We've got a big global smart prediction model for ocean phenomena, 琅琊 2.0.

News from IT House on June 6, according to press reports, today (6 June) was launched by the Large Model for Global Intelligence Forecasting of Marine Flows, an autonomous model developed by the Institute of Oceanography of the Chinese Academy of Sciences, known as 琅琊 2.0. The large ocean forecasting model will provide intellectual technological support for marine disaster prevention and mitigation, safety and security of shipping, etc.

Global ocean forecasting has been reported to have relied on numerical models in the past, resulting from the transformation of ocean processes into numerical equations and large-scale calculations. This model is relatively low-precision, high-cost calculations and low-frequency updates. The large smart forecast model addresses two problems, one faster and the other more accurate.

IT House noted that the “twilight” 2.0 was based on the “twilight” 1.0 issued in 2024, which further expanded from ocean state variable forecasting to intelligent forecasting of complex ocean phenomena, with six vertical models developed around six types of phenomena, such as typhoons, precipitation, storm surges, sea ice, and so on, building multi-species, systematized prediction capabilities.

For example, the Typhoon Forecast Model, which integrates information on the historical evolution of the atmospheric marine environment field, satellite cloud maps and typhoons, increases the ability to forecast the speed and intensity of typhoons 24 hours a day in response to fast-enhanced and abrupt turnovers; the Rainwater Forecast Model, which is based on global precipitation satellite observations, learns about the temporal and spatial evolution of precipitation and predicts future changes to support early warnings of heavy rainfall and extreme precipitation; and the Sea Ice Forecast Model, which is oriented towards the need for passage through the Arctic channel and allows for the rapid prediction of Arctic sea ice at a 3 km resolution and above the monthly scale, which supports the analysis of sea ice margins, sea ice range statistics and sea lane safety studies.

Extend Reading

The version of “琅琊” 1.0 was released on 28 December 2024 and developed by the Institute of Marine Research of the Chinese Academy of Sciences as a new generation of large artificial intelligence models for forecasting ocean state variables, allowing one-off forecasts of future global ocean state variables such as temperature, salinity and currents for 1 to 7 days.

In recent years, we have launched a number of large-scale weather- and ocean-related models, such as the first global South China Sea-air two-way coupling smart model, Flying Fish-1.0, released in February 2026, and the Global Weather Magnificent Model, Volvo 2.0, released in June 2024, which is oriented towards weather navigation.

In the area of meteorological AI, on 19 December 2025, the China Meteorological Administration released the weather artificial intelligence science model, Wind Source, which allows direct global weather forecasting based on observational data and has an autonomous intellectual property right to end-to-end scientific model base.

Advertising statements: The external jump links (including not limited to hyperlinks, 2D codes, passwords, etc.) contained in the text are used to convey more information and save time for selection purposes only for reference purposes, which are included in all IT House articles.