Physics-guided deep learning model for daily groundwater table maps estimation using passive surface-wave dispersion
Monitoring groundwater tables (GWTs) is challenging due to limited spatial and temporal observations. This study presents an innovative approach utilizing supervised deep learning, specifically a Multilayer Perceptron (MLP), and continuous passive-Multichannel Analysis of Surface Waves (passive-MASW...
Gespeichert in:
Veröffentlicht in: | Water resources research 2024-04 |
---|---|
Hauptverfasser: | , , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | Water resources research |
container_volume | |
creator | Teixeira, José Cunha Bodet, Ludovic Rivière, Agnès Hallier, Amélie Gesret, Alexandrine Dangeard, Marine Dhemaied, Amine Gaboriau, Joséphine Boisson |
description | Monitoring groundwater tables (GWTs) is challenging due to limited spatial and temporal observations. This study presents an innovative approach utilizing supervised deep learning, specifically a Multilayer Perceptron (MLP), and continuous passive-Multichannel Analysis of Surface Waves (passive-MASW) for constructing 2D GWT level maps. The study site, geologically well-constrained, features two 20-meter-deep piezometers and a permanent 2D geophone array capturing train-induced surface waves. For each point of the 2D array, dispersion curves (DCs), displaying Rayleigh-wave phase velocities (V_R) across a frequency range of 5 to 50 Hz, have been computed each day between December 2022 and September 2023. In the present study, these DCs are resampled in wavelengths ranging from 4 to 15~m in order to focus the monitoring on the expected GWT levels (between -1 and -5 m). Nine months of daily V_R data around one of the two piezometers is used to train the MLP model. GWT levels are then estimated across the entire geophone array, generating daily 2D GWT maps. Model’s performance is tested through cross-validation and comparisons with GWT level data at the second piezometer. Model’s efficiency is quantified with the root-mean-square error (RMSE) and the coefficient of determination (R²). The R² is estimated at 80% for data surrounding the training piezometer, and at 68% for data surrounding the test piezometer. Additionally, the RMSE is impressively low at 0.03 m at both piezometers. Results showcase the effectiveness of DL in estimating GWT level maps from passive-MASW data, offering a practical and efficient monitoring solution across broader spatial extents. |
doi_str_mv | 10.22541/essoar.171322609.99979575/v1 |
format | Article |
fullrecord | <record><control><sourceid>hal</sourceid><recordid>TN_cdi_hal_primary_oai_HAL_hal_04716694v1</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>oai_HAL_hal_04716694v1</sourcerecordid><originalsourceid>FETCH-LOGICAL-h891-73e5cf7d4a0e81f66aa1b518963d08aafe49206c91ca89d3e3972e544eb8437f3</originalsourceid><addsrcrecordid>eNotT8FOwzAUixBIjME_5MKBQ7akSZPmOE3AkCbBYffqrXndgrq2yms37e8pgpNly7Zsxp6VXGRZbtQSiTpIC-WUzjIr_cJ773zu8uVZ3bCZ8sYI552-ZTMpjRZKe3fPHoi-pVQmt27Ghq_jlWJF4jDGgIEHxJ43CKmN7YGfuoANr7vEA8Tmyg-pG9twgQETH2DfID9BTxxpiCcYYtfykX5zPRDFM3IaUw0VigtMJETqMdHkemR3NTSET_84Z7u31916I7af7x_r1VYcC6-E05hXtQsGJBaqthZA7XNVeKuDLABqND6TtvKqgsIHjdO7DHNjcF8Y7Wo9Zy9_tUdoyj5NE9O17CCWm9W2_NWkccpab85K_wAWyGVy</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Physics-guided deep learning model for daily groundwater table maps estimation using passive surface-wave dispersion</title><source>Access via Wiley Online Library</source><source>Wiley-Blackwell AGU Digital Library</source><source>Wiley Online Library (Open Access Collection)</source><creator>Teixeira, José Cunha ; Bodet, Ludovic ; Rivière, Agnès ; Hallier, Amélie ; Gesret, Alexandrine ; Dangeard, Marine ; Dhemaied, Amine ; Gaboriau, Joséphine Boisson</creator><creatorcontrib>Teixeira, José