Journal of Marine Research, 56, 329–347, 1998

Deep water properties, velocities, and dynamics

over ocean trenches
by Gregory C. Johnson1

ABSTRACT
Observations of water properties and deep currents over several trenches in the PaciŽ c Ocean
central basins give consistent evidence for recent ventilation of water below the trench sills and
cyclonic sense of circulation over the trenches. A dynamical argument for this pattern is advanced.
First, a review of previous analyses of hydrographic data shows that the trenches are well ventilated
by dense bottom water, that within the trenches this bottom water generally spreads away from its
source, and that a cyclonic sense of circulation is suggested over some trenches. Then, this cyclonic
sense of circulation over the trenches is further documented using deep current meter and  oat data.
Finally, bathymetry is used to motivate a simple dynamical framework for  ow over trenches. If the
trench sides are sufficiently steep and the trench is sufficiently removed from the equator to ensure a
region of closed geostrophic contours, then any upwelling over that region will drive a strong deep
cyclonic recirculation in the weakly-stratiŽ ed abyss through vortex stretching. The magnitude of this
recirculation is limited by bottom drag. Ageostrophic  ow in a bottom Ekman layer into the trench
balances the water upwelled over the trench. The cyclonic recirculation is much stronger than the
upwelling-driven  ow predicted across blocked geostrophic contours by the linear planetary
geostrophic balance.

1. Introduction
The classical theoretical framework for the thermohaline circulation of the abyssal
ocean assumes localized sinking at high latitudes forced by buoyancy loss from air-sea
exchange. This sinking is balanced by widespread upwelling at lower latitudes that in turn
counters the effects of downward diffusion of buoyancy from the surface (Stommel and
Arons, 1960). If upwelling increases with distance from the sea  oor and bathymetry is
sufficiently gentle to avoid regions of closed geostrophic contours, the planetary geostro-
phic balance dictates overall poleward interior  ows. With a boundary condition of no net
 ow into the eastern edge of each basin, the interior  ow is fed from deep western-
boundary currents. These dynamics result in recirculation regions near the poles where
interior  ow feeds deep westward- owing boundary currents at poleward walls. The deep
western-boundary currents near these recirculations sometimes  ow back toward the high
latitude sources. Predictions of locations and strengths of deep boundary currents have

1. PaciŽ c Marine Environmental Laboratory, NOAA, 7600 Sand Point Way, N.E. (Bldg. 3), Seattle,
Washington, 98115-0070, U.S.A.
329
330 Journal of Marine Research [56, 2

been conŽ rmed by observations in locations throughout the World Ocean (Warren, 1981).
Many variations on this framework have been investigated, and a small sample are given
below. Interior zonal jets have been observed and modeled  owing away from an eastern
deep-water source (Warren, 1982) and toward an eastern deep-water sink (Warren and
Speer, 1991; Speer et al., 1995). The distortion of the circulation by gentle bathymetry has
been discussed (Warren and Owens, 1985). Abyssal spin-up and nonuniform interior
upwelling in uences have been modeled (Kawase, 1987). Effects of a fully nonlinear
continuity equation (Speer and McCartney, 1992) have also been investigated.
The focus of this work is abyssal circulation over deep trenches, speciŽ cally along the
western and northern margins of the PaciŽ c Ocean central basins. The simple Stommel-
Arons framework must be modiŽ ed substantially here because steep bathymetry results in
regions of closed geostrophic contours over the trenches, a situation predisposed to strong
recirculations (Welander, 1969). Theoretically, a dense bottom water source introduced
into a stratiŽ ed rotating deep basin shaped more like a bowl than a box can drive an
anticyclonic interior circulation. This tendency, coined the hypsometric effect, results from
decreasing vertical velocity with decreasing depth owing to the increasing area of the basin
with decreasing depth (Rhines and MacCready, 1989). However, 1-12 and 2-12 layer
numerical models as well as analytical results demonstrate that in situations where
stratiŽ cation is weak, upwelling over closed geostrophic contours is associated with a
strong cyclonic recirculation tangent to the contours. The results are similar for elevated or
depressed topography in sink-driven experiments with prescribed uniform upwelling
(Straub and Rhines, 1990) and in source-driven basin-Ž lling experiments (Kawase and
Straub, 1991; Kawase, 1993a,b). The upwelling within the region of closed geostrophic
contours results directly in vortex stretching that drives a strong cyclonic recirculation,
limited only by bottom drag. A bottom Ekman layer associated with this drag transports
mass into the region normal to the closed contours. If the circulation is in steady-state, the
bottom Ekman transport balances the mass lost to upwelling within the region.
In the following, Ž rst a literature review shows several inferences about this deep  ow
from hydrographic data. Water properties below the trench sills reveal that bottom water
there is not stagnant. That is to say, each series of trenches is actively ventilated by bottom
water. In addition, water properties in some locations suggest  ow patterns. In the PaciŽ c,
the inferences from the water properties are that the bottom water in these trenches is
spread relatively rapidly away from its southern source and often that a cyclonic sense of
circulation is evident. These two inferences are not contradictory since the southern bottom
water signature can be carried rapidly away from its source on one branch of the cyclonic
circulation around the trench. Next, velocity measurements are used to document a
cyclonic sense of circulation over several trenches. Finally, the trench bathymetry is shown
to result in closed geostrophic contours, and this fact is used to place the observations in a
theoretical framework.
1998] Johnson: Deep circulation over ocean trenches 331

Figure 1. Deep trenches at the western and northern edges of the PaciŽ c Ocean central basins
revealed by the GEBCO 7000-m isobath (IOC, IHO, and BODC, 1994). The labeled boxed areas
correspond to individual panels of Figure 5, and contain the trenches extensively discussed, (a)
The Kermadec and Tonga Trenches, (b) the southern end of the Mariana Trench, (c) the
Izu-Ogasawara, Japan, and Kuril Kamchatka Trenches, and (d) the Aleutian Trench.

