news

SASIP November Newsletter

Follow Us

Science Spotlights

Generative Diffusion for Regional Surrogate Models From Sea-Ice Simulations

In their new paper published in Journal of Advances in Modeling Earth Systems, Tobias Finn et al. introduce generative AI, which has similarly revolutionised image generation from text, to predict the evolution of sea ice in a region north of Svalbard. Trained on more than 20 years of neXtSIM data, the AI significantly enhance the accuracy and detail compared to other forecast models. Furthermore, it accurately reflects physical processes similar to the physical model neXtSIM. Therefore, they are optimistic that this technology can lead to fast and physically accurate predictive models derived purely from data.

Investigating ocean circulation dynamics through data assimilation: A mathematical study using the Stommel box model with rapid oscillatory forcings

Theoretically climate change could result in a reversal of the Atlantic Meriodional Overturning Circulation resulting in decreasing temperatures over Europe. In this new article co-authored by Ivo Pasmans and published in Chaos, they assimilated ocean temperature and salinity observations in a simple Stommel to find the model parameters and determine how likely it is that such a reversal happens before 2100.

Science Update

Data-driven emulation of melt ponds on Arctic sea ice

The formation and evolution of sea ice melt ponds are complex, insufficiently understood and are not traditionally included in climate models. In this pre-print submitted to the Cryosphere, Simon Driscoll et al. create the first observationally based machine learning model of melt pond fraction that is also a ready and viable candidate to be included in these models. They set up the problem avoiding necessity of conservation properties of more poorly observed quantities and target the most crucial aspect of melt ponds for albedo. Their emulator captures shifts in melt pond fraction that may be associated with a climate change caused regime shift in sea ice. They discuss how this work can be extended.

News

Welcome to the new team members!