controlgoogle

IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017158INFORMATIONACQUISITIONANDDECISIONSUPPORTFORSUBURBANDEERMANAGEMENTG.
KentWebb,SanJoseStateUniversity,g.
webb@sjsu.
eduABSTRACTAweb-baseddecisionsupportsiteforpublicmanagementofdeerwascreatedfromanintensivedailyinternetsearchandfromtargetedsearches.
Relyingonnewsandothersources,informationhasbeenorganizedbasedonkeydecisionissues.
Thisinformationhasalsobeenusedtobuildadecisionsupportsimulationusingdetaileddemographicandotherdatarelatedtodeerpopulationmanagement.
InformationacquisitionissuesandanapplicationofthesimulationareillustratedusingacasestudyinSanJose,California,whereadeersterilizationprojecthasreducedthepopulationbelowadesiredtargetandappearsdestinedtoresultineradicationofthelocaldeer.
Trailcameraswereusedtogathersitespecificinformation.
Previoussimulationapproacheslackeddemographicdetailandwerebasedonverynarrowgeographicsamples,resultinginunreliablepredictionsforSanJose.
Thesimulationmodelpresentedhereisbeingtestedagainstaverylargegeographicsampleofcases.
AnanomalouscasereportedatCornell,NewYork,iscontradictedbyotherresultsandmaybearesultofdataissues.
Keywords:DecisionSupportSystems,Simulation,KnowledgeManagement,DataAcquisition,WebInformationSystems,TrailCamerasINTRODUCTIONInarecentarticle,wildliferesearchersMcCance,Decker,Colturi,Baydack,Siemer,CurtisandEason(2017,page1)concludethat"urbanwildlifemanagementrankswithspeciesimperilmentasoneofthegreatestconservationchallengesofourtime"becauseproblemswithurbanmanagementcouldleadtoabacklashagainstgeneraleffortsatwildlifeconservation.
Newspaperarticlesroutinelyreportthatissuesrelatedtosuburbandeermanagementareamongthemostdivisivepublicdecisionsmadeintheircommunities.
IntheUnitedStatesdeerarelegallyownedbythecitizensofeachstatesomanagementofsuburbandeeristypicallyacommunitydecision.
Thedebateamongcitizensabouthowtobestmanagedeeristypicallysupportedbyinformationtakenfrominternetsearches,buttheinternetofferscontradictoryinformationthatoftenconfusesthepublicdebate.
Informationcanbedifficulttofindwithoutextensivesearching.
Simulationmodelsfordeerpopulationmanagementhavebeenconstructed.
However,thereisnoevidencefromnewssourcesreportingonthepublicdebate,monitoredaspartofthisresearch,thatthesemodelsareactuallyusedtosupportthedecisionprocess.
Oneproblemisthatconditionsmayvarydramaticallyacrosslocations,makingitdifficulttocreatemodelsthatcanbegenerallyappliedtoavarietyofsituations.
Also,theuserinterfacesdevelopedaresomewhatchallengingsomaybetoodifficultforgeneraluse.
Agoalofthisprojectistoprovideawebsitewhereextensiveinformationregardingdeermanagementisorganizedaroundkeydecisionissuesinaneasytouseknowledgebase.
Anothergoalistoprovideademographicallydetailedsimulationmodelthatcaneasilyincorporatelocaldatawhenavailable.
Theorganizationalstructureofthewebsiteandasummaryofpreviousacademicdecisionsupportandsimulationeffortsareprovidedinfollowingsectionsofthispaper.
Thedecisionprocess,datacollectionissues,andanapplicationofthesimulationmodelareillustratedusingacaseexamplefromSanJose,California.
DataissueswithacaseattheCornell,NewYork,campusareexaminedastowhytheresultsconflictwithothercasesandrepresentanunlikelysimulationoutcome.
SanJoseCaseStudy,ACameraSurveytoCollectInformationFigure1showsaphotographtakenbyatrailcameraofadoewithaneartagforvisualidentification,andaradiocollarforremotemonitoring–threemethodsofcollectinginformation.
Developmentoflowcost,motionsensingcameras--trailcameras--thatcanbeoperatedwithbatteriesatremotelocationsstoringvideo,infrarednightvideo,andphotosonmemorycardshaveopenedupasignificantnewfieldofcollectinginformationforwildlifemanagement(Hofmeester,Rowcliffe,&Jansen,2016).
Originallydevelopedhttps://doi.
org/10.
48009/2_iis_2017_158-170IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017159primarilyforhunters,effortsareunderwaytodeveloppatternrecognitionsoftwarethatcanconvertimagestodata(Tack,West,McGowan,Ditchkoff,Reeves,Keever,&Grand,2016)dramaticallyimprovingtheeffectivenessofthistool.
ThedoeinFigure1wasrecentlyphotographedinSanJose,California,wherein2007agatedcommunityconsideredusingarcherstoreducethedeerpopulation.
Twoissuesthatwerecitedinfavorofthisaction,damagetolandscapingandtheriskofdeer-vehiclecollisions,arecommonlyraisedacrosscommunities.
Theplantousearcherswasabandonedafterangryprotestsaboutkillingthedeerandpublicsafety,alsoacommonresponsetoaproposeddeercull.
Instead,acontractorwashiredtosterilizethedeer.
AcountofthedeerwastakeninJanuary,2013,asthesterilizationprocesswasunderwayshowing175deer,70malesand105taggedfemales.
Femalesweretaggedwithidentifyingnumbersandsomewerefittedwithradiocollarstohelpevaluatetheprogressoftheproject.
Deertagsaresometimespulledout.
