Efficiency of One-Visit Censuses of Bird Communities Breeding on Small Islands

Publisher Website
Google Scholar
Abstract
Breeding land bird communities of 2 islands in the Aland archipelago (Finland) were studied. On Gasholmen (14 ha) a 1-visit census covered the whole island, while on Bockholm (38 ha) 2 line transect censuses were made through the island and conversion coefficients were used to derive estimates of total numbers. The results were compared with those of mapping censuses. The proportion of breeding species observed in the 1-visit censuses was 89% on Gasholmen and 88% on Bockholm; the percentages agree with theoretical expectations derived by rarefaction. On Gasholmen, 70% of the breeding pairs were observed in the 1-visit census. On Bockholm, the estimate of the total density was 95% of the mapping result; the high apparent efficiency was probably a consequence of a biased distribution of habitats along the transects. Shannon''s diversity, rarefaction curves, and the species-abundance distributions derived from the logseries model were similar in both censuses on both islands. One-visit censuses are thus, adequate for quantitative comparisons of community parameters of island bird communities. A special problem is posed by surplus species (and individuals), either visitors from neighboring islands or non-breeding individuals. The possible existence of transients in the observation data of 1-visit censuses must be taken into account in interpreting the results.