Studying Correlation in the AERONET AOD Time Series Using Multifractal Analysis

In this work, we analyze the AERONET’s solar photometer Aerosol Optical Depth from four stations in Brazil to investigate the scale-invariant behavior using the multifractal detrended fluctuation analysis. The selected stations have different levels of exposure to pollution in a year and exhibit long-term correlation and multifractal behavior. All three measures, viz., the generalized Hurst exponent, mass exponent, and multifractal spectrum quantifies the multifractal behavior. For all the four stations, the fluctuations quantified by the Hurst exponent varying between 0.6 and 1.8 indicate long-term correlations in the data. Non-linear variation of the mass exponent with negative and positive moments, and the width of the multifractal spectra, between 0.5 and 1.27 confirm the multifractal behavior. To summarize, for the chosen stations the multifractal analysis is able to distinguish between sites influenced by biomass burning emissions from the urbanized ones. Results for all sites show the sensitivity to larger temporal scales.