Radiocollarsarelargeandpresentsomerisktodeersurvivalandqualityoflife.
Theyalsohavelimitedbatterylife,butthetechnologyisrapidlyimproving.
Thecontractor,WhiteBuffalo,usedadistancesamplingmethod(Horcajada-Sanchez&Barga,2015)infollowupsurveysofthedeertomonitorthepopulationchange.
Volunteersatthegatedcommunityorganizedaphysicalcountofthepopulationthatinvolveddividingtheareaandassigningpairsofobserverstodrivearoundingolfcartswithbinocularstocountthedeer.
Itsoonbecomeclearthatthevolunteercountwasprovidingagoodestimateofthepopulationandwaslessexpensivethanusingacontractor.
Aspartofthisresearchstudy,atrailcamerasurveyoftheareafromJuly,2016toJuly,2017verifiedthatthevolunteerswereindeeddoingagoodjobofdeterminingthedeerpopulation.
Nouncounteddeerwereidentified.
Toencouragepublicparticipation,acommunitywebsitewascreatedfortheprojectwherecameravideoandprojectupdatesaresharedatwww.
sjdeer.
com.
ForSanJose,thepopulationreductioneffortwastoosuccessful.
Allofthecommunitydoeshadbeensterilizedandnonewdoesweremigratingin.
Bucksweremigratingout.
Giventhattrend,itappearedthepopulationwasheadedforeradication.
Thecamerasurveyconfirmedthatonesterilizeddoe,Doe100,hadafawn,butthedoediedafewmonthslater.
Thefawnsurvived,butturnedouttobeabuck,notwhatwasneededforreproduction.
Thecommunityhadagreedonatargetof70to90deer.
Mostcommunitieswantsomedeer,perhapsjustfewerdeer.
Bythefallof2016,thepopulationcountfellbelowthistarget.
Residentdeerweremovinginandoutofthecommunityusingthelargeopenfrontentrance,buttherestoftheperimeterwasfenced.
Tomakeiteasierfordeeroutsidethecommunitytoenterandhelpsustainthepopulation,cameraswereusedtoidentifylocationsforopeningsinthefencethatwouldsafelyallowoutside,untaggeddoestoenterthecommunity.
Camerasdocumentedthattagged,residentdeerroutinelyusedtwolargeopeningsmadeinthefence,butformonthsnoneoftheuntaggeddoesfromoutsidethecommunitywouldenter,althoughtheywereseeninspectingtheopenings.
Finally,anuntaggeddoecamein,onlytobechasedoutseveraltimesbyoneofthetaggedresidentdoesdefendingherterritory.
AWEBSITETOPROVIDEINFORMATIONFORKEYDECISIONISSUESInvolvementintheSanJosecasewasanoutcomeofaprojectstartedin2010tocollectinformationaboutsuburbandeermanagementandorganizeitonawebsitesothatcommunitydecisionmakerscouldeasilyshareandfindinformation.
Informationforthewebsiteiscollectedfromadailyinternetsearch(Webb,2016)usingabout60relatedkeywordsandGoogleAlerts,atoolthatdeliverssearchinformationatspecifiedtimeperiodsonkeywords.
ABingsearchisalsoconductedtoincreasethechancesoffindingrelevantinformation.
Keyinformationfromeachsource,mostlynewsarticles,issavedontherelevantwebpagesubjecttocopyrightrestrictions.
SeveralGoogleScholarsearchesarealsorunningusingGoogleAlertssonewresearchistypicallydiscoveredasitbecomesavailableonline.
ThesitemapinFigure2forwww.
deerfriendly.
com/deer-population-controlhasbeendevelopedbasedonnewsaccountsofissuesthatareraisedinthepublicdebate.
Thetopwebpageinthesitehierarchyprovidessomesummaryinformationforkeydecisionissues,grayboxesinFigure2,discoveredtobeasignificantbasedonanexaminationofthenewsreportsandresearch.
Linksareprovidedtosub-pageswheredetailedinformationoneachtopiciscollected.
Inadditiontoprovidingaknowledgebasefordecisionmakers,thecollectedinformationprovidesabasisforcreatingrulesinthesimulationmodelthatisdesignedIssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017160toletdecisionmakersevaluatetheimplicationsofdifferentpoliciessuchasthepopulationimpactandcostofnon-lethalandlethalpopulationcontrolmethods.
Emailexchangeswithcommunitymembersfromacrossthecountryinvolvedinpubicmanagementoftheirlocaldeerprovideencouragementthatthewebsitehasbeenusefulasasourceofinformation.
FrompagecreationinJune,2011,toApril,2017,thetoppagewasviewedover29,000timesbyabout25,000users.
Figure2.
SiteLayoutforWebPagesProvidingInformationonKeyIssues(WhiteBackgroundBoxesareWebPages,GrayBoxesareLinkstoPagesbyTopic)REVIEWOFRESEARCHONSUBURBANDEERMANAGEMENTDECISIONSUPPORTANDSIMULATIONWiththerecoveryofwhite-taildeerpopulationsandthedevelopmentofsuburbsinthelatterhalfofthetwentiethcentury,therewasincreasedinterestinmanagingsuburbandeer.
Somecomputerbaseddecisionsupportsystemshavebeendevelopedandempiricalanalysishasbeenconductedtoprovideinsightintomanagementissues.
