Magnetotelluric response and geoelectric structure of the Great Slave Lake shear zone

Magnetotelluric response and geoelectric structure of theGreat Slave Lake shear zone

Xianghong Wu a;*, Ian J. Ferguson a, Alan G. Jones b

a Department of Geological Sciences, University of Manitoba, Winnipeg, MB, R3T 2N2, Canadab Geological Survey of Canada, 615 Booth St, Ottawa, ON, K1A 0E9, Canada

Received 31 July 2001; received in revised form 8 November 2001; accepted 10 November 2001

Abstract

The Great Slave Lake shear zone (GSLsz) is a northeast-trending 25-km-wide dextral continental transform faultfrom the foothills of the Rocky Mountains in northeast British Columbia to the southeast side of Great Slave Lake.Based on its magnetic expression the GSLsz can be correlated for at least 1300 km, mostly in the subsurface.Magnetotelluric (MT) soundings were made at 60 sites in the southwest part of the Northwest Territories, Canada,along the LITHOPROBE Slave^Northern Cordillera Lithospheric Evolution (SNORCLE) Transect Corridors 1 and1A, in the Summer and Autumn of 1996. Of these, 15 sites were along Corridor 1A which crossed the GSLsz, the GreatBear magmatic arc and Hay River terrane to the northwest of the shear zone, and the Buffalo Head terrane to thesoutheast. The primary objective of the Corridor 1A deployments was to image the structure of the GSLsz. Analysis ofMT data indicates that along the Corridor 1A the resistivity structure is approximately 1D at shallow depth (6 1 km)corresponding to Phanerozoic sedimentary rocks, 2D in the upper to mid crust with a strike VN30³E, andapproximately 2D in the lower crust to lithospheric mantle with a strike of VN60³E. The direction in the upper crust isinterpreted to represent the local-scale (6 50 km) horizontal strike of the GSLsz whereas the direction in the mantle isparallel to the larger-scale strike of the GSLsz. 2D inversions of the MT data reveal that the GSLsz forms a crustal-scaleresistive zone (s 5000 6m) that is spatially correlated with a magnetic low. The GSLsz comprises greenschist togranulite facies mylonites. Its high resistivity is interpreted to be due to the resistive nature of the granitic protolith ofthe mylonites and that mylonites are dominated by rocks deformed in the ductile regime. To the northwest of GSLsz theMT profile reveals crustal conductors beneath the Great Bear magmatic arc and Hay River terrane. The enhancedconductivity occurring beneath the Great Bear magmatic arc is interpreted to be caused by electronic conduction withindeformed and metamorphosed rocks of the Hottah terrane or the Coronation Supergroup. The MT results also reveal amantle conductor beneath the margin of the Hottah terrane and Great Bear magmatic arc that is interpreted to beassociated with the subduction of oceanic lithosphere. A second mantle conductor to the southeast is truncated at theGSLsz suggesting an older source for the enhanced conductivity and that the GSLsz includes significant strike^slipmotion of sub-crustal lithosphere. ß 2002 Elsevier Science B.V. All rights reserved.

Keywords: Great Slave Lake; shear zones; Paleoproterozoic; Rae Province; lithosphere

0012-821X / 02 / $ ^ see front matter ß 2002 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 1 ) 0 0 5 9 4 - 5

* Corresponding author. Fax: +1-204-474-7623.E-mail addresses: [email protected] (X. Wu), [email protected] (I.J. Ferguson), [email protected]

(A.G. Jones).

EPSL 6090 7-3-02 Cyaan Magenta Geel Zwart

Earth and Planetary Science Letters 196 (2002) 35^50

www.elsevier.com/locate/epsl

1. Introduction

The LITHOPROBE Slave^Northern CordilleraLithospheric Evolution (SNORCLE) TransectCorridor 1A, located south of Great Slave Lake,Northwest Territories, Canada (Fig. 1), crossesProterozoic terranes interpreted to lie beneaththe thin veneer of Paleozoic sediments. These ter-ranes are the Great Bear magmatic arc, the HayRiver terrane and the Bu¡alo Head terrane, andthe interpretation is based primarily on aeromag-netic data, chips from basement-reaching bore-holes, and extrapolations of surface outcrop tothe north [1]. The Paleoproterozoic Great SlaveLake shear zone (GSLsz), arguably the mostdominant linear feature on the aeromagneticmap of Canada, is located between the Hay Riverand Bu¡alo Head terranes and was proposed byHanmer [2] as the type example of a crustal-scalecontinental shear zone.

One speci¢c objective of the SNORCLE tran-sect investigations was to determine the geometryand lithospheric extent of the GSLsz, given itscontinental scale, so an appropriately designedmagnetotelluric (MT) survey was undertaken.The MT method has been used successfully toinvestigate major fault systems, and examples in-clude the Fraser [3] and the Slocan Lake faults [4]in southern British Columbia, the Tintina-north-ern Rocky Mountain Trench fault in northernBritish Columbia and the Yukon [5], the Denalifault of the Yukon and Alaska [6], the San An-dreas fault in California [7^9], and the Alpinefault on the southern island of New Zealand[10,11].

In MT investigations depth information is ob-tained using the period variation of the response:signals of di¡erent period penetrate to di¡erentdepths into the earth. Short-period MT signals(1034^1033 s) penetrate several hundred metersinto the earth whereas long-period signals (103^104 s) penetrate 100 km or more into the uppermantle. The period range in the present study ex-tends from 1034 to 104 s allowing imaging of thesurface Paleozoic sedimentary rocks, the Precam-brian crust, and the lithospheric mantle.

In this paper we will present analyses and in-terpretation of the MT data acquired across the

GSLsz. We will examine the variation of the geo-electric strike direction with depth, variation ofthe resistivity across the GSLsz, and lithosphericstructure. We will demonstrate that the shear zoneis characterized by a resistive zone at crustaldepths, that the shear zone extends through thecrust, and that there is an associated electricalconductivity signature in the underlying mantlelithosphere. In a companion study to this oneEaton et al. [12] examined the teleseismic responseon a pro¢le crossing the GSLsz. Their preliminaryresults are discussed below.

