Potential of Satellite-Based Models for Land Surface Evapotranspiration Estimation

Use of satellite remote sensing data for the estimation of surface evapotranspiration (ET) is emerging as a promising approach with a range of application potentials. Unlike traditional point-based approaches, some of these satellite-based approaches have been shown to have bounded error. This is a significant advantage of such ET estimation approaches compared to ground-based or hybrid (i.e. combining ground and remote sensing data) approaches, particularly over large heterogeneous domain where representation of key distributed parameters from ground-based observations are questionable. For clear days, main components relevant to surface ET estimation such as evaporative fraction (EF) and net radiation (Rn) can be derived using data from AVHRR (Advanced Very High Resolution Radiometer) and MODIS (Moderate Resolution Imaging Spectroradiometer) sensors. Use of temporal stability of EF and a mathematically convenient and physically realistic, at least for clear sky days, diurnal model for Rn may allow us to get a reasonably accurate estimates of daily ET over large heterogeneous areas. Extension of these results for cloudy days and related challenges will be discussed with particular examples from South Florida and Southern Great Plains using data from MODIS and AVHRR.