Abstract:In recent years, ground-level ozone has become a severe ambient pollutant in major urban areas of China, which has adverse impacts on population health. However, in-situ measurements of the ozone concentration before 2013 in China are quite scarce, which cannot facilitate the assessment of the long-term trends and effects of ozone pollution. In this study, we used daily maximum 8-hour average (MDA8) ozone observations from 2013 to 2017 combined with concurrent ozone retrievals, aerosol reanalysis, meteorological parameters, and land-use data to establish a nationwide MDA8 prediction model based on the eXtreme Gradient Boosting (XGBoost) algorithm. The model achieves high prediction accuracy compared with other studies, with R2 values for the by-year, site-based, and sample-based cross-validation (CV) schemes of 0.61, 0.64, and 0.78, respectively, at the daily level. External testing with regional measurements from 2005 to 2012 and nationwide data in 2018 have shown that the model is robust and reliable for historical data prediction, with external model testing R2 values ranging from 0.60 to 0.87 at the month level in different years. Using the final estimator, we obtained nationwide monthly mean ozone concentrations from 2005 to 2012 and daily MDA8 ozone concentrations from 2013 to 2017 at a resolution of 0.1° × 0.1°. According to the average number of days exceeding the standard and the average of the 90th percentile of the MDA8 ozone concentrations, the Beijing-Tianjin-Hebei (BTH), the Yangtze River Delta, the Pearl River Delta, the Jianghan Plain, the Sichuan Basin, and the Northeast Plain regions were identified as pollution hotspots. During the research period, the overall ozone levels fluctuated slightly, and their trends were not spatially continuous. There was a significant increasing trend in the BTH region by 1.37 (95% CI: 0.46,2.29) μg/m3/year between 2013 and 2017. In 2017, 26.24% of the population lived in areas exceeding the Chinese grade II national air quality standard, which shows that ozone pollution has posed an obvious threat to population health in China. Our products will provide reliable support for future long-term nationwide health impact studies and policy-making for pollution control and prevention.

What problem does this paper attempt to address?

This paper attempts to address the issue of the long-term spatiotemporal distribution patterns of surface ozone concentration in China. Specifically, due to the very limited surface ozone observation data in China before 2013, it has been difficult to assess long-term trends and impacts. Therefore, this paper utilizes daily maximum 8-hour average (MDA8) ozone observation data from 2013 to 2017, combined with ozone retrievals, aerosol reanalysis, meteorological parameters, and land use data from the same period, to establish a national MDA8 prediction model based on the XGBoost algorithm. This model aims to fill the historical data gap and provide high-resolution surface ozone concentration estimates nationwide from 2005 to 2017, thereby supporting future health impact studies and pollution control policy formulation. ### Main Issues: 1. **Lack of Historical Observation Data**: Surface ozone observation data before 2013 is very limited, making it difficult to effectively assess long-term trends and impacts. 2. **Need for High-Precision Long-Term Estimates**: High-precision long-term surface ozone concentration estimates are needed for environmental health research and pollution control policy formulation. 3. **Model Stability and Reliability**: A model capable of stably predicting historical data needs to be established, and its reliability must be ensured through external validation. ### Solutions: 1. **Data Integration**: Utilize observation data from 2013 to 2017, combined with satellite retrievals, aerosol reanalysis, meteorological observations, and land use data. 2. **Machine Learning Model**: Use the XGBoost algorithm to establish a prediction model, which performs well in handling complex nonlinear relationships. 3. **Cross-Validation**: Employ by-year cross-validation (by-year CV), site-based cross-validation (site-based CV), and sample-based cross-validation (sample-based CV) strategies to ensure the model's stability and generalization ability. 4. **External Validation**: Use regional data from 2005 to 2012 and national data from 2018 for external validation to further test the model's reliability and generalization ability. ### Main Findings: 1. **Model Performance**: The model performs well on daily, monthly, and seasonal scales, with particularly small errors on monthly and seasonal scales. 2. **Spatiotemporal Distribution**: Regions such as Beijing-Tianjin-Hebei (BTH), the Yangtze River Delta, the Pearl River Delta, the Jianghan Plain, the Sichuan Basin, and the Northeast Plain are identified as pollution hotspots. 3. **Long-Term Trends**: From 2013 to 2017, ozone concentration in the BTH region increased significantly, with an annual growth rate of 1.37 μg/m³. 4. **Health Impact**: In 2017, 26.24% of the population lived in areas exceeding China's secondary air quality standard, indicating a significant threat to population health from ozone pollution. Through these methods and findings, this study provides important data support for future atmospheric pollution research and policy formulation.

Spatiotemporal distributions of surface ozone levels in China from 2005 to 2017: A machine learning approach

Near Real-Time Distribution of Ozone in China from 2013 to 2020 and Agricultural Impacts

Satellite-Based Long-Term Spatiotemporal Patterns of Surface Ozone Concentrations in China: 2005–2019

Evaluating the spatiotemporal ozone characteristics with high-resolution predictions in mainland China, 2013–2019

Uncovering the evolution of ozone pollution in China: A spatiotemporal characteristics reconstruction from 1980 to 2021

Spatiotemporal Variations of Air Pollutants and Ozone Prediction Using Machine Learning Algorithms in the Beijing-Tianjin-Hebei Region from 2014 to 2021.

Estimating Monthly Surface Ozone Using Multi-Source Satellite Products in China Based on Deep Forest Model

Full-coverage mapping and spatiotemporal variations of ground-level ozone (O3) pollution from 2013 to 2020 across China

Multidecadal ozone trends in China and implications for human health and crop yields: a hybrid approach combining a chemical transport model and machine learning

Machine Learning‐Driven Spatiotemporal Analysis of Ozone Exposure and Health Risks in China

Estimating Spatiotemporal Variation in Ambient Ozone Exposure during 2013–2017 Using a Data-Fusion Model

Understanding the spatial and seasonal variation of the ground-level ozone in Southeast China with an interpretable machine learning and multi-source remote sensing

Estimating ground-level high-resolution ozone concentration across China using a stacked machine-learning method

Spatio-Temporal Prediction of Ground-Level Ozone Concentration Based on Bayesian Maximum Entropy by Combining Monitoring and Satellite Data

Prediction of daily PM 2.5 and ozone based on high-density weather stations in China: Nonlinear effects of meteorology, human and ecosystem health risks

Analysis of severe ozone–related human health and weather influence over China in 2019 based on a high-resolution dataset

Machine-learning-based corrections of CMIP6 historical surface ozone in China during 1950-2014

Development of a High-Performance Machine Learning Model to Predict Ground Ozone Pollution in Typical Cities of China.

Estimation of near-surface ozone concentration and analysis of main weather situation in China based on machine learning model and Himawari-8 TOAR data

Distinct spatiotemporal variation patterns of surface ozone in China due to diverse influential factors

Ozone Concentration Estimation and Meteorological Impact Quantification in the Beijing‐Tianjin‐Hebei Region Based on Machine Learning Models