• Aucun résultat trouvé

The formation of Laurentia: Evidence from shear wave splitting

N/A
N/A
Protected

Academic year: 2021

Partager "The formation of Laurentia: Evidence from shear wave splitting"

Copied!
9
0
0

Texte intégral

(1)

Contents lists available atScienceDirect

Earth

and

Planetary

Science

Letters

www.elsevier.com/locate/epsl

The

formation

of

Laurentia:

Evidence

from

shear

wave

splitting

Mitch

V. Liddell

a

,

,

Ian Bastow

a

,

Fiona Darbyshire

b

,

Amy Gilligan

c

,

Stephen Pugh

a

aImperialCollegeLondon,UnitedKingdom bUniversitéduQuébecàMontréal,Canada cUniversityofAberdeen,Scotland,UnitedKingdom

a

r

t

i

c

l

e

i

n

f

o

a

b

s

t

r

a

c

t

Articlehistory:

Received1April2017

Receivedinrevisedform5September2017 Accepted12September2017

Availableonline3October2017 Editor:P.Shearer

Keywords:

shearwavesplitting Paleoproterozoic seismicanisotropy Trans-Hudsonorogen Laurentia

cratons

The northern Hudson Bay regionin Canada comprisesseveralArchean cratonicnuclei, assembled by a number of Paleoproterozoic orogenies includingthe Trans-Hudson Orogen (THO)and the Rinkian– NagssugtoqidianOrogen.Recentdebatehasfocusedontheextenttowhichtheseorogenshavemodern analogues such as the Himalayan–Karakoram–TibetOrogen. Further,the structure ofthe lithospheric mantlebeneaththeHudsonStraitandsouthernBaffinIslandispotentiallyindicativeofPaleoproterozoic underthrusting of the Superior plate beneath the Churchill collage. Also in question is whether the Laurentiancratonicrootisstratified,withafast,depleted,Archeancoreunderlainbyaslower,younger, thermally-accretedlayer.Plate-scaleprocessthatcreatestructuressuchastheseareexpectedtomanifest asmeasurablefossilseismicanisotropicfabrics.Weinvestigatetheseproblemsviashearwavesplitting, and presentthemostcomprehensivestudytodateofmantleseismicanisotropyinnorthernLaurentia. Strong evidenceispresented formultiplelayersofanisotropy beneathArcheanzones,consistentwith theepisodicdevelopmentmodelofstratifiedcratonickeels.WealsoshowthatsouthernBaffinIslandis underlainbydippinganisotropicfabric,whereunderthrustingoftheSuperiorplatebeneaththeChurchill has previously been interpreted. This provides direct evidence of subduction-related deformation at 1.8 Ga,implyingthattheTHOdevelopedwithmodernplate-tectonicstyleinteractions.

©2017TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).

1. Overview

The geological record of the northern Hudson Bay region in Canada exceeds 2 billion years, including several Archean nuclei and a series of Paleoproterozoic orogens that culminated in the assembly of the cratonic core of North America, Laurentia. The largestof theseis theTrans-Hudson Orogen (THO), whichmarks the

1.8GacollisionbetweentheArcheanSuperiorcratonandthe Churchillplate(Fig. 1;Hoffman,1988).Structuraland thermobaro-metricstudies suggest the THO was similar inscale andstyle to themodern-dayHimalayan–Karakoram–TibetOrogen(HKTO)(e.g., St-Ongeetal., 2006),afindingcorroborated byseismicstudies of the crust (Thompson et al., 2010; Pawlak et al., 2011; Gilligan et al., 2016), and recently by the discovery of low-temperature, high-pressure eclogite rocks within the THO indicative of plate-scale subduction (Weller and St-Onge, 2017). Farther north are the remnants of the 1.7Ga Nagssugtoqidian Orogen (NO; Fig. 1) whichrecordsthecollisionoftheNorthAtlantic,Superior,andRae cratons with plate-scaledeformation distinct fromthat imparted

*

Correspondingauthor.

E-mailaddress:m.liddell14@imperial.ac.uk(M.V. Liddell).

from thenearly contemporaneous THO tothe south.Laurentia is alsocharacterisedbyoneofthedeepestlithosphericroots(‘keels’) on Earth,withthe lithosphericmantlereaching depths

>280

km in places (e.g., Bao and Eaton, 2015; Porritt et al., 2015). Re-cent debate has centred on whether the root is stratified, with a seismically fast, depleted, upper layer underlain by a younger, slower, thermal lithosphere (e.g., Yuan and Romanowicz, 2010), andwhethersuchalayer isrestrictedtoArcheandomainsor ex-tends beneath Proterozoic regions as well (e.g.,Darbyshire et al., 2013).

Increasingourknowledgeoftheseismicstructure ofthe Hud-son Bay region is central to our understanding of the assembly of Laurentia. Modern-style plate tectonics would have imparted measurable,plate-scale,seismically-anisotropicfossilfabricsinthe lithosphere. Forexample,plate-scaleunderthrusting ofthe Supe-rior lithospherebeneaththe Churchillplatecould be expectedto createdippinganisotropiclayerswithfastdirectionsperpendicular tothe directionofcollision,whilekeelstratificationshouldresult inmultiplelayersofanisotropy(e.g.,YuanandRomanowicz,2010; Darbyshireetal.,2013).

When a radially-polarised shear wave encounters seismically anisotropicmedia,itwillsplitintotwoorthogonalshearwaves po-larised alongthefastandslowaxesofthematerial.The splitting http://dx.doi.org/10.1016/j.epsl.2017.09.030

(2)

Fig. 1. GeologicalmapofnorthernHudsonBaywithSKSsplittingresults.Dashed yellowbarsindicatestackedφandδt values,solidyellowbarsindicateunstacked individualmeasurements.Redbarsarenullmeasurementswith90◦ambiguity. Ab-soluteplatemotionarrowsarepurple.Theinsetglobalmapshowsthelocationof thereceivernetwork,thereddotsareearthquakesusedinthisstudy.BaS,Baffin Suture;CI,CoatsIsland;HP,HallPeninsula;MI,Meta-Incognita;SI,Southampton Island;SRS,SoperRiverSuture;STZ,SnowbirdTectonicZone;THO,Trans-Hudson Orogen;NBTB,NortheastBaffinThrustBelt;NO,NagssugtoqidianOrogen;LL,Lynn LakeFault;MISH,Meta-Incognita-Sugluk-Hall-Peninsula.(Forinterpretationofthe referencestocolourinthisfigure,thereaderisreferredtothewebversionofthis article.)

parameters

φ

(the polarisation direction of the fast shear wave) and

δ

t (thedelay timebetweenthetwo waves)canthenbeused to characterise the crust and mantle anisotropy beneath the re-ceiver(SilverandChan,1991).

