Journal of cave and karst studies

Citation
Journal of cave and karst studies

Material Information

Title:
Journal of cave and karst studies
Series Title:
Journal of Cave & Karst Studies
Alternate Title:
Continues NSS bulletin (OCLC: 2087737)
Creator:
National Speleological Society
Publisher:
National Speleological Society
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Geology ( local )
Genre:
serial ( sobekcm )
Location:
United States

Notes

Restriction:
Open Access - Permission by Publisher
Original Version:
Vol. 76, no. 3 (2014)

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
K26-02068 ( USFLDC DOI )
k26.2068 ( USFLDC Handle )
20950 ( karstportal - original NodeID )
0146-9517 ( ISSN )

USFLDC Membership

Aggregations:
Added automatically
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item has the following downloads:


Full Text

PAGE 3

INTEGRATEDGEOPHYSICALANDGEOLOGICAL INVESTIGATIONSOFKARSTSTRUCTURESIN KOMBEREK,SLOVAKIARENE PUTIS KA1*,DAVIDKUS NIRA K1,IVANDOSTA L1,ALEXANDERLAC NY 2,ANDREJMOJZES 1,JOZEFHO K2, ROMANPAS TEKA1,MARTINKRAJN A K1,ANDMARIA NBOS ANSKY 1Abstract: Acomplexofgeophysicalmethodswereusedtoinvestigateasmallkarstarea aimedattheproductionofdetailedgeologicalmapping,toconfirmgeological localizationofknownsinkholes,andtofindpossiblecontinuationsofcavesandvoids belowthesurface.Thedipoleelectromagneticprofilingandradiometricmapping(the gamma-rayspectrometrymethod)wereappliedtodeterminethespatialdistributionof hardcarbonaterocksandweatheredvalley-fillsediments.Detailedhigh-definition magnetometrywascarriedoutatselectedsitesinthestudiedregionwiththeaimof distinguishingbetweensinkholesandman-madelime-kilns,pitswherelimestonewas heatedandtransformedintolime.Themicrogravityandtheelectrical-resistivity tomography(ERT)methodswereusedtocreatehigh-resolutionimagesofthe undergroundcave.TheresultsofERTandthegeologicalsurveywereusedasan initialmodelforgravitymodeling.Subsurfacecavitiesofvarioussizesarecontrasting geophysicalobjects,andtheelectricalresistivitycanrangefromveryconductiveto relativelyresistivedependingonthecompositionofthefillingmaterials.The interpretationofresistivitypropertiesisnotalwaysstraightforward.Wemust distinguishair-filled(high-resistivity)andloamywater-filled(low-resistivity)cavities andfractures.Thecombinedgeophysicalmethodologypermitsustodetermineamore accuratenear-surfacegeologicalmodel,inourcasetheparallelinterpretationofastrong conductiveanomalyintheERTinversionandapredominantdensitydecreaseinthe gravitymodellingyieldthepresenceofcavitiesatdepthsapproximatelyof50to60m belowthesurface.INTRODUCTIONExploredcavesareonlyalimitedportionofthose actuallyexistingunderground(White,1990;Fordand Williams,2007).Toobtaininformationaboutthesehidden caves,orunknownorinaccessiblecontinuationsofknown caves,wemustuseindirectmethods(PariseandLollino, 2011;Margiottaetal.,2012;Pepeetal.,2013).Nearsurfacegeophysicalmethodshaverecentlybecomean importanttoolinkarst-cavesresearch.Theideabehind mostofthesegeophysicalmethodsisamaterialpropertyof thevoidthatissignificantlydifferentfromthesurrounding hostrockandthusmakesamaterialcontrast.Thismaterial contrastcanthenbedetectedusingaspecificgeophysical technique(e.g.,Butler,1984;Gibsonetal.,2004;El-Qady etal.,2005;DobeckiandUpchurch,2006;Nyquistetal., 2007;Mochalesetal.,2008;Margiottaetal.,2012;Putis ka etal.,2012a).Amongsomeofthemostfrequentlyused geophysicaltechniques,electrical-resistivitytomography andmicrogravitycanbementioned,butadditional methodscanprovideveryusefulinformationaswell. Thestudiedregionislocatedinthenortheastern portionoftheMaleKarpatyMts.,inthewesternpartof Slovakia,andbelongstotheKuchyn a-Oresanykarst.The northeasternpartofKuchyn a-Oresanykarstisrepresented bytheKomberekkarst,whichisanotlarge,about1km2, butinterestingkarstplain,wheretwocaves,theStrapek CaveandtheZavrtovaPriepast Cave,werediscovered. Besidesthecaves,morethanseventyterraindepressions werefoundinthearea(Fig.1),butsomeofthemarenot karstlandforms,astheyareman-madelimekilns.Thelime kilnswerecreatedduringthefifteenthandsixteenth centuriestoproducequicklimethroughthecalcinationof limestone.Thekilnsareupto3mindiameterand1-mdeep, whichisverysimilartosomeofthesmallernaturalsinkholes inthearea.Therefore,wetriedtofindamethodtoreliably distinguishthem.Severaldepressionsarefilledwithmud andwater,whichmakesitdifficulttodistinguishwhether theiroriginisnaturalornot.Thetermmudholesisusedfor thesestructuresinthefollowingtext. Avarietyofgeophysicaltechniquescanbeusedto detectthepresenceofcavesandvoidsbelowthesurface (Cardarellietal.,2010;Gambettaetal.,2011;Kaufmann etal.,2011;Lac nyetal.,2012;AndrejandUros,2012). *CorrespondingAuthor:putiska@fns.uniba.sk1DepartmentofAppliedandEnvironmentalGeophysics,FacultyofNatural Sciences,Mlynska dolinaG,84215Bratislava,SlovakRepublic2DepartmentofGeologyandPaleontology,FacultyofNaturalSciences,Comenius University,MlynskadolinaG,84215Bratislava,SlovakRepublicR.Putiska,D.Kusnirak,I.Dostal,A.Lac ny,A.Mojzes,J.Hok,R.Pasteka,M.Krajn ak,andM.BosanskyIntegratedgeophysical andgeologicalinvestigationsofkarststructuresinKomberek,Slovakia. JournalofCaveandKarstStudies, v.76,no.3,p.155. DOI:10.4311/2013ES0112JournalofCaveandKarstStudies, December2014 N 155

PAGE 4

Eacharebasedonaphysicalcontrastbetweenacaveand thesurroundingrocks.Karstfeaturesareprevalent throughoutthestudyarea.Sinkholesdevelopbyacluster ofinter-relatedprocesses,includingbedrockdissolution, rockcollapse,soilsapping,andsoilcollapse.Anyoneor moreoftheseprocessescancreateasinkhole.Thebasic classification(Walthmanetal.,2005)hassixmaintypesof sinkholes,solution,collapse,caprock,dropout,suffusion, andburied,thatrelatetothedominantprocessbehindthe developmentofthesinkhole.Thedominanttypeof sinkholeswithintheKomberekkarstareaissolution, formedwheresurfacewaterandsoilwaterdissolve bedrocknearthesurfaceasitflowstowardspointswhere itcansinkintofissuredorcavernousground.As mentionedabove,theprimaryfocusofourgeophysical researchwastomapthesiteandtodetectgeological anomalies.Onthebasisoftheseresultstheplanfordeeper investigationswassetup.Theaimofthesurveywasto detectkarstfeatureslikesinkholesorfracturedzonesthat couldcommunicatewiththeundergroundnetwork.New cavesorextensionsofknowngallerieswereexpectedtobe foundtoo.Thelocationofthegeophysicalsurveyisshown inFigure1.GEOLOGICALSETTINGTheKomberekkarstarea,aswellastheMaleKarpaty Mts.,areintegralpartsoftheWesternCarpathians orogenicbelt.TheMaleKarpatyMts.geologicalstructure consistsofseveraltectonicunits(Polaketal.,2011).The Tatricumtectonicunitisthemostautochthonousunit comprisingthePaleozoiccrystallinebasementandthe Mesozoicsedimentarycover.TheFatricumandHronicum tectonicunitsbelongtothenappestructuresoftheWestern CarpathianstectonicallyindividualizedduringtheAlpine orogenyduringtheCretaceousperiod.TheTatricumunit inthestudiedareacontainsonlytheuppermostsynorogenicflyschmember,clayeyshales,andturbiditicsandstonesofAlbian-Cenomanianage.AnUpperCretaceous thrustplaneseparatesunderlyingTatricfromFatricunitin itshangingwall.TheKomberekarea(Fig.2)isbuiltup almostexclusivelybytheFatricunitsMiddletoUpper TriassicandJurassicmembers.MiddleTriassicdark-grey toblackthick-beddedGutensteintypelimestoneisthe prevailingrocktypeinthisarea,anditisalsothe lowermostpartofFatricunit.Thetectoniccontact betweentheTatricumandtheflyschsedimentsofthe Figure1.Topographicmapofthestudyareashowingthegeophysicalsurveysitesandkarststructures.ERTiselectricalresistivitytomography,MGismicrogravity,andDEMPissurfaceelectromagneticconductivity.CoordinatesinFigures1 areUTMzone33N.INTEGRATEDGEOPHYSICALANDGEOLOGICALINVESTIGATIONSOFKARSTSTRUCTURESINKOMBEREK,SLOVAKIA156 N JournalofCaveandKarstStudies, December2014

PAGE 5

Fatricumunitsislinkedtotherauhwackes.Rauhwackes (cornieulesorcargneules)arebrecciaswithacalcareous matrixandmainlydolomiticcomponentsthatweatherto formcavernousrocks.Theyareveryoftenassociatedwith tectoniccontacts.Theoriginofrauhwackesisstill controversial,buthasbeenattributedtotheweathering andalterationofdolomite-bearingevaporites,thetectonizationofdolomites,orotherprocesses(Krauter,1971; Schaad,1995).Itissupposedthattherauhwackesare tectonicallyderivedfromtheGutensteinlimestones.The GutensteinlimestonesshouldbeoverlayedbyRamsau dolomites,butinthislocalitygreythick-beddeddolomitizedlimestonescropout.Variegatedclayeyshales,quartzy sandstones,andquartzstones,andalsocavernousgrey dolomites(rauhwackes)ofUpperTriassicagebelongingto CarpathianKeupermemberoverlietheGutensteinand Ramsaucarbonaticcomplex.Inthenorthwesternpartof theKomberekareathereareJurassicgreycrinoidalcherty andpinknodularlimestones.TheKomberekareais disruptedbynorthwest-southeastorientednormalfaults activeduringtheNeogene(Polaketal.,2011).The sinkholesarelocalizedalongdistinctlinessituatedalong thelitologicalandtectonicdiscontinuites.Themain lithologicaldiscontinuityisbetweentheCarpathianKeuperFormationandtheGuttensteinlimestones.The northwest-southeasttectoniclineisthereasonforthe occurrenceofthesinkholesarraythatfollowsit.Anedited geologicalmapisshowninFigure2;becauseofthevery flatmorphologyofthestudiedareaandthescarcityof outcrops,geologicalmappingisverydifficult.GEOPHYSICALMETHODSAcomplexofgeophysicalmethodswasdesignedto gatheralargeamountofgeophysicalinformationfromthe studyarea.Themethodscanbedividedbypurposeinto twoparts.Thefirstpartwasaninitialfieldinvestigationof thewholearea.Inthiscase,electromagnetic-conductivity, magnetometry,andgamma-rayspectrometrywereusedto mapthesiteandtodistinguishnear-surfacegeological settingsinthearea.Theinformationobtainedwasusedto characterizethegeologicalsettingoftheKomberekarea (Fig.2).Thesecondpartofthegeophysicalfieldwork includedelectrical-resistivitytomography(ERT)andmicrogravitysurveyonaselectedprofileacrossthekarstarea toobtaininformationaboutunknownorinaccessible continuationsofknowncaves.Theprofilewasbasedon theresultsfromthegeophysicalworkdoneduringthefirst partoftheinvestigationandalsocoversasmanysinkholes andmudholesaspossible(Fig.1).Duringthisstage,a Figure2.Geologicalmapandcross-sectionoftheKomberekarearevisedonthebasisofthegeologicaland geophysicalinvestigations.R.PUTIS KA,D.KUS NIRA K,I.DOSTA L,A.LAC NY ,A.MOJZES ,J.HO K,R.PAS TEKA,M.KRAJN A K,ANDM.BOS ANSKY JournalofCaveandKarstStudies, December2014 N 157

PAGE 6

detailedmagnetometrysurveywasperformedonselected sinkholes,todistinguishbetweennaturalandhuman-made sinkholes. Allgeophysicalworksweredonesimultaneouslyby measuringthecoordinatesusingGPSoratotalstation. ThecoordinatesystemusedinthefiguresisUTM,zone 33N. Aprofilesurveybythegroundgamma-rayspectrometrymethodwasusedtostudytheradioactivityofrocks, soils,andcoversinthestudyarea.Thismethodallowedus todeterminefourmeasuresofgamma-rayactivityofnearsurfacerockandsoilhorizonateachmeasuredstation: totalgamma-rayactivityeUt[Ur](Urisaunitof radioactivity,1Ur 1ppmeU),concentrationof40K [%K],concentrationof238U[ppmeU],andconcentration of232Th[ppmeTh]wherethelettererepresents equivalent.Thedepthrangeisrelativelyshallow,nomore than1mfromthesurface,butthemethodgivesuseful informationmainlyforgeologicalmappingpurposes.In situmeasurementswerecarriedoutusingaportable256channelgamma-rayspectrometerGS-256(Geofyzika Brno,formerCzechoslovakia)with3 933 9 NaI(Tl) scintillationdetectorusingatraditionalgroundsurvey procedure:grass,oldleaves,andthethinuppermosthumus soillayerwereremovedandgroundsurfacewaslevelledin acircleareaof1to1.5mindiameterateachmeasured station.Timeofmeasurementwas2minutesperstation.In total,226stationsweremeasuredalongeightparallelwesteastprofiles(PF1PF8)approximately100-mapartand onetransversenorthwest-southeastprofile(PF9)(Fig.1) witha40-mstepbetweenmeasurements.Theother geophysicalmethodswerecarriedoutatthesamestations ontheprofiles,controlledbyGPSmeasurements. Electromagneticconductivitymapping(DEMP,for DipoleElectroMagneticProfiling)wasperformedonthe eightparallelprofileswithlengthsfrom900mto740m. Thesamplingintervalofthemeasurementwas20m,sothe wholedatasetcovers289points.Themeasurementwas donewithaCMD-4instrumentmanufacturedbyGF InstrumentsInc.(CzechRepublic),whichhasdipolecenter distanceof3.77m,sothemediandepthoftheelectromagneticinvestigationwasaround6.0m.Thedepthrange formostofthepointsallowedustoreachthebedrockin theareawithanegligibleeffectfromthesediments.The mostimportantadvantageofelectromagneticconductivity isthepossibilityofobtainingquickandusefulresultsthat matchverywellwithDCresistivitymethods. The2Delectrical-resistivitytomographyline(Fig.1) wascollectedusinganARESinstrument(GFInstruments Inc.).Thesurveywas1006.5-mlongandconductedwitha dipole-dipolearraywith5.5-melectrodespacing.The88 electrodeswereusedsimultaneously,withalternationof twocurrentandtwopotentialelectrodesandaroll-along survey.Forpost-processinganddatainterpretation,the inversionprogramRES2DINV(LokeandBarker,1996) wasapplied.Itgeneratesatopographicallycorrectedtwodimensionalresistivitymodelofthesubsurfacebyinverting thedataobtainedfromelectricalimaging(Putiskaetal., 2012a).Arobustinversion(L1norm)wasusedbecauseitis moresuitablefordetectingcavesandsharpeninglinear featuressuchasfaultsandcontactswithincomplex geologicalsettingsofkarstregions. ThesameprofilefromtheERTmethodhasbeenused forthemicrogravitysurveyandthestationswereplaced nexttoeachelectrode.Theinstrumentusedforthismethod wasasingleScintrexCG-5unitwitharesolutionof1mGal (102 8ms2 2).Duetothickvegetationcoverinthearea,it wasnotpossibletomeasurethepositionsofthestations usingdifferentialGPS,thereforethelocationsofthe gravitystationsandnearbytopographywasobtained Figure3.Resultofsurfaceelectromagneticconductivityandradiometrymapping.Intheleft-handpartisthemapofapparent resistivityandtheright-handpartisthemapoftotalgamma-rayactivity(Urisaunitofradioactivity,1Ur 1ppmeU).INTEGRATEDGEOPHYSICALANDGEOLOGICALINVESTIGATIONSOFKARSTSTRUCTURESINKOMBEREK,SLOVAKIA158 N JournalofCaveandKarstStudies, December2014

PAGE 7

usingTrimbleM3andTrimbleS8totalstationstyingto severalGPSpointsinthevicinityoftheprofile.Atotal numberof184gravityreadingswereprocessedtoremove instrumentdrift,whichwasdeterminedbyrepetitive measurementatthebasestationlocatedinthecenterof theprofileoveraperiodof2hoursandafterwards processedintoformofcompleteBougueranomaly(Fig.1). Detailedhigh-definitionmagnetometryusingaCsvapourmagnetometerTM-4with0.2-msamplingstepon lines1-mapartwasrealizedatselectedsitesofthestudied regionwiththeaimtodistinguishbetweensinkholesand man-madelimekilns.Selectedsinkholesarehighlightedin theFigure1withgreyrectanglesthatreflecttheactualarea surveyedbydetailedmagnetometry.Thesurveyrectangles fordetailedmagnetometryonthreesiteswerechosendue topracticalreasons.Thefirsttwositeswereplacedonthe largestsinkholeintheKomberekareaandonaknown man-madelimekilnidentifiedbytheburnedlimeresiduals andashesinsideit.Asthelastsite,aswarmofterrain depressionsofuncertainorigin,waschosen.RESULTSANDINTERPRETATIONTheapparentresistivitymap(Fig.3)wasobtained usingtheelectromagneticconductivitymethod(DEMP). Theresultsshowagoodcontrastbetweenthehighresistivityareascomposedofdolomiteandlimestoneand thelow-resistivityonesthataretheeffectofQuaternary sedimentsconcentratedinthetopographydepressions.The majorityofthelargestsinkholesareconcentratedonthe boundarybetweenlowandhighresistivityareas.This boundarywaschosenformicrogravityandERTsurveys. Resultsfromtheradiometricmappingareshownasamap oftotalgamma-rayactivity(Fig.3).Thecontourmapof thetotalgamma-rayactivityeUtdistributionshowsvery goodagreementwiththepictureoftheapparentresistivity distributionforrocksandsoilsinthestudyarea,as obtainedbysurfaceelectromagneticconductivitymapping (DEMP)measurement(Fig.3).Lower-radioactivityareas correspondtohigher-resistivityareasinmapsandvice versa.Simultaneouslybothphysicalfieldsshowgood correlationwithterraintopographyinthestudyarea (Fig.3).Positivetopographicfeaturesaremostlycharacterizedbyhigherelectricalresistivityandlowertotal gamma-rayactivityvalues,whereasnegativeterrain featuresmostlyshowlowerresistivityandhighertotal gamma-rayactivity.Thisrelationshipis,ofcourse,strictly determinedbythegeologicalstructureofthearea(Fig.2), sincethehardcarbonatesformtopographicalelevations characterizedbyhigherelectricalresistivityandlowertotal gamma-rayactivity.WeatheredrocksandQuaternary sedimentsfilltopographicdepressionsinthecentraland southwesternpartsofthestudyarea,whicharecharacterizedbylowerelectricalresistivityandhighertotal Figure4.Resultsfromoneofthedetailedmagnetometrysurveys.Terraindepressionboundariesareshownbydashedlines. Thedepressioninthelowerrightisshowntobeanoldlimekilnbytherelativelystrongdipoleanomaly.R.PUTIS KA,D.KUS NIRA K,I.DOSTA L,A.LAC NY ,A.MOJZES ,J.HO K,R.PAS TEKA,M.KRAJN A K,ANDM.BOS ANSKY JournalofCaveandKarstStudies, December2014 N 159

PAGE 8

gamma-rayactivity.Generally,thetwomaps,theresistivityandtheradioactivity,arequantitativelyopposedto eachother.Inthisway,theelectromagneticandradiometricsurveyssuccessfullyoutlinedtheboundarybetween valleyfillandthecarbonaterocksdolomiteandlimestone. Thetotal-fieldmagnetic-anomalymapofthelastofthe surveyrectanglesisshowninFigure4.Theaimofthe magneticsurveywastorevealcontrastbetweensinkholes andman-madelimekilns.Duetotheremnantmagnetizationproducedduringtheheatingprocesseswewereableto easilyrecognize,basedonqualitativeinterpretation,the limekilnsashavingrelativelystrongdipoleanomalies,as inthelowerrightofFigure4. Gravityandelectrical-resistivitytomographymethods wereemployedaftertheresistivityandradioactivitystudies todetectpossiblecavernousstructuresintheselected profileacrossthewholearea.Suchcavescanbeempty, full,partlywater-filled,orfilledwithadifferentkindof sediment.Air,water,orsediment-filledvoidshaveamuch lowerdensity(air0kgm2 3,water1000kgm2 3,or sediment 2000kgm2 3)thanthehostrockforwhichthe densityof2670kgm2 3hasbeenassigned.Thisdifference indensityisverysignificantandcanbeeasilytracedby gravitysurvey.Gravity,asanintegralmethod,provides onlyinformationaboutthebulkcompositionofthe subsurface.Thereforeadditionalconstraintsareneeded tomodeltheBougueranomalywithappropriatestructures.Electricalresistivitytomographyisareasonable choiceduetothelowcostofthesurveyandthehigh resistivitycontrastthatexistsbetweenanair-filledcavity Figure5.Resultsoftheelectrical-resistivitytomographyandgravitymodeling.Theupperpartofthefigure(a)showsthe correlationbetweentheobservedresidualBougueranomalyandthemodelshowninthemiddlepart(b).TheERTcross-section imageisinthelowerpartofthefigure(c).INTEGRATEDGEOPHYSICALANDGEOLOGICALINVESTIGATIONSOFKARSTSTRUCTURESINKOMBEREK,SLOVAKIA160 N JournalofCaveandKarstStudies, December2014

PAGE 9

andthesurroundingformation(Zhouetal.,2002;Andrej andUros,2012).Cavitiescanbealsopartiallyor completelywater-orsediment-filledand,dependingon thecompositionofthewater,theymightshowaresulting electricalconductivityrangingfromveryconductiveto relativelyresistivecomparedtothehostrock(Putiskaetal., 2012b).TheresultsofERTandgeologicalsurveyswere usedasentriesforthegravitymodelling(Fig.5).Gravity modellingwasperformedusingtheresidualBouguer anomalycomputedforareferencedensityof2670kgm2 3. Thezeroelevationhasbeenusedasalowerboundaryfor themodel,with2-Dgeometryexceptthecavitybody, whichwasmodelledasa2.5-Dobjectwithalateral extensionof70manditscenterplacedontheprofile.To keepthemodelassimpleaspossible,onlythemost importantfeaturesweretakenintoconsideration,where themainstructureofthelithologicalunitswasadopted fromthegeologicalmapofthearea.Receiveddifferential densitiesusedforthemodelareshowninFig.5.Two importantlowdensityandlowresistivitystructuresthat canbeimportantfortheidentificationofkarststructures weredetected. Figure5bshowsthefinalmodel,obtainedfromthe combinedinterpretationofthemicrogravity(Fig.5a)and ERTmeasurements(Fig.5c).Frommicrogravity(Fig.5a) twoimportantlowdensitystructureshavebeendetected. ThefirstBouguergravityminimum,withitscenterlocated onprofilelength225mandamplitudeof 2 0.2mGal,is associatedwiththegeologicalsettinginthearea.The negativeanomalyisproducedfromthelowergravityeffect oftherauhwackes,whichareporousandthereforehave lowerdensity(Fig.2,Fig.5b).Thesecond,moresignificantminimumontheresidualBougueranomalycurve, withamplitudelowerthan 2 0.35mGalandcenterlocated ataround575m,seemstobemoreinterestingfroma speleologicalpointofview,asnogravityanomalies resultingfromgeologywereexpectedthere.Comparing theresultswiththeERTimage,astrongconductive anomalyisvisibleatthesamepointontheprofileasthe maingravityminimum(Fig.5a,c),whereweinserteda Figure6.Thesinkholeclosetotheelectrical-resistivitytomographyandmicrogravitylinethroughtheKomberekarea. Diameteranddepthofthesinkholeare 25mand9m,respectively.R.PUTIS KA,D.KUS NIRA K,I.DOSTA L,A.LAC NY ,A.MOJZES ,J.HO K,R.PAS TEKA,M.KRAJN A K,ANDM.BOS ANSKY JournalofCaveandKarstStudies, December2014 N 161

PAGE 10

bodyrepresentinganemptycaveinFigure5b.Empty spaceusuallyproducesahigh-resistivityanomalyinan ERTimage,butinsomespecialcases,aswhenthecaveis partlyfilledwithconductivematerialsuchasclay,the overallanomalyproducedfromtheemptyspacecanbe conductive(Putiskaetal.,2012b).Thecavebodyinthe finalmodel(Fig.5b)isconnectedtothegroundbya communicationchannel,anditspresenceisalsovisible fromtheground,whereoneofthelargestsinkholes (diameter 25m)intheKomberekareawasfound (Fig.6).TheinvertedERTmodel(Fig.5c)showscarbonaterockwithasignificantlyhigherresistivity(500Ohmm) thantheloamymaterialinthesinkhole,becauseofits considerablysmallerprimaryporosityandfewerinterconnectedporespaces.Loamymaterialscanholdmore moistureandhavehigherconcentrationsofionstoconduct electricity;therefore,theirresistivityvaluesarebelow100 Ohmm.Thehighcontrastinresistivityvaluesbetween carbonaterockandloamymaterialmakesitpossible touseelectricalresistivitytodeterminetheunderground structure.CONCLUSIONSAgeomorphologicalanalysisoftheKomberekarea identifiedmorethan70topographicdepressions.However, notallofthemarekarstlandforms,asman-madelime kilnsarealsopresentinthearea.Detailedhigh-definition magnetometrywassuccessfullyemployedtodistinguish betweennaturalsinkholesandman-madelimekilns (Fig.4).Resultsfromtheradiometricmappinganddipole electromagneticprofiling,supportedbygeologicalmapping,allowedustorefinethegeologicalboundariesofthe lithologicalunitswithintheKomberekkarstarea(Fig.2). Bymeansoftheresistivitytomographyandmicrogravity methods,thefinalgeologicalcross-sectionmodelofthe areawasconstructed(Fig.5b).TheBougueranomaly curve(Fig.5a)showstwodominantnegativeanomalies thatwereinterpretedbyintroducingERTinversemodel andgravityforwardmodelling.Thefirstnegativeanomaly, withcenterlocatedatprofilelocation225mandamplitude of 2 0.2mGal,isassociatedwithapresenceoftheporous rauhwackesformationandseemstobeunimportantfroma speleologicalpointofview.Thesecondmajornegative anomaly,withamplitudemorethan 2 0.35mGaland centerlocatedat 575m,correlateswithaconductive anomalyintheERTinverseimageatdepthof 60m, whereacavitywasdetected.Lateralplacementofthis anomalousareaislinkedthepresenceofthelargest sinkholeintheKomberekkarstarea.Accordingtothe resultsobtainedfromthisstudy,wecanconcludethat microgravitytogetherwithelectricalresistivitytomography haveprovedtobeeffectivetoolsforimagingsubsurface cavitiesinlimestoneatshallowdepths.Thus,webelieve thatthepresentedmethodsandevaluationtechniques couldbesuccessfullyappliedtootherkarstareasand potentiallyhelpinidentifyinghiddenvoidsthatpossibly constitutekarsthazards(seePariseandGunn,2007;De Waeleetal.,2011andreferencestherein).ACKNOWLEDGMENTSTheauthorswouldliketothankstoananonymous reviewerforhishelpfulcommentsthatimprovedthefinal manuscriptandtotheSlovakResearchandDevelopment AgencyAPVV(GrantNos.APVV-0194-10,APVV062511,APVV-0099-11,APVV-0129-12)andtheSlovakGrant AgencyVEGA(GrantNos.1/0095/12,2/0067/12,1/0747/ 11,1/0712/11,1/0131/14)forthesupportoftheirresearch.REFERENCESAndrej,M.,andUros,S.,2012,Electricalresistivityimagingofcave Divaskajama,Slovenia:JournalofCaveandKarstStudies,v.74, no.3,p.235.doi:10.4311/2010ES0138R1. Butler,D.K.,1984,Microgravimetricandgravitygradienttechniquesfor detectionofsubsurfacecavities:Geophysics,v.49,p.108496. doi:10.1190/1.1441723. Cardarelli,E.,Cercato,M.,Cerreto,A.,andDiFilippo,G.,2010, Electricalresistivityandseismicrefractiontomographyto detectburiedcavities:GeophysicalProspecting,v.58,p.685. doi:10.1111/j.1365-2478.2009.00854.x. DeWaele,J.,Gutierrez,F.,Parise,M.,andPlan,L.,2011,Geomorphologyandnaturalhazardsinkarstareas:areview.Geomorphology, v.134,no.1,p.1.doi:10.1016/j.geomorph.2011.08.001. Dobecki,T.L.,andUpchurch,S.B.,2006,Geophysicalapplicationsto detectsinkholesandgroundsubsidence:TheLeadingEdge,v.25, p.336.doi:10.1190/1.2184102. El-Qady,G.,Hafez,M.,Abdalla,M.A.,andUshijima,K.,2005,Imaging subsurfacecavitiesusinggeoelectrictomographyandground-penetratingradar:JournalofCavesandKarstStudies,v.67,no.3, p.174. Ford,D.C.,andWilliams,P.,2007,KarstHydrogeologyandGeomorphology:Chichester,JohnWiley&SonsInc.,576p. Gambetta,M.,Armadillo,E.,Carmisciano,C.,Stefanelli,P.,Cocchi,L., andTontini,F.C.,2011,Determininggeophysicalpropertiesofa near-surfacecavethroughintegratedmicrogravityverticalgradient andelectricalresistivitytomographymeasurements:JournalofCave andKarstStudies,v.73,no.1,p.11.doi:10.4311/jcks2009ex0091. Gibson,P.J.,Lyle,P.,andGeorge,D.M.,2004,Applicationofresistivity andmagnetometrygeophysicaltechniquesfornear-surfaceinvestigationsinkarsticterranesinIreland:JournalofCaveandKarstStudies, v.66,no.2,p.35. Kaufmann,G.,Romanov,D.,andNielbock,R.,2011,Cavedetectionusing multiplegeophysicalmethods:Unicorncave,HarzMountains,Germany:Geophysics,v.76,no.3,p.B71B77.doi:10.1190/1.3560245. Krauter,E.,1971,ZurGeneserauhwackigerBreccienderalpinenTriasan BeispielenausderSchweizundO sterreich:Geologisch-PalaontologischeMitteilungenInnsbruck,v.1,no.7,11p. Lac ny,A.,Putiska,R.,Dostal,I.,andKusnirak,D.,2012,Vyuzitie metodyERTpriprieskumejasky n vHavranejskale(Plaveckykras): SlovenskyKras(ActaCarsologicaSlovaca),v.50,no.1,p.41. Loke,M.H.,andBarker,R.D.,1996,Rapidleast-squaresinversionof apparentresistivitypseudosectionsbyaquasi-Newtonmethod: GeophysicalProspecting,v.44,p.131.doi:10.1111/j.13652478.1996.tb00142.x. Margiotta,S.,Negri,S.,Parise,M.,andValloni,R.,2012,Mappingthe susceptibilitytosinkholesincoastalareas,basedonstratigraphy, geomorphologyandgeophysics:NaturalHazards,v.62,no.2, p.657.doi:10.1007/s11069-012-0100-1. Mochales,T.,Casas,A.M.,Pueyo,E.L.,Pueyo,O.,Roman,M.T., Pocov ,A.,Soriano,M.A.,andAnson,D.,2008,Detectionof undergroundcavitiesbycombininggravity,magneticandground penetratingradarsurvey:AcasestudyfromtheZaragozaarea,NEINTEGRATEDGEOPHYSICALANDGEOLOGICALINVESTIGATIONSOFKARSTSTRUCTURESINKOMBEREK,SLOVAKIA162 N JournalofCaveandKarstStudies, December2014

PAGE 11

Spain:EnvironmentalGeology,v.53,p.1067.doi:10.1007/ s00254-007-0733-7. Nyquist,J.E.,Peake,J.S.,andRoth,M.J.S.,2007,Comparisonofan optimizedresistivityarraywithdipole-dipolesoundingsinkarst terrain:Geophysics,v.72,no.4,p.F139F144.doi:10.1190/ 1.2732994. Parise,M.,andGunn,J.,eds.,2007,NaturalandAnthropogenicHazards inKarstAreas:Recognition,AnalysisandMitigation,London, GeologicalSociety,London,SpecialPublicationno.279,202p. Parise,M.,andLollino,P.,2011,Apreliminaryanalysisoffailure mechanismsinkarstandman-madeundergroundcavesinSouthern Italy:Geomorphology,v.134,no.1,p.132.doi:10.1016/ j.geomorph.2011.06.008. Pepe,P.,Pentimone,N.,Garziano,G.,Martimucci,V.,andParise,M., 2013,LessonslearnedfromoccurrenceofsinkholesrelatedtomanmadecavitiesinatownofsouthernItaly, in Land,L.,Doctor,D.H., andStephenson,J.B.,eds.,Proceedingsofthe13thMultidisciplinary ConferenceonSinkholesandtheEngineeringandEnvironmental ImpactsofKarst,Carlsbad(NewMexico,USA),6May2013, Carsbad,NationalCaveandKarstResearchInstitute,p.393. Polak,M.,Plasienka,D.,Kohut,M.,Putis ,M.,Beza k,V.,Filo,I., Olsavsky,M.,Havrila,M.,Buc ek,S.,Maglay,J.,Elec ko,M.,Fordina l, K.,Nagy,A.,Hras ko,L .,Nemeth,Z.,Ivanic ka,J.,andBroska,I., 2011,Geologickamaparegio nuMaly chKarpatvM1:50000(GeologicalmapoftheMaleKarpatyMts.Region,1:50000),Bratislava, MZ PSR,S tatnygeologicky u stavDiony zaS tura. Putiska,R.,Dostal,I.,Mojzes,A.,Gajdos,V.,Rozimant,K.,andVojtko, R.,2012a,TheresistivityimageoftheMura n faultzone(Central WesternCarpathians)obtainedbyelectricalresistivitytomography: GeologicaCarpathica,v.63,no.3,p.233.doi:10.2478/v10096012-0017-3. Putiska,R.,Nikolaj,M.,Dostal,I.,andKusnirak,D.,2012b, Determinationofcavitiesusingelectricalresistivitytomography: ContributionstoGeophysicsandGeodesy,v.42,no.2,p.201. Schaad,W.,1995,DieEntstehungvonRauhwackendurchdieVerkarstungvonGips:EclogaeGeologicaeHelvetiae,v.88,p.59. Waltham,T.,Bell,F.G.,andCulshaw,M.G.,2005,Sinkholesand Subsidence:KarstandCavernousRocksinEngineeringand Construction.Chichester,UK,Praxis,382p. White,W.B.,1990,Surfaceandnear-surfacekarstlandforms, in Higgins, C.G.,andCoates,D.R.,eds.,GroundwaterGeomorphology:The RoleofSubsurfaceWaterinEarth-SurfaceProcessesandLandforms: GeologicalSocietyofAmericaSpecialPaper252,p.157. Zhou,W.,Beck,B.F.,andAdams,A.L.,2002,Effectiveelectrodearrayin mappingkarsthazardsinelectricalresistivitytomography:EnvironmentalGeology,v.42,p.922.doi:10.1007/s00254-002-0594-z.R.PUTIS KA,D.KUS NIRA K,I.DOSTA L,A.LAC NY ,A.MOJZES ,J.HO K,R.PAS TEKA,M.KRAJN A K,ANDM.BOS ANSKY JournalofCaveandKarstStudies, December2014 N 163

