the Indian Ocean exhibit large-scale changes, with
elevated sea levels in the eastern Pacific Ocean (the
maximum reached over 30 cm in December 1997)
and the western Indian Ocean.
Chambers et al. (1999) computed the empirical
orthogonal functions for the T/P data during
1992–98 after removing an annual and a semiannual sinusoid as well as a linear trend. They
found that the ENSO-related variability (appearing as the leading mode) in the Indian and Pacific
Oceans are highly correlated with no significant
phase lags. They also demonstrated that the wind
anomalies in the western tropical Pacific and the
eastern Indian Ocean were significantly correlated
with each other and opposite in direction. As the
wind anomalies (westerly anomaly in the western
Pacific and easterly anomaly in the eastern Indian
Ocean) became intensified during the initial phase
of the 1997–98 El Niño, downwelling (thermocline depressed/sea level elevated) Kelvin waves
started propagating eastward in the Pacific,
whereas downwelling Rossby waves started propagating westward in the Indian Ocean. Such
processes led to the buildup of the high sea levels
in the western Indian Ocean and the eastern
Pacific Ocean during the peak of El Niño. Refer to
Section 3.3.3.3 for the roles of the Kelvin and
Rossby waves in the formation of El Niño in the
Pacific.
After El Niño evolved into La Niña in mid-1998,
the entire North Pacific Ocean became progressively abnormal. In 1999, the western part of
the basin was characterized by a large pool of high
sea level north of New Guinea and west of the
Philippines, as well as by a series of bands of high
sea level extending from the western boundary of
the basin to the central basin from subtropical to
high latitudes. The two bands of high sea level emanating from east of Japan are most notable. The
eastern part of the basin exhibits low sea levels from
the Gulf of Alaska to the southern tip of Baja California. This strip of low sea level connects to the
huge pool of low sea level associated with La Niña
in the central equatorial Pacific. This pattern of sealevel change has drawn significant attention from
the climate research community, because it bears a
strong resemblance to a phase of ENSO-like, longterm variability of the Pacific Ocean (Zhang et al.,
1997). The record is obviously too short to lead to
any conclusions, but something apparently very
interesting has been occurring in the Pacific Ocean.
The 1997–99 ENSO event is probably the most
intensively observed to date in comparison with
previous events. Satellite altimetry has played a
significant role in providing a unique global perspective to complement in-situ observations, which
are mostly concentrated in the equatorial regions. In
fact, T/P data have been incorporated into the data
stream used by NOAA’s National Center for Environment Prediction for short-term climate forecast.
The merits of altimeter data in improving the skills
of prediction models have been demonstrated in
hindcast experiments and reported in Ji et al. (2000).
There is evidence that ENSO also affects the
remote Southern Ocean (Peterson and White, 1998)
via the so-called Antarctic Circumpolar Wave,
which is an eastward ocean–atmosphere-coupled
wave circling around Antarctica. This wave is
originated in the western subtropical South Pacific
Ocean due to ENSO activities. This wave creates
a slow oceanic teleconnection mechanism that
spreads the influence of ENSO to other ocean
basins 6–8 years later. Jacobs and Mitchell (1996)
showed earlier evidence of such a wave in the relatively short GEOSAT altimeter data. Using the
much longer record of T/P data, W. White (personal communication) found clear evidence of this
wave pattern.
The long duration and high accuracy of the
T/P data record have created opportunities to study
the interaction of variabilities of different spatial
and temporal scales. A number of studies have
showed evidence of transfer of energy between eddy
and gyre scales. The seasonal and interannual variability of gyres, boundary currents, and their associated eddy fields is apparently linked to the exchange
of energy among them. The readers are referred to
Sections 3.3.4.4 and 3.3.4.6 for a discussion.
3.3.3.7 Global mean-sea-level change
Sea-level variability at the largest scale is the variation of the global mean sea level. There is a high
level of interest in mean-sea-level change because it
is a clear indicator of the consequence of global
warming through the melting of polar ice and
mountain glaciers as well as the thermal expansion
of seawater. Estimation of mean-sea-level change
in the past was based on tide gauge observations
(Douglas, 1991). However, the sparse distribution
of tide gauges presents a serious sampling problem
that requires long averaging time to reduce the
sampling errors. The global coverage of altimetry
SECTION 3 NEW WAYS OF OBSERVING THE OCEAN
160
elevated sea levels in the eastern Pacific Ocean (the
maximum reached over 30 cm in December 1997)
and the western Indian Ocean.
