GLOBAL SEA RISE: A REDETERMINATION
BRUCE C. DOUGLAS
Department of Geography, University of Maryland, College Park, MD 20742, USA
Abstract. It is well established that sea level trends obtained from tide gauge records shorter than
about 50-60 years are corrupted by interdecadal sea level variation. However, only a fraction ( 25%)
of even the long records exhibit globally consistent trends, because of vertical crustal movements.
The coherent trends are from tide gauges not at collisional plate boundaries, and not located in or
near areas deeply ice-covered during the last glaciation. Douglas (1991), using ICE-3G values for the
postglacial (PGR) rebound correction, found 21 usable records (minimum length 60 years, average
76) in 9 oceanographic groups that gave a mean trend for global sea level rise of 1.8 mm/yr 0.1
for the period 1880–1980. In that analysis, a significant inconsistency of PGR-corrected U.S. east
coast trends was noted, but not resolved. Now, even after eliminating those trends, more (24) long
records (minimum 60 years, average 83) are available, including series in the southern hemisphere
not previously used. The mean trend of 9 groups made up of the newly-selected records is also
1.8 mm/yr 0.1 for global sea level rise over the last 100+ years. A somewhat smaller set of longer
records in 8 groups (minimum 70 years, average 91) gives 1.9 mm/yr 0.1 for the mean trend. These
values are about an order of magnitude larger than the average over the last few millennia. The recent
(in historical terms) dramatic increase in the rate of global sea level rise has not been explained,
and no acceleration during the last century has been detected. This situation requires additional
investigation and confirmation. VLBI/GPS/absolute gravity measurements of crustal motions can be
employed to correct many long (60+ years) tide gauge records not now usable because of vertical
crustal movements, improving the geographic coverage of sea level trends. Direct altimetric satellite
determinations of global sea level rise from satellites such as TOPEX/POSEIDON and its successors
can provide an independent estimate in possibly a decade or so, and thereby ascertain whether or not
there has been any recent change in the rate of global sea level rise.
<
1. Introduction
The issue of global sea level rise (GSLR) has aroused much interest because it is of
both great practical and scientific importance. As a practical issue, GSLR has major
impacts on most coastal regions. For discussions of these impacts, see Bird (1993),
Warrick et al., (1993), and Nicholls and Leatherman (1994). They document the
serious consequences of even a few-mm/yr increase of sea level. As a scientific
issue, GSLR is a unique indicator of global climate change, potentially providing
a means for evaluating climate models via their hindcasts and forecasts.
Summaries and reviews of the issue of global sea level rise (GSLR) normally
state that the value over the last 100 years or so lies between 1 and 2 mm/yr (e.g.,
Warrick et al., 1995). Douglas (1995) reviewed the more than one dozen studies and
determinations of GSLR from tide gauge data made since 1980, and noted that all
but one of the most recent estimates (1989 and later) conclude that global (eustatic)
sea level has risen during this century at a rate much closer to 2 mm/yr than 1 mm/yr
(Peltier and Tushingham, 1989; Trupin and Wahr, 1990; Douglas, 1991). This is
Surveys in Geophysics 18: 279–292, 1997.
c 1997 Kluwer Academic Publishers. Printed in the Netherlands.
280
BRUCE C. DOUGLAS
not to say that there is a consensus concerning the rate of rise of eustatic sea level;
some authors do not agree that it can be measured at all. Barnett (1984), Pirazzoli
(1989), Emery and Aubrey (1991), and Groger and Plag (1993) all argue that the
existing tide gauge record is inadequate for the task of determining a global value
for sea level rise. Douglas (1995) considered their arguments, and concluded that
in each case the authors depended on sea level records of insufficient length and/or
from unsuitable sites to reach their conclusions. These issues are considered again
below.
It is interesting that an accurate estimate of GSLR may not in fact present an
accurate picture of the thermal expansion of the oceans and addition of melt water.
Chao (1991) and Gornitz et al. (1995) calculate that during the last 40–50 years an
additional amount of water equivalent to 0.7–0.9 mm/yr of GSLR has been stored
in large and small reservoirs and other sinks, so that a much higher contemporary
rate of GSLR than derived from tide gauge trends is being masked.
The contemporary value of GSLR stands in sharp contrast to the rate during
the previous several millennia. During that time, GSLR was apparently about
an order of magnitude less (Flemming, 1978, 1982; Flemming and Webb, 1986;
Kearney and Stevenson, 1991; Shennan and Woodworth, 1992; Varekamp et al.,
1992; Peltier (this issue); Gornitz, 1995b; Kearney, 1996). In contrast, for the next
century various authors contend that global sea level will rise at a faster rate than
at present because of global warming. The Intergovernmental Panel on Climate
Change (IPCC) report (Houghton et al., 1990) gives for the “business-as-usual”
scenario of global warming an additional sea level change of 18 cm by 2030 and
44 cm by 2070 (the latest IPCC Sea Level assessment (Warrick et al., 1995) gives
a somewhat smaller, but still significantly increased value.) Church et al. (1991),
calculate a rise of 35 cm by 2050. Woodworth (1990) and Douglas (1992) have
shown that such increases require an acceleration of sea level an order of magnitude
greater (about 0.1–0.2 mm year 2 ) than the negligible acceleration observed in the
global tide gauge record of the last century. It is essential to further evaluate
the GSLR during the late Holocene through more field investigations, and refine
and verify estimates for the period of the recording tide gauge. But tide gauge
results will always suffer from the fact of their poor geographic distribution, as
eloquently discussed by Groger and Plag (1993). Indeed, the very existence of
altimeter satellites such as TOPEX/POSEIDON is due to the inherent limitations
of any conceivable tide gauge network! Satellite altimetry of TOPEX/POSEIDON
quality has both the requisite precision and geographic coverage to determine the
current rate of GSLR in a reasonable time (perhaps one or two decades), and further
detect variations in global and regional SLR that might accompany global warming.
