Flow forecasts in a non-perennial river of an arid basin using neural networks. Issue 2 (2nd July 2016)

• Record Type: Journal Article
• Authors: Zemzami, Mahmoud
• Is Part Of: Journal of applied water engineering and research. Volume 4:Issue 2(2016)
• Page Start: 92
• Page End: 101
• Publication Date: 2016-07-02
• Subjects: artificial neural networks -- arid watershed -- non-perennial river -- flow forecasting -- Morocco; Water-supply engineering -- Periodicals; Water-supply engineering; Periodicals; 627.05
• DOI: 10.1080/23249676.2015.1072851
• Languages: English
• ISSNs: 2324-9676

Abstract

Abstract : The hydrological modeling is a complicated problem in arid areas with a non-perennial river system, where conceptual or physically based models are difficult to implement. In such situations, where explicit knowledge of the internal hydrologic system is not available, artificial neural networks (ANNs) can be a good tool to model the hydrological behavior. In this paper, the prediction of daily flows, which is required to provide information about a wide range of problems related to design and river management, was investigated in an arid watershed. The results show that the optimization of the different network parameters has significantly improved the model performance (the Nash-Sutcliff coefficient was increased by 10 points) and thus increases its ability to catch the hydrological processes. It was also shown that the ANN is a promising method of flow forecast in arid watershed and can provide an effective approach for forecasting flows in the current situation of the study area.