The difficulty of predicting an ice-free Arctic

After examining both high- and medium-level carbon dioxide emission modeling scenarios for the rest of the 21st century, the study found that it is not possible to reduce the uncertainty window for an ice-free Arctic to a period of less than 21 years due to the inherently chaotic nature of the climate.

The study also found that commonly-used metrics of past and present sea ice thickness, extent and volume are not predictive enough to reduce this long-range uncertainty.

The new findings were published online today in the American Geophysical Union journal Geophysical Research Letters.

Scientists typically define an "ice-free Arctic" as having fewer than 1 million square kilometers worth of ice cover, which would leave the Arctic Ocean virtually clear while some pockets of ice would remain in the northern reaches of Canada and Greenland.

"When it comes to predicting the timing of an ice-free Arctic, climate models show a large spread of over 100 years. Many studies have attempted to narrow this wide range to as little as five years in some cases," said Alexandra Jahn, an assistant professor in CU-Boulder's Department of Atmospheric and Oceanic Sciences (ATOC) and lead author of the new research. "Here, however, we find that the low bound of our predictive ability is significantly longer due to inherent climate variability."

The study, which employed a large collection of simulations from the Community Earth System Model developed by the University Corporation for Atmospheric Research (UCAR) in Boulder, Colorado, also found that consecutive ice-free summers would become common after 2060 under the high emission scenario while remaining the exception in the medium emission scenario.

"Overall, these results serve as a sort of caution against over-narrowing the long-term sea ice predictions from climate models" said Jahn, who is also a fellow in CU Boulder's Institute for Arctic and Alpine Research (INSTAAR).