Reports

Oceanic Mass Transport by Mesoscale

sent work, the outmost closed PV con-
tours on isopycnals are used as the cri-
terion to determine the boundary of

Eddies fluid trapped by eddies.

Regarding the second challenge of
capturing the three-dimensional struc-
tures of mesoscale eddies, our recent
Zhengguang Zhang,1 Wei Wang,1* Bo Qiu2 study has demonstrated that they can be
1
Physical Oceanography Lab, Qingdao Collaborative Innovation Center of Marine Science and Technology, reconstructed by combining the high-
Ocean University of China, Qingdao, P. R. China. 2Department of Oceanography, University of Hawaii at resolution satellite altimeter data and
Manoa, Honolulu, HI, USA concurrent Argo profiling float temper-
*Corresponding author. Physical Oceanography Lab, 5 Yushan Road, Qingdao 266003, P. R. China. E- ature/salinity data (14). Once the three-
mail: wei@ouc.edu.cn dimensional potential density (density
hereafter) field is reconstructed from
Oceanic transports of heat, salt, and fresh water, dissolved CO2 and other tracers the altimeter sea surface height (SSH)
regulate global climate change and distribution of natural marine resources. While measurements, we are able to evaluate
the time-mean ocean circulation transports fluid as a conveyor belt, fluid parcels the eddy-perturbed PV field and finally
can also be trapped and transported discretely by migrating mesoscale eddies. By quantify the fluid volume trapped by
combining available satellite altimetry and Argo profiling float data, we show that the moving eddies (15). To verify our

