• Alexeev, V. A., V. V. Ivanov, R. Kwok, and L. H. Smedsrud, 2013: North Atlantic warming and declining volume of Arctic sea ice. Cryosphere Discuss., 7, 245265, https://doi.org/10.5194/tcd-7-245-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bintanja, R., R. G. Graversen, and W. Hazeleger, 2011: Arctic winter warming amplified by the thermal inversion and consequent low infrared cooling to space. Nat. Geosci., 4, 758761, https://doi.org/10.1038/ngeo1285.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, H. W., R. B. Alley, and F. Zhang, 2016: Interannual Arctic sea ice variability and associated winter weather patterns: A regional perspective for 1979-2014. J. Geophys. Res. Atmos., 121, 14 43314 455, https://doi.org/10.1002/2016JD024769.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davini, P., C. Cagnazzo, R. Neale, and J. Tribbia, 2012: Coupling between Greenland blocking and the North Atlantic Oscillation pattern. Geophys. Res. Lett., 39, L14701, https://doi.org/10.1029/2012GL052315.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, https://doi.org/10.1002/qj.828.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Deser, C., and H. Teng, 2008: Evolution of Arctic sea ice concentration trends and the role of atmospheric circulation forcing, 1979-2007. Geophys. Res. Lett., 35, L02504, https://doi.org/10.1029/2007GL032023.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Deser, C., J. E. Walsh, and M. S. Timlin, 2000: Arctic sea ice variability in the context of recent atmospheric circulation trends. J. Climate, 13, 617633, https://doi.org/10.1175/1520-0442(2000)013,0617:ASIVIT.2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ding, Q., J. M. Wallace, D. S. Battisti, E. J. Steig, A. J. E. Gallant, H.-J. Kim, and L. Geng, 2014: Tropical forcing of the recent rapid Arctic warming in northeastern Canada and Greenland. Nature, 509, 209212, https://doi.org/10.1038/nature13260.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ding, Q., and Coauthors, 2017: Influence of high-latitude atmospheric circulation changes on summertime Arctic sea ice. Nat. Climate Change, 7, 289295, https://doi.org/10.1038/nclimate3241.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Enfield, D. B., A. M. Mestas-Nuñez, and P. J. Trimble, 2001: The Atlantic multidecadal oscillation and its relationship to rainfall and river flows in the continental U.S. Geophys. Res. Lett., 28, 20772080, https://doi.org/10.1029/2000GL012745.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flanner, M. G., K. M. Shell, M. Barlage, D. K. Perovich, and M. A. Tschudi, 2011: Radiative forcing and albedo feedback from the Northern Hemisphere cryosphere between1979 and 2008. Nat. Geosci., 4, 151155, https://doi.org/10.1038/ngeo1062.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gagné, M.-È., N. P. Gillett, and J. C. Fyfe, 2015: Impact of aerosol emission controls on future arctic sea ice cover. Geophys. Res. Lett., 42, 84818488, https://doi.org/10.1002/2015GL065504.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gibson, P. B., S. E. Perkins-Kirkpatrick, P. Uotila, A. S. Pepler, and L. V. Alexander, 2017: On the use of self-organizing maps for studying climate extremes. J. Geophys. Res. Atmos., 122, 38913903, https://doi.org/10.1002/2016JD026256.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gong, T., S. Feldstein, and S. Lee, 2017: The role of downward infrared radiation in the recent Arctic winter warming trend. J. Climate, 30, 49374949, https://doi.org/10.1175/JCLI-D-16-0180.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hanna, E., and Coauthors, 2014: Atmospheric and oceanic climate forcing of the exceptional Greenland ice sheet surface melt in summer. Int. J. Climatol., 34, 10221037, https://doi.org/10.1002/joc.3743.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hoskins, B. J., and D. J. Karoly, 1981: The steady linear response of a spherical atmosphere to thermal and orographic forcing. J. Atmos. Sci., 38, 11791196, https://doi.org/10.1175/1520-0469(1981)038<1179:TSLROA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, Y., X. Dong, B. Xi, E. K. Dolinar, and R. E. Stanfield, 2017: The footprints of 16 year trends of Arctic springtime cloud and radiation properties on September sea ice retreat. J. Geophys. Res. Atmos., 122, 21792193, https://doi.org/10.1002/2016JD026020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kapsch, M.-L., R. G. Graversen, T. Economou, and M. Tjernström, 2014: The importance of spring atmospheric conditions for predictions of the Arctic summer sea ice extent. Geophys. Res. Lett., 41, 52885296, https://doi.org/10.1002/2014GL060826.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kapsch, M.-L., R. G. Graversen, M. Tjernström, and R. Bintanja, 2016: The effect of downwelling longwave and shortwave radiation on Arctic summer sea ice. J. Climate, 29, 11431159, https://doi.org/10.1175/JCLI-D-15-0238.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kim, B.-M., and Coauthors, 2017: Major cause of unprecedented Arctic warming in January 2016: Critical role of an Atlantic wind storm. Sci. Rep., 7, 40051, https://doi.org/10.1038/srep40051.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kohonen, T., 2001: Self-Organizing Maps. 3rd ed. Springer, 501 pp.

