Multifractal Characteristics in Air Pollutant Concentration Time Series

Abstract

One-year series of hourly average air pollutant concentration(APC) observations, including O3, CO, SO2 and NO2,were analyzed by means of statistical tools: histogram, variancemethod, and multifractal analysis, to examine the possible scale-invariant behavior and the clustering characteristics. It was found that all APC measurements exhibit the characteristicright-skewed unimodal frequency distribution and long term memory. A monodimensional fractal analysis was performed by thebox counting method. Scale invariance was found in these time series and the box dimension was shown to be a decreasing function of the threshold APC level, implying multifractal characteristics, i.e., the weak and intense regions scaledifferently. To test this hypothesis, the APC time series weretransferred into a useful compact form through the multifractalformalism, namely, the τ(q)-q and f(α)-α plots. The analysis confirmed the existence of multifractal characteristics in the investigated APC time series. It wasconcluded that the origin of both the pronounced right-skewnessand multifractal phenomena in APC time series may be interpretedin terms of the Ott (1995) proposed successive random dilution (SRD) theory and the dynamics of APC distribution process can bedescribed as a random multiplicative process.