Collaboration projects

Seven projects conducting data-intensive research between research groups in Finland and Colorado, and Finland and Japan finished in September 2025. Read more about the projects below.

Four projects on environmental research, material discovery and moon regolith engaged researchers in Colorado

Dr. Tuula Aalto from the Finnish Meteorological Institute’s (FMI) Carbon Cycle group led one of the selected projects on CO2 Sources and Sinks. Her team collaborated with a research group from the Cooperative Institute for Research in the Atmosphere (CIRA) and Colorado State University (CSU). The project developed advanced modelling tools that are specifically designed to make use of emerging high-intensity satellite data streams, which until now have been difficult to fully exploit. By integrating these datasets into more efficient and faster inversion systems, researchers will be able to produce carbon budget estimates with much higher spatial and temporal accuracy than previously possible. The societal benefits of creating these models are various: more accurate emission reporting, stronger environmental governance, better targeting of mitigation measures, and ultimately faster progress toward a green transition.

Scientific publications and public datasets:

Isomäki K., McGrath M.J., Backman L., Leskinen J., Berchet A., Broquet G., Fortems-Cheiney A., Junttila V., Leppänen A., Lindqvist H., Mengistu A., Mäkelä A., Raivonen M., Thölix L. & Aalto T. 2024: Assessing the role of terrestrial ecosystems in Finland’s total CO2 balance through a comparison of top-down and bottom-up estimates. Boreal Env. Res. 29: 77–102. https://www.borenv.net/BER/archive/pdfs/ber29/ber29-077-102.pdf

A. Mikkonen, H. Lindqvist, J. Peltoniemi, J. Tamminen: “Non-Lambertian snow surface reflection models for simulated top-of-the-atmosphere radiances in the NIR and SWIR wavelengths”, Journal of Quantitative Spectroscopy and Radiative Transfer, 315, (2024). A. Mikkonen, A. Koskinen, J. Tamminen, and H. Lindqvist, “Scattering graph method for 3D radiative transfer,” Opt. Express 33, 35489-35509 (2025).

Thanwerdas, J., Berchet, A., Constantin, L., Tsuruta, A., Steiner, M., Reum, F., Henne, S., and Brunner, D.: Improving the ensemble square root filter (EnSRF) in the Community Inversion Framework: a case study with ICON-ART 2024.01, Geosci. Model Dev., 18, 1505–1544, https://doi.org/10.5194/gmd-18-1505-2025, 2025.

Tsuruta, A., Kuze, A., Shiomi, K., Kataoka, F., Kikuchi, N., Aalto, T., Backman, L., Kivimäki, E., Tenkanen, M. K., McKain, K., García, O. E., Hase, F., Kivi, R., Morino, I., Ohyama, H., Pollard, D. F., Sha, M. K., Strong, K., Sussmann, R., Te, Y., Velazco, V. A., Vrekoussis, M., Warneke, T., Zhou, M., and Suto, H.: Global CH4 fluxes derived from JAXA/GOSAT lower-tropospheric partial column data and the CarbonTracker Europe-CH4 atmospheric inverse model, Atmos. Chem. Phys., 25, 7829–7862, https://doi.org/10.5194/acp-25-7829-2025, 2025.

The TURSCA software: https://github.com/amikko/tursca

OCO2-MIP data: https://gml.noaa.gov/ccgg/OCO2_v11mip/

Professor Antti Karttunen from the School of Chemical Engineering at Aalto University led a project with the aim to use high-performance computing to accelerate materials discovery for clean energy and zero emission vehicles. His Inorganic Materials Modelling research group worked in cooperation with researchers from the Department of Physics, Colorado School of Mines. The project successfully developed and applied new computational methodology that made it possible to investigate the heat transfer properties of thermoelectric alloy materials, where computational workflows combining quantum chemistry and machine learning were implemented on the LUMI supercomputer. The developed methodology can be used to discover new inorganic materials for clean energy. The project also established a completely new and fruitful collaboration with the Colorado School of Mines, and thanks to its success the researchers are committed to continuing it also in the future.

Scientific publications and public datasets:

Heat transport properties of PbTe1–xSex alloys using equivariant graph neural network interatomic potential, Conley, K.; Gerber, C.; Novick, A.; Berriodi, T.; Toberer, E. S.; Karttunen, A. J. Mater. Horiz. 2025, 12, 8084–8094. https://doi.org/10.1039/D5MH00934K

Research data related to the publication: https://zenodo.org/records/15718966

Dr. Antti Penttilä from the Department of Physics at the University of Helsinki led a project addressing the multi-scale problem in (computational) light scattering and collaborated with partners in the Space Science Institute in Colorado. The project developed advanced tools that help simulate how light behaves when it hits granular surfaces, improved models for surface reflectance, and conducted experiments that show how the polarization of light from simulated lunar soil depends mainly on the angle between the light source and the observer. These results can contribute to addressing all kinds of remote sensing problems, correct interpretation of planetary science data, and potentially even to space resource utilization and manned lunar exploration, which require the capability to map compositional information.

Scientific publications and public datasets:

Photopolarimetric characterization of rough surfaces in regolith simulants by Frattin E, García-Izquierdo FJ, Martikainen J et al. (incl. Videen G, Penttilä A, Muinonen K, 2025), Astronomy and Astrophysics vol. 697, number 9.
Scattering matrices of particle ensembles analytically decomposed into pure Mueller matrices, Muinonen K and Penttilä A (2024). Journal of Quantitative Spectroscopy and Radiative Transfer vol. 324, number 8.

Research professors Christina Williamson and Risto Makkonen from the Finnish Meteorological Institute (FMI) and the University of Helsinki worked together with Coloradan partners from NOAA Chemical Sciences Laboratory, University of Colorado and the National Centre for Atmospheric Research (NCAR), to conduct research on understanding and constraining aerosol forcing. They created emulators for Earth System Model EC-Earth and developed and applied methods to understand uncertainties in simulated aerosol forcing and to further constrain it with a comprehensive suite of real-world observations of atmospheric aerosols. The project made great advancements in running Earth System Models in massive GPU systems such as LUMI, which will make it easier to utilise these models in LUMI and other EuroHPC systems in the future. Aerosols can influence how much sunlight reaches the planet, and thus the results on reducing the aerosol uncertainty in EC-Earth will improve climate prediction, which is key to mitigating and adapting to climate change.

Three projects with researchers in Japan studied topics related to ice sheet modelling, machine learning and supernovae

Dr. Rupert Gladstone’s project (Arctic Centre of the University of Lapland) developed computational models to address the question of how big an impact can the effect of global warming on polar ice sheets have on sea level rise, collaborating with various researchers from Japan and beyond. The work focused on subglacial hydrology and ice dynamic models on the scale of the whole Antarctic, as well as coupled ice dynamic-hydrology models and ice-ocean processes in more limited regions. These methods can contribute to better estimates and projections of sea level rise in the future, which enables policymakers to implement more comprehensive plans to address its ramifications.

Scientific publications and public datasets:

Chen Zhao, Rupert Gladstone, Thomas Zwinger, Fabien Gillet-Chaulet, Yu Wang, Justine Caillet, Pierre Mathiot, Leopekka Saraste, Eliot Jager, Benjamin K.Galton-Fenzi, Poul Christoffersen, and Matt A. King (2025) ‘Subglacial water amplifies Antarctic contributions to sea-level rise’ Nat. Commun., vol 16, article 3187. https://doi.org/10.1038/s41467-025-58375-4

Qin Zhou, Chen Zhao, Rupert Gladstone, Tore Hattermann, David Gwyther, and Benjamin Galton-Fenzi (2024) ‘Evaluating an accelerated forcing approach for improving computational efficiency in coupled ice sheet-ocean modelling’, Geosci. Model Dev., Vol 17, Issue 22, Pages 8243-8265. https://gmd.copernicus.org/articles/17/8243/2024/

Yu Wang, Chen Zhao, Rupert Gladstone, Thomas Zwinger, Ben Galton-Fenzi, and Poul Christoffersen (2024) ‘Sensitivity of Future Projections of the Wilkes Subglacial Basin Ice Sheet to Grounding Line Melt Parameterizations’, The Cryosphere, Vol 18, Issue 11, Pages 5117-5137. https://tc.copernicus.org/articles/18/5117/2024/

Yufang Zhang, John C. Moore, Liyun Zhao, Mauro A. Werder, Rupert Gladstone, Michael Wolovick (2024) ‘The role of hydraulic conductivity in the Pine Island Glacier’s subglacial water distribution’, Science of the Total Environment, vol 927, page 172144. https://doi.org/10.1016/j.scitotenv.2024.172144

Seroussi, H. and ISMIP6 members including Gladstone, R.M. (2024) ‘Evolution of the Antarctic Ice Sheet over the next three centuries from an ISMIP6 model ensemble’, Earth’s Future. https://doi.org/10.1029/2024EF004561

Professor Antti Honkela from the University of Helsinki and Professor Samuel Kaski from Aalto University, both part of the Finnish Center for Artificial Intelligence (FCAI), led a project to create an AI assistant to help develop privacy-preserving versions of deep learning models, working together with Japan’s RIKEN. The project developed a novel interactive method to allow users to find the optimal privacy–utility trade-off and explained how to find the best training settings to improve model accuracy while keeping data private by fine-tuning a foundation model using a sensitive dataset. These methods make efficient privacy-preserving machine learning using differential privacy more accessible to scientists and other developers training models on sensitive data, likely leading to broader adoption and improved results.

Scientific publications and public datasets:

Python package jaxdpopt and source code for reproducing the experiments of the scalable training work: https://github.com/DPBayes/Towards-Efficient-Scalable-Training-DP-DL

Dr. Hanindyo Kuncarayakti from the Stellar Explosions research group at the University of Turku led a project on supernovae explosions that aimed to construct physically realistic, state-of-the-art models of SN explosion and interaction with the surrounding medium. His team collaborated with researchers from Kyoto University, University of Tokyo and the National Astronomical Observatory of Japan. The project was the first time such a physical setup was simulated with detailed multi-dimensional radiation-hydrodynamic simulations, and showed that the models could reproduce the observed behavior of the actual supernovae. These results can contribute to other efforts to compare observational data with model predictions, and the model grids developed in the project are expected to be extensively used by the astronomical community following the public release, as they are readily usable without the need of further computing resources.