Temporal changes in groundwater quality of the Saloum coastal aquifer

(1)

JournalofHydrology:RegionalStudies9(2017)163–182

ContentslistsavailableatScienceDirect

Journal

of

Hydrology:

Regional

Studies

jou rn a l h om ep a ge :w w w . e l s e v i e r . c o m / l o c a t e / e j r h

Temporal

changes

in

groundwater

quality

of

the

Saloum

coastal

aquifer

Ndeye

Maguette

Dieng

a,c,∗

_,

_Philippe

_Orban

a

_,

_Joel

_Otten

a

_,

_Christine

_Stumpp

b

_,

Serigne

Faye

c

_,

_Alain

_Dassargues

a

a_Hydrogeology_&_{Environmental}_Geology_–_Dpt_ArGEnCo,_Aquapole,_University_of_Liège,₄₀₀₀_Liège,_Belgium

b_Institute_of_Groundwater_Ecology,_Helmholtz_Zentrum_München,_{Ingolstädter}_Landstrasse_1,_D-85764_Neuherberg,_Germany

c_Department_of_Geology_–_FST,_University_of_Cheikh_Anta_Diop,_Dakar,_Senegal

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received27April2016

Receivedinrevisedform7December2016

Accepted10December2016

Keywords:

Groundwatersalinization

Selforganizingmaps

Geochemicalprocesses

Environmentalisotopes

Salouminverseestuary

a

b

s

t

r

a

c

t

Studyregion:GroundwaterinthesouthernpartoftheSaloumBasininSenegal.

Studyfocus:TheSaloumestuaryisahypersalineand‘inverse’estuarywherethesalinityof riverwaterincreasesintheupstreamdirection.Thisregionisproblematicinthatdueto theunderlainsuperﬁcialContinentalTerminalaquiferborderedbythehypersaline estu-aryconstitutestheuniquefreshgroundwaterreservoirforwatersupplyforitsestimated 466,000residentslivingin18ruraldistricts(belongingtotheregionsofFatick,Kaolack andKaffrine).ThisisofhighvaluegiventhatthedeepMaastrichtianaquifer(200–300m depth)issaline.Thisstudyaimstodescribeandunderstandtemporalchangesinthe chem-icalandisotopiccompositionsofgroundwater,thegeochemicalprocessesandespecially thegroundwatersalinization.

Newhydrologicalinsightsfortheregion:Theanalyticaldatawerediscriminatedinto3groups onthebasisofthewatertypes.Na-Cl,Ca-ClandCa-SO4richwatersderivedfromsaline waterintrusionatthevicinityoftheSaloumRiveraccompaniedbyionexchangereactions andpollutiondominatethefirstgroup.Thesecondgrouplocatedmainlyinthecentre andeasternpartsoftheregionisfeaturedfreshgroundwaterofCa-HCO3derivedfrom calcitedissolutionreactions.ThethirdgroupofNa-HCO3typeandlessmineralizedindicates fresheningprocessesbyrecentlyinfiltratingrainwaters.Slightseasonalchemicalvariations areobservedduetonewinfiltratingwaterreachingthewatertable.

Highvariationinrainfallbetweenthe2referenceyears(2003and2012)alsochanges chemicalpatternsinthegroundwater.Chemicalevolutionofthegroundwateris geographi-callyobservedandisduetoacombinationofdilutionbyrecharge,anthropiccontamination andseawaterintrusion.Theresultsofenvironmentalisotopes(␦18_O,_␦2_H)_compared_with thelocalmeteoriclineindicatethatthegroundwaterhasbeenaffectedbyevaporation pro-cessesbeforeandduringinﬁltration.Theresultsalsoclearlyindicatemixingwithsaltwater andanevolutiontowardsrelativefresheningbetween2003and2012insomewellsnear theSaloumRiver.

∗ Correspondingauthorat:DepartmentofGeology–FST,UniversityofCheikhAntaDiop,Dakar,SenegalandHydrogeology&EnvironmentalGeology

–DptArGEnCo,Aquapole,UniversityofLiège,4000,Liège,Belgium..

E-mailaddresses:nmdieng@doct.ulg.ac.be,ndeye81.dieng@ucad.edu.sn,diengmaguette@yahoo.fr,mdmaguettedieng@gmail.com(N.M.Dieng).

http://dx.doi.org/10.1016/j.ejrh.2016.12.082

(2)

164 N.M.Diengetal./JournalofHydrology:RegionalStudies9(2017)163–182

Introduction

TheContinentalTerminal(CT)aquiferisauniquefreshwaterreservoir,whichsupportsvariouswaterneedsintheSaloum regionlocatedmid-WestofSenegal.Unfortunately,thisresourceisthreatenedbysaltwaterintrusionderivedfromboththe oceaninthewesternpartandfromthehypersalineSaloumRiverwhichlimitsthesysteminthenorthernpart.TheSaloum hydrologicsystemisatide-influencedinverseestuaryinwhichhighevaporationratescombinedwithlowflowinthis contextofflattopographyresultinsaltdepositionandhighsalinityofthesurfacewaterbody.AsdefinedbyPritchard (1967),ininverseorhypersalineestuarysystems,thesalinityoftheriverwatersubstantiallyexceedsseawatersalinity. Thispatternisgovernedbythenetlossoffreshwaterusuallyresultingfromanexcessofevaporationoverrunoffand precipitation.Thesesystems,forwhichsalinityincreasesupstreamfromtherivermouthtowardsinlandregions,occur mostlyinaridorsemi-aridpartsoftheworld(Largieretal.,1997),suchasAustralia(Wolanski,1986)andWestAfrica (Barusseauetal.,1985;Diop,1986;PageandCiteau,1990;SavenijeandPagès,1992;RiddandStieglitz2002;Mikhailovand Isupova,2008).Therefore,groundwatersalinizationintheseareasiscloselylinkedtodirectsaltwaterintrusionandhigh evaporation.PreviousstudiesinAustraliaevidencedgroundwatercontaminationprocessesthroughestuarinesaltwater intrusionintheSwan-CanningEstuary(SmithandTurner,2001)andinthePioneerValley,NorthernQueensland(Werner andLockington,2004).IntheSaloumregion,Noël(1975)thenDiluca(1976)investigatedthegeometry,flowregime,and chemicalcharacteristicsofgroundwaterintheCTaquifer.Salinitypatternandassociatedgeochemicalprocessesusingbasic ionchemistryweredescribedbyFayeetal.(2001,2003,2004,2009,2010).Theresultspointedoutthatthegroundwater resourceisconstrainedbytheoccurrenceofmoderatetohighsalinityinthenorthern(neartheSaloumRiver)andwestern (coastalarea)partsoftheaquifer.There,manyshallowwellshavebeenimpactedbytheoccurrenceofsalinewater.Further investigationsweredonebyFayeetal.(2004)usingminorelements(B,Br,Sr)andisotopicindicatorstogetherwithsome majorionsforconstrainingsource,andrelativeageandsorption/desorptionreactionsrelatedtorefresheningandsalinization processes.Inthislatterstudy,boronconcentrations(7–650␮g/L)weretestedagainstthebinarymixingmodel.Results showedsalinizationprocessesareevidencedbyBsorptionandNadepletionwhentheSaloumRiverwaterintrudesthe aquifer(salinization)inthenorthernpartoftheregion.Conversely,BdesorptionandNaenrichmentoccurasthefresh groundwaterflushingdisplacesthesalinegroundwatersinthecoastalstrip(refreshening).Inthecentralzonewhereancient intrusionprevailed,theprocessoffresheningofthesalinegroundwaterwasindicatedbythechangesinmajorionchemistry aswellasBdesorptionandNaenrichment.TherelativelystableisotopeandtritiumcontentsoftheCTaquiferindicatedthat groundwaterisderivedfromrecentprecipitationandmixingwithrecentlyinfiltratingwatersandevaporationcontribute tothechangesinisotopicsignature.

Therefore,salinisationconstitutesamajorissueintheSaloumregionduetothefactthatpotabledrinkingwaterneeds relysolelyontheCTgroundwaterresourcegiventhatthedeepMaastrichtianaquifer(200–300m)exhibitshighchloride contents(500–1000mg/L)andsomehazardouselements,suchasﬂuorideandboron,whichreachconcentrationsof2–4 and1–2mg/L,respectively(Travi,1988).Therefore,theCTaquiferrepresentsauniqueaccessiblefreshwaterresourceinthe region.

Despiteofthesevaluableresults,temporal(bothseasonalandlongterm)effectswerenotadressed.Therefore,thepresent studyaimsatreassessingthegroundwaterqualitychangesoftheCTaquiferinthelightofthenewconstraints(e.g.changes inclimaticconditionsandpumpingconstraints),andthenewgroundwaterflowpatternconfiguration.Specifically,this studyintendsto:(1)identifyseasonalvariationsinthechemicalpatternsinrelationtogeochemicalreactionsanddilution usingstatisticaltools,suchasSelfOrganizingMaps(SOMs),and(2)assessthetemporaltrendofgroundwaterchemistry andsalinizationusingchemicalandisotopicgroundwatercompositions.

2. Studyarea

2.1. Locationandgeomorphology

ThestudyareaislocatedinthewesternpartofSenegalandcoversthesouthernpartoftheSaloumhydrologicBasin exceptforthemangrovesystem.Thesurfaceareaisapproximately3380km2 ₍_Fig.₁₎_and_is_limited_to_the_North_by_the SaloumRiver,totheWestbythemangroveareaandtotheSouthandEastbytheGambiaRivercatchment.

Thegeomorphologicsettingwhichconsistsofagentlyslopingplainthatextendstowardthecoastisincisedbyanetwork ofdepressionswhere temporarystreamsdrainrun-offwaterduringtherainyseason.Adigitalelevationmodel(_±4m precision)thatwascreatedfromtopographicalmaps(1:50,000)andmeasuredpointsshowselevationsrangingfrom−1m intheWestandtheNorth(nearthemangrovesandtheSaloumRiver)to+50mintheSouthandEasttowardstheplateau zoneborderedbytheGambianBasin(Fig.1).InthenorthernandwesternregionsalongtheSaloumRiverandthecoast, respectively,asuccessionofhighareas(10–30m)anddepressions(0–10m)formsthelandscape.

Landuse/landcoverofthestudyareawereinvestigatedusingOctober2010Landsatimage(ETM)andlineardiscriminant classiﬁer(LDC)supervisedclassiﬁcationmethod(Diengetal.,2014).Surfacefeaturesgeneratedcompriserainfedagricultural crops(52%),naturalvegetation(40%)(composedofsavannah,woodedpark,bushlandandshrubsteppetrees),dispersed habitationspredominantlyvillages(3.5%),saltybarrensoilslocallycalled“tanne”(2%),surfacewaterbodies(1.5%)and mangroveecosystemareas(1%).

(3)

N.M.Diengetal./JournalofHydrology:RegionalStudies9(2017)163–182 165

Fig.1. Locationoftheinvestigatedareaandmonitoringnetwork.Thedarklineindicatesthelocationofthecross-sectionA-A’presentedinFig.2.

2.2. Climate

Theclimateoftheregionistropicalwithtwodistinctseasons:adryseasonfromNovembertoMayandarainyseason fromJunetoOctober.Theaverageannualrainfallfrom1960to2012variedfrom750mminthesouthernpartofthebasinat Niorostationto615mminthenorthernpartatKaolackstationwithaSouth-Northdecreasinggradient.Therainfallrecord (Fig.3)atKaolackstationcomparedtothemeanreference1960–1990showsahumidperiodfrom1928to1971followed byadryerperiodfrom1971totheendofthe1990s,whichoccurredthroughmostoftheSahelzone.InsomeSub-Saharan countries,suchasthestudyarea,thisdryperiodwasfollowedbyprogressivelywetteryears(LebelandAli,2009).

