Abstract
Abstract
This research analyzes the minimum daily flows of each month for 52 years from Eskikalak and Dokan gauge stations onGreaterZabriverand Lesser Zab river respectivelyto predict their lowflowsusingtwo Stochastic Models;(ARIMA)model (Autoregressive Integrated Moving Average) and (Thomas-Fiering)model.A comparison is conducted for the results of both methodsutilizing theMean Average Error(MAE) and Root Mean Square Error (RMSE)ofpredicteddatafrom bothmodelsto verify theirperformanceandreliability. The results showed that the ARIMA model has given more appropriate forecasts and harmonic direction comparing to the same historical time series than those given by (Thomas-Fiering)model. Minitab version 14package is utilizedto assess the statistical side of the research.
Key word: Lowflow, GreaterZab, Lesser Zab, ARIMA, Thomas-Fiering
