基于CatBoost-SHAP-MCM模型的关中地区PM_(2.5)浓度的气象影响因素研究  

Research on the Meteorological Influencing Factors of PM_(2.5) Concentration in Guanzhong Region Based on CatBoost-SHAP-MCM Model

作　　者：苏佳[1] 聂达文 李晓萌 张新生[1] 宋金昭[1] 董明放 SU Jia;NIE Dawen;LI Xiaomeng;ZHANG Xinsheng;SONG Jinzhao;DONG Mingfang(College of Management,Xi'an University of Architecture and Technology,Xi'an 710055,China)

机构地区：[1]西安建筑科技大学管理学院,陕西西安710055

出　　处：《环境科学研究》2025年第4期787-797,共11页Research of Environmental Sciences

基　　金：陕西省教育厅科学研究计划项目人文社科专项(No.24JK0131);国家社会科学基金年度项目(No.23BTJ034);中国博士后基金项目(No.2020M683433)。

摘　　要：为研究关中地区PM_(2.5)浓度变化及其复杂因素间的非线性关系,基于2020年1月−2023年12月的气象数据,从年、季和月不同时间尺度深入分析关中地区PM_(2.5)的空间分异特征;采用最大信息系数分析关中地区PM_(2.5)与其他大气污染物的关系,同时利用CatBoost-SHAP-MCM模型识别PM_(2.5)浓度的关键气象影响因素。结果表明:①关中地区PM_(2.5)浓度呈明显的空间分布和季节变异性。年际PM_(2.5)浓度在2021年最低,为42.93μg/m^(3),在2022年最高,达49.09μg/m^(3)。季度和月际变化较为相似,均呈冬季高、夏季低的特征,冬季污染最严重,PM_(2.5)浓度达84.35μg/m^(3),夏季最轻,为21.42μg/m^(3)。西安市、咸阳市和渭南市为高污染城市,铜川市和宝鸡市为低污染城市。②PM_(2.5)浓度与PM10浓度的相关性最高,与CO浓度、SO2浓度相关性均较低。③露点温度、气温、海平面气压、降水量和地面气压为关键气象影响因素,其在各城市表现出显著的影响作用,对关中地区整体和各城市的影响基本保持一致。④在低露点温度、低气温以及低露点温度、高海平面气压和高地面气压等特定因素组合下,其对PM_(2.5)浓度的影响更为显著。研究显示,关中地区PM_(2.5)浓度具有明显的空间分异特征和季节性变化特征,且与露点温度、气温、海平面气压、降水量和地面气压等气象因素密切相关,在特定气象组合条件下PM_(2.5)浓度波动更为显著。In order to study the nonlinear relationship between PM_(2.5)concentration variations and complex factors in Guanzhong region,the spatial distribution characteristics of PM_(2.5)on different time scales such as year,season and month were deeply analyzed based on the meteorological data from January 2020 to December 2023.The maximum information coefficient was used to examine the relationship between PM_(2.5)and other air pollutants in the region,and the CatBoost-SHAP-MCM model was applied to identify the key meteorological factors affecting PM_(2.5)concentrations.The results show that:(1)The PM_(2.5)concentration in the Guanzhong region exhibits obvious spatial distribution and seasonal variation characteristics,with the lowest annual average PM_(2.5)concentration of 42.93μg/m^(3)in 2021 and the highest of 49.09μg/m^(3)in 2022.Seasonal and monthly variations are similar,both showing heavy pollution in winter and low pollution in summer.Pollution is most serious in winter,with PM_(2.5)concentration reaching 84.35μg/m^(3),and the lowest pollution in summer,at 21.42μg/m^(3).Xi′an,Xianyang and Weinan are high-pollution cities,while Tongchuan and Baoji are low-pollution cities.(2)PM_(2.5)shows the highest correlation with PM10 concentrations,while correlations with CO and SO2 concentrations are relatively low.(3)Dew point temperature,air temperature,sea level pressure,precipitation and surface pressure are the key meteorological influencing factors.Their influence is still significant in each city,and their effects are consistent across both the entire Guanzhong region and individual cities.(4)Under specific combinations such as low dew point temperature with low air temperature,and low dew point temperature with high sea level pressure and high surface pressure,the effects on PM_(2.5)concentrations become more pronounced.The study demonstrates that PM_(2.5)concentrations in the Guanzhong region exhibit distinct spatial differentiation and seasonal variation patterns,and are closely related to meteorological fac

关 键 词：关中地区 PM2.5 影响因素 CatBoost-SHAP-MCM模型 

分 类 号：X321[环境科学与工程—环境工程]