万盛大气污染特征分析及污染预报模型的建立  

Analysis of Wansheng Air Pollution Characteristics and Establishment of Pollution Prediction Model

作　　者：何佳玥 韩余[2] 陈道劲 贺靖 HE Jia-yue;HAN Yu;CHEN Dao-jin;HE Jing(Chongqing Wansheng Economic Development Zone Natural Disaster Early Warning&Prevention Office,Chongqing 408200,China;Chongqing Meteorological Bureau,Chongqing 401147,China;Ningxia Hui Autonomous Region Shizuishan Meteorological Bureau,Shizuishan,Ningxia 753000,China)

机构地区：[1]重庆市万盛经开区自然灾害预警预防办公室,重庆408200 [2]重庆市气象局,重庆401147 [3]宁夏回族自治区石嘴山市气象局,宁夏石嘴山753000

出　　处：《四川环境》2020年第2期65-73,共9页Sichuan Environment

摘　　要：利用2017年4月1日~2019年3月31日万盛经开区万东北路站点的空气质量日均值监测数据进行分析发现,万盛大气污染具有很强的季节变化特征,冬半年主要污染物为PM 2.5,夏半年主要污染物为O 3,冬半年污染重于夏半年,颗粒物污染重于O 3污染。万盛污染以轻度污染为主,仅有冬季会出现中度以上污染天气,其首要污染物均为PM 2.5。利用多元回归模型和差分自回归移动平均模型(ARIMA)建立了万盛PM 2.5、PM 10与O 3的预报模型。通过对模型得出的预报值与实况值的比较来看,预报与实况的变化趋势基本一致,均可以较好的指示未来AQI的变化趋势。多元回归预报模型中,O 3的预报效果要远好于PM 2.5和PM 10;而ARIMA预报模型三者预报效果接近。总体来说,ARIMA(p,d,q)预报模型对颗粒物污染的预报效果要远好于多元回归预报模型,而O 3则两种模型预报效果接近。Based on the monitoring data of daily mean air quality of Wandong North Road Station in Wansheng economic and Technological Development Zone from April 1,2017 to March 31,2019,it is found that the air pollution of Wansheng has a strong seasonal change feature,PM 2.5 is the main pollutant in the winter half year,O 3 is the main pollutant in the summer half year,the pollution in the winter half year is more serious than that in the summer half year,and the particle pollution is more serious than O 3 pollution.Wansheng pollution is mainly light pollution,only in winter the moderate pollution weather occurs,and the main pollutantis PM 2.5.The prediction models of Wansheng PM 2.5,PM 10 and O 3 are established by using multiple regression model and differential autoregressive moving average model(ARIMA).Through the comparison of the predicted value and the actual value of the model,we can see that the change trend of the predicted value and the actual value are basically the same,which can better indicate the change trend of AQI in the future.In the multiple regression model,the prediction effect of O 3 is much better than that of PM 2.5 and PM 10,while in the ARIMA model,they are close.In general,ARIMA(p,d,q)model is much better than multiple regression model in the prediction of particulate pollution,while in O 3 prediction,the two models are similar.

关 键 词：大气污染 特征分析 预报模型 

分 类 号：X513[环境科学与工程—环境工程] X-1