Thomas Jones, CIMMS / NSSL

Satellite data assimilation within a prototype Warn-on-Forecast system

Kent Knopfmeier, Dustan Wheatley, Patrick Minnis, and Rabindra Palikonda


The goal of the NOAA Warn-on-Forecast (WoF) system is to provide probabilistic short-term (0-3 h) forecast guidance for high impact weather events such as tornadoes, hail, high winds, and flash flooding. The prototype system WoF system employs an ensemble adjustment Kalman filter (EaKF) data assimilation technique on a convection-permitting (3 km) grid. High resolution surface based radar and geostationary satellite products are assimilated over a regional domain at 15 minute intervals for the duration of a particular severe weather event. Versions of this prototype system have been run in realtime during the 2015 and 2016 Hazardous Weather Testbed (HWT) located in Norman, Oklahoma. This research describes the impact of assimilating satellite observations into the WoF system and how it can improve forecasts relative to only assimilating radar data. The primary satellite product assimilated is cloud water path (CWP), which represents the total column liquid and frozen hydrometeor content of a cloud. CWP and corresponding cloud height and phase information are retrieved from GOES Imager observations. Assimilating CWP retrievals produces a more accurate analysis of the cloud properties in the model analysis that has the downstream impact of improving the analyzed thermodynamic conditions. These changes can impact whether or not an analyzed storm maintains its intensity or weakens as it moves into an environment that is or is not supportive of convection. The other major benefit of assimilating CWP is that it allows for a quicker spin-up of convection within the model. Since the satellite is sensitive to non-precipitating clouds, CWP retrievals can act to increase hydrometeor content in developing cumulous prior to the onset of precipitation and its measurement by radar data. Examples from several experiments showing the impact of assimilating CWP will be provided.

File available here