Accuracy, Precision, and Optimal Sampling Duration of Low‐FrequencyElectrofishing for Sampling Reservoir Flathead Catfish Populations

Flathead Catfish Pylodictis olivarisare popular among anglers; however, information about their sampling is limited. Low‐frequency electrofishing (LFE) is the most used method for sampling Flathead Catfish, but LFE data quality (precision and accuracy) has not been previously studied. Therefore, we evaluated accuracy, precision, and optimal sampling duration for maximizing precision of LFE sampling for Flathead Catfish. To quantify accuracy, we created known populations by tagging Flathead Catfish in Lake Carl Blackwell, Lake McMurtry, and Boomer Lake, Oklahoma, with numbered modified Carlin dangler tags and calculated their capture probabilities from recapture data with a Cormack–Jolly–Seber model, with water temperature as an environmental covariate and fish size as an individual covariate. Capture probability was negatively correlated with increases in fish length for Lake Carl Blackwell and Lake McMurtry but was positively correlated with increases in fish length for Boomer Lake. Capture probability was highest at warmer temperatures at Lake Carl Blackwell and Lake McMurtry but was highest at lower water temperatures at Boomer Lake. Therefore, catch rate and size bias varied by system, but size bias was still relatively consistent at all temperatures within lakes (i.e., lake‐specific differences in slopes were subtle even though significantly different), indicating that LFE could be used to detect relative changes in size structure if temperatures were standardized. Catch rates were highest and most consistent from June to September when water temperatures were ≥24°C. The number of 5‐min LFE efforts needed to achieve a relative standard error ≤25% was lowest when water temperature was ≥20°C from months between May and September. Catch rates and size structure did not differ between LFE efforts (5, 10, or 15 min), suggesting that any LFE effort would produce similar relative abundance estimates.