In our upcoming Bayesian Working group meeting on Friday (09/27), 2:10 - 3 pm in Snedecor 2113, we will have Dr. David Delaney from Natural Resource Ecology and Management department presenting. The details are as follows.

Title: An ecologist's move to Bayesian statistics for more flexible modeling of wildlife density

Summary: Distance sampling is a method of estimating the population size of unmarked animals when detection probability is imperfect. Packages and software to analyze such data have been developed for study sites that are rectangular or circular in shape. We assessed the shape of 99 viewsheds (visible area from observation roadways) in Iowa that are visited annually to count deer. Every viewshed decreased in area with distance from the roadway, at varying magnitudes, because of barriers to visibility such as trees and topography. We moved to a Bayesian analytic framework using JAGS in R because we were able to account for non-traditional viewshed shapes in the JAGS/BUGS coding structure. The flexibility of JAGS code structure has at least two additional areas of improvement over frequentist methods. We'll explore the data structure, current code, and attempt to code in these additional improvements.