摘要 为提高西安市ρ(PM2.5)及ρ(O3)预报准确率,更好地服务西安市预报预警工作,以CAMx模式预报结果为基础,结合中尺度WRF气象预报数据、ρ(PM2.5)及ρ(O3)观测数据,基于多元线性回归、岭回归、lasso回归、决策树、随机森林以及支持向量机6种机器学习优化模型,对西安市2019年PM2.5及O3模拟结果进行优化.结果表明:①CAMx模式对污染物的预报存在偏差,优化模型明显修正了CAMx模式的系统性偏差,提高了预报精度.②ρ(PM2.5)及ρ(O3)的均方根误差(RMSE)由174.00、37.11μg m3分别降至34.36~39.37、24.77~28.82μg m3,相关性系数(R)由0.63、0.78分别提至0.70~0.78、0.83~0.88.③不同模型对模拟值的订正优势不同,随机森林对PM2.5优化效果显著,优化提高率为80%;支持向量机对O3的优化效果最理想,优化提高率为36%;线性回归方法对O3的优化效果较好,但对PM2.5的优化效果相对较差.研究显示,机器学习模型显著优化了CAMx模拟结果,反映了利用机器学习修正空气质量数值模式预报结果的研究意义和可行性. In order to improve the prediction accuracy ofρ(PM2.5)andρ(O3)in Xi′an and better serve the prediction and warning work of Xi′an,based on the prediction results of the CAMx model,combined with the mesoscale WRF weather prediction data,ρ(PM2.5)andρ(O3)observation data,this study optimized the simulation results ofρ(PM2.5)andρ(O3)in Xi′an in 2019 based on multiple linear regression,ridge regression,lasso regression,decision tree,random forest and support vector machine model.The results showed that:(1)The CAMx model had bias in the prediction of pollutants,and the optimization model could obviously correct the systematic deviation of the CAMx model and improve the prediction accuracy.(2)The RMSE values ofρ(PM2.5)andρ(O3)decreased from 174.00 and 37.11μg m3to 34.36-39.37 and 24.77-28.82μg m3,respectively.The R values increased from 0.63 and 0.78 to 0.70-0.78 and 0.83-0.88,respectively.(3)Different models had different advantages in correcting the simulated values.The random forest model had a significant effect onρ(PM2.5)optimization,with an optimization improvement rate of 80%.The support vector machine model had the best effect onρ(O3)optimization,and the optimization improvement rate was 36%.The linear regression method had good optimization effect onρ(O3),but poor optimization effect onρ(PM2.5).The research results show that the machine learning algorithm has significantly optimized the CAMx simulation results,reflecting the research significance and feasibility of the machine learning algorithm to modify the results of the air quality numerical forecast model.
出处 《环境科学研究》 CAS CSCD 北大核心 2021年第4期872-881,共10页 Research of Environmental Sciences
基金 国家重点研发计划项目(No.2017YFC0212202)。
