Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability

The sediment transport, involving the movement of the bedload and suspended sediment in the basins, is a critical environmental concern that worsens water scarcity and leads to degradation of land and its ecosystems. Machine learning (ML) algorithms have emerged as powerful tools for predicting sedi...

Descripción completa

Detalles Bibliográficos
Autores principales: El Bilali, A., Brouziyne, Youssef, Attar, O., Lamane, H., Hadri, A., Taleb, A.
Formato: Journal Article
Lenguaje:Inglés
Publicado: Springer 2024
Materias:
Acceso en línea:https://hdl.handle.net/10568/149323
_version_ 1855542172895412224
author El Bilali, A.
Brouziyne, Youssef
Attar, O.
Lamane, H.
Hadri, A.
Taleb, A.
author_browse Attar, O.
Brouziyne, Youssef
El Bilali, A.
Hadri, A.
Lamane, H.
Taleb, A.
author_facet El Bilali, A.
Brouziyne, Youssef
Attar, O.
Lamane, H.
Hadri, A.
Taleb, A.
author_sort El Bilali, A.
collection Repository of Agricultural Research Outputs (CGSpace)
description The sediment transport, involving the movement of the bedload and suspended sediment in the basins, is a critical environmental concern that worsens water scarcity and leads to degradation of land and its ecosystems. Machine learning (ML) algorithms have emerged as powerful tools for predicting sediment yield. However, their use by decision-makers can be attributed to concerns regarding their consistency with the involved physical processes. In light of this issue, this study aims to develop a physics-informed ML approach for predicting sediment yield. To achieve this objective, Gaussian, Center, Regular, and Direct Copulas were employed to generate virtual combinations of physical of the sub-basins and hydrological datasets. These datasets were then utilized to train deep neural network (DNN), conventional neural network (CNN), Extra Tree, and XGBoost (XGB) models. The performance of these models was compared with the modified universal soil loss equation (MUSLE), which serves as a process-based model. The results demonstrated that the ML models outperformed the MUSLE model, exhibiting improvements in Nash–Sutcliffe efficiency (NSE) of approximately 10%, 18%, 32%, and 41% for the DNN, CNN, Extra Tree, and XGB models, respectively. Furthermore, through Sobol sensitivity and Shapley additive explanation–based interpretability analyses, it was revealed that the Extra Tree model displayed greater consistency with the physical processes underlying sediment transport as modeled by MUSLE. The proposed framework provides new insights into enhancing the accuracy and applicability of ML models in forecasting sediment yield while maintaining consistency with natural processes. Consequently, it can prove valuable in simulating process-related strategies aimed at mitigating sediment transport at watershed scales, such as the implementation of best management practices.
format Journal Article
id CGSpace149323
institution CGIAR Consortium
language Inglés
publishDate 2024
publishDateRange 2024
publishDateSort 2024
publisher Springer
publisherStr Springer
record_format dspace
spelling CGSpace1493232025-12-08T09:54:28Z Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability El Bilali, A. Brouziyne, Youssef Attar, O. Lamane, H. Hadri, A. Taleb, A. sediment yield forecasting machine learning algorithms modified universal soil loss equation soil erosion models consistency sensitivity analysis watersheds sediment transport datasets neural networks case studies The sediment transport, involving the movement of the bedload and suspended sediment in the basins, is a critical environmental concern that worsens water scarcity and leads to degradation of land and its ecosystems. Machine learning (ML) algorithms have emerged as powerful tools for predicting sediment yield. However, their use by decision-makers can be attributed to concerns regarding their consistency with the involved physical processes. In light of this issue, this study aims to develop a physics-informed ML approach for predicting sediment yield. To achieve this objective, Gaussian, Center, Regular, and Direct Copulas were employed to generate virtual combinations of physical of the sub-basins and hydrological datasets. These datasets were then utilized to train deep neural network (DNN), conventional neural network (CNN), Extra Tree, and XGBoost (XGB) models. The performance of these models was compared with the modified universal soil loss equation (MUSLE), which serves as a process-based model. The results demonstrated that the ML models outperformed the MUSLE model, exhibiting improvements in Nash–Sutcliffe efficiency (NSE) of approximately 10%, 18%, 32%, and 41% for the DNN, CNN, Extra Tree, and XGB models, respectively. Furthermore, through Sobol sensitivity and Shapley additive explanation–based interpretability analyses, it was revealed that the Extra Tree model displayed greater consistency with the physical processes underlying sediment transport as modeled by MUSLE. The proposed framework provides new insights into enhancing the accuracy and applicability of ML models in forecasting sediment yield while maintaining consistency with natural processes. Consequently, it can prove valuable in simulating process-related strategies aimed at mitigating sediment transport at watershed scales, such as the implementation of best management practices. 2024-07 2024-07-31T06:05:15Z 2024-07-31T06:05:15Z Journal Article https://hdl.handle.net/10568/149323 en Limited Access Springer El Bilali, A.; Brouziyne, Youssef; Attar, O.; Lamane, H.; Hadri, A.; Taleb, A. 2024. Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability. Environmental Science and Pollution Research, 31(34):47237-47257. [doi: https://doi.org/10.1007/s11356-024-34245-2]
spellingShingle sediment yield
forecasting
machine learning
algorithms
modified universal soil loss equation
soil erosion models
consistency
sensitivity analysis
watersheds
sediment transport
datasets
neural networks
case studies
El Bilali, A.
Brouziyne, Youssef
Attar, O.
Lamane, H.
Hadri, A.
Taleb, A.
Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability
title Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability
title_full Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability
title_fullStr Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability
title_full_unstemmed Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability
title_short Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency, sensitivity, and interpretability
title_sort physics informed machine learning algorithms for forecasting sediment yield an analysis of physical consistency sensitivity and interpretability
topic sediment yield
forecasting
machine learning
algorithms
modified universal soil loss equation
soil erosion models
consistency
sensitivity analysis
watersheds
sediment transport
datasets
neural networks
case studies
url https://hdl.handle.net/10568/149323
work_keys_str_mv AT elbilalia physicsinformedmachinelearningalgorithmsforforecastingsedimentyieldananalysisofphysicalconsistencysensitivityandinterpretability
AT brouziyneyoussef physicsinformedmachinelearningalgorithmsforforecastingsedimentyieldananalysisofphysicalconsistencysensitivityandinterpretability
AT attaro physicsinformedmachinelearningalgorithmsforforecastingsedimentyieldananalysisofphysicalconsistencysensitivityandinterpretability
AT lamaneh physicsinformedmachinelearningalgorithmsforforecastingsedimentyieldananalysisofphysicalconsistencysensitivityandinterpretability
AT hadria physicsinformedmachinelearningalgorithmsforforecastingsedimentyieldananalysisofphysicalconsistencysensitivityandinterpretability
AT taleba physicsinformedmachinelearningalgorithmsforforecastingsedimentyieldananalysisofphysicalconsistencysensitivityandinterpretability