A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Rainfall-runoff modelling – a comparison of Artificial Neural Networks (ANNs) and Hydrologic Engineering Centre-Hydrologic Modelling System (HEC-HMS)
This article presents comparison of Artificial Neural Networks (ANNs) and Hydrologic Engineering Centre-Hydrologic Modelling System (HEC-HMS) model for rainfall-runoff (R-R) process. Aim of the present work is to forecast runoff one day ahead at Shivade station of Upper Krishna Basin, India, using 17 years of daily rainfall and discharge data. The R-R modelling can be exercised using various traditional methods which generally require exogenous data in the form of basin parameters. Unavailability of such data becomes major impediment in applying these models at many basins. In such situations, soft computing techniques like ANNs have been extensively applied to model R-R process. Though ANN is now an established tool in hydrology, compared to HEC-HMS, its results are viewed with suspicion owing to its data-driven nature rather than a model-driven nature. In this study, ANN model performed reasonably well, with a higher correlation coefficient (0.87) and the lowest Root Mean Square Error (136.28 m3/s) when compared with HEC-HMS (0.76, 139.8 m3/s) respectively. Novelty of the present work lies in model development using restricted basin data. Both models showed less accuracy in predicting extreme events. Finally, it is concluded that ANN model can be used as a supplementary technique along with HEC-HMS for this phenomenon.
Rainfall-runoff modelling – a comparison of Artificial Neural Networks (ANNs) and Hydrologic Engineering Centre-Hydrologic Modelling System (HEC-HMS)
This article presents comparison of Artificial Neural Networks (ANNs) and Hydrologic Engineering Centre-Hydrologic Modelling System (HEC-HMS) model for rainfall-runoff (R-R) process. Aim of the present work is to forecast runoff one day ahead at Shivade station of Upper Krishna Basin, India, using 17 years of daily rainfall and discharge data. The R-R modelling can be exercised using various traditional methods which generally require exogenous data in the form of basin parameters. Unavailability of such data becomes major impediment in applying these models at many basins. In such situations, soft computing techniques like ANNs have been extensively applied to model R-R process. Though ANN is now an established tool in hydrology, compared to HEC-HMS, its results are viewed with suspicion owing to its data-driven nature rather than a model-driven nature. In this study, ANN model performed reasonably well, with a higher correlation coefficient (0.87) and the lowest Root Mean Square Error (136.28 m3/s) when compared with HEC-HMS (0.76, 139.8 m3/s) respectively. Novelty of the present work lies in model development using restricted basin data. Both models showed less accuracy in predicting extreme events. Finally, it is concluded that ANN model can be used as a supplementary technique along with HEC-HMS for this phenomenon.
Rainfall-runoff modelling – a comparison of Artificial Neural Networks (ANNs) and Hydrologic Engineering Centre-Hydrologic Modelling System (HEC-HMS)
Deulkar, Aparna M. (author) / Londhe, Shreenivas N. (author) / Jain, Rakesh K. (author) / Dixit, Pradnya R. (author)
ISH Journal of Hydraulic Engineering ; 30 ; 478-488
2024-08-07
11 pages
Article (Journal)
Electronic Resource
English
Hydrologic modelling for runoff determination
| Taylor & Francis Verlag | 1991
Hydrologic vs Hydraulic Rainfall-Runoff Modeling
| British Library Conference Proceedings | 2022
RAINFALL—RUNOFF MODELLING USING ARTIFICIAL NEURAL NETWORKS
| Taylor & Francis Verlag | 2009
Hydrologic System Identification Using Artificial Neural Networks
| British Library Conference Proceedings | 2001
Artificial neural networks for rainfall-runoff modelling in urban drainage
| British Library Conference Proceedings | 1998