三维空气污染物系统的多元多重分形分析  

Multivariate Multifractal Analysis of Three-dimensional Air Pollutant System

作　　者：贾娜 王宏勇[1] JIA Na;WANG Hongyong(School of Applied Mathematics, Nanjing University of Finance & Economics, Nanjing 210023, China)

机构地区：[1]南京财经大学应用数学学院,江苏南京210023

出　　处：《环境科学与技术》2019年第3期214-221,共8页Environmental Science & Technology

基　　金：教育部人文社科规划基金项目(12YJAZH020);江苏省研究生科研创新计划基金项目(KYCX17_1203)

摘　　要：运用最新提出的多元多重分形去趋势波动分析法(MV-MFDFA)研究北京和上海两地的三维空气污染物(PM_(2.5)、PM_(10)、NO_2)浓度指数系统的非线性波动特征及动态复杂性。实证结果表明,两地污染物系统及其分量序列均存在长程相关性和时变的多重分形特征,相比较而言,北京污染物系统序列的多重分形强度更大,而上海污染物系统的长程相关性更强。此外,两地污染物系统在不同时间标度下呈现出不同的多重分形特征。最后,研究了两地污染物系统序列具有多重分形性的原因,指出胖尾概率分布和长程相关性是导致序列多重分形性的共同原因,且胖尾概率分布的贡献更大。Using the newly proposed method of the multivariate multifractal detrended fluctuation analysis (MV-MFDFA), the nonlinear fluctuation characteristics and dynamic complexity of the series of the three-dimensional air pollutant (PM2.5, PM10,NO2) concentration index systems for Beijing and Shanghai were studied. The empirical results show that both pollutant systems and their component series have the long-range correlation and time-varying multifractality. In comparison, the multifractal strength of the series of Beijing pollutant system is greater than that of Shanghai, while the long-range of Shanghai pollutant system is stronger than that of Beijing. In addition, it was found that the two pollutant systems present different multifractal features at different time scales. Finally, the reasons for the multifractality of the two pollutant systems were investigated, and it was found that both the fat-tailed probability distribution and long-range correlation are common causes for the multifractality, and the fat-tailed probability distribution has a major effect on the multifractality.

关 键 词：三维污染物系统 MV-MFDFA 长程相关性 多重分形性 

分 类 号：X513[环境科学与工程—环境工程]