Academic Journal
Graph Fourier Deep Learning for Spatiotemporal and Hydrogeological Interpretation of Groundwater Levels in the Yellow River Basin.
| Τίτλος: | Graph Fourier Deep Learning for Spatiotemporal and Hydrogeological Interpretation of Groundwater Levels in the Yellow River Basin. |
|---|---|
| Συγγραφείς: | Wu, Zhenjiang, Yao, Yingying, Guo, Shuitao, Yang, Shuai, He, Xin, Lancia, Michele, Zheng, Chunmiao |
| Πηγή: | Water Resources Research; Nov2025, Vol. 61 Issue 11, p1-20, 20p |
| Θεματικοί όροι: | Hydrogeology, Water table, Deep learning, Watersheds, Spatiotemporal processes, Climate change, Prediction models |
| Γεωγραφικοί όροι: | Yellow River (China), China |
| Περίληψη: | Groundwater systems exhibit strong spatial heterogeneity and nonlinear dynamics, complicating the parameterization of physically based hydrological models and limiting adaptability at large scales. Deep learning (DL) offers a promising data‐driven alternative but faces challenges in prediction accuracy and interpretability. This study proposes a novel Graph Fourier Network (GFN) model that integrates Fourier domain modeling, dynamic graph structure generation, and multi‐head graph attention. Hydrogeological units are embedded as prior information, leveraging data from 473 groundwater monitoring wells across the Yellow River Basin (YRB), China. We show that the proposed GFN model significantly outperforms baseline models, including Long Short‐Term Memory (LSTM), Extreme Gradient Boosting (XGBoost) and Spatio‐Temporal LSTM (ST‐LSTM), achieving an R2 of up to 0.90. By incorporating hydrogeological prior knowledge, the GFN model enhances the extrapolation capability of groundwater level predictions, particularly for extended lead times up to 25 days. The predicted groundwater level dynamics show lagged and nonlinear responses to climatic and ecological drivers, with their spatial heterogeneity shaped by hydrogeological conditions. These response patterns include the prolonged influence of precipitation, and spatially varying effects of evapotranspiration and vegetation across hydrogeological units. These findings highlight the GFN model informed by hydrogeological information effectively captures characteristics of regional groundwater system. This study provides a fast and reliable approach for short‐ and long‐term regional groundwater prediction under varying hydrogeological conditions. Plain Language Summary: Groundwater levels are crucial for water resource management but are difficult to predict due to complex geological structures and climate variability. In this study, we propose a new deep learning model for basin‐scale groundwater level prediction. This model incorporates information from hydrogeological zoning maps representing local groundwater conditions, learns frequency patterns from well records, and adaptively modulates spatial influence among wells. We applied the model to the Yellow River Basin in China, a globally representative watershed, using data from 473 monitoring wells for validation and testing. The model achieved excellent predictive performance, capturing approximately 90% of the observed patterns and trends, and outperformed existing deep learning and machine learning approaches. Incorporating hydrogeological information strengthens performance at longer time steps, providing reliable predictions up to 25 days. The model also reveals the influence of hydroclimatic drivers on groundwater level fluctuations. For example, precipitation exerts a sustained influence on groundwater levels, whereas the effects of evapotranspiration and vegetation vary spatially depending on local geology. Our model provides a fast and reliable AI‐based tool for groundwater level prediction under diverse hydrogeological conditions, supporting both short‐ and long‐term water resource planning. Key Points: The Graph Fourier Network (GFN) model outperforms deep learning baselines in predicting basin‐scale groundwater levelsThe hydrogeological units embedded in the GFN enhance the extrapolation capability of groundwater level predictionsThe GFN model reveals groundwater dynamics in the Yellow River Basin are driven by spatial hydroclimatic and ecological factors over time [ABSTRACT FROM AUTHOR] |
| Copyright of Water Resources Research is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Βάση Δεδομένων: | Biomedical Index |
| FullText | Text: Availability: 0 CustomLinks: – Url: https://resolver.ebsco.com/c/fiv2js/result?sid=EBSCO:edm&genre=article&issn=00431397&ISBN=&volume=61&issue=11&date=20251101&spage=1&pages=1-20&title=Water Resources Research&atitle=Graph%20Fourier%20Deep%20Learning%20for%20Spatiotemporal%20and%20Hydrogeological%20Interpretation%20of%20Groundwater%20Levels%20in%20the%20Yellow%20River%20Basin.&aulast=Wu%2C%20Zhenjiang&id=DOI:10.1029/2025WR041215 Name: Full Text Finder (for New FTF UI) (ns324271) Category: fullText Text: Full Text Finder MouseOverText: Full Text Finder |
|---|---|
| Header | DbId: edm DbLabel: Biomedical Index An: 189523779 RelevancyScore: 1023 AccessLevel: 6 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 1023.09326171875 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: Graph Fourier Deep Learning for Spatiotemporal and Hydrogeological Interpretation of Groundwater Levels in the Yellow River Basin. – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Wu%2C+Zhenjiang%22">Wu, Zhenjiang</searchLink><br /><searchLink fieldCode="AR" term="%22Yao%2C+Yingying%22">Yao, Yingying</searchLink><br /><searchLink fieldCode="AR" term="%22Guo%2C+Shuitao%22">Guo, Shuitao</searchLink><br /><searchLink fieldCode="AR" term="%22Yang%2C+Shuai%22">Yang, Shuai</searchLink><br /><searchLink fieldCode="AR" term="%22He%2C+Xin%22">He, Xin</searchLink><br /><searchLink fieldCode="AR" term="%22Lancia%2C+Michele%22">Lancia, Michele</searchLink><br /><searchLink fieldCode="AR" term="%22Zheng%2C+Chunmiao%22">Zheng, Chunmiao</searchLink> – Name: TitleSource Label: Source Group: Src Data: Water Resources Research; Nov2025, Vol. 61 Issue 11, p1-20, 20p – Name: Subject Label: Subject Terms Group: Su Data: <searchLink fieldCode="DE" term="%22Hydrogeology%22">Hydrogeology</searchLink><br /><searchLink fieldCode="DE" term="%22Water+table%22">Water table</searchLink><br /><searchLink fieldCode="DE" term="%22Deep+learning%22">Deep learning</searchLink><br /><searchLink fieldCode="DE" term="%22Watersheds%22">Watersheds</searchLink><br /><searchLink fieldCode="DE" term="%22Spatiotemporal+processes%22">Spatiotemporal processes</searchLink><br /><searchLink fieldCode="DE" term="%22Climate+change%22">Climate change</searchLink><br /><searchLink fieldCode="DE" term="%22Prediction+models%22">Prediction models</searchLink> – Name: SubjectGeographic Label: Geographic Terms Group: Su Data: <searchLink fieldCode="DE" term="%22Yellow+River+%28China%29%22">Yellow River (China)</searchLink><br /><searchLink fieldCode="DE" term="%22China%22">China</searchLink> – Name: Abstract Label: Abstract Group: Ab Data: Groundwater systems exhibit strong spatial heterogeneity and nonlinear dynamics, complicating the parameterization of physically based hydrological models and limiting adaptability at large scales. Deep learning (DL) offers a promising data‐driven alternative but faces challenges in prediction accuracy and interpretability. This study proposes a novel Graph Fourier Network (GFN) model that integrates Fourier domain modeling, dynamic graph structure generation, and multi‐head graph attention. Hydrogeological units are embedded as prior information, leveraging data from 473 groundwater monitoring wells across the Yellow River Basin (YRB), China. We show that the proposed GFN model significantly outperforms baseline models, including Long Short‐Term Memory (LSTM), Extreme Gradient Boosting (XGBoost) and Spatio‐Temporal LSTM (ST‐LSTM), achieving an R2 of up to 0.90. By incorporating hydrogeological prior knowledge, the GFN model enhances the extrapolation capability of groundwater level predictions, particularly for extended lead times up to 25 days. The predicted groundwater level dynamics show lagged and nonlinear responses to climatic and ecological drivers, with their spatial heterogeneity shaped by hydrogeological conditions. These response patterns include the prolonged influence of precipitation, and spatially varying effects of evapotranspiration and vegetation across hydrogeological units. These findings highlight the GFN model informed by hydrogeological information effectively captures characteristics of regional groundwater system. This study provides a fast and reliable approach for short‐ and long‐term regional groundwater prediction under varying hydrogeological conditions. Plain Language Summary: Groundwater levels are crucial for water resource management but are difficult to predict due to complex geological structures and climate variability. In this study, we propose a new deep learning model for basin‐scale groundwater level prediction. This model incorporates information from hydrogeological zoning maps representing local groundwater conditions, learns frequency patterns from well records, and adaptively modulates spatial influence among wells. We applied the model to the Yellow River Basin in China, a globally representative watershed, using data from 473 monitoring wells for validation and testing. The model achieved excellent predictive performance, capturing approximately 90% of the observed patterns and trends, and outperformed existing deep learning and machine learning approaches. Incorporating hydrogeological information strengthens performance at longer time steps, providing reliable predictions up to 25 days. The model also reveals the influence of hydroclimatic drivers on groundwater level fluctuations. For example, precipitation exerts a sustained influence on groundwater levels, whereas the effects of evapotranspiration and vegetation vary spatially depending on local geology. Our model provides a fast and reliable AI‐based tool for groundwater level prediction under diverse hydrogeological conditions, supporting both short‐ and long‐term water resource planning. Key Points: The Graph Fourier Network (GFN) model outperforms deep learning baselines in predicting basin‐scale groundwater levelsThe hydrogeological units embedded in the GFN enhance the extrapolation capability of groundwater level predictionsThe GFN model reveals groundwater dynamics in the Yellow River Basin are driven by spatial hydroclimatic and ecological factors over time [ABSTRACT FROM AUTHOR] – Name: Abstract Label: Group: Ab Data: <i>Copyright of Water Resources Research is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.) |
| PLink | https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edm&AN=189523779 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1029/2025WR041215 Languages: – Code: eng Text: English PhysicalDescription: Pagination: PageCount: 20 StartPage: 1 Subjects: – SubjectFull: Yellow River (China) Type: general – SubjectFull: China Type: general – SubjectFull: Hydrogeology Type: general – SubjectFull: Water table Type: general – SubjectFull: Deep learning Type: general – SubjectFull: Watersheds Type: general – SubjectFull: Spatiotemporal processes Type: general – SubjectFull: Climate change Type: general – SubjectFull: Prediction models Type: general Titles: – TitleFull: Graph Fourier Deep Learning for Spatiotemporal and Hydrogeological Interpretation of Groundwater Levels in the Yellow River Basin. Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Wu, Zhenjiang – PersonEntity: Name: NameFull: Yao, Yingying – PersonEntity: Name: NameFull: Guo, Shuitao – PersonEntity: Name: NameFull: Yang, Shuai – PersonEntity: Name: NameFull: He, Xin – PersonEntity: Name: NameFull: Lancia, Michele – PersonEntity: Name: NameFull: Zheng, Chunmiao IsPartOfRelationships: – BibEntity: Dates: – D: 01 M: 11 Text: Nov2025 Type: published Y: 2025 Identifiers: – Type: issn-print Value: 00431397 Numbering: – Type: volume Value: 61 – Type: issue Value: 11 Titles: – TitleFull: Water Resources Research Type: main |
| ResultId | 1 |