A comparison of South Pacific Antarctic sea ice and atmospheric circulation reconstructions since 1900

Abstract

The recent changes and record lows in Antarctic sea ice extent illustrate the need for longer estimates beyond the short satellite observation period commencing around 1979. However, Antarctic sea ice extent reconstructions since 1900 based on paleo-records and those generated based on instrumental observations from the Southern Hemisphere midlatitudes are markedly different, especially prior to 1979. Here, these reconstructions are examined with the goal of understanding the relative strengths and limitations of each reconstruction better so that researchers using the various datasets can interpret them appropriately. Overall, it is found that the different spatial and temporal resolutions of each dataset play a secondary role to the inherent connections each reconstruction has with its implied atmospheric circulation. Five Southern Hemisphere pressure reconstructions spanning the 20th century are thus examined further. There are different variabilities and trends poleward of 60∘ S between proxy-based and station-based 20th century pressure reconstructions, which are connected to the disagreement between the Antarctic sea ice extent reconstructions examined here. Importantly, reconstructions based on only coral records provide the best agreement between the early pressure reconstructions, suggesting that a contributing role of tropical variability is present in the station-based pressure (and therefore sea ice) reconstructions. In contrast, ice-core-only reconstructions provide a local, high-latitude constraint that creates differences between the proxy-based and station-based reconstructions near Antarctica. Our results reveal the greatest consistencies and inconsistencies in available datasets and highlight the need to better understand the relative roles of the tropics versus high latitudes in historical sea ice variability around Antarctica.