to months. Thus (for instance) a southward mean
flow at 1000 m that falls with distance offshore
from 0.08 m s
91 at the inshore mooring may not
be significant. However, combining two years of
data near the 2000 m isobath yields an alongshore
component of 0.041<0.029 m s
91 to the southwest, with an onshore flow of 0.029<0.013 m s
91
.
Annual means at 1500 m, 2000 m and 3000 m
are nearly zero. However, a strong seasonal
flow occurs at 3000 m: 0.103<0.018 m s
91 to the
northeast in summer (June to September),
0.066<0.027 m s
91 to the southwest in winter.
Wacongne and Pacanowski (1996) compared these
currents with their model results and suggested
that this deep seasonal reversal may indicate a
mass-conserving response to meridional Ekman
transports, which are southward (northward) in
summer (winter), on both sides of the equator.
Marked differences in salinity patterns on the
1.8°C isotherm (about 2600 m) under the length of
the Somali Current were found when data from
August 1964 (Warren et al., 1966) were compared
with surveys in April 1985 and April 1986 (Fieux
et al., 1986; Schott et al., 1989). These differences
are consistent with the previously mentioned seasonal reversal observed at 3000 m on the equator
under the Somali Current in 1984–85 (Schott et
al., 1989). However, regional hydrographic surveys spanning December 1986–January 1987 and
repeated in July–August 1987 failed to reveal a
seasonal cycle in deep salinity or dissolved oxygen
distributions (Warren and Johnson, 1992). This
difference may have been owing to the remarkably
weak southwest monsoon in 1987, which could
have been insufficiently strong to reverse a presumed prevailing southwestward flow along the
western boundary.
Below 3000 m, water mass properties and
geostrophy do indicate some mean flow in the
equatorial regime, at least within the western deep
basin system. The very narrow Amirante Passage
(9°S, 52°E) admits a bottom water flow northward
into the Somali Basin with transport estimates
ranging from 1<0.5 Sv based on two station pairs
(Fieux and Swallow, 1988) to 4<1 Sv based on
three short sections (Barton and Hill, 1989). The
most recent estimate based on an analysis of
WOCE and historical hydrographic data suggests
1.0–1.7 Sv (Johnson et al., 1998). This flow apparently continues north to at least 3°S along the
western boundary, but may turn eastward at the
equator (Johnson et al., 1991a; Warren and
Johnson, 1992), possibly consistent with inviscid
boundary current dynamics (Johnson, 1993). The
heat budget for the bottom water of the Somali
Basin gives an estimate of vertical eddy diffusivity
from 2.5 to 1010
94 m
2 s
91 (Fieux and Swallow,
1988; Barton and Hill, 1989) with the range
mainly a result of the inflow transport estimate
range. After the bottom water warms through
diabatic processes, at least some of it may return
southward through the Amirante Passage (Johnson
et al., 1998), to complete the deep meridional
overturning cell, although there also is evidence
for northward flow out of the Somali Basin
through the Owen Fracture Zone (10°S, 57°E) to
supply the bottom waters of the Arabian Basin
(Johnson et al., 1991b; Quadfasel et al., 1997).
Luyten and Swallow (1976) were the first to
discover equatorial deep jets in the Indian Ocean.
They were further investigated by Ponte and
Luyten (1990), who found no clear spectral peak
in the zonal velocity, though visual inspection suggests a wavelength of about 500 m (in ‘stretched
metres’, i.e. depth adjusted by a factor of
(N/N 0 )
1/2 , with N the buoyancy frequency, N 0
is the depth-averaged buoyancy frequency of 1
cycle h
91
). Dengler and Quadfasel (2001) examined hydrographic and current profiles along the
80°30ЈE current meter section discussed earlier,
during three cruises. Once low vertical modes were
removed from the data, all three sections showed
largely zonal, vertically alternating jets, with
speeds of up to 0.12 m s
91
. These jets were within
a degree of the equator. Vertical length scales were
about 300 m, lengthening to 1000 m at depth; relatively strong vertical shears occurred between the
jets. The jets were largely reversed in July 1993
and September 1994, relative to December 1990;
this is consistent with the semiannual reversals
found by Luyten and Roemmich (1982) continuing to substantial depths. CTD data showed layers
of low stratification, often 50 m thick, which
coincided with the bands of large vertical shear.
Richardson numbers within these layers were
often less than 0.33, suggestive of active turbulence. This was confirmed by examining shear
spectra (from the Pegasus current profiler) and
strain spectra (from the CTD temperature profiles). Spectral densities were about five times
greater on the equator than at 2° on either side,
implying large internal wave activity with vertical
SECTION 4 THE GLOBAL FLOW FIELD
244
flow at 1000 m that falls with distance offshore
from 0.08 m s
91 at the inshore mooring may not
be significant. However, combining two years of
data near the 2000 m isobath yields an alongshore
component of 0.041<0.029 m s
91 to the southwest, with an onshore flow of 0.029<0.013 m s
91
.
