Proceedings of the 12th International Congress of Speleology. Volume 2: Symposium 1, 6th Conference on Limestone Hydrology and Fissured Media

Citation
Proceedings of the 12th International Congress of Speleology. Volume 2: Symposium 1, 6th Conference on Limestone Hydrology and Fissured Media

Material Information

Title:
Proceedings of the 12th International Congress of Speleology. Volume 2: Symposium 1, 6th Conference on Limestone Hydrology and Fissured Media
Alternate title:
12th International Congress of Speleology. Volume 2
Alternate title:
Symposium 1, 6th Conference on Limestone Hydrology and Fissured Media
Creator:
Jeannin, Pierre-Yves (editor)
Publisher:
International Union of Speleology
Swiss Speleological Society
Universite of Franche-Comte, Sciences & Techniques de l'Environnement
Publication Date:
Physical Description:
1 online resource

Subjects

Subjects / Keywords:
Speleology ( lcsh )
Caves ( lcsh )
Karst ( lcsh )
Genre:
Conference papers and proceedings ( TGN )

Record Information

Source Institution:
University of South Florida
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
K26-05420 ( USFLDC DOI )
k26.5420 ( USFLDC Handle )

USFLDC Membership

Aggregations:
Added automatically
Karst Information Portal

Postcard Information

Format:
Book

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

11 fl 1 .. 11111 LA ( H A U XDEf ONDS II ~ SWITZERLAND ~ bGJ ~~ : 10 t h -17 th AUGUST 1997 PROCEEDINGS Of THE 12 th INTERNATIONAL (ONGRESS Of SPELEOLOGY VOLUME 2 6 th (ONFERENCE ON LIMESTONE HYDROLOGY AND f ISSURED MEDIA 6e (OLLOQUE D'HYDROGEOLOGIE EN PAYS (AL CAIRE ET EN MILIEU f ISSURE INTERNATIONAL U NION OF S PELEOLOGY/ S wiss S PELEOLOGICAL S omTY

PAGE 2

The Centre of Hydrogeology University of Neuchatel Switzer l and The Laboratory of Structural and Applied Geology University of Franche Comte Besanc;on France The Swiss Speleological Society Proceedings of the 12th International Congress of Speleology Volume 2 Actes du se Colloque d 1 Hydrologie en Pays Calcaire et en Milieu Fissure Proceedings of the 6 th Conference on Limestone Hydrology and Fissured Media La Chaux de Fonds, Switzerland, 10 -17.08.1997

PAGE 4

Pierre-Yves Jeannin, editor Symposium 1: 6 th Conference on Limestone Hydrology and Fissured Media Scientific board: Pierre Chauve Denis Slant Cyril Delporte Jacky Mania Jacques Mudry lmre Muller Franc;ois Zwahlen Editorial board: Scientific Committee: Sabine Robert Gisele Salomon Cedric John Jacques Peguenet Dr. M Bakalowicz Montpellier France Dr M Bouzelboudjen Neuchatel Switzerland Prof. W. Balderer Zurich Switzerland. Prof B Blavoux Avignon France Prof 0 Bonacci Split Croatia Prof. L. Bouchaou Agadir Morocco. Prof A. Burger, Neuchatel Switzerland Prof P Chauve Besanc;on France Prof W. Dreybrodt Bremen Germany. Prof C Drogue Montpellier France Prof J.P Faillat Brest-Plouzane France Prof. D Ford Hamilton Canada. Prof H Hotzl Karlsruhe Germany. Prof L. Kiraly, Neuchatel Switzerland Prof R. Langguth Aachen Germany Prof C Leibundgut Freiburg Germany. Dr A. Mangin Moulis France Prof J Mania Besanc;on France. Prof B Mijatovic Novi Sad Yugoslavia Prof I. Muller Neuchatel Switzerland. Prof J Mudry Besanc;on France Prof AN Palmer, Oneonta, USA. Prof A. Pulido-Bosch Granada Spain Prof J Prester Ljubjana Slovenia Dr. J M. Puig Avignon France Prof. M. Razack Poitiers France Dr. B Reichert Karlsruhe Germany. Dr. Y. Rossier, Meyzieu France Prof C Rousset Marseille France Dr. M. Sauter Tubingen Germany. Prof P Smart Bristol UK Prof G Teutsch Tubingen Germany. Dr F Vuataz Neuchatel Switzerland Prof W B. White Pennsylvania USA Dr S. Worthington Hamilton Canada Prof. G M Zuppi Torino Italy Prof F Zwahlen, Neuchatel Switzerland. Proceedings of the 12 111 International Congress of Speleo logy, 1997, Switzerland Volume 2 II/

PAGE 5

IGU UGI EID) International Union of Speleology Union lnternationale de Speleologie lnternationale Union tor Spelaologie Swiss Speleological Society (SSS) Societe Suisse de Speleologie (SSS) Schweizerische Gesellschaft tor Hohlenforschung (SGH) United Nations Educational, Scientific and Cultural Organisation Organisat i on des Nations Unies pour education la science et la culture United Nations Educational Scientific and Cultural Organisation International Association of Hydrogeologists (IAH) Association lnternationale des Hydrogeologues (AIH) International Association of Hydrological Sciences (IAHS) Association lnternationale des Sciences Hydrologiques (AISH) Swiss Academy of Sciences (SAS) Academie Suisse des Sciences Naturelles (ASSN) Schweizerische Akademie tor Naturw i ssenschaften (SANW) International Geographic Union Union Geographique lnternationale International Association of Geomorphologists (I.A G.) Associat i on lnternationale des Geomorphologues (A.LG.) ISBN 2 88374-007-0 (Vol. 2) ISBN 2 88374-012 7 (Vol. 1-6) ISSN 0759 7517 (Sciences & Techniques de l environnement) Aknowledgements for sponsoring to: Publisher: Cover: Sales: University of Franche-Comte at Besangon lnter-Unec, Besangon University of Franche-Comte, Sciences & Techniques de Environnement Titanengang, Holloch, Switzerland (Photo Ballmann/Widmer) Bibliotheque de la Societe Suisse de Speleologie, c/o Bibliotheque de la Ville, Rue du Progres 33, CH-2300 La Chaux-de-Fonds Fax : 021 947 53 78, email: ssslib@vtx ch Printed in Switzerland Papers published from the camera ready copies, prepared by the authors after reviewing by the members of the scientific committee. Despite this, the scientific board wishes to make clear that it shall take no responsability for any mistakes and omissions, or for the opinions stated by the authors. SCIENCES & TECHNIQUES DE L ENVIRONNEMENT Universite de Franche-Comte memoire hors-serie n Q 12 1997 IV Proceed i ngs of the 12 '" Internat i onal Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 6

Table of Contents Symposium 1 of the 12 th International Congress of Speleology: 6 th Conference on Limestone Hydrology and Fissured Media Session 1: The effects of natural catastrophes and large scale construction on karst aquifers Impact d'evenements exceptionnels ou de grands travaux sur le comportement des aquiferes karstiques Bayle Christian Emblanch Christophe, J Mudry, R Lastennet & J M Puig Horat Peter A. Faeh, F Naef, S Scherrer & Ch. Zurbrugg Jeanbourquin Pascal Lastennet Roland, J Mudry & Ch Emblanch Shenjifang Wanjunwei Y & X C C Liyanyun Incidences d'une galerie realisee a travers un massif avec karst... .... .... .. .. .. Evolution du regime des ecoulements karstiques lies a une impulsion pluvieuse exceptionnelle Exemple de la source de Notre Dame des Anges (Vaucluse, France) .. .... .... . .................. .. ..... .. ........... .. ..... ....... .. Wie grosse Hochwasser kann die Schuss bringen? Uber die Entstehung extremer Hochwasser im verkarsteten Einzugsgebiet der Schuss .. ... ........ Eaux souterraines et evaluation des risques lies a !'exploitation des carrieres en pays karstiques .. ......... .......... .. . . .... .. ..... .. ................. .. ........ La structure du karst: apport par l'etude des ecoulements lors de phenomenes pluvieux exceptionnels ......... .. .. .. ............... ... .. .. ... .. .. ... ... The karstic collapse earthquake induced by water storage of Geheyan reservoir in the Quingjiang River, Hubei China ....... .. ......... .... .. .... .......... .. Session 2: Development of measurement techniques in caves and its meaning for the study of karst. Evolution des techniques de mesure en grotte et apport pour l'etude du fonctionnement du karst. Bakalowicz Michel Hydrogeologue versus speleologue, ou de qui releve l'etude et 5 9 13 17 21 !'exploration des eaux souterraines karstiques?.. .. .. .. .. ...... .. ....... .. .. .. ...... ... 23 Bonacci Ognjen Hauselmann Philipp & M Otz Hoblea Fabien Le Pennec Robert Martin-Rosales W., M. Lopez-Chicano, A. Pulido-Bosch & A. Vallejos Schnegg Pierre-Andre & N Doerfliger Smart Chris C & L. Zabo Meus Philippe M Bakalowicz W Kass B. Barczewsk & G Schmid Role of speleology in karst hydrology and hydrogeology... .. ........ .... .......... 27 Wasserfarbung im Gebiet Gemmenalp Sieben Hengste: Ziele und Resultate . ............ ..... . .. .... . .. ... .......... . .. ... . . ... ... .. .. .... ........ . ... ...... .. 31 Speleologie exploitation et protection des eaux souterraines dans les karsts des environs de Chambery (Savoie lsere France)... .. ................ .. .. 35 Delimitation du bassin versant karstique de la Bienne (Haut-Jura France) par la geologie et les essais de trac;age ...................... .... ........ .. ........ .... .. .. 39 Contr61e automatique de plusieurs parametres dans la Grotte des Merveilles (Aracena SW de l'Espagne) ... .. .. .. .. ....... ............. .. ......... ........ .. 43 An inexpensive flow-through field fluorometer .. .......... .... .. ....... .............. ... 4 7 Experimental design, technique and protocol in fluorimetric tracing of ground water . . .. . .. . . . .... .... .. . .. . . . . . . . .. ... . .. .... . .. .. . .. .. . .. . .. .. . .. . 51 Perspectives offertes par la mesure en continu des traceurs fluorescents dans le karst ... ... ..... ..... .... .... ..... .. .... ..... .. .... .. .... .... ........ .. ... ..... . .... .. ... 55 Proceed i ng s of the 12 '" In t erna ti onal Congress of Speleo l ogy 1997 Sw i tzer l a n d V ol ume 2 V

PAGE 7

Session 3: Rock solution kinetics and speleogenesis. Cinetique de dissolution et speleogenese Borsato Andrea Brouquisse Franc;o i s Clemens Torsten D Hi.ickinghaus M Sauter R Liedl & G. Teutsch Destombes Jean-Luc M Cordonnier J Y. Gadat & J J Delannoy Dreybrodt Wolfang & J Siemers E i senlohr Laurent B Madry & W. Dreybrodt Fa i llat Jean Pierre Grasso Domenico A. & P Y Jeannin Paj6n-Morej6n Jesus M. A. V. de la Cruz J. J. Valdes A. E Garcia C M Cruz Perrin Jerome Sanchez-Moral S., J C Cariaveras E Sanz-Rubio M Hoyos & V Soler Sauter Martin R Liedl, T Clemens & D.Hi.ichinghaus Somlette Luc J P Faillat & G Le Bee Vanara Nathalie Menne Benjamin van Beynen P E., V.A. Toth D C. Ford & H P. Schwarcz Xuewen Zhu Yuan Daoxian Dripwater monitoring at Grotta di Ernesto (NE-Italy): a contribution to the understanding of karst hydrology and the kinetics of carbonate dissolution Donnees hydrogeochimiques sur quelques karsts du sud-est asiatique .. .. Simulation of the evolution of maze caves .. ...... .. ................ .. .................. .. Periodic and aperiodic forcing of water flow through sodastraw stalactites (Choranche, Vercors France) ............ .. .... .......... . ..... .. .. .. .. .................... Early evolution of karst aquifers in limestone : Models on two-dimensional percolation clusters .. ................. ... .. .. ....... . .... ... ..... .. ... .... .. ..... .. .... ... .. .. Changes in the dissolution kinetics of limestone by intrinsic i nhibitors adsorbing to the surface ...... ............ ... ....... .... .... .. .. .................... .......... .. Abondance et origine des composes carbones inorganiques des sources karstiques du Languedoc-Roussillon .. .. ........... .. ................... .... ....... ... ... .. Approche statistique de l'impact des variations climatiques sur la reponse chimique des sources karstiques ...... .. .. .. ................................................ Geochemical and geomathematical control of the karstic waters composition of the pan de Guajaib6n and surroundings Sierra Del Rosario Pinar Del Rf o Cuba ................ ................................................. Geologie et geochimie des eaux dans le reseau du Grand Cor (Vala i s Suisse) ......... ...... .. .. .......... ........ ......... ........................................ ..... Hydrogeochemical characterization of Tito Bustillo cave (northern Spain) .. Karst aquifer genesis Modell i ng approaches and controlling parame t ers .. Influence d'intercalations carbonatees en milieu sedimentaire a dominance siliceuse sur le chimisme des eaux (Finistere France) .............................. Dissolution et speleogenese en contexte tectonique actif : Le massif des Arbailles (Pyrenees Atlantiques Fr ) ...... .... ............ . ................. . .. ...... .. .. Microbiological processes at the cave development and karstificiation .. .. ... Seasonal Fluxes of Humic Substances in Cave Drip Waters Marengo Cave Southern Indiana ...... ...... ................... ............ ... ................. ........ .. .. .. Systematic evolution of subterranean river caves ........................ ........ .. Characteristics of hydrochemical responses to environmental change in a carbonate rock aquifer ... .. . ......... .. . ... . ...... .............. ....... .. ... ............. VI P roceedings of t h e 12 m Int ern ational Con g ress of Spe leology, 19 9 7, Switz e rl a nd Volum e 2 57 61 65 69 75 81 85 91 95 99 103 107 111 115 119 120 121 123

PAGE 8

Session 4: Hydraulics and mass transport in karst aquifers Hydraulique et transport de masse dans Jes aquiferes karstiques Atteia Olivier Bundschuh Jochen Doerfliger Nathalie Guglielmi Yves C. Mangan J Mudry & A. Reynaud Hauns Michael F Hermann & 0 Atteia H0ckinghaus Dirk R. Liedl & M Sauter Jeannin Pierre-Yves & J C. Marechal Liedl Rudolf S. Renner & M Sauter Maucha Laszlo Mohrlok Ulf J Kienle & G. Teutsch Mohrlok Ulf & M Sauter Motyka Jacek & A. Zuber Pascu Mar i a R.D Gaspar Plagnes Valer i e Smart Chris C Werner Andreas H Hotz! & w. KaB Wh i te W i lliam B & E. L. White Worthington Stephen R.H & D C Ford Zechner Er i c D Genereux J D.A. Guard i ario & J E Saiers Nature and transport of natural colloids in a karstic aquifer .......................... 125 Temporal variations of spring water temperatures in relation to the extents of the heat transport modes occuring in the karstif i ed lower GypsumKeuper aquifer (Karnian southern Germany) ............................................ 129 Results of a study about tracing tests transfer functions variability i n karst environment . .. . . . .. . .. . . . .. . .. . . . .. . .. . . . . . .. . .. . . . . . ... .. . .... .. ... . . 1 33 Evolution piezometrique et qualitative d un aquifere carbonate soumis a un fort pompage de longue duree : exemple du graben de Pinchinade (Mouans-Sartoux 06) .. .... .... ..... .. .. .. ... ............. .. .. . .. ..... ....... ....... .... ..... 137 Applicat i on of a computational fluid dynamics model to cave river hydrodynamics . . . .. . . .. . .. .. . . . .. .. .. .. .. . . ... .. .. .. .. .. . .. . .. .. .. ..... .. ... 141 Characterization of karst aquifers by heat transfer .. .. .. .. ... .. .. .. .. .. .. .. .. .. .. .. .. .. 1 45 Dispers i on and tailing of tracer plumes in a karstic system (Milandre JU Switzerland) .. . .. . . .. .. .. ... . ... . .. .. ... . . .. .. .. .. . .. .. .. .. . .. . .. .. .... .. .. .. .. 149 Obtaining information on fracture geometry from heat flow data .. .. .... .. .. .. .. 1 53 Special spring discharge changes in the Aggtelek karst region in Hungary.. 157 Parameter identification in double-continuum models applied to karst aquifers . .. .. . . .. . . .. ... .. . ..... .. . .. .. .. . .. . ... .. . . . . .. . . .. ... . .. .. . .. .. ... 1 63 Modelling groundwater flow in a karst terrane using discrete and double continuum approaches importance of spatial and temporal distr i bution of recharge . . .. .. . .. . .. . . . ... ... ... .. . . .. . .. .. .. .. . .. . . .. .. . .. .. . . .. ... . . .. 167 Water levels as response funct i ons for different porous fissured karstic systems ............. .. ...... . .... .. ....... .. ... .. ....... ... ..... . .. ... ......................... .. . 171 Application of flow models to the prediction of karst behaviour and pollutant migrat i on in a groundwater system .. .......... .... .......... .... ... .. .......... .... ..... ..... 175 Le transport de matiere dans les aquiferes karstiques .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 1 79 Modelling of suckwell recharge in carbonate terrain .. .. ... .. .... .. ... .. .. .. .. .. .. .. 1 83 The i nterpretation of a High Water Tracer Test i n the Danube-Aach System (western Swabian Alb Germany) ........... .... ............................................... 187 Storm pulses thresholds and fluid mechanics in the transport of elastic sediments in limestone aquifers .. ..... ...... .. .. ......... ...... ............................. 191 Borehole tests for megascale channeling in carbonate aquifers ................. 1 95 Estimation of aquifer parameters in h i ghly permeable limestone (B i scayne Aquifer Florida). .. ......... .......................... .... ........ .................. ............. 199 Pr oceedin g s of th e 12 th I n ter na t ional Con gress o f S peleology, 1 99 7, S wi tze rl a n d Volu me 2 V II

PAGE 9

Session 5: Vulnerability and protection of karst aquifers Vulnerabilite et protection des aquiferes karstiques Burri Ezio R. Massoli-Novelli Vulnerability of karstic aquifers in Abruzzi (Italy). Two different case-studies : & M Petitta Gran Sasso mountain and Fucino plain ............. ..... ... .................. ....... ... 201 Cimino Antonio, R. Abbate, M Macaluso & S Orecchio Doerfliger Nathalie & F Zwhalen Eiswirth Matthias & H Hotzl Gaspar Radu & I. Paunica Meier Edi P Huggenberger, W. Stiefelhagen I. Muller, & R Christe Melloul Abraham J. &S Wollman Montandon Paul Etienne, S Gogniat P. Rognon & J.J Miserez Motyka Jacek & A. Postawa Muller Elisabeth, K.P. Seiler & A. Hartmann Muller lmre & P Turberg Karst hydrogeology and vulnerability in a coastal sector of Nebrodi Mts. region (Northern Sicily) ... .. . .. .. .. ... .. . . .. . . . . ....... .. .. ..... ...... .. .. ...... ... .. ... 205 EPIK, methode de cartographie de la vulnerabilite des aquiferes karst i ques pour la delimitation des zones de protection . .. ... .. . ... ... ... ... .. ..... ... ....... .. 209 Contaminant transport from leaky landfills in karst areas ....... .... ... ... ... ... .. ... 213 South Dobrogea karst vulnerability by assessment of some geophysical and geochemical parameters ................... ............................................... 217 Geophysical methods as a tool for speleological and geotechnical investigation in karst Maira, Plateau de Bure (JU), Switzerland .... . . . .. ....... 221 Contamination of carbonate aquifer and evidence of karst effects in Israel ... 225 Nitrification des eaux rejetees par la STEP de La Chaux-de-Fonds Suisselors de leur passage dans le systeme karstique de la Ronde .... ..... .......... .. 229 Variability of c1ion in water inflows to karst-fissured limestone "Zakrzowek" quarry (Cracow, Poland) ....... ... .. .. ....... .... .. . .. .. .. . ... ..... .. . .. ..... .. . .... .. .. .. 235 Microbiologic activies in open karst aquifers with matrix porosity and consequences for ground water protection in the Franconian Alb, Germany 239 "L'image" de la fracturation et de la karstification dans le Jura suisse par prospection magnetique .. . . .. . .. .. . . . .. .. . .. .. .. . . .. . . .. .. .. . .. . .... .. . .. .... 243 VIII Proceedings of the 12 th International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 10

Session 6: Karst hydrogeology Hydrogeologie du Karst Abdesselam Malek H Lami Drainage karstique de la chaine calcaire du Djurdjura (Grande Kabylie J Mania J Mudry C Aigoun Algerie) Alimentation et interaction entre composantes evaporit i que et & P Chauve carbonatee .. . .. ...... . .. ...... . .. .. ... .. .. ....... ...... .. . . . .. .. .. ......... .. ................. 24 7 Andrea B., J Mudry F Carrasco & I. Vadillo Utilisation des traceurs meteoriques (c1-, 18 0) a l etude des aqu i feres carbonates des sierras Blanca et Mijas (Cordillere Betique Sud de l Espagne) . . . . . . . . . . . . . . .. . .. . . . . . .. . . . . . . . .. . . . .. . . . . . .. . . .. 251 Benderev Aleksey Karst waters from the region of Bosnek (West Bulgaria) : Characteristics and & S Shanov condtions of format i on .. .. .. .... ...... .. .. .. .. .. .. .... .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .... ...... ... 255 Bodin Jacques & M Razack Application du concept de surface elementaire representative (S.E.R.) a l'etude comparee entre karstification et tectonique dans le departement de La Vienne (France) ..... .. .. ... .. .. .. .. .. .. .. ... .. .. ..... .. .. .. .. .. ........ .... .. .. .. .. .. .. .. .... .. 259 Collignon Bernard Etude geophysique et piezometrique d un karst cotier a Zanzibar .... .. ...... .. 263 Dublyansky Yuri V. Transition between hydrothermal and cold-water karst .. .. ..... ... .. ... ....... .. 267 Guyot Jean-Loup Regime hydrologique et flux de matieres dissoutes du karst de Sao L. Melo Filho & A. Auler Domingos Goias Bresil .......................................................................... 271 Herold Th. W Balderer & P Jordan Knez Martin Larocque Marie, A Mangin, M Razack & 0 Banton Lauriol Bernard I.D. Clarck & J Cinq-Mars Marechal Jean-Christophe M. Looser & A. Dematteis Mijatovic Borivoje F Mouret Claude & F Brouquisse Reichert Barbara H Hotzl W Stichler & P Trimborn Rousset Claude Sbai Abdelkader Smart Chris C GorbunovaK.A., N G Maximovich & S M Blinov Lauritzen Stein Erik Maucha Laszlo The influence of pre existing and orogenetic faults to the structures on modern groundwater circulation of two karstic aquifers in the southernmost anticline (Weissenstein) of the folded Jura (Switzerland) .. ...... .. ............ .. .... 275 Speleogenesis of phreatic channels in bedding-planes in the frame of karst aquifer (Skocjanske jam Caves Slovenia) .. .. ... .. .. .. .... .... .... .... .. .... .. .. .. .. .. 279 Caracterisation du karst de La Rochefoucauld (Charente France) a l'aide des analyses correlatoire et spectrale .. ...... .... .. .. .. .... .. .. .. .. .. .. .... .. .. .. .. .. .... .. 283 Hydrologie karstique en region de pergelisol: l'exemple du Yukon septentrional Canada ................................................... .. .. ......... ........... 287 Caracteristiques hydrogeologiques de deux aquiferes calcaires des Prealpes romandes .......... ....... ... .... .. .. . .. .. .......... . ... .. ..... .... .. ............... 291 Aspect hydraulique de l'ecoulement karstique et ses consequences sur le fonctionnement du systeme aquifere ... .. ..... .. .. .. .. .. .... ........ .... .. .... .. .. ...... .. 295 Les ecoulements karstiques du polje de Ban Vieng, Laos central .......... .... 299 Hydrodynamic Behavior of a Karst Aquifer (Trnovski gozd Plateau Slovenia) . .. .. .. . . . . . .. .. . .. .. . . . . . . . .. . . .. .. . . . . . . .... . . . . . .. .. . . . . 303 Structure tectonique et typologie karstique : cas des karsts de Vaucluse et de Port-Miou (region Provence Alpes cote d Azur SE de la France) .. .... 307 Approche de la structure et du fonctionnement d un bassin hydrologique en pays calcaire : exemple du bassin du Suran Jura meridional ... .......... .... 311 Hydrogeology of glacial and subglacial karst aquifers : Small River British Columbia Canada .. ..... .. . . .. ... . . . . . .. .. . . ... . . .. .. . ... . . .. .. . .. . . . . ... 31 5 Underground water level regime of the Kungur cave region .. ........... .... .. .... 31 9 Groundwater recharge in thermoglacial karst springs, South Spitsbergen ... 320 Studies on the mechanism of tidal phenomena observed i n the Aggtelek karst region... .. ... .. ... .. .. .. ..... .. .. .. ..... ... ............. .................................... 321 P roceedin g s of the 12 'h International Congress of Speleology, 1997, Switzerland Volume 2 IX

PAGE 12

Incidences d'une galerie realisee a travers un massif avec karst par Christian BAYLE Horizons 75 rue M. Berthelot Antelios C F.13858 Aix-en-Provence Cedex 3 Abstract A tunnel, 3,5 km long, in the Alps of South of France, cross folded and breaked limestone and marl. The karstic water-level, not the same than others aquifers, had repercussions on working sites : stop mining while draining karst, drying up public spring, layout and setting up borehole for drinking water, in a limited area where underground water has high mineralisation Resume Une galerie de 3,5 km de long, dans Jes Alpes du Sud, traverse des calcaires et marnes plisses et fractures. La nappe karstique des calcaires, bien distincte des autres aquueres, a eu des repercussions sur les travaux : arret du percement de la galerie pendant le drainage du massif, tarissement d'une source communale, implantation et creation de forages pour captage d'eau potable dans un domaine restrei.nt ou la mineralisation de l'eau s'est trouvee elevee 1. Situation geographique L'ouvrage conceme se trouve dans Jes Alpes du Sud de la France La galerie, a !'altitude 630 m NGF, traverse, sur 3,5 km de longueur, le massif de Javanon qui culmine a 870 m NGF 2. Geologie Les terrains font partie des chaines subalpines du domaine autochtone Deux entites morphologiques distinctes appartiennent aux formations sedimentaires du Secondaire 2.1. Calcaires du Bajocien et calcschistes du Bathonien Ils caracterisent le relief du massif de Javanon. Les calcaires ont un debit en plaquettes, separees par uu fin litagt:: marneux. lls ont ete exploites pour la construction dans tme carriere sur le bard occidental du massif 2.2. Les Terres Noires Ce sont des marnes schisteust::s du Jurassique moyen a superieur (Bathonien Sup, Oxfordien) Ellt::s se d.:veloppt::nt largement dans la region, (anticlinorium de Laragn..:) sur plus de 1000 m d'epaisseur et donnent un rt::lit::fmou 2.3. Quatemairc Differentes fom1ations du Ouaternaire sont presentes localement : placages morainiques de cailloutis et argiles depot d argiles varvees fluvio-lacustres, eboulis, a la base des versanls calcaires de Javanon depots de tufs 2.4 Geologic structurale Au point de vue structural, le massif de Javanon forme un dome anticlinal deverse et chevauchant a l'Est. Le contact, materialise par une faille et quelques affieurements de gypse, constitue un accident tectonique regional, de direction N 145E Les deformations subies par les terrains calcaires ont provoque une fracturation intense. Deux directions de failles, Jes unes assez proches de celle de !'accident regional, Jes autres perpendiculaires, delimitent des compartiments Ceux ci sont bascules Jes uns par rapport aux autres autour d'un axe de direction N 145 E 3. Hydrogeologie 3.1. Calcaires et calcschistcs Ces terrains durs se sont fractures lor s des mouvements tectoniques et par le jeu des failles Le s fissures ont pu etre agrandies par dissolution, facilitant 'i nfiltration et la circulation des eaux meteoriques L 'e n semb le est penneable en grand et forme un karst. C 'es t la seule ressource aquifere consequente du secteur. Avant 1990, son niveau s 'e tablissait aulour de la cote piezometrique 687,50 ; sa principale emergence alimentait la source S 2, captage communal 3 2. Terres Noires Les Terres Noires, pratiquement inlpenneables, ne sont aquiferes que dans les zones fissurees ou la partie superftcielle alteree Les failles, d'extension limitee, forment des compartiments relativement independanls au point de vue aquifere Ainsi deux piezometres distants de 200 m ant des niveaux decales entre eux de 12 a 22 m et des amplitudes de variations comprises entre 5 et 14 m 6 '" Conference on Limestone Hydrology and Assured Media 1

PAGE 13

Les niveaux actuels sont perches 25 a 50 m au-
PAGE 14

Vieux youions 0 100m \ 6,870,1 JAVANON \ \ \c Puzn de situation des poinJs d' eaJl Deux forages, tubes 0 168/4 mm, ont ete realises et equipes, de preference a un seul ouvrage de gros diametre, pour disposer de deux pompes (securite) 4.4. Realisation et resultats des forages Une plate-forme a ete tail!ee a la cote 702 m NGF, dans !es calcschistes a la base du versant Nord de Javanon a forte pente Les forages y ont ete implantes a 24 m de distance l un de l autre Les premieres venues d'eau sont apparues entre 80 et 85 m de profondeur Les debits importants, de 20 a 25 111 3 /h se sont produits a plus de 100 m lorsque les terrains sont passes progressivement a des calcaires en plaquettes Les debits obtenus sont trois fois plus eleves que le debit de l'ancienne source La foration a ete arretee a 150 m dans FI et 140 m dans F2 a cause d un terrain bro ye difficile Toult!fo1s, les caracterist1ques de F2 sont meilleures Pour un rabattement de I m, le debit est de 20 m 3 /h contre 5 dans FI Les tubages sont crepines snr 40 m au-dessus du fond Le niveau d 'ea u s'est stabilise auteur de 68 rn de profondeur soil la cote 634 m NGF, tandis que le radier de la galerie se trouve a 630 m NGF 4 5. Qualite de l'eau On a vu la similitude de la conductivite des eaux drainees par la gakrie au pm 2120 et de l'ancien captage communal S2 (460 cr) Par centre dans les forages, bien que positionnes a mi distance entre les deux points ci-dessus, les eaux se sont rcvclcc s plus chargees Parrni Jes composants les plus significatifs, on a : conductivite 700 cr, durete 45, S04 : 180 mg/I, Fe : 0,95 mg/I, oxygene dissous : 1,45 Cette eau fortement mineralisee, seleniteuse proche de la !unite de potabilite, et avec une teneur en fer superieure a la norme, necessite un traitement de deferrisation pour la distribution publique Cependant au point de vue bacteriologique, l'eau est exempte de germe 4.6. Hypothese explicative La mineralisation elevee de l 'eau des forages peut s'expliquer par une faible teneur en oxygene dissous La zone captee dans le karst se trouve confinee so us un recouvrement de mames et de schistes Le karst est plus ouvert dans la partie traversee par la galerie ; ii n'y eut aucun depot de fer ni trace de rouille a l'exutoire pendant toute la periode de plus de six mois de vidange et de drainage du karst. De merne l ancienne source commw1ale S2 d01mait une eau potable sans fer. En effet la circulation de l'eau a travers des eboulis, depuis le griffon jusqu'au captage, pennettait aeration et precipitation de la surcharge minerale, comme en attestent localement plusieurs depots de tu( 4.7. Protection Les forages son! cimentes sur les dix metres superieurs et crepines entre 100 et 150 rn de profondeur L'eau des forages, desti.nee a l'alirnentation hwnai.ne sera traitee pour abaisser la teneur en fer. 6 1 Conference on Umestone Hydrology and Fissured Media 3

PAGE 15

L'Hydrogeologue agree a defmi les perimetres de protection : immediat avec cloture (0 20 a 40 m), rapproche de 150 a 200 m, excluant toute construction. Le forage F20, prive et captant un aquif'ere sans relation avec le karst sc trouve hors des perimetres. L'independance entre la nappe du karst et l'eau de riviere transilant par la galerie a ete demontree dans ditTerentes configurations : la nappe du karst reste en charge par rapport a la galerie, meme en saison d'etiage ; le debit d'exploitation des forages (7 m 3 /h) a 175 m de distance, provoque un rabattement nu! ou inferieur a 4 cm au droit de la galerie, sans risque de communication de l'eau de la galerie vers le karst. 4.8. Impact de la galerie L'ensemble de la galerie est revetu de beton Une aureole speciale a ete realisee dans la traversee des gypses, sur 50 m de longueur. Il n'y a pas eu traitement 1u injection du karst, dans la zone de grossc arrivee d'eau Apres misc en service de la galerie, le niveau piezometrique s'est retabli dans le piezometre Ja 10 situe dans un compartiment de Terres Noires. Par contre, le niveau du karst reste draine et se mai.ntient a environ 4 m au-dessus du radier de la galerie (piezometre Ja 12) Lors d une visite tecluuque de la galerie mise a se;:c apres cinq ans de fonctiom1ement, aucune venue d'eau majeure n'a ete constatee Seu!, un faible filet d'eau sort au pm 2110 proche de la traversee du karst. Quelques fissures, suintements ou pelits jets en radier ont ete observes dans la traversee des Terres Noires, essentiellement a l aval (pm 665 a 685) Les debits drai.nes par la galerie, sur 3,5 km, totalisent 10 1/s Cette valeur est du meme ordre de grandeur que les debits evacues !ors du percement sans campier Jes eaux venant du karst Celles-d, a la fin de la periode de drainage, donnaient a elles seules 12 a 15 1/s Le bilan de l'exutoire de la galerie sc trouve inferieur de moitie a la valeur reelle attendue. Le niveau du karst reste rabattu sans qu'un debit manifeste apparaisse ni a l'interieur ni au.tour de la galerie D'ailleurs, !'aureole de ci.rnentation autour du gypsc bloque tout ecoulement eventuel. Cependant aucune rernontee de la nappe du karst n'a ete observee depuis la misc en service de la galerie Le percement d une galerie a travers un massif comportant des calcaires fractures a eu plusieurs impacts : impacts directs : sur la nappe karstique qui a ete rabattue et tlrainee, arret et retard des travaux de percement tarissement d tme source communale nece ss iwnt la creation d 'tm captage par forage ; impacts indirects : sur Jes petits aquiferes locaux, notamment dans Jes placages morauuques, les eboulis ou Jes couches d alteration superficielles Les ressources en eaux souterraines, en partie drainees par les fissures du massif sous-jacent ont tli.rninue References BAYLE C. 1992 A 1996 Rapports intemes Faisabilite d'un forage d'eau Reconnaissance Assistance Technique Misc en exploitation Influence eventueUe galerie/forage Qualite des eaux BORGHINO ET CAMPION! 1982 DEA Etude du massif de Javanon VASKOU PH. 1985 OGG Eludes et reconnaissances geologiques/projet de galerie 4 Proceed i ngs of the 12 th Internat i onal Congress of Speleology 1997, Switzerland Volume 2

PAGE 16

Evolution du regime des ecoulements karstiques lies a une impulsion pluvieuse exceptionnelle. Exemple de la source de Notre Dame des Anges (Vaucluse, France) par Christophe Em blanch*, Jacques Mudry**, Roland Lastennet*** & Jean-Michel Puig* *Laboratoire d'hydrogeologie, Faculte des sciences, 33 rue louis Pasteur, 84000 Avignon. **Laboratoire de Geologie Structurale et Appliquee, Faculte des Sciences, 1 place Leclerc, 25000 Besan~on. *** CDGA Bordeaux I, Batirnent Geologie, Avenue des Facultes, 33405 Talence Abstract Karst systems are generally considered as a stable systems. However they can be widely modified during exceptional rainfall. These modifications can concern the shape of the outlet which deeply transform the hyrodynamical behavior. We studied the consequence of an exceptionnal rainfall (>300 mm in 4 h). We observed immediate morphologic variation. The study of the recession coefficient shows an hydrodynamic evolution increasing stormflow, karstification possibilities and decreasing the reserve. Resume Generalement Jes systemes karstiques apparaissent comme relativement stables A l'echelle de temps humaine. Les modifications qui tendent a atteindre l equilibre soot lentes. Cependant a la faveur d'evenements pluvieux exceptionnels ii peut se produire des evolutions brutales Ces transformations rapides peuvent atteindre la forme de l'exutoire, mais aussi transformer profondement l'hydrodynamique du systeme Dans le cadre de ce travail, nous avons etudie Jes consequences d'un episode pluvieux exceptionnel en intensite (plus de 400 mm en 4 h), nous avons constate des variations morphologiques irnportantes immediates, mais aussi des evolutions rapides (quelques annees) induites par ce phenomene. Par l'intermediaire de l'etude des courbes de recessions, nous avons montre une evolution hydrodynamique du systeme qui a conduit A une augmentation des debits de crues, de nouvelles possibilites de karstification et une diminution des reserves. 1. Introduction Les precipitations a caractere exceptionnel marquent la memoire collective par leurs manifestations immediates telles que Jes inondations. Dans les aquiferes karstiques, ces evenements provoquent egalement des modifications qui, meme si elles ne sont pas toujours aussi spectaculaires, vont affecter le systeme en profondeur. Ainsi dans la partie septentrionale de la Provence, l'orage du 22/(1)/1992 (plus de 300 mm en 4 heures) reste dans toutes Jes memoires comme la catastrophe de Vaison-la-Romaine. Pour les hydrogeologues, cet orage aura aussi permis de mettre en evidence des evolutions brutales d aquiferes karstiques, evolution physique spectaculaire mais aussi hydrodynamique. Apres avoir commente les differentes modifications de l'exutoire et leur relations pos sibles avec l'evenement exceptionnel, nous ferons une etude hydrodynamique du systeme au travers des coefficients de tarissement. 2. Presentation du site d'etude Le systeme de Notre-Dame des Anges, que nous avons suivi dans le cadre de cette etude, est localise dans le S-E de la France, A la lirnite NE du Mont Ventoux (figure 1). Son fonctionnement semble se faire essentiellement par effet piston avec un role important de la zone non saturee (LASTENNET, 1994). Outre la reaction de l'exutoire principal et des differents griffons qui se soot actives !ors de l'orage du 22/(1)/1992, Lastennet a constate une evolution dans le comportement hydrodynamique de ce systeme entre Jes crues qui precedent et celles qui suivent cet orage (LASTENNET & MUDRY, 1995). Nyons 300m y lsoh~(mm) Eme:icnc:e Fig. I : Jsohyetes de l'orage du 22/09/1992 ( d'apres CJRAME modifie) localisation du bassin versant de Notre-Dame des Anges (LASTENNET et MUDRY, 1995). 6 1 h Conference on Limestone Hydrology and Rssured Med ia 5

PAGE 17

Il est a noter qu'a !'occasion de la crue qui a suivi l'orage du 22/09/1992, de profondes modifications sont apparues a l'exutoire. Au griffon principal, par lequel s'ecoulait plus de 80% de l'eau, est venu s'ajouter un second griffon qui a immediatement draine une partie consequente des ecoulements. Il faut egalement tenir compte du fait que !ors de cette crue, un exutoire considere alors comme inactif (grotte situe 5 m au dessus de l'exutoire et qui permet un acces a la zone noyee), a fonctionne pendant deux jours avec un debit de pointe estime a 15 m3/s et des eaux a forte turbidite (qui sont la preuve d'un decolmatage important). 3. Les modifications physiques Les evolutions Jes plus directement observables sont les modifications physiques a l'exutoire. Le nouveau griffon qui est apparu a quelques metres du precedent draine quatre ans apres plus de la moitie du debit. La forme de ce nouvel exutoire (cylindrique et de plusieurs centimetres) montre qu'il s'agit de la reactivation d'un ancien conduit qui avait certainement ete colmate par le Toulourenc (riviere dont le lit borde les exutoires de la source de NDA) Ce nouveau griffon perenne est situe plus bas que l'ancien. Sachant qu'avant le 22/09/1992 ii etait colmate (aucun ecoulement n'existait alors a cet endroit), on peut en deduire qu'a un moment donne le niveau de base etait remonte (]ors du colmatage de ce griffon). L'evolution des aquiferes karstiques peut done se faire de f~n brutale en rapport avec des conditions exceptionnelles, mais ces simples observations nous ont egalement permis de mettre en evidence des evolutions inverses a !'evolution globale. En effet, dans le cas de la source de Notre-Dame des Anges, on peut penser que le niveau de base s'abaisse au fur et a mesure que le Toulourenc s'enfonce dans les gorges qu'il creuse au sein des calcaires. Un autre fait remarquable de la crue du 22/09/1992 est que depuis !'apparition brutale du nouveau griffon, nous avons une evolution permanente de ]'ensemble de l'exutoire. A la faveur d'un joint de stratification ou d'une fissure colmatee, il se developpe progressivement une ligne de sources qui tend a rejoindre Jes deux griffons principaux. Outre les modifications immediates du systeme, Jes evenements exceptionnels peuvent provoquer des modifications plus lentes a l'exutoire (quelques annees). Un systeme qui apparaissait comme stable a J'echelle de temps humaine peut done se retrouver en pleine evolution a la suite d'une impulsion pluvieuse exceptionnelle. 4. les modifications hydrodynamlques Toutes ces modifications visibles a l'exutoire, et notamment l'abaissement du niveau de base peuvent avoir modifie l'hydrodynamique du systeme. L'etude de R LASTENNET (1994) faisait appara'itre des evolutions sensibles du coefficient de tarissement dans les crues suivant immediatement celle du 22/09/1992. S'agissait-il d'un phenomene momentane, d'une evolution du systeme ou d'un artefact lie au fait qu'il ne ma'itrisait plus ]'ensemble des debits? Les donnees d'un suivi realise en 1995-1996 vont nous permettre de repondre a ces questions. L'etude des coefficients de tarissement passe par une bonne connaissance des debits Avant 1992, la quasi totalite de ceux ci s'ecoulaient au !ravers d'un deversoir triangulaire, il etait done facile d'etablir une courbe de tarage. Pour l'etude de 1995-1996, le systeme est plus complexe puisqu'il comporte deux exutoires. Pour le griffon qui s'ecoule a travers le deversoir ii n'y a pas eu de probleme, pour le second par contre aucun equipement n'etait disponible. Afin d'etablir une courbe de tarage complete, nous avons profile dune forte crue pour proceder a de nombreux jaugeages simultanes sur Jes deux griffons. On a ainsi pu etablir une relation entre les deux griffons (figure 2). Connaissant Jes debits a I'exutoire equipe du seuil, nous avons ensuite deduit celui du total deux exutoires. Par securite, nous avons regulierement procede a des jaugeages de verification qui, bien que le systeme apparaisse comme en pleine evolution, n'ont montre aucune modification sensible de la relation etablie entre Jes deux griffons. Definition des coefficients et des volumes de tarissement Le coefficient de tarissement a : les courbes de tarissement peuvent etre decrites par une fonction exponentielle de la forme (MAILLET, 1905) : Q ( t) = Qo e a t avec Q(t) debit a l'exutoire au temps t, Q 0 debit de tarissement a un instant to a le coefficient de tarissement (pente de la droite issue de la relation Ln Q temps). Le volume de tarissement est defini a partir de !'integration de la formule de Maillet, on obtient : roo Qo V = C I Qo e-at dt = ------C J a to avec c facteur de correction de !'unite de temps, si Q en m3/s et t en jours, alors c = 86400. 0 0 05 0,1 0 ,1 5 0 2 debits au seull en m 3 /s Fig. 2 : Relation entre les deux exutoires de Notre-Dame des Anges 6 Proceedings of the 12 1h International Congress of Speleology 1997, Switzerland Volume 2

PAGE 18

Le terme de volume de tarissement (CAST ANY & MARGAT, 1977) est defini comme le volume d'eau deplace dans l'aquifere (reserve noyee, drains, eau de la zone non saturee), susceptible d'etre renouvele, et s'ecoulant a l'exutoire en suivant une fonction exponentielle de type Maillet (ROQUES, 1972, MANGIN, 1975, BAKALOWICZ & MANGIN, 1980, SOULIOS, 1991, BONACCI & ZIV ALlEVIC, 1993). Le volume calcule depend du debit initial Qo choisi a un instant toII est preferable de prendre ce temps to au debut du tarissement et non au moment du pie de crue, les ecoulements de la reserve pendant les crues ne suivant pas une relation exponentielle de type tarissement (BLAVOUX & MUDRY, 1983, DREISS, 1989, HARUM & al., 1992). Determination des coefficient de tarissement et des volumes dynamiques : Pour la periode 1991-1992, nous utiliserons Jes resultats de R. LASTENNET (1994). La figure 3 montre clairement !'evolution qu'il a pu observer sur Jes coefficients de tarissement entre les crues precedant et suivant la crue du 22/09/1992. 200 ] Debit (1.s"') 100 a. I 0 a"' 0 0 g 2 "-~, ,.,,_.,_,....N 10/0Qlll'".cj 221'09192 excep11onn.a. 10111Y92 cru poaterieure 10 12 Temps uours) Fig. 3 : Courbes de recession (Q=f(t)) faisant suite a quatre crues de l'automne 1992 a Notre Dame des Anges, echelle semi logarithmique (d'apres LASTENNET & al., 1995). Comme nous l'avons dit precedemment, apres le 22/09/1992, ii ne maitrisait pas tous les debits, mais la relation quasi-lineaire entre les deux griffons que nous avons obtenue en 1995 laisse supposer que meme avec des debits partiellement maitrisfs, la pente de tarissement calculee est la bonne, par contre, on ne tiendra pas compte du Qo et du volume de tarissement calcules dans ces conditions. Avec la chronique enregistree en 1995-1996, nous allons verifier si ces modifications hydrodynamiques apparues apres la crue du 22/09/1992 etaient temporaires ou durables. Pour cela nous avons choisi trois tarissements sur la chronique que nous possedons (figure 4). Notre choix s'est porte sur les tarissements Jes moins perturbes par des precipitations !ors de la decrue. Nous allons comparer des crues d'automne (1992) a des crues d'automne et de printemps (1995-1996). La crue du 28/11/1995 est la premiere crue d'automne et les parametres obtenus sont identiques a ceux du printemps 1996. Nous pouvons done comparer les deux chroniques sans tenir compte de leur difference de position par rapport au cycle hydrologique. 2 II) ;;-1,5 E C: 1 C1I 0,5 0 It) It) It) It) ::i, 0 ::i, )ij Q Q 5 0) ;;; C 0 ,a, C E = :2. :2. 0 0 C "0 .:!. 'tl E Fig. 4 : Courbes des debits de l'ensemble des exutoires de Notre-Dame des Anges pour la periode d'octobre 1995 a aoQt 1996. 5. Discussion L'ensemble des resultats est recapitule dans le tableau 1. On constate immediatement que le debit de tarissement Qo a considerablement augmente entre les deux periodes de mesure. Cette variation peut etre attribuee a la baisse du niveau de base liee a !'apparition du nouveau griffon, ou au fait que la periode 1995-1996 se trouve dans une succession d'annees humides alors qu'en 1991-1992 on avait des annees seches. crues debit debit Qo a Volume de max (1/s) a to en Vs tarissement (mJ) 22/8/92 107 7 69.4 0 0099 0.606.10 6 10/9/92 106.2 73 2 0.0108 0.586. 10 6 22/9/92 0.0713 10/10/92 0 063 28/11/95 403 121 0 04 0.261.10 6 24/1/96 900 165 0 03 0.475 10 6 3/6/96 l 100 151 0 03 0.434 10 6 Tab.I Parametres tires des courbes de tarissement des crues a Notre-Dame des Anges avant et apres la crue de Vaison la Romaine. Apres une periode de crue, le systeme pourra soutenir le debit d'etiage a !'aide de ses reserves. Le coefficient de tarissement doit done diminuer. C'est exactement !'inverse qui se produit entre les valeurs precedant l'orage du 22/09/1992 et celle de 1995-1996. On peut done bien attribuer Jes variations hydrodynamiques a l'exutoire aux variations physiques observees sur celui-ci. II ne peut pas s'agir d'un phenomene lie aux conditions hydrologiques. De plus, on observe un doublement du debit de tarissement Qo qui, Jui aussi, ne peut etre attribue qu'a une modification hydrodynamique permanente du systeme. 6 '" Conference on Umestone Hydrology and Rssured Media 7

PAGE 19

Dans la meme periode, le volume de tarissement a plutot tendance A diminuer. La baisse brutale du niveau de base provoquee par l'apparition du nouveau griffon a modifie les conditions d'ecoulement A l'exutoire L'augmentation du coefficient de tarissement et la baisse du volume de tarissement montrent que la vidange des reserves du systeme A ete facilitee par les consequences de la crue du 22/09/1992. La baisse du volume de tarissement montre bien que !'evolution du systeme provoque par cette crue tend A diminuer les reseives par un meilleur drainage. Ces modifications impliquent de nouvelles possibilites d'accroissement de la karstification. Par erosion souterraine en liaison avec !'augmentation des vitesses d'ecoulement, ou par arrivee en profondeur d'eau demeuree corrosive, la consequence est une plus grande facilite d'introduction d'eau nouvelle encore agressive, qui permettra, de dissoudre la roche profondement dans le systeme (WILLIAMS, 1983). Conclusion Au !ravers de cette etude de la modification physique et hydrodynamique d'un systeme karstique faisant suite A un episode pluvieux exceptionnel, nous avons pu mettre en evidence la relative fragilite des caracteristiques de ce type d'aquifere. Il ne s'agit pas d'une sensibilite A telle ou telle pollution, mais de celle A l'equilibre hydrodynamique. Un systeme qui paraissait stable A l'echelle de temps humaine s'est vu totalement transforme du jour au lendemain. L'apparition d'un nouveau griffon qui abaisse le niveau de base et un fort decolmatage ont radicalement modifie l'exutoire. Ensuite cette evolution physique A l'exutoire a provoque une amelioration du drainage de l'aquifere et a ainsi multiplie les debits de crues par 5 ou 10, il s'ensuit une diminution sensible des reseives. On imagine aisement les problemes qui se poseraient si une telle evolution se produisait sur un aquifere capte et surexploite en etiage. Enfin nous avons pu mettre en evidence que non seulement !'evolution des systemes karstiques peut etre brutale, mais qu'elle peut egalement etre inversee. Le nouveau griffon qui est apparu est en fait un ancien conduit qui avait ete obstrue par les depots alluviaux du Toulourenc. On peut d'ailleurs imaginer que des conditions climatiques particulieres (phase erosive sur le bassin versant marneux du Toulourenc par exemple) amenent a un nouveau comblement de ce griffon. References BAKALOWICZ M. & MANGINA. 1980: L'aquifere karstique Sa definition, ses caracteristiques et son identification Mem. hors. serie. Soc. Geol. France 11, 71-79. BLAVOUX B. & MUDRY J. 1983: Separation des composantes de l'ecoulement d'un exutoire karstique A l'aide des methodes physico-chimiques. Hydrogeol. Geol. Ing., BRGM, Orleans, 4 : 269-278. BONACCI 0. & ZIV ALJEVIC R. 1993: Hydrological explanation of the flow in karst: example of the Crnojevica spring: J. Hydro!., 146: 405-419 CASTANY G & MARGAT J 1977: Dictionnaire fran~ais d hydrologie. BRGM, 249p. CIRAME (Centre d'information Regional Agro Meteorologique et Economique de Vaucluse). 1992 : Bulletin climatique et agroclimatique de Vaucluse, 73, sept. 92, 27p. DREISS S. J. 1989: Regional scale transport in a karst aquifer. l. Component separation of springflow hydrographs. Water Ressources Research, 25(1): 7125. HARUM T., ZOJER H., LEDITZKY H.P. & al. 1992: Short term investigations by means of nature! tracers. 6th symposium on Water Tracing, Karlsruhe, 64-77. LASTENNET R 1994: Role de la zone non saturee dans le fonctionnement des aquiferes karstiques Approche physico-chimique et isotopique du signal d'entree des exutoires du massif du Ventoux (Vaucluse). These, Univ. Avignon, 205 p LASTENNET R. & MUDRY J. 1995: Impact d'un evenement pluvieux exceptionnel sur le fonctionnement d'un systeme karstique. Cas de l'orage du 22/9/92 A Vaison la Romaine (Vaucluse, France). C. R. Acad. Sci., Paris, 320 (II a) : 953-959. LASTENNET R., PUIG J.M., EMBLANCH C. & BLA VOUX B 1996: Influence de la zone non saturee sur le fonctionnement des systemes karstiques. Mise en evidence dans les sources du Nord-Vaucluse. Hydrogeologie, BRGM, 4: 57-66. MAILLET E. 1905: Essais d'hydraulique souterraine et fluviale. Herman, Paris, 218 p. MANGIN A. 1975: Contribution A l'etude hydrodynamique des aquiferes karstiques. These Sc. Univ. Dijon, 258 p. Ann. Speleol., 29: 283-332, 495-601; 30: 21-124. ROQUES H. 1972: Sur une nouvelle methode graphique d'etude des eaux naturelles. Ann Speleol., 27: 79-92. SOULIOS G. 1991: Contribution A l'etude des courbes de recession des sources karstiques : exemples du pays Hellenique. J. Hydro!. 124: 29-42. WILLIAMS P. W. 1983: The role of the subcutaneous zone in karst hydrology. In BACK W. et LAMOREAUX P E., eds V T Stringfield Symposium. Processes in Karst Hydrology. J Hydro!., 61: 45-67. 8 Proceed ings of the 12 "' International Congress of Speleology 1997, Switzerland-Volume 2

PAGE 20

Wie grosse Hochwasser kann die Schuss bringen? Ober die Entstehung extremer Hochwasser im verkarsteten Einzugs gebiet der Schuss P.Horat, A. Faeh, F.Naef, $.Scherrer, Ch. Zurbrugg Versuchsanstalt fiir Wasserbau, Hydrologie und Glaziologie (VA W), ETH Zentrum, CH-8092 Zurich Abstract Large parts of the Schuss catchment are dontinated by karst. In the wake of the 1990 and 1991 floods, during which the capacity of the Schuss channel in Biel was almost exceeded, the size of extreme floods for this catchment had to be estimated. By means of different methods (investigation of historical floods, analysis of hydrograpbs, inte gration of geological and pedological information, experiments with artificial rain, rainfall-runoff-modelling) the catchment reaction was determined. The karstic parts of the catchment react slowly to rainfall and are able to retain large amounts of water. Only prolonged rainfall events in combination with snow melt and frozen soil conditions are therefore able to generate large floods An unfavourable combination of these factors could lead to floods significant ly larger than any of those observed during the last 130 years. Zusammenfassung Grosse Teile des Einzugsgebietes der Schuss sind verkarstet. Nach zwei Hochwassern 1990 und 1991, bei denen die Kapazitat des Schusskanals in Biel nahezu erschopft war, stellte sich die Frage nacb der Grosse von Extrem hochwassern. Aus verschiedenen Teiluntersuchungen (bistoriscbe Hochwasser, Ganglinienanalyse, Umsetzung geolo giscber und pedologischer Grundlagen, Beregnungsversu che, Niederschlag-Abflussmodellierung) ergab sich ein Bild von den massgebenden Abflussprozessen bei Hochwasser. Die von Karst dorninierten Teile des Einzugsgebietes rea gieren langsarn auf Niederschlage und verrnogen grosse Wassermengen zu speichern. Daher werden grosse Hocbwas ser nur von langandauernden, ergiebigen Niederscblagsereignissen in Verbindung rnit Schneescbrnelze und Boden frost erzeugt. Bei einer ungunstigen Kombination dieser Faktoren sind wesentlich hohere Abflusse rnoglich, als in den letzten 130 Jahren beobachtet wurden. 1. Einleitung Durch die Stadt Biel am Fusse des Juras fliesst die kanalisierte Schiiss. Grosse Teile des Einzugsgebietes sind verkarstet. 1990 und 1991 traten kurz hintereinander zwei Hochwasser auf, die zu Uberschwernmungen im Einzugs gebiet fiihrten. Bei beiden Hochwassern wurde die Kapazitat des Schiisskanals in Biel voll ausgelastet und zurn Teil erschopft. Ein nur geringfiigig hoherer Abfluss hatte in Biel zu Uberschwernmungen mit grossen Schadensfolgen ge fiihrt. Um entscheiden zu konnen, ob ein Ausbau des Kanals sinnvoll ist, musste abgekHirt werden, ob die beobachte ten Hochwasser seltene Ereignisse waren, die zufallig kurz hintereinander auftraten, oder ob in Zukunft mi t Hochwas sern in der beobachteten Grossenordnung oder ntit noch grosseren Hochwassern gerechnet werden muss. schnell reagierende Flachen mittelschnell ragierende Flachen langsam reagierende Flachen Ortschaft Beregnungsversuch Quelle Einzugsgebietsgrenze Fluss N 01cm 41cm 8km Abb.1: Einzugsgebiet der Schiiss bis zum Pegel Sonceboz oberhalb von Biel. Ebenfalls eingezeichnet sind Fliichen, auf denen Beregnungsversuche durchgefiihrt wurden, und die raumliche Verteilung von Flachen mit unterschiedlicher Abflusscharakteristik, sogenannte Abflusstypen. 6 1h Conference on Limestone Hydrology and Fissured Media 9

PAGE 21

Die Abflusse der Schuss werden seit 1961 am Pegel Sonceboz oberhalb von Biel gemessen Dieser Pegel erfasst zwei Drittel des 213 km 2 grossen Einzugsgebietes der Schuss (Abb l). Die Messreihe ist jedoch zu kurz, um mit statistischen Untersuchungen allein auf die Grosse seltener Extremhochwasser und das Risiko fur die Stadt Biel zu schliessen. Die Hydrologie der Schuss musste deshalb in einem grosseren Rahmen betrachtet werden. Die massgebenden Abflussprozesse im Einzugsgebiet der Schuss bei grossen Hochwassern wurden untersucht. Im folgenden wird gezeigt, wie aus verschiedenen Teilaspekten (Historische Hochwasser, Ganglinienanaly s e, Geologie und Pedologie, Beregnungsversuche) Erkenntnisse tiber Abflussentstehung, Speicherund Retentionseigenschaften der Boden Ausdehnung der beitragenden und nichtbeitragenden Flachen und weitere hochwasserauslosende Fakto ren gewonnen wurden Mit dem hydrologischen Modell das anschliessend vorgestellt wird, konnten diese Vorstellun gen uberprtift, prazisiert und erganzt werden Das Modell erlaubte es auch, die massgebenden Abflussprozesse zu extrapolieren und Extremhochwasser abzuschatzen 2. Massgebende Prozesse bei der Entstehung extremer Hochwasser Auswertung historischer Hochwasser Erste Hinweise auf die Reaktionsweise des Einzugsgebietes lieferte eine Untersuchung der historischen Hochwas ser. Aus Archiven und Zeitungen wurden lnformationen uber Hochwasser zusammengetragen Das fri.iheste in den Zeitungen erwahnte Hochwasser fand 1863 statt. Durch Berucksichtigung der historischen Hochwasser konnte der uberblickbare Zeitraum somit auf 130 Jahre ausgedehnt werden. In dieser Zeitspanne traten insgesamt sechs Hoch wasser auf deren Spitzenabfluss in der gleichen Grossenordnung wie bei den Hochwassem von 1990 oder 1991 lagen. Alle grossen Hochwasser fanden im Winterhalbjahr statt, meist verbunden mit Schneeschmelze. Das Abflussverhalten bei beobachteten Hochwasserereignissen Die grossen bekannten Hochwasser der Schuss wurden durch mehrere Tage dauernde Niederschlage verursacht (Abb 2). Bei Gewitterregen reagiert nur ein kleiner Teil des Einzugsgebietes und verursacht einen rasanten Anstieg gefolgt von einem schnellen Ruckgang des Abflusses. Grosse Teile des Einzugsgebietes reagieren auf kurze Nieder schlage nicht. Der Spitzenabfluss ist daher bei einem solchen Ereignis unbedeutend Der Knick im abfallenden Ast der Abflussganglinie ist typisch fur ein Einzugsgebiet mit unterschiedlich reagierenden Speichern (Abb 2 ). ::::, E. -0 C: (II iii
PAGE 22

Geologie und Pedologie Die Hugel und Talflanken des Einzugsgebietes be s tehen au s s ehr durchlas s i g en Kalkformationen. Eher undurcb lassige tertiiire Ablagerungen bilden den Talboden Aufgrund geologisch e r, hydrogeologischer und bodenkundliche r Kriterien wurde das Einzugsgebiet in hydrologisch iihnlich reagierende Fliichen sogenannte Abflusstypen eingeteilt (Tab I ). Die riiumliche Verteilung dieser Flachen ist in Abb l dargestellt. Abflusstyp Unt e rgrund Standort s chnell tonige Grundmorline iiber tertiliren Sedimenten gerinnenahe Bereiche oder Mulden und Tliler die al s bevorzugte Wasserleitlinien dienen mi ttelschnell tonige Grundmorline iiber tertiliren Sedimenten gerinneferne Fllichen oder Talauen mit guten Reten mit teilweiser Gehiingescbuttiiberdeckung tionseigenschaften Iangsam stark verkarstete Malmkalke gerinneferne Fllichen mit unterirdischer Entwlisserung, Hochplateaus . .. Tab.I. Hydrologisch ahnlich reagierende Flii.chen im Einzugsgebiet der Schuss Beregnungsversuche Um das Abflussverhalten der Boden besser beurteilen zu konnen, wurden auf je einer Fliiche des sclmell und de s mittelschnell reagierenden Abflu s styps Beregnungsversuche durchgefuhrt. Die Versuchsfliiche in Scmvilier repriisen tiert den schnell reagierenden Abflusstyp, diejenige in St. Imier den mittelschnell reagierenden Beregnungsversuche auf einer verkarsteten Flache in Blauen ( Kanton BL, Koordinaten : 606 300/254'625 ) dienten als Grundlage fur die Beurteilung der Abflussreaktion der langsam reagierenden Flachen (Abb l ). Beregnet wurden Fliichen von 60 m 2 uber mehrere Stunden mit Ni e derschlagsintensitiiten zwi s chen 65 und 100 mm/h. Gemessen wurde der Oberfliichenabfluss und der Abfluss im Boden. Abb.3 zeigt den unterschiedlichen Verlauf der Abflussreaktion auf den drei Abflusstypen Die beiden Beregnungsversuche im Einzugsgebiet der Schus s bestiitigten die Zuteilung der beregneten Flachen zum schnellen bzw mittelschnellen Abflusstyp. Die Flache in Son vilier reagierte deutlich schneller und intensiver als diejenige in St. Imier. Bei beiden Flachen ist Oberfliichenabfluss der dominierende Abflussvorgang. Auf der verkarsteten Flache in Blauen wurde auch nach mehrstundiger Beregnung mit einer Gesamtniederschlagssumme von 250 mm praktisch kein Oberfliichenabfluss gemessen Solche Boden ver mogen grosse Wasserrnengen zu speichern. Die Beregnungsversuche wurden im Rahmen eines Nationalfondsprojektes durchgefiihrt ; die Methodik sowie wei tere Beregnungsversuche sind in SCHERRER (1996) und FAEH (1997) beschrieben .... Q) C
PAGE 23

3. Hydrologisches Modell Aufgabe und Aufbau des Modells Um die erarbeiteten Vorstellungen von der Hochwasserentstehung im Einzugsgebiet zu uberprufen wurde ein Schneeschmelz-Niederschlag-Abfluss Modell entwickelt Dieses berechnet die Abflusse der Schuss beim Pegel Son ceboz auf der Basis von Niederschlag auf verschiedenen Hohenstufen unter Berucksichtigung von Schneeschmelze und Evapotranspiration Speicherung und verzogerte Abgabe des abflus s wirk s arnen Niederschlags dur c h s chnell rea gierende, verzogert reagierende und stark verzogert reagierende Flachen wurden dur c h drei lineare Speicher simu liert. Jedem Speicher wurde ein maximales Speichervolumen zugeordnet. Gefiillte Speicher haben keine Retention s wirkung mehr und geben den abflusswirksamen Niederschlag ohne Verzogerung an den Vorfluter ab Berechnungen Mit dem Modell wurden acht Hochwasser nachgerechnet. Zusatzlich zu den Abflussmessungen der Schuss bei Sonceboz standen ab 1978 Abflussmessungen der zwei Quellen Source de la Raisette und Source de Ja Dou zur Ver fugung (Abb. l ). Sie widerspiegeln den Ausfluss aus dem langsamen Speicher und wurden bei der Modelleichung mitberucksichtigt. Nicht befriedigen konnte die Simulation der beiden grossten Hochwasser 1990 und 1991. Die berechneten Abflu ss e der Hochwasser waren wesentlich tiefer als die beobachteten Eine Analyse der Daten zeigte, dass diesen Ereignis s en eine Kalteperiode voranging rnit Bodentemperaturen unter 0 C. Da sich Infiltration und Wasserleitfahigkeit eines Bodens bei Frost vermindern vor allem wenn der Boden einen hohen Feuchtegehalt aufweist (KANE & STEIN 1983 ), kann angenommen werden, dass Frost in Teilen des Einzugsgebietes eine Erhohung des Abflusses bewirkte. Die weiteren Ergebnisse der Berechnungen !assen sich wie folgt zusarnmenfassen : Die schnell und die mittel schnell reagierenden Speicher werden schon von Niederschlagen kleiner Jahrlichkeit gefiillt und verlieren ihre Wir kung Die Abflussbildung wird vom langsam reagierenden Speicher rnit seinem grossen Speichervermogen gesteuert. Da sich die schnellen Speicher rasch entleeren und der langsame Speicher nie vollstandig gefiillt ist beeinflusst ein Vorregen die Grosse eines Hochwassers nur unwesentlich. Niederschlag, Schneeschmelze und Frost als hochwasserauslosende Faktoren Grundsatzlich !assen sich drei Typen von Hochwassern unterscheiden : Hochwasser die
PAGE 24

Eaux souterraines et evaluation des risques lies a l'exploitation des carrieres en pays karstiques par Pascal Jeanbourquin Ch de la Reine-Berthe 10 CH-1009 Pully Abstract One can postulate that the environmental impact of open quarries is important particularly in regards to groundwater. Therefore the Swiss Water Law has recently forbidden quarrying in groundwater protection zones of springs and water catchments (zones S Art 44 Leaux 1991) In the karst massifs, evaluate the groundwater vulnerability is a puzzling problem because of the strong heterogeneity and anisotropy of the massifs that contain the aquifers. The consequence is that the groundwater protection areas show a very large extent and they interfere a lot with the human activities Therefore the r i sk assessment of diverse anthropic activities needs a special attention in such karst areas The present work is a short term expertise (three months) of the technical dangers and the geological and hydrogeological vulnerabilities of four limestone quarries in the Jura bernois (Switzerland). Its ambition is to point out the specific hazards of the quarrying work, and possibly to compare them with other human activities such as the agriculture or the traffic of vehicles. Hence, this work is divided in five parts : 1the risk assessment, 2the description of dangers 3the geological approach of the groundwater vulnerability 4 the assessment of dangers, hazards and risks of four quarries and 5a comparison, focused on the hydrocarbures with the traffic and the agriculture. The risk assessment has been divided in six levels (1-6) or three main levels : Aworking hazard (1= danger 2= probability of occurrence), Bgeological vulnerability (3= seepage sensitivity and 4= karst vulnerability) and Cthe economic risk (5 and 6, the socio-economic potential consequences). This approach allows to compare different hazardous activities at several levels. This study brings to light two main hazards of the quarrying work : 1the use of large amounts of hydrocarbures ( mostly diesel) 2the filling and the rehabilitation of the site. The presence of HC is a potential risk for the water resources in the karst. However, the respect of elementary safety rules and the regard of usual precautions in manipulating HC allow to reduce the geohazard to a very low level. Comparing with the geohazard of vehicles traffic and agriculture, the quarrying geohazard is lower and it has the advantage to be well localized in space and time and therefore to permit easy supervision and action The experience shows that the filling and the rehabilitation of the site presents a high potential of risk for the future because the checking of the incoming materials is still weak and loose. Simultaneously, the authorized infilling materials are not well defined. Consequently it is urgent to establish quality rules for this infilling activity and to set up a control procedure that would permit to bring closer the theory and practical reality In a general way quarrying in the karst does gain a reputation for sound quality in respecting a few simple technical and environmental rules. 1. Introduction et buts Les exploitations rocheuses telles que les carrieres en pays calcaire, font peser un risque particulierement important sur I' environnement et sur Jes eaux souterraines Cette affirmation qui transparait notamrnent au travers des lois sur l'environnement et de leur application (ex en Suisse: LEAUX 1991) repose souvent sur un acquis de donnees fragmentaires non-exhaustives et une demarche peu objective. Dans une etude systematique de quatre carrieres du Jura bernois (Suisse) on a essaye de nuancer cette affirmation par une approche systematique du risque et une comparaison avec d'autres activites anthropiques telles que le trafic routier ou Jes exploitations agricoles Le present article est la synthese de cette etude qui a ete realisee pour le WEA du canton de Berne ; elle a ete presentee a l'EPFL dans le cadre d'un cycle post-grade (Jeanbourquin 1996) 2. Evaluation des risques En raison de l'ambiguite de la conception actuelle en matiere de risques naturels (ou anthropiques) et de la terminologie qui l'accompagne ("danger alea, risque "hazard and risk", Gefahr), il a ete developpe une approche du risque adaptee aux eaux souterraines du milieu karstique qui est basee sur trois niveaux, chacun d'eux etant subdivise en deux sous niveaux (fig 1) 6 th Con f erence on Li mestone Hydro l og y and Fiss u red Med i a 13

PAGE 25

Niveau 1 : .. :., .... : :... ',, -~' -~;:::: ~:=::::~~=:: .. .. :;. :::::: ;:::;_ ............. Repertoire des dangers e t de 1 eur impo~ ~~ ...... : I\ =, Niveau la Niveau 2 Niveau 2a Sensibilite de l'infiltration D;, i = l... n pour n dan ge rs P;, i = 1 ... n pour n dan ge r s Risquepur de l'activite Rp = L Rpi av ec Rp; = D ; P ; I ; / \ Risque pur de pollution des massifs rocheux Rpm =Rp F si Nivean 2b Sensibilite des eaux sonterraines (du kar.st) Facteurs de sensibilit e Fsk Risque pur de pollution des eaux soute"aines Niveau 3 R F F ('JIIIIIIIIIIIIIIIIIIBllllll8'11111111111BUIIIIIIIIIIIIIIIIIIIIBIIIBll811111111111mlllllll:18111111111B;' pe*I., s~cij Niveau 3a Vulnerabilite des installations Fvj entr e = 0 e t 1 Ee I Niveau 3b Importan c e strategique de la ressource en eau I potable valeur soci ~on o miqu e Fsocej I ave c j = 1-m. pour m s ourc es ou capra ges Figure 1: la pyramide du risque. Evaluation des risques lies a l'exploitation des carrieres en pays calcaire. Niveau 1: -risque pur de l'activite bumaine. Analyse des causes de danger, de l'importance de cbaque danger et de la probabilite d'occurrence des evenements dangereux. aL evaluation d'une carriere a la fois technique et environnementale permet de reconnaitre Jes causes possibles d'accidents ou d activites regulieres qui presentent des dangers potentiels pour I environnement. Il faut ensuite specifier importance de chacun des dangers repertories par un facteur de proportionnalite. On obtient ams1 une valeur D pour chaque danger pris individuellement d'ou Di i =l...n pour n dangers Exemple : D 1 = 1 = ecoulement sur le terrain de quelques litres d hydrocarbure (HC) (fuite de diesel <20 I. lors re manipulations) ou encore D 2 = 5 = fuite sur le terrain re quelques milliers de litres de HC suite a un acc i dent important (Ai par exemple eclatement d une citerne) bApres avoir defini Jes causes il faut evaluer le probabilite d occurrence de chacun d evenements dangereux decrits ci-dessus. Suivant les cas on peut se baser soit sur la frequence historique des accidents soit sur le rythme re travail ou encore sur la qualite de I' exploitation. Par exemple pour le danger D 1 (ci-dessus), la probabilite d'occunence est fonction de la qualite d'exploitation (manipulation des HC dans une aire protegee adequate) alors que pour le danger D 2 il faut tenir compte de la frequence historique et de la qualite d exploitation. On note P = probabilite d'occurrence de l'evenement dangereux Di avec Pi i = l.. n pour n dangers Exemple : pas de citerne P = 0 sinon 0


PAGE 26

caracteristiques des exutoires et/ou des forages existants et selon les proprietes physico-chimiques du polluant (viscosite, rensite solubilite caractere hydrophobe ... ) qui sont aussi tres determinantes a ce niveau Les valeurs proposees vont de 0 5 a 3 avec un increment de 0 5 Par exemple on a choisi F s k = 0.5 a 1 pour du diesel dans un massif sain dont les sources principales refletent un transit dans le karst noye et F s k = 2 a 3 dans le cas d'un polluant soluble dans l eau et dont le massif montre des relations directes et rapides avec les exutoires Niveau 3: facteurs influen\:ant Jes risques socio-economiques (Fsocon). aF vj = vulnerabilite des installations (1 si il y a des captages, 0 si la source n'est pas captee avec des valeurs intermediaires envisageables suivant les cas) J = 1 a m pour m sources captees ou captages et F v etant la somme de tous les F v j Exploitation Massif rocheux Carriere 1 Fsi = 1 Ai citeme 1 1 Ai. camion c 2 2 Am machine 3 3 Am transport 2 2 Incident manipul 2 2 R. materiaux 0 OU 1 Somme 10 OU 11 10 OU 11 Carriere 2 20 (Fsi=2)40 Carriere 3 26 (Fsi = 1) 26 Carriere 4 22 (F. ; = 2) 44 Agriculture F.;= 2 Ai. citeme 5 10 Ai. camion c 6 12 Am machine 5 10 Am transport 0 0 Inc manipul. 6 12 L. materiaux 0 OU 1 Somme 22 44 Trafic F.;2 Ai camion c 6 12 Am transport 6 12 Somme 12 24 b F socej = importance strategique de la ressource en eau potable valeur socio-economique a long term e (0_q socej~l ) Ce facteur socio economique est surtout dependant dJ domaine politique et de la demarche socio-economique qui l accompagne. La combinaison de ces differents facteurs autorise une grande souplesse d utilisat i on. Elle permet de definir differents types de risques a differents niveaux et pour des dangers specifiques Ainsi il est possible de comparer des risques specifiques d une activite anthropique a differents niveaux Dans la figure 1 quatre types de risque, les plus utiles on ete retenus par exemple : risque pur de l activite (utile pour comparer des activites anthropiques) risque pur de la pollution des massifs rocheux (utile pour des evaluations a long tenne de l etat de pollution d un massif) risque pur de la pollution des eaux souterraines, risque total socio-economique. Eaux souterraines Socio-econo Fsk = 2 2 4 6 4 Source de la Birse : F --"-' = 2 4 Source de la Cuchatte: F ._.._. = 0.3 20-22 Birse = 40, Cuchatte = 6 tot. = 46 (F .. k = 3) 120 I Birse = 240, Cuch = 36 tot. = 276 (F,.k = 2) 52 I (F.=6rnn = 2,3)total = 166 (F .. k = 3) 132 (F.-6,,nn = 0 3)total = 46 F ,.k = 2 20 24 20 0 24 Birse = 176, Cuch = 26, tot.= 202 F = 2 24 24 48 Birse = 96, Cuch = 14 tot.= 110 Fig. 2 Synthese des valeurs de risque pour 4 carrieres du Jura bernois, /'agriculture et le trafic (Ai = accident important, Am= accident moyen. 6 th Conference on Limestone Hydrology and Rssured Med ia 15

PAGE 27

3Resultats obtenus pour quatre carrieres du Jura bernois. Les quatre carrieres etudiees sont toutes situees en zone re protection de sources et captages du canton de Berne (Zones SIi or Sill) Pour la comparaison on a choisi la route cantonale de Pierre Pertuis (trafic) et la region agri c ole de la Montagne du Droit situee au Sud de l ax e Tramelan-Tavannes. Pour chacune de ces situations l etude a porte sur : atechniques et formes de l' e xploitation ; ceci a permis re reconnaitre les dangers dfis a } exploitation et d'etablir une liste de dangers et de leur probabilite d occurrence au niveau I ainsi que d evaluer provisoirement le facteur F S I de la sensibilite a l infiltration ; bla configuration des assises geologiques locales et les comportements hydrogeologiques locaux et regionaux. Ceci a conduit a evaluation des facteurs du niveau 2 i e le facteur F s k (sensibilite des eaux souterraines) et a la correction du facteur de sensibilite a l infiltratiOD' cla reconnaissance des captages et sour c es de ia region Etant donne que l approvisionnement en eau potable w Jura bemois est en pleine evolution les facteurs du niveau 3 n ont ete discutes que succinctement (F = F soceco n VJ F socej ) Parmi tous les dangers que font courir les carrieres sur l environnement, deux principaux ont ete ret enus: Ales accidents de carburants fossiles (essentiellement le diesel) Bl'accumulation de materiaux de remblais lors de la remise en etat. En comaparaison avec ces deux evenements, le risque engendre par les autres activites des carriei-es parait mineur notamment en regard des effets sur les eaux souterraines. A partir de ces deux dangers principaux (A et B c i-dessus) des comparaisons ont ete etablies apres avoir quantifie les facteurs de risques intervenant a tous les niveaux en les calibrant d une maniere assez fruste. Un meilleur calibrage est toujours possible mais necessite plus de temps. L'ensemble des resultats est resume dans le tableau de la figure 2. 4. Discussion Le tableau (fig. 2) permet de comparer les elements desires directement. ll s en degage le fait suivant: les carrims montrent des risques purs a exploitation du meme ordre de grandeur voire meme parfois tres inferieur au risque pur de l 'activite agricole ou du trafic routier Par contre le trafic routier entraine un risque pur considerablement plus eleve que toutes les autres activites etudiees dans ce travail si seul le trafic integral des vehicules est pri s en compte (poids lourds et vehicules legers pollution par diesel et en plus avec de l essence). Par ailleurs ce tableau montre que si l exploitant d'une carriei-e respecte scrupuleusement un formel de normes de qualite durant l exploitation, le risqu e pur d'exploitation devient tres faible bien inferieur a toute autre activite anthropique courante (carriei-e 1). Concemant les carrieres le risque pur d exploitation resulte d une activite suggeree par la loi: le remplissage et la rehabilitation du site. En effet le controle de qualite w remplissage est tres diffi c ile a gerer ; les carriei-es sont encore trop souvent synonyme de decharge Les remplissages deja effectues doivent etre consideres comme un risque majeur pour le maintien de la qualite des eaux re source en pays calcaire. Les carrieres en voie re remplissage presentent egalement un risque inferieur mais cependant tres aleatoire car il est difficile et coiiteux de controler l'apport des materiaux. Aussi il est capital d ameliorer et de respecter les formels de c ontrole des remplissages actuels en reunissant plusieurs organes re c ontrole independants: exploitants et transporteur s, administrations (cantonales et federates), bureaux prive s (techniqueou surveillance). 5. Conclusion Si les exploitations de materiaux rocheux respectent des normes drastiques de qualite, l acbarnement environnementaliste qui a c cable actuellement les exploitants de carriei-es n est absolument pas ju s tifie notamment en comparaison avec des activites bumaines courantes acceptees par tous comme les deplacements dl vebicules automobiles ou l'exploitation agricole du sol. 6. Remerciements Je remercie !'office de l economie hydraulique et energetique du canton de Berne (OEHE/WEA) et l ecole polytechnique federate a Lausanne (EPFL) qui ont partiellement finance cette etude. Que toutes les personnes qui ont contribue a elaboration de ce travail soient egalement remerciees 7. Bibliographie JEANB0URQUIN P 1996. Eaux souterraines et evaluation des risques lies a i:exploitation de carrieres en pay s karstique Memoire 3..,,,., cycle EPFL. LEAUX 1991. Loi federate s ur la protection des eaux. Suisse, ref. 814.20 PARRIAUX A. 1991. Le probleme des carrieres dans la protection globale de l environnement avec des exemples de la Suisse. Bollettino della Associa z ione Mineraria Subalpina XXVID 731 743 16 P roceedin g s of th e 12 th Inte rn ational Cong r ess o f Spel eo logy, 19 9 7, Switzerla nd Volume 2

PAGE 28

La structure du karst: apport par l'etude des ecoulements lors de phenomenes pluvieux exceptionnels R. Lastennet (1), J. Mudry (2), C. Emblancb (3). (1) CDGA Bordeaux I, Bat. de Geologie, Av. des Facultes 33405 Talence (2) Laboratoire de Geologie Structurale et Appliquee, Fae des Sciences, 1 place Leclerc, 25000 Besan~on (3) Laboratoire d Hydrogeologie Fae. des Sciences 33 rue Louis Pasteur 84000 Avignon Abstract Studies which were undertaken on the Notre-Dame des Anges karst system (Vaucluse, southeastem France) show that in highly hierarchised karst aquifers, with a well-developed epikarstic and infiltration zones, an important part of the outlet discharge is originating in the unsaturated zone This continuous contribution( hanging reserve), because of an early and efficient drainage, allows the recharge of permeable areas of the saturated zone upstream the system Hydraulic head is transferred to the karst conduits from these areas to the outlet. This involves that annex-to-drainage areas of the downstream system are poorly drained with as a consequence an increased clogging and extensive hydraulical discontinuities These subsystems are slowly emptying, only during low-water periods Their recharge is ensured either from conduits during high water pressure increases or from local infiltration when karstic organization of the susystem allows it. Primordial informations can be obtained about this functioning during exceptional rain events, which allow to activate these annex-to-drainage systems and sometimes to modify in the depth the functioning ofkarst. Resume Les etudes effectuees sur le systeme karstique de Notre-Dame des Anges (Vaucluse), montrent que dans les aquiferes karstiques fortement hierarchises ayant des zones epikarstique et d'infiltration developpees une part importante des ecoulements a l'exutoire provient de la zone non saturee Cette contribution permanente (reserve suspendue) grace a un drainage precoce et efficace permet de recharger des secteurs permeables de la zone noyee en amont du systeme. Le potentiel hydraulique est transrnis a partir de ces secteurs dans les conduits karstiques jusqu'a l'exutoire Ceci entraine que Jes zones annexes dans la partie basse du systeme sont mal drainees avec pour consequence un colmatage accru et des discontinuites hydrauliques importantes Ces sous-systemes ne se vidangent que lentement et seulement en periode d'etiage. Leurrecharge provient soit des drains lors de fortes raises en charge soit elle est locale si l'organisation karstique du sous-systeme le permet. Des informations cruciales sur ce fonctionnement peuvent etre obtenues lors d'evenements a caractere exceptionnel qui permettent d'activer ces systemes annexes et parfois de modifier en profondeur le fonctionnement du karst. 1. Introduction Les etudes effectuees dans le Vaucluse en France montrent que dans les systemes karstiques fortement heterogenes le fonctionnement devient complexe en raison des contributions variables des differents reservoirs aux ecoulements. Ces participations, rnises en evidence par les traceurs chirniques et isotopiques naturels (MUDRY et al. 1994 ), dependent des conditions hydrologiques qui existent a une echelle de temps reduite mais aussi de l 'organisation geometrique de l'aquifere, heritage de !'evolution anterieure du systeme Dans les aquiferes a forte denivellation les ecoulements retardes provenant de la zone epikarstique et de la zone d'infiltration ont un role quantitatif et qualitatif important en hautes eaux comme en basses eaux (WILLIAMS, 1983; LASTENNET 1994) La zone noyee, quanta elle, est compartimentee en systemes annexes (MANoIN, 1975) plus ou moins permeables C'est avant tout la qualite du drainage mettant en relation les differents secteurs de cette zone qui de:finira la contribution de ces sous-systemes aux ecoulements La caracterisation d'une telle organisation dans la zone noyee reste difficile a demontrer en particulier dans les systemes fortement hierarchises pour lesquels l'exutoire est souvent unique Ceci entraine qu'il existe peu d'inforrnation directe (outre les forages) sur les secteurs peripheriques ou eloignes du drain terminal L'acces a cette information est cependant possible dans certains cas lors de crues a caractere exceptionnel. Nous avons pu en effet observer dans un karst du Vaucluse a la suite d'un orage a l'origine des inondations de Vaison-la-Romaine en septembre 1992 (300 mm en 4 heures) la rnise en activite de plusieurs sorties d'eau, autre que la source principale perenne, en aval du systeme Le suivi de ces nouveaux griffons au cours de cette periode a fourni des informations interessantes sur le fonctionnement et l'organisation du karst de Notre-Dame des Anges. 2. Le systeme karstique de Notre-Dame des Anges Le bassin d 'alimentati on de cette emergence estime a 20 km2, est difficile a delimiter en raison d 'une geologie complexe (calcaires Barremo-Bedoulien) Ilse situe dans le Vaucluse, au nord du moot Ventoux en peripherie du bassin versant de la Fontaine de Vaucluse (1100 km 2 ). La denivelee est d 'environ de 1600 metres entre les cretes du Mont Ventoux (1909 m) et l'exutoire (290 m). Ce dernier se situe en aval des gorges du Toulourenc, a une cote proche du lit de la riviere, ce qui a entraine des difficultes importantes pour le suivi des debits. Il faut signaler la presence d'une grotte surmontant le griffon principal perenne, extrernite du drain terminal reconnu par les speleologues jusqu'a une profondeur de 100 metres (Jolivet et Isler en 1990 et 1992) 6 th Conference on Limestone Hydrology and Fissured Media 17

PAGE 29

Le suivi hydrochimique et isotopique realise en 1991 et 1992 (figure 1) a montre !'importance des apports provenant de la z one d infiltration ( reserve suspendue Ila mis en evidence Jes fortes discontinuites de la zone noyee (LAST E NNET & MUDRY 1997) Nous avons ainsi constate que l'eau a l'exutoire etait systematiquement en desequilibre thermique avec 1' encaissant. La temperature mo y enne ( 12 1 C) calculee sur deux ans est de 1 degre plus froide que la temperature mo y enne de l air (13C) ceci malgre une circulation profonde des eaux dans la zone noyee Du point de vue h y drochimique l etiage prononce de I ete 1991 (6 mois) a permis d identifier un type d eau qui pourrait correspondre a la vidange de sous-systemes rarement sollicites Ces eaux a long temps de sejour se caracterisent par des teneurs elevees en silice et en magnesium et faibles en traceurs provenant de la surface (nitrate et chlorure) La temperature encore fluctuante a la fin de l etiage se rapproche des conditions d equilibre (12 8C). Au cours du cycle 1992 la pluviosite reste mo y enne mais les pluies sont mieux reparties qu'en 1991. Pendant toute l annee 1992 l'ecoulement est influence par des eatLx d origine superficielle marquees en nitrates et chlorures et riches en carbonates :: l r o o o Preclp l tatlona mm 100 I r r' ,.,. I "I I "n I (3 00 mm ) F M A M J J A S O N D J MAM J JAS O N [1 II' I I IJI I' ~~:=tr--1 :~: Q '"'' T ('C) : : j 11 \11 Concentration m g / 1 34 0 320 HCO 3 00 2 80 210 240 220 3 No; 3 2 ,.. ern u1I HCO" 24 / 10 / 9 I 22 10 9 / 92 DJ PMAIIJJAIONDJ PIIAMJJAION 1990 1991 1992 Figure 1 : Evolution de la chimie a N.-D. des Anges au cours des cycles 1991 et 1992. 6 1 -1 -7 6 I 6 II O !I, -1 6 9 9 Concentr1tlon1 u mg/I 12 IO so c11 s 1 0, I M g 2 Ces observations tendent a montrer que le conduit terminal draine majoritairement des eaux situees plus en amont du systeme. Elles proviennent de secteurs transmissifs recharges en permanence par Jes eaux de la zone non saturee plus froide et marquee plus negativement en oxygene 18. Dans la partie basse du systeme les zones peripheriques au drain terminal ne participent aux ecoulements qu en periode d'etiage prononce en raison de leur mauvaise connexion hydraulique au drain En effet, en periode influencee le potentiel hydraulique est irnprirne a v ant tout par les eaux de recharge de l'amont du systeme Dans le cas de systemes a zone d'infiltration importante nous pensons que cette recharge par ecoulement retarde est perrnanente 11 est probable que les variations observees dans les coefficients de tarissement de certains aquiferes karstiques viennent de cette alimentation retardee (LAS TE NNET et al ., 1995) 18 P roc eed ings of the 12 th International Cong r ess of Spe leology 1997 Switzerland Volume 2

PAGE 30

3. Observations lors de la crue du 22/09/1992 Cette crue fait suite a un evenement pluvieux d'une rare intensite (300 mm en 4 heures) a l'origine des inondations de triste memoire de Vaison-la-Romaine L'impact sur le systeme a ete considerable puisque la forte mise en charge de l'aquifere est a l'origine d'un changement brutal des conditions d'ecoulement. Celui-ci est consecutif a un decolmatage important qui a entraine une modification de la configuration de l'exutoire avec l'apparition d'un nouveau griffon tres productif Outre la modification du regime karstique de cet aquifere (LASTENNET & MUDRY 1995) nous avons pu observer l'activation d'autres sorties d'eau clans la partie aval du systeme Un suivi hydrochimique et isotopique a ete realise au cours de cette periode (figure 2) sur deux griffons, l'un situe en amont (200 metres) de la riviere Toulourenc l'autre situe a hauteur de l'exutoire principal du systeme de Notre-Dame des Anges mais sur l'autre rive de la riviere II faut signaler que le debut des chemiogrammes de la figure 2 sont legerement decales, le 21/09/92 pour l'exutoire principal le 22/09/92 pour le griffon rive droite et le 23/09/92 pour le griffon amont. 13 8 5 5 13 8 T("c) 13 4 -6 13 2 -6 5 13 12.8 -7 12.8 12.4 12.2 -8 12 11 a I:() 5 10 15 20 25 5 10 15 20 25 110 8 5 8 Mg'mg/1 105 5 100 5 4.5 95 4 90 3 5 3 85 ca 2 mg/I 2 5 80 2 t-0 5 10 15 20 25 I:() 5 10 15 20 25 4 5 5 5 imgll 5 4 4 5 3 5 3 3 5 2 5 1 2 I:() 5 10 15 20 25 t:0 5 1 0 15 20 25 2119192 1tv10/ll2 2119192 1 tv10/ll2 -Exl.to lr e principal o Griffon rtve droile Griffon amont Figure 2: Evolution geochimique aux exutoires du systeme de N.-D. des Anges au cours de la crue du 22/09/1992 Exutoire principal : On observe le jour de la crue une augmentation brutale des carbonates et du calcium Le calcul des indices de saturation a cependant montre des eaux fortement sous-saturees Ces observations indiquent que Jes eaux ecoulees ne proviennent pas directement de I' orage mais etaient en transit dans le systeme en desequilibre avec la matrice calcaire Ceci entraine qu'un potentiel de dissolution important peut etre transmis dans la zone noyee lors d'une forte rnise en charge. Ces eaux, necessairement recentes provenaient de }'infiltration des pluies de deux orages precedant l evenement du 22/09/92 Ces deux episodes pluvieux (90 mm et 42 mm) n'ont en effet provoque qu'une tres faible 6 th Conferen c e on Limestone Hydrolog y and Assured Med i a 19

PAGE 31

augmentation de debits a l'exutoire ce qui tend a prouver que ces eaux etaient encore en transit dans le systeme L evolution de l oxygene 18 corrobore cette hypothese Nous avons observe en effet un enrichissement spectaculaire de cet element au cours de la crue or les deux orages precedents avaient des teneurs tres peu negatives en isotopes lourds (4 35 8 %o) L evolution des chlorures et surtout des nitrates indique que des eaux d origine superficielle parviennent a l e>..'Utoire seulement 5 ou 6 jours apres l orage Griffon amont: Celui-ci se presente sous la forme d un boyau circulaire situe deux metres au-dessus du lit de la riviere Il draine episodiquement un secteur de la zone no y ee en aval du systeme de Notre-Dame des Anges Un pompage d essais effectue en 1980 par la DDA a permis de relier hydrauliquement ce griffon au drain terminal du systeme Les observations faites au cours d un suivi realise recemment (EMB LAN CH these en cours) montrent qu il s agit d un sous systeme a l'equilibre thermique (eau proche de 13C) tres stable du point de vue isotopique et tres peu influence par !'infiltration rapide d ou des teneurs en chlorures et nitrates faibles et constantes Ces caracteristiques se rapprochent fortement des eaux analysees a la fin du grand etiage de 1991 (figure 1) Ceci laisse penser que ces sous-systemes ne sont sollicites par le drain terminal qu en basses eaux Cette relation s inverse lorsque le drain est fortement rnis en charge apres une pluie. Ainsi le lendemain de la crue du 22/09/92 on peut observer que les eaux sont tres voisines en temperature en chimie et surtout en oxygene 18 II est clair que ce secteur est fortement influence par les eaux du drain terminal en raison d un potentiel h y draulique tres superieur a la normale dans ce dernier Enfin, les faibles variations des chlorures et des nitrates en cours de decrue indiquent que ce sous-systeme est peu recharge par les ecoulements differes en provenance de la zone epikarstique Griffon rive droite : Les similitudes observees dans les differents griffons lejour de la crue indiquent qu il s agi t du meme type d eau et qu il existe bien une communication entre le drain principal et ce secteur de la zone noyee Cependant a l inverse du sous-systeme precedent, ce dernier est nettement plus influence par les ecoulements venant de la zone d'infiltration Ainsi on note le lendemain de la crue une forte augmentation des chlorures et des nitrates qu i temoigne de l'arrivee d une eau superficielle Le dephasage de quelques jours constate avec le drain principal tend a demontrer un fonctionnement independant de ce sous-systeme Celui-ci semble nettement plus karstifie d ou une recharge plus locale et un fonctionnement propre Conclusion: Les observations effectuees dans le systeme de N.-D des Anges au cours des cycles 1991 et 1992 pendant les crues de septembre 1992 et plus recemment en 1995 montrent les discontinuites importantes qui regnent dans ce type d aquifere caracterise par une zone d infiltration developpee et une forte hierarchisation Cette organisation permet une contribution importante et permanente d eau de la zone non saturee aux ecoulements Le drainage des eaux s effectue en grande partie dans la zone d infiltration ce qui permet une recharge constante de secteurs contributifs a forte permeabilite de la zone noyee Ce soot ces derniers qui alimentent preferentiellement le drain vers I'exutoire Dans ce t exemple la temperature et les isotopes permettent de localiser cette zone en amont du systeme a une cote superieure de 200 a 300 metres a l exutoire. Dans la partie basse du systeme le colmatage (bones rouges precipitation de calcite depots de la riviere .. ) est a l origine des mauvaises connections hydrauliques entre les systemes annexes et le drai n terminal Ce processus semble cependant reversible en cas de forte mise en charge La recharge de ces zones demeure encore a etudier. II semble cependant que dans certains cas elle est surtout locale (griffon rive droite) et que dans d autres elle s effectue par les drains (griffon amont) lorsqu ils soot suffisamment en charge References ISLER, 0. 1992 Informations lnfo-Plongee 57 : 16-17 JOLI VE T P 1990. Explorations du C L.A.C (36) lnfo-Plongee 5 3: 7-9. LAST E NNET R. 1994 Role de la zone non saturee dans le fonctionnement des aquiferes karstiques Approche par l'etude physico-chimique et isotopique du signal d'entree et des exutoires du massif du Ventoux (Vaucluse) These Doct U ni v. d 'Av ignon, 239 p LAS TE NNET R. & MUDRY J. 1995 Impact d un evenement pluvieux exceptionnel sur le fonctionnement d un systeme karstique Cas de I'orage du 22/09/92 a Vaison-la-Romaine (Vaucluse France) C. R. A cad Sc i Paris 320-II : 953-959 LAST E NNET R. & MUDRY J. 1997 Influence of the flow conditions on the functioning of heterogeneous karstic systems E nv i ronmental Geology (a paraitre) LAS TE NN ET, R. ; Pt.JIG J.-M .; EMBLANCH C & BL AV O UX, B 1995 Influence de la zone non saturee sur le fonctionnement des systemes karstiques Mise en evidence dans les sources du Nord-Vaucluse Hydrogeologie 4 : 57-66 MANGIN A. 1975 Contribution a l'etude h y drodynarnique des aquiferes karstiques Th e se Doct d' E tat Uni v. Dijon A nn Speleol ., 30 : 21-124 MUDRY J. ; LASTENNET R. ; Pt.JIG J.-M .; BL A VOUX B 1994 Use of natural tracing to understand how an aquifer works Intake area recharge fluxes transit. An example from the Ventoux-Vaucluse karst systems (southeastem France) COST 65 Crampon, N. et Bakalowicz M ., eds Basic and Applied H y drogeologica/ Research in Fren c h Areas p 27-35 EC Brussels WILLIAMS P W 1983 The role of the subcutaneous zone in karst hydrology in: W. Back and P .E. LaMoreau x (Guest Editors ), V. T. Stringfield Symposium Processes in Karst H y drology J Hydro/ ., 61 : 45-67 20 Pr oceedi n gs of t he 12 '" I n t erna t ional Congress of Spe leology, 1997, Switzerland Volume 2

PAGE 32

The karstic collapse earthquake i nduced by water storage of Geheyan reservoir in the Qingjiang River, Hubei China Shenjifang Wanjunwei Yuqingchun China University of Geosciences Wuhan Liy a nyun Xio n gzhaoyu Chenru n fa Caijinyan The company of limited liability of hydroelectric developments of Qingjiang river Hubei Abstract The Geheyan Reservoir is located in karstified carbonate rock areas of South-Western Hubei, China During the initial stage of water storage, distinct seismic events occurred along with large-scale rises of water level of reservoir in four places on both banks In less than three months after sealing up the bottom of holes monitoring network of the reservoir 282 earthquakes were recorded with a maximum magnitude of 3.8. Combining the spatial distribution pattern of the seismic activities with the geologic environment in the epicenter areas we consider that Geheyan Reservoir induced earthquakes are caused by the action of underground karst collapses. It is different from that due to surface karst collapses described in the literature. This paper approached the forming mechanism condition and controlling factors of underground karst collapse and considered the development tendency of earthquake as a result of karst collapse during reservoir operation. 1. Introduction The Geheyan reservoir located in carbonate areas of the south-western Hubei is the first phase of the major project for stepped exploitation of the Qingjiang River a tributary of the Yangtze River. During the initial stage of water storage in 1993 distinct seismic events occurred along with large-scale rises of water level of the reservoir in four places on both banks In less than three months after sealing up the bottom of holes monitoring network of the reservoir 282 earthquakes were recorded of which 118 have located epicenters the largest magnitude being about 3.8 ( M L) Only by comparing the earthquake above magnitude 1.5 with the regional data over 33 years (the monitoring network has been working since 1959), frequencies per year are 30 times over the perennial average observation value, but the energy released is only 1/10-1/50 the average value from the Xiannushan fault. Combining the spatial distribution pattern of the seismic activities with the geologic environment in the epice n ter areas it is considered that the seisms during the initial stage of water storage of the Geheyan Reservoir should belong to reservoir induced seismicity, which is mainly represented by karstic collapse earthquake 2. Characters of seismic activities and its geologic environment According to the macroscopic survey and the instrumental records of the seismic monitoring network there are four relatively gathering epicenter areas. The strongest shock was felt in the Xiacun village where occurred the strongest earthquake of magnitude 3.8 in the Geheyan Reservoir area and a following-up aftershock succession ( nearly 60 times ) during May 30 to 31. Deformations were not seen at the surface, except that some local houses were destroyed But thi s seismic regime was not found in the exposed karst area originally monitored from the Xiannushan fault to the dam site The Xiacun seismic areas is located in the western sector of the Xiwan anticline in the southern bank of the Qingjiang River. The exposed strata are easily dissolved carbonate rocks in the core of the anticline Water-conducting faults and basically enclosed boundary of water catchment are favorable to the development of karst. Although the overall development degree of karst is only limited compared with that of the region, because the topography and the tectonic pattern does not completely match the main direction of the underground runoff. But where the Qingjiang River crosscuts the Ordovician strata at the Bashanxia Gorge groundwater is centralized and discharged in the two sides of the intensely developed fracture belt of the Xiannushan fault. In the Xiacun collecting area the karst in the lower Ordovician limestone buried under the middle and upper Ordovician marl is intensely developed, as well as caves and underground rivers up to 1-3 km long have been found Other sections close to the river banks do not show similar conditions; therefore the 6 th Conference on Umestone Hydro l ogy and Rssured Med i a 2 1

PAGE 33

Xiacun seismic activity is typical for karst collapse earthquake induced by water storage of reservoir and also i s the most obvious consequence reservoir-induced seismicity in the Geheyan Reservoir area 3. Forming mechanism and controlling factors of underground karst collapse The karst collapse earthquake induced by water storage of the Geheyan Reservoir is due to the following processes. Dry ca v es originally buried in the vadose zone are rapidly filled with water simultaneously with the rise of water level of the reservoir. The confining beds of cave roofs are submitted to positive hydrofracturing and high pressure impact of compressed air which result in rock blasting and collapse The impact-blasting collapse of buried underground karst which occurs in special enclosed environment with relatively impermeable overlying formations is different from that of surface karst. The forces resulting in collapse are mainly growing gas and water pressure. The former depends on cave scale and increasing velocity of water level while the latter principally depends on the range of water level fluctuations The collapse process goes from developing to gradually atten u ating with a remarkable and lasting vibrating effect. While the original large-scale dry caves in the vadose zone are filled with water promptly along with the increase of water level of the reservoir the air in the caves is compressed and gathers internal energy then makes a positive impact on the upper cave walls When the force increases to exceed the permissible compression of the rock masses, the roof rocks are brittly destroyed and collapse with high sound and vibration consequences. While the ascending water press u re reaches a certain value, pressure burst is made on the wall rocks and then closed microfissures are made open, and pressure on the rock masses is reduced Impact of the twofold air pressure is superimposed and results in the roofs destruction and collapse. The internal energy of the air is realeased and reduced with the roofs collapsing and the caves enlarged. Water pressure is also lowered with higher roofs but intermittent impacts and destructions to the roofs continue but their strength i s gradually weakened When the forces reach an equilibrium with the rock mass intensity the impact-blasting collapse s stop The development process of the karst collapses mentioned above shows that it is possible that water storage o f reservoir in carbonate rocks area induces karstic impact-blasting colJapse together with earthquakes as long as there i s a certain number of caves in the original vadose zone below the water level of the reservoir. These processes chiefly happened in the stage of large rise of water level during the initial stage of water storage in the reservoir The controlling factos are : ( 1 ) Scale, buried depth and enclosed degree of the karst cave (2) Lithological character and tectonic features of the roof rocks. ( 3 ) Ascending velocity and range of the water level in the reservoir. 2 2 Proceed i ngs of the 12 th International Congress of Spe l eo l ogy 1997 S wi tzer l and Vo l ume 2

PAGE 34

Hydrogeologue versus speleologue, ou de qui releve l'etude et l'exploration des eaux souterraines karstiques ? par Michel Bakalowicz charge des programmes de recherche CNRS-BRGM en hydrogeologie karstique, CNRS UMR 5569, Geofluides Bassins Eau, Case 057 F-34095 MONTPELLIER CEDEX 5 (France) Abstract. Tracing tests and sometimes investigations such as exploration of karst conduits for instance are done by cavers in France in the aim of defining protection areas for karst water supply or exploitation conditions of karst groundwater The cavers who promote such investigations over all consider that the only direct approach that of the cave explorer, gives the real knowledge about karst and allows to understand, to exploit and to protect karst groundwater. After defining the originality of karst and its function as an aquifer one explains the part of the potential actors of its knowledge, of its exploration and of its management and especially the part of cavers and of hydrogeologists That paper tries to define the boundaries of the domain in which the actors take a part It also indicates to decision makers and to managers that their responsability is directly engaged when defining the conditions of each other contributions Investigations and management of groundwater in karst areas absolutely need an obvious definition of the part plaid by cavers and by hydrogeologists. Cavers should not play a direct part in hydrogeological investigations as it occurs sometimes but they should contribute to them, in co-operation with hydrogeologists. Proposals are made about cave and karst feature data base and about tracing test experiments. Resume Des experiences de coloration et parfois meme des etudes (reconnaissances de conduits karstiques par exemple) sont entreprises par des speleologues dans le but de definir des perimetres de protection de captages ou d'exploiter des eaux souterraines du karst. Les promoteurs de cette demarche considerent avant tout que seule I'approche directe celle de l'explorateur, permet de connaitre, de comprendre, d exploiter et de proteger les eaux souterraines karstiques A partir d un rappel de l'originalite du karst et de sa fonction d'aquifere, le role des differents acteurs potentiels de sa connaissance, de son exploration et de sa gestion est precise, en particulier celui du speleologue et celui de l'hydrogeologue. Le but est de tenter de definir les limites d'intervention des unset des autres et de preciser aux decideurs et aux gestionnaires que leur responsabilite est engagee par leur definition des conditions d'intervention des uns et des autres Il y va de la credibilite de chacun des acteurs de l'etude et de la gestion des eaux souterraines karstiques. Plutot que d intervenir directement dans des etudes hydrogeologiques cornrne cela se produit parfois les speleologues devraient etre associes a des hydrogeologues, chaque fois que necessaire et dans le cadre de leurs competences. II est desormais devenu banal de definir le karst cornme un milieu particulier dont la complexite ne peut pas supporter une simplification qui conduirait a une generalisation des caracteres observes ou supposes dans un modele universe!. En fait, l originalite du karst tient d'abord a son ambigwte intrinseque : s agit-il d'un paysage, offrant les spectacles naturels parrni les plus extraordinaires de la planete? ou bien d un monde souterrain paysage Jui aussi, mais renfermant mythes et legendes de l'humanite ? ou d un aquifere, un reservoir, comme ii en existe dans d'autres roches mais avec des particularites dues au fait que l eau qui le parcourt modifie considerablement, et rapidement a l'echelle des temps geologiques, Jes caracteres et Jes proprietes originels des roches aquiferes carbonatees ? II est evidemrnent les trois en meme temps, mais le karst est un paysage sou terrain et superficiel parce qu'il est d'abord un aquifere parcouru par des eaux souterraines. Pour ces raisons, l'ambigui:te du karst tient tout autant a cette imbrication des definitions qui peuvent en etre donnees qu aux auteurs memes de ces definitions, qui sont aussi Jes acteurs de sa connaissance ( MARTEL, 1921; GEZE 1965, 1974 ; MANGIN, 1974, I 975) Dans la realite, chacun tend a etendre son champ d'actions au-dela de son domaine, au-dela de ses competences en considerant que son approche I' autorise a intervenir dans tous Jes domaines concernant le karst. 1 La conception speleocentrique du karst Pour E A MARTEL (] 921 ) le karst sou terrain, dont ii fut le premier explorateur et observateur attentif le premier speleologue, n etait qu un tube reliant une perte a une source a !'image de la grotte de Bramabiau, dont ii fit un s ymbole Sa vision du karst evolua ensuite vers celle de l' egout le jour ou ii decouvrit le tout-au-karst et la pollution des sources karstiques (MARTEL, I 925 ). Du karst, ii donna !'im a ge d un milieu tre s vulnerable peu apte a etre exploite OU a etre arnenage ; ses rapports aux Ministeres de I' Agriculture et de la Sante publique ainsi que ses conferences en temoignent abondarnment. Les con s equences de sa vision largement vulgarisee du karst s ont encore d'actualite malgre pres d un siecle d explorations d etudes et de travaux de toute sorte sur le karst. En effet, Jes regions karstiques sont encore souvent considerees par Jes decideurs et Jes amenageurs, aussi bien que par Jes speleologues et certains defenseurs de l'environnement, comme des regions particulierement fragiles ou tout amenagement doit provoquer des alterations irremediables de l environnement en general et des ressources en eau en particulier. De ce fait le developpement des regions karstiques a souvent ete neglige alors qu e!les offrent un interet economique certain, mais impliquant une politique d amenagement du territoire adaptee a leurs caracteristiques. Par exemple, dans de nombreuses regions d'Europe, Jes ressources en eau des aquiferes karstiques ont ete negligees au profit des autres eaux souterraines et des eaux de surface. Alors que toutes sont exploitees au maximum, sinon meme surexploitees et/ou contarninees plus ou mains gravement 6 th Conference on Limestone H y drology a nd Fissured Media 23

PAGE 35

les eaux des karsts constituent desormais la principale ressource potentielle permettant d' envisager un developpement economique. Aussi, malgre les reticences et les difficultes conceptuelles et methodologiques d etudes des aquiferes karstiques, les eaux du karst deviennent un enjeu de la strategie de developpement economique Ceci est particulierement vrai pour les regions mediterraneennes dont !'expansion est importance et les besoins en eau difficiles a satisfaire du fait du contexte climatique particulier. C est dans ce contexte que sont engagees des etudes des ressources en eau des karsts et de leur vulnerabilite, ainsi que leur exploitation Cette evolution, imposee par Ja necessite, fait se tourner les decide u rs, commanditaires de ces travaux vers tous ceux qui sont concernes par le karst, qu'ils soient specialises ou non dans Jes questions d'ecoulements souterrains 2 Consequences C'est ainsi que des experiences de coloration et parfois meme des etudes sont entreprises par des (groupes de) speleologues dans le but d'aider a definir des perimetres de protection de captages ou en vue d'exploiter des eaux souterraines. Ces travaux s'inscrivent clairement dans la conception martelienne du karst. Ils doivent en outre etre rapprochees de prises de position publiques de representants de groupes speleologiques affirmant que seule l'approche directe, celle de J'explorateur, permet de connaitre le milieu souterrain karstique et, par consequent, de comprendre, d'exploiter et de proteger les eaux souterraines qu'il contient. Bien qu'excessif et caricatural, et le fait d un petit nombre de speleologues, ce discours est malheureusement pris en consideration par certains decideurs mal informes, ou plus preoccupes d'etudes peu coilteuses et/ou de complaisance. Ces derniers font parfois meme appel a certains speleologues comme conseillers en matiere d'eau et d'environnement, alors que leurs competences en ces domaines sont inexistantes Ces speleologues, explorateurs et connaisseurs des seules cavites souterraines naturelles, outrepassent evidemment leur role, comme le faisait deja E AMARTEL Mais a la difference de cette epoque, le karst et les circulations d'eau qui le creent et le parcourent sont beaucoup mieux connus, en partie grace a ces explorations ; en outre, ii existe desormais des specialistes hydrogeologues maitrisant Jes concepts et les methodes qui permettent d'etudier les eaux souterraines du karst et de proposer des scenarios de leur exploitation et de leur protection, c'est-a dire de leur gestion durable 3. Critiques de la conception speleocentrique du karst Ces critiques portent sur les concepts eux-memes : quelle est !'importance et la representativite des vides parcourables par l'homme, par comparaison a ensemble des vides effectivement parcoums par I' eau souterrains ? Elles portent aussi sur Jes methodes d'etude du milieu : l'approche directe est-elle necessaire et suffisante pour comprendre le fonctionnement hydrologique de J'ensemble? En ce qui concerne J'importance des vides speleologiques, ii est connu depuis longtemps qu'ils ne representent en general en volume que quelques pour-mille, exceptionnellement un pour cent, du volume total des vides occupes par l'eau En outre, la plupart du temps, Jes cavites parcoumes par Jes speleologues soot abandonnees par Jes circulations hydrologiques actives correspondant au fonctionnement actuel de l'aquifere : la plus grande partie des grottes connues sont des restes de reseaux de drainage karstique, abandonnes a la suite de modifications environnementales Elles informent sur l'histoire du karst, aident parfois a comprendre certains aspects particuliers de son fonctionnement (trop pleins, captures), mais ne permettent pas de decrire le fonctionnement de I aquifere ni de definir les ressources en eau ni de predire !'existence ou la position de reserves noyees et leur vulnerabilite aux pollutions : les v ides speleologiques ne sont pas representatifs de ensemble du milieu karstique. En revanche, c est la connaissance speleologique du milieu qui a conduit a concevoir le milieu karstique differemment des milieux poreux et fissure et, ainsi, de contraindre les hydrogeologues a construire des modeles conceptuels prenant en compte c es observations indispensables En effet malgre Jeur modeste contribution au volume total des vides les vides speleologiques constituent des elements remarquables dans lesquels Jes lois d'ecoulement different tres fortement de celles regnant dans Jes vides de petites dimensions Les vides speleologiques introduisent des discontinuites majeures dans la structure et dans le fonctionnement de l'aquifere karstique. Cette absence de representativite des vides speleologiques a pour consequence que Jes observations faites directernent a Jeur niveau ne peuvent pas etre generalisees a !'ensemble de l aquifere. Seule l'approche indirecte par une "auscultation depuis la surface, est susceptible d analyser l'aquifere Mais la reside la veritable difficulte offerte par le karst. Les milieux poreux et fissures eux aussi ne peuvent etre analyses que par une approche indirecte depuis la surface; mais leur constitution est telle qu'ils peuvent etre conceptualises ou schematises a partir d un volume de terrain ideal de dimension reduite (un cube de quelques dizaines a quelques centaines de metres de cote) dans lequel Jes caracteristiques physiques, et par consequent hydrauliques peuvent etre considerees comme etant homogenes C'est le concept de volume elementaire representatif ( V.E.R ) Des mesures locales, sur les formations geologiques (porosite, fracturation) et dans les forages (caracteristiques hydrauliques ), peuvent etre generalisees au V.E.R. qui est !'element de base perrnettant de constituer le milieu equivalant au milieu reel. Or precisement !'exploration speleologique fournit des element s qui permettent d affirmer que, dans le cas des massifs karstiques cette conception n'est pas acceptable C est de ce constat qu'est parti MANGIN (1975) pour definir le systeme karstique, unite fonctionnelle de drainage correspondant au bassin d'alimentation d'une source karstique Le systeme karstique ne peut pas etre reduit a !'assemblage d'un certain nombre d'elements relativement semblables de fai;:on a le representer dans sa totalite C'est pourquoi ii ne peut etre aborde que grace a une approche indirecte non pas de la structure de ses vides par trop non predictible, mais de son fonctionnement d'ensemble II est bien etabli que le fonctionnement de !'ensemble est expression de !'organisation des ecoulements, eux-memes responsables de celle des vides (MANGIN, 1975) Aussi !'organisation des ecoulernents et des vides du systeme karstique qu ils empruntent est reconstituee a partir de I' analyse du fonctionnement de l' ensemble, ainsi que de certaines de ses parties quand cela est possible. Cette dernarche est strictement identique a celle du medecin qui par I' auscultation du patient definit se s caracteristiques internes et ses eventuels troubles fonctionnels La difficulte majeure de cette approche reside dans la necessite de mise en reuvre d un nombre important de methodes correspondant aux variables susceptibles d'inforrner sur le fonctionnement du systeme et de reconstituer ainsi son organisation Cette approche, qui analyse le fonctionnement pour reconstituer de fa~on indirecte la structure est profondement differente de celle qui fut pratiquee par les premiers speleologues a !'initiative de Martel. Cette derniere decrit la partie penetrable 24 Proceedings of the 12 th Internat i onal Congress of Speleology 1997 Sw i tzer l and Vo l ume 2