Cunha ; Bodet, Ludovic ; Rivière, Agnès ; Hallier, Amélie ; Gesret, Alexandrine ; Dangeard, Marine ; Dhemaied, Amine ; Gaboriau, Joséphine Boisson</creatorcontrib><description>Monitoring groundwater tables (GWTs) is challenging due to limited spatial and temporal observations. This study presents an innovative approach utilizing supervised deep learning, specifically a Multilayer Perceptron (MLP), and continuous passive-Multichannel Analysis of Surface Waves (passive-MASW) for constructing 2D GWT level maps. The study site, geologically well-constrained, features two 20-meter-deep piezometers and a permanent 2D geophone array capturing train-induced surface waves. For each point of the 2D array, dispersion curves (DCs), displaying Rayleigh-wave phase velocities (V_R) across a frequency range of 5 to 50 Hz, have been computed each day between December 2022 and September 2023. In the present study, these DCs are resampled in wavelengths ranging from 4 to 15~m in order to focus the monitoring on the expected GWT levels (between -1 and -5 m). Nine months of daily V_R data around one of the two piezometers is used to train the MLP model. GWT levels are then estimated across the entire geophone array, generating daily 2D GWT maps. Model’s performance is tested through cross-validation and comparisons with GWT level data at the second piezometer. Model’s efficiency is quantified with the root-mean-square error (RMSE) and the coefficient of determination (R²). The R² is estimated at 80% for data surrounding the training piezometer, and at 68% for data surrounding the test piezometer. Additionally, the RMSE is impressively low at 0.03 m at both piezometers. Results showcase the effectiveness of DL in estimating GWT level maps from passive-MASW data, offering a practical and efficient monitoring solution across broader spatial extents.</description><identifier>ISSN: 0043-1397</identifier><identifier>EISSN: 1944-7973</identifier><identifier>DOI: 10.22541/essoar.171322609.99979575/v1</identifier><language>eng</language><publisher>American Geophysical Union</publisher><subject>Computer Science ; Earth Sciences ; Environmental Sciences ; Geophysics ; Global Changes ; Hydrology ; Mathematical Software ; Sciences of the Universe</subject><ispartof>Water resources research, 2024-04</ispartof><rights>Distributed under a Creative Commons Attribution 4.0 International License</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><orcidid>0000-0002-6002-3189 ; 0000-0003-0305-4146 ; 0000-0003-2271-3223</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>230,314,780,784,885,27924,27925</link.rule.ids><backlink>$$Uhttps://hal.science/hal-04716694$$DView record in HAL$$Hfree_for_read</backlink></links><search><creatorcontrib>Teixeira, José Cunha</creatorcontrib><creatorcontrib>Bodet, Ludovic</creatorcontrib><creatorcontrib>Rivière, Agnès</creatorcontrib><creatorcontrib>Hallier, Amélie</creatorcontrib><creatorcontrib>Gesret, Alexandrine</creatorcontrib><creatorcontrib>Dangeard, Marine</creatorcontrib><creatorcontrib>Dhemaied, Amine</creatorcontrib><creatorcontrib>Gaboriau, Joséphine Boisson</creatorcontrib><title>Physics-guided deep learning model for daily groundwater table maps estimation using passive surface-wave dispersion</title><title>Water resources research</title><description>Monitoring groundwater tables (GWTs) is challenging due to limited spatial and temporal observations. This study presents an innovative approach utilizing supervised deep learning, specifically a Multilayer Perceptron (MLP), and continuous passive-Multichannel Analysis of Surface Waves (passive-MASW) for constructing 2D GWT level maps. The study site, geologically well-constrained, features two 20-meter-deep piezometers and a permanent 2D geophone array capturing