Detection probability of the Collared Flycatcher Ficedula albicollis during quick, multiple surveys: a case study in an isolated population in northern Poland

Abstract

We investigated factors affecting the detection probability of the Collared Flycatcher Ficedula albicollis, a secretive forest passerine, whose range is confined to central and eastern Europe. With a robust design sampling protocol, 78 sites were surveyed four times in spring 2012 for the presence of singing males in the Las Warmiński nature reserve (northern Poland). Multi-season occupancy models fitted to the data showed that detection probability varied considerably as a function of habitat quality. Probabilities of detection of at least one singing male during a single survey were very similar under various models: 0.383-0.419 at sites with single males, 0.815-0.837 at sites with two males, 0.970-0.974 and 0.999 at sites with three and four males. Detection probability did not depend on time of day, indicating no decline in singing activity as the day progresses. The local extinction probability was zero, suggesting that sites occupied early in the season are unlikely to be deserted - males were actively singing several days after arrival. Since mated males reduce their singing activity very much, these results may suggest low mating success. If this holds true, the colonization probability invariably estimated at 1.0 is best explained by the late arrival of some males rather, than by within-season movements of males already present. This study emphasizes that occupancy models are useful for the simultaneous investigation of the effects of multiple covariates not only on detection but also on other, biologically important parameters.