Evapotranspiration estimation over agricultural plains using MODIS data for all sky conditions

Abstract

Agricultural, water resource, and forest management require evapotranspiration (ET) data over a range of spatial and temporal scales. Several methods, with reasonable accuracy, exist for computing ET for clear-sky conditions using satellite remote sensing data. However, computation of ET for cloudy pixels remains a challenge. Here, we have developed an ET estimation algorithm for all sky conditions using primarily remote sensing data. The approach is based on an extension of the Priestley–Taylor model with contextual interpretation of remotely sensed surface temperature and vegetation index. It requires estimation of evaporative fraction, soil heat flux, and net radiation that can be derived from the Moderate Resolution Imaging Spectroradiometer (MODIS). For cloudy pixels, a regression-based algorithm utilizes MODIS cloud data product for cloud-top temperature, cloud fraction, cloud emissivity, cloud optical thickness, and land-surface temperature (LST) to calculate net radiation and thereafter ET. The algorithm was tested and validated using data from the South Great Plains in the USA with reasonable accuracy of ranging between 54.5% and 69.6% and an average root mean square error of 40.3 . The methodology was tested to estimate ET over the agricultural plains of the Ganges Basin in India. The proposed ET algorithm has capabilities to estimate ET over a range of space and time scales in data-scarce regions.