An international research team led by the Leibniz Institute for Tropospheric Research (TROPOS) has identified a critical gap in current climate models regarding urban aerosols. The study, published in Communications Earth & Environment, reveals that aerosols in polluted regions grow faster and absorb more water than previously assumed, a factor that significantly impacts cloud formation and global cooling trends. While clouds play a vital role in Earth's energy balance, their ability to reflect solar radiation depends on hygroscopicity, or the capacity of particles to absorb water. Until now, climate models have represented this process too simplistically, particularly in chemically complex environments. The researchers utilized explainable machine learning to estimate size-dependent hygroscopicity across diverse atmospheric conditions. By integrating data on chemical composition, particle number size distribution, and meteorology from ten global sites spanning over a decade, the team captured the complexity of aerosol mixing states where conventional methods fail. The dataset included measurements from highly polluted megacities such as Delhi, Cairo, and Beijing, alongside remote locations in Germany, Namibia, and the Atlantic Ocean. The findings indicate that in heavily polluted urban areas, externally mixed particles interact with aged aerosols to cause rapid hygroscopic growth. This increased water uptake enhances the particles' ability to scatter sunlight and form clouds, potentially explaining why regions like South Asia and parts of Africa are warming more slowly than predicted. Dr. Ajit Ahlawat of TU Delft noted that these dynamics also have implications for public health, as increased water absorption contributes to smog formation, a factor confirmed by drone measurements in Delhi. Current models often assume ideal internal mixing of particles and ignore variations in particle size and source, leading to the largest errors in urban regions. In contrast, the new data-driven approach provides a more accurate representation of aerosol behavior. Prof. Mira Pöhlker of TROPOS emphasized that these regional estimates can alter direct radiative forcing calculations by up to plus or minus 0.1 watts per square meter. While this value may seem small, it is substantial on a global scale and could change both the magnitude and direction of aerosol-radiation interactions in future simulations. The study builds on previous work establishing that aerosol hygroscopicity is primarily determined by the proportion of organic and inorganic substances. However, the new method moves beyond global averages to offer region-specific parameterizations essential for reducing uncertainties in next-generation climate models. The research team hopes their algorithm will be integrated into global models, ultimately leading to more precise climate projections and a better understanding of the regional cooling effects observed in some of the world's most polluted areas.