A multi-scale analysis of the relationship between land surface temperature and PM2.5 in different land use types

The high land surface temperature (LST) and fine particulate matter (PM2.5) pose significant threats to urban ecological environment. LST and PM2.5 are closely related, and land use/cover change (LUCC) has a significant influence on them. However, there is a lack of multi-scale analysis of their relationship under the background of LUCC. Therefore, this study designed a multi-scale analysis framework, adopted the Pearson correlation analysis, bivariate Moran's I and coupling coordination degree (D) to examine the relationship between LST and PM2.5 in different land use types based on remote sensing data, and analyzed the influencing factors by Geodetector. The results showed that LST and PM2.5 have a positive correlation, and their correlation would decrease with the research scale expansion, with the absolute values of Moran's I ranging from 0.229 to 0.577, 0.415–0.754 and 0.551–0.852 at national, urban agglomeration and city scales, respectively. LST, PM2.5 and their correlation were lower in forest and water at the three scales, while they were the highest in construction land, and the difference in D values was about 0.3. Furthermore, the normalized difference vegetation index (NDVI) and construction land area played the most important role in spatial patterns of this relationship, followed by population and GDP. At last, the cost analysis of environmental pollution reduction was performed, and this study could provide support for proper public health policy-making and environmental protection in different regions.