Predictions of Tornado-like Vortex Embedded in the 8 May 2003 Oklahoma City Tornadic Supercell Initialized from the Sub-Kilometer Analysis Produced By the GSI Based EnVar Data Assimilation System: Methodology and Sensitivity to Horizontal and Vertical Grid Spacing and Physical Parameterizations

Previous studies often initialize a high-resolution tornado-like vortices (TLVs) prediction from an analysis at a coarser resolution, mostly larger than or equal to the 1 km grid-spacing. Such a coarse resolution initial conditions (ICs) may not capture the fine scale characteristics and circulation associated with the TLVs. To investigate the impact of the sub-kilometer (grid spacing < 1km) analysis to initialize the prediction of TLVs, an efficient approach to generate the high-resolution analysis is proposed by extending the GSI based dual resolution (DR) EnVar system with direct radar data assimilation capability (Wang and Wang 2017). In the DR approach, the control analysis is at a sub-kilometer resolution (e.g., 500-m) whereas the ensemble is at a coarser kilometer resolution (e.g., 2-km). The single coarse resolution experiment (SCR) as a benchmark is also conducted to generate the kilometer analysis where both control and ensemble are at a coarser resolution (e.g., 2-km). In both experiments, the free control forecast is run with a 500-m horizontal resolution. The 8 May 2003, Oklahoma City, tornadic supercell storm is selected for the study.

The prediction covers the entire period of two observed tornado outbreaks. The two strong surface vorticity swaths derived from DR fits the observed tornado damage track well in both locations and timing. In comparison, SCR only produces one strong surface vorticity swath and fails to capture the development of the second observed tornadic vortices. Such significantly differently simulated TLVs indicate the importance of initializing the prediction with a fine, sub-kilometer analysis. Further diagnostics and additional sensitivity experiments suggest the analyzed dynamic fields determine the timing of first TLV’s dissipation and the second TLV’s re-intensification; the analyzed thermodynamic fields affect the longevity and strength of the second TLV.

Additional tests are further conducted to investigate the impact of vertical levels and physics schemes configurations on the predictions of TLVs. Initial results show these configurations can strongly affect the longevity and strength of simulated TLVs. Such sensitivity motivates additional research on these configurations for both the initialization and prediction of TLVs. More experiments are ongoing and the comprehensive results will be presented on the conference.