Vertical and lateral water properties and density structure within the Arctic Ocean are intimately related to the ocean circulation, and have profound consequences for sea ice growth and retreat as well as for propagation of acoustic energy at all scales.

Recently, there has been significant arctic warming, accompanied by changes in the extent, thickness distribution, and properties of the arctic sea ice cover. Summertime sea ice extent has been declining since at least 1979.

Sea ice has become younger alongside the decreases in extent. Sea ice thickness typically increases with age, such that the combined trends toward decreasing extent and younger mean age point to a persistent loss of sea ice volume. Thinner, younger ice tends to be weaker, more subject to deformation and fracturing, and thus more mobile and more likely to provide efficient coupling between the atmosphere and upper ocean. Furthermore, the growing summertime expanses of open water provide periods when the dynamics might more closely resemble those that govern the upper ocean at lower latitudes.

The need to understand these changes and their impact on arctic stratification and circulation, sea ice evolution, and the acoustic environment motivate the Office of Naval Research (ONR) Stratified Ocean Dynamics of the Arctic Departmental Research Initiative (SODA DRI).

What are the causes and consequences of the changing upper ocean stratification (mixed layer and Pacific Water)?

Vertical: What is the interplay among wind entrainment, convection, solar heating, and buoyancy input?

• On synoptic, seasonal, and year-long integral timescales?
• How does this balance vary with sea ice conditions?

Lateral: What is the importance of heterogeneity in mesoscale and submesocale processes?

• On synoptic, seasonal, and year-long integral timescales?
• How spatially variable is this balance? (Is it more important at some locations, e.g., ice edge, mesoscale fronts, etc.?

How is the wind stress partitioned within the ice–ocean system (depth and frequency; surface waves; ice motions; mixed layer and deeper acceleration; internal waves)?

• How do sea ice properties affect this partition?
• How do the buoyancy flux and stratification affect the partition?
• How do lateral contrasts in forcing or ocean structure affect the partition and create secondary circulations?
• How does the phenology of the air–ice–ocean system affect the partition?

What is the fate and impact of the significant increase in upper ocean heat (and the associated sound channel)?

What processes control the near surface temperature maximum formation and release?

• How does it impact the fall sea ice freeze-up?
• How does it vary with sea ice conditions?

How difficult is it for the mixed layer to entrain and access heat carried in the Pacific Water, and how does it vary with sea ice conditions?

Water mass exchange between the Arctic and subpolar Atlantic and Pacific oceans (and the inputs of shelf waters along the perimeter of the deep basin), and the local momentum and buoyancy transfers between the atmosphere, ice, and upper ocean govern Arctic Ocean stratification and circulation.

The primary sources for arctic water masses are the saline water from the Atlantic (S~35) entering through Fram Strait and the shallower seas to its east, fresher seawater from the Pacific (S~32) entering through Bering Strait, freshwater runoff, predominately from the large Eurasian rivers, and a net positive precipitation minus evaporation (P–E).

Among the most prominent features of the present-day Arctic is the amplified seasonality of sea ice extent that exposes vast regions to a broad range of ice conditions over an annual cycle. The physical processes that govern the coupled atmosphere–ice–ocean system evolve with this annual cycle in ice cover and sea ice properties.

The combination of ice cover, which modulates momentum and buoyancy transfer between the atmosphere and upper ocean, and the strong vertical density contrast created by the fresh mixed layer and cold halocline, inhibit the processes that drive diapycnal mixing.

Subsurface water mass modification thus occurs slowly along circulation pathways, and arctic sea ice has been largely insulated from subsurface heat carried within the Atlantic and Pacific inflows. But near the surface, variability in sea ice properties imprints onto upper ocean structure by providing a time-varying buoyancy source (fresh water and brine) and by modulating the coupling between the atmosphere and ocean (momentum and heat).

Changes in the efficiency of air–sea momentum and heat transfer will drive changes in the processes that govern lateral and vertical exchanges in the water column. More efficient coupling between the atmosphere and upper ocean could enhance entrainment at the mixed layer base and internal wave generation. Given the contrasting water masses present in the upper ocean, enhanced vertical exchanges associated with these processes will impact stratification and circulation.