PAGE 12

POSSIBLECAVERN-FORMINGACTIVITYATMILLENNIAL TIMESCALESANDITSIMPACTONVARIATIONSIN SUBMARINECAVEENVIRONMENTSANDHABITAT AVAILABILITY,OKINAWA,JAPANAKIHISAKITAMURA1*,CHIKAKOTAMAKI1,YOSUKEMIYAIRI2,YUSUKEYOKOYAMA2,ANDHIDEKIMORI3Abstract: Weexaminedthespatialdistributionsand14Cagesofsessilemarineorganisms inthesubmarinecaveGinamaatOkinawa,Japan.Thecaveisthenorthernmostandonly knownsubmarinecavewithanairchamberinthenorthwesternPacificregion.Theupper limitoflivingindividualsofthecorallinesponges Acanthocaeteteswellsi andthebivalve Pycnodontetaniguchii inthecaveislocatedatadepthofapproximately10mrelativetothe surfaceofthepoolinthecave.Lowtemperaturesandlowsalinitiespreventthese organismsfromcolonizingwaterdepthsshallowerthan10m.Ontheotherhand,theupper limitoffossilindividuals,datedbetween5117and387calyrBP,isatadepthof2.5m, implyingthatatpresenttheinfluenceoffreshwateronthecavepoolisstrongerthanithas beenatothertimesinthepast5000years.Thisincreaseinthefluxoffreshwatermaybe explainedbycontinuouscavern-formingactivitiessuchasdissolution.INTRODUCTIONAnumberofworkershaverecentlyusedarchival materialsfromsubmarinecavestoreconstructenvironmentalchangesduringthelateQuaternary(Antonioli etal.,2001;Kitamuraetal.,2007,Yamamotoetal.,2008, 2009a).Forexample,oxygen-isotopiccompositionsof corallinespongesandmicro-bivalvesfromsubmarinecaves havebeenstudiedasproxiesforpalaeotemperature(Bo hmet al.,2000;Haase-Schrammetal.,2003;Yamamotoetal., 2008,2009b,2010;Kitamuraetal.,2013).Inaddition,the historyofsealevelchangeshasbeenexaminedusingboth speleothemsandchangesinfaunalassemblagespreservedin cavesediments(Doraleetal.,2010;Tuccimeietal.,2010;van HengstumandScott,2011,2012,vanHengstumetal.,2011). However,fewstudieshaveexaminedtemporalchanges infaunalcommunitiesinsubmarinecaves.InMediterraneansubmarinecaves,ChevaldonneandLejeusne(2003) documentedthatpopulationsofcoldstenothermalspecies ofmysids(Crustacea)werereplacedbycongenersofwarmer affinitiesduringaperiodofregionalwarminginthe summersof1997and1999.Parravicinietal.(2010) examinedchangesinsessilecommunitiesinaMediterranean submarinecaveusingphotographstakenin1986and2004, revealingthatmassivenumbersoforganismsexperienced highratesofmortalityduetothermalanomaliesduringthe summerheatwavesof1999and2003andthatthesegroups werereplacedbyencrustingorganisms. Thenorthward-flowingtropicalKuroshioCurrent allowscoralreefstoformintheRyukyuIslandsof southwesternJapan,northwesternPacific(Fig.1a).The islandsextendfromTane-gaIsland(30 u 44 9 N,131 u 00 9 E)in thenortheasttoYonaguniIsland(24 u 27 9 N,123 u 00 9 E)in thesouthwest.Manysubmarinelimestonecavesoccuron theislands(HayamiandKase,1993).However,previous studiesofchangesinthecommunitiesofsubmarinescaves haveonlyconsideredmillennial-scalevariationsinthe speciescompositionsofbivalves(Kitamuraetal.,2007; Yamamotoetal.,2009a)andalgalsymbiont-bearinglarge benthicforaminifers(Omorietal.,2010)inDaidokutsu submarinecave(waterdepth,29m)atIeIsland,Okinawa (Fig.1a).Inbothstudies,thespecieslivinginthe innermostareasofthecavewereseentohavebecome increasinglydominantoverthepast7000years,whilethose livingnearthecaveentranceshavedeclinedinabundance. Thestudiesconcludedthatthechangeswerecausedbya declineinfoodsupplyandlightintensityinthecave associatedwiththefillingofcavitieswithinthereefduring atleastthepast6500years.However,temporalchangesin thecommunitiesofsessileorganismsinsubmarinecaves haveyettobeexaminedinthenorthwesternPacific. Thisstudyexaminedmillennial-scalevariationsinthe compositionsofsessilemarinefaunalcommunitiesinthe submarinecaveGinamaatOkinawa,Japan(Fig.1bd). Thecaveisuniqueinbeingthenorthernmostandonly knownsubmarinecavewithanairchamberinthe northwesternPacificregion.STUDYAREAThesubmarinecaveGinamaoccursinTriassiclimestone(Ishibashi,1974)onthenorthernmostcoastof OkinawaIsland(Fig.1b).Thecoastlineischaracterized *CorrespondingAuthor:seakita@ipc.shizuoka.ac.jp1InstituteofGeosciences,ShizuokaUniversity,Shizuoka,422-8529,Japan2AtmosphereandOceanResearchInstitute,UniversityofTokyo,Chiba,277-8564, Japan3DivisionofTechnicalService,ShizuokaUniversity,Shizuoka,422-8529,JapanA.Kitamura,C.Tamaki,Y.Miyairi,Y.Yokoyama,andH.MoriPossiblecavern-formingactivityatmillennialtimescalesandits impactonvariationsinsubmarinecaveenvironmentsandhabitatavailability,Okinawa,Japan. JournalofCaveandKarstStudies, v.76, no.3,p.164.DOI:10.4311/2013PA0109164 N JournalofCaveandKarstStudies, December2014

PAGE 13

byseacliffswithaheightofapproximately10m.The entrancetothecave,whichisapproximately15-mbelow sealevel,is5mhighand10mwide.Thecaveconsistsofa narrow40-mlongupward-slopinggalleryapproximately 2mhighand2mwideandadark,innerair-filledchamber withaheightof15m(Fig.1c).Manystalagmitesare presentbothaboveandbelowthewatersurfaceinthe innerchamber. Noemergentcoastallandformsarepresentinthearea, suggestingnonetupliftoftheareaduringtheHolocene.The subsidencerateofthestudyareaiscurrently0.04mmy2 1, basedongeographicrecordsfrom1979to2006(GeographicalSurveyInstitute,2007).HongoandKayanne(2010) reconstructedtheHolocenesealevelcurvebasedoncoral reefsofIshigakiIsland,RyukyuIslands,whichislocated approximately500kmsouthwestofthepresentstudyarea. Basedonthereconstruction,theentrancetoGinamaCaveis thoughttohavebecomesubmergedatapproximately 8000yearsBP.Tothebestofourknowledge,noresearches havebeenconductedonlocalgroundwater,andnowater wellsarepresentinthearea.METHODSWatertemperatureandsalinityweremeasuredoutside GinamaCaveclosetotheseasurfaceandinthepool belowthechamberat1mintervalsfromthesurfaceofthe pooltoadepthof12m;allmeasurementswereobtained on10July2012.Wecarefullyobservedandrecordedall mega-fossilsofsessileorganismsontherockwallofthe cavepoolatdepthsshallowerthan10mon124October 2011andagainon10July2012,atwhichtimewe collectedfossilshellsofthreeindividualsofthecoralline sponge Acanthocaeteteswellsi andtenindividualsofthe bivalve Pycnodontetaniguchii .Bothspeciesaresessile organismsthatliveinsubmarinecaveorcryptichabitats (Jacksonetal.,1971;JacksonandWinston,1982;Hayami andKase,1992).Thebivalve P.taniguchii attachesto hardsubstrateswithathickenedleftvalve(Hayamiand Kase,1992).Theshellsofboth A.wellsi and P.taniguchii arecomposedof100%calcite(HayamiandKase,1992; ReitnerandGautret,1996).Manysessilemicro-bivalves, mostofwhicharelessthan5mminlength,suchas Cosa Figure1.(a)LocationmapofOkinawa,Japan,showingthelocationsofdrillcoresA7(Sunetal.,2005;Xiangetal.,2007), MD403(Linetal.,2006),KY07-04-01(Kubotaetal.,2010),andDonggeCave(Wangetal.,2005);KC:KuroshioCurrent; inset,areaofpart(b)and1:DaidokutsuCaveoffIeIsland.(b)LocationmapofsubmarinecaveGinamaonthenortherntipof Okinawa.(c)Simplifiedcross-sectionofthecave.Noteverticalexaggeration.Thewhitearrowshowsthedirectionoftheview in(d).(d)Photographofthecaveentrance,lookingoutwardfrominsidethecave.A.KITAMURA,C.TAMAKI,Y.MIYAIRI,Y.YOKOYAMA,ANDH.MORIJournalofCaveandKarstStudies, December2014 N 165

PAGE 14

waikikia Cosakinjoi ,and Parvamussiumcrypticum are presentinthesubmarinecavesofOkinawa(Hayamiand Kase,1992).However,becausethesespeciesattachtorock surfacesbymeansofabyssus,theshellsdropoff immediatelyafterdeath.Inthisstudy,wedidnotinvestigate thesessilemicro-bivalvespecies. Collectedfossilspecimenswerecutalongtheaxisof maximumgrowthusingalow-speedsaw.Thespecimenswere checkedfordiageneticalterationusingthin-sectionobservationswithanopticalmicroscope.Wedeterminedradiocarbon agesofallthirteenspecimensof A.wellsi and P.taniguchii Thesamplesweregraphitized,andthetargetgraphiteswere analyzedusingacceleratormassspectrometryattheUniversityofTokyo,Japan.Theresultswerecorrectedusinga reservoirageof400years,andtheagesweretransformedtoa calendartimescaleusingtheprogramOxCal4.1(Bronk Ramsey2009),basedoncomparisonswithMarine13data (Reimeretal.2013),afterapplyinga D R valueforthe Okinawaregionof29 6 18years(Yonedaetal.,2007).RESULTSWatertemperaturesandsalinitiesinthepoolincreased withwaterdepth(Fig.2).Atadepthof12m,thevaluesof bothparameterswerenearlyequaltothoseofsea-surface wateroutsidethecave. Alivingindividualof Acanthocaeteteswellsi wasfound atawaterdepthof11.3m.Wealsofoundayoung individual Pycnodontetaniguchii atadepthof10.5m. Threefossilindividualsof A.wellsi andtenfossil individualsof P.taniguchii (leftvalvesonly)weredistributedatdepthsofupto2.5mwithinthebrackish-waterlens (Figs.3and4).Thin-sectionobservationsshowthatthese specimenswereunaffectedbydiageneticalteration,alFigure2.WatertemperatureandsalinityinGinamaCave, showingthewaterdepthoftheupperlimitofliving individualsofthebivalve Pycnodontetaniguchii (1)andof thecorallinesponges Acanthocaeteteswellsi (2). Figure3.Photographofafossilofthebivalve Pycnodontetaniguchii (sampleGinama9)atawaterdepthof6.5m.POSSIBLECAVERN-FORMINGACTIVITYATMILLENNIALTIMESCALESANDITSIMPACTONVARIATIONSINSUBMARINECAVEENVIRONMENTSANDHABITAT AVAILABILITY,OKINAWA,JAPAN166 N JournalofCaveandKarstStudies, December2014

PAGE 15

thoughapartiallycalcifiedfillingcanbeobservedonthe surfacelayerof A.wellsi (Fig.5a);thisportionwasexcluded from14Cdating.Theradiocarbonagesobtainedforthe fossilspecimensarelistedinTable1andplottedinFigure6. Allspecimensof A.wellsi andsevenindividualsof P. taniguchii fallintoanolderagegroup,from5117 6 243(2 s ) to3105 6 232(2 s)calyearsBP.Threeindividualsof P. taniguchii fallintoayoungeragegroup,from566 6 75(2 s ) to387 6 96(2 s)calyearsBP(Fig.6).DISCUSSIONThecorallinesponge Acanthocaeteteswellsi isa componentofcrypticsessilecommunitiesonmodern Indo-Pacificreefs(ReitnerandGautret,1996).According toGrottoli(2006),thehabitatsof A.wellsi areunaffected byrunofffromland.Thisspecieshasalsobeencollectedin submarinecavesonthesouthernmostcoastoftheOkinawa mainlandandonKumeIsland,Okinawa(Ohmorietal., Figure4.Photographsofspecimens.1: Acanthocaeteteswellsi (samplenos.Ginama4-3,6-1,5-2).4: Pycnodonte taniguchii 4-13;(samplesGinama3,8,9,10,7-1,7-2,7-3,19-1,19-2,19-3).A.KITAMURA,C.TAMAKI,Y.MIYAIRI,Y.YOKOYAMA,ANDH.MORIJournalofCaveandKarstStudies, December2014 N 167

PAGE 16

2008;Higaetal.,2010).Noneofthe A.wellsi sitesare reportedtobeinfluencedbyrunofffromland.We thereforethinkthat A.wellsi thrivesinconditionsof normalsalinity.Toourknowledge,theGinamaCaveisthe northernmostcaveinwhich A.wellsi hasbeenfound. Liveindividualsofthebivalve P.taniguchii havebeen foundatmanycavernicoloussitesofthefollowingislandsin thewesternPacificandeasternregionsoftheIndianOcean: Ryukyu(Miyako,Okinawa,andYonaguniislands),Bonin, Palau,Philippines(Luzon,Cebu,Boholislands),Malaysia, Vanuatu,Fiji,TongaandThailand(nearPhuket)(Hayami andKase,1992,1999).AccordingtoHayamiandKase (1992,1999),individualsof P.taniguchii alwaysgrowunder conditionsofnormalsalinityandtemperature.GinamaCave andDaidokutsuCave,locatedapproximately45kmsouthwestofthestudyareaoffIeIsland(Fig.1a),areamongthe northernmostsitesatwhich P.taniguchii hasbeenfound. AccordingtoYamamotoetal.(2010),seasonalchangesin watertemperatureinDaidokutsuCaverangefrom29 u Cin AugustSeptemberto21 u CinFebruary.Datafromthe JapanOceanographicDataCenter(2007)showthat,at waterdepthsof20m,similartothoseofthestudyarea,water temperaturesrangefrom28.2 6 0.6 u CinSeptemberto22.0 6 0.9 u CinFebruary,whichindicatesthat P.taniguchii can surviveattemperatureof21 u Cduringthewinterseason.No studieshavereportedonthelowerlimitsofdissolveoxygen forthesurvivalofeither A.wellsi or P.taniguchii Asnotedabove,livingindividualsof A.wellsi and P. taniguchii werefoundinGinamaCaveatdepthsof11.3m and10.5m,respectively.Atdepthsshallowerthan10m, bothwatertemperatureandsalinitydecreaseupward.Itis thereforelikelythatrelativelylowtemperaturesand Figure5.Thin-sectionphotomicrographs(cross-polarizedlight).(a)Specimenofcorallinesponge Acanthocaeteteswellsi (sampleGinama4-3;age,4623 206calyearsBP);arrowsshowcalcifiedfillings.(b)Specimenofthebivalve Pycnodonte taniguchii (sampleGinama3;age,5064 221calyearsBP). Table1.Resultsof14Cdating;uncertaintiesare2 s. SampleNumberSample Depth,mConventional14CAge,BPCalendarAgeRanges,calBP Ginama4-3 Acanthocaeteteswellsi 2.5 4491 6 83 4623 6 206 Ginama6-1 Acanthocaeteteswellsi 4.2 4863 6 85 5117 6 243 Ginama5-2 Acanthocaeteteswellsi 5.1 3454 6 83 3279 6 222 Ginama3 Pycnodontetaniguchii 4.0 4814 6 86 5064 6 221 Ginama8 Pycnodontetaniguchii 5.5 3966 6 85 3932 6 247 Ginama9 Pycnodontetaniguchii 6.5 4356 6 86 4474 6 271 Ginama10 Pycnodontetaniguchii 6.5 4304 6 82 4394 6 251 Ginama7-1 Pycnodontetaniguchii 8.3 4622 6 88 4806 6 262 Ginama7-2 Pycnodontetaniguchii 8.3 3313 6 86 3105 6 232 Ginama7-3 Pycnodontetaniguchii 8.3 4591 6 158 4779 6 439 Ginama19-1 Pycnodontetaniguchii 9.7 784 6 47 387 6 96 Ginama19-2 Pycnodontetaniguchii 9.7 996 6 46 566 6 75 Ginama19-3 Pycnodontetaniguchii 9.7 896 6 43 489 6 86POSSIBLECAVERN-FORMINGACTIVITYATMILLENNIALTIMESCALESANDITSIMPACTONVARIATIONSINSUBMARINECAVEENVIRONMENTSANDHABITAT AVAILABILITY,OKINAWA,JAPAN168 N JournalofCaveandKarstStudies, December2014

PAGE 17

salinitiespreventedthesesessilemarineorganismsfrom colonizingsurfacesatdepthsshallowerthan10m.Onthe otherhand,14Cagesobtainedinthepresentstudyshow thatfossilspecimensof A.wellsi and P.taniguchii are presentatdepthsshallowerthan10m.Accordingto HongoandKayanne(2010),amid-Holocenehighstand occurredatapproximately5000cal.yearsBP,atalevelof approximately3 6 2.5mabovepresentmeansealevel, basedonverticaldistributionsofcoralsonIshigakiIsland, Okinawa.Atapproximately5000cal.yearsBP, A.wellsi (sampleno.Ginama6-1)and P.taniguchii (sampleno. Ginama3)weredistributedatdepthsof4.0and4.2m, Figure6.Top:Agedataforfossilsfromthispaper.Middle: d18Oandcorrespondingtemperaturevaluesof Carditella iejimensis shellsfromcoredsamples(Kitamuraetal.,2013).Bottom:OncommonscaleMg/Ca-derivedSSTsand d18OswrecordsfromcoresA7(Sunetal.,2005),MD403(Linetal.,2006),andKY07-04-01(Kubotaetal.,2010)andstalagmite d18O recordsfromDonggeCave,southeastChina(Wangetal.,2005).ThelocationsofallsitesareshowninFigure1.A.KITAMURA,C.TAMAKI,Y.MIYAIRI,Y.YOKOYAMA,ANDH.MORIJournalofCaveandKarstStudies, December2014 N 169

PAGE 18

respectively(Table1).Assumingthattherateofsubsidenceduringthepast5000yearsequalsthepresentday surveyedrateof0.04mm/year,thetotalsubsidenceduring thisperiodisestimatedtobe0.2m.Basedonthis subsidencerateandsealevelatapproximately5000cal. yearsBP(3 6 2.5mabovepresentmeansealevel),the estimateddepthsof A.wellsi and P.taniguchii werefrom 4.3to9.5m,andfrom4.5to9.7m,respectively.From 566 6 75to387 6 96calyearsBP,theupperlimitof survivalofindividualsof P.taniguchii (depth,9.7m; samplenosGinama19-1,-2,and-3)wasshallowerthan thatofpresentlylivingindividuals.Wethereforesuggest thattheinfluenceoffreshwaterintheGinamaCavepool atthepresentdayisthestrongestithasbeeninthepast 5000years,althoughwenotethatdataaremissingforthe period3105calyearsBP.Inaddition,changesinthe fluxoffreshwaterintothecavecannotbeexplainedby relativesealevelchanges. Therainyseasoninthestudyareaoccursduringthe summermonsoon.PreciselydatedstalagmiteoxygenisotoperecordsfromChinarevealthataHolocene weakeningofthesummermonsoonsince7000yearsBP correspondstoanorbitallyinducedreductioninsummertimesolarinsolationintheNorthernHemisphere(e.g., Dykoskietal.,2005;Wangetal.,2005)(Fig.2).Sucha weakeningofthesummermonsoonshouldcausea decreaseinthefluxoffreshwaterintothecave. ManystudieshaveexaminedHoloceneoceanographic changesintheEastChinaSea,basedongeochemical analysesoftheplanktonicforaminifera Globigerinoides ruber ,whichremainsatwaterdepthsof2to50mduringits lifecycle(e.g.,Fairbanksetal.,1982;Hemlebenetal., 1989;Linetal.,2004)andwererecoveredfromdeep-sea sedimentarycores(Jianetal.,2000;Ijirietal.,2005;Sun etal.,2005;Linetal.,2006;Kubotaetal.,2010).Sunetal. (2005),Linetal.(2006),andKubotaetal.(2010)reported nochangesinsea-surfacetemperaturesorsea-surface salinitiesintheEastChinaSeaduringthepast7000years. Jianetal.(2000)proposedthatadecreaseintemperatures duringtheperiod46000cal.yearsBPwasrelatedtoan intensificationofthewintermonsoon,althoughthey reportednochangesintemperaturesorsalinitiesduring thepast2700years. Yamamotoetal.(2010)measured d18Ovaluesoffossils ofthemicro-bivalve Carditellaiejimensis ,whichhave heightandlength 3.5mmanddwellonthesediment surface(HayamiandKase,1993),fromcoresediments collectedfromDaidokutsuCave,IeIsland,Okinawa.The resultsalsoshownoclearlong-termtrendsinthe d18O valuesofbivalvesduringthepast7000years.More recently,Kitamuraetal.(2013)analyzed d18Ovaluesof 50living C.iejimensis specimensfromDaidokutsuCave (Fig.6)andconcludedthatthe d18Ovaluesrepresentthe meanannualtemperatureand d18Ovalueofseawater.In summary,Kitamuraetal.(2013)confirmedthatno changeshaveoccurredineithersea-surfacetemperatures orsalinitiesduringthepast7000years.Theseresultsare consistentwiththefindingsofpreviousstudiesintheEast ChinaSea(Jianetal.,2000;Ijirietal.,2005;Sunetal., 2005;Linetal.,2006;Kubotaetal.,2010). Bothoceanographicdataandstalagmiteoxygenisotope recordsshowthatagradualweakeningofthesummer monsoonhasnotbeenasignificantinfluenceonseasurfacetemperaturesandsalinitiesintheareaofOkinawa overthepast7000years.Thus,theincreasedfluxoffresh watertoGinamaCavecannotbeexplainedbyclimate variables.Alternatively,wesuggestthatthedevelopment ofcrevicesandpassageswithinthecavecausedtheincrease inthefluxoffreshwaterovertime.Totestthis interpretation,afurtherstudyshouldmeasuregrowth ratesofstalagmitesintheairchamberwithinthecave. Ourhypothesisisoppositetothatproposedfornearby DaidokutsuCave,whichisthatcontinuousfillingof cavitieswithinthereeffoundationofthecavehas influencedsubmarinecavecommunities(Yamamoto etal.,2009a;Omorietal.,2010).Thedifferencebetween thetwocaves,thatis,thepossibledevelopmentofcavities versusfillingofcavities,indicatesvariabilityofmillenniascalecavernformingprocessesinsubmarinecavescaused bydifferenceintheirverticalpositionrelativetosealevel.ACKNOWLEDGEMENTSWegreatlyappreciatetheassistanceofKoushin YasumuraandFumioTamamurainthecollectionof samples.Wethanktwoanonymousreviewers,whose commentsandsuggestionsimprovedtheoriginalmanuscript.WethankA.StallardforimprovingtheEnglishin themanuscript.ThisstudywasfundedbytheMitsubishi Foundation.REFERENCESAntonioli,F.,Silenzi,S.,andFrisia,S.,2001,TyrrhenianHolocene palaeoclimatetrendsfromspeleanserpulids:QuaternaryScience Reviews,v.20,p.1661.doi:10.1016/S0277-3791(01)00012-9. Bohm,F.,Joachimski,M.M.,Dullo,W-C.,Eisenhauer,A.,Lehnert,H., Reitner,J.,andWorheide,G.,2000,Oxygenisotopefractionationin marinearagoniteofcorallinesponges:GeochimicaetCosmochimica Acta,v.64,p.1695.doi:10.1016/S0016-7037(99)00408-1. BronkRamsey,C.,2009,Bayesiananalysisofradiocarbondates: Radiocarbon,v.51,p.337. Chevaldonne,P.,andLejeusne,C.,2003,Regionalwarming-induced speciesshiftinnorth-westMediterraneanmarinecaves:Ecology Letters,v.6,p.371.doi:10.1046/j.1461-0248.2003.00439.x. Dorale,J.A.,Onac,B.P.,Fornos,J.J.,Gines,J.,Gines,A.,Tuccimei,P., andPeatel,D.W.,2010,Sea-levelhighstand81,000yearsagoin Mallorca:Science,v.327,p.860.doi:10.1126/science.1181725. Dykoski,C.A.,Edwards,R.L.,Cheng,Hai,Yuan,Daoxian,Cai,Yanjun, Zhang,Meiliang,Lin,Yushi,Qing,Jiaming,An,Zhisheng,and Revenaugh,J.,2005,Ahigh-resolution,absolute-datedHolocene anddeglacialAsianmonsoonrecordfromDonggeCave,China:Earth andPlanetaryScienceLetters,v.233,p.71.doi:10.1016/j.epsl. 2005.01.036. Fairbanks,R.C.,Sverdlove,M.,Free,R.,Wiebe,P.H.,andBe,A.W.H., 1982,Verticaldistributionandisotopicfractionationofliving planktonicforaminiferfromthePanamaBasin: Nature ,v.298, p.841.doi:10.1038/298841a0.POSSIBLECAVERN-FORMINGACTIVITYATMILLENNIALTIMESCALESANDITSIMPACTONVARIATIONSINSUBMARINECAVEENVIRONMENTSANDHABITAT AVAILABILITY,OKINAWA,JAPAN170 N JournalofCaveandKarstStudies, December2014

PAGE 19

GeographicalSurveyInstitute,2007,CrustalMovementsintheKyushu andOkinawaDistricts:ReportoftheCoordinatingCommitteefor EarthquakePrediction,Japan,v.78,p.561(inJapanese). Grottoli,A.G.,2006,Monthlyresolvedstableoxygenisotoperecordina Palauansclerosponge Acanthocaeteteswellsi fortheperiodof1977 2001, in Suzuki,Y.,Nakamori,T.,Hidaka,M.,Kayanne,H., Casareto,B.E.,Nadaoka,K.,Yamano,H.,andTsuchiya,M.,eds., eds.,Proceedingsofthe10thInternationalCoralReefSymposium: Tokyo,JapaneseCoralReefSociety,p.572. Haase-Schramm,A.,Bohm,F.,Eisenhauer,A.,Dullo,W-C.,Joachimski, M.M.,Hansen,B.,andReitner,J.,2003,Sr/Caratiosandoxygen isotopesfromsclerosponges:TemperaturehistoryoftheCaribbean mixedlayerandthermoclineduringtheLittleIceAge:Paleoceanography,v.18,no.3,doi:10.1029/2002PA000830. Hayami,I.,andKase,T.,1992,AnewcrypticspeciesofPycnodontefrom RyukyuIslands:Alivingfossiloyster:TransactionsandProceedings ofthePalaeontologicalSocietyofJapan,NewSeries,No.165, p.107089. Hayami,I.,andKase,T.,1993,SubmarinecavebivalviafromtheRyukyu Islands:systematicsandevolutionarysignificance:TheUniversity Museum,TheUniversityofTokyo,Bulletin35,133p. Hayami,I.,andKase,T.,1999,GeographicDistributionof Pycnodonte taniguchii :NewsletteroftheMalacologicalSocietyofJapan,v.30, p.5(inJapanesewithEnglishAbstract). Hemleben,C.,Spindler,M.,andAnderson,O.R.,1989,Modern PlanktonicForaminifera:NewYork,Springer,363p. Higa,K.,Fukuyama,Y.,Sekiya,M.,West,R.R.,Nagai,K.,and Sugiyama,T.,2010,Habitatofalivingfossil, Acanthocaetetes (Class Demospongiae),alongtheKomesucoastinthesouthernpartof OkinawaIsland:FukuokaUniversityScienceReports,v.40, p.233(inJapanese). Hongo,C.,andKayanne,H.,2010,Holocenesea-levelrecordfromcorals: reliabilityofpaleodepthindicatorsatIshigakiIsland,RyukyuIslands, Japan:Palaeogeography,Palaeoclimatology,Palaeoecology,v.287, p.143.doi:10.1016/j.palaeo.2010.01.033. Ijiri,A.,Wang,Luejiang,Oba,T.,Kawahata,H.,Huang,Chen-Yue,and Huang,Chi-Yue,2005,Paleoenvironmentalchangesinthenorthern areaoftheEastChinaSeaduringthepast42,000years:Palaeogeography,Palaeoclimatology,Palaeoecology,v.219,p.239. doi:10.1016/j.palaeo.2004.12.028. Ishibashi,T.,1974,OntheTriassicFormationatHedo-misakiin Okinawa-jima:TheJournaloftheGeologicalSocietyofJapan, v.80,p.329.(inJapanese). Jackson,J.B.C.,Goreau,T.F.,andHartman,W.D.,1971,Recent brachiopod-corallinespongecommunitiesandtheirpaleoecological significance:Science,v.173,p.623.doi:10.1126/science.173. 3997.623 Jackson,J.B.C.,andWinston,J.E.,1982,Ecologyofcrypticcoralreef communities.I.Distributionandabundanceofmajorgroupsof encrustingorganisms:JournalofExperimentalMarineBiologyand Ecology,v.57,p.135.doi:10.1016/0022-0981(82)90188-5. JapanOceanographicDataCenter,OnlineData,2007,http://jdoss1.jodc. go.jp/cgi-bin/1997/bts.jp.[AccessedOctober10,2013] Jian,Zhimin,Wang,Pinxian,Saito,Y.,Wang,Jiliang,Pflaumann,U., Oba,T.,andCheng,Xinrong,2000,Holocenevariabilityofthe KuroshioCurrentintheOkinawaTrough,northwesternPacific Ocean:EarthandPlanetaryScienceLetters,v.184,p.305. doi:10.1016/S0012-821X(00)00321-6. Kitamura,A.,Kobayashi,K.,Tamaki,C.,Yamamoto,N.,Irino,T., Miyairi,Y.,andYokoyama,Y.,2013,Evidenceofrecentwarmingin theOkinawaregion,subtropicalnorthwesternPacific,fromanoxygen isotoperecordofacave-dwellingmarinemicro-bivalve:PaleontologicalResearch,v.17,p.58.doi:10.2517/1342-8144-17.1.58 Kitamura,A.,Yamamoto,N.,Kase,T.,Ohashi,S.,Hiramoto,M., Fukusawa,H.,Watanabe,T.,Irino,T.,Kojitani,H.,Shimamura,M., andKawakami,I.,2007,Potentialofsubmarine-cavesedimentsand oxygenisotopecompositionofcavernicolousmicro-bivalveasalate Holocenepaleoenvironmentalrecord:GlobalandPlanetaryChange, v.55,p.301.doi:10.1016/j.gloplacha.2006.09.002. Kubota,Y.,Kimoto,K.,Tada,R.,Oda,H.,Yokoyama,Y.,and Matsuzaki,H.,2010,VariationsofEastAsiansummermonsoonsince thelastdeglaciationbasedonMg/Caandoxygenisotopeofplanktic foraminiferainthenorthernEastChinaSea:Paleoceanography,v.25, PA4205,doi:10.1029/2009PA001891. Lin,Hui-Ling.,Wang,Wei-Chial,andHung,Gwo-Wei,2004,Seasonal variationofplanktonicforaminiferalisotopiccompositionfrom sedimenttrapsintheSouthChinaSea:MarineMicropaleontology, v.53,p.447.doi:10.1016/j.marmicro.2004.08.004. Lin,Yu-Shih,Wei,Kuo-Yen,Lin,In-Tian,Yu,Psi-Sen,Chiang,HongWei,Chen,Chen-Yin.,Shen,Chaun-Chou,Mii,Homg-Sheng,and Chen,Yue-Gau.,2006,TheHolocene Pulleniatina MinimumEvent revisited:geochemicalandfaunalevidencefromtheOkinawaTrough andupperreachesoftheKuroshiocurrent:MarineMicropaleontology,v.59,p.153.doi:10.1016/j.marmicro.2006.02.003. Ohmori,K.,Watanabe,T.,Shirai,K.,andKan,H.,2008,SkeletalmicrostructuresandthevarietyoftraceelementsinPacificsclerosponges: ChikyuMonthly,v.30,p.322(inJapanese). Omori,A.,Kitamura,A.,Fujita,K.,Honda,K.,andYamamoto,N., 2010,Reconstructionoflightconditionswithinasubmarinecave duringthepast7,000yearsbasedonthetemporalandspatial distributionofalgalsymbiont-bearinglargebenthicforaminifers: Palaeogeography,Palaeoclimatology,Palaeoecology,v.292, p.443.doi:10.1016/j.palaeo.2010.04.004. Parravicini,V.,Guidetti,P.,Morri,C.,Montefalcone,M.,Donato,M., andBianchia,C.N.,2010,Consequencesofseawatertemperature anomaliesonaMediterraneansubmarinecaveecosystem:Estuarine, CoastalandShelfScience,v.86,p.276.doi:10.1016/j.ecss. 2009.11.004. Reimer,P.J.,Bard,E.,Bayliss,A.,Beck,J.W.,Blackwell,P.G.,Bronk Ramsey,C.,Buck,C.E.,Cheng,Hai,Edwards,R.L.,Friedrich,M., Grootes,P.M.,Guilderson,T.P.,Haflidason,H.,Hajdas,I.,Hatte C.,Heaton,T.J.,Hoffmann,D.L.,Hogg,A.G.,Hughen,K.A., Kaiser,K.F.,Kromer,B.,Manning,S.W.,Niu,M.,Reimer,R.W., Richards,D.A.,Scott,E.M.,Southon,J.R.,Staff,R.A.,Turney, C.S.M.,andvanderPlicht,J.,2013,IntCal13andMarine13 radiocarbonagecalibrationcurves,0,000yearscalBP:Radiocarbon,v.55,no.4,p.1869.doi:10.2458/azu_js_rc.55.16947. Reitner,J.,andGautret,P.P.,1996,Skeletalformationinthemodernbut ultraconservativechaeteidsponge Spirastrella ( Acanthochaetetes ) wellsi (Demospongiae,Porifera):Facies,v.34,p.193. doi:10.1007/BF02546164. Sun,Youbin,Oppo,D.W.,Xiang,Rong,Liu,Weiguo,andGao,Shu, 2005,LastdeglaciationintheOkinawaTrough:subtropicalnorthwest PacificlinktoNorthernHemisphereandtropicalclimate:Paleoceanography,v.20,PA4005,doi:10.1029/2004PA001061. Tuccimei,P.,Soligo,M.,Gines,J.,Gines,A.,Fornos,J.,Kramers,J.,and Villa,I.G.,2010,ConstrainingHolocenesealevelsusingU/Thagesof phreaticovergrowthsonspeleothemsfromcoastalcavesinMallorca (WesternMediterranean):EarthSurfaceProcessesandLandforms, v.35,p.782.doi:10.1002/esp.1955. vanHengstum,P.J.,andScott,D.B.,2011,Ecologyofforaminiferaand habitatvariabilityinanunderwatercave:distinguishinganchialine versussubmarinecaveenvironments:JournalofForaminiferal Research,v.41,p.201.doi:10.2113/gsjfr.41.3.201. vanHengstum,P.J.,andScott,D.B.,2012,Sea-levelriseandcoastal circulationcontrolledHolocenegroundwaterdevelopmentinBermudaandcausedameteoriclenstocollapse1600yearsago:Marine Micropaleontology,v.90,p.29.doi:10.1016/j.marmicro.2012. 02.007. vanHengstum,P.J.,Scott,D.B.,Grocke,D.R.,andCharette,M.A.,2011, Sealevelcontrolssedimentationandenvironmentsincoastalcaves andsinkholes:MarineGeology,v.286,p.35.doi:10.1016/ j.margeo.2011.05.004. Wang,Yongjin,Cheng,Hai,Edwards,R.L.,He,Yaoqi,Kong,Xinggong, An,Zhisheng,Wu,Jiangying,Kelly,M.J.,Dykoski,C.A.,andLi, Xiangdong,2005,TheHoloceneAsianmonsoon:linkstosolar changesandNorthAtlanticclimate:Science,v.308,p.854. doi:10.1126/science.1106296. Xiang,R.,Sun,Y.,Li,T.,Oppo,D.W.,Chen,M.,andZheng,F.,2007, PaleoenvironmentalchangeinthemiddleOkinawaTroughsincethe lastdeglaciation:Evidencefromthesedimentationrateandplanktonicforaminiferalrecord.Palaeogeography,Palaeoclimatology, Palaeoecology243,p.378.doi:10.1016/j.palaeo.2006.08.016. Yamamoto,N.,Kitamura,A.,Irino,T.,Kase,T.,andOhashi,S.,2008, ReconstructionofpaleotemperaturesinNorthwestPacificoverthe past3,000yearsfrom d18Ovaluesofthemicro-bivalvia Carditella iejimensis foundinasubmarinecave:GlobalandPlanetaryChange, v.62,p.97.doi:10.1016/j.gloplacha.2007.12.001.A.KITAMURA,C.TAMAKI,Y.MIYAIRI,Y.YOKOYAMA,ANDH.MORIJournalofCaveandKarstStudies, December2014 N 171

PAGE 20

Yamamoto,N.,Kitamura,A.,Irino,T.,Kase,T.,andOhashi,S.,2010, ClimaticandhydrologicvariabilityintheEastChinaSeaduringthe last7,000yearsbasedonoxygenisotoperecordsofthesubmarine cavernicolousmicro-bivalve Carditellaiejimensis :GlobalandPlanetaryChange,v.72,p.131.doi:10.1016/j.gloplacha.2010.01.025. Yamamoto,N.,Kitamura,A.,Ohmori ,A.,Morishima,Y.,Toyofuku,T., andOhashi,S.,2009a,Long-termchangesinsedimenttypeand cavernicolousbivalveassemblagesinDaidokutsusubmarinecave, OkinawaIslands:evidencefromanewcoreextendingoverthepast 7,000years:CoralReefs,v.28,p.967.doi:10.1007/s00338-009-0536-2. Yamamoto,N.,Sakai,S.,andKitamura,A.,2009b,Evaluationofthe d18Ovalueofthesubmarinecavernicolousmicro-bivalve Carditella iejimensis asaproxyforpalaeotemperature:PaleontologicalResearch, v.13,p.279.doi:10.2517/1342-8144-13.3.279. Yoneda,M.,Uno,H.,Shibata,Y.,Suzuki,R.,Kumamoto,Y.,Yoshida, K.,Sasaki,T.,Suzuki,A.,andKawahata,H.,2007,Radiocarbon marinereservoiragesinthewesternpacificestimatedbyprebomb molluscanshells:NuclearInstrumentsandMethodsinPhysics ResearchB:BeamInteractionswithMaterialsandAtoms,v.259, p.432.doi:10.1016/j.nimb.2007.01.184.POSSIBLECAVERN-FORMINGACTIVITYATMILLENNIALTIMESCALESANDITSIMPACTONVARIATIONSINSUBMARINECAVEENVIRONMENTSANDHABITAT AVAILABILITY,OKINAWA,JAPAN172 N JournalofCaveandKarstStudies, December2014

PAGE 21

MODERNPOLLENRECORDONBATGUANODEPOSIT FROMSIJUCAVEANDITSIMPLICATIONTO PALAEOECOLOGICALSTUDYINSOUTHGAROHILLSOF MEGHALAYA,INDIASADHANK.BASUMATARY*ANDSAMIRK.BERABirbalSahniInstituteofPalaeobotany,QuaternaryPalynologyDivision,53,UniversityRoad,Lucknow-226007,UttarPradesh,IndiaAbstract: Meghalayaiswellknownforitsrichtropicalbiodiversityandnumerous naturalcaves.TheSijuCaveofMeghalaya,alsoknownascaveofthebat,isamongthe longestcavesontheIndiansubcontinent.Thepalynologicalstudyoffiftybat-guano samplesfromSijuCaveandofthirtysurfacesoilandmosscushionfromthearea immediatesurroundingthecavereflectstheclosesimilaritybetweenthemodernpollen andvegetationintheregioninthetwoareassampled.Theresultingpalynodatacomprise mainlythenativeflora,dominatedbyripariantaxalike Duabanga Syzygium Careya and Ficus, alongwithevergreenanddeciduouselementsintheregion.Theevergreentaxa Mesua Elaeocarpus ,and Garcinia ,alongwith Impatiens ,reflectthehighprecipitationin theregion.Theclimateoftheregionandthevegetationarestronglyinfluencedbythe SimsangRiver,andtheheavyrainfallresultsinthedominanceofripariantaxaandother high-rainfallindicatortaxa.Theoccurrenceofpollenfrom Nepentheskhasiana,an endemicandendangeredplantofMeghalaya,inbatguanoissignificantandmaybedue toitsinsectivorousandentomophilousnature.However,medicinalplantslike Swertia chirata Cinchona ,and Rauvolfia arenotencounteredinthebat-guanosediments,despite theirluxuriantgrowtharoundthecave,andthesespeciescouldbeavoidedbytheinsects duetotheiralkaloidcontainandbittertaste.Thepresenceofhighlandtaxasuchas Pinus Abies Picea ,and Larix inthebatguanodepositissignificantandsuggestiveof highwindsfromhigheraltitudes,orthepollenmaybeintroducedbythereturnof migratorySiberianbirdsduringwintertothenearbySijuwildlifeandbirdsanctuary. Therecoveryofcerealiaalongwith Arecacatechu and Citrus pollenindicatethehuman activityintheregion.Theabundanceoffungalremains,namely Meliola Glomus ,and Microthyriaceaealongwithdegradedpalynomorphsaresuggestiveofstrongmicrobial activityunderwarmandhumidconditionsduringsedimentationintheregion.Themain objectiveofthisstudyistoidentifythepotentialofbatguanoforpalaeoecological researchandassupportivedataforsurfaceandsedimentarysoilprofilesintheSouth GaroHillsofMeghalaya.ThepalynodatafromthebatguanooftheSijuCaveprovides ausefulsourceforpalaeoecologicalinformationfortheSouthGaroHills,where intensivenaturalandhuman-causedforestfires,heavyrainfall,andsoilerosionoccur everyyearandthereisarelativescarcityofthelake,wetland,andswamphabitatsthat normallypreservepollen.INTRODUCTIONThestateofMeghalayaincludesasignificantportionof bothHimalayaandIndo-Burmabiodiversityhotspots (Mittermeieretal.,2005)andisgloballyknownfor Cherrapunjee,withthehighestrainfallonearthandthe endemicandendangeredplant Nepentheskhasiana Hk.f., thesymbolicplantofMeghalaya.ThefloraofMeghalaya istherichestinIndiaandprobablyinthewholeofAsia, withaclosesimilaritytothefloraofsoutheastAsiaand southernChina(Hooker,1905).Thisregionoftheworldis consideredbybotanistsandgeographersasoneofthe nuclearareasofearlyplantdomestication(Vivilov,1951; Sauer,1952;Harris,1972).ThewholeoftheGaroHills, especiallytheSouthGaroHills,isalsoknownasthe ecologicalcanvasofMeghalayabecauseofitsunique biodiversityandnumerousbeautifulnaturalcaves(Anonymous,2006).Thequantityandlengthofcavesin Meghalayafarexceedsthatofanyotherknownkarst regionofIndia.Untilrecentlyonlyafewcaveshadbeen explored,butrecentexplorationhasresultedinover 320kilometersofcavepassagemappedandovera thousandcaveentrancesdocumentedinMeghalaya (Harriesetal.,2008).Ascavesarenaturalopencavities intheearth,theyfunctionasnaturalsedimenttraps *Correspondingauthor,sbasumatary2005@yahoo.co.inS.K.BasumataryandS.K.BeraModernpollenrecordonbatguanodepositfromSijuCaveanditsimplicationtopalaeoecological studyinSouthGaroHillsofMeghalaya,India. JournalofCaveandKarstStudies, v.76,no.3,p.173.DOI:10.4311/2013PA0119JournalofCaveandKarstStudies, December2014 N 173

PAGE 22

(White,2007).Incaveswithlargebatpopulations,guano (batexcreta)ispresentinsufficientquantitiestobe classifiedasaclasticsedimentandservesasasourceof usefulclimaticrecords(LeroyandSimms,2006;White, 2007).Ingeneral,therearemainlythreemechanismsfor theincorporationofpollenintobatguano:batscan consumelargenumbersofflyinginsects,mayconsume pollenthemselves,andmayflythroughapollen-laden environmentsothatpollenandotherdustparticlesoften adherestotheirbody(Pendletonetal.,1996).Pollenin fragmentsofbatskinandhairmaybelostduring grooming,anddustcontainingpollenmaybebrought intothecaveanddepositedontoguanobyaircurrents (ColesandGilbertson,1994).Wheretherearelarge coloniesofbatstheirexcrementaccumulatesonthecave floorbelow(Maher,2006). Manypalaeoecological-researchstudieshavebeen carriedoutonpalynologyofcavesediments,including batguanoandcoprolitesfromothertaxaandsurfacesoil indifferentpartoftheglobe,suchasinRomania(Popand Ciobanu,1950;BoscaiuandLupsa,1967a,b;Feurdeanet al.,2011;Geantaetal.,2012),theUnitedStates(Searsand Roosma,1961;Davis,1990;Nieves-Rivera2003;Maher, 2006;BatinaandReese,2011),Switzerland(Groner,2004), Austria(Kral,1968;Draxler,1972),theUK(Colesetal., 1989;McGarryandCaseldine,2004),Spain(Carrio n, 1992;Carrionetal.,1999,2006;NavarroCamachoetal., 2000,2001),Belgium(Bastin,1978;Bastinetal.,1986), Nepal(Denniston,etal.,2000),China(Qinetal.,1999; Zhangetal.,2004),andotherpartsofAsia(Huntand Rushworth,2005).InIndia,nopreviousstudieshave addressedtheuseofbatguanoincavesediments forpalaeoecologicalstudy.WehavediscoveredtheSiju Cavetobeoneofthebestsitestostudybatguanofor palynologicalresearchintheSouthGaroHillsof Meghalaya.IntheGaroHills,intensivenaturaland human-causedforestfireseveryyearburnseveralcentimetersofsoilintheregion(Fig.2F).Inaddition,other factorslikethehighrainfallandtremendoussoilerosion, alongwiththehillyterrain,makedifficulttheproper investigationandinterpretationofmodernpollenand vegetationrelationshipsintheGaroHills.Therelatively scarcityoflake,wetland,andswampshaslimitedattempts torecoversedimentcoresfromtheGaroHillsfor palaeoecologicalresearch.AlthoughtherearesomescatteredsmalllakeslikeNapakLakeintheSouthGaroHills, thesearenotsuitableforpropercoringforpalynological study.Previously,onlyafewpreliminarypalynological studieshavebeencarriedoutinMeghalaya(Guptaand Sharma,1985;BasumataryandBera,2007,2010,2012, Basumataryetal.,2013).Consideringthesepreviouslimits topollenstudiesintheregion,weinitiatedthepresent palynologicalstudyonbatguanoinSijuCaveintheSouth GaroHills.Themainobjectiveofthestudy,asthefirstto beundertakenasanewinnovationinpollenstudiesin India,wastoidentifythepotentialofbatguanoasa reliablesourcetodocumentmodernpollenandvegetation inrelationtoclimateintheSouthGarohills.PHYSIOGRAPHYANDSOILTheSouthGaroHillsofMeghalayaareaunique featureofphysiographywheretheTuraandArabella Rangesrunparallelinaneast-westdirection.TheTura RangerunsfromSijutoTura,andtheArabellaRangeis tothenorthoftheTurarangeandgraduallyincreases inheight,eventuallyjoiningtheTura,whichstartsin theWestGaroHillstothesouth.TheGaroHillsof Meghalayaaredrainedbymanyriversandstreamlets.The soiloftheregionisgenerallyred-loamy,butsometimes variesfromclaytosandyloam,andisrichinorganic carbonwithhighnitrogen-supplyingpotentialbutis deficientinphosphorusandpotassium.ThesoilpHranges fromacidic(pH5.0to6.0)tostronglyacidic(pH4.5to5.0) (DirectorateofAgriculture,Meghalaya,2012).CLIMATETheclimateoftheregioniscontrolledbysouth-west andnorth-easternmonsoons.Itiswarmandhumidin summerandcoldanddryinwinter.Themaximum temperatureduringsummeris36 u Candtheminimumin winteris4u C.Therelativehumidityrangesfrom70to98%. Rainfalloftheregionrangesfrom3900to6800mmyr2 1(DirectorateofAgriculture,Meghalaya,2012).STUDYSITEANDVEGETATIONThestudysite,SijuCaveintheSouthGaroHills,is locallyknownasDobakol,caveofthebat,asitisthehome forthousandsofbats.Accordingtofieldobservationand previousrecords(Sinha,1994,1999;BatesandHarrison, 1997)inSijuCave,thedominantbatspeciesare Rhinolophussubbadius R.pusillus R.pearsoni ,and Miniopterusschreibersii ;thesearemainlyinsectivorous. TheSijuCaveissituatedatlat.25 u 21 9 16.17 0 Nandlong. 90 u 41 9 08.09 0 Eat95masl(Fig.1).Itwasfirstexploredby theBritishGeologicSurveyin1920.Thiscaveisoneofthe longestcavesintheIndiansubcontinentandcontainssome ofthefinestriverpassagestobefoundanywhereinthe world.Thecavehasmanyunexploredchambersand labyrinths(Fig.2A).Themajesticformationsinsidethe cave,especiallyinPrincessDisChamber,areamongthe specialattractionsofthecave.Inthenortheastdirectionof SijuCave,SimsangRiver(Fig.2B),originatesfromthe Turapeakareaandsupportsluxuriantgrowthofmarshy andaquaticplants.NaphakLakeandtheSijuwildlifeand birdsanctuaryarelocatedclosetoSijuCave.During winter,Siberianducksmigratetothebirdsanctuary. PreviousstudiesontheSijuCavefaunaarelistedin Table1,butthisisthefirstattemptofapalynological studyonbatguanoinrelationtomodernpollen,MODERNPOLLENRECORDONBATGUANODEPOSITFROMSIJUCAVEANDITSIMPLICATIONTOPALAEOECOLOGICALSTUDYINSOUTHGAROHILLSOFMEGHALAYA,INDIA174 N JournalofCaveandKarstStudies, December2014

PAGE 23

Figure1.MapshowingthelocationofSijuCaveintheSouthGaroHills,Maghalaya.Thenumbersshowthelocationsofthe thirtysurfacesamples.S.K.BASUMATARYANDS.K.BERAJournalofCaveandKarstStudies, December2014 N 175

PAGE 24

Figure2.A.TheentranceofSijuCave,SouthGaroHills,Meghalaya.B.ThehangingbridgeovertheSimsangRiver connectingtheSijuCaveareaandtheSijuwildlifeandbirdsanctuary,Meghalaya.C.Aviewof Nepentheskhasiana growing luxuriantly,intermixedwith Gleicheniadichotoma ,intheentranceofSijuCave.D.Aviewofthebatcolonyintheirresting chamberinsideSijuCave.E.BatguanocollectioninsideSijuCave.F.AviewoftheforestfloorburningduringwinterinSouth GaroHills.MODERNPOLLENRECORDONBATGUANODEPOSITFROMSIJUCAVEANDITSIMPLICATIONTOPALAEOECOLOGICALSTUDYINSOUTHGAROHILLSOFMEGHALAYA,INDIA176 N JournalofCaveandKarstStudies, December2014

PAGE 25

vegetation,andclimate,anditwillbegreatlyhelpfulto furtherpalaeoecologicalresearchintheSouthGaroHillsof Meghalayaandtoplacethisregionwithinaglobalcontext. Thevegetationoftheregionisdominatedbyriparian taxa,including Duabangagrandiflora Syzygiumpolypetalum Terminaliabellirica Schimawallichii Careyaarborea and Ficuspyriformis ,intermixedwithevergreenand deciduouselementssuchas Mesuaferrea Elaeocarpus rugosus Garciniapaniculata Dilleniapentagyna Emblica officinalis Artocarpuschaplasha ,and Adhatodavasica .The endemicpitcherplant Nepentheskhasiana, locallyknown asmemang-kokshibytheGaropeople,growsintermixed withthepteridophyticflora(Fig.2C).Thepteridophytes aredominatedby Lycopodiumclavatum Selaginella selaginoides Dryopterisfilixmas Gleicheniadichotoma Adiuntumcaudatum Cyatheagigantea Angiopterisevecta and Blechnumoccidentale .Theforestfloorisformedby densehumus,islitter-laden,andisoftencoveredbygrasses suchas Saccharumspontaneum S.arundinaceum Cynodon dactylon ,and Neyraudiareynaudiana associatedwith sedgeslike Mariscussumatrensis Kyllingamonocephale Fimbristylisdichotoma, Cyperuszollingeri ,and Scleria terrestris .Duringtherainyseason,Zingiberaceaelike Curcumaaromatica andotherdicotyledonousherbssuchas Justiciasimplex Rungiapectinata Amaranthusaspera Evolvulusnummularius ,and Polygonumorientale are presentinlargenumbers.MATERIALSANDMETHODSThelocationofSijuCavewasrecordedbyGPSwith coordinatesbasedonWGS1984.Thecavewassurveyedin detail,andfiftysamplesofapproximately50gofbat guanowerecollectedrandomlyfromthefloorofthebats restingchamberinsidethecave(Figs.2D,E).Atotalof thirtysurfacesamplesofbothsoilandmosscushionwere collectedrandomlyfromtheareaimmediatelyoutsideSiju Cave.Wealsocollectedpolliniferousplantmaterialsfor properidentificationoftaxaintheregionbytheirpollen morphology.ThespecimenswereplacedintheBirbal SahniInstituteofPalaeobotany(BSIP)Herbarium. Thebatguanoandsurfacesampleswereprocessed employingthestandardacetolysismethod(Erdtman, 1953).Thesamplesweretreatedwith10%aqueousKOH solutiontodeflocculatethepollenandsporesfromthe sedimentsfollowedby40%HFtreatmenttodissolvesilica content.Thentheconventionalprocedureofacetolysiswas followedusingtheacetolysismixture9:1anhydrousacetic acidandconcentratedH2SO4.Finallythematerialwas keptina50%glycerinsolutionwithadropofphenol. Between450and600pollenandsporespersamplewere countedtoproducethepollenspectra.Plantelementsin thestudywerecategorizedintoarboreals(trees,shrubs, andepiphytes),nonarboreals(terrestrialherbsandmarshy taxa),highlandtaxa,ferns,andfungalremainswith degradedpalynomorphs.Forthepreciseidentificationof fossilpalynomorphsinthesediments,weconsultedthe referencepollenslidesavailableatBirbalSahniInstituteof Palaeobotany(BSIP)herbariumofIndia,aswellasthe pollenphotographsinthepublishedliterature(Chauhan andBera,1990;Nayar,1990;Beraetal.,2009).Photodocumentationofthepalynomorphswasmadeusingan OlympusBX-61microscopewithDP25digitalcamera under403magnification(Fig.3).Thepollenspectrawere madeusingMicrosoftExcelprogramandmodifiedin CorelDraw-12software.Thefrequencypercentageofthe recoveredpalynomorphshasbeencalculatedintermsof thetotalpalynomorphs.RESULTSThemodernpollenspectraofbatguanofromSiju Caveandfromthesurfacesoilandmosscushionfromits immediatesurroundingsarediscussedbelowandprovide anoverviewofmodernpolleninrelationtovegetation andclimateintheregion.Thetotaloffiftymodernbat Table1.ListoftherecordedfaunaltaxafromSijuCave, SouthGaroHills,Meghalaya,withreferences. RecordedTaxa Reference Coleoptera Andrewes(1924) Molluscas AnnandaleandC hopra(1924) Coleoptera Blair(1924) Diptera Brunetti(1924) Coleoptera Cameron(1924) Collembola Carpenter(1924) Diptera Edwards(1924) Araneids Fage(1924) Lepidoptera Fletcher(1924) Coleoptera Fleutiaux(1924) Tartarides Gravely(1924) Pisces Hora(1924) Rhynchota KempandChina(1924) Diptera Lamb(1924) Lepidoptera Meyrick(1924) Gyrinidae Ochs(1925) Diptera Patton(1924) Hymenoptera Rohwer(1924) Myriapoda Silvestri(1924) Oligochaeta Stephenson(1924) Hymenoptera Wheeler(1924) Chiroptera BatesandHarrison1997 Chiroptera Sinha(1999) Teleostei Kottelatetal.(2007) Arachnida,Brachyura, Palaemonidae,Isopoda, Diplopoda,Orthoptera, Dictyoptera,Coleopteran, DipteraandPiscesHarriesetal.(2008) Diptera Disney(2009)S.K.BASUMATARYANDS.K.BERAJournalofCaveandKarstStudies, December2014 N 177

PAGE 26

Figure3.Palynoassemblagerecoveredfr omthebatguanoandsurfacesamplefromSouthGaroHills,Meghalaya.1. Mesua ,2.Schima ,3.Emblica ,4.Syzygium ,5.Terminalia ,6.Albizia ,7.Duabanga,8.Dipterocarpaceae,9. Areca,10. Elaeocarpus ,11. Dillenia ,12. Dendropthoe ,13. Nepenthes,14.Pinus ,15. Quercus ,16. Alnus ,17. Betula ,MODERNPOLLENRECORDONBATGUANODEPOSITFROMSIJUCAVEANDITSIMPLICATIONTOPALAEOECOLOGICALSTUDYINSOUTHGAROHILLSOFMEGHALAYA,INDIA178 N JournalofCaveandKarstStudies, December2014

PAGE 27

guanosamplescollectedrandomlyfrominsidethecave aredominatedbyarboreals(42%),followedbynonarboreals(35%),fungalremainswithdegradedpalynomorphs (10%),ferns(6%),andhighlandtaxa(5%)(Fig.4).The majorarborealsinclude Duabanga Schima Syzygium and Careya andarerepresentedbymaximumvaluesupto 10%.Otherassociatedtaxa,suchas Schima Semecarpus Albizia Adhatoda,and Nepenthes havevaluesof1to6%. NonarborealslikePoaceae,Cyperaceae,andTubuliflorae havevaluesof1to15%.Ferns,bothmonoleteandtrilete, exhibitanaveragevalueof4%and5%respectively. FungalremainslikeMicrothyriaceae, Glomus Diplodia and Meliola arerepresentedbymaximumvaluesupto6%. Amongthehighlandtaxa, Pinus isrecordedatvaluesof 2%,whereasitsassociatedtaxalike Abies and Picea are onlyencounteredintraceamounts(Fig.4). Thethirtysurfacesoilandmosscushionssamples collectedrandomlyfromtheimmediatevicinityofthe cavearedominatedbyarboreals(40%),followedby nonarboreals(37%),ferns(12%),fungalremainswith degradedpalynomorphs(7%),andhighlandtaxa(4%) (Fig.5).Themajorarborealsinclude Duabanga, Syzygium ,Anacardiaceae,Moraceae,and Careya andare representedbymaximumvaluesupto8%.Theother associatedtaxasuchas Mesua Elaeocarpus Schima Dillenia Terminalia Adhatoda,and Nepenthes havevalues of1to5%.AmongnonarborealslikePoaceae,Tubuliflorae,Onagraceaeand Impatiens thevaluesrange between1to18%.Ferns,bothmonoleteandtrilete,have anaveragevalueof5%and6%respectively.Fungal remainslikeMicrothyriaceae, Cookeina Tetraploa, and Glomus arerepresentedbyamaximumvalueof5%.The highlandtaxa Pinus isrecordedatthevalueof3%, whereasassociatedtaxa Abies and Picea arerepresented byvariablevalues.DISCUSSIONTherearesomedifferencebetweenpollenpercentages preservedinthebatguanoandthesurfacesamplesfrom thevicinityofSijuCave,although,ingeneral,the palynodatareflectaclosesimilaritybetweenthebatguano andsurfacesamples:(i)thepollendiversityfrombatguano ishigherthanthesurfacesamples,(ii)thenumberof arborealtaxaishigherinthebatguanosamples,(iii)the Dipterocarpaceaepollenismarkedlypresentinthebat guano,butisnotfoundinthesamplescollectedfromthe immediatevicinityofthecave,(iv)thefernsporesare Figure4.PollenspectraofbatguanosamplesfromSijuCave,SouthGaroHills,Meghalaya. r 18.Chenopodiaceae,19.Convolvulaceae,20.Tubuliflorae,21.Liguliflorae,22.Poaceae,23.Cerealia,24.Cyperaceae,25. Polygonaceae,26.Monolete,27.Trilete,28. Lycopodium ,29. Glomus ,30. Meliola .S.K.BASUMATARYANDS.K.BERAJournalofCaveandKarstStudies, December2014 N 179

PAGE 28

comparativelyfewinthebatguanosamples,and(v)the fungalremains,alongwithdegradedpalynomorphs,are comparativelyhigherinthebatguanosamples.RelationshipofPalynoassemblage,Vegetation,and ClimateThebatguanofromSijuCaveandtheforestsurface samplesofmosscushionandsoilcollectedfromthe immediatevicinitybothcontainpollenindicativeofthe tropicalriparianforestintermixedwithevergreenand deciduouselementsunderwarmandhumidclimatic condition.Amongthearboreals,thesamplesprincipally reflecttheproximityofripariantaxalike Duabanga Syzygium Careya ,and Ficus inthepalynoassemblagethat growsalongtheedgeoftheriverSimsang.Theassociated evergreenanddeciduouselementslike Mesua Elaeocarpus Garcinia Schima Dillenia Albizia ,andSapotaceaealso exhibit30to42%totalsindicativeoftheheavyrainfallin theregion.Thepresenceofevergreentaxa,alongwith PiperaceaeandEuphorbiaceae,suggestsheavyrainfallin theregion(Nairetal.,2010),whichisreflectedbythe observedpalynoassemblageinthesedimentsfromboth insideandoutsideofthecave.Theoccurrenceof Dendrophthoe (epiphyticplants)polleninthepalynoassemblageissignificantandreflectstheexistenceofaprimary forestthatreceivesheavyrainfallintheregion.The presenceof Areca (betelnut)pollenalongwithcerealia and Citrus (orange)arestronglyindicativeofthehuman activityinthearea.Thepresenceofthehighlandtaxa Pinus Betula Abies Picea ,and Larix inthebatguano depositissignificantandsuggestiveofstrongwindsfrom higheraltitudes,ortheymayhavebeenintroducedbythe migrationofSiberianbirdstothenearbySijuwildlifeand birdsanctuaryduringwinter.Thepalynodataofthe surfacesamplesfromtheimmediatevicinityofthecave alsoreflecttheriparianforestandareanexactmatchwith thecavesampleofbatguano.Thatboththevegetationand climateoftheregionarestronglygovernedbytheSimsang River,alongwiththeheavyrainfall,isindicatedbythe dominanceoftheripariantaxa Duabanga Ficus ,and Careya andotherheavy-rainfallindicatortaxalike Mesua Syzygium ,and Elaeocarpus .Thesamplesfrominsidethe cavealsoreflectlocalandregionalflorasquitewell(Burney andBurney,1993;Carrionetal.,2006),andthepollen containedinthebatguanomatcheswellthatinthesurface sedimentsoflakeandpeat(Maher,2006)thatsupportsour palynodata.Theabundanceof Nepentheskhasiana inbat guanoissignificantandsuggeststhepresenceofinsectivorousandriparianhabitatgrowinginverylimitedpockets inandaroundthearea.Thepresenceof Nepenthes pollenis stronglysuggestiveofheavyrainfallandaperennialwater systemintheregion,because Nepentheskhasiana generally growsintheshadealongriversandstreamletsin Meghalaya(HaridasanandRao,1985).Thepreservation ofDipterocarpaceaepolleninthebatguanoisinformative withregardtothesizeoftheareassampledbythebats,as itdoesnotgrowneartheSijuCaveandtheclosestplants areseveralkilometersdistancefromthestudyarea.Leroy Figure5.PollenspectraofsurfacesamplesfromtheimmediatevicinityofSijuCave,SouthGaroHills,Meghalaya.MODERNPOLLENRECORDONBATGUANODEPOSITFROMSIJUCAVEANDITSIMPLICATIONTOPALAEOECOLOGICALSTUDYINSOUTHGAROHILLSOFMEGHALAYA,INDIA180 N JournalofCaveandKarstStudies, December2014

PAGE 29

andSimms(2006)notedthatbatscanmigratehundredsof kilometersfromtheirshelterandissupportedbyour pollendata,whichstronglysuggestthatthebatsflyseveral kilometersfortheirfood.Thefernspores,especially Cyathea Dryopteris,and Pteris ,inthebatguanoarelocal inoriginandgrewintheimmediatevicinityofthecave. Theirpresencemayalsobeduetotheinfluenceofwind activity.Caveswithalargeentranceandstrongair circulationarelikelytohavehigherpollenpresence resultingfromwindthancaveswithsmallerentrances andminimalaircirculation(BurneyandBurney,1993). However,theoccurrenceofterrestrialferns,especially Gleichenia Dryopteri s,and Lycopodium ,signifieshigh rainfallandhumidclimaticconditionsinthearea (Basumataryetal.,2013),whichisexactlyreflectedinthe palynoassemblageofthestudiedsamples.Theabundance offungalremainssuchas Meliola Cookeina ,Microthyriaceae,and Diplodia, alongwithdegradedpalynomorphs, isstronglyindicativeofwarmandhumidclimatic conditionduringsedimentationintheregion.Thepresence of Glomus withhyphaeinsurface-soilandmoss-cushion samplesisstronglyindicativeofhighsoilerosioninthe region.Medicinalplantslike Swertiachirata and Rauvolfia serpentina arenotencounteredinthebatguano,although theyarepresentintheluxuriantgrowthinthevicinityof thecave.Theirabsencemaybeduetotheirbitterintaste andavoidancebyinsectsthatmightinturnbeeatenbythe bats.CONCLUSIONSThepalynologicalstudyonbatguanofromSijuCaveis thefirsttobeconductedinIndia.Thepalynodatafrom boththemodernbatguanoandsurfacesamplesfromthe immediatevicinityofthecaveareverysimilarandindicate theexistenceofmainlyriparianforestintermixedwithboth evergreenanddeciduoustaxathatexactlycoincideswith theextantvegetation.ThebatguanodepositinSijuCave canbeconsideredasareliablesourceofpalaeoecological datathatcanbeusedtosupportdatafromsurfaceand sedimentarysoilprofilesandtosubstituteforthescarcity oflakes,swamps,andwetlandsinMeghalaya,withthe caveatthatbatscanforagelongdistancesfromtheirroost siteinthecaveandmayaccumulatepollenfromplantsnot foundintheimmediatevicinityofthecave.Lastly, multidisciplinarystudiesthatintegratepollendatafrom caveswithpaleontological,archaeological,zoological,and geologicaldatacouldplayanimportantroleinany palaeoecologicalstudyoftheSouthGaroHillsof Meghalayaandatthegloballevel.ACKNOWLEDGEMENTSAuthorsthanktotheDirector,BirbalSahniInstituteof Palaeobotany(BSIP),Lucknow,India,forinfrastructure facilityandpermissiontopublishthepaper.Wealsothank anumberofforestofficialsfortheirhelpduringfieldwork.REFERENCESAndrewes,H.E.,1924,ColeopteraoftheSijucave,GaroHills,Assam:I Carabidae:RecordsoftheIndianMuseum,v.26,p.115. Annandale,N.,andChopra,B.,1924,MolluscasoftheSijucave,Garo Hills,Assam:RecordsoftheIndianMuseum,v.26,p.33. Anonymous,2006,MeghalayaresourceF\BIOSPHERE_RESERVES. http://meghalaya.nic.in/tourism/cave.htm[accessed20November 2006]. Bastin,B.,1978,Lanalysepolliniquedesstalagmites:unenouvelle possibilitedapproachdesfluctuationsclimatiquesduQuaternaire: AnnalesdelaSocie te Ge ologiquedeBelgique,v.101,p.13. Bastin,B.,Cordy,J.-M.,Gewelt,M.,andOtte,M.,1986,Fluctuations climatiqueenregistreesdepuis125000ansdanslescouchesde remplissagedelaGrotteScladina(ProvencedeNamur,Belgique): BulletindelAssociationFrancaisepourletudeduQuaternaire,v.23, no.2,p.168.doi:10.3406/quate.1986.1808. Basumatary,S.K.,andBera,S.K.,2007,Modernpollen-sporeassemblage fromsedimentoftropicalmoistdeciduousforest,EastGaroHills, Meghalaya:JournalofPalynology,v.43,p.111. Basumatary,S.K.,andBera,S.K.,2010,Developmentofvegetationand climatechangeinWestGaroHillssincelateHolocene:pollen sequenceandanthropogenicimpact:JournaloftheIndianBotanical Society,v.89,p.143. Basumatary,S.K.,andBera,S.K.,2012,Vegetationsuccessionand climatechangeinwesternGaroHills,Meghalaya,Indiasince 11,643yearsBP:apalynologicalrecord:InternationalJournalof EarthScienceandEngineering,v.5,no.4,p.749. Basumatary,S.K.,Dixit,S.,Bera,S.K.,andMehrotra,R.C.,2013, ModernpollenassemblagesofsurfacesamplesfromCherrapunjee anditsadjoiningareas,Meghalaya,northeastIndia:Quaternary International,v.298,p.68. Bera,S.K.,Basumatary,S.K.,andDixit,S.,2009,Studiesonpollen morphologyandphenologicalcharacteristicsofsomeeconomically importantarborescenttaxaoftropicalforestlowerBrahmaputra Valley,Assam,NorthEastIndia:JournalofPalynology,v.43,p.1. Blair,K.G.,1924,ColeopteraoftheSijuCave,GaroHills,Assam:IV Histeridae,Hydrophilidae,Erotylidae,Lathridiidae,Tenebrionidae andHylophilidae:RecordsoftheIndianMuseum,v.26,p.120. Brunetti,E.,1924,DipteraoftheSijuCave,GaroHills,Assam:I. Tipulidae,Tabanidae,Anthomyidae,AcalyptrateMuscidaeand Phoridae:RecordsoftheIndianMuseum,v.26,p.99. Bates,P.J.J.,andHarrison,D.L.,1997,BatsoftheIndianSubcontinent, Kent,HarrisonZoologicalMuseum,268p. Batina,M.C.,andReese,C.A.,2011,AHolocenepollenrecordrecovered fromaguanodeposit:RoundSpringCavern,Missouri,USA:Boreas, v.40,no.2,p.332.doi:10.1111/j.1502-3885.2010.00186.x. Boscaiu,N.,andLupsa,V.,1967a,PalynologicalresearchintheGrota HaiducilorCaveneartheHerculaneumspa(Romania):Revue RoumainedeBiologie,SeriedeBotanique,v.12,p.137. Boscaiu,N.,andLupsa,V.,1967b,CercetaripalinologicenPesteraVeteranidinDefileulDunarii,ContributiiBotanice,p.39. Burney,D.A.,andBurney,L.P.,1993,Modernpollendepositionincave sites:experimentalresultsfromNewYorkState:NewPhytologist, v.124,p.523.doi:10.1111/j.1469-8137.1993.tb03844.x. Cameron,M.,1924,ColeopteraoftheSijuCave,GaroHills,Assam:II. NewspeciesofStaphylinidae:RecordsoftheIndianMuseum,v.26, p.118. Carpenter,G.H.,1924,CollembolaoftheSijuCave,GaroHills,Assam: RecordsoftheIndianMuseum,v.26,p.285. Carrion,J.S.,1992,LateQuaternarypollensequencefromCarihuela Cave,southeasternSpain:ReviewofPalaeobotanyandPalynology, v.71,p.37.doi:10.1016/0034-6667(92)90157-C. Carrion,J.S.,Munuera,M.,Navarro,C.,Burjachs,F.,Dupre,M.,and Walker,M.J.,1999,Thepalaeoecologicalpotentialofpollenrecords incaves:thecaseofMediterraneanSpain:QuaternaryScience Reviews,v.18,p.1061.doi:10.1016/S0277-3791(98)00002-X. Carrion,J.S.,Scott,L.,andMarais,E.,2006,Environmentalimplications ofpollenspectrainbatdroppingsfromsoutheasternSpainand potentialforpalaeoenvironmentalreconstructions:ReviewofPalaeo-S.K.BASUMATARYANDS.K.BERAJournalofCaveandKarstStudies, December2014 N 181

