Spatial features of PM2.5 concentration in the Yangtze River Delta in 2016 were analyzed using remote sensing data. Selecting factors among meteorology, topography, vegetation, and emission list of air pollutants, factors and their interaction effects on the spatial distribution of PM2.5 concentration were studied based on GAM, with an evaluation unit of 0.25°×0.25° for the grid. It showed that:① With a more significant difference between the north and south, PM2.5 concentration was generally higher in the north and west but lower in the south and east. In the southern part of the delta, the concentration was mostly lower than 35 μg·m-3, with noncompliance of the PM2.5 concentration scattered in urban areas like islands. Meanwhile, PM2.5 concentration is generally over 35 μg·m-3, and the pollution appeared like sheets. ② Besides, PM2.5 concentration showed an apparent positive spatial autocorrelation with "High-High" PM2.5 agglomeration areas in the north of the delta and "Low-Low" PM2.5 agglomeration areas in the south. ③ Based on GAM, hypsography, temperature, and precipitation negatively affected PM2.5 concentration, whereas pollutant emissions positively affected it. The effect of wind was minor when its speed <2.5 m·s-1, and more negatively significant when its speed ≥ 2.5 m·s-1. Hypsography, temperature, and precipitation were higher in the southern part of the delta, but they were lower in the northern part, leading to a higher PM2.5 concentration in the northern parts and lower in the southern parts. A higher wind speed in the east and lower in the west also led to a concentration difference between them. ④ All factors had passed a significant pair interaction test, except for hypsography and PM2.5 emission, and they all showed a significant interaction effect on the distribution of PM2.5 in the Yangtze River Delta.
Keywords: PM2.5; Yangtze River Delta; factors; grid; interaction effect; spatial distribution.