email: zerbini@astbo1.bo.cnr.it
Tide gauges measure sea-level changes as variations in the relative position between the crust and the ocean surface. These measurements are difficult to interpret because they are influenced by several phenomena inducing vertical crustal movements. At present, vertical crustal motions at tide gauges can be measured to high accuracy independently of the sea-level reference surface by means of space techniques, therefore it will be possible to separate the crustal motions from the absolute sea-level variations. Tide gauge measurements are difficult to compare because tide gauges are referred to local reference systems and they have not yet been connected on a common datum. However, it should be pointed out that several international efforts are underway both at global (IOC, 1990) and regional scales which aim to overcome this problem. Nowadays, space geodesy techniques provide the possibility to connect tide gauges on a common datum by measuring via GPS their position to subcentimeter accuracy with respect to a well defined geocentric reference system such as the one, for example, established by the global network of Satellite Laser Ranging (SLR)/Very Long Baseline Interferometry (VLBI) fiducial stations. Moreover, it is available the International Terrestrial Reference System (ITRS) which has been established through an ensamble of SLR, VLBI, Global Positioning System (GPS) and Lunar Laser Ranging (LLR) stations (Boucher and Altamimi, 1993a, 1993b). The use of the ITRS has been recommended by the Commission on Mean Sea Level and Tides of the International Association for the Physical Sciences of the Ocean (IAPSO) in their well known Woods Hole report (Carter et al., 1989) and in the more recent report of the Surrey Workshop (Carter, 1994).
The SELF (SEa Level Fluctuations: geophysical interpretation and environmental impact) project, approved by ILP and financed by the Commission of the European Communities within the framework of the Environment program and developed in the time frame 1990-1992, was a pilot study (Zerbini et al., 1991) which provided a fundamental contribution to a better understanding of global change by providing a necessary base to succesfully approach the measurement of sea level fluctuations and to reliably assess the factors causing sea level rise. It involved several Institutions in 4 member states of the Community (Italy, Fed. Rep. of Germany, Greece and United Kingdom), Switzerland and Poland. The Institutions worked together to achieve the following objectives: to select, in the Mediterranean region, fiducial reference stations and well established tide gauges and to provide GPS links between the (SLR/VLBI) fiducial stations of the global network and the tide gauges; to improve GPS measurement procedures by using Water Vapor Radiometers to reduce vertical uncertainties to 1 cm or less; to perform measurements of absolute-g both at fiducial sites and tide gauges to monitor, with an independent system, vertical surface elevation changes; to perform, in selected areas of the Mediterranean basin, observations of geologic sea level markers of the past; to collect, analyze and interpret tide gauge data; to develop realistic models for tidal loading and tectonics in the Mediterranean region; to define corrections for the Earth's surface deformation due to exogenic causes and to study long-term variability of relative sea level.
Figure 1 illustrates schematically the measuring approach which has been adopted in the SELF project. Tide gauge measurements are referred to a permanent shore mark, the Tide Gauge BenchMark (TGBM). Simultaneous GPS observations were performed to tie the TGBM and the selected fiducial stations of the global reference system. Whenever the TGBM proved not to be suitable for GPS observations a new benchmark was installed and linked to the TGBM by means of high precision levelling. The global reference system adopted is the one provided by the SLR/GPS solution SSC(DUT)94C01R derived by the Delft University of Technology (Noomen et al., 1994). Several campaigns were carried out to link the selected tide gauges to the nearest fiducial reference station all across the Mediterranean basin from the Straits of Gibraltar as far as the Black Sea (Figure 2).
Simultaneous observations at tide gauges and reference stations were performed by using dual-frequency GPS receivers for, at least, 48 hours continuously with a sampling rate of 30 seconds. Simultaneously with GPS observations, two dual-frequency ground-based, transportable Water Vapor Radiometers (WVR) have been used in France, Italy and Greece to improve the estimation of the tropospheric path delay due to the water vapor content in the atmosphere. In the analysis, the WVR data have been combined with the GPS data in order to determine the station height to the best achiavable accuracy. Table 1 lists the GPS marker height in the SSC(DUT)94C01R reference system, the height difference between the GPS benchmark and the TGBM measured by means of high precision levelling, and finally the TGBM in the adopted reference.
Table 1. GPS benchmark height and TGBM height in the SSC(DUT)94C01R reference system, Height difference between the GPS benchmark and the TGBM.
