Page 1
Eff icacyof birddeterrentdevices inEff icacyof birddeterrentdevices inagricultural areasof theFraserVal leyagricultural areasof theFraserVal leyof Brit ishColumbia: a pi lot studyofBrit ishColumbia: a pi lot study January2009Preparedfor:TheMinistryofAgricultureandLands(MAL),ProvinceofBritishColumbiaandTheFraserValleyRegionalDistrict(FVRD)PrincipalInvestigator:ProfessorKarenM.M.Steensma Co‐director,EnvironmentalStudiesProgram AssistantProfessor,DepartmentofBiology TrinityWesternUniversity
Studentassistants:AmandaEdworthy,KatherineHartline,DeniseWong,BeckyKernandAmyGardner
Page 2
Steensma - Efficacy of bird deterrent devices 2
ExecutiveSummary
AseriesofbirdcountstationsweresetupintheSumasPrairieregionofBritishColumbia,
CanadainJune2008inordertogaugetheefficacyofbirddeterrentdevicesoverasingle
blueberryharvestseason.Birdswerecountedintheearlymorningandlateafternoonatsix
differentstationsfortwoweekspriortotheinstallationofdeterrentsinJuly.Threestations
remainedascontrolcountsitesthroughoutthestudy.Theremainingthreesiteseachhada
differentdeterrentinstalled:anauditorybirddistresscaller,apropanecannon,andahawk
kite.CountscontinuedthroughAugustwithdeterrentsinplace,andonthroughSeptember
afteralldeterrentswereremoved.Statisticalanalysisrevealedthatboththehawkkiteandthe
propanecannonshowedsignificantdecreasesinstarlingpopulationnumbersinitially,but
starlingsslowlyreturnedtohighernumbersaftertheinitialintroductionofthesedeterrents.
Thehawkkitedeterrenteffectlastedlongerthanthepropanecannoneffectwithinthispilot
study.Thebirddistresscallershowednostatisticallysignificantdeterrenteffectalthough
problemsattheparticularsiteapparentlyskewedresultsforthatdeterrenttrial,asinitial
starlingpopulationsatthatsiteweresignificantlyhigherthanthecontrolpopulations.Though
theoverallstudywasofapilotnatureandextremelylimitedinscope,theresultspointtosome
possiblerecommendationsforstarlingmanagementandtothepotentialvalueoffurther
researchonthistopic.
Page 3
Steensma - Efficacy of bird deterrent devices 3
ListofFiguresandTablesFigure1.BirdcountsitesinSumasPrairieregionofBritishColumbia………..……………….…..page7Figure2.StarlingpopulationnumbersinresponsetointroductionofaBirdGardaudibledistresscalleratasiteinSumasPrairie,BC…………………………………….….page10Figure3.StarlingpopulationnumbersinresponsetointroductionofapropanecannonatasiteinSumasPrairie,BC…………………………………………………………….….page11Figure4.StarlingpopulationnumbersinresponsetointroductionofahawkkiteinSumasPrairie,BC…………………………………………………………………………………….….page12Figure5.PercentdifferenceinbirdnumbersattransitionpointsforintroductionandremovalofvariousdeterrenttechniquesatsitesinSumasPrairie,BC……………….……page13Figure6.PercentdifferenceinbirdnumbersattransitionpointsforintroductionandremovalofvariousdeterrenttechniquesatsitesinSumasPrairie,BC,withcontrolpopulationchangessubtracted………………………………………………………………..………..page14Table1.Summaryofeffectivenessofbirdcontroltechniques…………………………………,..…page18ListofAppendicesAppendix1.LandownercontactletterAppendix2.SampledatacollectionsheetsAppendix3.Summarystatisticalanalysis
Page 4
Steensma - Efficacy of bird deterrent devices 4
Introduction TheinvasiveEuropeanStarling(Sturnusvulgaris)reproducesprolificallyandishighly
adaptable,whichhasresultedinnuisancepopulationsacrossNorthAmerica(Currieetal.
1977).Knownfortheirintelligence,theseglossyblackbirdsalsospreaddiseases,suchas
gastroenteritis,footandmouthdisease,salmonella,aviantuberculosisandhistoplasmosis
(Feare1984,JohnsonandGlahn1994).Starlingscanbefoundinflocksofupto1.5millionbirds
andcauseanestimated$800millionindamageintheUnitedStatesannually(Norrisetal.
2003).
Starlingsposeaseriousthreattotheagriculturalindustry,especiallyberryfarmers.
Starlingsdescendonberryfields,consumingtheripefruitandsometimesdestroyingentire
crops(PrittsandHancock1992,Gough1994,Sweeney2003).InthePacificNorthwest,which
producesathirdoftheworld’sblueberrysupply,farmersreportlossesofupto30%annually
(WhatcomFarmFriends2008).Othersectorsofagriculture,mostnotablydairiesandorchards,
arealsoaffectedbythesebirds.Withdairycommoditybarnsandstoredforagesproviding
winterfeedandblueberriesprovidingsummerfeed,starlingsparticularlythriveatberry‐dairy
interfaceareasinthisregion(GilkesonandAdams1996,2000;Steensmaetal.2007).
Numerousmethodsofstarlingcontrolhavebeenattempted,butnonehasproventobe
theidealsolution.Auditorydeterrents,suchasaircannonsorelectronicdistresscalls,and
visualrepellentssuchasreflectivetape,havebeenusedtofrightenstarlingsaway,butthebirds
mayquicklyreturnoncetheyrealizetheyarenotharmed(Johnsonetal1985;Bomford1990;
BomfordandO’Brien1990;JohnsonandGlahn1994;GilkesonandAdams1996,2000;Harris
andDavis1998;Norrisetal.2003).Additionally,auditorydevicesincreasethepotentialforand
incidenceofconflictswithneighbours(PrittsandHancock1992,Gough1994,Johnsonand
Glahn1994).Trapscanbeusedtolegallyremovestarlings,buttheyareonlyeffectiveagainsta
fewbirdsratherthanentireflocks(PrittsandHancock1992,HarrisandDavis1998).Chemicals,
suchasMethiocarbandAvitrol,canefficientlyeliminatestarlings(Conover1982,Conover
1984,TobinandDolbeer1987,Bruggeretal.1993,Sweeney2003)thoughnaphthaleneshows
norepellency(Dolbeeretal1998).Thesechemicalsarenotavailableforuseonfruitcropsdue
Page 5
Steensma - Efficacy of bird deterrent devices 5
totoxicitytohumansandothernon‐targetspecies.Reflectivetapecanbesomewhateffective
againstsimilarbirds,incertaininstances,butbirdsmaybecomehabituatedtothistechnique
(Dolbeeretal1986).Nettingoffieldscanbeaneffectivemethodofpreventingcroploss,butit
isexpensiveandinterfereswithharvesterequipment(Conover1982;JohnsonandGlahn1994;
GilkesonandAdams1996,2000;FlintandDreistadt1998;Norrisetal.2003;Sweeney2003).
ConcernsaboutdamagetoagriculturalproductioninthelowermainlandofBritish
Columbia,combinedwiththeconflictsposedbyauditorydeterrentsusedinaregionofhigh
urban‐ruralresidentialgrowth,havecreatedademandforaccurateinformationontheefficacy
ofvariousbirddeterrentdevicesintheregion.Thepurposeofthispilotstudywastoassess
backgroundstarlingpopulationsandthentesttheeffectivenessofthreeparticulardevices–an
audiblebirddistresscaller,apropanecannon,andahawkkite–indeterringstarlingsandother
birdsfromagriculturalareasoftheFraserValley.
Methods
TheFraserValleyRegionalDistrictandtheBCMinistryofAgricultureapprovedfunding
basedonaproposalforthisprojectinspring2008.Faculty,studentsandstaffofTrinity
WesternUniversity(Langley,BC)andinternswiththeinternationalconservationgroupA
Rocha,assistedinthedesignanddatacollectionfortheproject.
Atimelineofeventsintheprojectoccurredasfollows:
• March‐April2008:ARochainternBeckyKernconductedvolunteerliteratureresearch
• May‐June2008:TWUstudentKatherineHartlineandIconductedfurtherliterature
research,begancontactinglandowners,mappedpotentialexperimentaltreatment
sites,andmappedthecropsgrowninthesurroundingareasusingaGIS(Geographic
InformationSystem).TWUstudentDeniseWongjoinedinconductinginitialbirdcounts
(allspecies)intheregionforbackgroundinformationonbirdpopulationsandtheir
habitsassiteswerechosen.
• July2008:DeniseandIcompletedlandownercontactsandprotocolwasestablishedat7
sites:4controlsites,1sitewithaPurivoxTriple‐Johnpropanecannonsetonlowest
Page 6
Steensma - Efficacy of bird deterrent devices 6
setting,1sitewithaBirdGardbrandaudibledistresscaller,and1sitewithahawkkite.
Allsiteswereprovidedwithadditionalfoodsources(beyondwhattheywouldforage
fromthesurroundingarea).Starlingsandotherincidentalspecieswerecountedforat
least2weekspriortointroductionofdeterrentdevicestoestablishbaselinepopulation
numbers.Siteswerecountedtwiceperday,onceintheearlymorning,andonceinthe
lateafternoon/earlyevening.Atleast10‐15minuteswerespentateachsitevisit.
• August2008:DeniseandTWUstudentAmandaEdworthy,andvolunteerARocha
internAmyGardner,conductedcountsduringthebusiestportionoftheblueberry
harvestseason.Countscontinuedwithdeterrentdevicesinplaceat3ofthe7
locations.Dataentrytospreadsheetsbegan.
• September2008:AmandaandAmycontinuedtodothecounts.Deterrentdeviceswere
stoppedtheweekendofSeptember5,astheblueberryseasonwaswindingdown.
CountscontinuedthroughSeptember24,togaugebothseasonalbirdbehaviourand
effectofdeterrentremoval.
• October2008:DataentrywascompletedbyAmandaandspreadsheetswere
reformattedforstatisticalanalysis.WeconsultedwithstatisticiansatbothTWUand
UniversityofBritishColumbiainconductingtheanalysis.
• November2008:Rawdatawereconvertedforentryintoastatisticalanalysisprogram
atUBC.Graphsandchartssummarizingstatisticalresultsweredeveloped.Mappingof
SumasPrairiewasfinalized.
• December2008:Draftfinalreportwaswritten.