The SODA DRI focuses on how such changes modify the transfer of momentum and buoyancy from the atmosphere into the upper ocean, and their role in governing upper ocean stratification, circulation, and acoustic propagation within the Arctic.

The surface circulation of the Arctic Ocean is traditionally characterized by the anticyclonic Beaufort Gyre in the west and the Transpolar Drift across the Arctic towards Fram Strait in the east. The Beaufort Gyre, composed mainly of Pacific-derived waters in the upper few hundred meters of the water column, is driven by the time-mean anticyclonic wind stress associated with the Beaufort High in atmospheric pressure. The geostrophic component of the Transpolar Drift is aligned with a watermass front between Atlantic-derived upper ocean waters on the Eurasian side of the Arctic Ocean and Pacific-derived upper ocean waters on the North American side of the Arctic Ocean. At depths of 200–400 m, Atlantic Water circulates cyclonically in boundary currents around the major basins of the Arctic Ocean descending below the fresher, lighter Pacific-derived waters in the Canada Basin.

The Beaufort Gyre circulation has an important role in storing the fresh water that is the source of its stratification (Proshutinsky et al., 2002; Proshutinsky et al., 2009). The anticyclonic wind stress of the Beaufort High forces convergence of the Ekman transport of relatively fresh surface water. This domes the surface so as to drive the anticyclonic ocean circulation. From the early to mid 2000s, this doming intensified, the surface layer freshened, and the freshwater storage increased (Farrell et al., 2012; Morison et al., 2012; Proshutinsky et al., 2009). The process has been linked to a shift of basin circulation to an increasingly anticyclonic mode as defined by the Arctic Ocean Oscillation Index (AOOI), equivalent to the sea surface height gradient across the Beaufort Gyre in a barotropic ocean model (Proshutinsky and Johnson, 1997; Proshutinsky et al., 2001). On annual to interannual time scales, atmospheric circulation (e.g., Arctic Oscillation index) plays a large role in controlling storage, redistribution, and export of fresh water in the Arctic (Carmack et al., 2016). The relative distribution of freshwater export between the major exit pathways, through the Canadian Archipelago and Nares Strait or Fram Strait, is unclear. These exits lead to the Greenland and Labrador seas — regions that drive the thermohaline circulation of the world’s oceans.

Mesoscale eddies represent another lateral transport mechanism in the ocean. Eddies are believed to derive from instabilities of the larger-scale baroclinic flow field, such as the Beaufort Gyre (e.g., Manucharyan et al., 2016) and transport cores of anomalous water properties over significant distances. Dynamically, lateral down-gradient eddy fluxes likely balance the time-mean wind-driven convergence of low-salinity surface waters of the Beaufort Gyre. Owing to the small Rossby radius of the Arctic, these vorticies are typically quite small, around 10 km diameter (Zhao et al., 2014), but can have remarkably strong azimuthal velocities (several 10s of cm s^–1) that may modulate vertical mixing in the upper ocean.

Assessment of the eddy inventory in the Canada Basin suggests that there are now more of these features than in the past (Zhao et al., 2016) — a possible consequence of the intensified Beaufort Gyre baroclinicity. On still smaller spatial and temporal scales, submesoscale motions may play a role in upper ocean dynamics in the Arctic (Timmermans et al., 2011). These instability mechanisms act to remove small-scale lateral density gradients in the surface layer, effectively restratifying the upper ocean. When a sea ice cover is present, the surface mixed layer is held close to the local freezing temperature, precluding the development of density compensated temperature–salinity variability often observed in ice-free waters.

Vertical heat, salt, and chemical fluxes in the stratified ocean occur principally through turbulent mixing driven by the breakdown of internal wave energy (e.g., Gregg, 1989; Gregg et al., 2003) and double diffusive convection over depths in the halocline where both salinity and temperature increase with depth (e.g., Padman and Dillon, 1987; Timmermans et al., 2008). The latter is manifested by a staircase-like thermohaline stratification consisting of O(1-m) thick layers separated by sharp gradient regions. The vertical temperature and salinity fluxes through these staircase stratifications, effectively set by the molecular diffusion through the staircase gradient regions, are believed to be very small.

