Closing the gap: More precise observations supporting sea ice forecasting in the far north

Forecasts of sea ice are a lifeline for the Alaskan fishing industry and America’s growing icebreaker fleet who navigate vessels within or near sea ice, as well as for Alaskan coastal communities who rely on sea ice for subsistence activities and travel.

A recent study by NOAA’s Physical Sciences Laboratory (PSL) and the University of Colorado Boulder’s Cooperative Institute for Research In Environmental Sciences (CIRES) establishes a new standard for air-sea ice observations in the Arctic that represents a significant step towards improving sea ice prediction models. The research is based on the 2019/2020 Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition in which PSL and CIRES were significantly involved.

Using observing systems developed by PSL, the NOAA Global Monitoring Laboratory (GML), and CIRES for the expedition, researchers tracked the growth and melt of sea ice by measuring the amount of energy it absorbed and released, calculating the sea ice’s “energy budget.” This approach was chosen because it mirrors the way that advanced forecast models, like NOAA’s Unified Forecast System (UFS), predict ice behavior. Achieving “closure,” where the measured energy budget matches the physical change in ice mass, is notoriously difficult and extraordinarily rare due to issues of logistics, sampling, and timing.

The new study demonstrates that the MOSAiC observations achieved remarkable closure to within about 5% over 10 months of seasonal changes by comparing the calculations based on the energy budget measurements to direct observations of the ice thickness made by numerous, independent collaborators.

The findings point to specific ways to improve the forecast models. The study introduces a new diagnostic tool, the “atmospheric state effect,” that will help evaluate and address long-standing model biases associated with Arctic clouds. Arctic clouds have a significant impact on the sea ice energy budget, and thereby also sea ice growth, so reducing these cloud model biases will improve the models’ skill in predicting how the sea ice will evolve through a forecast. The study also highlights specific areas for future focused model improvement. For instance, model components responsible for calculating how the ice is influenced by atmospheric turbulence are found to be more sensitive to certain weather conditions and physical relationships than previously thought.

Ultimately, this work represents a significant step towards closing the precision measurement gap in the far north, improving the performance of models. With improved models comes more accurate sea ice prediction, strengthening the nation’s industries and communities in the Arctic.

Acknowledgements

This research is supported by the NOAA Global Ocean Monitoring and Observing Program, as well as the U.S. Department of Energy Atmospheric System Research Program , the U.S. National Science Foundation Arctic System Science Program , and the NOAA Physical Sciences Laboratory.

From field to data

The surface energy budget measurements from this study came from Atmospheric Surface Flux Stations, targeted, semi-autonomous instrument suites built on skis developed by PSL/CIRES for the MOSAiC expedition and other applications. Learn more

A PSL/CIRES-developed Atmospheric Surface Flux Station on ice in the Arctic with the German icebreaker RV Polarstern in the background during the MOSAiC expedition in 2020 (Photo: Matthew Shupe, CIRES/PSL)

Posted: April 15, 2026
Back to PSL News
Signup for PSL email updates