GLOBAL SEA RISE: A REDETERMINATION
281
2. Determining GSLR from Tide Gauge Data
The 13 values of global sea level rise determined from tide gauge data published
since 1980 range in magnitude from 1.15 to 3 mm/yr (Gornitz, 1994). At the
same time, the formal uncertainties associated with most of these results were
reported as only one or a few tenths of a mm/yr. The tide gauge data used for these
estimates were supplied to investigators by the Permanent Service for Mean Sea
Level (PSMSL) (Spencer and Woodworth, 1993), so that differences in estimates
and especially their uncertainties reflect the authors’ approach to the problem, and
not the data.
The difficulty in obtaining a consensus on the value of GSLR can be attributed
largely to the presence of two kinds of “geophysical noise” in the data, and how
they have been treated. Douglas (1991) (hereafter referred to as D91) showed that
the treatment of (1) vertical crustal movements and (2) low frequency variations
of sea level has a profound impact on the results obtained for GSLR. Obviously,
vertical crustal movements directly affect the measured value of SLR at a site. The
effect of low-frequency (i.e., interdecadal) oceanographic sea level change is no
less important. The latter causes a systematic error in the estimate of SLR that
depends critically on record length. Douglas (1991, 1992) showed that records
shorter than about 60 years are not useable for determining GSLR because of this
effect. In this paper we shall see that even much longer records are desirable.
The use of tide gauge records of inadequate length by many authors arises from
the scarcity of long records. Peltier and Tushingham (1989) and Groger and Plag
(1993) have shown how the number of tide gauge records is “traded off” against
record length. A comprehensive review in D91 of the monthly mean sea level
records distributed by the PSMSL found only 21 records (through 1980) longer
than 60 years suitable for determining GSLR. These divided morphologically into
nine oceanographic regions, with no representation in the southern hemisphere
(D91 used data only through 1980 for consistency with earlier estimates). The
situation has improved since that time, with new long records now available from
Argentina and New Zealand, and more than a decade of additional data has been
used for most sites.
Concerning vertical crustal movements, the most ubiquitous source of regional
submergence or emergence at tide gauge sites is the Post Glacial Rebound (PGR)
that continues from the last deglaciation. It is manifest over the entire planet, not just
at locations ice-covered at the last glacial maximum. Vertical crustal movements due
to PGR at most sites in or near formerly ice-covered regions are of approximately
the same magnitude as the global (eustatic) rise, and in many cases greater. Even in
the “far field” of PGR, that is, areas well away from rebounding deglaciated areas
and the adjacent collapsing peripheral bulge, the PGR can be a significant, although
small, fraction of the GSLR. Peltier and Tushingham (1989) provide an excellent
short summary of their ICE-3G model for PGR, and a particularly clear treatment
of the relevance of PGR to the sea level problem. An exhaustive treatment of PGR
282
BRUCE C. DOUGLAS
and the basis of the ICE-3G model is found in Tushingham and Peltier (1991). It
is obvious that any attempt at estimating global sea level rise from tide gauge data
must take PGR into account, as well as other vertical crustal movements.
In addition to the need for a PGR correction and an adequate record length,
there are other desirable criteria for selecting sea level series taken by tide gauges.
D91 examined all long sea level records available from the PSMSL, and concluded
that to be the most useful they should meet the following requirements. These
are: (1) be at least 60 years in length, (2) not be from sites at collisional tectonic
plate boundaries, (3) 80% complete or better, (4) in reasonable agreement (at low
frequencies) with records from nearby gauges that sample the same water mass,
and (5) not from areas deeply covered by ice during the last glacial maximum. For
this paper the last of these has been strengthened to eliminate also records from
sites in the area of the peripheral bulge immediately adjacent to formerly deeply
ice covered regions.
The reason for the first of these criteria, the use of long records, has been
discussed above and in great detail in D91. The second criterion is important
because vertical crustal movements are the norm at colliding plate boundaries and
are impossible to model at present. The third requirement, for a high level of
completeness, exists because sea level records are not stationary time series; large,
unpredictable fluctuations lasting a decade or more can occur (Douglas, 1992),
so that data gaps have a significant influence on derived trends if they are large.