PAGE 30

botanyandPalynology,v.140,p.175.doi:10.1016/j.revpalbo. 2006.03.007. Chauhan,M.S.,andBera,S.K.,1990,Pollenmorphologyofsome importantplantsoftropicaldeciduoussal( Shorearobusta )forests, districtSidhi,MadhyaPradesh:Geophytology,v.20,no.1,p.30. Coles,G.M.,andGilbertson,D.D.,1994,Theairfall-pollenbudgetof archaeologicallyimportantcaves:CreswellCrags,England:Journalof ArchaeologicalScience,v.21,p.735.doi:10.1006/jasc.1994.1073. Coles,G.G.,Gilberton,D.D.,Hunt,C.O.,andJenkinson,R.D.S.,1989, Taphonomyandthepalynologyofcavedeposits:CaveScience,v.16, p.83. Denniston,R.F.,Gonzalez,L.A.,Asmerom,Y.,Sharma,R.H.,and Reagan,M.K.,2000,SpeleothemEvidenceforChangesinIndian SummerMonsoonPrecipitationovertheLast;2300Years.QuaternaryResearch,v.53,p.196. Davis,O.K.,1990,Cavesassourcesofbioticremainsinaridwestern NorthAmerica:Palaeogeography,Palaeoclimatology,Palaeoecology, v.76,p.331.doi:10.1016/0031-0182(90)90119-R. Disney,R.H.L.,2009,Scuttleflies(Diptera:Phoridae)fromcavesin Meghalaya,India:JournalofCaveandKarstStudies,v.71,no.1, p.81. Draxler,I.,1972,PalynologischeUntersuchungenanSedimentenausder Salzofen-hohleimTotenGebirge:AnnalendesNaturhistorischen MuseumsinWien,v.76,p.161. Edwards,F.W.,1924,DipteraoftheSijucave,GaroHills,Assam:II. Chironomidae:RecordsoftheIndianMuseum,v.26,p.107. Erdtman,G.,1953,AnIntroductiontoPollenAnalysis,NewYork, RonaldPress,239p. Fage,L.,1924,AraneidsfromtheSijuCave,GaroHills,Assam:Records oftheIndianMuseum,v.26,p.63. Feurdean,A.,Persoiu,A.,Pazdur,A.,andOnac,B.P.,2011,Evaluating thepalaeoecologicalpotentialofpollenrecoveredfromiceincaves:A casestudyfromScarisoaraIceCave,Romania:ReviewofPalaeobotanyandPalynology,v.165,p.1.doi:10.1016/j.revpalbo.2011. 01.007. Fletcher,T.B.,1924,LepidopteraoftheSijucave,GaroHills,AssamI Pyralidae:RecordsoftheIndianMuseum,v.26,113p. Fleutiaux,E.,1924,ColeopteraoftheSijucave,GaroHills,AssamIII Elateridae:RecordsoftheIndianMuseum,v.26,119p. Geanta,A.,Tantau,I.,Tamas,T.,andJohnston,V.E.,2012,Palaeoenvironmentalinformationfromthepalynologyofan800yearoldbat guanodepositfromMaguriciCave,NWTransylvania(Romania): ReviewofPalaeobotanyandPalynology,v.174,p.57. doi:10.1016/j.revpalbo.2011.12.009. Gravely,F.H.,1924,TartaridesfromtheSijucave,GaroHills,Assam: RecordsoftheIndianMuseum,v.26,p.61. Groner,U.,2004,Palynologyandsedimentdatafromthehighalpine karstcaveonJungfraujoch,Switzerland:EclogaeGeologicaeHelveticae,v.97,p.237.doi:10.1007/s00015-004-1122-9. Gupta,H.P.,andSharma,C.,1985,Pollenanalysisofmodernsediments fromKhasiandJaintiahills,Meghalaya,India:JournalofPalynology,v.21,p.167. Haridasan,K.,andRao,R.R.,1985,ForestFloraofMeghalaya,volI, Dehradun,India,BishenSinghMahendraPalSingh,p.1. Harris,D.R.,1972,Theoriginofagricultureinthetropics:American Scientist,v.60,p.180. Harries,D.B.,Ware,F.J.,Fischer,C.W.,Biswas,J.,andKharpran-Daly, B.D.,2008,AreviewofthebiospeleologyofMeghalaya,India: JournalofCaveandKarstStudies,v.70,no.3,p.163. Hooker,J.D.,1905,HimalayanJournals:NotesofaNaturalistinBengal, theSikkimandNepalHimalayas,theKhasiaMountains&c,London, Ward,Lock,BowdenandCo.,639p. Hora,S.L.,1924,FishoftheSijucave,GaroHills,Assam:Recordsofthe IndianMuseum,v.26,p.27. Hunt,C.O.,andRushworth,G.,2005,Pollentaphonomyandairfall sedimentationinatropicalcave:theWestMouthofTheGreatCave ofNiahinSarawak,MalaysianBorneo:JournalofArchaeological Science,v.32,p.465.doi:10.1016/j.jas.2004.11.005. Kemp,S.,andChina,W.E.,1924,RhynchotaoftheSijucave,GaroHills, Assam:RecordsoftheIndianMuseum,v.26,p.93. Kottelat,M.,Harries,D.R.,andProudlove,G.S.,2007, Schistura papalifera ,anewspeciesofcaveloachfromMeghalaya,India (Teleostei:Balitoridae):Zootaxa,no.1393,p.35. Kral,F.,1968,PollenanlytischeUntersuchungenzurFragedesAltersder EisbildungeninderDachstein-Rieseneishohle:DieHohle,v.19, p.41. Lamb,C.G.,1924,DipteraoftheSijucave,GaroHills,AssamIII Dolichopodidae:RecordsoftheIndianMuseum,v.26,p.108. Leroy,S.A.G.,andSimms,M.J.,2006,Ironagetomedievalentomogamousvegetationand Rhinolophushipposideros roostinSouth-Eastern Wales(UK).Palaeogeography,Palaeoclimatology,Palaeoecology, v.237,p.4.doi:10.1016/j.palaeo.2005.11.025. Maher,L.J.,Jr.,2006,Environmentalinformationfromguanopalynology ofinsectivorousbatsofthecentralpartoftheUnitedStatesof America:Palaeogeography,Palaeoclimatology,Palaeoecology,v.237, p.1931.doi:10.1016/j.palaeo.2005.11.026. McGarry,S.F.,andCaseldine,C.,2004,SpeleothemPalynology:an undervaluedtoolinQuaternarystudies:QuaternaryScienceReview, v.23,p.2389.doi:10.1016/j.quascirev.2004.06.007. Meyrick,E.,1924,LepidopteraoftheSijucave,GaroHills,AssamII Tineidae:RecordsoftheIndianMuseum,v.26,114p. Mittermeier,R.A.,Gil,P.R.,Hoffmann,M.,Pilgrim,J.,Brooks,T., Mittermeier,C.G.,Lamoreux,J.,anddaFonseca,G.A.B.,2005, Hotspotsrevisited:EarthsBiologicallyRichestandMostEndangered TerrestialEcosystems,Arlington,Virginia,ConservationInternational,392p. Nair,K.M.,Padmalal,D.,Kumaran,K.P.N.,Sreeja,R.,Limaye,R.B.,and Srinivas,R.,2010,LateQuaternaryevolutionofAshtamudie-SasthamkottalakesystemsofKerala,southwestIndia:JournalofAsianEarth Science,v.37,p.36172.doi:10.1016/j.jseaes.2009.09.004. NavarroCamacho,C.,Carrio n,J.S.,Navaro,J.,Munuera,M.,and Prieto,A.R.,2000,Anexperimentalapproachtothepalynologyof cavedeposits:JournalofQuaternaryScience,v.15,p.603. doi:10.1002/1099-1417(200009)15:6 587::AID-JQS533 3.0.CO;2-2. Navarro,C.,Carrion,J.S.,Munuera,M.,andPrieto,A.R.,2001,Cave surfacepollenandthepalynologicalpotentialofkarsticcave sedimentsinpalaeoecology:ReviewofPaleobotanyandPalynology, v.117,p.245.doi:10.1016/S0034-6667(01)00095-1. Nayar,T.S.,1990,PollenFloraofMaharashtraState,India,NewDelhi, TodayandTomorrowsPrinters&Publishers,157p. Nieves-Rivera,A .M.,2003,MycologicalSurveyofR oCamuyCaves Park,PuertoRico:JournalofCaveandKarstStudies,v.65,no.1, p.23. Ochs,G.,1925,DescriptionsofnewAsiaticGyrinidae:Recordsofthe IndianMuseum,v.27,p.193. Patton,W.S.,1924,DipteraoftheSijucave,GaroHills,AssamIV Nycteribiidae:RecordsoftheIndianMuseum,v.26,112p. Pendleton,M.,Bryant,V.M.,andPendleton,B.B.,1996,Entomopalynology, in Jansonius,J.,andMcGregor,D.C.,eds.,Palynology:Principles andApplications,Dallas,AmericanAssociationofStratigraphic PalynologistsFoundation,ContributionsSeries,v.3,p.9393. Pop,E.,andCiobanu,I.,1950,AnalizedepolennghiatadelaScarisoara, AnaleleAcademieiRepubliciiPopulareRomane,seriaGeologie, Geografie,Biologie,StiinTehnicesiAgricole,III,p.23. Qin,Xiaoguang,Tan,Ming,Liu,Tungsheng,Wang,Xianfeng,Li, Tieying,andLu,Jinpo,1999,Spectralanalysisofa1000-year stalagmitelamina-thicknessrecordfromShihuaCavern,Beijing, China,anditsclimaticsignificance:TheHolocene,v.9,p.689. doi:10.1191/095968399671019413. Rohwer,S.A.,1924,HymenopteraoftheSijucave,GaroHills,AssamII DescriptionofanewBraconid:RecordsoftheIndianMuseum,v.26, p.124. SauerCarl,O.,1952,AgriculturalOriginandDispersals,NewYork, AmericanGeographicalSociety,BowmanMemorialLecturesseries2, 131p. Sears,P.B.,andRoosma,A.,1961,AclimaticsequencefromtwoNeveda caves:AmericanJournalofScience,v.259,p.669.doi:10.2475/ ajs.259.9.669. Silvestri,F.,1924,MyriapodafromtheSijucave,GaroHills,Assam: RecordsoftheIndianMuseum,v.26,p.71. Sinha,Y.P.,1994,OccurrenceofKashmirCaveBat Myotislongipes (Dobson,1873)inMeghalaya,India.GeobiosNewReports,v.13,p. 72.Sinha,Y.P.,1999,BatsoftheSijuCave,SouthGaroHills District,Meghalaya,India:taxonomyandbionomics:Recordsofthe ZoologicalSurveyofIndia,v.97,p.101. Stephenson,J.,1924,OligochaetaoftheSijucave,GaroHills,Assam: RecordsoftheIndianMuseum,v.26,p.127.MODERNPOLLENRECORDONBATGUANODEPOSITFROMSIJUCAVEANDITSIMPLICATIONTOPALAEOECOLOGICALSTUDYINSOUTHGAROHILLSOFMEGHALAYA,INDIA182 N JournalofCaveandKarstStudies, December2014

PAGE 31

Vivilov,N.I.,1951,TheOrigin,Variation,ImmunityandBreedingof CultivatedPlants,Waltham,Massachusetts,ChronicaBotanica, 364p. Wheeler,W.M.,1924,HymenopteraoftheSijucave,GaroHills,AssamI Triglyphothrixstriatidens Emeryasacaveant:RecordsoftheIndian Museum,v.26,p.123. White,W.B.,2007,Cavesedimentsandpaleoclimate.JournalofCaveand KarstStudies,v.69,p.76.Zhang,Meiliang,Yuan,DaoXian,Lin, Yushi,Qin,Jiaming,Bin,Li,Cheng,Hai,andEdwards,R.L.,2004,A 6000-yearhigh-resolutionclimaticrecordfromastalagmitein XiangshuiCave,Guilin,China:TheHolocene,v.14,p.697. doi:10.1191/0959683604hl748rp.S.K.BASUMATARYANDS.K.BERAJournalofCaveandKarstStudies, December2014 N 183

PAGE 32

GROUNDPENERATINGRADARINVESTIGATIONOF LIMESTONEKARSTATTHEODSTRZELONACAVEIN KOWALA,S WIETOKRZYSKIEMOUNTAINS,POLANDMIKOAJYSKOWSKI,EWELINAMAZUREK,ANDJERZYZIETEKAGHUniversityofScienceandTechnology,FacultyofGeology,GeophysicsandEnvironmentalProtection,A.Mickiewicza30Ave.,30-059Krakow,Poland, lyskowski@geol.agh.edu.plAbstract: GroundPenetratingRadar(GPR)isoneofthemosteffectiveandrapidtypes ofgeophysicalsurveysmethods.Thevarietyofitsusesislimitedonlybytheavailability ofcomponentssuchasantennaswithdifferentfrequenciesofemittedradiowaves.The applicationofGPRrangesfromengineeringapplicationsandgeologicalinvestigations toassessingtheconditionofoldtrees.TheauthorstestedGPRmeasurementsina geologicalexplorationofanewlydiscoveredlimestonecaveintheS wietokrzyskie Mountainsregion,theOdstrzelonaCaveinKowala,nearthevillageofthatname. InvestigationswereorderedbythemunicipalityofSitkowka-Nowiny,inthe S wietokrzyskiedistrictinPoland.GPRsurveyswereconductedinthevicinityofa knowncaveentrancetoseeifinformationcouldbegainedonnearbykarstfeatures. GPRinvestigationsdocumentedtwoadditionalchambersandafewmoreopeningsin thelimestone.Unfortunately,theyareprobablytoosmalltoexplore.INTRODUCTIONGroundPenetratingRadar(GPR)isafastgeophysical methodthatallowspreciseandeffectiverecognitionof geologicalformationsbelowthesurface.Dryormoderatelywetsolidrocksarethebestmediumforsuchasurvey. ReviewoftheavailableliteratureshowsthattheGPR methodisagoodgeophysicaltoolfordetectingkarst phenomenaandcaves.DoolittleandCollins(1998) comparedtheelectromagneticinductionmethodtoGPR resultsandobtainedverygoodresults.Chamberlainetal. (2000)gaveexamplesofcavedetectioninlimestones.Their surveysgavegoodresultsandthankstoagridofparallel profiles,itwaspossibletopresentthemindifferentways, asbothclassicalprofilesandtimeslices.Anotherexample oftheapplicationoftheGPRmethodindetectingcavesis presentedbyBeresetal.(2001),inwhichtheGPRmethod wascomparedtomicrogravimetricmeasurements.The obtainedresultsaregoodexamplesofGPRprofiles. AnomaliesinGPRechogramswereconfirmedbythe gravitymeasurementresults. Investigationscarriedoutbytheauthorsandpresented inthispaperwerefocusedonacavenewlyrediscoveredin 2008nearthecityofKielce,centralPoland,andon detectingkarstphenomena.Thecavewasprobably discoveredmanyyearsbeforeduringexplorationwork performedbytheworkersofthequarryStaraTrzuskawica, whoshotawaytheceilingintheentrancetunnelofthe cave.Forsafetyreasons,inthe1970sthecavewasdrilled andfilledwithlimestonerubblemadefromsurrounding rocks(Grzelak,2012).Asaresult,thecavewasunavailable formanyyearsfollowingtheclosureofthequarry. Speleologistswhoinvestigatedthequarrynamedit OdstrzelonaCaveinKowala(Polishword odstrzelona meansshootaway; Kowala isthevillagewherethecaveis located).Afterthediscoveryoftheentrance,theywanted toseekkarstareasadjacenttothecave.Speleologists proposedGPRmeasurements,becauseofpreviousexperiencewiththatmethod.Beforethesurvey,researchers analyzedavailablegeologicalinformationandmadealocal inspectiontodetermineifuseofGPRwasfeasible.The surveywasdoneinNovember2010.Thedatafromthe GPRmeasurementswereprocessedandinterpretedvery carefullytoprovideabasisforfuturespeleological investigations,aswellasforevaluatingrisktopublic safety.Wehaveobtainedresultsofaverygoodqualitythat allowedustopreciselylocatekarstobjectsinthelimestone andthelocationofunknownchambers,voids,andcracks.STUDYAREAIntermsofPolishgeographicalregionspresentedby Kondracki(2011),theareawheretheOdstrzelonaCavein KowalaislocatedbelongstotheKielceUplandmacroregionlocatedincentralPoland.Moreprecisely,itisthe S wietokrzyskie(HolyCross)Mountainsmezzo-regionand theChecin skieHillsmicro-region.Thisareaiscomposedof limestoneanddolomiteoftheMiddleDevonianperiod. ThecaveislocatedintheclosedquarryStara TrzuskawicanorthofthevillageofKowalainthe municipalityofSitkowka-Nowiny.Rocksformingthe wallsofthecaveareDevonianlimestones.Theoriginof thecaveisassociatedwithatypicalkarstprocesses.The entranceisartificial,createdduringexplorationworkin thequarry.Onlyasmallpartofthecaveisaccessible. Accesstotherestofthecaveisnotpossiblebecauseofthe collapseofceilingandwalls,whichtookplaceafterthe blastingatthequarry.ThecollapsedtunnelsarefilledwithM.yskowski,E.Mazurek,andJ.ZietekGroundPeneratingRadarinvestigationoflimestonekarstattheOdstrzelonaCavein Kowala,S wietokrzyskieMountains,Poland. JournalofCaveandKarstStudies, v.76,no.3,p.184.DOI:10.4311/2014EX0001184 N JournalofCaveandKarstStudies, December2014

PAGE 33

limestonerubble.Theopenparthasbeenexploredandhas alengthof28.8mandaheightreaching3.8m(Grzelak, 2012).Figure1showsthelocationofthecaveandthe surroundinggeology. Becauseofterraintopographytypicalofanopencast quarry,itwasnotpossibletoconductmeasurementsina parallelgrid.Surveyshadtobecarriedoutalongspecially cutpathsinthevegetationsurroundingthearea.Additional Figure1.AGeologicalmapoftheS wietokrzyskie(HolyCross)Mountainsmezzo-region,withinsetshowingitslocationin Poland(modifiedfromUrban,2007).BAsurfacegeologicalmapoftheOdstrzelonaCaveinKowala(Grzelak,2012).C PhotofrominsideofthecavetakenbyMr.MaciejGrzelak.M.YSKOWSKI,E.MAZUREK,ANDJ.ZIE TEKJournalofCaveandKarstStudies, December2014 N 185

PAGE 34

problemswerecausedbythemultiplelevelsoftheclosed quarry.ThedirectionandlocationofGPRprofileswas determinedbythelocalspeleologistswhohadrediscovered thecave.Cutpathsweremadeinplacesinterestingforthem andwheretheypredictedcontinuationofthekarstcave.The GPRprofilesareshowninFigure2.BASICSOFGPRSURVEYAGPRsystemincludescentralprocessingunit,signal generator,transmitter,transmittingandreceivingantennas,receiver,andmeasuringgear.Todayssystemshavean onboardlaptopcomputerthatfacilitatesdataacquisition andpre-processinginthefield(Reynolds,1997).The modernsystemusesthebistatictypeofantennasinwhich twoseparateantennasareused,withoneservingasa transmitterandtheotherasareceiver(Annan,2001).The GPRsystemsproducedtodayuseantennaswithfrequenciesfrom10MHzto6000MHz(BritishcustomGPR systemGroundvue5C).Basedontheirstructure,bistatic antennascanbedividedintothetwomostcommonlyused types,shieldedandunshielded.Inthecaseofthefirsttype, transmittingandreceivingantennasareencasedinan electromagneticshield.Thehousingmakesthemlessprone toexternalinterference,suchasreflectionsfromtrees,and isalsodesignedforaimingtheelectromagnetic(EM)wave directlyintothemedium.Unshieldedantennasarenot shieldedfromtheinterferenceandarenotphysically integrated.Separationoftransmittingandreceiving antennasmakesitpossibletocarryoutinvestigationswith differentspacingdistancesbetweenthedipolesandwith differentpolarizationsofthedipolesinrelationtothe profileline.Unshieldedantennasalsoallowthespecial kindofsurveycalledWARRprofiling(Wide-Angle ReflectionandRefraction,thefinalresultofwhichisan echogramthatisusedfordeterminationoftheEMwave propagationvelocityinthemedium(yskowskiand Mazurek,2013)). TheGPRsystemisbasedonatransmittergeneratinga pulseofradiowavesatafrequencydeterminedbythe characteristicsoftheantennas.EMwavesareemittedfrom asourcetodetectanobjectatadistanceanddeterminethe directiontotheobjectaswellasthedistancetotheobject. Inorderforanobjecttoreflectradiowavesitmusthave differentdielectricpropertiesfromthesurroundingmaterial,referredtoasthemedium(Annan,2001).Emitted radiowavestravelataspeedthatdependsonthematerial inwhichtheypropagate.Eachscanlastsaslongasthe totaltwo-waytraveltimerange.Thisvalue,definedasthe timewindow,canbesetbytheoperatorfromafewtensto severalthousandsofnanoseconds(Reynolds,1997). Theelectromagneticcharacteristicsofmaterialsare relatedtotheircompositionandwatersaturation.Dueto thehighfrequencyofradarwaves,theyaresensitivetoboth changesinconductivityanddielectricpropertiesofmaterials,whichaffectthespeedofradio-wavepropagationand theattenuationofelectromagneticwavesinthemedium (Reynolds,1997). WaveattenuationdeterminesthedepthrangeofGPR method.Waveattenuationmeansthegradualreductionof anelectromagneticpulsesamplitudealongthetravelpath. Itdependsonfourbasicfactors,thegeometricaldivergence ofthewavefrontfromthepointsourceofthewave, scatteringrelatedtononhomogeneityofthemedium, propagationdispersionofthewaves,whichdependson frequency,andtheelectricalconductivityofthemedium, usuallymainlyduetoporefluids(Annan,2001). Thefrequencyofthesignalemittedbytheantennasalso definesthedepthrangeandtheresolution.Thelowerthe frequencyofemittedEMwave,thegreaterthedepth penetration,butthelowerresolutionofthesurvey.Onthe otherhand,withincreasingfrequency,theresolutionis higher,buttherangedecreases. FortheGPRmethod,theconceptofresolutionis important,butquitecomplex.Wecandefineverticaland horizontalresolutions.Thesimplestdefinitionofboth resolutionssaysthatitistheminimumdistancebetween twoobjectswiththesamecrosssectionthatarevisibleon theechogramasseparateanomalies(Annan,2001).METHODSTheGPRsurveyspresentedinthispaperwereobtained usingaProExGPRunit(MALAGeoscience,Sweden) witha250MHzshieldedantenna.Thisequipmentallowed theacquisitionofinformationtoatheoreticaldepthrange ofabout13m.Testmeasurementswitha100MHz unshieldedantennaweredonealongasingleprofile,profile 16,thatwaslocatedabovethecave.Theobtaineddepth penetrationreachedabout18m.Thewholecasestudy consistsofsixteenprofiles.Theircoursesandlocationsare presentedinFigure2onaschematicmapofthesurvey site,alongwithsampleechograms. TheechogramisafinalresultoftheGPRsurvey.Itis composedofthetracesthatregisterthereflectedEMwave amplitudesasafunctionoftime.Thetracesarearranged onebyonealongtheX-axis.Thelengthoftheprofileis calculatedbythecomputerbyrevolvingthemeasuring gearconnectedtotheantennas.Thesemechanismscan generatesmallerrorsindeterminingtheprofilelength. Despitethefactthatthemeasurementisdoneinsteps declaredincentimeters,itiscalledcontinuous. RawdatacollectedfromtheGPRsurveyswere recordedinnanoseconds.Toobtaintheverticalaxisof theechogramsasadepthscaleinmeters,itwasnecessary todoatime-to-depthconversion.Thisprocessneedsthe valueofEMwavevelocityinthemedium,whichcanbe obtainedfromWARRprofiling.Thevelocitycanbealso calculatedfromtherelativedielectricconstant( er)taken frompublisheddata(e.g.Reynolds,1997,p.682; OYO,1988;MALA ,2009).Todoso,theEMwave velocityisgivenbythespeedoflightinvacuum,0.3mns2 1,GROUNDPENERATINGRADARINVESTIGATIONOFLIMESTONEKARSTATTHEODSTRZELONACAVEINKOWALA,S WIE TOKRZYSKIEMOUNTAINS,POLAND186 N JournalofCaveandKarstStudies, December2014

PAGE 35

Figure2.SchematicgeographicalmapofsurveysitecontainingpositionandcourseofGPRprofiles(blacklineswitharrows) andthelocationofOdstrzelonaCaveinKowala(center);Twoechograms,fromprofiles8and10,withvisibleanomaliesfrom karstfeatures.Notethatthedistancescalealongtheprofilesinthetopechogramisbackwards.Lightgraylinesareisolines (elevation),anddarkgraylinesarecontoursoftheexcavationinthequarryStaraTrzuskawica.M.YSKOWSKI,E.MAZUREK,ANDJ.ZIE TEKJournalofCaveandKarstStudies, December2014 N 187

PAGE 36

dividedbythesquarerootoftherelativedielectricconstant ofthemedium(Karczewskietal.,2011).Duetothe topographyoftheterrainandthevegetation,theauthors determinedthevelocitybyusinganestimationfroma table.DATAPROCESSINGANDRESULTSTherawdatawereprocessedusingthesoftware ReflexW(SandmeierScientificSoftware).Data-processing work-flowincludednoiseremoval,frequencyfiltering, gaining,andsmoothing.Itisalsopossibletoapplyan approximatetopographicshapeofthearea(profile8in Figure2).Thedatapresentedinthispaperwereprocessed asfollows,usingtheprocedurenamesintheprogram (ReflexW,2009):movestarttimeacceptsafixedvaluein nanosecondsthatisthetimeofthefirstsignalinthe echogram.subtract-DC-shiftcalculatesthemeanvalue ofthetrace(theDCvalue)forthegiventimerangeand subtractsthisvaluefromalldata.substract-mean (dewow)createsamovingtimewindowalongtheZaxis, subtractingthemeanvaluesfromtheactualdata. bandpassbutterworthperformsbandbassfrequencyfilteringinthetimedomainofeachtrace.timecutremoves tracevaluesbelowaspecifiedtimevalue.gainfunction strengthensreflectionswithatime-dependentmultiplier. averagexy-filterisasmoothingfilterthatcalculatesthe averageforaspecifieddatarange.Andstaticcorrection, appliedonlytoprofile8inFigure2,appliessurface topographytotheechogram. Thedata-processingwork-flowpresentedabovewas chosenasthemosteffectivefromseveralthathavebeen tried.Time-depthconversionoftheGPRdatawasdone withtheuseofpublisheddatafortherelativedielectric constantoflimestone.Theassumedvalue er5 9 correspondstoelectricalconductivityofmedium-wet limestone(OYO,1988).ItgivesanEMwavevelocity calculatedas v 5 0.1mns2 1.Theconversionfrom nanosecondstometersdepthisthenafactorof20,since theechohastraveledbothways. Profile8showninFigure2containsinformationabout thetopography.Pleasenoticethatthehorizontalaxisin thisplotisininverseorder.Thevisibledippingreflections, whichstartfrom25malongtheprofile(belowabbreviated r-m,forrunningmeters),canbeassignedtothegeological layersoflimestonesorkarstfissures.Between11and17rmtherearetwoanomaliesatthedepthof7and10.5mthat couldoriginatefromkarstvoids.Atabout4r-matthe depthof4m,theanomalyofavoidisclearlyvisible,and anotheronethatcouldbeinducedbydeepcracksinthe limestone,isvisibleatadepthof9m. Atthebeginningofprofile10thereisavisibleanomaly forakarstcave,whichcorrespondstotheonevisibleon profile9inFigure3at30r-m.Between10and17r-mat thedepthofabout1to4mwecanseeananomalyfrom looserock. Profile2,showninFigure3,istheclosestonetothe OdstrzelonaCaveinKowalaandshowsacracklocated neartheopenedfragmentofthecave.Thewholeseriesof smallhyperbolasbetween5and35r-moftheprofile, startingatthedepthof1m,aretheresultofkarst processes.Theanomalyat45r-matthedepthof6mand thicknessofabout1mcanbeassignedtoakarstvoid,as canbetheanomalylessvisibleatadepthabout8.5m. Adistinctanomalyfromakarstvoidappearsonprofile 3(Fig.3)at10r-mandadepthof6.5m.Underneathwe canseepoorlyvisibleanomaliesfromcracksintheold quarry.Near5r-matthedepthof2masmallanomalycan beseen,whichprobablyisalsoassociatedwithkarst processes.At14r-moftheprofileonemoreanomaly, elongatedintimewithheightofabout8m,canbe observed.Wesuggestthatitoriginatedfromakarst chimney. Profile12(Fig.3)wasdesignedinordertoseeif OdstrzelonaCaveinKowalacontinuestothenorth.Inthe first10r-moftheprofiletherearemanyanomaliesrelated tocracksneartheopenedfragmentofcave.Fromabout12 r-matabout1to3.5mdepth,ananomalywithasimilar originappears. Theveryinterestingprofile9(Fig.3)showswithout doubtfourbigandclearlyvisibleanomalies.Theyare picturesofakarstcavewithtwochambers.Thefirstone startsat23r-mofprofileandcontinuesforthenext11r-m, until34r-m.Thesecondonecanbefoundat52r-mandis visiblefor10m.Theirceilingsstartatdepths3.5mforthe firstoneand3mforthesecondone.Measuredthickness reachesupto2.5m.Brightanomaliesbelowthesetwo,ata depthofabout7.5to8mandlocatedat25r-mand52 62r-m,respectively,andthicknessreachingupto1.5mare probablyfloorsofthesechambers.Attheendoftheprofile wecanseeasimilaranomalytotheoneinprofile12.The reflector,about8-mhigh,isakarstchimney.DISCUSSIONANDCONCLUSIONComparingthequalityofobtainedresultstoechograms fromastudyperformedbyBeresetal.(2001)showsour surveystobeequallysatisfactoryandprecise. DuetothelimitationsofGPR,thereareseveral problemswithconductingmeasurementsandtheirinterpretation.Thesurveyhasconsiderablespacerequirements. ForcarryingoutGPRmeasurements,antennasneedclose orevendirectcontactwiththebedrocksurface.Treesand bushes,andevencutbrush,areagreatobstacle.Another limitationisthedepthrangeandresolution.Those parametersaredirectlyconnected.Forgreaterdepth penetration,theresolutionislowerandthesizeandweight ofantennasgrows;forexample,an800MHzshielded antennahassizeapproximately0.430.2mandweighs 2.6kg,anda250MHzshieldedantennaisabout0.830.5mand8kg.IfyouconsideranidealbedrockforGPR measurements,suchasdrylimestone,forreachingdeeperitGROUNDPENERATINGRADARINVESTIGATIONOFLIMESTONEKARSTATTHEODSTRZELONACAVEINKOWALA,S WIE TOKRZYSKIEMOUNTAINS,POLAND188 N JournalofCaveandKarstStudies, December2014

PAGE 37

Figure3.Echogramsfromprofiles2,3,9,and12containingclearlyvisibleanomaliesfromkarstprocesseffects:disruptionof thebedrock,asinprofile2between5and35m,orvoids,asinprofile9,whichhasfourlargeanomalies,twobetween23and 34r-wandtwobetween52and62r-m,eachpairinterpretedastheceilingandfloorofavoid.M.YSKOWSKI,E.MAZUREK,ANDJ.ZIE TEKJournalofCaveandKarstStudies, December2014 N 189

PAGE 38

isnecessarytousealowerfrequency,requiringalarger antennaandprovidinglowerresolution.Useofa250MHz antennagavegood-qualitydataandprovedtobetheright choice.Theresolutionandusabledepthpenetrationrange obtained,upto10m,weresatisfactory.Theterrain coveredwithtreeswherethesurveyswerecarriedoutalso madethatthechoiceofantennafrequency,size,andweight optimal.Usageoftheshieldedtypeantennasalsoreduced thenoise,asourceofwhichcanbetrees,so-called reflectionsinair.Beforeanalyzingtheresultstheauthors repeatedlyprocessedtherawdata.Theoptimalsequenceof processingwasappliedtoallprofiles.Processingalldatain thesamewayguaranteesconfidenceofinterpretation.Only profilenumber16,becauseoftheuseofanunshielded 100MHzantenna,wasprocessedindividually.The proceduresusedremovedunwantednoiseandkeptthe interferenceintherawdataataminimum. Satisfactorydetectionofvoidsandcracks(e.g.,Xuetal., 2010),caves,orkarstobjectsispossiblebyusingonlyGPR. Therearemanypublishedexamplesofsuchexploratory works(e.g.,Chamberlainetal.,2000).Forthepurposesofa fullinventoryofacaveitisnecessarytoperforma comprehensivestudy,includingtheuseofothergeophysical methodsandaprecisegeodeticgrid.Unfortunately,the topographyandthedenseforestoftheoldquarrymadeit impossibletoconductsuchasurvey.Complementaryuseof othermethodssuchasmicrogravity(Bereset.al,2001)ora high-resolutionengineeringseismicsurveywouldgivemore precisedataaboutthedepthandsizeofthecave. Unfortunately,thesemethodsaremuchmoreexpensive andtime-consuming.Despitetheuseofonlyonegeophysicalmethod,informationaboutvoidsandkarstthataffect theinvestigatedlimestonesgaveamorecomprehensive pictureforfurtherexplorationoftheareaoftheclosed quarry.Useofonlyonemethodalwaysprovidesambiguous results.However,precisemeasurementsandthought-out profileslines,goodacquisitionparameters,andoptimal processingofrawdataminimizetheimpactofthelackofa secondmethod.ACKNOWLEDGEMENTSWeexpressaspecialthankstothemayorofmunicipality Sitkowka-Nowiny,Mr.StanisawBarycki,whogave permissionforthepublicationoftheresultsofthesurveys. WewouldliketothanktheAssoc.Prof.SallySuttonfrom ColoradoStateUniversity,UnitedStates,andour colleaguefromAGHUniversityofScienceandTechnologyinKrakow,Poland,Ph.D.Eng.KamilaWawrzyniakGuzforsupportduringthewritingofthispaper. Additionalthanksareexpressedtothereviewersfortheir commentsthathelpedimprovethemanuscriptandtoMr. MaciejGrzelakforaccesstothephoto. MeasurementswereconductedbyM.Sc.Eng.Mikoaj yskowskiunderthesupervisionofPh.D.Eng.Jerzy Zietek.Presentationandanalysisofthesurveysresults weremadebyM.Sc.Eng.MikoajyskowskiandM.Sc. Eng.EwelinaMazurek,whoarePh.D.studentsatAGH UniversityofScienceandTechnologyinKrakowinthe DepartmentofGeophysics.REFERENCESAnnan,A.P.,2001,GroundPenetratingRadarWorkshopNotes: Mississauga,Ontario,Sensors&SoftwareInc. Beres,M.,Luetscher,M.,andOlivier,R.,2001,Integrationofgroundpenetratingradarandmicrogravimetricmethodstomapshallow caves:JournalofAppliedGeophysics,v.46,p.249.doi:10.1016/ S0926-9851(01)00042-8. Chamberlain,A.T.,Sellers,W.,Proctor,C.,andCoard,R.,2000,Cave detectioninlimestoneusinggroundpenetratingradar:Journalof ArchaeologicalScience,v.27,p.957.doi:10.1006/jasc.1999.0525. Doolittle,J.A.,andCollins,M.E.,1998,AcomparisonofEMinduction andGPRmethodsinareasofkarst,Geoderma,v.85,p.83. doi:10.1016/S0016-7061(98)00012-3. Grzelak,M.,2012,Jaskiniewregionies wietokrzyskim:Jaskinie,v.1, no.66,p.25. Karczewski,J.,Ortyl,.,andPasternak,M.,2011,Zarysmetody georadarowej,Wydaniedrugiepoprawioneirozszerzone:Krakow, AGHUniversityofScienceandTechnologyPress,346p. Kondracki,J.,2011,GeografiaregionalnaPolski:Warsaw,Wydawnictwo NaukowePWN468p. yskowski,M.,andMazurek,E.,2013.Analizakonsekwencjidoboru nieodpowiedniejpredkos cipropagacjifalielektromagnetycznejw trakcieinterpretacjiinz ynierskichpomiarowmetodageoradarowa (Englishabstract):Logistyka2013nr4,suplement:CDLogistyka nauka,p.330. MALA Geoscience,2009,ProExProfessionalExplorerControlUnit. OperatingManualV.2.0. OYO,1988,OYOGeoradarIManual:Tsukuba,Japan,OyoCorporation,50p. Reynolds,J.M.,1997,AnIntroductiontoAppliedandEnvironmental Geophysics:WestSussex,England,JohnWiley&Sons,796p. SSS,2009,ReflexwManual,UserGuide:Karlsruhe,Germany,Sandmeier ScientificSoftware. Urban,J.,2007,PermiantoTriassicpaleokarstoftheS wietokrzyskie (HolyCross)Mts.,CentralPoland:Geologia(Krakow),v.33,no.1, p.5. Xu,Xingxin,Zeng,Qiaosong,Li,Dong,Wu,Jin,Wu,Xiangan,andShen, Jinyin,2010,GPRdetectionofseveralcommonsubsurfacevoids insidedikesanddams:EngineeringGeology,v.111,p.31. doi:10.1016/j.enggeo.2009.12.001.GROUNDPENERATINGRADARINVESTIGATIONOFLIMESTONEKARSTATTHEODSTRZELONACAVEINKOWALA,S WIE TOKRZYSKIEMOUNTAINS,POLAND190 N JournalofCaveandKarstStudies, December2014

PAGE 39

THREE-DIMENSIONALMOBILEMAPPINGOFCAVESROBERTZLOT*ANDMICHAELBOSSEAutonomousSystems,CSIRO,Brisbane,AustraliaAbstract: Existingmethodsofcavesurveyaretime consumingandrequiresignificantly moretimethannaturallymovingthroughtheca ve.Theefficiencyofthesemethods,evenin thecaseofstate-of-the-artlaser-scanningtechnology,isfundamentallylimitedbythe requirementthatmeasurementsbetakenatstaticlocations.Wepresentamobileapproachto cavemapping,inwhichalightweight3Dlasers canneriscarriedbyasingleoperatorwhile walking,climbing,orcrawlin gthroughacaveatanaturalpa ce.Themobilityofthesystem meansthatitisstraightforwardandefficienttogenerateahigh-resolution3Dmapconsisting ofmillionsofpointsinalmostanyenvironm entahumancanreach.Wepresentresults demonstratingthetechnologyintwocavesystem slocatedindifferentpa rtsofAustralia,what webelievearethefirstinstancesofmobileLiDARmappingbeingutiliz edinnaturalcaves.INTRODUCTIONTraditionalmethodsofcavemappinginvolvemanual measurementsofrangeandbearingbetweenasequenceof stationstypicallyspaceduptoafewtensofmetersapart (Warild,2007,chap.10;Kershaw,2012).Themost commoninstrumentsformeasuringbearingarecompass andclinometer,whileforrange,afiberglasstapemeasure ortopofilarecommonlyused.Detailisobtainedthrough hand-drawnsketchesofthelocalcavepassagerenderedat someorallofthestations,aswellasleft-right-up-down distancemeasurementstothewalls,roof,andfloor.The dataarelatermergedintoacavemapbasedonthe registeredsurveystationlocations,oftenusingcomputer softwarepackages.Thesemappingtechniquesarenotonly highlytimeconsuming,butrelyonmanualacquisitionand recordingofthemeasurementsandsketches,whichare pronetohumanerrorsincludinginstrument-sightingerror, stationerrors,andtranscriptionerrors(Hunter,2010). IntegratedsystemssuchasDistoX(Heeb,2008)can eliminatesomeofthemanualsourcesoferrorbylogging measurementsdirectlyfromelectronicinstrumentstoa handheldmobiledevice,ratherthanpaper.Recently, handheldlaserdistancemetershavebeenincreasingly employedinsteadoftapetoacquirerangemeasurements betweensurveystations(Dryjanskii,2010).Theodolite systemshavelongbeenutilizedinsomeinstances(Middleton,1991;DavisandLand,2006;Rutheretal.,2009), butareconsideredtoocumbersomeorimpracticalfor manycavesurveyapplicationsduetotheirsize,fragility, andweight(Warild,2007;Slavova,2012). Oneofthemostcompellingrecentexamplesof traditionalcavesurveyingistheJenolanCavesSurvey Project(Jamesetal.,2009).Thisprojectproduceda tremendouslycomprehensive3Dmodelofthecavesusing dataprimarilyacquiredwithtotalstationandlaser distancemeasurements.Toimprovetheresolutionofthe modelinareaswithlargevoids,distancemeasurements weretakenintwelve-pointcrosssectionsspacedat10m intervals.Whiletheresultingmodelishighlydetailedand accurate,anextraordinaryamountofsurveyinganddata processingeffortwasinvestedinthemulti-yearproject, whichwascarriedoutbetween1987and2005. Inrecentyears,terrestriallaserscanning,orLiDAR, technologyhasbeenusedtocreatehigh-resolution3D mapsofanumberofcaves(e.g.,Rutheretal.,2009; McIntire,2010;Sadieretal.,2012).Terrestrialscannersare typicallymountedonastationarytripodandacquire millionsofpreciserangemeasurementsofthesurfaces surroundingthestationoveraperiodofafewminutes. Datafrommultiplestationscanbecombinedifthereis sufficientoverlapbetweenthescannedsurfaces,thoughin practiceitismorecommontomeasurethestationposition andorientationusingstandardsurveyingtechniques,orto placeknowntargetsintothescans(Rutheretal.,2009). Despitethehighqualityofthedataresultingfrom terrestrialLiDAR,thetechniquehasseenrelativelylimited useincaves,mostlikelyduetothehighcostofthe scanners,aswellasthesize,weight,andfragilityofthe equipmentmakingitdifficulttotransportthroughdifficult terrainandtightsqueezes.Inaddition,theoftencomplex geometryofcavesmayprescribethatalargenumberof scansbeacquiredtoachievesufficientcoverageandavoid shadowsduetoocclusion.McIntire(2010)reports,the mosttime-consumingpartofthescanwasmovingfrom stationtostationandshootinginthetargets....Every setupwasachallengetodeterminewherethepreviousand upcomingscansshadowswouldoccurandlocatingthe bestcombinationofscancoverageandsetupefficiency. Mobilemappingisatechniquewherebymeasurements oftheenvironmentareacquiredwhilemovingcontinuouslythroughit.Commercialsolutionsexistthatacquire LiDARscansfrommovingaircraft,watercraft,orstreet vehicles(Petrie,2010);however,thesesystemstypicallyrely onexpensive,bulkyequipmentandonglobalnavigation satellitesystems(GNSS)suchasGPSforpositioning. Miningapplicationshaveprovidedamarketfortransitioningmobilemappingtechnologyunderground.Duetothe *CorrespondingAuthor:Robert.Zlot@csiro.auR.ZlotandM.BosseThree-dimensionalmobilemappingofcaves. JournalofCaveandKarstStudies, v.76,no.3,p.191.DOI: 10.4311/2012EX0287JournalofCaveandKarstStudies, December2014 N 191

