Scientiﬁc objectives of current and future WEGENER activities et al wegener 1998.pdfELSEVIER Tectonophysics 294 (1998) 177–223 Scientiﬁc objectives of current and future WEGENER

ELSEVIER Tectonophysics 294 (1998) 177–223

Scientific objectives of current and future WEGENER activities

Hans-Peter Plag a,Ł, Boudewijn Ambrosius b, Trevor F. Baker c, Gerhard Beutler d,Giuseppe Bianco e, Geoffrey Blewitt f, Claude Boucher g, James L. Davis h, John J. Degnan i,Jan M. Johansson j, Hans-Gert Kahle k, Irina Kumkova l, Iginio Marson m, Stephan Mueller n,

Erricos C. Pavlis i, Michael R. Pearlman o, Bernd Richter p, Wim Spakman q, SuriyaK. Tatevian r, Paolo Tomasi s, Peter Wilson t, Susanna Zerbini u

a Norwegian Mapping Authority, Hønefoss, Norwayb Faculty of Aerospace Engineering, Delft University of Technology, Delft, Netherlands

c Proudman Oceanographic Laboratory, Bidston, Merseyside, UKd Astronomical Institute of Berne, Berne, Switzerland

e Agenzia Spaziale Italiana, Matera, Italyf Department of Surveying, University of Newcastle upon Tyne, Newcastle upon Tyne, UK

g Institut Geographique National, DR5, Saint-Mande, Franceh Harvard–Smithsonian Center for Astrophysics, Cambridge, USA

i NASA, Goddard Space Flight Center, Greenbelt, USAj Onsala Space Observatory, Chalmers University of Technology, Onsala, Sweden

k Institut für Geodäsie und Photogrammetrie, ETH-Hönggerberg, Switzerlandl Institute of Applied Astronomy, Russian Academy of Sciences, St. Petersburg, Russia

m DINMA, University of Trieste, Trieste, Italyn Swiss Federal Institute of Technology (ETH), ETH-Geophysics, Zürich, Switzerland

o Smithonian Astrophysical Observatory, Cambridge, USAp Institut für Angewandte Geodäsie (now Bundesamt für Kartographie und Geodäsie), Frankfurt, Germany

q Faculty of Earth Sciences, Geodynamic Research Institute, Utrecht University, Utrecht, Netherlandsr Space Geodesy Department, Astronomical Council, Russian Academy of Sciences, Moscow, Russia

s Istituto di Radioastronomia CNR, Area di Ricerca del CNR, Bologna, Italyt Geoforschungszentrum Potsdam, Potsdam, Germany

u Department of Physics, University of Bologna, Bologna, Italy

Received 10 May 1997; accepted 15 December 1997

Abstract

The WEGENER group has promoted the development of scientific space-geodetic activities in the Mediterraneanand in the European area for the last fifteen years and has contributed to the establishment of geodetic networksdesigned particularly for earth science research. WEGENER currently has three scientific objectives which are related toplate-boundary processes, sea-level and height changes, and postglacial rebound. In a full exploitation of the space-geodetictechniques, namely SLR, VLBI and GPS, the individual scientific projects do not only pursue these objectives butalso contribute to improving and developing the observation techniques as well as the modelling theories. In the past,particularly SLR observations within WEGENER-MEDLAS have provided a fundamental contribution to determine

Ł Corresponding author. Tel.: C47 (32) 118 474; Fax: C47 (32) 118 401; E-mail: [email protected]

0040-1951/98/$19.00 1998 Elsevier Science B.V. All rights reserved.PII: S 0 0 4 0 - 1 9 5 1 ( 9 8 ) 0 0 1 0 0 - 0

178 H.-P. Plag et al. / Tectonophysics 294 (1998) 177–223

the regional kinematics of the tectonic plates in the Mediterranean with high precision. With GPS, spatially denser sitedistributions are feasible, and in several WEGENER projects detailed studies of tectonically active areas were possible onthe basis of repeated episodic GPS observations. Current projects associated with WEGENER are successful in separatingcrustal movements and absolute sea-level variations as well as in monitoring postglacial rebound. These tasks requirehigh-precision height determinations, a problem central to all of the present WEGENER activities. In these projects,continuously occupied GPS sites are of increasing importance. Time series of heights observed with continuous GPS canbe determined with a few centimeters RMS error thus enabling the reliable estimates of vertical rates over relatively shorttime intervals. Regional networks of continuous GPS sites are already providing results relevant, for example, for thestudy of postglacial rebound. The Mediterranean area is an extraordinary natural laboratory for the study of seismotectonicprocesses, and the wealth of observations acquired in previous WEGENER projects together with new space-geodetic ob-servations will allow the test of geophysical hypotheses linking three-dimensional deformations of the Earth’s surface to thedynamics of the Earth’s interior. In particular, it is anticipated that WEGENER projects will aim at a test of the slab-detach-ment hypothesis. The complex investigations on sea-level fluctuations presently carried out at basin scale from the Strait ofGibraltar to the Black Sea make it possible to study the present and recent past interactions of ocean, atmosphere and solidEarth, as well as to develop appropriate models to assess future aspects. 1998 Elsevier Science B.V. All rights reserved.

Keywords: Space Geodesy; gravity; sea level; post-glacial rebound; African-Eurasian plate boundary; 3-D deformation ofthe Earth

1. Introduction

The ‘Working group of European Geoscientists forthe Establishment of Networks for Earth-Science Re-search’ (WEGENER) was established in the begin-ning of the 1980s as an inter-disciplinary group cen-tered on the application of space-geodetic and othertechniques to the study of geodynamics. A descriptionof the evolution of WEGENER over the last one and ahalf decades is presented by Wilson (1998). In 1991,the ‘WEGENER Project: Geodetic Investigations Re-lated to the Kinematics and Dynamics of the African,Arabian and Eurasian Plates’ was established as Spe-cial Commission SC6 of the International Associa-tion of Geodesy (IAG). It is the responsibility of WE-GENER to organize parts of the international geodeticand geophysical community in a concerted effort toproduce high-accuracy and coherent data and valuableresults relevant to the three objectives defined below.

The present general study fields were basicallyworked out during meetings in 1990 and 1991 inresponse to NASA’s ‘Dynamics of the Solid Earth’(DOSE) Announcement of Opportunity and as anatural development of the scientific activities car-ried out by the WEGENER group in the courseof NASA’s Crustal Dynamics Project (CDP) whichbegan in the early eighties.

After the initial period of growth in both the num-ber of scientists involved or related to WEGENER

activities and the geographical areas covered byprojects carried out within its frame, WEGENERhas been restructured in 1995 following the electionby the IAG of a new president (S. Zerbini) and theformation of a new directing board. During the firstmeeting of the new board (Bologna, October 1995),the scientific objectives were reviewed and revisedaccording to the most recent developments in thescientific areas of interest to the project.

Achieving the WEGENER scientific objectivesrelies very much upon the acquisition of high-accu-racy data in the experiments. Therefore, it is a ma-jor concern of the group that the most appropriatetechniques continue to be available or will be devel-oped whenever possible. WEGENER is maintaininga close contact to the agencies and institutions re-sponsible for the development and maintenance ofthe global space-geodetic networks in order to makethem aware of the scientific needs and outcomes ofthe project which might have an influence on thegeneral science policy trends.

The purpose of this paper is to clearly definethe objectives, identify scientific hypotheses to betested, and outline the approaches to be taken andthe improvements needed to achieve these objectives.In that, the paper (1) outlines past results and currentfocus of three prominent areas of geosciences and (2)demonstrates to the wider scientific community thecapabilities of space-geodetic techniques to provide

H.-P. Plag et al. / Tectonophysics 294 (1998) 177–223 179

useful constraints for the scientific problems in thesefields. Moreover, the paper documents the width andinterdisciplinarity required in empirical studies ofgeoscientific problems.

2. Scientific objectives of WEGENER: anoverview

The present main objectives of WEGENER are:(1) to study the three-dimensional deformations andgravity along the African–Eurasian plate boundariesand in the adjacent deformation zones in order tocontribute to a better understanding of the associatedgeodynamical processes; (2) to monitor the three-dimensional deformations in a large region centeredaround Fennoscandia in order to determine the mag-nitude and extent of the present-day postglacial re-bound in that area, thereby extending our knowledgeabout the viscoelastic properties of the Earth; (3) toinvestigate height and sea-level variations in orderto identify and separate the processes contributing tothese variations.

In parallel to these objectives, the mutual im-provements of the measurement techniques, the test-ing of new technological means and proposals fornew missions, as well as the synthesizing of ob-servable quantities and the inversion of the obser-vations for geodynamically relevant parameters areparts of all the WEGENER activities. In combi-nation, the three objectives make extensive use ofthe most advanced space-geodetic and gravity tech-niques to contribute to scientific fields associatedwith three-dimensional surface deformations, whichare not only relevant to basic science but also havepotentially significant economic and social benefits.

Plate boundaries are, in general, geographical ar-eas associated with high risks of natural disasters.A principal key to understanding the plate bound-ary processes, including the driving forces, is adetailed knowledge of the kinematics and of theassociated gravity changes. In particular, a synthesisof the structural information derived, for example,from seismic tomography and the present-day kine-matics determined with the geodetic measurementswill establish more strict or even novel constraintsfor geodynamical models of plate boundary pro-cesses. Moreover, these data contribute to the frame-

work required for an assessment of natural hazardrisks.

