Research Article

Glacial ocean circulation and stratification explained by reduced atmospheric temperature

See allHide authors and affiliations

PNAS January 3, 2017 114 (1) 45-50; first published December 19, 2016; https://doi.org/10.1073/pnas.1610438113

  1. Edited by Mark H. Thiemens, University of California, San Diego, La Jolla, CA, and approved November 7, 2016 (received for review June 27, 2016)

Significance

To understand climatic swings between glacial and interglacial climates we need to explain the observed fluctuations in atmospheric carbon dioxide (CO2), which in turn are most likely driven by changes in the deep ocean circulation. This study presents a model for differences in the deep ocean circulation between glacial and interglacial climates consistent with both our physical understanding and various proxy observations. The results suggest that observed changes in ocean circulation and stratification are caused directly by atmospheric cooling or warming, which has important implications for our interpretation of glacial–interglacial transitions. In particular, the direct link between atmospheric temperature and ocean circulation changes supports the notion of a positive feedback loop between atmospheric temperature, ocean circulation, and atmospheric CO2.

Abstract

Earth’s climate has undergone dramatic shifts between glacial and interglacial time periods, with high-latitude temperature changes on the order of 5–10 °C. These climatic shifts have been associated with major rearrangements in the deep ocean circulation and stratification, which have likely played an important role in the observed atmospheric carbon dioxide swings by affecting the partitioning of carbon between the atmosphere and the ocean. The mechanisms by which the deep ocean circulation changed, however, are still unclear and represent a major challenge to our understanding of glacial climates. This study shows that various inferred changes in the deep ocean circulation and stratification between glacial and interglacial climates can be interpreted as a direct consequence of atmospheric temperature differences. Colder atmospheric temperatures lead to increased sea ice cover and formation rate around Antarctica. The associated enhanced brine rejection leads to a strongly increased deep ocean stratification, consistent with high abyssal salinities inferred for the last glacial maximum. The increased stratification goes together with a weakening and shoaling of the interhemispheric overturning circulation, again consistent with proxy evidence for the last glacial. The shallower interhemispheric overturning circulation makes room for slowly moving water of Antarctic origin, which explains the observed middepth radiocarbon age maximum and may play an important role in ocean carbon storage.

The deep ocean today is ventilated mainly by two water masses. North Atlantic deep water (NADW) is formed in the North Atlantic before flowing southward at a depth of about 2–3 km and eventually rising back up to the surface in the Southern Ocean. Antarctic bottom water (AABW) is formed around Antarctica and spreads northward into the abyssal basins. The AABW that makes it into the Atlantic then slowly upwells into the lower NADW before returning southward and resurfacing again around Antarctica (1, 2). The glacial equivalent of NADW was likely confined to shallower depths, leaving more of the deep Atlantic filled with water masses originating primarily from around Antarctica (3⇓–5). Moreover, the two water masses appear more distinct, with less mixing between them (6).

Multiple studies have pointed toward the potential importance of sea ice and surface buoyancy fluxes around Antarctica in controlling changes in the deep ocean stratification and circulation between the present and Last Glacial Maximum (LGM) (7⇓⇓⇓–11). Specifically, we recently showed that enhanced buoyancy loss around Antarctica is expected to lead to an increase in the abyssal stratification, an upward shift of NADW, and a clearer separation between NADW and southward-flowing AABW (11)—all in agreement with inferences made for differences in circulation and stratification between the present and LGM (3, 6, 12, 13). This gives rise to the hypothesis that strong cooling around Antarctica led to an increased net freezing rate. The resulting net salt flux into the ocean amounts to an increased buoyancy loss, which then led to the observed changes in deep ocean stratification and circulation. This hypothesis is tested in this study.

We test the connection between atmospheric temperature and ocean circulation and stratification changes, using idealized numerical simulations, which allow us to isolate the proposed mechanism. We use a coupled ocean–sea-ice model, with atmospheric temperature, winds, and evaporation–precipitation prescribed as boundary conditions (Materials and Methods). The model uses an idealized continental configuration resembling the Atlantic and Southern Oceans, where the most elemental circulation changes have been inferred (3, 5, 6, 12).

