G E O M A T I C A

CRUSTAL MOTION AND DEFORMATIONMONITORING OF THE CANADIAN LANDMASS

Joseph A. Henton, Michael R. Craymer and Rémi FerlandGeodetic Survey Division Natural Resources Canada, Ottawa, Ontario

Herb Dragert and Stéphane MazzottiGeological Survey of Canada, Natural Resources Canada, Sidney, British Columbia

Donald L. ForbesGeological Survey of Canada, Natural Resources Canada, Dartmouth, Nova Scotia

The science of geodesy and the corresponding reference systems it develops have increasingly beenapplied to measuring motions and slow deformations of the Earth’s crust driven by plate tectonics.Improvements to geodetic methodologies have therefore enabled better understanding of the Earth’s systems,including improved modelling and forecasting of changes that may affect society. These geophysical process-es also systematically affect the reference frames used as standards for geodetic surveys. Reference framestherefore must not only define the system of coordinate axes (including orientation, origin, and scale), butalso characterize the time-evolution of spatial coordinates on the Earth’s surface. When evaluating the effecton reference standards within a given area, it is also important to realize that geodynamic processes operateon various spatial scales. In this paper we summarize some of NRCan’s efforts to monitor contemporarycrustal dynamics across Canada. Progressing from continental to smaller regional scales, we outline therationale, techniques, and results. The observational data and interpretations presented are fundamentallydependent on the Canadian Spatial Reference System yet in turn also contribute to the incremental improve-ment of its definition and maintenance.

1. IntroductionToday the Earth Sciences Sector (ESS) continues torespond to the Order in Council that created theGeodetic Survey of Canada in 1909 tasking the thenMinister of Interior “to determine… the positions ofpoints throughout the country… which may form thebasis of surveys for all purposes, topographical, engi-neering or cadastral, and thereby assist in the surveywork carried on by other departments of the DominionGovernment, by Provincial Governments, and bymunicipalities, private persons or corporations.”

[passage from Canadian Spatial ReferenceSystem (CSRS) service logic model].

Geodynamic monitoring is demanding interms of observational accuracy and stability over awide range of spatial- and temporal-scales. Anappropriate reference frame is a requisite tool thatenables better understanding of geodynamic systemsthat in turn provides better constraints to predictedimpacts on society. However, the effects of geody-namic and geophysical crustal deformationprocesses can systematically bias the referenceframes used for geodetic surveys. Kinematic

GEOMATICA Vol. 60, No. 2, 2006, pp. 173 to 191

La science de la géodésie et les systèmes de référence correspondants ont été de plus en plus utilisés pourmesurer les mouvements et les lentes déformations de la croûte terrestre causés par les plaques tectoniques.L’amélioration des méthodes géodésiques nous a donc permis de mieux comprendre les systèmes de la Terreen nous permettant, entres autres, de mieux modéliser et prévoir les changements qui risquent de nous touch-er. Ces processus géophysiques modifient aussi systématiquement les cadres de référence qui servent de normesaux levés géodésiques. Les cadres de référence doivent alors non seulement définir le système des axes descoordonnées (incluant l’orientation, l’origine et l’échelle), mais doivent aussi définir l’évolution temporelledes coordonnées spatiales sur la surface terrestre. Lorsqu’on évalue leur effet sur les normes de référence dansune zone donnée, il est aussi important de réaliser que les processus géodynamiques se produisent à plusieurséchelles spatiales. Dans cet article, nous résumerons certains des efforts de surveillance de la dynamiquecontemporaine de la croûte terrestre canadienne par Ressources naturelles Canada. De l’échelle continentaleà l’échelle régionale, nous présenterons un survol du besoin, des techniques et des résultats. Les données et lesinterprétations observationnelles présentées dépendent fondamentalement du Système canadien de référencespatiale tout en contribuant à l’amélioration de sa définition et à sa maintenance.

G E O M A T I C A

processes affecting the Earth’s surface can systemat-ically affect geodetic measurements and ultimatelylead to inconsistencies between observations at dif-ferent epochs. As measurement and processingaccuracies and requirements have increased, theCanadian Spatial Reference System (CSRS) is nowin an era in which the time-variability of its outputsmust be evaluated.

For Canada the largest observed deformationrates are at the level of 1-2 cm/yr. Millimetre-per-yearresolution, and better, is required to investigatemany processes of the dynamic Earth and conse-quently to better quantify their potential impacts onsociety (e.g. natural hazard and climate changeeffects). Measuring these rates requires very high-precision geodetic techniques coupled with suitablylong observation windows. The frequency of obser-vations is largely dependent on the magnitude ofthe deformation process and the precision of themeasurement technique. Survey design and dataprocessing must also consider any quasi-periodicsignals (e.g. the annual signals resolved by GPS) thatcould bias estimates of geodynamic motions.

Within ESS of Natural Resources Canada(NRCan), regions of geophysical interest (particu-larly hazards) have warranted densified geodeticinfrastructure. For ESS, long-term topics of interestfor geodetic investigations include studies of post-glacial isostatic adjustment across the Canadianlandmass, the Saint Lawrence seismic zone in east-ern Canada, the active plate boundary above theCascadia subduction zone along Canada’s westcoast, and the active fault margin of the QueenCharlotte Islands. Additionally, investigations ofthe vulnerability of Canada’s coasts to climatechange processes have increasingly relied on precisegeodetic observations.

Targeted regional studies within ESS augmentthe CSRS and thus contribute to aiding and strength-ening international services to the geophysical andgeomatics communities. The international- andnational-scale geodynamic monitoring efforts, inturn, support regional deformation studies. Within aspecific region there are often different kinematicprocesses operating on various spatial scales. At thelargest scale, rigid tectonic plate motions may needto be considered. Then, measuring long-wavelengthsignals at a broader (e.g. national) scale provides theregional deformation “background signal” which isuseful when assessing more localized deformationsignals during regionally-targeted kinematic studies.

This paper outlines a number of efforts withinESS that contribute to monitoring crustal motion anddeformation at differing scales across the Canadianlandmass. It is not a complete review, but ratherintends to provide a well-rounded cross-section of

Sector studies. Progressing from larger to smallerspatial scales, the rationale, methodology, and results(for established studies) are summarized for thegeodynamic investigations. Although this articleemphasizes the work of ESS, a considerableamount of data, analyses, and interpretations havebeen dependent on the work and productive collab-orations of numerous other groups and agencies.Further information and details may be found in thepublications referenced in each section.

2. Global-Scale PlateKinematics and the ITRF

With the wide-spread use of space-based tech-niques for positioning in the scientific and engineer-ing communities, a unified and consistent globalspatial reference system has long been necessary.The International Terrestrial Reference System(ITRS) was proposed and implemented over twodecades ago. Its physical realization, theInternational Terrestrial Reference Frame (ITRF),provides a high accuracy and easily accessibleglobal coordinate system for the study of Earthdynamics at all scales [Altamimi et al. 2001].Several updates of the ITRF have been made avail-able with each one improving on its predecessor.Although originally developed for and by the sci-entific community, this system eventually becamethe most widely used reference system for posi-tioning. For example, some countries have adoptedversions of the ITRF as the national standard andeven WGS84, the native reference system for GPS,is now based on ITRF. Additionally these framesprovide velocity information and use kinematicconstraints in their realizations.

Any ITRF is composed of a set of station coor-dinates at a given epoch and their velocities in threedimensions with appropriate covariance informa-tion. The frame is currently defined via four tech-niques: (1) Very Long Baseline Interferometry(VLBI), (2) Satellite Laser Ranging (SLR), (3)Global Positioning System (GPS), and (4)Détermination d’Orbite et RadiopositionementIntégré par Satellite (DORIS) [Boucher et al. 2004].A combination of the techniques is necessarybecause none alone is sensitive to all the parame-ters necessary to define a reference frame. Thecombination takes advantage of the strength of eachtechnique; e.g. the SLR contribution is essential forthe determination of the origin while VLBI and SLR

174

…a unifiedand consis-tent globalspatial refer-ence systemhas longbeennecessary.

G E O M A T I C A

contribute to determine the scale. The combinationof the networks from each technique also requiresthat they be connected at as many collocated sites aspossible. Repeated local surveys are required toprecisely measure the three-dimensional offsetsbetween the markers for each collocated techniqueand detect any potential change. All of the tech-niques contribute to the velocity field, which is usedto solve for the time evolution and the no-net rota-tion condition of the system. For the general usercommunity, it is the GPS products (coordinates andsatellite orbits) that enable easy access to the ITRF.