Herewe presenttheresultsofa shearwave splittingstudyof SKSand SKKS (hereafter referred to asSKS) phases recorded by 43seismograph stations in the northern HudsonBay region. Re-gionalSKSsplittingstudiesoftenutiliseonly1–2 yrofdata, limit-ingthebackazimuthalcoverageofhigh-qualitymeasurements,and precluding the possibility of interpreting more complicated dip-ping or multi-layer structures. However, our data set comprises stations withrecording timesbetween 4 and23 yr. The Hudson BayLithosphericExperiment(HuBLE;e.g.,Bastowetal.,2011) sta-tions inthe HudsonStrait andon BaffinIsland were active from 2007to2011.ThePortableObservatoriesforLithosphericAnalysis andResearchInvestigatingSeismicity(POLARIS;Eatonetal.,2005) network began deployment in 2004 in western Hudson Bayand northernQuebec.StationsFCCandFRBfromtheCanadianNational SeismographNetworkhavebeenactivesince1994.

2. Tectonicbackground

Ourstudyregioncontainslargeportionsof3Archeanprovinces (Rae,Hearne,Superior:Fig. 1)thatcomprisemuchoftheCanadian Shield (Hoffman, 1988). The nucleiof these Archean regions are thoughtto have originally grown by lateral accretion and wedg-ingofproto-continentsina pre-subductionsetting(Snyder etal., 2016). The Rae and Hearne are divided by the Snowbird Tec-tonic Zone (STZ), potentially a 1.9 Ga collision zone (Berman et

al., 2007), andtogether comprise thebulk of the Churchillplate. TheChurchill–Superiorcollisionisthoughttohavebeencomplex, trapping several smaller micro-platesbetween the principal cra-tons before terminal collision at 1.8 Ga (Corrigan et al., 2009). SouthernBaffinIslandhasbeenpostulatedtobeanamalgamation of some of these micro-continents, including the Meta-Incognita (MI)block,theSuglukblock,andtheHallPeninsulablock,together dubbed the ‘MISH’ block (Snyder et al., 2013). The BaffinSuture (St-Ongeetal., 2006) markstheboundarybetweenthesoutheast RaecratonandtheMeta-Incognita(MI)microcontinentthatmakes up much of southern Baffin Island. The northward trending fea-turesandrelativelyshallowburialdepthoftheregionindicatethat Meta-IncognitawastheupperplateintheRae-MIcollision (Corri-ganetal.,2009).

NorthernBaffinIslandincludesthewesternextensionof Green-land’s Paleoproterozoic Rinkian fold belt along the SE-NW ori-ented Northeast Baffin Thrust Belt(NBTB; Fig. 1), which exerted southwesterly pressure and strikes roughly perpendicular to the structural deformation patterns to the south, overprinting sev-eralArcheanandPaleoproterozoicprovinces(JacksonandBerman, 2000).Thisgenerallynorth–southstrikingfoldbelthasbeenlinked totheeast–westorientedplate-scaleNagssugtoqidianOrogen(NO) ofsouthern Greenland (e.g.,Connelly etal., 2006). The combined Rinkian–Nagssugtoqidian orogen issimilarly asymmetric, and po-tentiallysimilarinscale,totheTHOtothesouth.

3. SKSsplittingwithclusteranalysis

Seismicanisotropyreferstothedirectionaldependenceof seis-mic wavespeed. When a shear wave encounters an anisotropic medium,itwillsplitintotwoshearwaves,orthogonallypolarised, one travelling faster than the other (e.g.,Silver andChan, 1991). Thepolarisationdirection(φ)ofthefastshearwave,andthedelay time (δt) betweenthem can be used to characterise the seismic anisotropy of the material. P-to-S converted phasessuch as SKS andSKKS,areideally-suitedforupper-mantleshear-wavesplitting studies; they are radially polarised atthe core–mantle boundary and thus record no source-side anisotropy (e.g. Silver and Chan, 1991; Long andSilver, 2009). Olivine, themostcommon mineral in Earth’s upper mantle, is highly anisotropic. A crystallographic preferredorientation(CPO)maydevelopintheolivineinresponse to strain, withits a-axis aligned parallel to thedirection offlow (e.g.Bystrickyetal.,2000;Tommasietal.,2000;ZhangandKarato, 1995),assumingsteady-state,onedimensionalshearflow (Kamin-ski andRibe, 2002). The shear wave splitting parameters,

φ

and

δ

t, can therefore be related to pre-existing ‘fossil’ anisotropy in the lithosphere (e.g. Silver and Chan, 1991; Vauchez and Nico-las, 1991), mantle convection patterns (e.g, Vinnik et al., 1989), absolute platemotion directions (e.g, Debayle andRicard, 2013), alignedmelt/fluid(e.g.,BlackmanandKendall,1997),orany com-binationthereof.

WeinspectedSKSphasesforearthquakesofmb

6occurring at epicentral distances of

88◦ from 2004 to 2017. For perma-nent stationsFRB andFCC,our search extendedback to 1993.In total 5483event–station pairs were processed, and406 were in-cluded in the final dataset. Data were filtered prior to analysis using a zero-phase Butterworth bandpass filter with corner fre-quencies 0.04 and 0.3Hz. Splitting parameters were constrained usingthe semi-automatedapproachof Teanbyetal.(2004), built on the Silver and Chan (1991) method. Horizontal components are rotated and time-shifted to minimise the second eigenvalue ofthecovariancematrixforparticlemotionwithinatimewindow aroundtheshearwavearrival.Thisisequivalenttolinearisingthe particlemotionandminimisingtangentialcomponentshearwave energy.Iftheparticlemotionislinearisedinitiallythisiscalleda ‘null’measurementandindicatesthattheanisotropicfastdirection

(3)

Fig. 2. High-qualitysplittingmeasurementexamplefromstationILON.(a)TherecordedseismogramshowingtheSKSphaseandtheinitialwindow.(b)Theseismogram rotatedinto radialand tangentialcomponentsbothbefore(toptwo)andafter(bottomtwo)processing.(c)TopL–R:closeupofthe SKSphasesfor thefastandslow waveformsbeforecorrection,aftercorrection,andaftercorrectionwithoutnormalisedamplitudes.BottomL–R:particlemotionbeforeandaftercorrection.(d)Contourmap showingstabilityofthesplittingparameters.Linesindicateonestandarddeviation.Thethicklineindicates95percentconfidence.(e)Splittingparametervariationsasa functionofthechangingwindow.(f)Clusteranalysisresultsforφandδt foreachofthe100windows.Thesevalueswereverystableoverthefullrangeofwindows.

iseitherperpendicularorparalleltothebackazimuthofthewave, or that no anisotropic material was encountered. Null measure-mentsthereforehaveaninherent90◦ ambiguity,andarereported inFig. 1asaredcross.TheSilverandChan(1991)approachtakes asingle,manuallypicked,shear-waveanalysiswindow.Inthe clus-teranalysisapproachofTeanbyetal.(2004),however,thesplitting analysisis performed fora range ofwindow lengths andcluster analysisisutilisedtofindmeasurementsthatarestableovermany differentwindows.Allsplittingparameters weredeterminedafter analysisof 100different windows.An example ofthe analysisis showninFig. 2.

