REPORTS ON RESEARCH AT THE WOODS HOLE OCEANOGRAPHIC INSTITUTION

Volume 37, Number 1 • Spring 1994

Atlantic Ocean

* .,' I •

Circulation

Oceanus

REPORTS ON RESEARCH AT THE WOODS HOLE OCEANOGRAPHIC INSTITUTION

t >i i nuii'< is published semi-annually by the Woods Hole
Oceanographic Institution. Woods Hole. MA (C~i4:i
508-457-2 i.cxi 271H

iii-ciniiix and its logo are ® Registered Trademarks ol the
Woods Hole Oceanographio Institution, All Rights Reserved

In North America, an Oceanns Package. including<Vivj//».s
magazine, Woods Hole Currents la l^-pagc tabloid-size
neusletiei ) .mil 1 1,; nu l-:.i/ilnr, i ( an 8-page junior-high-selnm!
educational publu :iin.n ) IN axailable for a sj.'i in In- Single

I ilples lit Drvr/HI/.v rust M 1)11 pills S| Illl shipping Outside \iirth

America, the annual fee fur fii-fniiin, inagaz uK is S:;ii. su

for single copies. To receive the publications, iili'ascfall (loll
free) 1-M»! s^.Vnni.l m HI ate: WHOI Oceanus Package Snrvice
Center, P.O. Box Wlii. Syracuse, NY 1:^17

riirrksslinulil hcilraui a I'S hank in US dollars and

mailr p,i\ahlr I" \\uinls Hole Oceaiioguplnc Institution.

When .sending rhange of address, please inrludr mailing label
( 'la mis lor missing numbers from ihe I'S will be honored within
Ihree months Ol publir'.il ovei -IMS, fl\e innntlis

I'eriin.ssioii lo phoidciipv for mlrriial or personal use or tin-
internal in personal use of specific clients is granted In <>,,'uiiu\

In hlil. Illi' . and othel llsns M-Hl,li'lril Hltll I lie I '"pM'mlll
I'lralanrr Cfiiln (('('(' I, prn\ idcd that tile base fee nf S^ p,-i rMp\
ul I be arlicle is paid directly to CCC, 21 Congress St i rcl, Salem,
M \ iilii^n Special requests should be addressed to Ckwiv//s

ISSN 0029 3182 - !$2.00* n.iGSTR l^s :!72 IHi7 Printed in
SAor in \'iU-d paper

Vol. 37, No. 1 • Spring 1994 • ISSN 0029-8182

Atlantic Ocean Circulation

The Atlantic Ocean 1

Progress in Describing and Interpreting Its Circulation
By Michael S. McCartney

A Primer on Ocean Currents 3

Mi'i'i^iiri'ii/fi/tK <niil Lin/in nfriii/xicn/ Oceanographers

Towards a Model of Atlantic Ocean Circulation 5

The Plumbing of the Cl/ii/nlr's Rnil/u/m-
By MichaelS. Mri''nrli/i://

Dynamics and Modeling of Marginal Sea Outflows 9

Out They Go, Doini, and Around tin' \\brlil
By James F. Price

Meddies, Eddies, Floats, and Boats 12

How Do Mediterranean ni/d Atlantic Waters .l/ii1.-'
By Amy Bower

Where Currents Cross 16

Interarlnni of the CiilfSln'am and the Deep Western Boundary Current
By Rolinl S. P/,i;iil

Giant Eddies of South Atlantic Water Invade the North 19

Disrupted Flou- andSwiriing H'ati'n
By Philip L Richardann

The Deep Basin Experiment 22

How Does the \\'<iii-i F/<>/i in tin' l>,r/> Si</ith Mini/tie?
By Nelson Hogg

Wave-Induced Abyssal Recirculations 26

Turnini'i This Wti/i and That Way
By Michael A. Spall

Editor: Vicky Cullen • Designer: Jim Canavan

Woods Hole Oceanographic Institution

Robert B. Gagosian, Director • Guy W. Nichols, Chairman of the Board of Trustees
James M. Clark, President of the Corporation • Charles A. Dana, III, President of the Associates

Woods Hole Oceanographic Institution is an Equal Employment Opportunity and Affirmative Action Employer

The Atlantic Ocean

Progress in Describing and Interpreting Its Circulation

Who was the first physical oceanographer,
the first to observe and ponder the cause
of ocean movements? The answer is lost
in history, of course, but the first writings
on the subject date to early Greek civilization. By mid
seventeenth century, expanding exploration of the New
World had yielded enough geographic information to
allow the Jesuit priest Athanasius Kircher to construct
a chart of the world as well as detailed renditions of
smaller regions, such as the Atlantic sector from his
hand on this page. Much on this chart is recognizable
today, both in regard to principal land masses and
aspects of surface currents.

These charts include the first interpretations of the

world oceans' vast subtropical gyres. However, the
ocean circulation was deduced from extremely sparse
data (nonexistent data in many regions). The voids
were filled in with considerable imagination, an
imagination almost completely unconstrained by
physical principals, as physicists had not yet focused on
the ocean. For example, Kircher was unaware of the
physical impossibility of two currents, like those
diagrammed east of Brazil, passing through one
another, and he was particularly fond of the notion that
subterranean passages linked maelstroms and currents,
dispersing their entrances all over the world's oceans.

The classic era of physical oceanography began
about 200 years later with recognition of the central

Athanasius
Kircher's
seventeenth
century view of
Atlantic Ocean
circulation.

OCEANUS

Henry Stammel, Atlantis IL 1963

•John Swallow, Erika Dan, 1962

Val Worth ingtoii, Atlantis, 1948

role Earth's rotation plays in the
physics of ocean motion and the
fielding of expeditions for the
explicit purpose of exploring ocean
circulation. In a 1954 pamplet
entitled "Why do our ideas about the
ocean circulation have such a
peculiarly dream like quality," Henry
Stommel described this classic era of
physical oceanography :
"( 1 ) a grand cruise, or expedition,
brings back many hydrographic
stations' worth of data;

(2) extensive plots, graphs, and
tabulations of the data are made and
published for the benefit of future
generations;

(3) certain of the more striking
features of the data plots are noted;

(4 ) some plausible hypotheses are
advanced to explain them.

This procedure usually exhausted
the energy of those involved, and
almost always the funds, and the
study usually stopped at this stage."

However, the real focus of
Stommel's pamphlet was a call to the
research community for a fifth step:

"All of these steps are absolutely
necessary, of course; otherwise we
would know nothing at all about the
ocean as it really is. But for the full
development of the science there
must be one additional step:

(5) the plausible hypotheses must
be tested by specially designed
observations. In this way theories
can be rejected or accepted, or may
be modified to become acceptable."

Shortly thereafter Stommel
provided a prototype for this modern
era of physical oceanography by
hypothesizing on theoretical grounds
the nature of the basic structure of
the abyssal circulation. The

hypothesis was almost immediately validated by the
direct measurement program of John Swallow (the
recipient early in 1994 of WHOl's first Henry Stommel
Medal in Oceanography) and Val Worthington (WHOI
Physical Oceanography Department Chairman from
1974tol981),

It is now about 37 years since Stommel published his
pamphlet, and we can happily report much progress
toward fulfilling his model for modern physical
occanographic research. We still feel very much like
explorers when we cany out our measurement
programs, but we conduct and interpret them within a
framework of increasingly sophisticated physical
understanding.

We initially planned this collection of reports on
current research to focus on Atlantic deep circulation.
The diversity of topics we ended up with is an
indication of two aspects of the basic nonlinearity of
physical oceanography. First is the high degree of
linkage among physical phenomena: As some particu-
lar phemenon is examined, its physics often turns out
to be directly linked to that of other phenomena.
Second is the significant linkage among the scientists
themselves: Because ocean physics is nonlinear, it is
nearly impossible to be a specialist using a narrowly
defined technique or working on a narrowly defined
topic. We don't have to be experts in everything, but
we do need a good understanding of a broad spectrum
of phenomena that modify the physics of the ones we
want to understand.

The WHOI Physical Oceanography Department has a
32-member scientific staff working with 6 postdoctoral
associates, a 25-member technical staff, and 34
graduate students working on Ph.D.s, all ably supported
by 36 assistants. The staff works on physical oceano-
graphic problems throughout the world's oceans,
spanning the full range of scales from millimeters to
megameters, and utilizing a wide range of techniques
from intensive field-measurement programs to equally
intensive theoretical and numerical modeling studies.
It is the largest concentration of physical oceanogra-
phers in the world. Seven of these scientists have
contributed reports on one each of their many indi-
vidual projects. They include a study of how two
currents cross — not through each other, as in Kircher's
drawing, but one beneath the other. One report deals
with the very real complexity of the currents east of
Brazil, and the rest with the actual "plumbing" of the
deep circulation, not Kircher's subterranean passages,
but instead an intricate system of abyssal circulation
pathways by which waters sink from discrete source
regions and move through the Atlantic Ocean.

Michael S. McCartney

Senior Scientist

Physical Oceanography Department

Editor's note: There are many references in this volume to
the work of Henry Stommel, a physical oceanographer
extraordinaire, who died in January 1992 at 71, a loss deeply
felt by his colleagues both at WHOI and throughout the world.
A special issue oWceanus entitled A Tribi/tr in H, •///•//
Sinniiiii'l was published by the Institution in 1992. The inside
back cover of this issue features the awarding of the first
medal in his name.

Atmospheric
Pressure

Oceanic

Dynamic

Height

70°

60=

50°

Note the similarity of these imagined circulation fields for the atmosphere and ocean. Atmospheric pressure contours show a low over
Canada and a high over southeastern United States. Arrows indicate direction of flow deduced from the geostrophic relation (that between
horizontal pressure variations and horizontal currents), with higher speeds where the contours are closer together. Contour lines in the
western North Atlantic illustrate an analogous field, the dynamic height at the sea surface. The compressed bundle emerging from the south
off the coast of Florida is the Florida Current, which initiates the Gulf Stream system. T)>e ocean surface is about a meter higher in mid basin
than it is inshore of the Gulf Stream, and this drives the flow, indicated by arrows. Contours that peel off to the south indicate a partial
clockwise circulation pattern around the high-elevation region, analogous to the clockwise flow around high pressures in the atmosphere.

