New Development of the Hybrid Data Assimilation and Forecasting System for the Warn-on-Forecast Project During the HWT Spring Experiment in 2019

The hybrid data assimilation system (WoFS-hybrid) for the Warn-on-Forecast (WoF) project has been developed at National Severe Storms Laboratory during the past several years. The WoFS-hybrid is designed to take both advantages of the ensemble-based Warn-on-Forecast System (WoFS) and the 3DEnVAR analysis method designed specifically for convective-scale applications (Wang et al. 2019). The ensemble-based WoFS has been extensively tested with HWT spring experiments in recent years. The variational analysis component was also tested as a high-resolution analysis system for severe weather, but has not been tested extensively as a short-term forecast system. The WoFS-hybrid system assimilates NEXRAD radar observations (both reflectivity and radial velocity), satellite retrieved cloud water path, and conventional observations to perform 15-minute data assimilation cycles together with the ensemble-based WoFS. Then one deterministic forecast based on the hybrid analysis with 3 hours forecast length are launched every 30 minutes. The WoFS-hybrid system has the features of weather adaptive, dual-resolution implementation and computational efficiency. The goal is to provide deterministic physically-consistent gridded analysis and forecast products to decision makers (to supplement the ensemble-based products) for making warning decisions in a timely manner. The deterministic short-term (0-3 hours) forecast experiments launched from the WoFS-hybrid analysis started in 2017. Since its implementation, several changes have been made, including radar data and conventional data format/preprocessing, pseudo water vapor assimilation (Lai et al. 2019) and assimilation of GOES satellite data. In 2019, the WoFS-hybrid is initialized from the ensemble mean of the High-Resolution Rapid Refresh Ensemble (HRRRE) members instead of the HRRR forecast used earlier. Furthermore, the legacy LAPS format for conventional dataset is replaced with the NCEP PreBUFR format and the satellite data assimilation includes now both GOES cloud water path and GLM observations. The error statistics for the radar data and conventional observations are adjusted and the assimilation parameters are also optimized based on sensitivity study with retrospective cases in 2017 & 2018 respectively. The impact of these changes on the analysis results and the 0-3 hours convective-allowing forecasts during 2019 HWT spring experiments will be evaluated and reported in the conference.

Lai et al., 2019: Assimilation of Radar Radial Velocity, Reflectivity, and Pseudo–Water Vapor for Convective-Scale NWP in a Variational Framework. Monthly Weather Review, Vol. 147, 2877-2900.

Wang et al., 2019: Test of a Weather-Adaptive Dual-Resolution Hybrid Warn-on-ForecastAnalysis and Forecast System for Several Severe Weather Events. Wea. Forecasting, submitted.