Prediction of Temperature and Precipitation Changes for Serbia Using Time Series Models with Machine Learning

Abstract

In the past few decades, there has been an evident change in climatic conditions worldwide as well as on the territory of Serbia. Extremely high temperatures, heavy floods, and sudden changes in the weather are increasingly frequent occurrences that bring great social and material damage. Climate change affects many economic sectors, like tourism and agriculture, which are potentially at risk. In Serbia, one of the vital economic sectors is agriculture. In order to act preventive, the main goal of this research was to predict the mean monthly temperature and precipitation for Serbia for periods 2021-2050 and 2071-2100. We collected a dataset titled ERA5 monthly averaged data on single levels from 1940 to present from the Climate Data Store. The dataset was analyzed and prepared to be used with SARIMA(X) and ARIMA(X) methods, which are utilized for prediction. The results that we identified are presented in this paper.