2. Deep water properties

The PaciŽ c Ocean is ringed around much of its margins by deep trenches (Fig. 1).
SigniŽ cant trenches are found elsewhere in the World Ocean. However, readily available
long-term deep current measurements over trenches are all found at the western and
northern edges of the PaciŽ c Ocean central basins. For the sake of brevity, only water
properties in this region will be discussed in detail. No bottom water forms in the North
PaciŽ c Ocean; instead it enters the PaciŽ c Ocean central basins in a deep western-boundary
current from the south. The following literature review starts closest to this bottom water
source, with water properties of the Kermadec Trench, moves clockwise around the PaciŽ c
Ocean, and ends with the water properties of the Aleutian Trench, farthest from the source.
Zonal transpaciŽ c hydrographic sections at 32.5 and 28S (Whitworth et al., 1996;
Warren, 1973) cross the Kermadec Trench. The sections show that the water below the sill
is about as cold (in potential temperature), fresh, dense, oxygen-rich, and silica-poor as
bottom water just to the east. These properties are characteristics that distinguish the more
recently ventilated modiŽ ed Antarctic Bottom Water from the less recently ventilated
332 Journal of Marine Research [56, 2

modiŽ ed North Atlantic Deep Water. These water masses are the two main components of
the Circumpolar Deep Water that ventilates the deep PaciŽ c Ocean, with the former being
dominant in the bottom water of the southern hemisphere, and the latter being dominant in
the bottom water of the northern hemisphere. Properties below the trench sill distinguish
water there as strongly southern in origin. This conclusion is not surprising, since the
trench is coincident with the northward- owing deep western-boundary current banked
against the Kermadec Island Arc (Whitworth et al., 1996; Warren, 1973). The oxygen-rich
nature of the water below the sill clearly shows that water there is not stagnant. This water
must be sufficiently ventilated to maintain this characteristic in the face of biological
consumption. In addition, a trough in the isotherms immediately over the trench is visible
to varying degrees in the four zonal sections near 32.5S and the single zonal section at 28S.
Assuming geostrophy and a reference velocity of zero somewhat above the trough, this
persistent feature is consistent with a bottom-intensiŽ ed cyclonic sense of circulation over
the trench, as noted by Whitworth et al. (1996).
Farther north, a zonal transpaciŽ c section across the north end of the Mariana Trench at
24N (Roemmich et al., 1991) shows that the water below the trench still is relatively cold,
salty, dense, oxygen-rich, and silica-poor. A section near the saddle point of the Izu-
Ogasawara and Japan Trenches at 35N (Kenyon, 1983) also shows these tendencies. In the
North PaciŽ c Ocean, these tendencies distinguish more newly arrived Circumpolar Deep
Water below from the older PaciŽ c Deep Water above. By this latitude the modiŽ ed North
Atlantic Deep Water component of Circumpolar Deep Water is dominant in the bottom
water, which is why the distinguishing salinity and silica criteria for southern origin bottom
water change from south to north. At both latitudes, water of more extreme southern
characteristics is found offshore. However, the relatively oxygen-rich and silica-poor
nature of water below the trench sills again argues for relatively recent ventilation with
southern bottom water. Neither section has station spacing adequate to reveal near-bottom
isotherm de ections over the trenches.
In zonal sections across the Kuril Kamchatka Trench at 42 and 47N, and a meridional
section across the Aleutian Trench at 152W (Talley et al., 1991), local water properties are
most extreme in their southern characteristics below the trench sills. Data from four short
meridional sections across the Aleutian Trench at 165E, 175E, and 175W, discussed
previously following similar lines of inquiry (Warren and Owens, 1985; 1988), also
support this hypothesis. The role of these trenches in spreading bottom water poleward
along the western boundary and eastward along the northern boundary of the North PaciŽ c
Ocean has been discussed (Talley et al., 1991; Talley and Joyce, 1992). A map of dissolved
silica on a near-bottom potential isopycnal in the North PaciŽ c Ocean (Talley and Joyce,
1992) clearly shows a tongue of silica-poor (and oxygen-rich) water. This tongue spreads
poleward over the trench along the western boundary and then eastward over the
equatorward side of the trench along the northern boundary. In addition, the northern and
western boundary sections in this region show a narrow band of high silica at the boundary
on the inshore side of the trench, implying southward and westward  ow, respectively
1998] Johnson: Deep circulation over ocean trenches 333

(Warren and Owens, 1985; 1988; Talley et al., 1991; Talley and Joyce, 1992). Finally, close
examination of the seven individual sections across the trenches reported in these studies
reveals slight troughs in isotherms directly over the trenches, again consistent with a
bottom-intensiŽ ed cyclonic sense of circulation over the trenches. However, with the weak
shears and minimal water-property guidance in these abyssal regions, the inference of
geostrophic  ow direction from thermal wind is perilously dependent on the assumed
reference velocities.
In summary, existing studies show that water below the sills of these trenches always has
relatively, and sometimes extreme, southern water properties, suggesting that the trenches
are well ventilated with bottom water of southern origin. Previous analysis of hydrographic
data suggest decreasingly southern bottom water properties below the deep trench sills
with increasing clockwise distance from the source in the southwest corner (progressing
northward along the western boundary and then eastward along the northern boundary).
Furthermore, evidence from the literature shows that the Kuril Kamchatka and Aleutian
trenches rapidly spread bottom water away from the southern source. Studies detail how
the deep silica distribution also suggests that there is a cyclonic sense of circulation over
and within these trenches. Finally, published vertical sections show that deep isotherms just
above the trenches are suggestive of a bottom-intensiŽ ed cyclonic sense of circulation.

