Predicting Inlet Flow to Jamishan Dam Using Time Series Models

As aquifer feeder and influential parameter in water balance equations and groundwater resources balance, accurate prediction of dams and rivers discharge plays an important role in planning managing and operating optimal and sustainable water resources. In this research, in order to organize the Jamishan catchment area. In order to predict the future natural hazards of the basin, the monthly discharge of this basin is predicted by time series analysis methods. In this regard Was used from monthly discharge data of entrance to jamishan dam in sonqor city of kermanshah province during the period (1360-1389). Initial analysis of data included a review of definitive series semantics (period, trend, jump) done on the time series and after assurance remove these semantics, data was normal and the data stagnation was made. By examining the correlation and partial correlation functions for fifty percent of the data, the self-correlated model (AR) and the moving average model (MA) were fitted for the calibration period to the time series and with the non-correlation test of Port-Manto and the normalization of the remainder, a number of models that did not have these conditions were eliminated, and the best models were identified among the remaining models with Akayek's test. In the verification stage, using the best model during the calibration period, for the fifty-second percent of the data, the prediction verification step was performed. And error validation values were evaluated using white nose, Barlett-Test (Cumulative Rotational), mean of remaining significance, and after the success of the model in the verification prototypes, it was used to predict the monthly discharge of the next 15 years. It can be concluded that the more the model is more static, the analysis of the series is easier and the model with less acacia gives a better answer.