Estimating Reproductive Costs with Multi-State Mark-Recapture Models, Multiple Observable States, and Temporary Emigration

摘要: Multi-state mark-recapture models have seen increased use in recent years for studies of reproductive costs. When individuals both breeding and non-breeding states can be observed, multi-state used to directly estimate costs by comparing state-specific estimates survival probabilities. The method assumes that each state an animal occupies is observable, assumption violated if some animals are absent one or more seasons thus, unobservable due temporary emigration. Previous research on the case a single observable has shown non-random (Markovian) emigration can, not accounted for, bias survival. Here, simulation study effects probabilities two state. Results clearly show cause probability biased ignored. Bias was either positive negative depending circumstances, sometimes severe (percent relative as high 67% probability). Accordingly, strengths limitations including analyses also considered. For situations, simply will adequate solution. But, those particularly interested temporal variation reproduction, it necessary collect other information avoid problems parameter constraints. Additional consist data from sub-sampling during primary sampling occasions, radio telemetry, ring recoveries.