USF Libraries
USF Digital Collections

Holocene climate and hydrologic changes recorded in Tufa and Lacustrine deposits in Southern Yemen


Material Information

Holocene climate and hydrologic changes recorded in Tufa and Lacustrine deposits in Southern Yemen
Physical Description:
Sander, Kirk M
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:


Subjects / Keywords:
Oxygen isotopes
Wadi Hadramawt
Dissertations, Academic -- Environmental Science and Policy -- Masters -- USF
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


ABSTRACT: Tufa and lacustrine deposits are useful paleoclimate archives in reconstructing the early to middle Holocene climate and paleohydrology of southern Yemen's Wadi Idim and Wadi Sana, which are north-flowing tributaries to Wadi Hadramawt. Numerical age estimates and oxygen-isotopes are used to assess the onset and cessation of tufa formation and reconstruct the environment of lacustrine sediment deposition in the region in order to understand the broader early to middle Holocene hydrologic system.Numerical age estimates from the studied wadis show a correspondence between early to mid-Holocene humid-phase sediment deposition and the northward shift of the ITCZ, as documented in paleoclimate records from other East Africa -- Arabia -- India continental and marine sediments. The interval between ca. 10-5 ka B.P. corresponds to a period of greater availability of moisture from the Arabian Sea region. Increases in precipitation allowed for a lake and wetland systems to develop, and increased spring discharge contributed to the formation of the tufa. Within the lacustrine sediments are ostracodes, mollusks, and flora casts that are found in a much wetter climate compared to today's hyper-arid environment. This early to mid-Holocene humid phase corresponds with a more northerly positioned ITCZ, which shifted south to its present day position around 5,000 yr B.P.Oxygen isotope measurements from ostracods show a range of isotope values from ~ -4.0%₀ at approximately 10 ka B.P. to ~ -6.0%₀ at approximately 5 ka B.P. Theses values represent the early to middle Holocene pluvial phase. Changes in the oxygen isotopic signature represent a change in evaporation or a possible change in source.The early to middle Holocene humid phase also corresponds with periods of agricultural activity, which are being investigated by the archaeological team of the Roots of Agriculture in Southern Arabia Project (RASA). Research into the effects of climate change on human activities, specifically agricultural processes, is the focus of RASA. Southern Arabia offers not only a convergence of three major agricultural regions, but also preserves a sedimentary record of the climate shift that affected the region during the period of study.
Thesis (M.A.)--University of South Florida, 2006.
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Kirk M. Sander.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 74 pages.

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:
oclc - 162585348
usfldc doi - E14-SFE0001651
usfldc handle - e14.1651
System ID:

This item is only available as the following downloads:

Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200409Ka 4500
controlfield tag 005 20070806134551.0
007 cr mnu---uuuuu
008 070806s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001651
1 100
Sander, Kirk M.
0 245
Holocene climate and hydrologic changes recorded in Tufa and Lacustrine deposits in Southern Yemen
h [electronic resource] /
by Kirk M. Sander.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: Tufa and lacustrine deposits are useful paleoclimate archives in reconstructing the early to middle Holocene climate and paleohydrology of southern Yemen's Wadi Idim and Wadi Sana, which are north-flowing tributaries to Wadi Hadramawt. Numerical age estimates and oxygen-isotopes are used to assess the onset and cessation of tufa formation and reconstruct the environment of lacustrine sediment deposition in the region in order to understand the broader early to middle Holocene hydrologic system.Numerical age estimates from the studied wadis show a correspondence between early to mid-Holocene humid-phase sediment deposition and the northward shift of the ITCZ, as documented in paleoclimate records from other East Africa -- Arabia -- India continental and marine sediments. The interval between ca. 10-5 ka B.P. corresponds to a period of greater availability of moisture from the Arabian Sea region. Increases in precipitation allowed for a lake and wetland systems to develop, and increased spring discharge contributed to the formation of the tufa. Within the lacustrine sediments are ostracodes, mollusks, and flora casts that are found in a much wetter climate compared to today's hyper-arid environment. This early to mid-Holocene humid phase corresponds with a more northerly positioned ITCZ, which shifted south to its present day position around 5,000 yr B.P.Oxygen isotope measurements from ostracods show a range of isotope values from ~ -4.0% at approximately 10 ka B.P. to ~ -6.0% at approximately 5 ka B.P. Theses values represent the early to middle Holocene pluvial phase. Changes in the oxygen isotopic signature represent a change in evaporation or a possible change in source.The early to middle Holocene humid phase also corresponds with periods of agricultural activity, which are being investigated by the archaeological team of the Roots of Agriculture in Southern Arabia Project (RASA). Research into the effects of climate change on human activities, specifically agricultural processes, is the focus of RASA. Southern Arabia offers not only a convergence of three major agricultural regions, but also preserves a sedimentary record of the climate shift that affected the region during the period of study.
Thesis (M.A.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 74 pages.
Adviser: Rick Oches, Ph.D.
Oxygen isotopes.
Wadi Hadramawt.
Dissertations, Academic
x Environmental Science and Policy
t USF Electronic Theses and Dissertations.
4 856


Holocene Climate and Hydrologic Changes Recorded in Tufa and Lacustrine Deposits in Southern Yemen by Kirk M. Sander A thesis submitted in partial fulfillment of the requirements for the degree of Masters of Science Department of Environmental Science and Policy College of Arts and Science University of South Florida Major Professor: Rick Oches, Ph.D. Philip van Beynen, Ph.D. Robert Brinkmann, Ph.D. Date of Approval: July 19, 2006 Keywords: ostracodes, oxygen isotopes, Wa di Hadramawt, RASA, ITCZ, geomorphology, paleoclimate Copyright 2006, Kirk M. Sander


Acknowledgments I thank my parents and sister for the support they have given me through my education path. I thank Rick Oches for th e experiences and guidance through the process of my thesis and graduate school. I thank Dr. Jo y McCorriston and the rest of the RASA team including our Yemeni colleagues from GOAMM an d Nexen Petroleum. I could not have completed all of my field sampling, especially without having someone to talk to in the field, without the help of Scott Anderson. The comments and guidance from my committee members Phil van Beynen and Bob Brinkmann helped greatly with understanding of the system. I want to acknowledge Terry Quinn and Ethan Goddard at the USF College of Marine Science Paleolab for the use of the M ass Spectrometer for oxygen isotope analysis. The time and expertise of Brandon Curry at the Illinois State Geological Survey for identification and consultation on the ostrocad es is greatly appreciated. I thank the other graduate students on the RASA project especi ally Mike Harrower and Josh Anderson. Finally, I thank the National Science Foundation for grant BCS-0211490 Without this support this research would not have been possible.


i Table of Contents List of Tables iii List of Figures iv Abstract vi Introduction 1 Study Area 2 Regional Climatology 5 Holocene Monsoon Variability 7 Regional Paleoclimate Records 9 Rub al Khali and Arabian Shield 9 Speleothems 11 Marine Records 13 East Africa 14 India 16 Geologic Setting 18 Tufa Formation 20 Human Inhabitation 22 Methods 22 Field Sampling 23 Charcoal and Shell 24 Optically Stimulated Luminescence Dating 24 Water Samples 25 Loss on Ignition 25 Oxygen Isotopes 26 Results 29 Wadi Sediment Deposition 29 Numerical Age Estimates 32 Wadi Idim 36


ii 05-GBY-02 38 Paleo Lake 42 Ostracode Identification 43 Oxygen Isotopes 45 Carbon Isotopes 46 Loss of Ignition 47 Water Samples 50 Discussion 50 Correlation of Wadi Silts, Tufa, and Lacustrine Deposits 50 Oxygen Isotopes 51 Carbon Isotopes 52 Loss on Ignition 53 Water Samples 54 Paleoclimate Interpretation 55 Comparison with Regional Paleoclimate Studies 55 Oxygen Isotopes 55 Modern Analogue 58 Recent Recharge 59 Conclusion 61 Future Research 63 References 65 Appendix 1 72 Appendix 2 74


iii List of Tables Table 1 Compilation of radiocarbon dates from the 1998, 2000,2004, 34 2005 field seasons Table 2 Optically Stimulated Luminescence ages from th e 2004 field 35 season Table 3 Profile of tufa capped lacustrine deposit 05-GBY-02 40 Table 4 Loss on Ignition Data 48 Table 5 Oxygen Isotope data from gr oundwater and rainfall during the 50 2005 field season.


iv List of Figures Figure 1. Location of the studay area on the southern edge of the Arabian 3 Shield Figure 2. Location of the study area with the regional towns listed 4 Figure 3. Wind patterns across the Arabian Peninsula 5 Figure 4. a. Precip itation isohyets 6 b. Evaporation isohyets 7 Figure 5. 18 O curve from a speleothem in Oman 12 Figure 6. Cross section of the Dhofar mountains in Oman north of the 19 research area Figure 7. Picture of the wadi silt contact with the gravel terrace 30 Figure 8. Example of hanging wadi silt s abutted against the wadi wall 31 Figure 9. Representative numerical ag e estimates from along Wadi Sana 33 from north to south Figure 10. Numerical age estimate s fom Wadi Idim tufa mounds 37 Figure 11. Tufa mound in Wadi Idim 38 Figure 12. Lithologic profile representing 05-GBY-02 39 Figure 13. 05-GBY-02 40 Figure 14. The approximate location of th e paleo lake and interbedded tufa 43 And lacustrine deposits Figure 15. The left valve of Sarcypridopsis aculeate 44 Figure 16. Oxygen isotope trend with a simple regression trend line 45 Figure 17. Carbon isotope plot with trend 47


v Figure 18. Loss on Ignition data for 05-GBY-02 49 Figure 19. Present day oxygen isotopic valus measured from GNIP 52 Figure 20. Wind patterns over the Arabian Sea 57 Figure 21. A modern analogue for the marshlands of the early to middle 59 Holocene in Wadi Sana Figure 22. Comparison of the decrease in date palms in Wadi Idim, A. 1998 60 B. 2005 61


vi Holocene Climate and Hydrologic Changes Recorded in Tufa and Lacustrine Deposits in Southern Yemen Kirk M. Sander ABSTRACT Tufa and lacustrine deposits are useful paleoclimate archives in reconstructing the early to middle Holocene climate and pale ohydrology of southern Yemens Wadi Idim and Wadi Sana, which are north-flowing tribut aries to Wadi Hadramawt. Numerical age estimates and oxygen-isotopes are used to assess the onset and cessation of tufa formation and reconstruct the environment of lacustrine sediment deposition in the region in order to understand th e broader early to middle Holocene hydrologic system. Numerical age estimates from the stud ied wadis show a correspondence between early to mid-Holocene humid-phase sediment deposition and the northward shift of the ITCZ, as documented in paleoclimate records from other East Africa Arabia India continental and marine sediments. The interv al between ca. 10-5 ka B.P. corresponds to a period of greater availability of moisture from the Arabian Sea region. Increases in precipitation allowed for a lake and wetland systems to develop, and increased spring discharge contributed to the fo rmation of the tufa. Within the lacustrine sediments are ostracodes, mollusks, and flora casts that ar e found in a much wetter climate compared to todays hyper-arid environment. This early to mid-Holocene humid phase corresponds with a more northerly positioned ITCZ, which shifted south to its present day position around 5,000 yr B.P.


vii Oxygen isotope measurements from ostraco ds show a range of isotope values from ~ -4.0 at approximately 10 ka B.P. to ~ -6.0 at approximately 5 ka B.P. Theses values represent the early to middle Holo cene pluvial phase. Changes in the oxygen isotopic signature represent a change in ev aporation or a possible change in source. The early to middle Holocene humid phase also corresponds with periods of agricultural activity, which are being investigat ed by the archaeologica l team of the Roots of Agriculture in Southern Ar abia Project (RASA). Research into the effects of climate change on human activities, specifically agricu ltural processes, is the focus of RASA. Southern Arabia offers not only a convergen ce of three major agricultural regions, but also preserves a sedimentary record of the climate shift that affected the region during the period of study.


1 INTRODUCTION During the early to middle Holocene, ~10ka to 5ka B.P., the climate in southern Yemen was significantly differe nt from todays hyper-arid environment. A northward shift of the ITCZ approximately 10 ka B.P. led to a greater influe nce of the Southwest Indian Monsoon in southern Arabia, causing an increase in preci pitation, corresponding to the Holocene climate optimum ~10 5.5 ka B.P. (Glennie, 1998; Fleitman et al., 2003a,b; Overpeck et al., 1996). This shift al lowed for an increase in available moisture in the southern Arabian Peninsula. Increases in precipitation enabled early pastoralists to raise bovine species and practice agriculture as evidenced by numerous archaeological remains in the Wadi Sana region of hi ghland Southern Yemen (Harrower, 2006). This research is part of the Roots of Agriculture in Southern Arabia (RASA) project lead by Dr. Joy McCorriston at The Oh io State University and Dr. Rick Oches at the University of South Florida. RASA is a NSF-funded cooperative archaeological and geologic research project to explore interactio ns of humans with the environment and the origins of agriculture in southern Arabia. This multi-year project is analyzing human activities in the cont ext of changing environmental condi tions during the early to middle Holocene. The archaeological record is represented by numerous water management structures, possible ritualisti c sites, housing structures and a diversity of other archaeological remains throughout the study area (Harrower, 2006). In order to assess the geologic record of ch anging climate and hydrology during the early to middle Holocene in the region, I analyzed a tufa-ca pped lacustrine sequence in the headwaters region of Wadi Sana and a transect of tufa deposits along the middle reach of Wadi Idim.