4. SKSsplittingresults

The resultsobtainedforSKSsplitting innorthern Canada dis-playedvisuallyinFig. 1aresummarisedinTable 1;afulllistofall splittingresultsandassociatederrorsisincludedinsupplemental materials.The averagesignaltonoise ratio(asdefinedby Restivo andHelffrich, 1999) forthesedata is15.5 witha standard devi-ation of7.1. Quality control was enforced by visual inspection of each split toensure linearisation ofthe particlemotion. Wealso enforced dataerror upperlimitsof

±

15◦ in

φ

and

±

0.5s in

δ

t,

although mostwere much lower (see supplemental materials for completedataset). Thisisamuch stricter limitthan theprevious study ofSnyder etal. (2013),where errors sometimes exceeded

±

30◦ in

φ

and2 sin

δ

t.Whererelevant,amodelclassisassigned to each stations to describes the first order type of anisotropy observedatthatstation (Table 1). Eachbasicclassofmodel (sin-glelayer,two-layer,dippinglayer) hasadistinctive backazimuthal pattern intheir splitting parameters. A dipping anisotropic layer willvaryrelativelysmoothlywith360◦periodicity.Ananisotropic model ofa vertical interface has 180◦ backazimuthalperiodicity, andatwo-layer modelhas90◦ periodicity(Fig. 3;Silverand Sav-age,1994). Adipping modelhaspeak-to-peak

φ

variations

90◦,

asopposedtointerfacesortwo-layer models,whichhavesharper changesin

φ

andpeak-to-peakvariationsapproaching180◦.

Themodel-classforeachstationinTable 1ischosentobethe simplestpossible that explainstheobservations.Formany ofthe stations in Fig. 1 a single, horizontalanisotropic layer could ad-equately explain the data. Data from such stations were stacked usingaprocedurebasedonthemethodofWolfeandSilver(1998) to obtain single pairs of splitting parameters (e.g., CTSN, DORN, MARN, inFig. 1; Table 1).Wecannot, however,preclude the pos-sibility that this assumption is invalid for stations where back-azimuthal coverage of earthquakes isinsufficient toresolve more complex dippingormulti-layerpatternsof anisotropy(e.g.,Silver and Savage,1994). Plotsof

φ

vsbackazimuth foreach station in Fig. 1 are in thesupplemental materials. For stationsthat exhib-itedclearvariationtypicalofamorecomplexanisotropicstructure, agridsearch ofrelevantmodelparameterswas performedto for-ward model the

φ

and

δ

t using the MSAT toolkit ofWalker and Wookey (2012), which is capable of modelling both multi-layer and dipping anisotropy.Bastow etal. (2011) suggested the pres-ence of dipping anisotropy based on backazimuthal variation of splitting parameters; we are the first to explore this hypothesis quantitatively.The‘N’columninTable 1compares thenumberof splitsusedtodefineeachstationinourdatasetdirectlytothatof Bastowetal.(2011),inbrackets.

5. Discussion

5.1. Causesofobservedanisotropy

Acrossour studyarea, we commonlyobserve

δ

t

1 s(Fig. 1, Table 1), implying amantle contributionto the observations: re-gionalcontinentalcrustis

40kmthickandcanonlyreasonably account for

δ

t

0.5 s(BarruolandMainprice,1993;Silver,1996). North Americanabsoluteplatemotion(APM)is

22mm/yr,well

(4)

Table 1

SKSsplitting parametersforstacked(top)andunstacked(bottom)stations.Subscriptvaluesrefertobackazimuthrangesforbinnedresults,orthebackazimuthofrecording fornullmeasurements.N isthenumberofsplits(includingnulls)forthatstationorbackazimuthalrange,bracketednumbersarethenumberofsplitsusedby

Bastow

etal. (2011)forthesamestation.Forunstackedresults,boldtypetextindicatesthefirst-ordermodelclassgroupsbasedon

Fig. 3

.

Station φ δt N Station φ δt N AKVQ 84±4 0.9±0.1 3 MANN 82±1 0.8±0.1 8 (2) AP3N 43±2 0.6±0.1 10 MARN −58±1 1.1±0.1 16 B1NU 47±8 0.5±0.1 3 MNGN −88±1 1.1±0.1 23 (11) CDKN −47±5 0.7±0.1 4 (2) MUMO 73±1 1.3±0.1 14 CHIN −49±3 1.6±0.1 3 NOTN 90±1 1.6±0.1 8 (6) CLRN 86±5 1.3±0.1 1 NUNN 84±3 0.9±0.1 3 (1) CMBN −20±1 1.3±0.2 2 (2) PINU −86±2 0.8±0.1 4 CTSN 63±1 0.8±0.1 13 (2) PNGN −71±2 0.7±0.1 6 (2) DORN −85±3 1.1±0.1 4 (3) QILN 67±1 1.2±0.1 6 (2) INUQ 71±2 0.6±0.1 5 SEDN 70±1 0.8±0.1 10 IVKQ −80±2 0.8±0.1 2 SHMN −77±3 0.7±0.1 5 (2) JENN 69±14 0.7±0.2 1 SILO 62±1 0.6±0.1 7 JOSN 59±5 1.0±0.1 2 SMLN 14±4 0.6±0.1 2 KASO 91±2 1.1±0.1 4 SRLN 44±1 0.7±0.1 5 KAJQ −71±8 0.3±0.1 2 STLN −9±2 1.0±0.1 3 KIMN −76±2 0.7±0.1 6 (3) VIMO 83±1 0.6±0.1 9 KJKQ 51±2 0.7±0.1 7 WAGN 54±1 1.5±0.1 8 (2) KNGQ −58±2 1.1±0.1 5 YBKN −27±5.25 0.3±0.1 4 LAIN 56±2 1.0±0.1 7 Multiple ILON300−360 140±5 1.1±0.2 1

ARVN0–90 46±4 1.1±0.09 1 ILON2 null – 2

ARVN90–180 29±5 1.5±0.10 1 ILON290 null – 3

ARVN180–275 45±2 0.8±0.02 7 ILON316 null – –

ARVN275–300 61±5 0.9±0.04 1 ILON327 null – –

ARVN300–315 −89±9 0.7±0.06 1 ILON330 null – 2

ARVN315–360 59±4 1.5±0.09 2 ILON334 null – 3

ARVN320 null – 2 SHWN90–180 0±4 1.4±0.1 1 BULN90–275 56±6 1.0±0.10 5 (2) SHWN180–360 128±9 0.5±0.1 1 BULN275–360 114±7 0.9±0.20 6 (–) SHWN282 null – – FCC0−90 55±4 0.9±0.1 2 SHWN259 null – – FCC90−180 2±10 0.7±0.1 2 Dipping FCC180–275 41±1 0.8±0.1 29 FRB0–90 126±2 1.1±0.1 3 (–) FCC275–300 77±6 0.6±0.1 2 FRB90–183 52±8 0.6±0.1 3 (1) FCC300–315 90±4 0.9±0.1 5 FRB183–270 62±7 0.9±0.1 3 (2) FCC315–360 41±1 0.8±0.1 6 FRB270–300 83±1 0.9±0.1 19 (10) FCC316 null – – FRB300–330 103±2 0.7±0.1 9 (6) FCC320 null – 3 FRB330–360 107±4 0.7±0.1 5 (1) FCC325 null – 3 HP0–90 123±1 1.3±0.1 5 FCC327 null – – HP90–180 70±18 0.6±0.2 3 Interface HP180–270 77±4 0.9±0.1 3 CRLN110–180 5±7 1.2±0.20 1 (2) HP270–300 84±1 0.9±0.1 21 CRLN300–360 97±11 0.8±0.50 9 (1) HP300–330 99±2 0.8±0.1 18 GIFN90–180 7±5 1.3±0.2 4 HP330–360 102±2 0.8±0.1 19

