Spatial Variation Analysis of Urban Air Pollution Using GIS: A Land Use Perspective

Does land-use and land-cover (LULC) impact the formation of spatial variation characteristics of urban air pollution? Aiming at this problem, we firstly retrieved land use/cover from Landsat 8 image and consequently used them to calculate and map landscape metrics and area ratio of each land use/cover type. Meanwhile, space mapping of NO₂, PM₁₀, O₃, and PM_2.5 concentrations as well as meteorological factors were conducted through inverse distance weighted interpolation method. After that, geographical detector has been introduced to analyze the influence of land use on air pollution quantitatively with/without meteorological elements as confounding factors. The results show that the NO₂ and PM_2.5 concentrations increased with the increment of area ratio of built-up area, contrast to the fact that it was negatively correlated with that of green-land. This situation was completely opposite to that of O₃. The air pollution concentrations were higher in the non-built-up area with greater Shannon diversity index and Perimeter-Area Fractal (PAFRAC) Dimension index values. The power of landscape factors is a little bit lower than that of LULC type factors in indicating air pollution. Pairs of LULC type factors were found to enhance each other to increase the air pollution concentrations, so were pairs of landscape factors. Meteorological factors were found to reinforce the control of air pollution of LULC type factors as well as landscape factors significantly. In view of the above considerations, we draw the conclusion that the spatial patterns of air pollution probed were closely related to land factors, either in LULC or landscape metrics. The land use and meteorological conditions are both factors that have to be considered in forecasting and mitigating urban air pollution in early urban planning.