NCEP Modeling Efforts to Assist with Nowcasting Challenges

Numerical weather prediction has great potential to assist with the major challenge of nowcasting. These very short-range forecasts can be greatly improved through the use of frequently-updated high-resolution models, and the assimilation of radar data gives them greater opportunities to start with a more-correct initial picture of the atmosphere and avoid the spin-up issues that affect models without assimilation of radar data.

NCEP currently runs the NAM nest every 6 hours, the ARW and NMMB high-resolution windows (HiResW) every 12 hours, and the High-Resolution Rapid Refresh (HRRR) hourly. The NAM nest and Hi-Res Windows are run over CONUS, Alaska, Puerto Rico, and Hawaii, while the HRRR is run over CONUS and will soon be run (at 3-hour intervals) over Alaska. All have horizontal resolution at or close to 3 km. The NAM nest and HRRR use radar reflectivity data in their intialization, while the Hi-Res Windows do not directly assimilate radar data, but the CONUS and Puerto Rico versions get their initial conditions from a model (RAP) that does.

Of course, with multiple high-resolution models that are frequently updated, there is a lot of data for a forecaster to process, and generating simplified output can greatly assist the forecasting process. NCEP will soon be implementing the Version 2 of the High-Resolution Ensemble Forecasting system (HREF) which generates mean and probability fields from several of the high-resolution forecasting systems, effectively matching the output from the popular Storm-Scale Ensemble of Opportunity (SSEO) generated by the Storm Prediction Center.

Going forward, NCEP plans to consolidate its modeling suite and potentially add frequently-updated high-resolution ensembles. Version 2 of the HREF will serve as a benchmark for future high-resolution ensemble systems to beat. Future NCEP modeling plans in support of nowcasting capabilities will round out the presentation.