Availableclimaticdata(i.e.,precipitation,temperature,relativehumidity,insulation,andwindspeed)fortheperiodfrom 1981to2012collectedfromtheSenegalNationalMeteorologicalAgencyattheKaolackandNioroweatherstationsshow thattheaverageannualtemperaturerangesfrom28to29◦C.Theaverageannualpotentialandactualevapotranspiration valuescalculatedusingthePenmanmethod(Allenetal.,1998)are2030and552mm/yrrespectively.

2.3. Hydrology

ThehydrologicsystemoftheregionismainlycomposedbytheSaloumRiveranditstwotributaries,theBandialaand DiombossrRivers.TheNemaRiver,whichistheonlyperennialfreshwaterstream,ﬂowstowardstheBandialaRiver(Fig.1). Itreceivesagroundwaterbaseﬂowofbetween0.012and0.03m3_/s_during_the_dry_and_rainy_seasons,_respectively₍_Ngom, 2000).Downstreamtheseriversandstreams,smallseawatercreekslocallycalled“bolons”occurinadditiontothelarge low-lyingestuarybearingtidalwetlands,andthemangroveecosystem.Asdiscussedpreviously,theSaloumestuaryisaninverse estuary.MeasurementsdataduringthedryseasonshowedthatthesalinityoftheSaloumRiverincreasesfrom36.7g/Lat themouthtomorethan90g/latKaolack(located120kmfromthemouth)(Diop,1986).PageandCiteau(1990)usingdata recordrevealedthatsalinityoftheSaloumRiverishighlysensitivetoclimatevariationswithagradientofapproximately 1.3g/lperyearbetween1950and1986(Fig.3).ThisobservationisinaccordancewithworksofSavenijeandPagès(1992) inwesternAfricaninverseestuarieswhichevidencedsharpincreaseinsalinitygradientsduringtheSaheldroughtsinthe 1960s.

2.4. Geologicalandhydrogeologicalcontext

ThestudyareabelongstotheSenegalese-Mauritaniansedimentarybasin,whichisthelargestcoastalbasininnorthwest AfricaandbearsformationsfromtheCretaceoustotheQuaternary(Bellion,1987).Thegeologicalcontextoftheregion

(4)

Fig.2.CrosssectionofthestudyareashowingthegeologicalformationandthepiezometriclevelofCTgroundwater(crosssectionindicatedinFig.1).

hasbeenstudiedusinglithologicalinformationfromboreholesandhand-dugwellsinadditiontoregionalpreviousstudies (Lappartient,1985;LePriolandDieng,1985)andarecentregionalgeologicalmap(1:500000)(Rogeretal.,2009).TheCT formations,whicharethefocusofthisstudy,arepresentthroughouttheregionandarecoveredbyathinandyounger Quaternarysandlayerandlocallybyalluvialdepositsintheriverplains.Theseformationsoverlietheimperviousmarl/clay Eoceneformations.

TheCTsedimentsaredetritalmarineCenozoic(Oligo-MiocenetoPliocene)originandshowevidenceofintenseferralitic alterationduringthelateMiocene,includingferruginousconcretions(Lappartient,1985;ConradandLappartient,1987). Theyarelaterallyandverticallyheterogeneousandcontaindiscontinuousinterbeddedsandstone,sandyclay,clayeysand,silt andclay(Noël,1975;Lappartient,1985;ConradandLappartient,1987)(Fig.2).Locally,claysandstonebedsandferruginous crustsarepresentatthetopoftheformations.Kaoliniteformsthecementinthenon-ferruginoussandstoneandsand sedimentsandrepresentsthedominantclaymineral.Goethiteispredominantandformsthecementofthequartzgrainsin theferruginoussandstonehorizon.Tracequantitiesofillitehavebeenidentiﬁedwhileintheeasternpartoftheregion,the quartzgrainsarecoatedbycalcite(Lappartient,1985).

Hydrogeologically,theCTformationsareconsideredasunconfinedaquiferwhichthicknessincreasesfromNNW(less than50m)toSSE(approximately100m)andfromWest(lessthan50m)toEast(approximately100m)(Fig.2).Theaquifer reservoirliesonanirregularandimperviousEocenesubstratecomposedofmarlandlimestonecoveredbyclay(Fig.2)(Lyand Aglanda,1991;Sarr,1995).Thehydraulicconductivityandstoragecoefficientvaluesobtainedfrompumpingtestdataunder depthaveragedconditionsrangefrom0.5to8.6×10−4m/sandfrom0.5to21×10−2,respectively(Diluca,1976). Thirty-oneboreholeswithpumpingratesbetween10and45m3_/h_and_numerous_traditional_wells_are_used_for_the_water_supply inthisregionandtotalwithdrawalisestimatedat8000m3_/day._Groundwater_recharge_occurs_mainly_through_rainwater infiltration.Computedvaluesusingchloridemassbalance(CMB)method(chlorideprofilesatdifferentsitesobtainedduring thecourseofthisstudyandnotshownhere)intheunsaturatedzone,andhydrologicalbudget(ETocomputedbyPenman formula)arevariableandrangefrom17to100mm/yearandfrom19to130mm/year,respectively.Theserechargevalues aresimilartothoseobtainedbyEdmundsandGaye(1994)(11–108mm/year)usingthesameCMBmethod.Thespatial variabilityislikelytobeduetolocalvariationsinsoiltypeandtexture(notablyclaycontents)andtovegetationcover.The CMBapproachuncertaintiesinourcontextcanbelinkedtopotentialClcontributionfromothersourcessuchassaltwater intrusionoranthropogenicinput.Watertablemeasurementswereobtainedduringthe2012campaignsinMay(endofdry season)andinNovember(endofrainyseason).TheywereconvertedtoheadvaluesthroughanaccurateDGPStopographic surveyofthesamplingsites.ItallowstoreassessgroundwaterflowpatternsshowninFig.4forthedryseason.Asimilarflow patternisobtainedfortherainyseason(notshownhere).Head(h)distributionfeaturesthreegroundwatermoundslocated inthenorthwest(9.8≤h≤11.7m),inthesouthwestalongtheupstreampartoftheNemaRiver(25.3≤h≤26.6m),andin theNorth(13.2≤h≤16m).AmarkeddepressionisobservedNorthandparalleltotheSaloumRiver,wherethemeasured

(5)

Fig.3. InterannualvariationoftheSaloumRiversalinity(inred)andofthemeanannualrainfallmeasuredatKaolacklocalityfrom1927to2012modiﬁed

fromPageandCiteau(1990).

Fig.4.GroundwaterpiezometricmapintheCTaquiferduringthedryseasonof2012.

headsarebelowmeansealevel(aslowas−4m).Thislatterdepressionisacommonfeatureinmanyunconﬁnedaquifers insub-SaharanAfrica,suchasinTrarza/Mauritania,inYaere/NorthCameroon,inGondo,Nara,andAzaouad/Mali,andin Kadzell/Niger(Archambault,1960;Degallier,1962;Aranyossyetal.,1989;Ndiayeetal.,1993;Favreauetal.,2002;Ngounou Ngatchaetal.,2007;Koussoube,2010).IsotopicstudiesperformedinMali(Aranyossyetal.,1989),Niger(Favreauetal., 2002)andSenegal(Ndiayeetal.,1993)suggestthatthecombinedeffectsofhighevapotranspirationandlowhorizontal permeabilityarethemaincausesofthistypeofpiezometricdepressionssincepumpingratesintheseareaarelow.

(6)

Fig.5.Rechargemapandpunctualobservedseasonalchangesinpiezometricheads.

Basedonthisconﬁguration,severalgroundwaterﬂowpatternscanbedepicted:

−fromthenorthwestandnorthernmoundstowardstheseaandtheSaloumRiver,wheretheheadvaluesdecreasenear theshorebutarestillpositive;

−fromthesouthwestmoundinalldirections;occurrenceofthegroundwatermoundinthezonesustainstheNema Riverﬂowduringlowstageperiod(Ngom,2000);

−fromtheSaloumRiverandthesouthernmoundtothepiezometricdepression.

Theseasonalandlongtermheadvariationsobtainedfrom2012(dryandrainyseason)and1973datasetsmayreﬂect seasonalrechargeandinterannualevolution,respectively.Seasonalheadvariationsrecordedin2012rangefrom0.1to2.8m withthehighestvaluelocatedfarfromthepumpingareainazoneofrelativehighrecharge(>50mm).Thelowestvalueof headvariationisobservednearthepumpingareainazoneofrelativelowrecharge(<30m)(Fig.5).

Comparisonofthe2012headvalueswithheadsmeasuredduringpre-developmentanddriestperiodof1973infew referencepointsshowwatertablerisefrom0.7to6.8m(mean4.2m)(Fig.6).Thisisobserveddespitethepumpingfrom morethan100boreholesandnumerousdugwellssincetheearly70s.Thislongtermwatertableriseprobablyreﬂects sensitivityofthesystemtoclimateandespeciallytoannualrainfall(440and847mmrespectivelyin1973and2012).

3. Materialsandmethods

3.1. Fieldcampaignsandlaboratoryprotocols

Tocompletetheexistingdataset(includingthedataacquiredin2003),twoﬁeldcampaignswereconductedinMay2012 (endofthedryseason)andNovember2012(endoftherainyseason).Themeasuredandsamplednetworkcomprises forty-fourgroundwatersamplingpoints(composedof5boreholes,35hand-dugwellsand4piezometers)inadditiontothree surfacewatersamplingsitesalongtheSaloumRiver(atFoundiougne(FL1),Lindiane(FL2)andKaolack(FL3))(Fig.1).Prior tosampling,depthtowatertableandphysicochemicalparameterssuchaspH,electricalconductivity(EC)andtemperature (T)weremeasuredintheﬁeldateachsiteusingamulti-parameterprobe.Thegroundwatersampleswerecollectedafter purgingthepiezometerswhilewellsandboreholesusedforwatersupplydidnotneedtobepurgedbecausetheywere pumpedcontinuously.

Forchemicalanalyses,asetoftwosampleswascollectedforeachpointandsubsequentlyfilteredthrough0.45␮m membranesintopolyethylenebottles.OneacidifiedtoapHlessthan2usinghighpurityHNO3forcationanalysis,andthe otheronenotacidifiedforanionanalysis.

Forstableisotopeanalysesofoxygen(18_O)_and_hydrogen₍2_H),_selected_points_of_the_network_and_the_three_Saloum_River pointsweresampledduringthewetseasonandtightlysealedin1-lpolyethylenebottleswithoutﬁltered.

(7)

Fig.6. DifferencesbetweenpiezometriclevelsmeasuredinNovember1973andNovember2012in16wells.

Recordsofmeasuredquantityofrainand2012rainwatersampleswerecollectedfrom7climaticstations,namelyKaolack (K),Nioro(N),WakhNgouna(W),Foundiougne(F),Djilor(D),Sokone(S)andToubacouta(T).

Thechemicalanalysesformajorions(Ca2+_,_Na+_,_Mg2+_,_K+_,_Cl−_,_HCO₃−_,_SO₄2−_and_NO₃−₎_were_performed_using stan-dardmethodsofﬂameatomicabsorption,potentiometric,andtitrationattheLiegeUniversityHydrogeologyLaboratory (Belgium).TheEnvironmentalisotopeanalyseswereperformedattheHelmholtzZentrumMünchen,Instituteof Ground-waterEcology(Neuherberg/Germany)usingdual-inletmassspectrometry(Coplen,1988)andresultsareexpressedinthe conventional␦-permil(␦‰)notationwithreferencetotheViennaStandardMeanOceanicWater(VSMOW)withprecisions of±0.1‰for␦18_O_and_±1‰_for_␦2_H.

