North Atlantic Ocean Circulation and Decadal Sea Level Change During the Altimetry Era

Abstract

Regional sea-level rise is characterized by decadal acceleration and deceleration periods that typically stem from oceanic climate variability. Here, we investigate decadal sea-level trends during the altimetry era and pin down the associated ocean circulation changes. We find that decadal subpolar gyre cooling (warming), strengthening (weakening), widening (shrinking) since the mid-2000s (early 1990s) resulted in negative (positive) sea level trends of -7.1 mm/yr ± 1.3 mm/yr (3.9 mm/yr ± 1.5 mm/yr). These large-scale changes further coincide with steric sea-level trends, and are driven by decadal-scale ocean circulation variability. Sea level on the European shelf, however, is found to correlate well with along-slope winds (R = 0.78), suggesting it plays a central role in driving the associated low-frequency dynamic sea level variability. Furthermore, when the North Atlantic is in a cooling (warming) period, the winds along the eastern boundary are predominantly from the North (South), which jointly drive a slowdown (rapid increase) in shelf and coastal sea level rise. Understanding the mechanisms that produce these connections may be critical for interpreting future regional sea-level trends.

Figure 1

Regional dynamic and steric sea level trends in the North Atlantic Ocean. The dynamic and steric sea level trends (mm/yr) as deduced from altimetry and EN4 hydrographic dataset, respectively, for the (a,b) 1993–2004 and (c,d) 2005–2016 periods. The global mean sea-level and steric height trend of each period have been removed before calculating the regional trends. The stipplings indicate the non-significant regions at the 95% confidence level using the modified Mann-Kendall test. The gray contours in all panels depict the 500 and 1000-m isobaths. (e) The averaged monthly dynamic sea level in the SPNA [45°W–5°W, 55°N–65°N] (see box in panel c) as calculated from altimetry. The global mean sea level (Fig. S1) has been removed beforehand. The black line is the smoothed SPNA sea level using a 25-month running mean in order to represent interannual-to-decadal timescales. Anomalously high/low sea levels do not only represent warming/cooling of the upper ocean but also the strength of the subpolar gyre. The figure was produced using the software Matlab R2014b, (https://www.mathworks.com).

Figure 2

Atmospheric-driven ocean changes. Trends of wind-stress curl (WSC) calculated using NCEP/NCAR reanalysis for (a) the 1993–2004 WP and (b) the 2005–2016 CP. The vectors in the upper panels represent the wind-stress trends. The circles in the upper panels indicate the regions that are significant at the 95% confidence level. Trends of mean kinetic energy (MKE) calculated using altimetry for (c) the 1993–2004 WP and (d) the 2005–2016 CP. The stipplings indicate the non-significant regions at the 95% confidence level. The light grey contours depict the time-invariant mean dynamic topography (CNES-CLS2013 MDT) that ranges from 0.8 m to −0.8 m with a spacing of 0.1 m. Three of these contours (−0.1, −0.2 and −0.3 m) are highlighted in black as they trace the different branches of the NAC that continue into the northeast Atlantic. The gray contours in all panels depict the 500 and 1000-m isobaths. The figure was produced using the software Matlab R2014b, (https://www.mathworks.com).

Figure 3

Low-frequency dynamic shelf sea level variability. (a) Time-latitude diagram of regional dynamic shelf sea level variability in the eastern Atlantic with no smoothing applied. (b) Averaged regional dynamic sea-level variability over the east Atlantic shelf (shallower than the 500-m isobath) calculated from altimetry for the 1993–2016 period (shading) overlaid by the latitudinally-averaged along-slope wind-stress (black). The time series have been deseasoned, smoothed with a 25-month running mean and normalized. Dark blue and red indicate anomalous periods (higher and lower than 0.5 standard deviation). (c) Composite analysis of sea level (shading) and wind-stress (vectors) calculated based on the difference between the anomalously high and low dynamic shelf sea-level periods shown in panel (a) but excluding the first and last year due to smoothing effects. (d) Monthly (thin lines) and smoothed (thick lines) relative sea-level variability from tide gauges in the North (red) and Norwegian (blue) Seas. The figure was produced using the software Matlab R2014b, (https://www.mathworks.com).