Initiallanduseanalysisandbirdsurveysduringthelatespringandearlysummer
showedthatdairyfarmsandberryfarmsweremostlikelytoharbourlargepopulationsof
starlings.Duetotheconfoundingfactorspresentedbyblueberrygrowers’ownuseof
deterrents,andvariableattractivenessofblueberrycropsindifferentfields,treatmentsites
weredeliberatelychosenawayfromexistingblueberryfarms(Fig.1).Sevensiteswereinitially
chosenandlandownerpermissionwasobtained.Allbutoneofthesesiteswereondairyfarms;
theseventhwasonasodfarm.Asthesodfarmsitedidnotshowanappreciablebird
population,thesitewaseventuallydroppedfromthestudy.
Page 7
Steensma - Efficacy of bird deterrent devices 7
Figure1.SumasPrairie,BCstarlingbehaviorstudysites
1=Cedarwal;controlsite 4=Meier;hawkkite
2=Driessen;controlsite 5=Luymes;controlsite
3=Driessen;BirdGardsite 6=Kielstra;propanecannonsite
Page 8
Steensma - Efficacy of bird deterrent devices 8
Asheetofplywoodwasplacedonthegroundateachsite,inproximitytobirdroosting
areasandtoawatersource.Theplywoodwaskeptevenlycoveredwithascatteringofcracked
cornandblueberriestoserveasastablefoodsourceforthebirds.Duringeachsitevisit,the
observerswoulddriveuptowithin15‐20mofthefeedingstationandremaininthevehicle,
usingthevehicleasablindsoasnottoalarmthebirds(asperHolmes1993). Birdsonand
withina100mradiusofthefeedingstationwerecountedusingbinocularsandbirdguidesto
identifyspecies.Anyunusualindividualorflockbehaviourswerealsonoted.
Atthetimeofdeterrentintroduction,threesiteshaddevicesinstalledasfollows:1)the
BirdGardwassetupatadistanceof20mfromthebirdfeedingstationandsetaccordingto
dealerspecificationstomakenoiseatafrequenttimeintervalduringdaylighthours;2)the
propanecannonwassetupatadistanceof50mfromthebirdfeedingstationandwasset
accordingtodealerspecificationstofireattheleastfrequenttimeintervalfrom6:30amuntil
8:00pm;3)thehawkkitewassetupatadistanceof20mfromthebirdfeedingstation.All
deterrentswereleftunchangedintheirpositionandinthefrequencysettingofnoisefromthe
timeofinitialset‐uptothetimeofremoval.
Datawereorganizedforpresentationshowingmeandailystarlingcountsgroupedinto
periodsofninetotwelvedays,showingtrendsandtransitionsresultingfromaddingand
removingthedeterrents.Thepropanecannonandhawkkitesiteshadtwopre‐deterrent
periods(July9toJuly29),fourperiodswhilethedeterrentwasdeployed(July30toSeptember
7),andtwopost‐deterrentperiods(September8toSeptember24).TheBirdGardwasremoved
slightlyearlier(September1)duetobatteryfailure,andthushadonlythreedeterrent‐
deployedperiods.Wecompareddatafromthedeterrentsitestodatafromthethreecontrol
sites,whichwereaveragedtoshowbackgroundtrendsinstarlingabundance.Inaddition,
deterrentsiteswerecomparedtocontrolsitesateachtimeperiodusingat‐test.Analysesand
graphicsweredoneusingthestatisticalpackageR(RVersion2.4.0;RDevelopmentCoreTeam
2006)atUBC.
Page 9
Steensma - Efficacy of bird deterrent devices 9
Percentdifferencewasusedtoassesstheoveralldegreeofchangeinstarling
abundanceateachoftwotransitionpoints:1)Thepre‐deterrenttodeterrent‐deployed
transitiontime,and2)thedeterrent‐deployedtopost‐deterrenttransitiontime.
Results
ThevastmajorityofbirdsseenfrequentingthetreatmentsiteswereEuropeanstarlings,
althoughoccasionalredwingblackbirds,robinsandotherbirdswereseen.Aswell,occasional
predatorbirdssuchasbaldeagles,redtailhawks,harriers,andmerlinswereseeninthe
vicinity.
Ofthethreedeterrentsused,theBirdGardaudibledistresscaller(Fig.2)appearedto
showtheleasteffect.Infact,thesitehadahighstarlingpopulationtobeginwith,andthis
becameastatisticallysignificantdifferencejustpriortodeterrentintroduction.Midway
throughtheuseofthedeterrent,thestarlingnumbersactuallyclimbedto110,muchhigher
thanbothcontrolandpre‐deterrentnumbers.Birdpopulationsdidbegintodropaftermid‐
August.Anotherfeaturetonotewasthatanonsitemanurelagoonaroundwhichthestarlings
werecongregatingwaspumpedoutonAugust27th,atwhichtimethebirdcountdroppedto
fewerthan20birds.Thus,duringthetimeitwasoperating,itappearedthattheoveralltrend
wasfortheBirdGardtoactuallyattractbirdstothesiteinitially,followedbyaprecipitousmid‐
deterrentdrop.However,thedatahereshouldnotbeconsideredvalidafterthedateofthe
pumpingincidentsincethatapparentlyremovedamajorfoodsourcethathadbeenattracting
thebirds.
Thepropanecannonshowedaneffectonstarlingnumbersatthesitetested,withpeak
effectoccurringwithinthefirst15daysafterintroductionofthedeterrent(Fig.3).Though
starlingnumbersremainedlowerthanbackgroundcontrolnumbers,theydidbegintoclimb
againafter15days,showingnosignificantdifferencefromthecontrolafterthepeak
effectiveness.
Thehawkkiteshowedthestrongesteffectofthethreedeterrents(Fig.4).Starling
numberstookthesteepestdiveduringtheperiodinwhichthekitewasintroduced,withpeak
effectoccurring15daysafterkiteintroduction,andbirdnumbersremainingsignificantlylower
thanbackgroundstarlingpopulationlevelsfor10moredays.Numberscontinuedtoremain
Page 10
Steensma - Efficacy of bird deterrent devices 10
lowerthancontrolsanddidnotfullyoverlapwithcontrolnumbersuntilabout15daysafterthe
kitewasremoved.
Figure2.Starlingpopulationnumbersinresponsetointroductionofa BirdGardaudibledistresscalleratasiteinSumasPrairie,BC.Filledtriangles= significantdifferencefromcontrol,open=notsignificant.
Page 11
Steensma - Efficacy of bird deterrent devices 11
Figure3.Starlingpopulationnumbersinresponsetointroductionofapropane cannonatasiteinSumasPrairie,BC.Filledtriangles=significantdifferencefrom control;open=notsignificant.
Page 12
Steensma - Efficacy of bird deterrent devices 12
Figure4.Starlingpopulationnumbersinresponsetointroductionofahawkkitein SumasPrairie,BC.Filledtriangles=significantdifferencefromcontrol;open=not significantlydifferent.
Page 13
Steensma - Efficacy of bird deterrent devices 13
Figure5.PercentdifferenceinbirdnumbersattransitionpointsforintroductionandremovalofvariousdeterrenttechniquesatsitesinSumasPrairie,BC.
Page 14
Steensma - Efficacy of bird deterrent devices 14
Figure6.PercentdifferenceinbirdnumbersattransitionpointsforintroductionandremovalofvariousdeterrenttechniquesatsitesinSumasPrairie,BC,withcontrolpopulationchangessubtracted.
Page 15
Steensma - Efficacy of bird deterrent devices 15
Percentdifferencesinstarlingpopulationsatthedifferentdeterrentsitesareshownin
Figures5and6.Theapparentlackofeffect,orseeminglypositiveinitialeffectoftheBirdGard
onstarlingnumbers,mayrelatetoconfoundingfactorsmentionedabove.Usingpercent
changenumbersillustratesevenmorestronglythesuccessofthehawkkite,andalsoshows
thatbirddecreasesinresponsetothecannonmayhavebeenmutedsomewhatbythegeneral
increaseinthecontrolpopulation.Nostatisticaltestingofpercentdifferencevalueswas
possibleduetothesmallscaleofthestudy.
Trendsinnine‐to‐twelve‐daymeans(asshowninFigs.2,3,and4)weremoreinteresting
thanactualt‐testsforsignificantdifference,thoughthetesthelpsgivemeaningtoerrorbars
andgivesanideaofhowmuchofadifferenceisneededforsignificance.Thet‐testswouldhave
beenmoremeaningfulifpre‐deterrentstarlingabundancehadbeenstatisticallyidenticalfor
controlanddeterrentsites.Controlsitesshowedconsistentgradualincreasesinstarling
numbersastheseasonprogressed.Thethreecontrolsites,whentakentogether,forman
indicationofbackgroundtrendinbirdnumbers,whichwasusefulforcomparisontotrendsat
siteswheredeterrentswereintroduced.Deterrenteffectscanalsobeexaminedbysimply
notingthechangesinpopulationlevelspre‐deterrent,deterrent,andpost‐deterrentin
Figures2‐4.
Discussion
Therelativelysimpletechniquesusedinthisstudy,andthelimitedscopeofthestudy
duetotimeandbudgetconstraints,shouldbestronglyconsideredwhenoneattemptsto
interpretresults.Inparticular,thelackofmultiplereplicationsofeachdeterrenttechniqueina
greaternumberoflocations,andoveranumberofgrowingseasons,withcomplete
randomizationoflocations,makesstatisticalinferencesdifficult(BomfordandO’Brien1990).
However,giventhepilot‐studynatureofthisresearch,severalstrengthsofourstudy
thathavebeenlackinginmanypreviousbirddeterrentresearchstudies,shouldbenoted:
1)Controlareaswereestablishedfarenoughfromdeterrentareastominimizepotentialof
“scared”birdsmovingintocontrolplots.Controlandtreatmentsitesdonotneedtobe
matched,althoughideallybothshouldbereplicated(BomfordandO’Brien1990).Onlythe
Page 16
Steensma - Efficacy of bird deterrent devices 16
controlswerereplicatedhere.2)Deterrentplotsdidnothavetheireffectsconfoundedby
surroundingfarmers’deterrentuse.3)Resultsofthestudywerenottiedtocropdamage
measurements,whichcanbehighlyvariablebasedonfieldproductivityandotherconfounding
factors.4)Sincethe“before”orpre‐deterrentphaseisnotacontrolonthe“after”orpost‐
deterrentphase(BomfordandO’Brien1990)duetoseasonalchangesinbackgroundbird
populationsandbehaviors,theestablishmentofcontrolsitesisimportantandallows
comparisonofdeterrentsitetrendstobackgroundtrends.Ourcontrolswerenotsurprisingin
showingoverallupwardtrendsinstarlingnumbersastheseasonprogressed,sincestarlings
typicallyfledgeseveralclutchesofyoungbirdsthroughouttheseason,whichthenjointhe
flocks(Bent1950,Elliot1964).Thustheinformationgainedfromourstudyhasmeritfor
examiningsingle‐caseeffectsofdeterrentsintheseparticularlocationsovertime,eventhough
replicationofdeterrentplotswasnotpossible.
