Reconstruction of PM2.5 Concentrations in East Asia on the Basis of a Wide–Deep Ensemble Machine Learning Framework and Estimation of the Potential Exposure Level from 1981 to 2020

作者：Shuai Yin, Chong Shi, Husi Letu, Akihiko Ito, Huazhe Shang, Dabin Ji, Lei Li, Sude Bilige, Tangzhe Nie, Kunpeng Yi, Meng Guo, Zhongyi Sun, Ao Li · 发表于：Engineering · 年份：2024 · DOI：10.1016/j.eng.2024.09.025 · 被引用次数：5 · 研究领域：Air Quality Monitoring and Forecasting、Air Quality and Health Impacts、Atmospheric chemistry and aerosols

• A wide-deep ensemble framework was proposed for PM 2.5 reconstruction in East Asia. • The framework effectively leveraged the advantages of satellite and model estimations. • The reconstructed PM 2.5 has a good agreement with ground measurements. • The study explored the variation of PM 2.5 and its associated exposure risks in East Asia. Satellite observations are widely used to estimate the concentrations of surface air pollutants, but the temporal coverage of these datasets is relatively short. To overcome this limitation, we propose a wide–deep ensemble machine learning framework to reconstruct the fine particulate matter (PM 2.5 ) dataset of east Asia (EA) over the past four decades (1981–2020). The results indicate that the framework effectively leveraged the advantages of satellite observations (higher accuracy) and model-based estimations (longer temporal coverage) of surface air pollutants. The reconstructed PM 2.5 concentrations agreed well with the ground measurements, with coefficient of determination ( R 2 ) and root-mean-square error (RMSE) values of 0.99 and 1.38 μg·m −3 , respectively, which outperformed the satellite-based PM 2.5 estimates. As more ground measurements were incorporated into the model for training, the average RMSE in Japan and the Korean Peninsula decreased to 0.83 and 1.50 μg·m −3 , respectively. Simultaneously, on the basis of the reconstructed datasets, we investigated the exposure level to PM 2.5 in EA from 1981 to 2020. Since 2000, the ...