Terzi, Özlem Süleyman Demirel Üniversitesi - Teknik Eğitim Fakültesi - Yapı Eğitimi Bölümü, Turkey , Köse, Mehmet Süleyman Demirel Üniversitesi - Teknik Eğitim Fakültesi - Yapı Eğitimi Bölümü, Turkey

FLOW FORECASTING OF GÖKSU RIVER WITH ARTIFICIAL NEURAL NETWORKS METHOD

35792

Nowadays, it is important issues such as the use and operation of water resources because of appreciably increasing in drought and global warming. The river flow is determined by flow measurement stations established by relevant institutions on rivers. However, it is a difficult to operate these stations in such cases the absence of data and failure of the stations. In such cases, in order to complete the missing data, the flow estimation of Göksu River was made with artificial neural networks (ANN) method that most widely used in water resources engineering in recent years. For this purpose, it was used to develop ANN models daily flow values for the years 1990-2010 from Karahacılı (1714), Kırkkavak (1719) and Hamam (1720) measurement stations on the Göksu River. It was used determination coefficient and the mean absolute error to evaluate performance of the developed models. Comparing performances of the models, it was shown that ANN method can be used to estimate river flow.

1

Flow , artificial neural networks , Göksu River

Sdu International Technologic Science

7