Lackofeffectiveness,orevenapparentcounter‐effectiveness,ofaudibledistresscallers,
correspondswithpreviousresearchindicatinglimitedsuccessandgreatvariabilityofthis
technique(Johnsonetal1985,Feare1989,BomfordandO’Brien1990,Conover1994,Harris
andDavis1998).Inparticular,Conover(1994)foundthatdistresscallsareeffectiveonlyfora
shorttimeatbest,andareideallycombinedwithanactualpredator‐causedfatality,whichwas
notimplementedinourstudy.Sinceintroductionofactualnativepredatorssuchassharp‐
shinnedorCooper’shawks,merlins,orkestrelfalconswasnotconducted,thereisnowayto
measurewhetherthiscombinationofdeterrentswouldhaveworked.
ItisworthnotingthattheoperatorofthedairyfarmatwhichtheBirdGardwasplaced,
indicatedneartheendofourprojectthathehadpreviouslytriedusingaBirdGardinanearby
locationafewyearsearlier.Hefoundittobeeffectiveforatimebutdeterminedthatthe
effectivenessdecreasedtonilafterafewmonths.Itispossiblethatresidualolderbirdsfrom
thattimewerestillpresentatthetimeofourstudy,andmayhavebeenresistanttothe
BirdGardasaresult.OveralltheproblemsattheBirdGardsite,includingthemanurepump‐
out,limitconclusionsthatcanbedrawnregardingaudibledistresscallersonthissinglestudy.
Habituationofbirdstopropanecannonshasbeendocumentedinpreviousstudies
(BomfordandO’Brien1990,HarrisandDavis1998).Inourstudytherelativelyinfrequentuse
Page 17
Steensma - Efficacy of bird deterrent devices 17
ofthecannon(firingfrequencyofnomorethan3times/hour)mayhaveincreasedthelengthof
itseffectiveperiod,ashasbeendocumentedelsewhere(HarrisandDavis1998).The
comparisonoffiringfrequencies,however,wasbeyondthescopeofthisstudy.Propane
cannoneffectivenesscanalsobeincreasedifitismovedrandomlyaroundthefieldand
combinedwithvisualdeterrentssuchasartificialpredatorsorpop‐upscarecrows(Cummingset
al1986,BomfordandO’Brien1990).Aswiththedistresscallers,however,thebestincreasein
effectivenessofcannonscomeswhentheyarecoupledwithactualpredator‐causeddeathsor
shootingdeathsofbirds(BomfordandO’Brien1990,Conover1994).Shootingisnotanoption
withinoraroundblueberryfields,however,becauseoftheriskofcontaminatingfruitwithshot.
Olderresearchhasshowncannonstobeeffectiveatleastintheshortterm(Stickleyaetal
1972,DecalestaandHayes1979,PotvinandBergeron1981).
Effectivenessofhawkkites,aswellaseventualhabituationofbirdstothekites,has
beenshownpreviously(Conover1982,1984;Nakamura1997;HarrisandDavis1998).
Increasedrandomizationofhawkkiteplacement,whichwasnotdoneinourstudy,keepspest
birdsfrombecominghabituatedtothekites.Ideallythekitesshouldbemovedfrequentlyand
irregularly,andcombinedwithotherdeterrents(BomfordandO’Brien1990,HarrisandDavis
1998).Thiswasbeyondthescopeofoursimplepilotstudybutwouldcertainlybeofinterestin
pursuingnon‐auditoryoptionsforeffectivebirdcontrol.Forexample,hawkkitesin
conjunctionwithreflectivetape,actualshootingfatalities,orfalconrymaybehighlyeffective.
Consideringthesimplicityofourexperimentwiththehawkkite,itssuccessinthislocalewas
remarkable.
Successofanytechnique,whetherauditory(distresscaller,cannon)orvisual(hawk
kites,reflectivetape)isdependentontheattentivenessandskilloftheoperatorandrequires
randomizationofintervals,location,andcombinationsoftechnique(BomfordandO’Brien
1990,HarrisandDavis1998).Ourresultsagreewithlarge‐scalesurveysofvarioustechniques
thatrankeffectiveness(Table1).Professionalfalconry,conductedbyanexpertfalconeronsite
throughoutdaylighthours,isavirtualguaranteeofeliminatingbirddamage(HarrisandDavis
1998).Howeverthisisanexpensivealternative.Aclearunderstandingofbirdbiologyand
habitatrelatedtotheuniquenessofeachindividualsite,asprovidedbyaprofessionalbiologist,
Page 18
Steensma - Efficacy of bird deterrent devices 18
isanimportantsteptoeffectivebirdcontrol.Habitatrecommendations,suchasprovisionof
predatorperchsitesandnestboxes,canthenbeimplemented(HarrisandDavis1998).
Table1.Summaryofeffectivenessofbirdcontroltechniques
Not recommended Limited recommendation Highly recommended High intensity, infra- or ultra-sound
Hawk kites and other visual predator simulations
Habitat modification to encourage natural predators such as birds of prey
Lasers Distress and predator callers such as BirdGard
Falconry
Lights Propane cannons Active shooting of pest birds
Microwaves Reflective tape Habitat modification to discourage pest birds
Smoke Scarecrows Netting, when feasible (as in small, handpicked fields)
(adaptedprimarilyfromHarrisandDavis1998.Recommendationsandeffectivenessarebasedon operationbyskilledpersonnelwithanunderstandingofbiologyofbothpestandpredatorbirds) .