GIFN180–360 140±4 0.7±0.1 1 Null Only

GIFN269 null – – KUGN258 null – 2

ILON90–180 30±2 0.8±0.1 8 KUGN319 null – –

ILON180–300 106±4 0.9±0.1 3 KRSQ346 null

Fig. 3. MSATcalculatedφresponses forthreebasicclassesofanisotropy. in-dicatesdifferenceinfastdirectionbetweenlayers.Layerthicknessandalignment fractionofolivinea-axisdoesnotaffectthepatterns.

belowthe

40mm/yrsometimesconsideredaminimumforbasal drag fabric development (e.g.,Debayle and Ricard,2013;

Martin-Shortetal.,2015).Weobservenoconsistent,network-wide align-ment with the currentAPM direction, which hasbeen relatively constantforthepast

50Myr(Mülleretal.,2016).Althoughwe cannot preclude thepossibility that olderbasaldrag fabrics exist beneaththeregion,asthenosphericfabricsduetoAPMorregional mantle flow (e.g., Forte et al., 2015) would, accordingto Fresnel Zone arguments (Alsina andSnieder, 1995), be expected to pro-duce onlygradualvariations in

φ

and

δ

t across ournetwork.Our observedsplittingparameters,infact,varyoverhorizontal length-scalesas

150km(e.g.,SouthamptonIslandstationsversusCTSN, Fig. 6).We therefore reject APMasan interpretation for our ob-served anisotropy, asopposedto Snyder etal. (2013) who inter-pretedAPMascontributingalowerlayerofanisotropy acrossthe entiretyof the HudsonBayregion. The paucityof tectonic activ-ity since

1.8Ga rules out alignedmelt asa plausible cause of the observed anisotropy, so our discussion proceeds on the as-sumptionthatourdataareprimarilysensitivetofossillithospheric

(5)

Fig. 4. SplittingparametersforHPblockstationsonsouthernBaffinIsland(Fig. 1). Nullmeasurementsareindicatedbydashedlineswithdiamondsatthe parallel/per-pendiculartobackazimuthpoints.

anisotropy,consistentwithearlierstudiesoftheregion(Bastowet al.,2011).

Several stations in Table 1 exhibit strong backazimuthal vari-ation in splitting parameters, implying the presence of complex structure,includingmultipleand/ordippinganisotropiclayers.We address these issues in the following sections on a region-by-regionbasis.

5.2. Evidenceforproterozoicplate-scaleunderthrustingBeneathBaffin Island

Fastdirectionsforstationsin theHudsonStrait andalongthe northerncoastofQuebecandsoutherncoastofBaffinIslandtend to parallel the strike of the THO (Fig. 1; Stations AKVQ, MANN, CTSN,IVKQ,NOTN,DORN,CHIN,KIMN,KNGQ).Thisinterpretation is supported by the general trend in anisotropic direction found by Darbyshire et al. (2013) using surface waves, andthe earlier SKSsplittingstudyofBastowetal.(2011).Unfortunately,the east-ernmost stations, KRSQ andKAJQ, hadvery few successful splits and so cannot be used to confirm or refute the THO hypothe-sis. Lithospheric-scale THO deformation clearly exerts first-order controlontheanisotropy.TheapproximatelyEWfastdirectionof PNGN parallels the Baffin suture to the west of the station. At CMBN,

φ

aligns more closely withthe SE turn of the suture. At stationPNGN,

φ

parallelstheAPMdirection,buttheshortlength scalechange comparedto stationCMBN, lessthan100 kmaway, suggestsstronglythat thesource oftheanisotropy ismorelikely within the

200km-thick lithosphere inthis region (Alsina and Snieder,1995;Porrittetal.,2015).

StationsFRB,CDKN,JENN,andMNGNliewithintheHall Penin-sula.Tofirstorder,

φ

atMNGNandCDKNparallelsthetraceofthe THO,butJENNisnearlyperpendicular(Fig. 1).Thereasonforthis becomes clearin Fig. 4;JENNrecordsa splitfromabackazimuth notsampledby CDKN orMNGN,butmirrorsobservations atFRB atsimilar backazimuths (Fig. 4), givingcredenceto theidea that thesestationssample similarstructure andwarrantconsideration as a composite ‘Hall Peninsula’ or ‘HP’ station with good back-azimuthal coverage. The stations included in the ‘HP’ stack were chosen basedon their proximity toFRB (which providesmostof the data and backazimuthal coverage to the stack) and because theysharesimilarlocalgeology.TheHallPeninsularegion(Fig. 1) ischaracterisedbyhighgrademetamorphism,andwaspartofthe leadingedgeoftheChurchillplateduringtheTHO(St-Ongeetal., 2006,2009).NearbystationsKIMN andDORNarenot includedin the HP group because KIMN lies on the other side of the Soper River suture from the rest of the stations, andDORN has lower grademetamorphismonthewesternreachesofBaffinIsland.The splitting parameters for all HP stations were stacked within ap-propriate backazimuth bins (Table 1, Fig. 5), revealing a clear, smoothly-varying 360◦ periodicity, without the sharp changes in

φ

that are characteristicofmulti-layer anisotropy (Fig. 3).

There-fore,thatclassofmodelwasusedtoinitialiseanMSATmodelling proceduretocharacterisetheanisotropicstructure.

Agridsearchvaryinglayerdip,up-dipdirection,andolivine a-axisazimuth(AAZ)orientationwas performedandtherootmean squared(RMS)misfitbetweenthemodelresponseforeachsetof parametersandtheamalgamHPstationdatawascalculatedto in-dicate the agreementbetweenmodel andobservations.RMSwas calculatedusingtheformula:R M S

=



1

n

(

x

2

1

+

x22

. . .

x2n

)

wheren is the number of observed data points, and x is thedifference be-tween thedatapointandthemodelledcurve,lesstheerror.Data points were assumedto haveerrors inbackazimuth of

±

5◦. The best fitting modelwas determined by thelowest combined RMS forbothfor

φ

and

δ

t.TheRMSvaluesforeachsplittingparameter were normalised for each cell inthe grid search, then combined (Fig. 5,top)such that themaximumvalue ineach grid is1.This methodensuresthatboth

φ

and

δ

t areincludedinthe determina-tionofthebestfittingmodelparameters.