Tide gauge site GPS marker height Height difference TGBM height in the SLR GPS-TGBM (m) in the SLR SSC(DUT)94C01R SSC(DUT94C01R ref. system (m) ref.system (m) Brindisi 42.648 +- 0.005 0.667 +- 0.001 41.981 +- 0.005 Catania 51.977 0.009 9.007 0.002 42.970 0.009 Pt. Corsini 39.820 0.006 -0.536 0.001 40.356 0.006 Venice 44.629 0.014 -2.031 0.002 46.660 0.014 Trieste 52.981 0.006 7.451 0.002 45.530 0.006 Genoa 49.306 0.002 2.518 0.002 46.788 0.003 Marseille 61.753 0.003 11.204 0.001 50.549 0.003 Preveza 28.403 0.007 0.525 0.001 27.878 0.007 Patrai 28.406 0.008 1.374 0.002 27.032 0.008 Kalamai 26.976 0.013 0.007 0.001 26.969 0.013 Soudhas 38.968 0.017 14.735 0.002 24.233 0.017 Piraieus 39.660 0.004 0.644 0.001 39.016 0.004 Siros 40.028 0.005 0.411 0.002 39.617 0.005 Katakolon 28.071 0.003 3.709 0.002 24.362 0.004 Rhodos 29.184 0.003 7.265 0.002 21.919 0.004 Iraklion 27.356 0.008 1.909 0.003 25.447 0.009 Katsively 30.467 0.007 5.685 0.002 24.782 0.007 Tuapse 18.722 0.004 0.000 0.000 18.725 0.004 Algeciras 44.537 0.003 -0.471 0.002 45.008 0.004 Tarifa 46.524 0.002 2.691 0.002 43.833 0.003 Cadiz 52.295 0.005 3.317 0.002 48.978 0.005 Ceuta 44.887 0.003 -0.673 0.002 45.560 0.004 Gibraltar 45.520 0.003 -0.480 0.002 46.000 0.004Absolute gravity has also been measured both at the tide gauges and at the reference sites (Table 2). The average precision of the measurements of absolute g is in the order of 1.5 ūgal; since the free-air vertical gradient of g is equal to 3 ūgal/cm it is possible to detect changes in gravity due to pure height variations of the order of 1 cm, which is comparable with the precision of the height determinations achieved with the GPS technique. The combination of space geodetic measurements and absolute g observations can therefore provide a vertical datum accurate to, at least, the 1 cm level which is a necessary requisite in order to be able to detect and control sea level fluctuations most likely induced by global warming and to separate this signal from that which might be originated by tectonic movements or exogenic forces.
Table 2. Absolute gravity values.
Site Epoch g sigma h0 (microgal) (microgal) (m) Basovizza 1991.85 980 568 052 0.8 0.903 Matera 1991.85 980 185 527 0.9 0.925 Medicina 1992.85 980 473 607 1.5 0.916 Medicina 1995.10 980 474 790 1.6 0.900 Marseille 1993.40 980 485 131 1.2 0.931 Grasse 1993.40 980 216 056 0.9 0.929 Noto 1993.50 979 992 652 1.5 0.948 Genoa 1993.40 980 558 119 1.0 0.931 Venice 1994.95 980 635 142 3.2 0.000 Naples 1994.40 980 257 805 4.0 0.982 Pt. Corsini 1995.10 980 483 313 11.5 0.950 Brindisi 1994.95 980 293 481 3.0 0.000 Catania 1994.85 980 044 166 3.5 0.000 Karitsa 1993.90 979 913 519 1.3 0.920 Askites 1993.85 980 250 029 1.0 0.935 Ermoupoli 1993.85 980 048 152 2.0 0.951 Katakolon 1993.85 979 897 950 2.1 0.948 Xrisokellaria 1993.85 979 854 674 0.7 0.942 Kattavia 1993.85 979 861 917 0.7 0.939 Roumelli 1993.75 979 827 682 0.9 0.947 Dionysos 1993.75 979 961 222 1.8 0.943 San Fernando 1994.20 979 826 339 3.5 1.312 Tarifa 1994.20 979 737 252 3.8 1.323 Ceuta 1994.20 979 755 018 3.1 1.318 Simeiz 1994.45 980 566 746 12.4 1.307 Tuapse 1994.45 980 542 128 28.2 1.316 Zelenchukskaya 1994.45 980 249 154 2.6 1.306 Baksan 1994.45 979 909 691 2.6 1.316Modelling the sea level decadal variations can improve the local trend estimates or even allow for a detection of an anthropogenic acceleration in the trend. Furthermore, interannual to decadal fluctuation in sea-level may in itself prove useful as an effective indicator of climate changes. A substantial fraction of the intraannual to decadal fluctations in coastal sea-level is non-steric and forced mechanically by the atmosphere. To some extent, coastal sea-level integrates over the atmospheric forcing, thus enhancing the long-period part over shorter periods. Thus, studies of the interannual fluctuations in coastal sea-level may reveal changes in the mechanical forcing of the atmosphere, i.e. in the regional atmospheric circulation, that cannot be detected directly from the meteorological data. Analysis of the tide gauge data has been performed in the framework of the SELF project. Mixed quality tide gauge data are available for the Mediterranean Sea. Although trend values can easily be assigned, the analysis indicates that records at least 40 years long should be used if errors less than 0.5 mm/yr are required. The analyses of the monthly sea level data reveal an unexpectedly large variability on the coastal seasonal tidal constituent which is spatially highly coherent. This variability on dacadal time scales most likely is associated with changes in the regional atmospheric circulation.
