Multifractal analysis of interactive patterns between meteorological factors and pollutants in urban and rural areas

Abstract

This paper seeks to enhance understanding of the cross-correlation patterns between meteorological factors and pollutants. The observed databases of daily meteorological elements (temperature, humidity and wind speed), as well as pollutants (CO, NOx, PM10 and SO2) levels during 2005–2014, is collected. Based on the database, the cross-correlation test is carried out firstly and the results indicate that cross-correlation behaviors exist statistically between them. Then the detrended cross-correlation analysis is performed for further analysis. With a detailed comparison, long-term cross-correlation behaviors are found to be more obvious in rural area. Beside, the influences of meteorological factors on multifractal property for pollutants are investigated. In contrast to humidity and wind speed, the long-term cross-correlation behaviors between temperature with pollutants are found to be more evident in both urban and rural areas. Furthermore, the difference of multifractal property for varied pollutants is explored. The strengths of multifractal spectra between meteorological factors with PM10 are strongest while the corresponding values between meteorological factors with SO2 are weakest. These findings successfully illustrate that the multifractal analysis is a useful tool for uncovering the interactive pattern in environmental issues.