6.4 Evaluation of Quantitative Precipitation Forecast of High-Resolution Rapid Refresh Model

Radhakrishnan, Chandrasekar, Haonan Chen, and V.Chandrasekar, Colorado State University, Fort Collins, CO

A continental-scale hourly updated assimilation and model forecast system, termed "Rapid Refresh" or "RAP" in short, has been developed by NOAA, in order to produce operational precipitation forecasts. Nowadays, the RAP system is one of the mainstream models running operationally in National Centers for Environmental Prediction (NCEP). As a complement of RAP, the 3km High-Resolution Rapid Refresh (HRRR) system is also updated hourly, but covering a smaller geographic domain. The HRRR is using the community-based Advanced Research version of the Weather Research and Forecasting (WRF) Model (ARW) and Gridpoint Statistical Interpolation analysis system (GSI) for assimilation. The model domain covers the Contiguous United States (CONUS) with 3 km resolution. The HRRR model initiates every one hour and provides the forecast for 18 hours, totally 24 forecasts are available for a day.

This study will evaluate the Quantitative Precipitation Forecast (QPF) of the HRRR model over Dallas-Fort Worth (DFW) urban region. This study also develops an algorithm which provides an optimal ensemble forecast (up to 6 hours) by combining the previous forecast for the same time. The high spatiotemporal resolution Quantitative Precipitation Estimation (QPE) from the Collaborative Adaptive Sensing of the Atmosphere (CASA) Engineering Research Center for the DFW urban radar network is used for the model QPF evaluation. The real-time CASA DFW QPE system uses both the high-resolution X-band radar network and the National Weather Service (NWS) S-band radar observations.