Thebestpossibledeterrentforpestbirdsisnaturalpresenceofnativepredatorbirds
(FlintandDreistadt1998,Daly2002,Steensma2008),whichmayexplaintherelative
effectivenessofthehawkkitehere.SeveralstudieshaveshownthattheAmericankestrel
(Falcosparverius)caneffectivelypreyon,orterritoriallydefendagainst,theEuropeanstarling
(Balgooyen1976,Ashkam1990,Village1990,Suhonenetal.1994,BechardandBechard1996,
Parrish2000,Dover2008).AsthekestrelisnativetothePacificNorthwestbutdecliningin
numbersinWesternWashingtonandtheBCLowerMainlandoverthepast20years(Wahl
1995,Ireland2008;Pike2008),enhancementofnesting,foragingandperchinghabitatforthis
speciescouldprovidealow‐costalternativecontrolmeasureagainststarlings(TolandandElder
1987,DawsonandBortolotti2000,Valdezetal2007).Apilotkestrelreleaseprogramhas
showninitialsuccessinWashingtonState(Steensma2008)andisplannedforexpansioninto
BritishColumbia.
Page 19
Steensma - Efficacy of bird deterrent devices 19
SummaryandRecommendations Basedonbothourliteratureandfieldresearch,wewouldsummarizeandmakethe
followingpreliminaryrecommendationsregardingbirddeterrenttechniques:
• BirddistresscallerssuchastheBirdGardmaybeeffectiveinsomesituationsbutwere
notshowntobeofuseinthisparticularsettingandapplicationduetositeproblems
• Propanecannonscanbeofuseinshorter‐termapplications
• Hawkkitescanbeofuseinshorter‐termapplications
• Combinationsofaudibledistresstechniquesandactualshootingofpestbirdsare
moreeffectivethanaudibledistressalone.Shootingisnotanoptionwithinor
adjacenttoblueberryfields,however,duetotheriskofcontaminatingfruitwithshot.
• Hawkkitesincombinationwithactualshootingorpredator‐causedfatalitiesaremore
effectivethanhawkkitesalone
• Alloftheabovetechniquesareenhancedthroughrandomizationand
combination/integrationofaudibleandvisualdevices
• Falconryandpresenceofnaturalpredatorsareconsideredhighlyeffectivedeterrents
• Enhancementofnativebirdsofpreythroughhabitatimprovement,andreleaseof
youngbirdsofpreyshouldbeatoppriorityforcost‐effectivebirdcontrolresearch
Acknowledgments
ThisstudywasmadepossiblebyajointgrantfromtheBritishColumbiaMinistryofAgricultureandtheFraserValleyRegionalDistrict,andspecialthanksaregiventoDirectorDickBogstieandMs.SiriBertelsonofFVRD,andtoMr.MarkSweeneyandMr.BertvanDalfsenofMAL.Wearegratefulforthegraciousnessofthefollowingfarmsinallowingourpresenceduringthecourseofthestudy:CedarwalFarms,DriessenFarms,BosSod,MeierDairy,LuymesDairy,andKielstraDairy.TheassistanceofMr.MarkSweeneyinobtainingplywoodandfeedwasverymuchappreciated.Theco‐operationofHi‐TechDistributorsinprovidingandsettingupthedeterrentdeviceswasappreciated.Ms.MelissaOakesofTWUandARochaCanadaprovidedbudgetoversightandDr.TracyStobbeoftheTWUSchoolofBusinessconsultedonthestatisticalanalysis.Mr.PatrickLilleyofRaincoastAppliedEcologyandMr.GlenCarlsonofARochaCanadaprovidedinitialadviceonstudydesign.Finally,variousmembersofourfamiliesprovidedcompanyonsurveydaysandpatiencethroughouttheprocess.
Page 20
Steensma - Efficacy of bird deterrent devices 20
Literaturecited
Askham,L.R.1990.Effectofartificialperchesandnestsinattractingraptorstoorchards.ProceedingsfromtheFourteenthVertebratePestConferenceinLincoln,Nebraska. BalgooyenT.G.1976.BehaviorandecologyoftheAmericankestrel[dissertation].Berkeley,California:UniversityofCalifornia. BechardM.J.andJ.M.Bechard.1996.CompetitionfornestboxesbetweenAmericankestrelsandEuropeanstarlingsinanagriculturalareaofsouthernIdaho.In:Raptorsinhumanlandscapes.1sted.SanDiego,California:AcademicPressInc. BentC.1950.LifehistoriesofNorthAmericanwagtails,shrikes,vireos,andtheirallies.U.S.NationalMuseumBulletin197. Bomford,M.1990.Ineffectivenessofasonicdevicefordeterringstarlings.WildlifeSocietyBulletin. Bomford,M.andO’BrienP.1990.Sonicdeterrentsinbirddamagecontrol:areviewofdevicetestsandeffectiveness.WildlifeSocietyBulletin18:411–422. Conover,M.R.1994.Howbirdsinterpretdistresscalls:implicationsofappliedusesofdistresscallplaybacks.Proc.SixteenthAnnualVertebratePestConference.Lincoln,Nebraska. CurrieF.A.,D.ElgyandS.J.Petty.1977.StarlingRoostDispersalfromWoodlands.ForestryCommissionLeaflet69. DalyM.A.2002.AmericanKestrel:TheBarnyardFriend.CollegePark,Maryland:UniversityofMarylandCooperativeExtension. DawsonD.R.andG.R.Bortolotti.2000.ReproductivesuccessofAmericanKestrels:theroleofpreyabundanceandweather.Condor102:814‐822. DoverB.S.2008.Owner,OrchardGuard,Yakima,Washington.Personalcommunication. ElliotH.N.1964.StarlingsinthePacificNorthwest.VertebratePestConferenceProceedingsCollection:Proceedingsofthe2ndVertebratePestControlConference,U.S.BureauofSportFisheriesandWildlife,Portland,Oregon. FeareC.1984.TheStarling.Oxford,England:OxfordUniversityPress. FlintM.L.andS.H.Dreistadt.1998.Naturalenemieshandbook.Berkeley,California:UniversityofCaliforniaPress.