2. Geological setting

The GSLsz is a northeast-trending 25-km-widedextral continental transform fault, with up to700 km of strike^slip motion, extending fromthe foothills of the Rocky Mountains in northeastBritish Columbia to the southeast side of GreatSlave Lake [13]. It is linked to the convergenceand collision between the Archean Slave andRae provinces in the Paleoproterozoic and devel-oped in arc environment associated with the ini-tial subduction of Slave lithosphere [13^15]. Basedon its magnetic expression the GSLsz can be cor-related for at least 1300 km, mostly in the subsur-face. Ductile displacement on the GSLsz is esti-mated to have been up to 700 km. U^Pb zirconages on syntectonic granites de¢ne a minimumduration for ductile shear of 2.03 to 1.95 Ga[13]. Where it is exposed along the southeast shoreof Great Slave Lake, the GSLsz comprises gran-ulite to lower greenschist facies mylonite belts[13,15]. In areas where it is exposed the GSLszzone has a width of about 25 km [2,15].

Post-collisional convergence of the Slave andRae provinces at ca. 1.8 Ga [16] produced anadditional 75^125 km of strike^slip motion onthe McDonald fault system (MF). The overalltrend of the MF is coincident with the GSLsz,but at a smaller-scale the MF consists of a seriesof an echelon brittle displacements connected bytranspressional transfer zones that cut the GSLsz[15]. Along Corridor 1A the GSLsz and MF arecovered by Devonian sedimentary rocks. The cor-ridor crosses a distinct magnetic low that has a


X. Wu et al. / Earth and Planetary Science Letters 196 (2002) 35^5036

Fig. 1. (a) Selected tectonic elements and structures of the northwest Canadian Shield (after Hanmer, 1988). Na: Nahanni ter-rane, FS: Fort Simpson terrane, HO: Hottah terrane, GB: Great Bear magmatic arc, HR: Hay River terrane, GSLsz: GreatSlave Lake shear zone, BH: Bu¡alo Head terrane, BFZ: Bathurst fault zone, MFZ: McDonald fault zone. The circles show thelocation of MT sites on Corridor 1 and 1A. (b) Location of sites along Corridor 1A. The location of the GSLsz is shown bydashed lines and is inferred from magnetic ¢eld data. The dotted line shows the location of the Pine Point (Presqu'ile) barrier.The mine symbol shows the location of the Pine Point Mine.


X. Wu et al. / Earth and Planetary Science Letters 196 (2002) 35^50 37

width of about 30 km, and, on the basis of itssimilarity to the magnetic response of the exposedGSLsz, this magnetic low is interpreted to be thelocation of the GSLsz.

At the location of the LITHOPROBE pro¢le,the geological terranes to the northwest of thefault are the Great Bear magmatic arc and theHay River terrane and the terrane to the south-east of the fault is the Bu¡alo Head terrane (Fig.1). The Proterozoic Great Bear magmatic arc is acontinental `calc^alkalic' volcano^plutonic de-pression dated at 1.875^1.84 Ga that uncomfort-ably overlies the Hottah terrane and/or (de-formed) Coronation margin strata [17,18]. The

geological structure of the Hay River terrane ispoorly known as the terrane is not exposed atthe surface. Drill-core and potential ¢eld resultsindicate that the terrane consists of granitic ma-terial and corresponds to a magnetic low [19]. TheBu¡alo Head terrane is a 2.4^2.0 Ga magmaticbelt [1]. It comprises metaplutonic and subordi-nate felsic metavolcanic rocks [1]. The north-trending Taltson magmatic zone (1.99^1.90 Ga)welds the Bu¡alo Head terrane to the east [20]and is a belt of granitic to dioritic plutons(1.99^1.92 Ga).

Along Corridor 1A the Precambrian rocks areoverlain by 300^700 m of Devonian sedimentary

Fig. 2. The Groom^Bailey regional strike and maximum phase split orientations along the SNORCLE Transect Corridor 1A at0.01 s, 10 s, 100 s and 2500 s. There is a 90³ ambiguity between the direction of the orientations and the true strike. The resolu-tion scale on the GB plot, represented by the length of the strike symbol, is inversely proportional to mis¢t M2 error between theGB decomposition impedances and the observed impedances. The strike angle plot excludes several sites for which the strike an-gle was unresolved.



rocks. These rocks dip gently to the west at 1.9 mper km [21]. To the northwest of the GSLsz Cor-ridor 1A obliquely crosses a 10-km-wide Middleto Upper Devonian barrier complex, the PinePoint barrier, which forms part of the larger-scalePresqu'ile Barrier separating the Devonian ElkPoint Basin in the south from the Mackenzie Ba-sin in the north. The Pine Point barrier hosts themajor Mississippi Valley type Pb-Zn Pine Pointdeposit that is located approximately 25 km northof the corridor [21]. The overburden in the surveyarea consists of V20 m of Holocene and Pleisto-cene sediments (e.g. [21]).

3. MT survey and data analysis

Soundings at 60 MT sites were completed alongSNORCLE transect Corridor 1 and Corridor 1Ain the southwest Northwest Territories, Canada inthe Summer and Autumn of 1996 [22]. There are15 sites on Corridor 1A (Fig. 1). Sites 155 and 156lie above the interpreted location of the GSLszbut the magnetic low extends over much of thedistance between 154 and 157 (Fig. 2). Sites 154and 153 lie on a sigmoidal-shaped magnetic highforming the eastern margin of the magnetic low.Sites 153 to 152 lie above or adjacent to the PinePoint barrier.

The MT recordings included Phoenix V5 mea-surements (1034^100 s) at all sites and GeologicalSurvey of Canada long-period MT responses(LiMS) measurements (20^30 000 s) at all sitesexcept 158 and 159. The MT recordings includedmeasurement of time-variations of the horizontalelectric and magnetic ¢elds in orthogonal direc-tions and of the vertical magnetic ¢eld. Record-ings from adjacent or nearby sites were used asthe remote reference for noise reduction.

The Jones^Jo«dicke approach [23,24] and theChave approach [25] were used for time seriesprocessing to yield the MT response functions.Particular attention was placed to correct theLiMS data for distortion due to non-uniformsource ¢elds associated with auroral geomagneticactivity [26]. The V5 and LiMS data were mergedin order to obtain an 8-decade-period range(1034^104 s). The MT responses determined

from the recorded time series include the MT im-pedance, apparent resistivity and phase. The MTimpedance is a tensor quantity relating the twohorizontal components of the electric ¢eld to thehorizontal components of the magnetic ¢eld in thefrequency domain. Each tensor term can be usedto estimate an apparent resistivity, a spatiallyaveraged resistivity over the penetration depthof the signals. The phase of the impedance, thephase lead of the electric over the magnetic ¢eld,also provides information on the underlying resis-tivity structure.

4. Geoelectric strike directions

The MT impedance tensor can be used to de-termine the dimensionality and strike of the resis-tivity structure. In the maximum phase split meth-od, whereby the impedances are rotated todetermine the greatest phase di¡erence betweenthe o¡-diagonal elements, the geoelectric strikeis determined as the orientation in which thephase di¡erence between o¡-diagonal impedancetensor elements is maximized. Electric charge ac-cumulation on near-surface heterogeneities candistort the measured MT response so that it nolonger accurately represents the larger-scale (re-gional) conductivity structure. In the Groom^Bai-ley (GB) tensor decomposition method [27,28] thegeoelectric strike is determined simultaneouslywith the near-surface distortion e¡ects using aleast-squares approach.

The geoelectric strikes in the study area werederived using GB tensor decomposition and themaximum phase-split method at periods of 0.01 s,10 s, 100 s and 2500 s (Fig. 2,3). The maximumphase split results are for a narrow band of peri-ods centered on these values, whereas the GB re-sults are for a one-decade-wide band. The pene-tration depth of the signals at these periods can beestimated using electromagnetic skin depth rela-tions or the method of Schmucker and Janknow-sky [29] which estimates the depth of the ``centreof mass'' of the subsurface electric current distri-bution.

The MT signals at period 0.01 s correspond topenetration depths of V300 m in the southeast of



the pro¢le and V50 m in the northwest. At thisperiod, the maximum phase di¡erences are small(6 10³) at all sites (Fig. 2) implying little lateralheterogeneity in the shallow subsurface beneatheach site. The GB regional strike direction isweakly de¢ned with an average direction ofVN60³E along the pro¢le (Fig. 3). At most sitesthe ratio of the apparent resistivity correspondingto the MT response in the two orthogonal direc-tions is between 0.85 and 1.20 [22] indicating thedepartures from a 1D or layered MT response aresmall. Because of inherent ambiguity the geoelec-tric strike may be perpendicular or parallel to theazimuth determined from GB analysis and thetrue strike may be VN30³W. This direction cor-responds well with the strike of the Phanerozoicsedimentary rocks (Fig. 2) so it is probable thatthe weak 2D response is associated with the gentlewestward dip of the sedimentary rocks.

A 10-s period corresponds to an upper crustalsignal penetration of 5 km to the northwest of theGSLsz and about 17 km for sites to the southeastof the GSLsz. At this period the GB regional

strike has an orientation of N32³E at most sitesbut an orientation closer to N45³E at sites 151and 152 near the middle of the pro¢le (Figs. 2and 3). The maximum phase orientation is wellresolved and is similar to the GB regional strike.A 100-s period corresponds to penetration of MTsignals to around 30 km depth in the northwest,and s 55 km to the southeast, of the GSLsz. TheGB regional strike direction is approximatelyN56³E at sites crossing the GSLsz but rotates toN29³E at the southeast end of the pro¢le (Figs. 2and 3). The maximum phase di¡erence directionis VN45³E at sites in the Great Bear magmaticarc, but again rotates to about N20³E at sites inthe Bu¡alo Head terrane near the southeast endof the pro¢le.

MT signals at periods of s 1000 s penetratethrough the deepest crust and into the sub-conti-nental lithospheric mantle. Because of the lowersignal to noise ratio at these periods, the GBstrike and maximum phase di¡erence directionsare more erratic (Fig. 2). The GB regional strikedirection has an average orientation of N65³E

Fig. 3. (a) Rose diagram of the GB regional strike angle for four period ranges based on the 15 sites along Corridor 1A. The an-gle bin is 10³. The results for the ¢rst two period bands include 15 resolved strikes angles, the results for band 3 include 14strikes, and for band 4 include 10 strikes. (b) Geoelectric strike angle from multisite, multifrequency extended GB decompositionthat best ¢ts all sites along Corridor 1A simultaneously.



across the pro¢le but has a more east^west orien-tation within the Great Bear magmatic arc. Theaverage maximum phase direction at southeastsites on the pro¢le is N140³E but because of aninherent 90³ ambiguity this direction is likely per-pendicular to the geoelectric strike.

At periods longer than 1 s the GB regionalstrike and maximum phase split results are mod-erately consistent (to within þ 15³) at di¡erentsites and at di¡erent periods and indicate an ap-proximately 2D structure along the pro¢le. How-ever, the results do resolve a change in strike an-gle with period from a strike of VN32³E at 10 sto VN65³E at s 1000 s indicating a 3D compo-nent in the data. The strike angle variation can befurther investigated using the McNeice^Jones [30]multisite, multifrequency extended GB decompo-sition to determine the geoelectric strike anglethat best ¢ts all sites along Corridor 1A simulta-neously. The results of this analysis are shown inFig. 3. Short periods (0.1^10 s) show a preferencefor a strike scattering in the range N30³^35³E,whereas long periods (s 60 s) scatter in the rangeN45³^75³E. Multisite, multifrequency analysis intwo period bands results in a short-period strikeof N33³E and a long-period strike of N62³E. Thetransition between these two directions initiates inthe period range 10^20 s, which is approximatelythe period range for sensitivity of the MT signalsto the lower crust.

5. MT apparent resistivity and phase responses

The GB tensor decomposition results provide ameasure of the distortion of the MT response: theGB shear provides a measure of the local polar-ization of the electric ¢eld. At sites along Corridor1A the shear angle is 6 5³ at periods 6 1 s, in-dicating an absence of near-surface inhomogene-ities. At periods s 30 s the shear at most sites isin the range 10³^30³ indicating moderate distor-tion. The GB distortion parameters vary consis-tently from site to site indicating the source of thedistortion is the large-scale structure rather thanlocal heterogeneities. Near-surface e¡ects cancause a static shift of the MT responses that can-not be resolved in the GB analysis. The static shift

was analyzed using methods proposed by Jones[31] and Sternberg et al. [32] that established thestatic shift using multiple sites along the pro¢le[22]. The resulting static shift correction factorsare small at all sites, typically in the range 0.75^1.3. Accordingly, we exclude static shift correc-tions in the following processing.

The MT response in a 2D structure can be di-vided into two independent modes involving elec-tric current £ow parallel to strike, the transverseelectric or TE mode, and electric £ow perpendic-ular to strike, the transverse magnetic or TMmode [33]. In the present study the TE componentis de¢ned to have an orientation of N41³E, theaverage strike angle over the period range 0.02^200 s. This represents an average of the short-period (N33³E) and long-period (N62³E) strikedirections shown in Fig. 3. Models obtained inthose strike directions showed essentially thesame features discussed below. Fig. 4 shows theapparent resistivity and phase responses for theTE and TM mode.

