using hydrographic data and direct velocity
observations.
Lagrangian measurements
Surface drifters have revealed large-scale convergences and divergences as well as the average
‘Ekman’ layer flow. Using surface drifters drogued
at 15 m, Ralph and Niiler (1999) produced the
first large-scale map of the ageostrophic component of near-surface flow and a complete Ekman
spiral, providing a basis in observations over a
very large region for one of the fundamental tenets
of large-scale oceanography, that of Ekman-like
response to the winds. They were able to show
that the ageostrophic component is proportional
to wind speed rather than wind stress, and
inversely proportional to the square root of the
Coriolis parameter. Their result has been confirmed with a much larger data set and the results
are shown in Figure 4.1.9 (Niiler, Chapter 4.1).
Subsurface floats were deployed on a global
scale for the first time during WOCE and are a
major element of WOCE circulation mapping.
Floats have already revealed or validated major
circulation patterns in many regions (Davis and
Zenk, Chapter 3.2). Results from this new and
growing global data set are being produced at an
increasing rate. The following are a few examples.
In the tropical Pacific, Davis (1998b) has shown
vigorous zonal circulation with a significant seasonal component throughout the tropical region in
strong contrast with the gyre circulations of the
mid-latitude South Pacific. He found general
agreement with the mid-latitude circulation patterns produced by Reid (1997) based on geostrophy and careful selection of reference velocities
based on tracers.
In the northern North Atlantic, where subsurface
float density exceeds that in any other region,
Lavender et al. (2000c) have shown that the cyclonic
boundary circulation of the Labrador Sea is accompanied by a weak but permanent anticyclonic circulation offshore of the boundary circulation (Fig.
3.2.8, see Plate 3.2.8, p. 172), with potential major
ramifications for the location of mid-depth convection producing Labrador Sea Water. The stability
of the North Atlantic Current and its meanders
along the western boundary and of its eastward
turn as it separates have been demonstrated with
subsurface floats (Rossby, 1996). In the South
Atlantic, Davis et al. (1996) used floats to examine
the validity of general circulation model results and
showed the importance of eddies in the confluence
of the Brazil and Falkland Currents. Boebel et al.
(1997) used floats and other direct velocity data
sets to show the westward flow of intermediate
waters towards the western boundary and bifurcation of the flow at the coast (Fig. 3.2.7), confirming
that the source of intermediate waters to the tropics
comes from within the subtropical gyre rather than
from the western boundary.
Geostrophic circulation analyses
Full assimilation of hydrographic and velocity data
with models is rapidly developing (see Section
7.1.4 below). However, methods employing
hydrographic observations short of full assimilation are and have been used fruitfully, including
geostrophic velocity reference level choices based
on geostrophic shear and water mass properties
(as in McCartney, 1992; Reid, 1994), box inverse
models (as in Wunsch, 1978), beta-spiral methods
applied over a large area (as in Olbers et al.,
1985), and grid-based adjoint inverses (as in
Schlitzer, 1995). These methods can also incorporate direct velocity measurements as constraints or
as initial velocity choices.
The first global box inverse models have been
produced recently, using zonal, pre-WOCE, hydrographic sections (Macdonald and Wunsch, 1996;
Macdonald, 1998) and with WOCE data including nutrients (Ganachaud and Wunsch, 2001).
Global mass transports from these two efforts are
shown in Figures 7.1.1 and 7.1.2 (see Plates 7.1.1
and 7.1.2, p. 588). The deep water formation rate
for the North Atlantic in both works is comparable to earlier estimates. Macdonald tested the sensitivity of the global inverse to variations in
Pacific–Indian transport through the Indonesian
archipelago, showing that the region of influence
of the throughflow is limited to the Indian and
South Pacific Ocean, without a significant effect on
the global over-turning cell magnitude. Ganachaud
and Wunsch’s new result of a bottom water formation rate of 22 Sv (1 Sverdrup:10
6 m
3 s
91
) for
the Southern Ocean is made possible by the new
circumpolar hydrographic observations, and indicates that the higher range suggested by Schmitz
(1995) is likely valid.
On basin scales, analyses including WOCE data
are well underway. Perusal of the abstracts from
any of the WOCE basin workshops indicates the
7.1 Towards a WOCE Synthesis
527
Talley, Stammer and Fukumori
observations.
Lagrangian measurements
Surface drifters have revealed large-scale convergences and divergences as well as the average
‘Ekman’ layer flow. Using surface drifters drogued
at 15 m, Ralph and Niiler (1999) produced the
first large-scale map of the ageostrophic component of near-surface flow and a complete Ekman
spiral, providing a basis in observations over a
very large region for one of the fundamental tenets
of large-scale oceanography, that of Ekman-like
response to the winds. They were able to show
that the ageostrophic component is proportional
to wind speed rather than wind stress, and
inversely proportional to the square root of the
Coriolis parameter. Their result has been confirmed with a much larger data set and the results
are shown in Figure 4.1.9 (Niiler, Chapter 4.1).
Subsurface floats were deployed on a global
scale for the first time during WOCE and are a
major element of WOCE circulation mapping.
Floats have already revealed or validated major
circulation patterns in many regions (Davis and
Zenk, Chapter 3.2). Results from this new and
growing global data set are being produced at an
increasing rate. The following are a few examples.
In the tropical Pacific, Davis (1998b) has shown
vigorous zonal circulation with a significant seasonal component throughout the tropical region in
strong contrast with the gyre circulations of the
mid-latitude South Pacific. He found general
agreement with the mid-latitude circulation patterns produced by Reid (1997) based on geostrophy and careful selection of reference velocities
based on tracers.
In the northern North Atlantic, where subsurface
float density exceeds that in any other region,
Lavender et al. (2000c) have shown that the cyclonic
boundary circulation of the Labrador Sea is accompanied by a weak but permanent anticyclonic circulation offshore of the boundary circulation (Fig.
3.2.8, see Plate 3.2.8, p. 172), with potential major
ramifications for the location of mid-depth convection producing Labrador Sea Water. The stability
of the North Atlantic Current and its meanders
along the western boundary and of its eastward
turn as it separates have been demonstrated with
subsurface floats (Rossby, 1996). In the South
Atlantic, Davis et al. (1996) used floats to examine
the validity of general circulation model results and
showed the importance of eddies in the confluence
of the Brazil and Falkland Currents. Boebel et al.
(1997) used floats and other direct velocity data
sets to show the westward flow of intermediate
waters towards the western boundary and bifurcation of the flow at the coast (Fig. 3.2.7), confirming
that the source of intermediate waters to the tropics
comes from within the subtropical gyre rather than
from the western boundary.
Geostrophic circulation analyses
Full assimilation of hydrographic and velocity data
with models is rapidly developing (see Section
7.1.4 below). However, methods employing
hydrographic observations short of full assimilation are and have been used fruitfully, including
geostrophic velocity reference level choices based
on geostrophic shear and water mass properties
(as in McCartney, 1992; Reid, 1994), box inverse
models (as in Wunsch, 1978), beta-spiral methods
applied over a large area (as in Olbers et al.,
1985), and grid-based adjoint inverses (as in
Schlitzer, 1995). These methods can also incorporate direct velocity measurements as constraints or
as initial velocity choices.
The first global box inverse models have been
produced recently, using zonal, pre-WOCE, hydrographic sections (Macdonald and Wunsch, 1996;
Macdonald, 1998) and with WOCE data including nutrients (Ganachaud and Wunsch, 2001).
Global mass transports from these two efforts are
shown in Figures 7.1.1 and 7.1.2 (see Plates 7.1.1
and 7.1.2, p. 588). The deep water formation rate
for the North Atlantic in both works is comparable to earlier estimates. Macdonald tested the sensitivity of the global inverse to variations in
Pacific–Indian transport through the Indonesian
archipelago, showing that the region of influence
of the throughflow is limited to the Indian and
South Pacific Ocean, without a significant effect on
the global over-turning cell magnitude. Ganachaud
and Wunsch’s new result of a bottom water formation rate of 22 Sv (1 Sverdrup:10
6 m
3 s
91
) for
the Southern Ocean is made possible by the new
circumpolar hydrographic observations, and indicates that the higher range suggested by Schmitz
(1995) is likely valid.
On basin scales, analyses including WOCE data
are well underway. Perusal of the abstracts from
any of the WOCE basin workshops indicates the
7.1 Towards a WOCE Synthesis
527
Talley, Stammer and Fukumori