PatternsofOrthopteraabundanceandlesserkestrel conservationinarable landscapes
Carlos Rodr´ıguezÆ Javier Bustamante
Abstract ThelesserkestrelFalco naumanniexperiencedamarkeddeclineduringthe secondhalfofthe20thcenturyduetochangesinlandusethatinfluencedbreedingsuccess by reducingtheabundanceandqualityof prey.However,thefactorsgoverningspatialand temporalvariationofpreyabundance aroundlesserkestrelcolonieshasnotyetbeen investigated. WesampledOrthopteraabundanceinthemaincroptypesandedgehabitats surroundingsixlesserkestrelcoloniesinsouthernSpain.SamplingsfocusedonOrthoptera becausetheyconstitute themainpreyduringthenestlingperiod.OnlythoseOrthoptera speciesthatareknowntobepreyedbylesserkestrelswereconsidered inthisstudy.We founddifferencesinpreydensityamonglocalities,andcroptypes.Semi-natural habitats suchasgrasslands,fallowland,andfieldmarginsheldthehighestdensities.However,prey abundanceshowedacomplexpatternthat wasnotpossibletoexplain solelyonthebasisof cropcompositionaroundcolonies.Factorsdeterminingproductivityinindividualfields likesoiltypeandproductivityorbiocideinput,andmeansizeofagriculturalfields con- tributedtoexplainthiscomplexpatternofpreyabundance. Ourresultshighlightthekey roleofsemi-natural andedgehabitatsinfarmedlandscapesaspreyreservoirsandcorri- dors.Higherconservationprioritiesforthesehabitatsaresuggestedtobenefitforaging lesserkestrels,butmanyotherfarmlandspeciesthatalsoexperiencedsteeppopulation declinesduetodecreasingfoodsupplyresultingfrommodernagriculture.
Keywords Agri-environmentschemes Falconaumanni Margins Preyabundance
Preybiomass
Introduction
Modernagricultureisconsideredtobeamajoranthropogenic threattobiodiversity, comparabletoclimatechange initsabilitytoaffectvastareas (Donald etal.2002). The populationsofmanyfarmlandbirds,includingthelesserkestrelFalco naumanni,have
C.Rodr´ıguez() J.Bustamante
Estacio´nBiolo´gicadeDon˜ana,CSIC,AvdaMar´ıaLuisas/n.,Sevilla,Spain e-mail:
experimentedmarkedpopulationdeclinesinlastdecadesacrossEurope(Fulleretal.1995; Donaldetal. 2001).There is compelling andextensiveevidence to showthat these declineshavebeendrivenbyagricultureintensification(e.g.Gregoryetal.2004;Mattison andNorris2005).
Oneofthemostcommonly arguedprocessoperatingbehindthisnegativerelationship betweenfarmlandbirdabundanceandagriculturalintensification isfoodavailability (Newton2004;Morrisetal.2005).Howeverfewstudieshavequantifiedtherelationships betweenarthropodabundanceandagriculturalpractices(butseeMcCracken etal.1995; Wilsonetal.1999),especiallyatthescaleofthehome-rangeofthespeciesinvolved(but seePotts1986;Brickleetal. 2000;Hartetal. 2006).Understandingthisrelationshipcould beofenormousappliedinterest,becauseitconstitutesthescientific basisforthedevel- opmentofnewagri-environmental schemesthatcontrarytopreviousresults(lowto moderatesuccessoftheseschemes;KleijnandSutherland 2003;Kleijnetal.2006),can guaranteepositiveresultsonfarmlandbiodiversity(seeAebischerandEwald2004).This isespecially importantin Mediterraneancountrieswhereagri-environmentalschemeshave beenlittleused(LlusiaandOn˜ate2005),andtheireffectivenessremainalmostunknown becauseofthelackofscientificevaluations(KleijnandSutherland2003;butseeKleijn etal.2006).
Thelesserkestrelisacolonial migratoryspeciesassociatedto agricultural landscapesin WesternEurope,breedinginholesofoldbuildings,suchaschurchesorcastleswithin urbanareas,andalsoundertiledroofsoffarmhouses (Negro1997).Forthisspecies,the reduction inboththeextentandqualityofforaginghabitatsinitsWesternPalaearctic breedingrangeappearsto betheprimarycauseofdecline(PeetandGallo-Orsi2000).This isespeciallytruefortheAndalusian population(Southern Spain),wherestarvationisthe maincauseofnestlingmortality(Negroetal.1993).Bothpreyqualityandavailability havebeendemonstratedtobepivotalindeterminingbreedingsuccess(Rodriguezetal.
2006),whichconstitutes akeystoneparameterforthestabilityofthispopulation (Hiraldo etal. 1996). However,this is the firststudy providingdetailed informationon prey abundanceindifferentcroptypeswithinlesserkestrelhome-range,andtoanalysethe patterns,andfactorsinvolvedinitsspatialandtemporalvariation.
WeexaminedtherelativeabundanceofOrthopteraatsixlocalitieswith arablelanduse holdinglesserkestrelcolonies.Weusedthisgroupofarthropods becausetheyconstitute thebulkoflesserkestrel’sdietduring thechicks rearingperiod(78%oftotalbiomass— Rodriguez 2004; Rodriguez etal. 2006). Our main aim is to identify those factors influencingpreyabundanceinlesserkestrelhome-rangeduringthisperiod(whenfood demandishighest).Thisknowledge couldbehelpfulindesigningandimplementingnew managementactionstobeconductedinthelesserkestrelbreedingrange.
Methods
Studyarea
DatawerecollectedintheGuadalquivirriverbasin(south-westernSpain),apredominantly flatpseudo-steppe(elevation range20–240m)wheretraditionalagro-grazingsystems havebeenreplacedbyasystemofcontinuous croppingwithfullmechanizationandhigh inputofchemicalfertilizersandbiocides(Ferna´ndez etal.1992;Dona´zaretal.1993).In thisarea,sixurbanlesserkestrelcolonieshavebeenmonitored since1988andtheirbasic breedingparametersrecordedannually(seeRodriguezandBustamante2003formore
123
details).Thesecolonies arelocatedinthevillagesofArahal,Lebrija,MairenadelAlcor, Manzanilla,Moro´ndelaFrontera,andPalmadelCondado(Fig.1).Themaincropsin theselocalitiesarewheat,sunflower,andolivegroves(Rodriguezetal.2006).Grassy marginsareusedtoseparateindividualfields.
Orthopterasampling
Orthopteraweresampledinaradiusof3kmaroundthesixabove-mentionedcolonies. Negroetal. (1993)foundthat about75%oflesserkestrelforagingtripsduringthe breedingseasonwerewithin3kmofthecolonyofMairena,andFrancoetal.(2004)also suggested thisradiusascriticalforaginghabitatforthespeciesinSouthernPortugal. Samplings wereconducted almostdailyduringthenestlingperiod(fromlateMaytomid July;seeRodriguezandBustamante2003)inthreeconsecutive years(1999–2001). Samplingdateswererandomlydistributedamongcolonies.Weusedtwodifferentsam- plingmethods.Thefirstconsisted intransectsof30minduration(Voisin1980)randomly distributedwithinthe 3-kmradius,andthusincludingthe mostrepresentativecroptypesof the studyarea(Table1).Wealsoconductedtransectsinnon-arablelandsuchasremaining patchesofgrasslandandfallowland(hereaftersemi-naturalhabitats).Minorcroptypes (e.g.sugarbeet,chickpea,ormaize)weregroupedasothers.Urbanlandwasnotcon- sidered.TransectswerewalkedcarryingaGlobalPositioning System(GPS)andtryingto keepaconstantspeed.Thedistancecovered inthe30-mintransectwasknownfromthe GPStrack.Beltwidthoftransect(definedasthewidthofgroundvisiblebytheobserver) wasalsorecordedtoallowestimatesofOrthopteradensity(range:0.5mindensesun- flowercrops—2minolivegroveswithbareground).Weuseddensitiesinsteadofcounts
Iberianpeninsula
C6 C4C3 C1
C5
Colony C1-Arahal C2-Lebrija
C3-Mairena
Mean field size
0.026Km2
0.037Km2
0.024Km2
C2
N
100km
C4-Manzanilla
C5-Moron
C6-Palma
0.01 Km2
0.037Km2
0.013Km2
Fig.1 Locationofsamplinglocalitiesandmeansizeofagriculturalfields
Table1 Meandensity±SE(individualsperhectare)ofpreyspeciesfoundindifferentcroptypesandlocalitiesduring30mintransectsandILAs(lastcolumn)
%Cover(range) / Cereal23(8–33) / Cerealstubble / Cotton
7(0–10) / Olive
19(0–46) / Others / Semi-natural
9(4–22) / Sunflower
17(5–30) / Vineyards
2(0–3) / ILAs
Arahal / 11.5±5.2 / 9.1±2.4 / – / 0.2±0.2 / – / 178.3±159.7 / 11.7±6.4 / – / 1117±316
Lebrija / 10.5±10.5 / 1.0±0.5 / 4±3.2 / – / 0 / 49.8±26 / 1.2±0.7 / – / 135±40
Mairena / 18.5±9.4 / 3.7±1.3 / – / 0 / – / 37.7±25.3 / 4.4±1.8 / – / 331±70
Manzanilla / 18.7±6.6 / 31.5±8.8 / – / 1.2±0.8 / 35.8±20.2 / 25±8.4 / 11.2±3.6 / 36±15.5 / 624±122
Moro´n / 9.1±4.7 / 11.9±4.7 / – / 1.2±0.6 / – / 15.4±6.1 / 6.2±4.4 / – / 544±198
Palma / 80.3±69.2 / 17.7±11.0 / 17±7.1 / – / – / 350.1±200 / 4.7±1.4 / 6.8±3.2 / 674±150
SAMPLE / 42 / 140 / 27 / 74 / 20 / 86 / 171 / 25 / 163
Valuesbasedonlessthanfivesampleswereomitted.Meancoverof each croptypeandrange(betweenparentheses)in athree-kmradiusaroundthecoloniesisprovided.For thiscalculation,absenceswereomitted
becausetheareacoveredin30-mintransectsvariedbetweensamplings,andonlydensities allowedforcomparisonbetweenthem.Ineachfield, transectswereconductedatamin- imumdistanceof10mfromitslimitstoavoidthepotentialinfluenceofneighbouring cropsormargins(Badihetal.1997).Whenfieldlengthwasnotenoughfora30-min transect,weconductedparalleltransectwithinthesamefield separatedaminimumdis- tanceof10m(Gottschalketal.2003).Inordertoprevent double counting,allOrthoptera specimens,exceptsmalllarvae,weretrappedwithanentomologicalnet.Occasionally, longflightsoflargeOrthopteraprevented ustocapturethemandtheywererecordedwith thegenericnameof‘‘locust’’.Iftheirnewpositionwasinthesamplingdirection,thiswas consideredtocountthemonlyonce.WhenOrthopteradensitywashighthedistance walkedduringthe30-mintransectwasshorttobeaccuratelymeasured withtheGPS.In thoseinstances,weuseda50mfibre-glasstapetomeasurethedistancecovered.
