3.1 e Miocene eustatic cyclesIn the Rhodanian-Provence foreland basin between the Fontaine de Vaucluse and the Rhne, the marine molasse records several eustatic cycles during the Miocene (Aquitanian, Burdigalian). e regression, which is linked with tectonic upli, follows valley entrenchment up to 100m-deep, with eventual ooding and lling with sediments by transgressions (B et al. 2005a, 2005b; P et al. 1997). Near the Fontaine de Vaucluse, a fossil pocket-valley ends exactly at the Valescure Sha, which displays characteristic chimney-sha features. e Valescure Sha used to be a vauclusian spring during the Burdigalian, following the lling of the pocket-valley with the molasse. An earlier outow should exist, buried beneath the molasse sediments. Figure 6: Le: Podtrao jeskyn, Moravian karst, Czech Republic, a 140-m high chimney-sha, the lowest part of which is ooded below the Beroukna valley (Bruthans & Zeman 2003). It could show a record of the base-level rise of the hydrologic network aer pre-Badenian entrenchment. Center: the Puits des Bans and the Gillardes Spring (French Alps). e basin ll (glacial, lacustrine, and uvio-glacial) has blocked the Gillardes Spring. In high water, the Puits des Bans, a 300m-high chimney-sha, oods and overows. Right: Lagoa Misteriosa (Brazil), a 200-m deep phreatic sha, a window in a karst aquifer ooded aer the continent subsidence (survey by G. Menezes).

PAGE 187

Speleogenesis 792 2009 ICS Proceedings 15th International Congress of Speleology In the Rhodanian-Provence basin, the speleogenetic inuence of the Miocene eustatic cycles is partly hidden by the imprint of the younger Messinian-Pliocene eustatic cycle. On the contrary, the Paratethys molassic basin of central Europe, at least in its northern part (Czech Republic, Slovakia, Poland, etc.), has not been aected by Messinian entrenchment. Consequently, the oldest eustatic cycles are better recorded. e transgressions of the Carpathian (i.e., Burdigalian) and especially of the Middle-Badenian (i.e., Langhian-Serravalian) follow continental erosional phases, which deepened valleys as much as 150-200 m, and which were later fossilized. Pre-Badenian karsts are well-known: tower karst in Zbrasov (Czech Republic) partly exhumed from the molasse; caves lled with molasse in Bohemian; caves and uviokarst morphologies in the Moravian karst (K et al., 2001). In the Bohemian karst, the Podtraov jeskyn (cave) is a chimney-sha partly drained and more than 100 m deep (Fig. 6). It is developed below the Beroukna Valley, which was entrenched before the Burdigalian and then exhumed during the Pleistocene (B & Z 2003, Fig. 6). If some caves in this area have a hypogenic origin, its chimney-sha features would have recorded the Miocene base-level rise by per ascensum speleogenesis.3.2 Glacio-eustatic transgressionPost-glacial sea-level rises have ooded the coastal karsts (Fig. 7), including the cave systems developed during previous low sea levels. It is evidenced by submerged speleothems, which have been observed down to -120m, around the Gulf of Mexico: Yucatan Peninsula, Bahamian blue holes, Wakula Spring in Florida, etc. Such types of karst discharge through vauclusian springs at the mouths of phreatic lis. In French Normandy, Pleistocene sealevel changes are well recorded in cave systems developed in chalk. e high conductivity of the chalk allows cave systems to adapt precisely to the slightest base-level changes, with chimney-shas less than 10 m high (R 1991; R et al. 2001).3.3 Fluvio-glacial oodingGlacial retreat leaves moraine dams across valleys. Behind them, lacustrine and uvio-glacial sedimentation occurs, sometimes up to several hundred meters high. Cave outlets connected to the valley bottom become plugged. Some chimney-shas are still developing, allowing phreatic lis from deep passages up to the uplied base level. e height of the chimney-shas corresponds to the height of the base level rise. e Puits des Bans (French Alps), is a 300-m high chimney-sha (Fig. 6).3.4 Base-level rise aer continental subsidenceIn Brazil, the Lagoa Misteriosa is a deep-phreatic sha explored to -220 m by scuba divers (Fig. 6). Regional subsidence (communication from A. A) can be considered a relative base-level rise that has ooded the karst.4. Conclusion Studies of the Messinian-Pliocene eustatic cycle in the Mediterranean allow us to design a model of karst adaptation to major oscillations of base level. Pliocene baselevel rise has ooded the karst and systematically produced phreatic lis chimney shas which feed vauclusian springs. Some cave systems remain ooded, and others have been partly or entirely drained aer Pleistocene reentrenchment of the valleys. Other causes of base-level rise (eustacy, uvial aggradation, continental subsidence), less signicant in amplitude, have the same eect on PAMS. Consequently, there should be a global genetic model for most deep-phreatic systems (Fig. 7). Some of them have a hypogenic origin (e.g., in South Africa, North America, etc.) Figure 7: Distribution and origin of the deepest phreatic cave systems in the world.

PAGE 188

15th International Congress of Speleology Speleogenesis 793 2009 ICS Proceedings (Audra 2007). However, most of them could correspond to a base-level rise inducing the PAMS, which rst ooded the karst and then allowed the development of phreatic lis, chimney-shas, and of vauclusian springs, ReferencesA P. (1997) Les rseaux noys profonds franais et leur origine. 7e  Rencontre doctobre, 27-31 A D.M. and Granger D.E. (2004) A Late Tertiary origin for multilevel caves along the western escarpment of the Cumberland Plateau, Tennessee and Kentucky, established by cosmogenic 26Al and 10Be. Journal of Cave and Karst Studies, 66:2, 46-55 A P. (2007) Karst et splogense pignes, hypognes, recherches appliques et valorisation. Habilitation diriger des recherches, Universit de N ice Sophia-Antipolis. 278  p.  B B. and G D. (1987) Compte-rendu hydrogologique de lopration Splonaute du 2/8/85, Fontaine de Vaucluse. Karstologia, 9, 1-6. B D., P ., Rr J.-L., A J.-P., Ar M.-P., B B., B W. A., C G., f P., Df Y., F N., Iff J., J-M G., L G C., Mf J., S K. Sf J.-P., R J.-Y., and W R. (2005) Un rseau uviatile dge Burdigalien terminal dans le S ud-Est de la France  : remplissage, extension, ge, implications. C. R. Geoscience, 337, 1045-1054. B D., B R., C G., C J.J., D F., Df Y., D C., G F., L S P., Mf P., O J., P O., R J.Y., and Rr J.L. (2005r) Les systmes oligo-miocnes carbonats et clastiques de Basse-Provence. Des tmoins de lvolution godynamiques de la marge provenale et du bassin davant-pays alpin. Livret-guide dexcursion c ommune ASF GDR   Marges Golfe du Lion  B J.-Y. (2002) Conduits ascendants dans les gorges de l Ardche  : les avens Cordier, Rochas et de Nol. 12e  Rencontre doctobre, 15-19 Bf J.-J. and M R. (1988) Le karst du massif des Calanques (Marseille-Cassis). Karstologia, 11-12, 17-24. B B., G E. and R C. (2004) Alimentation et origine de la salinit de lmergence karstique sous-marine de Port Miou, Marseille Cassis Bouches-du-Rhne. C.R. Geosciences, 336, 523-533. B J. and Z O. (2003) Factors controlling exokarst morphology and sediment transport trough caves: comparison of carbonate and salt karst. Acta carsologica, 32:1, 83-99 C H. (2003) Valles et rseaux karstiques de la bordure carbonate sud-cvenole. Relations avec la surrection, le volcanisme et les paloclimats. se, Bordeaux. 675 p. Cavalera T. (2007) tude du fonctionnement et du bassin dalimentation de la source sous-marine de PortMiou (Cassis, Bouches-du-Rhne). Approche multicritres. se, Marseille. 403 p. C G., P J.-M. and G J.-L. (1997) Manifestations karstiques induites par le creusement messinien: exemples rhodano-duranciens. Runion   Gomorphologie quantitative et palogographie dans le domaine karstique mditerranen  La Sainte-Baume, Livret-guide, 33  p. C-G J. (1996) Prhistoire et karst littoral : la grotte Cosquer et les Calanques Marseillaises, Bouches-du-Rhne, France. Karstologia, 27, 27-40 C A. (1992) Les cavits du Comtat venaissin (Vaucluse). Spelunca, 45, 25-32 Df M. and F L.-H. (1993) En plonge sous les Calanques. Port Miou et le Bestouan. Splo, 12, 3-6. F D. C. (1977) Genetic classication of solutional cave systems. 7th International Congress of Speleology, 189-192 G (2001) Compilation danciennes mesures de dbit Port Miou. Apport lhydrogologie de la Provence. 7e Colloque dhydrogologie en pays calcaire et milieu ssur, 157-160. G and A P. (2004) Les lithophages pliocnes de la fontaine de Vaucluse (Vaucluse, France). Un argument pour une phase messinienne dans la

PAGE 189

Speleogenesis 794 2009 ICS Proceedings 15th International Congress of Speleology gense du plus grand karst noy de France. C. R. Geosciences, 336:16, 1481-1489. G D. E., Fr D. and P A. N. (2001) PlioPleistocene incision of the Green River, Kentucky, from radioactive decay of cosmogenic 26Al and 10Be in Mammoth Cave sediments. GSA Bulletin, 113:7, 825-836 H P., G D. E., J P.-Y. and Lauritzen S. E. (2007) Abrupt glacial valley incision at 0.8 Ma dated from cave deposits in Switzerland. Geology, 35:2, 143-146 K J., Hf H., B V., Sr P., D J. F. and G D. (2001) Cenozoic history of the Moravian karst (northern segment): cave sediments and karst morphology. Acta Mus. Moravi, Sci. geol., LXXXVI, 11-160. Mff L. (2007) Les manifestations godynamiques externes et internesde la crise de salinit messinienne sur une plate-forme carbonate prim diterranenne  : le karst de la Basse-Ardche (Moyenne valle du Rhne  ; France). se, Aix-enProvence. 221 p. Mff L., C G., B J.-Y. and B P. (2006) Geodynamic evolution of the perimediterranean karst during the Messinian and the Pliocene: evidence from the Ardche and the Rhne Valley systems canyons, Southern France. Sedimentary Geology, 188-189, 219-233. Mff L., A P., B J.-Y., C G., B E., P O. and M P. (2008) e Messinian Salinity Crisis manifestations on landscape geodynamic (karst surface, river piracies, and cave levels: Example of the Lower Ardche River (Rhne Mid-Valley). Geomorphology [in press] P A. N. (1987) Cave levels and their interpretation. e NSS Bulletin, 49, 50-66 P O., Rr J.-L. and J C. (1997) Architecture et gomtrie des corps tidaux bioclastiques comblant les palovalles miocnes au sud-est du bassin de Carpentras. Publ. ASF, 27, 215-216. R J. (1991) La craie et ses karsts. se dtat, Paris IV. 560 p. R J., M N., L B. and D J.P. (2004) e karstic delta as a morphological consequence of base level variations. Example of a chalk karst system in the Western Paris Basin (Normandy, France). Journes europennes de lAFK Le karst de la craie en Normandie, 64-65. Sf A., C G. and Af J.-P. (2001) Mouvements post-messiniens sur la faille de N mes  : implications pour la sismo-tectonique de la Provence. Bulletin de la Socit Gologique de France, 172  : 6, 697-711.

PAGE 190

15th International Congress of Speleology Speleogenesis 795 2009 ICS Proceedings THE PATTERN OF HYPOGENIC CAVESPHILIPPE AA UDRADRA1, LUDODO VIC MO O CO O CHA A INN2, JEAN AN -YVES BIGOTOT3, and Je E ANCl L Au U De E NO NO BECO O URT RT3 1PolytechN N ice-Sophia, Engineering School of NN ice Sophia AA ntipolis University, 1645 route des Lucioles, 06410 Biot, France2University of AA ix-Marseille, CER R EGE, Europle de lA A rbois, BP 80, 13545 AA ix-en-Proence, Cedex 4, France & Centre de Sdimentologie Palontologie Gologie des systmes carbonats, 13331 Marseille, Cedex 03, France 3French AA ssociation of Karstology e hypogenic cave pattern reects the speleogenetical processes. Processes vary according to the depth in the aquifer, involving mixing corrosion by convergent ux and with meteoric water, cooling, sulfur oxidation, carbon dioxide degassing, and condensation-corrosion. Cave patterns are: isolated geodes, 2D and 3D multistorey following joints and bedding planes, giant phreatic sha, water table mazes, isolated chambers, upwardly dendritic spheres, water table cave, and smoking shas.1. Indroductione development of caves by hypogenic processes (i.e. hypogenic speleogenesis) corresponds to the formation of caves by water that recharges the soluble formation om below, driven by hydrostatic pressure or other sources of energy, independent of recharge om the overlying or immediately adjacent surface (Ford 2006). Hypogenic caves oen referred to as thermal caves or sulfuric acid caves were oen considered as an exotic side of the normal (i.e. meteoric) caves. Palmer (1991) estimated that about 10% caves have hypogenic origin. Recent studies (overview in Klimchouk 2007) have emphasized the specic hydrogeological background and shown that hypogenic caves are much more common than previously thought. e extreme diversity of settings (carbonic, sulfuric, thermal, cold, deep phreatic, shallow phreatic, vadose...) in dierent geological or geomorphological contexts produces a puzzling impression: each hypogenic cave seems to be unique, with few characteristics in common with the other hypogenic caves regarding their pattern.2. MethodA data base of more than 350 hypogenic caves was constructed from the literature, comparing geological structure, hydrology, morphology of caves at dierent scales (wall features, passages morphology, and cave pattern), mineralogy, deposits... Field study of the most representative hypogenic caves, combined with the information in the literature, show that the apparent dissimilarity in shape can be overcome. Taking into account the diverse settings (hydrologic, geologic) and the speleogenetic processes, we obtain a conceptual model of a cave pattern, integrating all kinds of hypogenic caves (Fig. 1) (A 2007). Patterns are subdivided into two main types: deep phreatic systems generally developed in a conned aquifer by transverse speleogenesis (sensu Klimchouk 2000), and cave systems developed above the water table, where condensationcorrosion plays a paramount role.3. Hypogenic Cave Patterns in Phreatic Condition 3.1 Isolated geodesAt depth, mixing allows complex dissolution and deposition processes. Large crystals (calcite, gypsum...) are deposited in slightly saturated water, together with diverse minerals (mainly metallic suldes) (Fig. 2).3.2 3D Multistorey maze cavese rising hypogenic ow uses alternatively joints and bedding planes, producing a 3D maze cave, in a staircase pattern. Generally, the cave has a main trunk where hypogenic ow was rising, surrounded by 3D mazes, smaller in size (Fig. 3). In Monte Cucco Cave system (Italy), the sulfuric water was rising toward the top of the anticline, where impervious covers are breached, allowing the discharge of the karst aquifer. Contiguous vertical passages correspond to discrete hypogenic trunks, inclined galleries follow dip, horizontal passages and some cave entrances record past base level positions (GALDENZI & MENICHETTI 1995). In the Black Hills (South Dakota), Jewel and Wind Caves range among the largest maze cave of the world. ere genesis is complex, involving several early phases (Palmer 2006). However, the pattern resulting from the main speleogenetical phase is simply a dense network of enlarged discontinuities, similar to the previous examples.3.3 2D maze cavesIf an aquitard is present, the cave develops below this impervious ceiling, as a 2D maze cave (Fig. 4). e passages are horizontal or inclined, according to the dip. e Denis Parisis system in the central part of Paris basin is horizontal.

PAGE 191

Speleogenesis 796 2009 ICS Proceedings 15th International Congress of Speleology Figure 1: Conceptual model of the hypogenic cave patterns, according to the geological structure, the groundwater recharge, and the speleogenetic processes. Figure 2: Isolated geodes. Le: geode lined with calcite spar, France. Center-right: cueva de los Cristales (Chihuahua, Mexico) was intersected and drained by the Naica mine (Bernabei et al. 2007). e gypsum swords in this cave are the largest crystals of the world. In Monte Cucco, the Faggeto Tondo develops below the inclined marly cover (cave indicated as no. 2 in Fig. 3, right). e 2D maze cave is a subtype of 3D maze cave; some parts of 3D mazes locally develop as 2D mazes, when a less permeable stratum is present.3.4 Deep phreatic shasIn active tectonic areas, the combination of rising warm

PAGE 192

15th International Congress of Speleology Speleogenesis 797 2009 ICS Proceedings Figure 3: Le: perspective view of the Pigette Caves (France), a 3D multistorey maze cave originating in the phreatic zone by hypogenic upwelling, following structural weaknesses (inclined dip and vertical joints) (sketch by J.Y. Bigot). Right: Monte Cucco system (Italy), probably the deepest hypogene cave system of the world (923 m / 31 km), with large shas of several hundred meters of depth and ramps inclined along the dip of the anticline (Cairoli et al. 1991). water, with CO2 and H2S outgassing concentrates speleogenetic processes along major fault lines, producing the deepest phreatic shas of the world: pozzo del Merro, Italy (-392 m); El Zacaton, Mexico (-329 m, Fig. 5); Hranica propast, Czech republic (-267 m).4. Hypogenic Cave Pattern Along or Above the Water Table 4.1 Upwardly dendritic cavesAbove thermal water, condensation occurs at the ceiling Figure 4: A sub-type of the 3D maze, the 2D maze conned under an inclined impervious roof. Grotte de Saint-Sbastien (France). e dip is towards the SE, water used to well up towards the le. Figure 5: Deep phreatic shas. Zacaton sha is one of the numerous windows in the karst aquifer of Tamolipas area (Mexico), reaching a depth of 329 m. Inestigation showed nearby olcanoes were the source of H2S. Microbial activity based on sulfur oxidation is present (Gary & Sharp 2006).

PAGE 193

Speleogenesis 798 2009 ICS Proceedings 15th International Congress of Speleology which is cooler. CO2 and H2S outgassing enhance aggressivity. By condensation-corrosion, cupolas develop upward as a dendritic pattern of stacked spheres (Audra et al. 2007). e development of two neighboring spheres will be divergent, toward the greatest potential heat transfer, because the rock in between the two spheres has less transfer potential and remains warm (Szunyogh 1990), giving the bush-like structure, as found in the Stork-puszta Cave, Hungary (Fig. 6).4.2 Isolated chambersWhen strong degassing occurs, upwardly dendritic spheres enlarge and join together, eventually producing large isolated chambers (Fig. 7) (Audra et al. 2002). With a moderate thermal gradient and pCO2, modeling shows that such volume can develop in a rather short time span, about 10  000 years (LLISmMONDE 2003). From Israel occurrences, FRUmMKIN & FISCHHENDLER (2005) assign the origin of isolated chambers to phreatic convections.4.3 Water table sulfuric acid caves Above the water table, sulfuric vapors and thermal convections produce strong condensation-corrosion and replacement gypsum crusts (EEgGEmMEIER 1981). e main drain develops headwards from springs (Fig. 8). Due to the sulfuric corrosion the cave has a low gradient (Fig. 9). Minor changes in base level cause the ow to migrate laterally making incipient mazes (g. 9) (AAUDRA 2007). Figure 6: Upwardly dendritic caves. Le: Stork-puszta Cave, Hungary, has been used to represent thermal speleogenesis. Upwardly dendritic spheres develop aboe a basal chamber (survey in Ford & Williams 1989). Center: stacked up spheres in Serpents Cave, France, which is close to Chevalley. Right: Chevalley sha, France (Audra et al. 2007). Figure 7: Isolated chambers. ermal hypogene ow degasses at shallow depth, the thermal conections with CO2 (or H2S) gas enhance strong condensation-corrosion and the development of large isolated chambers. ey tend toward a hemispherical shape. Simultaneously, a massive calcite deposit occurs in the lake, oversaturated aer CO2 degassing; Champignons, France (sketches J.-Y. Bigot).

PAGE 194

15th International Congress of Speleology Speleogenesis 799 2009 ICS Proceedings Condensation domes develop upward and may breach to the surface (Fig. 8). e most demonstrative water table sulfuric caves are Cueva de Villa Luz (Mexico), Chat Cave (France), Kane Caves (USA). Because of major base level lowering, successive horizontal cave levels develop: Frasassi Cave (Italy).4.4 Smoking shas in the vadose zoneAbove thermal aquifers, the rock is signicantly heated by the geothermal gradient. In winter the atmosphere of open shas is unstable: the cold air sinks inside the sha and expels the warm air out of the sha which condenses, giving the impression that the sha is smoking. e air ow follows ceiling channels where condensation-corrosion focuses. Eventually, it produces condensation ceiling cupolas and channels, which could lead to misinterpreting them as phreatic in origin (Vapeur Sha, France; Nasser Schacht, Austria; Fumarollas and Vapor Shas, Spain). e origin of the sha is generally a mechanical fracture; the hypogenic role of the thermal gradient is indirect and limited to the etching of the wall features. Figure 8: Water table Sulfuric cave. Headwards evolution by condensation-corrosion along the water table, supplied with major sulfuric upwelling along a acture. Simultaneously, hydrothermalism lis the hot air, condensation-corrosion occurs, bells and chimneys develop, some nally break through to the surface. e white arrows indicate the direction of cave development (inspired om Cueva de Villa Luz, Mexico). Figure 9: Chat Cave, France. e long prole with very low gradient (0.7%) results om the sulfuric ows. Incipient mazes beside the main drain result om lateral migration of the ow due to minor changes in base level (plan om M. Rousseau, prole Ph. Audra).

PAGE 195

Speleogenesis 800 2009 ICS Proceedings 15th International Congress of Speleology 5. Conclusion e diversity of hypogenic caves is now placed in a global model, explaining all kinds of patterns, depending on the geological structure, the groundwater recharge, and the speleogenetic processes. Beyond hypogenic caves developed at depth by mixing corrosion and rising ow, some hypogenic caves are developing in the atmosphere at -or abovethe water table, mainly by condensation-corrosion, due to the combination of thermal convection, sulfuric and carbonic corrosion.ReferencesAudra P. (2007) Karst et splogense pignes, hypognes, recherches appliques et valorisation. Habilitation esis, University of Nice, 278p. Audra, P., Bigot, J.-Y., and Mocochain, L. (2002) Hypogenic caves in Provence (France). Specic features and sediments. Acta Carsologica, 3, 33. Audra, P., Hobla, F., Bigot, J.-Y., and Nobcourt, J.-Cl. (2007) e role of condensation-corrosion in thermal speleogenesis. Study of a hypogenic suldic cave in Aix-les-Bains. Acta Carsologica, 2, 185. Bernabei, T., Forti, P., and Villasuro, R. (2007) Sails: a new gypsum speleothem from Naica, Chihuahua, Mexico. International Journal of Speleology, 36:1, 23. C, E., C, M., J, K., L, C., G, M., Mf, M., R, R., S, F., Gr, C., G, M., M, M. (1991) Massiccio del Monte Cucco Guida naturalistica ed escursionistica. Centro Nazionale di Speleologia, Costacciaro, 151 p. Ford, D.C., and Williams, P. (1989) Karst geomorphology and hydrology. Chapman & Hall, London, 601 p. E, S.J. (1981) Cavern development by thermal waters. National Speleological Society Bulletin, 43:2, 31. Ford, D.C., (2006) Karst geomorphology, caves and cave deposits: A review of North American contributions during the past half century. In: Perspectives on Karst Geomorphology, Hydrology and Geochemistry. Harmon, R.S., and Wicks, C.W. (Eds.), Geological Society of America, Special Paper, 404, p. 1. F, A., Ff, I., (2005) Morphometry and distribution of isolated caves as a guide for phreatic and conned paleohydrological conditions. Geomorphology, 67:3, 457. G, S., Mf, M. (1995) Occurrence of hypogenic caves in a karst region: examples from central Italy. Environmental Geology, 26, 39. Gary, M.O., and Sharp, J.M. (2006) Volcanogenic Karstication of Sistema Zacatn. Geological Society of America Special Paper, 404, p. 79. Kf, A.B. (2000) Speleogenesis under deep-seated and conned settings. In Speleogenesis. Evolution of karst aquifers. National Speleological Society, Huntsville, p. 244. Kf, A. (2007) Hypogene speleogenesis. Hydrogeological and morphogenetic perspective. National Cave and Karst Research Institute, Special Paper Series, 1, 77. L, B. (2003) Limestone wall retreat in a ceiling cupola controlled by hydrothermal degassing with wall condensation. Speleogenesis and Evolution of Karst Aquifers, 1:4, 3 p. Palmer, A.N. 1991 Origin and morphology of limestone caves. Geological Society of America Buletin, 103: 1. Palmer, A.N. (2007) Cave geology. Cave Books, Dayton, Ohio, 454 p. Szunyogh, G. (1990) eoretical investigation of the development of spheroidal niches of thermal water origin Second approximation. Proceedings of the 10th International Congress of Speleology, Budapest 1989, III, 766.

PAGE 196

15th International Congress of Speleology Speleogenesis 801 2009 ICS Proceedings MORPHOLOGY AND GENESIS OF CAVES IN IRON-RICH ROCKSAA UGUSTO TO S. AA ULER R and LUS B. PIL Instituto do Carste. RR ua Kepler 385/04, Belo Horizonte, MG, 30360-240, Brazil, aauler@terra.com.br Abstract Caves in iron-rich rocks, although mentioned in the Brazilian geological literature since the 19th Century, have been subject to very few detailed studies. At present, nearly 2000 caves have been identied in the two major Brazilian mineral provinces, i.e., the Iron uadrangle area of southeastern Brazil and the Carajs plateau in eastern Amazonia. e large majority of known occurrences in Brazil are developed either in iron ore or in canga, a surcial iron-rich rock composed of varying quantities of detrital fragments cemented by a ferruginous matrix. Caves in iron-rich rocks are usually small, the longest surveyed cave being only 350 m long. Cave patterns comprise a number of irregular chambers connected by smaller passages. Entrances are also small, suggesting that these caves evolved initially as isolated entranceless chambers, with a later connection to the surface and between chambers. Speleogenesis involves the generation of porosity through the leaching of more soluble constituents, such as silica, creating a low-density zone that can slowly be enlarged through the colloidal removal of iron constituents. Following the opening of an entrance, the sloping nature of these caves tends to favor the removal of clastic material through piping processes. In iron-rich rocks the cave-forming process is unequivocally linked to the genesis of the high-grade ore. e management of these caves thus involves a delicate balance between cave preservation and iron mining.

PAGE 197

Speleogenesis 802 2009 ICS Proceedings 15th International Congress of Speleology THE ALKALI SPELEOGENESIS OF RORAIMA SUR CAVE, VENEZUELAHA A ZEL AA. BARTONARTON1, PA A ULA A SUAR AR EZ2, BRR ITTAN TTAN Y MUEN NCH1, JUAN AN GIARR ARR IZZOO1, MAR AR K BRORO ER R INNG1, ER R IC BANAN KS1 and KAT ATHSURR I VEN N KAT ATESWARAN ARAN3.1DDepartment of Biological Sciences, NN orthern Kentucky University, Highland Heights, Ky.2DDepartamento de Biologa de OOrganismos, Universidad Simn Bolvar, Caracas, Venezuela3Biotechnology and Planetary Protection Group, California Institute of TT echnology, Jet Propulsion Laboratory, Pasadena, Calif. It has been known for almost two decades that microbial species can contribute to the formation of caves, although such activities generally occur through the production of acids within carbonate rock. Other processes of biogenic speleogenesis are less easy to explain, such of the formation of Roraima Sur Cave, Roraima Tepui, Venezuela. is 16.4 km long cave, formed within generally insoluble orthoquartzite contains a noticeable microbial population and unusual opal formations. While no current conditions within the cave suggest extreme acidity, we believe that microbial activity is leading to an increase in pH and silica dissolution. Empirical evidence suggests that ammonia is accumulating within the silica rock at sites of primary dissolution, which correlate with high levels of microbial activity within the cave. Numerous bacterial species isolated from the cave have demonstrated nitrogen xation abilities, with a concurrent accumulation of ammonia within the media, suggesting that this is the source of possible alkaline conditions. Although nitrogen xation is an energetically expensive process, we believe that microbial hydrogen oxidation and methanogenesis are generating the energy necessary to drive this process. Together these data suggest a microbially-driven alkaline-speleogenesis in the formation of Roraima Sur Cave.1. Introductione Tepui Mountains of Venezuela are a spectacular range of isolated table mountains formed within the Matau Formation (Briceno et al. 1990). e age of the Matau Formation is still under debate, primarily due to the lack of zircon and other minerals that can be used for radiometric age determination. Nonetheless, it is known that it formed from arenitic sandstones of the Trans-Amazonian Mountains, which were deposited via braided uvial, tidal and wave action into the same ancient basin that formed the Roraima Supergroup. Santos et al. (2003) noted that the Matau Formation is separated from the Roraima Supergroup by a regional unconformity (the Capas de Abarn unit), that may represent the later development of a secondary basin. As a result, Santos et al. (2003) have estimated the age of the Matau to be similar to the Serra Surucucus Formation, at 1552 6 Ma. Subsequent metamorphosis led to the formation of the orthoquartzite massifs that currently represent the Matau Formation, which varies from 200 850 m in thickness. e orthoquartzite nature of the Matau Formation makes it particularly resistant to dissolution, which is evident by the dramatic table mountains such as Auyan-tepui, Chimant and Roraima Tepui, which oen rise over 1000 m above the surrounding eroded terrain (Fig. 1). Yet these mountains are known for the large number of cave systems they contain, with 12 caves exceeding 1 km in length and 16 caves exceeding 200 m in depth (Auler 2003). Recently a cave was discovered in Roraima Tepui, a 32 km2 massif of the Matau Formation, located between the borders of Venezuela, Brazil and Guyana. is cave, originally named Sistema Roraima Sur by Venezuelan speleologists (also known as Sistema Ojos de Cristal), exceeds 16 km in length. During a recent reconnaissance expedition (2005), a high level of microbial activity was found within the cave, much higher than observed in other cave systems, despite the nutrient limitation and lack of soil on Roraima Tepui. We therefore decided to investigate whether there was a link between the high amount of microbial activity and the extreme length of this quartzite cave. Figure 1: e Kukenan Tepui rises 1000 m aboe the Gran Sabana plain in Venezuela.

PAGE 198

15th International Congress of Speleology Speleogenesis 803 2009 ICS Proceedings 2. Materials and MethodsSmall rock samples were collected from the ceiling of the cave in three locations, 30 m, 120 m and 500 m into the cave: Cricket Pool, Red River and Largo Grande, respectively. Water samples were also collected from the stream owing through the cave near each sample location. ese samples were analyzed in the eld for pH, dissolved silica, ammonia and nitrate. Rock samples were analyzed for silica by adding 2 volumes of distilled water (pH 6.9; dissolved silica 0.0 ppm) to 1 volume of crushed rock (vol/vol). e sample was then shaken and the amount of dissolved silica (SiO2) was assayed using a pocket colorimetric assay (Hach Company, Loveland, CO). Samples were also preserved in 4% paraformaldehyde for scanning electron microscopy (SEM). As a control, water and rock samples were collected from outside the cave.3. ResultsFor a long time speleologists have argued whether caves in the tepui mountains of Venezuela represent karst or psuedokarst; an argument based on the relative contributions of dissolution and erosion to the speleogenesis of this system (Wray 2003). To determine whether silica in the cave was being dissolved by chemical or physical weathering we used SEM analysis of quartz grains. e results (Fig. 2) clearly demonstrate the classic etch pits that are indicative of chemical weathering of quartz grains (physical weathering leads polishing of the grains) (Bennett and Siegel 1987). ese data favor the role of dissolution in speleogenesis of Roraima Sur Cave. Due to the stability of the Si-O bond, acidic silica dissolution only occurs under highly acidic conditions (pH 2), yet this same bond is susceptible to nucleophilic attack, allowing dissolution under more moderate alkaline conditions (pH 8.5). Within Roraima Sur Cave we did not nd any geochemistry that could justify the presence of strongly acidic conditions. However, microbial activity can potentially raise local conditions above pH 8.5, by the production of ammonia (NH3) or amines (NH2). Given the large amount of microbial activity in the cave, we wondered if these species were able to locally alter the pH conditions, leading to quartz dissolution. To test this hypothesis, we examined the pH and dissolved silica concentration at each sample site within the cave, as well as control samples from the surface of the tepui (Table 1). As a control, the pH of the stream owing along the oor of the cave near each sample site was also recorded. While the surface stream pH was measured at pH 6.5 on the plateau, before this entered the cave system it owed through a large amount of decaying plant material. is process causes the stream to pick up humic acids and become more acidic (pH 5.5), similar to its pH in Cricket Pool 30 m inside the cave. As this stream continued to ow through the cave, it became more acidic (to a pH 5.0), which also coincides with an increase in dissolved silica content (the saturation index of silica is ~15 ppm under these conditions (Wray 2003)). Nonetheless the pH remains well above the pH necessary for acidic dissolution of silica. No evidence of recent ooding (within the last 100+ years) was seen in the system, primarily through the decomposition state of dead animals in the cave (a near complete quadamundi skeleton was seen ~1 m above base level suggesting that ooding events within the cave are rare). Figure 2: An SEM image of quartz sand grains taken om Roraima Sur Cave. e triangular-shaped pits are indicative of chemical weathering. A cluster of what appear to be bacterial cells is present in the upper-middle. StreamRock Sample pH Dissolved silicapHDissolved silicaSurface6.502 ppm 6.875 ppm Cricket Pool5.564 ppm 7.0172 ppm Red River5.106 ppm 7.1688 ppm Largo Grande4.9715 ppm 7.6864 ppm Table 1: pH and dissoled silica present in surface samples and at each sampling site within Roraima Sur Cave

PAGE 199

Speleogenesis 804 2009 ICS Proceedings 15th International Congress of Speleology While dissolved silica increased slowly in the stream (to 15 ppm), there was a remarkable amount of mobilized silica in the rock (Table 1), especially when compared to surface rocks (5 ppm) and exceeding the SI of silica. Our rudimentary pH measurement also indicates that as the sample sites become further from the entrance, the pH of these samples increases. While this number again does not reach the basic conditions necessary for dissolution, they are within 0.82 pH units of reaching the solubility threshold for silica. To determine if microbial activity was responsible for this dissolution, we cultivated ~500 bacterial species on low nitrogen (5 mM) media containing silica from the three sites. Of these isolates, a large number (83%) belonged to genera known to be either involved in nitrogen xation or nitrogen cycling within the environment (ammonia oxidation, nitrate reduction, etc). When a selection of these species were grown in a media containing amino acids, an AA rthrobacter sp. isolate demonstrated hyperammonia production, signicantly increasing the pH of the surrounding media. erefore, to determine whether ammonia and nitrate played a role in the geochemistry, we measured the ammonia and nitrate levels at each sample site (Table 2). Although the eld measurement techniques were crude, ammonia and nitrate were only found in the cave locations where dissolved silica levels were also high. Due to the nitrogen starved nature of the tepui environment, the ammonia detected is derived from microbial nitrogen xation. 4. DiscussionCurrent theories on the development of caves within silicate rocks are based on arenization, wherein dissolution plays a critical role in the process of speleogenesis by removing the cements between quartz grains (Martini 1979). Erosion of this loosened material by surface streams then leads to the formation of cavities and enlargement to produce the cave systems, although the relative contribution of each to speleogenesis in silicate rocks remains controversial (Aubrecht et al. 2008; Wray 2003). Our results suggest that microbial activity may be involved in the initial dissolution of these silica cements in the orthoquartzites of the Matau Formation. While nitrogen xation by bacteria leads to the generation of ammonia, the high energetic cost of acquiring this substrate means that it would be tightly regulated by the cell. Nonetheless, some nitrogen xing bacteria produce amino acids to balance oxidative stress (Gonzlez-Lpez, et al. 2005) on aromatic carbon sources. e consumption of these amino acids by other bacteria (such as the AA rt hrobacter isolate we obtained) will also cause ammonia to be excreted into the environment, further increasing pH. A more detailed molecular phylogenetic analysis of the microbial community found at each site conrms both the presence of nitrogen xing bacteria and an active nitrogen cycle that includes ammonia oxidizing bacteria and archaea (see Giarrizzo, these Proceedings ). e decrease in pH of the stream as it ows through the system is also of interest. e water on the surface of the tepui has a slightly acidic pH (6.5), which decreases as the water ows through decaying plant matter. is material contains an excess of soluble humic acids produced by the microbial degradation, which are normally removed by the Stream Rock Sample AmmoniaNitrateAmmoniaNitrate Surface0.00 ppm0.00 ppm0.00 ppm0.00 ppm Cricket Pool0.00 ppm0.00 ppm~ 0.25 ppm~ 0.25 ppm Red River0.00 ppm0.00 ppm~ 0.25 ppm~ 0.25 ppm Largo Grande0.00 ppm0.00 ppm~ 0.25 ppm~ 0.25 ppm Table 2: Ammonia and nitrate om surface samples and at each sampling site within Roraima Sur Cave Figure 3: In areas of active stream ow, smoothing of the rock demonstrates the humic-acid dissolution process.

PAGE 200

15th International Congress of Speleology Speleogenesis 805 2009 ICS Proceedings absorptive properties of soil (Kirk 2004). In the absence of soil these humic acids remain in solution and give the water of the tepuis their distinctive amber-brown coloration. Humic acid, due to its multifunctional organic structure, has been shown to chemically erode quartz in groundwater (Bennett and Siegel 1987). Our results suggest that as this humic acid-rich water travels through the cave, it mobilizes silica from the streambed, deprotonating the humic acid and increasing the amount of dissolved silica (Vandevivere et al. 1994). An important role for dissolution by the stream is seen by the smoothing and rounding of the streambed throughout the cave (Fig. 3). In areas of more turbulent ow, an aerosol of these silica-rich organic material is created, which is presumably the source of the opal and chalcedony formations seen in the cave. While the organic etching of these caves by humic acids likely plays a signicant role once speleogenesis has begun, it does not account for the breakdown of the orthoquartzite to the much more friable sands seen throughout Roraima Sur Cave. is breakdown is particularly prominent in areas with active microbial activity (Fig. 4), where it is possible to push a pencil many cm into the rock. ese loose areas still overlay orthoquartzite, with a gradual transition to the harder rock. Due to the instability of the Si-O bond to nucleophilic attack, especially by amines, subaerial microbial weathering can be proceeding at a rate many times greater than humic dissolution (Bennett and Siegel 1987; Hiebert and Bennett 1992; Icenhower and Dove, 2000). While the eect we observed is only sucient to raise the pH of the rock by <1 pH value, microbial species can induce local or ion eects that dramatically increase the microenvironment surrounding the cell (Bennett and Siegel 1987; Hiebert and Bennett 1992; Vandevivere et al. 1994). Given these ndings, we therefore propose an adaption to the arenitization model by (Martini 1979) for the speleogenesis for the tepui caves of Venezuela. In this model (Fig. 5), microbial activity in orthoquartzite beds leads to the production of ammonia and amines. is leads to a local change in pH and dissolution of the silica cements between the quartz grains, increasing the porosity of the deposit. Eventually meteoric water, rich in humic acids, can enter and lead to the formation of an open conduit. Microbial activity continues to degrade the silica cements of the orthoquartzite, weakening the integrity of the rock on the walls and ceiling. Flood events or stoping may then erode these layers, dramatically increasing the size of the cavity. Eventually the cavity becomes large enough that only down-cutting by humic-rich water continues speleogenesis. Whether this hypothesis is correct requires additional examination of the geology, microbial activity and geochemistry of these remarkable cave environments.AcknowledgmentsSampling permission was obtained from the Venezuelan Environmental Ministry and the Vice Ministry of Environmental Management and Administration, Caracas, Venezuela. anks are given to members of the Sociedad Venezolana de Espeleologa (SVE) and Oxford University Caving Club for helpful advice on Roraima Sur Cave and eld maps for research activities.ReferencesAubrecht, R., Lanczos, T., Smida, B., Brewer-Carias, C., Mayoral, F., Schlogl, J., Audy, M., Vlcek, L., Kovacik, L., and Gregor, M. (2008) Venezuelan sandstone caves: a new view on their genesis, hydrogeology and speleothems. Geologia Croatia, 61, 345-362. Auler, A. (2003) uartzite caves of South America. In Encyclopedia of Cave and Karst Science, Gunn, J. (Ed.), Routledge, New York, p. 611-613. Figure 4: In areas of active microbial activity, the orthoquartzite is continually degraded to a loose sand-like consistency.

PAGE 201

Speleogenesis 806 2009 ICS Proceedings 15th International Congress of Speleology Bennett, P.C., and Siegel, D.I. (1987) Increased solubility of quartz in water due to complexing by organic compounds. Nature, 326, 684-686. Briceno, H., Schubert, C., and Paolini, J. (1990) Table mountain geology and supercial geochemistry, Chimanta Masif, Venezuelan Guyana Shield. Journal of South American Earth Sciences, 3, 179-194. Hiebert, F.K., and Bennett, P.C. (1992) Microbial control of silicate weathering in organic-rich ground water. Science, 258, 278-281. Icenhower, J.P., and Dove, P.M. (2000) e dissolution kinetics of amorphous silica into sodium chloride solutions: eects of temperature and ionic strength. Geochimica et Cosmochimica Acta, 64, 4193-4203. Gonzlez-Lpez, J., Rodelas, B., Pozo, C., Salmern-Lpez, V., Martnez-Toledo, M.V., and Salmern, V. (2005) Liberation of amino acids by heterotrophic nitrogen xing bacteria. Amino Acids, 28, 363-367. Kirk, G. (2004) e biogeochemistry of submerged soils. John Wiley and Son, Ltd., Chichester, 291 p. Martini, J.E.J. (1979) Karst in black reef quartzite near Kaapsenhoop, Eastern Transvaal. Annals of the South African Geologic Survey, 13, 115-128. Martini, J. E. J. (2003) Silcate karst. In Encyclopedia of Cave and Karst Science, Gunn, J. (Ed.), Routledge, New York, p. 649-653. Figure 5: Model for the microbially inuenced speleogenesis of Venezuelan tepui caves. Microbial metabolic activity leads to the dissolution of the silica cements of the orthoquartzite. is dissolution leads to a loss of structural integrity of the orthoquartzite and the production of neosandstones and sand. e increase in porosity allows the entry of humic acid-rich surface waters, which chelate silica and lead to cavern enlargement through dissolution. e cave continues to increase in size as the loose material is remoed by erosion and stoping events. In very large systems, humic acid dissolution becomes the primary mechanism of cavern enlargement (adapted om Martini 2003).

PAGE 202

15th International Congress of Speleology Speleogenesis 807 2009 ICS Proceedings Vandevivere, P., Welch, S.A., Ullman, W.J., and Kirchman, D.L. (1994) Enchanced dissolution of silicate minerals by bacteria at near-neutral pH. Microbial Ecology, 27, 241-251. Wray, R.A.L. (2003) uartzite dissolution: karst or pseudokarst? Speleogenesis and Evolution of Karst Aquifers, 2, 1-9.

PAGE 203

Speleogenesis 808 2009 ICS Proceedings 15th International Congress of Speleology RELATIONS BETWEEN SPELEOGENESIS AND SURFACE MORPHOGENESIS OF AN EXHUMED KARST PLAIN (THE SLUNJ KARST PLAIN, DINARIC KARST, CROATIA)NN EVEN N BOO ID Department of Geography, Faculty of Science, University of Zagreb, Maruliev trg 19,10000 Zagreb, Croatia e Slunj karst plain is developed in Mesozoic carbonate rocks. In some places they are covered with Neogene clastic lacustrine sediments, usually occurring as denudational remnants. Initially, Neogene sediments covered most of the plain and a well developed drainage system formed. Aer erosion of the Neogene sediments, the Mesozoic carbonate rocks were exhumed and karstication led to disorganization of the surface drainage. In such conditions many ponors and associated caves developed. At rst cave genesis was mainly lateral, as phreatic tubes and/or as vadose canyons. Change of the stress direction in the neotectonic phase led to the genesis of pull-apart basins and hills with a pop-up structure. Because of the neotectonic upli of pop-up structures, some lateral cave channels were abandoned and new vertical shas started to develop. ese shas represent the youngest speleogenesis phase. is paper compares the development of the underground and surface features.1. Introduction, Geologic and Geomorphologic Settingse research area is situated in the framework of the Dinaric mountain system next to the Pannonian Basin. It has an area of 336 km2, and is a part of a larger area called the Karlovac karst plain, which extends to Slovenia in the south-west and to Bosnia and Herzegovina in the southeast (Fig. 1). e largest part of the area is made of more or less permeable, karstied, carbonate deposits of Mesozoic age (from Middle Triassic dolomites to Upper Cretaceous limestone). Only a small part of the research area is made of the oldest rocks, Permian sandstones (Korolija et al. 1979, 1981; Polak et al. 1976 1981). Miocene lacustrine deposits are preserved only in tectonic downwarps or as smaller denudation remnants on the plain. In a structural sense the area represents a series of structures and faults of mainly Dinaric direction, but with a marked bending of fault lines caused by a change in stress orientation. e result is that there are sequences of pop-up and pull-apart structures in the plain. e lowest height above sea-level is 196 m, and the highest is 658 m. e average altitude of the area is 353 m. A change in the stress direction has caused a prominent right horizontal component of the fault movement, as well as the development of most of the secondary morphostructures, because of the change to a new tectonic regime. Recent active faults have a signicant eect on the landscape forming steep steps and escarpments, elbow-like valley bends and linear sections of active and dry valleys. Karst and uviokarst morphogenetic types are the most represented exogenous processes and forms. e most signicant surface karst forms are dolines, grikes, residual hills and shallow uvalas. Areal and contact uviokarst processes and forms have been found. Dry ponor valleys and active canyon sections of the Korana and Slunjica River valleys are especially prominent among them. e karst plain is the most spacious and probably oldest relief form of this area. Fluvial and uvio-denudational processes highly inuenced the karstication processes of the Slunj Plain, because the Korana River represents the erosion base-level for a large part of this area. Many scientists have written about the relief development in the Dinaric karst plains: e.g., Cviji 1921; Rogli 1951, 1957; Gams 1986, 2001; Herak 1986; Bahun (1990); Mihevc (2007) and others. Garai (1984, 1991a) worked out the speleogenetic problems of this area several times; Boi (2003a, 2003b), Boi et al.. (2003), Boi and Baurin (2004) and Boi (2009) researched the relation between speleomorphology and the surface landscape. Data about caves are collected from eld explorations and the published and unpublished results of many speleological explorations of this area (e.g., Poljak 1914; SDH 1965; epelak, R. 1965; epelak, M. 1983; Boi 1973; Garai 1984, 1987; Jelini 1998; Kuhta 2001; Baurin et al. 2004; Boi & Baurin 2006;2. Speleomorphologic CharacteristicsData from about 103 caves of the Slunj Plain have been used in this study. Out of 101 caves (data are not available for two of them), 75 are horizontal and 26 are pits. A great part of the horizontal caves (almost 75%) support the fact that lateral circulation of the underground water played a dominant role in the speleogenesis of the known caves, while gravitational recharge through the vadose zone had

PAGE 204

15th International Congress of Speleology Speleogenesis 809 2009 ICS Proceedings a dominant role for a smaller number of caves (pits, about 25%). In the Slunj Plain most caves are simple (69 %), while branchwork caves (20%) are also signicant. Only 2% are caves with distinct levels, 7% are knee-shaped with alternating horizontal and vertical sections, and 2% are cave systems composed of two or more connected caves. Consequently, in the Slunj Plain simple caves dominate, most probably they are parts of larger active and fossil underground karst conduits (White and White 2003). A preponderance of one morphologic type points to relatively homogenous speleogenetic conditions, which is dierent from the representation of morphologic types in the whole karst area of Croatia (Garai 1991), where heterogeneity of speleogenetic conditions is more prominent, as well as a greater representation of the kneelike type connected chiey with pits. e cave system Panjkov ponorVariakova, 12,385 m long, is the longest in the study area. At the same time it is the second longest cave in Croatia (aer the cave system ulin ponor-Medvedica, 16,396 m long). Except for this system, there are two more caves longer than 2,000 m in the study area, then two caves longer than 1,000 m, six longer than 500 m, 17 caves longer than 100 m, while other caves are shorter than 100 m (51 caves). By considering of length of 79 caves (there are no information for other caves) the total length of all known caves of the Slunj Plain is 28,906.5 m. e average length of a particular cave is 365.9 m, consequently, the density of the cave passages is 85.9 m/km2. If we take into consideration only the area of developed karst and uviokarst relief (278.4 km2), the density of the known cave passages is Figure 1: Location of the research area.

PAGE 205

Speleogenesis 810 2009 ICS Proceedings 15th International Congress of Speleology 103.8 m/km2. e average length of a single cave, calculated without the length of the Panjkov ponor-Variakova cave (because its length signicantly stands out from the other caves length) is 211.8 m, which points to a high degree of speleogenesis. In the continuation of the Slunj Plain in Bosnia and Herzegovina, at the edge of the study area, there are some more caves. ey are: the cave Runica, 1,052 m long, then the cave Gatica (1,195 m), and the cave arieva, 1,378 m long (Garai 1991b). Data about those caves were not considered in this analysis. e deepest caves in the study area are: Volovska jama (129 m deep), Kojina jama (109 m) and Kunina jama (103 m), then Barieva pilja (102 m). ey are, at the same time, the only caves in the Slunj Plain deeper than 100 m. Out of the available data about the depth of 29 caves in this area (mainly pits) the average depth of 38 m was calculated. ese data point to the fact that the depth of the caves in the Slunj Plain is relatively small in relation to the depths in other parts of the Croatian karst (the deepest is the pit system Lukina jama-Trojama, 1,392 m deep). However, compared with the other parts of the shallow, so called Kordun karst, the Slunj Plain has somewhat deeper caves, so the Volovska jama is the deepest known cave in the Kordun karst. e entrances of the caves range in altitude from 210 to 550 m high (Fig. 2). e relationship between the individual caves and their heights above sea level is fairly linear, except for a somewhat smaller number of caves between 400 and 450 m. More signicant dierences are observed when their depths are plotted (i.e., the heights above sea-level of the deepest points). It appears that greater depths are characteristic of the caves at higher elevations, and the elevations of the deep points increase with the heights of entrances. Certain conclusions about this areas speleogenesis can be made on the basis of morphologic data and eld studies. Speleogenesis developed at several stages. At the rst stage it was conditioned by lateral movement of the underground water. Channels developed in phreatic tubes and shallow vadose canyon passages, and speleogenesis developed mostly in length. Neotectonic upli played a very important role. It Figure 2: Plot of cave entrance elevations (1) sorted by elevation aboe sea level (dots), and (2) elevations of lowest points in caves (rectangles), plotted directly beneath their corresponding entrance points. Most caves are too shallow for their deep points to be distinguished on the graph.