    • Crossref
    • Export Citation
  • Lee, S., and S. B. Feldstein, 2013: Detecting ozone- and greenhouse gas–driven wind trends with observational data. Science, 339, 563567, https://doi.org/10.1126/science.1225154.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, S., T. Gong, N. C. Johnson, S. B. Feldstein, and D. Pollard, 2011: On the possible link between tropical convection and the Northern Hemisphere Arctic surface air temperature change between 1958 and 2011. J. Climate, 24, 43504367, https://doi.org/10.1175/2011JCLI4003.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liu, Z., and A. Schweige, 2017: Synoptic conditions, clouds, and sea ice melt onset in the Beaufort Chukchi seasonal ice zone. J. Climate, 30, 69997016, https://doi.org/10.1175/JCLI-D-16-0887.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lynch, A. H., M. C. Serreze, E. N. Cassano, A. D. Crawford, and J. Stroeve, 2016: Linkages between Arctic summer circulation regimes and regional sea ice anomalies. J. Geophys. Res. Atmos., 121, 78687880, https://doi.org/10.1002/2016JD025164.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahajan, S., R. Zhang, and T. L. Delworth, 2011: Impact of the Atlantic meridional overturning circulation (AMOC) on Arctic surface air temperature and sea-ice variability. J. Climate, 24, 65736581, https://doi.org/10.1175/2011JCLI4002.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78, 10691079, https://doi.org/10.1175/1520-0477(1997)078<1069:APICOW>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mills, C. M., and J. E. Walsh, 2014: Synoptic activity associated with sea ice variability in the Arctic. J. Geophys. Res. Atmos., 119, 12 11712 131, https://doi.org/10.1002/2014JD021604.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Min, S.-K., X. Zhang, F. W. Zwiers, and T. Agnew, 2008: Human influence on Arctic sea ice detectable from early 1990s onwards. Geophys. Res. Lett., 35, 213236, https://doi.org/10.1029/2008GL035725.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mortin, J., G. Svensson, R. G. Graversen, M.-L. Kapsch, J. C. Stroeve, and L. N. Boisvert, 2016: Melt onset over Arctic sea ice controlled by atmospheric moisture transport. Geophys. Res. Lett., 43, 66366642, https://doi.org/10.1002/2016GL069330.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Notz, D., and J. Marotzke, 2012: Observations reveal external driver for Arctic sea-ice retreat. Geophys. Res. Lett., 39, L08502, https://doi.org/10.1029/2012GL051094.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ogi, M., K. Yamazaki, and J. M. Wallace, 2010: Influence of winter and summer surface wind anomalies on summer Arctic sea ice extent. Geophys. Res. Lett., 37, L07701, https://doi.org/10.1029/2009GL042356.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Park, W., and M. Latif, 2008: Multidecadal and multicentennial variability of the meridional overturning circulation. Geophys. Res. Lett., 35, L22703, https://doi.org/10.1029/2008GL035779.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pithan, F., and T. Mauritsen, 2014: Arctic amplification dominated by temperature feedbacks in contemporary climate models. Nat. Geosci., 7, 181184, https://doi.org/10.1038/ngeo2071.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Polyakov, I. V., and Coauthors, 2010: Arctic Ocean warming contributes to reduced polar ice cap. J. Phys. Oceanogr., 40, 27432756, https://doi.org/10.1175/2010JPO4339.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, https://doi.org/10.1029/2002JD002670.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reusch, D. B., R. B. Alley, and B. C. Hewitson, 2005: Relative performance of self-organizing maps and principal component analysis in pattern extraction from synthetic climatological data. Polar Geogr., 29, 188212, https://doi.org/10.1080/789610199.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rigor, I., J. M. Wallace, and R. L. Colony, 2002: Response of sea ice to the Arctic Oscillation. J. Climate, 15, 26482663, https://doi.org/10.1175/1520-0442(2002)015<2648:ROSITT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sato, K., J. Inoue, and M. Watanabe, 2014: Influence of the Gulf Stream on the Barents Sea ice retreat and Eurasian coldness during early winter. Environ. Res. Lett., 9, 084009, https://doi.org/10.1088/1748-9326/9/8/084009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schudeboom, A., A. J. McDonald, O. Morgenstern, M. Harvey, and S. Parsons, 2018: Regional regime-based evaluation of present-day GCM cloud simulations using self-organizing maps. J. Geophys. Res. Atmos., 123, 42594272, https://doi.org/10.1002/2017JD028196.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sedlar, J., and Coauthors, 2011: A transitioning Arctic surface energy budget: The impacts of solar zenith angle, surface albedo and cloud radiative forcing. Climate Dyn., 37, 16431660, https://doi.org/10.1007/s00382-010-0937-5.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J., J. Zhang, E. Watanabe, M. Ikeda, K. Mizobata, J. E. Walsh, X. Bai, and B. Wu, 2009: Is the dipole anomaly a major driver to record lows in Arctic summer sea ice extent? Geophys. Res. Lett., 36, L05706, https://doi.org/10.1029/2008GL036706.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Woodgate, R. A., T. J. Weingartner, and R. Lindsay, 2012: Observed increases in Bering Strait oceanic fluxes from the Pacific to the Arctic from 2001 to 2011 and their impacts on the Arctic Ocean water column. Geophys. Res. Lett., 39, L24603, https://doi.org/10.1029/2012GL054092.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yu, L., and S. Zhong, 2018: Changes in sea-surface temperature and atmospheric circulation patterns associated with reductions in Arctic sea ice cover in recent decades. Atmos. Chem. Phys., 18, 14 14914 159, https://doi.org/10.5194/acp-18-14149-2018.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yu, L., S. Zhong, J. A. Winkler, M. Zhou, D. H. Lenschow, B. Li, X. Wang, and Q. Yang, 2017: Possible connection of the opposite trends in Arctic and Antarctic sea ice cover. Sci. Rep., 7, 45804, https://doi.org/10.1038/srep45804.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, J., D. A. Rothrock, and M. Steele, 1998: Warming of the Arctic Ocean by a strengthened Atlantic inflow: Model results. Geophys. Res. Lett., 25, 17451748, https://doi.org/10.1029/98GL01299; Corrigendum, 25, 3541, https://doi.org/10.1029/98GL02634.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, R., 2015: Mechanisms for low-frequency variability of summer Arctic sea ice extent. Proc. Natl. Acad. Sci. USA, 112, 45704575, https://doi.org/10.1073/pnas.1422296112.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    The SOM patterns of the anomalous melt-season (June–August) daily sea ice concentration on a 3 × 3 grid for the 1979–2016 period. The percentages at the top left of each panel indicates the frequency of occurrences of the pattern.

  • View in gallery

    Time series of occurrence number for each SOM pattern in Fig. 1.

  • View in gallery

    The spatial patterns of the (a) first and (b) second modes of the EOF analysis of the anomalous melt-season sea ice concentration.

  • View in gallery

    (a) Total, (b) SOM-explained, and (c) residual JJA sea ice concentration trends (decade−1). (d) The ratio of SOM-explained to total sea ice concentration trends. The blue–red colorbar is used in (a)–(c), and the jet colorbar is used in (d). Dots in (a) indicate above 95% confidence level.

  • View in gallery

    Trends in sea ice concentration explained by each SOM pattern (decade−1). The fraction of the total sea ice trend represented by each node is shown in left corner of each panel.

  • View in gallery

    Composite of anomalous 500-hPa geopotential height (gpm) for each node. Dotted regions denote above 95% confidence level.

  • View in gallery

    As in Fig. 6, but for the anomalous 850-hPa wind field (m s−1). Shaded regions denote above 95% confidence level.

  • View in gallery

    As in Fig. 6, but for the anomalous surface downward longwave radiation (105 W m−2). Dotted regions denote above 95% confidence level.

  • View in gallery

    As in Fig. 6, but for the anomalous low-level temperature (1000–750 hPa; °C). Dotted regions denote above 95% confidence level.

  • View in gallery

    As in Fig. 6, but for anomalous low-level specific humidity (1000–750 hPa; g kg−1). Dotted regions denote above 95% confidence level.

  • View in gallery

    Anomalous SST (°C) regressed into the time series of occurrence number for nodes 1, 3, 7, and 9.

  • View in gallery

    As in Fig. 11, but for the anomalous 500-hPa geopotential height (gpm).

  • View in gallery

    As in Fig. 11, but for the anomalous 850-hPa wind field (m s−1).

  • View in gallery

    As in Fig. 11, but for the anomalous surface downward longwave radiation (105 W m−2). Dotted regions denote above 95% confidence level.

  • View in gallery

    As in Fig. 11, but for the anomalous low-level temperature (1000–750 hPa; °C). Dotted regions denote above 95% confidence level.

  • View in gallery

    As in Fig. 11, but for the anomalous low-level specific humidity (1000–750 hPa; 10−1 g kg−1). Dotted regions denote above 95% confidence level.

  • View in gallery

    Time series of the normalized summertime (a) PDO and (b) AMO indices. Blue lines denote the 20-yr low-pass filtered time series.

  • View in gallery

    The probability distribution of regression coefficients from the regressions of the anomalous daily 500-hPa height fields onto composited anomalous 500-hPa height fields for nodes 1, 3, 7, and 9 over the periods of 1979–98 and 1999–2016. The difference of probability density over two periods is statistically significant at the above 95% confidence level.

