acceleration of sea-level rise during the twentieth
century. However, an analysis of the few multicentury records indicate an acceleration of
0.3 mm yr
91 century
91 to 0.9 mm yr
91 century
91
(Woodworth, 1999).
One of the difficulties in establishing the globalaverage rate of sea-level change is the poor spatial
distribution of tide gauges. However, satellite
altimetry (Fu, Chapter 3.3) now provides near
global and homogeneous measurements of sea level
with respect to the centre of mass of the earth every
10 days. The most recent estimates of global average sea-level rise are 2.6<0.2 mm yr
91 (Cazenave
et al., 1998), 3.1<1.3 mm yr
91 and 2.5<1.3
mm yr
91 (Nerem, 1999). These estimates used tide
gauge data to remove small time-variable biases in
the TOPEX/POSEIDON altimeter data, and made
allowance for drifts in the water vapour corrections. The uncertainty of <0.2 mm yr
91 for the
Cazenave et al. estimate does not include allowance
for uncertainty in instrumental drift but only
reflects the temporal variations in measured global
sea level. These variations correlate with globalaverage sea surface temperature, perhaps indicating the importance of steric effects through upper
ocean heat storage (Cazenave et al., 1998; Nerem
et al., 1999). The altimeter data suggest a rate of
sea-level rise during the 1990s greater than the
mean rate of rise for much of the twentieth century. Whether this indicates a recent acceleration, is
the result of systematic differences between the two
measurement techniques, or is merely the result of
the short record is not yet clear. With high-quality
in-situ data to support the altimeter observations,
satellite altimeters will in the future be the prime
means of determining global sea-level change.
Components of twentieth and twenty-first century
sea-level rise
One of the major contributors to sea-level change
during the twentieth and twenty-first centuries is
ocean thermal expansion. Warmed surface waters
are advected (subducted) into the ocean interior
(Price, Chapter 5.3) such that the upper part of the
water column is warmed (Sections 1.2.6.4 and
1.2.6.5; Dickson et al., Chapter 7.3). As the ocean
warms, the density decreases and thus the volume
of the ocean increases. Salinity changes within the
ocean have a significant impact on the local density and thus local sea level, but have little effect
on the global-average sea level.
Evidence of large-scale ocean warming and thus
thermal expansion at rates of order 1 mm yr
91 has
come from comparing recent WOCE sections with
historical data and time series stations. The evidence is most convincing for the North Atlantic,
for which the longest records (up to 73 years) and
most complete oceanographic data sets are available (e.g. Parrilla et al., 1994; Joyce and Robbins,
1996; Arhan et al., 1998; Joyce et al., 1999) (Fig.
1.2.11, see Plate 1.2.11, p. 44). The warming also
extends into the South Atlantic (Fig. 7.3.4, Dickson
et al., Chapter 7.3). The only area of substantial
cooling is in the subpolar North Atlantic (Read
and Gould, 1992; Dickson et al., Chapter 7.3).
Observations from the Pacific and Indian
Oceans cover a shorter period and thus estimates
of change are less certain and may result from
decadal variability rather than indicating a longterm trend. However, water mass properties indicate that these basin-scale changes are not solely a
result of vertical thermocline heave but are consistent with surface warming and the resultant subduction of this water into the main thermocline
(Wong et al., 2001). The warming in the Pacific
and Indian Oceans (Bindoff and McDougall,
2000) is confined to the main thermocline (mostly
the upper 1 km) of the subtropical gyres. This contrasts with the North Atlantic, where the warming
is seen at depths of 1000–2000 m. This is consistent with our understanding of the different basins
described earlier – deep convection and larger
meridional overturning in the North Atlantic contrasting with much shallower gyres in the Pacific.
Most recently Levitus et al. (2000) and Antonov
et al. (2001) have estimated global ocean heat
uptake of 10
22 J over the period 1955 to 1995 and
a consequent global average rate of thermal
expansion of 0.55 mm yr
91
.
Realistic simulation of ocean thermal expansion
requires coupled Atmosphere–Ocean General Circulation Models (AOGCMs) of the climate system
(Wood and Bryan, Chapter 2.3). A number of
model simulations of the twentieth century have
recently been completed using realistic greenhouse
gas and aerosol forcings starting in 1900 or earlier.
The results indicate that the average rate of change
due to thermal expansion over the last hundred
years was of the order of 0.3–0.8 mm yr
91
, and
0.6–1.1 mm yr
91 in recent decades, similar to
the observational estimates (Church et al., 2001;
Gregory et al., 2001).
