Seasonaluseofcoastalresourcesbyotters:Comparing sandyandrockystretches
MiguelClaveroa,b,Jose´Prendaa,*,MiguelDelibesb
a DepartamentodeBiolog´ıaAmbientalySaludPu´blica,UniversidaddeHuelva,CampusUniversitario deElCarmen,Avda.Andaluc´ıa s/n,
21071Huelva,Spain
b DepartamentodeBiolog´ıaAplicada,Estacio´nBiolo´gicadeDon˜ana,CSIC,Pabello´ndelPeru´,Avda.Mar´ıaLuisas/n,41013Sevilla,Spain
Abstract
Seasonalpatternsofuseofcoastalresourcesbyottershavebeenstudied,simultaneously performingdietanalysesandspraintcounts(ex- pressedasmarkingintensity,MI).Inspiteofthesmallsizeofthestudyarea,twodifferentsectorscouldbeclearlyidentified,asandyandarocky one.Theorigin(freshwaterorcoastal)ofmostpreytypescouldalsobedetermined.Surveyswereperformedbimonthlyandclassifiedas autumnewinter(AeW)orspringesummer(SeS).
In thesandysector,eels,flatfishandgreymulletsweremorefrequentlyconsumedinAeW,whilecrayfishwasmorefrequentlyconsumedin SeS.Noneoftheidentified preytypesintherockysectorsshoweddifferentfrequencyofoccurrencevaluesbetweenthetwoperiods.Inthis sectorotterfedmainlyoncoastalpreythroughouttheyear,whileinthesandyonetherewasaclearshiftfromcoastalpreytofreshwateronesin SeS.Thedecreaseofcoastalpreyconsumption inSeSinthesandysectorcoincidedwiththelowestMIvalues,showingthatinthisperiod ottersusecoastalareaslessintensely.Thoughshiftsinotterdietanduseofspaceinthesandysectorareprobably influencedbythepopulation dynamicsofestuarinefishspeciesandcrayfish, theincreasedtouristdisturbanceduringspringandsummerinthesandysectorcouldbean importantdeterminantoftheobservedpatterns.Finally,theresultsshowthatMIcanbeusedasareliablemeasureofhabitatusebyotters, atleastinsmallareas.
©2005ElsevierLtd.Allrightsreserved.
Keywords:environmentalheterogeneity;predation;dietcomposition;fishassemblages;crayfish;Mediterraneanstreams
1.Introduction
TheEurasianotter(Lutralutra)isasemi-aquaticpredator specialisedinobtaining virtuallyallitsfoodinthewater (Carss,1995).Asformanyothercarnivores,theavailability oftrophicresourcesisthe principal determinant ofotters’ life-historyfeatures.Seasonalfluctuations inpreyabundance and/oravailabilityaffectaspectslikemortalityorreproduction timing(Kruuketal.,1987;HeggbergetandChristensen,1994; Beja,1996a).
Seasonalassessments ofotterdietinmarinehabitatsare available in the literature (Kruuk and Moorhouse, 1990;
*Correspondingauthor.
E-mailaddress:(J.Prenda).
Watt, 1995;Beja,1997;Kingstonetal.,1999).However,stud- iesoncoastalotters’dietareusuallyperformedinrocky envi- ronments.Intheselocationsthereareseasonalchanges inthe proportionofthedifferentpreyfishspecies,butthedietis uni- formlycomposedofmarinefishspeciesallovertheyear.Inan analysisofotterdietindifferentcoastalenvironments inPor- tugal,Beja(1991)showed thatawayfromtherockyshores otterdietcanshowstrongseasonalchanges, incorporating freshwaterprey.
Inthisworkweanalysetheseasonal variationinotterdiet compositioninaheterogeneous coastalareawhichfeatures adjacent sandyandrockysectors.Thetwocoastal sectors have verydistinctfeatures.Intherocky sectorthesteeprelief favourstheformationofmanywatercourses(over20),mostof themephemeral.Freshwaterfish andcrayfishpopulationsare
extremelyscarceorabsent inthesesmallstreams. Streams in thesandysectorofthestudyareahavecomparativelylarger drainageareas(onlytworivermouthscanbefoundalong thissector)andfreshwaterfishand,speciallycrayfish, are presentandinsomecasesabundant (Claveroetal.,2002). Streams inbothsectorsofthestudyarea,however,sufferex- tremeseasonalchanges followingthetypicalMediterranean climatecycleofautumnewinterfloods andsummerdroughts (GasithandResh,1999).
Itwaspredictedthatintherockycoastalsectorotterdiet wouldshowlittleseasonalvariationbeingconsistentlycom- posedofcoastalfishalongtheyear,sincestreamsinthissector featurelowavailabilityoffreshwaterprey.Otterdietwouldbe moreseasonally variableinthesandysector, incorporating bothcoastalandfreshwaterprey.
Theintensityofotter’suseofthecoastbyperiodicotter spraintcountswasalsoindirectlyassessed (MasonandMac- donald,1986).Theuseofotterspraintnumbersasindicators ofhabitat useand/orpopulationdensityisacontroversialissue (Kruuketal.,1986;Conroy andFrench,1987;Masonand Macdonald, 1987).Someworkshave,however,shownthere- lationbetweennumberofotterspraintsanddifferenthabitat features(Basetal.,1984;PrendaandGranado, 1996;Hutch- ingsand White, 2000).Thisworkrelatesthepossibleseasonal patternsofinlandandcoastaltrophicresourcesusebyottersto theintensity ofuseofthecoastalareaasassessed byspraint counts.These relations have not been previouslyanalysed andcouldaddsomelighttothisdiscussedissue.
2.Studyarea
ThisstudywascarriedoutinthesurroundingsofTarifa
(Ca´diz,SSpain).Theareacomprisesacoastalbandofabout
30kmlongthatincludesfourmainwatercourses:ElValle,La JaraandLaVegainthewestern(sandy)sectorandGuadal- mes´ıintheeastern(rocky)one(Fig.1).Thesandysector
features asofterreliefthantherocky one,wheretherearenu- merousephemeralstreams.Averageannualrainfallisaround
1000mm,withhugefluctuationswithinthestudyareadue totheabrupt relief.Meanannual temperatureisrathercon- stantamongstations,beingaround17.5oC(Ibarra,1994). ForadetaileddescriptionoftheareaseeClaveroetal.(2004).
3.Methods
Dietofotterswasstudied throughtheanalysis ofitsfaeces (spraints).OtterspraintswerecollectedbimonthlyfromDe- cember1999toDecember2001infive600-mlongcoastal transects,locatedinthelowerstretchesofthefourmainrivers and inthecommonestuaryofriversLaJaraandLaVega (Los Lancesbeach).Spraintsampleswerealsocollectedinsmall streammouthsalongtherockysectorofthestudyareafour times,inAugust andDecember 2000and2001.Allotter spraints and other otter marks were counted in the five
600-mtransectsineachsurvey,allowing thecalculationofan indexofMarkingIntensity(MI,numberofspraints600m—1) (MasonandMacdonald,1986).Onlyoneofthesetransects
waslocatedintherocky sector(seeFig.1),duetothesmall sizeandephemeral characterofotherwatercoursesinthesec- tor.Spraintcountwasnotperformed inOctober 2000,dueto heavy rains,thoughspraintswerecollectedthatmonthinthe rockysectorfordietanalysis.WealsorecordedMIinother four 600-m transects located in the main rivers’ upper stretches(Fig.1),inordertocontrolforpossible general changes inMIwithin thewholestudyarea.Thedifferentsur- veyswereclassifiedintotwoperiods:AutumneWinter (Ae W;October, DecemberandFebruary)andSpringeSummer (SeS;April,JuneandAugust).
Spraintanalysisfollowedstandardprocedures(Beja, 1997), themethodology beingthoroughlydescribedinClaveroetal. (2004).Eachidentified preytypeinaspraintwasconsidered an‘‘occurrence’’,and generaldietwasexpressedasfrequency
*
Sandy coast
*
*
Rocky coastN
36º05´ N
TARIFA
0510 kilometres
36º00´ N
Fig.1.Mapofthestudyareashowingthelocationoftwodifferentcoastalsectors.Transectsinwhichotterspraintcountswereperformedaremarked (asterisks, coastaltransects;circles,upperstreamtransects).
ofoccurrence(FO,numberofoccurrencesofacertainprey typedividedbynumberofspraintsanalysed) (Masonand Macdonald,1986).Resultsfromsandysectorandrockysector samples werepooled,sincepreviousanalyseshadshownthat dietcomposition wasquitehomogeneous withinsectors (Claveroetal.,2004).
Each prey type was classifiedas coastal or freshwater (Table1).Thecategorycoastal includedbothpuremarine preyas well as estuarineones.Acertain preytypewasconsid- eredcoastalwhenitoccurred insaltwaterandwasnever lo- catedintheupperstretches ofthestudiedstreamsduringanyof thefish surveysperformedinthearea(Claveroetal.,2005). Somepreytypesthatwereoccasionally caughtinmedium stretches,assandsmelts(Atherinaboyeri)oreels(Anguilla anguilla),werenotclassifiedinanyofthetwocategories, eventhoughtheyweremuchmoreabundantinthecoastal area.TheFOofcoastalandfreshwater prey,eachoneconsid- eredasasinglepreytype,wasalsocalculated.
