- Email: [email protected]
Absolute sea level measurements, climate change and vertical crustal movements
Global and Planetary Change, 8 (1993) 149-159 Elsevier Science Publishers B.V., Amsterdam
149
Absolute sea level measurements, climate change and vertical crustal movements T.F. Baker
Proudman Oceanographic Laboratory, Bidston Observatory, Birkenhead, Merseyside L43 7RA, UK (Received May 18, 1992; revised and accepted September 25, 1992)
ABSTRACT Baker, T.F., 1993. Absolute sea level measurements, climate change and vertical crustal movements. Global Planet. Change, 8:149-159 Annual mean sea level observations from tide gauges around the world, usually show interannual and decadal variations of order 5-10 cm. Because of these variations, several decades of data are normally required for a reliable determination of the secular trend in mean sea levels. Tide gauges only give relative sea level trends, since a rise in sea level cannot be distinguished from a subsidence of the crust at the tide gauge and thus estimates of the "global" rise of mean sea levels must be corrected for these vertical crustal movements. A brief review is given of recent work on relative mean sea levels, which shows the importance of these land movements. Modern space geodetic techniques and absolute gravity have now achieved the equivalent accuracy of a few centimetres that is compatible with the above variability in annual mean sea levels. Measurements of vertical crustal movements at tide gauges using these techniques are now being carried out in various parts of the world. A summary is given of the recommendations of an international working group on the geodetic fixing of tide gauge bench marks and some of the measurement errors that are now being investigated are discussed. These measurements are of interest to oceanographers working on climate change and to geophysicists working on vertical crustal movements.
Introduction Sea level is m e a s u r e d at over 1000 tide gauges a r o u n d the world. T h e most c o m m o n l y u s e d syst e m is a float o p e r a t i n g in a stilling well (shown schematically in Fig. 1), a l t h o u g h i n c r e a s i n g use is n o w b e i n g m a d e of p r e s s u r e a n d acoustic transducers (Pugh, 1987). A s c a n be s e e n in Fig. 1, the sea level is m e a s u r e d with respect to a local tide gauge b e n c h m a r k ( T G B M ) . This is a n e a r b y b e n c h m a r k e i t h e r in b e d r o c k or in a solid structure such as a h a r b o u r wall. T h e tide gauge d a t u m relative to the T G B M is c h e c k e d regularly a n d this should b e d o n e to m i l l i m e t r i c accuracy, b u t with m a n y tide gauges this is difficult to achieve. M e a n sea level is the average sea level over a suitable interval of time which is c h o s e n to s m o o t h o u t m a n y of the h i g h e r f r e q u e n c y v a r i a t i o n s due, for example, to tides a n d surges. M o n t h l y a n d
a n n u a l m e a n sea levels are n o r m a l l y u s e d in most work. T h e P e r m a n e n t Service for M e a n Sea Level (PSMSL), located at the P r o u d m a n O c e a n o graphic Laboratory, U K , is the i n t e r n a t i o n a l d a t a b a n k which collects a n d distributes worldwide m o n t h l y a n d a n n u a l m e a n sea levels. It is clear from Fig. 1 that m e a n sea level (MSL) d e t e r m i n e d from a tide gauge is a relative m e a n sea level, since a rise of sea level c a n n o t b e d i s t i n g u i s h e d from a s u b s i d e n c e of the crust at the T G B M . I n this paper, after s u m m a r i s i n g some of the r e c e n t work o n relative m e a n sea levels, a review is given of the m e t h o d s that are n o w b e i n g u s e d to s e p a r a t e vertical crustal m o v e m e n t s from sea level variations, using m o d e r n geodetic m e a s u r e m e n t s n e a r tide gauges. I n this review a summary will b e given of the m a i n c o n c l u s i o n s from a r e c e n t i n t e r n a t i o n a l w o r k i n g g r o u p o n geodetic fixing of tide gauge b e n c h marks ( C a r t e r et al., 1989).
0921-8181/93/$06.00 © 1993 - Elsevier Science Publishers B.V. All rights reserved
150
r.F. BAKER ROD
GPS RECEIVER
HT ABSOLUTE GRAVIMETER MSL SCHEMATIC OF TIDE GAU___GETO MEASURE ABSOLUTE SEA LEVEL
Fig. 1. A schematic diagram of a tide gauge to measure absolute sea level. The tide gauge bench mark (TGBM) is connected by spirit levelling to a network of bench marks at distances of a few h u n d r e d metres. At one of these bench marks a receiver is used to provide a connection via the Global Positioning System (GPS) satellites to a similar receiver at a Satellite Laser Ranging (SLR) station or Very Long Baseline Interferometry (VLBI) radio telescope. Absolute gravity m e a s u r e m e n t s both near the tide gauge and at the S L R / V L B I station give a completely independent determination of the vertical crustal movements,
The separation of sea level variations and vertical crustal movements is of interest to geophysicists working on crustal movements and to oceanographers working on climate change. Table 1 lists the main reasons for the interest in fixing tide gauge bench marks at the centimetric level. In some parts of the world, m e a n sea level observations from tide gauges give valuable time series from 1900, or even earlier, to the present day. They therefore contain important information on vertical crustal movements which can be used for testing models of post glacial rebound. In tectonic areas they contain information on crustal movements associated with earthquakes (Fig. 2). The mean sea level records may also contain information on subsidence due to sediment loading and compaction, subsidence due to, for example, extraction of ground water or oil, or local stability problems associated with the pier or harbour. As will be seen in the next section, the vertical crustal movements at T G B M s are of a similar order of magnitude to the global rise in m e a n sea levels and it is therefore necessary to correct the estimates of the global rise of mean sea levels for the crustal movements at the tide gauges. Global models of climate change should be able to reproduce the change in mean sea levels over the past century, in order to give increased confidence in their predictions of future sea level rise.
Changes in ocean circulation can be found by determining the changes in geostrophic flow between islands or across straits (e.g. Spencer et al., 1993). The changes in the sea levels recorded at pairs of tide gauges are used to monitor long term variations in flow between the tide gauges. However, it is important to monitor the vertical
NEZUGASEKI ANNUAL MEAN SEA LEVELS
i
50rnm
•
0 •
o0o
•
•
•
•
0000000
1960 1970 1980 Fig. 2. Annum mean sea levels from the tide gauge at Nezugaseki in Japan. The 20 cm jump in mean sea level is due to crustal subsidence during the Niigata earthquake in June 1964 (magnitude 7.5, distance 40 kin). Note that before the earthquake, the annual mean sea levels show evidence of crustal uplift.
ABSOLUTE SEA LEVEL MEASUREMENTS, CLIMATE CHANGE AND VERTICAL CRUST MOVEMENTS
151
TABLE 1 Geophysical and oceanographic requirements for the geodetic fLxing of tide gauge bench marks Vertical crustal movements
Climate change
(1) Local subsidence
(1) Global rise of mean sea levels (corrected for crustal movements) (2) Changes in ocean circulation (sea levels across straits or between islands) (calibration, validation, sea surface topography) (3) Satellite altimetry (calibration, validation, sea surface topography)
(2) Tectonics
(3) Postgalical rebound
crustal movements at the tide gauges to ensure that any changes in the sea levels can definitely be interpreted as long term changes in geostrophic flow. The last item in Table 1 concerns satellite altimetry. The radar altimeters on Geosat, ERS-1 (and, in the near future, TOPEX-Poseidon) are proving to be valuable techniques for measuring changes in sea surface topography on a global scale. This is particularly important for monitor-
ing changes in ocean circulation for the World Ocean Circulation Experiment (WOCE). Improved measurements of ocean circulation are needed for testing the future models of ocean circulation that are required for climate prediction. It should be noted that determining the mean ocean circulation from sea level observations is considerably more difficult than measuring the variations about the mean. This is because the
¢ OCEAN SURFACE- ~
LASER SITE
ISUAND TIOE GAUGE
Fig. 3. A schematic diagram of the measurement of sea level by satellite radar altimetry. If the tide gauge is connected to the satellite laser ranging station by GPS measurements, then the sea level can be determined in the same reference frame that is used for the altimetric sea levels.
