| dc.description.abstract | The topography of Taiwan is complex. Taiwan is surrounded by sea, where no observation data can be found; besides, tall mountains make the difficulty higher in weather forecast. Thus, radar plays a crucial role in the observation. The Variational Doppler Radar Analysis System (VDRAS) developed by National Center for Atmospheric Research (NCAR) is a system that uses the 4DVAR technique to assimilate the radar reflectivity and radial wind observations, which is capable to inverse the three-dimensional kinematic and thermodynamic fields within a weather system. Then, the forecast comes out. Previous studies in which the analysis fields from VDRAS were merged with MRF showed prominent improvement in results. The purpose of this study is to continue the model of combing VDRAS and WRF. But a few problems arise in the process of the improving. In fact, problems like how to put the pressure field derived from VDRAS into WRF, how to eliminate wrong reflectivity, and how to improve the weakening of radar reflectivity are waiting to be fixed. The study is expected to promote the short-term of Quantitative Precipitation Nowcasting.
A real case of Mei-Yu front occurred on 14 June 2008 during Southwest Monsoon Experiment (SoWMEX) IOP8 is selected. In the first part of experiment, the relative humidity in the background field of the VDRAS before assimilation would be adjusted, which is done to avoid wrong reflectivity from boundary of model and make reflectivity not affected. The second part is to correct the VDRAS analysis field merged with WRF model. Previous studies do not update the WRF pressure field. The pressure-distributed dynamic structure can only be affected by other variable fields, so the results come out slowly. Since pressure is the diagnostic variables in the WRF model, this study corrects the formula, adding a updated pressure filed to the WRF model. On the other hand, two models merged will produce the problem of the weakening reflectivity, so relative humidity of VDRAS is adjusted when they are merged. The promoted experiment turns out to adjust back the domain where the reflectivity is weakened, and suggests an effective way of Quantitative Precipitation Nowcasting.
| en_US |