Earth scientists rely on ITRF products formany investigations related to large-scale massredistribution, including characterization of theEarth’s interior and climate change studies. In par-ticular, space-geodetic techniques have revolution-ized the study of plate tectonics. At one time, platemotion models were primarily constrained by sea-floor magnetic anomalies and other geologicalinformation with rates averaged over millions ofyears. With GPS it is possible to directly measurecontemporary tectonic plate movement on a timescale of years. The motion of any rigid body on thesurface of a sphere can then be represented by arate of rotation about an axis (i.e. Euler pole) andexpressed as a rotation vector. The North AmericanPlate, for example, undergoes a counter-clockwiserigid body rotation around an Euler pole locatednear equatorial, northwestern South America (referto Figure 1). The associated tectonic plate motionsfor Canadian sites are on the order of 2 cm/yr.

When accurately defining the rotation vectorof a plate, it is important to ignore stations biasedby local site effects and those in known deformingzones unless reliable models of such deformationsare available. Ideally, only the sites that best repre-sent the rigid parts of the tectonic plate should beused. High-precision GPS has allowed scientists todiscriminate subtle differences in velocities withinwhat had been considered individual tectonic units,and in the process identify new plates or sub-platesand quantify their individual rotation vectors. Theimproved resolution of plate rotation vectors is alsovery useful for investigations that utilize kinematicinformation (e.g. fault and other crustal deforma-tion studies). However, the comparison of preciseGPS-determined results is often complicated byinvestigators using different techniques to producea plate-fixed reference frame in which to expresstheir results. For North America the SNARF initia-tive (discussed in the next section) will provide aconsistent reference system in which scientific andgeomatics results (e.g. positions in tectonicallyactive areas) can be produced and inter-compared.

3. Continental & NationalScale Monitoring Efforts

3.1 NAREF Densification of ITRF inNorth America

Since the beginning of 2000, ESS has beenplaying a leading role in the North AmericanReference Frame (NAREF) Working Group of theIAG Sub-Commission 1.3c (Regional ReferenceFrames for North America) [Craymer andPiraszewski 2001]. One of the primary objectivesof this working group is the densification of theITRF in North America using continuously operat-ing GPS (CGPS) stations. Following the distributedprocessing approach advocated by the InternationalGNSS Service (IGS) [Blewitt 1997], the NAREFworking group members have been independentlycomputing weekly regional coordinate solutions forCGPS sites throughout North America.

175

Figure 1: North American plate rotation pole and predicted veloc-ities. The blue vectors represent the motion for those sites predict-ed from North American rotation about its pole. The arrow lengthsare proportional to the rates at their origin point and are given forCanadian continuous GPS sites present in the IGS cumulativesolution (GPS week 1345). The North American rotation pole isfrom ITRF2000 [Altamimi et al. 2002].

G E O M A T I C A

NRCan has been producing three such weeklyregional solutions for Canada. Two are being gener-ated by NRCan’s Geodetic Survey Division (GSD)on a regular basis for redundancy and quality con-trol. One of these solutions incorporates additionalstations just outside Canadian borders for furtherredundancy and continuity with American solutions.In addition, NRCan’s Geological Survey ofCanada—Pacific Division has also been contribut-ing weekly solutions for their Western CanadaDeformation Array (see later section) on the westcoast. Other solutions include a preliminary versionof the Plate Boundary Observatory from the ScrippsInstitution of Oceanography, and the entire USCORS (Continuously Operating Reference Stations)network from the US National Geodetic Survey thatincludes nearly 700 stations. More recently, effortsare underway to include stations from provincialGPS networks, including those of British Columbia,Quebec and, soon, New Brunswick.

Following internationally accepted densifica-tion methodologies [Ferland et al. 2003], these dif-ferent regional solutions are being combined into asingle NAREF weekly solution. NAREF combina-tion solutions beginning the first week of 2001have been submitted to the IGS data archives andare usually available with about a four-week latency.There are presently over 800 stations in the com-bined NAREF network (refer to Figure 2). Manystations are included in more than one solution toprovide redundancy checks and to allow for thecorrect weighting of the different solutions relativeto each other and to the global solutions. The goal isto have all stations included in at least two differentsolutions. All regional solutions agree with eachother and with the global ITRF and IGS solutions atthe level of a few millimetres.

The weekly NAREF solutions have recentlybeen combined into a single so-called cumulative(multi-epoch) solution to provide estimates of bothstation coordinates and their velocities with respectto a consistent reference frame throughout NorthAmerica. Velocity estimates with accuracies of lessthan 1 mm/yr are expected to be achievable in thenear future with the accumulation of several years ofdata. This relatively long time series is required inorder to estimate and remove systematic annual andsemi-annual signals due to seasonal effects and dis-continuities caused by changes in equipment or soft-ware (refer to Figure 3). Nevertheless, even today,these solutions can be used to support geodynamicsstudies of both large and small scale crustal motions,including the direct measurement of the motion ofthe North American tectonic plate and intra-platedeformations such as post-glacial rebound as dis-cussed later in this paper. A preliminary version of

176

Figure 2: NAREF network of continuous operating GPS receivers. The redcircles represent stations in the IGS global network while the blue circles rep-resent densification stations.

Figure 3: Time series for station INVK in Inuvik, NWT illustrating relativelylarge seasonal fluctuations primarily due the effects of snow on the antennaphase centre. Note also the step (discontinuity) in the time series just after2003.5 due to the addition of an antenna dome.

G E O M A T I C A

this NAREF velocity solution has also been used tohelp define a North American plate-fixed referenceframe for studies of intra-plate crustal motions (seenext section).

3.2 A Stable North AmericanReference Frame (SNARF)

As discussed earlier, regional reference framesfixed to the stable part of a tectonic plate are oftenrequired to facilitate geophysical interpretation andinter-comparison of geodetic solutions of crustalmotions. In 2003, the Stable North AmericanReference Frame (SNARF) Working Group wasestablished under the auspices of UNAVCO, Inc.and IAG Regional Sub-Commission 1.3c for NorthAmerica especially to address needs for the US-ledEarthScope project. The goal is to define a regionalreference frame that is consistent and stable at thesub-mm/yr level throughout North America.

The SNARF Working Group identified andaddressed several issues that must be dealt with toproperly define such regional frames, including (1)the selection of “frame sites” based on geologic andengineering criteria for stability, (2) the selection ofa subset of “datum sites” which represent the stablepart of the plate and will be used to define a no-netrotation condition, (3) the modelling of any signifi-cant intra-plate motions using a relatively dense GPSvelocity field, and (4) the generation and distributionof reference and deformation products for generaluse [Blewitt et al. 2005]. The SNARF vertical datumis consistent with ITRF2000 in that the centre ofmass of the whole Earth system is taken to be theorigin while the horizontal datum differs by a rota-tion rate that brings the rotation of the stable part ofNorth America to rest.

The first release of SNARF provides a rotationrate vector that transforms ITRF2000 velocitiesinto the SNARF frame, and an initial referenceframe is defined via a list of selected sites, epochcoordinates and velocities in a Cartesian system.This rotation effectively defines the SNARF refer-ence frame in relation to the ITRF. It was comput-ed using only stable sites from a combination ofvelocity solutions from GSD for the NAREF net-work and the Canadian Base Network (see nextsection), and a velocity solution from PurdueUniversity for a selection of US CORS stations.Horizontal intra-plate velocities in the SNARFframe exhibit a pattern that is more consistent withexpected deformations from post-glacial reboundthan velocities obtained from other estimates; e.g.,ITRF2000 (refer to Figures 4a & 4b). In addition,these velocity solutions were also used to determine

a semi-empirical model of post-glacial reboundbased upon a novel assimilation technique thatcombines GPS velocities with a geophysical model[Blewitt et al. 2005] (see Figure 5). Over the nextfew years SNARF will be incrementally improvedthrough further research and as more accuratevelocity solutions become available.

177

Figure 4a: Residual horizontal intra-plate velocities of CBN network usingNorth American plate motion estimates (rotations) from ITRF2000. The platemotion estimate absorbs much of the horizontal motion from post-glacialrebound that is expected to radiate outward from the areas of maximumuplift. Little outward horizontal motion is seen in the intra-plate velocities.

Figure 4b: Residual horizontal intra-plate velocities of CBN network usingNorth American plate motion estimates (rotations) from SNARF 1.0. In thisreference frame defined by the SNARF plate motion estimate, the outwardpattern of intra-plate velocities is more clearly visible. The SNARF platemotion estimate is affected less by post-glacial rebound.

