基于土地利用与卫星数据建立PM2.5与含碳组分的暴露评价模型

PM2.5 has become the primary challenge of public health in China. The carbonaceous matter are the most important constituents of PM2.5 that could cause significant adverse health effects. It therefore has become a need to assess health risk of PM2.5 and its carbonaceous constituents. However, interpretation of findings from health studies has been hampered by uncertainties in exposures. Air pollutants exposure assessments have traditionally relied upon ambient concentration measurements, and ignored the spatial variation of targeted air pollutants which may cause exposures misclassification. The aims of this study are 1) to develop exposure assessment models with fine spatial and temporal resolutions by incorporating Land-Use Regression model and satellite data for spatial surfaces and temporal variation prediction of targeted air pollutants at a city-wide scale; 2) to discuss the applicability and validity of model methodology on PM exposure assessment, especially in an environment with various emission sources and heavy air pollution, and 3) to investigate the emission sources and potential factors which may impact the exposure characterization of targeted air pollutants. The results will provide improvements on the methodology of air pollution exposure assessment as well as accuracy of estimation on health effect of exposure. Moreover, this study will help on design strategies for pollution control on PM2.5 and its carbonaceous constituents and for their public health prevention.

PM2.5污染导致的健康危害已成为我国面临的主要公共卫生挑战之一。其中PM2.5的含碳组分是引发健康危害的主要成分之一。开展PM2.5及其碳组分污染的健康风险评估已成为应对这一挑战的迫切需求。但由于缺乏准确有效的暴露评价方法，制约了大气污染流行病学研究的开展。传统基于固定监测点的暴露评价方法，忽视了大气污染物的空间异质性，易造成暴露错分。本研究拟结合土地利用回归模型和卫星遥感监测，①建立具有高时空分辨率的PM2.5与含碳组分的暴露评价模型，模拟城市内目标污染物的浓度曲面及时间动态变化规律；②探讨该建模方法在复合性、高浓度环境下的颗粒物暴露评价的适用性和有效性; ③探索影响人群颗粒物暴露特征的主要来源和相关因素。本研究预期的研究结果有助于弥补我国传统大气污染研究在暴露评价方面的不足，提高颗粒物暴露-健康效应关系研究结果的准确性，对于制定大气污染防控措施和公共卫生预防策略均具有重要意义。

细颗粒物(粒径小于2.5µm的颗粒物，PM2.5)污染已成为我国面临的主要公共卫生挑战之一。开展PM2.5及其碳组分污染的健康风险评估已成为应对这一挑战的迫切需求。但由于缺乏准确有效的暴露评价方法，制约了大气污染流行病学研究的开展。传统基于固定监测点的暴露评价方法，忽视了大气污染物的空间异质性，易造成暴露错分。本研究基于加密观测，结合土地利用回归模型(LUR)和卫星遥感数据，建立了空间分辨率为1km×1km的暴露评价模型，模拟了研究地区颗粒物的浓度曲面及其时间动态变化规律。最终模型共纳入了三类变量：监测点1000米缓冲区内水域用地面积, 1500米缓冲区内一级道路长度, 100米缓冲区内农业用地的面积，提示交通污染排放和秸秆燃烧是该地区人群暴露的主要来源。本研究建立的模型调整R2与留一交叉验证R2分别为0.65和0.60，模型估计的均方根误差为3.11μg/m³，提示模型预测值与实际值的一致性较高。本研究的研究结果表明LUR建模方法在复合性、高浓度环境下的颗粒物暴露评价中具有较好的适用性和有效性，为泰州队列研究提供了更为准确可靠的暴露评价基础，并对于当地制定相关的大气污染防控措施和公共卫生预防策略均具有一定参考价值。