A Printer
on Ocean Currents

Measurements and Lingo of Physical Oceanographers

Volumes of Water Moving in Currents

Oceanographers express current flow in "millions of
cubic meters per second," a term difficult for most
people to comprehend. A large current, such as the Gulf
Suvam south of Nova Scotia, transports more than 150
million cubic meters per second, and typical transports
for the smaller deep western boundary currents are 10
to 20 million cubic meters per second. Various dense
overflows from marginal seas such as the Mediterra-
nean are even smaller, 1 to 3 million cubic meters per
second. For comparison, the sum of all the rivers
flowing into the Atlantic is about 0.6 million cubic-
meters per second. The Amazon contributes about a
third of that total, while the Mississippi River, whose

rampages plagued the midwest last summer, accounts
for only about 0.02 million cubic meters per second,
roughly one ten-thousandth of the Gulf Stream's
transport!

Properties of Seawater

Salinity: Though seawater's salinity is expressed in
grams of dissolved solids per kilogram of seawater,
today's methods measure seawater's electrical conduc-
tivity and then employ a well-known conversion
algorithm to determine salinity. While coastal waters
can exhibit a wide range of salinity as a result of
freshwater runoff, most of the world ocean lies in the
narrow salinity range 33.8 to 36.8.

OCEANUS

Temperature: Oceanographers commonly refer to
the "potential" temperature of a parcel of water. This
recognizes that a parcel of water sinking from a surface
source will, if it does not mix or exchange heat with
surrounding waters, become slightly warmer as the
pressure on it increases with depth. For example, if a

//parcel of surface seawater starts with a temperature of
0°C and salinity 35 and descends to 3,000 meters as part
of an overflow, it may warm as much as 0.3°C. An
instrnrtmnt lowered into this overflow would sense this

armer temperature, which is called the "in situ"
temperature,/ However, for most calculations and
analyses ooeanographers use the potential temperature,
which co/rects for this effect of pressure and thus
remain/0.0°C.

Tha North Atlantic is the warmest and most saline of
the world's oceans, having a mean potential tempera-
ture of i.08°C and mean salinity of 35.09, compared to
the globbl average of 3.51°C and salinity of 34.72. Most
of the waVmer andoKffe saline waters of the world are
concentratecTnTme upper kilometer of the subtropical
and tropical circulation regimes in what is called the
main thermocline (a region of rapid decrease in
temperature with depth, in the North Atlantic typically
Ot^uppfen kilometer). About 77 percent of world-ocean

'volume is colder than 4°C, with salinities in the
relatively narrow range 34.1 to 35.1. At the sea surface,
only about 26 percent of the surface area is colder than
4°C, and it is within this area that the large volume of
cold water acquires its characteristics before sinking
and traveling along paths like those discussed in this
publication.

Water Masses

The large volume of cold water described above
comprises the "deep" and "bottom" waters. In the
Atlantic there are several sources for these waters in
both hemispheres. Each hemisphere's sources are
blended by circulation and mixing. The net effect of
northern sources dominates the deep water, while the
net effect of southern sources dominates the bottom
water. In regions where the two water masses coexist,
the bottom water lies beneath the deep water, although
because of their mixing there is no sharp demarcation
between the two water masses away from their sources.
Oceanographers often recognize these dominant
sources through the proper names North Atlantic Deep
Water and Antarctic Bottom Water. The Mediterranean
outflow descends through the thermocline to form a
mid-depth saline "tongue" of Mediterranean Water,
which mixes downward through the uppermost part of
the North Atlantic Deep Water.(Authors Bower and
Price discuss the Mediterranean outflow.) Other
sources for this uppermost part of the North Atlantic
Deep Water are found in the Labrador Sea, where
winter cooling produces a large volume of water at
potential temperatures between 3.0°C and 4.5°C. Even
colder waters are produced in the Nordic Seas

(Greenland and Norwegian seas), the major source for
the rest of the North Atlantic Deep Water. The Nordic
Seas' sources are confined by the ridge system that
connects Iceland to Greenland and Europe, but spill
over the ridge to form dense overflows. The last source
for the North Atlantic Deep Water is the Antarctic
Bottom Water: Some of this water flows from the Brazil
Basin in the South Atlantic across the equator into the
North Atlantic (see articles by Hogg and Spall), where
its ultimate fate is to upwell into the lower part of the
North Atlantic Deep Water (see Pickart's article). Major
currents like the Gulf Stream , and eddies like the
North Brazil Current retroflection eddies (see
Richardson's article), may penetrate the deep water
and mix or disrupt the deep-current flow.

Current Velocity Measurements

Oceanographers measure current speed in two ways.
For direct measurements, moored current meters
record the speed of water flowing past them at intervals
of hours or days, and surface or subsurface drifters
reveal the pathways of the parcels of water in which
they were launched. Indirect estimates use a relation
between horizontal pressure variations and horizontal
currents (called the geostrophic relation). This is
analogous to a meteorologist's use of atmospheric
pressure charts to deduce the wind field. In both cases
the geostrophic velocity (of the wind or the ocean
current) is parallel to the pressure contours on the
chart and inversely proportional to the contour spacing.
For the ocean, the pressure is determined from the field
of density as deduced from an equation of state that
links density to the actual measurements of tempera-
ture and salinity. However, this calculation in the ocean
yields the velocity difference between the depth of the
calculation and a "reference level." To convert this
velocity difference to the absolute velocity, tin-
reference level velocity must be determined by direct
measurement, estimation, or some other means. Since
the number of direct velocity measurements is generally
very small relative to the hydrographic database
available for application of the geostrophic relation,
more often than not the "other means" are needed.
These include:

• "Budgets," where a reference-level velocity is deduced
by the requirement that the the sum of all the flows into
and out of a region must be zero (since sea level isn't
changing).

• "Water-mass analysis," where water-mass distribution
suggests flow pattern velocities. (For example, a given
source region may cause a property "tongue" along
which the source's characteristic properties propagate
and then mix with the surrounding water. This may
show the flow direction and allow estimation of a
speed.)

• "Level of no motion inference," where a zone of zero
speed is assumed to lie between two zones of water
flowing in opposite directions.

SPRING 1994

Towards a Model of
Atlantic Ocean Circulation

The Plumbing of the Climate's Radiator

Michael S. McCartney

Senior Scientist, Physical Oceanography Department

The theme of physical oceanographers'
research is the exploration and explanation of
ocean circulation as a natural physical
phenomenon with scales from millimeters to
megameters. In this context, the word "model" usually
brings to mind a theoretician analytically or numeri-
cally solving some system of equations that might
capture the essence of the physics of some particular
part of the ocean circulation. In other words, the
theoretician's model attempts to explain the causes and
mechanisms of observed phenomena in the ocean, or
perhaps predict unobserved ones.

Those of us who make ocean circulation observa-
! ions often get involved in a different sort of model.

While in some sheer numerical sense thKe_arej^ot of
ocean circulation measurements, tbj^j/culatior
remarkably complex, so the measurllmeru^flfe actiy
rather sparse relative to what is needea
phenomena unambiguously. Thus we also become
modelers, for even a basic description of ocean
circulation is a model: an attempt to provide in words
and pictures a synthesis, or simplified interpretation, of
the measurements.

One such model, the popular "conveyor belt"
interpretation (overleaf), has emerged from the
increasing focus (over roughly the last decade) on the
ocean's role in global climate. Its underpinnings are,
however, traced to some of the oldest observations and

artney

conduc-

tivfty/terkgeraifyre/
dejitlt water
mpling rosette.

The conveyor belt
image of the
Atlantic Ocean's
role in exchanging
warm and cold
waters with the
rest of the world's
oceans. Developed
by W. S. Broecker of
the Columbia
University's
Lamont-Doherty
Earth Observatory,
it represents the
Atlantic as a
radiator that
converts imported
warm water to
exported cold
water. The
intensity of the
conversion rate is
generally esti-
mated at between
10 and 20 m ill/on
cubic meters per
second.

An Oceanic Conveyer Belt

their early interpretations as expressed by Benjamin
Thompson, Count of Rumford, in 1797 and elaborated
by William Carpenter and Joseph Prestwich in the
1870s. This very simple conveyor-belt model is based on
a few fundamental circulation observations:

• Cold water with identifiable North Atlantic character-
istics can be traced in diluted form through most of the
world ocean.

• To conserve mass, an equal volume of warm water
must replenish the cold water draining out of the North
Atlantic.

• There is a large liberation of heat from the ocean to
the atmosphere at middle and high North Atlantic
latitudes.

The model, as far as it goes, is consistent with these
observations. In particular, the estimated replenish-
ment rate, the estimated intensity of heat loss, and the
observed temperature change of the warm-to-cold
conversions are in harmony.

An observationist builds on a simple model like the
conveyor belt by checking its consistency with observa-
tions other than the ones on which it was based. The
broader the base of such consistency, the more "real"
the model is perceived to be, and the more valuable it
becomes as a benchmark for testing a theoretician's
model or as a guide for theoretical models — for
example, what is the circulation that the model should
duplicate? But when consistency with observations is
lacking, the observationist must augment the simple
model by building in missing pieces — or perhaps
discard it as a flawed starting point.

The figure opposite is an example of an augmented
version of part of the conveyor-belt model. WHO I
Senior Scientist Bill Schmitz and I developed it over
the last few years to describe the field of horizontal
warm-water circulation in the North Atlantic (see
publication footnote). Such circulation schematics
attempt to integrate all relevant measurements and
interpretations, in this case both our own and those of
our many colleagues around the world, present and
past. Such circulation schematics have an even longer
history than the conveyor-belt interpretation, for we

'ollow the footsteps of
centuries of ocean
explorers who have tried
make maps of Atlantic
Ocean surface circula-
tion. One of the oldest of
these, by the Jesuit
Athanasius Kircher, dates
from 1678 (seepage 1).

It is important to
recognize two things
about circulation
schematics. First, they
are almost instantly
2 rendered obsolete by new
< measurements, for our

science is still in a basic

exploration phase. Second, they represent a subjec-
tive synthesis: Other oceanographers might empha-
size different aspects, that is, draw the picture and
estimate the transport amplitudes somewhat
differently — the product is very much "in the eye of
the beholder!"