PAGE 36

par l'homme et cherche a generaliser ces observations locales a !'ensemble ; elle suppose alors que Jes conditions d'ecoulement observees ponctuellement assimilees a celles de rivieres ou de lacs sont predominantes et representatives de !'ensemble Pour aisee qu elle soit a expliquer cette approche n'en donne pas pour autant une representation exacte ; loin s'en faut. Il est clair que seule l approche fonctionnelle peut permettre de construire progressivement un modele conceptuel coherent du systeme karstique etudie Seule cette approche doit conduire lorsqu'elle est bien menee a une connaissance suffisante du systeme qui fournisse tous les elements necessaires a son exploitation et a la definition des regles de sa gestion durable Une approche inadaptee, corn me I' est la conception speleocentrique s'appuyant sur une methodologie impuissante a resoudre les problemes complexes poses par exploitation des ressources du karst, deconsiderera evidemment par ses echecs, les auteurs de l etude et des recommandations proposees Mais et ceci est bien plus grave, ces echecs inciteront les decideurs, comme ce fut deja la cas a negliger les eaux souterraines karstiques et a leur preferer des amenagements de surface, peu respectueux de l'environnement ou mettant en jeu des solutions techniques lourdes et cofiteuses pour la collectivite 4. Les eaux souterraines karstiques : l'affaire des hydrogeologues avec la contribution de speleologues C est pour eviter ces consequences prejudiciables finalement a notre environnement qu il faut a la fois tenter de definir les limites d intervention des unset des autres (BAKAL0WICZ, 1994 ) et preciser aux decideurs et aux gestionnaires que leur responsabilite est engagee par leur definition des conditions d intervention des uns et/ou des autres ( cahier des charges des etudes et consequences sur !'exploitation et la gestion des ressources ). II y va de la credibilite de chacun des acteurs decideurs, hydrogeologues et speleologues, de !'exploration de ) exploitation et de la gestion des eaux souterraines karstiques Un seminaire national "Pour une gestion active des ressources en eau d'origine karstique (Montpellier 1996) a perrnis d aborder la question de la connaissance et de l etude du karst en tant qu aquifere avec tous les acteurs de son exploration, de son exploitation et de sa gestion en France : services publics, agences de l'eau et collectivites territoriales La place des speleologues dans ces actions est admise et reconnue par chacun d'eux Mais elle apparait parfaitement definie dans le cadre des activites sportives et d exploration qui relevent de leurs competences. Ces competences concement Jes domaines suivants. La connaissance des phenomenes karstiques fait partie des informations de base necessaires a la definition de !'architecture du systeme karstique et contribuant a la connaissance de sa structure de drainage et de son histoire geologique et environnementale Cette connaissance s'appuie, entre autres, sur Jes inventaires de cavites et sur leurs caracteristiques morphometriques c'est-a-dire sur des documents etablis exclusivement par Jes speleologues Dans une demarche premonitoire et de pionnier H.Paloc crea dans Jes annees soixante, sous l egide du Bureau de Recherches Geologiques et Minieres (BRGM) une banque nationale de donnees relatives au milieu souterrain karstique ( cavites et sources) Cette banque de donnees malheureusement delaissee actuellement devrait etre reactivee remise a jour et informatisee car elle presente un interet national certain pour toutes Jes operations d'amenagement des territoires karstiques Mais contrairement a la demarche initiale, reposant sur des contributions individuelles et volontaires, cette tache devrait etre menee dans un cadre contractuel sur des objectifs selectionnes en fonction de leur interet environnemental et/ou economique Des cadastres speleologiques sont en cours de constitution dans c ertains pay s europeens oli la masse de donnees existantes reste cependant tre s inferieure a celle disponible en France Par ailleurs la haute technicite souvent necessaire pour parcourir les cavites naturelles en pleine securite impo s e de recourir a assistance de s peleologue s confirmes pour as s urer l encadrement de specialistes, hydrogeologues notamment devant realiser des mesures ou des experiences sous terre Dans le cadre des etudes hydrogeologiques, ii est le plus souvent fait appel a des speleologues pour realiser ces travaux En revanche, moo experience me conduit a refuser tout credit aux experiences de tra~age ou "colorations reali s ees par des speleologues Pourtant l'activite de tra~age artificiel tient une place importante en speleologie, depuis ses origines Mais les nombreux resultats fournis aussi bien par les speleologues que par la recherche dans ce domaine montrent a I' evidence combien les experience s de tra~age arti ficiel sont difficiles a realiser techniquement pour fournir des donnees dignes de confiance En outre Jes exigences environnementales imposent de plus en plus frequemment que les teneurs en traceur des eaux soient faibles ce qui necessite le recours a des methodes analytiques lourdes et cofiteuses et a la prise de precautions tres strictes afin d eviter tout risque de contamination Ces exigences soot totalement incompatibles avec les moyens techniques dont disposent les speleologues C est ainsi que les resultats de nombreuses experiences realisees au cours de ces trente dernieres annees ont ete recemment remis en cause, notamment en A utri che (COST 65, 1995) et en France ( voir par exemple RICARD & BAKAL0WICZ 1994), pays pionniers en matiere de speleologie et de tra~ages artificiels. Enfin la necessite de realiser des tra~ages dans le cadre stricte d'explorations peut etre conteste, dans la mesure ou le fonctionnement hydrologique actuel est souvent different de l agencement de cavites correspondant a des etats anciens de I' environnement du reseau speleologique C est ainsi que l'inventaire des tra~ages commande par I Agence de l Eau Rhone Mediterranee Corse, et mis a jour recemment, sous le pretexte d'etre exhaustif rassemble toutes les experiences de la litterature speleologique, y compris celles en nombre non negligeable, dont les resultats sont manifestement errones ; cet inventaire n a pas ete I' objet d' analyses critiques, puisqu ii a ete realise par les speleologues eux-memes, qui ne disposent pas des moyens ni des connaissances, notamment hydrogeologiques, pour assurer cette critique De plus l'accroissement des etudes sur le karst en relation avec la definition des perimetres de protection de captages et ave c quelques grands chantiers ( autoroutes voies ferrees et canaux ) conduit a une multiplication des tests de tra~age et par consequent, a !'apparition de risques d interferences entre des experiences conduites simultanement et sans concertation II est evident que dans ces conditions, les experiences de tra~age ne devraient pas pouvoir etre entreprises sans qu une concertation s etablisse entre Jes differents intervenants du milieu karstique II faudrait envisager une declaration des projets de tra~age aussi bien par les groupes speleologiques que par les bureau d etudes et les services publiques d une part pour eviter toute interference provoquant un echec, d'autre part pour optimiser des operations lourdes et souvent cofiteuses 5. Conclusions Au debut de la structuration des groupements de speleologues, les geologues hydrogeologues et geomorphologues specialistes du 6 th Conferen c e on Limestone Hydrolog y and Fissured Med i a 25

PAGE 37

karst etaient peu nombreux, mais proportionnellement nombreux et actifs dans ces groupements Leur contribution aux travaux speleologiques a ete considerable. L'evolution des connaissances et des methodes d' etude, de mesures et d' analyses a progressivement donner une importance moindre aux donnees speleologiques dans I' acquisition des connaissances fondarnentales sur le karst. Parallelement, l'accroissement de la demande sociale en matiere d'etudes et d'expertises sur I'environnement karstique a incite certains speleologues a etendre Ieur champs d'activites au-deli\ de leurs competences Un dialogue devrait s'instaurer entre Jes differents acteurs, et tout specialement entre des speleologues ou leurs representants, et Jes hydrogeologues, de service public comme de bureaux d'etudes, pour eviter la prise de decisions unilaterales, contraignantes (par exemple, demandes d'autorisation d'exploration ou de realisation de tra~ages) et pour que soit pris en consideration et valorise le fruit des explorations speleologiques En contrepartie, Jes speleologues devraient veiller ace qu'aucun d'entre eux ne pretende intervenir directement dans une etude du milieu karstique, qui conceme Les eaux souterraines, leur gestion et leur protection Par le dialogue, hydrogeologues et speleologues doivent parvenir a une reconnaissance mutuelle de leurs competences, qu'ils devraient, le cas echeant, etre capables d' associer. References. Bakalowicz, M. 1994 La place des speleologues dans la connaissance du karst Introduction au debat. Spelunca Memoires. 21 : 40-42. Bakalowicz M & J. Ricard. 1994 Karstic aquifers of the Grands Causses (France). Water reserves that must be protected, p 137148 Crampon, N. & M Bakalowicz, COST 65 Action "Hydrogeological aspects of groundwater protection in karstic areas". Basic and applied hydrogeological research in french karstic areas. Montpellier-Millau Workshop, may 5-8, 1994. European Commission, Directorate-General, Science, Research and Development, Bruxelles. Bakalowicz, M. & J Ricard 1994 Gerer et proteger les ressources en eau du Larzac septentrional. Revue Agence de l'Eau Adour-Garonne 61 : 19-24. COST Action 65 1995. Karst groundwater protection. Final Report EUR 16547 EN, 446 p. Geze, B. 1965 La speleologie scientifique. Le Seuil, Paris, 190 p Geze, B 1974. Le "trou" et son environnement. Spelunca Memoires 8 : 43-47 Mangin, A 1974 Notion de systeme karstique. Spelunca Memoires. 8 : 65-68 Mangin, A. 1975 Contribution a l'etude hydrodynamique des aquiferes karstiques Doctorates Sciences naturelles, Universite de Dijon, 260 p. Martel, E.A. 1921. Nouveau traite des eaux souterraines Doin ed., Paris, 840 p Martel E.A. 1925 La contamination geologique des eaux souterraines. Annales d'Hygiene publique, industrielle et sociale l : 1-48 26 Proceed i ngs of the 12"' International Congress of Speleology 1997, Switzerland Volume 2

PAGE 38

Role of Speleology in Karst Hydrology and Hydrogeology Ognjen Bonacci. Faculty of Civil E n gineering, University of Split, Matice hrvatske st. 15, 21000 Split, Croatia Abstract Technology of speleological investigations enables special measurements of features of underground conduits, caves, pits, jamas, shafts and the other solutional sculpturings such as scallops and cave rocky relief. Speleologic investigations can reveal the positions, dimensions and interactions of underground and surface karst features and water flow in the karst and on its surface. Speleologists are capable of investigating the hydra u lic conditions under which laminar or turbulent flows occur in conduits and small and narrow karst fractures. From s u ch investigations crucial parameters for hydraulic, hydrologic and hydrogeologic modelling such as dissolution-bedform and hydraulically-transpo r ted sediment, can be obtained. For these reasons, the role of speleology in karst hydrology and hydrogeology should (and undoubtedly will) in future be given much more importance. This paper briefly explains the main theoretical aspects and gives some practical examples and experiences from Dinaric and others karst regions. Resume Ce role sera, prochainement, bien plus important que jusqu'a present. En utilisant les technique speleologiques, ii est possible d'effectuer Jes mes u res speciales des formes souterraines karstiques telles que: tubes karstiques, gouffres, grottes et rugosites du fond et des parois des conduits de circulation d'eau Pour l 'etablissement des modeles hydrologiques et hydrogeologiques, ii est de plus haute impo r tance de connaitre la rugosite des parois karstiques et Jes caracteristiques d'alluvion transporte par l'eau a travers de conduits karstiques. A !'aide de ces informations, il est possible d'identifier le type et Jes caracteristiques de l'ecoulement souterrain karstique Par Jes techniqus speleologiques on decouvre Jes positions, Jes dimensions et !'interaction des formes karstiques de surface et souterraines, done Jes principes d'ecoulement d'eau en surface et dans le souterrain du karst. La presente etude donne Jes explications theoriques de base et cite de nombreux cas pratiques du karst des Dinarides. 1 Introduction The role of speleology in karst hydrology and hydrogeology as well as in ka r st enginee r ing today is ve r y well-known and unavoidable. The division of the main applicable speleologic explorations in the karst regions can be made in following way: 1 Explorations near traffic structures and quarries; 2 Foundation of civil engineering constructions; 3 Explo r ations along dams and rese r voirs; 4 Grouting in underground hollows; 5 D iscove r ies d u ring the tunnel excavations; 6 Water intakes from the karst underground; 7 Explorations dealing with karst aquifer protection against pollution. In this paper accent will be put on the specia l and relatively new role of speleology. The modern speleology may and should play key role in explaining hydraulic, hydrologic, and hydrogeologic processes in karst undergro u nd essential fo r flow modelling. The limestone aquifers are two-component systems in which the major part of storage is in the form of tr u e ground water in narrow fissures, where laminar flow prevails. On the other hand the majority of water is transmitted through the karst underground by turbulent flows in solutinall y enlarged conduits. Slow or so-called diffuse flow occ u rs through karst fissures of small dimensions gene r ally in the laminar regime. Turbulent fast flow, or conduit flow, occurs in large fissures through irregular karst conduits, with dimensions varying from 1 cm to few meters. Speleological investigations enab l e the penetration of investigations into karst unde r ground. The space in which the spe l eologist investigator can enter should have the dimensions appropriate to human body. Because of that, speleology is generally usefull in exploration of fast tu r bulent flow. The diffuse flow characteristics can be controlled very well by piezometers. Today, primarily thanks to speleology, karst is not "black box" neither for laymen no r for experts. Intensive use of speleology in combination with other scientific branches can yield precise answers to the position, composition and dimensions of underground karst forms, which serve as main input in complex hydrologic-hydrogeologic models. Shapes of speleologic objects depend on: 1 Lithologic and stratigrafic characteristic of the rocks in which they originated; 2 Intensity and type of tectonic activity; 3 Underground water activity and regional meteorological conditions; 4 Ka r stification rate and speleogenesis The relationships between karst hydrology and hydrogeology and spe l eologic objects is known for a long time The general knowledge should be investigated in detail because of the specific hydraulic function of each individual speleologic object. Introductory remarks will be finished by quotation of types of speleologic objects: 1 Caves; 2 Pits Uamas or shafts) and 3 Conduits (channels, passages) According to the shape and distribution of connections in speleologic objects next morphologic types can be defined: I Simple speleologic object; 2 Branching speleologic object; 3 Speleologic object with different levels; 4 Knee formed speleologic object; 5 Cave or pit systems. Each of these objects can have specific hydra u lic, hydrogeologic and hydrologic function. 2 Inflow-Outflow Relationship Water circulation in karst terrain is characterized by existance of inflow-outflow relation shown in figure 1 (BONACCI, 1987). The speleology can help to resolve some of these complex situations. Figu r e 2 rep r esents hydraulic scheme of the simplest inflow-outflow re l ationship gives as type I on figure 1 A. Under the assumption that water flow through karst conduits is as in the pipes, Bernoully's equation can be applied : 6 "' Conference on Limestone H ydrology and Fissured Media 27

PAGE 39

v 2 L v 2 v 2 H 1 -H 0 =1; 1 -+;\.--+1; 0 (1) 2g d 2g 2g where: 1; 1 and 1; 0 are dimensionsless coefficients defining a losses of velocity head at the entrance-inflow (I) and exit outflow (0) to and from karst conduit; v=Q/A is mean water velocity through karst conduit in m s 1 ; Q is discharge through karst conduit in m 3 s 1; A is cross section of the karst conduit in m 2 ; L is the conduit length in m; d is the conduit diameter in m; g is the acceleration of gravity in ms 2 ; ;\. is dimensionless friction factor. Under the assumption that flow regime in karst conduits is complete roughly turbulent (confi rmed by many investigators (BOGLI, 1980; GALE, 1983; WHITE & DEIKE, 1989 )) the friction factor ;\. is defined by Prandtl-Karman 's equation : (2) where E is absolute roughness of the boundary in m. Speleologic investigation can supply us with exact or at least reliable data of L, d or A, E and v and can give us a glimpse of the karst underground It is well know that karst cond uits have an irregular cross section, differing more or l ess from a circu l ar section (WHITE & DEIKE, 1989; BONACCI, 1995; SLABE, 1995). I I 0 0 0 TYPE I TYPE IT TYPE m TYPE !Va TYPE IVb TYPE V I (C) I main karst system underground storage 0 0 Fig ur e 1: Posible types of r elation bet ween inflow (I) and outflow (0) in karst system A Types of flow network (B ROWN & F ORD 1973); B Presence of underground sto rag e; C Connection between main system and its s ub system It is also necessary to take into account irregularities of cross section in large karst conduits resulting from rock slides and existance of siphons. In addition, occasionally large caves and significant narrow sections occur allong the conduit. Hydraulic concequences of these undergr ound features can be very important for water flow and thus for its mathematical modelling. Let us return to the equation I. The most significant influence on the flow governed by it has the term ;\. (LID) where the conduit lenght is dominant factor. The determination of exact underground karst lenght conduit is extrem ly complicated task which could, for practical engineering purposes, be satisfactory fulfiled using speleologically investigated part of conduit, RICHARDSON's theory (number) (1961 ), SUSTERSIC's (1985) and BONACCI's ( 1995) suggestions. Fig ur e 2: Schematic presentation of inflow (I) -o ut flow (0) karst hydraulic system The essential role for hydraulic conditions of flow in carst conduit is also played by scallops (defined as shallow cuspate pockets dissolved into cave or conduit walls (WHITE & DEIKE, 1989)) and flutes. Scallops and flutes (CURL, 1974; BOGLI, 1980; SERBAN, 1987; WHITE & DEIKE 1989; SLABE, 1995) determine absolute roughness of the boundary E In figure 3 is presented the foto of scallops taken by Slabe in a underground karst conduit in Dinaric karst of Slovenia. The foto is taken in a karst cond ui t fulfiled by water, during one of the speleological investigations. WHITE & DEIKE (1989) on page 233 give the ralation of the scallop lenght to mean channel flow velocity for conduits of several diameters. Sensitivity ana ly sis of equations 1 and 2 on their parameters shows that main role is played by diameter d and discharge Q. Both of them can be succesfully determined only through the speleological measurements. 3 Other Specific Assistance of Speleology Many specific and, for engineering practice, key hydrologic and hydraulic problems could be solved by the fo l l owing procedure: 1 Determination of the flow type (open surface flow or flow under pressure; 2 Determination of hydraulic functioning and role of syphons; 3 Determination of exact ground water level; 4 Determination of hydraulic and hydrologic role of caves; 5 Determination of changes of dimensions and roughnes of karst conduits, and many other. For example, if in karst conduit exists a short narrow passage it could cause rapid drop of energy line significantly greater than drop of energy line caused by absolute roughness of the boundary in a long conduit sector. 28 Proceedings of the 12 '" Internat ional Congress of Speleology, 1997, Switzerland Volume 2

PAGE 40

Using measurements in Mendip Hills (U.K.) ATKINSON (1977) defined 1.22 m energy loss of head in karst conduit with lenght L = 620 m and diameter D = 1.433 m. Figure 4 gives schematic presentation of idealized circular karst cond u it with lenght of L = 620 m and diameter of D = 1.433 m. Supposing that flow velocity is v 1 = 0.047 m/s as in ATKINSON (1977) case m table l, BOJANIC (1994) calculated values of energy Joss: at entrance ti.H 1 at exit ti.Ho and in reduction ti.HR, for different values of Ls, D 5 and v 2 (see figure 4). Figure 3: Scallops on intraclastic limestone (Slovenia), scale 15 cm (taken by S/abe) ( Ls ) L = 620m Figure 4: Schematic presentation of idea l ised circ ul ar karst conduit (pipe) with narow passage From those results it is possible to see that great part of complete 1.22 m loss of head could be caused by morphologic characteristics of reduction. These characteristics should be defined by spe l eologic investigations especially in ka r st conduit with diameter in excess of 1 m. Figure 5 shows a cross-section of one real ponor (jama) in Croatian karst. Due to narrow passage designated by A water level in this ponor after the heavy r ain and during the wet period is over it. Real naximum gro u nd water level is about I 00 m lower. These statements were confirmed by speleological investigations. Ls Ds V2 LOST HEAD ml rml rml rm/sl ti.H1 M-Io t.HD 1.0 0.160 3.770 0.072 0.706 0.453 2.0 0.172 3.260 0.054 0.527 0.631 10.0 0.217 2.050 0.021 0.204 0.987 Tab l e I Losses of head in karst cond u it red u ction showed in figure4 Speleology could play key role in investigation of submarine springs (vruljes) and coastal brachish karst springs. (BONACCI et al., 1995; BONACCI & ROJE BONACCI, 1997). I 1~ _ls ~ I. I l -, 1 ll,5 1 1i I I ~ri I I I I I 1 1 I r _L I I I ::i, I I 1 ',l I ; _;1 I I I LI I_L LI I I I r1 1..:::r1 -1 __ c I I I __ _ A . J -t=1 ... L I _I_ ..._ -I I -, Scale 0 10 Figu r e 5 : C r os-section thro u ght the pono r in Dinaric karst of Croatia Before approximately 25000 years ago, the level of the Adriatic Sea was about 100 m lower than at present. Then, it represented the erosion basis for the karst area so that the 6 1h Conference on Limestone Hydrology and Fissured Media 29

PAGE 41

karst springs appeared at that level. As the sea level was raised the springs come under the sea level, and thus new conduits werw simultaneously formed in the karst, togeter with the new, higher positions of the coastal springs. The change in the position of the springs along the coast depends upon the local geologic structure of the coastal belt. Each actual coastal karst spring has a few conduits. Some of them are active and some are inactive as it is shown schematically in figure 6. The function and dimensions of all conduits is posible to check exclusively by speleologic measurements which should be closely connected with hydrometry. Modern speleology should and could serves for complex multy-and inter-disciplinary mvestigations of karst. The mechanism of sea water 1ntrus1on is relatively well known but the problem of the karst springs desalinisation is not solved practically Speleological submarine measurements have showed that morphologically the vruljes correspond lo the inundated fossile dolines from the pre-alluvial age formed in the continenta l phase of the Dinaric coastal karst. The vruljes frequently are only one element of a larger coastal spring system. Due to the fact that the mechanism of the sea water intrusion is directly connected with the position of the karst conduits through which the sea water penetrates into spring exit. 1t 1s of great importance to detect exact positions and dimensions of these conduits. The role of speleology in these investigations is indispensable. When the researchers are supplied with real field data it is possible lo make many hydraulic and hydrologic computations. Today these hypothetical, but reality based analyses are very fast and are not expensive. This should be the first step in an attempt to explain exceptiona lly complex, interesting and important coastal karst phenomenon FRESHWATER LENS INACTIVE CONDUlTS CLAY CORKS SEA WATER AQUIFER TRANSITION ZONE WITH BRACKISH WATER Figure 6: Schematic presentation of idealized circular karst conduit (pipe) with narow passage 4. Conclusions The role of speleology in engineering practice and theory should be given much more attention than until now. Speleology represents an exceptionally important source of a wide range of information necessary for all types of investigations related to the regime of karst ground water circulation. Speleologic mvestigatins represent a real "wi ndow" into the underground karst medium and make possible detailed insight which is absolutely necessary for theoretical and applied investigations. The speleolog1c measurements should be planed, organised and conducted carefuly as an unavoidable part of complex multiand inter disciplinary karstologic investigations. References ATKINSON, T., C. 1977. Diffuse now and conduit flow in limestone terrain in Mendis Hills, Somerset (Great Britain). J.Hydrol. 35: 93-110. BOGLI, A. 1980. Karst Hydrology and Physical Speleology. Springer Verlag Berlin: 284 p. BOJANIC, D. 1994. Identifikacija hidroloskohidraulickih elemenata otjecanja u krsu (Identification of hydrologic-hydraulic elements of flow in the karst). M. sc. Thesis. University of Split: 1 00p. BONACCI. 0. 1987. Karst Hydrology. Springer Verlag, Berlin: 184 p. BONACCI, 0. 1995. Ground water behaviour in karst: example of the Ombla pring (Croatia), J. of Hydrol. 165: 113-134. BONACCI, 0. 1995. Primjena koncepta deterministickog kaosa prilikom amnalize vodnih resursa (Ap lication of deterministic chaos concept for water resources analysis). Zbornik radova 1. l lrvatske konferencije o vodama, Dubrovnik, knjiga 1: 253-265. BONACCI. 0., FRITZ, F. & DENIC, V 1995. Hydrogeology of Slanac spring, Croatia. Hydrogeology Journal 3 (3): 31-40. BONACCI, 0. & ROJE-BONACCI, T. 1997. Sea water intrusion in coastal karst springs: example of the Blaz pring (Croatia), Hydrology Sci. J. (in press) BROWN, M. C. & FORD. D. C. 1973 Quantitative tracer methods for investigation of karst hydrological system. Trans. Cave Res. Group GB 13 (1): 37-51. URL, R. L. 1974. Deducing flow velocity in cave conduits from scalops. Natl. Spcleol. Soc. Bull. 36/2: 1-5. GALE, S J. 1983 The hidraulics of conduit fl ow 111 carbonate aquifers. J. of Hydrol. 70: 309-327. RICHARDSON. L. F. 1961. The problem of contmguny: An appendix of statistics of deadly quarrels. Gen Syst. Yearb. 6: 139-187. SERBAN, M. 1987. Wall microrelief in caves-effect of turbulence. Theoretical and Applied Karstology 3, Buchuresti, 1-30. SLABE, T. 1995. Cave Rocky Relief and its Speleological Significance. Znanstvenoraziskovalni Center SAZU, Ljubljana: 128 p. SUSTERCIC, F. 1985. Dolocenje dolzine neznanih jamskih rovor s pomocjo fraktalne analize (Use of fractal analysis for determination of lenght of noninvestifated part of the karst underground conduits.) Nas Krs XI (18-19), Sarajevo: 89-83 WHITE, W.B. & DEIKE, G.H. 1989. Hydraulic geometry of cave passages. In W.B White & E.L. White (eds.). Karst Hydrology, Concepts from the Mammot Cave Area. Van Nostrand Reinhold, New York: 223-258. WHITE, W.B. 1988. Geomorphology and Hydrology of Karst Terrains. Oxford University Press, New York: 464 p. 30 Proceedings of the 12 th International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 42

Wasserfarbung im Gebiet Gemmenalp-Sieben Hengste: Z i ele und Resultate Pb. Hii u se l ma nn und M. O tz Ph. Hauselmann Haldenstr. 32, CH-3014 Bern (Schweiz) M. Otz, Neubriickstr. 17, CH-2555 Briigg (Schweiz) Zusammenfassung Mit einer im Jahre 1996 durchgeftihrten Wasserfarbung wurde versucht, das Einzugsgebiet der Beatushohle von dem des benachbarten Systems der Sieben Hengste (nachfolgend Reseau genannt) zu trennen Gleichzeitig wurde die Farbung von 1984 partiell wiederholt da einige Zweifel an den damals gemachten Aussagen bestanden. Im nachfolgenden Artikel werden die Resultate aufgezeigt. Abstract With a coloration in 1996 we tried to separate the two catchments of the Beatus cave and the neighbouring Cave System of the Sieben Hengste (called Reseau below) At the same time the coloration of 1984 has partially been repeated because of apparently doubtful results. In the following article the actual results are shown. Geographischer und geologischer Oberblick Die Kette des Niederhoms Sieben Hengste Hohgant liegt nordlich von Interlaken und gehort zur helvetischen Voralpenkette. Sie weist eine recht regelmassig nach SE einfallende, flache Flanke und einen senkrechten Felsabbruch im NW auf. Die ftir die Hohlenentwasserung massgebliche Schichtfolge setzt sich ( von unten nach oben) aus folgenden Gesteinen zusammen: Zuunterst liegt der bis zu 200 m machtige Kieselkalk. Er ist schlecht verkarstungsfiihig; es finden sich nur wenige Hohlenteile darin. Dies ist zu einem Grossteil auch durch den Oberliegenden Drusbergmergel begrtindet. Dieser Drusbergmergel bildet in den meisten Fallen die undurchlassige Unterlage auf der die H ohlengewasser abtliessen. Er ist bis zu 40 m machtig und weist in seinen obersten Partien kalkige Einschaltungen auf. Es folgt der 160 200 m machtige Schrattenkalk, der das Leitgestein ftir die Hohlenbildu n g darstellt. Nach einer grossen SchichtlOcke, die die oben beschriebenen unterkretazischen Gesteine vom Tertiar trennt, folgen die Nummulitenkalke und der Hohgantsandstein, die stellenweise Ober 200 m macht i g werden und faziell sehr kompliziert ausgebildet sind In den kalkigen Bandem des Sandsteins bilden sich oftmals Hohlen die aber nicht notgedrungen einen Zugang zum oft viel tiefer liegenden Schrattenkalk darstellen mOssen. Die nach SE einfallende Platte ist von einigen Abschiebungen und Blattverschiebungen zerhackt. Die wichtigste Abschiebung ist die Hohgant-Sundlauenen-Verwerfung, die im S 500 m Versatz erreicht und die Einzugsgebiete des Siebenhengste Hohlensystems und der B eatushohle trennt. Die obersten Teile der Schrattenkalkplatte sind stark verkarstet ; dementsprechend ist die Hohlendichte gross Eine weitergehende Obersicht geben u.a. HAUSEL-MANN & JEANNIN ( 1997), HOF et al ( 1985) oder auch BITIERLI (1989) Fri.ihere Farbungen Schon frti h wurden im Gebiet Farbungen untemommen, um den unterirdischen Fliesswegen auf die Spur zu kommen Einen guten Oberblick Ober die vorgenommenen Farbungen gibt WILDBERGER et al ( 1982). Die letzte Farbung wurde 1984 durchgeftihrt (KU)TZLI, 1985) Es stellte sich wie vermutet heraus, dass Fitzlischacht und Haliloch in die Beatushohle entwassem. Es blieben allerdings einige Fragen bestehen, so zum Beispiel trat Fitzlischacht-Farbe in drei scheinbar unabhangigen Bachen auf. Aus diesen Farbungen resultierte die vorgenommene Trennung in zwei Einzugsgebiete, in dasjenige der Beatushohle (mit Wa l dhei m hoh l e, Fitzlischacht, H aliloch als Fa r bepunkte, A u stritt: Beatushohle) sowie dasjenige des Reseaus (mit Barenschacht und Faustloch als die zwei sOdlichsten Farbepunkte Austritt : Batterich und Gelberbrunnen), so wie auf der Figur I dargestellt. Figur 1 : alte h y drologische Karte der Region (nach Bitterli 1989) Ein grosses Gebiet allerdings wurde nie untersucht, weil sich darin keine grossen Hohlen fanden, da der Schrattenkalk an dieser Stelle von Sandstein bedeckt ist. Doch gerade dieses Gebiet zwischen H a li loch, Barenschacht und A2 (Figur 2) ist interessant, da sich irgendwo darunter die Grenze der zwei Einzugsgebiete befi n det. Eine Feldbegehung zeigte mehrere Schlucklocher und Dolinen Eine Farbung ware also machbar 6 th Conference on Limestone Hydro l ogy and Fissured Med i a 31

PAGE 43

(NAPH) (SUL) /) s :-;: Buelbach Q:j r Fitzlischacht .(LIS) Laubloch/ t Akkordloch Balterich /Gelberbrunne .A. Neuhaus lombach Figur 2: Plan der Region, Hohlen, Fiirbepunkte, Entnahmestellen Die Farbung 1996 Dank der Unterstlitzung des Hydrologiebiiros Otz in Bellmund bei Biel war es uns moglich, diese Farbung im Jahre 1996 zu realisieren. Nach mehreren Feldbegehungen entschieden wir uns fiir eine gleichzeitige Eingabe von sieben Farbstoffen Ge 200 gr Lissamin, Pyranin, Eosin, Duasyne, Sulforhodamin, 1 kg Fluorescein und 2 kg Naphtionat) in die auf Figur 2 eingetragenen Farbepunkte Die Farbung im Fitzlischacht wurde wegen der oben erliiuterten Fragen wiederholt. Um moglichst alle Farbe zuriickzuerhalten und die Fliessysteme moglichst genau zu erfassen, wurden an elf Stellen Proben gezogen, niimlich an fiinf Stellen in der Beatushohle im Batterich, im Bad Neuhaus sowie in den Oberfliichenbiichen Lombach Biielbach, Sundbach und Rischerengraben. Zur Organisation und technischen Durchfiihrung der Farbung sei auf Hii.USELMANN ( 1997) verwiesen da eine Aufzahlung hier zu weit fiihren wiirde. Bei der ersten Begehung und am Farbetag wurden bei den Quellen jeweils die Wassertemperatur und die elektrische Leitfiihigkeit gemessen. Die letztgenannte Messung gibt den Gehalt an dissoziierten Ionen eines Fliessgewassers wieder, sie ist temperaturabhiingig. Die elektrische Leitfahigkeit kann wertvolle lnformationen beziiglich der Wasserherkunft liefem. Farbepunkt e Ent nahmest ellen 1 km by Toporobot Auswertung des Flirbversuches Perkin Elmer FilterFluorometer LSB: Das LS -2 B hat auswechselbare Kantenfilter im Anregungsbereich und einen motorisch beweglichen Interferenzfilter (300-800nm) im Emissionsbereich. Eine mit linearer Frequenz gepulste Xenonentladungslampe dient als Lichtquelle Der Fotomultiplier ist dabei die optische Yerstiirkerrohre. Anschliessend folgt elektronische Yerstarkung rnit digitaler Anzeige und Ausdruck aufThermodrucker. Farbstoffe werden rnit einer Empfindlichkeit von 10 6 gtm 3 Wasser nachgewiesen. In der Regel werden die Proben mit diesem Geriit seriell mit festen Einstellungen" ausgemessen, z B. fiir den Tracer Fluorescein wird eine Anregung von 480nm (Priizisions-Kante nfilter ) und eine Emission von 515nm Onterferenzfilter) eingestellt etc. Mit diesem Gerat wurden alle Proben untersucht. Proben mit hohen Peak Werten wurden rnit dem Doppel-Scanning-Yerfahren rnit dem Geriit "Shimadzu RF500 I-PC" iiberpriift : Jeder Stoff hat eine charakteristische Anregung und Emission Man fiihrt von tiefen Wellenliingen zu hohen, wobei der Wellenliingen untershcied zwischen Anregung und Emission konstant gehalten wird Damit erscheinen die Emissions-Maxima der Tracer an bestimmten Stellen. Die Darstellung erfolgt im Konzentrations/Zeit-Diagramm. 32 Proceedings of the 12 '" International Congress of Speleology 1997, Switzerland Volume 2

PAGE 44

(1) Xenon Lamp, 1 50\V (12) Cell (21 Condense r Ellipsoidal Mirror (SiO 2 Coated) (1) faci1a11 o n Slit Assembly (13) L1gh1 collec 1tn g Lens (14) Emission Slit Assembly (4) Concave Mirror (5) Conawe Diffraction Grating (for Excitation) (6) Beam Spli tter Quartz Plate (15) Concave Diffraction Grat ing (for Emission) (16) Concave Mirror (7) Ten on Rcncc1or No I (17) Photometric Photomulliplier, R452 01 (IR) Focus (8) Tenon Rcncc1or No.2 (9) Optical A11cnua1or (19) Entrance Slit (20) Exi t Slit (10) M o n ito r Photomul11plicr, R21 2-09 (11) A Couple o f Ligh1-collcc1ing Lenses (2 1 ) Light Beam Balance Ap ert ure Figur 3: Shimadzu R F5001-PC. Die Funktionsweise und der Aufbau entspricht anniihemd dem des LS-2B Quantitative Mengenbestimmungen: Zu den einzelnen Markierstoffen wurde mit dem LS-2B eine Eichkurve erstellt. Die ausgemessenen Werte um den Wert des "Grundrauschens" = Nullwert reduziert und in die entsprechenden Quantitaten umgerechnet. Mit HPLC (High Pressure Liquid Chromatography) wurden letzte Unklarheiten beseitigt : Gegeni.iber dem Messverfahren mit Fluorometem (Kantenfilterund Gittergeraten) und dem Doppel scanning (Shimadzu RF500 I-PC) ist dies eine unterschiedliche Messmethode Es ist durchaus moglich, dass Huminsauren und andere Fremd stoffe organischer Natur, evtl. sogar andere Tracer in hohen Konzentrationen, einen Farbstoff so maskieren, dass eine Fluoreszenzbestimmung, auch bei chemischer und physikalischer Yorbehandlung, verunmoglicht wird Hier hilft nun HPLC. Einer Tragerfli.issigkeit (mobile Phase) wird eine kleine Menge Probenwasser beigemischt und anschliessend mit hohem Druck durch eine Silicagelsiiule (stationare Phase) gepresst. Die elu ierten Substanzen werden von der mobilen Phase in den Detektor transportiert und vom Schreiber als Gausskurven registriert Qualitativ: Die Rete n tionszeit ist bei gleichen chromat ografischen Bedingu n gen stets gleich gross. Retentionszeit ist die Zeit, welche vom Einspritzen der Probe bis zum Erscheinen des Signalmaximums auf dem Schreiber verstreicht. Die chromato grafischen Bedingungen sind : Trennsaule, Zusammensetzung der mobilen Phase, Fliessgeschwindigkeit der mobilen Phase, Probengrosse und Temperatur. Zur Identifikation eines Peaks konnen wir daher die in Frage kommende Substanz rein einspritzen und die Retentionszeiten vergleichen Quantitativ : Die Flache des Peaks ist der eingespritzen Stoff menge proportional. Wenn wir verschiedene Losungen genau bekannter Konzentrationen einspritzen, die zugehorigen Peak flachen bestimmen und eine Eichkurve zeichnen, konnen wir aus der Peakflache einer unbekannten Probe deren Konzentration bestimmen. Die Resultate Wie erwartet wurde die Farbe des Fitzlischachtes (Lissamin) in der Beatushohle wiedergefunden. Es zeigte sich, dass die i.ibereinander liegenden Gange Hoher Nordgang und Nordgang beide Farbe aufwiesen, der Hohe Nordgang aber etwa in fi.inffacher Konzentration. Die oberhalb des Biirenschacht eingegebenen Farben Pyranin und Duasyne wurden ebenfalls in der Beatushohle wiedergefunden 6 '" Conference on Limestone Hydrology and Fissured Media 3 3

PAGE 45

Das in der Niihe der Oberberg-Alp in eine Doline eingegebene Sulforhodamin fliesst eine Strecke unterirdisch, gelangt dann wieder ans Tageslicht und fliesst via Biihlbach und Lombach ab Das ebenfalls in der Niihe der Oberberg-Alp eingegebene Naphtionat fliesst ebenfalls oberirdisch ab, verschwindet aber zuweilen unter der Erdoberfliiche. Das unterhalb der Alp Biireney eingegenene Eosin zeigt dasselbe Verhalten. Das bei der Alp Biireney eingegebene Fluorescein fliesst komplett ab und tritt wieder im Biitterich und Gelberbrunnen aus Der Ostgang sowie einige sporadisch beprobte Zubringer der Beatushohle, der Sundbach und Rischeregraben sowie die Karstquelle Neuhaus sprachen nicht an Interpretation Die Fiirbung des Fitzlischachtes bestiitigte die Resultate von 1984 nur teilweise Speziell die 1984 nachgewiesene Fiirbung des Ostganges liess sich nicht reproduzieren Die Tatsache dass der Nordgang eine niedrigere Konzentration als der iiberliegende Hohe Nordgang aufweist, konnte mit einer Diffluenz des Hohen NordgangW assers erkliirt werden. Diffluenzen wurden innerhalb der Beatushohle bereits mehrfach festgestellt. Der Fund von Pyranin und Duasyne in der Beatushohle erweitert deren Einzugsgebiet um ein Betriichtliches gegen Norden da die eingefarbten Bache vom Niederhorngrat aus bis zur Burgfeldtluh obertliichlich abfliessen Die sporadisch beprobten Zubringer erbrachten erwartungsgemiiss kein Resultat, da die meisten derartigen Zubringer ziemlich direkt von der Obertliiche her einsickern diirften was auch durch chemische Analysen des Wassers bestiitigt wird (miindliche Mitteilung der SGHI) Der darauffolgende Abschnitt (mit Sulforhodamin Naphtionat und Eosin) entwiissert vollkommen oberirdisch und stellt somit die eigentliche Grenze der Einzugsgebiete dar. Die Tatsache, dass diese Biichlein zuweilen oberirdisch und dann wieder unterirdisch verlaufen, hat ihre Erkliirung in der unterschiedlichen Yerkarstungsfahigkeit der Hohgantsandstein Lagen die faziell sehr stark wechseln konnen. Die Grenze gegen die Beatushohle zu ist durch die Topographie gegeben (oberfliichliches Einzugsgebiet entweder gegen die Burgfeldtluh oder gegen Oberberg). Die Grenze gegen das Reseau ist etwas unscharf, da in dieser Region der Lithotharnnienkalk als Karstkorper fungiert und sich an der Basis desselben mehrere unterschiedliche Schwinden befinden konnen, die entweder gegen das Reseau oder gegen den Biihlbach zu entwiissern konnen Ab der Linie Biireney Gemmenalphorn Nordtlanke ist das Einzugsgebiet eindeutig das Reseau. Die Quelle Neuhaus, an der Grenze zwischen der alluvialen Bodeli-Aufschiittung und dem Schrattenkalk gelegen, wird aufgrund ihres Chemismus und Schiittungsverhaltens als Karstquelle gedeutet (miindliche Mitteilung der SGHI). Ein Zustrom von Grundwasser ist nicht auszuschliessen. Sicher ist nun jedenfalls, dass die Neuhaus-Quelle nicht mit dem Biitterich verbunden ist; es konnte sich hier um eine lokale Karstquelle handeln die nur Wasser aus dem unmittelbar benachbarten Kalk bringt Als Synthese prasentiert sich die Karte der Einzugsgebiete so, wie auf Figur 4 dargestellt. Figur 4 : Neue hydrologische Ka rte der eingefiirbten Region Bibliographie BrITERLI, T. (1989): Das Karstsystem Sieben Hengste-Hohgant Schrattenfluh, Versuch einer Synthese / Le reseau karstique Sieben Hengste-Hohgant-Schrattenfluh, essai de synthese. Stalactite 38 (1/2), 1988, 10-22 HOF et al (1985) : Le Reseau Hohlenforschung im Gebiet Sieben Hengste-Hohgant 0, 1985 HAUSELMANN Ph. (1997) : Wasserfarbung 1996 Die Organisation. Jahresbericht der SGH Bern 45, 1997 HAUSELMANN, Ph. & JEANNIN, P.-Y. (1997): The karstic Region north of the Lake of Thun Actes du Xlle Congres International de Speleologie, La Chaux-de-Fonds KL0TZLI, U. (1985) : Wasserfiirbung.Jahresbericht der SGH Bern 33, 1985, 47-50 WILDBERGER et al ( 1982): Markierversuche und weitere hydrologische Untersuchungen im Karst des Brienzergrates und des Briinigpasses Akten des 7 Nationalen Kongresses fiir Hohlenforschung, Schwyz 34 Proceed i ngs of the 1 2 Internat i onal Congress of Speleo l ogy 1997, Switzerland Volume 2

PAGE 46

Speleologie, exploitation et protection des eaux souterraines dans les karsts des environs de Chambery (Savoie-lsere, France) Fabien Hoblea CAGEP-URA 903 du CNRS, lnstitut de Geographie, 29 Avenue Robert Schuman, 13621, Aix-en Provence Cedex, France Abstract Several examples, located in Savoie (France), of speleologic actions concerning water management are given : water tracing, exploration and caves surveying etc ... Applied aspects are just one of the multiple aims of such actions, which are also led to progress in the fundamental knowledge of the drainage, to guide speleological exploration and to teach cavers the practice of scientific speleology Cavers proceed in association with karsl scientists and local administration offices That is the condition to obtain good results and to make these actions well agreed Resume Au travers d'exemples savoyards d'actions speleologiques concemant la connaissance et la gestion des aquiferes karstiques (participation a des tra93ges, exploration de cavites et fourniture ou realisation de releves topographiques en vue de captages ou d'etablissement de perimetres de protection ), sont examinees les conditions de realisation garantissant a la fois des resultats exploitables et une bonne acceptation de ces operations tant par les scientifiques que les amenageurs institutionnels, ce qui passe par la concertation et le partenariat entre les divers acteurs, mais aussi ('information et la formation des speleologues 1. Introduction hydrogeologie et speleologie A border les rapports entre !'elude scientifique fondamentale et appliquee des aquiferes karstiques et la speleologie n'est pas un sujet nouveau. Suite a !'evolution des preoccupations en matiere de gestion des ressources en eau, ce theme s'est trouve revigore au debut des annees 90, comme en temoigne le contenu des actes de congres speleologiques nationaux A !'occasion du Congres de Montpellier, organise en 1991 autour d'un symposium sur "les eaux souterraines en milieu calcaire", pas mains de onze communications sur le theme furent ainsi presentees Dev ant la multiplication des operations sur les eaux karstiques impliquant des speleologues, devant ('importance des enjeux, qui depassent le plus souvent le cercle de la speleologie, ii a ete juge necessaire de recenser et definir le role et la place des speleologues dans la connaissance et la gestion des aquiferes karstiques (A VIAS, 1991 ; BAKALOWIC'Z, 1994) A partir d'exemples savoyards repertories sur la figure 1, sont ici presentees quelques actions fondees sur une concertation entre speleologues et specialistes ou amenageurs, dans l'optique de concilier les interets de chacun Les speleologues de la region chamberienne, a !'image de ceux d'autres departements karstiques (SAUVEL, 1991), ont ete ces demieres annees amenes a contribuer a plusieurs operations de captage ou de protection des aquiferes karstiques en Savoie et en Isere, soit dans le cadre de leurs activites propres, soil sur commande. Bon nombre d'entre elles s'inscrivent de plus localement dans le programme de recherche conventionnel entre Agence de l'Eau et structures speleologiques federales, mis en place a la fin des annees 80 (LAURENT, 1991) 2. Tra~ages realises par des speleologues L'operations "Granier 94" En juin 1994 le geosysteme karstique du Mt Granier en Chartreuse, a fait l'objet d'un multitra9age repondant a trois objectifs complementaires : Faire progresser dans le cadre d'une recherche universitaire doctorale, l'etat des connaissances, encore fragmentaires sur !'organisation et le degre de hierarchisation des ecoulements souterrains dans ce massif. Former les speleologues a la technique du multitra9age en couplant un stage scientifique federal (regional) a ('operation ( HOBLEA 1994 ). Ameliorer la connaissance de la vulnerabilite des circulations karstiques d'une part dans le cadre de la collaboration instituee a l'echelon regional entre structures speleologiques et Agence de l'Eau, d'autre part a la demande de l'hydrogeologue departemental agree, dans le cadre du valet hydrogeologique de l'etude d'impacts du projet de ligne TGV Lyon-Turin (tunnel sous le Granier risquant d'affecter des sources soup9onnees d'etre en relation avec le karst) Dans le cadre du stage scientifique les speleologues ont participe a une partie de la preparation du tra9age (analyse du contexte geologique, reconnaissance des sources mise en place du dispositif de surveillance et preparation des traceurs) !Is ont effectue les injections dans l'endokarst en fin de fonte des neiges Suivi de la restitution et analyses des prelevements d'eau, demandant une disponibilite et des moyens specifiques, ont en revanche ete effectues par des universitaires qui ont ensuite fait part des resultats aux stagiaires Des resultats qui interessent aussi bien les explorateurs que les amenageurs : ce multitra9age a en effet permis de montrer l'unicite de ('emergence du massif en basses eaux, contrai rement au schema admis auparavant d un compartimentage du massif en unites hydrogeologiquement independantes de part et d'autre de grands decrochements transverses Les mesures de protection reglementaires du captage de Chapareillan pourront tenir compte de cette donnee nouvelle De plus, les hydrogeologues de l'universite de Savoie, charges de l'etude d'impacts sur le trace TGV, ont pu utiliser ce tra9age par le biais de reinfiltrations plus a l aval de !'emergence dans les calcaires juras s iques sous-j a cents draines vers des s ources sen s ibles ou le traceur est b ien reapparu 6 '" Conferen c e on Ume s t o ne H y drolog y and Fissured Med i a 35

PAGE 47

L'operatioa "Margeriaz 95" Le second exemple recent est le trac;:age de mai 1995, realise par le Comite Departemental de Speleologie de la Savoie dans la Tanne des Biolles (HOB LEA et al 19%). Cette cavite s'ouvre dans la montagne du Margeriaz au coeur du massif subalpin des Bauges. Elle se developpe en suivant vers le nord-est le pendage des calcaires urgoniens qui plongent vers un vat contenant une imjX>rtante zone noyee Le franchissement d'une tremie qui bloquait les explorations a .50 m depuis 1985 a permis d'explorer a partir de 1993 le principal affluent souterrain du collecteur noye de ce massif, suivi jusqu'a une nouvelle tremie a 530 m. Le jX>tentiel depasse les 1000 m de denivelee. La "coloration" est destinee au depart a s'assurer de liaisons internes dans la cavite et de la rapidite du transfert jusqu'a la source vauclusienne du Pissieu, connue comme seule emergence du systeme. Mais la encore, des mobiles non speleologiques viennent renforcer l'utilite de ce trayage : ii se trouve en effet qu'une source captee sur l'autre versant de la montagne connait des problemes de pollution qui coincident avec des epandages de fertilisant (fumier) sur les pistes de ski du Margeriaz en cours de revegetalisation Malgre un contexte geologique peu compatible avec l'idee d une liaison, la DDAS S de la Savoie s'est associee a la surveillance et au dejX>utllement du trac;:age afin de tester l'hypothese d'un lien hydrogeologique entre l'impluvium du Margeriaz et ces sources captees De plus, des travaux d'extraction de granulats dans le lit majeur du chera~, le ~ri~c(pal cours d'eau des Bauges, qui passe a I extrem,te sud du Margeriaz, ont recemment libere des ecoulements (source dite du V i llaret) dont les analyses indiquent une origine karstique S agit-il d'une ancienne exsurgence colmatee par le remblaiement alluvial ? Tant les speleologues que les scientifiques du karst comptent sur ce trac;:age jX>Ur apporter des elements de reponse ... L injection a lieu au plus fort de la fonte mvale a 70 m, tandis qu'une equipe de pointe descendue au petit matin en basses eaux se laisse volontairement pieger dans le karst profond pour suivre le cheminement du colorant et observer le comportement du reseau en crue "limitee" (cycle journalier permettant une remontee en fin de nuit !ors du regel exterieur) Ce trac;:age revele le caractere tres direct et rapide de cet axe de drainage vers !'emergence, branche tres a l'aval de la zone noyee, avec une vitesse ma,cimale de circulation vers le Pissieu de l'ordre de 800 m / h (contre une vitesse de l'ordre de 300 m / h enregistree precedemrnent en crue d'orage estival dans une cavite branchee plus a l'amont sur la zone noyee) De plus, la source du Villaret a delivre une faible quantite de traceur a travers son placage alluvial. ce qui renforce l'hypothese d'une ancienne exsurgen~e (reste a s'assurer que la fluoresceine detectee ne prov1e~t pas d'infiltrations de la riviere de surface a l'aval du P1ss1eu via la terrasse alluviale ... ) Si cette hypothese s'avere la bonne, un decolmatage artificiel par les carrieres de granulats entrainerait un assechement de la pittoresque emergence-cascade du Pissieu but de promenade balisee et site nature! d'interet patrimonial majeur en Bauges .. Enfin, la surveillance des sources polluees sur l'autre versant de la monlagne s est revele negative Tout hydrogeologue salt qu'un tel resultat ne permet aucunement de trancher de maniere peremptoire et definitive. Neanmoins, compte-tenu de la configuration structurale, ii serait sage de rechercher les causes de pollution de ces sources ailleurs que sur l impluvium du systeme du Pissieu ... 3. Exploration, topographie et aide a la localisation de drains karstiques a capter ou proteger Le forage AEP de La Feclaz en 1987 La station de ski de La Feclaz domine la ville de Chambery sur le rebord occidental du massif des Bauges Au milieu des annees 80, un pogramme d'extension de la capacite d'hebergement de la station prevoyait un renforcement significatif de l'approvisionnement en eau par un forage direct dans un drain collecteur appartenant au reseau speleologique reconnu sur plus de 20 km de developpement sous la surface du plateau de La Feclaz (reseau Garde-Cavale) Le projet a pris corps en 1987 II a necessite !'utilisation conjointe des donnees et techniques speleologiques et des methodes de pointe de reperage magnetostatique de cavites developpees a l'epoque par le LET!, laboratoire du Centre d'Etudes Nucleaires de Grenoble Sans entrer dans les details de cette importante operation, largement exposes par ailleurs (FANTOLI, 1990 ; HOBLEA, 1990), soulignons le role tenu par les speleologues, qui ont mis a disposition du ma'itre d'oeuvre les topographies detaillees du reseau souterrain dans lequel devait aboutir le forage II est clair que sans les decouvertes speleologiques et les documents en rendant compte, resultant de plusieurs decennies de patient labeur et d'un nombre d'heures incalculable passees sous terre ou devant la table a dessin (c'etait avant Toporobot), ce forage de precision (cible de 5 m de large a !ravers 200 metres d'epaisseur d'Urgonien) n'aurait guere ete envisageable De plus, des speleologues locaux ont ensuite ete employes par la societe Hydrokarst jX>Ur assurer les travaux souterrains necessaires au positionnement electromagnetique de la cible. puis a la construction d une retenue d'eau souterraine, a plusieurs heures de progression de l'entree du reseau la plus proche Exploration et etude d'une cavite hydrothermale accideateUement ouverte sur le chantier des nouveaux thermes a Aix-les-Bains. Plus recemment, une autre operation lourde d'amenagement en terrain karstique a necessite !'intervention des speleologues savoyards, dans le contexte insolite (du moins en France) des circulations hydro thermales d'Aix-les-Bains Les travaux de terrassement prealables a la construction d'un nouvel etablissement de cure ultramoderne ont demarre des 1996, sur les terrains du pare de la villa Chevalley, accroches au versant d'un anticlinal faille d'urgonien enchasse dans la molasse (cette antiforme represente le contre-coup du chevauchement du front subalpin sur ('avant-pays savoyard) Le carreau de chantier est situe 200 m a l'amont des sources thermales integrees a l'etablissement actuel en centre-ville une trentaine de metres en contre-haut. Comme I 'elude d'impacts en prevoyai t l'eventualite, les travaux de terrassement ont trepane une cavite aveugle, s'ouvrant par une verticale d'une dizaine de metres. inexplorable sans equipement speleo Par souci de securite tant pour la poursuite du chantier que pour la qualite des eaux thermales, les responsables _d~ l'amenagement ont done fail appel au Com1te Departemental de Speleologie pour explorer et topographier la cavi te, baptisee gouff re Che valley Si ce dernier s est avere de taille modeste et inoffensif pour la stabilite du substrat. ii a en revanche 36 Pro c eed i ngs of the 12 th Internat i onal Congress of Speleolog y, 1 997 Sw i tzerland Vo lu me 2

PAGE 48

permis d'atteindre la surface piezometrique de l'aquifere thermal. Vers la cote -27 m, la suite de cette cavite au developpement essentiellement vertical est en effet ennoyee (cheminee d'equilibre) La temperature elevee de l'eau (42C en moyenne) n'a permis qu'une prudente reconnaissance par le speleonaute Jean Bottazzi en decembre 96. Une courte plongee de 4 mn en volume etanche, Jui a permis de progresser d'une quinzaine de metres, avec une visibilite nulle, dans une diaclase plongeante, avant de buter sur un retrecissement qui aurait necessite un decapele, bien risque dans ces conditions . Parallelement a cette tentative de plongee, la surveillance du niveau du plan d'eau dans le gouffre lors d'un pompage a l'occasion du nettoyage annuel de la source thermale captee a l'aval, a permis de verifier le lien direct entre l'eau visible dans la cavite et celle sortant a la source. De plus, des prelevements d'eau realises par les speleos au fond du gouffre ont fait l'objet d'analyses bacteriologiques de premiere importance pour l'etude des contaminations periodiques des eaux thermales constatees a la source. Avant la decouverte du gouffre Chevalley, ii n'etait pas possible de trancher sur l'origine des germes indesirables. Ce regard sur l'amont de la circulation thermale a montre que des foyers de contamination peuvent exister avant la source dans ce type de cheminee d'equilibre Enfin, la reconnaissance de la configuration de la zone noyee par plongee et de son fonctionnement par surveillance du pompage permet d'envisager, sous reserve de trailer le probleme de pollution episodique, un captage direct de l'eau chaude a !'aplomb meme du futur etablissement thermal .. 4 Des actions concer t ees a v ocations multiples Un partenariat entre speleologues, scientifiques et amenageurs. Du cote des speleologues, ces actions ne se sont pas realisees a l'initiatives de pratiquants isoles ou meme de clubs mais au niveau departemental (Comite Departemental de Speleologie de la Savoie), voire au niveau du Comite Speleologique Regional Rhone-alpes et de !'Ecole Fran9aise de Speleologie pour !'organisation du stage scientifique "Granier 94", qui a rassemble autour des speleologuess savoyards une quinzaine de stagiaires de departements voisins. Toutes les actions citees ont fait l'objet d'un encadrement ou d'un soutien scientifique par un ou plusieurs chercheurs en speleo-karstologie et / ou hydrogeologie karstique, mobilisant des membres du CAGEP-URA 903 du CNRS d'Aix-en-Provence (F. Hoblea et Ph. Audra), du Laboratoire de Geologie Structurale et Appliquee de l'Universite de Savoie (G. Nicoud et M. Dzikowsky), du Laboratoire d'Hydrogeologie de l'Universite d'Orleans (M. Lepiller) et de l'Equipe d'Hydrobiologie et d'Ecologie Souterraine de l'Universite Lyon I (M.-J. Turquin et R. Laurent). Chaque operation a fait l'objet de rapports et de notes, ainsi que d'articles et communications associant scientifiques et speleologues, permettant a la communaute scientifique de juger de la validite des methodes comme des resultats (HOBLEA et al., 1996) "' Toutes les operations mentionnees ont aussi associe en partenariat des arnenageurs institutionnels : Agence de l'Eau (LAURENT, 1989), DDASS, DDAF, Societe d'Amenagement de la Savoie, intervenus soit pour le pret du materiel technique (preleveurs automatiques .. ) ou le financement des traceurs, soit pour collecter et utiliser les resultats, transmis dans certains cas aux usagers concernes : collectivites locales (communes, districts, SIVOM etc .. ) ou organismes publics (Thermes Nationaux d'Aix-les-Bains). Ces usagers ont toujours ete les premiers avertis des actions projetees, et ils ont ete consultes sur leur opportunite. Des actions polyvalentes. Autre caracteristique commune aux exemples presentes, la polyvalence des objectifs et des centres d'interet: "' Orientation et progres des investigations speleologiques : si bon nombre de decouvertes concement des galeries "fossiles", deconnectees du drainage actuel, les investigations dans les drains actifs et noyes restent un axe de developpement essentiel de l'exploration speleologique. Amelioration des connaissances sur la structure et le fonctionnement du karst : les recherches hydro-speleologiques peuvent apporter des donnees aux scientifiques du karst, qui etudient !'ensemble du systeme karstique, dont les reseaux souterrains sont une des composantes non negligeable. Les donnees seront exploitables a condition que les recherches soient realisees dans Jes regles de l'art, ce qui suppose aujourd'hui pour les tra9ages de proscrire l'usage des seuls fluocapteurs a charbon actif pour surveiller la restitution, et done de mettre en oeuvre des techniques non directement accessibles aux speleologues. Une telle collaboration entre speleologues et scientifiques peut ainsi debaucher sur des progres decisifs en matiere de karstologie, du mains sur le plan des connaissances des caracteristiques locales de tel ou tel systeme karstique : L'exemple du gouffre Chevalley est a ce propos edifiant, en livrant aux scientifiques une des rares cavites hydrothermales actives fran9aises, a laquelle seuls les speleologues peuvent acceder Notons au passage que, au vu des rapports soulignant l'interet scientifique de cette cavite, les Thermes Nationaux d'Aix-les-Bains ont decide d'en preserver l'integrite et d'en faciliter l'acces comme l'etude. L'exemple du multitra9age du Granier illustre egalement eel aspect : l'etude hydrogeologique des sources du massif avait conclu a la multiplicite de systemes karstiques independants, la ou un multitra9age dans l'endokarst a permis de montrer la hierarchisation du drainage vers une emergence commune. Finalement, l'idee est qu'avec un minimum d'encadrement et de methode, Jes speleologues peuvent contribuer efficacement a l'etude des circulations et du milieu karstiques. Cette efficacite depend aussi grandement de l'effort de formation et d'information que les scientifiques sont prets a fournir aupres des speleologues interesses Formation des speleologues aux approches scientifiques du karst : par le biais de stages scientifiques orientes vers l'hydrogeologie karstique en general et les tra9ages en particulier, a !'image du stage greffe sur )'operation "Granier 94", ii ne s'agit pas de vouloir faire des speleos des chercheurs de pointe au fail des derniers developpements d'une science que seul un cursus academique complet permet de ma1triser, mais plutot de veritables "equipiers scientifiques", sur le modele des equipiers secours. Ces demiers, sans etre des secouristes professionnels, n'en maltrisent en effet pas moins les bases necessaires a !'insertion dans une operation de secours souterrain, grace a une formation adaptee et continue. 6 '" Conference on Umestone Hydrology and Fissured Media 37

PAGE 49

Pour une formation en speleologie scientifique, nous pensons qu'il est preferable, a !'image de la plupart des stages de terrain universitaires. de travailler sur des operations de recherche effectives, afin d'eviter de perturber inutilement le milieu par des exercices sans application (on imagine maJ se mettre dans chaque region en quete de "karsts-ecole" a "sacrifier" pour la seule pedagogie) Comme les stages de formation au secours souterrain, la formation d'equipiers scientifiques serait organisee par les structures speleologiques federaJes, et encadree dans ce cas par des scientifiques du karst patentes, federes OU non ., Utilisation des donnees pour l'amenagement, ['exploitation et la protection de l'aquifere : tous les exemples cites ont debouche sur une insertion des resultats dans des etudes en cours : determination de perimetres de protection (Granier), localisation de point de forage (La Feclaz), protection active de l'aquilere (thermes d'Aix). A l'interet speleologique se joignent ainsi a chaque fois un interet scientifique, pedagogique et utilitaire en termes de gestion des ressources en eau Des actions cependant limitees L'exemple savoyard tend a montrer que des actions bien cadrees et bien encadrees, effectuees dans le karst profond penetrable, peuvent satisfaire plusieurs objectifs, et mobiliser divers partenaires en permettant un groupement d'interets. II ne faudrait toutefois pas en deduire que ce type d'actions peut s'appliquer dans toutes les situations et constitue la principale approche des ressources endokarstiques. Ne serait-ce que sur le plan methodologique, des limites apparaissent concernant l'idee meme d'utilisation de l'endokarst comme point d'injection des traceurs De meme qu'il est illusoire de fonder la modelisation du fonctionnement des hydrosystemes karstiques sur la seule etude des ecoulements observes dans les reseaux speleologiques (MANGIN & BAKALOWICZ, 1989), ii est pour le moins delicat de vouloir extrapoler d'un trayage realise directement dans des drains collecteurs du karst profond les modalites de propagation d'un polluant qui a en realite toutes les chances d'affecter l'aquifere karstique depuis la surface Les simulations de propagation des polluants doivent done etre effectuees dans la mesure du possible depuis les points de diffusion potentiels en surface, sous peine de pertes d'informations capitales Cette regle elementaire pour l'hydrogeologue merite d'etre rappelee aux speleologues. Conclusion Les actions savoyardes ont ete menees dans le souci de produire un resultat rigoureux, en s'entourant des collaborations necessaires, tant sur le plan scientifique qu'administratif. Ces operations s'inscrivent ainsi en complement, et non en concurrence, des travaux hydrogeologiques menes depuis la surface Cette demarche doit permettre aux representants decentralises de la Federation Fran9aise de Speleologie de se positionner comme acteurs et interlocuteurs serieux dans la gestion des ressources et du milieu karstique, et d'etre associes, comme cela se fait en Savoie, aux procedures d'amenagement local. notamment lorsqu'il est question de reglementer l'acces a l'endokarst, dans le cadre par exemple de mesures de protection des aquiferes Les donnees manquent a ce propos quant a !'impact reel de la frequentation speleologique sur la qua!ite des eaux souterraines, sujet d'etude qui reste a defricher par les hydrogeologues ... References AVIAS, J. 1991. De !'importance de la speleologie dans l'amenagement et la valorisation des ressources en eau des terrains karstiques Les eaux souterraines en milieu calcaire. Actes X!Xeme Congres National de Speleologie. Spelunca Memoires 19 : 135-144 BAKALOWICZ. M. 1994 La place des speleologues dans la connaissance du karst. Journees Norbert Casteret, XXeme Congres FFS, 1992. Spelunca Memoires 21 : 40-42 FANTOLI, J.L. 1990. Campagne de recherche magnetostatique par le LET! sur la localisation depuis la surface d'un repere en cavite karstique. XlXeme Congres regional Rhone-Alpes de Speleologie, 6 p HOBLEA, F. 1990 Les amenagements touristiques sur karsts d'altitude : problemes et impacts. Mem D.E.A. Geographie, Univ Lumiere-Lyon 2, 234 p HOBLEA, F. 1994 Le stage scientifique regional "Granier 94" Actes Quatrieme Rencontre d'Octobre SCParis : 69 81. HOBLEA, F. DooEUN, Ch LASSERRE, D BOTTAZZI, J. MANIEZ, P. 1996 La tanne des Biolles : un axe de drainage majeur sous le massif du Margeriaz (Bauges, Savoie, France). Karsto/ogia 27 : 41-56 LAURENT, R. 1989 La Federation Fran9aise de Speleologie, interlocuteur privilegie des agences de l'eau Spelunca 35 : 85-86 LAURENT, R. 1991. Travaux et reponses que peuvent apporter Jes speleologues aux collectivites locales. Les eaux souterraines en milieu calcaire Actes XIXeme Congres National de Speleologie Spe/unca Memoires 19 : 153 160 MANGIN, A. & BAKALOWICZ, M.1989 Orientations de la recherche scientifique sur le milieu karstique InOuences et aspects perceptibles en matiere de protection Spe/unca 35 : 71-79 SAUVE.., C. 1991. Quelques exemples de collaboration entre hydrogeologues et speleologues dans le departement de l'Herault. Les eaux souterraines en milieu calcaire Actes XIXeme Congres National de Speleologie Spe/unca Memoires 19 : 149-152 Bauges N .... 0 lirnite du -limite des karsts Chambery etudies fGnlnier )I tra~age 0 localisation cible de forage pour captage 0 reconnaissance aquifere Figure 1. Quelques operations speleologiques recentes concemant la protection et la mise en va/eur des aquiferes karsti.ques en Savoie (1987-1997). 38 Proceedings of the 12 1 h International Congress of Speleolog y, 19 97 Switzerland Volume 2

PAGE 50

Delimitation du bassin versant karstique de la Bienne (Haut-Jura, France) par la geologie et les essais de tracage par Robert Le Pennec Association Speleologique de Saint-Claude 11 rue du Bel v edere F 3 9200 SAINT-CLAUDE Abstract: The Bienne is a left bank tributary of the river Ain whose upper topographical watershed i s including a significant number of closed basins and watercourses or lakes leakages Since the last 30 years a 72 tracing tests have been performed and enabled us to appreciate the actual size of the karst area which is feeding the upper valley of the torrent. While replacing these tracing tests in the structural framework of the folded Upper Jura is is possible to divide this area into 14 hydrogeological units The largest of them the Grandvaux synclinorium, is more than 30 km long and is including the longest underground flow in the Franche-Comte originating in the Chaumusse sinkhole and the Enrage spring Resume: La Bienne est un affluent de la rive gauche de l'Ain dont le haut bassin versant topographique comporte de nombreux bassins fennes et des pertes de cours d'eau ou d'ernissaires de lacs Depuis plus de 30 ans 72 essais de trac;age ont perrnis d'apprecier l etendue reelle de la zone karstique alimentant la haute vallee du torrent. Replaces dans le contexte structural du Haut-Jura plisse ces tra9ages pennettent une subdivision en 14 unites h y drogeologiques La plus grande d'entre-elles le synclinorium du Grandvaux fait plus de 30 km de long et inclut la plus longue circulation souterraine de Franche-Comte entre la perte de la Chaumusse et la source de l'Enrage 1. Bassin versant topographique Vers le Nord (fig 1) le bassin de la Bienne est separe de celui de l'Orbe par le village des Rousses et le mont Risoux. Vers l'Ouest le secteur des Crozets est draine de maniere aerienne vers la Bienne mais le bassin du Grandvaux est ferme autour du lac de l'Abbaye Le ruisseau d'Heria se jette dans la Bienne a Jeurre mais le lac d'Antre est situe dans un bassin fenne La lirnite ouest, avec le bassin de l'Ain, est done difficile a percevoir Vers l'Est, Jes reliefs des Molunes et Bellecombe separent la Bienne de la Valserine Au Sud la lirnite topographique passe entre Oypnnax et Viry Enclavee dans ce secteur la depression du lac de Lamoura n'a pas d'exutoire aerien Ce grand secteur ains i defini presente done un nombre important de zones problematiques 2. Structure Le Haut-Jura (fig l et 2) est une succession de plis complexes d'orientation generale NE-SW Les anticlinaux de Jurassique superieur ou moyen chevauchent generalement vers l'Ouest des synclinaux a creur cretace et tertiaire souvent empates de depots glaciaires L'ensemble des plis est tron9onne par deux grands accidents transverses : celui de More z et celui de Saint-Claude et les plis sont affectes transversalement par deux abaissements axiaux : celui de More z qui fait chuter Jes axes de 200 m et celui du Flumen (de 400 m) La haute vallee de la Bienne traverse en cluse les reliefs du Haut-Jura : cluses de Morez de Vaux-Jes-Saint-Claude de Jeurre et de Dortan-Lavancia mais elle suit un axe synclinal entre Morez et Saint-Claude. Les series calcaires atteignent 1200 m d'epaisseur et la Bienne Jes entaille sur 800 m, dans des gorges sauvages. 3. Tra~ages et determination des limites Ensemble du bassin : Les sources karstiques sont situ e es au sommet des differents niveaux mameux les plus importantes etant a la base du Sequanien Plusieurs sources vauclusiennes ont un conduit terminal de plus de 50 m de profondeur : Bief Noir 85 m ; Doye Gabel 77 m ; Bief Goudard 60 m ; Brive 56 m D'autres emergences peuvent etre masquees sous les depots glaciaires surtout vers Lavancia ou le reseau des Bracelettes est tres peu connu Les tra9ages effectues dans le Haut-Jura (fig I) montrent une tendance generale au drainage dans l'axe des plis : le secteur des Lacs de Bellefontaine et des Mortes resurge a la source de la Doye Gabet sur l'abaissement axial de Morez la 6 t h Conference on Umestone H ydrology and Fissured Media 39

PAGE 51

reg1on comprise entre le secteur de Lamoura et la Pesse sur les emergences de celui du Flumen (MUDRY & ROSENTHAL 1977) A partir des differents essais de tre,;:age des cours d eau principaux et de la structure (sommets d anticlinaux dans le sens longitudinal zones d elevation axiale des plis dans le sens transversal) on peut subdiviser l e Haut-Jura en quatorze unites a drainage karstique (fig 1) : PRE.NOVEL LA AIN Le Lange A-A : Trace des coupes de la figure 2 Fig. 1: Les 14 zones hydrogeologiques du Haut-Jura et le bassin versant de la Bienne. 40 Proceedings of the 12 1 h I nternational Cong r ess of Speleology 1997 Switzerland Volume 2

PAGE 52

I Larnoura la Pesse ill Grandvaux V Risoux lac des Rousses VII Vulvoz IX : foret d' Avignon XI : Choux XIII : Montcusel Chancia Synclioorium du Grandvaux : II IV VI VIII X XII XIV : Longchaumois Cinquetral Bellefontaine Marez Foret Mt-Noir Jouvencelles Longchaumois la Mouille cret du Surmontant Ranchette Mt Chabot Viry Oyonnax Jeurre lac d' Antre Veyziat Uffel Les quatre zones majeures de drainage karstique du Haut-Jura sont le systeme de la Doye Gabet a Marez le secteur de Larnoura-la Pesse le systeme d'Oyonnax -Viry et le Grandvaux. Le synclinorium du Grandvaux est une zone relativement plane de 900 a 1000 m d'altitude. Les sources situees dans la vallee de la Bienne s'etagent entre 510 et 350 m Un premier long tra9age effectue depuis la perte du lac de l'Abbaye (FRACHON, 1965; MUGNIER & CHATELAIN 1969) est reapparu a la source de !'Enrage. La source de !'Enrage qui est situee dans la cour de l'usine Breuil a Chassal, a comme trop-plein la grotte de I 'Enrage. En 1994 30 kg de fluorescei'ne ont ete injectes au gouffre de la Tane a la Chaumusse (LE PENNEC, 1987 a et b) dans le but de delimiter )'unite en direction de !'accident de Marez 25 jours plus tard, le traceur atteignait la resurgence de Brive (27 km) et le lendemain la source de l'Enrage (distance : 29,5 km, vitesse apparente : 47,2 m/h) L'eau a ete tracee pendant 3 jours, mais non visible a l'oeil nu Cette relation souterraine, la plus tongue de Franche-Comte, a montre une faible vitesse de circulation comparativement aux autres tra9ages du Haut-Jura qui ont circule a 100 m/h en moyenne Le Grandvaux est un synclinorium large au niveau de Saint-Laurent, ou sont situees les pertes de (fig. 2, coupe AA). II se retrecit au niveau de Lescheres (fig 2, coupe DD) ou il doit etre chevauche par I 'anticlinal du bois de Cuttura La zone de sortie se situe done dans le coeur d un anticlinal pour !'Enrage (fig. 2 coupe EE), sur le flanc ESaint pour Brive, sur la rive gauche de la Bienne, a la faveur des ecailles du Trunet. Cette unite hydrogeologique va done de !'accident de Marez au Nord a celui de Vuache-Molinges au Sud sans qu on puisse savoir si ces failles jouent un role de lirnite. 4. Conclusion L'utilisation combinee du tra9age arti:ficiel et de l'etude structurale du Haut-Jura perrnet la delimitation de 14 unites hydrogeologiques dans le bassin versant de la Bienne Le drainage karstique se fait principalement dans la direction d allongement des plis, polarise par Jes grandes structures transversales : accident et zone d'abaissement axial de Morez accident de Saint-Claude et zone d abaissement axial du Flumen, accident de Vuache-Molinges Le synclinorium du Grandvaux apparait ainsi comme une grande structure triangulaire de plus de 30 km de longueur, de 100 km 2 de surface, a drainage exclusifvers la vallee de la Bienne References ALABOUVETTE B 1965 : Etude geologique de la region de Saint-Laurent (Jura). These 3 cycle Geologie Besan9on 117 p ., 14 pi HT. AUBERT D 1965 : Calotte glaciaire et morphologie jurassienne Eel. Geol. Helv. vol. 58 1, 555-578, 9 fig. BADIN L. & MAYA C 1971 : Etude des circulations souterraines par fluorimetrie SRAE Franche-Comte 35 p ., 10 fig. inedit. B.RG.M. Cartes geologiques l/ 50 000 : Moirans-en-Montagne Marez, Saint-Claude. CHAUVE P. DUBREUCQ F., FRACHON J.C GAUTHIER A METTETAL J.P & PEGUENET J. 1987 : Atlas des circulations souterraines reconnues par tra9age en Franche-Comte 35 p ., 9 fig ., 17 pl. Ann Scientif. Univ Besan9on Geologie Mem. H.S. n 2 FRACHON J C. 1965 : Le reseau de !'Enrage (Jura fran9ais) : une percee hydrogeologique de 21 km Le Boueux bull. trimest. de la Section de Geneve de la Soc Suisse de Speleol. III 4, 5-10. I.G N Cartes topographiques : 1/ 100 000 Annecy-Lausanne; 1/ 25 000 : Marez et Saint-Claude. LE PENNEC R 1987 : Inventaire des colorations. Assoc Speleol. de Saint-Claude, Bull. Special n 3 74 p LE PENNEC R. 1987 : Inventaire des colorations du Pare Nature! Regional du Haut-Jura. Actes 8 Congres Nat. Soc Suisse de Speleol., 125-134. MUDRY J. & ROSENTHAL P 1977 : La Haute-Chaine du Jura entre Marez, Saint-Claude et la Pesse These 3 cycle Geologie Besan9on 205 p ., 95 fig MUGNIER C & CHATELAIN D 1969 : Les recherches hydrogeologiques sur la percee karstique lac de l'Abbaye resurgences de I Enrage (21 km) et la recente coloration du Speleo-Club Salinois (Jura) Ann Speleol. CNRS t. 24 2, 299-316 2 fig. 6 th Conference on Limestone Hydro logy and Fissured Med i a 41

PAGE 53

des Bez See de La Doye 30 KM St-Laurent Tres SOURCE DE L'ENRAGE .IJ~~ ~ 1 ,,, .,a .... ,., E:2:l 1 s. ~ ,"."' f:J:) Oh1<"9 I' ~ : A'1"' .. .,,-.. .... ,..,,;, 8 ,e.11-w-.,, c=J u~$"' ~ a./A .. hj,eiu, IT:I3 : a.,r&,_,-.,.. 1110.-., a 11 .,,.,, wu rill ..,.,.,.,,_ ,.,.,,, E:3 ~,l,.,k, ... ,. m:]] ,,.,,11 EI:l3 ,c,-,_., .. ,., .. ,,_ ,~s ~ Les Baptaillards D oy e Gabel Les Frasses Pontoise Bief Goudard Noirecombe Marais de Grand Essa rt :J RESURGE 1 CE DE BRIVE COUPE: B ALABOUVETTE 1965 ATELIER DE L'ENVIRONNEMENT 1994 POINT D INJECTION TRAJET DU COLORANT Fig. 2 : Coupes geologiques et trafages dans le synclinorium du Grandvaux (Unite Ill) 42 Proceedings of the 12 '" Interna t ional Congress of Speleology, 1997 Switzerland Volume 2

PAGE 54

Controle automatique de plusieurs parametres dans la Grotte des Merveilles (Aracena, SW de l'Espagne) W. Martin-Rosales, M. L6pez-Chicano, A. Pulido-Bosch et A. Valltios Groupe de Recherche "Ressources Hydriques et Geologie de I 'Environnement". Universite de Grenade. Espagne Abstract Within Cave of the Marvels, the arr humidity, temperature and variations in the piezometric level of an interior lake were continually monitored; these parameters were all related to the number of visitors to the cave, a tourist attraction, and their influence on the environment. The continual measurements of CO2 levels were halted after a few days because of problems with the sensors. A comparison of the results obtained from a chamber that is relatively distant from the tourist route with those from visited areas provides an estimate of the influence of visitors and of the illumination, which emits both heat and light, thus stimulating the development of algae and other micro organisms. The existing adverse conditions of humidity, CO 2 and temperature have led to the alteration of numerous speleothemes. An electron microscope study has shown the alterations produced in the crystals. Resume La Grotte des Merveilles fait l'objet d'un suivi continue! de l'humidite de !'air, de la temperature de !'air, de !'evolution du niveau piezometrique mesure dans un lac interieur, le tout en rapport avec le nombre de visiteurs a la grotte amenagee au tourisme et son influence sur l'environnement. Les mesures continuelles de CO 2 ont d0 etre arretees apres quelques jours etant donne les problemes de l'appareillage. L'humidite, la temperature et les precipitations sont aussi mesurees a l'exterieur de la cavite. La comparaison des resultats obtenus dans une salle eloignee du parcours touristique et ceux obtenus dans les lieux visites permet d'estimer !'influence des visiteurs, ainsi que du systeme d'eclairage qui apporte de la chaleur et de la lumiere, favorisant le developpement d'algues et d'autres microorganismes. Les conditions adverses atteintes dans l'humidite, le CO 2 et la temperature ont favorise !'alteration de plusieurs speleothemes, dont le contraste peut etre fait avec les secteurs non visites qui restent inalteres Une etude par microscope electronique a permis de verifier les alterations induites dans les cristaux Introduction La Grotte des Merveilles est une des nombreuses cavites karstiques espagnoles amenagees pour le tourisme. Elle est situee dans le village d 'Aracena, au nord-est de la province de Huelva et a 70 km de la frontiere hispano-portugaise (figure 1) Bien que sa decouverte date du siecle dernier, elle ne fut ouverte au public qu'en 1911; plus de 160.000 personnes par an la visitent actuellement. Creusee dans un petit affleurement de marbres du Cambrien inferieur sur lequel s'est etabli le noyau urbain d'Aracena, elle a une longueur totale connue de 2130 m; elle est constituee essentiellement par trois niveaux superposes; les deux inferieurs sont amenages pour le parcours touristique, d'une longueur d'environ 1.000 m; le troisieme -plus eleveest inaccessible au public, et montre une beaute spectaculaire de grande variete et richesse de concretions. Le developpement probable en profondeur n 'est pas connu etant donne, entre autres raisons, que la cavite est inondee dans son niveau plus inferieur 0 2 a: 0 I Q.. I Ayamonte \ N + Penarroya Cazalla Figure 1.Situation glographique d Aracena, ou se trouve la Grotte des Merveilles. L 'exploitation intensive du complexe souterrain a des fins touristiques a eu des repercussions negatives sur l'environnement nature), similaires a celles decrites dans d'autres cavites du monde (FERNANDEZ et al., 1986; ANDRJEUX, 1988 ; CIGNA, 1993; MANGIN & D"HULST, 1995) Les autorites municipales, conscientes des problemes d'environnement, demanderent en 1993 la realisation d'une etude detaillee pour !'identification des principaux impacts, le degre d'alteration actuelle et !'elaboration d'une serie de mesures protectrices et correctrices face a ces affections. Les premiers resultats de cette etude ont revele !'existence de secteurs ou !'alteration des concretions devenait plus importante (MARTIN ROSALFS et al 1994), coincidant avec des zones de dimensions reduites et ventilation deficiente; il a aussi detecte une grande abondance d algues, liee a un reseau dense de projecteurs de lumiere artificielle a grande energie calorifique. 6 th Conference on Umestone Hydrology and Fissured Media 43

PAGE 55

Salle des coquilles Galeries S11perie111es (non visitees) Enregistreur de temperature fr:I Enregistreur de CO, n Enregistreur du ;;J et humidite relative l..l!J LI niveau d'eau Eau Figure 2.SchJma de la cavill et du rlseau d observatwn de parametres. Base topographique llaborle par le Grupo de Exploraclones Subtemfneas (GES) de la Sociedad Excurslonlsta de Malaga (SEM). Le reseau d' observation Le reseau de controle et suivi de parametres du milieu comprenait a l 'o rigine cinq senseurs pour mesurer la temperature et l humidite de !'air a un intervalle de 15 minutes Plus tard, furent enregistres le contenu en CO 2 de 'air et !'evolution du niveau piezometrique dans Jes lacs au moyen de deux data logger (figure 2) Les valeurs e levees d humidit e relative qui se presentent hab it uellement dans la plus grande partie de la cavite presque 100 %rendirent inutiles Jes equipements d 'e nregistrement du CO 2 atmospherique peu de mois apres leur mise en fonctionnement. Verrerie de Dieu 20 5 ------------~3000 E 20 ., 19 -ii 18 5 c.. E 18 17 5 1 4 7 10 13 18 19 22 25 28 Periode du 01/08/94 au 31/08/94 2500 5 f 1500 .. l -8 1000 E 0 z 20 5 ~----------,3000 2500 ? 1500 .. : ) 1000 E 500 Galeries Superieures 2: 18 15 5 1 4 7 10 13 16 19 22 25 28 31 Periode du 01/08/94 au 31/08/94 18 5 ~----------, 2: 18 17 5 .. "C i 17 -41 16 5 c.. E 16 15 5 1 4 7 10 13 18 19 22 25 28 31 Priode du 01/08195 au 31/08/95 Figura 3.Enreglstrements continus de templrature de l'air et entrle de visileurs dans la Ve"erle de Dieu et les Galerles Suplrleures, co"espondants au mots d aotlt 1994 et aotlt 1995. 44 Proceed ings of the 12 "' International Congress of Speleology 1997, Switzerland Volume 2

PAGE 56

Parallelement ont ete enregistres de far;on automatique la temperature et l'humidite relative de !'air, ainsi que les precipitations tout cela a l'exterieur de la cavite De plus, on a realise de nombreux profils avec des instruments portatifs tout au long de la grotte, dans lesquels on a mesure l'humidite relative, la temperature et la vitesse de !'air Etant donne qu'il etait necessaire de comparer les parametres mesures avec ceux des secteurs non influences par l'homme, on a installe des appareils au troisieme niveau de karstification, bien eloigne du parcours touristique L 'air est l'un des milieux ou les affections ont lieu de fa~on plus immediate apres l'entree des visiteurs, comme le montre la figure 3. On y presente Jes deux enregistrements continus de temperature, correspondant aux Galeries Superieures -non visiteeset a la salle appelee Verrerie de Dieu Cette derniere salle presente de nombreuses concretions parietales, au toit et au sol, formees par des cristaux d'aragonite, excentriques, gours et formes globulaires ; dans cette salle, la temperature peut augmenter de pres d'un degre apres l'entree d'un groupe de visiteurs L'affection la plus remarquable se produit au mois d'ao0t, le nombre de visiteurs etant plus important, ainsi que le contraste thermique interieur-exterieur Cet effet thermique est, dans une certlline mesure, cumulatif pour de courtes periodes, en fonction de la distance a l'entree de la Grotte; on a egalement constate l'effet thermo regulateur des lacs, dont les eaux attenuent la fluctuation therrnique a leurs alentours (PULIDO-BOSCH et al ., 1997) L 'augmentation de la temperatures 'accompagned 'une diminution de l 'humidite du milieu qui peut etre reduite de 3 % Le systeme actuel d'eclairage a un effet identique a celui des visiteurs, puisqu 'ii provoque I 'augmentation de la temperature en raison de la grande energie libereeet la diminution de l'humidite relative e Q. .9: 0 (.) C G) C 0 :::, !!! 'E :s C 0 <..) 4000 2000 \J\i\J 0 0 19.5 e... e ::J 19 iii 18.5 t F 60 40 20 8. ::J e CII '----' -.&...-!---MULIIJL.....L.-_.. ...... -'----~ 0 00 : 00 17 julllet 00 : 00 18 julllet 00:00 19Julllet 00 : 00 20julllet Figure 4.Concentration en CO, de l'air, temperature de l'air et nombre de visileurs par groupe dans la Ve"erie de Dieu. La concentration en CO 2 est habituellement de far;on naturelle, plus elevee dans les cavites qu'a l'exterieur (EK & GEWELT, 1985), etant donne la ventilation plus faible et l'apport de CO 2 a partir du sol et de la zone radiculaire; on a pu constater cette affirmation dans cette cavite, ou la concentration proche a l'entree est de 600 ppm (zone la plus ventilee), et atteint 1500 ppm dans Jes secteurs plus eloignes Le passage des visiteurs fait monter les concentrations a 5000 ppm (figure 4) On constate en partie une origine biologique non humaine, due probablement a la respiration vegetale, pour quelques pointes de C~ enregistrees durant la nuit, apres la fermeture de la cavite au public. Une augmentation aussi considerable de la pression partielle de C~ dans la cavite a un impact sur le milieu, etant donne qu'il s'agit d 'une atmosphere beaucoup plus agressive, en ce qui concerne la roche encaissante et done Jes speleothemes. L 'attraction principale de la Galerie des Lacs est !'existence de grandes masses d'eau qui ont ete ornementees de nombreuses lampes submergees Le niveau atteint par l'eau coincide avec le niveau phreatique de l'aquifere lie a l'affleurement carbonate ou se developpe la Grotte; cet aquifere est alimente essentiellement par l'eau de pluie, la precipitation moyenne etant d'environ 900 mm/an (MARTIN-ROSALES a al., 1994). L'extreme secheresse de la periode 1993/1995 et !'existence d'un forage proche qui pompe 7 I.s 1 oat contribue a faire baisser sensiblement le niveau de l'eau, meme au-dessous de la position des lampes Les concretions manifestent des alterations visibles etroitement liees aux impacts deja indiques En effet, le changement des parametres du milieu -augmentation de CO 2 et temperature, diminution de l'humidite, grande quantile d'energie lumineuseet le transport vers l'interieur de la cavite de spores et de microorganismes dans les chaussures et les vetements des visiteurs ont provoque la proliferation d 'algues, lichens et champignons, specialement abondants aux alentours des points d'eclairage. Ces microorganismes alterent les speleothemes de far;on aussi bien mecanique que chimique, comme le suggere VILES ( 1987) L 'elude comparee de 25 echantillons par Microscope Electronique de Balayage ("Scanner") a permis de detecter !'existence de corrosions multiples dans les cristaux qui, si les mesures adequates ne sont pas prises, pourraient deteriorer de maniere irreversible de grandes etendues de la cavite (PULIDO BOOCH etal., 1997) Considerations finales Les effets derives des visites touristiques a la cavite et de !'infrastructure mise au point dans ce but, sont visibles dans !'air -augmentation de la temperature de !'air et du contenu en CO 2 et diminution de l'humidite du milieu-, dans l'eau -baisse du niveau piezometrique-, et dans la roche Dans celle-ci, les impacts se manifestent par le developpement d 'algues, champignons et lichens qui trouvent un milieu favorable d0 a l'energie calorifique foumie par Jes systemes d'eclairage Suite a l'etude realisee on a propose differentes mesures correctrices en vue de reduire la degradation de la cavite (MARTIN-ROSALES el al., 1996); parmi ces mesures on peut souligner la diminution du nombre de visiteurs par groupe et de la frequence d'entree de ceux-ci, le remplacement progressif du systeme classique d'eclairage par un autre constitue par Jes "lumieres froides", plus co0teux mais de meilleur rendemeot et de moindre impact; la paralysation des pompages dans le forage proche de la cavite; traitement des niveaux plus colonises par la microflore avec une solution desinfectante non agressive a la roche; et le maintien d'un reseau d'observation de parametres de l'environnement, et de suivi de la recuperation du milieu 6 th Conference on Limestone Hydro logy and Assured Med i a 45

PAGE 57

Remerciements Nous remercions la Mairie d'Aracena, pour les facilitees donnees pour la realisation des etudes ainsi que les geologues Isabel G6mez, Rosario Herrera y Antonio Arriaza; A la CICYT, projet AMB95-0493 Blbliographie ANDRIEUX C 1988 Influence de l'homme sur l environnement climatique souterrain A c tes des Journees Felix Trombe Moulis Vol. I : 96-12 2. CIGNA A 1993 Environmental management of tourist caves : the examples of Grotta di Castellana and Grotta Grande del Vento Italy Env Geol., 21 : 173 180 EK C & GEWELT M 1985 Carbon dioxide in cave atmospheres New results in Belgium and comparison with some others countries Earth surface processes and landforms, 10 : 173 187 FERNANDEZ P L., GUTIERREZ I. QUIND6S, L S ., SOTO, I. & VILLAR E 1986 Natural ventilation of the Paintings Room in the Altamira Cave Nature, 321 : 586-588 MANGIN, A & D'HULST, D 1995 Frequentation des grottes touristiques et conservation Methode d approche pour en etudierles effets et proposer une reglementation Simposio Jnternazionale Grotte Turistiche e Monitoraggio Ambientale, Italia : 117-145 MARTfN ROSALES W ., RODRfGUEZ NAVARRO, C.M ., L6PEZ CHICANO, M PULIDO BOSCH, A & VALLEJOS A. 1994 Analisis de la influencia antr6pica sobre la Gruta de las Mara villas de Aracena (Espaiia) In ( J R Fagundo A Pulido Bosch J E Rodriguez Eds ). El Karst y los acufferos karsticos Ejemplos y metodos de estudio Cuba : 279-290 MARTfN ROSALES W PULIDO BOSCH A ., L6PEZ CHICANO, M ., RODRfGUEZ NAVARRO, C M & VALLEJOS A 1996 Principales impactos antr6picos en la Gruta de las Maravillas ( Aracena, Huelva) Geogaceta, 20 ( 5 ): 116 '. :?-1164 PULIDO BOSCH, A ., MARTfN ROSALES, W ., L6PEZ CHICANO, M ., RODRfGUEZ NAVARRO, C M & VALLEJOS A 1997 Human ~pacts in a tourist karstic cave (Aracena, Spain ). Env Geol (sous presse) VILES, H 1987 A quantitative scanning electron microscope study of evidenc e for lichen weathering of limestone Mendip Hills Somerset. Earth surface processes and landforms 12 : 467 473 46 P roceedings of the 12 "' International Cong r ess o f S pe leology, 1997, Switzerland Volume 2

PAGE 58

An inexpensive flow-through field fluorometer by Pierre An d re Sch n egg a n d N a t h alie Doerfliger 2 Groupe de geomagnetisme de l'Universite, Rue de l'Observatoire 58, CH-2000 Neuchatel, Suisse 2 Centre d'hydrogeologie de l'Universite, Rue Emile-Argand 11, CH-2007 Neuchatel Suisse Abstrac t In order to simplify sampling dye tracers in water streams without having to invest money in expensive systems we developed a highly sensitive flow-through field fluorometer. Thjs apparatus was tested successfully in the field with fluorescein and sulforhodamine. A submersible probe is connected to a digital data logger and the data is written to a PCMCIA memory card allowing 2 weeks of unattended recording at I sample every 4 minutes A sensitivity as low as 0.05 ppb has been achieved. Simultaneous concentration measurements of 2 tracers is possible, at the expense of reduced sensitivity, however. Resume Dans le but de faciliter et d'automatiser la mesure in situ de la concentration de divers traceurs fluorescents, nous avons developpe un fluorimetre de terrain dont la sensibilite avoisine celle d'un appareil de laboratoire Ce fluorimetre a ete teste avec succes lors de tra9ages a l uranine et a la sulforhodamine. La sonde submersible est reliee a un bo1tier d acquisition et les donnees sont ecrites sur une carte PCMCIA permettant deux semaines d enregistrement a raison d une mesure toutes les 4 minutes. Le niveau de sensibilite est de 0 05 ppb. La separation de 2 traceurs simultanes est possible, moyennant une dimjnution de la sensibilite. 1 Introduction The design of a new flow-through field fluorometer, able to be put in a stream or a spring to measure small dye concentrations was initiated at the request of researchers at our university Similar equipment is already commercially available but at a cost prohibitive Lo academic grants. We designed an automatic, inexpensive device, with the idea of replacing a mechanical sampling system. The advantage of our system is that it replaces the sampler in most situations, provided there is no need to keep the water samples for further analyses. Although it was initially designed for uranine (fluorescein) dye tracer, we also tested it with sulforhodamine, and with a mixture of both. 2. Description of the fluorometer The apparatus is made up of two parts a data acquisition box and a waterproof metallic cylinder which contains the optical system A 15 metre cable, with waterproof connectors links the 2 parts. Set-up of the fluorometer is possible in most of the conditions found at the water s edge (Photo I ) The Ou o r ometer The optical system used for the fluorescence excitation and detection is a standard one : The light emitted by a lamp is filtered to select only the wavelenghts corresponding to the excitation band of the dye The light is then focused on a transparent tube through the middle of which the water flows freely (Fig. I) The natural waterstream is sufficient to insure an almost instantaneous movement of the flujd across the optical system with whatever orientation the fluorometer has relative to the stream The tube is bent a t both ends to prevent daylight from entering The fluorescence from the dye is collected at a lens on the photodetector. Excitation light residuals which always occur in spite of different excitation and emission axes are removed by another filter. The photodetector current is amplified within the sensor box The output signal is fed to the datalogger through the cable Tests done with calibrated concentrations have a linearity and reproducibility that are close to l %. ,... @] = = = D = D I[ Q [ U = = D = = = @] ,... Fig. 1 Fluorometer optical system with 2 perpendicu l ar excitation a11d detection axes. Water flows perpendicularly in the central tube. Practical experience indicates that it is often necessary to measure the water turbidity, since it can considerably alter the measurement in the following ways: I) by reducing the excitation light and fluorescence through absorption by suspended particles 2) scattering of excitation light toward the detector. To measure the turbidity, we mounted a second detection axis in which the emission filter was replaced by an excitation filter. In clean water even in the 5 h Conference on Umestone Hydrology and F i ssured Med i a 47

PAGE 59

presence of dye, the signal of this axis is very small. It corresponds to natural scattering of water. The presence of suspended particles scatters light in all directions, giving rise to a signal proportional to the turbidity In water tracing experiments it is often interesting to use two dyes simultaneously. In order to separate their fluorescence signals, we installed a second excitation axis, identical to the first, but with a different filter. The new filter is designed to favor the excitation of the second dye at the expense of the first. Thus 4 independent values can be sampled by combining the two light sources and the two photodetectors Data acquisition The datalogger box contains the acquis1t1on electronics and a battery Every 4 minutes, a 2-second measurement cycle is initiated. The lamps are successively switched on and off. Six values are collected: the 4 discussed above, and the values at zero excitation (this measure gives the photodetector baseline) All the information is written to a PCMCIA memory card. This data medium can be easily and safely read in a portable computer Battery and memory card capacities allow for 2 weeks of unattended data recording. 3. Performance The detection threshold of the fluorometer ( used in the real situation of a quiet stream) is 5 x I 0' 11 g/ml for uranine Sensitivity is 200 mV for a concentration of I O' g/ml. In clean water, the residual noise level is 0.1 mY The analogue-digital conversion dynamics is 16 bits Several tracer tests have been done during which the data from the fluorometer was compared to the samples taken and measured su bsequently in the laboratory The results of two of them are shown in Figs 2a and 2b. The first difference concerns time resolution The high sampling rate of the fluorometer allows detection of faster variations in dye concentration. Second the tracer arrival time and the position of the maximum 120 A la Mllandnnll OSJOQl22 t,, ""''"'"" '\ -luommlr ) 25 T {houri) concentration reach an acc uracy of as short as 4 minutes A summary of the properties of the sampler a nd the fluorometer is given in the Table I. Advantages of the field fluorometer compared to samples measured subsequently in the laboratory The autonomy of operation, the readings every 4 minutes and not having to visit the site are the principal advantages of the field fluorometer. ln the case of the mechanical sampler, however the bottles filled with the water samples must typically be handled at the site at the end of each series This operation is always error prone and there is a risk of contamination even if the greatest care is taken. Then, the samples must be analysed by bringing them about to a laboratory which costs more and has inevitable delay. Moreover, the storage of the samples prior to analysis can lead to variations in the dye concentration. Also, in the case of tracer tests using several dyes, chemical cross-reaction can occur. It is also been documented that sulforhodamine G is adsorbed by clay in an hour (MDAGHR I ALAOUI, 1992 ). Two advantages of the fluorometer relate to mechanical aspects: unlike the water sampler, the fluorometer has no moving parts, and therefore no wear. Also, because it is always completely submerged, freezing cannot prevent water from flowing through the system window. Measuring fluorescence is done on the flowing water. The flow may have particles particularly during a fast rise in water level. Water turbidity then becomes an influencing component of the dye concentration measurement. This is never the case with laboratory analyses, where care can be taken to allow the water to settle before analysis Up to some threshold the turbidity effect can be removed from the signal, since one of the photodetectors is permanently measuring it. The sensitivity threshold is directly related to water clarity. Therefore laboratory measurement will detect a smaller dye concentration compared to in situ field fluorometer measurements. lA Tln, 96106/10 \ I:!. labo ralOf)' -fluo,ometer \ \_ '----.. "' eo .. "'' T(hour ) Fig. 2 a & b: Comparison of the results from the fluorometer (line) and the sampler (triangles) at 2 sites. 48 Proceedings of the 12 th International Congress of Speleology 1997, Switzerland Volume 2

PAGE 60

Sampler Fluorometer Time resolution bottle filling time 2 seconds Field duration number of bottles x sampling period 2 weeks Work expense large: manv visits necessarv low: set-up and removal Subsequent handling sample analvsis in the laboratory none Result availability up to several davs immediate Error by contamination possible impossible Errors bv tracer deeradation possible impossible Moving parts (wear) yes none Freezine: sensitivity yes no Turbidity sensitivity none if samples filtered yes Saturation threshold none 1.25 x 10"' g/ml Usable tracers all, incl. biological dyes Number of simultaneous tracers >2 2 Detection threshold (uranine) 2 x I011 g/ml 5 x 10 11 g/ml Table 1: Comparison of a water sampler and field fluorometer performance. Disadvantages are shaded. Disadvantages The maximum detection level of the fluorometer is determined by the amplifier gain. In a tracer test, this gain is kept fixed. However it can be decreased before the test, if it is desirable to measure larger conce ntrations. Using a water sampler allows analysis for all tracers, including the biological ones (JEANNIN et al., 1995). Furthermore, the separation of simultaneousy injected tracers (dyes and others) is a straightforward task using samplers. The fluorometer works well with one dye tracer at a time but also can separate uranine and sulforhodamine values. Simultaneous use of 2 tracers To be able to separate 2 tracers, the fluorometer must be calibrated. The coefficients c,,, c 12 correspond to a sig nal under excitation by source I of the uranine and sulfor hod amine at a concentration of I 01 g/ml, whereas c 21 c,, corresponds to source 2. The expressions [u] = C 22 X1 C1 2 X 2 Cl 1C22 C12C21 [s]= C11X 2 -C 21X1 Cl 1C22 C12C21 give the co ncentration of uranine (u) and sulforhodam ine (s). The values x" x, are the vo lta ges measured on the photodetectors. Note that careful choice of excitation and detection filters is important to ensure a good separatio n of the tracers. Combinations that result in too small val ue s of the denominator must be avoided However, the same filter set can be used for various combinations of dye tracers, as long as the denominator rule is observed. Laboratory tests show that it is possible to separate 2 tracers with good accuracy, provided their individual response is known (Fig. 3) 100 80 CL 60 ., C: ., -0 40 :, "' ., ., ::;; 20 a .. 0 I 0 20 I Sepan1tionot2t,.0e1"1 e su l olhodamm, D wanhe a ,i I 40 60 Uranne [PPB] a I 80 a 100 Fig. 3: Separation of 2 tracers Determination of their concentration i11 the mixture. The real 11ra11ine concentration is 011 the horizantal axis (inverted scale for sulforhodamine) and measured co11centratio11 011 the vertical axis Conclusion The flow-through field fluorometer represents an econo mical alternative to commercial fluorometers and to mechanical sampling syste m s. Compared to the latter, it has the advantage of dramatically reduced survei llance and maintenance (one set-up and one removal). Its sho rt time resolution sho uld allow an new approac h to hydrogeological issues that where neglected with the use of samp lers 5 th Conference on Limestone Hydrology and Fissured Media 49

PAGE 61

The data availability is immediate. If needed, the content of the memory card can be read into a laptop computer at any moment, and lhe tracer test still carried on. Laboratory delays and expenses are not needed Simultaneous use of several tracers is current practice. To develop further this instrumentation, an effort should Photo 1 : Flllorometer with its daJa acquisition box. References JEANNIN, P.-Y., A WILDBERGER & P ROSSI. 1995 Multitracing-Versuche 1992 und I 993 im Karstgebiet der Silberen (Muotatal und Kloental, Zentralschweiz ). Beitraege zur Hydrogeologie Graz, 46: 43-88. MDAGHRI ALAOUI, A 1992 Traceurs artificiels en hydrogeologie: Limites d'application et critiques des methodes d'analyses In: Hydrogeologie des milieux discontinus sous climats arides, Marrakech, Maroc. still be made to optimise the excitation/detection filter sets In addition, the advent of laser diodes in Lhe green part of the spectrum will allow new sources for the selective excitation of uranine. ..... \. i ~ .. f ,, 50 Proceedings of the 12 th International Congress of Speleology 1997 Switzerland Volume 2

PAGE 62

Experimental design technique and protocol in fluorometric tracing of ground water b y C h ris C. S ma rt & L. Z a b o Department of Geography, University of Western Ontario, London, Ontario, N6A SC2, Canada. SMART@SSCL.UWO.CA Abstract Detector-spectrofluorometer and water-filter fluorometry provide two complementary tools in tracer testing Current practice largely relies on the former technique despite some limitations in quality assurance A set of technical enhancements have been undertaken on the Turner Designs Model 10 Series filter fluorometer to redress this imbalance by improving efficiency. In addition, improvements in protocol are suggested A preliminary framework for optimal design of tracer tests is presented Nevertheless there remain barriers to wider deployment of filter fluorometers notably in complex traces and borehole tests 1. Introduction The majority of ground water traces in Anglo-America are used for determining ground water trajectories in karst aquifers (QUINLAN 1986) and employ fluorescent dyes collected on activated charcoal detectors, extracted using an alkaline-alkali elutant and analysed on a spectrofluorometer. The technique is extremely cost effective in tackling basic questions of hydrogeography ", where many sampling sites are under assessment. Moreover spectro fluorometry permits relatively precise separation of tracers so that multi-tracing can be undertaken, i.e. several tracers can be used simultaneously The charcoal-spectrofluorometer approach has become widely established and accepted in hydrogeological disputes However there are some disadvantages to the charcoal-spectro fluorometer technique which deserve consideration Detectors can be contaminated in handling and processing Only minute quantities of dye concentrate are needed to contaminate parts per trillion (ppb) samples Replicate tracing is ambiguous, because residual dye may well cause an apparent positive analysis. Furthermore the inte grated natu r e of the fluorescence measurement provides limited proof against contamination; a given signal may just as easily arise from contamination as reception of the true tracer A related shortcoming is the lack of quantitative information on time concentration and mass The processes of tracer adsorption. exchange and desorption are poorly understood and controlled rendering the technique qualitative despite the sophisticated analytical tools emp l oyed Activated charcoal concentrates many fluoresce n t compounds often increasing background and undermin ing the 'amplification" attributed to such detectors Furthermore, the high pH of elutants can cause significant shifts in fluo r escence spectra (KASS 1992) compromising the sensitive spectral separation of m u ltiple tracers Adoption of appropriate analytical protocol allows defence against many of these weaknesses. However it is seldom considered cost effect i ve to run t h e multi-level system of controls and calibrations required in field sampling for low concentration fluorescent tracers A well resolved tracer breakthrough curve is extremely unlikely to arise from accidental contamination, or background from previous traces, and therefore provides a far stronger demonstration of a trace than a single positive detector Trace r breakthrough curves also p ermit hydraulic analysis to be undertake n (e g. SAUTER 1992). Se r ial quantitative analys i s allows estimation of tracer recovery a powerfu l test of the completeness of a trace Furthermore time mass data provide the key to developing contaminant transport and dynamic discharge models Spectrofluorometry of multiple water samples has not been adeq u ately exp l ored, presumably the costs are considered too high. Fi l ter fluorometry is an excellent complement to detector spectrofluoro m etry, providing serial quantitative concentration data In combination with discharge data, tracer recoveries can be estimated. Accidental contamination is usually self evident in filter fluorometric data, and is virtually impossible w i th continuous flow filter fluorometry The primary disadvantage of filter fluorometers is the considerable capital and logistic cost involved in deployment of water samplers or fluorometers in sample handling and process ing and in data base management. Filter fluorometers have very good sensitivity but separation of multiple tracers and background is much more difficu l t. Hydraulic analysis and tracer recovery require monitoring of discharge, another time demanding field technique As intensive users of filter tluorometers we have been attempting to improve fluorometric techniques, experimental design and protocol to maximise efficiency and information yield The remainder of this paper identifies some tech n ical problems and solutions associated with the Mode l I 0 some general improvements in protocol in filter fluorometry and a first attempt at formalising the design of simple tracer experiments 2. Filter fluorometric technique The majority of our work has involved the use of the Turner Designs Model I O Series fluorometer either in continuous flow mode or for laboratory ana l ysis of grab samples collected automatically or manually The Model I O is a robust, rather dated instrume n t which remains the most common filter fluorometer used in fluorometric tracing in Anglo America. T h e field version of the Model 10 is weather proof. Power requirements are flexible but high. The instrument has excellent dynamic range obtained by automatic switching through x Jx 3 16x I O-x 31.6 and manual x 1x 100 ranges (The x J 16 multiplier is the square root often and allows flexible calibration and conversion of fluorescence to concentration .) The instrument i s read on a two range analogue panel meter. Operational calibration requires a blank and single standard, although more thoroug h tests should be run periodically Excellent telemetry connections allow attachme n t of a programma ble data logger not only to collect data, but to substantial l y upgrade the performance of the instrument. We have used the Campbell Scientific CRIO and 21x, although any good programmable data logger co ul d be used. Table I provides a summary of techn i cal pro b lems we have encountered with filter fluorometry of grab samples and in continuous flow mode, using the Turner Model I 0 In addition to the external logger sample handling and data processing protocol has been improved and a preliminary attempt made at developing tools to aid in test design Furthe r details can be obtained from SMART & ZABO ( 1997), or the senior author. 6 th Conference on Limestone Hydrology and Fissured Media 51

PAGE 63

I I Problem I Mode I Solution I I. Limits on sampling frequency and data CF / Grab Record to external logger. transfer 2 Output noisy and inefficient for readCF / Grab ND conversion and averaging using logger. ing and calibration 3 Conversion to ppb complex CF / Grab External processing using IO!!.!!.er 4. Slow range change (esp on rising CF External control using logger limb) 5 Automatic range limited CF Automate using logger and servo motor 6 Filter changes slow Grab Allen panel screws multiple tluorometers 7 Power consumption too high for reCF/Grab Micro-hydro, (solar) mote areas Table 1. Summary of problems with the Tumer Designs Model JO Seriesjluorometer a11d proposed solutions. CF implies applications with continuous flow jluorometry, "Grab" indicates a11alysis of discrete water samples. 1. Data frequency and transfer The Model 10 is conventionally read from an analogue panel meter or (in continuous tlow mode) drives an analogue chart recorder These data displays are invaluable for quick visual interpretation but do not lend themselves to compilation of accessible high frequency data Telemetry connections permit easy external logging to obtain a crude digital voltage record of concen tration and range time series Machine readable data are a great improvement over chart or manual recording Automated recording of grab sample analysis is less advantageous, at the very least a manual record is kept to clarify inconsistencies in the logger record 2. Noise rejection and calibration Instrument noise makes reliable reading or analogue recording of the panel meter difficult especially at low concentra tions or if the sensitivity adjustment is set too high Much of the instrument instability is random or white noise which can be rejected by compiling averages The external logger is easily programmed to convert the analogue signal to digital form and to output the average of a swath of readings, and this has been demonstrated to improve the precision of our results Logger averaging has also been applied successfully to grab sample analysis using a manually-triggered burst of readings and digital display External averaging also significantly improves instrument accuracy by allowing more precise calibration. 3. Conversion to concentration. Converting panel reading or external voltage readings to concentration requires some work and can be a source of error especially when processing uncorrelated grab samples or when averaging across range changes in continuous tlow mode Further more the conversion algorithm depends on the current calibration settings A robust logger algorithm has been developed to convert external readings of the analogue signal and range output to parts per billion. The appropriate multiplier is selected empirically during calibration Processing to ppb eliminates errors arising when compiling averages across range changes ; ppb is continuous across the break whereas voltage is not. Cross range calibration may drift on older or abused instrument. Such error can be compensated for within the conversion algorithm rather than by expensive internal refurbishment. Raw data are recorded should there be doubts about the reliability of the processing algorithm 4. Slow range change Automatic range changing is a major advantage in continuous flow mode However, range-changes take several seconds especially during rising tracer concentrations when the instrument steps slowly through two redundant ranges. Instrument readings are spurious during range changes and this can become critical with dilution gauging or very rapid tracer breakthrough. The solution is to use the logger to force stepping through redundant ranges rapidly and to suspend data collection during the adjustment phase 5. Limited Automatic Range Despite excellent dynamic range the manual x J -x I 00 limits automated operation The full dynamic range is needed where peak tracer concentrations cannot be anticipated or controlled by tracer mass adjustments The logger has bee programmed to identify conditions in which manual switching is required This can be u s ed to trigger an alarm if an operator is available. For full automation an external reversible motor is attached to the x I to x I 00 switch. A timed burst of appropriately polarised o c power drives the motor for the required number of revolutions This is the most arduous adjustment to the instrument, but requires only modest machine and electronic expertise 6. Filter changes Switching filters can take ten minutes or more. An inexpensive adaption (courtesy, P.L. Smart University of Bristol) is to replace the slot screws on the front panel with Allen screws (hex-sockets) which allows use ofa speed driver. (Light lubrication is essential as the screws are tapped into aluminium.) 7. Power consumption The Model 10 consumes some 24W of power which requires mains supply or independent power generation Continu ous tlow pumps are more power-demanding Gasoline generators add considerably to the logistic load and are best avoided Adequate 52 Proceedings of the 12 th Internat i onal Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 64

Detector-Spectrofluorometry Water SamplersContinuous Discharge Filter Fluorometers Flow Filter Fluorometry Detectors SpectroWater SamFilter Fluoro(n ~ 3) Stage sites/ fl uorometers pies meters calibrations Hydrogeography N.n v I N v f 1..n N.n N l l..10 Time of Travel N n v f I N v f l..n N n N l l..10 Time-Cone" n a N v f 1..n N n N I 1..10 Hydraulics n a N.v.f l..n N n N I 1..10 Recovery n a N.v f 1..n N n N I 1..10 Table 4. A matrix to allow design of appropriate resource a/location for tracer tests. Columns list instrument options; rows research questions. n.a. indicates that a technique is not applicable. N is the number of sample sites; n is the number of tracers; v the number of replicates,fthe sampling frequency. J .. n indicates that a single jilter jluorometer can be used and reset n times, or (more efficiently, n jluorometers can be used in parallel. The number of discharge calibrations depends on the variability of stage at the N sites. photovoltaics are prohibitively expensive and ill-suited to tempo rary deployment. Micro hydro has proven effective given adequate head ( 1 m with large discharge) A lower power fluorometer 3. Handling and processing protocol I. Water sample containers Grab water samples and elutant samples are normally decanted into a rinsed cuvette for analysis and where sample is in short supply returned to the sample bottle for subsequent analysis or archiving. This is time consuming and with alkaline elutant hazardous Handling of exposed samples invites contamination which is especially serious when a single charcoal detector provides a solitary data point. Serious fluorometric contamination is positive and occurs when the operator handles both dye concentrates and samples or sampling equipment. A solution (adopted from C Alexander University of Minnesota) is to use sample bottles that can be inserted directly into the fluorometer. Direct insertion of s amples has significant advantages in processing efficiency and reduction of contamination There is some loss of precision in using inexpensive glass bottles but optically flawed bottles can be rejected and the large increase in efficiency allows more samples to be collected in compensation lfreusing sample bottles to reduce waste care should be taken that dye does not become lodged in cardboard cap liners ; we use a cap liner of cling-film to prevent this. 2. Temperature compensation Fluorescence for the rhodamine dyes has significant inverse temperature coefficients of -2-3% per C (KASS 1992) which may be significant in quantitative applications especially when working in the field In continuous flow fluorometry a temperature sensor can be added to the fluorometer outfall and a temperature compensation subroutine added to the data processing This requires the temperature of calibration and the correction factor (KASS 1992) to be recorded in the logger. Alternatively temperature can be recorded and compensation undertaken during post-processing Paradoxically, temperature compensation is more difficult with grab samples where some attempt can be made to homogenise sample temperature Problems arise because the sample compart ment is warm, and fluorescence declines as a sample warms up Temperature cannot be read during analysis because samples become thermally stratified and the temperature sensor must not interfere with the light path Our approach is first to immerse samples and standards in a circulating water bath at ambient temperature We have adapted our sample compartment to allow forced ventilation of outside air. lnsertable sample bottles are an essential part of this strategy This approach not only bypasses the temperature compensation problem, but it also encourages frequent reference to calibration standards 3. Cross Fluorescence Many tracer tests involve the use of multiple tracers or demonstrate significant natural background Unfortunately the characteristic fluorescence spectra of tracers can overlap Fluores cent background typically peaks in the blue regions of the emission spectrum but can still be significant at green and red wavelengths. Fluorescence reading can therefore be ambiguous especially with filter fluorometers It is possible to correct for "cross-fluorescence ," providing arbitrary standards are available for the relevant fluor es cent compounds However there is little evidence for such prac tices. Corrections for cross-fluorescence are provided in SMAR T & ZABO ( 1997) and KASS ( 1992) Cross fluorescence corrections are simple to incorporate into machine data processing and should be made routinely for both filter and spectro-fluorometry whenever multiple tracers are in use even if only one tracer is present. This is considered good protocol providing a secondary level of robustness in analytical procedures Cross-fluorescence correction also provides an excellent method of coping with variable concentrations of background providing it is possible to characterise background with a suitable standard .'' 4. Design of tracer tests l. Experimental Design Efficient tracer test design is often attributed to expert intuition but some elements are amenable to scientific design For example WORTHINGTON et al ( 1995) provide an efficient algorithm for predicting tracer mass based upon distance spring discharge and peak concentration. The following is a first attemp t to develop a simple expert system" to aid the design of tracer tests. Compound tracing can be defined as tracing involving more than a single sampling site (N> I) Multi-tracing implies the use of multiple tracers (n > l ) usually implying multiple injection sites. Replicate tracing implies repetition of particular traces ( v > I) The other operational parameter is sampling frequency (f) crudely 6 'h Conference on Umestone Hydrology and Fissured Media 53

PAGE 65

presented as samples per trace. The resources required for tracer testing constitute the equipment listed in Table 4, plus consumables such as dye, and manpower. The nature of a test depends upon the question being asked, the complexity of the system, avai l able information and budget. Table 4 has been developed to formalise appropriate resource allocation for distinctive questions, and as a guide to management of key issues in design of appropriate instrument deployment. For example, hydrogeographic questions are best tackled using detector-spectrofluorometry In contrast, the number of continuous flow fluorometers to be deployed in a quantitative test is n x N, so that the highest level of efficiency is obtained for replicate, high resolution in simple systems (n and N are minimised) 2. Prac t ical optimisa t ion of tracer tests Table 4 is not intended to over-simplify tracer testing Fundamental groundwork in reconnaissance hydrogeology is essential, and remains highly dependent on the experience, skill and insight of the operator. Equally, it is inefficient to tackle quantita tive problems without an established hydrogeography. There are also many practical ways to optimise tracing For example, G Davies (Cambrian Ground water Inc., Oak Ridge Tn.) advocates a series of single tracer injections progressing headward from a continuously monitored spring site This allows efficient compilation of knowledge with a minimum of sample processing. Similarly, it may be possible to analyse water from adjacent sampling sites sequentially using electronically controlled valves (e.g. FOUNTAIN 1994) In distributary karst systems springs may be classified into "family groups" exhibiting similar tracer breakthrough It may only be necessary to sample a single member of a spring family. Stage monitoring and discharge rating are expensive, but can be rationalised Monitoring should concentrate on high mass (=discharge x concentration) springs, and overflow members in distributary spring systems Underflows" exhibit little variation and do not require continuous monitoring of stage The gains in sample processing speed described above make the water sampler-filter fluorometer strategy most efficient for quantitative compound multi-tracing Continuous flow fluorometry may also be constrained by problems like turbid water or inadequate power In complex traces however, the real-time data from a master continuous flow fluorometer can be invaluable in management of sampling tactics. In addition to the minimalist strategies presented in Table 4 data verification must also be considered Continuous flow fluorometry does not preserve samples for archiving or verification, and instrument failure can resu l t in irremediable data loss. Apart from designing a robust monitoring strategy, charcoal detectors and manual water samples can be collected and archived at little additional cost (providing samples are protected from bioand photo-degradation ; KASS 1992) Verification of detector-spectro fluorometric tracing is more difficult and requires multi-level blanks for background, charcoal, elutant etc. 3. Borehole deploy m ent While design of tracer tests for conventional karst tracing is relatively easy, many practica l problems involve "mid-aquifer" applications and boreholes The arbitrary position of boreho l es and forced gradients means that borehole tracer tests may not reveal solutionally enhanced permeability. Nevertheless, the adap(ed Model 10 is well suited to forced gradient experiments, and has been routinely used to continuously monitor water pumped from a borehole. However, natural gradient tracer experiments are much more difficult, and should be conducted with small samples collected with replacement (e.g. NOYAKOWSKI 1992) The pressing need is for the development of inexpensive versions of in situ fluorescence sensors similar to those described by BARCZEWSKI & MARSCHALL ( 1992) and BENJSCHKE & LEITNER ( 1992) 5 Conclusions The enhancements suggested in this report permit the venerable Turner Model I O to be operated as a leading edge filter fluorometer comparable to the new Turner 10 AU, and may make filter fluorometry more viable as an applied tool. Recommendations on protocol and test design should be considered by any practitioner interested in undertaking robust and defensible water tracing. Despite the complementarity of filter and spectrofluorometer techniques, there remain critical technical and operational limita tions in tracer testing. Low power, inexpensive fluorescence sensors should be able to reach sub ppb sensitivity and are seen as comple mentary to the modem arsenal of water tracing tools Acknowledgments Field and laboratory work reported here has been generously funded by the Natural Sciences and Engineering Research Council of Canada, and the University of Western Ontario Aspects of this work have involved collaboration with C Alexander (University of Minnesota), S. Worthington and D C. Ford (McMaster University) and G. Davies (Cambrian Ground water Inc ). Many students have p articipated in field experiments in fluorometric tracing. References BARCZEWSKI R., & MARSCHALL P., 1992. Development and applica tion of a lightfibre fluorometer for tracer tests. In, HOTZL H., & WERNER, A Tracer Hydrology Balkema, Rotterdam 33--40. BENISCHKE R & LEITNER A ., 1992. Fibreoptic fluorescence sensors-an advanced concept for tracer hydrology In HOTZL H ., & WERNER, A Tracer Hydrology Balkema, Rotterdam, 33--40. FOUNTAIN A.G ., 1994 Automated system for measuring and record ing fluorometric data from multiple sources. U.S. geol. Surv Water Supply Paper 2340. KASS, W 1992. Geohydro/ogische Markierungstechnik, Gebriider Borntrager, Stuttgart, 519 pp. NOYAKOWSK.J K ., 1992 The analysis of tracer experiments con ducted in divergent radial flow fields. Water Resources Research 28, 3215-3225 QUINLAN, J F., 1986 Discussion of Ground water tracers" by Davis et al. ( 1985) with emphasis on dye tracing especially in karst terranes Ground water 24 253 259; 396--397 SAUTER, M ., 1992. Quantification and forecasting of regional ground water flow and transport in a karst aquifer (Gallusquelle, Malm, S.W. Germany) Tubinger Geo wissenschaflliche Arbeiten (TGA) C 13, 251 pp. SMART c.c & ZABO, L., 1997. Some developments in fluorometric tracing of ground water. VI Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst. Ln press WORTHJNGTON R.R H. D C FORD & c c. SMART, 1995. Guidelines for the quantity of dye needed for quantitative tracing to springs in carbonate rocks ln G GUNA Y L TECZAN A A TILA, International Symposium and Seminar on Karst Waters and Environmental Impacts Beldibi Turkey 54 Proceedings of the 12 th Internat i onal Congress of Speleology, 1997, Switzerland Volume 2

PAGE 66

Perspectives offertes par la mesure en continu des traceurs fluorescents dans le karst par P. Meusl, M. Bakalowicz 2 W. Kass 3 B. Barczewski et G. Schmid 4 1 Laboratoire Souterrain du CNRS, F-09200 Moulis 2 CNRS, URA 1767 "Geofluides, Bassins, Eau", Groupe de Geochimie, F-34095 Montpellier et BRGM, Direction de la Recherche, F-34000 Montpellier 3 Miihlematten 5, D79224 Umkirch 4 VEGAS, Institut fur Wasserbau, Universitat Stuttgart, D-70550 Stuttgart Abstract A continuous tracer test lasting several months was performed between an annex-to-drain system (La Peyrere cave) and the main outlet (Las Hountas) of the Bagel karst system. Uranine at the spring was recorded with a 10 mn interval] thanks to a flow-through cell automatic fluorimeter. The breakthrough curve of the tracer reflects the dynamics of the system. 1. Introduction Un des problemes majeurs dans la prediction du transport des polluants dans le karst est sans aucun doute la maniere dont varient Jes reponses impulsionnelles. En raison de la structure particuliere du drainage, non seulement ces reponses varient suivant l'etat hydrodynamique du systeme mais de plus Jes variations peuvent etre discontinues. Pour ces raisons, la plupart des aquiferes karstiques ont un comportement typiquement non lineaire. Divers auteurs ont tente de tenir compte d'une variation de la reponse impulsionnelle avec le debit de tra~age (NIEMI, 1977; WBER, 1986; DZIKOWSKI, 1995). Siles premieres solutions proposees s'averaient trop restrictives (entre autres la condition de volume constant), en revanche Jes demieres generalisations de l'integrale de convolution permettent a l'heure actuelle une approche plus realiste. L'experience decrite ci-dessous consistait en un tra~age en continu, pendant une periode la plus longue possible, sur un systeme karstique, de fa~on a tester, avec une fonction d'entree parfaitement connue, Les reponses a un eventail le plus large possible d'etats du systeme, et ainsi pouvoir trailer ces demieres d'un point de vue statistique. Elle a ete rendue possible grace a !'utilisation d'un fluorirnetre de terrain specialement con~u pour le dosage de l'uranine en continu a la source. 2. Le systeme du Baget et le tra~age en continu Le debit moyen du systeme du Baget est de 0,5 m 3 /s. Ses reserves (volume dynamique moyen de l'ordre de 1,5 10 6 m 3 ) proviennent d'une part de la recharge continue par Jes ecoulements allochtones qui se perdent au contact des calcaires et, d'autre part, du stockage de l'eau dans Jes zones laterales au drain, dans des vides karstiques parfois importants appeles Jes "systemes annexes". L'un de ces systemes annexes, le gouffre de la Peyrere, avait fait l'objet d'un pompage a gros debit ( BAKALOWICZ CROCHET, D 'HUL ST MANGIN, MARSAUD, RICARD et ROUCH 1994) mettant en evidence !'absence de reserve importante en dehors du conduit karstique lui-meme. C'est precisement ce gouffre qui a ete choisi comme lieu d'injection du traceur dans le but de preciser Jes mecanismes de Ja recharge et de la vidange de ces systemes annexes. L'injection d'uranine a eu lieu, depuis une cuve de 3000 I placee dans le gouffre, grace a une pompe doseuse equipee d'un clapet de maintien depression a 3 bars pour s'affranchir des fluctuations du plan d'eau d'environ 25 m au-dessus du point d'injection. Le debit constant etait de 55 rnl/mn et la concentration de la solution mere de 3 g/1, plusieurs essais et modelisations ayant perrnis de trouver un compromis entre une concentration genante a la source et une concentration mesurable (> 0,1 ppb) malgre la dilution par Les crues. Le fluorimetre comporte une cellule a flux continu equipee d une diode comme source d'excitation et reliee au bottier electronique de mesure via un systeme de fibre optique ( BARCZEWSKI KASS, SCHMID und WERNER, 1996). 3. Interpretation du tra~age et conclusion Une courbe de restitution de l'uranine a l'exutoire de Las Hountas a ainsi ete obtenue au pas de 10 mn pour la periode de decembre 1996 a avril 1997. Cette courbe, comparee aux fluctuations naturelles du systeme (pluie, debit, conductivite, temperature) egalement mesurees en continu, permet une interpretation particulierement fine de la reponse. Elle confirme en outre !'existence de non linearites qui peuvent s'expliquer par la deconnexion de parties du reseau lorsque le niveau d'eau devient inferieur a celui de certains conduits preferentiels. L'experience montre que, malgre ces non linearites, le modele de convolution a debit variable de DZIKOWSKI (1995) est en bon accord avec le fonctionnement observe du systeme karstique. En conclusion, Ja mesure en continu des traceurs fluorescents injectes artificiellement dans Jes systemes karstiques constituent un excellent moyen de caracteriser la dynamique de ces systemes. Cette approche phenomenologique renseigne sur le degre de predictibilite, une notion dont la connaissance apparai't de plus en plus comme indispensable pour une gestion correcte des ressources en eau. NB : Cette recherche a ete fioancee dans le cadre du programme Capital Hu.main et Mobilite de la Commuoaute Europeenne. Nous tenons a remercier Mme J. Daffis et Mr G. Fontanini, techoiciens au Laboratoire Souterrain, pour leur aide precieuse sur le terrain 6 th Conference on Limestone Hydrology and Fissured Med i a 55

PAGE 68

Dripwater monitoring at Grotta di Ernesto (NE Italy) : a contribution to the understanding of karst hydrology and the kinetics of carbonate dissolution Andrea Borsato Museo Tridentino di Scienze Naturali via Calepina 14 1-38100 Trento ItaJy Ab stra c t Dripwater monitoring at Grotta di E rnesto (NEItaly) revea l s the different hydrological functioning of slow and fast-drippin g s talactites The conductivity record of fast dripping stalactites is punctuated by several abrupt negative peaks following strong rainf a ll epi s odes with a response time of a few hours During these infiltration event the discharge increases immediatel y, and the dripwater becomes strongly undersaturated The recovery of the previous condition needs from I O day s up to two months which indicate s th e progressive increase of the residence time of the water in the aquifer. In low-dripping soda straws no infiltration events are re c orded and the water is near the saturation state throughout the year Nevertheless, the conductivity trend is similar to that of fast-dripping s talactites which indicates a direct connection of the slowand fast-dripping stalactite plumbing systems R i assunto Il monitoraggio della acque di percolazione nella Grotta di Ernesto ( Trentino-Italia) ha evidenziato ii diver s o comportam e nt o idrogeologico di stalattiti con differenti velocita di gocciolamento La registrazione annuale della conducibilita di stalattiti a per c olazione veloce e caratterizzata da improvvisi picchi negativi connessi all'infiltrazione di acqua poco mineraliz.zata a seguito d i forti piogge Il ripristino dei valori di conducibilita precedenti l evento infiltrativo richiede da 10 a piu di 30 giomi e riflette ii progressivo aurnento del tempo di residenza dell'acqua nell acquifero Stalattiti a cannuccia a percolazione lenta non rcgistrano alcun evento di infiltrazione e l acqua e vicina alla saturazione durante tutto l anno Tuttavi a, l'andamento della conducibilita e s imile a quello di stalattiti a gocciolamento veloce, e testimonia una diretta connessione tra i due sisterni di alimentazione 1 Introduction and aims The recent introduction of automatic data-loggers to record cave air and water phisyco-chernical properties (MANGIN 1985 ; V IGNA, 1993 ; CI GNA, 1995 ; MAsS EN et al ., 1995 ; CELI & SAUR O; 1 9 9 6 ) has greatly improved the quality and number of researches on karst hydrology In the last few years this techniques has been applied also to the monitoring of low-dripping morphologies such as fissures, cone stalactites and soda straws (BORSATO, 1996 G ENTY et al ., 1997) in order to detect the delay time of dripwater after rainfall events, and the seasonality of dis s olution precipitation phenomena Caves which develop a few tenth of meters below the surface hav e the property of recording annual to seasonal climatic information through the physico-chemical characteristics of their dripwaters and speleothem carbonates (BAKER et al ., 1993; B ORS A TO, 1995 ; G E NT Y & Q UTNIF, 1996 ; FAIR C HILD et al. 199 6). Furthermore for temperate karstic areas shallow caves dripwater is subjected to seasonal fluctuations of the saturation s tat e from undersaturation to supersaturation due to the limited thickness of the rock above the cave that reduce the residence time of the water in the aquifer (U GG ERI 1992 ; B o RsATO 1995) Consequentely such caves are particularly suitable for h y drological studies in order to detect the kinetic of carbonate dis s olution in the epikarst zone The present study informs on the results from one y ear monitoring of two drippings at Grotta d i E rnesto 2 Cave site and environmental setting Grotta di E rnesto opens at an elevation of 1167 m a s l. (L ong 11 39 28" Lat. 45 58 37 ) in Valsugana valley ( Trentino NE-Italy) (DALMERI, 1985) and developed in partially dolornitised oolitic limestones of the Calcari Grigi Formation (Lias) which exhibit a good porosity due to the volume reduction during dolornitisation process. The cave consists of a single downdipping gallery 2 to 5 m wide, up to 4 m high and 72 m long, which developed along a NW trending subvertical fault 10 to 25 m below the surface The original syngenetic morphology of the cave is now blurred by collapse phenomena and the whole floor is covered by angular blocks, partially coated by flowstones (B o R S A TO, 1995 ) The cave contains a lot of active and fossil speleothems the age of which range s from 150 000 years to the Present (BoRSATO 1995 ; FRISIA et al ., 1997) and preserves important Mesolithic findings (D AL MERI 1985) At present, water enters the cave through several drippings, with low (from 0 I to 2 ml/min u te) to medium discharges (up to 50 ml/minute) feeding small muddy or concretional pools Pools show seasonal water level fluctuations : in summer some are dry in autumn and spring all pools fill up and several overflow The cave lies below the timberline, within the deciduous forest vegetation belt consisting of Fagus Carpinus Lanx decidua and Abies alba The climate is temperate humid with a mean annual precipitation between 1000 and I 500 mm From December to March most of the precipitation is snowy and the snow cover remains until March Between March and the fir st half of April snowmelting occurs The cave entrance was discovered end enlarged in 1983 and subsequently closed by an iron door which inhib i ts air flow rn the cave All this characteristics, make the cave particularly interesting for paleoclimatic and h y drogeological studies 3 Methods and b ackground Two single stalactite drips were monitored from 19/06/95 to 19/06/96 by means of two data-logger which measure water 6 1 h Conferen c e on Ume s tone H y dro l og y and Fi ss ured Med i a 57

PAGE 69

290 0 0 0 N 270 ro E 0 -Cl) :i 250 ---;;: +J 0 :, ,:, 230 C 30 0 0 210 Snow melting 20 10 +--L..-f L-L-1_..YL .J.. 0 19111195 19112195 18101196 17102196 18103196 17104196 17105196 16106196 Figure 1.: Co11ductivity record for G2 stalactite compared with precipitation (meteoro/.ogical station Bieno 806 m a.s L). At each rainfall event higher than 10 mm/day correspond a abrupt negative shift in dripwater conductivity. temperature and conductivity. The data-loggers are the Thermos-data b y DAS-ENEA_ which are stored in impermeable boxe s, powered b y internal batteries and connected with a double-channel probe The characteristics of the probes are the followings : Temperature : range -10 to +40C, resolution 0 01 accuracy +I0 2C Conductivity : range 10 to 1000 Siem ; resolution I or 0.1 Siem; accuracy +l-2 Siem The two monitored drips located on the roof of the main chamber of the cave which developed about 15 m below the surface are a slow dripping soda straw (Gl) and a fast dripping cylindrical stalactite (G2 ). The drip of the soda straw Gl which fed an active cone stalagmite (stalagmite ER76 see FRISIA et al 1997) was collected with a funnel positioned at the tip of the straw This soda straw is actually growing like others straws in the cave which exhibit clear annual growth band ing (Tooru et al. 1996) The drip rate during the whole monitoring period, varied from 0 07 to 1 37 ml/minute (mean 0.43 ml/m) The drip of the cylindrical stalactite G2 which is directly feed by a subvertical fissure was collected with a funnel positioned at the splash-site, about 2 5 m from the stalactite tip. The drip rate throughout the monitoring period varied irregularly from 3 33 to 34 6 ml/minute (mean 12 7 ml/m) At present, the impact point of G2 drops identifies the apex of a large flowstone generated from it. For both drippings the interval between two measurements was I hour (reduced to 3 hour in successive elaborations) At the end of the monitoring period the bottom of the samplers were sediment-free and no precipitation occurred within them During the monitoring period, several water samplings allowed to measure the exact drip-rate and drop volume and analyse the ph y si c o-chemical properties of the waters This also allowed for the calibration of the conductivity probes which were affected bv instrumental drift due to adhesion of air bubbles on the electrodes 4. Results Fig I shows the seven-month record of G2 stalactite The most striking feature in its conductivity record are several abrupt drops of about -20 to -50 Siem, followed by gradual increase in conductivity By comparing this record with the precipitation data for the same period from the meteorological station of Bieno (806 m a s l. 12 km WNW from the cave ) it results that each negative peak in conductivity corresponds to a precipitation event exceeding 10 mm/day that occurred on the same day or on the day before The recovery of the previous conductivity values takes place in two separate phases : an initial rapid increase which lasts about one week, and a successive slow increase which can last 2 months or more and tends towards the asymptotic value of about 280 Siem In most of the cases the rise in conductivity does not reach this asymptote but it is interrupted by the successive abrupt decrease 0 280 .,-----------------~ C) N ro 260 E 0 -Cl) 240 :i ;;: 220 :, ,:, C 0 0 6 6 0 0 Q) ... 6 5 :, iii ai a. 6.4 E Q) co co co co co co co co f0) 0) 0) 0) 0) 0) 0) 0) -------v v I{) I{) I{) co co 0 0 0 0 0 0 0 0 ------c<5 -v v v v v v M 0 N 0 ..N 0 ..Figure 2.: Conductivity and temperature records for G2 stalactite. To each conductivity drop corresponds a sudden increase in temperature. The secondary positive temperature peaks in May and June are due to tourist visits. 58 Proceed i ngs of the 12 th International Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 70

0 0 0 N (ll E 0 284 280 276 ;; ,3 272 C 0 (.) lO 0) co 0 00 ..lO lO lO 0) 0) 0) -0 -co N 0 ....-co -r-lO ....6 7 (.) 6.6 ::, (ll ..... 6.5 , ;; :;; 0 230 ::, -0 C 0 (.) 210 19111195 19112/95 18101196 17102196 By comparing the two data-logger records (Fig. 4) it is possible to observe an overall seasonal similarity for both temperature and conductivity, with a gentle decrease in T 0 e X from January to March and a more rapid increase during April. For both O I and 02 the lowest conductivity values ( except for the infiltration event in G2) are recorded during late winter and early spring, and culminate at the end of March This period corresponds to the progressive melting of the snow cover, which is completed by early April. 5. Discussion The conductivity of dripwater is a direct function of the total mineralisation, i e of the Ion Activity Product (IAP) which, in pure limestone aquifers, represents fairly well the total CaCO 3 in solution (BAKALOWTCZ, 1974) In Grotta di Ernesto G I and G2 drippings the lAP is mostly identified by Ca 2 + (2 5 to 2 95 meq/1) and HCO 3 (2 8 to 3 1 meq/1), the sum of which commonly exceeds 90% of the total ion cocentration (B oRSATO, 1995) This means that strong fluctuations in conductivity can only be explained by differences in the total dissolved CaCO 3 Furthermore 02 dripping is commonly undersaturated with a Saturation Index for calcite from -0.4 to +0 2 (before the drop splashes), which indicates that, during several periods the water can potentially dissolve more carbonate This is also recorded by the conductivity record of G l dripping which is generally over I O Siem higher than that of the fast dripping G2 site (Fig 4) The overall trend of G2 conductivity directl y reflects the rainfall pattern, which strongly influence the calcite saturation state: during infiltration events the drip-rate increases immediately and dripwater becomes strongly undersaturated Water slowly shifts towards the saturation state during dry periods when conductivity increases towards the asymptote and the drip-rate slows down The first rapid conductivity increase probably corresponds to the soil and rock drainage after the complete soil imbibition, whereas the slow positive trend towards the asymptote identifies the progressive increase of water residence time in the aquifer Conductivit y ofG2 responds 6 8 6 7 0 0 6 6 6 5 ro Qi 6.4 a. E 6 3 f18103196 17/04/96 17/05/96 16/06/96 Figure 4.: Combined conductivity and temperature (daily running mean) records for GJ (bold lines) and G2 (thin lines) stalactites. Note thaJ conductivity of Gl dripping is generally over 10 Siem higher than thaJ of the fast dripping G2 6 '" Conference on Limestone Hydrology and Fissured Med ia 59

PAGE 71

immediately (probably less that 12 hours) to the infiltration events i.e. to precipitation higher than 10 mm/day, as shown by the abrupt decreases of -20 to -50 Siem. This sharp conductivity drop indicates the absence of "piston effects" (VIGNA, 1993) i.e. sudden conductivity increases before the negative drop Piston effects are common in aquifers where waters have a relatively long residence time and "old water with higher conductivity is pushed into the stalactite-feeding network by the pressure induced by the infiltration event. The immediate response of the system and the lack of piston effects indicate that the mean residence time of G2 drip is really short, probably less than 2 week, and is directly correlated with the conductivity. During the in.filtration events, the water residence time is probably lower than a day, and progressively increases as conductivity rises towards the asymptote. For Gl there is no direct relationship between drip-rate and conductivity/calcite saturation state: no infiltration event is recorded by conductivity drops and the only remarkable fenomena are the low-conductivity values reached during winter and spring snow melting The overall conductivity trend of G l is similar to those of G2 record, therefore we can infer that the feeding system is the same for both stalactites, but also that the plumbing system of G 1 is too narrow to allow for the sudden arrival of the infiltration event waters at the soda straw Because of this correlation, is possible to infer that the mean residence time of Gl water is only a little bit longer than that of G2, probably between one and two month However, this time-span, which is related at the good porosity of the rock above the cave is enough buffer the conductivity fluctuation. In this case, the most important mechanisms controlling the water saturation state are seasonal episodes like snow melting in spring and higher CO2 production during summer and autumn 6. Conclusion Dripwater monitoring gives precious insight upon karst hydrogeology and the feeding mechanisms of stalactites. In Grotta di Ernesto the conductivity and saturation state of fast dripping stalactites is directly controlled by the infiltration events and, as a whole, the longer is the time elapse from the infiltration event the higher is the conductivity and the saturation state On the other hand, in slow dripping soda straws (and cone stalactites) the saturation state is not directly controlled by the rainfalls: the conductivity values are quite constant throughout the year and the saturation state is only controlled by the CO2 concentration in the aquifer Acknowledgements: This work is part of the E U project: "Holocene-Late Pleistocene high resolution climate reconstruc tion from continental carbonates" contract EV5V CT94-0509 References BAKALOWICZ, M J. 1974. Geochimie des eaux d'aquiferes karstiques 1. Relation entre mineralisation et conductivite Ann Speleol 29 (2) : 167-173 BAKER, A, SMART, P EDWARDS, R L., & RICHARDS, D.A. 1993 Annual growth banding in a cave stalagmite Nature 364 : 518-520 BoRSATO, A. 1995 Ambiente di precipitazione e analtst microstratigrafiche di speleotemi in grotte delle Dolomiti di Brenta e Valsugana (Trento) : interpretazioni genetiche e implicazioni paleoclimatiche Unpublished Ph D Thesis Universita di Milano, 175 p BoRSATO, A. 1996. Analisi chimico-fisiche e monitoraggio di acque di percolazione nella Grotta del Calgeron (Valsugana) : primi risultati e implicazioni idrogeologiche Studi Trenti/11 di Scienze Natura Ii Acta Geologica 70 ( I 993 ) : 79-84 CELI, M & SAURO U 1996 Primi risultati del monitoraggio delle acque della sorgente de! Cogol dei Siori (Fiume Oltere Valle de! Brenta) Studi Trentini di Scienze Naturali, Acta Geologica 70 (I 993) : 69-78 CIGNA, A.A. (ed ) 1995 Grotte turistiche e monitoraggio ambientale Show caves and environmental monitoring Preprint Int. Symposyum, Frabosa Soprana, Cuneo 288 p DALMER1, G 1985. La Grotta di Ernesto : un insediarnento preistorico di grande interesse per la conoscenza del Paleolitico finale nell'area trentino-veneta (Colle
PAGE 72

Donnees hydrogeochimiques sur quelques karsts du sud-est asiatique par Fran~ois Brouquisse, Association Pyreneenne de Speleologie 14 Cite Foch 65000 Tarbes. France Abstract Between 1985 and 1996 some 120 samples were collected from karst environments during APS caving e,q,editions in Thailand, Indonesia and Laos. Analytical results are given: ions contents and computed variables for CO2-CaCO3 system Extensive sampling is representative. Factor analysis bears evidence for different water types according to karst hydrology and soil and vegetation factors Hydrochemical characteristics of these water samples don't show difference with those of karst waters in temperate climates Resume Au cours d'explorations speleologiques de l'APS de 1985 a 1996 plus de 120 prelevements d'eau ont ete analyses sur divers sites karstiques de Thailande lndonesie et Laos Les principaux resultats sont presentes : ions majeurs et rnineurs parametres calcules du systeme calco--carbonique L'analyse factorielle des donnees conduit a identifier plusieurs farnilles d'eaux notanunent en fonction de la dynamique du systeme hydrologique et des caracteristiques de couverture vegetale et pedologique Les caracteres hydrochimiques de ces eaux ne Jes distinguent pas de celles des karsts de regions temperees 1. Cadre general Peu de travaux sur la chirnie des eaux et la dynamique du CO2 ont ete menes sur Jes karsts de Thai:lande et d'Jndonesie On citera entre autres ceux de PITMAN (1978) sur Phangnga et de BALAZS (1968) en Indonesie. Cette note resume Jes principaux resultats obtenus entre 1985 et 1996 surplus de 80 sites differents Ceux-ci assurent une bonne representativite : ecoulements de surface pertes emergences percolations et gouttages en cavites circulations au niveau du karst noye (Tab l) Cinq regions ont ete plus particulierement etudiees : Chiang Dao au nord de Chiang Mai: la Nam Lang pres de Mae Hong Son Kanchanaburi et Phangnga en ThaHande ; le karst de Maros a Sulawesi en Indonesie Les secteurs de Chiang Dao de la Nam Lang et de Kanchanaburi sont plutot montagneux. Le karst y est developpe dans des calcaires penno--carboniferes parfois dolomitises ; la couverture pedologique silicatee d'origine detritique est relativement importante et tend a empiiter le relief. Le karst a tourelles de Phangnga s'eleve au dessus d'une plaine alluviale en bordure de mer. A Maros le karst structure par les nombreux corridors developpes a la faveur de dykes de roches eruptives presente une variabilite morphologique importante ; ses calcaires d'iige eocene-rniocene ont ete en grande partie degages de leur couverture volcano-detritique 2. Mesures de terrain et analyses en laboratoire Les prelevements sont effectues en flacon de polyethylene de 125cc Quatre parametres sont mesures sur place : la temperature (thennometre a mercure au l/5C ou numerique Quick-Novodirect a 0 1 ; le pH (pHrnetre de terrain Aqua Data ou Bioblock a electrode epox 'Y tampons pH 6 87 et 4 01 precision:+/0 01 pH) ; le TAC et le TH (coffrets d'analyse et procedure MERCK avec prise d'essai doublee (10cc au lieu de 5cc) et moyenne de 3 essais par analyse : AquaMerck Durete carbonatee ref: 8048 titrage acidimetrique pipette a 0 2d ; AquaMerck Durete totale ref. 8039 titrage complexometrique. pipette a 0 2d) Les mesures de pCO2 de !'atmosphere (cavite et sol) sont effectuees a la pompe DRAEGER (tubes reactifs: anhydride carbonique pour teneurs de 0 1 a 6% ou de 0 5 a 10%) La conductivite est mesuree a 20 C (Siem) Sont analyses au laboratoire : les cations Mg Na K par spectrophotometrie d'absorption ; Ca est obtenu par difference entre TH et Mg (ou chromatographie ionique) : les sulfates par turbidimetrie ou cluomatographie ionique les nitrates et la silice par colorimetrie et les chlorures par colorimetrie ou potentiometrie Sont calcules : HCO3 a partir du TAC: la mineralisation totale (somme des concentrations des solutes ): l'erreur de balance ionique B en% (difference entre la somme des cations et celle des anions en meq/1 di, isee par leur somme) ; la pCO2 en % (pression partielle equilibrante de CO2) ; le dpH difference entre le pH mesure et le pH a l'equilibre (pour une eau a chimisme calco-carbonate) Le calcul des equilibres chimiques est realise avec le programme SOLUTEQ au Laboratoire de Moulis (pCO2 dpH B. mineralisation) L imprecision de certaines mesures ou Jes differences entre methodes d'analyse qui n'ont pas ete toujours les memes de 1985 a 1996 peuvent masquer ou decaler les v ariations naturelles faibles et done rendre plus delicate !'interpretation de certains resultats 3. Comportement des variables La balance ionique est < 10% pour 90 % des echantillons mais < 5% dans 50% seulement des cas : elle est e n general 6 h Conferen c e on Limestone Hydrology and Fissured Media 61

PAGE 73

Tableau 1 H y drogeochimie : resultats anal yt iques d'echantillonnages de 1985 a 1996 dans le sud-est asiatique Code Date lieu Teau pH TH HCOJ Cl S04 NOJ Ca M<1 Na K SiO2 Minera pCO2 dpl-l B% c t ma/I mg/I ma/I mg/I mg/I mg/I mg/I ma/I mall ma/I % A2 I 06 07 85 TCD em ICM) 21.2 7 23 22 50 210 45 1.05 1 50 0 40 85 60 2 70 0 15 0 20 2.10 305 0 1.15 --0.01 12 54 AJ I 06.07 85 TCD T4 sph Tour (CM) 21 6 7 01 21 60 225 70 1.20 0 50 1.40 80.40 2 95 0 25 0 45 2.70 316 0 2.06 --0 21 7.37 A4 I 06 07.85 TCD T 10=St2 (CM) 22.2 6 92 34 80 408 70 1.05 0 50 0 10 135.00 2.55 0 20 0 20 3 60 553 0 4 50 0 13 1.67 A7 I 09 07 85 T Bf Bet (CM) 2 1 .6 7 41 7 85 89.06 1.10 2 50 0 08 26.20 3 10 2 50 1 55 12 80 139.0 0.33 --0.64 5.43 AS I 09 07.85 DCD ruisseau (CM) 20.9 7 47 11 25 134.81 1 20 2 50 0 15 35 60 5 70 3 90 0 90 10 40 195 0 0.43 --0 30 3 14 A11 16 07.85 TCD G1 sph Guano ICM 21.4 7 21 20.00 232 41 1.05 1.00 0 35 75 00 3 00 0 20 0 20 2 25 316 0 1 34 --0 03 1 85 A17 # 19 07.85 N Rue Takhean em ICM 23 5 6 66 43 40 464 82 1.10 1 50 0 10 152.20 13.00 0 35 0 24 5 40 640 0 9.37 --0 02 6 26 A23, 22.07.85 T. Pi em IMHS) 22 9 6 65 23.55 261 08 1.25 1 00 0 13 73 60 12 50 2 60 1.85 12.50 367.0 5 53 --0 54 5.78 A26 09 08 85 T Sam IPl 24 8 10 35 117 12 3.90 1.50 1 00 37 20 2 60 2 80 1.70 8 70 A28 11.08 85 T Phra ~m (P) 24.7 13.75 134 81 4 10 2 50 0.13 51 60 2 10 2.65 1 40 9 90 A29 19 08 85 T.Thong (P) 24 0 19.30 183 00 4 65 0 50 0 11 61.40 9 70 1 30 0 15 2.90 A32 23 07.85 TPM rivlere (MHS) 21 6 8 03 15.00 164 70 1 10 6 00 0 10 46 80 8 00 1 50 1.50 9.90 240.0 0 14 0 44 4 09 A33 I 23 07 85 TPM perte IMHSl 21 8 7.91 12.50 147.62 1.10 9.50 0 10 37.40 7.60 1.35 1.55 9.90 217.0 0 17 0.20 -1.18 A34 I 23 07.85 TPM af RD !MHSl 21 3 7.36 22 15 243.39 1 50 0 50 0 07 73.80 9 00 1.90 1.07 11.00 343.0 0.98 0.12 5.TT A36 24 07.85 TPM res (MHS) 21.8 7 28 19 45 221 43 1 15 7 00 0 10 63 00 9 00 1 30 1.45 9.70 315.0 1 09 --0 06 2 17 A38 # 29 07 85 TPM af gours RG (MHS) 22 0 7 07 39 30 456 28 1 10 0 50 0 10 142 00 9 30 0 15 0 87 4 10 616.0 3 52 0 34 2.31 A40 # 29 07 85 TPM af RG (MHS) 22 3 7 08 25.00 286 70 1 25 4.50 0 13 83 20 10 20 1 55 1.08 10.60 400 0 2.23 --0 04 2 63 B1 06 07.86 Lomp0batang IS) 17.5 7.42 6 15 21 80 2 10 5.75 24 30 0 15 1 75 2.70 17.20 75 8 0 08 -1 32 44 25 B3 07 07.86 Riv Bantlmuruna ISl 25 2 8 14 15 95 180 80 1 05 8.60 56 80 4 00 3 85 2 65 19 80 2TT.6 0 13 0 71 6 58 B5 09 07.86 GSK-K31Sl 24 6 6 93 18.90 217.20 1.20 5.70 69 70 3 50 2.30 2.25 14.80 326.1 2.64 --0 33 -3.04 B6 12.07 86 GBHD-G11S) 26.0 7.22 15 95 183 00 1 10 6.20 58 30 3 10 2.40 1 50 11.00 266 6 1 10 --0 16 2.33 B7 13 07.86 Towakkalak M dev.av (S) 25 0 8 10 15.95 185 10 1.05 8.80 56 80 4 00 3 75 2 60 21.00 283.1 0.14 0.68 2.28 B11 22.07.86 GSK-SA(S) 23 9 7 14 14 50 167 70 9.75 54 70 1 80 1 50 2.10 237 5 1 18 --0 33 0 78 B12 22.07.86 GSK-SB(Sl 25.0 6 96 16 30 187 30 10 75 58 20 4 00 4.15 2 70 267 1 2.02 --0 43 2.86 B13 23.07 86 GSK-SCISl 24 9 7.45 16 45 187.30 1 10 9 20 58 90 4 00 3 95 2 65 21 40 288 5 0.65 0.06 3 24 B14 25.07.86 Jamala em. (S) 25 3 7.08 16 85 191 70 1 80 9 65 62 60 2 65 3.85 2.40 18 10 292.8 1 57 --0 26 2 60 B17 05.08 86 Batu Lu bang fl. (H) 23 3 7 30 13.95 152 50 2 40 5 50 53 20 1.45 1 10 0 30 3.10 219 5 0 74 --0 22 2.73 B18 16 08 86 Tilanga em. (Sl 21 0 6 96 14 95 161 20 0 80 6 50 56 60 1 75 1 15 0 55 5.20 233.8 1 66 --0.55 4 06 B50 # 10.06 86 N Tok res. (Pl 18 7 7 41 15 60 180 20 3 80 13.25 29.90 19 50 2.90 1.40 16 40 267.4 0 63 --0 38 -1.21 B51 # 11.06 86 T Lot af IPl 23 9 7.82 23 30 183 00 4 20 13.25 63 80 17 50 1 70 0 25 3.80 287 5 0 26 0.42 16 72 B52 # 11.06 86 T Lot res IPl 25 1 8 04 9.30 113.30 3 75 17.75 21.40 9 50 2 75 2 00 20.00 190.4 0.10 0 02 -7.39 B53 # 11 06 86 T Phet res. (P) 26 0 7.75 6 45 78 40 4 10 20 15 18 20 4 50 3 25 1.85 18 30 148 8 0.14 --0 46 -11.02 B54 I 11.06 86 T Phet gour (P) 24 9 8 12 26 15 287 50 4 60 9.00 71 90 19 50 1 95 0 50 2 90 397 9 0 20 0.95 2 64 B58 17.06 86 Sal Yok Nol em (Kl 24.5 6.35 39.90 448 70 2 80 11.00 82 30 46.50 2.95 1 40 12.30 608 0 18 74 --0.60 2 85 B59 # 18 06.86 T Sal Yok Nol (Kl 24 5 6 25 40 45 462 80 2 90 10.75 91 00 42 50 2 95 1 40 13 20 627.5 24 32 --0.64 1.99 B60 # 19.06 86 T. Nam Erawan (Kl 21.3 6 53 37 25 413.80 2.80 6 80 98.80 30 00 2 60 0 90 12.00 567.7 11.01 --0.41 3.76 B61 20 06 86 D Pul-cascadelTl 21 6 4.60 0 45 14 20 1 15 6 55 1 60 0 05 4.00 2.00 24 20 53 8 39 65 -5 35 -12.16 B62 23.06.86 D lnthanon tourbiere 1, 12 4 3.20 0.65 7 60 0 50 6 00 2 10 0 30 0.25 0 65 5 80 23 2 ~ 31 ~.19 B63 23 06 86 D lnthanon Varichata (T 16 0 6 70 2.10 31 60 1 00 6 25 7 50 0 55 2 60 1 65 14 40 65 5 0 58 -2.38 ~.17 B64 # 26 06.86 T. Lot perte (MHSl 25 8 6 65 4 45 59 90 1 15 6 30 12 40 3 30 4 10 2.20 17.50 106.9 1 39 -1 82 --0 91 B65 27 06 86 T. Hud perte (MHSl 23.7 6 70 11 45 125 20 1.30 5 50 28.30 10 50 8 50 3 40 182.7 2 45 -1 17 10 87 B66 I 28.06 86 T. Nam Hu ema. IMHSl 23.2 6.60 31 70 370 30 0.55 9.10 75.90 30 50 4 20 1 75 15 10 507.6 8 67 --0 46 1 98 B67 # 30.06.86 T Plaa em (MHSl 23 2 6.60 18 60 223.20 0 50 6 25 60 10 8 50 2 45 1.70 13 40 316.1 5.36 --0 73 0.61 C1 23.06.87 T. Nam aval (MHS) 23.3 6.89 46 30 486.50 2 40 6 50 0 05 148.50 22.30 2 34 0 95 11 30 681 3 5 72 0 20 6 98 C2 # 23.06 87 T Nam af. RG (MHS) 24 0 6.81 43 10 473 50 2 30 6 70 0 10 135 30 22 50 2.35 0 80 11 10 654 6 6.TT 0 09 4 69 C3 23.06 87 T Nam amont (MHS) 23 3 7 34 37.20 401 80 2.20 6 90 0 05 112 30 22 20 2 57 1 00 13 40 562 1 1.69 0.47 5.61 C4 # 01.07.87 T N. Ru Hua Koa (MHSl 26 3 7.22 22.30 249 80 0 75 5 00 0 05 72.40 10 10 1 54 2.10 12 00 354 1 1 49 0 05 3.94 C8 22.07.87 TPC res IPl 25 6 8 11 9 20 98.80 4.75 6 00 0.35 26.80 6 00 4 35 3 40 10 50 161 0 0.08 0 15 5 69 C9 # 22.07 87 TPC af.1 IPl 26.0 7.26 16.30 158 60 5.55 2.00 0 10 47.60 10.70 2 45 0 14 1.60 228.8 0.87 --0.26 9 32 C10 # 22 07 87 TPC af 2 (P) 25 8 7 86 16 90 168 30 5 65 2 60 0 05 44.60 14 00 2 36 0 13 1 80 239 6 0 23 0 32 8 06 C11 # 19 07 87 TPC percolation (Pl 25 1 8 17 11 30 110 80 3 80 2 80 0 05 25 60 11 90 2 20 0 12 2 30 159 8 0 07 0 22 8 69 C12 22.07.87 TPC aval case (P) 25 4 8.1 8.00 91 20 4 70 7 40 0 65 24 00 4 90 4 70 3 90 13 50 155 3 0.07 0 06 2 97 C13 22 07.87 TPC amont case. (Pl 25 4 7 75 8 20 97 70 4.70 6 20 0 20 24 70 4.90 4 70 3 90 13.40 160 3 0.17 --0 25 2.00 C14 I 22 07 87 TPC perte RG IPl 25 4 7 46 7 30 82 50 4 95 7.50 0 35 21 30 4 80 5.15 5 00 15.70 147.7 0 29 --0 66 4.33 C15 22 07.87 TPC perte RD (P) 25 4 7 43 6 70 73 80 4.25 3.30 0 05 19.30 4.50 4 20 2.60 9 60 122 2 0 28 --O n 5.84 D1 23 07 88 Riv Sagea BL 1 (H) 23 6 7 69 10 80 110 80 3.20 1 90 0.07 16 30 16 50 1 40 0 80 21 20 172.4 0 22 --0 46 7 26 D2 26 07 88 Batu Lubang BL2 (Hl 23 4 8 06 14 60 154 20 2.60 1 00 0 31 57 40 0 60 1 12 0.66 1.65 219 1 0.13 0.56 6.58 D3 07.08.88 GSKS4(S) 25 3 7 28 15 70 171 60 2 20 8 10 0 05 52 80 6 00 4 50 2.90 24 00 271 4 0 89 --0 18 5 82 F1 27 07.90 GSK-sph extr amont (Sl 23 8 6 68 27 32 271 00 0 30 12 70 1 TT 93.00 9 80 1 50 1 36 12 40 405.2 5.39 --0.40 7 79 F2 27.07.90 GSK-sDh.amt.15 ao0t (Sl 25 2 6 30 16 25 176 00 0 44 4 20 tr 56 00 5 40 5 30 2 50 28.80 279.1 8 75 -1.12 8 42 F3 27 07 90 GSK-af Pterodac (S) 24.0 6 14 20 54 214 00 0 46 2 10 0 75 72 00 6.10 1 46 1 26 10.00 308.8 15.01 -1 12 7 88 F7 07 08 90 G Pisc i ne N3 (S) 29 8 7 94 22 32 239 00 0 76 4 20 0.04 73.00 9.60 8 00 6 00 21 20 362.5 0 28 0 79 10 35 F8 07 08 90 G Patta N7 (S) 28 8 7 38 19 29 217 00 0 26 1 90 tr 69 00 4 50 5 50 1 90 25.20 326 2 0.93 0 18 6 37 F9 16 08 90 Maroanglng TA& em (S) 24 1 7 83 19 82 206 00 0.86 0 40 0 04 73 00 3.80 3.25 0 65 11 20 300 2 0 29 0 56 9 34 H1 # 23 07 92 N Mut res (Kl 23 5 7 56 11 90 124 00 0 36 3.20 tr 39 80 4 80 1 00 1 10 8 30 181 8 0 33 --0 17 7 94 H2 # 04 08 92 N X i em Ill 29.4 6 87 30 80 330 00 0 26 2 65 1 68 116 10 4 30 0 70 0 70 7 00 464 1 4 53 0.04 6 01 HJ # 11 08 92 T Baa Nam Tip aour ITl 23 8 8 21 19 80 182 00 2 33 10 30 9 39 76 40 1 60 0.60 0 20 7 10 291 0 0 10 0 88 7 65 11 02 10 93 B Sangki (SM) 23 2 7.45 16 60 182 00 2 30 2 40 0 52 59 60 4 00 1 70 1 10 12.60 267 3 0 62 0.04 4.74 12 05 10 93 N Anta Bung (SM) 27 2 7 .5 3 3 40 56 00 0 00 1 80 0 00 12 40 0 70 6 30 2 00 29 50 106 3 0 17 --0.94 1 59 13 11 10 93 N .. Pelayangan (SM) 23 4 7 24 12 30 141 00 0 04 2 00 0 01 43 20 3 60 3.20 1 70 29 50 225 2 0 79 --0 40 5 26 J1 07 08 94 G.Sawi em (Sl 26 4 6 95 59 36 300 17 917 50 121 00 0.66 190 10 66 50 497 00 21 45 11 55 2125 4 2 94 0 05 5.55 J2 10 08 94 G. Kacic i riv. 3eme reg {S 26 3 6 79 28 48 302 99 1 80 19 40 0 00 96 20 10.80 3 50 2 90 30 50 468 1 4 80 --0 20 4 51 J3 11 08.94 S.Alorok -em. IS) 25 0 7 26 17 80 197 65 1 60 4 80 1 37 60 30 6 60 1 60 1 30 8 00 283 3 1 06 --0 09 3 57 J4 15 08 94 Towakkalak -em. ISl 25 0 8 06 17 67 196 35 1 50 8 00 0 65 57 70 7 90 4 40 2 60 21.60 300 7 0 16 0 67 5 00 L1 29 02.96 T Boum Bam em Ill 20 6 7 35 21 80 233 00 2 70 2 10 1.49 2 00 1 32 6 00 L2 03 03 96 T Kam Hom em (L) 22 0 7 93 17 30 193 00 4 20 3 10 3 64 3.47 2 13 4 00 L3 04 03 96 Hat Mouet Pan em Ill 22.0 7.54 19 50 209 00 1 30 1.70 0 22 0 39 0 23 5 00 L4 05 03 96 T Louk Kouang em Ill 22 8 7 75 22.50 235 00 1 00 7 70 tr 0 75 0 51 10 00 . . LS 1 08 03 96 Ban Tham Kuai em (L) 26 8 7 85 16 40 181 00 1 20 3 10 1 12 0 49 0 63 4 00 . L6 08 03 96 Koun Ban Nat em (Ll 24 0 7 08 25 50 290 00 0 62 . 0.71 0 32 2 00 . . L7 10 03 96 T Pia Seua Ill 25 7 7 46 13 00 143 00 . tr . 0 33 0 22 16 50 . Abrev i ations: Thallande Chiang Ma r Mae Hong Son Kanchanaburi Phangnga Sula w es i SuMatra Halmahera Laos emergence resurgence siphon affluent Rive Gauche Rive Droite reaard deversoir aval # : echantillon retenu pour ACP I I 62 Proceedings of the 12 th International Congress of Speleology 1997, Switzerland Volume 2

PAGE 74

positive : le deficit d'anions peut etre pour partie impute a une sous-estimation des sulfates (series C a L) Les temperatures sont representatives du climat: voisines de 21 a 22C dans le nord de la Thai"lande de 25 a 26C a Phangnga et dans le karst de Maros A !'exception des eaux non carbonatees pour lesquelles le pH descendjusqu'a 3.2 les valeurs sont comprises entre 6.14 et 8.17 Ces variations ne peuvent etre interpretees independamment des deux autres paramertres pCO2 et dpH Les plus faibles valeurs correspondent a de fortes pCO2 La mineralisation varie de 23 a 2125 mg/I avec une majorite des echantillons entre 100 et 400 mg/I TH et TAC suivent les memes variations Compte tenu des valeurs de pH les bicarbonates s'identifient au TAC: de 8 a 486 mg/I Calcium et magnesium associes au travers du TH, representent la contribution principale des cations a la mineralisation totale: 1.6 a 190 mg/I pour Ca et 0 05 a 66.5 mg/I pour Mg dont !'importance est liee a une dolomitisation partielle comme le confirment Jes rapports Ca/Mg. Sauf pour l'echantillon JI tout-a-fait particulier sodium et potassium sont en faible quantile (resp 0 15 a 8.5 mg/I et 0 12 a 6 7 mg/I) ; avec la silice (1.5 a 30 5 mg/I) ils representent la composante exogene du chimisme des eaux Les teneurs en chlorure faibles varient de 0 8 a 6 mg/I Les sulfates sont compris entre 0 1 et 20.15 mg/I. Les nitrates sont marginaux., en general < 1 mg/I. dpH et pCO2: pour Jes echantillons dont !'analyse est complete et significative les dpH varient entre -2.38 et 0 95, Jes pCO2 entre 0 let 24.3%. 4. Interpretation En dehors de l'examen direct des donnees diverses methodes ont ete utilisees pour degager Jes correlations entre variables et caracteriser Ies echantillons De fa9on generate on retrouve Jes fortes correlations entre HCO3 et Ca d'une part pH et pCO2 d'autre part directement liees a l'equilibre calco-carbonique et celles reliant Na K et SiO2 qui representent !'influence des sols ou des terrains non carbonates Ainsi le tableau 2 donne les valeurs des coefficients de correlation obtenus par analyse sur 30 sites de Thailande L'ACP correspondante permet d'expliquer 82 5% de la variance totale sur Jes trois premiers axes principaux Le premier (36 4%) oppose HCO3 et Ca a Na K. SiO2 et differencie la composante chirnique acquise et evoluant en surface (peripherie non carbonatee ou couverture detritique superficielle) de celle acquise a l'interieur de J'aquifere Le second (29 6%) oppose le pH et la pCO2 et represente done le facteur CO2 : ii conduit a differencier Jes eaux riches en CO2 provenant de la zone noyee (H2 Al 7 C2) ou de percolations lentes (A4 A38) des eaux plus pauvres d'ecoulements de surface allochtones (CI4 B64 B52 B53) ou de percolations rapides (Cl I) Le troisieme (16 5%) est lie a Cl et Mg : !'evolution du chirnisme et l'origine de ces deux ions etant a priori differente ii semble que cet axe puisse rendre compte de l'eloignement du domaine maritime: on retrouve en effet une diminution des teneurs en chlorure de Phangnga a Chiang Dao. pH HC03 Cl S04 Ca Mg Na K Si02 pC02 dpH pH 1 CXXJ HC03 -0 616 1 CXXJ Cl 0 483 -0 285 1 CXXJ S04 0 319 -0 245 0.419 1.CXXJ Ca -O f/J7 0.922 -0.376 -0.405 1 CXXJ Mg -0.375 0 557 0 237 0 275 0 23) 1.CXXJ Na -0 065 -0 359 0 368 0.400 -a ere 0 ~1 1.CXXJ K -0 121 -0 ~ 0 025 O :D:l -0.429 -0 041 0 666 1 CXXJ Si02 -0 204 -0 210 -0 094 0 575 0.410 0 148 0 666 0 747 1 CXXJ pC02 -0 732 0 719 -0.167 0 006 0 495 0 750 0 063 -0 013 0 141 1 CXXJ dpH 0.585 0 222 0.200 -0 018 0 352 -0 029 -0 588 -0 546 -a ere -0 205 1 CXXJ Tableau 2 Coefficients de correlation entre variables (sur 30 sites de Thai"/ande) Globalement en dehors d'eau.\: tres peu mineralisees de terrains non karstiques (BI. B61. B62 B63. < 80 mg/I) on peut distinguer trois groupes aux limites floues : Le premier (106 195 mg/I) correspond a des ecoulements d'origine allochtone qui se chargent peu a leur traversee du karst: pertes (A7 B64 Cl4) et resurgences (B53 B52. 12. HI DI) en general sous-saturees ou a des eaux de percolation rapide (Cl I) Le 2eme groupe (217 400 mg/I) est represente par des eaux de percolation plus lente moyennement mineralisees (B 17 B54. D2. H3) et la plupart des ecoulements de la zone de transfert horizontal (A2. A34. B6 B7 B14. B51. C9 ClO 11. 13. 13 ) : on y rencontre egalement quelques fortes pCO2 generalement associees a des dpH ncgatifs( Arn. B67 F2 F3) Le dernier groupe est constitue des eaux les plus mineralisees (405 a 681 mg/I): percolation lente. ruissellement de paroi ou gouttage de stalactites (A4 A38). eaux de circulations et emergences au niveau du karst noye (C2 Al 7. B59 B60 B66 FI. H2 12). presentant de fortes pCO2 associees souvent a la presence d'un epais couvert pedologique. Ce sont des eaux a temps de sejour long 6 1h Conference on Limestone Hydro l og y and Fi s s u red Med i a 63

PAGE 75

5. Discussion Sur Phangnga nous recoupons les resultats de PITMAN (1978) qui identifie 3 types d'eau dont des Jessivats de sol tres sous-satures a forte teneur en CO2 et faiblement mineralises: nos ecoulements de surface derivent de cette categorie par perte du CO2 excedentaire, diminution du dpH et enrichissement en carbonates. En lndonesie BALAZS (1968) analyse pres d'une quarantaine d e sources; nos mesures sur le karst de Maros et Sumatra sont voisines des siennes Des mesures conjointes de pCO2 de !'atmosphere ont ete faites au droit de sites preleves; elles conduisent au constat general suivant: a/ Jes plus fortes teneurs se rencontrent la ou de fortes pCO2 existent dans l'eau, et en presence d'un couvert pedologique e pais; b / a quelques exceptions pres, la pCO2 de l'eau est egale ou superieure a celle de !'atmosphere qui la surmonte; c/ lorsque ii existe un gradient mesurable, les pCO2 atmospheriques sont plus fortes pres de l'eau. On retrouve Ja differenciation regionale avancee par DEHARVENG (1986, 1987) entre les karsts a couvert pedologique Ii.mite ou squelettique {Phangnga, Maros), et ceux a couvert detritique epais (Nam Lang et Kanchanaburi) avec des valeurs de pCO2 plus fortes dans ces demiers ("cavites a CO2"). Aucune correlation avec la pluviometrie n'a ete notee. La comparaison avec d'autres donnees publiees, notamment par RENAULT ( 1979), montre qu'il n'y a pas de difference dans la gamme des teneurs mesurees pour une pluviometrie plus faible. Les travaux de CROWTHER ( 1983), menes sur des sols karstiques de Malaisie, mentionnent egalement !'influence determinante (mais non unique) de la couverture pedologique dans la disponibilite du CO2 au niveau du sol. 6. Conclusion L'eau, moteur de Ja karstification est aussi vecteur d'information, et par ce qu'elle transporte en solution elle garde la trace de son origine et de son histoire. Cet aspect est particulierement important pour l'approche des karsts de regions difficiles d'acces C'est ainsi que le suivi individualise d'une crue ou un echantillonnage extensif de sites representatifs permet d'apporter la base de donnees quantitatives qui fait trop souvent defaut a certains travaux et permettrait de nuancer certaines interpretations. Nos resultats restent a developper mais ii est possible d'en degager Jes points suivants: La mise en evidence de plusieurs types d'eau marquees par leur origine et leur histoire L'absence de differenciation par Jes caracteres hydrochimiques et en particulier le CO2 de ces karsts avec ceux des regions temperees. L'importance de la nature de la couverture pedologique dans la limitation de la diffusion vers !'atmosphere du CO2, rendant celui-ci plus disponible pour son transfert vers la base de l'aquifere par )'infiltration rapide. En ce qui conceme les implications karstogenetiques, qu'il n'a pas ete possible de developper ici ii semble enfin que rien ne permette d'affinner que l'evolution des karsts tropicaux de ces regions dont certains sont d'ailleurs des paleokarsts soit plus rapide qu'ailleurs References BAKALOWICZ, M. 1979. Contribution de la geochimie des eaux a la connaissance de l'aquifere karstique et de la karstification, These Doct. Sciences, Universite P et M Curie 269 p BALAzS, D. 1968 Karst regions in Indonesia Karst-es Barlangkutatas : 3-61. BROUQUISSE, F ; BAKALOWICZ, M. 1986. 16 Hydrogeochirnie. In Expedition Tha'i-Maros 85 APS Toulouse: 136143 BROUQUISSE F .; DALGER, D 1987. 8 Hydrogeochimie ln Expedition Thai'-Maros 86 APS Toulouse: 85-110. BROUQUISSE F. ; DALGER D. ; BAKALOWICZ M. 1988 Resultats hydrogeochimiques des expeditions Tha'i-Maros 86 et Thai" 87 In Expeditions de l'APS en Asiedu sud-est travaux scientifiques-1. APS Toulouse: 5-16 BROUQUISSE, F. ; DALGER, D .; DEHARVENG L. 1992 6 Sulawesi: resultats scientifiques In Expedition Indonesie 90 APS Toulouse: 83-94. BROUQUISSE F 1995. 7 Quelques donnees hydrogeochimiques et climatologiques In Expedition Sumatra 93 APS Toulouse: 45-52 CROWTHER, J. 1983 Carbon dioxyde concentrations in some tropical karst soils West Malaysia Cat e na 10 : 27-39 DEHARVENG L. ; BEDOS, A 1986 17 Gaz carbonique. In Expedition Tha"i-Maros 85. APS Toulouse: 144-152 DEHARVENG L. 1987 6 Nouvelles donnees sur le gaz carbonique des sol s et des cavites de Tha"ilande et de Sulawesi In Expediton Tha"i-Maros 86 APS Toulouse : 97-110 PITMAN J.I. 1978. Carbonate chemistry of groundwater from tropical tower karst in South Thailand. In Water Resources Research vol. 14 N 5 ( 1978) RENAULT, P 1979. Mesures periodiques de la pCO2 dans les grottes fran9aises au cours de ces dix demieres annees In Actes du symposium international sur l'erosion karstique Aix-Marseille-Nimes 10-14 sept. 1979 : 17-33. Remerciements La collecte et !'analyse de s d o nnees a ete realisee avec l'aide du Laboratoire S o uterrain du CNRS de M o ulis du Laborato i re d'H y drobiolo g ie de l'Universite P Sabatier de Toulou s e et de nombreux collegue s et ami s, en parti c ul i er M BAKALOWI C Z D D HULST C MUR, D DALGER L. DEHARVENG A BEDOS R BROUQUISSE D RIGAL Le. DEHARVENG P L EC LERC Me DEOTTO de la s ociete MERCK 64 Pr oc eed in gs of the 12 th Internat i onal Congress of Spe l eo l og y 1 997 S w i tz e rl and V olu me 2

PAGE 76

Simulation of the evolution of maze caves Torsten Clemens, Dirk Hiickinghaus, Martin Sauter, Rudolf Liedl, Georg Teutsch University ofTiibingen, Applied Geology Sigwartstr 10, D-72076 Tiibingen Germany Abstract The development of cave systems in carbonate rocks depends on a variety of boundary and initial conditions Among the cave systems two main types of geometries can be distinguished : the dendritic and maze pattern A numerical model has been developed capable of modeling the genesis ofkarst systems in complex geological environments It is applied to simulate the development of the above mentioned two different types of cave geometries. The results confirm that a prerequisite for the development of maze caves is evenly distributed recharge (White 1969 ). However more important for the development of maze caves is that flow through the system is restricted by an overlying less c onductive horizon e g a sandstone caprock Thus the feed back mechanism of higher flow rates leading to higher dissolution rates and therefore the preferential development of a small number of tubes does not dominate the evolution of the karst aquifer Thi s hydrauli c restriction furthers the development of other conduits also to achieve a significant diameter. Zusammenfassung Die Entwicklung von Hohlensystemen hiingt von vielen Anfangsund Randbedingungen ab. Zwei grundlegende Geometrien von Hohlen konnen unterschieden werden : dendritische Hohlen und Netzwerkhohlen Es wird ein numerisches Modell vorgestellt da s e s ermoglicht die Genese von Karstaquiferen rnit komplexen geologischen Strukturen zu simulieren Es wurde eingesetzt um die steuernden Faktoren die die Genese von Netzwerkhohlen bestimmen zu untersuchen Di e Ergebnisse bestiitigen daB eine Voraussetzung filr die Entwicklung von Netzwerkhohlen eine gleichmiiBig verteilte Grundwasserneubildung ist (White 1969) Wichtiger filr die Entwicklung von Netzwerkhohlen ist die Beschriinkung des Durchflusse s durch einen geringer durchliissigen Horizont wie z.B einen den Kalkstein iiberlagernden Sandstein Wiihrend in Systemen ohne iiberlagemde Sandsteinschicht der FlieBwiderstand der Rohren durch die Eiweiterung abnirnmt bestimmt in Systemen mit iiberlagernder Sandsteinschicht der kr-Wert des Sandsteines die FlieBgeschwindigkeit in den Rohren Die positive Riickkopplung der Zunahme des Durchflusses auf die Fliefigeschwindigkeit in den Rohren und die damit erhohten Losungsraten ist dadurch unterbrochen und Rohren mit kleineren Durchmessern werden schneller erweitert als Rohren mit groBeren Durchmessem : somit entstehen Netzwerkhohlen Introduction Tite mapping of more than 20 000 km of caves ( COURBON & CHABERT 1986 ) in the last decades shows that two dominant geometries of caves can be distinguished : a dendritic and a maze pattern (PALMER 1975 PALMER 1991 ). The dendritic pattern is characterised by conduits that converge in downflow direction whereas the maze pattern consists of a labyrinth of intersecting pas s ages. Th e genesis of a dendritic cave requires a strong selective enlargement of some of the numerous initial fractures of a karst aquifer. On the other hand labyrinth caves should be the result of simultaneous enlargement of fractures A conceptual model for the development of maze caves was introduced by WlilTE ( 1969) He found that limestones which are sandwiched in insoluble rock exhibit the typical network pattern WlilTE ( 1969 ) concluded that the lack of concentrated recharge from overlying beds causes the development of the maze caves in the Limestones PALMER (1975) distinguished different processes for the development of maze caves According to PALMER (1975 ) maze caves develop due to diffuse infiltration from an overlying insoluble caprock beneath isolated hills of limestone or by floodwater recharge The most of the investigated maze caves ( 75 %) are capped by thin permeable sandstone (PALMER 199 I ). Another theory of maze cave development states that mixing of thermal and meteoric water beneath a sandstone lead to the development of tube network s (BAKALOWICZ et al. 1987 ; FORD & WILLIAMS 1989) Besides from the conceptual models about cave genesis the enlargement of fractures was analysed mathematically C oupling the flow and dissolution kinetics WEYL ( 1958 ) c oncluded that the penetration distance of water to reach saturation is short in comparison to the length of the flow path in natural catchments This problem of enlargement of fra c tures far from the water inlet was solved by WlilTE ( 1977). WHITE ( 1977 ) concluded that due to the reduction in the dissolution rates c lose to equilibrium the fracture s will also be enlarged in the downgradient portion DREYBRODT ( 1988 1990 ) and PALMER ( 1991 ) developed one-dimensional numerical models which included the reduced dissolution rate close to the equilibrium These models were used to determine the sensitivity of various parameters for the enlargement of a single fracture A two dimensional tube network model for the simulation of laminar flow conditions was introduced by GROVES & HOW ARD ( 1994 ). The result s of the simulations show that a strong selective enlargement of initial preferential flow paths occurs The model results of HOW ARD & GROVES (1995) lead to a formation of maze caves if large initial tube diameters or high hydraulic gradients are specified A numerical model coupling the flow in the fissured system of a karst aquifer with the flow in the conduit s ystem wa s developed by CLEMENS et al. ( 1996 ). The model CAVE ( .C,arbonate Aquifer Yoid Evolution ) simulates the evolution of carbonate aquifers under a variety of boundary conditions and laminar as well as turbulent flow conditions In the study presented here the development of maze caves is simulated using a two layer model. The upper layer represents an insoluble caprock and the lower layer the limestone formation Theory Flow in carbonate aquifers occurs in a le ss conductive high storage fissured system as well as in the highly conductive low 6 th Conference on Limestone Hydrology and Fissured Media 65

PAGE 77

storage conduit system In order to calculate the flow in carbonate aquifers the flow is separately calculated in the smaU fissures and in the conduit system Toe flow in the fissured system is simulated using a continuum model and the flow in the fractures with larger apertures by a tube network model. Both systems are coupled by a linear exchange term. Toe tubes are enlarged by dissolution of limestone Thus the hydraulic heads in the system must be calculated after every time step in order to obtain the new flow velocities in the tubes The flow in the fissured system is described mathematically by the groundwater flow equation (HUY AK.ORN et al. 1983) _!___(rohc)+_!___(rohc)=sohc -y -r (1) o x ox oy oy 0 1 where h is the hydraulic head in the continuum, T and S are the transmissivity and storage coefficient respectively y is the volumetric rate of fluid transfer from the fissured system to the pipe network per unit area and r is the groundwater recharge per unit area For the calculation of the flow in the tube network Kirchhoff s rule is applied, stating that the sum of inflow and outflow at any node i of the network is zero (2) : n""' In 0= IQu+ IQ u +R ; +r; (2) j=l j=l with R, the direct recharge at node into the pipe network and 11 .'.,,, and n,'"rb the numbers of the tubes with laminar and turbulent flow connected to node i Qy is the flow through the tubes connecting nodes i and j and can be either laminar or turbulent. For laminar flow the Hagen-Poiseuille formula is used, while turbulent flow is calculated using the Colebrook White equation. Toe exchange flow rate f; between the fissured system and the tube network is calculated by (3) where O.o.j is the exchange term, h, is the hydraulic head in the tube network at the node i and h ,., is the hydraulic head in the continuwn at the node i. Calcite dissolution far from equilibrium has been investigated by a number of authors (e.g. BERNER & MORSE 1974 PLUMMER & WIGLEY 1976) BUHMANN & DREYBRODT (1985) showed that the dissolution far from equilibrium can be approximated by a first-order rate law However, c lose to the equilibrium the dissolution rate is reduced significan tly and can be described by a fourth-order rate law (PLUMMER & WTGLEY 1976 SVENSSON & DREYBRODT 1992) The dissolution rate R is given by R = kn (Ceq c)" (4) where n is the reaction order c.,, is the equilibrium concentration of calcium and k.,, is the rate constant for the first order or fourth-order reaction rate For laminar flow the value for k 1 oaml is corrected in order to account for the diffusion of the species into the sol ution (DREYBRODT 1990) Toe flow of the water directly into the conduits is given by the vertical leakage and the head difference between the two layers. l11e vertical hydraulic conductance is multiplied by the dimensionless factor Yilir (here 5 10-') in order to determine the part of the water flowing directly into the conduit system This factor accounts for the lesser cross section for the water flowing directly into the conduit system than from the upper into the lower layer. Model Structure In order to model the evolution of a maze cave a rectangular model domain discretised into two layers, was used (Figure I). The modeled domain measures 1050 m by 1350 m. The hydraulic conductivity of the upper layer is 105 m/s Toe vertical leakance V co nt is 5 107 s- and a constant recharge of 400 mm/a is applied. Titree sides of the upper layer are no-flow boundaries and at one side a constant head boundary is specified, representing a river Iltis configura tion means that the water input as well as the water output are affected by the sandstone caprock. Fig11re 1: Two layer model with the Limestone in the lower and sandstone caprock in the 11pper layer The co nstant head bo1111dary is specified in the 11pper layer, i.e. the water which flows through the Limestone must cross the sandstone before it reaches the river A regular network is assumed in the lower layer (Figure 2) The initial diameter of the tubes conneting row 9 10 and I 1 (thicker lines in Figure 2) is I mm and 0 4 mm for all other tubes Toe larger initial diameters were used in order to investigate whether an initial preferential flow path leads to the development of dendritic caves. Toe transmissivity of the lower layer is l 04 m 2 /s, the thickness of the lower layer is JO m and the coefficient of the linear exchange is equal to 105 m 2 /s. For the lower layer four no-flow bo1mdaries are specified 0 C 1, ,( I < le I I'\. / I '\ I/ I'll,. / I'\. I/ I'. I/ I/ I X I'. I V / i)( ..... I JI IJ ll I V I/ IX I .J< I/ I)( I'I A I'. I/ I'\. X X / I'. I/ i)( I/ I'. [)( X / I's I/ illf: "' ,. ll / I 'I/ I)( I"\ I/ 1 'I)( I)( I/ I'\ I/ l >C I'I/ I 's I)( I)( I/ I"\ I/ "' no flow IA IA IA T ,. I/ : V I I's V I I's VI X X )( I I/ 1 '/ I 's V I'. i)( XI X I "> I / I'\ I/ ;', I/ I'\ I / )( )( X I'\ I/ I 'V I'\ I/ I\ Ill Ill llil II , ... i...i L.,11 " I'\ I/ 1 V I'. I/ I'IX !X I ,x "> I / I'\ VI I'\ I/ ' / IX I i)( I X I's. I/ 1 'V I'. V I'. I JO: lJOi IX I'> V I 'I/ I'I/ I'. 'c/ I)( I IX IX I's. lL I 'V I 'V I I"' V I IV no flow Figure 2: Geometry of the com/11it system in the lower layer uf the model domain. The i11itial diameters uf the tube s co1111ecti11g row 9, JO a11d 11 are I mm am/ of the remaini11,: t11bes 0 .J mm. 66 Proceedings of the 12 th International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 78

It is assumed that the water within the fissured system is slightly undersaturated (c=0 9 ceq) Toe water entering the conduit system from the upper layer is assumed to have a concentration of O mol/L of calcium Higher concentrations of the water entering the conduit system from the upper layer or from the fissured system lead to longer periods of time necessary for the development of the conduit system but the geometry of the conduit system develops in the same way as for lower concentrations Toe equilibrium concentration of calcium is C eq = 2 10 3 mol/L the kinetic rate constants far from the equilibrium ( n = l ) are k1(lam ) = 2 5 10 5 m/s kl(111r ) = 5 10 5 m/ s Close to the equilibrium the kinetic rate constant is (n=4 ) k.i = 1.2 10 13 cm 1 0 mor 3 1, as suggested by DREYBRODT ( 1990 ) Toe change in the kinetics from the first-order to the fourth-order rate law is assumed to occur at c/ c eq = 0 9 The recharge data and the hydraulic parameters for the lower layer are derived from a karst aquifer syste.JU in SW Germany ( SAUTER 1992 ). Results and Discussion The results of the simulations are depicted in Figures 3 and 4 Initially the tubes with the larger initial diameters are enlarged more rapidly than the other tubes (Figure 3) After 5500 years the enlargement rate of the tubes with the small initial diameters is higher than the enlargement rate of the tubes with the larger initial diameters This set up leads to the development of a maze c ave system as shown in Figure 4 The reason for this type of development is the low hydraulic conductivity of the overlying layer This causes that the positive feed back mechanism of the e nlar ge ment of the tubes leading to higher flow velocitie s and accordingly higher dissolution rates not to take place L.. (I) 0... 0... ::J C en 0) C c 0... en 100.00 300.00 500.00 700 00 900.00 1100 00 Figure 3: Geometry of the conduit system (lower layer) after 2500 years The thicker lines indicate tubes with a diameter > 0 002 "' 100 00 100 00 300 00 500 00 700 00 900 00 1100 00 Figure 4: Geometry of the co11d11it system (lower layer) after 5875 years. The thickest lines i11dicate tubes with a diameter> 0 02 m In order to investigate the importance of the feed ba c k mechanism addressed above a simulation is performed using a one-layer model i e without the caprock In this model configuration aU the recharge infiltrates into the limestone i e there is no surface runoff The geometry of the conduit s ystem is shown in Figur e 2 except that a constant head boundary is applied on the left hand boundary Recharge is distributed uniformly whereby l % of the total recharge is injected dir ec tly into the conduit system For the above described model run with the caprock runoff is allowed and direct recharge increases from 0 004 % at the beginning of the s imulation to 0 5 % at the end of the simulation depending on the drainage capa c ity of the conduits The aquifer in the simulation without th e c aprock i s a s sumed to b e un c onfined and the c o e ffi c i e nt of the Ii.near exchange term i s equal to 0 0001 m 2 /s. The re s ult s of th e model run are displayed in Figur es 5 and 6 The tubes with the higher initial parameter s are enlarged mor e rapidly than the tubes with smaller initial diameter s. Due to th e highest hydraulic gradient in the tube adjacent to the constant head boundary this tube is the first one to experien c e turbulent flow c onditions (/l Cl C c a. (/l 200 oo aoo oo 600 o o eoo oo 1000 oo 1200 oo Figure 5: Geometry of the co11d11it system and heads i11 th e fissured system after 2000 years The thicker li11es s how tube s with a diameter> 0. 02 111. 6 th Conference on Limestone Hydrology and Fissured Media oT

PAGE 79

200 oo oo oo soo oo ao o oo 1000 oo 1 200 oo Figure 6: Geometry of the conduit system and heads in the fissured system after 2500 years. The thickest lines show tubes with a diameter> 0.2 m. The low flow resistance of the enlarged tube causes the highest hydraulic gradient to be obsetved in the adjacent tube in upgradient direction In that way the conduit system evolves in an upgradient direction The continuous enlargement of the tubes change the flow field and lead to the development of a distinguished drainage area for the conduit. The results oftl1e model indicate tliat the unifonn distribution of recharge beneath insoluble caprock is a prerequisite for the development of maze caves as suggested by W1-IlTE (1969) and PALMER ( 1975). However we believe that even a more important factor could be the disrupting effect of the low conductive caprock on the flow-dissolution feed back mechanism Further simulations will be conducted to analyse this in more detail. Conclus i ons The genesis of maze caves was investigated using a numerical model. Tue numerical model integrates the dissolution kinetics of calcite as weU as the flow in the fissured and conduit system of a karst aquifer. A model domain consisting of two layers was chosen whereby the upper layer represents an insoluble caprock and the lower layer the limestone aquifer The model results show that for these conditions maze caves evolve due to the low hydraulic conductivity of the caprock The low hydrauli c c onductivity of the caprock disrupts the positive feed back mechanism of the enlargement of the tubes on the flow velocitie s Simulations with a single layer model allow the feed back mechanism to occur Therefore the enlargement of the tubes leads to a reorganisation of the flow field and the evolution of a dendritic cave system Acknowledgements This study is supported by the German Research Foundation ( DFG ) as a part of the Collaborative Research Centre 275 ( Sonde1forschungsbereich 275 ) and by the European Union ( Contract CEC EV5V-CT94-0471 ). The autllors would like to thank Prof Dreybrodt (U niversity of Bremen ) for helpful c omments on the manuscript. R e ferences BAKALOWICZ M ., FORD D C MJLLER T.E. PALMER, AN and M.V. PALMER 1987 Thermal genesis of solution caves in the Black Hills South Dakota G e ol. So c A m B u ll 99 : 729738 BERNER, R.A. and J W MORSE 1974 Dissolution kinetics of c alcium carbonate in sea water JV theory of c alcite dissolution A m J Sc i e n ce 274 : 108-134 BUHMANN D and W DREYBRODT 1985 The kineti cs of calcite dissolution and precipitation in geologicaUy relevant situations ofkarst areas 2 Closed systems C h e m. Geo l 5 3: 109-124 CLEMENS T. HUCKINGHAUS D SAUTER M. L[EDL R. and G. TEUTSCH 1996 A combined continuum and discrete network reactive transport model for tlle s imulation of karst development. !AH S Pu b / no 2 3 7 : 3 093 18 COURBON P and C CHABERT 1986 Atla s d es grande s c avites mondiales U n io n Jnt e rn a t io nal e d e S p e l eolog 1 e. Federation Francaise de Speleologie : 22 5 p DREYBRODT W 1988 Proces s e s in karst s y stem s. S pn nge r B e rlin H e id e l be r g New Yo rk : 288 p DREYBRODT W 1990 Th e rol e of di ss olution kin e ti cs i n th e development of karst aquifers in limestone : A mod e l s imulation ofkarst evolution J. Ge olo gy 98 : 6 39 -65 5 FORD D C and P W WILLIAMS 1989 Karst geomorpholo!, 'Y and hydrology U n wi n H y man Lo nd o n : 60 I. GROVES G G and AD HOWARD 1994 Early dev e lopment of karst sy stems : I. Preferential flow path en l arg e ment wider laminar flow H u t e r H e.w 11r l? es 3 0 ( 10 ) 2 8 3 72 84 6 HOWARD AD and C G GROVES 1995 Early developm e nt ofkar s t systems : 2 Turbulent flow lf'at e r l? es l?esearch 3 1 ( I ): 19-26 HUYAKORN P S .. LESTER B H and C.R FAUST 198 3 Finite element techniques for modeli.ng !,JJ oundwat e r flow in fractured aquif e r s. Wat e r R eso 11r R es 1 9( 4 ) : 101 9 -10 35. PALMER AN 1 9 7 5 The origin of maze c ave s \a/ Speleol. Soc Bull 3 7 : 5676 PALMER AN 1991. Origin and morphology of lim es ton e caves Ge ol. S o c A m Bull. 103 : 1-21. PLUMMER L.N and T.M L. WIGLEY 1976 The di ss olution of calcite in C{)z-saturated solution s at 2 5 C and I atm total pressure Geoc h i m e t C os m oc him Ac t a 40 : 1 9 12 0 2 SAUTER, M 1992 Quantification and forecasting of r egi onal groundwater flow and transport in a karst aquif e r ( Gallusquelle Malm SW Germany ) Ti i b in ge r Geow i sse n sc ha.ftli c h eA rb e il e n C 1 3 : 150 p SVENSSON U and W DREYBRODT 1992 Di ss olution kinetic s of natw a) c alcite m i neral s in C {)z -wat e r sys t e m s approaching calcite equilibrium C h e m Geol. 100 : 1 29 -14 5. WEYL P K. I 958 Tue solution kinetics of calcit e J. Ceo/ 6 6 : 163-176 Wl-IlTE W B 1969 Conceptual models for carbonate aquifers G roundwat e r 7 : 1 5 -21. WHITE W B. 1977 Role of s olution kineti cs in th e development of karst aquifer s Inf e rn a l Assoc Hydrogeol Memoir 12 : 503-517 68 Proceedings of the 12th lntamational Congress of Speleology 1997 Switzerland Volume 2

PAGE 80

Periodic and aperiodic forcing of water flow through sodastraw stalactites (Choranche Vercors France) Jea n -L u c D estom b es1, Michel C ord onnier1, Jean-Yves Gadat 1 and Jean Jacques De l a n noy 2 I) Laboratoire de Spectroscopie Hertzienne, URA CNRS 249 Centre d'Etudes et de Recherches Lasers et Applications Universite des Sciences et Technologies de Lille 59655-Villeneuve d '. Ascq cedex, France 2) Institut de Geographie Alpine, URA CNRS 903, Universite Joseph Fourier 17 rue Maurice Gignoux, 38031-Grenoble France Abstract During the last few years speleothems have appeared to be of increasing interest for high resolution paleo-en v ironmental applications Among the many parameters which govern the speleothem growth rate the drip water discharge is of The uppermo s t importance Since May 1996, high temporal resolution flow rate measurements are carried out by the drop counting method on five stalactites located in the Choranche caves (Vercors, France) Even if the recording period is still too short for characterising long term evolution, recordings already show interesting periodic variations of the flow rate of 2 sodastraw stalactites Preliminary Fourier analysis clearly show the existence of at least two major forcings with diurnal and semi-diurnal periods, confirming recordings obtained more than 20 years ago in a small cave of the same area Possible relations with earth and atmospheric tides are discussed A preliminary analysis of a spectacular stalactite swelling is also presented Resume Au cours des demieres annees Jes speleothemes sont apparus comme des outils susceptibles de foumir des informations paleoenvironnementales a tres haute resolution temporelle Parmi Jes nombreux facteurs qui gouvement leur taux de crois s ance le debit de l alimentation en eau est clairement de prerniere importance. Depuis mai 1996, des mesures de debit sont menees par comptage de gouttes sur 5 stalactites des Grottes de Choranche (Vercors France). Meme si la periode d'etude est encore trop courte pour une etude des tendances a long terme les enregistrements mettent d ores et deja en evidence des variations periodiques interessantes L analyse de Fourier fait clairement apparattre une modulation periodique diume et semi diume du debit de 2 fistuleuses confirmant des observations faites ii y a plus de 20 ans dans une cavite de l a meme region. Les relations possibles avec J es marees terrestres et atmospheriques s ont discutees Une analyse preliminaire d une spectaculaire crue de stalactites est egalement presentee 1. Introduction There is currently a growing interest in the study of the speleothems considered as a memory of the paleo environmental conditions over the continents Speleothem abundance is indeed closely related to climatic and environmental factors such as temperature, water availability, atmospheric and soil CO 2 concentration, presence of an active vegetation cover and insolation (GASCOYNE, 1992 ; HARMON et al ., 1978; HENNIG et al ., 1983 ; MAIRE & QUINIF, 1988; GORDON e t al., 1989 ; LAURITZEN et al 1990 ; BASKARAN & KRISHNAMURTHY, 1993; SZABO et al ., 1994 ; GOEDE, 1994 ; BAKER et al 1995 ; GENTY & QUINIF 1996 ; LAURITZEN 1996a, and references therein) Moreover KASHIWAYA et al (1991 ) showed unambiguously that global climatic changes have been recorded in the speleothem growth in the United Kingdom. As part of the search for a transfer function between the environmental conditions and the physico-chemical climate related signals potentially recorded in speleothems (growth rate, morphology and petrography isotopic ratios, tephras dust pollen content, concentration of trace elements and organic compounds, . ), a good knowledge of the depositional mechanisms of the modem underground calcite formations is clearly needed Among the various parameters which govern the growth of a speleothem and depend directly on the climate, the amount of seepage water available is of first importance. To get insight in the growth of modem speleothems the water flow rate of 5 stalactites is presently studied with a high temporal resolution in the Coufin cave (Choranche, Vercors, France ) by the technique of drop counting This cave is indeed a very favourable site for the observations of alpine karstic phenomena and several studies of the present and past calcite deposits are currently under way by various spectroscopic methods ( see two other communications by PERRETTE et al at this Colloquium ). Comparison between microscopic ( fis s ure network) and macroscopic ( rivers ) hydrology will be possible since a measurement station (pH, water resistivity, river discharge, temperature) is now implemented in the cave and will be soon completed with a meteorological station installed on the surrounding plateau After a short description of the Coufin cave we will present the method and the experimental set up The analysis of the preliminary results will be focused on a spectacular stalactite swelling and on diurnal and sernidiumal variations The s e latter ones appear during the summer period on 2 sodastraw stalactites confirming the observations made more than 20 years ago in a small cave of the same geographical area 2. Presentation of the Coufin cave The touristic Coufin cave opens in the NW part of th e Vercors at an altitude of 581 m down below the Cirque d e Choranche cliffs, a vast amphitheatre formed of urgonian limestones, 200 to 300 m thick surrounding the imperme a ble hauterivian marls The visited part of the Coufin cave is only a small part of one of the two major cave systems which drain the Massif des Coulmes anticlinal the Goumier system and the Coufin-Chevaline system with total developments of more than 29 km and depths of 411 m This massif of moderate altitude (around 800 m), covered with oakand beech groves is characterised by a conekarst related to a hot and wet pliocene period (DELANNOY et al 1988; DELANNOY 1991 ) It has been deeply cross cut by the Bourne river ( present altitude 300 m ) 6 1 h Conference on Limestone Hydrology and Fissured Media 69

PAGE 81

during the quaternary era. The underground networks develop at the contact between the urgonian limestones and the hauterivian marls Recent studies (DELANNOY, 1997) have shown the existence of two types of networks a suspended one that the lower galleries are contemporary with the quaternary climatic episodes while the vast higher galleries are related to a paleo-drain of pliocene age (DELANNOY et al., 1988 ; DELANNOY, 1991). The amazing diversity of speleothems of the Coufin cave (especially a great number of delicate sodastraw stalactites up to 3 m long) were formed during quaternary interstades Owing to the huge number of available stalactites, choosing five of them for the present study appears to be like a poker stroke Nevertheless two sodastraws with moderate flow rate at the time of the decision were chosen according to the previous 1969-1971 summer records which were carried out on this type of stalactite and showed up clear semi-diurnal oscillations (figure I ) The three others are more conventional stalactites which are associated with stalagmites and /or flowstones. They are all located between 20 and 90 m from the cliff i e. within the decompression zone 160 T(s) 150 140 130 120 110 100 90 Time (h) 12 1 8 24 6 12 18 24 August 'f' 1970 August 811i, 1970 Figure 1: Example of the se m idiurnal oscillatio11s observed 011 the flow rate of a sodastraw (Grotte du Sourcier, Sainte Eulalie -e11Royans, Drome) during su mmer 1970. T (ill s) is the mean time between two drops. T is then i11versely proportio11al to the flow rate. Stars indicate the st ro11g perturbations induced by nearby thunderstorms, which were dr y at the cave location. 3. Principle of the method and experimental set-up The easiest way to measure the temporal variations of the stalactite flow rate with a high temporal resolution is to measure the time interval between two successive drops. This is the technique which one of us used for obtaining the 1969-1971 recordings (figure 1). lt has been later used in the Coufin cave (DELANNOY, 1986) and is currently working elsewhere (LAURITZEN, 1996b ; GENTY et al 1996). The fundamental requirement of this method is that the volume of the drops is independent of the flow rate The problem of determining the volume of a pendant drop has been the subject of many investigations. Empirical and semi theoretical formula giving the volume of primary drops detaching from tubes in a quasi static experiment are available (TATE, l 864; RAYLEIGH, 1899 ; HARKlNS & BROWN, 1919 ; FORDHAM, 1948; GARANDET et al., 1994). However, the study of syngenetic origin with paragenetic one still active. ofthe sediment big dimensions, and a smaller Morphological studies and dating fillings show of the dynamics of drop formation ( i .e. when the flow rate is not vanishingly small) is much more recent and is still actively investigated both experimentally (PEREGRINE et al., 1990 ; SHJ et al 1994) and theoretically ( SCHULKES 1994 ; EGGERS 1995 ; ZHANG & BASARAN 1995 ; and references therein ). From these works, the formation of a drop can be described a s follows : during the growth process the volume of the drop hanging from the stalactite increases according to the flow rate of the feeding water and the geometry of the drop slowly evolves to pear-shaped. Once a critical volume is exceeded the evolution becomes much more rapid and just before the bifurcation, a spherical drop connected to the water remaining in the stalactite by a conical thread is formed The rapid increase of the length of the thread leads to a break which releases a free drop called the primary drop However one or more satellite drops can be produced by dynamical effects as the remaining water filament retracts after the rupture The details of these events critically depend on the water flow and at very high flow rate chaotic behaviour can even be observed ( SHA w, 1984 ; Wu et al 1989 ; DREYER & HICKEY, 1991 ). The dependence of the drop volumes with the flow rate has been experimentally studied by ZHANG & BASARAN ( 1995 ). According to their results, the volume of a primary drop of pure water is nearly independent ( within 5 %) of the flow rate as soon as this one is lower than about I drop/s. In these conditions the volume of the satellite drops is less than 2 10-3 of the primary drop volume. It can be then considered that the measurements are quantitative with an accuracy of 5 %. It means that the measurement of the time interval between two drops could be converted into flow rate measurement. However as the volume of a drop depends on the unknown diameter of the stalactite tip (TATE s law), it could be measured only by volume calibration This has not yet been done and we present directly the mean period T versus time, T being of course inversely proportional to the flow rate Drops are detected by their impact on detectors positioned under the stalactites of interest at a distance from about half a meter to several meters Each sensor consists of a piezoelectric gauge measuring the shock-induced deformation of a thin circular plastic membrane mounted on top of a flat hollow watertight container. These detectors are suitable for remote measurements because they do not require any power supply and they are insensitive to the specific environment of a cave Output signals of the order of I 00 mY are directly transmitted over -100 m through ordinary cables to a distant acquisition system located in the cave itself, in a wood box close to a mains plug After amplification, the damped sine wave delivered by a detector after a drop falls on it s surface is converted into a TIL pulse as soon as its amplitude becomes higher than an adjustable threshold defining the sensitivity of the detection This sensitivity is adjusted according to the strength of the drop impact which depends on the height of the fall. Changes in the logic state are detected using the centronic s port of a personal computer (PC). The P C printer port is indeed an inexpensive platform providing eight TIL outputs five inputs and four bi-directional leads for implementing projects dealing with peripherals The signals delivered by the detectors are applied to the inputs which correspond to the status lead s from the printer (BSY /ACK, PE, SELECT, /ERROR) and can be fetched by reading the five most significant bits of the status port. Note that TTL outputs can be used to extend the number of 70 Proceedings of the 12 1 h International Congress of Speleology 1997, Sw i tzerlandVolume 2

PAGE 82

detectors by multiplexing the inputs. A program has been developed to register each event and to evaluate the mean period T between two drops, averaged over a period of time ~t predefined by the user. The acquisition system has been carefully designed to avoid the loss of data in case of mains failure Data are referenced in time stored on a hard drive and regularly gathered into a distant database For this work a PC 386SX16 was handling five detectors at an individual maximal counting speed set to -5 drops/s 4. Data analysis Owing to the principle of the method the averaging time ~t has to be chosen according to the lowest flow rate expected for the stalactites First chosen as 3 min, it has been for example necessary to increase it to 6 min to take into account the drying during autumn In any case, after some months of recording, such short averaging durations produce a huge number of data points which are handled with a home made database and a commercial grapher. Owing to the lack of place we do not give here the whole recordings which will be presented on our poster and we focus on two particular phenomena clearly observed in our recordings : the transient response of the stalactites following a swelling and a periodic modulation of the flow rate Transient response following swelling On 7 111 and 8~' of July 1996, very heavy rains were observed on the Vercors and caused important floods and damages especially on the Bourne river but also in caves like the Gouffre Berger the Trou qui Souftle (LISMONDE, 1996 ), or the Coufin cave itself Unfortunately the river discharge measurement station was not yet operating at this time On July 7 t h the Autrans and Rencurel meteorological stations, distant from the Coufin cave basin by not more than 18 km and situated at about the same altitude recorded 168 and 115 4 mm respectively July 8 lh still showed important precipitations with 58 and 72 5 mm respectively These two highly rainy days followed three days of moderate rain (a total of 37 and 24.3 mm respectively for the July 4'\ 5 111 and 6 111 ) which themselves followed a relatively dry period (12.4 and 13 7 mm respectively during the 12 previous days) (figure 2 ) A completely dry period started after these two rainy days and lasted for more than two weeks. Such a situation where heavy rains are concentrated on a short period of time and are followed by a dry period is clearly ideally suited to study the response of the fissure network to a pseudo transient excitation. All the 4 sensors running during this period clearly detected the arrival of a swelling. The behaviour of the two sodastraws ( stalactites 3 and 5 ) was especially impressive with a very fast increase of the water flow ( figure 2) Unfortunately, sensor 3 was out of order immediately after the beginning of the swelling because it has been shifted by an important river overflow. In the case of stalactite 5 the swelling feature appeared as a strong dip in a regularly increasing curve. Fitting of this curve showed that the long term flow rate evolution was very well represented by an exponential law with a time constant of 26.4 days which then represents the characteristic time for draining the fissure network which feeds this stalactite. This is the result expected for the draining of a reservoir in a laminar regime. The onset of the stalactite swelling was very sharp and the transition from normal flow to maximum flow rate, which follows approximately an exponential law appeared to be especially short about 8 hours The maximum delay between this onset and the beginning of the rainy days was about 40 hours the uncertainty coming mainly from the low temporal resolution of the rain data (I day). The return to the normal flow rate was also well timely definite and lasted less than 5 hours The total duration of the high water flow rate was about 30 hours, shorter than the rain duration and also very short compared to the characteristic draining time. Ts{s) ____ h (mm) 60 350 50 40 30 20 10 300 250 200 150 100 140 150 160 170 180 190 200 210 Day (1996) Figure 2: Recording of the flow rate of sodastraw 5, from May 2 8'" to July 2 4'" I 996 showing the swelling feature of July 8'". T 5 (in s) is the mean time between two drops, averaged over 3 minutes. T 5 is the11 inversely proportional to the flow rate. Rain data (h i11 mm) are from the Autra11s meteorological station. The recordings of the two other stalactites also unambiguously indicated the time of the swelling : the flow of the stalactite 4 already very fast (T 4 = 0.5 s) became nearly continuous within less than JO hours and stayed at this high discharge for about 3 days Stalactite I recording definitely showed a discontinuity at the same time but with an amplitude which did not exceed that of the other discontinuities which seem to be a characteristic of this stalactite Note that such discontinuities are actually observed more or less often in all the recordings ( see for example sensor 5 on figure 3 ) ; they are characterised by very steep rising and falling edges which occur in less that the averaging time i.e less than 3 minutes Semidiurnal and diurnal variations Similarly to what is observed in some types of aquifers (Rm1 et al ., 1991 ) the response of the water flow through karstic microfissures under periodic forcing can contain informations on the fissure network properties. As shown on figure I such a modulation of the flow rate of a sodastraw were obvious in the summer recordings of 1969-1971 ( DESTOMBES 1971 ). The period was close to 12 hours and the variations could represent as high as 40 % of the mean flow rate At that time however the duration of the recording period (< l month) was too short to allow any grounded interpretation Figure 3 shows that such periodic variations are also observed on the two sodastraw of the Coufin cave Howe v er their amplitudes are much weaker in the range of 3-6 %. After baseline substraction, a Fourier analysis was performed on the two sets of data and gave evidence for two major forcings with 6 '" Conference on Umestone Hydrology and Fissured Media 71

PAGE 83

periods of 23.8( 0.60) hours and 12 02( 0.20) hours, with an amplitude ratio of 2 3 This result contrasts with the 196971 T (s) 18 16 14 12 10 8 6 4 2 0 210 215 220 225 230 235 240 Day(l996) Figure 3 : Periodic variatio11s of the flow rate of the two sodastraws, from July 28' h to August 27' h 1996. To allow the two recordings to be put 011 the same figure, a11 ordi11ate tra11slatio11 has bee 11 applied as followi11g : T, = T + 26 s a11d T 5 = T + 46 s. observations since in this latter case the main forcing was definitely semidiurnal. The low accuracy of the period determination is mainly due to the relatively bad signal/noise ratio of the periodic signals and to the limited duration of the period where the oscillations can be observed According to the data of stalactites 3 and 5, these oscillations were indeed recorded only during about one month mainly in August (figure 3), the same month where sernidiumal oscillations were observed in 1970 During the same time, stalactite 4 showed up clear discontinuity features with a period of about 24 hours Atmospheric and earth tide forcings differ in their frequency spectrum (TOMASCHEK, 1957 ; SIEBERT 1961 ; MELCHIOR 1978). Earth tides occur at the principal lunar and solar frequencies : 0 1 (25 82 h, relative coefficient R = 0 377), K 1 (23.93 h, R = 0 531), M 2 (12.42 h, R = 0 908), S 2 (12 00 h R = 0 423) whereas the atmospheric pressure tides occur at the solar frequencies K 1 and S 2 since they are mainly thermally driven Comparison with our results suggests that the forcing observed in the Coufin cave is primarily of atmospheric origin which is consistent with the fact that the periodic oscillations seem to be observed only during the hot summer months. These observations are to be compared to the periodic pressure fluctuations inducing cave breathing which have been known for a long time They were theoretically interpreted by WIGLEY (1967) who showed that the observation of unexpected large wind velocities induced by periodic pressure variations could be used in conjunction with the theory to gain information on the inaccessible volume of a cave and/or to estimate the combined matrix and fracture permeability of limestones Nevertheless earth tide forcing deserves more investigations since it is clearly displayed in the response of well-aquifer systems through the observation of the water level in wells ( BREDEHOEIT, 1967 ; RITzI e t al., 1991 ; and references therein) Moreover there is at least one observation of a significant semidiumal movement along a bedrock joint in a cave, with an amplitude of about 0.4 m ( DAVIS & MOORE 1965) 5. Conclusions and perspectives The preliminary results presented in this study demonstrate that the flow regimes of stalactites chosen at random in a same cave present some common behaviours like periodic oscillations and fast response to a swelling That makes them powerful tools to study the hydraulic properties of the karstic microfissure network and ultimately to quantify the water flow and the transport of dissolved substances like trace elements humic/fulvic acids and pollutants In particular these result s establish that the response of the fissure network to a sudden and important water excess can be very fast with an impulse response less than 10 hours. Of course this time is not the transfer time of water and more measurements are needed to understand how such short delays could be related to a flu s hing effect similar to what is observed on a macroscopic scale in underground rivers. Although the comparison between the old results of th e Grotte du Sourcier and the Coufin cave ones al o sugge s ts the existence of general behaviours at least in the geographical area studied here it shows important differences in the order of magnitude of the phenomena which have to be understood and correlated with geological and geomorphological parameters There is indeed important differences between these two c a ves : the Coufin cave develops at the basis of a massive limestone whereas the Grotte du Sourcier is a sub-cutaneous cave developing in a consolidated slope scree. Although the high temporal resolution attainable with our data acquisition system has not yet bee n fully exploited, it is clear that it will be useful in the detailed study of swellings as well as periodic forcings It will allow us to correlate dripping rates with environmental parameters such as temperature atmospheric pressure, rains. insolation, river discharge, physico-chemical properties of water ... High temporal resolution is also needed for the study of the flow rate discontinuities which seem to be a characteristic of the percolation regimes of all the stalactites studied in this work. Promising preliminary results are currently obtained in the analysis of these features by the deterministic chaos techniques. If atmospheric forcing is confirmed it should be also of interest to extend the W!GLEY's theory to transient pressure variations in order to try to interpret the strong 1970 perturbations induced by dry thunderstorms Such perturbations have not yet been observed in the Coufin cave recordings but it must be recalled that the periodic o s cillations recorded so far in this cave are much smaller than those observed in the Grotte du Sourcier cave. A c knowledgemenrs : The Centre d Etudes et de Recherches Lasers et Applications ( CERLA) is supported by the Ministere charge de la Recherche, the Region Nord/Pas de Calais and the Fonds Europeens de Developpement Economique des Regions. We wish to gratefully acknowledge Gilbert Mantovani for 72 Proceedings of the 12 th International Congress of Speleology 1997 Switzerland Volume 2

PAGE 84

allowing tis to install our equipment in the Coufin cave, and many thanks are also due to all the guides of this cave for their constant availability. We also thank Michel Garneau who built the electronic devices and Gilles Marcou who carried out, as a summer student, many laboratory experiments on the drop volume. References BAKER A ., SMART, P L. & R L. EDWARDS 1995 Paleoclimate implications of mass spectrometric dating of a british flowstone. Geology 23 : 309-312. BASKARAN, M & R V KRISHNAMURTHY 1993 Speleothems as proxy for the carbon isotope composition of atmospheric C02 Geophys Res L ett. 20 : 2905-2908 BR EDE HOEFT J. D 1967 Response of well-aquifer syste ms to Earth tides J Geoph y s Res 72 : 3075-3087. DAVIS, S N & G W MOORE 1965 Semidiumal movement along a bedrock joint in Wool Hollow cave, California Bull Nat Speleological Soc 27 : I 33-142. DELANNOY J J 1986. Contribution a l'etude des circulations aquiferes dans le geosysteme Coulmes-Choranche. Presentation du site experimental de la Grotte de Coufin. Re v. Geogr. Alpine, T LXXIV 83-92. DELANNOY J J ., GUERDON J L. & Y QUINIF 1988 Les remplissages speleologiques : un apport a la connaissance de la karstogenese du massif des Coulmes (Vercors, Alpes). Ann. Soc Geol. B elg. 111 : 21-38. DELANNOY J J 1991 .: Vercors Histoire du relief. Carte geomorphologique commentee. Editions du Pare nature! regional du Vercors D ELANNOY, J J. 1997 These de Doctorat d'Etat, Grenoble DESTOMBES J L. 1971. Unpublished results. DREYER, K. & F. R HICKEY 1991. The route to chaos in a dripping water faucet. Am. J Phys 59: 619-627. EGGERS, J 1995 Theory of drop formation. Ph ys. Fluids 7 : 941-953 FORDHAM, S I 948. On the calculation of surface tension from measurements of pendant drops Proc R. Soc. London A 190 : 1-16 GARANDET, J P ., VINET, B & P GROS 1994. Considerations on the pendant drop method : A new look at Tate 's law and Harkins' correction factor. J Colloid lnte,fac e. Sci 165 : 351-354. GASCOYNE M. 1992 Paleoclimate determination from cave calcite deposits. Quatern Sci Re v. 11 : 609-632 GENTY D & Y QUINIF 1996 Annually laminated sequences in the internal structure of some belgian stalagmites Importance for paleoclimatology J Sediment. Res 66 : 275288 GENTY, D ., DEFLANDRE G. QUINIF, Y. & S VERHEYDEN 1996 Les !amines de croissance des speleothemes : origine et interet paleoclimatique. Ann. Soc Geol Belg in press GOEDE, A 1994. Continuous early last glacial palaeo environmental record from a tasmanian speleothem based on stable isotope and minor element variations Quatern. Sci. Re v. 13 : 283-291. GORDON, D ., SMART P L., FORD, D C., ANDREWS J N., ATKINSON, T. C., ROWE, P J & N. S. J CHRISTOPHER. 1989. Dating of late pleistocene interglacial and interstadial periods in the United Kingdom from speleothem growth frequency. Quatem. Res 3 I : 14-26. HARMON R S ., THOMPSON P ., SCHWARCZ, H P. & D C. FORD. 1978 Late pleistocene paleoclimates of North America from stable isotope studies of speleothems. Quatern. R es. 9 : 5470 HENNIG G J ., GRON, R. & K BR UNNACKER. 1983 Speleothems, travertines and paleoclimates Quatern R es. 20 : 1-29 HARKINS W. D & F E. BROWN 1919. The determination of surface tension and the weight of falling drops : the surface tension of water and benzene by the capillary height method J. Amer Chem. Soc. 41 : 499-524. l
PAGE 86

Early evolution of karst aquifers in limestone : Models on two-dimensional percolation clusters by Wolfang Dreybrodt & Jorg Siemers Institute of Experimental Physics, University of Bremen, 28334 Bremen Germany Abstract Two-dimensional nets of initial fractures are constructed on a square-lattice by occupying the lines between nearest neighbour sites by a water leading fissure of width a 0 and length l with an occupation probability p. For p > 0 5 percolating nets occur which lead water. To simulate cave genesis we calculate the water flow rates driven by the hydraulic head h through all fissures By employing nonlinear dissolution rates of the type F=kn(l -c /ceq? the widening of the fractures is obtained. At the onset of karstification flow is evenly distributed on all fractures. As the system develops solutional widing creates preferred pathways, which attract more and more flow, until at breakthrough both widening and flow increase dramatically. We discuss the evolution of karst aquifers for natural conditions and also upon human impact at dam sites where steep hydraulic gradients may generate water leading conduits below the dam in times comparable to the lifetime of the structure. 1. Introduction To understand processes of the early state of karst genesis, one-dimensional models have been developed which couple widening of primary fractures in limestone by dissolution, and flow rates of the water driven from the input to the output (DREYBRODT, 1990, 1996; PAL MER, 1991; GROVES & HOWARD, 1994). These models show that the width of the fracture at its exit increases slowly by dissolution until it has been doubled From then on a positive feedback loop is enhanced, and within an extremely short time, width and flow rates increase dramatically The time, when this happens is termed as break through time and terminates early evolution. This time can be estimated to be TB a 0 /F(l,0) [years], where ao [cm) is the initial width of the fracture and F(l,0) is the limestone dissolution rate in cm year1 at the end of the developing channel at the onset of its evolu tion. From this an expression has been derived which quantifies breakthrough time as a function of the pa rameters, which determine karstification (DREYBRODT, 1996) These are the initial width a 0 the breadth b 0 the length l of the fracture, and the hydraulic head h between the input and the output, which is assumed as constant until breakthrough occurs. Furthermore the kinetics of dissolution plays a role of utmost import ance It is quantified by the equilibrium concentration Ceq with respect to calcite, which can be attained by the aggressive solution, and the constants determining the dissolution rates as a function of the calcium concentra tion c in the solution For narrow initial fractures with widths between 5 103 cm up to 1 101 cm these rates are given by a linear relation F 1 (c) = k 1 (1-c/c. 9 ) (DREYBRODT et al., 1996). Close to equilibrium at c 0.8 Ceq the rates, however, switch to a higher order law Fn=kn(l-c/c. 9 t, where n=ll for karstifying lime stone (SVENSSON & DREYBRODT, 1992; EISENLOHR et al ., 1997). Without this switch to a higher order, kar stification for natural systems, i e evolution of conduit aquifers, would not be possible at all. The work described so far enables one to a first un derstanding on the early evolution of karst in its dimen sions on space and time. More realistic models, however, are required. In this work we report on two-dimensional models, which incorporate the statistical distribution of primary fractures We will give some examples for na tural karst systems. We will also model karst evolution for man-made conditions, such as dam sites, where ex tremely steep hydaulic gradients favor the development of karst channels in short times. 2. Model structure To model the random distribution of initial fractures in limestone penetrable by water, a percolation network is constructed by the following rules: Fig. 1 depicts an array of squares, shown by their corner points Each side Figure 1 : Fracture network with an occupation probability p=0.5. The fat lines represent frac tures, which transport water flow. of a square represents a location, which could possibly 5 h Conference on Limestone Hydro logy and Fissured Med ia 75

PAGE 87

host a fracture. These locations are randomly occupied with an occupation probability p > 0.5. The full lines in Fig. 1 represent such fractures and form a percola ting network. Depending on the boundary conditions, some of those fractures transmit water from the input to the output (fat lines), others do not (thin lines) In Fig 1 the lower and upper side of the network are as sumed as impervious. The left hand side is subject to a hydraulic head h, whereas the right hand side is at low hydraulic head h=0. Only the percolating pathways (fat lines) carry water. Such percolating pathways can be constructed with various occupation probabilities p, to model the degree of fracturing in the rock. For va lues p < 0.5 no percolating pathways will occur. With increasing p the density of the network increases. Fig 1 represents a net with p=0.5 Later on in this work we will use networks with p=0.7. Compare Fig 2. To calculate the flow rates in each fracture a yet un known hydraulic head h; is assigned to each node, where two or more fractures are connected. Assuming lami nar flow, which is verified by calculating the REYNOLD 's number, a set of linear flow equations is obtained, which depend on the geometric profile of each fracture. From these the heads h; are found. Then we calculate the dis solutional widening for all tubes in the following way: A. First we select an input at the boundary of the net work. We specify the concentration Cin of the solu tion, which flows in to the input fracture. Then we apply the one-dimensional transport-dissolution model (DREYHRODT, 1996) to calculate the concentration pro file along this channel, including the concentration of the solution leaving it. Furthermore we calculate the new profile of the channel after a time step li.t. We re peat this procedure for all input points at the boundary. B In the next step all those nodes are selected, where the concentrations of all the inflowing solutions are known. We assume complete mixing of these solutions before they are transfered into those conduits which transport the flow away from the corresponding node For these, as described above, the one-dimensional transport-dissolu tion model is used to obtain the concentration of the solution at their exits, and also their new profiles after the time step li.t. Then we repeat procedure B until the new profiles of all conduits have been obtained We thus have obtained a new network, to calculate the new flow rates at time li.t and the new profiles at time 2.6.t. By repeating procedures A and B the pro files of all conduits and their flow rates can be obtai ned as a function of time. It should be noted that the boundary conditions and the input concentrations can be changed after each time step, thus considering exter nal changes by climate or geological changes (up-lifting, down-cutting of base level) 3. Results Natural conditions Natural karst systems are characterized by low hy draulic gradients h/1, and large values ofl. Fig 2 depicts such a system The length of the square is 3 km. We assume all sides of the square to be impervious, with the following exceptions. At the left hand side we have defined three input points marked by arrows at a hy draulic head of h=150 m. There is one output point at the upper corner of the right hand side, located at base level, h=0. The following parameters have been used in this model, p=0.7, ao=l02 cm, bo=lO0 cm, Cin=0, c. 9 =2 mmole 11 c,=1.5 mmole 11 k1=4 1011 mole cm2 s1 k 11 =4.2 105 mole cm2 s1 and n=ll. Fig. 2a shows the fracture widths which have evolved after 280285 years. The thin lines (-) depict fractures with widths < 0.1 cm. Correspondingly (-) depicts fractures between 0.1 and 1 cm, and (-) fractures with widths between 1 cm up to 10 cm. Fi nally the fattest lines (-) denote fractures larger than 10 cm. From Fig. 2a one visualises that a karst channel has developed from the upper input towards the output, pe netrating half their distance. The other two inputs lo cated a higher effective distance to the output, less suc cessfully, have created only short channels. Only 1300 years later, however, the upper channel has reached the output (Fig. 2b), causing a dramatic increase in the flow rate, as is depicted by Fig. 3. Since these high flow ra tes cannot be delivered from the storage of the system the boundary conditions at the upper input are chan ged to a constant inflow of 500 cm 3 s1 At such high flow the concentration at the output, close to saturation prior to breakthrough, drops to low values (see Fig 3) and the upper channel widens with a rate of several cm in hundred years. Fig. 2c shows the system after further 10500 years. Due to the low hydraulic head established in the upper channel by restricted input, the second in put point has directed its channel evolution upward and thus created a second breakthrough event as shown by Fig. 3. Therefore also at this input point the inflow be comes limited (500 cm 3 s1 ), whereas the lowest input remains at constant head. Fig. 2d shows the situation only 800 years later at 292023 years. The l0west chan nel has succeeded to reach the upper network. From this moment on there is almost equal widening everywhere in the conduits, and the structure of the conduit system will remain unchanged if the boundary conditions re main unchanged It should be noted that the evolution of the first channel in Fig. 2a takes 280000 years, whereas the further development needs only 12500 years. Fig. 3 shows the evolution of flow rates Q leaving the exit Fur thermore the width a of the exit conduit and also the concentration of the water at the exit are illustrated From this one sees that the first breakthrough event ta kes the system into the state of adolescence going to ma turity a short time later Prior to the first breakthrough the width and the flow rates increase slowly. The con centration at the exit is high and therefore dissolution rates are low. After breakthrough the concentration c drops and dissolution rates become high 76 Proceed ings of the 12 '" International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 88

a) 280285y b) 281585y c) 292023y d) 292823y Figure 2: Evolution of a karst aquifer of 3x3 km 2 with 3 inputs (left hand side) at h=150 m and one output at base level (h=O). (high clip mo del). (a) First karst channels develop until (b) at a first breakthrough event the upper channel has reached base level. (c) The next channel is directed towards the upper channel and a next breakthrough event occurs. (d) Finally the lo west channel is inte~rated into the network. 1 0 0 8 0 6 0.4 0 2 amax = 57 cm ceq = 2 mmole/1 Qmax = 1500 emfs 0 0-1----------------; a/amax _.... ....... ..... ... ... ... // I I / Q/Qmax 2 5x10 5 2 6x10 5 2 7x10 5 2.8x10 5 2 9x10 5 T [a] Figure 3: Time dependence of flow rate Q(t), frac ture width a(t) at the exit, and concentration c( t) of the water at the exit. All values are normalized to the maximal values at the end of the ruu. The arrows designate the breakthrough events in Figs. 2b,c, and d. Therefore a(t) increases quickly The nonlinear be haviour of a(t) results from the fact, that at larges a(t) the dissolution rates are controlled by diffusional trans port. One should note at this point that the breakthrough time for each event depends critically on the initial frac ture widths by Tn ex ao 2 3 (DREYBRODT, 1996, SIE MERS and DREYBRODT, 1997). Thus upon a change of ao from 0.01 cm to 0.02 cm breakthrough times would be reduced from about 290000 years to about 60000 years 6 "' Conference on Limestone Hydrology and Rssured Med ia 77

PAGE 89

a) 25463y b) 26364y c) 44896y d) 45196y Figure 4: Evolution of a karst aquifer (low dip mo del). The input points at the left hand side and inside the network ( crossed circles) are at equal head. Therefore in a first step karst channels are created only from the inner input points. (a) Af ter breakthrough of the upper channel the lowest (b) reaches base level and the channels evolving from the two inner inputs are quickly integrated. From then on ( c) the channels grow from the left hand side and are integrated into the fist conduit system. Our model resembles the high dip model of EWERS (1982) and confirms EWERS' prediction of cave evolution as derived from models of Paris plaster (see also the textbook of FORD and WILLIAMS, 1989). As a second example Fig. 4 illustrates the evolution of EWERS' low dip model. In this case several ranks of multiple inputs exist with equal heads To obtain such a case we have changed the boundary conditions of the network depicted in Fig. 2. The upper and lower sides of the square are still impervious. The left hand side is at a hydraulic head of h=l50 m and has various inputs. The right hand side at base level (h=0) exhibits many output points. Finally, we have introduced inside the net a row of 4 input points with fixed head h=l50 m, depicted by circles. This is a situation often encounte red in karst. The plane square represents an extented horizontal bedding plane, where water descends to the input points from an upper plateau through open joints The Mammoth-Flint Ridge System, Kentucky, USA is a representative example for this case (FORD and WIL LIAMS, 1989). Fig. 4a shows the system after the first breakthrough from one of the inner inputs to base level. Channels have also developed from the other inner inputs but have not yet succeeded to reach base level. This is achieved 500 years later at 26363 years as illustrated by Fig. 4b. Due to the restricted input the hydraulic head is lowered at all inner input points Therefore karst channels grow 78 Proceedings of the 12 '" International Congress of Speleology, 1997 Sw i tzerland Volume 2

PAGE 90

downhead from the inputs at the left hand side. They reach, as seen in Fig. 4c those conduits already created in a further 19000 years. If there are further arrays of input points, this scenario is repeated. Thus dissolution creates channels growing downhead. Integration of the cave system, however, proceeds uphead. The reduction of breakthrough times in comparison to the case presen ted by Fig. 2 results from the fact that breakthrough times are related to the hydraulic head and the effec tive length between the input and the output by Tn ex (l 2 /h) 1 1 (DREYBRODT, 1996, SIEMERS & DREY BRODT, 1997). Fig. 5 as a further example, shows a vertical cave in its final state after 5700 years. The boundary conditi ons are: The left hand side and the low er side of the aquifer are impermeable. The upper side represents a plateau 3 km wide at an elevation of 300 m with various input points. The right hand side illustrates a vertical cliff with various output points at different heights, the lowest at base level, e.g. a valley We have used a 0 =0.01 cm, everything else is unchanged Note that the verti cal dimension now is only 0.3 km. The system evolves as follows: First the input nearest to the edge creates breakthrough at the upper most outlet after 3'196 years. Its input is then restricted. Other inputs, still at high hydraulic head, develop channels integrating to the first channel system. Furthermore vertical shafts are created simultaneously. Figure 5: Final state of a vertical cave system. In conclusion the examples presented, show that the structure of the conduits in a karst aquifer are deter mined by the initial setting of the fracture system and by the boundary conditions. After a long time of initia tion until the first breakthrough event, complex patterns arise in comparably short times Furthermore realistic evolution times of karst systems are obtained by use of realistic geometrical parameters a 0 b 0 I, and hand che mical dissolution kinetics, as obtained from the labora tory. Man-made conditions Close to hydraulic structures, such as dam sites, un naturally steep hydraulic gradients are created, and a sufficiently large supply of water driven by these high gradients through fractures beneath the foundation and across the abutment rock is always provided. There fore the question arises, whether within the lifetime of the dam, dissolutional widening might create channels sufficiently wide to cause serious loss of water. We therefore have applied our model to a situation illustrated by Fig. 6. a) 343y b) 99 y Figure 6: Evolution of conduits below a dam. {a) With a grouting curtain 200m deep, breakthrough occurs after 343 years. (b) Redu cing the depth of grouting to 100m results in breakthrough after 99 years. In both cases the channel widths at the exit have increased by a factor of ten to 0.16 cm. 6 1 Conference on Limestone Hydrolog y and Fissured Med ia 79

PAGE 91

We have constructed a percolation net with p=0 7 and imposed the following boundary conditions A grou ting curtain below the dam creates an impervious layer 10 m wide and 200 m deep, and thus blocks flow This is depicted by the shaded area in Fig. 6a The basal width of the dam is 50 m and impervious (shaded area). To the left of the dam water is impounded at a height of 200 m and sevel input points are available To the right hand side h=O. The remaining sides of the 300 m wid e square are assumed as impervious for simplicity Th e width of the initial fractures is a 0 =0 015 cm Fig 6a represents the channels which have evolved at breakthrough after 343 years. Fig 6b illustrates the situation if grouting is reduced to 100m. From th e n on turbulent flow sets in and dissolutional widening is in the order of 0 2 cm year1 From this by use of the DARCYWEISBACH equation water losses on the order of several m 3 s1 are estimated at chann e l widths of 5 cm Thus after breakthrough only 25 y e ar s a rc necessary to create serious water loss If the d e ptl1 of the grouting curtain is reduced to 150, 100, or 50 m, the corresponding breakthrough times are 255, 99 and 27 years, respectively Thus sufficiently deep grouting curtains are not only necessary to prevent flow through wide fractures already existing, but also to avoid the formation of new ones. It should be noted at that point, that similar cal culations on 1-dimensional models (DREYBRODT, 1992 1996; PALMER, 1988) yield breakthrough times s e v e ral times larger than those obtained on two dimensional nets. This is caused by two reasons W e have u s ed hig her order kinetics with n=ll, which seems to be mor e appropriate from recent experimental data (EISENLOIIR et al 1997) than kinetics with n=4, used in the one dimensional model. Furthermore, many parall e l con duits arise on two-dimensional networks (Fig. 6), which reduce resistance to flow in contrast to a one-dimensional channel. 4. Conclusion We have presented a two-dimensional model of the early evolution of karst aquifers for natural and man made conditions, based on a high order non lin ea r di s solution kinetics close to equilibrium. In all cas e s aft e r a long initiation time a sudden and dramatic incr e as e of widening of the fractures and flow through the syst e m ends the initial state. After this breakthrough event all fractures employed widen at an even pace by several 102 cm year1 under natural conditions, when flow is laminar If flow becomes turbulent the annual wid e ning increases by enhancement of diffusional mass transport (DREYBRODT and BUHMANN, 1991) up to about 101 cm. The latter occurs when after breakthrough a suffi cient amount of water is provided, as is the case at dam sites For natural conditions, depending on the boun dary conditions, complex conduit patterns arise, which give a realistic simulation of cave geomorphology and its dimensions on space and time. Application of the model to man-made conditions, such as dam sites in karst, provides strong suspicion that new karst channels may develop below the dam, which cause serious water losses within its lifetime. References DREYBRODT, W 1990 The role of dissolution kine tics in the development of karstification in limestone: A mod e l simulation of karst evolution J. Geol 98: 639655 DREYBRODT, W. 1992 Dynamics of karstification: A model applied to hydraulic structures in karst terra nes. J\ppl. Hydrogeo., 1 : 20-32. DREYBRODT W 1996 Principles of early develop ment of karst conduits under natural and man-made conditions revealed by mathematical analysis of nume rical models Water Resour. Res., 32(9) : 2923-2935 DREYBRODT w & D BUIIMANN. 1991. A mass tran s fer model for dissolution and precipitation of cal cite from solutions in turbulent motion Chem Geol ., 90: 107-122 DREYBRODT, W.; LAUCKNER J ; LIU Z ; SVENS SON U. & B BUHMANN 1996 The kinetics of reac tion C02+Il20-+ n++HC0 3 as one of the rate limi ting steps for the dissolution of calcite in the system II20 C02-CaC03. Geochim Cosmochim. Acta, 60(18) : 3375-3381. EISENLOIIR, L. ; MADRY B & W DREYBRODT 1997 Changes in the dissolution kinetics of limestone by int rinsic inhibitors adsorbing to the surface This volume EWERS, R 0 1982 Cavern development in the di m e nsions on length and breadth Pit. D Thesis Mac Ma s t e r University, Hamilton, Ontario, Canada fORD, 0 C. & P W. WILLIAMS. 1989: Karst g e morphology and hydrology. Unwin Hyman, Boston, 60lp GROVES, C G. & A D HOWARD 1994 Early de velopment of karst systems, 1, Preferential flow path enlargement under laminar flow Water Resour. Res ., 30(10): 2837-2846. PALMER, A N. 1988. Solutional enlargement of ope ning in the vicinity of hydraulic structures in karst regi ons Jn 2nd Conference on Environmental Problems in I{ arst Terranes and their solutions, Assoc of Ground water Sci and Eng Nashville Tenn., 16-18. PALMER, A N 1991. The origin and morphology of lim e stone caves Geol Soc. Am Bull., 103 : 1-21. SIE:MERS, J & W DREYBRODT 1996 Early dev lopment of karst aquifers on percolation network of frac tures in limestone Submitted to Water Resour R e s SVENSSON U. & W DREYBRODT. 1992 Dissolu tion kinetics of natural calcite minerals in C0 2 -water syst e ms approaching calcite equilibrium. Chem. G e ol 100: 129-145 Acknowledgements: We thank the Deutsche Forschungsgemeinschaft for financial support. 80 Proceed in g s o f the 12 '" Interna tio na l Con g res s of Speleo l og y, 1 997, Switz erland V ol ume 2