PAGE 40

difficultyofoperatingwithoutGNSScoverage,undergroundminemappinghaslargelyremainedaresearch problem(Nuchteretal.,2004;Morrisetal.,2006;Fairfield etal.,2010),and,withtheexceptionofotherrecentwork wehavedonebasedontechnologysimilartowhatis presentedinthispaper(ZlotandBosse,2014),nosystems ofwhichweareawarehavebeendemonstratedtobe capableofefficientlymappingalarge-scaleminein3D.In anycase,theplatformsthathavebeenfieldedhavebeen mountedonwheeledvehiclesthatrelyontheexistenceof roadwaysinthemineandwouldbeunsuitableformost naturalcaves.Indoormobilemappingtechnologycould alsobeconsideredrelevanttomappingcaves,butexisting systemsinvolvewheeledplatforms(e.g.,CanterandStott, 2011)orrequireplanarstructuresintheenvironmentsuch aswalls,floors,andceilings(Liuetal.,2010). Toaddresssomeofthelimitationsofexistingmobile mappingsystems,wehavedevelopedportableLiDARbasedmobilemappingtechnologythatdoesnotrequire externalpositioningorartificialinfrastructure,isrelatively inexpensive,andcaneasilybecarriedbyhandbyasingle operatorthroughthechallengingenvironmentspresented bynaturalcaves(Fig.1).Therawdataacquiredare convertedbysoftwareintogloballyconsistentand metricallyaccurate3Dpointcloudsorsurfacemodels consistingofmillionsofpointsortriangles,aswellasan estimateofthesensortrajectorythroughthecave.The combinationofmobilityandmaneuverabilityfacilitates efficientdataacquisition,becauseacavecanbemappedin thetimeittakestowalk,crawl,squeeze,orclimbthrough it;ahighdegreeofcoverage,becauseshadowsdueto occlusionareavoidedbyvirtueofthemotion;and versatility,becausemostterrainsthroughwhichahuman cantraversecanbemapped.Inaddition,thesystemisfully automated,eliminatingthehumanerrorsinherentto manualsurveyingtechniques,despiterequiringnearlyzero trainingtime.Easyacquisitionofdenseandaccuratepoint clouddatacanprovidemodelsusefulnotonlyfor navigation,butalsoforavarietyofscientificapplications previouslynotpossibleoreconomical. Ourhandheldmobilemappingsystemhasbeen deployedinseveralcaveandundergroundminesites aroundtheworld.Toourknowledge,ourbodyofwork representsthefirstinstanceofmobilemappinginnonsubmergedcaves(Stoneetal.(2000)andGaryetal.(2008) describemappingunderwatervoidsusingsonar),anda significantlymoreefficientandcompletemethodof surveyingcavescomparedtothestate-of-the-art.After describingtheequipmentanditsuse,wepresentresults obtainedfromextensivescanningofsignificantpartsofthe JenolanCavesandKoonaldaCaveinAustralia.EQUIPMENTANDMETHODSThekeyenablersofourmobilecavemappingtechnologyarealightweighthandheldlaser-scanningdevice coupledwithdataprocessingsoftwarecapableofaccuratelyestimatingthepositionandorientationofthe scannerovertimeasitismovedthroughtheenvironment. Thescannermeasurestensofthousandsofrangesper secondfromthesensororigintopointsonvariousphysical surfacesusingnarrowinfraredlaserpulses.Givenan accurateestimateofthescannersmotion,thesetofrange measurementscanbeprojectedinto( x, y z )pointsina commoncoordinateframe,therebygeneratingaconsistent pointcloudmodelofthecaveandsurrounds.THEZEBEDEEMOBILEMAPPINGSYSTEMZebedee(Fig.1)isahandheld3Dmobilemapping systemconsistingofa2Dlaserscannermountedonaspring (Bosseetal.,2012).TheinfraredlaserscannerisaHokuyo UTM-30LX,which,at370g,islightenoughtobecarriedby hand.TheUTM-30LXemits905nmlaserpulsesatahigh frequencythatreflectoffsurfacesintheenvironmentand returntothesensor.Thescannerinternallyconvertsthe signaltoarangemeasurementbasedonthetimeofflight. Withinthescanner,thelaserpulsesarespreadacrossaplane byaspinningmirrorrotatingat40Hz(100Hzinthenewest Zebedeehardware).Measurementsareacquiredwithina fieldofviewof270 u atquarter-degreeangularresolution, resultingin43,200pointspersecond.Themaximumrange ofthescannerisapproximately35minthecave environmentandsurfacesbeyondthatrangearenot registeredasmeasurements.Therangeprecisionistypically 1to3cm,dependingonthedistanceandincidenceangleto thesurface,aswellassurfacereflectivity. AuniquedesignfeatureofZebedeeisthespringon whichthelaserscannerismounted.Thepurposeofthe springistopassivelyconvertthenaturalmotionof theoperatorcarryingthedeviceintorotationalmotionof thescanner.Thenon-deterministic,looselyswayingmotion ofthescannertypicallyresultsina150to180 u out-of-scanplanefieldofview.Sweepingthedeviceinthismanner effectivelyextendstheinherenttwo-dimensionalfieldof viewoftheHokuyoscannerintoathree-dimensionalview oftheenvironmentacquiredroughlyeverysecond. AMicroStrain3DM-GX3industrial-grademicroelectromechanical(MEMS)inertialmeasurementunit(IMU) ismountedbeneaththescanner,andprovidesmeasurementsofangularvelocitiesandlinearaccelerations.The inertialmeasurementsareusedbytheprocessingsoftware, alongwiththeLiDARdata,toestimatethescanner trajectory.TheIMUalsocontainsathree-axismagnetometerthatcanfurtheraidthesolutionbyconstrainingthe absoluteheadinginenvironmentswithminimalmagnetic interference,whichisoftenthecaseincaves. Inadditiontothehandhelddevice,theZebedee hardwaresystemincludesasmalllaptopforoperating thesensorsandloggingdata.Alithium-ionbatterypack powersboththesensorsandlaptop.Batteriesofvarious capacitiesareavailable;a1kgbatteryprovidesmorethan tenhoursofoperation.Thelaptop,powerpack,andspareTHREE-DIMENSIONALMOBILEMAPPINGOFCAVES192 N JournalofCaveandKarstStudies, December2014

PAGE 41

batteriesaretypicallycarriedinabackpack,whichcan easilyaccommodateothertoolsandsupplies.Operation canalsobefullycontrolledandmonitoredviaa smartphoneinterface,eliminatingtheneedtoaccessthe laptopbetweendatasets.Toextendtheoperatorsreach,a cameramonopodcanbeconnectedtothebottomofthe Zebedeehandleviaastandard J -inchtripodscrewsocket.DATAACQUISITIONDatacollectionwithZebedeeisacontinuousprocedure,butcanbebrokendownintoanumberof manageabledatasets,whicharetypicallyintherangeof 10to90minutesinduration.Acquiringdataforan individualdatasetinvolvespoweringuptheequipment, startingtheloggingsoftwareusingawebbrowserona laptopormobiledevice,pickingupthescanningdevice, followingadesiredpath,thenputtingdownthedeviceand terminatingtheloggingprocess.Iftheobjectiveistomerge multipledatasetsintoaglobalmodel,somerepeated coverageisrequiredbetweenthescannedareasinorderfor theprocessingsoftwaretoautomaticallydetectthe matchingareasandalignthepointclouds.Typicallya Figure1.(a)MappingKoonaldaCavewiththeZebedee3Dmappingsystem.Thescanningdeviceisheldintheoperators righthand,withabatterypackandsmalllaptopforrecordingdatacarriedinabackpack.(b)Thecomponentsofthe Zebedeesystem.R.ZLOTANDM.BOSSEJournalofCaveandKarstStudies, December2014 N 193

PAGE 42

fewmetersofoverlappingtrajectoryissufficient.If multipleZebedeeunitsareavailable,multipleoperators cansimultaneouslycollectdataindifferentareastobe combinedlaterintoacommonmap.Thedataacquisition softwareusestheopen-sourceRobotOperatingSystem (ROS)middlewareplatform,andstorestherawdatain ROS-native.bagfilesatarateofapproximately100MB every7.5minutes.Thedataprocessingsoftwareisrun post-acquisitionandoutputsthe3Dmapsandtrajectory followedinstandardpointcloudfileformats,including.laz (compressed.las)and.ply.Thesizeofapoint-cloudfile storedin.lazformatisapproximately1MBforevery 325,000points,orabout7.75MBperminuteofdata,soa typicalcavemodelisexpectedtobeanywhereintherange oftensofmegabytestoagigabyteinsize.DATAPROCESSINGTheverynatureofmobilemappinginvolvesasensor platformthatiscontinuouslyinmotionasmeasurements areacquired.However,inordertogenerateaconsistent map,allofthepointmeasurementsarerequiredtobe transformedintoacommoncoordinateframe.Therefore,a mappingsolutionmustbeabletoaccuratelyestimatethe trajectoryofthelaserscanner,acontinuousfunction specifyingthescanners3Dpositionandorientationatall timesduringacquisition.Althoughthereisnoexternal positioningsystemavailable,thetrajectorycanbeestimatedbasedentirelyontheLiDARandinertialmeasurements. Thechallengeofconcurrentlyestimatingthetrajectoryofa sensorandamapoftheenvironmentisafundamental problemknownintheroboticsliteratureassimultaneous localizationandmapping,orSLAM(Durrant-Whyteand Bailey,2006). AnessentialrequirementofSLAMsolutionsisthe observationoffeaturesintheenvironmentmultipletimes. Asasimplifiedexampleofhowmotioncanbeestimated throughexternalobservations,imaginetakingasinglepointmeasurementofthedistancetoawall.Aninitial measurementof5misrecorded,followedbyameasurementof3mtakenataslightlylatertime.Assumingthat thetwomeasurementscanbeassociatedwiththesame physicalsurface,onecaninferfromtheseobservationsthat inthistimethesensorhasmoved2minthedirection perpendiculartothewall.Byaggregatingthousandsof similarmeasurementsofmatchedsurfaceswithvarious orientationsintheobservedenvironment,the3Dmotion ofthesensorcanbeestimatedtoahighdegreeofprecision. ThealgorithmicframeworkunderlyingtheZebedee trajectoryestimationsolutionisbasedonageneralization oftheaboveprinciple(Bosseetal.,2012).Asthelaser scannerswingsaboutonthespring,itsweepsthroughits fieldofview,capturinga3Dscanofthelocalenvironment roughlyoncepersecond.Withinatimewindowofafew seconds,thereisaconsiderableamountofoverlapbetween thepartsoftheenvironmentscannedineachsensorsweep. Surfaceelements,whichcontainapositionandnormal direction,areextractedfromlocalpatchesofscanpoints fromwithinthesesweeps.Bymatchingpairsofsurface elementsacquiredatdistincttimes,thetrajectorybetween thosesamplesisdeterminedbytherelationbetweenthe surfacegeometries.Theinertialandmagnetometermeasurementsareusedtogeneratefurtherconstraintsonthe scannertrajectoryovertheseshorttimewindows.An optimizationroutinesolvesforthetrajectorythatminimizesthedifferencesamongthevariousconstraints startingfromaninitialestimatederivedfromtheinertial measurements.Byshiftingthetimewindowateachtime step,thetrajectoryofthescannerisincrementally generatedasmoredataarecaptured. Asthetrajectorylengthincreases,smallerrorscanbuild up,resultinginadriftofthesolutionovertime.While theseerrorstendtobesmall,notaddressingthemwould notonlyresultinglobalinaccuracies,butalsoinpoint cloudswithapparentfuzzinessordoublingofsurfacesdue tomultipleimagesofasurfaceobservedatdifferenttimes. Therearealsosituationswherethereisahigherriskthat largelocalerrorscouldbeintroducedintothesolution, suchasverytightsqueezeswherethescannerhasalimited viewoftheenvironment.TheSLAMalgorithmusedfor estimatingtheinitialscannertrajectorycanalsobeusedto applycorrectionstothetrajectorybyoptimizingallofthe dataoveranentiredatasetratherthanoverafew-second timewindow.Thisnon-rigidglobaloptimizationstep appliessmallcorrectionsalongthetrajectorythatresult inconsistent,registeredsurfacesthroughoutthemap.This processissomewhatanalogoustoapplyingloopclosure constraintsinatraditionalsurvey.However,akey differenceinthissolutionisthattheloopclosuresdonot occuratdiscretestations(astherearenostations),but ratheraredetectedautomaticallyandcontinuouslyasparts oftheenvironmentarerescanned. Ifthebuildupoferrorsisrelativelylarge(forexample, inalargedataset),aplacerecognitionalgorithmcan improvetheinputtrajectoryprovidedtotheglobal optimizationbyidentifyinglocationsintheenvironment thathavebeenscannedmultipletimes(BosseandZlot, 2013).Placerecognitionalsoprovidesthefacilityto automaticallyalignmultiplemapstogether,providedthere issufficientoverlap,typicallyafewmeters,amongthem. Thiscapabilityisusefulformergingdatacollectedat differenttimesorwhenmultipleoperatorsarescanninga cavesystemsimultaneously. Theentiredataprocessingpipelineisrunautomatically:TherawLiDARandinertialmeasurementstreamsare takenasinput,andatrajectoryand3Dpointcloudare generatedasoutput.Theprocessingtimerequiredto computeasolutionislessthanthetimespentcollecting thedata;thereforeitispossibletobuildamapinreal-time duringacquisition.Moredetailedtechnicaldescriptionsof thealgorithmsareavailableinpreviouspublications (Bosseetal.,2012;BosseandZlot,2013;ZlotandBosse, 2014).THREE-DIMENSIONALMOBILEMAPPINGOFCAVES194 N JournalofCaveandKarstStudies, December2014

PAGE 43

RESULTSWehavedeployedour3Dmappingsystemsinseveral cavesaroundtheworld.Herewedescribethemost significantundertakingsandresultsfromtheJenolan CavesandKoonaldaCaveinAustralia.JENOLANCAVESTheJenolanCavesarelocatedinNewSouthWales, approximately110kmwestofSydney.Arecentstudyhas datedsomeareasofthecavestobe340millionyearsold, makingJenolantheoldestcomplexcavesystemaccessible tohumans(Osborneetal.,2006).TheJenolansystem consistsofanumberofinterconnectedcavesgenerally runninginanorthsouthdirectionandatvariouslevels vertically(Fig.2),withseveraloftheentranceslocatedat theGrandArch.Thecavesextendingakilometertothe southoftheGrandArchcontainseverallargechambers connectedbyaseriesofpassages,thelongestofwhich includesanundergroundriverflowingthroughapassage belowLucas,TempleofBaal,andOrientCaves.Tothe northoftheGrandArch,thecavesaregenerallycontsistof longpassagesandrelativelynarrowerlocalvoids.Awide varietyofattractivespeleothemsarepresentthroughout manyofthecaves.TheJenolanCavesareapopular touristattraction,andanumberofthecaveshavebeen convertedintoshowcaveswithpavedpathways,stairs, handrails,andlighting.Othercaveshavegenerallybeen leftintheirnaturalstate,containingruggedterrainand tightsqueezes. Ourinitialinvolvementinmappingthecaveswasto supportresearchbeingconductedbyresearchersatthe AustralianNuclearScienceandTechnologyOrganization (ANSTO).ANSTOisinvestigatingtheuseofspeleothem compositionandgrowthpatternsforinterpretingthe palaeo-climaticrecordthroughmeasurementandanalysis ofisotopes(Waringetal.,2009).Varioussensorshavebeen placedinseveralofthecavestomonitorthecompositionof gases,dripwater,andairflow.Large-scale,high-resolution 3Dvolumetricmodelsarenecessarytomoreaccurately modelairflowandgrowthpatternsthroughthecave system. OurfirstmappingtriptotheJenolanCavestookplace overtwodaysinSeptember2010.Atthetime,theZebedee systemwasinaveryearlystageofdevelopmentandwas notreadyfordeployment.However,theinitialrequirementsspecificallyinvolvedmappingseveralofthetourist showcaves,whichcontainpavedpathwaysandstairs. Therefore,weconstructedawheeledmobile-mapping platform,calledHannibal,consistingofhardwarethat Figure2.Three-dimensionalpointcloudmapoftheJenolanCavesprojectedasoverheadandelevationviews.Themapis generatedfrom15.5hoursofLiDARdatacollectedoverfourvisitsin2010(withHannibaldevice),2011(Zebedeedevice), 2012(Zebedee),and2013(Zebedee).Adifferentcolormapisusedtocolordatafromeachtripaccordingtoelevationas indicatedontherightsideoftheimage.Thefullresolutionpointcloudfromthesedatasetsconsistsofover2.7billionpoints. Significantcavesarelabeled.R.ZLOTANDM.BOSSEJournalofCaveandKarstStudies, December2014 N 195

PAGE 44

hadpreviouslybeendevelopedforuseonvehicles.This systemusedamotorizedplatformtospina2Dlaser scannerataconstantrate(BosseandZlot,2009).The scanner,aSICKLMS291,measuresrangesinaplane, similartotheHokuyo,withseveralkeydifferencesinthe specification,mostnotablyasignificantlyincreasedmassat 4.5kg.Inaddition,theLMS291scansat75Hz,producing 180pointsperscanwithina180 u angularrange,andhas withan80mmaximumrange.Bycontinuouslyrotating thescanneraboutthecenterscanrayat30rpm,a hemispherical3Dfieldofviewcontaining13,500pointsis obtainedoncepersecond.ThespinningLiDARsystemis mountedtoanupright,two-wheeledfurniturecart1.3min heightand54cmwide(Fig.3).AMicroStrain3DM-GX2 IMUisrigidlymountedtothecart,providingmeasurementsofangularvelocityandlinearaccelerationat100Hz usedtoprovideadditionalreliabilityaswellasstabilization ofpitchandrollangleswithrespecttogravity.Atthe bottomendofthecartarethreeracksstoringsealedleadacidbatteries,electronics,andasmallnetbookfor controllingthesensorsanddatalogging.Asthesystems overallmassisapproximately60kg,aconsiderable physicaleffortisrequired,oftenbytwopeople,tomove thecartupanddownlongstairways.Dataacquiredfrom Hannibalareprocessedwiththesamecoresoftwareused forZebedeedata. Overthecourseoftwodays,wecollecteddataovera pathlengthofmorethan9kmwithinChifley,Imperial, Lucas,TempleofBaal,andOrientCaves,aswellasseveral areasofthesurfaceaboveandbetweentheentrancesofthe caves.Thetotalacquisitiontimewasjustundertenhours atanaveragespeedof0.7km/hinsidethecaves.Inavery smallnumberofareasthefootpathsorstairwaysbecame toonarrow,andHannibalwasbrieflyraisedandcarriedby twoormorepeopleforafewmeters. Anumberofstationsfromprevioussurveys(James etal.,2009)aremarkedinthegroundsurfaceofthecaves usingstainlesssteeldisks.Wethereforeelectedto incorporatethestationcoordinatesintooursolutionfor addedreliabilityandgeoreferencing.Todoso,wesimply Figure3.TheHannibalmappingcartinoperationinLucasCave,Jenolan.THREE-DIMENSIONALMOBILEMAPPINGOFCAVES196 N JournalofCaveandKarstStudies, December2014

PAGE 45

stoppedthecartdirectlyaboveanysurveymarkersseenon theground,leavingtheequipmentstationaryforapproximatelytenseconds.Theseeventsarethenstraightforward toidentifylaterinthedatastreamasstationarypoints inthetrajectory.Amatchingstepfoundageometric transformationbetweenthetrajectoryandtheground controlpoints,whichwerethenusedaslooseconstraints withinthenon-rigidoptimizationalgorithm.Duetoits continuousmotion,itisnotasstraightforwardtostop Zebedeeatsurveystations.Wearecurrentlyexperimenting withmethodstoallowustoautomaticallydetectwhenwe areatoneofthesepointstoenablegeoreferencingof Zebedeedatainasimilarmanner. Inadditiontothepointcloudmodelsproduced, watertightsurfacemodelswererequiredforuseinthe researchersairflowanalysis.Softwareforgenerating3D surfacemodelsfromthepointcloudswasdevelopedfor thispurpose(Holensteinetal.,2011).Onerealityof mappingwiththisequipmentinconfinedspacesisthatitis inevitablethatmembersofthemappingteamwill occasionallyneedtoenterthescannersfieldofview, therebyintroducingspuriouspointsintothepointcloud. Thesespuriouspointscanberemovedduringsurface reconstructionbymodelingeachmeasurementasaspacecarvingray.Ifaraysuccessfullypassesthroughthe locationofapointatothertimes,thenthepointcanbe disregardedascomingfromamovingobject.One disadvantageofconvertingthepointclouddatatoa smoothsurfaceinthismanneristhatmuchofthefine structureofthecavefeaturesislostinthesurfacemodel. AnexamplesurfacereconstructionfromChifleyand ImperialcavesisillustratedinFigure4.Detailedwatertightsurfacemodelscanbefabricatedintophysicalscale modelsusing3Dprintingtechnology(Baselgiaetal.,2014). Althoughsuccessful,clearlythesize,weight,and wheeledbaseofHannibalsuggestthatitisunlikelytobe Figure4.Detailed3DwatertightsurfacemodelofChifleyandImperialCavesgeneratedfromLiDARdataacquiredfromthe Hannibalplatform.ChifleyisaC-shapedpassageatalevelapproximately10maboveImperial.WhilescanningtheWilkinson Branch,datawerelostforabout30secondsduetoaUSBcablebecomingunplugged.Thedatalossresultedinakinkinthe trajectory,indicatedbytheredcircle.RescanningofthissectionwithZebedeeinthe2012visithasproducedacorrectedmodel ofthisarea(seecalloutbox).R.ZLOTANDM.BOSSEJournalofCaveandKarstStudies, December2014 N 197

PAGE 46

appropriateforthevastmajorityofnaturalcaves. However,thisinitialdeploymentonlyrequiredaccessto pavedpathwaysandstairsinshowcaves,which,withsome physicaleffort,isachievable.Theplatformalsoservedasa proof-of-conceptforLiDAR-basedmobilemappingin naturalcavesandfurthermotivatedthedevelopmentofthe morepracticalZebedeesystem. SubsequenttripstotheJenolanCaves,withZebedee systems,occurredinAugust2011,November2012,and May2013.In2011,severalcaveswerescanned,including Orient,River,PoolofCerberus,Lucas,Imperial,Jubilee, andEldercaves.In2012,Chifley,Imperial,Nettle,Elder, andMichaelmascavesweremapped,aswellasseveral kilometersofpathsonthesurfaceoutsideandabovethe caves.In2013,Orient,Baal,Lucas,River,andJubilee caveswerescanned.Onthelattertwotrips,multiple Zebedeeunitswereavailable,allowingdifferentareastobe mappedsimultaneouslyandlaterautomaticallycombined bytheprocessingsoftware.Figure2showsacombined mapmergingasubsetofthedatafromthe20112013scans ofJenolan;someofolder,redundantdataandafewareas onthesurfacehavebeenexcluded.DatafromtheTemple ofBaalcavescannedwiththeHannibalcartsystemin2010 arealsoincludedinthefigure,astheheightofthemain chamberisbeyondthemaximumrangeofZebedeeslaser scanner.TheresultillustratedisbasedonlyontheLiDAR dataanddoesnotincorporateanygroundcontrolpoints fromprevioussurveys.Aclose-upofasectionoftheOrient CavepointcloudgeneratedwiththeZebedeesystemis presentedinFigure5. Keydatacollectionstatisticsforthemodelpresentedin Figure2arepresentedinTable1.Overall,themap includesabout15.5hoursofcavescanning,collectedover 17.1kmoftraverse.Itshouldalsobenotedthattherewasa higherthanrequireddegreeofoverlapbetweenthe datasets,sothetotalscanningtimeandtrajectorylength couldbesignificantlyreducedwhilestillachievingthesame coverage.Weestimatethattwooperatorsfamiliarwiththe cavelayoutcouldcoverthesameareainasingleday. Withintheshowcaves,theaveragerateoftraverseusing Zebedeewas1.7km/h,approximately2.5timestheaverage speedachievablewithHannibal. TheZebedeesystemisnotlimitedtotouristshowcaves containingpathways,platforms,andstairs.Eveninthe caseoftheshowcaves,manysectionscouldnothavebeen navigatedwithHannibalduetosteepstairs,ladders, narrowpassageways,androughormuddyterrain.In addition,theterraininothercavesthatwerecoveredis morenatural.Forexample,ElderandMichaelmasCaves primarilyconsistofsmallchambersconnectedbynarrow passagesandsqueezes.Adetailedviewofthe3Dmodelof ElderCaveispresentedinFigure6.TheterraininElder Caveisrelativelyrugged,requiringscrambling,squeezing, andclimbingtotraverseit.Thecavegenerallyruns verticallyfromasinkholeopeningatthesurfacedownto aconnectionwithImperialCave.Our2011traverse followedthisroute,includingscanningwhileabseiling downthesinkhole.In2012,theroutestartedfrombelow, ascendingtothesinkhole,andthenbackdowntoImperial alongaslightlydifferentpath,takingabouttwohoursin total.Atafewofthemostchallengingsqueezesandclimbs, theZebedeeunitwaspassedthroughtheopeningtoa secondpersontoallowtheprimaryoperatoruseofboth hands.Thesehand-offscouldbecompletedwithout interruptingthedatastream,thustherearenobreaksin themap.ThepointcloudmodelofElderCavecanalso beenseenaspartoftheoverallmodeloftheJenolanCaves inFigure2,whereitmatcheswithImperialCaveatthe bottomandthesinkholeareaasscannedfromthesurface above.TheresultsfromElderCavehighlighttheversatility Figure5.PointcloudofasectionofOrientCave.Someof thewallsurfaceshavebeencutawaytorevealnatural formations,platforms,staircases,andhandrails. Table1.KeystatisticsfromJenolanCavesdatasetsincludedinthemodelshowninFigure2. TripYear TotalPatha,km CavePathb,km CaveTimeb,h:min 2010 0.4 0.4 0:49 2011 2.2 2.2 1:33 2012 11.4 5.3 4:56 2013 11.4 9.1 8:08 Total 25.5 17.1 15:26aTotalPathindicatestrajectorylengthcombiningbothcaveandexteriordatasets.bCavePathandCaveTimerefertothetrajectorylengthandscanningdurationwithinthecaves(excludingexteriordatasets).THREE-DIMENSIONALMOBILEMAPPINGOFCAVES198 N JournalofCaveandKarstStudies, December2014

PAGE 47

androbustnessofthesysteminmappingmorechallenging caveenvironments. TheJenolanCaveshadpreviouslybeenthesubjectof extensivesurveysusingmoretraditionalmethods.In Figure7,theZebedee-generatedmodeliscomparedtoa mapcompletedin1925byOliverTrickett(Middleton, 1991)andwith59surveystationsfromthemorerecent JenolanCavesSurveyProject(Jamesetal.,2009).The layoutofthecavesisgenerallyinagreementwiththe previoussurveys,thoughtherearesomeclearmisalignmentsbetweentheoutlines.Themostobviousdiscrepancy canbeseenontheleftsideoftheimage,where accumulatingyawdriftalonga500mbranchofImperial andJubileeCaveshasintroducedalargeerror.Inthiscase, oneofthedatasetscoveringthissectionofImperialwas collectedin2011,usinganolderversionoftheacquisition softwareinwhichnomagnetometerdatawasrecordedand timinginformationwaslessaccurate.Asaresult,thereis increaseddrifterror,andwithoutanopportunitytoclose thetraverse,therearenoconstraintstocorrecttheerror. Weplantorescanthissectioninanupcomingtrip,which weexpectwouldimprovethisareaofthemodel.Elsewhere inthecavesystem,thedifferencesareofamuchsmaller magnitude,andthefactthatthemajorityofthesurvey stationmarkersfallwithintheZebedeemapoutlineshow thatthemodelsagreetowithinafewmeters.Ingeneral,the Zebedeemapsarelocallyaccurate,anditisprimarilyover largeopentraversesthattherelativeerrorcanbecome significant.Althoughtheabsoluteaccuracyofthetraditionalsurveysappeartobesuperior,theadvantagesofthe Zebedeesystemarethattheresultscanbeproducedmuch moreefficiently,withinadayortwoforJenolan; automatically,astherequiredoperatorexpertiseis minimal;andatamuchhigherresolution,billionsof measurementsratherthanhundredsorthousands. ThedrifterrorapparentinthebranchthroughImperial toJubileeCaveinFigure7isindicativeofpotentialsystem performanceinsituationswheretheonlywaytosurveya caveorpassageisbyfollowingalong,opentraverse(i.e., whereitisnotpossibletoclosealoop).Wecanquantify thedriftbycomparingtheopen-looptrajectoryfromthe firstphaseofprocessingtothetrajectoryfromthefinal optimizedmodelinplaceswherethereareloopsavailable. Theanalysisiscarriedoutbycalculatingthedifference betweenfixed-lengthsegmentsoftheopenandclosed trajectoriesoftheZebedeehandle(whichreflectsthe distancetheoperatorhaswalkedratherthanthedistance thelaserhasmoved),byfirstaligningthesegmentsatthe startandthenrecordingthepositionalerroraccumulated bytheend.Theroot-mean-square(RMS)errorsare plottedasafunctionoftraverselengthinFigure8.The observedRMSerrorsgrowlinearlywithdistance,atarate of2to5percentofdistancetraveled.Inaccuracyin headingisthelargestcontributingsourceoferror.Twoof thedatasets,LucasandMonsMeg,exhibitrelativelylarger errorgrowthrates,possiblyduetothenatureofthe environments.Notethatthesameanalysiscarriedoutfor thelaserscannertrajectoryratherthanthehandle trajectoryresultsindriftratesofaroundtwo-thirdsofa percentofdistancetraveled.Theperformanceofthe SLAMalgorithmdependslargelyontheamountandtype of3Dstructurepresent.Thisbehaviorissomewhat differentthaninthecaseoftraditionalsurveymethods, wheretheaccuracydependsmoreheavilyontheequipment usedandthesurveyorsskill.KOONALDACAVEKoonaldaCaveisanarchaeologicallysignificantcave intheremoteNullarborPlaininSouthAustralia.Thecave consistsoftwolargeinterconnectedchambersontwo differentlevels(Fig.9).Theupperchamberisafairly linearpassageabout250minlengthand15to30mwide. Theceilingisdomed,withatypicalheightrangingbetween 4and20m,andthefloorcontainsmanyrockpiles primarilyduetoceilingcollapse.Thelowerchamberis T-shaped,withonemajorsectionrunningnorthsouthand anothereastwest,anditcontainsseveralsmalllakes.The ceilingsaregenerallyhigherthanintheupperchamber, typicallyrangingbetween10and30m.Attheirmain junction,thelowerchambersfloorisonlyafewmeters belowtheupperchambers,butceilingcollapseinthe upperchamberhasresultedinmuchofitsfloorbeing raisednearly20-mhigher.Thewesternendofthelower chambercontainsalakewitha23mhighroof,nearthetop ofwhichasmallbalconyconnectstotheupperchamber throughanarrowsqueeze.TheentrytoKoonaldaCaveis atthebottomofa20mdeepsinkhole,whichhasan openingtothesurfaceofapproximately60mby35m. Figure6.PointcloudmapofElderCavegeneratedfromthe 2012data.The3Drepresentationatcenteriscoloredby height.Projectionsontotheaxis-alignedplanesareshownin grey.Attheuppermostregionofthemodelisthesinkhole openingtothesurface.Thelongpassageatthebottomispart ofImperialCave.Theactualresolutionofthedataishigher thantheresolutionofthisdisplay.R.ZLOTANDM.BOSSEJournalofCaveandKarstStudies, December2014 N 199

PAGE 48

UnlikemanyofthecavesatJenolan,KoonaldaCavedoes notcontainanysignificantspeleothemformations,and typicallyitisfairlydry,withthenotableexceptionofthe lakesinthelowerlevel. KoonaldaCavehasbeenthesubjectofarchaeological andspeleologicalstudyforseveraldecades.Weare workingincollaborationwitharchaeologistsfromthe SouthAustralianMuseumandFlindersUniversitywho arestudyingthecaveandsurroundingregion.Thecave containsevidenceofhumanpresence,includingflint mining,believedtohaveoccurredbetween30,000and 10,000yearsago.Thesevisitorsalsoleftmarkingsonsome ofthesoftrocksurfacesusingfingersandothertools.The fingerflutingsformgroovesabout3to5mmdeepthat coverhundredsofsquaremetersofthewallsurfaces,which areinmanyplacesmadeupofasoft,powderycalcite material.Itisexpectedthathigh-resolution3Dmodelsof thecavewillbevaluableinprovidingdataforremote archaeologicalandgeomorphologicalstudy,aswellas providingavirtualmodelthatcanbeinteractively explored,asthecaveisnotaccessibletothepublic. OurinitialmappingexpeditiontoKoonaldatookplace inNovember2011,duringwhichasingleearlyZebedee systemwasavailableandonlytheupperlevelwasscanned. WereturnedtothecaveinDecember2012,whenmore completescanningofbothlevelswasperformedwith multipleZebedeesystems.Havingmultipleoperators enabledthemappingteamtoscandifferentroutesthrough thecavesimultaneously,providingbothredundancyand efficiency,aswecouldcoverdifferentsidesofthemany rockpilesmorequicklyinthismanner.Weestimatethatit wouldbefeasibletogetanoverallmodelofthecave structureusingasingleZebedeesysteminunderanhour, butwewereaimingforfairlydenseandcompletecoverage ofallvisiblesurfacesandthereforecoveredallreachable areasofthecavemeticulously.Inadditiontothemapping team,theexpeditionincludedateamofarchaeologists fromtheSouthAustralianMuseum,aphotogrammetry researchercapturingtheartworkinhigh-resolutiondetail, andrepresentativesfromtheMirning,whoarethe traditionalownersofthelandinwhichKoonaldaresides. Asignificantportionofthedatacollectionwasperformed Figure7.ComparisonofJenolanresultswithprevioussurveys.TheZebedeemodelisshadedinblue(caves)andgreen (exterior)andoverlaidonthemapproducedbyTrickett(1925).Theredspotsdenote59surveystationsfromtheJenolan CavesSurveyProject(19872005).TheZebedeetrajectoryhasbeenrigidlyalignedwiththesurveystationsusingarobust IterativeClosestPointalgorithm,andthesurveystationshavebeenalignedtotheTrickettmapmanually.Theregistration betweenthesurveystationsandtheZebedeetrajectoryhasbeenappliedusingarigidmodeltoensurethatthepresented Zebedeemapisbasedonindependentmeasurementsonly.THREE-DIMENSIONALMOBILEMAPPINGOFCAVES200 N JournalofCaveandKarstStudies, December2014

PAGE 49

byoneoftheMirningrepresentatives,whosuccessfully capturedoveranhourofdata,predominantlycoveringthe lowerlevelofthecave,afteronetotwominutesof instructiononhowtooperatethesystem. Figure9illustratesafull3Dmapofthescannedareaof thecave,includingthesinkhole,theupperandlowerlevels, andthesqueezearea.Thesqueezeisa3mlongpassage about30cmhighconnectingthenorthwestendoftheupper leveltoabalconyaboveoneofthelakesinthewesternend ofthelowerlevel.Thedatawererecordedbythreeoperators inatotalofthreehoursofacquisitiontime,butlesselapsed timeduetothemultipleunits.Theoperatorsgenerally traversedthecaveataslowwalkingpacethatvaried somewhatbasedonlocalterrain.Visualizingthemapin Figure9asa3Dmodelonscreencanfacilitatescientific modelingandanalysisofvariousaspectsofitshistory, includinggeomorphology,extentofformertwilightzones, andpossibleformerentrances.Acomparativeoverlay betweenthe1976surveymapandtheZebedee3Dmapis presentedinFigure10.Thetwomapsaregenerallyin agreement,thoughthereisaslightangulardifferenceatthe squeeze.Itisdifficulttoknowwhichisclosertothetrue structurewithoutfurtherindependentmeasurements;we note,however,thattheZebedeemodelclosesaloopthrough measurementsofthelowerlevelfromthebalcony,andwe areunawareofwhetherthetraversewasclosedinthe1976 survey.Anexampleclose-upviewofanareaintheupper levelknownastheRampartsispresentedinFigure11.This renderingillustratesthelevelofdetailavailableinthe3D pointclouddata.Ingeneral,thedensityofthepointcloudis afactorofhowquicklytheoperatortraversesthecaveand howmuchtimeisspentscanningaparticularareaof interest.Thelakesonthelowerlevelofthecave(notshown inthefigure)appearasemptyareasinthepointcloud, becausethebeamfromtheinfraredlaserisforthemostpart absorbedorspecularlyreflectedbywater,thoughitcan penetratewherethewaterissufficientlyshallow.CONCLUSIONSWehaveintroducedanewsystemthatenablesthe applicationofmobileLiDARmappingtechnologyto surveyingnaturalcaveenvironmentsforthefirsttime.The Zebedee3Dmappingsystemhasbeendemonstratedina varietyofcaves,withthemostextensiveresultsinJenolan CavesandKoonaldaCaveinAustralia.Theproposed methodoffersimprovementsovercurrentpracticeand state-of-the-arttechnologyinanumberofways.Mobility increasesefficiencybytransformingthemappingprocess intoacontinuousoneinwhichasinglepersoncansurveya caveinapproximatelythesametimeittakestotraverse throughit.Theportabilityandflexibilityoftheequipment ensuresthatitcangonearlyeverywhereitsoperatorcan, includingthroughtightsqueezes,upladders,anddown abseils.Coverageoftheenvironmentisachievedthrough mobility,ratherthanworryingaboutviewpointpositioning asinthecaseofstaticterrestrialLiDAR.Theequipment canbeoperatedbynon-expertswithalmostnotraining.In general,abriefcoachingsessionissufficient.Inaddition, workflowautomationforbothdataacquisitionand processingenablesnon-expertstogenerate3Dmodels directlyfromrawdataandpreventserrorsthatcanoccur withmanualtechniques.The3Dpointcloudmapsthatare generatedandthesurfacemodelsthatarecreatedfrom themaresignificantlymoredetailedandaccuratelocally comparedtotraditionalhandsketchesandcoarse3D modelsbasedonleft-down-up-rightmeasurementsat stations.Mapscanalsobetransformedintoageoreferencedcoordinateframeifsuitablecontrolpointswithinor GPSmeasurementsoutsidethecaveareavailableandcan beassociatedwiththeexistingdata.Whileroughmaps fromtraditionalsurveymethodscanbesuitableforgeneral navigationthroughacave,forsomeapplications,suchas scientificresearchandenvironmentalassessment,greater detailandresolutionarerequired. Therearesomelimitationstothecurrentsystem, severalofwhicharebeingaddressedasthetechnology progresses.Overlargescales,theaccuracyofthesystem canbelowerthantraditionalmethodsappliedwiththebest currentequipmentandexpertise.Weareworkingon advancementstothealgorithmsandhardwarethatshould improvethesystemperformanceovertime.Thecentimeter-scaleprecisionoftherangemeasurementsofthe Hokuyoscannerprecludethecaptureoffinedetailssuch Figure8.Positionalroot-mean-square(RMS)errorcurves calculatedforfivedatasets,eachofwhichformspartofa closedloopbutisnotitselfaloop.Thecurvesshowthe observederrorasafunctionoftheoperatorstraverselength, thatis,thedistancewalkedratherthanthedistancethe swayinglaserscannertraveled.Foreachdataset,theerroris computedbasedonthedifferencebetweentheendpointsof theopen-andclosed-loopsolutionsfordifferentsegment lengthsoftheZebedeehandletrajectory.R.ZLOTANDM.BOSSEJournalofCaveandKarstStudies, December2014 N 201

PAGE 50

Figure9.Three-dimensionalmapofKoonaldaCavegeneratedusingtheZebedeesystem.Theeyeiconindicatestheviewing directionofthesideelevation.Severalareasofinterestaremarked.Themodelconsistsofapproximately300millionpoints, eachofwhichiscoloredaccordingtotherelativelocalheightabovethecavefloor.Thepointcloudwasgeneratedusingfive separatedatasetsrepresentingunderthreehoursofdatacollection,someofwhichwasdoneinparallelbymultipleoperators. Thesurveyofthenorthpassagewasnotcompleted,asitcontainsdeeperlakesandwouldhaverequiredaboatorother equipmenttoproceed.Twoarchaeologicaltrenchesarevisibleinthesoutheastelevationview.THREE-DIMENSIONALMOBILEMAPPINGOFCAVES202 N JournalofCaveandKarstStudies, December2014

PAGE 51

asthefingerflutingsatKoonaldaandverythinspeleothem featuresatJenolan.However,incaseswherethesefeatures arerelativelylocalized,othertechniquessuchasphotogrammetryorstaticlaserscanningcanbeusedtoproduce high-resolutionmodelsthatcanbemergedintothepoint cloud(photogrammetryiscurrentlybeinginvestigatedfor modelingthefingerflutingsatKoonalda).Improvements inLiDARtechnologycouldeventuallyresultina lightweightscannerwithmillimeterprecisionandother improvedfeaturesthatcouldseamlesslybeincorporated intotheZebedeesystem.Theexistingequipmentcannotbe usedunderwater,thoughthecorealgorithmscouldbe adaptedforusewithsonarorothersuitablesensing modalities.Intheory,therearesometypesofenvironments Figure10.Overheadviewofa3DmapoftheupperlevelofKoonaldaCavegeneratedwiththeZebedeesystem.Aline drawingfroma1976map(usingtacheometer,5mmgraduatedstaff,andSuuntocompass)hasbeenmanuallyoverlaidfor comparison.Somedifferencesinthewalllocationscanbeattributedtothefactthatthe1976surveywassketchedata particularheight,whereasthisviewofthe3Dmaphighlightstheouterhullofthecavewalls.Notetheslightdifferences betweenthetwomapsintheupperleftnearthesqueezearea.The3Dmapconsistsofapproximately150millionpoints,eachof whichiscoloredaccordingtotherelativelocalheightabovethecavefloor.R.ZLOTANDM.BOSSEJournalofCaveandKarstStudies, December2014 N 203

PAGE 52

thataretheoreticallytroublesomeforthistechnology,but thesearenotexpectedtooccurincaves.Forexample,a verylarge(relativetothemaximumscanningrange) geometricallyfeaturelessvoidoralong,smoothtunnellikeenvironmentwouldmakeitdifficulttoestimatethe scannersmotioninalldimensions.Finally,thehardware currentlycoststhousandsofdollars,whichmaylimit affordabilityforsomecavesurveyapplications. Ourdatacollectionstrategythusfarhasprimarilybeen intendedtomapthecaves,andwedidnotspecificallyplan inexperimentalproceduresthatwouldprovideastraightforwardwaytoquantitativelycomparetheresultswith previoussurveys.Weintendtoaddressthisinupcoming fieldtripswhenwewillcollectdedicateddatasetsspecificallyforcomparisonpurposes.Wealsoplantoreplace someofourearlierdatasetswithdatacollectedfrommore up-to-datehardware,whichweexpectwillimprovethe overallaccuracyofthemaps. Two-dimensionalplanandsectionmapsarewidely usedfornavigationthroughcaves,andhigh-resolution3D modelsarenotnecessarilysuitableforthispurpose. Furtherprocessingcanbedonetoconvertthe3Dmodels intothestandard2Dsymbolicrepresentationsforprinting outonpaper.Anotherpossibilityisthat3Delectronicor evensolidrepresentationsofcavescouldbecomea standardnavigationtoolinthefuture.Wearefurther investigatingmethodsforcolorizingthepointclouds accordingtothevisualappearanceofthecaves,andhave recentlygeneratedpreliminaryresultstowardsthisgoalby addingasmallcameratothehandheldunit. AlthoughZebedeehasbeendeployedacrossawide rangeofmappingapplications,includingforests,mines, interiorsandexteriorsofbuildings,andindustrialsites,the conceptwasinitiallyinspiredbyimagininghowwecould adaptlarger,vehicle-bornetechnologyintoaformsuitable formappingcaves.Weenvisionthattheavailabilityofthis technologywillcreatenewopportunitiesforscientific studiesofnaturalcavesthatwerepreviouslyimpossible. Thefactthatthesystemcanbefullyautomatedalsoopens upthepossibilitythatsimilarhardwarecanbedeployedon roboticvehiclesincaveenvironmentstoodifficultor hazardousforhumanexploration.ACKNOWLEDGEMENTSWeacknowledgetheassistanceandsupportofanumber ofindividualsandorganizationsinmakingthiswork possible.PaulFlickislargelyresponsibleforthemechanFigure11.Viewofthe3DpointcloudviewedfrominsideKoonaldaCaveatthebottomoftheRamparts.Intheforeground areseveralboulders,beyondwhichisasteepslopeformedbysignificantceilingcollapse.Thepointcloudhasbeen downsampledto3cmrevolution.THREE-DIMENSIONALMOBILEMAPPINGOFCAVES204 N JournalofCaveandKarstStudies, December2014

PAGE 53

icalandelectronicdesignandassemblyofboththe ZebedeeandHannibalhardwareplatforms.Wewishto acknowledgethetraditionallandownersofbothofthe cavesitesvisited. Our2010fieldtriptotheJenolanCaveswaspartially supportedbytheAustralianNuclearScienceandTechnologyOrganisation(ANSTO).Wewishtoacknowledgein particularthecontributionofChrisWaringtherein helpingtocoordinatethefieldworkformultipletrips, assistingwithdatacollection,andintroducingtheopportunitytomapthecaves.WethanktheJenolanCaves ReserveTrustandthemanyguideswhohaveassistedour effortsovertheyears.Thesurveylocationsillustratedin Figure7arecourtesyoftheJenolanCavesSurveyProject. WealsothanktheSydneySpeleologicalSocietyand JenolanCavesHistoricalandPreservationSocietyfor assistanceinlocatingsourcematerials. WewishtothanktheMirningPeoplefortheirsupport andinvolvementwiththisprojectatKoonalda.Ourwork therehasbeenpartiallysupportedbytheDepartmentof Industry,Innovation,Science,ResearchandTertiary EducationthroughanInspiringAustraliaUnlocking AustraliasPotentialgrantandbytheSouthAustralian Museum.WewishtoacknowledgeKerynWalsheofthe SAMuseumandFlindersUniversityforinspiringthis effortandcoordinatingthefieldwork.WealsothankSA Parks&WildlifeandtheSouthAustralianStateEmergencyServiceforprovidingassistancewithaccessand entrytothecave.The1976surveydataappearingin Figure10wasproducedbyI.D.LewisandK.R.Mott, CaveExplorationGroupSouthAustraliaandtheSA MuseumandprovidedcourtesyoftheSAMuseum Archives/Archaeology/Gallus/Koonalda. WealsowishtothankandacknowledgeClaudeHolenstein,CirilBaselgia,andHansMoorkensofthe CommonwealthScientificandIndustrialResearchOrganisation(CSIRO)fortheircontributionsindevelopingthe 3Dsurfacereconstructionsoftwareusedtogeneratethe mapsinFigure4;PeymanMoghadam(CSIRO),Stuart Hankin(ANSTO),andMichaelLaing(Mirning)for assistancewithdatacollection;EliseBossefordeveloping theweb/smartphoneinterfacetotheZebedeesystem;and AaronMorrisforinitialinspirationtowardsaddressingthe challengeofmobilecavemapping.REFERENCESBaselgia,C.,Bosse,M.,Zlot,R.,andHolenstein,C.,2014,Solidmodel reconstructionoflarge-scaleoutdoorscenesfrom3DLiDARdata, in Yoshida,K.,andTadokoro,S.,eds.,FieldandServiceRobotics: Berlin,SpringerTractsinAdvancedRobotics92,p.541. doi:10.1007/978-3-642-40686-7_36. Bosse,M.,andZlot,R.,2009,Continuous3Dscan-matchingwitha spinning2Dlaser, in ICRA:IEEEInternationalConferenceon RoboticsandAutomation,p.431219.doi:10.1109/ROBOT.2009. 5152851. Bosse,M.,andZlot,R.,2013,Placerecognitionusingkeypointvotingin large3DLiDARdatasets, in 2013IEEEInternationalConferenceon RoboticsandAutomation(ICRA),p.2677.doi:10.1109/ ICRA.2013.6630945. Bosse,M.,Zlot,R.,andFlick,P.,2012,Zebedee:Designofaspringmounted3Drangesensorwithapplicationtomobilemapping:IEEE TransactionsonRobotics,v.28,no.5,p.110419.doi:10.1109/ TRO.2012.2200990. Canter,P.,andStott,A.,2011,Mappinginteriorspaceswithspeed,ease& accuracy:TheAmericanSurveyor,v.8,no.4 Davis,D.G.,andLand,L.,2006,RecentlydiscoveredpassagesinFort StantonCave,NewMexico,andimplicationsforspeleogenesisand regionalgeomorphicprocessesinthenorthernSacramentoMountains, in Land,L.,Lueth,V.W.,Raatz,W.,Boston,P.,andLove,D.L.,eds., CavesandKarstofSoutheasternNewMexico:NewMexicoGeological Society57thAnnualFieldConference,p.2196. Dryjanskii,M.,2010,ThesubterraneanworldofEasterIsland:GeoInformatics,v.13,no.1,p.6. Durrant-Whyte,H.,andBailey,T.,2006,Simultaneouslocalizationand mapping(SLAM):PartItheessentialalgorithms:IEEERobotics& AutomationMagazine,v.13,no.2,p.99.doi:10.1109/MRA. 2006.1638022. Fairfield,N.,Wettergreen,D.,andKantor,G.,2010,SegmentedSLAM inthree-dimensionalenvironments:JournalofFieldRobotics,v.27, no.1,p.85.doi:10.1002/rob.20320. Gary,M.O.,Fairfield,N.,Stone,W.C.,Wettergreen,D.,Kantor,G.,and Sharp,Jr.,J.M.,2008,3DmappingandcharacterizationofSistema ZacatonfromDEPTHX(DEepPhreaticTHermaleXplorer), in Yuhr, L.B.,Alexander,Jr.,E.C.,andBeck,B.F.,eds.,Proceedingsofthe 11thMultidisciplinaryConferenceonSinkholesandEngineeringand EnvironmentalImpactsofKarst:AmericanSocietyofCivilEngineers GeotechnicalSpecialPublicationno.183,p.202.doi:10.1061/ 41003(327)20. Heeb,B.,2008,Paperlesscavinganelectroniccavesurveyingsystem, in Gonon,T.,ed.,Proceedingsofthe4thEuropeanSpeleological Congress,Vercors2008:Lyon,Fe de rationfrancaisedespe le ologie, SpeluncaMemoires33,p.130. Holenstein,C.,Zlot,R.,andBosse,M.,2011,Watertightsurface reconstructionofcavesfrom3DLiDARdata, in 2012IEEE/RSJ InternationalConferenceonIntelligentRobotsandSystems, p.3830.doi:10.1109/IROS.2011.6095145. Hunter,D.,2010,Afieldtrialofcommonhand-heldcavesurvey instrumentsandtheirreaders,BullitaCaveSystem,July2010:Caves Australia,no.183,p.10. James,J.M.,Martin,D.J.,Tunnock,G.M.,andWarild,A.T.,2009,A cavesurveyforresearchandtouristcavemanagement, in White, W.B.,ed.,Proceedings15thInternationalCongressofSpeleology: Huntsville,NationalSpeleologicalSociety,v.3,p.1381. Kershaw,B.,2012,Managingthesurveyinformationofthecavesof Judbarra/GregoryNationalPark,NorthernTerritory:Helictite, v.41,p.87. Liu,T.,Carlberg,M.,Chen,G.,Chen,J.,Kua,J.,andZakhor,A.,2010, Indoorlocalizationandvisualizationusingahuman-operated backpacksystem, in Mautz,R.,Kunz,M.,andIngensand,H.,eds., Proceedingsofthe2010InternationalConferenceonIndoor PositioningandIndoorNavigation:IEEE,10p.doi:10.1109/ IPIN.2010.5646820. McIntire,D.,2010,Laserscanningmushpotcave:TheAmerican Surveyor,v.7,no.9,p.18. Middleton,G.J.,1991,OliverTrickett:DoyenofAustraliasCave Surveyors,184734:SydneySpeleologicalSocietyOccasionalPaper no.10,156p. Morris,A.,Ferguson,D.,Omohundro,Z.,Bradley,D.,Silver,D.,Baker, C.,Thayer,S.,Whittaker,C.,andWhittaker,W.,2006,Recent developmentsinsubterraneanrobotics:JournalofFieldRobotics, v.23,no.1,p.35.doi:10.1002/rob.20106. Nuchter,A.,Surmann,H.,Lingemann,K.,Hertzberg,J.,andThrun,S., 2004,6DSLAMwithanapplicationinautonomousminemapping, in Proceedings,2004IEEEInternationalConferenceonRoboticsand Automation,ICRA,p.199803.doi:10.1109/ROBOT.2004. 1308117. Osborne,R.A.L.,Zwingmann,H.,Pogson,R.E.,andColchester,D.M., 2006,CarboniferousclaydepositsfromJenolanCaves,NewSouth Wales:implicationsfortimingofspeleogenesisandregionalgeology: AustralianJournalofEarthSciences,v.53,no.3,p.377. doi:10.1080/08120090500507362.R.ZLOTANDM.BOSSEJournalofCaveandKarstStudies, December2014 N 205

PAGE 54

Petrie,G.,2010,Mobilemappingsystems:Anintroductiontothe technology:GeoInformatics,v.13,no.1,p.32. Ruther,H.,Chazan,M.,Schroeder,R.,Neeser,R.,Held,C.,Walker,S., Matmon,A.,andHowritz,L.K.,2009,LaserscanningforconservataionandresearchofAfricanculturalheritagesites:Thecase studyofWonderwerkCave,SouthAfrica:JournalofArchaeological Science,v.36,p.1847.doi:10.1016/j.jas.2009.04.012. Sadier,B.,Delannoy,J.-J.,Benedetti,L.,Bourle`s,D.L.,Jaillet,S., Geneste,J.-M.,Lebatard,A.-E.,andArnold,M.,2012,Further constraintsontheChauvetcaveartworkelaboration:Proceedingsof theNationalAcademyofSciencesoftheUnitedStatesofAmerica, v.109,no.21,p.800206.doi:10.1073/pnas.1118593109. Slavova,T.,2012,Modernmethodsanddevicesformappingunderground galleriesandnaturalcaves, in Proceedingsofthe4thInternational ConferenceonCartographyandGIS.7p. Stone,W.C.,amElde,B.A.,Wefer,F.L.,andJones,N.A.,2000, Automated3Dmappingofsubmarinetunnels, in Stone,W.C.,ed., Robotics2000:FourthInternationalConferenceandExposition/ DemonstrationonRoboticsforChallengingSituationsandEnvironmentAmericanSocietyofCivilEngineers,p.148.doi:10.1061/ 40476(299)19. Warild,A.,2007,Vertical,5thedn.:AlanWarild,206p. Waring,C.,Wilson,S.,Hurry,S.,andGriffith,D.,2009,Cave speleothemgrowthresponsetoexternalweatherfromcontinuouscaveatmosphere(CO2)anddrip-waterchemistry(DIC)isotopicmeasurement[abs.]:GeophysicalResearchAbstracts,v.11, 11778p. Zlot,R.,andBosse,M.,2014,Efficientlarge-scalethree-dimensional mobilemappingforundergroundmines:JournalofFieldRobotics,v. 31,no.5,p.758.doi:10.1002/rob.21504.THREE-DIMENSIONALMOBILEMAPPINGOFCAVES206 N JournalofCaveandKarstStudies, December2014

PAGE 55

IMPROVEDKARSTSINKHOLEMAPPINGINKENTUCKY USINGLIDARTECHNIQUES:APILOTSTUDYINFLOYDS FORKWATERSHEDJUNFENGZHU*,TIMOTHYP.TAYLOR,JAMESC.CURRENS,ANDMATTHEWM.CRAWFORDKentuckyGeologicalSurvey,UniversityofKentucky,504RoseStreet,228MMRB,Lexington,Kentucky40506Abstract: TheexistingsinkholedatabaseforKentuckyisbasedonlow-resolution topographicmapscreatedmorethanfiftyyearsago.LiDAR(LightDetectionand Ranging)isarelativelyrecenttechniquethatrapidlyandaccuratelymeasuresfeatureson earthssurfaceinhigh-resolution.TotestthefeasibilityofusingLiDARtomap sinkholesinKentucky,wehavedevelopedamethodofprocessingLiDARdatato identifysinkholesandtestedthemethodinportionsoftheFloydsForkwatershedin centralKentucky.Themethodconsistedoffoursteps,creatingahigh-resolutiondigital elevationmodel(DEM)fromLiDARdata,extractingsurfacedepressionfeaturesfrom theDEM,inspectingthedepressionfeaturesforprobablesinkholes,andverifyingthe probablesinkholesinthefield.Atotalof1,683probablesinkholeswereidentifiedinthe studyarea,comparedto383previouslymappedforthesamearea.Wefield-checked121 randomly-selectedprobablesinkholesandconfirmedthat106ofthemwerekarst sinkholes.Thismethodincreasedthenumberofsinkholesbyafactoroffourwitha successratebetween80%and93%forthestudyarea,demonstratingthattheLiDAR sinkhole-mappingmethodisreliableandefficient.Thismethodidentifiedapproximately 55%ofthepreviouslymappedsinkholes,andapproximately98%ofthemissedsinkholes appearedtobefilledorcoveredforurbandevelopmentandagriculturepurposes.The nextstepistoextendthismethodtoprovidehigh-resolutionsinkholemapsforother karstareasinKentuckywhereLiDARdatabecomeavailable.INTRODUCTIONDetailedmappingofsin kholesiscriticalinunderstandinghydrologicalproces sesandmitigatinggeologicalhazardsinkarstlandscapes.Sinkholesaresurface depressionsthatforminplaceswherecarbonaterocks aredissolvedfromwaterandoverlyingsoilparticlesare carriedawayunderground,causingthesurfacetosubside gentlyorcollapsesuddenly(FordandWilliams,1989; Currens,2002;Brinkmann,2013).Therearethree generaltypesofsinkholes,diss olutionsinkholes,coversubsidencesinkholes,andcover-collapsesinkholes(Tihansky,1999).Sinkholesserveasamajorconnection betweensurfacewaterand groundwaterbycollecting rainfallanddrainingitinternallyintothesubsurface. Sinkholescancausedamagetoprivatepropertyandcivil infrastructuresuchasbuildingsandroads.Covercollapsesinkholes,whichoccurwhenthematerial overlyingsubsurfacevoidscollapses,cancausedamage tobuildingsandroads,farmponds,andfarming equipment(Currens,2002).Becauseoftheirfixed cross-sectionarea,sinkhol esarepronetooverflowand flooding.Somesinkholescanactasspringsand dischargewatertothesurfaceduringintensestorms (Currens,2002).Dingeretal.(2007)estimatedthe damagesassociatedwithsinkholesinKentuckywereapproximately$23milliondoll arsperyear.Consequently, existingland-useplanninginkarstareasoftenrelieson detailedmappingofsinkholes(Fleury,2009). SomesinkholescanberecognizedfromtheUSGS 1:24,000scaletopographicmaps.Thesetopographicmaps includecloseddepressionfeatures,oftenindicativeof sinkholesinkarstterrains.Inthelastfewdecades,several stateshavedevelopeddigitalsinkholedatabasesbasedon thetopographicmaps(Beck,1984;Floreaetal.,2002; Payloretal.,2003;Alexanderetal.,2013).Developinga sinkholedatabasefromtopographicmapsfirstrequires digitizationofthecloseddepressions,whichisoftenlabor intensivebecausethenumberofsinkholesiscommonlyin thethousandsonaregionalscale(Floreaetal.,2002).In addition,thetopographicmapshaveelevationcontour intervalsof3m,6m,orhigher,resultinginshallowand smallsinkholesbeingoverlooked.Furthermore,most USGStopographicmapswerecreatedpriortothe1970s, andmanynewsinkholesmayhavedevelopedsincethen. Althoughpeoplerecognizethatnotallcloseddepressions illustratedinthesetopographicmapsaresinkholes, extensivefieldverificationofthedepressionsrarelyoccurs, becausetheprocessisslowandcostly. Remote-sensingdatahavelongbeenrecognizedas usefulinlocatingsinkholes(Newton,1976).Highresolution,high-accuracydataobtainedfrommodern *Correspondingauthor:Junfeng.zhu@uky.eduJ.Zhu,T.P.Taylor,J.C.Currens,andM.M.CrawfordImprovedkarstsinkholemappinginKentuckyusingLiDARtechniques:a pilotstudyinFloydsForkWatershed. JournalofCaveandKarstStudies, v.76,no.3,p.207.DOI:10.4311/2013ES0135JournalofCaveandKarstStudies, December2014 N 207

PAGE 56

remote-sensingtechnologyprovideopportunitiestoimprovesinkholemapping.Forexample,Littlefieldetal. (1984)appliedLandsatimagestostudytherelationship betweenlineamentsandsinkholesinwest-centralFlorida. Dingeretal.(2007)usedaspectrumenhancementmethod on1mresolutionnatural-colorimagestoextractcircular shapesthatrepresentedareaswithdifferentvegetation signaturesthansurroundingareas.Someofthecircular shapeswerefoundinthefieldtobeactivesinkholes.Inthis study,weusedLiDAR(LightDetectionandRanging)to improvesinkholemapping.LiDARisaremote-sensing techniquethatrapidlyandaccuratelymeasuresfeatureson theearthssurfacebysendingoutshortlaser-lightpulses andmeasuringtheirreturnsfromanaircraftoraterrestrial platform.Alaserpulsecanhaveoneormultiplereturns becausethepulsecanencountermultiplereflectionsurfaces whenittravelstowardtheearthssurface.Collected LiDARdata,calledpointclouds,areoftenpost-processed toclassifythepointsintoseveralcategories,including ground,vegetation,building,andwater.LiDARexcelsin revealingsmallsurfacefeaturesandhasbeenwidelyusedin studyingnaturalresourcesandtheenvironment(Evans andHudak,2007;Floydetal.,2011;Crawford,2012). LiDARhasalsobeenappliedinstudyingsinkholesinsome otherstates.Seale(2005)andSealeetal.(2008)used LiDAR,alsocalledairbornelaserswathmapping,tomap sinkholesinPinellasCounty,Florida.Theysuggestedthat contemporaneousaerialphotographsshouldbeusedin conjunctionwithLiDARforreliablesinkholemapping. Rahimietal.(2010)andRahimiandAlexander(2013) appliedLiDARtoverifysinkholesmappedinthe1980s and1990sinWinonaCounty,Minnesota.Theyfoundthat mostoftheinventoriedsinkholesthathadnotbeenfilled laterforagriculturaluseswerevisibleusingLiDAR. MukherjeeandZachos(2012)usedasink-fillingmethod toidentifysinkholesfromLiDARandfoundanexcellent matchbetweenLiDAR-identifiedandactualsinkholesin Nixa,Missouri.TotestthefeasibilityofLiDARin providingaccurateanddetailedsinkholeinformationfor Kentucky,wedevelopedasinkhole-mappingmethodbased onLiDARpointcloudsandappliedthemethodinasmall karstwatershedincentralKentucky.STUDYAREAThestudyarea,FloydsForkWatershed,islocated approximately16kmeastofLouisville,Kentucky(Fig.1). Thewatershedconsistsoftwo10-digitUSGShydrologic unitsanddrainspartsofBullitt,Henry,Jefferson,Oldham, Shelby,andSpencercounties,coveringapproximately 736km2.TheFloydsForkstreamoriginatesinthe southwesternportionofHenryCountyandflowssouthwesttotheSaltRiver,whichflowstotheOhioRiver.The areahasasubtropicalclimatewithaverageannual precipitationof117cm(NationalDroughtMitigation Center,2013). MostoftheFloydsForkwatershedisintheOuter Bluegrassphysiographicregion,andasmallsouthwest portionofthedownstreamwatershedisintheKnobs region(Fig.1)(Woodsetal.,2002).TheOuterBluegrass regionhaslowtomoderatereliefwithvariablesoildepth rangingfromthickoverlimestonetothinovershales (Newell,2001).Theregionisunderlainbylimestones, dolomites,andshalesofLateOrdovicianandSilurianage. Themajorformationsare,fromoldesttoyoungest,the GrantLakeLimestone,theBullForkFormation,the DrakeFormation,theOsgoodFormation,theLaurel Dolomite,theWaldronShale,andtheLouisvilleLimestone.TheOsgoodFormationandtheWaldronShaleare composedofmostlyshale(90%orhigher)andverylittle dolomite.Theremainingunitsarecarbonaterocks(i.e., limestoneanddolomite)withsmallamountsofcalcareous shale.Mostkarstdevelopmentoccursintheseformations. TheKnobsregionisdominatedbyroundedhills,ridges, andnarrow,high-gradientvalleys(Woods,etal,2002). MostoftheKnobsregionisnon-karstandisunderlainby diverseshale,mudstone,andlimestonesedimentaryrocks ofSilurianandMississippianage.DATAANDMETHODThedatausedinthesinkhole-mappingmethod includedmainlyLiDARpointcloudsandaerialphotography.TheLiDARdatawereprovidedbytheLouisville/ JeffersonCountyInformationConsortium(LOJIC) throughtheKentuckyDivisionofGeographicInformation andcoverBullitt,Jefferson,andOldhamCounties.LiDAR datawerecollectedinMarch2009withanaveragepoint spacingof1mandaverticalroot-mean-squareerrorof 8.8cm.TheLiDARpointswerepost-processedinto severalcategories.Thecategoriesassociatedwithphysical featuresincludeground,lowvegetation,mediumvegetation,highvegetation,building,andwater.Theactualstudy area,whichexcludednon-karstareas,wasapproximately 580km2,or79%ofthewatershed(Fig.1).BingMapswas theprimaryaerialphotographyusedforthisstudy.Data fromBingMapswereimporteddirectlyintoArcMap10.1 (ESRI,2012)asbasemaps.InadditiontoBingMaps,we alsousedGoogleEarthhistoricimagesandaerial photographycollectedbyLOJICattwodifferenttimes, onein2009andtheotherin2012. Thesinkhole-mappingmethodhasfoursteps,building adigitalelevationmodel(DEM)fromLiDARpoint clouds,extractingsurficialdepressionfeaturesforthe DEM,inspectingthedepressionfeaturesforprobable sinkholes,andfield-checkingtheprobablesinkholes.The firstthreestepswerecarriedoutinArcMap10.1. Inthefirststep,theLiDARgroundpointswereusedto createaDEMwithacellsizeof1.5musinganaverage binningmethod.Theaveragebinningmethodcalculates theelevationforeachcellbyassigningtheaveragevalueof allpointsinthecell.MoresophisticatedinterpolationIMPROVEDKARSTSINKHOLEMAPPINGINKENTUCKYUSINGLIDARTECHNIQUES:APILOTSTUDYINFLOYDSFORKWATERSHED208 N JournalofCaveandKarstStudies, December2014

PAGE 57

methods,suchaskriging,couldpotentiallybeusedforthis application.BecausethecellsizeoftheDEMwaslarger thantheLiDARpointspacing,meaningatleastone measurementisavailableforeachcell,theaveragebinning methodwasconsideredsufficientforthisstudy. DepressionfeatureswereextractedfromtheDEMat thesecondstep.AfilltoolinArcGISwasusedtoidentify depressionfeaturesontheDEM.Thefilltoolwas originallydevelopedtoremovesmalldepressionsresulting fromdatanoise;hereitwasusedtofindnatural Figure1.Locationandgeologyofthestudyarea.TheFloydsForkwatershedisindicatedbytheirregularblackline,withthe bedrockgeologyshownforthekarstportionthatisthestudyarea.ThethinpurplelineistheboundarybetweentheBlueGrass andKnobsregions.J.ZHU,T.P.TAYLOR,J.C.CURRENS,ANDM.M.CRAWFORDJournalofCaveandKarstStudies, December2014 N 209

PAGE 58

depressions.ThetoolfillsdepressionsontheDEMwithan optionaluser-specifiedmaximumsinkdepth.Alldepressionsthatarelessthanthemaximumdepthandarelower thantheirlowestadjacentneighborwillbefilledtothe heightoftheirpourpoints.Weusedamaximum depressiondepthof6m,whichallowsidentificationof sinkholesthatarelessthan6mdeep.Weconsideredthis depthsufficienttoidentifymostnaturalsinkholesinthe studyarea.MukherjeeandZachos(2012)foundthata4-m depththresholdwassufficientforidentifyingexisting sinkholesinNixa,Missouri.Thefilltoolgeneratedanew filledDEM,andthedepressionswerethenextractedby subtractingthefilledDEMfromtheoriginalDEMto createadepressionraster. Thedepressionsidentifiedinthefillprocedureincluded depressionsthatwerenotsinkholes.Furthermore,wewere interestedinlocatingsinkholesthatwereconsideredto haveasignificanthydrologicimpactintheFloydsFork watershed,thatis,sinkholeswithrelativelylargedrainage areas.Forthisreason,weselectedthedepressionswithan arealargerthan46m2anddepthofgreaterthan0.3mfor furtherprocessing.Theselecteddepressionswerethen convertedfromrasterformattopolygonformat.These polygonswerefurthersmoothed,andtheholesinsidesome polygonsthatwereartifactsoftheraster-to-polygon conversionprocesswereremoved.Theseproceduresfor generatingdepressionpolygonsfromLiDARcanbe accomplishedbysequentiallyusingseveralArcGIScommands,includingLASDatasetToRaster,Fill,Raster Calculator,RastertoPolygon,SmoothPolygon,and EliminatePolygonPart.Tostreamlinetheseprocedures, webuiltamodeltoolwithasingleuserinterface.The modeltoolaskstheusertoprovideasingleinput,the LiDARdataset,andthenexecutestheaforementioned commandswithdefaultparametervalues(DEMcellsize, filldepth,depressionarea,smoothtolerance,etc.)automaticallytocreatedepressionpolygons.Thetoolalso allowsausertochangetheparametervaluesonthesame userinterface. Inthethirdstep,everypolygonwasvisuallyinspected andmanuallyclassifiedintooneofthreecategories, probablesinkholes,suspicioussinkholes,andnon-sinkholes.Ashaded-reliefmapwith53verticalexaggeration wascreatedfromtheLiDARDEMtoamplifytheshape anddepthofthedepressionfeatures.Theshaded-relief mapalongwithaerialphotographywasusedtoclassifythe polygons.Sealeetal.(2008)andAlexanderetal.(2013) alsousedaerialphotographytohelpidentifysinkholesin theirstudies.Toensureaconsistentclassification,the polygonclassificationwascarriedoutbyaprocedure consistingofaninitialclassification,areview,and discussion.Theinitialclassificationandthereviewwere conductedbydifferentindividuals.Thereviewresultswere thendiscussedtoreachthefinalclassifications.Although manypolygonsneededtobeinspected,manyofthem wereunambiguouslystreamchannels,water-filledponds, swimmingpools,anddrainagestructuresandwerevery easilyandquicklyidentifiedasnon-sinkholes.Ontheother hand,naturalsinkholestendtohaveacircularorelliptical shapeandmanyofthemhaveoneormoreinternal drainagepoints(i.e.,throats)thatarereadilyvisibleonthe shaded-reliefmap.Onoccasions,theclassificationprocedurecouldnotleadustoadecision,andtheseambiguous polygonswereassignedtosuspicioussinkholes. Inthefourthstep,probablesinkholeswererandomly sampledforfield-checking.Tocreatearandomsampleof probablesinkholesovertheentirearea,wefirstdividedthe areabycreatinga3,000-by-4000ftgrid,producingasetof cellslargerthanthenumberofprobablesinkholes,and thenrandomlyselectingonesinkholefromeachcellthat containedatleastone,creatingapoolofprobable sinkholesfromwhichthosetobefield-checkedwere randomlyselected.FieldinvestigatorsusedaGPS-enabled iPadwithamapshowinglocationsoftheselectedprobable sinkhole.TheiPadtrackedlocationsofthefieldinvestigatorsinreal-timeinrelationtothelocationofeachtargetto minimizelocationerrors;sincesinkholesinthestudyarea generallyoccurinclusters,itiseasytocheckthewrong location.Tocheckaprobablesinkholeinthefieldwe consideredwhetherthefeaturewasadepression,whether drainholesexistedinsidethefeature,whetherman-made structure(s)existedwithinthefeature,whethertherewas vegetationwithinthefeature,andwhetherwaterexisted withinthefeature.Thesamesamplingandfield-checking methodswerealsousedforsuspicioussinkholes.RESULTSANDDISCUSSIONWeextractedapproximately10,720depressionpolygonsinthestudyareafromtheDEMcreatedfromthe LiDARdataset.Amongtheextractedpolygons,1,696were classifiedasprobablesinkholesand282assuspicious sinkholes.Approximately10%oftheprobablesinkholes fromBullittCountyand5%fromJeffersonandOldham Countieswereselectedforfield-checking.Excludingthe samplesthatwereinaccessible,mostlyduetoabsentlandowners,wefield-checked121probablesinkholesand confirmed106ofthem(88%)assinkholes(Fig.2).We alsorandomlyselectedandfield-checked18suspicious sinkholesandfound5ofthemwereactualsinkholes.The totalnumberofactualsinkholesdetectedintheLiDAR datawouldbe,basedonthefield-checkingstatistics,1563. TheLiDAR-derivedsinkholecoverageisavailabletothe publicontheKentuckyGeologicalSurveysonlinemap service(http://kgs.uky.edu/kgsmap/kgsgeoserver/viewer.asp). Thelargenumberofpolygonsgeneratedindicatedthat thedepression-extractionprocedurewaseffectivein locatingsurfacedepressions.Althoughsomeofthe polygonswereassociatedwithsinkholes,morethan80% ofthemwerestreamchannels,ponds,orroaddrainsor otherman-madestructures.Thenumberofsinkhole-like depressionscanbereducedbyusinganautomatedIMPROVEDKARSTSINKHOLEMAPPINGINKENTUCKYUSINGLIDARTECHNIQUES:APILOTSTUDYINFLOYDSFORKWATERSHED210 N JournalofCaveandKarstStudies, December2014

PAGE 59

procedure.Forexample,Miaoetal.(2013)illustratedone suchprocedurebyusingarandomforestalgorithmthat usesshapeanddepthparameterstoextractcircular-or elliptical-shapedsinkholesfromdepressions.Inourstudy area,wefoundthatalthoughmostsinkholeshavea circularorellipticalshapes,somesinkholeshavemore complicatedshapesandmaypotentiallybeexcludedbyan automatedprocedure. Thepolygon-classificationprocesswasactuallyquite fastandeffective,becausemanypolygonswereeasyto identifywhenusingshaded-reliefmapsandhigh-resolution aerialphotography(Fig.4).Polygonsassociatedwith streamchannelsweretheeasiesttoidentifyasnonsinkholes.Onaerialphotography,eachhadanelongated shapeandoverlappedstreamchannels;onashaped-relief map,eachhadasmoothandflatbottom.Thesmoothed bottomswereartifactsontheDEMresultingfromLiDAR beamsbeingabsorbedatthewatersurface.Polygons associatedwithwater-filledpondswerealsoeasilyidentifiableasnon-sinkholes,becausethesepolygonsalsohad flatbottomsonshaded-reliefmaps.Polygonsassociated withman-madestructuresthathaveunnaturaland irregularshapeswereeasilyidentifiableasnon-sinkholes fromaerialphotography.Ontheotherhand,some polygonsassociatedwithcover-collapsesinkholeswere readilyidentifiable.Onashaded-reliefmap,thosepolygons hadaninternaldrainthatshowedasaholeorthroatinside thedepression.Suchpolygons,whenshownbyaerial photographytobeinaforestedareaofclusteroftrees surroundedbygrassland,werelikelytobetruecoverFigure2.Field-checkingresultsforrandomlysampledprobablesinkholesidentifiedfromtheLiDARdata.J.ZHU,T.P.TAYLOR,J.C.CURRENS,ANDM.M.CRAWFORDJournalofCaveandKarstStudies, December2014 N 211

