teleseismic studies of the lithosphere below the abitibi ... · teleseismic studies of the...

Teleseismic studies of the lithosphere below theAbitibi–Grenville Lithoprobe transect1,2

Stéphane Rondenay, Michael G. Bostock, Thomas M. Hearn, Donald J. White,Hua Wu, Guy Sénéchal, Shaocheng Ji, and Marianne Mareschal3

Abstract: In the past decade, the Abitibi–Grenville Lithoprobe transect has been the site of numerous geological andgeophysical surveys oriented towards understanding the lithospheric evolution of the southeastern Superior and adjoin-ing Grenville provinces. Among the different geophysical methods that have been employed, earthquake seismologyprovides the widest range of information on the deep structures of the upper mantle. This paper presents a review ofstudies, both complete and ongoing, involving teleseismic datasets that were collected in 1994 and 1996 along thetransect. A complete shear-wave splitting analysis has been performed on the 1994 dataset as part of a comparativestudy on electrical and seismic anisotropies. Results suggest a correlation between the two anisotropies (supported byxenolith data) and favour a lithospheric origin for the seismic anisotropy. The two anisotropies are believed to representthe fossilized remnants of Archean strain fields in the lithospheric roots of the Canadian Shield. Preliminary splittingresults for the 1996 experiment suggest that theS-wave azimuthal anisotropy may be depth dependent and laterallyvarying. Ongoing receiver function analysis and traveltime inversion studies provide velocity models of the crust andupper mantle beneath the study area. Preliminary receiver function results reveal the presence of anS-velocity increaseat -90–100 km depth which appears to be laterally continuous over 200 km. Traveltime inversion models indicate thepresence of an elongate, low-velocity anomaly beneath the southern portion of the 1996 array which strikes obliquelyto major geological structures at the surface (e.g., Grenville Front). Preliminary interpretation relates this anomaly tothe same process (e.g., fixed mantle plume, continental rifting) responsible for the emplacement of the MonteregianHills igneous province.

Résumé: Au cours de la dernière décennie, de nombreuses campagnes géologiques et géophysiques ont été effectuéesle long de la traverse Abitibi-Grenville du projet Lithoprobe, dans le but de comprendre l'évolution lithosphérique de lapartie sud-est de la Province du Supérieur et de la Province de Grenville. Parmi les différentes méthodes géophysiquesemployées, la sismologie passive est sans doute celle qui procure le plus d'informations sur les structures profondes dumanteau supérieur. Cet article passe en revue les études, complétées et en cours, s'appuyant sur les données télésismi-ques enregistrées en 1994 et en 1996, le long de la traverse. Une analyse complète de biréfringence des ondesS a étéeffectuée sur les données de 1994, dans le cadre d'une étude comparative entre l'anisotropie électrique et sismique. Lesrésultats obtenus suggèrent une corrélation entre les deux anisotropies (observation supportée par l'analyse de xénoli-thes), et favorisent une anisotropie sismique d'origine lithosphérique. Les deux anisotropies semblent associées à deschamps de déformation archéens fossilisés dans les racines lithosphériques du Bouclier canadien. Les résultats prélimi-naires d'analyse de biréfringence sur les données de 1996 suggèrent une anisotropie azimutale d'ondesS variant latéra-lement et en fonction de la profondeur. Les études en cours, comprenant l'analyse des fonctions de transfert etl'inversion des temps de parcours, permettent d'obtenir des modèles de vitesses sismiques pour la crôute et le manteausupérieur, sous la région d'étude. Les résultats préliminaires de l'analyse des fonctions de transfert révèlent l'existenced'un saut positif de la vitesse d'ondesS, situé à une profondeur d'environ 90–100 km et latéralement continu sur unedistance de 200 km. Les modèles obtenus par inversion des temps de parcours indiquent la présence d'une anomalie al-longée de faible vitesse, qui croise la portion sud du dispositif d'écoute de 1996 et présente une orientation oblique auxprincipales structures de la région (e.g., Front de Grenville). L'interprétation préliminaire des résultats relie cette ano-malie de faible vitesse au même processus que celui étant à l'origine de la mise en place de la province ignée des col-lines Montérégiennes (e.g., panache mantellique bloqué, rift continental).

[Traduit par la Rédaction] Rondenay et al. 426

Can. J. Earth Sci.37: 415–426 (2000) © 2000 NRC Canada

415

Received April 29, 1998. Accepted October 13, 1998.

S. Rondenay4 and M.G. Bostock. Department of Earth and Ocean Sciences, The University of British Columbia, Vancouver,BC V6T 1Z4, Canada.T.M. Hearn. Physics Department, New Mexico State University, Box 30001, Department 3D, Las Cruces, NM 88003-0001, U.S.A.D.J. White. Geological Survey of Canada, 615 Booth St., Ottawa, ON K1A 0E9, Canada.H. Wu and M. Mareschal. Génie minéral, École polytechnique de Montréal, P.O. Box 6079, Succursale Centre-Ville, Montréal,QC H3C3A7, Canada.G. Sénéchal.I.P.R.A., Université de Pau et des Pays de l’Adour, 64000 Pau, France.S. Ji. Département de géologie, Université de Montréal, P.O. Box 6128, Succursale Centre-Ville, Montréal, QC H3C3J7, Canada.

1Lithoprobe Publication 962.2Geological Survey of Canada Contribution 19982063Deceased on July 11, 1995.4Corresponding author (e-mail: [email protected]).

Introduction

In recent years, a variety of studies has greatly increasedour knowledge of lithospheric evolution in the region en-compassed by the Abitibi–Grenville Lithoprobe transect, anarea which has remained tectonically stable for the past1000 Ma. Among these studies, earthquake seismology is anessential tool in probing deep Earth structures. Coupled withdata from xenoliths that are brought to the surface by local-ized and infrequent kimberlite volcanism, teleseismic signalsprovide important information on the nature and structure ofthe lithosphere and the tectonic processes that have contrib-uted to its evolution.

This paper presents a review of the teleseismic studiesthat have been performed along the Abitibi–Grenvilletransect, where two separate arrays were deployed (one in1994 and the other in 1996). Data from the former arraywere used for shear-wave splitting analysis and published inSénéchal et al. (1996) and Ji et al. (1996); a review of theirresults and interpretations is presented here. Other studies,involving receiver function analysis and traveltime inversionand shear-wave splitting on the 1996 dataset, are still inprogress. Preliminary results from these latter studies arediscussed.