In addition, data gaps not only modify the trend in a statistical sense, they are a
“caution flag” that there has been an instrumental problem with the possibility of a
datum shift (P.L. Woodworth, private communication, 1996). The fourth condition,
for agreement of nearby gauges, is desirable because there is really no other way
to evaluate the quality of individual tide gauge records other than by comparing
them with records made by their neighbors. This comparison is useful because
at low frequencies oceanographic phenomena tend to be large in spatial scale, so
that gauges even several 100 km from each other will usually record similar low
frequency sea level variations and trends. Finally, gauge records were excluded in
D91 from deeply ice covered areas because the ICE-3G PGR correction for these
sites is very large, and even small errors in the correction are significant compared
to the actual GSLR.
This last point deserves further comment. PGR values are largest at the rebounding, formerly deeply ice-covered regions, reaching in some cases a cm/year or more.
In the adjacent collapsing peripheral bulge areas the rate is less, but still comparable
to the GSLR, so that any errors in the PGR corrections will have a major impact
on estimates of corrected SLR. A major disparity in long ICE-3G PGR-corrected
sea level trends along the U.S. east coast was noted in D91, but causation was a
matter of speculation at that time. In retrospect the problem seems certain to have
arisen from inadequacies of the ICE-3G model of Peltier and Tushingham (1989).
Mitrovica and Davis (1995) and Davis and Mitrovica (1996) attack the problem of
the U.S. east coast by making adjustments to the mantle viscosity used in the PGR
GLOBAL SEA RISE: A REDETERMINATION
283
model, and obtain an estimate of about 1.5 0.3 mm/yr for SLR there. However,
they also compute an estimate of SLR using their revised PGR corrections for
the non-east coast sites in D91, and obtain an average value of SLR of 1.4 mm
per year for them. It is interesting that this value is just what D91 yields for the
average SLR for those same sites without any PGR correction. The implication
is that their revision to the PGR correction, while producing greater coherence of
PGR-corrected trends on the U.S. east coast, gives an essentially null correction
for the far field of PGR. Clearly, this matter is not settled.
Peltier (this issue) takes a different approach to the problem of SLR on the U.S.
east coast. He argues persuasively from a comparison of geological estimates of
east coast SLR during the last millennium to contemporary sea level trends that
SLR over the twentieth century is very consistent on the entire east coast at about
1.9 mm/yr 0.2 (s.e.) greater than during the last 1000 years.
In contrast to the situation for the north American east coast and northern
Europe/Fennoscandia, ICE-3G PGR-corrected sea level trends are highly consistent
from suitable gauge sites well away from rebound and collapsing peripheral areas.
Especially it is unlikely that the signs of the ICE-3G corrections are wrong in the
far field of PGR. The correction for the effect of PGR as computed by Peltier
and Tushingham (1989) for the far field is to require an increase by a small but
significant amount (up to 0.5 mm/yr) to obtain the true value of SLR for stations
there.
3. A New Determination of GSLR
For this paper, PSMSL monthly mean sea level records were processed as in D91,
but subject to the revised selection criteria discussed above. In D91, records were
considered only to 1980 for the sake of consistency with most other published
reports. For this new analysis, the longest data records available have been used
(extending in many cases to 1991–93), and new records have been added including
southern hemisphere results from Argentina and New Zealand.
Table I presents the groups of sea level trends that survive the revised criteria
of this paper. There are several differences from the list of 60+ year records used
in D91. Most prominent is the elimination of any consideration of sea level trends
along the U.S. east coast north of Florida. It was pointed out in D91 that ICE-3G
PGR-corrected trends along the east coast of N. America fell into two groups,
divided at a latitude of about 38–40 N. An oceanographic explanation associated
with the N. Atlantic subtropical gyre had a certain appeal because of the nearness
to the latitude where the Gulf Stream departs from the coast and meanders toward
Europe. However, maintaining a trend difference requires an acceleration of the
gyre, an improbable occurrence over a time span approaching 100 years. The
more plausible explanation lies in the ICE-3G PGR correction used in D91. The
peripheral bulge collapse in this area north of approximately South Carolina is very
284
BRUCE C. DOUGLAS
large, ranging from 1–2 mm/yr. Any errors in the estimate of PGR from the ICE
3G model will have a relatively large effect on the PGR-corrected sea level trends.
As discussed above, that such errors exist is confirmed by Peltier (this issue). His
examination of the differential sea level trends between those derived from geologic
data over the last 1000 years and tide gauge data from the 20th century shows no
separation into groups, and a stable value (1.9 mm/yr) of the excess modern SLR
over that of the last millennium along the entire coast. Note in Table I that the
average corrected trend for Florida, where the PGR corrections are small, is little
different from that determined by Peltier for the entire coast.
Another difference in Table I from that of D91 is the absence of records from
the northern U.K. These were also eliminated because of their proximity to the
peripheral bulge. Table I is notable also for now including trends from the southern
hemisphere, in contrast to D91. At the time the calculations were done for D91,
all southern hemisphere records failed the requirements of that investigation for
inclusion. The sea level trends in Table I for the four New Zealand sites are taken
from Hannah (1990). Both he and Gornitz (1995b) argue that there is little evidence
for significant tectonic effects or subsidence or uplift during the period of the tide
gauge records. In addition, Gornitz concludes that there is evidence for only a very
minor amount of isostatic adjustment (0.0 to 0.1 mm/yr) over the last several
millennia. Peltier and Tushingham (1989) did not compute explicit values for
these sites because of the tectonic complexity of the region (W.R. Peltier, private
communication). But ICE-3G PGR values in the region are negative, and would
have the effect of pushing the estimate for SLR at New Zealand in the direction
of 2 mm/yr, rather than 1 mm/yr. Note that the results for GSLR are not changed
by a significant amount by whether or not New Zealand trends are included in the
calculation. Their significance lies in the fact that they tend to confirm that GSLR
is closer to 2 mm/yr, rather than 1 mm/yr.
