2013-2020年京津冀地区PM_(2.5)浓度时空变化模拟及趋势分析  

Simulation and Trend Analysis of Spatiotemporal Variation of PM_(2.5) Concentrations in the Beijing-Tianjin-Hebei Region from 2013 to 2020

作　　者：杨晓辉 肖登攀[1,2,3] 柏会子 唐建昭 王卫 郭风华[1] 刘剑锋[1] YANG Xiao-hui;XIAO Deng-pan;BAI Hui-zi;TANG Jian-zhao;WANG Wei;GUO Feng-hua;LIU Jian-feng(Institute of Geographical Sciences,Hebei Academy of Sciences/Hebei Technology Innovation Center for Geographic Information Application,Shijiazhuang 050011;College of Geography Science,Hebei Normal University,Shijiazhuang 050024;Hebei Key Laboratory of Environmental Change and Ecological Construction,Shijiazhuang 050024,China)

机构地区：[1]河北省科学院地理科学研究所/河北省地理信息开发应用技术创新中心,河北石家庄050011 [2]河北师范大学地理科学学院,河北石家庄050024 [3]河北省环境演变与生态建设实验室,河北石家庄050024

出　　处：《地理与地理信息科学》2022年第4期58-67,共10页Geography and Geo-Information Science

基　　金：河北省科学院高层次人才培养与资助项目(202201)。

摘　　要：为研究京津冀地区PM_(2.5)浓度时空变化趋势,利用多角度大气校正(MAIAC)气溶胶光学厚度(AOD)产品,结合气象和土地利用等数据,构建线性混合效应(LME)和地理加权回归(GWR)组成的两阶段统计回归模型,建立了2013-2020年1 km空间分辨率的PM_(2.5)浓度数据集。结果显示:模型交叉验证后的决定系数(R^(2))、斜率、均方根预测误差(RMSPE)和相对预测误差(RPE)范围分别为0.85~0.95、0.87~1.05、7.87~29.90μg/m^(3)和19.19%~32.71%,数据质量较高;2013-2020年京津冀地区PM_(2.5)浓度呈现出明显的时间特征(冬季高、夏季低)和空间特征(南部平原高、北部山区低);相对2013年,2020年PM_(2.5)高浓度区域明显缩小,年均浓度下降54.04%,全域降至55μg/m^(3)以下,由于政府对污染物排放的严格控制,2015-2017年冬季PM_(2.5)浓度出现大幅下降;相对2017年,2018-2020年PM_(2.5)浓度下降不明显。研究结果可为京津冀及周边地区空气污染防治提供科学依据。Particulate matter with an aerodynamic diameter of less than 2.5μm(PM_(2.5))has a significant impact on air pollution,atmospheric visibility and human health.In recent decades,China′s economy has developed rapidly,especially in the Beijing-Tianjin-Hebei(BTH)Region,where air pollution problems have become more and more serious.The most important step in regional air pollution control is to obtain high-resolution and high-quality PM_(2.5) data from satellite remote sensing products.In order to study the spatiotemporal changes and trends of PM_(2.5) in the BTH region,the data of multi-angle implementation of atmospheric correction(MAIAC)aerosol optical depth(AOD)products,meteorology and land use were used.A two-stage statistical regression model was constructed by linear mixed effects(LME)and geographically weighted regression(GWR),and a high-spatial resolution(1 km)and high-quality PM_(2.5) datasets from 2013 to 2020 was established.The model was fitted annually,and the coefficient of determination(R^(2)),slope,root-mean-squared prediction errors(RMSPE),and relative prediction error(RPE)ranges of the model′s cross-validation results were 0.85~0.95,0.87~1.05,7.87~29.90μg/m^(3) and 19.19%~32.71%,respectively.Overall,the PM_(2.5) concentrations displayed obvious temporal characteristics(high in winter and low in summer)and spatial characteristics(high in the southern plains and low in the northern mountains)in the BTH region from 2013 to 2020.During the study period,the area with high concentrations in 2020 significantly reduced compared with 2013,the annual average PM_(2.5) concentrations dropped by 54.04%,and the study areas reduced to below 55μg/m^(3).In terms of winter,the PM_(2.5) concentrations dropped sharply from 2015 to 2017 due to the government′s strict control of pollutant emissions.Compared with 2017,the decrease in PM_(2.5) concentrations from 2018 to 2020 was not obvious,mainly because the low-polluting area had a tendency to expand.Therefore,more attention should be given to winter when the po

关 键 词：PM_(2.5) 气溶胶光学厚度(AOD) 两阶段统计回归模型 标准差椭圆 时空变化 京津冀地区 

分 类 号：X513[环境科学与工程—环境工程]