Assimilating radar and satellite data in an operational environment for Warn-on-Forecast

Previous research has studied the impacts of assimilating either radar reflectivity and radial velocity or satellite cloud property retrievals in convection permitting numerical weather prediction (NWP) models, but assimilating both together is much less well understood. To explore this issue, a NWP system has been designed that assimilates both data sets using an ensemble adjustment Kalman filter (EAKF) approach on a 3 km grid. The system was designed to operate in a real time environment assimilating data as it comes in at 15 min intervals for a several hour period during the afternoon and evening. This system was run for several events during the spring of 2013 and 2014 comprising multiple storm modes and environments. Details of this system are provided in a companion abstract. This research discusses the impacts of assimilating satellite cloud water path (CWP) retrieved from the operational GOES imagers for these events. The variety of events allows for a more complete analysis of the advantages and disadvantages of assimilating CWP than possible from a single case study. To determine the effectiveness of assimilating CWP, simulated CWP from model forecasts will be compared with observations for experiments with and without CWP assimilated. Further verification will occur by comparing model surface variables such as temperature, humidity, and shortwave flux with observations where available. The goal will be to show where satellite data improve the analyses and forecasts and determine where improvement in the assimilation methods are required to limit potential failure modes when combined with radar data assimilation.