2 These deposits record changes in the amount and source of ra infall in the region prior to the climatic shift to the present hyper-arid environm ent ~5 ka B.P. This research was conducted as part of the RASA field season in March 2005. Wadis Idim and Sana were surveyed for tufa and lacustrine deposits in order to provide data to further understand the paleoclima te of the region and build upon the 2004 field seasons research contributing to hydrologi c modeling of Wadi Sana (Joshua Anderson, M.S. thesis, in prep). Initial numerical age estimates led to the following hypothesis and research questions, which are explored in this thesis: Research Hypothesis: Onset and cessation of spring discharge a nd tufa formation in southern Yemen record shifts in the Arabian monsoon system and associated change in regional climate between semi-arid and hyper-a rid precipitation regimes. Research questions: 1. When and under what climatic conditions did tufa form in Wadi Idim and lacustrine deposits accumulate in Wadi Sana? 2. Was there a temporal lag between the mi d-Holocene precipit ation decline and cessation of tufa deposition and lacu strine and fluvial sedimentation? 3. Can relative contributions of groundwater discharge and meteoric inputs be assessed using oxygen isotopes measured in ostracode shells from lacustrine sediments inter-bedded with the tufa? 4. How does middle Holocene reconstructed paleoclimate in Wadi Idim and Wadi Sana compare with other paleoclimate r econstructions from the broader region? STUDY AREA Wadi Sana and Wadi Idim are locate d in the Hadramawt governate in southwestern Yemen. These wadis flow north into the larger Wadi Hadramawt, which


drains the interior of southern Arabia, ultimately discharging into the Arabian Sea (Figure 1). Wadi Hadramawt is also referred to as Wadi Masila east of the town of Sayun. Figure 1. Location of the study area on the Southern edge of the Arabian Peninsula. Wadi Sana, the focus of my research and of the RASA project, is approximately 80km long. Wadi Idim drainage basin is adjacent to Wadi Sana, although the main channel is ca. 50 km west. The villages of Ghayl bin Yumain and Dhabiah are indicated by the purple dots in Wadi Sana and Wadi Idim, respectively. 3


Figure 2. Location of the study area with the regional towns listed. Wadi Sana is indicated by the number 1 and Wadi Idim by the number 2. Note the regional physiographic features (Modified from McCorriston et al., 2002). Tufa mounds and tufa-capped lacustrine deposits in Wadis Sana and Idim are located near the villages of Ghayl bin Yumain and Dhabiah, respectively (Figures 1, 2). The tufa-capped lacustrine deposits in Wadi Sana are located about a kilometer north of Ghayl bin Yumain, which is 60 km south of the confluence of Wadi Sana with Wadi Hadramawt. Tufa mounds representing fossil spring deposits in Wadi Idim are located near the town of Dhabiah, which is approximately 20 km south of the village Tarim at the confluence of Wadi Idim with Wadi Hadramawt. 4


REGIONAL CLIMATOLOGY The Arabian Peninsula is affected by a monsoonal climate, fed by the Southwest Monsoon from the Arabian Sea and the Shamal from the northwest, blowing southwest across the Rub al Khali region (Figure 3). Effects of each of these wind systems are controlled by the position of the ITCZ. Figure 3. Wind patterns across the Arabian Peninsula. The Shamal blows from northwest southeast, across the Arabian Peninsula then turns to the southwest across the Rub al Khali (Glennie et al., 2002). Presently the region is characterized by a hyper-arid climate and receives less than 100 mm of precipitation annually. The rain comes during the spring and fall monsoonal seasons, if at all. Southern Arabia has maximum temperatures in the 20s C in January and in the 40 C in July with temperatures up to 56C in the summer (Harrower, 2006; WorldBank, 2004). Evapotranspiration in the region is approximately 100mm. This creates a hyper arid desert environment that is influenced by flash floods (Sayls) that 5


limit the amount of infiltration. Average precipitation in the study region, measured at the town of Seyun from 1981 to 2002, is 78 mm (Harrower, 2006). A Figure 4. a, Precipitation isohyets. Our research area is indicated by the red dot, located within the 100mm precipitation and evaporation isohyets (Al Sharhan et al., 2001). 6


B Figure 4b. Evaporation isohyets. Our research area is indicated by the red dot, located within the 100mm precipitation and evaporation isohyets. (Al Sharhan et al., 2001) Holocene Monsoon Variability The ITCZ was located further north during the middle to early Holocene, allowing for a greater influence of the summer monsoons in southern Arabia (Braconnot et al., 2000; Noblet et al., 1995; Kutzbach and Liu, 1997, Fleitman et al., 2003a,b; Rodrigues, Abell, and Krpelin, 2000; Abell and Hoelzmann, 2000; Thamban, Rao, and Schneider, 2002). Increases in precipitation over the Arabian Peninsula occurred with more water vapor coming from Africa and greater evaporation over the Arabian Sea 7


8 (Noblet et al., 1995). Based on climate m odeling by Noblet and others (1995), the northward shift of the ITCZ caused a 25% in crease in rainfall over eastern Africa and southern Arabia. The shamal in winter and the southwes t monsoon (Indian Ocean Monsoon) during the summer are the two major wind systems th at control rainfall in the region (Glennie and Singhvi, 2002; Weldeab et al., 2005; Rodrig ues et al., 2000; Abell and Hoelzmann, 2000). Woods and Imes projected rainfall increases to 200 mm/yr during the middle to early Holocene (Woods and Imes, 1995). This increase would provide sufficient rainfall to recharge aquifers and allow for the di scharge of springs (Woods and Imes, 1995). The return to more arid conditions du ring the middle Holocene, around 5 ka B.P., resulted from a southerly shift of the ITCZ to its current location (Braconnott et al., 2000; Glennie and Singhvi, 2002; Maye wski, 2004; Thamban et al., 2002). Greater influences of the Shamal over southern Arabia resulted from the southern migration of the ITCZ during the middle Holocene (Braconnott et al., 2000; Glennie and Singhvi, 2002; Mayewski, 2004; Thamban et al., 2002). Changes in the ITCZ greatly reduced the amount of precipitation by forcing the humid Southwest Monsoon to move parallel to the southern Arabian Coast (Glennie and Singhvi, 2002; Weldeab et al., 2005; Rodrigues et al., 2000; Abell and Hoelzmann, 2000). Current rainfall is well below 100mm/y ear (Clarke et al., 1987). Currently no dryland agriculture is practiced, although limited groundwater-irrigated cultivation is maintained in the region. With the weaken ing of the monsoon approximately 5 ka B.P., lower precipitation rates re duced groundwater recharge and increased effective evapotranspiration.


9 REGIONAL PALEOCLIMATE RECORDS Studies have been conducted throughout th e Arabian Peninsula, Arabian Sea, East Africa, and India to understand the early to middle Holocene c limate of the region. Each geographic region provides a glimpse into key regional changes and their timing that occurred between approximately 10 5 ka B.P. The regional climate saw a shift from hyper-arid to semi-arid from the Younger Dryas to the Holocene climate optimum, then back to the hyper-arid environment that currently ch aracterizes the region. Rub al Khali and Arabian ShieldSo uthern Arabia and Indian Ocean Climate change during the early to middl e Holocene is recorded in sand deposits, lake levels, and archeological sites throughout the Arabian Peninsula (Braconnot et al., 2000; Mayewski et al., 2004; Kutzbach a nd Liu, 1997; Enzel et al., 1999; Makhnach et al., 2004; Gupta et al., 2003; Jain and Tandon, 2003; Gle nnie and Singhvi, 2002; Wood and Imes, 1994; Haynes, 2001; Nicoll, 2003; Jorgensen and al Tikiriti, 2001; Abell and Hoelzmann, 2000; Hailemichael et al., 2002). Paleoclimate proxies from archives including speleothems, ocean cores, and lacust rine sediments, indicate that the climate changed from hyper-arid during the Pleistocene to a pluvial phase during the early to middle Holocene. The climate then shifted back to the current hyper-arid phase that has persisted since the middle Holocene. Archaeological evidence from the Rub al Khali provide a proxy for early to midHolocene increases in lake levels and an in flux of freshwater based on preserved hunting camps and remains of animals that are no longer found in the region, including jackal, fox, ostrich, gazelle, cheetah, and wolf (Frank, 1977). Other archaeologic evidence for a


10 hunting-and-gathering society are indicated by the number of flint implements and hearths found in the region (Fra nk, 1977). A hearth located in the southwest region of the Rub al Khali was 14 C dated to 5,090 200 (uncalibrated) (Frank, 1977). This coincides with the higher lake levels from freshwater inputs (Frank, 1977). In comparison to the Pleistocene lakes that were br ackish water, the Holocene lakes were freshwater (Frank, 1977). As the freshwater input was reduced, th e lakes may have turned brackish prior to drying. The extensive archaeological evidence for greatly expanded human activity throughout the Arabian Peninsula during the ear ly to middle Holocene may be considered a proxy for increased precipitation, as pres ent-day habitation occu rs only around oases. Analysis of lake sediment and obs ervation of high strandlines provides sedimentologic and geomorphic evidence of increased lake levels and influxes of freshwater during the early to middle Ho locene (Frank, 1977). Lacustrine sediments found in the Wahiba Sand Sea in the Sultana te of Oman show an expanse of grey sediment indicating remnants of ancient lake s (Raides et al., 2004). Within these grey sediments are numerous gastropod shells that require freshwater and are evidence of permanent lakes with vegetated banks (Rai des et al., 2004). Unco rrected radiocarbon ages on the gastropod Melanoides tuberculata cluster around 8,800 14 C yr B.P. (Raides et al., 2004). Sandstone deposits of the Mesa unit were dated by infrared stimulated luminescence (IRSL) to between 9,800 800 and 5,400 400 yr B.P. (Raides et al., 2004). The Mesa unit contains higher amounts of silt and clay with large burrows and termite galleries (Raides et al., 2004). These are interpre tated to represent increased sedimentation and precipitation during the early to middle Ho locene (Raides et al., 2004).


11 Aeolian sediments in the region show an inverse relationship to the early Holocene climate optimum. An increase in aeolian sediments correla tes to a reduction in precipitation throughout the Arabian Peninsul a. Dune formation was a three-stage process, beginning with a gradual accretionary phase starting at approximately 7 ka B.P., a middle phase of very rapid vertical accretion at 2-3 ka B.P., and the third phase of stasis since ca. 2 ka B.P. (Bray and Stokes, 2004). Transition from phase one to phase two is associated with a decrease in vegetation resu lting from a decrease in precipitation (Bray and Stokes, 2004). With less vegetation, ra pid accumulation of sediment for dune formation was occurring ~ 4.7 ka B.P. Complete desiccation of se diment created large volumes of material for dune formation (Bra y and Stokes, 2004 and Glennie and Singhvi, 2002). Data evaluated by Bray and Stokes ( 2004) show a reactiva tion of the desert system since 6 ka B.P. Speleothems Isotopic studies of speleothems provide a well established proxy for temperature, rainfall variability, atmospheric circulation changes, and vegetation changes (McDermott, 2004). Samples collected in Hoti and Qunf caves in southern Oman provide a record at approximately the same latitude as Wadi Sana (Fleitmann et al., 2003a, b). Oxygen isotopes were analyzed in order to interpret temperature and rainfall changes (Fleitmann et al., 2003 a, b, Fleitmann et al., 2004). Records from speleothems provide a 4 to 5 yr resolution record that can reconstruct the climate back 330 kyr (Fleitma nn, 2003a). Data from Qunf Cave in Oman representing the last ca. 10,000 years are s hown in Figure 5 (Fleitmann et al., 2003a).


Increases in precipitation between 10.3 ka and 9.5 ka B.P. are indicated by a sharp decrease in 18 O from -0.8 to -2, which remained level until about 7.5 ka B.P. A gradual enrichment of the signal from -2.0 to -0.9 occurred from 7.5 ka to approximately 2.7 ka B.P., when the speleothem growth ceased (Fleitmann et al., 2003 a, b). These values are below the measured present day oxygen isotopes values in the region (Bowen and Wilkinson, 2002; Bowen and Revenaugh, 2003; IAEA/WMO, 2001). Figure 5. 18 O curve from a speleothem in Oman. Note the decrease towards zero per mill during the Middle Holocene. The blue indicates increases in precipitation during the early to middle Holocene compared with present (modified from Fleitmann et al., 2003a). Oxygen-isotope m easurem ents of the speleo them are an indication of the am ount effect; which is directly related to the am ount of m o isture present in the air (F leitm a nn et al., 2003a). Modern com p arison of cave dr ip waters oxygen isot opic signature show that the spe l eothem s were deposited at or c l ose to isotopic eq uilib rium with the environm ent of the cave (Fleitm a nn et al., 2003a). The depleted 18 O during the early 12


13 Holocene indicates a northerly position of the ITCZ, contri buting increase d precipitation from the Indian Ocean (Fleitman et al., 2004; Fleitmann et al., 2003a, b). Marine Records Numerous sediment cores have been extracted from the Arabian Sea and have been analyzed to reconstruct the paleoc limate of the region using oxygen isotopes, laminated sediments, preserved aquatic fauna, and foraminifera upwelling (Sarkar et al., 2000; Gupta et al., 2003; Overpeck et al ., 1996; Staubwasser et al., 2002; Doose Rollinskin et al., 2001; Thamban et al., 2002) These indicators show that at approximately 10 ka B.P. there was an in crease in monsoon strength over the Arabian Sea (Overpeck et al., 1996, Gupta et al., 2003; Sarkar et al., 2000; Gupta et al., 2005). Records not only show what is occurring in the Arabian Sea, but how the global climate system is interacting with the region. Based on the 18 O signature in two sediment core s recovered off of the Indian sub-continent, a decrease in evaporation over pr ecipitation is interpreted from 10 ka to 6 ka B.P. (Sarkar et al., 2000). This signature is further reinforced by the 14 C age records of Globigerinoides sacculifer that constrained the 18 O record of Globigernoides ruber found in the northern Arabian S ea (Staubwasser et al., 2002; G upta et al., 2003; Gupta et al., 2005). Oxygen isotope measurements of fora minifera in the cores show an abrupt change to a stronger monsoon at 10 ka B. P., corresponding to a decrease of 1 in 18 O in G.ruber (Staubwasser et al., 2002). This decrease in 18 O is attributed to fluctuations in the sea-surface temperature of the Arabia n sea from influences of solar insolation (Staubwasser et al., 2002).