Downloaded from www.sciencemag.org on June 26, 2014

the eddy-induced zonal mass transport can reach a total meridionally integrated method of reconstructing the eddy den-
value up to 30-40 Sv (1 Sv = 106 m3 s1), and it occurs mainly in subtropical regions sity field, we use independent mooring
where the background flows are weak. This transport is comparable in magnitude observations from different regions of
with that of the large-scale, wind- and thermohaline-driven circulation. the world oceans: the Labrador Sea, the
Arabian Sea and the Kuroshio Exten-
sion (Fig. 1). The reconstructed eddy
Heat and material transports in oceans play a dominant role in regulating density structures are found to compare favorably with the mooring re-
Earths climate and in controlling the oceanic absorption of greenhouse sults for both the warm- and cold-core eddies. The relative errors of the
gases responsible for the global warming. Large-scale wind- and ther- reconstructed density displacement are estimated to be 20% of the vari-
mohaline-driven ocean circulation has traditionally been regarded to ance (15), an uncertainty level acceptable for estimating the fluid volume
constitute the major part of the oceanic transport, and its spatio-temporal trapped by mesoscale eddies.
variations to have profound climate and biogeochemical impacts. How- For illustration, three isopycnal surfaces of an observed warm-core
ever, by tracking water masses (1) and radiocarbons (2), observational eddy in the subtropical North Pacific are presented in Fig. 2A. The iso-
studies have revealed that the large-scale circulation alone cannot ex- pycnal surfaces of this warm-core eddy exhibit a concave shape. The
plain the oceanic transports. Since the detection of their ubiquitous pres- upper isopycnal surface outcrops at the sea-surface and envelops a bulk
ence in the oceans more than three decades ago, oceanic mesoscale of warm water. All PV contours on this isopycnal are closed and the
eddies with length scales of O(50-300 km) have been recognized as key boundary of the trapped water coincides with the outcrop line. On the
contributors in transporting heat, dissolved carbon, and other biogeo- middle isopycnal of the eddy (at ~500 m depth), the PV contours near
chemical tracers (36). the eddy core are closed whereas those outside the eddy core are open; in
Unlike the large-scale circulation that transports fluid and their prop- this case, the eddy only traps the fluid within the outermost closed PV
erties continuously, mesoscale eddies can trap fluid parcels within the contour. As the eddy perturbation decays with depth, the area enclosed
eddy core and transport them discretely (7, 8). Our theoretical under- by the closed PV contours diminished on denser isopycnals, and the
standing of the processes involving mesoscale eddies remains, however, eddy loses its ability to trap water in these deep layers. The outmost
incomplete. For example, many existing studies have treated transport by closed PV contour on each isopycnal forms a three-dimensional conic
eddies as a simple diffusion process that tends to obscure the redistribu- surface (the transparent black surface in Fig. 2A), and it is this surface
tion processes of dynamical and biogeochemical tracers (9). In addition that delineates the boundary of the trapped fluid by this warm-core eddy.
to the theoretical understanding, it is equally important to quantify the For comparison, a cold-core eddy detected in the same geographical
eddy-induced fluid volume transport based on available observational region is presented in Fig. 2B, and the fluid trapped by this cold-core
data. There exist, however, two major challenges for such quantification: eddy exhibits a somewhat different vertical enveloping shape.
First, there is no widely accepted criterion that defines the volume of From the sequential SSH maps from the satellite altimeter measure-
fluid trapped by an eddy and, secondly, accurate vertical structures of the ments, we evaluate the propagation speed of individual eddies by adopt-
mesoscale eddies are often observationally inaccessible. ing a Lagrangian tracking method (16, 17). With the eddy propagation
In defining the fluid trapped by eddies, all existing criteria are kine- speed and its volume of trapped fluid at hand, the eddy-induced mass
matic in nature; they are inherently descriptive and depend on the choice transport is readily quantifiable (15). Figure 3A shows the global distri-
of reference frame (68, 10). From a dynamical point of view, move- bution of the eddy-induced zonal transport based on the altimeter SSH
ment of a rotating fluid can be depicted by a dynamically conserved data of 19922010. A remarkable feature is that the regions of large
quantity: potential vorticity (PV), its tendency to spin. PV contours have westward eddy transport are located in all latitude bands between 20 to
long been used to identify the trajectory of fluid particles in the ocean 40. Also noteworthy is that an enhanced eddy-induced eastward
(11), and fluid tends to be trapped within the closed PV contours (12). A transport occurs along the Antarctic Circumpolar Current (ACC) path
classical demonstration of this constraint is the Taylor Column in the south of 40S. When integrated over the entire latitude range from 80S
homogeneous fluid rotating tank experiment (13). For an adiabatic strati- to 80N, the total zonal eddy-induced transports add up to 3040 Sv
fied ocean, fluid motion is constrained on isopycnal surfaces. If a PV westward and 510 Sv eastward (see Fig. 3B). The global distribution of
contour on an isopycnal is closed due to eddy perturbations, the fluid eddy-induced meridional transport is shown in Fig. 4A. Geographically,
trapped inside this closed contour will move with the eddy. In the pre- the meridional eddy transport has a poleward tendency within the tropi-

/ http://www.sciencemag.org/content/early/recent / 26 June 2014 / Page 1 / 10.1126/science.1252418