The sea level trends shown for the Argentine stations Quequen and Buenos
Aires are also an addition to the list of D91. These sites are about 400 km apart, do
not show even qualitative correlation at low frequencies, and give rather different
values for the trend of sea level. Part of this difference is due to the disparity of
record length. The Buenos Aires series shows a steeply lower sea level before the
Quequen series begins, and a rapidly rising trend after it ends. Once again the need
for the longest possible series is underscored.
A final comment concerns Australian sea level trends, missing in Table I. D91
noted a major inconsistency between Sydney and Newcastle III, and was unable
to determine which was correct. Table II presents the trends for the Sydney and
Newcastle III series over four intervals. These are (1) their entire lengths, (2)
the period 1950–1989 where they are highly correlated (the Newcastle III series
obtained from the PSMSL ends in 1989, the Sydney series in 1993), (3) the period
1925-50 where they do not show even qualitative agreement, and (4) their common
period 1925–89. The trends are in exact agreement from 1950 onward, but disagree
very substantially over any other interval. The difference between the Sydney and
285
GLOBAL SEA RISE: A REDETERMINATION
Table I
Sea level trends for selected records longer than about 60 years. The mean of the ICE-3G
PGR-corrected group trends is 1.8 0.1 mm/yr. The median of the group trends is 1.8 mm/yr.
The average record length is 83 years.
Groups
Trend
(mm/yr)
NEWLYN
BREST
1.7
1.4
CASCAIS
LAGOS
TENERIFE
PGR
T-PGR
Start
End
Span
(yrs)
0.1
0.1
1.8
1.5
1915
1880
1991
1991
76
111
1.6
1.3
1.5
1.5
0.5
0.4
0.0
1.8
1.9
1.5
1882
1909
1927
1988
1990
1991
106
81
64
1.7
MARSEILLE
GENOVA
TRIESTE
1.2
1.2
1.2
0.2
0.2
0.3
1.4
1.4
1.5
1885
1884
1905
1991
1989
1991
106
105
86
1.4
AUCKLAND
DUNEDIN
LYTTELTON
WELLINGTON
1.3
1.4
2.3
1.7
1.3–1.4
1.4–1.5
2.3–2.4
1.7–1.8
1904
1900
1904
1901
1989
1989
1989
1988
85
89
85
87
1.7–1.8
HONOLULU
1.5
0.4
1.9
1905
1991
86
1.9
SAN FRANCISCO
SANTA MONICA
LA JOLLA
SAN DIEGO
1.5
1.4
2.1
2.1
0.4
0.7
0.6
0.6
1.9
2.1
2.7
2.7
1880
1933
1925
1906
1991
1991
1991
1991
111
58
66
85
2.4
BALBOA
CRISTOBAL
1.6
1.0
0.3
0.3
1.9
1.3
1908
1909
1970
1970
62
61
1.6
QUEQUEN
BUENOS AIRES
0.8
1.5
0.4
0.7
1.2
2.3
1918
1905
1983
1988
65
83
1.8
PENSACOLA
KEY WEST
FERNANDINA
2.2
2.2
1.8
0.2
0.4
0.2
2.4
2.7
2.0
1923
1913
1897
1991
1991
1991
68
78
94
2.4
Estimated by Gornitz (1995b) to be in the range 0.0 to
Group
Trend
(T-PGR)
0.1 mm/yr (see text). New Zealand
trends from Hannah (1990).
Newcastle III trends over their entire length, but their consistency from 1950onward is especially interesting. This divergence of record prior to 1950 in these
two series was noted in D91, and remains unexplained. Averaging these and other
Australian trends without regard to their record length or other problems, as done
by Gornitz (1995b) to obtain a mean value for Australian sea level rise, does not
seem appropriate. Unfortunately, it is not possible to unravel this situation from
286
BRUCE C. DOUGLAS
Table II
Sea level trends in mm/yr for Sydney and Newcastle III tide gauge records. The
Newcastle III series available from the PSMSL begins in 1925 and ends in 1989.
Note the exact agreement of trends from 1950-forward, but poor agreement
over their common time period of 1925–89. The problem arises from their
discordant sea level differences between 1925 and 1950. The trends derived
from their complete records actually disagree by a factor of 3 (0.7 mm/yr vs
2.1 mm/yr).
SYDNEY
NEWCASTLE III
1897–1993
1925–89
1925–50
1950-forward
0.7
–
1.4
2.1
0.2
1.5
0.7
0.7
Rate for entire record.
the published data, so that Australian sea level trends are excluded from further
consideration in this paper.
Overall, the value of 1.8 mm/yr 0.1 obtained for GSLR shown in Table I is
not different from the value given in D91 from 60+ year records. The use of longer
(70+ year) records, discussed below, makes a significant difference from the D91
results for the longest records.
