Testing hypotheses on distribution shifts and changes in phenology of imperfectly detectable species

Abstract/Summary

With ongoing climate change, many species are expected to shift their spatial and temporal distributions. To document changes in species distribution and phenology, detection/non-detection data have proven very useful. Occupancy models provide a robust way to analyze such data, but inference is usually focused on species spatial distribution, not phenology.
We present a multi-season extension of the staggered-entry occupancy model of Kendall et al. (2013), which permits inference about the within-season patterns of species arrival and departure at sampling sites. The new model presented here allows investigation of species phenology and spatial distribution across years, as well as site extinction/colonization dynamics.
We illustrate the model with two datasets on European migratory passerines and one dataset on North American tree frogs. We show how to derive several additional phenological parameters, such as annual mean arrival and departure dates, from estimated arrival and departure probabilities.
Given the extent of detection/non-detection data that are available, we believe that this modeling approach will prove very useful to further understand and predict species responses to climate change.