全球变化知识资源中心

Aim We analysed elevational and microclimatic drivers of thermal tolerance diversity in a tropical mountain frog clade to test three macrophysiological predictions: less spatial variation in upper than lower thermal limits (Bretts' heat-invariant hypothesis); narrower thermal tolerance ranges in habitats with less variation in temperature (Janzen's climatic variability hypothesis); and higher level of heat impacts at lower elevations. Location Forest and open habitats through a 4,230-m elevational gradient across the tropical Andes of Ecuador. Method We examined variability in critical thermal limits (CTmax and CTmin) and thermal breadth (TB; CTmax-CTmin) in 21 species of Pristimantis frogs. Additionally, we monitored maximum and minimum temperatures at the local scale (tmax, tmin) and estimated vulnerability to acute thermal stress from heat (CTmax-tmax) and cold (tmin-CTmin), by partitioning thermal diversity into elevational and microclimatic variation. Results Our results were consistent with Brett's hypothesis: elevation promotes more variation in CTmin and tmin than in CTmax and tmax. Frogs inhabiting thermally variable open habitats have higher CTmax and tmax and greater TBs than species restricted to forest habitats, which show less climatic overlap across the elevational gradient (Janzen's hypothesis). Vulnerability to heat stress was higher in open than forest habitats and did not vary with elevation. Main conclusions We suggest a mechanistic explanation of thermal tolerance diversity in elevational gradients by including microclimatic thermal variation. We propose that the unfeasibility to buffer minimum temperatures locally may explain the rapid increase in cold tolerance (lower CTmin) with elevation. In contrast, the relative invariability in heat tolerance (CTmax) with elevation may revolve around the organisms' habitat selection of open- and canopy-buffered habitats. Secondly, on the basis of microclimatic estimates, lowland and upland species may be equally vulnerable to temperature increase, which is contrary to the pattern inferred from regional interpolated climate estimators.