Simplified spatially distributed, coupled soil moisture–groundwater model, evaluated for Flanders (Belgium)

Flanders region of Belgium. In hydrological studies, there is a need for computationally fast, spatially distributed models involving the unsaturated zone soil moisture, groundwater and their interactions, while maintaining good accuracy. This study builds on an approach based on the AquaCrop model for the unsaturated zone and a 2D explicit groundwater flow model. The model’s speed stems from its simplicity, while AquaCrop’s wide use provides a basic database for parameterization. Capillary rise, often not (explicitly) assessed in distributed models, is considered. For coupling the unsaturated and saturated zones better, the capillary rise algorithm by AquaCrop was modified. Groundwater flow to the rivers is represented combining Darcy’s law and the linear reservoir concept. Surface runoff modelling is modified to better incorporate the effects of urban environments. The model was parameterized using publicly available data (involving minimal calibration) and tested for Flanders. Despite its simplicity, the model often performed acceptably. Its accuracy was generally higher for the degree of saturation in the upper 70 cm of the soil (R=0.90, RMSE=0.14) than the first 10 cm (R=0.68, RMSE=0.17). The groundwater levels for wells with observed depths less than 5 m were simulated better (R=0.82, RMSE=0.72 m) than the deeper ones (R=0.63, RMSE=4.59 m). Capillary rise was found to be a potentially substantial source of water to the crops. • A distributed hydrological model, aiming at combining accuracy and speed, is set up and tested for northern Belgium. • The unsaturated zone is modelled largely based on the AquaCrop model methodology, while a 2D groundwater model is used. • Capillary rise is considered. • Despite its simplicity, the model often performs decently regarding the degree of soil saturation and the groundwater levels.