3. Deep velocities
Here data from six deep current meter arrays and one deep deployment of  oats are
discussed and shown to provide more evidence for a cyclonic sense of circulation over
deep ocean trenches, greatly augmenting the information from water-property and density
distributions. Most of these trenches are the same ones on the western and northern
margins of the PaciŽ c Ocean central basins discussed in the previous section. These
measurements are reviewed working from south to north along the western boundary and
then from west to east along the northern boundary, as before. Current directions here are
reported as °T, in degrees clockwise from north. For example, 90°T signiŽ es eastward  ow.
Recently, a current meter array was deployed across the Kermadec Trench at 32.5S to
quantify the transport of the deep western-boundary current there. These data have been
analyzed and discussed at some length (Whitworth et al., 1996). A deep cyclonic sense of
circulation is evident over the trench, superimposed on the equatorward  ow of the deep
western-boundary current. Twenty moorings were deployed over 22 months from 178.753W
to 168.225W, with current meters at nominal depths of 2500, 4000, and 6000 m (or
near-bottom where the bottom is shallower than 6000 m). Seven moorings west of 177W
are located over the steep sides of the trench and the adjacent eastern  ank of the Kermadec
Island Arc. The axis of the observed mean currents in this region (weighted by record
length), is near 27°T. The isobaths, for both the trench and the island-arc eastern  ank, have
roughly the same orientation. Vertical proŽ les of the mean velocity rotated in this direction
(Fig. 2) show a strong (exceeding 9 cm s2 1 ) positive (nominally equatorward)  ow along
isobaths banked against the island arc eastern  ank. This feature is the core of the deep
334 Journal of Marine Research [56, 2

Figure 2. Vertical proŽ les of velocity components [cm s2 1, positive values shaded] parallel (left
panel) and perpendicular (right panel) to the axis, 27°T, of mean velocity vectors over the
Kermadec Trench from a current meter array along 32.5S (Whitworth et al., 1996). The data are
objectively mapped assuming a Gaussian covariance with correlation lengths of 1750 m in the
vertical and 28 km in the horizontal and an error-to-signal energy of 0.01. The correlation lengths
used are roughly the mean instrument spacings in the region and the error-to-signal energy is
unrealisticallylow, so contours are faithful at and between current meters (black dots), but should
be regarded with care outside array limits. Vertical exaggeration is 100:1. The 5-minute resolution
ETOPO5 bathymetry is used. See Figure 5a for mooring locations and Figure 1 for trench location.
1998] Johnson: Deep circulation over ocean trenches 335

western-boundary current, consisting of Circumpolar Deep Water (Whitworth et al., 1996).

Offshore and shallower (around 2500 m), a weaker (exceeding 2 cm s2 1 ) negative (nomi-
nally poleward) return  ow of PaciŽ c Deep Water is found (Whitworth et al., 1996). Most
relevant to the point at hand, at deeper levels (4000 m and near-bottom) the  ow remains
nominally equatorward over the west side of the trench, but switches to a weaker (just
exceeding 2 cm s2 1 ) negative (nominally poleward)  ow over the east side of the trench.
This pattern is consistent with the geostrophic velocities implied by the isotherm trough
immediately over the trench (Whitworth et al., 1996). The velocities perpendicular to the
weighted mean current axis (Fig. 2) are everywhere small and less organized, as might be
expected for the component roughly perpendicular to the isobaths.
Data are available from a small but useful current meter array at the southwest end of the
Mariana Trench that also provide evidence for a cyclonic sense of circulation there.
Statistics for these current meter records are available on the World-Wide-Web home page
for the Division of Physical Oceanography in the Ocean Research Institute at the
University of Tokyo (http://dpo.ori.u-tokyo.ac.jp:81/). Three current meter moorings, each
with two current meters, were deployed across the southwest end of the trench. Here the
trench axis is oriented roughly east-west, and the mooring line crosses it nominally along
142.6E (see Fig. 5b for mooring locations). Deployments started in July 1995 with record
lengths from 370 to 441 days, and an average length of 424 days. Mooring latitudes range
from 10.998 to 11.600N and instrument depths range from 6095 to 9860 m. The (record
length weighted) mean velocity vectors for both current meters at each mooring are very
closely aligned with the isobaths, and again the  ow (not shown) is cyclonic in sense
around the trench axis. On the north side of the trench, over the 6960-m isobath, the  ow is
nominally westward with a magnitude of 1.3 cm s2 1. At the center of the trench, over the
10286-m isobath, the velocity magnitude is a minimum at 0.1 cm s2 1. On the south side of
the trench, over the 6520-m isobath, the  ow is nominally eastward with a magnitude of
0.5 cm s2 1.
The next useful current meter array is at the north end of the Izu-Ogasawara Trench.
These velocity data provide excellent evidence of the cyclonic sense of deep circulation
over the trench. Again, statistics are available at http://dpo.ori.u-tokyo.ac.jp:81/. The focus
here is on 30 current meter records nominally along 34N, longer than 100 days, with
deployments starting as early as November 1987 and as late as May 1995. The record
lengths range from 266 to 536 days, with an average of 401 days. Their longitudes range
from 141.170 to 142.547E and their depths range from 3830 to 8961 m. The bathymetry
here is more complicated than at the array over the Kermadec Trench, with a less uniform
orientation of the trench walls. However, the moorings are roughly clustered at six
locations, and the deep velocities are similar for most depths and deployments at each
location. Therefore, the (record-length weighted) mean velocity vectors at each location
are simply presented over the bathymetry (Fig. 3). The mean directions are very closely
aligned with the isobaths, and again the sense of circulation is cyclonic around the trench
axis. On the west side of the trench the  ow is nominally equatorward with magnitudes of
336 Journal of Marine Research [56, 2