PAGE 92

Changes in the dissolution kinetics of limestone by intrinsic inhibitors adsorbing to the surface b y Laurent Eisenlohr, Britta Madry & Wolfgang Dreybrodt Institute of Experimental Physics, University of Bremen, 28334 Bremen Germany Abstract We have measured the dissolution rat es of natural carbonate-CaCO 3 minerals by employing a free-drift experiment, where sized particles were suspended by turbulent stirring in a H 2 O-CO 2 solution The experimental system was closed with respect to CO2 as is the case in the evolution of karst aquifers. At low calcium concentrations c we find a rate law F 1 =k1 (1-c/ c 09 ) n,, which switches to a higher order law F2=k2 (1-c/ c 09 t 2 above a concentration c,. The values of the constants k1, k2, of n 1 n2 and of c,, depend systematically on the ratio V / A, where V is the volume of the solution and A the surface area of the dissolving mineral. This result can be exp lain ed by assuming that inhibitors are present in the bulk of the mineral, which upon dissolution accumu lat e on its surfa ce. Thus inhibition increas es with the thickness of the layer removed from the limestone which is proportional to th e V / A-ratio. These results are discussed with respect to previous experiments in the lit e ratur e and with respect to the genesis of limestone aquifers 1. Introduction In order to model the time-space relationship of karst environments, exact knowledge on the dissolution kine tics of limestone is rigorously required (DREYBRODT, 1988, 1990, 1996; PALMER, 1991; GROVES & HOWARD, 1994). Extensive experimental work has been publis hed on the dissolution kinetics of calcite in the system CaCO 3 -H 2 O-CO 2 under open system conditions with re spect to CO2. PLUMMER & WIGLEY (1976) showed that dissolution rates observed on Iceland spar cou ld be des cribed by an empirical relation : F1 = k1(l c/ceqr I for C c,, and (1) F2 = k2(l c/c 09 t 2 for c > c, (2) k 1 k 2 are rate constants, c is the concentration of Ca 2 + in the solution, and c 09 the equilibrium concentration with respect to calcite. c, is the concentration, where the rate law switches from order n 1 to order 112, n2 > n1. They observed n 1 ::::: 2, c,/c 09 ::::: 0.9 and 112::::: 4, but in some runs also values up to 8. Using the experimenta l data of PLUMMER et al. (1978) from dissolution rates on Iceland spar at various Pco, and temperatures, PAL MER (1991) found a similar rate equation. SVENSSON & DREYBRODT (1992) have carried out measurements of dissolution on natural limestones, marbles, and cal careous deep-sea sediments which have confirmed the rate laws above on these materials. They found values of n 1 in the range from 1.2 up to 2.5. Beyond a value of c,/c 09 in the range from 0 .6 up to 0.9 the dissolu tion rates switched to higher order with 2.5~n2~4 .9. In contrast, dissolution rates obtained experimentally on synthetic calcite (NBS) exhibited a linear dissolution be haviour with n 1 :::::1 whi c h is closely relat ed to th e theo retical dissolution rates predicted by the PWP-model (PLUMMER e t al 1978) This fundam e ntal differ e nc e between dissolution of natural material and synth e tic calcite has been explained by th e presenc e of inhibitors which are origina lly in the bulk of the mineral. During dissolution those trace elements as e.g. Cu 2 +, Pb2+ or PO!are absorbed to the calcite surface and block dis solution processes. So far all experimental rates r eported have been obtai ned under conditions of a system open with repect to CO2 The aim of this study is to investigate dissolu tion kinetics of limestone under closed system conditi ons with r epect to CO 2 since these conditions determine initial processes in karst evolution. 2. Material and method Limestone from the Swiss Jura Mountains was used in all the experiments. This material was ground, size fractioned, treated with diluted HCl (0.02 molar) clea ned in bi distilled water, in acetone and finally dried at 30C. The size fractions selected for the dissolution ex periments were 100-125 m, 125-180 180-250 250-355 m, 355-500 and 500-710 m. The surface area was estimated from the weight of th e sized material and the mean dimension of the particles assuming each particle as a rhomboh e dron. Dissolution experiments hav e also been perform e d on lim esto ne disks from the same material. These disks were 7 mm thick slabs with a diameter of 5.8 cm. The lim esto ne face exposed to dissolution was smooth ed and polished using progres sively 1200 2400 and 4000 silicon ca rbid e paper th e n treated in HCl (0.02 molar) cleaned in bidistilled water and dried The disks wer e mounted into the bottom of the reaction vessel. Fig. 1 shows the experimental set up. The experiments w e r e carried out as batch runs, using the free-drift techniqu e. Th e solution was contai ned in a Teflon vessel (1) with a volume V=262 cm 3 thermostated at 10 .1 C. No furth e r space for air was available. Particles were kept in supension by turbu l e ntly stirring the solution with a propeller at 350 rpm (3) and (5). The experiment was started by filling the Teflon vessel with bidistill ed water in eq uilibrium with a CO 2 atmosphere of 5.102 atm partial pressure. Then a given amount of sized lim esto n e was added and th e Teflon vessel was seal e d imm ed iat e ly to b e airtight (2). The calcium concentration whi c h developed by dis so6 111 Conference on Limestone Hydrology and Fissured Media 81

PAGE 93

3 4 1 5 Figure 1 : Sch em a tic vi e w o f t h e e xp e rim e n t al a paratus (see t e xt ). lutio n was followed by measuring the conductivity ( 4) as a function of time for periods in the range of 3 up to 48 days, depending on the surface area of the added mineral. At the end of the run, the final calcium concen tration was ana l ysed by E D TA titration and I CP. Using the linear relation between the conductivity O' [S] and the calcium concentration c [mmo l e 11 ], c is obtai ned as a function of time by the relation c=5.35 103 0' 0.053 ( D REY B R O DT et a l 1996) Calcite dissolution rates have been calculated using the relation: de R=V/A dt (3) where V is the volume of water, A the surface area of limestone in contact with water. U n its are mole cm s. Correspondly concentrations are in mole cm3 3. Results and discussion We have measured dissolution rates for different V / ratios in the range from 0 15 up to 9 cm. The maximum V / A ratio was obtained by employing limestone disks A typica l experimental resu l t of the observed l imestone dissolution rates is shown in Fig 2. By plotting logarith mica ll y, disso l ution rate versus (1-c/ Ceq) a kinetic switch from a lower to a higher reaction order is well seen at c,. This higher reaction order is independent of the stir ring rate. Log dissolution rate was plotted against log (1-c/ceq) for more than 40 experiments. From these th e empirical relations (1) and (2), n 1 n2, k 1 k2and c, w e re found by regression ana l ysis as shown in Fig. 2 Th e steep decrease of the rates at concentrations abO\e c, ind i cates that an inhibition mechanism is activ e. As a working hypothesis we assume that inhibitors such as PO~are incorporated in t h e bulk of the limestone and are absorbed to the actual surface, when they becom e mobi l e during des in teg r at i on of t h e l att i ce by d isso l tion. In that case the concentration of these inhibitors at the surface should be dependent on the depth d of the layer removed by disso l ution. This is related to the V / A-ratio by: d = V/A Cend 0.037 (4) Cend is the concentration in mole 1 1 when the run was completed. From the considerations above one ex pects va r iations in the disso lu tion r ate with i nc r easing d. We have therefore mess u red the rates w it h varying V / A cove r ing the r egion fr om d '.:::'. 0.2 10 4 t o 5 104 cm fo r the particles and d '.:::'. 6 10 4 cm for t h e disks. Fig. 3a,b, Fig. 4a,b and F ig. 5 depict t h e r es ul ts o f 42 dif-5,--------------------, T : 1 0"C ; pC0 2 : 5 1 o2 atm V/A:0 36cm 'j -6 ., t E -7 0 0 8 E E ..__, -9 e "' .2 -10 run : JCS10 2 {\\ C/C-i -11 -h-r.-'T"'T.,...,.....,...,....,...,...,..........,....rr,rr,..,....,..,.....-r.-..,...,....,...,...,.....,...,....,.,,.......,-t -2.0 1 7 -1 + -1.1 -0 8 -0 5 -0.2 0 99 l og( 1 -C/Cr,q) 97 95 .90 c/c9Cl 75 50 0 0 Figure 2: l og o f t h e di sso lu t i on r a t e v e r s u s l og ( 1c/ ceq) fo r Jur a l i m esto n e ferent runs, each point representing an average of about 2-8 identical experiments. As can be visualized from Fig 3a,b, Fig. 4a,b and Fig. 5 all parameters show a similar variation with the thickness d removed fr om the mine r al. After a steep i ncrease within a few 10 4 cm (n1, n 2 and k 2 ) or a steep decrease (k 1 and c,) saturation is reached. This indicates that the surface concentration of the inhibitors also attains a saturation concentration, when a ll possib l e sites for absorption are occupied. It should be noted that the depth of 3 104 cm corre sponds to about 3000 atomic layers With a concen tration of 104 moles of inhibiting species per mole of limestone, and all of those absorbed at the surface, one expects a coverage of about 30 %. We have also per formed experiments on marb l e and flowstone deposited in a cave First results indicate that the trends obser ved in the Jura limestone apply genera ll y for all natural calcium carbonate minerals There is a further conse quence of the variation of the kinetic constants, espe cially c,, n 2 and k 2 : Due to the effective inhibition of dissolution the approach to thermodynamic equilibrium with respect to calcite becomes infinite l y slow and the solution attains an apparent quasi-equilibrium Fig 6 illustrates the apparent equilibrium concentration Capp as a function of d After a steep decrease beyond a va lue of about 1.5 104 cm this concentration reaches a constant va l ue Capp '.:::'. 1.78 mmole J1 The values of the apparent equi l ibrium concentration shown in Fig. 6 82 Proceedings of the 12 th International Congress of Speleology, 1997 Switzerland Volume 2

PAGE 94

A)2.o c 1 8 ... G) 'C 0 1.6 1.2 1.0 s{5 12 11 N c10 ... "E 9 0 8 C o 7 g 6 5 4 3 I I) I I I I~! I I t CLO~ED SYSTEM 2 T : 10 C; pC0 2 :5 atm i 2 3 4 5 6 e I I e e !m CLO~ED SYSTEM 1 T:10 C ; pC0 2 : 5 otm i 2 3 4 5 6 thickness removed (10-4cm) Figur e 3 : A) Reaction ord e r n 1 as a function o f removed limestone by dissolution B} R e action order n 2 as a function of r e mov e d lim es ton e by dissolution A) 4 5 f'4 0 CLOSED SYSTEM 2 T : 10C; pC0 2 : 5 o otm (I) b 7* 3 5 E 'a 3.0 Q) ] 2.5 ,..E 2 0 I 0 m ,::.1.5 -:Z1.o I e e I 0.5 4-~-~-~~--.--.----.--~~--.--,---l 0 2 3 4 5 10~ ~------------------~ 8) e e e 1 c orr es p o nd to di ss olut io n r ates b e l ow 1014 mo l e cm 2 s1 N ot e t h e r e s e mbl ance b etween the b ehavio u rs of t h e Capp ( Fi g. 6) a nd the k i n et i c parameters (Fig. 3a,b, F i g 4a, b a nd Fig. 5). R ece n t ly we have pe rform ed sub0 90 ~-----------------~ 0.85 uo.ao 0 75 0 e I CLOSED SYSTEM 2 T:10C ; pC0 2 :510otm 2 3 4 5 thickness removed ( 10-4cm) I 6 F i g ur e 5: c as a function of th e thickn es s of r e m o v e d lim e ston e by di ss olution seq u ent d isso l ut i on experiments increas in g the va lu e of d w i th each r epea t e d ru n on the same l imestone u de r ope n system co n ditions wi th r espect to C 0 2 and under tu r b ul ent motio n The resu l ts, not shown here, are in good agreement with our inte r pretation. Pre viously, in h ibition was observed a l so i n open systems with Pco 2 = 1 atm by PLUMMER & WI GLEY (1976) and by H ERMAN (1982). In a batch exper i me n t PLUMMER & WIGLEY ( 1 976) observed nonlinea r k inetics with or der n2=8 Using a rotating d i s k tec h nique (V / A = 60 cm) HERMAN ( 1 982) observed n 2 = 10. S u mma r iz i ng we have found that the d issolut i on kinetics o f natura l cal cium carbonate mine r als depends on t h e depth of the l ayer a lr eady remove d by disso lu tion T h erefore experi menta l data obtained from freshly broken mate r ia l are not adequate to be used in d i scussing a r ea l geo l ogical situation. Due to the accum ul ation of inhibito r s, ra tes can be l ower by orders of magnitude. Further work must be done to investigate the nature of the inh i bitors and of the inhibition mechanism. To corroborate the existence o f this i nh ibition effect one s h ou l d note the fact, that it is miss in g comp l ete l y in u l t r apure synthetic calcite (N B S), which sho u ld be free of i nhibitors. This h as been observed by SVENSSON & DREY B RODT (1992), who showed that fo r this mate r ia l the d i ssolution r ates are we ll described by the r ate equation of PLUMMER et 'I' f :'. :: ;:icalions to karst evolution -:2 f In the evo l ution of karst aquifers by calcite aggres10 i::,, pri~~102 atm siv e aqueous C0 2 -so l utions percolating through pr i mary ~ 0 '---~~~....,...-...2 -~-. 3 -....,...4 ..-----.--, 5 -~--1 5 fractures of li mestone rocks, one has to consider that thickness removed (10--4 cm) l ayers of about severa l 104 cm will be removed from the Fig ur e 4: A) Kin e ti c c on s t a nt k 1 as a functi o n of r e mov e d lim e ston e by di ss olution. B) Kin e ti c con s tant k 2 a s a function of r e mov e d lim es ton e by di ss olution. freshly bro k en joint surfaces within a short time compa red to the evo l ution of the karst conduit (DREYBRODT 1996). T h erefore the disso l ution kinet i cs in the inhibi ted reg i on w i th 11 2 c'::'. 11 seems to be more appropriate i n mode ll ing the ea rl y evo l ution of karst conduits. Up 6 th Conference on Limestone Hydrology and F issured Media 83

PAGE 95

2.2 C CLOSED SYSTEM eq:2.14 T:10C; _EC~ : 52 otm -----------------------------------------------I_ 2.1 '1II I) 0 E 2.0 .., I 0 I ....... I f 1.9 II r N I 8 1.8 e ...... 0 I 1.7 0 2 3 4 5 6 thickness removed (10--4cm) Figure 6: Ca 2 + concentration as a function of the thickness of removed limestone during dissolu tion experiments; open circles: Calcium concen tration at the end of the runs; stars: calcium concentration at dissolution rat e :10 11 mmol e cm2 s1 to now dissolution rates with n 2 = 4 have been used in all these modelling efforts. We therefore have used our mode l (DREYBRODT, 1996) to calculate breakthrough times for a karst channel consisting of a p l ane parallel fracture with initial aperture widhts a 0 from 5 10 3 cm up to 3 102 cm subject to an hydraulic ratio (12 /h). I is the l ength of the fracture in [cm) and h the hy draulic h (cm) acting, to drive wate r through it. Thus 1=1 km and h=50 m correspond to an hydrau li c ratio of 2 10 6 cm. Br eakthrough times mark the end of early conduit evolution by a dramatic increase in the flow rates and are a measure for the intensity of kar stification (DREYBRODT, 1988 1990 1996) In Fig 7 the dashed curves give breakthrough times for the inhi bited kin etics with n2=ll taken from our experiments, for various values of ao and (12 /h). The full lines illu strate the breakthrough times for the kin etics with n2=4 taken from experiments obtained on freshly broken ma terial (SVENSSON & DREYBRODT 1992) From this figure one can read that in cases, especially when kar stification times are extremely larg e for kinetics with n 2 =4 (e.g. ao= 5 103 cm, 1=10 km and h=50 m), these are reduced by one order of magnitud e upon the action of the inhibition effect (n 2 =11). Thus the evolution of karst aquifers over large distances of se\eral km in joints with small aperture widths of about 50 m becomes feasibl e in time scales of l ess than one million years. This shows that karstification is much l ess restric ted than believed up to now, where aperture widths of at least 0.01 cm were supposed as a nec essary condition for the evolution of larg e karst aquifers in geo l ogica lly reasonable times. References DREYBRODT W. 1988: Proc esses in karst systems Physics, chemistry and geo l ogy. Spr in ger Series in Phy sical Environment, Springer Berlin New Y ork, 288 p. --=-10 1 n2:4 0 n2:11 m10 8 I-: 10 5 2 7 .c -c,, 10 4 ... 2 5 .c 10 3 ... .a 10 2 5 s 7 II 10 -l 2 3 4 initial aperture aa (cm) Figure 7 : Breakthrough tim e as a function of in itial aperture a 0 widths for kin e tics with n 2 =4 (full line s) and n 2 =11 (dashed lines) at variou s hydraulic ratios. DREY BRODT, W. 1990. The r ole of dissolution kine tics in the development of karstification in limestone : A model simulation of karst evolution. J. Geol., 98: 639655. DREYBRODT, W .; LAUCKNER J .; LIU Z .; SVENS SON U. & B BUHMANN 1996 The kinetics o f reac tion C02+H20-+ H++HC0 3 as one of the rate l imi ting steps for the dissolution of calcite in the system H20-C02-CaC03 Geochim. Cosmochim. Acta, 60(18): 3375-3381 DREYBRODT, W. 1996 Principles of early develop ment of karst conduits under natural and man-made conditions revealed by mathematical analysis of nume rical models. Water Resour Res., 32(9): 2923-2935. GROVES, C. G. & A. D HOWARD 1994 Early de velopment of karst systems, 1, Pr efere ntial flow path enlargement under l aminar flow Water Resour. Res ., 30(10): 2837-2846. HERMAN J S. 1982. The dissolution kineti cs of cal cite, dolomite and dolomite rocks in carbodioxide water system. Ph D Thesis, Pennsylvania State University. PALMER A. N. 1991. Th e origin and morphology of lim estone caves. Geol. Soc. Am. Bull., 103 : 1-21. PLUMMER, L N. & T. M. L WIGLEY. 1976. The dissolution of calcite in C0 2 -satured solutions at 25 C and 1 atmosphere total pressure. Geochim. Cosmochim Acta, 40: 191-202 PLUMMER, L. .; WIGLEY, T. M. L. & D. L. PARK HURST 1978 The kin et i cs of calcite dissolution in C02water systems at 5C to 60C and 0.0 to 1.0 atm C02. Am. J. Sci., 278: 537-573. SvENSSON, U. & W DREYBRODT 1992 Dissolu tion kinetics of natural calcite minerals in C02-water systems approaching ca l cite equilibrium. Chem. Ceo/., 100 : 129-145. A c knowl edgeme nts W e thank the D e uts c h e Forsch ungsgemeinschaft (W.D. and B.M .) and the Fonds na tional de la recherche scientifique suisse (L.E.) for finan cial support 84 Proceed ings of the 12'h International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 96

Abondance et origine des composes carbones inorganiques des sources karstiques du Languedoc-Roussillon par Jean Pierre Faillat Laboratoire d Hydrogeologie. ISAMOR. Technopole Brest-Iroise. 29280 Plouzane. France Abstract The compilation and interpretation of hydrochemical data (major constituents) from 650 karstic springs of Languedoc-Roussillon, recently completed with physico-chemical chemical and isotopic analysis of & 13 C-TDIC vs PDB of about thirty springs aJlow to settle levels of anomalies for HCO3 pCOz and pH above which deep CO2 is necessary to explain the HCO3concentrations. Diversity of the sources of this deep CO2 is placed in a prominent position. the factors of control are indexed and their relative importance briefly evaluated Resume La compilation et !'interpretation de donnees hydrochimiques (ions majeurs) anciennes concernant 650 sources karstiques du Languedoc-Roussillon completes r e cemment par des suivis physico-chimiques et chimiques et par de s analyses isotopiques du & 13 C-CITD vs PDB sur une trentaine de ces sources, a permis de fixer des seuils d'anomalies pour HCO3 -, pCO2 et le pH au dela desquels !'intervention de CO2 profond" est necessaire pour expliquer Jes concentrations en HCO3 observees. La diversite des origines de ce CO2 profond est mise en evidence. les facteurs de cont:role sont repertories et leur importance relative evaluee. 1. Introduction Les implications medicales ou economiques de fortes teneurs en bicarbonates associees au dioxyde de carbone dans les eaux souterraines. ont depuis longtemps suscite des etudes en grand nombre. Celles-ci portent indifferemment sur des milieux sedimentaires ou endogenes. Dans cette note situee en Languedoc-Roussillon, seul le milieu sedimentaire carbonate est pris en compte directement. puisque ne sont utilisees que des sources karstiques au nombre de 650 {PURIDIMAJA 1991) et de quelques forages. Le choix de cette echelle d'observation et de milieu est justifie par le souci d une approche systematique non orientee; de generalisation et d'homogeneite des donnees traitees. Le but de ce travail est de determiner l'origine du bicarbonate et du dioxyde de carbone dans les eaux karstiques et de reconmutre les facteurs qui influent sur leurs concentrations et compositions i s otopiques. 2. Contexte regional Hormis sur sa peripherie montagneuse. le Languedoc-Roussillon est soumis a un climat mediterraneen. De geologie complexe. on retiendra que l'on y trouve d epaisses series carbonatees (jusqu a 1000 m d'epaisseur d'un seul tenant) intensement fracturees lors des phases tectoniques du Cenozo"ique. Elles permettent !'existence de puissants aquiferes karstiques. cela d autant plus qu il y a eu plusieurs occasions favorables a leur developpement depuis le Cretace: climat tropical a certaines epoques du Cenozoi"que. regressions importantes abaissant le niveau marin du Messinien au Quaternaire, et enfin, tectonique distensive, notamment a l'Oligocene et au Miocene. Lars de la phase de compression alpine, au Miocene superieure, se produit le debut de l'exhaussement des zones montagneuses mouvement qui se prolonge actuellement. II y aura des zones saturees a reserves permanentes sous le niveau des emergences grace aux structures plissees. effondrees ou ennoyees par des sediments argilo-sableux Par contre, les structures exhaussees tectoniquement et/ou isolees par !'erosion ne permettront que la realisation de karsts perches. a reserves regulatrices situees au-dessus des emergences, principalement en zone non saturee. Du fait de la diversite extreme des conditions hydrogeologiques et du grand nombre de sources utilisees, les caracteristiques de celles-ci sont egalement tres variees, avec notamment des altitudes d'emergence de -3 l m a 2250m et des debits d etiage de O a 800 Ls 1 3. Methodologie Les donnees physico-chimiques et chimiques utilisees proviennent de 57 eludes de natures diverses realisees entre 1964 et 1992 Elles sont la base d'un inventaire de 650 sources karstiques (PURADIMAJA, 1991; FAILLAT & PURADIMAJA 1995). completees par 30 forages Un lot de 32 sources a ete suivi tout particulierement pour elargir Jes informations fournies par cet inventaire. Des valeurs de & 13 c du carbone inorganique total dissous (CITD = COr + HCO3+ COz) proviennent d'etudes aux environs de Montpellier (YERRIAH. 1986. GRILLOT & BLAVOUX 1986). Les points d'eau concernes etaient tres types par leur situation geologique (proximite de grandes failles crustales) et leurs caracteristiques hydrochimiques et gazeuses. Aussi, pour completer ces donnees. ii a ete fait en 1989 (PURADIMAJA, 1991) puis en 1991 (ESTIV AL, 1991) 14 prelevements sur des sources ou des forages karstiques ordinaires Sept sont situes sur le versant sud-est du Massif de la Gardiole. cinq autres. a debits faibles. avec des temperatures de 14 a 18C, dans une region a anomalies en HCO3 du 6 '" Confe r enc e on Li meston e H ydrology and F issured M edia 85

PAGE 97

Cretace au Nord de la faille de Nimes. Les deux autres prelevements ont ete pris clans deux sources cotieres des Corbieres (23 37), fortement marquees par un apport lagunaire et par de legeres anomalies thenniques positives. A cela viennent s ajouter 6 valeurs concemant les Corbieres obtenues en 1987 (BAKALOWICZ. 1988) Les prelevements ont ete recueillis clans des flacons en polyethylene de 250 cc. Deux centimetres cubes de NaOH a 2 N sans CO 2 ni sulfate de fer et un exc e s de BaCl2 ont ete rajoutes a l'eau immediatement et le flacon bouche avec soin. Les dosages du 13 c ont ete realises au laboratoire d 'Hydrologie et de Geochimie isotopique de l'Universite de Paris-Sud. 4. Donnees disponibles sur les eaux karstiques Dans le cadre de ce travail seuls HCO3 pCO2 et le pH sont employes Leurs distributions statistiques sont donnees dans la figure 1. La teneur en CO2 dissous est le principal facteur de controle des equilibres calco-carboniques. C'est aussi celui dont ii est le plus difficile de donner des ordres de grandeur transposables a l'echelle d'un systeme hydrogeologique karstique a cause de la grande heterogeneite de sa distribution spatiale e t temporelle. Une centaine de sources de la region possedent des donnees qui pennettent le calcul de la pCO2. Cepenclant, apres elimination des donnees douteuses ou des eaux excessivement mineralisees seules 73 ont pu etre employees. L essentiel des valeurs varie de 0.001 a 0.5 atmosphere pour cet echantillon statistique reduit. Deux modes apparaissent vers 0.006 et vers 0.02. Les teneurs moyennes en HCO3 (650 donnees disponibles) peuvent varier de 0 5 a 20 0 meqJ 1 avec un mode a 4-6 meqJ 1 (350 donnees). Ence qui conceme le pH, 203 sources ont des valeurs moyennes s'etenclant de 6 5 a 8,3 avec un mode a 7,4-7 6. Au total 29 valeurs de o 13 C-CITD sont disponibles qui correspondent directement ou indirectement a des eaux de nappes karstiques prelevees clans des forages ou des sources Jes aquiferes principaux etant cretaces ou jurassiques Les o 13 C-CITD varient de -2.5 a -15.4 o/oo vs PDB (tableau 1 ). N' FI F2 F 3 F 3 F 4 F S 627 6 25 626 SS 56 6 2 R3 94 147 2 3 37 8 1 5 !!2... R73D -~ RR7 924 G9 CIS C23 G2S G28 G30 G) \ 86 J 10 NO M F Font d Aurc llc F ENSA F u sn. dtbu 1 F us n. fin F. Av. d'Occilam c F Ai aucloni u c La R oub i n c La M adel e in e Font Fo rte La M oui ll ~rc La Fo u Las O o ux F o u r ada d c F o nl des A v,uc,: Ca Can .man y Cro u. S t V i n cen t La Va lcnl i nc Mas Pl a nt ~ 1!gc de L K'luc ~~ Clf~ ron1 11nn Fa nl lldc G ro ll c F Fillot F Chambon F Sica rd F V J F Rcmv F Karland -2 10 Rftio n s Montpcllicr Gardiol c C orbibc."' ., .. .. N im cs .. G1 r d1olc . TC o HT 6 80 6 70 6.75 7 30 6.6 5 6 30 6.40 6.20 24. 1 7 11 26. 1 6 357 13 2 7 33 21.0 6 93 22 0 6,95 18.7 6 807 2 0 6 7,40 18.5 7 04 20 S 6.96 13 R 7 14 14.3 6 88 1 4 0 6,95 13 9 6 90 1 5 R 6 94 16 S 6,70 16 2 6,68 1 4 6 6 6 1 IS ,2 6 7 9 I S.4 1 IS 1 5.5 6 60 17 -1 10 peo 2 (atm) H COJ nf" O2 aim IJC F orma t ion 6 80 0,022 12 60 Pl i oc~nc 7 80 0.033 14.20 7 20 0,031 9.90 Oligattn c 1 2 00 2.50 S R0 0.018 11.30 J uru s 1qu c 9 60 0, 1 04 6 70 .. 1 2.90 0 059 7J1J 1 2.10 0 144 -5, 7 0 1 9.50 0.425 -4 10 .. 3. 36 0 014 1 0. 80 C r C la c C 3.10 0 074 7 R 60 4.46 0 011 12.50 3.56 0 019 1 3 .10 .. 4.56 0 027 13.40 Jura~ s iquc 4 .40 0 0377 -7. 80 Crftad 5, 8 0 0 0 1 1 R.47 S 99 0 025 9.40 6,7 8 0 04 11.JB Ju u ,siquc 7 06 0.026 14 .2 6 C rf t acf 6 76 0,045 14.49 8 33 0 047 15.41 7.66 0 048 1518 .. 6.50 0.037 10 00 ~a ~. 1quc 8 60 0 092 9 90 8 20 0 085 11 63 1 1.05 0 131 7 .3 8 7 .30 0,065 1 1 .52 S 60 0.02 13 73 1 0.5 0 0.13 7.55 '-' 0 I> H CO : f (m e .1 1 ) Figure I : Histogrammes des pCO2 en atm (73 souces), des HCO3en meq.11 (650 sources) et des pH (203 sources) Onginc YERR )A H. 1 9 86 BAKALOWI CZ. 1981 .. .. PU R AD IMAJA 1991 .. .. .. .. ESTIV AL 1991 .. .. .. Tableau I : Donnees chimiques et isotopiques concernant les sources karstiques dosees pour le o 13 C-CITD P roceedings of the 12 'h I nternational Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 98

5. Interpretation et discussion Dans l'air atmospherique la tension en CO2 est de 0 0003 atm en moyenne. Ce sera celle de l'eau de pluie et d une maniere generale. de toute eau largement au contact de l'air atmospherique depuis un temps suffisant pour que l'equilibre soit atteint (cours d eau ) Une telle tension de CO2 ne pennet d atteindre qu une concentration de 2 meqJ 1 e n HCO3 -, valeur presque toujours depassee par Jes eaux en milieu carbonate Aussi. l'origine la plus frequente et la plu s largement adrnise du CO2 des eaux est l'air des sols (TROMBE, 1952; SCHOELLER 1962, 1969; ROQUES 196 3 ; RENAULT 1968.1971 1972 ; RIGHTMIRE & HANSHAW. 1973 ; BAKALOWICZ. 1979; etc ... ). En fonction des teneurs en CO2 rencontrees le plus souvent dans ceux-ci en certains points. ainsi que dans Jes eaux et en tenant compte d'un effet d echelle, nous avons retenu un seuil d anomalie en pCO2 egal a 0.03 atmosphere pour les emergences karstiques du Languedoc-Roussillon. II est interessant de constater qu une telle valeur de pCO2 correspond pour une eau a 10-20 C tres nettement bicarbonatee calcique a l'equilibre vis-a-vis de la calcite a des teneurs en HCO3 de 6-7 meq.1 1 et un pH de 6 97 1. Cela joint a l allure de la distribution statistique des valeurs de HCO3 et des pH (figure l) nous amene a fixer a ces valeurs Jes seuils d anomalie On suppose que pour depasser ces seuils (HCO 3, pCO2 et H +) un apport de CO2 a forte pression partielle est toujours necessaire Cela ne signifie cependant pas que du CO2 profond ne participe jamais a l a fonnation de HCO3 au-dessous du seuil 6 9-7,1 meqJ 1 mais ii n est pas alors absolurnent necessaire pour expliquer la concentration atteinte. Dans ce travail le CO2 du sol sera la source dite "nonnale en ce compose. Pour preciser l origine multipolaire du CO 2 qui a participe aux equilibres calcocarboniques ii sera fait appel au o 13 CITD > dont les ordres de grandeurs ont ete compiles l pour differents milieux ( CRAIG 1953; 0 s RANKAMA, 1963 ; PILOT 1974 ; BARNES G iii 1979 ; HOEFS, 1987) et reportes dans la '"' figure 2. c.() -~ Celle-ci montre qu il y a globalement une relation entre le 0Bc1TD u d'une part et d autre part HCO3 -. Les 29 points representatifs adoptent cependant 5 l allure d une croix qui indique une origine multipolaire du carbone. La branche A correspond a des eaux dont le carbone est enrichi en 13c et cela d autant plus que HCO3et pCO2 soot forts et le pH faible. II s agit la de la classique courbe de melange 10 8 qui met en competition le CO2 du sol et le CO 2 profond magmatique ou metamorphique (FONTES et al. 1963; ALLARD 1979: PILOT 1974 ; FRICKE, 1978 ; etc . ). Des valeurs aussi fortes pourraient aussi s expliquer par la fennentation methanique d'apres BLA voux 1 5 et LETOLLE (1995). Ce mecanisme etait deja propose par WIGLEY et al (1978), en plus E "' aj E 56 ,, 5 5 ,, de celui de precipitations et dissolutions ; concomitantes (WIGLEY, 1976), qui peuvent 1 :-: 1 mener a des o 13 C de l'ordre de -5 a -3 % 0. :" 1 ~ I mais sans accroissement de la quantite de 2 0 -t-:-:'.'TT"--,-..---,--,---,---+--,--,----.-,---,--,---,-..---,--,---,--~ carbone, ce qui n est pas le cas ici 5 1 o 1 5 H C03 ( m1 H1) : ~: La branche B correspond a des eaux O lal ( 8 17 88 7 924 ) au dela du seuil d'anomalie Figure 2 : ol3C-CITD (%0 vs PDB) en fonction des 0 en HCO 3-, mais sans enrichissement en HC03en meq.l" 1 avec indication des domaines de 13c. Le pole distal responsable de la position de cinq points est dil au co 2 variation des 0 13 C de divers composes carbones profond biogenique produit par la matiere organique fossile. La branche C, peu marquee est materiaJisee par les deux sources littorales (23 et 37). Le melange de l eau karstique avec de l'eau d etang a fort o 13 C-CITD mais peu de HCO3 peut entrainer un deplacement de ces points hors de la courbe de melange A. avec de plus modifications des concentrations des especes carbonatees. de l'equilibre et du pH (WIGLEY et PLUMMER. 1976). Ici. le CO2 excedentaire doit etre profond et inorganique. La position des points 56 et 147 pourrait s'expliquer par la dissolution de gypse par de l eau a l'equilibre avec la calcite, ce qui pennettrait la dissolution d e dolomite et la precipitation de calcite (WIGLEY, 1978 ). 80% d'apport en gypse suffirait a expliquer les o 13 c observes. mais toutes les eaux du meme type ne semblent pas concemees ... 6 1h Co nfe r e nc e on Li mestone H ydrology and Fissured M edia 87

PAGE 99

Enfin la branche D. mal definie, est le domaine des eaux karstique banales. Jes plus nombreuses bien que peu representees ici, car tres rarement choisies Le CO2 qui intervient dans les equilibres calco-carboniques de ces derni e res provient tres majoritairement du CO2 produit dans le sol et peu de l'air atmospherique. Malgr e l'origine plurimodale du carbone et les m e langes possibles la complexite des processus isotopiques ( effets isotopiques. fractionnement) chimiques et biochimiques (cinetique dissolution precipitation variations de temperature degazage du CO2 conditions redox confinement) mis en jeu. qui toutes peuvent dans certaines conditions mener a un enrichissement isotopique (WIGLEY 1976 et 1978) Jes eaux Jes plus caracteristiques peuvent etre identifiees en ce qui concerne l'origine de leur carbone. surtout si l'on tient compte d'un ensemble de pararnetres tels que pH Eh t 0 pCO2 ions majeurs ... II est remarquable qu aucun point ne tombe entre Jes branches A et B, ce qui semble indiquer pour l'echantillon disponible. que l'une ou l autre source de CO2 profond dominera exclusivement. Cela pourrait etre du a la plus grande profondeur des apports en CO2 magmatiques et metamorphiques. qui. la ou ils sont en contact avec la matiere organique fossile des ensembles sedimentaires lors de leur trajet vers la surface. s appauvriraient en 13 c par echange avec celle ci. Des travaux regionaux recents, dont ceux de GRILLOT & YERRIAH-MASSON (1986), GRILLOT & SCHOELLER (1989), BLAVOUX & DAZY (1990) et d ARTHAUD et al. (1994 ) apportent des renseignements sur la presence de CO2 profond entre Rhone et Pyrenees. Pres d une vingtaine de forages pouvant atteindre 5000 metres de profondeur sont connus pour presenter des venues gazeuses constituees en proportions variables de CO2 et Cl-4 et d autres gaz. ll en est de meme pour certaines mines de charbon ou de metaux ainsi que pour des s ources carbo-gazeuses du socle varisque. BLA VOUX e t DAZY (1990) ont rassemble des valeurs de o 13 C CO 2 provenant de ces divers milieux Ces o 1 3 c sont compris entre -3 % o et 14 % 0 vs PDB. Les valeurs les plus faibles resulteraient "soit du melange entre CO 2 profond s echappant par les fractures du socle et un CO2 organique fossile piege dans les couches de charbon. soit d un simple echange isotopique entre ce CO2 profond et le carbone organique ". Les hypotheses emises sont done en accord avec celles de ces auteurs L 'explication geologique des venues du CO2 profond serait done de deux types principaux : Oxydation de la matiere organique contenue dans les roches disseminee ou en gisements ( charbon petrole). Degazage des magmas ou metamorphisme des roches carbonatees et remontee du gaz a la faveur d accidents La distribution spatiale des concentrations en ions bicarbonate (figure 3), disponibles pour Jes 650 sources et Jes 30 forages. permet de reperer a l'echelle regionale Jes principales zones d anomalies en HCO3 (70 sources et forages) et done en pCOz et en pH donnees bien moins disponibles La principale est situee dans la region des Garrigues de Nimes et d Avignon (l) constituee de terrains cretaces calcaires et marneux. plus ou moins plisses. On y trouve Jes sources 815 817 873D 887 924. Leurs faibles valeurs de o 13 CITD (-15.4 a -14%0 une seule a -11%0) et les teneurs elevees en HCO3, qui necessitent des pCOz superieures a celles que permettent Jes sols a l'echelle regionale, exigent des apports en COz profond. mais a faibles o 13 c. done d'origine biogenique (gisement de charbon. de petrole .. ) Une telle origine n est pas liee forcement a des failles importantes. Cela pourrait expliquer une distribution spatiale apparemment aleatoire dont le controle structural est difficile a mettre en evidence. bien que certain. D autres secteurs contiennent des sources au dela du seuil. D une part dans la region entre Beziers et Lodeve ( 2 ), d autre part entre Montpellier et Sete. le versant sud-est de la Gardiole (3) et entre Narbonne et lamer, le Sud du Massif de la Clape (4). II est remarquable de constater sur la figure 3 que Jes sources 625 626, 627 et Jes forages peu profonds de la partie NE du versant SE de la Gardiole sont situes a proximite de la grande faille normale de Nimes cachee par d epais sediments plio-quaternaires dans la partie du massif parcourue par de nombreuses failles N-S Le Languedoc-Roussillon ayant ete soumis a des phases de distension avec amincissement crustal a l'Oligo miocene. des apports en CO 2 magmatique et metamorphique ont pu se faire a cette occasion (GRILLOT et al., 1983). C e CO2 a pu etre Stocke dans des structures geologiques OU se d e placer lentement a travers l'epaisse serie sedimentaire (ARTHAUD et al.. 1994). Les failles E-0 et NE-SO sans doute voies de cheminement privilegiees a cette epoque, ne joueraient plus actuellement qu un role secondaire de ce point de vue. a cause des contraintes horizontales maximales orientees N-S a NO-SE. Le role principal appartiendrait done aux accidents N-S a NO-SE. qui beneficient de nos jours de mouvements en extension ou en decrochement (GRILLOT & SCHOELLER 1989). 6. Conclusion Plus de la moitie des sources karstiques du Languedoc-Roussillon presentant des anomalies hydrochimiques sont dans ce cas du fait des bicarbonates Les pCO2 calculees pour atteindre de telles concentrations sont incompatibles avec Jes pC02 en grand des sols. Les exced e nts de CO2 n e cessaires ne peuvent avoir. pour l'essentiel. qu une origine profonde. Le s teneurs en 13 c du CITD montrent q ue cette origine profonde peut e tre variable soit en relation avec du carbone biogenique ( charbon petrole etc ... ) a o 13 c faibles soit avec du carbone mineral. magmatique ou metamorphique, a o 13 c forts. Dans le premier type d'apport profond magmatique et "metamorphique le o 13 c des s ources karstiques peut etre superieur a -10%0 et dans le second type d apport profond "biogenique fossile le o 13 C des sources sera inferieur a -lW'oo en premiere approximation. Des o 13 C inferieurs a lW' oo ne signifient done pas obligatoirement une origine biogenique superficielle mais peuvent indiquer des apports en CO2 biogeniques fossiles profonds. surtout si Jes teneurs en HCO3 sont au-dessus du seuil de 6 9-7 1 meq.1 1 En fail les melanges en provenance de differentes sources de carbone et Jes processus en j eu peuvent etre plus complexes pour certains points. II ne sera done pas toujours possible de proposer des origines determinees pour le o 13 c de certaines sources Dans Jes Garrigues de Nimes et de Montpellier. ii semble que ces differents apports puissent jouer un role dans de nombreuses sources et dans Jes nappes karstiques qui les alimentent. On est en droit 88 P roceedings o f t he 12 th Inte rn ational Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 100

de se demander si !'augmentation des valeurs de pCO2 vers le bas dans nombre de cavites (RENAULT 1972 ) ne s expliquerait pas principalement par une telle origine Ce phenomene serait alors generalisable a d autres regions karstiques limitrophes du Massif Central tels le Quercy et le Perigord ou des valeurs de pCQi superieures a 0 03 peuvent etre observees dans !'atmosphere du fond des grottes et avens. Le dosage du ol3C-CO2 de l'air des cavites devrait permettre de tester cette hypothese, si Ies difficiles conditions de prelevement le permettenl Sources et forages ( Gardiole ) a v ec HC0 3 >7meq.1 1 Failles normales oligo -a quitaniennes Li,.:~-,,,,, D Cenozoi"que et anthropozotque Calcaires et mames s urtout e ocenes ----,-~ E3 Permien et ~ esozo"ique. Calcaires. mames et gres TTTl Paleozotque Calcaires dolornitique s IL.LJ cambriens et devoniens 2 51cm Figure 3 : Carte de localisation des sources karstiques du Languedoc Roussillon et de forages (Gardiole) ayant des concentrations en HC03 superieures a 7 meq.J 1 Remerciements : Ceux-ci s adressent au Professeur W. BALDERER pour ses remarques pertinantes qui ont perrnis d ameliorer !'interpretation des donnees. References ALLARD, P. 1974. 13 c; 12 c and 34 s;3 2 s ratios in magmatic gases from ridge volcanism in Afar. Nature 282 56-58. ARTHAUD F. ; DAZY J. & GRILLOT J.C. 1994 Distribution of deep carbon dioxide in relation to the structure and tectonic evolution of south-east France. Geodinamica Acta Paris 7 2 86-102 BAKALOWICZ M 1979. Contribution de la geochimie des eaux a la connaissance de l'aquifere karstique et de la karstification. These Doc Sc nat Paris 6 257 p. BAKALOWICZ, M. 1988. Geochimie des sources carbonatees des Corbieres et formations de travertins, U A. 903 CNR S et A T P Piren, Aix-en-Provence 209-225 6 th Confe r ence o n L imestone H ydrology an d Fissured Media 89

PAGE 101

BARNES H.L. 1979 Geochemistry of hydrothennal ore deposits. Second edition A Wile y lnters c ien ce publication, John Willey & Sons. New York 567 p BLAVOUX B & DAZY. J. 1990. Caracterisation d'une province a C0 2 dans le bassin du Sud-Est de la France. H y dr o geologie 4. 241-252 BLA VOUX B. & LETOLLE, R. 1995 Apports des techniques isotopiques a la connaissance des eaux souterraines. G e ochronique, 54 12-15 CRAIG H 1953. The geochemistry of the stable isotopes. G e o c him Cosmochim Acta 3 53-92 ESTIV AL V. 1991. Etude des phenomenes induisant des anomalies hydrochimiques et isotopiques dans les eaux de magasins carbonates de la partie Sud-Est de la Gardiole (Herault France) Mem D .E.S Montpellier II 59 p. FAILLAT J.P. & PURADIMAJA DJ. 1995 Evaluation a l'echelle regionale des contraintes hydrochimiques sur la gestion des eaux souterraines karstiques. Exemple du Languedoc Roussillon. Hydrog e o/ogie, 1995, n 1, 97-112 FONTES J C. GLANGEAUD L. GONFIANTINI R. & TONGIORGI E 1%3 Composition isotopique et origine des eaux et gaz thennaux du massif central. C R Acad S c. Paris t. 256 472-474 FRICK E, K. 1978. 13 ct 12 c investigation of borehole of dry COz gas, with remarks on the development of output and pressure with in 10 years, as well as references to other 13 ct 12 c data Conferences of Cieplice SPA (Poland ), Warsaw 204-211. GRILLOT J.C .. BOUSQUET J.C. PHILIP H. & DAZY J 1983. Hydrothennalisme carbo-gazeux. radio activite et sismotectonique: exemple du Languedoc-Roussillon (France). Bull BRGM ., 4 313-320 GRILLOT J.C. & BLAVOUX, B 1986. Approche geologique et geochimique des eaux minerales carbo-gazeuse : quelques exemples du S.E. de la France (eaux minerales carbo-gazeuse) J Fran,ais d H y drologie 17 fasc. 3 263-272 GRILLOT J C. & YERRIAH-MASSON, J. 1986. Apropos du geothennalisme carbo-gazeux dans le S.E de la France. C.R A c ad S c. Paris t. 303 serie II no 7. 623-626 2 fig. 1 tab!. GRILLOT, J C. & SCHOELLER, M. 1989. Exemple d approche pluridisiplinaire dans la caracterisation d eaux thermales carbo-gazeuses. Revue des Scien c es de/' eau 2, 211-248 HOEFS J. 1987 Stable isotope geochemistry. Third edition SpingerVe rlag Berlin-Heidelberg, 241 p. PILOT J. 1974. Les isotopes en Geologie (Methodes et Applications ) Dain editeurs Paris 174 p. PURADIMAJA DJ. 1991. Differenciation hydrochimique et isotopique des emergences karstiques du LanguedocRoussillon (France). These de do c toral Montellier II. Nov. 1991 267 p. RANKAMA K 1963 Progress in isotope geology. lnter s cience publishers a division of John Wiley & Sons. New York, 188-245 RENAULT, P. 1968. Speleologie physique. Sur la distinction de plusieurs regions karstiques en raisons de la teneur en anhydride carbonique des atmospheres de grottes. C.R A c ad Sci Paris serie D t. 267 22882290 RENAULT P 1971. La teneur en anhydride carbonique des atmospheres de grottes. Bull de I' Association des Geographesfran,ais no. 389-390 241-245. RENAULT P. 1972. Le gaz des carvemes. S c ience progres Decouverte, n 3443 mars 1972 8-12 RIGHTMIRE C.T. & HANSHAW B.B. 1973. Relationship between the carbon isotope composition of soil CO2 and the carbonate species in groundwater Water Resources Resear c h 9 4 958-%7 ROQUES H. 1963 Sur la repartition du C02 dans les karsts. Ann speleol ., t. XVII, fasc 2 141-184 SCHOELLER H 1962. Les eaux souterraines. Traite, Masson Ed Paris, 641 p. SCHOELLER H. 1969. L acide carbonique des eaux souterraines. Bull BR G M ., sect. III 1 1-32 TROMBE, F. 1952. Traite de speleologie Paris Pa y ot Ed. 149 p. WIGLEY T.M.L. 1976. Effect of mineral precipitation on isotopic composition and 14 c dating of groundwater. Nature Vol. 263 219-221. WIGLEY T.M.L. & PLUMMER L.N 1976. Mixing of carbonate waters G e o c him Co s mochim Acta 40 989995. WIGLEY T M.L.. PLUMMER L.N. & PEARSON F.J. 1978. Mass transfer and carbon isotope evolution in natural water systems. G eo chim Cosmochim. A c ta 42 1117 1139. YERRIAH. J 1986. Le thermomineralisme carbo-gazeux du S.E. de la France ( domaine sedimentaire) dans son contexte sismotectonique. Th e se 3eme cy cl e. Montpellier II 108 p .. l carte h t. 90 Pr oceedings of t h e 1 2 h In te rn ati on a l Co ngr ess of Speleology, 1997, Switzerland Volume 2

PAGE 102

Approche statistique de l impact des variations climatiques sur la reponse chimique des sources karstiques Par Domenico Alessandro Grasso & Pierre-Yves Jeannin Centre d hydrogeologie de l Universite, Rue Emile-Argand 11 CH-2007 Neuchatel, Suisse Res u me Une methode d analyse statistique a ete developpee afin de tenter de detecter Jes facteurs Jes plus importants qui influencent la reponse chimique des sources karstiques L'analyse en composantes principales a permis de deceler l importante influence de I'etat hydraulique du systeme (debit de la source) sur la reponse chimique Une relation mathematique permet de simuler la variation du TAC en fonction du debit. Les ecarts entre la chronique ainsi simulee et la chronique observee montrent !'existence de variations s aisonnieres qui semblent resulter des variations de la production de CO 2 dans les sols, c'est-a-dire dependre du climat. Cette influence est cependant nettement subordonnee a celle du debit. La methode presentee permet en principe de mettre en evidence influence d'autres parametres sur la reponse chimique ou encore !'influence a long terme des variations climatiques Abstrac t A statistical method has been developped to attempt to detect the factors playing a major role in the chemical response of karst springs. A principal component analysis (PCA) of the data illustrated the importance of the spring discharge ( hydraulic conditions) on it s chemistry. A methematical relationship allows one to simulate the variation in alkalinity with respect to discharge Residuals between the simulated and the observed time series shows that season depending processes (variations in CO 2 production in soils depending on climate) are also playing a role Meanwhile this role is clearely subordinate to the one of the discharge Le method presented here basically allows one to detects the influence of any other parameter on the water chemistry or also the long term influence of climate Introduction La reponse chimique d une source represente !'evolution dans le temps de !'ensemble des elements chimiques dissous dans l'eau Nous no u s proposons d'evaluer l 'effet des variations climatiques sur la reponse chimique des sources karstiques qui est fonction de plusieurs facteurs concomitants : hydra u liques, geologiques climatiques, anthropiques, etc Nous presentons ici une approche statistique qui a pour objectif de mettre en evidence l es effets respectifs de chacun de ces processus L effet predominant du debit (l'etat hydraulique du systeme) sur la reponse chimique est etudie en premier, puis soustrait des reponses observees afin de mettre en evidence l'effet des autres parametres et du climat plus particulierement Donnees a disposition Les donnees recueillies a la source de l'Areuse (NE, Suisse) ont principalement servi pour cette etude Les donnees sont representees par des mesures en continu, sur cinq annees du debit, de la temperature et de la conductivite electrique, ainsi que par des analyses mensueUes des elements chimiques majeurs et des isotopes stables de l'eau. Les parametres mesures sont done : debit, Temp ., Cond, pH, TAC, Ca, Na K, Mg, NO , Cl, SO 4 Les donnees chimiques mesurees par KIRALY & MULLER (1979) entre octobre 77 et septembre 79 ont egalement pu etre utilisees. L'avantage de ces donnees est qu'elles presentent un echantillonnage tres serre (900 analyses sur deux ans), foumissant des valeurs de debit, T, Cond, TAC Ca Na, K, Mg L'Areuse prend naissance pres du village de St-Sulpice, dans la partie orientale du Val-de-Travers, une importante vallee du Jura neuchatelois ( coordonnees 532'250/l 95'975/790 m ). Le debit moyen annuel de la source est de 5 0 m /s. Analyse en composantes principales effet de l etat hydraulique du systeme Pour detecter !'influence respective de differents facteurs sur le comportement chimique des sources karstiques ii est necessaire de rechercher des correlations entre ceux-ci et la reponse chimique de la source. Nous utiliserons !'analyse en composantes principales (ACP) appliquee aux series d analyses chimiques pour deceler Jes variations non aleatoires de la reponse chimique des systemes karstiques Suivre !'evolutio n de la composition chimique d'une source revient a observer dans un espace a N dimensions (N egal au nombre de parametres) la trajectoire dans le temps d'un point representatif de sa composition. Dans notre analyse nous avons retenu 12 parametres: T Cond, pH TAC, Dto Ca Na, K, Mg N0 3 Cl, SO 4 L ACP (DAVIS 1986) permet de remplacer l'espace defini par les 12 parametres mesures par un espace a douze dimensions dont les axes (facteurs) sont positionnes selon les directions d'allongement principales du nuage des points mesures (vecteurs propres du nuage) Ces facteurs sont analytiquement independants (tous perpe n diculaires les uns a u x autres). Le premier facteur est positionne dans la principale direction d allongement du nuage II represente ainsi le p lu s fort po u rcentage de la variance totale du nuage Chaque facteur consecutif represente un pourcentage de plus en plus faible de la variance totale du nuage Dans l e cas de la source de I' Areuse Jes cinq premiers facteurs representent plus du 95 % de la variance totale de la serie Ainsi les plans factoriels formes du premier facteur couple aux quatre autres ont ete analyses L'axe factoriel I (ou FI) est bien correle avec tous Jes parametres representatifs de l a mineralisation totale : l es cations ( surtout le Ca), la durete totale, le TAC, la conductivite, et dans une moindre mesure Jes sulfates. On peut done considerer le facteur I comme un axe mineralisation" Le facteur II (ou F2) est representatif surtout des Nitrates et du pH (dans une moindre mesure). II peut done etre admis que ce facteur est lie a l'activite biologique dans Jes sols (dependant du climat) et a l'activite anthropique SeuJ le plan factoriel defini par les axes I et II a permis de mettre en evidence des variations non aleatoires de la reponse chimiq u e. 6 1 Conference on Limestone Hydrology and Fissured Media 91

PAGE 103

Pour analyser l'effet eventuel du debit sur la chimie des eaux, l'hydrograrnrne des sources karstiques a ete schematiquement decompose en quatre "e tats hydrauliques" differents : la montee, la decrue rapide, la periode de transition, le tarissement. A partir de cette distinction, Ja figure I montre qu'il existe une relation nette entre J'etat hydraulique et la mineralisation, c'est-a-dire entre le temps de sejour de l'eau dans le systeme et sa mineralisation En effet, le nuage des analyses correspondant a la decrue rapide est place dans la partie gauche du plan factoriel (FI F2) ; les points ont done des valeurs faib le s selon FI, ce qui correspond a une faible mineralisation Ces eaux ont des temps de sejour tres faibles et n'ont pas pu atteindre un equilibre chimique avec Jes roches encaissantes. Au fur et a mesure que le debit diminue le nuage des observations se deplace vers la droite dans le plan factoriel. Des eaux avec une mineralisation de plus en plu s importante sont evacuees, ce qui correspond a des temps de sejour de plus en plus longs et peut etre explique par Ja vidange des zones a permeabilite de plus en plus faible. 20 Dtbits des tchantillonages chimiques classts selon l'ttat hydraulique de la source --0.ah-_,~ 5 ~-~-.~, ~~~--0-.~S~ F. ~ c, ~ ou ,~, ~o~-."-~,--,o.s~~~~~ Figure 1 : La comparaiso11 entre Jes deux graphiques mo11tre la correlatio11 exista11te e11tre l'etat hydraulique du systeme (fig. a) et la mi11eralisatio11 a la source (fig. b). Le report, sur le plan factoriel, des analyses chimiques prises pendant la montee des crues, montre une forte dispersion qui ne peut pas etre mise en relation avec le debit seulement. Cette dispersion est liee a Ja poussee des eaux stockees anterieurement dans le systeme par Jes eaux des precipitations (effet piston) Cet effet depend du comportement hydraulique propre du systeme, des facteurs extemes comme l'intensite de la pluie et de l'etat hydraulique du systeme avant la crue. La presence de nuages bien distincts dans le plan factoriel entre decrue et tarissement montre !'existence d'une relation claire entre debit et mineralisation L'effet de l'etat hydraulique sur Jes pararnetres chimiques est tres important et rend difficile la mise en evidence d' autres influences (par ex em pie, les variations climatiques saisonnieres, l'activite anthropique, etc.). En effet, les plans factoriels n ont pas mis en evidence d'autres variations non aleatoires de la reponse chimique des sources, ce qui montre que l'effet de l'hydraulique sur les variations chimiques est largement predominant. A ce stade, Jes variations de la composition chimi que liees a d'autres agents ne peuvent done pas etre detectees Pour tenter de mettre en evidence les effets d 'a utres agents, ii faut analyser Jes residus de la relation entre reponse chimique et debit, et pour cela, ii faut soustraire de la reponse chimique observee la partie "expliquee" par la relation entre chimie et debit. Dans ce but les hypotheses suivantes doivent etre admises : la variation observee de la reponse chimique est fonction : du regime hydraulique de la source; d'autres agents (cli mat, activites humaines, etc.); des erreurs (aleatoires) de mesures Relation concentration debit L'existence d'une relation non aleatoire entre mineralisation et debit a ete demontree ci-dessus; ii convient maintenant de definir I 'allure de cette relation. Admettons que la concentration des elements chimiques a la source est influencee par trois processus : la dissol'ution, la dilution et le transport convectif. Par hypothese simplificatrice, ii est admis que ces processus sont instantanes et homogenes dans !'ensemble du systeme. L'efficacite de ces processus est fonction de la vitesse et du volume d'eau qui transite dans le systeme. La concentration instantanee ll. la source est egale ll. : o/s C cl) = m = masse transportee; v = volume d'eau M ctl Ccl)=-Q (1) M el) = masse instantanee transportee (de bit massique ); Q co = debit a la source (debit volumique). (I) (2) La masse M el) est fonction du debit. A un debit Q (l) eleve correspond une vitesse elevee de l 'ea u dans le systeme et un volume important de roche en contact avec l'eau Par consequent, la quantile de roche dissoute est plus importante et done le transport de masse vers l'exutoire egalement. La relation entre la masse transportee par unite de temps et le debit, trouvee empiriquement ll. partir des mesures experimentales, est de la forme suivante: M ctl =A* Q<,/3 (3) A = masse transportee pour un debit unitaire (concentration); Selon la valeur prise par P on peut envisager trois cas di fferents : -P < I !'augmentation du debit est plus importante que celle de la masse transportee; -P = I !'augmentation du debit augmente lineairement avec celle de la masse transportee; -P > 1 !'augmentation du debit est moins importante que celle de la masse transportee ; Des valeurs de p negatives (diminution de la masse transportee alors que le debit augmente) n'ont jamais ete observees a partir de nos mesures. En introduisant la relation (3) dans la (2) on obtient : A *Qc,/3 A Ccl) = Q = (I) Q(l)(I-/J) et en remplac;ant a= (I P) A -aLn(Q(t)) C(I) =-a-=A*e (4) Cr ) La concentration est une fonction exponentielle du logarithme du debit. 92 Proceedings of the 12 1 h International Congress of Speleology, 1997 Switzerland Volume 2

PAGE 104

e ll Afl -a*Ln(Q(I)) (1) "V) e (7) Cette courbe a alors une amplit ude tres voisine de celle de la courbe mesuree (figu re 4) Moyennant d'admettre Jes hypotheses inherentes a cette methode, ces simulations indiquent que Jes variations du TAC sont fonction principalement du debit mais aussi, dans une moindre mesure, des saisons, done des variations climatiques II subsiste de petites differences entre Jes deux courbes, principalement pendant Jes crues. Elle sont dues au parametre a qui est considere comme constant dans la simulation, mais qui en realite varie legerement d'une crue a l'autre, principalement en fonction de l'etat hydraulique du systeme precedant la recession Pour tester la validite de la methode, ii convient de te ster la ressemblance entre la serie simulee et la serie observee, en mettant en evidence Jes frequences caracteristiques ou Jes periodicites des series. L'analyse spectrale des series temporelles permet de tester la similitude entre deux series Cette methode perrnet de decomposer la variance totale d'une chronique en fonction de ses frequences (Box &JENKINS 1970, MANGIN 1984, GRASSO & JEANNlN 1994). Figure 4 : Chroniques TAC mesure, TAC simule en fo11ctio11 du debit et des variations saisonnieres. Les residus sont aleatoires. Les spectres des series mesurees et simulees (Fig. 5) presentent trois pies distincts qui correspondent aux memes frequences caracteristiques pour le trois chroniques (fl=0.0028, f2=0,007 l4, f3=0 0229) Ces frequences correspondent a des periodicites de 350 jours (cycle hydrologique annuel), 140 jours (cycle saisonnier) et 44 jours (intervalle moyen entre Jes crues les plus importantes). L'amplitude des pies de la deuxieme simulation (serie c' Q ,A 1)/ ) est tres proche de celle des donnees observees La presence des meme pies de frequence dans la serie mesuree et dans les deux series simulees represente un indice clair de la relation fonctionnelle predominante entre la reponse chimique et l'etat hydraulique du systeme. L amelioration de l'ajustement entre la courbe observee et la courbe simulee !ors de la deuxieme simulation atteste de l'effet des variations climatiques saisonnieres sur les variations du TAC dans les eaux de la source. Cet effet apparrut cependant nettement subordonne au premier. 200 1,0 Spectre TAC mNUr' -Spectre TAC 1 MT1ullllon 100 SptN:t,e TAC r am u\al ioo OOJ F r 6quenc:e (1/)oura) 005 006 0 0 7 Figure 5 : La presence des memes periodicites dans le spectre du TAC observe et simule confirme la relation fonctiormelle entre la reponse chimique et l'etat hydraulique du systeme. Conclusions generales Le meme type d'analyse a ete applique a trois autres sources du Jura et des Preealpes suisses Malgre des contextes hydrogeologiques et climatiques bien differents, Jes resultats generaux obtenus sont restes pratiquement Jes memes. II ressort done de cette etude que : Le debit joue un role tres important sur la reponse chimique des sources karstiques. Ceci indique clairement que analyse du comportement chimique d'une source karstique sans prendre en consideration son comportement hydraulique n'a pas de sens. Pour tester !'influence generalement faible d'autres facteurs que le debit sur la reponse chimique des systemes karstiques, en particulier du climat, la methode presentee ici consiste a modeliser l'influence du debit sur la reponse chimique, puis a soustraire cette influence de la reponse observee et a analyser la reponse residuelle La methode presentee ici a ete appliquee au TAC en vue de tester l'hypothese d'une influence des variations de la pression de CO 2 dans les sots dependant du climat sur la mineralisation des sources Les resultats confirment cette influence, qui est cependant nettement subordonnee a celle de l'etat hydraulique La methode presentee peut etre utilisee pour tester d'autres hypotheses; ii faut cependant relever que les residus analyses apres retranchement de l'effet du debit sont faibles, entaches de variations aleatoires et dependent des simplifications admises dans le modele utilise. Les effets recherches sont alors du meme ordre de grandeur que Jes incertitudes inherentes aux mesures et a la methode d'analyse. La methode proposee ici est en principe utilisable pour detecter Jes variations a long terme de la reponse chimique des sources karstiques liees aux variations climatiques Cependant par rapport aux variations annuelles du climat, Jes variations climatiques a moyen terme (5 a I 0 ans) sont tres faibles et de ce fait probablement pas detectables Sur les cinq a sept annees de chroniques disponibles aux 4 sources etudiees, aucune tendance significative n'a pu etre degagee. Seules des chroniques nettement plus longues (30 ou 50 ans) pourraient fournir d'eventuels resultats interpretables. Bibliographie Box, G P & G M. JENKINS 1970. Time series analysis, forecasting and control. Ed. Holden-Day, San Francisco, 553 pp. DAVIS, J C. 1986 Statistics and data analysis in Geology (second edition) John Wiley & Sons, inc, New York 646, pp. GRASSO D A. & P.-Y JEANNIN I 994 Elude critique des methodes d'analyse de la reponse globale des systemes karstiques : Application au site de Bure (JU, Suisse) Bull d'Hydrogeologie 13-1994 Neuchatel : 87-113. KLRALY, L. & I. MOLLER 1979. Heterogeneite de la permeabilite et de l'alimentation dans le karst : effet sur la variation du chimisme des sources karstiques. Bull. Centre d'Hydrogeologie Univ Neuchiitel, 3-1979 : 237-282. MANGIN, A. 1981. Utilisation des analyses correlatoire et spectrale dans l'approche des systemes hydrologiques C.R. Acad. Sci. 293, Paris : 401 404 SCHOELLER, H.-J I 980. Influence du climat, de la temperature sur la teneur en HCO et H CO, des eaux souterraines. J of Hydrology 46: 365-376. 94 Proceedings of the 12 '" International Congress of Speleology, 1997 Switzerland Volume 2

PAGE 105

Geochemical and geomathematical control of the karstic waters composition of the pan de Guajaib6n and surroundings, Sierra Del Rosario Pinar Del Rio Cuba by J e sus M. P a j 6n -M orej6 n *, Arge li o V. d e l a Cr u z** J ulio J. V ald es**, A lbe rt o E. Gar cia & C a r lo s M. C ru ~* Instituto de Geofisica y Astronomia Ministerio de Ciencia, Tecnologia y Medio Arnbiente ** Instituto de Cibemetica, Matematica y Fisica. Ministerio de Ciencia, Tecnologia y Medio Ambiente Abstract Cluster and Factorial analysis methods were applied to the chemical-physical data in the winter and summer geochemical periods 1-2/1984 and 9/1984, with the purpose to characterize the geochemical and chemical-physical behavior in the Pan de Guajaib6n and surrounding waters, thus as the study of the relation between typology defined by geomathematical methods and geological and hidrogeological criterion. The original matrix was conformed with the original variables; T pH, SPC (25C), CO 2 (mg/1) and ionic contents (as mg/1) of Ca 2 +, Mg 2 +, Na + +K +, HCO 3 so t and c1 Was studied the seasonal variation of some index and ionic contents in representatives points of each group of waters, during the period 1/1984 and 2/ 1986 From the correlation analysis the signification and coherence between the original chemical-physical variables is analyzed. The characterization and punctual analysis, for the dynamic of dissolution-precipitation in the Pan de Guajaib6n and Sierra Chiquita karstic massifs considering the general tendency of the saturation index a nd CO 2 contents, suggest the occurrence of complex mechanisms of water-rock interaction, due to the frequent and intense mixing-waters effects in such massifs Resumen Con el objetivo de caracterizar el comportamiento geoqufmico y quimico-ffsico de la s aguas del Pan de Guajaib6n y sus inmediaciones, asf como estudiar la relaci6n entre la tipologfa de las aguas definida por los metodos geomatematicos y la definida segun criterios geo16gicos e hidrogeo16gicos, se aplicaron metodos de clasificaci6n numerica y analisis factorial a los datos qufmico-ffsico s correspondientes a la campafias geoqufmicas 1-2/1984 y 9/1984, las cuales caracterizan los perfodos de invierno y verano del referido afio. La rnatriz original se conform6 con las variables originales: T, pH. SPC (2 5 C), CO 2 ( mg/I) y los co ntenidos i6nicos (expresados como mg/!) de Ca 2 +, Mg 2 +, Na++K+, HCO 3 so t y CL Se estudi6 la variaci6n estacional de algunos fndices y contenidos i6nicos, en puntos representativos de cada grupo re aguas, durante al periodo co mprendido entre el l /1984 y el 2/1986. Se analiz6 la significaci6n y coherencia de las asociaciones entre las variables qufmico-ffsicas originales, a partir del analisis de la correlaci6n. La caracterizaci6n y analisis puntual de la dinamica de disoluci6n precipitaci6n en los rnacizos karsticos Pan de Guajaib6n y Sierra Chiquita, teniendo en cuenta las tendencias generales de los fndices de saturaci6n y contenidos de CO 2 sugieren la ocurrencia de complejos mecanisrnos de interacci6n agua-roca, dado los frecuentes e intensos efectos de rnezclas de aguas que ocurren en tales macizos. 1 Introduction The present paper contains a brief compress of the geochemical and geomathematical resu l ts obtained in the Pan re Guajaib6n massif in the We s tern Karst of Cuba. These results are included in the scope of International Program for Genesis and Evolution of Kar s t (PIGEK), which took place during 1984-1992, in order to study the dynamic of the Contemporary Karstic Processes in a representative Tropical Karst of Cuba as well as the formulation of water-rock interaction models and the resultant morphological effects (PULINA et al., 1984) 2 Characteristics of the karstic area The Pan de Guajaib6n is a karstic massif (699 m a.s.l., A= 13 Km 2 ) located in the most northwestward part of Sierra de) Rosario mountain of the Guaniguanico Cordillera, in the province of Pinar de! Rfo There are two geological formations: The Pan de Guajaib6n formation (al bian-cenomanian ) and Chiquita formation (campanian-maastrichtian). The geological, geomorphological, hidrogeological and climatological characteristics of the st udied area can be ee in details in ( PULINA et al.. I 984) and (RODRIGUEZ et al., 1985) 6 1 h Conference on Limestone Hydrology and F iss ured Med ia 9 5

PAGE 106

3. Results and discussion Geochemical and Geomathematical Characterization of the Karstic Waters: The examined matrix was conformed with the o ri ginal physico-chemical variables, a naly zed 'in situ", according the analytical techniques described by (MARKOWICZ & PULINA 1979) : T, pH, SPC (2 5 11 C) CO 2 (mg/I) and ionic contents (in mg/I) of Ca 2 +, Mg 2 \ Na + +K +, HCO 3 -, SO / and c1 A D ata Base was created with the information of six hydrochemical field-work campaigns in dry and wet periods. The Computer Programs used were AGMAR (FAGUNDO et al., 1986) CLUSTE R and ANFACT (VALDES et al., l98l)(DE LA CRUZ & VALDES, 1985). With the purpose to study the relation between the karstic waters typology defined by geomathematical methods and geological-hydrogeological criterion, as well as to make deep the internal structure of the phenomenons, C l uster and Factorial Analysis were applied LO the physico-chemical data. The figure I show the dendrogram of the numerical classification by the Ward method using the dissimilarity measure (square Euclidean distance ) for 35 individuals and I O variab le s of the campaign 9/ 1984. From the combined analysis and interpretation between campaigns, we can deduce t h e presence of five gro up s of waters: Group I: Non-karstic surface waters (low mineralization and insaturate values of RSC RSD and RSY) Group 2: Waters of the aeratio n zone of Pan de Guajaib6n massif (oversaturate values of RSC a nd RSD ; insatura t e values of RSY ; low CO 2 contents). Group 3: Waters of the aeration and discharge zone of Sierra Chiquita massif (oversaturate values of RSC and RSD ; high values of CO 2 respect to the group 2 Group 4: Waters of the discharge zone of P an de Guajaib6n massif (mixing-water effect). Group 5: Spring-surface karstic waters in the zone Pan de Gu ajaib6n-Sierra Chiquita (oversaturate va lu es of RSC and RSD ; high values of CO 2 INDIVIDUOS RES.ARR.CANILLA RES.ARR.CANILLA RES.ARR.CANILLA RES.ARR.CANILLA RES.ARR.CANILLA 2 3 4 5 6 7 8 91011121314151617181920212223 CUEVA CAN.TEND. RES.ARR CANILLA RES.ARR.CANILLA 2--~~ RES.ARR.CANILLA 3 RES.ARR.CANILLA 1---~ RES.ARR.CANILLA 4-------------~ MAN.CONUCO.VIJA 21--------------------~ SUG ANCON 12------~--~ CUEVA ANCON SIF 25-------' AGUADA "ACAGUA 24--------~ CUEVA ANCON ARD 27 CUEVA ANCON CAB 29 CUEVA ANCON TAN 28 SURG.ANCON 11-----' CUEVA ANCON DUC 26 ---------' Grupo IV Grupo Ill Grupo I ARR. DE LA RED 17---------------,------------------, AGUA LLUVIA 35------------~ CUEVA LEON GOUR 3------.-----, CUEVA LEON DUCH 33 Grupo II CUEVA LEON GOUC 31-------' AGUADA MACAGUA 20------------' MAN.BERRO 18--------------------------, AG. C. SAGUA 150M I 23-----------, Grupo V ARR.FLUORESCEIN 14 __________ ..__ ______ ---r-----, SUM ARR.MANEY 16-----------------~ ARR.MINEROS 17---~-~ AGUADA.C SAGUA 22-----' ~---------------. ARR. LAS PI2AS 15-------' MAN TRAVERTINA 19---------------.---~ CUEVA LECH.ESC 30-----------------' 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 2122 23 Figure 1.Dendrogram with the hierarchical classification by the Ward method, for 35 individuals and I O variab l es (campaign 9/1984). 96 Proceedings of the 12 '" International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 107

To study the influence of the different geodynamic factors on the physico-chemical properties of the waters (BAKALOWICZ, 1979), was applied the Factorial Analysis Mode R, with varimax rotation, according the extraction of three factors to apport around 76% of the total variance in the original variable. Factor 2 SPC H co ; 1 Factor 1 /,.__ ............. ______ .f, f',v C l Figure 2.Factorial Analysis. Distribution of the original variables, according the rotate factors (cam paign 9/84). In the figure 2 are plotted three rotated factors, suitable to the summer campaign 9/84. From the ensemble analysis, is possible arrive to the following conclusions: associate to the first factor appear the variables SPC pH, Ca 2 +, HCO 3 and et (it is related with the limestone rock, with strong control in the winter period). With an strong positive contribution to the second factor, the Mg 2 +, SPC, HCO 3 and less the CO 2 (strong control of the limestone-dolomitizate rocks in the summer period on the variables Mg 2 +, HCO 3 SPC and CO 2 ), which is according with the results of (JACOBSON & LANGMUIR 1972). The SO }and Na++K + are associate to the third factor and the T to the fourth factor. An special study has been development, to obtain an adequate interpretation to the factors three and four (influence of biogenetical and degradative processes of organic matter in soils-humic/fulvic acids and CO 2 production. The figure 3 show the dynamic of dissolution for the Pan de Guajaib6n massif in a punctual-time point of the water-rock interaction process with a general increase of the saturation index and mixing-water effect. The correlation matrix between the original variables show a good coherence concerning the main variables of the COrHiO-CaCO 3 system. The chemical denudation calculated for both catchment area of Pan de Guajaib6n-Sierra Chiquita massif during the hydrological year 1984 are (PULINA et al., 1984): Ancon system (112 m 3 /k m 2 .yea r) and Canilla system (40 m 3 /km 2 .year) These results are in concordance with the results of the experimental simulation study with carbonate blocks of the studied area (P AJON & V ALOES, 199 I). A genera l discussion is located in this paper. Acknowledgments We would like to express our greatitude to the colleague Lie. Ismael Hernandez for very helpful discussions and the composition of this paper. 6 t h Conference on Limestone Hydrology and F issured Med ia 97

PAGE 108

Agua de Ll u via Cueva de Luchuza CD t1 cha) T = 26,8 l 0 0 r T = 22 2 SPC = 10 SPC = 2108 pH = 6, 15 I I I 65 p H = 8 log pCOz = 1,05 67 I 1,11111 I I I I 11 log pCOz = 2 95 RSC = 5,56 J I I 11 RSC = 0 39 RSD = -7,40 I I RSD = 0,07 Sunidero Arroyo Maney + T = 26,8 l SPC = 410 pH = 8 22 logpCOz = 2 78 48 :RSC = 0,94 J \ :RSD = 1.,39 Resolladero Arroyo Arroyo de la Sed Cani I la T = 24,0 l Q(l/s) r T = 22 4 SPC = 110 SPC = 225 pH = 7 I 10 pH = 7 50 log pCO z = 2,19 49 400 42 log pCOz = -2 30 RSC = 1 37 J l RSC = -0 19 RSD = 2,75 300 RSD = 0 76 200 100 0 20 22 24 26 28 Fi g ur e 3 D y n a mic of Di ss olution in the Pan de Guajaib6n ma ss if (c ampaign 9/84) References DE LA CRUZ A. & VALDES J.J. 1985 Estudio de las aguas karsticas del Pan de Guajaib6n y su s inmediacione s mediante metodos matematicos de analisis exploratorio de datos Rev Vol. Hid 68, Aiio XXII 1985 pp 25 34 FAGUNDO J.R., VALDES J J. CARDOSO M.E. DE LA CRUZ A. 1986 Algoritmo para el calculo de para.metro s e fndices q u fmico-ffsicos y geoqufmicos en aguas naturales altamente mineralizadas. Rev Ciencias Qufmicas, 17 1-2 pp JACOBSON R.L. & LANGMUlR D 1972. An accurate method of calculating s aturation levels of groundwater with re s pect to calcite and dol o mite Tran s. Cave Re s. Group G B. Vol 14 No. 2, 104 108 pp MARKOWICZ M. & PULINA M. 1979. Ilosciowa polmikroanaliza chemiczna wod wobszarach krasu weglanowego Katowice Poland 67 PAJON J M & VALDES J.J 1991. Simulacion qufmica de la disoluci6n de rocas carbonatadas del macizo karstico re Pan de Guajaib6n, Sierra del Rosario, Cuba Rev. LAPIAZ, Monograffa ill Valencia, Die. 1991 25-37 pp. PULINA M. FAGUNDO J R VALDES J R. RODRIGUEZ J ., KOZIK A. LESZKIEWICZ J GLOWACKI P ., PAJON J.M., DE LA CRUZ A. GARCIA A. 1984. The dy n amic of the contemporary karstic processes in the tropical area of Cuba. Prelimin a ry report of the field investigations performed by the Expedition Guajaib6n'84 in the winter s ea s on 1984 Univ Sla s ki Sosnowiec, 40 pp RODRIGUEZ J E ., CUTIE F ., CRUZ C ., FRANCO E ., FAGUNDO J.R. 1989 Hidrologfa carsica de l macizo Pan re Guajaib6n Si e rra del Ro sa rio Aiio hidrol6gico Nov 1984-Oct. 1985. De A ca demia La Habana 1989 61 pp. VALDES J.J ., FAGUNDO J R. PAJON J M I 981. Aplicaci6n de metodos num e r i cos al estudio hidrogeoqufmico de la c uenc a del rfo Cuyaguatej e Re v. Ing Hid Vol. II, No. 3 275 286 pp 98 Proceed i ngs of the 12 '" Internat i onal Congress of Speleology 1 99 7, S witz e rl and Vo lu me 2

PAGE 109

L'equation (4) peut etre rendue lineaire par une simple operation algebrique : Ln(C(I)) = Ln(A)a* Ln(Q(I)) Les pararnetres A et ex peuvent etre determines au moyen d'un graphique ln(C c 1 > ) ln(Q c l)) Ln(A) est la valeur du logarithme nature! de Ja concentration pour un debit unitaire, et ex est la pente de la droite de regression. On peut calculer ex avec la forrnule suivante : Ln(C 1 )Ln(C2) a=--~~ Ln(Q2)Ln(Q,) Le coefficient ex represente globalernent I' action de ea u dans le systeme en fonction du debit (processus de dissolution et de dilution simultanement). Le coefficient ex est :51. Plus ex est proche de I, plus l effet de dilution est important. Une augmentation importante du debit provoque une forte diminution de la concentration Plus ex s eloigne de I, plus l'effet de dilution est faible Une augmentation du debit a alors comme consequence une faible diminution de la concentration. Pour ex=O Ja concentration est constante, done independante du debit (mais Ja r,nasse transportee par unite de temps varie en fonction du debit) Pour cx ) represente la somme de plusieurs causes : la variation reelle du parametre A w les variations eventuelles du pararnetre ex, les erreurs d'analyse chimique, etc Les erreurs d'analyse pouvant etre considerees comme aleatoires ii est possible de debarrasser la chronique de ces variations en calculant la moyenne mobile de (A\ 0 ) On obtient alors la chronique A\) En introduisant le parametre A\ > (moyenne mobile de A\l)) dans !'equation (4) on obtient la chronique du TAC en fonction de A"(I) et de Q ( l): 6 'h Conference on Limestone Hydrology and Fissured Media 93

PAGE 110

Geologie et geochimie des eaux dans le reseau du Grand Cor {Valais-Suisse) par Jerome Perrin, Institut de Mineralogie, Universite de Lausanne, Dorigny, CH-1015 Lausanne Resume Le reseau du Grand Car est une cavite de haute altitude (2700 m ) s'ouvrant sous la Dent de Moreles (Valais-Suisse) et developpant quelque 4500 metres de conduits dans le flanc inverse de la nappe de Moreles (domaine Helvetique). Sa profondeur de 600 metres autorise quelques observations geologiques interessantes; parallelement une etude de la geochimie des eaux a ete entreprise, e!le a permis d'apprehender !'evolution du chimisme dans la zone vadose de ce karst haut alpin. Abstract The Grand Car system is situated below the Dent de Moreles in the Valais Alps (Switzerland) at an alti t ude of2700 meters ; the 4500 meters of conduits developed in the inverted limb of the Moreles nappe (Helvetic domain) The 600 meters depth of this system allows some interesting geological observations ; also a study of the water geochemistry was made with the aim of better understanding the chemical evolution in the vadose zone of this high alpine karst. 1 NW SE mames berriasiennes calca i res urgoniens olithique et congl. du Roe Champion calcaires greseux nummuti t iques P129 calcaires reconstitues a e l ements dolomitiques cristallin reconstitue R e sea u du Grand Cor ( Fu ll yNS ) Coordonnees : 572 5 95/115 740 2670 m DEV: 4670 m D ENIV : 598 m. Coupe projetee N135" Echelle : '::.. o =,g 1 00 m C==1 =ecoulements de chenaux C::::) =ecoulements de fissures Fig ur e 1: Co u pe geologiq u e d u Resea u d u G ra11 d Co r p r ese 11 ta11 t la sit ua tio 11 d es p o in ts d'ec lt a 11 tillo 111ia ge 1. Preambule La majorite des etudes concernant l'hydrochimie du karst se cantonnent aux sources karstiques, elles foumissent des informations sur le comportement hydrodynamique et sur Jes particularites geochimiques des aquiferes karstiques Les etudes hydrochimiques portant sur la zone vadose du karst restent elonnarnment peu nombreuses alors que l'on sail que le chimisme des eaux karstiques s'acquiert principalement dans cette zone "superficielle" (MAIRE, 1990) L'etude presentee dans eel article a pour objectif d'apporter une modeste contribution a cette problematique meconnue Le choix du t errain s'est porte sur le reseau d u Grand Car pour deux raisons: ce gouffre est situe en milieu haut alpin, absence de couverture pedologique permet de s'affranchir de !'influence du sol sur l a chimie des eaux souterraines (AITEIA, 1992) ; Jes atteintes anthropiques sur le milieu sont minimisees D'autre part, ce reseau traverse des roches au chimisme fort different (PERRIN 1996) susceptibles d'influencer le chimisme des eaux (figure I). L'evolution temporelle du chimisme a ete deliberemen! laissee de cote devant l'impossibilite d'effectuer des prelevements serres 6 th Conference on Limestone Hydrology and Fissured Media 99

PAGE 111

2 Situation geographique et geologique Le reseau du Grand Cor est situe dans Jes Alpes valaisannes (Suisse occidentale) au-dessus de la petite ville de Martigny L unite tectonique dans l a que!le ii se developpe est l'Helvetique et plu s precisement la nappe de Moreles. Cette nappe est peut etre la plus "celebre" des Alpes elle a fail l'objet de nombreuses publications (voir par exemple BADOux, 1972 ; MAssoN et al., 1980 ; RAMsA y. 1981 ) Elle presente un flanc normal et un flanc inverse, la serie sedirnentaire qui La constitue va de l'Aalenien au Tertiaire Le reseau du Grand Cor se developpe uniquement dans le flanc inverse, sa profondeur (-598 metres) permet de traverser l'integralite de la serie sedirnentaire Cretace-Tertiaire ; le fond du gouffre arrive a proximite du plan de chevauchement de la nappe ( figure 1 ). La serie sedirnentaire traversee montre une grande variabilite de roches : calcaires, calcaires siliceux, marnes, conglomerats a elements calcaires dolomitiques, mameux, cristallins La Chx-de-Fonds Berne SUISSE Figure 2: Sit 11 atio1t geograp h iq u e d u resea 11 d u G r a nd Co r Ch i mie des eaux : methodologie Les sites de prelevements ont ete places dans la mesure du possible legerement en aval de chaque changement de lithologie de plus, en certains endroits des ecoulements lents de fissures ont ete preleves (figure 1 ) II faut remarquer que Jes sites d echantillonnage ne sont pas tous situes sur le meme drain ; en effet ii existe plusieurs ruisseaux dans le gouffre issus de conduits impenetrables a l'homme et se perdant en aval dans des fissures elles-memes trop exigues II s agit de distinguer trois drains differents : Le premier regroupe Jes points d'echantillonnage perte-GC11-GCl2-GC l 3-GCl4GCl6-GCl8 auxquels s ajoutent Jes ecoulements de fissures GCl5 et GCl 7 Le deux.ieme drain est represente par le site GCl 11, ce drain avail ete considere comme un ecoulement lent de fissure et c est pour cela qu il a ete preleve, cependant les analyses chimiques ont montre qu il s agit en fail d un ecoulement tres rapide Le troisieme drain est constitue par Jes points d echantillonnage GCI 12-GC113-GCI 14, alors que le site GCl15 est soit le prolongement aval du premier drain soil un drain encore different. Un seul ecoulement de fissure a ete observe dans la zone profonde, c'est le site GCI 16. Les eaux du reseau resurgent a la source de la Sarvaz au niveau de la va!lee du Rhone Le chimisme de cette source est tres particulier mais ne sera pas aborde ici ( voir PERRJN & T ACCH! N l 1997) Les eaux prelevees ont ete conditionnees dans des bouteilles neuves en polyethylene de 250 ml rincees prealablement deux fois avec l'eau a analyser. Deux bouteilles ont ete remplies pour chaque echantillon, une bouteille etant acidifiee par 1 ml d HN"O3 65% suprapur afin de garder les cations en solutions Toutes les eaux ont ete filtrees a !'aide d'un filtre Nalgene 40 permettant I elimination des matieres en suspension, a exception des colloides Les boutei!les ont ete stockees moins de 24 heures apres le prelevement dans un endroit frais (4C) Les elements majeurs ont ete analyses par chromatographie ionique (Dionex 300) et Jes elements traces ont ete doses par ICP-MS couple AES (TJA-Poems). 4. Les elements majeurs Les eaux analysees sont de type bicarbonate-calcique, cependant quelques echantillons montrent une tendance a etre sulfates, c'est le cas par exemple pour le point de prelevement GC 19 Les trois e l ements montrant le comportement le plus significatif sont discutes dans la suite, il s agit du calcium, du magnesium et des sulfates. Differents facteurs sont a la source des variations chimiques illustrees a la figure 3, nous allons Jes passer en revue: D istance d'eco ul ement Les concentrations augmentent clairement avec la profondeur d echantillonnage dans le gouffre Cette evolution est bien entendu correlee a la distance d ecoulement: les eaux Jes plus profondes sont celles qui ont parcouru le plus long cheminement au sein de la zone vadose. L'evolution des concentrations est lineaire en premiere approximation, a exception du calcium qui semble suivre une evolution legerement parabolique. Dans tous les cas, toutes Jes eaux analysees sont sous-saturees vis a vis des mineraux qui pourraient precipiter dans un tel milieu (calcite, gypse). En realite on trouve de tels speleothemes actifs dans la cavite, mais l eau sursaturee responsable de ces precipitations n a pas pu etre echantillonnee Le gypse precipite en fines pe!licules microcristallines vraisemblablement a partir d eau de condensation Les concretions calcitiques (fistuleuses et excentriques) sont situees au droit de fissures peu permeables et siege d'ecoulements tres lents et satures en calcite Ecouleme n ts de fiss u res-Ecoulements de chenaux Les ecoulements de chenaux montrent des concentrations en elements majeurs nettement plus basses que Jes ecoulements de fissures. Les eaux circulant par Jes fissures moins permeables ont plus de temps pour se charger en differents ions Les eaux du site GCI 9, issues d'une fissure, illustrent remarquablement cette difference : l'ecart des concentrations entre GC 19 et le site GC18 (ecoulement de chenaux situe a meme altitude) s eleve a 24% pour Ca, 56% pour Mg et 60% pour SO4 L'ecoulement fissural GC17 montre aussi des concentrations sensiblement plus elevees que Jes ecoulements de chenaux mais en proportions moindres L ecoulement de fissure GCJ 16, situe dans la zone profonde du gouffre n a malheureusement pas pu etre echantillonne, mais la conductivite mesuree etait de 406 Siem alors que l'ecoulement de chenaux le plus proche montre une conductivite de 113 Siem L'ecoulement GC15 sort d une fissure centirnetrique, mais son chimisme est en tout point semblable aux ecoulements de chenaux : certaines fissures peuvent done etre le siege d ecoulements forts rapides ... Variations saison n ieres !ices a u x changements h ydrodynamiques Ces variations sont illustrees par Jes deux senes d echantillonnage : l'une effectuee en aoOt lors de la fonte des 1 00 Proceedings of the 12 th International Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 112

neiges, l autre en novembre alors que le massif est deja sous l emprise du gel. Les prelevements lors de la crue d'ete sont proportionnellement moins concentres l'eau de fonte dilue fortement les solutions II est probable que les vitesses d ecoulement plus rapide s en crue favorisent egalement la dilution ( dissolution moins efficace ) En etiage, les concentration s en calcium dans les ecoulements de chenaux se rapprochent tres nettement des concentrations dans les ecoulements de fissures Les sulfates montrent un ecart plus important que le magnesium cette difference doit tenir aux modalites de mise en solution de ces deux ions Anomalies Jiees au chimisme de la roche environnante Les site d'echantillonnage GC 19 (ecoulement de fissure) montre des concentrations tres elevees en Mg et SQ4 (respectivement 3 5 et 25 mg/I) De telles concentrations doivent provenir de la circulation de cette eau dans les gres siderolithique a pyrite, la pyrite constituant un apport de S04 non negligeable 4000 30 00 10 00 + + 0 0 o =echanUllonnage du 19 08 95 + =echanlillonnage du 02 11 95 + + + + + 0 0 0 00 -+------,,----r----.--.--,----,--,----, 4 00 3 00 i 2 00 C) ::;; 1 00 2000 00 2200 00 2400 00 Altitude [m] o =echantillonnage du 19 08 95 + =echanlillonnage du 02 11 95 + + + + 2600 00 + + ,;0 0 00 -+---,---,-----.--.----,--,--2000 00 2200 00 2400 00 2600 00 Altitude [m] 2800 00 2800.00 25 00 o =echantillonnage du 19 08 95 + =echantillonnage du 02 11 95 20 00 15 00 i + + .., 0 (/) + + 10.00 5 00 0 + 0 + 000-+------,----~---2000 00 2200 00 2400 00 2600 00 2800 00 Altitude [m] Figure 3: Evolutio11 de la co11ce11tratiou e11 eleme11ts majeurs (Ca, Mg, S04) e11 fo11ctio11 de l'altitwle et des co11ditio11S /,ydrody11amiq11es. 5. Les elements traces La figure 4 represente !'evolution concentration-altitude pour Jes trois elements traces Jes plus significatifs, a savoir Si Sr et Ba Les facteurs influen1rant le chimisme des eaux developpes dans le chapitre precedent restent valable pour Jes elements traces Quelques points meritent toutefois d etre soulignes pour chacun de ces elements : La silice La silice augmente lineairement avec la profondeur dans le gouffre Les eaux d etiage sont legerement plus concentrees que les eaux de crue Les ecoulements de fissures se marquent par une legere anomalie positive La silice est presente dans toutes les roches traversees generalement sous forme de quartz authigene ou detritique, les teneurs varient sensiblement d une formation a I autre Le strontium Les concentrations en strontium augmentent avec la profondeur selon une courbe vaguement parabolique. L ecoulement de fissure GCl9 montre une anomalie positive tres nette ; les echantillons d etiage sont faiblement plus concentres Le comportement de cet element se rapproche significativement du magnesium Le baryum Le baryum voit ses concentrations augmenter lineairement avec la profondeur les echantillons d etiage sont faiblement plus concentres Par contre, les ecoulements de fissures ne sont que tres legerement enrichis Le baryum et le strontium sont present en traces dans la calcite, ii est d autre part possible que le strontium provienne en partie de la strontianite ( cependant aucune occurrence de ce mineral n a pu etre observee sur le terrain) Ces deux elements peuvent dans tous Jes cas etre consideres comme de bons traceurs des aquiferes carbonates 6 '" Conference on Limestone Hydrolog y and Fissured Med i a 101

PAGE 113

r iii ::0 0.. .9, ., co ::0 a. .9, ui 1 20 o =echantillonnage du 19 08 95 + =echantillonoage du 02 11.95 + 0 0 80 + + + + 0 + $ 040 + + 0 00 2000 00 2200 00 2400 00 Altitude [m] 2 50 o =echantilloonage du 1!! 08 95 + =echantillonnage du 02 11 95 2 00 + + 0 0 150 0 0 1 00 & 0 0 50 0 00 200000 2200 00 2400 00 Altitude [m] 160 00 o =l!chantillonnage du 19 08 95 + + =echantillonnage du 02.11 95 + 120 00 + 80 00 i. 0 0 0 0 40 00 0 + + + i 0 + 2600 00 + + + 0 + 2600.00 + + + 0 0 o 0 0 2800 00 0 + 2800.00 0 0 00 --~--~-~-~-~-~--~-~ 2000 00 2200 00 2400.00 Altitude [m] 2600 00 2800 00 Figure 4: Evobltio11 de la co11ce11traJio11 e11 eleme11ts traces (Si, Ba, Sr) e11 fo11ctio11 de l'altitude et des co11ditio11s hydrody11amiq1tes. 6. Conclusions-Perspectives Les differents ions discutes plus haut voient leur concentration augmenter avec la distance de ruissellement dans la zone vadose, cependant aucun ion ne peut etre considere comme un traceur sOr de la fonnation geologique traversee En effet, les ions presentant une evolution (Ca, Mg Sr, Ba, Si, HCO3 et SO4) se retrouvent, sous forme minerale, dans !'ensemble des roches traversees. La seule exception est l'ecoulement de fissure GCl 9 : cette eau a ete en contact avec les gres siderolithiques a pyrite et montre de fortes concentrations en SO4, Mg, Sr, (Si) On pouvait s'attendre a un marquage des formations telles que le Berriasien marneux par Na-K-Al-Si l'Hauterivien siliceux par Si, les niveaux dolomitiques par Mg, or ii n en est rien Cette absence de marquage est liee au type d ecoulement de cet aquifere karstique : les ecoulements sont trap rapides et empechent toute dissolution de mineraux dont la cinetique de mise en solution est Jente; seuls les ecoulements vraiment lents (tel que GCl 9) echappent a cette regle et montrent des anomalies en certains ions La majorite des elements doses dans les eaux du reseau du Grand Cor sont bien en dessous des concentrations medianes calculees par DEMATTEIS (1995) a partir d'une centaine de sources karstiques; toutefois les sulfates font exception (mediane = 9,7 mg/1) et le strontium s'approche de la mediane (160 ppb) pour l ecoulement de fissure GC 19 Malgre ces faibles concentrations, !'evolution d elements tels que Ca, Mg, HCO3, Sr, Ba, SO4, Si est tout a fait significative dans la zone vadose d'un tel karst. Les autre elements doses (par exemple K, Na, NO 3 Mo, B V) sont trap peu concentres pour apporter des infonnations utiles Un suivi temporel detaille des ions interessants permettrait d'obtenir des donnees originales sur le fonctionnement hydrodynarnique et hydrochimique de la zone vadose du karst. Remerciements J'ai pu beneficier du soutien de plusieurs speleologues tout au long de ce projet, en particulier: B Ducluzaux, explorateur de la premiere heure dans le reseau, D Christen et P Tacchini lors des penibles campagnes d echantillonnage L'Institut de Mineralogie et le Centre d analyses minerales de l'Universite de Lausanne ainsi que l'Institut Fore! rattache a l'Universite de Geneve m'ont accueilli chaleureusement dans leurs laboratoires, qu'ils en soient vivement remercies. References ATIEIA 0. 1992 Role du sol dans le transfert des elements traces en solution Application a quelques systemes d'altitude These 1031, EPF-Lausanne BADOUX H. 1972. Tectonique de la nappe de Moreles entre Rhone et Lizeme Mat. carte geol. Suisse NS 143, 78 p DEMATTEIS A 1995 Typologie des eaux des aquiferes carbonates des chames alpines d'Europe centrale et meridionale These EPF-Lausanne, 226 p MAIRE R. 1990 La haute montagne calcaire Karstologia memoires n, 731 p MAssoN H., HERB R & STECK A 1980 Helvetic Alps of western Switzerland In: Geology of Switzerland, part B Wepf, Basel : 109-153. PERRIN J 1996. Etude geologique et hydrogeologique de la region Dent de Moreles-Grand Chateau Diplome Instituts de Geologie et Mineralogie, Univ. Lausanne, 125 p Non publie PERRIN J & TACCHINI P 1997 Le systeme karstique grotte du Poteu-source de la Sarvaz (Valais-Suisse) Ces actes RAMSAY J G 1981. Tectonics of the Helvetic nappes In : Thrust and Nappe Tectonics The Geological Society of London 102 Proceedings of the 12 th International Congress of Speleology, 1997, Switzer l and Volume 2

PAGE 114

Hydrogeochemical characterization of Tito Bustillo cave (northern Spain) S Sanchez-Moral 1 J. C. Caii a veras 1 E. Sanz-Ru b io 1 M . Hoyos 1 and V So l er 2 1 Dept. Geologfa. Museo Nacional Ciencias Naturales. Jose Gutierrez Abascal 2. E-28006 Madrid (Spain ) 2 Instituto de Productos Naturales y Agrobiologfa de Canarias. C.S.I.C. E-38206 La Laguna, Tenerife Spain. Abstract The geochemical characteristics of karstic water in Tito Bustillo Cave are the result of two main types of recharge : authigenic recharge via percolating and infiltrating water throuhg conduits and fractures and allogenic authigenic recharge via San Miguel river that sinks (Gorgocera ponor ) into the karst system The chemical composition of San Miguel river reflects that its drainage basin is located on siliceous (shales and sandstones) and calcareo u s lithologies ; likewise, the influence of tides is also pesent. The chemistry of infiltrational waters specially Pco2, show a direct relation with climatic conditions (rainfal regime) and soil development. Hydrogeochemical data, in conjunction with others data, such as rnicroclimatic conditions, are necessary to assess the deterioration induced by man action in show caves. Resumen Las caracteristicas geoquimicas de !as ag u as karsticas de la Cueva de Tito B ustillo reflejan dos tipos de recarga : recarga autigenica por la percolaci6n-infiltraci6n via conductos y fracturas y recarga alogenica-autigenica por medio del rio San Miguel que se sume en el sistema karstico en el paraje del surnidero de la Gorgocera La composici6n quimica del rio San Miguel muestra que su cuenca de drenaje atraviesa litologfas siliceas (pizarras y areniscas) y calcareas ; asimismo tambien denota la influencia del regimen de mareas La composici6n q u imica de las aguas infiltracionales, especialmente en los valores de Pco2 esta directamente relacionada con el regimen de lluvias y con el desarrollo de cobertera vegetal. Los datos hidrogeoquirnicos junto con otros datos como las codiciones microclimaticas de la cueva son necesarias para una apropiada valoraci6n del deterioro antr6pico en cuevas turisticas I. Introduction Multidisciplinary studies including hydroche mi s try and climatic characterization are nece s sary for the conservation of valuable geological biological and cultural ( archaeological ) elements in caves and karst This paper focuses on the study of the geochemical characterization on karstic waters in Tito Bustillo Cave ( northern Spain ) which hosts one of the most important paleolithic rock-paintings of western Europe This is part of a multidisciplinary study focused on making the suitable conservation of rock paintings comp a tible with Tito Bustillo Cave tourism indu s try Tito Bustillo Cave is located in Ribadesella ( Asturias northern Spain ), in the western riverside of Sella Ria ( Cantabrian Sea ). This cave belongs to the Ardines karst which is a relict polygenic karst system initially developed during the Lower Pliocene o n a carboniferous calcareous relief ( Ardines Massif) This massif is formed mainly by a thick sucession of gray to beige-coloured limestones that s hows a homogeneous mineralogical composition ( 93-100 % calcite ; 0-7 % clays and quartz ). The San Miguel River sink into the Ardines karst ( Gorgocera ponor ) and intermitently flow ( aproximately 600m ) throught Tito Bustillo Ca v e 's passages, towards its mouth in the Sella R ia ( fig I ). Th e Gorgocera ponor is located 30 m high above sea level. II S urface W a t er Two main types of recharge can be distinguished in the Ardines karst system : disperse authigenic recharge vi a percolating and infiltrating water through dolines and k a rren fields from rainfall ; and authigenic-allogenic recharge via San Miguel river drainage basin The San Miguel d r ainage basin is composed of several smaller sub-basins ( for instance, the Sardedo and San Juan sub-basins which are mainly composed of shales and s a ndstones ) and show an 2 elongated shape with a total surface of 9 7 km The slope s are pronounced, favouring high runoff and low infiltration rates in the upper drainage basin The drainage density i s low specially in the zones with calcareous lithology The c h emical composition of waters in Tito Bustillo Cave the lower zone of the drainage basin is very influenced by the fact that the San Miguel River pass through zones with different lithology as is expressed below 6 '" Conference on Limestone Hydro l og y and Fissured Med i a 1 03

PAGE 115

1V A .A Gours --~ San Miguel River stream-sink Dripping water ~Q La Gorgocera o~-~-50m Figure 1 Location of sampling points in Tito Bustillo Cave The composition of surface waters related to San Miguel river was achieved by chemical analyses (major elements) of a total of 18 samples corresponding to stream-waters and underground springs (Martfnez, 1990 and this work). The chemical composition of these waters reflect the lithological characteristics of the zone where they flow. A) Samples from springs in calcareous zones are of HCO3-Ca type These water are close to calcite saturation and are similar to those of Gorgocera ponor, but with higher Ca and HCO3 contents. B) Samples from the upper drainage basin (San Miguel River, San Juan River and San Juan Creek; siliceous lithology) are also of the HCO3-Ca type, but with higher SO4 and Cl contents and low values of total dissolved solids. The chemical composition of waters from San Miguel river before their entrance in the interior of Tito Bustillo Cave reflects an effect of mixing between authigenic underground discharge and allogenic recharge from the upper drainage basin In order to test this phenomenon a theoretical nuxrng was performed using average compositions of waters of type A and B. The result is shown in figure 2; the proportions used in the theoretical mixing are 60% and 40% respectively. III. Groundwater A total of 29 samples of water corresponding to gours, dripping-points and San Miguel river during three differents sampling campaigns (July 1995; Sepetember-October 1995; May 1996) were collected (see fig 1 for sampling I ocati ons and table I for dates of sampling) The last sampling campaign was carried out after a high rainfall period (70 2 1/m ) and the infiltration rates were perceptibly higher than in the previous sampling campaings. The temperature, = electrical conductivity, pH and CO2 HCO3contents were measured at the time the sample was taken. These samples were stored at a constant temperature (12-14ooC) and taken to the laboratory where Ca, Mg, Na, K, SO4, Cl NO3 were analyzed two days later. The saturation states and Pco2 values of waters were calculated from the temperature and the chemical composition of the water sampling. The speciation calculation necessary for determining the mineral saturation indices (table n was made using the PHRQPTIZ computer code (Pl ummer and others 1988). .San Miguel river (in the interior of the cave). The samples SUM-BI reflect a strong marine influence due to its proximity to the Sella Ria with high values of Cl, Na and total dissoved solids, higher than the rest of the samples collected in the interior of the cave Concerning saturation states, these waters are very near to equilibrium respect to the main carbonatic mineral phases. The two samples termed SUM-B2 are waters of HCO3-Cl-Ca type; in low tide periods (september 1995) a decrease in the Cl-content can be observed Both samples was undersaturated with respect to calcite and aragonite, which are the most common carbonatic mineral phases in the host rock and in the speleothems of the cave . Gours These waters are of HCO3-Ca type, with low concentration in total dissolved species. The samples with the lowest Pco2 values are located in zones where air renovation rates are higher (in the proximity of gorgocera ponor, where an entry of the cave is located) and subsequent outgassing effect is also higher . Dripping waters These waters are of HCO3-Ca type, with SO4 and Mg contents ranging from I to 5% ; Cl and Na contents 104 Proceedings of the 12 th Internat i onal Congress of Speleology 1997 Sw itze rland Volume 2

PAGE 116

'aQJ E M 0 u I 6 0 5 0 4 0 3 0 2 0 SPRINGS (calcareous zone) ;\ 1 0 I SAN MIGUEL RIVER Average composition nearby to the cave THEORETICAL MIXING SAN MIGUEL RIVER TITO BUSTILLO CAVE SAN MIGUEL RIVER upper drainage basin (siliceous zone) 1.5 2 0 2S04 + Cl (meq/1) Figure 2 HCO 3 vs (SO/ +CJ) plot diagram sho w ing chemic al c omposition of San M iguel Ri ver ranging from I to 12% and from 4 to 11 % respectively These relatively high values in Cl and Na contents in the infiltration waters reflect the input of sea salts near the coast by means of aerosol effect ( Drever, 1982) Samples corresponding to zones with active speleothem formation (calcite and aragonite in mineralogy; samples GB-3 ; GB-4 and GBI) are oversaturated in main carbonatic mineral phases. Samples GB-2, GB-5 y GB-8 are undersaturated in these minerals showing low water-rock interaction rates reflecting low residence times and high infiltration rates of the waters. Samples GB-9 GB-10 and GB -I I show an anomalous chemical composition respect to the rest of the sampled waters with a concentration of SO4 ranging from 13 to 31 %. This sulphate-rich character coincides with the existence of gypsum crystal growth affecting (and deteriorating) the pigments of paleolithic rock-paintings Concerning Pco2 values these samples show a great variability. The higher Pcoz values correspond to samples collected in May 1996, that is, after a high rainfall period as previously mentioned. The CO2 produced in the soil by biological activity in the uppermost zone of the Ardines Massif is dissolved by rainwater which filters into it, this being the vehicle for transporting the CO2 to the interior of the karst; when the infiltration waters reaches the cave, partial outgassing of this is produced, COz being released into the cave atmosphere The lowest Pcoz values correspond to summer season with a slight increase in autumn due to the begin of a high rainfall period. IV. Conclusions The chemical composition of the waters of San Miguel river before its sinking into Tito Bustillo cave reflect an effect of mixing between authigenic underground discharge and allogenic recharge from upper drainage basin. In the interior of the cave this chemical composition is affected by the composition of infiltrational waters and the effect of tides dynamic. This multiple influence in the karstic water must be considered in order to assess the deterioration processes induced by man actions ( polution of karstic waters, mass-tourism, etc .). Infiltrational waters show a direct relation to rainfall regime and soil development in the surface of the Ardines karst system. In this way, the highe st Pco2 values correspond to spring seasons as a response of higher infiltration rates. Although most of these waters are of HCO3-Ca type, we have detected Ca-SO4 enriched dripping-waters affecting the conservation of rock paintings Table I Saturation indices of main carbonate mineral phases and Pco 1 values of the ana/yzed waters (R: river : G: gour ; D : dripping water) Sample Date Type Aragonite Calcite Dolomite Gypsum PC02 SUM-B 1 11 / 07 / 95 R SUM-B 1 29 / 09 / 95 SUM-B2 11 / 07 / 95 R SUM-B 2 29 / 09 / 95 E-B 1 22 / 05 / 96 G E-B 2 29 / 09 / 95 G E-B 3 29 / 09 / 95 G E-B 12 22 / 05 / 96 G G-B 1 11 / 07 / 95 0 G-8 1 30 / 09 / 95 G-B 1 22 / 05 / 96 G-B2 11 / 07 / 95 0 G-B 2 22 / 05 / 96 G-B3 11 / 07 / 95 0 G-8 3 29 / 09 / 95 G-B 3 30 / 09 / 95 G-B 3 1 / 10 / 95 G-B 3 22 / 05 / 96 G-B4 11 / 07 / 95 0 G-B 4 30 / 09 / 95 G-B 4 22 / 05 / 96 G-B5 11 / 07 / 95 0 G-B 6 30 / 09 / 95 0 G-B 6 22 / 05 / 96 G-B 7 30 / 09 / 95 0 G-B8 1 / 10 / 95 0 G-B 9 22 / 05 / 96 0 G-B 1 o 22 / 05 / 96 0 G-B-11 22 / 05 / 96 D +0 09 -0 29 -0 32 -0 33 -0 14 -0 37 0 04 -0 42 +0 20 +0 09 -0 20 -0.46 -0 64 +0 79 +0 68 +0.52 +0 55 +0 11 +0 29 +0 19 +0 16 -0 35 -0 21 -0 47 0 18 -0 68 -0.48 -0 39 -0.40 +0 28 -0 09 -0 13 -0 .1 3 +0 05 -0 18 +0 15 -0 23 +0.40 +0 28 o -0 27 -0.44 +0 99 +0 88 +0 71 +0 74 +0 31 +0.49 +0 38 +0 36 -0.15 -0 02 -0.27 o -0 49 -0 28 -0 20 -0 21 +0.74 -0 09 -0 92 -0 94 -1.08 -1 51 -0 85 1 18 -0.48 -0 69 -1 30 -1.53 -1 94 +1 06 +0 92 +0 59 +0 62 -0 26 -0 09 -0 34 -0 56 -1 27 -1.22 -179 -1 03 -1 94 -1 37 -1.18 -1 02 -1.43 -1 53 -2.23 2 12 -2 23 -2.38 -2 71 -2 73 -2.41 -2 41 -2 35 -2.96 -2 86 -1 97 -1 95 -1 99 -1 97 -2 01 -2.20 2 19 -2 11 -2 72 3 97 2 66 3.41 -2 89 -2 02 -1 93 -1 39 10 .6 9 10 40 10 17 10 18 1 o 39 1 o 31 1 o 86 10 69 10 91 10 63 1 o 58 10 .2 91 1 o 67 10 19 10 87 1 o.73 10 10 .3 8 10 95 10 10 41 1 o 18 10 85 10 2 59 10 03 1 o 73 1 o 84 10 88 10 2 93 6 1 h Conference on Limestone Hydrology and Fissured Media 1 05

PAGE 117

References Drever, J L. 1982 The geochemistry of Natural Waters : 3 88 pp Prentice Hall Englewood Cliffs New Jersey Hoyos, M 1979 El karst en Asturias durante el Pleistoceno superior y Holoceno Unpublished Ph Thesis. 413pp Martfnez E 1990 Prevenci6n de inundaciones en la Cueva de Tito Bustillo Inforrne para la Consejeria de Cultura del Principado de Asturias 118 pp Plummer, L.N ; Parkhurst, D L. ; Fleming G W & Dunkle, S A 1988. PHRQPITZ a computer program incorporating Pitzer's equations for calculation of geochemical reactions in brines US Geol Surv Water Res Inv 88-4153 : 310 pp Acknowledgements This work has been supported by the Council for Education Culture, Sports and Youth of the Principality of Asturias by means of a Scientific Cooperation Agreement with the C.S.I C The authors wish to express their gratitude to Prof J Fortea, (Univ Oviedo ) for his collaboration throughout the whole period of the investigation Hydrogeochemical analyses and mineralogical deter minations were conducted with the assistance of M A Vallejo, M.I. Ruiz and R. Gonzalez ( MNCN-CSIC) 106 Proceed i ngs of the 1 2 th I n ternat i onal Congress of Speleolog y, 1997 Switzer l and Vo l ume 2

PAGE 118

Karst aquifer genesis Modelling approaches and controlling parameters Martin Sauter, Rudolf Liedl, Torsten Clemens, Dirk Hiickinghaus Applied Geology, University of Tiibingen, Sigwartstrasse 10, 72076 Tiibingen, Germany Abstract The quantification of regional groundwater flow and the transport of dissolved substances in a karst system generally poses problems in parameter identification as well as in the modelling of the above processes. On the other hand, available qualitative and quantitative geological information on the history of the development of karst aquifers which can provide information on the distribution of karstified horizons within the modelled domain are generally not considered. A model has been developed, capable of simulating the interactions between the different processes i.e. flow, transport coupled with carbonate dissolution, taking into account the varying boundary conditions, e.g. changes in geological and geomorphological (base level) and climatic conditions (recharge depth temperature) A sensitivity analysis of the duration of karstification to changes in climatic parameters and physico-chemical constants of carbonate dissolution shows that the equilibrium concentration of calcium and the initial diameter of the conduits are very dominant factors. Zusammenfassung Die Quantifizierung von Grundwasserstromung und Stofftransport in Karstsystemen ist gekennzeichnet durch Probleme in der Erkundung und durch einen Mangel an angepafiten Modellwerkzeugen. Dagegen wird vielfach verfugbare geologische geomor phologische, palaohydrologische Information iiber die Entstehung dieser Systeme nicht verwendet, obwohl sie Aufschlufi iiber Parameterverteilung und Aquifergeometrie zur Verfugung stellen konnte. In diesem Artikel wird ein numerisches Modell vorgestellt das die Entstehung und zeitliche Entwicklung von Karstsystemen unter Beriicksichtigung der komplexen Stromungsund Karbonat losungsprozesse fur zeitlich variable Randbedingungen (Vorfluterniveau Grundwasserneubildung, Temperatur) simuliert. Mittels Sensitivitatsanalyse konnte gezeigt werden dafi die Gleichgewichtskonzentration von Kalzium und der initiale Durchmesser der Karstrohren zu den fur die Verkarstungsdauer wichtigsten Grofien gehoren 1 Introduction The understanding of the genesis of karst aquifer systems has been the objective of research for some time For a long period this field of research had been approached from a more descrip tive point of view. With the understanding and quantification of the carbonate dissolution kinetics (PLUMMER AND WIGLEY' 1976 ; RAUCH AND WHITE, 1977; PLUMMER et al, 1978) it was possible to investigate the main factors involved in karst develop ment (PALMER, 1981; 1991: DREYBRODT, 1990 ; 1996) The application of these principles however was up to now only applied to simple boundary conditions (JAMES AND KlRK PATRICK, 1980; GROVES & HOWARD 1994; HOWARD & GROVES, 1995 DREYBRODT, 1996) e.g. near hydraulic struc tures The understanding of the karstification processes at a regio nal, i e catchment scale could be of most valuable importance for the quantification of the geometry of karst aquifers as well as their parameter distribution in space. This information would be most useful for the prognosis of flow and transport in karst systems. Generally, geological, geomorphological and paleo hydrological information is available and could be used for the specification of boundary conditions in space and time. In this paper, a new modelling approach is presented integra ting the chemical dissolution kinetics of carbonate rocks as well as the special hydraulic characteristics of karst systems thereby allowing the simulation of the development of karst systems at a catchment scale With this model, the sensitivity of the develop ment of karst aquifers to a change in various parameters is de monstrated 2 Modelling Approach and Processes Figure 1 summarises the most important processes for karst aquifer genesis. Groundwater flow is represented by a dualistic flow system, i.e. a highly conductive, low storage (0.0001) conduit system and a much less permeable but high storage (0 02) fissured sys tem The conduit system is characterised by high flow velocities ( up to several 100 m/h) Depending on the flow velocities laminar or turbulent flow conditions can occur. Flow between the two systems is controlled by the difference in potential. spring direct recharge conduit (flow, calcite dissolution, Ca 2 -transport) Figure 1: Flow system and processes 6 "' Conference on Limestone Hydrology and Fissured Media 107

