Behavioral correlates of supplementary feeding of wildlife: Can general conclusions be drawn?

Behavioral correlates of supplementary feeding of wildlife:

Can general conclusions be drawn?

Sam M.J.G. Steyaert

, Jonas Kindberg

, Klemen Jerina

, Miha Krofel

, Matija Stergar

, Jon E. Swenson

, Andreas Zedrosser

aDepartmentofEcologyandNaturalResourceManagement,NorwegianUniversityofLifeSciences,NO-1432,Ås, Norway

bDepartmentofWildlife,Fish,andEnvironmentalStudies,SwedishUniversityofAgriculturalSciences,SE-90183 Umeå,Sweden

cDepartmentofForestryandRenewableForestResources,BiotechnicalFaculty,UniversityofLjubljana,Veˇcnapot 83,SI1000Ljubljana,Slovenia

dNorwegianInstituteforNatureResearch,NO-7485Trondheim,Norway

eDepartmentofEnvironmentalandHealthStudies,TelemarkUniversityCollege,NO-3800,Bø,Norway

fInstituteofWildlifeBiologyandGameManagement,UniversityofNaturalResourcesandLifeSciences,A-1180, Vienna,Austria

Received24April2014;accepted5October2014 Availableonline12October2014

Abstract

Supplementaryfeedingisacommon,butcontroversial,toolinwildlifemanagement,becauseit canbenefitbothhumans andwildlife(e.g.,increasedwildlife densities),buthascertaindownsides(e.g.,increased diseasetransmission).Forspecies thatareofteninvolvedinhuman-wildlifeconflicts,twoopposingparadigmswithrespecttosupplementaryfeedingexist,i.e., (i)that supplementaryfeedingisefficienttolureanimalsawayfromundesiredplaces(i.e.,diversionaryfeeding;hypothesis 1),and(ii)thatsupplementaryfeedingstimulates‘nuisance’behavior(i.e.,increasedtoleranceforhumansandselectionfor humanfacilities;hypothesis2).Weformulatedanalternativehypothesis(hypothesis3);i.e.,thatbehavioralvariationamong individualsdilutespopulation-wide,generalpatterns withrespecttosupplementaryfeeding.BasedonGPS relocationdata andresource selection functions, we show that neither of the two opposing management paradigms (hypothesis 1 and2) holdinaparticularly‘conflictrich’species,thebrownbear(Ursusarctos),becauseindividualvariationinselectionbehavior withrespecttosupplementaryfeedingdilutedpopulation-widepatterns(hypothesis3),evenunderverydifferentenvironmental contexts(Swedenvs.Slovenia;i.e.,differenthumanandbearpopulationdensity,historyandintensityofsupplementaryfeeding, topography,etc.).Ourresultsemphasizethatindividualvariationisanimportantcomponentofbehavioralecologyandshould beconsideredinwildlifemanagementandconservation.

Zusammenfassung

Ergänzende Fütterungen sind ein häufiges,aber umstrittenesMittel beim Managementvon Wildtieren, weil siesowohl demMenschenalsauchdenWildtierennützenkönnen(z.B.durcherhöhteSiedlungsdichtenderWildtiere),aberesgibtauch Nachteile(z.B.erleichterteÜbertragungvonKrankheiten).BeiArten,diehäufigmitdemMenscheninKonfliktgeraten,gibt eszweientgegengesetzteAnsichten:(i)dasssichmitFütterungenTiereeffektivvonunerwünschtenPlätzenweglockenlassen

∗Correspondingauthor.Tel.:+4764965385.

E-mailaddresses:sam.steyaert@nmbu.no,samsteyaert@gmail.com(S.M.J.G.Steyaert).

http://dx.doi.org/10.1016/j.baae.2014.10.002

1439-1791/©2014TheAuthors.PublishedbyElsevierGmbH.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/

licenses/by-nc-nd/3.0/).

(ablenkendeFütterung; Hypothese1) und (ii) dassFütterungen Problemverhaltenanregt (zunehmendeToleranzgegenüber demMenschenundAufsuchenvonmenschlichenEinrichtungen;Hypothese2).WirformulierteneinealternativeHypothese (Hypothese3),diebesagt,dassdieVariabilitätimIndividualverhaltenallgemeine,diePopulationumfassendeVerhaltensmuster hinsichtlichderFütterungenabschwächt.MithilfevonGPS-LokalisationundRessourcenauswahl-Funktionenzeigenwirfür den,,konfliktreichen”Braunbären(Ursusarctos),dasskeinsderentgegengesetztenManagementmodelle(Hypothesen1und2) zutrifft.DerGrundist,dassindividuelleVariationimAuswahlverhaltenbezüglichderFütterungenpopulationsweiteReaktions- musterabschwächte(Hypothese3)unddiessogarinstarkunterschiedlichenUmweltkontexten(SchwedenundSlowenien,d.h., beiunterschiedlichenBären-undBevölkerungsdichten,unterschiedlicherGeschichteundIntensitätderFütterung,Topographie usw.).UnsereErgebnisseunterstreichen,dassindividuelleVariationeinewichtigeKomponentederVerhaltensökologieistund beimWildtiermanagementund-schutzberücksichtigtwerdensollte.

