large amount of synthesis work, including a number of box inverse models. Reid (personal communication) has nearly completed an Indian Ocean
circulation analysis similar to his previous Pacific
and Atlantic works (e.g. Reid, 1994), and incorporating all WOCE Hydrographic Programme
(WHP) data along with historical data. Ganachaud
et al. (2000) focus on the Indian Ocean portion of
their global inverse, showing a somewhat higher
Indonesian Throughflow and lower inflowing deep
water than in previous estimates, thus somewhat
reducing the strength of the overturning circulation compared with previous results. They also use
the inverse model results to calculate surface heat
and fresh water fluxes, providing a useful constraint on climatologies. Sloyan (1997) used
hydrographic sections, including several WHP sections, in a box inverse of the Southern Ocean,
including air–sea and dianeutral fluxes to constrain the water mass transformations. She diagnosed the formation rate of SubAntarctic Mode
Water (SAMW) north of the Antarctic Circumpolar Current (ACC), and showed that the sources of
SAMW are thermocline water north of and surface
water south of the ACC. The southward meridional heat transport across 30°S was dominated by
the Indian Ocean.
A consortium in the US and Australia (Robbins,
Wijffels, Toole, Johnson and Talley) is currently
making a box inversion of the Pacific WHP data
using initial reference velocity fields based on
examination of all of the sections, and incorporating constraints from direct velocity measurements.
Initial results for heat transports and convergences
are encouraging. Robust results can be obtained
despite the use of intersecting, non-synoptic sections, showing that zonal heat redistribution is
larger than meridional in the South Pacific. Heat
flux convergences for the many boxes formed by
the intersecting sections have been compared with
many different air–sea flux climatologies, showing
that model-data-based climatologies are better than
data-only climatologies in the data-poor southern
South Pacific, and providing bias estimates for
the net heat fluxes computed from the various
climatologies.
For the South Atlantic, Holfort and Siedler’s
(2001) inversion of WOCE sections provides results
for mass, heat, fresh water and nutrient transports
and the conclusion that the warm water that feeds
the North Atlantic Deep Water overturning is both
upper ocean and Antarctic Intermediate Water,
with a higher fraction of the upper ocean waters.
B. A. King (personal communication) is using a
box inverse to study the southwestern Atlantic circulation; results thus far suggest that excess intermediate water enters at Drake Passage and excess
surface water leaves the box, supporting Rintoul’s
(1991) result of net upwelling in the southwestern
Atlantic.
Cunningham (2000) used a Bernoulli inverse to
construct streamlines for the North Atlantic, with
emphasis on the North Atlantic Current and subpolar region. He found a northward influx of
14 Sv into the eastern subpolar region and that the
Irminger Basin circulation is equally fed from the
North Atlantic Current west of the Reykjanes
Ridge and westward flow south of Iceland. Constraints on the outflow from the Mediterranean
are being produced using an inverse model in
a box around the outflow, with initial results
showing almost all flow exiting to the north and
west despite the prevalence of Meddies to the
southwest (D. Slater and H. L. Bryden, personal
communication).
Work on global inversions will continue with
the complete WOCE data set, providing improved
estimates of air–sea fluxes, diapycnal velocities,
and complete circulation schemes. These inversions are of increasing complexity, which in its
limit is equivalent to state estimation. For instance,
a group at the Alfred-Wegener-Institut in Germany
is using a number of complementary inversion and
data assimilation techniques to study the mean
and time-varying circulation of the Atlantic. Continuing section and basin-oriented work will contribute to further improvement of the global-scale
work. Results may also eventually be used to test
basic hypotheses about ocean dynamics.
