Low-flow information is essential for water-resources management and drought mitigation, yet its simulation remains challenging because low flows emerge from complex interactions among climate drivers, catchment processes, and human regulation. This study evaluates the capability of a physically based land surface model (LSM) to simulate low flows in a regulated basin, with emphasis on soil-depth representation. The Simple Biosphere model including Urban Canopy (SiBUC), coupled with a flow-routing model, was applied to the Toyo River Basin in Japan. Simulations were conducted using two soil-depth datasets: a global product and locally derived borehole measurements. Under the default configuration, the model reproduced evapotranspiration and streamflow reasonably well, with further improvement after accounting for water-management effects. Low flows, defined as the 10th lowest discharge in the flow duration curve, were underestimated; however, this bias was reduced when locally measured, shallower soil depth was used. The resulting increase in simulated low flows was linked to enhanced subsurface runoff arising from reduced storage capacity and more frequent near-saturation of the lower soil layer. These findings indicate that physically based LSMs such as SiBUC can be effectively applied to low-flow prediction in regulated basins and highlight the importance of improved representation of subsurface and groundwater processes.
Tinumbang et al. (Thu,) studied this question.