152
geoid is still relatively poorly known and will only be improved to the accuracy of a few centimetres that is required for ocean circulation work, when future satellites are launched containing gravity gradiometers. Figure 3 gives a schematic diagram of an altimeter satellite. The satellite orbit is determined from tracking by the global network of satellite laser ranging (SLR) stations, together with a global network of radio tracking devices. Therefore the sea surface as determined from the altimeter is measured with respect to the geocentre or reference ellipsoid. The geocentre is determined to within a few centimetres by the global network of SLR measurements to the L A G E O S satellite, which was launched in 1976. If the tide gauge in Fig. 3 is geodetically connected to the SLR station, the sea level as determined from the tide gauge will be in the same reference frame as that determined from the altimeter. This can be used to calibrate the altimeter, reduce the altimeter orbit errors and to validate the altimetric sea level measurements in different areas. Orbital errors for G E O S A T and ERS-1 are typically 3 0 40 cm and for T O P E X - P o s e i d o n an orbital accuracy of order 10 cm is planned. Even geocentrically fixing tide gauges at the 10 cm level will provide important information for satellite altimetry. Relative mean sea levels
Annual mean sea levels from all parts of the world contain interannual and decadal fluctuations of order 5 - 1 0 cm. These fluctuations are usually very coherent between nearby tide gauges and therefore they are real sea level variations rather than measurement "noise". Understanding these interannual and decadal variations is important for work on ocean circulation and climate change. Changes in wind stress, salinity and ocean circulation are considered to be important factors (Thompson, 1986). In the tropical Pacific, sea level variations associated with the E1 Nifio Southern Oscillation are important in climate research (Wyrtki, 1975). The interannual and decadal variations cause significant problems for determining a reliable
LF. BAKER
estimate of the secular trend in mean sea level (Pugh, 1987). Ideally, over 60 years of mean sea level data are needed to obtain an estimate of the secular trend that is not severely affected by the decadal variations. Although the PSMSL contains data from over 1400 tide gauges worldwide, there are fewer than 100 stations with reliable mean sea level data for 60 or more years. Very few of these longer data sets are from the southern hemisphere and the largest number of long records are from Europe and North America (Woodworth, 1991). In order to overcome this geographical bias in future work. the Intergovernmental Oceanographic Commission (IOC 1989) have initiated the Global Sea Level Observing System (GLOSS). This consists of about 300 primary tide gauges with approximately 1000-km spacing along the continental coastlines and a gauge in every island group (Fig. 4). It should be noted that at present only about 2 / 3 of these gauges are operational (Woodworth, 1991). Figure 5 shows plots of annual mean sea levels from some of the longest tide gauge records available in Europe (Woodworth et al.. 1990). It can be seen that the mean sea levels at Cascais (Portugal), Brest (France), Newlyn (UK) and Hoek Van Holland (Netherlands) are rising by 15-20 cm per century (i.e. 1.5-2.0 m m / y r ) . This rate of rise is typical of that found in other parts of the world. However, it is also clear from Fig. 5 that in Northern Europe the mean sea level trend is markedly different. At Stockholm (Sweden) the mean sea level is falling by about 40 c m / c e n t u r y . Scandinavia is still rising due to post glacial rebound following the last ice age, which ended about 7000 years ago. At Aberdeen (Scotland) the mean sea level trend is very small and it appears that the post glacial rebound at Aberdeen is of a similar magnitude to the sea level rise. Models of post glacial rebound have been developed (Tushingham and Peltier, 1991; Lambeck, 1990) which use viscoelastic models of the Earth's interior and s p a c e - t i m e histories of the ice sheets that loaded the Earth. By fitting the model results to relative sea levels over the past 18,000 years, as determined from radiocarbon dating of previous shorelines, both the Earth
20"[
?0
60"£
i
40"[
~0'
,3+ ae|3_l~ V
IO0"E
0
~11
~
60"[
80"E
i lO0"E
120'E
,proposed elations i l i i
,1~0
160*E
ill
~ 441
i
,o1.
,I.II4
180"
160'W
130
• 143
• I.
160".
~ ILl ".114,
11 .III
• 111~
b
...
180"
11..,,I.%,
~61
140'lr
i
':
n f
160"E
:11411P
140'(
A'.~I";
s
II:
n
I20"E
Global Sea-Level OblerTln I System S t a t u s D e c e m b e r 1968
~,Z3
21.
;' 1, ' t ~ , . ,
41"-
80'£
• existing stations i i i i
.21
.24 1,8
I~ !P., °;+
[3 ]
t
40"(
i 140*W
i
.138
,14,2
140'~
120*W
,°,
120'.
i lO0*Id
'137
• 1115
I00"~
175
80"W
20~
~17
40%/
~O."
Lll3 300
/-
,180
.103
53
20"I.1
+
,,
20"M
.,. •go~;8' Pillo
40"~
1~..
.305
i~
60*W
17ll~k
1 7 ~
#174
• 130 ~ _ ~
171
177
~173
/~0
103
60".
~ | 0 4
+.
80"U
O*
O'
'log
20*E
01+01 lO
b
20"E
Fig. 4. The Global Sea Level Observing System (GLOSS) network of primary tide gauges (Intergovernmental Oceanographic Commission). Regional networks are also being developed in areas of particular interest for sea levels or crustal movements (e.g. the North Sea, Mediterranean and Caribbean).
O"
20"N
40"N
20"(
Z
< r~
@
,.q
tc~
t-n
t~ <
P~
c'3
t-n
t'n
t-n >
~r
t" t~ < t"
-] t'n
o
154
r.F. BAKER
the ice sheets, give important spatial variations in present day sea level trends. Even in equatorial areas, post glacial rebound models predict a present day sea level fall of about 0.4 m m / y r (Peltier and Tushingham, 1989, 1991). Many papers have been published which use the PSMSL data set to estimate the secular trends in global mean sea levels. The most recent of these use the output from the global models of post glacial rebound to correct the tide gauge
models and the knowledge of the ice distribution can be improved. The models include the loading deformation of the Earth due to the ice sheets, the loading of the ocean floor by the meltwater and also allow for the gravitational effects due to the redistribution of mass in the Earth and oceans. In Europe and North America, crustal uplift in the areas which were covered by ice during the last deglaciation (18,000-7,000 yr B.P.) and crustal submergence in the peripheral region adjacent to 9000--
8800
--
8600
--
8400
--
STOCKHOLM
ESBJERG ~ 8200
t
--
ABERDEEN
E
.~• ~
8000--
7800
--
7600
--
7400
--
7200
--
7000
--
6800
..... 1800
i ..... 1830
i ..... 1860
i ..... I 890
i ..... 1920
i ..... 1950
i' 1980
Fig. 5. Some examples of long time series of European annual mean sea levels (each record has been given an arbitrary vertical offset for presentation purposes). Note the interannual and decadal fluctuations of sea level of the order of _+5 to _-t-10 cm. Because of these variations, several decades of data are normally required for a reliable determination of the secular trend in mean sea levels.
