Construction of a virtual PM_2.5 observation network in China based on high-density surface meteorological observations using the Extreme Gradient Boosting model

With increasing public concerns on air pollution in China, there is a demand for long-term continuous PM_2.5 datasets. However, it was not until the end of 2012 that China established a national PM_2.5 observation network. Before that, satellite-retrieved aerosol optical depth (AOD) was frequently used as a primary predictor to estimate surface PM_2.5. Nevertheless, satellite-retrieved AOD often encounter incomplete daily coverage due to its sampling frequency and interferences from cloud, which greatly affect the representation of these AOD-based PM_2.5. Here, we constructed a virtual ground-based PM_2.5 observation network at 1180 meteorological sites across China using the Extreme Gradient Boosting (XGBoost) model with high-density meteorological observations as major predictors. Cross-validation of the XGBoost model showed strong robustness and high accuracy in its estimation of the daily (monthly) PM_2.5 across China in 2018, with R², root-mean-square error (RMSE) and mean absolute error values of 0.79 (0.92), 15.75 μg/m³ (6.75 μg/m³) and 9.89 μg/m³ (4.53 μg/m³), respectively. Meanwhile, we find that surface visibility plays the dominant role in terms of the relative importance of variables in the XGBoost model, accounting for 39.3% of the overall importance. We then use meteorological and PM_2.5 data in the year 2017 to assess the predictive capability of the model. Results showed that the XGBoost model is capable to accurately hindcast historical PM_2.5 at monthly (R² = 0.80, RMSE = 14.75 μg/m³), seasonal (R² = 0.86, RMSE = 12.28 μg/m³), and annual (R² = 0.81, RMSE = 10.10 μg/m³) mean levels. In general, the newly constructed virtual PM_2.5 observation network based on high-density surface meteorological observations using the Extreme Gradient Boosting model shows great potential in reconstructing historical PM_2.5 at ~1000 meteorological sites across China. It will be of benefit to filling gaps in AOD-based PM_2.5 data, as well as to other environmental studies including epidemiology.