用中尺度模式输出污染气象的拉格朗日湍流统计量

Output of the Lagrangian Statistics of Turbulence in Air Pollution Meteorology by Using Mesoscale Weather Prediction Model

大气对污染物的稀释和扩散能力随气象条件的改变而发生巨大的变化,而各种湍流特征参数能反映出大气对污染物的扩散能力,因而计算这些参数对污染物浓度的预报有很大的帮助.以往,各种湍流特征参数都是通过局地测量获得的,而要获得较大尺度的扩散背景场,单靠单点测量是做不到的.试图利用MM5模式中的中尺度模式输出量来计算一些湍流特征参数,这些参数可以代表网格上的平均值.选择了MM5中的Mellor＿Yamadalevel2.5的边界层参数化方案,用输出的q2(湍流动能的两倍),再通过解一个代数方程组,求得另外的各种二阶湍流矩量u2、v2、w2.然后利用这些二阶矩量来计算拉格朗日积分时间尺度、自相关系数,尝试用中尺度模式来输出这些量.经计算结果的分析表明,模式输出的这些拉格朗日湍流统计量是合理的,因而利用这种方法来计算拉格朗日湍流统计量是可行的,进一步,这种方法可应用于空气污染扩散的预报中.

The capability of diluting of the air varies with weather conditions, and various kinds of turbulence characteristic parameters reflect the dilution capability of the air. So, it will do a great favor in forecasting the concentration of the polluting matters to calculate these turbulence parameters. The traditional way of collecting these parameters is using local observing methods, yet it's hard to get large scale diffusion fields through single point observation. This paper proposes an approach of making use of mesoscale outputs from MM5 to compute some turbulence characteristic parameters. We have used q2 (which is twice as much as the turbulence energy) from the Mellor_Yamada level 2.5 boundary layer parameterization scheme in MM5 to derive other kinds of turbulence second order moments (u2、v2、w2) by the way of solving a set of algebraic equations, (and σi=ui2, so by using the output from the models, we can get σi, the diffusion parameters, which play an important role in air pollution meteorology, for they reflect the diffusion ability of air pollutes). We use these outputs to calculate the atmospheric diffusion parameters — Lagrangian integral time scale、correlation coefficient, making attempt to get these quantities from mesoscale forecasting model. The simulated results show that the outputs of Lagrangian statistics of turbulence are quite reasonable, thus it is feasible to calculate the diffusion parameters by this method, and further it can be used to predict air pollution. These methods extend the applications of mesoscale weather model into the air pollution meteorology, making it possible to acquire the air diffusion character at different time and different grid points in condition of lacking the suitable planetary boundary layer observation data. Moreover, T\-L and R\-L are mainly used in stochastic diffusion models, like Monte_Carlo methods, so the methods used in this paper can be applied into the field of stochastic diffusion simulation. 