© 2014 The Authors. Published by Elsevier GmbH. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Keywords: Behavioralstrategies;Brownbear;Habituation;Human-wildlifeinteractions;Individualvariance;Supplementaryfeeding;Ursus arctos;Wildlifemanagement

Introduction

Supplementary foods are provided towildlife wherever humans and wildlife coexist (Beckmann &Berger 2003), eitherintentionallyformanagementorrecreationalpurposes, or unintentionally,for exampleas garbage.Supplementary feedingcaninfluencewildlifebehavior(e.g.,movementpat- terns,reproductivestrategies),demography(e.g.,population growth),andlifehistory(e.g.,reproduction),andmayalter communitystructures(e.g.,speciesdiversity)(Boutin1990;

Robb, McDonald, Chamberlain, & Bearhop 2008). These potentialinfluencescanbeappliedtowildlifemanagement and conservation. For example, supplementary feeding is used to increase the productivity and density of wildlife populations (Boutin 1990), or to support the recovery of endangeredspecies, such as the kakapo (Strigopshabrop- tilus)(Clout,Elliott,&Robertson2002),ortheIberianlynx (Lynxpardinus)(López-Bao,Rodríguez,&Palomares2008).

Supplementaryfeedingisoftenusedtoredistributewildlife populations(i.e.,diversionaryfeeding)toreduceforestdam- age (Ziegltrum &Russell 2004) or traffic collisions (Rea 2003).Supplementaryfeedingisalsoappliedforrecreational andhuntingpurposes,i.e.,toattractelusivespeciestospe- cificplacesforobservationorharvest(i.e.,baiting)(Bischof, Fujita,Zedrosser,Söderberg,&Swenson2008)ortoimprove trophy size (e.g., antlers in Cervidae) (Putman & Staines 2004).

However,supplementaryfeedingcanalsohaveundesired effects on wildlife and habitats (Boutin1990; Robb et al.

2008),andisthereforeconsideredasacontroversialpractice (Putman&Staines2004).Undesiredpotentialeffectsinclude elevated risk for disease transmission or parasite burdens (Putman & Staines 2004), altered sex ratios (Clout et al.

2002), potential risks to human health (Kavˇciˇc, Adamiˇc, Kaczensky, Krofel,& Jerina2013), concerns about selec- tive harvest at bait sites (e.g. when certain sex and age classesmakedisproportionateuseofbaitsites)(Bischofetal.

2008), increased interspecific predation (Cortés-Avizanda,

Carrete, Serrano, & Donázar 2009), and habitat degrada- tion(Putman&Staines2004).Anadditionalconcernisthat animalsmayrelatesupplementaryfeedingwithhumans(i.e., becomefood-conditioned)andlosetheir‘normal’wariness (i.e.,habituation)towardspeople(Woodroffe,Thirgood,&

Rabinowitz2005).Animalswithincreasedtolerancetowards humans maybecome a‘nuisance’,andcan—dependenton the species—be a threat to human safety. Such species include elephants (O’Connell-Rodwell, Rodwell, Rice, &

Hart2000),bears(Elfström,Zedrosser, Støen,&Swenson 2014),felids(Saberwal,Gibbs,Chellam,&Johnsingh1994), and canids (Orams 2002). The potential tocondition ani- mals on certain foods and/or habituate them to humans alsohighlightsthefactthatsupplementaryfeedingmaycut both waysas amanagement tool, andraisesthe question:

does supplementary feeding facilitate nuisance behavior, or can it efficiently redistribute wildlife in relation to humans?

Here,wetestifandhowselectionforsupplementaryfeed- ing correlateswithmanagementefficacy(i.e.,diversionary feeding)andpotentialnuisancebehaviorina‘conflict-rich’

species, the brown bear (Ursus arctos). Brown bears are large omnivorous opportunists and are often perceived as a‘problemspecies’becausetheysometimesdamage prop- erty and kill livestock, and occasionally attack and kill people(Elfström,Zedrosser,Støen,etal.2014).Supplemen- tary feedingis commonly used as awildlife management tool, for example to bait animals for hunting purpose (i.e., population regulation) (Bischof et al. 2008), or to lureanimalsawayfromundesiredplaces(i.e.,diversionary feeding) (Elfström, Zedrosser, Støen, et al. 2014). How- ever, supplementary feeding is also generally presumed to stimulate‘nuisance’ behavior inbears (Herrero, Smith, DeBruyn,Gunther,&Matt2005;Elfström,Zedrosser,Støen, et al.2014).Thedichotomous perceptionsamong wildlife biologists, managers, and the general public on the func- tionality of supplementary feeding is hotly debated, and canleadtoopposingmanagementapproaches.Forexample,

Fig.1. Conceptualrepresentationthethreehypothesestoevaluatebehavioralcorrelatesbetweentheselectionforsupplementaryfeedingand theselectionforhumanfacilitiesinbrownbears.Hypothesis1(H1)predictsthatbearsgenerallyselectforsupplementaryfeedingandselect againsthumanfacilities.H2predictsthatselectionforsupplementaryfeedingsitesispositivelycorrelatedwithselectionforhumanfacilities.

