Estimating fine-scale transpiration from UAV-based thermal imagery and atmospheric profiles

Accurate estimates of plant transpiration are critical for understanding how ecosystems will respond to climate change, yet there are currently limited ways to observe transpiration across scales. Existing remote sensing approaches lack the spatial resolution to resolve individual plants, while traditional *in situ* measurements are difficult to upscale. Furthermore, the sources of uncertainty in these methods are not well-constrained. We present a novel approach for independently observing fine-scale transpiration using thermal imagery and a suite of environmental sensors mounted on an unmanned aerial vehicle (UAV) platform. The approach uses a combined surface energy balance and atmospheric profiling algorithm to estimate transpiration at sub-canopy scales. Here, we will discuss the sources of uncertainty in the algorithm and highlight an application of the UAV-based approach to study seasonal and diurnal responses of dryland trees to changes in water availability and atmospheric demand at sites in Southern California and Southern Africa. Our results suggest that downregulation of transpiration occurs at lower vapor pressure deficit (VPD) under dry conditions and that plant sensitivity to both water availability and atmospheric moisture demand varies by species and tree size. These findings show how high-resolution observations of transpiration rates across spatial and temporal scales can reveal the mechanisms driving plant and ecosystem responses to hydroclimatic change.