Forexample,adecisionsupportsystemcalled"SuburbanDeerPopulationModel"developedinthelate1990'sandupdatedasrecentlyasMarch,2017,isavailableatontheinternetfordownloadattheClearingHouseforEcologySuburbanDeerPopulationControlHabitatManagementPopulationManagementManagementGoalsBenchmarksTopicWebPageDecisionIssuesDogsFenceDeerResistantPlantsLethalNon-LethalPublicSafetyDeerFeedingBanRecurringCullFertilityReboundWoundingRiskCostExamplesHuntingEthicsContraceptionContraceptionResearchSterilizationCasesRelocationCasesVariationsinDeerDensityTargetsDeerDensityResearchTransportationResearchLymeDiseaseDeerSurveysPeopleSurveysCaseExamplesbyStateDeerGeneticsIssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017161Software(Etter&VanDeelen,2004).
ThemodelwasdevelopedandrunsusingStella5.
0softwareforWindowscomputersbutcanberunonMacintoshcomputersalso.
Themodelhasthreerequiredinputs:removalsbysex,initialpopulationsize,andsizeofmanagementarea.
Optionalinputsincludethesexratio,recruitment(birthrate),andmortalityrates.
Thedefaultvalueforthesexratioissetat60:40femalestomales,basedonsuburbancullingandmanagementexperience.
Datausedtoestimateotherinputssuchassurvivalandmovementinformationcamefrom147radio-collaredwhite-taileddeertakenfromasingle,geographicalsamplewheremortalityandmigrationcharacteristicsweresomewhatunique,verydifferentfromtheSanJosecase.
Somerelationshipsaredeerdensitydependent,butthereislittledemographicdetail.
Researchrelatedtomodeldevelopmentincludesa2002paperfromEtter,Karmen,VanDeelen,Ludwig,Chelsvig,Anchor,andWarnerdocumenting10logisticmodelsrelatedtosuburbandoesurvival.
Datafordispersalrates(migration)werealsoincludedinthispaper.
Duringthespringdispersalrateswere7%fordoes,6%forfawns,and50%forbuckswithallbutone(buck)stayingwithin9km(page1).
Anearlierproceedings(Etter,VanDeelen,Ludwig,Hollis,Chelsvig,&Warner,2000,p.
198)showslittledispersalfordoesand20%foradultbucks.
Birthrateswereinverselyrelatedtodeerdensity–doesproducemorefawnswhendeerdensityislowcomparedtotheavailablefoodsupply.
Theempiricalestimatefromthisproceedings(Etter,VanDeelen,Ludwig,Hollis,Chelsvig,&Warner,2000,p.
200)wasincludedwithothercitedresearchtoobtainthebirthrateequationforthemodeldevelopedinthispaper.
RondeauandConrad(2003)generateanequationsystemusingdatafromastudyoftheIrondequoit,NewYork,deerpopulation.
Theyrelyonasingleequationtoestimatepopulation,createanequationtoestimatedeermortalityasaresultofdeer/vehiclecollisions,andincludeavarietyofcostestimatesrelatedtodeerdamagesuchastovehiclesandvegetation.
Thedecisionrecommendationresultingfromaverysophisticatedmathematicalanalysisoftheseunderlyingequationswas:"theoptimalmanagement…dictatesthatauthoritiesharvestasmanyanimalsassafetyconstraintswillallow,whenthestockexceedsandendogenouslydeterminedthreshold.
"Theyconcludedthattryingtomaintainastablepopulationwouldbemoreexpensive.
Themodeldoesnotconsidernon-lethalalternativeswhichtheauthorssuspectaremoreexpensive,butmaybeconsideredforethicalreasons(p.
280).
AlsousingdatafromIrondequoit,researchersPorter,Underwood,andWoodward(2004)createdasimulationmodelthatexaminedthecullingorcontraceptiveraterequiredtomaintaindeerpopulationswhichincludedensitydependentdispersalandrecruitmentrateestimates.
Theauthorsconcludethatdispersalisimportantwhenmanagersattempttomaintainlowdeerdensitylevels,butlessimportantwhentargetdeerpopulationsarehigher.
Theyalsoconcludethatthedatashowthatlocalmanagementofdeerpopulationsisfeasiblegivenobserveddispersalrates,butthatmanagersshouldbewaryof"thestandardassumptionthatemigrationequalsimmigration"(Porter,Underwood,andWoodward,2004,p.
255).
A2004suburbandeersimulationmodel(Kilpatrick,H.
J.
,LaBonte,A.
M.
,Barclay,J.
S.
,&Warner,G.
)focusesontheuseofbowhuntingtocontrolpopulations,generallyignoresdensitydependence,butprovidessomeusefuldatafromGreenwich,Connecticut.
Allofthesemodelsarebuiltonanisolated,geographicalsampleswithsparsedemographicdetail.
Thesimulationmodeldescribedhereattemptstocombineallresearchcitedinthispaper.
TheDeerDensityBirthRateRelationshipDeerhaveadaptedtodramaticchangesinfoodavailabilitycausedbyfactorssuchasweather,snowanddrought.
Inordertorapidlyrepopulatewhenconditionsimprove,birthratesincreasewhenthereisgoodfoodavailabilitybutdecreasewhenfoodisscarce.
Thisproduceswhatiscommonlycalledthe"reboundeffect"inhunteddeer.
Areductioninthedeerdensitymakesmorefoodavailableforsurvivingdeer,increasingthebirthrate.
Thebirthrateequationusedinthesimulationmodelattemptstosummarizeresultsfromthestudiesidentifiedinthispaper,expressedasapercentofcarryingcapacity"K",themaximumsustainablepopulation.
A1985studybyRichter&Labiskyreportsthat"incidenceoftwinningwas38%onhuntedsitesand14%onnon-huntedsites"(p.
964).
Verme(1969,p.
881)reports"Productivityoflow-dietyearlingsandprime-ageanimalsamountedto0.