H3postulatesthatthevarianceinselectionforsupplementaryfeedingisnotassociatedwithselectionforhumanfacilities,andthatvariance inselectionforsupplementaryfeedingsitesismostlyexplainedbyvariationinbehavioramongandwithinindividuals.

supplementaryfeedingbrownbearsisstronglydiscouraged inseveralcountries,regions,ornationalparks(e.g.,Scandi- navia,YellowstoneNationalPark,DenaliNationalPark,etc.), whereaslegallymandated(e.g.,Sloveniauntil2004,Croatia), orrecommendedinothers(Craighead,Sumner,&Mitchell 1995; Robbins, Schwartz, & Felicetti 2004; Kavˇciˇc et al.

2013).

Weformulated three hypotheses to addressour general objective.Hypothesis1postulatesthatdiversionaryfeeding efficiently mitigates conflicts between bears and humans, and predicts that (a) selection for supplementary feeding sitescorrelatesnegativelywithselectionforhumanfacilities and (b) that the majority of bears select for supplemen- tary feeding sites (Fig. 1.). Hypothesis 2 postulates that supplementaryfeedingstimulatespotentialnuisancebehav- ior, and predicts that selection for supplementary feeding sitescorrelates positively withselection for human facili- ties(Fig.1).Becauseindividualbehavioraldifferencesare commonamongmammals(Wolf&Weissing2012),hypoth- esis3postulatesthatindividualvarianceinbehaviordilutes population-wideselectionpatterns,andpredictsthat selec- tionforsupplementaryfeedingsitesdoesnotcorrelatewith selectionforhumanfacilities(Fig.1.).Wetestedourhypothe- sesintwobrownbearpopulations(i.e.,SwedenandSlovenia) withverydifferentenvironments(densityofbears,humans, andsupplementaryfeedingsites)tocontrolforcontingencies andtorevealgeneralitiesinbehavioralcorrelatesinrelation tosupplementaryfeeding.Becausehuman-bearconflictsare oftensuggestedtocorrelatewiththeannualvariationinthe availabilityofnaturalfoods(lowavailability∼highconflict rates) (Mattson, Blanchard, & Knight1992), we also test theimportanceofannualvariationinsupplementaryfeeding siteselection.Becausemoredominant sexandageclasses candominatesupplementaryfeedingsites(Craigheadetal.

1995),wealso evaluatedapotential effectofreproductive status(i.e.,acombinationofsexandageclasses,andpres- ence or absence of young) on supplementary feeding site selection.

Material and methods

The Swedish study area encompassed approximately 13,000km² of intensivelymanagedboreal forestin south- centralSweden(61^◦N,15^◦E).Thehumanpopulationdensity (4.1– 7.1inhabitants/km²)isoneofthe lowestwithin the Europeanbrownbearrange,andthebearpopulationdensity isapproximately 30bears/1000km² (Bellemain,Swenson, Tallmon,Brunberg,&Taberlet2005).Supplementaryfeed- ingwasextensivelyusedtobaitbearsforhuntinguntil2001, whenitwasbanned.Weweregrantedpermissiontomain- taintwoexperimentalsupplementaryfeedingsitesbetween 2008and2012,whichwererestockedweeklywith5kgof gamemeatorfish,5kgofcorn,5kgofsugarbeetpulp,and 5lofmolasses(Zedrosser,Steyaert,Brunberg,Swenson,&

Kindberg2013).Approximately1.5%ofallharvestedbears inSwedenareconsideredproblembears.Problembearsare generallyyoungerthannon-problembearsinSweden,and theoccurrenceofproblembearsisnotrelatedtobodycon- ditioninbears(i.e.,andproxyforfoodavailability)orbear populationdensity(Elfström,Zedrosser,Jerina,etal.2014).

The Slovenian study area encompassed approximately 3800km² of extensively managed forest in south-central Slovenia(45^◦N,14^◦E).Thehumanpopulationdensityaver- ages54inhabitants/km²,andthebearpopulationcanlocally reachextremelyhighdensities(>400bears/1000km²).Sup- plementaryfeedingsitesoccuratdensitiesof1/400–700ha andhavebeenmaintainedwithcontinuoussuppliesoflarge amounts(annualaverage:70–280kg/km²)ofpredominantly cornandcarrionforseveraldecadesinsomeareas(Kavˇciˇc etal.2013).About14%ofallharvestedbearsinSloveniaare consideredproblembears.AsinSweden,however,Slovenian problembearsaregenerallyyoungerthannon-problembears, andtheincidenceofproblembearsisnotrelatedtobodycon- ditionorbearpopulationdensity(Elfström,Zedrosser,Jerina, etal.2014).

