Abundance estimation from point counts and territory mapping: comparing different approaches for two Regulus species

Abstract

The effectiveness of surveys of breeding birds varies due to multiple factors, with the primary being imperfect detection, which is particularly severe for elusive species. For example, the territory mapping method requires surveying an area multiple times a season to compensate for missing individuals during single surveys. Novel methods require much less effort in the field and include estimation of both detection probability and abundance corrected for individuals that went undetected. The aim of this study was to check if point counts and model-based results provide estimates similar to the ones from the territory mapping method. We studied the abundance of two forest birds—Goldcrest Regulus regulus and Firecrest R. ignicapilla—on three permanent census plots in the Białowieża Forest (E Poland). We compared abundance estimates resulting from the territory mapping method in its ‘standard’ (~ 10 visits) and intensive (~ 20 visits) approaches. We also performed point counts at the same plots using distance sampling methodology and hierarchical models in an attempt to get unbiased estimates by correcting for imperfect detection. We found that the standard territory mapping method produces much lower abundances than model-based estimates, which was particularly evident for the more numerous Firecrest. At the same time, results from point counts were more consistent with numbers from the intensive territory mapping. Our findings suggest that applying point counts and distance sampling models meet modern standards by considering various effects in abundance, availability and detection processes along with providing uncertainty of their estimates. We assume that our results might be applicable to other elusive species.