Improving Reservoir Management in North Central Texas using Ensemble Forecasting

The Dallas-Fort Worth (DFW) Metroplex is one of the fastest growing metropolitan areas in the US in terms of population. Being largely dependent on surface water, raw water suppliers such as the Tarrant Regional Water District (TRWD) operate systems of reservoirs interconnected through extensive pipe networks to meet the demand. For operational decision support, TRWD uses a computer model of their reservoir system, which currently uses forecasted inflows generated based on climate variables or transition probabilities. The goal of this research is to infuse skillful ensemble inflow forecasts forced by weather and climate forecasts into the TRWD’s model to improve the water supply operations in North Central Texas. To assess the value of medium- and long-range forecasts in TRWD’s decision support, a set of hindcasting and verification experiments is carried out. Using the Global Ensemble Forecast System and the Climate Forecast System Version 2, inflow ensemble hindcasts are generated using the National Weather Service’s Hydrologic Ensemble Forecast Service. TRWD’s RiverWare-based reservoir model is then run using the inflow hindcasts to generate outflow hindcasts for each reservoir. Verification of the hindcasts is carried out using the Ensemble Verification System. To assess the value of the inflow forecasts into the decision support tool, cost-benefit analysis is carried out. In this presentation, we share the verification and preliminary cost-benefit analysis results.