Performance of Three Alternative Estimators of Stream Residence Time Based on Live and Dead Counts of Salmonids

摘要

Data from live and dead fish counts at salmonid spawning areas have been used for decades to estimate stream residence time. Simulation studies indicate that the peak-to-peak and median-to-median estimates of stream residence time will have at least a 50% negative bias under the best of circumstances. A new estimation technique called the 'expectation-to-expectation' method is shown to be approximately unbiased over a range of arrival and lifetime distributions. The method also allows corrections for incomplete detection probabilities and carcass retention rates less than 1.0. An important byproduct of the approach is a concurrent estimate of total escapement.