兰州城区空气污染预报的动力统计模型

The dynamic statistics model of air pollution in Lanzhou proper

从空气污染扩散方程出发 ,经过简化推导 ,得出污染预报动力统计模型的数学形式 .在模型中既考虑了气象条件的作用 ,又考虑了污染排放量和起报日的污染浓度 ,与以往的空气污染预报统计模型相比 ,所依据的物理基础更可信一些 .利用兰州城区 1988～ 1992年 3种主要空气污染物 SO2 ,CO,NOx 浓度资料和同期兰州气象站的地面气象要素、低空气象参数及天气图 ,通过逐步回归 ,确定了主要污染物 SO2 ,CO,NOx 浓度 2 4 h变率的预报方程 .拟合与试报表明 ,本方案在预报技巧上有一定的提高 ,技巧得分可达 0 .0 5～ 0 .89.

By simplifying the equation of air pollutant diffusing,the mathematic form of dynamic statistics model of air pollutant forecast has been worked out.In this model the effect of meteorological element,output and concentration of air pollutant were considered,so it′s physical foundation may be more believable than pure statistics model.By using the three main air pollutant concentrations of SO 2,CO,NO x from Lanzhou Environmental Station,surface and lower layer meteorological element of Lanzhou Meteorological Station and the weather map during 1988～1992,the regression equations of 24 h variation ratio of air pollutant concentrations of SO 2,CO,NO x were established.Tests has showed that the method in this paper has a certain improvement on forecast skill.