Theassociationbetween freshwaterprey,coastalpreyand eelremainsinotterspraintswasassessedforeachsectorby pairwisecomparisonsin2x2contingencytables,usingc2 test.Whenever multiplecomparisonswereperformedsignifi- cancelevelswerecorrectedusingthesequential Bonferroni technique(Rice,1989).
Aprincipalcomponents analysis(PCA)wasperformedto describe themainsourcesofvariationinotter’sdietcomposi- tion,bothspatiallyandtemporally.ThePCAwasappliedtoan FOmatrix,whichwaspreviouslyarcsinetransformed.Other variablesinvolvingproportionswerealsoarcsinetransformed, whileMIvalueswerelogarithmically(basee)transformed
Table1
Seasonalvariationinotterdietcompositioninthetwosectorsdefinedinthe studyarea.Resultsareexpressedasfrequencyofoccurrence.Theclassifica-
tionofeachpreytypeascoastal(C)orfreshwater(F)isalsoshown
SandycoastRockycoast
AeW SeS AeW SeS
priortoanalysis. Therelationbetween theproportionsofthe differentpreytypes,theprincipalcomponentscoresandthe MIvalues werestudiedthroughcorrelationanalysis.Toassess differencesbetweensectorsorperiodstwo-samplecomparison testswereemployed.
4.Results
4.1.Dietdescription
Overthestudyperiod1186otterspraintswereanalysed,
812fromthesandysectorand374fromtherocky one.There werecleardifferencesinotterdietcompositionbetweenthe twosectors (Table1).Only6.4%ofthespraintscollectedin therockysectorcontainedremainsoffreshwaterprey,more than95%containing coastalpreyremains.Inthesandysector theproportions ofotterspraintscontainingcoastalandfresh- waterpreywere verysimilar,beingaround50%.Redswamp crayfish(Procambarusclarkii)wasbyfarthemostimportant freshwater preyinthissector.ThedifferenceintheFOof coastalpreybetweenthetwosectorswasstrikingwhenthe
12 surveys performed were analysed (t¼7.2; d.f.¼22;
P0.001).Amoredetaileddescriptionofotterdietinthe areaisprovidedbyClaveroetal.(2004).
4.2.Seasonalvariationindietcomposition
ThepossibleseasonaldifferencesintheFOsofthemain preytypeswereassessed bymeansofmultipletwo-sample comparisonsineachofthetwosectorsseparately(Table1). ToapplythesequentialBonferronicorrectionthose preytypes thatwerenotpresentonacertainsectorwerenotconsidered (n¼13preytypesinbothsectors).NoneoftheFOsshowed significant changesbetweenperiodsintherockysector,even withoutapplying thesequentialBonferroni correctionofsig- nificancelevels(P0.25in all cases).Inthe sandysectorflat- fish(P¼0.00007),loaches(P¼0.0009)andeels(P¼0.004) were more frequently consumed during the AeW period,
CAeW(P¼0.006),almostreachingstatisticalsignificanceaf-
ChubSqualiuspyrenaicus 1.5 4.3 e e F LoachCobitispaludica 2.7 0.3 e e F Otherfish 3.5 2.7 19.8 19.1 C Redswampcrayfish Procambarusclarkii 31.1 74.5 1.7 e F Marinecrab 7.7 2.0 6.8 5.1 C Smallcrustaceans 13.5 11.7 3.4 1.5 e Amphibians 8.1 12.4 1.7 5.9 F Reptiles 1.2 7.4 e 1.5 F Insects 2.5 3.7 2.5 3.7 F
Totalcoastal67.325.597.592.6
Totalfreshwater36.680.95.18.8
Nspraints 517 298 237 136
rocky onetheirvalues remainedquitestablealong theyear (Fig.2).Inthislattersectortherewasaremarkableconcor- danceinthefrequencyofthesethreepreytypesinotterdiet inthedifferentsurveys(Kendal’scoefficient ofconcordance; W¼0.82;c2 ¼19.6;d.f.¼2;P0.001).Nosignificantcon- cordanceamongsurveyswasfoundfortheFOsofcoastal prey,freshwaterpreyandeelsinthesandysector(W¼0.06; c2¼1.6;d.f.¼2;P¼0.45).Therewasaclearparallelsea- sonalvariationoftheFOofeelsandthatofcoastalpreyin thesandysector(r¼0.78;P¼0.003).Infact,eelandcoastal preyremainswerepositivelyassociatedinotterspraintsand both of them were negatively associated with freshwater
100
Sandy coastRocky coast
100
8080
6060
4040
2020
0
D F A J A O D F A J A OD
0
D F A J A O D F A J A O D
2000 / 2001 / 2000 / 2001Coastal prey / Freshwater prey / Eel
Fig.2.Bimonthlyvariationofthefrequencyofoccurrence(FO)ofcoastalprey,freshwaterpreyandeelinotterdietinbothsectorsofthestudyarea.
preyremains(Table2).Thisrelationshowed that,though eels canbefoundintheareabothincoastalandinupstream envi- ronments,ottersusuallycatchtheminthesameforagingbouts inwhichthey capturecoastalprey,andrarelyinfreshwater stretches. Intherockysector,whereeelsarealessimportant preytype(seeTable1),eelsandcoastalpreyremainsareneg- ativelyassociated(Table2),suggestingthatinthatsectoreels aremorefrequentlytakeninfreshwaterstretches.