PAGE 119

There are cwo different modes of infiltration of groundwater recharge: a) direcc recharge, which reaches the phreatic zone within a few days via highly conductive vertical shafts and do lines, b) indirect recharge arriving at the water table con siderab ly delayed. It is important to distinguish between these two types of recharge components, because of their different quantities and degree of undersaruration with respect to calcite. Dissolution of carbonate occurs along che conduits and conse quently their diameters are en lar ged with time This increase in diameter leads 10 an increase in hydraulic conductivity of the conduit network and to an enhanced calcite di ss olution i e a positive feedback mechanism between flow and dissolution pro cesses. Further, it has to be taken into account that some of the parameters important for the karstification process and conse quently on the karstification period, such as rate constants e quilibrium constants viscosity also depend on temperacure The above described processes were implemented in a nu merical model (CA VE, ~arbonate ~quifer Yoid gvolution) Flow in the fissured syscem is modelled using a continuum ap proach (BERKOWITZ 1988) whereas flow in the conduit necwork is discribed by flow in a rube network, which follows Kirch hoff' s rule saying that the sum of inflows and outflows equal 10 zero at every node (HORLACHER & LUDECKE 1992). The ex change of water between the continuum and the tube network systems is controlled by their potential difference and an ex change coefficient, determined by the hydraulic conductivity of the fissured system and geometric factors Transport of Ca 2 + -Ions is modelled using a one-dimensional advection equation expanded by a source term allowing for the increase in ionic concentrations a l ong the flow path due to car bonate dissolution Further source terms describe the input of calcium-ions from recharge. The quantification of the carbonate dissolution is based on experimental results from BUHMANN & DREYBRODT (1985) According to the authors, different reaction rates could be deter mined depending on the concentration in solucion Rapid dis solution prevails for concentrations far from (first order rate law) and slow dissolution close to (fo urth order rate law) calcite equilibrium concentration Details of the model its verification and sample runs are presented in CLEMENS et al. (1996), CLEMENS et al. (1997) and HOCKINGHAUS et al. (1997) 3. Controlling factors in karst development The degree of karstification and the variety of different karst features are determined by the interaction between differem processes which are controlled by "intrinsic" and "extrinsic" factors. "Extrinsic" factors comprise factors such as climate (precipitation, evapotranspiration, temperature ), level of dis charge (absolute height and relief, level of ea water lakes and rivers), type of vegetation (humid tropical, arid etc.) and espe cially the time period available for karstification The term "in trinsic" factors summarises the lithological and petrographic characteristics and also the type and geometry of structural elements. The dominant determining factor for carbonate dissolution the erosional and hydraulic processes is the climate and know ledge of its temporal variation is an important prerequisite for reliable predictions of the intensity of karstification and the spatial distribution of the karstified zones. Figure 2 (changed after FORD & WILLIAMS, 1989) shows the importance of the influence of the climate on karstification and illustrates the dif ferent feed-back mechanisms inherent in such a complex system Climate Temperature, Precipitation, Evapotranspiration Soil Vegelation a, u. Hydrology, Geomorphology Relief Discharge Level Regol~h a, u. Aquifer Geome~ Petrol ., Lilhol. Distribut~;,J:~one Fm Groundwater Recharge Depth Distribution HydrochemicaJ Processe Solution Precipilalion Kinetics Distribution of 'Pores' Level of Karstification Hydraulic Conductivity Porosity Groundwater Row Flow Velocity, Hydraulic Potential Figure 2: Influence of the climate on karst development and important feedback mechanisms (after FORD & WILLIAMS, 1989) 4. Sensitivity analysis for the duration of karsti fication By varying the magnitude of the relevant process parameter one at the time the sensitivity of the duration of the karst de velopment was examined The model used was described by CLEMENS et al (1996). The flow model is rectangular and mea sures 1350 1250 m 2 There are no flow boundaries on 3 sides and a fixed head boundary on the fourth side representing e.g a river. The conduit system is discretised as a cen~ral series of tubes connected to the fixed head boundary with initial dia meters of 0.4 mm. Groundwater recharge was 400 mm/a with 1 % as direct recharge. The input concentration of the direct recharge was O mol/L. This data set forms the "standard" of Table 1 and is taken from laboratory experiments (e.g. BUH MANN AND DREYBRODT, 1985) and field investigations in S Germany The parameters are varied within a range that is be lieved to occur in nature (Table 1) The karstification periods specified in the following represent the times when turbulent flow conditions prevail throughout the tube network Here only the main results are summarised (Fig. 3). In this analysis, the Ca 2 + -equilibrium concenrracion and the initial tube diameter have been identified as the controlling fac tors for karstification A reduction in the equilibrium concen tration by 50% leads to an increase in the duration of karsti fication by 18646 years. Similarly a reduction in the initial dia meter by 50% leads to a prolongation of karstification by more than 10000 years 108 Proceedings of the 12 th International Congress of Speleology, 1997, Switzer l and Volume 2

PAGE 120

Table 1: Data set for sensitivity analysis Parameter Ca 2+ Equilibrium Concentration (mmol / L ) initial tube diameter ( mm ) Dissolution rate constant close to equilibrium ( cm 1 0 mo1 3 s 1 ) Dissolution rate constant far from equilibrium ( cm / s ) Groundwater recharge ( mm ) kinematic viscosity ( m 2 / s ) percentage of direct recharge (%) Minimum Standard 1.0 2.0 0 2 0.4 2 5 10 12 1.2 10 13 2 5* 10 6 2 0*10 5 160 400 1.8* 10 6 1.3*10 6 0 5 Maximum 3 0 1.0 5 0 10 13 1.5 10 4 1829 1.0*10 6 2 The sensitivity of karstification with respect to the kinetic rate constants and groundwater recharge is considerably less Within the range of parameters analysed, the s patial distribution of groundwater recharge between conduit and fissured system and the viscosity of the water play a negligible r o le for the duration of karstification 2 3 4 5 6 7 Figure 3: Results of se11sitivity analysis for parameters: (1) ca2+ -equilibrium concentration, (2) initial tube diameter, (3) rate constant for concentrations close to equilibrium, (4) groundwater recharge, (5) rate constant for concentrations far from equilibrium, (6) kinematic viscosity (7) distribution of groundwater recharge (grey: maximum, black: minimum value of respective parameter, cf. Table 1) However during the course of aquifer development, the per centages assumed for direct recharge might well change by an order of magnitude to 10% and more leading to much reduced duration Further the input concentrations of direct and indirect re charge ( cdir> cFK) and the concentration threshold for the change between first order dissolution and fourth order dissolution k s w were varied as well The results are summarised in Table 2 An increase in the input concentrations obviously leads to an exten sion of the karstification period whereas a reduction in the above threshold value k s w (standard condition k s w = 0 9 ) speeds up the increase in the tube diameter because undersaturated water can penetrate further into the tubes due to the reduced dissolution rate under conditions of fourth order (slow) kinetics Table 2: Effect of a change in cdirl cFK and k,,., on karstification Parameter Value Change in karstification period lnput c oncentration 1.4 mmol/L + 654 years direct recharge cdir Input concentration 1.88 mmol/L + 323 years indirect recharge cFK concentration 0 6 2804 years threshold k s w ( fraction of equilibrium concentration) The parameters cdir> cFK, and ksw however need further inves tigation in the field in order to determine their effect on the duration of the karstification 5. The role of the climate In the above sensitivity analysis, only a single parameter was varied at the time However, climatic changes evoke a multitude of factors to vary which are important for karstification such as temperature, evapotranspiration and precipitation. A change in these meteorological parameters brings about changes in the type of vegetation and therefore also in the carbon dioxide production in the soil. Changes in the carbon dioxide concentration and precipitation in turn change the rate of erosional processes which also affect discharge levels In Figure 4, the effect of a change in temperature on the duration of karstification is illustrated as an example how com plex the system can be affected by a change in a single master variable. In sum: With all other variables constant an increase in temperature by ten degrees reduces karstification time from approximately 50000 years to ea. 10000 years. With an increase in temperature the first order dissolution rate constant is in creased thereby prolonging the duration of karstification the viscosity of the water is decreased leading to shorter duration the partial pressure of carbon dioxide is increased with shorter durations and the equilibrium concentration of calcium is de creased with a consequential increase in the duration of karsti fication The dominant parameter, however is the increased production rate of carbon dioxide, which overall decreases the time required for the development of a mature karst system (Table 3 ). 6 th Conference on Limestone Hydrology and Assured Media 109

PAGE 121

Table 3: Effect of an increase in temperature on variables, controlling duration of karstiftcation Parameter Viscosity of water Dissolution rate constant far from equilibrium Partial pressure of carbon dioxide in the soil Equilibrium concentration c 60000 ., E a. 50000 .2 ., > ., 40000 "'O ;;; 30000 -""' ., -:S 20000 '0 C: 10000 9 ::, 0 "'O 0 Change in Parameter + + 5 Change in duration of karstification + + 10 15 temperature in C 20 Figure 4: Change in karstiftcation period as a result to a change in temperature Acknowledgements This study was supported financially by the Deutsche For schungsgemeinschaft (DFG) within the context of the Collabora tive Research Centre 275 (SFB 275) and by the European Union Contract EV5V-CT94-0471 References BERKOWITZ B., BEAR, J. & BRAESTER, C. 1988. Continuum models for contaminant transport in fractured porous formations Water Res Res. 24 1225-1236 BUHMANN D & DREYBRODT W 1985 The kinetics of calcite dissolution and precipitation in geologically relevant s ituations of karst areas 2 Closed systems Chem Geol. 53 109-124 CLEMENS, T H0CK.INGHAUS, D SAUTER M L!EDL R. & TEUTSCH G 1996 A combined continuum and discrete network reactive transport model for the simulation of karst development. LAHS Publ., 237 309-318 CLEMENS, T H0CK.JNGHAUS D., SAUTER M ., LIEDL R & TEUTSCH G 1997. Simulation of the development of karst aquifers by using a pipe flow model coupled to a continuum model : 2. Model verification and sensitivity analysis Water Res Res submitted DREYBRODT W 1990. The role of dissolution kinetics in the development of karst aquifers in limestone: a model simulation of karst evolution J. Geol., 98, 639-655 DREYBRODT W 1991, Principles of early development of karst conduits under natural and man-made conditions revealed by mathematical analysis of numerical models Water Res. Res 32 2923 2935 FORD D C & WILLlAMS P W 1989 Karst Geomorphology and Hydrology Unwin London 60lp GROVES, C G. & HOWARD, A.D. 1994 Early development of karst systems, 1. Preferential flow path enlargement under la minar flow Water Res Res ., 30, 2837-2846 HORLACHER H.-B. & LUDECKE, H.-J. 1992 Stromungs berechnung fur Rohrsysteme Expert Verlag 218p HOWARD A D & GROVES C.G. 1995. Early development of karst systems 2. Turbulent flow. Water Res. Res. 31 19-26 H0CKINGHAUS, D ., CLEMENS T ., LIEDL R SAUTER M & TEUTSCH G. 1997. Simulation of the development of karst aquifers by using a pipe flow model coupled to a continuum model: I. Theory. Water Res. Res., submitted JAMES, A.N & KlRKPATRICK, I.M 1980 Design of foun dations of dams containing soluble rocks and soils Quat. J. Eng. Geol ., 13 189 198 PLUMMER, L N. & WIGLEY, T.M.L. 1976 The dissolution of calcite in CO 2 -saturated solutions at 25 C and 1 atmosphere total pressure. Geochim Cosmochim. Acta 40, 191-202 PLUMMER, L.N., WIGLEY T.M L. & PARKHURST, D.L. 1978 The kinetics of calcite dissolution in CO 2 -water systems at 5 to 60'C and 0 0 to 1.0 atm CO 2 Am J. Sci, 278, 179-216. PALMER A N 1981. Hydrochemical controls in the origin of limestone caves: 8th Int. Spel. Congr. Bowling Green, Kentuck y 120-122 PALMER, A.N. 1991. Origin and morphology of lim estone caves. Geol. Soc. Am. Bull., 103 1-21. RAUCH H W. & WHITE, W.B. 1977. Dissolution kinetics of carbonate rocks 1. Effects of lithology on dissolution rate. Water Res Res 13 381-394. 11 O Proceedings of the 12 '" International Congress of Speleology 1997 Switzerland Volume 2

PAGE 122

Influence d'intercalations carbonatees en milieu sedimentaire a dominance siliceuse sur le chimisme des eaux (Finistere, France). par Luc Somlette, Jean-Pierre Faillat et Gael Le Bee Laboratoire d Hydrogeologie, ISAMOR, Technopole Brest Iroise 29280 PLOUZANE FRANCE Abstract In Brittany water generally has acid and agressive characteristics Nevertheless water can be locally neutral to basic and have a HCO 3 -Ca hydrofacies as a consequence of carbonate intercalations in siliceous and aluminosilicate sedimentary rocks. The difficulty is to observe those intercalations because of the thickness of the ground, of the vegetation, and of the weak extension of carbonate outcrops This study deals with a river basin near Plougastel-Daoulas (Finistere France) constituted with Devonian sediments of schistes, sandstones quartzites and carbonates. We focused our investigation on the hydrochemical characterization of major ions in order to identify and locate those carbonates The methodology cons is ted in measurin~ electrical conductivity and pH of all the brooks and sources, and in taking samples to analyse ions (Ca2 +, Mg 2 +, Na 'K +, er sot HCO3 NO3"). Three families of water can be deduced from the statistical analysis of chemical results. Firstly a Na Cl hydrofacies due to the influence of the proximity of the sea secondly an HCO 3 -Ca hydrofacies due to the presence o f carbonate layers and finally, an intermediate hydrofacies. The localisation of those three families generally agrees with the geological map Yet, this method precises and completes it. For example, some samples have bicarbonate calcium facies whereas there is no outcrop of carbonates We conclude that it is possible to easily identify and locate the presence of carbonate intercalations with this method using hydrochemical characterization of water in spite of the absence of outcrop or of access difficulties. Resume Difficilement observables a l'affleurement du fait du couvert pedologique et vegetal continu et de leur faible extension, les intercalations de bancs carbonates et de schistes du Devonien de l'extremite Ouest du Massif Armoricain influencent fortement le chimisme des eaux souterraines qui presentent alors un hydrofacies bicarbonate Cet hydrofacies les differencie nettement des eaux exclusivement au contact de roches siliceuses ou aluminosilicatees (gres, schistes) En consequence, il est ainsi possible de reperer aisement la presence de niveaux carbonates qu en bien meme leur presence etait insoup~onnee et d avoir des eaux peu agressives dans une region ou cette caracteristique est inbabituelle. 1Introduction et but En Bretagne les eaux superficielles et souterraines presentent generalement un caractere agressif et acide suite a leur contact avec des roches siliceuses ou aluminosilicatees. Mais localement, ces eaux ont un pH neutre a alcalin et ont un hydrofacies carbonate calcique. Cette particularite signe la presence d'intercalations de roches carbonatees aisement dissoutes par une eau agressive. L'existence d'un couvert pedologique et vegetal continu la faible extension de ces roches rendent difficiles leur reperage L'etude menee sur le bassin versant de Pennaster a permis de tester l'utilisation des ions en solution comme outil facilitant identification et la localisation d'intercalations carbonatees en environnement schisto-grescux au niveau d'un petit bassin versant. 2Situation geographique, geologique, et hydrogeologie Le bassin versant de Pennaster est situe au Sud Ouest de Plougastel-Daoulas Sa supcrficie est de 40 Km 2 (Fig.I). L activite est essentiellement maraichere (fraises, tomates ... ) il n'y a pas d industries et !'habitat est disperse Le climat est oceanique avec une pluviometrie et une temperature moyenne annuelles de 1100 a 1300 mm de 11 a 12C 11 est occupe par des formations sedimentaires s' etageant du Silurien Superieur au Devonien moyen (CHAURIS ET PLUSQUELLEC, 1980), constituees de schistes, gres et quartzites, avec de rares intercalations calcaires que l'on rencontre surtout dans les schistes et calcaires de l Armorique (Gedinnicn superieur S i egenien moyen (Fig. l). La fracturation est marquee par l existence de deux farnilles de fractures NW-SE et NE-SW. Les premieres observations concemant les phenomenes karstiques dont ces roches sont le siege ont ete fa i tes par PLUSQUELLEC (1982) Plus recemment, WAKEFIELD (1993) et AUBRY-DELAF0SSE (1995) ont realises les releves topographiques de deux cavites ayant des developpements d'une centaine de metres (Pennaneac'h-Rozegat) et d une soixantaine de metres (Pointe de l'Armorique). L orientation de ces galeries est principalement influencee par des 6 th Conference on Limestone Hydrology and Fissured Media 111

PAGE 123

fractures N40-50 et N80-90 Les concretions y sont assez rares (quelques planchers stalagmitiques, stalagtites draperies ... ) 11 en est de meme du modele karstique de surface, reduit a quelques dolines dans les environs de Runavel. Cependant, les chenaux sont bien representes en front de mer et dans les carrieres. On ne connait pas de sources karstiques dans cette reg i on. Le gres Armoricain le gres de Landevennec, et les quartzites gediniennes constituent des aquiferes de fissures. LEGENDE Schistes et calcaires de l'Armorique Faille Contour g6ologique Rbeau hydrographique de Pennaster 1, 2, .. 28B R6ferences des points 6chantillonnb ( P = Puits, S ,. Source) I HCo, J : 0-1 m6.r 1 ( 0-61 mg r ) RADE DE BREST e 1-3 m6 r 1 ( 61-183 mg r 1 1 e 3-5 m6S 1 ( 183-305 mg.1' 1 ) PLOUGASTEL DAOULAS 0 Fig.I : Localisation geographique et geologie du bassin versant de Pennaster. Position et concentration en HCO 1 des points echantillonnes 3Methodotogie 1km Les prelevements ont ete realises apres plusieurs jours sans pluie Les eaux de surface et les sources ont ete recueillies dans des flacons propres prealablement rinces avec l eau a analyser puis hermetiquement fermes en evitant la presence de bulles d air Des mesures d electroconductivite de pH et de temperature ont ete pratiquees sur le terrain La conservation des echantillons s'est faite a l obscurite et a une temperature de 4C. Des l'ouverture ont ete mesures l'electroconductivite, le pH, HCO 3 par neutralisation acide, Ca 2 + et Mg 2 + par complexometrie, puis NO 3 par spectrophotometric en U V so/ par turbidimetrie er par la methode de Mohr et Na + et K + par spectrometric d absorption atomique Le controle de la qualite de la conservation des echantillons a ete effectue par comparaison des electroconductivites in situ et au laboratoire et la qualite des analyses par le calcul de la balance ionique et sa comparaison avec I' electroconductivite Au total les parametres physico-chirniques de 39 points (sources, puits, et cours d'eau) ont ete mesures et 26 d entre eux ont fait l'objet d'un echantillonnage pour dosage des ions au laboratoire (Tableau 1). On constate une evolution sensible de l'electroconductivite de 3 echantillons et la balance ionique est comprise entre -3% et +2% pour 22 prelevements sur 26 (LE BEc, 1995) 112 Proceedings of the 12 th Internat i onal Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 124

N i nv 1 2 3 4 5 P6 S7 P8 S10 11 12 14 15 16 17 19 20 P22 23A 23B 24 25 26 27 28A 28B T('C) 11 2 8 1 12 5 12 5 11 8 12 12 2 12 0 10 3 12 5 10 6 12 5 11 1 11 6 12 2 10 7 10 8 1 1 1 1 2 5 1 2, 3 ECTelT 706 3-41 43-4 431 289 338 230 395 35' 456 510 3-49 3-45 256 306 352 291 378 419 561 613 EC Labo 5'7 477 586 537 581 720 3-40 430 430 289 3-40 231 395 326 537 518 355 3-47 261 308 35' 292 327 418 5'19 61 5 pH Te 3 m6.r' ( CrJ > 1,5 m6 r' 0 I K+ J > 0,15 m6.r' ___ [!J2 ROCHES [!J'221!lS10 SIUCEUSES) 11 ~5 23Al26 5 Z 15 e19 1 .00 2.00 -1. 00 0 00 1.00 AXE 1 Fig.2 : Analyse en Composantes Principales des prelevements ,, --... 'POLLUTION I , IIPB'-, AS-, (AEROSOL' ,MARINS 1 ..... ___ .,, 2 00 3.00 Ainsi, a I'echelle du bassin versant, l'influence des roches carbonatees sur la composition chirnique des eaux est nettement mise en evidence. Les ions HCO 3 et Ca 2 + se distinguant clairement des autres elements, nous avons etudie la part de HCO 3 dans la concentration totale en anions (Fig 3). On distingue 3 familles d'eaux: les eaux mi la part de HCO 3 est superieure a la moitie, definissant ainsi un hydrofacies bicarbonate calcique. les eaux ou la part de HCO 3 est inferieure au tiers, correspondant a un hydrofacies chlorure-sodique. les eaux intermediaires. 6 '" C onference on Limestone Hydrology and Fissured Med i a 113

PAGE 125

5 4 .... -Q) 3 .. ..... .., 0 2 0 .... 1 0 0 16 14 15 20 24 S10 S7 23A P22 2 5 _.....23B 12 11 28B 1/2 19 2 P6 1 2 3 4 5 6 7 8 SOMME DES CONCENTRATIONS EN ANIONS (me.1" 1 ) Fig.3 : Relation entre la somme des anions et la concentration en HCO/ L'ACP et la part de HCO 3 clans la somme des anions montrent que les eaux en contact avec Jes roches carbonatees ont un facies hydrochimique se detachant nettement de !'ensemble des eaux du bassin versant, qui ont globalement une composition chimique homogene. On a reporte sur la carte (Fig.I) la position et la concentration en Hco 3 de chaque prelevement, en les comparant a la position des affieurements carbonates deja connus (CHAURIS ET PLUSQUELLEC, 1980). On note que globalement, Jes plus fortes concentrations en Hco 3 coincident avec des zones contenant des roches carbonatees ce qui verifie la relation evoquee. Cependant, ii existe 2 cas ou cette generalite n'est pas verifiee. La source S7 et les puits P6 et PS, situes sur une zone a carbonates, ont des concentrations en HCo 3 plus faibles que prevues (1.69, 2.05 et 0.57 meS 1 respectivement). Cela peut correspond.re a 2 situations: surevaluation de !'extension des roches carbonatees, ou pour le puits, recoupement d'une nappe situee dans des alterites ou des alluvions. Les points 5, 15, 19, 20, 27, 28A et 28B, ont des concentrations en HCO 3 de 2.13 a 4.25 me.r 1 alors que la presence de roches carbonatees n'est pas indiquee et que Jes points echantillonnes a proxirnite ont des concentrations en HCO 3 nettement plus faibles Ceci signifie que !'extension des zones a roches carbonatees est plus importante que signale sur la carte, ou qu il existe au sein d un ensemble sedimentaire siliceux des intercalations carbonatees de faible extension mais suffisantes pour marquer chimiquement l eau. Ces intercalations carbonatees au sein de formations siliceuses se rencontrent en plusieurs points du Finistere comme a Pont de Buis-Loperec, St Segal, lrvillac et Scrignac-Trinivel. Une source karstique existe clans cette demiere localite, d un debit de 15 a 50 1.s 1 et d'une concentration en Hco 3 de 2 me.r 1 Des dolines, la presence d'anciennes carrieres ayant alimente des fours a chaux, ou l' existence de sources a facies bicarbonate-calcique sont autant d'indices qui marquent la presence de niveaux carbonates meme si leur presence en surface n'est pas reperee 5Conclusion Pour le bassin versant etudie on observe que Jes intercalations carbonatees au sein de formations schisto greseuses influencent nettement le chimisme des eaux en leur donnant un hydrofacies bicarbonate calcique par ailleurs peu frequent clans le Massif Armoricain L'utilisation dans l'exemple presente de l'hydrochimie des ions majeurs comme marqueurs de particularites geologiques cachees constitue done un outil interessant et facile a mettre en oeuvre. References AUBRY-DELAFOSSE A S 1995 : Etude du controle fissural de la karstification et implications hydrodynarniques Presqu ile de Plougastel. Juin 1995. Memoire de Maitrise Dpt des Sciences de la Terre, U B.O ., 30 p ., 29 annexes. CHAURIS L. et PLUSQUELLEC Y. 1980 Carte geologique de Brest au 1 / 50 000 274 IV-17 LE BEC G 1995. Identification hydrochimique de la presence de niveaux carbonates clans les series a dominante siliceuses du Paleozoi'que de la presqu ile de Plougastel. Juin 1995. Memoire de Maitrise, Dpt des Sciences de la Terre U.B.O. 24 p ., 3 annexes PLUSQUELLEC Y 1982. Morphologie karstique clans le Paleozoi'que du Finistere. Penn Ar Bed Fev.1982, Vol.13, N: 91-96. WAKEFIELD B. 1993. Leves topographiques des grottes de Pennaneach-Rozegat et de la Pointe de l 'Armorique. Inedit. 114 Proceedings of the 12 1 International Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 126

Dissolution et speleogenese en contexte tectonique actif Le massif des Arbailles (Pyrenees-Atlantiques, F) par Nathalie Vanara Laboratoire de Geographie Physique Appliquee, Institut de Geographie, Universite Bordeaux III, F-33405 Talence Cedex Abstract The "Arbailles" massif (700-1200 m) is a folded area of Jurassic and Lower Cretaceous limestones belonging to the orth Pyrenean Zone and characterized by an oceanic mountain climate (2000 mm/yr) We observe four types of aquifers : I / the Lower Cretaceous aquifers (bicarbonated-calcite springs) ; 2 1 the Jurassic aquifers (bicarbonated-calcic lightly sulfated and magnesian springs) ; 3 / the complex aquifers (bicarbonated-calcic and sulfated springs) ; I the deep aquifers (chlorinated-sodic spring) The mineralization of springs indicates an evolution opposite to that of the discharge, classic in the mountain karsts Waters are agressive during summer (low waters, biogenic CO2) and near the equilibrium or lightly oversaturated during winters (high waters). early 90 % of dissolution occurs between 0 and 100 m deep The plio-quatemary uplift is responsible for the creation of genetic cave levels and underground stream piracies The superficial dissolution is the cause of the disappearance of more than 200 m of limestone since the beginning of Pliocene Resume Le massif des Arbailles (500-1200 m) est situe dans Ja zone nord pyreneenne II forme une unite plissee de 165 km 2 constituee de calcaires jurassiques et cretaces, soumis a un climat oceanique montagnard (2000 mm/an) On y distingue quatre aquiferes karstiques separes par des ecrans plus ou moins impermeables : I / les aquiferes du C retace inferieur (sources bicarbonatees calciques), 2 1 les aquiferes du Jurassique (sources bicarbonatees calciques legerement sulfatees et magnesiennes) 3 / Jes aquiferes complexes (sources bicarbonatees calciques et sulfatees), 4 / les aquiferes profonds (source chloruree sodique) La mineralisation des so4rces montre une evolution inverse a celle du debit, classique pour un karst montagnard Les eaux so ot agressives durant l'ete en periode d'etiage (CO2 biogenique) et a l'equilibre ou legerement sous saturees pendant le reste de l'annee (hautes eaux). Pres de 90 % de la dissolution s'exerce entre 0 et l 00 m de profondeur La surrection plio-quatemaire a entraine un etagement genetique des niveaux de galeries et des captures souterraines d'un bassin karstique a un autre, la tranche de calcaire dissoute en surface etant d'au moins 200 m 1 Situation et problematique une unite plissee de 165 km 2 appartenant a la zone nord pyreneenne (figure 1) II s'agit d'un karst de moyenne montagne (500 1200 m) soumis a un climat oceanique montagnard (1500-2000 mm/an) La forte surrection au cours du Plio-quatemaire, estimee a plus de 1000 m a entraine un enfoncement des reseaux, un etagement genetique des niveaux de galeries et des captures souterraines d'un bassin karstique a un autre On observe egalement une correspondance entre Jes anciennes galeries noyees etagees et Jes grands niveaux perches des vallees seches (VA ARA, MAIRE & LACROIX, 1997) Les recherches speleologiques ont permis de decouvrir plus de 600 cavites et une vingtaine de rivieres souterraines (DELAITRE, 1995) Draine par 138 sources ce massif possede des aquiferes diversifies dans le Jurassique et le Cretace inferieur (figure 2) La majorite des sources presente un debit moyen inferieur a I 1 / s avec tarissement en ete ; elles sourdent soil dans les calcaires marneux albiens, soil a la limite des calcaires urgoniens (Aptien superieur) et des mames de Sainte-Suzanne ( eocomien-Aptien inferieur) Generalement Jes petites sources se perdent rapidement pour alimenter les systemes karstiques draines par 15 grandes emergences (> 5 1/s) Ces demieres jaillissent au bas du massif et marquent le niveau de base des grandes vallees (200 m) Les lirnites des bassins d'alimentation des principales sources ont ete precisees par Ja realisation de 10 colorations (figure 1) (BA UE R, 1995 ; VANARA, 1996) Les parametres physiques et les principaux elements chimiques ont ete mesures pour les huit principales emergences au cours de l'annee 1995 / 96 Ces resultats soot completes par des mesures de durete totale et de Ca 2 + dans differents ecoulements souterrains afin de permettre une prerniere appreciation de la dissolution endokarstique actuelle 2. Fonctionnement hydrochimique actuel des sources Aquiferes karstiques et facies chimiques La serie lithostratigraphique et les facies chimiques des sources permettent de distinguer quatre families d'aquifere s karstiques separes par des ecrans plus ou moins impermeables (figures 2 et 3) : I / Les aquiferes du Cretace inferieur (calcaires marneux albiens et / ou calcaires massifs urgoniens) sont limites au mur par Jes marnes de Ste-Suzanne (Neocornien-Aptien inferieur) Les eaux presentent un facies bicarbonate calcique classique : HC03(155 a 161 mg / I) et Ca 2 + (50 a 54 mg / I) representant 90 % du total de la mineralisation (Grande et Petite Bidouze) ; 2 / Les aquiferes du Jurassique (calcaires du Dogger et calcaires et mames du Lias) sont lirnites au mur soit par les marnes d'Hosta (Oxfordien) soil par les mames du Lias moyen et superieur. Les eaux presentent un facies bicarbonate caJcique legerement sulfate et magnesien (Zahaguy, Urondoa Haute, U thurbietta) : HCOJ(155 a 159 mg / I) et ea 2 + (51 a 53 mg / I) soot predominants (83 et 88 %), mais SO4 2 (6 a 19 mg / I) et Mg 2 + (2 a 4 mg / I) sont egalement presents ; 3 / Les aquiferes complexes de la bordure sud soot caracterises par la liaison hydrologique entre les differents aquiferes dOe a la reduction d'epaisseur des couches (Trias a Cretace) Les eaux presentent un facies bicarbonate calcique et sulfate (Arhanzeta, Cent Sources et Garaybie ) HCOJ ( 152 a 181 mg / I), ea 2 + (60 a 73 mg / I) et SO 2 (50 a 101 mg / I) sont en proportion importante (86 a 88 %). Les aquiferes profonds sont caracterises par la forte concentration des ions HCOJ ( 183 mg / I) Ca 2 + (494 mg / I) Na+ ( 368 mg/I) SO4 2 (1065 mg / I) et er (6922 mg / I) ( moyenne de deux analyses de la source thermale de Camou a facies chlorure sodique 6 th Conference on Limestone Hydro logy and Fissured Media 115

PAGE 127

Vallee seche i,:-Cours d'eau + + + Axe anticlinal A~e synclinal Faille normale Calcaire maroeux -VVFaille inverse ou chevauchante D Relation prouvee par tra~ge CJ A -'"' qw.iere urgomen o Source (comp chimique non defwie) Aquifere jurassique Ava111-pays dujlysch Source bicarhonatee calcique Source bicarhonatee calcique Jegerement sulfaltee et magnesienne Source bicarhonatee calcite et sulfatee Source cWoruree sodique j:. ::.: J Niveau350-l00 m Niveau680 750m Niveau 800-850 m Niveau 900 -9 50 m Massif d'lgou111ze Figure J : Croquis hydrogeologique du massif des Arbailles. Pararnetres physiques et regimes chirniques La moyenne des temperatures des huit emergences etudiees est de J l 0 C. Le pH moyen est de 7 6 Les variations aisonnieres sont nettes Le pH est inferieur a 7 en juillet ao0t et superieur a 7 le reste de l annee Le CO2 dissous (moyenne de 22 mg /I ) augmente l'ete en fonction de l'activite biologique La conductivite des huit sources varie de 27 a 66 i em La durete totale etroitement liee au debit, nuctue de 136 (Petite Bidouze) a 225 mg / I (Ar hanzeta ). Le TH augmente en periode de basses eaux Uuillet, ao0t novembre) et diminue en periode de hautes eaux (mai, septembre, hiver) C omme pour le TH, !'evolution du Ca 2 + est tributaire des debits : valeurs maximales atteintes en etiage, valeurs rninimales en crue (phenomene de dilution par des eaux peu chargees venue s rapidement de la surface) La moyenne annuelle du ea 2 + est de 57 mo i l (50 a 73 mg / I) sur les Arbailles. Les rapports entre le maxi~um et le minimum vont de 1 ,3 (Petite Bidouze) a 2,2 (Arhanzeta) La moyenne annuelle du Mg 2 + des sources va de 0, ( U thurbiena) a JO mg / I (Garaybie), pour une moyenne de 5 mg / I sur !'ensemble du massif. La moyennedu HCO3 est de 143 mg / I pour !'ensemble des sources etudiees Les valeurs des moyennes pour chaque source vont de 152 (Arhanzeta) a 181 mg / I (Garaybie) Les valeurs extremes montrent un minimum de 131 et un maximum de 20 I mg / I pour des rapports de 1,2 a 1 4 Les fortes teneurs en ions sulfates se manifestent !ors des etiages notamment pour trots sources aux aquiferes complexes (Arhanzeta Cent Sources et Garaybie, moyenne annuelle de 71 mg / I) L'aquifere jurassiqu e draine en partie par U rondoa Haute, U thurbietta et Zahaguy possede une moyenne de 10 mg / I. S04 2 es t absent de l'aquifere Cretace (figure La moyenne annuelle des ions Na+, K +, c i, :--103 et PO est respectivement de 2 0,5, 6, 3 et 0,03 On est en presence d'une dissolution relativement classique dans un lcarst de moyenne montagne forestiere du domaine tempere humide avec des eaux agressives l'ete et a l'equilibre ou legerement sursat urees le reste de l'annee en fonction de l'activite des sols et de la vegetation (CO2 biogenique) 3. La dissolution epikarstique et actuelle specifique, souterraine Le suivi des principales emergences des Arbailles montre que la moyenne de la mineralisation (alcalinite) est d'environ 160 mg / I de HCO) et 165 mg / I de CaCO3 (TH) A l'altitude moyenne de 900 m (T moy = 8,9 C), Jes precipitations annuelles sont de 2029 mm et l'evapotranspiration reelle de 540 mm, soil une tranche ecoulee de 1489 mm/an On estime done la dissolution specifique (superficielle et profonde) a environ 92 m3 t lcm 2 t an, soil 0,09 mm / an La durete totale depuis la zone d'absorption jusqu'aux emergences passe par deux phases d'evolution principaJes (figure 5) : I) une augmentation rapide dans la zone epilcarstique en rai _so n de la dissolution d'origine pedologique et du regime d'ecoulement par percolation et suintement. La durete passe ainsi en moyenne de Oen surface a 140 mg / I a -JO m 116 Proceed ings of the 12'" International Congress o f Speleolog y, 1997, Switzerland Volume 2

PAGE 128

C r e a C e J u r a s s i q u e '}T"--I --T-_. ~;:ii;: ~n:Cu:, I I I J ~2'.~!f~4i~ict1i,_ I I e 7 I I ( n l-l-l-l,l ASpiculcs ______ --r...__....., ________ , '\ p T J..: N t M a I :: .. f L i a I I \ I I I -J l\rtan.10--caJcaiJ"CS "" -1-1-1----E 1-1-1-r -1-1-14 I (-6-ic de N11bol~8UY I c td'lspu.c:,oa) -1-1-1'_______ J nt,h nt,a n4 II II II 1-1-1-1-1-1-1II IJ Aquif~rc des cn l cafrcs A ( aciCs urg o ni c u (s"1ric ca lC11in :: d Ascunc e t du L.Abott) I Aquift,;1-c ba,~micn e t k:in1n1~1i.dgic11 r, i-:.:-~::::t:::::::I=~~ "'." '1-fi_ _ 1 bcmn dci:nan"1CS 17 9 I I d:!!"!."' I t Aquilcrc du + Ooggcr Tri a.~ ,n.qM!ricur L7 Pcm1,o1 nn s LI 2 ,._ II II II I Carbo nif C1'C h S 200m Figure 2 : Echelle lit/1ostratigraphique (d'apres Fabre & Navarre, 1992 ; modifie), aquiferes et ecrans hydrauliques. 2) une augmentation globalement lente dans les rivieres souterraines La durete y passe de 140 a 170 mg/I environ Cependant, on observe des valeurs tres dispersees selon les types d'ecoulements (percolation, suintement, riviere) allant de 100 a 200 mg/I (les percolations etant plus mineralisees que les rivieres) On peut done considerer que la dissolution se repartit ainsi : 80 a 85 % dans la zone epikarstique et 15 a 20 % dans la zone eodokarstique Pres de 90 % de la dissolution s'exerce entre 0 et JOO m de profoodeur. Ces resultats soot en accord avec cetL~ de DELANNOY ( 1983) qui estime que, dans les karsts forestiers du Vercors, la corrosion preferentielle se produit dans la tranche superficielle du karst : 55 % de la dissolution totale entre 0 et -50 m et jusqu'a 80 % entre 0 et I 00 m Dans Jes karsts supraforestiers nus, la dissolution entre 0 et LOO m represente 65 % de la dissolution totale ( 15 % pour la dissolution superficielle) La porosite karstique diminue en profondeur eanmoins, la concentration des ecoulements permet de conserver un potentiel de dissolution suffisant pour creuser des galeries importantes qui s'organisent en quelques collecteurs orincioaux 4. Les etapes de la speleogenese (Fig. 1) Le karst des Arbailles est une surface heritee, a elements de plateaux et buttes residuelle s (l 000 1250 m) perchee 800 m au-dessus du niveau de base regional Les reliefs residuels sont separes par un systeme de vallees seches et par des depressions et des megadolines Plusieurs families de cavites. dont les plus anciennes sont mises au jour par !'erosion attestenl de l'anciennete de la karstification Certaines presentent des temoins sedimentaires cornme celle du Belchou avec sa coulee stalagmitique tongue de 17 m et epaisse de 70 cm situee au sornmet, a l 127 m d'altitude Ce temoin de grotte subsistant au sornmel d'une grande butte indique !'importance de la tranche calcaire erodee depuis le eogene, date a laquelle se situeraient ces anciennes karstifications Au eogene et au Pleistocene l'ensemble du massif se souleve de 1000 m en plusieurs etapes Les accidents majeur s des Arbailles sont jalonnes par des sources minerales et / ou thermales. Les sources sulfureuses et ferrugineuses de Garaybie sont situees sur le trace de la faille chevauchante de la bordure nord La so urce chloruree sodiq ue de Carnou jaillit a 30 C sur le trace du grand accident decrochant de la bordure orientate (zone sismiquement active) La temperature et la mineralisation indiquent le le ssivage vers 900 m de profondeur des marnes gypsiferes du Keuper. Les vallees seches et les reseaux souterrains ont enregistre les etapes de la su rrection par des niveaux etages reajustements des ecoulements en fonction de l'enfoncement progressif des vallees bordieres Plusieurs niveaux subhorizontaux sont visibles : 380 m 700 m, 850 m et 950 m Le niveau majeur de 700 m est ii lustre par des troni;;ons de vallees seches de 2 a 3 km de long, a faible pente (I a qui representent un important stationnement du niveau de base regional conserve par immunite karstique apres la surrection Dans le reseau souterrain d'Etxanko Zola. le niveau fossile de 675 m (galerie des Planches a C lous) correspond au niveau subhorizontal de 700 rn de la grande vallee seche d'Ithe A la suite de la surrection, la galerie s'est assechee de meme que la conduite forcee de la zone d'entree La capture des eaux souterraines du reseau d'Etxanko Zola vers le nouveau niveau de base de l'Apoura (Cent sources) s'est effectuee a la faveur du pendage general vers le Set le SW -~tVhandta --o-Gr.and Bidouzt Uthl.rt)ittta --+-C.nt sourOH P#tit Elidouz ZohOIJUII G.r11,1bi. Urondoo Hout --c.mou 1000 -,-----------------------, 100 )( ... .. .... "><.. / 10 ,; '\ /' ; 1 ,01 ', ; OOl-'-~TH'--~C.J++-~HJ++--N~ ... --K~ -C~r~~S -04-...__HC03_,_ _ N03 ___ ---' Figure 3 : Diagramme de Schoeller-Berkaloff Les facies chimiques de neuf emergences des Arbailles. 6 1 h Conference on Umestone Hydrology and Fissured Med i a 117

PAGE 129

800 120 C ooductiiti a,++ -Ar:banz.cta Peate B i ~c I' S/ cm mg/I lw&UY G,nyb, 7 00 UIJ:mbietta m.tScx.-ces U roadooflaule Gnod< Bulouu 100 600 220 HCOJ mg/I 500 200 /\ II) I \ 400 ,' ff/\ 180 300 ic;r, .. .... 160 : 60 200 r 140 1 00 Figure 4 : Evolution de la conductiviU, des ions ca2 +, HC03_ et S04 2 de huit g r andes emergences des Arbailles (Annie 95-96). Plus bas, un autre jalon est donne par le niveau subhorizontal de 365 m du Nebele, dont la morphologie indique un creusement en regime noye en correspondance avec un ancien niveau de base de la vallee d'Ithe A la suite de la surrection, la galerie s'est assechee permettant le depot de speleothemes En datant la calcite de la generation de stalagmites la plus ancienne du niveau de 365 m, on date le debut de \'assechement et par consequent l'age de la surrection La generation la plus ancienne a un age U / Th de 407 000 ans (+ 78 700, 46 600) (analyse Y Quinif, CERAK, Mons) On estime done que le massif s'est souleve de plus de 150 m en 400 000 ans environ, soit un taux de surrection de 0,4 mm/an. TH (mg/I de CaC03) 300 ,.--,-r,-rrn,r--,-r,-rrn,r--,-r,-n-r
PAGE 130

Microbiological processes at the cave development and karstification Benjamin Menne BfUS Hartfeld s tr.32 D-754 I 7 Mlihl a cker ( G e rman y) Abstract The process of karstification can be regarded as part of the carbon-cycle of the earth. The assumption that microorgani s m s could play a role at the Speleogenesis was repeated in the past. To a guest extent these con s iderations have been intensively contradicted Microbiological sediment researches in European cave s led however to new cognitions Th e possible role of microorganisms at the karstification is represented with the models of the "Carbonatol y se" and "Bioconservation" Basis of these models is the "Festbettreaktor" (trickling filter) a common type of sewage water treatment. The formation of biofilms is described both as model and as experimental examinations Data from some cavesystem s are introduced The consequences of the results for the karstwater and its use are briefl y explained. Der Vorgang der Verkarstung kann als Tei! des Kohlenstoffkreislaufes betrachtet werden. In der Vergangenheit wurde wiederholt die Vermutung geauBert, daB Mikroorganismen eine Rolle bei der Seplaogene s e spielen konnten. Die s en Oberlegungen ist teils heftig widersprochen worden Mikrobio l ogische Sedimentuntersuchungen in europaischen Hohlen filhrten je d och zu neuen Erkenntnissen. D i e mogliche Rolle von Mikroorganismen bei der Verkarstung wird anhand der Modelle der "Carbonatolyse" und der "Biokonservienmg" dargestellt. Grundlage dieser Modelle ist der Festbettreaktor. Die Bildung von Biofilmen wird sowoh l modellhaft als auc h experimentell unters u cht. Oaten aus a u sgesuchten Hohlensystemen werden vorgestellt. Die Konsequenzen der Befunde filr das Karstwasser und seine Nutzung werden kurz erlautert Results and Discussion It is obvio u s that the corrosion of limestone is a surface-chemistry mechanism For this the structure a nd co n dition of the inner surfaces of the karstsystem are most important. Surprisingly we know only a little about the real circumstances of the crack and fissure surfaces. Duri n g the last decade we examine d h u ndreds of sediment samples of European caves. We found that there is a high variety and diversity of the subterraneous microbiocoenosis but there are as well clearly perceptible common facts. We discovered for example that the karstic rock-mass is divided into some mainly vertical microbiological reaction zones : Subcutum Epiklasal, Hypok l asal and Hydroklasal (MENNE 1997 ). The complex three dimensional structure of microbiocoenosises within the karst systems leads to questions about the occurrence of biofilms at the crack-surfaces We realized that the existence of biofilms is clearly indicated There are also high-comple x microbiological structures on the fissure surfaces. Biofilms cover surfaces The first step in the process of bacterial attachment and biofilm formation on crack-surface s i s adhesion-mechanism (adsorption). Three-dimensiona l growth is possib l e Even if the biofilms including perhaps abiotic components (for example clay -p articles) develop on the fiss ur e-surface without attacking the limestone the corro s ion conditions change dr astically We term this hypothetica l situation ,,Bioconservation ". On the other hand generally all kinds ofmicroorganisms are ab l e to attack and degrade materia l s. The deterioration is caused by the excretion of metabolic products (SAND 1996) One of the most common products is carbondioxide Within biofilms relativly h i gh concentration s of carbo n dioxide are possible The production of other acid substances was observed b y technical methods The destruction of natural stones by means of biodeterioration processes are proved (W ARSCHEID & KRUMBEIN 1996 ). Thi s indicates that these processes can a l so take pl ace in karstsystems We te rm t h i s ,, Carbonatolyse ". Literature MENNE B. (1997) : Myxobakterien in der Rettenbachho hl e und in einigen Kleinhohlen des Sengsengebirges Eine karstmikro b iologisc h e S tu die.Karstprogramm 1 996 unver. Fo r sc h ungsber. d Nationa l parks Kalkalpen (Oberosterreich ); 34 s SAND W. (1996): Microbial Mec h anisms in : HEITZ E ; FLEMMING H.-C. ; SAND W (Eds.) : Microbiall y Influenced Corrosion of Materials. Springer WARSCHEID TH. ; KRUMBEIN W E (1996) : Case Histories Inorganic Non-Metallic Materials Genera l Aspekt s and Selected Cases in : HEITZ E .; FLEMMING H C.; SAND W (Eds ): Microb i ally l nfl u e n ced Corrosion of Materials Springer 6 '" Conference on Limestone Hydrolog y and Assu re d Media 119

PAGE 131

Seasonal Fluxes of Humic Substances in Cave Drip Waters, Marengo Cave, Southern Indiana. P.E. van Beynen, V.A. Toth, O.C. Ford and H.P. Schwarcz Schoo l of Geography and Geology, ~k11aster1 1 niversity Hamilton, Ontario L8S -1Kl, Canada Abstract Water samples were collected for 15 months and analyzed for their fluorescence and IX>C content. \ positiYe relation was apparent between total fluorescence and the DOC content of the cave water lhrough four sites within the caYe Howe, er, some sites produced stronger relations than others Seasonal variabilit) is evident, but again this varied from site to site. Spectroscopic studies suggest the fluorescence is attributable to organic acids most particularly fulvic acid (F.\). l 1traftltration experiments add weight to this finding as most of the organics present in the ca, e water were less than 1000 ~fWCO, a range characteri tic of F.'\. Rationale The paleoenvironmental significance of this study is related to annual bands of calcite growth in speleothem which have been discovered using luminescence microscopy These detenninations are only possible if the seasonal fluctuations in organics are being recorded in the calcite Drip water entering the cave which feeds the spe leothem must contain a seasonally variable organic concentration Through weekly collection of cave water and its analysis using techniques outlined above. any variability can be ascertained Comparisons between variability of organic content and climate factors could result in the creation of relations between climate and fluorescence of speleothem. Future testing of these relations using speleothem of known modern age from similar climatic regions could he undertaken Study Area l\farengo Cave, Southern Indiana (latitude 38'22"N, longitude 86'20"W) is optimal due to ease of access, seasonal variability of locale, cooperation of cave staff in sampling, and a lush temperate deciduous forest above the cave producing significant amounts of organic matter. ~1arengo Cave is contained within the Ste. Genevieve limestone with the Paoli limestone overlying it and capped by the \Yest Baden Sandstone. Cave water was collected at four sites within the cave over a 15 month period. Methods DOC Ie, els were attained using a Dohnnann DC 180 Carbon Analyzer. ,\ ~1odel 10 Turner fluorometer was used to establish the total fluorescence intensity of all samples ; \ Perkin -E lmer LS-5 fluorescence Spectrophometer recorded the spectra as collected with 5nm s lit width for the monochromator llltrafiltration was carried out using OA5um, 10.000, and 1 000 ~f\\ CO pore size ~fillipore Pellicon cassette membranes l\folecular weight fractionations coupled with DOC val ue s help ascertain which dominant sizes are present. Results DOC and total fluorescence analyses reveal that spring is the period of greatest infhu of organics into the cave 1bis coincides with warmer temperatures and associated snow melt. ,\ reasonable positive correlation is also evident between fluorescence and organic concentrations of drip water. Spectroscopic analysis suggests that FA is the dominant organic substance in the water Peak fluorescence emission values were recorded at 425nm a wavelength characteristic of FA although at lower excitation wavelengths than expected (280nm) Molecular weight separations revealed through DOC \ alues suggest that the soluble organics are mostly below 1000 l\lWCO demonstratiYe once again of FA Relationships to climatic variables have still to be determined, though preliminary results show the influence of storm events on fluore cence Conclusion The above result provide important information for the paleoenvironmental work to be done on interpreting fluorescence in speleothem Seasonality in cave waters lends support to the existence of annual bands in speleothem r-or the interpretation of the fluorescence generated by speleothem, meaning is given by the relation between fluorescence and organic carbon content in drip water The characteristics of this organic matter is revealed through spectroscopy as probably fulvic acid and lhrough ultrafiltratioo as small molecular weight compounds again indicative of fulvic acid. 1 20 Proceed i ngs of the 12 t h International Congress of Speleology 1997, Switzerland Vo lum e 2

PAGE 132

Systematic Evolution of Subterranean River Caves Abstract Zhu Xuewen Institute of Karst Geology, Chinese Academy of Geologic Sciences ; 40 Qixing Road, Guilin 541004 Guangxi P R.China There are thousands of underground river cave systems in karst regions in Southern China. The author based his study on an input-output system of material and energy (usually is a hydrologic system), proposes a new plan for the classification of underground river caves ; believes that caves of the different parts of the system should be developed by reciprocal inhibition and synergy under the control of the entirety of the system ; and explains the characteristics and the evolution pattern of the caves of all types Cave genetics take a very important position in Karstology Davis (1930 ) proposed a two-cycle genetic model for limestone caves and believed that caves was first formed in the phreatic zone and then lifted up to the vadose zone at later s tage Nowaday s, genetic theories concerning karst caves such as vadose-origin phreaticorigin and underground water table-origin are also well known to us Ford (1989 ) has put forward a FourState Model to describe how lime s tone caves are developed from phreatic zone to ground water table. Till now the exi s ting speleo genetic theories are still isolated from one another. Noticing that there are a great many of subterranean river systems developed in karst regions in Southern China and because the formation and evolution of karst caves could definitely be ascribed to a certain input-output s ystems of material and energy we have established a new model to explain the formation and development of subterranean river caves on the basis of the re s ults of the simulation experiment by Ewers ( 1982) First, we distinguish the input output movement of material and energy within an independent genetic system into lateral, planar, and multiple input three ba s ic types ; con s equently caves of the genetic system could consequently be classified into input, output, and aquifer caves of three fundamental types ( Fig I) Input caves mainly consi s t of vadose and water table cave, output cave i s dominated by water table cave and aquifer cave is normally characterised by phreatic cave (including deep phreatic zone) In this new model, we bring every basic cave-genesis theory into an organic genetic system. In geomorphic history it is quite usual to observe caves of all types being uplifted into vadose zone. The research shows that the development, enlargement, and extension of input cave usually start from the input point; and the principal input cave normally has a relatively fast development and high developing intensity in comparison with output cave within the same underground river system During the formation of an underground river cave, the retrogressive erosion only occurs in the late development stage of the output c ave when conduit water flow becomes turbulent. The development of an aquifer cave is of high degree of selection in different stages and most aquifer caves will s top developing and gradually be abandoned as they will h a ve no water flow, as water flow in the aquifer tends from initial disorder to an ordered state. The main evolution s tages of a n underground river cave system are as follow s ( Fig I ): I Early stage of the formation of a karstic hydrologic input-output sy s tem : with the formation of an independent relatively stable input-output hydrosystem, the hydrologic and hydrogeologic process has completed preliminary ordering ; the principal input cave has been ascertained and accelerated development; output c ave still occurs as karst s pring ; and aquifer caves are dispersed and in a disordered state. 2 Main s tage of the ordering development of the inputoutput hydrosystem: Input cave speeds up the extension and enlargement process ; output cave starts to be formed and develops into the inner part of the a quifer with retrogressive erosion ; and aquifer caves are selected to develop intensely. 3 Breakthrough stage of the input-output cave s ystem : Output cave and the principal input cave are connected with each other and become a welldrained passage in the aquifer; the underground hydrologic network becomes mature and the underground cave s ystem has been formed; most of the early developed aquifer caves are abandoned; Water table cave i s the dominant landscape in cave development. The collapse skylights occurred on the surface along the main underground river cave and even part of the underground river course exposed to the surface 6 '" Conference on Limestone Hydrology and Fissured Media 121

PAGE 133

+ t Input t t Inflow cave Aquifer cave ------Outflow cave A Output .D Inflow cave --. -Aquifer cave Outflow cave o Inflow cave Aquifer cave Outflow cavec Output D m Input. -----~~-Ill Inflow cave Outflow cave cave ----~ Outflow cave Aquifer cave ~f~w1 Auifer ---leave Outflow cave Fig. I The model of cave development and its systematic evolution in an underground river system. A-lateral input; B-Planar input, C-Multiple input; I, II, III -Evolution stages The s ystematic evolution model of underground river cave is very useful to study and distinguish the natural feature and the evolution stage of underground river systems in Fengcong mountain are a s of South China and helpful for the exploration and exploitation of karst water resources. References BOGLI A. 1980. Karst Hydrology and physical speleology Berlin: Springer. DAVIS W M 1930 Origin of limestone caverns. Geol.Soc Amer.Bull. EWERS R.O I 982. Cavern development in the dimensions of length and breadth Ph D thesis McMaster University FORD D C. and WILLIAMS P W 1989 Karst geomorphology and hydrology Unwin Hyman Ltd London. ZHANG REN. 1993 New consideration on classification of karst caves. Proceeding s of the international congress of speleology. Beijing. ZHU XUEWEN and ZHANG REN I 993 The formation and evolution of Nandong underground river system Yunnan Proceedings of the international congress of speleology Beijing ZHU XUEWEN. ZHANG REN. ZHANG YUANHAI 1995 Karst and caves in Xingwen stone forest area, Sichuan Car s ologica Sinica, Supplement(in Chinese ) ZHU XUEWEN. ZHANG YUANHAI. ZHANG REN et al. 1996. Cave development and its systematic evolution in an underground river sy s tem Carsologica Sinica. No.1-2 Vol. I 5 122 Proceed i ngs of t he 12 "' Internat i onal Congress of Spe l eo l ogy 1 997 Switzerland Vo lu me 2

PAGE 134

Characteristics of hydrochemical responses to environmental change in a carbonate rock Aquifer Yuan Daoxian The Institute of Karst Geology Guilin Guangxi China 541004 Carbonate rock the biggest carbon reservoir on the Earth is the result of the processes in geological history that reduced the CO 2 content in the atmosphere. By this way it has played an important role to make the Earth s environment favourable for the development of life and human beings. However it is considered that this big carbon reservoir is no longer active in the modem carbon cycle since the time scale of mankind Recent works show that carbonate rock is still sensitive to global carbon cycle and has impact on Man s environment both macroscopically and microscopically. These ideas could be explained with a minor example. In Fangtang village Liujiang county Guangxi China people suffered from the acidic water of a spring which is their major source of water supply and recharged by a siliceous rock aquifer of Upper Permian However in a farmer s private well sunk in the same aquifer just 50m away from the spring the water is neutralized quickl y and improved remarkabl y because the well is lined by limestone blocks (Fig. I). reservoir pH: 5 11 DH: 0.8(HCO3") T: 20.8c siliceous rock limestone ..... rain water pH : 4.05-5.59 average of 1983-1989 spring pH: 4.42 DH:1.4 ~-if r : 20c OH : bicarbonate hardness in German degree imestone lining Fig.] Hydrochemical contrast between waters from the same siliceous rock aquifer, but different in limestone lining, Fangtang village, Liujiang county, Guangxi, China Pr e cl H C03 mm/ mg/t 28 250 2 400 23 2200 2 1 0 2000 190 1800 170 1600 150 wet per1oo-t----dry period ---+-we t period 1 4 130~--------~ ----~ c a 2 (mg/l) 80 70 60 50 Fig.2 General trend of HCO 3 -,Ca 2 + rising in a Devonian limestone aquifer of Guiin,China during the past decade, and its impact from change in precipitation(after Liu Jingrong, 1996) The increase in atmospheric CO 2 following the increasing use of fossil fuel will result in the intensification of carbonate rock dissolution and high content of ca2 +, Mg 2 +, and HCO 1 in water. However the process is highly influenced by climatic factor Long term hydrochemical monitoring in Guilin( 1981-1994) Guangxi China shows not only a general increasing trend of Ca 2 \ Mg 2 \ and HCO 1 in limestone or dolomite aquifer but also their higher content in the dry period(l 984-1989), and lower at the wet periods (Fig 2 Liu Jingrong 1996) The hydrochemical responses of carbonate rock aquifers to acidic water from mining areas such as pyrite contained coal mines mainly appeared as the increase in sulfaes hardnes s. The over exploitation of water resources from an evaporite-mixed carbonate rock aquifer has the similar r e sult but their contributions can be distinguished by an isotopic-hydrological model based on sulphur isotopic data The behaviour of heavy metal pollutants such as Cd 2 +, Cu 2 +, Pb 2 +, Zn 2 + etc. in a carbonate rock aquifer are mainly reflected as to their absorption by the wall rocks. The phenomena are observed both in laboratory experiments or field practices. However where do they stay how and when will they be mobilized again remain a concern. 6 th Conference on Limestone Hydrology and Fissured Med i a 123

PAGE 135

. r

PAGE 136

Nature and transport of natural colloids in a karstic aquifer Olivier ATTEIA Centre d Hydrogeologie Rue Emile Argand 11 CH-2007 Neuchatel Switzerland Abst r act This p a per pre s ent s a study of colloids and particles in a karstic a quifer of the Swiss Jura The studied site has been chosen for the presence of cl a y layers and peats, area able to generate colloids The study of particle size distributions (PSD ) shows that all curves can be fitted by the same model distinguishing large particles the number of which depends on flow conditions and small ones influenced b y pH The breaking point is considered to be the limit between colloids and particles The analysed colloids consist mainly of clays, quartz and complex particles made of organic matter Ca and Fe The formation and evolution of colloid s and their role on contaminant transport are discussed Resume Nous presentons ici J etude de collo'ides et de particules dans un aquifere karstique du Jura Suisse. Le site a ete choisi en raison de la presence de marnes et de tourbieres qui peuvent liberer de s quantiles notables de collo'ides naturels Les courbes de di s tribution de taille des particu!es rencontree s peuvent etre ajustees par un modele s eparant grosses particules dont le nombre depend des conditions hydrodynamiques alors que Jes petites sont reLiees au pH Ceci nous permet de separer collo'ides et particules Les collo'ides analysees sont principalement constitues d argiles, de quartz et de particules complexes imbriquant matiere organique, Fe et Ca La formation et !'evolution des collo'ides sont discutees ainsi que le role eventuel des particules dans Le transport de polluants 1. I n tr oduc ti o n Particles are generally considered as colloids when their size implies a long settling time in water, typically from days to months In natural waters colloids are usually clays, iron ( oxihydr ) oxides silica, organic matter and bacteria Concerning their chemical properties colloids are often defined as charged species and solids having large specific areas ( BUFFLE & LEPPARD 1995 ). This implies a high reactivity which justifies their ability to transport contaminants. Colloid studies on karstic aquifer s are of major interest for two main reason s: ( i ) these aquifers are highly vulnerable to contaminants ( DOERFLIGER & ZWAHLEN 1994 ) a nd therefore any potential contaminant carrier must be studied and (ii ) the discharge variations are very large ( KJRALY 1988 ) thu s enabling the study of colloidal behaviour in very different hydrodynamic c onditions The uniqueness of k a r s t a nd the de v eloping concern for colloids incited the study presented here This paper gathers two aspects of colloids : ( i ) the evolution of Particle Size Distributions ( PSD) and (ii ) their chemical and mineralogical characteristics. In fact PSD and particle type a re both nece s sary to understand the role of colloids on contaminant tran s port The a nalysis of PSD variation is very u s eful as no other information exi s ts on karstic aquifers Beside we would like to analyse the role of the specific hydrodynamic properties of karstic aquifers on the particulate and colloidal transport. The characterization of colloids is of primary importance for contaminant transport because the sorption properties strongly differ between colloid types The analysis of colloid mineralogy, composition and morphology at several points of the basin should also outline the effect of solution chemistry on the colloid behaviour 2. Site The 60 krn 2 Noiraigue basin is located in the Swiss Jura mountains at an altitude ranging from 1000 to 1350 m Average rainfall, over the last 70 years, was about 1500 mm/y. The catchment area is represented by ( i) a flat valJey ('Vallee des Ponts ) covered after the last glaciation by peat-bogs and ( ii ) the surrounding calcareous anticlines The Noiraigue Spring is the only outlet of this karstic aquifer and lies at 750 m flowing at the outcrop of a regional fault. Fine Molasse sandstone covers the limestone in the centre of the valley with a maximum thickness of 300 m. These tertiary sediments of low permeability are overlaid by thin glacial till and post glacial marly deposits Finally during the last 6000 years peat wa s formed on these low permeability deposits covering approx. 20 km 2 The other two-thirds of the catchment basin are composed of limestone anticlines covered by thin organic and Carich soils The Bied brook draining the remaining peat areas and the agricultural soils of the valley ends in a kar s tic sink.hole located in the Cretaceous rocks The karstic network is well developped with water velocities varying typically from I O to 60 m/h and discharge ranging from 0 5 to 20 m 3 /s The quaternary marly layers covering the valley are potential sources of colloidal clays Peat areas should be able to deliver very large amounts of fulvic and humic acids as they contain more than 90 % of organic matter (MATTHEY 1986 ). When ferrous iron coming from the anoxic peat layers will enter in contact with oxic waters, it will precipitate and can therefore form colloids The site is equipped for continuous measurements of water level temperature electric conductivity, turbidity and rainfall on the Bied brook and at the Noiraigue spring Samples were taken almost weekly from May 94 to November 95 for chemical a n d particle analysis During some high flow events, samples were taken automatically at shorter time intervals 3. Methods Size a n a l ysis The PSD data presented here were obtained by single particle counting The used CIS Galai apparatus counts individual particles from 0 5 to 60 m by measuring the duration of the shadowing of a rotating laser Light by particles There is no need for any pre-treatment or fractionation of the sample This technique has been applied under continuous flow (50 ml/min) and stirring conditions Three repetitions were 6 th Co nferen ce o n Li me s tone H y drolog y a nd Fi ss ured Med ia 1 25

PAGE 137

always carried out (JANTSCI-IlK ET AL. 1992) and the analysis was done until the reproducibility was sufficient. Different tests were performed to assess the quality of the PSD determined by CIS Three cultures of bacteria were counted by epifluorescence and by CIS ; the standard deviation between the two techniques was lower than 20 % for a range of more than one order of magnitude For one campaign CIS data were compared with particle counting on Nuclepore filters of 2, 0 8, 0.2 and 0 05 Three pictures of each dried filter where taken by scanning electron microscopy (SEM ) 1500 particles were then measured and counted by using an image analysis software. The two techniques agree well, with a difference lower than 25 % for measurements over several orders of magnitude Particle composition Filtration has been used to fractionate the samples Knowing the important constraints on filtration ( BUFFLE et al. 1992, MCDOWELL-BOYER et al. 1986) we developed a cascade stack with Nuclepore 2 0 0 8 0 2 and 0.05 m filters ( 47 mm 0 ). For the three initial filtration steps, the suspension is filtered by gravity us i ng a 20 cm water pressure; then pressure i s imposed on the system to flow through the 50 nm filter. The mineral composition of the water samples from the first campaigns was determined by XRD using centrifugation at 1800 g during 20 min This allows the deposition of the > I m fraction For further campaigns we used a continuous ultracentrifuge ( Sharple s s Tl) at 13'000 g We obtained classes of the following calculated sizes : 0 2 0.4 0.4-0 65 0 65-1.3 and 1.3-5 m The limit of the classes are approximate due to the uncertainty on density and to turbulent motion existing in the tube. The analysis of individual particles was processed with a SEM-EDS (Jeol JSM-35 ) Particles from the Bied brook and the Noiraigue spring were recovered at 4 sampling dates (19.07 26 07 09 08 and 16 08 1995) onto Nuclepore polycarbonate filters (2 and 0.8 ) Approximately 1000 particles were randomly chosen and analysed under identical conditions for Mg Al Si P K Ca Ti Mn and Fe The bacteria flora was enumerated by using the following two techniques : ( i ) liquid cultures by serial dilution of sample s in microtiter plates or incubation of filters on solid media and ( ii ) Epifluorescence using a photonic microscope with fluorescent light source 4. Factors controlling particle size distribution {PSD) A review of analysed PSD in natural media (A TTEIA & KOZEL 1997 ) showed that the amount of colloidal/particulate matter in Noiraigue Spring falls intermediate between deep groundwaters and large rivers The peaks of the volume weighted distributions occur at fairly large sizes compared to any other groundwater. More than 55 samples were analysed for the Noiraigue spring and 35 for the Bied brook Each curve consists of more than 100 values before the data is treated and 18 values after class aggregation In order to analyse the whole dataset modelling of the curves was necessary Most of the PSD curves are composed of two parts The break point i s generally s ituated near 5 The first part of the curve is best fitted by a model in which the number of particles is a power law of particle size s ometime called a Pareto distribution (LERMAN 1979 ) dN =a q, -b d q, where N is the particle number and qi the particle diameter b is the slope and a the intercept. The second part of the curve can be fitted by a Pareto law or an exponential law dN dq, =a exp ( -b qJ ) The symbols have the same signification as above The criterion of choice between the two curves was the mean square error In this paper, we deal only with the values of the model at specific sizes (0.5, 4 and 10 ). These will be referred as N o s N 4 and N 1 The values are given in number of particles/ml/m The dm ax parameter is the size, in m, at which the model crosses the value of I O particles/ml/m. At Noiraigue, the most influential parameter on PSD curves is the effect of spring discharge on the shape of the second part of the curve The dm ax parameter increased with the square of the discharge, as shown in Fig I However some samples show higher dm ax for the same value of discharge These points were generally sampled during the steep ris i ng of the hydrogram Other environmental factors such as temperature or electrical conductivity seem to play an insignificant role on the parameters of the second part of the model. 45 f(x) = 0 155 x 2 0.54 x + 10 5 40 2 = 0 81 35 m 30 E ::i 25 X 20 ml "' E 7 'O 15 I Others 1, m Rising limb 10 "1 1 1 I i' ltn I 11 I 5 I 0 0 2 4 6 8 10 12 14 16 Discharge (m3is) Rising limb : samples collected during the rising limb of s torm hydrogram Other s: samples coll ec ted during other periods 18 Fig 1: Relation between dmax parameter and discharge at Noiraigue. At high flow water i s able to transport larger particles due to the high ratio of horizontal velocity to stokes settling velocity But discharge appears to have a different effect during rise and recession period of a storm event. This effect may be interpreted as a resuspension phenomenon Large particles should sediment in the aquifer during recession periods When significant rainfall occurs, the water flux and velocity swiftly increase and generate turbulence in the previously quiet zones This turbulence may resuspend the large particles, which can then remain in suspension and be transported if velocities are sufficient in the whole system As the large particles are washed out their content at the spring quickly decreases. 126 Pr oc eed i n gs o f the 1 2 "" Internat i ona l Congre s s o f Speleolog y, 1997 S wi tzer l and V ol ume 2

PAGE 138

E ::1. "' ci z C) 0 ...J 6.8 6.6 6.4 6.2 6 5.8 5.6 7.3 I I I I 7.4 I I I I I I I I , ,, I 7.5 7.6 pH I I I I I 7.7 7.8 Fig. 2: Relation water and number between pH of the of small particles Noiraigue The number of small particles (i.e., the first part of the curve) is independent of discharge but clearly decreases with increasing pH (Fig 2) Particle number is an order of magnitude higher at pH 7.4 than 7 .8. This pH effect can be interpreted as a coagulation effect. In fact the aggregation properties of clays or other colloids is highly dependent on pH. When pH is close to the zero point of charge (zpc), the surface charge decreases and colloids can form larger aggregates diminishing the number of small particles. However, it is difficult to predict the effect of this process on a PSD curve since no model of aggregation has been successfully applied to natural PSD. Phyllosilicates Mica Smectite+I/S lll8 Kaolinite Paragonite 1111 Chlorite Serpentine 0 20 40 60 80 100 % Fig. 3: Mineralogy of boreholes drilled on surface sediments (a: Molasse, b: Tills, c: Marls) and suspensions from Bied brook and Noiraigue spring (lines show different sampling campaigns). Pl,yllosilicates are presented in proportions of total phyllosilicate. The role of environmental factors on the Bied brook PSD curves is more difficult to describe. The effect of discharge on each parameter seems to be negligible The only dependence found is the effect of temperature on the number of small particles. This could be explained by the influence of temperature on the global biological activity and particularly on the bacterial growth. Again data sampled during the steep rising of the hydrograph form a separate population with higher particle numbers for the same temperature. 5. Nature of transported particles The s uspensions from the Bied brook and the Noiraigue river contained mainly quartz, calcite and dolomite grain minerals and mica, chlorite and kaolinite phyllosilicates. Both locations show similar composition of the phyllosilicate fraction, i.e. ea. 50 % of mica, ea. 25% of kaolinite and ea. 25% of chlorite, phyllosilicates also existing in the quaternary deposits. However smecrite and mixed-l ayers, well represented in the deposits are absent from the sampled waters (Fig 3). Suspensions are also enriched in feldspars relative to deposits. The content of calcite relative to quartz is higher in the Noiraigue waters than in Bied ones. On the size fractionated samples, the sedimented fraction contains only quartz, calcite, dolomite and sometime feldspars. 60 10 Mixed 50 Clays m "Quartz' 5 I Fe-Ca Partic .' 40 m 30 m 'Iii 10 20 20 ... 10 m m m m mmm m m m mm 20 40 60 80 100 120 Si 16-,-----------,----.-------, 12 .. . <1S 8 (.) \ I I I It 'm ,_ .mm:, I ... m i I 4 ,m.,._',I, l. In ,~ :1 ::n. f~ / I , Ill 1 .} l U I "'111111:S...~'11 I~ I ;~,,m.,... I I la I .l i"1, 1 -o --....--....---,---,-, ..--~..--,--.-, .--...-,...-,--,...-,----....-,...-, ---; 0 4 8 12 16 Fe Fig. 4: SEM-EDS intensities (relative units) measured 011 particles from the Bied brook: a) Si and Al, b) Fe and Ca 6 '" Conference on Umestone Hydrology and Fissured Media 127

PAGE 139

Ultracentri fuged fractions show phyllosilicates, quartz dolomite in some samples, and scarce and small peaks of feldspars. Quartz peaks decrease with decreasing size of particles, and dolomite peaks almost disappear down to ea. 0 65 Mica, kaolinite and chlorite are not significantly differentiated by size, as opposed to mixed-layer clays and talc which appear only at small particle size. Mixed-layer clays were detected only after fractionation : they might have been hidden by larger particles on raw samples. In the SEM-EDS analysis the main difference between Bied and Noiraigue are the Fe intensities being more than twice higher in the Bied brook The use of Al and Si intensities allowed the differentiation of four groups of particles : Clays Quartz, Fe-Ca particles and mixed particles (Fig 4) All Bied particles have the same range of Ca and Fe intensities and the same Ca/Fe ratio Beside if particles are differentiated according to campaign number, a clear Ca/Fe correlation appears independently of the particle types At Noiraigue the particles present very small amounts of Fe and thus no correlation between Fe and Ca content has been established It has been established that the numerous Ca-Fe particles encountered at the Bied mainly consist of organic matter covered by Ca and Fe It seems that these particles come from organic matter network existing in peats on which iron precipitates when the media becomes oxidizing. During this process the organic matter network is also condensed The ratio of the bacteria over total number of particle varies from I to I 00 % at the Bied with a median at 30 %. At Noiraigue, the figures are completely different as the ratio varies from Oto 22 % with a median at 2 %. One major part of the variability in number of bacteria can be related to temperature effects. This effect can easily be understood due to the general influence of temperature on Jiving organisms development. It seems to be the only measured environmental factor playing a significant role on number of bacteria, while the total number of particles is not influenced by thi s parameter. Approximately I 0 % of the total bacteria were identified by cultivation methods Within these 10 % the large majority are heterotrophic while only a negligible proportion corresponds to enterococci. 6 Discussion and Conc l us i on This first characterization of the particulate and colloidal matter of a karstic aquifer underlines the complexity of such media By modelling the PSD curves it has been possible to clearly distinguish two parts of the PSD curves The large particles behaviour seems to be governed by physical factor : the shape of the second part of the curve and the largest particle size are highly dependent on discharge However this effect is not unique as the particle content is much higher in the steep rise of the event than in the recession part We are presently working on turbulence models in order to analyse the resuspension effects in the complex karstic geometry It appears that a detailed analysis of PSD evolution with discharge could be a powerful tool to characterize flow in a n unknown network The small particles, below 4-5 are mainly dependent on chemical conditions. This effect i s attributed to surface charge an important characteristics of colloids in aqueous media That is why we think that the 4-5 m size is the limit between colloids and particles in the studied aquifer. The surprisingly large size of colloids ari s es from high water velocity allowing transport over long distances and turbulent flow enhancing the aggregation process The particle composition revealed a majority of clay s as in many natural systems These clays, stable in low temperature environments are not modified during their transport. We have few means to detail the evolution of the composition of their surface The Ca-Fe particles, already described in lakes are of major intere s t a s they seem to be the prominent passways for Fe and organic matter in the environment. They also seem to be retained in the aquifer despite the very small change in pH between the Bied sinkhole and the Noiraigue spring These particles reveal the complexity of colloidal matter for which the interaction of physical and chemical conditions are crucial. The s e particles might be modified by ageing process typical of colloidal mixtures or by turbulent flow driven aggregation Bacterial populations were difficult to study as most of the enumerated bacteria are unknown their transport seems to be quite different from the one of the other particles The role of colloids in biogeochemical cycles is evident for the elements s uch as Al Si incorporated in clays but, a s determined above also for Fe and organic matter. Beside, the importance of colloids in contaminant transport can be extrapolated from this s!Udy. Particle counting s howed the large number of particles present in natural samples and electron microscopy evidences the tremendous surface areas free for contaminant adsorption However, we showed a bove that the particle s were complex and often made of mixtures of several particle types The main chemical colloidal characteristics, namely adsorption capacity and surface charge must therefore be determined directly on natural samples. References ATTEIA 0 & R KOZEL ( 1997) Particle s ize distribution s in waters from a karstic a quifer: from particles to colloid s. A c cepted b y J. Hydrol. BUFFLE J. & G.G LEPPARD (1995 ) Characterization of a quatic colloids and macromolecules. I Structure and behaviour of colloidal material. En v ironmental S c i. and T ec h 29 n 9 : 2169 2175 BUFFLE J. D PERRET & M NEWMANN (1992 ). The use of filtration and ultrafiltration for size fractionation of aquatic particles, colloids and macromolecules In : J Buffle and H P van Leewen ( Eds), Environmental particles volume I. IUPAC Series on environmental analytical and physical chemistry Lewis Publisher. 171-230. DOERFLIGER, N & F. ZWAHLEN (1994 ). Vulnerabilite des aquiferes karstiques : principes de la methode EPIK et presentation d un cas d'application Conference on Karst Fractured Aquifers Vulnerability and Sustainabilit y. Katowic e, JI14 septembre 1995 ., JANTSCHIK R ., F. NYFFELER & O F X DONA RD ( I 992 ). Marine particle size measurement with a stream-scanning laser s y s tem Mar. G e o/. 106 : 239-250 KIRALY, L. ( 1988) Large scale 3-D groundwater flow modelling in highly heierogeneous geologic medium Jn : E. Cu s todio and a others ( Eds), Groundwater Flow and Quality Modelling Reidel Publishing Company 761-775 : LERMAN A ( 1979) Geochemical processes in water and sediment environments John Wiley and Sons MATTHEY Y ( 1986). Etude phytosociologique du complexe de tourbieres du Bois-des-Lattes. Bull Soc. Neuchareloise Sci Nat. ( Suisse) 109 : 137-145 MCDOWELL-BOYER L.M J.R. HUNT & N SITAR (1986 ) Particle transport through porous media Water R e sour Res ., 22 : 1901-1921. 128 Pr oceedin g s of t h e 12 '" In t e rnation a l C o ng r e ss o f Spe leology 1997 Switzerlan d Volume 2

PAGE 140

Temporal variations of spr i ng water temperatures in relation to the extents of the heat transport modes occuring in the karstified lower Gypsum-Keuper aquifer (Karn i an southern Germany) Jochen Bundschuh, Technical University ofDannstadt, Institute of Geology Schnittspahnstraf3e 9 64287 Dannstadt, Germany present address: University of Salta, Institute of Hydrogeology INASLA Buenos Aires 177 4400 Salta, Argentina A b stract The importance of different heat transport modes was studied b y anal ys is of the t e mporal s equence s of spring water tempera tures in a karstic aquifer. For the aquifer with a horizontal groundwater flow and flow velocitie s of more than I 00 m yea( 1 model calculations indicate that horizontal heat transport in the aquifer occur s predominantl y b y convection. For v ertical heat transport from the earth's surface, conduction and convection are implicated in varying degree s: in order to indicate these the temporal se ries of the spring water temperatures were separated into their conductivel y and con v ecti v el y component s. Generally this can yield valuable information about the properties of thin surface-near aquifers. Thu s, the quantification of a heat component trans ported vertically by convection into the aquifer can provide evidence of h y draulic conducti v ities as potential infiltration route s for contaminants Z u s amm e nfa ss un g Die Bedeutung verschiede n er W!irmetransportmechanismen wurde in einem verkarsteten Aquifer durch Zeitreihenanal y se von Quellwassertemperaturen untersucht. FUr den Aquifer mit horizontaler Grundwa s serstromung und Flie8geschwindigkeiten von Uber 100 m Jahr 1 haben Modellrechnungen gezeigt, da8 in dem Aquifer der horizontale W!irmetran s port ilberwiegend kon duktiv erfolgt. Am vertikalen Wfirmetransport von der Erdoberfltiche her. sind Konduktion und Konvektion zu verschiedenen Ma8en beteiligt ; um diese zu bestimmen wurden die Zeitreihen der Quellwas s ertemperaturen in ihre konduktiv und konvektiv erzeugten Komponenten aufgetrennt. Generell erlaubt dies Hinweise auf Eigenschaften geringmllchtiger oberfl!ichennaher Aquif ere ; so kann die Quantifizierung einer konvektiven Yertikalkomponente Hinweise auf mogliche lnfiltrationswege fur Schadstoffe liefem 1. Introduction Climatic changes on the earth's surface. e s peciall y changes through time in temperature and precipitation. influ ence u n derground temperature in a comp l ex manner b y means of conductive, convective and dispersive heat tran s port In addition to those alread y mentioned parameter s and cau s e s influencing heat transport (and also underground tempera ture) are as follows: depth-to-water table thickness and ar ticulation ( h ydraulic conductivity) of the l ayers overl y ing the aquifer materia ll y dependent parameters and parameter functions of aquifer components (heat conductivity heat ca pacity, density. viscosity, etc.) groundwater recharge geo logical hydrogeological and hydrological data pertaining to the aquifer and its environment. Input factors such as precipitation the earth' s surface temperature, etc are known to be involved as are too temporal series of groundwater temperatures as output factor s. Un known however is the extent to which the various heat transpo r t modes are involved in overall heat tran s port from the earth's surface into the aquifer (thereby altering the latter's temperature) Such knowledge might provide important in sights for solving hydrogeological questions. Transfer of thermal energy in aquifers involves several. reciprocally linked transport mechanisms (HOLMANN 1976 IBELE 1 963, JISCHA 1 982 KAY & NEDDERMAN 1974 KNUDSE N & KATZ I 958) Co n d u ctive transport along an existent tem perature gradient, is found both in the grain skeleton and in fluids Heat transport through moving groundwater can be described by a convective and a dispersive term In convective transport heat propagation in a fluid occu r s with an average interstitia l velocity. The transport mechanism of dispersion comprises all processes that can be att r ibuted to the effective velocity deviating from the average interstitial velocity The various heat transport mechanisms dominate under different hydrogeological conditions (flow velocities disper sivities, t h ermally open or closed system etc.) Their role in overall heat trans p ort was quan t itative l y determined by the a u thor ( Bu ndschuh, 1 993). He fo u nd that for one-d i mensional heat propagation in a I O m thick, horizontally percolated sand aquifer which wa s thermally insulated both above and below and for average interstitial velocities of up to I m year onl y conduction i s significantly implicated while for more than I 00 m y ea( 1 convection accounts for 90 % of heat tran s port. For a v erage inter s titial v elocities of over 10 m year 1 the dis p e rsive component i s le ss than 1 0 % for a longitudinal dis per s ivity of I O m If the horizontally percolated aquifer is not thermally insulated above and below and restrained by imper meable layers significant amounts of heat can be removed from or fed into the aquifer via these layers by means of vertically oriented conduction The importance of the different heat transport modes was s tudied b y anal y si s and interpretation of the temporal se quences of spring water temperatures and other hydrogeologi cal and hydrochemical parameters of a partially karsted g y sum aquifer in the G!iu region (southern Germany) 2. Field examples Area description and basic investigations Morphologically the G!iu comprises the broad floodplain expanses formed by mollusk bearing limestone (Muschelkalk mo). loam keuper (Lettenkeuper ku) and lower gypsum Keuper (Gipskeuper. km 1 ) in the approaches to the south western Swabian Alb ; while the younger gyp s um-Keuper form s the rise to the Keuper mountains. The 18 m thick loam Keuper (ku) con s i s ts of a series of water-retaining clay marls and clefted. water-conducting sandstones and dolomites ( s pring discharge up to 3 I s 1 ) The overlying unleac h ed I 00110 m thick gypsum Keuper can be subdivided in terms of its hydrogeological properties. The l ower part which constitutes the principal groundwater conductor is formed of20 m thick main gypsum (Grundgips) layers. which consist to 60-65 % of gypsum / anhydrite (spring discharge up to 50 I s1 ) In the up per part, which consists predominantly of relatively thin. clefted dolomite marl banks a n d maximally I m thick gyp sum / anhydrite deposits form groundwate r conductors of lim6 '" C onferen c e on Li m e s tone Hydro logy and Fi ssu red M e d ia 1 29

PAGE 141

59 53 58 57 56 SS 54 53 52 Schilfsondstein to Lias (km2-l o< ) Gipskeuper (km 1 ) : middle and upper Grundglpsschichten (GG) (oufcrops of gypsum) Lettenkeuper (ku) and Muschelkolk (mo) 34 69 70 71 34 74 75 76 77 \ ,, 78 .......... '\ \ 79 80 58 57 towns: Bf = Bergfelden, Bo = Bochingen, Em = Empfingen, F i = Fischingen, Gr = ____ fault Hz = Heil ig enz im mern, Rf = Renfrizhousen Su = Sulz, Tu = Tubingen, Vo = Vohringen springs : Hungerbrunnen Urenbrunnen A Goringquelle Mostkrugquelle D Reinerlquelle Fig I: The area under investigation (Gliu between Fisc!,i11ge11 am/ Viiltringe11): Geological layer organhatio11, layer packing, l,ydrogeology, and location of springs used as examples. ited significance (spring discharge up to 4 I s1 ). Karst formation indicators (episodic springs dry valleys, stream infiltration sink holes, etc .) are predominantly re stricted to the main gypsum levels (GGS) since the gypsum beds in higher levels only rarely attain a thickness of I m The discharge coefficient of the springs studied ranged from 0.2 to 0.6, i.e. at times significantly above those of the carbonate karst (mostly around 0.1). Springs whose GGS aq uifers are not superposed by higher Keuper layers have dis charge coefficients of 0 2 while values of 0 5 are obtained in corresponding cases of superposition of aquifers. Semi-quantitative laboratory and field studies on the solubility of gypsum showed that mineralization of precipi tated water occurs very rapidly (from within hours to a few days) upon penetration into the underground Balanced be havior through time by electrical conductivity does not therefore necessarily suggest long residence times Determination of the outflux of geo~enic sulfate yielded values between 220 and 360 t 1an 2 year, which corresponds to average leaching values for gypsum of 0 10 and 0 17 mn year1 respectively. Spring water temperatures: an overview From the temporal series of groundand spring water tem peratures after correction for trends the annual periodic component was isolated by applying Fourier analysis. This share predominantly reflects the slow process of conduction, through which periodic temperature changes on the earth's surface can be vertically transported down into the aquifer and after a time lag transfered to it in mitigated form. The remaining component comprises shorter-lived temperature changes that are due to convectively vertically injected heat resulting from discrete precipitation events. The respective components ob tained for the different field examples are graphically depicted in Fig. 2 Applying as a differentiation criterion the changes in aq uifer and spring water temperatures as a result of heat input from periodic climatic variations and short-lived precipitation events, for the springs studied in G!iu the following cases can be distinguished: Only short -li ved discrete events with no discernible periodicity neither discrete events nor periodicity discernible periodic changes and discrete events occur only periodic changes occur Each of these cases will be described in detail by a representa tive example (Fig. 2) Example 1: Only discrete events occur Example of this spring type: Urenbrunnen/VOhringen. Temperature graph: It does not contain any unambigu ous periodic components. Changes have their origins in short lived events with only strong discrete events such as pro longed heavy precipitation or thawing episodes exercising a pronounced influence on the temperature of spring water. This produces temperature anomalies for the months of January through March, while as a result of low groundwater re charge a constant temperature progression is mostly onl y found during the rest of the year. Thus if we resolve the spring water temperature into a Fourier series we obtain for the periodic components an amplitude of 0 3 K (which is sur130 Proceedings of the 12"' Internat i onal Congress of Speleo logy, 1997, Switzerland Volume 2

PAGE 142

N SW N ... (j) 0 560 540 ,,o 540 540 ~10 520 520 500 500 500 ..., 400 400 00 111,0 460 RW : l 7 800 1750 RW: 1, 7 ns RW: 34 76 6$0 lfW: l 77 4 0 HW : S3 SJ 075 HW: "5 8$0 HW: Sl 51 420 HW: 5l 57 000 11 11 t1mp1ratur1 of spring water 12 10 10 11 To(t)fCJ 11 10 10 spring woftr : sum of overage temperature and annual variation 10 10 11 spring water: sum of overage temperature and annual Yorlatlon k==---r:'.,M=!,:':,.~..,~.~=--=i:m==--"""=~,~-~.v",~.:-, -:::,.:-,l 101--------~-------------------110 f"vi 10 0J c / ..... tu ..,. t;to. r.e,01 ""Ii """ cH(M 1 ti .,~., "' > ..... . llM...,.. J.J/l'1.0U T~t)f+ 8 1 (C) 4 y 2 -.tJJ1' 1 wt(w t 6t'l p 1 ) (t<] temperolurt of spr ing wot1r : variation r om th annual periodic +I function .T(t)=T 0 (t)-T 1 (t) [K] {K] ftmperoture of spring 'WOler : variation from the annual ptriodic I +I luncf;on .T(t)=T 0 (t)-T 1 (t) [K] +I 1 RW: J4 71 IIO HW : 5-J 61 GOO 11 temperalur of spring wafr 10 T0(1l["C) 10 spring wolr. sum ol 011rog ... .. Ull' 1 :i .. [K) temperature ol spring water : I function .T(l)=T0(t)-T 1 (t) -I 20 discharge of spring 16 (t)[1 .. J 12 40 20 -1 I electrica l conductiv i ty LF(Q[S1cm] -I ,000 2&00 1--------U(-'1"'"1--------~-------------!"' 17!10 tmperotur end annual AW: 34. 79 820 HW : 5J 59 la7 420 11 tmprature of spring wolr ~----i 10 10 T0(()["C) variation 10 10 fAy-1.!c 110-r+e, ( 1 "v1 .u1., c .. ( .. I 6 ....... ,, variation from fh annual period i c [K] +I +I -I -I electrical conductivity JOOO LF(Q [Siem] 2100 750 AW : 34. 77 94.0 HW : 53 58 54.0 ,.,,. 2200 440 ,20 11 10 Fig. 2: Spring examples 1 to 4: Cross-section through the main inflow area of the springs (above). The three diagrams set out below describe temporal temperature changes in the spri11g water: represented are: the measurement series as discrete values of the actual temperature progression TQ(t); the primary (anual) harmonic 81 (periodic change with amplitude AMPJ) ob tained by Fourier analysis from TQ(t), representing (together with t!,e mean annual temperature 1) t!,e measurement values of periodic progressio11 T 1 (t) (L1(f)Q indicates the phase differe11ce between earth's surface and spring water temperature; L1(f)L the phase difference of t!,e eartl, 's surface temperature referred to January I st) ; and t!,e deviations tiT(t) of t!,e measurement values from the periodic progression T1(t). These describe the temperature progression freed of periodic behavior, resulting from s!,ort-lived events. Both lowermost diagrams show spring disc!,arge Q(t) and electrical conductivity LF(t) for spring water and for local precipitation N(t). 6 '" Conference on Limestone Hydrology and Rssured Media 131

PAGE 143

prisingly high for such a case). The scant phase displacement of this primary harmonic of only 24 days is a further proof that its minimum is induced by corresponding negative temperature anomalies as a result of discrete events. The snow thaw of December 17 1988 led to a sinking of the aquifer tem perature, with the minimum being attained on Jauray 14 at the same time as the discharge masimum. It should be noted that the relevant minimum of conductivity was already attained on December 19 In ter pre t at io n : This aquifer of mostly constant heat con tent (constant temperature) is primarily convectively fed with heat during the snow thaw and extreme summer thunderstorms. which leads to a temperature change in the groundwater (after a short period. however. this reverts to its original value). Groundwater surfacing in Urenbrunnen onl) needs 2 to 3 days to attain its conductivity minimum. However. the dis charge maximum and the temperature maximum/minimum are only attained after 10-20 days The reason for the large timelag between the conductivity minimum and the temperature mini mum (and the spring discharge maximwn) is the excellent solubility of gypsum and resultant mineralization of water infiltrated during the snow thaw (which takes no more than a few days to complete). In wide areas of the predominantly south-situated catchment area, in which the karstified main gypsum partly reaches to the surface. permitting rapid infiltra tion to occur, the groundwater depth-to-water table is more than 10 m At this dept h the conductive l y transported tem perature wave is for all practical purposes no longer discerni ble. Example 2: Neither discrete events nor periodic ity discernible Exa mpl es of t h is s pr i n g ty p e: Reinertquelle (Q3 I). Hungerbrunnen/Renfrizhausen (Q34 ). Te mp e r a ture g raph s: Both these springs from the main gypsum aquifer are characterized by constant water tempera tures ranging from 10 3 C (Q34) to I 0.0 C (Q3 l) 0.06 K. In contrast to the previous example, discrete events do not result in temperature change. nor do periodic variations occur. Likewise severe events such as the snow thaw of December I 7 1988 prod u ce no discernible effect here Temperatures respec tively lie at 2 1 (Q34) and 1.8 K (Q31) above the annual mean air temperature. In both springs, the electrical conductivity fluctuated by a maximum of 4 % (referenced against their mean value). Inte r p r e t a ti on : These springs from the middle cross section of the main gypsum layer are fed with water from the mountain side, i.e coming from an easterly direction There the main gypsum layer is superposed by predominantly clayey marl layers of higher -l ying gypsum Keuper which prevent any rapid, convective heat input due to infiltrating seepage water. Since superposition of the aquifer already reaches more than 10 m in close proximity to the localized seep. conductive heat transport due to periodica ll y fluctuating air temperatures is no longer perceptible. Instead, there is a temperature in crease of around 1.8 to 2 K. coming from the lower part of the aquifer and caused by vertical conductive heat transport from the earth's interior. Under quasi-stationary conditions, this temperature increase corresponds to a mean aquifer superposi tion of 20 to 40 m Neither convective nor conductive heat transport imposes temperature changes on the aquifer. Depend ing on depth, a heat flow emanating from the earth's interior leads to raised tempe r atures that produce spring temperatures of 10 10.5 C. Example 3: Only periodic changes Ex ampl e o f thi s s prin g ty pe: Mostkrugquelle. Te mpe rature g r a ph s: It only exhibits periodic changes, not the effects of short-lived discrete events. Likewise a severe event like the snow thaw of December 17 1986 exerted no discernible influence on spring water temperature. This indicates that Mostkrugquelle draws on a north situated catch ment area, with none of its water directly deriving from the area of the main gypsum layer not superposed by younger layers. This is also confirmed by the electrical conductivity of the spring water, which in the period studied deviated from its mean value by a maximum of 8 %. The annual mean of the spring water temperature is approx I K above the mean annual air temperature In ter p re t ation: This temperature behavior is caused by layers of low hydraulic permeability in the vadose zone of the main gypsum aquifer. which prevent infiltration (convection) into the latter. The aquifer is affected by periodic temperature changes due to conductive, vertical heat transport. Example 4 : Periodic and discrete events occur Examp l es of t hi s s pr i n g t y p e: Goringquelle Te mp erat u re grap h: At Goringquelle, changes occur in spring water temperature due to both periodic and short lived events. Hence at Goringquelle the above-mentioned snow thaw of December 17 1986 resulted in the spring water tem perature being lowered by 1.0 K, with the minimum being reached on December 1 9 1986 concurrently with the associ ated discharge maximum and the conductivity minimum. I nter pr e t a t i on: The spring surfaces in the margina l dolo mites of the loam-Keuper obtain a part of their water from the overlying karstified main gypsum layers. This makes rapid infiltration into the marginal dolomites possible duri n g flood conditions. which explains why this spring reacts so rapidly to discrete events. This rapid, convective heat transport is overlain by a conductive heat transport which imparts to the aquifer periodic variations in the earth's su r face temperature. 3. Results The field examples studied are characterized by m ore or less horizontally percolated aquifers. The flow velocity is more than I 00 m year'. According to the results of numerical model calcu l ations this means that for horizontal heat t r ans port in the aquifer the convective components heavily outstrip the conductive ones. For vertical heat transport from the earth's surface. on the contrary conduction and convection are implicated in varying degrees. Springs with only periodic temperature components (example 3) only manifest conductive vertical heat transport. Example I, which involves exclusively short-lived temperature variations, points to predominantly convective, vertical heat transport. In Example 4, on the other hand conduction and convection are superposed; as a result. the periodic progression of spring water temperature is over lain by short-term variations. Separating the temporal series of ground and spring water temperatures into their conductively and convectively created components can yield valuab l e informatio n concerning the properties of surface near aquifers of limited thickness. Thus for example. demonstration of a heat component transported vertically and convectively into the aquifer can provi d e evi dence of hydraulic conductivities as potential infiltration routes for contaminants 4. Bibliography B UNDSCHUH J I 993. Modeling annual variations of spring and groundwater temperatures associated with shallow aquifer systems. J. Hydro/ 1 42: 427-444. HOLMAN. J. P 1976. Heat transfer. McGraw-Hill Book Comp .. ew York. IBELE, W (Ed) 1963. Modem developments in heat transfer. Academic Press, New York. JISCHA, M. 1982. Konvektiver Impuls Wlirrne und Stoffaus tausch. Vieweg Verlag: Braunschweig Wiesbaden, 367 pp. KAY J M. & NEDDERMAN R. M. 1 974. An introduction to fluid mechanics and heat transfer. Cambridge University Press: Cambridge. KNuoSEN, J. G. & KATZ, D. L. 1958. Fluid dynamics and heat transfer. McGraw-Hill Book Comp., New York. 132 Proceedings of the 12 1 International Congress of Speleology 1997, Switzerland Volume 2

PAGE 144

Results of a study about tracing tests transfer functions variability in karst environment. par Nathalie Doerfliger Centred' Hydrogeologie, Universite de Neuchatel, I I rue E-Argand, 2007 Neuchatel, Switzerland Abstract Artificial tracing tests are often used to simulate migration of a point-source contaminant under various hydrological conditions in karst hydro geological impact assessment or to define groundwater protection zones. Due to economic reasons, it is rather difficuh to carry out adequate tracing tests to determine what are the possible recovery curves over range of discharges at the outlet, are the lracer test results representat i ve of the spring watercatchment being protected ? Our objective was to characterize the !racing systems in a karst environment by a mean transfer function; such transfer function may be used to predict the breakthrough curve of a point-source contamin a nt taking into account an error factor. A Jura mean transfer function with + and 95% interval confidence functions can be established and differentiated from the Alps mean 1ransfer function. The use of this transfer function to predict the response of a point source contaminant requires considerations of water catchment size, thickness or the aquifer and discharge at the outlet. The results of this variability analysis confirm that the transfer functions by themselves may not be used to protect the whole karst spring water catchment, as this one is affected by the heterogeneity of the physical parameters At the scale of a water catchment, transfer functions are not the major tool to protect the groundwater. But with a multiattribute approach of vulnerability mapping, transfer functions contribute to the development of groundwater protection strategy Resume Les essais de trac;:age sont souvent utilis e s dans le cadre d etude d impact ou de delimitation de zones de protection afm de simuler la migration d'unc contamination ponctuellc et ccci pour des conditions hydrologiques differentes Pour des questions cconomiques, ii n est pas po s sible d effectuer le nombre d'essais de trac;:age suffisant pour d e terminer la variation des courbes de restitution pour des conditions de d e bit differentes et connaitre la courbe de restitution caracteristique de la source du bassin a proteger. Notre objectif est ainsi de caracteriser les systemes-trac;:ages en milieu karstiquc par une fonction de 1ransfert moyenne; une telle fonction pcut crre utilisee pour predire la courbe de restitution d un contaminant d'une source ponctuelle prenant en consideration un facteur d'erreur. 1. Introduction Artificial tracing tests are often used to simulate migration of a point-source contaminant under various hydrological conditions in karst hydrogeolog i cal impact assessment or to define groundwater protection zones Due to economic reasons, it is rather difficult to carry out adequate tracing tests to determine what are the possible recovery curves over range of discharge at the outlet. Our objective was to characterize the tracing-systems in a karst environment by a mean transfer function If transfer functions can be readily obtained from i nput-response models, can they be used to predict contamination migration for part or all of karst water catchments ? Can we differentiate a a-ansfer function developed for a-acer tests in the Alps from those derived from tests in the Jura ? Do hydrological conditions affect the transfer functions ? To study the variability of transfer functions, the results of 96 tracer experiments have been used 14 were done in the Swiss Alps, 12 in the Folded Jura and 70 in the Tabular Jura (D0ERR..IGER, I 996) 2. Methodology Interpreting tracing tests in terms of transfer functions is based on having an input-response conceptual model. The real system, the karst aquifer or the tracing system, is a closed box with an ena-an c e and an exit. The internal sa-ucture is not determined as in a deterministic mod e l. Transfer fun c tions can be obtained from tracer breakthrough curves and a black-box input respon s e model. It can also be dir e ctly obtained, without any mathematical treatment. from a breakthrough curve if 1) the tracer input is instantaneous. simulating a Dirac and 2) there is a continuous record of the discharge at the sampled spring. When these conditions exist or are assumed to be a-ue, the normalized breakthrough curve is a transfer function referred to a Rm (Residence Time Disa-ibution). In practice, however, instantaneous injection of tracers is not possible for several technical reasons Consequently, the directly observed response of an injection at the outlet does not correspond to the impulse response. In that case, the impulse response has to be obtained by a mathematical "treatments" referred as deconvolution. Input-response model : deconvolution operator Input-response model or black-box model permits analysis of the karst system or sub-systems, the tracing system Considering a linear and steady-state tracing-system, even if in the real world, hydrogeological systems do not only response in a linear manner and their behaviour is not consistent in time with varying changes, for example in flow rates, transfer functions are obtained using a black-box model of deconvolution (DZIK0WSKJ, 1992; GUIZER.IX, 1988). The output function is deconvoluted by the input function; transfer function is calculated using the numerical code DYEBOX, a global mass transport FORTRAN 77 code (KIRALY & R0SSIER, 1992) The discretisation of the deconvolution can be written as follows : l(j)= 1 I S(j) I l(j-k} E(k}j E(O)l_ ~t k =1 ( 1) with j from O 10 N; k a variable differenJ than O; tff is the choosen discretisation time step 6 '" Conference on Limestone H ydrology and Fi ssured M edia 1 33

PAGE 145

Several 1ypes of lransfor functions (figure 1) can be described: the Impulse response corresponds to the syslem responding lo an impulse wilh Dirac condi1ions. The resultanl unil response is independent of the injection duration. the normalized response is the system responding to a step function (or Heaviside function). ll is the results of the convolution of the impulse response with a step function. This type of response is used in order to compare transfer functions to each other. In this situation, the normalized response is equivalent 10 a statistical distribution function and the derived normalized response directly defines, the impulse response For example, the time corresponding to the 0.5 percentile is the time at which 50% of the tracer has passed through the system (figure 2) the rererence response (KIRALY & Ross1ER, 1992) is equal to the system responding to a "reference impulse" which has been or can be charactzerised with an arbitrary duration and shape. The reference response corresponds to the tracer travel time (RTD) of a tracing tesls. This response is opposite to an impulse response, in that it includes the duration to the input. C 0 .:; co L. C Q) () C 0 Impulse response () __ .,,.,__...:,,, __ ,. time time Ref ere nee response time Figure I: Various types of transfer functions: response of a system in relation to the type of input function 0 003 0 0025 i 0 002 l 0 0015 E ;: 0 001 0 0005 0 ls 0.8 = (I! 0 6 C 8 0.4 i 0 2 a: 0 0 0 l1Tpulse response 1 ntensly (derivaie gradienl al 0 5) 1enslty (de
PAGE 146

Variance analysls A variance analysis was carried out on the modal velocity and the 0.5 percentile gradient 10 compare the following conditions pertinent to each tracer test: geological setting, hydrodynamic conditions, nature of injection point, catchment size, the tracer distance and the outlet discharge. These conditions were compared to determine if their man values (gradient at 0.5 and modal velocity) were significantly different. A single factor variance analysis (ANOV A) was carried out on each group lo determine if the categories of each conditions influenced the transfer functions Multiple factor variance analysis (MANOVA) was also carried out: in theses cases, two or more variations in each of the tests conditions were simultaneously testsed to determine if either cause significant differences between treated parameters means. Each parameter is assumed to have an normal distribution around its mean; all parameters have the same variance This analysis allows calculation of a whole model test considering the experimental values and the predicted ones and taking into consideration influences of each considered group. The significant difference is related to the results of the effect test: a probability > F equal to 0 05 or lower than 0.05; this value is the significance level a 3. Transfer functions varlablllty Resulls of the single factor variance analysis arc summarised in the table I There is no highly significant difference among any of the conditions and their categories for the normalized 0.5 percentile variance analysis For other relationships, about half of the differences are moderate and the other half of low significance. The low significance is related principally lo geological selling, hydrological conditions and injection point conditions, where as moderate significance characterizes the catchment size. tracer distance and discharge rate conditions. The hydrological conditions provoke no major difference within each class of the geological context, contrary to the low degree difference among all the tracer tests between base flow conditions and rising limb conditions Generally, the gradient at 0.5 percentile which represents the mean values, for significant class of considered parameters, varies between 65 with a slope of 0.46 to 88.4 degrees. with a slope of 0 028 The slope at the percentile is calculated as the tangent of ~. where = 90gradient. PARAMETERS Geological Mydrologocal Inject ion nature Su,1ace ol Tracing distance Discharge (V s) context ondil ions we basin (m) cE cE -f 0 ] :) 3: c'5 E :) -, ::; 0 0 ..., E E .. -, ::; 0 .c 0 0 0 .,, 0 0 0 0 .. .. ,.. ,.. (:)" a a: 2 .c C> .,, 0 0 .. .. ,.. ,.. 'I,!) :s l 0 ..., .., .., ,.. ,.. 0 0 ...J UJ :::; .!l .2 0 ;g ..., ..., ,._ 0 0 0 0 0 c( iil .c f 0 0 0 ...J 0 ] f 0 8 8 8 8 b b c( co a: i J, :;i ..., .2 8 0 ..., .,, ., ..... >< R X X X 0 0 0 0 t_ X XX IT XX x~ xx 0 XX 0 00 0 X X Al ,_ L LJ I I u LU tr&cing L w I I-le&ts (Q6) 7U IH 10' II 7' 7'2 ll 1 &U 76' ll.> .... &2.& 7& 1 v 71 0 2< 0,, 0 16 0 11 0 25 0 2< 0 ,2 0 0 24 0 o o, 0 12 0 2, o / 0 0 0 0 0 f X X _J xx T XX ~x XX 0 Jt;X 0 00 0 X X Jura 1 l I L I I LJ (82) ll.> ......, ll & ... 7U ll > ,u ,. 7U &U U I 17 & 0.2 0 11 0" 0 2 0 1 o 0.2< 0 1' 0 0 0 12 0 D O o :~ r.fm1 0 0 0 0 0 0 0 0 0 0 / / X 0 x,x 0 00 00 0 Alps 0 2 (1') :) : .. ,u 0 <6 1'1. 0 xx / ./ XX / x~ ~x 0 0 0 0 / IY Foldo ,. ,. IJ.> 0 o ,6 o ,6 0 0 1 r O< O < : 0 0 0 / / X X l. LJ / X 00 I/ /,/ V Tabulat (70) 10 7 "' ... ,.. ... ~ .... IJ I n NI Jura :x ::::::~-:--:::-." 0 16 0 2 < 0 0~ 0 I 0 2< o o, 0 0 2 0 21 L~nd XXX a,gMc.,.ty diNuenl ~ 1.4>p,tr or to O 75 XX 1,19 nilcarcty dN.rnt l.4>pe,i' o, equ,al lo 0 25 and lower than O 75 X a,grdcanlty dl,Nerenl F\q., ~uel to O 1 and lower 1 han O 25 --g,oop of pamttn &ttuerc "om .,.,,t.t, L___J d.Netenc:e betwHn ths and thanl 0 ro s,9,-kant d1fle nrc al 5 mean vA of the vaditnl a. the s,.rctn1' 0 5 mean v lu tor 3000 10 5000 m mean vw tor upp.r 5000 m chtanct l 4 mHnYue olt hit 1.ng t,whtn b11 equ,el 10 90 o grlldi4rc o. Whengrad,trt t ends 10 90 tat'9!S,ltl"'h to O WMngr~rc otnd& lo ,s 0 1.ngt,ltndsto 1 ,,,,,,,,,.,,,,. no a-i ... ablt dalt Table I: Significant and non significant differences of gradient means at 0.5 percentiles. 5 ,h Conference on Umestone Hydrology and Fissured Media 1 35

PAGE 147

Among all tracer tests, a low significant difference in the geological context and in the hydrological conditions parameters appears. The mean characteristic values (modal velocity and gradients) or the transfor functions of tracer experiments permit to distinguish the Alps transfer functions from the Jura transfer functions. The mean transfer function of the Tabul ar Jura is not significantly different than for the Folded Jura (figure 3). Grad 0.5 By Geol cont 8 0 -ci 7 "' c'!, 70 65 60 5 ot ed jura alp tabular jura Gaol cont Figure 3: Variance analysis of the geological context for the 0.5 percentile gradient. Source Model Error Sum of squares 279 19 3443.74 Mean square 139 6 37 F ratio 3.77 Prob>F 0 0267 For the hydrological conditions, the base flow mean transfer function differs from both the falling limb (gradient) or the rising limb (modal velocity). The rising limb (gradient), the [alling limb (modal velocity) transfer functions are rather similar to the base flow mean transfer function. The difference between base flow and rising limb is not significant (Prob > F higher than 0 05) In the Alps, tracer distance characterizes transfer functions. Other conditions are important to classify the characteristic modal velocity of tracer tests. Major differences result from changes of injection point The injection point, tracer distance and catchment size are the only conditions whose classes cause differences in transfer functions characteristics. In the Jura, the injection point plays an important role in characterizing transfer functions In the Folded Jura, surface tracer test transfer functions are different from those in swallowholes. In the Tabular Jura, the main difference is between tracer tests carried out into boreholes and artificial holes and those with a direct sprinkling onto the surface. Transfer functions are influenced by the catchment size in the Tabular Jura. This influence is directly linked to the compounds of the tracer system and mainly to the ratio of the unsaturated to saturated subsystem Tracer distance is therefore also significant in the Tabular Jura and is also related to the tracer system composition. Finally, discharge is significant in both the Folded and the Tabular Jura when there is no major difference in hydrological conditions. 4. Mean transfer function From transfer functions variance analysis, a mean transfer function and a corresponding "95% confidence intervals" function can be determined for the Jura and the Alps geological seuings (figure 4 ). 1 -Alpes 0 9 C 0 8 0 7 f c 0.6 ----Alpes+ 95% Alpes 95% .. 0.5 u C 0 0 4 u -O-Jura .. 0 3 --Jura + 95% 0.2 oi 0 1 a: --Jura 95/o 0 0 5000 10000 Tlma (minutes) Figure 4: Calculated mean transfer J unctwns for the Alps and the Jura with the 95 % interval of confulence. Within the Alps context and in the Jura context., where significant differences in the variance analysis appeared, mean transfer functions can be calculated. This is true for the 5000m and lO'OOOm tracer distance parameters in the Alps, for the injection point, the catchment size and the tracer system hydraulic conductivity in the Tabular Jura 5. Conclusion and prospects Variations in transfer functions can be distincL On the economic purpose, we would like to determine mean transfer functions of a water catchment that is identified by some parameters such as its size, it mean transit distance, the thickness of the aquifer, the nature of the injection point, the discharge rate range,... But such functions do not exist No significant difference of the considered parameters is relevant. Nevertheless, mean transfer functions from tracing tests corresponding to a given geological context or to a given injection point can be differentiated with significance. The variability of the transfer functions within a water catchment is important and depends principally on the injection point, and to the proximity of the injection point to the karstic network. At the scale of a karst water catchment., tracing tests are not the single tool used to protect groundwater; a multiattribute approach of vulnerability mapping may be considered as well. References DoERFUGER N. 1996. Advances in karst groundwater protection strategy using tracing tests analysis and multiattribute vulnerability mapping (EPIK method). Thesis, University Neuchatel, Switzerland, 225pp. DZIKOWSKI M. 1992. L'analyse des systemes-trayages a debit variable et volume constanL Possiblit.es d'application en milieu karstique. These, Universite des Sciences et technologies de Lille, fevrier 1992, 182pp. GUIZERIX J.1988. A geometrical interpretation of tracer experiments in non steady (now, volume) systems. 4th conference on Radioisotope Application and Radiation Processing in Industry. Leipzig-G.D.R., sept. 1988. KIRALY L. & Ross!ER Y 1992. Interpretation quantitative des essais de trayage dans Jes aquiferes karstiques. Rapport inteme du CHYN, pour le SHGN/OFEFP. 136 Proceedings of the 12 th International Congress of Speleology 1997 Switzerland Volume 2

PAGE 148

Evolution piezometrique et qualitative d'un aquifere carbonate soumis a un fort pompage de longue duree : exemple du graben de Pinchinade (Mouans-Sartoux, 06) par Yves Guglielmi**, Christian Mangan*, Jacques Mudry** & Anne Reynaud** *Hydrogeologue conseil, 1 Bd Carabacel, 06 000, Nice. **Laboratoire de Geologie Structurale et Appliquee, place Leclerc, 25 000, Besan~on Abstract Pinchinade hydrogeologic unit (1 km 2 ) is a Jurassic calcareous graben lowered into Triassic dolomite, clay and gypsum. A long time pumping test in the graben lead to the study of the emptying of the aquifer and to the measurement of the impact of Triassic groundwaters, rich in sulfate. Specific yield has been calibrated from 950 000 m3 pumped (in a 3 years time) and it appears that the pumped volume exceeds recharge capacity of the aquifer from effective rainfall. This overexploitation causes a lowering of nitrate concentrations due to surface waters infiltration as well as an increase in sulfate concentrations due to Triassic waters. Resume L'unite hydrogeologique de Pinchinade (1 km 2 ) est constituee par un graben de calcaires jurassiques abaisses dans les formations de dolomies, argiles et gypses triasiques. Un pompage d'essai de longue duree pratique dans ce graben a perrnis d'etudier la vidange a long terme du reservoir jurassique et de mesurer son impact sur la realimentation par les eaux riches en sulfates du reservoir triasique. L'exhaure de 950 000 m3 (en 3 ans) permet d'apprecier le volume specifique et de dire que le prelevement excede Jes potentialites de realimentation de l'aquifere par la pluie efficace. Cette surexploitation se traduit par une diminution de la composante nitratee superficielle, en meme temps qu'une augmentation des eaux sulfatees calciques profondes en provenance du Trias. 1. Introduction La regie municipale des eaux de Mouans-Sartoux (Alpes-Maritimes, 06) exploite une petite structure aquifere constituee par un graben de calcaires jurassiques karstifies emboite dans des formations triasiques argileuses et gypsiferes. Les etudes engagees successivement depuis 1990 permettent une evaluation du volume de la ressource en eau disponible, des limites du bassin versant d'alimentation et des conditions du renouvellement de cette ressource. En particulier, nous avons evalue le risque de provoquer un drainage depuis l'aquifere triasique qui se traduirait, a tenne, par une augmentation des teneurs en sulfate des eaux. Les recherches reposent sur des essais de vidange prolongee (exhaure de 950 000 m3 en un peu plus de trois ans) a debit variable completes par des suivis piezometriques et chimiques de la nappe. La correlation du rabattement en fonction du volume pompe (COLLIGNON, 1986) est couplee a !'analyse des traceurs naturels de l'eau (MUDRY, 1987) pour estimer le risque de contamination triasique. 2. Contexte hydrogeologique L'aquifere exploite se situe dans les calcaires bajociens et hettangiens ainsi que dans les marnocalcaires rhetiens a la base de la serie jurassique de l'avant-pays proven~al. Ces terrains reposent sur les marnes, gypses et dolomies du Keuper localement aquiferes mais qui peuvent etre considerees comme relativement impenneables a l'echelle regionale. La serie presente un pendage general de 10-20SE fortement perturbe par un reseau de failles nonnales d'age oligocene orientees selon trois directions principales: N0-160, N20-30 et N90. Cette fracturation delimite des zones de horsts et de grabens globalement submeridiennes. Dans ce contexte, le drainage souterrain de l'aquifere du Jurassique inferieur se fait dans son ensemble de l'est vers l'ouest et localement selon la direction submeridienne des structures (MANGAN, 1982). Ainsi, la station de pompage de Mouans-Sartoux se situe a la cote NGF 250 au droit dun graben de 150-250 m de large et de 2-2,5 km de long, le graben de Pinchinade, constitue de calcaires bajociens effondres dans les formations rhetiennes (Figure 1 A et B).Le contact avec le Keuper a ete reconnu par sondages a 120 m de profondeur (MANGAN, 1990 et 1994). Les bordures laterales du graben fonnees en surface par Jes calcaires rhetiens et en profondeur par Jes formations gypsiferes du Keuper contiennent des nappes d'eau souterraines en communication avec celle du graben. Des piezometres installes dans le Rhetien montrent en effet un rabattement de la nappe lie au pompage dans le graben. Ces donnees nous conduisent a prendre en compte un impluvium d'alimentation d'environ 1 Km 2 nettement plus large que la surface meme du graben (0,2 Km2) (Figure lA). En raison de la lithologie marno-calcaire du reservoir Rhetien, les transmissivites sont plus faibles que dans le reservoir calcaire jurassique du graben. Le grand volume de ce reservoir lateral favorise en revanche l'emmagasinement des eaux souterraines. 6 th Conference on Limestone Hydrology and Fissured Med ia 137

PAGE 149

A ... . . COUPE AA' Outst Reservoirs annexes \ / 1-----------1 ,-..._Groben ---' \de Pinchinode I Est Source de I Forora; I ~-i100m 0 ........ . .. ) :'.) . n:=::neajacion ] 1 (I:] "'qu1ere t
PAGE 150