At periods of 6 0.1 s corresponding to signalpenetration to several hundred meters both theTE and TM response show a decrease in resistiv-ity with increasing period (and thus depth). Thephase exceeds 45³ as expected for cases in whichthe resistivity decreases with depth. At periods of0.1^1000 s corresponding to signal penetrationinto the Precambrian crust the TE and TM ap-parent resistivities are signi¢cantly higher than atshorter period. The TE and TM responses arequite similar and both indicate that the apparentresistivity in the Bu¡alo Head terrane (s 4006m) is higher than that in the Great Bear mag-matic arc (6 400 6m). The TE and TM responsesdi¡er from each other near sites 155 and 156 in-dicating stronger 2D features. In the period range0.3^3 s the part of the pro¢le between sites 153and 156 is characterized by relatively high TEphase (with a local low at site 155) and by rela-tively low TM phase. The apparent resistivity ex-hibits corresponding e¡ects over the period range3^100 s with the TE response being relatively con-ductive (locally more resistive at 155) and the TMresponse being relatively resistive. The responsessuggest the presence of resistive crust beneath site155.



At very long periods (s 1000 s) there is a de-crease in the apparent resistivity beneath the Hot-tah terrane and the Great Bear magmatic arc anda corresponding increase in the phase. These re-sponses indicate the presence of a conductivelayer at mantle depth beneath these geologicalunits. The same MT responses are not visible inthe Bu¡alo Head terrane indicating an absence of,or greater depth from, this conductor.

6. Resistivity structure

The 2D OCCAM [34] and non-linear conjugategradient (NLCG) [35] inversion methods wereused to determine the resistivity structure of thePrecambrian lithosphere. As discussed above, thegeoelectric strike direction varies with period andtherefore depth around the GSLsz. In order toexamine the strike dependence of MT models

Fig. 4. Apparent resistivity and phase pseudosections for theTE and TM modes along Corridor 1A. The pro¢le is per-pendicular to regional geoelectrical strike N41³E. Crossesshow the location of data points. The data have been cor-rected for distortion using a GB model.

Fig. 5. Four inversion models along Corridor 1A.(a) NLCGinversion for data at periods 6 10 s at 30³ strike angle, (b)NLCG inversion for data at periods s 30 s at 60³ strike an-gle, (c) NLCG inversion for data at periods 0.001^3000 s at41³ strike angle. (d) Occam inversion for data at periods0.001^3000 s with 41³ strike angle. There is no signal pene-tration in the blanked area.



from the upper crust to the upper mantle acrossthe GSLsz, inversions of MT data from the di¡er-ent period ranges and with the di¡erent strikeangle were completed (Table 1). An inversionmodel obtained using short-period data (6 10 s)and a N30³E rotation angle is designed to resolveupper to middle crustal structure (Model A). Theresolution depth for a period of 10 s is less than20 km. An inversion model obtained using long-period data (30^3000 s) and N60³E rotation angleis designed to resolve middle to lower crust andupper mantle structure (Model B). Very similarresults are obtained when the starting model forthis inversion includes near-surface resistivitystructures (6 1 km) de¢ned by the higher fre-quency (N30³E) inversion, when it includes uppercrustal resistivity structures (6 20 km) de¢ned bythe higher frequency inversion, and when theupper crustal structure is ¢xed to the results ofthe higher frequency inversion. Finally, inversionsobtained using a full period range (1033^3000 s)and average rotation angle N41³E are designed toresolve the structure from the near-surface tomantle (Models C and D) and to stitch the resultsfrom the other two inversions.

Multiple inversions were done using di¡erentinversion parameters and di¡erent subsets of thedata. Fig. 5 shows the four ¢nal inversion modelsobtained using both TE and TM data. The re-sponses of Model C are shown for representativesites in Fig. 6. The model provides a reasonablygood ¢t to the data: the root mean square mis¢t,based on error £oors of 5% for the apparent re-sistivity data and 1.4³ (equivalent to 2.5% in ap-parent resistivity) for the phase data, is 4.5. Thismeans that we are ¢tting the data to V6³ inphase, on average. The TE data show the poorest¢ts, at periods longer than 10 s at sites 150, 151,154, 155, and particularly 156. Investigation of

the site 156 response shows that for a range of TEdirections, between N30³E and N60³E, the GBdistortion parameters at periods exceeding 10 sare relative high and frequency-dependent. Also,in a series of 2D inversions, e.g. using decreasederror £oors relative to other sites, it was not pos-sible to obtain a good combined ¢t to both TEand TM responses at site 156. These results sug-gest the presence of a more complex crustal struc-ture near the center of the pro¢le. Modelling stud-ies have shown that the TM response is morerobust than the TE response in the presence oflocal 3D structures [36,37] so the superior ¢t tothe TM response provides increased con¢dence inthe models.

The structures in the ¢nal models are robustfeatures that appear in most of the inversions.The major geoelectric structures in the modelare labelled A^D and are discussed below.

b There is a conductive zone (V30 6m) in thelower crust beneath the boundary of the Hot-tah terrane and Great Bear magmatic arc (la-belled A). Seismic re£ection results indicate thedepth to the Moho on Corridor 1 at its inter-section with Corridor 1A is V40 km [17]. TheMT images suggest that the conductive zoneextends into the mantle but the depth of itsbase within the mantle is not well resolved.There is a sharp eastern edge to the conductivezone at site 147.

b There is a conductive zone (labelled B) with itstop at V15 km depth below sites 149 and 150.The conductivity of the body is 6 50 6m inmodels B, C and D and its thickness is at least20 km. Its presence is supported by the obser-vations in the apparent resistivity pseudosec-tions (Fig. 4) which show moderately low resis-tivity at periods of 10^1000 s beneath sites 149and 150. The conductor is located near the in-terpreted position of the Hay River terrane andGreat Bear magmatic arc (Fig. 1).

b There is a conductive zone (labelled C) at man-tle depths beneath sites 150^154. The resistivityof the zone reaches values of 6 100 6m in allof the models. The upper surface of the zoneappears to have an eastward dip. In some mod-els this conductor connects with the previous

Table 12D Inversions for Corridor 1A data using di¡erent strike an-gles

Model Inversion period Rotation angle Inversion method

A 1033^10 N30³E NLCGB 30^3000 N60³E NLCGC 1033^3000 N41³E NLCGD 1033^3000 N41³E OCCAM



conductor but this aspect is not well resolved.The southeastern margin of this conductor atsites 154 and 155 consists of a sharp transitionto more resistive zone to the southeast.

b There is a sub-vertical resistive zone (V10 0006m) beneath sites 155 and 156, labelled D. For

all inversions of data sets extending to periodsof at least 100 s, and for data rotations ofN30³E, N41³E, and N60³E, the resulting mod-els include a resistive zone extending throughthe crust. The exact con¢guration of the resis-tive zone depends on the rotation of the data

Fig. 6. Comparison of the NLCG inversion model apparent resistivity and phase responses with the observed data at some repre-sentative sites. Solid line: TM response, dash line: TE response, b : TM observed data, a : TE mode observed data. The RMSmis¢t for each site for Model C (Fig. 5) is shown.



inverted. For models derived from data with aN30³E rotation angle (model A) the resistorextends further to the northwest than for mod-els derived from data with a N60³E rotation(model B). It is of note that none of the models¢ts the more conductive TE response at site156, but as noted above, the models do providea good ¢t to the TM mode data at this site andthe mis¢t to the TE mode is interpreted to bedue to local 3D structures. Overall, the resultspermit the conclusion that the MT data requirethe presence of a resistive zone centered on sites155 and 156.

7. Geological interpretation

7.1. Great Bear magmatic arc^Hay River terrane

The Great Bear magmatic arc has been inter-preted by Cook et al. [18] as the product of east-ward subduction of oceanic lithosphere beneaththe Hottah terrane at 1.84^1.87 Ga. Cook et al[18] suggest that the Great Bear magmatic arc isrelatively thin (V3^4.5 km) and lies above eitherHottah crust or imbricated rocks of the Corona-tion margin. Therefore, the enhanced conductivityin the middle and the lower crust beneath theboundary between the Hottah terrane and GreatBear magmatic arc (labelled A in Fig. 5) corre-sponds to deformed and metamorphosed rocks ofeither the Hottah terrane or the CoronationSupergroup. The source of the enhanced conduc-tivity could be either carbon or conductive miner-als concentrated during the deformation andmetamorphism. In studies of rocks from the Ka-puskasing uplift [38] and Trans-Hudson orogen[39], petrophysical models and MT results havesuggested graphite can be a source to enhancethe conductivity of mid-lower crust. Alternatively,Gupta and Jones [40] and Jones et al. [41] discussextensive conductivity anomalies in the crustcaused by interconnected sulphide mineralization.In either case, the enhanced conductivity is causedby electronic conduction in interconnected meta-sediments. The source of the high crustal conduc-tivity observed to the southeast is less certain,particularly because of uncertainty in the bound-

ary between the Great Bear magmatic arc and theHay River terrane.

7.2. GSLsz

2D MT modelling results reveal a sub-verticalresistive zone (s 3000 6m) in the upper-lowercrust beneath sites 155 and 156, which is coinci-dent with the magnetic anomaly low (Fig. 5, 7).The resistive zone is interpreted to represent theelectrical signature of the GSLsz.

Interpretation of the resistive response of theGSLsz requires consideration of the geology ofthe shear zone. Where the GSLsz is exposed tothe northwest it has been mapped as a bundle ofupright belts of mylonites [15]. The mylonites areinterpreted to have been formed from a mixedprotolith of hornblende-biotite, magnetite-bear-ing, granite and granodiorite that was intrudedboth pre- and syntectonically. The oldest mylon-ites are of granulite facies and formed a belt witha width of s 10 km, possibly up to 25 km. Themetamorphic grade of the mylonites decreaseswith decreasing age with the decreasing grade re-£ecting both decreases in temperature and pres-sure [2]. The subsequent strain produced progres-sively narrowing belts of upper and loweramphibolite grade mylonite and lower green-schist-facies chlorite-bearing mylonites. The late

Fig. 7. Comparison of magnetic ¢eld data, long-period MTazimuths, and SKS fast directions. The H and L symbols re-fer to magnetic highs and lows respectively. The MT strikesare for the period band 20^500 s.



stage of deformation evolved through the ductile^brittle transition and the latest stages involveddilational faulting and development of quartzstockworks [2,15].

The relatively high resistivity of the GSLsz sug-gests that at the location of the LITHOPROBECorridor all of the mylonite units have high elec-trical resistivity. Within the GSLsz on the exposedshield greenschist facies mylonite occurs in a net-work of belts up to 2 km wide and 70 km long. Ifsuch greenschist facies mylonites occur within thestudy area they must have relatively high resistiv-ity. This result is consistent with a detailed studyby Olhoeft [42] who found that structural waterprovided little contribution to the electrical con-ductivity. It is also possible that the e¡ective elec-trical conductivity of the lower grade mylonitebelts is also reduced by the presence of a quartz`stockwork' which consists of vertical veins up to25 m wide and up to 40 km long that cut themylonite structures [2].

Studies of major transcurrent faults have pro-vided evidence for both conductive and resistivefault zones. The Tintina fault in the northern Cor-dillera is a major fault with up to 450 km ofdextral transcurrent movement [5]. MT resultsshow that the fault is associated with a 20 kmwide resistive zone (s 400 6m) at depths exceed-ing 5 km. The Denali fault in Alaska is also asso-ciated with relatively resistive rocks at upper crus-tal depths [6] and the San Andreas fault atCarrizo Plain is associated with a resistive zoneat midcrustal depths [7]. In contrast the MT re-sults from elsewhere on the San Andreas fault [8]and from other major transcurrent faults includ-ing the Alpine fault in New Zealand [10] and theFraser fault in British Columbia, Canada [3] havea conductive signature.

A variety of sources of the enhanced conduc-tivity within fault and shear zones have been pro-posed. These include clays and other minerals as-sociated with the faulting process [8], graphiteprecipitated from £uids within the fault [3], andin the case of active faults, £uids present in thefault zone. Serpentinite has also been proposed inthe case of faults involving oceanic rocks and adegree of dip slip movement [8]. The results fromthe GSLsz imply the absence of any of these com-

ponents over a region of V20 km width extend-ing throughout the crust.

The high resistivity of the GSLsz can be attrib-uted to the composition of the protolith, the con-ditions of the deformation and to subsequent £uidmovements and reactions. The protolith for themylonitized rocks, at least to the northwest ofthe study area, is horneblende-biotite granite [2].In its undeformed state this rock would be rela-tively resistive [21]. The rocks within the GSLszare dominated volumetrically by mylonitesformed at granulite to amphibolite metamorphicgrade within a ductile rheology. Deformation oc-curring under these metamorphic conditionswould be less likely to create enhanced conductiv-ity than brittle faulting at lower metamorphicgrade that could be associated with the formationof clay minerals with high cation exchange capaci-ties and enhanced conductivity. The granulite fa-cies metamorphism may have released £uidsthrough dehydration reactions but such £uidscould have either been consumed in retrogradereactions [43] or may have dissipated over thelong time period since the deformation.

