Harvard University, School of Public Health, Department of Environmental Health, 401 Park Drive, Landmark Center 4th Floor West, Boston, MA 02115, United States.
Harvard University, School of Public Health, Department of Environmental Health, Boston, MA, United States.
Environ Int. 2018 Oct;119:152-164. doi: 10.1016/j.envint.2018.06.021. Epub 2018 Jun 26.
Precise population information is critical for identifying more accurate environmental exposures for air pollution impacts analysis. Basically, there are two methods for estimating spatial distribution of population, choropleth and dasymetric mapping. While the choropleth approach accounts for linear distribution of population over area based on census tract units, the dasymetric model accounts for a more heterogeneous population density by quantifying the association between the area-class map data categories and values of the statistical surface as encoded in the census dataset. Environmental epidemiological studies have indicated the dasymetric mapping as a more accurate approach to estimate and characterize population densities in large urban areas. However, investigations that have attempted to compare the exposure estimates from choropleth versus dasymetric mapping in environmental health analysis are still missing. This paper addresses this gap and compares the impact of using choropleth and dasymetric mapping in different exposure metrics. We compare the impact of using choropleth and dasymetric mapping in three case studies, defined here as case study A (relationship between urban structure types and health), case study B (PM emissions and human exposure), and case study C (distance-decays of mortality risk related to PM emitted by traffic along major highways). These case studies represent previous investigations performed by our research group where spatial distribution of population was an essential input for analysis. Our findings indicate that the method used to estimate spatial distribution of population impacts significantly the exposure estimates. We observed that the choropleth mapping overestimated exposure for the case study A and B, while for the case study C the exposure was underestimated by the choropleth approach. Our findings show that the dasymetric model is a preferred method for creating spatially-explicit information about population distribution for health exposure studies. The results presented here can be useful for the environmental health community to more accurately assess the relationship between environmental factors and health risks.
精确的人口信息对于识别更准确的环境暴露对于空气污染影响分析至关重要。基本上，有两种方法可以估计人口的空间分布，即等值线图和密度分配制图法。虽然等值线方法根据普查区单位计算人口在区域上的线性分布，但密度分配模型通过量化区域-类别图数据类别与统计表面值之间的关系来更准确地描述人口密度，该统计表面值编码在普查数据集中。环境流行病学研究表明，密度分配制图法是估计和描述大城市地区人口密度的更准确方法。然而，在环境健康分析中，尝试比较等值线与密度分配制图法的暴露估计的研究仍然缺乏。本文旨在填补这一空白，并比较在不同暴露指标中使用等值线和密度分配制图法的影响。我们在三个案例研究中比较了使用等值线和密度分配制图法的影响，这里将其定义为案例研究 A（城市结构类型与健康之间的关系）、案例研究 B（PM 排放和人体暴露）和案例研究 C（与交通沿线主要高速公路排放的 PM 相关的死亡率风险的距离衰减）。这些案例研究代表了我们研究小组之前进行的研究，其中人口的空间分布是分析的重要输入。我们的研究结果表明，用于估计人口空间分布的方法显著影响暴露估计。我们发现，对于案例研究 A 和 B，等值线制图法高估了暴露，而对于案例研究 C，等值线方法低估了暴露。我们的研究结果表明，密度分配模型是创建与健康暴露研究相关的人口分布空间信息的首选方法。这里呈现的结果对于环境健康界更准确地评估环境因素与健康风险之间的关系可能是有用的。