ABSOLUTE SEA LEVEL MEASUREMENTS, CLIMATE C H A N G E AND VERTICAL C R U S T MOVEMENTS
records. Trupin and Wahr (1990) used the results of an earlier rebound model of Peltier (1986). Using data from 84 tide gauges with a minimum record length of 37 years, they obtained a global sea level rise of 1.75 + 0.13 ram/year. Douglas (1991) used the rebound model results of Tushingham and Peltier (1991) to correct the trends from 21 tide gauges situated on non-convergent plate boundaries and with a minimum record length of 60 years and obtained a sea level rise of 1.8 + 0.1 ram/yr. These results agree within the quoted errors, which are the standard errors of the mean trend. Peltier and Tushingham (1989, 1991) used their model to correct the trends from 40 tide gauges with a minimum length of 50 years data and obtained a global trend of 2.4 + 0.9 m m / y r (standard deviation). However, they also show that the estimated trend is sensitive to the choice of record length and the time period used for the analysis. The sea level trend determined from the above tide gauge data is consistent with estimates of the various contributions to sea level rise due to climate change. The Intergovernmental Panel on Climate Change (IPCC, 1990) reviewed the published evidence and estimated a sea level rise of 10 cm over the last century but with an uncertainty ranging from - 0 . 5 cm to + 22 cm. The estimated contributions from thermal expansion of the oceans and melting of low latitude glaciers each amount to about 40% of the total. The contributions of the Greenland and Antarctic ice sheets are particularly uncertain due to the lack of observational data. The IPCC also reviewed the published papers on the predictions of future sea level rise due to greenhouse induced climate change. Their best estimate, assuming a business-as-usual scenario, is a sea level rise of 66 cm by the end of the next century. Although this is less than some previous estimates, it is still 3-6 times faster than the rate of rise over the last century. A b s o l u t e sea level m e a s u r e m e n t s
Although the above estimates of sea level rise using tide gauge data have been corrected for post glacial rebound using global models, it is
155
dearly desirable to measure the vertical crustal movement at a tide gauge. This would enable the true sea level variation to be decoupled from the crustal movement and thus provide valuable data for testing both models of glacial rebound and models of sea level variations due to climate change. Although post glacial rebound models are now improving rapidly, suitable models are not available for tectonic movements or subsidence at tide gauges. Some authors have pointed out that estimates of global sea level rise may be biased since subsidence is more common than uplift in oceanic and coastal areas (Pirazzoli, 1986). Emery and Aubrey (1991) conclude that most of the trends in relative sea levels are due to land movements. Advances in modern geodetic techniques now offer the possibility of directly measuring the crustal movement at tide gauges. These measurements will determine mean sea level at a tide gauge in a global geocentric reference frame and hence they will give absolute mean sea levels, rather than mean sea level relative to each local bench mark. The U.S. Department of Defense, over the last few years, has been launching satellites as part of a satellite based global navigation system called the Global Positioning System (GPS). When the constellation of satellites is complete in the early 1990s, it will consist of 21 satellites (and 3 spares) at an altitude of 20,000 km (12 hour orbital period) arranged so that at any one time at least 4 satellites will be visible from any point on the Earth's surface. The satellites transmit coded modulations on two carrier frequencies (carrier wavelengths of 19 and 24 cm). With access to the codes, a user with a GPS satellite receiver can determine his real time position to an accuracy of the order of 10 m. The key development that is now giving the accuracies required for crustal deformation work is to use the phases of the two carrier waves rather than the codes. By using pairs of dual frequency GPS receivers, relative vector positioning has been achieved at the centimetric level for baselines of up to 1000 km in length. The reader is referred to the articles by Dixon (1991), Hager et al. (1991) and Bilham (1991) for a review of the advances in differential
156
GPS measurements and the application to the measurement of crustal deformations. An international working group was set up by the International Association for the Physical Sciences of the Ocean (IAPSO) under its Commission on Mean Sea Level and Tides to recommend a strategy for global fixing of tide gauge bench marks using the latest geodetic techniques. The report of the working group (Carter et al., 1989) recommends that the global absolute sea level monitoring system should be based upon the primary satellite laser ranging (SLR) stations and Very Long Baseline Interferometry (VLBI) radio telescopes of the International Earth Rotation Service (IERS) Terrestrial Reference Frame. Many of the 30-40 station positions in this network are now known to within 2 cm (Ray et al., 1991; Carter and Robertson, 1990). The addition of more stations and further improvements in accuracy are expected in the next few years. SLR observations have already been used to determine the vertical motion of stations to within 1 m m / y r (Kolenciewicz et al., 1992). The recommended procedure is to connect the T G B M to the nearest primary SLR or VLBI site using differential dual frequency GPS. If satellite visibility is restricted at the TGBM, then a new bench mark may have to be installed nearby for the GPS measurements and connected to the T G B M by primary spirit levelling (see Fig. 1). The report also recommends that absolute gravity measurements should be made at the S L R / V L B I stations and in the vicinity of the tide gauge, This will give an important, completely independent, cheek upon the vertical crustal movements at both the tide gauge and the IERS sites. At remote sites, such as on oceanic islands that are far removed from the V L B I / S L R stations, absolute gravity may be the only feasible method of determining the vertical crustal movement at the tide gauge. For a review of the recent advances in absolute gravimetry, the reader is referred to Marson and Faller (1986) and Torge (1989). The principle of the absolute gravimeter is the measurement of the acceleration of a mass in free fall (or rise and fall) in a vacuum using a laser length standard and a rubidium frequency time standard. The mass is a retro-reflector which forms
r v . BAKER
one arm of a Michelson laser interferometer. A lot of effort has been put into reducing or eliminating various sources of systematic error. A great deal of experience has been gained during the past few years using portable absolute gravimeters built by the Istituto di Metrologia Colonnetti, CNR, Torino (Marson, 1989) and the Joint Institute of Laboratory Astrophysics (JILA), Boulder, Colorado (Torge et al., 1987; Peter et al., 1989; Lambert et al., 1989). The gravity value is obtained by making repeat drops over one or two days at each site and corrections are made for tides and atmospheric pressure variations. At good sites repeat visits show that a precision of about 2 p, gals can be achieved. The absolute accuracy is harder to estimate but is believed to be about 6 ;zgals. After more developments to reduce the errors still further, a new portable absolute gravimeter is being built by the University of Trieste (Cerutti et al., 1992) and a new absolute gravimeter is being produced commercially in 1.992 by the AXIS Instruments Company, Boulder, Colorado. The specifications for the AXIS instrument are a precision of +_1/zgal and an accuracy of + 2 ~gals. The gravity gradient in free air, at the Earth's surface, is 3 /xgal/cm. In practice, for crustal deformation work, since a large area of the Earth's surface is usually displaced simultaneously, the measured gravity change is of the order of 2 / z g a l / c m . Thus, it can be seen that absolute gravity and space geodetic techniques are both approaching the equivalent accuracy of 1 cm that is required for measuring vertical crustal movements.
Conclusions and future developments Table 2 lists the main techniques that are required for geodetic fixing of TGBMs (Carter et al., 1989). The first requirement, should already be undertaken by each National Authority as part of their programme for the reliable determination of relative mean sea levels, This is a network of 6-10 local bench marks which serve as both witness marks for the T G B M but will also show whether any movement of the T G B M is either extremely local or represents a more regional
ABSOLUTE SEA LEVEL MEASUREMENTS, CLIMATE CHANGE AND VERTICAL CRUST MOVEMENTS TABLE 2 Techniques required for geodetic fixing of tide gauge bench marks (TGBMs) Technique
Required accuracy
(1) l o c a l network of bench marks for relative sea levels. (Primary levelling or GPS)
0 - 1 km: < l m m 1-10 kin: < 1 cm
(2) GPS from T G B M to S L R / V L B I reference frame
(3) Absolute gravity at S L R / V L B I sites and near tide gauges
< 2 p, gal
crustal movement. The local network of bench marks should be resurveyed by spirit levelling or GPS at least once per year. The working group report does not discuss the dimensions of the local network. However, many countries use a network of very local bench marks extending over a few hundred metres, which show any instabilities in the construction of the tide gauge or harbour area. Other bench marks up to 10 km away are used to check that the vertical movements at the tide gauge are regionally representative. For the bench marks that are within a few hundred metres, primary spirit levelling should achieve an accuracy of better than 1 mm. For distances of up to 10 km, spirit levelling or GPS should aim for sub-centimetre accuracy. It is important to note that all these measurements are made with respect to the TGBM. The earlier practice of relating the TGBM, and hence mean sea level, to a National datum point by spirit levelling is not recommended for this work. Conventional spirit levelling has significant systematic errors over distances of greater than a hundred kilometres and relevellings, or adjustments, of the national levelling network often lead to spurious apparent jumps in mean sea level. In order to avoid the higher microseismic noise for gravity measurements immediately adjacent to the coastline, the report recommends that the absolute gravity measurements should be made at sites 1-10 km inland. The gravity site (which is normally in a building with reasonable tempera-
157
ture control) has also then to be connected to the T G B M by spirit levelling or GPS. Inland sites also enable a higher accuracy to be achieved for the calculation of ocean tide loading and attraction corrections. However, measurements for a few months with a well calibrated continuously recording relative gravimeter should enable corrections to be made to a few tenths of a microgal at any distance from the coastline (Baker et al., 1991). Although most of the measurement errors in the space geodesy and gravity measurement techniques have now been reduced to the level required to begin the measurements of vertical crustal movements outlined in this review, there are still some significant problems that remain to be addressed. The future programmes of measurements will identify the various approaches to reducing these problems. There have been great improvements in GPS receivers and analytical techniques during the last few years. Receiver noise and set up errors are typically a few millimetres. Improved GPS satellite tracking from fiducial (VLBI or SLR) stations will ensure that errors due to uncertainties in the orbits are reduced to the order of 1 part in 108 of the baseline length. Dixon (1991) reviews the precision and accuracy currently available with GPS as a function of baseline length. The vertical component is the most difficult to determine with GPS, since there is a geometric limitation due to the satellites only being observable in the upper hemisphere. It is also approximately a factor of 3 more sensitive to errors in the tropospheric calibration. Tropospheric errors due to the wet component depend upon baselength for the first 20 km or so (depending upon the tropospheric correlation length) and are then independent of baselength. Dixon (1991) presents measurements from California which, for the vertical component, have a repeatibility of the order of 2 cm for baselines from 20 to 500 km. He also shows that the repeatability is significantly worse in tropical areas. Future improvements in GPS measurements in the vertical will depend upon improved troposphere calibration and estimation techniques and may require the use of line of sight water vapour radiometers.
