Determining the Impact of Assimilating Satellite Radiance Data for Forecasts within a Mesoscale Ensemble Kalman Filter

Advancements in computer technology and remote sensing have created a unique scenario for the data assimilation community as satellite data assimilation is now possible within storm-scale ensemble systems. While observation impacts within coarser global models have been more heavily investigated, it is unclear how satellite data assimilation affects forecasts of high-impact events on a range of scales in finer scale ensemble systems. The overarching goal of this research is to analyze the impacts of all-sky satellite radiance observations on the predictability of high-impact weather events including severe convection, heavy precipitation, and winter weather within a high resolution EnKF framework. Fifteen events within the last decade were chosen as a first set for this analysis. A 42-member ensemble initialized 48 hours prior to each observed high-impact event, and spun up to achieve flow-dependent covariances, was simulated with the assimilation of all available conventional and radiance observations every 6 hours. The relative contribution of satellite radiance assimilation on forecasts of severe convection is quantified through several verification metrics and compared to assimilation runs assimilating only conventional observations. The strategy for future experiments is discussed in light of the results.