Abstract
Exposure to fine particulate matter (PM(2.5)) remains a worldwide public health issue. However, epidemiological studies on the chronic health impacts of PM(2.5) in the developing countries are hindered by the lack of monitoring data. Despite the recent development of using satellite remote sensing to predict ground-level PM(2.5) concentrations in China, methods for generating reliable historical PM(2.5) exposure, especially prior to the construction of PM(2.5) monitoring network in 2013, are still very rare. In this study, a high-performance machine-learning model was developed directly at monthly level to estimate PM(2.5) levels in North China Plain. We developed a random forest model using the latest Multi-angle implementation of atmospheric correction (MAIAC) aerosol optical depth (AOD), meteorological parameters, land cover and ground PM(2.5) measurements from 2013 to 2015. A multiple imputation method was applied to fill the missing values of AOD. We used 10-fold cross-validation (CV) to evaluate model performance and a separate time period, January 2016 to December 2016, was used to validate our model's capability of predicting historical PM(2.5) concentrations. The overall model CV R(2) and relative prediction error (RPE) were 0.88 and 18.7%, respectively. Validation results beyond the modeling period (2013-2015) shown that this model can accurately predict historical PM(2.5) concentrations at the monthly (R(2) = 0.74, RPE = 27.6%), seasonal (R(2) = 0.78, RPE = 21.2%) and annual (R(2) = 0.76, RPE = 16.9%) level. The annual mean predicted PM(2.5) concentration from 2013 to 2016 in our study domain was 67.7 μg/m3 and Southern Hebei, Western Shandong and Northern Henan were the most polluted areas. Using this computationally efficient, monthly and high-resolution model, we can provide reliable historical PM(2.5) concentrations for epidemiological studies on PM(2.5) health effects in China.