G E O M A T I C A

3.3 Geodynamic Observations fromthe CBN

Initiated in 1994, the Canadian Base Network(CBN) is a national network of high-stability pillarmonuments with forced-centering mounts for GlobalPositioning System (GPS) receiver antennas. Theinitial role of the CBN was to complement theActive Control System (ACS) of the CSRS by pro-viding easily accessible 3D reference coordinatesites, with a reasonable distribution across Canada.In order to maintain the accuracy of the CSRS ref-erence frame realization, CBN sites have been re-occupied to confirm initial positions and to detectcrustal movements. By combining nearly 10 years ofrepeated multi-epoch (episodic) GPS measurements,GSD has begun to estimate velocities at the CBNsites in order to provide an increased spatial sam-pling of crustal deformation throughout Canada. Todetermine individual station velocities, regionalCBN solutions for each measurement epoch aresystematically combined into a single Canada-wide,multi-epoch cumulative solution (i.e. similar to theprevious discussions for NAREF and SNARF). Inorder to generate time series of consistent, high-accuracy coordinates for velocity estimation, it isnecessary to ensure consistency of the integrationinto the reference frame. This is accomplished byaligning each of the individual CBN solutions to asubset of stations from a recent cumulative solutionfor the IGS global network in ITRF. Fortunately,there are many IGS stations in Canada and mostwere included in each regional CBN solution tostrengthen the connection to reference frame and

ensure consistency between epochs. Consistent andrealistic weighting of the individual CBN solutionsis improved through the estimation of separate vari-ance components relative to the IGS global solu-tion. After the individual CBN solutions are alignedand weighted, they are combined together in asimultaneous cumulative solution to producevelocities at each site [Henton et al. 2004].

On the national scale, glacial isostatic adjust-ment is the most significant geodynamic processdriving vertical deformation [e.g., Tushingham andPeltier 1991; Peltier 1994]. Preliminary resultsfrom the combination of CBN regional solutions inCanada exhibit a spatially coherent pattern of upliftconsistent with the expected post-glacial rebound(PGR) signal (refer to Figure 6). Regions of high-est uplift rates are generally consistent with areas ofgreatest ice accumulation during the last period ofcontinental glaciation [e.g., Dyke 2004; Peltier1994]. Horizontal velocities associated with PGRare also spatially coherent (typically directed radi-ally outward from regions of highest uplift) buthave smaller rates. These modest horizontal veloc-ities can provide an important additional constraintto PGR models. However, the horizontal vectorsmay be significantly biased by differing referencesystem rotation-rate vectors (e.g., Figures 4a & 4b).The definition of SNARF should minimize thisissue. Eventually the deformation maps, coupledwith PGR model predictions, may allow NRCan tointerpolate or estimate coordinates at differentepochs for regions that display spatial coherence inthe velocity field.

3.4 Using Absolute Gravity toMonitor Uplift at CBN Sites

The integration of geodetic techniques is desir-able when monitoring geodynamic processes (andsimplifies any connections between correspondingreference standards). Absolute gravimetry (AG),which is independent of GPS, has demonstrated thatit plays a complimentary role to GPS especiallywhile measuring vertical crustal motions [e.g.,Lambert et al. 2001]. In addition, repeated GPS andAG observations at common sites provide insightinto the geophysical processes that drive theobserved deformation since AG is sensitive tointernal mass changes and not just deformationalone. Therefore, issues such as mass redistributionor changes in density contrasts within the Earth maybe addressed by monitoring positional changes (i.e.,primarily height changes) and integrating theseobservations with gravitational variations. However,the observed rates of gravity change resulting from

178

Figure 5: Post-glacial rebound rates from the SNARF model (courtesy of JimDavis, Harvard-Smithsonian Center for Astrophysics [after Blewitt et al. 2005]).This empirical model of post-glacial rebound employs a novel technique thatcombines observed GPS velocities with a geophysical model.

G E O M A T I C A

post-glacial rebound in Canada are no greater thanapproximately -2 µGal/yr (one µGal is approxi-mately one part-per-billion of the Earth’s surfacegravity). Observing such a small signal requires asuitable combination of high-precision and obser-vation length. The principle of most modernabsolute gravimeters is to observe (using laserinterferometry) the acceleration of a free-fallingtest mass in a vacuum. This technique thereforerequires very precise measurements of length andtime over repeated “drops.” Despite the simple fun-damental principles, high-accuracy absolutegravimeters involve a great deal of instrumental andelectronic sophistication. Properly operated and aftercareful processing, AG can provide the value of theEarth’s gravity at a point with an accuracy of onepart-per-billion, and the instruments can be main-tained nearly drift-free. Also the AG instruments ofthe CSRS are tied to international standards throughcomparisons at the Bureau International des Poids etMesures (BIPM). Continued improvements inabsolute gravimetry have made these instrumentsmore compact, robust, and efficient.

In order to monitor the temporal variations ingravity resulting from regional glacial isostaticadjustment, a set of absolute gravity measurementsites has been established in northern Quebec, col-located near pillars of the CBN. The NouveauQuebec-Labrador region of eastern Canada was thesite of one of the major ice domes of the LaurentideIce Sheet [e.g., Dyke 2004; Peltier 1994] and is cur-rently experiencing post-glacial rebound. For east-ern Canada the highest uplift rates are in the vicin-ity of James Bay through to southwestern Labrador.Rates decrease to the south and become negativetowards the coastal Atlantic margins. For post gla-cial rebound, surface gravity measurements sensethe effect of increasing distance from the centre ofmass of the earth (i.e. gravity decreases) coupledwith the redistribution of mass due to viscous flowat great depth (increases gravity). The resultingregional gravity decrease associated with uplift dueto post-glacial uplift is approximately -0.15µGal/mm [Lambert et al. 2001]. Preliminaryabsolute-gravity trends for this region showdecreasing gravity with time. Values range from

179

Figure 6: Contour map of observed CBN vertical rates. Preliminary results from the combination of CBN regional solutions in Canadaexhibit a spatially coherent pattern of uplift consistent with the expected post-glacial rebound (PGR) signal. Black dots represent thelocations of CBN sites.

G E O M A T I C A

about -2 µGal/yr (i.e. ~13 mm/yr for regional PGR)near Kuujjuarapik and Radisson and decrease south-ward to approximately -0.5 µGal/yr (~3 mm/yr)near Val d’Or. These preliminary results exhibitgeneral agreement between the rates for GPS upliftvelocities and gravity trends. Additionally, theobserved (GPS & AG) rates are generally consistentwith predictions of vertical crustal motion frompost-glacial rebound models. To better contribute tothe definition of the vertical component of a highlyaccurate, multi-purpose, active and integratedCSRS, efforts are now underway to create a sparsearray of absolute gravity observation sites collocatedwith geometric reference (i.e. VLBI, continuous andepisodic GPS) stations across Canada.

4. Monitoring Relative Sea Levelfor Coastal Impact Studies

4.1 Introduction

As potential impacts of climate change andsea-level rise on coastlines depend in large measureon the localized rates of relative sea-level heightchange, it is important to quantify vertical defor-mation velocities in vulnerable coastal regions. Themagnitude of the rate of relative sea-level changecan be used to better understand future impacts (e.g.flooding, storm-surge levels). Sea level heights havetraditionally been determined by tide gauges whichmeasure the relative sea level height with respect toland. The long-term changes in relative sea levelobserved at tide gauges reflect both the verticalcomponent of regional crustal kinematics and thechange in the regional (and/or global) sea level sur-face. Through this linkage with relative sea levelobservations, understanding the kinematics of post-glacial rebound in vulnerable coastal areas is animportant adjunct to relative sea level studies.

4.2 Eastern Canada

While most of the Canadian landmass is current-ly experiencing uplift associated with post-glacialisostatic adjustment, the Maritimes are experiencingsubsidence. This is primarily due to collapse of theperipheral bulge following the deglaciation of theLaurentide Ice Sheet (refer to Figure 7), in additionto the effects of rising post-glacial sea level loadingthe continental shelf. The subsidence rates are notparticularly large, typically on the order of -1 to -2mm/yr. However they are of the opposite sign tosea level height change (on the order of +2 mm/yr)and consequently relative sea-level rise (withrespect to land) is regionally more rapid. Sea-levelrise can produce significant impacts in the coastalzone, particularly for low-lying parts of theMaritime Provinces [e.g., Forbes et al. 2004b].These include storm impacts on the coast (waves,surges, and flooding), sediment movement and ero-sion hazards, impacts on ecological systems (e.g.coastal wetlands and fisheries), and damage to pri-vate or commercial property and public infrastruc-ture [e.g., O’Reilly et al. 2005]. Work is underwaywithin ESS to better quantify these hazards throughthe installation of additional geodetic infrastructure(e.g. new continuous GPS sites at select regionaltide gauges).

180

Figure 7: Generalized process of post-glacial rebound. The mantle acts like anextremely viscous fluid; it continues to flow back to regions where heavy icesheets had forced out portions of the mantle 18,000 years ago. These regions,like northern Quebec, experience “post-glacial rebound.” Around the edges ofthe ice sheet the crust flexes upward creating a peripheral fore-bulge. Afterdeglaciation, the fore-bulge slowly collapses resulting in subsidence forregions such as the Maritimes.

