Impact of the 4D-Var Assimilation of Airborne Doppler Radar Data on Numerical Simulations of the Genesis of Typhoon Nuri (2008): Radial velocity versus 3-D wind analysis

Initial conditions have a significant impact on forecasts of tropical cyclone (TC) genesis in numerical weather prediction. In this study, data assimilation experiments were conducted using the Weather Research and Forecasting (WRF) model and its four-dimensional variational data assimilation (4D-Var) system to examine the impacts of radar data assimilation on numerical simulations of the genesis of Typhoon Nuri (2008). Radar observations from the ELDORA airborne radar were assimilated into the WRF model for improving the atmospheric conditions in the pre-genesis phase of Typhoon Nuri during the Office of Naval Research's Tropical Cyclone Structure 2008 (TCS-08) field program. Two ways of assimilating the radar wind data were performed: one assimilated the radar wind analysis retrieved from a 3-dimensional variational method; the other directly assimilated the radar radial velocity. Results show that the 4DVAR assimilation of radar wind observations significantly improves the model simulations of Nuri's genesis. The simulations with the radar data assimilation (both assimilating the radar wind analysis and the radial velocity) produce Nuri's genesis with enhanced storm intensity, while the control experiment (without radar data assimilation) fails to predict Nuri's genesis. In addition, by comparing two assimilation methods, it is found that the assimilation of the radial velocity leads to more improvement in Nuri's intensity forecasts, while the track forecasts are better improved with the assimilation of the radar wind analysis. Further diagnoses indicated that the radar data assimilation results in the enhanced middle-level vortex and moist conditions in pre-Nuri environments thus provide favorable conditions for the development of deep convection.

Additional experiments were also performed to examine the impacts of model resolution on data assimilation results. Results from 4DVAR radar data assimilation were also compared with those from the 3-dimensional variational data assimilation method.