The schematic opposite points out a problem
inherent in the process of sorting out ocean
circulation's role in climate. We show 13 million cubic
meters per second of warm water entering the North
Atlantic across the equator; this is our estimate of the
magnitude of cold-water production in the North
Atlantic and thus of the intensity of the conveyor belt.
However, the Gulf Stream south of Nova Scotia
transports about six times this amount! The "recirculat-
ing gyre" components of measured North Atlantic warm-
water transport thus greatly exceed the amount that is
estimated to convert from warm to cold. (The recircu-
lating gyre can be seen in the figure opposite as the Gulf
Stream water that moves across and down the middle of
the Atlantic to rejoin the northward flowing currents.)
Thus, the part of the circulation field most important
for climate issues is a small fraction of the field actually
being measured. The physical phenomenon responsible
for the northward transport of warm water through the
North Atlantic is a linked set of gyre flows rather than
the simple image of the upper limb of the conveyor belt.
The physics of the western intensification of these gyres
was first deduced by Henry Stommel in 1948, and their
relationship to the wind and thermal and hydrological
forcing remains a very active research topic to this day.
The synthesis of observations, their interpretation, and
their physical modeling proceed in parallel.

Similar degrees of complexity are emerging from
studies of the cold-water part of the system, studies
that illustrate the model-building process. The
dominant components of the meridional flow of cold
water are western intensified both within the Atlantic
and within its subbasins, which are defined by abyssal
ridges. In Kircher's old circulation-system map on page
1, water was imagined to descend into subterranean
passages and to reemerge in faraway places. The real

SPRING 1994

plumbing of the system is internal lo the ocean itself,
lint it achieves a similar result: Waters sink in a few
places to inid-depth or the seafloor (but not with the
\ lolcncc of maelstroms!) and are earned by current
systems to the far reaches of the world ocean by deep
currents. The width of these currents is restricted In
the internal physics of the ocean and steered by the
alnssal topography (not subterranean passages!)

The physics of this deep western intensification was
first elucidated by Henry Stommel in the mid 1950s
(see page 22). Such intensified flows are "easier" to
measure because they are geographically small.
Measurements of the intensity of these southward flows
show a magnitude often twice or more than that
expected from the simple conveyor-belt intepretations.
Flow measurements in the basins' interior regions have
revealed that a gyre component is responsible for the
larger-than-expected boundary current flows. This
system of gyres is more broken up by abyssal bathymetry
(top figure overleaf) than are the large warm-water
gyres. This model preserves the intensity of the
cnmeyor-belt conception's net meridional flows of cold
\\ ater. but it is also consistent with measurements of
the intensity of deep western boundary currents and
interior flows. Combining this cold-water schematic
with the warm-water circulation of the figure below
gives a representation of the full three-dimensional
ocean circulation that the quasi-two-dimensional
conveyor belt simplifies. Perhaps it is more aptly
described as a baggage carousel than a conveyor belt,
with water parcels mostly going round and round rather

than directly towards their destinations.

The latest work on the Atlantic circulation model
involves the cold-water system near the equator. For
this, MIT/WHOI Joint Program graduate student Marjy
Friedrichs, Associate Scientist Mindy Hall, and I
examined the southward cold-water transport in the
tropics of the North and South Atlantic. We found the
North Atlantic's transport dominated by a colder water
mass than the South Atlantic's. The shift does not occur
as a smooth north-to-south transition, for the contrast
in the western basin persists even quite close to the
equator. Measurements indicate that a zonal upwelling
cell in the cold water along the equator achieves the
transition (bottom figure overleaf).

One of the difficulties we face in interpreting our
data is visualization of the three-dimensional structure
of the circulation. In the top figure overleaf, the vertical
structure of the deep circulation was averaged to focus
on the overall horizontal flow. In the bottom figure, part
of the vertical structure of that deep flow is shown in a
perspective view, with the ocean above 2,000 meters
removed, and with the view looking northwest towards
the continental slope of Brazil from a point in the
middle South Atlantic south of the equator. The result
looks a bit like an expressway cloverleaf, but with no
connections between the two levels. We believe that the
indicated seven million cubic meters per second
eastward flow of the colder level upwells in the eastern
tropical Atlantic to supply the indicated seven million
cubic meters per second of westward flowing water on
the warmer level of the cloverleaf.

100

A simplified
rendition of the
system of linked
recirculation gyres
and flow pathways
of imported warm-
water as it makes
its way from the
equator to the
subpolar basins of
the North Atlantic,
with flow rates in
millions of cubic
meters per second,
indicated by
circled numbers.
Partition of the 13
million cubic
meters per second
estimated net
northward warm-
water floiv into
distinct cold-water
production regions
is given by the
numbers in
squares: 7 million
cubic meters per
second sinking in
the Labrador Sea,
4 million cubic
meters per second
(pink squares)
entering the
Nordic Seas to
return as dense
overflows into the
region south of
Iceland, and 2
million cubic
meters per second
entrained into
these overflows.

OCEANUS

Principal path-
tnii/.v of flow for the
combined deep and
bottom water
(all water
colder than
3°C)

circulation
hi the North and
tropical Atlantic,
with a perspective
view of the seafloor
bathymetry that
contains and
shapes the
pathways. Red
pathways indicate
southward
trending flows,
and blue north-
ward. The
parallel
black
strips
are the
axes of
the intense
deep recirculai-
ing gyres to either
side of the Gulf
Stream (counter-
clockwise to its
north and
clockwise to its
south). Additional
counterclockwise
recirculating gyres
are shown in
green. Tfiese gyres
mix the character-
istics of the
northward and
southward flowing
waters, and
intensify the
southward flou<
along the western
boundary.

2.4°Cto3.2°C
(near 2500 meters)

1.8°Cto2.4°C
(near 3500 meters)

The temperature change involved in this equatorial
upwelling cell, about 0.5°C, is small, but trie recognition
of the circuitous route that cold water from the North
Atlantic takes in order to move southward across the
equator has implications for the mode of the system's
response to climate change. A perturbation in the
warm- to cold-water conversion process that provides
the "northern source" water to the cold-water limb of
the conveyor belt does not simply move through the
Atlantic relatively undiluted in a deep western
boundary current "filament." The recirculation gyres
dilute the source water's characteristics with those of
the basin's interior water mass. The diluted
deep water that reaches the equator is mostly
diverted into the interior and further altered
before returning to the western boundary to
continue southward through the South
Atlantic — where we are finding additional
recirculation gyres. These new observa-
tions and interpretations present a new
modeling challenge to the theoreticians,
and, for we observationists, it points to
new measurements that will improve
our interpretive circulation model.

This work was supported by the Office of
\ni nl h'lwarch, the National Science Foundation
mi,l lln \nlionat Oceanography unit Mn/n-
x/i/in-n- Ail in i miration. One of the most recent
ir/ri'iml /i/ih/icationsonitis "North Atlantic
Circulation" (W.J. Schmitz and M.S. McCartney,
Reviews of Geophysics 31, paiji's 29 In 49).

Mike McCartney came to WHOI 20 years
ago with degrees in theoretical fluid dynamics.
Within two weeks of his arrival Fritz Fuglister,
one of the real pioneers of Atlantic physical
oceanography, took him to sea on the
institution's research vessel Chain. It changed
his outlook1 Recently Mike added up his
research cruise time and found that he has
averaged a month per year, including eight
crossings of the Atlantic He says that the main
change over the years is that the cruises come
in clumps, for example none in 1993 and a
scheduled 90 days in 1994 for crossings of the
South Pacific and of the Antarctic Circumpolar
Current Between the cruises and the fun of
untangling the data's message about the
circulation and the physics, he further relaxes
by taking long solo cruises on his old gaff
cutter and untangling all her rigging He
doesn't understand why people think it odd
that he spends so much time on the water

An interpretation of tropical cold-water circulation in the Atlantic, looking
northward/ram beloiv the equator. Numbers indicate flow rates in millions
of cubic meters per second. In the North Atlantic, a deep western boundary
current is dominated by deep water near 2°C. Tliis southward current
mostly turns eastward, away from the continental slope, after crossing the
equator, and is warmed by downward diffusion of heat in the equatorial
zone, causing an "upwelling" into the wanner layer. Westward/low of this
layer in the equatorial zone (dominated by deep tvater near2.7°C) diverges
near the western boundary info northward and southward flows, with the
latter dominating the deep South Atlantic iceatern boundary current.

SPRING 1994

Dynamics and Modeling
of Marginal Sea Outflows

Out They Go, Down, and Around the World

James F. Price

Associate Scientist, Physical Oceanography Department

The waters of the deep ocean carry a huge
volume of heat and dissolved gases along
paths that cross ocean basins and may
encircle the globe. These paths all originate
in one of a handful of polar marginal seas where cooling
or evaporation makes water sufficiently dense that it
can sink to great depths in the open ocean. The paths
can be thought of as terminating where the deep water
returns to the sea surface, usually after many centuries,
and after being slowly modified by mixing with warmer
waters and by a rain of organic material from the upper
ocean. In special regions where deep water comes to
the surface in greater volume, along the equator and in
coastal upwelling zones, the resulting nutrient load of

the deep water leads to the highest productivity found
in the oceans. This global transport aspect of deep-
ocean circulation has long been recognized as an
important element of the ocean's geochemistry and has
been studied intensively by oceanographers since at
least the 1920s.

It is a sign of the times that we now also think of the
deep ocean as a storage site, inadvertent or planned, for
anthropogenic substances such as radioactive wastes
and greenhouse gases including carbon dioxide and
fluorocarbons. Most of these are harmful to the
atmospheric environment, and some of these sub-
stances may also harm sea life. To evaluate the risks, we
need to know with some confidence where and in what

Jim Price
discusses the
Mediterranean
over/loin

9y OK,,-

5s'Pat/0n

Profiles of
currents, salinity,
and turbulent
energy dissipation
through the core of
the Mediterranean
outflow (which is
the saline bottom
layer approxi-
mately 100 meters
thick). T)ie Strait
of Gibraltar is to
the right, and the
outflow current is
moving southwest
into the Gulf of
Cadis;.