PAGE 60

collapsesinkholes,butifthosepolygonswerecloseto residentialhousesorroads,theycouldbeeithersinkholes orman-maderetentionbasinswithdrainsthatappearedas holes.Polygonsassociatedwithsubsidencesinkholeswere hardertoscreen.Onashaded-reliefmap,theyappearedas shallow,bowl-shapeddepressions;onaerialphotography, theycouldbeinthemiddleofafarmfieldorclosetoa residentialarea.Thesepolygonscouldhavebeeneither sinkholesorpondsthatwereemptywhenLiDARwas flown.Forthesepolygons,weusedhistoricalaerialimages todetermineifthesefeatureswerenaturalsinkholesor man-madeponds. Thefield-checkingofprobablesinkholesshowedan 88%successrate.However,thefield-checkedsinkholes wereapproximately7%ofallthesinkholesweidentified. Tounderstandtheoverallsuccessrateandthemarginof errorforthestudyarea,weconsideredthisproblemasa binomialdistributionwithtwopossibleoutcomes,sinkhole andnon-sinkhole,andusedsamplestatisticstoestimate populationparameters.Theestimatedproportion,i.e.,the successrate( p )andstandarddeviation( ss)are(Zar,1999): p ~ X n and ss~ p 1 { p n { 1 1 { n N r 1 where N issizeofthepopulation, n isthenumberof samples,and X isthenumberofsuccessinthesamples.We identified1,696probablesinkholesfromLiDARandfieldchecked121.Amongthe121field-checkedsinkholes,there are106sinkholes,9non-sinkholes,and6inconclusive. Consideringalltheinconclusiveasnon-sinkholes,the Figure3.ComparisonbetweensinkholesmappedfromtheLiDARdataandthosepreviouslymappedfromtopographicmaps.IMPROVEDKARSTSINKHOLEMAPPINGINKENTUCKYUSINGLIDARTECHNIQUES:APILOTSTUDYINFLOYDSFORKWATERSHED212 N JournalofCaveandKarstStudies, December2014

PAGE 61

Figure4.Examplesofpolygonclassification,showingpolygonsoverlainwithshaded-relief(left)oraerialimages(right):a) Examplesofprobablesinkholes.b)Examplesofnon-sinkholes.c)Examplesofsuspectedsinkholes.Theshaded-reliefmapsare 53verticallyexaggerated.J.ZHU,T.P.TAYLOR,J.C.CURRENS,ANDM.M.CRAWFORDJournalofCaveandKarstStudies, December2014 N 213

PAGE 62

estimatedsuccessrateandstandarddeviationare0.88and 0.03,respectively. Toestimatethemarginoferror,wecalculated95%of confidenceintervalbasedonZar(1999),whichcalculates thelowerconfidenceintervalusing L1~ X X z n { X z 1 F0 : 052 u1, u2, 2 where u1~ 2 n { X z 1 u2~ 2 X ,and F0 : 052 u1, u2isthe2tailed0.05criticalvaluefora F distributionwithdegreesof freedom v2and v1;andtheupperconfidencelimitusing L2~ X z 1 F0 : 052 u0 1, u0 2n { X z X z 1 F0 : 052 u0 1, u0 2, 3 where u01~ u2z 2, u02~ u1{ 2,and F0 : 052 u0 1, u0 2isthe2-tailed 0.05criticalvaluefora F distributionwithdegreesof freedom v2and v1.Usingequations(2)and(3),thelower andupperconfidencelimitsare0.80and0.93.Inanother words,wehave95%confidenceinstatingthatthesuccess rateofthemethodforthestudyareafallsbetween80%and 93%. Thesuccessratebetween80%and93%forthestudy areasuggestedthatthismethodisreliableandpromising. Tofurtherimprovethereliabilityofthemethod,we exploredtherelationshipbetweenthefield-checkeddepressionsandtruesinkholesbyexaminedtwoprominent depressionfeatures,bermsandholes,bothofwhichwere prominentontheshaded-reliefmaps.Abermisa prominentridgealongtherimofadepression,andahole isaspotatthebottomofthedepressionthatappearsmuch deeperthanitssurroundings.Allofthefield-checked depressionsfitintooneofthreecategories:with-berm-nohole,no-berm-no-hole,andno-berm-with-hole.Thenumberoffield-checkeddepressionsineachcategoryandthe successrateforeachcategoryaresummarizedinTable1. Theno-berm-with-holecategorymadeup59%ofthe overallfield-checkeddepressions,andthiscategoryhadthe highestsuccessrate(97%)amongthethreecategories. Mostoftheno-berm-with-holefeatureswererevealedin thefieldascover-collapsesinkholes,withtheremainder beingsinkholeswithverticalrockopeningsorcoversubsidencesinkholes.Theno-berm-no-holecategorymade up28%oftheoverallfield-checkeddepressions.This categoryproved88%successful,andmostsinkholesinthis categorywerecover-subsidencesinkholes.Thewith-bermno-holecategorymadeup13%oftheoverallchecked sinkholesandhadthelowestsuccessrate(44%).Thebermlikeshapeprovedtobeaman-madestructureforawaterholdingpond.Butthroughtime,asresidualinsolublefillin rockjointsiserodedintounderlyingconduits,manyponds startedtoleakandeventuallywereunabletoholdwater, thusfunctioningassinkholes.Forthistypeofdepressionit wasdifficulttodistinguishbetweenapondthatholdswater periodicallyandapond-turned-sinkhole. Amongthefifteenprobablesinkholesthatwerenot confirmedastruesinkholesinthefield,nineofthemwere confirmedasnon-sinkholes;theywereeitherstream meandercutoffs,pondswithwaterortrash,orman-made drains.Theotherfivefeatures,shownasinconclusivein Table1.Summaryoffield-checkingofprobablesinkholes,showingthreetypesofdepressioncharacteristicsandtheir successrates. IMPROVEDKARSTSINKHOLEMAPPINGINKENTUCKYUSINGLIDARTECHNIQUES:APILOTSTUDYINFLOYDSFORKWATERSHED214 N JournalofCaveandKarstStudies, December2014

PAGE 63

Figure2,couldnotbedeterminedinthefield.They appearedasamixofnaturalkarstfeaturesdisturbedby humanactivities. Sincewefocusedondepressionfeaturesthatwere46m2andlarger,smallerpotentialsinkholeswerenotincluded. Fromtheshaded-reliefmaps,wenoticedthatinsomeareas wherelargesinkholeswerepresenttherewerealsosmaller depressionfeaturesthatappearedtobesinkholes.Figure5 showsanexampleofsuchareas.Themethodwedeveloped canbereadilytailoredtoidentifythesmallersinkholes whenresourcesbecomeavailable. TheexistingKentuckysinkholecoverage,whichwas derivedfromtheUSGStopographicmaps,had383 sinkholesforthesamearea(Fig.3).Amongthe383 sinkholes,215(56%)ofthemwerealsodetectedfromthe LiDARdata,and168(44%)ofthemweremissed.Sixteen ofthesinkholesfoundinbothdatabaseshadslightly differentlocationsbutobviouslycorrespondedtothesame features,judgingfromtheshaded-reliefmap.Avisual inspectionofthesinkholesmissedbytheLiDARanalysis usingrecentaerialimagesshowedthatapproximatelyhalf ofthemoverlappedwithman-madestructures,suchas roads,buildings,parkinglots,andquarries,andtherest werelocatedonopenfields,suchaspasture,buthadeither nooraveryshallowdepressionassociatedwiththem.We speculatethatmanyofthosesinkholesmayhavebeenfilled foragricultureorotherpurposes.Thiscomparisonshowed thatanysinkholecoverageneedstobeupdatedfrequently, becausesinkholesaretemporaryfeaturesandcanbeeasily enhanced,destroyed,oralteredbyhumanactivities.CONCLUSIONSInthisstudy,wedevelopedasinkhole-mappingmethod thatuseshigh-resolutionLiDARandaerialphotography tomapkarstsinkholesindetail.Weappliedthemethodto partsoftheFloydsForkwatershedincentralKentucky andrevealedfourtimesasmanysinkholesastheexisting databaseforthesamearea.Field-checkingsuggestedthat thesuccessrateofthismethodwasbetween80%and93% Figure5.Shaded-reliefmap(53verticallyexaggerated)showingexamplesofpotentialsmallsinkholesnotprocessedin thisstudy.J.ZHU,T.P.TAYLOR,J.C.CURRENS,ANDM.M.CRAWFORDJournalofCaveandKarstStudies, December2014 N 215

PAGE 64

forthestudyarea,indicatingthemethodisaccurateand reliable. High-densityandhigh-accuracyLiDARdataprovidea greatopportunityformappingkarstsinkholesinhigh resolutionandwithgreatdetail.Inparticular,bare-earth elevationdatainLiDARpointcloudsrevealedsinkholesin forestedareasthatwereundetectableusingonlyaerial images.Thedepression-extractionprocedurewaseffective inlocatingsurfacedepressions,butitdidnotdistinguish sinkholesfromotherdepressions,resultingintheneedfor additionalvisualscreening.Shaded-reliefmaps,especially withverticalexaggeration,revealeddepressionfeaturesin greatdetailandservedastheprimarytoolforthevisual screeningprocess.Examiningaerialimagesfromdifferent sourcesandtimeperiodswasalsocriticaltodistinguishing sinkholesfromotherdepressionfeatures.Shapeanddepth characteristicsofthedepressionswerecloselyrelatedtothe physicalfeaturestheyrepresented.Mostnon-sinkhole depressionscanbeeasilyidentified.Furthermore,fieldcheckingsuggestedthatno-berm-with-holedepressions weremostlikelytobesinkholesandthewith-berm-noholedepressionscouldbeeithersinkholesorponds.ACKNOWLEDGEMENTSThisstudywassupportedbytheKentuckyGeological Survey,andwewouldliketothanktheLouisville/Jefferson CountyInformationConsortiumandtheKentucky DivisionofGeographicInformationforprovidingthe LiDARdata.ThanksalsogotoLizAdams,CalebEssex, BaileeHodelka,ChaseLockhart,MikeLynch,Brittany Shelton,RichardSmath,andPatrickWhalen,whoassisted inthefield-checking.Wealsothankthreeanonymous reviewersfortheirconstructivecomments,whichgreatly improvedthemanuscript.REFERENCESAlexander,S.C.,Larson,E.,Bomberger,C.,Greenwaldt,B.,Alexander, E.C.,Jr.,andRahimi,M.,2013,CombiningLiDAR,aerial photography,andPictometry H toolsforkarstfeaturesdatabase management, in Land,L.,Doctor,D.H.,andStephenson,J.B.,eds., ProceedingsoftheThirteenthMultidisciplinaryConferenceon SinkholesandtheEngineeringandEnvironmentalImpactsofKarst: Carlsbad,NationalCaveandKarstResearchInstitute,Symposium2, p.441. Beck,B.F.,1984,AComputer-BasedInventoryofRecordedRecent SinkholesinFlorida:Orlando,SinkholeResearchInstitute,UniversityofCentralFlorida,Rept.No.84-85-1,12p. Brinkmann,R.,2013,FloridaSinkholes,ScienceandPolicy:Gainesville, UniversityPressofFlorida,256p. Crawford,M.M.,2012,UsingLiDARtoMapLandslidesinKentonand CampbellCounties,Kentucky:KentuckyGeologicalSurvey,ser.12, ReportofInvestigations24,12p. Currens,J.C.,2002,KentuckyisKarstCountry!WhatYouShouldKnow aboutSinkholesandSprings:KentuckyGeologicalSurvey,InformationCircular4,SeriesXII,35p. Dinger,J.S.,Zourarakis,D.P.,andCurrens,J.C.,2007,Spectral enhancementandautomatedextractionofpotentialsinkholefeatures fromNAIPimageryinitialinvestigations:JournalofEnvironmental Informatics,v.10,no.1,p.22.doi:10.3808/jei.200700096. ESRI,2012,ArcGISDesktop:Release10.1SP1:Redlands,California, EnvironmentalSystemsResearchInstitute. Evans,J.S.,andHudak,A.T.,2007,Amultiscalecurvaturealgorithmfor classifyingdiscretereturnLiDARinforestedenvironments:IEEE TransactionsOnGeoscienceandRemoteSensing,v.45,no.4, p.1029.doi:10.1109/TGRS.2006.890412. Florea,L.J.,Paylor,R.L.,Simpson, L.,andGulley,J.,2002,KarstGIS advancesinKentucky:JournalofCav eandKarstStudies,v.64,p.582. Floyd,C.T.,Syverson,K.M.,andHupy,C.M.,2011,UsingLiDARdata andArcGIStoevaluatesubtleglaciallandformsassociatedwiththe EarlyChippewaandEmeraldPhaseIce-Marginpositions,Barron County,Wisconsin[abs.]:InstituteonLakeSuperiorGeology Proceedings,57thAnnualMeeting,Ashland,WI,v.57,p.35. Fleury,S.,2009,LandUsePolicyandPracticeonKarstTerrains:Living onLimestone:NewYork,Springer,187p.doi:10.1007/978-1-40209670-9, Ford,D.C.,Williams,P.,1989,KarstGeomorphologyandHydrology: London,Unwin-Hyman,601p. Littlefield,J.R.,Culbreth,M.A.,Upchurch,S.B.,andStewart,M.T., 1984,Relationshipofmodernsinkholedevelopmenttolarge-scale photolinearfeatures, in Beck,B.F.,ed.,Sinkholes:TheirGeology, Engineering&EnvironmentalImpact:Rotterdam,A.A.Balkema, p.189. Miao,Xin,Qiu,Xiaomin,Wu,Shuo-Sheng,Luo,Jun,Gouzie,D.R.,and Xie,Hongjie,2013,Developingefficientproceduresforautomated sinkholeextractionfromlidarDEMs:PhotogrammetricEngineering& RemoteSensing,v.79,no.6,p.5454.doi:10.14358/PERS.79.6.545. Mukherjee,A.,andZachos,L.G.,2012,GISanalysisofsinkhole distributioninNixa,Missouri[abs.]:GeologicalSocietyofAmerica AbstractswithPrograms,v.44,no.7,549p. NationalDroughtMitigationCenter,2013,AnnualClimatology:Louisville, KY(SDF),http://drought.unl.edu/archive/climographs/LouisvilleANC. htm,accessedOctober24,2013. Newell,W.L.,2001,Physiography, in McDowell,R.C.,ed.,TheGeology ofKentuckyATexttoAccompanytheGeologicMapof Kentucky:U.S.GeologicalSurveyProfessionalPaper1151-H, p.79. Newton,J.G.,1976,EarlyDetectionandCorrectionofSinkholeProblems inAlabama,withaPreliminaryEvaluationofRemoteSensing Applications:AlabamaHighwayDepartment,BureauofResearch andDevelopment,ResearchReportno.HPR-76,83p. Paylor,R.L.,Florea,L.J.,Caudill,M.J.,andCurrens,J.C.,2003,AGIS CoverageofSinkholesintheKarstAreasofKentucky:Kentucky GeologicalSurvey,metadatafileandshapefilesofhighestelevation closedcontours,1CDROM,(http://kgs.uky.edu/kgsweb/download/ karst/ksinks.zip). Rahimi,M.,Alexander,S.C.,andAlexander,E.C.,Jr.,2010,LiDAR mappingofsinkholes:WinonaCounty,MN[abs.],GeologicalSociety ofAmericaAbstractswithPrograms,JointMeetingNorth-Central/ South-CentralSections,v.42,no.2,p.107. Rahimi,M.,andAlexander,E.C.,Jr.,2013,LocatingsinkholesinLiDAR coverageofaglacio-fluvialkarst,WinonaCounty,MN, in Land,L., Doctor,D.H.,andStephenson,J.B.,eds.,Proceedingsofthe ThirteenthMultidisciplinaryConferenceonSinkholesandthe EngineeringandEnvironmentalImpactsofKarst:Carlsbad,National CaveandKarstResearchInstitute,Symposium2,p.469. Seale,L.D.,2005,Creation,analysis,andevaluationofremotesensing sinkholedatabasesforPinellasCounty,Florida[M.S.Thesis]:Tampa, UniversityofSouthFlorida,55p. Seale,L.D.,Florea,L.J.,Brinkmann,R.,andVacher,H.L.,2008,Using ALSMtoidentifycloseddepressionsintheurbanized,coveredkarst ofPinellasCounty,Florida1,methodologicalconsiderations: EnvironmentalGeology,v.54,p.995.doi:10.1007/s00254-0070890-8. Tihansky,A.B.,1999,Sinkholes,west-centralFlorida, in Galloway,D., Jones,D.R.,andIngebritsen,S.E.,eds.,LandSubsidenceinthe UnitedStates:U.S.GeologicalSurveyCircular1182,p.121. Woods,A.J.,Omernik,J.M.,Martin,W.H.,Pond,G.J.,Andrews,W.M., Call,S.M.,Comstock,J.A.,andTaylor,D.D.,2002,Ecoregionsof Kentucky(colorposterwithmap,descriptivetext,summarytables, andphotographs):Reston,VA.,U.S.GeologicalSurvey(mapscale 1:1,000,000). Zar,J.H.,1999,BiostatisticalAnalysis,4thed.:UpperSaddleRiver,NJ: PrenticeHall,929p.IMPROVEDKARSTSINKHOLEMAPPINGINKENTUCKYUSINGLIDARTECHNIQUES:APILOTSTUDYINFLOYDSFORKWATERSHED216 N JournalofCaveandKarstStudies, December2014

PAGE 65

SINKHOLESANDADISAPPEARINGLAKE: VICTORYLAKECASESTUDYDR.TAMIEJ.JOVANELLYAssociateProfessorofGeology,BerryCollege,P.O.Box495036,MountBerry,GA30149,tjovanelly@berry.eduAbstract: Human-inducedsinkholecollapsecanresultindrasticchangestolandscape aestheticsandpresentchallengestolandmanagersseekingtodeterminetheplausibility ofrestoration,theamountoffinancialinvestmentneeded,andthelong-term sustainabilityoftamperingwithkarsticenvironments.Alterationofgroundwaterflow inakarsticenvironmentexpeditedtheformationoflargesinkholesinthesouthernend ofman-madeVictoryLake,causingittodrainimmediately.Soonafterthelakeemptied in1986,twounsuccessfulattemptsweremadetorestorethe13ha(32ac)lake.The sinkholesformedinthesouthernbasinwerecompletelyin-filled,eliminating3ha(8ac) oflakebasinandsignificantlyalteringtheoriginallakemorphology.Sometwenty-seven yearslater,VictoryLakeisholdingsomewaterintheshallowbasinatthenorthernend andwouldprimarilybeclassifiedasamarshywetland.Thisstudywasinitiatedto investigatethecurrentrelationshipbetweenthegroundwaterandsurfacewateratthe lakesalteredbasintodeterminethepotentialforittoberestoredfullyorpartiallyasa recreationalfocalpointoftheBerryCollegecampus.Overthecourseofoneyearwe measuredtheinputs(streamflowandprecipitation)andoutputs(evaporationand surfacewateroutflow)ofthelakesystem.Wewereabletoconcludethatgroundwateris notlikelycontributingtothelake,basedoninorganicandstableisotope(18Oand2H) waterchemistryanalysisandthedeeppositionofthegroundwatertablerelativetothe lakebottom.Fromtheresultsofdye-tracertestsconductedinthelake,weconcluded thatbasinwatermaynotbeescapingdownwardatmeasurableratesbecauseofitsclay bottom.Ouroverallwater-budgetanalysisconfirmsanadequatewatervolumeentering byrainfallandephemeralstreaminflow;nearly90%ofthewaterleavesVictoryLake throughsurface-wateroutflow.Waterlossthroughevapotranspirationduringspringand summermonthsovercomesthegainaccomplishedduringwetterandcoolermonths, particularlyFebruary.Throughaninvestigationofwaterlevelrecordskeptforcampus monitoringwellsfrom1998through2012weconfirmthatthegroundwatertablehas stabilizedandtheimmediatethreatofnewsinkholeformationisminimal.Restorationof VictoryLaketoitsoriginalpicturesquemeetingspotmaybepossiblethroughcreative engineeringstrategiesandprojectfinancing.However,wequestionthelongevityof managingakarsticenvironmentandconsiderthepotentialriskstoinfrastructure, groundwater,andhumanhealthshouldlakebottomfailureoccuragainoncampus.INTRODUCTIONDuringthelate1980sgeologistsbegantoreportan increasingpaceofhuman-inducedsinkholesintheeastern UnitedStates,withparticularfocusonthechanging landscapesinGeorgia,Alabama,andFlorida(MacIntyre, 1986;Newton,1987).Morerecently,stressonandoveruse oftheFloridaaquiferhasledtoseveraloccurrencesof emptiedlakes,includingLakeJackson(1619ha;4,000ac) andLakeScott(115ha;285ac)(Penson,2002;McBrideet al.,2011).Human-inducedsinkholesoftenresultfrom dewateringbywells,quarries,andminesinlimestone environments(Newton,1987;Fidelibusetal.,2011).The processesformingsinkholescanbeenhancedbyhumaninducedchangeinthegroundwaterhydrologicregimeby eitherinflowsoroutflowsresultingfrompumpingactivities (Benitoetal.,1995;Martinezetal.,1998;Floreaetal., 2009).Surfacesubsidencecandevelopwithinamatterof dayswhenhighlysolublerocksdissolveduetoanthropogenicpressures(Martinezetal.,1998).Thelocationof sinkholecollapseandspeedofformation,ratherthanthe diameterordepth,dictatethethreattohumanlifeandthe potentialforeconomicloss(Newton,1987).Todate,the FederalEmergencyManagementAgencyreportsthat insuranceclaimsrelatingtosinkholes,eithernaturalor human-induced,intheUnitedStatestotals$100million annually.Sinkholeformationiscloselyrelatedtolocal hydrologicconditions,andhuman-inducedchangestothe localhydrologycommonlyacceleratetheprocess.Ithas beenshownthatareasintheeasternUnitedStatesthat haveahighersinkholedensitytendtoshowalowerwater qualityduetothedirectpathwayofcontaminationfrom surfacetothegroundwatertable(Lindseyetal.,2010).An understandingofgroundwaterandsurfacewaterinterac-T.J.JovanellySinkholesandadisappearinglake:VictoryLakecasestudy. JournalofCaveandKarstStudies, v.76,no.3,p.217. DOI:10.4311/2012ES0272JournalofCaveandKarstStudies, December2014 N 217

PAGE 66

tionsinkarsticenvironmentsisessentialforaquifer protectionandthedevelopmentoflocalwaterresources (PraiseandGunn,2007;McBrideetal.,2011). Previousstudieshaveindicatedtheusefulnessof creatingwaterbudgetsforlakesinkarsticterrains(Dalton etal.,2004;McBrideetal.,2011;Auritetal.,2013).More specifically,thegreatseasonalvariabilityoflakeswater budgetstiedtotheFloridaaquiferandtherepercussionsto surroundingcommunitieshasbeenthefocusofrecent UnitedStatesGeologicalSurveyinvestigations(Daltonet al.,2004;Spechler,2010;McBrideetal.,2011;Sepulvedaet al.,2012).TheresearchconductedatLakeSeminole (Daltonetal.,2004)providedtheframeworkforthewater budgetcompletedinthisstudy.Daltonliststhechallenges tomeasuringevaporationorevapotranspirationovera hydrologicyearandemphasizestheimportanceofitas adriverinseasonallakevolume.Humanimpacton groundwaterresourcesmayhavebeenbesthighlightedby Spechler(2010)inhisstudyoftheFloridaaquiferinsouthcentralFlorida.Spechlerwasabletoillustratethestresses ofpopulationgrowthonthegroundwaterlevel.In addition,Spechlersresultsshowthehighvariabilityover timeinaquiferwaterqualityfromalargesamplingof129 areawells.ThestudybyMcBrideetal.reportsafour-year waterbudgetofLakePanasoffkeeinwest-centralFlorida andfocusesonthegroundwaterandsurfacewater connectednessofakarsticsystem.Intheirstudy,environmentalisotopesofstrontium,oxygen,andhydrogen confirminterplaybetweengroundwaterandsurface-water systems.Inaddition,theywereabletodiscernthroughthe stableisotopeanalysisthatrainfallwasaprimarysourceof groundwaterrechargewithintheLakePanasoffkeewatershed.TherecentUSGSpublicationbySepulvedaetal.uses measuredwaterbudgetparametersofrunoff,infiltration, lakewaterlevels,streamflows,andevapotranspiration measuredfrom19956ineast-centralFloridato computetheinteractionofgroundwaterflowsystemwith thesurfaceenvironmentusingMODFLOW-2005.This modelallowedthemtomakesomepredictionsaboutlongtermgroundwaterrechargeandwithdrawalrates. Awaterbudgetsystematicallyquantifiesthegain,loss, andstorageofwaterinthewatercycleusingtheprinciple oftheconservationofmass.Typicalfieldmeasurements forawaterbudgetincludeprecipitation,groundwater inflow,surfacewaterinflow,evaporation,transpiration, groundwateroutflow,andsurfacewateroutflow.For water-supplyplanningandmanagement,waterbudgetsfor aquifersandwatershedsareanimportanttoolusedto determinefluctuationsandstressonthesystem(Winstanleyetal.,2006;Healyetal.,2007).Asallcomponentsof thewatercycleareconnected,estimatingfuturewater budgetsallowsplannersandmanagerstoevaluatewater availabilityandtheimpactsofwithdrawalsonthesystem. Datacollectedforwaterbudgetsareoftenusedfor understandingcurrenthydrologicalconditionssothat futureoutcomescanbeforecast. Awaterbudgetcanbeusedtoexploretheimpactona singleparameterortheentiresystemtoavariablesuchas temperature.Manyresearchersoptfordefiningawater budgetbasedontheresponseofoneinfloworoutflow parameter(LoagueandFreeze,1985;Winter,1985; Deevey,1988;HebbertandSmith,1990).Theisolationof asingleparameter,suchasgroundwater,maymakeit easiertoidentifytheresponseofassociatedstreams,lakes, orwetlands(Winter,1999).Forexample,Winstanleyand Wendland(2007)usedawaterbudgettoinvestigate responsetoclimatechangeovertime,andtheyshowed theinfluenceoftemperatureonwateravailability. Waterbudgetsarenotlimitedbygeographicalscale. Somerangefromasmalldrainagebasin,asinthisstudy (11km2),toconsiderablylargerareaslikethe618km2exploredbyShusteretal.(2003).However,aswasthe focusofapaperbyHamilton-Smith(2006)andreiterated byHorvatandRubinic(2006),thestartofaqualitywater budgetbeginswiththeaccuratedelineationofthetotal catchmentarea. Despitevariationsinprojectsize,scope,andregionof study,twoquestionspersistinwaterbudgetanalysis:What isthebestwaytomeasureevaporation,andwhatisthe errorinsuchestimates?Winter(1981)foundthatannual averageshadsmallererrors(2to15%)thanmonthly averages(2to30%).Itappearsthatthepreferenceinhow evaporationismeasureddependsontheamountofdata availableortheamountoftimeandmoneyastudyhasto commit.UsingClassApanevaporationmeasurementsand GeorgiaAutomatedEnvironmentalMonitoringNetwork (GAMEN)dataascontrols,Daltonetal.(2004)evaluated sixdifferentmethodsofcalculatinglakeevaporationrates. Thecomplexityandtheamountofdataneeded,suchas windspeed,humidity,andsolarradiationforthesemethods variesgreatly.Daltonetal.(2004)andRosenberryetal. (1993)determinedthattheenergy-budgetmethodwas8to 26%morereliablethanempiricallyderivedequationsand ultimatelyprovidedthebestmatchifrawdataare unavailable.SITEDESCRIPTIONVictoryLake(34 u 17 9 54 0 N,85 u 12 9 07 0 W)isarecreationalman-madelakethatwascompletedin1928(Fig.1).This picturesquespotontheBerryCollegecampusprovideda placeforthecollegecommunitytowalk,picnic,andcanoe. IthasbeenimportanttoBerryCollegeanditsalumnifor decades.In1953,BerryCollegesoldmineralrightstoa limestonequarryoperatorapproximately0.8kmwestof VictoryLake.Thequarryhadsuccessfulminingoperations forthenextthirty-threeyears(Fig.2a).ByJune1986,the open-pitquarrywas110-mdeepanddailypumpingrates topped52,616m3s2 1(RichardFountain,personal communication).Atthistime,therecordeddepthtothe watertableontheBerryCollegecampuswas44m. Overnight,fourlargesinkholesmorethan6mindiameterSINKHOLESANDADISAPPEARINGLAKE:VICTORYLAKECASESTUDY218 N JournalofCaveandKarstStudies, December2014

PAGE 67

formedinthesouthernbasinofVictoryLakethatcaused thelaketoempty(Fig.2b).Presumably,aconeof depressionformedfromtheintensepumpingatthequarry thatcausedthebottomofthatpartofthelake,whichwas overlimestone,tocollapse.Thesinkholeswerecompletely filledwithearthmaterialsafterafewfailedattemptsatlake restoration(Fig.2c).Thelakebasinwasnowreducedby 3ha(8ac),andwaterthatoriginallydrainedthrougha tributarynowleftbywayofaculvert.Sometwenty-seven yearslater,siltation,plantencroachment,andbeaverdams haveturnedVictoryLakeintoashallowbasinfullof vegetation.Currently,thebasinisonlyknowntohold waterforashortdurationfollowingahardrainfall (Fig.2d). Thepurposeofthisstudyistodeterminethepossib ilityof restorationofalakeinfluencedbykarsttopography.Todo this,weconductedafourteen-monthinvestigationtomeasure theinflowandoutflowcomponentsofthepresent-day VictoryLakeswaterbudget.Thegoalsofthewaterbudget weretoidentifyhowlakevolumerespondstocurrentpatterns inseasonalchangesinprecipitationandtemperature,to determineifgroundwaterissubstantiallyaddingtoorleaving thesystem,todetermineifcurrentbasinmorphologyis influencingtheamountofwaterthatthebasinholds,and, ultimately,todeterminethepotentialoflakerestoration.STUDYAREAANDGEOLOGICSETTINGKarsticfeaturesresultingfromlimestonedissolutionare commonplaceinthenorthwestGeorgiaphysiographic regioncalledtheRidgeandValley(Hubbard,1988).The nearlytwohundredcavesmappedintheCumberland PlateaunorthofMountBerryareunlikeRidgeandValley undergroundcavernsbecausetheyaretypicallyconnected anddonotremainsolitaryundergroundvoids(Jenkins, 2009;Buhlmann,2001).MostofthevoidsintheRidgeand Valleyregionareformedbysolutionprocessesalong fractures,joints,andbeddingplanes(Weary,2005). Twotypesofsinkholesoccurmostcommonlyinthe RidgeandValleyregion:cover-collapseandcover-subsidence(Hubbard,1988).Cover-collapsesinkholesform whenthesurficialsedimentscontainalargeamountof clay.Theclaybindsthesoilsothatitcanbridgesmall cavities,butnotlargeones.Cover-subsidencesinkholes formwhensurficialsedimentsfilterintocavitiestogradually formsurfacedepressions.(Floreaetal.,2009).Although cover-subsidencesinkholesareknowntobemoredestructive,bothsinkholevarietiescanposerisktohumanhealth andeconomicrisktourbanplanners,developers,homeowners,andinsurancecompanies(Scheidtetal.,2005). Thesetypesofnaturalsinkholedevelopmentaregenerally notpredictable,althoughsinkholescanbeexpectedwhere limestoneformationsarefound(Newton,1987). Thegeologyexposedatthesurfacearoundthe perimeterofVictoryLakevaries.Thebedrockfound underthewesternsideofVictoryLakeispredominately Mississippian-agelimestone.Thislimestoneispartofthe ConasaugaFormation.TheConasaugaFormationconsistsofsiltstone,claystone,shale,andlimestoneandwas describedbyAnderson(1993).Theformationiseasily identifiablewithinthenumeroussinkholesfoundinthe forestedarealiningthewesternsideofthelake.The contactbetweentheshaleandthelimestoneunitsofthe ConasaugaFormationcanbeseenatthequarry,however, thelowercontactoftheConasaugaFormationonBerry Collegesmaincampusisnotexposed.Fromourreviewof thewelllogsfromthetwenty-twomonitoringwellson campusthatweredrilledin1998,theboundaryisnotclear. Themaximumdepthdrilledinthesemonitoringwellsis 93mbelowthesurface.Theeasternsideofthelakehas exposuresofCambrianagesandstone.Thissandstoneis partoftheRomeFormation,whichconsistsofsandstone, siltstone,andclaystone.Thisformationisknownfor tightlyfoldedandsteeplytiltedbeds(Anderson,1993). ThegeologicmapoftheRome,Georgia,areawas completedbyTomCrawfordin1990,butnotpublished. SomerevisionsandadditionsweremadebyC.Williamsin 1993.CrawfordandWilliamsidentifiedanormalfault, referredtoastheRomeFault,thatcrosscutsthesouthern portionofVictoryLake.Totheeastofthisfault,andto thesouthofthequarry,CrawfordandWilliamsalso identifiedseveralverticaljointsetsexposedinoutcrops.HYDROLOGICALSETTINGVictoryLakeislocatedinatopographicallylowarea comparedtoregionsimmediatelyadjacenttoit.Therimof thequarryselevationis189masl,whereasVictoryLakes Figure1.LocationmapofVictoryLakenearMountBerry, Georgia,showingitsproximitytothequarry,surface-water samplinglocations,monitoringwells,andotherdatasources. GeologicunitsandfaultpositionestimatedfromCrawford (1990)[fromTomCrawfordsunpublished1990geologic mapoftheBerryCollegecampus.].T.J.JOVANELLYJournalofCaveandKarstStudies, December2014 N 219

PAGE 68

Figure2.AerialphotographyofVictoryLake.TheXinphotosA,B,Cindicatesthesoutheastern-mosttipofVictoryLake.The photographshowninDwastakenfromthesoutheastern-mostposition(X)lookingnorth.A.Thisaerialphotographshowsearly stagesofquarryexcavationin1964andafullVictoryLake.B.Thisaerialphotograph(1987)showsanemptyVictoryLake.Notice thatvegetationhasnotyetenteredthebasin.C.This2012aerialphotograph(GoogleEarth)illustratescurrentconditionsatVictory Lake.D.AphotographofVictoryLakeafterMarch(2008)rains.SINKHOLESANDADISAPPEARINGLAKE:VICTORYLAKECASESTUDY220 N JournalofCaveandKarstStudies, December2014

PAGE 69

elevationisapproximately181masl.Undernormal conditions,thegroundwaterflowpatternintheareais fromwesttoeastandfollowstheregionaltopographyof thelandscape. Anephemeralstreamwithdrainagebasinapproximately 5.0km2entersVictoryLakeonitsnorthernend.Pre-1986, thiscreekcontinuedfromthesouthernendofVictoryLake tomergewithBigDryCreek.Today,aculvertfunnelswater fromVictoryLakeintoBigDryCreek. Twenty-twomonitoringwells(PZ26AandPZ16A)were installedonthemaincampusinJune1998followingthe collapseofseveralcampusbuildingsintosinkholes.The locationsoftwowellsusedforgroundwatersamplingin thisstudyareshownonFigure1.Thewaterleveldata recordedfrom19980providesvaluableinsightasto thesignificantdepletionoftheaquiferduringandafterthe timeofpumping;thequarryofficiallyclosedin2000.The lowestdepth-to-watermeasurement(47m)wasrecordedin thewellatthesouthernendofVictoryLakeinJuly1998. Thewatertableimmediatelyreboundedafterquarry operationswerehalted.Thecurrentaveragedepthtothe watertableis10m.CLIMATEThemeantemperatureforMountBerryoverathirtyyearperiodis16.5 u C.Augustistypicallythewarmest month,averaging32.2 u C;Januarythecoldest,averaging 11.8 u C.Averageannualrainfallis1425mm;generally,the highestmonthlyrainfalloccursduringMarchandthelowest inOctober.Thehydrologicandclimaticdatacollectedfor thispaperrepresentsonehydrologicyearstartingOctober1 andendingSeptember30;seeFigures3and4.MATERIALSANDMETHODSESTABLISHINGLAKEWATERBUDGETThewaterbudgetequationusedinthisstudyis( P + I ) 2 ( ET + O ) 5 DS ,where P isprecipitation, I isinflow, ET isevapotranspiration, O isoutflow,and DS ischangein storage.Allwaterbudgetmeasurementsareconvertedto m3.Annualandseasonalwaterbalanceswerecomputed basedonamethodfromMcCarthyetal.(1991)and Chescheiretal.(1995).Theexpressionforpercentage closureerroris( D Scalc2 DSmeas)/F3100%,where F isthe systemflux(m3)expressedas F 5 ( P + I + O + ET + | DScalc|)/2and D Scalcistheresidualstorageand DSmeasis themeasuredstorage. Inflowandoutflowstreamvelocityweremeasured weeklyusingaPriceAAflowmeter.Theinflowvelocity wasmeasuredattheconfluenceofatributaryandVictory Lake,andtheoutflowwasmeasuredattheculvertwhere thelakewaterhasbeendivertedtopreventroadflooding. Stagegaugeswereplacedintheselocationstomeasure changesinwaterdepthovertime. Figure3.Averagemonthlytotal(19102009)precipitation(mm)measuredforRome,Georgia,(blacksquares)comparedto thetotalmonthlyprecipitation(mm)measuredatVictoryLakeforOctober2007toSeptember2008(greydiamond).Average monthlytotaldatawereretrievedfromtheGeorgiaAutomatedEnvironmentalMonitoringNetworkwebsitewww. georgiaweather.net.T.J.JOVANELLYJournalofCaveandKarstStudies, December2014 N 221

