Application of seasonal autoregressive integrated moving average (SARIMA) for flows of River Kaduna

Abstract

Using 25 years monthly discharge data (1988 to 2013), the discharge of River Kaduna was investigated for the possibility of accurate forecast. With the aid of ADF (Augmented Dickey-Fuller) test, auto-correlation and partial auto-correlation functions the discharge was found to exhibit a stochastic non-stationary seasonal time series behavior which becomes stationary after first seasonal differencing. Based on this, the study predicted the discharge of the river from 2014 to 2018 using seasonal autoregressive integrated moving average model and validates this with the actual discharge of the river for the corresponding period. Hence, the study concluded that SARIMA (1, 0, 1) (0, 1, 1)12 mode is the most appropriate based on the selection criteria, and could adequately predict the discharge of River Kaduna with minimal errors.