OPTIMIZING RESOURCE MANAGEMENT USING SEASONAL ARIMA RAINFALL FORECASTING MODEL FOR ODEDA, NIGERIA

Abstract

Forecasting rainfall is essential for countries with economies hinged on agriculture, engineering, and tourism, as well as for the supply of water and sanitation. In this study, a seasonal ARIMA model was used to forecast rainfall events in the Odeda region, Southwestern Nigeria. The rainfall records of the OgunOshun River Basin Development Authority weather station from January 1982 to December 2016 were employed to develop the model. A visual inspection of the autocorrelation function (ACF) and partial autocorrelation (PACF) of tentative models was done using the Box-Jenkins approach. The selected preliminary models were subjected to selection criteria of coefficient of determination (R2), Bayesian information criterion (BIC), and Root Mean Square Error (RMSE) and diagnostic checking. The selected best model was Seasonal ARIMA (3,1,3) (2,1,1)12 with a coefficient of determination of 0.68 for predicting rainfall events from January 2017 to December 2018. Hence, the model was utilized to forecast rainfall events till 2022. With the increasingly changing patterns of rainfall occurrence worldwide due to climate change, this study highlights the importance of analyzing and forecasting rainfall.