Study area

Tectonic settingThe data used in this paper were collected along two ar-

rays in the region shown in Fig. 1. This area includes a di-versity of geological domains ranging in age from lateArchean in the north to Paleozoic in the south and a numberof major tectonic boundaries.

The northern half of the study area comprises the SuperiorProvince of the Canadian Shield, the largest Archean prov-ince in the world (Card 1990). The Superior Province con-sists of a sequence of generally east–west-trending volcano-plutonic (granite–greenstone), metasedimentary, plutonic,and high-grade gneiss subprovinces (Card and Ciesielski1986). The Opatica plutonic belt is the northernmostsubprovince of the Superior Province shown in Fig. 1. It isinterpreted as the deeply eroded core of an Archean orogen(Benn et al. 1992; Sawyer and Benn 1993) onto which theyounger terraces to the south were accreted by a complexprocess of northward thrusting of the upper crust and north-dipping subduction of the lower crust (Calvert et al. 1995).To the south, the Abitibi subprovince is composed of alter-nating east–west-trending volcanic (greenstone) and plutonicbelts (Ludden and Hubert 1986). This assemblage is be-lieved to have formed from northward accretion (with north-dipping or perhaps northwest oblique subduction) of oceanicplateaus and (or) paired arcs and sedimentary prisms(Kimura et al. 1993; Desrochers et al. 1993). The southern-most Archean terranes traversed by the teleseismic arraysform the Pontiac metasedimentary subprovince, interpretedby Kimura et al. (1993) as a late-stage synorogenic turbiditefan associated with a late Archean prograding orogen involv-ing the accreted Superior continent. To the west of the mainteleseismic arrays, the Kapuskasing Structural Zone (KSZ)transects the greenstone belts of the Abitibi and Wawasubprovinces of the Superior Province. The northeast-

trending structure represents a section of high-gradeArchean lower crust, thrust southeastward during the EarlyProterozoic (1.95–1.85 Ma) (Percival and West 1994).

The southern half of the study area includes the GrenvilleProvince, which comprises the youngest terrains of the Pre-cambrian Canadian Shield. It was accreted to the SuperiorProvince during the Grenvillian orogenic cycle, between1160 and 970 Ma (Rivers et al. 1989). The Grenville and Su-perior provinces are separated by the northeast–southwest-trending Grenville Front (GF), a major tectonic zone wherethe Grenville rocks were thrust over the Superior. TheGrenville is divided into three main belts according to theage of the rocks, their origin, and the episodes of deforma-tion they have experienced: the Parautochthonous Belt, theAllochthonous Polycyclic Belt, and the AllochthonousMonocyclic Belt (Rivers et al. 1989). The 1994 and 1996teleseismic stations located in the Grenville Province covermainly the Parautochthonous and Allochthonous belts. Thesouthernmost stations sit on Paleozoic sedimentary rocks,which were deposited unconformably on basement rocks ofthe Grenville Province.

Finally, Late Jurassic to Early Cretaceous diamond-bearing kimberlites have been found at two locations in thestudy area: the “Rapide des Quinze” kimberlites (maximumage of 126 Ma; Ji et al. 1996), and the Kirkland Lakekimberlite field (age ranging between 147 and 158 Ma;Meyer et al. 1994, and references therein).

Geophysical coverageIn the past decade, the study area has been the locus of in-

tensive geophysical surveys, as part of the Abitibi–GrenvilleLithoprobe transect. Seismic reflection profiles have shedsome light on the tectonic evolution of the crust and the up-permost mantle (Green et al. 1990; Jackson et al. 1990;Bellefleur et al. 1995; Lacroix and Sawyer 1995; Ludden etal. 1993; Kellett et al. 1994; Calvert et al. 1995). Seismic re-fraction profiles shot over the same regions have providedwell-constrained velocity models (Grandjean et al. 1995;Winardhi and Mereu 1997). The aforementioned studies re-veal variations in Moho depth along the corridor traversedby the teleseismic array. A thinning of the crust occurs nearthe latitude of the Grenville Front, where the thickness is32–34 km, relative to 39–43 km in the Grenville Provinceand 34–36 km in the Superior Province (Winardhi andMereu 1997; Kellett et al. 1994; Grandjean et al. 1995;Martignole and Calvert 1996).

A number of magnetotelluric (MT) stations were deployedalong the Abitibi–Grenville transect to obtain electrical con-ductivity models of the crust and upper mantle (Kellett et al.1992, 1994; Mareschal et al. 1995). One important result tocome out of the MT study is the discovery of deep-seatedelectrical anisotropy beneath the southeastern CanadianShield. This anisotropy was first detected below the Pontiacsubprovince by Kellett et al. (1992), who interpreted it as adeep crustal feature. Later studies helped constrain the pres-ence of electrical anisotropy, relocating it within the uppermantle, between 50 and 150 km depth (Kellett et al. 1994;Mareschal et al. 1995; Sénéchal et al. 1996). The source ofthe electrical anisotropy is interpreted by Mareschal et al.(1992, 1995) as being due to interconnected films of graph-ite deposited along grain boundaries. Mareschal et al. (1995)

© 2000 NRC Canada

416 Can. J. Earth Sci. Vol. 37, 2000

© 2000 NRC Canada

Rondenay et al. 417

Fig. 1. Simplified geotectonic map of the study area with the location of teleseismic and magnetotelluric (MT) stations. The upper in-set shows the location of the study area with respect to Canada, and the lower inset focuses on the area of the 1994 experiment. ABI,Abitibi; CNSN, Canadian National Seismograph Network; SOSN, Southern Ontario Seismic Network.

© 2000 NRC Canada


(a) (b)

(c)

Fig. 2. Azimuths and amplitudes of electrical anisotropy (a) and shear-wave splitting calculated on the 1994 (b) and 1996 (c) datasetsbeneath the study area. For electrical measurements, the solid bar indicates the most conductive direction, and its length is proportionalto the phase difference between the most conductive and orthogonal directions. For teleseismic data, the solid bar indicates the polar-ization direction of the fast shear wave, and its length is proportional to the delay time between the two split waves.

extended the observation of electrical anisotropy to a largerarea of the Canadian Shield, including the Grenville and Su-perior (Pontiac, Abitibi, KSZ, Wawa) provinces. The direc-tion of maximum conductivity is generally east–west andparallel to the lineations defined by the major deformationzones that have affected the area during Precambrian time.Along the corridor traversed by the teleseismic arrays, theelectrical anisotropy shows a fairly uniform orientation ofN80°E (Kellett et al. 1994; Sénéchal et al. 1996). These re-sults are presented in Fig. 2a.

