Artificial Neural Network Models Investigation for Euphrates River Forecasting & Back Casting

The development of stream flow forecasting model is one of the most important aspects in water resources planning and management , since it can help in providing early warning of river flooding as well as in short term operation of water supply system. In this research the best ANN artificial neural networks model for simulation and forecasting of Euphrates river flow downstream Al-Hindiyaha barrage was investigated by applying different architectures of ANN models to the monthly flow of Euphrates river at the mentioned site depending on the previous months data from the same site (this was called Single Site ANN model SSANN )and also by investigating the dependence of Euphrates river flow downstream Al-Hindiyaha barrage on two other sites (Husaybah , Hit)discharges of the same river (which was called Multi Site ANN model MSANN). Another Trial was achieved by investigating different ANN models to predict a missed monthly data for Euphrates river flow at Husaybah gauging station. This trail was achieved first by investigating different MSANN models then by reversing operation of the best forecasting model which was found, this operation was called Back casting technique. The current study had demonstrated a promising application of ANN –stream flow forecasting and back casting for different gaging stations of Euphrates river .