Impact of GPS RO refractivity or excess phase data assimilation on WRF simulation of Hurricane Helene
Xue Meng Chen, University of California, Davis, Davis, CA; and S. H. Chen, S. Y. Chen, and C. Huang
Assimilation of refractivity profiles, a sub-product of the Doppler excess phase measured by spaceborne Global Positioning System (GPS) Radio Occultation (RO) system, has proven to have a positive effect on mesoscale numerical forecasts of hurricanes. In this study, four experiments were conducted to discern the impact of GPS RO data assimilation on the prediction of Hurricane Helene using the Weather Research and Forecasting model. First, a 4-day model simulation, initialized by the National Centers for Environmental Prediction's final analysis, was compared to a simulation initialized by 12 hours of data assimilation of conventional observations, including surface and radiosonde measurements. Second, atmospheric profiles of refractivity under moist conditions, available from the Constellation Observing System for Meteorology Ionosphere and Climate (COSMIC) project website, were added to assimilated observations. Refractivity profiles are treated as column measurements by the data assimilation system, when in fact the perigee points change position with height (i.e., potentially crossing model grid points horizontally). The error due to this assumption was studied in a third experiment. A fourth experiment studied the difference between assimilating refractivity and the line-integrated excess phase within the model domain. Thinning of observations in the vertical was applied to all experiments.
Results show that Helene intensified too quickly in the WRF simulation without data assimilation. Assimilation of surface and sounding observations delayed the development and intensification of Helene by providing a more accurate initial environment. The addition of refractivity observations had a positive impact, mostly on the prediction of intensity, but also on the hurricane's track during the first 60 hours. All simulations overestimated the hurricane's intensity beyond 3 days simulation, an error often encountered in numerical model forecasts of weak hurricanes.
Poster Session 1, Poster session I
Monday, 1 August 2011, 2:30 PM-4:00 PM, Marquis Salon 3
Previous paper Next paper