3.2. Dataanalysisandmultivariatestatisticalanalysis

Whiletheelectricalconductivity(EC)isroutinelyusedtomeasurethemineralisationofwater,conventionalmethods involvingmajorand minorionsanalysis,ionicratios,and isotopicmethodsareusedinassessinggroundwaterquality (Ghesquièreetal.,2015;Nicolinietal.,2016)andtheinteractionbetweenfreshwaterandsalinewater(Magaritzetal., 1981;Mercado,1985;Kimetal.,2003).

Multivariateanalysismethods,suchasFactorAnalysis(FA),PrincipalComponentAnalysis(PCA)andHierarchical Clas-sificationAnalysis(HCA),arealsooftenusedtoidentifymajorgroundwatergroupsandfactorsthataffectthegroundwater chemistryinaquifers(Mudry,1991;Güleretal.,2002;Belkhirietal.,2012;GambleandBabbar-Sebens,2012;Montcoudiol etal.,2015;Ghesquièreetal.,2015;Ibrahimaetal.,2015).However,thesemultivariateanalysesaregenerallybasedon linearprinciples(Mudry,1991;GambleandBabbar-Sebens,2012)andcannotovercomedifficultiesarisingfrombiases duetocomplexityandnon-linearityindatasetsandfromtheinherentcorrelationsbetweenvariables.TheSOMsmethod (Kohonen,1982)appliedhereisanalternativetotheselattermethodstoefficientlydealwithdatasetsthatareruledby complex,non-linearrelationshipsintheanalysisanddiagnosisofdynamicsystems(HongandRosen,2001;Hongetal., 2004;GarciaandShigidi,2006;Peetersetal.,2007;GambleandBabbar-Sebens,2012;Gning,2015).Itallowsclustering asetofhydrochemicaldataintotwoormoreindependentgroups(Kohonen,1995;).Thismethodoutperformstheother statisticaltoolsinthatitcansuccessfully:1)dealwiththenonlinearitiesofthesystem;2)bedevelopedfromdatawithout requiringmechanisticknowledgeofthesystem;3)handlenoisyorirregulardataandbeeasilyandquicklyupdated;and4) interpretandvisualizeinformationfrommultiplevariablesorparameters(Kohonen,2001).

ThecomponentplanesandU-matrixarevisualizationmapsthatuseahexagonaltopologytorepresentthemainoutput results.Thecomponentplanesshowhighandlowvaluesonastandardizedmaptoidentifycorrelationsbetweendifferent parameters.Ultimately,intheSOMsanalysis,allthesecomponentmapsarecombinedtocreateauniﬁedmatrix(U-matrix) thatcharacterizestheEuclideandistancesbetweeneachnode(Peetersetal.,2007;GambleandBabbar-Sebens,2012).The

(8)

clusteringisbasedonavisualinspectionoftheU-matrixwhereclusterscorrespondtodifferenttypesofgroundwater withrespecttothecontributionofeachelement.TheinformationobtainedfromSOMsapplicationcanthenbeusedto highlightwatermineralizationprocessesthatoccurintheaquiferbasedonthedegreeofsalinityandthespatiotemporal hydrogeochemicalvariability.

Inthisstudy,11chemicalvariables(EC,pCO2,pH,Ca2+,Mg2+,Na+,K+,Cl−,HCO3−,SO42−andNO3−)areprocessedusingthe SOMsToolboxoftheMATLABTM_code_to_ﬁrst_investigate_the_variables_that_affect_the_salinity_and_their_{relationships.}_Classical tools,suchasPiperdiagramsandbivariateanalysis,arethenusedtoconﬁrmtherelationshipsbetweenthevariables.

4. Resultsanddiscussion

4.1. Hydrochemicalcharacteristics

Theresultsfromthein-situphysico-chemicalsurveysand thechemicaland isotopesanalysesofgroundwater, sur-facewater,andseawatersamplesarepresentedinTable1,andthemeanconcentrationsfromtherainwateranalysesare presentedinTable2.Allsampleshaveionicelectricalbalanceslessthan5%,whichconﬁrmreliableanalyses.

ThesurfacewatersamplesthatwerecollectedatthethreesitesalongtheSaloumRiverhaveaveragepHvaluesfrom7.8 (dryseason)to8.3(wetseason),whichindicateslightlyalkalinecharacteristics.Theaveragesurfacewatertemperatures areclosetotheambienttemperatureandrangebetween28and30◦C.TheECvaluesvaryalongtheriverfrom64,700(dry season)to50,800␮S/cm(wetseason)atFoundiougne(10kmfromthemouth),from91,700(dryseason)to60,400␮S/cm (wetseason)atLindiane(90kmfromthemouth)andfrom89,000(dryseason)to55,200␮S/cmatKaolack(120kmfromthe mouth).TheseresultsshowthatthesalinityoftheSaloumRiversubstantiallyexceedsseawatersalinityvaluesandsupport theinverseestuarycharacteroftheSaloumRiverwithgraduallyupstreamincrease ofsalinity.However,thedecrease observedatKaolackisprobablyduetotheinﬂowoffreshgroundwaterasshowninthepiezometricmap(Fig.4).

Temperaturesofthegroundwatersamplesrangefrom27.7to30.9◦Candfrom26.4to33◦Cinthedryandwetseasons, respectively.ThepHvaluesvarybetween3.7and7duringthedryseasonandbetween4.6and8.8duringthewetseason. Thelowestacidicvalues(pH3–5)arefoundinwellS49,whichislocatednearthemangrovearea,andcouldbedueto theoxidationofpyrite,whichischaracteristicofmangrovedegradationgroundwaterconditions(Fordetal.,1992).The measuredECvaluesrangefrom74to3430␮S/cmandfrom47to2040␮S/cminthedryandwetseasons,respectively. AccordingtoWorldHealthOrganisationWHO(2008),thisbroadrangeofECvaluesdifferentiatestwotypesofgroundwater inthesystem:

1-freshwaterwithECvalueslowerthan900_␮S/cm.Thisgroupcompriseswatersampledfromwellsinthewesternand centralpartsoftheregion;

-salineand/orpollutedwaterswithECvalueshigherthan900␮S/cm.Thesesamplesoriginatefromthecoastalareas, neartheSaloumRiverandfromafewwellsinthecentralpartofthestudyarea.

4.2. Groundwaterqualityassessment

Theevaluationofthechemicalcharacteristicsofthegroundwatersamplesusingthedescriptivestatisticapproachshows anorderofrelativeabundance(inmeq/L)ofCa2+_>_Na+_>_Mg2+_>_K+_and_of_HCO

3−>Cl−>NO3−>SO42−.Itisthecaseforboth campaigns(dryandrainyseason).Ahighvariabilitybetweentheminimumandmaximumvalues(Fig.7)isobserved.The concentrationsofSO42−andNO3−,andtoalesserextentCa2+,Mg2+andHCO3−,arehigherforthegroundwatersampled duringtherainyseason.Thiswasprobablytheresultoftheinfiltrationofpollutedwaterandthedissolutionofcalciteand dolomiteduringinfiltration.TheconcentrationsofNa+_and_Cl−_in_these_groundwater_samples_decrease_less,_which_is_probably duetodilutionbyrechargewater.MostofthemajorionconcentrationsarewithintheWHOrangesfordrinkingwaterexcept forsomecasesinwhichthehighestvaluesofsodiumNa+_,_Cl−_,_NO₃−_and_SO₄2−_exceed_the_drinking_water_threshold._This suggestssalinewaterintrusionandanthropogeniccontaminationasshownbytheECvalues.Theshallowestwells(S23,S49 andS52)arecharacterizedbyhighNO3-contents(195,200and327mg/L,respectively)duetolocalcontaminationfrom domesticwasteandpastoralactivitiesaroundthewells.ThehighestSO42−concentrationof582mg/LisrecordedinwellS49 andisassociatedwithhighconcentrationsofCl−(215mg/L),Na+₍₁₁₁_mg/L),_Ca2+₍₂₁₄_mg/L),_Mg2+₍₅₃_mg/L),_K+₍₄₆_mg/L) andNO3−(200mg/L).ThiswellislocatedatNemavillageclosetothemangroveecosystemprobablyreflectstheimpactof salinewaterintrusionandassociatedprocesses(gypsumdissolutionand/orpyriteoxydation)andanthropogenicpollution atwellhead.

4.3. GroundwaterclassiﬁcationusingtheSOMsmethod

TheSOMsmethodappliedtobothdatasets(rainyanddryseasons)exhibitssimilarpatternandinthispaperweonly describeresultsobtainedfromthedryseasoninFig.8aandb.ThecomponentmapsofCl−,Na+_,_Ca2+_and_Mg2+₍_Fig.₈_a)_show similardistributionpatternstoECmap,whichindicatesthattheseionsarehighlycorrelatedwithECandthatdilutionisthe mainfactorinducingvariability.ThehighcorrelationbetweenNa+_and_Cl−_suggests_a_common_origin_from_the_dissolution ofhaliteand/orsaltwaterintrusion.Toalesserdegree,HCO3−correlatestoCa2+.ThepHandHCO3−alsohavesimilarmaps but,asiscommonlyfoundingroundwater,areinverselycorrelatedwiththeCO2pressure.

(9)

N.M. Dieng et al. / Journal of Hydrology: Regional Studies 9 (2017) 163–182 171 Table1

Physico-chemicalmeasurementsandresultsforgroundwatersamplesandSaloumRiversamples(FL1,FL2andFL3),fordryandrainy(wet)seasonscampaigns,andgroundwaterisotopicdataforthewetseason.

