基于LUR-卫星数据耦合模型的中国PM2.5浓度时空分布研究

A National Satellite-Based Land Use Regression for PM2.5 Pollution in China

摘要: 以PM2.5为代表的大气污染已经成为造成全球疾病负担最重要的环境因素之一，PM2.5浓度的时空分布模拟是其健康风险效应分析的基础。土地利用回归法(Land Use regression, LUR)是模拟PM2.5时空分布的一种有效手段，但一直局限于城市小尺度的研究。近年来，部分研究者结合卫星遥感数据，利用LUR方法成功构建出国家尺度的模型，但在中国相关研究并无报道。本文基于LUR方法，对2013~2015年中国的PM2.5浓度的时空分布进行模拟：加入遥感数据和不加入遥感数据模型的R2分别是0.7和0.55；模型中气象变量居多，在影响PM2.5分布中产生重要影响；不同年份之间差异较大，通过引入时间因子获得不同年份之间的纵向比较；绘制出全国的PM2.5浓度分布图，为后续中国的PM2.5健康效应分析提供污染物暴露依据。 The air pollution such as PM2.5 has been one of the most important environmental factors that cause the global disease burden. The simulation of the spatiotemporal distribution of PM2.5 concentration is the basis for its health risk analysis. Land Use regression (LUR) is an effective means to simulate the spatiotemporal distribution of PM2.5, but it was limited at inner urban scale. In recent years, some researchers have successfully constructed the national LUR model with the satellite remote sensing data. In China, however, there have been no relevant reports. In this study, the spatiotemporal distribution of Chinese PM2.5 concentration in 2013-2015 were simulated based on the satellite-based LUR model: R2 of LUR model with and without remote sensing are 0.7 and 0.55; the meteorological variables play an important role in this model; PM2.5 decreases year by year and introduction of year factor achieved good results; the PM2.5 concentration distribution in China is drawn, providing the basis for the subsequent effects of exposure to PM2.5 in China.