14 The use of G. bulloides a cold water foraminifera, has been correlated using modern sea-floor samples and sediment traps to understand the paleo-temperature (Gupta, 2003). Upwelling of cold water in the Arabian Sea is recorded by the increases in G. bulloides An increase in precipitation during the early Holocene is correlated to the increase of G. bulloides (Staubwasser et al., 2002). As a proxy, G. bulloides is directly related to the summer monsoon upwelling in th e Arabian Sea and is not influenced from other sources of input to the system. There is a linear correlation between surface cooling from upwelling and a strong sensitivity to monsoonal atmospheric pressure gradient and wind sp eed across the Arabian Sea (Gupta et al., 2003). An increase in upwelling during the early to middle Holocene corresponds to an increase in strength of the monsoonal winds. As the ITCZ moved to its current southern position during the middle Holocene, there wa s a decrease in winds, resulting in a decrease in upwelling (Overpeck et al., 1996; Gupta, 2003; Gupta, 2005). Analyses of pollen found in marine sedi ment cores indicate influences of source winds at the time of deposition (Overpeck et al., 1996). With an increase in the influence of the SW monsoon during the early Holocene there was an increase in the amount of pollen from East Africa preserved in the core s (Overpeck et al., 1996). Weakening of the SW monsoon correlates to a sout hward shift of the ITCZ a llowing for an increase of pollen from the Asian steppe to be transpor ted to the study area (O verpeck et al., 1996) East Africa Paleolimnological studies from Africa indica te that lake levels were much higher than present approximately 9.7 ka B.P., with both playas and groundwater-fed lakes


15 forming in Sudan (Haynes, 2001). In a st udy of tufa mounds at Kharga, Egypt, Smith and others (2004) found evidence of increased la ke levels indicated by the erosion of the tufa mounds. The pluvial phase did not crea te enough recharge to produce tufa during the Holocene as opposed to the pluvial phase of the Pleistocene (Smith et al., 2004). The period of increased moisture was followed by a reduction in moisture that dried playas and reduced the size of groundwater fed lake s (Nicoll, 2004). Reduction in rainfall resulted in diminishing of vegetation, lead ing to increased aeolia n deposition (Nicoll, 2004). Pollen, diatoms, and invertebrate rema ins help to reinforce interpretations of a moister climate that ranged from a tropical savanna woodlands to grassland savanna during the early to middle Holocene (Haynes, 2001). Human habitation in Egypt and Sudan is indicated by the increased number of artifacts that are locat ed around lake deposits (Nicoll, 2004). Other indications of a more humid environment include vegetation not adap ted to arid conditions and the remains of many large and small mammals; including rh inoceros, elephants, oryx, and rodents (Nicoll, 2004). The aquatic snail Etheria elliptica was studied from early Holocene deposits in eastern Africa (Rodrigues et al., 2000). Environmental tolerances for Etheria elliptica are very constrained with intolerance to sa linity, alkalinity, and desiccation. These constraints indicate that sufficient am ounts of freshwater were present for Etheria elliptica to thrive (Rodrigues et al., 2000). Geoc hemical signatures i ndicate isotopically depleted water of -10 to -12 during the ea rly to middle Holocene. This water is thought to originate from the Atlantic as ev idenced by the isotopic depletion. Influences from the Atlantic were attributed to the nor thward movement of the ITCZ (Rodrigues et


16 al., 2000). Mollusks from other lakes in eastern Africa indicate oxygen-isotope data from mollusks show a lake level rise of 150 m dur ing the early Holocene in the central Afar lake system, and a nearby lake rose an es timated 80m (Hailemichael et al., 2002). On average the 18 O values averaged -2.1, compared with present day values of +16.00 (Hailemichael et al., 2002). India In the Indian subcontinent increases in precipitation during the early and middle Holocene are recorded in lake sediments, pollen, stable carbon isotopes, and river sediments (Gupta et al., 2003; Ramesh, 2001; Gasse and van Campo, 1994; Jain and Tandon, 2003). Lake levels throughout Asia, from the Indi an sub-continent to the Tibetan Plateau, indicate a climate fluctuation to a more pl uvial setting during the early Holocene, with increasing aridity from the middle to late Holocene (Gasse and van Campo, 1994). Increases in aquatic organisms, ostracodes, and diatoms, at ~10-9.5 ka B.P. in Lake Sumxi and Lake Bongon on the Tibetan Plat eau correspond with the increases in precipitation, in contrast to the lack of aqua tic organisms preserved in the sediment from ~11 to 10 ka B.P. (Gasse and van Campo, 1994). Pollen data from the lake corroborates the increase in precipitation along with 18 O data from the lake. Maximum humidity is recorded at 9 to 8.7 ka B.P. with a slight return to aridity from 8.6 to 7.5 ka B.P. A shift back to more humid conditions from 7.3 to 6.3 ka B.P. is inferred from the preserved biological activity record ed in the sediment (Gasse and van Campo, 1994). The lakes then experienced a reduction in water level


17 from ~6 to 3.8-3.7 ka B.P., based on the pollen record and 18 O values (Gasse and van Campo, 1994). The Tibetan Plat eau has not received significa nt monsoonal input to the reduction in rainfall at ~6 ka B.P. (Ga sse and van Campo, 1994). Lake levels in northwest India indicate increa ses in water levels at ~10ka B.P. (Gasse and van Campo, 1994). Pollen-rich sediment levels indicate greater floral produc tion (Gasse and van Campo, 1994). Increases in precipitation enab le floral growth and increases in lake levels (Gasse and van Campo, 1994). The highest lake levels were recorded from ~7.2 to 6.2 ka B.P., and complete desiccation of La ke Lunkaransar occurred after ~3.8ka B.P. (Gasse and van Campo, 1994). Abundance of shrub and tree pollen from 9.3 to 6.2 ka B.P. indicates greater precipitation (Gasse and van Campo, 1994). An onset of aridity occurred at 5.2 ka B.P. and continued til l 3.8 ka B.P. (Ramesh, 2001; Gasse and van Campo, 1994). Based on the pollen record fr om Lake Lunkaransar a 30% increase in precipitation was interpreted at 9 ka B.P. co mpared to modern precipitation. Prior to 10.5 ka B.P. and after 3.8 ka B.P. a decrease of 40% versus modern precipitation occurred (Gasse and van Campo, 1994) Fluvial features in the Thar Desert reveal increased gravel bedload during the wet phase from 12-8 ka B.P. (Jain and Tardon, 2003) Sheet flow aggradation is interpreted from 9 to 5 ka B.P. with arid aeolian dunes after 3 ka B.P in the Luni Basin. (Jain and Tandon, 2003) Rivers in the upland of Maharashtra show a deposition of gravel at 14 to 13 ka B.P. (Jain and Tandon, 2003). Increased erosion occurred from 10 to 4 ka B.P. with the increase precipitati on (Jain and Tandon, 2003). Gravel beds that were deposited


18 during the wet phase were deposited by fl ood outbursts in the basin (Jain and Tandon, 2003). Early to middle Holocene climate in west ern India indicates an increase in precipitation at approximately 10 to 9.5 ka B.P., with a peak in aridity at approximately 3.7 ka B.P. (Jain and Tandon, 2003; Ramesh, 2001; Gasse and van Campo, 1994). This correlates with the southern migr ation of the ITCZ. A slight offset from 4 to 4.5 ka B.P. to 3.7 ka B.P. is attributed to a greater in fluence of the SW Indi an Monsoon on the Indian sub continent. GEOLOGIC SETTING The Wadi Hadramawt basin is part of the Hadramawt Arch, which formed during the uplift of the Southern Ar abian Paleocene carbonate plat form, emerging in its present form at the end of the Eocene (Alsharan et al ., 2001). Uplift of the platform created the faulted and tilted uplands that channel flow of Wadis Idim and Sana to the north, instead of flowing south towards the Arabian Sea. The uplift of the south Hadramawt Arch exposed the Paleocene Rus, Jezza, and Umm er Radhuma Formations in the study area (Figure 6) (Clarke et al., 1987). These limestone units are all exposed in Wadi Sana. The Umm er Radhuma is currently exposed al ong the wadi channel through downcutting, with the Jezza Formation above and the Ru s Formation only appearing as isolated remnant knobs in the highlands.


Figure 6. Cross section of the research area. The geologic units are calcareous limestone with some dolomite. The Umm er Rhadhuma is a solution fissured dolomitized sparry limestone at the bottom and highly solution fissured fossiliferous limestone at the top (Clarke et al., 1987). The sparsely exposed Rus formation overlies the Jezza and consists of chalky limestone, marls, and some evaporate beds (Clarke et al., 1987). The uplift of the carbonate platform has caused deep incision, forming a north flowing drainage network. In the floors of these deep bedrock valleys formed sandy silts that filled the valley bottoms during the Late-Pleistocene-Holocene showing a period of aggradation and then incision during the late Holocene. Further discussion of the paleohydrology of Wadi Sana is in the forthcoming masters thesis by Josh Anderson. Aggradation of the wadi silts occurred during the early Holocene when paleochannels in the silts indicate a more sinuous channel representing a depositional environment (Anderson et al., 2005). Numerical age estimates from the base of the silts indicate deposition began 10,200 to 9,200 14 C yr B.P. (uncalibrated). Suspended silts in caves and erosional notches along the bedrock channel walls indicating the upper extent of 19


20 aggradation have age estimates as young as 4,800 to 4,500 14 Cyr B.P. Further discussion of the numerical age estimates can be found later in the data section. Source material for the three-to-ten meters of accumulated silt is believed to have originated from the southern parts of Wadi Sana. The silts are deposited on a gravel terrace that is identified th roughout Wadi Sana. The gravel terrace underlying the wadi silts serves as a reference horizon for st ratigraphic correlation throughout the wadi. Optically stimulated luminescence ages on sand lenses within the upper part of the gravel terrace deposits are as young as 10-11 ka B.P. Aggradation of over-lying wadi silts is believed to correspond with the onset of a moister climate and continues until approximately 4.8 ka B.P., when incision began. TUFA FORMATION Tufa formation is a geochemical process that occurs at ambient temperatures, when meteoric water becomes saturated with CaCO 3 through the underlying carbonate substrate and degasses from the reduction in pressure or during further downstream degassing with rapids or waterfalls forming tufa (Ford and Pedley 1996; Andrews, 2006). This is similar to the formation of spel eothems, but above ground (Andrews, 2006). Since tufa formation is above ground, it usually incorporates algal ma terial and encrusts higher order plants around the margins of the water body (Andrews, 2006). Vegetation contributes to the formation of tufa through encrustation of calcite around the stems and roots of wetland plants (A ndrews et al., 2000). Each type of flora produces different molds; pinnacle shaped structures around the fringes, stem


21 encrustations, or deposits of fine mud that settle in the calm pools formed by the flora (Andrews et al., 2000). A classification scheme was proposed that de lineates tufa into four types: perched spring line, paludal, fluvial ba rrage, and lacustrine (Pedley et al., 2003). These types are based on several properties, including sedi mentological, geomorphic, botanical, process based, petrology, facies association, or lithofaci es characteristics (Pedley et el, 1996). The geomorphic, hydrologic, and geochemical characteristics of tufa are useful for paleoclimate research in drylands. Studies of tufa have been carried out in Egypt, Spain, England, and throughout the world (For d and Pedley, 1996; Ga rnett et al., 2004; Smith et al., 2004; Ordonez et al., 2005; A ndrews et al., 2000; Benson et al., 1995). However, most research has examined ge ological and depositional processes. Paleoclimates of the Pleistocene have b een reconstructed from tufa deposition at the Kharga Oasis in Egypt, showing that ev aporation rates remained high and climatic conditions did not vary signifi cantly among pluvial phases (Smith et al., 2004). Holocene aged tufa should produce a more accurate record in reconstructing the climate, due to less weathering or other alterations Through dating of charcoal and geochemical and faunal analysis of ostracodes preserved in th e tufa bodies, the paleohydrology can be reconstructed. The tufa deposits in Wadis Idim and Sana southern Yemen provide an opportunity to investigate the geologic and pa leoclimatic significance of these geologic formations in a little studied region with si gnificant paleoclimatic and archaeological context.