Results

We first focus on the model’s ability to reproduce key features of the modern (interglacial) ocean state and circulation. Fig. 1A shows the sea surface temperature (SST) over the simulated domain, with boundary conditions resembling present-day atmospheric forcing. The chosen forcing gives SSTs in broad agreement with observations, varying from about 28°C in the tropics to the freezing point (∼−2°C) around Antarctica, where sea ice forms (Fig. 1A, white contours). The North Atlantic is somewhat warmer and ice free. Fig. 2 shows zonally averaged depth–latitude sections of potential temperature (Fig. 2A) and salinity (Fig. 2C), which reproduce the basic features observed in the present-day Atlantic. In particular, we see a tongue of salty water at middepth, penetrating southward into the Southern Ocean and leading to a reversal of the salinity stratification in the abyss (2). The associated overturning streamfunction (Fig. 3A)reveals that this salty tongue is associated with NADW that flows southward as part of the Atlantic meridional overturning circulation (AMOC). Below the clockwise AMOC cell lies an anticlockwise abyssal cell, representing the pathway of AABW. The peak overturning transports of +17.5 Sverdrup (SV) and −5.2 SV are roughly consistent with the observed rate of NADW formation and the transport of AABW into the North Atlantic (1, 2).

Fig. 1.
Fig. 1.

(A and B) Sea surface temperature (colors) and sea-ice concentration (white contours), for simulations with boundary conditions representing present-day conditions (A) and LGM conditions with reduced atmospheric temperature (B). Contour interval for sea-ice concentration is 20%.

Fig. 2.
Fig. 2.

(A–D) Zonal mean temperature (A and B) and salinity (C and D) for simulations with boundary conditions representing present-day forcing (A and C) and simulations with reduced atmospheric temperature resembling LGM conditions (B and D). Contour intervals are 1°C in A and B and 0.05 g/kg in C and D. Note that the colorbar ranges have been cropped to focus on the deep ocean.

Fig. 3.
Fig. 3.

(A and B) Meridional overturning streamfunction (colors) and potential density referenced to 2 km depth (black lines), for simulations with boundary conditions representing present-day forcing (A) and LGM conditions with reduced atmospheric temperature (B). The overturning streamfunction is computed from the sum of the Eulerian zonal mean velocity and the parameterized eddy-induced bolus velocity. [Note that the isopycnal overturning transport includes an additional component associated with standing meanders (11), which is not included in Fig. 3. The contribution of standing meanders largely cancels the apparent diapycnal zonal-mean transport in the channel region.]

We now analyze the response of the model’s equilibrium solution to a reduction in atmospheric temperature. Consistent with proxy data for the LGM, the prescribed atmospheric cooling is polar amplified, ranging from 2°C in the tropics to 6°C around Antarctica (14, 15). All other boundary conditions are held fixed. The reduction in atmospheric temperature leads to a reduction in ocean surface temperature and an expansion of sea ice around Antarctica, as well as the appearance of sea ice in the North Atlantic (Fig. 1B). Moreover, the model suggests a cooling and salinification of AABW, leading to a strong salt stratification in the deep ocean, which replaces the reversed salinity stratification observed with present-day forcing (Fig. 2B). The strong salt stratification is consistent with pore-fluid data for the LGM (12). (To compare the absolute salinity values here to those inferred for the LGM, one needs to account for the increased bulk ocean salinity resulting from the eustatic drop in sea level, which is not included in the simulations and would add around 1 g/kg globally.) The AMOC cell weakens and shoals (Fig. 3B), again consistent with inferences for the LGM (3⇓–5). The abyssal cell slightly strengthens and becomes somewhat more confined to the abyss, leading to an increased separation between the two overturning cells, which may have played an important role in the increased ocean carbon storage (6).

In the real ocean the abyssal overturning cell is distributed over multiple basins and currently overlaps in depth and density with the AMOC cell, leading to a continuous exchange of water between the two cells in the Southern Ocean (1, 2). Whereas this interbasin overlap between the present-day overturning cells cannot be modeled in a single-basin model, the simulated contraction of both the AMOC and abyssal cells is likely to be robust (11). In a multibasin configuration, this contraction of both cells is expected to remove or reduce the overlap between the two cells, thus generating a more isolated abyssal water mass. The rearrangement of the overturning circulation would likely resemble that sketched in Ferrari et al. (9), although the mechanism proposed here is somewhat different.

