Abstract
The estimation of extinction dates from limited and incomplete sighting records is a key challenge in conservation (when experts are uncertain whether a species has gone extinct) and historical ecology (when the date and mechanism of extinction is controversial).
We introduce a spatially-explicit method of interpolating extinction date estimators, allowing users to estimate spatiotemporal surfaces of population persistence from georeferenced sighting data of variable quality.
We present the R package spatExtinct, which produces spatially-explicit extinction date surfaces from geolocated sightings, including options for custom randomization schemes to improve accuracy with limited datasets. We use simulations to illustrate the sensitivity of the method to parameterization, and apply the method to identify potential areas where Bachman’s warbler (Vermivora bachmanii) might be rediscovered.
Our method, and the spatExtinct package, has the potential to help describe and differentiate different drivers of extinction for historical datasets, and could be used to identify possible regions of population persistence for species with an uncertain extinction status, improving on non-spatial or imprecise methods that are currently in use.
