An Evaluation of Mark–Recapture Estimates of Black Crappie Population Size Determined by Trap-Netting during Fall and Spring in Minnesota Lakes

Abstract

We evaluated mark–recapture population estimates of black crappies Pomoxis nigromaculatus 200 mm and longer made with trap-net catches during fall or spring in eight Minnesota lakes (18–168 ha). We estimated population size N during fall by use of marking and recapture in fall (FF estimates) and marking in fall and recapture during the following spring (FS estimates). We estimated spring N by use of marking and recapture after ice-out in early spring (ES estimates) and marking after ice-out and recapture during late spring (LS estimates). The FF and ES estimates were multiple-census estimates, whereas the FS and LS estimates were two-census estimates even though sampling occurred over many days. The most precise estimates were the LS estimates (median 95% confidence limits (CLs) were within 36% of 1/N), and FF estimates were least precise (median 95% CLs were within 51% of 1/N) among sampling scenarios. The median trap-net effort needed to achieve 95% CLs within 25% of 1/N estimates ranged from 160 (ES estimates) to 490 trap-net sets (FS estimates), and the effort needed for this level of precision decreased with increasing density of marked fish at large. The ES and LS estimates for the same lakes and years differed significantly (P < 0.05), and FF estimates did not exceed LS estimates from the following spring, which indicates that either the estimates from one or more scenarios were biased or that the catchable portions of populations differed temporally. Probable sources of bias included unequal capture probabilities, unequal mortality, and unequal emigration between marked and unmarked individuals during or after marking, as well as immigration of unmarked individuals after marking. Although marking occurred over short time periods, mortality, emigration, and immigration were possible, and therefore populations may not have been closed. Comparisons of N estimated from the same sampling scenario are the most meaningful because the similarity in biases or catchable portions of populations should be highest.