G E O M A T I C A

4.3 Great Lakes

The Great Lakes is an important economic cor-ridor where lower water levels can have significantimpacts such as the need for expensive dredging ofports and harbours, and reduction in hydro-electricpower generation. Additionally, elevated erosionrates have been evident in regions of the Great Lakesduring previous higher water levels. These potentialimpacts ultimately depend in large measure on therate of change of relative water-level heights withrespect to land. Again, although the rate of crustaltilting driven by PGR is not particularly large, itsteadily accumulates over time and can be clearlyobserved in water-level gauge records [e.g.,Mainville and Craymer 2005]. This regional verticaldeformation must be considered when evaluatinglong-term impacts on datums, water management,infrastructure, and basin ecology.

The regions north of the Great Lakes are risingfaster than the regions to the south. In fact, much ofthe southern areas are part of the collapsing glacial

“forebulge” and are experiencing subsidence. Thison-going process of crustal “tilting” (see Figure 8)results in a pattern of slowly declining lake levels onnorthern shores with a commensurate rise in lakelevels on southern shores. These relative changes ofwater level for a single lake can be resolved quiteprecisely from a long history of water gauge obser-vations [e.g., Mainville and Craymer 2005].However, it is difficult to accurately relate levels onone lake with those of another.

GPS stations have recently been established atGreat Lakes water gauge sites in Canada by GSD incollaboration with the Ohio State University, and inthe US by the US National Geodetic Survey.Combining these with other GPS measurements (e.g.the CBN) is enabling the determination of an accu-rate and spatially coherent pattern of absolute crustalvelocities that is consistent with the expected rates ofglacial isostatic adjustment. These results will enablelakes to be linked to each other as well as to sea levelin support of bathymetry, hydraulic operations, andhydrological studies in the Great Lakes Basin.

181

Figure 8: Glacial isostatic effects along lake shorelines. In addition to the effect of “tilting” generalized in this illustration, on-goingchanges in average water level may also contribute to the impact on vulnerable coastal regions.

G E O M A T I C A

4.4 Arctic Canada

Observed climate warming in the Arctic maylead to increases in Arctic Ocean sea level[Proshutinsky et al. 2001]. Sea-level rise increasesthe flooding risk and is a clear concern for coastalcommunities such as Tuktoyaktuk [Manson et al.2005]. Furthermore the coastal area subject to risingrelative sea level is expected to expand [Forbes etal., 2004a]. With recent infrastructure developmentassociated with hydrocarbon production in theMackenzie River delta region, this work has becomemore timely. As the impacts on coastlines ultimatelydepend in large measure on the localized rate of rel-ative sea level height change, it is important to betterquantify crustal deformation velocities throughoutcoastal northern Canada. Inherent to this study is theneed to better map the observed pattern of PGRthroughout northern Canada in order to more tightlyconstrain post-glacial isostatic adjustment models.

To provide a framework in which to monitorthese changes, a consistent velocity field will bedetermined from GPS observations throughoutNorth America, including the Canadian ArcticArchipelago and the Mackenzie Delta region. Todetermine the contribution of vertical motion tosea-level rise under climate warming in theCanadian Arctic, ESS in collaboration with theCanadian Hydrographic Service has establishedcollocated tide gauges and continuous GPS at anumber of sites across the Canadian Arctic, fromAlert and Qikiktarjuaq in the east to Ulukhaktok(Holman) and Tuktoyaktuk in the west. The contin-uous GPS sites have been augmented with multi-epoch (episodic) sites in places such as Kugluktuk(Coppermine) and Mercy Bay (northern BanksIsland). This expanded network will furtherenhance regional geophysical studies including thediscrimination of crustal motion, other componentsof coastal subsidence, and sea-level rise.

5. Campaign GPS Constraintson Regional Tectonicsand Seismicity

5.1 Introduction

GPS and other geodetic measurements ofcrustal deformation provide an important new toolfor estimating earthquake hazard. In addition to thestandard method of earthquake hazard based on pastearthquake statistics, it is now possible to relatecurrent deformation rates to the frequency of large

earthquakes. Both continuous and campaign GPSstations have been established for earthquake hazardin four areas, the western Canada subduction zone,the Queen Charlotte transform fault zone, the Yukoncrustal deformation region, and the eastern Canadaregion of high seismicity.

As a complement to a network of continuousGPS stations, campaign GPS surveys provide aneconomical way to obtain a denser map of crustaldeformation. Major limitations associated withcampaign GPS measurements are: (1) lack of reso-lution of short-term (less than one year) deformationepisodes, and (2) less accurate estimates of long-term velocities. Relative velocities across tectonical-ly active regions are typically of the order of 5 - 50mm/yr, which can generally be resolved with cam-paign GPS data acquired every year over a three tofive year period. In the following sub-sections, twoexamples of campaign surveys in western Canada(Queen Charlotte Margin and Northwestern Canada,investigated in collaboration with University ofVictoria) are described, where the relative motionsare associated with plate boundary interactions andstrain transfer within the Canadian Cordillera.

In contrast, relative velocities across intraplate(tectonically stable) regions are typically less than 1mm/yr. Such relative motions are at the current limitof resolution of GPS measurements. An example ofa GPS survey (in collaboration with UniversitéLaval) of intraplate deformation at the eastern edgeof the Canadian Shield along the lower SaintLawrence Valley is presented. In all of the exampleswithin this section, typical surveys consisted ofmeasurements of two to three full days at each site.The data, along with data from nearby continuousstations, were processed at the Pacific GeoscienceCentre (Geological Survey of Canada—PacificDivision, Sidney, BC) with the Bernese GPSSoftware package. Processing details can be found inthe publications referenced in each section.

5.2 The Queen Charlotte Margin

Since mid Eocene (ca. 42 Ma), the QueenCharlotte (Q.C.) margin has been primarily under astrike-slip tectonic regime associated with thePacific-North America relative motion [Hyndmanand Hamilton 1993]. At about 5 Ma, a small changein the Pacific-America relative motion resulted inthe current oblique convergence along Q.C. margin.A wide range of geophysical data indicates a com-ponent of convergence and possibly under-thrustingalong the southern Q.C. margin (e.g. high offshoreheat flow, seismic structure studies) [Smith et al.2003]. The seismicity shows primarily strike-slipfaulting (e.g. the great M=8.1 earthquake of 1949),

182

G E O M A T I C A

although there are some thrust mechanisms near thesouthern end of the islands.

In 1998, NRCan established a small local networkof five campaign GPS sites in the central and north-eastern Q.C. Islands. These sites were re-surveyedin 1998 and 2001 [Mazzotti et al. 2003a]. In 2002,three new sites along the west coast of the islands andone on the mainland at Prince Rupert were then added[Bustin et al. 2004]. The full network (nine sites) hasbeen surveyed every year from 2002 to 2005. Thesecampaign GPS network sites were augmented by thecontinuous station at Williams Lake, to provide astable tie to the North America plate, and campaignmeasurements from the Canadian Base Network sitesin northern British Columbia, to provide some con-straints for the kinematics of the central Cordillera. TheGPS velocity field, with respect to North America,indicates oblique convergence along the Q.C. margin,with relative motion of 5 - 10 mm/yr directed mostlynorthward (refer to Figure 9). The landward gradient inleft lateral shear (i.e. decrease in margin-parallelmotion) shows that the Q.C. fault is currently lockedand is building up strain until the next large earthquakerupture [Mazzotti et al. 2003a]. The results also showthat a small component of the Pacific-North Americaoblique convergence (~10 per cent) is probablyaccommodated by distributed shear across the Q.C.margin and possibly in mainland [Mazzotti et al.,2003b]. A possible interpretation is that this obliqueconvergence is accommodated mainly by partitioninginto strike-slip earthquakes on the Q.C. fault andinfrequent large under-thrusting earthquakes beneaththe margin [Bustin et al., 2004]. Thus, the earth-quake hazard along the Q.C. margin may be higherthan current estimates based solely on an activetranscurrent Q.C. fault.

5.3 The Yukon & Northwestern Canada

Northwestern Canada is known for the intensetectonism during the Mesozoic (ca. 250 to 65 Ma)associated with the accretion and deformation of thedifferent terranes that form the Canadian Cordillera.No sign of significant tectonic activity has beenrecorded along the major faults and deformationzones in most of northwestern Canada for the last~50 Myr. In contrast, the southwestern YukonTerritory and adjacent Alaska Panhandle region isthe locus of intense deformation due to the collisionof the Yakutat block, a composite oceanic-continen-tal terrane that is currently being accreted in the cor-ner of the Gulf of Alaska. This collision produces thehighest mountain ranges in Canada (e.g. St. Eliasrange and Mount Logan). Earthquake activity is veryhigh in this collision zone including several M~8events, but surprisingly substantial seismicity also

occurs in the far-field Mackenzie and RichardsonMountains, around 800 kilometres northeast of thecollision zone [Mazzotti and Hyndman 2002;Hyndman et al. 2005]. The Mackenzie Mountainsappear to be undergoing NE-SW shortening, whileN-S right-lateral strike-slip deformation is occurringin the Richardson Mountains.

