Efficiently estimating salmon escapement uncertainty using systematically sampled data

American Fisheries Society Symposium
By: , and 


  • The Publications Warehouse does not have links to digital versions of this publication at this time
  • Download citation as: RIS | Dublin Core


Fish escapement is generally monitored using nonreplicated systematic sampling designs (e.g., via visual counts from towers or hydroacoustic counts). These sampling designs support a variety of methods for estimating the variance of the total escapement. Unfortunately, all the methods give biased results, with the magnitude of the bias being determined by the underlying process patterns. Fish escapement commonly exhibits positive autocorrelation and nonlinear patterns, such as diurnal and seasonal patterns. For these patterns, poor choice of variance estimator can needlessly increase the uncertainty managers have to deal with in sustaining fish populations. We illustrate the effect of sampling design and variance estimator choice on variance estimates of total escapement for anadromous salmonids from systematic samples of fish passage. Using simulated tower counts of sockeye salmon Oncorhynchus nerka escapement on the Kvichak River, Alaska, five variance estimators for nonreplicated systematic samples were compared to determine the least biased. Using the least biased variance estimator, four confidence interval estimators were compared for expected coverage and mean interval width. Finally, five systematic sampling designs were compared to determine the design giving the smallest average variance estimate for total annual escapement. For nonreplicated systematic samples of fish escapement, all variance estimators were positively biased. Compared to the other estimators, the least biased estimator reduced bias by, on average, from 12% to 98%. All confidence intervals gave effectively identical results. Replicated systematic sampling designs consistently provided the smallest average estimated variance among those compared.

Publication type Article
Publication Subtype Journal Article
Title Efficiently estimating salmon escapement uncertainty using systematically sampled data
Series title American Fisheries Society Symposium
Volume 54
Year Published 2007
Language English
Publisher American Fisheries Society
Publisher location Bethesda, MD
Contributing office(s) Alaska Science Center
Description 9 p.
First page 121
Last page 129
Online Only (Y/N) N
Additional Online Files (Y/N) N
Google Analytic Metrics Metrics page
Additional publication details