利用臭氧探空数据评估卫星及再分析资料在粤港澳地区的适用性  

Assess the applicability of satellite and reanalysis data in Guangdong-Hongkong-Macao regions based on ozonesonde measurements

作　　者：龚宇 李婷苑 沈劲 陈靖扬 GONG Yu;LI Ting-yuan;SHEN Jin;CHEN Jing-yang(Guangdong Ecological Meteorological Centre,Pearl River Delta Center for Environmental Meteorology Prediction and Warning,Guangzhou 510640,China;Guangdong Provincial Observation and Research Station for Atmospheric Environment and Carbon Neutrality in Nanling Forests,Guangzhou 511443,China;Key Laboratory of Regional Air Quality Monitoring,Guangdong Environmental Protection Key Laboratory of Secondary Air Pollution Research,Guangdong Ecological Environmental Monitoring Center,Guangzhou 510308,China)

机构地区：[1]广东省生态气象中心(珠三角环境气象预报预警中心),广东广州510640 [2]广东省南岭森林大气环境与碳中和野外科学观测研究站,广东广州511443 [3]广东省生态环境监测中心,国家环境保护区域空气质量监测重点实验室,广东省环境保护大气二次污染研究重点实验室,广东广州510308

出　　处：《中国环境科学》2025年第4期1858-1868,共11页China Environmental Science

基　　金：广东省基础与应用基础研究基金项目(2023A1515110536);广东省科技计划项目(2024B1212040006);广东省重点研发计划项目(2020B1111360003);中国气象局青年创新团队项目(CMA2023QN13);广东省气象局软科学项目(Z202304);广东省气象局科学技术研究项目(GRMC2022Q10)。

摘　　要：基于2022~2023年粤港澳地区臭氧(O_(3))探空观测资料,分析O_(3)浓度垂直分布特征变化趋势,并结合自组织映射神经网络(SOM)分型方法,评估了Aqua卫星大气红外探测器(AIRS)O_(3)垂直廓线产品以及ERA5再分析资料的O_(3)垂直廓线产品在粤港澳地区的适用性.结果表明:①粤港澳地区O_(3)垂直分布季节性差异较为显著,春夏冬季呈单峰分布结构,峰区分别位于700,950和300hPa附近,秋季呈双峰分布结构,峰区位于925和400hPa附近.站点间差异较小,广东省各站点与香港站点的O_(3)廓线年平均偏差介于-3.2%~11.0%之间.②对流层内秋冬季AIRS和ERA5数据质量好于春夏季.850~200hPa,AIRS和ERA5数据质量相对较好,各季节相对平均偏差(Rad)平均值分别介于16.5%~25.8%和15.1%~25.7%之间,相关系数(r)平均值分别介于0.47~0.75和0.23~0.74之间;而在850hPa以下,AIRS和ERA5数据质量相对较差,AIRS数据略好于ERA5数据,Rad平均值分别为31.6%和40.9%,r平均值分别为0.34和-0.15.③将O_(3)垂直分布分为5型,其中,AIRS与ERA5资料在1型分布结构下数据质量最好,而在2型和3型分布结构下数据质量最差.1型在秋冬季出现频率较高(分别为43%和61%),2型和3型在夏季出现频率较高(共66%),4型和5型在春季出现频率较高(共85%).Based on the ozonesonde data in Guangdong,Hong Kong,and Macao regions from 2022 to 2023,the vertical distribution characteristics of O_(3)concentration were analyzed,and the applicability of Aqua satellite AIRS O_(3)vertical profile product and ERA5 reanalysis O_(3)vertical profile product were evaluated by using self-organizing map neural network(SOM)method.The vertical distribution of O_(3)in the Greater Bay Area exhibited significant seasonal variations.In spring,summer,and winter,the ozone vertical distribution displayed a unimodal structure,with peak concentrations located near 700,950,and 300 hPa,respectively.In contrast,the vertical distribution in autumn showed a bimodal structure,with peaks near 925 and 400hPa.The vertical differences in O_(3)between stations were relatively small,with deviations between Guangdong and Hong Kong stations ranging from-3.2%to 11.0%.The quality of AIRS and ERA5 data in autumn and winter within the troposphere was better than that in spring and summer.At 850~200hPa,both AIRS and ERA5 data showed relatively good quality,with seasonal relative average deviations(Rad)ranging from 16.5%to 25.8%for AIRS and 15.1%to 25.7%for ERA5.The average correlation coefficients(r)for the seasons ranged from 0.47 to 0.75 for AIRS and 0.23 to 0.74 for ERA5.Below 850 hPa,the quality of AIRS and ERA5 data was relatively bad,with average value of r were 0.34 and-0.15.The vertical distribution of O_(3)was categorized into 5 types.Among these,the data quality of AIRS and ERA5 was best under the type 1distribution structure,while it was worst under the type 2 and type 3 structures.Type 1 occurred more frequently in autumn(43%)and winter(61%),whereas type 2 and type 3 were more common in summer(66%).Type 4 and type 5 occurred more frequently in spring(85%).

关 键 词：粤港澳 臭氧(O_(3)) 大气红外探测器(AIRS) ERA5 评估 自组织映射神经网络(SOM) 

分 类 号：X511[环境科学与工程—环境工程]