The present-day deformations and changes ingravity associated with the ongoing postglacial re-bound are an important augmentation of the exist-ing data-sets related to glacially induced deforma-tions such as the Pleistocene and Holocene sea-level changes, secular polar motion, and secularchanges of the low-degree geopotential. The inter-actions of glacial loads, crustal deformations andsea-level changes over the past 100 ka constitute cru-cial boundary conditions for paleo-climate models.The uncertainties in our knowledge of the rheologyof the Earth’s mantle are a basic limitation for thequality of, for example, reconstructed paleo-topogra-phies or geophysically determined ice models. Suchknowledge is required as input for paleo-climate re-constructions, and as such it is of utmost significancefor the quality of paleo-climate reconstructions.

Sea-level variations and in particular secularchanges of sea level play a prominent role in the cli-mate-change discussion. Climate variability on timescales from decades to centuries is presently cominginto focus but is not very well understood (see for ex-ample, Crowley and Kim, 1993; Rind and Overpeck,1993). However, it is clear that at these time scales,the ocean is a major component of the climate system,and studying the sea-level variability will contributeto an improvement of our understanding of the rele-vant processes. Europe has coastal areas of consider-able extent and ecological and economical value. Un-derstanding sea-level variations on time scales of upto centuries is crucial for an integrated and sustain-able management of coastal zones. The anticipatedanthropogenic environmental changes are likely to in-duce significant changes in future sea levels, be it inthe frequency of storm surges, tidal ranges or seculartrends. Thus, the need to develop the capability of pro-viding current rates and predicting future variations insea level on local, regional and global scales is fullyunderstood by the international scientific community,as is expressed in recently developed projects such asLOICZ (Holligan de Boois, 1993) and internationalactivities of the IAPSO Commission on Mean SeaLevel and Tides (Carter, 1994), the ILP (Fard, 1994),and the IOC-Euro-Gloss project (Baker et al., 1996).

Sea-level records longer than a decade originateexclusively from coastal tide gauges. To determine


secular changes from such records is problematic,because tide gauges provide sea levels only relatedto a benchmark on land. In order to obtain ab-solute sea levels, crustal movements and sea-levelvariations have to be separated, which necessitates amonitoring of the crustal height variations. Thus, theheight determination problem is closely related to theobservation of coastal sea level. Moreover, in orderto interpret correctly the observed height variations,crustal movements resulting from tectonic forces andpostglacial rebound need to be known. This clearlydemonstrates the synergistic nature of the three WE-GENER objectives: the sea-level problem cannot besolved correctly without solving the two other prob-lems attacked by the WEGENER activities.

In Section 3, we will first summarize the currentstatus in monitoring three-dimensional deformationsand gravity with respect to the various techniquesand the time scales of the variations. Section 4considers geodetic networks from the WEGENERperspective with special emphasis on the interactionsbetween the WEGENER activities and the networks.The three main objectives are considered separatelyin Sections 5–7, each starting with a summary of theproblem, then discussing scientific hypotheses to betested, which might contribute to a better understand-ing of the problem, and, after a short description ofongoing projects, concluding with an identificationof future approaches and activities apt to contributeto the objectives. The final section summarizes thecurrent status related to the WEGENER objectivesand outlines the major scientific developments to beexpected for the near future.

3. Monitoring three-dimensional deformationsand gravity

3.1. Space geodesy: techniques, organisations, anddata products

Space geodesy went through a remarkable de-velopment during its 35 years of existence. Threeorganisations dealing with space geodesy shall bementioned briefly, namely the International EarthRotation Service (IERS), the International GPS Ser-vice for Geodynamics (IGS), and the Commissionon International Coordination of Space Techniques

for Geodesy and Geodynamics (CSTG). The firsttwo are service-type organizations, the last is a jointcommission of the IAG and of COSPAR, the aimof which is to coordinate existing and stimulate newactivities in the research involving space techniques.Very close links between the three organizations arerequired to guarantee a meaningful development inthis area of research and application. More informa-tion concerning the IERS is given in the series oftechnical reports issued by the IERS, while Beut-ler et al. (1996) provide an overview of the IGS,and Beutler and Drewes (1996) discuss the currentCSTG policy and important issues like the role ofdifferent space techniques today and in the future.

Whereas optical astrometry was the primary ob-servational tool in the pioneer era, Satellite LaserRanging (SLR) and Very Long Baseline Interferom-etry (VLBI) were the main contributors in the lateseventies and in the eighties for high-accuracy tasks.The IERS Terrestrial Reference Frame (ITRF) up tothe version ITRF92 almost uniquely relied on SLRand VLBI, and the IERS Celestial Reference Frame(ICRF) relies completely on the VLBI technique.Almost everything we know today about the Earth’sgeocenter and gravity field stems from the SLR ob-servation technique. Since the end of the seventiesuntil the early nineties the transformation parametersbetween the two systems ITRF and ICRF were alsoalmost uniquely determined by VLBI and SLR.

In the late eighties and the early nineties a verypowerful contributor to regional and global geody-namics — with the advantage of a low-cost terres-trial segment (as compared to both, VLBI and SLR)— came up in the form of the Global PositioningSystem (GPS). Particularly since the IGS started itsoperations in June 1992, GPS may be used in a veryconvenient and productive way for regional geody-namics projects. Today’s IGS products are estimatedto be accurate on the level of 10 cm for the satelliteorbits, 0.1–0.2 ms for the pole positions x and y, and0.03 ms (time) in the length of day. This accuracyis comparable to those of VLBI and SLR. The IGSproducts allow a relative positioning in regions ofcontinental (or even global) size with an accuracy ofabout 0.5 cm in the horizontal, and of about 1 cmin the vertical position with site occupation times ofa few days provided there are some permanent IGSsites in the region of interest. Because IGS products


of highest accuracy are available rather rapidly (11days is the maximum delay) after the observations,there is indeed an almost unlimited number of sci-entific applications made possible through GPS. Theavailability of IGS products are of major importancefor the WEGENER objectives.

It should be mentioned that recently other (thanGPS) micro-wave satellite systems like DORIS,GLO-NASS, and PRARE started to contribute toregional and global geodynamics. It can be expectedthat more of these systems will be realised in thenear future. Depending on the availability of thedata, the Technology Committee of the WEGENERboard will promote the integration of these systemsin present and future activities within WEGENER.

Taking into account the achievements of thesenew systems, IERS progressively included GPS andthen DORIS among the contributing techniques.Consequently, the ITRF92 solution included for thefirst time GPS contributions and ITRF94 includedfurthermore DORIS contributions. This greatly im-proved both the IERS network by giving it a substan-tially better geographical coverage and improvingits reliability as well as the accuracy of the linksbetween the frames underlying each technique.

It is nowadays widely recognized that the high-precision determination of particularly vertical crustalmotion (as required, for example, in sea-level studies)requires a colocation of space-geodetic positioningmethods and gravity. While the first techniques pro-vide high-accuracy kinematic data, the variations ofthe gravity field provide information on the dynamicsof the process (i.e., constraints on mass movements)as well as an extensive check on the vertical surfacemovements with a completely independent system.Several resolutions or conclusions of scientific bod-ies state in fact that “absolute gravity measurementsshould be made at all the space-geodetic primary sta-tions (VLBI, SLR and GPS) and near as many indi-vidual tide gauges as possible” (Carter, 1994). Thefocus of attention is on absolute gravity because ab-solute gravimeters provide a powerful tool for studiesof long-term changes of the gravity field. It shouldbe stressed, however, that the spatial and short-termtemporal variability of the gravity field needs to bemonitored as well, which can be accomplished by us-ing high-resolution relative gravimeters such as thesuperconducting ones.

3.2. Reference frames

The study of deformations requires the adoptionof consistent reference systems. Therefore the pre-cise definition and realization of such systems is akey element for those investigations. It is clear thatthe uncertainty of such realizations should be lessthan the level of the signals ascribed to the deforma-tion processes. Though regional reference frames canbe used, most of the time, a global system regardingthe whole Earth is better. Such systems are in factalmost mandatory as soon as space techniques areused. The main kinematical characteristics are theirorigin, scale and orientation, as well as their timeevolution. Such problems have already been inves-tigated by groups which have produced global platetectonic models (e.g. DeMets et al., 1990, 1994). An-other recent example of such connections betweenreference frames and scientific results is the directlink between global sea-level variations derived fromsatellite radar altimetry and the secular scale changeof the relevant terrestrial reference system.

Reference frames are also of physical interestas soon as they are clearly linked to the Earth asa massive deformable body. In particular, the linkbetween its origin and the physical geocenter is ofspecial interest. In addition, the motion in space ofsuch a frame is also of major interest for globalgeodynamics (precession, nutation, Earth rotation).

As mentioned in Section 3.1, space-geodetic tech-niques such as VLBI, SLR and GPS provide high-accuracy three-dimensional station coordinates. Eachof these techniques contributes to the definition of aterrestrial reference frame which might consist of thestation coordinates at a given epoch and of the relevantstation velocities. Recent comparisons of different in-dividual solutions determined by using either SLR orVLBI data have demonstrated that the solutions agreeat the subcentimeter level in the horizontal compo-nents and at the 20-mm level in the vertical one (Ray etal., 1991; Himwich et al., 1993; Watkins et al., 1994).

As already mentioned, since its creation in 1988IERS has produced a series of terrestrial referenceframes from a combination of all submitted solutionsusing the currently available space techniques. Thedefinition of the system and its realizations are fullydescribed in the IERS publications and specificallythe IERS Conventions (McCarthy, 1996).