Parameters EC T pH Ca2+ _Mg2+ _Na+ _K+ _HCO

3− Cl− SO42− NO3− ␦18O ␦2H

Units (␮S/cm) ◦_C _mg/l _‰

ID Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet Wet

S1 688 551 30.3 29.9 5 7.4 22.4 30 8.5 8.8 58.9 55.6 3.6 3.4 9.8 19.5 120.1 106.1 0.9 1.9 59.2 67.3 −4.5 −32.5 S2 612 360 30 29 5.6 6.8 43.9 37 4.9 4.4 38.9 19.3 5.8 2.9 65.9 25.6 76.4 58.5 8.3 1.8 65.5 51.4 S3 493 453 29.5 30.2 6.7 7.4 58.3 70 2.6 3.4 11.5 10.6 1.9 3.2 177.6 188.8 25 26.7 2.6 3.4 1.7 18.5 S4 369 289 29 30.8 6.2 7 39.1 41 2.9 2.9 15 15.7 0.7 0.5 131.4 126.8 23.5 22.8 1.4 1.3 5.9 5.2 S5 388 317 29.4 30.8 6.6 7.6 44 50.3 1.7 1.6 12.9 13.9 2.6 2 149.8 155.6 15.9 17 3.7 3.7 2.2 4.8 S6 185 225 29 29.2 6 7.9 17.8 30.3 1 1.7 12.9 13.7 1.1 2 58.5 106 8.6 9.4 1.1 1.8 9.8 6.7 S17 191 231 28 29.8 5.5 7.2 16.5 27 1.1 1.6 13.7 16.7 0.5 1 36.6 59.7 11.1 14.3 0.2 0.6 28.2 43.9 −5.1 −35.2 S18 570 500 30 31.1 6.6 7.5 81.3 86.3 5 5.3 14.2 14.3 1.8 1.9 300.2 298 9.4 11.9 0 0.6 0 3.4 −5.2 −35.9 S19 341 313 30 30.6 6.5 7.6 50.6 59.4 0.9 0.8 5.3 5.9 4.1 3.7 163 183.2 5.7 7.6 1.3 1.4 0.8 2.2 S20 900 748 29 30.1 7 8.1 116.4 109 9.1 9.3 32.9 30.7 1.1 1 332.9 314.9 79.8 72.2 3.5 3.8 0 3.8 −5.1 −34.7 S21 268 194 29.1 29.7 6.5 7.4 34.3 30.7 1.4 1.8 8.1 7.7 0.9 0.9 101 81.2 8.6 8.1 1.4 0.9 14.1 22.3 S22 980 781 29 29.6 5.5 5.9 51.2 52.1 6.2 7.4 81.1 83.8 1.4 1.3 65.9 64.7 185.4 180.4 10.6 11.4 5 7 −5.1 −35 S23 1040 890 29 29.2 6.2 7.5 109.1 106 11.1 12.1 39.6 40.4 7.9 8.4 162 130.4 86.7 94 3.5 3 176 196 S24 669 575 29.5 29.2 6.5 7.7 72.4 75.2 7.1 6.7 32.7 33.3 1.4 2 227.5 219 58.9 64.5 3.8 5.4 9.4 9.3 S25 296 231 29 33 5.4 7.4 12.6 13.3 1.9 1.8 26.1 25.8 0.5 0.4 17.1 22 25.9 23.6 0.5 0.7 57.5 57.8 S26 391 333 29.2 30.4 6 7.5 39.6 44.8 8.1 8.9 9.8 10.5 3.5 3.8 152.4 156 15.5 19.3 2.1 2.2 9.7 13.5 −5.1 −34.9 S27 294 224 29 31 5.9 7.9 19.3 18.8 2.8 2.8 20.7 20.1 0.5 0.5 46.3 46.3 30.9 31.5 0.9 0.7 22.9 23.2 S28 1160 908 29.2 30.1 7 7.4 113.3 102 17.2 15.2 41.9 39.3 9.3 11.3 194 169.3 191 178.1 0 2.7 0 4.5 S29 490 492 29.8 30.1 6.4 7.3 59.2 71.7 6.1 6.1 13.7 18.8 1.9 5 181.4 222.2 23.5 34.5 2.3 1.8 18.8 10.1 S30 138 98 29.2 31.1 5.5 8 8.4 8.3 3.2 3.4 6.3 6.3 0.5 0.6 26.9 29.3 7.1 8 9.6 10.8 1 1.1 −5.2 −35.5 S32 125 101 29 28.9 5.3 7.8 5.1 5.1 1.1 1.1 14.1 13.3 0.8 1.1 15.9 14.7 6.6 6.3 0.9 0.9 25.3 28 −5.2 −36.7

Parameters EC T pH Ca2+ Mg2+ _Na+ _K+ _HCO

3− Cl− SO42− NO3− ␦18O ␦2H

Units (␮S/cm) ◦_C _mg/l _‰

S33 362 308 30.2 30.3 6 7.6 39 43.5 4 4.3 14 14.9 0.8 0.6 104.8 104.8 23.8 26 1.1 1.5 24.6 29

(10)

172 N.M. Dieng et al. / Journal of Hydrology: Regional Studies 9 (2017) 163–182 Table1(Continued)

Parameters EC T pH Ca2+ Mg2+ _Na+ _K+ _HCO

3− Cl− SO42− NO3− ␦18O ␦2H S35 827 995 28 28.7 5.8 8.4 28.4 38.5 7.5 11.2 97.3 130.1 4.1 6.6 24.4 22 184 250.3 10 10.9 24.5 32.2 −4.1 −31.2 S36 480 550 29 29 5.8 6.9 31.3 41.5 3.8 5.7 39 55.6 0.7 1.6 39 89 47.2 78.3 4.9 13.2 78.7 54.5 −4.1 −31.7 S37 419 197 29.2 30.6 4.1 8.8 19 25.3 6.1 2.6 27.7 10.6 2.4 2.4 1.9 42.6 62.1 23.9 0 1.2 66.3 24 −4.2 −31.3 S38 730 760 29 30.1 6.1 8 58 74.8 4.7 6 58.4 62.8 0.9 1 87.7 126.7 116.3 139.9 1.6 3.1 39.2 40.1 S39 253 217 30 29.4 5.7 8.6 15 17.9 1.9 2 20.8 22.6 0.5 0.4 48.8 50 23.9 23.3 0.6 0.9 25.2 30.8 S40 184 175 30.5 33.5 4.9 8.4 6.9 10 1.9 2.4 11.9 14.5 6.1 7.4 7.4 11 26 29.1 0.5 1 24.1 28.3 S41 280 286 29.4 29 4.7 8 9.3 18.1 1.5 2.7 23.7 29.4 0.3 0.4 3.9 12.3 31.6 36.5 0.3 1 49.5 68.4 −5 −33.9 S42 528 1583 27.7 26.4 5.2 8.1 28 124.4 6.3 22.5 33.8 138.5 5.6 6.5 13.5 60.9 108.2 390.2 2.3 8.8 12.7 90.3 −4.5 −31.1 S43 216 278 28.7 30 5.4 8.6 11.6 20.2 1.2 2.4 17.3 29.5 0.3 0.3 22 22 23.6 46.3 0.3 1 27.8 48.7 S44 355 332 29.8 29.4 4.9 7.4 14.7 18.5 3 3.9 28.3 34.5 0.9 1.1 14.8 13.5 32 39.1 0 0.6 71.9 86.2 S45 185 147 29.3 29.3 5.7 7.9 12.4 11.4 0.7 0.7 16 15.3 0.5 0.6 36.6 36.6 13.6 13.4 0.7 0.6 17.3 18 S46 342 286 30.1 29.8 4.4 7.2 6.1 7 2.8 3.3 35 37.5 1.4 1.4 5.2 3.8 30.2 34 0 0.6 72.8 80.8 S47 74 58 30.9 31.1 5.2 7.6 1.5 6 0.2 0.3 6.2 4.4 0.6 0.5 12.3 14.7 2.9 3.6 0.6 0.6 7.9 15.1 −5.1 −33.7 S48 3430 239 27.7 28.8 6.7 7.9 260.4 30.9 38.3 0.4 324.3 13.8 3.7 0.1 323.4 34.1 870 47.4 34 3.5 0 0.5 S49 2191 2040 28.9 27.9 3.7 4.6 124.8 213.7 37.5 52.9 91.9 110.6 58.8 45.8 16.3 7.6 173.5 214.9 426.8 583 203.8 200 −4.7 −32.1 S50 111 125 29.6 29.8 5.4 7.5 6.7 14.6 0.5 0.6 7.9 9 0.2 0.1 20.8 48.8 11.5 11.5 0.9 1.1 0.9 0.2 −5.4 −34.9 S51 81 47 30 28.3 5.3 7.8 3.2 2.1 0.2 0.4 6.8 7.3 0.2 0.2 13.5 9.8 2.6 2.8 0.2 0.1 12.1 12.3 −5.2 −34.2 S52 1140 995 28.2 28.7 3.8 5.9 49.2 53.8 14.1 16.6 98.1 90 2.3 4 13.6 6.5 110.1 104.9 1.2 1.6 296.7 328 S53 176 98 29.5 29.3 5.1 7.8 7.2 9.8 1.1 0.9 9.2 8 0.9 0.7 22 24.4 9.2 7.2 1 0.9 12.1 15.7 Parameters EC T pH Ca2+ _Mg2+ _Na+ _K+ _HCO3 _Cl− _SO₄2− _NO₃− _␦18_O _␦2_H Units (␮S/cm) ◦_C _mg/l _‰

S62 407 388 29.6 29.6 5.2 6.1 23.3 31.9 5.5 6.8 25.2 26.5 3.6 3.6 25.6 40.2 32.8 34 0.8 1 86 93.3 S64 75 54 29.1 31.3 4.8 6.2 5.8 5.7 0.8 0.8 4 3.9 0.2 0.2 28.2 27 3.6 3.5 0.1 0.1 0.4 0.4 −5.4 −36 Min 75 47 27.7 26.4 3.7 4.6 1.5 2.1 0.2 0.3 4 3.9 0.2 0.1 1.9 3.8 2.6 2.8 0 0.1 0 0.2 −5.4 −36.7 Max 3430 2040 30.9 33 7 8.8 260.4 213.7 38.3 52.9 324.3 138.5 58.8 45.8 332.9 314.9 870 390.2 426.8 583 296.7 328 −4.1 −31.1 Avg 554.18 465.8 29.3 29.8 5.63 7.49 41.49 45.73 5.7795 6.339 36.28 33.6 3.609 3.998 81.03 80.29 70.19 63.11 12.54 16.3 40.35 49.8 −4.1 −31.1 FL1 64700 50800 27 28.9 7.8 8.1 690.7 424.7 2010.7 1226 17455 10389 718.6 383.7 174.8 142.8 29641 16918 4155 2435 0 0 −0.1 −5.4 FL2 91700 60400 30 31.3 7.8 8.3 956 570.1 3679.3 1510 30053 12707 1086 399.8 206.4 151.1 52933 20857 7458 2984 0 0 0.3 −5.3 FL3 89000 55200 28 29.5 7.7 8.4 1162 574.2 4036.6 1241 32652 11443 1195 481.9 215.8 147.1 60183 18518 8540 2840 0 0 0.3 −3.2

(11)

N.M.Diengetal./JournalofHydrology:RegionalStudies9(2017)163–182 173 Table2

Physico-chemicalmeasurementsandresultsforseawatersample(sw)andmeanvaluesofconcentrationsfromrainwater((K,N,F,D,S,TandW)analyses.

Parameters CE pH Ca2+ _Mg2+ _Na+ _K+ _HCO 3− Cl− SO42− NO3− Units (␮S/cm) (/) mg/l SW 39250 7.7 338.3 952 7998.9 356.2 149.1 13014 1911.7 0 K 105 6.6 3.5 0.4 2.7 1 19 6.7 2.3 5.1 N 44 6.7 2.7 0.4 0.7 0.2 14 2.3 2.8 3.1 F 17 5.2 0.7 0.1 0.7 0.3 2.2 1.7 2.3 0.8 D 27.2 5.3 1.1 0.2 0.6 1.1 2.9 1.6 4.9 1.2 S 39.8 5 2.3 0.3 0.8 5 3.7 1.5 5.9 2.7 T 24.6 6.2 2.6 0.1 0.4 0.3 8.2 1.3 2.2 1.2 W 34.4 6.7 0.7 0.2 0.5 1.8 19 1 0.7 0.9

Fig.7.BoxplotofmajorionscompositioninCTgroundwaterandcomparisonbetweendryandwetseasonsof2012.

ThreedistinctgroupswereobtainedwiththeSOMsmethod,whichcorrespondtodifferentgroundwatertypeswith

respecttothecontributionofeachchemicalparameter(Fig.8b).Thephysico-chemicalandchemicalcharacteristicsofthese

3groupswereevaluatedusingadescriptivestatisticapproachinwhichtheminimum,maximum,andaveragemedianof theECandmajorionswerecomputedforeachcampaign(Table2).Inadditiontothestatisticalapproach,aPipertrilinear diagram(Piper,1944)thatdescribesthewatertypes(Fig.9),bivariatediagrams(Fig.11aandb)andcomputedSaturation Index(SI)values(Fig.11c)wereusedtofurtherinferthevariousprocessesthatcontrolthechemicalcompositionsofthe groundwatergroups.