Notonlycrops,butalsofieldmarginsandedgesconformtheagriculturallandscape,and thelatterareknowntobeimportantforarthropods (MarshallandMoonen2002).In margins,vegetationwastoodensetobeefficientlysampledwiththepreviousmethod,so weusedasecondmethodexclusivelyappliedtotheselinearhabitats,whereOrthoptera tendtohidebehind broadleaves andstemsmorethantojumpastheyusetodoincrops (Voisin1980,1986).Thismethodconsistedinaminimumoffive1091mquadrats wherevegetationwascarefullysearchedforOrthopteraspecimens. Eachsetofquadrats wascalledIndicelineared’abundance (ILA;seeVoisin1986fordetails),andtheywere conductedduring2000and2001.
Forbothmethods,samplingsweremadefrom10amto8pm(localtimes),when temperaturesinthestudyareaarewarmenough(above20°C)forinsectactivity(Willott
1997).IneachindividualsamplingwenotedaspotentiallyexplainingfactorsofOrthoptera abundance:croptype,cropdevelopment,cropcover,windspeed,andtemperature(see Voisin1980;Willott1997;Gottschalketal.2003).Allofthemwereestimatedbyeye andcategorized.Categoriesforcropdevelopmentandcoverweregroupedintofive cat- egories: 0=0%(recentlysownfieldsornovegetationcover); 1=1–20%;2=21–40%;
3=41–60%;4=61–80%;5=81–100%(cropsfullydevelopedornobaregroundfrom aboveview). Windspeed wasrelativelylowinthestudyarea(98%outof2,702obser- vationswerebelow20knots—37 kmh-1;SpanishInstitutoNacionaldeMeteorologia), andthenwe consideredwindasalogicalvariable(true/false).Usingtheentomologicalnet aswindsock,weconsideredwindydayswhenthewindwasabletofillthenet(catego- ries3,4,and5inBeaufortwindscale,approximately).Meantemperature duringsampling datesaveraged25.1°Cwithlowvariationamongdays (SD=2.9°C),andamonglocalities (SD=1.3°C).Allindividualscapturedwerecounted,andreleasedinthefieldimmedi- atelyaftersampling,exceptasub-sampleofoneindividualforeachcombination of species, sizeandsexfoundineachsamplinglocation.Thissub-samplewasbrought tothe labandweightedfresh(tothenearest10-2 g)toobtainestimatesofOrthopterabiomass.
Orthopteraspecimenswereidentified tothegenusorspecieslevelsusingdichotomic keys (Harz 1969; Clemente etal. 1987), and with the help of Orthoptera experts (Dr.A.Badih.University ofGranada,Spain,andS.Gadoum.ParcNaturelRe´gionaldu Vexinfranc¸ais,France).AllsampledindividualsbelongtocommonOrthopteraspecies, andtheywerenotincludedinanycategoryofconservation concern(IUCN1983; Gangwereetal.1985).Fromthewholesetofspecimenssampled,weonlyconsidered thosebelonging tospeciesactuallyconsumed bythelesserkestrelasidentifiedinathree- yeardietarystudyonthesamesixcoloniesofthisstudy(Rodriguez2004;seeRodriguez etal.2006forasummary).This includedthetreelocustAnacridiumaegyptium,thedesert locustCalliptamussp.,thewhite-faced bushcricketDecticusalbifrons,thesaddle-backed
bushcricketEphippigerephippiger,themigratorylocustLocustamigratoria, andgrey bushcricketsPlatycleisspp.(allof themreferredasprey or preyspecieshereafter).Large, unidentifiedlocusts(seeabove)werealsoconsideredaslesserkestrel’sprey.
