Newell s shearwater population modeling for Habitat Conservation Plan and Recovery Planning

Transcription

1 The Hawai`i-Pacific Islands Cooperative Ecosystem Studies Unit & Pacific Cooperative Studies Unit UNIVERSITY OF HAWAI`I AT MNOA Dr. David C. Duffy, Unit Leader Department of Botany 3190 Maile Way, St. John #408 Honolulu, Hawai i Technical Report 176 Newell s shearwater population modeling for Habitat Conservation Plan and Recovery Planning December 2011 Adam M. Griesemer 1 and Nick D. Holmes 2 1 Kauai Seabird Habitat Conservation Plan, Department of Land and Natural Resources, Division of Forestry and Wildlife 2 Kauai Endangered Seabird Recovery Project, Department of Land and Natural Resources, Division of Forestry and Wildlife

2 PCSU is a cooperative program between the University of Hawai`i and the Hawai`i-Pacific Islands Cooperative Ecosystem Studies Unit. Organization Contact Information: Kauai Seabird Habitat Conservation Plan, Department of Land and Natural Resources Division of Forestry and Wildlife. 4272B Rice Street, Lihue, Hawaii (808) office Recommended Citation: Griesemer, A.M. and N.D. Holmes Newell s shearwater population modeling for Habitat Conservation Plan and Recovery Planning. Technical Report No The Hawai`i-Pacific Islands Cooperative Ecosystem Studies Unit & Pacific Cooperative Studies Unit, University of Hawai`i, Honolulu, Hawai`i. 68 pp. Key words: Newell s shearwater, petrels, endangered, population modeling, light attraction, powerline collision, predation, management, Save Our Shearwaters, minimization, predator control, predator eradication, chick translocation Place key words: Hawai`i, Kauai, Poipu, Kapaa, Upper Limahuli Preserve, Kalaheo, Kilauea Point National Wildlife Refuge Editor: Clifford W. Morden, PCSU Deputy Director ( cmorden@hawaii.edu) About this technical report series: This technical report series began in 1973 with the formation of the Cooperative National Park Resources Studies Unit at the University of Hawai'i at Mnoa. In 2000, it continued under the Pacific Cooperative Studies Unit (PCSU). The series currently is supported by the PCSU and the Hawai'i-Pacific Islands Cooperative Ecosystem Studies Unit (HPI CESU). The Pacific Cooperative Studies Unit at the University of Hawai'i at Mnoa works to protect cultural and natural biodiversity in the Pacific while encouraging a sustainable economy. PCSU works cooperatively with private, state and federal land management organizations, allowing them to pool and coordinate their efforts to address problems across the landscape. The Hawaii-Pacific Islands Cooperative Ecosystem Studies Unit is a coalition of governmental agencies, non-governmental organizations and universities that promotes research, education and technical assistance to support better stewardship of imperiled natural and cultural resources within the Pacific. The HPI CESU is one of 17 cooperative ecosystem studies units across the U.S.

3 AdamM.Griesemer andnickd.holmes UniversityofHawaii PacificCooperativeStudiesUnit & StateofHawaii DepartmentofLandandNaturalResources DivisionofForestryandWildlife

4 2 Newell sshearwaterpopulationmodeling Newell s shearwater population modeling for HCP and RecoveryPlanning Summary TheNewell sshearwater(puffinusauricularisnewelli),aniucnandesalistedspecies,facesterrestrial threats from predation, fallout (attraction to artificial lights) and collision with powerlines. Various indicessuggestthepopulationhasdeclinedby~75%inthepasttwodecades.populationmodelingis requiredforhabitatconservationplan(hcp)andrecoveryplanningtoconsiderthebenefitsofexisting andproposedmanagementactionstothekauaipopulation. Population scenarios modeled here included a) stable, realistic and optimal growth; b) threats of predation, fallout and powerline collision; and c) management actions of minimizing fallout and powerlinemortality,thesaveourshearwaterrescueprogram,predatorcontrol,predatoreradication andchicktranslocation.thegrowthrate(lambda)producedinourworstcasethreatscenarioforall threats(0.906)fellwithintherangeofannualchangesuggestedbyornithologicalradardatafrom using only Newell s shearwater traffic (0.899), and Save Our Shearwater data of Newell s shearwaterfledglingsfrom (0.905). When considered independently, fallout and powerline minimization have the potential to increase growth rate by up to 0.5% and 0.3%, respectively, but would prove more effective in areas where concurrent colony management is planned.the Save Our Shearwater program could theoretically increasegrowthratebyupto0.8%iffledglingsrecoveredandreleasedexperiencethesamesurvivalas fledglings notexperiencingfallout.whileexperiencingpowerlinestrikeandfalloutthreats,predator control with 90% effectiveness produced population growth rates from to and predator eradication produced growth rates from to 1.00, suggesting additional growth of 0.2%2.4% wouldberequiredtostopadeclineintheseareas.weallowedtranslocatedpopulationstogrowat 1.012becausechickswouldlikelyonlybemovedtoareasfreeofterrestrialthreats. Combinedmanagementscenarioswith1)10%populationsubjectedtopredatorcontrol,5%topredator eradication,and100chickstranslocatedoverfiveyears,and2)20%populationsubjectedtopredator control,10%topredatoreradication,and400chickstranslocatedover10years,wouldprovideabenefit of 2,000 and 4,000 birds over 25 years, respectively, compared to no management undertaken. The benefitsfromtheseactionshaveclearpotentialtooffsetincidentaltakeproposedinhcpplanning. Arecovered(i.e.stable,selfsustaining)Newell sshearwaterpopulationin25yearswilllikelybemuch reducedinsizefromwhatexiststoday.however,itiscertainlyrealisticthatsubsetsofthepopulation can achieve positive growth rates when predation, fallout and powerline mortality are removed. Combined management actions likely to provide the greatest potential benefit should be targeted to preventthisspeciesfrombecomingextinct,includingpredatorcontrolofnorthwestcolonies,predator eradication projects in combination with aggressive powerline and fallout minimization in the same region,andchicktranslocationtothreatfreeenvironments.

5 3 Newell sshearwaterpopulationmodeling Keywords:Newell sshearwater,petrels,endangered,populationmodeling,lightattraction,powerline collision, predation, management, Save Our Shearwaters, minimization, predator control, predator eradication,chicktranslocation

6 Newell sshearwaterpopulationmodeling 4 Contents SUMMARY INTRODUCTION METHODS MODELINGAPPROACH Stable,GrowthandOptimalmodels Elasticityandsensitivityanalyses ThreatModels ManagementModels MODELINGASSUMPTIONS DATAQUALITY MODELINPUTS PopulationSize LifeHistoryParameters Ageoffirstbreeding Breedingprobability Reproductivesuccess Subadultsurvivorship Adultsurvival Threats Predation Fallout/Lightattraction Powerlinestrike Managementactionsasindependentscenarios FalloutMinimization PowerlineStrikeMinimization SaveOurShearwatersprogram Predatorcontrol Predatoreradication Chicktranslocation Managementactionsascombinedscenarios RESULTS STABLEREALISTICANDOPTIMALGROWTHMODELS Elasticityandsensitivityanalyses THREATMODELS ComparisontoOrnithologicalRadarandSOStrends MANAGEMENTMODELS Managementactionsasindependentscenarios Falloutandpowerlinestrikeminimization SaveOurShearwatersprogram Colonymanagementactions AdditionalGrowthrequiredtoachieveastablepopulation Managementactionsascombinedscenarios Kauaipopulationwithmanagementactions...37

7 5 Newell sshearwaterpopulationmodeling 4 DISCUSSION STABLE,REALISTICANDOPTIMALGROWTH THREATSTONEWELL SSHEARWATERS ComparisonofprojecteddeclineswithOrnithologicalRadarandSOStrends Comparisontopreviousmodelingefforts Whatvitalrateshouldmanagementeffortstargettomaximizerecovery? MANAGEMENTSCENARIOS ACKNOWLEDGEMENTS REFERENCES...44

8 6 Newell sshearwaterpopulationmodeling ListofTables DATAQUALITYFORMODELINPUTS...15 TABLE21 TABLE22 BREEDINGPROBABILITYINPROCELLARIFORMES...17 TABLE23 PREVIOUSESTIMATESOFADULTSXINSMALLSHEARWATERSANDPETRELSUSEDINMODELING...19 TABLE24 ADJUSTMENTSFROMSTABLEFORTHREATMODELS...21 TABLE25 BREEDINGSUCCESSFOR44STUDIESOFBURROWINGPETREL3YEARS.INDIVIDUALSSTUDIESDETAILEDINAPPENDIX1.22 TABLE26 FALLLOUTASAPROPORTIONOFTOTALANNUALNEWELL SSHEARWATERCHICKPRODUCTION,BASEDONCHICKPREDATION INTHECOLONYANDDISCOVERYRATEOFFALLOUTBIRDSBYTHEPUBLIC,ANDA75%POPULATIONDECLINE...24 TABLE27 NEWELL SSHEARWATERREPRODUCTIVESUCCESSADJUSTMENTSFORFLEDGLINGFALLOUTMORTALITYUSEDINTHREATAND MANAGEMENTMODELS...24 TABLE28 NEWELL SSHEARWATERREPRODUCTIVESUCCESSADJUSTMENTSFORFLEDGLINGFALLOUTMORTALITYUSEDTO TABLE29 TABLE31 TABLE32 TABLE33 INVESTIGATEEFFECTOFSOSMANAGEMENTACTION...26 COMBINEDMANAGEMENTACTIONSCENARIOSMODELED...29 GROWTHRATESFORNEWELL SSHEARWATERUNDERLOW,MEDIUM,ANDHIGHPROJECTEDLEVELSOFFALLOUT, POWERLINEANDPREDATIONTHREATS,ASANADJUSTMENTFROMSTABLE...31 MODELTERMSANDOUTCOMESFORFITTINGMODELSOFSOSANDORNITHOLOGICALRADARDATEBYYEAR EFFECTOFSAVEOURSHEARWATERSPROGRAMONPOPULATIONGROWTHRATESUNDERLOW,MEDIUM,ANDHIGH PROJECTEDLEVELSOFFALLOUT,POWERLINE,ANDPREDATIONTHREATS...33 TABLE34 GROWTHRATESWITHCHICKTRANSLOCATION,PREDATORERADICATION(100%EFFECTIVENESS),ANDPREDATORCONTROL (90%EFFECTIVENESS)ATLOW,MEDIUM,ANDHIGHPROJECTEDLEVELSOFFALLOUT,POWERLINE,ANDPREDATIONTHREATS...34 TABLE35 GROWTHRATESREQUIREDTOACHIEVESTABLEPOPULATIONINMANAGEMENTAREASUNDERLOW,MEDIUM,ANDHIGH PROJECTEDLEVELSOFFALLOUT,POWERLINE,ANDPREDATIONTHREATS.RESULTSINITALICSARETHOSEEQUALTOORGREATER THANMODELEDREALISTICGROWTHRATE,ANDRESULTSINBOLDITALICSAREGREATERTHANOPTIMALGROWTHRATE...35