62and1.
36fawnsperdoe,respectively,comparedtoratesof1.
63and1.
80forhigh-dietdeer.
"RondeauandConrad(2003,p.
269)takeasimilarapproachtoprovideamoregeneralpopulationestimateusingcarryingcapacity,K,toestimatetheannualchangeinthedeerpopulation:F(X)=0.
5703X(1-X/K)basedontheIrondequoit,NewYork,deerpopulation,X.
IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017162Anotheradaptiontovariabilityoffoodavailabilityisthetendencytoproducemorefemalefawnswhenfoodisabundantandthehabitatcansupportmoredeer,buttoproducemoremalefawnswhenfoodisscarce.
A1969studybyVerme(1969,p.
881)observesthat"Malescomprised70percentofthebirthsfromphysicallymaturemothersonlowdietwhenbred,whereasmalesconstituted46.
7percentoftheoffspringconceivedbydoesonhighdiet.
"The2004studybyPorter,UnderwoodandWoodward(p.
252)providesanestimateofthenumberoffemalefawnsproducedbydoesatdifferentlevelsofthepopulationsizeasaratiotoK,carryingcapacity.
However,intheiranalysisthecarryingcapacitywassetat20deer/km2(p.
250),quitealownumberforthebiologicalcarryingcapacityofasuburbanenvironment.
Theyreportthatanadultfemaleproducedonaverageabout0.
57femalefawnsperyearwhendeerwereatcarryingcapacityandabout0.
80peryearwhenat48percentofcapacity.
Table1reportstheresultsfortwoequationsrelatedtoK,carryingcapacity,forbirthrateandproportionofmalefawns.
TheequationswereestimatedusingdatafromcitedstudiesincludingthosesummarizedinTable2.
ThetwoequationsprovidedthebestfitofalternativesincludingQuadratic,Cubic,Compound,Growth,Exponential,andLogistic.
SincethedataforK,thecarryingcapacity,wasoftenprovidedinsubjectiveterms--suchashigh,medium,orlow–thenumericalvalueofKwasoftensubjectivelyestimated.
Asaresult,testsofsignificancearenotrigorous.
Theseequationsareincludedasdefaultsinthesimulationmodel.
Table1.
BirthRateonProportionofMaleFawnsEquationsfortheSimulationModelIndependentVariable:ProportionofCarryingCapacity,KDependentVariableConstantParameter(s)FR-SquareSignificance*NMaximumBirths,DoesperFawn(Logistic)0.
026108.
2402650.
960.
000132.
1ProportionofMaleFawns(Cubic)0.
446b1=0.
222b2=-0.
448b3=0.
3652360.
9870.
0001370%*Thedataisnotarandomsample,buthasbeentakenfromacademicandotherresearchreports.
Thevalueoftheindependentvariablewassubjectivelyestimatedinmostcasesbasedonstatementssuchashighorlowpopulationcomparedtocarryingcapacity.
AmaximumvalueforthebirthsvariableisarequirementoftheLogisticmodel.
FortheCubicmodelestimatingtheproportionofmalefawns,amaximumvaluewassetinthesimulationmodelat70%sincethisisthehighestvaluereportedintheliterature.
ThecubicmodelwouldpushthisnumberhigherathighlevelsofK.
MortalityThefollowingmortalityrelationshipsareprovidedasdefaultsinthesimulationmodel.
Etter,VanDeelen,Ludwig,Hollis,Chelsvig,andWarner,(2000,p.
202)reportsonthesignificanceofage-specificreproductionandsurvivalwithdeeringoodsuburbanenvironmentstypicallynotlastingbeyond10years.
Delgiudice,Fieberg,andSampson(2006,p.
1556)observedaUshapedsurvivalpatternina13yearsurvivalstudyoffemalewhite-taileddeer,havinghighmortalityrateatthebeginning—fawnsurvival—andhighlyincreasingratesinolderdeer.
Migration(Dispersal)Dispersalofdeer,ormigration,providesamechanismformaintaininggeneticdiversityindeerpopulations.
Populationshinderedbylackofmigratorycorridorsarelikelytosufferfromgeneticdrift.
Aswithbirthrates,dispersalratesaredensitydependentandalsorelatedtohabitatquality.
Kilpatrick,Spohr,andLima(2001,p.
949)reportthat"Annualhomerangesizeduringhighdeerdensities(88to91deer/km2)werelargerthanduringperiodsofmoderate(20deer/km2)andlowdeerdensities(11deer/km2).
"ExamplesourcesofdataformigrationthatareprovidedforuserinputintothesimulationmodelappearinTable2.
Sincemigrationisverygeographicallydependent,themodelusesanaveragedefaultratebutusersareprovidedwithexamplestohelpadjustthemodelwhichisillustratedinFigure3.
IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017163INTERACTIVEWEB-BASEDSUBURBANDEERMANAGEMENTSIMULATIONMODELTable2.
Inputs,Data,andRelationshipsfortheDeerSimulationModelModelInputsDescriptionandrepresentativedatasourcesMaximumDeerCapacityNumberofdeer.
Knownas"K",thebiologicalcarryingcapacitycanbeover100deerpersquaremileinsuburbs(Lien,2000,p.
1).
InSanJose,theherdwas93persquaremile.
SuburbanAreaInsquaremiles,usedtostandardizesomedensityequationsandtoestablishconstraintsonsomevariableinputs.
CurrentDeerPopulationNumberofbucks,does,andfawnsagedlessthanoneyear.
Asadefault,theadultbucktodoeratioinasuburbanenvironmentisabout40to60.
(DeNicola,Etter,&Almendinger,2008,p.
