中国环境科学
 
 
中国环境科学  2023, Vol. 43 Issue (12): 6225-6234    DOI:
大气污染与控制 最新目录| 下期目录| 过刊浏览| 高级检索 |
空气质量模拟与观测机器学习NO2浓度预报
黄泳熙1, 朱云1, 谢阳红2, 李海贤2, 张志诚1, 黎杰1, 李金盈1, 袁颖枝1
1. 华南理工大学环境与能源学院, 广东省大气环境与污染控制重点实验室, 广东 广州 510006;
2. 佛山市生态环境局顺德分局, 广东 佛山 528300
Forecast of NO2 concentrations based on coupled air quality model simulations and monitoring data using machine learning method
HUANG Yong-xi1, ZHU Yun1, XIE Yang-hong2, LI Hai-xian2, ZHANG Zhi-cheng1, LI Jie1, LI Jin-ying1, YUAN Ying-zhi1
1. Guangdong Provincial Key Laboratory of Atmospheric Environment and Pollution Control, College of Environment and Energy, South China University of Technology, Guangzhou 510006, China;
2. Foshan Ecology and Environment Bureau, Shunde Branch, Foshan 528300, China

全文: PDF (1382 KB)   HTML (1 KB) 
输出: BibTeX | EndNote (RIS)      
摘要 在空气质量模拟预报数据基础上,采用套索算法(Lasso)将前馈神经网络(FNN)与基于污染物浓度及气象实时观测值搭建的长短期记忆网络(LSTM)组合,形成了模拟与观测机器学习(SOML)预报模型,开展了佛山市顺德区NO2未来3d 10个镇街空气质量监测点位逐日浓度预报.结果显示:SOML3d的准确性均优于WRF-CMAQ及其它单一模型,其中第一天SOML平均绝对误差(MAE)为4.99 μg/m3,改进幅度达66.18%;SOML不同季节适用性均较强,四季预报效果均较WRF-CMAQ明显提升(MAE分别降低42.18%、42.89%、61.04%、50.91%),其中秋冬季改善幅度更好;相比WRF-CMAQ,SOML预报结果能较好反映顺德区内各站点NO2浓度实际空间分布和数值水平,有效提升了浓度预报精准度.
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
黄泳熙
朱云
谢阳红
李海贤
张志诚
黎杰
李金盈
袁颖枝
关键词 NO2浓度预报机器学习预报模型WRF-CMAQ模型空气质量监测    
Abstract:In this study, built upon the WRF-CMAQ air quality model simulations, a novel machine learning method based on simulations and observations (SOML) that integrating feedforward neural network (FNN) and long short-term Memory network (LSTM) through the Lasso method was developed for forecasting NO2 concentrations, where LSTM was derived based on real-time pollutant and meteorological data. This innovative method was then applied to forecast the NO2 concentrations for three consecutive days for ten air quality monitoring stations in Shunde, Foshan to evaluate the model performance. Our results show that: Compared to WRF-CMAQ and other individual models, SOML gave higher accuracy in the three-day forecast of NO2 concentrations, with the mean absolute error (MAE) of first day at 4.99μg/m3, decreasing up to 66.18%; The accuracy of SOML predictions has significantly improved compared with that of WRF-CMAQ, indicating SOML’s suitable applicability to all seasons (MAE decreased by 42.18%, 42.89%, 61.04% and 50.91%, respectively), particularly in autumn and winter; and Compared with WRF-CMAQ, SOML appears to provide better forecasting accuracy of the spatial distribution as well as the NO2 concentration levels at each station in Shunde.
Key wordsNO2 concentration forecast    machine learning    forecast model    WRF-CMAQ model    air quality monitoring   
收稿日期: 2023-04-23     
PACS: X511  
基金资助:高端外国专家引进计划项目(G2023163014L)
通讯作者: 朱云,教授,zhuyun@scut.edu.cn     E-mail: zhuyun@scut.edu.cn
作者简介: 黄泳熙(1999-),女,广东汕头人,华南理工大学硕士研究生,主要从事空气质量模拟与预报研究.1663795613@qq.com.
引用本文:   
黄泳熙, 朱云, 谢阳红, 李海贤, 张志诚, 黎杰, 李金盈, 袁颖枝. 空气质量模拟与观测机器学习NO2浓度预报[J]. 中国环境科学, 2023, 43(12): 6225-6234. HUANG Yong-xi, ZHU Yun, XIE Yang-hong, LI Hai-xian, ZHANG Zhi-cheng, LI Jie, LI Jin-ying, YUAN Ying-zhi. Forecast of NO2 concentrations based on coupled air quality model simulations and monitoring data using machine learning method. CHINA ENVIRONMENTAL SCIENCECE, 2023, 43(12): 6225-6234.
链接本文:  
http://www.zghjkx.com.cn/CN/      或     http://www.zghjkx.com.cn/CN/Y2023/V43/I12/6225
中国环境科学
友情链接: 中国科学技术协会 中国环境科学学会 中国知网 万方数据
版权所有 © 2010 《中国环境科学》编辑部  地址:北京市海淀区红联南村54号(100082)  电话、传真:(010)62215145  
E-mail:zghjkx1981@126.com;zghjkx1981@chinacses.org;zghjkx1981@188.com
本系统由北京玛格泰克科技发展有限公司设计开发  技术支持:support@magtech.com.cn