An assessment of carcass counting surveys with increasing time lapse following a simulated fish kill on a small upland stream

Abstract

AbstractA simulated fish kill was conducted on a small upland stream in Northern Ireland by planting out hatchery‐produced brown trout Salmo trutta L. carcasses of various size categories. Standard, post‐fish kill, assessment walkover surveys were conducted over time intervals to determine the number of carcasses visible. The sample variance between individual surveyors was generally low, with good agreement between the observed counts for the three, discrete, size fractions of fish up to 72 hr after the simulated fish kill. Despite low discharge rates, shallow water and good accessibility to the experimental stream, only 52% of the small category fish (<8 cm LF) were recorded 4 hr after the start of the simulated fish kill. Larger carcasses (>17 cm LF) were more visible, and >90% were detected 48 hr after the start of the simulated fish kill. After 96 hr, all size fractions of carcasses had reduced significantly, and the variability between replicate surveys increased markedly.