TheU-matrix(Fig.8b),Piperdiagram(Fig.9)andTable3showthat:

a)Group1,whichisrepresentedbysamplesS1,S2,S22,S23,S34,S35,S36,S37,S38,S42,S49,S52andS62,ischaracterized byhighvaluesofEC,Cl−,Na+_,_NO

3−andSO4−andrelativelylowvaluesofHCO3−.ItisgenerallyencounteredneartheSaloum RiverandinonesampleneartheNemaRiver(Fig.10).TheECvaluesrangefrom491to2191duringthedryseasonwith anaverageof872␮S/cm.Thewatertypes(Fig.9)aredominantlyNa–Cl(42%)andCa–Cl(42%),andtwosamples(S23and S49)exhibitingCa–HCO3(8%)andCa/Na–SO4(8%)facies,respectively.Thisgroupincludessalinewater(S22,S28,S35and S42)thatischaracterizedbyhighCl−contents(oftenabove150mg/L)aswellaspollutedwater(S23,S34,S49andS52)with highNO3−contents(above100mg/L).Itindicatescontaminationofthegroundwaterandrecentinfiltration(Currelletal., 2010).Salinegroundwater,seawaterandtheSaloumRiverhavetypicalcommoncharacteristicsofNa-Cltypewater(Fig.9) andmolarratiosclosetoseawatervalue(0.86)(Fig.11a).Thisfactconfirmsthemarineoriginoftheions(Jonesetal.,1999; Vengoshetal.,1999;Fayeetal.,2001,2003,2004;Vengosh,2013).However,somesamplesexhibitingarelativedeficiencyin Na+_relative_to_Cl_probably_reflect_reverse_ion_exchange_reaction._In_this_case_Na+_is_taken_up_by_the_clay_exchanger_and_Ca2+_is releasedtothesolution.MagaritzandLuzier(1985)statedthatgroundwatersamplescontaininglessthan15%seawaterare

(12)

Fig.8.a)SOMsindividualcomponentsmapsofthechemicalparametersintheCTaquiferduringthe2012dryseasonandb)SOMsU-matrixandclustering

in3groups.

enrichedinNa+,andasthepercentageofseawaterincreases.Infact,thewaterbecomesincreasinglyenrichedinCa2+_and Na+_is_depleted._This_process_leads_to_groundwater_facies_evolving_from_Na-Cl_to_Ca-Cl₍_Mercado,_1985;_Appelo_and_Postma, 2005).Inaddition,thesampledpointsarealignedalongaslopeof₋₁inthe(Ca+Mg)–(SO4–HCO3)versus(Na+K)-(Cl-NO3) diagram(Fig.11b).Thisdiagramexcludeshalite,calciteanddolomitedissolution(JankowskiandAcworth,1997;Garcia etal.,2001;Abidetal.,2011)andevidencedionexchangereactionasadominantgeochemicalprocess..

ThecorrelationbetweenCa2+_and_SO42−₍_Fig.₈_a)_and_the_SI_value_(-0.97)_with_respect_to_gypsum₍_Fig.₁₁_c)_in_the_highly mineralizedwatersfromwellS49(ofCa-SO4watertype)showadissolutionofgypsum.Thisisobservedinmangrovesoil, andwouldreleaseCa2+_and_SO

42−intothegroundwater(Eq.(1)).Otherwise,consideringthelowpH(3–5),theenrichment ofSO42−mayalsoresultfrompyriteoxidation(Eq.(2)),usuallyoccuringinmangrovedegradationsoils(LebigreandMarius, 1984;Kristensenetal.,1991) CaSO4⇔Ca2++SO2−4 (1) FeS2+ 7 2O2+H2O⇒CO2+⇔CH2CO3⇔F 2+ e +2S024−+2H+ (2)

b)Group2isrepresentedbywellsamplesS3,S4,S5,S18,S19,S20,S21,S24,S26,S28,S29andS33,locatedintheeastern partoftheregion,andbysampleS48(Fig.10),locatedinthecoastalstrip.SampleS48characteristicsdifferfromother samplesingroup2inthatitexhibitshighvaluesofCl−(870mg/L)andNa+_(324.3_mg/L)_and_is_a_Na-Cl_facies._Its_EC_values varyseasonally(3430and239␮S/cmduringthedryandrainyseasons,respectively).Itissuggestingthatthiswelltapsthe

(13)

Fig.9.Piper’sdiagramshowingtherelativechemicalcompositionsofgroundwater,SaloumRiverwaterandseawatersamplescollectedduringthe2012

dryseason.GroundwatersamplesbelongingtothethreegroupsdeﬁnedwiththeSOMsmethodareplottedwithdifferentcolors.

Table3

SummaryofstatisticalhydrochemicaldataofﬁeldcampaignscorrespondingtodryandwetseasonsforgroundwatergroupsobtainedwiththeSOMs

analysis. Parameters CE T pH Ca2+ _Mg2 _Na+ _K+ _HCO 3- Cl- SO42- NO3 -␮S/cm) ◦_C _mg/l Dry 1 min 419 28.2 3.7 19 3 25 0.6 1.9 32.7 0.8 5 max 2191 30.3 6.2 124 14 98.1 58.7 161.9 187.3 426 296.7 median 730 29 5.5 43.9 6.3 58.4 3.6 25.6 110.1 3.5 66.3 average 872 29.1 5.5 50.8 9.7 62.1 8.6 48.0 110.7 39.5 100.7 2 min 268 27.7 6.2 39.1 0.9 5.2 0.7 100 5.7 0.1 0.1 max 1160 30 7 116.4 17 41 9.3 332.9 191 3.8 24 median 490 29.5 6.5 54.5 4.5 13.9 1.8 170.3 23.5 1.8 4 average 533.4 29.5 6.5 62.3 5.5 17.7 2.5 184.7 40 1.9 7.8 3 min 74.4 28 4.4 3.2 0.2 3.9 0.1 3.8 2.9 0.1 0.3 max 342 30.9 5.9 19.3 3.2 34.9 6.1 58.5 31.6 9.5 72 median 185 29.5 5.35 8.8 1.2 13.9 0.5 21.4 12.5 0.5 23.5 average 197 29.5 5.3 10 1.5 15.6 0.9 24.3 16.7 1.0 25.9 wet 1 min 388 26.4 4.6 31.9 4.3 28.5 9 6.5 33.9 1.5 6.9 max 2040 30 8.4 213.7 52.8 138.5 45.8 169.2 390.1 582.1 314 median 890 29 7.4 53.8 11.2 62.8 4 60.8 139.9 3.131 67.3 average 948.2 29.0 7 79.3 14.8 74.3 10.0 65.7 157.8 50.9 113.7 2 min 194 29.2 7 43.5 0.7 7.6 0.6 81.1 8.1 0.5 3.1 max 748 30.8 8.1 108.7 9.2 30.7 5 314.9 72.2 5.3 29 median 325 30.1 7.6 54.9 3.8 14.1 2.0 169.6 21.1 1.8 8.0 average 395.6 30.2 7.6 59.3 4.4 15.8 2.2 179.7 26.7 2.3 10.7 3 min 47 28.3 6.2 5 0.3 3.8 0.05 9.7 3.5 0.6 3.4 max 332 31.1 8.8 30.8 3.3 35.5 7.3 59.7 46.2 5.3 22.3 median 197 30.2 7.8 13.3 1.8 14.5 0.5 24.4 23.3 0.9 24.0 average 180.2 30.2 7.8 14.2 1.8 17.0 1.0 27.5 21.3 1.4 30.7

(14)

Fig.10.Spatialdistribution,duringthe2012dryseason,ofthethreegroundwatergroupsasobtainedwiththeSOMsanalysis.

Fig.11.Relationshipbetweenthecontentsofa)Na/ClvsCl,b)[(Ca+Mg)-(SO4+HCO3)]vs[(Na+Cl)-(Cl+NO3)]andc)saturationindexvalueinthe ground-watersamplescollectedduringthe2012dryseason.ThethreedistinctgroupsarethosedeﬁnedwiththeSOMsmethod.

(15)

N.M.Diengetal./JournalofHydrology:RegionalStudies9(2017)163–182 177 Table4

Summaryofstatisticalhydrochemicaldataofﬁeldcampaignsduringthewetseasonsof2003and2012forgroundwatergroupsobtainedwiththeSOMs analysis. Parameters CE Ca2+ _Mg2 _Na+ _K+ _HCO 3- Cl- SO42- NO3 -␮S/cm) mg/l 2003 1 Mean 2099.4 133.7 37.9 278.6 16.5 82.4 513.4 64.8 131.6 Median 1769.0 140.4 29.5 202.6 8.8 61.0 386.3 12.6 59.6 Min 1013.0 98.1 6.5 114.5 1.9 30.5 174.0 0.5 2.0 Max 4920.0 189.8 106.2 791.7 61.8 158.6 1498.0 368.0 259 2 Mean 486.0 41.8 4.0 48.6 1.3 134.2 88.9 4.0 10.0 Median 358.0 46.1 3.5 21.4 1.2 134.2 23.1 0.9 7.8 Min 35.0 5.3 0.2 2.9 0.6 39.6 1.5 0.2 1.6 Max 1686.0 50.0 6.2 234.0 2.6 225.7 75.3 23.6 29.2 3 Mean 216.0 16.8 3.7 20.2 2.6 58.0 20.5 1.0 49.0 Median 172.5 11.6 1.9 14.7 0.8 47.3 19.6 0.3 26.7 Min 62.0 3.8 0.2 6.8 0.1 27.4 2.6 0.1 5.9 Max 299.0 56.5 17.6 37.1 19.3 131.1 30.8 5.0 199.6 2012 1 Mean 1250.6 96.0 20.7 90.3 19.8 46.2 194.4 132 124.9 Median 1260.0 99.6 16.8 101.3 6.5 23.8 194.2 9.9 78.8 Min 360.0 30.0 4.3 19.3 0.9 7.6 58.4 1.7 32.1 Max 2191.0 213.0 52.8 138.5 45.8 91.1 390.0 582.4 200.1 2 Mean 365.0 54.1 3.1 15.1 2.2 150.7 30.2 2.5 9.7 Median 313.0 59.4 2.9 13.8 2 183.2 26.7 1.8 9.3 Min 194.0 30.7 0.3 5.9 0.6 34.1 64.5 0.8 0.5 Max 575.0 71.7 6.7 33.3 5.0 222.1 8.1 5.0 22.3 3 Mean 170.5 11.6 1.8 17.3 1.3 27.1 19.6 0.7 29.6 Median 161.0 10.7 1.4 14.9 0.6 25.7 18.3 0.7 20.6 Min 47.0 5.0 0.3 3.8 0.1 9.7 3.5 0.9 3.4 Max 286.0 18.7 3.9 37.5 7.3 49.9 34.0 1.0 86.8

dynamicseawaterinterfacecharacterizedbyseawateradvanceduringdryseasonfollowedseawardinterfaceretreatduring

rainyseason(i.e.whengroundwaterheadsarehigher).AttheexceptionofS48,thisgroupischaracterizedbyhighCa2+_and

HCO3−contents,relativelylowCl−andNa+contentsandvariablemineralizationrangingfrom268to1160␮S/cmwithan

averageof533␮S/cmduringthedryseason.Thisgroupisfeaturedbyfreshanduncontaminatedgroundwaterthatisnearly

freeofmarinewatercontamination,andthegroundwaterfaciesaredominantlyCa–HCO3.ComputedSIvalueswithrespect

tocalcite(rangingbetween−1.7and0.7)anddolomite(rangingbetween−4and0.6)(Fig.11c),andCa/HCO3molarratios

(1)inadditiontooccurrenceofcalcitemineralinthereservoirmatrix(asevidencedbyLappartient,1985)reﬂectcarbonate mineralsdissolution.Thisreactionisthemaingeochemicalprocesswhichwouldbeconcomitanttoionexchangeinwhich Ca2+_replaces_Na+_in_the_clay_matrix₍_Fig.₁₁_a).

c)Group3,whichisrepresentedbywellsamplesS6,S17,S25,S27,S30,S32,S39,S40,S41,S43,S44,S45,S46,S47,S50, S51,S53,andS64,ischaracterizedbylowmineralizedgroundwater.TheECvaluesarerangedfrom74.4to342␮S/cmin thedryseason.Morethan70%ofthesamplesexhibitlowECvalues(below250␮S/cm)suggestingalowermineralisation andtheirwatertypesaredominantlyNa–HCO3(33%),Na–Cl(33%),Ca–HCO3(22%)andCa–Cl(12%).Thesecharacteristicsof lowmineralisationandvariouswatertypeslikelydepicttherecentlyinﬁltrationrainwaters.Theyarereachingtheaquifer accompaniedbybothrapidcation(NaforCa)andanion(ClforHCO3)exchangesintheclaymatrixoftheaquiferreservoir inadditiontocalcitedissolution.Thiswouldresultinaconsistentgroundwater-typegradationoflowmineralisedCa-Cland Na-HCO3typesevolvingtoNa-Cl,Ca-HCO3toCa-ClandNa-Clindicatingchangesalonggroundwaterﬂowpathsinthese zones(Mercado,1985;Hidalgoetal.,1995).