  • View in gallery

    The (a) leading pattern and (b) its time series of EOF analysis of melt-season Arctic sea ice concentration anomalies. Prior to applying EOF, the sea ice concentration data are 20-yr low-pass filtered and detrended. The PDO and AMO are 30-yr low-pass filtered and detrended.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 46 46 16
PDF Downloads 39 39 15

Revisiting the Linkages between the Variability of Atmospheric Circulations and Arctic Melt-Season Sea Ice Cover at Multiple Time Scales

View More View Less
  • 1 SOA Key Laboratory for Polar Science, Polar Research Institute of China, Shanghai, China
  • 2 Department of Geography, Environment and Spatial Sciences, Michigan State University, East Lansing, Michigan
  • 3 SOA Key Laboratory for Polar Science, Polar Research Institute of China, Shanghai, and National Marine Environment Prediction Center, Beijing, China
  • 4 National Center for Atmospheric Research, Boulder, Colorado
  • 5 SOA Key Laboratory for Polar Science, Polar Research Institute of China, Shanghai, China
© Get Permissions
Full access

Abstract

The sharp decline of Arctic sea ice in recent decades has captured the attention of the climate science community. A majority of climate analyses performed to date have used monthly or seasonal data. Here, however, we analyze daily sea ice data for 1979–2016 using the self-organizing map (SOM) method to further examine and quantify the contributions of atmospheric circulation changes to the melt-season Arctic sea ice variability. Our results reveal two main variability modes: the Pacific sector mode and the Barents and Kara Seas mode, which together explain about two-thirds of the melt-season Arctic sea ice variability and more than 40% of its trend for the study period. The change in the frequencies of the two modes appears to be associated with the phase shift of the Pacific decadal oscillation (PDO) and the Atlantic multidecadal oscillation (AMO). The PDO and AMO trigger anomalous atmospheric circulations, in particular, the Greenland high and the North Atlantic Oscillation and anomalous warm and cold air advections into the Arctic Ocean. The changes in surface air temperature, lower-atmosphere moisture, and downwelling longwave radiation associated with the advection are consistent with the melt-season sea ice anomalies observed in various regions of the Arctic Ocean. These results help better understand the predictability of Arctic sea ice on multiple (synoptic, intraseasonal, and interannual) time scales.

© 2019 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Dr. Lejiang Yu, yulejiang@sina.com.cn

Abstract

The sharp decline of Arctic sea ice in recent decades has captured the attention of the climate science community. A majority of climate analyses performed to date have used monthly or seasonal data. Here, however, we analyze daily sea ice data for 1979–2016 using the self-organizing map (SOM) method to further examine and quantify the contributions of atmospheric circulation changes to the melt-season Arctic sea ice variability. Our results reveal two main variability modes: the Pacific sector mode and the Barents and Kara Seas mode, which together explain about two-thirds of the melt-season Arctic sea ice variability and more than 40% of its trend for the study period. The change in the frequencies of the two modes appears to be associated with the phase shift of the Pacific decadal oscillation (PDO) and the Atlantic multidecadal oscillation (AMO). The PDO and AMO trigger anomalous atmospheric circulations, in particular, the Greenland high and the North Atlantic Oscillation and anomalous warm and cold air advections into the Arctic Ocean. The changes in surface air temperature, lower-atmosphere moisture, and downwelling longwave radiation associated with the advection are consistent with the melt-season sea ice anomalies observed in various regions of the Arctic Ocean. These results help better understand the predictability of Arctic sea ice on multiple (synoptic, intraseasonal, and interannual) time scales.

© 2019 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Dr. Lejiang Yu, yulejiang@sina.com.cn

1. Introduction

A sharp decline in Arctic sea ice has been observed in recent decades, and various theories based on statistical or numerical analyses have been proposed to explain the downward trend. These theories include contributions from local processes, such as ice-albedo feedback (Flanner et al. 2011) and water vapor and cloud radiative feedback (Sedlar et al. 2011), development of surface-based temperature inversions (Bintanja et al. 2011), and changes in the atmospheric lapse rate (Pithan and Mauritsen 2014), as well as large-scale processes, such as temperature and moisture advection associated with anomalous atmospheric circulations.

As a fundamental forcing, increasing greenhouse gas emissions and the associated warming could contribute to the observed Arctic sea ice loss. Several studies have shown that the increase in the atmospheric CO2 concentrations (Min et al. 2008; Notz and Marotzke 2012) and the overall decrease in aerosol emissions (Gagné et al. 2015) have contributed to the decrease in the Arctic sea ice extent.

Besides the increased greenhouse gas emissions, natural variability of the climate system may have also contributed to the Arctic sea ice depletion. The downward trend of the Arctic sea ice prior to the 1990s has been linked to a positive trend in the North Atlantic Oscillation (NAO) index (Deser et al. 2000), and this linkage is expected to extend into the first decade of the twenty-first century (Ogi et al. 2010). The thinning Arctic sea ice has also been attributed to a positive trend in the Arctic Oscillation (AO) index (Rigor et al. 2002) and the trend in the Arctic dipole (AD) anomaly index toward higher values (Wang et al. 2009). In addition, the multidecadal variability of sea surface temperature (SST), as characterized by the Pacific decadal oscillation (PDO) in the North Pacific Ocean and the Atlantic multidecadal oscillation (AMO) in the North Atlantic Ocean, has also been found to influence the Arctic sea ice (Ding et al. 2014; Park and Latif 2008; Yu et al. 2017), but these multidecadal-scale variability modes do not explain the recent trend in the Arctic sea ice (Deser and Teng 2008).

The Arctic sea ice loss has also been linked to heat transport associated with North Atlantic warming (Zhang et al. 1998; Polyakov et al. 2010; Alexeev et al. 2013), the Atlantic meridional overturning circulation (AMOC; Mahajan et al. 2011), and the increasing oceanic fluxes into the Arctic through the Bering Strait (Woodgate et al. 2012). Zhang (2015) indicates that oceanic heat transport in North Atlantic and Pacific Ocean and AD contributes substantially to the September Arctic sea ice decline.

Apart from the aforementioned changes in SST and large-scale atmospheric circulations, local factors can also influence Arctic sea ice through thermodynamic mechanisms. Several recent studies have noted that cloud and longwave radiation strongly influence sea ice concentration (Kapsch et al. 2014; Mortin et al. 2016; Huang et al. 2017), and their effects can change significantly with season (Kapsch et al. 2016). A recent study (Liu and Schweiger 2017) has related the sea ice melt over the Beaufort and Chukchi Seas to fewer low-level clouds and high precipitable water associated with changes synoptic conditions in the Arctic.

To date, the majority of climate analysis studies of Arctic sea ice have used monthly or seasonal sea ice data, which limits the ability of the analyses to adequately capture the synoptic intraseasonal-scale Arctic sea ice variability. In this study, we utilize daily sea ice data for a period of 38 years (1979–2016) to further examine and quantify the contributions of atmospheric circulation changes to the Arctic sea ice variability at multiple time scales.

Most previous climatological studies of Arctic sea ice variability have used monthly or seasonal mean data that fail to resolve variability shorter than seasonal scale. Lee and Feldstein (2013) showed that the interdecadal Southern Hemisphere poleward jet shift is realized through changes in the frequency of poleward jet shift events with time scale of less than 10 days. Lee et al. (2011) also noted that the surface air temperature trend from 1958 to 2001 is realized through changes in the frequency of occurrences of atmospheric circulation patterns with time scales of less than 10 days. The primary time scales of Arctic sea ice variability are synoptic and intraseasonal scales, and it is necessary to resolve these time scales in order to fully capture the Arctic sea ice variability. In this study, we utilize daily sea ice data for a period of 38 years (1979–2016) to further examine and quantify the contributions of atmospheric circulation changes to the Arctic sea ice variability. The use of daily sea ice data enables the capturing of sea ice variability at all important time scales from synoptic to intraseasonal to annual and beyond.