The most recent solution is ITRF94 (Boucher etal., 1996) which provides a set of positions andvelocities for more than a hundred sites distributedover the globe and obtained by a combination ofvarious single-technique solutions (4 VLBI, 2 SLR,3 GPS, and 3 DORIS). The accuracy is estimated tobe better than one centimeter for the best sites.

3.3. Current accuracies and limitations

Based on space-geodetic methods, three-dimen-sional deformations can be monitored with a rapidlyimproving accuracy. In this section, a short summaryof the status of the three most widely used techniquesis given, namely GPS, SLR and VLBI. In addition,the status of absolute gravimetry is discussed.

3.3.1. GPSThe accuracies of coordinates, Earth Orientation

Parameters (EOPs), etc. quoted in Section 3.1 so farapply to mean values over one or even several days.There are, however, indications (but not yet conclu-sive evidence) that the GPS will be as successful ingiving a high sub-diurnal temporal resolution as it isin giving mean values.

For most scientific purposes, GPS is used asan interferometric technique, i.e. highest accuraciesare achievable only if the difference of the original(phase) observables are analysed. The principles ofobservation and of processing are closely related tothose of VLBI (radio-band of the electromagneticspectrum, interferometric approach). VLBI is abso-lute in the sense that no orbits have to be modeled.Only station positions (and velocities), the effects ofEarth rotation, and the tropospheric refraction haveto be modeled by VLBI (once the celestial referenceframe is fixed). Because some perturbing accelera-tions, in particular those related to radiation pressure,may not be known with sufficient accuracy (but haveto be estimated instead), the GPS-derived LOD-val-ues may even be biased due to correlations with theestimated radiation pressure parameters. Therefore,GPS (as every satellite geodetic technique) will al-ways have to rely on VLBI for calibration purposes.

GPS and VLBI are making use of the radio-bandof the electromagnetic spectrum. Both techniques aretherefore fully exposed to the so-called wet tropo-spheric delay which is very difficult to predict on

the accuracy level required. In both techniques thetropospheric delay has to be estimated. The modelsare getting more and more complex as they evolvefrom conventional time- and site-specific parametersto stochastic processes, from pure zenith-distance tozenith- and azimuth-dependent models. The morecomplex the model, the better the observable maybe represented. On the other hand, the danger ofcorrelations of troposphere parameters with otherparameters is growing, too, with the complexity ofthe model. In GPS we know in particular that thereare strong correlations between station heights andtroposphere parameters (see Section 3.5). In this con-text SLR is an ideal calibration tool: SLR is workingin the optical part of the spectrum, where the tropo-spheric delay may be modelled with an accuracy wellbelow the centimeter level even with modest knowl-edge of the atmosphere surrounding a given observa-tion site. Therefore, at this high-accuracy level GPSmust rely on SLR as a calibration tool, for example,for the definition of a height datum or for the estima-tion of the geocenter. This calibration aspect will beof even greater importance if high time resolution isrequired, for example, when trying to model the sub-diurnal station motions or the sub-diurnal motion ofthe ephemeris pole on the Earth surface.

It is also predictable that GPS will be the work-ing horse in all future regional geodynamics projects(see also Bevis et al., 1997). Thus, it is evident thatthis technique will play a central role in all futureWEGENER projects. While currently a large frac-tion of GPS observations is still originating fromcampaigns with episodic occupations of sites, thereis a clear trend towards more continuously oper-ating GPS (CGPS) sites. Moreover, strategies forcombining CGPS stations providing high temporalresolutions with spatially dense campaigns can beexpected to be developed (e.g. Bevis et al., 1997).

3.3.2. SLRThe round trip time-of-flight of an ultra-short (a

few picoseconds) laser pulse from a ground station toa satellite is certainly the most straightforward and ac-curate observable of all the competing space-geodetictechniques. In the best systems, systematic opticaland electronic propagation delays in the instrumenta-tion are usually assessed through colocation of two ormore systems prior to deployment and then controlled


at the few-millimeter level via frequent ground cali-brations to well-surveyed targets. Fits of SLR nor-mal point data around a short orbital arc generallyhave an RMS of 1 to 3 mm for the best systems, butperformance varies widely within the existing globalnetwork of approximately 45 stations.

Of all the space-geodetic techniques, SLR suf-fers the least from propagation delay variabilitiesin the atmospheric channel. Compared to the mi-crowave frequencies utilized by VLBI, GPS, DORISand PRARE, SLR’s optical frequencies are relativelyinsensitive to the two most dynamic (and henceleast predictable) components of the atmosphericrefraction delay, i.e. the ionosphere and water va-por distribution. Ions are too heavy and sluggish torespond to optical frequencies in the 300 to 900 ter-ahertz range, and laser wavelengths are typically farfrom strong absorption features in the water vaporspectrum. Specifically, the effect of water vapor onlaser ranges is roughly 70 times smaller than onmicrowave distance measurements (Hauser, 1989).Thus, the so-called ‘dry’ component of the atmo-sphere is the principal contributor to the propagationerror in SLR, and subcentimeter atmospheric correc-tions can be applied relatively easily (for elevationangles greater than 20º) through onsite meteorologi-cal measurements and spherical shell models of theatmosphere (Degnan, 1993).

Another advantage of SLR is the simplicity andlow cost of the space segment. Unlike microwave-based satellite systems which all require power, an-tenna structures, and occasional maneuvers to sup-port these subsystems, SLR uses relatively inexpen-sive spherical satellites equipped with passive retrore-flectors. Geodetic laser-ranged satellites, such as LA-GEOS I and II and ETALON, are heavy and compactand placed in relatively high orbits (6000 to 20,000km) to minimize the effects of drag and other non-conservative forces. Lower-altitude satellites of simi-lar design (e.g. GFZ-1, Stella, and Starlette) are oftenused to provide better sampling of higher-order staticgravity field components and the tidal part of the field.Furthermore, their simple spherical shapes allow anyresidual nonconservative force to be easily modelled.

The aforementioned three advantages — unam-biguous data type, insensitivity to atmospheric vari-ability, and simplified satellite force models — makeSLR the preferred approach for determining the

ITRF origin (geocenter) and scale (GM) as wellas the terrestrial gravity field. Geocenter motionsdetermined with LAGEOS observations have indi-cated annual and semi-annual signals of 2–4 mmamplitude (Watkins et al., 1995), and variations ofthe Earth’s center of mass due to ocean tides canbe resolved with sub-millimeter accuracy at sometidal frequencies. Estimates of GM have improvedfrom 50 parts per billion (ppb) (Lerch et al., 1978)to 2 ppb (Ries et al., 1992). SLR also providesthe highest absolute accuracy in determining stationheights relative to the geocenter. Height estimatesfrom the strongest SLR stations can be resolved to 2mm using observations spanning eight years (Dunnet al., 1993). Furthermore, SLR is the only commontracking system for the three altimetric satellitescurrently measuring sea and ice surface height (i.e.ERS-1, ERS-2, and TOPEX=POSEIDON) and isexpected to support future altimetry missions suchas the U.S. Navy’s GEOSAT Follow-on (GFO-1),TOPEX-2, and NASA’s Geoscience Laser AltimeterSystem (GLAS).

On the negative side, SLR is not an all-weathertechnique, and the current global distribution of sta-tions is far from optimum. Specifically, there are toofew high-quality permanent stations in the SouthernHemisphere and Central Asia. From a WEGENERperspective, SLR coverage in central and southernEurope is good but the surrounding regions (i.e.,Scandinavia, northern Africa, and Central Asia) arenot well monitored by SLR. Some of these issues arecurrently addressed by relocating SLR systems.

3.3.3. VLBIWhile in the past the focus in VLBI was exclu-

sively on horizontal movements, in the last few yearsthe attention has also been drawn to the vertical com-ponent (see for example, Mitrovica et al., 1994a,b;Sovers, 1994; MacMillan and Gipson, 1994; Van-Dam and Herring, 1994). However, the error of thiscomponent is still about three times that of the hori-zontal ones, and significant effort is being devoted toreducing this error.

Limitations of VLBI accuracy, in particular forthe vertical, are certainly due to tropospheric effects,but in some cases also ‘modifications’ of the atmo-spheric structure due to temperature variations haveto be taken into account.


During 1994, the European network for VLBIwas enlarged by including the stations of Crimea(Simeiz) and Ny-Alesund. The current accuracy forhorizontal velocities in Europe obtained with VLBIis about 0.3 mm=a for the core network sites (On-sala, Wettzell, Madrid, Medicina, Matera, Noto andEffelsberg; see also Fig. 2), 1 mm=a for Ny-Alesund,and 2 mm=a for Crimea. For the vertical component,these figures are about three times larger for moststations, while for Ny-Alesund and Crimea they areprobably four to five times larger at the moment.

3.3.4. GravityAbsolute gravimeters measure the acceleration

due to the gravity force of the Earth by studyingthe free fall of a mass in a vacuum. Space and timeare directly measured by means of a laser interfer-ometer and an high-precision time interval counter.The measurements are made with respect to metro-logical standards of space (iodine stabilized lasers)and time (rubidium frequency clocks). In this sensethe measurement is absolute. It is worth noting thatsince the ‘ruler’ is based on atomic properties, theabsolute technique is the only one suitable for long-term studies. Typically, some thousands (for free fallinstruments) or one hundred (for rise and fall in-struments) independent measurements are taken inone or two days measuring sessions. The standarddeviation is normally of the order of 1–3 µGal.The main sources of errors are changes in the vac-uum conditions, errors in the laser wavelengths andin the frequency standards, frequency-dependent er-rors in the electronic components, microseismicityand soil recoil effects. A recent intercomparison ofeleven absolute gravimeters showed that a realisticoverall error budget for a single gravimeter couldbe of the order of 3.2 µGal for a single session(Marson et al., 1995). This means that discrepanciesbetween gravimeters of the order of 6–9 µGal couldbe expected (corresponding to about 20–30 mm inheight). The detection of systematic errors or themeasurement of their changes can be achieved bycomparison with other absolute gravimeters and fre-quently repeated measurements at a reference stationmonitored by a superconducting gravimeter.

Absolute gravimetry is influenced by environmen-tal effects such as: (1) solid Earth and ocean tides;(2) polar motion; (3) attraction and loading of atmo-

spheric masses; (4) attraction and loading of oceanmasses; (5) changes in water table and soil water con-tent. Accurate mathematical models or direct mea-surements of the quantities which cannot be modelledhave to be employed in order to remove these perturb-ing effects. A suitable solution for the improvement ofthe models, and eventually for the direct measurementof the environmental gravity effects, can be providedby the use of superconducting gravimeters. Ideally,a measuring site should be equipped with a perma-nently installed superconducting gravimeter, periodi-cally colocated with an absolute one.

Absolute gravity applied to the determination ofcrustal deformations provides: (1) an independentcontrol for vertical displacements and velocities de-termined by space-geodetic techniques; (2) informa-tion on the forces causing crustal deformations; (3)boundary conditions for the verification of crustaldeformation models. In order to fulfil these tasks,secular gravity changes have to be measured witha precision of 0.5 µGal per year over decades. Atthe present level of precision (2 µGal), this could beachieved by semiannual measurements in 5 years.

3.4. Anticipated improvements of the measurementsand corrections

The repeatability of the space-geodetic measure-ments of surface positions or deformations is notonly limited by the accuracy achievable for onesample of observations but also by the level of un-derstanding of the geophysical signal at time scalescomparable to the measurement duration requiredto obtain one sample. In this section we will firstoutline improvements on the measurement side andthen, in the next section, turn to the problem of mod-elling short-period deformations that, if unmodelled,might reduce the repeatability or accuracy of themeasurements.

3.4.1. GPSThe major improvements in high-accuracy appli-

cations of GPS probably will be due to (1) a betterglobal coverage of the IGS network, (2) a better and(even) more reliable next generation of receivers andantennas (azimuth- and elevation-dependent phasecenter variations are still an important issue), and (3)several improvements in the modelling part.


It can be expected that a continued and intensiveobservation of the GPS satellites PRN5 and PRN6with SLR (both satellites are equipped with a LASERreflector array) will help to break up some of the exist-ing correlations (troposphere, station heights, antennaphase patterns). On the other hand it has to be men-tioned that presently the 11-year solar cycle is at aminimum of its activity. This may lead to an overes-timation of accuracy of precise positioning with theGPS because it is (1) easier for the receiver-internalsoftware to maintain the phaselock (which means thatfewer cycle slips occur) and (2) easier for the post-processing software to ‘repair’ the remaining cycleslips. However, the next generation of receivers maywell be apt to cope in a better way with the challengesrelated to a highly dynamic ionosphere.

3.4.2. SLRFuture improvements in the SLR data product

are expected to come from several sources includ-ing: (1) a better geographic distribution of stations;(2) a more uniform performance (i.e. few-millime-ter precision, subcentimeter accuracy, and high datayields) at these global sites; (3) the extended tem-poral coverage provided by fully automated stations;(4) improved handling of the atmospheric correctionthrough the use of more sophisticated gradient mod-els and=or the implementation of two-color rangingat selected sites; and (5) the deployment of a newgeneration of satellite targets designed to supportmillimeter accuracy ranging.

Efforts are already underway to transfer severalNorthern Hemisphere stations to French Polynesia,South Africa, Tunisia, and other Southern Hemi-sphere sites to provide a more balanced network insupport of the ITRF. In addition, the three major SLRnetworks (NASA, EUROLAS, and WPLTN) are pro-viding funds as well as technical consultation andsupport to improve the performance of SLR stationswhich are at important geographic locations but arenot operating at an acceptable state-of-the-art level.New, state-of-the-art, high-precision SLR=LLR sta-tions, such as the Matera Laser Ranging Observa-tory, being developed by the Italian Space Agency(Bianco and Varghese, 1994) are being colocatedat fundamental stations in close vicinity of VLBI,GPS and PRARE geodetic stations, in order to pro-vide a better connection between reference systems

defined by concurrent techniques. New automatedSLR stations, such as the SLR 2000 system be-ing developed by NASA (Degnan, 1994; McGarryet al., 1996) and the Keystone stations being de-veloped for use in Japanese geodetic monitoring,will greatly reduce the cost of SLR operations, pro-vide round-the-clock tracking coverage, and allowSLR operations in remote areas (e.g. island sites)which have little or no technical infrastructure. Two-color laser ranging systems, which provide a directmeasurement of the atmospheric refraction delayand promise 2 to 3 mm absolute range accura-cies (Abshire and Gardner, 1985; Degnan, 1993),are already being field-tested in the United States,Europe, and Australia. Multi-color laser ranging togeodetic satellites, in addition to improved rangeaccuracies, will contribute to a better understand-ing of the atmospheric refraction, which is currentlythe dominant error source in space-geodetic observa-tional techniques. In the interim, the proliferation ofglobal meteorological databases on the World WideWeb (WWW) permits the consideration and incor-poration of more sophisticated atmospheric modelswhich include the presence of horizontal gradientsat individual single-color SLR sites. Finally, a newgeneration of geodetic satellites are being designedwhich preserve the narrow pulse-width of the laser(Degnan, 1993) and allow true millimeter-accuracymeasurements. Experimental examples of these newtarget designs include the ADEOS Retroreflector inSpace (ADEOS=RIS) experiment launched by Japanin August 1996 and the WPLS geodetic satellitelaunched by the Western Pacific Laser Tracking Net-work (WPLTN) in early 1997. The majority of thecurrent SLR sites are colocated with GPS receiversof the IGS network. Combination solutions with thetwo data types have been reported as early as 1993(Pavlis and Braatz, 1993).

3.4.3. VLBIMain improvements for VLBI are to be expected

in the observing technique, where a better sky cov-erage and improved data quality due to a larger bandavailable on several stations (Mark IV) are antici-pated. Modelling of the troposphere is planned. Moststations of the VLBI network have now a permanentGPS receiver on site. Comparison of VLBI resultswith these continuous GPS observations and the use


of the GPS-derived troposphere for VLBI solutionsare likely to further improve the accuracy. Moreover,combined solutions of different techniques, includ-ing VLBI, shall be regularly computed in the future.

Concerning the European VLBI measurements,continuation of the present rate of six measurementsper year are likely to improve the results particularlyin the vertical. After almost five years of regularobservations only for 1994 and 1995 six observa-tions per year are available. Therefore, it is not yetclear whether apparent short-term variations are realmotions or are attributable to some source of noise.

3.4.4. GravityOver the last five decades, gravimetry has made

an impressive progress. The precision of both abso-lute and relative measurements has been improvedby almost three orders of magnitude to presently10�9. A new concept of measuring the temporalvariations of the gravity field by means of super-conducting gravimeters has been introduced as welland this has allowed a continuous acquisition ofthe gravity signal at a given site with the impres-sive precision of 10�10. A similar trend cannot beexpected for the next years. In terms of precision,gravimetry has reached a limit which will be diffi-cult to overcome. The same is not true with respectto accuracy. Further studies and developments arerequired in the design of more accurate calibrationtechniques for superconducting gravimeters and inthe reduction or even elimination of any drift. Abso-lute gravimeters still suffer from sources of potentialsystematic errors. Intercomparisons of instruments,frequent colocation of instruments of different typesare powerful means to detect at least biases due tosystematic errors. In this respect, the development ofinstruments based on different measuring conceptsor techniques should be encouraged.

In order to detect the weak gravity signals associ-ated with vertical crustal movement of the order ofmm=yr, it is necessary to measure g with a relativelyhigh sampling rate over decades. This task puts se-rious constraints on the accuracy of the instrumentsto be used, as well as on the measurement strategyand on the modelling of perturbing effects. The low-frequency components can only be observed by anabsolute gravimeter equipped with a laser stabilisedon an iodine absorption cell (the only one theoret-

ically free of drift), while the high-frequency partof the spectrum, which is very important for themodelling of environmental perturbing effects, canbe observed by superconducting gravimeters. Thedesign of a cost-effective measuring strategy whichcombines the two kinds of instruments will also be achallenging task for the future.

3.5. Modelling exogenic deformations and gravityvariations

In most geodynamical applications of geodeticmeasurements, exogenic deformations and associ-ated gravity variations are a significant source ofnoise. At periods of a day or less, Earth and oceantides are major sources of exogenic deformations. Atperiods of days, storm surges induce vertical motionof several centimeters, however, with associated de-formations restricted mainly to areas up to 100 kmfrom the coast. For periods larger than a day and upto a year, atmospheric loading is dominant and mayresult in large signals (several centimeters) at certainlocations and times, depending on climatological andmeteorological conditions. Additionally, long-periodtides contribute to the spectrum of surface defor-mation at the known tidal frequencies (fortnightly,monthly, semi-annual and annual).

3.5.1. Earth and ocean tidesAt any location on the Earth’s surface, the de-

formations due to forces at tidal frequencies can beconsidered as the sum of the direct effect of solidEarth tides and secondary effects due to ocean tidalloading, nutation, and tidal variations of the Earth’sorientation and rotation. The respective models thatshould be used in space-geodetic analyses are spec-ified in the IERS Conventions (McCarthy, 1996).In principle, all the analysis centers in all space-geodetic techniques should use the algorithms givenor recommended in this document. In practice, thesituation is not as ideal as it could or should beparticularly for the younger contributors like GPS,where the emphasis (especially at the beginning ofthe activities) had to be put on operational and not onmodelling aspects. However, the situation is improv-ing. The IERS Conventions are being observed moreand more closely, and exceptions are made knownby the analysis centers.


In general, the current algorithms are based onspherically symmetric models, which often are non-rotating. However, the theory for computing tidesand other exogenic deformations for laterally hetero-geneous Earth models is available (see for example,Wang, 1991; Plag et al., 1996), and as three-dimen-sional structural and compositional data (based, inparticular, on seismic tomography) for the Earth be-come more reliable, the application of three-dimen-sional (3-D) Earth models in space-geodetic tech-niques is near to being feasible.

3.5.2. Atmospheric and non-tidal ocean loadingMass movements within the atmosphere and the

ocean load and, consequently, deform the Earth. Inthe atmosphere, the pressure variations associatedwith mass movements have spatial scales rangingfrom a few hundred kilometers for tropical cyclonesto a few thousand kilometers for continental anticy-clones and typical time scales of a few days to a fewweeks. In addition, large-scale variations are foundat the seasonal time scale. The largest surface pres-sure variations are associated with weather patternsof 1000 to 2000 km, with increasing ranges towardshigher latitudes.

Atmospheric and ocean loading mainly deformsthe Earth in the vertical direction. Theoretical stud-ies of deformations and gravity variations due toatmospheric loading based on half-space or spher-ically symmetric models (for example, Rabbel andZschau, 1985; VanDam and Wahr, 1987; Merriam,1992) suggest that maximum vertical displacementsof more than 25 mm and 10 mm can be expected fordaily and seasonal time scales, respectively, whilehorizontal displacements are typically only one-thirdto one-tenth of the vertical ones.

In recent years, several attempts have been madeto identify the atmospheric loading signal in GPS,VLBI and SRL observations (see for example, Van-Dam et al., 1994; MacMillan and Gipson, 1994;Dunn et al., 1994). In most of these studies, themodel calculations are based on a Green’s functionapproach (Farrell, 1972), where the deformationsare computed from a convolution of the surfacemass density with the Green’s function for a spher-ically symmetric, non-rotating, elastic and isotropic(SNREI) Earth model. The atmospheric loading iscomputed using the ECMWF or NMC gridded global

meteorological data sets, which are based on theanalysis of surface meteorological data. From theseorganisations data are now available in near real-timewith temporal and spatial resolutions of down to 6 hand 2.5º, respectively.

According to VanDam et al. (1994), weekly av-eraged vertical station positions determined fromcontinuous GPS observations can reach precisionsof approximately 7 mm at the best sites. Thus, theatmospheric loading signal should be recorded inthese station height variations. However, using aSNREI model, VanDam et al. (1994) find for most ofthe investigated stations about 65% of the modelledpressure load to be present in the GPS vertical po-sitions. As sources for the remaining discrepanciesbetween model and observations, they suggest, “(1)anisotropic effects in the Earth’s loading response,(2) errors in the GPS estimates of tropospheric de-lay, (3) errors in the surface pressure data, or (4)annual signals in the time series of loading andstation heights”. Their first source should actuallynot be termed ‘anisotropic’; it can be separated intoeffects due to lateral heterogeneities particularly inthe lithosphere and upper mantle, and effects dueto the largely unknown response of the ocean toatmospheric forcing. VanDam et al. (1994) empha-size the high correlation between local air pressureand the measured height variations, and, therefore,the second error source may originate from a partialassimilation of the deformations in the troposphericdelay. Related to the third error source it shouldbe mentioned that the spatial resolution of the sur-face pressure data is low compared to the typicalwavelength of the Earth’s topography. Especially inmountainous areas, pressure determined from thegrid poorly resembles the actual air pressure at agiven topographic height, which represents the loadon the crust. Moreover, the accuracy of the availablegridded air pressure data is spatially highly vari-able depending on the availability of observations.In general, the accuracy is higher over land thanover oceans. Over land, it is highest over Europe andNorth America and lowest over most of Asia, theformer Soviet Union and the Indian subcontinent.

For VLBI, early attempts by VanDam and Wahr(1987) and Manabe et al. (1991) based on theGreen’s function approach failed to find a correlationbetween model displacements and observed vertical


movements at VLBI stations. However, more re-cently, VanDam and Herring (1994) concluded thatabout 60% of the computed loading effect is presentin the VLBI baseline length measurements. UsingVLBI station air pressure, MacMillan and Gipson(1994) found a pressure sensitivity between 0 and�0.6 mm=mbar with a general trend to smaller sen-sitivities for stations near the coast, which wouldbe expected from an inverted barometer response ofthe ocean to atmospheric loading. A simple loadingmodel based on local approximations to the regionalpressure field was found by Dunn et al. (1994) tomake a significant contribution to the height signa-ture observed at SLR stations.

3.6. Proposed activities

As outlined in the previous section, the potentialcorrelation of exogenic station-height variations andatmospheric effects particularly for GPS and VLBImight constitute a serious limitation for further im-provements of the accuracies of these techniques.Therefore, a special programme should be initiatedon studies of the atmospheric loading effect on theheight component of site coordinates obtained byGPS. In these studies, a comparison between GPS-and SLR-determined station heights is mandatory fora separation of geophysical and tropospheric effectsin GPS. The experiment currently being performedwithin the SELF II project (Zerbini, 1995) can beconsidered as a first example in this field. In thisexperiment, two GPS receivers are operating contin-uously at the Medicina reference site near Bologna,Italy, together with a cryogenic gravimeter periodi-cally controlled by three absolute gravimeters. Theexperiment will be conducted for one year. Therelevant high-accuracy atmospheric data are beingcollected on-site and from the regional meteorolog-ical service for a network of stations at distancesof up to 70–80 km around the reference site. Ad-ditionally, a GPS receiver is observing continuouslyfor one year at the Porto Corsini tide-gauge stationnear Rovenna, Italy, on the Adriatic Sea. The PortoCorsini site is also periodically checked with anabsolute gravimeter.

Modelling the displacements due to atmosphericloading for laterally heterogeneous Earth models re-quires the implementation of the theory outlined in

Plag et al. (1996), together with the setting up of aglobal 3-D model of the Earth’s lithosphere and man-tle. A 3-D model of the crust is now available with aspatial resolution of 5º (Mooney et al., 1998), whilea 3-D mantle model can be based on results of globalseismic tomography (e.g. Trampert and Woodhouse,1996). As discussed by Plag et al. (1996), based onglobal GPS-measured atmospheric loading the 3-DEarth model could be iteratively improved.

As a result of such a programme, recommen-dations for position determinations in regions withdifferent geomorphology or varying atmospheric andclimate conditions could be given. Moreover, theprogramme should be aiming at a service supplyingcontinuously the deformations due to atmospheric(and at a later stage, non-tidal ocean) loading, whichcould be directly utilized, for example, in the analy-ses of GPS observations.

4. Geodetic networks: the WEGENERperspective

Networks constitute the connection to the funda-mental reference systems for all geodetic investiga-tions and are therefore an essential tool for WE-GENER activities. They are established at three levelsof relevance for WEGENER. Efforts are made in eachcase to maintain the highest level of achievable accu-racy and precision, in accordance with the objectivesand requirements of the study being performed, thephysical and logistical feasibility and the availablefunding. The three categories may be defined as: (1)global networks; (2) regional networks; (3) local net-works. Individual measurements will always be madein relationship to one or another of these.

Global networks serve to define the InternationalTerrestrial Reference Frame (ITRF) and establish thebasis for connecting terrestrial networks (DORIS,GPS, SLR, VLBI etc.) to the Inertial Reference Frameand three-dimensional space. They are made up ofpermanent, continuously operated stations, many ofwhich are occupied by more than one (colocated) ob-serving system. Global networks build the frameworkinto which all subsidiary networks need to be fitted.The individual stations are required to report theirobservations daily in accordance with internationallyagreed standards to an international data collection


agency. They therefore require a good infrastructureand need to have access to especially good communi-cation facilities. The data are analysed continuouslyand the reported results are made available to the usercommunities at frequent intervals.

Regional networks comprise stations covering arestricted region of the Earth’s surface, for exam-ple, the eastern Mediterranean or South-East Asia,which are under investigation. The individual sta-tions may be permanently or repeatedly occupied,dependent on the availability of instrumentation,technical know-how, man-power and infrastructure.The connection to the global reference is establishedby a core of permanent stations within or on thefringes of the network making observations coinci-dent with the periodic repeat measurements. Underwell designed experimental conditions, this core con-trols the continuity of the observed phenomena fromone observation period to the next.

Local networks serve to make detailed periodic orcontinuous recordings addressing specific geophysi-cal questions particularly characteristic of the localarea. In general, the observing systems will only beinstalled temporarily, but every effort is made to re-late the observations to at least two reference stationsof a global or regional network.

