Scientists living in ice camps during an entire year in 1975 (top). Automated instruments attached to sea ice in 2006–2012 (bottom).
New research could help explain why sophisticated climate models struggle to accurately capture the increase in Arctic Ocean freshening. Researchers used yearlong, below-ice observations collected by scientists living in ice camps for an entire year in 1975 (top) and by automated instruments attached to sea ice in 2006–2012 (bottom). Credit: Les Nakashima, National Snow and Ice Data Center (top), Mary-Louise Timmermans (bottom)
Source: Geophysical Research Letters

The surface of the Arctic Ocean is getting fresher. As climate change progresses, shifts in processes such as precipitation and ice melt are reducing the salinity of Arctic surface waters, which could disrupt marine ecosystems and carbon storage. However, computer models designed to help predict the effects of climate change do not accurately reflect real-world observations of Arctic surface salinity—and it’s unclear why.

Now, research by Rosenblum et al. suggests that two generations of one of the most widely used state-of-the-art climate models misrepresent the mixing of fresh Arctic surface waters with deeper waters. This error, the authors say, contributes to the underestimation of declining surface salinity in the region.

These new insights arose from an analysis of two sophisticated Earth system models—Community Earth System Model (CESM) versions 1.1 and 2—that are used to help scientists better predict Earth’s future. The researchers used the models to estimate the salinity of the Arctic Ocean’s Canada Basin in 1975 and from 2006 to 2012 and compared those estimates to real-world measurements of salinity for the same region and time periods.

Unlike previous studies that have explored underestimation of Arctic salinity, the researchers considered season-by-season changes in salinity, not just annual data. They used yearlong, below-ice observations that were collected by scientists in 1975 during the Arctic Ice Dynamics Joint Experiment and by automated instruments attached to sea ice from 2006 to 2012.

The analysis showed that both models’ estimates did not match the observations. Although the estimates are accurate for 1975, they overestimate salinity from 2006 to 2012. To figure out why the model estimates did not match the observations, the authors first analyzed whether the models accurately capture seasonal sea ice volume and processes that add fresh water to the ocean at its surface, including melting ice, river runoff, and precipitation. They found these estimates to be accurate, suggesting that these aspects of the models are not to blame.

However, the researchers showed, the models deviate significantly from observations in capturing how fresh surface waters mix with deeper ocean waters. Specifically, the models overestimate the depth of mixing, and this unrealistically deep mixing range ultimately contributes to the models’ underestimations of surface salinity.

According to the authors, the findings of this study can inform refinements of the two analyzed models, as well as other climate models, so they can more accurately predict future declines in Arctic surface salinity and the resulting effects on sea ice, ecosystems, and the planet. (Geophysical Research Letters,, 2021)

—Sarah Stanley, Science Writer

Citation: Stanley, S. (2021), Capturing how fast the Arctic Ocean is gaining fresh water, Eos, 102, Published on 8 December 2021.
Text © 2021. AGU. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.