Verification of the NOAA/NWS MMEFS Operational Hydrologic Ensemble Forecasting System in the Ohio River Valley

U.S. National Weather Service (NWS) River Forecast Centers (RFCs) have produced monthly to seasonal hydrologic forecast ensemble outlooks for many years using historical precipitation and temperature as the forcing data for the ensembles within the Advanced Hydrologic Prediction System (AHPS). This has worked well for many water resources applications relying on long lead-time ensembles of greater than thirty days. In the eastern United States, there is considerable interest in shorter lead-time ensembles, particularly in the period of one to seven days that convey hydrologic forecast uncertainty resulting directly from the uncertainties in precipitation and temperature forecasts, with a focus on flood forecasting. This paper demonstrates the use of numerical weather prediction (NWP) model ensembles as the forcing for hydrologic models within the NWS hydrologic forecasting system, the Deltares Flood Early Warning System (FEWS) based Community Hydrologic Prediction System (CHPS), utilizing Ensemble Streamflow Prediction (ESP) system techniques. This approach makes it possible to produce probabilistic hydrologic forecasts that incorporate the skill and uncertainty of atmospheric models. Significant issues exist with directly using the meteorological ensembles, including biases and spatial resolution of the available data, but the approach allows for the generation of valuable information for end-user decision makers, covering a time frame for which there is little current probabilistic hydrologic guidance available. This paper presents a description of a joint project among four NWS RFCs to incorporate ensemble forecasts from a variety of meteorological model ensembles. These include the National Oceanic and Atmospheric Administration (NOAA) NWS National Centers for Environmental Prediction (NCEP) Global Ensemble Forecast System (GEFS), North American Ensemble Forecast System (NAEFS), and Short Range Ensemble Forecast System (SREF). Verification results are presented for operational NAEFS and SREF based MMEFS real-time probabilistic hydrologic forecasts covering multi-year period. Improvements to the MMEFS are also recommended.