Related Experiment Video
Updated: Jun 21, 2025
10:32
Image-based Lagrangian Particle Tracking in Bed-load Experiments
Published on: July 20, 2017
9.0K
Physics-informed machine learning algorithms for forecasting sediment yield: an analysis of physical consistency,
Ali El Bilali1,2, Youssef Brouziyne3, Oumaima Attar4
1Faculty of Sciences and Techniques, Hassan II University of Casablanca, Mohammedia, Morocco. ali1gpee@gmail.com.
Summary
Physics-informed machine learning (ML) models accurately predict sediment yield, outperforming traditional methods. The Extra Tree model showed the best consistency with physical sediment transport processes, aiding watershed management.
Area of Science:
- Environmental Science
- Hydrology
- Data Science
Background:
- Sediment transport is a major environmental issue impacting water resources and ecosystems.
- Machine learning (ML) offers potential for sediment yield prediction but lacks physical process consistency.
- Existing ML models raise concerns among decision-makers regarding their reliability in environmental applications.
Purpose of the Study:
- To develop a physics-informed machine learning (ML) approach for predicting sediment yield.
- To enhance the accuracy and physical consistency of ML models in sediment transport modeling.
- To provide a framework for improving ML applicability in watershed management and sediment mitigation strategies.
Main Methods:
- Generated synthetic hydrological and sub-basin datasets using Gaussian, Center, Regular, and Direct Copulas.
- Trained various ML models including deep neural network (DNN), conventional neural network (CNN), Extra Tree, and XGBoost (XGB).
- Compared ML model performance against the Modified Universal Soil Loss Equation (MUSLE) and conducted interpretability analyses (Sobol, Shapley).
Main Results:
- All ML models significantly outperformed the MUSLE model in predicting sediment yield, with Nash-Sutcliffe efficiency (NSE) improvements ranging from 10% to 41%.
- The Extra Tree model demonstrated superior consistency with the physical processes of sediment transport, as evaluated by interpretability methods.
- DNN, CNN, Extra Tree, and XGB models showed notable improvements in predicting sediment yield compared to the process-based MUSLE model.
Conclusions:
- Physics-informed ML models offer enhanced accuracy and reliability for sediment yield prediction.
- The Extra Tree model's physical consistency makes it a valuable tool for understanding and managing sediment transport.
- This framework supports the development of effective watershed-scale sediment mitigation strategies and best management practices.
Related Concept Videos
01:30Typical Model Studies
354
Fluid mechanics model studies often utilize scaled-down systems to predict fluid behavior in full-scale environments, such as river flows, dam spillways, and structures interacting with open surfaces. Maintaining Froude number similarity in river models is crucial, as it replicates surface flow features like wave patterns and velocities.
354
01:12Modeling and Similitude
261
Scaled modeling is a fundamental technique in engineering, enabling the study of large and complex systems by creating smaller, manageable replicas that recreate critical characteristics of the original. In hydrology and civil infrastructure, for example, scaled models of dams help analyze water flow, turbulence, and pressure. This method allows for accurate predictions of real-world behavior within a controlled environment, significantly reducing the cost and time involved in full-scale...
261
01:29Mechanistic Models: Compartment Models in Algorithms for Numerical Problem Solving
48
Mechanistic models play a crucial role in algorithms for numerical problem-solving, particularly in nonlinear mixed effects modeling (NMEM). These models aim to minimize specific objective functions by evaluating various parameter estimates, leading to the development of systematic algorithms. In some cases, linearization techniques approximate the model using linear equations.
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...
48
01:24Rapidly Varying Flow
57
Rapidly varying flow (RVF) in open channels is characterized by abrupt changes in flow depth over a short distance, with the rate of depth change relative to distance often approaching unity. These flows are inherently complex due to their transient and multi-dimensional nature, making exact analysis difficult. However, approximate solutions using simplified models provide valuable insights into their behavior.Key Features of Rapidly Varying FlowRVF is commonly observed in scenarios involving...
57
01:23Mechanistic Models: Compartment Models in Individual and Population Analysis
36
Mechanistic models are utilized in individual analysis using single-source data, but imperfections arise due to data collection errors, preventing perfect prediction of observed data. The mathematical equation involves known values (Xi), observed concentrations (Ci), measurement errors (εi), model parameters (ϕj), and the related function (ƒi) for i number of values. Different least-squares metrics quantify differences between predicted and observed values. The ordinary least...
36
01:21Design Example: Creating a Hydraulic Model of a Dam Spillway
157
Scaled hydraulic models of dam spillways provide a practical way to replicate and study the intricate flow dynamics of these structures. Often built to a 1:15 ratio, these models allow for observing critical water behavior, such as velocity distribution, flow patterns, and energy dissipation.
157