9 7 Newell sshearwaterpopulationmodeling ListofFigures FIGURE21 STAGEBASEDMODELFORNEWELL SSHEARWATERS.NUMBERSREFERTOSTAGE(0THROUGH5REPRESENTSUBADULT AGES,6+ADULTS),A)ISSURVIVORSHIPSANDB)FECUNDITY...10 FIGURE22 NEWELL SSHEARWATERPOPULATIONMATRIXUSEDFORMODELSIMULATIONS.NTISTHESTAGESPECIFICPOPULATION SIZEATTIMET,SISTHEAGESPECIFICSURVIVALATAGES0,1,2,3,4,5ANDADULT(6+),ANDFISFECUNDITY,THEPROBABILITYOF PRODUCINGAVIABLEONEYEAROLDINTHENEXTGENERATIONANDDEFINEDASF=FXBXC,WHEREFISTHEPROPORTIONOFTHE POPULATIONTHATISFEMALE(ASSUMEDTOBE0.5),BISBREEDINGPROBABILITY,ANDCISTHEPROPORTIONOFCHICKSTHAT FLEDGEDFROMTHEEGGSLAIDBYAFEMALEPARENT...10 RELATIONSHIPBETWEENADULTSURVIVALANDLN(MASS).SX= LN(MASS) FIGURE23 FIGURE31 FIGURE32 FIGURE33 FIGURE34 SIMULATIONOFSTABLE,GROWTHANDOPTIMALBREEDINGMODELSFORNEWELL SSHEARWATERS...30 ELASTICITYANDSENSITIVITYOFLAMBDATOVITALRATESINSTABLEPOPULATIONMODEL...30 EFFECTSOFPREDATION,POWERLINEANDFALLOUTTHREATSONGROWTHRATE,ASSUMINGASTABLEPOPULATION...32 PREDATORCONTROLFOR120%OFKAUAIPOPULATIONASSUMINGLOWFALLOUTWITH0%SOSEFFECTIVENESSANDNO POWERLINETHREATS(=0.985,INCLUDINGA)GROWTHRATEOFPOPULATIONRECEIVINGMANAGEMENT,ANDB)BENEFIT EXPRESSEDASTHENUMBEROFBIRDSADDEDTOTHETOTALPOPULATIONCOMPAREDTONOMANAGEMENT(=0.913)...36 FIGURE35 PREDATORERADICATIONFOR110%OFKAUAIPOPULATIONASSUMINGNOFALLOUTANDNOPOWERLINETHREATS (=1.000),INCLUDINGA)GROWTHRATEOFPOPULATIONRECEIVINGMANAGEMENT,ANDB)BENEFITEXPRESSEDASTHENUMBER OFBIRDSADDEDTOTHETOTALPOPULATIONCOMPAREDTONOMANAGEMENT(=0.906)...36 FIGURE36 NEWELL SSHEARWATERSINATRANSLOCATEDPOPULATIONOFUPTO20CHICKSPERYEARFORFIVEYEARS,21TO40 CHICKSPERYEARFOR10YEARS,AND41TO60CHICKSPERYEARFOR10YEARS,INCLUDINGA)PROJECTEDGROWTHRATE (=1.012),ANDB)TRANSLOCATIONBENEFITEXPRESSEDASTHENUMBERSOFBIRDSADDEDTOTHETOTALPOPULATIONCOMPARED TONOMANAGEMENT(=0.906)...36 FIGURE37 TOTALKAUAIPOPULATIONWITHCOMBINEDMANAGEMENTACTIONS,INCLUDINGA)GROWTHRATEINSIDEMANAGEMENT AREAS,B)GROWTHRATEOUTSIDEMANAGEMENTAREAS,C)GROWTHRATEFORTOTAL(INSIDEPLUSOUTSIDE),ANDD)BENEFIT EXPRESSEDASTHEADDITIONALNUMBERSOFBIRDSADDEDTOTHETOTALPOPULATION.REDINDICATESBASELINEGROWTHWITHNO MANAGEMENTACTIONS.GREENINDICATESMANAGEMENTCOMBINATION1WITH10%POPULATIONSUBJECTEDTOPREDATOR CONTROL,5%TOPREDATORERADICATION,AND100CHICKSTRANSLOCATEDOVERFIVEYEARS.BLUEINDICATESMANAGEMENT COMBINATION2WITH20%POPULATIONSUBJECTEDTOPREDATORCONTROL,10%TOPREDATORERADICATION,AND400CHICKS TRANSLOCATEDOVER10YEARS...37 FIGURE41 PTERODROMACAHOWPOPULATIONTRENDSFROMBROOKEETAL.(2010)...39 Appendices Appendix1: Appendix2: Appendix3: Appendix4: Appendix5: Appendix6: Breeding success of petrels <1000 g where nonnative predators absent, nonnative predators present,andnonnativepredatorscontrolledoreradicated. Ainleyetal.(2001)growthratesincludingpowerline,falloutandpowerlinemortality SaveOurShearwatersandOrnithologicalRadartrendsonKauaiHawaii, BreedingsuccessandburrowingoccupancyfromNewell sshearwatercoloniesatkalaheo( , )andKilaueaPoint( ) AdultsurvivalandmassinProcellariiformes FalloutMortalityRateswith100%SaveOurShearwatersProgramEffectiveness

10 8 Newell sshearwaterpopulationmodeling 1 Introduction The Newell s shearwater Puffinus auricularis newelli is an endemic Hawaiian seabird listed as ThreatenedundertheUSEndangeredSpeciesAct(ESA)andendangeredontheInternationalUnionfor theconservationofnature(iucn)redlist(iucn2010,pl93205).newell sshearwatersarenocturnal, burrow and crevice nesting, colonial breeding petrels (Procellariformes), with key life history characteristicsreflectingakselecteddemographicstrategy,delayedbreeding,lowfecundity,highadult survivalandonlyreturningtolandtobreed(warham1990,warham1996). Asforsomanyseabirdsthatevolvedonislands,Newell sshearwatersareentirelynaïvetothethreats posed by predatory mammals or humans (Nettleship et al. 1994). Known terrestrial threats include predation(nonnativecatsfeliscatus,ratsrattussp.,barnowlstytoalba),habitatmodificationfrom invasive plant species and ungulates (pigs Sus scrofa, goats Capra hircus), and attraction to artificial lightsandcollisionwithartificialstructures(ainleyetal.2001,sincockandswedberg1969). PopulationindicesofNewell sshearwatershavedeclinedprecipitouslyinthepasttwodecades.based on atsea observations from , their population was estimated at 84,000 birds, with ~90% nestingonkauai(spearetal.1995).between1993and2009,populationindicesofornithologicalradar, andthenumbersofdownedfledglingscollectedafterattractiontoartificiallights(falloutbirds)declined by~75%(dayetal.2003,holmesetal.2009).concurrently,abreedingrangecontractionappearstobe occurringbasedonreducedactivityatthreenewell sshearwaterbreedingcoloniesknowntobeactive between (Holmes et al. 2009). Atsea threats are poorly understood and require investigation,howeverminimizingterrestrialthreatsiscrucialtotherecoveryofthisspecies. PopulationdemographymodelsandPopulationViabilityAnalyses(PVA)areusedbywildlifemanagers toinformdecisionmakingprocesses(caswell2001,keedwell2004).pvamethodsarewellestablished andcanhaveahighdegreeofaccuracyoverashorttermperiod(i.e.2030yrs).acomparisonoffive PVA software methods showed high shortterm predictive accuracy when compared to observed population changes amongst 21 longterm studies of wildlife populations (Brook et al. 2000). For managers, PVA can be used to compare the impact of different threats, the efficacy of different management scenarios, and evaluate which life stages should be the focus of management efforts (Morrisetal.1999). Previous population modeling of Newell s shearwaters were undertaken in 1995 (Ainley et al. 2001), whenthedramaticdeclineinornithologicalradartargetsandfalloutbirdswasnotyetevident.based on estimated life history parameters and measures of mortality from predation, light attraction, and collision with humanmade structures, model outputs predicted annual declines of 3.2 to 6.1 % ( of 0.968to0.939)ordecadaldeclinesof3060%(Ainleyetal.2001). Light attraction and collision of Newell s shearwater with humanmade structures is considered incidentaltake 1 undertheesaandhawaiirevisedstatutessection195dwhereimpactstothespecies 1 Theterm take asdefinedbytheesameanstoharass,harm,pursue,hunt,shoot,wound,kill,trap,capture,or collect,ortoattempttoengageinanysuchconductpl stat.884,16u.s.c Endangered SpeciesActof1973..PerHRS195D4, Take isdefinedasto: harass,harm,pursue,hunt,shoot,wound,kill,trap,

