This article was downloaded by: [University of Texas Libraries]On: 26 November 2014, At: 03:45Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

New Zealand Journal of Geology and GeophysicsPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/tnzg20

Upper crustal structure beneath the easternSouthern Alps and the Mackenzie Basin, NewZealand, derived from seismic reflection dataD. T. Long a d , S. C. Cox a , S. Bannister a , M. C. Gerstenberger b & D. Okaya ca Institute of Geological & Nuclear Sciences , P.O. Box 30 368, Lower Hutt, New Zealandb Swiss Seismological Service , ETH‐Hoenggerberg/HPPP , Zurich, CH 8093, Switzerlandc Institute of Geophyics , University of Southern California , Los Angeles, CA,90089–0740, USAd Petroleum Geo‐Services , Level 4, IBM Centre, 1060 Hay St, West Perth, WA, 6005,AustraliaPublished online: 21 Sep 2010.

To cite this article: D. T. Long , S. C. Cox , S. Bannister , M. C. Gerstenberger & D. Okaya (2003) Upper crustal structurebeneath the eastern Southern Alps and the Mackenzie Basin, New Zealand, derived from seismic reflection data, NewZealand Journal of Geology and Geophysics, 46:1, 21-39, DOI: 10.1080/00288306.2003.9514993

To link to this article: http://dx.doi.org/10.1080/00288306.2003.9514993

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose ofthe Content. Any opinions and views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be reliedupon and should be independently verified with primary sources of information. Taylor and Francis shallnot be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

New Zealand Journal of Geology & Geophysics, 2003, Vol. 46: 21-390028-8306/03/4601-0021 $7.00/0 © The Royal Society of New Zealand 2003

21

Upper crustal structure beneath the eastern Southern Alps and the MackenzieBasin, New Zealand, derived from seismic reflection data

D. T. LONG*

S. C. COX

S. BANNISTER

Institute of Geological & Nuclear SciencesP.O. Box 30 368Lower Hutt, New Zealand

M. C. GERSTENBERGER

Swiss Seismological ServiceETH-Hoenggerberg/HPPPCH 8093, Zurich, Switzerland

D. OKAYA

Institute of GeophyicsUniversity of Southern CaliforniaLos Angeles, CA 90089-0740, USA

*Present address: Petroleum Geo-Services, Level 4, IBMCentre, 1060 Hay St, West Perth, WA6005, Australia.

Abstract A 65 km seismic reflection transect was shot in1998 across Mackenzie Basin to Mount Cook Village, SouthIsland, New Zealand, to provide a detailed image of thecrustal structure in the central Southern Alps. The first 5 s(two-way time (TWT)) of data were processed separately,with a maximum offset of 14 km for each shot, to image theupper 12-15 km of the crust. Data were processed as onecontinuous section, although the line was physicallysegmented due to an area of relatively steep topography withno vehicular access. No major, continuous regional-scalefeatures > 10 km are present in the data, but numerous 2-3km scale reflections and discontinuities occur which areconsistent with the known geology of monotonousgreywacke sequences overlying schist. Strong, well-definedreflections mark the active Irishman Creek Fault and confirmit to be a southeasterly dipping reverse fault with c. 1300-1700 m of Late Cretaceous-Pleistocene sediments preservedin the footwall and an uplifted greywacke basement "high"in the hanging wall. Some evidence exists for active faultsbeneath latest Quaternary gravels at the Jollie valley andTekapo River. Oppositely dipping reflections anddiscontinuities define a large, c. 15 km wavelength antiformbeneath Tasman valley and Mount Cook that is imaged to10 ± 2 km depths (3.5 s TWT).

Two "end-member" interpretations are consistent withthe seismic data observations, velocity models, andconstraining features of exposed geology, and extendexisting geological cross-sections to 10-15 km depth. Oneinterpretation assumes imaged structures are primarily

G01035; published 21 March 2003Received 29 October 2001; accepted 24 September 2002

backthrusts developed in response to distributed Cenozoicdeformation southeast of the Alpine Fault plate boundary,incorporating features observed in contemporary geodeticstrain and numerical plate boundary models. The secondinterpretation assumes structures are mostly Mesozoic, eitherreactivated or preserved by late Cenozoic deformation. Themain difference between the interpretative cross-sections isthe degree to which active structures link into basaldetachment and high-strain zones at depth.

Keywords Southern Alps; seismic reflection; SIGHT;Mackenzie Basin; faults; tectonics; deformation

INTRODUCTION

The South Island Geophysical Transect (SIGHT) project isa multidisciplinary, multinational study of the South Islandof New Zealand with the purpose of obtaining a betterunderstanding of continent-continent collisional processesacross the Australian-Pacific plate boundary. During 1998,a detailed seismic reflection survey (called SIGHT98) wascarried out in the Lake Pukaki region (Fig. 1) as part of theSIGHT project complementing earlier experiments done in1995 and 1996 (Davey et al. 1995, 1998; Smith et al. 1995;Kleffman et al. 1998). The reflection survey follows a 65km transect nearly perpendicular to the Alpine Fault plateboundary. The survey line extends from Mt Cook Village(26 km southeast of the Alpine Fault), down the Tasmanvalley, across the Mackenzie Basin to Burke Pass (Fig. 1).The line was positioned to traverse a prominent negativegravity anomaly (c. 100 mgal) thought to be caused bycrustal thickening and development of a root beneath theSouthern Alps (Stern 1995).

The principal objective of SIGHT98 was to image theMoho and crustal root. Survey parameters were chosen toobtain the best possible information of these features,expected to occur at depths around 40 km and greater. Thedata, however, also contain information on the shallow crust,albeit not of the quality that would be returned from a surveydesigned specifically for the purpose of imaging shallowfeatures. This present study examined seismic reflections inthe first 5 s two-way time (TWT) of the data, correspondingto the uppermost c. 12-15 km of crust.

Interpretation was constrained by geologicalobservations in rocks exposed nearby, and by modelling ofseismic velocities in the uppermost 2-3 km. The studyprovides the first detailed seismic image of upper crust inthe central Southern Alps region. This paper describesobservations of features in the data and separates these fromtwo interpretive models/explanations. The proposed crustalarchitecture presented by this paper provides importantconstraints on the mode by which distributed plate-boundarydeformation may be occurring southeast of the AlpineFault.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

22 New Zealand Journal of Geology and Geophysics, 2003, Vol. 46

Fig. 1 Geological map of centralSouth Island and location of theSIGHT98 seismic reflectionsurvey. The dashed line traces thedistribution of shots and receiversduring seismic data acquisition,whereas the solid line representswhere the acquired data wereprocessed to form a single line.Major faults and the strike ofsteeply dipping bedding inbasement rocks (white formlines)are also shown. Inset: The AlpineFault offset and distributed dextralbending of the Maitai Terrane(M).

I Laic O-nlacoaus Miotcnn

faull -actvE cLr* or Mismt ahals

IkiulCDP

TECTONICS AND GEOLOGY

Tectonic setting

Terranes in the South Island of New Zealand have been bothoffset and bent by the Pacific-Australia plate boundary(Norris 1979; Spörli 1979; Sutherland 1995, 1996, 1999;Molnar et al. 1999; Mortimer et al. 1999). Plate recon-struction constrained by satellite altimetry suggests the 440-470 km offset on the Alpine Fault accounts for c. 55% ofthe late Eocene-Recent plate displacement, with c. 45%

accommodated by distributed dextral shear (Sutherland1999). Late Quaternary strike-slip displacement rates alongthe Alpine Fault (27 ± 5 mm/yr) make up 70-75% of thefault-parallel interplate motion, and are relatively constantcompared with dip-slip rates (0 to >10 mm/yr) which aregreatest adjacent to the highest mountains (Norris & Cooper2001). Geodetic surveys show that rocks in the immediatevicinity of the Alpine Fault are currently storing elastic strainenergy that corresponds to 50-70% of the plate motion rate(Pearson 1994; Pearson et al. 1995; Beavan et al. 1999).

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island 23

Importantly, a significant component of both long-term andcontemporary deformation appears to be distributed in theSouthern Alps beneath the seismic reflection line.

Quaternary slip is clearly demonstrated on structureswithin the Mackenzie Basin and along the southeasternmargin of the Southern Alps (e.g., Ostler and Irishman CreekFaults). Uplift rates calculated from faulted and foldedglacial and postglacial landforms correspond closely withthose measured by geodetic survey (0.4-1 mm/yr) (e.g.,Blick et al. 1989; Van Dissen et al. 1993), but these are arelatively insignificant component of the total plate motion(Lensen 1975; Walcott 1998). A major backthrust near thedrainage divide crest of the alps, the Main Divide Fault Zone,juxtaposes rocks with different thermochronological agesindicating significant late Cenozoic vertical displacement(<5 Ma) (Cox & Findlay 1995). However, definitiveevidence for active fault slip and slip-rates within the centralSouthern Alps is lacking—probably due to extreme erosionrates and consequent lack of preserved tectonic features. Thepresence of elevated topography and earthquakes providesindirect yet strong evidence that at least some of the observedcontemporary deformation must be permanent in this region(Pearson 1993; Leitner et al. 2001). Estimated uplift andexhumation rates vary across the Southern Alps, from c. 1mm/yr in the southeast near Lake Pukaki (beneath theseismic line), to c. 6-10 mm/yr immediately adjacent to theAlpine Fault where the rainfall and erosion rates are muchhigher (e.g., Wellman 1979; Adams 1980a,b; Simpson et al.1994).

Irishman Creek Fault is the only fault with a known,active surface trace that is traversed by the seismic line (Fig.1). It has uplifted and tilted Pliocene-Pleistocene sedimentsand displaces an old glacial outwash surface (Maizels 1989;Black 2000; Mildenhall 2001). Irishman Creek Fault isinterpreted to be a reverse fault dipping c. 55° to the southeastwith a vertical slip rate of c. 0.6 mm/yr (Fox 1987; Chetwin1998). It is possible that other active faults and folds lieburied beneath young glacial deposits with no surfaceexpression. One objective of this study was to delineate anysuch faults in the subsurface.

Permian-Triassic basement lithology and structureRocks forming the mountain ranges either side of the seismicline are mostly mid-late Triassic grey wackes of the TorlesseTerrane, with Permian sequences occurring in the south andeast about Burke Pass (Gair 1967). They comprisecompositionally monotonous sandstone, interlayered withsiltstone and argillite in submarine-fan-type sequences, withrare red and black argillites and conglomeratic units. Beddingis typically <1-50 m scale, with large-scale variations inlithological proportions at c. 1 km and c. 3 km scales that inplaces are pronounced enough to define lithologicalassociations or units (e.g., Spörli et al. 1974; Hicks 1981;Oliver et al. 1982).

