article · Water Resources
Accurate rainfall-runoff modelling is vital for managing water resources and mitigating flood risks. A comparative study evaluated several hydrological modelling techniques within a watershed in north-central Algeria. The assessment included conceptual models, specifically Génie Rural (GR5J) and Hydrologiska Byråns Vattenbalansavdelning (HBV), alongside standalone machine learning algorithms, namely Multilayer Perceptron Neural Networks (MLPNN) and Random Forest Regression (RFR). These methods were compared against hybrid models that combine Variational Mode Decomposition with the machine learning algorithms, designated as VMD-MLPNN and VMD-RFR. The findings demonstrated that the hybrid VMD-MLPNN model delivered superior predictive performance compared to both the conceptual approaches and the standalone machine learning models. It achieved the highest accuracy, recorded through a correlation coefficient of approximately 0.990 and a Nash-Sutcliffe Efficiency value of 0.964.
Reliable streamflow predictions are essential for planning water supplies and protecting communities from flood hazards. By demonstrating that hybrid machine learning methods outperform conventional hydrological models, this research highlights how advanced data decomposition techniques can significantly enhance forecasting accuracy in watershed management.
The work could enable improved flood forecasting tools and water resource management systems for watershed managers and environmental agencies. Tested in a specific catchment setting, the methodology represents applied research that requires further operational integration before reaching near-market commercial or public deployment.
AI-generated from the published abstract. Always read the original work before citing.
Abstract Improving rainfall-runoff (RR) modeling aims to refine streamflow predictions using more accurate data and methods, which is crucial for effective water resource management and reducing flood risks. Machine learning models and conceptual models have been utilized in an attempt to perform rainfall-runoff modeling. This study aims to compare the performance of conceptual models; Génie Rural (GR5J) and Hydrologiska Byråns Vattenbalansavdelning (HBV), machine learning models Perceptron; Neural Network (MLPNN), Random Forest Regression (RFR), and hybrid machine learning (ML) models based on Variational mode decomposition (VMD), VMD-MLPNN, and VMD-RFR for rainfall–runoff modelling of watershed in north-central of Algeria. It was obtained that the performance of the hybrid ML models VMD-MLPNN is better than conceptual models and stand-alone machine learning models, with the highest values for correlation coefficient (R) and Nash-Sutcliffe Efficiency (NSE) of approximately 0.990 and 0.964, respectively.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1134/s0097807824604850
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
