Modelling the dynamics of evapotranspiration using Variable Infiltration Capacity model and regionally calibrated Hargreaves approach

Computation of reference evapotranspiration (ETO) at different spatiotemporal scales is constrained by limited in situ meteorological data availability which, in turn, led to alternate ET estimation methods using the commonly available meteorological data of maximum and minimum temperatures. In this study, the Hargreaves–Samani model and the water budget approach of the Variable Infiltration Capacity (VIC-3L) land surface model was evaluated for grid-scale actual ET estimation using the benchmark FAO-56 Penman–Monteith (PM) equation and crop coefficient relationships. These approaches were field-tested in the Kangsabati River basin in eastern India, a tropical monsoon-type climate region with dominant paddy land uses. The results revealed that the VIC model could estimate the grid-scale ET reasonably well; however, the corresponding estimates by the Hargreaves method were highly overestimated. To enhance the field applicability of the Hargreaves method for data-scarce regions, this method was coupled with a genetic algorithm-based bias correction approach that improved the Nash–Sutcliffe efficiency significantly. Hence, this study reveals that there is a need for regional-scale standardization of the Hargreaves ET estimates using the FAO-56 PM, lysimeter data or the VIC-3L model