Urban centers face unprecedented challenges due to climate change and increasing environmental volatility. In this study, we developed a hybrid framework to evaluate climate-related stress in Ankara, Türkiye. To achieve this aim, a long-term monthly dataset from 1997 to 2025 is analyzed. Collected via the Meteostat library, the data were processed through the Entropy Weighting Method to objectively rank the impact of six variables, including temperature extremes and precipitation. By building a composite Climate Stress Index, we quantified monthly risk levels. After that, we applied a Random Forest regressor to capture the complex, non-linear patterns within the weather data. Results reveal that precipitation is the primary factor driving climatic stress in Ankara; also, the machine learning model performed with remarkable precision. Performance metrics show an outstanding result. In that aspect, an R2 score of 0.9881 and an RMSE of 0.0193 are observed. These results demonstrate that the proposed hybrid approach can serve as a practical, automated tool for local authorities to monitor and manage regional climate risks more effectively.

Keywords

Climate Stress Index (CSI), Entropy-Based Methods, Random Forest Regression, Regional Climate Assessment, Ankara.