The changes in the deep ocean circulation and stratification here result from an increased buoyancy loss rate around Antarctica, which in turn results primarily from enhanced brine rejection associated with sea-ice formation and export. Sea-ice export is proportional to the product of the ice load (here defined as the time- and zonal-mean mass of sea-ice and snow per unit area) and equatorward transport velocity. Both increase as the atmospheric temperature is reduced, with the dominant role played by differences in the ice load, which (near its maximum) goes up from about 300 kg/m2 in the “present” simulation to about 800 kg/m2 in the “LGM” simulation. The ice export velocity also increases, as sea ice extends farther northward where the westerly winds are stronger. As a result of the larger ice load and export velocity, the peak ice export rate from around Antarctica increases from about 3 ×107 kg/s to about 14 ×107 kg/s.

To compute the effective net buoyancy loss around Antarctica it is important to consider the nonlinearity of the equation of state and in particular the pressure dependence of the thermal expansion coefficient (16). If surface buoyancy fluxes are computed using the surface haline and thermal expansion coefficients, virtually no buoyancy loss around Antarctica is found in the present-day–like simulation. This lack of buoyancy loss would appear to be at odds with the presence of an abyssal cell and the transformation of upwelling circumpolar deep water (CDW) to AABW. The apparent contradiction can be resolved by noting that the density increase associated with the transformation of CDW to AABW is dominated by a cooling and counteracted by a freshening. Whereas cooling has a small effect on surface densities at cold temperatures, the temperature effect is amplified as a water parcel sinks into the deep ocean. We can estimate the effect of heat and salt fluxes on a parcel at depth by computing buoyancy fluxes based on changes in potential densities referenced to 2 km depth (consistent with the potential densities shown in Fig. 3), which yields an integrated surface buoyancy loss rate around Antarctica for the present-day–like simulation of about 4.4 ×103 m4⋅ s−3 (Materials and Methods). In the LGM simulation the integrated surface buoyancy loss rate increases to 2.1×104 m4⋅ s−3.

The larger buoyancy loss rate around Antarctica in the LGM simulation gives rise to the observed changes in deep ocean circulation and stratification. Because surface buoyancy loss around Antarctica has to be balanced by vertical diffusion in the basin, the deep ocean stratification is expected to depend approximately linearly on the buoyancy loss rate (11). This relationship explains the increase in the deep ocean stratification between the present and LGM simulations (although a quantitative comparison of changes in buoyancy loss and stratification is somewhat complicated by the nonlinearity in the equation of state). The increased stratification in the LGM then leads to an upward shift of NADW, consistent with the results of Jansen and Nadeau (11).

Sensitivity Experiments

To test the robustness of the results to additional modifications in the boundary conditions, we consider a number of sensitivity experiments, varying the wind stress and the vertical turbulent diffusivity, as well as the spatial structure of atmospheric temperature change. The results of these simulations are summarized in Table 1 and are briefly discussed in the following.

View this table:
Table 1.

Summary of results from sensitivity experiments

In a seminal paper, Toggweiler et al. (17) proposed an equatorward shift in the latitude of the Southern Hemisphere surface westerlies as a potential mechanism for differences in the ocean circulation between the present and LGM. Observational evidence, however, allows for an equatorward shift of at most about 3∘ (18, 19), which in turn is here found to have negligible impact on the solution (experiment “LGM windN” in Table 1). Proxy observations and climate models instead indicate a slight poleward shift and strengthening of the surface westerlies over the Southern Ocean (18, 19). A simulation incorporating a 3∘ southward shift and 20% strengthening of the Southern Hemisphere westerlies shows a moderate increase in ice export and associated buoyancy loss around Antarctica (experiment “LGM windS” in Table 1). The increased buoyancy loss rate amplifies the differences between the present and LGM simulations discussed above.

The loss of shallow shelf seas during the LGM has likely led to increased tidal energy dissipation in the deep ocean, which in turn may have caused enhanced vertical mixing (20, 21). On the other hand, the increased deep ocean stratification during the LGM may have suppressed vertical mixing, as more turbulent kinetic energy dissipation would be required to mix the more stratified water column (22). In our simulations, which assumed unchanged vertical diffusivities, the implied energy input to mixing below the upper thermocline (300 m) almost doubles between the present and LGM simulations (Table 1, last column).