Bedding in the vicinity of the seismic line is moderateto steeply dipping (60-90°). A marked swing in averagestrike occurs along the line from dominantly NW-SE in thesoutheast, through N-S to NNE-SSW in the northwest (Fig.1). The swing in strike matches the geometry and bendingof terrane boundaries in the southern part of the South Island(Fig. 1 inset) (Spörli 1979). By analogy with argumentspresented by Spörli (1979) and Sutherland (1999), basementrocks at the southeastern end of the seismic line are near to

their original Mesozoic orientation, whereas those closer tothe Alpine Fault probably have a Cenozoic overprintassociated with distributed plate-boundary deformation. Inaddition to regional variation, bedding swings in orientationlocally at 1-4 km scales about steeply plunging folds andacross faults (Fig. 1) (Spörli & Lillie 1974; Spörli et al. 1974;Findlay & Spörli 1984). There is considerable variation inthe angular relationship between basement rocks and theseismic line as a result of (1) regional bending, (2) kilometre-scale folds, and (3) subtle changes in direction of the seismicline. Parts of the seismic line are oriented at a high angle tobedding, whereas others are nearly parallel, which may inpart explain variations in the signal:noise ratio in thereflection data.

Faults are abundant within the Torlesse greywackes (e.g.,Spörli & Lillie 1974; Cox & Findlay 1995). Many aresubparallel to bedding, others result in small angulardiscordance between beds, and less commonly major faultsresult in marked changes in orientation of bedding. Themajority of regional-scale faults exposed in mountain rangeseither side of the seismic line are moderate or steeplydipping, with NNE-SSW striking traces (Fig. 1) (Cox &Findlay 1995). Faults are commonly spaced (horizontally)with frequencies c. 2 and c. 6 km, resulting in relativelycontinuous, coherent blocks that are c. 1 and c. 5 km thick(Fig. 1). The age of faults is poorly known, and can generallyonly be inferred depending on the nature of fault rocks (e.g.,assemblages of mineral growth, or clay gouge, breccia andcataclasite versus more ductile fabrics) (Spörli 1979). Thefaults show a complete range in age from syn-sedimentary/syn-accretionary faults to others associated with Mesozoictectonics and metamorphism, to more recent faults associatedwith late Cenozoic uplift of the Southern Alps. Evidencefor late Cenozoic fault reactivation is widespread (Spörli1979; MacKinnon 1983) with clay-rich pug zones andcataclasites overprinting earlier Mesozoic fault rocks withmetamorphic quartz+chlorite assemblages.

Differential erosion and exhumation across the SouthernAlps have exposed metamorphic transitions across the SouthIsland that are almost parallel to the Alpine Fault (Grapes &Watanabe 1992). Non-schistose prehnite-pumpellyite faciesgreywackes form the mountains in the southeast and the areabeneath the seismic line, pass through to greenschist faciessemischists at and immediately west of the Main Divide,and then to amphibolite facies schists immediately adjacentto the Alpine Fault (Fig. 1). The geology across the SouthIsland approximates a structurally disrupted section downthrough the upper crust, rotated by Cenozoic collision, withthe deepest rocks exposed beside the Alpine Fault in theregion of greatest exhumation and total uplift.

The geological section beneath the seismic line can beinferred from differentially uplifted rocks exposed acrossthe South Island, by applying a knowledge of expectedpressure and temperature conditions at different meta-morphic grades (Grapes & Watanabe 1992) and assuming ageothermal gradient (e.g., 30°C/km). The expected verticalprofile of different rock types is outlined in Table 1, withdepths recalculated into seismic two-way travel times usinga 5.0-6.5 km/s velocity range that has been experimentallymeasured for these rock types (Garrick & Hatherton 1973;Okaya et al. 1995; Godfrey et al. 2000). Fission tracks inapatite and zircon help to constrain the section in Table 1.Basement rocks at the northwesternmost end of the seismic

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

24 New Zealand Journal of Geology and Geophysics, 2003, Vol. 46

line (Mt Cook Village) contain apatites with totally healedfission tracks indicating rocks were at >125°C (Gallagheret al. 1998) or depths of 3-5 km (corresponding to 1-2 sdepths TWT) before late Cenozoic uplift (Tippett & Kamp1993a,b). Apatite fission tracks are partially annealed inrocks for the next 10 km southeast (i.e., uplifted from <125°Ctemperatures and slightly shallower depths) (Tippett & Kamp1993a,b). Annealed zircons occur in the hanging wall of theMain Divide Fault Zone, only 3 km northwest of the seismicline (Cox & Findlay 1995). Late Cenozoic exhumation ofthese rocks has occurred from c. 175 to 250°C (Gallagheret al. 1998) (i.e., depths of c. 5 km or seismic depths of c.1.3-2.8 s TWT, assuming 30°C/km geothermal gradient,velocities of c. 5.5-6.0 km/s). A considerable shallowing ofthe geological section in Table 1 is expected toward thenorthwestern end of the seismic line.

Belts of cleaved greywacke and semischist c. 2 km thickalso crop out near the southeastern end of the seismic linenear Burke Pass in the Rollesby and Two Thumbs Ranges,where they are unconformably overlain by Late Cretaceous-Miocene sediments. The unconformable relationshipindicates some semischist belts southeast of the Main Dividewere uplifted before the Tertiary, presumably associated withMesozoic tectonics.

Late Cretaceous-Miocene sequenceA cover sequence of Late Cretaceous-Miocene sedimentaryrocks is exposed east of the seismic line forming theCanterbury Basin of South Canterbury and North Otago,which lies unconformably on both Torlesse greywackes andsemischist (Field et al. 1989). In the vicinity of the seismicline, there are isolated non-marine remnants preserved infault zones but no significant exposures of the sequence (Gair1978). However, there is potential for Late Cretaceous-Miocene sediments to underlie the seismic line in MackenzieBasin, buried beneath Pliocene-Pleistocene and lateQuaternary fluvioglacial gravels and moraines.