7.3. Sub-crustal lithosphere

The MT responses and 2D inversion models allshow a region of enhanced conductivity at depth.The maximum depth to the 150 6m contour inthe 2D Occam models is V150 km and occurs inthe Bu¡alo Head terrane. This depth is inter-preted to correspond to the base of the litho-sphere with the enhanced conductivity at greaterdepth explained by a small component of moreconductive partial melt in the asthenosphere.The enhanced conductivity observed at shallowmantle depths beneath the boundary of the Hot-tah terrane and Great Bear magmatic arc (la-belled A in Fig. 5) and to the northwest of theGSLsz (labelled B in Fig. 5) is more localized andis therefore interpreted to represent increased con-ductivity within the lithosphere rather than a de-crease in lithospheric thickness.

Seismic re£ection results show delaminationstructures extending to 100 km depth in the man-tle beneath the Great Bear magmatic arc provid-ing evidence that lower crustal rocks were em-



placed in the mantle during subduction [17,18]associated with collision and accretion of the Hot-tah terrane to the western margin of the Slavecraton between 1.94 and 186 Ga. Bostock andCassidy [44] also suggest on the basis of teleseis-mic and heat£ow data that the lithospheric man-tle beneath the east and southeast of the Slaveprovince has been reworked through processesassociated with subduction and continental colli-sion. The source of the enhanced conductivity inthe mantle beneath the western Great Bear mag-matic arc may be either hydrogen or carbon in-troduced into the mantle through the subductionprocess.

There is a signi¢cant boundary in the mantleconductivity at the GSLsz with the conductivitybeing higher to the northwest of the boundary.The precise geometry of this boundary is ratherpoorly constrained by the MT data because it isoccurring at the longest periods available in theresponse and is obscured in part by the overlyingcrustal structures. Nevertheless the existence ofthe conductive mantle to the northwest of theGSLsz is supported by the observed high phasesand low apparent resistivity at periods exceeding103 s at sites northwest of 154 (Fig. 4). The trun-cation of the mantle conductor near the GSLszsuggests signi¢cant strike^slip movement of man-tle lithosphere as well as the crust. However, themovement on the GSLsz (2.03^1.95 Ga) predatesthe recorded orogenic activity associated with thecollision of the Hottah terrane and Slave province(1.94^1.86 Ga) [13]. If the high mantle conductiv-ity truncated by the shear zone was caused by thesubduction of oceanic lithosphere it suggests that,either some subduction occurred before 1.94 Gaor else, there was relative motion at lithosphericdepth on the shear zone subsequent to 1.95 Ga.The most probable explanation is that the trunca-tion of the conductor is explained by the move-ment on the younger (ca. 1.8 Ga) MF.

8. Discussion

The MT responses de¢ne clear azimuthal de-pendence in the area of the GSLsz. The geoelec-tric strike changes from VN33³E at periods cor-

responding to the upper crust to N62³E at periodscorresponding to lithospheric penetration. Thegeoelectric strike in the crust is more north^souththan the regional (s 40 km scale) strike of theGSLsz south of Slave province which is N60³E[15]. However, the magnetic low also suggests anazimuth similar to the MT results (Fig. 7). Theresults suggest that the strike of the GSLsz whereit crosses Corridor 1A is locally close to N30³E.De£ections of the shear zone away from theregional N60³E azimuth are observed on theexposed portion of the GSLsz to the northwest.Hanmer [2] attributes such de£ections to latestage deformation wrapping around constric-tions.

There is the possibility that the magnetic anom-aly on Corridor 1A represents a transpressionaltransfer zone of the MF rather than the GSLsz.The transfer zones strike between N30³E toN50³E, consistent with the strike of the magneticanomaly. However, the 30 km width of the anom-aly, and its similarity with the magnetic anomalyof GSLsz on the exposed shield, together supportthe interpretation of the GSLsz as the sourcerather than the MF. The presence of an electri-cally resistive structure strengthens this interpre-tation, as it is more di¤cult to explain the ob-served high resistivity if it is associated withbrittle faulting.

The geoelectric strike at periods correspondingto signal penetration to lithospheric depths isN62³E. This orientation is parallel to both thelarger-scale strike of the GSLsz to the south ofthe Slave province and to the larger-scale strike ofthe MF. The result suggests the tectonic motionassociated with these faults has been recorded bythe resistivity structure of the mantle lithosphere.

In a companion study to the present one, Eatonet al. [12] resolve signi¢cant shear wave splittingnear the GSLsz with di¡erences in SKS arrivaltimes from di¡erent azimuths of 1.1^1.5 s. Thereis consistency between the fast axis at each sitewith an average direction of approximatelyN50³E (Fig. 7). This direction is sub-parallel tothe absolute plate motion direction of 225³ sug-gesting that the observed seismic anisotropy maybe due to present-day asthenospheric £ow. Thesmall obliquity between the directions is possibly



due to localized de£ections caused by topographyat the base of the lithosphere [12].

These shear wave splitting data are interpret-able in terms of two-layer seismic anisotropy[12] and the di¡erent geoelectric strike directionsdetermined at short and long periods (Fig. 3) pro-vide support for such a two-layer interpretation.Preliminary two-layer inversions of the seismicdata (D. Eaton, personal communication, 2001)suggest an upper layer with a polarization sub-parallel to the short-period MT strike at shortperiods and a lower layer with a polarizationoblique to both the long period MT strike andto absolute plate motion.

At sites near the Grenville Front in easternCanada, an average 23³ obliquity between SKSand MT responses has been interpreted by Ji etal. [45] as a kinematic indicator of dextral shear-ing. The seismic and MT anisotropies are thoughtto be caused by lattice-preferred and shape-pre-ferred orientation of mantle minerals respectively[45]. In the area of the GSLsz there is an obliquitybetween the average MT strike direction ofN62³E, at periods corresponding to mantle pene-tration, and the seismically determined SKS fastaxis for a single layer of N50³E. This obliquity isparticularly evident within the magnetic low inter-preted to represent the GSLsz and within themagnetic high to the northwest (Fig. 7). Thisobliquity we would interpret to be a consequenceof the seismic anisotropy being due to present-dayasthenospheric £ow and the MT strike being as-sociated with Paleoproterozoic lithospheric struc-tures.

However, if two-layer seismic analysis leads tothe seismic anisotropy being interpreted in termsof a lithospheric source, then comparison of theMT and seismic data will provide an importantassessment of the model of Ji et al. [45].