19 ........................... ...... .. .................. ... .. ............ ....... .... ............ .. .............. .. .......................................................... 190 "' A A A A A A 14 ....... :_ccc---.-~~cc::c:~=~~ ~ ~ ~ ~~ ~ ~ -----.---r;"'"' N ... 0 II) L...,j 9-'----,~::::::.....---=--~--------r--------,.------,-------~-----+90 05/rn/' 21/01/94 05/08/94 25/02/95 13/05/95 31/03/96 VOLUME POMPE CUMULE (ml) ml 0 100.000 200.000 300 000 400 000 500000 600000 )00.000 800.000 0+----..__---.c-:--:_.__ ___ .__ __ __. ___ ___._ ___ ___.._ ___ -L ___ ......1... ___ __,__......, .s-20 ,30 l:: f:\ -60 70 ' ' Volume pomp cumul, (ml) 0 1000 2000 3000 4000 o..,__ __ .__ __ .__ ___. __ __., __ s E :::-10 C j-1s 1-2 0 "' -25 'a, E ";;" 6 C !! ,o 20 8 A --A--A---A---~ o~--~---.------.------,Jour Volume pomp cumule ( ml) 0 1000 2000 3000 4000 Or-----'~ ___. __ __._ __ _,___~ ..... s E -1 c I 0 .0 a "' -25 -30 Figure 2 : Evolution des teneurs en sulfates et nitrates des eaux comparee au rabattement de la nappe en fonction du volume pompe depuis la mise en exploitation (en encadre, teneurs en nitrates et en sulfates des eaux comparees au rabattement pour deux periodes d'exploitation de la nappe). 4. Effet du pompage sur la chimie des eaux Sur toute la periode d'essai, l'evolution de la chimie des eaux se marque par une opposition nette entre la diminution progressive des teneurs en nitrates (element marqueur des eaux dinfiltration) depuis une valeur de 19,4 mg/I. jusqua 12 1 mg/I. et l'augmentation des teneurs en ions S042(element acquis dans le reservoir) de 32 mg/I. jusqua 194 mg/I. (Figure 2). Cette evolution reguliere se correle mal avec les remontees piezometriques liees a la recharge par les pluies ef ficaces et traduit plutot une modification a long terme de la chimie de l'aquifere. Un suivi chimique plus precis comprenant des prelevements effectues tous les jours permet d e comparer deux epoque s d'exploitation sur des durees semblables de 3-4 jours, mais pour des etats hydrodynamiques differents (Figure 2 encadres). En mai 1993 lors de lessai de pompage precedent la mise en s ervice du puits d e finitif la nappe est testee a un debit constant de 49 m3/h pour un rabattement final de 14 m. Apres un abaissement rapide dans les premieres heures du pompage correspondant a la vidange des zones les plus drainantes du reservoir la vitesse de rabattement ralentit traduisant la vidange des zones moins drainantes. La chimie des eaux evolue peu dans le temps. Les teneurs en N03varient selon l'infiltration 6 'h Conference on Umestone Hydrology and Rssured Med i a 139

PAGE 151

et demeurent relativement elevees (20-25 mg/I) tandis que les teneurs en SO42restent faibles (30 mg/I). L'eau pompee correspond done a de l'eau de reserve en relation avec !'infiltration dans la zone non saturee ; En septembre 1996 lors dun pompage plus intensif du a la mise en service simultanee des deux forages la nappe est testee a un debit constant de 60 m3/h. La profondeur initiale est de 43 met la profondeur finale de 75 m Un niveau plus profond de l'aquifere est ainsi teste. L'evolution du rabattement pr e sente les memes caracteristiques que !ors du premier essai. Cependant, si la vidange rapide initiale restitue des eaux a teneurs en NO3relativement fortes et a teneurs en SO42faibles correspondant aux zones drainantes en relation avec l'infiltration de surface la vidange lente se traduit par une forte augmentation des teneurs en SO42(jusqua 205 mg/I) et une diminution des teneurs en NO3-. L eau pompee provient alors de zones moins permeables faiblement influencees par l'infiltration recente Ces eaux apparaissent plus anciennes que celles prelevees en debut d'exploitation et/ou plus melangees a des eaux riches en SO42provenant des alimentations depuis les aquiferes bordiers. Les teneurs en sulfates de l'aquifere rhetien, prelevees au niveau des piezometres et des sources environnant le graben (Fig.I), varient entre 30 mg/I (piezometre) et 100 mg/I (source Saurin). Aucun prelevement n a pu etre fait dans les eaux du Keuper. Cependant, la source des Breguieres situee au niveau d une faille certainement injectee de Keuper presente une teneur de 199 mg/I. Les teneurs intermediaires des eaux pompees actuellement dans le graben correspondent ainsi a un melange entre des apports lateraux en provenance du Rhetien et du Keuper. Conclusion L'exploitation du graben karstique de Pinchinade se traduit par une modification des caracteristiques de l'aquifere sur le long terme. La recharge par les precipitations insuffisante en volume et trop discontinue dans le temps, ne provoque qu un remplissage partiel des zones les plus permeables de l'aquifere La reserve stockee dans la matric e moins permeable n est pas renouvelee. Au contraire, au fur et a mesure de l'abaissement de la nappe des parties plus profondes et plus anciennes de cette reserve sont sollicitees. La difference de charge avec les nappes laterales induit des apports importants depuis celles-ci. Comme ce phenomene se produit surtout en periode d absence de recharge de ces aquiferes (secheresse dete) les volumes d'eau deplaces correspondent done aussi a des eaux anciennes plu s ou moins riches en sulfates selon l'aquifere considere References COLLIGNON B. 1986. Hydrogeologie appliquee des aquiferes karstiques des monts de Tlemcen (Algerie). These de Doctorat Universite dAvignon et des pays de Vaucluse 2tomes 282 pages. MANGAN, C. 1982. Geologie et hydrogeologie karstique du bassin de la Brague et de ses bordures (Alpes-Maritimes France). These de specialite, Universite de Nice 2 tomes MANGAN, C. 1990. Commune de Mouans-Sartoux (06) Etude des disponibilite s en eau souterraine pour AEP en partie orientale de la commune. Rapport inedit du Cabinet Mangan. MANGAN C 1994. Commune de Mouans-Sartoux (06). Dossier synthetiqu e sur le forage de Pinchinade. Rapport inedit du Cabinet Mangan. MUDRY J. 1987. Apport du tra~age physico-chimique nature! a la connaissance hydrocinematique des aquiferes carbonates. These d'etat, UFR des Sciences et Techniques de Franche Comte-Besan~on 381 pages. Remerciements : Les auteurs remercient les responsables de la Regie Municipale des Eaux de Mouans-Sartoux pour leur soutien efficace (prelevements et mise a disposition des donnees existantes). 140 Proceedings of the 12 th International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 152

Application of a computational fluid dynamics model to cave river hydrodynamics M Hauns1, F. Hermann 2 an d 0 Atteia 1 1 Centre d'Hydrogeologie de l'Universite de Neuchatel, 11 rue Emile Argand, CH-2007 Neuchatel 2 Versuchsanstalt fur Wasserbau der ETH Zi.irich, ETH-Zentrum, CH-8092 Ziirich A b s tr ac t We show that in some cases the use of a 3-d model solving the Navier-Stokes equations is necessary to simulate flow in karst drains The presented model is based on the flow solver CFX-F3D. We describe the fundamental set of equations and the finite volume numerical method Availability of precise cave topography data is a necessary condition. A procedure to generate the grid from field data is outlined This model has been applied to calculate the relation between velocity and discharge for a flowmeter in the Rinquelle (Kanton St. Gallen, Switzerland) First results show that the model is able to deal with complex cave geometries With a high flexibilty of the solver to add user modules we expect to be able to resolve tracer and particle transport problems in the future. Key words computational fluid dynamics, turbulence, free surface, physical speleology (cave topography), tracer experiments, discharge measurements Introduction and objectives In a karst aquifer water flows both in the block structure and in drains Flow in the porous part can be described with models based on Darcy's law (Rossier, 1992) When considering models based on loss of head for drains, the formula of Louis is most adequate for phreatic karst conduits, while the formula of Manning-Strickler works better for vadose conditions (White 1988) The res u lts are good while the assumptions on which such models are based are valid in the considered case. When simplifying assumptions are not valid any more, higher-dimensional models have to be applied : there are cases where the currents in a cave m u st be known exactly to make reliable predictions. Simulations of tracer experiments in a drain with a pseudo 2-d river model (FLORJS) did not return correct results (Jeannin, 1996). To obtain recirculation and turb ul ent dissipation a higher dimensional model is necessary The velocity distribution is important for sedimentation and resuspension of particles and also for the physico-chemical processes between them It may be interesting to map the topography of a site, usually for low water conditions and then examine numerically what happens at high water: for example there may be recirculation zones that do not exist at normal water levels and w h ere material is deposited at high and falling water level conditions One-dimensional models are usually restricted to a few parameters thus drain tubes will have the same cross section for a range of discharges, and redistribution of flow to an increasing number of tubes with higher l evel cannot be taken into account (Jearmin & Wildberger, 1995) The redistribution can be resolved with a local three-dimensional model. The model presented here is expected to give im p roved answers to some of the questions me n tioned above. The fundamental set of the Navier-Stokes equations, the used numerical method and a procedure to generate the grid from field data are described An example is given to demonstrate the usefulness of t h e model for a flow prob l em in a cave. 1. Influence of the velocity distribution in a drain on tracer experiments The importance of velocity distribution on tracer transport can be seen on Fig I, which shows schematically the difference between a 1-d simulation and a 2-d simulation for a channel transporting tracer dye. 1n the 1-d model the velocities will reduce for a wider cross section but no eddies will develop in a 2-d mo d el a recirculation zone exists Some tracer substance will be captured therein and when released give a retardation that deforms the measured concentration peak s~ 1. ~-~ --=--2 -d Fig. 1 : t o r es ol ve t ra ce r s ub s tan ce r et a r din g w h en ca ptur e d in a n e d dy 2 -d mod e llin g is n ecess a ry 2. Cave hydrodynamics modeling with CFX The hydrodynamic problem of water flow is described by the Navier-Stokes equations Topography mapped with sufficient accuracy is needed to solve correctly the equations numerically W hil e most of the water flows in the main branch of a gallery, sufficient accuracy might be obtained by mapping cross sections every five meters, and the bottom profile to a precision of five centimeters. To deal with very complicated structures like a strainer a precision of 5-10 cm will be needed in all three d imensions. For mula tio n o f the N a v i e rS t okes eq uati o n s The Navier-Stokes equations can be derived from a general transport equatio n for ex t e n sive quan t i t ies (Honerkamp and R omer, 1986) For a fluid or gas dynamics problem the Navier-Stokes equations in coordinate-independent vector notation are : a dt p + V ( p 1J.) = 0 o) a dt ( p yJ + V ( p ~ ~) = V p + b. (~)+ B (2) l! p velocity vector p fluid density pressure dynamic viscosity 6 th Conference on Umestone Hydrology and Fissured Media 14 1

PAGE 153

B body forces Equation (1) comes from mass conservation, equation (2) describes the behaviour of an infinitesimal volume element in the fluid exposed to inner and outer forces Viscous friction pressure gradient and inertia are examples for inner forces, buoancy, magnetic or centrifugal are examples for outer forces. Simplifying assumptions which are generally based on an priori integration of the equations for a direction where little change occurs (e g the z direction) lead to lower dimensional models Definition of the fluid dynamic problem Liquid flow is in general treated as incompressible flow therefore equation (1) will simplify in the liquid phase Water is a newtonian fluid, that means the stress tensor will be isotropic. As the temperature of the water changes very little and slowly we consider the problem to be isothermic, there is no boyancy force due to temperature gradients Buoancy forces may be present due to density variations caused by a high salinity Turbulence modelling is important since colloid particle diffusion and mixing depend on it. Except for sumps and high water conditions, there will be a free surface Free surfaces can be modelled with a moving grid, but for such beautyful things like breaking waves and bubbly boiling pots after a waterfall multiface modeling is necessary, hence the Navier Stokes equations must as well be solved for the air phase. Turbulence modeling with CFX-F3D Turbulent flows are very complex time-dependent flows that are governed by the laminar flow equations presented above in the smallest scales($; 0 1 mm). The size of the smallest eddies depends on the Reynolds number Re. The resolution required grows with Re 91 4 ( Moharnmadi & Pironneau, 1993). For most turbulent problems a grid that resolves the small eddies is not feasible with current technology Therefore a turbulence model is introduced which solves the transport equations for the Reynolds-averaged quantities: t+lit (t) = 2~ I ( -r)d-r (3) t-lit represents any averaged quantity, e g. velocity, pressure It is important that the frequency of turbulent fluctuations and the frequency of mean value fluctuations are well separated. So 6t in eq. (4) can be chosen large relative to the time scale of turbulent fluctuations, but small relative to the time scale of the problem we wish to solve Under this assumption fields can be split into their mean and fluctuating parts: =+~ w With the laws for the averaged quantities, equation (2) can be rewritten in Reynolds-averaged form: a dt (p~) + V (p~@ ~) = -V p + fl(~) V (p ti @~') (5) where the bars for all variables as well as the body force term have been dropped for clarity. The additional term V (p ~ ~:)describes the influence of the turbulent eddies on the mean flow and has the dimension of a stress One possibility to model it is by introducing turbulent viscosity by setting !{ !{ = v,(V~+ V!/). Free surface modeling with CFX-F3D In a partially filled drain the free surface of the water must be computed Free surfaces may be modelled with a moving grid, but in torrential flow mixing with air is important so we use a multiphase formulation The free surface is then defined as the interface between the two phases A surface sharpening algorithm helps to overcome the numerical diffusion at the surface Particle and tracer transport modeling In CFX-F3D tracer transport can be expressed by means of an additional scalar equation while the tracer particles have no inertia and the same density as the surrounding liquid and will therefore always move with the same velocity as the fluid. When the particles have a different density a two fluid model must be applied For high particle concentrations there is a coupling between the momentum equations, but for a tracer experiment and colloid particle transport modelling this can be neglected Numerical Solution The set of equations described above has to bediscretized and solved within a computational domain This may be done with a finite difference, finite volume or finite element method. We chose to work with the finite volume method because it is fast and offers a body-fitted grid There is a computational grid that is always topologically rectangular in this numerical method and a physical grid which adapts to the shape of the gallery A procedure to create the physical grid from field data is described later. For simple calculations the physical grid is identical to the computational grid, but for most of our applications it will be distorted and has to be mapped to the computational grid. Fig. 2a shows a scheme of the two grids The advantage of a body fitted physical grid is that the velocity vectors are more likely to be parallel to the cell bounds than in a rectangular grid and that no cells are wasted in complex geometries. The transformation between the two grids involves elements of tensor calculus and will later influence the differencing scheme for the equations The Navier Stokes equations will contain additional force terms originating from the transformation of an isotropic advection-diffusion problem in physical space to an anisotropic advection-diffusion problem in computational space. The physical grid has always to be structured, therefore the geometry is not as arbitrary as is is with unstructured grids To describe complex structures it is necessary to join several subgrids. Fig 2b shows an example for a composite grid in a conduit. The volume between neighbouring points will be referred to as a volume element, generally a parallelepiped. In CFX F3D the variables are discretized at the centers of these elements, and the equations are discretized treating these elements as control volumes for the conservation laws The Navier-Stokes equations have to be satisfied for each control volume simultaneously The default differencing scheme is UPWIND, but several higher order schemes can be chosen The SIMPLEC algorithm is used for velocity-pressure coupling 142 Proceedings of the 12 th International Congress of Speleology 1997 Switzerland Vo l ume 2

PAGE 154

a) b) \ ~II Fig. 2: a) physical grid fitted to an irregular channel and corresponding computational grid. b) body fitted grid for varying channel geometries blocks are shaded (see jig. 3 for flow pattern classification). 3. Grid generation from field data Underground rivers flow in galleries that may be located both in the the vadose and phreatic zone of karst aquifers. We may focus on the vadose (non-saturated) zone where field data are obtained more easily Non-saturated drains show a wide variety of flow patterns with frequent change between fluvial and torrential flow Basic flow patterns are classified in Fig 3. Big lakes are separated by series of cascades, ramp-type rapids and calm channels follow each other (Marechal, 1994) =<@=~~ (a} lncgllarly shaped channels (b) Core-shaped entrances to basins Holes in the bed (c) TIited plate ard small step-typ:: rapds (d} Wa. t.erfaJJ into II basin 'M th varying length (e} Elbow curves and influence of sinoosity Fig. 3: basic flow patterns in vadose karst conduits There are a few important rules to obey for grid generation: a body-fitted grid should be as orthogonal as possible, especially where gradients are high. The spacing of the grid may vary to be more precise at "hot spots" (vo lumes where the variables change fast), but the variation has to be smooth. The physical grid is created from sections of the cave and a profile in geographic xyz coordinates of the cave floor. The sections may be digitized from drawings with an arbitrary number of points. A polygon is fitted to each cross section and placed onto the profile of the cave floor. The volume between two subsequent sections is called a block 4. Application to cave hydrodynamics The 'Rinquelle' cave is located north of 'Walensee' in eastern Switzerland It consists of a long (approx. 1 km) underwater gallery with two branches. One of these feeds a spring, which is active during summer and inactive in winter time the other one starts about 300m after the entrance of the sump and is the main outflow branch of the system. It's explored length is approx. 800m It 's emerging point lie s at the bottom of the Walensee To gain insight into the flow regime of the cave, flow meters were installed in the spring and in the other outflowing branch. The latter is a flow velocity meter, located roughly in the middle of the gallery. With the assumption of uniform flow, the discharge may be calculated by Q=V*A, where A is the cross section and V the mean flow velocity. However, the topography of the gallery suggests that the assumption of a uniform velocity distribution might not be fulfilled In order to find out more about that, a numerical simulation of the flow through a part of the gallery was performed and the resulting flow field analysed. The grid consists of ea. 19000 cells The calculation took slightly more than an hour on a modem workstation Turbulence was modelled with the commonly used k-E model. Fig. 4a and 4b show the computational grid used, Fig 5 gives an impression of the flow behaviour by the use of streamlines and grey-coded cross-sections (lowest velocity dark grey) and Fig. 6 shows a velocity profile modelled at the place of the velocity meter. The streamlines in Fig 5 start in a plane near the entrance, covering the whole vertical extent. Note, that the upper two groups of streamlines are not affected by the deep part of the conduit, while the lower ones are kept in the large recirculation zone. Some of them escape towards the outlet. Fig. 4a: Outline of the computational domain which represents a small part of the rinquelle system Inflow is at the square face to the right, outflow in the upper left corner. The dimensions are 15 x 7 x 3.2 meters, the inflow velocity is lmls. Another interesting observation may be made at the outlet, where all the streamlines concentrate at the right side of the duct. The grey shaded cross-sections show the velocity distribution at the end of the deep part and at the outlet. In 6 th Conference on Limestone H ydrology and Fissured Media 143

PAGE 155

Fig. 4b: Grid structure of a part of Fig. 4a. The grid consists of several blocks glued together. The spacing in the streamwise direction is 50 cm. Fig. 5: Flow field in the duct of Rinquelle, visualised by streamlines and grey-shaded cross-sections. See explanations in the text. 1.00 ~-~---~---~ -~~ 0 0 1 0 2.0 he i ght [ml Fig. 6: Velocity (streamwise component) distribution along a vertical line in mid-duct at the position of the velocity-meter, located slightly behind the first cross section of Fig. 5. The straight line shows the mean value of the velocity, which is lower than the local value in mid-duct. the first cross section, the highest velocities are in the middle of the conduit (see also Fig 6), while they are near the right boundary at the outlet. The black zone in the second cross section indicates low velocities suggesting together with the streamlines that the outlet cuts through another recirculation zone Fig 6 shows the velocity profile co mputed at the position of the velocity meter. Th e plot shows that the velocity taken in the middle of the conduit (where the measuring device is located) is higher than the mean value of the whole curve The sa me could be observed in the other direction of the cross section The overestimation of the discharge from the measured velocity is approx. 7%. The streamline plot shows that the direction of the streamlines near the position of the velocity meter are roughly parallel to the main direction of the gallery gtving confidence that the measured values are correct to the accuracy of the device 5. Conclusion Comparison of tracer experiments with ex1stmg double porosity and 1-d or pseudo 2-d models has shown that they do not represent correctly the hydrodynamics and flow patterns in a drain The presented approach gives a physical solution for the flow and the shape of the free surface. This is useful for velocity/discharge calculations but also for the interpretation of tracer experiments in drains and for particle dynamics modeling Due to computational ressources and data required, this model is well-suited to solve flow locally and to obtain information about physical variables that influence other processes of interest, e g. particle dynamic s in the water. The deterministic model will be used to detail the water velocities in the specific cave geometries described above. These effects will then be lumped and introduced into a simplified 1-d network model which will be able to describe the whole cave References BR0GER K. Duct Profiles around the position of the velocity-meter. Personal communication 1991. CFDS CFX 4.1 Flow Solver User Guide. Computational Fluid Dynamics Services, Harwell Oxfordshire UK HONERKAMP J & ROMER H. 1986 Einfiihrung in die theoretische Physik. Springer Heidelberg New York Berlin JEANNIN P .Y 1996 Structure et comportement hydraulique des aquiferes karstiques These, CHYN 1996 JEANNIN P -Y. & Wll..DBERGER A 1995 Modelisation des ecoulements dans le reseau du Holloch. Congres national suisse de Speleologie (Breitenbach oct. 1995) KLINGENFUSS B 1977 Die Rinquelle Stalactite, 27: 6792 MARECHAL J -C 1994 Etude et modelisation de l'hydraulique et du transport dans Jes drain s karstiques Travail du diplome, CHYN 1994 M0HAMMADI B & PIR0NNEAU 0 1993 Analysis of the K-Epsilon Turbulence Model. Chichester: Wiley R0SSIER Y & KIRALY L. 1992 Effet de la dilution sur la determination des dispersivites par interpretation des essais de trar;age dans les aquife res karstiques. Bull du Centre d hydrogeologie de Neuchatel I I: 1 15. TR0SCH J AND ZURBR0GG c 1995, Turbulent flow in high permeable karst. Numerical simulation, Bulletin d'Hydrogeologie No 14 235-240. WHITE W B 1988 Geomorphology and hydrology of karst terrains Oxford University Press Oxford 464 pp 144 Proceed i ngs of the 12 '" Internat i onal Congress of Speleology 1997, Sw itz erland Vo l ume 2

PAGE 156

Characterization of karst aquifers by heat transfer by Dirk Hiickinghaus, Rudolf Liedl & Martin Sauter Applied Geology, University of Tu bingen, Sigwartstr 10, 72076 Ttibingen, Germany Abstract The paper presents a modelling approach which couples the hydraulically complex flow system and the heat transport processes within karst aquifers. Using this model together with quantitative measurements of flow and temperature in karst springs it will be possible to obtain detailed information about the geometry (surface, diameter etc .) of the conduit system Zusammenfassung In diesem Beitrag wird ein Modellansatz vorgestellt, der die komplexe Hydraulik eines Karstaquifers und dessen Warmeflufi vereinigt In Verbindung rnit quantitativen Durchflufiund Temperaturmessungen in Karstquellen wird es mit diesem Modell moglich sein, mehr und detailliertere Informationen iiber das Rohrensystem eines Karstaquifers zu erfahren 1 Introduction Groundwater flow in karst aquifers is controlled by two coupled hydraulic systems, i e. a conduit network consisting of relatively large fractures or tubes and a system of small fissures within the rock matrix These two flow systems show very different kinds of flow behaviour The high storage system responds rather slowly to recharge events while at the same time a fast response can be attributed to the low storage high flow velocity conduit system. ATKINSON (1977) demonstrated that it is mainly the flow in the conduit system that dominates the groundwater flow Due to the high flow velocity (up to 100 m/h) in the tubes information about the geometry of the conduit network is an essential prerequisite for groundwater risk assessment. Recent investigations have shown that it is possible to obtain information about the network geometry from temperature deviations in karst springs after recharge events (BENDERITTER et al 1993 ; ROY & BENDERITTER. 1986 ; RENNER, 1996) This paper presents a new modelling approach which takes into account the two different flow systems and the heat transport processes in the conduit system and the rock matrix It is intended to utilise this model in order to obtain more information about the structure of the conduit network and about the geometrical parameters of the tubes (diameter, length) based upon measurements of karst spring discharge and temperature 2. Observations After recharge events characteristic responses of temperature have been observed in karst springs (Fig 1). In many karst systems the temperature deviations are independent of seasonal atmospheric temperature fluctuations (ALEY, 1970 ; HESS et al ., 1988 ; DAVIES, 1992) In order to produce this characteristic phenomenon three conditions have to be fulfilled (RENNER, 1996 ): ( 1 ) a sufficiently thick unsaturated zone to provide significant temperature differences between the atmosphere and the aquifer (2) a sufficient residence time in the upper cooler region (epikarst) of the unsaturated zone so that the recharge water can adjust its temperature to the surrounding rock, and (3) a rapid seepage from the epikarst through the vadose zone to retain a difference in temperature between the seepage water and the water in the phreatic zone (Fig 2) : I lL I ],,m~,itttTit 1 3 88 15 9 88 1 4 89 15 10 89 1 5 90 15 11 90 time Figure 1: Time series of spring water temperature and discharge variation in the Gallusquelle (Swabian Alb SW Germany) karst spring (RENNER, 1996 ; data from Sauter 1992) 6 th Conference on Limestone H ydrology and Rssured Med i a 145

PAGE 157

Consequently in this case the temperature of the seepage water reaching the groundwater table is lower than the aquifer temperature disregarding seasonal fluctuations of the atmospheric temperature Then the seepage water is collected by large fractures or tubes and transferred to the karst spring at flow velocities in the order of 100 m/h Therefore the adaptation of the recharge water temperature to the higher aquifer temperature is delayed and the temperature pulse of the recharge event reaches the spring, where it can be measured .,,. karst vadOH zone phr-lc -..zone le\19 and a(fer w!ter 9 ( aq<.ifer temperaue) Figure 2 : Seepage through the unsaturated zone : the recharge water accumulates in the epikarst reservoir before seeping through the vadose zone along highly permeable fractures (RENNER, 1996 ). 3. Objectives and conceptual model The field observations summarised above promote the idea of characterising the conduit system of karst aquifers by using spring water temperature data RENNER ( 1996 ) has demonstrated that this is possible for aquifers with a highly interconnected conduit system which can be represented by a single equivalent' fracture If there is a temperature difference between the inflow in the tube and the surrounding soil matrix a heat transfer between tube and matrix occurs The transfer is controlled by the flow velocity the thermal characteristics of the surrounding rock and by the geometry of the tubes Hence the measured temperature in the spring behaves analogously to tracer transport If we know the input and rock temperature and the flow velocity in the tube it is possible to obtain infonnation about the tube geometry Recent numerical calculations by RENNER ( 1996 ) show that it is even possible to distinguish between differently shaped tube systems using temperature measurements (Fig 3 ). At present a research project is in progress in order to transfer this approach to more realistic discrete tube networks In particular the main objectives of this project are ( a ) to identify the geometrical parameters of the fast flow system (b ) to characterise the maturity and the type (" shallow' or deep ") of a karst aquifer (c) to establish a mathematical model of heat transport in the conduits and the rock matrix and ( d ) to implement this heat transport module into an existing groundwater flow model for karst aquifers ( CLEMENS et al ., 1996 ) in order to apply it to karst catchments in the Swabian Alb ( SW Germany) In order to reach objective ( c ) a conceptual model of flow and heat transport in karst systems has to be established The flow model is based upon a distinction between a slow and a fast flow system The slow flow in the smaJl fissures is modelled by a continuum approach using the Boussinesq equation (BEAR & VERRUIJT 1987 ). The large conduits are represented by a pipe network consisting of cylindrical tubes The flow in this conduit system is modelled by regarding the conservation of mass according to Kirchhoff's rule ( CLEMENS et al ., 1996) The two flow systems are coupled by an exchange term This approach is used in the program system CA VE ~arbonate _Aquifer Yoid ~volution) which was developed to simulate the genesis of a karst system (HUCKINGHAUS et al ., 1997 ; CLEMENS et al ., 1997 ). For heat transfer modelling it is assumed that convection is the dominant process within the tubes whilst heat transport in the rock matrix is governed by conduction perpendicular to the tubes (Fig 4) Additionally heat exchange between the rock matrix and the conduits has to be considered oI I a. fractures Q-J 41 fr Input ... .a T p CII Gi 0.. E 0 To t= time t breakthrough curves Q 41 Cl Qt iii .c 0 ,,, 'e 02t=O t1 time t 1e3 1 e4 1 e5 time t-V/Q 2 (s) Figure 3 : RENNER (1996) shows that it is possible to identify fracture systems with different geometry but identical volume ( V 1 = V 2 + V 3 ) of the systems with variable discharge 146 Proceedings of the 12 th International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 158

r l conduction convection I d X Figure 4 : Important processes of heat transfer present in the rock and tube system 4. Mathematical model Heat conduction in the rock matrix is described by the differential equation ( cylindrical coordinates) (1) where T sot is the temperature of the rock, t is time, r is the radial coordinate in the cylindrical system A.so, is the thermal conductivity of the rock, P so t is the density of the rock and cso, is the specific heat capacity ( CARSLA W & JAEGER, 1959) The convective transport of heat in the flow direction x along each tube can be described by with the fluid temperature in the tube Tfl, the flow velocity v, the density of the fluid in the tube Pf/ the specific heat capacity c fl of the fluid, the tube diameter d the volume of the tube V and the surface of the tube A The first term on the right hand side represents the exchange of heat between the rock matrix and the conduit water At the surface of the tube both temperatures are identical, i e d Tsd(x, r=, t )= Tfl (x t ) 2 ( 3 ) For a single tube it is possible to find solutions of eqs (1) (3) in the Laplace space for a uniform initial temperature T 0 and a temporally constant water temperature T ""P at the inlet of the tube lf the symbol /\ denotes Laplace transformation we obtain A Ttmp T 0 Tfl(x,s)=T 0 +---* s and {-< A-,., A K 1 ( pd I 2 ) p )x} exp s + p .,c., V Ko( pd 1 2 ) v A A K 0 (rp) T sd (x r,s) = Tfl ( x s)----'-K0(pd I 2) ( 4 ) ( 5 ) where p = .J p "'c "' s I A "' and s is the Laplace variable The functions Ko and K 1 are modified Bessel functions of integer order (ABRAMOWITZ & STEGUN 1970) The temperatures T fl and T sot are obtained by applying the inverse Laplace transform to the equations ( 4) and ( 5 ) A computer program using the numerical inversion of Laplace transforms is currently being developed 5. Summary and outlook It has been shown that field observations support the idea of characterising the fast flow system of karst aquifers by analysing spring water temperatures To this end, a new numerical model is being developed which couples the two different flow systems in a karst aquifer to beat transfer processes in the conduit system and in the rock matrix It is planned to continue the research activities as follows : in a first step the numerical solutions for a single tube will be verified Next this solution will be implemented in the CA VE program in order to obtain a modelling tool for the simulation of flow and transport in a discrete conduit network which is coupled to a continuum model representing the rock matrix. Finally this combined model will be applied to karst areas in the Swabian Alb ( SW Germany ) where presenting measurements of spring water temperature are being carried References ABRAMOWITZ M. & I. A. STEGUN 1970 : Handbook of Mathematical Functions Dover Publishing Co ., New York ALEY T J. 1970 Temperature fluctuations of a small Ozark spring C aves and K ar s t V ol. 1 2, N o 4 ATKINSON T C 1977 Diffuse flow and conduit flow in limestone terrain in the Mendip Hills Sommerset ( Great Britain ). Joum. H y dro/ ., 35 : 93 100 BEAR, J. & A. VERRUIJT 1996 : Modelling groundwater flow and pollution D Reidel Publishing Compau Dordrecht 414 p BENDERITIER Y ; ROY B & A. TABBAGH 1993 Flow characterization through heat transfer evidence in a carbonate fracture medium : First approach Wate r R esou r R es., 29 (11 ): 3731-3747 CARSLA W H. S & J. C. JAEGER 195 9 : Conduction of heat in solids 2 nd ed Oxford University Press London 510 p CLEMENS T. ; HUCKINGHAUS D .; SAUTER, M .; LlEDL R. & G TEUTSCH 1996 A combined continuum and discrete network reactive transport model for the simulation of karst development. IAHS Pub!. 237 : 30 9 318 CLEMENS T. ; HUCKINGHAUS D .; SAUTER, S .; LlEDL R & G TEUTSCH 1997 Simulation of the 5 m C on ferenc e on Limestone Hydrology and Fissured M edia 147

PAGE 159

development of karst aquifers using a pipe flow model coupled to a continuum model 2 Model verification and sensitivity analysis ., Water Resour Res ., in prep DAVIES G J. 1992 Water temperature variation at springs in the Knox Group near Oak Ridge, Tennessee Proceedings of Ground Water in Karst Terranes HESS J. W & W B WHITE 1988 Storm response of the karstic carbonate aquifer of Southcentral Kentucky Journ Hydro/ ., 99 235 252 HUCKINGHAUS D. ; CLEMENS T .; LIEDL, R ; SAUTER, M & G TEUTSCH 1997 Simulation of the development of karst aquifers by using a pipe flow model coupled to a continuum model 1. Theory Water Resour Res ., in prep RENNER S 1996 Warrnetransport in Einzelkliiften und Kluftaquiferen Untersuchungen und Modellrechnungen am Beispiel eines Karstaquifers University of Tiibingen, Tiibinger Geowissensc haflli che Arbeiten, C 30 : 89 p ROY B & Y BENDERITTER 1986 Transferts thermiques naturels dans un systeme aquifere carbonate fissure peu profond. Bull Soc. geol France ( 8) t. II n 4 SAUTER, M. 1992 Quantification and forecasting of regional groundwater flow and transport in a karst aquifer (Gallusquelle, Malm, SW Germany) University of Tiibingen, Tiibinger Geowissenschaflliche A rbeiten C 13 150p 148 Proceedings o f the 12 th Internat i onal Congress of Speleology 1997, Sw i tzerland Vo lu me 2

PAGE 160

Dispersion and tailing of tracer plumes in a ka r stic system (Milandre JU Switzer l and ) Pierre-Yves Jeannin 1 & Jean-Christophe Marechall 2 1 Centre d'hydrogeologie de l Universite de Neuchatel Rue Ernile Argand 11 CH-2007 Neuchatel Switzerland 2 Now at Ecole Polytechnique Federale de Lausanne (EPFL), Geolep CHIOI 5 Lausanne Switzerland Abstract A large number of tracing experiments have been carried out in a karstic aquifer in the Swiss Jura. These allow to observe the evolution of a tracer plume along the length of a karst conduit. The method of Sauty was used to make possible the comparison between all the observed breakthrough curves The flow velocities and the dispersivities obtained are extremely variable. The dispersivities measured at different point s along the length of an underground stream in the course of the same tracing experiment increase with di s tance (scale effect) If the fit of theoretical Sauty curves on the experimental curves works well for rising limbs, this is not always the case for falling limbs: a tailing effect or Jag of the experimental curves compared to the theoretical ones is often observed. Micro-tracings have shown that the lag effect is linked more to the karst conduit geometry than to the types of flows (turbulent or laminar). Measurable tailing effect is induced by the presence of a single conduit enlargement (also called pool) Further the experiments have shown that a succession of enlargements along the length of the underground stream causes a clear increase in the dispersivity and a "homogenisation" of the recovery curve which shows up by the apparent disappearance of the lag effect. These observations show clearly the influence of the heterogeneity of the karst conduit geometry on the breakthrough curves. This effect might be considered when one interprets the shape of the breakthrough curves especially for dispersivity estimation Resume De nombreux essais de trac,age ont ete realises dans un aqui.fere karstique du Jura suisse. lls permettent d'observer ) evolution dJ nuage de traceur le Jong d'un conduit. La methode de Sauty a ete utilisee pour rendre Jes di verses courbes de restitution comparables Les vitesses et Jes dispersivites obtenues soot extremement variables. Les dispersivites mesurees en differents points le long du ruisseau souterrain au cours d'un meme trac,age augmentent avec Ja distance (effet d'echelle) Si l'ajustement des courbes theoriques de Sauty sur les courbes experimentales fonctionne bien sur Ja montee, ii n'en va pas toujours de meme pour la descente: un effet de retard (tailing effect) des courbes experimentales par rapport aux courbes theoriques est souvent observe Des micro trac,ages ont montre que l'effet de retard est davantage lie a la geometrie des conduits karstiques qu'a Ja nature d es ecoulements (turbulents ou l aminaires) Un effet de retard mesurable est induit par la presence d'un seul large bassin Les experiences ont egalement montre que la succession de plusieurs bassins le long du ruisseau souterrain induit une nette augmentation de la dispersivite (effet d echelle) et une "homogeneisation" de la courbe de restitution se traduisant par une disparition apparente de l'effet de retard. Ces observations montrent ('influence des variations de Ja geometric des conduits karstiques sur Jes courbes de restitution des traceurs Cet effet devrait etre pris en consideration )ors de I 'interpretation des courbes de restitution notamment lors de l'estimation de la dispersivite du milieu 1. Introduction In karstic aquifers the dispersivity values measured on breakthrough curves are often high (MEus 1993 ), scale depending and the shape of the breakthrough cu rves is frequently strongly asymmetric (MARECHAL 1994). These results do not fit well with results of a theoretical 1-D homogeneous model, assuming that conduits are thin but very conductive features (pipes). The discrepancies between the 1-D model and the observations leads many authors to consider further processes generating dispersion namely adsorption desorption (DE MARSILY 1986) a Large molecular diffusion (SEU.ER et. al 1989 MALoSZEWSKI & ZUBER 1989) interaction with stationary water (BrVER 1993) the effects of dilution (ROSStER& KIRALY 1992) etc .. This paper is focussed on describing the relation between the flow conditions in karst conduits and the breakthrough curves. The low permeability volumes i. surrounding the conduits in karst aquifers are not under consideration here. Breakthrough curves could be observed along the path of an underground karstic flow and provide then informations about the dispersion processes which could be respo n sible for the strong asymetrical breakthrough curves frequently observed in karst systems 2. The test-field of Milandre The drainage basin of the Milandrine ( I 3km 2 ), situated in the north of Switzerland contains an underground river the Milandrine which can be followed upstream from the emer gence point for 4.6 km (actual distance in the cave, equivalent to 3 km as the crow flies figure I) The exsurgence of the Milandrine is made up of two principal springs one perennial the Saivu, with a discharge of 20 150 1 / s and the other ephe meral La Bame with a discharge ofO to 1500 1 / s Four measu ring stations were set up along the length of the underground rive~. allowing to measure the passage of tracers along the section to be observed The measuring stations are situated at 4 6 km from the springs (station I upstream of the cave) at 4 km (station I ) 3 km (station 2) 1 5 km (station 3 ) and O km (station 4, springs themselves) respectively. The flow is open along the largest part of the underground river with only the downstream part of about 500 metres in length being constricted About sixty tracing experiments were carried out on this site by injecting NaCl (50 to 500 kg) at different points in the drainage basin upstream from the known part of the cave The breakthrough curves are obtained at each station by measuring the electrical conductivity of the water and by transforming it into NaCl concentrations 6 th Conferen c e on Limestone Hydro l ogy and Rssured Med i a 149

PAGE 161

= J' Underground River of Milandre : dlspersiv ity and lag Station 1 1, Discharge al low water (19 10 1994) cl o.s I <:! 0 L--L.--.9' ----6..-7.10~..._.,_..L~J...._ __ too t i me alter Injection lh) Figure I : The tracers are generally observed at four or five stations. The tracing experiment (Tracing 42 in table I) presented here show a clear increase in dispersivity ( al.) along the underground course of the river (scale effect). The tailing effects vary a great deal along the length of the flow. 3. Aims and methods A frequent tailing effect (strongly assymetrical breakthrough curves) has been observed in Milandre Does this retardation come from adsorption desorption processes, from diffusion of tracer in the matrix, from dilution or from any other process ? Direct observations in the cave, make it possible to be almost sure that none of these processes are significant: there is no clays for adsorption desorption, the porosity of the matrix is very low (<2%) and the flow velocity too high for a substancial diffusion in the matrix or towards immobile water, there is no tributary, at least along certain sections, causing a dilution effect. As scale and tailing effects are observed anyway, we will discuss here some further processes which may be responsible for. In order to be able to compare the many breakthrough curves we measured, only those carried out in similar hydraulic conditions (at low water levels) have been considered. The changes of the parameters as a function of discharge could be the subject of a future paper. The breakthrough curve Following the injection of a tracer, the concentration C(X,t) of a solute obtained at an observation point situated al a distance Xfrom the injection point is expressed as a function of time. This is the breakthrough curve. The modal transit time (tm) represents the largest number of tracer particles: it is therefore the time value with the highest probability This time is used for calcuating the modal velocity (X/t.,) which is the velocity given in table l. The Sauty method In order to compare the breakthrough curves, the simpliest possible model has been used (Sauty method). This model is easy to calibrate and does not introduce any of the dispersion processes described before. It provides physical parameters (dispersivity) but in this study, it is utilised as a tool of comparison, not as a physical model of the reality We mainly focus on the tailing effect (difference between the model and the breakthrough curve) and the spatial variation of the calculated dispersivity (apparent dispersivity). SAlJfY (1977) provides a series of theoretical curves that have been calculated using a uniform uni-dimensional, laminar flow model. F.ach of the theoretical curves corresponds to a hydrodynamic Peclet number. The experimental breakthrough curves are superimposed on the theoretical curves and adjusted to the one giving the best fit. From the appropriate Peclet number we easily calculate the dispersivity which is equal to the distance divided by the Peclet Number. The retardation or tailing effect To obtain a dispersivity value, Sauty suggest to fit both the theoretical curves and the experimental curves on the rising limb of the peak (figure 2) In most cases, the curve fitted on the rising limb does not then adjust correctly on the fall. It is therefore possible to make several successive fits and to determine several longitudinal dispersivities for different parts of the falling limb of the curve. Generally the experimental curve remains above the theoretical curve of Sauly in the falling limb of the peak. This is the so called tailing effect 0B k e o.6 u u 11 a: 0.4 u oz 0 3 0 4 0 .5 0 6 0 7 0 8 0 9 reduced time of the grid (IA) 5 6 7 8 9 10 lime alter lnjecllon [h ) Figure 2 : Fitting of the curve on Sauty grids. The fitting is carried out on the rising limb of the curve and gives a value for despersivity aL. A lag effect of the experimental curve with respect to the Sauty model is visible. 4. Results and discussion Apparent dlspersivlty: scale effect The dispersivities measured on the rising limbs of the breakthrough curves at different points along the Milandrine during the same tracing (Table 1) almost systemetically rise as a function of the distance. This increase in dispersivity shows the existence of a scale effect: if the distance increases, the probability of encountering many and large dispersive phenomena also increases This scale effect is well known in heterogeneous media (PrAK & TarrSCH 1994 SEllER et al. 1989, SAITTY & KINZELBACH 1988). The results presented here clearly show that this scale effect also exists in karstic conduits. Tailing effects OBSERVATIONS Tailing effects appear very often at certain measurement points (Figure 1). The important tail observed at the station I (upstream) is probably due to the injection conditions and Lo the transfer through the unsaturated zone. As this part of the route cannot be observed directly, it is not considered here. The tail at station 2 is always very reduced but the apparent dispersivity is always substancially higher than the one in station I. At station 3, a tail is always observed and at the 150 Proceed i ngs of the 12 "' International Congress of Speleology 1997, Switzerland Volume 2

PAGE 162

exsurgence (Saivu pt 4)), although the major part of the tracings observed have shown no lag, the tracing presented on figure 1 shows a lag which cannot be ignored. Note that at a given station, the size of the tailing effect is variable from one experiment to another. Slltlon 1 Slltlon 1' Slltlon 2 Slltlon 3 Slltlon 4 Trtelng 1 408m 15 1/e 1110m 3805m 5173 m a=2 7 V=26 a=6 V=57 a= 14 V=B:I a =26 V= 105 Trlclng 2 "m 13 1/e 1104m 5167m a=4 V:28 a=11 V:52 a=26 V:73 Trtelng 3 13 1/e mm 4765m Q:4 V= 124 a=48 V= 100 Trtelng 11 203m 32 1/e 4ll68 m a=20 V:21 a =50 V= 135 Tracing 12 173m 00 1/e 4938m a=S V=S2 a=5 V= 157 Tracing 14 455m 21 1/e 1157m a=3 V=40 a=19 V=84 I Tracing 23 176m 32 1/e 3573m 4841 m a =2 V:88 a=12 I V=114 a = 14 V= 174 Tracing 24 176m 25 1/e 3673m 4841 m a=2 V:69 a =12 V= 166 a =25 V= 143 Tracing 25 461 m 24 1/e 2026m 3858m a=5 V:42 a=5 V=123 a= 19 V= 132 Tracing 40 176m 00 1/e 1741 m 3573m a =? V:306 a= 17 V:229 a=18 V:205 Tracing 41 40m 26 Ill 1605m 3437m a=1 V:200 a =16 V=217 a=17 V= 187 Trtelng 42 40m 22 1/e 1605m 3437m 4805m a=0 4 V:44 a=16 V:178 a= 17 V=149 a=24 I V=161 Table 1 : Dispersivities ( a) in m and flow velocities (V) in mlh measured at different distances along the length of the river's underground course. The dispersivities always increase with the measurement scale (scale eff eel). The discharge is given at station 1 TURBUl.ENT FLOW AS A DISPERSIVE PROCESS ? The flow are turbulent on all the length of the underground stream (JFANNTN & MARECHAL 1996). Can turbulent flow cause the observed tailing effect ? The existence of turbulent flow zones, downstream from which the theoretical and experimental curves can be fitted perfectly seems to indicate that the tailing effect is not due to the turbulent flow regime Further, the station 2 where the tailing effect is always reduced is located after a section of torrential flow. To confirm this hypothesis, several "micro tracings" (tracing experiments on sections of a few tens of metres in length) have been carried out. Let us consider the "micro tracing" (A) depicted on figure 3. lt was carried out in a zone 80 metres long of turbulent torrential flows in a conduit of constant cross section (no major enlargements) The fitting of the experimental curve onto the Sauty model is excellent. This experiment empirically confirms that the turbulent flows are not the cause of the lag effect. EFFECTS OF THE VARIATIONS IN CROSS-SP.CTION The "micro tra cing" (B) was designed to evaluate the effect of an enJargment of the flow cross-section on the dispersion and tailing (figure 3) The tracer was injected upstream of an enlargement ( 1 5 m deep, 4 m large) The breakthrough curve just before the enlargment does not show any tailing effect. Nex.t to the outflow of the pool the curve shows an important tail. The presence of a large and deep cross-section (pool) appears to produce a tailing effect. Other similar experiments have been carried out, all showing the important influence of variations in cross sectional area on tailing effects Tracing A :o a E 12 0 6 'if O. a: 0 0.2 0 1--
PAGE 163

RELATION BEIWEEN TAIUNG AND DISPERSIVITY Tailing effects appears to vanish along certain flow sections where the apparent dispersivity markedly increases For example the "micro-tracing" (B) has been measured at station 2 situated 850 m downstream of the injection point. The fit with the Sauty model is perfect at this station, despite the strong tailing effect measured some hundred meters upstream (figure 3 b). Along the same section, the dispersivity increases markedly The pathway is made up of alternating enlargemements and smaller conduits with sections of torrential flow The flow through this pathway bas "homogenized" the tracer plume. These observations seem to indicate that one enlargement provoques a tailing effect but several successive enlargements can cause effects which cancel out Intuitively, the relation between tailing effect and scale effect of the apparent dispersivity can be considered as follow : tracer molecules slowed in whirls of a first enlargement -or poolcan use a faster flow route across a following enlargement. After traversing several pools, the molecules will have all passed by slow and fast zones, and their average velocity distributions will have regained the form of a Sauty curve, but the overall apparent dispersivity will then be higher One can simulate the statistical decrease in the lag effect across several enlargements using a cuive of residence time distributions (RTD) of a tracer cloud which has crossed one enlargement (large tailing effect), and then convolve it several times with itself. Figure 5 gives an impression of the form of the breakthrough cUJVe as a function of the number of convolutions (number of pools). After six enlargements, the cUJVe has a form very similar to the theoretical cUJVes of Sauty the dispersivity measured on the rise of the cUJVe increases from one cUJVe to the next. o 24 1 enlargement 2 enlargement s 6 enlargement s ---......__ .. .. o.J..__J_~~.....__..., ____::, __.:; _______ .._ __ 0 20 40 80 80 100 Temps (T ( 1 ~nlarg 2 enlarg 3 anlarg 6 anlarg (au = X-'00) = 2au) = -luu) (tFl2ou l 1 0 0 8 J o a 'i : (.) 0.4 02 0 0 1 0 2 0 3 0 4 0 5 0 6 0 8 1 IT Figure 5 : "Black box" model used to simulate the evlution of the lag effect after several enlargements. After 6 enlargements, the lag effect has disappeared, and the dispersivity increased. 4. Conclusion The observations made along the length of the undeground river Milandrine document the large variability in velocity (between 17.5 and 365 m/h) and dispersivity (between 0.4 and 49 7 metres) of the tracers along the length of their route. All the experiments carried out allow us to make four observations I) A scale effect is associated with the dispersivity measurements which increase with the scale of observation 2) Turbulent but regular flows on a channel of constant cross-section do not give rise to a measurable tailing effect the experimental curves obtained at the end of such sections fitting perfectly to the Sauty curves. 3) The presence of a large basin (stro ng variation of the flow cross-section) causes a clear lag effect. This effect is very probably linked to the strong velocity gradients found inside the basins between slow and rapid zones 4) The succession of big basins seems to cause an increase in dispersivity and the disappearance of the lag effect: one single basin produces a lag but several successive ones produce effects which cancel out tqwards the tailing, but produce a substancial increase of the apparent dispersivity Tailing effects are seen very often in the breakthrough curves measured at karst springs They might come from the existence of several independent pathways, from a transient flow regime during the experiment, or from the injection con ditions (strong tailing effect between the injection point and the point of arrival in the karst conduits). This paper shows that they might also be produced by strong heterogeneity of the flow cross-section along the karst conduits. References BNER, P. 1993. Etude pbenomenologique et numerique de la propagation de polluants miscibles dans un milieu a porosite multiple. These Sciences Univ of Liege, Belgium, 389 p. DE MARSILY, G 1986. Quantitative hydrogeology. Groundwater hydrology for Engineers. Academic Press, 440 p HAUNS M. ; HERMANN, F. & ATTFJA 0 1997. Application of a CFD model to cave river hydrodynamics This volume JE'ANNIN, P .Y & MARECHAL, J .-C. 1996 Lois de pertes de charge dans les conduits karstiques : base theorique et observations. Bulletin d h y drogeologie No I 4 149-176 MALOSZEWSKI P. & ZUBER, A. 1989 Mathematical models for interpretating tracer experiments in fissured aquifers In: "The application of isotope techniques in the study of the hydrogeology of fractured and fissured rocks" IAEA : 287-30 I MARECHAL, J.-C. 1994. Etude et modelisation de l'hydraulique et du transport dans Jes drains karstiques Master Thesis in Hydrogeology Univ of Neuchatel unpublished 128 p. MEus Ph 1993 Hydrogeologie d'un aquifere karstique dans les calcaires carboniferes (Neblon-Anthisnes Belgique ). Apports des tra~ages a la connaissance des milieux fissures et karstiques These Sciences Univ. of Liege Belgium 323 p PrAK T & TEUTSCH G. 1994 Forced and natural gradient tracer tests in a highly heterogeneous porous aquifer : instrumentation and measurements J. Hydro[. 159: 79-104. ROSSIER Y & KIRALY L. 1992. Effel de la dilution sur la determination des dispersivites par interpretation des essais de tra~age dans les aquiferes karstiques Bull. du Centr e d hydrogeologie de Neuchatel 11: 1 15. SAUTY, J.-P. & KINZELBACH, W. 1988 On the identification of the parameters of groundwater mass transport. In : (E. Custodio et al. ed.), Groundwater flow and Quality modelling: 33-56 SAUTY, J .-P. 1977. Contribution a ]'identification des parametres de dispersion dans Jes aquiferes par interpretation des experiences de tra~age. These Universite scientifique el medicale et Institut National Polytechnique, Grenoble, 157 p SEILER, K .P ., MALOSZEWSKI, P. & BEHRENS H 1989. Hydrodynamic dispersion in karstified limestones and dolomites in the upper Jurassic of Lhe Franconian Alb F.R.G J. H yd ro/ 108 : 235-247 152 Proceedings of the 12 th Internat i onal Congress of Speleology 19 97 Sw itz erland Volume 2

PAGE 164

Obtaining information on fracture geometry from heat flow data Rudolf Lied.I, Sven Renner, Martin Sauter Applied Geology, University of Tuebingen, Sigwartstrasse 10, 72076 Tuebingen, Germany Abstract ln this paper it is shown that changes of water temperature recorded at the outlet of a fracture depend on the spatial structure of the fracture geometry if the discharge is time-variant. This result is used to characterise the geometric properties of a conduit network in a Swabian Alb karst aquifer. Zusammenfassung Dieser Beitrag zeigt daB Anderungen der Wassertemperatur am Ausgang einer Kluft voo der riiumlichen Struktur der Kluftgeo mecrie abhiingen, sofern ein zeitabhiingiger DurchfluB vorliegt Dieses Ergebnis wird verwendet, um die geometrischen Eigenschaften des Rohrenoetzwerks eines Karstaquifers der Schwiibischen Alb zu charakterisieren 1 Introduction For modelling purposes fractured aquifers are frequently assu med to consist of two hydraulic systems which are coupled to each other : a network of large fractures, e g karst conduits, with high flow velocities and low storage, and the rock matrix with slow flow but a comparatively high storage coefficient. The high vulnerability to contamination of fractured aquifers is mainly a result of the fast transit time of dissolved substances in th e fracture network However data on these fractures are rare due to the inaccessibility of these pathways For single frac ture BENDERTIT E R et al. (1993) could characterise flow in a car bonate medium by using heat transport data ln this paper it is shown that unsteady flow conditions, e g during recharge events, are necessary in order to get insight into and to obtain quantitative information on the geometry of the fracture system This approach is applied to a karst catchment in the Swabiao Alb (SW Germany) The paper can be regarded as a brief summary of the results achieved by RENNER (1996) 2. Modelling Approach lo order to establish a mathematical heat transport model it is convenient to consider a single fracture first. It is assumed that the fracture geometry can be approximated by a parallel-plate model, i e V II A 2 (1) with V volume o aperture, and A contact area of the fracture According to the conceptual model shown in Pig 1 heat is tr a nsported by convection along the fracture and by conduction in the rock matrix Conductive heat transfer is modelled by oTS ).s iPTS (2) at = c,,p,, az 2 (CARSLAW & JAEGER, 1959) with T temperature in the matrix, p rock density c specific heat capacity, and As heat conduction coefficient of the matrix For heat flow in the fracture conductive processes are neglected so that the transport equation can be written as art +v( t) art=_!:__!!_~ ars (x, 0 t) ( 3 ) at ax CtPt V az (CARSLAW & JAEGER, 1959) with Tr water temperature, c r specific heat capacity and Pr density of water. ln eq (3) the convective velocity v depends on time because unsteady flow conditions are considered The right-hand side represents the heat exchange between the fracture and the rock due to the tempera ture gradient at the surface of the matrix (z=O) z 0 heat conduction in the rock matrix heat convection in the fracture t X ----1 v=v(t) Figure 1: Conceptual model of flow and h eat transport in a single fracture with constant aperture ln order to complete the mathematical model appropriate ini tial and boundary conditions have to be specified In the rest of this study heat transport problems with initially uniform tempera ture distributions are discussed, i. e (4) The boundary conditions for the conductive heat flow in the ma trix are given by The boundary condition at the fracture inlet can be written as 6 th Conference on Limestone Hydrology and Assured Media 153

PAGE 165

(6) In eq. (6) Tp(t) denotes the time-dependent water temperature at the fracture inlet. RENNER (1996) solves eqs (1)-(6) numerically u sing a finite difference technique His approach also allows for (i) non-uni form fracture apertures, (ii) fracture intersections. (iii) inflow from the vadose zone. With these model extensions the flow ve locity in the fracture also depends on x 3. Parameter Studies In a first example heat transport is studied for the three frac tures shown in Pig. 2 having identical volume and contact area but different apertures (RENNER, 1996) The water temperature at the fracture inlet has a constant value TP. The flow rate is equal to Q 1 until t=ti, then it drops to Qi< Q 1 The breakthrough curves (BTCs) at the fracture outlet are identical until t-V /Qi =t 1 For larger times they differ significantly. o-.iic l~ci1EJ o~::= ===~07~2 fractures o~,--1 --~ fr input BTCs T p Q) a. E Q) t=0 T o --ti-m-;e t t=O t, time t 0 I' a. 0 It::, Q) :i iii .... Q) a. E 2 0 1e2 1e3 1e4 1e5 time t-V/O2 (s) Figure 2: BTC s for thr ee different fractures for time depetidem discharge In order to explain this behaviour it sho uld be mentioned first that the temperature must decrease after the reduction of the dis charge leading to an increase of the residence time of the water in the fracture so that more heat is transferred into the rock ma trix Secondly, the decrease of temperature depends linearly on the surface volume ratio AIV= 2/o (eq (1)). As a r esult, the temperature decreases more slowly for larger apertures and the BTC of fracture 1 starts to decrease less significantly after V /Qi =L1 because the wider part of fracture 1 is close to the out let This reduction of water temperature is enlarged as soon as water coming from the narrow part of fracture 1 reaches the outlet. Por fracture 2, which is a mirror image of fracture 1, the temperature starts to decrease more steeply because in this case the narrow part is close to the outlet. The BTC of the uniform fracture 3 shows an intermediate behaviour. Finally, all BTCs approach a BTC that would have occured if the discharge had been equal to Qi for the who l e time This example clearly shows that temperature changes of spring waters emerging from fractured and karstified aquifers may reveal information on the fracture geometry only if there is unsteady flow which is likely to occur during recharge events. The non-uniformity of the fracture is rendered by the shape of the BTC Next, the importance of fracture intersections is demonstrated (RENNER, 1996) Pig. 3 shows some BTC s for a system of two interconnected fractures which have the same length but different apertures (0/0 2 = 1.86). Por an increasing number of intersections it can be seen that the BT C of an "e quivalent fracture" with con tact area A=A 1 +A 2 and aperture o=(o 1 +oz)/2 is approached 0 8 a. t 0 t--;0 6 t:.. ::J ro a; 0.4 c. E 0,2 n=O -/ -n=1 n=4 ./ -n=9 . equivalent ---i nterconnected 0 ...__ ............ __. __ ...__...._ __. __ .,__ _. 1E-2 1E-1 1E+0 1E+1 1E+2 1E+3 1E+4 time t V/Q (s) Figure 3: BTCs for two interconnected fractures with ,i ititersectiotis (ti=0 1,4,9) It is obvious that an increasing number of intersections must have the same effect on heat flow as a permanent mixing of wa ters along two parallel fractures and that this system can be re presented by a single fracture with equivalent properties REN NER (1996) presents a similar example showing that an equiva lent fracture can also be found for a system of two fractures with the same aperture but different lengths provided that there is a sufficient number of intersections. 4. Application The approach outlined above is applied to the Gallusquell e catchment which is part of the Swabian Alb karst in SW Germa ny (Pig. 4) Time series of spring discharge and water tempera ture exhibit a strong correlation (Pig 5) A very important fea ture of the temperature record is that recharge events always lead to a decrease of spring water temperature with respect to the background value of 8 75 C 154 Proceed ings of the 12 1 h Internat i onal Congress of Speleo logy, 1997 Switzerland Volume 2

PAGE 166

O~km I I I \ ~----I Borehol es IJ Spring Trac er t est Equipotentia l s N I m above sea leve l Catchment a r ea Fig u re 4: G all us quell e c aJ c hm e 11 J a re a ( RENNE R & SAlffER, 1 997) This observation can be explained by a conceptua l model (Fig. 6) which is based on the assumption that there is an epikar stic reservoir with a low vertica l permeability {RENNER & SAU TER, 1997). Accordingly, the recharge water remains in this do main long enough to adapt the rock temperature which is about 6 C Then, the water flows very rapidly through vertical shafts so that there is no temperature equilibration with respect to the geothermic gradient. R eaching the groundwater table the water temperature is always less than 9 C These temperature differences are found again in the time series shown in Pig. 5 Similar observations are reported by JEANNIN (199 0 ) A model of heat transport in the aquifer is established by ma king use of the equivalent fracture approach mentioned at the end of section 3 The fracture geometry is determined by calibrating the model with respect to spring water temperature measured du ring a recharge event in June 1988 (Fig 7) It is found that the temperature diffe r e n ces at the spring can be simulated quite ac c u rately by ass umin g that the eq u ivalent fracture consists of three segments {RENNER, 1996). The volume V, the aperture o and the length L of each fracture segment are summarized in Pig 8 0 5 ,-----------------0 4 0 3 f<] 0 2 0 1 0 1E + 6 I June 88 I m ea sured -s i m u la te d -=---' 3E+ 6 4 E+ 6 6E + 6 time (s) 7E+6 Figur e 7 : Mod e l c alibralio11 wit h r es p ect to s p ri ,i g waJer te mp e ratur e These parameters are now used in order to simulate spring water temperatures after a recharge event in Pebruary 1990 The absolute values of th e deviation from the background temperature are shown in Pig 9 where only the input temperature T P has been used as a calib r ation parameter It can be seen that the mo del results and the measured values match very well. The slight discrepancy at the peak of the B TC may be anributed Lo the in fluence of the fluctuations of the groundwater level which are neglected in the model {RENNER 1996) S ome more examples of model validation, provided by RENNER (1996) and RENNER & S AUTE R (1997), also show a c l ose agreement of the simulations and the measure m ents so that the approximation of a karst con duit system by a single equiva l e n t fracture seems to be justified for this aquifer. 2 5 .-------___:.-------------------------------------, 2 1 5 0 5 discharge (m3/s) o L-------=--=::::... ___ ___;_ ___ :....:..:=========--==::::.. __ ....._ __::: .=.._..:._ __ __J 8,8 ------------------------------------------, 8 6 8 4 8,2 temperature ( ) 8 L-----------------------------------------------------' 1 3 88 15 9 88 1.4 89 Fig ur e 5: S pri11 g water te mperaJur e a,id discharg e aJ the G allu s qu e ll e k a rs t s pri,i g time 15 10 89 1 5 90 15 11 90 6 '" Conference on Umestone Hydrology and Fissured Med i a 155

PAGE 167

epi karst vadose zone phreatic zone epikarstic reservoir shaft gro u nd wate r level condu i t Fig ur e 6: Co11 ce ptual mod e l for th e c han ge i 11 t e mp e ratur e of see pa ge w aJ e r V=20000m 3 V=l0000m 3 V=l2000m3 o=O lm o=0 12m 6=0.4m L=5000m L=5000m L=2500m Fig ur e 8: E qu iv al e 11 t fra c tur e for th e G al lus qu e ll e ka rst c aJ c hmen t 5. Summary and Conclusions I t has been demonstrated that the geometrical propert i es of a karst conduit system in the Swabian Alb can be characterized by analysing spring water temperatures. It is important to note that the temperature data have to be recorded for unsteady flow con ditions, e g. during o r shortly after recha r ge eve n ts. As has been show n also by numerica l experiments, tempe r atu r e signals at the spring can provide information on the fracture geometry only if the discharge is time-dependent. I n the examples p r esented above the concept of an equivalent fracture representing a fracture network is employed This ap proach is valid only if there is a s u fficient number of fracture in tersections Of course, this approach leads to "bulk values for the geometric paramete r s Further research will have to investi gate whether more detailed information can be obtained by using a discrete description of the fracture netwo r k An approach based on this idea is described by H OCKJNGHAUS et al (1997) A con ceptual model of heat transport including the vadose zone is be ing developed by J EANNIN et al (1997) .__~ ____ ....1,_.;.a __ n __ n~ua~l--t;,;;e,;.;m;,::.P fluctuat i ons temperature epikarstic reservoir geothermal grad i ent tem pe r ature of w ate r r un ni ng t hr o u g h h ig h capacity shafts t empe r ature d i fference between fresh w ater and aquifer water 9 (aquifer temperature ) 0 ,7 -------------------0 6 0 5 0 ,4 u 0 I0 3
PAGE 168

Special spring discharge changes in the Aggtelek karst region in Hungary I... Maucha Water Resources Research Centre (VITUKI) Plcl095 Kvassai Jen6 ut I. Budapest Hungary Abstract The Aggtelek Karst is renowned for the number of its special hydrologic features The surges of the Lofej spring result from the combined action of three siphon systems These waves are transferred to, and superimposed upon. as an attenuated pulsation on the yield time series of the similarly siphon c:ontrolled Nagy-Tohonya spring. Tidal inOucnccs on the fluctuations of karst ,wter table, spring discha~ and lithoclase fluctuation \\ere noted with surprise The influence of surface air temperature variations was also detected in spring discharge fluctuations during the SflO\\lnclt period No specific effects of earthquakes and atmospheric pressure could be identified, but statistical analyses have shm\11 these to play a role in triggering discharge changes The precipitation-yield correlation was also studied and its mechanism explained as the se\'enth influence Zuzammenfassung Des ~elelcer Ks:rstgebcbiet ist von vielen spe:zielJen hydrologischen Erscheimmgen bekannt. Die Ausbrtlchc der Lofej-Quelle \\mien von einl.m l.>rcil.-r-Saughchcr-Systcm verursacht. Dicsc An
PAGE 169

Q SIPHON OUTBURSTS OF THE L0FEJ-SPRING Tht\-.dl,m,ln ,. 11 10 1974. Z3 Z4 zsl zel 21l 2a Z9 :! 30 I 31 j 11,1 I Z ., Qmu Vmin(daly-I 1. I'--... Activity of the afphon A. B, malriy 0,,_, 3300 I disdwge) ..... P I .. Qmln 1000 I EI < I ilil a: ... m ,r r, ~m1 rf, :~il ~~, lNhM W\ >al c., ~Ew < c,:, ll.1 ;,:,ou. 7 10 11 1Z 13 14 YfflU ~~mln 18 111 < l2 r;-,, Acllvtty of the C only 0 ,,_, I and A,B,C a: .~mln I 400 1 I l w (9 ~'+.. M ::c: > I '\JI~ -~ c,:, < I Iii 10 11 1Z 13 the-=~ A."': and A,;~C 11 111 I ZV Z1 w ....,,, Activity C 0mu MOO Vmin ::r: ..J I EI 3. l fJ~ lr,I :k. u ..J l l l l l -= ', 1lil ;io ZI D Z3 Z4 Z5 ze 27 :ze Zlil au 31 03 1975. 4. Activity of the aiphcin A.B,C 0 mu lXlO Vmln 0 ll'INII 380 I hJ a:: .u .... 0mln 300 I L ll .. .... J .. w "8 .. I= ""'"""'~ I ......... P"''" r 0 E, z 4 " ., TI 1Z .I 14 07. 1975 < I Actlvtty of the llpllOI, A.B,C Q mu -450 I( mln 5. 0 .....,, 110 I ll l 0min .30 1 l I l z iu .. \. II, < w 02. 5 ., 8 Iii 10 TI 12 14 ... .I 18 17 I 6. Activity 0 the siphon C only 0mu <450Vmin 0"I I ... 0min 30 I 1 .. I DI I. .. ,._ .. a: I :zo Z1 Z2 Z3 Z4 Z5 I 28 27 28 Z9 30 .1 31 OL1 w I 08. ~5. 0mu l/mln El ktMty of ttw aipl,on C only I 0--10 I < 0mln -20 I :r; ": : 09. 1975. s 4 a e 1 e 10 11 12 1:1 I 14 15 8 0 mu L/mln Qn-, I 0 I lnflltratlor otthej'c~ 0mln 1-10 I ..J I "" 0 0 Thooaand 1/min SIPHON OUTBURSTS OF THE NAGY-TOHONYASPRING a: 08" 30 05.1 Z 3 ~=1[0 'l0 10 w .. .. !. E.. 04. 1978. A < "' )/ -"'-""< vJ ,. _J 9 \I .,_,...,_____ ~l,...r-::c: ..... 0max /mln (dally_..gad~) I c., I 0 maall 9990 I 5= Qmln 3000 I a:: I w .. r 13 14 15 17 18 11 I .. r Z1 zz 23 24 E" 05.il976. J\ < ., 0max l10090 I 10. I A A a- 5920 I \ --vQmln 4580 I I ___, ,._ I A a: w E.. 06. 1976. 23 Z4 Z5 28 Z7 28 21 :so 07.1 2 3 4 5 < 11 Qmu 15190 I\ ., (~ I I 11. Q..-i 3980 I 0min 3170 I I A 0 I ..l Fig. 2. Recorded disclaarge tune series of siplaon conlrolkd springs during higl,water, medium and ww jloK'S sequence of small releases from the small siphon C is suddenly interrupted by the draw
PAGE 170

while in the second case by the siphon C alone The cause thereof is the circumstance that the crest of the di!;Charge limb of siphon C' is ot a lower elevation that in the case of siphon n. which needs a steep increase in now to trigger release. It is for this reason that at low Oows the siphons A. C are full and the water spilling or leaking over siphon C \\ithout release feeds the spring (be erruptions from the Nagy-Tohonya spring which appear al considtrnble higher mean Oows and discharge volumes larger by one order of magnitude (2500 m 3 ) are illustrated by the times series 9-11 in Fig. 2 The highest frequency thereof is four releases per week As implied by the rntio of the mcnsured siphon volume and the spring yield. further by the water temperature drop follo\\ing the front of the erruption ( Gidoros, 1971 ) the siphon is believed lo operate on a by-pass The higher frequency of em1ptions (identified on the L6fej spring) is presumably limited by the circumstance that at higher water le\'els in the main stem the lateral arm containing the siphon becomes inundated 3. The phenomenon of indirect siphon effect Tracer studies and the observations of P Szilny alike have provided hydraulic evidence demonstrating that the yield fluctuations of the Nagy-Tohonya spring arc affected by the siphon releases of the L6fej spring As will be perceived from the time series 9 10 and 11 in Hg 2 the em1ptions of 5-15 hours duration and a lew hundred 1/s discharge (A) follow with an average delay of four hours the em1ptions of the L6foj spring This is illustrated in Fig 3 The flow diagram of the phenomenon is shO\m in the upper part of the figure the lower diagram demonstrating the close correlation between the appearance of the phenomenon considered on the lwo springs These pulsation s are the strongly attenuated micro-em1ptions of the L6fej spring emptions which travelled a distance of 3 km and ore superimposed on the yield variations of the Nagy Tohonya spring and are capable of modifying appreciabl y the hydrograph of the erruptions (A') Siphons spaced longer distances apart are capable of inflm .. 'llc.:ing the yield changes of karst springs in this Wa\' 4. Tidal phenomena Statistical anal y ses hove revealed the erruptions to start at 6 12 18 and 24 hours \\ilh 30 % frequency at both siphon controlled springs mentioned above (M1ucha 1966 1977 Barth 1967) In round moon phases these are the times when the extreme values of lunisolar changes occur, so that this observation was the first si[?Jl implying that tidal effects in underground streams are also involved in triggering yield erruptions The first assumption was corroborated by the finding that the si:x-hour tidal fluctuation most frequently registered on non siphon controlled springs has similar time-statistics (fig. 4) !be six-hour period is impossible lo register on the L6fej spring, where the original shape of the yield hydrograph is modified by the siphons 'Ille existence of the karst tidal phenomenon could be demonstrated by the registered tidal-period fluc.:tuations of the karsl water table (Fig 5 and Gerber, Errupt i ons o f the L6t'e j spring 10 .. = ~11 0 ,.J u 10 -= 0 .. I : ,; I I;i .... .,,. :: i :.:; ... ,, .. ~I I : ~ .: o~ ....... L.;,"T'T ............... ,......._....,....,~O 5 10 11 :! U Hours of th~ Nagy Tohonya spring Fig. J Co"elation bdweat the erruptions of the Lofej spring md the initial times of yield variatioru of the Nagy-Tollonya spring Erruplions of 1he L6fej spring ~k 05 I 970 2 4 5 6 7 8 9 days Fig. 4. Starting erruptions on the Lof q spring by tidal tYll'iatioru in the stream filling the siphon. The yield t'flriations of the Lof q spring inaccasibk to measurnnent are rq,laced by tidal flactuation.t nf the Kis-Tohonya spring. The potential starting times of siphon release are indicated by arrows. H ,., I I I \ m ,o ,~ I 1 1 I I I JI I I "' I; I ~ I I I I 1 r "' "' 1 I \ J I I "'" ii I 1 i 1\ i : ~ j i 1 1 1, \f\ ~ ,,, I I: 1I11 1 I~ 1 n o ~ j I I I 1 I J / I I I I I I I "' j ii I I I I 1 f; i\ n .. i I i; : u : i 1' VI\! \ I ::: j I I I :; J "' 1996 ?J ~ l r ch Fig 5. Tidal fluctuations ,if the kantwater ln't!l in the ohser,'tdion well sunlt in dolomite aboi'f! the Kis Tohonya spring during February-March,J,96 6 1 Conference on L i mestone Hydro l ogy and Fissured Med i a 159

PAGE 171

1965), spring yields (Fig. 6) and karst void volwne, called also lithoclase fluctuation (l<'ig. 7, Maucba, 1966, Gadoros, 1969, Maucba-Sarvary, 1970) The existence of a close correlation between these changes and the Junisolar changes in the strength of the terrcstriul gravitational field (Maucba, 1977) provided conclusive evidence of the phenomenon and of the tact that tidal defonnations of the lllJStified rock trigger hydrologic changes by changing the void volwne 5. Earthquake effects The potential role of earthquakes in triggering spring yield fluctuations was studied Statistics were analysed to detennine how many of the 355 remors registered over the 1965-69 period coincided with the various types of yield changes of the Nagy-T ohonya spring The different hydrograph types have been plotted on the abscissa of Fig. 8, while the frequency by nwnbcr of the tremor scale corresponding to each type on the ordinate The white colwnns indicate the simple frequencies, the black ones the frequency divided by duration of the types of yield change The largest Wlit nW11ber of earthquakes (21) was found to coincide with the "start of siphon activity" (Fig. 8, time series 9) water yield oscillations period (14) and peak yields (5) Of these only the first, conspicuous frequency hos been adopted as reliable, in that level oscillations and peak yields can be produced also artificially on the record by some disturbance of the rm:asuring system The yield wave front, on the other hand, may actually be related to earthquakes Another analysis has revealed, in fact, that the steep tidal yield drops after the erruption peaks on the Kis-Tohonya spring (lt'ig. 6) appeared with 40 % frequency at the time of earthquakes These changes, detectable also on the Nagy-Tohonya spring (l<'ig. 2, initial section of the times series 9), are capable of causing water-level drops in lhe main slem of the spring, which trigger such phenomena prematurely (c.f the facts in Chapter 2) By relieving suddenly the tensile stresses built up on the rising limb of the tide, the earthquakes increase the void volwne in the karstic rock, the yield drop induced thereby triggering siphon activity 6. Atmospheric pressure effects The variations in atmospheric pressure have also been studied for their possible influence on spring yields, because no perceptible evidence ( comparable to that of tidal phenomena) was found of 2~7 day fluctuation in the yield records In Hungary it was Gerber, ( 1965) who first demonstrated opposite variation of the kllJ'St water table and air pressw-e, but this could have been caused by infiltration in low pressure period, too For this reason statistical analyses wear again resorted to explore any possible influence of this effect on the yield of kllJSt springs Since the SL"llsitive transducers mounted on the siphons have made the detection of tidal effects possible (Maucba, I 970), so that the same approach was used in the attempt at demonstrating the barometric pressw-e effect. For this purpose the starting times of the 216 siphon triggered erruptions of the Nagy-Tohonya spring over the period 1965-69 were checked for the frequency at which these coincided whit the diverse phases of the barometric change periods In Fig. 9 the abscise show the phases of the full period of barometric pressure change, the ordinates the frequency scale of occurrence Two distributions are presented in the colwnn diagram, the Q~IDJittt8. I .I. I .I. IJ IJf I. I. I .I]. I 1 ll I. UJJULJ Fig. 6. Tidalyiel4fb,clualion observed regularly in tl,e Kis-Tol,onya spri11g on tl,e recessiu11 limb of erruptions. A-A: steep yield drops, 8-B: app. 6-l,111u /b,cllUllion, C-C: yield fb,ct11Uliu11 lull'arib tl,e e11d uf die err"P1iun (April, May, 1979) 1971 Apr a 9 Fig. 7. /,illu>c:lase Jb,clualio11 regi.Jtered ;,, tl,e I "~ Imre cave. The time series are fullowing: ltmuspheric pressue, precipilt11w11, Kis-Tol,onyu spril,g )'ield, sl,earing movement uf N-S directiu11, l't:rliciu :!il1euri,.g move,r,ent, W-E dilatation mu,-eme,u, eurtl,9ualt.es, IMni.solar gravitation cht11,ges. ( April, 1971) Frequency 114 D t-llEQl"E."IC"Y BY "'UMIU 70 TI C"O UECTED FIUCQUE..,.n .. WATER USCILL\TIO.~ .. .. ,..!!! J_ ;! ,. I! HOUR YtEUJ l"EAIC. !l .. 14 10 .i-l 7 r.:. 5 2 ..:. ,. ,. z Q 0 t Q ffi "" 0 > Ii: z .1i 0 6 z Q ,., 0 0 ::, Q 0 ... .. "' g ... .. < "' ;:: < z 9 "' "' "' "' < g ... .. z ... "' "' ... "' "' 0 ... ..., ::, < "' w :,: 0 ::, ... ..., ... ... ... z ... "' ..., < ;;; z 0 "' G "' ;:: > ::, 3: 3: ... ;l > ;:: :s 0 ... ;:: g 3: a ... ::, < ... "' ;;; "' '.::: 3 "' "' w :i: :,: "' z ... ... < 3: Fig. I. Earth9uake fre9uency dMring periods of different typa of yield cl,ange of tl,e Nagy-1'ulaunya spring (1965-1969) days 160 Proceed i ngs of the 12 1 h Internat i onal Congress of Speleology 1997, Switzerland Vo l ume 2

PAGE 172

foll line being the frequency distribution of ull erruptions, the dashed line that of the 35 1-'ITUptions in the winter period at temperatures below the freezing point. The second is considered relevant in that it docs not include the elfoct of infiltration caused by rains falling in low pressw-e periods The changes in atmospheric pressure were thus found to influence the yield of karst springs in that the highest erruption frequency was fow1d in the periods of steepest pressure drop around the point of counterflexure In the total distribution 20 % in the sub-zero distribution 43 % of the erruptions were noted in these periods at the 10-11 12 phase values lbe result obtained for the infiltration-free period is comparable to that arrived at by Gerber Moreover, on a single occasion we have succeeded in registering also the lithoclase change caused by a rapid pressure drop immediately before a stonn In this case the pressure drop was found to increase the v oid content decreasing the spring y ield This is believed to imply that changes in atmospheric pressure induce changes in the spring yield by changing the volume of voids (lithoclase fluctuation) The only safe conclusion is however that the air pressure ellect is capable of triggering syphon release even in periods without infiltration 7. Air temperature effects During snowmelt periods daily spring yield fluctuations (24-hour period) were repeatedly observed The phenomenon occWTed during anticyclone (high pressure) periods, when infiltration caused by snowmelt around noon is interrupted by freezing during the night. A typical sequence of these cycles i s illustrated in Fig 10 the days figuring on the abscissa the scales of the corresponding time series on tl1e ordinate axis The upper two curves indicate the variations in air temperature and tl1e reduction of snow cover thickness at the nearby J6svafo meteorological station The lower two curves are the time series of karst water table and the yield in the water system of the Nagy-Tohonya spring 8. The rainfall effect The analysis spring yields bas raised the question of the mechanism by which the precipitation falling is transformed into spring yield Relying on these studies the actual spring yield is interpreted as the cumulative sum of the discharge time series of erruptions originating from rainfall groups (in winter snowmelt ) during the antecedent months and subject to different influences The major factors involved in the development of erruptions are : 1 the infiltration percentage (1%) which varies St:aSOtui!ly as a result of changes in air temperature and vegetation cover 2 the per cent deviation from the normal mean (So; 0 ) of k.arst saturation during the period preceding the particular rainfall group and 3 the per cent ch.-crcase (Do; 0 ) of the peak yield caused by depletion up to the day of the yield under consideration A simple method of taking into account the major factors influencing the precipitation effect i s illustrated in Fig. 11 The depths of the rainfall groups (in mm units) during the two months preceding the day of the spring yield considered are reduced by the foregoing percentages, than summed up The resulting swn is termed the precipitation effect index P e The pairs of index nwnbers aud spring yields considered were used to plot the correlation points of the diagram shown For Mui mu m I IOI plC\e '"'tbunts 11, p.e *l~N Fre-q u ency I Po im of Mi n imum inll ex ion r,D 111r! bu1!on ~r J 1 I r I ..... r, t;l J J J n;u pl i ~ns I D m ribu 1io n o f errupt i oas lr, I ; at: s ubz ero t empera tu res I I m i I r ru L. LJ u --, I i: r I j I .I ,. h I .. I I lL.J I I ,, r J i J I r ,,. l 2 3 4 s 7 8 9 1 0 1 1 12 1314 IS I ii 17 18 19 2021 PHASE VALUES Fig 9. Fre9,uncy of erruptions on tie Nugy-Toionya spring "'4ring diff erenl phases of almospheric pressure ciange (1965-1969) A e C 0 +10 0 -10 ~c! o .c E E 4 ~:: .,: i5 a. 3 "0 l ::. 0 ,. .!! ; 0 a. e 3000 .. l February 1171 11 17 18 19 20 21 Days Fig. JO. lnf"'ence of air temperalure on the lant woler tabk and spring yielJ/ in the Nagy-Tol,onya spring system (February,1976 It Q Q Pe = L Pi-l"iS,,i-D.,,i ( mm) / m in mm /y ear i=-1 40 a:,a lO 100 ., e_l> CU .c 10 u "' 'o 400 b -~ 0 10 200 0 ............. -.-...--,.....,... ......... ......,r-r-.--r-i,........-.-0 o u P I ndex of precip i tat i on eff e ct Fig. J J. Correlation between the daily spring yields tl,e precipitation effect index in tl,e case of ,,,e Nagy Tohonya spring (l96S-1969) 6 '" Conference on Limestone H y d r o l og y and Fi ssu r ed Med i a 161

PAGE 173

plotting the diagram 80 pairs of figures (excluding s110\\mell) wtm: taken from the 1965-69 yield lime series of the Nagy-Tohonyu spring The precipitation i ndex hus been plotted on the ubscis::ia, the yields on the ordinutc uxis lbe fitting curve is 11 parabola up to 18 000 llmin yield the potential infiltration increasing exponentiully with precipitation Beyond this limit tbe curve deflects to the right, reJlc:cting the effect of surfui:e rw101T, whii:h n:
PAGE 174

Parameter identification in double-continuum models applied to karst aquifers U. Mohr l ok J. Kienle, G. Teutsch Applied Geology, Geological Institute, University of TU bingen, Germany Abstract One modelling approach which proved successful in describing the groundwater flow within karst terraines is based on the double continuum concept. This concept was first introduced by TEUTSCH ( 1988) and subsequently used by TEUTSCH & SAUTER ( 1991 ) SAUTER (1992), LANG ( 1995) MOHRLOK ( 1996) and others to describe the ambivalent characteristics of karst aquifers. However the approach has the drawback that the double-continuum model parameters can be determined only through model calibration (inverse approach) i.e. so far the model parameters cannot be related directly to physical field measurements Therefore, in order to develop a better understanding of the physical significance of hydraulic parameters within double-continuum systems, a detailed numerical modelling study was conducted For this purpose a number of synthetic but realistic karst aquifer network geometries were generated and analysed The response of the karst network to recharge events was simulated using a detailed discrete fracture flow model with the resulting head and spring flow variographs being subsequently assumed as field measurements This measured data was then used fo r the calibration of a double-continuum model and the resulting parameters were compared to the original karst network geometry data This comparison was used to develop mathematical / physical relationships between the discrete karst network geometry representing reality and the double-continuum parameter representation of it. Kurzfassung Ein bisher erfolgreicher Weg zur Quantifizierung der Grundwasserstromung in Karstgebieten ist unter dem Stichwort Doppel Kontinuum Ansatz bekannt geworden Dieser Ansatz wurde zuerst von TEUTSCH ( 1988) eingefiihrt und anschlie8end von TEUTSCH & SAUTER ( 1991 ) SAUTER ( 1992) LANG (1995) MOHRLOK ( 1996) und anderen dazu verwendet, den ambivalenten Charakter der Grundwasserstromung in Karstaquiferen zu beschreiben Leider konnen dabei die Modellparameter jedoch nur durch Kalibrierung (Inversmodellierung) ermittelt werden und sind somit sehr schlecht mit den im Gelande tatsiichlich ermittelten Parametem korrelierbar. Um dieses Problem zu losen wurde im Rahmen eines Forschungsvorhabens eine detaillierte numerische Simulationsstudie durchgefiihrt, bei der ein synthetischer aber realitiitsnaher Karstaquifer einschlie8lich des Karstrohrennetzwerks generiert und analysiert wurde Das Ansprechen dieses synthetischen Karstsystems auf Niederschlagsereignisse wurde unter Verwendung eines diskreten Kluftmodells simuliert und die sich dabei ergebenden Grundwasserstandsund Quellabflu8ganglinien als im Feld gemessene Werte betrachtet. Diese gemessenen Gro8en wurden anschlie8end zur Kalibrierung eines Doppel-Kontinuum Modells verwendet und die dabei ermittelten Werte mit den Ausgangsgro8en, d h der Karst-Netzwerkgeometrie verglichen. Dieser Vergleich diente dazu mathematiscb/physikalische Beziehungen zwischen dem diskreten Karstnetzwerk und den Doppel-Kontinuum Parametem zu entwickeln 1. I ntroduction Frequently the hydraulic properties of karst aquifers can be approximated assuming (a) a lower permeability fissure system with a high storage capacity which is drained by (b) a high permeability but low storativity conduit network. This simplified dualistic view of a karst system can be supported through observations of well and spring hydrograph behaviour which frequently exhibit two major response frequencies to recharge input a fast event response and a slow seasonal variation The simplest way to mathematically represent groundwater flow within such a system is the double-continuum approach where the fissure system as well as the conduit network are represented by a continuum each (TEUTSCH 1988) Even though the concept of a representative elementary volume (REV ; BEAR 1972) may not be applicable for the conduit network within smaller catchments the characteristic hydraulic features of several karst catchments on the Swabian Alb plateau in Southern Germany could be adequately reproduced using double-continuum models (TEUTSCH 1988; SAUTER, 1992) However the drawback of this approach is that the double-continuum model parameters cannot be directly related to the geometry and hydraulic conductivity of the karst system, which means that these parameters cannot be directly derived from field measurements but need to be obtained through model calibration (inverse approach). In order to develop a better understanding for the physical meaning of the double-continuum model parameters and to improve the prediction capability of the models a detailed numerical analysis was performed Due to the impracticability of a detailed karst geometry analysis at catchment scale, a synthetic karst system network was used as a reference system to represent reality The hydraulic response of this synthetic karst system to recharge events was calculated using a discrete numerical model which represented in detail the karst network geometry The resulting well heads and s pring hydrographs were subsequently treated as measured field data Consequently this data was then used as calibration target for the double-continuum model. The calibration itself was performed in two steps First, under steady state conditions a single-continuum model was used to fit the piezometric head surface which was obtained through interpolation based on a set of measurement points (boreholes) The resulting transmissivity distribution was then normalised to a base value and a spatially variable multiplier. Second the doublenow at the Institute for Hydromechanics, University ofKarlsruhe, Germany 6 th Conference on Limestone Hydro l og y and Fissu r ed Med i a 163

PAGE 175

continuum model was calibrated under transient conditions using the previously determined transmissivity multipliers, leading to spatially variable storage and exchange coefficients. 2. Synthetic karst system Approach Within a discrete karst conduit network, the turbulent groundwater flow can be described by the non-linear Darcy Weisbach equation: (I) ve is the flow velocity, d the diameter, h e the piezometric head in the conduit, g the acceleration due to gravity and A. a friction coefficient of the conduit. For the fissure system the assumption is made that a REV based continuum approach may be applicable Hence, the flow in the fissured system may be represented through a standard porous media groundwater flow equation (here in 2D) : (2) T 1 is the transm1ss1V1ty S 1 the storage coefficient, h 1 the piezometric head of the fissure system and N 1 the recharge to the fissure system The exchange flux q,x between the conduits and the continuum is defined by: (3) K 1 is the hydraulic conductivity of the fissured system A,x represents the interfacial area between the conduit and the fissured system V, 01 is the bulk volume of the adjacent fissured block and V h_J stands for the hydraulic gradient still within the fissured system but at the interface (x 1 = x e )The numerical calculations were performed using the finite element code ROCKFLOW (W0LLRATH et al ., 1991). [m] synthetic karst system 1000 0 1000 2000 [m] Figure 1: Discretisation of the synthetic karst system and positions of the measurement points. Discrete model The design of the synthetic karst system was based on the assumption of a dendritic conduit network This type of karstification has been postulated for numerous karst catchments on the Swabian Alb carbonate plateau It is believed to develop where a continuos hydraulic outflow control e g through a spring or a river, exists over a long period of time. The parameters used for this synthetic karst study are equivalent to those determined for the Gallusquelle' catchment area (SAlJTER, 1992; KIENLE et al ., 1997). Only the shape of the catchment area and some details of the conduit network geometry were simplified (Fig I) The recharge time series was calculated based on climatic data from the Gallusquelle' catchment. The relation to reality of the synthetic karst system is demon strated by comparing the characteristics of the spring discharges of the synthetic karst system and the 'Gall usquelle ', both calculated with a discrete numerical approach (Fig 2) The latter represents quite well the observed spring discharge (Fig. 2a). a) [Vs] 2500 spring discharge 'Gallusquelle' --measured 2000 1500 calculated 1000 500 0 )~\l-t-\1-) )t-':>O~O )~\l-t-\1-) )t-':>0~0 1988 1989 b) spring discharge rvsJ synthetic karst system 500 400 300 200 100 O Q )~\l-t-\1-)) t-':>O~O )~\l-t-\1-)) t-'=>0 1988 1989 Figure 2: a) Observed and computed spring discharge at the 'Gal/usquel/e' (KIENLE et al., 1997); b) computed spring discharge for the synthetic karst system. 3. Double-continuum model Approach In the double-continuum model approach, it is assumed that the two continua which represent the fissures and the conduit network overlap each other indistinguishable. Groundwater flow is described in both continua using Darcy s law (TEUTSCH 1988): 8h 1 J; V 2 h1 = S1 81 -Ni +q ex 8h 2 T2 ,:;;72h 2 = S2 atN2 -q *ex. (4) 164 Proceedings of the 12 '" International Congress of Speleology 1997, Sw i tzerland Volume 2

PAGE 176

T 1 and Ti are the transmissivities s 1 and Si the storage coefficients, h 1 and h i the piezometric heads and N 1 and N i the recharge for the respective continuum The exchange flux ,x between the two continua is determined based on a steady-state exchange coefficient a and the piezometric head difference (BARENBLAIT et al ., I 960) : (5) The double-continuum concept has been implemented as an extension to the well-known numerical code MODFLOW (MCDONALD & HARBAUGH 1984). Calibration For the synthetic karst system an equivalent double continuum model was calibrated using a two step procedure First a single-continuum model was calibrated under steady state conditions Then the double-continuum model was calibrated under transient conditions For the steady-state calibration the 'measured piezometric head surface was interpolated based on the available borehole locations (comp Fig I) These borehole locations were selected without prior knowledge of the location of the conduit network The heads at the borehole locations were calculated using the previously described discrete model which represents the synthetic karst system. The interpolated piezometric surface was then used as the calibration target for the steady-state single-continuum model which leads to the transmissivity distribution described in Figure 3 The minimum transmissivity obtained for the catchment area was 8 0* 10-4 m 2 / s which is what was defined as the so called base level transmissivity The other (higher) transmissivity values within the catchment were normalised to this base level i e. leading to a multiplier factor( > I) [m] 3000 2000 1000 factor distribution : 0 50 0 20 0 10 0 E\1 5 0 D 2 0 D 1 0 transmissivities 0 -+----r----r----r--r--r-.-.-.-i=lfr=-lri.......,--r---r-.---r--r---r-,--.--.0 1000 2000 [m] 3000 4000 Figure 3: Transmissivity distribution of the single-continuum model In a second step the transient calibration of the double continuum model was performed keeping the transmissivity multipliers constant and adapting the base level transmissivity only. This calibration strategy proved successful and was used to calibrate also the exchange coefficients For the calibration of the storage coefficient the square roots of the multiplier values were used. The results of the calibration procedure are shown in Figure 4 It is seen that the 'measured' spring discharge from the discrete model representing the synthetic karst system can be fitted very well by the double-continuum model. Similarly the 'measured piezometric head time series were fitted for the pre-defined borehole locations It is interesting to note that the quality of fit to one of the two continua depends mostly on the distance between the observation point (borehole) and the next conduit. For instance at the borehole location P4 which is located close to a conduit the measured data fits well the head hydrograph representing the conduit continuum (Fig. 4) Whereas, at the borehole location PI which is located far from the next conduit the measured data fits nicely the head hydrograph representing the fissure continuum (Fig. 4). 4. Results The quality of fit of the spring and well hydrographs to the measured data originating from the synthetic karst system show that double-continuum models may well reproduce the observed ambivalent features of karst systems Since the 'true geometric and hydraulic parameters are perfectly known for the synthetic karst system a set of functional relationships could be derived between those and the double-continuum parameters obtained through model calibration These relationships are summarised in Tabet 1 The transmissivity of the conduit continuum is determined by a zone L with an equivalent flux compared to the respective conduit (Fig 3) The relationship for the exchange coefficient is derived by the approximation of the hydraulic gradient in the fissure continuum adjacent to the conduit. The width L of the catchment area of a single conduit determines mainly the relationships for these two parameters The storage coefficient of the conduit continuum is determined by a part of the continuum block representing the catchment area of a single conduit. This part with width L shows a similar hydraulic response as the conduit. A more detailled discussion of the relationships described in Tabet 1 is given by MOHRLOK (1996) References BARENBLAIT G. E ., ZHELTOV I. P & I. N KOCHINA 1960 Basic concepts in the theory of the seepage of homogeneous liquids in fissured rocks J. Appl Math. Mech. (USSR) : 12861303 BEAR J 1972. Dynamics of fluids in porous media. Elsevier, New York 764 p. KIENLE J. MOHRLOK U & G TEUTSCH 1997 Beschreibung der Grundwasserstromung im Einzugsgebiet der Gallusquelle (Schwabische Alb) mit Hilfe eines diskreten Modellansatzes in pr e p LANG U 1995. Simulation regionaler Stromungsund Transportvorgange in Karstaquiferen mit Hilfe des Doppelkontinuum-Ansatzes : Methodenentwicklung und Parameterstudie. PhD-thesis, Institute for Hydraulics and Groundwater University of Stuttgart, 179 p McDONALD M G. & A W HARBAUGH 1984 A modular three-dimensional finite-difference groundwater flow model. USGS Reston 527 p MOHRLOK, U. 1996 Parameter-Identifikation in Doppel Kontinuum-Modellen am Beispiel von Karstaquiferen PhD thesis Tubinger geowissenschaftliche Arbeiten C3 I 125 p SAUTER, M 1992. Quantification and forecarsting of regional groundwater flow and transport in a karst aquifer (Gallusquelle Malm SW Germany). PhD-thesis Tubinger geowissenschaft /iche Arbeiten Cl3 150 p. TEUTSCH G 1988. Grundwassermodelle im Karst: Pralctische Ansatze am Beispiel zweier Einzugsgebiete im Tiefen und 5 t11 Conference on Limestone Hydrology and Fissured Med i a 165

PAGE 177

spring discharge [Vs] discrete model double-continuum model 0 )~"'""' l )!'-':>O~O )~"'""') )!'-':>O 1988 1989 [m] 600 550 500 piezometric heads P1 discrete model double-continuum model, condutt cont. ...... .. double-continuum model fissure cont. l I \/ 1 I I :n i ~ : .....,..,.. ,_ ~ -~::: .~ :::, ~ ~~.: .~ -::/' ~ 0) ~"'I'-"')) I'-':> 0 0) ~"'""'))I'-':> 0 1988 1989 [m] 600 550 500 piezometric heads P4 0) ~"'I'-"')) I'-':> O 0) ~"' ""') l I'-':> O 1988 1989 Figure 4: Spring discharge and piezometric head time series at the measurement points P 1, P4 of the calibrated double-continuum model hydraulic parameters exchange coefficient transmissivity of conduit cont. storage coefficient of conduit cont. transmissivity of fissure cont. storage coefficient of fissure cont. geometric und physical relationships of the parameters approximation of 1 T k hydr. gradient: ex = T L 2 f = 0 2 0 5 flux equivalent d 512 to conduit: T 1 = \J iiVh L similar hydraulic behavior as conduit: identical cont. approach: subdivision of whole storage volumen: zones determined by calibration catchment area of conduits: L'=26 ... 450m L=45 ... 500 m L*= 15 m Table J: Relationships between the hydraulic parameters of the double-continuum model and the geometric and hydraulic parameters of the discrete model Seichten Malmkarst der Schwiibischen Alb PhD-thesis at the Geologic-Palaeontologic Institute University ofTUbingen, 220 p TEUTSCH G & M SAUTER 1991. Groundwater modelling in karst terranes : scale effects, data acquisition and field validation Proc. Third Conference on Hydrogeology, Ecology, Monitoring and Management of Ground Water in Karst Terranes Dec 1991 Nashville. WOLLRATH, J & R HELMIG 1991. SM-2 Stromungsmodell flir inkompressible Fluide Theory and user manual, Institute for Fluidmechanics University of Hannover technical report Acknowlegments The presented study has been supported by the German Science Foundation (DFG). 166 Proceedings of the 12 Internat i onal Congress of Speleology 1997, Switzerland Vo lume 2

PAGE 178

Modelling groundwater flow in a karst terrane using discrete and double-continuum approaches importance of spatial and temporal distribution of recharge U. Mohrlok', M. Sauter Applied Geology, Geological Institute, University ofTtibingen, Gennany Abs t rac t Groundwater flow had been modelled in the karst catchment area 'Gallusquelle' (Swabian Alb, SW-Germany) using two different types of modelling approaches. The discrete and the double-continuum model differ in their respective representation of the conduit network and the formulation of the exchange flux of groundwater between fissured system and conduits In the case of the discrete approach this exchange is determined by local hydraulic properties adjacent to the conduits The double-continuum approach represents this exchange using a 'steady state' lumped parameter. As a result of this fundamental difference between the two approaches the temporal distribution as well as the percentual allocation of groundwater recharge to conduits and fissured system plays a major role in the respective model calibration Kurzfassung Die Grundwasserstromung im Karsteinzugsgebiet der Gallusquelle (Schwabische Alb SW Deutschland) wurde mit zwei Modellansatzen beschrieben Diskretes und Doppel Kontinuum-Modell unterscheiden sich in ihrer jeweiligen Beschreibung des Karstrohrennetzes und der Formulierung des Austausches von Grundwasser zwischen Feinkluftsystem und Karstrohrennetz. Beim diskreten Ansatz wird dieser Austausch durch die hydraulischen Eigenschaften in der Umgebung der Karstro h ren bestimmt. Der Doppel-Kontinuum-Ansatz reprasentiert diesen Austausch mit Hilfe eines 's tationaren Parameters Als Ergebnis dieses fundamentalen Unterschieds zwischen beiden Ansatzen spielt die zeitliche Verteilung sowie die Aufteilung der Grundwassemeubildung auf die Karstrohren bzw. das Feinkluftsystem eine wesentliche Rolle bei der jeweiligen Modellkalib r ierung 1. Introduction The hydraulic properties of most karst aquifers can be characterised by typical discharge and piezometric head time series The observed features of these time series are a steep initial increase and decrease afte r recharge events and a more or less long recession period Further groundwater level fluctua tions increase with increasing distance from the spring These observations can be explained by the hierarchical structure of void volume in a karst aqui fe r. A network of highly permeable karst conduits has developed as a resu l t of carbonate dissolution in the original fissured carbonates These conduits drains the permeable high storage fissured system. Different model approaches describing groundwater flow in s u ch dual permeabi lity systems are summerized by TEUTSCH & SAUTER ( 1991 ). In the presented study the discrete and t h e double co n tinuum approach were compared describing groundwater flow in the Gallusquelle karst catchment area (Fig I) They differ in their representation of the conduit network Therefore the mathematical formulation of the exchange flux of groundwater between the fissured system and the conduit network has to be different for the two approaches Generally it is assumed that a zone exists near the s u rface the epikarst (WILLIAMS 1983) acting as a buffer for gro u ndwater recharge This horizon effects strongly the temporal and spatia l distribution of groundwater recharge. A fraction of the i nfiltrated water flows rapidly towards doline structures and is conveyed to the conduit system of the aquifer via vertical shafts The rest slowly drains to the p h reatic zone. Depending on the type of the selected groundwater flow model different temporal distrib u tion as well as different percentual allocation to conduits and fissured system for the recharge input functio n have to be chosen. In this paper it is discussed whether this phenome n on can be used to derive the recharge mechanism in a karst terrane or if it is related to the respective model ap p roach 5352 ..-,=------~-----. 50 48 46 44 42 3502 04 06 08 10 12 14 16 A N Fig ur e 1: Lo c ati o 11 of th e Gal/u s quel/ e' ca t c hme11 t a r e a 2. The Gallusquelle catchment area Geology hydrogeology The Swabian Alb consists of a sequence of Upper Jurassic limestones, which vary between bedded and massive bioherm facies The catchment area of the Gallusquelle is situated in the Institute for Hydromechanics, University of Kar l sruhe Germany 5 m Conference on Limestone Hydrology and Fissured Media 167

PAGE 179

massive facies of the Upper Jurassic (ki2/3). At the surface typical karst features such as dry valleys and doline structures are visible. In the past, a number of geological and hydrogeological investigations had been performed, especially in context with a large water supply project during the sixties. Information about the catchment boundarie s is summarized by SAUTER ( 1995). Additionally, aerial photographs and satellite images were analysed in order to identify lineaments as zones of major karstification and potential conduits (BRUX, 1981 ). Recharge It is assumed, that the epikarstic horizon plays a major role in the temporal and spatial distribution of groundwater recharge (SAUTER, 1994 ). Water enters the subsurface and collects in a highly porous and highly conductive horizon There, a perched water table is formed and a certain percentage of the recharge water flows towards vertical shafts and is rapidly conveyed to the groundwater table This way, the epikarst system can function on the one hand as a high storage buffer and on the other hand as a highly transmissive system for rapid infiltration. The recharge is calculated using a soil moisture balance approach. Potential evapotranspiration was computed after HAUDE (1955) based on recorded meteorological data The delaying effect of snow melt and epikarst is accounted for seperately. A certain quantity of recharge is allocated directly to the conduit network. The rest infiltrates into the fissured system 3. Groundwater flow modelling Discrete approach The conduit network is represented by a geometric discrete pattern. Flow in the conduits is calculated using the friction law of Darcy-Weisbach. The fissured system is represented by a continuum. Since at the regional scale of the catchment area the number of fissures is large enough, a representative elementary volume (REV; BEAR 1972) can be defined and Darcy's law is 3510 5345 ... N 0 2km .... : ... ,,.-::,. 3510 used to describe groundwater flow in the continuum The exc hange flux qex between the conduits and that continuum is defined by: (I) K 1 is the h ydraulic conductivity of the fissured system A,x the interface between the conduit and the fissured system, V, 01 the bulk vo lume of the adjacent continuum block and V h 1 the hy draulic gradient in the fissured system at the interface (x 1 = x c ). This approach is implemented in the finite element code ROCKFLOW (W0LLRATH et al 1991) Using this approach a two-dimensional model was built for the 'Ga llusquelle catchment area. The conduit network was derived from available informations (Fig. 2). For discrete modelling, no intermediate storage within the epikarst was assumed and 95% of the recharge was allocated directly to the conduit network (Fig. 3). Double-continuum approach This approach repre sents both subsystems, the fissured system as well as the conduit network, by a continuum where groundwater flow i s described by Darcy's law. This is the simplest way to overcome the lack of knowledge of the exact geometry and positions of the conduits and still to represent a fast flow system. The two continua overlap in space. The exchange flux q,x between them is defined by a 'steady state' exchange coefficient a 0 and the piezometric head difference between the two subsystems (BARENBLATT et al., 1960) : (2) This approach is implemented in a further development of the numerical code MODFLOW (McDONALD & HARBAUGH, 198 4) and was successfully applied to several karst areas on the Swabian Alb (TEUTSCH, 1988 ; SAUTER, 1992) 3515 I 3515 5345 l""v-J \: --\__ -'-y ~ valleys spring dolines dry valleys fault lineament inflow to planned water pipe line Figure 2: Conduit network derived from available information, and discretisation of the discrete model 168 Proceedings of the 12 th International Congress of Speleology 1997, Switzerland Volume 2

PAGE 180

double continuum model epikarst vadose zone ph r eatic zone discrete model sp ring gro u nd w ate r r echa r g e 15 10 5 discrete model: to conduit network ---to fissured system O --J-...c.._ __ __._.cl..<= "--..1., ...._____._+-'----_,___--'___ ____ ---, IIAR A PR [ mrn/d ] 20 double-continuum model : 15 to conduit continuum 10 f1i ,_r J i.,_ r ..... to fissured continuum 5 I '"1-,_J -----0 4== ----11 A R """"c!::,:=-_,..J=---====A=P=R =====; 1984 Figure 3: R echarge allocation to the two subsystems a n d tempora l dist r ib u tio n of a sing l e recharge event for bo t h mode l s SA UTE R ( 1992) used a one-dimensional model to simulate groundwater flow and transport in the Gallu s quelle catchment area (Fig 4). In order to account for the buffering role of the epikarst he de v eloped a retention function and allocated only 7% of the ca l culated recharg e to the continuum representing th e conduit network (Fig. 3) Results Both model s are able to fit the time series of the s pring di s charge well (Fig 5) In the discrete model the distances between the piezometric head measurement point s and th e conduits are not represented adequatel y, since the location of the single conduits were not varied during the calibration process On the other hand in the neighbourhood of the original measurement point nodes exi s t, where the calcul a ted piezometric head time series shows similar characteristics as the measured one (Fig 5 ). In the double-continuum model the measured piezometric head i s assumed to be that of the fissured continuum becau s e the probability that the borehole taps a conduit is very low (F i g 5) 4. Discussion Comparing the different models it is obviou s that both were able to reproduce the observed discharge and pie z ometric heads e v en if different recharge input function s were used. These input function s were not cho s en arbitraril y Their di s tinct choice in combination with the determination of the model parameters was necessary to calibrate the respective model s u cces s fully mas/ 740 Fissured System 1 l I I I l f Ftt:= / ?ro 1 un~wa 1 t er 1 Lev t el (;iss ure~ s ys~;m\ 27 1 3 8, 8 ) G roundwater Level (fissured s 14 1 2 89 700 680 660 Co nduit Sys tem ' 1 I I I I I I G roundw ater Le vel (co nd uit Sys t e m) 1 4 12 8 9 I I I I I I I I I I ~"'l.-G r oundwa t er L eve l (con duit S ste m) 2 7 3 88 10000 8000 6000 4000 2000 0 Distance from Spring ( m) Figu r e 4: Discretisa t ion of the do ub le co n ti nuum model Both models show that there i s a large quantity of recharged g roundwater that had to be stored within an intermediate time s cale At the s ame time a certain amount of water has to discharge rapidl y and the rema i ning water has to be stored for a long time producing a l ow reces s ion. In the discrete model most of the water is allocated to the conduit network It flows into the fissured contiuum as a result of the s teep hydraulic gradient betwe e n the discrete conduits and the fis s ured sys tem When th e pre ss ure in the conduits decreases this water flows back from the fissured continuum resulting as low recession In this way characte r istic piezometric head time series can be modelled in the fi s sured sy s tem for certain d i stances to th e conduits. Jn the double-continuum model this proces s doe s not occur since the two s ubs y stem s are coupled directly at every node Therefore the equivalent s torage proce ss had to be assumed to occur in the epikarst. Both interpretations storage in the karst aquifer for the di s crete approach and in the epikarst for the double-continuum approach are in agreement with the respective conceptual models of the rec h arge process and the h y draulic properties of a karst aquifer However it s hould not be caused by the groundwater flow mode l assumptions which kind of model for the recharge proce s s has to be used Further research is needed to answer the question where the recharged water is really stored. Then a re a listic model approach can be applied and the nature of the ex change proce s ses can be identified References BARENBLATI G E ., ZHELTOV I. P & I. N KOCHINA 1960 Basic concepts in the theory of the s eepage of homogeneou s liquid s in fis s ured rocks J. A ppl Math. Me c h. (USSR) : 12861303 B E AR, J 1972 Dynamics of fluids in porou s media Elsevier New York 764 p BRUX C 1981. Vergleichende Untersuchungen der Bruchtektonik im Bereich des Hohenzollernund Lauchertgrabens (Schwabi s che Alb) im Luftbild Satellitenbild und Gelande Unpublished diploma thesis University of Karl s ruhe 66 p 5 h Conferen c e on Umestone H y dro l ogy and Fissured Med i a 16 9

PAGE 181

(Vs) 2500 2000 1500 discrete model spring discharge (Vs) 2500 2000 1500 1000 500 double-continuum model spring discharge measured calculated 0 1984 1985 1 986 1987 1988 1 989 1984 1985 1986 1987 1988 1989 lmJ piezometric heads 87 Im) piezometric heads 87 670 660 I'll 650 87 measured 87 calculated 87a calculated 1984 1985 1 986 1987 1988 1 989 1984 1985 1986 1987 1988 1989 1ml piezometric heads 814 I'll Im] piezometric heads 814 700 700 690 680 670 660 .. J 814, measured 814 calculated 814a, calculated 660 1984 1985 1 986 1987 1988 1 989 1 984 1985 1986 1987 1 988 1989 Figure 5: Spring discharge and piezometric heads of the two calibrated models HAUDE W 1955 Zur Bestimmung der Verdunstung auf moglichst einfache Weise. Mitt Dt Wellerdienst 11 : 1-24 McDONALD, M G & A. W. HARBAUGH 1984 A modular three-dimensional finite-difference groundwater flow model. USGS Reston 527 p SAUTER M 1992 Quantification and forecarsting of regional groundwater flow and transport in a karst aquifer (Gallusquelle, Malm SW Germany) PhD-thesis Tubinger geowissenschaft liche Arbeiten CI3 150 p SAUTER M. 1994 Areal-source transport in a karst system Proc Water Down Under '94, Nov. 1994 Adelaide Australia. SAUTER M 1995. Delineation of a karst aquifer using geological and hydrological data and information on landscape development. Carbonates and Evaporiles, 10 / 2 : 129-139 TEUTSCH, G 1988 Grundwassermodelle im Karst: Praktische Ansiitze am Beispiel zweier Einzugsgebiete im Tiefen und Seichten Malmkarst der Schwiibischen Alb PhD-thesis at the Geologic-Palaeontologic Institute University ofTiibingen 220 p TEUTSCH, G & M SAUTER 1991. Groundwater modelling in karst terranes : scale effects data acquisition and field validation. Proc Third Conference on Hydrogeology Ecology Monitoring and Management of Ground Water in Karst Terranes, Dec 1991 Nashville WILLlAMS P W 1983 The role of the subcoutanous zone in karst hydrology. J. of H y d. 61 : 45-67 WOLLRATH, J. & R HELMIG 1991. SM-2, Stromungsmodell flir inkompressible Fluide Theory and user manual Institute for Fluidmechanics University of Hannover technical report 170 Proceedings of the 12 th International Congress of Speleology 1997 Switzerland Volume 2

PAGE 182

Water levels as response functions for different porous fissured-karstic systems. by Jacek Motyka(l) and Andrzej Zuber(2), (1) University of Mining and Metallurgy, PL-30065 Cracow, Mickiewicz.a 30, Poland, (2) Institute of Nuclear Physics PL-31342 Cracow, Radzikowskiego 152, Poland Abstract In carbonate rocks the following three basic types of porosity exist: matrix porosity, fissure porosity and cavern (karstic) porosity Existence of these three porosities in different proportions and charactemed by different hydraulic conductivities influences the dynamic behaviour of the water system. That behaviour can be observed by changes in water level as a function of time. Different case studies will be presented. 1. Introduction In general, for carbonate rocks, three systems can be distinguished, which clearly differ geometrically and have different hydraulic characteristics : the cavernous (karstic), fissured, and porous spaces Forms filled by secondary material can be included into these systems, because they are very often unstable and can change their geometry and hydraulic parameters when the hydraulic gradient changes Neither an exact definition nor an exact determination of particular spaces is possible, because they do not differ in origin and overlap one another. According to Choquette and Pray (1970), the porous space is represented by empty spaces of any origin, which are observable in a small rock sample, say, of the order of several tens of cm 3 Fissures can also be of different origin, but they are charactemed by space orientation and two dimensions (length and width) which are much larger than the third dimension (aperture). Caverns are arbitrarily defined as empty forms with the cross-section area on an exposed surface, which is not smaller than 0 001 m 2 or with length in a borehole not shorter than 0.1 m (Motyka and Szuwarzynski, I 994). The presence of these three different spaces in carbonate rocks influences the dynamics of water and solute transport Within this work fluctuations of the water table are discussed as a response to the importance of particular spaces in different aquifers of triple porosity The discussion is based on case studies from Cracow Olkusz area, southern Poland 2. Water table in the area of a well developed paleokarst (Olkusz region) In the period of 1962-65 the water table level was measured in about 60 boreholes and several observation wells (piezometers) in dolomites and limestones of the Lower and Middle Triassic The water table appeared to be very irregular (Fig. IA). In the central part, it was possible to observe an elongated and narrow drainage zone with an axis having the W direction The recharge of the area is from the eastern and south-eastern directions The carbonate formations have the average thickness of about 100 m and lie at the depth of about 30 to about 200 m below the ground level. On a large area they are covered by Keuper clays overlain by Jurassic limestones (Fig lB). During drillings about 50 caverns were discovered with the length along the borehole axis of about 0 3 to 4.5 m. Some of them were filled with breccias regarded as being of karstic origin The roof surface of the carbonates is hilly with a distinct depression directed from EES to WWN, and l.5 to 2.0 km wide That depression has a relatively flat bottom with a number of local elliptical funel depressions up to 30 m deep. According to Wilk et al. ( 1989) it is a paleorelief characteristic for advanced karstification processes As a result of that, the main role in the development of the shape of natural water table with numerous irregularities (before the exploitation of mines has started) was played by the internal drainage system of the karstic caverns and channels within the Triassic formations mul LOO l JOO 200 100 SW 1-1 NE 0 200m ._____, Fig. I. Hydrogeology of the eastern pat of the Olkusz region A. Water table in the Trassic aquifer. B. fu!ological cross-section. 1-Permian, 2-Triassic aquifer, 3-Keuper, 4-Jurassic, 5-Quaternary, 6-faults, 7-water table contours, 8--0bservation wells, 9-cross-section line, 10-directions of groundwater now. 3. Water table fluctuations in a zone of different hydraulic connections between carbonate rocks (Olkusz region) That example is also related to the Olkusz mining area of Zn-Pb ores, where as the result of drainage by mines and an intensive exploitation of water, a large depression cone has formed, mainly in the Triassic limestones and dolomites By different forms of hydraulic connections that cone locally 5 1 h Conference on Umestone Hydrology and Fissured Media 171

