Assimilating satellite retrievals and radar observations in a convection permitting model

Assimilating geostationary satellite derived cloud retrievals such as cloud water path (CWP) have been shown to successfully improve the characterization of convection within storm-scale numerical weather prediction (NWP) models. Combining satellite retrievals with radar reflectivity and radial velocity observations also showed promise, but many questions remained un-answered. This research builds upon previous findings, using an improved satellite forward operator and additional radar vs. satellite assimilation tests. Clear vs. cloudy regions are now handled separately and are assigned different observation errors. A correction factor has also been included that adjusts cloud heights in regions of deep convection where the satellite retrieval becomes saturated. The combined satellite – radar assimilation experiments are tested using a Southern Plains severe weather event occurring on 24 May 2011. This event generated several tornado producing supercells in central Oklahoma along an eastward progress dryline. GOES-13 CWP retrievals generated at a 4 km, 15 minute resolution are assimilated along with Oklahoma WSR-88D Doppler radar reflectivity and radial velocity into a 2 km resolution domain. Data are assimilated using an Ensemble Kalman Filter (EnKF) approach through the Data Assimilation Research Testbed (DART) software combined with the WRF-ARW model version 3.4.1. Data are assimilated at various time intervals between 1800 – 2100 UTC with forecasts generated thereafter. Simulated satellite and radar products are compared with observations to assess the overall effectiveness of the assimilation. Several experiments are performed, assimilating various combinations of satellite and radar data to determine which represents the ideal formulation. Time permitting, additional experiments using data from the recent 2013 tornado events in Oklahoma will also be tested using the same model configurations.