Estimating Fine-Scale Transpiration From UAV-Derived Thermal Imagery and Atmospheric Profiles

Accurate and timely observations of individual-scale transpiration are critical for predicting ecosystem responses to climate change. Existing remote sensing methods for measuring transpiration lack the spatial resolution needed to resolve individual plants, and their sources of uncertainty are not well-constrained. We present two novel approaches for independently quantifying fine-scale transpiration using thermal imagery and a suite of environmental sensors mounted on an unmanned aerial vehicle (UAV) platform. The first is a surface energy balance (SEB) approach designed for fine-scale thermal imagery; the second uses profiles of air temperature (T-a) and humidity (h(r)) to calculate transpiration from the Bowen Ratio. Both approaches derive the energy equivalent of transpiration, latent heat flux (lambda E), solely using data acquired from the UAV. We compare the two approaches and their sources of uncertainty using data from several flights at a grassland eddy covariance site in 2021 and 2022 and using typical diurnal conditions to evaluate the uncertainty of lambda E estimates for each approach. The SEB approach generated independent, UAV-based estimates of lambda E within similar to 20% of eddy covariance measurements and was most sensitive to surface temperature and resistance to heat transfer. lambda E calculated from the Bowen Ratio approach was similar to 30% higher than tower values due to inaccuracies in T-a and h(r), the main sources of uncertainty in this approach. The Bowen Ratio approach has a lower overall potential uncertainty, indicating its potential for improvement over the SEB approach. Our results are the first physically-based observations of transpiration derived solely from a UAV platform, with no ancillary data inputs.