Gravity and heat-flow surveys (Antonuk and Mareschal1992; Guillou et al. 1994) complete the main geophysicalcoverage of the study area. The results from these methodshelp constrain and improve the tectonic interpretations ob-tained from other methods by providing information oncrustal composition and thickness (through modelling). Re-sults show a general increase of Bouguer gravity anomalyand crustal heat flux from east (Grenville) to west(Kapuskasing), whereas the mantle heat flux stays relativelyconstant across the different provinces and subprovinces ofthe southeastern Canadian Shield (10–14 mW·m–2; Guillouet al. 1994).

Finally, at the continental scale,S-velocity models showthe Canadian Shield to be characterized by a thick litho-sphere (i.e., lithospheric root), extending to depths of 200–250 km beneath the Abitibi–Grenville transect (Grand 1994;Van der Lee and Nolet 1997).

Teleseismic projects

Plans for a teleseismic experiment to investigate mantlestructure below the Abitibi–Grenville Lithoprobe transectwere originally motivated by the discovery in the area ofdeep-seated upper mantle electrical anisotropy from MTmeasurements (see previous section). The teleseismic projectwas designed as a two-phase experiment, performed in 1994and 1996.

The Abitibi-94 experimentThe first phase of data acquisition took place from July to

November 1994, using 10 short-period seismographs fromthe French Lithoscope program. Two types of recording in-struments were used, namely HADES (25 Hz) and TITAN(31.5 Hz) data-acquisition systems. All stations wereequipped with portable, short-period, three-componentLennartz 3D/5s seismometers, with a natural frequency of0.2 Hz (5 s), permitting recovery of frequencies between 0.2and 12.5 Hz. Stations were deployed along a north-northwest–south-southeast line crossing the Grenville Front(Fig. 1), with an average spacing of-25 km. The array cov-ered parts of the Pontiac subprovince and the northernGrenville Province (Fig. 1).

After installation, the teleseismic stations were servicedonce every 3 weeks, for collection of the data, synchroniza-tion of the HADES recorders internal clocks, and trouble-shooting for technical problems. The recovered dataset wasprocessed into a “workable” format at École polytechniquede Montreal and Laboratoire de geophysique interne et detectonophysique.

The Abitibi-96 experimentThe second phase of the Abitibi–Grenville Teleseismic

Project took place between May and November of 1996.This second deployment was cofunded by Lithoprobe, theU.S. National Science Foundation, and the Geological Sur-vey of Canada. The equipment (seismometers, recorders, andwork stations) was provided by the Program for the ArraySeismic Studies of the Continental Lithosphere of the Incor-porated Research Institutions for Seismology (IRIS), basedat the Lamont Doherty Earth Observatory of Columbia Uni-versity (Palisades, N.Y.).

This array, which overlapped with the Abitibi-94 line, in-cluded a total of 28 stations (Fig. 1). A relatively small sta-tion spacing of-20 km was selected to facilitate work onconverted-phase profiling and local shear-wave splittinganalysis. Two stations were deployed west of the main lineto provide additional spatial coverage of two enigmatic geo-logical features, the Kirkland Lake kimberlites and theKapuskasing uplift.

The instrumentation used in the Abitibi-96 array consistedof RefTek data-acquisition systems and Streckeisen STS-2portable, three-component, broad-band seismometers. Theseseismometers have a flat response to ground velocity from0.0083 to 50 Hz (120–0.02 s). The data were recorded usinga 20 Hz sample rate, permitting recovery of frequencies be-tween 0.0083 and 10 Hz.

The entire array of stations was serviced approximatelyonce every 4 weeks for maintenance of the instruments andcollection of the data. After servicing, the data were broughtback to headquarters in Rouyn-Noranda, Quebec, where theywere preprocessed and backed up on multiple devices. Inten-sive quality control was performed on the recovered data todetect possible internal problems with the instruments in thefield. The data were then sent to the IRIS – Data Manage-ment Center (DMC), for final processing and archival.

The following sections present the processing techniquesemployed on the two teleseismic datasets and results ob-tained to date.

Shear-wave splitting analysis

Shear-wave splitting analysis is a technique that is widelyapplied to teleseismic data. It provides information on theazimuthal anisotropy of the upper mantle, which is generallyinterpreted as due to the lattice preferred orientation (LPO)of olivine (Babuška and Cara 1991; Vinnik et al. 1986; Sil-ver and Chan 1988, 1991). The exact location of the aniso-tropy within the upper mantle is poorly constrained. In thecase of stable continental regions, like the Canadian Shield,some researchers interpret it to reside primarily within thelithosphere (Silver and Chan 1988, 1991), whereas otherssuggest it is located mainly in the flowing asthenosphere(Vinnik et al. 1992, 1995). Physically, the anisotropy causesa plane-polarized, vertically propagating shear wave to splitinto two waves, which are orthogonally polarized along thetwo main axes of anisotropy, and travel at different veloci-ties. The splitting is generally parameterized by two quanti-ties: (i) a time shift, which depends on the thickness of theanisotropic layer and the strength of the anisotropy; and(ii ) the polarization direction of the fast wave, which charac-terizes the orientation of the anisotropy. Various methods

© 2000 NRC Canada

Rondenay et al. 419

have been developed to recover these parameters using in-verse operators (involving rotation and time shifting) to re-produce a single polarized wave from the two split wavesrecorded at the surface (e.g., Vinnik et al. 1986; Bowmanand Ando 1987; Silver and Chan 1991). These calculationsgenerally assume the existence of a single horizontal and ho-mogeneous layer of anisotropy. Such an assumption may besimplistic for an area like the Canadian Shield, which hasbeen subjected to numerous episodes of deformation and re-working, but it is often necessary due to limited back-azimuthal coverage in the data.

