Monitoring Passerine Reproduction by Constant Effort Ringing: Evaluation of the Efficiency of Trend Detection

摘要

TheDutchConstantEffortSite(CES)programmehasbeenoperatingsince1994andisespeciallydesignedfortheanalysisofdemographicparameters.Currently,itworkswith40activemist-nettingsitesandhasadatabasewithmorethan250,000recordsatitsdisposal.Here,weaskwhethertheeffortinvestedintheprogrammeisappropriatetodetecttemporaltrendsofproductivityofpasserinebirdpopulationsacrossTheNetherlands.WespecificallyaskiflesseffortwouldproducethesameresultsorifthenumberofCESsshouldbeaugmented.Toevaluatethesequestionsweusedaresamplingapproachinwhichwedecreasedsamplesizeinstepsoffiveanditeratedresampling1000timesforeachstep,thussimulatingCESprogrammesinwhichfewersiteswereoperated.Forthetwentymostabundantspeciesandforeachsamplesize,linearregressionsoftheproductivityindex(alogisticgeneralizedlinearmodeloftheproportionofcapturedjuvenilebirds)onyearwerecomparedwiththeregressionobtainedusingthecompleterecords.Theproportionofsamplesthatyieldedsignificantpositiveornegativeslopeswasdetermined,aswell.Forthemajorityofspecieswithoutsignificanttemporaltrends(18),theproportionsofsignificantslopesatsmallersamplesizeswerelowerthan20%,indicatingamoderateriskofcommittingTypeIerrors(detectinganegativeorpositivetrendalthoughproductivitydidnotchangeovertime).FortheGardenWarblerSylviaborin,whichshowedasignificantpositiveproductivitytrend,theprobabilityofcommittingTypeIIerrors(notdetectingexistingtrends)increasedrapidlywithdecreasingsamplesize.WeconcludethattheDutchCESprogrammeworkswithasufficientnumberofsitestodetectreliabletemporaltrendsformostofthemoreabundantpasserinespecies.However,increasingthenumberofsiteswouldallowforamoresecuredeterminationofproductivitytrendsforthosespeciesthatcurrentlyshowambiguousresults(ReedWarblerAcrocephalusscirpaceus,WillowWarblerPhylloscopustrochilus,WillowTitParusmontanus).