Spatio-temporal variability and persistence of PM2.5 concentrations in China using trend analysis methods and Hurst exponent

Abstract

Studying the persistence and spatial-temporal trends of air pollution is beneficial for determining the pollutant risk area and enables the development of associated prediction tools and models. Relying on the PM2.5 concentrations data retrieved via remote sensing from 2000 to 2018, the spatial and temporal pattern, variation tends, and persistence is determined through the Theil-Sen median trend analysis, Mann-Kendall, and Hurst exponent. We combine the Theil-Sen Median + Mann-Kendall and Hurst to quantitatively and qualitatively predict the future trends of China's PM2.5 concentrations as a new perspective. Results reveal that PM2.5 concentrations increased at first and then decreased significantly, with 2009–2011 as the turning point for PM2.5 pollution changes, particularly in Central China and the Southeast Coastal Area. The area where PM2.5 concentrations were below 10 μg/m3 account for 29.75% of China's total territory, reaching the annual average criterion value determined by the World Health Organization. The areas presenting a continuous increase (15.69%) and decline (17.46%) of PM2.5 concentrations were almost equal. As a result, the constant monitoring of the variance in PM2.5 concentrations in the sustainably increased and underdetermined regions, such as Tibet and Northeast China, is needed. This study used simulated PM2.5 concentrations data as a valuable complement to China's ground monitoring stations, thus compensating for a shortage of long-term series data. Grid data analysis can more finely show the interior disputes in PM2.5 concentrations. The algorithm codes can be freely downloaded and become a helpful tool for analyzing the spatio-temporal variation characteristics of primary air pollutants.