Figure 3. Mean deep current vectors over GEBCO bathymetry (IOC, IHO, and BODC, 1994). The
Japan coast (thin lines), and isobaths at 1-km intervals (thickest line at 1 km to thinnest line at
9 km) are shown. Data from moorings along 34N (available on the World-Wide-Web home page
for the Division of Physical Oceanography in the Ocean Research Institute at the University of
Tokyo, http://dpo.ori.u-tokyo.ac.jp:81/) near the north end of the Izu-Ogasawara Trench are
records exceeding 100 days. All are below 3800 m. Data near 36N (Hallock and Teague, 1996) are
between 2000 and 4200 m at the south end of the Japan Trench. Vectors are averaged at each
location (record length weighted) from instruments within these depth intervals. Data near 1900
and 4900 m at the eastern location near 36N are excluded for reasons discussed in the text, so the
vector there is representative of 2900 m only. See Figure 5c for average mooring locations and
Figure 1 for trench locations.

3.6, 4.6, and 2.4 cm s2 1, roughly over the 4500, 6000, and 9000-m isobaths, respectively.
At the center of the trench the velocity magnitude is a minimum at 0.8 cm s2 1, nominally
equatorward. On the east side of the trench the magnitude is 3.0 and 12.8 cm s2 1, but
nominally poleward, roughly over the 9000 and 6000-m isobaths, respectively.A few more
deep current meter records (again from http://dpo.ori.u-tokyo.ac.jp:81/) to the south (near
33 and 29N) also show nominally equatorward  ow of similar magnitudes on the west side
of the trench roughly over the 5000-m isobath, but these data are not shown here.
At around 36N is another deep current meter array at the western boundary spanning the
Japan Trench near its southwestern end (Hallock and Teague, 1996) as well as RAFOS
1998] Johnson: Deep circulation over ocean trenches 337

 oats deployed along the array at 3000 m. These data are mostly consistent with a cyclonic
sense of deep circulation over the Japan Trench (Fig. 3). The current meter data have been
used to document evidence of a weak, 1.3 and 1.6 cm s2 1, nominally equatorward, deep
western-boundary current below 2000 m over the 3300 and 4600-m isobaths on the west
side of the trench (Hallock and Teague, 1996). These authors have also summarized
various other current meter records farther to the northeast at similar depths that show
similar nominally equatorward velocities on the west side of the trench. The mean  ow
directions are all strongly aligned along isobaths, as in the previous examples. On the east
side of the trench a stronger, 5.2 cm s2 1, nominally poleward  ow exists near 2900 m over
the 6400-m isobath (Hallock and Teague, 1996). Near 4900 m a relatively strong,
3.2 cm s2 1  ow is found toward 318°T. This location is very close to Kashima 1 Guyot,
which rises to around 3500 m, so the  ow at 4900 m may be heavily in uenced by this
seamount and is not included in Figure 3. Measurements near 1900 m report an even
stronger, 8.2 cm s2 1, nominally poleward, 65°T  ow, in accord with expectations, but it
may be sufficiently shallow to be unrepresentative of the abyss, so it is also not included in
Figure 3. Neutrally buoyant RAFOS  oats deployed at a nominal depth of 3000 m during
current meter operations all show nominally poleward  ow on the east side of the trench
and, with the exception of the  oat deployed farthest west, weaker nominally equatorward
 ow on the west side of the trench (S. C. Riser, personal communication, 1997).
Deep current meter data over the Kuril Kamchatka Trench come from a recent
deployment of nine moorings ranging in position along a line from 36.40N, 146.11E to
42.30N to 150.23E (W. B. Owens and B. A. Warren, personal communication, 1997). Yet
again, the available current meter records are consistent with a deep cyclonic sense of
circulation over the trench, with westward  ow on its poleward side and eastward  ow on
its equatorward side. Current meters were deployed at nominal depths of 2000, 3000, and
4000 m for a length of 744 to 758 days, with one exception of a 379-day record. The
northwestern mooring is located roughly over the 4000-m isobath on the poleward side of
the trench axis. The magnitude of the mean velocity at 2000 and 3000 m is only 0.2 and
0.5 cm s2 1, but there is a 4.3 cm s2 1  ow toward 259°T, nominally westward and roughly
parallel to the isobaths, at 4000 m (see Fig. 5c for mooring locations). The next mooring to
the southeast is located between the 7000 and 6500-m isobaths just on the equatorward side
of the trench axis. This mooring has an extra instrument at a nominal depth of 5500 m and
the exceptionally short record at 3000 m. The  ow directions and magnitudes at this
mooring are similar at all depths, and the record-length average is 8.1 cm s2 1 toward 59°T,
nominally eastward, again roughly parallel to the trench axis.
Finally, another current meter array was occupied over the Aleutian Trench along 175W
(Warren and Owens, 1985; 1988). Since only two moorings are located over the trench, a
cyclonic sense of circulation around the trench is only suggested by these data. This
circulation may be shifted poleward of the trench axis because there is eastward  ow over
the axis. Five moorings were deployed over a fourteen-month period from 45.972 to
50.990N, with current meters at nominal depths of 2000, 3000, and 4500 m. The three
338 Journal of Marine Research [56, 2

Figure 4. Vertical proŽ les of velocity components [cm s2 1, positive values shaded] oriented
perpendicular (left panel) and parallel (right panel) to the axis, 77°T, of the mean velocity vectors
over the Aleutian Trench from a deep current meter array along 175W (Warren and Owens, 1985;
1988). The data are objectively mapped assuming a Gaussian covariance with correlation lengths
of 1250 m in the vertical and 145 km in the horizontal and an error-to-signal energy of 0.01. The
correlation lengths used are roughly the mean instrument spacings in the region and the error-to-
signal energy is unrealistically low, so contours are faithful at and between current meters (black
dots), but should be regarded with care outside array limits. Vertical exaggeration is 100:1. The
5-minute resolution ETOPO5 bathymetry is used. See Figure 5d for mooring locations and
Figure 1 for trench location.

