Precision of diving and canoe-based visual estimates of Atlantic salmon ( Salmo salar) abundance

Abstract

Precision (repeatability) of visual counts of Atlantic salmon ( Salmo salar L.) by divers and canoeists was evaluated in the Northwest Upsalquitch River, New Brunswick, in October 1995. Diver counts were more precise (coefficient of variation, CV 3.9–19.5%) than canoe-based counts ( CV 22.0–97.3%). Diver counts of large salmon were more precise ( CV 3.1–11.3%) than counts of grilse ( CV 7.6–19.5%), probably because of differences in fish behaviour. CVs of total counts were usually lower than CVs of grilse or large salmon counts, suggesting that observers varied more in their ability to correctly identify fish than in their ability to locate fish. Diver counts provided a reasonably precise index of abundance (95% confidence limits were within 16% of the mean in trials with three or more observers), but canoe counts were highly dependent on observer experience.