Journal cover Journal topic
Hydrology and Earth System Sciences An interactive open-access journal of the European Geosciences Union
https://doi.org/10.5194/hess-2016-347
© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.
Research article
01 Aug 2016
Acclimatizing Fast Orthogonal Search (FOS) Model for River Stream-flow Forecasting
Abdalla Osman1, Mohammed Falah Allawi2, Haitham Abdulmohsin Afan2, Aboelmagd Noureldin1, and Ahmed El-shafie3 1Electrical and Computer Engineering, Royal Military College of Canada , Kingston, Canada
2Civil and Structu ral Engineering Dept. Universiti Kebangsaan Malaysia, Malaysia
3Department of Civil Engineering, Faculty of Engineering, University Malaya
Abstract. River stream-flow is well-thought-out as an essential element in the hydrology studies, especially for reservoir management. Forecasting river stream-flow is the key for the hydrologists in proposing certain short or long-term planning and management for water resources system. In fact, developing stream-flow forecasting models are generally categorized into two main classes; process and data-driven model. Different model techniques based on empirical methods, such as stochastic model or regression model, more recently, Artificial Intelligent (AI) models have been examined and could provide accurate stream-flow forecasting. However, AI models experienced crucial difficulty is the necessity to utilize appropriate pre-processing methods for the raw data. In addition, the AI model should be augmented with proper optimization model to adjust the model parameters to achieve the optimal accuracy. In this paper, a novel model namely; Fast Orthogonal Search (FOS) model is proposed to develop river stream-flow forecasting. FOS is basically structured for recognizing the difference equation and its functional expression model for the mapping between the model input and output. The major advantage of using FOS is the waiver of the requirement of data pre-processing and optimization model for model parameters adjustment as these procedures are performed implicitly inside FOS. In addition, pole-zero cancellation procedure within FOS process can detect the over-fitted models and avoid them. The proposed FOS method was adopted in this research to perform stream-flow forecasting model at Aswan High Dam using monthly basis for130 years. Results showed outstanding performance for stream-flow forecasting accuracy compared to other AI models developed during the last 10 years.
The discussion paper was formally withdrawn.


Citation: Osman, A., Allawi, M. F., Afan, H. A., Noureldin, A., and El-shafie, A.: Acclimatizing Fast Orthogonal Search (FOS) Model for River Stream-flow Forecasting, Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2016-347, 2016.
Abdalla Osman et al.
Abdalla Osman et al.

Viewed

Total article views: 562 (including HTML, PDF, and XML)

HTML PDF XML Total BibTeX EndNote
257 288 17 562 12 20

Views and downloads (calculated since 01 Aug 2016)

Cumulative views and downloads (calculated since 01 Aug 2016)

Saved

Discussed

Latest update: 22 Aug 2017
Publications Copernicus
Download
Short summary
The current research is significant not only for researchers but also for decision-makers for water resources system management more specifically for dam and reservoir operations. Fast Orthogonal Search (FOS) model might result in a system that could be used in several river basins. Supported by efficient FOS-based pattern recognition model, accurate information about future river stream flow could be utilized by the decision-maker to formulate better water resources system management.
The current research is significant not only for researchers but also for decision-makers for...
Share