moorings north of 49N are located over the steep topography of the trench and the rise just
to the south. The axis of the observed mean currents in this region (weighted by record
length), is near 77°T. The isobaths, for both the trench and the rise just to the south, have
roughly the same orientation. Vertical proŽ les of the mean velocity rotated in this direction
(Fig. 4) show a moderate (exceeding 3 cm s2 1 ) negative, nominally westward,  ow along
isobaths on the poleward side of the trench, banked against the Aleutian Island Arc. This
 ow is the expected deep northern-boundary current (Warren and Owens, 1985; 1988).
Over the trench and on its equatorward  ank is a slightly weaker (exceeding 2 cm s2 1 )
positive, nominally eastward, return deep  ow previously without interpretation (Warren
and Owens, 1985; 1988). The velocities perpendicular to the mean current axis (Fig. 4) are
again everywhere small and less organized, as might be expected for the component
roughly perpendicular to the isobaths.
In summary, available velocity measurements from long-term deep current meters and
 oats over the Kermadec, Mariana, Izu-Ogasawara, Japan, Kuril Kamchatka, and Aleutian
Trenches are nearly all consistent with cyclonic gyres of at least a few cm s2 1 magnitude
around the trenches. These direct measurements are consistent with the inferences about
the velocity Ž eld from hydrographic property distributions and geostrophic shear in the
1998] Johnson: Deep circulation over ocean trenches 339

studies reviewed in the previous section. The measurements interpreted here to imply
cyclonic gyres above the trenches could also imply unconnected current-countercurrent
pairs aligned with the trench axes, but the theoretical framework advanced below argues
against this possibility.

4. Theoretical framework
For all of the trenches discussed in the previous two sections, the bathymetry unambigu-
ously results in regions of closed planetary geostrophic, or f/h, contours. Here f is the local
Coriolis parameter and h is the total thickness of the water column. Regional plots
presented for the Kermadec and Tonga Trenches (Fig. 5a); the southern portion of the
Mariana Trench (Fig. 5b); the Izu-Ogasawara, Japan, and Kuril Kamchatka Trenches
(Fig. 5c); and the Aleutian Trench (Fig. 5d) conŽ rm this assertion (see Fig. 1 for Fig. 5
panel locations). In some studies h is deŽ ned as the thickness of some deep layer beneath
an arbitrary isobath or isopycnal (Kawase and Straub, 1991; Kawase, 1993a,b). DeŽ ning h
as the total water-column thickness is a more stringent test for closed geostrophic contours.
If a thinner abyssal layer alone were to feel the in uence of the isobaths, the regions of
closed contours over the trenches would be larger than those shown. At any rate, linearized
planetary geostrophic theory requires blocked geostrophic contours, so it is not applicable
where contours are closed. Hence, the trenches are special dynamical regions, as discussed
below.
Numerical simulations of Ž nite-depth abyssal layers under inŽ nitely deep resting surface
layers have shown that over regions with closed geostrophic contours upwelling results
directly in vortex stretching that drives a cyclonic recirculation around these contours
(Straub and Rhines, 1990; Kawase and Straub, 1991; Kawase, 1993a,b). Unlike eddy-
driven  ows, which have a direction dependent on the contour gradient, these recircula-
tions are cyclonic in sense over both hills and trenches that result in regions of closed
geostrophic contours, since they are driven by upwelling-induced vortex stretching. These
experiments cannot reach a steady state because they rely on prescribed upwelling or a
supply of mass to the abyssal layer for forcing and do not include vertical mixing.
However, the abyssal upwelling or mass source can be kept small so the net abyssal layer
thickness changes very slowly. The vortex stretching interior to a given closed geostrophic
contour owing to the upwelling is balanced by the bottom drag along that contour, so the
velocities within the regions of closed geostrophic contours reach a nearly steady balance.
For a similar dynamic balance to hold over the trenches, the mechanics must be along
the following lines: A net upwelling over the trench is balanced by vertical diffusion at the
top of the abyss. This upwelling maintains relatively low pressure just over the trench. In
the abyss, where stratiŽ cation is weak and  ow nearly barotropic, conservation of potential
vorticity tends to prevent water from  owing across geostrophic contours and eliminating
the pressure gradient, so water recirculates cyclonically around the trench in geostrophic
balance. Restated, the upwelling results in vortex stretching, driving a cyclonic recircula-
tion around the trench. Water can cross geostrophic contours in a bottom Ekman layer or
340 Journal of Marine Research [56, 2

Figure 5. Absolute magnitude of planetary geostrophic, or f/h, contours [at 2 3 102 9 m 2 1 s2 1

intervals] where f is the local Coriolis parameter and h the total water depth, plotted for three
different trench systems, (a) the Kermadec and Tonga Trenches [from 0.6 to 1.6 3 102 8 m2 1 s2 1 ]
with mooring locations for Figure 1 (1 ’s); (b) the southern portion of the Mariana Trench [from
0.2 to 1.2 3 102 8 m2 1 s2 1 ] with the three mooring locations (1 ’s); (c) the Izu-Ogasawara, Japan,
and Kuril Kamchatka Trenches [from 0.8 to 1.8 3 102 8 m2 1 s2 1 ] with average locations for
Figure 3 and the two moorings over the Kuril Kamchatka Trench (1 ’s); and (d) the Aleutian
Trench with mooring locations for Figure 4 (1 ’s) [from 1.6 to 2.6 3 102 8 m2 1 s2 1 ]. The 5-minute
resolution ETOPO5 bathymetry is used. Some of the contours over each trench (lightest regions)
are closed. The white regions are land masses. See Figure 1 for location of each panel.
1998] Johnson: Deep circulation over ocean trenches 341

