SARIMA-BASED FORECASTING OF MONTHLY RAINFALL FOR AGRICULTURAL PLANNING IN JOS, NIGERIA

Agriculture in Nigeria, particularly in the North-Central region, is predominantly rain-fed, making it highly susceptible to the increasing variability of climatic patterns. This study develops a Seasonal Autoregressive Integrated Moving Average model to forecast monthly rainfall in Jos, Plateau State, a region critical for the cultivation of specialised temperate-like crops. Utilising a 30-year dataset (1996–2025) from the Nigerian Meteorological Agency, the research identified significant rainfall variability, with a mean monthly precipitation of 104.91 mm and a high coefficient of variation (103.75%). The Augmented Dickey-Fuller test confirmed the stationarity of the seasonally differenced series (P < 0.05). Among several candidate models, SARIMA was selected based on the lowest Akaike Information Criterion (AIC) of 3430.71 and a Root Mean Square Error (RMSE) of 35.56. The results demonstrate that the model effectively captures the seasonal cycles and temporal dependencies of the rainfall series. The implications of these forecasts for agriculture are profound, providing a quantitative basis for synchronising planting calendars with actual moisture availability, mitigating the risks of "false onsets," and enhancing food security for moisture-sensitive crops, such as Irish potatoes and maize. The study recommends the integration of such stochastic models into regional agricultural planning and the development of climate-smart infrastructure to bolster resilience against climatic shocks.