Periodic Reporting for period 1 - CleanCloud (Clouds and climate transitioning to post-fossil aerosol regime)
Reporting period: 2024-01-01 to 2025-06-30
Aerosol-cloud interactions (ACI) remain the largest source of uncertainty in past, present, and future radiative forcing, impeding credible climate projections. ACI effects are expected to change dramatically as we enter a post-fossil world, characterized by strong reductions in anthropogenic aerosol emissions but with increasingly larger impacts from natural aerosols. Although we expect cleaner clouds compared to today, ACI in this post-fossil state may considerably differ from preindustrial conditions, owing to shifts in climate and changes in sources region characteristics. CleanCloud will address the major gaps impeding robust ACI assessments, improve their representation in current and next generation kilometer-scale climate models , quantify and understand their regional and temporal effects, and how they will evolve in the transition to the post-fossil regime. To accomplish this, CleanCloud will carry out targeted field experiments in European climate hotspots, develop state-of-the-art algorithms to obtain new proxies and diagnostics for key ACI-related processes, contribute to the calibration and validation of upcoming satellite missions, improve and better constrain kilometer- and large-scale climate models using advanced machine learning, data assimilation, confronting perturbed physics ensembles with existing and new satellite and in-situ data, assess the role of aerosols in the life cycle of convective systems, focusing on precipitation formation and the impacts on the hydrological cycle, and enhance the exploitation of data centres, measurement programs, international campaigns, laboratory studies, and models.
CleanCloud conducted extensive field campaigns to study aerosol–cloud interactions (ACI) in two key regions: the Arctic and the Mediterranean. At the Villum Research Station in northeast Greenland, seasonal campaigns in spring and summer captured the region’s sharp seasonal contrasts and its vulnerability to long-range pollution. The Arctic campaign advanced understanding of organic aerosol sources (biogenic, wildfire, Arctic haze) and demonstrated reliable retrieval of cloud droplet number concentrations through remote sensing and surface aerosol measurements.
The Mediterranean campaign, CHOPIN, was carried out at Mt. Helmos in Greece, a location influenced by diverse air masses and aerosol types including wildfire smoke, pollution, sea salt, dust, and biological particles. Initially planned for three months, it was extended to eight months and attracted broad collaboration. Key findings include distinguishing aerosol origins (boundary layer vs. free troposphere), identifying cloud-processed organic aerosols, and developing a novel parameterization for ice-nucleating particles (INPs) that outperforms existing models. The campaign also demonstrated the potential of lidar and other remote sensing methods for INP retrieval. Data from these campaigns, along with chamber studies, are being submitted to ACTRIS.
CleanCloud became part of ESA’s EarthCARE Cal/Val program through the EC3 project, providing data from Arctic and Mediterranean campaigns. Kilometer-scale global and regional models have been prepared with remote sensing forward operators and convective life-cycle constraints. Source–receptor analyses traced aerosol origins, showing global models often underestimate natural aerosol sources in the Arctic. To improve representation, new reanalysis datasets and receptor models were developed, including a high-resolution CCN dataset interpolated onto air mass trajectories. Studies also explored how land cover changes and emission reductions affect BVOC emissions and SOA formation, underlining feedbacks between climate and aerosols.
New parameterizations were developed for Earth System Models, including machine-learning emulators of ACI in convective precipitation and updraft velocity schemes based on large-eddy simulations. Global INP datasets informed empirical parameterizations revealing regional source differences. Models and observations were integrated through Perturbed Parameter Ensembles in ECHAM-HAM, ICON-HAM, and EC-Earth, with ICON-HAM using nudging for improved meteorological alignment. Historical analyses of aerosol radiative forcing combining satellite and CMIP6 data showed a shift from increasing to decreasing forcing in the early 21st century, with strong regional variation. Machine learning highlighted cloud fraction and droplet concentrations as key drivers of top-of-atmosphere albedo changes.
Near-term climate projections from RAMIP simulations assessed regional climate effects of aerosol reductions, notably in East Asia where air quality improvements contribute to accelerated warming. Source–receptor analyses with UKESM1.0 further revealed future changes in aerosol transport and sources.
Finally, CleanCloud developed the CleanCloud Metabase (CCM), a search engine hosted at AU-ENVS that connects users to open aerosol–cloud datasets from multiple databases. While it does not host data directly, CCM enables integrated access across platforms. Currently in development and accessible upon request, it will be made public via the CleanCloud project webpage once finalized.
