In our first Stats Toolbox Seminar we introduced species distribution models (SDMs). Occupancy models are often used for very similar purposes as SDMs, e.g. predicting the distribution of species and determining habitat suitability. One of big differences between SDMs and occupancy models is that the latter require repeat observations. This has the advantage though that one is then able to deal with imperfect detection of a species, i.e. localities that are observed to be unoccupied may in reality be occupied. Another advantage of occupancy models is that one is able to model covariates at both the site level (factors that don't change between repeat visits, e.g. percent grassland cover) and at the survey level (factors that are unique to the particular survey, e.g. weather).

In our second Stats Toolbox Seminar, Res Altwegg provided an extremely useful introduction to occupancy models. Below you can find the lecture slides and R scripts from this seminar:

Presentation slides

R scripts (these are in a zip file. "occupancy.R" uses a simulated dataset to introduce occupancy models. "bald_ibis_web.R" produces an occupancy model for the Bald Ibis. "baldibisdata.csv" is the data needed for the aforementioned code).