To test the sensitivity of our results to the rate of vertical mixing, we consider two additional LGM simulations in which the vertical diffusivity is reduced or increased by 50% (“LGM κ −50%” and “LGM κ + 50%”, respectively). Due to the stronger stratification, the implied energy input to mixing below 300 m is still slightly larger than for present-day conditions in LGM κ −50%, and the energy input is almost tripled in LGM κ + 50%. Reduced vertical mixing weakens and shoals the AMOC cell during the LGM and increases the abyssal stratification—thus amplifying the effects of atmospheric cooling. Enhanced vertical mixing instead strengthens and deepens the AMOC cell and reduces the abyssal stratification—thus counteracting the effects of atmospheric cooling. However, even the simulation with enhanced mixing maintains a stronger stratification and weaker and shallower AMOC cell compared with the simulation representing present-day conditions, indicating that the effect of atmospheric cooling remains dominant.

Significant uncertainty also exists in the spatial pattern of atmospheric temperature change between the present and LGM (14, 15). We consider two sensitivity experiments, which represent extreme cases: one where atmospheric cooling is restricted to the Southern Hemisphere (experiment “LGM dTSH” in Table 1) and one where a globally constant cooling is applied (experiment “LGM dTconst”). Both experiments show roughly similar results as the LGM reference case with symmetric polar amplified cooling, as long as the reduction in atmospheric temperature at high southern latitudes remains about the same. This result confirms our interpretation that circulation and stratification changes are controlled primarily by differences in the surface boundary conditions around Antarctica.

The simulations here are highly idealized and, among other things, do not include a seasonal cycle, which may affect even the mean sea-ice growth and export rate around Antarctica. To analyze the effect of seasonality, the present and LGM simulations were repeated using seasonally varying air temperatures, but leaving the annual mean temperatures unchanged (experiments “Present seas” and “LGM seas” in Table 1 and Fig. S1). Even though the idealized representations of surface boundary conditions, mixed layer dynamics, and sea-ice thermodynamics make this model arguably less suited to represent the seasonal cycle, the simulations do exhibit strong seasonality in sea-ice cover around Antarctica (Fig. S1B). Whereas the addition of a seasonal cycle causes some minor changes in the deep ocean circulation and stratification in both the present and LGM simulations, the main results regarding differences in the stratification and circulation between the present and LGM remain unchanged by the inclusion of a seasonal cycle (Fig. S1C).

Fig. S1.
Fig. S1.

Results from simulations with seasonally varying air temperature. (A) Air temperature as a function of latitude for each calendar month (Inset) for present (Top) and LGM (Bottom) simulations. (B) Annual mean SST (colors) and sea-ice line (defined as 15% contour of sea-ice concentration) for July (solid line) and February (dashed line). (C) Overturning streamfunction (colors) and potential density (referenced to 2 km, black contours).

To address the potential implications of our results for long-term future climate change, we finally examine a “global warming” simulation (experiment “Warm” in Table 1), which shows essentially reversed results from the cooling experiments: Above-freezing temperatures around Antarctica lead to ice-free conditions and net buoyancy gain. As a result AABW formation is shut down, leaving the entire deep ocean filled with nearly unstratified water of North Atlantic origin (Fig. S2). The effect of these potential rearrangements on ocean carbon storage represents an important topic for future research.

Fig. S2.
Fig. S2.

(A–C) Zonal-mean potential temperature (A), salinity (B), and overturning streamfunction (C) in the warm simulation. Contour intervals in A and B are the same as in Fig. 2, although the color scale range differs. Black contours in C show potential density referenced to 2 km depth. Note that the anticlockwise (negative) overturning streamfunction near the channel–basin interface is limited to a virtually unstratified region and is driven entirely by the GM eddy parameterization, which is poorly defined in the limit of very weak stratification.

Discussion and Conclusions

The results of this study suggest that atmospheric cooling alone—via a modification of the buoyancy loss rate around Antarctica—leads to a strong increase in the deep ocean stratification, a shoaling of NADW, and an increased separation between NADW and the underlying abyssal overturning cell. The simulated response of the deep ocean circulation and stratification to atmospheric temperature change is consistent with differences between the present and LGM inferred from paleo-proxy observations (3, 6, 12, 13). A series of sensitivity experiments suggests that the dominant control over circulation and stratification changes is exerted by the atmospheric temperature around Antarctica. Temperature changes in other regions, as well as differences in the atmospheric wind stress, instead play only a relatively minor role.

