Estimating Ground‐Level PM2.5 by Fusing Satellite and Station Observations: A Geo‐Intelligent Deep Learning Approach

• Published in: Geophysical research letters Vol. 44; no. 23; pp. 11,985 - 11,993
• Main Authors: Li, Tongwen, Shen, Huanfeng, Yuan, Qiangqiang, Zhang, Xuechen, Zhang, Liangpei
• Format: Journal Article
• Language: English
• Published: Washington John Wiley & Sons, Inc 16.12.2017
• Subjects: Air monitoring; Air pollution; Architecture; Deep learning; Environmental monitoring; geo‐intelligent; Neural networks; Observational learning; Particulate matter; Particulate matter monitoring; PM2.5; Pollution; Pollution monitoring; Satellite observation; satellite remote sensing; Satellites
• ISSN: 0094-8276; 1944-8007
• DOI: 10.1002/2017GL075710

Fusing satellite observations and station measurements to estimate ground‐level PM2.5 is promising for monitoring PM2.5 pollution. A geo‐intelligent approach, which incorporates geographical correlation into an intelligent deep learning architecture, is developed to estimate PM2.5. Specifically, it considers geographical distance and spatiotemporally correlated PM2.5 in a deep belief network (denoted as Geoi‐DBN). Geoi‐DBN can capture the essential features associated with PM2.5 from latent factors. It was trained and tested with data from China in 2015. The results show that Geoi‐DBN performs significantly better than the traditional neural network. The out‐of‐sample cross‐validation R2 increases from 0.42 to 0.88, and RMSE decreases from 29.96 to 13.03 μg/m3. On the basis of the derived PM2.5 distribution, it is predicted that over 80% of the Chinese population live in areas with an annual mean PM2.5 of greater than 35 μg/m3. This study provides a new perspective for air pollution monitoring in large geographic regions. Key Points A deep learning architecture is established to estimate ground‐level PM2.5 by fusing satellite and station observations A geo‐intelligent model is developed to incorporate geographical correlation into deep learning for performance improvement This model shows a superior estimation accuracy (R2 = 0.88, RMSE = 13.03 μg/m3) at national scale