Prediction of Rainfall by using Seasonal ARIMA Model in Namtumbo District, Tanzania

Abstract

Several scholars have attempted to predict rainfall every month at various points worldwide by utilizing various time series methodologies. Because rainfall varies randomly in space and time due to various dynamic and environmental factors, rainfall prediction at the monthly time scale is challenging. This study examines the monthly rainfall in Tanzania's Namtumbo District from January 2000 to December 2020. The average rainfall in Namtumbo District, Tanzania, was predicted using Seasonal Auto-Regressive Integrated Moving Average (SARIMA) models since rainfall time data indicate seasonal components. Considering the model with the lowest Akaike's Information Criteria (AIC) value, the best SARIMA model was chosen. It was discovered that the most effective time series forecasting model for capturing the current pattern of monthly rainfall data in Namtumbo District is SARIMA (1, 1, 1) (0, 1, 1)₁₂. After that, the model was employed to forecast monthly rainfall values between January 2022 and December 2030. The forecasting results showed that Namtumbo District will be experiencing high rainfall from November to April and will not be experiencing rainfall from May to September each year. Also, results show that rainfall will have a seasonal pattern in the future where rainfall rises and falls in the predicted data values that repeat regularly over the same period. The study will help the Namtumbo district's decision-makers set goals and plans based on the challenges caused by rainfall variability.