The relative influence of abiotic and biotic factors on suitable habitat of Old World fruit bats under current and future climate scenarios

Abstract

There is growing evidence that biotic factors such as predator-prey interactions play significant roles in driving species distribution across large spatial scales. The relative influence of abiotic and biotic factors on species distribution, however, may change under climate change. We investigated the relative influence of abiotic and biotic variables on the potential current and future distributions of three fruit bat species, Epomophorus angolensis (Gray, 1870), E. wahlbergi (Sundevall, 1846) and Rousettus aegyptiacus (E. Geoffroy St.-Hilaire, 1810), in southern Africa. We tested three hypotheses, namely that bat species’ distribution is primarily driven by (1) productivity; (2) physiological tolerance to climate; and (3) biotic interactions, specifically fig distribution. We adopted an ensemble niche modelling approach to project the suitable habitat of fruit bat species for current and future climate scenarios, and assessed variable importance in the models using a randomised variable shuffle procedure. We predicted that both biotic and abiotic factors influence suitable habitat of fruit bats, the relative influence of factors on habitat suitability of bat species are taxon specific, and the relative influence of abiotic and biotic factors will change from current to future climate scenarios. Abiotic variables associated with productivity were the primary determinants of habitat suitability for E. wahlbergi and E. angolensis under both current and future conditions. By contrast, suitable habitat of R. aegyptiacus was primarily mediated by temperature under current climatic conditions yet by freestanding fig distribution under both moderate and extreme future climate change scenarios. Freestanding fig distribution was also the most significant factor of habitat suitability for E. angolensis under the extreme future climate change scenario. Our results were congruent with our predictions and suggest that biotic variables play important roles in determining habitat suitability of species at relatively large spatial scales, contrary to the conventional assumptions of the Grinnellian niche.