National Engineering Research Center for Geoinformatics, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China; University of Chinese Academy of Sciences, Beijing 100049, China.
National Engineering Research Center for Geoinformatics, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China; University of Chinese Academy of Sciences, Beijing 100049, China; Zhejiang-CAS Application Center for Geoinformatics, Jiashan 314100, China.
Sci Total Environ. 2021 Mar 25;762:144502. doi: 10.1016/j.scitotenv.2020.144502. Epub 2020 Dec 14.
Assessing short-term exposure to PM requires the concentration distribution at a high spatiotemporal resolution. Abundant researches have derived the daily predictions of fine particles, but estimating hourly PM is still a challenge restrained by the input data. The recent aerosol optical depth (AOD) product from Himawari-8 provides hourly satellite observations informative to modelling. In this study, we developed separate random forest models with and without AOD and combined the estimates to obtain a full-coverage hourly PM distribution. 10-fold cross validation R ranged from 0.92 to 0.95 and root mean square errors from 14.1 to 16.9 μg/m, indicating the good model performance. Spatial convolutional layers of PM measurements and temporal accumulation effects of meteorological features were added into the model. They turned out to be of the most important predictors and improved the performance significantly. Finally, we mapped hourly PM at a 1-km resolution in Beijing during a pollution episode in 2019 and studied the pollution pattern. The study proposed a method to obtain 24-h full-coverage hourly PM estimates which are useful for acute exposure assessment in epidemiological researches.
评估 PM 的短期暴露需要高时空分辨率的浓度分布。大量研究已经推导出了细颗粒物的日预测值，但由于输入数据的限制，估计每小时的 PM 仍然是一个挑战。来自 Himawari-8 的最新气溶胶光学厚度 (AOD) 产品提供了对建模有帮助的每小时卫星观测。在这项研究中，我们分别开发了带有和不带有 AOD 的随机森林模型，并结合这些估计值来获得全面的每小时 PM 分布。10 折交叉验证 R 范围从 0.92 到 0.95，均方根误差从 14.1 到 16.9μg/m，表明模型性能良好。PM 测量的空间卷积层和气象特征的时间累积效应被添加到模型中。它们被证明是最重要的预测因子，显著提高了模型性能。最后，我们在 2019 年的一次污染事件中以 1 公里的分辨率绘制了北京每小时的 PM 图，并研究了污染模式。该研究提出了一种获取 24 小时全面覆盖的每小时 PM 估计值的方法，这对于流行病学研究中的急性暴露评估非常有用。
