The researchers’ approach for modeling sea ice extent transforms complex Arctic data (left) into simpler rhythmic patterns (center), which are then modeled as coupled oscillators (right) to generate accurate forecasts, mimicking how different natural cycles interact. Credit: Kondrashov et al.

Arctic sea ice has large effects on the global climate. By cooling the planet, Arctic ice impacts ocean circulation, atmospheric patterns, and extreme weather conditions, even outside the Arctic region. However, climate change has led to its rapid decline, and being able to make real-time predictions of sea ice extent (SIE)—the area of water with a minimum concentration of sea ice—has become crucial for monitoring sea ice health.

New real-time forecasts for sea ice

In Chaos, researchers from the United States and the United Kingdom report accurate, real-time predictions of SIE in Arctic regions. Sea ice coverage is at its minimum in September, making the month a critical indicator of sea ice health and the primary target of the work.

"Indigenous Arctic communities depend on the hunting of species like polar bears, seals, and walruses, for which sea ice provides essential habitat," said author Dimitri Kondrashov.

"There are other economic activities, such as gas and oil drilling, fishing, and tourism, where advance knowledge of accurate ice conditions reduces risks and costs."

The researchers’ approach treats sea ice evolution as a set of atmospheric and oceanic factors that oscillate at different rates—for example, climate memory at long timescales, annual seasonal cycles, and quickly changing weather—while still interacting with one another.

Regional subdivisions (left) and seasonal decline of total Arctic sea ice extent during summer melt (right), highlighting 2024, 2023, and the record low of 2012. Credit: Kondrashov et al.

They used the National Snow and Ice Data Center’s average daily SIE measurements from 1978 onward to find the relationships between these factors that affect sea ice.

Testing their prediction method live in September 2024, and retroactively for Septembers of past years, the group confirmed their technique is generally accurate and can capture effects from subseasonal to seasonal timescales. They predicted SIE ranging from one to four months out and found their predictions outperformed other models.

Accuracy, limits, and future improvements

In general, long-term climate forecasts tend to be easier and more reliable than short-term predictions. However, by incorporating regional data into their model, the researchers were able to improve short-term ice and weather estimates.

"The model includes several large Arctic regions composing pan-Arctic," said Kondrashov. "Despite large differences in sea ice conditions from year to year in different regions, the model can pick it up reasonably accurately."

The group plans to improve their model by including additional oceanic and atmospheric variables, such as air temperature and sea level pressure.

These variables can cause fast changes and short-term fluctuations that are not currently reflected in the model, and the researchers hope these additions will further enhance the predictability of summertime Arctic sea ice.

Publication details

Accurate and robust real-time prediction of September Arctic sea ice, Chaos: An Interdisciplinary Journal of Nonlinear Science (2026). DOI: 10.1063/5.0295634

Journal information: Chaos

Citation: Accurately predicting Arctic sea ice in real time (2026, February 3) retrieved 3 February 2026 from https://phys.org/news/2026-02-accurately-arctic-sea-ice-real.html

This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.