Leipzig. Aerosols and clouds play a key role in the Earth’s climate budget. However, the extent to which they reflect solar energy depends heavily on how much water the particles can absorb. This so-called hygroscopicity has so far been represented in a simplified manner in climate models. An international research team led by the Leibniz Institute for Tropospheric Research (TROPOS) has now demonstrated through a global study that the models are not precise enough, particularly in urban regions. In chemically complex and polluted regions such as Delhi or Cairo, there is likely to be greater hygroscopic growth and higher water uptake, which could partly explain the observed regional cooling trends or the slower warming on the Asian and African continents, the researchers write in the journal Communications Earth & Environment, published by the Nature Publishing Group.
Particles in the atmosphere have a significant impact on the Earth’s radiation balance: on the one hand, these aerosols directly reflect sunlight and thermal radiation. On the other hand, they also act as cloud condensation nuclei. The amount of water vapour that adheres to the particles has a major effect on cloud formation. The hygroscopicity of aerosols (κ) is one of the key parameters in calculations of radiative forcing and influences the uncertainties in climate projections. Although cloud condensation nuclei have been studied for a long time, the hygroscopic growth of aerosols at sub-saturated conditions remain poorly characterised, particularly in remote and pristine regions.
To address these gaps in knowledge, the researchers developed a method using explainable machine learning (ML) to estimate the size-dependent κ in various atmospheric environments, incorporating observations from ten sites and across several particle sizes ranging from 50 to 300 nanometres. By integrating chemical composition, particle number size distribution and meteorology, the complexity of aerosol mixing states could be captured whilst simultaneously enabling the imputation of data gaps. “Unlike previous regional ML studies, our approach was extended and evaluated using geographically diverse and regionally resolved datasets, thereby improving predictive accuracy and interpretability,” explains Shravan Deshmukh from TROPOS. Machine learning allowed more data to be analysed than usual, and a wide range was covered through a diverse array of measurements. Hygroscopicity measurements using Hygroscopicity Tandem Differential Mobility Analysers (HTDMA) at ground stations span several continents and over a decade: Beijing (China, 2016/17), Cairo (Egypt, 2019/20), Delhi (India, 2020), Goldlauter (Germany, 2010), Henties Bay (Namibia, 2017), Houston (USA, 2021/22), Mahabaleshwar (India, 2020), Melpitz (Germany, 2015), Paris (France, 2022) and the Atlantic (R/V Polarstern, 2011/12).
A significant influence of externally mixed particles on κ was observed, particularly in urban and populated areas where new emissions interact with aged aerosols. “In heavily polluted regions such as megacities in Egypt or India, the particles are likely to grow faster and absorb more water. This could explain why these regions warm up less quickly. Increased hygroscopic growth in such regions also has potential implications for public health due to smog, as we were able to demonstrate through drone measurements in Delhi,” explains Dr Ajit Ahlawat, Assistant Professor at TU Delft. In such areas, conventional models exhibit the greatest errors, as they assume ideal internal mixing and disregard size and source variabilities. This underscores the importance of the chemical composition of the particles. As early as 2023, the team was able to show that, on a global average, hygroscopicity is essentially determined by the proportion of organic and inorganic substances in the aerosol composition.
Building on previous work, our regional estimates provide an improved, data-driven representation of aerosol hygroscopicity. This approach leads to more accurate estimates of negative radiative forcing and offers an alternative to conventional uniform parameterisations,” emphasises Prof. Mira Pöhlker of TROPOS and the University of Leipzig. “Our results highlight the importance of region-specific aerosol parameterisations as a crucial step towards reducing uncertainties in the estimation of direct radiative forcing in next-generation climate models. The use of estimates such as ours can typically alter regional direct radiative forcing by up to ±0.1 watts per square metre, which would be significant on a global scale.” The researchers, therefore, now hope that their new algorithm will be integrated into global models, which could potentially alter both the magnitude and the sign of aerosol-radiation interactions. This could make future climate models more accurate. Tilo Arnhold
Communications Earth & Environment
Meta-analysis
Not applicable
Regional aerosol hygroscopicity influences radiative forcing globally.
7-May-2026
The authors declare no competing interests.