In D91, an estimate of GSLR of 1.7 mm/yr was obtained from twelve 70+ year
(average length 86 years) records in 7 groups. The additional data now available
yields 17 longer records (average 91 years) in 8 groups that include southern
hemisphere series. Table III presents the results for GSLR for these longest records.
The value shown of 1.9 0.1 mm/yr is consistent with the value obtained in this
paper from 60+ year records, and is 0.2 mm/yr higher than the value shown in D91.
This estimate is more robust than the similar estimate in D91 because it has more
geographic representation and longer records.
Table III shows that something else that remains a problem with results from
tide gauge analyses. Note that the trends in the Mediterranean, which are computed
from some very long records, are systematically lower than the trends elsewhere.
Table IV gives the reason for this. It shows that average sea level has been stable
or falling slightly on both sides of the Italian Peninsula for at least 40 years in
comparison to the average record of the twentieth century. Clearly, true global
coverage of SLR is needed, and only satellite altimetry can supply that need.
4. Late Holocene Determinations of GSLR
It has been noted above that Peltier (this issue) has found an excess modern SLR of
1.9 mm/yr for the US east coast compared to that of the last millennium. He compared modern tide gauge trends with those obtained from geologic (radiocarbondated) sea level data. Gornitz (1995b) has carried out a similar analysis, but using
sea level data over the last 6–8 millennia. She found an excess SLR of 1.5 mm/yr
287
GLOBAL SEA RISE: A REDETERMINATION
Table III
Sea level trends for selected records longer than about 70 years. The mean of the group
trends 1.9 0.1 mm/yr. Note that the Mediterranean group trend is significantly lower than
the others. Average record length is 91 years.
Groups
Trend
(mm/yr)
NEWLYN
BREST
1.7
1.4
CASCAIS
LAGOS
PGR
T-PGR
Start
End
Span
(yrs)
0.1
0.1
1.8
1.5
1915
1880
1991
1991
76
111
1.6
1.2
1.5
0.5
0.4
1.8
1.9
1882
1909
1987
1990
106
81
1.8
MARSEILLE
GENOVA
TRIESTE
1.2
1.2
1.2
0.2
0.2
0.3
1.4
1.4
1.5
1885
1884
1905
1991
1989
1991
106
105
86
1.4
HONOLULU
1.5
0.4
2.0
1905
1991
86
2.0
1.3–1.4
1.4–1.5
2.3–2.4
1.7–1.8
1904
1900
1904
1901
1989
1989
1989
1988
85
89
85
87
1.7–1.8
Group
Trend
(T-PGR)
AUCKLAND
DUNEDIN
LYTTELTON
WELLINGTON
1.3
1.4
2.3
1.7
BUENOS AIRES
1.6
0.7
2.2
1905
1988
83
2.2
SAN FRANCISCO
SAN DIEGO
1.5
2.1
0.4
0.6
1.9
2.7
1880
1906
1991
1991
111
85
2.3
KEY WEST
FERNANDINA
2.2
1.8
0.4
0.2
2.6
2.0
1913
1898
1991
1991
78
93
2.3
Estimated by Gornitz (1995b) to be in the range 0.0 to 0.1 mm/yr (see text). New Zealand
trends from Hannah (1990).
for the same region, a bit less than that determined by Peltier. This difference
probably results from her analysis method. Gornitz (1995b) states that her use of
a linear fit to geologic sea level records over 6000–8000 years will overestimate
the correction for PGR because the isostatic adjustment is not linear over this time,
but in fact decreases by a significant amount. The result of neglecting this effect
is to decrease the estimate of the excess contemporary east coast SLR. Including
the effect would tend to make her estimate of the excess SLR more nearly equal to
that of Peltier (this issue).
In the matter of Australian trends, it has been noted already that trend results
from Australian tide gauge records are unreliable. Gornitz (1995b) found an excess
contemporary SLR of about 1 mm/yr, but with an equal uncertainty derivable from
the inconsistency of the sea level trends alone. The situation is much more tractable
for New Zealand. The evidence gathered by Hannah (1990) and Gornitz (1995b)
288
BRUCE C. DOUGLAS
Table IV
Mediterannean sea level trends from 1950–1992. Note
the average slight fall of sea level over this extended time
period. The Marseille record begins later because of a
“spike” in the data during 1950–52.
Station
Trend
MARSEILLE (1953-92)
GENOVA
TRIESTE
ROVINJ
BAKAR
SPLIT RT MARJANA
SPLIT HARBOUR
DUBROVNIK
0.77
0.28
0.44
0.49
0.04
0.36
0.94
0.33
Average
0.25
is persuasive that there has been very little subsidence/emergence at the tide gauge
sites during the late Holocene. The average value for SLR in New Zealand of 1.7–
1.8 mm/yr lies in the higher part of the published range of contemporary excess of
SLR over that of the late Holocene.
A very thorough investigation of the excess contemporary SLR for the UK and
a nearby part of Europe over the late Holocene has been made by Shennan and
Woodworth (1992). They conclude that the excess SLR there is 1.0 mm/yr 0.15.
In their paper, they express concern over two major issues, viz., the short (a few
decades) span of some of the tide gauge records from which they derived trends,
and the nonlinearity of PGR over the 6000–8000 years of the Holocene sea level
record.
