Recent efforts to predict climate change impacts on biodiversity, and ectotherms in particular, across large geographic scales have emphasized the importance of physiology and climatic gradients. Results from such studies have noted the possible high vulnerability of tropical species to future warming, a conclusion highlighted in the most recent IPCC report. However, habitat and environmental variation within the tropics is also believed to impact warming vulnerability of species, though how such gradients might contribute to thermal tolerance remains understudied and not well understood. Here we present two case studies of a skink species (Trachylepis affinis) and a butterfly species (Bicyclus dorothea) in Cameroon and compare the thermal performance and tolerance of rainforest populations to ecotone (forest/savanna transition) populations. We reared individuals of B. dorothea from rainforest and ecotone populations in the lab for two generations under common conditions and measured thermal tolerance (CTmax and CTmin). We also measured sprint speed of T. affinis over a range of temperatures and measured thermal tolerance of the skinks from rainforest and ecotone populations. Finally, we used this data to parameterize a mechanistic model of climate change impacts using NicheMapR to assess landscape scale consequences of thermal performance variation for warming vulnerability.
For the butterfly species, we found that 2nd generation individuals originating from ecotone populations had a higher CTmax, by nearly 3°C, relative to individuals originating from the rainforest populations. For the skink species, we found no difference in tolerance limits between populations but did find preference for lower temperatures in rainforest skinks; sprint speed thermal optimum in ecotone skinks was 3°C higher than rainforest skinks. Through the application of the biophysical and mechanistic model based on these results, we also found that rainforest skinks are projected to incur greater losses in activity time under warming conditions (and therefore have higher vulnerability to climate change) relative to ecotone skinks. For both species, physiological performance varies significantly across the forest-savanna gradient. Based on thermal performance alone, ectotherm rainforest populations will likely be more vulnerable to future warming impacts. Additional information on predicted land use changes and whether there is genetic variation in physiological traits sufficient to permit an evolutionary response to climate change, will be necessary for informed conservation of these biodiverse Central African ecosystems under assault from global change.