22 HUMAN HABITATION Evidence of human habitation is prom inent throughout Wadi Sana; stone house structures, check dams, water management structures, tombs, and other artifacts are abundant across the wadi bottoms. The transformation from a hunting and gathering society to a more sedentary ag ricultural society is partiall y based on the amount of water which was more prevalent in the region during the early Holocene. Rock shelters in the wadi have been dated to 7,723 14 C years B.P. with a continuation of habitation for at least the next 1.2 ka (McCorri ston et al., 2002). Other shelters have been studied near, the Khuzma a distinct bedrock butte, located at the confluen ce of Wadi Sana and its tributary Wadi Shumluya, dated to 6,352 14 C yr B.P. The Khuzma is easily located on satellite imagery and has been used for centu ries for navigation. Ashy layers, some of which are traceable into hearths, are found throughout the stratigraphic profile, indicating possible anthropogenic-started fire s (McCorriston et al., 2002). Sedentary structures and activities cease at approximately the same time as the shift in the monsoonal wind sy stem (McCorriston et al., 2002). Human activity at the time of the monsoonal shift also appears to transition back to a hunting and gathering. This is based upon interpretations from numerous cairns and rock wall graffiti (McCorriston et al., 2002). METHODS Integration of a number of geologic techniques was used to obtain the data to understand the early to middle Holocene climate of southern Arabia. Field samples of bulk sediments and water for isotopic analysis were taken for further analysis in the


23 United States. Loss on ignition measurements were made on bulk sediments to estimate organic carbon and carbonate co ntent, and ostracodes were ex tracted for stable isotope measurements. Oxygen-isotopes were measured on water samples as a basis for interpreting the stable isotopes of the os tracodes and determining current isotopic composition of recharge. Field Sampling Samples from thirteen tufa mounds and tufa capped lacustrine deposites were taken for charcoal and sediment analysis throughout Wadis Idim and Sana. Water samples were taken from shallow aquifer wells in the villages of Dahbiah and Ghayl bin Yumain, and two samples were taken near our base camp during each of the rain events that occurred during the fiel d season. The samples were shipped to the United States where they were processed. All samples re mained sealed at ambient temperatures following sampling. Lithologic descriptions were complete d for each tufa mound and sedimentary profile. Stratigraphic section 05-GBY-02 wa s described and sampled at constant intervals of approximately 20cm. Samples, charcoal and bulk sediment, and descriptions were then used to integrate into the greater sedimentological system in Wadi Sana and will be discussed later in the paper. Multiple samples taken from Wadi Idims numerous tufa mounds and were used for radiocarbon da ting to understand the timing of formation and cessation of the tufa de posits in Wadi Idim.


24 Charcoal and Shell Charcoal from hearths that are buried in and around the tufa throughout the study area where used to constrain the ages of the tufa mounds using AMS radiocarbon ages on the charcoal. Origins of charcoal around th e tufa mounds are from anthropogenic hearths or from natural fires. Radiocarbon age es timates measured from charcoal provides a minimum age estimate for the enclosing sediment and associated tufa deposits. Charcoal that has been deposited in lakes could be re worked resulting in an unconformity of ages with younger ages at the edges of the basin (Haynes, 2001). Charcoal samples were taken where there was a sufficient amount of charcoal to sample. Samples from the 2005 season were se nt to Beta Analytical, Inc and samples from the 1999, 2000, and 2004 field seasons were sent to the University of Arizona accelerator lab for numerical age estimates and were calibrated using IntCal 98 (Table 1) (McCorriston et al., 2002). Optically Stimulated Luminescence Dating Sediment samples were taken duri ng the 2004 field season for optically stimulated luminescence dating. Precautions such as limiting exposure to light were taken to assure sample integrity. Optically stimulated luminescence provides a method of dating when carbon and U/Th met hods are not available, and is well suited to the sand lenses within the gravel terrace deposits, in Wadi Sana. The samples were analyzed by Manfred Frechen at Institut fr Geowi ssenschaftliche Gemeinschaftsaufgaben in Hannover, Germany; results are in Table 2.


25 Water Samples Four water samples were collected for isotopic analysis using cleaned plastic bottles. Two groundwater samples were taken from wells in the vi llages of Ghyal bin Yumain in Wadi Sana and Dhabiah in Wadi Idim, after being pumped from the shallow aquifer. During the field season, we had the rare occurrence of experiencing two rain storms, which produced a sayl, or flash fl ood. Two modern precipitation samples were collected which provide a co mparison between meteoric water and groundwater. All samples remained sealed until processed for 18 O measurements by the Institute of Arctic and Alpine Research stable isotope lab at the University of Colorado by Bruce Vaughn. There were no flowing springs or other surf ace water available at the time of sampling. Loss on Ignition In the case of 05-BGY02, organic-rich sediments interpreted to have been deposited in a lacustrine e nvironment were analyzed for organic and inorganic carbon by loss-on-ignition. The amount of organic and in organic carbon in the sediments reflects biological productivity and sediment preservation conditions. Many methods can be used to determine percentages of organic and inor ganic carbon in sediment s. Loss-on-ignition provides a method that is effici ent, reproducible, and is a cost effective means to measure the carbon content in the sediment. The met hod outlined by Heiri et al. (1998) was used for this study. Sediment samples were weighed to betw een 3-5 grams, put in ceramic crucibles, and dried at 105C overnight. Samples were weighed and placed in a muffle furnace for four hours at 550C to combust organic car bon, cooled in a dessicator, and weighed.


26 Next the samples were heated at 950C for two hours to determine the amount of inorganic carbon (Table 4). LOI calculations for determining the organic carbon content are: LOI 550 =((DW 105 DW 550 )/DW 105 )*100 Eq 1 The percentage for inorganic, or carbonate ca rbon is calculated based on the following: LOI 950 =((DW 550 DW 950 )/DW 105 )*100 Eq 2 Oxygen Isotopes The paleoenvironment was reconstructed in part by measuring stable isotopes of oxygen and carbon in ostracodes shells sampled from lacustrine sediments in upper Wadi Sana. Identification of ostracodes can be used for biostratigraphic correlation and broad environmental interpretations, although that is beyond the sc ope of this investigation (Curry, 1999; Smith, 1992; Colin and Lethie rs, 1988). Stable isotopes measured on ostracodes provide a quantitative analytical interpretation of such factors as source, chemistry, and temperature of the water in which the organisms lived (Griffiths and Holmes, 2000; Griffiths et al., 1996). Stable isotopes of oxygen measure the enrichment or depletion of 18 O relative to 16 O in the carbonate shell of the ostracodes (Faure, 1991). When the formation of carbonate occurs in the ocean the 18 O value is near 0 (Faure, 1991). Non-marine water bodies are depleted in 18 O according to latitude and distance from the moisture source (Faure, 1991). The ratio of 18 O/ 16 O in the ocean is not static but dynamic based on local evaporation, freshwater input s, and global glaciations. Oxygen isotope values can be used to determine paleotemperature based on the following equation:


27 t=16.9-4.2( C W ) + 0.13( C W ) 2 Eq 3 where C is the 18 O value of calcite, W is the 18 O of water and temperature (t) is in degrees C (adopted from Faure, 1991). Equation three cannot be used to determine both temperature and 18 O value of water, because three va riables are present and only one can be measured (Faure, 1991). This equation makes the assumption that biogenic calcite was precipitated in equilibrium with th e surrounding environment. Mollusks, foraminifera, and ostracodes have been show n to precipitate in equilibrium (Faure, 1991; Schwalb and Dean, 2002). Oxygen isotope va lues cannot be used to determine a numerical temperature, but they can be used to discern general temperature trends. When paleo 18 O is combined with a modern analog for a region, a better interpretation of paleotemperature shifts in the region can be made. An enrichment of 18 O could mean an increase in evaporation, possibly resulting fr om an increase in temperature, while a decrease in 18 O could mean an influx of meteor ic water (Faure, 1991). Increased evaporation, however, could be interpreted either as an increase in temperature, a decrease in precipitation, or a combination of both processes. A dditional evidence would be needed to resolve specific factors. Comparison of the 18 O signatures of the water from the wells to the rain events will indicate if the groundwater is of Pleisto cene to early Holocene origin. Indications of old water being discharged under artesian conditions have been found in Oman to be 4 ka to 30 ka in age (Clarke et al., 1987). In 1998, springs in Wadi Idim were flowing under artesian conditions. Analysis of the groundw ater should give an approximation of the 18 O signature of the rainfall at the time of de position of the wadi silts and tufa in the region.


28 Changes in temperature may be recorded in the value of 18 O of the ostracodes in the system (Garnett, 2004; Andrews et al ., 2000; Schwalb et al., 1999; Clarke et al., 1987). Temperature increases of the surrounding air masses move 18 O towards a more positive value (Garnett, 2004). A change of one degree Celsius in air temperature would correlate into a 0.24 alteration in 18 O (Craig, 1965 in Garnett, 2004). Following procedures described by Pigati et al. (2004) and Holmes et al. (1997), the ostracodes were cleaned and readied for identification and isotopic analysis. Water fluctuations can be interpreted from the sp ecies of ostracodes present at a particular stratigraphic level (Alin and Cohen, 2003). Samples were processed in late March 2006 at the Paleolab of the College of Marine Science at USF by Ethan Goddard. Sixty three samples were analyzed in a Dual Inlet IRMS (Thermo Delta+XL with Keil III Carbonate Device) stable isotope ratio mass spectrometer. Seven samples did not yield usable data, either failing register a valu e or showing a discrepancy with the signal intensity. Four replicates from each sampled level were at tempted, but this was not always possible depending on the amount of testable material. We obtained four re plicates at 60-200 cm, 295 cm, 400 cm, and 410 cm. Samples at 240 and 395 cm had four replicates, but one replicate of each sample location did not produce a readable signal due to a lack of signal intensity. The samples that were analyzed pr ovide a complete stratig raphic profile of the tufa-capped lacustrine deposit, 05-GBY-02, isotope data are compiled in Appendix 1.


29 RESULTS Wadi Sediment Deposition Silt deposits throughout Wadi Sana and Id im are found atop a gravel terrace that has been dated to approximately 10 ka B.P. using OSL and 14 C techniques at the contact with the wadi silts. The silts are fine to co arse grain sandy silt sediments, deposited by in a low-energy fluvial environment, interprete d as slackwater deposits. The gravel terrace provides a useful reference horizon throughout the wadis. Based on OSL ages, the gravel terrace fo rmed during the Late Pleistocene, when aridity was similar to the present. Rainfall events that occurred in that hyper-arid environment would create sayls with the force needed to transport cobbles that comprise the terrace landform. These quick moving floods provided the energy to move and deposit the gravel terrace sediments. The wadi silts were deposited during the moister early to middle Holocene. Accumulation of silts buried a vast number and variety of archaeological features, indicating that a much la rger population than present was living along the wadi floodplain. This environment would also have provided a fertile landscape for agriculture.


Figure 7. Picture of the wadi silt contact with the gravel terrace. The top of the wadi silt is 2.5 m above the contact of the gravel. Increasing precipitation during the early Holocene allowed for an increase flow through the wadi, transporting and depositing the wadi silts. Deposition of the silt continued to approximately 4.5ka B.P. The youngest ages of wadi silts were determined on samples collected from remnant or hanging silts on the walls of the wadi, representing the highest level of flooding and sediment deposition (Figure 8). 30


Figure 8. Example of hanging wadi silts abutted against the wadi wall. The wadi silts are highlighted. These represent the latest phase of wadi silt deposition and are dated at ca 4.8-4.5 ka B.P.( 14 C, uncalibrated) (Photo Eric Oches) The age of the gravel terrace is estimated from optically stimulated luminescence dating of sand lenses found within the gravel terrace. Age estimates put the youngest part of the gravel terrace at 10.6 to 11.2 ka B.P. (Tables 1 and 2). Flash floods or sayls, witnessed in todays hyper arid environment, correspond to the arid environment during the Younger Dryas (Overpeck et al., 1996). The force of the flash flood would provide the force necessary to transport cobbles. While in the field during the 2005 field season we had the experience of having two rainstorms. These rainstorms demonstrated how a small quick rain storm, less than two hours, can make an impact on the wadi. Prior to the two rain events there had not 31


32 been rain in the wadi for approximately fi ve years. The day-to-day drying of the sediment surface in the wadi produced a hard semi-impervious surface reducing infiltration, leading to the flash flood. As the climate changed from the aridity of the Younger Dryas to the climate optimum of the early to middl e Holocene (Overpeck et al., 1996), there was a shift from incision to deposition in Wadi Sana. The wa di silts in Wadi Sana are located midway between of Ghyal bin Yamain to the confluen ce with Wadi Hadramawt. Source material for the silts is the large uplifted basin locat ed in the headwaters region of Wadi Sana (Anderson, thesis in prep, 2006) The wadi silts fluctuate in height abov e the gravel terrace throughout the wadi. Based on numerical age estimates from outcrops of wadi silts, inferences can be made that there was not a consistent sedimentati on rate through the wadi. At the top of the gravel terraces, age estimates are consistently at approxima tely 10ka B.P. (Table 1). There are few to no correlated sedimentation layers in the wadi silts above the gravel terrace until the present day erosional surface. Based on superposition, the age estimates are then correlated to the relative height of th e current top erosional surface and the top of the gravel terrace. Due to the lack of defi ned stratigraphic layers, the relative distance from the gravel terrace can be an indication of the age of the wadi silts. Numerical Age Estimates Numerical Age estimates were determined in samples taken from outcrops of the wadi silts throughout Wadi Idim and Wadi Sa na to understand the timing of aggradation and incision. A compilation of two dating techniques, 14 C of charcoal and shell, and


optically stimulated luminescence, from three field seasons have been combined to provide a more complete picture of the system. Numerical Age estimates from Wadi Sana are shown in a representative profile in figure 9 to illustrate the range of results obtained by 14 C and OSL and table 1 and 2 have the compiled results of age estimates. Figure 9. Representative numerical age estimates from along Wadi Sana from south to north. The dates in red are from shell, black are from charcoal, blue are OSL, and the white line represents the contact between the wadi silts and gravel terrace. The background picture is to represent the wadi silt and gravel terrace contact. The samples for numerical age estimates were taken throughout Wadi Sana and were not confined to this actual outcrop (Photo Eric Oches, 2004). 33