Shear-wave splitting analysis was the first method em-ployed on the Abitibi-94 dataset. The results that were ob-tained and their interpretation are presented in Sénéchal etal. (1996), Ji et al. (1996), and Rondenay (1996). Splittingparameters were calculated independently for a series ofevents and phases (S, SKS, SKKS, ScS). The average valuesfor each station are presented in Fig. 2b. For the entire array,the average direction of fast polarization is N101°E ± 10°,and the average splitting is 1.46 ± 0.21 s. Using this timeshift and anS-velocity anisotropy of 3.2% inferred from lo-cal xenoliths samples, the thickness of the anisotropic layeris estimated at approximately 200 km (Ji et al. 1996). Overthe study area, the direction of seismic anisotropy correlateswell (within a consistent bias of +20°) with the calculateddirection of electrical anisotropy (N80°E; see section titledGeophysical coverage). Given the depth of the electrical ani-sotropy (50–150 km; see section titled Geophysical cover-age), a possible link between the two anisotropies would

constrain the location of the seismic anisotropy within thethick lithosphere present below the area (Sénéchal et al.1996), which is consistent with the interpretation of Silverand Chan (1988, 1991). The two kinds of anisotropy areboth interpreted to manifest the same horizontal stress fieldthat originally caused the now fossilized deformation in thesubcrustal lithosphere (Sénéchal et al. 1996). This deforma-tion resulted in alignment of a grain boundary conductivephase (source of electrical anisotropy) and LPO of olivine(source of seismic anisotropy). The different thickness esti-mates of the anisotropic layers obtained with electrical(100 km) and teleseismic (200 km) methods can be recon-ciled if the stability field of the conductive phase extends toa maximum depth of approximately 150 km (e.g., graphite;Kennedy and Kennedy 1976). The seismic anisotropy canthen extend to greater depths within the thick cratonic litho-sphere present beneath the study area.

Although Sénéchal et al. (1996) did not consider the con-sistent-20° obliquity observed between the two anisotropies(see Figs. 2a, 2b), it was the basis of the interpretation pre-sented in Ji et al. (1996). Xenolith samples from the Rapidedes Quinze locality present a similar obliquity between theLPO of olivine, which is responsible for the seismic aniso-tropy, and its shape-preferred orientation (SPO), which canbe associated with the electrical anisotropy; this obliquity isdue to finite noncoaxial strain (Ji et al. 1996). Since MT dataand shear-wave splitting analysis provide a comparable char-acterization of horizontal strain-induced anisotropy withinthe upper mantle, Ji et al. (1996) speculate that the obliquitybetween seismic and electrical anisotropy can be employedto indicate the movement of transcurrent shear zones in themantle. Thus, the 20° separation between the twoanisotropies beneath the Abitibi–Grenville may reflect adextral shear sense in the mantle, which is in agreement withthe last cycle of regional deformation inferred from the sur-face geology (Ji et al. 1996, and references therein). This in-terpretation further supports the location of seismicanisotropy within the lithospheric mantle.

The data from the Abitibi-96 deployment are currently be-ing analyzed for shear-wave splitting. Parameters have beenobtained using a multiple splitting method (Vinnik et al.1989; Wolfe and Silver 1998), in which a suite of events areconsidered simultaneously for each station. This methodprovides results which are more robust than those obtainedwith single events. The event coverage available for the pe-riod of the experiment is relatively poor. In fact, only sevenevents are suitable forSKS splitting analysis, and all areclustered within the small back-azimuthal range of 285–335°(i.e., mainly northwest). Preliminary results from the analy-sis of SKSphases are presented in Fig. 2c) (results fromSandSKKSphases obtained to date are of poorer quality andare not shown here). The average direction of fast polariza-tion for all stations is N93°E ± 18°, which corresponds tothe same general orientation as the main belts and deforma-tion zones traversed by the array. We observe a progressiverotation of the fast axis from east-northeast at the northernend of the array to east-southeast in the southern portion ofthe line. This rotation may imply a slight difference in direc-tion for the last major deformations to have affected north-ern Abitibi with respect to southern Abitibi, the Pontiacsubprovince, and the Grenville Province. The splitting delay

© 2000 NRC Canada


Fig. 3. Observed (solid line) and synthetic (broken line) receiverfunctions for sites AB01 and AB05. The synthetic receiver func-tions were calculated using velocity models represented inFig. 4.

times vary between 0.25 and 1.05 s over the array, with anaverage of 0.57 ± 0.22 s. This broad range of delay times isdifficult to interpret in terms of a single uniform anisotropiclayer, since it would imply large variations of the layer’sthickness (35–150 km, with 3.25%S-wave anisotropy andignoring any crustal contribution) over small lateral dis-tances. It is more likely that these variations are related to apossibly depth-dependent, laterally varying azimuthal aniso-tropy, which cannot be fully sampled by the limited back-azimuthal coverage of the data. Along the section where thetwo arrays overlap, the estimates of fast polarization direc-tions are consistent within errors, whereas the Abitibi-96 de-lay times are smaller than those of Abitibi-94. Thisdiscrepancy can be explained by the incorporation of a ma-jority of S-wave (vs. SKS) observations within the 1994dataset, which may have suffered source-side contamination.This has a more serious effect on delay time measurementthan on fast axis orientation (Saltzer et al. 1998).

Receiver function analysis

Receiver function analysis (Langston 1979; Owens et al.1984; Cassidy 1995) has been applied to the Abitibi-94dataset to identifyS-wave velocity discontinuities in thelithosphere beneath the array of three-component stations(H. Wu and D.J. White, in preparation). The analysis com-prised three steps: (i) deconvolution of the vertical compo-nent of motion from the radial and transverse components toobtain the radial and transverse receiver functions, (ii ) sum-mation of individual receiver functions at a given station to

enhance the signal-to-noise ratio, and (iii ) modelling of theresultant receiver function wave forms to obtain anS-wavevelocity model. The first step eliminates the earthquakesource and instrument response effects so that the receiverfunctions consist primarily ofP- to S-converted arrivalswhich originate atS-wave velocity contrasts beneath the ar-ray.

