Massive monitoring data requires effective statistical analysis methods. This paper aims to visualize the spatiotemporal characteristics and spillover effect of air pollutants. Ground-based PM2.5 data in 336 cities across China revealed a tough but improving situation. The PM2.5 average annual concentrations in 2016 and 2017 were 47 ± 18 µg/m3 and 44 ± 16 µg/m3 respectively, but a worse, or at least a not-improving PM2.5 situation happened in winter. A slight declining north-south disequilibrium and a growing east-west disequilibrium exhibited in 2017, along with an increasing weight of eastern and southern pollution in the proportion of the overall pollution level. North-south disequilibrium existed stably throughout the year but east-west disequilibrium was erratic. Nearly half of the cities exhibited significant spillover effects, presenting 2 clusters with spillover by high concentrations and 3 clusters with spillover by low concentrations. Most cities in Hebei, Shandong and Henan provinces showed a high but decreasing spillover effect, but increasing trend happened in the cities in Anhui and Shanxi provinces. A significant correlation appeared between the city population and PM2.5 concentration. Population density explained about 25% of the PM2.5 concentration change, and the explanation ability increased in 2017. A higher influence of population on PM2.5 concentration happened in the early stage of city development, and the influence exhibited spatial differences. The city population and PM2.5 spillover effect existed an overall positive correlation, but the population only addressed about 10% of the spillover effect change. Our findings provide important information for the joint prevention and control of air pollution in China, and the approach proposed in this paper is applicable to other fields.
Keywords: Cluster detecting model; Gravity center model; PM(2.5); Visualization expression.
Copyright © 2019 Elsevier B.V. All rights reserved.