2. Datasets and methods

We first apply the self-organizing map (SOM) technique (Kohonen 2001) to daily sea ice concentration data to determine the variability and trends of melt-season (June, July, and August) Arctic sea ice for the period 1979–2016. We then apply the regression and composite analysis techniques to explain the sea ice variability in the context of anomalous atmospheric circulations.

The analyses utilize three types of gridded data, including daily sea ice data from the U.S. National Snow and Ice Data Center (NSIDC; ftp://sidads.colorado.edu/DATASETS/nsidc0051_gsfc_nasateam_seaice/final-gsfc/north/daily), monthly SST data from the Met Office Hadley Centre (http://www.metoffice.gov.uk/hadobs/hadisst; Rayner et al. 2003), and daily atmospheric data from ERA-Interim (https://apps.ecmwf.int/datasets/data/interim-full-daily/levtype=sfc/; Dee et al. 2011). The sea ice data are on a polar stereographic grid at 25 km × 25 km resolution, while the other datasets are on a latitude and longitude grid at 1° resolution for SST and 1.5° resolution for atmospheric variables. We also download the monthly PDO (Mantua et al. 1997) and AMO (Enfield et al. 2001) indices (from http://ds.data.jma.go.jp/tcc/tcc/products/elnino/decadal/pdo.html and https://www.esrl.noaa.gov/psd/data/correlation/amon.us.data, respectively).

The SOM technique, which is a type of artificial neural network, utilizes unsupervised learning to detect patterns in data and is used to reduce multidimensional data into a two-dimensional array. Each node in the array has a reference vector that displays a spatial pattern of the input data. All patterns in the two-dimensional array represent the continuous variability of the input data. The distance between any two nodes, which is measured by root-mean-squared differences of all spatial grid points in the two nodes, represents effectively the level of similarity between their patterns. Unlike other commonly used pattern-recognition and dimension-reduction methods [e.g., the empirical orthogonal function (EOF) method], the SOM technique is not limited by stringent requirements such as orthogonality of two spatial patterns and a priori assumption on data distribution (Reusch et al. 2005). For detailed description of the SOM algorithm, refer to Kohonen (2001).

In this study, the SOM technique is used to identify the spatial patterns of anomalous daily sea ice concentration north of 50°N in the boreal melt season (June–August). SOM analysis has been utilized in previous Arctic sea ice studies. For example, Mills and Walsh (2014) identified synoptic patterns, which are closely related to the Arctic sea ice variability. Lynch et al. (2016) analyzed separately Arctic circulation regimes associated with anomalous September sea ice states in the Atlantic and Pacific sectors of the Arctic. Using the SOM method, Chen et al. (2016) grouped the Arctic into different subregions and examined the interannual variability of sea ice concentration at these subregions.

Prior to applying SOM, daily anomalies are calculated at each grid point by subtracting the overall average of daily data for the entire study period 1979–2016 (i.e., the 38-yr climatology) from the original daily data. It is also necessary to determine the number of SOM nodes to be used in the analyses. The general rule for choosing the proper number of SOM nodes in an analysis is that the number should be large enough to adequately capture the variability of the variable analyzed, but also small enough to avoid overly similar patterns that tend to distract the analysis with unnecessary details (Lee and Feldstein 2013; Gibson et al. 2017; Schudeboom et al. 2018). Following this rule, we tested different numbers of SOM nodes. For a specific number, we assign the Arctic sea ice anomaly pattern for each melt-season day to the best matching SOM pattern on the basis of minimum Euclidean distance, and calculate spatial correlation between the daily Arctic sea ice anomaly pattern and the best matching SOM pattern. Table 1 shows the averaged correlation for all daily data across all SOM nodes. There is an increase in the spatial correlation from 0.29 to 0.39 as the number of SOM nodes increases from 3 (3 × 1 grid) to 16 (4 × 4 grid). The gain in the correlation is relatively small from 9 (3 × 3 grid) to 16 (4 × 4 grid) and, thus, a 3 × 3 grid seems to satisfy the aforementioned rule in grid selection and will be used in the analyses.

Table 1.

Spatial correlations between the daily melt-season Arctic sea ice concentration and the corresponding SOM pattern for each day from 1979 to 2016.

Table 1.

The contribution of each SOM pattern to the spatial pattern of Arctic sea ice trends at each grid point is calculated by the product of each SOM pattern and its frequency trend normalized by the total number of melt-season days (92; Lee and Feldstein 2013). The sum of these contributions represents the trends in the Arctic sea ice explained by the SOM patterns. Residual trends are calculated by subtracting SOM-explained trends from the total trends. The significance of the trends in the time series of the frequency of occurrences for each SOM pattern is calculated using the two-tailed Student’s t test.

3. Results

a. Spatial patterns and trends

The melt-season daily Arctic sea ice concentration anomalies are depicted by the nine SOM nodes and the spatial patterns for all nine nodes are shown in Fig. 1. Note that there are no sea ice data close to the pole. The anomalies are predominantly negative for nodes 1, 4, and 7 and positive for nodes 3, 6, and 9. Nodes 2, 5, and 8 display a dipole structure with opposite variability between the area of 150°E–120°W and the rest of the Arctic, indicating a transition from a predominantly negative state for nodes 1, 4, and 7 to a nearly positive state for nodes 3, 6, and 9. The nine nodes occur at different frequencies (the number on the top-left corner of each panel). The highest frequencies are for the node 3 (20%)–7 (19.4%) pair, followed by the node 1 (12.6%)–9 (14.6%) pair, with the rest of the nodes occurring less than 10% of the time. The positive anomaly pattern depicted by node 3 happened exclusively prior to 2000, while the negative anomaly pattern represented by node 7 appeared after that with the exception of 1995 (Fig. 2), indicating a transition from above-normal to below-normal sea ice concentration in Arctic seas especially in the Barents Sea and the Kara Sea, where the maximum anomalies occur. Similarly, the period of occurrences is also clearly separated for the next most frequent pair, with node 9 happening before 2000 and node 1 appearing afterward except for 1993 (Fig. 2), revealing a transition from above- to below-normal sea ice concentration for the Beaufort and the East Siberian Seas. The spatial pattern and temporal distribution of nodes 6 and 4 also reflect the transition from above-normal to below-normal sea ice cover in these Arctic seas. Derived from the temporal distribution, there are significantly positive temporal trends in the frequencies of occurrences for nodes 1, 4, and 7 and significantly negative trends for nodes 3, 6, and 9 (Table 2). However, there is no separation and no trend in the temporal distribution of the occurrences for the nodes with a dipole pattern (nodes 2 and 8).

Fig. 1.
Fig. 1.

The SOM patterns of the anomalous melt-season (June–August) daily sea ice concentration on a 3 × 3 grid for the 1979–2016 period. The percentages at the top left of each panel indicates the frequency of occurrences of the pattern.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 2.
Fig. 2.

Time series of occurrence number for each SOM pattern in Fig. 1.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Table 2.

Trends in the frequency of occurrences for each SOM node (day yr−1). Asterisks indicate the above 95% confidence level based on two-tailed Student’s t test.

Table 2.

EOF analyses of the Arctic sea ice concentration anomalies are also performed, and the results are compared to those of SOM analyses. The spatial patterns of the first two EOF modes (EOF1 and EOF2) are shown in Fig. 3. EOF1, which explains 18.4% of the total variance of melt-season Arctic sea ice concentration anomalies, shows all positive values except for the northeastern coast of Greenland. The EOF1 pattern closely resembles the SOM node 6 pattern (spatial correlation of 0.96) as well as the sum of the SOM nodes 3 and 9 patterns (spatial correlation of 0.98). EOF2, which accounts for 7.3% of the total variance, displays a dipole pattern that closely resembles the SOM node 2 pattern (spatial correlation of 0.95). The above comparison indicates that the leading EOF patterns are included in the SOM nodes, but the SOM nodes allow for more continuous variations and capture more details of the spatial and temporal variability.

