1,1,2,1,3,4,5,7,6,8,6,5,9,7,1,10- 1Barcelona Supercomputing Center, Earth Sciences, Barcelona, Spain
- 2Universitat Politècnica de Catalunya, Project and Construction Engineering, Barcelona, Spain
- 3Finnish Meteorological Institute, Helsinki, Finland
- 4Leipzig Institute of Meteorology, Leipzig University, Leipzig, Germany
- 5École Polytechnique Fédérale de Lausanne, School of Architecture, Civil and Environmental Engineering (ENAC), Laboratory of Atmospheric Processes and their Impacts (LAPI), Lausanne, Switzerland
- 6Finnish Meteorological Institute, Kuopio, Finland
- 7Royal Netherlands Meteorological Institute, De Bilt, Netherlands
- 8University of Eastern Finland, Kuopio, Finland
- 9Foundation for Research and Technology, Center for the Study of Air Quality and Climate Change (C-STACC), Institute of Chemical Engineering Sciences (ICE-HT), Patras, Greece
- 10ICREA, Catalan Institution for Research and Advanced Studies, Barcelona, Spain
Secondary ice production (SIP) is increasingly recognized as a key regulator of ice crystal number concentrations and cloud phase in mixed-phase clouds (MPCs). In the EC-Earth3-AerChem model, we recently showed that SIP, implemented via a machine-learning-based parameterization, may strongly amplify ice crystal numbers in MPCs, particularly in regions with weak primary ice nucleation such as the Southern Ocean, and can substantially modify cloud phase partitioning and radiative effects. However, those findings were obtained within a model configuration with known limitations in cloud microphysics and supersaturation treatment, motivating their re-examination in the next generation of EC-Earth.
Here we present the implementation of SIP in EC-Earth4, using its new atmospheric core OIFS48r1, which features major updates to mixed-phase cloud microphysics. Building on the EC-Earth3 framework, we implemented in OIFS48r1 aerosol-aware immersion freezing with a machine-learning-based SIP parameterization (RaFSIP), allowing ice multiplication to respond dynamically to cloud thermodynamic and microphysical conditions. This configuration provides, for the first time in EC-Earth4, a physically consistent link between aerosol-controlled primary ice formation and secondary ice amplification.
The model is evaluated against satellite-derived cloud properties from MODIS and CALIPSO, and radiative fluxes from CERES-EBAF. The experimental design enables us to quantify how SIP modifies ice crystal number concentrations and liquid–ice phase partitioning relative to both temperature-based and aerosol-aware primary ice nucleation. Before introducing aerosol-driven ice nucleation and SIP, OIFS48r1 already shows substantial baseline improvements relative to EC-Earth3, including reduced biases in liquid and ice water paths across latitudes. These improvements provide a more robust framework for isolating the climatic role of SIP in a next-generation model.
By extending the SIP analysis from EC-Earth3 to EC-Earth4, this work establishes a consistent modelling framework to assess how secondary ice production interacts with aerosol-controlled primary ice formation, paving the way for more reliable projections of mixed-phase cloud feedbacks in future climate simulations.
How to cite: Costa-Surós, M., Chatziparaschos, M., Gonçalves Ageitos, M., Vacondio, S., Bergman, T., Georgakaki, P., Holopainen, E., Huijnen, V., Kokkola, H., Laakso, A., Le Sager, P., Nenes, A., van Noije, T., Wu, L., and Pérez García-Pando, C.: The Role of Secondary Ice Production in Shaping Mixed-Phase Clouds in EC-Earth4, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17587, https://doi.org/10.5194/egusphere-egu26-17587, 2026.
