The Robustness of Mark-Recapture Methods: A Case Study for the Northern Spotted Owl

A simulation study was carried out to investigate various aspects of modern methods for estimating survival probabilities from data on the recaptures of marked birds. The emphasis was on a study of the northern spotted owl in the Pacific Northwest of the United States where many captures and recaptures are of male and female nesting pairs, but the methods and many results apply generally to the study of nesting birds and denning mammals. The primary objectives of the study were to examine the bias in estimates, if any, due to temporary emigration, different capture probabilities for nesting and nonnesting birds, and nonindependent captures of males and females. Secondary objectives were to examine the properties of model selection using the quasi-Akaike information criterion (QAIC) and the power of goodness-of-fit tests for detecting temporary emigration, heterogenous capture probabilities, and dependence in capture probabilities. It was found that (i) the effects on survival estimates of the deviations from the usual assumptions for mark-recapture models are generally quite minor; (ii) the QAIC usually selects a model that is close to the correct one, although the difference may be important in terms of the conclusions from an analysis; and (iii) for the specific sample sizes considered, tests for goodness of fit have little power to detect errors in assumptions.