WEGENER has contributed, and will continue tocontribute to each of the foregoing network typesand exchanges data in accordance with well estab-lished international principles. The facilities offeredby the Crustal Dynamics Data Information System(CDDIS), the IGS and the European SLR DataCentre (EDC) are used extensively both for dataarchiving and exchange purposes. Specifically, WE-GENER members contribute a number of stationsto the global networks (though this has never beenan objective of the group), and they have initiateda number of regional and local studies. By doingso, WEGENER continues to demonstrate its basicphilosophy of taking advantage of existing facilitiesand coordinating activities in consultation with othergroups, to address problems jointly that no individualorganisation could solve on its own.

4.1. Global networks

The results of the WEGENER investigations mustultimately be tied to a global reference frame in or-

der to maximize the scientific contributions of thedata. Each of the experimental techniques employedcontributes something unique to the investigation,and one must be able to freely use all data typesin the scientific analysis. This can only be accom-plished if adequate networks of colocated stationsare maintained and analyzed to provide the appropri-ate coordinate transformations for each space-geode-tic technique (i.e. SLR, VLBI, GPS, GLONASS,DORIS, and PRARE, see Figs. 1–5).

VLBI is a purely geometric technique. It providesa unique link to the celestial reference frame and iscapable of providing continuous and unambiguousmeasurements of the full suite of EOPs includingnutation, precession, polar motion, LOD, and UT1.The competing satellite techniques are insensitive tonutation and precession (parameters which are notparticularly relevant to the WEGENER science pro-gramme), whereas successful long-term monitoringof LOD and UT1 are ultimately limited by uncertain-ties in the satellite force models and require periodicupdates from VLBI. Among EOPs, polar motion issomewhat unique in that it can be effectively mon-itored by all of the principal space-geodetic tech-niques. Variations in polar motion, LOD, and UT1can signify large-scale movements of mass and=orexchanges of angular momentum between the solidEarth, ocean, and atmosphere systems.

As a geodetic tool, VLBI can provide relativestation positions and velocities with a high precisionunencumbered by a need for accurate satellite forcemodels, but the resulting polyhedrons are totally de-coupled from the geocenter by the very nature ofthe measurement. Satellite techniques, on the otherhand, are sensitive to the geocenter. However, be-cause VLBI is independent of the Earth’s gravityfield and its secular variations, it has the potentialto provide an independent constraint on the scaleof the global polyhedron, although VLBI scale sta-bility may ultimately be limited by the quality oftropospheric modelling.

Accurate monitoring of station heights and verti-cal movement rates is crucial to the WEGENER sci-entific objectives. On a local or even regional scale,geocenter motion (real or instrumentally induced)can often be treated as a common-mode error withlittle or no impact on observed relative vertical posi-tions and rates or the resulting scientific conclusions.


Fig. 1. Current global network of SLR stations.

However, without a properly constrained geocenteras provided by a global network of satellite trackingstations, measurement of absolute rates (e.g. of meansea-level rise) would not be possible, and many ob-served phenomena could not be properly interpretedfrom a global perspective.

Satellite techniques are also the primary source ofinformation on the intermediate- to long-wavelengthterms of the gravity field (and hence the geoid) andtheir secular variations. Combined with spacebornealtimetry, satellite measurements can further refinethe spatial resolution of the geoid over the openoceans and other large bodies of water, such asthe Mediterranean, and reveal important topographicfeatures on the ocean floor resulting from tectonicprocesses.

For reasons given in Section 3.3, SLR is bestsuited to define the origin (geocenter) and scale(GM) of the terrestrial reference frame, but theother satellite systems (GPS, GLONASS, DORIS,and PRARE) also sense the geocenter with vary-

ing degrees of accuracy. Over the past eight years,the University of Texas Center for Space Research(UTCSR) computations of the SLR geocenter, basedon successive 15-day LAGEOS orbits, have beenstable to better than 5 mm (RMS) in X and Yand better than 15 mm (RMS) in Z . Tidally co-herent motions have been reported by Pavlis (1994)and are in good agreement with theoretical modelsbased on TOPEX=Poseidon-tides (Schrama and Ray,1994). Long-term secular changes in the lower-orderzonal harmonics of the Earth’s gravity field, as ob-served through SLR tracking of LAGEOS and otherpassive geodetic satellites, have long been corre-lated with geophysical processes such as postglacialuplift. Presently, SLR is the only tracking systemcommon to the three ice and ocean altimetric mis-sions presently in orbit — ERS-1 (SLR only), ERS-2(SLR and PRARE) and TOPEX=POSEIDON (SLRand DORIS plus limited GPS). SLR therefore pro-vides a crucial link between the geodetic and alti-metric databases and permits all of the topographic


Fig. 2. Current global network of VLBI stations.

results to be expressed in a unified global referenceframe. The subcentimeter absolute range accuracy ofSLR combined with the relatively good global dis-tributions of the GPS, DORIS, and PRARE groundnetworks (which are often colocated at SLR andVLBI sites) can provide high-quality orbits (fewcentimeter absolute accuracy) for altimeter missions,clearly an important component of the WEGENERsea-level initiative. The optical wavelengths usedby SLR are also useful in calibrating out residualionosphere and water vapor effects in the altimetrymeasurement and in monitoring long-term drifts inthe instrumentation.

The relatively low-cost and high-precision relativepositioning of GPS receivers permits the spatial den-sification required to study adequately the complextectonic interactions between the Eurasian and sur-rounding plates, the details of uplift in Fennoscandia,and the integration of tide-gauge networks into thegeodetic=altimetric dataset for mean sea-level stud-ies. Through the auspices of the IGS, a global net-work of over 80 high-quality GPS receivers contin-

uously monitors the transmissions from the nominal24-satellite GPS constellation. Global IGS analysiscenters produce and distribute daily and high-accu-racy ephemerides for each of the satellites with alatency of about a week. Direct comparisons ofthese precise ephemerides with SLR measurementsto the retroreflector-equipped GPS-35 and 36 satel-lites typically indicate a radial agreement to betterthan 10 cm as do the orbits computed by the variousIGS Analysis Centers (Pavlis, 1995). The preciseephemerides and site positions obtained from theglobal analysis can then be used in the treatmentof regional and local GPS networks. For optimumresults, three or more global IGS sites are usedto constrain the regional analysis, while three ormore regional sites further constrain the local anal-ysis. Colocation of the IGS global GPS receiverswith the other space-geodetic techniques, combinedwith careful millimeter level local surveys betweensystems, allows the entire GPS network (global, re-gional, and local) to draw on the inherent, and oftenunique, strengths of the other space-geodetic tech-


Fig. 3. Current global network of permanent GPS stations supporting the IGS.

niques and aids in the identification, quantification,and correction of residual instrumentation or analysiserrors.

4.2. Regional and local networks

An important aspect of WEGENER, being linkedto one of the main scientific objectives of theproject, is the investigation of horizontal and ver-tical crustal motions in the boundary region betweenthe Eurasian, African and Arabian tectonic plates.This zone extends more than 5000 km in longitudefrom the Azores to the Caspian Sea, and 3000 km inlatitude from the Caspian Sea to the Gulf of Aden.For a careful study of the kinematics of the entire re-gion at the level of individual tectonic units, ideallycontinuous observations from an extremely densegeodetic network with a spacing of about 10–100km would be required. The obvious technique to useis GPS, since it is the only ‘affordable’ techniquecapable to provide such dense coverage at present.Examples of medium-density networks can be found

in Southern California in the US (Bock et al., 1993;Bock and Williams, 1997) and in Japan (Tsuji et al.,1995).

However, it is not feasible to cover the entirearea with GPS. First of all this would be very costlybecause of the large number of receivers required(for a cost estimate, see Prescott, 1996a). Secondly,much of the area is covered with large sea and oceanbasins where GPS points can only be established ona few islands. Finally, political problems and lack ofan adequate infrastructure also make it difficult toestablish a dense network in most of the remainingland areas.

Therefore, a distinction is made between regionalnetworks and local networks. Regional networksmay be considered as the first level of densificationof the global network. Their function is to provide re-gional connections to the global reference frame andto monitor the evolution of medium-scale tectonicprocesses. Local networks primarily serve to studysmaller areas in more detail, such as deformationzones near active faults.


Fig. 4. Current global network of DORIS stations.

A good example of the first category is the WE-GENER-MEDLAS network. It was established inthe mid-eighties with mobile SLR (which was theonly suitable technique at that time), to provide infor-mation on the overall tectonic motions in the easternand central Mediterranean regions. Since the earlynineties, it has also been observed with GPS andtogether, the observations of both techniques haveresulted in an accurate kinematic model for this area(see e.g. Noomen et al., 1996). However, this wasachieved on the basis of epoch-type campaigns. Withthis approach it is difficult to discriminate betweencontinuous deformations and instantaneous displace-ments due to, for example, earthquakes. Also, therepeated installation (and removal) of equipment,which may also change between campaigns, bearsthe risk of introducing errors.

Therefore, regional networks, consisting of fixed,permanently operating systems are favoured. Ideally,such networks consist of evenly spaced stations atmutual distances of typically 250 km, fitted withstandardized GPS equipment. To ensure a proper

connection with the global reference frame, someof these stations should also be part of the globalIGS network, in particular those colocated with othertechniques such as SLR, VLBI, DORIS and PRARE.