train-induced surface waves. For each point of the 2D array, dispersion curves (DCs), displaying Rayleigh-wave phase velocities (V_R) across a frequency range of 5 to 50 Hz, have been computed each day between December 2022 and September 2023. In the present study, these DCs are resampled in wavelengths ranging from 4 to 15~m in order to focus the monitoring on the expected GWT levels (between -1 and -5 m). Nine months of daily V_R data around one of the two piezometers is used to train the MLP model. GWT levels are then estimated across the entire geophone array, generating daily 2D GWT maps. Model’s performance is tested through cross-validation and comparisons with GWT level data at the second piezometer. Model’s efficiency is quantified with the root-mean-square error (RMSE) and the coefficient of determination (R²). The R² is estimated at 80% for data surrounding the training piezometer, and at 68% for data surrounding the test piezometer. Additionally, the RMSE is impressively low at 0.03 m at both piezometers. Results showcase the effectiveness of DL in estimating GWT level maps from passive-MASW data, offering a practical and efficient monitoring solution across broader spatial extents.</description><subject>Computer Science</subject><subject>Earth Sciences</subject><subject>Environmental Sciences</subject><subject>Geophysics</subject><subject>Global Changes</subject><subject>Hydrology</subject><subject>Mathematical Software</subject><subject>Sciences of the Universe</subject><issn>0043-1397</issn><issn>1944-7973</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNotT8FOwzAUixBIjME_5MKBQ7akSZPmOE3AkCbBYffqrXndgrq2yms37e8pgpNly7Zsxp6VXGRZbtQSiTpIC-WUzjIr_cJ773zu8uVZ3bCZ8sYI552-ZTMpjRZKe3fPHoi-pVQmt27Ghq_jlWJF4jDGgIEHxJ43CKmN7YGfuoANr7vEA8Tmyg-pG9twgQETH2DfID9BTxxpiCcYYtfykX5zPRDFM3IaUw0VigtMJETqMdHkemR3NTSET_84Z7u31916I7af7x_r1VYcC6-E05hXtQsGJBaqthZA7XNVeKuDLABqND6TtvKqgsIHjdO7DHNjcF8Y7Wo9Zy9_tUdoyj5NE9O17CCWm9W2_NWkccpab85K_wAWyGVy</recordid><startdate>20240416</startdate><enddate>20240416</enddate><creator>Teixeira, José Cunha</creator><creator>Bodet, Ludovic</creator><creator>Rivière, Agnès</creator><creator>Hallier, Amélie</creator><creator>Gesret, Alexandrine</creator><creator>Dangeard, Marine</creator><creator>Dhemaied, Amine</creator><creator>Gaboriau, Joséphine Boisson</creator><general>American Geophysical Union</general><scope>1XC</scope><orcidid>https://orcid.org/0000-0002-6002-3189</orcidid><orcidid>https://orcid.org/0000-0003-0305-4146</orcidid><orcidid>https://orcid.org/0000-0003-2271-3223</orcidid></search><sort><creationdate>20240416</creationdate><title>Physics-guided deep learning model for daily groundwater table maps estimation using passive surface-wave dispersion</title><author>Teixeira, José Cunha ; Bodet, Ludovic ; Rivière, Agnès ; Hallier, Amélie ; Gesret, Alexandrine ; Dangeard, Marine ; Dhemaied, Amine ; Gaboriau, Joséphine Boisson</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-h891-73e5cf7d4a0e81f66aa1b518963d08aafe49206c91ca89d3e3972e544eb8437f3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science</topic><topic>Earth Sciences</topic><topic>Environmental Sciences</topic><topic>Geophysics</topic><topic>Global Changes</topic><topic>Hydrology</topic><topic>Mathematical Software</topic><topic>Sciences of the Universe</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Teixeira, José Cunha</creatorcontrib><creatorcontrib>Bodet, Ludovic</creatorcontrib><creatorcontrib>Rivière, Agnès</creatorcontrib><creatorcontrib>Hallier, Amélie</creatorcontrib><creatorcontrib>Gesret, Alexandrine</creatorcontrib><creatorcontrib>Dangeard, Marine</creatorcontrib><creatorcontrib>Dhemaied, Amine</creatorcontrib><creatorcontrib>Gaboriau, Joséphine Boisson</creatorcontrib><collection>Hyper Article en Ligne (HAL)</collection><jtitle>Water resources research</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Teixeira, José Cunha</au><au>Bodet, Ludovic</au><au>Rivière, Agnès</au><au>Hallier, Amélie</au><au>Gesret, Alexandrine</au><au>Dangeard, Marine</au><au>Dhemaied, Amine</au><au>Gaboriau, Joséphine Boisson</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Physics-guided deep learning model for daily groundwater table maps estimation using passive surface-wave dispersion</atitle><jtitle>Water resources research</jtitle><date>2024-04-16</date><risdate>2024</risdate><issn>0043-1397</issn><eissn>1944-7973</eissn><abstract>Monitoring groundwater tables (GWTs) is challenging due to limited spatial and temporal observations. This study presents an innovative approach utilizing supervised deep learning, specifically a Multilayer Perceptron (MLP), and continuous passive-Multichannel Analysis of Surface Waves (passive-MASW) for constructing 2D GWT level maps. The study site, geologically well-constrained, features two 20-meter-deep piezometers and a permanent 2D geophone array capturing train-induced surface waves. For each point of the 2D array, dispersion curves (DCs), displaying Rayleigh-wave phase velocities (V_R) across a frequency range of 5 to 50 Hz, have been computed each day between December 2022 and September 2023. In the present study, these DCs are resampled in wavelengths ranging from 4 to 15~m in order to focus the monitoring on the expected GWT levels (between -1 and -5 m). Nine months of daily V_R data around one of the two piezometers is used to train the MLP model. GWT levels are then estimated across the entire geophone array, generating daily 2D GWT maps. Model’s performance is tested through cross-validation and comparisons with GWT level data at the second piezometer. Model’s efficiency is quantified with the root-mean-square error (RMSE) and the coefficient of determination (R²). The R² is estimated at 80% for data surrounding the training piezometer, and at 68% for data surrounding the test piezometer. Additionally, the RMSE is impressively low at 0.03 m at both piezometers. Results showcase the effectiveness of DL in estimating GWT level maps from passive-MASW data, offering a practical and efficient monitoring solution across broader spatial extents.</abstract><pub>American Geophysical Union</pub><doi>10.22541/essoar.171322609.99979575/v1</doi><orcidid>https://orcid.org/0000-0002-6002-3189</orcidid><orcidid>https://orcid.org/0000-0003-0305-4146</orcidid><orcidid>https://orcid.org/0000-0003-2271-3223</orcidid><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | ISSN: 0043-1397 |
ispartof | Water resources research, 2024-04 |
issn | 0043-1397 1944-7973 |
language | eng |
recordid | cdi_hal_primary_oai_HAL_hal_04716694v1 |
source | Access via Wiley Online Library; Wiley-Blackwell AGU Digital Library; Wiley Online Library (Open Access Collection) |
subjects | Computer Science Earth Sciences Environmental Sciences Geophysics Global Changes Hydrology Mathematical Software Sciences of the Universe |
title | Physics-guided deep learning model for daily groundwater table maps estimation using passive surface-wave dispersion |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-23T05%3A42%3A13IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-hal&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Physics-guided%20deep%20learning%20model%20for%20daily%20groundwater%20table%20maps%20estimation%20using%20passive%20surface-wave%20dispersion&rft.jtitle=Water%20resources%20research&rft.au=Teixeira,%20Jos%C3%A9%20Cunha&rft.date=2024-04-16&rft.issn=0043-1397&rft.eissn=1944-7973&rft_id=info:doi/10.22541/essoar.171322609.99979575/v1&rft_dat=%3Chal%3Eoai_HAL_hal_04716694v1%3C/hal%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |