High-resolution spatiotemporal prediction of PM2.5 concentration based on mobile monitoring and deep learning

Yi-Zhou Wang; Hong-Di He; Hai-Chao Huang; Jin-Ming Yang; Zhong-Ren Peng

doi:10.1016/j.envpol.2024.125342

High-resolution spatiotemporal prediction of PM_2.5 concentration based on mobile monitoring and deep learning

Environ Pollut. 2024 Nov 18;364(Pt 2):125342. doi: 10.1016/j.envpol.2024.125342. Online ahead of print.

Authors

Yi-Zhou Wang¹, Hong-Di He², Hai-Chao Huang³, Jin-Ming Yang⁴, Zhong-Ren Peng⁵

Affiliations

¹ Center for Intelligent Transportation Systems and Unmanned Aerial Systems Applications, State-Key Laboratory of Ocean Engineering, School of Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China; Data-Driven Management Decision Making Lab, Shanghai Jiao Tong University, Shanghai 200240, China.
² Center for Intelligent Transportation Systems and Unmanned Aerial Systems Applications, State-Key Laboratory of Ocean Engineering, School of Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China. Electronic address: hongdihe@sjtu.edu.cn.
³ Center for Intelligent Transportation Systems and Unmanned Aerial Systems Applications, State-Key Laboratory of Ocean Engineering, School of Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China.
⁴ MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University, Shanghai, 200240, China.
⁵ International Center for Adaptation Planning and Design, College of Design, Construction and Planning, University of Florida, Florida, 32611-5706, USA; Healthy Building Research Center, Ajman University, Ajman, United Arab Emirates.

PMID: 39566710
DOI: 10.1016/j.envpol.2024.125342

Abstract

Obtaining the high-resolution distribution characteristics of urban air pollutants is crucial for effective pollution control and public health. In order to fulfill it, mobile monitoring offers a novel and practical approach compared to traditional fixed monitoring methods. However, the sparsity of mobile monitoring data still makes it a challenge to recover the high-resolution pollutant concentration across an entire area. To tackle the sparsity issue and fulfill a prediction of the spatiotemporal distribution of PM_2.5, a high-resolution urban PM_2.5 prediction method was proposed based on mobile monitoring data in this study. This method enables prediction with a spatial resolution of 500m × 500m and a temporal resolution of 1 h. First, a Light Gradient Boosting Machine (LightGBM) was trained using mobile monitoring of PM_2.5 concentration and exogenous features to obtain complete spatiotemporal PM_2.5 concentration. Second, a model consisting of Convolutional Neural Network and Transformer (CNN-Transformer) with a customised loss function was established to predict high-resolution PM_2.5 concentration based on complete spatiotemporal data. The method was validated using real-world data collected from Cangzhou, China. The numerical results from cross-validation showed an R² of 0.925 for imputation and 0.887 for prediction, demonstrating this method is suitable for high-resolution spatiotemporal prediction of PM_2.5 concentration based on mobile monitoring data.

Keywords: CNN-Transformer; Customised loss function; High-resolution; Mobile monitoring; PM(2.5).