Occupancy sampling, modelling and estimation: what, why, and how?
Occupancy estimation has gained tremendous popularity over the last decade, largely due to the relative ease and low costs of carrying out occupancy surveys combined with availability of easy-to-use, freely downloadable software for analysing occupancy data. The approach allows both estimation of a key (meta-)population parameter and, more importantly, assessment of species-habitat relationships while accounting for confounding observation processes such as imperfect and heterogeneous detectability, among others. Perhaps inevitably, the approach is sometimes poorly or wrongly applied, such as when there is a mismatch between the question and the spatial scale, or between the ecological/sampling situation and the model(s) used for inference.
In this workshop, we will first examine occupancy estimation in the context of population parameter estimation and statistical inference, spending some time on trying to intuitively understand maximum likelihood estimation. Occupancy models for a few different ecological/sampling situations will be covered in some detail, including single and multi-season models, covariate modelling, abundance-induced heterogeneity, multi-scale multi-method occupancy and spatially dependent presence. Having understood how occupancy models work, we will look at the relationship of the framework to other approaches such as presence-only modelling and logistic regression and the assumptions required by each. The design of occupancy surveys and common pitfalls to avoid in designing, implementing and analysing occupancy surveys will be covered in some detail. Finally, there will be practical sessions to carry out analyses of sample datasets using spreadsheet (e.g. MS-Excel) and freeware software programs (PRESENCE).
No prior skills, training or experience are assumed or required for the workshop. Participants are strongly urged to bring laptops (fully charged!) with MS-Excel (or an equivalent spreadsheet application) and program PRESENCE (available at http://www.mbr-pwrc.usgs.gov/software/presence.html) installed in advance.
Devcharan Jathanna, Senior Research Associate, Centre for Wildlife Studies