618 The Impact of GPSRO Observations on HWRF Forecast Accuracy for Hurrican Gonzalo

Tuesday, 9 January 2018
Exhibit Hall 3 (ACC) (Austin, Texas)
Michael D. McAtee, The Aerospace Corporation, Offutt AFB, NE; and S. Diener

Handout (579.1 kB)

The community version of the Hurricane Weather Research and Forecasting (HWRF) model system was used to perform an observation system experiment for Hurricane Gonzalo. Forecasts were initialized on October 14, 2014 12 UTC and were made using the operational configuration of the system. In this configuration, the HWRF forecast model component is run employing a stationary parent and two moving nest domains with approximately 18 km, 6km and 2km grid spacing respectively. Model initialization is accomplished using both a vortex improvement procedure and through the assimilation of observations using the Gridpoint Statistical Interpolation (GSI) system’s three-dimensional (3D) hybrid ensemble-variational (VAR) data assimilation technique. Data assimilation is performed only on the middle and inner nest domains. The HWRF forecast model component is coupled with the Message Passaging Interface (MPI) parallel version of the Princeton Ocean Model (POM) for Tropical Cyclones (MPIPOM-TC). A control run was accomplished where Global Positioning System Radio Occultation (GPSRO) observations of bending angle were assimilated on the middle nest and a 5-day forecast was run. The control run was then compared to a run where the GPSRO observations were not assimilated. Verification of the two runs will include comparison of forecast cloud features to satellite imagery.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner