Modeling Atmospheric Variables Influencing Extreme Rainfall Events in Makassar City During February 2025

Extreme rainfall events have become increasingly frequent and intense in tropical urban areas, posing serious challenges for disaster management and early warning systems. This study aims to identify the key atmospheric variables influencing extreme rainfall intensity in Makassar City, Indonesia, during the period of 7–15 February 2025, focusing on a major event that occurred on 11 February. Using secondary meteorological data from NASA POWER, a Multiple Linear Regression (MLR) model was constructed with a stepwise selection method in MATLAB to determine the most significant predictors. The analysis revealed six significant variables: relative humidity, air pressure, wind speed, wind direction, temperature, and specific humidity. The model demonstrated strong predictive performance, achieving a coefficient of determination (R²) of 0.676 and a Root Mean Square Error (RMSE) of 11.265 mm. Pearson correlation analysis confirmed a strong linear relationship between observed and predicted rainfall values. Relative humidity, specific humidity, and wind speed showed positive correlations with rainfall intensity, while air pressure and temperature were negatively correlated, and wind direction exhibited negligible influence. These results indicate that atmospheric moisture and dynamic variables play a crucial role in triggering extreme rainfall events, while some factors such as temperature and wind direction have a lesser impact. The findings highlight the potential of using stepwise regression modeling as a tool for short-term rainfall forecasting and urban flood mitigation in tropical coastal environments.