INTERANNUAL VARIABILITY OF SEA-ICE COVER IN THE ARCTIC AND SUBPOLAR REGIONS: RELATED TO NAO and ENSO?
Jia Wang
FRPGC and IARC, University of Alaska Fairbanks, Fairbanks, Alaska
1. INTRODUCTION
The nature of the sea-ice cover in the Arctic and subpolar regions associated with NAO (North Atlantic Oscillation) and ENSO (El Nino-Southern Oscillation) episodes is investigated using historical data from 1900 to 1993. A principal component (or empirical orthogonal functions, EOFs) analysis was applied to SLP, SAT, and sea-ice areas in the high latitudes for the period of 1953-1993. The results indicate that the first EOF mode is associated with NAO, while the third mode with ENSO. The statistical significance test was also conducted to identify the regions of over 95% significance level. Spectral analysis and Monte Cairo simulation were also conducted to the analyzed data, indicating 5-7 year and decadal variabilities.
In section 2 the data analyzed are briefly described. Section 3 presents an empirical orthogonal (EOF) analysis of SLP, SAT fields and sea-ice areas whose two modes explain very well the NAO-induced dipolar (or seesaw) structure in both SAT and sea-ice anomalies and ENSO-ιnduced pattern. In section 4 conclusions and suggestions for fuτure work are given.
2. The DATA
The monthly sea-ice concentration (SIC) data analyzed are similar to those used in Wang et al. (1994, 1995) and Mysak et al. (1996) and have been clearly described by Walsh and Johnson (1979). This dataset has been updated to 1990. The data was available at the time this research was carried out extend over the period 1900-1990 and cover the Arctic Ocean and the adjacent seas with a 1° by 1° latitude-latitude grid. The monthly sea-level pressure (SLP) and surface air temperature (SAT) data used were
directly from the National Center for Atmospheric Research (NCAR), archived on 5° by 5°global grids. These two monthly data sets extend over the periods of 1850-1993 for SAT and 1899-1993 for SLP. The data during the period of 1900-1993 were used in this study.
3. RESULTS: EOF ANALYSIS
The Arctic seaice coverage is divided into 8 subregions: 1) the Bering Sea, 2) the Beaufort and Chukchi Seas, 3) Hudson Bay,4) Baffin Bay, David Strait and the Labrador Sea, 5) the Greenland and Norwegian Seas, 6) the Barents and Kara Seas, 7) the East Siberian and Laptev Seas, and the Sea of Okhotsk (Wang et al. 1995). The ice-covered area was calculated for each area. The time series of the eight regions were input to the EOF analysis package developed by Wang et al. (1995) with the means removed.
Eight sea-ice modes were derived. The first three modes contain 21%, 19.5%, and 17.2% of total variance, respectively. The first and third modes are physically plausible, while the second mode is not well understood. The first EOF mode of the sea-ice areas in the Arctic explains 21% of the total variance. The mode can explain the NAO (seesaw) impact, because the sea-ice anomalies in Hudson Bay (subregion 3), Baffin Bay, David Strait and the Labrador Sea (subregion 4) were out of phase with the anomalies in the Greenland and Norwegian Seas (subregion 5) and Barents and Kara Seas (subregion 6) (Fig. 1, upper panel). This first sea-ice EOF mode matches very well the first SAT EOF mode. Therefore, the NAO-induced
. Corresponding author address: Jia Wang, The Frontier Research Program for Global Change (FRPGC) and International Arctic Research Center (IARC), University of Alaska Fairbanks (UAF), P.O. Box 757220, Fairbanks, AK 99775 USA; email: jwang@ims.uaf.edu