Thedifferencesinotterdietcompositionbetweenthetwo
coastalhabitatfoundinthestudyarea(seeTable1) areclearly resumedbythefirstcomponent(PC1)ofthePCA,whichex- plained 47% of the observed variance (eigenvalue¼7.55) (Fig.3).Asaresultofthehigherotterpredationoncoastal preyintherockysectortherewasa strongcorrelationbetween theFOofcoastalpreyandPC1scores(r¼—0.79;P0.001). Therewasnosignificanteffectoftheperiodoftheyearon PC1scores,neitherconsideringbothsectorstogethernorana- lysingeachoneseparately(P0.3inallcases).
PC2(eigenvalue¼1.88;11.8%explainedvariance)sepa- ratedsurveyswithahighfrequencyofgreymulletsinotter dietfromthosewithahighfrequencyofcrayfish(Fig.3). Rockycoastalsurveysshowedverylittlevariationalongthe gradientdefinedbyPC2,andtheperiodoftheyeardidnot have any effect on their scores (P¼0.88). However,PC2 clearlydiscriminatedAeWsurveys fromSeSonesinthe sandycoastalsector(t¼4.1;P0.01).Therewasastrong correlation between PC2 scores and predation on coastal preyforsandycoastalsectorsurveys(r¼0.89;P0.001).
4.3.Marking intensityandotterdiet
Duringthestudyperiodtherewerestrongvariationsinthe
MIintensityvaluesincoastaltransects,bothintherockyand
inthesandysector(Fig.4).Though clearseasonalpatterns werenotapparent,anannualminimumMIwasrecordedin Augustinthetwoyearsandinbothsectors. Thelowvalues recordedinFebruary2001areprobably anartefactproduced bythestrongrainsthatoccurredbeforethesurvey.Inthe sandysectorMIwashigherinAeW surveysthaninSeS ones, though significance was only marginal (t¼1.96; d.f.¼10; P¼0.08). When the data from February 2001 werenottakenintoaccountthedifferencebetweenthetwope- riodsbecamesignificant(t¼2.56;d.f.¼9;P¼0.03).These differenceswerenotobservedintherockycoast,evenafter eliminatingdatafromFebruary2001(P0.8inbothcases). MIvariationwasrathersimilar inthefourcoastaltransects studiedinthesandysector,withthevaluesobtainedinthedif- ferentsurveysbeing significantlyormarginallypositively cor- related(P0.1inallthesixpossiblepairedcorrelations).
Amongthetransectslocatedin upperstreamstretches,only
thatinElVallerivershowedsignificantMIdifferencesbe- tweenperiods(t¼3.45;P¼0.006),withmorespraintsbeing foundinSeSsurveys thaninAeW.Intheotherthreetrans- ects the variationinMIcould not berelatedtothetwoperiods consideredinthisstudy(P0.25inthethreecases)(Fig.5).
When thetwosectors intheareawereconsidered separately,therewereclear relationsbetweenthe main sources ofvariationinotterdietcomposition(definedbyPCs)andthe
Rockycoast
Greymullets
Table2
Associationsbetweencoastalprey,freshwaterpreyandeelsinotterspraintsin bothsectorsofthestudyarea.Neg,negativeassociation;Pos,positiveassoci- ation;Ind,nosignificantassociation.P0.001inallsignificantassociations
Crayfish
Wrasses
Blennies
Autumn-Winter
Spring-Summer
Sandycoast
Eel
Flatfish
Rocklings
Smallcrustaceans
Crayfish
Fig.3.Distribution ofthedifferentsurveysinrelationtothefirstcomponents ofthePCA.
100
80
60
40
20
0
Sandy coast
Rocky coast
DFAJAOD FAJAOD
20002001
freshwaterpreywereconsumed, inthesandycoastlowerMI valueswererecorded whentheFOof freshwaterpreywas higher.Theseoppositepatternswerestronginbothsectors, buttheyhadaclearseasonalcomponentonlyinthesandy one(Fig.5).Alltheserelationsremainedsignificantafterre- movingdatafromFebruary 2001,becomingstronger inthree outoffourcases.
