There, the vertical mixing of waters of different
temperature does occur along the Somali coast
during the southwest monsoon season, and in the
Arabian Sea (McCreary and Kundu, 1988). The
processes that maintain the thermocline in the
Indian Ocean need to be explored.
The tropical–extratropical exchanges in the
Atlantic Ocean differ from those in the Pacific in
being far more asymmetric relative to the equator.
Water that subducts off southwestern Africa does
reach the equatorial thermocline, but the water
that subducts north of the equator in the Atlantic
does not (Fine et al., 1987, 1994; Harper, 2000).
Consistent with this result is the tritium distribution, which shows a sharp front on the southern
boundary of the North Atlantic subtropical gyre
(Sarmiento et al., 1982; Kawase and Sarmiento,
1985). This strong asymmetry relative to the equator in the Atlantic could have several reasons. One
is the thermohaline circulation in which there is a
net northward flow of warm surface waters across
the thermocline (Schmitz and McCartney, 1993).
Another possible factor is the coastal geometry,
i.e. the small zonal extent of the Atlantic just north
of the equator. Waters that subduct off northwestern Africa are likely to join northward-flowing
western boundary currents before they can travel
far south towards the equator.
The dynamics of subduction temperature variability remain to be studied. Up to now, studies of
tropical–extratropical exchanges have focused on
water masses that amount to passive tracers. Little
attention has thus far been given to the fate of
temperature anomalies that subduct in the subtropics. (These anomalies have been described in
recent observations: Deser et al., 1996; Zhang
et al., 1998a; Schneider et al., 1999b.) Part of a
temperature anomaly could be salinity-compensated with no density signature (Miller et al.,
1998). This part of a salinity-compensated temperature anomaly should behave exactly like a passive
tracer, whose evolution is determined by mean
subduction flow and small-scale mixing. It is possible that a temperature anomaly is generated to
be accompanied by a compensation salinity. For
example, an anomalous westerly wind in the midlatitude could generate both a cold surface temperature anomaly and a low surface salinity anomaly
by inducing a southward Ekman flow advection.
The subtropical cold surface temperature anomaly,
by reducing evaporation, could further freshen the
surface water and therefore compensate an additional part of the cold temperature anomaly. However, a sea surface temperature anomaly usually
will not be completely salinity-compensated.
It is more likely that a large part of a sea surface
temperature anomaly in the subtropics is not
salinity-compensated. The non-compensated temperature anomaly will induce density disturbances
that propagate as dynamic waves. On interannual
and shorter time scales, the Rossby waves that are
of primary importance are refracted equatorward
(Schopf et al., 1981; Chang and Philander, 1989).
On longer time scales, the subduction anomalies
can be understood in terms of higher baroclinic
modes of planetary waves that tend to be advected
by the mean current (Liu, 1999a,b). The evolution
of these planetary waves are controlled by wave
dynamics as well as mean advection and mixing,
and therefore could differ significantly from a passive tracer. Preliminary studies suggest that, in the
subtropics, the amplitude of a density anomaly
could decrease along the subduction pathway due
to the divergent group velocity (Liu, 1993), or
increase in the far downstream of a subtropical gyre
due to planetary wave instability (Liu, 1999a). Furthermore, the speed of the density anomaly is usually somewhat slower than the mean flow, because
of the northward propulsion by the eastward mean
potential vorticity gradient in the ventilated zone (in
the northern hemisphere) (Liu, 1999b; Stephens
et al., 2000). A subducted temperature anomaly
also has a more complex spatial structure than a
subducted passive tracer anomaly (Liu and Shin,
1999). In addition, the subduction density anomaly
has distinctively different dynamic responses to the
surface wind stress as well as buoyancy forcing
(Huang and Pedlosky, 1999; Liu, 1999b; Schneider
et al., 1999b). Finally, if density disturbances reach
a western coast, they can generate coastal Kelvin
waves that propagate to the equator (Lysne et al.,
1997; Liu et al., 1999b; Shin and Liu, 2000).
The climate implication of tropical–extratropical water exchange remains speculative at the
present stage. The exchange is recognized to be
critical in affecting the mean equatorial thermocline, and in turn the surface temperature through
equatorial upwelling. The influence of tropical–
extratropical exchange on the ocean–atmosphere
interactions and climate variability remains yet to
be quantified. These resultant climate fluctuations
have only recently been given attention.
