Overview of Quick OSSEs in Support of the Warn-on-Forecast Project

The first study utilizes an Observing System Simulation Experiment (OSSE) to explore the combined impact of assimilating GOES-R Advanced Baseline Imager (ABI) 6.95 μm brightness temperatures and WSR-88D Doppler radar reflectivity and radial velocity observations in an ensemble data assimilation system.  A high-resolution nature run was used to create synthetic radar and satellite observations corresponding to a severe weather event that occurred across the Central Plains on 4-5 June 2005.  This and the following experiment employ the WRF-ARW model combined with the ensemble adjustment Kalman (EaKF) filter included in the Data Assimilation Research Testbed system. Assimilating GOES-R ABI clear and cloudy sky brightness temperatures along with radar observations resulted in superior short-term forecast skill compared to only assimilating one or the other observation types.

The second study used an OSSE approach to generate synthetic temperature and humidity profiles from a hypothetical geostationary-based sounder from a nature run of a high impact weather event in central Oklahoma on 20 May 2013. The synthetic observations are then assimilated using an EaKF approach with hourly and 15 minute cycling to determine their effectiveness at improving the near storm environment. Results indicate that assimilating both temperature and humidity profiles reduced mid-tropospheric bias and error compared to assimilating conventional observations alone. While hourly cycling was generally effective, 15-minute cycling generally produced the lowest errors while also generating the best 2-4 hour updraft helicity forecasts of ongoing convection. 

These projects represent two examples of how satellite based remote sensing observations may be used in WoF type forecast models. As increasing amounts of real data become available over time, the assimilation techniques used in these OSSE experiments will prove valuable in making the transition to an operational product with real observations.