11 9 Newell sshearwaterpopulationmodeling areunavoidable.thehawaiidivisionofforestryandwildlife(dofaw)isaddressingtheincidentaltake through the development of the Kauai Seabird Habitat Conservation Plan (KSHCP). The KSHCP is an islandwideplanformultipleentitiesapplyingforauthorizationofincidentaltakeviafederalandstate permits (KSHCP 2010).Hawaii state permit issuance criteria require that HCPs must provide a net benefit 2 tothespeciesinadditiontooffsettingunavoidableincidentaltake.biologicalactionstooffset unavoidabletakeandprovidenetbenefitasspecifiedinthehcpsaredefinedas mitigation.proposed mitigation activities of the KSHCP include removing predators from breeding colonies and controlling habitatmodifiers(kshcp2010). Theactionstominimizeandaddressincidentaltakearepartofalargerscopeofrecoveryplanningthat addressesterrestrialandmarinethreatstothisspecies population(holmesetal.2011,usfws1983). Recovery planning requires determination of the most effective management strategy to stop the decline,andallowthepopulationtogrow. TheoverallaimsofthecurrentpaperweretoguideKSHCPmitigationgoalsandrecoveryactionsfor Newell sshearwatersbya)describinglifehistory,b)quantifyingeffectsofthreatsonthepopulation,c) comparingoutcomesofthreatsonpopulationtoexistingpopulationindicesofradarandfalloutdata,d) identifying potential outcomes of management actions, including facility based actions of minimizing lightandpowerlinemortality,thesaveourshearwaters(sos)falloutrescueprogram,andcolonybased managementactionsofpredatorcontrol,predatoreradicationandchicktranslocation.thisdocumentis for use by government agencies, land managers, researchers, and other stakeholders with technical expertise in conservation biology that are involved in HCP mitigation planning and/or Newell s shearwater recovery planning.importantly, in regard to HCP planning these results should not be consideredacomprehensiveassessmentofnetbenefitfromamelioratingpredationthreats,butrather astartingpointforselectinganappropriatemitigationinvestment.investmentsshouldbeconsidered tentativeuntilhcpmonitoringprovidesactualmeasuresofproductivityinmanagementareas. 2 Methods 2.1 Modelingapproach For population modeling, we used stagebased Lefkovitch matrices, based on Caswell (2001) using PopToolsv3.1(Hood2009)andRAMASMetapopv5.0(AkçakayaandRoot2007).Matrixmodelingwas basedonlifehistoryparametersofsurvivalandfecundity(ageoffirstbreeding,breedingprobability, reproductivesuccessandsubadultandadultsurvivorship)(figure21andfigure22). capture,orcollectanythreatenedorendangeredspecies,includingplants,animals,birds,freshandmarinewater species. Incidentaltakeoccursunintentionallyandincidentaltotheuseofotherwiselawfulfacilities. 2 HRS 195D30: Allhabitatconservationplans,safeharboragreements,incidentaltakelicenses,andsubsequent actionsauthorizedunderthoseplans,agreements,andlicensesshallbedesignedtoresultinanoverallnetgainin therecoveryofhawaii'sthreatenedandendangeredspecies. Federalissuancecriteriadoesnotrequirenetgain inrecovery,howeverrequiresthatthetakingwillnotappreciablereducethelikelihoodofsurvivalandrecoveryof thespeciesinthewild(pl93205,section10(a)(2)(b)).

12 10 Newell sshearwaterpopulationmodeling Figure21 Stagebased model fornewell s shearwaters. Numbers refer to stage (0 through 5 representsubadult ages,6+adults),a)issurvivorshipsandb)fecundity. Figure22 Newell s shearwater population matrix used for model simulations. Nt is the stagespecific population sizeattimet,sistheagespecificsurvivalatages0,1,2,3,4,5andadult(6+),andfisfecundity,the probabilityofproducingaviableoneyearoldinthenextgenerationanddefinedasf=fxbxc,wheref istheproportionofthepopulationthatisfemale(assumedtobe0.5),bisbreedingprobability,andcis theproportionofchicksthatfledgedfromtheeggslaidbyafemaleparent. Models were deterministic and made no allowance for natural variation in vital rates over time (i.e. stochasticmodels).reproductivesuccessisthemostwelldocumentedaspectofpetrelandshearwater biology that demonstrates this natural variation, typically showing wide variation from year to year (Appendix1)becauseoffluctuationsinoceanographicconditionsaffectingfoodsupply(Warham1990). A practical outcome of this variation is that it may make a small population more susceptible to extinction (i.e. several bad years in a row). For our modeling purposes, we deliberately chose life history values that represented longterm means, and thus incorporated an element of longterm variationintomodels Stable,GrowthandOptimalmodels Wedevelopedastablemodelbasedonestimatedlifehistoryvaluesforapopulationwithnonegativeor positive growth. We also developed a growth and optimal breeding model to estimate a biologically reasonable and maximum growth rate, the latter under ideal conditions (i.e. no marine or terrestrial threatsandarecoveringpopulation).notethatwedidnotconsiderthisoptimalgrowthrateachievable underanyofourmanagementscenarios,andthepurposewastodemonstratethelimitedopportunity

13 11 Newell sshearwaterpopulationmodeling for growth in this species even under optimal breeding conditions, which is similar for other Procellariformes(Bonnaudetal.2009,CuthbertandDavis2002b,Keittetal.2002). StableModel: =LifeHistory stable RealisticGrowthModel: =LifeHistory growth OptimalGrowthModel: =LifeHistory optimal Elasticityandsensitivityanalyses To determine which vital rates are most significant to the population growth rate we conducted elasticityandsensitivityanalysesusingthestablemodel.elasticityandsensitivityanalysesarecommon modelingtoolstoassesswhichvitalratehasthegreatestimpactongrowthrate.formanagers,these analysesareimportanttoolstoidentifywhatvitalratetotargetwithvariousmanagementactions,and achievethemostimpactforrecovery(keedwell2004).forexample,isitmoreimportanttoreduceadult mortalityfrompowerlinesorfledglingmortalityfromlightattraction? ThreatModels Toinvestigatetheeffectsofpowerlinestrike,fallout,andpredationatthecolony(independently,andin combination)onpopulationgrowthrate,weadjustedthestablemodelforthesethreethreatsunder fourlevels:none,low,medium,andhigh(threatmodels).forcomparison,weprovideainley smeasures ofthesethreethreats(ainleyetal.2001)recalculatedinalefkovitchmodelinappendix2. ThreatModels: =LifeHistory stable Threats To provide a comparison to population modeling results of threats, we compared growth rates producedinthreatmodelstotrendsfromthesaveourshearwatersprogramandornithologicalradar datacollectedonkauaisince1993(appendix3).radardataandsosweremodeledovertime(year)and the parameter estimate. The 95% Confidence Intervals for Year provided an annual measure of populationchange,asacomparativemeasuretoassessgrowthrates()producedinthethreatmodels ManagementModels We investigated the potential impact of management actions including minimization of fallout (light abatement) and powerline strike, the SOS program to recover and rerelease fallout fledglings, plus colonymanagementactionsofpredatorcontrolanderadication,andchicktranslocation. Weconsideredeachofthesemanagementmodelsindependently,andincombination.Weexpressed thepotentialimpactasachangeingrowthrate(lambda),andlambdarequiredtoreachastablegrowth rate. For a subset of combined management scenarios we expressed potential effect on those proportionsofthepopulationbeingmanaged.theeffectwasexpressedasthedifferenceinpopulation sizecomparedtonomanagementactions(i.e.baseline).weusedabaselinepopulationexperiencing highratesofpowerlinecollision,fallout,andpredationasthisgrowthbestfitannualpopulationchange suggested by ornithological radar and SOS trend modeling (Section and Appendix 3).Further, becausetheserepresentedourmostconservativeestimatesofcurrentthreatstonewell sshearwaters, thiseffectivelycreatedaworstcasescenariofromwhichtoplanmanagement.

14 12 Newell sshearwaterpopulationmodeling Effortstoreduceartificiallightshaveincreasedinthelastfiveyears(ErichsenandGriesemer2009),but effortstoreducepowerlinecollisiononkauaihavebeenminimal.however,theeffectivenessofboth typesofminimizationeffortshavenotbeenquantified.forourmodelingweconsideredthenone,low and medium values produced in the threat models representative of minimized powerline strike and falloutmortality. For the SOS program and Colony Management Actions (CMA), we made Management Models by adjustingthreatmodels.toestimatetheeffectsofterrestrialthreatsandthebenefitsofsosandcma onpopulationgrowthrate,lambdaorpopulationgrowthrateshouldideallybemeasuredinsituations where threats and removal of threats are occurring independently and then compared (Natividad HodgesandNagata2001,Pierce2002,Rayneretal.2007,Zinoetal.2008).Thisopportunitydoesnot existfornewell sshearwatersbecausefewcmahavebeenofferedtoanynewell sshearwatercolony, andtheeffectivenessofthesosprogramhasnotbeenquantified.inthisstudy,colonymanagement modelswerecreatedbyadjustingstablelambdavaluesforestimatedeffectsofterrestrialthreats;and benefitsof,whereby: SOSManagementModel(Stable): =LifeHistory stable Threats+Management SOS ColonyManagementModel(Stable): =LifeHistory stable Threats+Management CMA Theoreticallyunderthismodelingapproach,themostbeneficialmanagementoutcomewouldremove all the threats and allow the population to return to a stable growth rate (i.e. the maximum growth achievablewouldbe=1.000).forexample,thisapproachwoulddemandthatevenif100%ofnewell s shearwater predation could be alleviated on Kauai, should there be any mortality from fallout and powerlinecollision,thepopulationwouldstilldecline(albeitslower). To identify what growth rates would be required to stop a decline under our threat models, we developed a second set of colony management models (and not SOS models) where fecundity and survivorshipvalueswereincreased(lifehistory growth )untilastablepopulationgrowthratewasachieved. We then compared this to our Realistic and Optimal Growth model to provide context for what is biologicallypractical. ColonyManagementModel(RequiredGrowth): Stable =LifeHistory growth Threats+Management CMA Wealsocreatedchicktranslocationmodelswherebychicksaremovedtoasitefreeofterrestrialthreats (nopredators,lightsorpowerlines)andsubsequentlyreturntobreedthere(miskellyetal.2009).we allowed these populations the life history values of our Realistic Growth Model because they were subjecttonoterrestrialthreats. TranslocationManagementModel(Growth): =LifeHistory Growth 2.2 Modelingassumptions Assumptions and limitations specific to model inputs are described in each relevant report section. Otherkeyassumptionsincludeuseofaclosedpopulationandastableagestructureatyearone.Input limitationsincludenoestimationofadjustmentsfromhabitatmodificationormarinethreats.

