Testing common assumptions in studies of songbird nest success

Abstract

We studied Ovenbird Seiurus aurocapilla and Golden-winged Warbler Vermivorachrysoptera populations in northern Minnesota, USA, to test two common assumptions instudies of songbird nest success: (1) that the condition of an empty nest on or near itsexpected fledge date is an indicator of nest fate; and (2) that the presence of a fledgling orfamily group within a territory confirms a successful nest in that territory. We monitoredthe condition of nests and used radiotelemetry to monitor juveniles through the expectedfledging date and early post-fledging period. Of nests that contained nestlings 1–2 daysbefore the expected fledge date, fates were misidentified using nest condition alone for9.5% of Ovenbird nests, but those misidentifications were made in both directions (suc-ceeded or failed), yielding only a small bias in estimated nest success. However, 20% ofGolden-winged Warbler nests were misidentified as successful using nest condition duringthe final visit interval, biasing the nest success estimate upward by 21–28% depending onthe treatment of uncertain nest fates. Fledgling Ovenbirds from 58% of nests travelledbeyond their natal territory within 24 h, rising to 98% after 5 days, and those fledglingstravelled up to 390 m from nests within 10 days of fledging. Fledgling Golden-wingedWarblers from 13% of nests travelled beyond their natal territory within 24 h, rising to85% after 5 days, and those fledglings travelled up to 510 m from nests within 10 days offledging. We conclude that nest condition and fledgling presence can be misleading indica-tors of nest fate, probably commonly biasing nest success estimates upward, and we rec-ommend that these assumptions should be tested in additional species.Keywords: fledgling, Golden-winged Warbler, Ovenbird, Seiurus aurocapilla, telemetry,Vermivora chrysoptera.Estimates of songbird reproductive success, typi-cally limited to nest data, are used to assess habitatquality (e.g. Weinberg & Roth 1998), model pop-ulation dynamics (e.g. Podolski et al. 2007), iden-tify source and sink populations (e.g. Donovan etal. 1995), and inform conservation and manage-ment plans (e.g. Woodworth 1999). Althoughsongbird population growth may be generally moresensitive to adult annual survival and fledgling sur-vival (Donovan & Thompson 2001, Streby &Andersen 2011), population growth is also sensi-tive to variation in nest success (Donovan et al.1995), and nest success is the only directly esti-mated parameter in most studies of songbirdreproductive success (Anders et al. 1997). Manypopulation models account for re-nesting (birdsnesting again after initial failure) and estimates ofnest productivity (number of young produced persuccessful nest). All such studies require accuratefield identification of whether each monitored nestsucceeded or failed in producing young. However,observational studies of songbird nests often