The authors requested a talk

Signal of Persistent Multidecadal Variability in Wintertime Sea-Ice Records: Linkages to the Atlantic Multidecadal Oscillation

Miles, Martin 1 ; et, al. 2

1 Uni Research / Bjerknes Centre
2 Several institutions

Recent satellite observations suggest an arctic sea ice–climate system in rapid transformation, yet its long-term modes of variability is poorly known. Here, we integrate and synthesize an extensive set of multicentury historical records of arctic–subarctic sea ice, supplemented with high-resolution paleo proxy sea-ice records. To identify patterns of multidecadal variability of sea ice, we evaluated a number of long historical and paleo proxy data and selected altogether seven historical sea-ice records spanning the subarctic–arctic Atlantic, from the Labrador Sea to the Barents Sea (Fig. 1), supplemented with two high-resolution paleo proxy sea-ice records: a terrestrial record calibrated for Greenland Sea ice extent and a marine record from north of Iceland. These paleo records are co-located with two of the historical records, thereby providing independent evidence of any signals in the historical records. In contrast to the recent multiproxy reconstruction of summer sea ice (Kinnard et al., 2011), most of the records here reflect conditions in the cold season (winter–spring), when decadal-to-multidecadal climate variability (e.g., the Early Twentieth Century Warming (ETCW)) is most pronounced in the Arctic.

The most salient feature to emerge from the time-series analysis is the presence of pervasive multidecadal variability, upon which interannual-to-decadal fluctuations are superposed. Four key aspects should be noted: (i) Less sea ice is generally seen in the 20th century; however, the changes are far from the monotonic recent negative trend indicated from hemispheric datasets - a common characteristic amongst several of the records is sharply reduced sea ice at the onset of the ETCW, which heralded the termination of the Little Ice Age in the region. The most reduced sea ice before the 20th century occurred in the late 16th and mid-to-late 17th century, as seen in both the Icelandic historical sea-ice and Western Nordic Seas proxy records. (ii) Multidecadal variability is apparent in all of the records (except for Baltic Sea, which is predominated by interannual variability). The wavelet-filtered signals have predominately 60–90 year time scales, which are most pronounced in the Greenland Sea. The multidecadal fluctuations amongst the records are essentially consistent in their periods and in phase (except for the Barents Sea). (iii) The multidecadal signal are persistent in all records where it is found – in no cases does the signal disappear or dissipate through time, although being only quasi-periodic rather than deterministically periodic, the signal naturally varies in amplitude and periodicity. The persistence of the signal through the centuries and longer strongly suggests that the multidecadal oscillation is a robust feature and thus must have an underlying physical mechanism. (iv) Multidecadal signals are strongest in the Greenland Sea region and weaker on either side, i.e., Newfoundland record and the Eastern Nordic Seas (Barents Sea). Further, this is consistent with model simulations of multidecadal climate-system variability that suggest the Greenland Sea to be a key region.

Covariability between sea ice and the Atlantic Multidecadal Oscillation (AMO) is clearly evident during the instrumental record, including an abrupt change during the early 20th century warming (ETCW). Similar behaviour through previous centuries is evident from comparison of longer historical records and paleo proxy reconstructions of sea ice and the AMO, demonstrating that arctic sea ice is a dynamic climate-system component intrinsically and robustly linked to Atlantic multidecadal variability. Further, we extend the analysis to include a cross-comparison with some newly published climate and sea ice reconstructions: the 1,450 year summer sea ice record (KInnard et al., 2011), a 1000-year series of winter surface air temperature variations in Svalbard and northern Norway reconstructed from ice-core data (Divine et al., 2011), and new sea-ice proxy records from Greenland waters.

Divine, D., et al., 2011, Thousand years of winter surface air temperature variations in Svalbard and northern Norway reconstructed from ice core data, Polar Research, 30, 7379, DOI: 10.3402/polar.v30i0.7379.

Kinnard, C., et al., 2011, Reconstructed changes in Arctic sea ice over the past 1,450 years. Nature, 479, 509–512, doi: 10.1038/nature10581.


Fig 1.

Approximate locations of the historical and paleo proxy wintertime sea-ice time series used in this study. White dots signify the seven historical records, and brown dots signify the two paleo proxy records.. Bathymetry and topography are shown in relief. Selected arctic marginal seas are indicated.


Fig 2.

Linking multidecadal fluctuations in sea ice to North Atlantic SSTs. Original non-smoothed time series (gray) and multidecadal 50–120 year component (blue) reconstructed from wavelet decomposition: (A) AMO index, not de-trended, i.e., North Atlantic SST anomaly. (B) AMO proxy index, not de-trended, 10-year running average. (C) Fram Strait ice export (km^3) reconstructed from historical arctic ice extent along SW Greenland. (D) Icelandic sea-ice severity index (1600–1870) and sea-ice incidence index (1880–2000) (sigma units). (E) Western Nordic Seas sea-ice extent (10^3 km^2) proxy reconstruction. The squares in D indicate sea-ice conditions from a marine core sea-ice proxy off North Iceland, with anomalies larger than ±1 s.d. in blue (positive) and red (negative). The color bar in (E) indicates periods of relatively less (red) and more (blue) ice, inferred from the multidecadal wavelet filter; the less-icy warm periods are projected onto the other sea-ice and AMO series (light red shading).



Questions? Check our FAQ
or email us at

Supported by
The US National Science Foundation
OPP- Arctic Natural Sciences Program

Site built and
hosted by