Comparative evaluation of classical and SARIMA-BL time series hybrid models in predicting monthly qualitative parameters of Maroon river

Abstract In the present study, seasonal autoregressive integrated moving average (SARIMA) time series models, nonlinear BL model, multi-layer perceptron artificial neural network and SARIMA-bilinear hybrid models were employed to predict quality parameters of total dissolved solids (TDS), sodium adsorption ratio (SAR), electrical conductivity (EC) in Maroon basin Khuzestan Province. For fitting mentioned monthly data used for calibration (1970–2000), confirmation (2001–2011) prediction model (2012–2018). The appropriate SARIMA, bilinear SARIMA-BL rationalized above-mentioned selected based on adequacy tests, such as Akaike criterion independence test residuals (Ljung–Box). To determine effective input network, partial mutual information (PMI) algorithm was three EC, TDS, SAR parameters. Also, output layers linear transfer function hidden layer various active functions with back-propagation learning modeling predicting this water Comparison showed tangible superiority than PMI algorithm, SARIMA bl performed all qualitative stages, training-validation-test basin.