How reliable are the evapotranspiration estimates by Soil and Water Assessment Tool (SWAT) and Variable Infiltration Capacity (VIC) models for catchment-scale drought assessment and irrigation planning?

Catchment-scale assessmen t of evapotranspiration (ET) becomes indispensable for evaluating agricultural drought hazard and irrigation water management. As direct estimation of ET is quite intricate under sparse data availability scenarios, the recent indirect ET estimation approaches heavily depend on the water balance-based ET estimates from the popular land surface models, such as, Soil and Water Assessment Tool (SWAT) and Variable Infiltration Capacity (VIC-3L) model or the MOD16A2 satellite product. Along with the land uses, the irrigation application methods of alternate wetting and drying (AWD) could also influence the actual ET (ETact) estimates. This study evaluates the usability of these indirect ET estimation methods with reference to the moisture-stress specific FAO-56 Penman-Monteith (PM) equation for catchment-scale drought and irrigation planning in a real paddy-dominated integrated reservoir-catchment command. The results reveal that the SWAT-derived ET in the non-irrigated paddy fields was consistent with the FAO-56 PM based ET estimates during two crop growing seasons; however, significant overestimation was observed by the VIC-3L, and MOD16A2 products showed a relatively better estimate. The SWAT and VIC-3L derived Standardized Soil Moisture Index (SSI) and Standardized Evapotranspiration Drought Index (SEDI) for the study area revealed that the VIC-3L based drought severity was continuously overestimated or underestimated; whereas the SWAT-derived SEDI indices were almost consistent with the benchmark MOD16A2-ET and ESA-CCI soil moisture derived estimates. Finally, this study endorses to use these indirect catchment-scale drought assessment tools with caution with a need to revise the existing ET subroutines in these hydrological models to deal with the crop dynamics and water management options realistically.