Statisticalanalyses
Toidentifywhetherthereweredifferencesinthedensity ofpreyandpreybiomass among localitiesholdinglesserkestrelcolonies,weusedgeneralizedlinearmodelsGLMs(see Crawley2002)wherethelocalitywasintroduced asfactor.Toascertainwhichvariables determinedthedensityofpreyandpreyabundanceinthetwosamplingmethods,colonyand yearwereintroducedasrandomfactorsinGeneralizedmixedlinearmodels(GLMMs)fitted byrestrictedmaximumlikelihood(seePinheiroandBates2002).Croptype,cropcover,crop development,wind,andtemperaturewereintroducedaspotentialexplanatoryvariables (fixedfactors).Date(asthenumberofdaysfromthefirst ofJanuary)wasintroducedas continuousvariable.Wealsotestedseconddegreepolynomialsfitted tovariabledateto checkforpotentialcurvilinearresponses.Becausetheaggregationoffieldsisoneofthemain processesinagriculturalintensification(Donaldetal.2002)andmeansizeoffieldshas demonstratedtobeoneofthebeststructuralcorrelatesoffarmlandbiodiversity(Baessler andKlotz2006),we alsoincludedmeanfieldsize (continuous)inouranalyses.Meanfield sizewascalculatedbydividingmunicipalitysurfacebythenumberofagriculturalfieldsin themunicipality.Originaldatawerelogtransformedtoachievenormalityoferrors. When necessary,heteroscesdasticitywascorrectedbyfittingdifferentvariancestructureforeach factorusingtheweightfunctionofS-plus(seePinheiro andBates2002fordetails). StatisticalassumptionswerecheckedusingTukey-AnscombeandNormal/QQdiagrams (seeCrawley2002;PinheiroandBates2002).Modelsstartedbysequentiallyincludingeach predictor,usingaforward-stepwise searchofthebestmodel.Ineverystep,wetestedthe significanceofvariablesbycomparingthecurrentfullmodelversusthereducedmodel withouteachparticularvariableusingP\0.05andAICvaluesascriteria.Thisprocess stoppedwhennoadditionalvariablescontributedtoincreasesignificantlytheexplanatory abilityofthemodel. AllmodelswererunwithS-plus2000(MathSoft2000).
Results
Weconducted atotalof585transects,and163ILAs.Thedistribution ofsamplingsper localityandyearisprovidedinTable2.
Table2 Numberofsamplingsperlocalityandyear
ILAs
Individuals
Transects30min.
Biomass
Arahal
Mairena
Morón
Arahal
Mairena
Morón
Lebrija
Individuals
Manzanilla
Palma
ILAs
Lebrija
Manzanilla
Palma
Biomass
Arahal
Mairena
Morón
Arahal
Mairena
Morón
Lebrija
Manzanilla
Palma
Lebrija
Manzanilla
Palma
Fig.2 Meanpartialeffectoffactorlocalityonthedensityofpreyandpreybiomassinboth30-min transects(ontop)andILAs(bottomplots).Broken linesindicatetheSEofthemean.Therugplotonthex- axesindicatesthedensityofdatapoints
Inbothmethods, thedensityofpreyandpreybiomassvariedamonglocalitiesholding lesserkestrelcolonies(P\0.0001inallcases),beingArahal,Manzanilla,andPalma thosewithhigherdensitiesinallcases(Fig.2).
TheminimumadequateGLMMbuiltforthedensityofpreyindividuals in30-min transectsincludedlocalityandyearasrandomfactors,andcroptype(P\0.0001),crop development (P=0.0006) and mean fieldsize (P=0.01) as explanatory variables. Parameter estimatesindicatedthatthedensityofpreyfollowsthesequence: semi-natural habitats[cereal crops[cotton crops[vineyards[others[sunflower crops[olive groves. Nonetheless,there were some variations among localities (see Table1) that accordingtothemodelcouldbeduetodifferencesincropdevelopment (positiveeffect), andmeanfieldsizeamonglocalities(negativeeffect).Itwasnotpossibletobuildthe GLMMincludingtheinteractionbetweencroptypeanddevelopment,sowebuiltaGLM includingthesamefixedfactorsthantheGLMM.Thisinteractionwasnotsignificant (P=0.2).
ThebestGLMM builtfordensity ofpreybiomass in30-mintransects wasverysimilar tothemodelfor preydensity.Itincludedlocalityandyearasrandomfactorsandcroptype (P\0.0001), crop development (P=0.0002) and mean field size (P=0.007) as explanatoryvariables.Parameterestimatesindicatedthatthepreybiomassfollowsthe sequence: semi-natural habitats[cotton crops[vineyards[cereal crops[others [sunflower crops[olive groves. Prey biomass was positively influenced by crop development,andnegativelyinfluencedbymeanfieldsizeineachlocality.Itwasnot possibleto buildtheGLMMmodelincludingtheinteraction betweencroptypeand
development,butthisinteractionwasnotsignificant(P=0.2)inaGLMincludingthe samefixedfactorsthantheGLMM.
ThebestGLMMmodelsforpreydensityandpreybiomass inILAsincluded locality andyearasrandomfactors andvegetationcover (P=0.002;P\0.0001,respectively)as the main explanatory factor. The model built for prey density also included wind (P=0.02),whilethemodelforpreybiomass alsoincluded meanfieldsize(P=0.04)as additionalexplanatory variables.Parameter estimatesindicatedapositiveeffectofcover andanegativeinfluenceofwindandmeanfieldsize.
Discussion
AhighvariabilityofOrthoptera densitieswasfoundamongdifferentcroptypesand localitiesinSouthern Spain.Becauseweonlyfocusedonthosespeciesthatareknownto bepartoflesserkestrel’sdietduringthechick-rearing period(seeRodriguez2004),our resultspointouttothemainfactorsdetermining preyabundanceforthisendangered speciesinthestudyarea.
