Cooperative Institute for Research in the Atmosphere

The simultaneous assimilation of underused radar and satellite observations to improve convection forecasts

Host: Dr. Keenan Eure

Presenter:

Location:

Accurate forecasts of the initiation and evolution of deep, moist convection in convection-allowing models (CAMs) are both a priority and a challenge of the numerical weather prediction and convective-scale data assimilation communities. Underused observations from dual-polarization weather radars and all-sky (clear and cloudy affected radiances) satellites have the potential to improve the forecasts of deep convection in CAM ensembles. Skillful convection initiation (CI) forecasts are dependent on accurate forecasts of the planetary boundary layer (PBL). There are many processes within the PBL that contribute to CI, including moisture, lift, and instability. Radars provide clear-air radial velocities within the PBL as well as novel PBL depth observations, obtained from quasi-vertical profiles of differential reflectivity (Z_DR). GOES-16 infrared brightness temperatures provide information on cloud structures and cover. The first part of this talk explores the value of these observations when assimilated jointly and separately using a 40-member Ensemble Kalman Filter (EnKF) for a case study. In addition to the PBL, the internal structures of convection are difficult to model, which can be important for storm mode, intensity, and longevity. One distinct dual-polarization signature in intense convection is the Z_DR column, which is a vertical protrusion of positive Z_DR values above the environmental melting level. These are significant for characterizing storm updrafts, which can dictate storm mode and severe hazards. For the second portion of the talk, the direct assimilation of Z_DR is assessed jointly and separately with all-sky infrared brightness temperatures in another case study. From both sets of experiments, some results may suggest radar alone provides modest benefits relative to satellite data assimilation alone; however, simultaneous radar and satellite data assimilation in these cases provide the most promising results.