Page 21
Steensma - Efficacy of bird deterrent devices 21
GilkesonL.A.andR.W.Adams.1996.IntegratedpestmanagementforstructuralpestsinBritishColumbia.4thed.Vancouver,BritishColumbia:BritishColumbiaMinistryofEnvironment,Land,andParks;PollutionPreventionandPesticideManagementBranch. GilkesonL.A.andR.W.Adams.2000.IntegratedpestmanagementmanualforlandscapepestsinBritishColumbia.3rded.Vancouver,BritishColumbia:BritishColumbiaMinistryofEnvironment,LandsandParks;PollutionPreventionandRemediationBranch GoughR.E.1994.Thehighbushblueberryanditsmanagement.1sted.NewYork,NewYork:FoodProductsPress. HarrisR.E.andR.A.Davis.1998.Evaluationoftheefficacyofproductsandtechniquesforairportbirdcontrol.LGLLimitedforAerodromeSafetyBranch,TransportCanada. HolmesE.L.1993.Responsesofwinteringgrasslandraptorstohumandisturbance.WildlifeSocietyBulletin21:461‐468. Ireland,J.2008.Director,ReifelBirdSanctuary,Delta,BC.Personalcommunication. Johnson,R.J.,P.H.Cole,andW.W.Stroup.1985.Starlingresponsetothreeauditorystimuli.J.WildlifeManagement49:620‐625. JohnsonR.J.andJ.F.Glahn.1994.Preventionandcontrolofwildlifedamage:Europeanstarlings.In:WildlifeDamageManagementHandbook.2nded.Lincoln,Nebraska:UniversityofNebraska. Nakamura,K.1997.EstimationofEffectiveAreaofBirdScarers.J.WildlifeManagement.61(3):925‐934 NorrisR.F.,E.P.Caswell‐ChenandM.Kogan,editors.2003.Conceptsinintegratedpestmanagement.1sted.UpperSaddleRiver,NewJersey:PrenticeHall. ParrishJ.W.2000.PossiblepreventionofEuropeanstarlingnestingbySoutheasternAmericanKestrelsatapowersubstationinsouthernGeorgia.J.RaptorResearch34:152. PikeD.2008.MasterFalconer,Lynden,Washington.Personalcommunication. PrittsM.P.andJ.F.Hancock,editors.1992.Highbushblueberryproductionguide.1sted.Ithaca,NewYork:NortheastRegionalAgriculturalEngineeringService. Steensma,K.M.M.2008.AdvancesinbirddeterrentmethodsforagriculturalareasofsouthwestBritishColumbiaandnorthwestWashington.Invitedpresentation,WSUWesternWashingtonSmallFruitWorkshop,Dec.16,Lynden,Washington.
Page 22
Steensma - Efficacy of bird deterrent devices 22
Steensma,K.M.M.,B.DoverandK.Hartline.2007.Useofkestrelsforbiologicalcontrolofstarlingsonfruitanddairyfarms.Invitedpresentation,WSUWesternWashingtonSmallFruitWorkshop,Dec.5,Lynden,Washington. SuhonenJ.,K.Norrdahl,andE.Korpimaki.1994.Avianpredationriskmodifiesbreedingbirdcommunityonafarmlandarea.Ecology75(6):1626‐34. Sweeney,M.2003.AnoverviewoftheBChighbushblueberryindustry.Abbotsford,BritishColumbia:BritishColumbiaMinistryofAgriculture,FoodandFisheries,IndustryCompetitivenessBranch. TolandB.T.andW.H.Elder.1987.Influenceofnest‐boxplacementonabundanceandproductivityofAmericanKestrelsincentralMissouri.WilsonBull.99:712‐17. ValdezU.,S.Robertson,B.RobertsonandK.L.Bildstein.2007.NestboxusebyAmericankestrels(Falcosparverius)andEuropeanstarlings(Sturnusvulgaris)ineasternPennsylvania.PennsylvaniaBirds14(3):150‐53. VillageA.1990.Thekestrel.London,England:TandAPoyser. Wahl,T.R.1995.BirdsofWhatcomCounty:statusanddistribution.Bellingham,WA:TRWahl.184pp. WhatcomFarmFriends.2008.Starlingcontrol.FebruaryFarmFlash.
