| 基于机器学习方法研究气象及排放变化对长三角地区主要城市大气污染物的影响 |
| |
| 引用本文： | 付文星,黄琳,丁嘉豪,秦墨梅,于兴娜,谢放尖,胡建林.基于机器学习方法研究气象及排放变化对长三角地区主要城市大气污染物的影响[J].环境科学,2023,44(11):5879-5888. |
| |
| 作者姓名： | 付文星 黄琳 丁嘉豪 秦墨梅 于兴娜 谢放尖 胡建林 |
| |
| 作者单位： | 南京信息工程大学环境科学与工程学院, 江苏省大气环境监测与污染控制高技术研究重点实验室, 江苏省大气环境与装备技术协同创新中心, 南京 210044;南京信息工程大学长望学院, 南京 210044;南京信息工程大学气象灾害预报预警与评估协同创新中心, 中国气象局气溶胶与云降水重点开放实验室, 南京 210044;南京生态环境保护科学研究院, 南京 210093 |
| |
| 基金项目： | 国家自然科学基金项目(42211530082)；;国家级大学生创新创业训练计划项目(202210300070Z)； |
| |
| 摘 要： | 选取2015～2021年长三角地区4个代表性城市污染物浓度，利用机器学习的气象归一化方法解耦气象因素对污染物的影响，量化气象和排放对污染物浓度变化的贡献.结果表明，长三角地区PM2.5、 NO2和SO2排放下降影响贡献较大(57.2%～68.2%、 80.7%～94.6%和81.6%～96.1%),抵消了气象因素带来的不利影响，致使污染物浓度降低.而气象条件对于臭氧日最大8 h(MDA8＿O3)的贡献强于其他污染物(23.5%～42.1%),其中气象因素促进污染物浓度上升(4.7%),排放变化促进污染物浓度下降(-3.2%). NO2和MDA8＿O3在2019～2021年降幅更快，主要原因是2019～2021年排放起到较2015～2018年更强的促进污染物浓度降低作用.PM2.5和SO2在2019～2021年的降幅较2015～2021年整体有所减弱.基于机器学习的气象归一化方法可以解耦气象对污染物的影响，量化排放...
|
| 关 键 词： | 机器学习 气象归一化 大气污染物 长期趋势 随机森林模型 |
| 收稿时间： | 2023/1/18 0:00:00 |
| 修稿时间： | 2023/2/2 0:00:00 |
Elucidating the Impacts of Meteorology and Emission Changes on Concentrations of Major Air Pollutants in Major Cities in the Yangtze River Delta Region Using a Machine Learning De-weather Method |
| |
| FU Wen-xing,HUANG Lin,DING Jia-hao,QIN Mo-mei,YU Xing-n,XIE Fang-jian,HU Jian-lin.Elucidating the Impacts of Meteorology and Emission Changes on Concentrations of Major Air Pollutants in Major Cities in the Yangtze River Delta Region Using a Machine Learning De-weather Method[J].Chinese Journal of Environmental Science,2023,44(11):5879-5888. |
| |
| Authors: | FU Wen-xing HUANG Lin DING Jia-hao QIN Mo-mei YU Xing-n XIE Fang-jian HU Jian-lin |
| |
| Institution: | Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing 210044, China;Changwang School of Honors, Nanjing University of Information Science & Technology, Nanjing 210044, China;Key Laboratory for Aerosol-Cloud-Precipitation of China Meteorological Administration, Collaborative Innovation Center on Forecasting and Evaluation of Meteorological Disasters, Nanjing University of Information Science and Technology, Nanjing 210044, China;Nanjing Municipal Academy of Ecological and Environment Protection Science, Nanjing 210093, China |
| |
| Abstract: | This study applied a de-weather method based on a machine learning technique to quantify the contribution of meteorology and emission changes to air quality from 2015 to 2021 in four cities in the Yangtze River Delta Region. The results showed that the significant reductions in PM2.5, NO2, and SO2 emissions(57.2%-68.2%, 80.7%-94.6%, and 81.6%-96.1%, respectively) offset the adverse effects of meteorological conditions, resulting in lower pollutant concentrations. The meteorological contribution of maximum daily 8-h average O3(MDA8_O3) showed a stronger effect than that of others(23.5%-42.1%), and meteorological factors promoted the increase in MDA8_O3 concentrations(4.7%); however, emission changes overall resulted in a decrease in MDA8_O3 concentrations(-3.2%). NO2 and MDA8_O3 decreased more rapidly from 2019 to 2021, mainly because the emissions played a stronger role in reducing pollutant concentrations than from 2015 to 2018. However, emissions changes had weaker reduction effects on PM2.5 and SO2 from 2019 to 2021 than from 2015 to 2018. De-weather methods could effectively seperate the effects of meteorology and emission changes on pollutant trends, which helps to evaluate the real effects of emission control policies on pollutant concentrations. |
| |
| Keywords: | machine learning de-weather method air pollutant long-term trends random forest model |
|
| 点击此处可从《环境科学》浏览原始摘要信息 |
| 点击此处可从《环境科学》下载免费的PDF全文 |
|
