Comparison Of Precipitation Of Tropical Cyclones By Using Different Cloud Microphysical Schemes

With the improvement of computational power,more and more scientists have begun to use numercial model to research and predict the generation and development of tropical cyclones as well as disasters like flooding and gale caused by the landfall of them.There are fourteen cloud microphysical schemes in current WRF model(weather research and forecast system,version 3.5).They can be briefly divided into three parts as single-moment, partially double-moment and double-moment cloud schemes. lin,WSM6,WDM6,Thompson and Morrison are chosen as the representative scheme and would be evaluated.All cases that generated during 2013 to 2014 and classified as typhoon or more intense category while making landfall at the coastline of China have been simulated using T-RAPS(Typhoon rapid assimilation and prediction system) model which is designed and improved by the center of severe weather research and simulation of Chinese Academy of Meteorological Sciences.Four of them(Supertyphoon Utor(1311),Supertyphoon Usagi(1319),strongtyphoon Fitow(1323) and Supertyphoon Rammasun(1409)) are selected as meeting certain requirements and evaluated to some extent(e.g. performance of simulated track,intensity as well as accumulated precipitation).Compared with various kinds of observation data derived and detected by radar,satellite and gauge-based station,the temporal-spatial distribution of hydrometeor and the evolution in different parts of Supertyphoon Rammasun is analysized and the crucial process(function) that describes the source-sink terms of the rain which is directly related to the precipitation are figured out.The impact of diabatic cooling related to the phase change of hydrometeor on the evolution of the intensity of Rammasun during its rapid intensification are examined as well.The results can be concluded as follows.Firstly,rare divergence has been found using selected cloud microphysical schemes,especially when the cases were on the ocean.The difference between the best-track and simulated track data using current cloud schemes are resulted from feedback within the microphysics process and large-scale environmental field.Although the effect of microphysics process are negligible in a short term during the simulation, some unique cases that obvious discrepancy can be found when using different cloud schemes need to be investigated as well.Secondly,almost all the selected scheme failed to simulate the rapid dissipation after the tropical cyclones made landfall and diverged from each other greatly. In terms of the simulated 24-hour accumulated rainfall and the probability distribution of different level of precipitation,Morrison scheme is better than the other ones with high tested score while it showed an overestimation(underestimation) of the strong(light) rainfall rate performed by WDM6 scheme in all four cases.In comparsion of the temporal-spatial distribution of the simulated content of hydrometeor in supertyphoon Rammasun,rare difference of vertical distribution of mixing ratios of solid hydrometeors(cloud ice,snow and graupel) has been figured out by using WSM6 and WDM6 cloud microphysical schemes because both of them are composed of same parameterization procedure which describe the process of the phase change within the development of the tropical cyclone.Compared with the TRMM/2A12 data,on one hand,the height simulated by Morrison scheme where maximum content of cloud ice exist is higher than the others and observation,on the other hand, the content of snow simulated by Thompson scheme is twice as much as others and the content of cloud ice is very little. The cloud top temperature that calculated by using the result of each test with different schemes show good agreement with the distribution of content of hydrometeor because this physical parameter is determined by the cloud top height that relates to the cloud ice and as a result, higher height where cloud ice exist simulated by Morrison scheme leaded to lower cloud top temperature compared with that derived from MODIS satellite.The value of radar reflectivity simulated by Thompson scheme after the landfall of Rammasun tended to be()from the upper to the lower level of troposphere while the higher radar echo simulated by Morrison scheme occurred in the upper level which may related to the mixing ratios, number concentration or the size distribution of the solid phase.Due to the different functions that bewrite the size distribution of all hydrometeors and phase change bewteen the solid and liquid ones in the selected cloud scheme,it is hard to figure out the key physical process that determined the performance of the simulation. WSM6 and WDM6 schemes are the alomost the same cloud microphysical scheme expect that the number concentration of rain is predicted in WDM6 while one is diagnosed in WSM6 and some parameterization methods depict the liquid phase change differ from each other.In the analysis of the generation of the rain, the melting of snow and graupel are the main source terms for rain and rain accretion by the solid species and evaporation are the main sink terms, using different parameterization functions makes the rain mixing ratios that simulated by WDM6 scheme has a peak value in the middle-lower troposphere and a steep decline when it reaches the ground. The mixing ratios of hydrometeors in the inner region of WSM6 scheme is larger than WDM6 scheme, but the mixing ratio of rain that simulated by WDM6 scheme in both the inner and outer region are always higher than WDM6, the influence and the impact due to this cooling rate arose by the melting or evaporation should be further checked.CAMS scheme is coupled in WRF(Version 3.6) model and used to simulate supertyphoon USAGI(1319) and compared with the optimized single-moment WSM6 and double-moment Morrison scheme.It is a mixed-phase double_moment scheme designed and improved by Chinese Academy of Meteorological Sciences.Like previous study, rare difference was found in the simulated track between three cloud scheme and the best-track data while the intensity is on the opposite. It also showed an overestimation(underestimation) of the strong(light) rainfall rate like performed by WDM6 scheme.As the three-dimensional radar reflectivity and cloud top temperature simulated by CAMS scheme is more matched with the observation data than WSM6 and Morrison scheme,we trully believe that CAMS scheme is suitable for the simulation of tropical cyclone.