山东省环境空气质量动力统计预报系统

Statistical Forecast System of Environmental Air Quality in Shandong Province

以山东省17个城市2014年至2015年环境空气质量监测指标和同步的气象观测数据为基础,采用线性回归和神经网络方法建立统计预报模型,构建山东省环境空气质量动力统计预报系统。该系统实现了业务化自动运行,对山东省17城市的6项污染物指标(PM_(2.5)、PM_(10)、NO_2、SO_2、CO日均浓度和O_3日最大8小时滑动平均浓度)和AQI指数进行逐日预报。预报结果能较好的反应各市空气质量的变化趋势,为预报业务提供参考。

Based on the correlation of the 17 cities and its historical environment change of air quality with meteorological factors in Shandong Province,the environmental air quality dynamic statistical forecast system is build by linear regression and neural network method,using meteorological data and air pollution history data. The automatic operation of the business system is accomplished through using observation data of air pollutant concentration and numerical weather prediction data, and provides Daily forecast of 6 pollutant concentration（ daily average concentration of PM2. 5,PM10,NO2,SO2,CO and daily maximum concentration of O38 hours moving average concentration） and AQI index of 17 cities. The prediction can indicate the trend of the air quality change,and provides a reference for the operation forecast.