In Europe, the core of such a network basically al-ready exists. It consists of a relatively dense networkof about 30 IGS stations, which are mainly located inthe central part of Europe. Also, there is an initiativeto increase the number of permanent stations as acontribution to the densification of IGS. This activityis coordinated by the EUREF subcommission of theIAG. In addition, in a growing number of countries,extensive networks of continuously operating GPSreceivers have been or are currently being estab-lished for reference purposes (see Table 1). Still, thescope of these networks is too limited to serve all thescientific investigations of WEGENER. This is illus-trated in Fig. 6, which shows the current networkof permanent IGS stations in Europe and adjacentregions. As can be seen, the main problem is thatthere are almost no stations on the African continentand in Arabia. Therefore, one of the major tasks of


Fig. 5. Current global network of PRARE stations.

WEGENER should be to establish new permanentstations in these areas.

Other important scientific areas of investigationaddressed by WEGENER are the postglacial reboundin Fennoscandia and sea-level fluctuations associ-ated with climate change. To study these phenom-ena, dense regional networks of continuous stationsare also required. For sea-level monitoring, many ofthese stations should be colocated with tidal stations.

Table 1Some continuously operated GPS networks in Europe

Network Country=Region Number of stations

SATREF Norway 11SWEPOS Sweden 20FINNNET Finland 12BIFROST Scandinavia 46GREF Germany ¾50AGRS Netherlands 5

Belgium 4Italy (IGS) 5

Both topics are primarily concerned with the verticalcomponent and to a certain extent they are interre-lated: vertical land motions manifest themselves assea-level changes at tidal stations. As stated in theprevious section, these studies require an accurate andstable vertical reference frame. Although GPS doesnot (yet) provide accurate height information in anabsolute sense, it may well be used to study time vari-ations and relative changes between stations in thiscomponent. The connection with an absolute globalvertical reference frame can be provided by colocat-ing a few points of the regional networks with SLR.

In Scandinavia, a relatively dense network of about40 permanent GPS stations already exists (see e.g.Plag et al., 1998). They were primarily established fordifferential navigation purposes, but the data are alsoused to study postglacial rebound. There are plansto further expand the networks, in particular to colo-cate more points with tidal stations. A selection of thestations has been proposed as a contribution to theEuropean Vertical Network (EUVN), which is beingestablished to provide an accurate vertical European


Fig. 6. Current network of IGS stations in Europe and adjacent regions.

reference frame. Still, although the current networkcovers the main uplift area quite well, more stationsare needed in the surrounding regions (Baltic states,Denmark, UK) to extend the coverage to the margins.Furthermore, the connection with the global SLR net-work should be improved. The colocation sites in Eu-rope are presently mainly concentrated in the mid-lat-itudes. Therefore, it would be quite useful to repeatthe 1991 occupation of Tromsø in northern Norwaywith mobile SLR.

To study the relation between sea-level fluctua-tions in Europe and global sea-level change, again,an extensive network of permanent GPS stationscolocated with tidal stations can provide signifi-cant information. At present, only a few of thesestations are operational. There are several projectsunderway (Euro-GLOSS, EOSS) which aim at den-sifying the network of tidal stations colocated withGPS. However, it may still be necessary to establishmore points, in particular in areas of special interestsuch as the Baltic and Caspian Seas. Until this hasbeen realized, epoch-type campaigns such as thoseconducted in the frame of the SELF projects arenecessary to get started.

Local networks are usually established in rela-

tively small areas with the aim to study a veryspecific phenomenon. They often consist of some30 to 80 points with a spacing of 20 to 50 km.In most cases these networks are observed duringepoch-type campaigns which typically last betweenone and two weeks and which are repeated at oneto two year intervals. GPS is ideally suited for thispurpose and it has been successfully applied formany projects within the WEGENER area. In Ta-ble 2 an overview of the most important projects ispresented. Sometimes it may be required to establishpermanent networks for such local applications, inparticular when discontinuous motions are expected.The only example within WEGENER at this momentis the Kephalonian Fault network. Currently, epochcampaigns remain necessary to further densify thenetworks and to study special localized phenomena.They also provide an useful alternative for regionswhere it is difficult to establish continuous stations.However, for a full exploitation of the capabilities ofGPS, a combination of a number of CGPS sites withother episodically occupied GPS sites may turn outto be most appropriate. This approach is called themultimodal occupations strategy (MOST) by Beviset al. (1997). A major task for the next years will


Table 2GPS campaign-type networks in the WEGENER area; status of the table corresponds to mid-1996

Project Region Number of stations Number of occupations

GIG-91 Europe ¾25 1WEGENER Eastern Mediterr. ¾15 2EUREF Western Europe >150 ¾10SAGET Central Europe ¾10 2W. ALPS Western Alps ¾10 1BULGARIA Bulgaria 15 2CERGOP Central Europe 31C5 2CCRODYN CroatiaCC 22 2

E Turkey C Cauc. ¾80 ¾5NAFAS Turkey (Mudurnu) 31 5W-TURKEY Western Turkey 35 3MARMARA Marmara Sea ¾55 3AEGEAN Aegean 29 3CENTR. GREECE Central Greece 66 4W-HELENIC ARC SW Greece=SE Italy 30 4WHAT-A-CAT SW Greece=S Italy 20–43 3CTANGO Azores ¾15 4PEKA Caucasus=Black Sea ¾20 2CATS Pamir–Tienshan >75 >3BAIKAL Baikal Sea 13 1C1=2SELF Mediterranean Sea 28 2BSL Baltic Sea 35 (C12) 2UKGAUGE UK 9–16 6EUROGAUGE UK=France=Spain=Portugal 16 2NEREF-MAREO The Netherlands 18 3

be to define the conditions for successful applica-tions of MOST and to illuminate the strengths andlimitations of this strategy.

4.3. Tide gauges

Several authors have pointed out that the rela-tively large interannual and decadal variability in sealevel requires long records of 50 years and more if re-liable estimates of (local) trends in mean sea level areto be determined (e.g. Pirazzoli, 1986; Warrick et al.,1996; Douglas, 1997). Moreover, only tide gaugeswith high-quality data and with a well documentedhistory of the tide-gauge benchmarks (TGBM) thathas been used as the local datum can provide re-liable trends. Permanent GPS measurements at, ornear, tide gauges now make it possible to measurethe vertical crustal movement to an accuracy of š1mm=a with about five years of data. Therefore, thisis an additional reason for concentrating the GPSand absolute gravity work at well established tidegauges, where the relative mean sea-level trend is

already known to this or a better accuracy. On theglobal scale, there is a well known bias in the spatialdistribution of mean sea-level information, with themajority of long sea-level records being from a lim-ited number of areas on the Northern Hemisphere.The IOC GLOSS (Global sea-level observing sys-tem) network of tide gauges was initiated specifi-cally to address this issue and to ensure that a betterspatial distribution of high-quality sea-level data willbe available in the longer term.

In Europe, there is a similar north–south bias inlong sea-level records. For example, in the Mediter-ranean and Black Sea, all five tide gauges withmore than 50 years of well-controlled sea-level dataare exclusively located on the northern coasts: PortTuapse, Bakar, Trieste, Genova and Marseille. Onthe Atlantic coast of the Iberian Peninsula, thereare five appropriate tide gauges: Lagos, Cascais,Vigo, La Coruna and Santander (and Santa Cruz onTenerife). Further north, Brest in France and fivetide gauges around the UK are suitable. However,in the Netherlands there are nine tide gauges with


very long records and for the countries around theFennoscandian postglacial rebound and subsidencearea (Denmark, Germany, Finland, Sweden and Nor-way) there are a large number of long tide-gaugerecords available with many of them having 80 yearsor more of data. In areas of coastal subsidence, suchas the southern North Sea or the northern Adri-atic, where coastal infrastructure and eco-systemsare under threat, a higher density of modern geodeticmeasurements is clearly required. Even if the near-est suitably long tide-gauge record is a few hundredkilometers away, then this can be used (together withGPS and gravity measurements at the tide-gauge lo-cations) to determine the regional absolute sea-leveltrend and the local GPS and gravity measurementscan be used to find the spatial variability of thecoastal subsidence, and hence the local relative sea-level trend required for coastal impact work.

The IOC-coordinated EUROGLOSS proposed anetwork of 135 tide gauges around the Europeancoasts (Baker et al., 1996) in order to realize a moreuniform coverage of high-quality tide gauges. Inparticular, this network will provide information onthe spatial variability of mean sea-level trends.

4.4. What WEGENER needs from and may return tonetworks

The present-day capabilities of space technologyallow to perform geodetic observations with an ac-curacy which is significantly higher than typical de-formations of the Earth’s crust occurring on global toregional or even local scales. These observations pro-vide, on the one hand, fundamental input to geophysi-cal studies which deepen insight into, for example, theresponse of the solid Earth to different forces, its in-ternal structure and density distribution. On the otherhand, basic kinematic information is also supplied forthe definition of the terrestrial reference system whichis affected by geodynamical processes.

The realization of a reference system is donethrough a reference frame, that is, by means of anetwork (or polyhedron) of stations whose coordi-nates implicitly define a spatial coordinate system.The terrestrial reference system is produced throughan ensemble of Cartesian station coordinates givenat a certain arbitrary reference epoch and obtainedby combining different space-geodetic observations

such as SLR, VLBI and GPS. The deformations ofthe Earth’s crust are taken into account, at least to thefirst order, by providing together with the station co-ordinates the associated velocity field deduced fromlong-term geodetic measurements. The connectionbetween the fundamental polyhedron and the terres-trial system is given by the deformation of the stationpositions (Bock, 1996).