Fig. 3.
Fig. 3.

The spatial patterns of the (a) first and (b) second modes of the EOF analysis of the anomalous melt-season sea ice concentration.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Trend analyses of the melt-season Arctic sea ice concentrations show downward trends everywhere in the region except for a small spot at the northeastern coast of Greenland (Fig. 4a). The strongest trends are found in the Kara Sea, with the maximum rate of −0.0197 yr−1, as well as the Beaufort and East Siberian Seas. The contributions of each SOM node to the total trend are calculated and vary considerably among the nodes (Fig. 5). The domain-averaged ratio of SOM-explained trend by each node to the total trend is shown in the top-left corner of each panel in Fig. 5. The node 3 (19.8%)–7 (18.8%) pair makes the largest domain-averaged contributions, followed by the node 1 (9.1%)–9 (8.1%) pair, and the rest of the nodes together explain less than 5% of total trend. A summation of the contributions from all nine nodes at each grid point result in a domain-averaged percentage of 41.5% explained by all SOM nodes, which is smaller than the summation of the domain-averaged percentage by each node (63.2%) because of the cancelation at some locations by opposite trends from different nodes. Figure 4d shows the percentage at grid points with significant trends. Larger ratios occur over Kara and Laptev Seas, while small values occur over Sea of Okhotsk and the coast of the Greenland.

Fig. 4.
Fig. 4.

(a) Total, (b) SOM-explained, and (c) residual JJA sea ice concentration trends (decade−1). (d) The ratio of SOM-explained to total sea ice concentration trends. The blue–red colorbar is used in (a)–(c), and the jet colorbar is used in (d). Dots in (a) indicate above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 5.
Fig. 5.

Trends in sea ice concentration explained by each SOM pattern (decade−1). The fraction of the total sea ice trend represented by each node is shown in left corner of each panel.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

b. Relationship to atmospheric circulations

To identify the types of atmospheric circulation patterns associated with the sea ice anomaly patterns identified by each of the nine SOM nodes, composites of anomalous fields of several atmospheric variables, including the 500-hPa geopotential height, 850-hPa wind, surface downward longwave radiation, and the 1000–750-hPa layer temperature and specific humidity, are generated based on the time series of the occurrences for each node (Figs. 610).

Fig. 6.
Fig. 6.

Composite of anomalous 500-hPa geopotential height (gpm) for each node. Dotted regions denote above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 7.
Fig. 7.

As in Fig. 6, but for the anomalous 850-hPa wind field (m s−1). Shaded regions denote above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 8.
Fig. 8.

As in Fig. 6, but for the anomalous surface downward longwave radiation (105 W m−2). Dotted regions denote above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 9.
Fig. 9.

As in Fig. 6, but for the anomalous low-level temperature (1000–750 hPa; °C). Dotted regions denote above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 10.
Fig. 10.

As in Fig. 6, but for anomalous low-level specific humidity (1000–750 hPa; g kg−1). Dotted regions denote above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Not surprisingly, the opposite anomalous sea ice concentration patterns depicted by the two groups of the SOM nodes (nodes 3, 6, 9 vs 1, 4, 7) are associated with completely opposite atmospheric circulation patterns. The nodes with all negative sea ice anomalies (nodes 1, 4, 7) correspond to an anomalous ridge over most of the Arctic region (Fig. 6). The anomalous low-level anticyclonic atmospheric circulations (Fig. 7) associated with the ridge transport warm and moist air from northern Canada into the Arctic, increasing low-level temperature (Fig. 9) and moisture (Fig. 10). The increased low-level moisture and temperature enhances surface downward longwave radiation (Fig. 8), which provides a positive feedback to low-level temperature (Ding et al. 2017). This temperature increase is consistent with the sea ice decrease in the region, as reflected by the corresponding SOM nodes, although the statistical analysis here cannot determine whether the warming resulted in sea ice reduction or vice versa. The anomalous southerly winds induced by negative height anomalies over Russia are favorable for warm and moist air being transported into the Pacific sector of the Arctic, potentially reducing sea ice there (Fig. 7). A similar case also occurs over the Barents and Kara Seas. The anomalous northerly winds over the Nordic seas favor the transport of sea ice from the Arctic, possibly reducing the Arctic sea ice cover. The anomalous atmospheric circulation patterns are reversed completely for the SOM nodes with positive sea ice anomalies (nodes 3, 6, 9).

Since the two paired nodes (nodes 3–7 and nodes 1–9), which are completely opposite in the spatial patterns and trends, account for nearly two-thirds of the total variability and close to half of the total trend in the melt-season Arctic sea ice concentrations, the analyses henceforward will focus only on these four nodes. To understand the atmospheric and oceanic background for the sea ice variability, regression analyses are performed where the anomalous seasonal SST and atmospheric variables are regressed onto the normalized time series of the occurrences for these four nodes (Figs. 1116). Before regression analysis, anomalous seasonal SST and atmospheric variables are obtained from monthly data.

Fig. 11.
Fig. 11.

Anomalous SST (°C) regressed into the time series of occurrence number for nodes 1, 3, 7, and 9.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 12.
Fig. 12.

As in Fig. 11, but for the anomalous 500-hPa geopotential height (gpm).

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 13.
Fig. 13.

As in Fig. 11, but for the anomalous 850-hPa wind field (m s−1).

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 14.
Fig. 14.

As in Fig. 11, but for the anomalous surface downward longwave radiation (105 W m−2). Dotted regions denote above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 15.
Fig. 15.

As in Fig. 11, but for the anomalous low-level temperature (1000–750 hPa; °C). Dotted regions denote above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Fig. 16.
Fig. 16.

As in Fig. 11, but for the anomalous low-level specific humidity (1000–750 hPa; 10−1 g kg−1). Dotted regions denote above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Not surprisingly, the completely opposite Arctic sea ice anomaly patterns depicted by each pair correspond to reversed regression patterns, though they are not mirrored. For node 1, the SST regression pattern in the Pacific Ocean shows warm anomalies in western and central North Pacific and cool anomalies off the coast of North America and in the eastern equatorial Pacific, resembling the cold-phase PDO (Fig. 11). The pattern is nearly reversed for node 9, although the regions of warm anomalies are much weaker compared to their counterparts for node 1. The SST regression pattern is also reversed over the North Atlantic. For both nodes 1 and 9, the anomalous SST in the western Pacific Ocean excites a wave train traveling from the northern Pacific northeastwards into the Arctic (Hoskins and Karoly 1981), but there are some noticeable differences in the phase and amplitude of the wave train (Fig. 12). For node 1, an anomalous Greenland high controls most of the Arctic Ocean, producing a stronger anticyclonic circulation in the Arctic (Fig. 13). Anomalous southeasterly winds over northeastern Canada transport warm and moist air into the Arctic, leading to mostly negative anomalies of the Arctic sea ice, especially in its Pacific sector (Fig. 13). Gong et al. (2017) noted that increased downward longwave radiation related to enhanced total column water results in a warmer Arctic. The increased downward longwave radiation due to increased humidity in the lower troposphere also elevates surface temperature in the Arctic, consistent with sea ice loss (Figs. 1416).