Fig. 5showstheresultsofthegridsearch forthethreemodel parameters. Initially, AAZ

=

0◦ (aligned with the dip direction), the simplestpossible scenario.TheminimumRMS valuegivesan updip direction of 265◦

±

10◦ anda dip angle of 70◦

±

5◦. These valuesarethenheldconstantwhilevaryingtheothertwo param-eterstoconfirmthebestfittingmodelvalues.Themiddleandright RMS surfacesinFig. 5 confirmthat fan AAZof0◦ is appropriate. The bestfittingmodelis thereforealayer dipping at70◦,witha dip direction ofN85◦E (updipdirection of 265◦). These parame-ters yield an RMS of 5.0◦ and 0.12 s, for

φ

and

δ

t, respectively. Layer thickness and olivine alignment fraction had no effect on the patternofthe splittingparameters; theeffectwas limited to a static shift of

δ

t, and is therefore not well defined. What can beconfidentlyconcludedisthatSKSsplittingresolvesasignificant amount of eastward dipping anisotropy beneath southern Baffin Island. Such a dipping layer is consistent withrecent findings of eclogiterocksintheTHOindicativeofdeepplate-scalesubduction (Weller andSt-Onge, 2017). Our HP block results therefore pro-videcompellingsupportforthehypothesisthattheSuperiorplate underthrust the Churchill plate in a modern-style plate tectonic collisionat

1.8 Ga.

5.3. Evidenceforlithosphericsubdivisions:implicationsforthe two-plateTHOhypothesis

Splitting parameters for stations on Southampton Island (SI) are clearly distinguished fromthe nearby HudsonStrait stations:

φ

variessharplybetween315◦ and180◦ backazimuth atstations CRLN, SHWN, andSHMN,while CTSNcan be explained easilyby a single,horizontalanisotropic layer(Fig. 6).Withnodatapoints atbackazimuths

180◦,wecannotconfidentlydeterminethe pre-cise anisotropic modelthat would best explain the observations. However,wecandiscountsomepossibilitiesbyobservingthatthe smoothly varying

φ

over 180◦ backazimuth (Fig. 6) is uncharac-teristic of a two-layer model (which has a periodicity of 90◦); peak-to-peak variations of

φ

(

135◦) are larger than expected (

90◦) fora dippinglayer (Figs. 3, 5).Further,thelow metamor-phic gradeofSouthampton Island rocks indicates thatSI hasnot undergone the same deformation and uplift as the higher-grade rocks on southern Baffin Island (St-Onge et al., 2009). Superior plate underthrusting, and a dipping layer, is thereforeless likely inthisregion.

A more plausibleexplanation for the Southampton Island ob-servationscouldbelateralvariationsinanisotropyduetoanearby terrane boundary. Thisinterpretation is inline with magnetotel-luric evidence of a thickening resistive crust to the northeast of Southampton Island,whichwas interpretedasapotential terrane boundarybySprattetal.(2012).Suchaboundarycouldpotentially

(6)

Fig. 5. (Top)NormalisedRMSmisfitsurfacesforφandδt combined.Leftsurfacesetsolivinea-axisto0◦relativetodipdirection.TheparameterswithlowestRMSmisfitare eachkeptconstantindividuallyinthenexttwosurfaces.(Bottom)Modelledresponseforthebestfittingdip,a-axisorientation,andup-dipdirectionplottedagainststacked HPblocksplittingparameters.

Fig. 6. SplittingparametersforthestackedSouthamptonIslandstationsandnearby CTSN, in Hudson Strait. Both groups are stacked in backazimuth bins: 0–90◦, 90–180◦,180–285◦,285–330◦, and330–360◦.Solidblacklineisthebestfitting HPmodelfrom

Fig. 5

.

be an extension ofthe Rae-Hearne boundary (Snowbird Tectonic Zone:Fig. 1), oran extensionof theMISHblock beneaththis re-gion. Similar Vp/Vs ratios and metasedimentary rocks between SIandmainlandRae(Thompson etal.,2010)support theformer, whilegeologicalmodelsbyCorriganetal.(2009)supportthe lat-ter.

While the nature of the boundary cannot be confirmedfrom ourdata,theextensionoftheSTZontoSouthamptonIslandisless likelythananinfluencefromtheMISHblockbecauseofthe sim-ilarity in

φ

between the Rae and the Hearne on the mainland (Fig. 1). Stations NUNN, WAGN, and QILN on the Rae mainland, andSEDN and JOSN on the Hearne (Fig. 1), parallel THO-related deformationratherthantheSIpatternshowninFig. 6.Therefore,

Southampton Island is anisotropically distinct both fromthe cra-tonicprovinces(Rae/Hearne)tothenorthandtheHudsonStraitto thesouth. Thisstronglysuggeststhat Southampton Islandis part ofalithosphericblockseparatefromtheSuperiorortheChurchill plates. Our observations in this region thus imply that the THO cannotbeconsideredasasimpletwo-plateProterozoiccollision.

5.4. NortherlyextentoftheTHOonMelvillePeninsula

The north-central Raedomain is dominated by granite-green-stone rocks withPaleoproterozoic deformation zones, potentially having been reworked by far-field THO pressure (Carson et al., 2004). With the notable exception of station BULN, the stations on the central Rae craton in Fig. 1 parallel these surface struc-turaltrends.TheaverageanisotropicfastdirectionfortheMelville Peninsula stationsLAIN, AP3N,andSRLN, is34◦,the same orien-tation found for mantle-depth geoelectric strike by Spratt et al. (2013) (Fig. 7). Studies by Berman et al. (2005) and Corrigan et al. (2009) also noted strongNE–SW geological trends in this re-gions, and relatedthem to late-Archean eventsfollowed by THO relateddeformation. StationsMARN andPINUon northern Baffin Island display

φ

values that parallel the Northeast Baffin Thrust Belt(JacksonandBerman, 2000), which isthe western extent of theNagssugtoqidianorogen(NO)preservedinmodern-day Green-land.

ILON and GIFN (Fig. 8) have strongly varying fast direction, however, we do not present anymodelling ofthese stations be-causeofthelackofreliableresultsfromthe0–180◦ backazimuth. Thepeak-to-peakvariation of

φ

forILON/GIFNinFig. 8precludes

(7)

Fig. 7. Geologicalmap ofMelville PeninsulaandnorthernBaffin IslandwithSKS splittingresults.Dashedyellowbarsindicatestackedφandδt values,solidyellow barsindicateunstackedindividualmeasurements.Redcrosses arenull measure-mentswith90◦ambiguity.Legendisthesameas

Fig. 1

.THO,Trans-HudsonOrogen; NBTB,NortheastBaffinThrustBelt.(Forinterpretationofthereferencestocolourin thisfigurelegend,thereaderisreferredtothewebversionofthisarticle.)

Fig. 8. Fastdirectionasafunctionofbackazimuthshowingsimilarities between GIFN/ILON and two single-layer-interpreted regions at different backazimuths. MainlandMelville stationsareAP3N/LAIN/SRLN.BaffinIslandstationsareMARN/ PINU.

thepossibilityofadippinglayer,andthedataarenot90◦periodic as expected for two-layer systems (Fig. 3). Although we cannot precludethepossibilitythatmulti-layeranisotropyexistshere,our datacoverage limitsustodiscussionof thevariable

φ,

and com-parisonwithnearby,lessvariable,stations.StationsGIFNandILON lie between the two orogen-dominatedregions anddisplay vari-able

φ

valuesthatshowinfluencefromboth(Fig. 8).Atsoutherly backazimuths,

φ

valuesare closetothoseobserved onthe main-landofMelvillePeninsula, butsignificantly changetoparallelthe BaffinIsland stationsasbackazimuth increasesto

300◦ (Fig. 8). Thesestations representa transect through deformation regimes, andthereforemarktheapproximatelimitofProterozoic deforma-tionbothfromthesouthandthenorth.Itisexpectedthat oroge-niesofthesizeoftheTHOandtheNOwill overprintanyprevious anisotropic signature. This isconsistent with theview that fossil lithosphericanisotropydocumentsthelastmajortectoniceventin a region (Silver andChan,1991). This hasbeen previously inter-preted for Melville Peninsula near ILON and GIFN (Spratt et al., 2013;Bermanetal.,2005).OurresultsindicateaTHO-related de-formationzone

700kmwide,similarto thewidthof

650km suggested by Gilligan et al. (2016) across the northern Quebec– southern Baffin transect, and comparable to the 400–1000 km widthofthemodern-dayTibetanplateau.