The Mediterranean campaign, CHOPIN, was carried out at Mt. Helmos in Greece, a location influenced by diverse air masses and aerosol types including wildfire smoke, pollution, sea salt, dust, and biological particles. Initially planned for three months, it was extended to eight months and attracted broad collaboration. Key findings include distinguishing aerosol origins (boundary layer vs. free troposphere), identifying cloud-processed organic aerosols, and developing a novel parameterization for ice-nucleating particles (INPs) that outperforms existing models. The campaign also demonstrated the potential of lidar and other remote sensing methods for INP retrieval. Data from these campaigns, along with chamber studies, are being submitted to ACTRIS.
CleanCloud became part of ESA’s EarthCARE Cal/Val program through the EC3 project, providing data from Arctic and Mediterranean campaigns. Kilometer-scale global and regional models have been prepared with remote sensing forward operators and convective life-cycle constraints. Source–receptor analyses traced aerosol origins, showing global models often underestimate natural aerosol sources in the Arctic. To improve representation, new reanalysis datasets and receptor models were developed, including a high-resolution CCN dataset interpolated onto air mass trajectories. Studies also explored how land cover changes and emission reductions affect BVOC emissions and SOA formation, underlining feedbacks between climate and aerosols.
New parameterizations were developed for Earth System Models, including machine-learning emulators of ACI in convective precipitation and updraft velocity schemes based on large-eddy simulations. Global INP datasets informed empirical parameterizations revealing regional source differences. Models and observations were integrated through Perturbed Parameter Ensembles in ECHAM-HAM, ICON-HAM, and EC-Earth, with ICON-HAM using nudging for improved meteorological alignment. Historical analyses of aerosol radiative forcing combining satellite and CMIP6 data showed a shift from increasing to decreasing forcing in the early 21st century, with strong regional variation. Machine learning highlighted cloud fraction and droplet concentrations as key drivers of top-of-atmosphere albedo changes.
Near-term climate projections from RAMIP simulations assessed regional climate effects of aerosol reductions, notably in East Asia where air quality improvements contribute to accelerated warming. Source–receptor analyses with UKESM1.0 further revealed future changes in aerosol transport and sources.
Finally, CleanCloud developed the CleanCloud Metabase (CCM), a search engine hosted at AU-ENVS that connects users to open aerosol–cloud datasets from multiple databases. While it does not host data directly, CCM enables integrated access across platforms. Currently in development and accessible upon request, it will be made public via the CleanCloud project webpage once finalized.
The CleanCloud CHOPIN campaign showed that biological ice nucleating particles (INPs) exhibit a strong diurnal cycle in the Boundary Layer, with late-afternoon peaks linked to biological activity, higher temperature, and humidity. Active at warmer temperatures (−13.5 °C), biogenic INPs can initiate ice formation and precipitation. Data confirmed variability in bioaerosol concentrations during cloud events and their influence on cloud ice content, supporting their role in triggering afternoon precipitation in orographic systems.
High-resolution modeling with WRF-EPFL (1 km domains) explicitly represented microphysics and secondary ice production, reproducing extreme precipitation validated by radar. While bulk parameterizations captured average rainfall, extremes with deep convection required fine-scale simulations. Fusion of radar with WRF outputs (via CR-SIM) showed skewness in fall velocity distributions, signatures of riming and aggregation, resolved only at high resolution.
New INP parameterizations were developed for Earth System Models. A Doppler lidar method, using aerosol backscatter and depolarization, matched in situ PINE data at colder, dust-dominated temperatures (−23.5 °C). A fluorescence lidar method quantified biological INPs at warmer regimes (−13.5 °C), showing strong agreement with WIBS-INP benchmarks. Together, these approaches demonstrate the feasibility of particle-type-specific INP retrievals, paving the way for satellite applications and improved climate model integration.
High-resolution modeling with WRF-EPFL (1 km domains) explicitly represented microphysics and secondary ice production, reproducing extreme precipitation validated by radar. While bulk parameterizations captured average rainfall, extremes with deep convection required fine-scale simulations. Fusion of radar with WRF outputs (via CR-SIM) showed skewness in fall velocity distributions, signatures of riming and aggregation, resolved only at high resolution.
New INP parameterizations were developed for Earth System Models. A Doppler lidar method, using aerosol backscatter and depolarization, matched in situ PINE data at colder, dust-dominated temperatures (−23.5 °C). A fluorescence lidar method quantified biological INPs at warmer regimes (−13.5 °C), showing strong agreement with WIBS-INP benchmarks. Together, these approaches demonstrate the feasibility of particle-type-specific INP retrievals, paving the way for satellite applications and improved climate model integration.