Deriving hourly evapotranspiration rates with SEBS; a lysimetric evaluation

Numerous energy balance (EB) algorithms have been developed to use remote sensing data for mapping evapotranspiration (ET) on a regional basis. Adopting any single or combination of these models for an operational ET remote sensing program requires a thorough evaluation. The Surface Energy Balance System (SEBS) was evaluated for its ability to estimate hourly ET rates of summer tall and short crops grown in the Texas High Plains by using 15 Landsat 5 Thematic Mapper scenes acquired during 2006 to 2009. Performance of SEBS was evaluated by comparing estimated hourly ET values with measured ET data from four large weighing lysimeters, each located at the center of a 4.3 ha field in the USDA-ARS Conservation and Production Research Laboratory in Bushland, TX. The performance of SEBS in estimating hourly ET was good for crops under both irrigated and dryland conditions. A locally derived, surface albedo-based soil heat flux (G) model further improved the G estimates. Root mean square error and mean bias error were 0.11 and -0.005 mm h (super -1) , respectively, and the Nash-Sutcliff model efficiency was 0.85 between the measured and calculated hourly ET. Considering the equal or better performance with a minimal amount of ancillary data as compared to with other EB algorithms, SEBS is a promising tool for use in an operational ET remote sensing program in the semiarid Texas High Plains. However, thorough sensitivity and error propagation analyses of input variables to quantify their impact on ET estimations for the major crops in the Texas High Plains under different agroclimatological conditions are needed before adopting the SEBS into operational ET remote sensing programs for irrigation scheduling or other purposes.