15 13 Newell sshearwaterpopulationmodeling Our modeling approach required assumption of a closed population because it was not possible to estimateimmigrationoremigrationfromotherislands.howeverweexpectimmigrationandemigration wouldhavelittleimpactontheoverallmodelingoutcomesbecauseofthehighsitefidelitycommonto Procellariformes, that 90% of the entire population breeds on Kauai, and that the remaining 10% breedingonotherhawaiianislandsarealsolikelysubjecttothesamethreats(mitchelletal.2005). We also assumed that the population age structure at year one was stable because no estimate of currentpopulationdemographyispossible.becausethepopulationisindeclineandthethreatsfacing birds affect different age classes disproportionally, this assumption is likely untrue. For example, if powerlinemortalityaffectsagreaternumberofsubadultsaged25,thenwewouldexpecttoseelower representationoftheseageclassesinthepopulation. Wecouldnotestimateatseathreatsbecauseoflackofinformation.Newell sshearwatersspendupto 80%oftheirlifeatsea.Thisspeciesisconsideredayellowfintuna(Thunnusalbacores)associateanda flockfeederintheeasterntropicalpacific(etp)(birdlifeinternational2010a,spearetal.2007).impacts from fisheries are unknown but are hypothesized to include reduced density of prey aggregations availabletonewell sshearwaters(spearandainley2007).suchafactorwouldincreaserequiredprey searching and foraging effort to provision offspring, ultimately affecting annual reproductive success andadultsurvival.newell sshearwatersarenotasefficientaflyerasother tunabirds (SpearandAinley 1997), b) and, therefore, increasing prey search is especially perilous to them. This is a critical assumptiontoourmodelingeffortsbecausefisheriesimpactsmayhavebeenoccurringsincetheadvent ofindustrializedfishinginthepacificinthe1950 s.nopopulationmeasuresareavailablenearthistime period, although measures of genetic diversity and stable isotopes (dc13 and dn15) should offer valuableinsights(wileyetal.2011).forourpopulationmodeling,anyfisherieseffectswouldoperate as an additional unquantified threat, potentially leading to an overestimation of modeled threats of predation,powerlinestrikeorfallout,orunderestimationoftotalpopulationgrowthrate.investigating possiblefisherieseffectsisimportanttorecoveryplanning. A second unquantified threat is climate change. Generalized effects from global climate change that could affect Newell s shearwaters include increased thermal stratification, lower ocean productivity (Sarmientoetal.2004),andpossiblyincreasedfrequencyofElNiñoconditionswhichwouldaffectocean productivity (Devney et al. 2009)). Determining the potential effects of climate change is critical to recoveryplanningfornewell sshearwaters. Although habitat modification is considered a threat to forest breeding Hawaiian petrel (Pterodroma sandwichensis)andnewell sshearwaters,noestimationwasmadeforhabitatmodificationbecauseofa lack of quantified estimates. When compared to active Newell s shearwater colonies, three inactive Newell s shearwater colonies had higher rates of nonnative vegetation (Holmes et al. 2009). Certain nonnativeplantspeciesposeathreatbyalteringvegetationandhabitatstructure,forexampleyoung, fastgrowingstrawberryguava(psidiumcattleianum)thicketsreducetheburrowinghabitatavailableto Hawaiianpetrels(Penniman2010).Further,fruitingspeciessuchasstrawberryguavahaveasynergistic interactionwithnonnativemammals(nogueirafilhoetal.2009).

16 14 Newell sshearwaterpopulationmodeling Further,wemadenoallowanceforeffectsofrarecatastrophiceffects,suchashurricanes,oreffectsof inbreedingdepressionforsmallpopulations. 2.3 DataQuality Thequalityofthedataandassumptionsusedinanypopulationmodelarekeytotheiraccuracyand predictivepower(keedwell2004,morrisetal.1999).obtainingadequatelifehistorydata,andother importantmodelingparameters,isakeyconsiderationbeforeundertakinganymodeling(coulsonetal. 2001,Morrisetal.1999).Keymodelinputsforourprocessincludedlifehistory,threatlevelsandcolony managementeffectiveness(table21).spatialandtemporallimitationsarediscussedherewithspecific limitationsidentifiedineachrelevantreportsection. Theminimumlifehistorydatarequirementsforundertakingsimpledeterministicpopulationmodeling include age or stage structure, age of first breeding, mean fecundity (i.e. breeding success, breeding probability)foreachageorstage,andmeansurvivalforeachageandstage(caswell2001,keedwell 2004).ForNewell sshearwatersdirectmeasuresoffecundityexistfromstudiesatthekalaheocolony (Telfer1986,Ainleyetal.2001),andmorerecentlyfromthreepairsofNewell sshearwatersatkilauea PointNWR(Zaun2007)(Appendix4).Importantly,however,theKalaheocolonymeasureswereinthe presenceofpredation,andthiscolonyappearsabandoned,suggestingthesefecunditymeasuresdonot represent stable breeding conditions. To estimate a stable population, we supplemented Newell s shearwater metrics with vital rates from analogous species under stable conditions. Survival has not beenestimatedfornewell sshearwaters,andweusedsurvivalfromanalogousspeciesaswasdoneby Ainleyetal.(2001).Usingmetricsfromanalogousspeciesisacommonapproachformanymodeling effortswithprocellariformes(bonnaudetal.2009,jones2002,keittetal.2002,martinezgomezand Jacobsen2004),andjustifiablegiventhatmanyotherpetrelspecieswillsharesimilarlifehistory(Brooke 2004). To account for the inherent natural variation in measures of survival and fecundity we used studiesfromanalogousspecieswheremorethanthreeyearswereavailable. For our modeling purposes, we also quantified threats of predation, fallout and powerline mortality. Importantlimitationsexistinourthreatestimations,includingtherepresentativenessoftheeffectsof predationobservedatthekalaheocolony,thespatialvariationinpredationacrosstheisland,changesin light and powerline infrastructure since the Ainley et al. (2001) mortality estimates, the population declineandproportionalrepresentationofthepopulationfacingpowerlineandfalloutthreatsonkauai, andchangesinthelikelihoodofcollectingbothdeadandlivenewell sshearwaters.noquantification existsofthesepotentiallimitations. The previous estimates of powerline and fallout mortality produced by Ainley et al. (1995, 2001) are likelylessrepresentativeofcurrentmortality.factorsthathavelikelydecreasedtherateofmortality includeincreaseduseofshieldedlightsacrosskauai(kiucsthcp2010),increasedtreegrowthadjacent to powerlines following Hurricane Iniki, and the lowering of some powerlines closer to vegetation or otherstructures(kiucsthcp2010).factorslikelytohaveincreasedcollisionmortalityincludeincreased development on the island especially near Poipu, Princeville and Kapaa, upgrading >20 km of major transmissionpowerlinesto>15.2minheightalongpowerlineroad,andinstallationofnewpowerlines ontheisland(kiucsthcp2010).

17 15 Newell sshearwaterpopulationmodeling TherateofdeadbirdsreportedtoSOSbythepublicalsomayhavechangedsince1993.Ainleyetal. (1995) noted that 750% of dead birds were reported by the general public. Since then, the general publicmaybelesslikelytoreportdeadbirdsgiventhatfindingthemhasbecomemuchlessfrequent;in additionlawenforcementpersonnelhaveincreasedtheireffortsince2004. In1993,Ainleyetal.(1995)estimatedthat80%oftheNewell sshearwaterpopulationbrednorthof Kealia.Since1993,coloniesontheEastandSouthShoresshowsignsofabandonment(Holmesetal. 2009).Thus,theproportionoftheKauaipopulationontheEastandSouthshoresmaynowbelessthan 20%,inwhichcasetheeffectofpowerlineandfalloutmortalityonthetotalKauaipopulationmaybe lowerbecauseofreducedshearwatertrafficontheseshores. Table21 Dataqualityformodelinputs Direct Indirect Measure Recent(>2000) Historic(<2000) Estimatedand analogous measures(report section) Baseline LifeHistory AdultSurvival Section and Appendix5 Juvenile/Subadult Section Survival BreedingSuccess KilaueaPoint97 09 Kalaheocolony80 84,9394 Section and Appendix1 Breeding Probability Ageoffirst breeding KilaueaPoint97 09 Kalaheocolony80 84,9394 Section Section PopulationSize Atseastudies Section ThreatRates Powerline EPRIstudies9394 Section Fallout SOSDatabase SOSDatabase , EPRIstudies9394 Predation Kalaheocolony80 84,9394 Colony Management Effectiveness Section Section AdultSurvival Section and Appendix5 Juvenile/Subadult Section Survival BreedingSuccess KilaueaPoint97 09 Kalaheocolony80 84,9394 Section and Appendix1 Breeding Probability KilaueaPoint97 09 Kalaheocolony80 84,9394 Section

18 16 Newell sshearwaterpopulationmodeling 2.4 ModelInputs PopulationSize TheonlyNewell sshearwaterpopulationestimateavailableis84,000individuals(19,300breedingpairs inthespringand16,700breedingpairsintheautumn)basedonatseaobservationsbetween1984and 1993(Spearetal.1995).Sincetheseatseaestimates,populationindices(e.g.ornithologicalradar,SOS fledglings)havedeclinedby75%from (dayetal.2003,holmesetal.2009).recently,pyle andpyle(2009)suggestedthebreedingpopulationofnewell sshearwatersonkauaiat10,000pairs, howeverthiswouldsuggestadeclineof4048%duringthepast15yearandisinconsistentwithdeclines evidentinindicesofsosandornithologicalradar(section3.2.1andappendix3). Forourmodelingpurposes,weadjustedinitialpopulationsizeto25%oftheSpearetal.(1995)estimate for Kauai, and assumed that 10% of the population bred elsewhere, producing 18,900 individuals, or 12,040birdsofbreedingageunderastableagedistribution(Ainleyetal.2001)(Section2.2). InadditiontothelimitationsinusingradarandSOSdatatoadjustpopulationsizeoutlinedabove,an additionallimitationisspatialcoverage.sosandradarpopulationtrendmeasurementsareproduced frominformationfromthesouthern,easternandnorthernshoresofkauai,withlimitedtonocoverage along the Northwestern shore where significant numbers of Newell s shearwaters breed (KESRP, unpublished data). While the colonies on the northwestern shore are subject to predation and light attraction, powerline collisions are minimal because of the lack of infrastructure on Kauai s northwesternshore.shouldpassageratesonthenorthwesternshorenotfollowthesetrends,wewould expect that our adjusted population size may be an underestimate. Repeating atsea surveys or determininganothermethodofpopulationestimationiscriticaltorecoveryplanning LifeHistoryParameters Ageoffirstbreeding Age of first breeding within Procellariformes species can vary between sites and between years (Warham 1996). Age of first breeding also can be a density dependent factor in Manx shearwaters (Puffinus puffinus), a closely related species, as a result of burrow competition (Brooke 1990) and in wandering albatrosses (Diomedia exulans) because of fishery related mortality (Croxall et al. 1990, WeimerskirchandJouventin1987). Weassumedtheageoffirstbreedingtobe6years,similartoAinleyetal.(2001)andcomparablevital rates from Manx shearwaters (Brooke 1990), and shorttailed shearwaters (Puffinus tenuirostris) (Bradleyetal.1989) Breedingprobability Breedingprobabilityistheprobabilitythatanyonebirdwillbreedinagivenyear(Newton1998).Thisis inherentlydifficulttomeasureinburrowingpetrelsbecauseitreliesondocumentingtheabsenceofa breedingevent,andcanbeconfoundedbymortality,failedbreeding,subadultburrowoccupancyand TypeIIerror(failuretoobserveabreedingeventwhentherewasone).Further,breedingprobabilityis dynamicandinseabirdscanbeinfluencedbyextrinsiceventssuchasclimaticconditions(e.g.elniño