Concern about the former is well justified. Sea level trends derived from the
relatively short records are corrupted by interdecadal fluctuations of sea level.
However, these trends have low weight in their overall solution for the excess SLR,
and offer at least qualitative support for their estimate. In fact, eliminating all of the
trends from shorter records would not appreciably affect their result for the excess
SLR.
In the matter of the nonlinearity of PGR during the late Holocene, Shennan
and Woodworth (op. cit.) also note that this effect will cause their value for the
excess modern SLR to be underestimated, and point out the necessity of further
work to ascertain the amount. It is tempting to apply the difference between the
results of Peltier (this issue) and Gornitz (1995b) for the U.S. east coast. That
difference, presumably due to the nonlinearity of PGR, was 0.4 mm/yr. However,
nearly all of the geologic records used by Shennan and Woodworth are less than
6000 years in length, and the nonlinearity effect will be smaller (I. Shennan, private
communication, 1996) than in the case of the analysis by Gornitz (1995b).
GLOBAL SEA RISE: A REDETERMINATION
289
Table V
Determinations of excess of SLR in the 20th century compared to the late Holocene.
Long Australian sea level records are inconsistent, and unusable for this analysis.
U.S. East Coast
Australia
New Zealand
U.K.
1.5 mm/yr (Gornitz, 1995b)
1.9 mm/yr (Peltier, 1995)
not used: see text
1.7–1.8 m/yr (Gornitz, 1995b)
1.1 mm/yr (Shennan and Woodworth, 1992)
5. Discussion
Table V summarizes the geological results for the excess modern sea level rise
over that of the late Holocene. Ignoring the values for Australia for the reasons
previously given, the range of values is 1.0–1.9 mm/yr. Thus a true consensus
remains elusive, although values closer to 2 mm per year rather than 1 mm per
year are favored, in good agreement with the value obtained in this paper of 1.8–
1.9 mm/yr, and all but one of the recent estimates from tide gauge data (Douglas,
1995). Of course the scientifically intriguing question as to why GSLR is an order
of magnitude greater now than the late Holocene average remains unexplained.
In addition to the problem of explaining the reason for the high modern value
of GSLR compared to that of recent millennia, it is also necessary to refine the
estimate of the GSLR over the last century in spite of the satisfyingly small uncertainty reported in this paper. This is necessary because many investigators remain
unconvinced that a robust value can be found. What is required to obtain entirely
convincing results is to eliminate the issue of modeling PGR or determining vertical
crustal movement rates from geologic data. This can be done by simply measuring
vertical crustal movements at the sites of tide gauges whose vertical motions are
unknown or controversial. Some good examples (in addition to the U.S. east coast)
are Bombay, Seattle, Tonura, and many others. That the necessary technology to
do this exists is well documented (Carter et al., 1989; Eden, 1990; Klopping et al.,
1991; Baker, 1993), and significant measurement programs are underway (IGS,
1995; Van Dam et al., 1995) that hopefully will be expanded in the future. But
even if further studies improve the understanding of modern vs late Holocene sea
level rise and yield a better determination of GSLR during the era of the recording
tide gauge, very important matters indeed, a problem remains. We are left with the
likelihood (Douglas, 1992; 1995) that detecting future change of the rate of GSLR
using tide gauges will simply take too long to be of value as an indicator of global
change. In addition, regional SLR changes that might occur will be missed in a
very great number of places because of the lack of coverage in the broad ocean
basins. Satellite altimetry offers a solution to this problem.
At first consideration, determination of GSLR by a satellite-borne radar altimeter
seems improbable. Prior to the TOPEX/POSEIDON (T/P) satellite, the error in
290
BRUCE C. DOUGLAS
the altitude of altimetric satellites (which maps directly as an error of sea level)
was of the order of tens of centimeters to a meter or more. But the potential
for determining low frequency variations of sea level was always better than the
error of the altitude. This occurs because the systematic error in the altitude of a
satellite has two forms; it is either periodic with the period of the orbit (Douglas,
et al., 1984), or correlated with geographic position (random error is negligible).
The former error obviously averages quickly, and the latter does not contribute
to measurements of change of sea level. Thus Seasat data, which had an error of
more than 1 meter from error in the satellite orbit, were used by Born et al. (1986)
to demonstrate a 10 centimeter consistency of global sea level from consecutive
three day repeat cycles. Later, Wagner and Cheney (1992) were able to improve
this consistency to a few cm for 17-day repeat cycles of Geosat data. The T/P
satellite has improved on its predecessors in every way. It uses a dual frequency
altimeter to determine the ionospheric correction, a water vapor radiometer for
direct measurement of the precipitable water in the path of the altimeter beam,
and advanced Global Positioning System (GPS), laser, and DORIS dual frequency
doppler tracking systems for orbit determination (Fu et al., 1994). Cheney et al.
(1994) concluded that TOPEX/Poseidon data are accurate at the level of 2 cm
for monthly mean changes on scales of a few hundred km, and by inference far
more accurate on a global, annual scale. Tapley et al., (1995) claim that a precision
of 1 mm or better is obtainable for measuring annual global sea level change,
supporting the results for GSLR of 3–5 mm/yr obtained by Wagner and Cheney
(1994), and Nerem (1995a,b,c).
