Use of Remote Sensed Products for Large-scale SWAT+ Model Calibration

Calibration of large-scale models comes with several challenges. Among these challenges are the availability of observation data and the computational cost of running the calibration. As such, some large-scale models are not calibrated. Yet calibration of impact models is crucial, as Krysanova et al. (2018) concluded. A calibration strategy focusing on hydrological mass balance can reduce calibration computational costs and improve the model application while global remote sensed products provide data for large-scale applications. This study presents the Hydrological Mass Balance Calibration (HMBC) methodology for SWAT+. We test the method using a remotely sensed ET product (WaPOR). We also use flow data from the Global Runoff Data Centre (GRDC) and compare projections made by the HMBC model and those without. We then apply the HMBC to a SWAT+ model for Africa. Results show that HMBC leads to improved simulation of discharge and evapotranspiration with fewer iterations than a full parameter calibration. Substantial spatial differences are also observed in projections made by the HMBC model compared to the uncalibrated model. Thus, it makes a difference if we apply HMBC. In addition, HMBC used with remotely sensed data can remedy some barriers to the calibration of hydrological models applied at large scales, as demonstrated by application to the Africa SWAT+ model.