158
For absolute gravity measurements, variations in ground water can produce gravity variations at some sites that are in excess of 10/zgals. Corrections for these variations will require monitoring of ground water using boreholes, but in some areas modelling of the complex hydrology will be difficult. Superconducting gravimeters now have a very low drift rate and can be used to provide continuous measurements of gravity variations between absolute gravity measurements at some primary sites (Goodkind, 1986). It may be necessary to avoid sites that are likely to have significant ground water variations by choosing sites on crystalline bedrock, but this option is not available in many areas. The repeatability of order 2 /zgals achieved by the present generation of absolute gravimeters at some sites shows that, with care, suitable sites can be chosen. Absolute sea level measurement programmes have recently commenced in many countries. Absolute gravity measurements near tide gauges are now being made in North America, Japan and Europe. GPS measurement programmes at tide gauges have commenced in North America, Hawaii, South Africa, Japan, Australia and Europe. Gravity and space geodesy measurements in areas of post glacial rebound (e.g. North America, Fennoscandia, Scotland) will also test the models that are currently being used for correcting tide gauge records and such measurement programmes are also now underway. Many years of careful observations will be required for a reliable determination of the vertical crustal movements. During the next 2-3 years the results from the first repeat campaigns will be available and it will be interesting to see the progress that has been achieved.
Acknowledgements I would like to thank Phil Woodworth and Bob Edge for useful comments on the manuscript.
References Baker, T.F., Edge, R.J. and Jeffries, G., 1991. Tidal gravity and ocean tide loading in Europe. Geophys. J. Int., 107: 1-11.
w.v. BAKER Bilham, R., 1991. Earthquakes and sea level: space and terrestrial metrology on a changing planet Rev. Geophys., 29: 1-29.
Carter, W.E. and Robertson, D.S., 1990. Definition of a terrestrial reference frame using IRIS VLBI observations: approaching millimetre accuracy. In: C. Boucher and G.A. Wilkins (Editors), Symposium, 105: Earth Rotation and Coordinate Reference Frames. Springer, New York, pp. 115-122. Carter, W.E., Aubrey, D.G., Baker, T.F., Boucher, C., LeProvost, C., Pugh, D.T., Peltier, W.R., Zumberge, M., Rapp, R.H., Schutz, R.E., Emery, K.O. and Enfield, D.B., 1989. Geodetic fixing of tide gauge bench marks. Woods Hole Oceanogr. Inst. Tech. Rep. WHOI-89-31/CRC-89-5. Cerutti, G., Alosia, F., Germak, A., Bozzo, E., Caneva, G.. Lanza, R. and Marson, 1., 1992. The absolute gravity station and the Mt. Melbourne network in Terra Nova Bay North Victoria Land, East Antarctica. In: Proc. Sixth Int. Symp. Antarctic Earth Sciences, in press. Dixon, T.H., 1991. An introduction to the Global Positioning System and some geological applications. Rev. Geophys., 29: 249-276. Douglas, B.C., 1991. Global sea level rise. J. Geophys. Res., 96(C4): 6981-6992. Emery, K.O. and Aubrey, D.G., 1991. Sea Levels, Land Levels and Tide Gauges. Springer, New York, 237 pp. Goodkind, J.M., 1986. Continuous measurement of non-tidal variations of gravity. J, Geophys. Res., 91: 9125-9134. Hager, B.H., King, R.W. and Murray, M.H., 1991. Measurement of crustal deformation using the Global Positioning System. Annu. Rev. Earth Planet. Sci., 19: 351-382. IOC, 1989. Global Sea Level Observing System (GLOSS): Proposed Implementation Plan. Intergov. Oceanogr. Comm. Rep,, IOC-XV/8 Annex 4, 67 pp. IPCC, (J.T. Houghton, G.J. Jenkins and J.J. Ephraums, Editors) 1990. The Intergovernmental Panel on Climate Change Scientific Assessment. Cambridge Univ. Press, 365 pp. Kolenkiewicz, R., Smith, D.E., Pavlis, E.C. and Torrence, M.H., 1992. Direct estimation of vertical variations of satellite laser tracking sites from SL8 LAGEOS analysis. Ann. Geophys., 10(1), p. 111. Lambeck, K., 1990. Glacial rebound, sea level change and mantle viscosity. Q. J. R. Astron. Soc., 21: 1-30. Lambert, A., Liard, J.O., Courtier, N., Goodacre, A.K., McConnell, R.K. and Failer, J.E., 1989. Canadian absolute gravity program: Applications in geodesy and geodynamics. EOS, 70: 1447-1460. Marson, I., 1989. Esperimenti ed idee nella realizzazione di un gravimetro assoluto trasportabile. In: Atti dell'Istituto Veneto di Scienze Lettere ed Arti, 146: 41-58. Marson, I. and Failer, J.E, 1986. g - - T h e acceleration of gravity: its measurement and its importance. J. Phys. E: Sci. Instr., 19: 22-32. Peltier, W.R., 1986. Deglaciation-induced vertical motion of the North American Continent and transient lower mantle rheology. J. Geophys. Res., 91(B9): 9099-9123. Peltier, W.R. and Tushingham, A.M., 1989. Global sea level rise and the greenhouse effect: might they be related.'? Science, 244: 806-810.
ABSOLUTE SEA LEVEL MEASUREMENTS, CLIMATE CHANGE AND VERTICAL CRUST MOVEMENTS Peltier, W.R. and Tushingham, A.M., 1991. Influence of glacial isostatic adjustment on tide gauge measurements of secular sea level change. J. Geophys. Res., 96(134): 6779-6796. Peter, G., Moose, R.E., Wessells, C.W., Failer, J.E. and Niebauer, T.M., 1989. High precision absolute gravity observations in the United States. J. Geophys. Res., 94: 5659-5674. Pirazzoli, P.A., 1986. Secular trends of relative sea level changes indicated by tide gauge records. J. Coastal Res., 1: 1-26. Pugh, D.T., 1987. Tides, Surges and Mean Sea Level: a Handbook for Engineers and Scientists. Wiley, Chichester, 472 pp. Ray, J.R., Ma, C., Ryan, J.W., Clark, T.A., Eanes, R.J., Watkins, M.M., Schutz, B.E. and Tapley, B.D., 1991. Comparison of VLBI and SLR geocentric site coordinates. Geophys. Res. Lett., 18: 231-234. Spencer, R., Foden, P.R., McGarry, C., Harrison, A.J., Vassie, J.M., Baker, T.F., Smithson, M.J., Harangozo, S.A. and Woodworth, P.L., 1993. The ACCLAIM programme in the South Atlantic and Southern Oceans. Int. Hydrogr. Rev., 71(1): 7-21.
159
Thompson, K.R., 1986. North Atlantic sea level and circulation. Geophys. J. R. Astron. Soc., 87: 15-32. Torge, W., 1989. Gravimetry. De Gruyter, Berlin, 465 pp. Torge, W., R6der, R.H., Schniill, M., Wenzel, H.-G. and Failer, J.E., 1987. First results with the transportable absolute gravity meter JILAG-3. Bull. G6ocies., 61: 161-176. Trupin, A. and Wahr, J., 1990. Spectroscopic analysis of global tide gauge sea level data. Geophys. J. Int., 100: 441-454. 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(B3): 4497-4523. Woodworth, P.L., 1991. The permanent service for mean sea level and the global sea level observing system. J. Coastal Res., 7: 699-710. Woodworth, P.L., Spencer, W.E. and Alcock, G., 1990. On the availability of European mean sea level data. Int. Hydrogr. Rev., 67(1): 131-146. Wyrtki, K., 1975. El Nino--the dynamic response of the Equatorial Pacific Ocean to atmospheric forcing. J. Phys. Oceanogr., 5: 572-584.
