Assessing the Value of Assimilating Surface, PBL, and Free Atmosphere Observations from TORUS on Storm-Scale Ensemble Forecasts

Data assimilation using the current meteorological observing network often struggles to properly represent storm-scale environmental features, which can result in storms spun up through radar DA evolving inaccurately in subsequent model forecasts. Assimilating unconventional observations may be helpful in mitigating this problem by improving storm-scale ensemble analyses of such features. This study examines the impact of assimilating near- and in-storm observations from the Targeted Observation by Radars and UAS of Supercells (TORUS) project on storm-scale ensemble forecasts of two supercells on 8 June 2019. TORUS observations from mobile mesonets, UAS, and special soundings are divided into subsets representing the surface, PBL, and free atmosphere. Assimilating all three subsets together leads to an improved representation of the evolution of the second supercell, as well as an improved representation of elevated convection which develops elsewhere in northwestern KS. The impact of assimilating observations from each subset individually on the accuracy of storm-scale ensemble analyses and forecasts for the 8 June case will be evaluated through data denial experiments to determine the contribution of each subset of TORUS observations to this forecast improvement.