Random forest model based fine scale spatiotemporal O3 trends in the Beijing-Tianjin-Hebei region in China, 2010 to 2017

Ambient ozone (O3) concentrations have shown an upward trend in China and its health hazards have also been recognized in recent years. High-resolution exposure data based on statistical models are needed. Our study aimed to build high-performance random forest (RF) models based on training data from 2013 to 2017 in the Beijing-Tianjin-Hebei (BTH) region in China at a 0.01 ° × 0.01 ° resolution, and estimated daily maximum 8h average O3 (O3-8hmax) concentration, daily average O3 (O3-mean) concentration, and daily maximum 1h O3 (O3-1hmax) concentration from 2010 to 2017. Model features included meteorological variables, chemical transport model output variables, geographic variables, and population data. The test-R2 of sample-based O3-8hmax, O3-mean and O3-1hmax models were all greater than 0.80, while the R2 of site-based and date-based model were 0.68–0.87. From 2010 to 2017, O3-8hmax, O3-mean, and O3-1hmax concentrations in the BTH region increased by 4.18 μg/m3, 0.11 μg/m3, and 4.71 μg/m3, especially in more developed regions. Due to the influence of weather conditions, which showed high contribution to the model, the long-term spatial distribution of O3 concentrations indicated a similar pattern as altitude, where high concentration levels were distributed in regions with higher altitude. [Display omitted] •Random forest models of O3 with 0.01° resolution in BTH region were built.•Three metrics of daily O3 concentration during 2010–2017 were simulated.•Reliable model performance was indicated by test-R2 over 0.8.•Long-term temporal and spatial O3 concentration trend were provided. Main finding: Multi-variable RF models of O3 with 0.01° resolution in Beijing-Tianjin-Hebei in China were built, with test-R2 over 0.80.