In order to understand the details of the wholenorthwestern Canada tectonics and seismicity,NRCan started a campaign GPS survey in 1999across the Yukon and Northwest Territories. The cur-rent network comprises 20 campaign GPS sites, sup-plemented by a larger network of 28 continuous GPSstations from British Columbia to Alaska [Leonardet al. 2005]. The campaign sites were surveyed atfour or five epochs since 1999. The interpretation ofthese results have been complicated by the occur-rence of the very large (magnitude M~8) Denaliearthquake in December 2002. Coseismic and post-seismic displacements related to this earthquake at

183

Figure 9: GPS velocities along the Queen Charlotte margin. Redand orange vectors show velocities with respect to stable NorthAmerica for campaign sites with five years of data or more andthree years of data, respectively. Uncertainty ellipses show the 95per cent confidence region. The large black vector shows themotion (from the rotation vector of DeMets and Dixon [1999]) ofthe Pacific plate relative to North America.

G E O M A T I C A

our GPS sites reach a few centimetres and create anoffset that needs to be removed in order to estimatethe long-term velocities of our sites. Overall, thecontinuous and campaign GPS velocities are con-sistent with the transfer of compressive strain (~5mm/yr) across the northern Canadian Cordillera tothe foreland fold-and-thrust belt with little inter-vening deformation [Mazzotti and Hyndman 2002;Leonard et al. 2005] (see Figure 10). These resultsindicate that the whole northern Cordillera is cur-rently part of the large-scale Pacific-North Americaplate boundary zone and accommodates about 10per cent of the full relative plate motion.

5.4 The Lower Saint Lawrence Valley

The Saint Lawrence Valley, Quebec, presentsone of the largest concentrations of earthquakes ineastern North America. Background seismicityextends over 900 kilometres from the Gulf of SaintLawrence to Montreal following the PaleozoicIapetan Rift system [e.g., Adams and Basham 1991;Lamontage et al. 2003]. Two main seismic zonesoccur along this trend: Charlevoix, the most activein eastern Canada and the locus of at least five M6+earthquakes in the last 350 years; and lower Saint

Lawrence, where the largest known earthquakes areabout M5. Integration of earthquake moment statis-tics in both zones indicates that the equivalent seis-mic deformation rates are 1.0 ± 0.5 mm/yr and 0.2± 0.3 mm/yr, respectively [Mazzotti and Adams2005; Mazzotti et al. 2005].

Unlike plate margins where the seismic activityis directly correlated with plate interactions, easternCanada earthquakes lie in the “stable” interior of theNorth American Plate. The driving mechanismsbehind of these earthquakes are thus more difficult todetermine. Additionally, the hazard posed by poten-tially devastating earthquakes for this region is notwell constrained due to limitations with the proba-bilistic seismic hazard estimation method (e.g. theinexact nature of extrapolating the rate of occurrenceof frequent small events to the occurrence of infre-quent larger events, relatively short instrumental andhistorical records, and limited paleoseismic evidencefor past large events).

In 2003, NRCan started a new program to studythe geodynamic aspects of earthquake hazard ineastern Canada. The study [Mazzotti et al. 2005]used the existing network and data for the CanadianBase Network and added an additional survey to anetwork of 16 stations surrounding the SaintLawrence Valley. These high-precision campaigndata provide relative velocities and strain ratesacross both the Charlevoix and lower SaintLawrence seismic zones based on three to fourcampaigns over the last seven to nine years. On aregional scale, horizontal strain rates are 0.5-2nanostrain per year of roughly NNW-SSE shorten-ing (refer to Figure 11). This strain pattern agreeswell with earthquake focal mechanisms. Horizontalvelocity vectors on both sides of the Saint LawrenceRiver suggest that this shortening corresponds to amaximum convergence of 0.5 ± 0.5 mm/yr betweenthe north and south shores, in general agreementwith the rate from earthquake statistics. Assumingthat seismicity in Charlevoix follows typicalGutenberg-Richter statistics, the GPS results con-strain the return of a potentially very damagingmagnitude M~7 earthquake to ~170 years.

6. Monitoring Crustal Motionsat an Active Plate Boundarywith Continuous GPS

6.1 Introduction

The coastal region of southwestern BritishColumbia comprises the northern portion of the

184

Figure 10: GPS velocities in northwestern Canada. Red and blue vectorsshow velocities with respect to stable North America for continuous and cam-paign GPS sites, respectively. The green vector shows the Yakutat site GPSvelocity. Uncertainty ellipses show the 95 per cent confidence region. Thelarge orange vector shows the Pacific/North America relative motion fromDeMets and Dixon [1999].

G E O M A T I C A

Cascadia Subduction Zone (CSZ) that forms theconvergent plate boundary between the Juan deFuca and North America plates. This boundary ismarked by extensive earthquake activity as well asa belt of active arc volcanoes stretching from north-ern California to southwest British Columbia.Special aspects of this subduction zone are the rel-ative youth of the subducting oceanic crust, whichranges from six to 10 Ma at the oceanic trench, anda modest convergence rate of under 40 mm/yr.Seismic, thermal, and geodetic studies have deter-mined that the shallower interface of the CSZ islocked, and that stress is accumulating for the nextgreat subduction thrust earthquake. From paleo-seismic evidence, such great events occur every500 to 600 years, the last having occurred in 1700.

The monitoring of crustal motions along thecoastal margin of southwestern BC using continuousGPS stations was initiated on Vancouver Island in1992 with sites established in Victoria (ALBH) andHolberg (HOLB) [Dragert and Hyndman 1995].These sites, along with a reference site nearPenticton (DRAO), form the basis of the WesternCanada Deformation Array (WCDA) which nowconsists of 16 stations in southwestern BC. Usinglow-multipath antennas mounted on stable geodeticmonuments, the relative position of these stationshundreds of kilometres apart can be monitored on adaily basis to an accuracy of a few millimetres. Thislevel of accuracy allows the resolution of long-termcrustal movement associated with major platemotions as well as the relative squeezing and buck-ling of the plate margins. More recently, GPS datafrom a number of WCDA stations have also revealedtransient crustal motions that are related to repeatedslow slip on the deeper subducting plate interface.

6.2 Long-term Deformation

Analyses of GPS data from WCDA sites haveconfirmed that long-term elastic deformationoccurs along the northern Cascadia Margin due tothe locking of converging plates across a portion ofthe subduction interface between the Juan de Fuca(JDF) plate and the overlying North America (NA)plate [cf. Flueck et al. 1997; Mazzotti et al. 2003c].The motion vectors shown in Figure 12, the largestexceeding 1 cm/yr, are based on the linear trends inthe time series of changes in horizontal positions ofGPS sites with respect to the reference site DRAO,located south of Penticton, British Columbia, andassumed fixed on the NA plate. The pattern of theregional crustal velocity field is a key constraint indetermining the location and extent of the lockedfault zone - i.e. that portion of the fault that willultimately rupture in a great subduction-thrust

earthquake [cf. Wang et al. 2003]. For the northernCascadia margin, the fully locked portion of thesubducting plate interface lies offshore beneath thecontinental slope and the “transition zone,” markedby a transition from fully-locked to free-slippingplates, terminates directly beneath the coastline ofVancouver Island and the western edge of theOlympic Peninsula. The landward extent of theincipient rupture zone for the next megathrustearthquake is a key parameter for estimating strongmotions that can be expected in the densely popu-lated areas of Vancouver and Seattle from such anearthquake.

6.3 Episodic Tremor and Slip

Improvements in the accuracy of IGS preciseorbits and the regional densification of continuousGPS coverage were two key factors leading to thediscovery of a “silent” slip event that occurredalong the Cascadia Subduction Zone in August1999 [Dragert et al. 2001]. Analysis of GPS datafrom 1994 to 2005 has revealed that the motions ofcontinuous GPS sites in northern Washington Stateand southern Vancouver Island are marked bynumerous, brief, episodic reversals. This is bestillustrated by the east-component time series at the

185

Figure 11: GPS horizontal strain-rates for the Saint Lawrence seismic zones.Dark cyan arrows show the average horizontal principal strain rates for thenetwork (spatial extent given by dotted grey polygon). Red arrows show thehorizontal principal strain rates for sub-networks around the Charlevoix andlower Saint Lawrence seismic zones (spatial extent given by dotted red poly-gons). Light grey circles indicate the pattern of regional seismicity [afterMazzotti et al. 2005].

G E O M A T I C A

Victoria GPS site (ALBH) where the motion relativeto DRAO is clearly characterized by a sloped saw-tooth function: For periods of 13 to 16 months, thereis eastward motion that is more rapid than the long-term rate, followed by a one to three week period ofreversed motion (refer to Figure 13). These transientreversals of motion are surprisingly regular events,having a recurrence interval of 445 ± 35 days. Thecumulative displacements over the two-week peri-ods of these transient events is generally less than sixmillimetres, in a direction opposite to the longer-term deformation motion. For the longer-lastingevents, station displacements were not simultaneousat all coastal margin sites but appear to migrate alongstrike of the subduction zone at speeds ranging from5 to 15 km/day. These transient events are accompa-nied by distinct, low-frequency tremors [e.g., Rogersand Dragert 2003], similar to those reported in theforearc region of southern Japan [Obara, 2002]. Theone-to-one correspondence of slip events with pro-nounced tremor activity prompted the naming of thisphenomenon as “Episodic Tremor and Slip” (ETS)[Rogers and Dragert 2003].

The transient surface displacements can bereplicated by simple slip-dislocation models if oneadopts the subduction interface geometry fromFlueck et al. [1997] and assumes that slip occurs atthe plate interface and consists of updip motion ofthe overlying crustal block. Figure 14 shows mod-elling results for four of these events. The goodagreement between observed and modeled dis-placements shows that the transient motions can berepresented by simple slip on the subducting plateinterface between depths of 25 and 45 km, with theslip region parallel to the strike of the subductingplate. The downdip boundary appears to be sharperwhereas the updip boundary is more diffuse, requir-ing a gradual tapering of the slip amplitude.Although the maximum slip is only a few centime-tres per event, the large area of slip generates anequivalent moment magnitude for these “slowearthquakes” ranging from 6.5 to 6.8.

Although the physical processes involved arenot well understood, Figure 15 outlines the concep-tual kinematic model for ETS on the northern CSZ.Both offshore and at greater depths (>50 km), thetwo plates converge steadily at ~4 cm/yr, the geo-logical average rate. Across the shallower interface,the plates are locked for centuries, moving cata-strophically past each other only at times of greatthrust earthquakes. At depths of 25 to 45 km, platesresist motion temporarily for ~14 months accumu-lating some stress, and then slip the equivalent of afew centimetres over periods of one to two weeksreleasing that small stress accumulation. Thisrelease is accompanied by unique seismic tremors

186

Figure 12: Long-term velocities of regional continuous GPS sites. Three to 10-year linear trends in the horizontal position with respect to Penticton (DRAO)for some of the sites of the Western Canada Deformation Array (WCDA: greensquares) and the Pacific Northwest Geodetic Array (PANGA: yellow squares)are plotted by green arrows with 95 per cent error ellipses. The position of thestrongly coupled zone determined from slip-dislocation models is indicated bythe locked and transition zones. The convergence vector of the Juan de Fucaplate is with respect to the North America (NA) plate, and the GPS referencestation DRAO is assumed fixed on the NA plate [from Dragert et al. 2004].

Figure 13: Record of slip and tremor activity observed for the Victoria area.Blue circles show day-by-day change in the east component of the GPS siteALBH (Victoria) with respect to DRAO (Penticton) which is assumed fixed onthe North America plate. Continuous green line shows the long-term (interseis-mic) eastward motion of the site. Red line segments show the mean elevatedeastward trends between the slip events which are marked by the reversals ofmotion every 13 to 16 months. Bottom graph shows the total number of hoursof tremor activity observed for southern Vancouver Island within a sliding 10-day period. Ten days corresponds to the nominal duration of a slip event.Pronounced tremor activity coincides precisely with slip events.

G E O M A T I C A

on and above the region of slip. In the context ofseismic hazard, the recognition of this ETS slipzone has two significant implications. First, sincethe repeated slip events define a region where littleor no shear stress accumulates over longer periodsof time, the updip edge of the slip zone can serve asa proxy for the maximum downdip (i.e. landward)extent for subduction thrust rupture. This can beused to improve estimates of strong motion in thedensely populated regions of southwest BC.Secondly, it can be shown that the occurrence ofslip on this deeper part of the plate interface addsstress to the shallower portion of the locked plateinterface, driving it closer to rupture in discreteincrements. Consequently, during the time of slip,the weekly cumulative probability of a megathrustearthquake is significantly greater than at timesbetween slip events [Mazzotti and Adams 2004].This provides, for the first time, a potential basisfor time-dependent seismic hazard estimates.

7. Summary

Geophysical processes systematically affect thespatial reference frames used for geodetic surveys.Geodynamic rates in most of Canada (i.e. away fromactive plate margins) are generally rather modest(typically 1-2 cm/yr). However, when highest accu-racy is required, the measurable effects of geody-namic processes must be considered. When evaluat-ing the effect on reference frames within a given

187

Figure 14: Models for slow slip events. The geometry of the subducting plateinterface established from structural studies [Flueck et al., 1997] is adopted,and simple linear slip distribution directed up-dip is assumed. The four panelsillustrate the results of elastic dislocation modelling [Okada, 1985] of four recentslip events. Blue arrows show observed horizontal surface displacements with95 per cent error ellipses that were used to constrain the models; green arrowsshow surface displacements obtained from separate analyses and not used inmodel constraint; yellow vectors show model displacements. Dark shading indi-cates fault areas with full slip whose magnitudes are shown in panel headers;light shading indicates fault areas where slip is tapered linearly from full to zeroat the up-dip end. Also shown are equivalent moment-magnitudes assuming arigidity of 40 Gpa [from Dragert et al. 2004].

Figure 15: Cross-section of a conceptual model for CSZ plate-interface kinematics. The blue region represents the subducting Juande Fuca plate and the green and brown regions, the overlying margin of the North America plate. Seaward of the trench (to the leftof the diagram) and landward of Victoria (to the right), the plates converge continuously. Beneath the continental slope, mostly com-prised of the accretionary prism, the plates are fully locked thereby causing an accumulation of stress that will be released in thenext megathrust earthquake. At depths of 25 to 45 km, plates resist motion temporarily for ~14 months and then slip a few cen-timetres over periods of one to two weeks, accompanied by distinct seismic tremors (yellow stars). The transition zone marks theregion on the plate interface where a transition from “locked” to “stick-slip” behaviour occurs.

G E O M A T I C A

area it is important to note how kinematic processesoperate on various spatial scales.

Within regions of particular interest (e.g. earth-quake hazards), NRCan has complemented theCSRS and installed additional geodetic infrastruc-ture to monitor specific geodynamic processes. Anobservation strategy for geodynamics based mainlyon GPS should optimize both the accuracy of resultsand the use of complementary data sets from inde-pendent techniques where tenable (e.g. absolutegravity, tide gauges). All geodynamic investigationsrequire a long-term commitment to systematicmonitoring in order to determine the relatively smallkinematic rates and to subsequently develop usefulquantitative models. Many years of observations aretherefore invaluable and often required to betterunderstand the particular regional geodynamicprocesses. Furthermore, the data and results can beused to improve the realization and maintenance ofnational NAD83(CSRS) reference frame for thegeneral georeferencing community.

For much of the Canadian landmass the mostsignificant crustal movement is vertical motion, asdemonstrated by current models and analyses ofpost-glacial rebound. Post-glacial isostatic adjust-ment, whose maximum uplift rates occur generallynear Hudson Bay, is a process that is evidentthroughout much of Canada. Crustal tilt associatedwith post-glacial rebound can, albeit slowly, affectwater resources and alter risks associated withflooding. Of specific practical importance arechanges in relative sea level for coastal areas includ-ing regions that may be particularly vulnerable toclimate change such as arctic regions. It is thereforehighly useful to monitor and confirm coastal heightvariations at geodetic reference stations over time.

Seismic activity in intraplate zones tends to berelated to the regional stress fields, with earth-quakes concentrated in regions of crustal weakness.Horizontal strain rates can be directly related to thefrequency of large earthquakes. Vertical motionsmay provide additional insight into the regionalseismic process. Thus the occurrence of large earth-quakes in active seismic regions of eastern Canada(e.g. lower Saint Lawrence Valley, Charlevoix, andOttawa Valley) can be better characterized throughlong-term precise geodetic monitoring.

Detailed regional surveys continue to map thecrustal deformation field associated with the geo-physical processes in Canada noted above. Forregions with active seismicity and/or faults, meas-urements of crustal deformation supply direct infor-mation on crustal-strain accumulation which isessential for studies of tectonic and seismogenicprocesses and is increasingly used in seismic hazardassessments. Particularly exciting was discovery of

the “silent-slip” phenomenon along the Cascadiasubduction-zone interface. This was first observedwithin the time-series of precise, continuous GPSmeasurements on Canada’s west coast and itsimplications continue to be explored.