9. Conclusions

Analyses of the MT data collected on a pro¢lecrossing the GSLsz along the SNORCLE Trans-ect Corridor 1A have shown that the geoelectricstrike direction near the GSLsz varies with depthfrom VN33³E at periods less than 20 s corre-

sponding to the upper and middle crust toVN62³E at period range of 20^1500 s corre-sponding to the lower crust and lithospheric man-tle. These results and 2D modelling show that theGSLsz is a lithospheric-scale feature. Within thecrust the GSLsz forms a resistive zone at least 20km wide and at the surface is correlated with amagnetic low. The rocks within the GSLsz consistof greenschist to granulite facies mylonites. Thehigh resistivity (s 5000 6m) is interpreted to bedue to the resistive nature of the granitic protolithof the mylonites and the fact that the GSLsz aredominated volumetrically by rocks deformedwithin the ductile regime.

The geoelectric strike at crustal depths in thevicinity of the GSLsz is more north^south thanthe overall azimuth of the shear zone to the southof the Slave province (N60³E) but is consistentwith the strike of the magnetic anomaly inter-preted to correspond to the shear zone. This strikeis interpreted to re£ect a local-scale (6 50 km)de£ection of the mylonite belts. The geoelectricstrike at periods corresponding to lower crustand mantle lithosphere depths is N62³E. This ori-entation is close to the large-scale (s 50 km) azi-muth of the shear zone south of the Slave prov-ince. In the present study the long-period MTresponses were ¢tted quite well by a resistivitymodel consisting of 2D isotropic structures butthe data could also be ¢tted by a model incorpo-rating both anisotropic conductors and structuralelements.

Acknowledgements

The ¢nancial support for this project camefrom Natural Sciences and Engineering Councilof Canada grants to LITHOPROBE and fromthe Geological Survey of Canada (GSC). Thedata from the contracted MT V5 survey werecollected by Phoenix Geophysics. The LiMSdata were collected by Mr. Nick Grant, with as-sistance of sta¡ from the GSC, the University ofBritish Columbia, and the University of Manito-ba. We would like to thank Dr. Randy Mackiefor advice on the use of his 2D inversion pro-gram. We used the 2D Occam inversion program



from S. Constable and C. deGroot-Hedlin as im-plemented in the Geotools software package. Themagnetic data were provided by the GSC. PaulHo¡man, Malcolm Ingham, Juanjo Ledo, andan anonymous reviewer are thanked for their con-structive comments. LITHOPROBE contribution1261, GSC contribution 2001158.[RV]

References

[1] G.M. Ross, R.R. Parrish, M.E. Villeneuve, S.A. Bowring,Geophysical and geochronology of the crystalline base-ment of the Alberta Basin, western Canada, Can. J. EarthSci. 28 (1991) 512^522.

[2] S. Hanmer, Great Slave Lake shear zone, CanadianShield: reconstructed vertical pro¢le of a crustal-scalefault zone, Tectonophysics 149 (1988) 245^264.

[3] A.G. Jones, R.D. Kurtz, D.E. Boerner, J.A. Craven,G.W. McNeice, D.I. Gough, J.M. DeLaurier, R.G. Ellis,Electromagnetic constraints on strike-slip geometry ^ TheFraser River fault system, Geology 20 (1992) 561^564.

[4] A.G. Jones, D.I. Gough, Electromagnetic images of crus-tal structures in southern and central Canadian Cordil-lera, Can. J. Earth Sci. 32 (1995) 1541^1563.

[5] J. Ledo, A.G. Jones, I.J. Ferguson, Electromagnetic im-ages of a strike-slip fault: the Tintina Fault-northernCanadian Cordillera, Geophys. Res. Lett. (2002) in press.

[6] W.D. Stanley, V.F. Labson, W.J. Nokleberg, B. Csejtey,M.A. Fisher, The Denali fault system and Alaska Rangeof Alaska: evidence for underplated Mesozoic £ysch frommagnetotelluric surveys, Geol. Soc. Am. Bull. 102 (1990)160^173.

[7] R.L. Mackie, D.W. Livelybrooks, T.R. Madden, J.C.Larsen, A magnetotelluric investigation of the San An-dreas Fault at Carrizo Plain, California, Geophys. Res.Lett. 24 (1997) 1847^1850.

[8] M.J. Unsworth, P.E. Malin, G.D. Egbert, J.R. Booker,Internal structure of the San Andreas Fault at Park¢eldCalifornia, Geology 25 (1997) 359^362.

[9] M.J. Unsworth, G.D. Egbert, J.R. Booker, High-resolu-tion electromagnetic imaging of the San Andreas Fault inCentral California, J. Geophys. Res. 104 (1999) 1131^1150.

[10] M. Ingham, C. Brown, A magnetotelluric study of theAlpine Fault, New Zealand, Geophys. J. Int. 135 (1998)542^552.

[11] F.J. Davey, T. Henyey, W.S. Holbrook, D. Okaya, T.A.Stern, A. Melhuish, S. Henrys, H. Anderson, D. Eber-hart-Phillips, T. McEvilly, R. Uhrhammer, F. Wu, G.R.Jiracek, P.E. Wannamaker, G. Caldwell, N. Christensen,Preliminary results from a geophysical study across amodern, continent-continent collisional plate boundary;the Southern Alps, New Zealand, Tectonophysics 288(1998) 221^235.

[12] D. Eaton, I. Asudeh, I.A. Jones, Mantle strain beneaththe Great Slave Lake Shear Zone, NWT from measure-ments of seismic and electrical anisotropy, in: F. Cook, P.Erdmer (Eds.), Slave-Northern Cordillera LithosphericEvolution (SNORCLE) Transect and Cordilleran Tecton-ics Workshop Meeting, February, University of Calgary,AB, LITHOPROBE Report, 72, 2000, 56^63

[13] P.F. Ho¡man, Continental transform tectonics GreatSlave Lake shear zone (1.9 Ga), northwest Canada, Geol-ogy 15 (1987) 785^788.

[14] R.A. Gibb, Slave^Churchill collision tectonics, Nature271 (1978) 50^52.

[15] S. Hanmer, S. Bowring, O. Breemen, R. Parrish, GreatSlave Lake shear zone, NW Canada: mylonitic record ofearly Proterozoic continental convergence, collision andindentation, J. Struct. Geol. 14 (1992) 757^773.

[16] B.D. Ritts, J.P. Grotzinger, Depositional facies and detri-tal composition of the Paleoproterozoic Et-Then Group,NWT, Canada: sedimentary response to intracratonic in-dentation, Can. J. Earth Sci. 31 (1994) 1763^1778.

