Can Single-Pass Electrofishing Replace Three-Pass Depletion for Population Trend Detection?

Since changes in climate and land use operate at broad spatial scales, efficient monitoring of temporal trends in fisheries resources over large geographic areas is vital to appropriate management. We compared the statistical power of single- versus three-pass electrofishing surveys in detecting temporal trends of age-1 and older Brook Trout Salvelinus fontinalis populations in western North Carolina. Empirical estimates of abundance and capture probabilities were obtained from annual three-pass depletion surveys at 14 headwater stream sites between 2012 and 2017. The CVs in abundance averaged 26% (SD = 14.2%) across study sites, and the mean capture probability per pass was 0.72 (range = 0.57–0.84, SD = 0.09). Captures from single-pass sampling and abundance estimates from three-pass removal sampling were highly correlated (r² = 0.98). In simulations, under the range of years sampled (5–25 years) and annual declines (2.5–7.5%) considered, the power to detect temporal trends was similar (∆ power < 0.1) between the two methods when five or more sites were monitored. An additional set of simulations with varying capture probabilities demonstrated that differences in power between the two methods increased as mean capture probabilities decreased (0.8, 0.5, and 0.2) accompanied by larger variation in capture probabilities among the samples, indicating results obtained with Brook Trout populations in North Carolina might not be applicable to other habitat types or species. Variation in fish abundance did not affect the difference in power between the two methods. Single-pass electrofishing surveys can be an efficient survey method to monitor temporal population trends for habitat types and species characterized with high capture probabilities and low variation among samples. However, single-pass data would not typically allow for inferences of capture probabilities and thus abundance. This can be problematic when environmental factors vary and when data sets collected using different protocols are compared. This trade-off should be carefully considered when designing monitoring programs.