Predicting rainfall amounts in Nablus city using time series

Israa  Naem Hasan; Hassan  Abu Hassan; Hussein Al-Rimmawi

doi:10.35192/jjoas-n.v17i2.795

Israa Naem Hasan Program of Applied Statistics and Data Science, Birzeit University, Palestine.
Hassan Abu Hassan Program of Applied Statistics and Data Science, Birzeit University, Palestine.
Hussein Al-Rimmawi Geography Department. Birzeit University, Palestine.

DOI: https://doi.org/10.35192/jjoas-n.v17i2.795

Abstract

Background: This research aims to predict the rainfall of the City of Nablus, Palestine, in the years to come to assist in managing water resources in an area where rainfall quantities vary yearly. Box-Jenkins model is used in analyzing the time series, characterized by high prediction accuracy.

Methods: The study is based on rainfall data (from the Palestinian Meteorological Department) of the city of Nablus over 51 years, i.e., from 1970 to 2020.

Results: It is clear from the study that the best rainwater forecasting model is ARIMA (0,07) because the data from this series are consistent over time. The model gave good predictions that are close to the actual values. However, the results indicated that quantities in the coming years will fluctuate. The fluctuation is consistent with the Mediterranean climate, which suffers from climate changes and variations in rainfall levels.

Conclusion: The study recommends using Box and Jenkins models in time series to predict hydrological phenomena. The study recommends using Box and Jenkins models in time series to predict hydrological phenomena.

Keywords: Nablus, Palestine, Prediction, ARIMA Models, Dickey – Fuller Test.