4.4. Interannualvariationsofgroundwatersalinitybetween2003and2012

Thetemporaltrendvariationsofgroundwatersalinitywereinvestigatedusingdatacollectedin2003and2012.Intotal, 28chemicalanalyseswereretrievedfromNovember2003(Fayeetal.,2004)andNovember2012(thiswork)asacomparison todepictchangesrelatedtoanthropogenicandnaturalstressinthisvulnerablesystem.SalinitypatternsoftheSaloumRiver waterECexhibitschangesfrom78,300to90,100␮S/cmin2003andfrom50,800to60,400␮S/cmin2012.Thesechanges muchlikelyderivefromrainfalleventswhichvariedfrom535mm(502.9mmeffectiverainfall)in2003–847mm(796.18mm effectiverainfall)in2012.Comparedwiththelong-term(1960–1990)averageof610mm,consideredasreferenceinSahel, the2003and2012rainfallrecordsrepresentrespectivelyadeclineof12%andanincreaseof38%.

Thishighvariationinrainfallbetweenthese2referenceyearswouldprobablyalsochangechemicalpatternsinthe groundwater.SOMsanalysesofthedifferentgroupsrevealed:

-InwellsS1,S2,S27,S34,S35andS38,whicharegenerallylocatedneartheSaloumRiverandcorrespondtothegroup 1intheSOMsanalysis,theECevolutionshowsadecreasingtrendwithvalueshigherin2003(averageofapproximately 2099␮S/cm)thanin2012(1250␮S/cm)(Table4).Theioniccontentschanges,showninthePiperdiagram(Fig.12),are featuredbyashiftofthegroundwatertypesfromNa-CltoCa-ClorCa-HCO3.ThisconﬁrmstheenrichmentofCarelative

(16)

Fig.12.Comparisonofboxplotdiagramsasobtainedfora)majorionscompositionandb)electricalconductivity(EC)inCTgroundwaterbetweenwet

seasonsof2003and2012.

toNa,andthereforethesofteningbydilutioninferredbyinﬂowsofbothinﬁltratedrainfallandSaloumRivercomparedto 2003(seeabove).

-InwellsS3,S19,S21,S22,S23,andS62,locatedinthecentralandeasternzonesmostly,andcorrespondingtogroup 2intheSOMsanalysis,theECvaluesincreasedfrom2003(meanvalueof353␮S/cm)to2012(averageofapproximately 587␮S/cm).ThisisalsoobservedintheioniccontentsofthegroundwatersamplesasconﬁrmedbythePiperdiagram(Fig.13). GroundwatertypesmostlyevolvedfromCa-HCO3toNa-ClorCa-Cl.Thesecharacteristicsofincreasedmineralisationand enrichmentofHCO3-,Cl−and/orNO3−insomewellslikelydepicttherecentlyinﬁltrationrainwaters.Itisreachingthe aquiferaccompaniedbybothcombinedcalcitedissolutionandpollutantsfromanthropiccontamination.

-NearthecoastandtheNemaRiver,notrendcanbedistinguished.TheECvaluesdecreasedinsomeofthewells(e.g., S40,S41,S48,andS50)(from1750toapproximately200␮S/cm)andincreasedinothers(e.g.,S42,S49,andS52)(from 974toapproximately1539␮S/cm).Inthisregion,thechemicalevolutionofthegroundwaterwasduetoacombinationof seawaterintrusionthatwasenhancedbypumpingactivitiesanddilutionbyrecharge.

4.5. Isotopecomposition

Often,isotopicresultsgreatlyhelpforunderstandingthestructuralcontrolsongroundwaterﬂowinaridregions(e.g. Daileyetal.,2015)andtherechargepatterns(amongmanyothers:ShahulHameedetal.,2015).Here,theresultsoftheisotope analysesofthesurfacewaterandgroundwatersamplesarepresentedinTable1.Isotopicanalyseswereonlyperformed onasubsetofsamplesthatwerecollectedinNovember2012,whichcorrespondstotheendoftherainyseason.The␦18_O and␦2_H_values_of_the_groundwater_samples_and_the_Saloum_River_water_are_plotted_against_the_Global_Meteoric_Water Line(GMWL)andtheLocalMeteoricWaterLine(LMWL)ofequation(␦2_H₌_7.93_␦18_O₊_10.09)_deﬁned_by_Travi_et_al.₍₁₉₈₇₎ (Fig.14a).

TheSaloumRiversamplesaredistinctlyenrichedwithvaluesrangingfrom−0.11to+0.32‰for␦18_O_and_from_−5.4_to −3.2‰for␦2_H._These_samples_deviate_{signiﬁcantly}_from_the_SMOW_values₍_Fig.₁₄_a)_conﬁrming_prominent_evaporation processesaffectthelowtostagnantSaloumRiverwater.Inthegroundwatersamples,the␦18_O_values_range_from_−5.4_to −4.1‰,andthe␦2_H_values_range_from₋₃₆_to_−31‰,_with_mean_values_of_−4.9‰_and_−33‰,_{respectively.}_Plotted_in_the ␦2_H_vs._␦18_O_diagram,_the_data_cluster_below_the_LWML_along_the_line_␦2_H₌_3.9_␦18_O₋_14.3_(R2₌_0.83)._The_samples_plot_to therightoftheglobalmeteoricindicatesanenrichmentofheavyisotopes,andsuggeststhattheyweresubjectedtovariable degreesofevaporation

Indeed, in the Saloum environment, where temperatures (mean 28,5◦C) are high as well as evapotranspiration (2030mm/year),itislikelythatrechargewaterfromrainfallevaporatesbeforeand/orduringafterinﬁltration.

Twodistinctclusterscanbedistinguishedinthegroundwatersamplesusingtheisotopicdata(Fig.14a).InclusterA,the stableisotopevaluesrangefrom−5to−4.1%for␦18_O_and_from₋₃₅_to_−31%_for_␦2_H._This_cluster_represents_samples_of_group

(17)

Fig.13. Piperdiagramshowinggroundwatercompositionchangebetweenwetseasonsof2003and2012.

1,whicharelocatedalongthecoastandneartheSaloumRiver.ClusterBischaracterizedbymoredepletedstableisotope valueswith␦18_O_ranging_from_−5.4_to_−5.1%_and_␦2H_ranging_from₋₃₆_to_−34%._This_cluster_represents_fresh_groundwater samplesofgroups2and3.ThegroundwaterofclusterAappearstobemoreaffectedbyevaporationorismoreinfluencedby theintrusionofevaporatedsurfacewater.Bycomparing␦18_O_and_Cl−₍_Fig.₁₄_b),_the_existence_of_the_two_previous_clusters isconfirmed.ThesamplesofclusterAarelocatedonamixinglinethatjoinsthesamplesofclusterB,whichrepresentfresh water,andthesamplesfromtheSaloumRiverandtheocean.ThisconfirmsthatthesalinizationofthesamplesofclusterA iscausedbymixingwithcoastalmarineorSaloumRiverwaters.

TheresultssuggestedthattheCTaquiferintheinvestigatedareashavebeenrechargedby

rainwaterinfiltrationandmixedwithsalinesurfacewaterfromtheSaloumRiver,whichisalsoaffectedbyhigh evapo-rationprocesses.Fayeetal.(2004)hypothesizedfiveprocessesmayleadtothisscatteredpattern:1)mixingprocessalong flowpathscombinedwithevaporationfromtheshallowwatertable;bothcanhomogenisethe␦18_O_values_and_increase theoverallsalinity;(2)theisotopiccontentsofrechargedwatersandtheSaloumRiverwatercanhaveawiderangeand alsohomogenizationof␦18_O_values;₍₃₎_{transpiration}_from_vegetation_cover;₍₄₎_dissolution_of_the_salt_minerals_may_lead_to changingboththeSaloumRiverhydrochemistryaswellastheinterstitialwaterchemistryincreasingthereforesalinityin thetransitionzone;(5)toalesserextent,thesaltsecretinghalophytesplantsofthemangrovesystemlocatedatthemouth oftheestuarymayleadtotheenrichmentofsalinityintheunsaturatedzonewithoutisotopicfractionation.

AsshowninFig.14c,theSaloumRiverwasnotablydepletedinheavyisotopesinNovember2012incomparisonto November2003,whichvaluesrangefrom1.5to1.7‰for␦18_O_and_from₀_to_3‰_for_␦2_H._Otherwise,_the_evaporated_trend ofthe2003groundwatersamplesdeviatesmoresigniﬁcantly(slopeof3.9)fromtheLMWLthanthoseofthe2012samples (slopeof4.9)(Fig.14c).TheseresultssuggestthedilutionoftheSaloumRiverwaterandthefresheningofgroundwaternear theriverduetotheincreasingamountoffreshwaterfromrainfallbetween2003and2012(Fig.14c).

5. Conclusions

TheCTaquiferrepresentsanimportantgroundwaterresourceinSenegalandparticularlyintheSaloumregion,whereit isthemainsourceofdrinkingwaterforthelocalpopulation.Therefore,thereassessmentofprocessesresponsibleforthe observedmineralizationanddegradedwaterqualityandofthegroundwaterqualitychangesoftheCTaquiferisofgreat importance.Thisreassessmentmustbedoneinthelightofthenewconstraints(e.g.changesinclimaticconditionsand pumpingconstraints),andthenewﬂowpatternconﬁgurationinthissystem.Thesimultaneousanalysisofhydrogeochemical and isotopicdatausingconventionalandstatisticalmethods allowstheassessmentofthegroundwaterquality inthe

(18)

Fig.14.Relationshipbetweena)␦2_H_and_␦18_O_and_b)_␦18_O_and_Cl_in_the_CT_groundwater_during_the₂₀₁₂_wet_season_and_c)_␦2_H_and_␦18_O_in_the_CT

groundwaterduringthewetseasonsof2003and2012.ThegroundwatersamplesareclassiﬁedusingthethreegroupsdeﬁnedwiththeSOM’smethod.

southernpartoftheSaloumBasin.Statisticalanalysesofthegeochemicaldatademonstratethespatialvariabilityofthe chemicalcharacteristicsofthegroundwater.Theyallowtodistinguishthreegroupsofgroundwater.Theﬁrstgroup(1),which islocatednearthecoastandneartheSaloumRiver,isaffectedbymixingwithsalinewaterintrusionand/oranthropogenic pollution,andthemainfaciesareNa-ClorCa-Cltypes.Thesecondgroup(2),whichismainlylocatedinthecentralandeastern partsofthestudyarea,islessmineralizedandcontainsCa-HCO3faciesasaresultofcalcitedissolutionduetointeraction withtheaquifer.Inthethirdintermediarygroup(3),somesamplesindicatedfresheningprocesseslikelydepictingtheeffect ofrecentinﬁltrationrainwaters.

Thesethreegroupsandtheirchemicalcharacteristicsaregenerallyobservedinboththedryandwetseasons.However, adecreaseofthemineralizationisgenerallyobservedduringthewetseasonduetotherechargeoflessmineralizedwater. Acomparisonofthegroundwaterqualitycharacteristicsintermsofsalinizationbetween2003and2012indicatesthat signiﬁcantchangeshaveoccurred.InthewellsthatarelocatedneartheSaloumRiver,decreasesoftheNa+andCl- concen-trationsandtheECvaluesareobserved.ThegroundwaterfacieshavegenerallyevolvedfromNa-CltoCa-ClorCa-HCO3, whichindicatesthedilutionofgroundwaterduetofresheninginthevicinityoftheSaloumRiver.Thisfresheningisdueto anincreaseofrainfallaswellastheintrusionoflessmineralizedwaterinrecentyearsasthesalinityoftheSaloumRiver hasalsodecreased.Anincreaseofmineralizationisobservedinsomewellslocatedinthecentralandeasternzonesofthe studyarea.GroundwatertypesmostlyevolvedfromCa-HCO3toNa-ClorCa-Clandindicaterecentinﬁltrationrainwaters reachingtheaquiferaccompaniedbybothcombinedcalcitedissolutionandpollutantsfromanthropiccontamination.No trendcanbedistinguishednearthecoastandtheNemaRiverwherethechemicalevolutionofthegroundwaterwasdueto acombinationofseawaterintrusionanddilutionbyrecharge.

Insummary,changesofgroundwaterlevelsandqualityhavebeenobservedintheCTaquiferbetween2003and2012 inresponsetovariationsintheprecipitationregimeandotherfactors.Additionalchangesinrechargeorintheﬂowregime willoccurinthenearfutureinresponsetoclimatechangeortoincreasesofgroundwaterpumping.Itisthusimportant toidentifytheevolutionofthegroundwaterqualitythatiscausedbythesenewchangestoavoidormitigatethefuture degradationofgroundwaterqualitybeforeitbecomesamajorobstacletothefutureeconomicdevelopmentofthiscoastal area.

(19)

Acknowledgments

TheauthorsthanktheInternationalFoundationofSciences,FrancophoneUniversityAgencyandtheANGLEprojectfor ﬁnancialsupporttocarryoutthiswork.TheyalsothanktheHydrogeologyLaboratoryoftheUniversityofLiège<(Belgium), theDepartmentofGeologyofCheikhAntaDiopUniversityinDakar(Senegal)andtheHelmholtzZentrumMünchen,Institute ofGroundwaterEcology(Neuherberg/Germany).Theauthorsaregratefultothetwoanonymousreviewerswhohelpeda lotforimprovingthepaper.

AppendixA. Supplementarydata

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ejrh.2016.12.082.

References

Abid,K.,Zouari,K.,Dulinski,M.,Chkir,N.,Abidi,B.,2011.HydrologicandgeologicfactorscontrollinggroundwatergeochemistryintheTuronianaquifer

(SouthernTunisia).Hydrogeol.J.19(2),415–427,http://dx.doi.org/10.1007/s10040-010-0668-z.

Allen,R.,Pereira,L.,Raes,D.,1998.CropEvapotranspiration–GuidelinesforComputingCropWaterRequirements–FAOIrrigationandDrainagePaper

56.FoodandAgricultureOrganizationoftheUnitedNations(FAO).

Appelo,C.A.J.,Postma,D.,2005.Geochemistry,GroundwaterandPollution,seconded.BalkemaRotterdam(656p).

Aranyossy,J.F.,Guere,A.,Sidoro,M.,1989.Étudeparlesisotopesdel’environnementdesdépressionspiézométriques;premièresdonnéessurdes

exemplesauMali.Hydrogéologie3,151–158.

Archambault,J.,1960.LesEauxSouterrainesd’AfriqueOccidentale.BergerLevrault,Nancy,139p.

Barusseau,J.P.,Diop,E.H.S.,Saos,J.L.,1985.Evidenceofdynamicsreversalintropicalestuaries,geomorphologicalandsedimentologicalconsequences

(Saloumandcasamancerivers,Senegal).Sedimentology32(4),543–552.

Belkhiri,L.,Mouni,L.,Boudoukha,A.,2012.Geochemicalevolutionofgroundwaterinanalluvialaquifer:caseofElEulmaaquifer,EastAlgeria.J.Afr.

Earth.Sci.66–67,46–55,http://dx.doi.org/10.1016/j.jafrearsci.2012.03.001.

Bellion,Y.J.C.,1987.HistoiregéodynamiquePost-paléozoïqueDel’AfriqueDel’Ouestd’aprèsl’étudeDeQuelquesBassinssédimentaires(Sénégal,

Taoudéni,Iullemmeden,Tchad).ThèseSciences.Univ.Avignon,296p.

Conrad,G.,Lappartient,J.R.,1987.TheContinentalterminal;itspositionwithintheCenozoicgeodynamicevolutionoftheSenegalo-mauritanianbasin.J.

Afr.EarthSci.6(1),45–60.

Coplen,T.B.,1988.Normalizationofoxygenandhydrogenisotopedata.Chem.Geol.72,293–297.

Currell,M.J.,Cartwright,I.,Bradley,D.C.,Han,D.,2010.RechargehistoryandcontrolsongroundwaterqualityintheYunchengBasin,NorthChina.J.

Hydrol.385,216–229,http://dx.doi.org/10.1016/j.jhydrol.2010.02.022.

Dailey,D.,Sauck,W.,Sultan,M.,Milewski,A.,Ahmed,M.,Laton,W.R.,Elkadiri,R.,Foster,J.,Schmidt,C.,AlHarbi,T.,2015.Geophysical,remotesensing,

GIS,andisotopicapplicationsforabetterunderstandingofthestructuralcontrolsongroundwaterﬂowintheMojaveDesert,California.J.Hydrol.: Reg.Stud.3,211–232.

Degallier,R.,1962.Hydrogéologiqueduferloseptentrional(Sénégal).MémoireBRGM19,44p.

Dieng,N.M.,Faye,S.,Dinis,J.,Gonc¸alves,M.,Caetano,M.,2014.CombinedusesofsupervisedclassiﬁcationandNormalizedDifferenceVegetationIndex

techniquestomonitorlanddegradationintheSaloumsalineestuarysystem.In:Diop,S.,etal.(Eds.),TheLand/OceanInteractionsintheCoastalZone ofWestandCentralAfrica,EstuariesoftheWorld.,10.1007/978-3-319-06388-12.

Diluca,C.,1976.HydrogeologyoftheContinentalTerminalAquiferBetweentheSineandtheGambia.TechnicalReport.BRGMDakar,33p.

Diop,E.S.,1986.TropicalHoloceneEstuaries.ComparativeStudyofthePhysicalGeographyFeaturesoftheRiversfromtheSouthofSaloumtothe

Mellcorée(GuineaRepublic).(PhD.Thesis).Univ.LouisPasteur,Strasbourg,France.

Edmunds,W.M.,Gaye,C.B.,1994.EstimatingthespatialvariabilityofgroundwaterrechargeintheSahelusingchloride.J.Hydrol.156(1–4),47–59,

http://dx.doi.org/10.1016/0022-1694(94)90070-1.

Favreau,G.,Leduc,C.,Marlin,C.,Guéro,A.,2002.UnedépressionpiézométriquenaturelleenhausseauSahel(Sud-OuestduNiger).C.R.Geosci.334(6),

395–401,http://dx.doi.org/10.1016/S1631-0713(02)01763-7.

Faye,S.,Evans,D.,CisséFaye,S.,2001.OriginanddistributionofsalinegroundwaterintheSaloum(Senegal)coastalaquifer,proceed.In:Second

InternationalConferenceonSaltwaterIntrusionandCoastalAquifers,Mérida,Yucatán,México.

Faye,S.,CisséFaye,S.,Ndoye,S.,Faye,A.,2003.HydrogeochemistryoftheSaloum(Senegal)superﬁcialcoastalaquifer.Environ.Geol.44(2),127–136,

http://dx.doi.org/10.1007/s00254-002-0749-y.

Faye,S.,Maloszewski,P.,Stichler,W.,Trimborn,P.,CisséFaye,S.,Gaye,C.B.,2004.GroundwatersalinizationintheSaloum(Senegal)deltaaquifer:minor

elementsandisotopicindicators.J.Sci.Total.Environ.,17,http://dx.doi.org/10.1016/j.scitotenv.2004.10.00.

Faye,S.,Diaw,M.,Ndoye,S.,Malou,R.,Faye,A.,2009.Impactsofclimatechangeongroundwaterrechargeandsalinizationofgroundwaterresourcesin

Senegal.InProc.GroundwaterandClimateinAfrica,Kampala,IAHSPubl,334:163–173.

Faye,S.,Ba,M.S.,Diaw,M.,Ndoye,S.,2010.ThegroundwatergeochemistryoftheSaloumdeltaaquifer:importanceofsilicateweathering,rechargeand

mixingprocesses.J.Afr.Environ.Sci.Tech.4(12),815–830,http://dx.doi.org/10.5897/AJEST09.180.

Ford,M.,Tellam,J.H.,Hughes,M.,1992.Pollution-relatedacidiﬁcationintheurbanaquifer,Birmingham,UK.J.Hydrol.140(1-4),297–312,

http://dx.doi.org/10.1016/0022-1694(92)90245-Q.

Güler,C.,Thyne,G.D.,McCray,J.E.,Turner,A.K.,2002.Evaluationofgraphicalandmultivariatestatisticalmethodsforclassiﬁcationofwaterchemistry

data.Hydrogeol.J.10(4),455–474,http://dx.doi.org/10.1007/s10040-002-0196-6.

Gamble,A.,Babbar-Sebens,M.,2012.Ontheuseofmultivariatestatisticalmethodsforcombiningin-streammonitoringdataandspatialanalysisto

characterizewaterqualityconditionsintheWhiteRiverBasin,Indiana.USA.Environ.Monit.Assess.184(2),845–875,

http://dx.doi.org/10.1007/s10661-011-2005-y.

Garcia,L.A.,Shigidi,A.,2006.Usingneuralnetworksforparameterestimationingroundwater.J.Hydrol.318(1–4),215–231,

http://dx.doi.org/10.1016/j.jhydrol.2005.05.028.

Garcia,M.G.,Hidalgo,M.,Blessa,M.A.,2001.GeochemistryofgroundwaterinthealluvialplainofTucumanprovince.ArgentinaHydrogeol.J.9(6),

597–610,http://dx.doi.org/10.1007/s10040-001-0166-4.

Ghesquière,O.,Walter,J.,Chesnaux,R.,Rouleau,A.,2015.ScenariosofgroundwaterchemicalevolutioninaregionoftheCanadianShieldbasedon

multivariatestatisticalanalysis.J.Hydrol.:Reg.Stud.4(B,246–266.

Gning,A.A.,2015.EtudeEtModélisationHydrogéologiqueDesInteractionsEauxDeSurface-EauxSouterrainesDansUnContexted’agricultureIrriguée.

(Ph.DThesis).Univ.Liege,285p.

