Harnessing ERA5 Reanalysis Data for Improved Long-Term Rainfall Forecasting in Southern Iran

Accurate long-term precipitation forecasts are important for water resource management and disaster prevention. Traditional precipitation forecasting is mainly a statistical analysis method. In recent years, methods based on machine learning have been widely used. However, it is an ongoing challenge to improve the accuracy of precipitation prediction. Therefore, this study proposes a long-term precipitation forecasting method based on decision trees intending to improve the accuracy of precipitation prediction. This study designed algorithm comparison experiments and evaluated various machine learning models for precipitation forecasting in Fars province located in the south of Iran, respectively. In this study, the ERA5 reanalysis dataset is used to make predictions of precipitation status. The results show that the prediction accuracy of XGBoost can reach 72% and the root mean square error is only 0.023K, which greatly improves the accuracy of monthly precipitation prediction in Fars province and also provides a new research idea for the development of machine learning in the field of precipitation prediction.