主办单位:中国气象局沈阳大气环境研究所
国际刊号:ISSN 1673-503X
国内刊号:CN 21-1531/P

气象与环境学报 ›› 2020, Vol. 36 ›› Issue (2): 98-104.doi: 10.3969/j.issn.1673-503X.2020.02.013

• 快报 • 上一篇    下一篇

应用机器学习算法的成都市冬季空气污染预报研究

孙苏琪1(),王式功1,2,*(),罗彬3,杜云松4,张巍4   

  1. 1. 成都信息工程大学大气科学学院/高原大气与环境四川省重点实验室, 四川 成都 610225
    2. 贵州省遵义院士工作 中心 气候环境与医疗康养重点实验室, 贵州 遵义 563000
    3. 四川省环境政策研究与规划院, 四川 成都 610041
    4. 四川省生态环境监测总站, 四川 成都 610091
  • 收稿日期:2019-03-25 出版日期:2020-04-30 发布日期:2020-03-03
  • 通讯作者: 王式功 E-mail:528121551@qq.com;wangsg@cuit.edu.cn
  • 作者简介:孙苏琪,女, 1994年生,在读硕士研究生,主要从事环境气象研究, E-mail:528121551@qq.com
  • 基金资助:
    国家自然科学基金重大研究计划重点支持项目(91644226);国家自然科学基金重大研究计划面上项目(41775147);四川省重大科技专项(2018SZDZX0023);国家重点研发计划“全球变化及应对”重点专项(2016YFA0602004)

Air pollution forecast in winter based on machine learning method in Chengdu

Su-qi SUN1(),Shi-gong WANG1,2,*(),Bin LUO3,Yun-song DU4,Wei ZHANG4   

  1. 1. Plateau Atmosphere and Environment Key Laboratory of Sichuan Province/College of Atmospheric Sciences, Chengdu University of Information Technology, Chengdu 610225, China
    2. Zunyi Academician Center, Chinese Academy of Sciences & Chinese Academy of Engineering, Zunyi 563000, China
    3. Sichuan Province Environmental Policy Research and Planning Institute, Chengdu 610041, China
    4. Sichuan Environmental Monitoring Center, Chengdu 610091, China
  • Received:2019-03-25 Online:2020-04-30 Published:2020-03-03
  • Contact: Shi-gong WANG E-mail:528121551@qq.com;wangsg@cuit.edu.cn

摘要:

利用2014年3月至2017年2月成都市8个环境监测站的PM2.5、PM10、SO2、NO2、CO、O3共6种污染物质量浓度资料以及T639全球中期数值预报模式产品,采用两种机器学习算法—递归特征消除法(Recursive feature elimination,RFE)和随机森林方法,构建了成都市冬季5种(O3除外,因为其冬季污染较轻)污染物浓度的预报模型,并对模型的预报效果进行了评价。结果表明:基于RFE模型的5种污染物预报值与实测值的均方根误差值分别为47.58 μg·m-3、72.10 μg·m-3、8.87μ·m-3、0.59 mg·m-3、19.84 μg·m-3;基于随机森林模型的5种污染物预报值与实测值均方根误差值分别为23.94 μg·m-3、20.98 μg·m-3、2.40 μg·m-3、0.16 mg·m-3、8.09 μg·m-3,随机森林模型对各污染物浓度的预报效果均优于RFE模型,说明该预报方法性能良好,可为成都市冬季空气质量业务化预报提供技术支持和防控依据。

关键词: 空气污染预报, 递归特征消除法, 随机森林方法

Abstract:

In this paper, based on the PM2.5, PM10, SO2, NO2, CO, O3 pollutant concentration data from 8 environmental monitoring stations of Chengdu and the T639 global medium-term numerical forecast model products from March of 2014 to February of 2017, the forecast model for the five of the above-mentioned six pollutant except O3 in winter in Chengdu was built by using the recursive feature elimination (RFE) and the random forest method which are superior to the traditional statistical method, and its forecasting performance was assessed.The results show that the mean squared error (MSE) of the values of five pollutants forecasted by the RFE model are 47.58 μg·m-3, 72.10 μg·m-3, 8.87 μ·m-3, 0.59 mg·m-3, 19.84 μg·m-3, and those by the random forest model are 23.94 μg·m-3, 20.98 μg·m-3, 2.40 μg·m-3, 0.16 mg·m-3, 8.09 μg·m-3, which proves that the performance of the random forest model is better than that of the RFE model in the pollutant concentration forecast, indicating that the random forest method has a good performance and can provide the technical support for the air quality forecast business and the basis for the air pollution prevention and control in winter in Chengdu.

Key words: Air pollution forecast, Recursive feature elimination, Random forecast method

中图分类号: