Intensification of Near‐Surface Currents and Shear in the Eastern Arctic Ocean

A 15‐year (2004–2018) record of mooring observations from the upper 50 m of the ocean in the eastern Eurasian Basin reveals increased current speeds and vertical shear, associated with an increasing coupling between wind, ice, and the upper ocean over 2004–2018, particularly in summer. Substantial increases in current speeds and shears in the upper 50 m are dominated by a two times amplification of currents in the semidiurnal band, which includes tides and wind‐forced near‐inertial oscillations. For the first time the strengthened upper ocean currents and shear are observed to coincide with weakening stratification. This coupling links the Atlantic Water heat to the sea ice, a consequence of which would be reducing regional sea ice volume. These results point to a new positive feedback mechanism in which reduced sea ice extent facilitates more energetic inertial oscillations and associated upper‐ocean shear, thus leading to enhanced ventilation of the Atlantic Water.


Introduction
Throughout much of the Arctic Ocean layers of colder water isolate the sea surface from warm (temperature >0°C) and salty water of Atlantic origin (Atlantic Water, AW), which is transported throughout the Arctic Ocean at intermediate depths (~150-900 m) as topographically steered boundary currents (e.g., Aagaard, 1989;Rudels et al., 1994). The AW holds enough heat to melt Arctic sea ice several times over but is separated from the near-freezing, relatively fresh water in the Arctic Ocean surface mixed layer by a cold halocline layer which has a negligible vertical temperature gradient but a large salinity gradient. The associated density gradient impedes vertical mixing of AW heat upward to the sea ice (e.g., Fer, 2009;Rudels et al., 1996).
In the absence of significant shear-driven mixing, the vertical structure of cooler and fresher halocline water overlying the warmer and saltier AW facilitates double diffusive convection which mediates vertical heat fluxes across the lower halocline into the seasonal convective layer (Carmack et al., 2015). Double diffusion is driven by the different molecular diffusivities of heat and salt and is evident in vertical hydrographic profiles as multiple layers of near-uniform temperature and salinity that are separated by strong-gradient, thin ©2020. The Authors. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. interfaces. These "thermohaline staircases" have been found over a large portion of the Arctic Ocean in the lower halocline water above the depth of maximum temperature in the AW layer (e.g., Guthrie et al., 2017;Polyakov et al., 2019;Shibley et al., 2017), and are laterally coherent for more than 800 km in the Canada Basin (Timmermans et al., 2008) and for more than 1,000 km in the Eurasian Basin (EB, Polyakov et al., 2019). In the presence of the staircases, the vertical heat flux out of the AW is limited to between O (0.1) W/m 2 in the central basins (Guthrie et al., 2015;Padman & Dillon, 1987;Sirevaag & Fer, 2012) and O(1) W/m 2 over the Laptev Sea slope (Lenn et al., 2009;Polyakov et al., 2012Polyakov et al., , 2019.

10.1029/2020GL089469
It is long established that the influence of the wind in driving mixing is significantly enhanced in open water compared to under ice through the generation of inertial oscillations in the upper ocean, with associated shear leading to shear instability and turbulent mixing (Lenn et al., 2011;Rainville & Woodgate, 2009). This led to the idea that the decline of seasonal sea ice area would result in increased mixing by the wind, thereby increasing the influence of AW in melting the sea ice (a positive feedback). However, the influence of the inertial oscillations is restricted to the upper ocean. For example, microstructure measurements made in the ice-free Canada Basin during the 2012 "perfect storm" show that while there is significantly enhanced turbulence in the upper 50 m of the water column in response to the storm, the AW heat remained isolated by the strongly stratified halocline, across which thermohaline staircase structures persisted (Lincoln et al., 2016).
In contrast, shear instabilities drive AW heat fluxes of up to 50 W/m 2 in the upper ocean in the western Nansen Basin near sloping topography (Carmack et al., 2015;Fer et al., 2010;Padman & Dillon, 1991;Rippeth et al., 2015), greatly diminishing the influence of the halocline and in consequence leading to a locally increasing influence of AW heat on the sea ice extent.
In the eastern (east of 70°E) EB the halocline (60-150 m) includes the cold halocline layer and lower halocline water associated with strong vertical temperature and salinity gradients directly above the AW layer. The eastern EB halocline has become warmer and saltier since the 1970s, with a coincident weakening in stratification (Polyakov et al., 2010;Steele & Boyd, 1998). By the mid-2010s, increased AW heat fluxes were found to affect sea ice loss in the eastern EB, with the disappearance of the cold halocline layer observed during the winters of 2013-2015 (Polyakov et al., 2017). These changes made this region structurally similar to the western (west of 70°E) EB, which is closer to the AW source in Fram Strait. The combination of weaker stratification and shoaling of the AW, coupled with net loss in sea ice, has allowed progressively deeper winter ventilation and larger upward AW heat fluxes in the eastern EB (Polyakov et al., 2017). This so-called "atlantification" represents a transition toward a new Arctic climate state, in which AW heat is exerting a substantially increased influence on the seasonal freeze/melt freshwater cycle, with a trend toward reduced sea ice volume.
The aim of this paper is to examine the impact of the atlantification in the eastern EB on time-varying currents in the upper ocean and, by implication, mixing. To this end, we present a unique 15-year time series of upper ocean currents from an array of moorings stretching from the shelf break in the eastern EB to the basin interior. The data set is analyzed to isolate time-varying currents in the inertial-tidal band which are then compared to local wind conditions and sea ice state to examine the relationship between the changing environmental conditions and the currents.

Mooring ADCP Measurement
We use observations of ocean currents from moorings deployed in the eastern EB from 2004 to 2018; see Figure 1a for locations and Table 1 for deployment periods and instruments. The longest record is from the M1 4 mooring site, with several collocated moorings deployed and recovered annually prior to 2009, and longer-duration deployments after that.
Seasonal cross-slope displacements of the boundary current were only observed over the upper slope (M1 1 and M1 2 moorings); see Baumann et al. (2018).
Most moorings used in this analysis included upward-looking 300-kHz Acoustic Doppler Current Profilers (ADCP, Teledyne RD Instruments) targeting the upper 50-60 m of the water column (Table 1). ADCPs

10.1029/2020GL089469
Geophysical Research Letters provided current velocities, averaged over 2-m (prior to 2013) or 4-m (after 2013) vertical cells, with 1-hr time resolution. The manufacturer's estimates for 300-kHz ADCP accuracies are better than 1 cm/s for hourly averaged (over 50 single profiles) speed and 2deg for current direction.

Winds
Daily 10-m wind output with a spatial resolution of 0.75°from the European Centre for Medium-Range Weather Forecasts reanalysis ERA-Interim (Dee et al., 2011) was used to evaluate the wind speed at the mooring locations.

Sea Ice Concentrations
Daily-averaged SMMR, SSM/I, SSMIS satellite observations of sea ice concentrations for 2004-2018 (NASA team algorithm; available from ftp://sidads.colorado.edu/pub/DATASETS/nsidc0051_gsfc_nasateam_ seaice/final-gsfc/north/daily/) are the primary data set used to estimate ice conditions at EB mooring locations. This data set is provided on a polar stereographic grid with a 25-km spatial resolution.

Calculating Vertical Shear of Horizontal Currents
We filtered higher resolution 2 m ADCP vertical velocity profiles collected prior to 2013 with a running-mean filter to reduce resolution to 4 m, equivalent to the 2013-2018 ADCP observations. The vertical shear was calculated every hour using differences over a 4 m vertical scale.

Cross-Correlation Analysis
Cross-correlation analysis between wind and sea ice forcing and ocean currents and shear is based on dailyaveraged data. A thirty-day running window is used to calculate correlations for each date: these correlations were then averaged to obtain monthly estimates. The typical number of data points used for correlation analysis of summer (<75%) and winter (>95%) sea ice concentration and currents and shear was about 110 and 200, respectively.

Amplification of Upper-Ocean Currents and Shear in the Eastern EB
The original hourly ADCP records of total current speed (|U|) and shear (|U z |) in the upper~30-50 m layer are shown in supporting information Figure S1. Annual, winter (November-July), and summer (August-October) data are shown in Figure 2. The mean value of |U| for each mooring deployment period Figure 3. Increasing coupling between wind and sea ice forcing and ocean currents and their shear. (a) Total duration when sea ice concentration (C ice ) was less than 75% (circles) and greater than 95% (diamonds) and correlations (color) between C ice and current shear averaged over these two periods. (b-d) Monthly mean wind speed and correlations (color) between wind speed and current shear in (b) winter, (c) summer, and (d) transition season. All correlations are computed using daily data and 30-day running window. Statistically significant correlations are marked by larger circles. Correlation patterns for |U| are identical to patterns for shear (a, b) and are not shown here.  (Figure 2c). There is no evident change of |U| over the same years in winter (Figure 2b). These results are consistent with findings of increased mobility of sea ice and geostrophic currents in recent decades (e.g., Armitage et al., 2017;Kwok et al., 2013;Rampal et al., 2009). Shear (|U z |) in the upper 50 m has increased by about 40% for annual values, and~90% for summer values, between 2004-2007 and 2013-2018 (Figures 2d-2f).

