Air Quality Forecast Model Based on Model Products of CMAQ in Fuzhou City, China

Abstract

Based on the forecast products of Community Multiscale Air Quality Model (CMAQ), observational data of air pollutants and the conventional meteorological observation data of surface from January 2007 to June 2010 in Fuzhou City, Fujian Province, China, the weather systems influencing on Fuzhou City are divided into 7 weather patterns, including continental high, subtropical high, shear, warm sectors convergence, upper trough, typhoon and tropical convergence. According to the forecast method of combination of dynamic and statistic, using multiple linear stepwise regression method, the forecasting models of daily pollutant concentration under different weather systems are established. The confidence level of the equations achieves 0.000 and so the models are statistically significant. Using the models, the prediction effect of the concentration of different air pollutants in Fuzhou City from July to December, 2010 is tested by back substitution. The forecasting accuracy on contamination index level of PM10 reaches to 71.3%, and the forecasting accuracy on SO2和NO2 both reaches to 100%. The comprehensive score of air quality daily forecast in the city comes to 88.8 points on average.