Fig. 1
Mean ice concentration field from 1979 to 2012, including the ice edge (15% ice concentration, solid black) and the mean ice concentration in the study region (48%, dashed black). The black box shows the study region north of Svalbard. The position of the West Spitsbergen Current (WSC) temperature measurements at Sørkapp is indicated in yellow, while the red arrow indicates the pathway of the Svalbard Branch of the WSC. The bathymetry is drawn as thin grey lines.
Fig. 2
Schematic of air-ice-sea interactions north of Svalbard. Northerly winds transport sea ice from the Arctic Ocean (slightly deflected to the right) and bring cold air masses, facilitating larger ice cover. Upwelling of warm Atlantic Water (AW, reddish) melts the approaching sea ice, and a fresh, cold layer forms below the ice (bluish). Depending on the vertical mixing below the ice, the freshwater layer reduces further ice melt. The large ocean-to-atmosphere heat flux, Q, is strongly reduced by the presence of sea ice. In winter, Q is the sum of net longwave radiation, latent heat and sensible heat. Excess heat is lost to space through longwave radiation. Ta: air temperature, and Tw: water temperature.
Fig. 3
Upper: Monthly averaged satellite-observed sea ice concentration (red line, September to September) including the standard deviation (grey region). The dashed line shows the annual mean ice concentration. Lower: The total 1979–2012 ice concentration reduction for each month is shown as black dots, with error bars indicating the 95% confidence intervals.
Fig. 4
Satellite-derived sea ice concentration anomalies north of Svalbard, 1979–2012. Dark blue is winter (a) and summer (b) means; and red shows the 3 yr running mean. The zero line represents the winter (56%) and summer (40%) 1979–2012 average. Linear trends are shown in light blue, and the shaded areas indicate the 95% confidence interval for the trend estimates.
Fig. 5
Contour lines of the 40% winter (DJFM) ice concentration north of Svalbard during the 1980s (dark blue), 1990s (light blue), and 2000s (green). The most recent winters are also included with dashed lines (2010: yellow, 2011: purple, and 2012: red). In addition, February 2012 is shown in grey (dashed) to indicate a period with especially low ice concentrations, and ice-free areas extending towards Franz Josef Land. The black box indicates the study region.
Fig. 6
Winter (DJFM) averaged sea ice concentration (a), AW temperature (b), air temperature (c), north–south wind component (positive from the south) (d), and east–west wind component (positive from the west) (e), 1979–2012. Note that the ice concentration (a) is inverted. Statistical significant linear trends (95%) are indicated by straight lines.
Table 1. Correlations between winter (DJFM) means
Ice— − 0.50 − 0.20 − 0.900.11v −0.37 — − 0.130.510.06AW −0.41 −0.17—0.27 − 0.05Ta −0.92 0.36 0.47 —0.03u0.090.05 −0.030.04—
[i] The upper right triangle (roman) represents correlations between detrended time series, while the lower left triangle (italic) shows correlations including trends. Significant correlations within a 95% confidence level are marked with boldface. Ice: ice concentration, v: north–south wind, AW: Atlantic Water temperature, Ta: air temperature, u: east–west wind.