Journal Article

Modeling and Forecasting Humidity with SARIMA Model Application in Selected Sudan Savanna States, Nigeria

Abstract

Forecasting humidity is very important to help maintaining comfortable environment as well as optimizing the efficiency of Heating, ventilation, and air conditioning (HVAC) System. The research aimed at modeling and forecasting humidity elements of weather in some Sudan Savanna State (SSS) in Nigeria which comprises of Sokoto, Kano, Katsina and Zamfara. Monthly humidity data were collected from World Historical Weather in those states from 2009 to 2022. SARIMA model was employed as the statistical tool and the analysis was carried out with R software. Augmented Dickey Fuller test results proved that there were no unit roots from the datasets and KPSS tests confirmed that all the data were stationary at 5% significant level. Ljung Box results revealed that there were no autocorrelation from the dataset. 80% of the datasets were trained for precision and Akaike Information Criterion (AIC) was used to obtain optimum model identification for each of the selected SSS states while the remaining 20% were used for the main forecast. Optimum models that were identified for years 2024, 2025 and 2026 forecasts are: SARIMA (0,0,2)(0,1,1)^12, SARIMA (0,0,0)(0,1,1)^12, SARIMA (0,0,0)(0,1,1)^12 and SARIMA (0,0,0)(2,1,1)^12 for Sokoto, Kano, Katsina and Zamfara respectively. The implication of the forecast results is that all the SSS may have the same sinusoidal patterns but with different optimum models.

Download pdf