Limits to species distributions on tropical mountains shift from high temperature to competition as elevation increases

Species turnover with elevation is a widespread phenomenon and provides valuable information on why and how ecological communities might reorganize as the climate warms. It is commonly assumed that species interactions are more likely to set warm range limits, while physiological tolerances determine cold range limits. However, most studies are from temperate systems and rely on correlations between thermal physiological traits and range limits; little is known about how physiological traits and biotic interactions change simultaneously along continuous thermal gradients. We used a combination of correlational and experimental approaches to investigate communities of Drosophila flies in rainforests of the Australian Wet Tropics, where there is substantial species turnover with elevation. Our experiments quantified individual-level and population-level responses to temperature, as well as the impact of interspecific competition under different temperature regimes. Species' distributions were better explained by their performance at extreme temperatures than by their thermal optima. Upper thermal limits varied less among species than lower thermal limits. Nonetheless, these small differences were associated with differences in the centered elevation of distribution. Low-elevation species were not those with the lowest tolerance to cold, suggesting that cold temperatures were not limiting their abundance at high elevations. Instead, under upland temperature regimes, abundances of these low-elevation species were reduced by competition with a high-elevation species, in both short- and long-term competition experiments. Our results demonstrate that high-elevation species are confined to their current ranges by high temperatures at lower elevations, indicating that their ranges will be highly sensitive to future warming. Counter to expectation, species interactions strongly influenced community composition at cooler, high-elevation sites. Together, these results raise the possibility that tropical communities differ from better-studied temperate communities in terms of the relative importance of biotic interactions and abiotic factors in shaping community composition and how the impact of these factors will change as temperatures increase.