19 17 Newell sshearwaterpopulationmodeling decreasing forage availability and breeding likelihood (Ainley and Boekelheide 1990), or changes in habitat availability (e.g. hurricanes damaging habitat reducing burrowing opportunities (Ainley et al. 1995)). Demographic studies of longlived seabirds have shown that measures of reproductive performance,includingbreedingprobability,andsurvivorshipareafunctionofindividualquality,age andexperience.forexample,thelikelihoodofbreedingincreaseswithage(ainleyanddemaster1980), andlesssuccessfulbreedersmaybreedlessoften(lescroeletal.2009). Accuratemeasuresofbreedingprobabilityforlonglivedseabirdstypicallycomefromlongtermstudies and are notably sparse. Amongst studies of shearwaters, breeding probability has been reported at ~86% (n>40 years) for shorttailed shearwaters (Bradley et al. 2000), 8690% (n=15 years) in Cory s shearwater (Calonectris diomedea) (Mougin et al. 1997), and ~80% (n=~14 years) for the Manx shearwater(brooke1990).amongst15estimatesofbreedingprobabilityinprocellariformes,ameanof 82%wascalculated(SD=8%,Table22). Table22 BreedingProbabilityinProcellariformes. Species Breeding Probability Reference Diomedeaexulans 87% Jouventin&Weimerskirch(1988) 1 Diomedeachlororhynchos 90% Jouventin&Weimerskirch(1988) 1 Diomedeamelanophrys 80% Princeetal.(1994) 1 Diomedeachrysostoma 72% Princeetal.(1994) 1 Diomedeaimmutabilis 81% Fisher(1969) 1 Phoebetriapalpebrata 75% Jouventin&Weimerskirch(1988) 1 Phoebetriafusca 83% Jouventin&Weimerskirch(1988) 1 Fulmarusglacialis 95% Hatch(1987) 1 Pterodromasandwichensis 89% Simons(1984) Calonectrisdiomedea 90% Mouginetal.(1997) Puffinustenuirostris 86% Bradleyetal.(1990) Puffinuspuffinus 80% Brooke(1990) Puffinusmauretanicus 74% Oroetal.(2004) 2 Bulweriabulwerii 83% Jouventin&Weimerskirch(1988) 1 Hydrobatespelagicus 69% Hemeryetal.(1986) 1 1. InMouginetal.(1997),reportedasproportionofsabbaticalbirds. 2. Reportedasskippingrate. AtKilaueaPointNationalWildlifeRefuge,threepairsofNewell sshearwatersarecurrentlyknownto breed.pairswerefirstidentifiedin1997,2001and2008,andsincethattimehavebreedfor12,9and2 years,respectively,producingabreedingprobabilityof100%(appendix4,table3).thisvalue,plustheir outstandingreproductivesuccess,islikelynotrepresentativeoftheentirebreedingpopulation,norasa realistic management target. Outcomes of 100% reproductive success and 100% breeding probability werenotidentifiedinanyofthe52studieswefoundforburrowingpetrels(table22andappendix1). TheseKilaueabirdsmaybe superbreeders (Lescroeletal.2010,Lescroeletal.2009):membersofthe populationthatwillconsistentlyhavehigherreproductiveoutputbasedoninherentindividualquality.

20 18 Newell sshearwaterpopulationmodeling Few direct measures of breeding probability exist for Newell s shearwaters. Telfer s (1986) study consisted of checking burrows previously known to have hosted breeding in 1980 or earlier, and he reportedthatfrom ,anaverageof46.6%oftheseburrowswereactive(appendix4,table1). Ainleyetal.(2001)revisitedTelfer s(1986)data,andassumedthatnotallactiveburrowswouldhave heldbreeders,andadjustedthisfigurebasedon15.9%ofthepopulationbeing4and5yearoldsthat wouldhaveoccupiedburrowsbutnotbred,toproduceabreedingprobabilityof54.7%. Breeding probability of 46.6 or 54.7% for Newell s shearwaters is low in comparison with values reportedforotherprocellariformesandpuffinusspeciesreportedintable22.when54.7%wasusedby Ainleyetal.(2001)intheirmodelingefforts,itwasconsideredakeyfactorlimitingpopulationgrowth(a valueof80%fromthemanxshearwater(brooke1990)wasneededtosimulateastablepopulation). Both Ainley et al. (2001) and Telfer (1986) identified mate loss from the high adult predation rates reportedatthecolonyascontributingtothisoutcome.mortalitycausedbynearbypowerlinesisalso likelytohavecontributedtothislowrate.ainleyetal(1995)reported10deadsubadultsandadults killedperkmfromnearbylawaiinthesameflywayaskalaheocolony. WefollowedAinleyetal.(2001)inusingabreedingprobabilityof80%fromthecloselyrelatedManx shearwater (Brooke 1990) to model a stable population. We also considered 80% a biologically reasonable management target (i.e. breeding under favorable conditions), and made reductions in breeding probability for areas not subject to predator control correspondingly (Section and ) Reproductivesuccess Reproductive success is the number of chicks fledged from eggs laid, and can vary widely among colonies and years (Warham 1996) because of food availability (Newton 1989) and intrinsic factors includingincreasedperformancewithage,individualqualityandexperience(woolleretal.1989). Brooke (2004) states that values between 40 70% are typical of shearwaters and petrels, with no consistent difference in this metric among the four families of Procellariformes.Mean reproductive successofburrowingprocellariformesforstudies3threebreedingseasonsinareaswherepredators never occurred was 0.59 ± 0.11SD (n=17) and 0.62 ± 0.08SD (n=9) for areas where predators were eradicated(table25andappendix1).thelongtermstudy(40years)ofshorttailedshearwaterson FisherIsland,Australia,providesaveragebreedingsuccessatacolonyintheabsenceofpredatorsas wellasinterannualvariation.woolleretal.(1989)reportedameanof61%(range:2583%),whichis similartothemeanvalueoftheabovestudies. ReproductivesuccessoftheNewell sshearwatersatthekalaheocolonyrangedfrom54 59%(Ainley etal.1995,telfer1986)(appendix4,table1and2).reproductivesuccessofthreepairsofneshatthe KPNWR was 100% for 23 breeding attempts (Zaun 2007) (Section and Appendix 4, Table 3). Similar to breeding probability, we considered 100% breeding success unrepresentative of the entire breeding population, and an impractical target to consider for colony management actions. In the closely related Manx shearwater, Brooke (1990) reported a mean reproductive success of 70.1% at SkolkholmIslandand53.9%atRhumIslands.

21 19 Newell sshearwaterpopulationmodeling To model a stable population, we used 60% breeding success, close to the mean from Table 25 of studiesinareaswherepredatorswereneverpresent,eradicated,orcontrolled.tomodelanoptimal population,weused70%breedingsuccessandweconsideredthistorepresenttheupperendofwhatis likelyforthespecieswhenaveragedoveraminimumof3years Subadultsurvivorship Subadultsarethoseyoungerthanbreedingage.Atbreedingcolonies,subadultsincreaseattendance as they age, investing more time to gain breeding experience with each year. Away from the colony thesebirdsmustgainexperienceinforaging(warham1990,1996).developingtheseskillscomesata cost,andsubadultsurvivalinthefirstfewyearsistypicallylowerthanadultsurvival.subadultsurvival isadifficultvariabletomeasurebecauseindividualsarenotreliablyassociatedwithspecificburrows. Brooke(1990)estimatedthat33.3%ofManxshearwaterssurvivedfromfledglingtobreedingage(age6 orlater).weusedagespecificsurvivalratesestimatedinainleyetal.(2001)toachievebrooke'soverall survival value.ainley et al. (2001) maintained a pattern of age specific survival rates of 0.654, 0.78, 0.89, and (years four, five and six), which we calculated at survival from fledgling to breedingage Adultsurvival Adultsurvivorship(Sx)istheproportionofadultsinthepopulationthatsurvivetothefollowingyear. Survivorship in petrel species increases with adult body mass (Warham 1996). For Newell s shearwaters,ainleyetal.(2001)usedadultp.puffinussxof0.905basedonbrooke(1990).thisvalue was also consistent with an estimated survivorship of Newell s shearwaters based on the allometric regressionof16petrelspecies(with5species<1000g)basedonsxdatafromgaillardetal.(1989)and massfromdunning(1992)(ainleyeta.2001). WeconsiderthatadultSxof0.905forNewell sshearwatersintheabsenceofthreatstobelow.indeed, withareproductivesuccessof0.6andabreedingprobabilityof0.8,thisproducesapopulationdeclining at1%peryear.anotablecomparisonisadultsxinhutton sshearwaterswasmeasuredat0.931(n=9 years)(cuthbertanddavis2002a).previousestimates(asopposedtodirectmeasures)usedinother modelingeffortsofadultsxin300600gshearwatersandpetrelsrangefrom (excludingthe Balearic shearwater (Puffinus mauretanicus) facing significant threats of fisheries bycatch adult mortality)(table23). Table23 PreviousestimatesofAdultSxinsmallshearwatersandpetrelsusedinmodeling Species Mass(g) Sx References Pterodromaatrata Brookeet.al(2010) Puffinusauricularis Kiettet.al(2002) Puffinusnewelli Ainleyet.al(2001) Puffinusyelkouan Bonnaudetal.(2009) Puffinusmauretanicus Oroet.al(2004) Puffinuscarnipes Priddelet.al(2006)

22 20 Newell sshearwaterpopulationmodeling Weupdatedtheallometricregressiontoinclude41datapoints(Appendix5),including23measuresfor petrels <1000 g, plus updated mass measurements (Dunning 2008). Initial data screening highlighted oneoutlier(pelecanoidesurinatrix,sx0.67)whichweremoved.adultsxshowedasignificantpositive relationshipwithln(mass)(f 1,38 =7.55,P=0.009),howeveradjustedR 2 wasonly0.144,highlightingthata largeproportionofvariationinadultsxwasnotexplainedbyln(mass)inourequation. Basedonthisequation,thepredictedadultSxofNewell sshearwaters(381g)is0.920,with95%cifor the mean at and (Figure 23). In combination with breeding success of 0.6 and breeding probabilityof0.8thisproducedapopulationjustabovestable(growthrateof0.05%peryear). Figure23 Relationshipbetweenadultsurvivalandln(mass).Sx= ln(mass) Threats Ouradjustmentsforpredation,lightandpowerlinemortalityforNewell sshearwaterswereprimarily baseduponainleyetal.(2001).adjustmentsfromstableforthreatmodelsaregivenintable24.

