Title: Meso- and storm-scale assimilation to improve very-short-term forecasts of hazardous weather within the RUC, Rapid Refresh, and HRRR models

Authors: Steve Weygandt, Stan Benjamin, David Dowell, Ming Hu, Curtis Alexander, Eric James, Patrick Hofmann, Tanya Smirnova, Haidao Lin, and John Brown – NOAA Earth System Research Laboratory, Boulder, CO

Accurate short-term prediction of local weather conditions (including weather hazards such as thunderstorms, low ceilings/fog, icing, localized heavy rain/snow, localized wind events, etc.) depends critically on both mesoscale and storm-scale data assimilation. Key foci for mesoscale data assimilation systems include 1) incorporating near surface observations (METARs, mesonets, boundary layer profilers, etc.) to provide more accurate depiction of near surface conditions and 2) analyzing cloud and precipitation systems. Storm-scale data assimilation has focused on use of Doppler radar data (reflectivity and radial velocity) to improve forecasts of small-scale precipitation elements, especially convection.

The mesoscale Rapid Update Cycle (RUC) and Rapid Refresh (RR) analysis and prediction systems, and the convective-scale High Resolution Rapid Refresh (HRRR) that runs as a nest with the RUC or RR, incorporate advanced real-time mesoscale / storm-scale assimilation procedures. Within both the RUC and the RR, an hourly cycled 3DVAR analysis is employed, augmented by twice daily partial cycle updates for the RR. In addition, both system use a novel approach for surface observation assimilation and include a detailed cloud / hydrometeor analysis that utilizes METAR and satellite-based cloud information to update cycled microphysics fields.

Of crucial importance for short-term thunderstorm and precipitation forecasts, both the RUC and RR also include a diabatic-digital filter-initialization based radar reflectivity analysis procedure (radar-DFI). Nested HRRR forecasts benefit from this cycled RUC/RR reflectivity analysis procedure (in addition to benefitting from the RUC/ RR traditional 3DVAR and cloud analysis procedures). At the workshop we will provide case study examples to illustrate how the radar-DFI procedure induces storm-scale divergence circulation patterns leading to improved shorn-term forecasts of precipitation systems by the HRRR.

We will also describe recent experiments to examine the benefit to HRRR forecasts of supplemental application of a 3-km latent-heating-based reflectivity assimilation procedure at sub-hourly intervals. Analysis of these results for active convective storms has highlighted to importance of both the mesoscale and storm-scale assimilation at different forecast lead times. These detailed results will complement an overall documentation of the forecast impact from the various aspects of the RUC / RR / HRRR mesoscale / storm-scale assimilation system, with an emphasis on what techniques and observations are most crucial for which specific weather hazards. Finally, we will discuss ongoing / planned work on new strategies and observing systems.