PIMS PDF Colloquium: Counting Sheep -- Bayesian Methods to Account for Time-Dependent Covariates in Open-Population Capture

  • Date: 12/10/2009
  • Time: 14:00
Simon Bonner (PIMS/UBC)

University of British Columbia


Capture-recapture methods are widely used to monitor endangered wildlife populations. A requirement of simple capture-recapture models is that all individuals alive on one sampling occasion have the sameprobability of capture. While this assumption may be reasonable insmall, isolated populations, there are many variables that mightinfluence an individual's catchability and estimates of survival ratesor the abundance will be biased if these differences areignored. However, covariates of the capture probability which varyboth between individuals and over time, like body mass, present achallenge in the analysis of capture-recapture data because 1) theirvalues can only be measured for the individuals captured on eachsampling occasion and 2) the unknown values are not missing at randomand cannot be ignored. I will present Bayesian methods to incorporatethe effects of such covariates in the Cormack-Jolly-Seber andJolly-Seber models -- the two most common models for open-populationcapture-recapture data.



My talk will begin with an introduction to Bayesian statistics,capture-recapture methods, and the problems associated withtime-dependent covariates. I will then describe my method forincluding such covariates in the Cormack-Jolly-Seber model to estimatesurvival rates and how this method can be extended to the Jolly-Sebermodel to obtain estimates of abundance. I will illustrate my methodsby application to data from the study of Soay sheep on the Isle ofHirta, Scotland, and conclude by discussing applications to morecomplicated models and comparisons with other approaches.


WMAX 216

Other Information: 

Tea & cookies afterwards!