Fig. 9. FastdirectionanddelaytimeasafunctionofbackazimuthforbothWestern HudsonBaystations.Nullmeasurementsaredashed lineswith diamondsatthe parallel/perpendiculartobackazimuthpoints.Thesimilarityissuchthatthesame structurecanbeinterpretedbeneatheachstation.Theblacklineisa2-layermodel witha50kmthickupperlayerwithφ=31◦overan80kmthicklayerwithφ=

79◦.

5.5. Implicationsfor2-stagekeelformation

Western HudsonBay stations FCC andARVN havewide back-azimuthal coverageandsimilarvariations in

φ

and

δ

t (Fig. 9).For both stations

φ

variesfrom

45◦ to

20◦ but increasessharply around 270◦ backazimuthto

90◦.Thispattern mostclosely re-semblesthe90◦periodicityofthe2-layersyntheticmodelinFig. 3. StationBULNhasa similar, butlessclear,patternofsplitting pa-rameters.Itpotentiallywarrantsdiscussionasamulti-layermodel also,butarelativelysimple2-layersystemisnotcapableof recre-ating the observations. The cratonic nucleusof the Rae province grewbyaccreting terranestoits boundaries(Snyderetal.,2016). The variabilityof

φ

observedatBULN, whichliesonone ofthese accreted terranes, could be fromcomplexanisotropic fabricsthat are the remains oftheseArchean-age accretionary processes,but nearbystationslackthebackazimuthalcoveragetocorroboratethis interpretation.

Snyder et al. (2013) published receiver function results and anisotropic

φ

valuesforatwo-layersystemforstationFCC(Table 2 andFig. 8ofSnyder etal., 2013).We usedthereceiverfunctions todefinepotentiallayerthicknessesfortheir

φ

valuesandcreated a modelwithwhichto compareourdata.Thismodelhasa

φ

of 31◦ overlying 79◦, withlayer depths of50 km and130 kmand is plottedasthe solid line inFig. 9.The model periodicity qual-itatively fits our data and supports the interpretation of 2-layer anisotropyinthisregion.Atbackazimuthsof

315◦,however,the dataarelesswellmatched,indicating thatthemodelmaybe too simplistic. More recentlypublished results fromDarbyshire etal. (2013) showthat thelithosphere–asthenosphereboundaryinthis regionisat

240kmdepth;avaluemuchdeeperthanthelower anisotropic layerusedtoproducethemodelinFig. 9.Further,the similaritybetweentheFCCandARVNdatainFig. 9linksthe struc-ture beneath these stations, thus precluding the Lynn Lake fault (local to station FCC) from being the upper layer of anisotropy. A layeredlithospheremodelresultingfromepisodiccratonic devel-opmentthusfitstheobservationsbetterthanonerequiringAPM.

Cratonicregionsaretypicallyunderlainwithexceptionallythick (

250km)highwavespeedlithospherickeels.Thehighly-depleted cratonic coreobtains a thermallyaccreted, morefertile boundary

(8)

layeratsomepointafterinitialcratonicdevelopment,probablyno earlier than the Paleoproterozoic (Yuan and Romanowicz, 2010; Bastow et al., 2013; Bao et al., 2016; Boyce et al., 2016). Sup-portforthisinterpretationcomesfromstudiessuchasYuanetal. (2011)andDarbyshireetal.(2013),whichhavefoundevidencefor astratifiedlithosphere, cratonic andotherwise,in NorthAmerica usingSKS splitting,long-period seismic wave inversion, and sur-facewavetomography.AvarietyofreceiverfunctionandSKS split-tingstudieshavealsosuggestedthepresenceofasharpchangein seismicvelocityatmid-lithosphericdepthsinNorthAmerica(e.g., MillerandEaton, 2010; YuanandRomanowicz,2010). Whileour resultsdo not support a simple two-layer model forthis region, theyareconsistentwiththepresenceofmultiplelayersexhibiting complexanisotropyandarethereforeconsistentwiththepresence ofan MLDpredictedbythe two-stagecratonicdevelopment the-ory.

6. Conclusions

We have investigated seismic anisotropy across much of the northern Hudson Bay and Baffin Island region of Canada using shear-wave splitting analysis of up to 23 yr of SKS and SKKS phases.Below theArchean western coast ofHudsonBaywe find strong evidence for at least two anisotropic layers in the litho-sphere. Our observations are interpreted to be due to cratonic stratification andsupport a multi-stage keel development theory (e.g.,YuanandRomanowicz,2010).Morebroadlyacrossthestudy area,however, we observea prevalenceof anisotropic fast polar-isation directions that parallel the strike of the Paleoproterozoic TransHudsonOrogen(THO),implyingthatitimpartedplate-scale anisotropic fabrics on the region at

1.8 Ga. Variations in

φ

on Southampton Island reveal a unique mantle anisotropy signature consistentwiththepresenceofamicro-continent(Meta-Incognita) caught between the colliding Churchill and Superior plates. This pointsto amore complexmodelthan asimpletwo-plate system forthe THO.Theperiodicityof

φ

and

δ

t forstationson southern BaffinIslandindicates thepresenceofdipping anisotropy.We in-terprettheanisotropytobeduetotheSuperiorplate underthrust-ing theChurchill at

1.8Ga. THO-parallel plate-scale anisotropic fabrics persist as far north as Melville Peninsula, implying THO deformation was as laterally extensive (

700 km) as in the Hi-malayas today. Our results thus constitute strong evidence that modern-styleplatetectonics were in actionduring, andwere re-sponsiblefor,Paleoproterozoicorogens.

Acknowledgements

The authors would like to thank A. Walker for invaluable helpunderstanding theMSAT forwardmodellingcode,aswell as A. Boyce, L. Petrescu, and C. Ogden of the ICcratons group for numerousenlighteningconversationsaboutCanadianPrecambrian geologyandbeyond.M.V.LiddellisfundedbyanImperialCollege President’sScholarship.F.A.Darbyshire issupported bythe Natu-ralSciencesandEnvironmentResearchCouncilofCanadathrough their Discovery Grant and Canada Research Chair programmes (341802-2013-RGPIN).

Appendix A. Supplementarymaterial

Supplementarymaterialrelatedtothisarticlecanbefound on-lineathttps://doi.org/10.1016/j.epsl.2017.09.030.