PAGE 183

influences the Carboniferous-Devonian, Jurassic and Quaternary aquifers (Wille and Motyka, 1977) In the south-eastern part of the region, the Carboniferous Devonian formations are elevated in the eastern direction and gradually are covered by Triassic and next by Jurassic (Malm) aquifers That part of the region is therefore characterized by complex hydraulic connections between rocks of different ages and hydraulic parameters (hydraulic conductivity and storage characteristics in different kinds of spaces). The Jurassic limestones are fractured and with well developed karstic forms In spite of a high matrix porosity (0 070.10), its hydraulic properties practically do not influence the hydraulic properties of the aquifer. Therefore, the aquifer is mainly of the double porosity type, i.e., a karstic-fissured system The Triassic limestones and dolomites are characterized by well developed systems of caverns and fissures as well by relatively high hydraulic conductivity and specific yield of the porous space Therefore, they are classified as a triple porosity type, i.e., a karstic-fissured-porous system (Motyka, 1988). The Carbonifereous and Devonian limestones and dolomites are practicaly permeable only due to the presence of fractures (though caverns are sporadically observed) because the porous space is of negligibly low hydraulic conductivity and specific yield (Borcz.ak et al., 1994 ) Therefore, that aquifer is of the single porosity type, i e., a fractured system (a quick response in 5 4 3 which are recharged indirectly from the Jurassic aquifer All these inake the whole zone to be quite complex (Fig 2) In that complex zone three piezometers have been installed at close distances at Jurassic, Triassic, and Devonian aquifers Unfortunately, the natural behaviour remains unknown because the water levels measurements were started when the Triassic aquifer was already strongly drained by mines. The results of observations in Triassic and Devonian aquifers are shown in Fig 2A as differences in the respect to the water level in the Jurassic aquifer These results can be explained both by considering the geological structure of the zone and by taking into account the differences in the hydraulic properties of the three carbonate systems. Namely, the Devonian limestones and dolomites being a fissured systems are characterized by relatively good hydraulic conductivity and low storativity (specific yield), and, therefore they quickly respond to changes in the water table in the Jurassic aquifer. The Triassic limestones and dolomites due to an important role of the porous space in the storage of water respond much slower and to a lower degree. The overlapping effects of the drainage in the Triassic aquifer, inflow from the Jurassic aquifer to the lower ones and the differences in the hydraulic properties lead to the behaviour of the water table responses which at the first glance may seem to be inconsistent (Fig 2A), though it may be explained by differences in characteristics of particular porous spaces ,,,,,,,,,,..,.,,.V __.,,..,.,,\ __,..,,__,..,, 2 //,.,,,,\ ? \(j) 9 9 3 E L 5 6 7 g --/----9 9 4 / \ 1 9 9 5 9 11 12 2 3 4 5 6 7 t 9 10 11 12 1 2 3 4 5 6 7 0 :c <] I. ..----------------------, ',,~, OBSERVATlON WELL \,, -1 /t' -2 ;./ .,.,,,,.,.I -3 I I -4 I -5 @ \ ' \ \ Fig. 2. A. Differences in water table levels (m) in the Triassic aquifer (T) and the Carboniferous-Devenian aquifer (D) in respect to the water table in the Jurassic aquifer B. Schematic cross-section and directions of groundwater Dow. water table changes, i e ., a very low inertia of the system). The above given classification is related only to the flow dynamics For transport parametrs the matrix porosity plays a dominant role (Zuber, Motyka, 1994; Motyka et al., 1994) As a result of a significant decline of the water table in the Triassic system ( about 80 to 100 m at the centre of the drainage area), that system is recharged in the south-eastern part by vertical flows from the Jurassic limestones at the local hydraulic connections. A significant role is also played by local horizontal flows from the Carboniferous and Devonian carbonate aquifers 4. Recovery of the water table in an abandoned quarry Malm limestones were exploited from a Zaknowek quarry in Cracow, which is situated about 600-900 m from the Vistula river whose flow is usually between 50 and 200 m 3 /s. Exploitation was performed to the depth of about 60 m below the ground surface, i e ., in the deepest parts of about 30 m below the river level (Fig. 3). In November 1992 the exploitation of limestone was stopped and the quarry has started to be filled in 172 Proceedings of the 12 1 International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 184

by inflowing groundwater. The recovery of the water table as a function of time is shown in Fig 4) At the beginning the increase of water table level was 2-3 cm/d and next for more than 700 days it was on the average equal to about 1 9 cm/d The final stages are characterized b y changes of about 1 cm/d Temporary faster changes were related to snow melt and intensive precipitation events These events are particularly well seen at the middle and final stages of the recovery curve (Fig 4 ) m a.sJ. 240 220 200 ,ea 160 1 l..0 12 0 w VISTULA RIVER 0 100 200m hydraulic network (Fig 3) The interbedding planes slightly dip at the angle of several degrees from the Vistula river to the quarry. Outflows from caverns existed only temporarily. The main feature of the recovery curve (Fig 4) is the existence of two straight parts, if temporal changes caused by intensive rains are neglected The hydraulic conductivity is governed by the fissure system, which explains the first straight part of the line When a sufficiently high water pressure was created in the fissure network the porous and microfissure ZAKRZOWEK OUARRY -------, I L __ E Fig.3. Geological cross-section through the Zakrzowek horst 1Quaternary, 2-Miocene clays, 3..Jurassic limestones, 4-water level during the exploitation. m a s l. ~o---------~----------.----------,,-------..,,------=.,... I I I 1 9 0 1 1 1 I 180 -1+ 1 I I I I .... 1 70 h,2 +-, -,2 .,... 3 -.-4 -,-5.,.-6 ., -, s -,9 .,... 0 11 -.-,2 .;......., 2 -.,...-J 4 -r5 -,6 -r7 's--r9 -r o r ..., ,2 ,, 2 ,-J ,4 r 5 .6 -,1 s :-r: 9 :r: 0T -:, r:,2 :r:, .:2 r: J :-,-:4 ,-:5 r:6:-r:7 ,:S r: g :--r::, o -r ::-, ',2 ::-1 199 "2 1 g 9 J 1 9 9 4 1 9 9 5 1 9 9 6 Fig. 4. Water table recovery curve the Zakrzowek quarry. There are three porous spaces in the Malm limestones of that region : cavernous fissured and porous The mean matrix porosity is 0 076 as determined on 418 samples whereas the cavernous and fissured porosities together are estimated visually to be below 0 01. The hydraulic conductivity is mainly governed by fissures and interbedding planes which form an orthogonal t i me systems most probably started to be filled in and in consequence, the second part of a slower recovery was observed 5 th Conf e rence on Umes t on e H y d r o logy a n d Fissur ed Me dia 173

PAGE 185

5. An example of changes in water table caused by pumping test Heterogeneities of karstic-fissured systems can also be observed during pwnping tests In the case of the presence of karstic fonns partly filled in with secondary material the depression curve does not correspond to the shape expected from the Theis or Theis-Jacob methods Very often the water level declines and partly recoveres. Till now no reliable methods exist for the interpretation of such results. An example of a pwnping test perfonned in the Triassic carbonate fonnation of the Ollrusz area is shown in Fig 5, where changes in the water table in an 0 6 @ 0.4 E (/) 0 2 0 10 2 10 3 t ( s) References BORCZAK, S., T LESNIAK, & J. MOTYKA, 1994 Hydrogeological properties of the pore space of the Lower Carbonifereus Czatkowice limestones and dolomites (SW Poland) (in Polish) Przegl. Geol. 8 : 653-657 CHOQUETTE, P.W & L C PRAY, 1970. Geologic nomenclature and classification of porosity in sedimentary carbonates. The Am. Assoc. of Petrol Geol. Bull 54, 2 : 207250 MOTYKA, J. 1988. Triassic carbonate sedimentes of Ollrusz-Zawiercie ore -bearing district as an aquifer (in PolisJv. I ,al. 10 5 Fig. 5. An example of irregular changes in the depression curve in an observation well (P) during a pumping test (well W). observation borehole (piezometer) are demonstrated In the first stage after the first 4 min the water table declined by several ems Next, after about I O min the water table returned to the initial level. In these stages, the changes resulted from the transmission of the water pressure After 20 min a systematic decrease of the water table level was observed However, even that systematic decrease was still very irregular, and it can be approximated by two straight lines Most probably it was a result of differences in the hydraulic conductivity of two systems of caverns and fissures with large apertures . In further stages the water table somewhat recovered, and next steeply declined, with a following recovery (Fig. 5). Such changes were most probably caused by the movement of the secondary material in some caverns due to an increased hydraulic gradient. As a result of that some conducting channels and fissures were temporarily opened and clogged intern with the secondary material An alternative explanation is that the system consists of a nwnber of cavern and channels characterized by different ability for transmission and storage of water i e., by having different inertions Zesz. Nauk. AGH, 1157, Geologia 36, Krakow, 109p MOTYKA, J. & M SZUWARZYNSKI 1994 Cavern porosity in the Triassic carbonate rocks of the Cracow-Silesian monocline (South Poland) Pub/. Serv Geo/. Luxembourg, 27 : 103-110 MOTYKA, J., S. WITCZAK & A ZUBER. 1994 Migration of lignosulfonates in a karstic-fractured-porous aquifer : Histroy and pronosis for a Znb-Pb mine, Pomorzany, southern Poland. Environ Geology, 24 z : 144-149 WILK, Z & J. MOTYKA.1977 Contacts between water bearing horizons (Olkusz ore mining region, near Cracow) (in Polish) Racz. Pol. Tow. Geo/ 47(1) : 115-143 WILK Z ., T. KAWALEC, J MOTYKA & K. ZUBER. 1989 Influence of karst passages upon the piezometric surface of the surrounding aquifer (in Polish) Kras i Speleologia 6 : 722 ZUBER,A.& I MOTYKA. 1994. Matrix porosity as the most important parameter of fissured rocks for solute transport at large scale J. Hydro/. 158: 19-46 174 Proceedings of the 12 th International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 186

Application of flow models to the prediction of karst behaviou r and pollutant migration in a groundwater system Maria Pascu, R.D.Gaspar Institute of Physics and Nuclear Engineering Bucharest Romania Abstract The most important karstic system from Valcan Mountains discharges through the Izvarna springs, 2.2 m 3 /sec which are tapped and piped for water supply to the city of Craiova The connections between the springs and the main inflows were identified with activable tracers.The karst behaviour and the vulnerability to pollution were investigated by using In-EDTA as a tracer. Hydraulic dispersion of the pollutants involves the transfer of a dissolved constituent from a higher concentration zone to a lower concentration zone by means of mixing and molecular diffusion. Dispersion has a very significant influence on the migration of solutes in large groundwater systems. Using the dispersion model the flow parameters as such velocity, length of the underground flow path, transit time and hydrodynamic dispersion coefficient were determined Resume Le jeu de sources d'eau d'Izvama, ayant un debit de 2.2 m' /sec est le systeme karstique le plus important de ceux appartenant aux Montaigne s Valcan. II est utilise pour approvisionner la ville Craiova On a utilise des traceurs activables pour mettre en evidence des rapports entre Jes deux groupes de sources captees et les trois points de pertes observes Le comportement du karst et sa vu l nerabilite aux polluant s ont ete examines en utilisant comme traceur le In-EDTA Comme regle generate, la dispersion hydraulique implique le transfert de la substance dissoute d'une zone de grande concentration vers une autre zone, de concentration plus basse par les moyens du melange et de la diffusion moleculaire Dans des grands systemes d'eaux souterraines, la dispersion pe u t jouer un role tres important dans la migration des solutes. En soutenant cette affirmation, no s dates experimenta l es pour les trois sub-systemes de sources sont p r esentees dans le texte qui suit. 1. Introduction One of the most complex hydrogeological structures and karst interconnection s from Valc a n Mount ai ns, Rom a ni a, extending on more than 500 s q km include s three karstic syste m s: lzvarna Jales and Patrun saPicui e l. The lz varna karst springs, with a flow roughly 2.2.cum./sec and variation coefficie nt of 1.4 2 0 are major sources of pot ab le water s upply for Craiova town. In the 15 last years, a lar ge deforestation a nd urb a nization caused a increasing pollution and a degrad atio n of water resources As a consequence a continuous lowering of the discharges was observed. The construction works on a major hydropower complex in the a rea which implied the builing of dams to create storage lake s with waters from the rivers that cross the zone posed the problem of the influence the retentions exert on capture. Therefore as the damming of water flows entails a diminution of the capture discharge. engineering solutions should be found to the problem. While the demand for water is in creas ing and its availability being reduced a nd effect of growing of this pollution superposed, new hydrological st udies were necessary to delineate the karst systems a nd to determine their vulnerability to po l lution. The connections between the s pring s and the main inflows were identified in the s tudied area with activable tracers and activation analysis. The karst behaviour and the vulnerability to pollution were investigated by u s ing In-EDT A as a tracer. 2. The tracer The ac tivable tracers have a great stability in time, reacts slowly with the l abe led environment, and most of them can be considered as co nser vat ive tracers. Thus ln EDTA has a stability consta nt of 24.9 and can be measured in routine determinations in concentrations down to 2. 10 12 g/ml. It ha s a very low toxicity permitting the direct labeling of the potable waters. ln-EDTA the most used tracer by us. is considered as a reference tracer. Indium is practically absent ( in the limits of the a nalytical method) from the investigated waters. S am ples are analysed in the laboratory lndium is extracted from the water samples ( I 00 ml) by precipitation The precipitate i s then filtered throu g h a nuc l ear membrane filter dried at room temper a ture a nd irradiated with a pneumatic rabbit tube at the nuclear reactor. ( GASPAR & all, 1984 ) The m a jor advantage of this tr ace r is that, owing to its extremely low detection limit it ca n label emergences of very low flow rates which ca nnot be labelled by other tracer, whether fluore sce nt. chemical or radioactive (w ith the exception of tritiated water) : a n a rtifici a l lake or natural reservoir of 10 6 m 3 m ay be l abelled using 100 g In a ln-EDTA complex. The detection limit can be lowered below 1 10 12 g/ml using a volume of water samples larger than I 00 cm' and a larger neutron flux for irradiation ( BEHRENS & all. 1977 GASPAR & al, 1985). s th C onfe rence o n Li mes t on e Hydro l og y a nd Fi s sured Med ia 175

PAGE 187

1-PATRUNSA SFRJNG 2 GRlPU SS: PON OR 3-GROPU CU APA PONOR 4FUSTEJCA PONOR 5-JALES SPRING 6 SU SITA VERDE PONOR 7-SUSlTA SEACA PONOR 8 -Gt,RLA VACJI PONOR 9IZVARNA SPRING 10-BALTA PONOR 11-BlSTRlTA PONOR 12-SOHODci PONOR J.; LABELLED SWALLET a WATER SUPPLY STATION LIMESTONES 0 1 2 3Km Figure.I.General sketc h of Southern Valca11 Mountains karst. 3. Flow models The data obtained from tracer measurements may be used either directly or in association with now models. Generally speaking, dispersion and mixing processes t aking place during flow in a real karst hydrol ogical system are far too comp l ex to be entire l y described mathematically: furthermore, it is difficult. if not impossible. to adapt a response function co rre spo ndin g to the system und er study. H owever, though idealized, the mathematical models and the digital simulation approach, which are currently used in hydrology (PRZEWLOCKI & YURT SEVER. 1974; YURT SEVER & PAYNE. 1978 ; MALOSZEWSKI & ZUBER 1984 ) may lead in certain cases, to a convincing description of the dynamic behaviour of the aquifer. The e models originate in two basic approaches, the convolution approach and the quasi-physical models employing a n interconnected array of mixing cells (compartmental models). Flow in karstic structures may be modelled if the input and the output functions are known. To obtain these f,rnctions is important as they provide an assessme nt of the transfer function of the karstic system It i easy to determine the input function for tracer tests when inj ections have the form of a pulse In th e case of diffuse losses through river beds or l ake bottoms. a deconvolution operation is needed for the assessment of the input function. In general, experimental data (repeated labellings, multitracing experiments) supply sufficient arguments for the selection of a now pallern with total displacement drain or branches short circuits. perfect-mixing zones (underground pools), stagnant zones ( underground lakes syphons), zo n es with dispersion diffusion and exchange (auxi li ary systems). The selection of flow models and the determination of parameters from tracer studies is based on the analysis of impulse respon se fu n ction. 176 Proceedings of the 12 t h International Congress of Speleology 1997, SWitzerland Volume 2

PAGE 188

This function is the Residence Time Distribution (RTD) obtained from the system 10 a tracer input in the form of a Dirac delta function (sudden injection). RTD = E(t) = C(t) f C(t)dt (I) where C(t) is the tracer concentration in water (i.e.mass or activity per unit.vo l ume of water) The impulse response functions were fitted with the dispersion function h(t) as: h(t) = 1 v ex ( L W )2 2 .J nDt p 4Dt (2) 4.The experiment The hydrological survey, combined with the study of the geology of the morphology and of the hydrography of the southern slope of Valcan Mountains resulted in the delineation of three karstic systems. The artificial labellings demonstrated convergent and divergent connections between ponors and springs term. The most important karst system flowing from north-east to the south-west discharges through the group of springs at Izvarna. (JURKLEW I CZ & MANGIN, 1996) The groundwater flow direction, the residence time distribution and the assessment of karst vulnerability to pollution were determined by tracer experiments in each system In this experiment, we have determined the mean transit time and the dispersion coefficients by repetitive checkings, through minimising a performance criterion. 2 I = f [h(t)-E(t)] dt (3) 0 The results obtained in cases of the three subsystems Valea Pargavului (figure 2) Bistrita (figure 3) and Balta (figure 4) are presented in the table I. SINKHOLE Yalea Bistrita Balta Pargavului Length (Km) 10.4 16.48 20 8 Mean transit time 14.85 1 3 18 14.09 (d) Velocity (m/d) 700 1250 1450 Dispersion 0.621 5.957 5.520 coefficient (m~/s) Tab l e I. Featurers of the J zvama karstic syste m In the figures 2 3 and 4 are depicted the impulse response functions for the three above subsystems. 0 .3 Elt) h( t) 0. 2 OJ h ( t ) 1 0 t i m e [d] 20 Fig u re 2 Tra n sfer fu n ctions E(t) a nd h(t) for t h e Valea P a r gavului expe r iment As it could be seen from these figures. the experimental data of the concentration of the tracers show some irregularities which are caused by either heterogeneities of the local geology or the existence of some different flowing pathways.(see fig.3). 0 .3 0 2 0 1 5 E ( t ) h ( t ) 1 0 Figure 3. Time variatio n and transfer fu n ctio n in of trace r co11ce11tration t h e Bist r ita expe r iment. Even in this case. each of the experimental curve can be splilted into many partial ones, every of these partial curve being filled to h(t) (ec 2) as above, t, (D/v), and q> 1 being determined. Here the subscript "i" stands for the number of peaks of the experimental curve and q>, is the partial recovery defined as the ratio of the area under the partial curve to the area of the total concentration curve 6 th Confere nce on Li me s tone Hydro l og y and F issur ed Med i a 1 n

PAGE 189

0 E(t) h(t l 1 0.8 0.6 O L. 0 2 10 20 t(d] 30 Figure 4. lmpulse-respo11se fu11ctio11s for the Balta experime11t. I C;(t)dt


PAGE 190

Le Transport de matiere dans les aquiferes karstiques par Valerie Plagnes, BRGM, Direction de la Recherche, Dep Hydrologie-Geochimie et Transferts 1039 rue Pinville, 34000 Montpellier, France Abstract How functionning (behaviour) and structure of karstic aquifers can be distinguished by studying the mass transport (geochemistry, natural tracing)? How this characterisation complete the study of energy transfert (hydrodynamical methods)? The aim of this purpose is to try to answer these questions. In fact mass transport results in the competition between several physical (convection, hydrodispersion) and chemical (dissolution ... ) phenomena from which we try to define their contribution to karstic underground flows. To complete this theoretical approach, a multi-disciplinary study has been realized on the Larzac Plateau (Aveyron France) Since the last five years, we collected many hydrogeochemical data on the most important karstic springs draining the area The most significative results are presented Resume Comment caracteriser le fonctionnement et la structure des aquiferes karstiques par l'etude du transport de matiere (geochimie, tra~age nature!)? En quoi cette caracterisation differe et complete celle realisee par l'etude c4J transfert d'energie (hydrodynamique)? L'objectif de notre etude est de repondre a ces questions En effet le transport de masse resulte de la competition de plusieurs processus physiques (convection, hydrodispersion) et chimiques (dissolution) dont nous cherchons a definir ('importance dans Jes ecoulements souterrains karstiques. Ces recherches reposent en plus d'une approche theorique, sur une etude multidisciplinaire realisee sur le Causse du Larzac (Aveyron, France). De nombreuses donnees hydrogeochimiques ont ete acquises ces demieres annees sur les principales sources karstiques qui drainent le plateau. Les resultats J es plus significatifs seront presentes. Une confrontation avec les donnees existant par ailleurs conduira a une tentative de generalisation de ces resultats a l'etude des systemes karstiques. 1. Introduction L'etude du karst impose de reconnaitre sa specificite et necessite par consequent d'employer des methodes particulieres adaptees a cette specifite Puisque la structure du reservoir ne peut pas etre determinee a priori, l'aquifere karstique doit etre considere dans sa totalite et etudie au moyen de methodes globales visant a comprendre son fonctionnement. La methodologie d'etude retenue s'appuie done sur une approche hydrogeologique fonctionnelle C'est pourquoi l'etude des systemes karstiques pour l'hydrogeologue, repose en principe, essentiellement sur analyse des ecoulements aux exutoires de ces systemes. Deux types de transferts caracterisent ces ecoulements : Le transfert d'energie qui correspond au passage d'une onde de pression a travers le systeme. A cette variation de pression est associee une modification du debit a l'exutoire et de niveau en tout point de la zone noyee. Les methodes hydrodynamiques (debit classes analyse des courbes de recession analyses correlatoires et spectrales ... ) sont utilisees pour etudier ces variations. Le transfert de matiere ou transport qui correspond au passage effectif de l'eau Ce deuxieme type de transfert est lie au precedent, mais ne Jui est pas necessairement correle Les tra~ages artificiels et l'hydrogeochimie (tra~ages naturels) sont utilises pour etudier ces deplacements d'eau A la suite de J. MARG AT, A MANG IN (1975) nommait tranfert le transfert d'energie et transit le transfert de masse (in BAKALOWlCZ, 1975) Notons que transfert d'energie et transport sont deux phenomenes distincts qui s'appliquent a des variables differentes (variations de debit et variations chimiques). Bien qu'ils ne soient pas regis par Jes memes lois et qu'ils possedent chacun leur propre evolution ces deux phenomenes ne sont pas pour autant independants. En effet, c'est le transfert qui provoque en general le transport Pourtant Jes variations chimiques dues au transfert de masse ne sont pas toutes causees par les modifications du debit. Seuls Jes processus determinant les variations du transport seront consideres ici Apres Jes avoir recenses nous verrons quels sont ceux que les differents traceurs chimiques permettent de caracteriser. 2. Bilan theorique des transferts de matiere Au cours du cycle hydrologique la composition chimique de l'eau qui s'ecoule aux exutoires des aquiferes karstiques peut presenter des variations remarquables La plupart d'entre elles sont determinees par la succession des evenements hydrologiques qui modifient la quantite et la qualite de l'eau presente dans l'aquifere Mais d'autres phenomenes entrent en jeu et particulierement les processus physico-chimiques intemes au systeme karstique En effet, les cinetiques des reactions chimiques qui se produisent dans l'aquifere sont du meme ordre de grandeur que les vitesses d'ecoulement. On retrouve done, en plus des phenomenes hydrodispersifs classiques (dispersion, 6 '" Conference on Limestone Hydro logy and Fissured Med ia 179

PAGE 191

diffusion ... ), des mecanismes specifiques au milieu karstique (dissolution, precipitation de carbonates). 2.1. La dissolution des calcaires La chimie des carbonates a suscite depuis longtemps l'interet de nombreux chercheurs parmi lesquels on note : H ROQUES (1962) M BAKALOWICZ (1979) et W. DREYBRODT (1988) entre autres II en resulte que "/'anhydride carbonique apparait comme le moteur de la dissolution des carbonates" (BAKALOWICZ, 1979). C'est pourquoi l'etat du systeme calco-carbonique vis a vis de l'equilibre determine le potentiel d'une solution a dissoudre ou precipiter des carbonates. Cet equilibre entre les trois phases du systeme calco-carbonique est decrit par un couple de variables : la pC02 equilibrante et le dpH ou lndice de Saturation vis a vis de la Calcite La cinetique de la reaction est tongue et compliquee. II en resulte que la dissolution des calcaires depend : de la quantite de CO2 dissous dans l'eau, du temps de reaction entre l'eau et la roche plusieurs facteurs entrent al ors en jeu : la cinetique des reactions chimiques les phenomenes physiques qui conditionnent l'etat de la solution a ( interface liquide-solide ainsi que la surface de contact, et surtout la vitesse des ecoulements dans le systeme. Tous ces phenomenes sont a tout moment en competition Leur interaction genere des eaux geochimiquement differentes selon les zones d'acquisition de la mineralisation (epikarst zone d'infiltration, zone noyee) selon les temps de contact les vitesses d'ecoulement et les possibilites de melange 2.2. Les phenomenes physiques qui interviennent sur le transport De nombreux auteurs les ont etudies plus particulierement en hydrogeologie : POREL (1988) HANICH (1991) MEUS (I 993) et d'autres encore ... Les deux principaux phenomenes rencontres sont la convection et la dispersion La convection Elle conceme le mouvement du fluide lui-meme qu1 entraine les elements dissous suivant sa vitesse propre. Cette vitesse est elle-meme soumise a une forte variabilite a la fois spatiale (vitesse differente selon les elements de la structure karstique parcourus) et temporelle (circulations beaucoup plus rapides en hautes eaux qu'en etiage) Bien qu'il ne puisse etre dissocie des effets de la cinetique chimique, ce phenomene conditionne le transport de masse. En effet ii permet le renouvellement de la couche d'eau en contact avec la roche (maintien d'un gradient de concentration) et ii assure !'exportation des matieres dissoutes. Si les deplacements de l'eau sont suffisamment lents pour que Jes reactions chimiques atteignent l'equilibre alors l'eau qui circule ne reagit plus avec son environnement. Par centre si les ecoulements sont plus rapides que les cinetiques chimiques les differents traceurs de l'eau ne seront pas equilibres avec leur milieu encaissant. Ce desequilibre renseigne sur certaines des conditions d'ecoulement a l'interieur du systeme. La dispersion hydrodynamique Elle conceme seulement ce qui se passe dans la phase mobile du fluide Sa contribution est relative face a !'importance des mouvements convectifs ces processus feront done l'objet d'une attention particuliere des lors que les ecoulements sont suffisamment lents On distingue : la diffusion moleculaire : transfert de matiere qui tend a equilibrer Jes concentrations entre Ies differentes parties d'un fluide selon Jes gradients de concentration Ce phenomene peut etre neglige dans de nombreux cas. Toutefois, comme en milieu karstique peuvent coexister des "paquets d'eaux" peu ou pas melanges possedant une origine differente la diffusion moleculaire doit etre prise en compte. la dispersion cinematique : due a la variabilite des vitesses d'ecoulement. C'est la structure du milieu qui est responsable de ces modifications-, ce phenomene est done accentue par la forte heterogeneite des aquiferes karstiques Des dilutions peuvent aussi perturber les reactions chimiques en cours De meme des phenomenes d'adsorption et de desorption de traceur a la surface d'une phase solide peuvent intervenir sur le transport de masse 2.3. Conclusion Que l'on etudie de tongues chroniques ou seulement !'impact d'un evenement hydrologique de courte duree (crue) aucune methode basee sur !'interpretation des donnees hydrochimiques ne permet en l'etat actuel des connaissances, d apprehender reellement le fonctionnement des karsts en revelant ('importance relative de chacun des processus intervenant sur le transport. La decomposition chimique de l'hydrogramme de crue des exutoires karstiques (BLA VOUX et MU DRY, 1983 ; MUDRY 1987) qui est une methode communement employee, ne peut donner d'informations sur ces phenomenes puisqu'elle tente d'expliquer les variations du chimisme par le seul phenomene de melange entre des eaux d'origines differentes (eau matricielle et eau de ruissellement souterrain) Aucun des processus physico-chimiques decrits ci-dessus n'est en effet pris en compte II est alors impossible, pour la plupart des cas, d'evaluer la contribution des conditions d'ecoulement face aux processus chimiques intemes au systeme dans les variations du chimisme aux exutoires karstiques. 3. Les differents traceurs hydrochimiques : de precieux informateurs Les variations du chimisme aux exutoires renseignent a la fois sur le fonctionnement et la structure de l aquifere Les informations fonctionnelles sont plutot liees a une notion de temps (vitesse d'ecoulement temp s de sejour inertie du systeme ... ) celles relatives a la structure sont davanta g e liees a la notion d'espace (zone d'acquisition du traceur 80 Pro c eed i ng s o f the 12 th Internat i onal Congre s s of Speleolog y, 1 9 97 Sw i tzerland V ol um e 2

PAGE 192

localisation des zones d infiltration pr e ferentiell e, estimation des zones de reserves .. ). La plupart du temps ces deu x types informations sont intimement lies Le tableau I resume l'apport de chaque traceur hydrochimique a la connaissance de l'aquifere karstique A tout moment, les informations Traceur Lieu d'acquisition Cond. electrique tout l'aquifere Temperature tout l aquifere dpH zone d'infiltration pC02 epikarst et sols Mg zone noyee encaissant dolomitique (epikarst ZN) Si02 zone noyee S04 zone noyee encaissant geologique Cl epikarst encaissant geologique N03 ep ikarst et so ls Na et K epikarst 180 pluie 3H pluie 02 dissous pluie relatives a la structure et au fonctionnement de la zon e noyee des karsts sont liees De plus un m e m e traceur peut marquer plusieurs phenomenes differents Information apportee temps de sejour, milieu encaissant origine profonde ou superficielle de l'eau vitesse de circulation temps de sejour vitesse de circulation (convection) temps de sejour ( cinetique) vitesse de circulation (convection) temps de sejour ( cinetique) temps de sejour de l'eau geologie du reservoir stockage dans un epikarst dolomitique . temps de sejour colmatage par remplissage detritique temps de sejour de l eau geologie du reservoir (gypse anhydrite) caracterise la zone d infiltration (reconcentration par ETR), indice de pollutions geologie du reservoir ( evaporites) conditions d'infiltration indice de pollutions caracterise la zone d infiltration (reconcentration par ETR), indice de pollutions adsorption, d e sorption temps de sejour altitude de la zone d infiltration temps de sejour conditions d'ecoulement (surface libre ou captive) Tableau I: Quels types d'informations apporte11t Les differe11ts traceurs chimiques? 4. Le Karst du Larzac : exemple Quelques 800 echantillons ont ete preleves ces demieres annees sur les sources karstiques qui drainent le Causse du Larzac (au pas hebdomadaire pour les sept principales et mensuel pour dix autres) Chaque echantillon est caracterise par des variables mesurees (conductivite electrique T pH HC03, Ca Mg, Na K Cl N03 SO4 SiO2) et des variables calculees (dpH pCO2 !EC .. ) ,., ., :.c .... Q 3 ~----------------, 4 5 0 2 5 Homecle 2 -, 1.5 0 5 -t ------, 4 4 5 -+ 4 40 I\ !I / 4 3 5 ] 430 'g_ C 425 ; 420 : ~ v 4 1 5 C 4 1 0 e 405 0 :!:::=::::=:==::::::::~::::::::==~--_J 4 00 Jun /95 A ou/95 Oct/95 Dec/95 Pour un meme evenement climatique les differentes sources suivies ne presentent pas toujours le meme comportement hydrochimique Cette difference peut etre observee non seulement sur la mineralisation globale de l eau ( Figure 1 ) mais aussi sur !'evolution des teneurs en certains element s chimiques (Figure 2) ~----------------, 640 0 9 ------~-------+-------~---0 8 1 0 7 : ; -:; : ,0 : : -;: : : 0 6 ______ ____ . ____ r") : I 1/ 1 ., 0.5 : _, fI I t :.C 0 4 Q I I 0 3 :~ Lavencou ~ 0 2 ~J ___ ~l _ ,_ 0 1 ______ -' _ 0 ='=======~===:::!..i: __J Jun /95 Ao0/95 Oc t / 95 D ec/95 630 E 620 C 610 ; ;;: 600 ~ "C C 590 e 580 Figure I : Variation de la conductivite e/ectrique pour deux sources karstiques du Larzac. 6 th Confere nce on Li m e s t o ne H y dr ology and Fissur ed Med i a 181

PAGE 193

20 18 16 tiD E 14 C .. l>Al ::;: 12 C .. "' C 10 0 .:: "' 8 .. = .. ,., 6 C 0 u 4 2 0 31/01/93 19 / 08/93 07 / 03 / 94 23 / 09 / 94 11/04 / 95 28/10 / 95 15 / 05 / 96 01/12/96 Figure 2 : Evolution des teneurs en Mg pour deux sources karstiques du Larzac La variabilite des teneurs en magnesium d'une source a l'autre peut etre reliee a de nombreux facteurs : existence de reserves dans la zone noyee de l'aquifere, stockage suffisamment long dans l'aquifere epikarstique dolomitique, effet de la dilution avec des eaux d'infiltration rapide augmentation du temps de sejour de l'eau dans la zone noyee. Le magnesium marque a Lui seul de nombreux phenomenes qui peuvent permettre d'apprehender le fonctionnement de la zone noyee des karsts etudies. Toutefois, ils sont la plupart du temps superposes et bien qu'ils soient tous a prendre en compte, leur importance est relative d'un systeme karstique a un autre. II n'est done pour l'instant pas possible de distinguer la participation de chacun de ces phenomenes dans la composition chimique de l'ecoulement. Nous sommes confrontes au meme probleme en ce qui conceme les autres traceurs chimiques. Par exemple, les chlorures peuvent caracteriser l'eau en provenance de la zone d'infiltration, mais aussi etre apportes par des activites anthropiques ou par la nature geologique du reservoir II en resulte qu'une bonne connaissance generate (geologique, structure ... ) des systemes etudies est indispensable. C'est ainsi qu'en procedant par elimination et en recoupant les differents indices apportes par chaque traceur chimique, Les processus qui dirigent le transport pourront etre mis en evidence et etre relies au fonctionnement global de l'aquifere 5. Conclusion Les demarches adoptees jusqu'a present pour interpreter les variations du chimisme aux exutoires des aquiferes karstiques ne permettent pas d'obtenir des informations sur le fonctionnement reel des systemes etudies Une approche differente semble alors necessaire. Elle devra prendre en compte tous les processus physico-chimiques qui se produisent a l'interieur des karsts pour rendre compte de la specificite de ce milieu References BAKALOWICZ, M 1975. L'evolution du chimisme de l'eau d'une source karstique. 3eme convegno internazionale sulle acque sotterranee, Palermo, 1-5 novembre 1975 BAKALOWICZ, M. 1979. Contribution de la geochimie des eaux a la connaissance de l'aquifere karstique et de la karstification. These de Doctoral d'Etat es Sci nat, Univ Paris 6, 269p BLAVOUX, B ., MUDRY, J. 1983 Separation des composantes de l'ecoulement d'un exutoire karstique a !'aide des methodes physico-chirniques. Hydrogeologie, 4, 1983 pp 269-278, 8jig, I tab/. DREYBRODT, W. 1988. Processes in karst systems. Ed. Springer-Verlag, Berlin, 288p, I 84 jig. HANICH, L. 1991. Les tras;ages d'essai en aquifere heterogene : methodes, interet et limites. These Univ Lille Flandre-Artois, l 55p MANGIN, A 1975. Contribution a l'etude hydrodynamique des aquiferes karstiques. These Doctoral Sci nat ., in Ann. Speleol., 29, 3,p283-332; 4,p495-60I; 30,I,p2I-I24. MEUS, Ph. 1993. Hydrogeologie d'un aquifere karstique dans Jes calcaires carboniferes (Neblon-Anthismes, Belgique). Apport des tras;ages a la connaissance des milieux fissures et karstiques. These Univ de Liege, 323p MUDRY J. 1987. Apport du tra9age physico-chirnique nature! a la connaissance hydrocinematique des aquiferes carbonates These de Doctoral d'Etat, Univ Besan9on Franche Comte. RICARD, J. et BAKALOWICZ, M 1996 Connaissance, amenagement et protection des ressources en eau du Larzac Septentrional Aveyron (France). Rapp BRGM R38 953 94p 200 annexes (format A4) 7 planches hors texte (format A3}, 5 planches hors texte (format Al) ROQUES, H. 1962. Considerations theoriques sur la chimie des carbonates. Ann. Speleol. 17, I pi 1-41 ; 2, p24!-284; 3 p463-46 7. 182 Proceedings of the 12 th Internat ional Congress of Speleology 1997, Sw i tzerland Volume 2

PAGE 194

Modelling of suckwell recharge in carbonate terrain by Chris C. Smart Department of Geography, University of Western Ontario, London, Ontario N6A 5C2, Canada. SMART@SSCL. UWO.CA Abstract Suckwells are artificial shafts excavated in soft carbonates to relieve surface flooding throughout the Caribbean They significantly enhance ground water recharge but may also convey surface contaminants directly to the unsaturated zone Physically based models of suckwell recharge are difficult to implement A simple linear model of suckwell drainage through the base of the well shows that drainage rate is determined by readily measured nominal volume and a recharge coefficient. Field slug injections confirm the generality of the model, but demonstrate systematic departure from predictions ; more water is drained from the borehole early in the test and less later. 1be error is interpreted as indicating radial drainage and return of some fraction of radially discharged water. The basal permeability coefficient can be determined by stepwise analysis of the data to zero head Radial discharge and permeability can be estimated at maximum head by subtracting calculated basal drainage from observed total drainage. In general the radial permeability coefficient is less than the basal permeability coefficient However the magnitude, time-constant and direction of the radial drainage needs direct investigation. There are indications of significant heterogeneity in the radial permeability, and there may be dependence on antecedent moisture Higher quality test data are necessary to adequately resolve the problem 1. Introduction Suckwells are artificial vertical shafts excavated in soft coral rock throughout the Caribbean to relieve ponding and flooding. They are commonly located in depressions or intermittent channels, and are excavated until a suitable enlarged fissure or suck is found Suck wells are typically cylindrical in form 2-20 m deep and 0 5-0 75 m in radius, but may enlarge at their base They are commonly protected by a low wall or grid at the surface and masonry-lined down to solid rock They typically contain sediment organic flood debris, and garbage Suckwells can significantly enhance ground water resources. By transmitting surface water rapidly to depths near the water table they enhance ground water recharge and reduce evaporative losses Interception of surface streams may prevent discharge of fresh water to the ocean However, Suckwells also allow surface water contaminants more direct transmission to the aquifer than under natural, soil-mediated recharge Increasing pressure on water resources makes it imperative to develop improved techniques of management and assessment. Smart and Ketterling (1997) report on suckwell management strategy The present paper attempts to increase understanding of Suckwell hydrology by development of a drainage model. 2. Suckwell model Point recharge in a highly heterogeneous unsaturated zone is beyond the scope of conventional ground water models In addition, the large volume, rough-hewn form and unsaturated position of Suckwells combined with uncertain moisture content confound their assessment using any conventional ground water testing technique. However, it is possible to develop simple models of Suckwells as a guide to understanding and management A suckwell can be considered a permeable cylinder of length h.: radius r and basal radius rb filled with water to height h (Figure 1) Surface runoff enters as Qin and the suckwell drains (Q...J by basal (QJ and radial (Q,) infiltration 2r A h Figure I. Model Suckwell During a runoff event the suckwell fills when Q .. exceeds QOOt Consider a simple field slug test, where a quantity of water has been injected into the well and is draining away Drainage is dependent on hydraulic gradients and unsaturated hydraulic conductivities which will vary throughout the well and with time The key physical variables defy field monitoring, and a simpler approach is required if widespread application is anticipated Suckwells are constructed to drain through their beds into the basal fracture This suggests a simpler approach might be to consider the suckwell as a cylinder with a leaky bed and 6 'h Conference on Limestone Hydrology and F i ssured Media 183

PAGE 195

impenneable walls. The head, his the primary driving force and the Q, is ignored. Alluvium on the floor implies Qb will be laminar so that the suckwell can be considered a linear reservoir (e g. Chow, 1988) in which (1) where ab is a basal infiltration coefficient with dimension of r 1 A useful management index is basal discharge at capacity (h=hJ which is given by abV c where Ye (=m/hJ is the nominal volume" of the suckwell. Linear reservoir drainage is exponential in fonn (e g Chow 1988). so that (2) The expected exponential fonn of water level during drainage is testable by observation. Significant departure from exponential fonn will indicate weakness in the simple basal drainage model. Equation (2) can be linearised, allowing a (= ab (rb /r) 1 ) to be obtained by linear regression 1/ a constitutes a "residence time ". Measurements of r and rb allow ab to be estimated from passive drawdown data 3. Field testing Slug tests were made by tnJecting a large volume (-13m3) of water from two water tankers (bowsers) provided by the Barbados Water Authority (BWA). and observing drawdown using a standard water level tape (Smart and Ketterling 1997 ). Suckwell radii and basal radii were estimated to 10 cm and used to calculate the cross sectional area, A (.1 m2) Tests made in Suckwells near Barbados Water Authority supply wells at Buttals and Codrington showed relatively rapid infiltration Replicate slug tests were run a few days later concurrent with tracer injection. Assuming an injected volume of 13 m 1 ( 3000 imperial gallons) the head as measurement began (ho ) suggests that 50% of injected water had already left the Suckwells as measurements began (Table 1 Vo) The lost volume multiplied by an assumed storage coefficient of 0 2 (Stanley 1972) gave an equivalent volume of saturation in the surrounding aquifer from 24-44 m 1 equivalent to a zone about 1 m in thickness around the well r A rh Ah hn V n Test m m l m m l m m l Buttals l 0 15 1.8 1.2 4.5 4 91 7 7 Buttals 2 0 15 1.8 1.2 4 5 4 72 8 2 Codrington 1 0 6 1.1 0 6 1.1 4 28 4 2 Codrington 2 0.6 1.1 0 6 1.1 4 50 4 9 6,-------r--.-,----.--r----.----,.----, Bu tt a ls Test 1 5 h = 4 91 e 0 045.:5 t 2 0 Time 4 (min ) 6 0 80 Figure 2 Example of suckweU drawdown data for Buttals Test 1, with fitted exponential curve. Exponential fits to the corrected water level (h(t) in m ) against time (t in minutes) gave correlation coefficients (r) in excess of 0 99 indicating good exponential fonn (e g ., Figure 2) a values obtained were from 0 04 to 0 08 min 1 (Table 1), giv i ng residence times (1/a ) from 12 to 22 min. ab varied between .01 and .08 min 1 Higher values on second tests suggest the antecedent moisture increases the basal penneability The results support the simple steady state basal drainage model. 4. Analysis Inspection of the drawdown curves suggests that the model predictions systematically depart from observations. Figure 3 shows the residuals between observed water level and the exponential model are initially positive become negative at heads of 3 1 and 4 2 m for Buttals and Codrington respectively then becoming positive at 2 and 1.2 m. This pattern suggests that radial flow is active near maximum head and reverses as water level falls in the suckwell Residual distribution appears to vary more between test sites than within replicate tests at each site This despite different initial volumes and antecedent conditions suggests that storage may well be influenced by zones of preferred radial penneability. a (Xh (Xh( oo) (X, (0) h min 1 min 1 min 1 min 1 m 0 0453 0 0177 0.018 0.018 2.45 0 0517 0 0202 0 023 0 0060 7 36 0 0622 0 0622 0 060 0 010 6 6 0 0796 0 0796 0.07 ? 0.002? 38.5 ? Tobie 1: Resula of the suckweU injection tests. rand A, and r and A are well and basal radius and area respectively. h 0 is the inilial head at time zero and V 0 is the initial volume at the start of the test. a' was detennined from regression; a. is the overall basal injiltraJion coeffident. a.( 00 ) is the basal injiltration coefficient obtained by extrapolalion. a, (0) is the radial infiltration coefficient obtained at the start of observations. h' is the head above which radial discharge dominates. 184 P r oceed in gs of t he 1 2 "' In te r na ti o n al Cong r ess o f Spe l eo l og y, 1997, Switz e rland Volum e 2

PAGE 196

Butta ls 1 I I ,.-._ 1.0 But t a l s 2 I I E I Codr i ngton 1 I '-" I Codr i ngton 2 I I I -0 I 0 I 0 5 I (l) I .c I I I I I 0 I ... ~ ::J ... ; ; ; ..,. .... .. .. ..... -0 !' .. \ .. -<'" .. ... (/) 0 0 / (l) .,, / Q'.'. -0 5 0 2 3 4 Head (m ) Figure 3. Distribution of residual head from exponential jiJ to suclcwell slug tests. An abundance of water dripping from the walls as head falls suggests that the concept of temporary radial storage is not unreasonable, but its magnitude is uncertain Ignoring heterogeneity. consider the model in Figure l The radial discharge Q, is obtained by integration of the hydraulic head over the range zero to h to give Q = 1t r a h 2 r r (3) Equations (1 ) and (3) show that Q, scales with rh 2 while Q b scales with rb 2 h. Therefore the relative influence of Q, is minimised at low head. Providing the radial reflux is not significant at the last stages of drawdown, it may be possible to extract a reasonable estimate of a.b from the last few data points Every observational pair {(h; t;),(h; + i t 1 + 1 )} allows an estimate of the basal permeability coefficient (4) Figure 4 shows a plot of stepwise estimates of a.b against head for the first test at Buttals At initial head, the a.b value is 0 068 min 1 but includes maximal rates of Q, At the limit as h -> O, a.b approaches 0 018 min 1 and this value will be incorporated into further modelling. The minimum value at 2m i s assumed to result from active radial reflux ( Q, < 0) The step-a. curves all show considerable irregularity. Part of this is not unexpected error in measurement and is characteristically oscillatory as both time and depth are "closed measurements; e g . a positive error forces a serial negative error. However, both Buttals tests show irregularity al around four metres possibly a zone of heterogeneity Total discharge can be estimated from 1he drawdown data Q,-= nr2oh/01. Qb can be escimated from equacion (I) using !he limicing value of a.b( 00 )=0.018 min 1 estimaced above Q can then be calculaced as the residual ; Q, =Q.Qb. Figure shows the respeccive discharge escimates for Buttals tes1 I The irregular nature of QT (and Q,) arises from the sensitivity of the step method to errors in measurement of h Q, appears to become negacive at around three metres head, when residuals are most negative a. is estimated from the first data point and shown in Table I as a.,(O) values are somewhal sensilive 10 seleccion of a. ~ bu1 are on !he same order or smaller I Simulations of suckwell recharge and drainage in response to a storm do not incorporate either radial reflux or antecedent moisture, as there is not yet any effective way of parameterising these processes Interestingly, the radial discharge dominates above a critical height, considerably below the top of the suck well. The critical height, h' can be estimate by equating (I) and (2) and solving for h = r/a.b / ra. (Table 1) 0 070 -,----.----r----,---~-----~ 0 060 0 050 E o o4o 0 0 030 .c ~.020 0 0 1 0 Butta ls Test 1 0 000 +----,----,--.----.----.--~ 2 3 4 5 H ead (m ) Figure 4. Stepwise estimales of a., basal penneabiliJy coefficient from observed drawdown data. A smoothed curve has been Jilted through the points. 30,----,---,----.---~--,---~ 25 .:'20 I (/) 1 5 ~10 .... _g 5 u (/) 0 0 0 I ----,; ~-: ........ \ ,, ,.., ,, , f' / \ I, I ---:::-~-: ,-;----2 4 6 Head (m) Figure 5. Observed total discharge Q 1 for Buttals test 1, with derived basal (QJ and radial discharge Q, so.-~--.--r---,--,---.----.--,~..-,--T-,---.--.-~ ,......._60 I (/) .......,,40 (I) 0, \.... 0 t 20 (/) 0 O-t0.:.......-, 2' 0 -r-4' 0~ -6-.-0-,-8,0-r--1 'o-,-_:.:.:;=;"" 1 ~411" 0 ~ Tim e ( m in. ) Figure 6. Simulaled stonn drainage through Buttals Suclcwell based on Test I parameters. 6 '" Conference on Limestone H y drolog y and Fi ss ur ed Med i a 185

PAGE 197

5. Conclusions The assumption of dominant basal penneability leads to a useful first approximation of suckwell drainage, allowing practical assessment and classification of Suckwells in terms of their residence time It also leads directly to management principles from which volume critical and contaminant critical sites can be identified (Smart and Ketterling 1997). The dual discharge model allows tentative estimation of basal and radial penneability coefficients and simulation of suckwell perfonnance It suggests that radial discharge may dominate under high head conditions This surprising conclusion arises from the assumption that radial hydraulic gradient depends only on the head in the well and that the radial permeability is constant at the rate detennined at maximum head in a slug test These assumptions are contradicted by reflux discharge when the radial hydraulic gradient is reversed as head falls Much more careful monitoring, analysis and modelling are needed if this phenomenon is to be better understood The analysis is dependent on data quality and a number of assumptions. Much of the noise in the analysis arises from inadequate drawdown data. A pressure transducer-logger system would provide superior data. A key assumption has been that the permeability is spatially unifonn and constant in time This is entirely unreasonable in a heterogeneous carbonate. Apart from the basal fracture to which the well is connected, there is evidence that there are zones of enhanced penneability down the well column Better drawdown data might allow identification of these zones Replicate tests showed higher basal permeability coefficient which implies that the permeability tenns depend on antecedent moisture Correspondingly, they would be expected to evolve during the duration of the test. Many of these constraints could be easily addressed with further field experiments. Acknowledgments This work would not have been possible without the assistance of the Barbados Water Authority and Klone-Krippen Ltd Financial support was also provided by the Natural Science and Engineering Research Council of Canada References Chow, V T., Maidment, D R. Mays, L.W ., 1988. Applied Hydrology McGraw Hill. Freeze, R A and Cherry, J C., 1981. Ground water. Prentice Hall. Senn, A 1946. Geological investigations of the ground-water resources of Barbados, B. W.l., Report to the British Union Oil Company. Smart, C C., and Ketterling, 8.8 1997 Preliminary assessment of the role of suckwells in karst water resources VI Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, Springfield Missouri. In press Stanley 1972 Barbados Water Resources Study Volume III, Water resources and Hydrogeology Unpublished Report to Barbados Water Authority. 186 Proceed in gs of the 12 '" Internat i onal Congress of Speleology 19 97, Sw i tzerland Volume 2

PAGE 198

The Interpretation of a High Water Tracer Test i n the Danube Aach-System (Western Swabian Alb Germany) Andreas Werne r 1, Heinz Hotzi2 Department of Applied Geology, University ofKarlsruhe, Kaiserstr. 12, D-76128 Karlsruhe Werner Kail Milhlematten 5, D-79224 Urnkirch Abstract A multi-tracer test was carried out 1995 in the Danube Aacb System (Western Swabian Alb, Germany) during a high water event. The evaluation was performed with different model approaches. The comparison of the results has shown that the used analytical models are able to describe mathematically the breakthrough curves of the used tracers The tracer transport could be restricted on convection and dispersion processes Due to the existing unsteady flow conditions the determined analytical values must be checked with a numerical model. Zusammenfassung In der westlichen Schwabischen AJb (Deutschland) wurde 1995 im Donau-Aach-Gebiet ein Multitracerversuch wahrend eines Hochwasserereignisses durchgefiihrt. Bei der Auswertung wurden verschiedene Modellansatze vergleichend eingesetzt. Es konnte gezeigt werden, daB die eingesetzten Modelle in der Lage sind die Durchgangskurve der Tracer mathematisch zu beschreiben Der Transport im Donau-Aach-System unterliegt dabei nur Dispersionsund Konvektionsprozessen Bei stark instationaren Verhii.ltnissen miissen jedoch die analytisch gewonnen Parameter numerisch iiberpriift werden 1. Introduction The Danube-Aach-System is located in the South-West of the Swabian AJb (Germany). The Danube river crosses this region from the West to the East. For about 160-180 days per year the water of the river is vanishing completely. This water reappears in the largest spring of Germany, the Aachquelle (Aach spring). It is located about 12 km south of the main sinking wholes. Consequently the region is an object of research since a long time (BATSCHE et al., 1970; HOTZL, 199 I). Thirty-nine tracer experiments have been carried out in the past 120 years. The first quantitative one was already done in 1877 with chloride About the interpretation of the tests in the area of the main sinking wholes was already reported (WERNER et al., 1997a). In the context of these investigations an additional tracer experiment was done in the summer of 1995. This experiment was carried out as a multi-tracer test. This means that two tracers were injected simultaneously In the following, the paper reports about the used methods and the resulting problems during the interpretation of the test. 2 The tracer test In connection of the reinterpretation of the tracer experiments of the Danube-Aach-System (WERNER et al. 1997) an additional test was performed to get more data for the mathematical interpretation This was necessary because much of the existing data is of qualitatively character The experiment was started during the time of total water absence of the Danube river followed by a high water event. Two tracers (uranine and lithium) were injected simultaneously (multi-tracer test) for recognising processes, which based on the different transport behaviour of the used tracers. The input (Table I) was performed in August 28 1 \ 1995 in the area of the main sinking wholes near Immendingen (.,Im Briihl") Firstly, the uranine was injected, immediately (ea. 15 minutes later) followed by the lithium (25kg lithiumchloride) At the same time the detection of the tracers was started with automatically working probe samplers at the Aach spring. During the injection an increasing Danube water level could be observed. A few hours later the injection wholes were flooded Email : andreas.wemer@bio-geo.uni-lcarlsruhe.de Email: heinz hoetzl@bio-geo.uni-karlsruhe de Due to this high water the first reappearance of the tracer could be detected after only 22 hours later. Exoerime n t o f 1 995 Inj ectio n .,Briihl" near Immendingen 28. August 1995 Tracers 8 kg Uranine 25 kg Lithium(-chloride) D etecti o n Aach Spring 28.August. 16 September 1995 Distance from ,,Briihl" 11.8 km Tracer Recovery 81% Uranine 35%Lithium Table I: The tracer experiment of 1995 The breakthrough curves of both tracers (Figure I) were characterised as a double peak About I 00 hours after the input the tracer concentrations have reached the values of background concentrations (lithium) or minimum detection limit (Uranine) The curve of the high water is shown in Figure 2. The discharge curve of the Danube is formed as two waves The Aach spring does not reflect such behaviour The passage through the karst system has a compensating effect. ,. ,---------------------, _ } Ulhlum .i 3 Tim (h] after Injection Figure I: Breakthrough curves of the injected tracers at the Aach Spring 6 th Conference on Limestone Hydrology and Fissured Media 187

PAGE 199

The double peak of the uranine breakthrough is not the result of the double wave of the high water. The second flood wave begins nearly at the same time as the second uranine peak. However, the time to flow from the injection place to the Aach spring takes at least about 20 hours i Aach Spring / i t u~ -.__._--1 ~ \ I \ Uranlne I / L. -,.cl~--~ --= ~=! =m=o:mzm:-.li . = .... Figure 2: The discharge curves of the Danube and the Aach sprint!. The uranine breakthrough curve is shown for comparison. 3. The model approaches The evaluation of the tracer test was carried out for three aspects: Comparison of the both tracers uranine and lithium Comparison of two different model approaches during the evaluation of the uranine breakthrough curve Comparison of steady and unsteady flow interpretation Altogether three different model concepts were used ln the following these methods are characterised briefly All of the used models are implemented in the computer program ,,Traci" 4 3.1. The Multi-Dispersions-Model (MDM) The Multi-Dispersion-Modell (MALOSZEWSKJ ET AL 1992) is a variation of the classical convection-dispersion model from Lenda & Zuber (1970) The resulting breakthrough curve of the tracer experiment is seen as the outcome of different flow paths. Step by step the breakthrough curves of the individual flow paths and the parameter of convection (mean transit time respectively mean flow velocity) and dispersion (dispersivity) processes are determined. The procedure during the evaluation is described in WERNER ET AL. (I 997a). The first application (MALOSZEWSKJ ET AL, 1992) in a karstic area was performed in Styria ( Austria ). 3.2. Single Fissure-Dispersion-Model (SFDM) This model (MALOSZEWSKJ, 1994 ; MALOSZEWSKJ ET AL 1994 ) was originally developed for the interpretation of tracer tests in fissured aquifers. The SFDM is a double porosity model which uses two parameters for convection and dispersion and one parameter for describing a possible diffusion into the rock matrix With the knowledge of the matrix porosity and the diffusion constant of the surrounding rocks it is possible to calculate a mean fissured width The SFDM is only valid if the mean transit times are less than one month. 3.3. Numerical 1-D-FE-Modell The analytical approaches after 3 1. and 3 2 are based on a stationary behaviour of the regarding system If the experiment is instationary (of discharges during time) mean values of the discharge will be used The numerical model was developed for The discharge dala is provided by the Landesanstalt fiir UmwellschulZ (LfU) Baden-Wiirttemberg Demo version of this program are available from the authors checking the analytical determined results. The model itself is described in WERNER ET AL. (1997b) 4. The evaluation of the tracer experiment 4.1. Comparison of the used tracers (Fig. 3) The multi-tracer test was performed under the intention to detect some possible differences in the transport behaviour of the used tracers. This comparison is only possible under identical hydraulic conditions Therefore the tracers must be injected simultaneously For technical reasons it was necessary to inject the tracers successively during this experiment. The breakthrough curves (Fig I) already show the parallel course of both tracers Be caused by the background concentration of lithium (about 2-3 mg/m 3 ) in the system, the lithium breakthrough curve shows an indistinct behaviour. Concentration [mg/m 3 1 10 8 6 I 4 \ I 2 /\ 0 20 40 60 Time [hi after lnjecktion Uranine PEAK I 29 7 397 Mean transit time [h] : Flow velocity [m/h]: Dis [m] : 31 Concentration [mg/m3I 5 4 3 2 0 20 30 40 50 Time [hi after Injection 80 PEAK II 39 2 301 49 60 Lithium PEAK I PEAK II Mean transit time [h]: 29.5 39.2 Flow velocity [m/h] : 397 301 Dis ersivit [m] : 31 0 59 7 100 70 Figure 3.: Results of the evaluation of the Multi-tracer-Test with the MDM ( a background value of 2.5 was subtracted from the lithium-values) The interpretation was performed with the MOM for both tracers The tailing of the uranine curve was not fitted (see 4 2) because the Lithium curve does not show this behaviour. The 188 Proceed in gs of the 12 th Internat i onal Congress of Speleo l ogy 1997, Sw itz erland Volume 2

PAGE 200

determined parameters (Figure 3) are comparable. There is only a difference in the dispersivity values for the second peak It is much higher for the lithium as for the uranine peak. The reason is not caused by the transport; the high background values in contrast to the maximal concentration (about 5 mg/m 3 ) are responsible for this effect. 4.2. Evaluation of the uranine breakthrough curve with the MDM and the SFDM Because of the extreme tailing the uranine breakthrough curve could not be fitted satisfactorily with two fit curves (Fig.3, above ) Therefore the SFDM was used next to MDM The SFDM is able to describe mathematically this kind of tailing effects. The evaluation (Fig. 4) showed that a complete description of the breakthrough curve is possible with both models: MDM: Fitting of the breakthrough curve by superposition of five single peaks SFDM: Fitting of the breakthrough curve by two peaks. The first one is calculated using the MDM, the second one using the SFDM What is now the ,, right model ? From the visual impression and without background knowledge the SFDM seems to be the better one However the following reasons show that the fitting with the MDM is more plausible: Older tracer tests (WERNER ET AL 1997, BARCZEWSKI ET AL 1996) never showed such tailing effects. These experiments were performed under flow conditions with lower flow velocities All experiments also showed a superposition of at least two flow paths The comparison of the two tracers uranine and lithium yield no different transport behaviour showing that there exist matrix diffusion processes Therefore, the high water is responsible for the tailing effect of the uranine breakthrough curve. The flood event would imply that the flow takes place in areas of the aquifer, which are normally not participating at the flow system The several peaks of the total breakthrough curve at the Aach spring resulted from these additional flow paths The tailing is a result of the overlapping of the peaks of the individual flow paths These results, together with the conclusions of 4 1 ., elucidate that only dispersion and convection processes determine the character of the transport behaviour in the karst system 4.3. Steady/unsteady flow evaluation of the tracer test The whole tracer test was characterised by the high water event. On the other hand the evaluations with the analytical solutions were performed with mean values of the discharge. For this reason a numerical model was used in addition to prove the determined analytical results Toe evaluation (Fig. 5) was limited to the first 50 hours of the experiment. The analytical fit with the MDM shows a adequate description of the breakthrough curve (CD). As expected, the verification with the numerical model with stationary discharge conditions (mean values) shows the same result(@). The unsteady numerical calculation with the same p~ameters lea~ to considerable differences A satisfactory fit 1s only possible with a modification of the parameters The differences between the mean transit times are negligible ( 28.3h c 29h respectively 39 2h c. 39.8h) The differences of the results of the different models are recognisable for the dispersivity values. The increasing flow velocity during the experiment causes increasing differences in the flow b ehaviour. This is reflected by the dispersivity values. These differences (31 3m c. 29m respectively 49.2m c. 80m) must be considered if they are important for the regarding problem Concentration [mg/m 3 ) 10 , 8 I 6 /1 4 1 I\ 2 \ - "O 20 40 60 80 100 Time [hi after Injection MDM PEAK I II m IV V Mean transit time [h]: 30 39 51 64 77 Flow velocity [m/h] : 397 301 232 1 84 151 Dis ersivit [m]: 31 49 47 50 20 Concentration [mg/m 3 1 10 I C B I 6 I 4 1 \ 2 )\ '-O 20 40 60 80 10( Time [hi after Injection SFDM PEAK I PEAK II MDM-Model SFDM-Model Mean transit time [h]: 30 39 Flow velocity [m/h]: 397 301 Dispersivity [m] : 31 49 Fissure width 2b [mm] 2-4 Figure 4.: Results of the evaluation of the uranine breakthrough curve with the MDM and the SFDM 5. Conclusions The interpretation of karst tracer tests is seldom unequivocal The experiment in the Danube-Aach-System shows that it is possible to describe the breakthrough curves correctly with different model approaches. The advantages and disadvantages of the different used models are difficult to define. It depends frequently on other results and on background information The models are on the other hand very important tools for the estimation of the dynamic of a karst system The analytical solutions (MDM and SFDM) are suitable for a fast evaluation of a tracer experiment. The determined results are applicable if the boundary conditions of the used model were considered The MDM is also useful under unsteady flow conditions. For the estimation of the error of the parameter it is necessary to execute a numerical calculation The determined results are also important for the understanding of the Danube-Aach-System. However, this is not the subject of this paper. It shall be referred to WERNER ( 1997). 6 th Conference on Limestone Hydrology and Fissured Media 189

PAGE 201

References BARCZEWSKI, B ; KASS, W. ; SCHMID G & A. WERNER (1996): Neue Moglichkeiten und Anwendungen bei der Grundwassermarkierungstechnik Wasserwirtschaft, 86, I 20-24, Stuttgart (Germany). B ATSCHE H., BAUER, F BEHRENS, H., BUCHTELA, K., DOMBROWSKI, H.J ., GEISLER, R ., GEYH, M A., HOTZL H HRIBAR, F KASS, W ., MAIRHOFER, J., MAURIN, V MOSER, H., NEUMAIER, F ., SCHMITZ, 1., SCHNITZLER, W.A. SCHREINER, A., VOGG, H., ZOTZL, J ( 1970) : Kombinierte Karstwasseruntersuchungen im Gebiet der Donauversickerung (Baden-Wiirttemberg) in den Jahren 1967-1969 Steir Beitr Z Hydrogeologie, 22, S. 5-165, Graz (Austria) HOTZL, H ( 1992):The lost Danube : Karst Hydrogeology of the Western Swabian Alb, South Germany. IAH, Int. Cont. to Hydrogeology, Vol. 13, 179-194 Hannover(Gennany). iLNDA, A. & ZUBER, A ( 1970): Tracer dispersion in groundwater experimentes Isotope Hydrology 619-641 IAEA, Wien (Austria). MALOSZEWSKI P (1994) : Mathematical Modelling of Tracer Experiments in Fissured Aquifers Freiburger Schriften zur Hydrologie, Band 2, I 07 S Freiburg (Germany). MALOSZEWSKI P HARUM, T., BENISCHKE R (1992): Mathematical Modelling of Tracer Experiments in the Karst of Lurbach System in BEHRENS H., BENISCHKE, R 8RJCEU, M HARUM, T ., KASS, W., KOS! G ., iLDITZKY, H.P ., UIBUNDGUT, CH., MALOSZEWSKI, P MAURIN, V I c !! ::::, 0 011. 70 \ 7 0\ 1 _\ I 0\ I a\ r D~ I I -But-Flt a lkanlM / "'~ /4 ").,. "\ "I( 0 DaoD ~ 0 i / D llme [hi alter Infection Analytical Model (steady flow) (D Mean transit time [h]: Dispersivity [m): 0 a 0 a PEAK I 29 7 31.3 PEAKII 39 2 49 2 I Beat-Flt a Uranln r 7 I 0 .. I I D 7 I I ::/ .\ 0 a a " \_ a D / ""D \ I "'' / '\: D Dao"I'---...__ ua .. .. llmo [h] attar lnJectlon Numerical Model (unsteady flow) with analytical values RAJNER V RANK, D REICHERT, B. STADLER H ., STICHLER W ., TRIMBORN, P ZOJER, H ZUPAN, M (1992) : Investigations with Natural and Artificial Tracers in the Karst Aquifer of the Lurbach System (Peggau-Tanneben-Semriach, Austria) in Steir. Beitr z Hydrogeologie 43, S. 116-143 Graz (Austria) . MA_LOSZEWSKI P.; BENISCHKE, R ; HARUM, T & H ZOJER (1994) : Esnmatton of Solute Transport Parameters in Heterogenenous Groundwater Systems of a Karstic Aquifer Using Artificial Tracer Experiments Water Down Unter'94 Vol.2 Part A, 105-111 ; Adelaide (Australia) WERNER, A .. ; HOTZL, H ; KASS, W & P MALOSZEWSKI (1997A) : Interpretation of tracer experiments in the Danube-Aach-System with analytical models Procededings of the 5"' Symposium on Karst Waters&Environrnental Impacts 1995, 153-160, Rotterdam (The Netherlands). WERNER, A. HOTZL, H & MALOSZEWSKI, P (19978) : Interpretation of Tracer Tests in Karst Systems with unsteady flow conditions. IAHS-Workshop W2: Karst Hydrology, Rabat (Marocco) (in preparation) WERNER, A (1997): Auswertung van ktinstlichen Markierungsversuchen zur Quanlifizierung der Abflu6dynarnik in Karstsystemen, Schrift. Angewandte Geologie Band 51, Karlsruhe (Germany). (in preparation) 10 [ E i' c I!!. ::::, 0 .. _/ 0 ':,-, / "\ 7 D\ J _\ I 0\ I a\ / "'q, r "~ ,(" I V I D a llme [h] attor Injection I -a .. t.ftt l I a Uranlne .... \ \a 0 Doan I ~ .. Numerical Model (steady flow) with analytical values (2) Mean transit time [h]: Dispersivity [m]: J'I / D \ PEAK I 29 7 31.3 1 l [. .s c ... 5 1 I I r I I \ \ / \a_ D I' '-._;/ PEAKII 39 2 49 2 I I -Boot-At I ..... \ \_ '-t. D "fI I I I D r-,._-ua i ,/ .. .. ., .. .. llmo [h) .,._ 1,ioctlon Numerical Model (unsteady flow ) Mean transit time [h]: Dispersivity [m]: PEAK I 29 7 31.3 PEAKII 39 2 49 2 Mean transit time [h]: Dispersivity [m]: Figure. 5: Comparison of the evaluation with steady and unsteady flow conditions 190 Proceedings of the 12 th International Congress of Speleology, 1997, Sw i tzerland Volume 2 PEAK I 28.5 36 PEAK II 35 4 71

PAGE 202

Storm pulses, thr~shold~, and f lui~ mechan i cs in the t r ansport of elastic sediments 1n limestone aquifers b y W illiam B W h ite & Eliza b eth L. White Department of G eosciences and Environmental R esources Research Institute T h e Pennsylvania State University, University Park PA 16802 USA Abstrac t Active conduit systems carry a flux of water from recharge regions to discharge downstream at karst spring s. They also carry a flux of elastic sediments. Transport of elastic sediments is episodic Sediment piles are moved mainly during extreme floods Relationships between sediment particle sizes and stream flow can be used to determine paleoflow characteristics of caves passages from the elastic sediments and also to calculate the magnitud e of flood pulses needed to maintain the sediment flux 1 I ntroduct i on All caves that are part of active karst drainage systems contain and transport a certain flux of elastic sediment. These are seen in the caves as deposits of clays sands silts and occasional cobb l es and boulders These are usually derived from sandstones and other rocks that have been transported into the cave system from non-local sources It was demonstrated long ago (WHITE & WHITE (1968) that any karstic drainage basin with a component of allogenic recharge must carry a sediment load Tributary surface streams remove weathered rock material and carry it into the underground system at their swallets If the transport processes for transporting the elastic material are not sufficient to handle the load being injected into the system by sinking streams the s u bterranean drainage system would simply clog up The underground drainage would then be forced to return to surface routes However it is apparent that the transport of these sediments is not continuous. The transport is episodic related to extreme storm events that inject very large quantities of water into the karstic system The sediments move through the karst system as a series of pulses rather than as a continuous flow Furthermore there should exist a threshold in the conduit flow velocity below which the sediment will not move. Under sub-threshold conditions the sediments will simply pile up in the conduit system Only when the threshold is exceede d will a pulse of elastic material be flushed through and out of the conduit s y stem Our purpose in the present paper is to establish some criteria for sediment movement for the threshold s, and s ome analysis of the episodic nature of sediment movement in terms of the repeat frequency of severe s torm s. 2. Clastic Sediments in Karst Aquifers C la s tic sediments consist of interbedded cla y s s ilt s, s ands gravels cobbles and boulders with widel y v ary in g distributions of particle sizes and widely varyin g degrees of sorting At any particular location in a cave the elastic sediments often exhibit a distinct sequence of beds (Fig I-A) However these bed sequen c e s can s eldom be traced for any great distance along cave passages Some beds are well sorted containing a relatively narrow distribution of particle sizes ; others are chaotic with sediment particles of sizes from clay to boulders intermixed A stratigraphic section can be constructed at any particula r point along a cave passages but these sections are not very useful for the interpretation of depositional processes In order to discuss the relation of elastic sediment deposits to the stream flow processes that transport them it is useful to classify the deposits into facies These are : Slackwater facies Channel facies Stormwater facies Extreme flood facies The typical appearance of these facies for some common sedimentary sequences from the low relief caves of eastern United States are illustrated schematically in Fig I. The channel facies represents sorted or partially sorted sediments that have been moved or deposited along the cave passage Figure I-A shows a filled side passage off Rose's Pass in Mammoth Cave (DAVIES & CHAO 1959) The passage is filled with bedded sands and gravels which result from variou s flow conditions at a time when the pas s age was an active streamwa y. The s e sediments were transported mainl y as bedload At the top of the section in Fig I -A i s a layer of laminated cla y which is called a slackwat e r facies When passa g e a re plugged with sediment or blocked b y breakdown they can still be flooded with muddy water when base levels rise In many cave s, the final sediment to be deposited is a layer of laminated cla y transported a s suspended load in th e b a cked-up floodwaters In some caves active streams have cut throu g h the channel facies to form a secondary stream channel with b e d mat e rial con s i s tin g o f g rav e l cobbl e s and b o ulders ( Fi g. 1-8). Thi s coar s e g rain e d 6 1 h Confere nc e on Limestone H y dro l og y and Fi s s u red Med i a 19 1

PAGE 203

8_ 7 6_ 5 4 3_ 2 0 -~--. --# :----.......... ----...... '> I ---= -:-....__ Slackwater facies Lom,noted cloy Silt 1 /4 inch 9rovel Ton sono with block stain bond 1 14 inch 9rovel Ton sand, with block stain bond Floor Channel facies Channel facies Stormwater facies Channel and breakdown facies Figure 1. Sketches showing elastic sedimentfacies: (AJ Mammoth Cave (from DAVIES & CHAO, 1959J, (BJ Generic chan11el form, (CJ Tytoo11a Cave, Pe1111sylva11ia, USA 192 Proceed ings of the 12 th International Congress of Speleology, 1997 Switzerland Volume 2

PAGE 204

material from which all of the sand and clay has been winnowed out is called a stormwater facies. Figure 1C shows a sketch of Tytoona Cave in central Pennsylvania, an active streamway, in which the stormwater facies occupies most of the passage with only some remnant banks of the channel facies Not shown in the sketches are unsorted sediments consisting of a chaotic mixture of all particle sizes from clay to boulders The Butler Cave Sinking Creek System Virginia USA contains large inflows of these chaotic deposits (WHITE & HESS 1982). Absence of sorting can be ascribed to sediment transport by extreme floods which entrain the entire sediment pile and move it along the cave passage The Butler Cave passages which contain the extreme flood facies are on the flank of a syncline and have much steeper channel slopes than other caves examined The facies described above are arranged in order of increasing stream power required for their transport The slackwater facies are transported and deposited as suspended load Rising water levels during floods can back-flood into all parts of the cave and which allows the slackwater clays to settle without the necessity of high flow velocities and with very little required stream power The channel facies sediments are moved mainly as bedload during flood events of moderate intensity Once the flow systems that deposited them have been reduced either by loss of recharge or by diversion of stream flows into other routes the channel deposits tend to remain as part of the sedimentary record The stormwater facies are those sediments that require flood events to be transported as bedload. Normal flows and floods of moderate intensity serve to winnow clay and sand sized particles leaving only the cobble and boulder size material as a form of bed armoring Extreme flood facies are found in caves with steep gradients and which are otherwise subject to catastrophic flooding 3. Mechanisms and Thresholds for Sediment Transport There is a large literature on sediment transport written mainly by civil engineers (GRAF 1971 ; VANONI 1975) The mechanisms and thresholds for sediment transport have been determined by laboratory experiments using flumes with uniform particle size beds ( homogeneous beds ). The critical shear stress for movement of bedload provides a threshold below which the sediment does not move Many experiments establish the relationship of critical shear stress to particle size (VANONI 1975) A fitted equation is "t C 0.067 D ~0 8 where c is the critical shear stress in k g/ m 2 and D 50 is the median particle size in mm The critical boundary shear is determined by flow depth, d, and channel slope, S -r y d S where y is the specific weight of water. These parameters are also connected to the stream power Q ywdVS )' Q s where w is the channel width V is the mean flow velocity (m / sec) Q is the volume discharge (m 3 /sec), and n is the stream power in watts (Joule-sec) Transferring concepts developed for surface streams with continuous slopes to conduit systems with irregular slopes and alternating reaches of open channel flow and closed conduit flow requires replacing channel slopes with overall hydraulic gradients Some hydraulic significance can be attached to the facies concepts introduced in this paper For any sediment transfer the threshold for bed movement must be exceeded The channel facies require flow volumes and flow velocities expected from annual high flows and from moderate floods Stormwater facies require high floods to provide the necessary stream power to move the cobble and boulder beds Moderate flows of more frequent recurrence serve to extract silt and sand size particles leaving only the very coarse part of the bedload to await in infrequent high floods A second threshold must be exceeded to achieve complete entrainment of the elastic sediment deposits The entrainment threshold is exceeded only during extreme floods (50 I 00 year recurrence interval typically) and results in the extreme flood facies consisting of unsorted sediments with a wide range of particle sizes 4. Conclusions Clastic sediments in caves can be classified according to the flow regimes required to transport them Examination of the flow velocities and stream powers required to transport elastic sediments suggests that flood events of various magnitudes provide the power source necessary to transport the flux of elastic sediments through karst aquifer s. References DAVIES W E & E C.T CHAO 1959 Report on the s ediments o f Mammoth Cave K e ntucky Administrative Report, U.S. Geol. Survey to Natl. Park Service 117 pp GRAF W H 1971. Hydraulic s of Sediment Transport McGraw-Hill New York 513 pp. VANONI. V A I 975 Sedimentation Engineering Am e r. Soc C ivil En g . New York 745 pp WHITE E L. & W B WHITE 1968 Dynamic s o f s ediment tran s port in limestone caves Natl Sp e l eo l S oc Bull 3 0 : 115-1 2 9 6 1 h Confere nc e on L i mestone Hydro l og y and Fissured Med i a 193

PAGE 205

WHITE W.B. & J.W HESS 1982 Geomorphology of Burnsville Cove and the Geology of the Butler Cave Sinking Creek System. Natl. Speleo/. Soc Bull 44 : 4767 194 P roceedings of th e 12 th International Con g ress of Speleology 1997, Switzerland-Volume 2

PAGE 206

Borehole tests for megascale channeling in carbonate aquifers by Stephen R.H. Worthington and Derek C Ford Department of Geography, McMaster University, Hamilton Ontario L8S 4Kl Canada. Abstract Dissolution processes in carbonate aquifers commonly result in an interconnected network of enlarged fractures or channels which make up only a very small fraction of the rock volume. This means that averages from hydraulic testing represent the relatively slow flow through the matrix and tectonic fractures, and give little information on channeling We have analyzed hydraulic conductivity, water level and water quality data from dolomite terranes in Ontario (Canada) and limestone terranes in Kentucky (USA) with known networks of channels Seven test methods in particular have been found useful in indicating the presence of extensive channels The combination of several of the above seven methods is an excellent way of detecting whether an aquifer has an extensive network of interconnected channels with rapid solute transport 1 Introduction Two conceptual models of flow in carbonate aquifers have been developed over the last hundred years. One model considers that solution is not normally of major importance in enhancing the permeability of carbonates so that carbonate aquifers may be studied in the same way as other bedrock aquifers Thus the fracture networks are tectonic in nature, cavities encountered in drilling are "vugs" rather than laterally persistent channels, and flow may be studied by equivalent porous medium (EPM) models This view is implicitly followed in general hydrogeology texts such as FREEZE & CHERRY (1979) The second conceptual model of flow in carbonate aquifers stresses that solution is commonly of considerable importance in enhancing the permeability of carbonate rocks. The landscapes over such aquifers are often so distinctive that they are termed karst landscapes This model is followed in specialist texts on carbonate hydrogeology such as WHITE (1988) and FORD & WILLIAMS (1989) The differences between the two models are fundamental and of considerable significance to the movement of contaminants. The two models represent end members of a spectrum There has been some discussion in the literature of this spectrum (e g WHITE, 1969), but there are no practical tests using borehole data that are currently in widespread use to differentiate where on the spectrum a given field site lies Given the considerable differences between the end members of th e spectrum there is a need for explicit borehole tests Seven possible tests are outlined below 2. Terminology: channels, conduits, macrofissures and karst Confusing and contradictory terminology is a major barrier to communication of ideas on flow in carbonate aquifers. Meteoric water circulating through an unconfined carbonate aquifer will tend to produce an integrated network of solu ti onally-enlarged fractures These enlarged fractures have been called channels (CHOQUETTE & PRAY 1970), macrofissures (REEVES, 1979), or secondary fissures (PRICE et al. 1993) and larger examples are known as conduits ( > 1cm diameter) or caves ( > lm enterable by people) (WHITE 1988 ; FORD & WILLIAMS, 1989) The interconnectivity of the enlarged fractures is of prime importance and this is emphasised in the term megascale channeling (WINOGRAD & PEARSON, 1976) This latter term will be used in the following account to describe interconnected solutionally enlarged fractures, which may range in aperture from less than 1mm t o more th a n 10m and in lateral extent up to tens of kilometres (WINOGRAD & PEARSON, 1976 ; QUINLAN & EWERS, 1989) The term karst is generally understood to describe a landscape which is distinctive beca use it is underlain by soluble rocks Aquifers in the soluble rocks underlying karst topography are often called karst aquifers or karstic aquifers. To many hydrogeologists caves are an essential component of a karst aquifer Thus a karst aquifer can be thought of as an extreme example of an aquifer with megascale channeling 3. Seven test methods that may indicate interconnected channels An importan t characteristic of carbonate aquifers with networks of solutionally-enlarged fractures is that these networks can function effectively at transporting solutes rapidly yet only occupy a very small fraction of aquifer volume. For instance, for the limestone aquifer in the Mendip Hills (England) ATKINSON (1977) calculated that the interconnected conduit network accounted for only 0 03 % of the aquifer volume Consequently, boreholes are unlikely to intersect the major arteries of the network Instead, tests should be designed to sample for the existence of the network of interconnected channels One proven method and six potentially useful methods are described below i) Well-to-well or well-to-spring tracer tests Tracer testing is a well proven method of measurin~ flow direction and velocity in carbonate aquifers (FORD & WILLIAMS, 1989, pp. 219-241) The technique is more than 100 years old, and more than 10,000 successful tests have been conducted Most tests h a ve been under natural gradient conditions from sinking streams to springs with boreholes used in a smaller number of tests Fluorescent dyes are the most popular tracers, combining low cost and low toxicity with high detectability. An example of a well to spring test in Ontario is shown in Figure l. The orange dye Rhodamine WT was used with 19mg of dye being injected at a well The dye was recovered at a spring 120m away demonstrating a solute velocity of 6 2 x 10 3 m/s. This measured velocity is some tho\lSand times faster than velocities calculated using hydraulic conductivity measurements and EPM assumptions Natural gradient well to well tests are not assured of success since the main channel network may not be intersected by the well Therefore moderate pumping is recommended to draw the tracer to the monitoring well 6 th Conference on Limestone H ydrology and Fissured Media 195

PAGE 207

100 C g E Q) Q. "' t: I Q. 50 3: G> C .E I "O 0 .s::. a: 0 0 6 12 18 Hours after dye injection Figure 1. Breakthrough curve of a Rhodamine WT tracer test in dolomite from a borehole to a spring. Typical features shown are /Ju rapid rise and exponential recession, and high sensiJivity of /Ju technique. ii) Combination of core, packer, slug and pump tests In an ideal porous medium, permeability is independent of the test scale. Thus the permeability calculated from core and pumping tests should be similar. This will not be true in many fractured media, and is especially untrue where there is extensive solution In such cases, core tests and packer tests across unfractured intervals only measure matrix permeability Packer tests across fractured intervals and slug tests will give higher values as they include both the matrix and some fracture permeability Pump tests will give yet higher values as they will also sample the channel network Thus the permeability increases with the scale of a test (QUINLAN et al 1992). Some examples of permeability tests in Paleozoic carbonates are shown in Figure 2 All five sites are broadly similar in packer or slug permeability, but other hydraulic tests give considerably different results 0 01 0 001 E.. 0 0001 1E-05 > 15 ::, 1E-06 "O C 1E-07 0 0 .2 1E-08 :5 I 1E-09 -u i' 1E-10 1E 11 0 1 2 16 50 84 98 99 9 Probability % --+Ontario 1 -Ontario 2 -eOntario 3 -England -Kentucky Figure 2. u,g-nonnal hydraulic conductivity tests in Paleozoic carbonates using packers (OnJario) and slug tests (England, Kentucky). In the central Kentucky limestones core tests have given hydraulic conductivity values of 10 11 m/s (BROWN & LAMBERT, 1963). but slug tests g i ve values a million tim es higher (Figure 2) A similar range of values is found in th e Ordovician dolomites in Ontario, from consultants reports at sit e s along the Niagara Escarpment. Pack e r tests across unfractured intervals give valu e s of 1 0 11 to 10 8 m/s while pump tests and packer tests across open fractures commonly give values of 10 5 to 10 3 m/s The site in Kentucky is in the Mississippian limestones of the Turnhole Spring catchment. Turnhole Spring lies within Mammoth Cave National Park, and Mammoth Cave is the most extensive cave in the world. Most of the borehole tests shown in Figure 2 were made within 50m of a megascale channel, which has been shown by tracer testing and flow measurement to be at least 18km long and to have an aperture of at least several metres (QUlNLAN & EWERS 1989) Primary porosity of the limestone is 3 3 % (BROWN & LAMBERT, 1963) The site in England is the Mendip Hills, a Mississippian limestone aquifer with O 8 % effective porosity. but with extensive surface karst features and caves (ATKINSON, 1977) The slug tests were undertaken in a quarry to which several tracer tests were made The tests showed velocities of 0 0010 003 m/s, indicating megascale channeling (EDWARDS et al. 1992). The three sites in Ontario in Figure 2 are in low porosity ( < 5 %) Ordovician dolomites Glaciation as recently as 13 000 years ago has removed or infilled most surficial karst features However, megascale channelling in the dolomite aquifer has been demonstrated by tracer tests (Figure 1) ili1 Variable rate pumping test HICKEY (1984) showed that the pumping rate should be proportional to the drawdown in observation wells if Darcy's Law is valid within the cone of depression In the carbonate aquifer in Florida this was found to be true for one series of tests. It has yet to be shown whether how sensitive this technique is for detecting departures from Darcy's Law in the vicinity of channel networks with turbulent flow iv) Matrix and fracture packer test to calculate fracture extent PRICE (1994) described a method for estimatin g the extent of interconnected fractures intersected by wells by using steady state packer testing. The contrast between the permeability of the matrix and of a fracture isolated by packers can be used t o estimate fissur e e xtent. The method uses some simplifying assumptions (e g homogeneous, isotropic matrix, parallel plate constant aperture fissures) but is useful for differentiating whether an open fracture intersected by drilling is an isolated vug or is connect e d to an extensive interconnected channel network. v) Symmetry of cones of depression at pumping wells The cone of depression at a pumping well is symmetrical in a homogeneous porous medium However, the cone of depression is likely to be irregular if there is extensive channeling nearby Figure 3 shows an example from a pumping test in Ordovician dolomite in Ontario Most of the observation well water levels do not closely correspond to the idealised con e of depression and between two of these wells (87-4B and Wl B) the hydraulic gradient is away from the pumping well However, the most interesting feature is the cascading of water into the well from a single fracture Most of the pumped water 196 P roceedings o f the 12 '" I nternational Congress of S peleology, 1997 Switz e rland Volume 2

PAGE 208

came from this single fracture indicating an extensive interconnected channel network -1 a, E2 QI -2 E C -3 0 -0 -4 a! Gi > -5 .!! ., I Water cascading .; -e :: from fracture PW-8 -7 0 10 20 30 40 50 60 Horizontal distance (metres) Figure 3. Water level profiles during a pumping test in dolomite in Ontario. Tiu curved lines represent the best-fit cone of depression from all nine observation wells. vl) Continuous water level monitoring Interconnected channel networks transmit water quickly, so a prompt water level response in boreholes can be expected following rainfall Such a response will not be detected by conventional weekly or monthly water level measurement programmes. Continuous water level monitoring is technically straightforward, and the response magnitude and lag following precipitation is valuable for indicating channel connectivity In Central Kentucky water levels in well often rise by several metres within hours of major rainfall events (QUINLAN & EWERS, 1989) vu) Frequent water quality monitoring Precipitation which rapidly infiltrates along channel networks commonly has a much lower solute concentration than long residence matrix water. Thus variation in solute concentration at a well should be an indicator of connectivi t y t o major channel networks Frequ e nt sampling is necessary to detect the rapid response following rainfall QUINLAN et al (1982) have shown how a sampling interval of hours rather than days may be necessary for an adequate understanding of solute changes following rainfall Pigure 4 shows trichlorobenzene concentrations in a bedrock well at a contaminated site on Ontario dolomites Daily measurements h e re reveal increasing concentrations during a runoff recession (November 1 24) and a hundredfold decrease following precipitation (November 27 December 4th) Toe lag here between precipitation and dilution at the well is about three days Traditional infrequent sampling such as quarterly would completely miss such short term variability in solut e concentrations 4. Discussion Tracer testing is the only on e of the seven techniqu e s that is likely to give simple, unequivocal evidence that megascale channeling is present This is because the groundwater velocities measured in carbonates are usually in the range 0 002 0.2 m/s Such velocities are orders of magnitude more rapid than .;;0 .D E >(I) ., :0 a. E: G> C G> N C G> .D 0 0 :c ~ .= 60-r---------------------,.20 40 f 20 '"r/ \,. ..... . : . l v 1 y \ o+==_..== ., -..;;, ===r--===--,,....1. __ ..,....:..i......~ .. ----+ o 01-Nov 08-Nov 15-Nov 22-Nov 29-Nov 06-Dec Date in 1987 Figure 4. Rapid changes in trichlorobenzeneconcenJrations in a dolomite tuJuifer in Ontario, showing lagged inverse correlation to runoff. predictions based on Darcy s Law and can only be explained by megascale channeling The remaining six techniques make use of, or are extensions of commonly collected data from boreholes Each of them can be used to investigate the possibility of megascale channeling ; together the evidence from several of these techniques may give strong evidence for megascale channeling Major karst hydrological features such as sinking streams caves containing streams, and large springs definitively show that megascale channeling is present However, there is no simple test that will show that megascale channeling is not present Some authors have suggested that limestones and dolomites have higher permeability and porosity values if they are karstic (e g (FREEZE & CHERRY, 1979, pp 26, 37) However, the permeability enhancement is likely to be minor especially if tested by low aquifer volume core, packer or slug tests. Furthermore, porosity enhancement due to megascale channeling is negligible; even for the well-karstified aquifer in the Mendip Hills, England, where e xtensive caves have been found th e porosity enhancemen t by channeling is only 0 03 % (ATKINSON 1977) (I) -E o 01 i r-::::=======================::::::1 / 0 001: : Moderately compacted Limestone and Dolomite f 0 0001 1 / "------.------------./ ( Cenozolcand Mesozoic) -0 C 0 0 1E-06 Highly compacte< Limestone and Dolomite (Paleozo i c and Mesozoic) 1E-0 71 '-~----1 E-08+---,.----.---.--~-~--,.----.---.--~--t 0 10 20 30 40 50 Porosity % Figure 5. General range of hydraulic conductivity and porosity in unconfined carbonate aJJuifers, as afunction of compaction and age. 6' h Co n ference on Li mestone H yd r ology and Fissured M edia 197

PAGE 209

Instead, the substantial range in permeability and porosity in carbonate aquifers is explained primarily by compaction (Figure 5) Most Paleozoic carbonates have been subjected to deep burial, resulting in low porosities ( often < 2 % ) and low permeabilities Conversely, most Cenozoic carbonates have been compacted less, and thus have both higher porosity and higher permeability (Figure 5). As well as compaction, the solubility of calcite and dolomite also plays a substantial role in porosity and permeability development at all stages in the life of a carbonate deposit. Prom deposition through burial, uplift and erosion carbonate rocks contrast strongly with siliciclastic rocks (CHOQUETTE & PRAY, 1970) Toe seven test methods outlined in this paper can help show that solute transport in carbonate rocks, and especially unconfined carbonates, is profoundly influenced by solution, which inexorably leads to megascale channeling References ATKINSON, T C 1977. Diffuse flow and conduit flow in limestone terrain in the Mendip Hills, Somerset (Great Britain). J Hydrol 35, 93-110. BROWN, R P & T.W. LAMBERT, 1963. Reconnaissance of ground water resources in the Mississippian Plateau region, Kentucky, U S Geol. Survey Water-Supply Paper 1837, 64p. CHOQUETTE, P.W., & L.C PRAY, 1970. Geological nomenclature and classification of porosity in sedimentary carbonates. Am Assoc. Petrol Geologists Bull 54, 207-250 EDWARDS, A.J S.L. HOBBS & P.L. SMART, 1992 Effects of quarry dewatering on a karstified limestone aquifer; a case study from the Mendip Hills, England Proc Third Conference on Hydrogeology, Ecology, Monitoring, and Management of Ground Water in Karst Terranes, National Ground Water Association, Dublin, Ohio FORD, D C & P. WILLIAMS, 1989. Karst geomorphology and hydrology Unwin Hyman, London, 60lp. FREEZE, R.A. & J.A CHERRY, 1979. Groundwater. Prentice-Hall, Englewood Cliffs, NJ, 604p. HICKEY, J.J., 1984. Field testing the hypothesis of Darcian flow through a carbonate aquifer. Ground Water, 22, 544-547. PRICE, M.; R.A. DOWNlNG & W M. EDMUNDS, 1993 Toe Chalk as an aquifer. In: Toe Hydrogeology of the Chalk of North-West Europe (Eds R.A. Downing, M. Price and G P. Jones), Clarendon, Oxford, 35-58. PRICE, M 1994 A method for assessing the extent of fissuring in double-porosity aquifers, using data from packer tests. Internat. Assoc. Hydro!. Sci., Pub!. No 222, 271-278. QUINLAN, J.F G J DAVIES & S R.H WORTHINGTON, 1992 Review of ground water quality monitoring network design. Journal of Hydraulic Engineering, 119, 1436-1442 QUINLAN. J.F & R.O. EWERS, 1989 Subsurface drainage in the Mammoth Cave area. In : Karst hydrology : concepts from the Mammoth Cave area, (Eds W B. White and E.L. White), Van Nostrand Reinhold, New York, p65-103 REEVES, M .J 1979 Recharge and pollution of the English Chalk: some possible mechanisms Eng. Geol 14, 231-240 WHITE, W B 1969. Conceptual models for carbonate aquifers Ground Water, 7, 15-21. WHITE, W B 1988. Geomorphology and hydrology of karst terrains Oxford Univ Press, 464p WINOGRAD, I.J & P J PEARSON, 1976 Major Carbon 14 anomaly in a regional carbonate aquifer: possible evidence for megascale channeling South Central Great Basin. Water Resources Research, 12, 1125-1143. 198 Proceed in gs of the 12 th Internat i onal Congress of Speleology, 1 997 Switzerland Vo l ume 2

PAGE 210

Estimation of aquifer parameters in highly permeable limestone (Biscayne Aquifer, Florida) Eric Zechner, David Genereux, Jose D.A. Guardiario, James E. Saiers Geology Department, Florida International University, Miami, FL 33199, USA Hydrogeological Setting The surficial Biscayne Aquifer (BA) extends from the Atlantic Coast of SE-Florida about 50-80 km westward into the Everglades. It is the source of most water supplies in this densely populated region and is considered one of the most permeable in the world. The extensive use of the BA with canal systems and large-scale pumping as a result of urban development leads to problems such as coastal saltwater intrusion, groundwater pollution, and drainage of the sensitive ecosystem of the Everglades marshes. In the studied field site at the eastern border of Everglades National Park, 50 km SW of Miami, the BA reaches a thickness of nearly 15 m. It consists of coralline, bryozoan, shelly, and in some parts sandy limestone of Pleistocene age. Groundwater flow is enhanced due to a strongly developed secondary porosity with frequent solution cavities. In zones of abundant solution holes hydraulic conductivity exceeds 10 1 m/s. The two formations making up the BA, the Miami Limestone and the underlying Ft. Thompson Formation are separated with a 0.3 m thick condensed micritic and less porous layer. Parameter Estimation Groundwater models have become a widely used and important tool to estimate the relevant hydraulic parameters. The accuracy of models are strongly related to the density of available data which are used in the parameter estimation procedure. The most common approach is still trial-and-error calibration on groundwater head observations. Inverse models may result in a more objective estimation of the relevant hydrogeologic parameters. However, sparse distribution of same type of data often result in insensitive parameters and several models with a different pattern of hydrogeologic parameters may reproduce the dataset equally well. This insensitivity can be reduced using additional field data like (1) direct measurements of the relevant hydraulic parameters with for example aquifer tests, (2) flow at stream boundaries, (3) geological and/or geophysical data, (4) and tracer concentrations In the BA groundwater flow is strongly controlled through the manipulation of the canal levels. The relevant hydraulic parameters to be estimated to calculate the aquifer-canal water flow exchange are aquifer hydraulic conductivity, specific yield, and the canal bed conductance Field measurements of hydraulic parameters have to be adapted to the highly permeable limestone. Pump tests for example would face considerable logistical problems with the use of very large pumps in an area of difficult access (Everglades) and the remote disposal of the water to avoid violating the assumptions required for the analysis. The alternative used was to stress the aquifer with a canal drawdown test. On a 2.4 km long S section of L3 l W, a about 17 m wide canal bordering the Everglades National Park, the canal level was lowered within hours of 0.4 mover a period of 10 days. The response of the groundwater head was monitored continuously in 3 E-W sections of totally 20 piezometers. In addition the canal flow was monitored on both ends of the section with acoustic velocity meters. An analytical model assuming I-dimensional flow estimated an average hydraulic conductivity for the BA of 9-10 2 m/s and a canal bed conductance of 5 10 4 m/s. A numerical 3D forward flow simulation of the drawdown test with these conductivities showed a fairly good agreement of the observed and simulated head (figure). With a currently developed inverse groundwater model the parameter estimation is gradually improved to quantify the canal seepage with data of canal flow measurements and tracer concentrations of chloride in groundand canalwater. 1.40 ~-----------------------_-_-_-_-_-_-_-___ -,--, 1 30 0 1 20 z i 1 10 Q) 1.00 0 90 20 21 22 23 Groundwater Head Piezometer P3s (ENP) 24 25 26 day January 1996 27 28 -+measured A simulated 29 30 31 Result of a numerical flow model based only on hydraulic field parameters (no calibration): Measured vs. simulated head of a piezometer in Everglades National Park (ENP), 50m west of canal L3HV, during a canal drawdown test in January 1996. 6 th Conference on Limestone Hydrology and Fissured Med ia 199

PAGE 212

Vulnerability of karstic aquifers in Abruzzi (Italy). Two different case-studies: Gran Sasso mountain and Fucino plain Ezio Burri, Raniero Massoli-Novelli & Marco Petitta Dipartimento di Scienze Ambientali, Universita degli Studi dell' Aquila, loc. Coppito, 67100 L' Aquila, Italia Abstract The Abruzzi Apennine's hydrogeology is characterized by karstic mountain aquifers and main springs located at the massif basal contact. In th e Gran Sasso mountain, infiltration is very high and concentrated in the big glacial-karstic ridges, and is determined by tectonic patterns, climate and numerous karst features In the 1970's, two highw ay tunnels intercepted the Gran Sasso regional aquifer and reduced the discharge of surrounding springs. Th e aquifer vulnerability is very high in infiltration areas (Campo Imperatore plain). The alluvial -lacustrine sediments that fill the Fucino plain contain an important aquifer, gained from the neighbouring carbonate mountains. Increased agricultural activity causes high pollution risk increasing deep water exploitation and impoverishment of groundwater resources. This si tuation has obviously interfered with the hydrogeological setting of the karstic aquifers, determining changes in the dynamic equilibrium of groundwaters Riassunto L'idro geologia dell'Appennino abruzzese e caratterizzata da estesi acquiferi carbonatici, che alimentano sorgenti basali di portata elevata e regime abbastanza stabile. Sul Gran Sasso, le sorgenti sono paste a notevole distanza dalle aree di infiltrazione, particolarmente concentrata in depressioni tettono-carsiche; ii regime idrodinamico e influenzato da un traforo autostradale, che drena la falda regionale causando ripercussioni sulle sorgenti piu vicine. La vuloerabili ta risulta particolarmente elevata nelle zone ad infiltrazione concentrata, come Campo Imperatore. Nella piana de! Fucino, i sed imenti quatemari che la colmano sono sede di una imponante falda, alimentata direttamente dai rilievi carbonatici circostanti. II maggiore grado di vulnerabilita si riscontra nella Pi a na dove l'intensa attivita agricola comporta immissione di inquinanti e depauperamento della falda a causa dei prelievi idrici. L'alterazione dell'equilibrio idrodinamico determina una maggiore vulnerabilita dei vicini acquiferi carsici. 1. Introduction In the Central Apennine the structu ral -geologica l setting was principally conditioned by the Alpine orogeny which formed the carbonate ridges, and at the same time a wide deposition of thick Pliocene and Quaternary sediments in lower areas. As a result a particular hydrogeological selling has been obtained in whic h karstic carbo nate ridges constitute the most important aquifers and the recent deposits locat ed at the base of the massifs represent their limits of permeability The carbonate aquifers are characterized by high permeability due to fractures and to karst and, despite the lithological and structural setting, each ridge can be considered as a single aquife r The lower permeability of the Plio quaternary deposits, composed mainly of sands. silts and clays, determines the main base level of the mountain aquifers. Groundwaters come to the surface through the main springs located at the base of the massifs Such springs. with high discharge (> I m'/s) and a rather steady regime. are limited in number and withou t the typical features of the mature karstic systems. Moreover the aquifers. inside the carbona t e ridges often interact with the aquifers of the Quaternary deposits, through l ateral losses that feed the porous aquifer in the bottom of the valley. In the car bonate Apennine the infiltration is strongly conditioned by karst on the ridges where numerous and mature karstic phenomena are present: tectonic-karstic depressions, dolina, endoreic basins and so metime s insurgence s. Obviously in these areas the infiltration of the rainfall is very high and the runoff can be considered almost zero. Consequently carbonate aquifers are greatly fed by a karstic system. In the mountain areas of the Apennines, particularly in Abruzzi, a double problem, co nne cted to the vulnerability of the aquifers occurs: the pollution risk of the karstic aquifers with high vulnerability in the areas having concentra ted infiltration; any possible interferences in the rech arge areas or in the unsaturated zone, can have consequences on all the springs of the ridge, because there is no direct karstic co nnection between the infiltration and the gro undwater s emergences; the interaction between carbonate aquifers and porous aquifers greatly exploited as drinking water for irrigation and for industrial supply. The piezometric variations inside the plains both positive (irrigation) and negative (pumping) determine continuous c h anges of the hydrodynamic equilibrium, thus affecting the carbonate aquifer. In this paper we illustrate two different case-studies: the Gran Sasso mountain one of the greatest regional aquifers and a recently founded National Park, and the Fucino plain, a wide basin with high hum an interference a nd with intensive agricultural activity. 2. The Gran Sasso mountain The Gran Sas so mountain. in Abruzzi. Central It aly, covers an area of about 800 km 1 and has the highest peak at 2.912 m a.s.l. (fig. I ). Geological features are a Mesozoic-Tertiary carbonate sequence in marginal or s l ope facies with a high overthrust toward the north The tectonic selling is the result of an initial stage, characterized by mainly co mpressive deformation and a subsequent stage with strong uplift and extensio n al fault activiiy (ADAMOLI et al., 1982) 6 "' Conference on Limestone Hydrology and Fissured Media 201

PAGE 213

The carbonate aquifer is formed by a series of inter communicating reservoirs, confined laterally by impermeable outcrops Infiltration is very high (> 800 mm/year), due to the tectonic patterns climate and numerous karstic features (polje, karstic basins, etc.), compared with the average rainfall on the massif (approx. I 200 mm/year). t N I 10 IS 20 25 km Fig ur e 1 : A r ea of the G r a11 Sasso mou11 t ai11 The m ain s prin gs (ci r cles), h ighway (do u b l e li11e) a11 d thei r d ra i na ge (s tar s) ar e Th e n um be r s of s pr ings r e f e r t o T ab l e 1. (g r ey) tu 1111 e l s shown The main springs are located at the massif basal contact with impermeable deposits There are 12 groups of such springs (tab I), with a mean discharge of about 27 m 3 /s : 5 groups are located on the northern side at an elevation of > l000 m with a total discharge of about 2 m 3 /s and substantial seasonal variat i ons ; 7 groups, which have a total discharge of about 25 m 3 /s, a r e on the southern side at an elevation of< 650 m In the I 970' s, two highway tunnels, I 0 15 km long, were bored in the mountain at an elevation of 900 m The highway connects Rome with L' Aquila and Teramo, from the Thyrrenian sea to the Adriatic coast. The highway tunnels have obviously interfered with the hydrogeological setting of the Gran Sasso mountain Hydrogeologic a l studies were carried out in particular by A. MONJOr E (1975). The tunnels i n tercepted the Gran Sasso regional aquifer initially on the southern s ide with a mean discharge of 0 5 m 3 /s Subsequently, the aq u ifer was also intercepted on the northern side, after about 3 km of boring ; in this case, the water seepage qu i ckly exceeded 2 m 3 /s When boring ended, between I 980 and I 982 drainage by the tunnels was 1.3 m 3 /s on the northern side and 0 5 m / s on the southern side practi ca lly the s ame as the curr e nt values ( ADAM0LI, 1990). The immediate con s equence of this underground drainage wa s a marked r e duction in the disch a rge of the springs on the northern side ( fig I tab I ) that were a l re a dy being captured almost entirely for drinking water. Because of this immediate step s were taken to channel the water from the tunnels to the aqueduct network Hydrogeological interpretation s regarding the variations brought about by the tunn e l s (MO N JOIE, 1975 ; CELI C O 1983; BONI et al. 1986 ; ADAMOLI, 1990 ; MASS0LI-NOVELLI & PETITTA in press) indicate that the Gran Sa s so aquifer has reacted to the i mp ac t of tunnel d r ain a g e. in the fir s t pla c e by rapidly changi n g it s hydrodynamic s etting (unsteady n o w ) and di s charging large quantitie s o f groundwater. Subsequently, the a quifer ad a pted to the pre se n c e of tw o new springs consist i n g of the drainage from th e tunnel s. until it achieved a new hydrodynamic balance after a few years (steady flow) which corres p onds to the p resent si tu a ti on. New monthly monitoring of spring discharge began in 1994 limited to the Vera springs near L'Aquila (PETITTA & MASS0LI-N0VELLI 1995) Si n ce 1996, this monitoring has been completed with surveys rega r ding the other main springs on the so u thern side (Capo Pescara S Calisto Tirino springs) (MASSOLI-NOVELLI & PETITTA in press) Exa m ination of the initial results of the new moni t oring of Vera springs produces two ma i n considerations : conditions are extremely stable, as proven by the almost constant discharge measured d uring the year; t his sta b i li ty was not to be found in pre-tunnel val u es ; the mean discharge (abo u t I m 3 /s) is considerab l y Jess than the discharge meas u red before and during tunnel construction (1.8 m 3 /s). The main charac t eristic of the Tirino R iver is that it is fed almost exclusively by groundwater (tab. I, springs n-9-I O) It receives an average of more t h an 12 m 3 /s of water (as much as 18 m 3 /s in the past) supplied by numero u s large groups of springs Insufficient data are availab l e at the present time to express meaningfu l assessments regarding the Ti ri no springs and the situation there can only be clapfied by continuing the monitoring as planned It has been asce r tained however, that discharge from the Capodacqua Tirino spring has decreased substantially over the past I 0I 5 years to the extent t h at it has been almost halved ; t his lack of water cannot on l y be justified by the t u nnel drainage, but may also depend on the ~uced rainfall. The on-going monitoring of springs on t he southern side of the Gran Sa s so provide s indispensable data to assess fully the new hydrogeological situation Only for the Vera springs, for which t h ree years of continuous observations are avai l able is it possible at the present time to assess the new hydrogeological situation determined by tunnel drainage at a high elevation The reduced discharge, which has certainly taken p l ace in basal sp r ings both on the northern and on the southern side (especially Vera) is currently accentuated by the reduced rainfall ascertained in recent years and, therefore, by the reduced recharge of the aquifer. It is not yet poss i b l e to establish the extent to which each of the two causes contributes to the current reduced discharge of the main springs e l ev atio mean n o spring nam e ll dis c harge (m a s.l .) (m 3 /s) I Vomano 1300 0,1 2 C h iarino 1320 0,5 3 Rio Amo 1520 0,1 4 R uzzo >1000 0 ,6 5 Vitella d oro / Mortaio 680 0,7 d Angri 6 Veto i o Boschetto 630 0 7 7 Vera 650 1, 8 8 Capodacqua Tirino 3 40 5 0 9 Presciano II ago 330 2,0 I 0 Basso Tirino 270 6,0 11 S C a listo 290 2 1 I~ Caoo Pesc ara 250 7,5 13 Tu n nels drainage (s outh) 960 0, 5 14 Tunnels drain a ge ( north) 8 90 1 ,2 Tabl e I : G r a 11 Sasso: mai n springs a n d thei r discharge in p r e tunnel period T11n11el discha r g e(*) refe r s to current situatio11 202 Proceedings of the 12 111 International Congress of Speleology 1997 Switzerland Volume 2

PAGE 214

The drainage from the Gran Sasso highway tunnels (tunnels that were undoubtedly necessary for the region's road system) has definitely determined a new hydrodynamic balance in the karstic carbonate aquifer. However, these changes cannot be considered solely negatve, i.e. the decrease in basal spring discharge, because it has to be born in mind that at the present time all the water drained from the tunnels (about 1.7 m 3 /s) at an elevation of around 1000 m a s l. is available for supply as drinking water, and this elevation permits better distribution of the water resources. The reduced discharge observed at the springs, to which a probably transient cycle of low rainfall has definitely contributed, is nevertheless not such as to compromise the ecosystems downstream. The vulnerability of the Gran Sasso aquifer, because of fracture network and karst features, can be considered very high, particularly in the sectors having concentrated infiltration, of which the most important is Campo Imperatore area. Because there is no direct connection between the basal springs and mature karstic zone, it is possible to suppose a lower vulnerability of these resources than the drained tunnel waters (900 m a s.l.), probably fed by Campo Imperatore depression ( 1600 m a.s.1.). The water drained from the highway tunnels is being captured almost entirely for drinking water and for this reason it is important continuously to protect the environment of the recharge areas. In fact, if the drainage level is higher than the base level of the aquifer, water autopurification can decrease, the water salinity is lower and residence time in the aquifer is shorter. The chemical and physical data collected in these last years are confirming this hypothesis : the drained waters from the tunnels have salinity lower than 0 1 g/1, while the value of the basal springs ranges from 0.15-0 4 g/1. 0 ~o 1cm --.. t====i, Artesian well zone Production well zone Dry well zone Spring .A Linear spring At present we haven't detailed data to analyse the inflow outflow diagrams that allow to compute the lag times of the Gran Sasso aquifer in the new hydrodynamic equilibrium 3. The Fucino plain The Fucino basin covers an area of about 900 km 2 on the carbonate mountains of the Apennine chain From the morphological point of view the basin is characterized by a wide alluvial plain (>200 km 1 ) where in the recent past the largest Apenninic lake has been localized The oval shaped plain has its greater axis, trending NW-SE, about 19 km long. Heterogeneity Quaternary alluvial-detrital-lacustrine sediments fill the Fucino plain and are over 1000 meters thick The Fucino plain and its wide lake didn't originally have a natural outflow. In the past several reclamations have been carried out to stabilize the level of the lake and then to drain it. The first tunnel was realized by the roman emperor Claudio, in the 1 century A.C .. Drainage canals and tunnels were dug to channel the waters of the Fucino Lake towards the nearby basin of the Liri river. In order to completely drain the lake, Prince A Torlonia built a second tunnel, which was completed in 1874 and at present it is utilized only in case of need. The drainage work ended in 1942 when one more tunnel, which is still working, was excavated (PETITIA & CAPELLI, 1994) In the Fucino plain all the springs are fed by the surrounding carbonate ridge (fig.2) A main spring group is located in the south-eastern side of the depression with a total discharge of I 2 m 3 /s Others natural springs with a total discharge of 0 5 m 3 /s are in the northern sector. The carbonate ridges also feed an aquifer made up of alluvial gravel-sand deposits (CEUCO & CEUC0 1994) By means of drainage canals these groundwaters have been partly channelled in river bed (fig.2), giving important linear springs ( > 1 m 3 /s). Pescina Fig. 2 Fucino plain: the alluvial-lacustrille deposits (white area) contain a groundwater aquifer exploited for human activity, 111ai11/y for agriculture Modified from CELICO (1983) 6 th Conference on Limestone Hydrology and Assured Med i a 203

PAGE 215

Humaninterference and drainage have obviously affected the hydrogeological and hydrological setting of the Fucino plain: agricultural activity is modifying the relationship between the carbonate aquifer of the ridge and the alluvial-detrital aquifer of the plain. In 1874-1875 the Fucino Lake was completely drained, and by means of an underground tunnel I billion m 3 of water was run into the Liri basin. As a result of this drainage a single vaste estate of 14.006 ha with 210 km of inner roads, 100 km of main canals and 648 km of drainage canals, rose. At present a very specialized agricultural activity is being carried on in this estate (BURRJ, 1990) with some growing environmental damages as follows: there is a widespread use of pesticides which exceeds 3 5 kg/ha, well above the italian national average of 12.8 kg/ha, and above and the general consumption of the Abruzzi Region which is 9.9 kg/ha; the system of sprinkler irrigation exceeds the water availability of the canals during the dry season. In order to solve this problem since the fifties more than 200 wells have been drilled, with a seasonal delivery of about 2.000.000 m 3 of water. As a consequence, with a long period of dry weather in the area, there is a remarkable groundwater lowering. In a emergency, 0.1 m 3 /s are taken from the neighbouring Liri river by means of the tunnel built a century ago In the north-western sector of the Fucino plain (fig.2) almost 50% of the wells have proved unproductive (CELICO, 1983); on the contrary, along the south-eastern side, where the main springs are located, the wells are very rich in water. ln the plain the groundwater is just a few meters deep and sometimes it is artesian. However three main features condition the hydrogeological and environmental setting of the Fucino area: the spring and groundwater exploitation for different purposes from the aquifer of the plain; the variability of the water demand during the year, due to irrigation and the incre ased drinking water supply, which reach the highest peak in summer; the existence of sprinkler irrigation system and the use of pesticides in agriculture. 4. Conclusions The Abruzzi aquifers, located in the carbonate ridges and the wide plains surrounded by mountains, are an important resource for the needs of the local population, but also a precious environmental heritage, which assures the survival of ecosystems both in river and humid areas. The vulnerability of these resources, as a consequence of the hydrogeological setting and of human interference is generally very high, and the ways of safeguarding the carbonate aquifers and the alluvial aquifers are quite different. As for as the Gran Sasso case-study is concerned the monitoring under way aims to estimate the discharge of the water drained from the springs and from the highway tunnels. At present we don't have sufficient data to make a meaningful assessment, of the connection between the highway tunnels drainage and the areas with concentrated infiltration, such as the Campo Imperatore area, characterized by high vulnerability. For this reason it is important lO protect continuously the recharge areas of the aquifers. It is also important to note that the waters drained from the highway tunnels are utilized as drinking water. In the case-study of the Fucino plain the alluvial aquifer is highly vulnerable owing to extensive agricultural activity which causes high pollution risks for this shallow carbonate aquifer. The possible consequences can be summarized as follows: reduction of the functionality of the operating wells because of the lowering of the piezometric levels. Thi s kind of problem is already evident and there are reasons to believe that, in some particular situation, it may cause economic and environmental damage, especially if the summer shortage of water gives a precarious hygienic situation in the canal network of the plain Moreover the pumping and the exploitation of the water are causing the subsidence of the soil, with gas emission. reduction of discharge has certainly taken place in normal and linear springs. This phenomenon involves both the lowering of the available supplies of drinking water and the regular downflow in the canals. increasing of the vulnerability of the surrounding carbonate aquifers; thi s problem, not yet sufficiently well studied, could determine infiltration of irrigation waters of poor quality into the karstic aquifers These phenomena can happen if the saturation level of the carbonate aquifers is l ower than the level of plain aquifer, even if this should happen only in some short period of the year. References ADAMOLI, L. 1990. ldro geologia de! massiccio carbonatico del Gran Sasso e conseguenze idrogeologiche degli scavi autostradali Ceo!. Teen. 3/90: 4-15. ADAMOLI, L.; BERTINI, T .; DEIANA, G .; PlERUCCINI, U. & A ROMANO 1 982. Ricerche geologiche sul Gran Sasso d'ltalia (Abruzzo). VI : primi risultati dello studio strutturale della catena del Gran Sasso d'ltalia Srudi Ceo/. Camerri. 7: 97-103. A.N A.S., 1980. II traforo autostradale. Unpublished report 463 p. BONI, C .; BONO, P & G CAPELLI 1986. Schema Idrogeologico dell'Italia Centrale. Mem Soc. C eo!. fr. 36: 9911012 B URRI, E., 1990 Storia di un lago. II Fucino in Abruzzo Terra, 3/10 : 42-52. CELICO, P. 1983 ldrogeologia dei massicci carbonatici, delle piane quaternarie e delle aree vulcaniche dell'ltalia centro meridion a le : Progetti speciali per gli schemi idrici nel M ezzogiorno. Quad Cassa Me:::.:.og 4/2, 225 p. CELICO, F & P. CELICO 1994 Aspetti idrogeologici del bacino del Fucino In (E. Burri, ed .): II lago Fucino e ii suo emissario. Carsa ed., Pescara: 35-45 MONJOIE A 1975 Hydrogeologie du massif du Gran Sasso (Apennin central). Coll. Pubbl Fae. Scien ze Appl. Univ. 53. Liegi. MASSOLI-NOVELLI, R. & M. PETrlTA in press. Hydrogeological impact of the Gran Sasso tunnels (Abruzzi, It aly). Int ernational Symposium on Engineering Geology and the Environment (Athens, Greece. 23-27 June. 1997). PETmA, M & G. CAPELLI 1994 lnqu adramento idrologico del bacino de! Fucino In (E. Burri. ed.): II lago Fucino e ii suo emissario. Carsa ed., Pescara: 46-61. PETmA, M & R. MASSOLJ-NOVELLI 1995. Primi risultati del monitoraggio geoambientale delle sorgenti del Fiume Vera (Gran Sasso) Quad. Ceo/. Appl. uppl.1/95: 93-101. Thanks to Co11sorzio Ri cerca Gran Sasso" for their co11trib11tio11 to this research. 204 Proceedings of the 12 1t1 International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 216

Karst hydrogeology and vulnerability in a coastal sector of Nebrodi Mts. region (Northern Sicily)(1) Antonio Cimino*, Rosario Abbate*, Massimo Macaluso*, Santo Orecchio ** (*) lstituto di Geofisica Mineraria, University of Palermo, Via Mariano Stabile 110, l-9013 9 Palenno Italy (**) Dipartimento di Chimica Inorganica, University of Palenno Via Archirafi 26. I-90139 Palenno Italy Abstract This paper describes the hydrogeological features of the Acquedolci karst area, in the Western sector of the Nehrodi Park (Sicily), where peculiar geostructural unit-, occur. This region is characterized by the presence of numerous springs, due to the very different penneability unit'> On the other hand, coastal aquifers represented by sandy-arenaceous fonnations are supplied by karst limestones; they are undergone to increasing sea-pollution phenomena The relationships among the hydrogeological units and the anthropic intervention have affected the actual productivity and quality of water resources principally in the coastal plain Geochemistry and geophysics have surveyed the main h y drogeological feature s of the investigated area, most of which are characterized by karst circulation : chemical analyses of groundwaters and apparent resistivity prospections have evidenced the main aspect'> affecting the water resources of the region The results of water geochemistry are presented in form of Langelier Ludwig and/or Piper diagrams, that show different groups of waters, also depending on the hypogeal limestone circulation. Apparent resistivity maps, perfonned in the coastal sectors, have been traced in order to follow the pollution occurrence and the pattern of the buried fonnations ; the most important of these ones are impervious clayey units and permeahle arenaceous units The final purpose of this research will be the elaboration of the first vulnerability map of the Acquedolci plain 1 Introduction This paper deals with the hydrogeology of an area belonging to the north western sector of the Nebrodi Mow1tains. It is executed in the ambit of the research program relevant to the study of water resources of Eastern Sicily The Nebrodi range is mainly characterized by scarcely permeable terrains (figure I) : as a matter of fact groundwaters are generally deficient or very sparse So, the water productivity results insufficient to the needs of the population. As a consequence, U1e study of the karst area of Acquedolci is very interesting because it might represent an app reciable water reservoir for this particular sector of Sicily Goal of this paper is to describe by means of selected methodologies the geomorphologic and hydrogeological characteristics of this area The acquired data have allowed us to know more deeply tile hydrogeological structures of the examined area and hence propose ilieir bener management and protection 2. Geomorphological outlines The considered territory, parallel to ilie coast line (figure 2) is delimited at North by Tyrrhenian Sea and at South by a group of carbonate relieves iliat sharply rise from the plain At West the Furiano torrent end lo tile sea in a wide coastal fan. This torrent presents, as the other ones in ilie territory regular courses, with narrow valleys deeply engraved on mount. Along ilie medium and lo wer courses it presents a considerably wide bed extending also more Ulan 300 m at ilie mouU1 The coastal plain and a great part of the inland territory are constituted by seven orders of marine terra ces (HUGONJE 1974 ) : they are diffi.1sed from this coasta l plain up to 700 m level. Sanfratello Mt. presents in its Noriliern sector a well squared geometrical trend wiili sub-vertical walls originated by two main fau l t sys ten1 s NNW-SSE and NNE-SSW directed Epigeal karst is characterized by scarcely developed morphologies uncompleted and rough!~ incised Hypogeal karst is represented by a group of cavities located along tectonic lines testifying old aquifers TI1ev evidence traces of Palaeolitic fauna, with rests of vertebrate skeletons too (San Teodoro Men in ilie omon~m grotto, see BACCI, 1989). Jn the considered area fonnations include portions of a pre-Mesozoic metamorphic basement togeilier wiili the relevant sedimentary overhurdens of ilie f,ongi-Taom,ina Stratigraphic Structural Unit (S.S U ) This S S U is the lowest unit of the overilirust belt of the Calabrian-Peloritani Arc (NIGRO 1992). The Oligocene-Miocene Numidian Flysch outcrops in the Western part of the region: it is constituted by a sequence of quartz.arenites with calcareous cement and argillites The Monte Soro Flysch, widely extending in the Southern sector is mainly represented by sequences of fine grain feldspatic quartzarenites and shales. The Longi Taormina Uni t includes a phyllite basement and an overburden consti tuted by conglomerates and quartzites, fine grain lime stones ,,..iili siliceo us element~ dolomites and marls Unconfonnable terrains of Capo d'Orlando Flysch occur in the region, formed by graded arenite banks with rare cla yey interbeddings and thick conglomerates These different formations have very complicated la yering relations due to the compressive tectonic interesting ilie Nebrodi belt and ilie successive Pliocene-Quaternary distensive tectonics that have divided the relieves in blocks and steps (DUEE, 1969 BONFIGLIO 1987 ) The described sediments, on tile whole, are covered by course detritus carried to ilie sea by ilie torrents (locally namedfiumare) and by the wave work They are successively re-deposed along ilie coast. 3. Hydrogeological setting Water resources of Nebrodi region are considered to be scarce due to ilie large e>,.1ension of flyshioid and metamorphic outcrops Neverilieless Authors consider ilie great importance of tile karst units, that have in this area a notable extension Karst grow1dwaters are not adequately explored and exploited leaving the crowded coasta l villages witilout a sufficient arnowlt of waters for the different purposes Synoptic table n I and figure I want to demonstrate ilie importa nc e of U1e karst I ) l?esearch l!:cec111ed wit h a gram of ,1,,., !ralia11 Naliona l Research Council Srm1egic Projec/ "Crilicism of wa f er availabi liti es 10 bi! 1t1iliz e d for po1able purposes" p11bl 110 ~I 6 th Conference on Limestone Hydrology and Fissured Media 205

