Inferring extirpation

After a gridded search effort density dataframe has been assembled, inference of extirpation risk proceeds in two stages.

In the first stage, we infer an posterior distribution for extirpation probability for each cell in historical and potential habitat.

In the second stage, these per-cell estimates are pooled into regional estimates for extirpation, forming quantified proxies for the IUCN extinction criteria, complete with uncertainty estimates.

At both stages, prior and posterior estimates of extirpation are modelled using Bayesian inference on distributions from the beta family.