Basin-scale property mapping as an aid to circulation analyses is included in some of the abovementioned works, particularly that of Reid. The
basin atlases, described below, will include property maps. The potential vorticity field is important for planetary wave propagation (e.g. de Szoeke
and Chelton, 1999) and also provides information
like that of other tracers for circulation. O’Dwyer
and Williams (1997) mapped the global potential
vorticity field using historical data, showing the
zonality of potential vorticity at mid-depth and
large regions of potential vorticity that deviate
from zonal at abyssal depths, including near
SECTION 7 INSIGHTS FOR THE FUTURE
528
circulation analysis similar to his previous Pacific
and Atlantic works (e.g. Reid, 1994), and incorporating all WOCE Hydrographic Programme
(WHP) data along with historical data. Ganachaud
et al. (2000) focus on the Indian Ocean portion of
their global inverse, showing a somewhat higher
Indonesian Throughflow and lower inflowing deep
water than in previous estimates, thus somewhat
reducing the strength of the overturning circulation compared with previous results. They also use
the inverse model results to calculate surface heat
and fresh water fluxes, providing a useful constraint on climatologies. Sloyan (1997) used
hydrographic sections, including several WHP sections, in a box inverse of the Southern Ocean,
including air–sea and dianeutral fluxes to constrain the water mass transformations. She diagnosed the formation rate of SubAntarctic Mode
Water (SAMW) north of the Antarctic Circumpolar Current (ACC), and showed that the sources of
SAMW are thermocline water north of and surface
water south of the ACC. The southward meridional heat transport across 30°S was dominated by
the Indian Ocean.
A consortium in the US and Australia (Robbins,
Wijffels, Toole, Johnson and Talley) is currently
making a box inversion of the Pacific WHP data
using initial reference velocity fields based on
examination of all of the sections, and incorporating constraints from direct velocity measurements.
Initial results for heat transports and convergences
are encouraging. Robust results can be obtained
despite the use of intersecting, non-synoptic sections, showing that zonal heat redistribution is
larger than meridional in the South Pacific. Heat
flux convergences for the many boxes formed by
the intersecting sections have been compared with
many different air–sea flux climatologies, showing
that model-data-based climatologies are better than
data-only climatologies in the data-poor southern
South Pacific, and providing bias estimates for
the net heat fluxes computed from the various
climatologies.
For the South Atlantic, Holfort and Siedler’s
(2001) inversion of WOCE sections provides results
for mass, heat, fresh water and nutrient transports
and the conclusion that the warm water that feeds
the North Atlantic Deep Water overturning is both
upper ocean and Antarctic Intermediate Water,
with a higher fraction of the upper ocean waters.
B. A. King (personal communication) is using a
box inverse to study the southwestern Atlantic circulation; results thus far suggest that excess intermediate water enters at Drake Passage and excess
surface water leaves the box, supporting Rintoul’s
(1991) result of net upwelling in the southwestern
Atlantic.
Cunningham (2000) used a Bernoulli inverse to
construct streamlines for the North Atlantic, with
emphasis on the North Atlantic Current and subpolar region. He found a northward influx of
14 Sv into the eastern subpolar region and that the
Irminger Basin circulation is equally fed from the
North Atlantic Current west of the Reykjanes
Ridge and westward flow south of Iceland. Constraints on the outflow from the Mediterranean
are being produced using an inverse model in
a box around the outflow, with initial results
showing almost all flow exiting to the north and
west despite the prevalence of Meddies to the
southwest (D. Slater and H. L. Bryden, personal
communication).
Work on global inversions will continue with
the complete WOCE data set, providing improved
estimates of air–sea fluxes, diapycnal velocities,
and complete circulation schemes. These inversions are of increasing complexity, which in its
limit is equivalent to state estimation. For instance,
a group at the Alfred-Wegener-Institut in Germany
is using a number of complementary inversion and
data assimilation techniques to study the mean
and time-varying circulation of the Atlantic. Continuing section and basin-oriented work will contribute to further improvement of the global-scale
work. Results may also eventually be used to test
basic hypotheses about ocean dynamics.
Basin-scale property mapping as an aid to circulation analyses is included in some of the abovementioned works, particularly that of Reid. The
basin atlases, described below, will include property maps. The potential vorticity field is important for planetary wave propagation (e.g. de Szoeke
and Chelton, 1999) and also provides information
like that of other tracers for circulation. O’Dwyer
and Williams (1997) mapped the global potential
vorticity field using historical data, showing the
zonality of potential vorticity at mid-depth and
large regions of potential vorticity that deviate
from zonal at abyssal depths, including near
SECTION 7 INSIGHTS FOR THE FUTURE
528