So <sub> 2 </ Sub> Concentration Of Numerical Simulation And Prediction Experiments,

In order to explore a wind model that is suitable for air pollution numerical model, we put in MM5 as coarse-mesh and put in wind model of Peilke and Mass as fine-mesh model. So it is used to forecast SO2 concentration of Nanjing in first ten days of January, as so as April, July, and October in 2000 , based on comparing to actual data of observation of Nanjing we can get to some results. The results show that: output of four wind models is agree with actual data of observation of Nanjing, put in MM5 can increase the result of pollution forcast, but needs further testing and verify. A method that output of MM5 is used as input of mass model is a suitable method. SO2 concentration of Nanjing in 2000 is no great change following season. Because of pollution source, high SO2 concentration area is in Dachang and Yuhua location.