香港港口近地面O3与氮氧化物浓度变化的多重分形特征

以2012年3月-2014年2月香港葵涌港口大气ρ（O₃）、ρ（NO₂）和ρ（NO_x）的小时均值为研究对象，运用多重分形去趋势互相关分析法，对香港港口近地面ρ（O₃）与ρ（NO₂）和ρ（NO_x）相互作用的多重分形特征进行研究.结果表明:香港港口O₃ vs NO₂ρ（O₃）vs ρ（NO₂），下同]的h（2）（广义Hurst指数）为0.80，O₃ vs NO_xρ（O₃）vs ρ（NO_x），下同]的h（2）为0.79，二者的h（2）均大于0.5，表明ρ（O₃）与ρ（NO₂）、ρ（NO_x）间均存在显著的长程交叉相关性.港口ρ（O₃）与ρ（NO₂）、ρ（NO_x）相互关系的多重分形特性在日际和季节上存在显著差异，其中夜晚的O₃ vs NO₂的Δα（分形强度指数）为0.92，O₃ vs NO_x的Δα为0.81，而白天二者的Δα分别为0.59和0.43，说明夜晚的多重分形特征明显强于白天；Δα为春季>夏季>秋冬，表明多重分形程度在春季最强，夏季次之，秋冬季最弱，可能与港口特殊的地理位置和气候条件有关.研究显示，在构建港口ρ（O₃）的预测模型时，需要考虑不同的时间尺度、不同气象条件下ρ（O₃）与ρ（NO₂）、ρ（NO_x）相关性的多重分形特征的差异性. 

Multifractal Property between Ground-level Ozone and Nitrogen Oxide Concentrations in Hong Kong Port

The time series of multifractal property between ground-level ozone and nitrogen oxide was investigated in Hong Kong Port area. ρ(O₃), ρ(NO₂) and ρ(NO_x) were observed hourly at a port monitoring station from March 2012 to February 2014. Based on the collected database, cross-correlation testing was carried out first to examine the cross-correlation; the results indicated that statistically significant correlations existed. MF-DCCA was subsequently performed. The generalized Hurst exponent h(2) of port O₃ vs NO₂ was 0.80, and h(2) of O₃ vs NO_x was 0.79. Both h(2) values were greater than 0.5, and as a result the multifractality was found to remarkably occur between ozone and its precursors, which features long-term memory. Besides, these multifractal properties were found to be different from each other. For example, the fractal intensity index Δα was 0.92 for O₃ vs NO₂ at night, and Δα of O₃ vs NO_x was 0.81, and the values of Δα were 0.59 and 0.43 for daytime. This indicated that the long-term cross-correlation behavior at night was verified to be stronger than that at daytime.Among the seasons, the long-term cross-correlation behaviors in spring were found to be the greatest, followed by summer, autumn and winter, because by comparing the Δα of the four seasons, the largest was in spring, followed by summer, autumn and winter. The difference of multifractal property might be influenced by the special geographical and climatic conditions in Hong Kong Port area. Hence, it is necessary to separate it into varied seasons or varied periods of time when constructing ρ(O₃) forecasting model. The findings illustrate that multifractal analysis could be a useful tool for describing the temporal scaling behavior of ozone time series and its precursors.