An Ensemble Hydrologic Modeling System for Runoff and Evapotranspiration Evaluation over an Agricultural Watershed

Modeling framework for simulation at a finer scale is important for long-term water resources planning for management. It has always been a challenge to select the appropriate model to simulate the hydrology of a watershed/river basin at a finer spatial resolution. Comparative evaluation of models based on field observations could help researchers to select the suitable model for their purpose. However, a single hydrologic model generally leads to simulation uncertainties due to poor input data, model structure, and model output uncertainty in large-scale exercises. The ensemble model approach could be a better decision-making tool to overcome uncertainty in modeling hydrological processes. In the present study, a widely used macroscale hydrologic model, the three-layer Variable Infiltration Capacity (VIC-3L), was employed to simulate runoff and evapotranspiration (ET) at 3′ × 3′ grids (~ 5.5 km) resolution over an agriculture-based Marol watershed (5092 km 2 ) of India. The VIC-simulated results were compared and assessed with the results obtained from the Hydrologic Response Unit (HRU)-based Soil and Water Assessment Tool (SWAT) hydrologic model. Further, the ensemble of VIC and SWAT outputs (EnSwaVi; averages of individual model-simulated datasets with equal weights) was also assessed. Simulated runoff and ET were evaluated using observed discharge data at the outlet of the watershed and the actual ET product (MOD16A2) of Moderate Resolution Imaging Spectroradiometer (MODIS), respectively. The simulated discharge values generated by the two models were closely matched with the observed flow. Conversely, ET simulated by VIC was found to be more precise as compared to SWAT. A minimal difference between two model results can be due to the difference in the model structure and runoff simulation method. In general, the ensembles of VIC and SWAT outputs (EnSwaVi) were found better than the individual model outputs. The ensemble modeling approach could provide more reliable assessments of hydrological processes for the planning and management of water resources.