The trends determined from the longer RLR records available in the Mediterranean are compiled in Table 3 together with the relative sea-level rise expected from isostatic compensation due to post-glacial rebound. In the Mediterranean, this effect is of the order 0.3 mm/a. Thus, at tectonically stable sites we should expect a relative sea-level rise close to the global one.
To improve local trend estimations from shorter records a better understanding of the response of sea level to the various forcing parameters is needed. Hydrodynamical models to simulate sea-level variability due to meteorological forcing and steric effects would be an appropriate tool to separate these effects from the long-term sea-level changes. However, in the absence of such models, other means can be utilized, which are based on the assumption, that the interannual to multidecadal sea-level variability is spatially coherent within properly defined regions. Thus, in regions, where a long and qualitatively good record is available, this record may be used as "base record''. For shorter records, trends are estimated from the differences of monthly (or annual) means to the base record thus eliminating the synchroneous part of the sea-level variations (Sjberg, 1987). For the western part of the Mediterranean, Marseille is a potential base station, while Trieste could serve as such for the eastern part. The trends determined with Marseille as base station tend to be slightly larger than those using Trieste (Table 3). Especially for records spanning the end of the last and the beginning of the present century (marked with an asterisk in Table 3), Trieste introduces a negative bias, which may be due to the small overlap of these records with Trieste. For records spanning the second half of the century, both Marseille and Trieste as base records improve the trends in the sense that the interstation scatter is reduced. At one of the two records at Split (Rt. Mar.), the effects of using Trieste and Marseille are opposite, possibly indicating some data problems. As it can be seen from Table 3, at least at the tide gauges included in the present study, crustal movements are small compared to the decadal to multidecadal sea- level variability discussed above but of the same order as the long-term trend in sea level, thus necessitating a careful monitoring if crustal movement is to be separated from the oceanographic contribution to relative sea-level changes.
Table 3. Local trends in relative sea level.
The stations are sorted according to the available data. N: number of available monthly sea-level values, t: sea-level trend in mm/a, sigmat: standard error in mm/a, tP: trend due to postglacial rebound as computed with the ICE-3G model (Peltier and Tushingham, 1989), tTr/tMa: trends in mm/a determined using Trieste and Marseille, respectively, as base record. For records marked with an asterisk, the overlap with Trieste is small, explaining the large differences compared to the trends derived with Marseille as base station. rC: crustal movement rates (positive for uplift) decontaminated for post-glacial rebound effects; these rates given are calculated from rC = -(tMa - tP - e), where e is the eustatic sea-level change. We assumed e = 1.8 mm/a (Douglas, 1992).