Trial and error simulation of the observed receiver func-tions was used to model theS-wave velocity structure. Anexample of observed and calculated receiver functions isshown in Fig. 3. Synthetic seismograms were calculated forS-wave models using a standard propagator matrix tech-nique. It is well known that the inversion of receiver func-tions for S-wave velocity structure is highly non-unique(e.g., Ammon et al. 1990), as observedP–S conversions in-dicate the presence ofS-wave velocity contrasts within thelithosphere, but provides little absolute velocity informationin the absence of simplifying assumptions. To reduce thenon-uniqueness in modelling the receiver functions, an ini-tial S-wave model was calculated from an existingP-wavemodel (assuming a Poisson’s ratio of 0.25) determined froma coincident seismic refraction analysis (Wu and White, inpreparation). The constraints applied in the modelling pro-cess were that theS-wave model (i) must be as close as pos-sible to the starting model, (ii ) produces synthetic responsesconsistent with the observed receiver functions, and (iii ) pro-vides a depth to the Moho which is consistent with the seis-mic refraction results. Average values of Poisson’s ratio forthe crust along the profile were determined using the averagecrustal P-wave velocity from the refraction model and the

© 2000 NRC Canada

Rondenay et al. 421

Fig. 4. Average two-dimensionalS-wave velocity model obtained from receiver function analysis on the 1994 teleseismic dataset. Ve-locity models from the individual stations have been averaged to emphasize velocity boundaries which are consistently observed acrossthe array. Shown are the receiver function velocity model (solid line) and the average initial velocity model calculated from the refrac-tion P-wave velocity model (dashed line). The dotted lines represent ±1 standard deviation of the average values.

average crustalS-wave velocity from the receiver functionmodel. The depth of the receiver function models was lim-ited to 120 km below the surface to avoid possible misinter-pretation of multiple reverberations which are dominant atgreater depths.

The main characteristics of the resulting velocity modelare shown in Fig. 4. The following observations can bemade: (i) significant velocity contrasts occurring at the topand base of the lower crust (-25 km and-34–43 km, re-spectively) correspond to layer boundaries observed in therefraction model, whereas the deeper boundary at-85 kmdepth provides new evidence for deep lithospheric layeringin this region; (ii ) an average crustal value of Poisson’s ratioof 0.263 is required to match the third constraint (see ear-lier), and the crustal value of Poisson’s ratio increases from0.250 beneath the Grenville Front to a maximum of 0.275 atthe south end of the array within the Grenville Province; and(iii ) lower crustal Poisson’s ratio values of 0.28–0.30 weredetermined, which in combination with lower crustalP-wavevelocities of 6.9–7.2 km/s from refraction modelling(Grandjean et al. 1995; Winardhi and Mereu 1997) are con-sistent with a lower crust of anorthositic and (or) maficgneiss composition (cf. Tables 1–3 in Holbrook et al. 1992).The southward increase in average crustal Poisson’s ratiofrom the Grenville Front into the Grenville Province is likelyassociated with the increasing thickness of the high-velocity,high Poisson’s ratio lower crustal layer.

Traveltime inversion

Traveltimes of teleseismic waves can be used to image lat-eral variations in mantle velocity structure beneath a seismicarray. Traveltime inversion, as the technique is known, relieson the ray theory approximation and is formulated as a lin-ear relation between measured delay times and perturbationsto a mantle velocity model. The implementation used herewas developed by VanDecar (1991) and involves cross-correlation of all pairs of wave forms for a given event to ex-tract accurate relative delay times (VanDecar and Crosson1990), followed by inversion of delay times for velocity per-turbations with respect to the one-dimensional (1D)iasp91model (Kennett and Engdahl 1991).

Of the two datasets, only the Abitibi-96 records are ame-nable to traveltime inversion, since the technique requiresaccurate timing. Traveltimes from Abitibi-96 were aug-mented by data from two broadband stations, GAC andSADO, from the Canadian National Seismograph Network(CNSN, Geological Survey of Canada) and six short-periodseismographs from the Southern Ontario Seismic Network(SOSN, University of Southern Ontario). The inversions areperformed independently onP- and S-wave relative arrivaltimes for the model volume shown in Fig. 5a. Figure 5bshows a preliminary result, in the form of a horizontal slicethrough the preferredP-velocity model at 200 km depth.This model was obtained with a dataset consisting of 2796

© 2000 NRC Canada


Fig. 5a. Traveltime inversion. Grid used for traveltime inversion, where highlighted volume represents the region of interest.

traveltime picks from 124 events (5.0≤ mb ≤ 6.6), providinga comprehensive back-azimuthal coverage. The colour scalegives the slowness (reciprocal of velocity) anomaly in termsof percent deviation from the 1Diasp91model. Regions ofpoor ray coverage (cells intersected by less than four rays)are shown in black. The most prominent feature to appear onthis depth slice is an elongate, low-velocity anomaly(−1.0%) that crosses the southern portion of the main line atlatitude 46°N. This low-velocity corridor is well defined be-tween depths of 50 and 300 km. We also note that the anom-aly is flanked on both sides by regions of relatively highseismic velocity. The 85 km discontinuity suggested by the

receiver function analysis cannot be observed in the travel-time inversion model, since the technique is not sensitive to1D horizontal layering. Synthetic tests (not presented here)indicate that model resolution is good from 30 to 450 kmdepth, within the areas of dense ray coverage (four or morerays per cell). Close to the surface, the area of dense raycoverage is restricted to a narrow band centred on the array.With depth, it widens substantially where there is a reason-able event distribution in back-azimuth and epicentral dis-tance. Three factors may contribute to the presence of a low-velocity corridor in the lithosphere: (i) thermal contrast,(ii ) compositional contrast, and (iii ) anisotropy.

© 2000 NRC Canada

Rondenay et al. 423

Fig. 5b. Traveltime inversion. Horizontal slice through the preferred model (within the region of interest) at 200 km depth. Colourscale gives theP-slowness anomaly with respect to the 1Diasp91model (Kennett and Engdahl 1991). Note the west-northwest-strikinglow-velocity corridor crossing the teleseismic array at latitude 46°N. The inset shows the location of the Monteregian (M) – WhiteMountains (WM) – New England Seamounts (NES) magma series with respect to the study area.

