Keywords = Rainfall Forecasting

Monthly rainfall forecasting using artificial neural network model and regression method

Volume 13, Issue 1, July 2026, Pages 125-133

https://doi.org/10.22116/jiems.2026.535479.1608

Paria Soleimani, Shahnaz Piroozfar

Abstract Rainfall plays a vital role in water resource management and is a critical variable for effective decision making and planning. Rainfall strongly controls the dynamic of the water resource. It supports ecosystems, regulates temperature, cleans the atmosphere by removing pollutants, and creates landscapes. Accurate estimation of monthly rainfall is essential for various applications such as flood forecasting, drought management, irrigation planning, and watershed management. In this study, monthly rainfall in Tehran city is predicted using an artificial neural network (ANN) model. ANNs allow to solve complex problems that might have difficulty been solved through the traditional algorithms. The optimal structure is selected based on the valid statistics after testing different numbers of parameters. The results are also compared with regression-based approach. Support vector regression (SVR) aims to find a function that can accurately predict the monthly rainfall. The dataset includes 60 years of meteorological records such as rainfall, humidity, temperature, total monthly sunshine hours, dew point temperature, and wind speed. The results show the artificial neural network method is more proper for predicting monthly rainfall in Tehran due to its lower error rate compared to the support vector regression method. The results of paper can help and show a guidance for the managers to choose the proper method for forecasting of rainfall leading to the management of primary source of fresh water.