Understanding movement from mark-recapture data

Background/Question/Methods

Mark-recapture data often provides a wealth of information on movement patterns of individuals, and the successful interpretation of mark-recapture data relies on understanding movement patterns to separate apparent survival into survival and emigration.  This issue is typically resolved by monitoring emigration or appealing to literature on the limits of movement for the study species.  We asked how this issue could be resolved relying primarily on mark-recapture data itself.  We implemented a state-space movement, survival, and recapture model for our study species (Atlantic salmon) and tested the model using simulated data covering a wide range of biologically realistic parameters as well as field data.

Results/Conclusions

Our model model can be successfully implemented in the BUGS language and provides unbiased estimates of survival, recapture, and the dispersal kernel under a wide variety of conditions where a simple CJS model produces negatively biased survival and recapture estimates.  For the dispersal kernel the model is able to estimate both scale and tail weigh parameters despite the usual problems arising from censored observation of long-distance dispersal.  Our simulation results describe the combination of parameters where the model begins to show bias and results from field data are comparable to those obtained using a time-intensive method of monitoring emigration.  We believe novel analyses can significantly the improve the utility of mark-recapture data for understanding movement.  Incorporating that understanding into state-space movement models will further extend the parameter where this type of model is reliable.