Century-scale perspectives on observed and simulated Southern Ocean sea ice trends from proxy reconstructions

• Published in: Journal of Geophysical Research: Oceans
• Main Authors: Hobbs, Will, Curran, Mark, Abram, Nerilie, Thomas, Elizabeth R.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2016
• Subjects: Antarctic; Southern Ocean; Weddell Sea; Ross Sea; Weddell; Antarc*; Sea ice
• Online Access: http://nora.nerc.ac.uk/id/eprint/514812/; https://nora.nerc.ac.uk/id/eprint/514812/1/Hobbs_et_al-2016-Journal_of_Geophysical_Research__Oceans.pdf; https://doi.org/10.1002/2016JC012111

Since 1979 when continuous satellite observations began, Southern Ocean sea ice cover has increased, whilst global coupled climate models simulate a decrease over the same period. It is uncertain whether the observed trends are anthropogenically forced or due to internal variability, or whether the apparent discrepancy between models and observations can be explained by internal variability. The shortness of the satellite record is one source of this uncertainty, and a possible solution is to use proxy reconstructions, which extend the analysis period but at the expense of higher observational uncertainty. In this work, we evaluate the utility for change detection of 20th century Southern Ocean sea ice proxies. We find that there are reliable proxies for the East Antarctic, Amundsen, Bellingshausen and Weddell sectors in late winter, and for the Weddell Sea in late autumn. Models and reconstructions agree that sea ice extent in the East Antarctic, Amundsen and Bellingshausen Seas has decreased since the early 1970s, consistent with an anthropogenic response. However, the decrease is small compared to internal variability, and the change is not robustly detectable. We also find that optimal fingerprinting filters out much of the uncertainty in proxy reconstructions. The Ross Sea is a confounding factor, with a significant increase in sea ice since 1979 that is not captured by climate models; however, existing proxy reconstructions of this region are not yet sufficiently reliable for formal change detection.