A good example for illustrating this association of the Arctic sea ice loss with anomalous atmospheric circulations was melt-season 2012 when SST and atmospheric circulations exhibited a pattern similar to the above, and greater Arctic sea ice loss and surface melt of the Greenland Ice Sheet were observed (Hanna et al. 2014). Along with the anomalous high, a negative-phase NAO occurs in the North Atlantic; this coupled relationship has been noted in several previous studies (Davini et al. 2012; Hanna et al. 2014). The cold flows induced by the anomalous northwesterly winds from the Arctic may result in a sea ice increase in the Greenland and the Barents Sea.

In contrast, node 9 shows negative height anomalies, and the accompanying anomalous northwesterly winds over northeastern Canada and Greenland (Fig. 12) reduce moisture and heat transport into the Pacific sector of the Arctic, potentially increasing sea ice concentration there. The southerly winds in the Atlantic sector reduce the transport of sea ice out of the Pacific sector rather than increase transport into the Pacific sector (Rigor et al. 2002; Fig. 13). The moist and warm air advected by the anomalous southerly winds could potentially lead to somewhat reduced sea ice concentration over the Nordic seas. The negative-phase AD in the Arctic also is unfavorable for the export of Arctic sea ice into the North Atlantic (Wang et al. 2009), possibly leading to sea ice increase in the Pacific sector and decrease in the Nordic and Barents Seas).

Similarly, the node 3–7 pair show an opposite anomalous SST regression pattern over the North Pacific and the North Atlantic (Fig. 11). For node 3, negative height anomalies dominate over the entire Arctic and the associated anomalous cyclonic circulation could reduce sea ice transport from the Arctic into the northern Atlantic and thus increase the sea ice cover especially in the Barents and Kara Seas. The anomalous northwesterly winds over northeastern Canada may also increase the sea ice cover there and decrease moist and warm air into the Arctic Ocean (Fig. 13). Decreased downward longwave radiation related to the reduced specific humidity reduces low-level air temperature, and a lower temperature is consistent with the Arctic sea ice increase (Figs. 1416). For node 7, positive height anomalies occur over most of the Arctic, and in particular, two blocking highs, one above Beaufort Sea and the Canadian Arctic Circle (90°W) and the other over the Greenland Sea and northern Europe (30°E) are developed. The anomalous circulations associated with the first blocking high increase surface temperature over the Beaufort Sea and northeastern Canada, consistent with the reduced sea ice concentration there. The anomalous anticyclonic circulation over the Bering Sea may also help reduce the sea ice in the Pacific sector of the Arctic. The anomalous low over the Nordic seas transports warm and moist air over northern Europe to the Barents, Kara, and Laptev Seas, potentially producing negative sea ice anomalies in these regions. This phenomenon was observed in winter of 2016 when a large increase in surface temperature and reduction of sea ice in the Barents and Kara Seas were observed as a strong blocking high developed in northern Europe (Kim et al. 2017). Sea ice depletion in the Barents Sea has also been linked to positive SST anomalies in the Gulf Stream (Sato et al. 2014).

The above relationship between the large-scale atmospheric circulations and the sea ice concentrations in different regions of the Arctic Ocean is consistent with the composite analyses where composites of the atmospheric variables for each of the four nodes (Figs. 610) exhibit similar spatial patterns to those of the regression patterns (Figs. 1216).

It is clear from the above analyses that the Arctic sea ice anomalies could be connected with the anomalous atmospheric circulations associated with changes in SST over the North Pacific Ocean. The main mode of SST variability over the North Pacific Ocean is characterized by the PDO index, which, as shown in Fig. 17, had a significant shift in phase in the late 1990s (Yu et al. 2017) that corresponds to the shift of frequency for nodes 1, 3, 7, and 9 (Fig. 2). An examination of the changes in the frequencies before and after 1998 (Table 3) of all SOM nodes shows that for the melt seasons before 1998, the frequencies of nodes 1, 4, and 7 are significantly lower than their climatology, while the frequencies of nodes 3 and 9 are higher than their climatology. The opposite occurs after 1998. The results suggest a potential connection between the phase shift of the summertime PDO and AMO indices and the occurrence frequencies of Arctic sea ice SOM patterns related to the downward trend in the melt-season Arctic sea ice.

Fig. 17.
Fig. 17.

Time series of the normalized summertime (a) PDO and (b) AMO indices. Blue lines denote the 20-yr low-pass filtered time series.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

Table 3.

Frequencies of occurrence (%) of the melt-season Arctic sea ice patterns in Fig. 1 for all summers (JJA) before 1998 and after 1998 summers for the period 1979–2016. Values with asterisks are significantly different from climatology above the 90% confidence level.

Table 3.

To examine the influence of the PDO and AMO phase shift on atmospheric circulations, the frequency of occurrence of composited anomalous 500-hPa geopotential height for the four nodes (nodes 1, 3, 7, and 9 in Fig. 6) is examined through regressing the daily anomalous 500-hPa height fields onto the composited anomalous 500-hPa height fields for each node, which yields a daily time series of regression coefficients. Figure 18 shows the probability distribution of the regression coefficients for the periods of 1979–98 and 1999–2016, respectively. There is a clear, statistically significant (95% confidence level) difference in the distribution for the two periods, which corresponds to before and after the PDO and AMO phase shift. For nodes 1 and 7, which are dominated by negative sea ice anomalies in the Arctic, the distribution tilts toward negative values for the 1979–98 period and positive values for the 1999–2016 period. The opposite occurs for nodes 3 and 9 that represent generally positive sea ice anomalies. This indicates that the atmospheric circulation pattern linked to negative sea ice anomalies (nodes 1 and 7) occurs more frequently over the 1999–2016 period, while the circulation patterns associated with positive sea ice anomalies (nodes 3 and 9) happens more frequently in the 1979–98 period. Above analysis further confirms that over the different phases of the PDO and AMO there is a significant difference in the frequency of occurrence of daily height fields, thus influencing the frequency of occurrence of different patterns of daily sea ice concentration.

Fig. 18.
Fig. 18.

The probability distribution of regression coefficients from the regressions of the anomalous daily 500-hPa height fields onto composited anomalous 500-hPa height fields for nodes 1, 3, 7, and 9 over the periods of 1979–98 and 1999–2016. The difference of probability density over two periods is statistically significant at the above 95% confidence level.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

c. Discussion

The analyses presented above use the original sea ice time series without filtering or detrending. Detrending would remove long-term sea ice trend due to factors such as global warming caused by increased greenhouse gas emissions. However, detrending would also eliminate effects on the sea ice trend by multidecadal atmosphere–ocean oscillations such as PDO and AMO. Removing long-term trend related to global warming while retaining multidecadal variability requires a time series longer than 38 years. Therefore, we obtain melt-season sea ice data from Met Office Hadley Centre for the period 1920–2015 and performed EOF analysis. Prior to applying EOF, the sea ice concentration data are 30-yr low-pass filtered and detrended, which eliminates long-term trend related to climate change but still keeps multidecadal variability. The spatial pattern and the time series of the leading EOF mode, which explains 45.3% of the total variance, are shown in Fig. 19. It is clear that the accelerated decrease since the 1990s is not simply a reflection of long-term trends associated with anthropogenic climate change, but likely is related to the phase reverse of the PDO and AMO indices. The EOF results support the potential teleconnection between the Arctic sea ice depletion and the decadal- to multidecadal-scale oscillations of sea surface temperatures in North Pacific and Atlantic.

Fig. 19.
Fig. 19.