At the “local” scale of the experiment, the anomaly strikesobliquely to the trend of major surface structures (e.g.,Grenville Front). Thus, it appears to be related to a sub-crustal process that has not significantly affected the surfacegeology. It is in considering the larger scale of northeasternNorth America that the low-velocity corridor appears in anew light, as it coincides with the northwestward extrapola-tion of the trend defining the Monteregian igneous province.The Monteregian Hills are a series of Cretaceous alkaline in-trusions that have been associated with the White MountainsCretaceous intrusions and the New England Seamounts (seeinset in Fig. 5b). Two candidate hypotheses have been sug-gested for the emplacement of these formations: (i) the pas-sage of the North American plate over a fixed mantle plume(Foster and Symons 1979; Crough et al. 1980; Crough 1981;Duncan 1984; Foland et al. 1986; Sleep 1990; Adams andBasham 1991), and (ii ) tension-induced rifting of the conti-nental lithosphere associated with the opening of the NorthAtlantic (Sykes 1978; McHone and Butler 1984; McHone etal. 1987; Bédard 1985). The extension of this magmatictrack to the northwest, beneath the Canadian Shield, hasbeen suggested by proponents of both hypotheses on the ba-sis of the spatio-temporal proximity of the Kirkland Lakeand Rapide des Quinze kimberlites (Crough et al. 1980), andthe seismicity associated with this trend (Sykes 1978; Ad-ams and Basham 1991). Hence, the low-velocity corridor isinterpreted, at this point of the study, as a zone of contrast-ing thermal–compositional–anisotropic properties related toa Cretaceous event that affected the deep lithosphere of theCanadian Shield down to depths of 300 km. This same Cre-taceous event is responsible for the emplacement of theMonteregian – White Mountains – New England Seamountsmagmatic series and possibly for the eruption of theKirkland Lake and Rapide des Quinze kimberlites. The in-terpretation of the high-velocity anomalies flanking the cor-ridor are, at this point, highly speculative. They could eitherrepresent zones of depleted residuum produced by mantlemelting (Saltzer and Humphreys 1997) or transitions in man-tle fabric (i.e., anisotropy) related to the Cretaceous processdiscussed above. Subsequent analysis of the data and deriva-tion of improved models will help constrain the location andinterpretation of the low-velocity corridor.

Conclusions

This review of teleseismic studies along the Abitibi–Grenville Lithoprobe transect is, in part, a progress report, asseveral studies are still underway at the time of writing(April 1998). Results from completed studies provide evi-dence for an anisotropic lithosphere. Indeed, the anisotropyparameters calculated from the Abitibi-94 data were juxta-posed with EM measurements and xenolith data to deter-mine (i) the probable location of the anisotropic layer withinthe lithosphere, (ii ) an estimated thickness of 200 km for theanisotropic layer, and (iii ) a dextral mantle shear strain re-lated to the last cycle of regional deformation in the area.Shear-wave splitting results obtained to date on the Abitibi-96 dataset suggest, however, that the single uniformanistropic layer assumption may not be appropriate for theAbitibi–Grenville area of the Canadian Shield. Future workwill be dedicated to identifying an anisotropic signature

from the 85 km deep mantle layer inferred from receiverfunctions. This may allow for more detailed modelling ofSKSsplitting.

Traveltime inversion provides a seismic velocity model ofthe upper mantle from teleseismic data recorded along theAbitibi-96 array. The preliminaryP-wave velocity modelshows the presence of a low-velocity corridor which, givenits alignment, may be associated with the same process re-sponsible for the emplacement of the CretaceousMonteregian – White Mountains – New England Seamountsmagmatic series.

Completed and ongoing teleseismic studies along theAbitibi–Grenville Lithoprobe transect reveal a mantle litho-sphere that is both anisotropic and laterally heterogeneous.Both of these characteristics reflect mechanisms which havecontributed to lithospheric evolution in the region, possiblyover a wide interval of time. In fact, the anisotropy appearsto be associated with the last cycle of regional deformationto have affected the area, during the Archean and (or) EarlyProterozoic. The low-velocity corridor appears to be associ-ated with a much more recent event, which has not affectedthe surface nor even the crust, but only the deep lithosphericroot present beneath the study area. Ongoing work is dedi-cated to unveiling in greater detail the nature of these pro-cesses and their contribution to the stabilization andmodification of the Canadian Shield.

Acknowledgments

The authors acknowledge everyone at IRIS Program forthe Array Seismic Studies of the Continental Lithosphere(Lamont Doherty Earth Observatory) and DMC, G. Poupinetat Laboratoire de geophysique interne et de tectonophysique,the Geological Survey of Canada, J. Goutier at Ministère desRessources naturelles du Quebec, and graduate students atÉcole polytechnique de Montréal and New Mexico StateUniversity for making the field experiments possible. Wealso thank J. Guilbert for his shear-wave splitting code;R. Mereu and B. Dunn (University of Western Ontario) foraccess to the SOSN data; T. Owens, C. Ammon, andJ. Cassidy for their receiver function analysis code;J. VanDecar for his traveltime inversion code; C-G Bank forhis general help; and J. Cassidy and M. Granet for construc-tive reviews. The authors made use of the Generic MappingTools (GMT; Wessel and Smith 1991) for some of the fig-ures. This work was supported by scholarships and grantsfrom the Natural Sciences and Engineering Research Coun-cil of Canada, the Fonds pour la formation de chercheurs etl’aide à la recherche du Quebec, and the U.S. National Sci-ence Foundation (Division of International Programs, grant9322499).

References

Adams, J., and Basham, P. 1991. The seismicity and seismo-tectonics of eastern Canada.In Neotectonics of North America.Edited by D.B. Slemmons, E.R. Engdahl, M.D. Zoback, andD.D. Blackwell. Geological Society of America, Boulder, Colo.,pp. 261–276.

Ammon, C.J., Randall, G.E., and Zandt, G. 1990. On thenonuniqueness of receiver function inversions. Journal of Geo-physical Research,95: 15 303 – 15 318.

© 2000 NRC Canada


Antonuk, C.N., and Mareschal, J.C. 1992. Preliminary gravitymodelling along the LITHOPROBE seismic reflection profiles28 and 29, Northern Abitibi Subprovince.In Proceedings,Lithoprobe Abitibi–Grenville Transect, Report 33, pp. 71–75.

Babuška, V., and Cara, M. 1991. Seismic anisotropy in the Earth.Kluwer Academic Publishers, Dordrecht, The Netherlands.

Bédard, J.H. 1985. The opening of the Atlantic, the Mesozoic NewEngland Igneous Province, and mechanisms of continentalbreakup. Tectonophysics,113: 209–232.