GPS-relocationdata

WecapturedandequippedbrownbearswithGlobalPosi- tioningSystemcollars(GPS;VectronicAerospaceGmbH) byaerialdarting withanimmobilization drugfromaheli- copterbetween2008and2012inSweden,andusingAldrich footsnares(MargoSuppliesLtd.)anddartingwithanimmo- bilizationdrugfromthegroundbetween2005and2012in Slovenia.TheSwedishbearsweremonitored ona30-min GPSrelocationschedule,whereaswemonitoredSlovenian bearsonanhourlybasis.Fordetailsoncaptureandhandling, refertoArnemoetal.(2011)andJerina,Krofel,Stergar,&

Videmsek(2012).Weclassifiedbearsintoadultmales(males

≥5years),lonefemales(≥5years,withoutyoung),family groups(femaleswithyoung),subadultmales(<5years),and subadultfemales(<5yearswithoutyoung).

Statisticalanalysis

We used resource selection functions (RSFs) to quan- tifythebehaviorofindividualbearswithrespecttoafixed set of landscape variables that are considered important inanimalresourceselection, includingbears(i.e.,normal- izeddifferencevegetationindex,forestvs.nonforest,terrain ruggedness, and distance to supplementary feeding sites, settlements,singlehouses,androads) (Martinetal.,2010;

Steyaert, Kindberg, Swenson, & Zedrosser 2013). Refer to Appendix A for details on the spatial data. The GPS relocationsand aset of random pointrepresent ‘use’ and

‘availability’ of resources, respectively, andserved as the responsevariableinlogisticregressionmodels.Wesampled use/availabilityina1:1 ratio,andwithinthe annual100%

minimumconvexpolygonofeachbear-yearthatoverlapped atleastonesupplementaryfeedingsiteoutsidethedenning period.Theparameterestimates(β)andstandarderrors(SE) foreachlandscapevariableincludedinthemodel revealif variablesareselectedfor,selectedagainst,orarerelatively unimportantinanindividual’sresourceselection(i.e.,behav- ioral responses)(Boyce,Vernier, Nielsen, &Schmiegelow 2002).We multiplied the parameter estimates of the ‘dis- tanceto’variableswith−1tofacilitateinterpretation;such thatpositivevaluesindicatedselection,whereasnegativeval- uesindicated avoidance.We used the parameterestimates generated bythe individual RSFs toevaluate the relation- shipbetweensupplementaryfeedingsiteselection(i.e.,the responsevariable),selectionforlandscapevariables,aswell as bear-year specific data (i.e., bear ID, year, and repro- ductive status) with linear mixed-effectregression models (Dingemanse&Dochtermann2013).Weincluded‘bearID’

asarandomfactor.Weusedakaikeinformationcriteriadif- ferences(AICc)andweights (AICcw) toselect themost parsimoniousmodelamongsevencandidatesdefinedapri- ori (Table 1). We considered models with AICc values

>4asinconclusive(Burnham,Anderson,&Huyvaert2011).

Wevalidatedthemostparsimoniousmodelsbyplottingthe

modelresidualsversusthefittedvaluestoevaluatepotential heteroskedasticity (Zuur, Ieno, Walker, Saveliev, &Smith 2009). We used R 2.15.0 for all statistical analyses (R DevelopmentCoreTeam2013).

Results

We obtained relocation data and behavioral estimates from24and33bearsinSwedenandSlovenia,respectively (Table2).Weremovedbehavioralresponsestoroadsfromthe Sloveniandatasetinthesecondstep,becauseofcollinearity withsettlements(r=−0.67)(Table1).

The most parsimonious model was the ‘null’ model for both Sweden and Slovenia (AICcw=1). Individual bear variance explained 33% and 43% of the total vari- ance in supplementary feeding site selection in Sweden (1.59/4.91×10⁻⁸)andSlovenia(1.96/4.75×10⁻⁷),respec- tively.Allothercandidatemodelswereinconclusive(AICc

values>54.4,Table1).BearsinSloveniagenerallyselected for supplementary feeding sites (β=0.589×10⁻³; 95%

bootstrapped confidence limits 0.484 – 0.896×10⁻³);

whereas Swedish bears generally did not select for or against supplementary feeding sites (μ=0.045×10⁻³;

−0.013−0.105×10⁻³).Noheteroskedasticitywasappar- entinthemodelresiduals.

Discussion

We found that individual behavior best explained the strengthanddirectionofselectionforsupplementaryfeeding sites(hypothesis3),andsuggestthatvariationinindividual behaviordilutespopulation-widepatternsrelatedtosupple- mentaryfeedingsiteselection.Selectionforsupplementary feedingsiteswasnotrelatedtoreproductivestate,year,and selectionforhumanfacilitiesinbothSwedenandSlovenia (Fig.2.). Thisindicates that diversionaryfeedinghasonly lowconflict-mitigationpotential(hypothesis1),andthatsup- plementary feedinggenerallyisunlikelytocausenuisance behavior(hypothesis1)inbrownbears.Ourresultsarecon- sistentinbothcountries,althoughbearsinSloveniagenerally selected for supplementary feeding siteswhereasSwedish bearsdidnot.

Supplementary feeding is common in wildlife man- agement and conservation, and has received considerable attentionintheliterature(Putman&Staines2004;Robbetal.

