Comparison of a Stratified, Instantaneous Count Creel Survey with a Complete Mandatory Creel Census on Escanaba Lake, Wisconsin

Abstract

Creel surveys often play a major role in fisheries management, yet there has been little research on the accuracy of such surveys. To address this shortage of information, we compared fishing effort and harvest rate estimates from a stratified, instantaneous count, access point creel survey (20 h/week) with known values from a complete creel census on Escanaba Lake, Wisconsin. The estimated total fishing effort was only 7% lower than the actual, with a relative standard error (RSE = 100s·SE/estimate) of 11%, and the 95% confidence interval (CI) of the estimate included the actual effort. Harvest rate estimates were made twice, once by using all hours of fishing and once by using only time spent fishing for specific species. The 95% CIs for all of the estimates over the entire period included the actual measurements, and RSEs ranged from 18% to 27%. However, the 95% CIs for some of the estimates from individual strata did not include actual measurements, and RSEs ranged from 20% to 71%. We found that small sample size (few interviews) may lead to large variance and poor confidence interval coverage for harvest rate estimates. We did not observe significant differences (P > 0.05) between the creel survey and the creel census in length distributions of fish harvested or in proportions of marked fish in the harvest. However, estimates of proportions of marked fish in the harvest were relatively imprecise because of small numbers of fish in the creel.