cal regions (20S-20N) and an equator-ward tendency in the subtropical
regions (20N-60N, 20S-60S). Interestingly, these meridional eddy
transport in many parts of the world oceans moves in the same direction
as the meridional wind-forced Ekman transport. The convergent latitudes
for the meridional eddy transport, however, appear equatorward of the
convergent latitudes for the meridional Ekman transport. Due to the
predominance of westward propagations by the mesoscale eddies (17),
the zonally integrated eddy-induced meridional transport is of the order
of several Sv, which is much smaller than the zonal eddy-induced mass
transport (Fig. 4B). The relative errors in these transport values are esti-
mated at about 20% (15).
As comparison, the transports of the wind-driven gyres in the Pacif-
ic, Atlantic and Indian oceans are 20-60 Sv, 20-50 Sv and 10-25 Sv,
respectively, and the transports of the thermohaline circulation in these
three oceans are 10-20 Sv, 9-20 Sv and 10-15 Sv, respectively (18, 19).
The 3040 Sv value of the eddy-induced zonal transport is comparable
in magnitude with the large-scale wind- and thermohaline-driven circu-
lations. Apart from the Southern Ocean, these zonal transports work to
accumulate water onto the western side of the ocean basins, implying
that the mesoscale eddies can exert a strong impact upon the tranports of
the western boundary currents. Indeed, observations of the Kuroshio
transport east of Taiwan reveal that its interannal variation has no obvi-
ous correlation with the large-scale wind fluctuations, but exhibits a
good correspondence with the eddy kinetic energy level east of the Tai-
wan island (20). As indicated in Fig. 3A, the region east of Taiwan is the
strongest hot-spot of eddy-induced tranport in the North Pacific Ocean.
With the amount of mesoscale eddies modulating on the interannual and
decadal time scales (21), the eddy-induced zonal transport is likely to
fluctuate as well. By altering the poleward mass and heat transport car-
ried by the western boundary currents, the eddy-induced transport fluc-
tuations can contribute significantly to the regional and basin-scale
climate variability (2224).
It is worth noting that there exist no strong currents in many regions
with large westward eddy-induced transport identified in Fig. 3A. Equal-
ly important, the eddy-induced transports can penetrate to the depth be-
low the main thermocline, where the wind-driven circulation is generally
weak. In these regions and depth ranges without strong background cir-
culation, the mesoscale eddies can dominate the transport of heat, fresh
water, dissolved carbons and other biogeochemical tracers, which are
important to understand the long-term climate change.
In climate models that simulate the global warming scenarios, the
transport by oceanic mesoscale eddies is generally underestimated
and/or improperly resolved due to their coarse grid resolution. As the
model resolution increases with enhanced computing power, mesoscale
eddies will be resolved progressively by future climate models, and the
impact of the eddy-induced transport will likely emerge. This makes the
task of establishing proper theories about the eddy-induced transport
particularly important and urgent.
References and Notes
1. G. Danabasoglu, J. C. McWilliams, P. R. Gent, The role of mesoscale tracer
transports in the global ocean circulation. Science 264, 11231126 (1994).
Medline doi:10.1126/science.264.5162.1123
2. S. Peacock, Debate over the ocean bomb radiocarbon sink: Closing the gap.
Global Biogeochem. Cycles 18, GB2022 (2004). doi:10.1029/2003GB002211
3. H. L. Bryden, E. C. Brady, Eddy momentum and heat fluxes and their effect on
the circulation of the equatorial Pacific Ocean. J. Mar. Res. 47, 5579 (1989).
doi:10.1357/002224089785076389
4. D. B. Chelton, P. Gaube, M. G. Schlax, J. J. Early, R. M. Samelson, The
influence of nonlinear mesoscale eddies on near-surface oceanic chlorophyll.
Science 334, 328332 (2011). Medline doi:10.1126/science.1208897
5. D. J. McGillicuddy Jr., L. A. Anderson, N. R. Bates, T. Bibby, K. O. Buesseler,
C. A. Carlson, C. S. Davis, C. Ewart, P. G. Falkowski, S. A. Goldthwait, D.
A. Hansell, W. J. Jenkins, R. Johnson, V. K. Kosnyrev, J. R. Ledwell, Q. P.
/ http://www.sciencemag.org/content/early/recent / 26 June 2014 / Page 2 / 10.1126/science.1252418