The results here are broadly consistent with some complex coupled climate model simulations (7, 8), as well as with global ocean-only simulations under LGM boundary conditions (10). Coupled LGM climate simulations using the Community Climate System Model (CCSM)1.4 and CCSM3 show a strongly increased abyssal stratification and shallower NADW (7, 23, 24). Consistent with the results discussed here, these changes have been attributed to enhanced sea-ice export around Antarctica (7) and ultimately reduced CO2 concentrations (8). However, whereas Shin et al. (7) argue that enhanced sea-ice export is caused “ultimately by increased westerlies,” (ref. 7, p. 1) we here suggest that cooling alone is sufficient and likely to be the dominant driver. Changes in freshwater fluxes around Antarctica have also been argued to explain the high deep ocean stratification in a decoupled Community Earth System Model (CESM)1.1.2 ocean simulation under LGM boundary conditions (10). Many other coupled climate models, however, show different and widely diverging changes in the deep ocean circulation and stratification between preindustrial and LGM simulations (25, 26). Possible reasons for this disagreement include differences between transient and fully equilibrated solutions (27, 28), insufficient increase in Antarctic sea-ice cover and formation (29), and/or compensating effects due to other differences in the boundary conditions, such as changes in the ice sheets (26, 30). A more detailed analysis of comprehensive LGM climate simulations is needed to better understand differences between them, but is beyond the scope of this study.

The dominant role of atmospheric temperature change as the driver of differences in the ocean circulation and stratification has important implications for our understanding of glacial–interglacial climate swings. Although the exact mechanisms are still debated, it is likely that changes in the ocean circulation and stratification have played a key role in modulating ocean carbon storage and thus atmospheric CO2 concentrations (31, 32). If ocean circulation changes are themselves directly driven by temperature swings, as argued here, they are likely to play a crucial role in a positive feedback loop between global temperature, ocean circulation, and atmospheric CO2. A better comprehension of this feedback loop will be central to our understanding of past and future climatic changes.

Materials and Methods

Numerical Model Configuration.

The numerical simulations use the Massachusetts Institute of Technology (MIT) general circulation model (MITgcm) (33), in a hydrostatic Boussinesq configuration. The idealized domain extends from 70°S to 65°N, covers 72° in longitude, and is 4 km deep. The horizontal resolution is 1∘×1∘ and the vertical resolution ranges from 20 m near the surface to 200 m in the deep ocean, with a total of 29 levels. The ocean is bounded by a 1∘ strip of land on all sides, which is interrupted only between 69°S and 48°S, where, above 3 km depth, zonally periodic boundary conditions give rise to a reentrant channel, representing the Southern Ocean.

Atmospheric temperatures, net freshwater flux (i.e., precipitation–evaporation), and surface winds are prescribed in the form of idealized analytical functions shown in Fig. S3. Except in the sensitivity simulations with seasonal cycle in air temperatures, all atmospheric forcing fields are constant in time. Momentum and sensible heat exchange between the atmosphere and the ocean are described using standard bulk formulas (34, 35). In addition to the sensible heat flux, an idealized radiative restoring is prescribed as Frad=σ(Ts4−Ta4), where Ts is the ocean/ice surface temperature, Ta is the prescribed atmospheric temperature, and σ=5.67×10−8 W⋅m−2⋅K−4 is the Stefan–Boltzmann constant. In reality the ocean experiences a net radiative heating, which is balanced primarily by latent heat loss. Both are ignored here, which in effect simplifies the thermal boundary conditions to a restoring to the prescribed atmospheric temperature—albeit with a restoring rate that is modified by wind speed and boundary layer stability. This idealization was chosen to avoid additional assumptions about changes in downward short- and long-wave radiation and boundary layer-specific humidity. Moreover, use of a bulk formula for moisture and latent heat transfer would require artificial flux adjustments to close the global salt budget.

Fig. S3.
Fig. S3.