The sequence in South Canterbury to North Otagorecords Late Cretaceous-Oligocene transgression followedby a Miocene regression in a passive margin setting. Thetotal thickness of the sequence does not exceed 1 km andappears to thin from east to west (Gair 1978; Field et al.1989). If Late Cretaceous-Miocene sediments are presentbeneath the seismic line, it is likely they will be less than1 s TWT thick (assuming 2 km/s minimum velocity, 1 kmmaximum thickness), very probably <0.5 s TWT thick.

Pliocene-Pleistocene gravelsPassive margin deposition ceased in the South Canterburyregion with the influx of terrestrial Pliocene-Pleistocenegravels of the Kurow Group (Field & Browne 1986), whichmark the onset of uplift and erosion of the Southern Alps.In the vicinity of the seismic line, Kurow Group sedimentscomprise: (1) basal clay-rich carbonaceous mudstones,pebbly sandstones, and coal measures, formed in lacustrineand/or meandering river setting with little topographic relief;overlain by (2) coarse-grained, fluvial and fluvioglacialgravels interbedded with minor lithic-rich sandstone beds(Black 2000; Mildenhall 2001). These terrestrial sedimentsare typically moderately well indurated or cemented, folded,and faulted to moderate and locally steep dips. There are nocontinuous stratigraphic sections exposed near the seismicline. Prior to faulting, folding, and erosion, Kurow Groupsediments probably formed a relatively continuous, butlaterally variable, sequence across the Mackenzie Basinanalagous to the current deposition of fluvioglacial gravels,tills, and lacustrine sediments. Local, incomplete sectionsare up to 200 m thick, but based on exposures elsewhere inCanterbury (Field & Browne 1986), the Kurow Group mayhave locally exceeded 500 m. In the absence of detailedpalynological studies, tilted dip is the primary method ofdistinguishing these Pliocene-Pleistocene gravels fromyounger, late Quaternary sediments, which are lithologicallyidentical.

Table 1 Geological section beneath the seismic line inferred from differentially uplifted rocks exposed across the South Island, byapplying a knowledge of expected pressure and temperature conditions at different metamorphic grades (Grapes & Watanabe 1992)and assuming a geothermal gradient (e.g., 30ºC/km). Depths are recalculated into seismic two-way travel times using the 5.0-6.5 km/svelocity range experimentally determined for these rocks.

Metamorphism and modeof deformation Dip and structure

Expected depth(TWT)

Greywacke

Semi-schist

Low-grade schist

High-grade schist

Prehnite-pumpellyite facies,220-320°C. Base of apatitefission track annealing zonec. 125°C (c. 1-2 s TWT).Brittle faulting.

Pumpellyite-actinolite facies.Base of zircon fission trackpartial annealing c. 175-250°C(c. 2-3 s TWT). Faulting andsolution transfer.

Greenschist facies, 320–420°C.Fabrics mostly the result ofdeformation by solution transfer.

Biotite zone, greenschist facies,420–480°C. Fabrics predominantlythe result of deformation bycrystal-plastic recrystallisation.

Steep or moderate dip. Folds 2-6 kmacross with isolated hinges. Faultsspaced 1 or 5 km.

Cleavage and bedding mostly parallel.Homoclinal with moderate or shallowdip. Possible open folds at 2 km. Mayhave numerous brittle shears.

Flat-lying foliation? Numerous smallscale buckle folds. Larger folds andshears expected.

Flat lying schistosity with shear zones.

Surface-2 s

1.5-3 s

>4s

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island 25

Late Quaternary fluvioglacial gravelsLate Quaternary alluvium and glacial tills cover the regioncrossed by the seismic line. The deposits can be highlylocalised with considerable lateral variation of internalcharacter. They record at least five glacial advances, as tills,alluvial outwash, and lake sediments (Speight 1963; Gair1967; Maizels 1989; Suggate 1990).

The thickness of late Quaternary sediment is poorlyconstrained. In the Tasman and Hooker valleys, it is unlikelyto be >800 and 300 m, respectively, based on valley widthand projected mountain slope profiles. A velocity modelusing arrival times from wide-angle reflections in refractionseismic data suggested these gravels thicken significantlyalong the shores of Lake Pukaki, from c. 600 m at the northend to c. 800 m (just below sea level) at the south end of thelake (Kleffman et al. 1998). Thickness in the MackenzieBasin is more uncertain, but is expected to be <1 km, basedon topography and sediment fill in the glacially carved lakes(Pickerill & Irwin 1983; Kleffman et al. 1998; P. Upton pers.comm).

DATA AND OBSERVATIONS

Seismic data acquisition and processingThe seismic data were acquired by GECO-Prakla in February1998, along the northwest-southeast survey line shown inFig. 1. The survey line was divided into two sections, witha gap occurring in an area of steep topography where novehicular access was possible. A total of 181 shots weredetonated along the survey line, shown as dashed lines inFig. 1. Large 50 kg powergel shots were placed at 1 kmintervals and buried 20 m deep. Smaller shots of 2.5 kg,buried 3 m deep, were located 250 m on either side of eachlarge shot. Shot triggering was keyed on GPS time, to allowsecondary piggy-back arrays to trigger without the need fordirect radio communication. A total of 1000 geophonereceiver group positions were used to record the shots, witha group spacing of 40 m. The number of recording channelsfor any one shot varied from a minimum of 400 to amaximum of 870. The maximum array aperture was nearly35 km and the maximum shot receiver offset c. 30 km,primarily targeting deep crustal structure. The raw seismicdata for each shot were recorded to 30 s time, at a 2 mssampling rate.

