Forecasting the Monthly Mean Temperature of Jos Northern Region of Plateau State using the SARIMA-GAM Hybrid Model

Abstract

In reality, meteorologists create their periodic weather forecasts for variables like humidity, average temperature, precipitation, and other atmospheric factors using a variety of statistical techniques. The goal of the current study was to examine the temporal variations in the monthly mean temperature of Jos City based on a hybrid SARIMA and Generalized Additive Models (GAM) technique (SARIMA-GAM), using data from the Department of Geography of the University of Jos Meteorology Unit from 1986 to 2023. This study evaluated a suitable forecasting model for the monthly and annual mean temperature of the Jos North Region. The SARIMA and GAM models are used to capture both the linear and nonlinear components, respectively. This study benefits from the strengths of both SARIMA and GAM to improve forecasting accuracy. The hybrid model provides a more comprehensive analysis of the time series data by capturing nonlinearity using the GAM model and linearity using the SARIMA model. In conclusion, popular metrics like MSE, RMSE, MAE, and MAPE are used to assess and compare the predicted performance of various models. Principal results show that the hybrid SARIMA-GAM model outperforms the individual models in predicting the region’s mean temperature in Jos City, Nigeria.