Unreliable population inferences from common trapping practices for freshwater turtles

Abstract

Fundamental questions in ecology and conservation require reliable data about population size and structure. For freshwater turtles, such data are often obtained via mark–recapture trapping, but commonly used trap types are biased in the sex and age classes they sample although these biases are seldom quantified. We present data from 11 populations of Western painted turtles (Chrysemys picta bellii; 1107 turtles total, n caught per pond 6–322) captured in hoop traps, dip-nets, and basking traps to examine bias in captures and the impact on estimates of population size and sex/age ratios. Hatchlings and juveniles were primarily captured in dip-nets, while hoop nets had the lowest capture rates for adults. Most turtles were caught only once; among recaptures, the majority were recaptured in the same trap type. Estimates of population size and sex/age ratios varied strongly when we calculated results from each trap type separately versus combining all captures. These results show clearly that turtle sampling that uses only one trap type will almost certainly mis-estimate population size and sex/age ratios. These results are troubling in the light of current practice: of population studies of North American turtles published during 2009–2014, 45% used only one trap type, and 49% of studies did not even mention possible sampling biases. The conservation implications are serious, as current trapping efforts probably result in erroneous population estimates and sex/age ratios, which may encourage management actions that are not needed or may obscure actions that are in fact necessary for viable populations.