convergence induced by the sea surface temperature anomalies. Such waves are self-sustaining and
propagating westward with a phase speed higher
than that of standard linear free Rossby waves.
The discrepancy between the observed phase speed
from the standard Rossby wave theory can be
largely accounted for by the coupled theory. The
results of White et al. (1998b) should be viewed
with the caveat that only 2 years of data were used
to study a biannual phenomenon. The theory
needs to be re-examined using longer data records,
which are now readily available.
An interesting scenario for the generation of
mid-latitude Rossby waves was demonstrated by
Jacobs et al. (1994). By using GEOSAT data and
an ocean general circulation model, they showed
that ENSO-induced perturbations in the tropics
could propagate poleward along the ocean’s eastern boundaries as coastal trapped waves. These
waves tend to evolve into Rossby waves and transmit the effects of ENSO into the mid-latitude
oceanic interiors. Such tropics–extratropics connection is an important factor affecting the longterm climate variability.
3.3.3.3 Tropical Kelvin and Rossby waves
Large-scale low-frequency oceanic waves play a
key role in the dynamics of the tropical ocean and
its interaction with the atmosphere. These waves
are essentially driven by winds, but the overlying
winds are also affected by the waves through air–
sea coupling mechanisms. The interplay of baroclinic Kelvin waves and Rossby waves has been
proposed to be an important mechanism for sustaining the semioscillatory behaviour of ENSO in
the delayed-action-oscillator theory (Battisti, 1988;
Schopf and Suarez, 1988). Altimetry data provide
a basin-wide perspective to test the validity of the
theory. The early studies with the GEOSAT data
(Delcroix et al., 1991; du Penhoat et al., 1992)
clearly showed the important roles of the Kelvin
and Rossby waves in the 1986–87 El Niño and the
1988–89 La Niña. White et al. (1990b) reported
evidence for the reflection of equatorial Rossby
waves into eastward Kelvin waves at the western
boundary of the tropical Pacific, a key requirement
for the delayed-action-oscillator theory. By projecting sea-level anomalies into Kelvin and Rossby
wave components, Boulanger and Fu (1996) showed
persistent reflection of Kelvin waves into Rossby
waves at the eastern boundary of the Pacific in
1993–95, while the reflection of Rossby waves at
the western boundary was relatively intermittent.
They also suggested that the reflection of Rossby
waves played a role in the onset of the warming of
1994–95.
Using a longer record of T/P data, Boulanger
and Menkes (1999) repeated the calculation of
Boulanger and Fu (1996) to study the properties of
the equatorial Rossby and Kelvin waves from
October 1992 to May 1998. Shown in Fig. 3.3.9
are longitude–time plots of the Kelvin and the first
mode Rossby wave coefficients. Near the western
boundary, the Kelvin waves generated by the
reflection of Rossby waves were highly damped
and rarely propagated past the dateline. Most of
the Kelvin waves east of the dateline were forced
by west wind bursts west of the dateline. In late
1997 during the peak of the 1997–98 El Niño,
however, the strong upwelling Rossby waves (negative sea-level anomalies) originated east of the
dateline created strong reflected Kelvin waves at
the western boundary. These reflected upwelling
Kelvin waves worked against the locally forced
downwelling Kelvin waves and eventually passed
the dateline and reversed the sea-level anomalies
all the way to about 100°W, leading to the demise
of El Niño in May 1998. The Rossby wave reflection into Kelvin waves seems to have played a
significant role in the termination of El Niño, consistent with the delayed-action-oscillator theory.
Near the eastern boundary, reflection of Kelvin
waves into the first mode Rossby waves is quite
persistent. The reflected Rossby waves produce
anomalously westward surface currents that tend
to slow down the eastward progression of the
eastern edge of the warm pool of El Niño. These
waves are thus an impediment to the growth of
El Niño, consistent with the conceptual model of
Picaut et al. (1997).
3.3.3.4 The relation between sea level and
subsurface variability
As noted earlier, a major advantage of altimetry
over other ocean remote sensing is the relation
between observed sea level and subsurface fields.
From the discussions in the preceding sections on
the various mechanisms of sea-level variations, it is
clear that sea-level variations reflect changes in
subsurface fields due to thermal and mechanical
processes. Except for steric change of sea level due
to buoyancy forcing, large-scale sea-level variability
3.3 Ocean Circulation and Variability from Satellite Altimetry
155
Fu
propagating westward with a phase speed higher
than that of standard linear free Rossby waves.
