by the Monsoon Index (MOI), which can be
defined as the sea-level pressures between Nemuro,
Japan, and Irkutsk, Russia (Hanawa et al., 1988b;
Watanabe, 1990). Suga and Hanawa (1995c)
demonstrated a good correlation between the MOI
and the potential vorticity and apparent oxygen
utilization of NPSTMW along the repeated 137°E
section. This means that severe wintertime cooling
in the formation region causes the high formation
rate of colder and oxygen-rich NPSTMW. Hanawa
and Yoritaka (2000) described the year-to-year
variation of NPSTMW core layer temperatures
using XBT data taken along 32°30ЈN once a year
from 1990 to 1993. Over the data period of 4
years, core layer temperatures in the western part
of the formation area between 140°E and 160°E
substantially changed, with a range of 0.7°C,
apparently associated with the MOI. These changes
were confirmed by Taneda et al. (2000) using data
taken around the Ogasawara Islands, which is the
downstream site of the Kuroshio recirculation.
Yasuda and Hanawa (1999) clarified the above
change in temperature and area distribution by
making composites of water properties with respect
to the MOI.
A local interannual signal of NPSTMW in the
region south of Honshu is due to the existence or
disappearance of the large meander of the
Kuroshio (Suga and Hanawa, 1995a). When the
Kuroshio large meander is present, NPSTMW is
blocked from advecting westward by the Kuroshio
recirculation into the region south of Japan. This
is also confirmed by the repeat hydrographic
observations at 137°E (Suga and Hanawa, 1995c).
For the South Pacific STMW, as mentioned in
the previous subsection, Roemmich and Cornuelle
(1992) described variations over 5 years using the
first WOCE high-density XBT survey. However,
the time series was too short to determine if the
time scale of the variations was actually interannual or longer.
An interannual signal for the North Atlantic
STMW (Eighteen Degree Water) has not been
sought using long time series, since the record
appears to be dominated by decadal variation (see
below). Klein and Hogg (1996) described the variation of water properties and formation rate from
1987 to 1990 using moored data as well as XBT
data in the eastern part of the recirculation region
of the Gulf Stream. They found that the severe
winter of 1988 led to the formation of a colder
type of mode water, followed by increasing temperature after the mild winter of 1989. They also
pointed out that since the relationship between
winter forcing and water properties was qualitative, advection could strongly influence the water
properties.
5.4.4.3 Decadal variations and abrupt shifts
The North Atlantic Eighteen Degree Water has
demonstrable decadal variations, possibly associated with variations in the North Atlantic Oscillation. Using Sea Surface Temperature (SST) records
in the Eighteen Degree Water formation area and
at Bermuda and Ocean Weather Station E, Fieux
and Stommel (1975) described a warming of late
winter SST from 1910–20 to a maximum in 1950–
60, and thereafter cooling. Talley and Raymer
(1982), Jenkins (1982) and Talley (1996b) documented the variability of Eighteen Degree Water
core and isopycnal properties using the hydrographic time series at Bermuda, which was started
in 1954 (Figure 5.4.4, see Plate 5.4.4, p. 428; from
Talley, 1996b). Joyce and Robbins (1996) and
Talley (1996b) continued the Bermuda time series
analysis, showing that salinity changes on isopycnals in the upper ocean at Bermuda, including
through the Eighteen Degree Water core, correlate
well with the North Atlantic Oscillation (NAO)
index. The Eighteen Degree Water was strong in
three periods – 1956–57, the late 1960s and the
early 1980s – coinciding with low NAO. The
mode water was weakest from 1975 to 1978, just
following a year with high NAO and low surface
salinity. Looking at potential vorticity as a marker
of the core of Eighteen Degree Water, it appears
that the core density, rather than being somewhat
cyclic like isopycnic salinity, changes discretely.
Such changes occurred around 1965, 1975, 1980
and 1985–87. During the intervening periods, the
density was rather stable, with the exception of the
very weak STMW period of 1975–78. The density
of the core for these periods has ranged from 26.4
to 26.55 .
Dickson et al. (1996) showed that SST in the
Sargasso Sea (Eighteen Degree Water formation
area) is out of phase on a decadal scale with SST
in the Labrador Sea and in phase with SST in the
Greenland Sea (see also Dickson et al., Chapter
7.3, in particular Fig. 7.3.5). Thus formation
of Eighteen Degree Water is out of phase with
convective formation of intermediate water in the
5.4 Mode Waters
385
Hanawa and Talley
defined as the sea-level pressures between Nemuro,
Japan, and Irkutsk, Russia (Hanawa et al., 1988b;
Watanabe, 1990). Suga and Hanawa (1995c)
demonstrated a good correlation between the MOI
and the potential vorticity and apparent oxygen
utilization of NPSTMW along the repeated 137°E
section. This means that severe wintertime cooling
in the formation region causes the high formation
rate of colder and oxygen-rich NPSTMW. Hanawa
and Yoritaka (2000) described the year-to-year
variation of NPSTMW core layer temperatures
using XBT data taken along 32°30ЈN once a year
from 1990 to 1993. Over the data period of 4
years, core layer temperatures in the western part
of the formation area between 140°E and 160°E
substantially changed, with a range of 0.7°C,
apparently associated with the MOI. These changes
were confirmed by Taneda et al. (2000) using data
taken around the Ogasawara Islands, which is the
downstream site of the Kuroshio recirculation.
Yasuda and Hanawa (1999) clarified the above
change in temperature and area distribution by
making composites of water properties with respect
to the MOI.
A local interannual signal of NPSTMW in the
region south of Honshu is due to the existence or
disappearance of the large meander of the
Kuroshio (Suga and Hanawa, 1995a). When the
Kuroshio large meander is present, NPSTMW is
blocked from advecting westward by the Kuroshio
recirculation into the region south of Japan. This
is also confirmed by the repeat hydrographic
observations at 137°E (Suga and Hanawa, 1995c).
For the South Pacific STMW, as mentioned in
the previous subsection, Roemmich and Cornuelle
(1992) described variations over 5 years using the
first WOCE high-density XBT survey. However,
the time series was too short to determine if the
time scale of the variations was actually interannual or longer.
An interannual signal for the North Atlantic
STMW (Eighteen Degree Water) has not been
sought using long time series, since the record
appears to be dominated by decadal variation (see
below). Klein and Hogg (1996) described the variation of water properties and formation rate from
1987 to 1990 using moored data as well as XBT
data in the eastern part of the recirculation region
of the Gulf Stream. They found that the severe
winter of 1988 led to the formation of a colder
type of mode water, followed by increasing temperature after the mild winter of 1989. They also
pointed out that since the relationship between
winter forcing and water properties was qualitative, advection could strongly influence the water
properties.
5.4.4.3 Decadal variations and abrupt shifts
The North Atlantic Eighteen Degree Water has
demonstrable decadal variations, possibly associated with variations in the North Atlantic Oscillation. Using Sea Surface Temperature (SST) records
in the Eighteen Degree Water formation area and
at Bermuda and Ocean Weather Station E, Fieux
and Stommel (1975) described a warming of late
winter SST from 1910–20 to a maximum in 1950–
60, and thereafter cooling. Talley and Raymer
(1982), Jenkins (1982) and Talley (1996b) documented the variability of Eighteen Degree Water
core and isopycnal properties using the hydrographic time series at Bermuda, which was started
in 1954 (Figure 5.4.4, see Plate 5.4.4, p. 428; from
Talley, 1996b). Joyce and Robbins (1996) and
Talley (1996b) continued the Bermuda time series
analysis, showing that salinity changes on isopycnals in the upper ocean at Bermuda, including
through the Eighteen Degree Water core, correlate
well with the North Atlantic Oscillation (NAO)
index. The Eighteen Degree Water was strong in
three periods – 1956–57, the late 1960s and the
early 1980s – coinciding with low NAO. The
mode water was weakest from 1975 to 1978, just
following a year with high NAO and low surface
salinity. Looking at potential vorticity as a marker
of the core of Eighteen Degree Water, it appears
that the core density, rather than being somewhat
cyclic like isopycnic salinity, changes discretely.
Such changes occurred around 1965, 1975, 1980
and 1985–87. During the intervening periods, the
density was rather stable, with the exception of the
very weak STMW period of 1975–78. The density
of the core for these periods has ranged from 26.4
to 26.55 .
Dickson et al. (1996) showed that SST in the
Sargasso Sea (Eighteen Degree Water formation
area) is out of phase on a decadal scale with SST
in the Labrador Sea and in phase with SST in the
Greenland Sea (see also Dickson et al., Chapter
7.3, in particular Fig. 7.3.5). Thus formation
of Eighteen Degree Water is out of phase with
convective formation of intermediate water in the
5.4 Mode Waters
385
Hanawa and Talley