Capture probability of fishes in Central European (Hungary) wadeable lowland streams

Neglect of imperfect capture efficiency leads to biased inferences on population abundance, and correspondingly, seriously affects ecological research, bioassessment, conservation, and fisheries management. To date, many research studies have studied capture efficiency of salmonid fishes, but the catchability of fishes living in non-salmonid streams has received much less attention. This paper estimates capture probability for seven fish species in densely vegetated lowland streams by using double-pass electrofishing data and an N-mixture removal model. Results show that capture probability can vary among species, and between-stream differences have a stronger influence on the abundance and the catchability than within-stream variability. Estimation uncertainty decreases with observed abundance, and the mean catchability tends to be the highest for the medium abundant species. These findings suggest that relative abundances from single-pass data are biased to a species- and habitat-specific degree. Therefore, plausible estimation of capture probability from double-pass electrofishing requires data collected from numerous sites that cover a wide range of the environmental gradient in lowland streams.