Oer - quantity these materials

are injected into the
deep sea, and where and
when they will reappear
at the sea surface. Our
present knowledge of
deep-ocean currents is
elementary, and reliable
estimates of deep-ocean
transport and storage
are very difficult to
make. Research aimed
\ at improved understand-
t ing of deep-ocean
'- circulation is thus likely

to be one of the focal
points for oceanography in
the coming decade. This
research effort will go forward at many
institutions using ideas and resources from
many disciplines (see, for example, the 1992
Paleoceznography Reports on Research, page 11).
One intriguing aspect of deep circulation is the
dynamics of the dense marginal-sea outflows that
initiate deep circulation. It is tempting to say that these
outflows "drive" the deep circulation, but the processes
that cause the deep water to return to the surface are
probably of equal importance in the long run. What we
can say with some confidence is that the transport,
temperature, and salinity of the outflow waters are
crucial factors in setting the properties of the deep
water, and they are clearly also crucial in the overall
problem of understanding deep-water circulation.

The first important step in my research program to
understand outflows was to make a detailed field survey

Deep Water Formation
in Marginal Seas

OPEN
OCEAN

A schematic showing three stages in the formation of deep water by a marginal
sea. A shallow and confined marginal sea acts as a concentration basin that
produces a comparatively small volume of very dense water. Before this dense
water can settle into the open ocean it must descend the continental shelf and
slope, and may mix significantly with oceanic waters along the way. The
water that finally reaches the deep sea may be quite different from the water
that first came out of the marginal sea.

of the Mediterranean outflow in the Gulf of Cadiz, in
collaboration with Tom Sanford of the University of
Washington and Rolf Lueck of the University of Victoria.
We measured current profiles through the outflow and
some properties of the turbulence as well (see figure at
left). These data were analyzed by Molly O'Neil Baringer,
then a Ph.D. student in the MIT/WHOI Joint Program,
and by Greg Johnson, a recent WHOI graduate then at the
University of Washington as a postdoctoral researcher.
The data showed that the Mediterranean outflow begins
with highly saline and very dense water and yet ends up
making intermediate (rather than deep) water because it
mixes intensely with oceanic water as it first begins to
descend the continental slope in the Gulf of Cadiz (see
figure below left). The large bottom slope combined with
the large initial density causes a very strong buoyancy
(gravitational) acceleration that increases the current
speed to more than a meter per second. This strong
current causes instability at the interface between the
North Atlantic water and the Mediterranean water, which
in turn causes mixing that dramatically reduces the
salinity and density of this outflow.

The end result is that mixed Mediterranean Water
settles into the North Atlantic thermocline at a depth of
about 1,000 meters. Mixing (or entrainment) causes
the transport to increase from about .7 million cubic
meters per second at the Strait of Gibraltar to about 2.5
million cubic meters per second in the North Atlantic.
The mixed Mediterranean water makes an important
salinity anomaly that spreads across the North Atlantic
basin (see Amy Bower's article on page 12).

A study of historic data on the subpolar outflows
(two flow from the Norwegian and Greenland seas and
one from the Weddell Sea) showed a similar pattern of
strong mixing as these outflows began to descend the
continental slope. However, these subpolar outflows all
produce bottom water despite beginning with a lesser
density than does the Mediterranean outflow. Why the
difference?

To understand outflow mixing in a quantitative and
ultimately predictive way, Baringer and I developed a
numerical model that simulates a single outflow
descending into a resting ocean. The elements of the
model dynamics were known from the field study: We
knew that topographic and Coriolis accelerations (an
effect of Earth's rotation that accelerates moving ocean
water to the right in the Northern Hemisphere and to
the left in the Southern Hemisphere) were essential,
and we had evidence that mixing occurred primarily in
response to strong currents. These and other insights,
when combined with the conservation laws for mass
and momentum, led to a simple but somewhat realistic
outflow model that we could use to interpret observa-
tions and explore outflow dynamics. The bottom figure
opposite shows observed and simulated density and
speed along the path of the Mediterranean outflow. The
model has some skill at simulating this outflow, and it
has proven useful for understanding the differences and
similarities between this and the other major outflows.

10 • SPRING 1994

Experiments with the model gave some surprising
results. The most important is the apparent high
sensitivity of marginal-sea outflows to the properties uf
the oceanic water through which the outflow descends.
If there is a large density contrast between the outflow
and the oceanic water (as there is in the Mediterranean
case because it enters the North Atlantic at a compara-
tively shallow depth, approximately Slid meters, before
descending to about 1,000 meters), then the outflow
current will accelerate to higher speeds, which causes
more intense mixing and increased loss of density. Such
an outflow is not likely to make bottom water, but it will
acquire an enhanced transport that may exceed the
initial transport by a factor of two or more. The
subpolar outflows begin at a much greater depth,
approximately 700 meters, and do not have such a large
density difference with respect to the surrounding
oceanic waters. While they too mix with the surround-
ing ocean waters, they nevertheless remain dense
enough to reach the seafloor.

From this work it appears that to understand the
product water of a marginal-sea outflow (and thus the
properties and transport of the water injected into the
deep-ocean circulation), we have to know the proper-
ties of the oceanic water through which the outflows
descend. This is a reminder that the ocean circulation
is a highly interconnected system; few components are
truly independent of the others. Detailed studies of
single components, in our case the outflows, are
essential to make headway on such complex systems,
but can yield only partial insight into the workings of
the overall system.

The other, holistic, approach to modeling and
analyzing ocean circulation is represented by the
general circulation models that attempt to include all
or most circulation processes in a single (gigantic!)
numerical calculation. It is these models that
will yield the estimates of deep-ocean
transport and storage needed to
guide environmental policy.
Because they are inclusive, these 250 200

general circulation models must
necessarily give up some resolution of time or
space that we can easily achieve in our single-
purpose outflow models. In the near term we
will work on ways to include outflow
models within the framework of
the general circulation models.
Our goal is to develop outflow
models that are compatible with
the lower resolution of the general
circulation models, and yet remain
responsive to changing bottom topography, sea level, or
climate. Once we have done this, we can analyze the
general circulation models to learn how and where
materials are absorbed and transported by deep-ocean
circulation, and perhaps learn where they may come to
the surface again.

'/'//« imrk truxf/ii/ili'il !>// Ha1
Office of.\'iinil Ht'ncnrrli. The Gut/
of Cadis results were descr/h* •</ / //

"Mi'tlitfi'i'inii'iiii iliillln/r Mi.rnni

and Dynamics (Science, 26

February 1993, pages 1277-1282),

and the niui/.i/i/n n-ork is

ilt wr/lii'd in a paper scheduled/or

1994 publication in Progress in

Oceanography.

Jim Price became interested in
physical oceanography while
studying plasma physics at the
University of Miami. At WHOI he
has worked on problems involving
the upper ocean, and more
recently the deep ocean circulation,
and has been active in the MIT/
WHOI Joint Program for graduate
education. He has three teenaged
children who can easily outrun him
at any distance, but he remains
about even in tennis, and claims to
be vastly superior at cooking
(anything spicey), painting (just
fish), and tending the household
pets (currently one rabbit).

Marginal Sea Outflows

80 W 60

40

200 1 50

Distance Downstream (kilometers)

Tlie observed (data points) and simulated (solid lines)
speed and density along the path of the Mediterranean
outflow, ne simulated density shows a very rapid
decrease where the outflow her/ins to descend the
continental slope (about 30 kilometers downstream
from the start). Tlie profiles in the top figure opposite,
made in this region, indicate very strong outflow
currents and intense turbulence due to miring.

0 20 40 80'E

Four major
outflows feed dense
water into the
North Atlantic. The
three subpolar
outflows (through
the Denmark
Strait and from the
Norwegian and
Weddell seas) all
produce bottom
water, while the
Mediterranean
outflow produces
intermediate
water despite
beginning as the
densest water.

Amy Bower works
with RAFOS floats.

V ^

T^^^^^___

M

_^^^^^_^^^ ,^^^^^

_ _ ^ ,^_ _ _ _

eddies, Eddies, floats
^ and Boats

Mediterranean and Atlantic Waters Mix?

Amy S. Bower

Assistant Scientist, Physical Oceanography Department

One of the most prominent features of the
North Atlantic Ocean is the tongue of salty
water that extends from Portugal westward at
mid depth. This tongue results from an
exchange flow between the Mediterranean Sea and the
Atlantic through the Strait of Gibraltar. Atlantic Water
flows through the strait into the Mediterranean, where it
is converted into saltier Mediterranean Water by an
excess of evaporation over precipitation. Mediterranean
Water, which is heavier than Atlantic Water clue to its
higher salt content, escapes into the North Atlantic
through the strait under the incoming Atlantic Water.
After leaving the strait, the dense Mediterranean Water
sinks and turns to the right, following the continental

slope of Spain and Portugal in a relatively narrow jet
known as the Mediterranean Undercurrent. Eventually,
the Mediterranean Water leaves the coast and spreads
out into the North Atlantic to form the tongue of salty
water shown in the top figure opposite.

Traditionally, physical oceanographers have thought
Mediterranean Water to be carried out into the North
Atlantic by steady currents, and further dispersed by
some random mixing. While these processes certainly
contribute to the spreading of Mediterranean Water, the
discoveiy of a new type of eddy in the Atlantic has forced
us to revise our thinking about how Mediterranean
Water is carried from its source. These eddies, found at
about 1,000 meters, are rapidly rotating lenses that

SPRING 1994

contain a con' of warm,
highly saline Mediterra-
nean Water. One of the
earliest observations of
such an eddy occurred in
lH7i; when researchers
making temperature and
salinity measurements
near the Bahamas
encountered a patch of
unexpectedly warm and
salty water. The water
property characteristics of
this patch or lens
suggested it was of
Mediterranean origin. An
acoustically tracked float
quickly deployed in the
lens showed a persistent
clockwise rotation.
This discovery
prompted a more rigorous
search for salty lenses in
the eastern Atlantic,
closer to the Mediterra-
nean. In the last decade.

