The Robust Design in Capture-Recapture Studies: A Review and Evaluation by Monte Carlo Simulation

Abstract

The use of capture-recapture methods in the estimation of population size, survival rates, and recruitment rates of wild animals is well established. The traditional approach has been to use the closed population models summarized by Otis et al. (1978) for short-term studies, and the Jolly-Seber open population model (Jolly 1965, Seber 1965), or some variation of it, for long-term studies. Pollock (1982) proposed a “robust” design that would allow the use of both closed and open population models in an ad hoc fashion for long-term studies. Potential advantages of this ad hoc method, over the use of the Jolly-Seber model alone, include: (1) abundance and other quantities can be estimated without assuming equal catchability, (2) all parameters can be estimated, (3) alternative estimators are available for most parameters, (4) recruitment can be separated into immigration and ’births’ with only two age-classes, (5) abundance and survival can be estimated with minimal sampling correlation, and (6) the design can be customized more finely to meet specific needs. This paper reviews the design and various methods of parameter estimation it allows for, and briefly compares the performance of available estimators using Monte Carlo simulation.