23 Table24 Adjustmentsfromstableforthreatmodels Survivorship BreedingProbability ReproductiveSuccess Age (adult) Stable Predation low med high Powerline low med Predation high Lowfallout none low med high Medfallout none low med high Highfallout none low med high

24 22 Newell sshearwaterpopulationmodeling Predation Predation by introduced mammals is a key threat to colonial breeding seabirds (Birdlife International 2010a).Catsandratsareconsistentlyidentifiedasmajorcontributorstothedeclineofseabirdsaround the globe (Keitt et al. 2002, Natividad Hodges and Nagata 2001). On Kauai, evidence of cat and owl predation on Newell s shearwaters comes from almost every colony on Kauai, including the most remote sites (Holmes et al. 2009, Telfer 1986). To account for predation we adjusted breeding probability,breedingsuccessandsurvivorship. Undernaturalconditions,divorceleadstoalowerbreedingprobabilityinthefollowingyearbecauseof thetimerequiredtoacquireanewmate(brooke1990,warham1996).predationcanhavethesame effectonapairbyrequiringtheremainingindividualtoalsosearchforanewmate.giventhedifficulty ofobtainingbreedingprobabilitiesevenunderthemostoptimalconditions,itisnotsurprisingthata reductionofbreedingprobabilityduetopredationisrarelyreported.however,giventhelowbreeding probabilityvalueobtainedatthekalaheocolony,andthatthiswaslikelyanartifactofthepredation recordedthere,weconsideredthisanimportantadjustment.wereduced0.8breedingprobabilityby 0.10,0.20and0.30reductioninbreedingprobabilityasaresponsetopredation,withourmediumto highmeasureprovidingacomparativemeasureto54.7%calculatedbyainleyetal.(2001). Chickandeggmortalityfromnonnativepredatorsiscommonlyreportedinburrowingpetrels(Warham 1996). We conducted a literature review to determine reproductive success in the presence and absenceofpredators(primarilycats,ratsandmice)andafterpredatorcontrolanderadication(n=72, Appendix1).Fromthisliteraturereview,forstudies3yearsinduration,themeanbreedingsuccessof burrowing Procellariformes in the presence of predation without predator control or eradication was 32% (n=17 studies, Table 25). While this broad scale effort did not isolate the effects of nonnative predationfromotherthreatsthatcan influencebreedingsuccess(e.g.,tramplingofburrowsbyferal donkeysorovergrazingbyrabbitsleadingtoerosionandburrowcollapse),predationwasakeyfactor resultinginreducedreproductivesuccess.whilethesepredationimpacts,andoutcomesfromcolony managementactions,willhavelimiteddirecttransfertonewell sshearwatersbecauseofdifferencesin predator species composition, habitat type, seabird species composition and other factors, we consideredtheseresultsrelevantasaguidetothescaleofimpactthatpredationcanhaveoncolonial breedingprocellariformesweighing<1000g,andforourmodelingefforts. Table25 Breedingsuccessfor44studiesofburrowingpetrel3years.IndividualsstudiesdetailedinAppendix1 X SD N Lower1SD Upper1SD Predatorsneverpresent Predatorspresent Predatorseradicated Predatorscontrolled

25 23 Newell sshearwaterpopulationmodeling ChickmortalityhasbeenrecordedforNewell sshearwatersatforestcoloniesonkauai;includingchicks killedinburrows(kesrp,unpublisheddata).atthenewell sshearwaterkalaheocolony,predationwas primarilyassociatedwithsubadultandadultbirdsandnotwithchicks.breedingsuccessatkalaheoin thepresenceofpredationwas59%(n=5, )and54%(n=2, ),notmuchlowerthan 60%consideredforastablepopulation,andsuggestinglittlepredationonchicksduringthesestudies. Howeverpredationofburrowingpetreleggsandchicksiscommonlyrecordedelsewhere(Appendix1) andreasonabletoexpectfornewell sshearwatersgiventhisageclasshasnomeansofescapefrom predatorsoncelocated,andarenaïvetothiskindofthreat.weadjustedbreedingsuccessby0.10, 0.25and0.40fromstableinresponsetopredation,similartooneSDaboveandbelowthemeanfrom studiesintable25. Adultandsubadultmortalityfrompredationatthecolonywillvarybasedonphenologyandageclass (Bonnaudetal.2009,CuthbertandDavis2002b,MartinezGomezandJacobsen2004).Despitelesstime spent at the colony, subadults are considered at higher risk from predation because they are more vocal.adultsarepredictedtobemostvulnerableduringarrivalandprospectingwhentheyvocalizeto reestablish pair bonds. Differential predation was estimated by Ainley et al. (1995) based on the numberofdeadsubadultsandadultsfoundcomparedtotheestimatednumberofbirdsatthecolony (600).Hereportedanestimatedpredationrateof2and3yearoldsat5%,for4and5yearoldsat10%, andbreedersat1%.weusedtheserates,andadjusted50%aboveandbelowthesevaluestosimulate higher and lower predation (7.5%,15%, and1.5%; and2.5%,5%, and0.05% by respective age classes) Fallout/Lightattraction FledglingfalloutmortalitywaspreviouslyestimatedbyAinleyetal.(2001)usingSOSnumbers,andwas considered to be a function of both discovery and morbidity, whereby discovery is the proportion of total fallout fledglings found and morbidity is the proportion of the total dead fledglings reported to SOS.Based on those estimates of mortality, Ainley et al. (2001) decreased reproductive success to model the effect of fallout, and we follow that method here. Because we considered predation to decreasereproductivesuccess(i.e.increasechickmortality),wecalculatedfalloutmortalityestimates foreachpredationlevel(none,low,medium,high). Weconsideredfourlevelsofdiscovery,where100%,80%,67%,and50%ofalldownedfledglingsare reportedbythesosprogram,followingainleyetal.(2001).thecurrent5yearaverage( )of thenumberofbirdsreportedtososwas280,leadingto280,350,418,and560falloutbirdsatthese discovery rates, respectively. Total island chick production under our no, low, medium and high predation models equaled 2889, 2107, 1264 and 602 fledglings produced respectively, and Table 26 showsthetotalfledglingfalloutasaproportionofthetotalfledglingproductiononkauaiunderthese scenarios.

26 24 Newell sshearwaterpopulationmodeling Table26 FallloutasaproportionoftotalannualNewell sshearwaterchickproduction,basedonchickpredationin thecolonyanddiscoveryrateoffalloutbirdsbythepublic,anda75%populationdecline. Predation None low medium High Discovery Downed SOS Fledglingsinpopulation fledglings Proportion of fledglings as falloutbirds Ofthe16estimatesoftheproportionoffalloutbirdsoftotalfledglingproduction,weselectedthe17 th, 50 th, and 83 rd percentiles of the range of fallout (280560) at each predation level to represent low, medium, and high fallout, similar to Ainley et al. (2001). For consistency in modeling threats as rate adjustments,wepresentthedecreaseinreproductivesuccessratesateachleveloffledglingpopulation sizeratherthantheproportionaldecreasesinreproductivesuccess(table27). Table27 Newell sshearwaterreproductivesuccessadjustmentsforfledglingfalloutmortalityusedinthreatand managementmodels Fallout No.ofdead Percentile of Predation Threatlevel fledglings valuerange None Low medium High Fallout Low Medium High Powerlinestrike Ainley et al. (2001) previously estimated mortality from powerline collision and we retained those estimatesherebecausenoquantifiedsearcheffortshavebeenundertakensince.theseestimateswere obtained based on standardized sampling (specific search routes under lines) and the dead birds collected. The numbers of dead birds found and reported were adjusted based on the probability of sighting a carcass, survey period and distance of powerlines not searched. The scaled adjustments resultedinlow,medium,andhighmortalityratesbasedontheproportionofdeadbirdsrepresentedin the total Kauai subadult and adult population. Mortality estimates for subadults (ages 2 to 5) were 0.60%,1.20%,and1.72%andforadultswere0.046%,0.092%,and0.131%respectively Managementactionsasindependentscenarios Management actions considered were a) fallout and powerline strike minimization, b) the Save Our Shearwaters fallout fledgling rescue program, and c) colony based activities of predator eradication, predatorcontrolandtranslocation.wecreatedarangeofeffectivenessscenariosforeachmanagement action.

27 25 Newell sshearwaterpopulationmodeling FalloutMinimization FalloutminimizationisachievedbyreducingtheamountofupwardartificiallightavailabletoNewell s shearwaterfledglingssusceptibletoconfusionbylights.thismanagementactionincludesturninglights off, retrofitting lights with shielded fixtures, replacing lights with full cutoff, angling lights down, changingdurationoflightusage(timers),orchanginglightcolor.previousstudiesofshieldingatresort facilities have demonstrated that fallout can be reduced by nearly 40% (Reed et al. 1986).Fallout minimizationactionscanbeaccomplishedthroughoutreachandthehcpprocess.onkauai,outreachto entitieswiththehighestriskoflightattractioniscurrentlyundertakenbykshcpstaff.between2005 and 2007 approximately 30 businesses on Kauai turned lights off and/or angled lights down to the ground(kshcp2010). Weconsideredthefalloutlevelsofnone,lowormediumasrepresentativeofimpactminimizationgiven thisisapreventativemeasure PowerlineStrikeMinimization Powerline strike minimization is removing (undergrounding) lines, reconfiguring lines, or placing bird divertersonlinestoeliminateorreducecollisionrisk.undergroundinglinespermanentlyeliminatesrisk ofcollision.reconfiguringlinesincludesdecreasingthelineheightdowntothelevelofthesurrounding vegetation, or lowering or switching from a vertical to horizontal grouping (Ainley et al. 1995). Bird divertertechnologieshavebeendemonstratedtohavepositiveeffectsforsomebirdspeciesbymaking linesmorevisible,therebyreducingcollisionrisk(yee2008).procellariformesdidnotoccurintheyee (2008) study, and while bird diverters have been installed on some powerlines on Kauai, no quantificationoftheireffectivenessfornewell sshearwatershasbeenmeasured. The utility cooperative of Kauai has recently identified potential reconfiguration options for line segments(kiucsthcp2010)aspreviouslysuggestedbyainleyetal.(1995).implementationofthese optionsisconditionalontheutilitycooperative sshorttermhcppermitissuance(kiucsthcp2010). Similar to fallout, we considered none, low and medium powerline strike representative of impact minimizationgiventhisisapreventativemeasure SaveOurShearwatersprogram The Save Our Shearwaters program is a public conservation effort whereby fallout fledglings are recoveredbymembersofthepublic,andthenrereleased.theprogramhasbeeninoperationsince 1979, with more than 30,000 Newell s shearwater fledglings collected. No effort to quantify the effectivenessoftheprogramhasbeenundertaken. ObservationstodeterminethesurvivaloffalloutfledglingsreleasedundertheSOSprogramarenotably sparse.from ,of1094adultandsubadult(i.e.1yearinage)mortalitiescollected,only19 (4.14%)werebandedinapreviousyearbySOSstaff.Ainleyetal.(1995)reportednopreviouslybanded SOSbirdsfrom42and33nestsmonitoredattheKalaheocolonyin ,despite1000 sofbirds bandedinthatregionfor13yearsunderthesosprogram.

