89th American Meteorological Society Annual Meeting

Thursday, 15 January 2009: 8:45 AM
Generating short term hydrologic ensembles from numerical weather prediction model ensembles
Room 127B (Phoenix Convention Center)
Thomas Adams, NOAA/NWS, Wilmington, OH; and R. Shedd and J. Ostrowski
National Weather Service (NWS) River Forecast Centers (RFCs) have produced hydrologic forecast ensembles for many years using historical precipitation and temperature as the forcing data for the ensembles. This has worked well for longer-term ensembles of thirty days or longer. However, at least in the eastern United States, there is more interest in shorter lead-time ensembles, particularly in the period of three to seven days that convey hydrologic forecast uncertainty resulting directly from the uncertainties in precipitation and temperature forecasts. This paper will demonstrate the use of numerical weather prediction (NWP) model ensembles as the forcing for hydrologic models within the NWS River Forecast System (NWSRFS) Ensemble Streamflow Prediction (ESP) system. The authors believe this approach will provide the opportunity to develop probabilistic hydrologic forecasts that incorporate the skill and uncertainty of atmospheric models. While there are significant issues in directly using the meteorological ensembles, including biases and spatial resolution of the available data, the approach allows for the generation of information for a time frame at which there is little current probabilistic hydrologic guidance available. This paper presents a description of a joint project among three NWS RFCs to incorporate ensemble forecasts from a variety of meteorological model ensembles including the National Oceanic and Atmospheric Administration (NOAA) NWS National Centers for Environmental Prediction (NCEP) Global Ensemble Forecast System (GEFS) and Short Range Ensemble Forecast System (SREF). The paper presents the data processing procedures to convert the incoming gridded ensembles into the necessary hydrologic inputs, describes the interpolation procedures for ensemble information, and discusses a variety of potential products that are being generated. Initial verification on some of the ensemble forecasts is also presented along with some case study examples. Finally, there is a discussion on how this project fits in with future ensemble forecast work including the NWS Office of Hydrologic Development (OHD), Hydrology Laboratory, Experimental Ensemble Forecast System (XEFS).

Supplementary URL: