The Skoltech Artificial Intelligence Center, together with the Shirshov Institute of Oceanology of the Russian Academy of Sciences (IO RAS), held a presentation at the “Digital Economy” ANO venue of Chione — an intelligent system for short-term forecasts of ice conditions and weather in the Arctic. The service is intended for shipping and oil & gas companies, ports, and situation centers, and combines modern numerical models of the ocean and atmosphere with AI algorithms. Chione produces short-term forecasts for key parameters — ice concentration, thickness, and drift, as well as meteorological and oceanographic conditions — on a horizon of up to 72 hours. The system is based on open models NEMO+SI3, WRF, and WW3.

Opening the event, Professor and Skoltech Vice President for AI Development Evgeny Burnaev spoke about the mission of the AI Center, key development trajectories, and the portfolio of industry projects. Vladimir Vanovskiy, who heads the Hybrid Modeling direction at the Skoltech AI Center, showed which parameters are included in the forecast — ice concentration, thickness, drift and compaction, meteorological and oceanographic conditions, and satellite imagery; explained how the employed models are linked with reanalysis data and satellite observations; and how AI methods, including data assimilation, are used to improve the accuracy and speed of computations.

Alexander Konovalov, who heads the development team at the AI Center, conducted a live demonstration of the service and video cases of practical use, and outlined access and integration formats — the web interface and API connectivity. About ten attendees received test access on site.

“Chione is an example of engineering AI: Rigorous physical models are combined with machine-learning methods to deliver an operational and reproducible forecast for industry. This approach reduces uncertainty in planning, shifts risk management into a proactive mode, and helps make real-time decisions — from ship routing and optimizing icebreaker support costs to supporting situation centers. We design the system to be scalable and integrable: a web interface for day-to-day work, an API for embedding into corporate processes, and on-premises deployment for those who require a sovereign environment,” emphasized Evgeny Burnaev.

“Today, as the Northern Sea Route becomes a major national and international priority, providing transportation between the western and eastern parts of Russia, as well as between Europe and Asia, forecasting not only ice, but also hydrodynamic conditions in the ocean and meteorological conditions in the atmosphere along the NSR route becomes a crucial challenge for scientists. Our joint work, the Chione system, includes hydrodynamic modeling developed at the Institute of Oceanology and machine-learning and AI technologies that help not only assimilate data, but also significantly increase the detail of the hydrodynamic processes. Chione cannot exist in a ‘frozen’ state: it must evolve and be regularly updated, because climate, ice conditions, and dynamics are changing in many parts of the Arctic Basin—all of which must be taken into account,” noted RAS Academician Sergey Gulev, Head of Laboratory at the IO RAS.

Access requests and integration inquiries: chione@skoltech.ru
Project page: https://events.skoltech.ru/chione_ru
System video: https://disk.yandex.ru/i/g9G_OgW8VW18sQ