34 Table 2: Compilation of radiocarbon dates from the 1998, 2000, 2004, and 2005 field season. Sample Measured 13C/12C Conventional Radiocarbon Age Ratio Radiocarbon Age(*) Strat. Info 05-GBY02-1.0 9920 +/40 BP -7.0 o/oo 10220 +/40 BP Wadi Sana 1.0 m below top of tufa cap 05-WIY-01-180 5940 +/50 BP -23.0 o/oo 5970 +/50 BP Wadi Idim 1.80 m below top 05-WIY-01-915 4530 +/40 BP -25.6 o/oo 4520 +/40 BP Wadi Idim 1.80 m below top 05-WIY-03D 4380 +/70 BP 24.2 o/oo 4400 +/70 BP Wadi Idim from sample 05-WIY-03 04WS-17(1) -10.67 4,633+-40 W. Sana PSI wadi silts, 20cm below surface; youngest silt deposition 04WS-17(4) -24.46 4,721+-56 W. Sana PSI wadi silts, 90cm below surface; youngest silt deposition 04WS-7(3.6) -22.64 5,842+-43 W. Shumlya, lower exposed silt in 98-WS3 profile; 3.6m below surface 04WS-6 -23.69 5,402+-42 Wadi Sana paleochannel top edge of channel filling 04WS-8 -25.83 5,970+-72 Wadi Sana paleochannel basal infilling of channel 04WS-7(0.7) -11.2 5,329+-42 W. Shumlya, uppermost burned horizon in 98-WS3 profile; 0.7m below surface 04WS-3(a) -26.75 6,387+-61 W. Sana charcoal from base of wadi silts (onset of silt dep.) 04WS-10(4b) -23.87 9,252+-52 W. Sana charcoal from lower 1/3 of wadi silts (1.4m above base; 4m below top) 04WS-10(4c) 1.03 post-bomb (bad sample) 04WS-3(b) -6 10,254+-55 W. Sana shells from base of wadi silts (onset of silt dep.) 04-WS-18 -25.57 5,765+-45 W. Sana tributary, middle of wadi silt deposition 04-WS-4(1) -23.32 4,545+-45 W. Sana uppermost wadi silt in PSI infilling P2000-8A-0.25 -13.6 5,485+-64 charcoal in wadi silts, 0.25m below surface, N-Khuzma P2000-8A-1.95 -23.5 6,246+-58 charcoal in wadi silts, 1.95m below surface, N-Khuzma 00-WI-5 -25.8 6,859+-57 W. Idim tufa, N-end (downstream) 00-WI-9 -23.5 6,586+-56 W. Idim tufa, N-end (downstream) 00-WI-10 -26.6 5,280+-52 W. Idim tufa, S-end (upstream) 00-WI-12 -25.6 5,288+-52 W. Idim tufa, S-end (upstream) = 05-WI-05 98-Hearth #13 680+-35 Wadi Shumlya Sana confluence, hearth along rock slope probably intruded, not insitu 98-CS1-0.3-0.4m 4610+-45 PSI hanging seds, 0.3-0.4m below surface, Wadi Sana 98-CS2-0.25m 4800+-60 PSI hanging seds, 0.25m below surface, Wadi Sana 98-Hearth #10 5750+-45 Wadi Shumlya hearth, middle of wadi silt (2m above GT)


98-Hearth #2 / 98-WS1 5870+-45 Wadi Shumlya hearth #2 (bell-shaped pit hearth), 98-WS1, 3.9m below surface; 1.1m above GT 98-WS2-+45cm 5880+-55 Wadi Shumlya burned horizon, ~ 2m below surface 98-Hearth #14 6070+-40 Wadi Shumlya Sana confluence, hearth, ~ 70cm below #13, apparently insitu 98-Hearth #16 6080+-55 Wadi Shumlay hearth, no waypoint, 1.1m above GT, ~ 4m below surface. AA38385 5288+52 Geological date on S. extent of fossil springs in Wadi Idim AA38380 5485+-64 Geological date on silt section E. across wadi from Khuzma as Shumlya, 0.25 m below surface--top of silts AA38381 6246+-58 Geological date on silt section E. across wadi from Khuzma as Shumlya, 1.95 m below surface AA38546 6352+-57 Hearth 2000-044-25-1in section ca. 1m below surface E. of 4 rockshelters, Khuzma as Shumlya AA38383 6586+-56 Geological date on N. extent of fossil springs in Wadi Idim AA38382 6859+-57 Geological date on N. extent of fossil springs in Wadi Idim Table 2. Optically Stimulated Luminescence ages from the 2004 field season. Sample Lab.Ident. U [ppm] Th [ppm] K [%] H 2 O [%] Paleo-Dose rate [Gy/ka] IRSL Age [ka] 04WS-3c 658 1.910.02 4.790.05 0.710.02 3.51.5 1447 2.300.18 04WS-11 659 1.750.02 2.340.02 0.420.01 21 1226 1.640.13 04WS-13 660 1.560.02 1.170.01 0.210.01 21 1387 1.200.10 04WS-14 661 1.750.02 1.430.01 0.190.01 21 1699 1.330.12 04WS-15 662 1.890.02 2.290.02 0.380.01 21 1206 1.650.14 04WS-19 663 1.770.02 2.570.03 0.480.01 21 1447 1.760.14 04WS-22-1 664 1.910.02 3.980.04 0.730.02 3.51.5 1116 2.190.16 04WS-22-2 665 2.060.02 2.540.03 0.520.01 3.51.5 1065 1.890.15 04WS-22-3 666 2.020.02 2.110.02 0.410.01 21 1015 1.690.14 35


36 Wadi Idim Wadi Idim is located ~50 km west of Wadi Sana and has numerous tufa mounds located along the east wall of the wadi. Tufa mounds were only found near the village of Dhabiah, approximately 20 km south of the to wn of Tarim (Figures 1 and 2). Tufa mounds were not found further upstream or downstr eam of the research area. The lack of tufa mounds outside of approximately a 5 km zone around the village of Dhabiah could be the result of anthropogenic destruction, bu rial by the wadi silts, or changes in hydrologic properties of the bedrock aquifer. The tufa mounds represent spring discharge during a period of increased moisture and provide insight to the d ecrease in precipitation through tufa cessation. Radiocarbon age estimates, gathered over the 2004 and 2005 field seasons, show uniform age estimates of between 4.4 ka B.P. and ~6,650 B. P. in our research area in Wadi Idim (Figure 10). These dates repres ent the greatest lateral extent of tufa formation in Wadi Idim and provide constraints for the cessation of tufa formation by dating the outer part of the tufa mound (Figure 11) The broad rang e of dates throughout the wadi reinforces a regional change in climate with cessation of tufa formation.


Figure 10. Numerical age estimates from Wadi Idim tufa mounds. The dates in the box are stratigraphically in context to each other. 37


Figure 11. Tufa mound in Wadi Idim. Numerical age estimate from the tufa mound is 5,28050 14 Cyr B.P. AA-38384 (Photo Eric Oches, 2000). 05-GBY-02 The tufa-capped lacustrine deposit that had the best and most complete stratigraphic profile was 05-GBY-02. Located approximately one kilometer to the north of Ghayl bin Yumain, 05-GBY-02, is a lacustrine deposit that was not only easily accessible, but also was intact from the gravel terrace to the tufa cap, and is 4.6 meters in height (Figure 12, 13, and Table 3). The site was sampled at 20cm intervals from the basal contact with the gravel terrace to the tufa cap for both bulk sediment and datable organics. When there was a significant amount of organic material present, a sample was taken to provide a possible radiocarbon age estimate. Within 05-GBY-02, ostracodes were picked from each sampled horizon. These sediments are interpreted to have been deposited in a former lake or wetland system extending across this part of Wadi Sana 38


(Figure 14). Isotopic and LOI analyses from this sequence are used to interpret the paleoenvironment in upper Wadi Sana. Figure 12. Lithologic profile representing 05-GBY-02 39


Figure13. 05-GBY-02 The tufa capped lacustrine deposit in Wadi Sana. Note the bottom white line indicating the contact with the underlyinggravel terrace, the top white line indicating the contact with the caprock tufa, and the red dot indicates the location of the radiocarbon date on shell of 10,22040 14 C uncalibrated B.P. (Photo Eric Oches, 2005). Table 3. Profile of tufa capped lacustrine deposit 05-GBY-02 at 4.6m in height. Color descriptions are those of the view of the author and not compared to a standard color chart. Depth Description 0-60cm Top of exposed section:Reddish grey, tufa limestone, vesicular, fine grain, well cemented, many shells, diffuse lower boundary 60cm-1m brown grey with iron mottling, even mixture of silt and clay, angular blocky peds, bioturbation and 1-5% shells, diffuse lower boundary 1-1.2m dark grey with iron colored mottling, well sorted silt, no shells, lack of sand, bioturbation, gradual lower limit, iron stains around bioturbation 40


41 1.2-1.3m light grey, silty with shells, and carbon, abrupt lower boundary 1.3-1.7m light to dark gre y, blocky ped structure, faint laminations, clayey, no shells, very small pieces of charcoal, no bioturbation, no shells 1.7-2.2m light grey, well sorted clayey silt, no shells, no bioturbation, no charcoal, defined lower boundary l2.2-2.6m cream to light grey, random speck of carbon, 1-3% shell, silty clay diffuse lower boundary 2.6-2.95m dark grey silt with some black laminated carbon, some shell material, clayey silt abrupt lower boundary 2.95-4.10m cream to light grey with shells and carbon, the silty clay and carbon are laminated with the carbon layers becoming more pronounced at the bottom. The gradation is pronounced carbon at the bottom, shells throughout, diffuse lower boundary 4.10-4.20m silty clay with a bundance of shells, light grey, abrupt lower boundary 4.20-4.60 gravel terrace; base of the exposed section Shell fragments from 1.0m below the top of 05-GBY-02 provided a 14 C age of 10,220 40 (uncalibrated). This age would i ndicate that the tufa-capped lacustrine deposit was much older than the Holocene clim atic optimum and is inconsistent with other numerical age estimates measured from the wadi. Based on numerical age estimates from shells, charcoal, and OSL ta ken from an outcrop from the wadi silts further south in Wadi Sana at sample location, 04-WS-3, 14 C ages of shells are


42 approximately 3 ka to 3.5 ka younger th an the date indicated of 10,220. The 14 C ages of the shell at 04-WS-3 are 10,254 55. This contrasts the OSL age estimate of 7.04 ka 1.01 and the charcoal age estimate 6,387 61. Based on the position of the gravel terrace at 05-GBY-02, we can be confident th at the lacustrine silts were deposited in stratigraphic concordance with wadi silts elsewhere in Wadi Sana. This would suggest that the lacustrine silts range in age from 9-5 ka B.P. and the 14C age on shells in 05GBY-02 is approximately 3-4 ka to old. Numerical age estimates from shells are dire ctly influenced by the source material and fractionation during uptake a nd secretion of the shell com ponents (Yates et al., 2000) Shells that have a large influence of bedroc k-derived carbonate will incorporate the older carbonate, which will have an influence on the numerical age estimate of the shell (Yates et al., 2000). An offset of as much as 1,500 yrs was found by Yates and other in land snails from England (Yates et al., 2000). Paleo Lake During the early to middle Holocene, upper Wadi Sana was a marshland environment near the sample location of 05-GBY-02, with a larger flood plain further downstream. The paleomarshlands located to th e south of the throat of Wadi Sana may have also been a shallow lake during times of higher rainfall. There is also evidence that tufa mounds may have created a dam, resu lting in a shallow lake through-out the southern expanses of Wadi Sana (Figure 14). Increases in rainfall would also increase infiltration rates that would in turn allow for greater infiltration. As th e ability for the silts to store water, plant growth could


increase. This increase produced a slack water marsh area with a main stream near 05-GBY-02. In the slow moving parts of the marsh, tufa was able to form through calcification around plant stems. Figure 14. The approximate location of the paleo lake and interbedded tufa and lacustrine deposits. The view is looking to the south with the furthest marked edge of the paleo lake at approximately one kilometer upstream (Photo Eric Oches, 2005). Ostracode Identification Basic identification and interpretation of the ostracodes and their environment was performed by Brandon Curry at the Illinois State Geologic Survey. Prior to examination by Curry, the ostracodes were divided into three categories A, B, and C. This division was based on visual morphological features of the ostracodes to an untrained eye. 43


Identification shows that four genera are present in 05-GBY-02. Samples from category A at 130 and 295 cm appear to be of the genus Cyprideis spp. with two possible species present. This is an ostracode that prefers fresh to brackish water. Sample category B from 4.0m, was identified as Sarscypridopsis aculeata, but appears to be larger than other documented species. Category A at 4.10m is similar to Hemicythere villosa. Ostracodes identified from category C are an unknown species. In order to have a better understanding of the species present and the environments represented, a more detailed study would needed. Each species sampled was analyzed for 18 O. Figure 15. The left valve of Sarscypridopsis aculeata Forester et al., 2005. 44


Oxygen Isotopes Oxygen isotopes, measured in ostracodes from 05-GBY-02 range from approximately -3 to less than 5.0 (Figure 16). Depletion in the samples indicate a change in the climate of the system. Depletion or enrichment of oxygen isotopes can be caused by either an increase in evaporation, change of source location, or a change in the composition of oxygen isotopes of the oceans through glacial/interglacial periods. At least four individuals of the uniformly present species were selected from most stratigraphic levels in order to replicate data and to assess the reproducibility of the 18 O values (Figure 16). Figure 16 shows how most of the data is clumped together with only one major outlier at approximately 200 cm. 0100200300400500 -12-10-8-6-4-20 18 O Depth (cm) Figure 16. Oxygen isotope trend with a simple regression trend line to show the change in 18 O. 45


46 Carbon Isotopes Measurement of carbon isotopes provides a view of a very complicated system, especially in a region of carbona te parent rock. In Wadi Sa na, the system has inputs from the Paleocene Umm er Rhadhuma, degassing of carbonate that formed the tufa, and organic carbon production from plants. This provides for multiple sources as shown in the wide range of the values on the scatter plot. There appears to be a trend of depletion of 13 C through 05-GBY-02. This corresponds with the decrease in inorganic carbon found in the loss on ignition data (Figure 17). Based on the error bars on the 13 C measurements, the trend may not exist and the multiple influences may contribute to a large scattering of carbon isotope values for the section. The trend line ranges from just below -4 at the base of the section to approximately -8 at the top.