2008).However,theimportanceofindividualbehaviorwith respect tosupplementaryfeedinghasbeen largelyomitted in research (Boutin 1990). Ecological research (includ- ingstudiesonsupplementaryfeeding)typically focuseson population-widegeneralitiesinbehaviorordemography,and rarelyontheindividuallevel(Boutin1990;Dingemanse&

Dochtermann2013).However,theimportanceofbehavioral types(forexampleshyvs.notshy),individualbehavior,and suites of correlatedbehaviors (i.e., behavioralsyndromes)

Table1. Modelselectiondiagnosticsforfivecandidatelinearmixed-effectregressionmodelstodeterminetherelationshipbetweenselection behaviorforsupplementaryfeedingsites,year,andreproductivestatus(Status),andselectionforenvironmentalcharacteristics(TRI=terrain ruggednessindex,forest/non-forest,NDVI=NormalizedDifferenceVegetationIndex,anddistancetoSettlements,Houses,andRoads)of GPS-markedbrownbearsinSweden(2008–2012)andSlovenia(2005–2010).Weincluded‘BearID’asarandomfactorinallcandidate models.AICcandAICcwindicatesecond-ordercorrectedAkaike’sInformationCriteriadifferencesandweights,respectively.√

indicates theinclusionofacertainmodelterminacandidatemodel.DistancetoroadswasneverincludedintheSloveniancandidatemodel,because ofcollinearitywithdistancetosettlements.BearIDwasalwaysincludedasarandomfactorontheintercept.

Candidatemodel Modelterm AICc AICcw

NDVI TRI Forest Year Status Distanceto: BearID

Settlements Houses Roads

Sweden Allinclusive √ √ √ √ √ √ √ √ √

202.77 0.0

Human √ √ √ √ √

78.11 0.0

Terrain √ √ √ √ √

142.35 0.0

Bear √ √ √

132.71 0.0

Status √ √

54.40 0.0

Year √ √

77.02 0.0

Null √

0 1.0

Slovenia Allinclusive √ √ √ √ √ √ √ √

243.43 0.0

Human √ √ √ √

87.10 0.0

Terrain √ √ √ √ √

139.26 0.0

Bear √ √ √

162.95 0.0

Status √ √

78.11 0.0

Year √ √

81.43 0.0

Null √

0 1.0

Table2. Descriptivestatisticsofbearrelocationdata inSweden andSloveniaasusedinthisstudy.

Slovenia Sweden

Nbears 33 24

Nfemales 18 10

Nmales 15 14

Nbearyears 43 54

Nrelocations(μ±SE) 2162±2089 3547±1177

Nrelocations(range) 221–6903 153–5351

Studyperiod 2005–2010 2008–2012

are becoming more prevalent in ecology and evolution (Dingemanse&Dochtermann2013).Wesuggestthatthere is considerable variation among individuals and selection strategies regarding selection for supplementary feeding sites, i.e.,someindividuals selectstrongly for supplemen- taryfeedingsites,whereasotherdonot.Thisselectionmay becorrelated(positivelyornegatively)withtheselectionfor humanfacilities by certainindividuals, butnot for others.

Thisdoesnotruleoutthatsupplementaryfeedingmaytrig- gernuisancebehaviorincertainindividuals;or,ontheother hand, that supplementary feeding may indeed be efficient tolurecertainindividualsawayfrom humanfacilities.We stress,however, that (i) the absence of a general relation- shipbetweenselectionforsupplementaryfeedingsitesand humanfacilitiesdoesnotwarranttheuseofsupplementary

feedingasanefficientmanagementtoolingeneral,and(ii) that the presumptionthatsupplementary feedinggenerally causesnuisancebehaviordoesnotnecessarilyhold.

Supplementaryfeedinggamespecieshasalongtradition inSlovenia(>100yearsincertainareas),andbearshaveyear- roundaccesstolargeamountsofhighenergysupplementary feed(i.e.,anannualaverageof70–280kg/km²,predomi- nantly corn).Kavˇciˇcetal.(2011)estimatedthat Slovenian bears obtain approximately 35% of their annual energy requirementsfromsupplementaryfeeding.Jerina,Jonozoviˇc, Krofel,&Skrbinˇsek(2013)suggested thatsuch long-term andintensivesupplementaryfeedingcanincreaseanareas’

carryingcapacity,whichcanexplaintheextremelyhighlocal bear densities in Slovenia (>40 bears/100km²) compared tootherEuropean (e.g.ItalianAlps,3 bears/100km²;Slo- vakian Carpathians,5– 11/100km²; Romania:9/100km²) (Swenson,Gerstl,Dahle,&Zedrosser2000;Rigg&Adamec 2007;Groff,Dlpiaz,Rizzoli,&Zanghellini2012)andinte- riorNorthAmericanpopulationaverages(<5bears/100km²) (Hilderbrand, Schwartz, Robbins, Jacoby, Hanley et al.

1999).OurresultthatSlovenianbearsgenerallyselectedfor supplementaryfeedingsiteswhereasSwedishbearsdidnot, suggeststhat long-term andintensive supplementaryfeed- ingcanconditionbearstosuchpredictable foodresources.

However, food conditioning does not necessarily result in nuisancebehavior(Elfström,Zedrosser,Støen,etal.2014).