Acknowledgments

The authors first wish to acknowledge andexpress our appreciation to the many dedicated indi-viduals of NRCan for their roles in providing theGPS and AG data sets. From site selection and mon-umentation to acquisition, validation and data pro-cessing, the calibre of results from continuous andepisodic sites reflects a collective effort to attain thehighest quality solutions possible. Universities andother government agencies have been tremendouspartners in these efforts. We would like to thank themultitude of national and international sources ofsupport that have made possible or enhanced manyof the studies presented in this paper. Detailedacknowledgments of support agencies and collabo-rative partners can be found in the publications ref-erenced within each section. Work reported here onArctic sea levels and coastal stability is supported byPERD and ArcticNet (Network of Centres ofExcellence Canada). Numerous graduate studentsworking under the academic guidance of GeorgeSpence (University of Victoria) have contributed tothe studies of crustal deformation along the westernmargin of Canada. Mike Schmidt (NRCan/GSC) isgratefully acknowledged for his long-standing com-mitment to the operation of the WCDA and techni-cal support for western Canada GPS surveys. Wealso would like to acknowledge Earl Lapelle andMike Piraszewski (NRCan/GSD) for their dedica-tion to GPS data processing and network adjust-ments and combinations. Many of the maps in thisdocument were drafted using GMT [Wessel andSmith 1998], an open source collection of ~60 toolsfor manipulating geographic and Cartesian datasets. We would like to thank Pierre Héroux(NRCan/GSD) for his role leading the CSRS-relatedcontributions collected in this special issue ofGEOMATICA. Finally, the authors wish to especial-ly thank Robert Duval, Calvin Klatt, Yves Mireault(NRCan/GSD), and Roy Hyndman (NRCan/GSC)for their constructive reviews of the manuscript.

References and Bibliographies

Adams, J. and P.W. Basham. 1991. The seismicity andseismotectonics of eastern Canada, in Neotectonics

188

G E O M A T I C A

of North America, Slemmons, D.B., E.R. Engdhal,M.D. Zoback and D.D. Blackwell (eds.), Geol. Soc.Amer. Decade Map Vol. 1, Boulder CO, pp. 261-276.

Altamimi, Z., D. Angermann, D. Argus, G. Blewitt, C.Boucher, B. Chao, H. Drewes, R. Eanes, M. Feissel,R. Ferland, T. Herring, B. Holt, J. Johannson, K.Larson, C. Ma, J. Manning, C. Meertens, A.Nothnagel, E. Pavlis, G. Petit, J. Ray, J. Ries, H-G.Scherneck, P. Sillard and M. Watkins. 2001. The ter-restrial Reference Frame and the dynamic Earth,Eos Trans AGU, 82(25), pp. 273, 278-279.

Altamimi, Z., P. Sillard and C. Boucher. 2002.ITRF2000: A new release of the InternationalTerrestrial Reference Frame for earth science appli-cations, J. Geophys. Res., vol. 107, 2214,doi:10.1029/2001JB000561.

Blewitt, G., D. Argus, R. Bennett, Y. Bock, E. Calais, M.Craymer, J. Davis, T. Dixon, J. Freymueller, T.Herring, D. Johnson, K. Larson, M. Miller, G. Sella,R. Snay and M. Tamisiea. 2005. A Stable NorthAmerican Reference Frame (SNARF): FirstRelease, Joint UNAVCO/IRIS Workshop(Proceedings), Stevenson, WA, USA, June 9-11.

Blewitt, G. 1997. IGS Densification Program, IGS 1996Annual Report, International GPS Service, pp. 24-25.

Boucher, C., Z. Altamimi, P. Sillard and M. Feissel-Vernier. 2004. The ITRF2000, (IERS TechnicalNote; 31) Frankfurt am Main: Verlag desBundesamts für Kartographie und Geodäsie, 2004.289 pp., paperback, ISBN 3-89888-881-9.

Bustin, A.M., R.D. Hyndman, J. Cassidy and S.Mazzotti. 2004. Strain Partitioning of ObliqueConvergence: Initiation of Subduction on the QueenCharlotte Margin?, Eos Trans. AGU, 85(17), Jt.Assem. Suppl., Abstract T31C-02.

Craymer, M.R. and M. Piraszewski. 2001. The NorthAmerican Reference Frame (NAREF): An Initiativeto Densify the ITRF in North America, Proceedingsof KIS 2001: International Symposium onKinematic Systems in Geodesy, Geomatics andNavigation, Banff, Canada, June 5-8.

DeMets, C. and T.H. Dixon. 1999. New kinematic mod-els for Pacific-North America motion from 3 Ma topresent, I: Evidence for steady motion and biases inthe NUVEL-1A model, Geophys. Res. Lett., vol. 26,pp. 1921–1924.

Dragert, H. and R.D. Hyndman. 1995. Continuous GPSmonitoring of elastic strain in the northern Cascadiasubduction zone, Geophys. Res. Lett., vol. 22, pp.755-758.

Dragert, H., K. Wang and G. Rogers. 2004. Geodetic andSeismic Signatures of Episodic Tremor and Slip inthe Northern Cascadia Subduction Zone, EarthPlanets Space, vol. 56, 1143-1150.

Dragert, H., K. Wang and T.S. James. 2001. A Silent SlipEvent on the Deeper Cascadia Subduction Interface,Science, vol. 292, pp. 1525-1528.

Dyke, A.S. 2004. An outline of North American deglacia-tion with emphasis on central and northern Canada,Quaternary Glaciations - Extent and Chronology,

Part 2, North America, ed. J.Ehlers and P.L.Gibbard, Elsevier.

Ferland, R., Z. Altamimi, C. Bruyninx, M. Craymer, H.Habrich and J. Kouba. 2003. IGS Position Paper:Regional Networks Densification, Proceedings of theIGS Network, Data and Analysis Center Workshop,Ottawa, April 8-11, IGS Central Bureau, JetPropulsion Laboratory, Pasadena CA, pp. 123-132.

Flueck, P., R.D. Hyndman and K. Wang. 1997. Three-dimensional dislocation model for great earthquakesof the Cascadia Subduction zone, J. Geophys. Res.,vol. 102, pp. 20539-20550.

Forbes, D.L., M. Craymer, S.M. Solomon and G.K.Manson. 2004a. Defining limits of submergence andpotential for rapid coastal change in the CanadianArctic, Rep. Polar and Marine Res (Alfred WegenerInstitute), vol. 482, pp. 196-202.

Forbes, D.L., G.S. Parkes, G.K. Manson and L.A. Ketch.2004b. Storms and shoreline retreat in the southernGulf of St. Lawrence. Mar. Geol., vol. 210, pp. 169-204.

Henton, J.A., M.R. Craymer, M. Piraszewski and E.Lapelle. 2004. Crustal Deformation Velocities FromEpisodic Regional Measurements at Canadian BaseNetwork Sites, Eos Trans. AGU, 85(47), Fall Meet.Suppl., Abstract G31B-0797.

Hyndman, R.D. and T.S. Hamilton. 1993. Queen Charlottearea Cenozoic tectonics and volcanism and theirassociation with relative plate motion along thenortheastern Pacific margin, J. Geophys. Res., vol.98, pp. 14,257-14,277.

Hyndman, R.D., P. Flück, S. Mazzotti, T. Lewis, J. Ristauand L. Leonard. 2005. Current tectonics of theNorthern Canadian Cordillera, Can. J. Earth Sci., vol.42, doi:10.1139/E05-023, pp. 1117-1136.

Lambert, A. N. Courtier, G.S. Sasagawa, F. Klopping, D.Winester, T.S. James and J.O. Liard. 2001. New con-straints on Laurentide postglacial rebound fromabsolute gravity measurements, Geophys. Res. Lett.,vol. 28, pp. 2109-2112.

Lamontagne, M., P. Keating and S. Perreault. 2003.Seismotectonic of the Lower St. Lawrence seismiczone, Québec: Insights from geology, magnetics,gravity, and seismics, Can. J. Earth Sci., vol. 40, pp.317-336.

Leonard, L. J., R.D. Hyndman, S. Mazzotti, J.F. Cassidy,G.C. Rogers and J. Adams. 2005. GPS measure-ments, seismicity, and current tectonics in the north-ern Canadian Cordillera: Northward stress transferfrom the Yakutat collision, Proceedings, IASPEIConference, Santiago, Chile, October 2-7.

Mainville, A. and M.R. Craymer. 2005. Present-day tiltingof the Great Lakes region based on water levelgauges, GSA Bulletin, vol. 117 (7/8),doi:10.1130/B25392.1, pp. 1070-1080.

Manson, G.K., S.M. Solomon, D.L. Forbes, D.E. Atkinsonand M. Craymer. 2005. Spatial variability of factorsinfluencing coastal change in the western CanadianArctic, Geo-Mar. Lett., vol. 25, doi:10.1007/s00367-004-0195-9, pp. 138-145.

Mazzotti, S. and J. Adams. 2005. Rates and uncertaintieson seismic moment and deformation in eastern

189

G E O M A T I C A

Canada, J. Geophys. Res., vol. 110, B09301,doi:10.1029/2004JB003510.

