Mapping ET at High Resolution in an Advective Semi-Arid Environment with Airborne Multispectral Imagery

Abstract

Abstract Periodic and accurate estimates of spatially distributed evapotranspiration (ET) are essential for managing water in irrigated regions and in hydrologic modeling. In this study, METRIC (Mapping ET at high Resolutions with Internal Calibration), an energy balance algorithm originally developed for application with Landsat imagery at a regional scale, was applied to very high resolution aircraft imagery (0.5-2 m pixels) in the Texas High Plains. ET predictions were evaluated using data from four large precision weighing lysimeters located in the USDA-ARS Conservation and Production Research Laboratory, Bushland, Texas. Statistical results indicated that METRIC worked better for crops with leaf area index greater than 2.5 m 2 m -2 . Potential limitations may have been the areal extent of the imagery, the surface roughness for the momentum transfer sub-model, and the lack of a cold pixel with characteristics similar to the reference crop, i.e. alfalfa. Introduction Remote sensing (RS) derived ET maps can potentially be used in the monitoring of spatially distributed crop water use, to schedule irrigations, and as input for hydrologic models. Also, spatially distributed seasonal ET may be used to assess the water use efficiency of irrigation projects. Spatially distributed ET has been estimated using land surface energy balance (EB) models using RS multispectral imagery and ground-based micro-meteorological data. Gowda et al. (2008) and Gowda et al. (2007) present a description and discussion on most of the RS-based EB models available in the literature. Most of these EB models are single source models, e.g. SEBI (Menenti and Choudhury, 1993), SEBAL (Bastiaanssen et al., 1998), SEBS (Su, 2002), METRIC (Allen et al., 2007a), ReSET (Elhaddad and Garcia, 2008). Of all these algorithms, METRIC may have an advantage under advective conditions. METRIC’s ET estimation errors were reported to be approximately 10 to 20% for