Conventionally, hydrological models are calibrated with streamflow data. However, streamflow-only calibration does not guarantee a reliable spatial representation of other hydrological fluxes and states with a distributed hydrological model. The increasing availability of satellite remote sensing (SRS) data has promoted the development of spatial hydrology and large-scale hydrological modelling. Multivariate calibration based on the simultaneous use of multiple SRS products can improve model performances.

This talk will present various multivariate calibration approaches based on the simultaneous incorporation of streamflow data and spatial patterns of multiple SRS products using the Mesoscale Hydrologic Model (mHM), implemented over the Volta River Basin (West Africa). The SRS products describe different hydrological processes (i.e. evaporation, soil moisture and terrestrial water storage change). A new bias insensitive spatial pattern metric is proposed as objective function. The impact of the choice of the SRS products, the calibration strategies and the meteorological forcing are investigated through various scenarios.

Results show an improvement in the prediction of spatial patterns of evaporation and soil moisture, while a deterioration is observed in the temporal dynamics of streamflow and terrestrial water storage change. Overall, it is found that spatial patterns of SRS products are their key feature, which can be used without absolute values to improve the predictive skill of hydrological models, thereby advancing the spatiotemporal prediction of floods and droughts.