COS 154-6 - Centimetre-scale mapping of animal operative temperatures in a rapidly degrading tropical landscape

Background/Question/Methods

Quantifying the thermal environments organisms experience is central to predicting their exposure to future environmental change and to understanding processes operating across ecological scales from individual behaviour to population and community dynamics. Traditional methods from thermal ecology can provide accurate estimates of animal operative temperature at precise locations, but sample only a small portion of the available environment. Alternatively, combined biophysical and microclimate models can provide continuous estimates across larger spatial extents but still, despite recent advances, at spatial resolutions that are far coarser than those experienced by individual organisms. Here, we integrate field data with remotely sensed measures of canopy structure from unoccupied aerial vehicles and satellites to map operative temperatures of the tropical lizard Anolis bicaorum at resolutions ranging from 1cm to 50cm, across spatial extents from 0.05 ha to several km². We use these models to 1) test how the extent and configuration of thermally suitable habitat influences the ecology of individuals and populations; and 2) evaluate changes in thermal habitat given rapid forest alteration across A. bicaorum’s entire range.

Results/Conclusions

We find that random forest models incorporating standard air temperature and remotely sensed canopy structure from unoccupied aerial vehicles can explain up to 82% of the variation in operative temperature across 0.05 ha plots, with an RMSE of 1.03°C. Combining cm-scale maps of operative temperature with A. bicaorum’s thermal niche limits suggested that population abundance may be weakly constrained by the area of thermally suitable habitat within plots (P=0.08), but not the number or configuration of thermal refugia (P >0.25 in all cases). Using satellite (Worldview-2) measures of canopy variation at 50cm resolution, we found that from 2018 to 2020, ca. 3% of thermally suitable habitat was lost from across A. bicaorum’s range, but much of the deforestation was restricted to already thermally unsuitable areas. Our work reveals the potential of linking thermal ecology and remote sensing to provide unprecedented information on changing thermal environments at organism-relevant scales across complete landscapes.