| Description |
Forecasting tropical cyclone (TC) genesis is a challenging problem. This dissertation attempts to understand the following questions through studying the genesis of Typhoon Nuri (2008) with numerical simulations and data assimilation: 1) What are the atmospheric conditions and processes that contribute to Nuri's genesis and early rapid intensification? 2) To what extent can data assimilation improve the forecasts of Nuri's genesis? To address the first question, numerical simulations of Nuri's genesis are conducted using an advanced research version of the Weather Research and Forecasting (WRF) model. First, initial and boundary conditions derived from two global analyses are found to lead to remarkably different simulations of Nuri's genesis in developing and nondeveloping cases. It is also found that the convective development into the pre-Nuri core region is a critical process for Nuri's genesis. A strong midlevel vortex and a moist environment provide the favorable conditions for the convective development. Induced by the persistent deep convection, diabatic heating at upper levels is produced from latent heat release. This substantial warming at upper levels results in the drop in Nuri's minimum central sea level pressure. Next, the sensitivity of numerical simulations of Nuri's genesis to the model horizontal resolution is examined. Results show that the simulation at a coarse-resolution (e.g., 12 km) better predicts Nuri's rapid intensification than that at a higher resolution (e.g., 4 km). Specifically, the simulation at the coarser resolution produces strong convective bursts and diabatic heating in the inner core region and also stronger warming in the upper atmosphere, thus leading to a lower minimum sea level pressure (MSLP). Further experiments suggest that an appropriate microphysics scheme (e.g., the twomoment Morrison scheme) and a later initialization time (after Nuri's early development) could help the high-resolution simulation better capture Nuri's rapid intensification. Finally, numerical experiments are conducted to examine the impact of radar data assimilation on numerical simulations of Nuri's genesis using a four-dimensional variational data assimilation (4D-VAR) method. The radar data assimilation results in significant improvements in the numerical simulation of Nuri's genesis. Several configurations of data assimilation are evaluated. Specifically, assimilation of radial velocity leads to more improvement in intensity forecasts, whereas track forecasts are better simulated by the assimilation of radar-retrieved wind components. Improved analysis and forecasts are obtained when both radial velocity and retrieved winds are assimilated. In addition, 4D-VAR performs better than three-dimensional variational data assimilation (3D-VAR) in radar data assimilation. The positive impact of radar data assimilation can be attributed to the improved simulations of convective evolution and the enhanced midlevel vortex and moisture conditions. |