1)BirthsThedefaultequationisdrivenbydeerdensity,expressedasapercentofdeercapacity,usinganequationsummarizingresultsfromthestudiescitedintheprevioussectiononrecruitmentandtheseadditionalsources.
Resultsoftheequationcanbeeasilyscaledforeachageofthedoe.
Examples:"anaverageof0.
69–1.
08offspringperyear"(Rutberg,2004,p.
246),"increasedsignificantlyfrom1.
33to1.
85fawnsperfemale"aspopulationdensitydeclined(Nielson,Porter,&Underwood,1997,p.
472),"Meanfawnrecruitmentratewas0.
88fawnsperdoe"athighpopulationdensity(Kilpatrick,LaBonte,Barclay,&Warner,2004,p.
1181).
InahighdensitydeerpopulationofCayugaHeights,NewYork,1.
038fawnsperdoe(Curtis,Boldgiv,Mattison,&Boulanger,2009,p.
121)Mortality(Natural)Percentageofbucks,does,andfawnsdyingfromnaturalcauseseachperiod.
Includesvehiclecollisions,disease,predation.
Naturalmortalityiscommonlyexpectedtoincreasewithdeerdensity.
Examples:Annualsurvivalrateof64percent(Porter,Underwood,&Woodard,2004,p.
247)so36percentannualmortality–highduetosignificantvehiclecollisions.
Averagemortalityrateofabout15%foradultdeerandabout72%forfawns,ahighfawnmortality(Nielsen,Porter&Underwood,1997,p.
473).
Annualadultsurvivalrateofabove80percent(Etter,VanDeelen,Ludwig,Hollis,Chelsvig,&Warner,2000,p.
200),somortalityofabout20%.
Migration(dispersal)Migrationratesinandoutoftheareaforbucks,doesandtheirfawns.
Defaultvaluesforquarterlydispersalratesarebasedonthedatadiscussedpreviously.
Datacanalsobeenteredtoreflectlocalconditions.
Forexample,dispersalratesarelowerwhenfoodisabundantanddeerhavenoincentivetosearch.
CullMortalityForpopulationstrategiesrelyingonlethalmethods,thedeercull,thepercentageofbucks,does,andfawnskilledinthecullcanbeenteredasapercentage.
CullCostRepresentativecostsforsharpshootersandotherlethalmanagementmethodsareprovidedbasedoninformationstoredonthewebsitefromcaseexamplesreportedinnewsmedia.
SterilizationPercentAlternativevaluesforthenumberofdoessterilizedcanbeentered.
Sterilizationisnearly100percenteffectiveandpermanent.
Inpracticeasmallpercentagedoesmaybecomepregnantordiefromsurgery.
SterilizationCostRepresentativecostsfordoesterilizationareprovidedbasedoninformationstoredinthewebsitefromcaseexamplesreportedinnewsmedia.
ContraceptionPercentAlternativenumbersforthenumberofdoesreceivingcontraceptioncanbeentered.
Contraceptioneffectivenessiscurrentlyaround90to95percent,reportedupto100%(Hernandez,Locke,Cook,Harveson,Davis,Lopez,…Fraker,2006,p.
1431)ContraceptionDurationTheeffectivenessofcontraceptivedrugsvariesfrom1toabout6years,withlongerlastingdrugsbeingmoreexpensive.
The6yearversion,SpayVac,iscurrentlydifficulttoacquire.
Repeateduseofcontraceptivesonthesamedeermaybedifficultsincecaptureisoftenrequiredanddeerlearntoevadecapture.
ContraceptionCostRepresentativecostsforcontraceptivesareprovidedbasedoninformationstoredinthewebsitefromcaseexamplesreportedinnewsmediaandacademicresearch.
RelocationPercentAlternativevaluesforthenumberofbucksanddoesthatarerelocatedoutoftheareacanbeentered.
Survivalratesandeffectivenessofrelocationvarywidelybasedonexperienceandlocalconditions.
RelocationCostRepresentativecostsforrelocationareprovidedbasedoninformationstoredinthewebsitefromcaseexamplesreportedinnewsmedia.
IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017164Figure3.
FlowDiagramforSuburbanDeerPopulationManagementSimulationIssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017165Inordertoprovideasenseoftheimpactandcostofalternativedeermanagementstrategies,asimulationmodelwasconstructedusingdatafromsourcesasdescribedinthispaper.
Alistofinputsthatausercansetforthemodel,formulasinthemodel,anddefaultvaluesformanyinputsareprovidedalongwithsourcesincitedinthispaper.
Themodelandavailableinputscontinuetobeupdatedbasedontheongoingsearchforinformation.
ThemodelrunsonlineinGooglesheetsandisembeddedinaGooglesiteforeasyaccess.
Theprimaryanalyticaldifferencesofthismodelcomparedtopreviouseffortsistheverydetaileddemographicanalysis.
Also,themodelallowsuserstoenterlocaldataorchoosefromexampledataandtoeasilymodifyequations.
Exampledataisprovidedtohelpuserswiththistask.
Resultsofthemodelarecomparedwiththemanyactualcaseexamplesobtainedthroughthesearchprocess,sothemodelundergoescontinuousimprovementasnewdatabecomesavailable.
Amongthemanycaseexamplesexaminedinconstructingthesimulation,oneseemsparticularlysurprising.
TheCornellAnomalyTheresultofasterilizationprojectatCornellinNewYorkreceivedconsiderablepublicattentioninaWashingtonPostarticletitled:"Tryingtolimitthenumberofdeer,withsurprisingresults"(Landers,2014).
