PARENT SESSION
Contributed Oral Session 20: Modeling Animal Populations
Monday, August 8, 1:30 PM - 5:00 PM, Meeting Room 513 C, Level 5, Palais des congrès de Montréal

Factors influencing the probability of detecting Vermont songbirds during point counts.

Mitchell, Brian^*,1, Donovan, Therese¹, ¹ USGS - Vermont Cooperative Fish & Wildlife Research Unit, Burlington, Vermont, USA

ABSTRACT- Point count data for songbirds is difficult to interpret when the factors affecting detectability are poorly understood. The lack of a particular species during a count could be due to the species' absence from the survey area, or it could reflect a failure to detect the species when it is actually present. Recent advances in mark-recapture software and methods now allow explicit modeling of factors that affect a species' detectability, and the incorporation of these factors into occupancy models. We are investigating the distributions of bird species throughout Vermont in an effort to better understand the effects of land use change. An integral part of this research is delineating important factors affecting the detection probability of each species so that we can adjust our occupancy models accordingly. Our data set includes over 2,400 10-minute point counts taken at over 700 locations throughout Vermont during the late spring and early summer of 2003 and 2004. We have modeled the effects of time of day, time of season, habitat, and count type (i.e. with or without the playback of a chickadee mobbing tape) for ten commonly detected species, and we will discuss how these factors affected detection probability.

Key words: removal model, detection probability, Vermont, songbirds