Evaluation of SARIMA Model for Rainfall Forecast in Ogbaru in Anambra State, Nigeria

Abstract

Ogbaru community is blessed with arable land suitable for crop production. However, this region cannot be fully harnessed without proper understanding of the rainfall pattern. Modelling and forecasting rainfall in this region is crucial considering the climate change that has brought a new narrative into the rainfall pattern nationwide. This study applied Seasonal Integrated Moving Average (SARIMA) models in modelling and forecasting rainfall in Ogbaru. The yearly rainfall data for this locality between 1995 and 2025 (30 years) was obtained from the NIMET. To assess the stationarity of the time series, initial exploratory analysis was conducted through graphical visualization. Formal stationarity was carried out using the Augmented Dickey–Fuller (ADF) test. Adopting a 5% significance level, a p-value below 0.05 (p < 0.05) was considered as been stationary. Following the establishment of stationarity, tentative Seasonal ARIMA (SARIMA) models were identified. The identification of this structure was guided by the inspection of Autocorrelation Function (ACF) and Partial Autocorrelation Function (PACF) plots. This manual method of model identification was cross-validated using the auto-arima procedure in Python, which optimizes parameters based on the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). Model parameters were estimated with 95% confidence intervals, and goodness-of-fit was evaluated using the Ljung–Box test and a p-value greater than 0.05 (p > 0.05). Based on the result of out-of-sample forecast performance, SARIMA (1,0,2) (1,0,1)₁₂ was found to be suitable for rainfall forecast. Finally, ten years foecast for Ogbaru was obtained using the optimal model. The findings show that the seasonal terms were statistically significant in the model which justified the use of SARIMA models in modelling rainfall in Ogbaru. Findings also revealed that SARIMA model is good for long-term forecasting of rainfall in this location.