Page 23
Steensma - Efficacy of bird deterrent devices 23
Appendix1.Landownercontactletter 30 June 2008 To: Select landowners in the Sumas Prairie area As you are likely aware, the presence of starlings has increased in our area in recent years. The resulting damage to farm crops, and the efforts to deter these birds through various devices such as propane cannons and other means have created some controversy in the community. The Ministry of Agriculture and the Fraser Valley Regional District have commissioned me to carry out a study this summer, to survey the overall bird population in the Sumas Prairie area, and to test the usefulness of several starling deterrent techniques. Assisted by senior-level Trinity Western University environmental studies and biology students, and by volunteers with specific expertise on starlings, I will undertake the following approach:
• Identify locations in which starlings are particularly numerous • Deliberately create feeding stations to attract starlings regularly • Count numbers of starlings frequenting the stations • Introduce various deterrent techniques to some of the feeding stations • Continue to count starling numbers at all stations through September • Analyze results statistically to determine effectiveness of these deterrents
This is a pilot study, and part of a larger overall study in the Fraser Valley and northwest Washington State, that is seeking solutions to starling problems. The goal is to identify solutions that are workable for farmers, farm neighbors, and the natural environment. Your land may provide a suitable location for one of the feeding stations, if you are willing to participate in this research. If you are interested in allowing us space for the project, we will attempt to minimize disruption to your property, and the study will be concluded by mid- to late September. Thank you for your consideration, and please feel free to contact me with any questions. Karen M.M. Steensma Assistant Professor of Biology Co-director, Environmental Studies Program [email protected] Other partners: BC Blueberry Council, Whatcom Farm Friends, Washington Department of Fish and Wildlife, Washington State University Cooperative Extension, A Rocha Canada & USA, and various BC & Washington dairy and blueberry farmers
7600 Glover Road Langley, BC V2Y 1Y1 Canada Tel: (604) 888-7511
Page 24
Steensma - Efficacy of bird deterrent devices 24
Appendix2.Sampledatacollectionsheets Luymes Time:9:21am Weather:Grayskies,coolbreeze FeedAdded?None Table7:AvianObservationonJuly23rd2008
Species Number OtherObservationsLatinName
CommonName
Onfeed
Withina100mradius
HouseSparrow 0 12
RockPigeon 0 3
Robin 0 1
EuropeanStarling 0 25
Nobirdsseenonfeed.Allblueberriesarestillpresentonfeed.
Kielstra Time:7:50am Weather:Cold,cloudy,noprecipitation FeedAdded?None Table13:AvianObservationonJuly31st2008
Species Number OtherObservations
LatinName CommonName Onfeed Withina100mradius
EuropeanStarling 0 102
FemaleBrewer'sBlackbird 0 2
RockPigeon 0 18
HouseSparrow 0 3
Cannonfired,birdsfledyetwithinaminuteofthelastblastbirdsreturned(20returnedtothepowerline)
Page 25
Steensma - Efficacy of bird deterrent devices 25
Appendix3.Summaryoft‐testvaluesfortimeperiodsshowninFigs.2,3,4
Characteristics of starling populations at deterrent and control sites throughout the experiment
Deterrent status (time period) mean ± SE
Test statistic df p*
Site Deterrent Control
Off (1) 41.1 ± 18.0 12.6 ± 6.4 1.55 5 0.18 Off (2) 52.1 ± 19.8 23.6 ± 10.9 1.37 6 0.22 On (3) 10.1 ± 7.4 27.5 ± 8.7 -1.93 14 0.07 On (4) 4.7 ± 1.8 42.6 ± 16.3 -3.96 21 0.00 On (5) 18.1 ± 7.1 69.3 ± 34.9 -2.33 19 0.03 On (6) 42.8 ± 16.8 69.7 ± 42.3 -0.91 13 0.38 Off (7) 57.8 ± 14.6 94.4 ± 46.4 -1.20 22 0.24
Kite
Off (8) 58.2 ± 18.3 82.7 ± 34.1 -0.91 17 0.38
Off (1) 21.2 ± 9.2 12.6 ± 6.4 0.86 7 0.42 Off (2) 62.1 ± 18.2 23.6 ± 10.9 2.00 6 0.09 On (3) 55.0 ± 21.2 27.5 ± 8.7 1.26 8 0.24 On (4) 13.4 ± 4.4 42.6 ± 16.3 -2.81 26 0.01 On (5) 43.2 ± 14.7 69.3 ± 34.9 -1.03 20 0.32 On (6) 28.5 ± 8.0 69.7 ± 42.3 -1.60 13 0.13 Off (7) 60.0 ± 14.8 94.4 ± 46.4 -1.12 22 0.27
Cannon
Off (8) 75.8 ± 27.6 82.7 ± 34.1 -0.21 11 0.84
Off (1) 56.2 ± 19.7 12.6 ± 6.4 2.17 5 0.08 Off (2) 87.8 ± 8.7 23.6 ± 10.9 5.99 11 0.00 On (3) 67.8 ± 15.9 27.5 ± 8.7 2.42 9 0.04 On (4) 115.0 ± 16.9 42.6 ± 16.3 3.74 10 0.00 On (5) 60.9 ± 17.2 69.3 ± 34.9 -0.31 18 0.76 Off (6) 23.4 ± 21.2 69.7 ± 42.3 -1.43 11 0.18 Off (7) 10.8 ± 3.8 94.4 ± 46.4 -3.09 18 0.01
Bird Guard
Off (8) 7.6 ± 3.0 82.7 ± 34.1 -3.77 18 0.00 *Values bolded if p ≤ 0.05; indicates significant difference between deterrent and control means.