guest ::
| Home > Publications database > HClimRep: A Foundation Model for Capturing the Atmosphere, Ocean, and Sea Ice Interactions |
| Home > Publications database > HClimRep: A Foundation Model for Capturing the Atmosphere, Ocean, and Sea Ice Interactions |
| Poster (After Call) | FZJ-2025-05488 |
; ; ; ;
2025
Abstract: Climate change presents critical challenges to ecosystems and human society. Accurate projections of climate change and its consequences are essential for assessing climate policies and developing proactive strategies to mitigate extreme weather events. While traditional climate models based on fluid dynamics and radiative transfer have provided valuable information, they face limitations such as inherent biases, coarse resolution, and structural errors. Moreover, these models are computationally intensive, dependentof physical constraints or governing equations, and it is therefore impossible to explore a wide range of policy scenarios and generate actionable climate information at the desired high resolutions. Building on the success of AtmoRep, a foundation model for atmospheric dynamics, we propose HClimRep, a novel fully data-driven global climate model that seeks to capture complex interactions of the atmosphere, ocean, and sea ice to create a realistic climate simulation, including the stratospheric ozone. Upon completion of phase 1 of the model development, we have successfully built a first HClimRep prototype that integrates unstructured ocean gridded data and includes stratospheric dynamics. The phase 2 of the model development leverages insights gained from phase 1, combined with the advancements of the newly launched ECMWF WeatherGenerator (WGen) prototype, to drive further progress. In particular, we use the WGen prototype to significantly facilitate the incorporation of new climate data modalities and achieve long iterative forecasts (rollouts). Further key challenges of our initiative include the establishment of stable seasonal-to-decadal climate forecasts, the accurate simulation of future climates by learning CO2 forcing as well as the employment of petabyte-scale data on multiple resolutions and grids. Our main project objective is the creation of a generalizable large-scale foundation model, which will serve as a basis for various downstream climate-related applications and products from stratospheric warmings forecasts.to tropical cyclone climatology and hydrological downscaling.
|
The record appears in these collections: |
PDF