Investigate the effects of urban land use on PM2.5 concentration: An application of deep learning simulation

Abstract

As the fine particulate matter (PM2.5) polluting seriously threat people's health, exploring its mitigation strategies has become an urgent issue to be studied. Urban land use, the carrier of urban functions and human activities, has been widely recognized as an important contributor of PM2.5 pollution. Taking Wuhan metropolitan area as an example, this study employs a deep learning simulation method to explore the effects of land use types and density on the spatial distribution of PM2.5 pollutants. The PM2.5 concentration data, raster-based land use data and meteorological conditions data are analyzed to identify their dynamic spatiotemporal characteristics. The meteorological conditions, including temperature and wind speed, are incorporated into the simulation platform, which improves the precision significantly. The simulation results show that PM2.5 concentration caused by construction land such as industrial, residential, transportation, logistics and warehousing, commercial, utilities, and public service sequentially decreases. The impact of FAR on PM2.5 concentration is spatially different. With the increase of FAR, some north construction pixels present PM2.5 mitigation effects while a few grids in the south appear heavier pollution. By analyzing the results of different simulation scenarios, specific spatial-based PM2.5 mitigation strategies and control measures are provided to promote healthy and sustainable urban development. This method can be transferred and applied to other metropolitans, so as to provide as a reference for policymakers and urban planners to promote effective air pollution mitigation strategies from the view of spatial planning.