Predicting spatiotemporal persistence of rare species: An example with North Atlantic right whales

Abstract Knowledge of when species remain in specified areas is essential for survey design, conservation, and management. Using species occurrence data to predict persistence in space and time (i.e., presence of one or more individuals of the species of interest within a defined spatial area over a duration of a specified number of days) may be possible with extensive survey effort and complex modeling, but such requirements pose challenges. Here, we present a method for estimating wildlife spatiotemporal persistence by (1) filtering data to contain detections and all surveys occurring 7 days following each detection within a spatial buffer around each detection (i.e., spatiotemporal detection buffer), (2) identifying redetections in each spatiotemporal detection buffer, and (3) grouping detection buffers temporally and spatially for bootstrap resampling. Our method avoids the need for mechanistic models of animal behavior while accommodating survey effort that may, at times, be sparse. We illustrate the approach using spatiotemporal data from 2010 to 2020 vessel‐based and aerial surveys of the North Atlantic right whale ( Eubalaena glacialis ), an endangered species that experiences various anthropogenic threats that are the focus of significant management actions. Our analyses suggested that persistence probabilities of North Atlantic right whales varied across time and space, which could guide management measures associated with forecasting or nowcasting spatiotemporal persistence for new detections. Our method can be applied to any species with repeated survey data and could facilitate dynamic management practices that effectively target conservation efforts. The presented method is especially helpful for rapid decision‐making.