149
Absolute sea level measurements, climate change and vertical crustal movements T.F. Baker
Proudman Oceanographic Laboratory, Bidston Observatory, Birkenhead, Merseyside L43 7RA, UK (Received May 18, 1992; revised and accepted September 25, 1992)
ABSTRACT Baker, T.F., 1993. Absolute sea level measurements, climate change and vertical crustal movements. Global Planet. Change, 8:149-159 Annual mean sea level observations from tide gauges around the world, usually show interannual and decadal variations of order 5-10 cm. Because of these variations, several decades of data are normally required for a reliable determination of the secular trend in mean sea levels. Tide gauges only give relative sea level trends, since a rise in sea level cannot be distinguished from a subsidence of the crust at the tide gauge and thus estimates of the "global" rise of mean sea levels must be corrected for these vertical crustal movements. A brief review is given of recent work on relative mean sea levels, which shows the importance of these land movements. Modern space geodetic techniques and absolute gravity have now achieved the equivalent accuracy of a few centimetres that is compatible with the above variability in annual mean sea levels. Measurements of vertical crustal movements at tide gauges using these techniques are now being carried out in various parts of the world. A summary is given of the recommendations of an international working group on the geodetic fixing of tide gauge bench marks and some of the measurement errors that are now being investigated are discussed. These measurements are of interest to oceanographers working on climate change and to geophysicists working on vertical crustal movements.
Introduction Sea level is m e a s u r e d at over 1000 tide gauges a r o u n d the world. T h e most c o m m o n l y u s e d syst e m is a float o p e r a t i n g in a stilling well (shown schematically in Fig. 1), a l t h o u g h i n c r e a s i n g use is n o w b e i n g m a d e of p r e s s u r e a n d acoustic transducers (Pugh, 1987). A s c a n be s e e n in Fig. 1, the sea level is m e a s u r e d with respect to a local tide gauge b e n c h m a r k ( T G B M ) . This is a n e a r b y b e n c h m a r k e i t h e r in b e d r o c k or in a solid structure such as a h a r b o u r wall. T h e tide gauge d a t u m relative to the T G B M is c h e c k e d regularly a n d this should b e d o n e to m i l l i m e t r i c accuracy, b u t with m a n y tide gauges this is difficult to achieve. M e a n sea level is the average sea level over a suitable interval of time which is c h o s e n to s m o o t h o u t m a n y of the h i g h e r f r e q u e n c y v a r i a t i o n s due, for example, to tides a n d surges. M o n t h l y a n d
a n n u a l m e a n sea levels are n o r m a l l y u s e d in most work. T h e P e r m a n e n t Service for M e a n Sea Level (PSMSL), located at the P r o u d m a n O c e a n o graphic Laboratory, U K , is the i n t e r n a t i o n a l d a t a b a n k which collects a n d distributes worldwide m o n t h l y a n d a n n u a l m e a n sea levels. It is clear from Fig. 1 that m e a n sea level (MSL) d e t e r m i n e d from a tide gauge is a relative m e a n sea level, since a rise of sea level c a n n o t b e d i s t i n g u i s h e d from a s u b s i d e n c e of the crust at the T G B M . I n this paper, after s u m m a r i s i n g some of the r e c e n t work o n relative m e a n sea levels, a review is given of the m e t h o d s that are n o w b e i n g u s e d to s e p a r a t e vertical crustal m o v e m e n t s from sea level variations, using m o d e r n geodetic m e a s u r e m e n t s n e a r tide gauges. I n this review a summary will b e given of the m a i n c o n c l u s i o n s from a r e c e n t i n t e r n a t i o n a l w o r k i n g g r o u p o n geodetic fixing of tide gauge b e n c h marks ( C a r t e r et al., 1989).
0921-8181/93/$06.00 © 1993 - Elsevier Science Publishers B.V. All rights reserved
150
r.F. BAKER ROD
GPS RECEIVER
HT ABSOLUTE GRAVIMETER MSL SCHEMATIC OF TIDE GAU___GETO MEASURE ABSOLUTE SEA LEVEL
Fig. 1. A schematic diagram of a tide gauge to measure absolute sea level. The tide gauge bench mark (TGBM) is connected by spirit levelling to a network of bench marks at distances of a few h u n d r e d metres. At one of these bench marks a receiver is used to provide a connection via the Global Positioning System (GPS) satellites to a similar receiver at a Satellite Laser Ranging (SLR) station or Very Long Baseline Interferometry (VLBI) radio telescope. Absolute gravity m e a s u r e m e n t s both near the tide gauge and at the S L R / V L B I station give a completely independent determination of the vertical crustal movements,
The separation of sea level variations and vertical crustal movements is of interest to geophysicists working on crustal movements and to oceanographers working on climate change. Table 1 lists the main reasons for the interest in fixing tide gauge bench marks at the centimetric level. In some parts of the world, m e a n sea level observations from tide gauges give valuable time series from 1900, or even earlier, to the present day. They therefore contain important information on vertical crustal movements which can be used for testing models of post glacial rebound. In tectonic areas they contain information on crustal movements associated with earthquakes (Fig. 2). The mean sea level records may also contain information on subsidence due to sediment loading and compaction, subsidence due to, for example, extraction of ground water or oil, or local stability problems associated with the pier or harbour. As will be seen in the next section, the vertical crustal movements at T G B M s are of a similar order of magnitude to the global rise in m e a n sea levels and it is therefore necessary to correct the estimates of the global rise of mean sea levels for the crustal movements at the tide gauges. Global models of climate change should be able to reproduce the change in mean sea levels over the past century, in order to give increased confidence in their predictions of future sea level rise.
Changes in ocean circulation can be found by determining the changes in geostrophic flow between islands or across straits (e.g. Spencer et al., 1993). The changes in the sea levels recorded at pairs of tide gauges are used to monitor long term variations in flow between the tide gauges. However, it is important to monitor the vertical
NEZUGASEKI ANNUAL MEAN SEA LEVELS
i
50rnm
•
0 •
o0o
•
•
•
•
0000000
1960 1970 1980 Fig. 2. Annum mean sea levels from the tide gauge at Nezugaseki in Japan. The 20 cm jump in mean sea level is due to crustal subsidence during the Niigata earthquake in June 1964 (magnitude 7.5, distance 40 kin). Note that before the earthquake, the annual mean sea levels show evidence of crustal uplift.
ABSOLUTE SEA LEVEL MEASUREMENTS, CLIMATE CHANGE AND VERTICAL CRUST MOVEMENTS
151
TABLE 1 Geophysical and oceanographic requirements for the geodetic fLxing of tide gauge bench marks Vertical crustal movements
Climate change
(1) Local subsidence
(1) Global rise of mean sea levels (corrected for crustal movements) (2) Changes in ocean circulation (sea levels across straits or between islands) (calibration, validation, sea surface topography) (3) Satellite altimetry (calibration, validation, sea surface topography)
(2) Tectonics
(3) Postgalical rebound
crustal movements at the tide gauges to ensure that any changes in the sea levels can definitely be interpreted as long term changes in geostrophic flow. The last item in Table 1 concerns satellite altimetry. The radar altimeters on Geosat, ERS-1 (and, in the near future, TOPEX-Poseidon) are proving to be valuable techniques for measuring changes in sea surface topography on a global scale. This is particularly important for monitor-
ing changes in ocean circulation for the World Ocean Circulation Experiment (WOCE). Improved measurements of ocean circulation are needed for testing the future models of ocean circulation that are required for climate prediction. It should be noted that determining the mean ocean circulation from sea level observations is considerably more difficult than measuring the variations about the mean. This is because the
¢ OCEAN SURFACE- ~
LASER SITE
ISUAND TIOE GAUGE
Fig. 3. A schematic diagram of the measurement of sea level by satellite radar altimetry. If the tide gauge is connected to the satellite laser ranging station by GPS measurements, then the sea level can be determined in the same reference frame that is used for the altimetric sea levels.
152
geoid is still relatively poorly known and will only be improved to the accuracy of a few centimetres that is required for ocean circulation work, when future satellites are launched containing gravity gradiometers. Figure 3 gives a schematic diagram of an altimeter satellite. The satellite orbit is determined from tracking by the global network of satellite laser ranging (SLR) stations, together with a global network of radio tracking devices. Therefore the sea surface as determined from the altimeter is measured with respect to the geocentre or reference ellipsoid. The geocentre is determined to within a few centimetres by the global network of SLR measurements to the L A G E O S satellite, which was launched in 1976. If the tide gauge in Fig. 3 is geodetically connected to the SLR station, the sea level as determined from the tide gauge will be in the same reference frame as that determined from the altimeter. This can be used to calibrate the altimeter, reduce the altimeter orbit errors and to validate the altimetric sea level measurements in different areas. Orbital errors for G E O S A T and ERS-1 are typically 3 0 40 cm and for T O P E X - P o s e i d o n an orbital accuracy of order 10 cm is planned. Even geocentrically fixing tide gauges at the 10 cm level will provide important information for satellite altimetry. Relative mean sea levels
Annual mean sea levels from all parts of the world contain interannual and decadal fluctuations of order 5 - 1 0 cm. These fluctuations are usually very coherent between nearby tide gauges and therefore they are real sea level variations rather than measurement "noise". Understanding these interannual and decadal variations is important for work on ocean circulation and climate change. Changes in wind stress, salinity and ocean circulation are considered to be important factors (Thompson, 1986). In the tropical Pacific, sea level variations associated with the E1 Nifio Southern Oscillation are important in climate research (Wyrtki, 1975). The interannual and decadal variations cause significant problems for determining a reliable
LF. BAKER
estimate of the secular trend in mean sea level (Pugh, 1987). Ideally, over 60 years of mean sea level data are needed to obtain an estimate of the secular trend that is not severely affected by the decadal variations. Although the PSMSL contains data from over 1400 tide gauges worldwide, there are fewer than 100 stations with reliable mean sea level data for 60 or more years. Very few of these longer data sets are from the southern hemisphere and the largest number of long records are from Europe and North America (Woodworth, 1991). In order to overcome this geographical bias in future work. the Intergovernmental Oceanographic Commission (IOC 1989) have initiated the Global Sea Level Observing System (GLOSS). This consists of about 300 primary tide gauges with approximately 1000-km spacing along the continental coastlines and a gauge in every island group (Fig. 4). It should be noted that at present only about 2 / 3 of these gauges are operational (Woodworth, 1991). Figure 5 shows plots of annual mean sea levels from some of the longest tide gauge records available in Europe (Woodworth et al.. 1990). It can be seen that the mean sea levels at Cascais (Portugal), Brest (France), Newlyn (UK) and Hoek Van Holland (Netherlands) are rising by 15-20 cm per century (i.e. 1.5-2.0 m m / y r ) . This rate of rise is typical of that found in other parts of the world. However, it is also clear from Fig. 5 that in Northern Europe the mean sea level trend is markedly different. At Stockholm (Sweden) the mean sea level is falling by about 40 c m / c e n t u r y . Scandinavia is still rising due to post glacial rebound following the last ice age, which ended about 7000 years ago. At Aberdeen (Scotland) the mean sea level trend is very small and it appears that the post glacial rebound at Aberdeen is of a similar magnitude to the sea level rise. Models of post glacial rebound have been developed (Tushingham and Peltier, 1991; Lambeck, 1990) which use viscoelastic models of the Earth's interior and s p a c e - t i m e histories of the ice sheets that loaded the Earth. By fitting the model results to relative sea levels over the past 18,000 years, as determined from radiocarbon dating of previous shorelines, both the Earth
20"[
?0
60"£
i
40"[
~0'
,3+ ae|3_l~ V
IO0"E
0
~11
~
60"[
80"E
i lO0"E
120'E
,proposed elations i l i i
,1~0
160*E
ill
~ 441
i
,o1.
,I.II4
180"
160'W
130
• 143
• I.
160".
~ ILl ".114,
11 .III
• 111~
b
...
180"
11..,,I.%,
~61
140'lr
i
':
n f
160"E
:11411P
140'(
A'.~I";
s
II:
n
I20"E
Global Sea-Level OblerTln I System S t a t u s D e c e m b e r 1968
~,Z3
21.
;' 1, ' t ~ , . ,
41"-
80'£
• existing stations i i i i
.21
.24 1,8
I~ !P., °;+
[3 ]
t
40"(
i 140*W
i
.138
,14,2
140'~
120*W
,°,
120'.
i lO0*Id
'137
• 1115
I00"~
175
80"W
20~
~17
40%/
~O."
Lll3 300
/-
,180
.103
53
20"I.1
+
,,
20"M
.,. •go~;8' Pillo
40"~
1~..
.305
i~
60*W
17ll~k
1 7 ~
#174
• 130 ~ _ ~
171
177
~173
/~0
103
60".
~ | 0 4
+.
80"U
O*
O'
'log
20*E
01+01 lO
b
20"E
Fig. 4. The Global Sea Level Observing System (GLOSS) network of primary tide gauges (Intergovernmental Oceanographic Commission). Regional networks are also being developed in areas of particular interest for sea levels or crustal movements (e.g. the North Sea, Mediterranean and Caribbean).
O"
20"N
40"N
20"(
Z
< r~
@
,.q
tc~
t-n
t~ <
P~
c'3
t-n
t'n
t-n >
~r
t" t~ < t"
-] t'n
o
154
r.F. BAKER
the ice sheets, give important spatial variations in present day sea level trends. Even in equatorial areas, post glacial rebound models predict a present day sea level fall of about 0.4 m m / y r (Peltier and Tushingham, 1989, 1991). Many papers have been published which use the PSMSL data set to estimate the secular trends in global mean sea levels. The most recent of these use the output from the global models of post glacial rebound to correct the tide gauge
models and the knowledge of the ice distribution can be improved. The models include the loading deformation of the Earth due to the ice sheets, the loading of the ocean floor by the meltwater and also allow for the gravitational effects due to the redistribution of mass in the Earth and oceans. In Europe and North America, crustal uplift in the areas which were covered by ice during the last deglaciation (18,000-7,000 yr B.P.) and crustal submergence in the peripheral region adjacent to 9000--
8800
--
8600
--
8400
--
STOCKHOLM
ESBJERG ~ 8200
t
--
ABERDEEN
E
.~• ~
8000--
7800
--
7600
--
7400
--
7200
--
7000
--
6800
..... 1800
i ..... 1830
i ..... 1860
i ..... I 890
i ..... 1920
i ..... 1950
i' 1980
Fig. 5. Some examples of long time series of European annual mean sea levels (each record has been given an arbitrary vertical offset for presentation purposes). Note the interannual and decadal fluctuations of sea level of the order of _+5 to _-t-10 cm. Because of these variations, several decades of data are normally required for a reliable determination of the secular trend in mean sea levels.
