Accurate rainfall forecasting is crucial for effective climate monitoring, agricultural planning, and disaster risk management. This study develops and evaluates a predictive model for rainfall forecasting in Lagos State using the Seasonal Autoregressive Integrated Moving Average (SARIMA) approach. A monthly rainfall dataset spanning January 2016 to December 2022 was obtained from the relevant government agency in Nigeria. The dataset was systematically divided into a training subset (January 2016 – December 2021) for model development and a testing subset
(January 2022 – December 2022) for performance evaluation. Time series analysis revealed significant seasonal patterns and long-term trends in Lagos State’s rainfall data. Stationarity was assessed using the Augmented Dickey-Fuller (ADF) test, and necessary transformations, including first seasonal differencing, were applied to stabilize the series. Model selection was performed using the Akaike Information Criterion (AIC), with SARIMA (2,0,4)(2,2,1)[12] emerging as the optimal model due to its lowest AIC value. Model adequacy was further evaluated using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), confirming its superior predictive performance. The forecasting results demonstrated that the SARIMA model effectively captured the seasonal fluctuations in rainfall, with periodic peaks and troughs. The findings underscore the model’s applicability for real-world decision-making and policy formulation in climate-sensitive sectors
Adetunji K. Ilori, Adebisi Michael, Fatunsin L. Modupe, Olaiya O. O, and Toyosi Adebambo, “Modelling Lagos State, Nigeria Rainfall using Real Life Data: An application of SARIMA Model,” International Journal of Multidisciplinary Research and Publications (IJMRAP), Volume 7, Issue 10, pp. 122-127, 2025.