Theleadbiologistontheprojectsummarizesthat:"Ithoughtthatsterilizationinanopenpopulationwherethingscanmoveinandoutwon'twork.
Maybeitwasworthdoingitinasophisticatedwaytosaywetriedinthebestpossiblewayanditdidn'tmakeadifference.
"Accordingtothearticle,althoughthenumberofdoesandfawnswasreduced,anoffsettingincreaseinthebuckpopulationresultedinnoreductionofthetotalpopulation.
Bucksweredrawntherebythesterilizeddoesthatcontinuedtogointoheat,attractingbucksaspartofthebreedingprocess.
Thepublishedstudyreported"a38%and79%decreaseoftotaladultfemalesandfawnsvisibleinsampledphotographs,respectively,andan873%increaseinadultmalevisitationtocameratraps"(Boulanger&Curtis,2016,p.
727).
Thedataforsampledphotographsincluderepeatvisitsofthesamedeer.
Malevisitationsinthefirstyearwereonly4%ofthetotal,forabucktodoeratioofaboutonetoeight.
Thesexoffawnsandmanydeerinthephotographscouldnotbedetermined.
Inasuburbanherdthebucktodoeratioistypicallyabouttwotothree.
The873%increasewasjustenoughtobringthebucktodoeratiobacktoitsnormallevel.
Thereisnoexplanationinthepaperfortheincrediblylowstartingnumberofbucks.
Althoughthepaperconcludesthattheexperimentfailedbecausethepopulationremainedapproximatelythesame,thenumberofdeerseeninsampledphotographsactuallydecreasedby32%(Boulanger&Curtis,2016,p732).
TheconclusionthatthedeerpopulationwasnotreducedbytheexperimentisbasedonanalysisofthephotographicdatausingthesoftwareNOREMARK(White,1996).
Whilenopopulationestimateswereprovidedforadultmales,thepopulationestimatesforadultfemalesshowanincreaseoverthestudy(Boulanger&Curtis,2016,p732),incontrasttothe38%decreaseintotalfemalesfromsampledphotographs.
Totalpopulationestimateslaterpresentedbytheresearchersappeartoshowthatthetotalpopulationtrendwasaboutthesamebutalittlehigherthanthefemalepopulationbasedonthesoftwareestimate,suggestinglittleeffectfromadultmalesonthetotalpopulationtrendestimate(Boulanger&Curtis,2017).
Thereisadramaticdifferenceinthepopulationtrendsbasedonthephotographicdataandthesoftwareinterpretation,yetbothresultsareusedtoselectivelydrawconclusions.
AnimportanttechnicalissueleftoutoftheWashingtonPostarticleisthattwosterilizationmethodswereused,ovariectomyandtuballigation,notjusttuballigation.
Doeshavingtheirovariesremoved,ovareictomy,typicallycannotgointoheat,soarenotlikelytoattractbucks.
Twoothersterilizationprojectswithpublicallyaccessiblewebsitesbothreportreduceddeerpopulations:inOhioatCliftonDeer.
org(Year2FieldOperationsReport,2016-17)andinMarylandatwww.
WildlifeRescueInc.
org(FirsteverNon-LethalWhite-tailedDeerBirthControlSterilization,n.
d.
).
TheSanJosecasealsousedbothsterilizationmethods,althoughamuchsmallerproportionoftuballigation,butasreportedinTable3resultedinan89percentdecreaseinthenumberofbucks.
Thesecountsweremadebydirectobservation.
TheauthorsoftheCornellstudyacknowledge"itisunclearwhymalenumbersincreased,""aninabilitytoreliablyestimatethemaleportionofthecampusdeerpopulation,"(Boulanger&Curtis,2016,p.
732)andthatthattheirresultofadramaticallyincreasednumberofbucksis"currentlyundocumentedintheliterature"(Boulanger&Curtis,p.
733).
Theincreaseinthenumberofbucksphotographedmayalsohavebeencausedbythefoodpilesusedasbaittoattractdoesforsterilization,bythefawnsbornintheareaduringthestudy,byachangeintheareasfrequentedbyagroupofbucks.
IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017166TheexperienceofaoneyearcamerasurveyinSanJosemakesveryclearthedifficultyofcountingunmarkeddeer.
ThecamerasaregoodforcountingmarkeddeerandwereusedinSanJosetoverifythephysicalcountdonebyvolunteers.
Amongthemarkeddeerwefoundthatthesamedeerwouldreturnfrequentlytothesamelocation,butthisvariedovertime.
Gettingaccuratepopulationestimatescontinuestobeachallengeforwildlifemanagementandisthesubjectofmuchresearchdevotedtoimprovingavailabletechnologyandmethodsinanenvironmentoftightfunding.
ThecontractorontheSanJoseProject,WhiteBuffalo,continuestoexperimentwithpopulationcontrolmethods,providingnewdataandpossiblealternativesolutions.
MODELOUTPUTFigure4showstheoutputfromthesimulationmodelusingdatafromtheSanJosecaseexample.
ThesterilizationbeganinJanuary,2013,withsomedoesrelocatedresultinginthedropinthenumberofdoesreportedinTable3forthebeginningoftheproject.
Thecamerasurveysindicatethatthevolunteercountsareextremelyaccurate.
Imagesfromthecamerasallowforevaluationofdeerbehavior,providegoodinformationaboutmigrationandbirths,andhavebeenusedinanefforttoencouragein-migrationofnewdoesinanefforttochangethecurrentforecastthatthedeerpopulationwillapproachzeroinaboutfiveyears.
Sofar,effortstoencouragedeermigrationintothecommunitybymodifyingfencingwherethereseemtobegoodopportunitiesforimmigrationhavenotbeensuccessful.