Station Begin End N t sigmat tP tTr tMa rC Marseille 1885 1989 1157 1.1 0.1 -0.2 1.03 1.10 0.5 Genova 1884 1988 994 1.3 0.1 -0.2 0.98 1.21 0.5 Trieste 1905 1990 960 1.1 0.2 -0.3 1.10 1.17 0.3 Lagos 1908 1989 863 1.4 0.2 -0.4 1.39 1.42 0.0 Tuapse 1917 1990 844 2.2 0.3 -0.1 1.99 2.17 -0.3 Bakar 1930 1990 600 0.9 0.3 -0.3 1.04 1.69 -0.2 Split Rt Mar. 1952 1990 450 0.1 0.5 -0.3 -1.29 0.73 0.8 Split Harbour 1954 1990 44 -0.8 0.5 -0.3 -0.05 1.46 0.0 Cagliari 1896 1934 433 1.3 0.4 0.2 1.15 1.74 0.3 Rovinj 1955 1990 424 -0.2 0.5 -0.3 0.57 2.21 -0.7 Dubrovnik 1956 1990 419 -0.1 0.5 -0.3 0.58 2.10 -0.6 Alicante II 1960 1987 332 -1.4 0.4 -0.3 -0.18 1.15 0.3 Koper 1962 1990 332 -0.5 0.6 -0.3 1.07 2.21 -0.7 Bar 1964 1990 18 1.4 0.7 2.57 3.25 P. Maurizio 1896 1922 309 1.1 0.6 -0.1 1.52 2.71 * -1.0 Civitavecchia 1896 1922 304 1.2 0.7 0.0 0.59 1.43 * 0.4 Alicante I 1952 1987 303 -2.2 0.4 -0.3 -2.71 -0.24 1.7 Napoli (Man.) 1896 1922 301 2.1 0.7 0.0 2.18 3.25 * -1.5 Palermo 1896 1922 294 1.0 0.7 0.3 -1.39 1.15 * 0.9 Venezia (S.St.) 1896 1920 288 4.4 1.2 -0.3 2.11 4.72 * -3.2 Port Said 1923 1946 287 4.7 1.0 5.20 4.20 Venezia (Ars.) 1889 1913 287 1.8 1.4 -0.3 3.58 4.60 * -3.1 Gibraltar 1961 1989 272 -0.7 0.8 -0.6 0.24 0.15 1.0 Napoli (Ars.) 1899 1922 263 2.5 1.5 0.0 1.94 3.00 * -1.2 Porto Corsini 1969 1972 45 -3.1 17.99 -0.97
References
Boucher C. and Z. Altamimi, 1993a. Development of a Conventional Terrestrial Reference Frame. In Contributions of Space Geodesy to Geodynamics: Earth Dynamics, American Geophysical Union, Geodynamics Series, 24:89-97.
Boucher C. and Z. Altamimi, 1993b. Contribution of IGS92 to the TRF. In proceedings of 1993 IGS workshop, Bern, March 25-26, 1993, 175-183.
Carter W.E., D.G. Aubrey, T. Baker, C. Boucher, C. LeProvost, D. Pugh, W.R. Peltier, M. Zumberge, R.H. Rapp, R.E. Schutz, K.O. Emery and D.B. Enfield, 1989. Geodetic fixing of tide gauge bench marks, Technical Report, CRC-89-5, Coastal Research Center, WHOI-89-31, pp. 46.
Carter W.E. (Edt.), 1994. Report of the Surrey Workshop of the IAPSO Tide Gauge Bench Mark Fixing Committee. Deacon Laboratory, Godalming, Surrey, United Kingdom, December 13-15, 1993, NOAA Technical Report NOSOES0006.
Douglas B.C., 1992. Global sea level acceleration. Journ. Geophys. Res., 97:12699 - 12706.
IOC, 1990. Global Sea Level Observing System (GLOSS): Implementation Plan, Intergovernamental Oceanographic Commission, Technical Series, 35, Unesco, pp. 90.
Noomen R., T.A. Springer, B.A.C. Ambrosius, K. Herzberger, D.C. Kuijper, G.J. Mets, B. Overgaauw and K.F. Wakker, 1994. Crustal deformations in the Mediterranean area computed from SLR and GPS observations, sumitted to Journal of Geodynamics.
Peltier W.R. and A.M. Tushingham, 1989. Global Sea Level Rise and the Greenhouse Effect: Might They Be Connected? Science, 244:806-810.
Sjoberg L.E., 1987. Comparison of some methods of determining land uplift rates from tide gauge data. ZfV, 2:69-73.
Zerbini S., T. Baker, H.G. Kahle, G. Veis and P. Wilson, 1991. SEa-Level Fluctuations: geophysical interpretation and environmental impact (SELF). Proposal submitted to the Commission of the European Communities for the Environment Programme, pp. 50.