非静力中尺度高分辨率模式模拟中的垂直坐标影响研究

近几年来,随着高性能计算机技术的高速发展,甚高分辨率的中尺度数值预报模式业务应用已成为可能,与之相关的一系列模式技术的新问题也随之提了出来,垂直坐标系的影响就是其中之一。文中借助于美国新一代数值预报模式WRF(WeatherReseachandForecast),比较了非静力中尺度模式高分辨率模拟应用的垂直坐标影响问题。研究表明,当选用几何高度(z)和气压(p)来构造地形追随坐标时,低层两坐标引起的误差基本一样,中高层高度地形追随坐标引起的误差小于气压地形追随坐标;而且分辨率越高差异愈大。高分辨率模拟结果也表明,这种差异趋势是存在的;此外,本文对天气过程的预报要素场进行了相关的分析。

IMPACT STUDY OF THE VERTICAL COORDINATE ON A NON-HYDROSTATIC MESO-SCALE HIGH-RESOLUTION MODEL

With the development of super computer techniques and the progress of correlative subjects,it expedites the steps of development on the non-hydrostatic meso-scale high-resolution models in home and overseas,such as GRAPES(Global/Regional Assimilation and Prediction System)in China Meteorological Administration,WRF(Weather Research and Prediction)in the United States,LM(Local Model)developed by German,Italy and Switzerland etc.and ALADIN model developed by France,which are all non-hydrostatic meso-scale high-resolu-tion models.High-resolution in non-hydrostatic meso-scale numerical models is the requirement of simulating the complicated and elaborate possible process of atmosphere,and it is also a research tool and approach probing into many uncertain mechanism of atmospheric process.However,some of techniques,such as vertical coordinate system impact the use of meso-scale non-hydrostatic numerical weather.Based on WRF(Weather Research and Forecast)model,the influence of vertical coordinates on the non-hydrostatic meso-scale high-resolution model is compared.It showed that the error of various coordinates in lower levels is almost the same when the geometry height(z)and the pressure(p)to set up a terrain-following coordinate;but the error of the height terrain-following coordinate in higher levels is smaller than that of the pressure terrain-following coordinate.The higher is the resolution,the bigger is the error.The results of the high-resolution simulation exhibited that the trend of the difference in the two coordinates existed.In addition,the correlative coefficient and standard error are also analyzed by comparison between the forecast fields and the corresponding analysis fields.