PAGE 206

15th International Congress of Speleology Speleogenesis 811 2009 ICS Proceedings caused the emergence of increasingly low and young levels, so the phreatic channels, which had found themselves in the vadose zone, cut keyhole-shaped cross sections. In the uplied zones lateral circulation was changed by vertical circulation of the underground water, which created relatively younger vertical vadose channels.3. Discussion Relations Between Speleogenesis and Surface MorphogenesisIn the Mesozoic, carbonate platform sedimentation dominated in this area. At the end of the Cretaceous, this sedimentation ceased, and the area was eroded to a plain. By the end of the Cretaceous and at the beginning of the Palaeogene, tectonic movement occurred during the Laramian orogenetic phase. Under its inuence particular parts of the carbonate platform were uplied exposing them to exogenous processes. ose processes could have lasted until the Middle Eocene. During the Pyrenean orogenic stage (Upper Eocene Lower Oligocene), Cretaceous carbonate rocks were thrust over Eocene ysch deposits. ose processes probably caused a partial destruction of the past exogenous activity, but they also caused new strong exogenous processes, which probably lasted until the Middle Miocene. In that period of approximately 20 million years there was intense karst denudation. e karst plain could have developed in that period. e occurrence of Neogene bauxites in other parts of the plain also point to karst denudation (inkovec et al. 1985). One can presume that it was one of the nal stages of the karst plain formation, which was aerward, according to Bahun (1990), for the most part covered by the Middle Miocene lacustrine clastics. Middle Miocene clastics were deposited during a transgression over the karsted bedrock (Kranjec and Prelogovi 1974). During Middle Miocene lacustrine sedimentation, exogenous processes were interrupted. Aer repeated upli, a dense uvial network was developed on the lacustrine clastics. at network determined uviodenudation and uvial characteristics of that period landscape. Primarily erosion and slope processes dominated in that area, and they led to thinning, and partially to complete denudation of the Middle Miocene clastics. at exposed the carbonate bedrock, so karstication could be activated again. During that process a network of valleys also developed on the carbonate bedrock. Gradually, the valleys on carbonates lost their ow to caves. In that way the post-Miocene speleogenetic processes were activated. Emergence of underground space was probably connected with earlier karstication stages, but such traces have not been discovered yet. e border between karst and nonkarst areas migrated retroactively because of denudation, so the karst area grew with time. During the Pliocene there was again a shorter lacustrine stage. Until the end of the Pliocene, speleogenesis developed mainly in the phreatic zone (because of its high position near the land surface). Cave channels most probably started to develop laterally, along the privileged directions of the underground owing. Input of water into the underground was primarily determined by the position of the carbonate-caprock contact. Channels developed in phreatic and shallow vadose conditions (canyon channels) and speleogenesis proceeded mostly longitudinally. e underground waters, as well as speleogenesis, were directed toward the intermediate layer and tectonic fractures, especially faults. e cave Kojina jama is among the most prominent examples. Its position and interior morphology point to the dominant impact of faults on cave genesis. Neotectonic movements were intensively revived by the end of the Pliocene (Prelogovi 1975; Veli et al.. 1982), and they were accompanied by the stress orientation change from the direction northwestsoutheasst to the north south. ere was the activation of older faults of the Dinaric orientation, but with a prevailing right horizontal shi component. e stress orientation change caused bending of the fault routes and emergence of local compression and extension regimes. ey led to the creation of pop-up and pull-apart structures, i. e., forming positive and negative morphostructures in the framework of the plain. Besides the bending of the longitudinal fault lines, there was also activation of diagonal faults and structure rotation. Neotectonic upli caused an intensied denudation of positive morphostructures, and neotectonic lowering protected the basin bottoms from denudation, so there are now well-preserved remnants of the Miocene clastics with a developed surface drainage network (e. g. Krlja basin). Tectonic lowering also caused cutting of favorable underground water ows, so springs developed (Bahun and Fritz 1987), and ponors could develop in basins (e. g., Panjkova and Variakova caves in the Krlja basin). Because of exhumation, more and more plain area was exposed to karstication, with sinking streams and the emergence of more dry inactive valleys, which were transformed into relict valleys owing to karstication. Upli of particular blocks caused development of more cave levels. Except in the phreatic and subphreatic zones, more vadose cave development occurred, leading to canyons and keyholeshaped cross sections. Neotectonic upli of the whole plain, probably during the Pleistocene, caused the Korana River ow to downcut 50 m below the surrounding plain. In that way the relief of the Slunj Plain was nally formed. It is primarily characterized by a planated surface with dolines and a fragmented network of dry and relict valleys with

PAGE 207

Speleogenesis 812 2009 ICS Proceedings 15th International Congress of Speleology residual hills. Streams and uvial erosion were preserved on the down-dropped morphostructures (e.g., Krljanska basin).On the uplied morphostructures, karst denudation intensied, and the underground ow was gravitationally directed into the vadose zone. As a consequence, the development of the youngest underground channels are vertical. ey cut the older phreatic channels, which became uplied and inactive. New channels developed, mainly along younger ssure zones, which had developed as the consequence of the stress change, i. e., adaptation of geologic structures to a new stress orientation. e old passages were lled with speleothems and undissolved residual material (e. g., at Kunina jama). In the Mavina region many relatively deep pits formed.. Several generations of huge collapsed dripstones in the Kojina jama are the result of tectonic movement. Lowering of the erosional base-level produced dolines and shallow pits in the plain. e terrain, lowering under the inuence of karst denudation, cut into caves, which created secondary entrances (e. g., Kojina jama, Jama Davorinka) and partially or completely destroyed others. e nal destruction of some caves has been noticed in the Mavina region (e. g. the caves RI-M2, RI-M3). In the Slunj Plain there has recently been neotectonic upli and intensive karst denudation, which causes seasonal sinking of the main river in this area, the Korana River, in its upper reaches. Parallel with karst denudation on the surface, complex speleogenetic processes are occurring underground.ReferencesBahun, S. (1990) Stupnjevi razvoja zaravni u Dinarskom kru. Kr Jugoslavije, Zagreb, 12:6, 147-158. Bahun, S., i Fritz, F. (1987) Postanak izvora u dinarskom akumuliranom orogenskom kru. Kr Jugoslavije, 12:2, 27-37. Boi, N. (2003a) Bacis morphogenetic characteristics of caves in the Grabovac valley (Slunj, Croatia). Geoadria, 8:1, 5-16. Boi, N. (2003b) Relation between karst and uviokarst relief on the Slunj plain (Croatia). Acta Carsologica, 32:2, 137-146. Boi, N., Baurin, ., i Mihali, A. (2003) Speleoloki objekti na podruju brda Mavina. Speleozin., 16, 17-21. Boi, N., i Baurin, (2004) Geomorphological Conditions of the Genesis of the Ponor Jovac Cave (Croatia). Acta Carsologica, 33:2, 107-113. Baurin, ., Boi, N., i Bala, Z. (2004) Ponor pod Kremenom i Barieve pilje kod Slunja. Speleozin, 17, 15-20. Boi, N., i Baurin, (2006) Ponor Jovac kod Slunja. Speleozin, 18. Boi, N. (2009) Geomorfoloke znaajke prostora Slunjske zaravni. Doctoral thesis, University of Zagreb, 270 p. Boi, V. (1973) Baraeve spilje nekada i danas. Speleolog, 20-21, 2-6. Cviji, J. (1921) Abrazione i uvijalne povri. Glasnik geografskog drutva, 6, 1-61. epelak, R. (1965) Barieva peina. Nae planine, 1965, br. 9-10, 199-202.epelak, M.(1983) piljski sustav Panjkov ponor Krlje. Speleolog, 30-31, 21-27. Gams, I. (1986) Kontaktni uviokras. Acta Carslologica, 14-15, 71-78. Gams, I. (2001) Notion and forms of contact karst. Acta Carslologica, 30:2, 33-46 Garai, M. (1984) Neotektonske aktivnosti kao jedan od uzroka geneze i morfologija jednog od najveih spiljskih sistema u Hrvatskoj. Deveti jugoslavenski speleoloki kongres, Karlovac, Zagreb, p. 457-466. Garai, M. (1987) Novi sifon u sistemu Mukinje i Panjkove spilje na Kordunu. Bilten SSJ, Sarajevo, 3:2, 19. Garai, M. (1991) Morphological and hydrogeological classications of speleological structures (caves and pits) in the Croatian karst area. Geoloki vjesnik 44, Zagreb, p. 289-300. Garai, M. (1991a) Karstikacija spiljskog kanala iza Zelenog sifona i njegova hidrogeoloka uloga u spiljskom sustavu Mukinje i Panjkove pilje na Kordunu. Spelaeologia Croatica, Zagreb, 2, 5-14. Garai, M. (1991b) Hidrogeoloka funkcija arieve pilje na Kordunu. Spelaeologia Croatica, Zagreb, 2, 2329.Herak, M. (1986) Geotektonski okvir zaravni u kru. Acta Carsologica, 14-15, 11-15.

PAGE 208

15th International Congress of Speleology Speleogenesis 813 2009 ICS Proceedings Jelini, I. (1998) Sustav Mateieva spilja Popovaka spilja. Speleozin, 8-9, 3-5. Korolija, B., ivaljevi, T., i imuni, An. (1979) Osnovna geoloka karta 1 : 100 000 list Slunj. SGZ, Beograd IGI, Zagreb Korolija, B., ivaljevi, T., i imuni, An. (1981) Tuma Osnovne geoloke karte 1 : 100 000 za list Slunj. 48 str., SGZ, Beograd IGI, Zagreb. Kranjec, V., Prelogovi, E. (1974) O paleogeografskim i neotektonskim odnosima u tercijaru i kvartaru na teritoriju SR Hrvatske. Geoloki vjesnik, Zagreb, 27, 95-112. Kuhta, M. (2001) Speleoronilaka istraivanja izvora Slunjice. Speleolog, Zagreb, 46-47, 1998-99, 30-34. Mihevc, A. (2007) e age of karst relief in west Slovenia. Acta Carsologica, 36:1, 35-44. Polak, A., Juria, M., parica, M., i imuni, A. (1976) Osnovna geoloka karta 1 : 100 000 list Biha. SGZ, Beograd IGI, Zagreb. Polak, A., Crnko, J., imuni, An., imuni, Al., parica, M., i Juria, M. (1981) Tuma Osnovne geoloke karte 1 : 100 000 za list Biha. 52 str., SGZ, Beograd IGI, Zagreb. Poljak, J. (1914) Peine hrvatskog kra II dio. Peine okolia Plitvikih jezera, Drenika i Rakovice. Prirodoslovna istraivanja JAZU, Zagreb, 1-25. Prelogovi, E. (1975) Neotektonska karta SR Hrvatske. Geoloki vjesnik, Zagreb, 28, 97-108. Rogli, J. (1951) Unsko-koranska zaravan i Plitvika jezera geomorfoloka promatranja. Geografski glasnik, Zagreb, 13, 49-66. Rogli, J. (1957) Zaravni u vapnencima. Geografski glasnik, Zagreb, 19, 103-134. SDH (1961) Speleoloki objekti na karti 1 : 50 000 Gospi 2, Speleoloko drutvo Hrvatske, Zagreb. inkovec, B., unjara, A., i Saka, K. (1985) Boksiti Korduna i susjednih podruja. Geoloki vjesnik, Zagreb, 38, 215-233. Veli, I., Soka, B., Galovi, I. (1979) Tektonsko i paleogeografsko znaenje novih nalaza senonskih vapnenaca i eocenskog ia u Kordunu (sredinja Hrvatska). Geol. Vjesnik, Zagreb, 31, 191-202, White, W.B., and White, E.L. (2003) Conduit fragmentation, cave patterns, and the localization of karst ground water basins: the Appalachians as a test case. Speleogenesis and Evolution of Karst Aquifers, 2:1, 1-15, www.speleogenesis.info

PAGE 209

Speleogenesis 814 2009 ICS Proceedings 15th International Congress of Speleology How OW can CAN ghost GHOST rocks ROCKS help HELP in IN karst KARST development DEVELOPMENT ?LAu U Re E NT BRuxelles UXELLES 1, Yves VES Qui QUI Nif IF2  and Michel ICHEL Wi I Ni I N3 1 INRANRA P (NN ational Institute of Preventive AA rchaeology), 561 RR ue Etienne Lenoir, KM DDelta, 30900 NN mes, France ; UMR R 5608 du CNR NR S CR R PPM/TRATRA CES2 GFA A Facult Polytechnique de Mons, 9 rue de Houdain, B-7000 Mons, Belgium 3 Parc NN ational des Cvennes et SCSP, GrandR R ue 30360 Vezenobres, France For several years, numerous examples of ghost rocks have been described in Belgium and in France. ey correspond to decalcied pockets but also include big weathered networks of so alterite. is phenomenon is present in several dierent geological and geomorphic contexts. ese features illustrate one step in the geological history of the area, and their genesis implies certain geological, hydrological and morphological conditions. Exploration of mines allows us to study many examples of ghost rocks, linked or not with mineralization. At the top of the alterite we have found a thick level of laminated sediments that shows the existence of previous water circulation in the gap between the ghost rock and the vaulted roof. In some places the alterite has been progressively removed by seepage, to form little rooms. In one mine, large karstic galleries have reused old weathered sections and eroded the initial ghost rocks. ese examples show how the formation of ghost rocks prepares some carbonate plateaus for more classical karstication by forming gaps that can be used by underground water ow. Hydrogeological systems can expand quickly and extend their catchment areas by utilizing these discontinuities.   1. IntroductionResearches of Belgian geologists have shown the existence of a special karstication: the ghost rock (Vergari and uinif 1997; Vergari 1998). Many examples are now well known all around the world (Schmidt 1974; Huselmann and Tognini 2005; Martini 1985; Tognini 2001; Willems 2002), in dierent rocks, carbonates or not, and in several geological and geomorphic contexts. We have also found and studied some of them in France (Bruxelles 2004). In this paper, we want to describe dierent examples and to show some links between ghost rocks and classical karstication.  2. Ghost Rock FormationGhost rocks consist of so and clayey materials situated within the rock. ey can appear in dierent ways. Sometimes they involve large surfaces, like those in the Hainaut Province (Belgium), but they can occupy pockets, big and small, or in corridors between two walls of competent rock. At rst glance they look like classical karstic llings with allochthonous sediments. But if we take a closer look, we note that we can follow the original stratication of the rock from one wall to the other, through the so material (Fig. 1). Lithologic features, such as fossils, joints, and sylolites, can be recognized inside this material. Ghost rock is an isovolumic alteration that occurs by disappearance of soluble components and conservation in situ of less-soluble materials (sparitic calcite, siliceous framework, dolomite, etc.). In addition, some clays have formed in situ by neogenesis. Contrary to classic karst, the volume initially created by dissolution is distributed as pores in the whole of the alterite (uinif 1999). It is formed by very slow percolation of water, which penetrates into the rock by fractures and carries the aggressiveness to depth. As the drainage is very slow, there is no possibility of exporting the residual material. is change takes place from the surface, per descendum, in a context of weak hydraulic gradient and extensive fracturing. In this case, ghost Figure 1: Ghost rock cut by a quarry in Belgium. Horizontal stylolites can be followed om the safe limestone trough the alterite.

PAGE 210

15th International Congress of Speleology Speleogenesis 815 2009 ICS Proceedings rocks are formed in a context of low relief and low erosive energy. Recently, ghost rocks have also been discovered in association with mineralizations in southern France (Bruxelles and Wienin, in press) and undoubtedly have a hypogenic origin. e weathered parts are organized into a network of corridors from tens to several hundred meters long. ey can coalesce and, in the case of big ghosts, form some huge volumes of alterite that can reach more than one hundred meters below the surface. At further depth, this type of structure extends into pseudo-endokarst, with features that look like lled galleries beneath a competent calcareous roof (Fig. 2). Nevertheless, this conduit was never empty and the lling consists of the alterite (Vergari 1998). e origin and development of the pseudo-endokarsts are still poorly understood. ey imply reactions of oxidation-reduction (suldes, organic matter), migration of silica and carbonates, and also maybe bacterial action. e pseudo-endokarst may achieve complex forms, oen labyrinthine, guided over large distances by faults and joints. erefore, we can distinguish several dierent kinds of ghost rock (Fig. 2): (a) large areas of ghost rocks, above which depressions are formed by compaction of the alterite; (b) farther down, many pockets along discontinuities and fractures; and (c) in some places, ghost rocks beneath a competent roof: the pseudo-endokarst..  3. Some Examples of Underground Ghost RocksOne of the well known examples of pseudo-endokarst, the Pic Glace, was described in Belgium by Anne Vergari (1998). is example is surprising because it looks like a classic cave section, some kind of meander, with various levels of detrital ll. We can even recognize the wall features, with projecting shelves that could represent dierent stages of excavation by an underground stream. But study reveals that, in fact, the lling is not allochtonous sediment. e best demonstration is that we can follow several chert beds from one wall to the other across the so material. is is an example of ghost rock formed under a limestone roof. is cave had never been empty, and there was never an underground river in this cave. e wall features correspond to dierential weathering of the limestone. e geometry of the alterite is also interesting. e dierent levels show a concavity more and more pronounced from the bottom toward the roof of the section. It is due to the removal of part of the limestone by weathering, which provoked the compaction of the alterite.  In France, we also have many examples of ghost rocks (Rodet 1996; Courrges 1997 and Bruxelles 2004). Some of them in the Grands Causse are spectacular. ey are developed in limestone with chert, but also in dolomite. is weathering has produced clays with int or dolomitic sand that cover part of the plateau and contribute to cryptocorrosion (Fig. 3). In the Cvennes (south of the French Massif Central), there are many mines in lead, zinc, and iron ores. ey are excavated in dolomite and limestone but, in many areas, they have intersected what was once thought to be karstic llings (Fig. 4). But we can recognize the stratigraphy of the initial rocks by the presence of sandy levels with chert fragments from the dolomite, and clayey zones from the limestone. We can even see some mineralized veins, initially formed in the limestone, and now preserved in the alterite (Fig. 4A). Figure 2: Schematic diagram showing the dierent types of ghost rocks. 1 competent rock; 2 large area of ghost rocks, with overlying oids formed by compaction of the alterite; 3 pockets and corridors developed along discontinuities and actures; 4 ghost rocks under a secure roof : the pseudo-endokarst. Figure 3: Section in the Bajocian cherty limestone in the Grands Causses (France). We can recognize pockets of clay with chert (dark grey formation), but also pseudo-endokarst under a roof of competent limestone.

PAGE 211

Speleogenesis 816 2009 ICS Proceedings 15th International Congress of Speleology 4. Links Between Ghost Rocks and KarsticationWe have seen before that the dierent levels of alterite within the pseudo-endokarst show an increasing concavity from the bottom toward the roof of the section. is is a very important point because this process creates a space between the ghost rock and the roof that can be used later by a stream of water. is is the case in the pseudo-endokarst of the Pic Glace in Belgium, and also in the mine we have studied in the Cvennes (La Grande Vernissire, Fressac; Bruxelles and Wienin, in press). In the eastern gallery of this mine, between the ghost rock and the roof, we have identied 10-30 cm of laminated clays (Fig. 4). ere is an erosional unconformity between the alterite and the laminated clays, and we can clearly see sedimentary features that indicate the circulation of water. e void between the alterite and the roof was used by an underground stream. e top of the ghost rocks was rst eroded before deposition of the clayey layers. At the roof, we notice some karstic features that can be linked to this process (Fig. 4B). So, by this example, we can see that compaction creates a discontinuity which can be used by classic karstication.  In the same mine, in several places, the articial gallery has cut the bottom of a pseudo-endokarst. is opening permits removal of the alterite by creating an exit for dripping water (Fig. 5). Since the excavation of the mine about 20 years ago, a void nearly 5 m high has formed and continues to grow each year. So alterite is very susceptible to erosion, and the appearance of a new hydraulic gradient, dierent from the one that led the weathering, provoked the hollowing-out of the pseudo-endokarst. Figure 4: Section of the East gallery of the mine of La Grande Vernissire (South of France) showing the link between weathering and classic karstication. A: Fluorite mineralization in the ghost rock; B: Classic karst feature. Figure 5: Schematic section of a vertical pocket. Ghost rock has partially eroded since the opening of the mine.

PAGE 212

15th International Congress of Speleology Speleogenesis 817 2009 ICS Proceedings In Belgian quarries we can observe the same phenomenon on another scale. During excavation of the limestone, much pseudo-endokarst has been intersected. As the quarries are situated below base level, many springs have appeared on top of the alterite. With time, these springs became larger and began to erode the ghost rocks. Caves could form within a few weeks, some of them of explorable size. ey look like big meanders and become larger each day because of the erosion of the alterite by the new underground streams. Another important point is that if we draw a map of the pseudo-endokarst, we can see that they constitute a real network with several ghost rocks interconnected. In fact, all the tensional fractures present during the weathering were aected (generally in two directions). As in the examples of Belgian quarries, during the upli of a region or the entrenchment of a valley, the hydraulic gradient becomes steeper. e presence of pseudo-endokarsts introduces an important discontinuity that is exploited preferentially by underground streams. From the spring, headward erosion permits the emptying of the alterite and the quick extent of the catchment area inside the plateau (Fig. 6). It can be responsible of the formation of some labyrinthine caves, with a high density of big galleries, like the Trabuc Cave (Cvennes, France).  On the French Grands Causses Plateau, entrenchment of canyons has permitted the appearance of some powerful karst springs. In the weathered dolomite, headward erosion removes the dolomitic sand and empties the highest joints. By this way, the catchment areas of springs can grow quickly throughout the plateau. On the surface, many pits appear and lead to the removal of the alterite by underground streams (Fig. 5). At this time, many sinkholes open regularly. e catchment area of the Durzon spring, for example, is capturing the catchment area of other springs that have drained this part of the plateau for a long time (Bruxelles 2004). is process is leading to destruction of old horizontal morphologies and development of the most typical landscape of the Grands Causses: the tower-like mega-lapiaz.  5. ConclusionGhost rocks are a special case of karstication formed per descendum, but also per ascensum in association with hypogenic circulation. Instead of forming a single void, they consist of porous alterite that retains the original shape and volume of the original rock. Pseudo-endokarsts can form an interconnected network that promotes the development of the future karst. When a new hydraulic gradient appears, springs are formed and permit the erosion of the alterite. Parts of the weathered networks are progressively emptied. Beneting from the existence of these discontinuities, the catchment areas of these springs can expand quickly. Re-use of pseudo-endokarst by underground streams result in classic karst morphologies, provokes drainage reorganization, and introduces, for the rst time, allochtonous sediment. is process of cave genesis should make us have another look at some of our classical cavities. It can also provide new interpretations concerning karstic morphologies and the functioning of underground streams. Furthermore, the Figure 6: Schematic diagram of evolution om a pseudo-endokarst to a classic karst. A Pseudo-endokarsts are interconnected along the tectonic network exploited by the weathering. ere is not enough energy to evacuate the alterite; B Upli or valley entrenchment increases the hydraulic gradient. e underground drainage exploits the ghosts and remoes the alterite by headward erosion; C e catchment area of the spring spreads quickly and allows a network of galleries to appear.

PAGE 213

Speleogenesis 818 2009 ICS Proceedings 15th International Congress of Speleology discovery of ghost rocks in a karstic area can help assess the risk of collapse and protect groundwater resources.Acknowledgments e authors would like to thank Art Palmer for the reference of Schmidt 1974 about phantom passages in the USA. Many thanks also to Mike Komaransky for the proofreading of this paper.ReferencesBruxelles, L. (2004) Dpts et altrites des plateaux du Larzac central : causses de lHospitalet et de Campestre (Aveyron, Gard, Hrault). Evolution morphogntiques, consquences gologiques et implications pour lamnagement. Document du BRGM, Orlans, 2004, 304, 286 p. + 5 cartes couleur. Bruxelles, L., and Wienin, M. (in press) Les fantmes de la mine de la Grande Vernissire (Fressac, Gard). Premires observations sur lorigine de certains karsts de la bordure cvenole. Proceedings of the International Conference: le karst, indicateur performant des environnements passs et actuels, Arette, 09-2008. Karstologia, 10 p. Courrges, M. (1997) Le crypto-karst de la pninsule du Mdoc. Crypto-altration, dissolution, karst sousmarin et volution quaternaire. uaternaire, 8:2-3, 289-304. Huselmann, P., and Tognini, P. (2005) Kaltbach cave (Siebenhengste, Switzerland): phantom of the sandstone? AA ct a Carstologica, 34:2, p. 383-396. Martini, J. (1985) Caves of South Africa. Karstologia, 5, 39-44. uinif, Y. (1999) Fantmisation, cryptoaltration et altration sur roche nue, le triptyque de la karstication. Proceedings of the European Conference Karst-99, p. 159-164. Rodet, J. (1996) Une nouvelle organisation gomtrique du drainage karstique des craies : le labyrinthe daltration, lexemple de la grotte de la Mansonnire (Bellou-sur-Huisne, Orne, France). C. R. Acad. Sc., 322:IIa, 1039-1045. Schmidt, V.A. (1974) e paleohydrology of Laurel Caverns, Pennsylvania: Proceedings of the 4th Conference on Karst Geology and Hydrology, Morgantown, W. Va., West Virginia Geological and Economic Survey, p. 123-128. Tognini, P. (2001) Lombard Southalpine karst: main features and evolution related to tectonic, palaeogeographic and palaeoclimatic regional history two examples of a global approach. In R R ap ports de RRecherches, Institut de Gographie, Universit de Fribourg: Cave genesis in the AA lpine belt, Huselmann, Ph., and Monbaron, M. (Eds.), p. 81-114. Vergari, A. (1998) Nouveau regard sur la splogense : le pseudo-endokarst du Tournaisis (Hainaut, Belgique). Karstologia, 31, 12-18. Vergari, A., et uinif, Y. (1997) Les palokarsts du Hainaut. Geodinamica Acta, 10:4, 175-187. Willems L., Pouclet, A., and Vicat, J.-P. (2002) Existence de karsts en roche cristalline silicate non carbonates en Afrique sahlienne et quatoriale, implications hydrogologiques. Bull. Soc. gol. France, 2002, 173:4, 337-345.

PAGE 214

15th International Congress of Speleology Speleogenesis 819 2009 ICS Proceedings 1. Introductione Portuguese mainland is composed of a Variscan massif bordered by western and southern Mesozoic basins, respectively the Lusitanian basin and Algarve basin. Aer about 135 Ma of extension (Wilson et al. 1990, 1996), with minor early transient inversions (Terrinha et al. 2002), the basins were subjected to Cenozoic compression as a result of convergence between the African plate and the Iberian microplate (Ribeiro et al. 1990). e Lusitanian basin was formed as a consequence of North Atlantic opening during the Mesozoic, following Pangaea fragmentation. It resulted from the breakup of the Iberian massif, which is part of the western arc of the Variscan orogen. Late Variscan northwest-to-southeast main fractures (Ribeiro et al. 1979; Ribeiro 2002) were subject to four riing episodes (Kullberg et al. 2006) taking place as follows: (i) Triassic Sinemurian (ca. 220-190 Ma), (ii) Pliensbachian Oxfordian (ca. 190-156 Ma), (iii) Kimmeridgian Lower Berriasian (ca. 156-145 Ma), and (iv) Upper Berriasian Upper Aptian (ca. 140-112 Ma). e rst ri deposits are a Triassic siliciclastic red-bed formation followed by Hettangian evaporites deposited in grabens or half-grabens (Rasmussen et al. 1998), of great importance in the tectonic evolution of the basin. e Sinemurian is represented by dolomites. e second riing event produced carbonate rocks, mainly marly limestones in the Lower Jurassic, and oolitic to bioclastic limestones in the middle and upper Jurassic. e deposits of the last riing events became in general more siliciclastic. e total thickness of the deposits in the center of the basin is about 5,000 m. e basin developed in an almost north-south direction, along which several faults delimited horsts and grabens. It can be subdivided further in the northern, central, and EVOLUTION OF PORTUGUESE KARST REGIONS IN A BASIN-INVERSION SETTING: IMPORTANCE OF FAULTING AND CONFINEMENT ON CAVE DEVELOPMENT AND SPRING LOCATIONSJO OS ANTANT N N IO O CR R ISPIMD Department and Centre of Geology, Faculty of Sciences, University of Lisbon Edifcio C6, Piso 3, Campo Grande, 1749-016 Lisboa Portugal e Portuguese Speleological Society e Mesozoic extensional break-up of the Iberian Variscan belt produced the Lusitanian basin, which was inverted during the Cenozoic. Late Variscan faults were then reactivated as frontal or lateral ramps in a transpressional regime. e resulting upli produced the Estremenho Massif, in which Jurassic limestone forms three elevated areas separated by grabens with the structural trends NW-SE and NNESSW. On the south, the plateaus and sierras are bounded by a NE-SW thrust. is is the Portuguese main karstic area, which includes the most important caves, karst springs and poljes. Extensive speleogenesis is associated with poljes and border faults, where caves reach lengths on the order of one to ten kilometers. Early speleogenesis is revealed by shallow isolated conduits with a close relation to bedding. ey are relics that preserve coeval underground deposits. Extensional faults were reactivated as strike-slip faults during basin inversion. ese were the rst to furnish preferred paths for concentrated inltration and therefore promoted the origin of cave networks. en, as fault movement continued, there was persistent displacement of broken conduits, although with non-uniform horizontal and vertical shiing. e rate of fault displacement was greater than that of karstication and completely disrupted the earlier networks and isolated segments both upstream and downstream. In the inner parts of rotated faultbounded blocks some fractures have increased permeability as a result of secondary traction and have enhanced karstication along these directions. Finally, maze caves developed at the contact of Mesozoic karst rocks with conning impervious Cenozoic sediments. ere the karstication exploited multiple fractures along thrust surfaces, creating an isotropic matrix where oblique strike-slip faults determined the direction of trunk conduits and therefore the origin of main karst springs.

PAGE 215

Speleogenesis 820 2009 ICS Proceedings 15th International Congress of Speleology southern parts, which are separated by ENE transverse faults. e central part is limited by the Nazar fault on the north and the Arrife fault (or, accordingly to other authors, Arrife, Lower Tagus, Tagus neck faults). During Mesozoic extension these strike-slip late Variscan faults behaved as normal or listric faults associated with rollovers, with the vector of maximum extension changing from northeastsouthwest to an almost east-west trend (Crispim 1993; Kullberg et al. 2006). Tectonic inversion during Cenozoic times has generated two alpine chains along the borders of the Iberia microplate: the Cantabrian-Pyrenean on the north, and the Betic on the south. Intraplate and distant reexes of these compressions were the reactivation of Variscan basement faults, namely along the Central Cordillera, and detachment of Mesozoic cover where the basal evaporitic complex was thick enough, as in the Lusitanian basin (Fig. 1). Africa-Eurasia Paleogene convergence was NNE-SSW, while Neogene convergence was NNW-SSE. Former NNE-SSW extensional normal faults reactivated as lateral ramps, whereas ENE-WSW faults behaved as frontal ramps, and NNW-SSE faults formed dextral dominos compatible with slight NW-SE shortening. Inversion also caused renewal of motion of Lower Jurassic evaporite layers along thrust faults, forming salt walls or diapiric anticlinal structures (Ribeiro et al 1990; Crispim 1993; Pinheiro et al. 1996; Kullberg et al. 2000; Ribeiro 2002). 2. Tectonic Inversion in a Selected Karst Regione Estremenho limestone massif is the main Portuguese karst area (Martins 1949; Crispim 1992). It constitutes the southern part of an inversion area on the Lusitanian basin, bounded by the Arrife fault, which is oriented NE-SW. Two convergent sets of faults partition the massif and compose the boundaries of two graben stripes, one NNE-SSW and the other NW-SE, along which the poljes of Alvados and Minde fomed. e westernmost geomorphologic unit of the massif is the Candeeiros sierra, structurally the ank of a NNE-SSW anticline, while the easternmost unit is the So Figure 1: Map of structures active in Portugal during the A A l pine collision (with minor modications, om RR ibeiro et al, 1990). 1 Strike-slip faults; 2 thrust faults; 3 Probable t hrust faults. NN F NN azar fault; AA F AA rrife fault; LT TF Lower TT agus fault; TNTN F TT agus neck fault; E M Estremenh o massif. P OO porto; L Lisbon; F Faro. DDashed square = area represented in Fig. 2. Figure 2: Sketch of Estremenho limestone massif. 1 TT opog raphic limit; 2 Fault; 3 Strike-slip fault; 4 AA rrife thrust fault; 5 AA ire anticlinal; 6 Monsanto syncline; 7 Karst spring; 8 AA ltitude asl (m); DDashed square AA : AA lados polje (represented in Fig. 4); DDashed square M: Minde polje (represented in Fig. 3); DDark grey texture: Jurassic limestone of Estremenho massif; Medium grey texture: mainly Mesozoic d etrital rocks; Light grey texture: TT agus basin Cenozoic format ions. AA CB AA lcobaa; AA CN N AA lcanena; BT TL Batalha; PMS Porto de Ms; RR MR R RR io Maior; TNTN V TT orres N N oas; VNONO OO urm. AA lm AA lmonda spring; AA l AA liela spring.

PAGE 216

15th International Congress of Speleology Speleogenesis 821 2009 ICS Proceedings Mamede plateau. Between the two sets of faults is the Santo Antnio plateau, whose triangular shape points northward (Fig. 2). e NW-SE set of faults, the Costa de Alvados fault on the northwest and Costa de Minde fault on the southeast, consists of two almost parallel side-stepping braided faults with an overlapping zone of about ve kilometers. is overlap zone dened a lazy S-shaped basin (Mann et al. 1983) that delineates the Alvados polje. It can be considered a pull-apart basin in a sinistral strike-slip regime, chiey extensional. During inversion a north-south compression induced a dextral strike-slip regime. e north and south borders of Alvados polje were then transformed, each in a push-up situated on its respective restraining bend (Fig. 3). e center of the polje is a graben in which the Upper Jurassic outcrops are surrounded by Middle Jurassic push-up structures (Crispim 1993). Minde polje is situated on the east block of Costa de Minde fault. is one, like Costa de Alvados fault, dips to northeast and is the footwall of a broad extensional rollover whose axis is in S. Mamede plateau (Manuppella et al. 2000). However, this roll-over is sliced by NW-SE faults parallel to Costa de Minde fault. e slices near Minde polje constitute its bottom and eastern ank. So, this ank is limited by two strike-slip faults, which moved dextrally during inversion (Fig. 4). Analysis of secondary fractures on this NW-SE block reveals the typical directions associated with a simple dextral shear system, such as synthetic and antithetic Riedl faults, as well as symmetrical to Riedl (Riedl 1929; Tchalenko and Ambraseys 1970; Wilcox et al. 1973; Bartlett et al. 1981). Both the Costa de Alvados and Costa de Minde faults behave as dextral lateral ramps on the inversion regime. eir movement is blocked against the Arrife fault, the abovementioned south-bounding fault of the main inversion area on the Lusitanian basin. So, the main Arrife fault was a frontal ramp during inversion, causing the Estremenho limestone massif to thrust over the Cenozoic formations of the Tagus basin. Situated respectively north and south of the lateral ramp, the Aire anticline and Monsanto syncline, both with NE-SW axes, are the main folds parallel to the thrust trend. e contact is very irregular because of the low-angle dip of the fault plane. e thrust zone exhibits secondary folds and is cut by faults with strikes near or oblique to the thrust trend. e broad movement also has a sinistral component. In conclusion, the initial setting for speleogenesis must rst take into consideration the former extensional regime with extension direction from northeast-southwest to east-west, which provided upli along the roll-over axis, eventually with exposure of soluble rocks. Second, compression on the Figure 3: Main actures on the east slope of Minde polje, interpreted according to a theoretical simple shear zone (adapted om Crispim 1993). RR Synthetic shear (R R iedel); RR AA ntithetic shear (conjugate R R iedel); P Secundary synthetic shear (P shear); E Extension acture (or tension acture, T T ). A A PCF AA lgar do Portal da Coadinha cave fault; MVF Moinhos Velhos cave fault; RR F RRegatinho cave fault; AA LF AA lgar da Lomba cave fault. e springs represented are Pena, on the le, and RRegatinho. Figure 4: Simplied sketch of strike-slip tectonics and karst feat ures around restraining bend of AA lados polje (modied om Crispim 1993). 1 AA nticlinal fold; 2 NN ormal reverted fault; 3 DDepressed block; 4 NN ormal separation; 5 RReverse separat ion; 6 DDextral slip, 7 Karst spring (Falsa); 8 OOther caves. R R SJ RRestraining junction; PD DZ Principal displacement zone (1 associated to Costa de AA lados fault; 2 associated to Costa de Minde fault). PFF Pena da Falsa fault; LMF L apa dos Morcegos fault; VCF Vale da Canada fault. VDD Ventas do DD iabo cave; AA L AA lgar do Ladoeiro cave.

PAGE 217

Speleogenesis 822 2009 ICS Proceedings 15th International Congress of Speleology north quadrant during inversion established transpressive regimes with dextral sense in the northwest-southeast lateral ramps and compressive to transpressive with a sinistral component in the northeast-southwest frontal ramp, while, in the inner parts of the domino blocks, related faults with simple shear were dominant.3. Speleogenesis Associated with Extensional StructuresSpeleogenesis associated with the former extensional regime must be searched for, rst of all at higher levels of the massif surface, and secondly on fractures related to the extension. Two types of features must be considered: (i) sediments, and (ii) speleological features. Sediments may be masses of sand or owstone, while the most signicant speleological features are horizontal or sub-horizontal passages. ey may be primitive isolated karst features, or they may evolve into integrated caves. Some represent ancient, in general weakly developed cave passages, later perched and truncated by massif upli and slope retreat. In other cases they continue performing a hydrologic role by themselves or by complementing successor cave systems. Examples of fossil inllings are spread widely over the study area. In general they do not allow depicting their structural origin, as they are limited remnants. Its distribution in altitude, when coupled with regional geomorphologic features and their correlative sediments, may be valuable in deciphering past phreatic levels. Examples in the study area are not related neither to the extensional phase nor to the Miocene main inversion regime, but instead with a Plio-Pleistocene littoral to which can be related a fringe of sands and dispersed owstones bordering some southfacing slopes. A few more extensive features, including rooess caves, are in contrast good indicators of ancient ow trends. Collapsed roofs are better identied in aerial photographs, since their surface morphology is in general made uniform by later slope evolution and vegetal cover. Several asymmetric open dolines settled on steep slopes at dierent heights are the result of roof collapse of phreatic trunk passages. Unlled penetrable caves are in general of modest diameter and extend for short distances. ey seem to be remnants of vadose or tributary passages, with no coeval sediments but thin layers of calcied clays and silts on the oor. Some are dip-oriented. Exceptional and controversial are presently active caves that formed along extensional faults. is is the case of Cova da Velha cave, which is located on a northwest-seoutheast fault parallel to the Costa de Alvados main extensional fault. It is a 1 km single passage with a linear pattern, which was intercepted by slope retreat in a pocket valley. In winter, water collected along the fault plane rises up through the entrance. is hanging spring is suspended about 250 m above the regional base level and is partially explained by fault connement of overhanging aquifer horizons dened by alternation of layers of dierent types of limestone. In neither the cave nor at the surface is there morphological evidence of recent displacement along the fault plane. So, it seems that the genesis of this cave has resulted directly from an extensional structure not reactivated by later inversion. However, two speleogenetic factors must be considered, which are rst the decompression of the massif, and later the fault cross-cutting, which allowed the water to ow. e existence of a latent proto-cave dating from extensional times seems implausible and not veriable. Several inactive caves, at higher levels adjacent to currently active caves, can be interpreted as remnants of ancient cave networks related to extensional structures. Distinguishing between piracy and independent cave networks may be dicult, because the passage trends and their concordance with regional dip are more straightforward criteria. In some cases the currently active network cut unexpected and untraceable passages lled with anomalous sediments. ey are surely signs of a former speleogenetic stage but are not exclusively attributable to extension.4. Cave Networks Installed on Structures Related to or Reactivated on InversionSome important caves partially follow faults that are related to, or have been reactivated by, inversion. In the case of the Minde polje shear strip, the most important cave of the area, Moinhos Velhos cave, has a trunk passage with more than 1 km along an extension fracture; the direction of Pena cave is close to the trend of the main strike-slip fault; Algar do Portal da Covadinha and Algar da Lomba caves are partly guided by Riedel shears; and passages in Regatinho cave follow a direction close to a secondary synthetic P shear (Fig. 3). Regatinho and Pena caves discharge at two springs at the foot of the east ank of the polje, and are less than 1 km long. ey have, like Olho de Mira and Contenda, a maximum discharge of about 2 m3/s and are intermittent. e largest springs of the Estremenho massif are Almonda and Alviela, which are situated at the points where oblique strike-slip faults cross the main thrust fault of Arrife. Here, maze caves have developed at the contact of Mesozoic karst rocks with conning impervious Cenozoic sediments. Cave development followed multiple fractures along the contact zones of thrust surfaces,

PAGE 218

15th International Congress of Speleology Speleogenesis 823 2009 ICS Proceedings creating an isotropic matrix where oblique strike-slip faults determined the direction of trunk conduits, and therefore the emplacement of the main karst springs. Almonda cave is about 12 km long and the spring discharges about 15 m3/s at its maximum, while Alviela spring discharges 17 m3/s maximum. is cave has been dived to a depth of 130 m through a total of 1.5 km of passages.5. Eects of Inversion on the Adaptation and Disruption of Cave NetworksAs mentioned above, in the study area inversion took place along a lateral ramp in a regime of dextral strike-slip and along a frontal ramp in a thrust regime with a slight sinistral component (Fig. 4). Falsa is a temporary spring located on the northern push-up of the Alvados polje. e cave is composed of an initial part with narrow passages that in general follow the trend of the main strike-slip dextral transpressive fault. Aer about a 1-km straight line in the WE direction, one suddenly reaches a trunk passage developed in the direction NE-SW. e southwestern extension of this passage is blocked against the strike-slip fault, and the water had to nd its way along the fractured strip contiguous to the fault. Some sand deposits found at the surface may be related to possible abandoned fragments of the trunk passage. On the south ank of the Alvados polje, uplied along the push-up structure related to the passage of the south strike-slip fault, other minor caves are also disrupted on both sides of the fault (e.g. Lapa dos Morcegos cave, Lapa do Necrial). Algar do Ladoeiro cave, south of the southern push-up of Alvados, is in an area where parallel tension fractures predominate. e upli in this compression zone and correlative erosion laid open several caves while relative slip on the faults break up old trunk passages and provide the path to new vertical speleogenesis. In the Minde polje, another remarkable example of disruption of a big trunk passage occurs in Moinhos Velhos cave, where Galeria Gmea ends abruptly against an impressive block chaos. e link between this fossil trunk passage and the active collector is a very narrow passage 200 m long (Crispim 1987). Along the top of the scarp of the Costa de Minde fault, and also on its west slope, are several caves cut o by slope retreat. Some, like Lapa da Ovelha and Ventas do Diabo, are big enough to be considered trunk passages whose upstream continuation has disappeared as a consequence of dierential upli and erosion along the fault line. e plunge of passages is important and proves that the block west of the fault was tilted aer the cave origin. is means that even if the extensional listric fault was the rst to disrupt the cave networks with a vertical component, the west member would retain its original dip. e eect of the transpressional regime was to accentuate vertical displacement, increasing the plunge of passages and causing a horizontal shi between the western and eastern sections.6. Conclusionse study area of Estremenho limestone massif is perhaps the most outstanding karst region in the world in which to study speleogenesis in the context of transpression associated with side-stepped overlapping strike-slip faults. e clearness of the tectonic structures, and the number of caves available, allow the outlining of an evolutionary pattern dating from remote extensional phases to the Miocene, when the inversion achieved its maximum expression. Even if somewhat misleading, relict caves and other karst features are valuable objects for the study of early speleogenesis. Several cases of truncated ancient collectors allow an interesting evaluation of the capacity of caves to keep up with tectonic evolution, and to investigate the modalities of hydrologic adaptation. ey can also provide a measure for displacement or shortening along main fault zones. It was shown that in the inner parts of the massif, i.e., in the area of poljes, the main strike-slip faults and their conjugates are very important not only to speleogenesis but also in the location of karst springs. e same is valid along the southern border of the massif. Here, the Cenozoic impervious rocks function as a conning layer in the syncline area or as a dam along the contact with the limestone massif. However the emplacement of the Alviela and Almonda springs is determined by oblique faults that also guide the development of associated caves.Acknowledgementsis was a contribution to the projects Interpretation of groundwater conuence and diuence at Minde polje as a tool to advise the accurate protection of resources Watermind (POCI/CTE-GEX/59086/2004) and Active and fossil phreatic caves of Portugal (SPE/TNC5).ReferencesBartlett, W.L., Friedman, M. and Logan, J.M. (1981) Experimental folding and faulting of rocks under conning pressure. Part IX. Wrench faults in limestone layers. Tectonophysics, 79, 255-277. Crispim, J.A. (1987) Evolution of the hydrology in Moinhos Velhos cave (Mira de Aire). Algar, Bol. Sociedade

PAGE 219

Speleogenesis 824 2009 ICS Proceedings 15th International Congress of Speleology Portuguesa de Espeleologia, 1, 3-8. (in Portuguese). Crispim, J.A. (1992) Karst characteristics of carbonated rocks of the Central region of the Estremenho limestone massif (Portugal). Algar, Bol. Sociedade Portuguesa de Espeleologia, 3, 9-18. (in Portuguese, English abstract). Crispim, J.A. (1993) Some considerations on the geological structure of Alvados and Minde poljes (Estremadura, Portugal). Algar, Bol. Sociedade Portuguesa de Espeleologia, 4, 13-26. (in Portuguese, English abstract). Kullberg, M.C., Kullberg, J.C., and Terrinha, P. (2000) Tectnica da Cadeia da Arrbida, in TT ect nica das R Regies de Sintra e AA rrbida. Memrias Geocincias, Mus. Nac. Hist. Nat. Univ., 2, 35-84. Kullberg, J.C., Rocha, R.B., Soares, A.A., Rey, J., Terrinha, P., Callapez, P., and Martins, L. (2006) A Bacia Lusitaniana: estratigraa, paleogeograa e tectnica. In Geologia de Portugal no Contexto da Ibria, Dias, R., Arajo, A., Terrinha, P., and Kullberg, J.C. (Eds.) Univ. vora, p. 317-368. Mann, P., Hempton, M.R., Bradley, D.C., and Burke, K. (1983) Development of pull-apart basins. Journal of Geology, 91, 529-554. Manuppella, G., Antunes, M.T., Almeida, C.A., Azerdo, A.C., Barbosa, B., Cardoso, J.L., Crispim, J.A., Duarte, L.V., Henriques, M.H., Martins, L.T., Ramalho, M.M., Santos, V.F., and Terrinha, P. (2000) Notcia Explicativa da Folha 27-A Vila Nova de Ourm, Inst. Geol. Mineiro, Lisboa, 156 p. Martins, A.F. (1949) Macio Calcrio Estremenho. Coimbra, 248 p. Pinheiro, L., Wilson, R., Pena dos Reis, R., Whitmarsh, R., and Ribeiro, A. (1996) e Western Iberia Margin: A Geophysical and Geological Overview. Proc. ODP. Sc. Results, 149, 3-23. Rasmussen, E.S., Lomholt, S., Andersen, C., and Vejbk, O.V. (1998) Aspects of the structural evolution of the Lusitanian Basin in Portugal and the shelf and slope area oshore Portugal. Tectonophysics, 300, 199. Ribeiro, A. (2002) So Plate and Impact Tectonics. Springer-Verlag, Berlin, 324 p. Ribeiro, A., Antunes, M.T., Ferreira, M.P., Rocha, R.B., Soares, A.F., Zbyszewsli, G., Almeida, F.M., Carvalho, D., and Monteiro, J.H. (1979) Introduction la Gologie Gnrale du Portugal. Serv. Geol. Portugal, Lisboa, 114 p. Ribeiro, A., Kullberg, M.C., Kullberg, J.C., Manuppella, G., and Phipps, S. (1990) A review of the Alpine tectonics in Portugal: foreland detachment in the basement and cover rocks. Tectonophysics, 184, 357-366. Riedl, W. (1929) Zur Mechanik geologischer Brucherschneinungen. Zentralblatt fr Mineralogie, Geologie, und Paleontologie, 1929B, 354-368. Tchalenko, J.S., and Ambraseys, N.N. (1970) Similarities between shear zones of dierent magnitudes. Geological Society of America Bulletin, 81, 16251640. Terrinha, P., Ribeiro, C., Kullberg, J.C., Lopes, C., Rocha, R., and Ribeiro, A. (2002) Short-lived compressive episodes during Mesozoic ri tectonics in the Algarve Basin, South Portugal, the cause of interruption of marine communication around the SW corner of Iberia in the Jurassic. Journal of Geology, 110:1, 101-113. Wilcox, R.E., Harding, T.P., and Seeley, D.R. (1973) Basic wrench tectonics. American Association of Petroleum Geologists Bulletin, 57, 74-96. Wilson, R.C.L., Hiscott, R.N., Willis, M.G., and Gradstein, F.M. (1990) e Lusitanian Basin of West Central Portugal: Mesozoic and Tertiary tectonic, stratigraphic and subsidence history. In Extensional T T ect onics and Stratigraphy if NN orth AA tlantic Margins. American Association of Petroleum Geologists, 46, 341-361. Wilson, R.C.L., Sawyer, D.S., Whitmarsh, R.B., Zerong, J., and Carbonnell, J. (1996) Seismic stratigraphy and tectonic history of the Iberian Abyssal Plain. Proc. ODP, Sc. Results, 149, 617-633.

PAGE 220

15th International Congress of Speleology Speleogenesis 825 2009 ICS Proceedings AGE OF CAVES IN THE CORDILLERA DE LA SAL (ATACAMA, CHILE)JO O DD E WA A ELE1, VINNCEN N ZOO PICOTT OTTI1, PAO AO LO O FORT ORTI1, GEOR OR GE BROOROO K2, CUCCHI FRANRANCO O3, ZINN I LUCA A3 1Italian Institute of Speleology, University of Bologna, Via Zamboni 67 40127 Bologna, Italy2 DDepartment of Geography, University of Georgia, AA thens GA A 30602, USA A 3 DDepartment of Geological, Enironmental and Marine Sciences, University of TT rieste, Italy Although the Cordillera de la Sal, close to San Pedro de Atacama (Chile), is one of the driest places on Earth, it contains extensive cave systems that have developed in halite. A detailed morphological study of these caves, combined with 16 AMS radiocarbon ages on wood and bone fragments recovered from cave ceilings and diamictons, have allowed us to dene when these systems formed and when sediments were emplaced. e sometimes huge cave passages appear to have formed in less than 2000 years by a succession of short-lived ash oods, probably aer single extreme rain events. 1. IntroductionRock salt or halite is at least three orders of magnitude more soluble than limestone. Because of this it rarely crops out extensively at the surface and where it does it is readily dissolved leaving insoluble residue (mainly clays and marls). Rock salt can only survive at the surface in an extremely arid climate and normally displays a large set of typical solution morphologies similar to those developed on limestone. Solution of rock salt also leads to the formation of true caves that can extend several km underground. Important salt karst areas with extensive cave systems include Mount Sedom in Israel (Donini et al., 1985; Frumkin, 1994, 1997, 1998; Frumkin and Ford, 1995), probably the best-studied example, Algeria (Sesiano, 1986) and the Zagros Mountains of Iran (Bosak et al., 1999; Bruthans et al., 2006). Recently, interesting salt caves have also been discovered in one of the driest places on Earth, the Atacama Desert, where mean annual rainfall averages 20-50 mm y-1 and where there may be no rainfall at all for several years at a time (Houston and Hartley, 2003; Sesiano, 2006). Close to the village of San Pedro de Atacama, North of the Salar de Atacama basin, there is an important NNESSW trending elongated anticlinal ridge composed of Oligo-Miocene evaporitic rocks known under the name Cordillera de la Sal. e thick salt beds of this ridge, even in this hyperarid climate, have been karstied by occasional rains showing a well developed surface karst geomorphology with extremely sharp rillenkarren oen isolating salt pinnacles of up to 15 m in height. In the past 15 years several cave expeditions have discovered and documented a subterranean karst drainage network with more than 4 km of caves and tunnels (Fryer, 2005; Maire and Salomon, 1994; Padovan, 2003; Salomon, 1995; Sesiano, 1998; 2006; Walek, 2005). A detailed morphological study of this area has been carried out both at the surface and in the most important caves of the Cordillera de la Sal with the aim of understanding the mechanisms responsible for their formation and evolution.2. e Cavesere have been cave expeditions to the Cordillera de la Sal since the early 1990s (Maire and Salomon, 1994) but only recently have American (Fryer, 2005; Walek, 2005), Italian (Padovan, 2003) and French caving teams (Sesiano, 2006) conducted detailed explorations. Mainly caves in the area close to San Pedro de Atacama have been surveyed, as the area south of the road leading to the Valle de la Luna is dangerous because of land-mines and as a result has largely been avoided. ere are only three explored caves in this southern area (Cueva Zorro Andina or cave of the Andean fox (Fryer, 2005) and Cueva de lElection (Sesiano, 2006) and Cueva Eclipses (Sesiano, 1998). In the ridge between San Pedro and the Valle de la Luna there are about 15 surveyed caves, although several others are known but have not been mapped. Six of these caves have been studied on maps and/or in detail: these caves are the Zorro Andina, Cueva Mina de Chulacao (Chulacao Mine Cave), Cueva Lechuza de Campanario (Barn Owl Cave), Cueva Paisaje del Sal (Salt Landscape Cave), Cueva Palacio de Sal (Salt Palace Cave) and Cueva Paredes de Vidiros (Fryer, 2005). ese caves were selected for study because of their extent and the presence of sometimes abundant speleothems. Cave locations are shown in Figure 1 and cave plans and proles are in Fryer (2005)2.1 MorphologyAll six caves, which are the best developed cave systems of the Cordillera de la Sal, are through caves, in that they can be followed from the stream input entrance to the outow

PAGE 221

Speleogenesis 826 2009 ICS Proceedings 15th International Congress of Speleology point. Upstream entrances are generally vertical (salt shas or collapses) and soon connect with sub-horizontal rock salt passages that slope gently to the outlet. e salt is dissolved initially along fractures but solution continues underground regardless of structure, oen cutting inclined to almost vertical salt layers. e sub-horizontal rock salt passages develop close to the local base level, at equilibrium gradients so that there is neither net erosion nor deposition. Zorro Andina and Lechuza del Campanario have a single elongated and oen meandering cave passage, while the others are branchwork caves with multiple stream entrances and rock salt passages of 1st and 2nd (Paredes de Vidiros, Chulacao, Palacio del Sal) up to 3rd order (Paisaje del Sal). Passages in all of the caves tend to widen downstream, indicating a gradually increasing ow of water that is not saturated with respect to halite probably because of high ow velocities and resulting rapid ow-through times. One of the most striking features of the caves is vadose canyons with laterally evolving meanders. ese meanders are generally not related to bedrock features (e.g. gypsum interbeds, fractures) and evolve freely, increasing sinuosity while downcutting. e cross-sections change shape vertically following the meander entrenchment, but the passage width remains more or less constant. In some upper (older) levels the width of the channel seems to be greater, suggesting a climatic control (these might have developed in more rainy conditions).2.2 Cave depositse halite caves of Atacama, like those in other halite karst areas (Donini et al., 1985; Bruthans et al., 2006), contain ephemeral speleothems that evolve relatively rapidly aer short wet periods. Most are composed of salt and have grown in fossil passages or well above the present cave oor, since sporadic ooding readily dissolves them. In sheltered areas near entrances, there are large white halite owstones probably deposited by evaporation of both inltrating and condensation waters. Most of these are relict features evidenced by chisel marks made by salt miners who frequented these caves several hundred years ago. Stream passages that contain water only aer rare major precipitation events have oors covered by a thin crust of halite. e most interesting cave deposits, however, are massive diamictons up to 3-4 meters thick that in places almost completely ll the cave tunnels. ese bedload deposits are the result of major ood events that occurred aer long dry periods, when channel oors and slopes were covered with debris that was readily transported by the ood runo. In Lechuza de Campanario cave two main diamictons have Figure 1: e Cordillera de la Sal showing the locations of caves examined (modied om Google Earth). 1. Cueva de l Election,  2.  Cueva Eclipses, 3. Cueva Zorro AA ndina, 4.  Cueva Palacio de Sal, 5. Cueva Paredes de Vidiros, 6.  Cueva Paisaje del Sal, 7.  Cueva Lechuza de Campanario, 8.  Cueva de Chulacao Fossil, 9. Cueva Mina de Chulacao.

PAGE 222

15th International Congress of Speleology Speleogenesis 827 2009 ICS Proceedings been recognized, while in other caves, with bigger drainage basins, similar events are preserved only randomly. Due to the extreme aridity, wood fragments and bones are preserved remarkably well inside these diamictons or in alcoves and in slackwater deposits on fossil benches (Fig. 2).3. Downcutting RatesIn order to determine how fast salt is dissolved in the Cordillera de la Sal area, a total of seven Micro Erosion Meter stations (MEM-stations) were installed in November 2007. ese stations consist of 5 stainless steel nails allowing measurement of 3 spots/station. Six stations were placed in Lechuza de Campanario cave: two on exposed horizontal rock salt surfaces, two on vertical bare rock salt walls, one on the roof, and one on a lateral horizontal bench 2 meters above the oor. A seventh station was placed on a salt block along the canyon oor of uebrada Lechuza (Table 1). Measurements were made in November 2007, at the time of emplacement and a few days aerwards, and in the period March-April of 2008 (Table 2). Based on the rst set of measurements, the mean erosion rate on horizontal surfaces is 4.8 mm y-1, on vertical walls 2.4 mm y-1 and in the cave generally around 1 mm y-1 (slightly greater on the oor than on the roof). A small ood between measurements had entrenched at least 2 cm into the salt block on the canyon oor of uebrada Lechuza, pulling out the stainless steel nails but leaving the deeper parts of the drill holes as evidence of where the nails used to be. is observation suggests that oods are at least one order of magnitude more important in entrenchment than is downcutting due to lms of condensation waters. A 5 meter high cave canyon passage might thus be developed in less than 1000 years. e mean downcutting rate of 5 mm y-1 is slightly lower than rates reported for the Mount Sedom area (Frumkin, 1995), which has a higher precipitation than the Atacama. It is useful to point out that entrenchment rates in caves can be exceptionally fast in the early stages of cave development, because the stream gradient, and therefore the cave prole, reaches equilibrium close to base level quite quickly, but down-cutting slows down once equilibrium is reached. Further entrenchment is possible only if boundary conditions change (e.g. base level lowering or upli). Figure 2: Examples of sample locations: AA : Cueva Lechuza de Campanario, bone in a diamicton; B: Cueva Palacio de Sal, sampling twigs stuck in the ceiling; C: Cueva Mina de Chulacao, small alcoe just beyond the main collapse entrance, where small mammal bones were obtained om barn owl (?) pellets.

PAGE 223

Speleogenesis 828 2009 ICS Proceedings 15th International Congress of Speleology StationAltitude (m asl) ExpositionDescription 12521Vertical Southwards 500 m East of Valle de la Luna (VdL), under an abandoned salt quarry, +/100 m S of the road. 22519HorizontalSame as before, on salt block on the ground at 4 meters from Station 1. 32460HorizontalAt the upper entrance (collapse doline) of Chulacao cave, at 50 m distance from ancient mine reservoirs. 42502Vertical Southwards 400 m upstream of Chulacao main doline entrance, on the northern bank of the small valley. 52437HorizontalOn salt block in the bed of the uebrada Lechuza, at little more than 200 steps from the cave entrance. 62405On cave roofCave roof in upper passage at 20 meter from the entrance (owstone) 72405On cave oor Floor of upper abandoned meander, at 3 m height (le side) and at 60 m from the owstone entrance climb. Table 1: Location and characteristics of the MEM stations. StationData + calibration (in mm) Measure (in mm) Surface lowering (in mm) Average Lowering (in mm) Days Annual lowering (in mm) 1st2nd1st2nd 124/11/07 8.84 29/03/08 9.12 10.379.660.99 1.031263.00 8.267.361.18 7.476.820.93 224/11/07 8.84 29/03/08 9.12 9.438.031.68 1.411264.10 6.425.531.17 7.836.621.39 330/11/07 8.80 30/03/08 9.12 4.272.741.85 1.801215.44 12.4311.940.81* 7.716.291.74 424/11/07 8.84 30/03/08 9.12 7.907.800.43 0.601271.73 9.348.900.77 8.528.250.60 5125/11/07 8.84 01/04/08 9.10 12.84Diss./ > 201127>20116.95Diss./ 12.28Diss./ 626/11/07 8.79 01/04/08 9.10 11.2611.260.31 0.301260.88 9.559.560.30 8.408.410.30 725/11/07 8.79 01/04/08 9.10 0.361261.04 10.1010.030.38 6.326.290.34 Salt surface of measuring point was overgrown with new halite crust. This measure was dis-Table 2: Lowering rates registered at the 7 MEM stations.4. Dating of CavesTo have a more precise idea of entrenchment rates and age of cave passages, a series of wood and bone fragments were recovered from diamictons and cave roofs for radiocarbon dating purposes. Ages of wood and bone fragments enclosed in diamictons date the sediment itself, while wood fragments le behind in alcoves or jammed in cracks in the roof represent the latest ood that passed through the passage and reached that height. In all cases the cave passage are older than the dated twigs and bones. Results are given in Table 3. e oldest dated wood fragment was obtained from a diamicton in Lechuza de Campanario cave and gave a

PAGE 224

15th International Congress of Speleology Speleogenesis 829 2009 ICS Proceedings corrected Libby age of 4025-3967 y B.P. (1 uncertainty). In the same cave a twig and a bone in a lower and younger diamicton gave ages of 1489-1439 and 1505-1459 y B.P., respectively. ese debris ows seem to belong to two distinct ash ood events that occurred aer long periods of drought, enabling surface runo to carry large amounts of bedload. Six wood and bone fragments from a small cave in another sector of the Cordillera gave ages of 1573-1673, 1220-1174 and 965-830 y B.P. A much larger cave in the same area but 15 m lower, provided ve ages younger than 1000 y B.P. (around 980, 740, 480 and a bone and a twig around 150 y B.P.) Finally, in another well-developed cave, with large underground stream passages, one wood fragment jammed in a crack in the ceiling 3.5 m above the oor and another in a wall crevice less than 1 m above the oor, gave corrected Libby ages of 2165-2119 and 1300-1254 y B.P., respectively.5. ConclusionsCaves in the Cordillera de la Sal appear to have formed in the past few thousand years. e oldest cave passage (Lechuza de Campanario) must be slightly older than 4000 years B.P. Downcutting appears to occur during very localized heavy storms, when rainfall is able to create sucient surface runo to temporarily form a river with turbulent ow that can transport centimeter-sized cobbles. Exceptional, but probably very localized rainfall/runo events (since they dont appear clearly in all caves), occurred around 4000, 2140, 1500 and 1280 y B.P. and deposited prominent diamictons and slackwater deposits. However, there is no evidence that similar-sized deposits were emplaced in recent times possibly suggesting that rainfall has been lower in the last millennium. All of the caves have subhorizontal passages, adjusted to the present base level except for Lechuza del Campanario where the outlet of the cave is 3 meters above the canyon oor. ere has been no important incision of this cave oor since the tectonic upli (< 1500 years) that caused downcutting in the canyon. Due to their very rapid evolution, perfect preservation of sediments (diamictons, wood and bones), and peculiar morphologies (cave cross-sections, meanders, ecc.), salt caves are clearly important recorders of past climates and environments.Acknowledgmentsis research has been made possible thanks to the Integrated Project Atacama of the Dipartimento di Scienze della Terra e Geologico-Ambientali of Bologna University. Special thanks to Kevin Downey (USA) and Elio Padovan (Italy) for their help in nding caves and giving useful information. Sample IDType NumberpmC1 d13C corrected Libby Age 1 d13C Palacio Del Sal Legn, +50-70 cm twigs R 03265118,27 0,34 -127723 -16,23 Palacio Del Sal Legn, Roof, Alti +35 twigs R 0326676,65 0,22 214223 -24,25 Chulacao fossil 1wood; small branchR 0326790,09 0,26 85222 -23,4 Chulacao fossil 2wood; branchR 0326881,60 0,23 165122 -22,88 Chulacao fossil 3wood; small branchR 0326981,88 0,24 161223 -24,27 Chulacao fossil 4wood; small branchR 0327086,105 0,256 119723 -25,56 Chulacao fossil 5small bonesR 0327189,769 0,245 94025 -15,92 Chulacao fossil 6small bonesR 0327282,712 0,229 159825 -15,9 Chulacao 1wood; small branchR 0327398,165 0,283 15923 -23,73 Chulacao 2wood R 0327489,646 0,249 98122 -12,32 Chulacao 3wood R 0327594,286 0,284 48324 -23,66 Chulacao 4wood R 0327691,652 0,27 73723 -20,49 Chulacao 5bone R 0327798,867 0,273 14325 -18,56 Lechuza 1 splintered woodR 0327861,008 0,224 399629 -21,78 Lechuza 2 twigs R 0327983,605 0,269 146425 -21,84 Lechuza 3 bone R 0328083,754 0,24 148223 -17,78 Table 3: Samples and their radiocarbon ages.

PAGE 225

Speleogenesis 830 2009 ICS Proceedings 15th International Congress of Speleology ReferencesBosak, P., Bruthans, J., Filippi, M., Svoboda, T., and Smid, J. (1999) Karst and caves in salt diapirs, SE Zagros Mts. (Iran). Acta Carsologica, 28:2, 41-75. Bruthans, J., Filippi, M., Zare, M., Asadi, N., and Zdenek, V. (2006) 3 N Cave: the longest salt cave in the world. National Speleological Society News, 64:9, 10-14. Donini, G., Rossi, G., Forti, P., Buzio, A., and Calandri, G. (1985) Monte Sedom. Milano, 135 p. Frumkin, A. (1994) Morphology and development of salt caves. National Speleological Society Bulletin, 56, 82-95. Frumkin, A. (1995) Rapid entrenchment of stream proles in the salt caves of Mount Sedom, Israel. Earth Surface Processes and Landforms, 20, 139-152. Frumkin, A. (1997) Classication and some morphometric features of salt caves. Proceedings of the 12th International Congress of Speleology, La Chaux-deFonds, Switzerland, p. 139-142 Frumkin, A. (1998) Salt cave cross-sections and their paleoenvironmental implications. Geomorphology, 23, 183-191. Frumkin, A., and Ford, D.C. (1995) Rapid entrenchment of stream proles in the salt caves of Mount Sedom, Israel. Earth Surface Processes and Landforms, 20, 139-152. Fryer, S. (2005) Halite caves of the Atacama. National Speleological Society News, 63:11, 4-19. Houston, J. and Hartley, A.J. (2003) e central Andean west-slope rainshadow and its potential contribution to the origin of hyper-aridity in the Atacama desert. International Journal of Climatology, 23, 1453-1464. Maire, R. and Salomon, J.-N. (1994) Les grottes du sel et du gypse dans le dsert dAtacama (Chili). uatrime Rencontre dOctobre, Pau (France), 1-2 Octobre 1994, p. 86-89. Padovan, E. (2003) Il sistema carsico della Cordillera de la Sal nel deserto di Atacama. Progressione, 48, 37-49. Salomon, J.-N. (1995) Le Chili. Pays des karsts extrmes. Karstologia, 24, 52-56. Sesiano, J. (1986) Diapir de sel karsti sur le anc sud de lAtlas Algrien. Hypoges Les Boueux, 53, 8-12. Sesiano, J. (1998) Phnomnes karstiques en zone aride. Le dsert dAtacama, au nord du Chili. Hypoges Les Boueux, 64, 48-52. Sesiano, J. (2006) Evolution actuelle des phnomnes karstiques dans la Cordillera de la Sal (Atacama, Nord Chili). Karstologia 47, 49-54. Walck, C. (2005) Observations on halite cave geomorphology. National Speleological Society News, 63:11, 20-21.

PAGE 226

15th International Congress of Speleology Speleogenesis 831 2009 ICS Proceedings KRAUSHHLE (AUSTRIA): MORPHOLOGY AND MINERALOGY OF AN ALPINE SULFURIC ACID CAVEJO O DD E WA A ELE1, LUKA A S PLAN AN2, PHILIPPE AA UDRADRA3, ANTONANTON IO O ROROSSI4, CHRR ISTO TO PH SPT TL5, VICTOR TOR PO O LYA A K6, and BILL MCINTONTOSH7 1 Istituto Italiano di Speleologia, University of Bologna, Via Zamboni 67 40126 Bologna, Italy 2 NN aturhistorisches Museum Wien, Karstund Hhlenkundliche AA bteilung, Museumsplatz 1/10 1070 Vienna, AA ustria 3 PolytechN N ice-Sophia, Engineering School of NN ice Sophia AA ntipolis University, 1645 route des Lucioles, 06410 Biot, France 4 DD ipartimento di Scienze della TT erra, University of Modena and RReggio Emilia, Largo S. Eufemia 19, 41100 Modena, Italy5 Universitt Innsbruck, Institut fr Geologie und Palontologie, Innrain 52 6020 Innsbruck, AA ustria 6 Earth and Planetary Sciences, University of NN ew Mexico, 200 Y ale Bld, NN orthrop Hall 87109 AA lbuqueque, NN .M., USA A 7 NN ew Mexico Bureau of Geology and Mineral RResources, N N ew M exico TT ech, 801 Leroy Place, 87801 Socorro, NN .M., USA A Kraushhle (Gams, Styria, Austria) is the only currently known sulfuric acid cave in the Eastern Alps. Cupolas, ceiling partings and portals, ceiling channels, replacement pockets, horizontal corrosion/ convection notches, sulfuric acid karren, blind chimneys, incomplete dissolution walls, drip holes and cup shaped hollows in the oor are the most striking morphological features in this cave. Mineralogical analysis showed the presence, besides calcite, of gypsum, gibbsite, opaline, jarosite, metalunogene, hydroxylapatite, halloysite, and alunite. e timing of speleogenesis was preliminarily determined using 40Ar/39Ar dating of alunite, a product of acid limestone weathering, to 80 Ka +/80 (the cave is thus younger than 160 Ka). Preliminary Udates of calcitic stalagmites indicate a minimum age of gypsum deposition of 52 Ka. Stable isotope data of these speleothem are consistent with an epigenic origin of the drip water at that time.1. IntroductionSulfuric acid corrosion in caves was rst identied in France (Socquet, 1801), then in Austria and Italy (Hauer, 1885; Pf, 1931; M, 1935). Sulfuric speleogenesis was discussed later in the American literature (M, 1968). e remarkable publication on Kane Cave (USA) by E (1981) suggested a speleogenesis entirely dependent on the eect of sulfuric vapor, which caused the replacement of limestone by gypsum. is cave, at that time considered to be an exotic form of speleogenesis, became a reference work aer the discovery of the famous Lechuguilla Cave in New Mexico (Hill, 1987; Hose et al., 2000; Engel et al., 2004). Considerable progress with regard to gypsum development and corrosion was made in Italy, mainly by the study of the Frasassi caves (Galdenzi and Menichetti, 1995). Following Egemeier, Audra et al. (2007) suggested that a major part of sulfuric speleogenesis may occur in the cave atmosphere (i.e., above the water table) by thermal convection and condensation-corrosion. e sulfuric acid process of speleogenesis is based on the oxidation of suldes to sulfuric acid (1), either directly, or by an intermediate reaction involving native sulfur (2). ese reactions are facilitated by sulfo-oxidant microbes. Sulfuric acid then reacts by dissolving the calcareous host rock, creating replacement gypsum and releasing carbon dioxide (3), which dissolves even more limestone according to the reaction (4): (1) H2S +2O2 <=> H2SO4 (SO4 -+ 2H+ in aqueous solution) (2) H2S + O2 => 2S + H2O and 2S + 3O2 + 2H2O => 2 SO4 -+ 2H+ (in aqueous solution) (3) CaCO3 + SO4 -+ 2H+ + 2H2O => CaSO4 2H2O (gypsum)!“b + CO2!‘b+ H2O (4) CaCO3 + CO2 + H2O => Ca++ + 2HCO3 -e general reaction of sulde oxidation thus requires oxygen in gaseous or dissolved form. Limestone undergoes a double dissolution, part involving the replacement

PAGE 227

Speleogenesis 832 2009 ICS Proceedings 15th International Congress of Speleology by gypsum. Dissolved sulfate and alkalinity are carried away by the karst water. e reactive paths of the sulfuric replacement corrosion develop from the deeper, phreatic parts of the aquifer to the vadose zone near the emergence according to the following succession: - At depth in the aquifer: microbial reduction of sulfates derived from evaporites into sultes; - in the aquifer body: possible addition of sulfate derived from the oxidation of pyrite, likely to be rereduced thereaer; at shallow depth: dissolved sulde of deep water partly becomes oxidized by mixing with oxygenated meteoric water, reacts with the limestone rock and produces sulfate in solution. In the conned atmosphere of the cave: H2S degasses and oxidation continues, forming sulfuric acid on the walls and in the cave pools: sulfuric acid dissolves limestone, which is replaced by gypsum. Above the water table: condensation runo along the walls dissolves gypsum. e detachment of the gypsum crusts makes them fall into the pools and adds to the washing out of sulfates; In the pools: the slowly owing water gradually removes alkalinity and sulfate produced by the sulfuric dissolution of limestone. In this paper the results of a multidisciplinary study on a remarkable sulfuric acid cave in the Eastern Alps, Kraushhle near Gams in Styria (Austria) are reported.2. Exploration History and Previous WorksFranz Kraus started exploration of a small and inconspicuous cave (Annerlbauerloch) in 1881 and discovered most parts of Kraushhle as it is known today. Due to rich speleothem decoration, but primarily because of massive gypsum deposits, he excavated a second entrance and opened Kraus-Grotte for visitors as early as 1882. In 1883 Kraushhle became the rst show cave in the world equipped with electric lights one year before Postojnska Jama. Hauer (1885) already linked cave genesis to a nearby sulde spring and suggested replacement of limestone by gypsum as a minor cave-forming process. Kraus (1891, 1894) supported this hypothesis (cave genesis due to metamorphism) and gave a detailed description of the cave and its morphology. Trimmel (1964) interpreted the origin of the Main Chamber by dissolution of a gypsum body. Puchelt and Blum (1989) examined the sulphur isotopic composition and validated the hypothesis of Hauer (1885), but also concluded that this process had only limited impact on the development of the cave. No detailed morphological studies had been carried out on this cave prior to our study. 3. Study SiteKraushhle is situated in the easternmost section of the Hochschwab karst massif which is part of the Northern Calcareous Alps (NCA). e area shows a complex tectonic style and is aected by major strike-slip faults. Whereas the NCA are dominated by Triassic carbonates the cave formed in the Hirlatz Formation, a red Jurassic crinoidal limestone. e cave opens at 616 m asl, 90 m above the Gams brook that carved a narrow gorge in its upstream section. According to a complete resurvey in 2008 the total length of the cave is 767 m and the vertical range is m (+31, m; Fig. 1). e cave consists of a Main chamber 50 m long, 6 to 15 m wide, and a volume of ca. 6000 m. It is connected with another chamber, the Elysium (25 x 10 m). Several conduits spread from these chambers that are partly parallel and interconnected. ey are mainly subhorizontal but also show vertical steps of up to 10 m. Both cave entrances are incidental intersections of the cavern with the erosion surface. e section close to the upper entrance shows a 3D maze pattern (Fig. 1). Several blind chimneys up to 4 m in diameter reach upward; the highest of them terminates 30 m above the oor of the main chamber. A step of this chimney, which is not covered by gypsum or clastic sediments like most of the other parts of the cave oor, shows up to 1 m deep karren. e walls and the ceilings of almost all parts of the cave are formed by cupolas and mega-scallops. Larger cupolas have diameters of up to 3 m and oen show portals that connect two of them. Notches with at roofs and convection niches formed at several levels, in particular in the Main chamber. Condensation-corrosion channels are mainly present close to the lower entrance. Especially the lower parts of the cupolas are sculptured by gypsum replacement pockets. In the southernmost gallery several cup-shaped corrosion features are present below triple junctions of cupolas. Both chambers, the interconnected galleries east of them and adjacent chimneys contain gypsum. Massive gypsum on the oor is up to a few meters thick and is oen perforated by drip holes. On the walls and ceilings gypsum crusts and crystals are present, which reach up to 30 cm in length in the main chimney where they cover several square meters. e cave also contains calcite speleothems including stalagmites, helictites, popcorn, a shield and locally extensive moonmilk.

PAGE 228

15th International Congress of Speleology Speleogenesis 833 2009 ICS Proceedings During strong rain or snowmelt several small rivulets enter the cave. Water drains towards the deepest point of the cave which is blocked by breakdown material. Parts of the cave that are aected by this modern vadose water lack gypsum deposits. Locally this epigenic overprinting is documented by allochtonous pebbles that are typical for the geological units in the upstream part of Gams Brook. A lukewarm H2S-rich spring (8.6 20.0C) emerges at the level of the Gams Brook, 71 m below the deepest known point of the cave. Isotope data show that the water is a mixture of cold karstic and thermal waters (Zetinigg, 1993). Presently there is no thermal anomaly in Kraushhle. e average annual temperature in the Main chamber is 7.6 (.2) C which corresponds perfectly with the average outside air temperature. 4. Morphologies and eir Genesise morphology of the cave is mainly due to its sulfuric origin by H2S degassing of water rising from depth. H2S oxidized to sulfuric acid either in shallow pools fed by both rising water and downowing condensation runo, or on walls and ceilings through condensation. ese two types of environments (i.e. aqueous/gaseous) produced two families of corrosion features. e sulfuric water probably ascended along fractures, which are no longer visible due to later clay deposition. In the Main chamber and in some adjacent passages shallow pools were present. e aggressive water body caused lateral corrosion, which formed a notch with a at roof, corresponding to the former pool level (Fig. 2A). e rising water was also probably thermal. At the rim of the warm pools, juxtaposed convection cells caused condensation-corrosion (Fig. 2B. Wall convection niches developed, which intersect in a blunted vertical edge (Fig. 2C). When deeply incised into the wall, their coalescence tends to form a notch with shallow embedded niches (Fig. 2B). e thermal gradient produced convection cells of bigger size, which carved the upper walls and the ceilings according to the airow paths. Condensation-corrosion produced wall niches, ceiling cupolas (Fig. 2H), condensationcorrosion channels, and blind chimneys, whose walls are covered by megascallops (Fig. 2G). As a consequence, and in contrast to epigenic caves which develop along the entire conduit length, sulfuric caves developing mainly by condensation-corrosion, expand at discrete places (Osborne, 2007). Adjacent passages intersect, creating larger passages which evolve into condensation domes. Remnants of wall partings form ceiling partings, portals (Fig. 2H), protruding corners, blades, etc. Condensation prevailed in the cool upper parts, causing diuse and dierential corrosion and giving rise to boxwork-like structures. On the contrary, the lowest parts of the walls were warmer and prone to evaporation. Sulfate precipitated as gypsum crusts, giving rise to corrosion below the gypsum deposits. At half-height of the wall, where condensation and evaporation competed, gypsum is restricted to replacement pockets (Galdenzi and Maruoka, 2003) (Fig. 2D). Downwards, gypsum crusts thickened as a result of increasing evaporation. ese crusts commonly detached from the walls and piled up as gypsum oors, locally in excess of 1 m (Fig. 2E). Primary formation of gypsum occurred as microcrystalline gypsum, and later recrystallization through condensation and evaporation produced large crystals (Fig. 2F). Calcite speleothems did not form during the hypogean phase, and are most likely entirely related to the epigene phase. Aer the cessation of the sulfuric phase, morphological changes due to meteoric inltrations also resulted in the deposition of clay. Consequently, diagnostic features of hypogenic speleogenesis such as the morphologic suite of rising ow (Klimchouk, 2007) are not clearly visible, being masked by sediments and speleothems. 5. Mineralogyirty samples were taken at dierent locations inside Kraushhle. Mineralogical analyses were carried out using a Philips PW 1050/25 X-ray diractometer, and, if only small amounts of sample were available, using a Gandol chamber (diameter 114.6 mm). In both cases the Figure 1: Plan view of Kraushhle with gypsum occurrences, sam ple sites and location in AA ustria.

PAGE 229

Speleogenesis 834 2009 ICS Proceedings 15th International Congress of Speleology experimental conditions included 40 kV and 20 mA, CuK radiation, and a Ni lter. In addition to calcite (derived from the host limestone) and gypsum, seven additional minerals were identied: opaline (SiO2 .xH2O), jarosite (KFe3(SO4)2(OH)6), gibbsite (Al(OH)3), metalunogene (Al2(SO4)3 .12H2O), halloysite (Al2Si2O5(OH)4H2O), alunite (KAl3(SO4)3(OH)6), and hydroxylapatite (Ca5(PO4)3(OH)). Gypsum is abundant in Kraushhle and forms large deposits or replacement pockets along the walls and on the roofs. Many of the replacement pockets still host the pseudomorph gypsum which preserved structures of the original limestone such as calcite veins and fossils. Opaline is a rather common mineral in lava caves, but is rare in limestone caves (Hill and Forti, 1997). Its presence indicates rather acid conditions, compatible with sulfuric acid speleogenesis. e three sulfate minerals (jarosite, metalunogene and alunite) and the silicate halloysite are also typical of such low-pH conditions and this mineral association has been reported from Frasassi Cave (Bertolani et al., 1973) and from Guadalupe Mountain caves, New Mexico (Polyak and Gven, 1996). ese minerals, together with the alluminium hydroxide gibbsite, are the products of alteration of clay deposits under acidic conditions (Polyak and Gven, 1996; De Waele et al., 2008). Hydroxylapatite was found in one single sample only and appears as transparent vitrous inclusions together with minor quartz, illite and gibbsite in a amorphous brown matrix. Its presence might be related to the alteration of organic material (bone?). Figure 2: Cave morphologies: AA) NN otches with at roofs carved by lateral corrosion of a sulfuric acid pool; B) NN otch formed by imbricated wall conection niches aboe the level of the thermal basin; C) Wall conection niches aboe the thermal pool; DD ) RReplacement pockets; E) Pile of gypsum om detaching crusts; F) Gypsum crystals; G) Megascallops in a blind chimney; H) Portal between two adjacent cupola.

PAGE 230

15th International Congress of Speleology Speleogenesis 835 2009 ICS Proceedings 6. AluniteAlunite was reported as a speleogenetic mineral in caves of the Guadalupe Mountains by Polyak et al. (1998) who successfully dated this mineral to yield the timing of speleogenesis of sulfuric acid caves in that area. e raw alunite-bearing Kraushhle sample was treated with 25% HF for one hour. Approximately half of the raw sample dissolved. XRD results showed the presence of gibbsite in the raw sample, but probably not enough to make up half of the sample. EDS showed the presence of Al, S and K (Fig. 3A). It is likely that amorphous materials make up some of the raw sample. e Kraushhle alunite crystals are similar in size and appearance to alunite from Carlsbad Cavern and Lechuguilla Caves (Polyak et al., 1998; Polyak and Provencio, 2000). ey are pseudo-cubic rhombs ranging in size from 1-10 m (Fig. 3B). Aer HF treatment, XRD results showed pure alunite (Fig. 3C). Rened unit-cell dimensions of A = 6.962 and C = 17.260 indicate that the sample consists of K-rich, near-end-member alunite. e speleogenetic Kraushhle alunite has an exceptionally young 40Ar/39Ar plateau age ( ka), suggesting that the timing of formation of the cave is recent (Fig. 4). is is the rst dated cave alunite outside of the Guadalupe Mountains. 7. Calcite Speleothems Kraushhle contains abundant stalagmites, many of which were removed by vandalism. Preliminary investigations suggest that these speleothems are low-Mg calcite and Figure 3: e Kraushhle alunite is the K-rich endmember. AA) the three detectable elements using ED DS are AA l, S, and K; B) cryst als are micron-sized rhombs; C) the XRD RD pattern of the HF-treated aliquot indicates that alunite is the only crystalline phase. Figure 4: 40AA r/39AA r age spectrum for a sample of ne-grained alunite om Kraushhle, AA ustria.

PAGE 231

Speleogenesis 836 2009 ICS Proceedings 15th International Congress of Speleology are composed of a dense, transparent fabric. Speleothem deposition clearly postdated formation and local subsequent dissolution of massive gypsum as shown by the location of stalagmites in depressions surrounded by remnants of thick gypsum. Active soda straws and owstone crusts on roots (close to the upper entrance) give evidence of rapid modern speleothem and moonmilk formation. Fragments of two inactive stalagmites were dated using Umethods and yielded ages of 52 and 36 ka, thus providing a minimum age constraint for the cessation of gypsum deposition and subsequent local removal. 13C values range from -7.9 to -6.9 and from -2.4 to +0.4 VPDB suggesting variable amount of biogenic (soilderived) C input into the seepage waters during this early stage of epigenic speleogenesis. 18O values range from -8.6 to -8.2 and from -10.3 to -8.5 VPDB. Modern drip water samples are required to relate these values to past environmental conditions, but it is likely that the large range in O isotope values of the second sample represents a paleoclimatic signal, consistent with the known highamplitude climate variability during the time interval of the last glacial.8. ConclusionsKraushhle is a remarkably young sulfuric acid cave with a minor epigene overprinting. e suite of morphologies such as cupolas, ceiling partings and portals, ceiling channels, replacement pockets, horizontal corrosion/convection notches, sulfuric acid karren, blind chimneys, incomplete dissolution walls, drip holes and cup-shaped hollows in the oor are diagnostic features of hypogenic speleogenesis due to sulfuric acid. Also mineralogy displays a distinctive set of sulfates and hydroxides typical of acid weathering. Alunite, in particular, has been dated using the 40Ar/39Ar method and has shown that the acid corrosion, responsible for the cave development, is younger than 160 ka. Two U/ dated stalagmites yielded ages of 52 and 36 ka and conrm the extremely young epigenic phase that followed the sulfuric acid speleogenesis. AcknowledgementsY. Dublyansky, I. Lenauer, R. Pavuza, G. Stummer and C. Tschegg gave valuable input to the discussion and helped sampling. C. Berghold, M. Hammer, S. Koppensteiner, and B. Tarmann helped resurvey the cave. A. Mangini provided Udates on two pilot speleothem samples. R. Pavuza provided temperature data of the cave. P. Provencio made the EDS and SEM images of the alunite sample and checked for impurities.ReferencesAudra, Ph., Hoblea, F., Bigot, J.-Y., and Nobecourt, J.-Cl. (2007) e role of condensation-corrosion in thermal speleogenesis. Study of a hypogenic suldic cave in Aix-les-Bains, France. Acta carsologica, 37:2, 185-194. Bertolani, M., Rossi, A., and Garuti, G. (1973) e speleological complex Grotta Grande del VentoGrotta del Fiume in the Frasassi Canyon (Ancona, Italy): a petrographical and mineralogical study. Proceedings VIth International Congress on Speleology, Olomouc, p. 357-366. De Waele, J., Frau, F., Muntoni, A., and Cannas, C. (2008) Ritrovamento di Halloysite nella Grotta Eraldo (Barega, Iglesias, Sardegna sud-occidentale). Memorie Istituto Italiano di Speleologia, 21 245-254. Egemeier, S.J. (1981) Cavern development by thermal waters. NSS Bulletin, 43:2, 31-51. Engel, A.S., Stern, L.A., and Bennet, P.C. (2004) Microbial contributions to cave formation: new insights into sulfuric acid speleogenesis. Geology, 32:5, 369-372. Galdenzi, S., and Maruoka, T. (2003) Gypsum deposits in the Frasassi caves, Central Italy. Journal of Cave and Karst Studies, 65:2, 111-125. Galdenzi, S., and Menichetti, M. (1995) Occurrence of hypogenic caves in a karst region: examples from central Italy. Environmental Geology, 26, 39-47. Hauer, F. (1885) Die Gypsbildungen in der Krausgrotte bei Gams. Verhandlungen der Geologischen Reichsanstalt, 1885, p. 21-24. Hill, C.A. (1987) Geology of Carlsbad cavern and other caves in the Guadalupe Mountains, New Mexico and Texas. New Mexico Bureau of Mines and Mineral Resources, 117, 150 p. Hill, C.A., and Forti, P. (1997) Cave minerals of the world. National Speleological Society, Huntsville, 463 p. Hose, L.D., Palmer, A.N., Palmer, M.V., Northup, D.E., Boston, P.J., and Duchene, H.R. (2000) Microbiology and geochemistry in a hydrogensulphide-rich karst environment. Chemical Geology, 169:3-4, 399-423.

PAGE 232

15th International Congress of Speleology Speleogenesis 837 2009 ICS Proceedings Klimchouk, A.B. (2007) Hypogene speleogenesis. Hydrogeological and morphogenetic perspective. National Cave and Karst Research Institute, Carlsbad, Special Paper Series, 1, 106 p. Kraus, F. (1891) Hhlenbildung durch Metamorphismus. Die Natur, 40, 197-199. Kraus, F. (1894) Hhlenkunde. Gerolds Sohn, Wien, 288 p. Martel, E.-A. (1935) Contamination, protection et amlioration des sources thermominrales. Congrs international des mines, de la mtallurgie et de la gologie applique, 2, 791-798. Morehouse, D. (1968) Cave development via the sulfuric acid reaction. National Speleological Society Bulletin, 30:1, 1-10. Osborne, R.A.L. (2007) Cathedral Cave, Wellington Cave, New South Wales, Australia. A multiphase, nonuvial cave. Earth Surface Processes and Landforms, 32:14, 2075-2103. Polyak, V.J., and Gven, N. ) Alunite, natroalunite and hydrated halloysite in Carlsbad Cavern and Lechuguilla Cave, New Mexico. Clays and Clay Minerals, 44: 843-850. Polyak, v V .j J ., and Provencio, p P .P. (2000) Summary of the timing of sulfuric acid Speleogenesis for the Guadalupe caves based on ages of alunite. Journal of Cave and Karst Studies, 62:2, 72-74. Polyak, V.J., McIntosh, W.C., Provencio, P., and Gven, N. (1998) Age and Origin of Carlsbad Caverns and related caves from 40 Ar/ 39 Ar of alunite. Science, 279, 1919-1922. Principi, P. (1931) Fenomeni di idrologia sotterranea nei dintori di Triponzo (Umbira). Le grotte dItalia, 1:5, 45-47. Trimmel, H. (1964) Die Kraushhle bei Gams (Steiermark). Hhlenkundliche Mitteilungen des Landesvereins f r Hhlenkunde in Wien und N 20, 70-75. Puchelt, H., and Blum, N. (1989) Geochemische Aspekte der Bildung des Gipsvorkommens der Kraushhle/ Steiermark. Oberrheinische geologische Abhandlungen, 35, 87-99. Socquet, J.-M. (1801) Analyse des eaux thermales dAix (en Savoie), dpartement du Mont-Blanc (Analysis of thermal waters at Aix[-les-Bains], in Savoy, MontBlanc Department), Cleaz, Chambry, 240 p. Zetinigg, H. (1993) Die Mineralund ermalquellen der Steiermark. Mitt. Abt. Geologie und Palontologie Landesmuseum Joanneum, 50-51, 1-335.

PAGE 233

Speleogenesis 838 2009 ICS Proceedings 15th International Congress of Speleology THE AGE OF CAVE SYSTEMS IN CENTRAL-EAST SARDINIA: PRELIMINARY DATAJO O DD E WA A ELE1 and DARRDARR YL GRAN RANGER R2 1Italian Institute of Speleology, University of Bologna, Via Zamboni 67 40127 Bologna, ITA TA LY 2DDepartment of Earth and AA tmospheric Sciences, Purdue University, Civil Engineering Building, RRoom 3277 AA 550 Stadium Mall DDrive, West Lafayette, INN 47907-2051, USA A e island of Sardinia has some of the most extensive cave systems in Italy, such as Codula Ilune (more than 42 km), Bue Marino cave (more than 17 km) and several others. Many caves seem to have developed over a long time span, and probably started forming during the Tertiary. A minimum age of these cave systems could previously be obtained indirectly because in certain areas Pliocene basalts sometimes ll karst conduits. To attempt to date some of the oldest cave branches quartz sediments were carefully mapped and nine samples for cosmogenic nuclide burial dating with 26Al and 10Be were taken in the summer of 2005 in two karst areas in Central-East Sardinia. e cosmogenic results indicate that all of the cave levels have similar ages, and are probably of Upper Pliocene age or maybe older. 1. Introduction Caves are important geomorphic features for studying past environments, because they can preserve sediments that are otherwise dicult to nd in surface deposits. Many caves are believed to have formed during uaternary, when climate and changing base levels (sea level, glacial erosion, etc.) were ideal for the development of karst systems. ere is, however, increasing evidence that many important accessible and still-active cave systems developed before the onset of the uaternary (e.g., Anthony and Granger, 2004; Huselmann and Granger, 2005). is includes several active cave systems in Central-East Sardinia (Insular Italy), which are known to have started forming during Tertiary, since well-developed conduits are lled with Pliocene basalts (De Waele, 2004). Dating caves is oen dicult and minimum ages can be obtained using isotopic methods on chemical (speleothems) and physical (sediments) deposits in caves. In this research the sediments of four inactive caves in the Taquisara valley and a fossil conduit in the Codula Ilune cave system have been studied and were dated using cosmogenic nuclides. is research allows some preliminary conclusions to be drawn on the geomorphic evolution of this part of Sardinia. 2. Study AreaIn Central-East Sardinia large Mesozoic carbonate mountains rest unconformably on a Palaeozoic basement complex composed of metasediments, metavolcanics and intrusive bodies. Surface and subsurface karst landforms are well developed in the largest of these carbonate areas: Supramonte, the Gulf of Orosei and several table mountains called Tacchi. e evolution of this interesting landscape, with the karstic areas towering above the basement complex and separated in dierent units by structurally controlled deep valleys and gorges, is believed to have occurred in the past 5 million years. With the aim of carrying out cosmogenic burial dating of cave sediments, two cave areas have been selected based on the abundance of quartz in well-documented cave sediments and the importance of underground karst in the framework of local landscape evolution. e rst area, the karstic Taquisara valley, is situated south of the Gennargentu mountains in the central-eastern part of Sardinia. It ows from NE to SW and is developed between the altitudes of 780 m asl. e valley dissects the Jurassic carbonate table mountains Tacco of Ulssai and Taccu Isara and almost reaches the Palaeozoic basement. e evolution of this valley and of the deeply incised Riu Su Pardu and Rio San Girolamo valleys, which separate the Tacco of Ulassai and Arba from the other table mountains, is reported to have occurred during Plio-uaternary (De Waele et al., 2005) based on geomorphic observations, but there are no exact time constraints relying on precise dating methods. Many caves are known along the borders of the valley, mainly characterized by sub-horizontal passages oen partially occupied by stream sediments. ese water table caves are situated at dierent heights along the valley borders, especially at elevations of 775 m, 815 m 850 m and 930 m asl on the SE side and 900 m and 950 m asl on the northwestern side (Fig. 1). e second area is located in the Codula Ilune river, which extends its drainage basin on granite rocks before cutting

PAGE 234

15th International Congress of Speleology Speleogenesis 839 2009 ICS Proceedings its canyon through the Gulf of Orosei Mesozoic carbonate rocks, reaching the Tyrrhenian at Cala Luna beach. is area hosts the most important cave systems of the island, with more than 42 km of mapped passages in the Codula Ilune cave system and 17 at Bue Marino cave. Pliocene basalts (2-3 My) ll ancient karst conduits cut by more recent cave branches, and the canyon has cut a more than 100 m deep gorge in the past 2 My as shown by the basalt plateau dissected by the canyon (De Waele, 2004) (Fig. 2).3. Caves and eir SedimentsTaquisara valley is one of the richest cave areas of Ogliastra (Bartolo et al., 1999). Six caves have been studied in detail and a total of eight quartzite pebble samples have been taken in four of these (two in each cave) (Fig. 3). In Genna e Ua cave (952 m asl), on the northwestern ank of the valley, the impressive main passage has a length of 60 m and is characterised by the presence of two underground collapse sinkholes that give access to an underlying cave level. e walls of these sinkholes reveal a >4 m thick section of quartzite conglomerates intercalated with owstone levels and overlying a 1 m thick sequence of clayey sands (Fig. 4AB). is sedimentary sequence is capped by an important owstone (Fig. 4C). In two places this owstone shows a thickness of more than 2 m and is extremely corroded. Samples were taken at the top and at the bottom of the conglomerate. At Taquisara cave (954 m asl), 500 m southwest of Genna e Ua, the underground river passage shows important cave sediments and a complex geomorphic history with an active cave level 70 meters below. Cave sediments are represented by quartz conglomerates with minor phyllite fragments sometimes occupying entire rooms, successively eroded Figure 1: 3DD view of RR iu Pardu valley looking SSE (om Google Earth). e TT aquisara valley cave area (bottom right), the capture of RR io Pardu (top le) and the abandoned RR iu QQ uirra (top middle) are indicated. Figure 2: Karst area of the Gulf of OOrosei (bricks) and major cave systems (black lines). e sample location is indicated by the star.

PAGE 235

Speleogenesis 840 2009 ICS Proceedings 15th International Congress of Speleology and transported to lower levels. No samples were taken in this cave. On the opposite side of the valley the big Serbissi cave (938 m asl) represents another underground river passage, most probably related to the same karstic cycle that generated the Genna e Ua and Taquisara caves. Although the passage is close to the surface (rock thickness above the cave might be less than 20 m, thus sediments might not be completely shielded from cosmic rays), two samples have been taken at dierent heights in the most internal part of the cave. In the underground stream passage of the Sa Bulverera cave (901 m asl), located 50 m below Genna e Ua, concretions are corroded and sediment relics occur along the walls at heights of almost 2 m, testifying that the passage was almost entirely lled with quartz conglomerates, successively removed during a re-activation period. e dimensions of this cave are less important and probably reect a shorter period of formation than the one that was responsible for the huge passages of Genna e Ua, Taquisara and Serbissi. Samples were taken at 2 m and 0.5 m above cave oor in the nal part of the cave. In the meandering Su Coloru cave (816 m asl), on the opposite side of the valley, the sedimentary sequence is more articulated, with alternating quartz conglomerates and owstones demonstrating cyclic erosion and depositional events (Fig. 4D). Dimensions are similar to those of Sa Bulverera, suggesting a comparable time span of formation. Also here two samples were taken, one at 0.5 m above cave oor, the second 3 m higher. At the same altitude several other interesting caves are known close by, documenting a stable base level. Twenty meters lower, the active Cabudu Abba resurgence (800 m asl) descends very rapidly and hosts several sumps located 15 m below the present Taquisara valley oor, containing sediments characterised by quartzite-carbonate sands deriving from the Genna Selole Formation. No samples were taken here since the cave is believed to be of very recent origin. An analysis of valley morphology did not reveal distinct river terraces, but the cave oors testify to dierent base level still stands. e Codula Ilune cave system is, with more than 42 km of total surveyed passages, the biggest cave system of Sardinia. It is composed of three main conuent underground rivers that form a principal drain that has its ultimate outlet at the Cala Luna resurgence, an underwater spring in the Tyrrhenian Sea. e upstream part of this cave system, Su Palu cave, is characterised by at least 5 levels of conduits, the highest of which is located 180 m above the present active river level (Fig. 3). It is here that quartz pebbles have been sampled on the oor of a fossil phreatic conduit. Sediments were characterised by sands (deriving from granite and carbonate rocks) containing quartz pebbles of some cm in diameter.4. Methods and ResultsBurial dating of cave sediments with 26Al and 10Be is one of the few radiometric methods that date lower uaternary and Pliocene deposits ranging in age from about 100,000 years up to 5 Ma. Burial ages indicate the time sediment has been underground, oen corresponding to the time in which the passage has developed or, in some cases, giving a minimum age of the passage. More details on the method are reported in Granger et al., 2001 and Granger and Muzikar, 2001. Cave sediments have been carefully mapped and samples were taken in the summer of 2005. All of the cosmogenic nuclide concentrations in the sediments were very low, indicating relatively high erosion Figure 3: NNNN W-SSE prole of TT aquisara valley at the height of Genna e Ua-Serbissi caves: only sampled caves are shown (except for Cabudu AA bba Spring).

PAGE 236

15th International Congress of Speleology Speleogenesis 841 2009 ICS Proceedings Figure 4: Samples: Genna e Ua cave, AA) quartz sediment with old speleothem layers exposed on a cave roof; B) detail of the thick sediment showing a owstone oor in between quartz deposits; C) e large ancient owstone that coers the entire quartz sedim ent sequence; DD ) quartz pebble sediments in carbonate cement attached to the wall of the river passage in Su Coloru cave. AA ll photographs by Laura Sanna. Cave Burial date (My) Genna e Ua (2 samples) 2.82 0.50 Serbissi (2 samples) undatable Sa Bulverera (2 samples) 2.46 0.53 Su Coloru (2 samples) 2.76 1.17 Codula Ilune (upper level) 2.37 0.47Table 1: Cosmogenic 26AA l/10Be burial ages of cave sediment.

PAGE 237

Speleogenesis 842 2009 ICS Proceedings 15th International Congress of Speleology rates in the sediment source area. Uncertainties in the burial ages are thus quite large. Of the ve caves of Taquisara, the sediments at Serbissi proved undatable due to insucient burial depth below the surface. Resulting ages are shown in Table 1. e cosmogenic results indicate that all of the cave levels have similar ages, and date at least to the Upper Pliocene. e horizontal cave passages of the Taquisara caves have formed during relative stable periods during which the Taquisara river slowed its incision, but the cosmogenic nuclide dating did not achieve sucient precision to distinguish these various still stands. e fossil Su Palu conduit has a similar age, and is thus also of Pliocene age or older. 5. Conclusionse Taquisara valley, according to these (few) dates, appears to have already achieved its present shape in the Pliocene (2-3 Ma). Its drainage basin almost certainly extended far beyond the actual outcrop of Jurassic limestones. e deep valleys such as Rio San Girolamo and Riu Pardu, instead, are younger than the Taquisara incision, and have presumably formed in the last 2 million years. e fossil levels of the Codula Ilune cave system (based on one cosmogenic date) also appear to be of Pliocene age and could be older. Karst conduits lled with Pliocene basalt are in agreement with this cosmogenic burial age. e deepening of the canyon, and thus also of the cave system, appears to have occurred aer the emplacement of the basalts, thus during uaternary. From these preliminary data the present landscape of Central-East Sardinia, with its isolated table mountains (TT acchi) or the rough mountains of Supramontes, resting on the Palaeozoic basement, seems to have started forming during Late Tertiary, with a major incision rate during the last 2 My. Further research is needed to conrm these dates and to relate these events to the incision of other main rivers of the region, that according to these preliminary data appear to be less than 2 My old.Acknowledgmentse rst surveys and geomorphic investigations in the Taquisara area have been carried out with the help of Roberto Follesa. is work has beneted from the collaboration of Salvatore Cabras, Barbara Ibba, Alberto Muntoni and Laura Sanna. Many thanks to Laura Sanna for sharing her cave pictures. Cosmogenic dating has been carried out at the Department of Earth and Atmospheric Sciences, Purdue University.ReferencesAnthony, D.M., and Granger, D.E. (2004) A Late Tertiary origin for multilevel caves along the western escarpment of the Cumberland Plateau, Tennessee and Kentucky, established by cosmogenic 26Al and 10Be. Journal of Cave and Karst Studies, 66:2, 46-55. Bartolo, G., Concu, P., Deidda, D., De Waele, J. Gratti, G., and Salis, T. (1999) Taccu dOgliastra. Ulssai, Osini, Gairo, Ussassi. Editrice sAlvure Oristano, 269 p. De Waele, J. (2004) Geomorphologic evolution of a coastal karst: the Gulf of Orosei (Central-East Sardinia, Italy). Acta Carsologica, 33:2, 37-54. De Waele, J., Di Gregorio, F., Follesa, R., and Piras, G. (2005) Geosites and landscape evolution of the Tacchi: an example from central-East Sardinia. Il uaternario, 18:1, 211-220. Granger, D.E., Fabel, D., and Palmer, A.N. (2001) PliocenePleistocene incision of the Green River, Kentucky, determined from radioactive decay of cosmogenic 26Al and 10Be in Mammoth Cave sediments. Geological Society of America Bulletin, 113:7, 825836. Granger, D.E., and Muzikar, P.F. (2001) Dating sediment burial with in-situ produced cosmogenic nuclides: theory, techniques, and limitations. Earth and Planetary Science Letters, 188, 269-281. Huselmann, P., and Granger, D.E. (2005) Dating of caves by cosmogenic nuclides: method, possibilities, and the Siebenhengste example (Switzerland). Acta Carsologica, 34:1, 43-50.

