A new study, led by CMCC researcher Lorenzo Zampieri, sheds light on the shortcomings of models in accurately simulating the heat conduction through the sea ice system during the winter months. Diagnosing these shortcomings using novel observations improves our understanding of the sea ice system and will enable better parameterization for climate models.

A new paper published in the scientific journal Geophysical Research Letters provides insight into the Arctic climate system. Lead author and CMCC researcher, Lorenzo Zampieri, worked with a team of international researchers by combining models and winter observations of Arctic sea ice from the Multidisciplinary Drifting Observatory for the Study of the Arctic Climate (MOSAiC).

Models used to predict the sea ice evolution cannot represent this element in all its physical complexity, and compromises are often needed. These compromises are motivated by the limited computational resources available to run the models and by our limited understanding of the small-scale physical processes taking place in the system. In practice, these compromises translate into parameterizations, empirical formulations utilized in sea ice models to describe the system in a simplified and sustainable manner, while retaining the core information useful for climate assessment.

The present study focuses on the winter heat exchange between the polar ocean, the sea ice, and the atmosphere, which is a key factor for determining how much sea ice forms every winter in the polar regions.

The team’s investigation shows how existing model parameterizations may be falling short in capturing the complexity of heat conduction through the sea ice system. The main reason is that current parameterizations do not account for the variations in sea ice thickness and snow depth at the meter scale and do not consider the heat conduction process as multi-dimensional but rather one-dimensional.

Specifically, current sea ice models assume that heat flows only vertically through the sea ice, whereas this process can also take place horizontally; a failure which the study revealed may cause an underestimation of the total conductive heat flux of 10% or more.

These analyses have been made possible through the availability of novel in-situ measurements from the MOSAiC expedition, a unique observational dataset that provides a comprehensive description of the physical properties of the sea ice during the winter months, a time when the Arctic sea ice is typically inaccessible to scientists.

Critical approaches to model deficiencies, that also result in quantifying unresolved processes, are a key aspect to providing improved sea ice representation. “Learning the nature of these errors is useful because we could formulate corrections for our models, investigate the occurrence of climate feedback mechanisms, and possibly provide more reliable predictions about the current state of the sea ice and its future evolution, which is currently heavily impacted by global warming,” write the authors of the report.

Accurate sea ice models are important tools for informing policymakers and stakeholders about the trajectory of climate change and its potential impacts on global ecosystems and communities. By addressing the deficiencies identified in this study, scientists hope to improve existing models and provide more reliable predictions about the future of Arctic sea ice.

The study is an outcome of the Multiscale Machine Learning In Coupled Earth System Modeling (M²LInES) project, funded by Schmidt Futures, to which Zampieri contributed during his postdoc at the National Center for Atmospheric Research. The results and experiences from this endeavor will be important to guide future sea ice modeling activities at the CMCC.

More information:

Modeling the Winter Heat Conduction Through the Sea Ice System During MOSAiC; Lorenzo Zampieri, David Clemens-Sewall, Anne Sledd, Nils Hutter, Marika Holland; First published: 17 April 2024; https://doi.org/10.1029/2023GL106760