Spatiotemporal variations and connections of single and multiple meteorological factors on PM2.5 concentrations in Xi'an, China

Abstract

As one of the main sources that cause air pollution, the investigation on the change of PM2.5 concentrations and its related factors becomes particularly crucial and important. Previous works have rarely considered the combined effects of multiple meteorological factors on PM2.5 at different scales over time. Based on this consideration, the purpose of this paper is to investigate the spatiotemporal changes of PM2.5 concentrations and the relationship between meteorology and PM2.5 based on hourly ground monitoring data in Xi'an from 2018 to 2021. Specifically, we quantify the overall effect of meteorological factors on PM2.5 by convergent cross-mapping, and then explore the connection between a single or multiple meteorological factors and PM2.5 using wavelet transform coherence and multiple wavelet coherence. The experimental results show that: (1) the PM2.5 concentrations in northern Xi'an is significantly higher than that in other regions, and it is decreasing year by year from 2018 to 2021. Especially in 2021, the reaching national standard rates of daily PM2.5 concentrations achieve 88.77%. (2) Mean temperature, relative humidity and atmospheric pressure are the main single factor resulting in changes of PM2.5 concentrations, while the influences of boundary layer height, wind speed and wind direction are relatively small. Moreover, it exhibits the causality between PM2.5 concentrations and meteorology have obvious seasonal characteristics. (3) The PM2.5 concentrations shows the most consistent and significant correlation with boundary layer height and wind speed, with negative correlations at all scales and periods. Among all combinations of meteorological factors, temperature-humidity, and temperature-wind speed-wind direction present the largest significant coherence area in multiple wavelet coherence experiment and are the best combination factors to explain the changes of PM2.5 concentrations. Furthermore, the coherence values of multiple meteorological factors are obviously higher than those of a single meteorological factor, which indicates that the change of PM2.5 concentrations is determined by multiple variables. This study provides new insights for understanding the long-term spatiotemporal changes of PM2.5 concentrations and the relationship between PM2.5 and meteorology.