ABSOLUTE SEA LEVEL MEASUREMENTS, CLIMATE C H A N G E AND VERTICAL C R U S T MOVEMENTS
records. Trupin and Wahr (1990) used the results of an earlier rebound model of Peltier (1986). Using data from 84 tide gauges with a minimum record length of 37 years, they obtained a global sea level rise of 1.75 + 0.13 ram/year. Douglas (1991) used the rebound model results of Tushingham and Peltier (1991) to correct the trends from 21 tide gauges situated on non-convergent plate boundaries and with a minimum record length of 60 years and obtained a sea level rise of 1.8 + 0.1 ram/yr. These results agree within the quoted errors, which are the standard errors of the mean trend. Peltier and Tushingham (1989, 1991) used their model to correct the trends from 40 tide gauges with a minimum length of 50 years data and obtained a global trend of 2.4 + 0.9 m m / y r (standard deviation). However, they also show that the estimated trend is sensitive to the choice of record length and the time period used for the analysis. The sea level trend determined from the above tide gauge data is consistent with estimates of the various contributions to sea level rise due to climate change. The Intergovernmental Panel on Climate Change (IPCC, 1990) reviewed the published evidence and estimated a sea level rise of 10 cm over the last century but with an uncertainty ranging from - 0 . 5 cm to + 22 cm. The estimated contributions from thermal expansion of the oceans and melting of low latitude glaciers each amount to about 40% of the total. The contributions of the Greenland and Antarctic ice sheets are particularly uncertain due to the lack of observational data. The IPCC also reviewed the published papers on the predictions of future sea level rise due to greenhouse induced climate change. Their best estimate, assuming a business-as-usual scenario, is a sea level rise of 66 cm by the end of the next century. Although this is less than some previous estimates, it is still 3-6 times faster than the rate of rise over the last century. A b s o l u t e sea level m e a s u r e m e n t s
Although the above estimates of sea level rise using tide gauge data have been corrected for post glacial rebound using global models, it is
155
dearly desirable to measure the vertical crustal movement at a tide gauge. This would enable the true sea level variation to be decoupled from the crustal movement and thus provide valuable data for testing both models of glacial rebound and models of sea level variations due to climate change. Although post glacial rebound models are now improving rapidly, suitable models are not available for tectonic movements or subsidence at tide gauges. Some authors have pointed out that estimates of global sea level rise may be biased since subsidence is more common than uplift in oceanic and coastal areas (Pirazzoli, 1986). Emery and Aubrey (1991) conclude that most of the trends in relative sea levels are due to land movements. Advances in modern geodetic techniques now offer the possibility of directly measuring the crustal movement at tide gauges. These measurements will determine mean sea level at a tide gauge in a global geocentric reference frame and hence they will give absolute mean sea levels, rather than mean sea level relative to each local bench mark. The U.S. Department of Defense, over the last few years, has been launching satellites as part of a satellite based global navigation system called the Global Positioning System (GPS). When the constellation of satellites is complete in the early 1990s, it will consist of 21 satellites (and 3 spares) at an altitude of 20,000 km (12 hour orbital period) arranged so that at any one time at least 4 satellites will be visible from any point on the Earth's surface. The satellites transmit coded modulations on two carrier frequencies (carrier wavelengths of 19 and 24 cm). With access to the codes, a user with a GPS satellite receiver can determine his real time position to an accuracy of the order of 10 m. The key development that is now giving the accuracies required for crustal deformation work is to use the phases of the two carrier waves rather than the codes. By using pairs of dual frequency GPS receivers, relative vector positioning has been achieved at the centimetric level for baselines of up to 1000 km in length. The reader is referred to the articles by Dixon (1991), Hager et al. (1991) and Bilham (1991) for a review of the advances in differential
156
GPS measurements and the application to the measurement of crustal deformations. An international working group was set up by the International Association for the Physical Sciences of the Ocean (IAPSO) under its Commission on Mean Sea Level and Tides to recommend a strategy for global fixing of tide gauge bench marks using the latest geodetic techniques. The report of the working group (Carter et al., 1989) recommends that the global absolute sea level monitoring system should be based upon the primary satellite laser ranging (SLR) stations and Very Long Baseline Interferometry (VLBI) radio telescopes of the International Earth Rotation Service (IERS) Terrestrial Reference Frame. Many of the 30-40 station positions in this network are now known to within 2 cm (Ray et al., 1991; Carter and Robertson, 1990). The addition of more stations and further improvements in accuracy are expected in the next few years. SLR observations have already been used to determine the vertical motion of stations to within 1 m m / y r (Kolenciewicz et al., 1992). The recommended procedure is to connect the T G B M to the nearest primary SLR or VLBI site using differential dual frequency GPS. If satellite visibility is restricted at the TGBM, then a new bench mark may have to be installed nearby for the GPS measurements and connected to the T G B M by primary spirit levelling (see Fig. 1). The report also recommends that absolute gravity measurements should be made at the S L R / V L B I stations and in the vicinity of the tide gauge, This will give an important, completely independent, cheek upon the vertical crustal movements at both the tide gauge and the IERS sites. At remote sites, such as on oceanic islands that are far removed from the V L B I / S L R stations, absolute gravity may be the only feasible method of determining the vertical crustal movement at the tide gauge. For a review of the recent advances in absolute gravimetry, the reader is referred to Marson and Faller (1986) and Torge (1989). The principle of the absolute gravimeter is the measurement of the acceleration of a mass in free fall (or rise and fall) in a vacuum using a laser length standard and a rubidium frequency time standard. The mass is a retro-reflector which forms
r v . BAKER
one arm of a Michelson laser interferometer. A lot of effort has been put into reducing or eliminating various sources of systematic error. A great deal of experience has been gained during the past few years using portable absolute gravimeters built by the Istituto di Metrologia Colonnetti, CNR, Torino (Marson, 1989) and the Joint Institute of Laboratory Astrophysics (JILA), Boulder, Colorado (Torge et al., 1987; Peter et al., 1989; Lambert et al., 1989). The gravity value is obtained by making repeat drops over one or two days at each site and corrections are made for tides and atmospheric pressure variations. At good sites repeat visits show that a precision of about 2 p, gals can be achieved. The absolute accuracy is harder to estimate but is believed to be about 6 ;zgals. After more developments to reduce the errors still further, a new portable absolute gravimeter is being built by the University of Trieste (Cerutti et al., 1992) and a new absolute gravimeter is being produced commercially in 1.992 by the AXIS Instruments Company, Boulder, Colorado. The specifications for the AXIS instrument are a precision of +_1/zgal and an accuracy of + 2 ~gals. The gravity gradient in free air, at the Earth's surface, is 3 /xgal/cm. In practice, for crustal deformation work, since a large area of the Earth's surface is usually displaced simultaneously, the measured gravity change is of the order of 2 / z g a l / c m . Thus, it can be seen that absolute gravity and space geodetic techniques are both approaching the equivalent accuracy of 1 cm that is required for measuring vertical crustal movements.
Conclusions and future developments Table 2 lists the main techniques that are required for geodetic fixing of TGBMs (Carter et al., 1989). The first requirement, should already be undertaken by each National Authority as part of their programme for the reliable determination of relative mean sea levels, This is a network of 6-10 local bench marks which serve as both witness marks for the T G B M but will also show whether any movement of the T G B M is either extremely local or represents a more regional
ABSOLUTE SEA LEVEL MEASUREMENTS, CLIMATE CHANGE AND VERTICAL CRUST MOVEMENTS TABLE 2 Techniques required for geodetic fixing of tide gauge bench marks (TGBMs) Technique
Required accuracy
(1) l o c a l network of bench marks for relative sea levels. (Primary levelling or GPS)
0 - 1 km: < l m m 1-10 kin: < 1 cm
(2) GPS from T G B M to S L R / V L B I reference frame
(3) Absolute gravity at S L R / V L B I sites and near tide gauges
< 2 p, gal
crustal movement. The local network of bench marks should be resurveyed by spirit levelling or GPS at least once per year. The working group report does not discuss the dimensions of the local network. However, many countries use a network of very local bench marks extending over a few hundred metres, which show any instabilities in the construction of the tide gauge or harbour area. Other bench marks up to 10 km away are used to check that the vertical movements at the tide gauge are regionally representative. For the bench marks that are within a few hundred metres, primary spirit levelling should achieve an accuracy of better than 1 mm. For distances of up to 10 km, spirit levelling or GPS should aim for sub-centimetre accuracy. It is important to note that all these measurements are made with respect to the TGBM. The earlier practice of relating the TGBM, and hence mean sea level, to a National datum point by spirit levelling is not recommended for this work. Conventional spirit levelling has significant systematic errors over distances of greater than a hundred kilometres and relevellings, or adjustments, of the national levelling network often lead to spurious apparent jumps in mean sea level. In order to avoid the higher microseismic noise for gravity measurements immediately adjacent to the coastline, the report recommends that the absolute gravity measurements should be made at sites 1-10 km inland. The gravity site (which is normally in a building with reasonable tempera-
157
ture control) has also then to be connected to the T G B M by spirit levelling or GPS. Inland sites also enable a higher accuracy to be achieved for the calculation of ocean tide loading and attraction corrections. However, measurements for a few months with a well calibrated continuously recording relative gravimeter should enable corrections to be made to a few tenths of a microgal at any distance from the coastline (Baker et al., 1991). Although most of the measurement errors in the space geodesy and gravity measurement techniques have now been reduced to the level required to begin the measurements of vertical crustal movements outlined in this review, there are still some significant problems that remain to be addressed. The future programmes of measurements will identify the various approaches to reducing these problems. There have been great improvements in GPS receivers and analytical techniques during the last few years. Receiver noise and set up errors are typically a few millimetres. Improved GPS satellite tracking from fiducial (VLBI or SLR) stations will ensure that errors due to uncertainties in the orbits are reduced to the order of 1 part in 108 of the baseline length. Dixon (1991) reviews the precision and accuracy currently available with GPS as a function of baseline length. The vertical component is the most difficult to determine with GPS, since there is a geometric limitation due to the satellites only being observable in the upper hemisphere. It is also approximately a factor of 3 more sensitive to errors in the tropospheric calibration. Tropospheric errors due to the wet component depend upon baselength for the first 20 km or so (depending upon the tropospheric correlation length) and are then independent of baselength. Dixon (1991) presents measurements from California which, for the vertical component, have a repeatibility of the order of 2 cm for baselines from 20 to 500 km. He also shows that the repeatability is significantly worse in tropical areas. Future improvements in GPS measurements in the vertical will depend upon improved troposphere calibration and estimation techniques and may require the use of line of sight water vapour radiometers.
