Convection-Allowing Ensemble Analysis and Prediction of Tornadic Thunderstorms with Assimilation of GOES-R All-Sky Radiances and WSR-88D Radar Observations

The capability of imaging atmospheric conditions of the Earth with high spatial resolution and temporal frequency of the ABI onboard the GOES-R series satellites provides new opportunities for the monitoring and prediction of severe thunderstorms and associate weather hazards, and it is of particular interests to incorporate all-sky satellite brightness temperature (BT) observations with NWP models using ensemble-based data assimilation techniques since they are able to provide flow-dependent background error covariances. This study presents the first practice of assimilating real-world all-sky ABI infrared BT observations using a WRF-EnKF data assimilation system, focusing on the analysis and forecast of a tornadic thunderstorm event across Wyoming and Nebraska on 12 June 2017 at a convection-allowing 1-km horizontal resolution. It is found that spurious clouds created by the WRF model before observed convection initiation are rapidly removed, and new clouds are also effectively triggered where storms are observed, when all-sky BT observations are assimilated with the adaptive observation error inflation (AOEI) and adaptive background error inflation (ABEI) techniques that are specifically designed for satellite BT assimilation. As a results, analysis and forecasts of the thunderstorms are significantly improved: better forecasts of the timing and location of convection initiation can be achieved after only 30 minutes of assimilating BT observations; both deterministic and probabilistic WRF forecasts of mid-level mesocyclones and low-level vortices, started from the final analysis with 100 minutes of BT assimilation, closely coincide with the tornado reports. Furthermore, when additional radar radial velocity and reflectivity observations are simultaneously assimilated with BT, a higher confidence in the probabilistic prediction of the mesocyclones and a longer forecast lead-times can be gained, as well as a better depiction of the thunderstorm’s structure which is concealed below the cloud tops when assimilating BT alone. These results indicate the great potential of assimilating BT observations from GOES-R satellites in improving severe weather forecasts and warnings.