5.Discussion
Otterdietcomposition anduseofspaceexhibitedverydif- ferentpatternsinbothsectorsofthestudyarea.Asinitially hypothesised,thoughnoseasonalchangeswereobservedin therockycoast,therewereclearseasonalpatternsincoastal
Fig.4.Bimonthlyvariationofthemarkingintensity(MI)valuesinthetwo
studied coastalsectors. Sandy sectorvaluesaremeansforthefourtransects anderrorbarsarestandarderrors.Shadedareascorrespondtoautumnewinter surveysandunshadedareastospringesummerones.DatafromFebruary2001 areencircled,andthesevalueswerelowprobablybecauseofheavyrains.
MIvaluesincoastaltransects(Fig.5).Significantcorrelations werealsoobtainedbetweenMIandtheFOsoffreshwater prey,thoughthetrendswereoppositeinthetwosectors.While intherockysectormoreottermarkswerefoundwhenmore
andfreshwaterpreyconsumptioninthesandysector.
5.1.Rockycoast
Thedominanceoflittoralmarinefishes inotterdiet throughoutthe yearinrockycoastalhabitatis a consistentfea- tureofthestudiedEuropeanpopulations(Heggberget, 1993; Watt,1995;Beja,1997;Kingston etal.,1999)thatwascon- firmedinthestudyarea.TheFOoffreshwaterpreynever
250
ValleJara
40
200
30
150
20
100
10
50
00
VegaGuadalmesí
6020
4515
3010
155
0
D F A J A O D F A J A O D
0
D F A J A O D F A J A O D
2000200120002001
Fig.5.Bimonthlyvariationof themarkingintensity(MI)valuesin the fourtransectslocated in upperstreamstretches.Shadedareascorrespondtoautumnewinter surveysandunshadedareastospringesummerones.
Rocky coast
5
Sandy coast
6
4
5
3
4
2
13
PC1 scoresPC2 scores
56
4
5
3
4
2
1
0102030
3
20406080100
Arcsin FOfreshwater prey (%)
Fig.6.Relationsbetweendietcompositionandmarkingintensity(MI) incoastaltransectsofthestudyarea.ThepointcorrespondingtotheFebruary2001survey ismarked(*)inallplots.OriginalMIunitsfromtherockysectorarespraints/600m,whilethosefromthesandysectorarespraints/2400m(sumofthefour transectsinthatsector).Rockycoast:PC1-MI,r¼0.86,P0.001;FOfreshwaterpreydMI, r¼0.78,P0.01.Sandycoast:PC2-MI,r¼0.83,P0.001; FOfreshwaterpreydMI,r¼—0.69,P¼0.01.
exceeded25%anditsimportanceinotterdietwasnotsea- sonal(seeFig. 3),reflectingthescarcityoffreshwaterfoodre- sourcesthatstreamsinthissectorofferfortheotter(Clavero etal.,2002).Itwasquitesurprisinganyway thatnoneofthe FOs ofthepreytypesconsumed inthissectorshowedsignif- icantseasonalvariations,sincethesehavebeendescribedin otherlocations(KruukandMoorhouse, 1990;Watt,1995). Beja(1997)showedseasonalpatternsintheconsumptionof someimportantfishspeciesin a rockycoastalhabitatin south- ernPortugal.Theidentificationofpreytypestothe familylev- elusedinthiswork could accountforthislackofseasonality, whichmighthavebeendetectedifpreyhadbeenidentifiedto species. Thedivisionoftheyearintoonlytwoperiodscould alsobetoosimpletodetectsomefinerseasonalvariationin dietcomposition(e.g.KruukandMoorhouse,1990).Never- thelesssomeoftheobservationsdifferfromthosepresented byBeja(1997), forexample,theminimumsummer predation onwrasseswasnotdetected,withwrasseconsumption peak- inginAugustandOctobersurveys.Itisthereforepossible thatseasonalpatternsindietcompositioninthestudyareadif- fer,beingprobably lessmarked,fromthoseinthesouth-west coastofPortugal.
Thestrongandpositive relationfoundbetween theoccur- rence offreshwaterpreyandthemarkingintensityintherocky coastalsectorisapparently acontradictorypattern.Otter spraintcountswere,however,performedinatransectplaced in the mouth of the main stream within the sector (see Fig.1).Beja(1996b)showedthatottersinhabitingasimilar environment inPortugalmovedmainlyalongthecoast,using watercoursesasdaytimeshelter.Theserest-siteswerenever locatedfarfromthecoast.Sincethistransectwasplacedat
themainsourceoffreshwaterpreyintherockysector,itis predictablethathighMI valueswouldcoincidewithmaximum predation onit.LowMIvalueswouldthenindicatethatotters weremainlyusingotherareasalongthecoast.
5.2.Sandycoast
Seasonalvariationinotter diet in the sandysectorwasclear andfollowedthe samepatterninthe twoyearsofstudy.Fresh- waterpreywasmorefrequentlyconsumed byottersin springesummer, coincidingwithaclearreductionincoastal preyandeelconsumption (seeFig.2).Inthissectorthesur- veysinwhichpredationonfreshwater preywashighcoin- cided with those with low MI (SeS surveys), indicating alessintenseuseofthecoastalarea.Inthisperiodotters werefeedingmainlyinfreshwaterstretches placedupstream. Beja(1996b)radio-tracked afemaleotterinanestuarine streaminPortugalinspringandsummer,periodduringwhich itwasdetectedonlyinupstreamareas.
Thereducedexploitationofcoastalresourcesbytheotterin springandsummer coincides withtheperiodwhenestuarine andmarshareasusuallyreachthehighestvaluesinfishabun- danceanddiversity(Yoklavichetal.,1991;Rebelo,1992;Laf- failleetal.,2000;GordoandCabral,2001).Summerisalso thetimewhenMediterraneanstreamssufferadrasticdrought, mostofthembeingreduced toisolated poolsthatactasrefu- gestofreshwaterfauna(PrendaandGallardo,1996;Gasith andResh,1999;Magalh~aesetal.,2002).Inthissituationit would bepredictablethatottersusedthecoastmoreintensely duringspringandsummer,whichistheoppositepatterntothe oneobservedinthestudyarea.
Theseapparentlycontradictory patternscouldnevertheless berelatedtothefactthatsummerpeaksinfishabundanceare oftenduetotheincreasednumberof juveniles(GordoandCa- bral,2001).Somegreymulletspecies(Mugilcephalus,Liza aurataorLizasaliens)spawnintheseaduringthesummer (Ben-Tuvia,1986;Cardona,2000),andthusbiggerindividuals arenotavailableinestuaries.Greymulletsarethemostimpor- tantotterpreyintheareaintermsofbiomass(Claveroetal.,
2004).Cabral(2000)showedthatintheSadoestuary(Por-
tugal)Soleidaeflatfishpopulationsfeaturedhighproportions oflargeindividuals inautumn andwinter, whensomespecies alsoreachedtheirmaximumabundances.Anotherfactorinflu- encingotter predationuponestuarinefishescouldbetheir skill toavoidotterattacks.Fishactivityisdependent ontempera- ture(FrederickandLoftus,1993), beingmaximuminsummer months,whenfishreachhighswimmingspeed,thusreducing theirvulnerabilitytootterpredation(Chanin,1985).
Ontheother hand, thespringesummerpeakintheFOof freshwaterismainlyduetocrayfishconsumption(seeFig.3) andcoincideswiththeperiodwhenitismoreabundantandbig-
gerindividualsareavailable(NiquetteandD´Abramo,1991;
Correia,1995;Beja,1996c). Similarseasonal variationsof predationuponcrayfish havebeenpreviouslyreportedinthe IberianPeninsula fortheotterandotherpredators(Beja,1996c; Correia,2001),withamaximumcrayfishconsumptionduring summerandaminimuminwinter.Duringthespringesummer period,whenestuarinefishcommunitiesaredominatedbyjuve- nilesandfishactivityismaximum,itcouldbecomemoreprofit- ablefortheottertopredateuponcrayfish,eventhoughcrayfish isclearlyalessenergeticallyrewardingprey(Beja, 1996c).
Itisalsopossiblethat seasonalchangesinotterdiet compo- sition anduseofthesandycoastalareaarenotonlyrelatedto preypopulations’characteristicsbutalsotovariationsinthe intensityofhumandisturbance. Thepresenceoftouristsin sandbeachesisaseasonalphenomenon,withaclear peakdur- ingsummerthatcouldinfluence otter’sforagingbehaviour.It hasbeensuggestedthatottersarenotverysensitive tohuman disturbance(MasonandMacdonald, 1986)andBeja(1996b) showedthatotters’ rest-sitesinthePortuguesecoastwere sometimeslocated in intenselyhuman-disturbedareas.But thesameauthor(Beja,1992)also foundthatthenumberof ot- terspraintsinheavilydisturbedstreammouthswas sensitively lowerthanthatinundisturbedstreammouths.
5.3.Marking intensityasindicator ofhabitat use
Inthisworkottermarkingintensitywasfollowedintwo adjacent,thoughverydifferent, coastalareasoveratwoyear period.Noattemptwasmadetocomparevaluesobtainedin thetwosectors. Previous works(e.g.Kruuk,1992)havesug- gestedthatotter’ssprainting activitydecreasesinsummer,re- sulting insmallMIvalues. Thatwasnotthecaseinthestudy area,wheretherewerenoclearseasonal trendsinMIvalues. SometransectsshowmaximumsinSeSwhileotherspeaked inAeW (Figs.4and5).Thetemporalvariationsinhabitat useintensityandinotterdietcompositioncouldberelated andthecoastalorfreshwateroriginofmostpreytypescould
beassessedincoastaltransects. Thisallowedrelatingthein- tensityofuseofacertainareawiththefrequencyofoccur- renceofpreycapturedinthatsameareaorfarfromit.The observedrelationsbetweendietcompositionandfrequency offreshwaterpreyandmarkingintensitywereclear(Fig.6). Itwasalsopossibletorelatetheobservedpatterns tothechar- acteristicsofbothstudiedsectorsandtohabitatusepatterns reported in similar areas using radio-tracking techniques (Beja,1996b).InagreementwiththerevisionbyHutchings andWhite(2000),itissuggestedthatotterspraintdensityis ausefultooltoassessotterhabitatuse,atleastinsmallareas.
References
Bas,N.,Jenkins, D.,Rothery,P.,1984. EcologyofottersinnorthernScotland V. The distributionof otter(Lutralutra)faecesinrelation tobanksideveg- etationontheriverDeeinsummer1981.JournalofAppliedEcology21,
507e513.
Beja,P.R.,1991.Dietofotters(Lutralutra)incloselyassociatedfreshwater, brackish andmarinehabitatsinsouth-west Portugal.JournalofZoology, London225,141e152.
Beja,P.R.,1992.Effectsoffreshwateravailabilityonthesummerdistribution ofottersLutralutrainthesouthwestcoastofPortugal.Ecography15,
273e278.
Beja,P.R.,1996a.SeasonalbreedingandfoodresourcesofottersLutralutra (Carnivora,Mustelidae) insouth-westPortugal:acomparison between coastalandinlandhabitats.Mammalia60,27e34.
Beja, P.R.,1996b.Temporalandspatial patternsof rest-siteusebyfourfemale ottersLutralutraalongthesouth-west coastofPortugal.JournalofZool- ogy,London239,741e753.
Beja,P.R.,1996c.AnanalysisonotterLutralutrapredationonintroduced
AmericancrayfishProcambarus clarkiiin Iberianstreams.Journalof
AppliedEcology33,1156e1170.
Beja,P.R.,1997.Predationbymarine-feedingotters(Lutralutra)insouth- westPortugalinrelationtofluctuatingfoodresources.JournalofZoology, London242,503e518.
Ben-Tuvia, A.,1986.In:Whitehead,P.J.P.,Bauchot,M.L.,Hureau,J.C., Nielsen,J.,Tortonese,E.(Eds.),FishesoftheNortheasternAtlanticand theMediterranean,vol.III.UNESCO,Paris,pp.1197e1204.
Cabral,H.N.,2000.Distributionandabundancepatternsofflatfishesinthe
Sadoestuary,Portugal.Estuaries53,351e358.
Cardona,L.,2000.Effectsofsalinityonthehabitatselectionandgrowthper- formanceofMediterraneanflatheadgreymulletMugilcephalus(Osteich- thyes,Mugilidae).Estuarine,CoastalandShelfScience50,727e737.
Carss,D.N.,1995.Foragingbehaviour andfeedingecologyoftheotterLutra lutra:aselectivereview.Hystrix7,179e194.
Chanin,P.,1985.TheNaturalHistoryofOtters.CroomHelm,London,179pp. Clavero,M.,Rebollo,A.,Valle,J.,Blanco,F.,Narva´ez,M.,Delibes,M., Prenda,J.,2002.Distribucio´nyconservacio´ndelaictiofaunacontinental
en pequen˜os cursosde aguadel Campode Gibraltar.Almoraima27,
335e342.
Clavero,M.,Prenda,J.,Delibes,M.,2004.Influenceofspatialheterogeneity onotter(Lutralutra)preyconsumption.AnnalesZoologiciFennici41,
551e561.
Clavero,M.,Blanco-Garrido, F.,Prenda,J.,2005.Fishehabitatrelationships andfish conservationinsmallcoastalstreamsinsouthernSpain.Aquatic Conservation:MarineandFreshwaterEcosystems15,415e426.
Conroy,J.W.H.,French,D.D.,1987.Theuseofspraintstomonitor popula- tionsofotters(LutralutraL.).SymposiumoftheZoologicalSocietyof London58,247e262.
Correia,A.M.,1995.PopulationdynamicsofProcambarusclarkii(Crustacea:
Decapoda)inPortugal.FreshwaterCrayfish8,276e290.
Correia,A.M.,2001.Seasonalandinterspecificevaluationofpredationby mammalsandbirdsontheintroducedredswampcrayfishProcambarus
clarkii(Crustacea,Cambaridae)inafreshwatermarsh(Portugal).Journal ofZoology,London255,533e541.
Frederick,P.C.,Loftus,W.F.,1993.Responsesofmarshfishes andbreeding wadingbirdstolowtemperatures:apossiblebehavioural linkbetween predatorandprey.Estuaries16,216e222.