(A) Surface air temperature as a function of latitude for the simulation with boundary conditions resembling present-day conditions (Present, solid red), for the simulations with reduced temperatures representing LGM conditions (LGM, solid blue), for the simulation with globally homogeneous 5° C cooling (LGM dTconst, dotted blue), for the simulation with only Southern Hemisphere cooling (LGM dTSH, dashed blue), and for the global warming simulation (Warm, dashed magenta). All temperature distributions are zonally symmetric. (B) Zonal–mean zonal (blue) and meridional (red) surface wind as a function of latitude. The dashed and dotted lines show the zonal wind for the simulations LGM windS and LGM windN, respectively. All other simulations use the zonal wind shown by the solid line, and all simulations use the same meridional wind. The zonal wind stress is zonally symmetric, but the meridional wind varies zonally as u(x,y)=umax(y)sin(πx/L), where x=0..L. The meridional wind component simulates katabatic winds from the Antarctic continent. The presence of katabatic winds helps to prevent buildup of unrealistically thick ice along the shore and to maintain a realistic abyssal stratification in the present-day–like simulation, but does not affect the main results for the effect of cooling on deep ocean stratification and circulation. The zonal variation in katabatic winds was later found to have no significant effect. (C) Net evaporation–precipitation as a function of latitude (zonally symmetric and identical for all simulations). At latitudes where the air temperature is below freezing, net precipitation falls as snow.

To test the sensitivity of the results to the details of the thermal boundary conditions, additional simulations were performed with a simple linear restoring of the ice/ocean surface temperature (36) (Fig. S4). The surface heat flux is computed as F=25 W/(m2K)(Ts−Ta). This amounts to a somewhat faster effective restoring than what is on average implied by the original boundary conditions. To still obtain a realistic present-day–like simulation the meridional air temperature gradient was slightly reduced, decreasing the air temperature at the equator by 1°C while increasing the temperature at the highest latitudes by 2.5°C (which keeps the global mean air temperature approximately unchanged). The air temperature difference between the present and LGM simulations, however, is identical to that in the simulations discussed in this article. The results suggest that the main conclusions are not sensitive to the details of the thermal boundary conditions (Fig. S4).

Fig. S4.
Fig. S4.

(A and B) The same as (A) Fig. 2 and (B) Fig. 3 in the main text, but for simulations with linear restoring of surface temperature (see Materials and Methods for simulation details). In A, the subpanels show zonal mean temperature (a and b) and salinity (c and d), for simulations with boundary conditions representing present-day forcing (a and c) and simulations with reduced atmospheric temperature resembling LGM conditions (b and d). In B, the subpanels show results for simulations with boundary conditions representing present-day forcing (a) and LGM conditions with reduced atmospheric temperature (b). All colorbars and contour intervals are chosen identically to the corresponding figures in the main manuscript.

Sea-ice dynamics and thermodynamics are described using the MITgcm sea-ice package, which is based on the viscous-plastic rheology introduced by Hibler (37) and modified by Zhang and Hibler (38) and Losch et al. (39). Sea-ice thermodynamics are based on Zhang et al. (40) and use a zero-heat–capacity approximation, but account for the different heat conductivities of ice and snow cover. Mesoscale eddy fluxes are parameterized using the Gent and McWilliams (GM) (41) and Redi (42) parameterizations, with a variable eddy diffusivity formulated following Visbeck et al. (43), with α=0.01 and mixing length l=160 km. The eddy diffusivity is capped between Kmin=200 m2⋅s−1 and Kmax=2,000 m2⋅s−1, and the GM parameterization is tapered in the presence of steep isopycnal slopes, following Gerdes et al. (44). Diapycnal mixing is represented by a vertical diffusivity, which is strongly enhanced in the abyss, but reduces to a background value of 2×10−5 m2⋅s−1 in the thermocline (45, 46). The diffusivity profile is shown in Fig. S5. Convection is parameterized using a diffusive adjustment with a convective diffusivity κconv=10 m2⋅s−1 whenever the stratification is statically unstable. Dissipation of momentum in the bottom boundary layer is represented by a linear bottom drag, with a piston velocity of 1 mm⋅s−1. No-slip horizontal boundary conditions are applied at the walls, and grid-scale noise is suppressed using Laplacian and biharmonic viscosities, with viscosity coefficients ν2=2×104 m2⋅s−1 and ν4=2×1013 m4⋅s−1, respectively.

Fig. S5.
Fig. S5.

Vertical diffusivity as a function of depth. The dashed line shows the profile used for LGM κ−50% and the dotted line shows the profile for LGM κ+50%; all other simulations use the reference profile shown in the solid line. Note the logarithmic abscissa. The reference profile is based on the global-mean internal wave-driven mixing profiles computed by Ferrari and Nikurashin (45), with an added background diffusivity of 2×10−5 m2/s.