Similarly, away from regions of strong internal tidal generation, the internal wave field in the Arctic and the associated turbulent mixing is significantly weaker than that at lower latitudes. Away from topographic features, energy levels are typically 1–2 orders of magnitude lower than observed in the mid-latitude open ocean (D'Asaro and Morison, 1992; Levine et al., 1985, 1987). The isolation of the ocean from the atmosphere by sea ice is often cited as inhibiting atmospheric momentum transfer into the Arctic Ocean. This is certainly the case in a rigid, compact ice pack because internal ice stress absorbs some of the atmospheric momentum, but in a looser pack ice can enhance the momentum transfer from the atmosphere into the ocean (Martin et al., 2014; McPhee et al., 1987; Morison et al., 1987).

Dissipation in the oscillating boundary layer under the sea ice has been hypothesized as an important sink of internal wave energy in the Arctic Ocean compared to ice-free seas (Morison et al., 1985). It has been suggested that this increased under-ice dissipation causes a much shorter effective energy decay timescale [24 days compared to 100 days in the open ocean based on the Desaubies formulation (Desaubies, 1976) of the Garrett–Munk (GM) internal wave spectra (Garrett and Munk, 1972, 1975)], implying the internal wave field in the Arctic Ocean cannot reach typical saturated GM internal wave energy levels.

Acoustic Doppler Current Profiler (ADCP) observations in the Canada Basin (Pinkel, 2005) suggest that most near-inertial internal waves in the Arctic exist in a one-bounce scenario, i.e., that they are generated at the surface, reflect once off the bottom and then dissipate in the under-ice boundary layer (Morison et al., 1985). Despite historic declines in sea ice thickness, extent, and concentration in the Pacific sector of the Arctic Ocean in the past 10 years, a number of observational studies suggest that Arctic Ocean internal wave energy is still low, and it remains unclear how this might change in the future. Internal wave energy levels and inferred mixing measured in the last 10 years with Expendable Current Profilers (XCP) compared with similar observations from the 1980s show little change in internal wave energy and mixing (Guthrie et al., 2013). Strong near-surface stratification associated with the accumulation of fresh water appear to continue to enhance the dissipation of internal wave energy in the under-ice boundary layer (Guthrie et al., 2013).

Changes to near-surface stratification on weekly to monthly timescales clearly alter the internal wave field (Cole et al., 2014). Comparisons of internal wave shear measured at identical locations in the Beaufort Sea by the ONR Seasonal Ice Zone Reconnaissance Surveys project show virtually no difference between ice-free and ice-covered conditions, with inferred mixing values consistently O(10–6) m2 s–1. However, internal wave vertical displacement estimates from Canada Basin Ice-Tethered Profiler (ITP) data (Dosser and Rainville, 2015) show a seasonal cycle in near-inertial internal waves that is related to both wind strength and ice characteristics. While the 10-year (2005–2014) time series of ITP estimates shows only a minor increasing trend in mean amplitude, the distribution of internal wave amplitude may be changing.

Measurements of mixing under arctic storms (Kawaguchi et al., 2015), and with better time and space resolution (e.g., MIZ and Sea State of the Emerging Arctic Ocean DRIs data), are emerging, which may lead to a better understanding of the integrated impact of internal waves and episodic processes driven by patchy air–ice–ocean fluxes. Unfortunately, it is difficult to capture the fate of near-inertial internal waves in the Arctic Ocean from generation to dissipation, because in the presence of ice cover, concurrent high-resolution fixed point measurements of wind speed, ice drift, and ocean velocities from the surface to depth are difficult to make.

The SODA measurement program builds on advances in autonomous observing from the Marginal Ice Zone DRI to employ a system consisting of four interrelated components (Figure, at right):

Complementary observing elements will sample through diverse atmospheric forcing and ice cover regimes, providing a wide dynamic range to address SODA science questions.