Li, D. A. Siegel, D. K. Steinberg, Eddy/wind interactions stimulate
extraordinary mid-ocean plankton blooms. Science 316, 10211026 (2007).
Medline doi:10.1126/science.1136256
6. C. Dong, J. C. McWilliams, Y. Liu, D. Chen, Global heat and salt transports by
eddy movement. Nat. Commun. 5, 3294 (2014). 10.1038/ncomms4294
Medline doi:10.1038/ncomms4294
7. G. R. Flierl, Particle motions in large-amplitude wave fields. Geophys.
Astrophys. Fluid Dyn. 18, 3974 (1981). doi:10.1080/03091928108208773
8. J. J. Early, R. M. Samelson, D. B. Chelton, The evolution and propagation of
quasi-geostrophic ocean eddies. J. Phys. Oceanogr. 41, 15351555 (2011).
doi:10.1175/2011JPO4601.1
9. M. M. Lee, D. P. Marshall, R. G. Williams, On the eddy transfer of tracers:
Advective or diffusive? J. Mar. Res. 55, 483505 (1997).
doi:10.1357/0022240973224346
10. F. J. Beron-Vera, Y. Wang, M. J. Olascoaga, G. J. Goni, G. Haller, Objective
detection of oceanic eddies and the Agulhas leakage. J. Phys. Oceanogr. 43,
14261438 (2013). doi:10.1175/JPO-D-12-0171.1
11. J. R. Luyten, J. Pedlosky, H. M. Stommel, The ventilated thermocline. J.
Phys. Oceanogr. 13, 292309 (1983). doi:10.1175/1520-
0485(1983)013<0292:TVT>2.0.CO;2
12. P. B. Rhines, W. R. Young, Homogenization of potential vorticity in planetary
gyres. J. Fluid Mech. 122, 347367 (1982). doi:10.1017/S0022112082002250
13. G. I. Taylor, Motions of solids in fluids when the flow is not irrotational.
Proc. R. Soc. London A Contain. Pap. Math. Phys. Character 93, 99113
(1917). doi:10.1098/rspa.1917.0007
14. Z. G. Zhang, Y. Zhang, W. Wang, R. X. Huang, Universal structure of
mesoscale eddies in the ocean. Geophys. Res. Lett. 40, 36773681 (2013).
doi:10.1002/grl.50736
15. Information on the observational data and analysis methods is available in the
supplementary materials on Science Online.
16. D. B. Chelton, M. G. Schlax, R. M. Samelson, R. A. de Szoeke, Global
observations of large oceanic eddies. Geophys. Res. Lett. 34, L15606 (2007).
10.1029/2007GL030812 doi:10.1029/2007GL030812
17. D. B. Chelton, M. G. Schlax, R. M. Samelson, Global observations of
nonlinear mesoscale eddies. Prog. Oceanogr. 91, 167216 (2011).
doi:10.1016/j.pocean.2011.01.002
18. W. J. Schmitz Jr., On the World Ocean circulation: Volume I: Some global
features/North Atlantic circulation. Tech Rep. WHOI-96-03. Woods Hole
Oceanographic Institute (1996).
19. W. J. Schmitz Jr., On the World Ocean circulation. Volume II: The Pacific
and Indian Oceans/A global update. Tech Rep. WHOI-96-08. Woods Hole
Oceanographic Institute (1996).
20. Y. L. Chang, L. Y. Oey, Interannual and seasonal variations of Kuroshio
transport east of Taiwan inferred from 29 years of tide-gauge data. Geophys.
Res. Lett. 38, L08603 (2011). 10.1029/2011GL047062
doi:10.1029/2011GL047062
21. B. Qiu, S. Chen, Concurrent decadal mesoscale eddy modulations in the
western North Pacific subtropical gyre. J. Phys. Oceanogr. 43, 344358
(2013). doi:10.1175/JPO-D-12-0133.1
22. S. Minobe, A. Kuwano-Yoshida, N. Komori, S. P. Xie, R. J. Small, Influence
of the Gulf Stream on the troposphere. Nature 452, 206209 (2008). Medline
doi:10.1038/nature06690
23. B. Qiu, Interannual variability of the Kuroshio Extension system and its
impact on the wintertime SST field. J. Phys. Oceanogr. 30, 14861502
(2000). doi:10.1175/1520-0485(2000)030<1486:IVOTKE>2.0.CO;2
24. B. Qiu, Kuroshio Extension variability and forcing of the Pacific decadal
oscillations: Responses and potential feedback. J. Phys. Oceanogr. 33, 2465
2482 (2003). doi:10.1175/2459.1
25. M. F. de Jong, A. S. Bower, H. H. Furey, Two years of observations of warm-
core anticyclones in the Labrador Sea and their seasonal cycle in heat and salt
stratification. J. Phys. Oceanogr. 44, 427444 (2014). doi:10.1175/JPO-D-13-
070.1
ACKNOWLEDGMENTS
We thank the reviewers for their constructive and detailed comments. This
research was supported by the National Basic Research Priorities Program of
China through grant 2013CB430303, the National Natural Science
Foundation of China under grant 41276014, and the National Global Change
Major Research Project of China through grant 2013CB956201. The sources
of the data used in this paper are contained in the supplementary materials.
Supplementary Materials
www.sciencemag.org/cgi/content/full/science.1252418/DC1
Materials and Methods
Figs. S1 to S4
References