Fig. 5. (Continued)

where small-scale topographic relief allows ageostrophic  ow, so the cyclonic recircula-
tion is sufficiently strong such that the bottom ageostrophic transport into the trench
balances the mass lost to upwelling. Restated, the magnitude of the upwelling-driven
cyclonic recirculation is limited by bottom drag. After ventilating the trench ageostrophi-
cally near the bottom, water slowly spirals upward as it upwells and recirculates
cyclonically around the trench.
More insight into the dynamical balance can be obtained by manipulating reduced-
gravity, f-plane, shallow-water equations for an abyssal layer. In addition to the geostrophic
balance, the momentum equation includes time-dependence, advection and a simple
Rayleigh term, R 5 (2vf )1/2 h 2 1, that multiplies the horizontal velocity u. This term
parameterizes friction as a bottom drag resulting from interior spin-down by vertical
velocity at the top of the bottom Ekman layer. Here vertical eddy viscosity is v, the Coriolis
parameter is f, and the abyssal layer thickness is h. The continuity equation is fully
nonlinear, retaining time-dependence and including a source term w with the dimensions of
velocity to represent upwelling at the top of the abyss. Manipulation of the momentum and
continuity equation gives a potential vorticity equation
w z
qt 1 u ·= q 5 q 2 R , (1)
h h
where the potential vorticity q 5 (f 1 z )/h, and the relative vorticity z 5 = 3 u.
342 Journal of Marine Research [56, 2

Fig. 5. (Continued)

Following Pedlosky (1987), a simple balance is obtained by multiplication of (1) by h

followed by integration over a region, Ac , bounded by a closed geostrophic contour, Cc ,

e e r
w
f dA 5 Ru · d l. (2)
Ac h Cc

The area integral of the upwelling drives vortex stretching that is balanced by bottom drag
on circulation around the closed geostrophic contour. In deriving (2) upwelling is assumed
to be balanced by vertical diffusion, so that the system is in steady-state and the
time-dependence vanishes. The vertical advective-diffusive balance at the top of the abyss
1998] Johnson: Deep circulation over ocean trenches 343

Fig. 5. (Continued)

coupled with the Ekman layer at the bottom allows application of the divergence theorem,
so upon integration the advective term also vanishes. The upwelling term in (1) remains as
an area integral in (2), with a simpliŽ cation resulting from the assumption of small Rossby
number so z ½ f. The frictional curl term in (1) reduces to a line integral in (2) after
integration and the application of Stoke’s theorem.
This dynamical balance can be applied to the velocity data over the trenches to get an
estimate of the magnitude of the necessary upwelling. For example, using the 5-minute
resolution ETOPO5 bathymetry, the Izu-Ogasawara Trench is enclosed by f/h 5 1.15 3
102 8 m2 1 s2 1, bounding an area of 7.2 3 1010 m2 (roughly 980 km in the meridional
direction by 73 km in the zonal direction) with a path integral of length 2.3 3 106 m. The
deep velocity tangent to this geostrophic contour over the Izu-Ogasawara Trench is roughly
0.03 m s2 1. The area-averaged value of f/h inside this geostrophic contour is 0.99 3
102 8 m2 1 s2 1; and the mean value of ( f )1/2 h 2 1 along the path is 1.3 3 102 6 s2 1/2 m2 1. The
least well constrained parameter is n , which could conceivably range from 1 3 102 5 to 1 3
102 2 m2 s2 1. Abyssal viscosity estimates are sorely lacking, so this range is based on a
range for microstructure-based estimates of abyssal diffusion (Polzin et al., 1997) assum-
ing a Prandtl number between one and twenty. Fortunately, this parameter is under a square
root, so the three-decade variability is reduced to a factor of thirty. These parameters yield
upwelling velocities from 4.0 3 102 7 to 1.3 3 102 5 m s2 1, a range roughly three to one
hundred times one estimated interior value for the abyss (Stommel and Arons, 1960).
While the present estimates are large, they fall well within the range of abyssal mass
budgets for sill over ows, such as a recent one made north of the Samoan Passage
(Roemmich et al., 1996). The net volume transport of bottom water ageostrophically
ventilating the trench required to feed the upwelling over it is reasonable. It ranges from
0.03 3 106 to 0.9 3 106 m3 s2 1, at most a tenth of the net in ow of Circumpolar Deep
Water into the North PaciŽ c Ocean (Johnson and Toole, 1993). Were the effective vertical
eddy viscosity and therefore the deep upwelling near the upper limits over most trenches,
then a signiŽ cant fraction of Circumpolar Deep Water could be cycled through these
trenches as it moves around the PaciŽ c Ocean.
344 Journal of Marine Research [56, 2