All of the results obtained for GSLR from T/P data are larger than the 1.8–
1.9 mm/yr given in this paper. The three years of data from T/P show that even global
sea level has interannual fluctuations; Nerem (1995c) speculates that about a decade
will be required to determine a meaningful trend of GSL that can be compared to the
historical value. That decade of data will be forthcoming from the T/P Extended
Mission and its successor satellites. In the meantime the regional seasonal-tointerannual variations of sea level revealed by T/P, reflecting thermal expansion
and contraction of the upper layer of the ocean, are giving an unprecedented and
surprising new picture of the variations of upper ocean heat content.
Acknowledgement
This investigation was supported by a grant from NASA.
References
Baker, T.F.: 1993, ‘Absolute sea level measurements, climate change and vertical crustal movements’,
Global and Planetary Change 8, 145–149.
Barnett, T.P.: 1984, ‘The estimation of “Global” sea level change: A problem of uniqueness’, J.
Geophys. Res. 89(c5), 7980–7988.
GLOBAL SEA RISE: A REDETERMINATION
291
Bird, E.C.F.: 1993, Submerging Coasts, John Wiley and Sons.
Born, G., Tapley, B., Ries, J., and Stewart, R.: 1986, ‘Accurate measurement of mean sea level changes
by altimetric satellites’, J. Geophys. Res. 91, 11775–11782.
Carter, W.E., Aubrey, D.G., Baker, T., Boucher, C., LeProvost, C., Pugh, D., Peltier, W., Zumberge,
M., Rapp, R., Schutz, B., Emery, K., and Enfield, D.: 1989, ‘Geodetic fixing of tide gauge bench
marks’, Woods Hole Oceanographic Institution Technical Report WHOI-89-31.
Chao, B.F.: 1991, ‘Man, water, and global sea level’, EOS, Trans. Am. Geophys. Union 72(45), 492.
Cheney, R., Miller, L., Agreen, R., Doyle, N., and Lillibridge, J.: 1994, ‘TOPEX/POSEIDON: The
2-cm solution’, J. Geophys. Res. 99(C12), 24555–24564.
Church, J.A., Godfrey, J.S., Jacket, D.R., and MacDougal, T.J.: 1991, ‘A model of sea level rise
caused by ocean thermal expansion’, J. Climate 4(4), 438–456.
Davis, J.L. and Mitrovica, J.X.: 1996, Nature 379, 331–333.
Douglas, B.C., Agreen, R.W., and Sandwell, D.T.: 1984, ‘Observing global ocean circulation with
SEASAT altimeter data’, Marine Geodesy 8(1–4), 67–83.
Douglas, B.C.: 1991, ‘Global sea level rise’, J. Geophys. Res. 96(C4), 6981–6992.
Douglas, B.C.: 1992, ‘Global sea level acceleration’, J. Geophys. Res. 97(C8), 12,699–12,706.
Douglas, B.C.: 1995, ‘Global sea level change: Determination and interpretation’, Revs. Geophys.
(supp.), pp. 1425–1432.
Eden, H.F. (ed.): 1990, Towards an Integrated system for Measuring Long Term Changes in Global
Sea Level, Joint Oceanographic Institutions, Inc., Washington, D.C., May, 1990.
Emery, K.O., and Aubrey, D.G.: 1991, Sea Levels, Land Levels, and Tide Gauges, Springer-Verlag.
Flemming, N.C.: 1978, ‘Holocene eustatic changes and coastal tectonics in the northeast Mediterranean; implications for models of crustal consumption’, Phil. Trans. Royal Soc. London
289(1362), 405–458.
Flemming, N.C.: 1982, ‘Multiple regression analysis of earth movements and eustatic sea-level
change in the United Kingdom in the last 9000 years’, Proc. Geol. Assn. 93(1), 113–125.
Flemming, N.C. and Webb, C.O.: 1986, ‘Tectonic and eustatic coastal changes during the last 10,000
years derived from archeological data’, Z. Geomorph. N.F. Suppl.-Bd. 62, 1–29, Berlin.
Fu, L.-L., Christensen, E., Yamarone, C.A., Lefebvre, M., Menard, Y., Dorrer, M., and Escudier, P.:
1994, ‘TOPEX/POSEIDON mission overview’, J. Geophys. Res. 99(24), 24,369–24,382.
Gornitz, V.: 1994, ‘Sea level rise: A review of recent past and near-future trends’, Earth Surface
Processes and Trends, in press.
Gornitz, V., Rozenzweig, C., and Hillel, D.: 1995, ‘Effects of anthropogenic intervention in the land
hydrologic cycle on global sea level rise’, private communication.
Gornitz, V.: 1995b, ‘A comparison of differences between recent and late holocene sea-level trends
from eastern North America and other selected regions’, J. Coastal Res. Special Issue 17, 287–
297.
Groger, M. and Plag, H.-P.: 1993, ‘Estimations of a global sea level trend: Limitations from the
structure of the PSMSL global sea level data set’, Global and Planetary Change 8, 161–179.
Hannah, J.: 1990, ‘Analysis of mean sea level data from New Zealand for the period 1899–1988’, J.