0100200300400500 -10-8-6-4 13 C Depth (cm) Figure 17. Carbon isotope plot with trend. Loss on Ignition Loss on ignition was measured to interpret the productivity of the lacustrine system in upper Wadi Sana. There are two pathways that carbon can enter the system, either from the older carbonate deposits of the Um er Rhaduma or from the biologic activity of plants growing at the margins of the system. With an increase in precipitation during the early to middle Holocene recharge increased as did the growth of flora and fauna (Clarke et al., 1987; Overpeck et al., 1996; Fleitmann et al., 2003a,b) LOI provides a method to interpret if the system was having a greater input from organic carbon from plant material or from the inorganic carbon from carbonate bedrock. 47


48 The LOI data shows an mirror trend for th e organic and inorganic carbon plots. As the organic carbon increases from 0.5% to just under 3%, the inorganic carbon decreases from just above 30% to approximate ly 22% (Figures 18a,b). Fluctuations in the data above and below the regression lin e show opposite trends from 410 cm to approximately 275 cm. These contrasting tr ends show an increase of organic carbon when there is a decrease in carbonate and vice versa. Changes in the type of carbon in the system could be the result of in crease/reduction of through flow or an increase/reduction of plant growth. Table 4. Loss on Ignition data. All measurements are in gram. LOI samples notation correspond with depth. Loi sample 550 percent 950 percent 60 3.95 29.56 80 1.83 19.73 100 2.81 24.70 130 1.74 19.60 160 1.95 22.38 270 0.62 35.88 295 1.16 26.25 324 0.73 32.04 340 0.58 30.98 375 0.70 27.52 395 0.86 26.68 410 0.92 25.20 410 0.52 36.15


a LOI 550C00.511.522.533.544.50100200300400500Depth (cm)Organic Carbon% b LOI 950C05101520253035400100200300400500DepthCarbonate% Figure 18. Loss on Ignition data for 05-GBY-02. The blue dots indicate sample location with the black line indicating the trend. A is the plot of LOI at 550C and B is the plot of LOI at 950C Water Samples Four water samples were collected during the 2005 field season. Two samples were from two rainfall events that occurred during the field season and the other two samples were taken from two open pit wells in the villages of Ghyal bin Yumain and 49


50 Dhabiah. The results show that the groundwater that has been pumped from the aquifer is very similar to the modern rainfall of th e region. The groundwater isotopic values are 1.48 and -1.10 and the rainfall isotopic values are -1.81 and 0.32 It should be noted that the sayl 2 lighter isotopic value of 0.32 was likel y affected by evaporation. There was a delay of several hours between rainfall and recovery of the open sample container. Table 5. Oxygen Isotope data from groundwater and rainfall during the 2005 field season ID 18 O Type Location 05-GBY-92 0.32 Sael 2 W. Sana 05-GBY-99 -1.81 Sael 1 W. Sana 05-GBY-01 -1.10 Well W. Sana 05-WIY-09 -1.58 Well W. Idim DISCUSSION Correlation of Wadi Silts, Tufa, and Lacustrine Deposits Radiocarbon dates from shells and char coal found in the tufa capped lacustrine deposits in Wadi Sana and tu fa mounds in Wadi Idim indicate that the onset and cessation of deposition are uniform throughout the wadis. This is indicated with the dates near the top of the tufa mounds that s hows cessation around 5.5ka B.P. Age estimates from the tufa mounds from their southern to nor thern extent of Wadi Idim show that tufa formation ceased during the middle Holocene. Age correlation of 05-GBY-02 with Wadi Sana show that dates from lacustrine deposits gravel terrace contact is estimated at approximately ~10ka B.P.


51 Optically stimulated luminescence and radiocarbon ages from throughout Wadi Sana indicate that si lt formation started at the begi nning of the Holocene and continued until approximately 4.5 ka B.P. Wadi silt formation coincides with the increase of rainfall in the region. The age estimates indicate that from th e top of the gravel terrace to the top of the hanging silts show a uniform ity of deposition from the early to middle Holocene. Deposition continued until a decrease in rainfa ll during the middle Holocene and incision began in the pr oceeding hyperarid climate. Oxygen Isotopes Oxygen-isotope data from ostracodes in 05-GBY-02 indicate that a weak trend toward depletion of the isotopi c signature from the base of the lacustrine deposit to the 0.60m level. The tufa capped lacustrine sedime nts represent a snap shot of the different isotopic influences of the region. Continuous depletion indicates that there is either a change in source water or an increase in precipitation. Based on other proxies we know that the climate was becoming more arid and that a change of source is the most likely cause for the depletion of 18 O. Current isotopic signatu res from around the region are listed in appendix 2 and shown in figure 19. An alternative interpretation may suggest that the interval represented by 05GBY-02 pre-dates the monsoon shift and corres ponding expected isotopic shit. It is arguable that no clear trend exists in the data and an essentially uniform isotopic signal is present throughout the sampled interval. The projected age of 05-GBY-02 section ranges from ~ 10 ka B.P. at the base to ~6 ka B.P. at the top, predating the ca. 5 ka B.P. ITCZ shift.


Figure 19. Present day oxygen isotopic values measured from GNIP (Bowen and Wilkinson 2002; Bowen and Revenaugh, 2003; IAEA/WMO, 2001). Clustering of the oxygen isotopic values indicates that there is a small variation within each stratigraphic level. This variation could be attributed to environmental variations or species variations. The general trend shows an apparent depletion of oxygen isotopes in the lacustrine system through time. Carbon Isotopes Interpretation of the carbon isotopes for sample location of 05-GBY-02 using a linear regression shows a trend from -4 to approximately -8, with an R value of 52


53 0.142. With such a low R value the distributi on of the replicates was evaluated. The values of the replicates show that the carbon isotopes do not show a distinct trend, but are scattered. The slight trend that is indicat ed by the linear regression coincides with the depletion of the oxygen isotopes. Carbon isotopes may serve as a proxy for the vegetation that is growing in the region. Such a wide scatter of isotopes would indicate that there was not one predominant type of vegetation. Based on our interpretation and reconstruction of the wadi, there would be mostly shrubs and sma ll trees, with some grasses which would be indicative of C3 plants. A strict C3 or C4 plant environment cannot be interpreted from the carbon isotopes due to a lack of a distin ct trend and the indi vidual replicates not clustered. The carbon isotopes did not provide any significant data to help with the interpretation. Loss on Ignition Loss on Ignition was performed to understa nd the productivity of the system at 05-GBY-02 in Wadi Sana. In the early Ho locene inorganic production of carbon was at its peak, comprising of approximately thirty percent of the sediment, and organic carbon was only 0.5 percent of the sediment. From the early Holocene to the middle Holocene inorganic production of carbon de creased and organic production of carbon increased. A decrease in inorganic producti on and increase in organic produ ction indicates that there is a decrease in the amount of water flowing from the aquifer as there is a change in the climate.


54 Increases in the percentage of organic production in the system indicate that plants were able to gain a foot hold as the paleo channel or lake shrank. Observations of the tufa caprock show calcification around plant remains at 0.60 m to 0.00m. Plant impressions and molds are not found elsewhere at that location. The increase of carbon coincides with the decrease in precipitati on of the region based on the correlation of depth with age. Reductions of precipitation would cause a shrinking of the paleochannel and reduction in stream velocity, allowing fo r an increase of near shore and aquatic vegetation. Water Samples Comparison of groundwater samples to modern rainfall events show that the groundwater of the Wadi is recharged at a mu ch quicker rate than previously thought. Groundwater from aquifers in Oman has been dated to as old as th e Pleistocene (Clarke et al., 1987). Rapid recharge could also be from the increased pumping in the wadi that would induce greate r throughflow. Oxygen isotopic values of the groundwat er pumped during the 2005 field season show a signal of -1.48 and -1.10. This si gnal is compared with todays isotopic signature from GNIP of -3, the ostracodes that range from ~-3 at the beginning of the Holocene and ~-5.5 at the middle Holocene, and the precipitation events from 2005 of -1.81 and 0.32 (Bowen and Wilkinson, 2002; Bowen and Revenaugh, 2003; IAEA/WMO, 2001).


55 Paleoclimate Interpretations The ITCZ migrated north during the ear ly Holocene, due to increased solar insolation, allowing for a greate r influence of the SW Indian Monsoon (Fleitmann et al., 2003a; Street-Perrot and Perrott, 1993). With more of the precipitation coming from the Arabian Sea, source the isotopic signature of the ostracodes was near -4 at the beginning of the Holocene. A more northerly ITCZ allowed greater influence from the SW Indian Monsoon to penetrate further across the Arabian Peninsula over the Hadramawt Arch. Our research site sits in the rain shadow of the Hadramawt Arch and only receives strong precipitation events fr om the Arabian Sea such as cyclones. Through the Holocene there is a continuous depletion of 18 O measured in the ostracodes. With the southern migration of the ITCZ af ter the climate optimum at ~8ka B.P., there was a change of precipitation source a ffecting the Arabian Peninsula. The ITCZ acts as a valve for the competing wind forces of the Shamal and the SW Indian monsoon. With a northern migra tion of the ITCZ, the SW Indian Monsoon can penetrate further across the Arabian Penins ula. Conversely a southerly shift to the present day position allows for a greater infl uence of the Shamal and a Mediterranean source of precipitation (We yhenmeyer et al., 2000). COMPARISON WITH REGIONAL PALEOCLIMATE STUDIES Oxygen Isotopes With a reduction of rainfall coming off the Arabian Sea, depletion in 18 O is contradictory to precipitation trends of oxygen isotopes. A model showing just a change in precipitation based on our isotopic si gnal would be contradictory to the body of


56 knowledge and understanding of the region. Base d on the wind patterns of the region, we propose a model that would provide an explan ation to the depletion our record and an enrichment of the Qunf Cave model (Figure 5). Oxygen isotopes measured in ostraco des were analyzed to understand paleorainfall and sources of precipitation in Wadi Sana. The standard curve for the region was proposed by Fleitmann et al. (2003a) from Qunf cave in Oman (Figure 5). The cave is located near the Arabian Sea with the majority of precipitation originating from mist and dense clouds that cover the region during the monsoon months of July to September. This curve represents a strong influence of the SW Indian monsoon (IOM). This is further supported by the hiatus of de position in the Qunf cav e at approximately 3 ka B.P. indicating a drop in precipita tion (Fleitman et al., 2003a) (Figure 5). Wind patterns over the Arabian Peninsula are influenced by the Shamal and the SW Indian Ocean Monsoon (Indian Ocean M onsoon). The Shamal comes off of the Mediterranean into the interior and makes a southerly turn across the Arabian Peninsula towards Ethiopia (Figure 20). The Southwes t monsoon originates south of the equator and moves across the Arabian Sea towards the Indian sub-continent


. Figure 20. Wind patterns over the Arabian Sea. (Glennie and Singhvi., 2002). Our research area is located in the precipitation shadow of the Hadramawt Arch in southern Arabia. During the Holocene influences of the southwest monsoon was probably the significant source of precipitation on our research area. Oxygen isotopes during the first part of the early Holocene from 05-GBY-02 indicate a similar isotopic signal of an Arabian Sea influence. The isotopic signal from our sample site is similar to the curve from Qunf cave during the beginning of the Holocene. Oxygen isotopic values are similar of approximately -2.00 for Qunf Cave and -3.00 from our site. At approximately 8 ka B.P. a gradual enrichment of the Qunf Cave sample begins, while our samples show a further depletion of isotopic 18 O. The Qunf Cave curve shows depletion until 2.7 ka B.P. At that point the speleothem stopped growing. The curve after the hiatus starts at near 0 and shows depletion towards approximately 57


