Drivers of the spatiotemporal variability in the thermal balance of forests during heatwaves and normal conditions.

Different land covers present contrasting changes in energy budgets as a response to heatwaves and droughts and thus the land feedback is expected to vary over the landscape. To date, the study of the biotic determinants of land-atmosphere feedbacks during heatwaves has been restricted to the consideration of different plant functional types. We used improved vegetation structural measurements at organizational levels lower than plant functional types (inter– and intra–specific) to estimate the impact of forests on the surface thermal balance. We combined space-borne measurements of the temperature of plants (ECOSTRESS) and the land surface (MODIS) with ground-based meteorological data to estimate the thermal balance of the surface (∆T) at a resolution of 70x70m in 615 forest plots, dominated by 28 different species. In each plot, forest structural variables were determined through LiDAR. We then analysed the spatiotemporal drivers of ∆T by quantifying the contribution of topographical, landscape, meteorological and forest structural variables on ∆T both during normal conditions and heatwave episodes. Canopy temperatures fluctuated according to changes in air temperature and were on average 1˚C warmer than the air. During heatwaves, canopies were relatively cooler than the air, compared to normal conditions in all but Mediterranean coniferous forests. The thermal response of canopies to heatwaves strongly varied as a function of environmental variables. Forests in rainy areas and in steep slopes presented the lowest ∆T, whereas forests in arid areas and flat terrain had the highest ∆T. Interestingly, there was a strong effect of forest structure, since forests with larger biomass kept a cooler thermal balance (lower ∆T). Indeed, the total effect of forest structural variables on ∆T was of equal magnitude as that of topography or meteorological conditions. The thermal balance of the surface (∆T) was not only different among the main forest types, but also, it strongly varied within forests dominated by the same species. Because ∆T is an important component of the surface energy budget, our results on its dependence on forest structure imply that forest management could be employed to modify the surface energy budget to promote negative (mitigating) feedbacks of forests during heatwave episodes. Further efforts concentrate on estimating changes in aerodynamic conductance between forests and their surroundings, and their potential influence on the land–atmosphere coupling and the feedback of forests on local temperatures.