Evaluating a Remote Sensing Approach to Estimate Root-Zone Soil Moisture Accounting For Regional Characteristics

Mapping root-zone soil moisture over large regions is valuable for improving crop productivity, managing water resources, predicting flood and drought scenarios, and many other hydrologic applications. A remote-sensing method based on optical and thermal satellite imagery has been proposed to estimate root-zone soil moisture at fine resolutions (e.g., 30 m x 30 m) over large regions (e.g., 170 km x 185 km). In this method, multispectral imagery from the Landsat satellite is processed using a surface energy balance model (ReSET) to estimate the evaporative fraction (Λ). Λ is the ratio of the latent heat flux (main driver of evapotranspiration) to the sum of the latent and sensible heat fluxes. The root-zone volumetric water content (θ) is then most commonly estimated using a single empirical relationship with Λ. However, multiple relationships have been proposed recently that aim to account for region characteristics. The objective of this study is to evaluate the optical/thermal estimates for soil moisture when the region-specific Λ-θ relationships are used. The study sites include the Walnut Gulch Experimental Watershed in Arizona, the Piñon Canyon Maneuver Site in Colorado, the Little Washita/Fort Cobb Experimental Watersheds in Oklahoma, and the Mississippi Delta region in Mississippi. These regions correspond to arid, semiarid, sub-humid, and humid climates, respectively. These regions were selected because they have extensive in-situ soil moisture observations available. The soil moisture estimates from the remote-sensing method that accounts for regional characteristics were compared to the in-situ measurements and found to outperform the estimates using the single empirical Λ-θ relationship.