A solution for perfect bioclimate envelopes that are imperfect for extirpated species

Abstract

Present-day species distributions modeled with climate variables cannot provide potential future climate space for species that have contracted in range due to extirpations, regardless of abundant sample sizes within current ranges. My objective was to examine effects of range contractions on modeling of species to determine suitable space under climate change, exploring different approaches to modeling based on historical range maps. As examples of this issue, I estimated the current and future bioclimate envelopes of American bison (Bison bison) and elk (Cervus elaphus) from their current reduced ranges in the western United States compared to historical ranges immediately before extirpation. One solution for bioclimate envelope modeling is to generate presence samples from the historical range and pseudoabsence samples from outside of the historical range. By using the fullest climate space, the models identified areas of future suitable climate space that otherwise would be underpredicted (10%-27% of climate space, for these two examples) based on current ranges of species that have been extirpated from their historical range. Range contraction substantially reduced predictions of suitable climatic space under climate change. Therefore, species need to be evaluated for range extirpation before determining potential impacts of climate change on biodiversity conservation.