摘要
PM2.5的精确预测是大气污染评价和治理的关键性工作。本文针对PM2.5浓度变化的时间序列分布特征,结合环境监测站提供的相关数据,应用自回归移动平均模型(ARIMA(p,d,q))预测短期PM2.5的日平均浓度。结果表明:由于PM2.5浓度变化受气象场、排放源、复杂下垫面、理化生过程的耦合等多种因素的影响,不同时段内的变化模式存在巨大差异,因此采用分时段序列预测模型可以提高PM2.5的预测精度;通过将分时段序列模型与灰色GM(1,1)模型和全年时间序列模型的预测结果进行对比,发现该模型预测效果更好。
The accurate forecast of PM2.5is key work of atmospheric pollutant assessment and management.According to the characteristics of time series of concentration chang of PM2.5and with data provided by the environmental monitoring station,this research adopts ARIMA（p,d,q）（Auto-regression Integrated Moving Average）model to forecast the daily average concentration of PM2.5.The results show that there are different patterns in different periods because the daily average concentration of PM2.5 varies greatly with time influenced by meteorological field,emission source,complex underlying surface,physical chemistry reaction and so on.So the paper chooses different models in different periods in order to improve the accuracy of forecast.Compared with the gray model and the annual time series model,the different models in different periods are proved to have better performance.
出处
《安全与环境工程》
CAS
北大核心
2014年第6期125-128,共4页
Safety and Environmental Engineering
基金
国家自然科学基金项目(91324201)