28 26 Newell sshearwaterpopulationmodeling Itisunlikely thatfalloutfledglingshaveequivalent survivaltofledglingsnot experiencing fallout.itis reasonabletoexpectthatmanyfalloutfledglingsreleasedthroughthesosprogramwouldhavehada decreasedsurvivorshipbecauseofundetectedinjuries,decreasedhealthvalues(weight,hydration),or secondarycomplicationsacquiredduringfallout(disease,parasites,decreasedwaterproofing).further, fledgingislikelyacritical learning periodinpetrellifehistory,particularlygiventhatbirdsspendthe next two years at sea, and it is unclear if the fallout experience may impede future navigation or phototropicfeedingabilities. The assumption that all fallout fledglings released through the SOS program die before reaching breedingage(or0%soseffectiveness)overtheentirecourseoftheprogramfrom isalso likely untrue. More recently, SOS release protocols were modified to increase likelihood of survival, includingfeedingbirdsfallingbelowminimummassthresholdsbeforetheycanbereleased(sos2010). Fledgling mass correlates highly with subadult survivorship in other seabirds and this may increase survivalofthesefledglings(dannandcullen1990). Whilewemodeled0%effectivenessfortheSOSprograminourcombinedmanagementmodels(Section ),wealsomodeled100%toinvestigatethetheoreticalrangeofeffectsthisprogramcouldoffer. Tomodel100%effectivenesswereducedfledglingsuccess(Table28),basedonthemortalityofbirds not being found. The mortality of the downed fledglings not found was estimated by combining the discoverylevels(section )withmorbidityvalueswhichproduced64mortalityratesforeachlevel ofpredation(appendix6).torepresentlow,medium,andhighfallout,weselectedthe17%,50%,and 83% percentile of the range of mortality values at each predation level (Table 28).The theoretical managementeffectwasthedifferencebetween0%effectivenessand100%effectiveness. Table28 Newell sshearwaterreproductivesuccessadjustmentsforfledglingfalloutmortalityusedtoinvestigate effectofsosmanagementaction FalloutwithSOS program(100% effectiveness) Fallout Threatlevel No.ofdead fledglings Percentile of valuerange Predation none Low medium High Low Medium High Predatorcontrol PredatorcontrolforNewell sshearwatersistheremovalofcats,ratsandbarnowlsfromacolonysite. Thisactivityisonlysuccessfulforthedurationofapplicationandisconsideredtohavenovalueonce stopped.examplesofpredatorcontroleffectivenessareoutlinedinappendix1andintable25. ExamplesofpredatorcontroleffortsforNewell sshearwatersrangeconsiderablybetweenthekilauea PointNationalWildlifeRefuge,KalaheocolonyandUpperLimahuliPreserve.AtKilaueaPointNational Wildlife Refuge, a fence keeps dogs at bay and predator control (rat bating, cat trapping and owl shooting) has ranged from none to fair since 1997 (fair being multiple traps set near burrows and checkeddailythroughouttheyear),withfaironlyapplicableforafewyears(brendazaun,pers.comm.). Notably,evenduringthe"fair"years,multiplephotosofcatswereobtainedattheburrowshighlighting luckandtimingplayingaroleinthesebirds continuedsurvival.atkalaheocolony,cattrappingwas

29 27 Newell sshearwaterpopulationmodeling implemented from onwards following a significant predation event, and trap type, frequency, trap success and method are unknown, however likely consisted of cage traps checked duringeverymonitoringvisit(12timespermonth,typicallybeginninginmidlateincubation).atthe Kalaheocolonyfrom199394,predatorcontrolconsistedofcagetrappingalsocheckedduringregularly (every23days)duringmonitoringvisits,with1catcaughtin1993(ainleyetal.1995). AtUpperLimahuliPreserve,predatorcontrolincludesanungulateexclusionfence,removalofnoxious weeds (both initiated to protect watershed and flora habitat values), and cat, rat and owl control throughout the course of the breeding season. Predator control is limited by spatial and temporal coverage,androdenticideapplicationisnotcurrentlyfeasibleunderfederalandstateguidelines(ntbg 2006). We considered predator control for Newell s shearwaters to adequately target each of these key predatorsinconjunctionwithhabitatmanagement(ungulateremovalandnoxiousweedcontrol).for our modeling purposes we considered 90% effectiveness as an upper limit. We expect that 100% of predationcannotbemitigatedbypredatorcontrolfornewell sshearwaters,givena)thedenseforested colonies and steep terrain will make it unrealistic to obtain complete spatial coverage of cats, b) effectiveratcontroliscurrentlylimitedtobaitstationsthuslimitingbothspatialandtemporalcoverage, andc)fencing,trappingandbaitingareineffectiveforbarnowls Predatoreradication Predator eradication at colony sites is the optimal management action, and has been achieved on offshoreisletswithpositiveeffectsfornumerousseabirdspecies(cloutandrussell2006,nogalesetal. 2004). At mainland seabird colonies, predator eradication has been achieved in combination with predator proof fences (Day and MacGibbon 2007) with an important exampleforhawaiicompletedatkaenapointonoahuin2011.onkauai,newell sshearwatercolonies areinsteepandextremelyruggedterrain,wherepredatorprooffencesmaybedifficulttoconstruct.to removerats,significantregulatoryhurdlesmustbeovercomebeforeaerialbaitingcanoccurtoremove ratsoncethefenceiserected,anderadicationviabaitboxeswillbeproblematicinmanyareasbecause of steep and rugged terrain. Predatorproof fences will not exclude barn owl predation on Newell s shearwaters and any such project would require investment in ongoing owl control. Despite these limitations,predatorprooffencesyieldgreatpromiseforrecoveryandhcpmitigationbecausetheywill a) protect against mongoose ingress should these predators become established on Kauai, b) likely provideacostsavingsoverthelongtermwhencomparedtoperpetualpredatorcontrol,andc)provide tremendous complementary benefits to other species affected by rats [e.g. Newcomb s snail (Erinna newcombi),pritcharidapalmspeciesandnativeforestbirds]. Largescalepredatoreradicationislikelynotachievableanytimesoon,howeversmallscaleprojectsmay bepossiblewithinthenext510years.weconsideredtheoutcomefrompredatorprooffencing(and subsequentungulateandpredatoreradication)plusbarnowlcontrol100%predatorremoval.

30 28 Newell sshearwaterpopulationmodeling Chicktranslocation Translocation of petrel chicks is a viable management action and recovery strategy for endangered petrels whereby a new population is established in a predator and threat free environment. Chicks moved~onemonthbeforefledging(i.e.beforetheyexittheburrowforthefirsttime)will imprint on thenewlocationandreturntothenewsitetobreed.thishasprovenasuccessfulstrategyinaustralia and New Zealand (Miskelly et al. 2009, Priddel and Carlile 2009) and has application for Newell s shearwaters. AsimilarprojectwaspreviouslyconductedwithNewell sshearwaterseggs.between1979and1981,90 eggsweremovedtokilaueapointnationalwildliferefuge,andplacedundersurrogatewedgetailed shearwater(puffinuspacificus)parents(byrdetal.1984).todayaminimumofthreepairsareknownto bebreedingattherefuge(zaun2007)andarelikelyprogenyofthecrossfosteredbirds.observations suggestthatadditionalbirdsareprospectingthearealikelybolsteredbytheimplementationofasocial attractionprojectin2007toincreaserecruitmenttothissite(haberetal.2010). Keybenefitsoftranslocationincludeexpandingtherange,establishingan insurance colonyandhaving acolonyinanenvironmententirelyfreeofterrestrialthreats.prospectivesitesforchicktranslocation includelehuaislet(onceratfree)andkaenapoint(followingremovalofpredators).forourmodeling purposesweinvestigatedtranslocatingupto600chicksovera10yearperiod. Keylimitationstoourtranslocationmodelwerenoallowancefornaturalrecruitment,immigration,and noadditionaladjustmentforvitalrates.itisreasonabletoexpecttherecruitmentofprospecting(i.e. first time) breeders to the translocation colony may also occur independent of the translocated individuals (Miskelly et al. 2009, Priddel and Carlile 2009), and this will likely depend heavily on the locationofthecolonywithinanexistingmarineorterrestrialflyway.shouldthetranslocatedcolonyfor example be Kilauea Point, we would expect this natural recruitment to be higher than at Kaena because of the higher passage rates of Newell s shearwaters at the former site. Secondly, it is also reasonabletoexpectthattheageoffirstbreedingmaydecreaseatagrowingcolony(brooke2010), because there are opportunities for younger birds to breed. Indeed, in July 2010 a three year old Newell sshearwaterwasrecordedprospectingwithaknownbreederatkilaueapointnationalwildlife Refuge,providingthepossibilityitmaybreedatfouryearsofagein Managementactionsascombinedscenarios Eachofthemanagementactionsmodeledarenotlikelytobedoneinisolation,andwillalsobesubject topracticallimitations(i.e.predatoreradicationfortheentireislandisnotfeasible).wesubsequently chosefivemodelstoidentifysomepotentialscenariosonkauai,withtwomodelsexpressingeffectfor theentirekauaipopulation(table29). ModelsincludedaNorthwestKauaiscenario,apredatorprooffence,andachicktranslocationproject. The Northwest of Kauai is an obvious choice to implement management actions because of the high density of breeding observations in this region (Holmes et al. 2009), this area naturally abates for collisionwithartificialstructures,andfledglingsfromthisregionmayonlybeexposedtolowartificial lightcoverage(troyetal.2011).thescaleandtopographyofmanyareasofthenorthwestofkauaiwill

