| Summary: | 碩士 === 國立中央大學 === 大氣物理研究所 === 100 === The NCAR Variational Doppler Radar Analysis System (VDRAS) is a system which uses the 4DVAR technique to assimilate the radar reflectivity and radial wind observations, and is capable of providing the three-dimensional kinematic and thermodynamic fields within a weather system. Since VDRAS is formulated on a Cartesian coordinate, its application to Taiwan where the topography is complicated is expected to be limited. Previous studies in which the analysis fields from VDRAS were merged with WRF showed encouraging results in improving the model Quantitative Precipitation Forecast (QPF). The purpose of this study is to test the sensitivity of the cycling configuration for VDRAS, and find a robust way of combining VDRAS and WRF.
A real case of Mei-Yu front occurred on 14 June 2008 during Southwest Monsoon Experiment (SoWMEX) IOP8 is selected. In the first set of experiment, three tests are designed to examine the sensitivity of the analysis and forecast with respect to VDRAS cycling configuration containing 1-3 cycles in assimilation processes, respectively. The results of the principal kinematic and thermodynamic features reveal that VDRAS with two cycles is better than the other designs. In the second set of experiment, the VDRAS analysis fields are merged with WRF model. Assuming that the synoptic scale influence can be neglected within a period of three to four hours, it is found that using a single domain instead of nested domains can effectively remove the noises generated along the domain boundaries. Two different ways are used to merge VDRAS with WRF, and they are “direct replacement” and “WRF 3DVAR”, respectively.
It is found that when the VDRAS analysis fields generated by 1-3 cycles are directly merged with WRF model, the two-cycle VDRAS analysis field produces the best results. The impact of using VDRAS outputs to improve the QPF can last for about four hours. The accuracy of the predicted 4-hour accumulated rainfall after merging VDRAS and WRF turns out to be significantly higher than that generated by using VDRAS or WRF alone. This can be attributed to the assimilation of meso- and convective scale information, embedded in the radar data into the VDRAS, and to a better treatment of the topographic effects by the WRF model simulation. However, the experiments of using “WRF 3DVAR” to merge VDRAS analysis fields with WRF are not successful, and this method needs a longer time to spin up.
This research suggests an effective way of using VDRAS to forecast rainfall under Taiwan’s mountainous situation.
|
|---|