Inafirst, simpleapproachwefounddifferencesinpreyabundanceandpreybiomass betweenlocalities,withPalmaandManzanillashowingthehighestdensities andLebrija andMairenathelowest.Althoughthisdidnotaccountfordifferencesincomposition of croptypesamongtheselocalities,samplingsinbothcropsandfield-marginswerecon- sistent inthisresult(withtheexceptionofArahal,which showedintermediatedensitiesof preyandpreybiomassincrops,butthehighestinmargins).Thissuggestsgeneralized differencesinpreyabundanceamonglocalities.
Oncerandomvariabilitybelonging tolocalityandyearwasincludedinGLMM,crop typewasthemostinfluentialvariableonpreydensityandpreybiomassintransects.We foundthatsmallremainingpatchesofsemi-natural habitatssuchasfallowlandsand grasslands heldthehighestdensitiesofpreyandpreybiomass,whilethelowestdensities werefoundinolivegroves.Themaincropsinthearea,cerealsandsunflowers, showed intermediatevaluesofpreydensityandbiomass,butcerealshadhigherpreydensitiesthan sunflowers,andpreyaccessibilityisprobably muchlowerinthelatter(becauseofgreater vegetation coverandheight; seeTable3).Theseresultsconfirmthe indirect evidencesofa previousstudyonlesserkestrelalsosuggestingthathighpreydensitiesinsemi-natural habitatswouldexplaintheirpositive selectionbyforaging lesserkestrels, aswellasthe negativeselectionof othercropssuch assunflowerand olivegroves(Dona´zaretal.1993). Infact,ourresultsindicatethatcroppreferenceforforagingbylesserkestrels(grass- lands[cereals[sunflowersandolivegroves)asfoundina radio-trackingstudyby
Table3 Vegetationcover
(in%)andmediandevelopment
(categories0–5)ofsampled croptypes
Dona´zaretal.(1993)inoneofourstudylocalities(Mairena) isinaccordancewithmean relativepreyabundance (Table1),butalsowithparameterestimatesofGLMMforthese crops.Although oursamplesizepreventsustoevaluatewhetherotherlocallyimportant cropssuchastraditionalvineyards orcottoncropscouldbesuitableforforaging kestrels, ourresultspointedinthisdirection(seealsoIsenmannandDebout2000).Nonetheless furtherstudiesshouldexploreindetailthesepreliminaryresults.
Cropdevelopment alsoexplaineddifferencesindensitiesofpreyandpreybiomassin transects.Inthestudyarea,cotton wastheonlycropstillgrowingatthetimeofsampling. For theothers,crop developmentconstitutesameasureoftheirpreviousgrowth.Although alllocalitiesbelongtothesecondbestcategory(outof8)ofsoilproductivity(except Moro´nthatwasinthethird;AndalusianGovernment),differentsoiltypes, slopesandpast agriculturalpractices(nitrification, croprotation,etc.)probablydeterminedifferencesin individual fieldproductivity.Thus,thepositiveassociationofpreyandpreybiomasswith cropdevelopmentwasprobablyduetohigherproductivityofthosefields showingmore developedcropsconsideringthegeneralizedherbivoredietofOrthoptera(Hendriksetal.
1999;Olmo2002).
Afterconsideringtheeffectofcroptypeanddevelopment,themeansizeoffields in eachlocalityexplains partoftheremaining variabilityinpreydensityandbiomass. Althoughcroptypescomposition differsbetweenlocalitiesandsomecroptypesmay potentiallycontributemorethanothers tomeanfieldsize,digitalizedmapsofarablefields intheareasuggestthattherearegeneralizeddifferencesinfieldsizeamongthestudy localities.Theincreaseinmeanfieldsizehastakenplacemainlybytheaggregationof smallfields, whichreducesedgesandlinearelementsassociatedtofield limits,butalso removessmallpatchesofnon-arablelandpreviouslypresentinthelandscape(Ferna´ndez etal.1992;BaesslerandKlotz2006).Thus,thenegativeinfluenceofincreasingmeanfield sizeonthedensityofpreyandpreybiomassin30-mintransectssuggestapositiveeffectof habitatheterogeneityandhigheravailabilityofmarginsonpreyabundance inarable landscapes(MarshallandMoonen2002;Bentonetal.2003).Thisisprobablybecause marginsactaspreyreservoirs(Dona´zaretal.1993;Brickleetal.2000;Weibulletal.
2003)andcorridorsthatallowfortherecolonizationoflocallyextinctpatchesinthe farmedlandscape (Berggrenetal.2002).Theyalsobuffertheharmful effectsofbiocide applicationsonarthropodpopulations(LemkeandPoehling2002;seealsoUrsu´aetal.
2005). Inaddition, models forILAsindicatedthatnotonlyincrops,butalsoinmargins, preyspecieswerenegativelyinfluenced byincreasingfieldsizes.Higherconnectivity margin-field-margininlocalitieswithsmallerfields,couldexplainthisresult(Berggren etal.2002).
Themainfactorexplainingpreydensityandespeciallypreybiomassinmarginsand edgeswasvegetationcover,probablyenhancedbymoreproductivesoils(Kleijnetal.
2006),lessherbicideapplication(Hartetal.2006)and/orbetterconservationofthese habitats.Afterintroducingvegetationcover,onlyanegativeinfluence ofmeanfield-size andsomecorrection forwindydays(probably becausethereducedabilityoftheobserver todetectOrthopterawithwind)werefound.
Conclusionsandmanagementimplications
Prey abundance for lesser kestrels in arable landscapes shows a complex pattern accountingforthealreadydeduced influenceofcroptypecomposition(Bustamante1997; Franco etal.2004),alsohighlightingtheimportanceofcropdevelopment,spatialfeatures
ofthelandscape(meanfieldsize),andquality ofmargins(vegetationcover).Thissuggests thatrelyingsolelyonthecompositionofcroptypesinagivenlocalityasameasureofits suitabilityforthelesserkestrelmayrenderconfusing results.Valuesofpreyandprey biomassfoundinArahalexemplifytheimportanceofthesefactors.Inthislocalityonly
25%ofthesurface withinthe3-kmradiusissuitable forforaging (duetothehigh proportionofunsuitableolive grovesandsunflowersoccupyingtherest).Inaddition,mean fieldsizewasrelativelyhigh(Fig1)somarginsarenotespeciallyabundant.However, field marginsinthislocalityshowedthehighestdensitiesofpreyandpreybiomass, probablydueto their highdevelopmentorconservationstatus.Althoughvegetationin well developed marginsmaydecreasepreyaccessibility,highselectionofmarginsbyforaging lesserkestrelsindifferentareas(Dona´zaretal.1993;Ursu´aetal.2005)suggestthatthey areabletocircumventthesehuntingdifficulties.Nonetheless,preyaccessibilitycouldbe animportantissueforforagingkestrelsthatshouldbeconsidered.Forinstance,Dona´zar etal.(1993)reportedhigherforagingefficiencyofkestrelsin stubblesthanincerealfields, andalthoughthemajorityofcerealfieldswerealreadyharvestedwhensampled(141outof
182;seeTable3),ourresultsindicatehigherpreydensityincerealfields thaninstubble (WilcoxonranktestZ=-2.5626,P=0.0104).Similarly,lesserkestrelsshowedhigher huntingpreferencesforgrazedfallowsinSouthernPortugal(Francoetal.2004)because ofthehigherpreyaccessibility.Inthisrespect,rearinglivestock (mainlysheepandgoats) isaminoractivityinourstudylocalities(range=0–0.23animalsperha;INE1999)and itsinfluenceonthesampledgrasslandsandfallowsisminimal.
Thepositiveeffectofbothquantityandqualityofmarginsonpreyabundance and biomassfortheendangeredlesserkestrelhighlightstheirroleaskey-componentofthe farmedlandscapeaspreviousstudiesalsosuggested(e.g.Wilsonetal.1999;Marshalland Moonen2002;Ursu´aetal.2005).Thus,weencouragemanagerstoincreasetheconser- vationeffortsorientedtoensurethelong-termpersistenceofsemi-natural habitatsinthe farmedlandscape. Thisshouldincludetherecoveryandenhancementofgrassymargins, whichinthestudyareahavedisappearedbetweenmanyindividualfields. Thismeasure (creatingnewfieldbordersorincreasingtheirwidth)hasbeenidentifiedasadesirable agriculturalpracticetobenefitpseudo-steppebirdsthathasneverbeenappliedinSpain (LlusiaandOn˜ate2005).Nonetheless,furtherresearcheffortsshouldbecarriedoutto evaluatetheimportanceofcover,andspatialdistribution oftheseenvironmental-friendly measures(margins,in-fieldstrips,etc.),theireffectsonfarmlandbiodiversity,andtheir costforfarmers.Suchkindofscientific, multidisciplinaryevaluationsarerequiredto determinewhetheragri-environmentalschemesarereallyusefulandcost-effective(Kleijn etal.2006;Watzoldetal.2006).
Acknowledgements A.Oses,M.Calvo,J.Rengel,andM.Garc´ıacontributedtoinsectcollectionand determination.A.BadihandS.Gadoumtrainedusinsamplingmethods,andOrthopteradetermination. AndalusianGovernprovided land-useinformation. ThisresearchwasfundedbytheREN2001-2134/GLO projectoftheCICYTandFEDERfundsoftheEU.Thefirst authorwassupportedbyapredoctoral scholarshipfromtheSpanishMinisteriodeCienciayTecnolog´ıa.
References
AebischerNJ,EwaldJA(2004)ManagingtheUKGreypartridgePerdix perdixrecovery:population change,reproduction,habitatandshooting. Ibis146:181–191.doi:10.1111/j.1474-919X.2004.00345.x BadihA,HidalgoJ,BallestaM,RuanoF,TinautA(1997)Distributionandphenologyofacommunityof Orthoptera(Insecta)inaduneecosystem ofthesoutheastern IberianPeninsula. ZoolBaetica8:31–42
BaesslerC,KlotzS(2006)Effectsofchangesinagriculturalland-useonlandscapestructureandarableweed vegetationoverthelast50years.AgricEcosystEnviron115:43–50.doi:10.1016/j.agee.2005.12.007
BentonTG,Vickery JA,WilsonJD(2003)Farmland biodiversity:ishabitatheterogeneity thekey?Trends
EcolEvol18:182–188.doi:10.1016/S0169-5347(03)00011-9
Berggren A,BirathB,Kindvall O(2002)EffectofCorridorsandhabitatedgesondispersalbehavior, movementrates,andmovementanglesinRoesel’sbush-cricket(Metriopteraroeseli).ConservBiol
16:1562–1569.doi:10.1046/j.1523-1739.2002.01203.x
BrickleNW,Harper DGC,AebischerNJ,CockayneSH(2000)Effectsofagriculturalintensificationonthe breedingsuccessofcornbuntingsMiliariacalandra. JApplEcol37:742–755.doi:10.1046/j.1365-
2664.2000.00542.x
BustamanteJ(1997)PredictivemodelsforLesserKestrelFalconaumanni.Distribution,abundance,and extinctioninSouthernSpain.BiolConserv80:153–160.doi:10.1016/S0006-3207(96)00136-X
ClementeME,Garc´ıaMD,PresaJJ(1987)Clavedelosge´nerosdesaltamontesibe´ricos(Orthoptera,
Caelifera).UniversidaddeMurcia,Murcia
CrawleyMJ(2002)Statisticalcomputing:anintroductiontodataanalysisusingS-plus.Wiley,Chichester
Donald PF,GreenRE,HeathMF(2001)Agriculturalintensificationandthecollapse ofEurope’s farmland birdpopulations.ProcRSocLondB268:25–29.doi:10.1098/rspb.2000.1325
DonaldPF,PisanoG,RaymentMD,PainDJ(2002)TheCommonAgriculturalPolicy,EUenlargementand the conservation of Europe’s farmland birds. Agric Ecosyst Environ 89:167–182. doi:10.1016/
S0167-8809(01)00244-4
Dona´zarJA,NegroJJ,HiraldoF(1993)Foraginghabitat selection,landusechangesandpopulationdecline intheLesserKestrel.JApplEcol30:515–522.doi:10.2307/2404191
Ferna´ndez R,Mart´ınA,OrtegaF,Ale´sEE(1992)Recentchangesinlandscapestructureandfunctionina
MediterraneanregionofSWSpain(1950–1984).LandscEcol7:3–18.doi:10.1007/BF02573953
FrancoAMA,CatryI,Sutherland WJ,PalmeirimJ(2004)Dodifferenthabitatpreferencesurveymethods producethesameconservationrecommendationsforlesserkestrels?AnimConserv7:291–300.doi:
10.1017/S1367943004001465
FullerRJ,GregoryRD,GibbonsDW,MarchantJH,WilsonJD,BaillieSR,CarterN(1995)Population declines andrangecontractionsamonglowland farmlandbirdsinBritain. Conserv Biol9:1425–1441. doi:10.1046/j.1523-1739.1995.09061425.x
GangwereSK,deViedmaMG,LlorenteV(1985)Librorojodelosorto´pterosibe´ricos.ICONA,Madrid
GottschalkE,GriebelerEM,WaltertM,MuhlenbergM(2003)PopulationdynamicsintheGreyBush CricketPlatycleisalbopunctata(Orthoptera:Tettigoniidae)—whatcausesinterpopulationdifferences? JInsectConserv7:45–58.doi:10.1023/A:1024706030658
GregoryRD,NobleDG,CustanceJ(2004)Thestateofplayoffarmlandbirds:populationtrendsand conservation status of lowland farmland birds in the United Kingdom. Ibis 146:1–13. doi:
10.1111/j.1474-919X.2004.00358.x
HartJD,MilsomTP,FisherG,WilkinsV,MorebySJ,MurrayAWA,RobertsonPA(2006)Therelationship betweenyellowhammer breedingperformance,arthropodabundanceandinsecticideapplicationson arablefarmland.JApplEcol43:81–91.doi:10.1111/j.1365-2664.2005.01103.x
HarzK(1969)DieOrthopterenEuropas.TheHague,TheNetherlands
HendriksRJJ,deBoerNJ,vanGroenendaelJM(1999)Comparing thepreferences ofthreeherbivore specieswithresistancetraitsof15perennialdicots:theeffectofphylogeneticconstraints.PlantEcol
143:141–152.doi:10.1023/A:1009832621516
HiraldoF,NegroJJ,DonazarJA, GaonaP(1996)A demographicmodelforapopulationof theendangered
LesserKestrelinSouthernSpain.JApplEcol33:1085–1093.doi:10.2307/2404688
IsenmannP,DeboutG(2000)VineyardsharbourarelictpopulationofLesserGreyShrike(Laniusminor)in
MediterraneanFrance.JOrnithol141:435–440
INE(1999)Censoagrario1999.INE(InstitutoNacionaldeEstad´ıstica),Madrid
IUCN(1983)TheIUCNinvertebratereddatabook.IUCN,Gland
KleijnD,SutherlandWJ(2003)HoweffectiveareEuropeanagri-environmentschemesinconservingand promotingbiodiversity?JApplEcol40:947–969.doi:10.1111/j.1365-2664.2003.00868.x
KleijnD,BaqueroRA,CloughY,DiazM,EstebanJ, FernandezF,GabrielD,HerzogF, HolzschuhA,Johl
R,KnopE,KruessA,MarshallEJP,Steffan-DewenterI,TscharntkeT,VerhulstJ,WestTM, YelaJL (2006) Mixedbiodiversitybenefitsofagri-environmentschemesinfiveEuropeancountries.EcolLett
9:243–254.doi:10.1111/j.1461-0248.2005.00869.x
LemkeA,Poehling H-M(2002)Sownweedstripsincerealfields:overwinteringsiteand‘‘source’’habitat forOedothoraxapicatus(Blackwall) andErigoneatra(Blackwall) (Araneae:Erigonidae).Agric EcosystEnviron90:67–80.doi:10.1016/S0167-8809(01)00173-6
LlusiaD,On˜ateJJ(2005) Aretheconservationrequirementsofpseudo-steppebirdsadequatelycovered by
Spanishagri-environmentalschemes?Anex-anteassessment.Ardeola52:31–43
MarshallEJP,MoonenAC(2002)FieldmarginsinnorthernEurope:theirfunctionsandinteractions with agriculture.AgricEcosystEnviron89:5–21.doi:10.1016/S0167-8809(01)00315-2
MathsoftI(2000)S-Plus2000.Seattle
MattisonEHA,NorrisK(2005)Bridgingthegapsbetweenagriculturalpolicy,land-useandbiodiversity.