Chambers et al. (1999) computed the empirical
orthogonal functions for the T/P data during
1992–98 after removing an annual and a semiannual sinusoid as well as a linear trend. They
found that the ENSO-related variability (appearing as the leading mode) in the Indian and Pacific
Oceans are highly correlated with no significant
phase lags. They also demonstrated that the wind
anomalies in the western tropical Pacific and the
eastern Indian Ocean were significantly correlated
with each other and opposite in direction. As the
wind anomalies (westerly anomaly in the western
Pacific and easterly anomaly in the eastern Indian
Ocean) became intensified during the initial phase
of the 1997–98 El Niño, downwelling (thermocline depressed/sea level elevated) Kelvin waves
started propagating eastward in the Pacific,
whereas downwelling Rossby waves started propagating westward in the Indian Ocean. Such
processes led to the buildup of the high sea levels
in the western Indian Ocean and the eastern
Pacific Ocean during the peak of El Niño. Refer to
Section 3.3.3.3 for the roles of the Kelvin and
Rossby waves in the formation of El Niño in the
Pacific.
After El Niño evolved into La Niña in mid-1998,
the entire North Pacific Ocean became progressively abnormal. In 1999, the western part of
the basin was characterized by a large pool of high
sea level north of New Guinea and west of the
Philippines, as well as by a series of bands of high
sea level extending from the western boundary of
the basin to the central basin from subtropical to
high latitudes. The two bands of high sea level emanating from east of Japan are most notable. The
eastern part of the basin exhibits low sea levels from
the Gulf of Alaska to the southern tip of Baja California. This strip of low sea level connects to the
huge pool of low sea level associated with La Niña
in the central equatorial Pacific. This pattern of sealevel change has drawn significant attention from
the climate research community, because it bears a
strong resemblance to a phase of ENSO-like, longterm variability of the Pacific Ocean (Zhang et al.,
1997). The record is obviously too short to lead to
any conclusions, but something apparently very
interesting has been occurring in the Pacific Ocean.
The 1997–99 ENSO event is probably the most
intensively observed to date in comparison with
previous events. Satellite altimetry has played a
significant role in providing a unique global perspective to complement in-situ observations, which
are mostly concentrated in the equatorial regions. In
fact, T/P data have been incorporated into the data
stream used by NOAA’s National Center for Environment Prediction for short-term climate forecast.
The merits of altimeter data in improving the skills
of prediction models have been demonstrated in
hindcast experiments and reported in Ji et al. (2000).
There is evidence that ENSO also affects the
remote Southern Ocean (Peterson and White, 1998)
via the so-called Antarctic Circumpolar Wave,
which is an eastward ocean–atmosphere-coupled
wave circling around Antarctica. This wave is
originated in the western subtropical South Pacific
Ocean due to ENSO activities. This wave creates
a slow oceanic teleconnection mechanism that
spreads the influence of ENSO to other ocean
basins 6–8 years later. Jacobs and Mitchell (1996)
showed earlier evidence of such a wave in the relatively short GEOSAT altimeter data. Using the
much longer record of T/P data, W. White (personal communication) found clear evidence of this
wave pattern.
The long duration and high accuracy of the
T/P data record have created opportunities to study
the interaction of variabilities of different spatial
and temporal scales. A number of studies have
showed evidence of transfer of energy between eddy
and gyre scales. The seasonal and interannual variability of gyres, boundary currents, and their associated eddy fields is apparently linked to the exchange
of energy among them. The readers are referred to
Sections 3.3.4.4 and 3.3.4.6 for a discussion.
3.3.3.7 Global mean-sea-level change
Sea-level variability at the largest scale is the variation of the global mean sea level. There is a high
level of interest in mean-sea-level change because it
is a clear indicator of the consequence of global
warming through the melting of polar ice and
mountain glaciers as well as the thermal expansion
of seawater. Estimation of mean-sea-level change
in the past was based on tide gauge observations
(Douglas, 1991). However, the sparse distribution
of tide gauges presents a serious sampling problem
that requires long averaging time to reduce the
sampling errors. The global coverage of altimetry
SECTION 3 NEW WAYS OF OBSERVING THE OCEAN
160