Geophysical Research Letters
This amplification of the upper ocean |U| and |U z | is associated with increasing coupling between wind, ice, and oceanic currents, as evidenced by the increase, in time, of (a) the negative correlation between sea ice concentration and shear and (b) the positive correlation between wind speed and shear at the M1 4 mooring site over 2004-2018 (Figure 3). During this period, annually averaged sea ice concentration in the eastern EB decreased by −0.5 ± 0.2% per year (not shown). This decline is dominated by the sea ice retreat in the summer (−0.6 ± 0.4% per year in September) with a more modest reduction in winter concentration (−0.10 ± 0.04% per year in April). Intensification of coupling expressed as an increase in time of the negative correlations between the sea ice concentration and currents and their shear is more pronounced in summer than in winter (Figure 3a), consistent with Rainville and Woodgate (2009). Over this period, there was no significant trend in wind speed. Analysis of seasonal changes of coupling between wind speed and |U| and |U z | reveals increasingly strong correlations in summer and, especially, during the sea ice transition seasons; however, in winter the increase of coupling is not as strong. This is consistent with the lack of a significant increase in the overall winter-average shear (Figure 2e).
For the 15-year duration of the mooring observations, the increase in |U| in the upper 50 m of the water column is almost completely accounted for by an increase in currents in a frequency band centered near two cycles per day (Figure 2). This band includes the semidiurnal tides and, at these latitudes, inertial oscillations forced by changes in wind stress. This frequency band often dominates Arctic Ocean current and sea ice velocity variability (e.g., Dosser & Rainville, 2016;Gimbert et al., 2012;Lenn et al., 2011;Rainville & Woodgate, 2009).
We band-pass filtered the observed hourly current records to isolate oscillations with periods 10-14 hr and so retain only the semidiurnal-band current (SBC). Near-surface SBCs and shears increased by a The largest SBCs and shears in the 10-50 m layer correspond to the ice-free season in all years presented (Figures 4 and S2), consistent with previous observations showing that compact sea ice cover dampens the semidiurnal ocean response to wind forcing (Lenn et al., 2011;Lincoln et al., 2016;Rainville & Woodgate, 2009). However, there were significant SBCs and SBC shears even under winter sea ice conditions between 2008 and 2018, in contrast to a lack of significant SBC energy from the earlier winters of 2004-2007.

Discussion
Measurements of currents from a 15-year duration mooring record in the eastern EB of the Arctic Ocean demonstrate that the previously identified weakening of stratification in the halocline (e.g., Polyakov et al., 2017Polyakov et al., , 2018 has been accompanied by increased upper-ocean current speeds and associated current shear. Most of this increased energy and shear is in the semidiurnal band, which includes baroclinic tides and wind-driven inertial oscillations, with little change of mean along-slope water transport . The increased shear presented in this research together with the weakening stratification identified earlier indicate a greater potential for shear-driven turbulent mixing, consistent with the recent transition in sea ice and upper ocean state to conditions previously unique to the western Nansen Basin (Polyakov et al., 2017).
We hypothesize that this increased coupling between AW heat and the sea ice may lead to a positive feedback between reduced sea ice and higher mixing rates as the longer periods and increased areal extent of open water facilitate more energetic wind-driven inertial oscillations (and, potentially, less damping of baroclinic tidal currents) and associated upper-ocean shear coinciding with weakening halocline stratification.
As sea ice declines, a new Arctic state is emerging which, due to the positive feedback mechanism outlined above, may be pushing the system toward a tipping point. Both observations (e.g., Polyakov et al., 2005;Schauer et al., 2004;Woodgate et al., 2001) and modeling results (Karcher et al., 2003) indicate that AW fluctuations in the Arctic Ocean interior are also linked to the highly variable nature of the AW inflows, with abrupt cooling/warming events. In future, such a pulse of AW may lead to a permanently Atlantic-dominated state in this region, wherein the hydrographic structure of the halocline no longer insulates the AW heat from the sea ice, even during later periods of weaker AW heat input. This transition of the 10.1029/2020GL089469

Geophysical Research Letters
eastern Arctic Ocean toward a new state points to the need for models including the region to resolve changing AW inputs as well as sensitivity of the time-varying currents to the evolving stratification and sea ice state.