All simulations are integrated to equilibrium for at least 4,000 y, and averages are taken over the last 500 y of each simulation. Whereas the simulation with present-day forcing reaches a steady equilibrium solution, the cold LGM simulation exhibits some decadal to centennial variability in the AMOC. The exact nature of this variability, which appears to be associated with sea ice in the North Atlantic, is beyond the scope of this study, but may be addressed in follow-up work. The magnitude of the variability is small compared with the differences obtained between the present and LGM simulations and is averaged out in Figs. 1–3. Nevertheless, the lack of a steady-state solution puts into question the use of an accelerated tracer time stepping (47), which has been used to speed up the equilibration. To test the impact of the accelerated tracer time step, the LGM simulation was initialized from the statistical equilibrium solution obtained with tracer acceleration and integrated for another 1,000 y without tracer acceleration. Whereas properties of the internal variability appear to be affected by the tracer acceleration, the mean state and circulation discussed in this paper are virtually unaffected. The sensitivity to tracer acceleration was also tested explicitly in the present seas simulation, which includes a seasonal cycle in the thermal forcing. Again, no significant effect on the deep ocean circulation and stratification (or on the seasonality of sea ice) was found. Due to the negligible effect of tracer acceleration on the mean state and circulation, all sensitivity experiments discussed in this study use tracer acceleration for the entire integration.

Computation of Buoyancy Loss Rates.

Buoyancy loss from the ocean is computed based on the net heat and freshwater fluxes asB=gαQρ0−1Cp−1+gβFSρ0−1,[1]where Q is the net heat flux and F the net freshwater flux out of the ocean (which include the effects of sea-ice formation and melt), α is the thermal expansion coefficient, β is the haline contraction coefficient, g is the gravitational acceleration, ρ0=1,035 kg⋅m−3 is a reference density, S is the salinity, and Cp is the heat capacity of seawater. Because we are interested in the effect of surface fluxes on the density of a parcel at depth, the thermal expansion and haline contraction coefficients are computed for an ambient pressure of 2,000 dbar.

B is generally positive (denoting ocean buoyancy loss) within a strip around Antarctica, but negative farther northward in the circumpolar channel. The “total buoyancy loss rate around Antarctica” is computed by integrating B over the entire region of buoyancy loss around Antarctica, defined to include every gridpoint south of 55°S where B>0 (because the areas of buoyancy loss and gain are well separated, the results are not sensitive to the exact choice of this latitude).

Acknowledgments

I thank Alice Marzocchi, K. Thomas, and two anonymous reviewers for their very valuable comments. The MITgcm configuration files and model output data are available from the author upon request. Funding for this work was provided by the National Science Foundation under Award 1536454, and computational resources were provided by the Research Computing Center at the University of Chicago.

Footnotes

  • ↵1Email: mfj{at}uchicago.edu.
Back to top
Article Alerts
Email Article
Citation Tools
Request Permissions
Glacial ocean circulation and stratification
Malte F. Jansen
Proceedings of the National Academy of Sciences Jan 2017, 114 (1) 45-50; DOI: 10.1073/pnas.1610438113
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Article Classifications

Proceedings of the National Academy of Sciences: 114 (1)
Table of Contents

Submit

Jump to section

You May Also be Interested in

Permafrost.
Permafrost methane and heat waves
Atmospheric methane levels increased over geological formations in Northern Siberia following a heat wave in 2020, suggesting that permafrost thaw releases methane from reservoirs.
Image credit: Wikimedia Commons.
Granular forces (black lines) at the same location in the soil before (left) and after (right) ant tunneling.
3D modeling of ant tunnels
Mechanics of ant tunneling could inspire the development of improved robotic mining techniques.
Image credit: Jose E. Andrade and David R. Miller (California Institute of Technology, Pasadena, CA).
Image from Hubble Telescope of white dwarf start Sirius A and its smaller companion Sirius B.
Inner Workings: Astronomers uncover new way to measure the speed of stars
New observations now allow the seemingly impossible: measuring the speeds of some stars along the line of sight without ascertaining their Doppler shifts.
Image credit: NASA, H.E. Bond, and E. Nelan (Space Telescope Science Institute, Baltimore, MD); M. Barstow and M. Burleigh (University of Leicester, UK); and J.B. Holberg (University of Arizona, Tucson, AZ).

Similar Articles