Communitydeercontinuetomigrateinandout,butdeeroutsidethecommunityarereluctanttoenterandareconfrontedbyresidentdeer.
ThetargetdeerpopulationbasedoncommunityconsensusisillustratedbythegrayareaintheFigure4.
Thepopulationhasalreadyfallenbelowthistargetof70to90deer.
IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017167Thesimulationmodelfitsthehistoricaldataverywellbecausethemodelallowsforinputoflocaldatatoimprovemodelperformance,notafeatureeasilyaccomplishedinprevioussimulationattemptsthatwereestimatedusingisolatedgeographicsamples.
Inthiscase,thereisverygooddataonstartingdeerdemographics,createdbythecontractorduringthesterilizationprocess.
Migrationrateshavebeenmonitoredcloselybycamerasandbymanyconcernedvolunteersthatsharedatathroughtheprojectwebsiteandalsothroughpersonalcontactsthatenhancecommunitysocialinteraction.
Effortstobuildthemodelwithafewequationsasusedinprevioussimulationattemptsresultedinsimulationforecaststhatwerecounterintuitive.
Deerpopulationspersistedfarintothefutureeventhoughasimpleunderstandingthatthedeeraredyingoutandthattheyrarelylivebeyond10or11yearssuggeststhattheresultabovefromthesimulationmodelshouldbeexpected–virtuallynodeerafter10yearsintotheproject.
Thepoorforecastofapersistingpopulationfromothermodelsisaresultoflimiteddemographicdetail.
Inthissimulation,deerpopulationsareestimatedforeachageinyears.
Themodeltracksdeerastheyageanddie.
SUMMARYAlthoughbuildingthesimulationmodelwasagoaloftheprojectfromthebeginning,modelconstructiontookmuchlongerthananticipatedbecauseofthecomplexnatureoftheproblemandthedatacollectionprocess.
Availableformalsimulationmodelsappeartohaveplayedlittleornoroleindecisionmakingforsuburbandeermanagement,noreferencestothesemodelshavebeendiscoveredinthenewsaccountsofthepublicdebateondeermanagement.
However,theweb-basedknowledgebasedevelopedtosupportdecisionmakingasillustratedinFigure2hasbeenusedbymanycommunitiestoassistintheirdecisionprocess.
Previouseffortsatbuildingsimulationmodelsrelyingonseveralequationsestimatedfromasinglegeographicsamplemaynotbeapplicableacrossdifferentlocations.
AmodelfollowingtheformatofseveralequationsfailedtoprovideareasonableforecastforSanJose.
Basedonthisexperience,extensivedetailwasaddedtothemodelwithdemographictablesforthepopulation,mortality,andmigration.
Also,thedatausedtoestimatetheseprevioussimulationswasverydifferentfromtheSanJosedata,particularlyformortalityandmigration,alsomakingthesesimulationsunreliable.
Basedonthisobservation,equationsusedinthismodelwereestimatedfromawidegeographicsampleanddesignedtobeeasilyscaledbytheusertoreflectlocalconditions.
Also,themodelallowsuserstousedefaultdata,peruseexamplesofdatathattheycanenterinthemodel,orenterdatabasedonobservationsfromtheirIssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017168location.
IntheSanJosecase,gooddatawasavailablefromobservationsontheagedistributionofdeer(demographics),migration,andmortalityrates.
TheCornellcaseillustratessomeofthedatachallengeswhenmeasuringtheimpactofpopulationcontrolmethods.
ThesimulationmodeldevelopedherecontinuestobetestedagainstthecasedataaccumulatedintheknowledgebaseandisbeingusedtosupportdecisionmakinginSanJose.
Althoughthemodelwillbeavailableforfreeusethroughanon-linewebsite,preliminarytestingindicatesthemodelmayappeartoocomplexforsomeuserstorun.
Workisunderwayonimprovingtheuserinterface.
Futureresearchwillinvolvetestingacceptanceofthemodelandusingoutputfromthemodeltoassistwithlocaldecisions.
REFERENCESBoulanger,J.
R.
,&Curtis,P.
D.
(2016).
Efficacyofsurgicalsterilizationformanagingoverabundantsuburbanwhite-taileddeer.
WildlifeSocietyBulleting,40(4),727-735.
Boulanger,J.
R.
&Curtis,P.
D.
(2017,June).
Surgicalsterilizationforsuburbandeermanagement:mythorrealityPaperpresentedattheInternationalUrbanWildlifeConference,SanDiego,California.
Curtis,P.
D.
,Boldgiv,B.
,Mattison,P.
M.
,&Boulanger,J.
R.
(2009).
Estimatingdeerabundanceinsuburbanareaswithinfrared-triggeredcameras.
Human-WildlifeConflicts,3(1),116-128.
Delgiudice,G.
D.
,Fieberg,J.
R.
,&Sampson,B.
A.
(2006).
Along-termage-specificsurvivalanalysisoffemalewhite-taileddeer.
JournalofWildlifeManagement,70(6),1556-1568.
DeNicola,A.
J.
,Etter,D.
R.
,&Almendinger,T.
(2008).
Demographicsonnon-huntedwhite-taileddeerpopulationsinsuburbanareas.
Human-WildlifeConflicts,2(1),102-109.
Firstevernon-lethalwhite-taileddeerbirthcontrolsterilizationspayingresearchprogram(n.
d.
)WildlifeRescueInc.
org.
RetrievedJuly9,2017,fromhttp://www.
wildliferescueinc.
org/nonlethal_deer_project.
htmlHorcajada-Sanchez,F.
