This one day workshop will provide participants with an overview of some recently-developed methods for modelling the spatial and spatio-temporal distribution of wildlife populations using data arising from the most widely-used wildlife survey methods. More specifically, it will cover methods of fitting Poisson point process models and log-Gaussian Cox process models to survey data from plot surveys, distance sampling surveys and capture-recapture surveys.

​

A feature of wildlife surveys that distinguishes them from most other kinds of surveys is that inclusion probabilities (the probabilities that individuals in the study population are included in the sample) are beyond the surveyors’ control, unknown and spatially varying. This creates difficulties for spatial modelling because spatially varying inclusion probability and spatially varying density may be confounded. Wildlife survey methods have tended to focus on estimating inclusion probabilities under simplistic assumptions about spatial distribution. Spatial modelling methods, on the other hand, have focussed on realistically complex models for spatial distribution while assuming known inclusion probabilities. One would really like methods that combines the strengths of these two approaches.

​

This workshop gives an overview of some recently-developed methods that integrate the two approaches, allowing inferences about spatially varying inclusion probabilities to be drawn simultaneously with inferences about the spatial distribution and abundance of wildlife populations.

​

The workshop will include some hands-on sessions to give participants (a) a taster of R software packages implementing some of the methods and (b) a break from being lectured at. Participants should bring their own laptop computers for these sessions, with R and (preferably) RStudio installed. Details of required R libraries will be provided in advance of the workshop. Please note that there is insufficient time in the workshop to provide anything more than a taster of the packages.