19 February 2014; accepted 16 June 2014

Published online 26 June 2014; 10.1126/science.1252418

/ http://www.sciencemag.org/content/early/recent / 26 June 2014 / Page 3 / 10.1126/science.1252418

Fig. 1. Comparison between observed and reconstructed eddy density fields. (A)
Potential density section across a warm-core eddy at 60.6N, 52.4W in the Labrador
Sea on 11-Oct-2008. Horizontal ordinate is the distance to the eddy center, and
vertical ordinate is depth. Colored shades denote the reconstructed density field, with
3 3 3
white curves represent the isopycnal 27.65 kg m , 27.70 kg m and 27.75 kg m .
The corresponding isopycnal surfaces from the mooring observations are presented
by black curves (25). (B) Potential density section across a cold-core eddy at 15.5N
61.5E in the Arabian Sea on 30-Nov-1994. White and black contours denote the
3 3 3
isopycnal 24.5 kg m , 25.5 kg m and 26.0 kg m . (C) Comparison of the
reconstructed isopycnal displacement with the observations at the eddy centers. Blue
3
points represent the 27.7 kg m isopycnal displacement of 12 eddies observed in the
3
Labrador Sea. Black squares represent the 26.0 kg m isopycnal displacement of
3
three eddies observed in the Arabian Sea. Red triangles represent the 27.4 kg m
isopycnal displacement of two eddies observed in the Kuroshio Extension (15).

/ http://www.sciencemag.org/content/early/recent / 26 June 2014 / Page 4 / 10.1126/science.1252418

Fig. 2. Three-dimensional structures of trapped fluid by mesoscale eddies and
potential vorticity (PV) distributions on representative isopycnals. (A) PV contours on
three isopycnal surfaces of a warm-core eddy identified by the altimetry data at
30.1N, 150.1E in the subtropical North Pacific on May-25-2005. Sea level anomaly
at the eddy center is 12 cm. The shape and depth of isopycnals are represented by
the three-dimensional surfaces. The potential density of the upper, middle and lower
3 3 3
isopycnals is 23.35 kg m , 26.59 kg m and 27.48 kg m , respectively. PV
distribution on the isopycnal is depicted by colored contours. The transparent black
surface, defined by the outmost closed PV contours, signifies the boundary of
trapped fluid by this warm-core eddy. (B) Same as (A) except for a cold-core eddy
identified by the altimetry data at 30.7N, 151.0E in the subtropical North Pacific on
May-25-2005. Sea level anomaly at the eddy center is -14 cm. The depicted
3 3 3
isopycnal surfaces are 23.99 kg m , 26.75 kg m and 27.56 kg m , respectively.

/ http://www.sciencemag.org/content/early/recent / 26 June 2014 / Page 5 / 10.1126/science.1252418

Fig. 3. Global distribution of the zonal transport of trapped fluid by mesoscale eddies.
(A) Eddy-induced zonal transport through meridional cross-section per degree of
latitude. Based on the statistical average of eddy propagation speed and trapped fluid
volume (15). (B) Distribuion of the total meridionally-integrated zonal transport induced
by eddies as a function of longitude. The solid blue/red curves represent the
westward/eastward tranports computed by integrating the westward transport in (A)
from 80S to 80N. The dash blue/red curves represent integrated total tranport
north/south of 40S, respectively.

/ http://www.sciencemag.org/content/early/recent / 26 June 2014 / Page 6 / 10.1126/science.1252418

Fig. 4. Global distribution of the meridional transport of trapped fluid by mesoscale
eddies. (A) Eddy-induced meridional transport through zonal cross-section per degree
of longitude. (B) Distribuion of the total zonal-integrated meridional transport induced
by eddies as a function of latitude.

/ http://www.sciencemag.org/content/early/recent / 26 June 2014 / Page 7 / 10.1126/science.1252418