,&Barja,I.
(2015).
Evaluatingtheeffectivenessoftwodistance-samplingtechniquesformonitoringroedeer(capreoluscapreolus)densities.
AnnalesZoologiciFennici,52(3),167-176.
Hernandez,S.
H.
,Locke,S.
L.
,Cook,M.
W.
,Harveson,L.
A.
,Davis,D.
S.
,Lopez,R.
R.
,…Fraker,M.
A.
(2006).
EffectsofSpayVaconurbanfemalewhite-taileddeermovements.
WildlifeSocietyBulletin,34(5),1430-1434.
Hofmeester,T.
R.
,Rowcliffe,J.
M.
&Jansen,P.
A.
(2016),Asimplemethodforestimatingtheeffectivedetectiondistanceofcameratraps.
RemoteSensinginEcologyandConservation.
Available:http://onlinelibrary.
wiley.
com/doi/10.
1002/rse2.
25/fullEtter,D.
R.
,Karmen,M.
H.
,VanDeelen,T.
R.
,Ludwig,D.
R.
,Chelsvig,J.
E.
,Anchor,C.
L.
,&Warner,R.
E.
(2002).
Survivalandmovementsofwhite-taileddeerinsuburbanChicago,Illinois.
TheJournalofWildlifeManagement,66(2),500-510IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017169Etter,D.
R.
,VanDeelen,T.
R.
,Ludwig,D.
R.
,Hollis,K.
M.
,Chelsvig,J.
E.
,&Warner,R.
E.
(2000)Overabundantdeer:Bettermanagementthroughresearch.
WildlifeDamageManagementConferences--Proceedings.
Paper29.
197-205.
Available:http://digitalcommons.
unl.
edu/icwdm_wdmconfproc/29Etter,D.
R.
,VanDeelen,T.
R.
(2004).
Suburbandeerpopulationmodel.
ClearingHouseforEcologySoftware.
Available:http://nhsbig.
inhs.
uiuc.
edu/wes/model_description.
htmlKilpatrick,H.
J.
,LaBonte,A.
M.
,Barclay,J.
S.
,&Warner,G.
(2004).
Assessingstrategiestoimprovebowhuntingasanurbandeermanagementtool.
WildlifeSocietyBulletin,32(4),1177-1184Landers,J.
(2014).
Tryingtolimitthenumberofdeer,withsurprisingresults.
TheWashingtonPost,September29,2014.
Availableat:https://www.
washingtonpost.
com/national/health-science/trying-to-limit-the-number-of-deer-with-surprising-results/2014/09/29/3c16f9dc-28a5-11e4-958c-268a320a60ce_story.
htmlutm_term=.
634d7ad2ee81Lien,R.
(2000).
UrbandeermanagementinWisconsin.
WisconsinUrban&CommunityForests,8(1),1-5.
McCance,E.
C.
,Decker,D.
J.
,Colturi,A.
M.
,Baydack,R.
K.
,Siemer,W.
F.
,Curtis,P.
D.
,&Eason,T.
(2017).
ImportanceofurbanwildlifemanagementintheUnitedStatesandCanada.
MammalStudy,42(1),1-16.
Nielsen,C.
K.
,Porter,W.
F.
&Underwood,H.
B.
(1997).
Anadaptivemanagementapproachtocontrollingsuburbandeer.
WildlifeSocietyBulletin,25(2),470-477.
Porter,W.
F.
,Underwood,H.
B.
,&Woodard,J.
L.
(2004).
Movementbehavior,dispersal,andthepotentialforlocalizedmanagementofdeerinasuburbanenvironment.
JournalofWildlifeManagement,68(2),247-256.
Rutberg,A.
T.
(2013).
Managingwildlifewithcontraception:whyisittakingsolongJournalofZooandWildlifeMedicine,44(4s),S38-S46.
Rutberg,A.
T.
,Naugle,R.
E.
,Thiele,L.
A.
,&Liu,I.
K.
M.
(2004).
Effectsofimmunocontraceptiononasuburbanpopulationofwhite-taileddeer.
BiologicalConservation,116(2),243-250.
Richter,A.
,&Labisky,R.
(1985).
Reproductivedynamicsamongdisjunctwhite-taileddeerherdsinFlorida.
JournalofWildlifeManagement,49(4),964-971.
Rondeau,D.
,&Conrad,J.
M.
(2003).
Managingurbandeer.
AmericanJournalofAgriculturalEconomics,85(1):266-281Tack,J.
L.
P.
,West,B.
S.
,McGowan,C.
P.
,Ditchkoff,S.
S.
,Reeves,S.
J.
,Keever,A.
C.
,&Grand,J.
B.
(2016).
AnimalFinder:Asemi-automatedsystemforanimaldetectionintime-lapsecameratrapimages.
EcologicalInformatics,36(November),Pages145–151Verme,L.
(1969).
Reproductivepatternsofwhite-taileddeerrelatedtonutritionalplane.
TheJournalofWildlifeManagement,33(4),881-887.
IssuesinInformationSystemsVolume18,Issue2,pp.
158-170,2017170Webb,G.
K.
(2016).
Internetsearchenginecapturesuccessratesandmortalitystatisticsforhyperlinkstopublicnewsarticles.
IssuesinInformationSystems,17(2),25-33.
White,G.
C.
(1996).
NOREMARK:populationestimationfrommark-resightingsurveys.
WildlifeSocietyBulletin,24,50–52.
Year2FieldOperationsReport(2016-17).
CliftonDeer.
Org.
RetrievedJuly9,2017,fromhttp://cliftondeer.
org/year-two-field-operations-report/