Occupancy modelling is increasingly being favoured in monitoring studies, particularly where the effort and costs associated with estimating the abundance of species across a large landscape over long time periods is prohibitively expensive. In addition to being relatively easier to execute and inexpensive, the detection-nondetection data generated from occupancy surveys can be analysed using open source software which adds to its appeal. Occupancy models not only estimate “proportion of sites occupied”/occupancy as the population parameter of interest, but can address conservation needs by identifying key ecological and management correlates of occupancy, while explicitly accounting for imperfect and heterogeneous detection. However, inadequate understanding of issues of spatial scale, autocorrelation, model selection and model uncertainty often leads to poor implementation of this approach. In this workshop, a combination of lectures, field and computer-based practical exercises and discussions will be used to provide a theoretical and analytical introduction to occupancy modelling techniques.
The participants attending this workshop will learn about:
Estimating population occupancy, Role of detection probability, Basic statistical background: probability; maximum likelihood estimation, logistic regression, covariate modelling and odds ratios, hypothesis testing, model comparison and multi-model inference, Single-season models, multi-season models , species co-occurrence models, multi-scale, multi-method models, spatially-dependent models, other extensions ( e.g. Incorporation of count data and estimates of abundance, marked animals), Design matrices and fitting occupancy models in PRESENCE using sample dataset
Information to Participants:
Participants must have a working installation of:
We urge the participants to test their software before coming for the workshop.
Capacity: 20 Students