PAGE 70

TomeasureprecipitationontoVictoryLake,postmountedbutyrate-plasticraingaugesmadetoU.S. WeatherServicespecificationswereplacedinanopenarea atthenorthernandsouthernendsofthebasin.Thesedata werecollecteddaily. AClassAAevaporationpanwasplacedadjacentto VictoryLakeinanopenfieldfreefromshadeandwind obstruction.Theevaporationpanwasvisiteddailyfrom October2007toSeptember2008.Theevaporationpanwas storedforthewintermonthsofDecembertoFebruary. DataforevapotranspirationcamefromtheGeorgia EnvironmentalMonitoringNetworkwebsiteforthecity ofRome. Inorganicwaterchemistrydatawasanalyzedfromboth thegroundwaterandsurfacewater.Groundwatersamples werecollectedattwomonitoringwellsincloseproximityto VictoryLake(Fig.1).Eachwellwaspurgedforthree hoursbeforesamplecollection.Sampleswereimmediately analyzedforgeneralwaterchemistryatBerryCollege. Thosesamplescollectedforenvironmentalisotopes(18O and2H)wereimmediatelyplacedondryiceandmailed overnighttotheUniversityofWaterlooinCanadafor analysis.Thestableisotopecompositionofwateris reportedwithreferencetotheStandardMeanOcean Water,inpartsperthousand(Craig,1961). ARhodamineWT(C.I.AcidRed388)dyetracertest wasconductedinMarch2008toestablishgroundwater outfloworlackthereof.Standardsandacalibrationcurve werecreatedusingconcentratedstocksolutionofRhodamineWTfortheModel10TurnerDesignsFluorometer. ThemeasuredinjectionofRhodamineWTwasbasedon thetotallitersofwatercalculatedtobeinVictoryLake duringMarch2008.Sixlitersofdye(1.428kg)was introducedbysluginjectionatthenorthernentranceof BigDryCreek.Groundwatersampleswereretrievedat monitoringwellPZ26A(Fig.1)hourlyfor24consecutive hours,thenonceevery4hoursfortwodays,theneveryday forthreeweeks.RESULTSGROUNDWATERTable1comparesinorganicwaterchemistryforthe groundwaterandsurfacewatersamples.Thegroundwater hashigheralkalinity(222to246mgL2 1)whencompared tothesurfacewatersamples(110to132mgL2 1).In addition,theCO2isconsiderablyhigherinthegroundwater(18mgL2 1)thaninthesurfacewatersamples(0to 5mgL2 1).Thereislittlevariationintheamountof dissolvedoxygenandnitratesamongthesamples.The warmersurfacewaterhasalowerpHthanthatofthe groundwater. Groundwaterandsurfacewatersampleswereanalyzed for18Oand2H(Fig.5).Allofthesurfacewatersamples containlargerfractionsof18Oand2Hthanthegroundwatersamples.Thegroundwatersampleshavealighter signatureforboth2H(rangingfrom 2 3.18to 2 3.94 % ) and18O(rangingfrom 2 18.54to 2 22.48 % ). Thewellsmonitored(PZ26AandPZ16A)duringthe rhodamine-WTdye-tracertestarebothlocatedatthe southernendofthelake(Fig.1).Ourfluorometeranalysis didnotdetectanyamountsofRhodamineWTdyeinthe Figure4.Averageminimum/maximummonthlytemperature(6C)measuredforRome,GA.Averagedatawasretrievedfrom theGeorgiaAutomatedEnvironmentalMonitoringNetworkwebsitewww.georgiaweather.net.SINKHOLESANDADISAPPEARINGLAKE:VICTORYLAKECASESTUDY222 N JournalofCaveandKarstStudies, December2014

PAGE 71

groundwatersamplescollectedthroughoutthethree-week samplingduration.PRECIPITATIONThesamplingmonthsofOctober2007through September2008wereclassifiedasmoderatedroughtto near-normalconditions(NationalOceanicandAtmosphericAdministrationwebsite,www.cpc.ncep.noaa.gov). FromJanuary2008toSeptember2008thetotalrainfall recordedatVictoryLakewas915.37mm.Thiswas 175.54mmlowerthantheyearlyaverageforRome, Georgia(Fig.3).Theaveragemeanmonthlytemperatures recordedduringthedurationofthestudycorrelateswithin 2degreestothelong-termmonthlyaveragetemperatures forallmonths,exceptOctober2007andDecember2007, whichwerewarmerthanhistoricaverage(Fig.4).EVAPORATIONBecauseoftheextensivein-fillingoftheremainingbasin withsedimentovertwenty-sevenyearsandtheresulting overgrowthofplants,wedeemevapotranspirationat VictoryLaketobemoreinfluentialthanevaporation. Forthisreason,thedatawepresentistheevapotranspirationreportedforthecityofRomefoundfromthe GeorgiaEnvironmentalMonitoringNetwork.INFLOWANDOUTFLOWBecauseBigDryCreekisephemeral,itonlyflows duringperiodsofheavyandintenserainfall.Figure6 comparesstreaminflowtodirectrainfallontothesurface ofthelakefromOctober2007toNovember2008.During periodsofhighrainfall,suchasthemonthsofDecember 2007throughMarch2008,streaminflowisthedominant Table1.Surfacewaterandgroundwaterchemicalanalyses. Test SurfaceWaterSamples GroundwaterSamples 1-VL 2-VL 3-VL PZ26APZ16A Alkalinity,mgL2 1132 112 110 246 222 CO2,mgL2 10 5 0 20 18 Nitrate-Nitrogen,mgL2 1, 4.4 0 0.2 2 1 Phosphate,mgL2 100021 Temperature, u C 27 26.5 27 17 16.8 pH 7 6.9 6.8 7.7 7.7 Figure5.Groundwater(samples16Aand26A)andsurfacewater(samplesVL1andVL2)analysesforstableisotopes d18O and d2H.T.J.JOVANELLYJournalofCaveandKarstStudies, December2014 N 223

PAGE 72

processaddingtothewatervolumeinthebasin.Rainfall ontoitssurfacewillbethemajorcontributortoVictory Lakeduringtherestoftheyear.Similarlytoinflow,water flow,inthiscasetodownstreamBigDryCreek,isthe dominantprocessofwaterlosswhenthereareperiodsof highandintenseprecipitation(Fig.7).WATERBALANCEOvertheone-yearstudy,VictoryLakesoverallwater budgetindicatesthatthebasinlostmorewatervia evapotranspirationandsurfacewateroutflow(50.3%) thanitgainedviarainfallandstreaminflow(49.7%) (Fig.8).Whichisthedominantvariable,streaminflow, Figure6.ThepercentagescontributedtothewatergaininVictoryLakebyinflowfromBigDryCreekandrainfalldirectly ontothesurfacefromOctober2007toNovember2008. Figure7.ThepercentagescontributedtothewaterlossfromVictoryLakebyflowthroughtheexitculvertand evapotranspirationfromOctober2007toNovember2008.SINKHOLESANDADISAPPEARINGLAKE:VICTORYLAKECASESTUDY224 N JournalofCaveandKarstStudies, December2014

PAGE 73

rainfall,evapotransiration,orstreamoutflow,varies seasonally.Duringthesummerandfallmonths,thelakes outflowwasdrivenbyevapotranspiration,67%and54%, respectively.Duringthewinter,waterleftthebasinby streamoutflow(48%).Waterleftthebasinduringthe springviaoutflow(41%)andevapotranspiration(37%). Gainsorlossesinstorageweredeterminedby comparingthemonthlyfluctuationstotheinitialvolume andsurfaceareameasuredforthelakeatthestartofthe study;theoverallchangefortheyearbeingnegative (Table2).Monthlyincreasesinstoragerangingfrom5%to 23%occurredfromOctober2007toJanuary2008 (Table2).FromOctober2007toJanuary2008thelake volumedoubles.ByFebruary2008,however,thereisa 62%decreaseinstorage.Thisreductionisfollowedby increasesinrainfalladditionsandsurfacewaterinflow fromspringrainevents.ByMarch2008,thelakevolume increasedto10timesthatoftheinitiallakevolume determinedinSeptember2007.TheMarchrainsprovided enoughprecipitationtocompletelyfillinthelakebasin. AftertheMarchstorms,thestorageinVictoryLakebegan todecreasedespiteMayprecipitationdoublingthat recordedinApril.Lossescontinuedtobecalculated generallyfromMaytoNovember2008,withaslight positivereboundinSeptember. Acomparisonofmonthlyinflowtypes(Fig.6)reveals thatdirectrainfalldominatedthegainduringmostofthe study;innineoffourteenmonthsrainfallontothelake contributed100%totheinflowwaterbudget(Fig.6). Duringthewinterandearlyspring,however,thestream providedmorethan80%ofthemonthlyinflowwater budget(Fig.6).Likewise,outflowthroughtheculvertis thedominantprocessmorethan90%ofthetime(Fig.7). Evapotranspirationisprominentatthestartofspring (March)andagaininJuneandJuly,withover95%of waterbeinglostthroughplantgrowththen(Table2).DISCUSSIONTheresultsfromthechemicalanalysesindicatethat thegroundwatershowsmuchhigherevidenceofalkalinity (222and246mgL2 1comparedto110to132mgL2 1) (Table1).Thisisatypicalcharacteristicofwaterpumped fromalimestoneaquifer(McBrideetal.,2011).In addition,CO2isalsoconsiderablyhigherinthegroundwater,18and22mgL2 1comparedtothesurfacewaters0 to5mgL2 1.ThispatternlikelyemergesbecauseCO2is moresolubleincoldwater,andthemonitoringwellisa closedsystem.Thevariationintheamountofdissolved oxygenandnitratesbetweensurfaceandgroundwateris slight.Asanticipated,thesurfacewateriswarmerandhas alowerpHthanthatofthegroundwater. TheCraigandGordon(1965)modelestablishedthat environmentalisotopes18Oand2Hrespondtochangesin temperature,thereforeanalyzing18Oand2Hisparticularly usefulinthestudyofgroundwaterandsurfacewater interactions.Severalauthorshavesuccessfullyapplied isotopicfractionationof18Oand2Hasameansto distinguishbetweengroundwaterandsurfacewaterin waterbudgets(Krabbenhoftetal.,1990a,1990b,1994; Yehdeghoaetal.,1997).Morerecently,McBrideetal. (2011)usedstableisotopesforsimilarpurposestoestablish inflowandoutflowinawaterbudget.Thus,ifthe groundwaterandsurfacewaterareconnected,the18O Figure8.Totalmonthlyprecipitationrecordedduringthestudy(mm)comparedtototalmonthlyVictoryLakeoutflow throughtheexitculvert(m3).T.J.JOVANELLYJournalofCaveandKarstStudies, December2014 N 225

PAGE 74

Table2.MonthlywaterbudgetandstorageforVictoryLake.Allvaluesincubicmeters.Lakevolumeisfiguredbasedonaninitialestimatevolumefor September2007. Month&Year GainstoVictoryLake LossestoVictoryLake Calculated Changein Volume,m3Lake Volume,m3Flowfrom BigDry Creek,m3Rainfall ontothe Lake,m3Total Gain,m3Flowout theCulvert, m3Evapotranspiration (GAEMN data),m3Total Loss,m3October2007 010,668.4710,668.47162.595,281.905,444.515,222.9624,637.96 November2007 04,505.394,505.39609.392,693.043,302.341,203.0525,841.01 December2007 19,407.054,136.5923,543.6514,912.27789.6015,701.877,841.7833,682.79 Janunary2008 73,084.697,486.1880,570.8777,169.211,812.7278,981.931,588.9435,271.73 February2008214,327.298,448.09222,775.38233,551.602,748.48236,300.08 2 13,524.7021,747.03 March2008 2,919,790.87,497.582,927,288.382,831,445.225,310.482,836,755.7090,532.68112,279.71 April2008 04,752.524,752.528,886.567,943.0416,829.60 2 12,077.07100,202.64 May2008 010,984.0410,984.04100,612.9610,232.88110,845.84 2 99,861.81340.83 June2008 01,923.821,923.82488.8412,902.4013,391.24 2 11,467.42 0 July2008 06,204.896,204.89 012,099.3612,099.36 2 5,894.46 0 August2008 12,439.6811,618.9724,058.6522,003.1810,341.2432,344.42 2 8,285.77 0 September2008 0 216.71 216.71 07,659.127,659.12 2 7,442.40889.17 October2008 05,877.925,877.92 04,988.764,988.76889.17135.83 November2008 01,475.181,475.18 02,228.522,228.52 2 753.34 0 StudyPeriodAverage2,31,360.686,128.31237,488.99234,988.696,216.54241,205.23 2 3,716.24 NNNSINKHOLESANDADISAPPEARINGLAKE:VICTORYLAKECASESTUDY226 N JournalofCaveandKarstStudies, December2014

PAGE 75

and2Hsignaturesshouldbesimilar.Conversely,ifthe groundwaterandsurfacewaterarenotconnected,the resultswillbedistinctlydifferent.Thedatashownin Figure5indicatethatthesurfacewatersamplesareheavier in18Oand2H.Thisoccurswhenthesurfacewatersamples haveundergoneevaporationprocessesthatselectively removethelighterisotopes(16Oand1H),leavinganexcess of18Oand2H.Thegroundwatersamplesshowthe oppositesignature.Thedistinctdifferenceintheenvironmentalisotopesignaturesofgroundwaterandsurface watersamplesindicatethatduringourstudyitislikelythat thesystemswerenotconnected.Inaddition,from monitoringwellsatthesouthernendofthelake,we measuredchangesinthewatertableweekly.Throughout thestudythegroundwatertablewasmorethan9mbelow thesurface.Withthisevidencewechosetoeliminate groundwaterinflowfromourwaterbudget. Lake-waterseepageintothegroundisachallenging parametertomeasure.Inthisstudyweconducteda RhodamineWTfluorescentdye-tracertesttocheckfor seepagefromsurfacewatertogroundwaterthroughaslug injectionatthenorthernendofVictoryLake.Duringour samplingintervalwedidnotmeasureanyRhodamineWT ingroundwatersamples.Next,weconsideredthevery distinctenvironmentalisotoperesultsweretrievedforthe groundwaterandsurfacewatersamples.Moreover,we consideredthat,priortohuman-inducedsinkholeformationatthesouthernendofVictoryLake,thelakebasin, whichisflooredwithclay,heldwaterfornearlysixtyyears. Forthepurposesofthiswaterbudgetweeliminated groundwateroutflowasaparameterduringthedurationof thestudy.Futurestudieswillincludegeophysicalinvestigationtoconfirmgroundwaterflowpaths. Theinflowvariablesmeasuredinthisstudyinclude inflowfromBigDryCreekdrainingintoVictoryLakeand rainfallontoit.Thisstudydeterminedthatthestream isephemeralandonlyflowsafterintenserainfallevents ( 35mm)orprolongedrainfall(12hoursorgreater).The drainagebasinoftheprimarystreamis5km2.Thearea adjacenttothestreamjustupstreamofthelakeisacypress wetlandthatisusuallysaturatedwhenwerecordconsiderableinflowenteringVictoryLake.Fortheoverallwater budgetofVictoryLakeinflowmakesup48%(Fig.3). However,whenlookingatmonthlypercentagesrainfall dominatedoverstreaminflowformostmonths.This discrepancyisexplainedbylookingattheamountofwater enteringthesystem.Whensubstantialprecipitationevents occur,thensurfacerunoffwillcontributetotheinflow volumeofthestream.Thestreamdoesnotflowatother times.Therefore,theeffectsofmajorprecipitationevents aremultipliedbythesurfacewaterrunoffintobythe stream.Thesejumpsinstreaminflowvolumewereseenin December2007throughMarch2008(Table2).Moreover, whenprecipitationeventsarelesserindurationand intensity,asfromApril2008toJuly2008,thestreamdoes notflow. Duringthewinterandearlyspringmonthstheoutflow leavingthroughtheculvertwasthedominantprocess occurring 95%ofthetime(Fig.7).Aswouldbeexpected, theotheroutflowcomponent,evapotranspiration,became moreprominentduringspringandsummermonths (Fig.7).SinceVictoryLakeemptiedin1986,thelake basinhasnotbeenfullforprolongedperiodsoftime.Asa result,fast-growingplantsandsedimentationhaveencroachedonthebasintoinfillmostportions.Thisiswhy weconsiderevaportranspirationinthisstudy.More influentialbyvolumeoftotallossthanevapotranspiration wastheamountofwaterlostthroughtheculvertatthe southernendofthelakebasin(49%overall).CONCLUSIONSNatureorhuman-inducedhydrologicchangescanalter thefundamentalcharacterofkarstsites.Worldwide, sinkholeshavebeenanincreasingproblem.Forexample, a60kmstretchoflandalongtheDeadSeacoasthasseen anincreaseinnewsinkholeformationattherateof150to 200peryear,causingdrasticalterationstonearbylakes andgroundwaterflowpaths(Yechielietal.,2006).Similar challengeshavebeendocumentedineasternEnglandand westernGreece(Cooperetal.,2013;Deligiannietal., 2013).Asseenmostrecentlywiththeemptyingofthelakes atFiveBluesLakeNationalPark,Belize,naturalor human-inducedhydrologicchangescanalterthefundamentalcharacterofkarstsites(DayandReynolds,2012). SimilartoFiveBlueLakeNationalPark,Floridianlakes JacksonandScotthavealsoexperienceddrainingdueto sinkholeactivity(Penson,2002;McBrideetal.,2011). VariouswaterbudgetscalculatedbyUSGSinvestigatorsin karsticareas,likeFlorida,havehelpedresearcherstobetter understandwatermovementandseasonalvariability (Daltonetal.,2004;Spechler,2010;McBrideetal.,2011; Sepulvedaetal.,2012).Thiscasestudypresentsawater budgetthatcombinestheapproachesusedbyDaltonetal., Spechler,andMcBrideetal.,butwasappliedtocapturea signatureoftheRomeFormationlimestoneaquiferinthe RidgeandValleyProvinceofnorthwestGeorgia. VictoryLakewasonceapopular,picturesquemeeting placeandafocalpointforvisitorsandrecreation.Sinkhole collapseatVictoryLaketwenty-sevenyearsagoresultedin drasticchangetocampusaesthetics.Therestorationofthe laketoitsoriginalstateisofinteresttothefaculty, students,andstaff.Fromthedatacollectedandanalyzed forthisstudywebelievethatrestorationofVictoryLakeis possible.Sincethequarryingoperationsconcludedin2000 therehasnotbeensignificantsinkholeformationoccurring oncampus,andthewatertablehasrebounded.Through thedevelopmentofthiswaterbudgetwewereableto confirmthatthegroundwatertablehasstabilizedandthat thecurrentbasincouldholdwaterwellenoughtosupport alake.Therearetwomainreasonsthattheoverall waterbudgetforthelakeshowedaloss.OverthepastT.J.JOVANELLYJournalofCaveandKarstStudies, December2014 N 227

PAGE 76

twenty-sevenyearsthelakehasnotbeenmaintainedor dredged.EphemeralinflowfromBigDryCreekhascaused sedimentationtooccurwithinthebasin,causingitto shallowfrom2.5-mdeeptolessthan1m.Theshallow basinmorphologyencouragessurfacewatertorunoff quickly.AsseeninFigures2cand2d,duringthegrowing seasonsofspringandsummerevapotranspirationdueto plantencroachmentintothelakedrasticallyreducesthe amountofwaterleftinthebasin.However,theauthor admitsthatonlylong-termmonitoringandcontinualdata collectionwouldconfirmthatthelakeperenniallyloses morewaterthanitgains,asthisresearchprovideslimited dataforonlyonehydrologicyear. Thechallengeindealingwithsituationsinvolvingthe rapidemptyingofkarstlakes,assuggestedbyDayand Reynolds(2012),isthatsuddenundergrounddrainageisa rarephenomenon.Theevidenceforsuccessofmanagement andrestorationofsuchlandscapestotheiroriginalformis inconclusive.Althoughseveralcreativeengineeringstrategies,suchasaclayorsyntheticlinerandasupplemental watersupply,havebeensuggestedtorebuildVictoryLake, itremainsdifficulttoweighthefinancialinvestment againstthelimitedguaranteesuggestedbykarsttopography.Moreover,therepercussionsofthelakeemptyingfor asecondtimecouldbethedestabilizationofdowngradientcover-collapseorcover-subsidencesinkholeson campus.Thislong-termrisktothehumanhealth,building infrastructure,andeconomiclossshouldbeweighed. DayandReynolds(2012)suggestthatappropriate managementstrategiesinkarstterrainsshouldacknowledge thetemporalvariabilityofthehydrologicregimeand preparevisitorsforanexperiencefallingwithinawide spectrumofhydrologicconditions.PerhapsthismanagementstrategyshouldbeconsideredatVictoryLake.The currentlandscapeatVictoryLakelendsitselfeasilytoa pleasingwetlandenvironmentthathostsvaryingwildlife seasonally.Inaddition,themysticofvisitingadisappearinglakemayprovideauniqueandinterestingenvironmentalinterpretationoflivinginadynamickarstlandscape.ACKNOWLEDGEMENTSTheauthorwouldliketothankthesponsorsoftheLaura MaddoxSmithEnvironmentalScienceGrantforproviding projectfundingandNSFgrant0620101forproviding resourcesandsupport.Gratit udeisextendedtotheBerry Collegestudentswhohelpedincollectingdata:MelissaKemm, RoySrymanske,JoshuaStevenson,andLaurynGilmer.REFERENCESAnderson,C.,1993,Indexandshortdescriptionofthegeologictermsused bytheGSS:BulletinoftheGeorgiaSpeleologicalSurvey,v.25, p.4. Aurit,M.D.,Peterson,R.O.,andBlanford,J.I.,2013,AGISanalysisof therelationshipbetweensinkholes,dry-wellcomplaintsandgroundwaterpumpingforfrost-freezeprotectionofwinterstrawberry productioninFlorida:PLoSONE,v.8,no.1,p.163.doi:10. 1371/journal.pone.0053832. Benito,G.,PerezdelCampo,P.,Gutierrez-Elorza,M.,andSancho,C., 1995,Naturalandhuman-inducedsinkholesingypsumterrainand associatedenvironmentalproblemsinNESpain:Environmental Geology,v.25,no.3,p.156.doi:10.1007/BF00768545. Buhlman,K.A.,2001,Abiologicalinventoryofeightcavesin northwesternGeorgiawithconservationimplications:Journalof CaveandKarstStudies,v.63,no.3,p.91. Chescheir,G.M.,Amatya,D.M.,andSkaggs,R.W.,1995,Monitoringthe waterbalanceofanaturalwetland, in Cambell,K.L.,ed.,Versatility ofWetlandsintheAgriculturalLandscape:St.Joseph,Michigan, AmericanSocietyofAgriculturalEngineers,p.451. Cooper,A.H.,Odling,N.E.,Murphy,P.J.,Miller,C.,Greenwood,C.J., andBrown,D.S.,2013,Theroleofsulfate-richspringsand groundwaterintheformationofsinkholesovergypsumineastern England, in Land,L.,Doctor,D.H.,andStephenson,J.B.,eds., Proceedingsofthe13thMultidisciplinaryConferenceonSinkholeand theEngineeringandEnvironmentalImpactsofKarst:Carlsbad, NationalCaveandKarstResearchInstitute,Symposium2,v.13, p.141. Craig,H.,1961,Isotopicvariationsinmeteoricwaters:Science,v.133, p.1702.doi:10.1126/science.133.3465.1702. Craig,H.,andGordon,L.I.,1965,Deuteriumandoxygen18variationsin theoceanandmarineatmosphere, in Tingiorgi,E.,ed.,Stable IsotopesinOceanographicStudiesandPaleotemperatures:Pisa, LaboratoriodeGeologiaNucleare,p.9. Dalton,M.S.,Aulenbach,B.T.,andTorak,L.J.,2004,Ground-Waterand Surface-WaterFlowandEstimatedWaterBudgetofLakeSeminole, SouthwesternGeorgia,andNorthwesternFlorida:U.S.Geological Survey,ScientificInvestigationsReport2004-5073,49p. Day,M.,andReynolds,B.,2012,FiveBluesLakeNationalPark,Belize: Acautionarymanagementtale:JournalofCaveandKarstStudies, v.74,no.2,p.213.doi:10.4311/2011JCKS0203. Deevey,S.E.,Jr.,1988,EstimationofdownwardleakagefromFlorida lakes:LimnologyandOceanography,v.33,no.6,p.1308. Deligianni,M.G.,Veni,G.,andPavlopoulos,K.,2013,Landuseand limitationsinthesinkholeandpoljekarstoftheKsiromeroRegion, WesternGreece:CarbonatesandEvaporites,v.28,no.1,p.167 173.doi:10.1007/s13146-013-0133-x. Fidelibus,M.D.,Gutierrez,F.,andSpilotro,G.,2011,Human-induced hydrogeologicalchangesandsinkholesinthecoastalgypsumkarstof LesinaMarinaarea(FoggiaProvince,Italy):EngineeringGeology, v.118,no.1,p.1.doi:10.1016/j.enggeo.2010.12.003. Florea,L.J.,Budd,D.,andBrinkmann,R.,2009,Cavesandkarstof West-CentralFlorida, in Palmer,A.N.,andPalmer,M.V.,eds.,Caves andKarstoftheUSA;Huntsville,NationalSpeleologicalSociety, p.189. Hamilton-Smith,E.,2006,Spatialplanningandprotectionmeasurefor karstareas:ActaCarsologica,v.35,no.2,p.5. Healy,R.W.,Winter,T.C.,LaBaugh,J.W.,andFranke,O.L.,2007, WaterBudgets:FoundationsforEffectiveWater-Resourcesand EnvironmentalManagement:USGeologicalSurvey,Circular1308, 90p. Hebbert,R.H.B.,andSmith,R.E.,1990,Hillslopeparameterestimation usingtheinverseprocedure:JournalofHydrology,v.119,p.307. doi:10.1016/0022-1694(90)90049-4. Horvat,B.,andRubinic,J.,2006,Annualrunoffestimationanexample ofkarsticaquifersinthetransboundaryregionofCroatiaand Slovenia:HydrologicalSciencesJournal,v.51,no.2,p.314. doi:10.1623/hysj.51.2.314. Hubbard,D.A.,Jr.,1988,SelectedKarstFeaturesoftheCentralValley andRidgeProvince:VirginiaDivisionofMineralResources, Publication83,1:250,000scalemapwithtext. Jenkins,M.,2009,Deepsoutherncaves:NationalGeographicMagazine, June,2009,p.124. Krabbenhoft,D.P.,Anderson,M.P.,andBowser,C.J.,1990a,Estimating groundwaterexchangewithlakes:2.Calibrationofathree-dimensional,colutetransportmodeltoastableisotopeplume:WaterResources Research,v.26,p.245562.doi:10.1029/WR026i010p02455. Krabbenhoft,D.P.,Bowser,C.J.,Anderson,M.P.,andValley,J.W., 1990b,Estimatinggroundwaterexchangewithlakes:1.Thestable isotopemassbalancemethod:WaterResourcesResearch,v.26, p.2445.doi:10.1029/WR026i010p02445.SINKHOLESANDADISAPPEARINGLAKE:VICTORYLAKECASESTUDY228 N JournalofCaveandKarstStudies, December2014

PAGE 77

Krabbenhoft,D.P.,Bowser,C.J.,Kendall,C.,andGat,J.R.,1994,Useof oxygen-18anddeuteriumtoassessthehydrologyofgroundwater/lake systems, in Baker,L.A.,ed.,EnvironmentalChemistryofLakesand Reservoirs:Washington,D.C.,AmericanChemicalSociety,p.67. doi:10.1021/ba-1994-0237.ch003. Lindsey,B.D.,Katz,B.G.,Berndt,M.P.,Ardis,A.F.,andSkach,K.A., 2010,RelationsbetweensinkholedensityandanthropogeniccontaminantsinselectedcarbonateaquifersintheeasternUnitedStates: EnvironmentalEarthSciences,v.60,no.5,p.1073.doi:10.1007/ s12665-009-0252-9. Loague,K.M.,andFreeze,R.A.,1985,Acomparisonofrainfall-runoff modelingtechniquesonsmalluplandcatchments:WaterResources Research,v.21,no.2,p.229.doi:10.1029/WR021i002p00229. MacIntyre,D.F.,1986,AquantitativereviewofthehydrologyofLake Jackson,Florida,1971-81,[MastersThesis]:Gainesville,Universityof Florida,238p. Martinez,J.D.,Johnson,K.S.,andNeal,J.T.,1998,Sinkholesin evaporiterock:AmericanScientist,v.86,no.1,p.38.doi:10.1511/ 1998.1.38. McBride,W.S.,Bellino,J.C.,andSwancar,A.,2011,Hydrology,Water Budget,andWaterChemistryofLakePanasoffkee,West-Central Florida:U.S.GeologicalSurveyScientificInvestigationsReport 201037,96p. McCarthy,E.J.,Skaggs,R.W.,andFarnum,P.,1991,Experimental determinationofthehydrologiccomponentsofadrainedforest watershed:TransactionsoftheASABE,v.34,no.5,p.2031. Newton,J.G.,1987,Developmentofsinkholesresultingfrommans activitiesintheEasternUnitedStates:U.S.GeologicalSurvey, Circular968,54p. Penson,G.,2002,Restoringadisappearinglake:LandandWater,v.46, no.5,p.1. Praise,M.,andGunn,J.,2007,NaturalandAnthropogenicHazardsin KarstAreas:Recognition,AnalysisandMitigation:London,GeologicalSocietySpecialPublications279,202p. Rosenberry,D.O.,Sturrock,A.M.,andWinter,T.C.,1993,Evaluationof theenergybudgetmethodofdeterminingevaporationatWilliams Lake,Minnesota,usingalternativeinstrumentationandstudy approaches:WaterResourcesResearch,v.29,no.8,p.2473. doi:10.1029/93WR00743. Scheidt,J.,Lerche,I.,andPaleologos,E.,2005,Environmentaland economicrisksfromsinkholesinwest-centralFlorida:Environmental Geosciences,v.12,p.207.doi:10.1306/eg.05130404009. Schuster,P.F.,Reddy,M.M.,LaBaugh,J.W.,Parkhurst,R.S.,Rosenberry,D.O.,Winter,T.C.,Antweiler,R.C.,andDean,W.E.,2003, Characterizationoflakewaterandgroundwatermovementinthe littoralzoneofWilliamsLake,aclosed-basinlakeinnorthcentral Minnesota:HydrologicProcesses,v.17,no.4,p.823.doi:10. 1002/hyp.1211. Sepulveda,N.,Tiedeman,C.R.,OReilly,A.M.,Davis,J.B.,andBurger, P.,2012,GroundwaterFlowandWaterBudgetintheSurficialand FloridanAquiferSystemsinEast-CentralFlorida:U.S.Geological SurveyScientificInvestigationsReport2012-5161,214p. Spechler,R.M.,2010,HydrogeologyandGroundwaterQualityof HighlandsCounty,Florida:U.S.GeologicalSurveyScientific InvestigationsReport2010,84p. Weary,D.J.,2005,AnAppalachianregionalkarstmapandprogress towardsaNewNationalKarstMap, in Kuniansky,E.L.,ed.,U.S. GeologicalSurveyKarstInterestGroupProceedings,RapidCity, SouthDakota,September12,2005:U.S.GeologicalSurvey ScientificInvestigationsReport2005-5160,p.93. Winstanley,D.,Angel,J.A.,Changnon,S.A.,Knapp,H.V.,Kunkel, K.E.,Palecki,M.A.,Scott,R.W.,andWehrmann,H.A.,2006,The WaterCycleandWaterBudgetsinIllinois:AFrameworkfor DroughtandWater-SupplyPlanning:Champaign,Illinois,Illinois StateWaterSurveyI/EM2006-02,114p. Winstanley,D.,andWendland,W.M.,2007,Climatechangeand associatedchangestothewaterbudget, in Dando,W.A.,ed.,Climate ChangeandVariations:APrimerforTeachers,v.1,NationalCouncil forGeographicEducation,PathwaysinGeography35. Winter,T.C.,1981,Uncertaintiesinestimatingthewaterbalanceoflakes: JournalofAmericanWaterResourcesAssociation,v.17,no.1, p.82. Winter,T.C.,1985,Approachestothestudyoflakehydrology, in Likens, G.E.,ed.,EcosystemApproachtoAquaticEcology:MirrorLakeand itsEnvironment,p.128. Winter,T.C.,1999,Relationofstreams,lakes,andwetlandsto groundwaterflowsystems:HydrogeologyJournal,v.7,p.28. doi:10.1007/s100400050178. Yechieli,Y.,Abelson,M.,Bein,A.,Crouvi,O.,andShtivelman,V.,2006, SinkholeswarmsalongtheDeadSeacoast:Reflectionof disturbanceoflakeandadjacentgroundwatersystems:Geological SocietyofAmericaBulletin,v.118,no.9-10,p.107587.doi:10. 1130/B25880.1. Yehdeghoa,B.,Rozanski,K.,Zojer,H.,andStichler,W.,1997, Interactionofdredginglakeswiththeadjacentgroundwaterfield: anisotopestudy:JournalofHydrology,v.192,p.247. doi:10.1016/S0022-1694(96)03102-2.T.J.JOVANELLYJournalofCaveandKarstStudies, December2014 N 229

PAGE 78

BOOKREVIEW CoastalKarstLandforms MichaelJ.LaceandJohnE.Mylroie(eds.),2013.Coastal ResearchLibrary,vol.5,Springer,Dordrecht,The Netherlands,429p.ISBN978-94-007-5015-9,$129(hardcover,7.2310.2inches),$99(eBookPDFformat). Thisbookispartofaseriesoncoastalresearchwritten forgeoscientistsandlandmanagers,butwhichwillalso appealtonon-technicalcavers.Mostofthetwenty-one authorsarefromthegroupofAmericanscientistsand cartographerscoordinatedbyJohnandJoanMylroieof MississippiStateUniversity.Thisbookisasummaryof theirworkonislandkarstoverthepastthirtyyears.Other contributorsincludespecialistsinvariousfieldsor geographicareas.Eachchapterhasseparateauthorship, andalthoughtheyvaryindepthandstyle,theyarewell integrated.Theresultisthemostaccessiblecoverageof thistopicandisanimportantadditiontothekarst literature. Thebookgivesapositiveimpression,withitsclear layout,strikingcolorphotos,well-draftedillustrations, andglossary.Itincludestwoparts:principlesofcoastal karstdevelopment,andselectedcasestudies.Topicsinthe firstsectionincludepseudokarstcaves,erosionaland depositionalfeatures,hydrologyandgeochemistry,coastalkarstdevelopmentincarbonaterocks,biologyand archeology,andkarstresourcesmanagement.Thechapter oncoastalkarstdevelopmentisperhapsthefocalpointof thesection,asitlaysoutthevariousmodelsofcaveorigin devisedbytheteam.Thisincludesthewell-known CarbonateIslandKarstModel,whichassociatesvarious karststyleswiththelocalgeologicsetting.Coverageof biologyisbrief.Itcouldhaveexpandedtoincludethe occurrenceofguano,whichonseveralislandshas spawnedanimportantindustry,andmicroorganisms, whichplayimportantrolesinredoxreactions. Part2describesspecificexamplesofcoastalkarstthat havebeenmuchcitedastypeexamples.Withitsdetailed cavedescriptions,thissectionwillbeofinteresttonontechnicalcavers.ThesechaptersincludetheBahama Islands,PuertoRicoanditsoutlyingislands,Barbados, Mallorca,theMarianaIslands,seacavesalongthe westernU.S.coast,Florida,andtheYucatanPeninsula. OnlargeislandssuchasPuertoRico,thecoverageof karstislimitedtocoastalfeatures,sincethegreatamount ofstrictlymeteorickarstliesoutsidethescopeofthe book.Inareaswherethereisageneticrelationship betweencoastalkarstandinlandfeatures,thecoverage includesboth.Especiallyrelevanttokarstresearchersis therevelationthatinMallorcaandtheYucatanPeninsula thetypicalspongeworkpatternofcoastalcavesisrelated topoorlycemented,younglimestonesandisnotspecific tosaltwater-freshwatermixing.Achapteroncoastalkarst intelogeneticlimestonesisinteresting.Whileother chaptersmakedistinctionsbetweentruekarstand pseudokarst,thischaptergivesexamplesofvoidsformed ininduratedlimestonethatmimictheshapesofflankmargincaves. Limitsonspaceandtimewereapparentinsome chapters.CoverageofcoastalcavesinAustraliahasa narrowfocus,asitcoversmainlyafewcavesvisitedbythe authors.Anoverviewofthecoastalkarstresearchby authorssuchasthelateJoeJenningswouldhavebeena goodaddition.Thereisnomentionofcavesinthehuge NullarborPlateau,manyofwhichopentotheoceanand arethoughttorepresentrelictsea-levelstands.Those interestedinthemixing-zonemodelmightwishfora chapteronchemicalfielddatathatsupportthesolutional models.Italsowouldhavebeeninstructivetoincludean overviewofthesignificanceofcoastalkarstininterpreting pastsea-levelstandsandtheirrelationshiptoothersealevelindicators. DOI:10.4311/2013ES0129230 N JournalofCaveandKarstStudies, December2014

PAGE 79

Thisbookisafineadditiontothekarstliterature.It coversatopicthathadreceivedonlysparsecoverage beforethisgroupofauthorsbegantheirlong-termfocuson coastalandislandkarst.Itsdetailsarelargelygeomorphic anddescriptive,whichmakesitattractivetoawide readership,butitincludesenoughconceptsanddetailed siteanalysestomakeitoftechnicalvalueaswell.ReviewedbyMargaretV.Palmer,619WinneyHillRoad,Oneonta,NY 13820(palmeran@oneonta.edu).BOOKREVIEWJournalofCaveandKarstStudies, December2014 N 231


printinsert_linkshareget_appmore_horiz

Download Options

close
Choose Size
Choose file type
Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.