Direct Assimilation of GOES-16 ABI All-Sky Radiances in HAFS Self-Cycled Dual-Resolution 3DEnVar System: System Description and a Case Study of Hurricane Laura (2020) Weather and Forecasting

Compared to utilizing temperature and mixing ratios of water vapor and hydrometeors as state variables (Standard_state), the direct assimilation of GOES-16 Advanced Baseline Imager (ABI) channel-10 all-sky radiances is advanced by additionally including brightness temperature (BT) as a state variable (BT_state) within the Hurricane Analysis and Forecast System (HAFS) self-cycled dual-resolution three-dimensional ensemble–variational (3DEnVar) data assimilation (DA) system. With a focus on the rapid intensification (RI) predictions for Hurricane Laura (2020), the ABI all-sky radiances and operational observations are simultaneously assimilated during the pre-RI to RI period. Statistical comparisons across multiple DA cycles demonstrate the superiority of BT_state over Standard_state in terms of objective alignment with observations for background, analyses, and forecasts. Compared to Standard_state, the closer fit of BT_state to observations benefits from better convergence and faster variational minimization. The performances of the background, analyses, and forecasts from Standard_state and BT_state are verified subjectively against various observations. During the DA period, the background and analysis from BT_state outperform Standard_state in capturing the storm size, surrounding clear-sky areas, and the wind field structure within the storm region. Additionally, BT_state yields higher forecast skill than Standard_state for the intensity, track, and main structural features of Laura. Detailed diagnostics suggest that the enhanced thermodynamic structure analysis in BT_state facilitates the RI prediction, and the improved large-scale environment analysis in BT_state contributes to the accurate track prediction. Grant no. NA22OAR4590183