References

Alsina,D.,Snieder,R.,1995.Small-scalesublithosphericcontinentalmantle defor-mation:constraintsfromSKSsplittingobservations.J.Geophys.Res. 115(9), 1–24.

Bao, X.,Eaton, D.W.,2015. Largevariations inlithosphericthicknessofwestern Laurentia:tectonicinheritanceorcollisionalreworking?PrecambrianRes. 266, 579–586.

Bao, X.,Eaton,D.W.,Gu,Y.J.,2016.Rayleighwaveazimuthallyanisotropicphase velocitymapsbeneathwesternCanada.J.Geophys.Res. 121,1821–1834. Barruol,G.,Mainprice,D., 1993.A quantitativeevaluationofthecontributionof

crustalrockstotheshear-wavesplittingofteleseismicSKSwaves.Phys.Earth Planet.Inter. 78(3–4),281–300.

Bastow, I.D., Thompson,DavidA., Wookey, J.,Kendall, J.M.,Helffrich,G., Snyder, D.B.,Eaton, D.W.,Darbyshire,F.A.,2011. Precambrianplatetectonics:seismic evidencefromNorthernHudsonBay,Canada.Geology 39(1),91–94. Bastow,I.D.,Eaton,D.W.,Kendall,J.M.,Helffrich,G.R.,Snyder,D.B.,Thompson,D.A.,

Wookey,J.,Darbyshire,F.,Pawlak,A.E.,2013.TheHudsonBayLithospheric Ex-periment(HuBLE):insightsintoPrecambrianplatetectonicsandthe develop-mentofmantlekeels.Geol.Soc.(Lond.)Spec.Publ. 389(1),41–67.

Berman,R.G.,Davis,W.,Pehrsson,S.,2007.CollisionalSnowbirdtectoniczone res-urrected:growthofLaurentiaduringthe1.9Gaaccretionaryphaseofthe Hud-sonianorogeny.Geology 35(10),911–914.

Berman, R.G., Sanborn-Barrie, M., Stern, R.A., Carson, C.J., 2005. Tectonometa-morphism at ca. 2.35 and 1.85 Ga in the Rae domain, western Churchill province,Nunavut,Canada:insightsfromthe structural,metamorphicandin situ geochronologicalanalysisofthesouthwestern CommitteeBaybelt.Can. Mineral. 43(1),409–442.

Blackman, D.,Kendall, J.M.,1997. Sensitivity ofteleseismicbody wavesto min-eraltextureandmeltinthemantlebeneathamidoceanridge.Philos.Trans. R.Soc. A 355,217–231.

Boyce,A.,Bastow,I.D.,Darbyshire,F.,Ellwood,A.G.,Gilligan,A.,Menke,W.,2016. SubductionBeneathLaurentiamodifiedtheNorthAmericanCratonEdge: evi-dencefromPandS-wavetomography.J.Geophys.Res. 121,5013–5030. Bystricky,M., Kunze,K.,Burlini,L.,Burg,J.-P., 2000.Highshearstrainofolivine

aggregates:rheologicalandseismicconsequences.Science 290,1564–1567. Carson,C.J.,Berman,R.G.,Stern,R.A.,Sanborn-Barrie,M.,Skulski,T.,Sandeman,H.A.,

2004.AgeconstraintsonthePaleoproterozoictectonometamorphic historyof theCommitteeBayregion,westernChurchillProvince,Canada:evidencefrom zirconandinsitumonaziteSHRIMPgeochronology. Can.J.EarthSci. 41(9), 1049–1076.

Connelly,J.,Thrane,K.,Krawiec,A.,Garde,A.,2006.LinkingthePalaeoproterozoic NagssugtoqidianandRinkianorogensthroughtheDiskoBugtregionofWest Greenland.J.Geol.Soc. 163(2),319–335.

Corrigan,D.,Pehrsson,S.,Wodicka,N.,DeKemp,E.A.,2009.ThePalaeoproterozoic Trans-HudsonOrogen:aprototypeofmodernaccretionaryprocesses.Geol.Soc. (Lond.)Spec.Publ. 327,457–479.

Darbyshire,F.,Eaton,D.W.,Bastow,I.D.,2013.Seismicimagingofthelithosphere be-neathHudsonBay:episodicgrowthoftheLaurentianmantlekeel.EarthPlanet. Sci.Lett. 373,179–193.

Debayle,E.,Ricard,Y.,2013.Seismicobservationsoflarge-scaledeformationatthe bottomoffast-movingplates.EarthPlanet.Sci.Lett. 376,165–177.

Eaton,D.W.,Adams,J., Asudeh,I., Atkinson,G.M.,Bostock,M.G.,Cassidy,J.F., Fer-guson,I.J.,Samson,C.,Snyder,D.B.,Tiampo,K.F.,Unsworth,M.J.,2005. Investi-gatingCanada’sLithosphereandearthquakehazardswithportablearrays.Eos Trans.AGU 86(17),169.

Forte,A.M.,Simmons,N.A.,Grand, S.P.,2015.Constraintson3-Dseismicmodels fromglobalgeodynamicobservables:implicationsfortheglobalmantle convec-tiveflow.In:Romanowicz,B.,Dziewonski,A.M.(Eds.),2ndedition.In:Treatise onGeophys.,vol. 1.Elsevier,Amsterdam,pp. 853–907.

Gilligan, A., Bastow, I.D., Darbyshire, F., 2016. Seismological structure of the 1.8 GaTrans-HudsonOrogenofNorthAmerica.Geochem.Geophys.Geosyst. 17, 2421–2433.

Hoffman,P.F.,1988.UnitedPlatesofAmerica,thebirthofacraton:EarlyProterozoic assemblyandgrowthofLaurentia.Annu.Rev.EarthPlanet.Sci. 16(1),543–603. Jackson,G.D.,Berman,R.G.,2000.Precambrianmetamorphicandtectonicevolution

ofNorthernBaffinIsland,Nunavut,Canada.Can.Mineral. 38,399–421. Kaminski,É.,Ribe,N.M.,2002.Timescalesfortheevolutionofseismicanisotropyin

mantleflow.Geochem.Geophys.Geosyst. 3(8),1–17.

Long,M.D.,Silver,P.G.,2009.Shearwavesplittingandmantleanisotropy: measure-ments,interpretations,andnewdirections.Surv.Geophys. 30(4–5),407–461. Martin-Short,R.A.,Richard,M.,Bastow,I.D.,Totten,E.,Richards,M.,2015.Mantle

flowgeometryfromridgetotrenchbeneaththeGorda–Juan deFucaplate system.Nat.Geosci. 8,1–21.

Miller,M.S.,Eaton,D.W.,2010.Formationofcratonicmantlekeelsbyarcaccretion: evidencefromSreceiverfunctions.Geophys.Res.Lett. 37(18),1–5.

Müller,R.D., Seton,M., Zahirovic,S.,Williams,S.E., Matthews,K.J.,Wright,N.M., Shephard,G.E.,Maloney,K.T.,Barnett-Moore,N.,Bower,D.J.,Cannon,J.S.,2016. Ocean basin evolution and global-scale reorganization events since Pangea breakup.Annu.Rev.EarthPlanet.Sci.Lett. 44,107–138.

