Over the last 10-15 years, site-occupancy models have become a popular method for data analysis in ecological studies. Specifically, I will consider ‘site-occupancy’ or ‘occupancy’ models to be hierarchical models that aim to separate the observation process (i.e., detection) from the true underlying state process (i.e., occupancy). Occupancy modeling has been used in a vast array of applications including habitat relationships, species distribution modeling, metapopulation dynamics, multi-species and community dynamics, disease dynamics, etc. These applications range across plants, invertebrates, amphibians, mammals, etc.
Occupancy modeling has gained in popularity likely due in part to the relative ease of data collection, the availability of software for implementation, and the continued development to relax model assumptions and address new questions. With over 1000 publications citing occupancy models in the last decade, now seems like good time to review the use of occupancy modeling in ecology and give consideration to shortcomings of the methodology. This begs the question: have occupancy models reached their limit in utility for ecological applications? Are we learning new and useful information from these applications?
Results/Conclusions
Many argue that there is an inherent gain in separating out the sampling bias from the underlying state process. But other research suggests that the gain is minimal relative to the required extra sampling or that results from occupancy models may be inaccurate for a suite of reasons. In this talk, I will examine the role of occupancy models in a set of ecological examples and discuss the implications of these approaches and speculate on future potential of occupancy modeling in ecology.