Simultaneous Multiscale Data Assimilation of GOES‐16 ABI All‐Sky Radiances and Radar Reflectivity on a Dryline Convection Event

Abstract This study investigates the impact of simultaneous multiscale data assimilation (MDA) of GOES‐16 Advanced Baseline Imager infrared all‐sky radiance and radar reflectivity on a dryline convection event on 26–27 May 2019 through comparison with single‐scale data assimilation (SDA). The dryline convection event, which included tornadic supercells and a bowing mesoscale convective system (MCS) over the continental United States, was influenced by large‐scale supporting environments, including an upper‐level jet and shortwave trough. The simultaneous MDA of all‐sky radiance and radar reflectivity is conducted using scale‐dependent localization and variable‐dependent localization within the ensemble variational DA system. The MDA configuration, including scale separations, localization lengths, and cross‐scale and cross‐variable covariances, is optimized by examining ensemble perturbations across multiple scales. Single observation experiments indicate that MDA of all‐sky radiance more effectively analyzes large‐scale features than SDA with various localization lengths. Although previous studies examined MDA through the assimilation of radar reflectivity that samples low‐level precipitation structures, this study highlights the impact of MDA assimilating both all‐sky radiance and radar reflectivity. Together these two observations sample mid‐ and low‐level storms at convective and mesoscale, synoptic‐scale inflow, upper‐level clouds, and large‐scale moisture features. Our cycled MDA experiments demonstrate improved multiscale analysis of low‐level cold pool, convergence, inflow, upper‐level jet, and shortwave trough, using MDA, compared to an optimally tuned SDA that uses a short 15‐km localization radius. These improved multiscale features enhance the predictability of the initiating MCS, its bowing process, and embedded supercells during the subsequent 6‐hr forecast.