The area of interest to WEGENER encompassesa wide region from the Azores triple junction inthe west to the Caspian Sea in the east, and fromFennoscandia in the north as far as to the northernborder of the African plate in the south. In this area,networks of space-geodetic stations, both at regionaland local scales, are already available for the realiza-tion of some of the scientific objectives. Others arepresently being developed. These networks benefitfrom the existence of reference systems such as thoseproduced by the IERS and IGS and will rely upon therealization of a dense reference frame which is rea-sonably uniform in distribution and quality, accurate(few mm) and readily accessible to GPS users such asthat proposed by Blewitt et al. (1993). This referenceframe could provide the kinematic boundary condi-tions for the WEGENER large-scale network and forthe local networks dealing with specific problems ofa more localized nature (Blewitt et al., 1995).

WEGENER is also concerned with the establish-ment and maintenance of a highly accurate verticalreference system. To this end, WEGENER supportsthe continuation and the development of the activ-ities initiated by projects such as SELF I and II(Zerbini, 1995; Zerbini et al., 1996) and to be devel-oped within MEDGLOSS. One of the results shallbe the separation of the vertical crustal rates at thetide-gauge stations from sea-level variations which,at a regional scale, can differ significantly from theglobal trend. This is being accomplished by a com-bination of space and terrestrial techniques (absolutegravimetry). To properly accomplish this task, fidu-cial stations are needed, and for these WEGENERrelies upon the IGS network and the IERS referencesystems. In order to further contribute to the reliabledefinition of the vertical crustal rates WEGENERis also supporting long-term deployments (at leastone year) of mobile SLR systems at Tromsø in Nor-way, for measurements of postglacial rebound, andMedicina in Italy for sea-level variation studies.


In turn, WEGENER provides a fundamental con-tribution to the reference frames and systems through,basically, two mechanisms. First, the large-scaleSLR=VLBI=GPS network operating in the specificregion supplies continuously high-accuracy observa-tions which are being analyzed and interpreted withindifferent WEGENER projects. Thus, WEGENERcontributes to the definition of the reference frameby providing highly accurate solutions for the coordi-nates of the stations as well as a reliable velocity fieldnecessary for the realization of the reference system.Second, numerous GPS projects, at local scales, arebeing carried out within WEGENER and they con-tribute, with campaign-type of observations, to real-ize a more uniform and dense IGS polyhedron. Thiscould not possibly be realized otherwise, since theestablishment of a very dense network of fixed sta-tions is neither economically nor logistically feasi-ble.

As regards this latter point, WEGENER, with theaim to achieve a better representation of the kinemat-ics of the plates and platelets in the Mediterraneanregion and to contribute to a deeper understandingof the dynamic processes involved, is proposing todevelop the following activities: (1) to densify theIGS network along the African=Eurasian=Arabianplate boundary by means of campaign-type of ob-servations and solicit the installation of, at least,two to three permanent GPS stations on the north-ern African border; (2) to densify the existing GPSnetwork on the Azores and to deploy, for a longoccupation (1–2 years), one of the available mobileSLR systems; (3) to establish three to four fixed sta-tions across the Strait of Gibraltar; (4) to establish anetwork in the Arabian Peninsula across the Red Sea;(5) to densify the existing network in the Caucasusand Caspian Sea; (6) to densify with 2–3 fixed sta-tions, for a limited amount of time (1–2 years), and30–40 stations with campaign-type of observations,the network between southern Italy and Greece.

5. The African–European collision zone

5.1. The geodynamic setting

The Alpine–Mediterranean region marks thebroad transition zone between the African=Arabian

and Eurasian plates (Fig. 7). In a generalized schemethe recent major tectonic processes occur within thelarge-scale kinematic framework of active sea-floorspreading in the Atlantic Ocean, the Red Sea=Gulfof Aden and northwest Indian Ocean. The higherspreading rate in the South Atlantic (40 mm=a) ascompared to that in the North Atlantic (20 mm=a)leads to a gradual counterclockwise rotation of theAfrican plate resulting in a north-northwestward-directed push against Eurasia, which in turn leadsto a lithospheric shortening of 5 to 6 mm=a in-creasing to 40 mm=a in active subduction zones.With northwest–southeastward-oriented spreading inthe North Atlantic the zone of the Eurasian=Africanplate contact in the Mediterranean is expected to beto a large extent under compression. Argus et al.(1989) have deduced a kinematic model in which theclosure of the current plate motion circuit around theAzores triple junction has been enforced. This mo-tion belongs to a large-scale plate boundary whichpredicts a right-lateral slip of about 4 mm=a acrossthe Azores–Gibraltar fracture zone and a west-north-west convergence beneath the Gibraltar nappe in theGulf of Cadiz.

Despite the overall shortening between Africa andEurope, the Alpine–Mediterranean region encom-passes large sedimentary basins that have in the pastand even now experienced major extension sincethe late Cenozoic, e.g. the western Mediterranean,Tyrrhenian, Aegean, and Pannonian basins (Dewey,1988). The kinematic displacement field obtainedso far for the central Mediterranean demonstratesthe complex nature of the tectonics of the region(Albarello et al., 1995). In particular, this is illus-trated by the presence of extensional tectonics inthe Pantelleria Rift, the Sicily Channel, the Sira-cusa and Apulia escarpments which all lie along andclose to the North African compressional boundary.These regional non-stationary plate boundary pro-cesses contribute significantly (of the order of a fewcm=a) to the global-scale relative motion betweenAfrica and Europe.

At the northeastern end of the African plate, theArabian plate is separating from Africa (Nubia) ina NNE direction from the spreading center of theRed Sea, originally proposed by Drake and Girdler(1964), and from Somalia in a NE direction from thespreading center at the Gulf of Aden (Laughton et al.,


Fig. 7. Simplified sketch-map of main tectonic plates and sub-plates in the Mediterranean area. From Kahle et al. (1995a). Compare alsowith Fig. 8.

1970). Based on the spreading rates of these youngoceanic basins which are derived from magneticseafloor-spreading anomalies and from slip vectorsof earthquakes, it is estimated that the rate of north-ward motion of the Arabian plate is in the order of10 mm=a along the Dead Sea Rift (Levant), and 15mm=a in the southeastern part of the Arabian platearound Oman.

The ridges of the northern Caucasus and the as-sociated forelands in the area between the Black Seaand the Caspian Sea form the northern limit of thedeformation zone associated with the active bound-ary between the Eurasian and the Arabian plates.In this region the plate boundary and the associateddeformation zones are characterised by continent–continent collision and the resulting subduction ofthe Arabian plate beneath Eurasia. These features areevidenced by continued mountain building, accom-panied by relatively strong seismicity. No estimates

are available for the magnitudes of the horizontalmotions, but there is geodetic and geological evi-dence that in some areas uplift is occurring at a levellarger than 10 mm=a.

Since 1985, relative movements in the centraland eastern Mediterranean have been measured withSLR using permanent and mobile tracking equip-ment. SLR measurements from nineteen trackingsites were analyzed with data from the global track-ing network (Smith et al., 1994a,b). From this studyit could be seen that the kinematic field in the east-ern Mediterranean is characterized by the westwardmotion of Anatolia and the southwest motion of theAegean. Similar findings by Noomen et al. (1993)led Le Pichon et al. (1995) to describe the velocityfield of the Anatolian–Aegean region in terms of anaverage counterclockwise rigid rotation (2.4º=Ma)about a pole located north of the Nile delta, innorthern Egypt.


Fig. 8. Seismicity in the Mediterranean and adjacent areas. The lines delineate the main plate boundary structures between the Eurasian,African and Arabian plates. Also included are the large Anatolian fault in Turkey and the Caucasus–Caspian fault. Dots representearthquakes with ISC magnitudes >1, squares those with magnitudes >5.

From improved focal mechanisms constrainedby P and SH body-wave modelling, Taymaz et al.(1991a) showed that distributed right-lateral strike-slip is prevalent in the central and eastern Aegeanregion, on faults trending NE to ENE, and with slipvectors directed NE. This faulting ends in the west-ern Aegean against NW- and EW-trending normalfaults in central and northern Greece. The modelshave in common that a pronounced motion of theAegean microplate of ³40 mm=a is postulated, ori-ented towards the southwest, relative to a Europefixed reference. While it is generally agreed thatthe northern Aegean trough and the northern mar-gin of the Marmara Sea form the boundary in thenortheast, uncertainty exists as to where the bound-ary passes through northern Greece, how it connectswith the west Hellenic subduction zone and whatthe relative displacements rates are. How this zoneconnects the active dextral Kephalonia fault (Kahleet al., 1995a,b) with the active transtensional dex-tral North-Aegean fault zone and Sea of Marmara(Straub and Kahle, 1995; Straub, 1996) is as yetunknown.

A summary of the stress distribution in Turkeyand Greece is illustrated by well-determined earth-

quake focal mechanisms (Oral et al., 1993). In east-ern and northern Turkey the seismic activity is pri-marily associated with the east and north Anato-lian strike-slip fault zones (Fig. 8). The east Ana-tolian and Dead Sea transform fault zones exhibita predominantly left-lateral motion (Taymaz et al.,1991b) while the north Anatolian fault zone mo-tion is go

Documents

Scientiﬁc objectives of current and future WEGENER activities et al wegener 1998.pdfELSEVIER Tectonophysics 294 (1998) 177–223 Scientiﬁc objectives of current and future WEGENER