Evaluation of SARIMA time series models in monthly streamflow estimation in Idanak hydrometry station

Abstract

prediction of hydrological variables is a highly effective tool in water resource management. One of the important tools for modeling hydrological processes is the use of time series modeling and analysis. River series production series can be used by time series models in various studies such as drought, flood, reservoir systems design and many other purposes For this purpose, monthly flow data of this station has been used for 30 years (2011-1363). By using the regression method, incomplete data estimation and homogeneity of data were investigated by sequencing test .Using the SARIMA model, the monthly time series of the Idenak station was Investigated and the best model was fitted to its data.. The models were confirmed by the diagram of autocorrelation and partial-bond correlation functions of the residues and the Pert-Manto criteria. for evaluation the models, the AIC, SBC criteria were used. The results show that SARIMA models (1.0,1) * (2,0,2) 12, SARIMA (2,0,2) * (2,0,2 (12) and SARIMA (1,0,2) * (2,0,2) 12 are respectively in the first, second and third priority in terms of accuracy in modeling the monthly discharge of the Idenak station.