长沙地区O_(3)污染潜势广义可加模型的季节分异性  

Seasonal Heterogeneity of Ozone Pollution Potential in Changsha Based on Generalized Additive Models

作　　者：花瑞阳 倪长健[1] 李蔚 叶成志 石荞语 HUA Ruiyang;NI Changjian;LI Wei;YE Chengzhi;SHI Qiaoyu(College of Atmospheric Sciences,Chengdu University of Information Technology,Chengdu 610225,China;Changchun Meteorological Observatory,Changchun 130051,China;Hunan Meteorological Observatory,Changsha 410000,China;Institute of Plateau Meteorology,China Meteorological Administration,Chengdu 610072,China)

机构地区：[1]成都信息工程大学大气科学学院,四川成都610225 [2]长春市气象台,吉林长春130051 [3]湖南省气象台,湖南长沙410000 [4]中国气象局成都高原气象研究所,四川成都610072

出　　处：《中国环境监测》2023年第5期125-133,共9页Environmental Monitoring in China

基　　金：国家重点研发计划项目(2018YFC0214004,2018YFC1506006);四川省科技厅应用基础研发项目(2021YJ0314)。

摘　　要：基于长沙市2016—2019年臭氧(O_(3))浓度的逐时监测资料以及该时段同时次的气象观测数据,首先利用相关性分析和方差膨胀因子相结合的方法排除了相对湿度、太阳辐射、气温、风速、气压之间存在多重共线性问题,进而构建了春、夏、秋、冬四季O_(3)日最大8 h滑动平均质量浓度(O_(3)-8 h)与上述气象因子的广义可加模型(GAMs),分析了O_(3)污染潜势GAMs模型的季节分异特征。结果表明:(1)春、夏、秋、冬四季O_(3)与相对湿度、太阳辐射、气温、风速、气压各变量之间多呈现出非线性关系(自由度大于1)。(2)春、夏、秋、冬四季多变量GAMs模型方差解释率(IRV)分别为80.7%、60.2%、83.0%、81.4%,调整判定系数R2分别为0.795、0.564、0.819、0.795,即不同季节气象因子在GAMs模型中对O_(3)的解释能力存在显著差异,秋季最好,冬春季次之,夏季最差。(3)相对湿度、太阳辐射、气温是决定春、夏、秋、冬四季O_(3)浓度变化最重要的气象要素,但其重要性排序随季节有所变化,对应的太阳辐射的F统计值分别为140.841、36.606、14.16、46.377,相对湿度的F统计值分别为3.291、4.158、15.82、8.105,气温的F统计值分别为7.030、2.113、15.79、3.340。该结论揭示了气象因子对O_(3)演化影响的复杂性,并为后续O_(3)污染潜势的预报奠定了基础。Based on hourly observational data of O_(3) concentration and corresponding meteorological data in Changsha from 2016 to 2019,firstly,multicollinearity problems among relative humidity,solar radiation,temperature,wind speed and air pressure were excluded by correlation analysis conbined with variance inflation factor.Then,the Generalized Additive Models(GAMs)of the O_(3)-8h and the above meteorological factors in spring,summer,autumn and winter were constructed,and the seasonal variation characteristics of the GAMs model for O_(3) pollution potential were analyzed.The results showed that:①In spring,summer,autumn and winter,the relationship between O_(3)-8 h and relative humidity,solar radiation,temperature,wind speed and air pressure variables mostly was nonlinear(degree of freedom greater than 1).②The variance explanation rate IRV of the multivariate GAMs model in spring,summer,autumn and winter were 80.7%,60.2%,83.0%and 81.4%respectively,and the adjusted determination coefficient R2 were 0.795,0.564,0.819 and 0.795 respectively,indicating that different seasonal meteorological factors had significant differences in the explanatory power of O_(3)-8 h in the GAMs model,autumn was the best,followed by winter and spring,and summer was the worst.③Relative humidity,solar radiation and temperature were the most important meteorological elements determining the change of O_(3) concentration in spring,summer,autumn and winter,but their order of importance varied with the seasons,and the F statistic value of corresponding solar radiation were 140.841,36.606,14.16,46.377,F of relative humidity were 3.291,4.158,15.82,8.105,and temperature were 7.030,2.113,15.79,3.340,respectively.The above conclusions revealed the complexity of the influence of meteorological factors on O_(3) evolution,and laid a solid foundation for the subsequent prediction of O_(3) pollution potential.

关 键 词：臭氧 污染潜势 GAMs模型 季节 分异性 

分 类 号：X515[环境科学与工程—环境工程] X823