PAGE 238

15th International Congress of Speleology Speleogenesis 843 2009 ICS Proceedings Preliminary RELIMINARY development DEVELOPMENT of OF a A STATISTICALLYBASED karst KARST classificatio CLASSIFICATIO system SYSTEM PHORMSDDaniel H. DDoctor1, Benjamin F. Schwartz2, Marcus OO Gary3 1United States Geological Survey, 12201 Sunrise Valley DDrive, RReston, VA A 20192 USA A2 DDepartment of Biology, TT exas State University, San Marcos, TT X 78666, USA A3 Zara Enironmental, LLC, 118 West Goforth DDr., Buda, TT X 78610, USA A A karst classication system is necessary in order to identify common processes of karstication in disparate regions. A robust classication scheme for karst terrains and aquifers should be grounded in 1) a well-constructed geologic framework and 2) the hydrogeologic processes of karst development taking place within that framework. Prior classications of karst have been largely descriptive, lacking a foundation in quantiable parameters. A classication of karst should avoid being based solely upon morphologic descriptions of the numerous geomorphic features recognized within karst terrains, and instead be linked to the processes and geologic attributes that give rise to karst features. Ranking such processes and attributes according to their importance for karstication allows for a statistical comparison of dierent karst regions, and ultimately a more quantitative classication of karst terrains. Here, we introduce the PHORMS karst classication method. PHORMS is an acronym for the six factors considered in the classication: Physiography and climate, Hydrology, Other conditioning attributes, Rock properties, Morphology of karst features, and geologic Structure. e method is designed to be as quantitative as possible. Each factor comprises several attributes that are numerically scaled with regard to their relative importance for karstication processes then summed. A 6 x n matrix results: 6 numerical PHORMS factor values for each of the n karst regions being compared. e karst regions are then classied through the statistical techniques of Hierarchical Cluster Analysis (HCA), and the importance of each of the PHORMS factors within the classication is assessed through Principal Components Analysis (PCA). e approach presented here is preliminary and subject to renement. Our goal is to provide a classication system based upon quantitative parameters that can be used to eciently compare karst terrains around the world. e PHORMS classication method is suciently exible to be used as an exploratory tool as well as a means of comparison among factors responsible for karstication in a wide range of environments. 1. IntroductionAttempts to classify karst extend as far back as the history of karst science. Early work by Cviji and Grund classied karst terrain according to the degree of development of morphometric and hydrologic features, resulting in the broad classications of holokarst (complete or true karst), merokarst (partial karst) and transitional karst (Sweeting WEETING 1973). uinlan UINLAN (1967) and Sweeting W EETING (1973) expanded upon this approach and attempted to classify karst based upon a range of geomorphologic factors. More terms were added to the list of karst types, including designations such as uviokarst, glacio-karst (also known as nival-karst or cryo-karst), cone and cockpit karst (kegelkarst), tower karst, interstratal karst, naked karst (nacktkarst), denuded karst, exhumed karst, covered karst (including variants within), relict or fossil karst, paleokarst, syngenetic karst, thermal karst and pseudokarst. In spite of these various designations of karst types, several universal criteria were recognized to be important for karst development: rock properties, geologic structure, climate, type of unconsolidated cover, physiography, and past and present hydrologic conditions. Recently, greater focus has been placed upon the processes of karstication as a means of classifying karst. Debate has turned from questions such as what is epikarst?, what is paleokarst?, or what is pseudokarst? to what are the criteria for epigenic and hypogenic karst development? e three former questions arise when comparing karst terrains on the basis of their geomorphic features;however, due to varying interpretations of processes that give rise to observable morphologic features, clear consensus is

PAGE 239

Speleogenesis 844 2009 ICS Proceedings 15th International Congress of Speleology hardly possible. e latter question, on the other hand, grounds the discussion in the processes of karstication and landscape evolution that give rise to the features we observe, many of which are common in seemingly disparate regions. uantitative approaches to classication of karst have been based largely upon aquifer characteristics (see, for example, Bakalowicz AKALOWICZ and AND Mangin ANGIN 1980; Smart M ART and AND Hobbs OBBS 1986; El L Hakim AKIM and AND Bakalowicz AKALOWICZ 2007). In spite of the potential success of such an approach, its application thus far has been largely conceptual rather than practical. A more comprehensive approach would incorporate the geologic and geomorphologic aspects of karst development (White H ITE 1999), but this, too, has yet to be formulated in a practical manner. 2. e PHORMS Classication SystemHere, we present a preliminary classication system designed to include both the geomorphologic and hydrologic aspects of karst in a quantiable manner that can be applied globally to any karst region where the requisite data exist. We call this the PHORMS classication. PHORMS is an acronym for the six Table 1. AA ttributes included in the PHOR OR MS classication and the values assigned to the karst regions being compared. Values include relative scales for some attributes and quantitative scales for others.

PAGE 240

15th International Congress of Speleology Speleogenesis 845 2009 ICS Proceedings factors considered in the classication: Physiography and climate, Hydrology, Other conditioning attributes, Rock properties, karst Morphology, and geologic Structure. Each factor comprises several attributes that are ranked with regard to their relative importance for karstication processes (Table 1). e physiography and climate (P) factor includes topographic relief, prevailing climate, and thickness of insoluble, unconsolidated overburden. e hydrology (H) factor includes the modality of discharge frequency distribution for an index spring, the percentage of allogenic recharge, and the baseow depletion coecient of the spring as a measure of storage within the aquifer of interest. A factor termed other conditioning factors (O) accounts for hydrogeologic processes that may inuence current karstication processes, such as paleokarst, hydrothermal ow, or strong geochemical drivers toward karstication such as mixing corrosion or the inuence of sulfuric acid on speleogenesis. e factor that describes morphology of karst features (M) includes estimates on the spatial density of dolines, length of caves, and depth of caves. Rock properties (R) include matrix porosity, purity, and thickness of bedding. Finally, the geologic structure (S) factor includes inclination of strata, fracture frequency, and degree of deformation as expressed by faults and folds. is preliminary classication system only considers karst in carbonate rocks; classication for karst within other rock types and for pseudokarst will be developed separately. is system is designed to be as quantitative as possible, but necessitates some degree of subjectivity and simplication to include as many relevant factors as required to generate a useful classication. We have followed a method of attribute ranking and weighting as is done in karst groundwater and fractured aquifer vulnerability assessments (Doerfliger O ERFLIGER et ET al AL 1999; Denny ENNY et ET al AL 2003). e approach ranks each factor attribute in terms of its perceived signicance to karst development, as well as permits relative weighting (integer multipliers) of the attributes of each factor. e weighted ranks of the attributes within a factor are then summed, and the resulting values of a particular factor are normalized among all karst regions being compared for statistical and graphical purposes. For example, doline density is an attribute of the Morphology factor. As with all attributes, we use a simple tiered ranking, with 0=none, 1=low, 2=medium, 3=high. A more quantitative ranking could be based on an actual value of dolines per square kilometer where data are available. e higher rank indicates a higher signicance for karstication. e other Morphology factor attributes are mean cave depth and mean cave length. e attributes values are then summed to provide a single numerical value for the factor. is factor value is then standardized by subtracting the mean and dividing by the standard deviation among all of the other M values assigned to the karst regions being compared. Standardization is necessary to place all of the factor values within the same numerical scale. e standardized values of all PHORMS factors are a matrix of 6 x n, with n being the number of karst regions compared. Two multivariate statistical methods (Davis A VIS 2002) were employed to explore the data: Hierarchical Cluster Analysis (HCA) and Principal Components Analysis (PCA). HCA classies the dierent karst regions according to hierarchical correlations among the values in the PHORMS matrix. PCA identies the components of the matrix that account for the greatest amount of variance in the dataset. Although PCA is not a technique that can be directly used for classication, it permits an examination of those aspects of the dataset that are most likely exerting strong control over the classication borne out by the HCA. 3. Resultse example data shown in Table 1 are preliminary and are used to demonstrate proof of concept only. For this example, we chose to weight all of the attributes equally. Addition or modication of attributes, including weighting, within each of the six PHORMS factors is expected as the method is rened. e HCA was performed twice: rst using only the values of the six PHORMS factors as variables, and a second time using all of the attributes included in the classication as variables (Fig. 1). is served to test the method of summing the attribute values into single PHORMS factors. e HCA results of the PHORMS factors (Fig. 1A) fall into two major groupings separated to the rst-order on the basis of hydrologic condition: those having deep or signicant phreatic storage, and those generally lacking such storage. To a second-order, the classication seems to further divide the rst-order groups on the basis of structural deformation or lack thereof. At the third-order, dierentiation among karst regions occurs more rapidly as other conditioning attributes, such as pre-existing paleokarst, strong acids, or hydrothermal activity come into play. In contrast, the results of the HCA performed on a matrix of all attributes as individual variables showed a dierent discrimination within the rst-order, placing those regions having high structural deformation as well as signicant phreatic storage into the same grouping as those with little phreatic storage (Fig. 1B). As before, the rst-order discrimination among the three groups appears to be largely based on the degree of

PAGE 241

Speleogenesis 846 2009 ICS Proceedings 15th International Congress of Speleology phreatic storage; however, the discrimination is stronger. Although there is some similarity between the two HCA results, the discrepancies are interesting. For example, the Shenandoah Valley, Basin and Range, and Edwards Plateau regions were shied out of the rst-order grouping reective of high phreatic storage when the analysis was performed on all attributes. In order to explain this, the results of the Principal Components Analysis (PCA) can be used to provide additional insight into the HCA classication. In the case of the Shenandoah Valley and Basin and Range, the shi in categorization might be explained by the lack of primary porosity in the indurated Paleozoic carbonate rocks of these regions, since this attribute has the highest factor loading within the rst component of the PCA (Table 2). For the Edwards Plateau, the explanation is likely a more complex combination of attributes. As with the HCA, the PCA was performed rst using only the six PHORMS factors. e rst two components account for 66% of the variance of the data. e projection of the 6-dimensional data cloud into 2-dimensional space may be visually misleading due to the collapse of some points near to one another that may, in fact, be separated in a space of greater dimensions (Fig. 2). For example, the vectors for physiography (P), hydrology (H), rock type (R), and morphology (M) all fall within a cluster. ese four factors would be more separated in a space of greater dimension, as indicated by the factor loadings provided in the full component matrix (Table 2A). e rst component accounts for 47% of the variance of the data, and the factor loadings show that the greatest inuence on this component is exerted by the morphology (M=0.85) hydrology (H=0.78) and rock type (R=0.75) factors. e second component accounts for an additional 19% of the variance in the data matrix, and its loading factors are most strongly weighted on other conditioning attributes (O = 0.82) and geologic structure (S = 0.65). e third component is most weighted on the physiography factor (P=0.77). e PCA using all of the karst attributes as variables required three principal components to explain the same amount of variance (69%) that two components explained using only the six PHORMS factors as variables. High factor loadings (>0.70) within the rst component were on structural attributes (dip of strata, fracture frequency, and degree of faulting and folding) and hydrologic attributes (discharge frequency distribution and baseow storage); however, the highest loading (0.89) was on rock porosity (Table 2B). Other attributes with high loadings within the rst component (in decreasing order) were topographic relief (0.83) and cave depth (0.75). Attributes of the rst principal component with moderate loadings (between 0.70 Figure 1. Hierarchical Cluster Analysis (HCA) results. A) HCA results using the six PHORMS factors (summed attribute values) as variables. B) HCA results when all of the attributes were used as individual variables in the HCA classication. Figure 2. Principal Components Analysis results of the PHORMS factors. e arrows are vectors of the factor loadings, indicating relative importance for position of points within the 2-component space.

PAGE 242

15th International Congress of Speleology Speleogenesis 847 2009 ICS Proceedings and 0.50) were the percentage of allogenic recharge and cave length. e second principal component was most heavily weighted on other conditioning factors such as the presence of paleokarst (-0.82) and hydrothermal activity (-0.69). e negative loadings on these attributes indicate an inverse relation between these attributes and others with moderate loadings within the second component such as allogenic recharge (0.68) and doline density (0.66), possibly reecting the dierent expressions of deep and shallow karstication.4. Discussionere are several advantages to the PHORMS classication system. e rst is that quantiable information common among many karst regions is used in order to provide as objective a classication as possible. Databases on karst are growing rapidly in dierent regions; however, these databases lack a standard structure or guidance as to the key parameters needed for karst classication. e PHORMS system may serve as a guide to summarizing data collected within a particular karst setting in order to place the karst region within the classication. Admittedly, the values shown in Table 1 are based partly on objective data from the literature and partly upon educated guesses of the authors; thus, the analysis presented here should only be considered as preliminary. Nevertheless, the exercise provides a framework for further renement. e second advantage is that it permits direct comparison of dierent karst regions as well as a structure for statistically exploring the empirical connections among index parameters. Finally, the matrix structure also allows one to explore predictions of karst attributes. For example, one might create a multiple regression model in which doline density is set as the dependent variable in order to assess the relative importance of the other attributes on the surface expression of karst. Although empirical, the exercise may provide useful insight and help steer new research directions concerning the underlying processes and controls on karstication. 5. Conclusione preliminary PHORMS classication system reects an initial step toward a comprehensive classication of karst. Whatever classication scheme is applied to karst, it should enable theoretical models of karst processes to be placed within the classication alongside well-characterized regions. e ability to compile quantiable aspects of karst regions around the world is increasing with increasing research. e PHORMS classication system attempts to take advantage of these data for practical application in karst research and possible inclusion into developing databases such as the Karst Information Portal (KIP) or other future and existing systems of karst information organization.ReferencesBakalowicz AKALOWICZ M. and Mangin ANGIN A. (1980) Laquifre karstiques. Sa dnition, ses charastristiques, et son identication. Mem. Soc. Geol. France 11, 71-79. Davis A VIS J.C. (2002) Statistics and D D ata A A nalysis in Geology, ird Edition John Wiley and Sons, New York, 638 p. Doerfliger O ERFLIGER N., Jeannin EANNIN P.Y., and Zwahlen WAHLEN F. (1999) Water vulnerability assessment in karst environments: a new method of dening protection areas using a multi-attribute approach and GIS tools (EPIK method). Enironmental Geology 39(2), 165176. Table 2. Component matrices of the PCA A results. Higher absolute value of a factor loading indicates a greater contribution of that variable to the overall component. Factor loading values with absolute value greater than 0.70 are highlighted in gray; a bsolute values greater than 0.50 are in bold. AA) PCA A results using only the values of the six PHOR OR MS factors as variables. B) PCA A results using rank values of each individual karst attribute as variables; only the rst six principal components are shown for clarity.

PAGE 243

Speleogenesis 848 2009 ICS Proceedings 15th International Congress of Speleology Denny ENNY S.C., Allen LLEN D.M., and Journeay OURNEAY J.M. (2007) DRASTIC-Fm: a modied vulnerability mapping method for structurally controlled aquifers in the southern Gulf Islands, British Columbia, Canada. Hydrogeology Journal 15, 483-493. El L Hakim AKIM M., and Bakalowicz AKALOWICZ M., 2007, Signicance and origin of very large regulating power of some karst aquifers in the Middle East: Implication on karst aquifer classication. Journal of Hydrology 333, 329-339. uinlan U INLAN J.F. (1967) Classication of karst types: a r eview and synthesis emphasizing the NN orth AA merican literature 1941-1966. National Speleological Society Bulletin 29(3), 107-108. Smart M ART P.L. and Hobbs OBBS S.L. (1986). Characterisation of carbonate aquifers: a conceptual base. In Proceedings of t he Enironmental Problems in Karst TT erranes and eir Solutions Conference, October 28-30, Bowling Green, Kentucky, U.S.A., National Water Well Association, Dublin, Ohio, p. 1-14. Sweeting W EETING M.M. (1973) Karst Landforms. Columbia University Press, New York, 362 p. White H ITE W.B. (1999) Conceptual models for karstic aquifers. Karst Modeling: KWI Special Publication 5, A.N. Palmer, M.V. Palmer, and I.D. Sasowsky (Eds.), e Karst Waters Institute, Charles Town, West Virginia (USA), p. 11-16.

PAGE 244

15th International Congress of Speleology Speleogenesis 849 2009 ICS Proceedings TECTONIC INFLUENCES ON SPELEOGENESIS IN THE GUADALUPE MOUNTAINS, NEW MEXICO AND TEXAS, USAHAR AR VEY RR DD uCHEN N E1 and KIMBER R LEY I. CUNNNN INNGHA A M2 1Consulting Geologist, PO O Box 362, Lake City, CO O 81235, USA A2DDeceased Sulfuric acid speleogenesis in the Guadalupe Mountains of New Mexico and Texas is a consequence of the rise of the Alvarado Ridge and subsequent opening of the Rio Grande Ri during Cenozoic time. Uplands of the late Laramide (~38-35 Ma) Alvarado Ridge provided an immense recharge area that supplied water to aquifers draining eastward into the Permian basin. Prior to, or during the early stages of the opening of the Rio Grande Ri, hydrostatic head in the Capitan aquifer caused strong water ow that displaced oil in traps in the southeastern corner of New Mexico. At this time, water also owed upward along fractures to artesian springs in the aquifer within the ancestral Guadalupe Mountains. is resulted in solution enlargement of fractures and development of early-stage caves that may not have involved H2S. Extensional faulting since 29 Ma fragmented the east ank of the ridge, progressively reducing the size of the upland recharge area and reducing hydrostatic head. Fresh water inux also introduced microbes into Artesia Group (Permian, Guadalupian) oil reservoirs, causing biodegradation of petroleum and generating copious H2S. e water table within the Guadalupe Mountains began to fall 14-12 Ma in response to erosion and tectonism. During this time, oxygen-rich meteoric water mixed with H2S water to form sulfuric acid, which enlarged passages and galleries at the water table. Tectonic spasms related to the opening of the Rio Grande Ri caused abrupt drops in the water table, shiing the locus of sulfuric acid dissolution eastward and downward. Cave levels formed by sulfuric acid record the position of the water table at a given time, and the elevation dierence between levels may correlate with episodes of Rio Grande Ri tectonism since 12 Ma.1. Introductione Guadalupe Mountains of southeastern New Mexico and west Texas lie on the north margin of the Permian Delaware basin (Fig. 1). Within these mountains, Tertiary upli and erosion exposed Permian (Guadalupian) strata that contain caves formed by sulfuric acid. e current model for the speleogenesis of these caves is a combination of ideas rst proposed by Davis (1980) and Egemeier (1981 1987), where H2S derived from petroleum deposits was oxidized to sulfuric acid that dissolved limestone. Hill (1987 1990) conrmed a petroleum source and modied Davis conjectures by suggesting that speleogenesis was dependent on migration of H2S from the basin to the reef. Palmer and Palmer (2000) showed that initial stages of cave development resulted from rising water that reached the surface through springs, emphasizing the need for oxygen to convert H2S to H2SO4. Once primary conduits were formed, episodic lowering of the water table resulted in enlargement of passages and galleries at the water table where oxygenated meteoric water was available to mix with suldic water and form sulfuric acid. Polyak et al. (1998) used radiometric dating to show that Guadalupe caves were formed 12 4 Ma, with the oldest caves found at the highest elevations. e decrease in age with elevation reects the progressive lowering of the water table over a span of 8 Ma (Polyak et al. 1998; Palmer and Palmer 2000).2. Tectonic Settinge Guadalupe Mountains are located on the eastern ank of the Rio Grande Ri, an intermontane and intracratonic extensional feature superimposed on the Cenozoic Alvarado Figure 1: Index Map of the Guadalupe Mountains and vicinity. A A is the cross section shown in Figure 4.

PAGE 245

Speleogenesis 850 2009 ICS Proceedings 15th International Congress of Speleology Ridge (Fig. 2) in southern New Mexico and west Texas (Eaton 1986 and 1987; Chapin and Cather 1994). ese mountains lie between the Salt Basin graben and the Pecos River valley, and rise southwestwardly from beneath Ochoan evaporites and Pecos valley uaternary ll to an elevation of 2767 m at Guadalupe Peak in Texas (Fig. 3). e western margin is the Border Fault zone of King (1948). e upland surface of the Guadalupes slopes 1.2 degrees northeast (DuChene and Martinez 2000) and is continuous with the extensive upland of the Llano Estacado (Bretz and Horberg 1949), a late Eocene surface capped by Ogallala gravels (Fig. 2). e Pecos River valley is incised into the Ochoan evaporite sequence between the Llano Estacado and the Guadalupe Mountains where it breached the Capitan Reef Complex (Fig. 2)( Bretz and Horberg 1949; Motts 1968; Hiss 1980; Bachman 1980). e Alvarado Ridge (Fig. 2) is a regional topographic feature extending from southern Wyoming to westernmost Texas and northeastern Mexico (Eaton 1986, 1987). e feature was caused by distributed subcrustal thinning and related extensional strain, a mechanism similar to the cause of known marine and continental ri zones. e ridge is characterized by youthful mountains enclosing axial ri valleys, and by eastward and westward concaveupward slopes where elevation decreases asymptotically away from mountain crests. Sedimentary cover on the slope is composed of Miocene and Pliocene uvial sediments interbedded with rhyolitic tus, with the youngest undisturbed lithologies belonging mostly to the Ogallala Formation (Fig. 2). ese sediments were derived from upland areas of the Alvarado Ridge (Eaton 1987) and transported downslope to sites of deposition on a regional Eocene planation surface (Gregory and Chase 1992). e distribution of Miocene-Pliocene sediments is a consequence of the tectonic history of the Alvarado Ridge, a feature that has persisted against supracrustal and subcrustal degradation for at least 38-35Ma (Gregory and Chase 1992). e age of sediments derived from the crest of the ridge correlates with periods of tectonic maxima in the southern Rocky Mountains with ri initiation at 29-27 Ma, a maximum phase of extension between 17-14 Ma, and ri culmination at 7-4 Ma (Seager and Morgan 1979; Eaton 1987; Chapin and Cather 1994).3. Hydrogen Sulde and SulfurHydrogen sulde (H2S) is common in subsurface formations in southeastern New Mexico (Bjorkland and Motts 1959; Hinds and Cunningham 1970). is H2S is a byproduct of microbially assisted degradation of hydrocarbons associated with alteration of anhydrite to calcite (Kirkland and Evans 1976; Wiggins et al. 1993). Oil and gas accumulations in Permian Artesia Group strata east of the Pecos River become progressively more degraded to the west (DuChene and McLean 1989), and are rare west of the Pecos River. e most common occurrences of native sulfur are diagenetic deposits derived from reduction of H2S (Machel 1992). Diagenetic sulfur and H2S are common in the subsurface on the north ank of the Guadalupe Mountains (Hinds and Cunningham 1970). Diagenetic sulfur also occurs in the Gypsum Passage of Cottonwood Cave, in the Big Room of Carlsbad Cavern, and four sites in Lechuguilla Cave (Davis 1973; Spirakis and Cunningham 1992). In these deposits, and in associated cave gypsum deposits, sulfur is isotopically light compared to the Canyon Diablo Troilite (CDT) standard (Hill 1987). Limited uid inclusion data and modeling of the water chemistry in Lechuguilla Cave suggest that some sulfur formed subaqueously at a geochemical interface that was probably controlled by the availability of dissolved oxygen. Water composition was a complex mixture of fresh water, Figure 2: Distribution of Ogallala-age sedimentary coer. Modied om Eaton, (1987, g. 4).

PAGE 246

15th International Congress of Speleology Speleogenesis 851 2009 ICS Proceedings salty water, and various gases including CO2, CH4, H2S and light (C1 C6) aliphatic hydrocarbons (Spirakis and Cunningham 1992).4. Cave Elevations and Water Tableere are numerous caves in the western part of the Guadalupe Mountains with entrances at high elevations. A typical example is Hell Below Cave, located 8.9 km from the western escarpment (Fig. 3). Hell Below is about 150 m deep and has morphologic and mineralogic characteristics typical of sulfuric acid caves (Palmer and Palmer 2000). e entrance is at 2043 m, near the top of the Seven Rivers Formation and lies 945 m above the salt pan near Dell City, Texas (Fig. 3). Veldhuis and Keller, (1980) estimate 500 m of bolson ll in the Salt Basin, so the contact between the Yates and Seven Rivers formations is at least 1400 m lower in the Salt Basin than at Hell Below Cave. Most of the stratigraphic units that comprise the Capitan aquifer of Hiss (1980) crop out on the western escarpment of the Guadalupes (King 1948) and are above the modern water table. Today, the water table is at an elevation of 945 m at Carlsbad Springs (Fig. 3), and is estimated at 963 m at deep points within Lechuguilla Cave (Jagnow 1989). Extrapolating the gradient westward to Hell Below Cave gives an estimated potentiometric surface of 991 m, which is 1052 m below the entrance. is estimate compares favorably with Polyaks estimate of an 1100 m decline in the water table since 12 Ma (Polyak et al 1998; Polyak and Provencio 2000). If Hell Below Cave was the site of a owing spring early in its development, then the entrance had to be at or below the water table at that time. e present structural position and topography of the Guadalupe Mountains precludes a water table at the level of the entrance to Hell Below. ere had to be aquifer continuity and sucient elevation gain to the west to support the hydrostatic head required for a owing spring. is means that the Salt Basin graben had not yet subsided when speleogenesis was active at Hell Below Cave. Polyak et al. (1998) did not report an age for Hell Below Cave, but did report that nearby Cottonwood Cave, which has an entrance at 2073 m, formed 12.3 Ma.5. DiscussionSulfuric acid enlargement of caves occurred 12-4 Ma in the Guadalupe Mountains. However, the caves that were modied by sulfuric acid speleogenesis formed earlier (Palmer and Palmer 2000). ese caves formed by solution enlargement of fractures below the water table, a process that may not have involved H2S. Age dates reported by Polyak et al. (1998) record the time of sulfuric acid dissolution at the water table, not the onset of cave development, so the absolute ages of Guadalupe caves are unknown (Palmer 2006). e Cenozoic tectonic history of the region provides some constraints on timing of speleogenetic events. e Alvarado Ridge began to rise in early Tertiary time, and by 38-35 Ma, an elevated regional erosion surface extending across New Mexico had developed (Gregory and Chase 1992). Prior to opening of the Rio Grande Ri, the ridge was an immense upland recharge area for aquifers where groundwater owed to the east. As pointed out by Lindsay (1998), there was strong hydrodynamic ow in the mid-Tertiary that swept oil from Central Basin Platform elds. To move oil from these reservoirs requires a stronger hydrodynamic ow than exists today, so the aquifer system must have extended farther west than the Border Fault Zone (Fig. 3). e presence of sulfuric acid caves in the high western part of the Guadalupes also indicates that the aquifer had to extend farther west. As the Rio Grande Ri developed, the immense recharge on the east ank of the Alvarado Ridge was progressively reduced in area by extensional faulting (Fig. 4). Evidence of early ri tectonism is not recognized in Guadalupe caves. However, the interval from 14-4 Ma ts well with the 12-4 Ma cave age dates reported by Polyak et al. (1998). During this time, the elevation of the water table in the Capitan aquifer fell at least 1100 m between the westernmost caves in the Guadalupe Mountains and the deepest points in Lechuguilla Cave. e locations of cave passages and Figure 3: Tectonic map of the Guadalupe Mountains and vicinity with key landmarks and caves. Modied om Hayes, (1964, g. 24).

PAGE 247

Speleogenesis 852 2009 ICS Proceedings 15th International Congress of Speleology galleries mark the position of the water table at each pulse of sulfuric acid speleogenesis. Progressively lower levels of sulfuric acid enlarged passages from west to east record the times when the water table was stable, but the vertical dierence between levels reects times when either there was no speleogenetic activity, or the water table dropped more rapidly. Subsidence of the Salt Basin graben on the western margin of the Guadalupes began sometime between 12-4 Ma and probably contributed to the lowering of the water table in the Guadalupes. Once the graben subsided below the elevation of the lowest major passages in Carlsbad Cavern, active sulfuric acid speleogenesis in the Guadalupes ceased. is implies that the Salt Basin graben reached essentially its present conguration 3.8 Ma, the youngest age date for Carlsbad Cavern (Polyak et al. 1998), and this event brought the era of sulfuric acid speleogenesis in the Guadalupe Mountains to an end (DuChene and Cunningham 2006).ReferencesBachman, G.O. (1980) Regional geology and Cenozoic history of the Pecos region, southeastern New Mexico. U.S. Geological Survey Open File Report, 80-1099, 116 p. Bjorklund, L.J. and Motts, W.S. (1959) Geology and water resources of the Carlsbad area, Eddy County, New Mexico. U.S. Geological Survey Open File Report, 59-9, 322 p. Bretz, J.H., and Horberg, C.L. (1949) e Ogallala Formation west of the Llano Estacado. Journal of Geology, 57, 477-490. Chapin, C.E., and Cather, S.M. (1994) Tectonic setting of the axial basins of the northern and central Rio Grande Ri. in B asins of the RR io Grande RR i: Structure, Stratigraphy, and TT ectonic Setting, Keller, G.R., and Cather, S.M., (Eds.) Geological Society of America Special Paper, 291, 5-25. Davis, D.G. (1973) Sulfur in Cottonwood Cave, Eddy County, New Mexico. National Speleological Society Bulletin, 35, 89-95. Davis, D.G. (1980) Cavern development in the Guadalupe Mountains a critical review of recent hypotheses. National Speleological Society Bulletin, 42, 42-48. DuChene, H.R., and Cunningham, K.I. (2006) Tectonic inuences on speleogenesis in the Guadalupe Mountains, New Mexico and Texas. New Mexico Geological Society, 57th Field Conference Guidebook, p. 211-218. DuChene, H.R., and, McLean, J.S. (1989) e role of hydrogen sulde in the evolution of caves in the Guadalupe Mountains, in Subsurface and outcrop examination of the Capitan shelf margin, northern D D elaware basin, Harris, M., and G.A. Grover (Eds.) Society of Economic Paleontologists and Mineralogists Core Workshop, No. 13, p. 475-481. DuChene, H.R., and Martinez, R. (2000) Postspeleogenetic erosion and its eects on caves in the Guadalupe Mountains. Journal of Cave and Karst Studies, 62, 75-79. Eaton, G.P. (1986) A tectonic redenition of the Southern Rocky Mountains. Tectonophysics, 132, 163-193. Eaton, G.P. (1987) Topography and origin of the southern Rocky Mountains and Alvarado Ridge. in C ontinental Extension TT ectonics, Coward, M., J.F Dewey., and Hancock, L. (Eds.) Geological Society Figure 4: Cross section A shows the Guadalupe Mountains region prior to development of the Rio Grande Ri. Cross section B shows todays structural conguration. Modied om Matchus and Jones, (1984), and Lindsay, (1998, g. 3).

PAGE 248

15th International Congress of Speleology Speleogenesis 853 2009 ICS Proceedings Special Publication 28, p. 355-369. Egemeier, S.J. (1981) Cavern development by thermal waters. National Speleological Society Bulletin, 43, 31-51. Egemeier, S.J. (1987) A theory for the origin of Carlsbad Caverns. National Speleological Society Bulletin, 49, 73-76. Gregory, K.M., and Chase, C.G. (1992) Tectonic signicance of paleobotanically estimated climate and altitude of the late Eocene erosion surface, Colorado. Geology, 20, 81-85. Hayes, P.T. (1964) Geology of the Guadalupe Mountains, New Mexico. U.S. Geological Survey Professional Paper 446, 69 p. Hill, C.A. (1987) Geology of Carlsbad Cavern and other caves in the Guadalupe Mountains, New Mexico and Texas. New Mexico Bureau of Mines and Mineral Resources Bulletin, 117, 150 p. Hill, C.A., 1990, Sulfuric acid speleogenesis of Carlsbad Cavern and its relationship to hydrocarbons, Delaware basin, New Mexico and Texas. American Association of Petroleum Geologists Bulletin, 74, 1685-1694. Hinds, J.S., and Cunningham, R.R. (1970) Elemental Sulfur in Eddy County, New Mexico. U.S. Geological Survey Circular, 628, 1-12. Hiss, W.L. (1980) Movement of ground water in Permian Guadalupian aquifer systems, southeastern New Mexico and western Texas. New Mexico Geological Society Guidebook, 31st Field Conference, TransPecos Region, p. 289-290. Jagnow, D.H., 1989, e geology of Lechuguilla Cave, New Mexico, in Harris, M., and Grover, G.A., Subsurface and outcrop examination of the Capitan shelf margin, northern Delaware basin. in Subsurface and outcrop examination of the Capitan shelf margin, n orthern D D elaware basin, Harris, M., and G.A. Grover (Eds.) Society of Economic Paleontologists and Mineralogists Core Workshop No. 13 p. 459-466. Kirkland, D.W., and Evans, R. (1976) Origin of limestone buttes, Gypsum Plain, Culberson County, Texas. American Association of Petroleum Geologists Bulletin 60, 2005-2018. King, P.B. (1948) Geology of the southern Guadalupe Mountains, Texas. U.S. Geological Survey Professional Paper, 215, 183 p. Lindsay, R.F. (1998) Meteoric recharge, displacement of oil columns and the development of residual oil intervals in the Permian basin. West Texas Geological Society Publication, 98-105, 271-273. Machel, H.G. (1992) Low-temperature and hightemperature origins of elemental sulfur in diagenetic environments. in NN a tive Sulfur DDevelopments in geology and exploration Wessel, G.R. and B.H. Wimberley (Eds.), American Institute of Mining, Metallurgical and Petroleum Engineers (AIME) Special Volume, p. 3-22. Matchus, E.J., and T.S. Jones (1984) East to west cross section through Permian basin west Texas. West Texas Geological Society Special Publication No. 84-79, 1 sheet. Motts, W.S. (1968) e control of ground water occurrence by lithofacies in the Guadalupian reef complex near Carlsbad, New Mexico. Geological Society of America Bulletin, 79, 283-298. Palmer, A.N. (2006) Support for a sulfuric acid origin for caves in the Guadalupe Mountains. New Mexico Geological Society, 57th Field Conference Guidebook, p. 195-202. Palmer, A.N. and Palmer, M.V., 2000, Hydrochemical interpretation of cave patterns in the Guadalupe Mountains, New Mexico. Journal of Cave and Karst Studies, 62, 91-108. Polyak, V.J., McIntosh, W.C., Guven, N., and Provencio, P. (1998) Age and origin of Carlsbad Cavern and related caves from 40Ar/39/Ar of alunite. Science, 279, 1919-1922. Polyak, V.J., and Provencio, P.P., 2000, Summary of the timing of sulfuric acid speleogenesis for Guadalupe caves based on ages of alunite. Journal of Cave and

PAGE 249

Speleogenesis 854 2009 ICS Proceedings 15th International Congress of Speleology Karst Studies, 62, 72-74. Seager, W.R., and Morgan, P. (1979) Rio Grande Ri in southern New Mexico, west Texas and northern Chihuahua. in RR io G rande RR i: TT ectonics and Magmatism, Riecker, R.E., (Ed.), American Geophysical Union, 87-106. Spirakis, C., and Cunningham, K.I. (1992) Genesis of sulfur deposits in Lechuguilla Cave, Carlsbad Caverns National Park, New Mexico, in NN a tive Sulfur D Developments in geology and exploration Wessel, G.R. and Wimberley, B.H. (Eds.), American Institute of Mining, Metallurgical and Petroleum Engineers (AIME) Special Volume, p. 139-145. Veldhuis, J.H., and Keller, G.R. (1980) An integrated geological and geophysical study of the Salt Basin graben, west Texas. New Mexico Geological Society Guidebook, 31st Field Conference, Trans-Pecos Region, p. 141-150. Wiggins, W.D., Harris, P.M. and Burruss, R.C. (1993) Geochemistry of post-upli calcite in the Permian basin of Texas and New Mexico. Geological Society of America Bulletin, 105, 779-790.

PAGE 250

15th International Congress of Speleology Speleogenesis 855 2009 ICS Proceedings speleogenesisSPELEOGENESIS of OF THE New EW England NGLAND marble MARBLE caves CAVESTT Rev EV OR FAulk ULK Ne E R Limestone RResearch Group, GEES, University of Birmingham, c/o Four OOaks, Wilmslow Park NN orth, W ilmslow, Cheshire, SK9 2BDD, UK e structural geology of the northern Appalachians in New England, USA, comprises a collage of some ten Caledonide terranes that have over-thrust older, Grenville-age, rocks of the Canadian Shield. Some of these terranes contain karstic metacarbonates (marbles) of Precambrian to Middle Ordovician age that are commonly highly dismembered into lenses and merokarsts, and which contain >150 known caves with >9km of passages. Because the geological inheritance is rather similar to that of the previously-studied region of Central Scandinavia, it is instructive to compare the caves from the two areas and to consider variations in their speleogenesis. In Central Scandinavia, it has been reported that cave development proceeds as a four-stage process that is governed by the repeated uaternary cycle of glaciation, deglaciation and inter-glaciation. New England has a similar glacial history and the area provides evidence both of the deglacial seismicity that is probably necessary to create tectonic inception fractures within the marbles and of the deglacial ice-dammed lakes that enable the fractures to enlarge into phreatic conduits and cave passages by dissolution. e low angles of foliation of many New England marbles give their cave surveys a relatively planar appearance, and their mean lengths, cross-sections and volumes are rather smaller. However, the mean vertical range is comparable and subsurface cave distances are also consistently less than one-eighth of the depth of local glacial valleys, as in Central Scandinavia, suggesting that inception fractures were produced by similar processes. ere are proportionately less mainly vadose caves, but those that exist have larger mean cross-sections, probably arising from a longer period of interglacial conditions. e New England marble caves comply with the principles of the TopDown, Middle-Outwards model of cave development that applies in Central Scandinavia, although they commonly have less vertical complexity. It is concluded that were produced by similar processes, but that there are few multi-cycle caves that pre-date the Wisconsin glaciation. e relict (phreatic) caves and most combination caves that contain both relict phreatic and mainly vadose passages are single-cycle caves that started to form during deglaciation aer the Last Glacial Maximum. e few mainly vadose caves are half-cycle caves, which enlarged to present dimensions primarily during the Holocene. Similar processes also appear to apply to the caves in the Grenville-age marbles of the Adirondack Mountains of New York state.1. Introduction e New England (NE) marble caves lie in the northern Appalachian part of the Laurentian Caledonides, along the western side of Vermont, Massachusetts and Connecticut. e Caledonide mountain chain was formed during the Silurian closure of the Iapetus Ocean, when sedimentary rocks were subducted to depths of tens of kilometres, metamorphosed, and then overthrust on to older basements (Gee and Sturt, 1985). e limestones, whose ages vary from Precambrian to Middle Ordovician, were converted into marble and lost their original structures. Later riing created the Atlantic, when the original Caledonides separated into various terranes on both sides of the ocean. Subsequently, the whole north Atlantic region has undergone repeated glaciations, probably since the late Miocene, and some marbles have been karstied. e purpose of this paper is to compare the speleogenesis of these marble caves with that of marble caves in the four Caledonide allochthons of Central Scandinavia (CS), which was studied by Faulkner (2005a; 2006a; 2006b; 2007a; 2007b; 2008). He concluded that those caves commonly experience a four-stage, top-down, middle-outwards (TDMO) cycle of cave development that is driven by the glacial cycle: 1: Rapid deglacial isostatic rebound causes seismicity and forms inception fractures to a maximum distance from the surface equal to one-eighth the depth of the local glacial valley (tectonic inception) 2: Phreatic passages enlarge beneath owing deglacial icedammed lakes (IDLs) over periods up to 2000 calendar years at relatively high wall-retreat rates, despite low

PAGE 251

Speleogenesis 856 2009 ICS Proceedings 15th International Congress of Speleology temperatures and low PCO2 (deglacial speleogenesis) 3: Mainly vadose passages entrench during interglacials (interglacial speleogenesis) 4: Glacial erosion removes whole caves or their upper and outer parts during the next glaciation (glacial removal)2. New England Carbonate Geologye structural geology of NE is more complex than that of CS, because the area comprises a collage of some ten Caledonide allochthons that have over-thrust older, Grenville-age, rocks of the Canadian Shield. e thrust slices culminate in the Taconic Allochthons along their western extremity (Keppie, 1985). Where they occur, the karstic metacarbonates are commonly highly dismembered into lenses and merokarsts with low angles of dip, rather than into the long northaligned steeply-foliated stripe karsts that dominate much of CS. Faulting, jointing and thrusting at a local scale also occur. Individual metacarbonate blocks were sporadically transported across more competent rocks, and unconformities between rock types are sharply delineated. For example, it is sporadically possible to insert a hand between the upper surface of a marble outcrop and an overlying phyllite. Metamorphism varies from high to low grade, in a southnorth direction, and intrusions are common at higher grades. Some (short) caves are recorded in Winooski and Dunham Dolomites (uick, 1994), although these probably refer to dolomitic limestones rather than to pure dolomite. e area is characterised by northsouth aligned valleys extending from altitudes below c. 200m up to vegetated ridges and peaks, some being above 1,000m. 3. New England Glacial Historye glacial history of NE is rather similar to that of CS. As with its counterpart in Europe, the Wisconsin glaciation also appears to have had less magnitude than the two previous glaciations (Andersen and Borns, 1994, p40). According to Dyke et al. (2002) and Marshall et al. (2002), the icesheet thickness at the Last Glacial Maximum (LGM) increased from zero o the coast at Boston, via 1500 2500m across the Caledonides, to >3000m above Hudson Bay (when it may have extended over many basins of warmbased glaciation with subglacial lakes). Northern America apparently experienced several subsequent deglaciation / reglaciation phases (Dyke and Prest, 1987; Johnson and Lauritzen, 1995), with northsouth ow-switching as the Great Lakes region alternately melted and froze between the LGM and the Holocene (Clark et al., 2001). is may explain why a few caves contain ve or six cycles of rhythmic deposition of clay sediments and larger material (R. Pingree, pers. comm., 2002). Wisconsin deglaciation was probably complete in NE by c. 1300014Ca BP (Dyke and Prest, 1987; Andersen and Borns, 1994), c. 400014Ca earlier than in CS, and the impact of Younger Dryas cooling was much attenuated inland (Cwynar and Spear, 2001). Depositional evidence for the later stages of IDL evolution adjacent to retreating ice margins that moved from SENW (Stone and Borns, 1986, Fig. 1) has been reported by (e.g.) Clark and Karrow (1984: glacial Lake Iroquois at 329m altitude); Bierman and Dethier (1986: Lake Bascom, at >317m initially); Parent and Occhietti (1999: Lake Candona, which coalesced with the 125km-long Glacial Lake Vermont); Ridge and Larsen (1990) and Ridge et al. (1999): Lake Hitchcock; Rayburn et al. (2007); and ieler et al. (2007). Prior to this stage, only LaRocque et al. (2003) appear to have discussed the top-down melting of ice from mountain ridges that created static nunatak IDLs (when there was little sediment to be deposited), which later evolved into active IDLs as the ice sheet lowered. e author is unaware of any local models of early deglaciation with high-level IDLs that equate to the work of Grnlie (1975), who studied geomorphological features to calculate the rate of ice sheet lowering in CS, or to the thesis of Faulkner (2005a), who showed that all parts of inland CS were submerged beneath lowering IDLs for periods up to 2000 calendar years during deglaciation, with outlet ows into englacial Rthlisberger and / or subglacial Nye channels. However, it is assumed in this paper that similar processes applied in NE, so that local IDL ow regimes could also integrate with any underlying karst hydrology. e Holocene upli for the area varies south to north from c. 60m (Andersen and Borns, 1994, p18). e Atlantic coast contains many non-carbonate sea caves at and above the present sea level, where Rubin et al. (2002) discussed evidence of raised sea levels (including elevated sea caves, sea stacks and boulder beaches) on Mount Desert Island, Maine. However, the sea caves all have entrances that are only a few metres high, with a complete absence of very tall entrances (Rubin, pers. comm., 2002). is may indicate that the sea froze here before there was signicant isostatic depression at the onset of the Wisconsin glaciation, suggesting that there was no glaciation marine limit equivalent to one in CS suggested by Faulkner (2005b). However, as the karst areas are c. 180km from the coast at elevations >200m, probably none of the caves could have been inundated by the sea during either glaciation or deglaciation events, and the construction of local isobase maps is less relevant in NE to an understanding of cave development. e welldocumented existence of many tectonic ssure, fracture,

PAGE 252

15th International Congress of Speleology Speleogenesis 857 2009 ICS Proceedings and talus caves in a variety of metamorphic rock types provides evidence that this area also experienced many severe seismic shocks following rapid deglaciation and upli, perhaps comparable to those experienced in eastern Sweden (e.g. Mrner, 2003). e seismicity in northern NE was probably greater than that in CS, because, being nearer to the centre of the icesheet (in a position more comparable to eastern Sweden) the thickness of ice removed was even greater. However, the majority of the karst caves are located in the southern part of NE, where deglacial seismicity was probably less severe.4. New England Karst CavesA desk-based study of 153 marble caves was completed in 2004 by extracting information from uick (1994) and from the NN E C aver magazine into a North American Caledonides cave database, which is based on MS Excel. e database is incomplete, because location and altitude information is commonly suppressed in northern America (to protect the interests of property owners). However, the recording of karst type, cave type, cave hydrological class, main dimensions, entrances and hydrology was achieved to the denitions used by Faulkner (2005a) in CS, mainly from the well-presented cave surveys and descriptions. It is anticipated that these data are representative of the state of knowledge prior to 2004. About 9 km of passages are known, and the completeness of exploration may be higher than in CS, because groups of active cave explorers live locally in NE, although many karst outcrops and potential cave entrances are covered by extensive vegetation and glacial till. e author also made brief eld trips to NE in November 1996 and June 2002, visiting ve of the caves. e NE cave surveys do not have the same feel as those in the steep stripe-karst outcrops of the Helgeland Nappe Complex in CS. e reason is apparent from Table 1, which shows that 52% of the NE caves occur in low angle karst (L, commonly monoclinal, with dip o). Some 12% are in angled stripe karst (A, o), and none occur in vertical stripe karst (V). e karst type for 37% of the caves is unknown (X), but many of these (shorter) caves are probably also in low angle karst. Consequently, NE cave surveys are commonly less linear than in CS, and, with the strong faulting and jointing, some caves display ssure network patterns (Palmer, 1991). With a mean length of only 59m, the caves are commonly shorter than those in CS: the longest (Aeolus Bat Cave, VT) is only 900m long. e mean cave cross-section (XS, 2.9m2) and volume (234m3) are correspondingly smaller. However, at 9.3m, the mean vertical range (VR) is remarkably similar, and slightly deeper caves probably occur in karsts with lower dip angles, as in CS. e deepest (Purgatory System, VT) has a VR of 82m and a maximum subsurface cave distance of c. 40m. e passage with the greatest distance to the surface (68m) is in a deep sump in Morris Cave (VT). is is situated in a glaciated valley whose oor is at about 210m between peaks above 1,000m. us, the known distance of its deepest passage from the surface is well-within the one-eighth constraint that was proposed for the deglacial creation of inception fractures in CS during early fast upli. e caves also commonly contain large amounts of breakdown on chamber oors away from entrance areas and there are few chambers with smooth oors, suggesting that the caves were subjected to large seismic shocks. e authors 2002 visit to Nickwackett Cave and Chaee Mountain No. 2 Cave (VT) revealed evidence of internal tectonic movements of up to 15cm. e breakdown there is commonly covered by clay, suggesting phreatic deposition aer the last deglacial earthquake to modify an existing passage. Cave passages sporadically occur at junctions of two marble lithologies, New England Central Scandinavia Karst Type (See text)VALCXALLVALCXALL Count 1879 56153235480127339884 % of caves 11.851.636.6100.026.654.314.40.34.4100.0 Total cave length (km) 1.376.331.399.0920.2537.7912.642.911.3074.88 % of total cave length 15.269.715.2100.027.050.516.93.91.7100.0 Mean cave length (m) 7680 255986791009683385 Mean cave VR (m) 9.711.2 Mean cave XS (m2) 1.63.0 Mean cave Vol. (m3) 234 474Table 1: New England and Central Scandinavian Karst Types compared.

PAGE 253

Speleogenesis 858 2009 ICS Proceedings 15th International Congress of Speleology where tectonic inception fractures are likely to be created, as at Eldons French Cave (MA) and at Morris Cave. e caves in CS were divided among three hydrological classes by Faulkner (2005a): relict caves (which are almost all phreatic and are not fed by present allogenic recharge), mainly vadose (MV) caves (which may contain sumps) and combination caves (which contain both relict phreatic passages and levels and MV passages). From Table 2, over half the caves in NE are relict, a third are combination caves, and only 13% are MV caves. is represents a much larger proportion of relict caves and smaller proportions of both combination and MV caves than in CS, suggesting that Holocene vadose development was less important than deglacial phreatic development in NE. However, although combination caves commonly have the largest dimensions, the few MV caves commonly have larger dimensions than the short and extremely epigean relict caves, whose mean VR is only 5.3m. Entrance occurrences decrease in combination caves in the order DD ry Entrance (DE): Sink Entrance (SE): RR esurgence Entrance (RE), as in CS, but with smaller mean frequencies. Both combination and MV caves have slightly more cave streams per cave (CR), but far fewer sump pools (SP). Whereas relict and combination caves have mean XSs 16% and 31% smaller than in CS, the relatively few MV caves are 62% larger. e reason for the greater development of the active vadose parts of caves is probably that NE interglacial conditions started earlier than in CS (above), giving more time for vadose enlargement and / or entrenchment, and reducing the number of sumps by extra chemical and mechanical erosion of sump roofs and by the deeper down-cutting of pocket valleys. ey may also have larger catchment areas and shorter periods of winter freezing (not studied), but, as in CS, there are no glaciers or perennial snowelds to provide sustained meltwater recharge in summer. As expected from their elevations above probable marine limits, the caves do not contain entrances that were obviously enlarged by marine action. e greater sizes of entrance passages with parallel walls compared with internal passages are diagnostic of a pre-existing passage that has enlarged by ice-wedging as an IDL lowered past it (Faulkner, 2007c). e entrances to Aeolus Bat Cave (VT), Skinner Hollow Cave (VT), and at the resurgence of Horse Farm Road Cave (VT) (uick, 1994) appear to satisfy these criteria, indicating their existence prior to nal deglaciation. However, many entrances in NE are vertical shas into New England Central Scandinavia INTERNAL CAVE ATTRIBUTE RELICT CAVES COMBINATION CAVES MV CAVES ALL CAVES RELICT CAVES COMBINATION CAVES MV CAVES ALL CAVES UNITS Count 815220153279360245884No. % of caves 533413100314128100.0% Total cave length 1.9986.1890.9009.0879.41758.9956.46974.881Km % of total cave length Mean cave length 251174559341642685M Mean cave VR 5.314.810.79.35.914.53.98.8M Mean cave XS cave Volume 75473237234131102062474m3Average of DE 1.160.681.000.971.410.970.400.95No. per cave Average of SE 0.000.380.800.240.000.500.580.37No. per cave Average of RE per cave Average of all entrances per cave Average of CR per cave Average of SP 0.000.470. per caveTable 2: New England and Central Scandinavian Cave Hydrological Class comparisons.

PAGE 254

15th International Congress of Speleology Speleogenesis 859 2009 ICS Proceedings lower passages, where ice wedging would be less eective, or are themselves steep, shattered, passages. Although an analysis of variation with altitude has not been attempted, it seems likely that carbonate outcrops, cave occurrences and cave dimensions are primarily independent of altitude, as in CS, especially above 300m, up to which uick (1994) reported that the thicker glacial deposits in the Vermont Valley extend. e absence of reports of entrances situated very near peaks or ridge summits suggests that, also as in CS, few caves exist in a situation that was only submerged by a static nunatak IDL during deglaciation. However, as the NE marble outcrops commonly dip at rather low angles, their vertical distribution may not be as random as in CS, and an absence of very high altitude caves could arise from a corresponding lack of limestone at such elevations (not studied). A study of the NE cave surveys found only six relict vadose passages, which is even less proportionately than in CS. As none appear to lie above subsequently-formed phreatic passages, they provide no evidence that any combination caves started their enlargement prior to the nal glacial cycle. Only one occurs in a relict cave (Bats Den Cave, MA), and so relict caves were predominantly formed phreatically, as in CS, and therefore before the area was completely deglaciated. ere is no reported dating of speleothems to give any non-geomorphological indication of passage age, and signicant speleothems are rare, as in CS, suggesting a Holocene age for those that do occur. e only indication of possible multi-cycle cave development may therefore be the diameter of some passages, as in Aeolus Bat Cave. is, and the vertical complexity of the cave, hints that it may have developed over several glacial cycles. However, the mean XS of both relict and combination caves is less than in CS, which suggests that the total time that most of these caves remained submerged by owing deglacial IDLs was less than the 1,000,000 years assumed in CS. is authors observation of two sizes of scallops in Nickwackett Cave (about 10 and 30 cm), giving approximate ow-rates of 40 and 13cms-1 (both southwards) are similar to the major rates deduced for ows beneath IDLs in CS. 5. New England SummaryDeglacial seismicity (as suggested by uick, 1994) is conrmed by the existence of many talus and fracture caves and by movements within the karst caves themselves, which are suggestive of tectonic inception, and the area provides much evidence of low-level deglacial IDLs. e mean cave length, XS and volume are rather smaller than in CS, but the mean VR is comparable and subsurface cave distances consistently lie within the one-eighth relationship. ere are proportionately less MV caves, but they have larger crosssections than the MV caves in CS, probably because of the longer period of interglacial conditions. Valley-deepening in CS is probably in the range 15m per 100ka glacial cycle (Lauritzen, 1990). A comparable gure for NE is not known, but even if it is as low as 10m, then >50% of these epigean caves will be removed in the next glaciation. Hence, most existing relict and combination caves are probably single-cycle caves that only enlarged aer the Wisconsin LGM. ey commonly have less vertical complexity than those in CS, and comply with the principles of the TDMO model of cave development. e few MV caves are half-cycle caves, which enlarged to present dimensions aer deglaciation and during the Holocene, perhaps aer a deglacial phreatic initiation. Multi-cycle caves that developed over more than one glacial-interglacial period seem rare in NE, but may include Aeolus Bat Cave, plus Morris Cave and uarry Cave. ese two caves also contain rare abundant clay deposits (uick, 1994) that suggest deposition in almost static water, perhaps in a preexisting passage beneath a subglacial lake at the height of glaciation, or during post-LGM reglaciation phases (above). It is concluded that the marble caves in NE were formed by processes similar to those in CS, but that they developed over several glacial cycles even more rarely.6. Caves in the Adirondack Mountains of New Yorke signicant caves in the marbles of the Canadian Shield were not included in the North American Caledonides cave database, being outside the Caledonide terranes. However, visits in 1996 and 2002 to six marble caves in the Adirondack Mountains, which are situated in 1.3Ba Grenville-age crystalline marbles with large grain sizes, support a conclusion that these marble caves also t within the conceptual Caledonide models described by Faulkner (2005a). For example, Crane Mountain Cave contains many large dykes and sills, presumably of amphibolite (as does Browns Cave), and one of these forms the roof of the downstream sump. Although the cave is primarily in a low angle karst, with a dip of about 30o NE, the rst two waterfalls occur where the rock is folded complexly. A vertical fracture at the entrance shows tectonic movement, apparently with broken calcite, and at the base of the second waterfall is a fault with slickensides 15cm long, weathered to black. e commercial Natural Stone Bridge and Caves (Fig. 1) consists of a large stream captured by a large, complex, phreatic series of sumps beside a normallydry limestone gorge. is has small rockmills in its oor,

PAGE 255

Speleogenesis 860 2009 ICS Proceedings 15th International Congress of Speleology indicating formation during deglacial outows. Rusty Stove Cave has an obvious tectonic movement along a fracture on the le side of its entrance. e cave itself has formed along a dyke wall, which is breached at an inner chamber. e nearby Burroughs Cave has two moved joints orthogonal to the entrance passage that are about 10m apart. It contains the large Breakdown Chamber, formed by upward stoping of collapsed blocks, with no dissolution evident above its lowest level. e karst cave with the greatest VR is Crane Mountain Cave (about 30 m). All these caves also t within the one-eighth constraint for depth of exploited fractures. e local presence of large numbers of talus and ssure caves also supports the concept of fracture generation by postglacial seismicity. ese caves were also likely inundated by deglacial IDLs, and thus these and other marble caves in the Canadian Shield probably also comply with the speleogenetic concepts of the TDMO model.7. Newfoundlande island of Newfoundland forms a tectonic structural link between the northern Appalachians and the British Caledonides (van Staal et al., 1998, p213), comprising an assemblage of some six terranes with similarities to those of Britain and Ireland. However, the Dalradian Supergroup of the British Isles, with its metamorphic carbonate outcrops, appears to narrow considerably in Newfoundland, either within the Notre Dame Subzone or in its outlying Fleur de Lys Supergroup. ese subzones do not appear to contain signicant marbles, and no karst caves are reported there. e scattered outcrops of the Taconic Allochthons on the west of the island mainly consist of igneous and plutonic rocks. e Humber Zone on the St. Lawrence promontory contains large outcrops of sedimentary carbonates of Cambrian and Ordovician age, similar to the Durness Group limestones of northern Scotland (RA Gayer, University of Cardi, pers. comm., 1998). Higham (2001) reported a 780m-long karst cave in this limestone, together with other exokarst features. He also noted the existence of many sea caves from all over the island, with entrances up to 15m high, but not elevated above sea level. e speleogenesis of caves in sedimentary limestones adjacent to the glaciated Caledonides in Newfoundland and elsewhere awaits detailed study.ReferencesAndersen, B.G., and Borns, H.W. (1994) e Ice Age World. Scandinavian University Press, 208p. Bierman, P.R., and Dethier, D.P. (1986) Lake Bascom and the deglaciation of northwestern Massachusetts. Northeastern Geology, 8 (1/2), 32-43. Clark, P., and Karrow, P.F. (1984) Late Pleistocene water bodies in the St Lawrence Lowland, New York, and regional correlations. Geological Society of America Bulletin, 95, 805-813. Clark, P.U., Marshall, S.J., Clarke, G.K.C., Hostetler, S.W., Licciardi, J.M., and Teller, J.T. (2001) Freshwater forcing of abrupt climate change during the last glaciation. Science, 293, 283. Cwynar, L.C., and Spear, R.W. (2001) Late glacial change in the White Mountains of New Hampshire. uaternary Science Reviews, 20, 1265. Dyke, A.S., and Prest, V.K. (1987) Late Wisconsinan and Holocene history of the Laurentide Ice Sheet. Gographie Physique et uaternaire, 41 (2), 237263. Dyke, A,S., Andrews, J.T., Clark, P.U., England, J.H., Miller, G.H., Shaw, J., and Veillette, J.J. (2002) e Laurentide and Innuitian ice sheets during the Last Glacial Maximum. uaternary Science Reviews, 21, 9. Faulkner, T.L. (2005a) Cave inception and development in Caledonide metacarbonate rocks. PhD esis. University of Hudderseld. Faulkner, T. (2005b) Modication of cave entrances in Norway by marine action. Proceedings of the fourteenth International Speleological Congress, Athens. Paper O-69, 259-263. Book of abstracts. Figure 1: Sink entrance at Natural Stone Bridge (NY) showing the common epigean nature of marble caves.

PAGE 256

15th International Congress of Speleology Speleogenesis 861 2009 ICS Proceedings Abstract O-69, 90. Faulkner, T. (2006a) Tectonic inception in Caledonide marbles. Acta Carsologica, 35 (1), 7. Faulkner, T. (2006b) Limestone dissolution in phreatic conditions at maximum rates and in pure, cold, water. Cave and Karst Science, 33 (1), 11. Faulkner, T. (2007a) e one-eighth relationship that constrains deglacial seismicity and cave development in Caledonide marbles. Acta Carsologica, 36 (2), 195-202. Faulkner, T. (2007b) e hydrogeology of crystalline rocks as supporting evidence for tectonic inception in some epigean endokarsts. Cave and Karst Science, 33 (2), 55-64. (For 2006). Faulkner, T. (2007c) Cave entrance modication by marine erosion and by ice-dammed lakes. First Baltic Speleological Congress, Gotland, Sweden, August 2007. Abstract: S9. Faulkner, T. (2008) e top-down, middle-outwards, model of cave development in central Scandinavian marbles. Cave and Karst Science, 34 (1), 3 3-16. (For 2007). Gee, D.G., and Sturt, B.A. (Eds.) (1985) e Caledonide Orogen Scandinavia and Related Areas. John Wiley, 1250 p. Grnlie, A. (1975) Geologien I Vefsnbygdene. Vefsn Bygdebok, p. 417. Higham, S. (2001) An introduction to the caves of Newfoundland. NE Caver, 32 (4), 134. Johnson, R.G., and Lauritzen, S-E. (1995) Hudson BayHudson Strait jkulhlaups and Heinrich events: a hypothesis. Palaeogeography, Palaeoclimatology, Palaeoecology, 117, 123. Keppie, J.D. (1985) e Appalachian collage, in Gee and Sturt. p. 1217. LaRocque, A., Dubois, J-M.M., and Leblon, B. (2003) Characteristics of late-glacial ice-dammed lakes reconstructed in the Appalachians of southern uebec. uaternary International, 99, 73. Lauritzen, S-E. (1990) Tertiary Caves in Norway: a Matter of Relief and Size. Cave Science 17 (1), 31. Marshall, S.J., Marshall, S.J., James, T.S., and Clarke, G.K.C. (2002) North American ice sheet reconstructions at the Last Glacial Maximum. uaternary Science Reviews, 21, 175. Mrner, N-A. (Ed.) (2003) Paleoseismicity of Sweden: a novel paradigm. Stockholm University, 320 p. Palmer, A.N. (1991) Origin and Morphology of limestone caves. Geological Society of America Bulletin, 103, 1. Parent, M., and Occhietti, S. (1999) Late Wisconsinan deglaciation and glacial lake development in the Appalachians of southeastern uebec. Gographie Physique et uaternaire, 53 (1), 117-135. uick, P.G. (1994) Vermont Caves: a geologic and historical guide. Paleoow Press, USA. 74 p. Rayburn, J.A., Franzi, D.A., and Knuepfer, P.L.K. (2007) Evidence from the Lake Champlain Valley for a later onset of the Champlain Sea and implications for late glacial meltwater routing to the North Atlantic. Palaeogeography, Palaeoclimatology, Palaeoecology, 246, 62-74. Ridge, J.C., and Larsen, F.D. (1990) Re-evaluation of Antevs New England varve chronology and new radiocarbon dates of sediments from glacial Lake Hitchcock. Geological Society of America Bulletin, 102, 889-899. Ridge, J.C., and 7 authors. (1999) Varve, paleomagnetic, and 14C chronologies for late Pleistocene events in New Hampshire and Vermont (USA). Gographie Physique et uaternaire, 53 (1), 79-106. Rubin, P.A., Engel, T., Nardacci, M., and Morgan, B. (2002). Geology and paleogeography of Mount Desert Island and surrounding area, Maine in ML Nardacci (Ed.). A guide to the geology and caves of the Acadian coast: 2002 NSS Convention, Camden Maine. 47.

PAGE 257

Speleogenesis 862 2009 ICS Proceedings 15th International Congress of Speleology Stone, B.D., and Borns, H.W. (1986). Pleistocene glacial and interglacial stratigraphy of New England, Long Island, and adjacent Georges Bank and Gulf of Maine. uaternary Science Reviews, 5, 39. ieler, E.R., and 6 authors. (2007) A catastrophic meltwater event and the formation of he Hudson Shelf Valley. Palaeogeography, Palaeoclimatology, Palaeoecology, 246, 120-136. Van Staal, C.R., Dewey, J.F., MacNiocaill, C., and McKerrow, W.S. (1998) e Cambrian-Silurian tectonic evolution of the northern Appalachians and British Caledonides: history of a complex, west and southwest Pacic-type segment of Iapetus. Geology Society of London Special Publication, 143, 199.

PAGE 258

15th International Congress of Speleology Speleogenesis 863 2009 ICS Proceedings THE GENERAL MODEL OF CAVE DEVELOPMENT IN THE METALIMESTONES OF THE CALEDONIDE TERRANESTT Rev EV OR FAulk ULK Ne E R Limestone RResearch Group, GEES, University of Birmingham, c/o Four OOaks, Wilmslow Park NN orth, W ilmslow, Cheshire, SK9 2BDD, UK A general cave development model is proposed that applies throughout the metacarbonates of the non-arctic Caledonide terranes in Scandinavia, New England (USA) and the British Isles, which have comparable geological structures and which experienced similar uaternary glacial events. It builds on a previously-reported four-stage process for the inception and development of >1000 caves in the repeatedly-glaciated metalimestones of Central Scandinavia. e rankings of maximum and mean cave length and vertical range, and mean cave cross-section are commonly in the same order for each of ve main Caledonide regions, and this ranking order is similar to that of local icesheet thicknesses at the Last Glacial Maximum, local Holocene uplis and maximum relief dierences. It is therefore concluded that the main control on the extent of karstication in the non-arctic Caledonides is the thickness of the local Pleistocene icesheets. us, the greater karstication in Northern Scandinavia arose partly because the thicker icesheets and the higher mountains caused greater deglacial and neotectonic seismic activity. is produced longer and deeper inception fractures and caused deeper deglacial ice-dammed lakes to form that enabled underlying fractures, conduits and cave passages to be enlarged by phreatic dissolution for longer periods of time, and sporadically over more glacial cycles.1. Introduction e rocks of the metamorphic Caledonides derive their composition and structure from a highly-complex system of mountain building associated with the plate tectonic opening and closing of the Iapetus Ocean, from Late Precambrian to Mid Palaeozoic times: the Caledonian Orogeny (Gee and Sturt, 1985). Aer the nal thrusting over older basement rocks in the early Devonian, the CaledonianAppalachian fold and thrust mountain belt formed a continuous linear chain extending some 10000km from what is now Spitsbergen to the modern Gulf of Mexico. Subsequent orogenies and the later opening and spreading of the Atlantic Ocean caused it to be broken up into some 20 geographically-dispersed terranes, which now reside on both sides of the Atlantic (Barker and Gayer, 1985). In Scandinavia, the tectono-stratigraphic structure comprises four allochthons that overthrust and rest unconformably on the Baltic Shield. e more westerly nappes were subjected to deeper subduction and highergrade metamorphism, so that the grade of the nappe pile generally increases from sub-greenschist facies at the base up to medium amphibolite facies at the top, in an E-W direction. e whole region was covered by an ice sheet 2km thick at the Last Glacial Maximum (LGM). e non-arctic terranes in Scandinavia, Shetland, Scotland, Ireland, and New England contain metamorphic carbonates (marbles), most hosting karst caves. Faulkner (2005) studied the speleogenesis of 884 well-reported marble caves in Central Scandinavia (CS), concluding that these caves commonly experience a four-stage, top-down, middleoutwards (TDMO) cycle of cave development that is driven by the glacial cycle (Faulkner, 2008): 1: Rapid deglacial isostatic rebound that follows retreating ice margins causes seismicity and centimetre-scale movements, which form inception fractures to a maximum distance from the surface of one-eighth the depth of the local glacial valley (tectonic inception: Faulkner, 2006a; 2007a; 2007b). 2: Phreatic passages enlarge from inception fractures at high ow rates beneath deglacial ice-dammed lakes (IDLs) at relatively high wall-retreat rates over periods up to ~2000 calendar years, despite low temperatures and low PCO2 (deglacial speleogenesis: Faulkner, 2006b). 3: Mainly vadose passages entrench at the lowest levels during interglacials at a rate constrained by the size of the local catchment area (interglacial speleogenesis). 4: Glacial erosion removes whole caves or their upper and outer parts during the next glaciation (glacial removal), but valley-deepening in the range 15m (Lauritzen, 1990) produces ever-deeper inception fractures at the next cycle.

PAGE 259

Speleogenesis 864 2009 ICS Proceedings 15th International Congress of Speleology Excel cave databases were constructed for most metacarbonate Caledonide regions from the best information available in 2004. ese are discussed briey and their data compared with CS. From the similarities in the geological and glaciological evolution of the various non-arctic Caledonide terranes, they appear to follow similar processes, providing a general model for cave development throughout the marbles of the whole nonarctic Caledonide system.2. Northern ScandinaviaNorthern Scandinavia (NS), north of the Helgeland Nappe Complex (HNC) of CS, contains many karst caves, several being much longer and / or deeper than any in CS. To analyse this large region would be a task even greater than that for CS. Instead, the, hopefully-representative, karstic valley of Grtdal (St.Pierre, 1966) is considered, together with a few of the longest and deepest systems. e Grtdal caves commonly lie in long, linear, NS aligned, outcrops of amphibolite grade marbles that lie along the valley oor and its lower western slope. e foliation commonly dips steeply into the west side of the valley, or is vertical. e outcrop width varies up to some 700m, so that, although caves are commonly strike-aligned, passages also exploit orthogonal joint systems. is geological setting is similar to that in the HNC, except that the western mountains are permanently glaciated. Some 7 km of passages in 42 caves occur in this area, which is characterised by large, almost over t, underground streams that are sporadically too powerful to be explored. e Grtdal caves have mean dimensions much greater than those in CS (Table 1): their mean length and vertical range (VR) are twice as great, and mean cross-section (XS) is about four times greater, so that this set of caves would not t comfortably within the zones into which CS was divided. However, the deepest cave, Rnlihullet (140m), is well within the one-eighth constraint (above), because it lies in a valley some 1100m deep below a valley wall that slopes down above the cave, reducing its maximum subsurface cave distance to c. 30m. As in CS, there are roughly equal numbers of caves in each of the three hydrological classes of relict caves (which are almost all phreatic and are not fed by present allogenic recharge), mainly vadose (MV) caves (which may contain sumps), and combination caves (which contain both relict phreatic passages and levels and MV passages). e relict caves have about twice the mean length and XS as those in CS, and the combination caves have twice the mean length and VR, and ve times the mean XS. e MV caves follow similar trends, with four times the mean length. e larger mean VRs of the active caves are accounted for by the greater tectonism visible in the valley (Olesen et al., 2004), which probably arose from its considerable depth that caused greater deglacial seismicity. Similarly, the large lengths and VRs in other areas of NS arose from the greater seismic activity north of Ranaord (Dehls et al., 2000). Increased frequency and magnitude of deglacial and neotectonic earthquakes caused by deeper valleys means that the one-eighth constraint can be approached more closely in more areas, so increasing the overall lengths and densities of inception fractures, thereby providing larger frameworks in which individual cave systems can develop. Because the mountains are higher, it is also likely that caves enlarged beneath deglacial IDLs for longer periods of time in NS, resulting in a larger mean XS for the relict phreatic passages. Holocene vadose entrenchment by enhanced summer recharge from glaciers and perennial snowelds in Grtdal was much more vigorous, creating the very large stream passages. e roofs of many previous sumps were raised above water level by a combination of chemical and strong mechanical erosion, and resurgence sumps were lowered by the faster down-cutting of the external pocket valleys, to create more caves of the late interglacial stage of the TDMO model (Faulkner, 2008). From the Grtdal observations and those of other northern caves, it seems likely that most of these caves developed generally as in CS, some with possible enhanced vadose entrenchment from Holocene glacial recharge. Elsewhere in NS, the shas and passages in the deepest cave, Rgge Javre Raige (>580m to a submarine resurgence), remain within 173m of the wall of a ord that is 455m deep below peaks at 1200m, and thus this cave also remains within the one eighth constraint. However, the constraint is dramatically breached in at least three deep caves. e Grekjelen / Gresprekka system is formed in high-grade complexly-folded marbles. Here, inception seems to be tectonic, with synclinal / anticlinal folding enhancing the formation of deep, probably open, joints, but this tectonic activity may have been caused by much longer-timescale, possibly aseismic, processes, such as the general upli of the Scandinavian landmass, or the spreading of the Atlantic Ocean, rather than by deglacial seismicity. us, with a completely dierent process involved, the one-eighth constraint need not apply. Such a mechanism may also contribute to the depths of Tjoarvekrajgge and Okshola / Kristihola, which occur in only medium grade low angle karsts, and which both form long and deep maze networks. e deepest cave in CS (Ytterlihullet: 180m deep) has also formed in low angle karst, although it is in high-grade marble and complies with the one-eighth constraint. us, it seems that endokarst formation in low angle marbles is more likely to favour fractures that are aligned with the foliation,

PAGE 260

15th International Congress of Speleology Speleogenesis 865 2009 ICS Proceedings CAVE CLASS and AREA No. of caves in class % of caves in class Total cave length in class (m) % of total cave length in class Mean cave length (m) Mean cave VR (m) Mean cave XS (m2) Mean cave volume (m3) RELICT Grtdal 133196313744.76.3168 C. Scandinavia 28032943713345.93.1131 S. Scandinavia 1123117 3 115.72.018 New England 8153199822255.32.675 Scotland 422874518185.91.944 Ireland 86714234 185.62.870 RC TOTALS 43534*1340113*315.73.0110 COMBINATION Grtdal 143349256735233.624.93372 C. Scandinavia 36041589957916414.54.81020 S. Scandinavia 1634345481 21617.65.91902 New England 523461896811714.83.1473 Scotland 5536273664508.92.8175 Ireland 43327666 6912.55.1314 CC TOTALS 50139*7657576*15314.55.0959 MAINLY VADOSE Grtdal 15361474209810.28.8808 C. Scandinavia 244286449 9263.92.162 S. Scandinavia 204366716 334.32.596 New England 2013900104510.73.4237 Scotland 553674418142.91.625 Ireland 00 0 0 00.00.00 MV TOTALS 35427*1023410*294.42.4100 ALL CLASSES Grtdal 423.3*73627.3*17516.313.41465 C. Scandinavia 88468.5*7488174.7*858.83.5474 S. Scandinavia 47 3.6* 4238 4.2*909.13.5692 New England 15311.9*90879.1*599.32.9234 Scotland 15211.8*42254.2*285.92.184 Ireland 12 0.9* 418 0.4*357.93.6151 GRAND TOTALS1290100100211100788.83.6437Table 1: Caledonide Caves hydrological classes and major dimensions. = % of all caves. Values in Tables 1 and 3 for areas with small sample size or lower quality data are shown in italics.

PAGE 261

Speleogenesis 866 2009 ICS Proceedings 15th International Congress of Speleology and can thus carry water to deep outlets in favourable local topography. In these cases, the fractures can act more like the inception horizons of sedimentary limestones (Lowe and Gunn, 1997), so that chemical inception may become more important than tectonic inception, especially if the limestone is only weakly metamorphosed. 3. Southern ScandinaviaSouthern Scandinavia (SS) comprises the Caledonide nappes south of CS. ese caves are reported poorly, but a database of 47 better-documented caves (out of c. 70) was constructed that includes some estimates of VRs and XSs. Nothing seems remarkable about these caves, which are assumed to follow the processes described for CS.4. New England e Caledonide area of interest in northern America comprises the northern Appalachians on the western side of Vermont, Massachusetts and Connecticut. Its >150 caves with >9km of passages are compared with CS in a companion paper (Faulkner, 2009). Deglacial seismicity is conrmed by the existence of many talus and fracture caves and by movements within the karst caves themselves, and local deglacial IDLs have also been reported. e mean length, XS and volumes are rather smaller than in CS, but the mean VR is comparable and subsurface cave distances consistently lie within the one-eighth constraint. ere are proportionately less MV caves, but they have larger XSs than those in CS, probably because this area was deglaciated c. 4000 years earlier than CS. 5. British IslesIn Scotland, the Caledonide Dalradian Supergroup (Grampian) terrane correlates with the eastern part of Shetland, with the Dalradian Supergroup in Donegal, and with a displaced Dalradian terrane in Connemara. Whereas the Grampian terrane and the Uppermost Allochthon of Scandinavia have never been correlated, it would be fair to say that they have similar tectonic status and position (R. Gayer, University of Cardi, pers. comm., 1998). is explains the great similarities between the metacarbonate outcrops and their karst caves in Scotland and those in the HNC. British and Irish glacial history followed the CS pattern but, being farther south in a more oceanic setting with smaller and thinner ice caps (especially at Shetland), the glaciations were less intense, and more dicult to interpret, and deglacial IDLs more short-lived, although perhaps more frequent. Several tectonic ssure or fracture caves support the evidence that this area also experienced seismic shocks and tectonic movements following rapid deglaciation and upli at the start of both the Windermere Interstadial and the Holocene (Davenport et al., 1989), and the author has always been successful when looking for signs of tectonic movement in the Scottish marble caves (e.g. Fig. 1). us, the formational processes are probably similar to those in CS. e Grampian Terrane also contains >150 karst caves, with >4km of passages. Although the proportion in each hydrological cave class is similar to that in CS, they have much smaller mean dimensions, with less vertical complexity. Only 12 caves are recorded in the Irish Caledonides, with a sparse written record. eir mean cave dimensions are slightly larger than those in the Scottish Caledonides, but no MV caves are recorded. Probably more caves wait to be found, and the dimensional similarities with Scotland will be strengthened. e absence of endokarst and the paucity of exokarst (with dolines <2m deep) on Shetland, conrmed during a visit in 1999, seem paradoxical. However, the greatest relief across a limestone valley is only 250m. us, the maximum depth of tectonically-produced inception fractures is only some 30m, from the one-eighth constraint, but even this is too generous, because the probable maximum thickness of ice of only 200m (Mykura, 1976) would only permit much lower intensity deglacial seismic shocks. An Figure 1: Tectonic moement in Poll Seomar. Moement of ~10cm aer the formation of the passage in angled stripe karst. Both photos by Ivan Young.

PAGE 262

15th International Congress of Speleology Speleogenesis 867 2009 ICS Proceedings even more important factor is that the sea level has been continuously rising during the Holocene, creating inland waterways (Mykura, 1976). Shetland lies in the forebulge area of both Scandinavia and Scotland, so that as these lands were depressed isostatically during each glaciation, Shetland actually rose. e process was then reversed during interglacials, with Shetland falling as Scandinavia and Scotland re-adjusted upwards. is interglacial depression of Shetland suppressed neotectonic seismicity in its immediate area (Bungum, 1989), as did the smothering eect of ice during its enforced isostatic upli (Johnston, 1987). Hence, Shetland did not experience the seismic activity necessary for the creation of inception fractures at any time during the last glaciation, could not enter the phreatic phase of passage enlargement, and has not been able to develop any conduits deeper than c. 2m during the Holocene. Mykura (1976) also reported that Shetland has fallen by at least 82m during the uaternary. us, previous glacial and interglacial conditions at Shetland were similar to those observed for the Devensian and Holocene, so that karst caves probably never developed on Shetland during that period.6. e General Caledonide Model for Cave Developmente similarities in many mean cave dimensions (Table 1) and in the numbers of entrances, cave streams and sump pools per cave (Faulkner, 2005) across all areas for each hydrological class do suggest that similar processes have operated across most Caledonide terranes, but with two extra processes applying to NS and a null-process applying to Shetland (Table 2). e rankings of maximum and mean cave length, maximum VR, and mean cave XS (Table 3, columns 2, 3, 4 and 6) are in the same order for each of the ve better-documented areas: NS (largest caves, using Grtdal as an example for mean cave dimensions); CS; New England; Scotland; and Shetland (zero caves). Only the ranking of the mean cave VR (column 5) is slightly dierent, probably because of the meticulous recording of small caves in CS. is uniformity in ranking of the major cave dimensions suggests that the dierences between the Caledonides Observations Processes and Controls Scandinavian -Northern Has the longest, deepest and largest caves, as exemplied in Grtdal. Mean cave dimensions for the whole of NS are unknown, but they are probably greater than those of CS. Commonly, high local relief caused large deglacial seismic shocks. Exceptionally, very deep tectonic movement violates the oneeighth constraint, as can also occur in extensive low angle karsts (which may utilise inception horizons). Locally, recharge from permanent glaciers produced larger relict passages, greater vadose entrenchment and fewer sumps. Scandinavian -Central e authors main study area (Faulkner, 2005; 2008). Provides the standard four-stage process against which other areas may be compared Scandinavian -Southern Less well studied, and small sample size. No caves in Vertical Stripe Karst. Follows the standard process, with controls similar to those in CS. Laurentian -New England Reduced cave dimensions (except VR) and proportionately more Relict and less MV caves compared with CS. No caves in VSK. Follows the standard process. is may also apply to caves in metacarbonates of the Grenville-age Canadian Shield. British -Scotland Reduced cave dimensions, compared with CS and New England. No caves in low angle karst. Follows the standard process, but with smaller phreatic enlargements under shorter-lived IDLs, and less vadose entrenchment from shorter spring melts. e one-eighth relationship is too generous in the east, where the Devensian icesheet was not continuous. British -Ireland Less well studied, and small sample size. Mean dimensions comparable with Scotland. Probably follows the standards process. Unknown reason for apparent absence of MV caves. British -Shetland ere are long metacarbonate outcrops in this terrane, but no karst caves. Low relief, thin icesheets, and continual interglacial isostatic depression has suppressed tectonic inception and always prevented cave formation. Table 2: Caledonide caves and karsts major observations, processes and controls. Processes and controls in bold are additional to the main processes that apply in Central Scandinavia.

PAGE 263

Speleogenesis 868 2009 ICS Proceedings 15th International Congress of Speleology ve areas are greater than the dierences within each of them. e placement of SS and Ireland is more dicult. Both these areas comprise several geographically-dispersed distinct regions, which would be better considered individually. However, if the quality of this data is improved in future, these two areas would probably t in the rank order shown in Table 3, with the mountainous Jotunheimen area ahead of the rest of southern Scandinavia, and Donegal ahead of Connemara in Ireland. e nal demonstration of these relationships is presented in Table 3, columns 7. ese show that the ranges of local icesheet thicknesses at the LGM, the local Holocene uplis, and the maximum topographic relief dierences also follow similar ranking orders. e prime conclusion is that the main control on the extent of karstication in the non-arctic Caledonides is the maximum weight and thickness of each of the various Pleistocene icesheets. e icesheets caused isostatic depression, and therefore the previous thickness determined the amount of postglacial upli. e greater this was, the faster was the initial acceleration of the upli, by Hookes Law. is, in combination with the change of local relief experienced by a retreating ice margin determined the magnitude of local deglacial earthquakes. ese in turn controlled the density and the depth of tectonic inception fractures that were wide enough to permit enlargement to explorable cave passages within the timescales of the deglacial and interglacial hydrological regimes that the karst subsequently experienced (Faulkner, 2006b). Hence, because the Pleistocene glacialinterglacial cycles were approximately synchronous globally, these timescales were similar for all Caledonide terranes. eir cave developments therefore kept in step and total explorable cave lengths, VRs and XSs are functions of previous local icesheet thicknesses. A supplementary mechanism for phreatic enlargement is that the more ice there was to melt at the end of each glaciation, the longer the caves and fractures remained submerged under IDLs, and the more water owed through them. Greater ows in turn caused a greater widening of cave passages, increasing their XS and permitting smaller conduits to enlarge to explorable size, so increasing the measured length of each cave. Caves in the Caledonide areas outside NS and CS tend to be simpler and smaller, indicating less development stages over shorter timescales. It is concluded that all these caves were commonly produced by similar processes within the TDMO model, but their more epigean nature, the rarity of both relict MV caves and complex passage tiers, their commonly-smaller phreatic passage XS, and absence of large speleothems (e.g. Fig. 2) suggests that their relict and combination caves are commonly single-cycle caves that only enlarged during and aer the Weichselian / Wisconsin / Devensian deglaciation. ey have few older passages: most previously-existing higher passages were eroded away during this (and earlier) glaciations, because valley deepening is greater than mean cave depth. e MV caves are half-cycle caves, which enlarged to present dimensions only aer deglaciation and during the Holocene. e caves may perhaps be described as epikarstic, but in this case it is an epikarst that can support a wide range of cave morphologies, without any passages lying in the deeper unfractured rock mass. e concept of a watertable has little validity in the metamorphic Caledonides, because water storage and ow is contained within discrete fractures, conduits and cave passages. In contrast, CS and especially NS have many longer, deeper, larger, more complex and older caves. Some caves in NS exceed the constraints of the standard process and there are probably more multiCALEDONIDE AREA Max. cave length (m) Mean cave length (m) Max. cave VR (m) Mean cave VR (m) Mean cave XS (m2) Max. local icesheet thickness (m) Local Holocene upli (m) Max. relief dierence (m) N. Scandinavia17000?580??2000401655 (Grtdal) 175 16.313.420001401174 C. Scandinavia5600851808.83.51800120900 New England90059829.32.9150060790 S. Scandinavia 56090469.13.5 12001201200 Scotland34028485.92.11000980 Ireland 22035257.93.6 500? 15?>300 Shetland 200-9250Table 3: Caledonide caves, glaciation, upli, and local relief dierence.

PAGE 264

15th International Congress of Speleology Speleogenesis 869 2009 ICS Proceedings cycle combination caves than in CS, as supported by the greater ages of some dated speleothems (up to 800ka: Lauritzen, 1993). e oldest Caledonide marble cave is perhaps Rgge Javre Raige, where upper passages may have survived for ~1Ma. Such passages may still exist in NS because reduced glacial erosion, away from the saddle area of CS that focussed E-W ice streaming, preferentially protected the older, higher, passages. ReferencesBarker, A.J., and Gayer, R.A. (1985) Caledonide Appalachian tectonic analysis and evolution of related oceans, in e TT ectonic Evolution of the Caledonian AA ppalachian OOrogeny, Gayer, R.A. (Ed.), Vieweg, p. 126. Bungum, H. (1989) Earthquake occurrence and seismotectonics in Norway and surrounding areas, in Earthquakes at NN orth-A A tlantic Passive Margins: N N eotectonics and Postglacial RRebound, Gregerson S and Basham P.W. (Eds), p. 501. Davenport, C.A., Ringrose, P.S., Becker, A., Hancock, P., and Fenton, C. (1989) Geological investigations of late and post glacial earthquake activity in Scotland, in Earthquakes at NN orth-A A tlantic Passive Margins: N N eotectonics and Postglacial RRebound, Gregerson S., and Basham P.W. (Eds), p. 175. Dehls, J.F., Olesen, O., Bungum, H., Hicks, E.C., Lindholm, C.D., and Riis, F. (2000) 1:3000000 Neotectonic map: Norway and adjacent areas. 1:3,000,000. Geological Survey of Norway. Faulkner, T.L. (2005) Cave inception and development in Caledonide metacarbonate rocks. PhD esis. University of Hudderseld. Faulkner, T. (2006a) Tectonic inception in Caledonide marbles. Acta Carsologica, 35 (1), 7. Faulkner, T. (2006b) Limestone dissolution in phreatic conditions at maximum rates and in pure, cold, water. Cave and Karst Science, 33 (1), 11. Faulkner, T. (2007a) e one-eighth relationship that constrains deglacial seismicity and cave development in Caledonide marbles. Acta Carsologica, 36 (2), 195-202. Faulkner, T. (2007b). e hydrogeology of crystalline rocks as supporting evidence for tectonic inception in some epigean endokarsts. Cave and Karst Science, 33 (2), 55-64. (For 2006). Faulkner, T. (2008) e top-down, middle-outwards, model of cave development in central Scandinavian marbles. Cave and Karst Science, 34 (1), 3 3-16. (For 2007). Faulkner, T. (2009) e speleogenesis of the New England marble caves. Proceedings of the eenth International Congress of Speleology. Gee, D.G., and Sturt, B.A. (Eds.). (1985) e Caledonide Orogen Scandinavia and Related Areas. John Wiley, 1250 p. Hossack, J.R., and Cooper, M.A. (1986) Collision tectonics in the Scandinavian Caledonides. Geological Society Special Publication, (19) 287. Johnston, A.C. (1987) Suppression of earthquakes by large continental icesheets. Nature, 330, 467. Lauritzen, S-E. (1990) Tertiary Caves in Norway: a Matter of Relief and Size. Cave Science, 17 (1), 31. Figure 2: Speleothems in Poll Seomar, Scotland. e small sizes and the agility of the stalagmites and stalactites suggest formation during the Holocene.

PAGE 265

Speleogenesis 870 2009 ICS Proceedings 15th International Congress of Speleology Lauritzen, S-E. (1993) Natural environmental change in karst: the uaternary record. Catena Supplement, 25, 21. Lowe, D.J., and Gunn, J. (1997) Carbonate Speleogenesis: An Inception Horizon Hypothesis. Acta Carsologica, 26/2 38, 457. Mykura, W. (1976) British Regional Geology: Orkney and Shetland. HMSO. 149 p. Olesen, O., Blikra, L.H., Braathen, A., Dehls, J.F., Olsen, L., Rise, L., Roberts, D., Riis, F., Faleide, J.I., and Anda, E. (2004) Neotectonic deformation in Norway and its implications: a review. Norwegian Journal of Geology, 84, 3. St.Pierre, D. (1966) e caves of Grtdal, northern Norway. Trans. Cave Research Group of Great Britain, 8 (1), 1.

PAGE 266

15th International Congress of Speleology Speleogenesis 871 2009 ICS Proceedings FLOW DISTRIBUTION AT EARLY STAGE OF KARSTIFICATION AND 3D GEOMETRY OF CAVE SYSTEMSM. FILIPPON ON I1 and P.-Y. JEANN ANN INN2 1Labo. de gol. de lingnieur et de leniron. (GEO O LEP), Swiss Fed Institute of TT echnology Lausanne (EPFL), Switzerland2Swiss Institute of Speleology and Karstology (SISKA A), La Chaux-de-Fonds, Switzerland Simple groundwater numerical modeling allowed us to understand the inuence of the primary permeability distribution on the ow conditions during early speleogenetic phases and to recognize the crucial role of the geometry of the inception horizons and of the landscape evolution for the nal geometry of the cave system. Results of the numerical simulations indicate that under (deep) phreatic conditions the vertical distribution of groundwater ow mainly depends on the permeability distribution and is independent of the geometrical setting as well as on the distance between inception horizons. erefore, karstication may take place at any depth within the rock mass but is mostly concentrated along horizons with the highest permeability. e hydrologic behavior changes dramatically when an inception horizon becomes close to the spring area, for instance aer the incision of a valley oor. Under these conditions the ow is no longer controlled only by the permeability distribution but also by the distance between spring area and inception horizons. Flow is increased in a zone close to the spring, which is typically some tens of meters thick and some hundred meters wide. e obtained ndings allowed us to propose speleogenetic zones with characteristic hydraulic properties and conduit characteristics. is is an important step toward a probabilistic prediction of karst occurrences.1. IntroductionAnalyses of the 3D geometry of some of the largest conduit networks in the World (almost 2000 km of analyzed cave conduits) showed that the development and position of karst conduits under phreatic conditions is related to a restricted number of inception horizons (e.g., Filipponi et al. 2008). An inception horizon is a part of a rock succession (usually in the order of some centimeters to decimeters) that is particularly susceptible to the eects of the earliest cave-forming processes by virtue of physical, lithological or chemical deviation from the predominant carbonate facies within the surrounding sequence (Lowe 1992, 2000). Probably fewer than 10% of the existing bedding partings of a limestone sequence are inception horizons but guide more than 70% of the phreatic conduits (Filipponi et al. 2008). It is also clear that the inuence of these horizons on the 3D geometry of cave systems is high (Fig. 1) (Filipponi and Jeannin 2008a). Permeability measurements on such preferential karstied stratigraphic horizons demonstrated that it is possible to identify distinct inception horizons that have a slightly higher primary permeability than the surrounding rock mass, as well as another type of inception horizon with lower permeability (Filipponi and Jeannin 2008b; Filipponi et al 2009). Also we can assume that the primary permeability distribution is one of the main factors controlling the early karstication and therefore determining the later geometry of the karst system (Filipponi 2009; Filipponi and Jeannin 2009). e inception horizon hypothesis (Lowe 1992, 2000) distinguishes between three dierent phases of speleogenesis (Fig. 1). e karstication of inception horizons is supposed to start during the so-called phase of cave inception; it can be dened as starting as soon as the permeability of the rock mass increases steadily due to dissolution processes (Filipponi and Jeannin 2008b; Filipponi et al. 2008). One may expect that, at this early stage, hydraulic gradients are low and not directly controlled by well-dened boundary conditions. is may correspond to unexposed limestone, to exposed limestone in at terrain, or to parts of a limestone mass located below a well-dened active karst system. Flow is laminar and poorly organized. In other words, dissolution is low, slow, diuse, and distributed within the whole volume of the phreatic zone (deep phreatic setting). In this context however, some horizons tend to increase their permeability slightly faster than others preparing the later development of karst conduits. ey are becoming inception horizons. When hydraulic gradients become strongly controlled by the respective positions of the

PAGE 267

Speleogenesis 872 2009 ICS Proceedings 15th International Congress of Speleology recharge/discharge areas, ow becomes organized in order to drain the aquifer. us the ow selects a few horizons which provide the weakest resistance to ow. From this point on, conduit development is faster and organized. is phase is called the phase of cave gestation. Aer a given time, some ow paths reach a sucient size (about 1 cm in diameter) for turbulent ow to occur, which strongly increases the dissolution kinetics. Conduits can thus reach human size within a few thousands of years (White 1988; Dreybrodt and Gabrovek 2000; Kaufmann 2002; Palmer 2002). is is the phase of cave development. e inception horizon hypothesis is based mainly on eld observations and speleogenetic hypotheses on conditions occurring in the early phase of speleogenesis. Direct verications being dicult, and the primary permeability distribution being signicant, groundwater ow modeling can help us understand how caves could develop in these conditions. is paper presents the results of simple hydrologic numerical modeling and addresses the following questions: What is the inuence of the primary permeability distribution on ow conditions within the inception phases? What characterizes the change between cave inception and gestation phases? What is the inuence of the geometry of the inception horizons and of the landscape evolution on the nal geometry of the cave system?2. Method To better understand the inuence of primary permeability distribution to the pre-karstication and cave inception phases, we analyzed dierent scenarios with 2D vertical nite-element ow models (FeFlow). ese model scenarios represent simplied generic settings that were not designed to account for features at specic sites, but were still useful to gain general insight into the ow conditions during the inception or gestation phases of speleogenesis. e design of the selected scenarios was done within the inception horizon concept, assuming that within a rock mass some stratigraphic horizons are particularly susceptible to the eects of the earliest cave-forming processes due to a slightly higher (or lower) primary permeability than the surrounding rock mass (e.g. Filipponi and Jeannin 2008b). erefore our numerical model consists of a homogeneous rock mass that is pervaded by horizons with a permeability contrast with respect to the surrounding limestone. e occurrence of fractures was neglected, with the assumption that at early stages, as well as at depth, major fractures have only a very low frequency (Hillis 1998; Ortega et al. 2006). Our basic model (e.g., Fig. 2) consists of a 5000 m long limestone block 500 m thick. e width is divided into a 800-m-long spring area and a 4200-m-long recharge area. e elevation of the spring area in the model is changed in the various scenarios, ranging between 350 and 260 m to dene dierent hydraulic heads at the spring area. e recharge area is usually held at 500 m. At the recharge area we keep the head constant at the land surface because we assume that at early stages of karstication the drainage Figure 1: Schematic 3D model of a karst conduit system: e geometry of phreatic conduits is determined by inception horizons, joints, and faults as well as by the hydraulic gradient. e inception horizon hypothesis expects that the karstication of a rock mass pass through dierent speleogenetic phases, whereas dierent parts of the rock mass are in dierent karstication states at the same time.

PAGE 268

15th International Congress of Speleology Speleogenesis 873 2009 ICS Proceedings capacity of the rock is lower than the precipitation rate, so the water table will stay near the surface. No-ow conditions are assumed for all other boundaries. e permeability of the limestone block was assumed to be constant and isotropic, the only exceptions being two (inception) horizons with thicknesses of 1 m. eir permeability is varied relative to the permeability of the rock mass with factors of 0.01; 0.1, 10, and 100. Real-world inception horizons have a thickness of some centimeters to decimeters (e.g. Filipponi 2009), but because of mesh-generation diculties, 1 m is a reasonable compromise. e permeability of the rock mass was set at 10-4 m/s to reduce computing time. In the real world the permeability is about 10-9 m/s; the contrast between primary permeability of the modeled inception horizons and the surrounding rock mass was up to to % (Filipponi 2009; Filipponi and Jeannin 2008b; Filipponi et al. 2009). However, note that the absolute value of permeability has only a subordinate role; more signicant are its contrasts to the inception horizons. We analyzed four geometric congurations of horizons: horizontal, dipping toward the spring, dipping in the opposite direction, anticline, and syncline structure. For each conguration a series of simulations has been run with dierent combinations of permeability distribution, distance between inception horizons, and distance to the spring area. We evaluated more than 100 simulations. Evaluation was based on head and ow distribution at steady state. 3. Results A rst set of numerical simulations was run to understand the inuence of the contrast between the (inception) horizon permeability and the rock mass in (deep) phreatic settings. erefore we used a simple geometric conguration: spring area at 350 m, recharge area at 500 m, horizon 1 at 100 m, and horizon 2 at 250 m. e permeability of the horizon was varied in the dierent runs between 0.01 and 100 times the permeability of the surrounding rock mass. AA s e xpected, results of the niteFigure 2: Selected model scenario with one horizontal inception horizon at 250 m (permeability = 100 times that of the surrounding rock mass), and a second inception horizon at 100 m (permeability = 10 times that of the surrounding rock mass). Spring area at model altitude is 350 m and recharge area is at 500 m. e plots illustrate the distribution of hydraulic head (top), ow tracks (middle) and ow rate along the inception horizons as well as at the rock mass at 200 m (bottom) at steady state. e model scenarios show that the vertical ow distribution is proportional to the permeability distribution and is independent of the order of the inception horizons.

PAGE 269

Speleogenesis 874 2009 ICS Proceedings 15th International Congress of Speleology element model show that the vertical distribution of ow depends on the permeability distribution (e.g., Fig. 2). is means that horizons with a given ratio of permeability compared to the surrounding rock will have ow rates that are directly proportional to the permeability contrast. Refraction of ow lines causes an increased horizontal ow within high-permeability horizons but a reduction of horizontal ow within horizons of lower permeability. In other words, ow is relatively horizontal in more permeable horizons and relatively vertical in less permeable horizons. Low-permeability horizons not only cause ow lines to be steeper within those horizons (relative to the permeability boundary), but also at shallower angles in the rock mass below. is should lead to a remarkable diminution of ow in the rock mass below the horizon. is decrease depends mainly on the low-permeability horizon characteristics (thickness, permeability), but also on the length of the recharge area. However, for the thickness of the horizons (1 m) assumed in our model, as well as for the length of the recharge area (length = 4200 m), the ow diminution is not considerable. erefore we can assume that the relative position of horizons with various permeabilities has no major eect on the vertical ow distribution. Model results show that the ow rate along the (inception) horizons, as well as in the surrounding rock mass, decreases with the distance to the spring area. In other words, the highest ow rate can be observed shortly upstream from the spring area and decreases exponentially at greater distances. Further scenarios with dierent distances between horizons and dierent geometry of inception horizons show that ow distribution is independent of distance between inception horizons. e geometric setting has no inuence on the vertical distribution of ow rates. Vertical ow is goerned by the permeability distribution only. In models presented so far, we simulated ow along inception horizons that are far from the spring area. However, one can assume that the hydrologic behavior will change dramatically when a given inception horizon becomes close or even directly connected to the spring area. To this purpose, we designed a series of simulations to understand the role of distance between the spring area and the inception horizons on delineating the spring inuence zone. To describe this inuence, we analyzed the vertical distribution of ow for a series of scenarios with various distances. We assumed a rock mass with two horizontal inception horizons (at 100 m and 250 m) with the same permeability (100 times higher than the surrounding rock mass). e distance between the spring area and the uppermost inception horizon (distance d in Fig. 2) was varied between 5 and 50 m (Fig. 3). e nite-element model shows that the ow along (inception) horizons in the area near the spring is no longer proportional only to the permeability of the horizon, but decreases with the distance between the spring and the upper horizon. Flow in the deepest horizon remains proportional to the permeability dierence to the surrounding rock. erefore, ow rates along the upper inception horizon depend on the permeability as well as on the distance of the horizon to the spring area. Spring inuence decreases with the distance between the inception horizon and the spring. us a zone of spring inuence can be dened. is zone has a thickness of around 20 to 30 m below the spring area (Fig. 3). e zone of increased ow is laterally restricted and becomes thinner with distance to the spring area. Based on simulations we can estimate that the lateral elongation of this zone is on the order of hundreds of meters. It depends on the hydraulic head and the contrast in permeability between the rock mass and the more permeable horizon.4. Discussion and ConclusionFlow simulations show that ow concentration is moderate, non-selective, and controlled only by the initial permeability of inception horizons as long as they are far enough from the discharge area. e hydrologic behavior changes dramatically when one inception horizon becomes close to the spring area, for instance following incision of a valley oor. Under these conditions, ow is no longer controlled only by the permeability distribution but also by the distance between the spring area and the inception horizon. Flow increases in a zone of some tens of meters below the spring area and a few hundreds of meters laterally. Figure 3: Relationship between ow rate and distance of an inception horizon to the spring area. e ow rate is expressed in proportion to the uninuenced lowermost inception horizon. e diagram shows that the ow rate strongly increases close to the spring. From a distance greater than about 30 m the inuence of the spring becomes negligible.

PAGE 270

15th International Congress of Speleology Speleogenesis 875 2009 ICS Proceedings e size of this zone depends on the hydraulic head and the contrast in permeability between the rock mass and the more permeable horizon. From a speleogenetic point of view the zone of spring inuence can be considered the zone of cave gestation. Later, when ow within karst conduits becomes turbulent, it is the zone where phreatic cave development takes place (Figs. 1 and 4). In the area below this zone, speleogenesis is in the cave inception phase (Figs. 1 and 4). Note the dierence between the cave inception/ gestation/development phases (Lowe 1992, 2000), which represent the state of development (dissolution seams for the inception phase; karst conduits for the gestation phase; cave conduits for the development phase), and the inception/ gestation/phreatic development/vadose development zones (Filipponi 2009), which represent zones of characteristic ow conditions: (a) the cave inception zone is characterized by laminar ow under low hydraulic gradient conditions; (b) the cave gestation zone is characterized by laminar ow under higher hydraulic gradient conditions caused by the spring inuence and a gradual development of a karst conduit network; (c) the phreatic cave development zone is characterized by turbulent ow and low hydraulic gradient in conduits; and (d) the vadose cave development zone is above water table, where cave passages are airand water-lled with permanent or occasional water ow during snowmelt or rain events. In this zone the ow is ultimately controlled by gravity, i.e. mainly vertical. As water can only ow downward, vadose conduits are mainly vertical shas or sloping, meandering canyons (e.g. Lauritzen and Lundberg 2000) and are guided by inception fractures as well as inception horizons (e.g., Filipponi et al. 2008). With time, it must be expected that within the dropshaped cave gestation zone the hydraulic gradients will atten down to the spring level because of gradual development of a karst conduit network. Successively the gestation zone will move upstream. is attening happens mainly when a victor tube is large enough to allow turbulent ow (breakthrough; e.g., White 1988; Dreybrodt et al 2005) and therefore at the change between the cave gestation and cave development phases (Lowe 1992). Once the gradient is at the steep gradient moves upstream and a new gestation zone will form upstream from the previous one. e karst conduit network will thus develop step by step in the upstream direction. N N o te that the only motor for karst development is the hydraulic gradient, which varies throughout the massif and changes over short (horizontal shiing of the gestation zone) and long (valley incision) time scales. is inuence the upstream migration of the gestation zone has also been noticed in various numerical speleogenetic Figure 4: Schematic development of a karst system in time and space (vertical section) with horizontal inception horizons (IH): Dierent parts of the rock massif are in different karstication zones at the same time. Shortly aer a valley incision event (time 1.0 and 2.0) the gestation zone is located near the spring. Karst conduit network begins to develop along the inception horizons within the gestation zone. As a conduit within this zone begins to allow turbulent ow, the water table drops. e gestation zone becomes a cave development zone and a new gestation zone develops upstream. is upstream shi of the gestation and development zones produces typical cave levels.