Geophys. Res. 95(B8).
Houghton, J.T., Jenkins, G.J., and Ephraums, J.J. (eds.): 1990, Climatic Change; the IPCC Scientific
Assessment, Cambridge University Press, Cambridge, England.
Kearney, M.S. and Stevenson, J.C.: 1991, ‘Island land loss and marsh vertical accretion rate evidence
for historical sea-level changes in Chesapeake Bay’, J. Coastal Res. 7(2), 403–415.
Kearney, M.S.: 1996, ‘Sea level change during the last thousand years in Chesapeake Bay’, J. Coastal
Res. in press.
Klopping, F., Peter, G., Robertson, D.S., Berstis, K.A., Moose, R.E., and Carter, W.E.: 1991, ‘Improvements in absolute gravity observations’, J. Geophys. Res. 96(b5), 8295–8303.
Leatherman, S.P.: 1989, ‘Impact of accelerated sea level rise on beaches and coastal wetlands’, in
Global Climate Change Linkages, Elsevier, pp. 43–56.
Leatherman, S.P.: 1991, ‘Modeling shore response to sea level rise on sedimentary coasts’, Progress
in Physical Geography 14, 447–464.
292
BRUCE C. DOUGLAS
Leatherman, S. P.: 1992, ‘Coastal land loss in the Chesapeake Bay region: An historical analogy
approach to global climate analysis and response’, in J. Schmandt (ed.), The Regions and Global
Warming: Impacts and Response Strategies, Oxford University Press, pp. 17–27.
Mitrovica, J.X. and Davis, J.L.: 1995, ‘Present-day post-glacial sea level change far from the Late
Pleistocene ice sheets: Implications for recent analyses of tide gauge records’, Geophys. Res.
Letters 22(18), 2529–2532.
National Research Council (NRC): 1990, Report on Sea Level Change, National Academy Press,
Washington, DC.
Nerem, R.S.: 1995a, ‘Global mean sea level variations from TOPEX/POSEIDON altimeter data’,
Science 268, 708–710.
Nerem, R.S., Rachlin, K.E., and Beckley, B.D.: 1995b, ‘Long term sea level change observed by
TOPEX/POSEIDON’, EOS Trans. AGU (supp.).
Nerem, R.S.: 1995c, ‘Measuring global mean sea level variations using TOPEX/POSEIDON altimeter
data’, J. Geophys. Res. 100, 25,135–25,152.
Nicholls, R.J. and Leatherman, S.P.: 1994, ‘Global sea-level rise’, in K. Strzepek and J.B. Smith
(eds.), As Climate Changes: Potential Impacts and Implications, Cambridge University Press,
Cambridge, in press.
Peltier, W.R., and Tushingham, A.M.: 1989, ‘Global sea level rise and the greenhouse effect: Might
they be connected?’, Science 244(4906), 806–810.
Pirazzoli, P.A.: 1993, ‘Global sea level changes and their measurement’, Global Planet. Change 8,
135–148.
Shennan, I., and Woodworth, P.L.: 1992, ‘A comparison of late Holocene and twentieth-century
sea-level trends from the UK and North Sea region’, Geophys. J. Int. 109(1), 96–105.
Spencer, N.E. and Woodworth, P.L.: 1993, ‘Data holdings of the permanent service for mean sea
level, Bidston, Birkenhead: Permanent service for mean sea level’.
Tapley, B.D., Shum, C.K., Guman, M., Chen, J., Ries, J.C., and Chambers, D.P.: 1995, ‘Monitoring
global sea level variations using satellite altimetry’, EOS, Trans. Am. Geophys. Union 76(46),
F159.
Trupin, A. and Wahr, J.: 1990, ‘Spectroscopic analysis of global tide gauge sea level data’, Geophys.
J. Int. 100, 441–453.
Tushingham, A.M. and Peltier, W.R.: 1991, ‘ICE-3G: A new global model of late Pleistocene deglaciation based upon geophysical predictions of post glacial relative sea level change’, J. Geophys.
Res. 96, 4497–4523.
Van Dam, T., Nerem, R.S., and Schenewerk, M.: 1995, ‘A geodetic network for monitoring sea level
in the Chesapeake Bay (BAYONET)’, EOS Trans. AGU (supp.).
Varekamp, J.C., Thomas, E., and Van de Plassche, O.: 1992, ‘Relative sea level rise and climate
change over the last 1500 years’, Terra Nova 4, 293–304.
Wagner, C.A. and Cheney, R.E.: 1992, ‘Global sea level change from satellite altimetry’, J. Geophys.
Res. 97(C10), 15607–15615.
Warrick, R.A., Le Provost, C., Meier, M.F., Oerlemans, J., and Woodworth, P.L. (eds.): 1995, Changes
in Sea Level, Chapt. 7 of the Second Intergovernmental Panel on Climate Change Scientific
Assessment.
Warrick, R.A., Barrow, E.M., and Wigley, T.M.L. (eds): 1993, Climate and Sea Level Change:
Observations, Projections, and Implications, Cambridge University Press, Cambridge.
Woodworth, P.L.: 1990, ‘A search for accelerations in records of European mean sea level’, Int. J.
Climat. 10, 129–143.