基于机器学习的珠三角秋季臭氧浓度预测  

Prediction of Autumn Ozone Concentration in the Pearl River Delta Based on Machine Learning

作　　者：陈镇 刘润 罗征 薛鑫 汪瑶 赵志军 CHEN Zhen;LIU Run;LUO Zheng;XUE Xin;WANG Yao;ZHAO Zhi-jun(Institute for Environmental and Climate Research,Jinan University,Guangzhou 511443,China;Guangdong-Hong Kong-Macao Joint Laboratory of Collaborative Innovation for Environmental Quality,Guangzhou 511443,China;School of Information Science and Technology,Fudan University,Shanghai 200438,China)

机构地区：[1]暨南大学环境与气候研究院,广州511443 [2]粤港澳环境质量协同创新联合实验室,广州511443 [3]复旦大学信息与工程学院,上海200438

出　　处：《环境科学》2024年第1期1-7,共7页Environmental Science

基　　金：广州市科技项目(202002020065);国家自然科学基金项目(91644222,92044302)。

摘　　要：基于2015~2022年珠三角地区的臭氧(O_(3))日最大8 h浓度平均值[MDA8-O_(3),ρ(O_(3)-8h)]的观测数据和气象再分析数据,运用支持向量回归(SVR)、随机森林(RF)、多层感知机(MLP)和轻量级梯度提升机(LG)这4种机器学习方法,建立MDA8-O_(3)预测模型.结果表明,对于全年MDA8-O_(3)预测而言,SVR模型的效果最好,决定系数(R^(2))达0.86,均方根误差(RMSE)和平均绝对误差(MAE)分别为16.3μg·m^(-3)和12.3μg·m^(-3);对于秋季MDA8-O_(3)预测而言,SVR模型的效果依然略优于LG和MLP,其R2、RMSE和MAE分别为0.88、19.8μg·m^(-3)和16.1μg·m^(-3),RF模型在秋季的预测效果最差.采用全年数据构建的模型对秋季MDA8-O_(3)的预测效果比仅采用秋季数据构建的模型效果好,R2相差0.08~0.14.Based on the observation data of the daily maximum 8-hour ozone(O_(3))average concentration[MDA8-O_(3),ρ(O_(3)-8h)]and meteorological reanalysis data in the Pearl River Delta Region from 2015 to 2022,four machine learning methods,i.e.,support vector machine regression(SVR),random forest(RF),multi-layer perceptron(MLP),and lightweight gradient boosting machine(LG)were used to establish MDA8-O_(3) prediction models.The results showed that the SVR model had the best prediction performance on MDA8-O_(3) during the whole year,the coefficient of determination(R^(2))reached 0.86,and the root mean square error(RMSE)and mean absolute error(MAE)were 16.3μg·m^(-3) and 12.3μg·m^(-3),respectively.The prediction performance of the SVR model in autumn was still slightly better than that of LG and MLP,with R2,RMSE,and MAE values of 0.88,19.8μg·m^(-3),and 16.1μg·m^(-3),respectively.The RF model performed the worst in the autumn prediction.In addition,the models trained by data from the whole year had better prediction ability on autumn MDA8-O_(3) than that of those only trained by autumn data,and the R2 differed 0.08-0.14.

关 键 词：珠三角(PRD) 臭氧(O_(3)) 日最大8 h浓度平均值(MDA8-O_(3)) 机器学习 预测 

分 类 号：X515[环境科学与工程—环境工程]