Producing regional estimates of population size for common and widespread breeding birds from national monitoring data

Abstract

Capsule Line transect data from large-scale common bird monitoring schemes can be used to provide population estimates for many common and widespread species at a regional level. Aims To test the sample sizes necessary to derive regional population figures and to explore the precision, accuracy and bias of the results. Methods As a case study, distance sampling methodology was applied to transect survey data from the ornithological recording area of Cheshire and Wirral (England), which constitutes about 1% of UK land area. We used data from 109 transects, either surveyed through the UK Breeding Bird Survey (BBS) in 2004 or 2005 or from a project in 2005 to monitor farms in the Entry Level Stewardship Scheme (ELS); this is 1 in 23 of the 1-km squares in the region. Results There were sufficient observations to fit robust detection functions and produce regional-scale population estimates for 61 species – half of the 122 regularly breeding species. Population estimates were further derived for another 16 less common species by applying a detection function from a more common surrogate species. However, some of these population estimates are not considered to be accurate, as judged by local knowledge, the species' distribution shown by a recent tetrad-scale atlas, and as a proportion of the national figures calculated from the national BBS data. The most precise estimates are for species that are evenly distributed across the study area, with decreasing precision for those with a more clumped distribution or species occurring in restricted habitats. Conclusion The present study highlights the potential of these methods and national bird monitoring data for deriving population estimates for many species at a smaller regional scale.