PAGE 271

Speleogenesis 876 2009 ICS Proceedings 15th International Congress of Speleology models (e.g. Gabrovek and Dreybdodt 2001). However it was not discussed in detail and not related to the concept of inception horizons. Speleogenetic models describe an upstream migration of the turbulent ow regime and a attening of the hydraulic gradient. However, most numerical models deal with karstication along an initial fracture network, with initial openings of some 0.1 mm (equivalent to a permeability of some 106 mD) and do not consider the cave inception phase (e.g., Groves and Howard 1994; Dreybrodt and Gabrovek 2000; Palmer 2002; Kaufmann 2003; Bauer et al. 2003). e lateral shi of the gestation zone causes a step-by-step adjustment of the ow path. is numerical observation is also conrmed by cave geomorphologists (e.g., Huselmann et al. 2003; Audra et al. 2007) who described in detail the development of phreatic cave systems and observed the same type of stepwise development of new conduits from downstream towards upstream (soutirage) (Fig. 4 time 2.1). In real systems this phenomenon is reinforced by the existence of an epiphreatic zone in which hydraulic gradients are steeper than in the water below. Based on the concept of speleogenetic zones deduced from our approach, we have been able to understand the 3D cave pattern of dierent types of cave systems by considering the position and orientation of inception horizons as well as the history of landscape evolution (Filipponi and Jeannin 2008a; Filipponi 2009). A concept taking into account the position of inception horizons in a rock mass as well as the reconstruction of the hydrogeologic history oers substantial progress toward a probabilistic prediction of dissolution voids for applied proposes (e.g., Filipponi and Jeannin 2008c). AcknowledgmentThis project is supported by the Swiss National Foundation for Scientific research (project number 200020-116207/1).ReferencesAudra, P., Bini, A., Gabrovek, F., Huselmann, P., Hobla, F., Jeannin, P.-Y., Kunaver, J., Monbaron, M., uteri, F., Tognini, P., Trimmel, H., and Wildberger, A. (2007) Cave and karst evolution in the Alps and their relation to paleoclimate and paleotopography. Acta Carsologica 36:1, 53-68. Bauer, S., Liedl, R., and Sauter, M. (2003) Modelling of karst aquifer genesis: inuence of exchange ow. Water Resources Research 39:10, 1285-1295. Dreybrodt, W., and Gabrovek, F. (2000) Dynamics of the evolution of single karst conduits. In: Speleogenesis, evolution of karst aquifers, Klimchouk, Ford, Palmer, &Dreybrodt, Eds., National Speleological Society, Huntsville, Ala., p. 184-193. Dreybrodt, W., Gabrovsek, F., and Romanov, D. (2005) Processes of Speleogenesis: a modelling approach. Carsologica, ZRC Publishing, Ljubljana, 375 p. Filipponi, M. (2009) Spatial analysis of karst conduit networks and determination of parameters controlling the speleogenesis along preferential lithostratigraphic horizons. PhD esis, Swiss Federal Institute of Technology Lausanne. Filipponi, M., and Jeannin, P.-Y. (2008a) Possibilities and limits to predict the 3D geometry of karst systems within the inception horizon hypothesis. Geophysical Research Abstracts, EGU General Assembly, Vienna, Austria: (CD-ROM). Filipponi, M., and Jeannin, P.-Y. (2008b) What makes a bedding plane favourable to karstication? e role of the primary rock permeability. Proceeding of the 4th European Speleological Congress, Spelunca Mmoires, 33, 32-37. Filipponi, M., and Jeannin, P.-Y. (2008c) Prediction of karst occurrences by interpreting borehole data within the Inception Horizon Hypothesis. Sinkholes and the Engineering and Environmental Impacts of Karst 2008, Proceedings of the Eleventh Multidisciplinary Conference, Geotechnical Special Publication 183, 120-130. Filipponi, M., and Jeannin, P.-Y. (2009) Lithological Properties of Inception Horizons e key to understand cave development along bedding planes. Geophysical Research Abstracts, EGU General Assembly, Vienna, Austria: (CD-ROM). Filipponi, M., Jeannin, P.-Y., and Tacher, L. (2008) Evidence of inception horizons in karst conduit networks. Geomorphology (in press); doi:10.1016/ j.geomorph.2008.09.010 Filipponi, M., Jeannin, P.-Y., and Tacher, L. (2009) Understanding cave genesis along favourable bedding planes e role of the primary rock permeability. Zeitschri fr Geomorphologie (submitted).

PAGE 272

15th International Congress of Speleology Speleogenesis 877 2009 ICS Proceedings Gabrovek, F., and Dreybdodt, W. (2001) A model of the early evolution of karst aquifers in limestone in the dimensions of length and depth. Journal of Hydrology 240, 206-224. Groves, C.G., and Howard, A.D. (1994) Early development of karst systems 1. Preferential ow path enlargement under laminar ow. Water Resources Research 30, 2837-2846. Huselmann, P., Jeannin, P.-Y., and Monbaron, M. (2003) Role of epiphreatic ow and soutirages in conduit morphogenesis: the Brenschacht example (BE, Switzerland). Zeitschri fr Geomorphologie 47, 171-190. Hillis, R.R. (1998) e inuence of fracture stiness and the in situ stress eld on the closure of natural fractures. Petroleum Geoscience 4, 57-65. Kaufmann, G. (2002) Karst aquifer evolution in a changing watertable environment. Water Resources Research 38:6, 26.1.9. Kaufmann, G. (2003) A model comparison of karst aquifer evolution for dierent matrix-ow formulations. Journal of Hydrology 283, 281-289. Lowe, D.J. (1992) e origin of limestone caverns: in inception horizon hypothesis. PhD esis, Manchester Polytechnic, United Kingdom. Lowe, D.J. (2000) Role of stratigraphic elements in speleogenesis: the speleoinception concept. In Speleogenesis, evolution of karst aquifers, Klimchouk, Ford, Palmer, Dreybrodt, Eds., National Speleological Soc., Huntsville, p. 65-76. Ortega,O.J., Marrett, R.A., and Laubach, S.E. (2006) A scale-independent approach to fracture intensity and average spacing measurement. American Association of Petroleum Geologists Bulletin 90:2, 193-208. Palmer, A.N. (2002) Speleogenesis in carbonate rocks. In Evolution of karst: om prekarst to cessation, Gabrovek, F., Ed., Carsologica, ZRC-SAZU, Postojna-Ljubljana, p. 43. White, W.B. (1988) Geomorphology and hydrology of karst terrains. Oxford University Press, New York, 464 p.

PAGE 273

Speleogenesis 878 2009 ICS Proceedings 15th International Congress of Speleology MORPHOLOGY OF TJOARVEKRAJGGE,THE LONGEST CAVE OF SCANDINAVIATORTOR ST TEINN FINNNN ESAND AND1 and RANRAN E CURR L2 1e TT jorve Project, Sle Solfengsvei 4, NN -0956 OOslo, NN orway. nnesand@getmail.no2DDept. Chemical Engng, Univ. of Michigan, AA nn AA rbor, MI 48109-2136. ranecurl@umich.edu Tjoarvekrjgge (Tjorve), with a surveyed length of 21,814 meters, the longest cave of Scandinavia, is found in one of four marble bands of stripe karst in Bon, some miles north of the Polar circle in Norway. e cave is a two-dimensional labyrinth system situated close to a shoulder of a U shaped valley. Morphometric and fractal analysis can be made with over 99 % of the passage dimensions. Morphometric parameters of Tjorve yield a passage density of 47.5 km/km2 and a cave porosity of 0.8 %, intermediate between the values of conned and unconned settings, and an areal coverage of 21.8 %, close to the values for conned settings. Values for the uppermost part of the cave (cave porosity: 3.6 %, areal coverage: 32.6 %) are closer to or within the values for conned settings. e values might reect a cyclic development of the cave over several glacial-interglacial cycles. Four levels in the cave can be discerned in vertical prole, possibly corresponding to ancient water tables that have been step-wise lowered in successive glacial periods. Tjorve may have developed over a long time-period, from perhaps the Tertiary. e Linked Modular Element (LME) method (Curl 1986: http://tinyurl.com/6o53kd) is applied to Tjorve to determine the distribution of cave passage sizes. e distribution of LME sizes t a power-law function from 1.8 to 5.9 m and exhibits a fractal dimension of 2.929 (s.d. 0.068), similar to Little Brush Creek Cave (LBC), Utah (fractal dimension 2.79). e proper modulus is near 1.1 m, compared to 0.6 m for LBC, indicating perhaps less complete exploration. 1. IntroductionTjoarvekrjgge (Tjorve), with a surveyed length of 21 814 meters and a depth of c. 497 m, the longest cave of Scandinavia is found in one of four marble bands of stripe karst (Horn 1937) in Bondalen, Nordland county, some kilometers north of the Polar circle in Norway. Bondalen is a north-south U-shaped valley, widened and deepened by the glaciers in the last 2.5 million years. Tjorve is situated close to the western shoulder of the valley. e marble band is 50 to 60 m thick, surrounded by insoluble mica schist. e marble dips 25 to 40 degrees to the south and southeast (following the local folding), adding depth to the cave system. e resurgence is at 84 masl, close to the valley bottom. Tjorve has ve known proper entrances. ey have no drainage area today. ere is a short cave above Tjorve, Stoppenlen (496 m long and 190 m deep), leading straight toward Tjorve, but without obvious proper connections. Tjorve has tubes, canyons, rock blocks, and clay. e tubes follow the Tjorve plane (Fig. 1) horizontal in an eastwest direction (the strike) and sloping in the dip direction. Canyons above the groundwater level also follow the dip. Large areas of the upper parts of Tjorve contain boulders, mostly from breakdown, but also injected during glaciations. e clay deposits are especially prominent in the phreatic tubes in the upper part of the cave, but can be found in most other places, including on breakdown. e survey is done to BCRA grade 5, using Suunto compass and clinometer, tape and in recent years laser meters and digital clinometers. Due to many loops and side passages, stations are placed on bedrock, boulders and clay, and are normally properly marked. e survey includes 214 loops, with an average loop closure of 2.1 %. Survey data are downloaded into an Excel le developed for the Tjorve project. Export can be done to Compass, erion and Excel workbooks for additional analysis. A total of 2 805 valid survey shots have been recorded. Over 99 % of the shots have passage dimensions, which allow morphometric and fractal analysis. 2. Morphometric AnalysisMorphometric parameters (Klimchouk, 2003) of Tjorve

PAGE 274

15th International Congress of Speleology Speleogenesis 879 2009 ICS Proceedings and parts of Tjorve (Table 1) can be calculated in dierent ways. Our denitions are: 1 Surveyed passages (column 2, 3, 4 and 5 in Table 1) use survey data: Length is given as surveyed (3D) and projected horizontally (2D). Surveyed area is the plan area of the passages (seen from above, i.e. passages situated above others (seldom found in Tjorve) are not included). Volume is the horizontal length multiplied with an elliptical cross-sectional area (from le-right and up-down (LRUD) measurements at stations). 2 Cave extent (column 6, 7 and 8 in Table 1) is the twoand three-dimensional area that the cave occupies. e two dimensional extent is a horizontal area (cave eld) calculated from a polygon surrounding the cave. e height is calculated at the location where the vertical distance between the lowest and highest point is largest (minimum value would be the highest passage). Volume is the part of the rock volume (cave eld x height) in which the cave is developed. e passage density (47.5 km/km2) and the cave porosity (1.0 %) are intermediate between the values of conned and unconned (aer Klimchouk, 2003) settings, and areal coverage (22.1 %) is close to the values for conned settings. Almost all (93 % of the length) of Tjorve lies in the Tjorve plane (Fig. 1), which has a higher porosity (2,2 %) than Tjorve. e missing 7 % is mainly the rst few hundred meters of passages from three entrances, which appear to be invasion systems (Fig. 1). Values for the uppermost and old part of the cave such as Galleries (cave porosity: 3.6 %, areal coverage: 32.6 %) are closer to or within the values for conned settings. ese intermediate values might also reect a cyclic development of the cave over several glacial-interglacial cycles. In a vertical prole (Fig. 2) one can possibly discern four levels in the cave, corresponding to ancient water tables that have been step-wise lowered by glacial erosion during the glacial periods. If this model holds true, Tjorve must have developed over a long time-period, perhaps originating in the Tertiary. Figure 1: Tjorve looking northeast (40o) and looking up a slope (+24o). From this view the Tjorve plane is a thin line. e inasion passages can be identied. (om Finnesand et al. 2007). Proper entrances are numbered. Table 1: Morphometry, derived om surveyed (proper) passages. Specic olume is the sum of passage olume divided by the sum of passage length (3D). Passage density is the sum of horizontal passage length (2D) divided by the cave eld. Areal coerage is the ratio of the sum of the horizontal passage area seen om aboe to the cave eld. Cave porosity is the ratio of the sum of the passage olume to the rock olume. Data 1993-2008.

PAGE 275

Speleogenesis 880 2009 ICS Proceedings 15th International Congress of Speleology e method by which the cave eld is calculated is of great importance to the passage density, areal coverage and cave porosity (Klimchouk 2003). We calculated the Minimum Horizontal Polygon Field dened such that any line segment of the polygon 1) is as long as possible and 2) does not cross a shot (Finnesand et al 2007). is polygon in Tjorve has 19 line segments. Perhaps more common (and easier) is to calculate the area from the smallest rectangle that includes the cave (the cave eld of Tjorve would then increase from 0.41 km2 to 0.63 km2). Klimchouk (2003) identied polygons by using a lot of line segments, the goal is to have a polygon that reasonably closely embraces the plan array of a cave. e polygon can be drawn in many ways, but in practice the cave eld would have values within 10-15 % (Klimchouk, 2003). Doing that in Tjorve, the cave eld become 0,34 km2 (75 stations), which would increase the three aected parameters by 23 %. In theory, one can increase the number of line segments in the polygon until the cave eld will be the sum of the passage area and the area within the loops which would occur in plan view. ere are still some passages to be surveyed in Tjorve, although the extent of the cave would probably not change. Average size of remaining passages are probably small, which will reduce the specic volume. e passage density, areal coverage and cave porosity would increase somewhat, in particular in the Galleries.3. Fractal AnalysisA fractal analysis of Tjorve was done to determine whether the cave exhibited self-similar fractal structure and, if so, to estimate the unsurveyable (non-proper) length and volume of the cave and possibly that of the entire karst terrain. e method used was that of Curl (1986 hereaer cited as RC) and discussed further in Curl (1999). is is done by placing virtual spherical linked modular elements (LME) of diameter cm) at survey stations and interpolating additional LMEs linearly between stations, as explained in RC (pp 776-777, Fig 7). e diameter of LME at stations is chosen as the lesser of the measured or estimated LRUD distances because this is what limits exploration and hence denes the limiting scale of the proper cave. Counts of LMEs were sorted into a histogram using equal logarithmic interval binning corresponding to 2 % dierences in with qi LME per bin centered at i. ese are shown in log-log coordinates in Figure 3. Tjorve yielded 15 768 LME between 10 and 1 305 cm. e qi fall o rapidly at smaller than at the data peak (at approximately 110 cm) because of the physical diculty or impossibility of surveying in smaller passages, and are also truncated at large because of such factors as limited strata thickness, rock strength, and extent of solution. e nearly linear slope in a range of larger than at the data peak suggests a power-law model of the form Figure 2: Tjorve prole, looking north. e bottom of the cave is closer to the observer and the top is further away. Level 1 is the river of today. Level 2 is the tubes in Down below. Level 3 is the lower part of the Galleries. Level 4 is in the main part of the Galleries (the Galleries have passages that extend towards the surface). Canyons are oen found between the levels, in particular between level 1 and 2, and level 2 and 3. Since the cave dips 25o-40o to the south, the plan map of Tjorve is quite similar to the prole map. 19 line segments are needed to identify a polygon that includes the centerline.

PAGE 276

15th International Congress of Speleology Speleogenesis 881 2009 ICS Proceedings qi c i D (1) where D is the actal dimension. is is characteristic of selfsimilar geometric fractals (Mandelbrot 1983). e data may also be plotted as the more conventional cumulative, dened as the number Q() of LME larger than a value as shown in Figure 4. Here Q() appears to also exhibit a power-law range, as in Eqn. 1 (but with a dierent constant c). Newman (2005) gives numerous examples of similar power-law distribution behaviors of number versus size, known as Zipfs Law, a Pareto distribution, and rankfrequency plots, for such phenomena as word frequencies, earthquake magnitudes, moon crater sizes and population data. e causes of power-law behavior have seldom been explained, but it is related to there being no unique dening scales for the phenomenon. e least-square slope of the apparent power-law part of the cumulative plot was used in RC (p. 778, Fig. 8) to estimate D. ere are, however, two problems with this: Q() data are not statistically independent or homoscedastic, and an upper cuto at large values of due to the factors noted previously. e rst problem has been addressed by using a maximum likelihood (ML) estimator for D from the histogram data. Geyer (2007) details the general ML theory. Newman (2005) derived an ML estimator for power-law data not truncated above. We have derived the following estimators for D and its standard deviation D for data truncated to the range l < < u where l and u are the lower and upper bounds of the chosen range of Derivations are included in Supporting Online Material: Figure 3: Histogram of LME number distribution with logarithmic interval binning of 2 % of size (cm). e apparently linear power-law range was chosen to be 184 to 591 cm for the estimation of the actal dimension of the cave. e peak is at the nominal proper modulus (1.1 m) of the cave. Figure 4: Number of LMEs, Q( larger than cm. e slope in the apparently linear range of a log-log plot approximates D, but is too high due to the truncation of Q() at large A correction for truncation is shown by line A, which has slope ) D

PAGE 277

Speleogenesis 882 2009 ICS Proceedings 15th International Congress of Speleology 1 D r Dln( r ) 1 r D 1 n qi l uln il (2) D Dn 1 D2r Dln r21 r D2 1 2 (3) where r = l/ and n is the sum of qi over the range. Eqn. (2) must be solved iteratively for the D estimator D An Excel spreadsheet and instructions for performing these calculations are in the Supporting Online Material. e second problem with using a least-squares regression for the data in Figure 4 to estimate D is the necessity of correcting for the truncation of Q() data at large : this correction is shown by line A, now with the lower slope ) D calculated by adding a derived constant to all the Q() up to D must be estimated from the histogram to apply this correction, so regression of Q() data themselves does not provide an unbiased estimate of D. e value for D is somewhat sensitive to the range (l, u) chosen because of the inaccuracy with which LRUD are measured at survey stations, oen only to the nearest integer meter in large passage. e values of L, R, U and D at each survey stations were randomized uniformly over a local interval of 5 % to reduce this eect. is adjustment is less than the precision of surveyed LRUD distances, but reduces the apparent scatter of the data in Figure 3. For l = 184 cm and u = 591 cm, with station LRUD randomization, D = 2.929 and D = 0.068. Previous analyses have reported values for of 2.79 (Little Brush Creek Cave, (LBC; Colorado: RC), and 2.5 (Stagebarn Crystal Cave, South Dakota: Curl and Nepstad 1991). e earlier applications were less thorough than the current one. e estimate of D permits estimating the total volume of non-proper cave (smaller than surveyable) if it is assumed that the known proper cave is fully connected. at is, that there exists no unknown connected cave passages larger than the proper modulus of the survey. is is true of the Menger Sponge (RC, p. 775, Fig. 6) but it is likely that if smaller passages could be explored, additional large proper passage would be found. erefore an extrapolation of the data in Figure 3 to = 0 will provide only a lower limit to the remaining volume in the karst terrain. e volume of known proper cave can be estimated from the sum of qi from l up. e estimated volume of cave below l from a theoretical extrapolation of to 0, is given by V (0, l) n Dl 3(1 r D)(3 D) (4) which gives V (0, l) 831 000 m3 (19 600 m3 known), compared to 108 800 m3 for the cave above = 1 m. 4. Discussion and ConclusionsOne of the major problems when surveying in Tjorve (as in most caves) is the oen ill-dened walls due to the sloping marble, and usually the width and height of dicult passages have been very roughly estimated. e labyrinthine nature of Tjorve, with numerous side passages, also add to the challenge. ere are still probably some more kilometers of minor passages unsurveyed and unexplored in Tjorve. Likewise, unresolved problems in the fractal analysis are the inaccuracy of passage prole measurements and the more general question of how a passage prole should be used as a local measure of passage size. e cave dened by the current LME mehod of analysis does not represent the complexity of cave passages, but rather some measure of proper size that is, an anthropomorphic size. Nevertheless the fractal analysis provides estimates of cave morphology that can otherwise not be measured. e relations between the parameters of the above morphometric and fractal analyses are unclear. A related question was asked some time ago (Curl 1963) when the modulus of a cave was rst dened: are there dening scales for speleogenesis that imprint themselves upon caves? If there were, one would expect to nd multiple peaks in the histogram of the distribution of passage size. From this and previous analyses, there appear to only be two: the size of explorers (which has nothing to do with speleogenesis) and the upper LME cuto above about 5 meters, probably due to limits in strata thickness, rock strength, and extent of solution. e proper modulus is near 1.1 m, compared to 0.6 m for LBC, indicating perhaps less complete exploration. In the power-law range there appear to be no dening scales, or

PAGE 278

15th International Congress of Speleology Speleogenesis 883 2009 ICS Proceedings so many that they overlap in a way to produce a power-law dependence of LME size. How this works has not as yet been determined. e total volume of the karst terrain including Tjorve and nearby caves, for any given modulus, cannot yet be estimated by the method of RC (p. 774, Eqn. 20) because the karst terrain has not as yet been analyzed to estimate the number and lengths of all caves in the terrain, with and without entrances.AcknowledgementsSince Tjorve was discovered by Torbjrn Doj and Johannes Lundberg 29 July 1993, surveying has been done while exploring. Acknowledgements are due to the 75 cavers from Norway, Sweden, and four other countries who surveyed the cave. ReferencesCurl, R. (1963) One the Dention of a Cave. NN ational Speleological Society Bulletin (USA A) 26 (1), 1-6. Curl, R. (1986) Fractal dimensions and geometries of caves. Mathematical Geology 18, 765-783. (http:// deepblue.lib.umich.edu/handle/2027.42/43195) Curl, R., and J.A. Nepstad (1991) Fractal analysis of Stagebarn Crystal Cave, South Dakota (abstract). P roceedings of 1991 annual meeting, NN ational Speleological Society, in NSS Bulletin, 53 (1), p. 44. Curl, R. (1999) Entranceless and Fractal Caves Revisited. Proceeding of the Karst Modeling Symposium (Karst Waters Institute), Charlottesville, p. 183-185. Finnesand, T., K. Davidsen, S. Grundstrm, B. E. Johansen, J. Lundberg and R. Solbakk (2007). Tjoarvekrajgge 1993-2007. NN o rsk Grotteblad 49, 28-71. Geyer, C.J. (2007) Stat 5102 Notes: Fisher Information and Condence Intervals Using Maximum Likelihood. (http://www.stat.umn.edu/geyer/old03/5102/ notes/sh.pdf) Klimchouk, A.B (2003) Unconned versus conned speleogenetic settings: variations of solution porosity. Speleogenesis and Evolution of Karst A A quifers 1 (2), www.speleogenesis.info. Mandelbrot, B.B., (1983) e actal geometry of nature. W.H. Freeman and company, New York,.468 p. Newman, M.E.J. (2005) Power laws, Pareto distributions and Zipfs law. 46, 323-351. (http://arxiv.org/abs/ cond-mat/0412004)Supporting Online Materialhttp://www-personal.umich.edu/~ranecurl/2009ICS/ SOM_09-0338.pdf Derivations Excel Spreadsheet

PAGE 279

Speleogenesis 884 2009 ICS Proceedings 15th International Congress of Speleology CORRELATION BETWEEN PASSAGE LEVELS IN THE BELLAMAR CAVE SYSTEM AND MARINE TERRACES SURROUNDING THE BAY OF MATANZAS, CUBAEs S Teb EB AN GRAu U GONz Z lez LEZ Queve QUEVE DO, Iv V ONNe E Vz Z Quez UEZ De E l L A T T ORRe E and Humbe UMBE RTO Fe E RNNDez EZ R R Am M Os S Speleological Society of Cuba speleomat@atenas.inf.cu; spleoivt@yahoo.com Abstract e region under study is located in a continental erosion surface named the Bellamar Surface. It is a at karstic area, at elevations between 50 and 70 m, although it can reach maximum altitudes of 100 m, and a minimum of 40 m. It constitutes the southern portion of the Bay of Matanzas, rimmed along its north ank by a system of marine terraces. A system of ssures is oriented almost parallel to the coast, and it guides the development of large cavities, which shows a correspondence with the genesis of the Bay. e main cave in the area is the Bellamar Cave System, which currently contains 27 km of surveyed passages. In order to study the correlation between terrace levels and cave passages, we used 1:500 surveys of the dierent caves of the area, several generations of aerial and satellite images, as well as topographic maps at 1:2000 and 1:25,000 scale. is analysis allowed us to make longitudinal proles in various places, in the surface as well as in the underground galleries. ree-dimensional digital models of the area were also used, obtained by the digital processing of the surface and the underground survey.

PAGE 280

15th International Congress of Speleology Speleogenesis 885 2009 ICS Proceedings HOW KARST WORKS IN GRAND CANYON, ARIZONA, USACARO ARO L AA. HILL and VICTOR TOR J. PO O LYA A KD Department of Earth and Planetary Sciences, University of NN ew Mexico, 200 Yale Bld., AA lbuquerque, NN M 87123, USA A Caves in Grand Canyon, Arizona, USA fall into two main categories: unconned (vadose) and conned (artesian). is study focuses on the artesian caves of the Redwall-Muav aquifer, where groundwater was conned by the overlying series of Supai rocks and by the underlying Bright Angel Shale. Discharge for artesian groundwater was, as it is today, primarily from the Redwall Limestone where it has been incised by the main canyon or its tributaries and where it has converged along a structural low or fault. Descent of the water table (potentiometric surface) over time is recorded by seven mine/cave episodes: (1) Cu-U ore episode, (2) iron-oxide cave episode, (3) cave-dissolution episode, (4) calcite-spar lining episode, (5) mammillary episode, (6) replacement-gypsum episode, and (7) subaerial-speleothem episode. Grand Canyon artesian caves are the result of: (1) being formed under conned conditions where the Redwall aquifer has carried most (>95%) of the water in the total system, (2) the mixing of epigenic (upper world) waters with hypogenic (lower world) waters so that undersaturation is achieved, and (3) discharge toward spring points that have reorganized and adjusted with respect to ongoing canyon incision.1. IntroductionGrand Canyon of Northwest Arizona (Fig. 1) is located within the Colorado Plateau Province of western North America. e stratigraphic section exposed by Grand Canyon is over 1.6 km (1 mile) deep and consists of rocks of Proterozoic to Mesozoic age. e elevation of the Colorado River at Lees Ferry (Mile 0) is 960 m (3116 ) and at Lake Mead (Mile 278) is 353 m (1157 ), which provides an average gradient of 2.5 m/km (8.3 /mile). Most of the tributaries draining into the Colorado River along its Grand Canyon section are intermittent. Where ow is perennial, it is supplied by groundwater (spring) discharge. e aquifer system in Grand Canyon is and has been for millions of years karstic. In our ten year-long study of the canyon, we have visited over 50 caves on both the North and South Rims of the canyon, along the Colorado River, and in the region surrounding the canyon.2. Past Worke person who has done by far the most work on the karst hydrology of Grand Canyon is Peter Huntoon. e primary water-bearing unit is the Redwall-Muav aquifer, where the Redwall and Muav Limestones behave as a single hydrostratigraphic unit, together discharging a vast amount of water. e Redwall-Muav aquifer contains both vadose (unconned) and artesian (conned) caves (Huntoon 2000a, b).2.1. Unconned or vadose caves Grand Canyon caves formed under unconned hydrologic conditions are simple linear drains in the vadose zone where water recharges on the Kaibab Plateau and moves down along faults or master joints usually (but not always) to discharge from the base of the Muav Limestone and above the Bright Angel aquiclude. Some examples of North Rim, Kaibab Plateau vadose caves are Tapeats, under River, and Roaring Springs, which discharge along the south edge of the North Rim/Kaibab Plateau. Vadose caves are hydrologically active and have water owing through them. ey do not contain phreatic speleothems such as calcite spar or water-table speleothems such as mammillaries, as do artesian caves. 2.2. Conned or artesian cavese majority of Grand Canyon artesian caves are found in the Redwall Limestone, primarily in the Mooney Falls Figure 1: Grand Canyon of northern Arizona, USA. LF = Lees Ferry, C = Conuence, DV = Desert View, GV = Grandview monocline, UGG = Upper Granite Gorge, KP = Kanab Point, LGG = Lower Granite Gorge.

PAGE 281

Speleogenesis 886 2009 ICS Proceedings 15th International Congress of Speleology Member. Modern conned-aquifer caves include those forming today in the Marble Canyon area where the Colorado River is near Redwall Limestone level. Relict conned-aquifer caves formed like modern conned caves (i.e., under artesian conditions), but have been dissected and dewatered by canyon erosion. ey are now exposed in the Redwall Limestone high above the Colorado River. Most artesian caves are not extensive (most are less than a kilometer long) and they are usually horizontally conned within the Mooney Falls Member of the Redwall Limestone. None of these caves contain deep vertical passages. Over the past decade we have studied the artesian caves, and not the vadose caves, because it is only in artesian caves that speleothems are preserved that record the descent of the water table and incision of Grand Canyon over time (Hill et al. 2001; Hill and Polyak 2005; Polyak et al. 2008).2.3. Groundwater basinsHuntoon (1974, 2000b) identied four modern groundwater basins in the central-eastern part of Grand Canyon: the Cataract, Black Mesa, Kaiparowits, and Kaibab Plateau. Huntoon (2000a, b) established two very important hydrologic principles for the groundwater basins in Grand Canyon: (1) the discharge for artesian spring water is from the Redwall Limestone where it has been incised by the main canyon or its tributaries, and (2) the discharge is along a structural low or fault. In addition, Huntoon (1995) applied another principle of karst hydrology to Grand Canyon: ow in karst aquifers can cross faults and folds, move opposite to dip, and go under or through structures as it pursues a path along the steepest hydraulic gradient to discharge. Still another important principle involves the geochemical aspects of groundwater. Saturated water cannot dissolve caves: they must be dissolved by water undersaturated with respect to calcite. is is especially true in a limestone-karst setting such as Grand Canyon where saturated artesian water converges on springs from large-area groundwater basins. It is only at points where the mixing of dierent geochemical waters has achieved undersaturation that the artesian caves of Grand Canyon have formed. Another important hydrologic model for Grand Canyon was introduced by Crossey et al. (2006), and this was their concept of upper world versus lower world waters. ese authors analyzed waters discharging from both above and below the Bright Angel Shale aquiclude and found vast dierences in their quality and quantity. Epigenic (upper) waters are characterized by cool temperature (<20C), high discharge, low conductivity, neutral to slightly alkaline pH, and low CO2 content. Endogenic or hypogenic (lower) waters, on the other hand, are associated with faults and typically exhibit warmer temperatures (20-35C), low discharge, high salinity, lower pH, high CO2, and mantlederived helium. Such dierences in water chemistry and temperature are typical of a hypogene speleogenesis (Klimchouk 2007). 3. Water Table Lowering as Evidenced by Cave DepositsIn a former ICS talk and paper, Hill and Polyak (2005) discussed the dierent ore and cave deposits displayed in the mines and caves of Grand Canyon: (1) Cu-U ore episode, (2) iron-oxide cave episode, (3) cave-dissolution episode, (4) calcite-spar lining episode, (5) mammillary episode, (6) replacement-gypsum episode, and (7) subaerial-speleothem episode. Since that talk, the cave episodes especially the mammillary episode have been used to determine the descent of the water table in the canyon over time. Only episodes (3), (4), (5) and (6) will be discussed in this paper since they oer information on the water table and its descent relative to canyon incision.3.1. Cave dissolution episode As convective water rises and cools, the solubility of calcite gradually increases so that small caves can dissolve in the deep solutional zone (Dublyansky 2000). It was in this phreatic regime that the artesian caves in Grand Canyon most likely formed. Figure 2 shows a graph of the solubility of calcite with depth for a thermal-water system, with an estimated maximum solubility at about 200 m below the water table. However, the depth of peak calcite solubility varies with PCO2, salinity, and temperature, so the solubility peak moves along the depth axis depending on actual conditions. e caves probably formed as rising groundwater reached its maximum solubility where it mixed with descending meteoric water.3.2. Cave spar lining episodeAs the water table descends, caves formed in the solutional zone are shied into the depositional zone because the solubility of calcite drops sharply due to the loss of CO2 (Fig. 2). Since the loss of CO2 is very slow in the deeper phreatic zone, large-spar crystals can precipitate, lining previously formed cave passages. en, as the water table drops, slow CO2 loss causes these saturated waters to precipitate their calcite. In this model, a small descent of the water table would have aected the equilibrium of the system so that solutions would have changed from aggressive to precipitative. e fact that spar linings directly overlie cave bedrock, without any type of a weathering episode in

PAGE 282

15th International Congress of Speleology Speleogenesis 887 2009 ICS Proceedings Figure 2: As conective water rises and cools, the solubility of calcite gradually increases so that small caves dissole in the phreatic solutional zone. As the water table descends, caves formed in the solutional zone are shied into the depositional zone. When this happens the solubility of calcite drops sharply due to the loss of CO2 and solutions change om aggressive to precipitative. Since the loss of CO2 is very slow in the deeper phreatic zone, large spar crystals line previously formed cave passages. Just below or near the water table in the shallow phreatic zone the loss of CO2 is much more rapid and therefore ne-grained mammillary coatings form instead of spar linings. At the water table-air interface folia and cave ras form, and just aboe the water table replacement gypsum rinds form. Aer Dublyanksy (2000). between, supports the idea of a close connection in time and origin between cave dissolution and spar precipitation. 3.3. Mammillary episode Mammillaries are water-table speleothems that coat bedrock walls and ceilings or other speleothems. e mammillaries in Grand Canyon caves are known to have formed near the water table because they are associated with two speleothem types cave ras and folia that form at the surface of the water table (Hill and Forti 1997). Figure 3 shows the type of speleothem deposited with respect to the lowering of the water table through a cave. Spar linings form when the cave is in the phreatic zone, the mammillary coatings form just below the water table, the folia and cave ras form at the surface of the water table, the gypsum rinds form just above the water table, and dripstone and owstone form above the water table aer the cave becomes air-lled. e Grand Canyon mammillary coatings are ne-grained as opposed to the coarse-grained calcite-spar linings. is ne-grained, densely packed, brous nature attests to the rapid degassing of CO2 and precipitation of calcite near the water table, and also makes them suitable for dating by the uranium-lead (U-Pb) method (Polyak et al. 2008) which is the necessary dating method because Grand Canyon mammillary ages exceed the ~600,000 YBP limit of the Useries method. Since this speleothem type is common in the artesian caves of Grand Canyon, and since they denote the approximate position of the water table (Fig. 3), Polyak et al. (2008) used them as their primary basis for dating Grand Canyon incision.3.4. Replacement gypsum episodeReplacement gypsum rinds in Grand Canyon caves oen directly overlie mammillary coatings, which sequence

PAGE 283

Speleogenesis 888 2009 ICS Proceedings 15th International Congress of Speleology Figure 3: A diagram illustrating the progression of speleothem deposition related to the descending position of the water table. Spar lines cave walls in the phreatic zone below the water table, mammillaries form just below the water table, folia and cave ras form along the surface of the water table when the water table descends to cave level, replacement gypsum rinds form just aboe the water table, and speleothems like stalactites and stalagmites form aboe the water table in the subaerial zone. e small-scale folia and cave ras are not illustrated. further supports a mammillary association with the water table (Fig. 3). Gypsum rinds up to 6 cm thick have been seen lining the walls of Grand Canyon caves or overlying earlierformed speleothems. ese rinds form from a speleogenetic mechanism rather than from a speleothemic mechanism. Speleogenetic gypsum rinds form where H2S in ascending water diuses upward into the cave air as the water table drops through the caves (Palmer 1991; Fig. 3). e oxidation of this H2S at or just above the water table forms sulfuric acid which dissolves the limestone and produces replacement gypsum as a by-product of speleogenesis. 4. Generic Model for Grand Canyon CavesGrand Canyon artesian caves are the result of: (1) being formed under conned conditions where the Redwall aquifer has transported >95% of the water in the total system, (2) the mixing of epigenic waters with endogenic (hypogene) waters, and (3) discharge toward spring points that have reorganized and adjusted with respect to ongoing canyon incision and dissection. In the following discussion refer to Figure 4 where the numbers (1) to (6) correspond to sections i to vi: (i) Surface water descends via fractures from the Kaibab Limestone down through the Supai Group of rocks to recharge the Redwall aquifer (Redwall Limestone + Muav Limestone). Because of this fracture-restricted permeability, there never was a connected water table in the units above the Redwall aquifer only a potentiometric surface determined by the Redwall aquifer artesian system. It was only when the artesian-head level reached the most karstic levels in the Redwall aquifer that a relatively at, semiconnected water table could have existed. is is where the mammillaries formed: in caves that were dissolved during an Figure 4: A generic model for cave formation in Grand Canyon. Meteoric groundwater descends along actures in the Kaibab to Supai units and recharges the Redwall aquifer. Most of the ow in the aquifer is concentrated along the Mooney Falls Member, while deeper ow is constrained downwards by the Bright Angel Shale aquiclude. Although water discharges at spring points where canyon incision has intersected the Redwall Limestone, artesian caves do not form directly at these spring points. ey form along brecciated zones of enhanced permeability and where lower world waters are ascending and mixing with upper world waters so as to achieve undersaturation with respect to calcite.

PAGE 284

15th International Congress of Speleology Speleogenesis 889 2009 ICS Proceedings earlier period of phreatic speleogenesis, but which were later subjected to water-table conditions. (ii) Once downward recharge along fractures reached the Redwall aquifer, some of this water circulated deep, but most of it stayed concentrated along the more porous Mississippian paleokarst horizon in the Mooney Falls Member. Deep circulation in the Redwall aquifer only descended as low as the top of the Bright Angel Shale because this impermeable unit prevented further downward circulation. e maximum solubility for cave dissolution was probably achieved where hypogenic ascending groundwater mixed with meteoric groundwater (Fig. 4). (iii) e Bright Angel Shale is an aquiclude that separates descending, meteoric upper world waters from ascending, hypogene lower waters (Crossey et al. 2006). Because of the presence of the Bright Angel aquiclude, most vadose caves and springs discharge to the canyon along the base of the Muav Limestone, just above the Bright Angel Shale. is impermeable shale barrier is also a prime horizon for travertine deposits in Grand Canyon, where descending, carbonate-saturated, vadose water emerges as springs along the top of the Bright Angel horizon. (iv) Since almost all Grand Canyon artesian caves have replacement gypsum rinds in them, and since sulfur isotope values on this gypsum do not match those of bedrock evaporite units in the overburden, there must have been a source of reduced sulfur from the lower world regime. e sulfur isotopes of the replacement gypsum in eastern Grand Canyon caves have negative 34S values, and the only known organic source corresponding to these light sulfur isotope values is the Chuar-Walcott hydrocarbon unit in the Precambrian basement of the eastern Grand Canyon. In the western Grand Canyon, there is no Chuar in the Precambrian crystalline basement, and the replacement gypsum rinds in these caves display positive sulfur isotope values typical of a volcanic/magmatic source. (v) Redwall artesian ow converges on springs along the Colorado River. However, these spring points are not the loci for artesian cave dissolution. e conclusion that Grand Canyon artesian caves did not form at spring outlets but rather along ow paths leading toward spring outlets is supported by the following factors: (a) ere are no features in these caves, such as scallops, that would indicate high-velocity ow from a spring. (b) e calcite spar that lines these artesian caves had to have formed under very stagnant aquifer conditions not at springs with ow rates of 410 l/sec (6500 gal/min), such as has been monitored by Huntoon (1981) at Fence Springs. (c) Mammillary speleothems, while formed near the water table, still have to precipitate from saturated water (d) It is typical of hypogene speleogenesis not to be associated with terminal discharge regimes of basinal groundwater ow systems, but with intermediate discharge limbs of these systems (Klimchouck 2007). is appears to have been the case for Grand Canyon artesian caves. Rather than at spring outlets, the locations of Grand Canyon artesian caves seem related to: (a) Permeable zones in the Redwall aquifer where water diused along a joint-dominated path to a spring outlet. One zone of maximum permeability seems to have been along the Mississippian paleokarst horizon in the Mooney Falls Member of the Redwall Limestone (Fig. 4), and a number of caves display paleokarst breccias in their walls or ceilings. (b) e caves are related to places where lower world H2SCO2-rich solutions must have ascended through the Bright Angel Shale along faults/joints because practically every artesian cave displays replacement gypsum rinds. is H2SCO2 could have helped caves dissolve in the phreatic zone by the process of mixing corrosion, and later in time it could have caused cave enlargement just above the water table by the process of condensation-corrosion, where high CO2H2S in the air weathers limestone (Palmer 2007). Let us now take our generic artesian cave and trace its evolution with respect to the lowering of the water table and canyon incision. e artesian cave shown in Figure 5 was below the potentiometric surface until the time that the canyon incised to the level of the Redwall aquifer; then once the Redwall Limestone horizon was reached, unconned water table conditions began to prevail. is incision also caused the aquifer to be breached and springs to ow towards a canyon outlet. Lets say that the mammillaries in this generic cave (formed just below the water table) have a U-Pb date of 3.5 Ma, as shown in Figure 5. us, neglecting dip and other local factors, the age of these mammillaries approximately equals the time it has taken for the canyon to incise from this 3.5 Ma level to where it exists today which concept was the basis for Polyak et al.s (2008) correlation of the age of Grand Canyon incision with the age of mammillary speleothems.

PAGE 285

Speleogenesis 890 2009 ICS Proceedings 15th International Congress of Speleology Acknowledgmentsis study was supported by NSF grant EAR-0518602.References:Crossey, L.J., Fischer, T.P., Patchett, P.J., Karlstrom, D.E., Hilton, D.R., Huntoon, P., Newell, D., and Reynolds, A. (2006) Dissected hydrologic system at Grand Canyon: interaction between deeply derived uids and plateau aquifer waters in modern springs and travertine. Geology 34:1, 25-28. Dublyansky, Y.V. (2000) Dissolution of carbonates by geothermal waters. In Speleogenesis: evolution of karst aquifers, Klimchouk, A.B., Ford, D.C., Palmer, A.N., and Dreybrodt, W. (Eds.), National Speleological Society, Huntsville, AL, Ch. 4.1.5, p. 158-159. Hill, C.A., Forti, P. (1997) Cave minerals of the world. National Speleological Society, Huntsville, AL, 2nd ed., 463 p. Hill, C.A., and Polyak, V.J. (2005) Progressive lowering of the water table in the Grand Canyon, Arizona, USA as recorded by cave and mine deposits: 14th International Congress of Speleology, Athens, Greece, Paper O-48, p. 192-196. Hill, C.A., Polyak, V.J., McIntosh, W.C., and Provencio, P.P. (2001) Preliminary evidence from Grand Canyon caves and mines for the evolution of Grand Canyon and the Colorado River system. In e Colorado R R i ver: origin and evolution, Young, R.A., and Spamer, E.E. (Eds.). Monograph 12, Grand Canyon Association, Grand Canyon, Ch. 22, p. 141-145. Huntoon, P.W. (1974) e karstic groundwater basins of the Kaibab Plateau, Arizona. Water Resources Research, 10:3, 579-590. Huntoon, P.W. (1981) Fault controlled ground-water circulation under the Colorado River, Marble Canyon, Arizona. Ground Water, 19:1, 20-27. Huntoon, P.W. (1995) Is it appropriate to apply porous media groundwater circulation models to karstic aquifers? In Groundwater models for resources analysis and management. El-Kadi, A.I. (Ed.), CRC Lewis Publishers, Boca Raton, p. 339-358. Huntoon, P.W. (2000a) Karstication associated with groundwater circulation through the Redwall artesian aquifer, Grand Canyon, Arizona, USA. In Speleogenesis: evolution of karst aquifers, Klimchouk, A.B., Ford, D.C., Palmer, A.N., and Dreybrodt, W. (Eds.), National Speleological Society, Huntsville, AL, p. 287-291. Huntoon, P.W. (2000b) Variability in karstic permeability between unconned and conned aquifers, Grand Canyon region, Arizona. Environmental and Engineering Geoscience, 6:2, 155-170. Klimchouk, A. (2007) Hypogene speleogenesis: hydrogeological and morphogenetic perspective. National Cave and Karst Research Institute, Special Paper #1, 106 p. Palmer. A.N. (1991) Origin and morphology of limestone caves. Geological Society of America Bulletin, 103, 1-21. Polyak, V., Hill, C., and Asmerom, Y. (2008) Age and evolution of Grand Canyon revealed by U-Pb dating of water-table-type speleothems. Science, 319, 13771380. Figure 5: A generic model for how artesian caves correspond to the water table and incision of Grand Canyon. Figure 5 shows the Colorado River having incised the canyon down to Redwall Limestone level, with a spring exiting om the Redwall aquifer to this incised canyon. Because the water table has dropped to cave level, this is where and when any mammillaries would have been deposited in this cave. erefore the age of the mammillaries is approximately equivalent to the time it took for the canyon to incise om this 3.5 Ma level to where it exists today.

PAGE 286

15th International Congress of Speleology Speleogenesis 891 2009 ICS Proceedings THE POSSIBLE SPELEOGENE-HYPOGENE ORIGIN OF THE WARDA IRON ORE MIME DEPOSIT (JORDAN)ST TEPHAN AN KEMPE1, AA HMAD AD AA L-MA A LA A BEH2 and HOR OR ST T -VOO LKER R HEN N SCHEL3 1Institute of AA pplied Geosciences, University of TT echnology DDarmstadt, S chnittspahnstr. 9, DD -64287 DDarmstadt, Germany, 2Hashemite University, DDepartment of Earth and Enironmental Sciences, PO O Box 150459, Zarka 13115, Jordan,3Henschel & RRopertz, AA m Markt 2, DD -64287 DDarmstadt, Germany e iron ore deposit of the historic Warda mine (District of Ajloun, northern Jordan) and its speleological importance is discussed. Former prospecting results show that the deposit has a size of 300 x 200 m with a maximal thickness of about 10 m. e Warda Iron Deposit was mined during the time of the crusades by the Arabs in the time of Saladin to supply the castle of Ajloun. e historic mine consists of two larger rooms, ca. 1000 m2 in area. Much of the mines oor is now covered with recent ood sediments (680 m2), up to over 2 m deep. e mine cuts small natural cavities with a few speleothems and a breakdown hall in limestone that may or may not have been created by the collapse of a mine cavity. Conservative calculation of the mine volume suggests that a total of about 1100 t of Fe may have been extracted. Mineralogical investigation (XRD) shows that the iron ore is goethitic/ limonitic with a noticeable hematite content. Presence of greigite may be indicated by a few XRD d-values and an S-content of >3% in one sample from the center of the deposit exposed in a modern quarry. Geochemical (XRF) analysis shows that the goethite is very pure; impurities of main elements sum up to only 1%. Textural, mineralogical and geochemical criteria suggest that the ore body could have a speleogenetic origin, i.e. deposited in a hypogene, deep phreathic setting, possibly before regional upli or even prior to the maximal burial depth. e geochemistry excludes a synsedimentary origin while the mineral-paragenesis excludes a hydrothermal origin. A possibly similar ore-body was described from the gigantic, sand-lled Lower Cretaceous cave of Wlfrath (North RhineWestphalia, Germany) (Drozdzewski et al 1998). e number of known limestone caves in Jordan is very small even though large sections of the country are underlain by carbonates. e largest one, AlDaher Cave, is also a hypogene cave forming a maze within an area of 70x70 m (Kempe et al 2006), underscoring the importance of hypogene speleogenesis for Jordan. e discussed model, developed according to the present state of knowledge, could serve as a starting point to investigate similar deposits and the Warda Iron Deposit could serve as a locus typicus for a new type of iron ore deposit not well studied yet.1. Introduction and Geologic SettingSince 2003 the authors have systematically explored caves and karst and their genesis in Jordan. While the eastern part of Jordan is a lava plateau with lava caves (Kempe et al., 2008), much of eastern Jordan is formed by Upper Cretaceous limestone. Nevertheless only a few dissolutional caves are known in Jordan at present (Kempe et al. 2006; Al-Malabeh et al. 2007). Here we report on a cavity mostly created by historic mining that extracted oxidic iron ores. Iron was an important commodity to ancient Near East cultures since the beginning of the Iron Age at 1200 r.f. Since Jordan is lacking largescale iron ore deposits (e.g., Bender 1974), small ore occurrences were mined locally and extracted at the surface in open pits or underground in short tunnels. Natural cavities, made it much easier to extract the ore. One of these areas is near Ajloun (Fig. 1), where a mine is still accessible and several entrances are found. is mine is, however, threatened by a nearby quarry and road development. Many prominent structural features are present in Jordan. ey are closely related to the regional geology and tectonics of the Eastern Mediterranean area. Several intraplate deformation phases aected the northern Arabian Plate between the Late Paleozoic and the Cenozoic. Major riing episodes occurred in the Late Carboniferous to Permian, Middle to Late Triassic, and at the end of the Early Cretaceous (e.g., Garfunkel et al. 1981; Freund et al. 1970; Barazangi 1983).

PAGE 287

Speleogenesis 892 2009 ICS Proceedings 15th International Congress of Speleology Major structural trends are N-S, E-W and NNE-SSW. Along these directions, the rocks have been substantially jointed and fractured. Field investigation shows that the Warda Iron Ore Deposit (WID) occurs as a belt extending in a NNE-SSW direction. Along this direction a fault with fault breccia is recorded. e ore occurs in Upper Cretaceous limestone of Cenomanian-Touranian age in what locally known as the Wadi Al-Sir Formation.2. Warda Iron Ore Deposite Warda Iron Ore Deposit (WID) is situated east of the Dead Sea Transform Fault at 32 15N and 35E at an altitude of about 620 m. e area was called locally Jabal Al-Aqra, i.e. bald mountain because of the former absence of trees. Now the area was reforested by the Ministry of Agriculture and it is called Jabal Al-Akdar, i.e. Green Mountain. Warda means rose in Arabic, denoting the brilliant colors of the iron ore. e pit mining exposed an ore body about 10 m thick. It occurs as a band that is about 300 m long and 150 m wide striking NE-SW to NNE-SSW. e iron ores are oxidic, mostly of brown colors with yellow and black sections. e WID seems to be the only iron ore deposit in Jordan extensively mined historically and today.3. Mine and Cave e entrance to the Warda Iron Mine is located below the road from Ajloun to the Jordan valley. A small agricultural eld is nested in a sharp W-turn in the road. e eld and a neighboring plot are partly surrounded by low clis that have several articial openings (st. 2, 5 and 6, Fig. 2). Only one of them, leading north underneath the road, gives access to an old mine (Fig. 2). e underground cavity consists of two larger, SE-NW striking, 40 m-long rooms, connected by a low crawl (st. 13). e entrance room leads steeply down over blocks to the smooth oor of the mine, 12 m below st. 1. is oor is the surface of consecutive ood deposits, washed into the mine during severe rains. Here reddish, silty sediments cover the original oor meter-deep, as shown by two archeological digs, one, in the entrance hall, being 2.2 m deep and the other, in the furthest corner of the second hall, being 1.5 m deep (Fig. 3). In the SE corner of the entrance hall, adjacent to breakdown blocks, jamming parts of the former, much larger entrance, two little mining chambers (st. 10 and 11) are found. Much of the lower part of the entrance chamber (st. 12) is developed in the massive ore body. e far end (st. 15) of the second chamber is also dug out of the massive ore body while its northern parts are dierent. To the west, a large pile of breakdown blocks leads upwards, exposing SW at ca. 40 dipping limestone strata in a large cupola. is part of the cavity does not look articial it may though be a collapse of the mine roof into a once larger and deeper mining chamber. Here lives a sizable colony of bats. Bags show that the local inhabitants mine the bat-guano for fertilizer. In the SE corner, several passages are encountered that follow natural ssures through a brecciated ore body. e walls of these ssures are sparsely covered with speleothems, showing that the miners found some small natural cavity when Figure 1: Location map of Warda Iron Deposit in the northern part of Jordan.

