Forests face accelerating threats due to increases in the severity and frequency of drought and heat stress associated with climate change, which manifest as changes in their extent and composition. Regeneration in forests is increasingly garnering the attention of research and management efforts given that this vital rate is considered uniquely sensitive to climate trends and weather extremes. As rates of disturbance (e.g, wildfire) increase, and the climatological and hydrological suitability of landscapes for forest regeneration are diminished, vegetation shifts from forest to non-forest are expected to become more frequent and widespread. Most efforts at predicting these shifts have used correlative approaches which lack the ability to generalize to novel temporal or spatial conditions. To reduce uncertainty when predicting forest regeneration, researchers need mechanistic models and predictors of seedling mortality based on physiological processes and hydraulic function.
Here, I introduce a novel environmental metric: potential land surface temperature (pLST), which is an estimate of land surface temperature in the hypothetical absence of overstory vegetation. Land surface temperature (LST) is a radiometric measure of the energy balance at the Earth’s surface: it is governed by net radiation and soil moisture, has strong effects on seedling physiology, and can be remotely sensed. Therefore, it is a key variable in many studies of how vegetation, hydrology, and climate interact. To develop and validate pLST as a potential metric of forest regeneration suitability, I first produce pLST estimates using a coupled ecohydrologic and plant hydraulic model in previously burned watersheds distributed across the western U.S. I then validate these estimates using remotely sensed measurements of LST and forest cover.
Results/Conclusions
Remotely sensed measurements of post-burn LST agreed with modelled pre-burn estimates of pLST, demonstrating the skill of this ecohydrologic modelling approach in predicting surface temperatures after disturbance. In addition, the extent of currently forested areas was restricted to pixels below pLST thresholds expected to result in seedling mortality, further validating modelled pLST estimates. This work leverages advances in ecohydrologic modelling and remote sensing to produce information about forest regeneration suitability at scales relevant to forest managers.