Estimating hourly crop ET using a two-source energy balance model and multispectral airborne imagery

Abstract

Efficient water use through improved irrigation scheduling is expected to moderate fast declining groundwater levels and improve sustainability of the Ogallala Aquifer. An accurate estimation of spatial actual evapotranspiration (ET) is needed for this purpose. Therefore, during 2007, the Bushland ET and Agricultural Remote Sensing Experiment (BEAREX07) was conducted at the USDA-ARS, in Bushland, Texas, to evaluate remote sensing (RS)-based surface energy balance models. Very high-resolution aircraft images were acquired using the Utah State University airborne multispectral system. Instantaneous ET was estimated using a two-source energy balance model (TSM). A minor modification was made in the calculation of sensible heat fluxes to improve ET estimation. Furthermore, a sensitivity analysis of selected variables was conducted to evaluate their effect on ET estimation. Data from four weighing lysimeters, planted to sorghum and corn, were used for evaluating ET predictions. Instantaneous ET was predicted with mean bias error and root mean square error of 0.03 and 0.07 mm h−1 (4.3 and 11.7%), respectively. Results indicated that crop height, roughness length for momentum transfer, clumping factor and soil resistance sub-models need to be refined. Nevertheless, the application of the TSM using high-resolution RS imagery in the Southern High Plains is promising.