Gasith,A.,Resh,V.H.,1999.StreamsinMediterraneanclimate regionseabi- oticinfluencesandbioticresponsestopredictableseasonalevents.Annual ReviewofEcologyandSystematics30,51e81.
Gordo, L.S.,Cabral, H.N.,2001. Thefishassemblagestructureofahydrolog- icallyalteredcoastallagoon: theO´bidoslagoon(Portugal).Hydrobiologia
459,125e133.
Heggberget,T.M.,1993.Marine-feedingotters(Lutralutra)inNorway:sea- sonalvariationinpreyandreproductivetiming.JournaloftheMarineBi- ologicalAssociationoftheUnitedKingdom73,297e312.
Heggberget,T.,Christensen,H.,1994.ReproductivetiminginEurasian otters onthecoastofNorway.Ecography17,339e348.
Hutchings,M.R.,White,P.C.L.,2000.Mustelidscent-markinginmanaged ecosystems:implicationsforpopulationmanagement.MammalReview
30,157e169.
Ibarra,P.,1994.NaturalezayhombreenelSurdelCampodeGibraltar: un ana´lisispaisaj´ısticointegrado.AgenciadeMedioAmbiente,Consejer´ıa deCulturayMedioAmbiente.JuntadeAndaluc´ıa,Sevilla,440pp.
Kingston,S.,O’Connell,M.,Fairley,J.S.,1999.DietofottersLutralutra on Inishmore,AranIslands,westcoastofIreland.Proceedings oftheRoyal IrishAcademy(BiologyandEnvironment)99,173e182.
Kruuk,H.,1992.Scent markingbyotters (Lutalutra):signalingtheuseofre- sources.BehavioralEcology3,133e140.
Kruuk,H.,Conroy,J.W.H.,Glimmerveen,U.,Ouwerkerk,E.J.,1986.Theuse
ofspraintstosurveypopulationsofottersLutralutra.BiologicalConser- vation35,187e194.
Kruuk,H.,Conroy,J.W.H.,Moorhouse,A.,1987.Seasonalreproduction,mor-
talityandfoodofotters(Lutra lutraL.)inShetland.Symposiumofthe
ZoologicalSocietyofLondon58,263e278.
Kruuk,H.,Moorhouse, A.,1990.Seasonalandspatialdifferencesinfoodse- lectionbyotters(Lutralutra)inShetland.JournalofZoology,London
221,621e637.
Laffaille,P.,Feunteun,E.,Lefeuvre,J.C.,2000.Compositionoffishcommu- nitiesinaEuropeanmacrotidalsaltmarsh(theMontSaint-MichelBay, France).Estuarine,CoastalandShelfScience51,429e438.
Magalh~aes,M.F.,Beja,P.,Canas,C.,Collares-pereira,M.J.,2002.Functional heterogeneityofdry-seasonfishrefugiaacrossaMediterraneancatchment: theroleofhabitatandpredation.FreshwaterBiology47,1919e1934.
Mason,C.F.,Macdonald,S.M.,1986.Otters:EcologyandConservation.Cam- bridgeUniversityPress,Cambridge,248pp.
Mason,C.F.,Macdonald,S.M.,1987.Theuseofspraintsforsurveyingotter
Lutra lutra populations: an evaluation. Biological Conservation 41,
167e177.
Niquette,D.J.,D’Abramo, L.R.,1991.Populationdynamics ofredswamp crayfish,Procambarusclarkii(Girad,1852)andwhiterivercrayfish,P.as- tacusastacus(Girard,1852)culturedin earthenponds.JournalofShellfish Research10,179e186.
Prenda,J.,Gallardo,A.,1996.Self-purification, temporalvariabilityandthe macroinvertebrate communityinsmalllowlandMediterraneanstreams re- ceivingcrudedomesticsewageeffluents.Archivfu¨rHydrobiologie136,
159e170.
Prenda,J.,Granado,C.,1996.Therelativeinfluenceofriparianhabitatstruc- tureandfishavailabilityonotterLutralutra,L.spraintingactivityin asmallMediterraneancatchment.BiologicalConservation76,9e15.
Rebelo,J.E.,1992.Theichthyofaunaandabiotichydrological environmentof theRiadeAveiro,Portugal.Estuaries15,403e413.
Rice, W.R.,1989.Analyzingtablesof statistical tests.Evolution43,223e225.
Watt,J.,1995.Seasonal andarea-relatedvariations inthedietofottersLutra lutraonMull.JournalofZoology,London237,179e194.
Yoklavich,M.M.,Cailliet,G.M.,Barry,J.P.,Ambrose,D.A.,Antrim,B.S.,
1991.Temporalandspatialpatternsinabundance anddiversityoffishas- semblagesinElkhornSlough,California.Estuaries14,465e480.