Mazzotti S. and J. Adams. 2004. Variability of near-termprobability for the next great earthquake on theCascadia subduction zone, Bull. Seism. Soc. Am., vol.94, pp. 1954-1959.

Mazzotti, S. and R.D. Hyndman. 2002. Yakutat collisionand strain transfer across the northern CanadianCordillera, Geology, vol. 30, pp. 495-498.

Mazzotti, S., R.D. Hyndman, P. Flück, A. J. Smith and M.Schmidt. 2003a. Distribution of the Pacific/NorthAmerica motion in the Queen Charlotte Islands-S.Alaska plate boundary zone, Geophys. Res. Lett., vol.30, 1762, doi:10.1029/2003GL017586.

Mazzotti, S., R.D. Hyndman and A. Bustin. 2003b. Strainand stress partitioning of the Pacific/North Americainteraction along the Queen Charlotte Islands-AlaskaPanhandle, Eos Trans. AGU, 84(46), Fall Meet.Suppl., Abstract T52A-0229.

Mazzotti S., H. Dragert, J. Henton, M. Schmidt, R.D.Hyndman, T.S. James, Y. Lu and M. Craymer. 2003c.Current tectonics of northern Cascadia from a decadeof GPS measurements, J. Geoph. Res., vol. 108,2554, doi:10.1029/2003JB002653.

Mazzotti, S., T.S. James, J. Henton and J. Adams. 2005.GPS crustal strain, postglacial rebound, and seismichazard in eastern North America: The SaintLawrence valley example, J. Geophys. Res., vol. 110,B11301, doi:10.1029/2004JB003590.

Obara, K. 2002. Non-volcanic deep tremor associated withsubduction in southwest Japan, Science, vol. 292, pp.1679-1681.

Okada, Y. 1985. Surface deformation due to shear and ten-sile faults in a half space, Bull. Seismol. Soc. Am.,vol. 75, pp. 1135-1154.

O’Reilly, C.T., D.L. Forbes and G.S. Parkes. 2005.Defining and adapting to coastal hazards in AtlanticCanada: facing the challenge of rising sea levels,storm surges, and shoreline erosion in a changing cli-mate, Ocean Yearbook, vol. 19, pp. 189-207.

Peltier, W.R. 1994. Ice age paleotopography, Science, vol.265, pp. 195-201.

Proshutinsky, A.Y., V. Pavlov and R.H. Bourke. 2001. Sealevel rise in the Arctic Ocean, Geophys. Res. Lett.,vol. 28, pp. 2237-2240.

Rogers, G.C. and H. Dragert. 2003. Episodic tremor andslip on the Cascadia Subduction Zone: The chatter ofsilent slip, Science, vol. 300, pp. 1942-1943.

Smith, A. J., R.D. Hyndman, J.F. Cassidy and K. Wang.2003. Structure, seismicity, and thermal regime of theQueen Charlotte Transform Margin, J. Geophys.Res., vol.108, 2539, doi:10.1029/2002JB002247.

Tushingham, A.M. and W.R. Peltier. 1991. ICE-3G: A newglobal model of late Pleistocene deglaciation basedupon geophysical predictions of post-glacial relativesea-level change, J. Geophys. Res., vol. 96, 4497-4523.

Wang, K., R. Wells, S. Mazzotti, R.D. Hyndman and T.Sagiya. 2003. A revised dislocation model of inter-seismic deformation on the Cascadia subductionzone, J. Geophys. Res., vol. 108, pp. 2009-2022.

Wessel, P. and W.H.F. Smith. 1998. New, improved ver-

sion of the Generic Mapping Tools released, EosTrans. Amer. Geophys. U., 79(47), pp. 329.

Authors

Joseph Henton has served as a Space Geodetic& Gravity Systems Research Geophysicist with theGeodetic Survey Division of Natural ResourcesCanada (NRCan) since 2002. While pursuing hisDoctoral degree at the University of Victoria (UVic),he studied and collaborated with researchers at theGeological Survey of Canada (GSC) in Sidney,British Columbia. He obtained his PhD from theSchool of Earth and Ocean Sciences of UVic in2000. He holds B.S. and M.S. degrees in solid-earthgeophysics from the New Mexico Institute ofMining & Technology. His interests currently lie inthe application of geodetic techniques (precise GPSand absolute gravimetry) to geoscience and geomat-ics issues. Within NRCan his work primarily sup-ports the Canadian Spatial Reference System, but hisresearch interests also contribute to a number of ini-tiatives and projects, including studies of Canada'svulnerability to climate change, groundwater, andearthquake hazards.

Mike Craymer has been with the GeodeticSurvey Division of Natural Resources Canadasince 1991 and is presently head of GeodeticNetworks & Standards. He received his PhD ingeodesy from the University of Toronto and hasauthored many papers on a variety of geodetic top-ics. He is an associate member of the InternationalGGNS Service and is presently co-Chair of IAGSub-Commission 1.3c on Regional ReferenceFrames for North America. He is also co-editor ofthe GPS Toolbox column for the journal "GPSSolutions," editor of the Literature Review columnfor the journal "Surveying and Land InformationScience," and publisher of the popular Tables ofContents in Geodesy <www.geodetic.org/tcg>.

Dr. Herb Dragert is a research scientist with theGeological Survey of Canada (GSC) at the PacificGeoscience Centre (PGC) in Sidney, BritishColumbia. He obtained his BSc degree inMathematics & Physics at the University of Torontoin 1968, and his MSc and PhD degrees inGeophysics at the University of British Columbia in1970 and 1973 respectively. He next held a NATOPost doctoral Fellowship at the Institute forGeophysics, Goettingen, Germany, and returned toCanada to teach at UBC in the fall of 1974. Hejoined the Gravity and Geodynamics Division of the

190

G E O M A T I C A

Earth Physics Branch in Ottawa in 1976, and movedto PGC at the end of 1978. He currently also holdsan adjunct professor position with the School ofEarth & Ocean Sciences at the University ofVictoria. His principal area of research has been thestudy of crustal deformation within active seismicareas on the west coast of Canada. In 1992, he ledthe establishment of the Western CanadaDeformation Array, the first continuous GPS net-work in Canada for the express purpose of monitor-ing crustal motions. It was data from this networkthat led to the discovery of "Episodic Tremor andSlip" in the Cascadia Subduction Zone.

Dr. Stéphane Mazzotti obtained his BSc andMSc at the University of Paris-Orsay (France), andafterwards he worked as a visiting scientist at theUniversity of Tokyo (Japan) in 1996 to 1997. Hethen returned to France and obtained his PhD in geo-dynamics at Ecole Normale Superieure, Paris, in1999. Dr. Mazzotti has been working in Canadasince, first as a post-doctoral fellow at the Universityof Victoria, now as a research scientist with theGeological Survey of Canada since 2003. Dr.Mazzotti specializes in the study of relationshipsbetween seismicity and lithospheric deformation.His main research focuses on the integration of GPS,earthquake, and other geophysical data to studycrustal dynamics, seismicity, and seismic hazards ofplate boundary and intraplate deformation regions.Currently, his principal areas of interest are westernCanada (Cascadia, Queen Charlotte Islands, Yukon)and the St. Lawrence and Ottawa valleys.

Rémi Ferland has been involved in variousGPS research and development activities sincemid-1980. With the Geodetic Survey Division of

Natural Resources Canada, he has been involved inthe early 1990s with the estimation of precise GPSorbit, station and Earth related products. Since1999, he has been actively involved as the coordi-nator of the IGS Reference Frame Working Group.His responsibilities include the weekly combina-tion of the IGS station position and velocity prod-ucts, and Earth rotation parameters along with theselection and maintenance of the stations whosepositions and velocities are included in the IGSrealization of the International TerrestrialReference Frame (ITRF). Rémi has also beeninvolved with the IERS combination activities andis a member of the NAREF Reference FrameTransformation Working Group, responsible fordetermining the relationship between regional,national, and international reference frames.

Dr. Donald L. Forbes is a Research Scientistwith the Geological Survey of Canada at the BedfordInstitute of Oceanography and an Adjunct Professorat Memorial University of Newfoundland. He holdsdegrees from Carleton University, the University ofToronto, and the University of British Columbia. Hisresearch interests range from seabed mapping andremote sensing for coastal change detection to sea-level change, vertical motion, coastal hazards, andclimate-change adaptation. In the mid 1990s, he wasseconded to the South Pacific Applied GeoscienceCommission working on coastal issues in Fiji, CookIslands, Niue, and Kiribati. Over the past decade, hehas played lead roles in projects on coastal impactsof climate change in the Canadian Arctic andAtlantic provinces. He has wide experience inapplied science studies of coastal systems for appli-cation to public policy, integrated management,hazard assessment and adaptation planning. o

191