Robustness of Capture-Recapture Estimators to Sample Biases in a Cyclic Snowshoe Hare Population

Abstract

A Monte Carlo simulation model was used to determine estimator robustness to trap saturation and other sample biases typical of a cyclic snowshoe hare population. Field studies showed that hare capture probability varied directly with the amount of nightly movement made by an individual. This relationship was simulated by using an individual-based model of individual hare movement and the trapping process. All estimation models in program CAPTURE except the jackknife heterogeneity (M h ) estimator showed increasing negative bias with increasing hare abundance in computer simulations. The jackknife estimator was robust to biases caused by trap saturation, and showed an acceptable coefficient of variation. The program CAPTURE model selection routine performed poorly by selecting estimation models of different bias for each simulated hare abundance. We conclude that for closed populations the jackknife estimator (M h ) is the most robust to sampling variation in a cyclic snowshoe hare population. These results suggest that it is optimal to use one model rather than change models for populations which exhibit large changes in abundance.