Bellefleur, G., Barnes, A., Calvert, A., Hubert, C., and Mareschal,M. 1995. Seismic reflection constraints from Lithoprobe line 29on the upper crustal structure of the northern Abitibi greenstonebelt. Canadian Journal of Earth Sciences,32: 128–134.

Benn, K., Sawyer, E.W., and Bouchez, J.-L. 1992. Orogen paralleland transverse shearing in the Opatica belt, Quebec: implica-tions for the structure of the Abitibi Subprovince. CanadianJournal of Earth Sciences,29: 2429–2444.

Bowman, J.R., and Ando, M. 1987. Shear-wave splitting in the up-per-mantle wedge above the Tonga subduction zone. Geophysi-cal Journal of the Royal Astronomical Society,88: 25–41.

Calvert, A.J., Sawyer, E.W., Davis, W.J., and Ludden, J.N. 1995.Archaean subduction inferred from seismic images of a mantlesuture in the Superior Province. Nature (London),375: 670–673.

Card, K.D. 1990. A review of the Superior Province of the Cana-dian Shield, a product of Archean accretion. Precambrian Re-search,48: 99–156.

Card, K.D., and Ciesielski, A. 1986. DNAG No. 1. Subdivisions ofthe Superior Province of the Canadian Shield. Geoscience Can-ada,13: 5–13.

Cassidy, J.F. 1995. Review: Receiver function studies in the south-ern Canadian Cordillera. Canadian Journal of Earth Sciences,32: 1514–1519.

Crough, S.T. 1981. Mesozoic hotspot epeirogeny in eastern NorthAmerica. Geology,9: 2–6.

Crough, S.T., Morgan, W.J., and Hargraves, R.B. 1980.Kimberlites: their relation to mantle hotspots. Earth and Plane-tary Science Letters,50: 260–274.

Desrochers, J.-P., Hubert, C., Ludden, J.N., and Pilote, P. 1993.Accretion of Archean oceanic plateau fragments in the Abitibigreenstone belt, Canada. Geology,21: 451–454.

Duncan, R.A. 1984. Age progressive volcanism in the New Eng-land Seamounts and the opening of the Central Atlantic Ocean.Journal of Geophysical Research,89(B12): 9980–9990.

Foland, K.A., Gilbert, L.A., Sebring, C.A., and Chen, J.-F. 1986.40Ar/39Ar ages for plutons of the Monteregian Hills, Quebec:evidence for a single episode of Cretaceous magmatism. Geo-logical Society of America Bulletin,97: 966–974.

Foster, J., and Symons, D.T.A. 1979. Defining a paleomagnetic po-larity pattern in the Monteregian intrusives. Canadian Journal ofEarth Sciences,16: 1716–1725.

Grand, S.P. 1994. Mantle shear structure beneath the Americas andsurrounding oceans. Journal of Geophysical Research,99(B6):11 591 – 11 621.

Grandjean, G., Wu, H., White, D., Mareschal, M., and Hubert, C.1995. Crustal velocity models for the Archean Abitibigreenstone belt from seismic refraction data. Canadian Journalof Earth Sciences,32: 149–166.

Green, A.G., Milkereit, B., Mayrand, L.J., Ludden, J.N., Hubert,C., Jackson, S.L., Sutcliffe, R.H., West, G.F., Verpaelst, P., andSimard, A. 1990. Deep structure of an Archean greenstoneterrane. Nature (London),344: 327–330.

Guillou, L., Mareschal, J.-C., Jaupart, C., Gariepy, C., Bienfait, G.,and Lapointe, R. 1994. Heat flow, gravity and structure of the

Abitibi belt, Superior Province, Canada: implications for mantleheat flow. Earth and Planetary Science Letters,122: 103–123.

Holbrook, W.S., Mooney, W.D., and Christensen, N.I. 1992. Theseismic velocity structure of the deep continental crust.In Con-tinental lower crust.Edited byD.M. Fountain, R. Arculus, andR.W. Kay. Elsevier, Amsterdam, pp. 1–34.

Jackson, S.L., Sutcliffe, R.H., Ludden, J.N., Hubert, C., Green,A.G., Milkereit, B., Mayrand, L., West, G.F., and Verpaelst, P.1990. Southern Abitibi greenstone belt: Archean crustal struc-ture from seismic-reflection profiles. Geology,18: 1086–1090.

Ji, S., Rondenay, S., Mareschal, M., and Sénéchal, G. 1996. Obliq-uity between seismic and electrical anisotropies as a potentialindicator of movement sense for ductile shear zones in the uppermantle. Geology,24(11): 1033–1036.

Kellett, R.L., Mareschal, M., and Kurtz, R.D. 1992. A model oflower crustal electrical anisotropy for the Pontiac Subprovinceof the Canadian Shield. Geophysical Journal International,111:141–150.

Kellett, R.L., Barnes, A.E., and Rive, M. 1994. The deep structureof the Grenville Front: a new perspective from western Quebec.Canadian Journal of Earth Sciences,31: 282–292.

Kennedy, C.S., and Kennedy, G.C. 1976. The equilibrium bound-ary between graphite and diamond. Journal of Geophysical Re-search,81(14): 2467–2470.

Kennett, B.L.N., and Engdahl, E.R. 1991. Traveltimes for globalearthquake location and phase identification. Geophysical Jour-nal International,105: 429–465.

Kimura, G., Ludden, J.N., Desrochers, J.-P., and Hori, R. 1993. Amodel of ocean–crust accretion for the Superior province, Can-ada. Lithos,30: 337–355.

Lacroix, S., and Sawyer, E.W. 1995. An Archean fold–thrust beltin the northwestern Abitibi Greenstone Belt: structural and seis-mic evidence. Canadian Journal of Earth Sciences,32: 97–112.

Langston, C.A. 1979. Structure under Mount Rainier, Washington,inferred from teleseismic body waves. Journal of GeophysicalResearch,84(B9): 4749–4762.

Ludden, J., and Hubert, C. 1986. Geologic evolution of the LateArchean Abitibi greenstone belt of Canada. Geology,14: 707–711.

Ludden, J., Hubert, C., Barnes, A., Milkereit, B., and Sawyer, E.1993. A three-dimensional perspective of the evolution ofArchean crust: LITHOPROBE seismic reflection images in thesouthwestern Superior province. Lithos,30: 357–372.