5. Discussion
The observed deep cyclonic sense of circulation over the trenches is consistent with a
theoretical framework predicting cyclonic recirculations driven by upwelling over regions
of closed geostrophic contours. Of course, without other evidence, these recirculations
could be driven by eddy potential-vorticity  uxes resulting from current meandering,
waves, and other time-dependent phenomena in the upper ocean (Haidvogel and Rhines,
1983). In fact, the eddy driving does not have to occur over the entire trench, since such
forcing over just a small region of the closed geostrophic contours has been shown to drive
recirculation gyres (Thompson, 1995). Since most of the trenches discussed are overlaid
somewhere by a near-surface boundary current or boundary current extension, this
mechanism cannot be discounted as a potential driving mechanism for the cyclonic
recirculations over the trenches. However, the relatively oxygen-rich and (in most
locations) silica-poor character of water below the trench sills argues that the trenches must
be relatively vigorously ventilated by bottom water, which must upwell over the trenches,
most likely driving the cyclonic recirculations there.
This discussion of course begs the question of what forces the cyclonic recirculation.
The simple vertical advective-diffusive balance posited for the deep ocean (Munk, 1966),
where upwelling of cold water from below balances diffusion of heat from above, may
dominate and allow a density gradient along the trench axis. Alternatively, the water at the
trench bottom may be warmed mostly from below by geothermal heating, allowing it to
upwell and be replaced by colder, denser, source water. The relative sizes of geothermal
heating from below and vertical diffusion of heat out of the abyssal layer above the trench
can be compared. There are few heat  ow measurements within trenches, but the values
around 0.050 W m2 2 in the Japan Trench (Yamano and Uyeda, 1989) are similar to the 25
measurements characterized as being within PaciŽ c Ocean trenches with depths greater
than 5000 m reported by Pollack et al. (1993). The diffusive  ux of heat from above is
given by k r cpu z, where r cp , 4 3 106 J m2 3 °C, and u z , 1 3 102 4 °C m2 1 for above the
trench sills. For the two quantities to be of comparable magnitude requires a vertical
diffusivity k of 1 3 102 4 m2 s2 1, certainly plausible for an abyssal value over steep
topography (Polzin et al., 1997). With the large uncertainty in k and u z, it is difficult to
speculate which effect might dominate.
Many long trenches exist, but no more of which we are aware have long-term velocity
measurements over them. In the PaciŽ c Ocean, the Philippine and Middle America
Trenches are both signiŽ cant trenches with closed geostrophic contours. These trenches are
located in marginal basins, and may not have sufficient bottom water ventilation to drive a
strong cyclonic recirculation. However, maps of bottom water properties including
dissolved oxygen and silica do suggest that neither is totally stagnant (Mantyla and Reid,
1983). The Peru Chile Trench is probably long enough to be locally ventilated by bottom
water at the southern end, supporting overall deep upwelling. This supposition is supported
by a study of the abyssal  ow in the region suggesting the Peru Chile trench is a primary
1998] Johnson: Deep circulation over ocean trenches 345

source of bottom water in ow for the Peru and Panama Basins (Lonsdale, 1976). The fact
that this trench is adjacent to an eastern boundary may not have a large effect on the nature
of the recirculation, except on how it is fed with dense water. However, the trench
intersects the equator, meaning the geostrophic contours are not closed. Thus, there should
only be a de ection of the normal planetary geostrophic  ow warped by the topography,
not nearly as strong as the cyclonic recirculations (Kawase, 1993a). In the Atlantic Ocean,
a long trench that might have a cyclonic recirculation similar to those discussed previously
is the South Sandwich Trench. In the Indian Ocean, the Java Trench, which intersects the
equator and sits on an eastern boundary, is probably dynamically similar to the Peru Chile
Trench in the PaciŽ c Ocean.
It would be interesting to quantify the strengths of upwelling and recirculation around a
deep trench. The recirculation strength in the trenches may dwarf the size of the net
along-trench  ow and upwelling, as suggested by the theoretical framework advanced in
the previous section. In addition, the presence of deep western-boundary currents fre-
quently confounds matters. These factors would make it difficult to sufficiently instrument
a trench with current meters to diagnose the recirculation, net  ow, and upwelling. Careful
hydrographic measurements do reveal water-property contrasts and geothermal shear.
However, the small signals, the potential for aliasing owing to the synoptic nature of the
measurements, and the difficulty in determining reference velocities all mean that diagnos-
ing an accurate absolute horizontal circulation much less upwelling over the trenches with
hydrographic data alone is virtually impossible. Deep  oat measurements would be
difficult because side walls present problems below trench sills and above trench sills the
 oats might wander away from the narrow trenches. One possibility for quantifying the net
 ow and upwelling strength is a deliberate tracer release experiment. The path length of
closed geostrophic contours and the magnitude of the deep cyclonic velocity for the
Izu-Ogasawara Trench (if the velocity measurements at the north end are typical) suggest
an advective time-scale of two years for  ow to cycle around the trench, a period that
makes it amenable to such a technique. For the range of vertical velocity estimated in the
discussion, a rise of between 30 and 1000 m would be expected in one circuit of the
hypothesized spirals described by deep water parcels over the trench. The horizontal and
vertical movement and spreading of a tracer patch released below the trench sill on one side
of the trench could be used to diagnose the deep circulation and upwelling there.

Acknowledgments. This work was funded by the NOAA Office of Global Programs through the
Climate and Global Change Program and the NASA Physical Oceanography Program. The work
would not have been possible without the current meter and  oat data collected, some of it quite
recently, and kindly provided prior to publication by Worth D. Nowlin, Jr., W. Brechner Owens,
Stephen C. Riser, Keisuke Taira, Bruce A. Warren, and Thomas Whitworth, III. Their efforts are
gratefully acknowledged. LuAnne Thompson helped with theoretical considerations.The comments
of Bruce Warren and an anonymous referee greatly improved the manuscript. This is PMEL
contribution number 1909.
346 Journal of Marine Research [56, 2