The (a) leading pattern and (b) its time series of EOF analysis of melt-season Arctic sea ice concentration anomalies. Prior to applying EOF, the sea ice concentration data are 20-yr low-pass filtered and detrended. The PDO and AMO are 30-yr low-pass filtered and detrended.

Citation: Journal of Climate 32, 5; 10.1175/JCLI-D-18-0301.1

4. Conclusions and discussion

The SOM method has been applied to daily sea ice data to investigate the melt-season Arctic sea ice variability in the context of atmospheric circulation changes.

The results from the SOM analyses suggest the existence of two main variability modes: the Pacific sector mode and the Barents and Kara Seas mode. The positive (negative) phase of the Pacific sector mode refers to positive (negative) sea ice anomalies over the Pacific sector of the Arctic and similarly for the Barents and Kara Seas mode. The two modes explain about two-thirds of the melt-season Arctic sea ice variability. For both modes, a phase reversal from mainly positive to mainly negative occurs around the late 1990s, which coincides with a major phase shift in the PDO index. There are significant trends in the time series of the occurrence of the two modes, and together they explain nearly half of the trends in the melt-season Arctic sea ice. The climatic background for the occurrences of the two modes is investigated via regression and composite analyses. As expected, the general patterns of the SST and the associated atmospheric circulations are reversed for the opposite phase of the two modes, although some differences exist in the position and strengths of highs and lows. The positive phase of the two modes usually corresponds to negative SST anomalies over large portions of the North Pacific and Atlantic; the opposite occurs during their negative phase. The relationship indicates that PDO and AMO may have both played a role in Arctic sea ice loss. The positive phase of the Pacific sector mode and the negative phase of the Barents and Kara Seas mode correspond to the opposite variability of Greenland high and NAO. An anomalous high occurs in northeastern Canada, Greenland, Europe, and eastern Siberia during the positive phase of the Pacific sector mode and the negative phase of the Barents and Kara mode. The temperature and moisture advection and their radiative feedback associated with the anomalous atmospheric circulations are in general consistent with the variability of Arctic sea ice for the two modes.

The current study is, to our knowledge, the first to apply the SOM method to daily sea ice data to examine the variability and trends of the Arctic sea ice concentrations and their regional differences. The use of daily data allows for the capturing of synoptic- and intraseasonal-scale forcing that contribute to Arctic sea ice trend and reveal much more detail in the spatial and temporal variations of the Arctic sea ice than those in previous studies using monthly data (Ding et al. 2017; Yu and Zhong 2018). We show that the daily atmospheric circulation patterns similar to the composite pattern corresponding to the positive (negative) phase of the two modes tend to occur more frequently for the 1979–98 (1999–2016) period. This suggests that the synoptic-scale forcing under the climate background produces atmospheric circulations that are more often in favor of above-normal sea ice concentrations before the late 1990s and below-normal sea ice after that. Our results, which underline the importance of natural contributions to Arctic sea ice loss over the past several decades, may help improve the predictability of melt-season Arctic sea ice cover on synoptic–decadal time scales.

Finally, it is important to keep in mind that results presented here are fully statistical, which are unable to determine causal relationship. For example, while warm anomalies in a region of the Arctic may be consistent with negative sea ice anomalies in the same region, it cannot be certain that sea ice loss is a result of warm anomalies, or vice versa. An understanding of the causal relationship requires dynamic modeling with coupled atmosphere–ocean–sea ice models, which is beyond the scope of the current study.

Acknowledgments

This study is supported by National Key R&D Program of China (Number 2017YFE0111700). The National Center for Atmospheric Research is sponsored by the U.S. National Science Foundation.

REFERENCES

  • Alexeev, V. A., V. V. Ivanov, R. Kwok, and L. H. Smedsrud, 2013: North Atlantic warming and declining volume of Arctic sea ice. Cryosphere Discuss., 7, 245265, https://doi.org/10.5194/tcd-7-245-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bintanja, R., R. G. Graversen, and W. Hazeleger, 2011: Arctic winter warming amplified by the thermal inversion and consequent low infrared cooling to space. Nat. Geosci., 4, 758761, https://doi.org/10.1038/ngeo1285.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, H. W., R. B. Alley, and F. Zhang, 2016: Interannual Arctic sea ice variability and associated winter weather patterns: A regional perspective for 1979-2014. J. Geophys. Res. Atmos., 121, 14 43314 455, https://doi.org/10.1002/2016JD024769.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davini, P., C. Cagnazzo, R. Neale, and J. Tribbia, 2012: Coupling between Greenland blocking and the North Atlantic Oscillation pattern. Geophys. Res. Lett., 39, L14701, https://doi.org/10.1029/2012GL052315.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, https://doi.org/10.1002/qj.828.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Deser, C., and H. Teng, 2008: Evolution of Arctic sea ice concentration trends and the role of atmospheric circulation forcing, 1979-2007. Geophys. Res. Lett., 35, L02504, https://doi.org/10.1029/2007GL032023.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Deser, C., J. E. Walsh, and M. S. Timlin, 2000: Arctic sea ice variability in the context of recent atmospheric circulation trends. J. Climate, 13, 617633, https://doi.org/10.1175/1520-0442(2000)013,0617:ASIVIT.2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ding, Q., J. M. Wallace, D. S. Battisti, E. J. Steig, A. J. E. Gallant, H.-J. Kim, and L. Geng, 2014: Tropical forcing of the recent rapid Arctic warming in northeastern Canada and Greenland. Nature, 509, 209212, https://doi.org/10.1038/nature13260.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ding, Q., and Coauthors, 2017: Influence of high-latitude atmospheric circulation changes on summertime Arctic sea ice. Nat. Climate Change, 7, 289295, https://doi.org/10.1038/nclimate3241.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Enfield, D. B., A. M. Mestas-Nuñez, and P. J. Trimble, 2001: The Atlantic multidecadal oscillation and its relationship to rainfall and river flows in the continental U.S. Geophys. Res. Lett., 28, 20772080, https://doi.org/10.1029/2000GL012745.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flanner, M. G., K. M. Shell, M. Barlage, D. K. Perovich, and M. A. Tschudi, 2011: Radiative forcing and albedo feedback from the Northern Hemisphere cryosphere between1979 and 2008. Nat. Geosci., 4, 151155, https://doi.org/10.1038/ngeo1062.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gagné, M.-È., N. P. Gillett, and J. C. Fyfe, 2015: Impact of aerosol emission controls on future arctic sea ice cover. Geophys. Res. Lett., 42, 84818488, https://doi.org/10.1002/2015GL065504.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gibson, P. B., S. E. Perkins-Kirkpatrick, P. Uotila, A. S. Pepler, and L. V. Alexander, 2017: On the use of self-organizing maps for studying climate extremes. J. Geophys. Res. Atmos., 122, 38913903, https://doi.org/10.1002/2016JD026256.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gong, T., S. Feldstein, and S. Lee, 2017: The role of downward infrared radiation in the recent Arctic winter warming trend. J. Climate, 30, 49374949, https://doi.org/10.1175/JCLI-D-16-0180.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hanna, E., and Coauthors, 2014: Atmospheric and oceanic climate forcing of the exceptional Greenland ice sheet surface melt in summer. Int. J. Climatol., 34, 10221037, https://doi.org/10.1002/joc.3743.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hoskins, B. J., and D. J. Karoly, 1981: The steady linear response of a spherical atmosphere to thermal and orographic forcing. J. Atmos. Sci., 38, 11791196, https://doi.org/10.1175/1520-0469(1981)038<1179:TSLROA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, Y., X. Dong, B. Xi, E. K. Dolinar, and R. E. Stanfield, 2017: The footprints of 16 year trends of Arctic springtime cloud and radiation properties on September sea ice retreat. J. Geophys. Res. Atmos., 122, 21792193, https://doi.org/10.1002/2016JD026020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kapsch, M.-L., R. G. Graversen, T. Economou, and M. Tjernström, 2014: The importance of spring atmospheric conditions for predictions of the Arctic summer sea ice extent. Geophys. Res. Lett., 41, 52885296, https://doi.org/10.1002/2014GL060826.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kapsch, M.-L., R. G. Graversen, M. Tjernström, and R. Bintanja, 2016: The effect of downwelling longwave and shortwave radiation on Arctic summer sea ice. J. Climate, 29, 11431159, https://doi.org/10.1175/JCLI-D-15-0238.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kim, B.-M., and Coauthors, 2017: Major cause of unprecedented Arctic warming in January 2016: Critical role of an Atlantic wind storm. Sci. Rep., 7, 40051, https://doi.org/10.1038/srep40051.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kohonen, T., 2001: Self-Organizing Maps. 3rd ed. Springer, 501 pp.