SECTION 1 THE OCEAN AND CLIMATE
28
century. However, an analysis of the few multicentury records indicate an acceleration of
0.3 mm yr
91 century
91 to 0.9 mm yr
91 century
91
(Woodworth, 1999).
One of the difficulties in establishing the globalaverage rate of sea-level change is the poor spatial
distribution of tide gauges. However, satellite
altimetry (Fu, Chapter 3.3) now provides near
global and homogeneous measurements of sea level
with respect to the centre of mass of the earth every
10 days. The most recent estimates of global average sea-level rise are 2.6<0.2 mm yr
91 (Cazenave
et al., 1998), 3.1<1.3 mm yr
91 and 2.5<1.3
mm yr
91 (Nerem, 1999). These estimates used tide
gauge data to remove small time-variable biases in
the TOPEX/POSEIDON altimeter data, and made
allowance for drifts in the water vapour corrections. The uncertainty of <0.2 mm yr
91 for the
Cazenave et al. estimate does not include allowance
for uncertainty in instrumental drift but only
reflects the temporal variations in measured global
sea level. These variations correlate with globalaverage sea surface temperature, perhaps indicating the importance of steric effects through upper
ocean heat storage (Cazenave et al., 1998; Nerem
et al., 1999). The altimeter data suggest a rate of
sea-level rise during the 1990s greater than the
mean rate of rise for much of the twentieth century. Whether this indicates a recent acceleration, is
the result of systematic differences between the two
measurement techniques, or is merely the result of
the short record is not yet clear. With high-quality
in-situ data to support the altimeter observations,
satellite altimeters will in the future be the prime
means of determining global sea-level change.
Components of twentieth and twenty-first century
sea-level rise
One of the major contributors to sea-level change
during the twentieth and twenty-first centuries is
ocean thermal expansion. Warmed surface waters
are advected (subducted) into the ocean interior
(Price, Chapter 5.3) such that the upper part of the
water column is warmed (Sections 1.2.6.4 and
1.2.6.5; Dickson et al., Chapter 7.3). As the ocean
warms, the density decreases and thus the volume
of the ocean increases. Salinity changes within the
ocean have a significant impact on the local density and thus local sea level, but have little effect
on the global-average sea level.
Evidence of large-scale ocean warming and thus
thermal expansion at rates of order 1 mm yr
91 has
come from comparing recent WOCE sections with
historical data and time series stations. The evidence is most convincing for the North Atlantic,
for which the longest records (up to 73 years) and
most complete oceanographic data sets are available (e.g. Parrilla et al., 1994; Joyce and Robbins,
1996; Arhan et al., 1998; Joyce et al., 1999) (Fig.
1.2.11, see Plate 1.2.11, p. 44). The warming also
extends into the South Atlantic (Fig. 7.3.4, Dickson
et al., Chapter 7.3). The only area of substantial
cooling is in the subpolar North Atlantic (Read
and Gould, 1992; Dickson et al., Chapter 7.3).
Observations from the Pacific and Indian
Oceans cover a shorter period and thus estimates
of change are less certain and may result from
decadal variability rather than indicating a longterm trend. However, water mass properties indicate that these basin-scale changes are not solely a
result of vertical thermocline heave but are consistent with surface warming and the resultant subduction of this water into the main thermocline
(Wong et al., 2001). The warming in the Pacific
and Indian Oceans (Bindoff and McDougall,
2000) is confined to the main thermocline (mostly
the upper 1 km) of the subtropical gyres. This contrasts with the North Atlantic, where the warming
is seen at depths of 1000–2000 m. This is consistent with our understanding of the different basins
described earlier – deep convection and larger
meridional overturning in the North Atlantic contrasting with much shallower gyres in the Pacific.
Most recently Levitus et al. (2000) and Antonov
et al. (2001) have estimated global ocean heat
uptake of 10
22 J over the period 1955 to 1995 and
a consequent global average rate of thermal
expansion of 0.55 mm yr
91
.
Realistic simulation of ocean thermal expansion
requires coupled Atmosphere–Ocean General Circulation Models (AOGCMs) of the climate system
(Wood and Bryan, Chapter 2.3). A number of
model simulations of the twentieth century have
recently been completed using realistic greenhouse
gas and aerosol forcings starting in 1900 or earlier.
The results indicate that the average rate of change
due to thermal expansion over the last hundred
years was of the order of 0.3–0.8 mm yr
91
, and
0.6–1.1 mm yr
91 in recent decades, similar to
the observational estimates (Church et al., 2001;
Gregory et al., 2001).
SECTION 1 THE OCEAN AND CLIMATE
28