Relative Humidity Forecasts in Tetouan (Morocco) with SARIMA model.

Abstract

Abstract In this study, a seasonal ARIMA model is built using the Box and Jenkins method (1970) to predict the long-term relative humidity in the city of Tetouan. To this end, the monthly average relative humidity data over the period of 1990 to 2022 from the Sania Ramel station are used to build and verify the model. The methodology used in this work toward the development of the model includes five steps: exploratory analysis, model identification, parameter estimation, model validation and prediction. The validity of the model is tested by using the standardized residual plots along with the appropriate statistical tests given by Box and Jenkins. In a second validation step, the predicted values of monthly relative humidity are verified through the use of real data series. After carrying out the necessary checks, the ARIMA(2,1,1)(2,0,0)[12] model proved to be the most effective. Keywords: Time Series, Seasonal ARIMA, Relative Humidity, Forecast.