Reliable water resource allocation decision-making demands a reliable flow forecasting system, specifically during lean flow periods. Interstate Commission on the Potomac River Basin helps manage the Washington metropolitan area water supply system by coordinating withdrawals from the Potomac River Basin (PRB) and off-river reservoirs and recommending releases from upstream reservoirs during drought. These reservoir operation rules rely on low-flow forecasts to meet water supply, environmental, and recreational needs. While most hydrological early warning systems target short-range forecasts and floods, low-flow forecast systems must effectively capture medium-range rainfall dry spells to simulate reliable soil moisture and groundwater contributions. In this study, we are developing a drought forecasting system tailored to the unique context and features of PRB. The system utilizes precipitation and temperature from the Global Ensemble Forecast System (GEFS) with an eco-hydrological model (SWAT+) augmented with a detailed grid-based groundwater component to forecast flows. The model modeling framework also included improved reservoir operation routines to implement seasonal and complex operation rules in PRB. We have developed a framework to verify flow forecasts using retrospective ensemble weather forecasts from 2000-2019. Initial validations through the Brier score indicated higher accuracy for rainfall occurrences than nonoccurrences. The threat score and hit rate were higher (>75%) for non-rainy and low-flow periods with less than 20% false alarm rate. This verified process-based flow simulation from operational weather forecast could be a guiding tool for water management in the PRB for the Interstate Commission on the Potomac River Basin.
