Grid-Based Model for Estimating Evapotranspiration Rates of Heterogeneous Land Surface

Abstract

The complementary relationship areal evapotranspiration (CRAE) model estimates actual evapotranspiration (ET) based on potential and wet environment evapotranspiration rates. The existing CRAE-based models that calculate ET using point-based climatic inputs often adopt an interpolation method to produce spatially continuous ET grids at a regional scale. However, the effect of land surface heterogeneity on the ET estimation process is not well captured by these approaches. In the present study, the authors have developed a grid-based CRAE model (PY-CRAE) that accounts for the spatial heterogeneity in the land surface processes for estimating ET at a 4 km scale. The developed PY-CRAE model was applied for the entire state of Nebraska with monthly averaged climatic input data sets during the period 2000–2014. The performance of the PY-CRAE model along with the point-based CRAE model (WREVAP-CRAE) and a moderate resolution imaging spectro-radiometer (MODIS)-ET data products were assessed against water balance-based ET values. The results indicate that the PY-CRAE model performance was superior when compared to WREVAP-CRAE and MODIS-ET models with mean absolute error and root-mean-square error values of 40.94 and 49.15 mm/year, respectively.