The discrepancy between the observed phase speed
from the standard Rossby wave theory can be
largely accounted for by the coupled theory. The
results of White et al. (1998b) should be viewed
with the caveat that only 2 years of data were used
to study a biannual phenomenon. The theory
needs to be re-examined using longer data records,
which are now readily available.
An interesting scenario for the generation of
mid-latitude Rossby waves was demonstrated by
Jacobs et al. (1994). By using GEOSAT data and
an ocean general circulation model, they showed
that ENSO-induced perturbations in the tropics
could propagate poleward along the ocean’s eastern boundaries as coastal trapped waves. These
waves tend to evolve into Rossby waves and transmit the effects of ENSO into the mid-latitude
oceanic interiors. Such tropics–extratropics connection is an important factor affecting the longterm climate variability.
3.3.3.3 Tropical Kelvin and Rossby waves
Large-scale low-frequency oceanic waves play a
key role in the dynamics of the tropical ocean and
its interaction with the atmosphere. These waves
are essentially driven by winds, but the overlying
winds are also affected by the waves through air–
sea coupling mechanisms. The interplay of baroclinic Kelvin waves and Rossby waves has been
proposed to be an important mechanism for sustaining the semioscillatory behaviour of ENSO in
the delayed-action-oscillator theory (Battisti, 1988;
Schopf and Suarez, 1988). Altimetry data provide
a basin-wide perspective to test the validity of the
theory. The early studies with the GEOSAT data
(Delcroix et al., 1991; du Penhoat et al., 1992)
clearly showed the important roles of the Kelvin
and Rossby waves in the 1986–87 El Niño and the
1988–89 La Niña. White et al. (1990b) reported
evidence for the reflection of equatorial Rossby
waves into eastward Kelvin waves at the western
boundary of the tropical Pacific, a key requirement
for the delayed-action-oscillator theory. By projecting sea-level anomalies into Kelvin and Rossby
wave components, Boulanger and Fu (1996) showed
persistent reflection of Kelvin waves into Rossby
waves at the eastern boundary of the Pacific in
1993–95, while the reflection of Rossby waves at
the western boundary was relatively intermittent.
They also suggested that the reflection of Rossby
waves played a role in the onset of the warming of
1994–95.
Using a longer record of T/P data, Boulanger
and Menkes (1999) repeated the calculation of
Boulanger and Fu (1996) to study the properties of
the equatorial Rossby and Kelvin waves from
October 1992 to May 1998. Shown in Fig. 3.3.9
are longitude–time plots of the Kelvin and the first
mode Rossby wave coefficients. Near the western
boundary, the Kelvin waves generated by the
reflection of Rossby waves were highly damped
and rarely propagated past the dateline. Most of
the Kelvin waves east of the dateline were forced
by west wind bursts west of the dateline. In late
1997 during the peak of the 1997–98 El Niño,
however, the strong upwelling Rossby waves (negative sea-level anomalies) originated east of the
dateline created strong reflected Kelvin waves at
the western boundary. These reflected upwelling
Kelvin waves worked against the locally forced
downwelling Kelvin waves and eventually passed
the dateline and reversed the sea-level anomalies
all the way to about 100°W, leading to the demise
of El Niño in May 1998. The Rossby wave reflection into Kelvin waves seems to have played a
significant role in the termination of El Niño, consistent with the delayed-action-oscillator theory.
Near the eastern boundary, reflection of Kelvin
waves into the first mode Rossby waves is quite
persistent. The reflected Rossby waves produce
anomalously westward surface currents that tend
to slow down the eastward progression of the
eastern edge of the warm pool of El Niño. These
waves are thus an impediment to the growth of
El Niño, consistent with the conceptual model of
Picaut et al. (1997).
3.3.3.4 The relation between sea level and
subsurface variability
As noted earlier, a major advantage of altimetry
over other ocean remote sensing is the relation
between observed sea level and subsurface fields.
From the discussions in the preceding sections on
the various mechanisms of sea-level variations, it is
clear that sea-level variations reflect changes in
subsurface fields due to thermal and mechanical
processes. Except for steric change of sea level due
to buoyancy forcing, large-scale sea-level variability
3.3 Ocean Circulation and Variability from Satellite Altimetry
155
Fu