Title: Estimating the Probability that a Pheasant will be Seen: Fitting N-Mixture Models in Stan and JAGS

Abstract: Management of hunted species, including pheasants, is informed by surveys that estimate abundance.  Pheasants are typically surveyed by driving down a country road and counting how many are seen.  One difficulty is that pheasants may be present but not seen.  The detection probability is anticipated to depend on weather conditions and other survey-specific characteristics.  N-mixture models provide a way to simultaneously estimate abundance and detection probability.  The current preferred estimators are Bayesian, in order to account for all sources of variability.  I helped Adam Janke and collaborators analyze survey data from multiple states.  The informal talk will tell you about the bird and data, N-mixture models, why they can be difficult to fit, and our success (or lack thereof) using two different R packages for fitting Bayesian models: Stan and JAGS.