AsimilarsituationaroseintheGreaterYellowstoneEcosys- tem,inwhichgrizzlybearswereconditionedtolarge-scaled

Fig.2. RelationshipbetweenbaitsiteselectionandselectionforsettlementsandhousesbyGPS-markedbrownbearsinSweden(2008– 2012)andSlovenia(2005–2010).Coefficientsofdetermination(r²)indicatethestrengthoftherelationships.Measuresofselection(β)were estimatedwithindividual-basedresourceselectionfunctions.Dashedlinesindicateβ=0.Whiskersindicateonestandarderroraroundthe parameterestimates.Selectionforhumanfacilitieswasnotincludedinthemostparsimoniousmodelsforbothcountries.

open pit garbage dumps that were maintained for several decades.Theseopenpitgarbagefacilitatedalargerbearpop- ulationthanthesysteminfactcouldhold.Suddenclosureof thesedumpsitesinthe1970sresultedinapopulationdecline throughincreasedmortalityanddispersal,andreducedrepro- ductiverates(Craigheadetal.1995).

Becausesupplementary feedingcanincrease population densities(Boutin1990),ithasalsothepotentialtopositively feed-backonhuman-wildlifeconflictrates(Woodroffeetal.

2005).InSlovenia,adensebearpopulation,coexistingwith arelativelydenseanddispersedhumanpopulationprobably resultsinhighhuman-bearconflictrates(∼14%ofallhar- vested bearsare consideredproblem bears). The Swedish bear population, on the other hand, has a relatively low density,coexistswithalowdensityandcentralizedhuman population; which probably results inlower conflict rates ascomparedtoSlovenia(∼1.5%ofallharvestedbearsare consideredproblembears).Wesuggestsupplementaryfeed- ingbearsmayincreasehuman-wildlifeconflictrates,butby increasingpopulationdensitiesratherthanthatsupplemen- taryfeedingwouldstimulatenuisancebehavior.

There appears to be no consensus among researchers whether or not supplementary feeding can mitigate con- flict (i.e., diversionaryfeeding and/or tofacilitateefficient harvest), or stimulate nuisance behavior. For example, in black bears, someauthors argue that diversionary feeding canbeefficient(Ziegltrum&Russell2004),withoutstim- ulating nuisance behavior (Rogers 2011), whereas others advocatetheopposite;i.e.thatsupplementaryfeedingshould not be practiced, because it results in problem behavior (Herrero 1985;Inglis 1992;Herrero etal.2005).The effi- cacy of supplementary feeding as a management tool has alsobeenquestionedforotherspecies(e.g.,wildboar(Sus scrofa)(Geisser, Reyer,&Krausman2004),moose (Alces alces) (Rea 2003), red deer (Cervus elaphus) (Putman &

Staines2004),andmaydependone.g.,naturalfood avail- ability,habitatquality,supplementaryfeedingintensityand history, spatial scale, etc. We suggest that in other sys- temswithoutapparentgeneralsupplementaryfeeding-related behavior, individual behavioral strategies may dilute gen- eralpopulation-widepatternswithrespecttosupplementary feeding.

Conclusions

Wefoundthatvariabilityamongindividualswasthesin- glemostimportantfactorexplainingthestrengthofselection for supplementary feeding sites by brown bears in Swe- denandSlovenia,twoenvironmentalextremesintermsof human density, bear density, and history and intensity of supplementaryfeeding.Non-explanatorymodelcomponents included year and reproductivestatus, and the individual- based selection coefficientsfor terrain ruggedness, NDVI, andforestedvs.non-forestedhabitat.In addition,selection for supplementaryfeedingsiteswas unrelatedtoselection forsettlements,buildings,androads.Ourresultsimplythat selectionforsupplementaryfeedingsitesdidnotgenerally stimulatenuisancebehaviorinbears,northatsupplementary feedingisanefficienttooltokeepbearsawayfrompeople.

Westress that individual variationin behavioralstrategies candilutepopulation-widebehavioralpatterns.Wesuggest thatmanagingsupplementaryfeedingsitescanhavedirect but nonetheless unexpected effects on a population (e.g., increased densities andpotential conflictrates; or popula- tiondeclinesafterreducingsupplementaryfeeding),andour resultsaddtothegrowingbodyofevidencethatindividual varianceis animportantcomponent of behavioralecology andshouldbeconsideredinwildlifemanagementandcon- servation.

Acknowledgments

WethanktheSwedishEnvironmentalProtectionAgency, Norwegian Directorate for Nature Management, Swedish Association for Hunting and Wildlife Management, the ResearchCouncil ofNorway,the AgencyforEnvironment ofSlovenia,andtheEuropeanUnionforfinancial support.

WethankMarcusElfströmandtwoanonymousreviewersfor theirvaluablecommentsonearlierdrafts.Thisisscientific articleno.177fromtheScandinavianBrownBearResearch Project.

Appendix A. Supplementary data

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

j.baae.2014.10.002.

References

Arnemo,J.M.,Evans,A.,Fahlman,Å.,Ahlqvist,P.,Andrén,H., Brunberg,S.,etal.(2011).Biomedicalprotocolsforfree-ranging brownbears,wolves,wolverinesandlynx.Evenstad,Norway:

HedmarkUniversityCollege.

Beckmann, J. P., & Berger, J. (2003). Rapid ecological and behaviouralchangesincarnivores:Theresponsesofblackbears

(Ursusamericanus) toalteredfood.JournalofZoology,261, 207–212.

Bellemain,E.,Swenson,J.E.,Tallmon,D.,Brunberg,S.,&Taberlet, P. (2005).Estimatingpopulationsizeofelusiveanimalswith DNAfromhunter-collectedfeces:Fourmethodsforbrownbears.

ConservationBiology,19,150–161.

Bischof, R.,Fujita, R., Zedrosser, A., Söderberg, A., & Swen- son,J.E.(2008).Huntingpatterns,banonbaiting,andharvest demographics ofbrownbears in Sweden.Journalof Wildlife Management,72,79–88.

Boutin,S.(1990).Foodsupplementationexperimentswithterres- trialvertebrates:Patterns,problems,andthefuture.Canadian JournalofZoology,68,203–220.

Boyce,M.S.,Vernier,P.R.,Nielsen,S.E.,&Schmiegelow,F.K.

A.(2002).Evaluatingresourceselectionfunctions.Ecological Modelling,157,281–300.

Burnham, K.,Anderson,D., &Huyvaert,K.(2011).AICmodel selectionandmultimodelinferenceinbehavioralecology:Some background,observations,andcomparisons. BehavioralEcol- ogyandSociobiology,65,23–35.

Clout,M.N.,Elliott,G.P.,&Robertson,B.C.(2002).Effectsof supplementaryfeedingontheoffspringsexratioofkakapo:A dilemmafortheconservationofapolygynousparrot.Biological Conservation,107,13–18.

Cortés-Avizanda, A., Carrete, M., Serrano, D., & Donázar, J.

A. (2009). Carcasses increase the probability of predation of ground-nesting birds:Acaveatregarding theconservation value of vulture restaurants. Animal Conservation, 12, 85–

88.

Craighead,J.J.,Sumner,J.S.,&Mitchell,J.A.(1995).Thegrizzly bearsofYellowstone:TheirecologyintheYellowstoneecosys- tem,1959–1992.WashingtonD.C:IslandPress.

Dingemanse, N. J., & Dochtermann, N. A. (2013). Quantify- ingindividual variationinbehaviour: Mixed-effectmodelling approaches.JournalofAnimalEcology,82,39–54.

Elfström,M.,Zedrosser,A.,Jerina,K.,Støen,O.-G.,Kindberg,J., Budic,L.,etal.(2014).Doesdespoticbehaviororfoodsearch explaintheoccurrenceofproblembrownbearsinEurope?The JournalofWildlifeManagement,78,881–893.

Elfström,M.,Zedrosser,A.,Støen,O.-G.,&Swenson,J.E.(2014).

Ultimateandproximatemechanismsunderlyingtheoccurrence ofbearsclosetohumansettlements:Reviewandmanagement implications.MammalReview,44,5–18.

Geisser,H.,Reyer,H.-U.,&Krausman.(2004).Efficacyofhunt- ing,feeding,andfencingtoreducecropdamagebywildboars.

JournalofWildlifeManagement,68,939–946.

Groff,C.,Dlpiaz,D.,Rizzoli,R.,& Zanghellini,P.(2012).Bear reportoftheforestryandwildilfedepartmentoftheautonomous provinceofTrento.Trento.

Herrero,S.(1985).Bearattacks:Theircausesandavoidance.Pis- cataway,NewJersey:WinchesterPress.

Herrero,S., Smith,T.,DeBruyn,T.D., Gunther,K.,& Matt,C.

A.(2005).Brownbearhabituationtopeople:Safety,risks,and benefits.WildlifeSocietyBulletin,33,362–373.

Hilderbrand,G.V.,Schwartz,C.C.,Robbins,C.T.,Jacoby,M.E., Hanley,T.A.,Arthur,S.M.,etal.(1999).Theimportanceof meat,particularlysalmon,tobodysize,populationproductivity, and conservation of North Americanbrown bears.Canadian JournalofZoology,77,132–138.

Inglis, J. E. (1992). An analysis of human-black bear con- flicts in Algonquin Provincial Park, Ontario (1973 – 1990).

11thEasternBlackBearWorkshop(pp.108–123).Waterville Valley,NH.

Jerina,K.,Krofel,M.,Stergar,M.,&Videmsek,U.(2012).Factors affectingbearhabituationtohumans:AGPStelemetrystudy.

Ljubljana:UniversityofLjubljana,BiotechnicalFaculty(Final report–summaryforusers)

Jerina, K., Jonozoviˇc, M., Krofel, M., & Skrbinˇsek, T.(2013).

RangeandlocalpopulationdensitiesofbrownbearUrsusarc- tosinSlovenia.EuropeanJournalofWildlifeResearch,59,459–

467.

Kavˇciˇc,I., Adamiˇc,M.,Kaczensky, P.,Krofel,M.,& Jerina,K.