REFERENCES

Haidvogel, D. B. and P. B. Rhines. 1983. Waves and circulations driven by oscillatory winds in an
idealized ocean basin. Geophys. Astrophys. Fluid Dyn., 25, 1–63.
Hallock, Z. R. and W. J. Teague. 1996. Evidence for a North PaciŽ c deep western boundary current. J.
Geophys. Res., 101, 6617–6624.
IOC, IHO and BODC. 1994. Supporting volume to the GEBCO Digital Atlas. Published on behalf of
the Intergovernmental Oceanographic Commission (of Unesco) and the International Hydro-
graphic Organization as part of the General Bathymetric Chart of the Oceans (GEBCO), British
Oceanographic Data Centre, Birkenhead. This volume accompanies a CD-ROM.
Johnson, G. C. and J. M. Toole. 1993. Flow of deep and bottom waters in the PaciŽ c at 10°N.
Deep-Sea Res. I, 40, 371–394.
Kawase, M. 1987. Establishment of deep ocean circulation driven by deep water production.J. Phys.
Oceanogr., 17, 2294–2317.
Kawase, M. 1993a. Effects of a concave bottom geometry on the upwelling-driven circulation in an
abyssal ocean basin. J. Phys. Oceanogr., 23, 400–405.
—— 1993b. Topographic effects on the bottom water circulation of the western tropical North
Atlantic Ocean. Deep-Sea Res. I, 40, 1259–1283.
Kawase, M. and D. Straub. 1991. Spinup of source-driven circulation in an abyssal basin in the
presence of bottom topography. J. Phys. Oceanogr., 21, 1501–1514.
Kenyon, K. 1983. Sections along 35°N in the PaciŽ c. Deep-Sea Res., 30, 349–369.
Lonsdale, P. 1976. Abyssal circulation of the southeastern PaciŽ c and some geological implications.
J. Geophys. Res., 81, 1163–1176.
Mantyla, A. W. and J. L. Reid. 1983. Abyssal characteristics of the World Ocean waters. Deep-Sea
Res., 30, 805–833.
Munk, W. 1966. Abyssal recipes. Deep-Sea Res., 13, 207–230.
Pedlosky, J. 1987. Geophysical Fluid Dynamics, Springer Verlag, New York, 710 pp.
Pollack, H. N., S. J. Hurter and J. R. Johnson. 1993. Heat  ow from the Earth’s interior: Analysis of
the global data set. Rev. Geophys., 31, 267–280.
Polzin, K. L., J. M. Toole, J. R. Ledwell and R. W. Schmitt. 1997. Spatial variability of turbulent
mixing in the abyssal ocean. Science, 276, 93–96.
Rhines, P. B. and P. MacCready. 1989. Boundary control over the large-scale circulation. Proc. Fifth
’Aha Huliko’a Hawaiian Winter Workshop on Parameterization of Small-Scale Processes, Hawaii
Institute of Geophysics, Honolulu, 75–97.
Roemmich, D., S. Hautala and D. Rudnick. 1996. Northward abyssal transport through the Samoan
passage and adjacent regions. J. Geophys. Res., 101, 14093–14055.
Roemmich, D., T. McCallister and J. Swift. 1991. A transpaciŽc section along latitude 24°N: the
distribution of properties in the subtropical gyre. Deep-Sea Res., I, 38, Suppl. 1, S1–S20.
Speer, K. G. and M. McCartney. 1992. Bottom water circulation in the western North Atlantic. J.
Phys. Oceanogr., 22, 83–92.
Speer, K. G., G. Seidler and L. Talley. 1995. The Namib Col Current. Deep-Sea Res. I, 42,
1933–1950.
Stommel, H. and A. B. Arons. 1960. On the abyssal circulation of the World Ocean—II. An idealized
model of the circulation amplitude and pattern in oceanic basins. Deep-Sea Res., 6, 140–154.
Straub, D. N. and P. B. Rhines. 1990. Effects of large-scale topography on abyssal circulation. J. Mar.
Res., 48, 223–253.
Talley, L. D. and T. Joyce. 1992. The double silica maximum in the North PaciŽ c. J. Geophys. Res.,
97, 5465–5480.
1998] Johnson: Deep circulation over ocean trenches 347

Talley, L. D., T. M. Joyce and R. A. DeSzoeke. 1991. TranspaciŽ c section at 47°N and 152°W:
distribution of properties. Deep-Sea Res. I, 38, Suppl. 1, S63–S82.
Thompson, L. 1995. The effect of continental rises on the wind-driven ocean circulation. J. Phys.
Oceanogr., 25, 1296–1316.
Warren, B. A. 1973. TranspaciŽ c hydrographic sections at Lats 43°S and 28°S; the SCORPIO
Expedition-II. Deep water. Deep-Sea Res., 20, 9–38.
—— 1981. Deep circulation of the World Ocean, in Evolution of Physical Oceanography, ScientiŽ c
Surveys in Honor of Henry Stommel, B. A. Warren and C. Wunsch, eds., The MIT Press,
Cambridge, Massachusetts, 6–41.
—— 1982. The deep water of the Central Indian Basin. J. Mar. Res., 40, Suppl., 823–860.
Warren, B. A. and W. B. Owens. 1985. Some preliminary results concerning deep northern-boundary
currents in the North PaciŽ c. Prog. Oceanogr., 14, 537–551.
—— 1988. Deep currents in the central subarctic PaciŽ c Ocean. J. Phys. Oceanogr., 18, 529–551.
Warren, B. A. and K. G. Speer. 1991. Deep circulation in the Eastern South Atlantic Ocean. Deep-Sea
Res., 38, Suppl., S281–S322.
Welander, P. 1969. Effects of planetary topography on the deep-sea circulation. Deep-Sea Res., 16,
Suppl., 369–392.
Whitworth, T., III, B. A. Warren, W. D. Nowlin, Jr., S. B. Rutz, R. D. Pillsbury and M. I. Moore. 1996.
On the deep western-boundary current in the Southwest PaciŽ c Ocean, Prog. Oceanogr., (submit-
ted).
Yamano, M. and S. Uyeda. 1989. Heat  ow in the western PaciŽ c, CRC Handbook of Sea oor Heat
Flow, J. A. Wright, K. E. Louden and J. R. Moore, eds., CRC Press, Inc., Boca Raton, Florida,
277–303.

Received: 6 October, 1997; revised: 6 January, 1998.