20 -

0° -

100L

80

60

40°

20°

35°

T

Hyeres ^
Seamounf

30

2000

many new lenses have been found and studied using
hydrographic measurements, velocity profilers, and
floats. Eastern Atlantic salt lenses, now known as
"meddies," for Mediterranean eddies, are typically 100
kilometers in diameter, centered at 1,000 meters, and
extend over about 800 meters vertically. The water in
the meddy cores is as
much as 4°C warmer
and 1 part per thousand

re saline than the
ackgro u nd fluid,
indicating that
form I'mii
Mediterranean
Undercurrent water
somewhere along the
coast of Spain or
Portugal. Due to their
rapid rotation, mixing
between the meddy
cores and the surround-
ing fluid is limited. Thus
meddies are capable of
carrying the salty
Mediterranean Water far
from its source. This is
illustrated in the figure
at right, which shows
the trajectories of three
floats launched in
eastern Atlantic
meddies. When first
identified, these

meddies were probably already at least one year old, and
had traveled over 1,000 kilometers from their formation
site. Meddy 1 was tracked with floats for two years, and
surveyed hydrographically four times. It traveled about
2,000 kilometers and decayed slowly over that time.

25

20

T

MEDDY 2

NOV85

T

Madeira

JULY 86 MEDDY 1

1 4000 '

JUN 85
Second Crui:

Float

Trajectories
in Meddies

JULY 86

2000

~' Canary ls.~

20°-

Salinity on the
10°C surface
(about 1,000
meters deep) from
a 1976 monograph
by Valentine
Worthington. The
high-salinity water
of Mediterranean
origin is shown in
shades of red.

OCT85
MEDDY 3

30

25

20

Trajectories of
three acoustically
trackedfloats
deployed in
meddies (Mediter-
ranean eddies).
Tfie rapid
clockwise looping
of the floats is
characteristic of
meddies. (Courtesy
of Philip
Richardson)

OCEANUS

John Kemp
prepares to deploy
a sound-source
mooring aboard
R/V Oceanus in
May 1993. Tim and
other sound
sources are used to
track RAFOS floats.

slowly over that time. Meekly 2, on the other hand,
dispersed its salty core catastrophically when it collided
with a seamount.

Although the general characteristics of meddles have
been well-documented, fundamental questions about
their formation remain. For example, where do meddles
form? What physical process is responsible for their
formation? How many meddles are born each year, and

Cape St Vincent
Vilamoura

Hypothetical

Meddy

Formation

Site

36 -

35

11 10 9 8 7

A schematic diagram showing the Mediterranean Undercurrent (arrows) and the expendable

bathythermograph (XBT) line/float-launch site south of Portugal,

how long does it take a meddy to come to life? Since
meddles play a significant role in carrying salty water
from the Mediterranean into the Atlantic, we must learn
more about their life histories if we are to understand
how the distributions of temperature and salinity are
maintained in the North Atlantic.

A new program, called A Mediterranean Undercur-
rent Seeding Experiment (AMUSE), is currently
underway to address these questions. This study is being
conducted jointly with Laurence Armi of Scripps
Institution of Oceanography, and Isabel Ambar of the
University of Lisbon. In this experiment, 40 subsurface
RAFOS floats are being launched in the Mediterranean
Undercurrent south of Portugal over a six-month period,
at the rate of two per week. They are being tracked
acoustically using an array of German, French, and US
sound beacons moored in the eastern North Atlantic.
(RAFOS is the reverse spelling of the acronym SOFAR,
for SOund Fixing And Ranging, and refers to the fact
that RAFOS floats listen for sound signals transmitted by
moored sound sources, rather than themselves transmit-
ting to moored listening stations, as SOFAR floats do.)

The floats will remain underwater for about one year,
following the movement of Mediterranean Water as it
leaves the coast and mixes into the Atlantic. Some floats
will most likely be trapped in newly formed meddies
(see schematic below) . At the end of their mission, each
float will automatically drop a ballast weight, rise tn the
surface, and transmit its stored acoustic tracking data,
as well as temperature and pressure data, by radio

signals to receivers on
two satellites maintained
by the French-run Service
ARGOS. The data will
then be sent automati-
cally over the Internet
network to computers at
WHOI. The floats
themselves are consid-
ered expendable, though
we have recovered a few
test floats that were set to
surface after a month.

It would be impracti-
cal, costly, and unneces-
sary to retain a large
oceanographic research
vessel for the float
deployments. Just 2
meters long and weighing
only about 10 kilograms,
the floats, which look like
giant test tubes, can
easily be lifted and
dropped over the side by
one or two people.
Vertical temperature
profiles are necessary to
locate the float launch

SPRING 1994

position within the Undercurrent, but these can be
obtained from a small vessel using a portable expend
able bathythermograph (XBT) system. A charter
arrangement was deemed the best alternative, and we
have engaged the 73-foot motor-sailing vessel, Kialoa II,
for the six months of float deployments. (A former ocean
racer, Kiulon II is owned and operated by Frank Robben,
a retired University of California, Berkeley, mechanical
engineering professor.) Each deployment trip takes
about 24 hours tn complete. The vessel leaves
Vilumoura, a small town on the south coast of Portugal,
at dawn, ami reaches the first of the 20 XBT station sites
by late morning. XBT stations are made approximately
e\en :!i> minutes while the vessel motors or sails along a
transect crossing the Undercurrent . When the center of
the current is located, two RAFOS floats are lowered by
hand over the side, about 5 kilometers apart. The XBT
stations are then continued to the offshore edge of the
Undercurrent, and the vessel returns to Vilamoura in
the early morning hours of the next day. Some cruises
are being extended to three or four days to recover test
t'loats or to make more extensive XBT surveys of the
Undercurrent.

Most of our results will not be available until the
middle of 1994, but some tantalizing observations have
been made with a few floats deployed for 30-day test
missions. The trajectory of one test float is shown at
right superimposed on the bottom topography. This float
drifted downstream in the Undercurrent for four days
until it reached Cape St. Vincent, at the southwest
corner of Portugal. It then started looping clockwise and
moved away from the coast into deeper water. The float
made eight complete loops before the end of its mission.
This looping action, and the temperature record along
the float trajectory, indicate that this float was caught in
a new meekly — the first time the birth of a meddy has
been documented.

\Vhen this experiment is complete, we will have
many more trajectories illustrating how Mediterranean
Water makes its way into the deep North Atlantic. We
have already seen that Cape St. Vincent is a site of
meddy formation. Other
locations have been
suggested, and our
experiment will help to
confirm or reject these
sites of meddy formation.
New sites may also be
discovered. Once the
major formation sites have
been identified, the next
step in understanding the
medily story will be to
make detailed hydro-
graphic surveys during the
formation process to
determine the mecha-
nisms responsible for this
remarkable phenomenon.

39

38

37°

PORTUGAL

8

o
o

-fc.

1

Cape
St. Vincent

Trajectory of a RAFOS float deployed from Kialoa II in July 1993. Tlie dots
indicate the float's daily position. The persistent clockwise looping of the float,
starting at the southwest corner of Portugal, represents the first documented
evidence of a neu' meddy 's birth.

Thi.\ irnrk ixjiimli'd by the National Science Foundation.

Amy Bower grew up in Rockport, Massachusetts, where her
interest in science was nurtured by countless visits to Boston's
Museum of Science, weather-watching, and high school math club.
She discovered her sea legs (and the field of physical oceanography)
as an undergraduate on FW Westward, Sea Education Association's
research schooner. Lessons learned on board Westward have
helped her conduct science under sail on Kialoa II

Amy Bower and
Joao Martins
(University of
Lisbon) recover a
test RAFOS float on
the charter vessel
Kialoa II hi July
199.1

Where Currents Cross

Intersection of the Gulf Stream

and the Deep Western Boundary Current

Robert S. Pickart

Associate Scientist, Physical Oceanography Department

The Gulf Stream is perhaps the most widely
known ocean current because its impacts
range from effects on trans-Atlantic shipping
to influences on the European climate. Near
its origin off the southeast coast of the US, the Gulf
Stream (or Florida Current, as it is known there) flows
just off the continental boundary in several hundred
meters of water. After paralleling the boundary for
roughly 1,000 kilometers the current abruptly turns
offshore near Cape Hatteras, NC, and flows northeast
toward Europe. Beyond Cape Hatteras the Gulf Stream
extends to several thousand meters depth and mean-
ders significantly, in marked contrast to the shallow,
nearly straight Florida Current. The impact of this

"separation" from the boundary is enormous, yet the
precise reasons for the Gulf Stream separation (and
that of other western boundary currents) remain
unclear. The physics of the separation process is a topic
of active research.

As the Gulf Stream turns offshore, into deeper water,
it crosses the Deep Western Boundary Current
(DWBC), which flows southwest toward the equator.
The DWBC, carrying water recently formed at high
latitudes, is the primary means for replenishing the
deep waters of the vast ocean basins. Historically this
crossing was thought to be of little consequence, with
the upper-layer Gulf Stream simply passing above the
deeper boundary current. However, new modeling

SPRING 1994

Study region of the Gulf Stream -Deep Western
Boundary Current (DWBC) crossover, offshore of Cape
Hatteras, NC. Tfie historic paths of the Gulf Stream
(from satellite observations) and Dtt'BC (from water
property measurements) are shoum, along with
locations of moored instruments.

results and observations reveal that the crossover has a
profound impact on both currents, and may in fact
contribute to the Gulf Stream separation mechanism
itself. My fieldwork to address the Gulf Stream-DWBC
interaction has included a shipboard survey of the
crossover region in collaboration with a three-year
moored instrument array (see figure above).