SECTION 4 THE GLOBAL FLOW FIELD
256
temperature does occur along the Somali coast
during the southwest monsoon season, and in the
Arabian Sea (McCreary and Kundu, 1988). The
processes that maintain the thermocline in the
Indian Ocean need to be explored.
The tropical–extratropical exchanges in the
Atlantic Ocean differ from those in the Pacific in
being far more asymmetric relative to the equator.
Water that subducts off southwestern Africa does
reach the equatorial thermocline, but the water
that subducts north of the equator in the Atlantic
does not (Fine et al., 1987, 1994; Harper, 2000).
Consistent with this result is the tritium distribution, which shows a sharp front on the southern
boundary of the North Atlantic subtropical gyre
(Sarmiento et al., 1982; Kawase and Sarmiento,
1985). This strong asymmetry relative to the equator in the Atlantic could have several reasons. One
is the thermohaline circulation in which there is a
net northward flow of warm surface waters across
the thermocline (Schmitz and McCartney, 1993).
Another possible factor is the coastal geometry,
i.e. the small zonal extent of the Atlantic just north
of the equator. Waters that subduct off northwestern Africa are likely to join northward-flowing
western boundary currents before they can travel
far south towards the equator.
The dynamics of subduction temperature variability remain to be studied. Up to now, studies of
tropical–extratropical exchanges have focused on
water masses that amount to passive tracers. Little
attention has thus far been given to the fate of
temperature anomalies that subduct in the subtropics. (These anomalies have been described in
recent observations: Deser et al., 1996; Zhang
et al., 1998a; Schneider et al., 1999b.) Part of a
temperature anomaly could be salinity-compensated with no density signature (Miller et al.,
1998). This part of a salinity-compensated temperature anomaly should behave exactly like a passive
tracer, whose evolution is determined by mean
subduction flow and small-scale mixing. It is possible that a temperature anomaly is generated to
be accompanied by a compensation salinity. For
example, an anomalous westerly wind in the midlatitude could generate both a cold surface temperature anomaly and a low surface salinity anomaly
by inducing a southward Ekman flow advection.
The subtropical cold surface temperature anomaly,
by reducing evaporation, could further freshen the
surface water and therefore compensate an additional part of the cold temperature anomaly. However, a sea surface temperature anomaly usually
will not be completely salinity-compensated.
It is more likely that a large part of a sea surface
temperature anomaly in the subtropics is not
salinity-compensated. The non-compensated temperature anomaly will induce density disturbances
that propagate as dynamic waves. On interannual
and shorter time scales, the Rossby waves that are
of primary importance are refracted equatorward
(Schopf et al., 1981; Chang and Philander, 1989).
On longer time scales, the subduction anomalies
can be understood in terms of higher baroclinic
modes of planetary waves that tend to be advected
by the mean current (Liu, 1999a,b). The evolution
of these planetary waves are controlled by wave
dynamics as well as mean advection and mixing,
and therefore could differ significantly from a passive tracer. Preliminary studies suggest that, in the
subtropics, the amplitude of a density anomaly
could decrease along the subduction pathway due
to the divergent group velocity (Liu, 1993), or
increase in the far downstream of a subtropical gyre
due to planetary wave instability (Liu, 1999a). Furthermore, the speed of the density anomaly is usually somewhat slower than the mean flow, because
of the northward propulsion by the eastward mean
potential vorticity gradient in the ventilated zone (in
the northern hemisphere) (Liu, 1999b; Stephens
et al., 2000). A subducted temperature anomaly
also has a more complex spatial structure than a
subducted passive tracer anomaly (Liu and Shin,
1999). In addition, the subduction density anomaly
has distinctively different dynamic responses to the
surface wind stress as well as buoyancy forcing
(Huang and Pedlosky, 1999; Liu, 1999b; Schneider
et al., 1999b). Finally, if density disturbances reach
a western coast, they can generate coastal Kelvin
waves that propagate to the equator (Lysne et al.,
1997; Liu et al., 1999b; Shin and Liu, 2000).
The climate implication of tropical–extratropical water exchange remains speculative at the
present stage. The exchange is recognized to be
critical in affecting the mean equatorial thermocline, and in turn the surface temperature through
equatorial upwelling. The influence of tropical–
extratropical exchange on the ocean–atmosphere
interactions and climate variability remains yet to
be quantified. These resultant climate fluctuations
have only recently been given attention.
SECTION 4 THE GLOBAL FLOW FIELD
256