PAGE 288

15th International Congress of Speleology Speleogenesis 893 2009 ICS Proceedings working the ore body. ese ssures and their strike are most probably parallel to a fault, governing the valley in the cave area. e mine may have had more passages originally that are now either buried under the red dirt or by the breakdown pile. e area of the surveyed cave is 980 m2 of which up to 640 m2 seem to be covered with ood sediments (entrance hall and southern section of inner hall). 230 m2 are occupied by the breakdown hall and the ssure-determined passages cover 100 m2. Much of the mines oor is now covered with recent ood sediments (680 m2), up to over 2 m deep. 4. Results 4.1 XRD mineral identicationA total of ve samples were investigated by XRD (Table 1) (courtesy H. Hofmann Techn. Univ. Darmstadt). e following minerals were detected: Calcite, hematite and goethite. In Sample W1 a mineral is present that could not be clearly identied (d-values = 2.983, 1.906, 1.749). In review it was suggested that this mineral may be greigite (Fe2+Fe3+)S4. Its d-values (relative percentage intensities in brackets) are: 2.980 (100), 1.746 (77), 2.470 (55), 3.498 (32), 1.008 (31), 1.901 (29), 1.105 (16) (quoted from http://rru.geo.arizona.edu/doclib/hom/greigite.pdf). 4.2 EDAX chemical analysisEDAX analysis was performed on the same samples that were analyzed by XRD. Samples were scanned at a magnication of 60 times at 25KV with a Fei-uanta 200FEG and an EDAX. Spectra accumulated over 30 s were recalculated for elemental composition by the ZAF method (Genesis Soware by EDAX). e following elements were included in the calculations: C, O, Na, Mg, Al, Si, S, K, Ca, Fe (Table 2). No other element peaks occurred. e results match those of the XRD analysis. Fe is by weight the most important element ranging between 24.6 and 71.3% followed by Ca with 0.9 to 24.0 %. e oxygen concentration is not in proportion to its real content; for Figure 2: Map and cross-sections of the historic Warda Iron Ore Mine/Ajloun District. Figure 3: View of inner hall and archeological dig.

PAGE 289

Speleogenesis 894 2009 ICS Proceedings 15th International Congress of Speleology example in W3, if one takes the atom ratio of Ca (ca. 6) and subtracts 6x3 oxygen (because in CaCO3 we have three O for each Ca) then only 19 atom% of O remains. at is not enough to accommodate the 45 atom% of Fe as either goethite or hematite. is mismatch is due to the fact that light elements (C and O) are generally not well represented in the EDAX analysis. Sample W1 has a high concentration of Al and S and a noticeable K peak is present. us one could also speculate that these elements may constitute the unidentied mineral found in the XRD spectrum. At rst one thinks of potassium-alum (KAl(SO4)2x12(H2O)) or a mixture of it with Na-alum. However, neither the d-values of the XRD agree with those of K-alum (strongest d-values: 4.298 (1); 3.25(0.55); 4.053(0.45)) or Na-Alum (d-values: 4.314 (1); 2.962(0.35); 3.526 (0.14)), nor do the elemental ratios of Al/S match (i.e. they should be 0.5 instead of ca. 2 as measured for W1). us greigite currently remains the best option.4.3 XRF chemical analysisMajor elements (Table 3) were determined for ve more samples using the XRF at the Geological Institute, Erlangen University, Germany (courtesy Prof. Dr. H.J. Tobschall). Sample IC 1 is from the ash ood sediment in the historic mine. Samples IC 2, 3 and 4 are samples of the ore body in the historic mine and sample IC5-calcite is a speleothem sample from the cave cut by the mine.5. Discussion of Results 5.1 e ore bodyIn places, the WID shows a brecciated nature, possibly a consequence of the NNE-SSW striking faults in the area. Smaller ssures were lled with multi-generation calcite veins. Larger, open fractures were partly lled with speleothems. Large natural chambers may have existed within the ore body, as suggested by the irregular form of the mine and by the existence of a large breakdown cupola in limestone. Sample W1W2 W3 W4 W5 dark brown orangedark orange brownblackish brownblackish ochre blackish ochre brown Calcite XXXCalcite XXXHematite XXXGoethite XXX Goethite XXX Greigite?? XXGoethite XXCalcite XX Calcite XX Hematite XXX Hematite XHematite XGoethite XXHematite X Calcite X (Goethite) Table 1: XRD Results (XXX = most prominent mineral; XX = abundant; X = clearly detectable; ( ) = possibly present). e second line is giving the color description of the ground sample according to Michel Farbenfhrer, 2000. SampleNumber W 1 W2 W3 W4 W5 Element Wt%At %Wt%At%Wt%At%Wt% At%Wt%At% C 11.9422.303.8710.092.888.528.5018.002.656.71 O 34.3248.1122.5744.1216.7837.2732.9152.2927.3351.99 Na 0.640.620.370.500.000.000.560.620.430.57 Mg 0.380.350.941. Al 7.406.150.850.980.781.030.370.351.491.68 Si 2.021.611.511.681.001.270.530.482.372.57 S 3.562.490. K 1.170.670.320. Ca 13.967.819.577.466.715.9410.026.360.910.69 Fe 24.609.8859.7933.4971.2645.3346.1221.0063.9534.85 Total 100.00100.00100.00100.00100.00100.00100.00100.00100.00100.00Table 2: EDAX analysis Samples W1 W5.

PAGE 290

15th International Congress of Speleology Speleogenesis 895 2009 ICS Proceedings Sample IC1(sediment) RF 9863 IC 2 RF 9856 IC 3 RF 9861 IC 4 RF 9862 IC5-calcite RF 9854 Oxide % % % % % Si026.4 0.1 12.0 <0.1 <0.1 Ti020.07 0.13 0.41 0.02 <0.01 AI2030.08 0.01 0.1 0.07 0.01 Fe203 t36.77 62.29 27.31 73.76 0.07 Mn0 0.004 0.003 0.003 0.004 0.002 Mg0 0.39 0.28 0.65 0.07 0.51 Ca0 29.41 17.63 32.84 0.06 55.13 Na20 <0.01 <0.01 <0.01 <0.01 <0.01 K20 <0.01 0.01 <0.02 <0.01 0.01 P2050.033 0.061 0.068 0.116 0.004 S030.39 0.32 0.05 0.17 <0.01 Cl 0.08 0.08 0.07 0.06 0.09 F 0.02 0.25 <0.01 0.48 <0.01 LOI 27.5 15.22 27.89 10.34 43.78 SUM 101.0 96.3 101.4 84.9 99.0 Table 3: Major elements (XRF) of the studied samples. (RF = Lab. No. Erlangen; LOI = Loss On Ignition; Fe2O3t = total iron). Bold = highest values; italics = lowest values of ore samples (excluding IC5-calcite).5.2 Geochemistrye major element composition of the WDI (Table 3) shows that it is mainly characterized by high Fe2O3 concentrations, ranging from 27.3 to 73.8 wt% in the rock samples corresponding to the high Fe-concentrations found in the EDAX-analysis (Table 2). e highest amount is found in sample IC4, representing the most characteristic and unaltered ore composition. Even the sediment sample still has 36.8 wt% Fe. e Fe-concentration falls within the averages reported for WID by dierent authors (e.g., Saarini 1988). e CaO ranges from 0.06 to 32.84 wt% in the rock samples (with IC4 having the lowest value) and is 29.41 wt% in the sediment. e speleothem sample expectedly has the highest CaO content, i.e., 55.1 wt%. e SiO2 content ranges from < 1 to 12.0 wt% in the rock samples and reaches 6.4 wt% in the sediment. It is interesting to note, that we did not nd any quartz in the XRD samples so that the SiO2 most probably is in the form of amorphous chert. e TiO2 ranges from 0.02 to 0.41 wt% in the rocks sample and reaches an intermediate concentration in the sediment, i.e. 0.07 wt%. MnO, Al2O3 and P2O5 are all low in concentrations ranging from 0.003 to 0.0004, 0.01 to 0.1, and 0.06 to 0.116 wt%, respectively. Similarly Na2O and K2O are very low, i.e., below < 1 wt%.6. ConclusionsMany models can be discussed regarding the potential genesis of the WID, among them: 1) syn-sedimentary deposition, 2) hydrothermal marine, 3) metasomatic, 4) weathering of a basaltic precursor or 5) speleogenetic (phreatic-hypogene). Model 1 can be excluded, because of the texture of the ores and the fact that the goethite is very pure. Model 2 can be excluded because of the low Sr values in our samples (Hendricks et al. 1969). Model 3 most probably can also be excluded since high-temperature metasomatic deposits have a dierent paragenesis, including originally deposited siderite, veins of quartz, barite and co-precipitated heavy metal suldes, all of them missing in the WID (e.g., Kempe 1998). e weathering of basalts (Model 4), a process that can also lead to pure iron oxides (i.e., Al-Malabeh and Kempe 2005), is also not very likely. us the last, the speleogenetic model remains. In speleology it is now an accepted fact (for an overview see Klimchouk, 2007, and review by Kempe, 2007) that many limestone caves, and among them some of the largest

PAGE 291

Speleogenesis 896 2009 ICS Proceedings 15th International Congress of Speleology known, were formed by ascending waters below the water table (i.e., deep-phreatic) by slow gravity-driven convection. In order to generate in-situ dissolutional capacity for these very large cavities, the bacterially mediated reaction, between an anaerobic, ascending phase and an aerobic descending phase at low temperatures is assumed. e reaction between both can then generate protons that serve to dissolve the limestone around the ascending plume of water. Water loaded with H2S or CH4 is not aggressive, but if oxidized, the resulting acids (H2SO3 and H2CO3) would provide dissolutional capacity. In the case of Carlsbad Cavern internal gypsum deposits show that the ascending gas was mostly H2S. CH4, its oxidation and the consecutive dissolution does, however, not leave any traces, possibly with one exception: Goethite. Under anaerobic conditions, when sulde is missing, iron could be mobilized as Fe2+. If mixed with sinking O2-containing water, the CH4 would be oxidized bacterially; CO2 would be produced to dissolve the limestone and at the same time goethite would be precipitated. us, it is conceivable that under such conditions, small-scale goethitic ore bodies could be formed speleogeneticly. Similarly, W4+O2 (solubility 10 mg/l not depending on pH if <10) is more soluble than W6+O3 (0.1-2 mg/l depending on increasing pH) (Dermatas et al. 2004), thus being co-precipitated with Fe upon oxidation and explaining the high tungsten concentration in sample IC4. To our knowledge there are several examples of goethite deposits in hypogene caves. None of them are well published. e most impressive one is the goethite deposit in the gigantic cave found near Wlfrath/ North RhineWestfalia/Germany. A quarry (Rohdenhaus-Sd) in Devonian massive limestone uncovered an enormous, at least 700 m long, 200 m wide and 20 m high cavity lled completely with quartz sand that contained lower Cretaceous plant remains (Drozdzewski et al. 1998). In the centre, a 10 m thick limonitic ore deposit at least 50 m wide was discovered, lling the space above the sedimentary deposits. is ore body could mark the site of the ascending waters, comparable to the deep sea chimneys of black smokers. is cave is one of the largest natural cavities yet discovered and clearly of hypogene origin, created by some sort of in-situ acid generating reaction. e iron ore has not yet been investigated in detail. In a more recent cave, the Bismarck Cavern, Frankonia/Germany, we also observed stalactitic or chimney-like goethite deposits on the oor. is cave is also of hypogene origin and the goethite may be the only clue as to its real formation mechanism. erefore, it is conceivable that the WID formed in a similar way. is speleogenetic lagerstaetten-model, while quite hypothetical, would account for the stratiform character of the deposit, for its situation within a marine limestone sequence, for the missing marine trace metal imprint in the ore and for the missing high temperature signature of the mineral paragenesis. e model would also account for the observed limestone blocks oating in the ore; they could be interpreted as roof collapse of the original cave. When looking at the geological development of the area and assuming the speleogenetic model to be correct, then the deposit could have formed at the end of marine sedimentation in the area, when plate stress began to open up fractures into the basement. From there, methane (and Fe2+) rich solutions could have ascended upward into the limestone series. Where they met sinking, oxygen-rich surface waters, bacterial oxidation and the resulting H2CO3 would have provided for the formation of a large cave. At the same time the ferrous iron would be oxidized and precipitated as waterrich goethitic limonite in the growing cave. As pressure and temperature increased, the bacterial reaction ceased and some of the goethite was dehydrated to hematite. Further increase of the N-S pressure on the plate led to a brecciation of the ore body and to the formation of calcite veins. Once lied to altitudes of several 100 m asl, water, circulating through the breccia, could form small cavities and speleothems could start to grow. For an extended discussion of the WID see AlMalabeh et al (2008). is model, developed according to the present state of knowledge, could serve as a start to investigate similar deposits and the WID could serve as a locus typicus for a new type of iron ore deposit not well studied yet. In a sense, the nding of Al-Daher Cave as described in the introduction and its hypogene nature, illustrates that eective speleogenetic systems involving ascending water plumes have existed in Jordan in the past. ey therefore may have played a much more important role than hitherto acknowledged.ReferencesAl-Malabeh, A., and Kempe, S. (2005) Origin of iron ore nuggets (Bohnerze) through weathering of basalt as documented by pebbles from the Herbstlabyrinth, Breitscheid-Germany. Acta Carsologica, 34(2), 459. Al-Malabeh, A., Kempe, S., Frehat, M., and Henschel, H.V. (2007) Geo-conservation of the newly discovered Kufranja Cave, N-Jordan, a potential natural heritage site. Abstracts Vol. Sixth International Symposium on Eastern Mediterranean Geology, Amman April 2, 2007, 174.

PAGE 292

15th International Congress of Speleology Speleogenesis 897 2009 ICS Proceedings Al-Malabeh, A., Kempe, S., Henschel, H.-V., Hofmann, H., and Tobschall, H.J. (2008) e Warda Iron Ore Deposit and Historic Mine near Ajloun, (Northern Jordan): Mineralogy and Geochemistry of a Potential Speleogene Iron Ore Deposit. Acta Carsologica, 37 (2-3), 241. Barazangi, M. (1983) A summary of the seismotectonics of the Arab region. In AA sse ssment and Mitigation of Earthquake RR isk in the AA rab RRegion, Cidlinsky, K., and Rouhban, B., (eds.), Paris, France, UNESCO, p. 43. Bender, F. (1974) Geology of Jordan. Borntrger, Berlin, Stuttgart, 196 p. Dermatas, D., Braida, W., Christodoulatos, C., Strigul, N., Panikov, N., Los, M., and Larson, S. (2004) Solubility, sorption, and soil respiration eects of tungsten and tungsten alloys. Environmental Forensics, 5(1), 5. Drozdzewski, G., Hartkopf-Frder, C., Lange, F.-G., Oestreicher, B., Ribbert, K.-H., Voigt, S., and Wrede, V. (1998) Vorluge Mitteilung ber unterkretazischen Tiefenkarst im Wlfrather Massenkalk (Rheinisches Schiefergebirge). Mitt. Verb. dt. Hhlenu. Karstforscher, 44(2), 53. Freund, R., Garfunkel, Z., Zak, I., Goldberg, M., Weissbrod, T., and Derin, B. (1970) e shear along the Dead Sea Ri. Philos. Trans. R. Soc. London, A267, 107. Garfunkel, Z., (1981) Internal structure of the Dead Sea leaky transform (ri) in relation to plate kinematics. Tectonophysics, 80, 81. Hendricks, R.L., Reisbick, F.B., Mahay, E.J., Roberts, D.B., and Peterson, M.N.A. (1969) Chemical composition of sediments and interstitial brines from the Atlantis II, Discovery and Chain Deeps. in H ot Brines and RRecent Heavy Metal DDeposits in the R Red Sea, AA Geochemical and Geophysical AA ccount, Degens, E.T., and Ross, D.A. (eds.), Springer Verlag, New York, p. 407. Kempe, S. (1998) Siderite weathering, a rare source of CO2 for cave genesis: e Eisensteinstollen System and adjacent caves in the Iberg, Harz Mountains, Germany. Proc. 1998 Nat. Speleol. Soc. Conv. Sewanee, TN, 3.-7.8.1998, 78. Kempe, S. (2007) Das neue Buch von Alexander Klimchouk zur hypogenen Spelogenese. Mitt. Verb. dt. Hhlenu. Karstforscher, 53(3), 90. Kempe, S., Al-Malabeh, A., Al-Shreideh, A., and Henschel, H.-V. (2006) Al-Daher Cave (Bergish), Jordan, the rst extensive Jordanian limestone cave: A convective Carlsbad-type cave? J. Cave and Karst Studies, 68 (3), 107. Kempe, S., Al-Malabeh, A., Frehat, M., and Henschel, H.-V. (2008) State of lava cave research in Jordan. Proc. 12th Intern. Symp. on Vulcanospeleology, Tepotzln, Mexico, 2-7 July, 2006, Assoc. for Mexican Cave Studies, Bull. 19 and Socieded Mexicana de Exploraciones Subterrneas Bol. 7, 209. Klimchouk, A. (2007) Hypogene Speleogenesis: Hydrogeological and Morphogenetic Perspective. National Cave and Karst Research Institute, Special Paper 1, 106 p. Saarini, G. (1988) Ferride elements abundance in the carbonate-hosted iron occurrences of Jordan. Dirasat (Science), 15(9), 190.

PAGE 293

Speleogenesis 898 2009 ICS Proceedings 15th International Congress of Speleology The HE deepest DEEPEST cave CAVE in IN the THE world WORLD in IN the THE Arabika RABIKA Massif ASSIF Western ESTERN Caucasus AUCASUS and AND its ITS hydrogeological HYDROGEOLOGICAL and AND paleogeographic PALEOGEOGRAPHIC significance SIGNIFICANCEAA .B. Klimch LIMCH Ouk UK G.V. SAm M Okhi KHI N, and Y.M. KAsi SI AN Ukrainian Institute of Speleology and Karstology, 4 Vernadsky Prospect, Simferopol 95007, Ukraine Arabika is an outstanding high-mountain karst massif in the Western Caucasus composed of Upper Jurassic and Lower Cretaceous limestones continuously dipping southwest to the Black Sea and plunging below the sea level. e central sector (elevations 1,900,700 m) is characterized by pronounced glaciokarstic landscape and hosts several deep caves, including the deepest cave in the world, Krubera (Voronja), recently explored to the depth of -2191 m. Dye tracing experiments conducted in 1984-1985 revealed that the Krubera Cave area is hydraulically connected with major springs at the shore of the Black Sea by submarine discharge, and with the ow directed across the strike of major fold structures. Krubera Cave has an extremely steep prole and reveals a huge thickness of the vadose zone. e lower boundary of the vadose zone (the top of the phreatic zone) is at an elevation of about 110 m at low ow, which suggests a low overall hydraulic gradient of 0.007-0.008. Low-TDS groundwater is tapped by boreholes in the shore area at depths of 40, 500, 1,750, and 2,250 m below sea level, which suggests the existence of a deep ow system with vigorous ow. Submarine discharge along the Arabika coast is reported at depths up to ~400 m bsl. A huge closed submarine depression is revealed at the sea-oor next to the Arabika, with its deepest point at ~400 m bsl. ese facts point to the possibility that the main karst system in Arabika could have originated in response to the Messinian salinity crisis (5.96.33 Ma) when the Black Sea (Eastern Paratethys) could have almost dried up, as did the adjacent Mediterranean where a dramatic sea-level drop of ~1,500 m is well established. Further development of the huge vadose zone and the super-deep caves has been caused by subsequent uplis during the Pliocene-Pleistocene, with a great dierence between the shore sector (0.1.2 km of total upli) and the central sector (2.5 km) of Arabika. It is not by chance that the deepest cave in the world, with its exceptionally high vertical range of almost 2,200 m, has been discovered in the Arabika massif. ere were unique geological and paleogeographic pre-conditions for that. e study of the Arabika karst system and exploration of the Krubera Cave constitutes new arguments supporting the response of the Black Sea to the Messinian salinity crisis.1. IntroductionArabika is one of the largest high-mountain limestone karst massifs in the Western Caucasus. Cave exploration in Arabika during the 1980s by speleologists from Ukraine, Russia, Belorussia and Moldova, and dye-tracing experiments performed in 1984 by the Institute of Geological Sciences of the Ukrainian Academy of Sciences, radically changed previous hydrogeological models for the massif and revealed its outstanding potential for deep caves (Klimchouk, 1990). During 1999 this potential was realized in full in Krubera (Voronja) Cave by the expeditions of the Ukrainian Speleological Association. e cave became the deepest in the world in 2001 (-1,710 m). In 2004, for the rst time in the history of speleology, the 2,000 m depth mark was passed in Krubera and the cave was pushed to -2,080 m. In 2006 a depth of -2,158 m was reached, followed by the current record of -2,191 m in 2007. Krubera Cave has an extremely steep prole and reveals a huge thickness of the vadose zone. e top of the phreatic zone is reached at an elevation of about 110 m asl. is, along with evidence for vigorous karst groundwater circulation deep below the present level of the Black Sea, point to a possibility that karst systems in Arabika could have originated in response to a base level at a much lower position in the past than is suggested by the Pleistocene glacioeustatic oscillations. 2. Location and Physiographye Arabika Massif is located in the Western Caucasus (Fig. 1, A-B), in Abkhazia, a republic which ocially belongs to Georgia but holds claim to being an independent state.

PAGE 294

15th International Congress of Speleology Speleogenesis 899 2009 ICS Proceedings To the northwest, north, northeast, and east, Arabika is bordered by the deeply incised canyons of Sandripsh, Kutushara, Gega and Bzyb rivers. e Bzyb River separates Arabika from the adjacent Bzybsky Massif, another outstanding karst area with many deep caves, including the Snezhnaja-Mezhonogo-Iljuzia system (-1,753 m) and Pantjukhina Cave (-1,508 m). To the southwest, Arabika borders the Black Sea, with limestones dipping continuously below the sea level (Fig.1, C). Arabika has a prominent high central sector with elevations above the tree line at ~1,800,900 m. is is an arena Figure 1: Location (A and B) and shaded map of the Arabika and Bzubsky massifs. e sea-oor topography is based on the global digital bathymetric data that combines depth soundings and high-resolution marine gravity data (Smith and Sandwell, 1997). Deep caves are indicated by dots and numbers (explained in the text) and major springs are specied by two-character indexes (explained in the legend). Arrows indicate hydrologic connections proed by dye-tracing experiments.

PAGE 295

Speleogenesis 900 2009 ICS Proceedings 15th International Congress of Speleology of classical glaciokarstic landscape, with numerous glacial trough valleys and cirques, with ridges and peaks between them. e bottoms of the trough valleys and karst elds lie at elevations of 2,000-2,350 m, and ridges and peaks rise to 2,500-2,700m. e highest peak is the Peak of Speleologists (2,705 m) but the dominant summit is a typical pyramidal horn of the AA r abika Mount (2,695m). Some middleto lowaltitude ridges covered with forest lie between the central sector and the Black Sea. A plateau-like middle-altitude outlier of the massif in its south sector is Mamzdyshkha, with part of the plateau slightly emerging above the tree line. 3. Deep CavesAmong several hundred caves known in the Arabika massif, een have been explored below 400 m and ve below 1,000 m (shown in Fig. 1, C). Several are located within the Ortobalagan valley, a perfectly shaped, relatively shallow, glacial trough of the sub-Caucasian stretch, which holds the advanced position in the central sector toward the seashore. Since 1980, Ukrainian cavers have been undertaking systematic eorts in exploring deep caves in the Ortobalagan Valley resulted in exploration of the Krubera (Voronja) Cave (number 1 on Fig. 1, C), the deepest cave in the world with its current depth of -2,191 m, and the Arabikskaja System (number 4), which consists of Kujbyshevskaja Cave (-1,110 m) and Genrikhoa Bezdna Cave (-965 m to the junction with Kujbyshevskaja). Another deep cave in the valley, located in its very upper part, is Berchilskaja Cave (-500 m; number 11) explored by Moldovian and Ukrainian cavers. e Ortobalagan valley extends along the crest of the Berchilsky anticline, which gently dips northwest. e cave entrances are aligned along the anticlinal crest (Figs. 2 and 3) but the caves are controlled by longitudinal, transverse, and oblique fractures and faults and comprise complex winding patterns in the plan view, remaining largely within and near the anticlinal crest zone. All large caves of the Ortobalagan Valley likely belong to a single hydrologic system. e direct connection of Krubera Cave with the Arabikaskaja System is a sound speleological possibility. e caves are predominantly combinations of vadose shas and steep meandering passages, although in places they cut apparently old fossil passages at dierent levels (e.g., at 2,100-2,040 m in Kujbyshevskaja and Krubera caves, 1,200-1,240 m and 980-1,150 m in the Non-Kujbyshevskaja branch of Krubera Cave, etc.). e antiquity of these passages is supported by the ages of speleothems falling beyond the 230 dating limit (>500 ka; Klimchouk et al. 2009). e deep parts of Krubera display a more pervasive conduit pattern with a mixture of phreatic morphology, characteristic of the zone of high-gradient oods, which can be up to 400 m above the low-ow water table, and vadose Figure 2: Major caves in the Ortobalagan valley. Dots indicate dolines.

PAGE 296

15th International Congress of Speleology Speleogenesis 901 2009 ICS Proceedings downcutting elements that are observed even below the water table. Other deep caves in Arabika include the Iljukhina System (-1,273 m; number 3) located in the center of the massif, Dzou Cave (-1,090 m; number 5) and Moskovskaja Cave (-940 m; number 6) in the north-eastern part, and Sarma Cave (-1,540 m; number 13). e latter, the second deepest Figure 3: Combined cave prole along the axis of the Ortobalagan valley. Krubera Cave is shown in white, other caves in light grey.

PAGE 297

Speleogenesis 902 2009 ICS Proceedings 15th International Congress of Speleology cave in Arabika, is located along the same anticline as Krubera, at the similar advanced position relative to the seashore. 4. Geology e Arabika Massif is composed of Lower Cretaceous and Upper Jurassic limestones that dip continuously southwest to the Black Sea and extend below the modern sea level (Fig. 4). In the central part of Arabika the Cretaceous cover is retained only in a few ridges and peaks, as well as in small patches within the trough valleys (Valanginian and Hauterivan limestones, marls and sandstones). In the south and southeast, the Cretaceous succession includes Barremian and Aptian-Senomanian limestones and marly limestones with abundant concretions of black chert. e core part of the massif is composed of the Upper Jurassic succession resting on the Bajocian Porphyritic Series, which includes sandstones, clays and conglomerates at the top, and tu, tu sandstones, conglomerates and breccia, porphyry and lava. e Porphyritic series forms the nonkarstic basement of Arabika, which is exposed only on the northern and eastern outskirts, locally in the bottoms of the Kutushara and Gega River valleys. e Upper Jurassic succession begins with thin-bedded Kimmeridgian-Oxfordian cherty limestones, marls, sandstones and clays, which are identied in the lower part of Krubera Cave. Above lies the thick Titonian succession of thick-bedded limestones with marly and sandy varieties. Sandy limestones are particularly abundant through the upper 1000 m sections of deep caves of the Ortobalagan Valley. e tectonic structure of Arabika is dominated by the axis of the large sub-Caucasian anticline (oriented NW-SE), with the gently dipping southwestern mega-ank, complicated by several low-order folds, and steeply dipping northeastern ank. e axis of the anticline roughly coincides with the ridge bordering the Gelgeluk Valley to the north. Located on the southwestern ank of the major anticline is another large one (Berchil), in which the crest is breached by the Ortobalagan Valley. ere are several smaller sub-parallel anticlines and synclines farther southwest, between Berchil and the coast. e plicative dislocation structure of the massif is severely complicated by faults, with the fault-block structure strongly controlling both cave development and groundwater ow (Klimchouk 1990). Major faults of the sub-Caucasian orientation delineate several large elongated blocks that experienced upli with dierent rates during Pliocene and Pleistocene. is had a pronounced eect on the development of deep groundwater circulation and of Krubera Cave in particular. Both longitudinal and transverse faults and related fracture zones play a role in guiding groundwater ow; the latter guide ow across the strike of major plicative dislocations, from the central sector toward the Black Sea. Figure 4: Schematic geological and speleo-hydrological section of the Arabika massif. See text for explanations.

PAGE 298

15th International Congress of Speleology Speleogenesis 903 2009 ICS Proceedings 5. HydrogeologyMajor on-shore karst springs with individual average discharges of 1 to 2.5 m3/sec are located at altitudes ranging from 1 m (Reproa Spring) to 540 m (Gegsky Vodopad). Two of them are located in the shore area; these are Reproa (average discharge 2.5 m3/sec; alt. 1 m asl) and Kholodnaja Rechka (1.2 m3/sec; 50 m asl). Other two major springs are located in the river canyons bordering Arabika to the east: Goluboe Ozero in the Bzyb canyon (2.5 m3/sec; 90 m asl) and Gegsky Vodopad in the Gega canyon (1 m3/sec; 540 m asl). ere are also several smaller springs in the Gagra town. Some boreholes located along the shore of the Black Sea yield karstic groundwater from depths of 40 m below sea level. Other much deeper boreholes tapped lowsalinity karstic waters at depths of 500 and 1,750 m in the Khashupse Valley near Tsandripsh and 2,250 m near Gagra (Buachidze and Meliva 1967; Meliva et al. 1969). is suggests the existence of deep karst system and vigorous karst groundwater circulation at depth. Submarine springs are known in the Arabika area, emerging from the oor of the Black Sea in front of the massif. Shallow springs at depths of 5 m can be reached by free dive near Tsandripsh. Kiknadze (1979) reported submarine springs near the eastern part of Gagra at depth of 25-30 m and Buachidze and Meliva (1967) revealed submarine discharge at depths up to -400 m by hydrochemical proling. Recently (Klimchouk 2006), an outstanding feature of the seaoor topography near Arabika has been revealed from a digital bathymetric map that combines depth soundings and high-resolution marine gravity data (Smith and Sandwell 1997). is is a huge submarine depression in front of the Zhovekvara River mouth, which has dimensions of about 5 x 9 km and a maximum depth of about 380 m. e Arabika Submarine Depression (ASD; shown on Fig. 1, C) is a closed feature with internal vertical relief of about 120 m (measured from its lowest rim) separated from the abyssal slope by the bar at a depth of about -260 m. It has steep northern and northeastern slopes (on the side of the massif) and gentle south and southwestern slopes. Its formation is apparently karstic. Presently ASD seems to be a focus of submarine discharge of the karst systems of Arabika. e existence of ASD, along with other lines of evidences, points to the possibility of much lower sea-level positions in the past than is suggested by Pleistocene glacioeustatic oscillations (Klimchouk 2006). e hydrogeologic model for the Arabika massif that dominated before the 1980s did not allow the possibility of a hydrologic connection between the central high sector and the coastal springs. e model implied that the aquifer structure in Arabika is subordinate to the sub-Caucasian synclines and anticlines, and that minor non-karstiable beds within the carbonate succession separate the system into several superimposed aquifers (Kiknadze 1972; 1979). According to this model, groundwater recharged within the central part of the massif ows to northeast, beneath the non-karstic cover, and southwest to the Goluboje Ozero spring. e recharge areas for the coastal and submarine springs were assumed to be the proximal low-altitude ridges. is model was disproved by speleological explorations and dye tracing studies during the 1980s under the coordination of the Institute of Geological Sciences of the of the Ukrainian Academy of Science (Klimchouk 1990). A series of large-scale dye-tracing experiments was conducted in Arabika in 1984 and 1985. Tracers injected in the Kujbyshevskaja and Iljukhina caves were detected in the Kholodnaja Rechka and Reproa springs, proving groundwater ow to the south-southwest across major tectonic structures over a distance of 13-16 km as the crow ies (Fig. 1, C). e tracer from Kujbyshevskaja was also detected in a borehole located between these two springs, which yields groundwater from a depth of 200 m below sea level. is has been interpreted as an indication of the connection of the cave with the submarine discharge. e large Central Karst Hydrologic System, which encompasses most of the southeastern ank of the Arabika anticline, had been identied in this way. e system became the deepest in the world with its overall vertical range of about 2,500 m (measuring to the borehole water-bearing horizon) or even 2,700 m (measuring to the deepest reported submarine discharge points). Another tracer was injected in the Moskovskaja Cave (-970 m) and detected at the Gegsky Vodopad spring, indicating the presence of a karst hydrologic system comprising the northeastern ank of the Arabika anticline (the Northern System). No connections have been revealed with yet another major spring, Goluboje Ozero in the Bzyb River canyon, although it apparently drains a large area of the eastern sector of the massif (the hypothetical Eastern Karst Hydrological System). It is not clear where Sarma Cave (-1,550 m) drains to, Goluboje Ozero to the southeast or Reproa to the southwest, at the shore. e results of the dye-tracing tests have radically changed notions of the hydrogeology of Arabika and revealed its outstanding speleological perspectives and strongly stimulated further eorts for exploration of deep caves. ey demonstrated that groundwater ow is not subordinate to the fold structure but is largely controlled by faults that cut

PAGE 299

Speleogenesis 904 2009 ICS Proceedings 15th International Congress of Speleology across the strike of major folds, and that the large part of the central sector of Arabika is hydraulically connected to the springs along the seashore and with submarine discharge points. 6. DiscussionKrubera Cave has an extremely steep prole and reveals a huge thickness of the vadose zone. e lower boundary of the vadose zone (the top of the phreatic zone) is at an elevation of about 110 m at low ow, which suggests a low overall hydraulic gradient of 0.0070.008. LowTDS groundwater is tapped by boreholes in the shore area at depths of 40-280, 500, 1,750, and 2,250 m below sea level, which suggests the existence of a deep ow system with vigorous ow. Submarine discharge along the Arabika coast is reported at depths up to ~400 m bsl. A huge closed submarine depression is revealed at the sea oor next to Arabika, with the deepest point of ~400 m bsl. It is dicult to interpret these facts in terms of the development of karst systems controlled by contemporary sea level, or within the range of its Pleistocene uctuations (up to -150 m). Klimchouk (2006) suggested that karst systems in Arabika could have originated in response to the Messinian salinity crisis (5.96.33 Ma) when the Black Sea (Eastern Paratethys) could have almost dried up, as did the adjacent Mediterranean, where the dramatic sea level drop of ~1500 m is well established. It has been demonstrated by the recent studies of French karstologists (e.g., Mocochain et al. 2006) that the Messinian crisis played a great role in karst development in the Mediterranean region, where deep conduits formed in response to lowering of the base level and imposed a strong inuence on subsequent karst evolution. e hypothesis that the dramatic sea-level drop could have take place in the Black Sea basin during the Messinian time had been put forward 30 years ago (Hs and Giovanoli 1979) mainly on the basis of deep-sea drilling data (DSDP Sites 379, 380, 381). It has been strongly corroborated by recent studies of regional geology, including data from bioand magnetostratigraphy of the key sedimentary sequences (Semenenko and Olejnik 1995; Clauzon et al. 2005; Popesku 2006; Snel et al. 2006), seismic proling (establishing the Messinian erosional surface in the Eastern Paratethys; Gillet 2003), studies of deep-water delta complexes, etc. Before the late Miocene, the present coastal Western Caucasus was a lowto middle-altitude mountain terrain. Temporary desiccation of the Black Sea in response to the Messinian salinity crisis in the Mediterranean established the base level at many hundreds meters below the present level and caused conduit initiation and development within deep sections of the Arabika massif. ese early systems were ooded aer the Pliocene transgression. Uplis of the Arabika area during Pliocene and especially Pleistocene were highly dierentiated by elongate zones (blocks) of the sub-Caucasian stretch (parallel to the coast). e total upli amounted to 2-2.5 km in the central sector of Arabika, whereas it was minimal (0.10.2 km) in the zone proximal to the coast. Hydraulic continuity was always maintained across the zones, between the main recharge area in the central sector of Arabika and the coastal zone and submarine springs. e presence of high conduit porosity of Messinian origin in the coastal/ submarine sector allowed the zone of high hydraulic gradient during upli to be pushed far inland, beneath the rising central sector (Fig. 4). is created favorable conditions for the enhanced conduit development at depth in the central sector, quick adjustment of the water table to new upli pulses, and eventual development of a huge vadose zone and extremely steep and deep cave systems such as Krubera Cave. is was further favored by recurring sea-level drops up to -150 m during the Pleistocene, which caused the steepening of hydraulic gradients beneath the central sector of Arabika and enhancement of conduit porosity in the upper part of the present phreatic zone. e evolution scenario outlined above is indirectly supported by 230 dating of speleothems from the deep parts of Krubera Cave (Klimchouk et al. 2009). Stalagmites from depths of -1640 m and -1820 m (elevations 640 m and 436 m asl) have yielded ages older than 200 ka (max. 276 ka), which suggests that the deep parts of the cave were already in the vadose zone before the Middle Pleistocene.6. ConclusionIt is not by chance that the deepest cave in the world with the exceptionally high vertical range of almost 2200 m has been discovered in the Arabika massif. ere were unique geological and paleogeographic pre-conditions for that. e study of the Arabika karst system and exploration of the Krubera Cave constitutes new arguments supporting the response of the Black Sea to the Messinian salinity crisis.Acknowledgementis study is an outcome of enormous eorts of several generations of cavers who have explored the deep caves in Arabika during past 30 years.

PAGE 300

15th International Congress of Speleology Speleogenesis 905 2009 ICS Proceedings ReferencesBuachidze, I.M., and Meliva, A.M. (1967) To the question of groundwater discharge into the Black Sea in the Gagra area. Trudy laboratorii gidgogeologii I inzhenernoy geologii Gruzinskogo politechnicheskogo institute, 3, 33-39. Clauzon, G., Suc, J.P., Popescu, S.-M., Mruneanu, M., Rubino, J.-L., Marinescu, F., and Melinte, M.C. (2005) Inuence of the Mediterranean sea-level changes over the Dacic Basin (Eastern Paratethys) in the Late Neogene. e Mediterranean Lago Mare facies deciphered. Basin Research, 17, 437. Gillet, H., Gilles, L., Renault, J-P., and Dinu, C. (2003) La stratigraphie oligo-miocene et la surface derosion messinienne en mer Noire, stratigraphie sismique haute resolution. Geoscience, 335, 907-916. Hs, K.J., and Giovanoli, F. (1979) Messinian event in the Black Sea. Palaeogeogr. Palaeoclimatol. Palaeoecol., 29:1-2, 75. Kiknadze, T. Z. (1972) Karst of the Arabika massif. Metzniereba, Tbilisi, 245 p. (in Russian). Kiknadze, T.Z. (1979) Geology, Hydrogeology and activity of limestone karst. Metzniereba, Tbilisi, 232 p. (in Russian). Klimchouk, A. B. (1990) Karst circulation systems of the Arabika massif. Peschery (Caves), Perm University, Perm, 6 (in Russian). Klimchouk, A.B. (2006) e deepest cave in the world in the Arabika Massif and the evolution of the Black Sea. Svet (Light), 2:31, 33-36 (in Rissian). Klimchouk, A.B., Samokhin, G.V., Cheng H., and Edwards R.L. (2009) Dating of speleothems from deep parts of the worlds deepest cave Krubera (Arabika Massif, Western Caucasus), this volume. Mocochain, L., Clauzon, G., Bigot, J.-Y., and Brunet P. (2006) Geodynamic evolution of the perimediterranean karst during the Messinian and the Pliocene: evidence from the Ardche and the Rhne Valley systems canyons, Southern France. Sedimentary Geology, 188-189, 219-233. Popescu, S.-M. (2006), Late Miocene and early Pliocene environments in the southwestern Black Sea region from high-resolution palynology of DSDP Site 380A (Leg 42B). Palaeogeography, Palaeoclimatology, Palaeoecology, 238, 64. Semenenko, V.N., and Olejnik, E.S. (1995) Stratigraphic correlation of the Eastern Paratethys Kimmerian and Dacian stages by molluscs, dinocyst and nannoplankton data. Rom. J. Stratigr., 76:7, 113 114. Smith, W.H.F., and Sandwell, D.T. (1997) Global Sea Floor Topography from Satellite Altimetry and Ship Depth Soundings. Science, 277, 1956-1962. Snel, E., Mruneanu, M., Macale, R., Meulenkamp, J.E., and Van Vugt, N. (2006) Late Miocene to early Pliocene chronostratigraphic framework for the Dacic Basin, Romania, in Agusti, J., Oms, O., Meulenkamp, J.E., Eds., Late Miocene to Early Pliocene Enironment and Climate Change in the M editerranean AA rea. Palaeogeogr. Palaeoclimatol. Palaeoecol, 238, 107. doi:10.1016/ j.palaeo.2006.03.021.

PAGE 301

Speleogenesis 906 2009 ICS Proceedings 15th International Congress of Speleology Hypogenic YPOGENIC speleogenesis SPELEOGENESIS in IN the THE Piedmont IEDMONT Crimea RIMEA range RANGE AA .B. Klimch LIMCH Ouk UK E.I. T T YMO O KHINA NA and G.N N AA MELICHEV Ukrainian Institute of Speleology and Karstology, 4 Vernadsky Prospect, Simferopol 95007, UkraineAbstracte intense development of the theory and criteria of identication of hypogenic speleogenesis during last few years has stimulated re-interpretation of karst phenomena in many regions of the world. Recent research strongly suggests that solution features in the Piedmont Range of the Crimean Mountains, previously believed as being the result of epigenic karstication, were in fact formed in hypogenic environment due to ascending transverse ow in a stratied artesian system. Tectonically, the Piedmont Range of Crimea is an edge of the Scythian Plate, uplied and partially eroded along the regional fault separating the plate from the folded region of the Mountain Crimea. e Cretaceous and Paleogene sequence dips 5 to 20o to the north and north-west, where it plunges beneath the Neogene cover. It is exposed within the Piedmont Range as a series of distinct cuestas generally facing southeast. Karst features are represented by 26 caves and abundant and diverse solution forms at the cuesta scarps. Most of karst features have developed in two distinct limestone units of Paleocene (Danian) and Eocene age, but some are present in the underlying Maastrichtian unit of the Cretaceous. ere is strong and systematic evidence that the caves have hypogenic origins and that most of the solution features at the scarps are remnants of morphologies of hypogenically karstied fractures, whose walls are now exposed by block-fall scarp retreat. e features in various beds demonstrate strong lithostratigraphic control in their distribution and are vertically stacked into transverse complexes. Caves are fracturecontrolled, linear, or crude maze clusters, demonstrating the complete suit of morphologies indicative of hypogenic origin. Isolated cavities, expressed in the contemporary scarps as grottos and niches, as well as zones of spongework porosity, developed where laterally conductive beds of higher initial porosity were crossed by vertical fractures that once conducted rising uids from a regional ow system. e Piedmont Range of Crimea was a part of the Plain Crimea artesian basin before the Middle Pliocene. Subsequent upli and initial erosional entrenchment through the Early and Late Pliocene caused the pattern of tectonically and geomorphically guided zones of upward cross-formational discharge and hypogenic speleogenesis to establish. Further valley entrenchment in the region during the Pleistocene shaped the modern cuesta-like relief and drained the Cretaceous-Paleogene sequence. Hypogenically karstied fractures and caves, sub-parallel to valleys, provide zones of structural weakness along which blocks fall at the cuesta scarps, exposing relict hypogenic morphologies. e Piedmont Crimea Range, with its perfect and extensive exposures of hypogenically karstied sequences, provides outstanding possibilities for studying patterns and morphologies of hypogenic speleogenesis, which is important for understanding its hydrogeological functioning and roles in reservoir formation, especially with implications about the adjacent Plain Crimea artesian basin.

PAGE 302

15th International Congress of Speleology Speleogenesis 907 2009 ICS Proceedings BATHYPHREATIC SPELEOGENESIS OF SOME LARGE KARST SPRINGS IN CROATIAMLAD AD EN N KUHTA TA1, ANDRANDR EJ STRO TROJ1, ELJKA A BRR KI1, and AA LAN AN KO O VA A EVI2 1 Croatian Geological Survey, Sachsoa 2, 10000 Zagreb, Croatia2 Cving Club DDDD ISKF, Frane AA lroia 13, 10000 Zagreb, CroatiaAbstracte Karst Dinarides cover about 26,000 km2 in Croatia, which is 46% of the national territory. ey are composed of carbonate deposits that represent relics of several vertically stacked carbonate platforms, in places more than 8000 m thick, with a stratigraphic range from Middle Permian to Eocene. e largest part of the carbonate succession was deposited on the Adriatic Carbonate Platform. e thickness of these deposits, formed during the 125 million years of the platform history (lower Jurassic to uppermost of Cretaceous) ranges from 3500 to 5000 m. e carbonates are interbedded with Paleozoic, Triassic and Paleogene clastic deposits. e entire area is intensely tectonized. e nal upli of the Dinarides reached its maximum in the Oligocene-Miocene. e recent geomorphology, as well as the speleogenesis and evolution of karst aquifers, have been strongly aected by Neotectonic deformation. e great thickness of carbonates, intense tectonic disturbance, and endogenetic processes result in very deep and irregular karstication and complex hydrogeology in the area. Besides many surface landforms, and more than 10,000 explored caves in the underground, the strong karstication of the carbonate rocks is conrmed by large karst springs, some which discharge more than 100 m3/sec. During the past decade, deep cave-diving explorations have been conducted in some of the most signicant springs. e results are impressive. e Una River spring was explored to the depth of -205 m, the Kupa River spring to 154 m, and the Sinjac spring to -155 m. In addition, signicant depths have been reached in the springs Glava (115 m, Cetina River), and Kamanik (95 m). All of these springs are of the ascending type, with simple sha morphologies that continue beyond the depth of exploration. Dierent morphological characteristics were determined in the Majerovo vrilo spring (Gacka River), where divers not only reached the considerable depth of 104 m, but also discovered a 942-meter-long phreatic cave system. A typical example of shallow-phreatic zone development is 22-meter-deep Zagorska Mrenica River spring, in which the total length of discovered horizontal channels exceeds 1,100 m. In this paper, the geomorphology of some the most signicant active springs of Croatian Karst Dinarides and discussion on their speleogenesis is discussed within the framework of regional geologic and hydrogeologic conditions. e most important factors that aected the internal dynamics of the karst aquifers and genesis of the bathyphreatic conduits are evaluated. Although information about the system is incomplete, it is suggested that the development of the bathyphreatic caves is strongly conditioned by regional compressional tectonics, i.e. the structural position of the less permeable or impervious deposits caused by overthrusting and reverse faulting. Moreover, there has been signicant impact caused by the dierence in elevation between the recharge and discharge areas, which produces a large hydraulic potential and consequently deeper water circulation.

PAGE 303

Speleogenesis 908 2009 ICS Proceedings 15th International Congress of Speleology First IRST elements ELEMENTS of OF the THE karstic KARSTIC evolution EVOLUTION of OF a A continental CONTINENTAL chalk CHALK catchment CATCHMENT area AREA of OF the THE Paris ARIS Basin ASIN Upper PPER Avre VRE River IVER Normandy ORMANDY France RANCE LAu U Re E NT MAg G Ne E1 and JOel EL R R ODe E T2 1 CN N EK (NN orman Center for Karst Studies)2 UMR R 6143 CNR NR S-Continental and Coastal Morphodynamics, Laboratory of Geology, University of RRouen, 76821 Mont Saint AA ignan, FranceAbstractIn the Western Paris Basin, the Avre River feeds the Eure River, a major tributary on the le bank of the Norman Seine River. Located 100 km from Paris and Rouen and 40 km from Chartres, its catchment covers approximately 400 km. It is developed over a long back slope of a cuesta in the Turonian chalk (Upper Cretaceous) at the southern limit of the Seine basin. e Upper Avre River joins the Vigne River approximately 40 km downstream from its sources in the Perche country. e short upstream Vigne River originates from karstic springs that provide approximately 15% of the total water supply of the City of Paris. At the catchment scale, supercial perennial and temporary ows represent a 157-km linear course and are deeply impacted by numerous sink points. Many tracer tests conducted over the last century demonstrate that these sink points supply both the Vigne resurgences and also several springs along the Upper Avre that function as estavelles, notably those near the town of Verneuil-sur-Avre. Geomorphic and hydrologic analysis reveals ow anomalies (disjunction between underground and surface circulation, specic ows that do not correlate with the sizes of catchment areas, heterogeneity in piezometric levels in wells, changing functions of karstic phenomena under various hydrological conditions). It also permits discrimination of elementary units within the hydrographic basin: a recharge zone to the Upper Avre functioning as a main surface collector, a recharge zone to the Vigne springs including the Butenay and Lamblore Creeks, and nally the conuence of these two previous elements resulting in the Lower Avre River, along which no karstic phenomena interfere with its course. At the current level of investigation,, a geomorphic evolution model in three main stages can be proposed: (1) e Avre Basin rests on two individual systems: in the North, the Upper Avre is elongated in the WSW-ENE direction toward the center of the Paris Basin. In the southeast is a small independent river, the Lower Avre. (2) Tectonic reactivation of a fault downstream from Verneuil dammed the ow of the Upper Avre, generating the development of a polje. Leakage by karst piracy to the Vigne springs now augments the Lower Avre. us, the Buternay and the Lamblore increasingly feed the Vigne at the expense of the Verneuil polje. (3) A very narrow valley located along N-S fold axes captures the course of the Upper Avre, reducing the activity of the Verneuil polje. e Lower Avre is directly sustained by the Upper Avre and by the springs of the Vigne. e karstic connections between the polje and the springs are only temporary, as demonstrated by hydrochemistry and tracer tests.