Since the upper portion of the DWBC extends to
shallower depths than the separating Gulf Stream, one
question concerns how this upper DWBC can survive its
crossing and progress past the Gulf Stream. The
shipboard survey measured density and various
chemical properties of the water including chlorofluo-
rocarbons (CFCs), which are used as refrigerants and
aerosol propellants. The newly formed DWBC water is
recognizable by its high CFG content, absorbed from the
atmosphere while at the surface in high latitudes before
sinking to form the DWBC. The lateral CFC distribu-
tions from the shipboard survey reveal that the upper
part of the DWBC is actually sheared in half by the Gulf
Stream. At shallow depths a tongue of high-CFC water
is pulled offshore, implying that this portion of the
DWBC turns with the Gulf Stream (top figure at right).
This is confirmed by the trajectories of water parcels
calculated from the density field. Below the separating
Gulf Stream, the CFC distribution and trajectories are
much different. Here a significant portion of the DWBC
does progress past the Gulf Stream (although it bends

30

Lateral distribu-
tions ofchloro-
fluorocarbons
'(CFCs) help
reveal the fate

Western
Boundary

Current as it
encounters
the Gulf
Stream.
CFC maps
at three
different
depths in
the Deep
Western
Bound-
ary

Current
are shown
in relation to the
water parcel
trajectories
calculated
from the
density field.
Crosses
denote
shipboard-
sun>ey
measure-
ment sites,
and dotted
lines
trace the
underly-
ing

bathym-
etry.

40'

N

38'

- Upper
Layer

500-800

36°

34°

lK£Slr

Average Crossover / Maximum Crossover / Minimum Crossover /

(angle of incidence - ^ 4°) /

270

(angle of incidence = 37°) /

90° 270

180°

(angle of incidence - 7 °) /

90° 270

DWBC Gulf Stream

Topography

Hi ree years worth
of moored
measurements lias
revealed that the
angle at which the
Gulf Stream and
Deep Western
Boundary Current
cross varies
continually in
time. Tlie Deep
Western Boundary
Current responds
to changes in Gulf
Stream orientation
with approxi-
mately a one-
month delay.

seaward, apparently due to the Gulf Stream above it -
middle figure, previous page). This "splitting" of the
DWBC obviously impacts the transport of properties by
the current to lower latitudes. The Gulf Stream's
entrainment of newly formed water from the DWBC is
an effective means of replenishing the interior basin.
This entrainment also alters the dynamical structure of
the Gulf Stream itself. Whether this contributes to the
actual separation — as recent models suggest — remains
to be sorted out.

The situation at depth is different still. The CFC
map and water parcel trajectories near the 3,000-meter
level show that the DWBC passes beneath the Gulf
Stream with minimal difficulty and, as the bottom
figure overleaf shows, there is also a significant inflow
of water from offshore. This is not to say, however, that
the Gulf Stream has no impact on the deepest portion of
the DWBC. The three-year moored measurements
simultaneously tracked the movement and orientation
of the Gulf Stream and measured the flow of the near-
bottom DWBC. These observations revealed that the
path and transport of the DWBC are continually altered
by fluctuations of the upper-layer Gulf Stream.

In particular, a change in the Gulf Stream's angle of
separation causes a change in the trajectory of the
DWBC (see the figure above), while an increase in
Gulf Stream transport leads to a weakening of the
DWBC. The effect of the Gulf Stream angle change is
in accord with recent theory, though the reason for the
DWBC weakening with increased Gulf Stream
transport is still unclear.

While these recent measurements have improved
our understanding of the complex Gulf Stream-DWBC
crossover process, they have also led to more intriguing
questions. For example, by what physics does the upper-

layer Gulf Stream affect the deeper portions of the
DWBC? The importance of these two boundary currents
to the general circulation of the North Atlantic and the
significant consequences of this crossing motivate
continued investigation of their interaction. This will
also shed light on similar crossovers in other areas.

77m- ii'urk mix supported jointly by (he National Science
Fan ndnlii in unit tlir OJJ/ri'ufXni nl /,'.»/,, // // issiinniinri-i'it in
mi (ii'l/i'lf fi/ntled "Hoiv Does the Deep Western Boundary Current
Cross the Guff Stream?" (Journal of Physical Oceanography,
December 1993, pages 2602-2616).

Robert Pickart was introduced to oceanography (and
sea sickness) in college as a participant in WHOI's Summer
Student Fellow program. Despite recent experiences, such
as being knocked off his feet by large waves, he enjoys
going to sea and the excitement of collecting new
observations. His primary research interests are deep
circulation and ventilation. His most challenging en-
deavor, however, is raising his four young children with
his wife Anne

SPRING 1994

Giant Eddies of South Atlantic
Water Invade the North

Disrupted Flow and Swirling Waters

Phil Richardson

hefts a RAFOS float

aboard

tf/FKnorr.

Philip L. Richardson

Senior Scientist, Physical Oceanography Department

In the equatorial region of the Atlantic, the North
Brazil Current follows the Brazilian coast
northwestward before turning sharply to the right
between 5°N and 10°N to cross the Atlantic as the
North Equatorial Countercurrent. In 1990, satellite
ocean color images were used to identify large, 400-
kilometer-diameter, clockwise-rotating eddies that
appeared to be separating from the North Brazil
Current as it turned sharply to the east, much as warm-
core Gulf Stream rings form from northward meanders.
Because these eddies originate from retroflections or
sharp changes in current direction, they are called
North Brazil Current retroflection eddies. Satellite
observations show them to be some of the largest eddies

in the Atlantic and raise many questions about their
numbers, life histories, and importance in the general
circulation. Do these eddies provide a path for
significant amounts of upper-layer South Atlantic water
to travel up the coast? Do they transport water
primarily in the near-surface layer or do they extend
deeply into and below the thermocline? Do they drive
recirculating flows in the underlying deep water as
eddies are thought to do in the Gulf Stream system?

From 1989 to 1992 we were able to track six
retroflection eddies for the first time using surface
drifters and subsurface floats that were trapped in the
eddies' closed circulation and looped in them for as
long as five months (see lower figure on page 20). The

OCEANUS

Retro/lection
eddies (orange)
are water parcels
that pinch off from
the North Brazil
Current and
continue up the
Smith American
coast, instead of
remaining with
the North Equato-
rial Countercur-
rent. Approxi-
mately three of
these eddies form
each year starting
in July, when the
North Brazil
Current takes a
sharp right turn,
or retroflection.
After an eddy
pinches off near
8°N, the retrofl.ec-
tion forms again
farther south, near
a latitude of 5°N to
6°N. Retroflection
eddies are about
WO kilometers hi
overall diameter
near the surface
and drift
northwestward at
around 10
centimeters per
second. Tliey are
thought to be
responsible for
carrying signifi-
cant amounts of
Stm tli Atlantic
water northward
into the Caribbean
Current.

North Brazil
Current
Retroflection
Eddies

North Equatorial
Countercurrent
(NECC)

North Brazil

Current

(NBC)

Tlte inset schematic map
shows the major tropical
currents between July
and September, when the
North Brazil current
retroflects and feeds into
the eastward flowing
North Equatorial
Countercurrent. In
con trast, from Jan uary
through July the
Countercurrent disap-
pears in the western,
tropics, westward
velocities are observed in
this area, and the North
Brazil current continues
up the coast as the
Guyana Current.

40°

looping trajectories were used to
describe the number, movement, and
characteristics of these eddies.

Looping surface trajectories had
diameters up to 250 kilometers with
swirl speeds as fast as 80 centimeters
per second, dropping to diameters of
140 kilometers at 900 meters with
swirl speeds of 35 centimeters per
second. The deepest looper was at
1,200 meters with a maximum
diameter of 100 kilometers and a swirl
speed of 20 centimeters per second.
The eddy shape, determined from tin-
loop diameters, appears to be an
inverted cone. The data suggest that
at least three such eddies form each
year from July to March. They move
northwestward along the South
American coast with a mean velocity
of 10 centimeters per second, and
seem to disintegrate when they
encounter a 1,000-meter ridge
between Barbados and Tobago. The
water advected northward by the
eddies probably then enters the
Caribbean Current.

Retroflection eddies appear to
carry a significant volume of South Atlantic water
northward into the North Atlantic, short-circuiting the
longer route around the gyre formed by the North
Equatorial Countercurrent and the North Equatorial
Current. Each eddy transports about a million cubic
meters of water per second; three eddies per year
account for as much as a quarter of the total northward
transport in the upper limb of the thermohaline
(temperature and salinity driven) circulation cell.

Float and Drifter
Trajectories

.. i I

A

i

40°

~\ — r

Eddy Trajectories

B

o

A) Composite of looping trajectories measured by
surface drifters (42,45,57, 71) and subsurface SOFAR
floats (22B, 22C, 28B) in retroflection eddies from 1989
to 1992. Surface drifters' positions were recorded by
orbiting satellites several times per day. Acoustic
signals transmitted daily by the floats were recorded
by a moored array of listening stations.

B) Inferred trajectories ofsi.r retroflection eddies.
Three different eddies were observed during the period
August '1989 to April 1990(288, 22B, 45/57).

SPRING 1994

Estimates place the total
uppcr-layr current transport
of water into the North
Atlantic at about 13 million
cubic meters per second.
This upper-layer northward
transport is balanced In ihr
southward flow of cold North
Atlantic Deep Water beneath
the eddies. (See figures on
pages 6 and 7.)

The discovery of numer-
ous eddies translating up the
coast helps explain a
discrepancy between two
earlier data sets, from
drifting buoys and historical
ship drifts, that show
continuous flow up the coast

to the Caribbean from January through June. From July
through December, however, all available surface
drifters (during 1983 to 1985) in the North Brazil
Current retroflected into the countercurrent (see figure
above). In these same months, ocean-color images also
showed Amazon Water flowing around the retroflection
into the countercurrent, implying a complete surface
disruption of flow up the coast. These drifters and
images contradict historical ship drifts, which show a
continuous northwestward current there during July
through December, with a branch feeding into the
countercurrent. An explanation for this discrepancy
involves retroflection eddies. The continuous current
seen in ship-drift maps is probably an artifact of
averaging many years of velocity measurements on the
inshore side of the mean path of eddies, where eddy
swirl velocity is
northwestward. This is
demonstrated by a map of
surface-drifter velocity,
including newer trajectories
in retroflection eddies (see
figure to right), that agrees
with the earlier ship-drift
data. The Guyana Current, if
it exists during the months of
July through December, is
not a smoothly flowing
current but instead consists
primarily of a train of
retroflection eddies.

