Investigation of multifractal cross-correlation surfaces of Hurst exponents for some atmospheric pollutants
Abstract
We present the results of a multifractal multiscale analysis of the cross-correlations between ozone and some precursors (nitrogen oxides (NOX), non-methane hydrocarbons (NMHC), total hydrocarbons (THC) and methane (CH4)) during the cold season and the cross-correlations of each of these gases with temperature. Both individual data and the cross-correlated pairs have multifractal nature and are positive long-term correlated. A proper fitting, using a four parameters logistic function is proposed for the Hurst curve dependences versus fluctuation order. We define a new parameter for the variation of multifractality across time scales and discuss it in coupling with the values for the strength of multifractality. The overall variation of Hurst surfaces along time scale intervals is considered in defining a new measure for the global strength of long-range correlations presented both in the pairs of data and in each individual series. Our results suggest new possible way for characterization and understanding the complex relationship between principal pollution elements and their relationship with temperature.
- Publication:
-
Physica A Statistical Mechanics and its Applications
- Pub Date:
- May 2020
- DOI:
- 10.1016/j.physa.2019.123799
- Bibcode:
- 2020PhyA..54523799S
- Keywords:
-
- Multifractal multiscale analysis;
- Hurst exponents;
- Cross-correlations;
- Atmospheric pollutants