Satellite-Based Estimates of Daily NO2 Exposure in China Using Hybrid Random Forest and Spatiotemporal Kriging Model

Abstract

A novel model named random-forest-spatiotemporal-kriging (RF-STK) was developed to estimate the daily ambient NO 2 concentrations across China during 2013–2016 based on the satellite retrievals and geographic covariates. The RF-STK model showed good prediction performance, with cross-validation R 2 = 0.62 (RMSE = 13.3 μg/m 3 ) for daily and R 2 = 0.73 (RMSE = 6.5 μg/m 3 ) for spatial predictions. The nationwide population-weighted multiyear average of NO 2 was predicted to be 30.9 ± 11.7 μg/m 3 (mean ± standard deviation), with a slowly but significantly decreasing trend at a rate of −0.88 ± 0.38 μg/m 3 /year. Among the main economic zones of China, the Pearl River Delta showed the fastest decreasing rate of −1.37 μg/m 3 /year, while the Beijing-Tianjin Metro did not show a temporal trend ( P = 0.32). The population-weighted NO 2 was predicted to be the highest in North China (40.3 ± 10.3 μg/m 3 ) and lowest in Southwest China (24.9 ± 9.4 μg/m 3 ). Approximately 25% of the population lived in nonattainment areas with annual-average NO 2 > 40 μg/m 3 . A piecewise linear function with an abrupt point around 100 people/km 2 characterized the relationship between the population density and the NO 2, indicating a threshold of aggravated NO 2 pollution due to urbanization. Leveraging the ground-level NO 2 observations, this study fills the gap of statistically modeling nationwide NO 2 in China, and provides essential data for epidemiological research and air quality management.