    • Crossref
    • Export Citation
  • Lee, S., and S. B. Feldstein, 2013: Detecting ozone- and greenhouse gas–driven wind trends with observational data. Science, 339, 563567, https://doi.org/10.1126/science.1225154.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, S., T. Gong, N. C. Johnson, S. B. Feldstein, and D. Pollard, 2011: On the possible link between tropical convection and the Northern Hemisphere Arctic surface air temperature change between 1958 and 2011. J. Climate, 24, 43504367, https://doi.org/10.1175/2011JCLI4003.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liu, Z., and A. Schweige, 2017: Synoptic conditions, clouds, and sea ice melt onset in the Beaufort Chukchi seasonal ice zone. J. Climate, 30, 69997016, https://doi.org/10.1175/JCLI-D-16-0887.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lynch, A. H., M. C. Serreze, E. N. Cassano, A. D. Crawford, and J. Stroeve, 2016: Linkages between Arctic summer circulation regimes and regional sea ice anomalies. J. Geophys. Res. Atmos., 121, 78687880, https://doi.org/10.1002/2016JD025164.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahajan, S., R. Zhang, and T. L. Delworth, 2011: Impact of the Atlantic meridional overturning circulation (AMOC) on Arctic surface air temperature and sea-ice variability. J. Climate, 24, 65736581, https://doi.org/10.1175/2011JCLI4002.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78, 10691079, https://doi.org/10.1175/1520-0477(1997)078<1069:APICOW>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mills, C. M., and J. E. Walsh, 2014: Synoptic activity associated with sea ice variability in the Arctic. J. Geophys. Res. Atmos., 119, 12 11712 131, https://doi.org/10.1002/2014JD021604.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Min, S.-K., X. Zhang, F. W. Zwiers, and T. Agnew, 2008: Human influence on Arctic sea ice detectable from early 1990s onwards. Geophys. Res. Lett., 35, 213236, https://doi.org/10.1029/2008GL035725.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mortin, J., G. Svensson, R. G. Graversen, M.-L. Kapsch, J. C. Stroeve, and L. N. Boisvert, 2016: Melt onset over Arctic sea ice controlled by atmospheric moisture transport. Geophys. Res. Lett., 43, 66366642, https://doi.org/10.1002/2016GL069330.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Notz, D., and J. Marotzke, 2012: Observations reveal external driver for Arctic sea-ice retreat. Geophys. Res. Lett., 39, L08502, https://doi.org/10.1029/2012GL051094.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ogi, M., K. Yamazaki, and J. M. Wallace, 2010: Influence of winter and summer surface wind anomalies on summer Arctic sea ice extent. Geophys. Res. Lett., 37, L07701, https://doi.org/10.1029/2009GL042356.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Park, W., and M. Latif, 2008: Multidecadal and multicentennial variability of the meridional overturning circulation. Geophys. Res. Lett., 35, L22703, https://doi.org/10.1029/2008GL035779.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pithan, F., and T. Mauritsen, 2014: Arctic amplification dominated by temperature feedbacks in contemporary climate models. Nat. Geosci., 7, 181184, https://doi.org/10.1038/ngeo2071.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Polyakov, I. V., and Coauthors, 2010: Arctic Ocean warming contributes to reduced polar ice cap. J. Phys. Oceanogr., 40, 27432756, https://doi.org/10.1175/2010JPO4339.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, https://doi.org/10.1029/2002JD002670.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Reusch, D. B., R. B. Alley, and B. C. Hewitson, 2005: Relative performance of self-organizing maps and principal component analysis in pattern extraction from synthetic climatological data. Polar Geogr., 29, 188212, https://doi.org/10.1080/789610199.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rigor, I., J. M. Wallace, and R. L. Colony, 2002: Response of sea ice to the Arctic Oscillation. J. Climate, 15, 26482663, https://doi.org/10.1175/1520-0442(2002)015<2648:ROSITT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sato, K., J. Inoue, and M. Watanabe, 2014: Influence of the Gulf Stream on the Barents Sea ice retreat and Eurasian coldness during early winter. Environ. Res. Lett., 9, 084009, https://doi.org/10.1088/1748-9326/9/8/084009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schudeboom, A., A. J. McDonald, O. Morgenstern, M. Harvey, and S. Parsons, 2018: Regional regime-based evaluation of present-day GCM cloud simulations using self-organizing maps. J. Geophys. Res. Atmos., 123, 42594272, https://doi.org/10.1002/2017JD028196.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sedlar, J., and Coauthors, 2011: A transitioning Arctic surface energy budget: The impacts of solar zenith angle, surface albedo and cloud radiative forcing. Climate Dyn., 37, 16431660, https://doi.org/10.1007/s00382-010-0937-5.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J., J. Zhang, E. Watanabe, M. Ikeda, K. Mizobata, J. E. Walsh, X. Bai, and B. Wu, 2009: Is the dipole anomaly a major driver to record lows in Arctic summer sea ice extent? Geophys. Res. Lett., 36, L05706, https://doi.org/10.1029/2008GL036706.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Woodgate, R. A., T. J. Weingartner, and R. Lindsay, 2012: Observed increases in Bering Strait oceanic fluxes from the Pacific to the Arctic from 2001 to 2011 and their impacts on the Arctic Ocean water column. Geophys. Res. Lett., 39, L24603, https://doi.org/10.1029/2012GL054092.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yu, L., and S. Zhong, 2018: Changes in sea-surface temperature and atmospheric circulation patterns associated with reductions in Arctic sea ice cover in recent decades. Atmos. Chem. Phys., 18, 14 14914 159, https://doi.org/10.5194/acp-18-14149-2018.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yu, L., S. Zhong, J. A. Winkler, M. Zhou, D. H. Lenschow, B. Li, X. Wang, and Q. Yang, 2017: Possible connection of the opposite trends in Arctic and Antarctic sea ice cover. Sci. Rep., 7, 45804, https://doi.org/10.1038/srep45804.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, J., D. A. Rothrock, and M. Steele, 1998: Warming of the Arctic Ocean by a strengthened Atlantic inflow: Model results. Geophys. Res. Lett., 25, 17451748, https://doi.org/10.1029/98GL01299; Corrigendum, 25, 3541, https://doi.org/10.1029/98GL02634.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, R., 2015: Mechanisms for low-frequency variability of summer Arctic sea ice extent. Proc. Natl. Acad. Sci. USA, 112, 45704575, https://doi.org/10.1073/pnas.1422296112.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save