It may seem surprising
that we are just beginning to
understand a major current
system like this. Our
ignorance has been caused
partially by a lack of good in
situ data sets and partially by
the intermittent character of
these powerful eddies and

their rapid movement. Our knowledge about many
subsurface currents is even more rudimentary, which
suggests that we will find many more surprises in our
ocean-circulation studies.

Funding for the work described was provided by the National
Science Foundation. Recent scientific journal articles on I In'
subject of this article include "Tracking Ocean Eddies" (Philip L.
Richardson, American Scientist, May^lune 1993, pages 261-27 1)
and "North Bras/I Ciinr/il Eddies" ('PL Richardson, G.E.
H afford, R. Limeburner, and W.S. Brown, Journal of Geophysical
Research, 1994, Vol. 99, pages 5081-5093).

Phil Richardson grew up on a cattle ranch in California
where he spent long hours chasing cows. He ran away to sea
and eventually earned a Ph.D in oceanography from the
University of Rhode Island. His early experience on the ranch has
proven valuable in his recent efforts to follow floats and drifters
in the oceans.

60° W

Mean velocity vectors calculated by grouping nil available surface-drifter
velocity measurements from 1983 to 1993 into l°-by-l° bins. The newer data,
which are similar to historical ship drifts, include trajectories in several
retroflection eddies and show that on average the North Brazil Current runs
continuously up the coast into the Guyana Current, with some flaw peeling off
and feeding into the countercurrent. The mean northwestward current is
partly an artifact of averaging the clockui.ie rotating swirl velocity of several
eddies as they drifted northwestward along the coast.

Surface-drifting
buoy trajectories
in the North
Brazil Current
retroflection
during 1983 to
1985. All, five
available drifters
retroflected
during the months
of July to Decem-
ber, implying that
there was a
complete break in
the north-
westward flow.

OCEANUS

Nelson Hofig with
RAFqSfloats that

are bei
hi the la/xii

The Deep Basin Experiment

How Does The Water Flow in the Deep South Atlantic?

Nelson Hogg

Senior Scientist, Physical Oceanography Department

Fresh from his success at explaining why
currents like the Gulf Stream are found along
the western boundaries of all ocean basins,
Henry Stommel proposed almost 40 years ago
that thjjlwere similar features in the abyss. According
to his scheme, these would be fed by water made dense
in IjXpolar regions during winter. This convective

dilation would be completed by a slow bleed of water
from the deep boundary currents into the ocean
interior, and a broad rising through the ocean ther-
mocline (a region of rapid decrease in temperature
with depth ) into the upper ocean where it would then
I"1 carried poleward (in. for example, the Gulf Stream)
to complete the circuit.

Almost immediately, using conventional hydrography
and a novel neutrally buoyant float that John Swallow
(UK Institute of Ocean Sciences) had recently invented,
Val Worthington (WHOI) and Swallow discovered the
"Deep Western Boundary Current" part of the scheme on
the Continental Rise south of Cape Cod. Then, with help
from Stommel and others and using these same floats,
Swallow set out to confirm the existence of the slow
interior circulation that was an integral part of the
scheme. Instead, they found that the deep ocean, away
from the boundary, was dominated by a vigorous eddy
field, much like the weather systems that dominate our
day-to-day existence on land. With the ship-based
technology then available, there was no hope of being

SPRING 1994

able to extract the very weak mean flow in the presence
of so much noise from the eddies.

Advances in technology over the last four decades
nun make it possible to revisit this issue, and a group of
US, German, French, and Brazilian scientists are
engaged in an attempt to reveal the secrets of the deep

0 -

10

20

30

40 —

BO°S -

60°W

50'

40°

30

20°

10°

Tlie bathymetry of the South Atlantic showing pathways for the flow of the two
main deep water masses, North Atlantic Deep Water (NADW) and Antarctic
Bottom Water (AABW).

circulation. We have centered our program on the Brazil
Basin for several reasons: French and German oceanog-
raphers had already planned extensive hydrographic
work in this region as their contributions to the World
Ocean Circulation Experiment, previous work in the
area suggests that the return circuit through the

thermocline is surprisingly
strong in this area, and the
geometry of the Brazil
Basin is relatively
simple — there are just
three major connecting
passages to neighboring
basins and the bottom and
boundaries are smooth,
except for the Mid-
Atlantic ridge portion on
the basin's eastern side.

The observational
attack is multipronged.
First we will complete an
extensive network of
"hydrographic lines,"
making temperature,
salinity, and other
measurements. The
sampling strategy is
designed to split the basin
into a number of boxes.
Using various conserva-
tion principles (such as
for mass, momentum,
heat, and salt), we hope
to determine whether the
rising process occurs

10°E

Upper (left) and
lower (right)
Atlantic Ocean
circulation scheme
proposed by Henry
Stommel in 1957.
Water is made
dense by winter
cooling, evapora-
tion, and ice
formation in the
polar regions and
then moves
equatorward as
narrow, relatively
swift flows along
the western
margins. Water
slowly leaks away
into the interior
and then upward
across the
thermocline to
eventually return
to the poles.

OCEANUS • 23

A schematic
illustration of how
the deepest water
entering the Brazil
Basin front the
south through the
Verna Channel can
conserve both mass
and heat by rising
and changing
density. Tlie water
entering the basin
below a certain
temperature does
not leave at that
temperature, but
instead rises and
warms through
downward
diffusion of heat.
Tfiis balance
allows estimation
of thermal
diffiisivity, a
measure of the rate
of mixing in the
ocean.

Tlie Deep Basin
Experiment field
program. Filled
circles locate
acoustic sound
sources that will
track flocks of
ni'iilrallt/ buoyant
floats, rectangles
indicate moored
I'linrtit-ineter
arrays, and lini's
show planned
shipboard
In/drographic
surveys.

uniformly across the basin, as in Stommel's visionary
scheme, or perhaps is localized along the boundaries,
others have suggested. Current-meter moorings will
also be deployed, mostly along the boundaries and
within the deep connecting passages. Progress in

as

Moored Arrays
Q Sound Sources

40

20°

electronics and mooring hardware now allows us to set
these moorings for at least two years and anticipate a
90 percent data return. A similar array, moored across
the Vema Channel some 15 years ago, was part of the
inspiration for the Deep Basin Experiment: It showed
that the deepest, or
"bottom," water, was
pouring into the Brazil
Basin but had no exit other
than to warm and rise.
Swallow's neutrally
uoyant floats have
evolved considerably over
the last 40 years. We will
be setting more than 200
floats, mostly of a variety
known as RAFOS floats
developed originally by
Tom Rossby at the
University of Rhode Island
and now commercially
produced in North
Falmouth, MA. These
floats listen for moored-
acoustic-source signals
that can travel great
distances in the ocean, as
much as 1,500 kilometers
for our instruments and
depths. Consequently, we
need only nine moorings
to ensure that a given

10

SPRING 1994

float will he able to hear at least two sound sources
Accurate clocks and knowledge of the speed of sound
in water then allows triangulation on the float to
determine its position to within about a kilometer. The
miniaturization of electronics allows these instruments
to be housed within what amount to large test tubes
about 2 meters long. At the end of their two-to-three-
\ear missions, they drop a weight and rise to the
surface where they report their information to home
base through a satellite link. With these flocks of long-
lived floats, we anticipate now being able to fulfill John
Swallow's dream of directly measuring the weak
interior flow.

At least two observational techniques are available
to us now that were unknown 40 years ago. Over that
time, industry has been producing ever-increasing
amounts of chemicals called chlorofluorocarbons for
use as refrigerants and aerosol propellants, among
other things. Chlorofluorocarbons disperse through the
atmosphere and from there dissolve in the surface
layers of the ocean and are incorporated in the sinking
polar waters that are the sources for deep boundary
currents. In addition, the 1950s hydrogen bomb tests
introduced radioactive tracers into the amosphere and
ultimately into the oceans, tritium being one measur-
able tracer that is taken up by the ocean. These become
markers of the source water and allow us to estimate
the time since that water was last in contact with the
sea surface (figure below).

We have had no control over the input of these
industrial and military pollutants to the oceans.
Recently, however, WHOI Associate Scientist Jim
Lcdwell pioneered the use of a deliberately released,
nontoxic chemical compound (known as sulfur
hrxafluoride) to directly infer vertical mixing rates in

I In1 ocean interior. Small but measurable amounts of
this material are injected into the ocean over a small
area at a specified depth. Over the next few months
eddy processes stir the injected material both horizon-
tally and vertically into a smooth distribution, much
like stirring cream in a coffee cup. The rate at which
the material spreads vertically is a direct measure of
the process invoked by Stommel so long ago. Following
very successful use of this technique in the thermocline
of the eastern North Atlantic in 1992, Ledwell is now
planning to perform a similar experiment in the depths
of the Brazil Basin to answer the question of how
rapidly processes away from the boundaries are able to
carry the deep water back up to the thermocline on
their return poleward journey.

The Deep Basin Experiment is in the early stages of
field work, with a great deal of instrumentation
installed in the Brazil Basin. Data collection will
continue for several years, and then we can begin to
attack the scientific issues that have been unanswered
for the past 40 years.

I'S (null/ mi fur I lif fleet i Hiix/i/ K.i-/n'i-/nii'/il K firm nlnl by the
Xiil/n/iai Science Foundation.

Nelson Hogg dreamt of being on warm, sunny beaches by the
sea while growing up in northern Manitoba. It wasn't until he had
a summer job during college at the Defense Research Establishment
in Dartmouth, Nova Scotia, that he had that opportunity during an
instrument-testing cruise to Bermuda That was enough to convince
him that oceanography had advantages over high-energy physics
For most of the time since then, he has been at WHOI

South Atlantic Ventilation Experiment, Leg 1

250

500 750

Distance (km)

1000

A section showing
tritium concentra-
tion along a line
starting near 10°S
on the Brazilian
Coast and angling
northeastward to
cross the equator
near20°Watthe
Mid-Atlantic
Ridge. T)ie section
ends at Africa.

OCEANUS

Mike Spall rtnixan
ocean circulation
model.