PAGE 217

reservoirs in the Nebrodi region with respect to the water resources of whole Sicily The hydrogeological units of the region are described in figure I : the most important formations are the limestones and the torrent and sea alluvions in the coastal sector s (Hydrogeological Unit of San Frat e llo and Hydrogeological Unit of Acq11edolci Plain) The Hydrogeological Unit of San Fratello can be con s idered an autonomous unit because it i s limited by impervious terrains that partially tampon karst groundwater flows towards the plain of Acquedolci Part of these waters rises in springs originated by contact or tectonic unconformity The most productive of springs is the Favaro spring, located at the boundary of the San Fratello karst comple"' in a zone characterized by formations with different permeability The yield of this spring that has an average value of 15 1/s is strongly influenced by rainfalls Th.is is likely due to the shor1 distance with the supply zone Tue Hydrogeological Unit of Acquedolci Plain is characterized by a sandy-gravelly-arenaceous aquifer The trend of groundwaters depends on the morphology of the impervious substratum as well as on the granulometry features of the alluvial deposits : these are subjected to frequent lateral and vertical facies changes mainl y along the foothill belt. These unit is supplied by direct precipitations as well a s limestone flows A further supply comes from the fluvial groundwaters circulating in the permeable beds of the Furiano Torrent course Springs located in the alluvial deposit s have a minor importance with a comprehensive yield of 1 5 1/s A lot of wells intensely exploit the aquifer mainly for irrigation purpose Overpumpings of groundwaters cause a sea encroachment, evidenced by the chemical analyses of well waters along the coast. In certain sectors groundwaters present an exceptional concentration of Na and Cl ions The alluvion s relevant (at left in figure 2) to the Furiano Torrent Total preclplt. I~ Sicily 18.7 10 100% .. Total lnflltroJlon 1.31.10 Rill 9 5.61.10 30% Evapotran~ 11.78.10 63% 7% Volcanltes 0.38.10 9 Carbonatlc '" rocks 0.57 10 9 Arenaceous rocks o.36.1 o Mesozoic limestones 0.36.10 9 Miocene limestones 0.21 10 9 L belong to this unit too with high porosity and penneabilit y: the y also supply wide sectors of the coastal belt. Grnundwater s in these torrent deposits close to fl y shioid outcrop s ( see figure 1 ) have a chemical trend toward s a HCOrNa type ; Cl-Na trend characterizes groundwaters close to the coastal line Cla y e y, metamorphic and flyshioid fonnations constitute the impervious units of the area with a scarce local wat e r circulation : they assume an effective importance when limit permeable units In particular as quoted above the Argill e Variegate Unit represents the impervious bed to the H y drogeological Unit of Acquedolci Plain Figure 3 shows the isophreatic lines deduced from well data relevant to a recent inventory The investigated portion of the Acquedolci area corresponds to the rectangle indicated in the upper part of figure 2 The higher gradient in the central sector of the map is related to the notable supply to the plain from the karst San Fratello comple"' in spite of the tamponing b y the clays outcrops of the Argille Variegate Unit The higher permeability of the torrent sediments (Furiano Torrent) with regard to sands and arenites causes a lower gradient of the isophreatic lines in the Western sector of the Plain 4. Geochemical and geophysical surveys This paper represents the second contribution tov.'lirds an approach to a better hydrogcological knowledge of the Acqucdolci area (ABBA TE et al ., 1994 ) Here Authors extend geochemical and geophysical prospcctings trying a correlation with the hydrogcological information. The interest offered b y this region is justified b y the local vulnerability of alluvial and karst units shown in figure 2 : they arc considered highl y vulnerable in consideration of their good permeability A contamination ri s k along the very peopled coa s tal belt i s present for sea intrusion from the North The permeability of the limestone units creates another contamination possibility Nebrodi gn 0 30 10 Unexploited groundwater resources 0 032 l 0 9 Nebrodl Mesozoic limestones 0.025 10 9 I Spr1ngyelds In Nebrodi limestones 0.095 10 9 Nebrodl precipitations 1. 109 Nebrodl Infiltration 0.07. 10 9 I Nebrodl Evapotranf p 0.63 10 Exploited groundwater resources 0.038 10 9 Wells 0.0119. 10 9 ..... Springs 9 0.0257 .10 Tab. 1: Synoptic table representing thlj water potentiality of N ebrodi limestones with respect to whole Sicily. Yearly average values are in m 206 Proceed i ngs o f the 12 th Internat i ona l Congress o f Spe l eolog y, 1 997 S w i tz e r l and Vo l ume 2

PAGE 218

in the mount zones, especially for the springs supplied by karst groundwaters In order to elaborate a vulnerability map of the area, inventories of the main hydrogeological and geophysical parameters have began in these last years. Table 2 shows the anal y ses of two water samplings of springs carried out in two different periods They generally indicate an i11terrnediate mineralization, with a Ca/Mg ratio always below unit. In particular the spring waters, relevant to the karst reservoir of San Fratello Mt., are classified as bicarbonate-alkaline earth. lrtdeed the prevalent concentrations of Ca and HCO 3 ions and the high pH values are characteristic of a carbonate aquifer 9 11 0 / / / / / / / / / / f / / / / / / / / / / 11" rr rr n1 2 ........ Crystalline rocks 12 12% Limestones .,,---~ 8 CX3% Clayey Alluvions fiysch 591% 3,86% Calcarenites 0 45% Fig. 1. Diagram showing the rock distribution in the Nebrodi region. / / / / / / / / / / / / / / / / / / / Fig. 2: Hydrogeological structures and elements of the Acquedolci area (Easl Sicily). 1: Debris (Holocene); 2: coastal andjluvial deposits (Pliocene-Quaternary); 3: sandy-gravelly complex (Pliocene-Quaternary), 4: clayey-marly-arenaceousjlyscli (Cretaceous-Miocene); 5: limestones a,ul 6: limesto11es-marls of the "Longi Taormina" S.S.U. (Lias-Eocene); 7: metamorphic complex of the "Longi-Taormina" S.S.U.; 8: clays of "Ar gille Variegate" unit (Cretaceous); 9: water-flow directions; 10: permeability boundaries; 11: sampled springs. Upper rettangle delimits area relevant to figure 3 V.E.S. (ohm.m, AB/2=20-t0m) 1. water wells (m above sea le,el) I km C Acquedolci village Figure 3 : Contour lines of groundwaters (black) and average apparent resistivity (gray) it1 the Acquedolci plain 6 '" Conference on Limestone Hydro l ogy and Fissured Med i a 207

PAGE 219

date 12/94 08/95 12/94 08/95 12/94 08/95 12/94 08/95 12/94 0 8 /95 12/94 08/95 12/94 08/95 12/94 08/95 Ump Temp. C pH Cond at Cameq/1 C0 2 meq/l Hardness S i 0 2 mg/l TDS at No 20 uS / cm 180c ma ll 1 18 1 18 7 3 7 35 541 500 3,26 2 80 0 32 traces 26,9 27 5 8 22 9 91 303 354 2 20 2 n d. 7 2 7 1 887 683 5 30 4,40 0 ,1 8 0 00 36 7 31,3 7 66 7 99 552 457 3 17 2 1 7,4 6,9 1492 1467 8 81 7 30 0 69 0 00 70 55 74 0 12,38 14 70 1097 1 1 90 4 n d n d 8 8 3 604 510 2 65 1 96 0 00 0 00 20 2 17 3 23 40 23 19 371 29 7 5 17 6 7 4 545 2 90 26 55 10 59 315 6 1 4 3 1 9 7,1 7 35 876 849 6 24 6 55 0 46 0 62 43 25 47 5 15 60 16,00 541 5-44 Mgmeq /1 Na meq/1 Kmeq/1 H CO m eq/1 C0 1 meq/1 Clmeq /1 SO meq /1 1 2 12 2 70 0,75 0 98 0 077 0 068 4 8 4 9 0 0 0 7 0 7 1 4 0 31 2 2 04 1 86 2 61 2 32 0 110 0 068 4 25 3 7 0 1 64 0 8 2 95 2 97 3 5 30 7 50 4 35 4 70 0 384 0 320 9 25 8,75 0 0 1 8 1 7 7 6 9 16 4 1 39 1 50 2 00 2 32 0 115 0 104 1 8 1 85 0 1 0 1 1 67 1 8 2 1 27 5 2 4 1 0 78 0 036 4 7 0 0 0 85 0 4 6 2,41 2 95 1 04 1 28 0 02 1 0 002 9 1 8 6 0 0 1 17 0 7 0 9 0 98 Tab. 2 : C h e mi c al analy ses of wat e r s ampl es r e l ev an t to th e Ac qu e dol ci area (s ee t ext ) nevertheless, SO 4 and SiO2 concentrations are over the average, suggesting a contribution of flyschioid terrains to the comprehensive groundwater circulation. Sampling and water analyses of springs, carried out in different periods (Dec 94 and Aug 95), seem to reveal seasona l fluctuations, especially for Mg and so 4 ions. The n otable chemical differe n ces ofnos. 3 and 4 samples are probably due to ionic exchanges between groundwater and clay minerals : indeed, sample no. 3 presents higher Na ion content, with an exceptional IDS value Geoph ys ical s u rvey The geophysical survey has been executed by means of Schlumberger V E S.' (Vertical Electrical Soundings), with maximwn spacing of 200 m V E S.' have been carried out in the sandy-grave ll y coastal deposits ; a small group of soundings have been located in the alluvial fan of the Furiano Torrent. The geoelectrical interpretation has evidenced a /roe resistivity of the aquifer ranging between 50 and 150 ohm m, with higher values up to 300-350 ohm m Values seem to increase in the upper part of the alluvial deposits of the Furiano torrent. However it has been possible to observe general l y high values of resistivity along most of torrent deposits and close to limestone relieves (debris) The lowest resistivity values (5 to JO ohm m) have been found in the central sector of the investigated area Geoelectrical quantitative interpretation has been relatively easy : indeed, the beddings of the area, at least for the upper strata, generally permit to apply the simplest model, based on the horizontally layered situation Nevertheless, in order to overcome the difficulties connected \vith the theoretical geoelectrical model Authors have chosen to perform a further representation of the geophysical measurements Figure 3 shows the apparent resistivity map (grey values) relevant to an average value of apparent resistivity computed for 40 and 80 m spacing (AB/2 values of 20 and 40 m) The selected range of AB/2 is related to the thicknesses of the upper aquifer of the plain (some tenths of meters) that has been exploited by numerous wells So, the investigation depth (up to 30 m) permitted by the chosen AB/2 range reaches the average depth of the aquifer also revealing th e eventual occurrence of the impervious basement (clays) The map allows us to interpret certain hydrogeological features It presents a minimum in the rruddle part of the surveyed area, where the aquifer reaches its lower depths Here resistivity values are less than 5 ohm m : in this last sector notable outcrops of clays occur too contributing to lower the average apparent resistivity !so-anomalous curves seem to reveal sea encroachment phenomena m the coastal central sector where resistivitie s values are less than 20 ohm m As a matter of fact \\ater-wells intensely exploit the sandy-arenaceous aquifer \\ilh the risk of chlorine contamination for a \\ide area 5. Discussion and conclusive consideration The groundwater circulation in the Acquedolci territory is strictly influenced by the structural relations between karst units and impervious formations The study of these resources with different water potentiality is carefully considered, owing to frequent drought periods Most of waters belonging to karst units (ahnost 26 1/s in total) have not yet been intercepted and exploited to adequately sup p ly this area It has to be co n sidered tha t the area of Acquedolci constitutes an important urban pole of the Nebrodi area Here nwnerous activities converge from the surrounding and inland zones, including the San Fratello Mountain range The increased water needs are actually satisfied by springs, mainly by the quoted Favara Spring (no. I in the table 2) Notwithstanding the yield of the presently exploited springs does not guarantee a constant supply to local aqueduc t s, also considering the great water demand during l arge part of the year for the inc r eased pop ul ation. Indeed, the coastal bell of Nebrodi region is greatly interested by tourism movements Goal of Nebrodi Project, to which this study belongs is the elaboration of vulnerability maps of the region Autho r s, by monitoring the quoted chemical and physical characteristics, consider necessary to survey and protect groundwaters : they could constitute, in a next future, the sole water resources of good quality of the whole region, especially in coastal sectors The first results, here exposed in form of tables and maps encourage towards a further collection of periodical data. References ABBATE R .; CIMINO A ; EMMA S ; MARTORANA TUSA, A & S ORECCI-IlO 1994. Lineamenti geomorfologici ed idrogeologici dell'acquifero carbonatico de! tcrritorio di Acqucdolci Messina Boll. Acc Gioenia Sci. Nat 27 (1994) : 579-597 BACCI G M 1989 La Grotta di San Teodoro e le sue emcrgcnzc paleonlologichc c paletnologichc Tradizioni c prospettive lppopotarni di Sicilia Sfameni Press ., Messina Italy : 13-21 BONFIGLIO L 1987 Nuovi elementi faunistici e stratigrafici
PAGE 220

EPIK, methode de cartograph i e de la vu l nerabilite des aquiferes karstiques pou r l a delim i tat i on des zones de protection par Nathalie Doerfliger et Fran~ois Zwhalen Centred' Hydrogeologie, Universite de Neuchatel, l l rue E-Argand, 2007 Neuchatel, Switzerland. R es um e La methode EPIK est une methode generale multicritere utilisee pour cartographier la vulnerabilite des aquiferes karstiques et pour servir de base a l'etablissement des zones de protection en milieu karstique. L'objectif de cette methode developpee avec la contribution de !'Office federal de l'environnement, de la foret et du paysage (OFEFP) est de produire des cartes de vulnerabilite pour des sources ou captages en milieu karstique. Les zones de vulnerabilite obtenues servent de base a la delimitation des zones de protection Apres avoir determine le bassin d'alimentation de la source, on procede en quatre etapes: I) cartographie de l'Epikarst (approche morphologique), 2) Cartographie de la couverture protectrice. 3) cartographie des conditions d infiltration et 4) caracterisation du developpement du reseau karstique Chacun des ces quatre criteres est subdivise en classes ponderees chacune par un coefficient theorique La "superposition des canes" par un systeme d'information geographique SIG des quatre criteres et le calcul de chacune des zones resultantes permet d'obtenir une carte finale de la vulnerabilite Cette methode a fait l'objet des tests sur plusieurs sites en Suisse dont quelques resultats sont presentes ici Abstrac t The EPIK method is a general multiattribute method used for the karst aquifer vulnerability mapping and to provide a base to assesss the groundwater protection zones in the karst environment. The goal of this method developed with the support of the Federal Officle for Environment, Forest and Landscape is to produce some vulnerability maps for karst spring watercatchments According to the selected attributes, the obtained vulnerability zones can be a base to outline the groundwater protection zones After having determined the spring watercatchment borderlines, we proceed in four steps: I) mapping of the epikar t (geomorphological approach), 2) protective cover mapping, 3) infiltration conditions mapping and 4) characterization of the karst network development. Each of this attribute is subdivided in classes that are weightd by a theoretical coefficient. The four attributes maps a r e overlayed using a GIS and for each zone vulnerability degree is calculated ; the resulting map is the vulnerability map This method was tested in Switzerland on several sites ,whose some results are here introduced 1 Introduction protection des captages en zone karstique n'est encore souvent Les ressources en eau en milieu karstique sont importantes pour l'approvisionnment en eau dans la plupart des pays du mondes. mais sensibles aux impacts anthropogenes elles sont considerees comme vulnerables Cette vulnerabilte s explique notamment par la structure fortement heterogene des aquiferes karstiques, avec d'une part des permeabilites tres elevees dans les conduits souterrains entoures de blocs peu permeables et d autre part des alimentations de surfaces localement tres concentrees. Vulnerab i l i te des aquiferes La vulnerabilite est une propriete intrinseque des aquifrees qui depend de la sensibilite de ces derniers aux impacts naturels et anthropogenes (FOSTER S.S.D. 1987). Elle est utilisee pour caracteriser, a !'aide d'information d'origine geologique et hydrogeologique. la sensibilite des aquiferes aux co n taminations anthropogenes, que ce soit sous forme ponctuelle ou diffuse. Protection specifique Du fait de leur fonctionnement specifique et de leur vulnerabilite, les aquiferes karstiques requierent une protection toute particuliere. La loi federale relative a la protection des eaux de 1993, exige la delimitation des zones de protection S pour tous captages utilises pour l'alimentation en eau potable : en milieu karstique leur delimitation repose principalement sur des criteres morphologiques et quelques vitesses de transit etablies par essais de tra-;age (Instructions pratiques pour la delimitation des zones de protection. 1977) Si la realisation des zones de protection de tous les captages en Suisse est pratiquement terminee, malgre ce grand effort, la qu'imparfaite. Au vu de cette situation, ii sest avere necessaire de proposer et de developper une methode de cartographie de la vulnerabilite des aquiferes karstiques a partir des differents criteres specifiques du comportement hydrodynamique du systeme karstique Cette methode se veut etre objective: elle est basee sur des criteres geologiques et hydrogeologique s et est independante de !'occupation du sol et des consideration s economiques 2. Approche multicritere : la methode EPIK L approche proposee ici pour evaluer la vulnerabiltie de s bassins d'alimentation des sources karstiques est une approche multicritere a index, appelee EPIK; elle prend en compte 4 criteres, correspondant a quatre caracteristiques specifique s du fonctionnement d'un aquifere karstique tel qu il est decrit c 1dessous Apres avoir determine les limites du bassin d alimentauon de la source, on procede en quatre etapes : I) cartographie de l'epikarst, 2) cartographie de la couverture protectrice 3) cartographie des conditions d'infiltration et 4) caracterisation du developpement du reseau karstique et attribution d un facteur global au bassin d'alimentation. La cartographie des critere s subdivises en indices s'effectue a !'aide de methodes directes ou indirectes, locales ou globales telles que l etude geomorphologique, !'utilisation d'un modele d altitude numerique. des sondages a la tariere, !'interpretation de photo s aeriennes de la geophysique des essais de trai;age 6 th Conference on Limestone Hydrology and Fissured Media 209

PAGE 221

Fonctionnement d'un aquifere karstique Les aquiferes karstiques sont caractfaises par des particularites geomorphologiques, des phenomenes hydrauliques tels que I'existence d'importantes sources, de pertes, l'absence de reseau de dra:nage de surface, l'existence de reseaux karstiques et d'hydrogrammes typiques de source A partir de ces caracteristiques, on peut proposer l image suivante d'un aquifere karstiqu e schematise A la figure I : "Un reseau de conduits connexes (reseau karstique) aboutissant A des exutoires drainant ou alimentant des volumes de roche fissuree et fracturee de faible permeabilite" Sol Epikarst 1 Source karstique Zone non saturee Zone saturee Niveau impermeable Ecoulement rapide ~Ecoulement lent Figure I: Representation schematique du fonctionnement d'un aquifere karstique (DOERFLIGER et al., 1995). Les criteres de la methode EPIK A partir de ce schema, 4 criteres des plus pertinents sont differencies comme suit : Critere E: Epikarst (morphologie karstique) E 1 Gouffres et depressions absorbants un cours d eau Dolines Lapiez Cuestas Aftleurements avec fracturation intense ( bord de route/ affleurement artificiel) E 2 Zones intermediaires entre les alignements de dolines Vallees seches E 3 Le reste du bassin d alimentation Critere P: Couverture protectrice A Sol reposant directement sur les formations calcaires aquiferes ou sur des formations detritiques grossieres tres permeables (Ex eboulis, moraines laterales ... ) P 1 0 20 cm de sol P 2 20 I 00 cm de sol P 3 > I 00 cm de sol B Formations geologiques peu permeables avec ou sans sol (Ex limons lacustres, argiles ... ) P 3 > 100 cm au total de sol et formations geologiques peu permeables P 4 -> sol avec d'epaisses [8metres] formations geologiques de Ires faible permeabilite [limons-argileux-silteux] (A verifier de maniere ponctuelle) Critere 1: Conditions d'infiltration 11 Pertes perennes et temporaires Lits et rivages des cours d'eau Cours d'eau perennes et temporaires alimentant une perte ou une doline Cours d'eau infiltrants A/ A l'jnterieur du bassjn versant des cours d eau alimentant des pertes 11 Partie du bassin versant drainee artificiellement 1 2 Partie du bassin versant non drainee artificiellement et dont la pente est superieure A I 0 % pour les zones de culture, 25 % pour les pres et les champs 13 -Partie du bassin versant non drainee artificiellement et dont la pente est inferieure A I 0 % pour les zones de culture, 25 % pour les pres et les champs B / Hors des bassins versants des cours d'eau alimentant des pertes 1 3 -Surfaces au pied d'une pente jouant le role de collecteurs des eaux de ruissetlement et pentes alimentant ces points bas (pentes superieures A I 0 % pour les zones de culture, 25 % pour les pres et les champs) 14 -Le reste du bassin d'alimentation Critere K: Reseau karstique K 1 -Reseau karstique bien developpe, A conduits decimetriques A metriques tres peu colmates et bien connectes K 2 -Reseau karstique ma! developpe, A drains ou conduits ma! connectes ou colmates. ou encore de dimension decimetrique et inferieure K 3 -Exutoire en milieu poreux avec effet protecteur (A verifier) -Aquifere fissure non karstifie Ponderation de la vulnerabilite Pour chacun des criteres E P I et K, on etablit une carte de distribution spatiale de leurs indices Ces 4 cartes sont digitalisees et transformees en format image (raster); ce format permet d'attribuer A l'aide d'un Systeme d'information geographique. les valeurs des classes en toute cellule du bassin Ce dernier est quadrille selon une maille de 20 metres de cote Puis les cartes sont supperposees les unes aux autres: les valeurs des indices des classes sont additionees et multipliees par le poids relatif, afin d obtenir une carte synthetique de vulnerabilite (DOERFLIGER. 1996). selon l'equation de ponderation ci-dessous : F = a Ei + Pj + y Ik + 8 Kt ( I ) avec F : facteur de protection Ei, Pj, lk et K1 : valeurs des indices de chaque classe des criteres / a, p, y o : coefficient de ponderation relative 21 O Proceedings of the 12 th International Congress of Speleology 1997, Sw i tzerland Volume 2

PAGE 222

Pour attribuer des valeurs aux coefficients de ponderation, nous avons effectue differents tests de sensibilite et pris en compte .1otamment Jes considerations suivantes et : -Une doline recouve.te d un sol epais (EI-P3) est plus vulnerable qu'ui,e dalle calcrire compacte surmontee d'une mince couve;tu~e pedoiogioue (E3-Pl) -Un cours d'eau a li mentant une perte (11) est tres vulnerable, tndepe ndamment de la couverture protectrice -Une vallee seche (E2) est aussi vulnerable qu'un point topographiquement bas jouant le role de collecteur des eaux de ruissellement En tenant compte de ces considerations et des differents essais de ponderation effectues, les valeurs suivantes ont ete retenues pour le ca!cul du facteur de protection : I Les indices des criteres E, P .I et K sont ponderes ainsi : E1 E2 E3 P1 P2 P3 P4 1 3 4 1 2 3 4 11 12 13 14 K1 K2 K3 1 2 3 4 1 2 3 Rappel : La valeur de ponderation la plus basse correspond a la situation la plus vulnerable Tableau 1 :Ponderation des indices des criteres E, P et 1. 2 Les criteres E et I jouant un role mains important sur la protection du karst que le critere P, nous leur avons attribue un poids relatif plus important (vu lnerabilite plus grande) Le poids attribue au critere E par rapport au critere I est identique. K a un poids relatif qui est situe entre celui de E et P. (X p y 8 3 3 2 Tableau 2 : Poids relatifs attribues aux criteres E, P et I. 3. Lors du calcul du facteur de protection F pour les differentes combinaisons possibles, on obtient alors des valeurs allant de 19 a 34. Les combinaisons sui vantes des classes des criteres donnes sont incompatibles : Ki avec EI I I et P3 ou P4 Determination des zones de protection La combinaison des di fferents facteurs de ponderation propres a chaque critere selon !'equation ci-dessus permet d'attribuer en tous points du bassin d alimentation l'une des trois zones de protection S F inferieur OU egal a 19 Zone SI F entre 20 et 25 Zone S2 F> 25 Zone S3 Tableau 3: Equivalence Vulnerabilite et zones de protection 3. Exemples d'application Deux exemples d'application sont presentes II s'agit de la cartographie de la vulnerabilite du bassin d alimentation des sources (Clarive et Tine) de la commune de St-Gingolph (YS) dans les Prealpes medianes plastiques et de la cartographie de la vulnerabilite d'une partie du bassin d'alimentation des sources du Saivu de la Font et de la Bame a Bure dans le Jura Tabulaire (NW de la Suisse). St-Gingolph Prealpes valaisannes Dans ce cas des pollutions frequentes d'origine agricole (fumier et purin) affectent la qualite des sources de la Clarive et de la Tine. L'application de la methode EPIK sur ce critere a pu mettre en evidence la relative importante vulnerabilite de ce bassin surface considerable de zones de vulnerabilite moyenne a elevee (Figure 2). Le bassin d'alimentation est caracterisee par la presence majoritaire de la zone de protection S2 (Fig ure 3). Bure Jura Tabulaire Dans le cadre de l'etude d impact de la future route nationale Nl6, le bassin karstique des sources de La Font, du Saivu et de la Bame a fait l'objet d'importantes etudes, hydrogeologiques notamment. Ce bassin (quelques 15 km 2 ) se situe en Ajoie dans le Jura tabulaire (aquifere developpe dans Jes calcaires du Sequanien inferieur). Les eaux de ce site sont drainees par un reseau karstique souterrain, La Milandrine Les trois criteres E, P, I ont ete caracterises a !'aide des donnees existantes a disposition (geophysique, !eves de forages, photos aeriennes, ... ) et des travaux de terrain (cartographie, sondages a la tariere a la main ... ). Des etudes complementaires (geophysique, forages. essais de tra,;age ... ).ont egalement ete realisees avec pour object if de tester de maniere approfondie la methode EPIK Les resultats obtenus suite a ces differents travaux de terrain ont permis d'apprecier Jes avantages de la nouvelle methode (DOERFLIGER et al., 1996) 4. Conclusions et perspectives La methode EPIK est un outil tout a fait adapte a la gestion des ressources en eau en milieu karstique Elle permet d'obtenir des cartes specifiques de vulnerabilite; ces canes constituent une nouvelle base pour l'etablissement des zones de protection en terrain karstique. Les zones de protection sont ainsi mieux ciblees, en particulier les zones les plus restrictives (SI). Combinee a une carte des risques potentiels. les canes EPIK devraient a l'avenir faciliter I'etablissement d un reglement approprie relatif aux zones de protection en milieu karstique Les exemples d'application de cette methode sur plusieurs sites tests. ont demontre la faisabilite de cette nouvelle approche en terrain karstique A ce jour, le concept d'etablissement de ces nouvelles cartes est relativement clair et les recherches actuelles s'orientent sur la caracterisation de l'epikarst (cf. PUECH 1997, in ce colloque) et le transit de contaminants donne s dans la couverture protectrice La contamination des aquiferes karstiques n 'es t pas une fatalite Une delimitation des zones de protection en adequation avec le fonctionnement hydrogeologique du karst. contribue grandement a la protection des sources et ressource s en eaux souterraines karstiques Remerciements Les auteurs de cet article remercient v i veme nt I Offi ce federal de / 'environnement, des forets et du paysage ainsi q11e le Service hydrologiq11e et geologique natio11al pour le soutien apporte lors du developpement de cette m e tlwde PIK, ainsi que Ies membres du groupe karst de Ja sociere s11isse d'Hydrogeologie Nos remerciements vont egalement au canton de Berne, Office de J economie hydra11/ique et energetique du canton de Berne 6 1 Conference on Limestone Hydrology and Assured Med ia 211

PAGE 223

References bibliographiques O0ER.-LIGER Net ZWAHLEN ?. I 995 EPIK : a new method for outlining of protect i on arr.as in karstic environment. lnternationa! Symposium on Ka:-st Waters and Environmental impacts, Anlalya I 0 20 September 1995 Sous presse chez Ba i kema D0Ef: LIGER N Hche E. et ZWAHLEN F I 996 EPIK, une method e pour cartographhr la vulnerabilite des aquiteres karstiques : appiication c!eux sites dans le Jura suisse ESRA'96 L eau souturaine en region agricole, colloque, Poitiers, 9-12 septembre I 996 S?-17-20 D0ERFLIGER N. 1996 Advances in karst groundwater protection strategy using artificial tracer test analysis and multi-at1ribute vulnerability mapping ( EPIK method) Thesis,University of Neuchatel, Switzerland, 225pp FOSTER S S D 1987 Fundamental concepts in aquifer vulnerability, pollution risk and protection strategy IN Vulnerability of soli and groundwater to pollutants (W van Duijvenbooden and H G van Waegeningh eds) TNO Committee on Hydrological Research, The Hague Proceedings and Information, No 38, pp 69-86 Instructions pratiques pour la determination des secteurs de protection des eaux, des zones et des perimetres de protection des eaux souterraines 1977, revision partielle en 1982 Office federal de la protection de l'environnement. F = Ei + Pj + Ik + K1 avec 1 = I, 3, 4 j = I, 2, 3, 4 k = I, 2, 3, 4 l= l 2, 3 Rappel: Valeur basse = tres vulnerable Valeur elevee = peu vulnerable 50 551 52 N ~E s Li mite des bassins versants topographique s 134 553 Figure 2: Carte de la vulnerabilite du bassin d'alimentation des sources de la commune de St-Gingolph (VS). S I = facteurs de protection entre 9 et 19 S2 = facteurs de protection entre 20 et 25 S3 = facteurs de protection entre 26 et 32 Legende SI S2 S3 D ..._~--. :1 Source de Clarive N oE s Figure 3: Carte des zones de protection du bassi11 d'alimentation des sources de la commune de St-Gingolph (VS). 212 Proceed i ngs of the 12 th Internat i onal Congress of Speleolog y, 1997 Sw i tzerland Vo lu me 2

PAGE 224

Contaminant transport from leaky landfills in karst areas Matt h ias Eisw irth & He in z Hotz ] Departme n t of Applied Geology, University of Karlsruhe, KaiserstraBe 1 2, 76128 Karlsruhe, Germany Abstract The transport and fate of contaminants from leaky landfills in karst areas have been investigated by soil gas s urveys and geophy s cal methods One investigation area was a municipal landfill above karstified limestones of Triassic age, Northern Black Fore s t South Germany Research work was focused on the detection of hypothetical contaminant flow paths from the landfill to a karst spring north of the landfill, which is subject to severe pollution Therefore detailed hydrogeological and geophysical investigations have been c arried out to detect contaminant flowpaths from the leaky landfill. The detailed soil gas screenings on the selected investigation profile s al lowed in combination with geophysical methods the detection of fault zones as well as the further run of preferential flow path s of leachate and contaminated gro u ndwater in the karstic limestones 1. Introduction Leac h ates leaking from damaged landfills represents a poten tial threat for groundwater and becomes more and more a matter of public concern. Old landfills with no or insufficient leachate collection systems constitute a frequent risk to groundwater qua lity Leachates and gase s percolate downwards and dilute and attenuate passing through the unsaturated zone beneath the site Such processes include both biological and physicochemical reactions A significant benefit of an unsaturated zone below leaky landfills is controlling the seepage of l eachate i n to t h e satu rated zone such that the residual contaminant load is balanced by a continuous supply of oxygen and dilution which may be avail able in t h e groundwater too Unsaturated zone calculations must therefore be coupled with determinations of the oxygen supply within the saturated zone. The l eachates from o l d l andfills typi cally contain specific organic pollutants as well as a substantial content of organic matter. The prediction of the total amount of leachate leaking to soil and groundwater is difficult, but the effects on groundwater might be severe The Department of Ap plied Geology AGK is carrying out lo n g-term monitoring of a number of leaky landfill s. This paper detai l s the results of the most recent investigations on the Grotzingen landfill above kar s tic limestone. One karst spring north of the landfill is subject to severe pollution. Water samples from this spring and groundwater observatio n wells indi cate the existence of an aquifer pollution by waste site leachate Therefore detailed hydrogeological and geophysical investiga tions have been carried out to detect the contaminant flowpaths from the leaky landfill 2. Geology and site investigations The investigation area was a municipal landfill above kar s ti fied limestones of Triassic age Northern Black Forest, Southern Germany (Fig l and F i g. 2). Fi g 1 : Map o f t/ 1e landfill Gro tzi n ge n in t h e N o rt h e rn Bla ck F or est, S outh G e rman y. T he i nv estig a ti o n pr ofiles 1, 2 a n d 3 a r e i nd c a te d as we ll a s t h e proj ecti on of th e ex t e n si on faul t z on e. 6 Conference on Limestone Hydro l og y and Fissured Med i a 2 1 3

PAGE 225

NW Rhine Graben landfill Grotzingen SE ;r mo mu mu Berghausen 0 100 200m Fig. 2: Geological cross sectio11 A-A' as i11dicated in Fig. 1. (q, q1,q1 = Quaternary sediments, mo = Upper Muschelkalk, mm = Middle Muschelkalk, mu = Lower Muschelkalk, so = Upper Buntsandstein). The Upper Muschelkalk limestones, quarried at Grotzingen, consists of fine-grained, micritic limestones, intercalcalated with numerous siltstone and mudstone bands Fracturing and fissuring within the limestone is extensive and regionally, secondary permeability is as significant as interstitial flow. The karstified limestones of the Middle and Upper Muschelkalk in the investi gation area strikes West to West-East and dips with 5 to 10 degrees North to North-East (Fig 2). For the hydraulic regime Geological investigations suppose a NNE striking extension fault zone underlying the landfill Landfilling of primarily domestic waste commenced in 1973 into the working limestone quarry of Grotzingen (Fig 3) His torical investigations showed that the former quarry bottom and walls were not correctly sealed before landfilling. Leachate leaks from this areas into the unsaturated limestones. The mean aver age Jeachate exfiltration is calculated to I 5000 m 3. year until 1991 when a I m thick impermeable layer was installed above the waste Even after the completition of this inserted layer about 5000 m leachate leaks every year into the underground. Fig. 3: Aerial view of the la11dfill Grotzingen (view from North to South). Water samples from springs north of the landfill and various groundwater observation wells indicate clearly an aquifer pollu tion by waste site leachate. The leachate front has migrated at a significant distance within the unsaturated zone and groundwater without dominant attenuation and dilution This indicates that leachate migration through the unsaturated zone was not in form of an intergranular flow (average flow rate 1-4 myear '; LEWIN et al ., 1992) Rapid fissure flow on discrete flow paths within the karstified limestones seems to be the dominant leachate transport mechanism below the Grotzingen landfill (EISWIRTH 1995 ). Reaching the groundwater the leachate plume is transported mainly on discrete karst flow paths and with the dipping of the limestones into northern-north-eastern direction Therefore most of the I O investigation cross-sections were established north of the landfill (Fig. I) 3. Methodology Soil gas Screenings The aerobic biodegradation of leachate compounds in the un saturated and saturated zone decreases 0 2 and increases CO 2 in the soil air and in the groundwater (HENDRY et al ., 1992; DEYO et al 1993) The degradation of leachate compounds on small preferential flow paths, such as a fault zone, leads to elevated CO 2 and depressed 0, concentrations (compared to background levels) in both groundwater and soil gas within the karstified limestones (e.g. RivETI & CHERRY, 1991). The CO,-content in soil air at any soil depth is mainly controlled by a combination of the following processes : (I) biological production (organic matter decomposition, root respiration), (2) diffusional transport, and (3) equilibration between transported CO and soil water. The average CO 2 residence times at various depths under a variety of soil conditions were calculated using a steady state soil CO 2 model (AMUNDSON et al., 1995) 1n this model it was as sumed that the average time of soil CO 2 at any soil depth interval was equal to the mass of CO 2 present (measured in the field) divided by the amount of CO, diffusing through the upper boundary of that interval The CO 2 -flux through the top of any soil depth interval was calculated with known atmospheric diffu sion coefficient D, soil temperature T soil porosity E and atmos pheric pressure p The results indicate that in the upper 50 to 100 cm of most soils the residence time of CO is on the order of a few minutes to hours while below that depth, residence times are in the order of tens of hours up to SOO hours. Soil gas screenings have been carried out in boreholes deeper than 100 cm to avoid atmospheric influences and to exclude root respiration and organic matter decomposition in the upper soil sequences as source of high soil CO 2 -concentrations Landfill gas monitoring and migration calcu lations indicates that there is no active gas migration from Grotzingen landfill into the surrounding ( METCALFE et al ., 1987 ; BARBER et al., 1990). Measured soil gas CO,-anomalies corre spond therefore to a high diffusional CO ,transport on fissured 214 Proceed in gs of the 12 '" International Congress of Speleology 1997 Switzerland Vo l ume 2

PAGE 226

zones, derived from biological organic matter decomposition on discrete contaminant flow paths. The radioactive noble gas 222 Rn (t 1 12 = 3.8 d) is produced by decay of 22 "Ra (t 112 = 1600 d), a trace component of all soils. Only a smalJ fraction of totally produced 222 Rn escapes from soil parti cles into soil air depending on the specific soil particle surface and grain size distribution Gas transport is mainly by molecular diffusion The 222 Rn flux and 222 Rn content at the soil surface depends on the 22 "R.a content, the permeability and diffusion resistance of the unsaturated soil zone and many other parameters e.g. soil temperature, soil moisture, atmospheric conditions (WASHINGTON & ROSE, 1990). The spatial pattern of the 222 Rn flux depends more on soil type than on the 22 "Ra activity of the soil material (DORR & M0NNICH, 1990) Shear zones are com monly enriched in clay material containing uranium minerals and can produce elevated radon concentrations in overlying soils. Elevated 222 Rn content in soil gas probes often refers therefore to underlying shear zones and fault zones with high Radon emana tion and high rock permeability. Soil gas surveys were performed on site using portable CO 1 0 1 CH and ~S detectors (Meta, Inc Altenberge, Germany) 222 Rn-activity of the soil gas was analysed with an electronic alpha-particle scintillometer G.B. 83 (Gewerkschaft Brunhilde, Uetze, Germany). With a plastic hydroderrnic syringe about 170 ml soil gas was injected into the silver endowed zinc sulfide coated cell of the scintillometer. The alpha-decays of 222 Rn within the cell were detected and counted by a photomultiplier. After calibration of the cell the registered counts per minute cpm can be converted into radon concentration of the soil air [Bq m ]. Gas samples were collected in glass containers or concentrated on charcoal traps To prevent atmospheric dilution of the soil gas samples a bentonit slurry was molded around the probes at the surface By sampling at depths > 90 cm the effects of barometric pressure, rainfall and air pollution are minimized Soil gas sampling was carried out in 10 investigation profiles (150 to 500 m length) in the surroundings of leaky landfills within 1.5 m deep, small diameter boreholes. The boreholes were closed against the atmosphere during the measurements with blo wed up packers The general distance between two measuring boreholes was 10 m Nearby the supposed extension fault the distance between the boreholes was reduced to 2 5 m Electromagnetic soundings The electromagnetic soundings on the cross sections were carried out with direct current process. Horizontal coplanar and vertical coplanar coil configurations with spacings of 10, 20 and 40 m between transmitter and receiver lead to 6 various depth penetrations (2 5 m to 40m depth) along the cross-sections. 4. Results The results wi11 be summarised by showing the investigations of profile I, 2 and 3 Further details on all profiles and on the techniques applied are given in EISWIRTH (1995). In all soil gas screenings both CH, and H S have been below the detection limits of the instruments used Profile 1 During the soil gas investigations on profile I soil gas CO maximas were detected at location 73 m and 85 m (Fig 4). These maximas corresponds clearly with the detected minima in the soil gas O ,concentrations. Both positive CO and negative O anoma lies refer to contaminant transport on the extension fault zone. The aerobic biodegradation of leachate compounds in the unsatu rated and saturated zone decreases O and increases CO in groundwater and soil air The elevated CO and depressed 0 1 concentrations are caused by biodegration of leachate in a dis crete flow paths within the karstified limestones and overlying high permeable fault zones The results of the 221 Rn investigations indicate also a main anomaly at location 73 m and three minor anomalies at location 10 m, 85 m and 130 m (Fig 4) The 222 Rn anomaly at location lO m and 130 m corresponds to areas with a high soil clay content. These areas contain many bricks, but nearly no organic material The main 222 Rn anomaly at location 73 m refers to the extension fault zone and is probably caused by a high Radon emanation in the fault zone m._ -(qo, ) 2 ;0i\l>l"l "' 20 ,.. II 17' 16 "" ll 12:! 10 1 00 .,, ,. 2:! +~....:::2~=#=:::::l;l:===1\l=df-~-.,........:~~-.,........::::i;t:==!lt::=:!;t:::~ o @) 0 10 ]D lO SO 60 10 a, to um 110 llO 1.10 l
PAGE 227

limestones, partly filled with either clay minerals or soil water. At location 130 m this low resistivity zone reaches up to the surface m._ 1 ..-1 2~"l ., "' II ""' 16 175 150 12 125 10 Ill) 75 ,. +---,-----..cl'f----,-~--.:>,i=::::::l!t::=:f,!:::::::::t,t:=:i;i=::---r-.------+ o 10 lO lO 40 .SO' 70 _, ,0 ux, 110 l lll llO 1'10 uo@ -l -10 B U 5 -20 .:-2l -JO -Jl -Jl +--~.------,---~--.-,-.....,::::::,,,,,--.~=,.----r""::::...+-,o 10 20 JO 40 lO 60 70 10 90 100 110 120 llO UO l lO lou1lo1 (m) Obm-11 I I I I I I I I I I I I Iii o 10 100 uo 200 uo 100 no oo oo Fig. 5: Results of the soil gas screenings and electromagnetic soundings on profile 2 north of the leaky landfill Profile 3 In contrast to the results from profile l and 2, on profile 3 no sharp soil gas anomaly was detected. The broad area between location 50 to 150 m shows high CO,-concentrations (up to 6 Vol.%) and low O -concentrations (min. 14 Vol. %). At location 80 m, 100 m and 120 m the soil CO -concentration was low due to very high soil moisture contents (Fig. 6). The broad soil gas anomaly between 50 m and 150 probably refers to a broad fault zone, but it seems that the main contaminant transport will follow the earlier mentioned extension fault zone in north-eastern di rection (Fig. I). m._ --1 ... 1 ooi;"""""I "' ""' 175 150 '" Ill) 75 ,. ., II 12 ID 0..i..:::i.t=d:!:==:!:t=:;:::._ ___ ~.----~:Jf..--___);j~.----~---,----+D 10.,., m m l50 150 rn l -10 a .u 5 -20 .: -2l -JO -Jl ------------------.0 "-.J)~( u -l 10 ll -20 -2l -JO -ll 0 10 20 JO 40 lO 60 70 10 90 100 11 0 1 20 llO 140 llO l out101 jmJ Obm-m I I I I I I I I I ! l I f 0 jO 100 UO 200 UO JOO HO 4 00 HO Fig. 6: Results of the soil gas screenings and electromagnetic soundings on profile 3 north of the leaky landfill. The results of the 222 Rn investigations on profile 3 shows a broad anomaly between location 115 m to 150 m (Fig 6) The 222 Rn anomaly at this location corresponds probably to a high permeable fault zone (Fig 6). The results of the electromagnetic soundings on profile 3 show a broad zone below 20 m with low apparent electrical resistivities. The low resistivities (< 100 Ohm m) refer to fis sured, partly water saturated limestones. The boundary between the Upper Muschelkalk limestones and the Middle Muschelkalk siltstones was not detectable in this profile (Fig. 6). 5. Conclusions With the combined application of soil gas screenings and geophysical investigations, an extensional, inclined fault zone was detected in the karstified limestones forming the under ground of the landfill Grtitzingen. Leachate water leaks from the Grtitzingen landfill into this fault zone and percolate down to the groundwater with low attenuation and dilution on discrete flow paths. Acknowledgements: This paper is presented with the permission of the City of Karlsruhe who operate Grtitzingen landfill The authors would like to express their gratitude to B Krauthausen, HYDROSOND for permission to publish the results and for funding the research work. The views expressed are those of the authors, and do not neces sarily represent those of either of the above parties. References AMUNDSON, R. & Y. WANG 1995 The relationship between the oxygen isotopic composition of soil CO, and water. in IAEA SM-336/2 proceedings of the International symposium on iso topes in water resources management, 20-24 March Vienna BARBER, C ., DAVIS, G B BRIEG EL, D & WARD, J K. 1990 Factors controlling the concentration of methane and other vola tiles in groundwater and soil-gas around a waste site. Journal of Contaminant Hydrology 5 : 155-169 DEYO B.G ., ROBBINS G.A. & BINKHORST G K 1993 Use of portable oxygen and carbon dioxide detectors to screen soil gas for subsurface gasoline contamination. Ground Water 31(4) : 598-604 DORR H & M0NNICH K.O. 1990. 222 Rn flux and soil air concentration profiles in West-Germany Soil 222 Rn as tracer for gas transport in the unsaturated soil zone. Tellus 42B: 20-28. EISWlRTH M. 1995 Characterisation and simulation of con taminant transport from sewerages and landfills. (in German) Sehr Angew Geol. Karlsruhe 38 : 258 Karlsruhe HENDRY M.J ., LAWRENCE, J.R ., KIRKLAND, R & ZANNYK, B.N 1992 Microbial production of carbon dioxide in the unsaturated zone of a meso scale model, in KHARAKA & MAEST (eds.) Water-rock interaction, Balk. 287-290. LEWIN, K., YOUNG, C P & BRADSHAW K. 1992 Impact of waste disposal on the chalk and Sherwood sandstone aquifers manuscript International Symposium on Groundwater pollution at waste disposal sites : Detection, attenuation and remediation 20-22 March 1992: 16-32. Copenhagen. METCALFE, D .E. & FARQUHAR, G.J. 1987. Modelling gas mi gration through unsaturated soils from waste disposal sites. Wa ter, Air, and Soil Pollution 32:, 247-259 RlvEIT M O & CHERRY J.A. 1991. The effectiveness of soil gas surveys in delineation of groundwater contamination : Con trolled experiments at the Borden field site. proceedings of the Conference on Petroleum Hydrocarbons and Organic Chemicals in Ground Water, National Water Well Association, Houston 20 22 November: 46-64. Texas. WASHINGTON, J W & ROSE A.W. 1990 Regional and tem poral relations of radon in soil gas temperature and moisture Geophys. Res. Let 17(6) : 829-832 216 Proceedings of the 12 1 h International Congress of Speleology 1997, Switzerland Volume 2

PAGE 228

South Dobrogea karst vulnerability by assessment of some geophysical and geochemical parameters b y Rado Gaspar & I. Paunica I n stit u te of Ph ys ics and Nu cle ar En gi n ee rin g, Buchar es t Romania Abstract The aquifer resource s from South Dobrogea karst system are influenced by anthropogenic process (discharge, heat tran sfer) and pollution (nitrates heavy metals, organic compounds) Hydrochemical methods and neutron activation analysis were used to determine pollutant concentrations. Hydrodynamical parameters were determined by using environmenta l isotopes methods and artificial tracer techniques. Using lumped parameter models of the groundwater dynamics from isotope and artificial tracer data, some parameters of karst vulnerability to pollution were assessed. R e sum e Les ressources aquiferes du systeme karstique de Dobrogea du sud sont intluencees par des processus anthropogenes (deversements, transfer! de chaleur) et pollution (nitrates metaux l ourds composes organiques etc.). Les methodes hydrochimiques et !'analyse par activation neutronique ont ete utilisees pour la determination de la concentration des polluants. Les parametres hydrodynamiques ont ete determines a 1 aide des isotopes de J environnent et des techniques utilisant traceurs artificiels. En utilisant les donnees obtenues avec traceurs et modeles d 'e coulement adequats on a pu evaluer la vulnerabilite de s eau" karstiques 1. Introduction Industrial, agricultural and other human act1V1t1es have had and continues to have a significant impact on groundwater in most countries especially in Eastern and Central Europe and in many developing countries. The South Dobrogea karst has a special hydrogeological significance for Romania there fore it has been and still is the object of the interest of the administrative factors as well as of the researchers In order to qualify survey of the karst aquifers the main reservoir for drinking water supp l y of the spa and littoral cities, starting with 1990 a regional monitoring system has been constituted with the Barremian Jurasic and Sarrnatian aquifers as main objectives. To determine the karst vu l nerability and the immigration potential of pollutants important studies has been performed. Hydrochemical methods and neutron activation analyses were used to determine pollutant concentrations Hydrochemical and hydrogeological parameters were determined using environmental isotopes methods and artificial tracer techniques 2 Regional hydrogeologic information Regional information were collected in order to identify the hydrogeologic subsystems which can be affected by pollution (e.g. in case of an accidenta l release either directly or indirectly through surface water). The information collected includes : climatological data. major hydrogeologic units, water-bearing characteri s tics of the units, recharge and discharge relationships and data on surface hydrology Data on the type and stratigraphic distribution of the various geological formations were collected in order to characterize the regional system and its relationship with the local hydrogeologic units This information was supplemented by data from previous exploratory work in the region. The hydrogeological subsystem from South Dobrogea consists ?f man y aquifers in connections between them located 111 Paleo zo ic. Jurassic Cretaceous. Sarmatian and Quaternary formations. The Paleozoic deposits has a big nonuniformity and a hvdraulic conductivity between 0.01-0.15 m i d some 0 horehok s water having a tt:mperature between 23.5 C and 27 0 C. Fig ur e I H yd rauli c co nn ec ti o n s d e t e rmin e d u si11g r a di o a c ti ve a nd ac ti vab l e tra ce r s in S o uth D o br ogea karst sys t e m 6 th Conference on Limestone Hydrology and Rssured Media 21 7

PAGE 229

The Jurassic is formed by fissured limestones and is characterized by artesian sulfurous geothermal waters The Barremian-Jurassic aquifer is the main aquifer and the main research objective of the South Dobrogea .The great extent of the natural recharge area of this BarremianJurassic aquifer is outside the national boundaries and cannot be directly surveyed Between the Eocene Cretaceous and Jurassic there is not continuous impermeable horizons and make a great complex aquifer with sulphide bearing waters The transmissivities of the shallow aquifer varies between 50 and 2000 m 2 /d The transmissivities of the deep aquifer vary between 200 and 150 .000 m 2 /d In the schematic representation of the groundwater flow model in the Barremian-Jurassic aquifer (fENU et all 1987) the flow direction is from SW to NE. In the karst subsystem from Mangalia including the Movile Cave the local flow direction is different. Thus using artificial tracers (ln-EDTA and NI f/ 2 Br) for labeling of the Kara Oban sinkhole, a divergent flow resu l ted (figure I). In this labeling which led to the establishment of a karstic diffluence part of the tracer was guided toward the Mangalia sulfurous lake The tracer was recovered in three subaquatic mezothermal springs (figure 2) From the types of concuration-time curves after the processing the experimental data (MALOSZEWSKI & ZUBER 1984) resulted : low longitudinal dispersivities are attributed to solution channels, in which disper s ion is poorly dependent upon flow distance ; high longitudinal dispersivities are attributed to flow in fissures and preferential diffusive tracer exchange within fissures ; the transverse dispersion in the studied area is pronounced t Vl '-.. u w 0:: z ...... 1z ::, 0 u 20 MANGA LI A SUBMARINE SPRING 30 L.0 TIME [hl 50 Figure 2. Tracer recovery i11 a mezothermal karstic spring from Mangalia Lake 3 Environmental isotope studies and hyd r ochemical characteristics of the groundwater resources Groundwater studies during the last 8 yea rs was performed in a monitorin g network of the boreholes and surface water s These studies including environmental isotope (T, D, 18 0) and hydrochemical and radioactivation analysis revealed the behaviour of shallow and deep aquifers Thus the T concentrations decrease 10-100 m below surface between 54 TIJ for Tatlageak lake and 5 TIJ for a deep borehole, in Paleozoic.(The average value in 1995 year for the Movile Cave was 32 TIJ) The stable isotope content D and 18 0 demonstrated the origin of shallow and groundwater In the f/ H versus 6 18 0 diagram (SARBU & POPA 1991) point s fall for fresh water on the "local meteoric water line Instead for the geothermal water line points fall below the local meteoric line For a synthetically spatial hydrochemical feature of different water types in their vertical succession we have chosen the mean anionic composition Thus the variation of TDS values between surface waters and the Barremian-Jurassic aquifer varies from 700 mg/I to 638 mg/I for the same interval the change of the anion concentrations is from 150 to 96 mg/I for er from 139 to 43 mg/I for SO 2 4 and for HCO 1 from 198 to 294 mg/I (TENU & all, 1991) For sulfurous waters the correlation between the variation of the TDS and H 2 S concentrations (m/Kg)tt 2 o with depth and temperature is presented in figure 3 The point represents the Movile Cave where a unique cave ecosystem is developed 0 E :r: ~500 D... w D 1000 oTOS(B) + TDS(H) ' M= Movile Cove \. \ "\ 0 \ I .._ / 1000 200 0 3000 TDS (mg/ l] --30 u 20 cp w 0:: ::> ti 10 0:: w D... w I IFigure 3. Corre/atio11 clouds between temperature and depth for South Dobrogea underg r ound waters. 4. Polluting sou r ces and pollutants nature The aquifer resources from Southern Dobrogea karst system are influenced b y anthropogenic processes (discharge heat transfer) and pollution (nitrates heavy metals organic compound s The higher vulnerability to pollution of the shallow aquifer consists not only in its recharge from irrigation system but al so in the leakage and the seepage lo ss of industrial estates and factories (ZAMFIRESCU & all. 1994) + J0 Specific pollution indicators are measured : NH 4 PO 4 N 2 NO 3 Fe 2 + Toxic components as organo-chlorinated pesticides herbicides hydrogenated hydrocarbon s have been determined In surface waters (Tatlageak brook) l'vln Cd ++ and Pb ++ were also determined (see table I) In the same time, a gradual increase towards the South of the NO 3 content was observed 218 Proceed in gs of the 12 1 n International Congress of Speleology, 1997, Switzerland Volume 2

PAGE 230

TDS Mn++ Cd++ Pb++ 0 (0C) (mg/I) / 1 ) (g/ 1 ) (g/1) Tatlageak 793 282 7 1 00 brook Borehole 11 88 1 4 5 8 38 25 4073 Mangalia lake 4480 24 9 I 0 spring Albesti brook 1 583 178 2 UDL Borehole 11 32 30 1 2 28 22 8 4074 Litora l sp r ing 1 282 27 1 2 5 Venus Factory 1229 7 7 30 23 boreho l e Movi l e Cave 949 UDL UDL UDL 1 9 Tabl e 1 S o me ph ys i co c h e mi c al chara c t e risti cs (U DL = u11d e r d e t ec ti o 11 limit ). References GASPAR,E & I.ORASEANU 1 987 : Natural and art i ficial tracers in t h e study of the h y drodynam i cs of karst. Theoretical and Applied Karstology 3 ; 31 107 SARBU S N &R POPA 1992 : A unique chemical antropogenic based cave ecosystem. The natural history of biospeology Ed Ana Isabel Comach M a drid TENU A.F.DAVIDESCU S SIMIONAS L.EICHINGHER B.BERTLEFF W M l CHEL. E.SECARA, Z CARADIMA & N PITU 1994. Asses s ment of the po ll ution impact on the Groundwate r Qua l ity in South Dobrogea, Proc Symp Impact of Industrial Activities on Groundwate r Constanza, 557. ZAMFIRESCU F V MOLOOVEANU C.DINU N PITU M.ALBU A.DANCHIV & H NASH, 1994 : Vu l nerabi l ity to pollution of karst aquife r system i n Southern Dobrogea Proc Symp.lmp a ct of Ind u strial Ac t ivities on Groundwater Constanza59 I 6 1 Conference on Umestone H ydrology and Fissured Media 219

PAGE 232

Geophysical methods as a tool for speleological and geotechnical investigation in karst Maira, Plateau de Bure (JU), Switzerland Meier Edi 1 ; Huggenberger Peter1; Stiefelbagen Wilbelm 3 ; Millier Imre3; Christe Romain 4 1 Edi Meier+ Partner AG, Hard 4, CH-8408 Winterthur, 100777.3453@compuserve com 2 EA WAG, CH-8600 Dilbendorf, Huggenberger@EAWAG CH 3 Universite de Neuchatel, Centre d'Hydrogeologie, CH-2007 Neuchatel 4 MFR Geologie-Geotechnique SA, Rue de Chaux 9, CH-2800 Delemont Abstract Ground penetrating radar (GPR) profiles have been recorded on the karst plateau of Bure in the NW part of Switzerland The NE part of this plateau is drained by a subsurface river. The axis of the planned highway N 16 crosses the karst plateau at the test-site near the locality of Maira. At this site, the water level is located about 50 m below the surface Speleologists early on recognized chimney systems several meters in diameter which come up near to the surface. These chimneys potentially present a major source of difficulty in the highway construction, e.g. for foundations, blocking of tunnelling machines etc. After profile-mapping the area every 5 m with continuously measured Very Low Frequency Electromagnetics (VLF-EM) method, a series of GPR measurements followed in summer 1995. The GPR experiments performed with sets of 50 MHz and I 00 Mhz antennea showed reflectors up to a depth of about 25 meters in the limestone sequences. The soil thickness in this case was in the order of decimeters. In areas of increasing clay-rich soil thickness, the penetration of the GPR waves is reduced to several meters. From both these methods, chimneys and the orientation of the main geologic structures can clearly be identified. A comparison of the results of the VLF-EM and GPR measurements allows us to establish some general statements on the suitability of the two methods to address the questions mentioned above 1. Introduction regional setting The Ajoie plateau is situated north of the external part of the Jura arc, near the French-Swiss border The backbone of the plateau is formed by an essentially calcareous, subhorizontal Upper Jurassic series (approximately 300 m thick) It is cut by north-south trending, subvertical faults, whose throw gives the plateau a "piano key" structure Karstification of the limestone is well-developed: the surface of the limestones is locally relatively karstified (epikarst) and covered by a soil which is generally relatively thin(< 0 5 m), but in some places, up to 15 m thick, several active caves are known, of these the largest is the Milandre cave, which is located just below the Swiss N 16 highway project. The Milandre cave network is situated 40 and 50 m beneath ground level. Speleological exploration has recently mapped chimneys going up to only a few meters below the surface. Some of these chimneys are located directly beneath the projected highway, especially in the area of the southern entrance of the tunnel. These karstified features occur either as voids or as caves filled with an admixture of limestone blocks and fine sediments These zones present collapse and settlement risks associated with the construction of the highway. Figure I: Location of the test-site at Maira 5 th Conference on Limestone Hydrology and Assured Media 22 1

PAGE 233

The following problem is to be solved by the engineers: karst phenomena such as those mapped by the speleologists (figure 2) probably occur elsewhere under the projected highway but near the surface voids can not be detected. thin cover I -I I -11-1 I -\ \-1 II \ -L -1 I I\ karstified zone \ \ I ? \ \-1\ I \ \ ,Figure 2: Cross-section of the Milandre cave network The aim of this investigation with geophysical methods was to see how well karst phenomena may be detected by using a combination of different geophysical methods : first a survey with VLFmethods (TuRBERG, 1993; MOLLER et al., 1995) has been carried out to define large scale geological structures (local regional geological faults : 10 to 1000 m scale) second to test the GPR method, which is suitable for the detection of shallow subsurface anomalies. 2. Basic principles of the applied methods Ground penetrating radar (GPR) and Very Low Frequency methods (VLF) react to changes in the electromagnetic properties of the geologic materials VLF uses electromagnetic energy in the longwave and very longwave band range as source and GPR electromagnetic waves in the HF/VHF band range The Very Low Frequency (VLF) Methods The VLF-EM method measures the relation between the primary horizontal magnetic field and the magnetic field induced in the subsurface Source for the primary field are terrestrial radio transmitters in the range of 15 to 300 kHz. It allows to record continuously measured parameters and to image the subsurface between 5 and I 00 m penetration depth depending on the transmitter frequency. The VLF-R method measures horizontal magnetic and electric fields. The relation between these fields is expressed as appearant resistivity and phase (TURBERG, 1993). This method allows us to calibrate the VLF-EM measurements and to estimate the penetration depth of GPR The advantage of the VLF-EM is greater sensitivity to vertical structures and the possibility to record the data continuously. This results in higher lateral dataresolution Extreme points and the position of the strongest derivative give hints where to search for geological anomalies The relation between the measured fields are complex numbers. The imaginary part is called ,,out of phase component and used for interpretation because it is more stable than the real component. The Ground Penetrating Radar (GPR) method The GPR system consists of a pair of separated transmitter and a receiver antennae The transmitter antenna emits short electromagnetic pulses of a certain central frequency and bandwidth Reflections occur when the electromagnetic waves meet boundaries between geological units of contrasting dielectric properties The reflected waves are recorded by the receiver antenna The elapsed time between the transmitted pulse and the received echo and the propagation velocities of the electromagnetic waves allow to estimate the depth of the reflections. 3. Results of the VLF-EM Survey The three-dimensional representation of the data measured at Maira illustrates the VLF-EM method The profiles in figure 3 have been chosen to compare the VLF EM data with the GPR data and the postition of the cave Several frequencies were measured in NS profiles Two transmitters in EW direction (18 3 and 162 kHz) are presented here This configuration is sensitive to EW anomalies Since this sensitivity depends on the transmitter direction NS anomalies should be mapped by transmitters in NS direction and profiles in EW direction In the figure the position corresponds to the first rising ramp of the 18 3 kHz curve at a position with maximal first derivative. This correlation is possible in each of the measured profiles between x =567' I 00 m and x =567' 170 m Even some minor variations can be seen in parallel profiles The effects observed for 162 kHz are much weaker Its penetration depth is smaller than for 18 3 kHz because higher frequency results in better resolution but smaller penetration depth. From the speleological investigations we know the location of the cave at y = 257'342 m toy = 257'345 m. Jn the 18.3 kHz curve we find a stong derivative at this point. This indicates a geological anomaly like a fault. A strong derivative can also be found in the 162 kHz curve, but at y = 257'355 m We can not deduce the precence of a cave but we have evidence for a north-south dipping fault 4. Results of the GPR Survey Evaluation of the suitability of the GPR method and optimization of the acquisition parameters For planning the GPR Survey a model calculation was performed. Based on earlier VLF-R Surveys (TuRBERG, 1993) the apparent resistivities and the lithologies of the geological formation at the test-site were known. With this prior information and the expected size and shape of the geologic inhomogeneities, we evaluated the optimum 222 Proceed ings of the 12 111 International Congress of Speleology 1997, Sw itzerland Volume 2

PAGE 234

Distance Y [m] 257300 257310 257320 257330 257340 257350 257360 257370 257380 257390 257400 4 I I 1 I I I I I l 2 0 :::.e e:. 2 4> Ill "' Q. -4 'S 0 -6 -18 3kHz --162 0kHz -8 L + a ) 10 Position [m] .... u:l crl tsl r,j u:l al r,j ..;. al Ji ..,. ..,. ..,. !I' """ I I I I I I I 111 I I It I 0 lliil@uruum1111111111111111111111111111111111m111u11111111111mn111111111!)llllllll!lllllllllllllllllllllllllll!lll!lllll_llOIIIIIIIL11UIIU1 U lllllllll ll!_llrr1111(!ll!!lll!!lllllllfj(lf1Un11111l!!IIUIW111U1m 0 C ti1l11ii11r11 Cla -rich sediments ........ _ ... 0 e 100 u ,i' N .... .s n 4> > E i= 200 !. -s Q. .?12'0 b ) 400 Figure 3: a) VLF-EM data recorded at Maira b) GPR data recorded at the same test-site as above. antennae frequency and the acqu1S1t1on parameters by using the radar range analysis proposed by ANNAN and DAVIS 1977 As a result of these calculations for the 50 MHz antennae the maximum penetration depth was expected to be on the order of 26 m in high resistance areas and 6 m in areas of low resistance (assumptions : no soil cover air-filled cavity height 15 cm, diameter 80 cm) Field Measurements Seven profiles were recorded across the caves known from the previous speleological investigations We recorded a radar trace every 25 cm We used the 100 MHz antennae for two of the GPR profiles 'and the 50 MHz antennae for the other five sections To determine the velocity distribution, a set of CMP-data (common mid point ROBINSON, 1989; MEIER & HUGGENBERGER 1992) was collected To minimize the dammage on the cultivated land, the long antennae axis were oriented parallel to the profile trace Presentation of the data GPR-data are presented in two-way-traveltime distance diagrams. The traces are displayed in the wiggle mode For the conversion of traveltime to depth a mean propagation velocity of 12 cm / ns was assumed Interpretation A continuous reflection ascends from a depth of I O m at position 14 m to a depth of 5 m at position 40 m Between 40 m and 42 m the reflections are discontinuous and dipping in both profile directions A clear hyperbolic shape at 41 m is interpreted as a cavity or a karst deposit at the depth of about 7 m The continuous reflections represent ,, undisturbed Upper Jurassic series Figure Jb 6' h Co n ference on Limestone H ydrology and Fissured M edia 223

PAGE 235

shows that for the 50 Mhz antennae the penetration depth of the GPR waves is in the order of 12 m Therefore, the penetration depth is less than the depth of the caves explored by the speleologists However, the near-surface cavities, which can be precisely located by the GPR method and thus this method is very important for geotechnical problems Dataprocessing for automatic cave detection To find a way to interprete GPR data independant on the visual reflection pattern, a fourier analysis of the signals for different profile sections was performed Figure 4 shows the result of the applied fourier analysis 80 "Karst Zone" Pos 38 46 m (100 350 ns) 10 \/ Pos 30 38 m (100 -350 ns) o-L-----+------+----+-----+--'-30 0 50 60 70 Frequency (MHz] Figure 4: Fourier analysis of the data shown in figure 3b References: using the traces from position 30 m to 50 m Obviously the amplitudes of the signals in the karst zones (38-46 m) are two to three times higher than in the neighbour sections (46-50 m and 30-38 m) 5. Comparison between the VLF an GPR Results Allthough the two investigation methods are based on electromagnetic waves, there are some important differencies GPR uses electromagnetic waves in the high frequency (HFNHF) domain. This frequencies are nearly 500 times higher than the frequencies used by the VLF method resolution of electromagnetic investigation methods also increases. The investigated depth decreases with increasing frequencies Therefore combining the two methods yields best results 6. Conclusion and further recommendations The Very Low Frequency Electromagnetics (VLF-EM) method developed at the Centre d Hydrogeologie Neuchiitel enables mapping geological faults. Along these faults water may be drained leading to karstification and excavations. In contrast to GPR, the VLF method is less sensitive to poor conductors like caves. Nevertheless an investigation for the presence of caves may be reasonable because the underlying geology is easily detected with this method Furthermore the coating of the caves may consist of a conducting material like clay Mineralized water drained by the cave may also cause conducting anomalies The aim of our measurements was to gain some idea of these effects and to test the applicability of electromagnetic methods for near-surface search of cavities The G PR method shows a very good spatial resolution of the detected anomalies This allows to recognize shallow karstified zones. In non-destructive investigations the GPR method has become a common method We suggest that amplitude analysis might be a useful tool helping to interprete GPR data in karst areas ANNAN, A P ., DAVIS J L. (1977) : Radar range analysis for geological materials, Report of activities, Part B; Geol. Surv. Can ., Paper 77-1 B DAVIS, J.L., ANNAN, A. P. (1989) : Ground -penetrating radar for high resolution mapping of soiland rock stratigraphy Geophysical Prospecting 37, 531-551 MEIER, E., HUGGENBERGER, P ( 1992): Georadar zur Kartierung der obersten I 0 bis 20 Meter des Untergrundes Wasser, Energie, Luft, 5 / 6, Verbandszeitung Schweizerischer Wasserwirts chaftsverband, Baden (Schweiz), 111-113 MOLLER, I., STIEFELHAGEN, W. and INTCHI A.R (1995) : Reflections sur les resultats obtenus par l'enregistrement en continues des parametres geophysiques, electromagnetiques (VLF-EM) et magnetiques, pour )'exploration hydrogeologique des aquiferes karstiques (Grotte de Milandre Jura Suisse). Bulletin de la Societe neuchiiteloise de sciences naturelles ROBINSON E. A ( 1989) : Seismic velocity analysis and the convolutional model. Int. Human Resources Dev Corp. Boston, 290 TURBERG P (I 993): Apport de la cartographie radiomagnetotellurique a l'hydrogeologie des milieux fractures Thesis CHYN Neuchatel, Switzerland 132 p. 224 Proceed in gs of the 12'" International Congress of Speleology 1997, Sw i tzerland Volume 2

PAGE 236

Abs tr ac t Contamination of carbonate aquifer and evidence of karst effects in Israel by Abraham J. Melloul and Stuart Wollman Israel Water Commission, the Hydrological Service P.O. Box 6381, IL-91063 Jerusalem ISRAEL Ground water contamination may be enhanced by different contaminants which percolate to the water table from ground surface land-usages. In arid a nd semi arid areas, as in Israel, wetting\drying cycles of ground water recharges can augment the fissuration and cracking processess which increase hydraulic conductivity in soil and the vadose zone. In carbonate aquifers, this leads to the development of karst formations, leading to preferential channels with heightened infiltration rates of the pollutants to the aquifer. The objective of this study is to determine if there is so me relationship between enhanced ground water contamination, and karstic development in a calcareous aquifer, located in an arid/semi-arid region. In the Western Galilee pollution arrived in the ground water from a sewage pond located on calcareous rock sited about 2000m upstream. The fingerprint indicator which characterizes this pollution was the sudden appearance of E. Coli bacteria The second case involves the calcareous aquifer near Jerusalem Here ground water contamination is characterized by relatively high turbidity levels in the course of the initial day of abstraction from a pumping well. This points toward the accumulation of dusty material in preferential channels over the winter in low pumpage period. On a regional scale, the behavior of the mountain Yarkon Taninim calcareous aquifer, subsequent to the unusually high rainfall year of 1991/92 is described A sharp rise in water levels combined with an unexpected increase in ground water salinity is explained by karstic phenomena development mostly in the upper portion of the aquifer. The existence of karst as a major factor enhancing groundwater contamination should be considered for the use of management and operational models of the aquifer. 1 Background This paper aims to illustrate and analyse the relationship between aquifer karstification and enhanced ground water contamination of a carbonate aquifer. The study will also focus on the arrival of pollutants from the ground surface to these aquifers vis-a-vis the ecological conditions affecting its management, and change in climate in these arid/semi-arid areas. This is shown by the analysis of chemical parameter behavior from the mountain carbonate Yarkon Taninim (YT) and Western Galilee (WG) aquifers (Hydro l ogic Report 1996). The Yarkon Taninim aquifer re s presents one of the major s sources of ground water exploitation in Israel. It is bounded in the west by the sea coast and on the east by the Judaean and Samaria Mountains (figure I ). This basin is recharged by precipitation over the permeable limestone rocks of the Judaean Samarian mountain ridge. The Western Galilee carbonate aquifer is a relatively small basin in comparison to the YT aquifer (figure I ). It is replenished from precipitation over the limestone outcrops aquifer which are most l y Cenomanian. Three case studies are presented. The first case relates to a d i sposal pond and a well located down gradient, in the Kefar Yassif area of the western Galilee The presence of E Coli (Esherica coli) indicates the presence of bacterial pollutant s (RCWQC, 1 972) i n carbonate aquifer ground water A second case deals with the appearance of high turbidity values during the first activation of wells in a carbonate aquifer located in the vicinity of Jerusalem The third case analyses the regional response of the YT carbonate aquifer to a climate change (the sharp appearance of a severe winter period). Data are mostly from the 1996 Hydrologic report and internal reports. 2. Results and discussion Figure 2 deal s with ground water contamination by E. Coli and E. Coli faecali bacteria in the Kefar Yassif well in the WG aquifer. This well is located 2 km westward and downgradient of the domestic sewage disposal pond of Yarka. Both facilities are s ited in dolomite rocks c haracterised by their cavities and other karstic characteristic. The disposal pond used for domestic sewage, was active until 1990 It was built in 1984 without any good protective lining on its bottom The K Ya if well became contaminated by E. Coli bacteria identified after a period or only one year after the date of the beginning of operation of the Yarka effluent reservoir. Ana l ysis of this case indicates that conditions encouraging the pollution of the well were : ( I ) the so urce of pollution was located upgradient of the ground water now direction relative to the well ; (2) the anthropogenic pollutant source in this case the Yarka emuent reservoir was not insulated from the aquifer rock below which has high transmis ivity now channels; and (3) the now path does not absorb the microbiological populations nor the various toxics and does not neutralize in any other way its detrimental innuence on the contamination of the water. Thus it is evident that the time duration of the pollutant in the aquifer seems to be very short (around I0m/day). 6 "' Conference on Limestone Hydro logy and Assured Med ia 225

PAGE 237

I I 2ao ..... + .. .. .. I : KEFAR lYASSIF I YAR~PAsf -r200 \ ....... ... .. ( ... ... .. ~_ j i j i l i ,, A INIM 16oj ... .... j.: :~ : f t.A .: .. . ( T) .. . .. j 0 : -~ ,, .. : utcrops-1_ ::, ... : : --. J 1111 : 120! ... .. . ... .. .. .. l-:"-"~I&~ .. ... j i i LEM i l i l oao >....... ..._.EIN KEREM CASE I ~--r-:s.# I I 040: .. i 0001 .. ........ . l r 960r ........ .. . .. .. . .... .. . .... .. 1 l i i : : I 1 .. I j j ... j i l l i l 92~ ~ 60 200" 240 Figure 1: Locatio11 map of Israel: aquifers a11d local case studies The second case presented on Figure 3, deals with ground water contamination by sediments in the Ein Kerem well No. 6 in the YT aquifer in the vicinity of Jerusalem. In this case the measurements for a short period of time indicate high turbidity that appeared in the initial stage of well abstraction. Analyses indicate that the turbidity is due to clay mixed with coarse and fine granular quartz. One supposition is that the turbidity is caused by southerly winds blowing across desert areas, for a few days each year. During the dry seasons these fine materials are deposited and accumulated in cavities and channels, but after the rainy seasons they are conveyed in the carbo n ate ground water aquifer through local karstic systems Another source of this kind of fine material may also arrive with percolation from the sewage water rich in fine material which may be found in the subsurface (Goldenberg et al., 1993). BACTERIAL APPEARANCE K YftSf:FWB..l. 250 g200t----------150 2100+----------- so-----------0 2/85 12/85 8/87 9/88 7/89 12/89 2/90 DATE DE. COLi FAECAL! I Figure 2: E. Coli cha11ges ill the Kefar Yassifl well 19.11.197 40 --------ater eve (m) 20 +-----------------1 \ /l.9~~9.J_~5 2 =1~o=i!li;==;'l12~!1'=1~4~~1~6 =::::itJ Turbidity OU! t-rus Figure 3 : Turbidity changes function of discharge and ground water Levels in Ein Kerem6 Well 226 Proceedings of the 12 "' International Congress of Speleology, 1997 Sw i tzerlandVo l ume 2

PAGE 238

The third case study deals with the extremely hjgh precipitation in the winter of 1991/92, and its regional impact on the groundwater quality of the mountain YT aquifer. This winter resulted in a mean annual average ground water level that was about twice the long term mean. Th en, water levels for th e aq uifer returned to their historical level as in 1950's. Conversely chloride l evels conti nued to increase in mo st part of the aquifers, and their rate of increase reached two or more times the normal year value (aro und 1.5 mg Cl/I per year) as s hown in figure 4 3. Conclusion The local and the large scale effects in the YT aquifer studies give us an indication about the large aquifer recharge and the high infiltration rate values. The results seem to be link ed with karst phenomena which enable contaminants to be trapped and accu mul ated during th e dry season in various cavities and channe l s in the vadose zone, and later released to ground water mrunly after a period of severe rai n or due to sewage or effluent l eakage. On e expla n ation of thi s highly developed Karst effect is the geologica l structure, with karst more developed in the upper (anticline) part of the Cenomanian carbo nat e rocks i n Israel (Me nd el, 19 72). In s u ch arid semi-arid areas, the c h ange in c lim a te and the wetting\drying cycles of ground water recharges ca n a u g ment th e fissuration a nd crackj n g processess in the subsurface which increase hydraulic conductivity in soil an d the vadose zone. Th is phenomrena activate grou ndw ater circ ul ations a nd karst effects, a nd enha nc e groundwater con t amination i n areas where there are pollutant sources. Du e to the i n crease in population, urbanization, density of wells, and high concentrations of anthropogenic activities that may be present in the se areas, danger of contami n ation ca n occur due to the high rate of infiltration of water with pollutants and even other toxic materials arriving from replenishment areas with untreated or partially treated sewage. Thus, the main operational conclusions for preventing the deterioration of ground water quality in such aquifers are to prevent the entrance of undesirable materials to the aquifer system by diminishing po llut ant sources by adequate management of the aq uifer good treatment of waste water, and by augmenting dilution with good quality water. 100 YARKON TANINIM BASIN _J 90 0) E 80 .c C 16 0 0 70 Q) :.;::::; Q) co E I-. 14 0 +-J 60 .c C (1) Q) 0 C > 12.0 0 50 Q) u Q) _J 40 CU 10 0 u s 30 196-4 1966 1968 1970 1972 19H 1976 1978 1980 1982 1984 1986 1988 1990 1992 199-4 Figure 4: Ground water changes and Chloride changes in the central area of the Yarkon Tani11im aquifer References EPA. 1972. Water Qual ity Criteria 1972. R eport of the Committee on Water Quality Criteria 1 972 Washington DC., 5758. GOLDENBERG L.C., MELLOUL A. AND S. El.HANAN I. 1993. M echanism of ground water pollution in the western montanuou s Jude a-Sa mari a aquifer. Assessment and Simulation: In Judea-Samaria R esearch book 20-25pp, (in Hebrew) H YROLOG I CAL SERVICE REPORT. 1 996. Evolution of ground water resources of Israel up to autumn 1995. ISSN 0793109 3 Jerusalem, 210 pp, (in H ebrew) MELLOUL, A. AND L.C. G OLDENBERG 199 3. The rainy winter of 1 991/92 and its influence on the grou nd water quality in the Coastal plain aq uifer of l rael, Water a nd Irrigation Journal, 3 19 : 51-54. (in Hebrew). MENDEL, S. 1966. A conceptual model of karstic erosion by gro und water. R eprint of Bulletin of the I.AS.H, XJC N I 1966, 5-7 6 "' Conference on Limestone Hydrology and Fissured Media 227

PAGE 240

Nitrification des eaux rejetees par l a STEP de La Chaux-de-Fonds Suisse lors de leur passage dans le systeme karstique de la Ronde Par Pa u l-Etien n e Monta n don, Stephane Gogniat, Pierre Rognon 1 et Jean-Jacques Miserez Service de !'Hygiene et de l'Environnement, Serre 23, 2300 La Chaux-de-Fonds, Suisse Resume Les eaux issues de la station d'epuration des eaux usees (STEP) de la ville de La Chaux-de-Fonds (40'000 habitants) Suisse se deversent dans la combe du Valanvron, formant le ruisseau de la Ronde. Ces eaux s'infiltrent progressivement dans un systeme karstique pour dispara1tre totalement apres un parcours aerien de 1500m Les resurgences de ce systeme se situent au niveau du Daubs Les recherches effectuees montrent que le parcours souterrain de la Ronde se deroule essentiellement en milieu a ecoulement libre et que le systeme karstique presente de reelles capacites d'auto epuration, en particulier la nitrification de !'ammonium present dans les effluents de la STEP A partir de !'analyse d un element conservatif (chlorures) on a pu montre que le taux de dilution des eaux de la Ronde entre la STEP et les resurgences varie entre 3 et 6 selon les conditions pluviometriques L'analyse des flux massiques des elements rejetes par la STEP montre que, pendant la periode de mesures le 86 % de !'ammoniu m disparaissant dans le systeme karstique est oxyde en nitrate. Entin !'ammonium, provenant de la STEP et parvenant au Daubs apres passage dans le systeme karstique represente le 50 % des apports de cet element sur le tron~on considere ou encore le 35 % de la quantite totale d'ammonium transportee par le Daubs en aval de ce secteur Summary Treated wastewaters coming from the sewage treatment plant (STP) of the city of La Chaux-de-Fonds ( 40'000 inhabitants) Switzerland flow down the Valanvron valley where they become a small river that progressively infiltrates the karstic system and totally disappears after a course of 1500m. The resurgences of the system are l ocated at the level of the Daubs river. The present research shows that the underground course of the Ronde river essentially flows though an unsaturated zone and that the karstic system has real self purification capacity namely nitrification of ammonium found in the STP effluents. By analyzing a conservative element (chlorides), we were able to estimate that the rate of dilution of the Ronde river betwee n the STP and the resurgences varies from 3 to 6 in function of the level of precipitations Analysis of the mass flow of pollutants rejected by the STP shows that 86 % of the ammonium which disappears in the karstic system is oxidized into nitrates during the period of measurements Finally we found that the ammonium coming from STP that reaches the Daubs river, after passing though the karstic system represents the 50 % of the input of that compound over the considered section or 35 % of the total amount of ammonium carried by the Daubs river downstream from this area Introduction En service depuis 1975 la STEP de La Chaux-de-Fonds (40'000 habitants) collecte toutes les eaux usees de la ville ainsi que les eaux pluviales du synclinal (24'600 m 3 / jour ou 285 1 / s en valeur moyenne annuelle). L'epuration des eaux comprend un traitement mecanique, un traitement biologique par boues activees et une decantation finale Du chlorure ferrique est ajoute a plusieurs etapes du traitement notamment pour la dephosphatation L'azote inorganique provenant de la biodegradatio n de la matiere organique n est pas nitrifie et est done rejete sous la forme d'ammonium Les effluents de la STEP empruntent le lit d'un ancien cours d'eau dans la combe du Valanvron ou ils forment un ruisseau de faible debit appele la Ronde qui s infiltre progressivement dans le sous-sol calcaire par plusieurs pertes diffuses pour disparaHre totalement apres un parcours aerie n de l '000 a 1 '500 m en r egime hydrique normal (figure) Lors de son parcours souterrain la Ronde reJomt le systeme d ecoulement karstique nature!. Apres quoi les eaux de la Ronde gonflees et diluees par les eaux du systeme ressortent par diverses resurgences au niveau du Daubs dont les principales sont celles de la Verrerie et de la Rasse Le temps de transit qui varie en fonction des conditions pluviometriques est de l'ordre de 3 a 7 jours En depit de la rapidite d'ecoulement des eaux. ce systeme karstique presente des capacites interessantes d'auto epuration en particulier l'oxydation en nitrates de !'ammonium rejete avec les effluents de la STEP (Montandon et al. 1 995) Lors de precipitations importantes l e trop plein des eaux auquel s'ajoutent les eaux provenant d'ecoulements peripheriques s'ecoule dans la combe du Valanvron passe par le Cul-des-Pres et rejoint le Daubs par l'etang de Biaufond. Dans le present travai l nous avons evalue les flux massiques des composes azotes inorganiques rejetes par la STEP et reapparaissan t au niveau d u Da u bs II s agissait d'etab l ir la part de ce type de pollution dans le Daubs imputable aux rejets de la STEP 1 Adresse actuelle : Ecole polytechnique federa l e de Lausanne departement de genie rura l Suisse 6 th Conference on Limestone Hydrology and Fissured Media 229

PAGE 241

Hydrogeologie La region de la Chaux-de-Fonds la Ronde qui appartient au domaine du Jura plisse (Bourquin et al. 1946 Bourquin et al. 1968 Favre 1911 ) se situe en domaine karstique La surface totale du bassin versant a ete estime a 44 km 2 (Montandon et al. 1995) Les calcaires plus ou mains fractures sont karstifies et done susceptibles d'etre le siege d'aquiferes. Le dagger est essentiellement compose de calcaires et mamo calcaires plus ou mains karstifies avec quelques niveaux peu permeables fonctionnant comme aquicludes locaux Taus les etages du Malm sont largement karstifies L oxfordien et l'argovien inferieur (couches de Birmensdorf) jouent probablement un role sur les circulations profondes L'argovien superieur qui est en continuite avec les calcaires du reste du Malm constitue l'aquifere le plus important de la region Le niveau de base de ce systeme est le Daubs qui coule dans le synclinal du Pissoux-les Praillats On peut noter que les sources du Dagger se situent au-dessus de celles du Malm en raison du chevauchement de l'anticlinal des cotes du Daubs Finalement, on relevera qu'il est difficile de jauger les sources du Malm, voire Periodes de mesure Factue de dilution Ammonium meme impossible car elles sont proches du Daubs (Rasse) ou arrivent directement dans le lit de la riviere (Verrerie) Methodes Les prelevements ainsi que les analyses des chlorures de )'ammonium, des nitrites et des nitrates ont ete effectuees comme decrit precedemment (Montandon et al. 1995) Le carbone organique total (COT) a ete dose par infrarouge apres mineralisation a chaud avec un appareil Dorman 190 Entin l'azote total a ete mesure par la methode de Kjeldahl modifiee L'azote organique est mineralise a chaud et en milieu acide ; !'ammonium est ensuite dose par colorimetrie Les valeurs donnees sous !'appellation "azote total" correspondent done a !'ensemble de l'azote organique et de !'ammonium. Resultats Afin d'evaluer )'impact de la STEP sur son milieu recepteur nous avons analyse le devenir des polluants rejetes par la STEP dans le systeme karstique Nous avons mesure les substances polluantes qui entrent dans le systeme karstique a la sortie de la STEP et celles qui en ressortent a la resurgence de la Rasse qui a ete consideree comme representative de toutes les autres resurgences de la Ronde STEP La Rasse Nitrites Nitrates Ammonium Nitrates Nitrates dans l'aquifere 1 (mg N/1)2 (mg N/1) 18 au 26 06 1992 3 NE 18,1 2,3 0 06 0 03 0 22 0 14 1 46 I 0 22 0 16 7 4 0 5 17 au 30 03 1993 35 2,8 3 0 16 7 6 1 0 26 0 18 0,41 0,28 1 49 0,46 0,024 0 01 6 75 0 98 04 05 au 11 05 1994 6 7 4 7 5,8 12,5 7 0,14 0 1 1 29 0 86 0,24 0 64 0 009 0,01 2 64 1 1 21.06 au 05 07 1994 6 4 0 4 2 20,7 4,4 0,27 0 48 0,64 0,86 1 27 0,73 0 42 0,39 5,42 1 28 28.04 au 15.06 1995 68 2,8 3,0 17,2 6 2 0,073 0 048 1,17,76 0 68 0 71 0 019 0 013 5 35 1 28 08 11 au 06 12 1995 9 2 8 3 0 17 9,2 0 089 0 083 0 572 0 648 0 975 1 34 0,015 0 009 4 94 1 51 Tableau I : Facteurs de dilution et valeurs moyennes des composes azotes inorganiques a Ja sortie de Ja STEP et a la Rasse 1 Les valeurs donnees correspondent a celles calculees pour une concentration en chlorures de 5 mg/I et respectivement I O mg/I dans les eaux de dilution 2 Mg N / 1 = milligramme d'azote par litre 3 Analyse effectuee a partir d'un echantillon joumalier moyen 4 NE : non effectue 5 Temperature et teneur en oxygene; Rasse : 9 2 0 39 C et I 13 0 24 mg d'oxygene/1; STEP : I 0, 7 I C et 4,9 I I mg d'oxygene/1. 6 Analyse effectuee a partir d'echantillons preleves sur une periode de 4 heures 7 Temperature et teneur en oxygene du 6 au 26 avril; Rasse : 8,6 0 3C et 4, 7 0 5 mg d'oxygene/1; STEP : 8 5 I C et 5,5 0,5 mg d'oxygene/1. 8 Temperature et teneur en oxygene ; Rasse : 9,25 0, 15 C et 4,65 mg d'oxygene/1 ; STEP : 11,35 l 75C et 6,9 0 3 mg d'oxygene / 1. 9 Analyse effectuee a partir d'echantillons preleves sur une periode de 4 heures (STEP) ou d'echantillons ponctuels preleves chaque jour ( Rasse ) 230 Proceed i ngs of the 1 2 '" Internat i onal Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 242

Nitrification de l'ammonium et diminution de la teneur en COT Dans un systeme karstique comme celui de la Ronde on connait le debit des effluents de la STEP mais ii est difficile de jauger les debits des resurgences en raison de leur proximite avec le Daubs ou parce qu'elles se situent directement dans le lit de la riviere. On ne peut done pas estimer si la diminution de la concentration mesuree d'un polluant dans le systeme provient d'une adsorption d'une transformation ou simplement d'une dilution Nous avons precedemment montre !'utilisation d'un element conservatif, les chlorures (Montandon et al. 1995) pour analyser la vitesse d'ecoulement et le taux de dilution des effluents de la STEP dans le systeme karstique Le tableau I presente les valeurs de ce taux calcule pour les differentes campagnes d'analyses ainsi que les valeurs de concentration des composes azotes inorganiques dans les effluents de la STEP et a la resurgence de la Rasse On constate une diminution de la concentration d'arnmonium et une augmentation de la teneur en nitrates, la forme oxydee des composes azotes inorganiques Ces resultats indiquent que ('ammonium rejete avec les effluents de la STEP est oxyde en nitrates au cours du trajet souterrain Cette oxydation ne s'effectue pas aux depens de l'oxygene dissous dans l'eau En effet les teneurs en oxygene a la Rasse se situent generalement a une valeur comprise entre 11 et 40 % de la saturation, alors qu'elles s'elevent a 70 % de la saturation dans les effluents de la STEP (tableau I). La quantile d oxygene dissous correspondant a la difference entre les teneurs mesurees a la sortie de la STEP et a la Rasse n'est pas suffisante pour oxyder ('ammonium etant entendu que la formation d'une mole de nitrates consomme 2 molecules d'oxygene moleculaire, ou 64 mg d'oxygene par milliequivalent de nitrate forme On peut conclure que le trajet souterrain de la Ronde se deroule essentiellement dans des conditions ventilees. Estimations des debits Atin d'aborder )'aspect quantitatif de cette etude, nous avons mene une campagne du 9 novembre au 6 decembre 1995, pendant laquelle nous avons analyse des echantillons preleves dans le Daubs en amont de la Verrerie et dans le lac de Biaufond, ainsi qu'a la Rasse et a la sortie de la STEP Le choix des sites de prelevement dans le Daubs se justitie par le fait que l'on n a pas trouve d'apports signiticatifs provenant de la STEP en amont de la Verrerie et au dela du pant, qui est situe en amont du lac de Biaufond (figure ; Gogniat resultats non publies) Les debits d'eau aux sites de prelevement ont ete determines ou estimes comme decrit ci-dessous Les debits qui sont mesures a la sortie de la STEP nous ont ete communiques par le responsable de !'exploitation; ils ont varie entre 0 15 et 0 92 m 3 /s avec une valeur journaliere moyenne de 0,27 0, I 9 m 3 / s Pour l'amont de la Verrerie le debit du Daubs correspond a la somme du debit de restitution du barrage et du debit turbine a l'usine electrique du Chatelot, qui sont situes en amont du point de prelevement, respectivement a 8 et a 4 km Pendant la periode de mesure les debits, qui nous ont ete fournis par le responsable de l'usine du Chatelot ont varie entre 0 4 et 43 m 3 /s, avec une valeur moyenne de 13 6 13,6 m 3 / s Les fluctuations importantes des debits sont dues a ('exploitation de l'usine hydro-electrique. On estime que le debit est le meme a la Verrerie qu'a l'usine du Chatelot, car ii n'y a ni apport significatif ni perte d'eau sur le troni,:on separant ces points Le debit moyen pour la periode consideree a ete estime a l aval du lac de Biaufond par !'analyse des chlorures en supposant que ('augmentation de la concentration des chlorures dans l'eau du Daubs entre l'amont de la Verrerie et le lac de Biaufond est due uniquement a l'apport de cet element provenant de la STEP de La Chaux-de-Fonds (Rognon, 1996) On admet que les apports en chlorures etrangers a la STEP sont peu importants. Le debit obtenu est de 18 m 3 /s, soit une valeur 1 3 fois plus elevee que le debit determine a l'amont de la Verrerie. Cette valeur est raisonnable, car Matthey ( 1986) a montre une augmentation du debit moyen du Daubs pour le mois de novembre de 1 5 entre le lac des Brenets situe a environ 12 km en amont de la Verrerie et la Combe des Sarrasins qui se trouve a 4 km en aval du lac de Biaufond D'autre part les valeurs des precipitations pour le mois de novembre 1995 se situaient a une valeur proche de la moyenne, c'est-a-dire a 127 mm ou 111 % de la valeur moyenne. Entin le debit moyen de la Rasse pour la periode consideree a ete evalue a 250 1/s, une valeur comprise entre le debit d'etiage (45 1/s) et le debit en hautes eaux (300 1/s; Matthey 1990). Determination des flux massiques des composes azotes inorganiques, du COT et des chlorures Les valeurs moyennes de concentration des parametres analyses tigurent dans le tableau 2 et les flux massiques calcules a partir de ces valeurs et des debits d'eau sont donnes dans le tableau 3 On peut deduire du flux massique des chlorures que le 33 % des eaux provenant de la STEP ressort a la Rasse pendant la periode de mesures, le 67 % restant reapparaissant principalement a la Verrerie et probablement par des apports diffus directement dans le Daubs (Gogniat, resultats non publies) On constate encore qu'en moyenne 430 kg d'ammonium ont ete deverses quotidiennement dans le systeme karstique et que seuls 19 kg reapparaissent a la Rasse; compte tenu de la proportion des eaux de la STEP parvenant au Daubs par la Rasse on peut calculer les apports totaux en ammonium arrivant au Doubs a 58 kg/j (tableau 4), soit 13 % des quantiles deversees par la STEP Par contre le flux massique des nitrates augmente fortement de la STEP a la Rasse (tableau 3) La comparaison des flux massiques de ('ammonium et des nitrates montre que le 86 % de ('ammonium disparaissant dans le systeme karstique est oxyde en nitrates ce qui contirme le pouvoir eleve d'auto-epuration. On peut noter, enfin, la diminution du flux massique des nitrites entre La Verrerie et le lac de Biaufond (tableau 3) malgre l'apport du a la STEP Ce resultat suggere qu'une partie relativement importante des nitrites presents dans le Daubs est oxydee en nitrates Part des polluants provenant de la STEP transportes par le Doubs a Biaufond Le tableau 4 montre les apports totaux de la STEP parvenant dans le Daubs On constate que les apports en ammonium et en nitrates representent environ la moitie des arrivees de ces elements dans le troni,:on La Verrerie Lac de Biaufond, le reste provenant principalement des maisons isolees et de l'activite agricole. Quant aux apports de COT issus de la STEP, its correspondent au 15 % des arrivees de ces composes dans le troni,:on considere Entin les quantiles d ammonium et de COT provenant de la STEP representent le 35% et respectivement le 4% de la masse de ces polluants transportee par le Daubs en aval de Biaufond 6 th Conference on Limestone Hydrology and Fissured Med i a 231

PAGE 243

Parametres Effiuents de la STEP Amont Verrerie Rasse Aval Biaufond Ammonium [mgN / 1) 18 15 7 0,044 0,01 0 88 1 ,3 0 103 0 07 Nitrates [mgN / 1) 0 5 0 5 1 4,2 4 9 1 6 1 5 0 3 Nitrites [mgN / 1) 0 1 ,1 0 024 0,009 0 016 0,009 0 015 0 008 Azote total [mgN/1) NE 0 78 0 5 1 1 ,6 I ,18 Chlorures [mg/I] 150 ,4 53 5 5 3 1 ,5 54 5 26 6 3 1 8 COT [mg/I) 12 1 2 4,6 1 4 4 2 1 1 4,6 1 2 Tableau 2 : Valeurs moyennes des teneurs en composes azotes inorganiques, en chlorures et en COT dans le Doubs, a la Rasse et dans les effluents de la STEP (9 nov. 6 dee. 1995) Parametres Effiuents de la STEP Amont Verrerie Rasse Aval Biaufond Ammonium [kgN / j] 430 52 19 160 Nitrates [kgN / j] 12 1649 106 2333 Nitrites [kgN / j) 2,4 28 0,34 23 Azote total [kgN / j] NE 929 23,7 1571 Chlorures [kg/j) 3561 6241 1177 9798 COT [kg/j) 286 5429 90 7 7092 Tableau 3 : Flux massiques des composes azotes inorganiques, des chlorures et du COT dans le Doubs, a la Rasse et dans Les effluents de la STEP (9 nov. 6 dee. 1995) Apports totaux parvenant au Doubs sur Apports provenant de la Proportion imputable a la Parametres le tron~on Amont STEP 2 STEP 3 Verrerie Aval Biaufond ....................... .......................... (kg/j) (kg/j) (%) Ammonium 4 102 58 54 Nitrates 4 684 321 48 Nitrites 4 / s 1 2 / s Azote total 4 642 73 11 4 Chlorure 3557 3623 JOO COT 1843 277 15 Tableau 4 : Apports dans le Doubs en composes azotes inorganiques en chlorure et en COT provenant de la STEP, apres passage dans le systeme karstique 1 NE : non effectue 2 Ces apports ont ete calcules en sachant qu'environ 30 % de ceux-ci rejoignent le Doubs par la Rasse 3 La part imputable a la STEP correspond au rapport des apports totaux de !'element considere provenant de la STEP sur les quantites de element transportees par le Doubs en aval de Biaufond 4 kgN /j 5 Le flux massique diminue malgre l'apport du a la STEP 232 Proceedings of the 12'" International Congress of Speleology 1997 Switzerland Vo l ume 2

PAGE 244

Discussion Nous avons montre que !'ammonium rejete dans les effluents de la STEP est oxyde en nitrate s au cours du trajet souterrain de la Ronde Les resultats indiquent que cette nitrification ne s'effectue pas aux depens de l'oxygene dissous et qu'elle a lieu pendant toute l annee On peut conclure que le systeme karstique est ventile et que le parcours souterrain se deroule essentiellement en milieu a ecoulement libre Dans un tel ecoulement !'elimination des polluants peut inclure des processus plutot physiques tels 'a dsorption ou le fractionnement d organismes vivants Des etudes precedentes ont revele une baisse importante de la concentration en bacteries fecales et en bacteriophages de E. coli dans ce sys teme karstique (Blant et Stettler 1982 ; Montandon et al. 1995 ), le taux de dilution n'etant pas suffisant pour expliquer ces pertes En ce qui conceme les bacteriophages, les analyses ont demontre que le taux d'inactivation des particules infectieuses depend du type de bacteriophage analyse, suggerant une inactivation d'ordre mecanique On admet generalement que la retention des substances polluantes est faible en milieu karstique (Baylet et al. 1987 ; Caumartin 1975 ; Miserez et al. 1976) en raison de la rapidite des ecoulements d'eau et de l'insuffisance des couches argileuses qui fixent les polluants et permettent leur destruction par des micro-organismes Or les resultats presentes montrent que le sys teme karstique de la Ronde possede des capacites d'auto epuration qui consistent en une oxydation et non une simple dilution des polluants par des eaux d'infiltration. Une etude recente (Montandon et al. 1995) a d'ailleurs mis en evidence la biodegradation des hydrocarbures aromatiques de la serie du benzene dans ce systeme karstique En effet une bacterie capable de degrader ces composes en condition de laboratoire a ete isolee a partir d'un echantillon d'eau preleve a la Rasse Les capacites d'auto-epuration de ce systeme se sont probablement developpees de maniere progressive depuis la mise en service de la STEP en 1975 Avant la construction de celle-ci les eaux usees, qui etaient deversees dans des pertes situees 1 5 km en amont de la STEP (figure) et ressortaient aux resurgences de la Verrerie et de la Rasse etaient contaminees par des teneurs elevees en ammonium (Schardt 1911 Schassmann 1965) suggerant que le systeme karstique n'etait pas capable de nitrifier ('ammonium a cause de la charge trop elevee en matiere organique des eaux usees non epurees. La presence d'une STEP done d'une source reguliere en eaux usees epurees et riche en micro-organismes de differents types physiologiques a probablement contribue au developpeme nt de conditions favorables a la retention et a l'ox y dation des polluants deverses dans le systeme On a, enfin, pu estimer !'impact de la STEP de La Chaux-de Fond s s ur le Daubs par !'analyse des flux massiques des polluant s. La part des polluants transportes par le Daubs en aval de Biaufond et imputable a la STEP est vraisemblablement inferieure aux va leurs presentees dans le tableau 4 En effet, nous avons neglige dans notre calcul du debit du Daubs les arrivees en chlorures etrangers a la STEP qui parviennent dans le Daubs avec les eaux du bassin versant. Par consequent la valeur de debit utilisee pour calculer les flux massiques a Biaufond correspond a une valeur moyenne minimale (voir section Estimation des debits ), done a une sous-estimation des flux massiques a cet endroit. En conclusion, cette recherche represente une premiere approche dans l 'e tablissement de !'impact de la STEP actuelle sur le Doubs. A terme nous pourrons determiner l'effet sur le milieu recepteur des ameliorations de traitement notamment la nitrification qui interviendront avec la renovation de la STEP actuelle. Remerciements Les auteurs remercient MM J Baumgartner responsable de l'usine hydroelectrique du Chiitelot, et R Boillat, responsable de la STEP de La Chaux-de-Fonds d avoir transmis les donnees concemant les debits du Doubs et a la sortie de la STEP Mme S. Daval MM J.-M Erard et P Erard, collaborateurs et stagiaire au Service de !'hygiene et de l'environnement de leur excellente collaboration tout au long de ce travail ainsi que Mlle D Arnoux pour la mise en page du manuscrit. References BAYLET R ., JOSEPH C ., RODIER C ., SOUUE M. SINEGRE f. : Variation de la qualite bacteriologique des eaux karstiques au cours d'un episode de crue : source du Lez (France) Novembre 1982 Bull. du Centre d'hydrogeologie de Neuchiitel 7 : 253-264 1987 BLANT J.D ., STETTLER R. : Survie des bacteries indicatrices de pollution fecale dans un cours d'eau Bull. Soc. Neuch Sci. Nat. I 05 89-105 1982 BOURQUIN P., SUTER H ., ET FALLOT P : Carte geologique de la Suisse feuille 15 Biaufond Les Bois La Ferriere St.-lmier 1946 BOURQUIN P ., BUXTORF R. FREI E. LUTHI E ., MUHLETHALER C ., RYNIKER K ., SUTER H .: Carte geologique I : 25 000 de !'atlas geologique de la Suisse feuille 5 I Val-de-Ruz, 1968 CAUMARTIN V .: Les mecanismes de recyclage des eaux en milieu karstique Actes du 5;me Congr. nat. de speleologie septembre 1974 Interlaken R Gigon (Ed.) Soc Suisse de Speleolog ie, Neuchiitel : 79-85 1975 FAVRE J .: Description geologique des environs du Lacie et de la Chaux-de-Fonds Eclogae geol. Helv ., XI : 369-475 1911 MATTHEY B : Les ressources en eau du canton de Neuchiitel dans le cadre de l'amenagement du territoire Matthey B (Ed ), Montezillon (NE) Suisse 1989. MATTHEY B : Eaux souterraines. Dans : Centre d'identification de traitement et de recyclage des dechets speciaux CITRED protection de l'environnement, etude de faisabilite ATESA (e d) Republique et canton de Neuchiitel, dep trav pub!. novembre 1990 MISEREZ J.J SIMEON! G.P ., BERNASCONI R ., BRANDT C. DEPALLENS 0. DOMONT G ., KELLENBERER Y.S. MOESCHLER 0. VOISIN G : Essai sur le pouvoir d'auto-epuration des eaux karstiques Stalactite 26/1 13-28 1976 MONTANDON P E ., MAGES J.F. MISEREZ J.F : Etude de l'ecoulement du systeme karstique Exemple de la vallee de la Ronde (NE Suisse) Bull. du Centre d'hydrogeologie de Neuchiitel 14 : 177-198 1995 ROGNON P : Impact sur le milieu recepteur d'une station d 'e puration dont les effluents transitent par un systeme karstique Travail de diplome, Ecole polytechnique federale de Lausanne Suisse 1996 SCHARDT H : Le cours souterrain de la Ronde. Bull. Soc Neuch Sci Nat. 37 : 3-18 1911 SCHMASSMANN H : Ursachen, Folgen und Bekampfung von Grundwasserverunreinigungen in der Schweiz In Intemationales Jahrbuch Chemische Industrie Druck und Verlag Vogt-Schild AG Solothum Schweiz p 3-16 1965 5 th Conference on Limestone Hydro logy and Rssured Media 233

PAGE 245

D Qualcmairc / Faille. chc / vauchcmcnt D Tcrtiairc Malm upcricur Malm infcricur Dogg r / (ob crvc(c}. uppo c(c)) Resurgence de la Ronde Source captagc Figure : Carte de situation avec esquisse tectonique, coupe geologique schematique zones d'infiltration ancienne (avant 19 7 5) et actuel/e de la Ronde, resurgences principales et autres sources ou captages : I: Chez Bonaparte ; 2 : Moulinet; 3 : Tunnels ; 4 : Dames; 5 : Grande et Petite Rondes ; 6: Bas de Combe; 7: Fief Les points de prelevement dans le Daubs sont indiques par un X (Modifie d'apres Bourquin et al 1946) 234 Proceedings of the 12 '" International Congress of Speleology, 1997 Switzerland Volume 2

PAGE 246

Variability of er ion in water inflows to karst-fissured limestone ,,Zakrz6wek" quarry (Cracow, Poland) by Jacek Mot yka and Adam Postawa University of Mining and Metallurgy PL 30065, Cracow Mickiewicza 30., Poland Abstract Chemical composition of water outflowing in the ,,2.akn6wek" quarry (Cracow, SW Poland) developed in fractured and karstified Upper Jurassic limestones is controlled by infiltration of polluted water from the Vistula River and by infiltration of meteoric water into elastic rocks in the neighbourhood oflimestone complex. The river water belong to Cl-Na type with IDS up to 2 5 g/L while the water from inflows in the quarry have 0 6 2.0 g/L IDS Highly mineralised ones belong to Cl-Na type. With the decreasing IDS percentage of SO4, Ca, Mg and HCO 3 increases which seems to be a result of the interference of various processes: dilution of polluted river water, leaching of aquifer rocks and ion exchange (Na-Ca). Concentration of er ion varies in wide range Significant variability is observed in both time and space which will be described in present paper. Authors attempt to explain reasons of this variability 1. Introduction Presently inactive ,,2.akn6wek" quarry is situated in a southern suburb of Cracow, in the area of small horst structure known as ,,The Twardowski Rocks" The structure comprises Upper Jurassic limestones which had been mined from 1953 till 1991 and utilised as a raw material for soda production in the neighbouring ,, Solvay chemical plant. The deepest exploitation level reached the depth of about 60 meters below surface, i.e about 25 meters below the water level of Vistula River which flows 600-900 meters west from the western comer of the quarry The hydrodynamic conditions facilitate the infiltration of river water into the pit. Jn a neighbourhood of the ,2.akrz6wek quarry there are located a few abandoned small open pits with their bottoms above the water level in the river. These pits played certain role in recharging of inflows in the quarry The Vistula river in Cracow is highly polluted, mostly by mineralised water originating from dewatering of coal mines in the Upper Silesian Coal Basin Brines released to the river ma y reach up to 370 g/L IDS (R6zkowski, Rudzinska, 1983) Significant pollution is provided also by agriculture (fertilisers, pesticides and insecticides) by sewages and industrial wastes Localisation of the ,,lakrz6wek" quarry in the neighbourhood of the river bed enabled the multidirectional hydrogeological studies. One of the most interesting problems of both the scientific and practical importance is variability of chloride ion concentrations in water flowing into the quarry This problem is described in the present paper. 2. Hydrogeological background The main water level of tl1e Zakrz6wek horst is located in fractured and karstified Upper Jurassic (Malm) linlestones The fault-bordered and uplifted limestone complex is surrow1ded b y Quaternary sediments (Fig. 1) which include water-bearing alluvia sands and gravels . TI1ese sedin1ents are in direct hydraulic connection with the Vistula river (Fig I). Average hydraulic conductivity of Quaternary sands and gravel s is k=2 8-10-4m/s (Motyka et al., 1995) Along the eastern and northern margins of the Zakrz6wek horst Quaternary sands are thin and dry. Further eastand northward these sediments grade into poorly permeable clays alluvial muds and deluvial pelites Along tl1e souU1em margins of the horst Quaternary sands and gravels are thicker and water-bearing and contact laterally witl1 the Upper Jurassic limestones (Fig. I) Fig. t. Geological map of the Zakrzowek horst Quaternary: 1 embankments and dumps, 2 sands, 3 alluvia, deluvia, loess. Tertiary : 4 clays. Cretaceous: 5 marls, marly limestones. Jurassic (Malm): 6 massive and bedded limestones 7 faults, 8 boundary of the Zakrzowek Quarry, 9 cross-section, 10 sampling points. In the limestones tree systems of water flow showing ditTerent hydrogeological propertie s can be marked out which are : pore space fissure s, caverns Pore space Chouquette & Pray ( 1970) sense, comprises all empty spaces present in a small rock san1ple regardless of their genesis Open matrix porosity, determinated by laboratory testing of 4 18 samples ranges from 0 008 to 0.166. The average poro s it y is p 0 = 0 076 (with standard deviation s = 0 041 ) 6 1 11 Conference on Limestone Hydrology and Fissured Med i a 235

PAGE 247

Fractures and interbedding planes are the main conduits of underground water now Wide fractures particularly favourable for water transfer occur in weathered parts of the horst and in tectonic zones On the walls of the quarry many are high-angle fractures (70 90 )in few meters spacing can be seen Probably they continue in long distances Dominating fractures belong to the two systems: NE-SW and NW-SE Limestone beds are usually 1-2 m thick and dip gently (at some degrees) to the east i e. from Vistula river to the quarry Main now system constituted of fractures and interbedding planes is supplemented by different karst features : anastomoses, caverns caves and large wells and cones filled up with elastic sediments (sands, clayey sands and clays) In the area of Zakrz6wek horst was discovered 9 caves with complicated systems of passages and chambers. The caves are developed on fissures and interbedding planes network. Main directions of passages and smaller caverns are in general confonnable to fissures strike Dominating are NW-SE and NE-SW directions. Minor role play W-E and NE-SW directions characteristic for smaller caverns 3. Main factors forming the chemical composition of water inflows to the ,,Zakrz6wek" quarry Chemical composition of water inflows to the ,,.Zakrz6wek" quarry is formed by two main factors which are : filtration of salt water from Vistula river and infiltration of meteoric water in the area of Zakrz6wek horst. The Vistula river is since many years highly polluted with mineral and organic substances. As was mentioned above the main reason of this pollution is dewatering of coal mines in the Upper Silesian Coal Basin For this reason Vistula river's water belongs to Cl-Na chemical type The total dissolved solids (IDS) ranges from 1.5 to 3.4 g/L and total hardness (IB) 8.4, 15 mval/L. Dominating anion is er and its share varies from 75 to 90 % mval. Depending on chloride concentration in brines released to the river and amount of precipitation in its catchment area the concentration of er ion in river water varies in a very wide range Detailed long tenn investigations of chemical composition of atmospheric precipitation in Cracow and its vicinity were carried out by Turz.anski (1991 ). Despite methodological problems wiU1 sample collecting and interpretation of chemical analyses results, he established that average IDS of meteoric water is about 50 mg/L Industrial pollution of the air in Cracow caused specific chemical composition of precipitation in this area Dominating anion is sol with average content of 60+ 70 % mval. Dominating cation is Ca (60+70 % mval) Average concentration ofCr ion is 6.5 mg/L During intensive dewatering of ,,.Zakrz6wek quarry numerous water inflows with different yields was observed Total number of the noted inflows was some about 60 The most intensive were the ones in northern part of the quarry (up to 7 Lis) 1l1ey forn1ed the continuous zone of man y, nearly located inflows flowing out of the single interbeddi.ng plane situated 8 meters above the open pit floor. Other inflows were observed in different parts of the quarry at higl1s from one to few meters above the bottom of the pit and the, have yields not exceeding the value of one litre per minute In the northern and north-eastern parts of the quarry Cl-Na Cl-Na-Ca and Cl-S04-Na-Ca water types dominated Water from No 16 inflow which showed the highest mineralisation (IDS) be l onged to Cl-Na type With decreasing IDS the contents of Ca and so ~ followed by HC03 increased In some inflows (No 3 and 12) of lower mineralisation Mg ion appeared In the western and southern parts of the pit where mineralisation varied from 0 6 to 0.9 g/L the water revealed polyionic character Most c01mnon were fiveand six-ion compositions (Ca-Mg-HC0 3 -S04-Cl or Ca-Mg-Na-HC03-Cl-S04) The appearance of magnesiwn ion in this part of quarry increased 4. Spatial variability of chloride concentration. Concentration of er ion determinated in water from inflows in ,.Zakrz6wek quarry showed very strong spatial variability In north-eastern part of the quarry chloride concentration was ,,mosaic-type" (Fig 2) Percentage difference of er concentration in adjacent inflows reached up to 300% In eastern part of the open pit chloride content was lower and more equalised ranging from 60 to 90 mg/L Concentrations increased south-wards of the quarry where reached 100+260 mg/L ~2 ~3 ., ~5 11111 s @1 GJ 8 O 50m ..____, Fig. 2. Chloride contents (in mg/L): 1 up to 100; 210()-;-200; 3 20()-;-300; 4 30()-i-500; 5 50()-i-700; 6 over 700, 7 boundary of the Zakrzowek Quarry, 8 sampling points 5. Variation of chloride content in time. Investigations of variability of chloride content in water from i.nl1ows in ,, Zakrz6wek quarry in time were carried oul in year s 1990 + 1992 At first samples were collected from 11 in.flows From January of 1991 number of sampling points increased to I 8 and from October to 20 Additionally samples from Vistula river and small artificial reservoir at Uu! surface in souU1em vicinit y of U1e quarry were collected 236 Proceed i ngs of the 12 th International Congress of Speleology 1997 Sw i tzerland Volume 2

PAGE 248

Superposition of complicated hydraulic system and two different input functions (varying chloride concentration in Vistula river and meteoric water) strongly influenced the variability of er concentrations in investigated inflows. Actually, in each single monitored inflow chloride content varied in different way however a few types of this variability can be pointed out. 500 ____________________ 500 400 300 200 100 Group B inflows 2, 3 and 19 (Fig. 38) Amplitudes of long-tem1 variations of chloride concentration in water from these inflows were relatively small but some temporary, very high variations were noted Rapid decreases of er concentrations appeared afier periods of downpours or snow melting I such periods intensive dilution of water recharging described inflows was the main factor determining their chemical composition 500 ____________________ B 450 400 350 300'-~--'---___J'--~-...J.._----L----'-~~_J 01 01 90 01 07 90 01 01 91 01 07