Pawlak,A.,Eaton,D.,Bastow,I.,Kendall,J.-M.,Helffrich,G.,Wookey,J.,Snyder,D., 2011.CrustalstructurebeneathHudsonBayfromambientnoisetomography: implicationsforbasinformation.Geophys.J.Int. 184(1),65–82.

Porritt,R.W., Miller,M.S.,Darbyshire, F.,2015. Lithospheric architecturebeneath HudsonBay.Geochem.Geophys.Geosyst. 16(7),2262–2275.

(9)

Restivo,A.,Helffrich,G.,1999.Teleseismicshearwavesplittingmeasurementsin noisyenvironments.Geophys.J.Int. 137(3),821–830.

Silver,P.G.,1996.Seismicanisotropybeneaththecontinents:probingthedepthsof geology.Annu.Rev.EarthPlanet.Sci. 24(1),385–432.

Silver,P.G.,Chan,W.W.,1991.Shearwavesplittingandsubcontinentalmantle de-formation.J.Geophys.Res. 96(B10),16429–16454.

Silver,P.G.,Savage,M.K.,1994.Theinterpretationofshear-wavesplittingparameters inthepresenceoftwoanisotropiclayers.Geophys.J.Int. 119,949–963. Snyder, D.B., Berman, R.G., Kendall, J.M., Sanborn-Barrie, M., 2013. Seismic

anisotropyandmantlestructureoftheRaecraton,centralCanada,fromjoint interpretationofSKS splittingand receiverfunctions. PrecambrianRes. 232, 189–208.

Snyder,D.B.,Humphreys,E.,GrahamPearson,D.,2016.Constructionanddestruction ofsomeNorthAmericancratons.Tectonophysics 694,464–485.

Spratt,J.E.,Craven,J.A.,Sanborn-Barrie,M.,2012.SouthamptonIsland magnetotel-luricsurvey:dataacquisitionandpreliminaryanalysis.Geol.Surv.Can.Open File6988,39pp.

Spratt,J.E.,Jones,A.G.,Corrigan,D.,Hogg,C.,Canada,N.R.,2013.Lithospheric ge-ometrybeneathMelvillePeninsula,Nunavut,revealedbydeep-probing magne-totelluricsurveying.Curr.Res.,Geol.Surv.Can. 12.14pp.

St-Onge,M.R.,Searle,M.P.,Wodicka,N.,2006.Trans-HudsonorogenofNorth Amer-icaand Himalaya–Karakoram–Tibetan orogenofAsia:structuraland thermal characteristicsofthelowerandupperplates.Tectonics 25(4),1–22.

St-Onge,M.R.,VanGool,J.A.M.,Garde,A.,Scott,D.J.,2009.CorrelationofArchaean andPalaeoproterozoicunitsbetweennortheasternCanadaandwestern Green-land:constrainingthepre-collisional upperplateaccretionaryhistoryofthe Trans-Hudsonorogen.Geol.Soc.Lond.,Spec.Pubs. 318(1),193–235. Teanby,N.,Kendall,J.-M.,van der Baan,M.,2004.Automationofshear-wave

split-tingmeasurementsusingclusteranalysis.Bull.Seism.Soc.Am. 94(2),453–463.

Thompson,D.A.,Bastow,I.D.,Helffrich,G.R.,Kendall,J.M.,Wookey,J.,Snyder,D.B., Eaton,D.W.,2010.Precambriancrustalevolution:seismicconstraintsfromthe CanadianShield.EarthPlanet.Sci.Lett. 297(3–4),655–666.

Tommasi,A.,Mainprice,D.,Canoval,G.,Chastel,Y.,2000.Viscoplasticself-consistent and equilibrium-basedmodelingofolivinelattice preferredorientations: im-plicationsfortheuppermantleseismicanisotropy.J.Geophys.Res. 105(B4), 7893–7908.

Vauchez,A.,Nicolas,A.,1991.Mountainbuilding:strike-parallelmotionandmantle anisotropy.Tectonophysics 185(3–4),183–201.

Vinnik,L.P.,Farra,V.,Romanowicz,B.,1989.AzimuthalanisotropyintheEarthfrom observationsofSKSatgeoscopeandNARSbroadbandstations.Bull.Seism.Soc. Am. 79(5),1542–1558.

Walker,A.M.,Wookey,J.,2012.Recordofmodern-styleplatetectonicsinthe Palaeo-proterozoicTrans-Hudsonorogen.NatureGeoscience 10,305–312.

Weller,O.,St-Onge,M.R.,2017.MSAT–anewtoolkitfortheanalysisofelasticand seismicanisotropy.Comput.andGeosci. 49,81–90.

Wolfe, C.J., Silver,P.G., 1998.Seismic anisotropyofoceanicupper mantle:shear wave splitting methodologies and observations. J. Geophys. Res. 103 (B1), 749–771.

Yuan,H.,Romanowicz,B.,2010.LithosphericlayeringintheNorthAmericancraton. Nature 466,1063–1068.

Yuan,H.,Romanowicz,B.,Fischer,K.M.,Abt,D.L.,2011.3-Dshearwaveradiallyand azimuthally anisotropicvelocitymodelofthe NorthAmerican uppermantle. Geophys.J.Int. 184(3),1237–1260.

Zhang,S.,Karato,S.I.,1995.Latticepreferredorientationofolivineaggregates de-formedinsimpleshear.Nature 375,774–777.

Figure

Fig. 1. Geological map of northern Hudson Bay with SKS splitting results. Dashed yellow bars indicate stacked φ and δ t values, solid yellow bars indicate unstacked individual measurements
Fig. 2. High-quality splitting measurement example from station ILON. (a) The recorded seismogram showing the SKS phase and the initial window
Fig. 4. Splitting parameters for HP block stations on southern Baffin Island ( Fig. 1 )
Fig. 5. (Top) Normalised RMS misfit surfaces for φ and δ t combined. Left surface sets olivine a-axis to 0 ◦ relative to dip direction
+2

Références

Documents relatifs

La maîtrise de la langue française, qui est dans plusieurs pays africains la langue nationale, de l’administration et de l’enseignement, au côté des langues

To assess the cell capture capabilities of the 3D microdevice we first flowed controlled suspensions of a human prostate cancer cell line (PC3), genetically-modified

As shown in subsection 1.2, from the finite element solution of the fluid problem it is possible to recover the rigid body dynamics of the swimmer... In what follows we present how

Drawing on a recent dataset of the Indian manufacturing industry for 1994 to 2008, this paper shows for eight sectors that core infrastructure and Information &

diffusion of the polymer chains, thus producing an increase of the distance between the LDH platelets until most of the informations linked to the stacking are lost.

Consecutive reactions of adducts, resulting from OH radicals and aromatics, with the tropospheric scavenger molecules O 2 , NO and NO 2 have been studied for benzene, toluene, m-

A une fréquence pure donnée, même dans un fond complexe (qui peut notamment comprendre des effets de résonance, d'ondes de surface...), la forme typique du coefficient de réflexion

Raising the Integral gain lowers the steady state error but one is then controlling the velocity through two integrations; first the velocity error is integrated