Hidalgo,M.C.,Cruz-Sanjulián,J.,Sanroma,A.,1995.GeochemicalEvolutionofundergroundwatersinasedimentaryriverbasin(water-bearingof

(20)

Hong,Y.S.,Rosen,M.R.,2001.Intelligentcharacterizationanddiagnosisoftheeffectofstormwaterinﬁltrationongroundwaterqualityinafracturedrock

aquiferusingartiﬁcialneuralnetwork.UrbanWater3,193–204,http://dx.doi.org/10.1016/S1462-0758(01)00045-0.

Hong,S.,Kim,H.,Kumar,N.,Kim,J.,Park,N.,2004.Experimentalinvestigationsongroundwaterﬂowincoastalaquifers.In:GroundwaterandSaline

Intrusion.,pp.15,21p.

Ibrahima,M.,Moctar,D.,Maguette,D.N.,Diakher,M.H.,Malick,N.P.,Serigne,F.,2015.Evaluationofwaterresourcesqualityinsabodalagoldmining

regionanditssurroundingarea(Senegal).J.WaterRes.Protect.7(3),247–6363,http://dx.doi.org/10.4236/jwarp.2015.73020.

Jankowski,J.,Acworth,R.I.,1997.Impactofdebris-ﬂowdepositsonhydrogeochemicalprocessesandthedevelopmentofdrylandsalinityintheYass

RiverCatchment,NewSouthWales.Aust.Hydrogeol.J.5(4),71–88,http://dx.doi.org/10.1007/s100400050119.

Jones,B.F.,Vengosh,A.,Rosenthal,E.,Yechieli,Y.,1999.Geochemicalinvestigations.In:SeawaterIntrusioninCoastalAquifers-Concepts,Methodsand

Practices.SpringerNetherlands.14,51–71.10.1007/978-94-017-2969-73.

Kim,Y.,Lee,K.S.,Koh,D.C.,Lee,D.H.,Lee,S.G.,Park,W.B.,Koh,G.W.,Woo,N.C.,2003.Hydrogeochemicalandisotopicevidenceofgroundwater

salinizationincoastalaquifer:acasestudyinJejuvolcanicisland.J.Hydrol.270(3),282–294,http://dx.doi.org/10.1016/S0022-1694(02)00307-4.

Kohonen,T.,1982.Self-organizedformationoftopologicallycorrectfeaturemaps.Biol.Cybern43(1),http://dx.doi.org/10.1007/BF00337288.

Kohonen,T.,2001.AnOverviewofSOMLiterature.InSelf-OrganizingMapsSpringerBerlinHeidelberg.30,347–371.10.1007/978-3-642-56927-210.

Koussoube,Y.,2010.HydrogéologiedessériessédimentairesdeladépressionpiézométriqueduGondo(bassinduSourou)–BurkinaFaso/Mali.(Ph.D

thesis).ParisVI−PierreetMarieCurieUniv.285p.

Kristensen,E.,Holmer,M.,Bussarawit,N.,1991.BenthicmetabolismandsulfatereductioninaSouth-EastAsianmangroveswamp.Mar.Ecol.Prog.Ser.

73,93–103.

Lappartient,J.R.,1985.TheContinentalTerminalandtheEarlyPleistoceneoftheSenegalo-mauritanianBasin.Stratigraphy,Sedimentology,Diagenesis,

Alterations,PaleoshoreReconstitutionsfromtheFerraliticFormations,(PhDThesis).Univ.Marseille,France,294p.

Largier,J.L.,Smith,S.V.,Hollibaugh,J.T.,1997.SeasonallyhypersalineestuariesinMediterranean-climateregions.Est.Coast.ShelfSci.45(6),789–797,

http://dx.doi.org/10.1006/ecss.1997.0279.

LePriolJ.,DiengB.,1985.SynthèsehydrogéologiqueduSénégal(1984–1985).Etudegéologiquestructuraleparphoto-interprétation.Géométrieet

limitesdesaquifèressouterrains.Rap.Min.Hydraul.0l/85fMHIDEH.

Lebel,T.,Ali,A.,2009.RecenttrendsintheCentralandWesternSahelrainfallregime(1990–2007).J.Hydrol.375(1–2),52–64,

http://dx.doi.org/10.1016/j.jhydrol.2008.11.030.

Lebigre,J.M.,Marius,C.,1984.Etuded’uneséquencemangrove-tanneenmilieuéquatorial,baiedelaMondah(Gabon).PhotoAnn.Trav.Doc.Géogr.Trop.

CGET,131–145.

Ly,A.,Aglanda,R.,1991.Lebassinsénégalo-mauritaniendansl’évolutiondesmargespériatlantiqueauTertiaire.Cah.Micropaléontol.6(2),47p.

Magaritz,M.,Nadler,A.,Koyumdjiski,H.,Dan,J.,1981.TheuseofNa/Clratiostotracesolutesourcesinasemiaridzone.WaterResour.Res.17(3),

602–608.

Mercado,A.,1985.Theuseofhydrogeochemicalpatternsincarbonatesandandsandstoneaquiferstoidentifyintrusionandﬂushingofsalinewater.

GroundWater23,635–645,http://dx.doi.org/10.1111/j.1745-6584.1985.tb01512.x.

Mikhailov,V.N.,Isupova,M.V.,2008.HypersalinizationofriverestuariesinWestAfrica.WaterResour.35(4),367–385,

http://dx.doi.org/10.1134/S0097807808040015.

Montcoudiol,N.,Molson,J.,Lemieux,J.M.,2015.GroundwatergeochemistryoftheOutaouaisRegion(Québec,canada):aregional-scalestudy.Hydrogeol.

J.23(2),337–396.

Mudry,J.,1991.Discriminantanalysisanefﬁcientmeansforthevalidationgeohydrologicalhypothesis.Rev.Sci.Eau.4(1),19–37,

http://dx.doi.org/10.7202/705088ar.

NdiayeB.,AranyossyJ.F.,FayeA.,1993.Lerôledel’évaporationdanslaformationdesdépressionspiézométriquesenAfriqueSahélienne;hypothèseset

modélisation.In:LesressourceseneauauSahel.EtudeshydrogéologiquesethydrologiquesenAfriquedel’Ouestparlestechniquesisotopiques,

IAEA-TECDOC-721:53-63.

Ngom,F.D.,2000.CaractérisationDesTransfertsHydriquesDansLeBassinDeLaNemaAuSineSaloum.These3emeCycle.Univ.Dakar,130p.

NgounouNgatcha,B.,Mudry,J.,Aranyossy,J.F.,Naah,E.,Reynault,S.J.,2007.Apportdelagéologie,del’hydrogéologieetdesisotopesdel’environnement

àlaconnaissancedesnappesencreuxduGrandYaéré(NordCaméroun).Rev.Sci.Eau.20(1),29–43.

Nicolini,E.,Rogers,K.,Rakowski,D.,2016.Baselinegeochemicalcharacterisationofavulnerabletropicalkarsticaquifer;Lifou,NewCaledonia.J.Hydrol.:

Reg.Stud.5,114–130.

Noël,Y.,1975.EtudeHydrogéologiqueDuContinentalTerminalDeSineGambiePremièrePhaseEtRapportDeSynthèse.BRGMDakar,30p.

Page,J.,Citeau,J.,1990.RainfallandsalinityofaSahelianestuarybetween1927and1987.J.Hydrol.113(1–4),325–341,

http://dx.doi.org/10.1016/0022-1694(90)90182-w.

Peeters,L.,Bac¸ao,F.,Lobo,V.,Dassargues,A.,2007.Exploratorydataanalysisandclusteringofmultivariatespatialhydrogeologicaldatabymeansof

GEO3DSOM,avariantofKohonen’sSelf-OrganizingMap.Hydrol.EarthSyst.Sci.11(4),1309–1321,http://dx.doi.org/10.5194/hess-11-1309-2007.

Piper,A.M.,1944.Agraphicalprocedureinthegeochemicalinterpretationofwateranalyses.Trans.Am.Geophys.Union25(6),

http://dx.doi.org/10.1029/TR025i006p00914.

Pritchard,D.W.,1967.WhatisanEstuary:PhysicalViewpoint.InLauff,G.H.(Ed.)Estuaries.Am.Ass.Advanc.SciPub.83,3–5.

Ridd,P.V.,Stieglitz,T.,2002.Dryseasonsalinitychangesinaridestuariesfringedbymangrovesandsaltﬂats.Estuar.Coast.Shelf.Sci.54(6),1039–1049,

http://dx.doi.org/10.1006/ecss.2001.0876.

Roger,J.,Barusseau,J.P.,Castaigne,P.,Duvail,C.,Noël,B.J.,Nehlig,P.,Serrano,O.,Banton,O.,Comte,J.C.,Travi,Y.,Sarr,R.,Dabo,B.,Diagne,E.,Sagna,R.,

2009.Programmed’AppuiAuSecteurMinierCartographiegéologiqueDuBassinsédimentaire.Projet9ACPSE009.

Sarr,R.,1995.Etudebiostratigraphiqueetpaléoenvironnementaledessériesd’âgeCrétacéTerminalàEocènemoyenduSénégaloccidental.In:

SystématiqueEtMigrationDesOstracodes.(Thèsed’Etat).Univ.C.A.D.Dakar,335p.

Savenije,H.,Pagès,J.,1992.Hypersalinity:adramaticchangeinthehydrologyofSahelianestuaries.J.Hydrol.135,157–174.

ShahulHameed,A.,Resmi,T.R.,Suraj,S.,UnnikrishnanWarrier,C.,Sudheesh,M.,Deshpande,R.D.,2015.Isotopiccharacterizationandmassbalance

revealsgroundwaterrechargepatterninChaliyarRiverbasin,Kerala,India.J.Hydrol.:Reg.Stud.4(A,48–58.

Smith,A.J.,Turner,J.V.,2001.Density-dependentsurfacewater-groundwaterinteractionandnutrientdischargeintheSwan–Canningestuary.Hydrolog.

Process.15(13),2595–2616,http://dx.doi.org/10.1002/hyp.303.

Travi,Y.,Gac,J.Y.,Fontes,J.C.,Fritz,B.,1987.ChemicalandisotopicsurveyofrainwatersinSenegal.Geodyn2(1),43–53.

Travi,Y.,1988.HydrogéochimieEtHydrogéologieDesAquifèresFluorésDuBassinDuSénégal.OrigineEtConditionsDeTransportDuFluorDansLesEaux

Souterraines.ThesisSci.Univ.ParisSud(Orsay),190p.

Vengosh,A.,Spivack,A.J.,Artzi,Y.,Ayalon,A.,1999.Geochemicalandboron,strontium,andoxygenisotopicconstraintsontheoriginofthesalinityin

groundwaterfromtheMediterraneansoastofIsrael.WaterResour.Res.3(6),1877–1894,http://dx.doi.org/10.1029/1999WR900024.

Vengosh,A.,2013.SalinizationandSalineEnvironments.In:SherwoodLollar,B.(Ed.),Environmentalgeochemistry,TreatiseinGeochemistry.SecondEd.

11,325–378.10.1016/B978-0-08-095975-7.00909-8.(Chapter9SalinizationTreatiseinGeochemistry).

WHO,2008.GuidelinesforDrinking-WaterQuality,thirded.1.Recommendations,WorldHealthOrganization(WHO),Geneva,Switzerland,515pp.

WernerA.D.,LockingtonD.A.,2004.Dynamicgroundwaterandsalttransportnearatidal,partiallypenetratingestuary.In:MillerC.T.,PinderG.F.(Eds).

Computationalmethodsinwaterresources.55(2),1535–1547.10.1016/S0167-5648(04)80164-3.

Wolanski,E.,1986.Anevaporation-drivensalinitymaximumzoneinAustraliantropicalestuaries.Estuar.Coast.Mar.Sci.22(4),415–424,