TrendsEcolEvol20:610–616.doi:10.1016/j.tree.2005.08.011
McCracken DI,FosterGN,KellyA(1995)Factorsaffectingthesizeofleatherjacket(Diptera:Tipulidae) populationsinpasturesinthewestofScotland.ApplSoilEcol2:203–213.doi:10.1016/0929-1393 (95)00048-P
MorrisAJ,WilsonJD,WhittinghamMJ,BradburyRB(2005)Indirecteffectsofpesticidesonbreeding
yellowhammer (Emberiza citrinella). Agric Ecosyst Environ 106:1–16. doi:10.1016/j.agee.2004.
07.016
NegroJJ(1997)LesserKestrelFalconaumanni.In:BirdsoftheWesternPalearticUpdate,vol1.Oxford
UniversityPress,Oxford,pp49–56
NegroJJ,Dona´zarJA,HiraldoF(1993)HomerangeofLesserKestrelduringthebreedingseason.In: NichollsMK,ClarkeR (eds)Biologyandconservationofsmallfalcons.The HawkandtheOwlTrust,
Canterbury,pp144–150
NewtonI(2004)Therecentdeclinesoffarmland birdpopulationsinBritain:anappraisalofcausalfactors andconservationactions.Ibis146:579–600.doi:10.1111/j.1474-919X.2004.00375.x
OlmoJM(2002)AtlasdeOrto´pterosdeCatalunya.GeneralitatdeCatalunya,Barcelona
PeetNB,Gallo-OrsiU(2000)ActionplanforthelesserkestrelFalconaumanni.CouncilofEuropeand
BirdLifeInternational,Cambridge
PinheiroJC,BatesDM(2002)Mixed-effectmodelsinSandS-plus.Springer,NewYork
PottsGR(1986)ThePartridge.Collins,London
RodriguezC(2004)FactoresambientalesrelacionadosconelexitoreproductivodelCernicaloPrimilla.
Cambioclimaticoeintensificacionagraria.Dissertation,UniversityofSalamanca
RodriguezC,BustamanteJ(2003)Theeffectofweatheronlesserkestrelbreedingsuccess:canclimate changeexplainhistoricalpopulationdeclines?JAnimEcol72:793–810.doi:10.1046/j.1365-2656.
2003.00757.x
RodriguezC,JohstK,BustamanteJ(2006)Howdocroptypesinfluencebreedingsuccessinlesser kestrels throughpreyqualityandavailability?Amodellingapproach.JApplEcol43:587–597.doi:10.1111/
j.1365-2664.2006.01152.x
Ursu´aE,SerranoD,TellaJL(2005)Doeslandirrigationactuallyreduceforaginghabitatforbreedinglesser kestrels?Theroleofcroptypes.BiolConserv122:643–648.doi:10.1016/j.biocon.2004.10.002
VoisinJ-F(1980)Reflexionsaproposd’unemethodesimpled’echantillonagedespeuplementsd’orthop- teresenmilieuouvert.Acrida9:159–170
VoisinJ-F(1986)Unemethodesimplepourcaracteriserl’abondancedesOrthopteresenmilieuouvert.
L’Entomologiste42:113–119
WatzoldF,DrechlerM,ArmstrongCW,Baumgartner S,GrimmV,HuthA,PerringsC,Possingham HP, ShogrenJF,SkonhofA,Verboom-VasiljevJ,WisselC(2006)Ecological-economicmodelingfor biodiversitymanagement:potential,pitfalls,andprospects.ConservBiol20:1034–1041
WeibullA-C,OstmanA˜,GranqvistA(2003)Speciesrichnessinagroecosystems: theeffectoflandscape, habitatandfarmmanagement.BiodiversConserv12:1335–1355.doi:10.1023/A:1023617117780
WillottSJ(1997)ThermoregulationinfourspeciesofBritishgrasshoppers(Orthoptera:Acrididae).Funct
Ecol11:705–713.doi:10.1046/j.1365-2435.1997.00135.x
WilsonJD,MorrisAJ, ArroyoBE, ClarkSC, BradburyRB (1999)Areviewoftheabundanceanddiversity ofinvertebrateandplantfoodsofgranivorous birdsinnorthernEuropeinrelationtoagricultural change.AgricEcosystEnviron75:13–30.doi:10.1016/S0167-8809(99)00064-X