All-sky Assimilation Of The Infrared Radiance Data From GOES-R ABI At Convective Scale

Infrared Imagers onboard geostationary meteorological satellites are able to provide continuous images of Earth's surface,ocean,atmosphere and cloud objects.A much higher temporal and spatial resolution than the observations from polar-orbiting satellites makes it possible to capture the evolution of weather phenomena at the convective scale.Therefore,infrared observations from geostationary meteorological satellites have great potentials in Numerical Weather Prediction(NWP)at convective scale.However,only clear-sky radiances are routinely assimilated in most operational centers currently.One key problem is how to obtain useful meso-and convective-scale information from the large amount of radiance that affected by clouds and precipitations.A better understanding is also needed of how the assimilation of cloudy and clear-sky IR observations impacts,both positively and negatively,the analyzed storms as well as their environment in terms of individual state variables and on subsequent forecasts.In addition,it remains unclear on how to better select different tunable DA parameters to optimize certain En KF configurations.For these purposes,the following work is carried out:To better utilize the infrared radiance data acquired from imager instruments in a convection-allowing model,this study first established a DA scheme for the assimilation of allsky radiance data from Advanced Baseline Imager(ABI)onboard the Geostationary Operational Environmental Satellite R series(GOES-R)in GSI-En KF system.Various modifications and supplements are made to calibrate ABI observations,improve radiative transfer model and DA system.Using the enhanced DA scheme,all-sky radiance data from ABI are assimilated into a cloud-allowing model(CAM).Assimilation experiments are then carried out for the study of a mesoscale convective system(MCS)case that occurred over the center U.S.plain to make a preliminary test of the DA scheme.Compared with the control experiment that assimilated no observations,it is found that assimilating all-sky ABI BT data can correctly build up observed storms and effectively remove spurious storms in the background through frequent DA,and free forecasts of up to 4hours are improved when verified against observed ABI BTs and radar reflectivity.Further analyses show that over-drying often occurs accompanied by the removal of model spurious clouds when the observation prior of ensemble mean state is used in the ensemble mean state update equation of En SRF,which can cause spurious storm decay in the forecast.The problem is reduced when ensemble mean of observation priors is used instead.Then,more sensitivity experiments with different data assimilation(DA)configurations are conducted using the perfect-model observing system simulation experiments(OSSEs).Several En KF parameters are examined in this study for an optimal configuration for ABI DA.With the best-tuned configuration,impact of assimilating synthetic ABI data on model prognostic variables under clear and cloudy sky conditions are investigated.The combination of a 6 km thinning interval for clear-sky BTs,a 30 km horizontal localization radius for clear and cloudy BTs,a 0.5 RTPP covariance inflation and updating wind variables is found to provide utmost improvement on the analysis and forecast of the targeted severe weather event.Comparison with nature run indicates that model variables are improved by assimilating allsky ABI radiance in both clear and cloudy scenarios.Finally,using the optimized En KF configuration,results of assimilating clear-sky and allsky ABI radiance are further compared to examine the added benefit of assimilating cloudy radiance.It is found that the analysis BTs are improved when only clear-sky radiance is assimilated,while assimilating cloudy radiance gives more accurate analyses because larger adjustment are made to the model variables within cloudy and precipitation regions.Forecast results indicate all-sky DA significantly improves forecast BTs,radar reflectivity and precipitation,compared to a relative limited impact brought by assimilating clear-sky data.