58 1.0. Current averages in the region are comp arable with the values found in the Qunf Cave speleothem. Our record diverges from the Arabian Sea isotopic proxy of the Qunf cave speleothem at approximately 250 cm which would correspond with an age estimate of approximately 7.5 to 8 ka B.P. Isotopic composition of the ostracodes shows a further depletion of 18 O from the base to the top of 05-GBY02. This is approximately the same time that the Qunf Cave sample begins a stea dy enrichment until the hiatus at 2.7 ka B.P. A weakening of the southwest monsoon through in crease influence of the ITCZ would be indicative of this enrichment of the Qunf Cave speleothem. The migration of the ITCZ during the middle Holocene led to a weakening influence of the SW Indian Monsoon on the Ar abian Peninsula. As the influence of the SW Indian monsoon decreases the effects of the Shamal increase. The path of the winds and 18 O values from the Ethiopian coast and 18 O values from the Arabian Peninsula suggest an increase influence of the Sham al (Bowen and Wilkinson, 2002; Bowen and Revenaugh, 2003; and IAEA/WMO, 2001). A furthe r depletion of the oxygen isotopes at our sample location occurred due to a change in source of precipitation. MODERN ANALOGUE The Hadramawt governate of Yemen is a dry landscape where date palms are dying do to lack of water from a dropping water table. That is ve ry different to the Holocene climate optimum when springs, flowi ng water, and marsh grasses were present throughout the same wadi. With the increase in rainfall, the region was greener with large animals such as cattle and irrigated farming present in the region (Ha rrower, 2006;


McCoriston et al., 2002). We imagine a region of main channel stream with marshlands surrounding the channel. A modern analogue was found near the Nexen Petroleum processing area. The marshland area has sprouted from the freshwater runoff from oil production. Figure 21 is a photo that shows how the input of a little water can turn a hyper arid region into a lush area (Figure 21). Figure 21. A modern analogue for the marshlands of the early to middle Holocene in Wadi Sana. Recent Recharge Over the past six years there has been a considerable drawdown from the shallow aquifer in Wadi Idim. In the 1999 field season the team was able to find flowing streams in Wadi Idim. During the 2005 field season there were no flowing springs and the open well in the village of Dhabiah had a water level was 15 m below ground surface. Over 59


irrigation was noted by the beginning of salt encrustations that was forming on the irrigation channels of the farm. The decrease in the amount of water is also noted from the death of the numerous palms in Wadis Idim and Sana (Figure 22). Decreased in water levels were also noted in the village of Ghayl bin Yumain where there was no flowing streams and death of date palms. a 60


a Fi gu re 2 2 C o m p ari s on of t h e dec r ease i n d a t e pal m s i n Wadi I d i m A: Pi ct ure of W a di I d i m i n 19 98 B Pictu r e o f W a di I d im in 20 05 CONCL US ION Migration of the ITCZ during the early to m i ddle Holocene had an im pact on the clim ate of Southern Arabia. This impact is recorded in the w a di silts, organic sedim e nts, and depletion of oxygen isot ope in ostracodes found in sam p les from 05-GBY-02. Continuous depletion of oxygen isotopes is usuall y indicative of an increase in m o isture, but in the case of southern Arabia, there was a change in the s ource of precipitation. The wadi silts were depositing throug h the early to m i ddle Holocene. The incre a se in p r ecip ita tion enabled f o r a rive rine system to for m This system provided the energy for a flooding and deposition cycle that deposited the w a di silts. This deposition occurred until ~4.5ka B.P. when the hyper-ari d flash flood environm ent of today becam e 61


62 established. Incision is still occurring today and has removed over 10m of silt from some locations. Carbon influxes into the system are comp licated with the combination of lake productivity and inflow from the carbonate Umm er Rhudhuma. The increase of organic carbon could indicate that either there is rapid sedimentation over dead plant material or a large amount of plant material growing. Decr eases in the amount of carbonate indicate that there is a decrease of wa ter from the carbonate aquifer. A change in source is linked directly to the migration of the ITCZ. Through the early to middle Holocene the ITCZ acted as a gate to the influences of the Shamal and SW Indian Monsoon for our research region. With the increase in solar insolation during the early Holocene, the northward shift of the ITCZ allowed fo r a greater influence of the SW Indian Monsoon (Overpeck et al., 1997). With a greater influence the SW Indian Monsoon could penetrate much further into th e Arabian shield correlating the isotopic signature of the Arabian Sea. As the Shamal became the dominant source of precipitation and the amount of precipitation decreased, there was a direct effect on size of the paleo lake and marsh in Wadi Sana. Inorganic carbon entering into the system decrease d through the early to middle Holocene showing a decrease in the am ount of through flow from the carbonate aquifer. With a decrease in precipitati on and through flow the paleo lake began to disappear. A southern migration of the ITCZ allowe d for a greater influence of the Shamal further depleting the oxygen isot opes signature of the ostracodes. The modern isotopic signature of the northern Arab ian shield and Mesopotamia are much more depleted than


63 the signature from the Arabian Sea by 9 to 12 (Bowen and Wilkinson, 2002; Bowen and Revenaugh, 2003; IAEA/WMO, 2001). FUTURE RESEARCH In order to further understa nd the paleoclimatic story of southern Arabia, further research needs to be conducted in a number of areas. To understand the groundwater flow of the region, a long term study of precipitation/evaporation in the region, ostracodes taxonomic study, and a broader study of the tufa mounds is needed. These studies would further constrain the paleo-environment. A groundwater flow study in Wadi Sana and Idim would help to understand the ground water flow pattern that allowed for the formation of tufa throughout Wadi Idim and Sana. This study would not only help to understand paleo-climate regimes but also help to understand the current wate r flow patterns. This in turn could help to combat the rapidly dropping water table in the region. In conjunction with the groundwater study would be a long term study of the precipitat ion/evaporation of the region. This would allow for the understanding the amount of wa ter available for recharge and evaporation. A study of the ostracodes of the region would not only help to understand the paleo environment from the data that is he ld within the shells, but also the further understanding of the type of animals present. The data that is he ld within the shells would further boost the body of knowledge in the region. A taxonomic study would be the first in Southern Arabia. The near est study was published on ostracodes from the interior of Saudi Arabia.


64 A larger in depth study of multiple tufa mounds would further constrain the data. This would also allow for a better comparis on with other tufa s ites throughout the world and support or providing another conclusi on for the change of source conclusion.


65 REFERENCES Abell, Paul, Philipp Hoelzmann. 2000. Holocene palaeoclimates in northwestern Sudan: stable isotope studies of mollusks. Gl obal and Planetary Change vol. 26. p.1-12. AlSharhan, A.S., Z.A. Rizk, A.E.M. Nairn, D.W. Bakhit, S.A. Alhajari, 2001. Hydrogeology of an Arid Region: The Arabian Gulf and Adjoining Areas Elsevier, Amsterdam. Alin, Simon and Andrew Cohen. 2003. Lake-l evel history of La ke Tanganyika, East Africa, for the past 2500 years based on ostracode-inferred water depth reconstruction. Palaeogeography, Pal aeoclimatology, Palaeoecology. vol. 199. p. 31-49. Anderson, Scott, Oches, Eric A., Sander, Kirk M., McCorriston, Joy, and Harrower, Michael, 2005. Fluvial Sediments Record Middle Holocene Climate Change in Southern Yemen. Geological Society of America Abstracts with Programs for 2005, (Annual Meeting) p. 178. Andrews, J.E., 2006. Paleoclimatic records from stable isotopes in riverine tufas: Synthesis and review. Earth-Science Reviews, vol. 75. p. 85-104. Andrews, Julian E, Martyn Pedley, Paul F. Dennis, 2000, Palaeoenvironmental records in Holocene Spanish tufas: a stable isotope approach in search of reliable climatic archives. Sedimentology vol. 47, p. 961-978. Benson, L., Kashgarian, M., and Rubin, M., 1995, Carbonate deposition, Pyramid Lake subbasin, Nevada; 2, Lake levels and polar jet stream positions reconstructed from radiocarbon ages and elevations of carbonates (tufas) deposited in the Lahontan Basin: Palaeogeography, Pal aeoclimatology, Palaeoecology, v. 117, p. 1-30. Bowen, G.J. and B. Wilkinson, 2002. Spatial distribution of 18 O in meteoric precipitation. Geology vol 30. p. 315-318. Bowen, G.J. and J. Revenaugh. 2003. Interpolating the isotopic composition of modern meteoric precipitation. Water Resour ce Research 39(10): 1299, doi:129/2003 WR 002086 Braconnot, P., S. Joussaume, N. de Noblet G. Ramstein. 2000, Mid-Holocene and Last Glacial Maximum African monsoon changes as simulated within the Paleoclimate Modelling Intercomparison Project. Global and Planetary Change vol.26, p.5166.


66 Bray, Helen, Stephen Stokes. 2004. Temporal pa tterns of arid-humid transitions in the south-eastern Arabian Peninsula ba sed on optical dating. Geomorphology. vol. 59. p. 271-280. Clarke, I.D., P. Fritz, O.P. Quinn, P.W. Rippon, H. Nash, and Sayyid Barghash bin Ghalib al Said, 1987. Modern and Fossil Gr oundwater in an Ar id Environment: A look at the Hydrogeology of Southern Oman, Isotope Techniques in Water Resources Development, Proceedings of a Symposium, Vienna, 30 MarchApril 3, International Atomic Energy Agency in cooperation with UNESCO, p. 167187. Colin, J.P. and Lethiers, F, 1988, The impor tance of Ostracods in Biostratigraphic Analysis, in Ostracoda in the Earth Scie nces, edited by Patrick De Dekker, JeanPaul Colin, and Jean-Pierre Peypouque t. Elsevier. Amsterdam. p. 27-44. Curry, B. Brandon, 1999, An environmental tolerance index for ostracodes as indicators of physical and chemical factors in aquatic habitats. Palaeogeography, Palaeoclimatology, Palaeoecology. vol. 148. p. 51-63. Doose-Rolinski, Heidi, Ulf Rogalla, Georg Sc heeder, Andreas Lckege, and Ulrich von Rad, 2001. High-resolution temperature and evaporation changes during the late Holocene in the northeastern Arabia n Sea. Paleooceanography. vol. 16. p. 358367. Enzel, Y., L.L. Ely, S. Mishra, R. Ramesh, R. Amit, B. Lazar, S.N. Rajaguru, V.R. Baker, and A. Sandler. 1999, High-Re soluction Holocene Environmental Changes in Thar Desert, Northweste rn India, Science vol. 284, p. 125-128. Faure, Gunter, 1998, Principles and Applications of Geochemistry 2 nd Edition. Prentice Hall, United States. Fleitmann, Dominik, Stephen J. Burns, Manfred Mudelsee, Ulrich Neff, Han Kramers, Augusto Mangini, Albert Matter, 2003, Ho locene Forcing of the Indian Monsoon Recorded in a Stalagmite from Sout hern Oman, Science, vol. 300, p. 1737-1739. Fleitmann, Dominik, Stephen J. Burns, Ulrich Neff, Augusto Mangini and Albert Matter, 2003 Changing moisture sources over the last 330,000 years in Northern Oman from fluid-inclusion evidence in speleo thems. Quaternary Research. vol. 60. p. 223-232. Fleitmann, Dominik, Stephen J. Burns, Ul rich Neff, Manfred Mudelsee, Augusto Mangini, and Albert Matter, 2004, Palaeo climatic interporetation of highresolution oxygen isotope profiles derive d from annually laminated speleothems from southern Oman. Quaternary Science Reviews. vol. 23. p. 935-945.


67 Ford, T.D. and H.M. Pedley. 1996, A review of tufa and travertine deposits of the world. Earth Science Review, vol. 41, p. 117-175. Forester, R.M., Smith, A.J., Palmer, D.F., and Curry, B.B., December, 2005. North American Non-Marine Ostracode Database "NANODe" Version 1,, Kent Stat e University, Kent, Ohio, U.S.A. Frank, Barbara, 1977, Quaternary Micropale ontology of the Rub al Khali (Empty Quarter), Saudi Arabia. Masters Thesis, University of Delaware. Garnett, E. R., J.E. Andrews, R.C. Preece, and P.F. Dennis. 2004, Climatic change recorded by stable isotopes and trace el ements in a British Holocene tufa. Quaternary Science Review vol. 19, p. 251-262. Gasse, Francoise and Elise van Campo. 1994. Abr upt post-glacial climate events in West Asia and North African monsoon domains. Earth and Panetary Science Letters v. 126. p. 435-456. Glennie, K.W., 1998. The desert of SE Arabia: a product of climatic change In: Alsharhan, A.S., Glennie, K.W., Whittle, G.L., Ke ndall, C.G.St.C. (eds), Quaternary Deserts and Climatic Change. Balkema Rotterdam. p. 109-116. Glennie, K.W. and A.K. Singhvi, 2002. Ev ent stratigraphy, paleoenvironment and chronology of SE Arabian deserts. Quat ernary Science Reviews, vol. 21, p. 853869. Glennie, K.W., A.K. Singhvi, N. Lancaster, and J.T. Teller, 2002. Quaternary Climatic changes over Southern Arabia and the Thar Desert, India, in P.D. Clift and others, eds., The Tectonic and Climatic Evolu tion of the Arabian Sea Region The Geological Society, London, Speci al Publications 195, p. 301-316. Griffiths, H.I. and J.A. Holmes. 2000. Non-Marine Ostracods and Quaternary Palaeoenvironments. Quaternary Research Association Technical Guide No. 8. Griffiths, H.I., K.E. Pillidge, C.J. Hill, J.G. Evans, and M.A. Learner. 1996. Ostracod gradients in a calcareous stream: Im plications for th e palaeoecological interpretation of tufas and trav ertines. Limnologica, vol. 26. p. 49-61. Gupta, Anil k., David M. Anderson, and Jona thon T. Overpeck, 2003, Abrupt changes in the Asian southwest monsoon during the Ho locene and their links to the North Atlantic Ocean, Nature, vol. 421, p. 354-357. Gupta, Anil K, Moumita das, and Davi d M. Anderson. 2005. Solar influence on the Indian summer monsoon during the Holocen e. Geophysical Research Letters. vol 32. p. L177703.


68 Harrower, Michael, 2006, Environmental verses social parameters, landscape, and the origins of irrigation in southwest Arabia(Yemen). PhD. Dissertation. The Ohio State University. Hailemichael, Million, James Aronson, Samuel Savin, Michael Tevesz, and Joseph Carter. 2002. 18 O in mollusk shells from Pliocene Lake Hadar and modern Ethiopian lakes; implications for the history of th e Ethiopian Monsoon. Palaeogeography, Palaeoclimatol ogy, Palaeoecology, vol. 186, p. 81-99. Haynes, Jr., Vance C., 2001. Geochronology and Climate Change of PleistoceneHolocene Transition in the Darb el Arbain Desert, Eastern Sahara. Geoarchaeology: vol.16, p. 119-141. Heiri, Oliver, Andre Lotter, and Gerry Le mcke. 2001. Loss on ignition as a method for estimating organic and carbonate conten t in sediments: reproducibility and comparability of results. Journal of Paleolimnology. vol. 25. p. 1010-110. Holmes, J.A., F.A. Street-Perrott, M.J. Alle n, P.A. Fothergill, D.D. Harkness, D. Kroon, and R.A. Perrott. 1997. Holocene palaeolimnology of Kajemarum Oasis, Northern Nigeria: an is otopic study of ostracodes, bulk carbonate and organic carbon. Journal of the Geological Society, London. vol. 154. p 311-319. IAEA/WMO. 2002. Global Network for isotopes in precipitation. The GNIP database. Jain, M, and S.K. Tandon, 2003. Fluvial response to Late Quaternary climate changes, western India. Quaternary Science Reviews, vol. 22, p.2223-2235. Jorgensen, Donald Walid Yasin al-Tikiriti. 2002. A hydrologic and archeologic study of climate change in Al Ain, United Arab Emirates, Global and Planetary Change, vol. 35. p. 37-49. Kutzbach, J.E., Z. Liu. 1997. Response of the African Monsoon to Orbital Forcing and Ocean Feedbacks in the Middle Holocene. Science vol. 278, p. 440-443. Makhnach, N, V. Zernitskaja, I. Kolos ov, and G. Simakova. 2004. Stable oxygen and carbon isotopes in Late Glaci al-Holocene freshwater carbonates from Belarus and their palaeoclimatic im plications. Palaeogeography, Palaeoclimatolgy, Palaeoecology vol. 209, p.73-101. Mayewski, Paul A, Eelco E. Rohling, J. Curt Stager, Wibjorn Karlen, Kirk A. Maasch, L. David Meeker, Eric A. Meryson, Franco ise Gasse, Shirley van Kreveld, Karin Homgren, Julia Lee Thorp, Gunhild Rosqvist, Frank Rack, Michael Staubwasser, Ralph R. Schneider, and Eric J. Stieg, 2004. Holocene Climate Variability. Quaternary Research, vol.62, p.243-255.


69 McCorriston, Joy, E.A. Oches, D.E. Walter, and K.L. Cole, 2002. Holocene Paleoecology and Prehistory in Highland Southern Arabia. Paleorient. Vol. 28. p. 61-88. McDermott, Frank, 2004. Palaeo-climate reconstr uction from stable isotope variations in speleothems: a review. Quaternary Science Reviews, vol. 23, pp. 901-918. Nicoll, Kathleen. 2004. Recent environmental change and prehistoric human activity in Egypt and Northern Sudan. Quaterna ry Science Review, vol. 23, pp. 561-580. Noblet, Nathalie de, Pascale Bracaonnot, Sy lvie Joussaume, and Valer Masson, 1995. Sensitivity of simulated Asian and African summer monsoons to orbitally induced variations in insolation 126, 115, and 6 k B.P. Climate Dynamics vol. 12, p. 589603. Ordonez, S., J.A. Gonzalez Martin, M.A. Ga rcia del Cura, and H.M. Pedley. 2005. Temperate and semi-arid tufas in the Plei stocene to Recent fluvial barrage system in the Mediterranean area; The Ruidera Lakes Natural Park (Central Spain). Geomorphology, vol.69 p. 332-350. Overpeck, Jonathon, Davis Anderson, Susa n Trumbore, and Warren Prell. 1996. The southwest Indian Monsoon over the last 18,000 years. Climate Dynamics vol. 12, p. 213-225. Pedley, M., Andrews, J.E., Ordonez, S., Garc ia del Cura,Maria Angela, Gonzalez Martin, J., and Taylor, D., 1996, Does climate control the morphological fabric of freshwater carbonates? A comparative study of Holocene barrage tufas from Spain and Britain: Palaeogeography, Palaeoclimatology, Palaeoecology, vol. 121, p. 239-257. Pedley, M., Gonzalez Martin, J.A., Ordonez Delgado, S., and Garcia del Cura, A., 2003. Sedimentology of Quaternary perched spri ngline and paludal tu fas; criteria for recognition, with examples from Guadal ajara Province, Spain: Sedimentology, vol. 50, p. 23-44. Pigati, Jeffrey, Jay Quade, Timothy Shahanan, and C. Vance Haynes Jr. 2004. Radiocarbon dating of minute gastropods and new constraints on the timing of late Quaternary spring-discharge depo sits in Southern Arizona, USA. Palaeogeography, Palaeoclimatol ogy, Palaeoecology. vol. 204, p 33-45. Raides, Dirk, Frank Preusser, Albert Ma tter, and Maria Mange, 2004. Eustatic and climatic contols on the development of th e Wahiba Sand Sea, Sultanate of Oman. Sedimentology, vol. 51, p.359-1385.


70 Ramesh, R. 2001. High resolution Holocene mons oon records form different proxies; An assessment of their consistency. Current Science, vol 81, Special Section: Climate, Monsoon and Indias water. p. 1432-1436. Rodrigues, Donald, Paul Abell, Stefan Krpelin. 2000. Seasonality in the Holocene climate in Nowrthwest Sudan: interpretation of Etheria elliptica shell isotopic data. Global and Planetary Change. vol. 26, p. 181-187. Sarkar, Ashish, R. Ramesh, B.L.K. Smayajulu, R Agnihotri, A.J.T. Jull, and G.S. Burr, 2000. High resolution Holocene monsoon record from the eastern Arabian Sea. Earth and Planetary Science Letters. vol. 177, p. 209-218. Schwalb, Antje, Stephen J. Burns, Kerry Kelts, 1999, Holocene Environments from Stable Isotope Stratigraphy of Ostracods and Authigenic Carbonate in Chilean Altiplano Lakes, Palaeogeography, Palaeoclimatology, Palaeoecology, vol. 148, p. 153-168. Smith, J. R and R. Giegengack., 2004. Spatial and te mporal distribution of fossil-spring tufa deposits, Western Desert, Egypt: Implications for palaeoclimatic interpretations. In Oasis Papers III Oxbow Press. G. Bowen and C. Hope, Eds., 103-112. Smith, Jennifer R., Robert Giegengack, Henry P. Schwarcz, 2004. Constraints on Pleistocene pluvial climates through stable-isotope analysis of fossil-spring tufas and associated gastropods, Kharga Oasis, Egypt. Palaeogeography, Palaeoclimataology, Palaeoecology, vol 206, p.157-175. Smith, Jennifer R., Robert Giegengack, Henr y P. Schwarcz, Mary M. A. McDonald, Maxine R. Kleindienst, Alic ia L. Hawkins, and Charles S. Churcher. 2004. A Reconstruction of Quaternary Pluvial Environments and Human Occupations Using Stratigraphy and Geochronology of Fo ssil-Spring Tufas, Kharga Oasis, Egypt. Geoarchaeology: vol.19. p 407-439. Staubwasser, Michael, Frank Sirocko, Pieter M. Grootes, and Helmut Erlenkeuser, 2002. South Asian monsoon climate change and radiocarbon in the Arabian Sea during early and middle Holocene. Paleoceanography. vol 17. p. 1063-1075. Street-Perrott, F and R.A. Perrott. 1993, Holocene Vegetation, Lake levels, and Cimate of Africa.inGlobal Flimates since the Last Glacial Maximum. Eds H.E. Wright Jr., J.E. Kutzbach, T. WebbIII, W.R.Ruddiman, F.A. Street-Perrott, and P.J. Barleinn. University of Minnesota Press. p. 318-356 Thamban, M, V. Purnachandra Rao, R.R. Schneider. 2002. Reconstruction of late Quaternary monsoon oscillations based on clay mineral proxies using sediment cores from the western margin of India. Marine Ge ology. vol. 186 p. 527-539.


71 Yates, T.J.S., B.F. Spiro, and C. Vita-Fin zi. 2002. Stable isotope variability and the selection of terrestrial mollusk shell samples for 14 C dating. Quaternary International. vol. 87, p. 87-100. Weyhenmeyer, Constanze E., Stephen J. Burn s, H. Niklaus Waber, Werner AeschbachHertig, Rolf Kipfer, H. Hugo Loosli, Albert Matter. 2000. Cool Glacial Temperatures and Changes in Moisture Source Recorded in Oman Groundwaters. Science. vol. 287. p. 842-845. Weldeab, Syee, Ralph Schneider, Martin Klling, and Gerold Wefer. 2005. Holocene African droughts relate to eastern equatorial Atlantic cooling. Geology. vol. 33, p. 981-984. Wood, Warren W. and Jeffrey L. Imes. 1995. How wet is wet? Precipitiation constraints on late Quaternary climate in the sout hern Arabian Peninsula. Journal of Hydrology. vol.164, p. 263-268. World Bank, 2004, Rural Access Program, Yemen DRAFT SECTORAL ENVIRONMENTAL ASSESSMENT VOLUME 1 TECHNIPLAN in collaboration with SHEBA Engineering Services Rome, E1067v1 ntServer/WDSP/IB/2005/01/12/000160016 _20050112154248/Rendered/INDEX/E10670v10MNA1EA1P0852311.txt


72 Appendix 1: Oxygen isotope values from ostracodes Type A Type Depth d18O PDB A060.1 60 -5.04 A060.2 60 -5.04 A060.3 60 -5.03 A060.4 60 -4.68 A120.1 120 -3.68 A120.2 120 -4.61 A120.3 120 -8.33 A120.4 120 -5.34 A130.1 130 -2.43 A130.2 130 -3.35 A130.3 130 -7.06 A130.4 130 -4.85 A140.1 140 -8.44 A140.1 140 -6.81 A140.2 140 -4.40 A140.2 140 -5.09 A140.3 140 -3.69 A140.4 140 -4.34 A160.1 160 -3.97 A160.1 160 -4.78 A160.2 160 -5.04 A160.2 160 -4.72 A160.3 160 -3.83 A160.4 160 -5.06 A200.1 200 -10.35 A200.2 200 -4.14 A200.3 200 -2.58 A200.4 200 -4.77 A210220.1 210 -4.33 A240.1 240 -3.74 A240.2 240 -4.55 A240.3 240 A240.4 240 -4.07 A270.1 270 -2.10 A280.1 280 -5.04 A280.2 280 -4.49 A295.1 295 -4.42 A295.2 295 -3.79 A295.2 295 -2.40 A295.3 295 -3.43 A295.3 295 A295.4 295


73 A295.4 295 -2.65 A324.2 324 -2.98 A375.1 375 -1.74 A375.2 375 -4.91 A395.1 395 -5.24 A395.1 395 -4.19 A395.2 395 -2.40 A395.3 395 A400.1 400 -3.47 A400.2 400 -4.88 A400.3 400 -1.24 A400.4 400 -5.19 A410.1 410 -2.36 A410.1 410 -4.24 A410.2 410 -3.05 A410.2 410 -1.72 A410.3 410 -3.57 A410.3 410 -4.16 A410.4 410 -4.04 A410.4 410 -2.95 Type B. B060.2 60 -5.73 B060.3 60 -5.09 B120.1 120 -1.45 B120.2 120 -2.97 B120.3 120 -3.23 B375.1 375 -4.04 B375.2 375 -3.67 B375.3 375 -2.20 B375.4 375 -2.57 B395.1 395 -2.27 B400.1 400 -4.33 Type C C060 60 -4.89 C295 295 -3.84 C324 324 -3.21


74 Appendix 2. Regional isotopic values fr om around the Arabian Sea (Bowen and Wilkinson 2002; Bowen and Revenaugh, 2003; IAEA/WMO, 2001). City Elevation 18O Lat N Long E Dire Dowa, Ethiopia 1000 -4.56 9.666666 41.7833 Hadibu, Yemen 5 -2.3 12.666 53.0833 Qal' at Bishah, Saudia Arabia 800 -3.5 20.0166 -42.5 Mecca, Saudi Arabia 295 -2.03 21.45 -39.75 Medina, Saudi Arabia 600 -2.54 24.33 -39.7 Riyadh, Saudi Arabia 630 -3.04 24.5166 46.7833 Beersheba, Israel 450 -2.86 31.25 -34.8 Amman, Jordan 733 -3.55 31.95 -35.95 Damascus, Syria 800 -3.95 33.5 -36.3 Hamah, Syria 500 -3.71 35.133 36.8833 Mwingi, Kenya 800 -4.24 -0.933 38.0667 Makindu, Kenya 800 -4.14 -2.2833 37.8167 Mziha, Tanzania 690 -3.62 -5.9 37.7833 Morogoro, Tanzania 475 -3.41 -6.18666 37.6667