Assimilation of Non-Local Radio Occultation Measurements and its Impact on WRF Model Prediction of Hurricane Earl in 2010

An algorithm to assimilate non-local spaceborne GPS Radio Occultation (GPSRO) observations, in the Three-Dimensional Variational (3DVAR) Data Assimilation (DA) system of the Weather Research and Forecasting (WRF) model has been previously developed and described in Chen et al. 2009. The study showed that the assimilation of path-integrated refractivity, i.e. excess phase, improved the prediction of track and precipitation of Typhoon Shanshan (2006). In the present study, a new option has been added to this algorithm to increase the number of data points assimilated per occultation profile. It uses the change in horizontal position of perigee points within one profile to allow an increase in vertical resolution during data assimilation under the assumption of non-correlated observations.

Four numerical experiments were conducted to assess the impact of assimilation of excess phase observations on simulations of Hurricane Earl 2010. GTS experiment has assimilation of conventional surface and upper-level sounding observations from the Global Telecommunication System, representing a reference model prediction of the hurricane. The other three experiments have assimilation of both conventional and space-borne GPSRO observations: G+GC has RO observations assimilated as column data, G+G has excess phase as drifting data but assimilating only one point per model level, and G+GD has drifting data with increased vertical resolution.

As expected, the increase in the number of assimilated GPSRO excess phase data points produced larger increments, i.e. a larger impact of these observations on the WRF simulations. Both higher vertical resolution and consideration of horizontal drifting during data assimilation changed significantly the positions and magnitude of maximum impact zones of GPSRO data, mainly in the South-North direction due to the sampling geometry of the profiles, affecting pressure, water vapor, temperature and wind fields at initial forecasting time. Forecast results indicate that the GTS conventional observations mainly contributed in improving the prediction of Earl's track, from a better capture of the large-scale pressure field. GPSRO observations mainly improved the initial moisture field, and improved the prediction of intensity. Validation of the model results was based on satellite total precipitable water observations, ECMWF-Interim reanalysis, and dropsondes collected during the Genesis and Rapid Intensification Processes (GRIP) field campaign.