Processing of the seismic data followed the sequenceshown in Table 2. Data were processed as one continuoussection by tracing the highest concentration of commonmidpoints (Fig. 1, solid line). Since the seismic surveygeometry was primarily designed to image deep structure(c. 30–40 km depth) beneath the Southern Alps, it was achallenge to retrieve clear consistent shallow reflections inthe upper 5 s of data. There were also some attenuationproblems resulting from the Quaternary sedimentary cover.For this study of the shallow structure, the recorded seismicdata were re-sampled at the start of processing, to involveonly data with a maximum offset of 14 km from each shot.

Noisy traces and refracted energy were muted inter-actively, together with the airwave from the shots. Staticcorrections were also applied to correct for elevation andsurface layer variation. Additional pre-stack processingincluded application of Wiener deconvolution, and afrequency domain bandpass filter. The frequency filtering

and deconvolution parameters were determined using paneltests. Velocity analysis was carried out using both semblanceand constant velocity stacks and was performed iterativelybefore and after residual static determination. The finalstacked section is shown in Fig. 2. The final stack wassubjected to different sequences of post-stack processing,each providing enhancement of a particular feature (Fig. 3,4). Interpretations were made using a combination of thesepost-stack processed sections.

Observations from seismic dataThe experiment has not imaged any major regional-scalereflections or discontinuities in the uppermost 5 s TWT. Thislack of regional-scale features is consistent with the knowngeology of the shallow crust, composed of steeply dippinggreywacke sequences overlying schists that have fewvariations in lithology and seismic velocity. The experimentdid, however, image a number of subtle smaller scalefeatures. Three types of observations were mapped on thedata:

(1) Strong, well-defined reflections (thick, solid lines in Fig.2C, 3,4), which are laterally continuous for 2-3 km andcan be observed in all post-stack processing routines.Some of these can also be seen in the original shot data;

(2) Strong discontinuities in the data (medium, solid linesin Fig. 2C, 3, 4), and trends in reflections/reflectivitythat have a marked change in orientation. The discon-tinuities can generally be observed in all post-stackprocessing routines;

(3) Relatively weak, discontinuous reflections and trends inthe data (thin, dotted lines in Fig. 2C; thin, solid lines inFig. 3, 4), which can only be observed in some selectedpost-stack processing routines. They are more subjective,but can be related to known features observed in thegeology.

The well-defined reflections and discontinuities aremostly continuous only over 2-3 km and no features couldbe traced with confidence for more than 10 km. Thereflections and discontinuities are thought to image mostly(1) bedding or schistosity in Torlesse greywackes and schists,particularly where kilometre-scale facies changes andcontacts occur between sandstone- or mudstone-dominated

Table 2 Seismic data processing sequence. Note that minoradditional filtering may have occurred during electronicreproduction for diagrams in this paper.

Demultiplex of field recordsGeometry inserted into trace headersElevation and refraction-layer static applicationEdit of noisy tracesMute of shot-related airwaveOffset-varying Wiener deconvolution (Filter length 500 ms,

gap 24 ms).Spatially varying bandpass filter: (9-13-40-50 Hz)Surgical mute of first break and refraction energyIterative velocity analysis using semblance and CVS panelsIterative residual static calculation and application500 ms AGC as trace amplitude equalisationNMO application51-fold normalised stackApplication of a (FX-domain) post-stack coherency filter1000 ms AGC before final plotTrace sum (2:1) for plotting

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

NWM M inin

SE

toON

n-Kiunlan rftrtgfta

StiJUlle l|[^j[t¥afay

%

5 kmFig. 2 A, Elevation of the seismic reflection line with respect to mean sea level and theprofile of nearby mountain ranges. Note the 1.5x vertical exaggeration and the gap/stepoverwhere there was no vehicular access and the line was divided into two sections. B, Seismicreflection data with no post-stack processing applied. Continous reflections anddiscontinuities are mostly only 2-3 km across and no features can be traced with confidencefor more than 10 km. The strongest reflections occur near the Irishman Creek Fault. C, Observations from seismic reflection section. Thick solid lines represent well-defined reflections. Thinsolid lines represent strong discontinuities in the data. Thin dotted lines follow relatively weak, discontinuous reflections and trends in the data.

StrariQ, well-defined reflectionSlrong, wen-dedred.djscon-irnultyWeak, diaoantinuaus reilections 3 dala lrarcds

§

aoo

9f

<

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island

CDP150Q 1300 1100

27

15OT 1100

1 ,

3 -*-±v.V -1

- .

CDP150Q 13C0 i ; no

•

- - %;4l1500 13OD T100

r\ Irish man Creek— := Fault

well-defined,— — StFong, well-defined, discontinuity——— Weak, discontinuous rsfteajons &. data trends

Fig. 3 Seismic reflection data across Irishman Creek Fault in four NW-SE sections displayed with different post-stack plottingparameters. A, Wiggle trace plot with relatively high gain, low percentage bias shading. B, Wiggle trace plot with medium gain,high percentage bias shading. C, Variable area plot with low percentage bias shading. D, Variable area plot with high percentagebias shading, overlain by key observations used in the interpretation.

units, or (2) faults that juxtapose rocks of differentmetamorphic grade, lithology, or orientation.