Mareschal, M., Fyfe, W.S., Percival, J., and Chan, T. 1992. Grain-boundary graphite in Kapuskasing gneisses and implications forlower-crustal conductivity. Nature (London),357: 674–676.

Mareschal, M., Kellet, R.L., Kurtz, R.D., Ludden, J.N., Ji, S., andBailey, R.C. 1995. Archaean cratonic roots, mantle shear zonesand deep electrical anisotropy. Nature (London),375: 134–137.

Martignole, J., and Calvert, A.J. 1996. Crustal-scale shortening andextension across the Grenville Province of western Quebec. Tec-tonics,15: 376–386.

McHone, J.G., and Butler, J.R. 1984. Mesozoic igneous provincesof New England and the opening of the North Atlantic Ocean.Geological Society of America Bulletin,95: 757–765.

McHone, J.G., Ross, M.E., and Greenough, J.D. 1987. Mesozoicdyke swarms of Eastern North America.In Mafic dyke swarms.Edited byH.C. Halls and W.F. Fahrig. Geological Association ofCanada, Special Paper 34, pp. 279–288.

Meyer, H.O.A., Waldman, M.A., and Garwood, B.L. 1994. Mantlexenoliths from kimberlite near Kirkland Lake, Ontario. The Ca-nadian Mineralogist,32: 295–306.

Owens, T.J., Zandt, G., and Taylor, S.R. 1984. Seismic evidencefrom an ancient rift beneath the Cumberland Plateau, Tennessee:

© 2000 NRC Canada

Rondenay et al. 425

© 2000 NRC Canada


a detailed analysis of broadband teleseismicP waveforms. Jour-nal of Geophysical Research,89(B9): 7783–7795.

Percival, J.A., and West, G.F. 1994. The Kapuskasing uplift: a geo-logical and geophysical synthesis. Canadian Journal of EarthSciences,31: 1256–1286.

Rivers, T., Martignole, J., Gower, C.F., and Davidson, A. 1989.New tectonic divisions of the Grenville Province, southeast Ca-nadian Shield. Tectonics,8(1): 63–84.

Rondenay, S. 1996. Anisotropie sismique et electrique dense man-teau supérieur, au niveau du Front de Grenville (Canada).M.A.Sc. thesis, École Polytechnique de Montréal, Montréal.

Saltzer, R.L., and Humphreys, E.D. 1997. Upper mantleP wavevelocity structure of the eastern Snake River Plain and its rela-tionship to geodynamic models of the region. Journal of Geo-physical Research,102(B6): 11 829 – 11 841.

Saltzer, R.L., Jordan, T.H., and Gaherty, J.B. 1998. How are shear-wave splitting measurements affected by variations in the aniso-tropy with depth?In Proceedings of the 1998 Spring Meeting,American Geophysical Union, May 26–29, 1998, Boston, Mass.Program with Abstracts, p. S212.

Sawyer, E.W., and Benn, K. 1993. Structure of the high-gradeOpatica belt and adjacent low-grade Abitibi subprovince andArchaean mountain front. Journal of Structural Geology,15:1443–1458.

Sénéchal, G., Rondenay, S., Mareschal, M., Guilbert, J., andPoupinet, G. 1996. Seismic and electrical anisotropies in thelithosphere across the Grenville Front, Canada. Geophysical Re-search Letters,23(17): 2255–2258.

Silver, P.G., and Chan, W.W. 1988. Implications for continentalstructure and evolution from seismic anisotropy. Nature (Lon-don), 335: 34–39.

Silver, P.G., and Chan, W.W. 1991. Shear wave splitting and sub-continental mantle deformation. Journal of Geophysical Re-search,96(10): 16 429 – 16 454.

Sleep, N.H. 1990. Monteregian hotspot track: a long-lived mantleplume. Journal of Geophysical Research,95(B13): 21 983–21 990.

Sykes, L.R. 1978. Intraplate seismicity, reactivation of preexistingzones of weakness, alkaline magmatism, and other tectonismpostdating continental fragmentation. Reviews of Geophysicsand Space Physics,16(4): 621–688.

VanDecar, J.C. 1991. Upper-mantle structure of the Cascadiasubduction zone from non-linear teleseismic travel-time inver-sion. Ph.D. thesis, University of Washington, Seattle.

VanDecar, J.C., and Crosson, R.S. 1990. Determination of tele-seismic relative phase arrival times using multi-channel cross-correlation and least squares. Bulletin of the Seismological Soci-ety of America,80(1): 150–169.

Van der Lee, S., and Nolet, G. 1997. Upper mantleS velocitystructure of North America. Journal of Geophysical Research,102(B10): 22 815 – 22 838.

Vinnik, L.P., Kosarev, G.L., and Makeyeva, L.I. 1986. Azimuthalanisotropy of the Earth’s interior from observations of long-period body waves. Izvestiya Earth Physics,22(11): 955–960.

Vinnik, L.P., Farra, V., and Romanowicz, B. 1989. Azimuthal ani-sotropy in the Earth from observations of SKS at Geoscope andNARS broadband stations. Bulletin of the Seismological Societyof America,79: 1542–1558.

Vinnik, L.P., Makeyeva, L.I., Milev, A., and Usenko, A.Yu. 1992.Global patterns of azimuthal anisotropy and deformations in the con-tinental mantle. Geophysical Journal International,111: 433–447.

Vinnik, L.P., Green, R.W.E., and Nicolaysen, L.O. 1995. Recentdeformations of the deep continental root beneath southern Af-rica. Nature (London),375: 50–52.

Wessel, P., and Smith, W.H.F. 1991. Free software helps map anddisplay data. EOS, Transactions of the American GeophysicalUnion, 72: 441, 445–446.

Winardhi, S., and Mereu, R.F. 1997. Crustal velocity structure of theSuperior and Grenville provinces of the southeastern CanadianShield. Canadian Journal of Earth Sciences,34: 1167–1184.

Wolfe, C.J., and Silver, P.G. 1998. Seismic anisotropy of oceanicupper mantle: shear wave splitting methodologies and observa-tions. Journal of Geophysical Research,103(B1): 749–771.

teleseismic studies of the lithosphere below the abitibi ... · teleseismic studies of the...

Documents