National scale spatiotemporal land-use regression model for PM2.5, PMio and NO2 concentration in China
National scale spatiotemporal land-use regression model for PM2.5, PMio and NO2 concentration in China
复制标题
DOI：
10.1016/j.atmosenv.2018.08.046
复制
发表时间：
2018-11-01
影响因子：
5
通讯作者：
Chen, Kun
中科院分区：
文献类型：
--
作者：
Zhang, Zhenyu;Wang, Jianbing;Chen, Kun
背景：空气污染流行病学研究越来越依赖于高分辨率的暴露预测模型。然而，到目前为止，这种类型的模型存在于中国使用。目的：我们制作了一个全国土地利用回归模型（LUR），以估计2014年至2016年中国的月平均PM2.5，PMio和NO2。方法：我们开发了一个时空半参数模型，使用广义加法混合模型。模型中包括各种预测变量：随时间变化的气象数据、Globaland 30提供的高分辨率土地覆盖数据、气溶胶光学厚度的卫星测量值以及地理信息系统衍生的预测变量。我们评估了两个交叉验证（CV）的方法，包括保持CV，和10倍CV的模型性能。结果：超过22,000个月的观测在1382监测点，包括估计空气污染暴露。时变空间项分别解释了87%、71%和69%的变异性，PM2.5、PM2.0和NO2模型的交叉验证R-2分别为0.85、0.62和0.62。模型表明，气象变量，人口密度，海拔，距离道路，和土地覆盖类型是重要的预测因子为空气污染exposure.Conclusions：我们已经开发出一个新的全国范围内的模型来估计居住水平的空气污染暴露，这可以用于研究的慢性不利影响的空气污染。
Background: Air pollution epidemiological studies increasingly rely on high-resolution exposure prediction models. However, to date, few models of this type exist for use in China.Objectives: We produced a national land-use regression model (LUR) to estimate monthly average PM2.5, PMio and NO2 from 2014 to 2016 in China.Methods: We developed a spatiotemporal semi-parametric model using generalized additive mixed models. A variety of predictor variables were included in model: time varying meteorological data, high resolution land cover data from Globaland30, satellite measures of aerosol optical depth, and Geographic Information System (GIS)-derived predictors. We assessed model performance with two cross-validation (CV) approaches, including hold-out CV, and 10-fold CV.Results: Over 22,000 monthly observations at 1382 monitoring locations were included to estimate the air pollution exposure. The time-varying spatial terms explained 87%, 71%, and 69% of variability with a hold-out cross-validated R-2 of 0.85, 0.62, and 0.62 for PM2.5, PM2.0 and NO2 models, respectively. Models show that meteorological variables, population density, elevation, distance to road, and land cover types were important predictors for air pollution exposure.Conclusions: we have developed a new nationwide model to estimate residence-level air pollution exposures, which can be used in studies of the chronic adverse effects of air pollution.