Calculating census efficiency for river birds: a case study with the White-throated Dipper Cinclus cinclus in the Pyrénées

Using the binomial law we modelled field data to estimate the probability (p̂) of detecting pairs of breeding White‐throated Dippers, and the population size (N̂± confidence limits). The model was divided into two parts according to whether the actual size of the population under study was known or not; in the latter case the truncated binomial model was used. Dipper abundance data were collected from three 4‐km‐long river tracts in the Pyrénées (France) during the breeding seasons of different years. Goodness‐of‐fit tests indicated that the binomial model fitted the data well. For a given visit during the survey, the estimated probability of detecting any pair of Dippers if they were present was always high (0.63–0.94) and constant from year to year but not between sites. Estimations (N̂) of the size of the population provided by the binomial model were very close to that derived from mapping techniques. This study provides the first ever quantification of the number of visits required to detect birds on linear territories: three visits were necessary to detect the whole breeding population.