31 29 Newell sshearwaterpopulationmodeling likely lead to only predator control being appropriate, and for this scenario we consider 0% SOS effectiveness.a predator proof fence for 110% of the existing population was modeled. With no powerlineandnofalloutthreatgiven,thelocationofsuchafenceisunknown,butweanticipateitwill beconstructedinaregionwheremaximumfalloutandpowerlineminimizationcanalsobeachieved. Chicktranslocationwasmodeledforupto60chicksmovedperyearfor10yearstoapredator,lightand powerline free environment. Finally we considered two models for all three of these scenarios combinedwith1)10%populationsubjectedtopredatorcontrol,5%topredatoreradication,and100 chicks translocated over five years, and 2) 20% population subjected to predator control, 10% to predatoreradication,and400chickstranslocatedover10years. Table29 Combinedmanagementactionscenariosmodeled ModelingScenario ColonyManagementAction Fallout threat Powerline threat SOS effectiveness NorthwestPC Predator Eradication(PE) ChickTranslocation High predation, 90% predator control effectiveness,120%ofthepopulation 100% predator control, 110% of the population Upto600 chickstranslocatedover 10 years low none 0% none none n/a none none n/a Allthreecombined Kauai1 10%populationpredatorcontrol low none 0% 5%populationpredatoreradication none none n/a 100chicksmovedover5years none none n/a Kauai2 20%populationpredatorcontrol low none 0% 10%populationpredatoreradication none none n/a 400chicksmovedover10years none none n/a 3 Results 3.1 Stablerealisticandoptimalgrowthmodels SimulationsofstableandoptimalmodelsaredepictedinFigure31.Thestablesimulationrepresentsan undisturbedpopulationwithareproductivesuccessof60%,anannualbreedingprobabilityof80%,and annual adult survival of 92%.The realistic growth model produced a growth rate of 1.2%, and the

32 30 Newell sshearwaterpopulationmodeling optimalmodel2.3%(table22,table23andtable25).startingwithapopulationof1000individuals, the realistic growth model produced 115 additional birds, and optimal growth model resulted in an additional255birds,aftertenyearscomparedtothestablemodel. Figure31 Simulationofstable,growthandoptimalbreedingmodelsforNewell sshearwaters Elasticityandsensitivityanalyses Lambdaorgrowthratewasmostsensitivetochangesinadultsurvivalinthestablemodel(Figure32). Figure32 Elasticityandsensitivityoflambdatovitalratesinstablepopulationmodel

33 31 Newell sshearwaterpopulationmodeling 3.2 Threatmodels FinallambdasforeachthreatmodelarepresentedinTable31.Predationhadthegreatestimpacton growthrate,withourlow,mediumandhighpredationratesproducingdeclinesof2.7%,5.5%and7.9% per year (Figure 33). Fallout and powerline mortality produced declines of 0.6, 0.8 and 1% per year versus 0.1, 0.3, 0.4% per year, respectively. Ainley et al. s (2001) recalculated threat models are presentedinappendix2. Table31 GrowthratesforNewell sshearwaterunderlow,medium,andhighprojectedlevelsoffallout,powerline andpredationthreats,asanadjustmentfromstable. Powerline None Low Medium High Predation None Fallout None Low Medium High Predation Low Fallout None Low Medium High Predation Medium Fallout None Low Medium High Predation High Fallout None Low Medium High

34 32 Newell sshearwaterpopulationmodeling Figure33 Effectsofpredation,powerlineandfalloutthreatsongrowthrate,assumingastablepopulation ComparisontoOrnithologicalRadarandSOStrends Save Our Shearwaters and Ornithological Radar trend data suggest declines of % per year. When assessed independently or in combination, fallout and powerline threats alone (Table 31) produced far less conservative declines than suggested by SOS and radar trends (Table 32). Threat modelsusingmediumorhighpredationlevelsproduceddeclinesequivalenttoradarandsostrends. The radar model using only data from the peak Newell s shearwater traffic period from (Radar13.3&4.9310)andtheSOSmodelusingdatafrom (SOS.8899)(Table32,Appendix3) indicated declines equal to or worse (0.899, 0.905) than the lowest lambda produced in our threat models(0.906).

35 33 Newell sshearwaterpopulationmodeling Table32 ModeltermsandoutcomesforfittingmodelsofSOSandOrnithologicalRadardatebyyear 1. Model Years AdjustedR 2 1+Parameter estimatedfor thetermyear() Lower95%CI Radar13.ALL Radar13.3& Radar13.ALL Radar13.3& SOS SOS Upper95%CI 1. RadarfollowsDayetal.(2001)methodology(alldata,anddatafromsuggestedpeakofNewell sshearwatertraffic) for and sosisthenumberofnewell sshearwaterfalloutfledglingsrecoveredperyearwithno adjustmentfordiscoveryrate. 3.3 Managementmodels Managementactionsasindependentscenarios Falloutandpowerlinestrikeminimization Minimizingfalloutincreasedthegrowthrateby0.4,0.4and0.5%undernone,lowandhighpredation threats,respectively(table31).minimizingpowerlinestrikeincreasedgrowthratesby0.1%underhigh predationthreatsto0.3%underlowpredationthreats,andby0.3%undernopredation(table31) SaveOurShearwatersprogram TheSOSprogramwith100%effectivenesshasthepotentialtoincreasegrowthrateby0.5%underno predationthreatsto0.8%underhighpredationthreats(table33). Table33 Effect of Save Our Shearwaters program on population growth rates under low, medium, and high projectedlevelsoffallout,powerline,andpredationthreats Powerline None Low Medium High Predation None Fallout Low 0.5% 0.5% 0.5% 0.5% Medium 0.6% 0.6% 0.6% 0.6% High 0.6% 0.6% 0.6% 0.6% Predation Low Fallout Low 0.6% 0.6% 0.6% 0.5% Medium 0.6% 0.6% 0.6% 0.6% High 0.7% 0.7% 0.7% 0.7% Predation Medium Fallout Low 0.6% 0.6% 0.6% 0.6% Medium 0.7% 0.7% 0.7% 0.7% High 0.8% 0.8% 0.7% 0.7% Predation High Fallout Low 0.7% 0.7% 0.7% 0.6% Medium 0.8% 0.7% 0.7% 0.7% High 0.8% 0.8% 0.8% 0.8%

36 34 Newell sshearwaterpopulationmodeling Colonymanagementactions In combination with powerline strike and fallout threat level, predator control produced population growthratesfrom0.976to0.991.predatoreradicationproducedpopulationgrowthratesfrom0.986to 1.00 (Table 34). Chick translocation produced a population growth rate of because of the translocationtoterrestrialthreatfreeareas. Table34 Growthrateswithchicktranslocation,predatoreradication(100%effectiveness),andpredatorcontrol (90%effectiveness)atlow,medium,andhighprojectedlevelsoffallout,powerline,andpredationthreats Powerline None Low Medium High Chick PredationNone Fallout N/A n/a n/a n/a Translocation Predator PredationAll(Low, Fallout None Eradication Medium,orHigh) 100%Effectiveness Low Medium High PredatorControl PredationLow Fallout None %Effectiveness Low Medium High PredationMedium Fallout None Low Medium High PredationHigh Fallout None Low Medium High AdditionalGrowthrequiredtoachieveastablepopulation Additional growth required to achieve a stable population under management scenarios in Table 34 areintable35.

37 35 Newell sshearwaterpopulationmodeling Table35 Growthratesrequiredtoachievestablepopulationinmanagementareasunderlow,medium,andhigh projectedlevelsoffallout,powerline,andpredationthreats.resultsinitalicsarethoseequaltoor greater than modeled realisticgrowth rate, and results inbold italics are greater than optimal growth rate. Powerline None Low Medium High Predator PredationAll(Low, Fallout None Eradication Medium,orHigh) 0.0% 0.1% 0.3% 0.4% 100% Low Effectiveness 0.6% 0.7% 0.9% 1.1% Medium 0.8% 0.9% 1.1% 1.2% High 1.0% 1.1% 1.3% 1.4% PredatorControl PredationLow Fallout None 0.2% 0.4% 0.6% 0.7% 90% Low Effectiveness 0.9% 1.0% 1.2% 1.3% Medium 1.1% 1.2% 1.4% 1.5% High 1.3% 1.4% 1.6% 1.7% PredationMedium Fallout None 0.6% 0.7% 0.9% 1.0% Low 1.2% 1.4% 1.7% 1.7% Medium 1.4% 1.6% 1.7% 1.9% High 1.6% 1.7% 1.9% 2.0% PredationHigh Fallout None 0.9% 1.0% 1.2% 1.4% Low 1.5% 1.7% 1.9% 2.0% Medium 1.7% 1.9% 2.0% 2.2% High 1.9% 2.1% 2.2% 2.4% Managementactionsascombinedscenarios Predatorcontrolat90%intheNorthwestKauaiscenarioproducedanincreaseinlambdafrom0.913to Thebenefitincreasedforthefirst25yearsbeforedeclining(Figure34).Thecontributiontothe population from predator proof fencing (predator eradication) increased over time then stabilized (Figure35).Chicktranslocationproducedadelayedcontributionthatgrewwiththepopulation(Figure 36).

38 36 Newell sshearwaterpopulationmodeling Figure34 Predatorcontrolfor120%ofKauaipopulationassuminglowfalloutwith0%SOSeffectivenessandno powerlinethreats(=0.985,includinga)growthrateofpopulationreceivingmanagement,andb)benefit expressedasthenumberofbirdsaddedtothetotalpopulationcomparedtonomanagement(=0.913). Figure35 Predator eradication for 110% of Kauai population assuming no fallout and no powerline threats (=1.000),includinga)growthrateofpopulationreceivingmanagement,andb)benefitexpressedasthe numberofbirdsaddedtothetotalpopulationcomparedtonomanagement(=0.906). Figure36 Newell s shearwaters in a translocated population of up to 20 chicks per year for five years, 21 to 40 chicksperyearfor10years,and41to60chicksperyearfor10years,includinga)projectedgrowthrate (=1.012),andb)translocationbenefitexpressedasthenumbersofbirdsaddedtothetotalpopulation comparedtonomanagement(=0.906).