At the northwestern end of the line, beneath the Tasmanvalley, there is an antiformal pattern defined by a changefrom northwesterly to southeasterly dipping reflections anddiscontinuities (Fig. 4). The crest of the "antiform" occursat c. 1.2 s TWT corresponding to c. 3 km in depth-convertedsections. The feature can be observed at least as deep as 3.5s in the data (10 ± 2 km depth) and appears to have ahorizontal half-wavelength of c. 15 km. The antiform shape

(shown with respect to two-way time in Fig. 4, not truedepth) cannot be attributed to "pull-up" or large-scale localvariation in seismic velocity, because cover sediments arerestricted to thin Quaternary gravels of relatively uniformthickness at a regional-scale. Instead, the antiform is thoughtto be a real feature representing a change in the geometry ofrock sequences and faults.

In the centre of the line, there are some prominentsoutheastward-dipping reflections and discontinuities thatextrapolate up to the surface trace of the Irishman Creek

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

28 New Zealand Journal of Geology and Geophysics, 2003, Vol. 46

3OQ0 23W £600

: V

Strong. weJI-defirved. refledtarSErang. wall-deftrred. discontinuityWeak, discontinuous relffldmns & dala trends

5 km

Fig. 4 Seismic reflection data at the northwestern end of the line, from Hooker valley across Tasman valley to the mouth of the Jollievalley. Two variable area NW-SE sections are displayed with different plotting parameters. A, Low percentage bias shading. B, Highpercentage bias shading, overlain by key observations used in the interpretation.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island 29

Fault trace (Fig. 3). The strong reflections are thought toimage the fault plane and sediments dipping toward thesoutheast, confirming the dip direction suggested by the faultscarp and gravity modelling (Fox 1987; Chetwin 1998). Thestrong reflectivity is likely due to marked changes of velocitybetween fluvioglacial gravel deposits, inferred LateCretaceous-Miocene sediments, and basement Torlessegreywacke (see Smith et al. 1995; Langdale & Stern 1998and references therein).

Individual discontinuities in the uppermost 3 s TWT alsoshow some marked small-wavelength changes in orientation.The most prominent occur immediately northwest and belowthe Irishman Creek Fault and 10 km southeast of theIrishman Creek Fault where discontinuities and reflectionsthat appear to trace prominent fold-like patterns at both 1-2 and c. 5 km scales (Fig. 2). On the whole, however, theobserved features are relatively planar, and those in the 3-5 srange, albeit more poorly defined, have an appearance ofbeing relatively flat-lying compared with the shallowerequivalents. The only well-defined features observed in thedeeper data are a series of strong reflections at the verysoutheastern part of the line between 3 and 4 s (Fig. 2) whichcorrespond to features at 9-12 km depth—approximatelythe depth of the base of the present-day seismogenic zonein this area (Leitner et al. 2001).

VELOCITY MODELLING

Background and methodsLate Quaternary and Pliocene-Pleistocene gravels arelithologically indistinguishable and are unlikely to be ableto be distinguished in terms of seismic velocity. The otherprincipal geological units traversed by the seismic line,however, appear to have distinct seismic velocities. LateQuaternary and Pliocene-Pleistocene fluvioglacial graveldeposits have measured and interpreted velocities in therange of 2-2.6 km/s at 500-1500 m depth (Smith et al. 1995;Stagpoole 1997; Langdale & Stern 1998). Late Cretaceous-Miocene sediments have measured and interpreted velocitiesin the range 2.1-3.5 km/s at 500-1500 m depth. Incomparison, basement Torlesse greywackes and schists havebeen measured in the laboratory to have velocities of 5.0-6.5 km/s (Smith et al. 1995; Okaya et al. 1995; Stagpoole1997; Langdale & Stern 1998; Godfrey etal. 2000). Becausethe seismic velocities of the formations are relatively distinct,velocity models can provide a proxy for the geology andaid seismic interpretation.

The velocity structure was estimated using the finitedifference tomographic inversion method of Hole (1992),in which the non-linear travel time equations are solvediteratively. For such an iterative scheme, the initial velocityconditions can sometimes play an important role. A varietyof initial velocity conditions were created by analysing thefirst breaks within the shot gathers, but all converged tosimilar velocity models. Although the technique can be usedin three dimensions, it was applied in two dimensions onlyfor this study by projecting source and receiver co-ordinatesonto a 2D grid and carrying out appropriate corrections totravel times. The forward computation of travel times wascarried out using the finite difference travel time codedescribed in Hole & Zelt (1995), in which raypaths arecomputed by following the gradient in travel time from thereceiver to the source.

A velocity model (Fig. 5) was created using 94 shots inthe southeastern segment of the survey and 73 shots fromthe northwestern segment. There were a total of 21 950refracted Pg arrivals. The inversion model was parameterisedusing 0.1 km cells, and the final velocity model was obtainedafter five iterations. The travel time r.m.s. decreased from0.32 s (for a 1D starting model) to 0.08 s for the final model.The model provides information to depths of c. 4 km belowsea level, but is subject to much greater errors anduncertainties at depth, so only the shallowest 3 km (i.e.,depths of 2000 m below sea level) are thought to be reliable.

InterpretationModelling of the reflection data gives velocities ranging from2.0 to 6.0 km/s for both the southeast and northwest linesegments (Fig. 5). There is a gradual increase in velocitywith depth. A similar gradient is present in seismic velocitymodels generated by inversion of local earthquake and activesource data (Davey et al. 1998; Eberhart-Phillips & Bannister2002). The distribution of modelled velocities agrees closelywith shallow structure predicted independently fromgeological mapping.

