Hyperstability in Electrofishing Catch Rates of Common Carp and Bigmouth Buffalo

Electrofishing catch per unit effort (CPUE) can be an effective index of relative abundance when CPUE is proportional to population density. However, the relationship between fish catch rates and density can be variable and not proportional because catchability may not be constant under a range of environmental conditions. Hyperstability occurs when catch rates change faster than population density, whereas hyperdepletion occurs when catch rates change slower than population density. We used capture-mark-recapture methods to estimate density of Common Carp Cyprinus carpio and Bigmouth Buffalo Ictiobus cyprinellus across 24 lake-years. Next, we developed log-linear mixed-effects models to assess the relationship between catch rates and density to predict electrofishing CPUE based on density and environmental covariates. Common Carp and Bigmouth Buffalo electrofishing CPUE exhibited hyperstability; CPUE initially increased faster at low density but reached an asymptote as density continued to increase. Common Carp catch rates increased with water temperature and were negatively associated with season; highest catch rates occurred in the warmest temperatures of spring, whereas the lowest catch rates occurred in the coldest temperatures of fall. Bigmouth Buffalo electrofishing CPUE was negatively related to water temperature but had a quadratic relationship with season; catch rates were highest in the spring and fall when water temperatures were coldest. Secchi depth, shoreline length, and maximum depth did not influence our model predictions in either species. Electrofishing had little power to detect simulated changes in Bigmouth Buffalo density regardless of sampling effort, whereas power to detect changes in Common Carp density increase with population changes and sampling effort. At low population densities for both species, electrofishing CPUE is expected to increase or decrease faster than density and can obscure conclusions about population trajectories but may still be useful for monitoring population fluctuations if hyperstability is accounted for by a correction factor.