Wave-Induced Abyssal
Recirculations

Turning This Way and That Way

Michael A. Spall

Associate Scientist, Physical Oceanography Department

It is generally believed that deep-ocean circulation
plays a fundamental role in the global climate
system. The abyssal circulation transports cold,
fresh water formed at high latitudes toward the
mid and low latitudes, where it upwells to maintain the
main thermocline (a region of rapid decrease in
temperature with depth) in the presence of downward
heat diffusion. This is thought to be accomplished
through a complex pattern of boundary currents and
interior flows, although the dynamics, and even the
paths, of these flows are not well understood. The
traditional view of abyssal circulation stems from Henry
Stommel's late 1950s and early 1960s theoretical work,
which assumes that deep waters are formed in very

small regions at high latitudes and uniformly upwell
throughout the lower latitudes into the upper ocean.
This simple model predicted that a series of deep
western boundary currents in the world's ocean basins
would carry the waters away from their regions of
formation, and that the basin interiors would be
characterized by very weak poleward flow.

However, recent basin-scale hydrographic measure-
ments in the North Atlantic and in the Brazil Basin of
the South Atlantic Ocean ( the location of the World
Ocean Circulation Deep Basin Experiment) indicate
that strong flows are not confined to the western
boundary regions. Instead, the abyssal ocean appears to
contain significant large-scale recirculation gyres,

26 » SPRING 1994

Schematic of the model domain representative of the
Brazil Basin in the South Atlantic Ocean. Tlie basin is
bounded to the west by South America, to the east by the
Mid-Atlantic Ridge, to the south by the Rio Grande Rise,
and to the north by the Ceara Rise. Narrow channels
allow Antarctic Bottom Water to flow into the basin
from the south and out to the north, where it crosses the
equator and enters the North Atlantic Ocean.

cross-basin flows, eastern boundary currents, and
topographic waves (see Mike McCartney's article on
page 5 for a discussion of the deep mean flows). The
presence of these currents implies that the deep ocean
is influenced by more complex physics than were
included in the simple early models. Understanding the
structure and forcing mechanisms of abyssal recircula-
tion gyres, and their relation to the deep western
boundary' currents and upwelling, is essential if we are
to fully understand deep-ocean circulation and its
role in the global climate system.

I have been using a numerical model to
investigate some of the consequences of more
complex physics on abyssal circulations. The
model approximates the continuously
stratified ocean as three constant-density
layers. The two deeper layers represent the
lower and upper Antarctic Bottom Water
(Antarctic Bottom Water is a cold, fresh water
mass that originates near Antarctica and flows
into the Brazil Basin from the south); the
shallowest layer represents the upper ocean. This
system is well suited for studying abyssal circula-
tion because it allows for steep and tall bottom
topography, time-dependent motions, such as waves
and eddies, and vertical mixing between water
masses. The effect of mixing between the layers is
paramaterized here as a uniform upwelling of water
between the deeper two layers.

The figure above shows the model configuration.
The domain is roughly patterned on the geometry of the

Brazil Basin, with bottom topography approximating the
deep bowl shape of the Brazil Basin between the South
American coast and the Mid-Atlantic Ridge. The basin's
southern and northern limits are partially blocked by
ridges, representing the Rio Grande Rise in the south
and the Ceara Rise to the north. The basin extends
approximately 3,000 kilometers in the north-south
direction and 2,000 kilometers east-west. We know that
Antarctic Bottom Water is formed at very high latitudes
in the southern oceans and spreads northward, passing
through the Brazil Basin into the North Atlantic. We
approximate this flow into the Brazil Basin by introduc-
ing a deep western boundary current of Antarctic
Bottom Water through a channel at the southern
boundary of the domain. The flow exits through a
channel adjacent to the western boundary at the
domain's northern limit. Transport within each outflow
layer may be different from that at the inflow, allowing
for overall warming of bottom waters.

We have conducted two sets of experiments, one
with a steady and one with an unsteady deep western
boundary current. We keep the physical configura-
tion— stratification, topography, and flow strength —
constant, and render the flow steady by increasing the
current's drag against the bottom. This approach allows
us to investigate the influence of time-dependent
motion on the mean state under otherwise similar flow
conditions and model physics. The schematic below
shows the mean flow of Antarctic Bottom Water in the
deep basin with a steady deep western boundary
current. Consistent with traditional models of deep
circulation driven by large-scale upwelling, the
dominant circulation features are the northward .
flowing deep western boundary current and a large-scale
cyclonic (clockwise in the Southern Hemisphere)

Schematic of the
mean circulation
of Antarctic
Bottom Waterfrom
the model with a
steady deep
western boundary
current. Colored
lines indicate
bottom water
depth (yellow for
shallow, red for
deep) while arrows
indicate flow
direction. The bold
blue arrows show
where the water
flow's into and out
of the basin. Most
of the water is
carried to the
north over the
sloping bottom
along the western
boundary in the
deep western
boundary current.
There is a cyclonic
(clockwise)
recirculation with
a depressed center
in the basin
interior.

OCEANUS » 27

Schematic of the
mean circulation
of Antarctic bottom
Water from the
model with an
unsteady deep
western boundary
current. Tliedeep
west em boundary
current still flows
to the north along
the western
boundary, but, now
the direction of
flow in the basin
interior is
ant /cyclonic
(counter-clock-
wise) with a raised
center. Hie
unsteady western
boundary current
entirely changes
the mean-flow
direction in the
basin interior.

recirculation gyre in the basin interior. The flow indicated
here is essentially void of any time-dependent motion.

To investigate the influence of time-dependent
motion, we repeat the calculation above with reduced
bottom drag, which allows small perturbations in the
mean flow to grow into large amplitude waves anil
eddies. The schematic above indicates that the mean
deep western boundary current continues to flow to the
north, but now the mean interior circulation is
anticyclonic (counterclockwise), opposite to that found
in the absence of time-dependent motions.

The nature of the variability near the western
boundary is indicated in the figure at right by a kinetic-
energy time series over the sloping bottom near the
boundary. We find the passage of high-frequency events
with velocities as fast as 15 to 20 centimeters per second
about once each year. These currents are much faster
than the mean speed of the deep western boundary
current, which is approximately 3 to 4 centimeters per
second. These energetic events are the signature of
topographic Rossby waves that are generated by
meandering of the deep western boundary current and
that propagate downslope into the deep basin interim1.
These waves carry energy from the deep western
boundary current into the basin interior and entirely
change the direction of flow around the deep basin.

Recent observations of Antarctic Bottom Water
circulation within the Brazil Basin indicate the
presence of a basin-scale anticyclonic recirculation gyre
with strength and distribution similar to that found in
the model when waves ;nv allowed to develop. Other

aspects of the observations agree with the model
circulation, including an eastward flow of lower and
westward flow of upper Antarctic Bottom Water. A key
to determining if the observed large-scale recirculation
gyre is a result of wave propagation over the western
slope, as predicted by the model, would be the presence
of topographic waves with energy propagation into the
basin interior. RAFOS floats and current meters
recently deployed as part of the World Ocean Circula-
tion Deep Basin Experiment should be helpful in
testing this hypothesis. (See Nelson Hogg's article on
page 22 for a discussion of the float experiment. )

Tliix ii'ork ix mijiportril hi/ the National Scienc.e Foundation
a ml I In- ivTii/iin. Bundesnrinisterfur Forschung und Technologic.
I'll, n/iiin imrl of the resfinr// f/,'vr//W l/c/r mix I'uiricil out
while Michael Spall was rixilinii ///<• li/xli/iiljiii- Mi -i n 'sk/inde in
Kiel. Germany. A manuscript describing this work has been
submitted for publication in the Journal of Marine Research.

Michael Spall became interested in oceanography while
working on his Ph.D in Applied Mathematics at Harvard
University. He came to Woods Hole in 1990 after spending a
postdoc at the National Center for Atmospheric Research in
Boulder, Colorado He enjoys spending time with his family,
sailing, fishing, playing volleyball, and studying oceanography.

c 400-

;

CJl (U

1i <~

c Si

m o.

•a "8

•-

200 -

4 6

Time (years)

Kinetic energy over the sloping bottom near the
western boundary. Tlie high-frequency variability
marks the passage of topographic waves that are
generated in packets by the meandering deep western
boundary current about once per year. Energy these
waves carry into the basin interior causes the large
scale recirculation gyres seen in the two previous
figures to change directions.

SPRING 1994

MBL WHO! LIBRARY

rich or within the refuge of a monastery. We can count ourselves fortunate

Institution Director Robert Gagosian reads the citation for the Henry Stommel
Medal in Oceanography before presenting the certificate to British oceanogra-
pher.John Su'alloiv. Elizabeth Stommel later presented the medal to Swallow.
Tlie medal likeness of Henry Stommel was sculpted by Judith Munk, a La Mia,
California artist and architect, and wife ofScripps oceanographer Walter Munk.

The Henry Stommel Medal
in Oceanography

Established by the Trustees of the Woods Hole Oceanographic Institution

Awarded To

John Crossely Swallow, <£&.$.

For fundamental and enduring contributions to observing and understanding ocean processes,

and in recognition of his exemplary seagoing oceanographic studies

of North Atlantic, Mediterranean, and Indian Ocean circulation

February 9, 1994

The first Henry Stommel Medal in Oceanography
was presented early this year to British oceanogra-
pher John Crossley Swallow, a long-time Stommel
friend and colleague who is best known for his
invention of the neutrally buoyant float (otherwise
known as the "Swallow float") and its pioneering use
in uncovering important elements of the ocean's
general circulation. Stommel and Swallow shared
many interests, but especially their desire to under-
stand ocean circulation.

Stommel Medal Award Committee Chair Nelson
Hogg described Swallow's capacity for hard work at
sea as "legendary." In her remarks before presenting
the medal named for her husband, Elizabeth Stommel
said of Henry: "It was important, he said, to live with
the ocean: to watch, to feel, to hear it, to confront it
mindfully each day, to let it permeate one's life." The
legendary combined work of Henry Stommel and John
Swallow, at sea and ashore, have contributed mightily
to the world's knowledge of ocean currents.

ffi
CD

GO

c-t-

O

3
3

ct>

Henry Stommel, 1920-1992

Woods Hole Oceanographlc Institution

Woods Hole, MA 02543

508457-2000