PAGE 304

First elements of the karstic evolution of a continental basin in chalk limestones of the Paris Basin: Upper Avre River (Normandy, France). Jol Rodet1*, Laurent Magne2 1UMR 6143 CNRS Continental and Coastal Morphodynamics, Laboratory of Geology, Unive rsity of Rouen, 76821 Mont Saint Aignan, France 2CNEK (Norman Center for Karst Studies) corresponding author: Abstract In the Western Paris Basin, the Avre River Basin develops i n the Turonian chalk limestones (Upper Cretaceous) and feeds the Eure River, tributary of the Seine River, 100 km far from Paris and Rouen, and 40 km from Chartres, at the southern limits of the Seine Basin. The catchment covering approximately 400 km is developed over a long reverse slope of cuesta. From the sources in the Perche country to the junction with the Vigne River, the course of the Upper Avre River is approximately of 40 km. Superficial flows both perennial and temporary represent all together 190 km of linear course. The Avre River develops eastwards and receives on its right bank all the tributaries running from the forest of Perche, a regional water reservoir. Surrounding the Vigne Avre junction, a very great number of springs are identified. The most significant of them are collected in order to feed up to approximately 15 % of the total water supply of the City of Paris. The management developments of these springs emphases the karstic mode of their drainage, largely confirmed by several tracer tests, those revealing flow anomalies (disjunction between underground/surface circulations, specific flows in dissension with size of elementary basins, heterogeneity in the piezometric variation in wells, change of function of karstic phenomena accor ding to hydrological modes estavelle). The geomorphological analysis made it possible to identify a storage area functioning as a polje, exhibiting a complex evolution of the basin and allowing the proposal of an evolution diagram in three principal pha ses. 1) The Avre Basin develops two individualized systems: at the North, the Upper Avre which runs out of the Perche country from the WSW towards the ENE while passing by Verneuil and is prolonged on the same axis towards the center of the Paris Basin. Upstream of Verneuil, one notes the junction of the basins of the Buternay and the Lamblore. In the SE, a small river develops, the Lower Avre. 2) The tectonic activation of a fault located downstream from Verneuil creates a dam along the flow of the Upper Avre, which generates the development of a polje which karstic leakage is carried out to the springs of the Vigne which benefits from now on to the Lower Avre. Thus the Buternay and the Lamblore feed more and more the sources of the Vigne at the expense o f the polje of Verneuil. 3) A very narrow valley located along the S N tectonic undulation captures the course of the Upper Avre, reducing the activity of the polje of Verneuil. The Lower Avre is then directly sustained by the Upper Avre and the springs o f the Vigne. The karstic connections between the polje and the sources are only temporary, which is shown by the hydrochemistry of tracer tests. Karstologic and hydrological studies are in progress to confirm this conceptual approach. 1. Introduction Lo cated in the Southeastern Normandy, the Avre River is one of the most significant affluent of the Eure River, main tributary on left bank of the Lower Seine River. The upstream catchment develops on a large karstic basin of approximately 400 km. The Upper Avre Basin has been for a long time the object of hydrological studies because it contributes since the beginning of the 20th century to the water supply of the City of Paris (BRARD, 1899). It thus offers a rare and valuable record over more than one cent ury of varied hydrological measurements. Though curiously, this basin has been the object of very rare karstologic studies despite that many phenomena were known since a very long time (DESCHESNES, 1675). 2Description of the catchment basin The Upper Av re Basin develops on two geological sets with an opposite behaviour (Fig. 1). In the Southwestern area, there is the formation of the "Sands of Perche" dating from the Upper Cenomanian, in fact a coastal accumulation of sands and clays, over all impermeable, which changes, in the Northeastern area, by side variation of facies, into sandy and porous chalk limestones. This opposition of facies is asserted in the hydrological behaviour on the one hand, by an upstream zone with dense surface drainage (Forest of Perche) and on the other hand, by a downstream zone with numerous karstic phenomena which develop along the principal rivers flowing from the Perche region. The overburden is mainly constituted of clay with flints (CWF). It rests in a bevel configuration over the chalk substratum and exhibits

PAGE 305

an increasing thickness as it follows the general dip of the layers towards the centre of the Paris Basin (LAIGNEL, 1997). In fact the chalk does not appear in any place in the basin. The tectonic framework primarily consists in faults intersecting the long reverse slope of cuesta which extends towards the plateaus of the Southern Eure County. Joints and lineaments show a correlation with many sink points located in the bed of the rivers (SYKIOTI et al., 1996). The U pper Avre Basin is under oceanic temperate climate, with an annual average temperature of about 11 C. The major rain period extends from the end of September to the end of April but can last until July according to the quantity of rain and the daily tempe rature. The pluviometry is about 645 mm in Rueil la Gadelire and 750 mm in the town of La Fert Vidame. The effective rainfall is of 254 mm, a third of total precipitations. One can note that for the same monthly volume of rainwater, the rivers flow is mo re closely related to the intensity and the frequency of rainy episodes than to a daily mean distributed over the month (3 x 15 mm are more effective than 30 x 1,5 mm). The basin does not react uniformly to precipitations. When it is raining over the entir e basin, every spring reacts (DIENERT, 1901). When it is raining on the Upper Avre Forest of Perche, only the springs of Erigny, Foisy, Gravires and Rivire react, and when it is raining on La Fert Vidame, the springs of Nouvet (Chesne, Gonord, Blaou) re act first then followed by all the other sources (Fig. 2). The extent and complexity of the basin required an analysis of the topographic space which we divided into four geomorphological areas. aT he Upper Avre subbasin (SW of the basin) with a dense surface drainage developed on the formation of the Sands of Perche, between the ponds of the Forest of Perche and Saint Christophe sur Avre, centred on several surface rivers (Avre). One can notice the presence of the higher points of the basin with altit udes ranging between 280 and more than 300 m ASL (Bubertr). bThe linear axis of main drainage of the Upper Avre (oriented SW NE) from Saint Christophe sur Avre to a vast flooding area near Verneuil sur Avre. There is no notable affluent along this course. In Verneuil, the Upper Avre River flows at 160 m ASL. cT he Vigne River Basin with a dense surface drainage developed on the formation of the Sands of Perche between the ponds of La Fert Vidame and Rueil la Gadelire, centred on several temporary su rface creeks (Buternay, Lamblore, etc.) and subterranean perennial rivers which feed the sources of the Vigne. After a few kilometres of surface circulation, the drainage is lost through many sink points into the chalky substrate and supplies the springs o f the Vigne more than 10 km downstream. These sources contribute to 15 % of the total water supply of the City of Paris. dThe meandered and incised course of the Upper Avre River between Verneuil sur Avre and the junction of the Vigne River involves a s outhwards postponement of the river in a counter dip configuration (piracy of the Upper Avre River). At the junction, at the downstream limit of the basin, the altitude of the Upper Avre River is 145 m ASL. Downstream from the junction, the river flows in a relatively incised but very meandering valley (Lower Avre River). The bottom of the valley is filled in up till upstream of the sources of the Vigne. 3A complex karstic behaviour Most of the rivers, as soon as they leave the formation of the Sands of Perche, present a porous bed bored with sinkpoints. The surface flow of the creeks is thus maintained until the passage over the CWF, except in the sectors where the sink points were clogged. If the sink points are very numerous, one can underline that not one meter of karstic conduit is accessible today. However marlpits and wells have trepanned several times significant functional drains (Fig. 3). Reports and field surveys show that karst with metric dimensions exists in marly chalk at approximately ten meters under the surface (DIENERT, 1901). Galleries of 3 4 m wide and 1 m height were recognized as well as drains on diaclase which present the same orientation than the tectonic anomalies visible on surface. However, the wells allowing underground acces s are nowadays filled in (Beauche, Bois Spert, Morvilliers, etc.). The comprehension of karstic behaviour thus depends on the geomorphological (water losses, springs, etc.) and hydrogeological data (chemistry, tracings, etc). The slope of the roof of the aquifer does not follow the surface topography. The levels vary in depth between 0 and 15 m in the formation of the Sands of Perche. This water table is not very productive (DESPREZ & MARTINS, 1972) and the surface flows remain lower than 5 l/s. The depth increases suddenly when arriving on the chalk basement and the aquifer descends from 15 m to 40 m under the surface in the centre of the basin. Upstream from the springs of the Vigne, it then rises up around 15 m. The sink points are localised in the z ones where the underground aquifer falls down quickly and close to the tectonic features. In the valleys, one can encounter a subsurface water table sometimes localised in the CWF as visible near Rueil la Gadelire. This water table also outcrops close to Verneuil, area subjected

PAGE 306

to frequent floods. In this sector, the Upper Avre River exhibits estavelles which are working either as sink point or as spring according to the hydrological cycle. It is notable that the water table follows a threshold functionin g. As soon as it reaches the CWF, peaks of turbidity affect the output of all the springs. Drainage water from the infiltration into ditches also contributes to the turbidity. Th e sudden rises of the water table involve a significant degradation of the bio chemical quality of underground waters (MAGNE & WELTE, 2007). The average content in chlorides is 7 mg/l for the rain and 17 mg/l in the chalk aquifer. Curiously, it goes up to 30 mg/l in the wells. Can these facts be interpreted as the incidence of a surficial aquifer in the covering layers as illustrated in the area of Rouen (cf infra Rodet et al., Proceedings of the 15th ICS)? The Upper Avre Basin is the site of very many dye tracings since more than one century. The tracing operation of September 8th, 1887 into the Lambergerie sinkpoints (downstream of St Christophe) has affected the Verneuil springs one day before the Vigne springs (FERRAY, 1896). The injections into the Lamblore sink points has initially affected the Nouvet sources, but never reache d the Verneuil emergences (Polay, Gonord). Other tests on the Buternay sink points, upstream of Rohaire hamlet and during high water periods, have attained the Gonord and Polay springs and also reached some sources of the Vigne. These facts are confirmed by piezometric measurements which give evidence for a diffluence of the underground flow. These operations give apparently contradictory results. On the one hand, the tributaries of the Vigne River feed in some part the sources of the Upper Avre River, ar ound Verneuil. On the other hand water losses of the Upper Avre River, upstream of Verneuil, resurge at the Vigne springs. These connections do not seem functional during all the hydrological cycle but depend on the rainy episodes. These elements demonstrate the complex exchanges of karstic type between the sources of the Vigne and the overflow plain of Verneuil, which we defined as a polje, and whose sources behave almost all as inversac or estavelle. Thus the water table flow damming downstream from Verneuil generates its rising and its discharge towards the Vigne, during high water periods. The piezometric statements confirm this relation: the sub surface water table is drained by the rivers then spreads out between Verneuil sur Avre and Rueil la Gadeli re. 4 model of karstic evolution in 3 phases Several elements attract attention. What is the role of the polje of Verneuil? What is the significance of these hydrological exchanges between the two groups of springs which invert according to the water leve l, etc? The geomorphological and hydrogeological analysis led to an assumption of karst genesis and evolution in three great phases for the large 'AvreVigne' Basin. 1The Upper Avre Basin is composed of two individualized systems. In the northern part, the Upper Avre River runs out from the WSW (Perche region), passes through Verneuil and is prolonged along the same axis towards the ENE towards the centre of the Paris Basin. Upstream of Verneuil, there is the junction of the two basins of Buthernay and L amblore creeks. In the SE, the small Lower Avre River develops towards the East. 2The tectonic activation of a fault located downstream from Verneuil creates a dam effect on the Upper Avre course, which generates the development of a polje whose water i s flowed out and links the springs of the Vigne. The Vigne catchment now feeds the Lower Avre River. The Lamblore and the Buternay creeks feed more and more the springs of the Vigne and thus reduce their contribution to the flow of the polje of Verneuil. 3A highly carved valley incises along the N S tectonic feature by retrogressive erosion and captures the water flow of the Upper Avre River, consequently reducing the activity of the polje of Verneuil. The Lower Avre River is now directly sustained by the Upper Avre River and by the springs of the Vigne. The karstic links between the polje and the springs of the Vigne only express temporarily, as it is demonstrated by the hydrochemical tracings. This gradual diagram must be completed and adapted to the many data in process of study. Some of these results require to be confirmed, which could be the object of further studies in karstology and hydrology. 5. Conclusions In 1900, dye tracing tests carried out during high water periods have reached the Nouvet springs but not the other emergences. In 1992, a new injection from the same introductive sinkpoint and under rather close hydrological conditions has attained all the resurgences. That seems to indicate that the flow conditions have been modified over t he past century just likewise the hydrochemistry which also rose in the same time. Field campaigns of calibration of the outlets carried out in the 1970s have disturbed the behaviour of the springs and induced a significant increase of the values in Esche richia coli and other particles. This anthropic deterioration of the quality of the aquifer seems to be established for a long time. The management and control of water supply requires taking into consideration the karstic dimension of these large aquifers

PAGE 307

Quickly evoked elements, such as tectonic constraints or hydrochemistry, are objects of ongoing studies which have been difficult to develop in this approach. The objective of this first contribution was not to present a succeeded study of this large basin in the chalk limestones of the Western Paris Basin. The mere interest was rather to propose several research orientations and to show that the chalk limestones of the Northwestern Europe offer a large variety of karst expressions, while these facies have for a long time been considered not to be sensitive to karstification because of their primary porosity. References BRARD, F. (1899). Etude des pertes de l'Avre et de ses affluents et sources en aval des pertes. Bulletin de la Socit des Ingnieurs C ivils DESCHESNES, G.T. (1675). Observations sur l'Avre DESPREZ, N. & C. MARTINS (1972). Prospection des captages de la Ville de Paris dans la rgion de la FertVidame (Eure et Loir) Rapport BRGM, Direction Dpartementale de l'Agriculture, 18 p. DIENERT F. (1901). Etude sur les sources de la Ville de Paris captes dans la rgion de l'Avre. Travaux des annes 1899 1900 sur les eaux de l'Avre et de la Vanne Rapport de la Commission Scientifique de Perfectionnement de l'Observatoire de Montsouris, Paris, 263 289. FERRAY, E. (1896). Hydrographie du dpartement de l'Eure. Ed. Hrissey, Evreux, 121 p. LAIGNEL, B. (1997). Les altrites silex de l'ouest du Bassin de Paris, caractrisation lithologique, gense et utilisation potentielle comme granulat Thse de l'Universit de Rouen, d. BRGM, 224 p. MAGNE, L. & B. WELTE (2007). The sensibility of the karst of chalk of the west of the basin of Paris: difficulty for the water manager, outlines. European Journal of Water Quality 38 (1), 7986. RODET, J. (1992). La craie et ses karsts Ed. CNEK Elbeuf & Groupe Seine/CNRS Caen : 560 p. SYKIOTI, O., B. DEFFONTAINES, J. CHOROWICZ, D. OBERT, G. DE MARSILY, J. LAUVERJAT, J. CARVALHO (1996). Imagerie numrique multisource de la surface topographique. Application la g omtrie d'un milieu karstique : Verneuilsur Avre (Perche). Bulletin de la Socit Gologique de France, 167 (2), 269 284 ========================== Figure captions Figure. 1 The Upper Avre Basin. a: Upper Avre river in Perche area, b: Upper Avre river in chalk area, c: Vigne Basin, d: Avre Vigne confluence, Vigne: Vigne Springs area, Be: karst of Beauche, Bo: karst of Bois Spert, Mo: karst of Morvilliers, Ro: hamlet of Rohaire; FT: town of La Fert Vidame. Dotted line: approximate limit between the for mation of the Sands of Perche (SW) and the chalk of the Upper Cenomanian (NE). Figure 2 Springs of the Upper Avre and Vigne Rivers. VA : town of Verneuil sur Avre; RG : town of Rueil la Gadelire; UA Upper Avre River; LA Lower Avre River; BU Buternay cr eek; LA Lamblore creek; VI Vigne River. Group of the Upper Avre springs : 1Gonord; 2Duchesnes; 3 Lesieur; 4 Poley; 5 Breuil; 6 Lavalette; 7 Trois Mulets; 8Launay. Group of the Vigne springs: 9 Rivire; 10 Foisy; 11 Graviers; 12 Erigny (Buternay creek); Le Nouvet (Lamblore creek): 13 Blaou; 14 Ganderolle; 15 Chne. Figure 3 La Brosse marl pit (Morvilliers). This picture of 1911 shows a flooded karstic drain trepanned by the marl gallery (coll. L. Magne). Figure 4 Upper Avre Basin evolution 1st st age: the Perche and Vigne catchments supply the Upper Avre River. The fault near Verneuil still does not affect the water course of the Upper Avre downstream from Verneuil. Figure 5 Upper Avre Basin evolution 2nd stage: the fault near Verneuil blocks the water course of the Upper Avre River and the downstream course is definitely disconnected. The polje of Verneuil concentrates waters from the two water catchments. The excess is conducted by karstic piracy to the springs of the Vigne River which supply the Lower Vigne River. The connection between the Vigne catchment and the polje reduces whereas the karstic piracy between the Vigne catchment and the Vigne springs increases. Figure 6 Upper Avre Basin evolution 3rd stage: the ultimate piracy of t he Upper Avre River made by a meandering and incised valley from the Lower Avre River reduces the functioning of the polje and the karst linkage between the polje and the springs of the Vigne River. Karst links from the Vigne catchment to the polje and fro m the polje to the springs of the Vigne River only connect during high water episodes.

PAGE 314

15th International Congress of Speleology Speleogenesis 909 2009 ICS Proceedings SPELEOGENESIS OF THE HYPOGENIC CAVES IN CENTRAL ITALYMARc C O Me E Niche ICHE TTi I Istituto di Scienze della TT erra, Universit di Urbino, I-61029 Urbino, Italy In Central Italy, from the Tyrrhenian to Adriatic coast through the Apennine karstic mountains, there are many limestone caves of various sizes and origins. Several of them consist of three-dimensional maze systems with several tens of kilometres of solutional phreatic passages with galleries and shas characterized by large rooms, cupola roofs, blind pits, anastomotic passages. ey contain fossil galleries as well as phreatic conducts with active streams of highly mineralized water, mainly rich in hydrogen sulde. Here the limestone walls are still corroding and transforming calcium carbonate to gypsum. is main cave-forming processes is related to the H2S oxidation to sulfuric acid, in the groundwater and in the atmosphere where the redox reactions involve chemotrophic microbial activity. e whole region is presently rich in volcanic, crustal, and mantle-derived CO2 and H2S emissions and outcrops of uaternary travertine deposits. Hypogean speleogenesis is the main process for the largest caves forming in Central Italy and can be supposed for many other surrounding karst systems with similar geology and uid chemistry.1. IntroductionIn the past few decades there has been an interest in speleogenesis related to deep-seated hydrogeologic recharge, where limestone corrosion is driven by endogenic agents (Klimchouck, 2007). Hypogenic caves are well known in dierent part of the world from Central Asia to North and South America (Palmer, 2007; Klimchouk, 2007), and especially in the underground fossil system in the Guadalupe Mountains in New Mexico and Texas (Du Chene et al., 2000). e important role played by microbial activity in sulfuric acid speleogenesis is recognized, as are the associated sulfur-redox bacterial communities that generate sulfuric acid as a metabolic product (Hose et al., 2000; Engel et al., 2004; Macalady et al., 2006). Caves in Central Italy north of Rome are probably the worlds best place to observe active sulfuric speleogenesis (Galdenzi and Menichetti, 1989; 1995). Here there are many limestone caves with dierent variety and size of morphologies composed of three-dimensional maze systems and deep wells with active sulfuric streams, where both active and fossil gypsum deposits can be observed. e whole region is presently rich in volcanic, crustal, and mantle-derived CO2 and H2S emissions, as well as many uaternary travertine deposits (Fig. 1). Karst systems associated with active sulfur springs are scattered in many other regions in Italy from marine springs of Capo Palinuro in southern Italy (Mattison et al., 1998) to southwestern Sardinia, Sicily, and Apulia (Fig. 1; Menichetti, 1994). e geological factors associated with sulfuric acid cave formation show great variety and unusual characteristics in central Italy. Although the main speleogenetic reactions are known, the geological, hydrogeological, and geochemical conditions need to be documented, especially the activities of the gases (H2S CO2) and their association with mineral species. Hydrogeology and especially hydrochemistry are essential for understanding the space-time evolution of these hypogenetic karsts.2. Geological Settinge geology of central Italy is mainly the result of the continental collision processes between the Corsica/ Sardinia and the subducted Adriatic plates during the Cenozoic. e geological and geophysical data highlight two main sectors within the region: a western Tyrrhenian side with Neogene-uaternary active back-arc extensional tectonics, and an eastern Adriatic side with an active compressional stress eld (Cavazza and Wezel, 2003). e karstic carbonate Apenninic fold-thrust belt is located in a transitional area between these two domains (Fig. 1). e Tuscan-Umbro-Marchean sedimentary cover, which hosts the caves and is part of the Meso-Cenozoic basin, consists of three main lithological units: the lower one is about 1 km of Upper Triassic dolomites and anhydrite unconformably overlying Paleozoic phyllitic basement rocks; an intermediate sequence of limestone and pelagic cherty-marly-carbonates, about 2500 m thick, dates from the Jurassic to the Paleogene. is in turn underlies Neogene turbidite foredeep sediments a few thousand meters thick. In western Tuscany, and especially in the Apuane Alps, the carbonate succession is metamorphosed to greenschist facies and represents the deep roots of the collisional orogen.

PAGE 315

Speleogenesis 910 2009 ICS Proceedings 15th International Congress of Speleology In the Rome area, since mid-Pleistocene (about 700 Ka), and remaining intermittently active until recent, there has been a rich K-undersaturated volcanism associated with some carbonatite magmas extruded in several places in the Apennine chain (Cavazza & Wezel, 2003). On the Tyrrhenian side of the Apennine belt, the results of the Neogene-uaternary backarc extensional process are a reduced thickness of the lithosphere, a system of NW-SE -striking normal faults and associated basins, and high heat ow, with many hydrothermal vents rich in H2S and CO2 (Minissale, 2004). On the Adriatic side of the Apennines, mud volcanoes, salt springs, and CH4 emissions are well documented with many data derived from hydrocarbon explorations (Conti et al., 2000). e largest caves in the area are developed in a thick 1000 m carbonate bank of Jurassic age, where a syngenetic porosity with sedimentary facies of packstone and grainstone is well developed. In places, small caves are hosted by a Cenozoic Figure 1: Map of the main karst features in central Italy. Upper right: location map of documented hypogenic caves in Italy (circles). Legend: a Stratigraphic sequence om uaternary to Cretaceous; b uaternary olcanic rocks; c Main calcareous sequence om Jurassic to Cretaceous; d metamorphosed calcareous stratigraphic sequence om Jurassic to Cretaceous; e major karst spring; f major mineralized spring; g important caves; h caves with gypsum deposits; i closed valley with karstic drainage; j spring rich in CO2 gas; k CO2-rich gas vent; l Apennines drainage divide; m Tuscany geothermal anomaly (thick dotted line represents 100mW/m, in Larderello geothermal eld; the heat ux reaches 400 mW/m).

PAGE 316

15th International Congress of Speleology Speleogenesis 911 2009 ICS Proceedings marly-carbonate sequence conned by sandstones and marls (Menichetti, 1987). uaternary travertine deposits are scattered throughout the area, especially in Tuscany and Latium but in only a few localities in the Apennine chain associated with both thermal and cold springs (Fig. 1; see Minissale, 2004). e main phase of Apennine upli related to cave development took place within the Pleistocene (Mayer et al., 2001).e primary tectonic features that control the underground Apennine karst morphology and the carbonate aquifer drainage are a system of transpressive faults with a N-S trend and networks of conjugated joint sets, distributed in a primary NE-SW and secondary NWSE directions. In particular, the N-S faults have guided the main galleries and rooms and controlled the development of the larger underground passages, while the joint systems locally guide the smaller solutional passages (Menichetti, 1987; Mayer et al., 2001).3. Hydrogeology and Geochemistrye regional drainage network is quite complex because of the vertical range between the highest elevations and the drainage divide, in consequence of the eastward migration, since Neogene, of the compressional and extensional tectonic stress eld (Fig. 1; see Cavazza & Wezel, 2003). In many cases there have been superimpositions, antecedences, and regressions of the river networks with respect to the anticlinal structures. e regional aquifers are located in the Jurassic carbonate strata, and groundwater supplies the springs in the lower valleys and along the main fault zones with an average base ow of 22 L/sec/km2 (Galdenzi & Menichetti, 1995). Aquicludes, represented by marly layers, are distributed at various levels in the stratigraphic succession. Groundwater ow in the transfer zone is controlled by karst conducts and ssures, while faults and joints guide the regional drainage in the carbonate reservoirs. e hydrodynamics of larger springs is inuenced mainly by the base ow, while in some cases rapid ows take place in the transfer zone. All karst systems have basal input points along faults and joints at the bottom of the oxidizing zone, where mineralized water rises from deep-seated ow systems with transit times of many months. Flow in the vadose zone is regulated by conduit systems, with transit times of a few weeks (Sarbu et al., 2000). In several karst areas, close to the water-table, are suldesodium-chloride mineralized groundwaters combined with CO2-rich meteoric circulation. e carbonate waters are from inltration and seepage from the surface, with about 500 mg/L of total dissolved solids. Mineralized waters, with temperature ranging from 14 to 40C, and more than 2000 mg/L TDS, rise from depth, involving Triassic anhydrite Figure 2: Geochemistryl of the karst waters in central Italy. Sti diagrams on the le and Piper diagram on the right. Location of the springs is shown in Figure1.

PAGE 317

Speleogenesis 912 2009 ICS Proceedings 15th International Congress of Speleology in a complex underground regional ow path. Here, the H2S can reaches concentrations up to 0.5 mMol/L and the endogenic CO2 in the water can have values close to 100 mg/L (Fig. 2). e 18O/D isotopic content shows that the groundwater circulation involves large structures, with a recharge area in the highest Apennines. e hydrothermal circulation can reach the Triassic anhydrites at depth, and aer a journey of some decades, rises to the springs located close to the master faults along the border of the main karstic structures (Menichetti et al., 2008). roughout central Italy are several H2S, CO2, and CH4 low-temperature gas emissions, many of them localized close to the important karst systems (Fig. 1). e volume of uid ow through the carbonate rocks is very large, especially considering a CO2 estimate in 1011 mol/yr (12,000 t/day) (Rogie et al., 2000) according with the relevant outcrop of travertine deposits. e origin of the gases is still debated, both for CO2 and H2S, but is oen associated with CH4 and He (Minissale et al., 2004). e origin of non-volcanic CO2 seems to be mantle degassing and subsequently by thermogenic reaction with carbonates. e H2S results from gas reactions and re-equilibration in rock/mineral-buered geothermal systems within the buried Triassic anhydrites (Minissale et al., 2000).4. Cave patterns and speloegenesisIn Central Italy the main caves consist of systems with a few tens of kilometres of phreatic and vadose solution passages and shas, with large rooms, cupola forms, blind pits, and anastomotic passages (Fig. 3). roughout the region all the developmental stages of the hypogean caves are present in still-active caves in the Frassassi Gorge, Parrano Gorge, Saturnia area, Acquasanta Terme area, and Pozzo del Merro. Here the rising of highly mineralized water rich in sulfur can be observed in underground streams and pools. Elsewhere are fossil caves such as those in Monte Cucco, where large gypsum speleothems are present more than 1000 m above the current regional water table. e Pozzi della Piana Caves are developed in uaternary travertine, where gypsum deposits are associated with the phreatic passages, along with cupola and blind pits (Fig. 2). Fossil gypsum deposits can also be found in many caves, such as Pozzi della Piana, M. Rotondo, and Citt Reale (Fig. 1). Many of these caves have no relation to external drainage networks, and several entrances have volumes that are inconsistent with classic carbon dioxide karst corrosion. e caves in both active and fossil systems consist of large rooms and mazes with galleries branching outward from them with frequent phreatic features and horizontal tubes. Phreatic passages, many of them anastomotic, also extend over large parts of the cave, where they constitute some network zones (Fig. 3b). Shas and ssures in the cave oor represent the original sources of these H2S-rich waters. Large rooms, such as the Abisso Ancona Hall in Grotta Grande del Vento-Frasassi (about one km3 in volume) and those in Monte Cucco (in which many galleries are voids > 0.2 km3) are usually regular in pattern and controlled by joints and faults that governed the ows of meteoric groundwater and rising sulfuric water. Very deep pits show a corkscrew solutional pattern in the fossil caves of Monte Cucco in the Apennines, and in the still-active Pozzo del Merro near Rome (Fig. 3a), which reaches 332 m below sea level (Caramanna and Malatesta, 2002). Rising sulfuric water is the main mechanism of the hypogenic corrosion of these shas along deep hydrothermal ow paths guided by the principal tectonic lineaments of the region. On the other hand, several caves, such as Monte Cucco and Frasassi, are developed at dierent levels with vertical ranges of 100 m linked to the evolution of the regional groundwater base level. Smaller karst systems have a ramiform pattern of several large rooms with wide ceilings that end abruptly in narrow passages or ssures (Fig.3b). Phreatic morphologies with bubble trails are widespread in both active and fossil karst systems, indicating the action of aggressive gases during the ascent into the upper portions of the phreatic zone. In the whole region, the main cave-forming processes can be related to H2S oxidation to sulfuric acid acting in the groundwater as well as in the atmosphere, where the redox reactions involve chemoautotrophic microbial activity (Galdenzi et al., 1999). e carbonate corrosion produces sulfate ions in the phreatic zone and gypsum replacement of the limestone walls of the vadose sectors of the caves (Menichetti et al., 2008). e sulfuric pools contain stratied water with signicant physico-chemical variations of temperature, TDS, and pH at the interface between carbonic and sulfuric groundwater (Sarbu et al., 2000). Surfaces of sulfuric streams contain much organic matter, such as white laments that in the deep ooded pits can result in meter-long seaweed-like strands. Sulfur-oxidizing bacterial communities are known in the Frasassi caves that use H2S as an energy source in chemoautotrophic aphotic ecosystems that support invertebrate life (Sarbu et al., 2000; Macalady et al., 2006). e role played by this microbial community on the limestone corrosion in not well dened, nor is the relationship between the presence of organic matter and the released H2S in the atmosphere in other karst systems (Engel et al., 2004). Close to the sulfuric streams the air is rich in H2S released from the groundwater, and this accounts for most of the limestone wall corrosion. e corrosion operates on

PAGE 318

15th International Congress of Speleology Speleogenesis 913 2009 ICS Proceedings limestone as small white spots about one cm in diameter, where calcium carbonate is replaced by gypsum. is microcrystalline gypsum is oen so and can fall easily to the oor, building larger deposits and possibly owing as gypsum glaciers. Otherwise, wall-replaced gypsum crusts can contain centimeter-size recrystallized selenite crystals from preexisting gypsum in solution. In both cases the resulting wall morphology consists of limestone solution pockets that give evidence for the corrosion by sulfuric acid. At the limestone/gypsum interface a sulfur-oxizing bacteria biolm plays an important role in the limestone corrosion. is biolm consist of organic mucous matter arranged in small strands like spider webs and thin stalactites with acid droplets of pH <1 (Galdenzi et al., 1999; Sarbu et al., 2000). e gypsum deposits have 34S ratios between -1 and 20 in the fossil deposits of Monte Cucco, Frasassi Gorge, Pozzi della Piana, and Acquasanta caves. e presently forming gypsum has 34S values ranging from 6 to 10, with lighter values in the recrystallized crusts. In the mineralized groundwater, 34S in the sulfate varies from +19 to +21, while 34S in sulfur is -14.3, which is close to the values of the actively forming gypsum. is gypsum is later derived from the oxidation of H2S released from the groundwater (Menichetti et al., 2008). e limestone corrosion rate by sulfuric acid was tested over a 5-year experiment with calibrated limestone tablets, located in the sulfuric branch of the Frasassi caves, both in air and in the water. e tablet surfaces were replaced with gypsum at a rate of limestone corrosion of about 0.05 mm/y (Galdenzi et al., 1997).5. DiscussionUnderstanding the unusual pattern and morphologies of the caves of central Italy needs to take in account the geology, hydrogeology, and the water and gas chemistry that leads to hypogean speleogenesis. Karst in the region is not homogeneously distributed, and in many places the cave patterns are not related to the surface geomorphology. In the Acquasanta and Saturnia caves, and those north of Rome, the thermal water, rich in H2S as well as endogenic CO2, plays a supplemental role in cave evolution. It is important to consider the positive feedback between H2S oxidation and the releasing of new CO2 in the upper part of the groundwater to provide a complementary aggressiveness toward carbonates. e cave patterns show that the oxidation zone for H2S is not restricted to shallow groundwater levels but can extended through the deep parts of the aquifer where there has been input of fresh water in complex regional ow systems. e presence of a large uid ows opens further questions about the role of the aggressive gases CO2 and H2S in the speleologensis processes. e values of PCO2 in groundwater of central Italy range from 0.03 to 0.1 atm, which provides more than ten times the CaCO3 solubility than normal karst waters. e breakthrough mechanism involving progressive fracture widening by epigenic CO2 corrosion is modied to a homogeneous corrosion of the fracture walls along their entire length by the presence of endogenic CO2. An increase of endogenic PCO2 of 0.002 atm, is sucient to reduce the Figure 3: Patterns of caves in central Italy. a cross section of the Pozzo del Merro compared with the deeper part of the Monte Cucco Cave; b plan view of major hypogenic caves compared with a few branches of the Grotta Grande del Vento (Frasassi) and Monte Cucco Caves.

PAGE 319

Speleogenesis 914 2009 ICS Proceedings 15th International Congress of Speleology breakthrough time for a fracture to about half (Gabrovsek & Dreybrodt 2000). Moreover, the cooling during the ascent of thermal water along the conduits increases the CO2 aggressiveness, with the corrosion acting almost uniformly along the surfaces, producing a dramatic increase in karst void development and ow rate (Andre & Rajaram, 2005). e complex chemical reactions between the dierent minerals in contact with the carbonate rocks, and the presence of signicant concentrations of Cl and Na in several karstic groundwaters, gas/water reactions, and the role played by organic matter are topic for future research.ReferencesAndre, B.J., and Rajaram, H. (2005), Dissolution of limestone fractures by cooling waters  : wearly development of hypogene karst systems. Water Resources Research, 41, w01015, doi 10.1029/2004WR03331. Conti, A., Sacchi, E., Chiarle, M., Martinelli, G., and Zuppi, G.M. ( 2000), Geochemistry of the formation waters in the Po plain (Northern Italy): on overview. Applied Geochemistry, 15, 51-65. Cavazza, W., and Wezel, F.C., Eds. (2003), Geology of Italy. Episodes, 26:3, 268 p. Caramanna, G., and Malatesta, R., 2002, Il Pozzo del Merro http://www.techdivers.net/. DuChene, H.R., Hill, C.A., Hose, L.D., and Pisarowicz, J.A. (2000), e caves of the Guadalupe Mountains: Research symposium. Journal of Cave and Karst Studies, 62:2, 1-159. Engel, A.S., Stern, L.A., and Bennet, P.C. (2004), Microbial contributions to cave formation: New insights into sulfuric acid speleogenesis. Geology, 32:5, 369-372. Gabrovsek, F., Menne, B., and Dreybrodt, W. (2000), A model of early evolution of karst conduits aected by subterranean CO2 sources. Environmental Geology 39, 531-543. Galdenzi, S., and Menichetti, M. (1989), Evolution of underground karst systems in the Umbria-Marche Appennines in central Italy. In Proceedings, 10th International Congress of Speleology, Hazslinszky, T., and Takacsne, K., Eds., Budapest, 3, 745-747. Galdenzi, S., and Menichetti, M. (1995), Occurrence of hypogenic caves in a karst region: Examples from central Italy. Environmental Geology, 26, 39-47. Galdenzi, S., Menichetti, M., and Forti, P. (1997), La corrosione di placchette calcaree ad opera di acque solfuree: Dati sperimentali in ambiente ipogeo. In Proceedings, 12th International Congress of Speleology, Jeannin, P.-Y., Ed., Le Chaux-de-Fonds, Switzerland, 1, 187-190. Galdenzi, S., Menichetti, M., Sarbu, S., and Rossi, A. (1999), Frasassi caves: A biogenic hypogean karst system? In Proceedings of European Conference on Karst 99, Audra, P., Ed., Universit de Provence, Etudes de Gographie physique, suppl. 28, 101-106. Hill, C.A. (1987), Geology of Carlsbad Cavern and other caves in the Guadalupe Mountains, New Mexico and Texas. New Mexico Bureau of Mines and Mineral Resources, Bulletin, 117, p. 150. Hose, L.D., Palmer, A.N., Palmer, M.V., Northup, D.E., Boston, P.J., and DuChene, H.R. (2000), Microbiology and geochemistry in a hydrogensulphide-rich karst environment. Chemical Geology 169, 399. Klimchouck, A. (2007), Hypogene speleogenesis: hydrogeological and morphogenetic perspective. National Cave and Karst Research Institute, Carlsbad, N.M., USA, Special Publication 1, 106 p. Macalady, J.L., Lyon, E.H., Koman, B., Albertson, L.K., Meyer, K., Galdenzi, S., and Mariani, M. (2006), Dominant microbial populations in limestonecorroding stream biolms, Frasassi cave system, Italy. Applied Environmental Microbiology 72, 5596. Mayer, L., Menichetti, M., Nesci, O., and Savelli, D. (2003), Morphotectonic approach to the drainage analysis in the North Marche region, central Italy. uaternary Intern., 101-102, 157-167. Mattison, R.G., Abbiati, M., Dando, P.R., Fitsimons, M.F., Pratt, S.M., Southward, A.J., and Southward, E.C., (1998), Chemoautotrophic microbial mats in submarine caves with hydrothermal suldic springs at Capo Palinuro, Italy. Microbial Ecology, 35 58-71.

PAGE 320

15th International Congress of Speleology Speleogenesis 915 2009 ICS Proceedings Menichetti, M., (1985), Caratteristiche chimico-siche delle acque carsiche dellAppennino umbro-marchigiano. Atti Conv. Naz. Biospeleologia, Citt di Castello, 33-73. Menichetti, M. (1987), Analisi spazio-temporale del sistema carsico del M. Cucco. Atti XV Cong. Naz. Spel. Castellana Grotte, Bari, 731762. Menichetti, M. (1994), Grotte ipogeniche. SpeleoCAI, Costacciaro Perugia, 5:13, 34-37. Menichetti, M., Chirenco, M.I., Onac, B., Bottrell, S. (2008), Depositi di gesso nelle grotte del M.Cucco e della Gola di Frasassi, Considerazioni sulla speleogenesi. Mem. Ist. Ital.Speleol., II, XXI, 308325. Minissale, A. (2004), Origin, transport and discharge of CO2 in Central Italy. Earth Science Reviews, 66, 89-141. Minissale, A., Magro, G., Martinelli, G., Vaselli, O., and Tassi, F. (2000), Fluid geochemical transect in the Northern Apennines (centralnorthern Italy): uid genesis and migration and tectonic implications. Tectonophysics 319, 199. Palmer, A.N. (2007), Cave Geology. Cave Books, Dayton, Ohio, USA, 453 p. Rogie, J.D., Kerrick, D.M., Chiodini, G., and Frondini, F. (2000), Flux measurements of non-volcanic CO2 emission from some vents in central Italy. Journal of Geophysical Research, 105, 8435-8445. Sarbu, S. M., Galdenzi, S., Menichetti, M., and Gentile, G. (2000), Geology and biology of the Frasassi Caves in Central Italy, an ecological multi-disciplinary study of a hypogenic underground ecosystem. In: Ecosystems of the world, Wilkens, H., et al., Eds., New York, Elsevier, p. 359-378.

PAGE 321

Speleogenesis 916 2009 ICS Proceedings 15th International Congress of Speleology UNIUE DISSOLUTIONAL MORPHOLOGIES IN CAVES: A RESULT OF CONVECTIVE FLOW REGIMES?JO O HNN E. MYLRORO IE and JOAN OAN RR MYLRORO IED Department of Geosciences, Mississippi State University, Mississippi State, MS 39762, USA A A major group of caves, conned hypogenic caves, form gradually by slow ow under conned conditions and are said to have a diagnostic suite of dissolutional features including, but not limited to: cupolas, domepits, ceiling channels, ssures and ri-like infeeders, and thin bedrock partitions. e existence of these dissolutional features in epigenic stream caves is interpreted to indicate that speleogenesis had earlier initiated under conned, and most likely deep, conditions and had been overprinted by shallow epigenic conditions. It has since been demonstrated that ank margin caves in the Bahamas and Isla de Mona contain identical dissolutional features, yet these caves formed in a shallow environment in carbonate rocks that are still eogenetic and have never been in a deep burial or conned setting. Other workers had previously considered many of the diagnostic speleogens to be indicative of backooding in epigenic caves. What are the ow characteristics that create such dissolutional features? Flank margin caves are produced by fresh-water lens ow similar to the slow ow described for conned soluble rock settings. Vertical uid transfer can therefore occur within dissolutional voids by density ow without disruption by lateral fast ow. e density dierences for vertical ow can be produced by thermal variation, solute dierential, or uid degassing and entraining, or convection can be forced by pressure ow. In shallow ank margin caves, thermal variation is unlikely, however mixing of vadose fresh and phreatic water at the top of the lens can create a uid with higher solute load, and therefore density, than either starting uid. at uid would descend within the ank margin cave to set up a vertical ow regime. Organic material collects at the fresh-water boundary with underlying marine water, where it has been demonstrated that organic decay produces excess CO2, and also H2S. Degassing as a result of organic decay, with vertical bubble migration, could also entrain uid ow into vertical motion, with renewed dissolution at the cave roof by mixing with descending vadose water. In epigenic caves, backooding and downstream hydraulic damming can create stagnate, slow ow conditions. Vertical ow driven by density dierences becomes possible when lateral fast ow ceases. In this case, the unique speleogens produced are an overprint of an epigenic process, as opposed to being a conned ow speleogen overprinted by later epigenic processes.1. Introductione recent development of a model to describe cave genesis in deep, conned karst aquifers by slow ow ascending across conning aquicludes proposes that many epigenic caves are relict hypogenic caves participating in modern fast-ow regimes (Klimchouk 2007 and references therein). is model, called herein the conned hypogenic cave model, is identied in the eld by a suite of dissolutional bedrock forms, or speleogens, including but not limited to: cupolas, domepits, ceiling channels, ssures and ri-like infeeders, and thin bedrock partitions. e model explains that the appearance of these features in epigenic stream caves is the result of inheritance from a previous mesogenetic phase. It has been demonstrated (Mylroie and Mylroie 2008; in press) that identical speleogens are found in ank margin caves (Fig. 1). ese caves develop in the eogenetic environment of the fresh-water lens contained in carbonate islands and coasts (Fig. 2), and have never been conned or undergone deep burial (Mylroie and Mylroie 2007 and references therein). e presence of these speleogens in ank margin caves as well as conned hypogenic caves raises the question of whether these diagnostic speleogens might also be generated in stream caves by epigenic processes. 2. e Flow Regime for Speleogen Developmente processes that create the diagnostic speleogens observed in conned hypogenic caves cannot be observed directly, but the conned hypogenic cave model indicates that ascending ow is the critical component (Klimchouk 2007). Many speleogens show a geopetal orientation, that is, a response to the earths gravity eld resulting in a distinct vertical aspect.

PAGE 322

15th International Congress of Speleology Speleogenesis 917 2009 ICS Proceedings Figure 1: Bedrock dissolutional morphologies, or speleogens, om ank margin caves in the Bahamas and Isla de Mona, Puerto Rico. ese dissolutional features are nearly identical to those described in Klimchouk (2007) as being produced solely by conned hypogenic conditions in karst aquifers. (A) Rising chain of cupolas, Cueva del Agua Punte los Engelses, Isla de Mona. Compare to Plates 6C and 6D in Klimchouk (2007). (B) Ceiling channel, Cueva del Agua Sardinera, Isla de Mona, Puerto Rico. Compare to Plates 6G and 7A in Klimchouk (2007). (C) Domepit (sensu Klimchouk, 2007), Cumulous Cave, Crooked Island, Bahamas. Ruler in right center is 10 cm long for scale. Compare to Plate 9J in Klimchouk (2007). (D) Cupola with outlet, Hamiltons Cave, Long Island, Bahamas. Compare to Plate 9D in Klimchouk (2007). (E) Large Cupola, Hatchet Bay Cave, Eleuthera Island, Bahamas. Compare to Plate 6D in Klimchouk (2007). (F) A variety of ceiling cupolas, Hatchet Bay Cave, Eleuthera Island, Bahamas. Compare to Plates 6B, 6 C, 6I, 9B, and 11A in Klimchouk (2007). (G) Fissure and ri-like passage, Cueva del Agua Sardienra, Isla de Moan, Puerto Rico. Compare with Plates 5H and 5I in Klimchouk (2007). (H) Bedrock partition, Goat Cave, Long Island, Bahamas. Compare to Plates 12 E and G in Klimchouk (2007).

PAGE 323

Speleogenesis 918 2009 ICS Proceedings 15th International Congress of Speleology Cupolas, domepits (sensu Klimchouk 2007) and ceiling channels in particular display dissolution upward in the vertical sense. As the speleogens found in ank margin caves cannot be the result of uid pressure forcing water upward through overlying conning layers, there must be another way to generate such geopetal speleogens. In both ank margin caves and in conned hypogenic caves, lateral ow is either non-existent or very slow. e lack of rapid lateral ow allows other ow regimes to express themselves. ese ow regimes are not commonly found when exploring active epigenic stream caves. In conned hypogenic caves, convection can be forced by the upward movement of water under pressure, to generate a partial return ow pathway. e return ow could be generated by the density dierence of the water at the cave roof as it becomes saturated with solute, such that it is now denser than the ascending, less saturated ow. Given these caves are believed to form by upward dissolution (Klimchouk 2007), saturated water is constantly being created at the cave roof. ermal gradients may also assist in creating density dierences, and hence convection, in conned hypogenic caves. Ascending water will cool as it travels upward, increase in density, and sink. e initial force to create ascending water may be the result of the pressure of aquifer connement, but once the water has moved upward, it will experience a lower temperature regime and thermal convection may express itself. In ank margin caves, however, neither forced upward ow nor thermal gradients are likely. Convection can still be established by two possible mechanisms, both a result of the condition that at the margin of the fresh-water lens, the cave extends across most of the thickness of the lens (Fig. 2). One mechanism begins with vadose ow to the top of the lens. Mixing dissolution of vadose and phreatic waters can result in renewed dissolutional aggressivity in carbonates (Bgli 1980). Such mixing at the top of the fresh-water lens would create a uid having more solute than either initial component, so in an isothermal environment, it would be denser than either initial uid. As shown in the cartoon in Figure 3, the entry of this denser water into a ank margin cave causes the water to sink to the cave oor at the approximate position of the halocline (Fig. 2). is descending water displaces the bottom water of the cave, causing dissolution at the base of the cave as a result of the dissolutional aggressivity obtained by mixing this new fresh water with sea water (Plummer 1975). e descending water requires a return ow to the top of the chamber, where mixing with incoming vadose water generates cupolas and related ceiling topography. At the oor of the chamber, mixing dissolution creates slots, basins and other dissolutional oor topography. Once this topography becomes established on both the oor and the ceiling, it would tend to lock in the descending and ascending ows, to create a closed convection cell. Dissolution of this type has been suggested as a possible mechanism for the formation of a very specic type of cupola in ank margin caves, the bell hole (Birmingham, et al. in press). A second possible mechanism for generating speleogens in ank margin caves involves the decay of organics. Organic material collects both at the top of the fresh-water lens, and at the base of the lens, as both locations are regions of density contrast. e geochemical eects of organic decay to promote carbonate dissolution has been well established for blue holes (e.g. Bottrell et al. 1991). Subsequent work demonstrated that similar dissolutional activity occurred in ank margin caves as a result of organic loading of the lens, and eventual decay (Bottrell et al. 1993). ese studies showed that organic decay produces not only excess CO2 that promotes additional carbonate dissolution, but that signicant organic loading can drive the system into anoxic conditions, such that H2S-mediated dissolutional activity occurs. e evolution of these gases from decay at the halocline can create bubbles and entrain water ow upward, and across cave roofs (Fig. 3). As a result, dissolution by both water mixing and by gas absorption can result at the cave ceiling. is activity can create ceiling notches or chains of cupolas if the lateral movement is signicant, as shown in Figure 3. While the mechanisms presented here are merely speculative, they demonstrate that logical arguments can be applied to produce a selection of Figure 2: Diagram of the esh-water lens in a carbonate island, showing the upper and lower lens boundaries and mixing areas, and the development of ank margin caves under the ank of the enclosing landmass, in the distal margin of the lens. e vertical scale has been exaggerated, the lens does not dip as steeply as shown.

PAGE 324

15th International Congress of Speleology Speleogenesis 919 2009 ICS Proceedings dissolutional ceiling morphologies that are extremely similar to those seen in conned hypogenic caves. Flank margin caves form in the fresh-water lens by the enlargement of touching vug porosity into chambers, which subsequently intersect other enlarging chambers and groups of chambers to create the cave complexity seen in the eld today (Labourdette, et al. 2007). One consequence of this cave development style is that it creates numerous thin bedrock partitions between chambers and passages, commonly with holes in them. Similar to the conned hypogenic cave model, the development of ank margin caves occurs with little or no competition between adjacent sites of speleogenesis. Passages and chambers intersect randomly, and this intersection does not change the overall hydrologic regime, as it would in epigenic stream caves.3. e Case for Epigenic Stream Cavese origin of the unique dissolutional morphologies found in both conned hypogenic caves and in ank margin caves can be explained as an outcome of slow ow, which allows other ow styles to be expressed. Vertical ow under pressure, or convective vertical ow, can explain much of what is seen in these two cave types. What of the similar dissolutional forms seen in epigenic caves? ese dissolutional forms are abundant in epigenic caves, and Klimchouk (2007) argues that they represent inherited dissolutional forms from an earlier conned hypogenic origin. One could argue as well that these dissolutional forms have been inherited from an even earlier eogenetic origin, though we consider the successful transit of these voids from the eogenetic to the mesogenetic to the telogenetic environment highly unlikely. In epigenic caves, some of these forms have been ascribed to ood water conditions (Palmer 1991), a condition that at rst seems quite dierent from the proposed slow ow environments discussed for ank margin caves and conned hypogenic caves. Flood waters and epigenic caves need to be examined from two perspectives, input ooding and output ooding (Fig. 4). During input ooding, water input is of large magnitude, and passages may be unable to handle the high discharge, especially if passage constrictions or obstructions exist (Fig. 4A). Hydraulic heads are high, and water ow velocities are not only turbulent but commonly very fast. Output ooding occurs when base level rises, hydraulic heads decrease to zero for many cave passages, and ow velocities can drop signicantly; the cave is hydraulically dammed (Fig. 4B). At this time, lateral ow is much decreased, and Figure 3: Diagrammatic representation of possible dissolutional processes within ank margin caves that can produce bedrock morphologies similar to those found in conned hypogenic caves. A) Mixing locations (le), and decay of organics trapped at the halocline to produce CO2 and H2S (right). B) Closed cell conection creates cupolas and oor dissolution (le), open cell conection creates ceiling channels and chains of cupolas (right). See text for a full discussion.

PAGE 325

Speleogenesis 920 2009 ICS Proceedings 15th International Congress of Speleology vertical convective ow can be expressed. e same vadosephreatic mixing hypothesized for ank margin caves can occur, to form cupolas and create solute-dense water near cave ceilings. During ood events, increased vadose ow can be expected from meteoric input, and mixing dissolution thusly increased. e presence of organic material, commonly introduced by the initial ood pulse, can decay to release CO2 to both increase dissolutional aggressivity and to form bubbles to help drive vertical water ow. While the onset and end of output ood cycles may have high velocities, the middle stagnant phase does not. e assumption that certain dissolutional passage congurations and orientations found in epigenic stream caves are the result of inherited conned hypogenic cave features is not necessarily true. It is certainly likely that conned hypogenic caves, as paleokarst features, have been intersected by epigenic cave systems. What is unlikely is that these intersected features play a signicant role in epigenic cave genesis and function. Conned hypogenic caves form as caves that developed independent of each other, and connections are fortuitous and have no eect on aquifer ow. ese criteria are a key point expressed by Klimchouk (2007) for conned hypogenic cave genesis. For such a cave to be congured so as to provide the most ecient and eective pathway for fast, turbulent epigenic ow is a low probability situation. Models that explain epigenic cave development, taking into account rock characteristics, structural features such as dip and strike, faulting, folding and jointing, and climatic history, do an excellent job (e.g. Palmer 1991). Output ooding is an alternative explanation for some of the dissolutional morphologies thought to be unique to conned hypogenic caves, especially for those caves where there seems to be a pattern of cave development and utilization of passage segments with these morphologies (e.g. Alexander 2008).4. ConclusionsConned hypogenic caves are an important aspect of cave development. at they are poorly understood, and poorly represented in the cave literature is also true. e explorational bias associated with hypogenic cave development in general has been discussed by Mylroie (2003) and Klimchouk (2007). Hypogenic caves are decoupled from surface processes; they appear in the epigenic realm as paleokarst. e typical evidence cavers use to nd and explore epigenic caves does not work for hypogenic caves. Epigenic caves are coupled to the surface hydrology. Sinking streams, springs and karst windows are observable features that assist the cave explorer to locate and enter epigenic caves. Hypogenic caves interact with the surface in a random fashion. Access points are not intuitive. Furthermore, because hypogenic paleokarst caves are not coupled to surface hydrology, they are vulnerable to segmentation, overprinting, and most of all, they are sediment traps. Not only are hypogenic caves hard to nd, most are blocked or plugged with sediment. It is almost a certainty that in humid environments, epigenic caves have intersected paloekarst hypogenic caves. Some unusual cave chambers and passage segments in epigenic caves could well be inherited from the hypogenic realm. However, widespread utilization of hypogenic paleokarst as a signicant control of epigenic cave development does not appear to happen. e three large hypogenic caves of the American West (Lechiguilla, Wind and Jewel) are notable for their lack of interaction with current epigenic hydrology, which has been a signicant factor in their survival and access for exploration. Explorational bias not only distorts our understanding of the location, abundance and extent of hypogenic caves, it inuences our understanding of how these caves form. It is nearly impossible to explore conned hypogenic caves in their environment of formation when they are actively developing. We must infer their mode of formation by examining evidence aer these caves have become paleokarst Figure 4: Comparison of input and output ooding. A) Input ooding can ll all or part of the cave, but high ow velocities may also be present because the hydraulic head (h) is large. B) O