158
For absolute gravity measurements, variations in ground water can produce gravity variations at some sites that are in excess of 10/zgals. Corrections for these variations will require monitoring of ground water using boreholes, but in some areas modelling of the complex hydrology will be difficult. Superconducting gravimeters now have a very low drift rate and can be used to provide continuous measurements of gravity variations between absolute gravity measurements at some primary sites (Goodkind, 1986). It may be necessary to avoid sites that are likely to have significant ground water variations by choosing sites on crystalline bedrock, but this option is not available in many areas. The repeatability of order 2 /zgals achieved by the present generation of absolute gravimeters at some sites shows that, with care, suitable sites can be chosen. Absolute sea level measurement programmes have recently commenced in many countries. Absolute gravity measurements near tide gauges are now being made in North America, Japan and Europe. GPS measurement programmes at tide gauges have commenced in North America, Hawaii, South Africa, Japan, Australia and Europe. Gravity and space geodesy measurements in areas of post glacial rebound (e.g. North America, Fennoscandia, Scotland) will also test the models that are currently being used for correcting tide gauge records and such measurement programmes are also now underway. Many years of careful observations will be required for a reliable determination of the vertical crustal movements. During the next 2-3 years the results from the first repeat campaigns will be available and it will be interesting to see the progress that has been achieved.
Acknowledgements I would like to thank Phil Woodworth and Bob Edge for useful comments on the manuscript.
References Baker, T.F., Edge, R.J. and Jeffries, G., 1991. Tidal gravity and ocean tide loading in Europe. Geophys. J. Int., 107: 1-11.
w.v. BAKER Bilham, R., 1991. Earthquakes and sea level: space and terrestrial metrology on a changing planet Rev. Geophys., 29: 1-29.
Carter, W.E. and Robertson, D.S., 1990. Definition of a terrestrial reference frame using IRIS VLBI observations: approaching millimetre accuracy. In: C. Boucher and G.A. Wilkins (Editors), Symposium, 105: Earth Rotation and Coordinate Reference Frames. Springer, New York, pp. 115-122. Carter, W.E., Aubrey, D.G., Baker, T.F., Boucher, C., LeProvost, C., Pugh, D.T., Peltier, W.R., Zumberge, M., Rapp, R.H., Schutz, R.E., Emery, K.O. and Enfield, D.B., 1989. Geodetic fixing of tide gauge bench marks. Woods Hole Oceanogr. Inst. Tech. Rep. WHOI-89-31/CRC-89-5. Cerutti, G., Alosia, F., Germak, A., Bozzo, E., Caneva, G.. Lanza, R. and Marson, 1., 1992. The absolute gravity station and the Mt. Melbourne network in Terra Nova Bay North Victoria Land, East Antarctica. In: Proc. Sixth Int. Symp. Antarctic Earth Sciences, in press. Dixon, T.H., 1991. An introduction to the Global Positioning System and some geological applications. Rev. Geophys., 29: 249-276. Douglas, B.C., 1991. Global sea level rise. J. Geophys. Res., 96(C4): 6981-6992. Emery, K.O. and Aubrey, D.G., 1991. Sea Levels, Land Levels and Tide Gauges. Springer, New York, 237 pp. Goodkind, J.M., 1986. Continuous measurement of non-tidal variations of gravity. J, Geophys. Res., 91: 9125-9134. Hager, B.H., King, R.W. and Murray, M.H., 1991. Measurement of crustal deformation using the Global Positioning System. Annu. Rev. Earth Planet. Sci., 19: 351-382. IOC, 1989. Global Sea Level Observing System (GLOSS): Proposed Implementation Plan. Intergov. Oceanogr. Comm. Rep,, IOC-XV/8 Annex 4, 67 pp. IPCC, (J.T. Houghton, G.J. Jenkins and J.J. Ephraums, Editors) 1990. The Intergovernmental Panel on Climate Change Scientific Assessment. Cambridge Univ. Press, 365 pp. Kolenkiewicz, R., Smith, D.E., Pavlis, E.C. and Torrence, M.H., 1992. Direct estimation of vertical variations of satellite laser tracking sites from SL8 LAGEOS analysis. Ann. Geophys., 10(1), p. 111. Lambeck, K., 1990. Glacial rebound, sea level change and mantle viscosity. Q. J. R. Astron. Soc., 21: 1-30. Lambert, A., Liard, J.O., Courtier, N., Goodacre, A.K., McConnell, R.K. and Failer, J.E., 1989. Canadian absolute gravity program: Applications in geodesy and geodynamics. EOS, 70: 1447-1460. Marson, I., 1989. Esperimenti ed idee nella realizzazione di un gravimetro assoluto trasportabile. In: Atti dell'Istituto Veneto di Scienze Lettere ed Arti, 146: 41-58. Marson, I. and Failer, J.E, 1986. g - - T h e acceleration of gravity: its measurement and its importance. J. Phys. E: Sci. Instr., 19: 22-32. Peltier, W.R., 1986. Deglaciation-induced vertical motion of the North American Continent and transient lower mantle rheology. J. Geophys. Res., 91(B9): 9099-9123. Peltier, W.R. and Tushingham, A.M., 1989. Global sea level rise and the greenhouse effect: might they be related.'? Science, 244: 806-810.
ABSOLUTE SEA LEVEL MEASUREMENTS, CLIMATE CHANGE AND VERTICAL CRUST MOVEMENTS Peltier, W.R. and Tushingham, A.M., 1991. Influence of glacial isostatic adjustment on tide gauge measurements of secular sea level change. J. Geophys. Res., 96(134): 6779-6796. Peter, G., Moose, R.E., Wessells, C.W., Failer, J.E. and Niebauer, T.M., 1989. High precision absolute gravity observations in the United States. J. Geophys. Res., 94: 5659-5674. Pirazzoli, P.A., 1986. Secular trends of relative sea level changes indicated by tide gauge records. J. Coastal Res., 1: 1-26. Pugh, D.T., 1987. Tides, Surges and Mean Sea Level: a Handbook for Engineers and Scientists. Wiley, Chichester, 472 pp. Ray, J.R., Ma, C., Ryan, J.W., Clark, T.A., Eanes, R.J., Watkins, M.M., Schutz, B.E. and Tapley, B.D., 1991. Comparison of VLBI and SLR geocentric site coordinates. Geophys. Res. Lett., 18: 231-234. Spencer, R., Foden, P.R., McGarry, C., Harrison, A.J., Vassie, J.M., Baker, T.F., Smithson, M.J., Harangozo, S.A. and Woodworth, P.L., 1993. The ACCLAIM programme in the South Atlantic and Southern Oceans. Int. Hydrogr. Rev., 71(1): 7-21.
159
Thompson, K.R., 1986. North Atlantic sea level and circulation. Geophys. J. R. Astron. Soc., 87: 15-32. Torge, W., 1989. Gravimetry. De Gruyter, Berlin, 465 pp. Torge, W., R6der, R.H., Schniill, M., Wenzel, H.-G. and Failer, J.E., 1987. First results with the transportable absolute gravity meter JILAG-3. Bull. G6ocies., 61: 161-176. Trupin, A. and Wahr, J., 1990. Spectroscopic analysis of global tide gauge sea level data. Geophys. J. Int., 100: 441-454. 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(B3): 4497-4523. Woodworth, P.L., 1991. The permanent service for mean sea level and the global sea level observing system. J. Coastal Res., 7: 699-710. Woodworth, P.L., Spencer, W.E. and Alcock, G., 1990. On the availability of European mean sea level data. Int. Hydrogr. Rev., 67(1): 131-146. Wyrtki, K., 1975. El Nino--the dynamic response of the Equatorial Pacific Ocean to atmospheric forcing. J. Phys. Oceanogr., 5: 572-584.