Observing Montane Forest Evapotranspiration and Water Limitations With Thermal-Based Models In Complex Terrain

Abstract:

Predicting forest ecosystem water use and stress are critical to managing future water resources. Recently, new modeling tools that incorporate remote sensing-derived land surface temperature (LST) have been developed to estimate ecosystem water use and stress. Yet, the ability of these new LST-based tools to detect fine-scale hydrological processes has not been validated. We use an experimental design that takes advantage of differences in water availability across topography to validate a variety of LST-based products at Sagehen Creek in the Sierra Nevada, USA. First, we use ground- and unmanned aircraft system (UAS)-based LST observations to estimate water use and limitation using a two source energy balance model (TSEB) across a hydrologic gradient. Our results suggest that early estimates of ET derived from NASA’s ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) are able to effectively capture different timing caused by hillslope-scale differences in water availability. These inferences are confirmed by UAS-based ET estimates using a TSEB . We also compared our fine-scale ET estimates derived from TSEB model with ECOSTRESS ET products. Our early analyses suggest that ET from ECOSTRESS are able to effectively capture water stress caused by differences in water availability at hillslope-scale. These inferences are confirmed by UAS-based ET estimates using our TSEB. Second, we use eddy covariance estimates of latent heat flux above and below canopy to validate daily and seasonal ET estimates derived from ECOSTRESS and Landsat LST estimates. At the scale of 30-70 m, validation depends on local heterogeneity in ET caused by variations in vegetation and subsurface water availability.