A PM2.5 spatiotemporal prediction model based on mixed graph convolutional GRU and self-attention network

IF 7.6 2区 环境科学与生态学 Q1 ENVIRONMENTAL SCIENCES
Zhao Guyu, Yang Xiaoyuan, Shi Jiansen, He Hongdou, Wang Qian
{"title":"A PM2.5 spatiotemporal prediction model based on mixed graph convolutional GRU and self-attention network","authors":"Zhao Guyu, Yang Xiaoyuan, Shi Jiansen, He Hongdou, Wang Qian","doi":"10.1016/j.envpol.2025.125748","DOIUrl":null,"url":null,"abstract":"The increase in atmospheric pollution has made it essential to develop accurate models for predicting pollutant concentrations. The current researches have faced challenges such as the neglect of significant information selection from local and neighboring stations, as well as insufficient attention to long-term historical data patterns. Therefore, this paper proposes a spatiotemporal prediction model called MGCGRU-SAN, which leverages long-term historical data to predict PM<sub>2.5</sub> concentration values across multiple stations and multiple time steps in the future. Firstly, we employ the Mixed Graph Convolutional GRU(MGCGRU) module to capture the spatiotemporal dependencies in short-term historical time series from various stations. Secondly, the long-term PM<sub>2.5</sub> historical time series (e.g. one week) is divided into uniformly sized segments and fed into the Self-Attention Network(SAN) module to capture the long-term potential temporal patterns. These enable the model to not only capture short-term fluctuations, but also identify and track long-term temporal patterns and trends in the prediction process. Finally, we conduct extensive comparative and ablation experiments using historical air pollutant and meteorological data from the Beijing-Tianjin-Hebei region. The experimental results demonstrate that the model, after capturing the long-term latent temporal patterns, achieve improvements of 9.62%, 6.33%, and 4.98% in the RSE, MAE, and RMSE evaluation metrics during multi-step prediction. Overall, the model outperforms the best baseline model by an average of 8.34%, 6.12%,4.06%, and 2.60% in RSE, MAE, RMSE, and Correlation metrics, respectively, showing superior performance in multi-station long-term predictions.","PeriodicalId":311,"journal":{"name":"Environmental Pollution","volume":"9 1","pages":""},"PeriodicalIF":7.6000,"publicationDate":"2025-02-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Environmental Pollution","FirstCategoryId":"93","ListUrlMain":"https://doi.org/10.1016/j.envpol.2025.125748","RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0

Abstract

The increase in atmospheric pollution has made it essential to develop accurate models for predicting pollutant concentrations. The current researches have faced challenges such as the neglect of significant information selection from local and neighboring stations, as well as insufficient attention to long-term historical data patterns. Therefore, this paper proposes a spatiotemporal prediction model called MGCGRU-SAN, which leverages long-term historical data to predict PM2.5 concentration values across multiple stations and multiple time steps in the future. Firstly, we employ the Mixed Graph Convolutional GRU(MGCGRU) module to capture the spatiotemporal dependencies in short-term historical time series from various stations. Secondly, the long-term PM2.5 historical time series (e.g. one week) is divided into uniformly sized segments and fed into the Self-Attention Network(SAN) module to capture the long-term potential temporal patterns. These enable the model to not only capture short-term fluctuations, but also identify and track long-term temporal patterns and trends in the prediction process. Finally, we conduct extensive comparative and ablation experiments using historical air pollutant and meteorological data from the Beijing-Tianjin-Hebei region. The experimental results demonstrate that the model, after capturing the long-term latent temporal patterns, achieve improvements of 9.62%, 6.33%, and 4.98% in the RSE, MAE, and RMSE evaluation metrics during multi-step prediction. Overall, the model outperforms the best baseline model by an average of 8.34%, 6.12%,4.06%, and 2.60% in RSE, MAE, RMSE, and Correlation metrics, respectively, showing superior performance in multi-station long-term predictions.

Abstract Image

求助全文
约1分钟内获得全文 求助全文
来源期刊
Environmental Pollution
Environmental Pollution 环境科学-环境科学
CiteScore
16.00
自引率
6.70%
发文量
2082
审稿时长
2.9 months
期刊介绍: Environmental Pollution is an international peer-reviewed journal that publishes high-quality research papers and review articles covering all aspects of environmental pollution and its impacts on ecosystems and human health. Subject areas include, but are not limited to: • Sources and occurrences of pollutants that are clearly defined and measured in environmental compartments, food and food-related items, and human bodies; • Interlinks between contaminant exposure and biological, ecological, and human health effects, including those of climate change; • Contaminants of emerging concerns (including but not limited to antibiotic resistant microorganisms or genes, microplastics/nanoplastics, electronic wastes, light, and noise) and/or their biological, ecological, or human health effects; • Laboratory and field studies on the remediation/mitigation of environmental pollution via new techniques and with clear links to biological, ecological, or human health effects; • Modeling of pollution processes, patterns, or trends that is of clear environmental and/or human health interest; • New techniques that measure and examine environmental occurrences, transport, behavior, and effects of pollutants within the environment or the laboratory, provided that they can be clearly used to address problems within regional or global environmental compartments.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信