[17] F.A. Cook, A.J. VanderVelden, K.W. Hall, B.J. Roberts,Tectonic delamination and subcrustal imbrication of thePrecambrian lithosphere in northwestern Canada mappedby LITHOPROBE, Geology 26 (1998) 839^842.

[18] F.A. Cook, A.J. VanderVelden, K.W. Hall, B.J. Roberts,Frozen subduction in Canada's Northwest TerritoriesLithoprobe deep lithospheric re£ection pro¢ling of thewestern Canadian Shield, Tectonics 18 (1999) 1^24.

[19] G.M. Ross, H.C. Annell, M.A. Hamilton, Lithology andU^Pb geochronology of shallow basement along theSNORCLE line, southwest Northwest Territories: A pre-liminary report, in: F. Cook, P. Erdmer, (Eds.), Slave-Northern Cordillera Lithospheric Evolution (SNORCLE)Transect and Cordilleran Tectonics Workshop Meeting,February, University of Calgary, AB, LITHOPROBE Re-port, 72, 2000, 76^80.

[20] R.J. Theriault, Nd isotopic evidence for Paleoproterozoicpre-Taltson magmatic zone (1.99^1.90 Ga) rifting of west-ern Churchill Province, in: G.M Ross (Ed.) LithoprobeAlberta basement transects workshop meeting, February,Ramada Hotel, Calgary, AB, LITHOPROBE Report, 37,1994, 267^269.

[21] D. Rhodes, E.A. Lantos, J.A. Lantos, R.J. Webb, D.C.Owens, Pine Point orebodies and their relationship to thestratigraphy, structure, dolomitization, and karsti¢cationof the Middle Devonian Barrier Complex, Econ. Geol. 79(1984) 991^1055.

[22] X. Wu, Determination of near-surface, crustal, and litho-spheric structure in the Canadian Precambrian Shield us-ing the time-domain electromagnetic and magnetotelluricmethods, Ph.D. Thesis, University of Manitoba, Winni-peg, MB, 2001.

[23] A.G. Jones, H. Jo«dicke, Magnetotelluric transfer functionestimation improvement by a coherence-based rejectiontechnique, Contributed paper at 54th Ann. Int. Meeting,Soc. Explor. Geophys., Atlanta, GA, Dec. 2^6, 1984, Ex-panded abstracts 51^55.



[24] A.G. Jones, A.D. Chave, G. Egbert, D. Auld, K.K. Bahr,A comparison of techniques for Magnetotelluric responsefunction estimation, J. Geophys. Res. 94 (1989) 14201^14213.

[25] A.D. Chave, D.J. Thomson, Some comments on magne-totelluric response function estimation, J. Geophys. Res.94 (1989) 14215^14225.

[26] X. Garcia, A.D. Chave, A.G. Jones, Robust processing ofmagnetotelluric data from the auroral zone, J. Geomagn.Geoelectr. 49 (1997) 1451^1468.

[27] G.W. Groom, R.C. Bailey, Decomposition of magneto-telluric impedance tensors in the presence of local three-dimensional galvanic distortion, J. Geophys. Res. 94(1989) 1913^1925.

[28] R.W. Groom, R.D. Kurtz, A.G. Jones, D.E. Boerner, Aquantitative methodology to extract regional magnetotel-luric impedances and determine the dimension of the con-ductivity structure, Geophys. J. Int. 115 (1993) 1095^1118.

[29] U. Schmucker, J. Janknowsky, Geomagnetic inductionstudies and the electrical state of the upper mantle, Tec-tonophysics 13 (1972) 233^256.

[30] G. McNeice, A.G. Jones, Multisite, multifrequency tensordecomposition of magnetotelluric data, Geophysics 66(2001) 158^173.

[31] A.G. Jones, Static shift of magnetotelluric data and itsremoval in a sedimentary basin environment, Geophysics53 (1988) 967^978.

[32] B.K. Sternberg, J.C. Washburne, L. Pellerin, Correctionfor the static shift in magnetotellurics using transient elec-tromagnetic soundings, Geophysics 53 (1988) 1459^1468.

[33] K. Vozo¡, The magnetotelluric methods, in: M.N. Na-bighian (Ed.), Electromagnetic Methods in Applied Geo-physics, vol. 2 Applications, Society of Exploration Geo-physicists, Tulsa, OK, 1991, 641^711.

[34] C. deGroot-Hedlin, S. Constable, Occam's inversion togenerate smooth, two-dimensional models from magneto-telluric data, Geophysics 55 (1990) 1613^1624.

[35] W. Rodi, R.L. Mackie, Nonlinear conjugate gradientsalgorithm for 2-D magnetotelluric inversion, Geophysics66 (2001) 174^187.

[36] A.G. Jones, The problem of `current channeling': a crit-ical review, Geophys. Surv. 6 (1983) 79^122.

[37] P.E. Wannamaker, G.W. Hohmann, S.H. Ward, Magne-totelluric responses of three-dimensional bodies in layeredearths, Geophysics 49 (1984) 1517^1533.

[38] T.J. Katsube, M. Mareschal, Petrophysical model of deepelectrical conductors: graphite lining as a source and itsdisconnection due to uplift, J. Geophys. Res. 98 (1993)8019^8030.

[39] A.G. Jones, J.A. Craven, The north American centralplains conductivity anomaly and its correlation with grav-ity, magnetic, seismic, and heat £ow data in Saskatche-wan, Canada, Phys. Earth Planet. Inter. 60 (1990) 169^194.

[40] J.C. Gupta, A.G. Jones, Electrical conductivity structureof the Purcell Anticlinorium in southeast British Colum-bia and northwest Montana, Can. J. Earth Sci. 32 (1995)1564^1583.

[41] A.G. Jones, J. Katsube, P. Schwann, The longest conduc-tivity anomaly in the world explained: sulphides in foldhinges causing very high electrical anisotropy, J. Geo-magn. Geoelectr. 49 (1997) 1619^1629.

[42] G.R. Olhoeft, Electrical properties of granite with impli-cations for the lower crust, J. Geophys. Res. 86 (1981)931^936.

[43] B.W.D. Yardley, Is there water in the deep continentalcrust?, Nature 323 (1986) 111.

[44] M.G. Bostock, J.F. Cassidy, Upper mantle stratigraphybeneath the southern Slave craton, Can. J. Earth Sci. 34(1997) 577^587.

[45] S. Ji, S. Rondenay, M. Mareschal, G. Senechal, Obliquitybetween seismic and electrical anisotropies as a potentialindicator of movement sense for ductile shear zones in theupper mantle, Geology 24 (1996) 1033^1036.



Documents

Magnetotelluric response and geoelectric structure of the Great Slave Lake shear zone