Annual means at 1500 m, 2000 m and 3000 m
are nearly zero. However, a strong seasonal
flow occurs at 3000 m: 0.103<0.018 m s
91 to the
northeast in summer (June to September),
0.066<0.027 m s
91 to the southwest in winter.
Wacongne and Pacanowski (1996) compared these
currents with their model results and suggested
that this deep seasonal reversal may indicate a
mass-conserving response to meridional Ekman
transports, which are southward (northward) in
summer (winter), on both sides of the equator.
Marked differences in salinity patterns on the
1.8°C isotherm (about 2600 m) under the length of
the Somali Current were found when data from
August 1964 (Warren et al., 1966) were compared
with surveys in April 1985 and April 1986 (Fieux
et al., 1986; Schott et al., 1989). These differences
are consistent with the previously mentioned seasonal reversal observed at 3000 m on the equator
under the Somali Current in 1984–85 (Schott et
al., 1989). However, regional hydrographic surveys spanning December 1986–January 1987 and
repeated in July–August 1987 failed to reveal a
seasonal cycle in deep salinity or dissolved oxygen
distributions (Warren and Johnson, 1992). This
difference may have been owing to the remarkably
weak southwest monsoon in 1987, which could
have been insufficiently strong to reverse a presumed prevailing southwestward flow along the
western boundary.
Below 3000 m, water mass properties and
geostrophy do indicate some mean flow in the
equatorial regime, at least within the western deep
basin system. The very narrow Amirante Passage
(9°S, 52°E) admits a bottom water flow northward
into the Somali Basin with transport estimates
ranging from 1<0.5 Sv based on two station pairs
(Fieux and Swallow, 1988) to 4<1 Sv based on
three short sections (Barton and Hill, 1989). The
most recent estimate based on an analysis of
WOCE and historical hydrographic data suggests
1.0–1.7 Sv (Johnson et al., 1998). This flow apparently continues north to at least 3°S along the
western boundary, but may turn eastward at the
equator (Johnson et al., 1991a; Warren and
Johnson, 1992), possibly consistent with inviscid
boundary current dynamics (Johnson, 1993). The
heat budget for the bottom water of the Somali
Basin gives an estimate of vertical eddy diffusivity
from 2.5 to 1010
94 m
2 s
91 (Fieux and Swallow,
1988; Barton and Hill, 1989) with the range
mainly a result of the inflow transport estimate
range. After the bottom water warms through
diabatic processes, at least some of it may return
southward through the Amirante Passage (Johnson
et al., 1998), to complete the deep meridional
overturning cell, although there also is evidence
for northward flow out of the Somali Basin
through the Owen Fracture Zone (10°S, 57°E) to
supply the bottom waters of the Arabian Basin
(Johnson et al., 1991b; Quadfasel et al., 1997).
Luyten and Swallow (1976) were the first to
discover equatorial deep jets in the Indian Ocean.
They were further investigated by Ponte and
Luyten (1990), who found no clear spectral peak
in the zonal velocity, though visual inspection suggests a wavelength of about 500 m (in ‘stretched
metres’, i.e. depth adjusted by a factor of
(N/N 0 )
1/2 , with N the buoyancy frequency, N 0
is the depth-averaged buoyancy frequency of 1
cycle h
91
). Dengler and Quadfasel (2001) examined hydrographic and current profiles along the
80°30ЈE current meter section discussed earlier,
during three cruises. Once low vertical modes were
removed from the data, all three sections showed
largely zonal, vertically alternating jets, with
speeds of up to 0.12 m s
91
. These jets were within
a degree of the equator. Vertical length scales were
about 300 m, lengthening to 1000 m at depth; relatively strong vertical shears occurred between the
jets. The jets were largely reversed in July 1993
and September 1994, relative to December 1990;
this is consistent with the semiannual reversals
found by Luyten and Roemmich (1982) continuing to substantial depths. CTD data showed layers
of low stratification, often 50 m thick, which
coincided with the bands of large vertical shear.
Richardson numbers within these layers were
often less than 0.33, suggestive of active turbulence. This was confirmed by examining shear
spectra (from the Pegasus current profiler) and
strain spectra (from the CTD temperature profiles). Spectral densities were about five times
greater on the equator than at 2° on either side,
implying large internal wave activity with vertical
SECTION 4 THE GLOBAL FLOW FIELD
244