(2013).Supplementalfeedingwithcarrionisnotreducingbrown beardepredationsonsheepinSlovenia.Ursus,111–119.

Kavˇciˇc,I., Adamic,M.,Krofel, M.,Jerina, K.,& Kaczensky,P.

(2011).Brownbearfoodhabitsinhumandominatedlandscapes of Slovenia: Importance ofintensive long-term supplemental feeding.Proceedingsof the20th internationalconferenceon bearresearchandmanagement(pp.162–163).Ottawa.

López-Bao, J. V., Rodríguez, A., & Palomares, F. (2008).

Behaviouralresponseofatrophicspecialist,theIberianlynx, tosupplementaryfood:Patternsoffooduseandimplicationsfor conservation.BiologicalConservation,141,1857–1867.

Martin,J.,Basille,M.,VanMoorter,B.,Kindberg,J.,Allainé,D.,&

Swenson,J.E.(2010).Copingwithhumandisturbance:Spatial andtemporaltacticsofthebrownbear(Ursusarctos).Canadian JournalofZoology,88,875–883.

Mattson,D.J.,Blanchard,B.M.,&Knight,R.R.(1992).Yellow- stonegrizzlybearmortality,humanhabituation,andwhitebark pine seed crops. The Journal of Wildlife Management, 56, 432–442.

O’Connell-Rodwell,C.E.,Rodwell,T.,Rice,M.,& Hart,L.A.

(2000). Livingwith the modern conservation paradigm:Can agriculturalcommunitiesco-existwithelephants?Afive-year casestudyinEastCaprivi,Namibia.BiologicalConservation, 93,381–391.

Orams, M. B. (2002). Feeding wildlife as a tourism attraction:

A review of issues and impacts. Tourism Management, 23, 281–293.

Putman, R. J., & Staines, B. W. (2004). Supplementary win- ter feeding of wild red deer Cervus elaphus in Europe and North America:Justifications,feeding practiceand effective- ness.MammalReview,34,285–306.

R Development Core Team. (2013). R: A language and environment for statistical computing. R Foundation for

Statistical Computing,Vienna, Austria. ISBN 3-900051-07-0, URLhttp://www.R-project.org/.

Rea,R.V.(2003).Modifyingroadsidevegetationmanagementprac- tices to reduce vehicular collisionswith moose Alces alces.

WildlifeBiology,9,81–91.

Rigg,R.,&Adamec,M.(2007).Status,ecologyandmanagement ofthebrownbear(Ursusarctos)inSlovakia.Liptovsk´yHrádok:

SlovakWildlifeSociety.

Robb,G.N.,McDonald,R.A.,Chamberlain,D.E.,&Bearhop, S.(2008).Foodforthought:Supplementaryfeedingasadriver ofecologicalchangeinavianpopulations.FrontiersinEcology andtheEnvironment,6,476–484.

Robbins,C.T.,Schwartz,C.C.,&Felicetti,L.A.(2004).Nutritional ecologyofursids:areviewofnewermethodsandmanagement implications.Ursus,15,161–171.

Rogers, L.L.(2011). Doesdiversionaryfeeding createnuisance bearsandjeopardizepublicsafety?Human–WildlifeInterac- tions,5,287–295.

Saberwal,V.K.,Gibbs,J.P.,Chellam,R.,&Johnsingh,A.J.T.

(1994).Lion-humanconflictintheGirforest,India.Conserva- tionBiology,8,501–507.

Steyaert,S.M.J.G.,Kindberg,J.,Swenson,J.E.,&Zedrosser, A. (2013). Malereproductive strategy explains spatiotempo- ralsegregationinbrownbears.JournalofAnimalEcology,82, 836–845.

Swenson,J.E.,Gerstl,N.,Dahle,B.,&Zedrosser,A.(2000).Action planfortheconservationofthebrownbear(Ursusarctos)in EuropeCouncilofEurope.Strasbourgh.

Wolf, M.,& Weissing, F. J.(2012). Animal personalities:Con- sequencesfor ecologyand evolution.Trendsin Ecology and Evolution,27,452–461.

Woodroffe,R.,Thirgood,S.,&Rabinowitz,A.(2005).Peopleand wildlife:Conflictorco-existence?Cambridge:CambridgeUni- versityPress.

Zedrosser,A.,Steyaert,S.M.J.G.,Brunberg,S.,Swenson,J.E.,&

Kindberg,J.(2013).Theeffectsofbaitingforhuntingpurposes onbrownbearsandtheirbehavior.Reportnumber2013-3from theScandinavianBrownBearResearchProject.

Ziegltrum,G.J.,&Russell.(2004).Efficacyofblackbearsupple- mentalfeedingtoreduceconiferdamageinwesternWashington.

JournalofWildlifeManagement,68,470–474.

Zuur,A.F.,Ieno,E.N.,Walker,N.J.,Saveliev,A.A.,&Smith,G.

M.(2009).Mixedeffectsmodelsandextensionsinecologywith R.NewYork:Springer.

Availableonlineatwww.sciencedirect.com

ScienceDirect