Storm-scale data assimilation and ensemble forecasting for Warn-on-Forecast

The NOAA Warn-on-Forecast (WoF) research project is tasked with the development of very short-range (0-1 h) probabilistic forecasts that accurately predict severe convective storms. Storm-scale data assimilation and ensemble forecasting in a future WoF system will be performed on very-high-resolution (grid spacing ~1 km or less), event-dependent grids. Development and testing of a WRF-based ensemble data assimilation system has begun on somewhat coarser convection-allowing grids, and results from several radar data assimilation experiments are presented.

Storm-scale ensemble analyses and forecasts of six tornadic events from springs 2013-14 are produced on a 3-km event-dependent grid. This storm-scale ensemble is nested within a 15-km continental United States (CONUS) ensemble constructed from initial and boundary conditions provided by members of the Global Ensemble Forecast System (GEFS) forecast cycle starting at 0000 UTC (as described more fully in a companion abstract by K. Knopfmeier). For each event, single-Doppler observations from several radars are assimilated—beginning just prior to convective initiation—with the ensemble Kalman filter (EnKF) approach encoded in the Data Assimilation Research Testbed (DART). A series of 1-h forecasts are launched during the 30-60 min period preceding the onset of storm reports.

Of particular interest is the ability of 0-1 h ensemble forecasts (initialized from storm-scale analyses) to reproduce the low-level rotational characteristics of supercell thunderstorms, as well as other convective hazards. Notably, the 11 May 2014 severe weather event over eastern Nebraska and western Iowa was marked by a supercell-to-MCS transition, with both convective modes associated with tornado and high wind reports. Additional experiments have been conducted to evaluate the impact of assimilating radar data and cloud water path retrievals at the storm-scale (as described in a companion abstract by T. Jones).