基于时间序列的一线城市空气微粒排放物建模分析与预测  

Time Series-Based Modeling Analysis and Forecasting of Air Particulate Emissions in Tier-1 Cities

作　　者：宋婉宁 覃红霞 吕萌 周影慧 高婉琪 陈佳男 白晓东 

机构地区：[1]大连民族大学理学院,辽宁 大连

出　　处：《应用数学进展》2024年第11期4990-5000,共11页Advances in Applied Mathematics

摘　　要：大气可吸入颗粒物污染已成为危害人类健康的主要因素之一。对城市空气微粒排放物的分析与预测可以为改善空气质量及制定污染防控措施提供科学依据。本文以2014~2018年北京市、上海市和广州市的空气质量指数(Air Quality Index,简称:AQI)、PM2.5浓度和PM10浓度的逐日数据为研究样本。数据经过异常值剔除和插值处理后,采用ARIMA和SARIMA模型对时间序列特征进行分析,并对未来三期空气质量进行预测。结果显示,真实值均位于95%置信区间内,大部分数据的相对精度低于10%,模型拟合度较高。此外,研究表明,北京、上海和广州的AQI、PM2.5浓度及PM10浓度呈下降趋势,短期内空气质量显著提升。Atmospheric particulate matter pollution has become one of the primary factors endangering human health. The analysis and forecasting of urban particulate emissions can provide a scientific basis for improving air quality and formulating pollution control measures. This paper uses daily data on the Air Quality Index (AQI), PM2.5 concentration, and PM10 concentration from 2014 to 2018 in Beijing, Shanghai, and Guangzhou as research samples. After removing outliers and performing interpolation, ARIMA and SARIMA models were applied to analyze the time series characteristics and predict air quality for the next three periods. The results indicate that the actual values fall within the 95% confidence interval, with most data exhibiting a relative accuracy of less than 10%, and the models show a high degree of fit. Additionally, the study reveals a downward trend in AQI, PM2.5 concentration, and PM10 concentration in Beijing, Shanghai, and Guangzhou, with significant short-term improvements in air quality.

关 键 词：PM2.5 PM10 空气质量指数(AQI) 时间序列分析 ARIMA模型 

分 类 号：X51[环境科学与工程—环境工程]