In models from this study, the slowest velocities(<2.6 km/s) form a thin layer across the top of the modelsand are interpreted to represent late Quaternary andPliocene-Pleistocene fluvioglacial gravels. Thicker regions(c. 500 m) of material with low velocity occur in the vicinityof the Tasman valley, beside Lake Pukaki, in the footwall ofthe Irishman Creek Fault, and in the vicinity of Tekapo River(Fig. 5).

Significant regions with velocities of 2.6-3.5 km/s occurin the Mackenzie Basin in the footwall northwest of IrishmanCreek Fault, suggesting between 800 and 1200 m of LateCretaceous-Miocene sediments are preserved beneath 500 mof Pliocene-Pleistocene gravel. Velocities of 2.6-3.5 km/salso occur at depth beneath Tekapo River, suggestingc. 600 m of Late Cretaceous-Miocene section may bepreserved beneath the late Quaternary river gravels.

Velocities >4 km/s map the distribution of basementTorlesse greywacke and schist. These different rock typescan be anisotropic by as much as 17% and have wide rangesin velocity (Garrick & Hatherton 1973; Okaya et al. 1995;Godfrey et al. 2000), making them difficult to distinguishin velocity models. Relatively elevated velocities (>4 km/s)occur close to the surface at the southeastern end of themodel, consistent with the local geology at Burke Pass wherethere are very thin alluvial fans mantling exposed ridges ofgreywacke and semischist basement.

Gravity data were collected in the same position as theseismic survey line (Chetwin 1998). The gravity observ-ations and modelling provide an independent check ofsediment thickness variations predicted by velocity models.Residual gravity data show marked steps every 4-6 km downthe Tasman valley, with local anomalies that reflectshallowing of basement and/or thickening of gravels. Gravitydata steps are located in the same position where >4 km/s(i.e., basement) velocities shallow in velocity models.Gravity data also show a local negative anomaly across theIrishman Creek Fault, with a shape indicating the presenceof relatively low-density rocks in the footwall and relativelyhigh-density basement rocks near the surface in the hangingwall. Velocity model evidence corroborates the gravitymodel for thick sequences of medium velocity, Late

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

30 New Zealand Journal of Geology and Geophysics, 2003, Vol. 46

NW2 km

CD*

Qa.s.l. -

-2 km

- Oa.s.l.

Velocity {km/s)£ 3 4 5 1.5 x verbcai 5 km

Fig. 5 Velocity models derived from finite difference tomographic inversion using the first breaks of the shot gathers from the seismicsurvey. Velocities <2.6 km/s are interpreted to be late Quaternary and Pliocene-Pleistocene fluvioglacial gravels. Velocities of 2.6-3.5km/s are believed to represent Late Cretaceous-Miocene sediments. Velocities >4 km/s represent basement Torlesse greywacke andschists.

Cretaceous-Miocene sediments preserved in the footwall ofIrishman Creek Fault.

SEISMIC REFLECTION INTERPRETATIONS

ContextA number of 2D and 3D models have recently been publishedthat claim to approximate continental collision across theSouthern Alps. Some consist of finite-element solutionsusing complex but realistic rheologies (Braun & Beaumont1995; Beaumont et al. 1996; Batt & Braun 1997, 1999;Koons et al. 1998), while others use analytical and numericalapplications of critical-wedge theory (Koons 1990, 1994;Enlow & Koons 1998), and others are experimental sand-box analogues (Koons & Henderson 1995). The models areable to reproduce the mean topographic shape of theSouthern Alps and provide a context for the interpretationof many of the observed structural features usingrelatively simplistic boundary conditions. Importantly,modelling suggests a close relationship exists betweentectonics and first-order topography (>15 km wave-length), and that local structure and strain may be relatedto, and can be predicted by, the mechanics of deformation(Koons 1994, 1995).

Models extended into three dimensions predict thatstrike-slip and convergent components of shear strain shouldvary differently as a function of distance from the AlpineFault (Koons 1994; Braun & Beaumont 1995; Koons &Henderson 1995; Enlow & Koons 1998; Koons etal. 1998).Such partitioning is now suggested to have occurred duringboth late Quaternary and contemporary deformation (Beavan& Haines 2001; Norris & Cooper 2001). The modellingpredicts different regions of characteristic deformation andcorresponding tectonic style will occur within the collisionalorogen (Koons 1990, 1994). These have been used to placestructural, thermochronological, geochemical, and geo-physical observations into a present context (e.g., Koons etal. 1998; Templeton et al. 1998; Batt & Braun 1999; Uptonet al. 2000; Wannamaker et al. 2002). Critical wedge modelsof Koons (1990, 1994), for example, predict inboard (nearAlpine Fault) and outboard (east of Main Divide) structuralzones. The seismic line lies across the outboard zone, whichthe Koons models predict should include steeply dippingoblique thrusts that sole into a major westerly dippingdetachment fault.

Numerical and analogue modelling has not yet examinedthe effects of any lateral variations in rheology that may bepresent, and it is difficult to extend models realistically overmultiple deformation events. Passive seismic experiments

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island 31

have demonstrated anisotropy in geophysical propertiesbeneath the Southern Alps (Davey et al. 1998; Eberhart-Phillips & Bannister 2002; Wannamaker et al. 2002) thatmay mark changes in mechanical/deformation properties.Anisotropy caused by pre-existing faults, bedding, ordifferent rock types, for example, is likely to have beenpresent before formation of the Pliocene-Recent SouthernAlps.

Field mapping indicates that considerable deformationand reorganisation of Torlesse greywackes occurred duringMesozoic as well as Cenozoic tectonism (Spörli et al. 1974;Spörli 1979; Ward & Spörli 1979; MacKinnon 1983; Findlay& Spörli 1984; Cox & Findlay 1995; Becker & Craw 2000).Earliest deformation probably occurred shortly afterdeposition (Permian to Late Triassic), in places before rockswere completely lithified, and was followed by pre- and syn-metamorphic deformation events during Jurassic to LateCretaceous before the late Cenozoic formation of the alps.Local structural histories generally involve imbrication,folding, and tilting along horizontal fold axes, followed byrefolding around steeply plunging axes, then latestmovements on brittle faults. The metamorphic peak inadjacent Alpine Schist, and corresponding development ofpervasive foliation and isograds, also occurred before themodern phase of oblique convergence and uplift (Little etal. 2002 and references therein). Although the Alpine Schistwas tilted and structurally reorganised during the lateCenozoic (Cox et al. 1997), schist fabrics contain only aminor penetrative neotectonic overprint (Little et al. 2002).

Seismic reflection studies image the present-daydistribution of different rock types and structure, which isthe end result of a considerably longer geological historythan the present (instantaneous) convergence addressed bynumerical and analogue modelling. The northwestern endof the Mackenzie Basin seismic line images a region wherebedding has been rotated into a north-northeast strike (Fig.1), and the geology appears to be dominated by active and/or late Cenozoic faulting (Cox & Findlay 1995). In contrast,the southeastern end of the seismic line crosses theMackenzie Basin, from a region where bedding strikesnorthwest in an apparently Mesozoic orientation. Active faulttraces are present in the southeast, but late Quaternarydisplacement rates are about an order of magnitude lowerthan in the high alps. It is unclear the extent to which thegeological "architecture" along the seismic line (i.e., thegeometry and distribution of different structures and rocktypes) is dominated by Cenozoic or Mesozoic tectonics.

Approach of this studyTwo end-member interpretations are presented below. Thefirst is based on an assumption in which the crustalarchitecture predominantly reflects Cenozoic processesassociated with distributed deformation southeast of theAlpine Fault. The second is based on an assumption in whichthe crustal architecture is predominantly the result ofMesozoic tectonics, with only minor readjustment andoverprint by Cenozoic deformation. Both interpretations cansatisfactorily account for the geophysical and geologicalobservations.

In both interpretations, the observed reflections anddiscontinuities in the seismic data are considered to be eitherfaults, depositional contacts between units with significantlithological contrast, or faulted contacts. The interpretative

geological section from Table 1 was used in both models.The major difference between the two interpretations is howthe active structures are inferred to link into basaldetachments and high-strain zones at depth.

Interpretation 1: Cenozoic architecture

The seismic interpretation presented in Fig. 6A fits surfacegeology and reflection observations (Fig. 2C) to the principalfeatures of contemporary 3D plate boundary collisionalmodels (Koons 1994,1995; Braun & Beaumont 1995; Koons& Henderson 1995; Enlow & Koons 1998; Koons et al.1998) and earlier conceptual models of the Southern Alps(Wellman 1979; Norris et al. 1990).

These 3D models have variations in the strike-slip andconvergent components of shear strain, which were used byKoons et al. (1998) to define four regions: (1) an outboard(eastern) region of oblique overthrusting where material isincorporated into the active orogen; (2) a detachment or high-strain zone at the base of the deforming zone, backthrustingin the opposite direction to motion across the plate boundary;(3) a transitional (Main Divide) region dominated by angularstrain with minor contraction and extension; and (4) aninboard (western) region of concentrated strain, rapid uplift,and exhumation adjacent to the plate boundary. The seismictransect can be placed in this context on the basis ofmeasured variations in contemporary strain (Beavan &Haines 2001). The seismic line traverses from wherecontemporary strains have outboard characteristics (in thesoutheast) through the transitional region to the eastern sideof the inboard region (northwest). The basal detachment/high-strain zone is interpreted to occur at approximately thesame depth, or slightly deeper, than the 12 ± 2 km base ofthe seismogenic zone (Leitner et al. 2001).

The active Fox Peak Fault (just to the east of the surveyline) lies in the outboard region of the orogen, and is inferredin this interpretation to shallow at depth into a low angle,westerly dipping detachment fault. Active deformationoccurs at least 40 km farther east, along the edge of theCanterbury Plains, where contemporary strain is accum-ulating (Beavan et al. 1999) and active fault traces andanticlines have been mapped (Oliver & Keene 1989; Cox1995; Barrell et al. 1996). An active fault shown beneaththe Fox Peak Fault is inferred to link to structures at theeastern side of the orogen along the edge of the CanterburyPlains.

A northwesterly dipping fault with a trace in the vicinityof Tekapo River is drawn through a series of strong near-surface reflections and weaker discontinuities at depth.Although there is no trace of this structure across the recentalluvial river terraces, older glacial outwash surfaces (Woldsand Balmoral age; Maizels 1989) nearby are warped byanticlines (D. Barrell, GNS unpubl. mapping), which arelikely to be the surface expression of active fault slip at depth.The distribution of cover sediments, as evidenced by velocitymodels (above), suggests a protracted history of movementwith reactivation likely on this fault.

The Irishman Creek Fault provides a challenge to thisinterpretation, which seeks an explanation for faults in termsof the late Cenozoic deformation field only. Strong seismicreflections corroborate gravity modelling and evidence fromthe orientation of Pliocene-Pleistocence sediments exposedin the hanging wall (Fox 1987; Chetwin 1998) to provideclear indication of a southeast dip (Fig. 3). The fault has a

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

NW

Lale Quaternary fluwoglacial gravels

PliaoHnB-Plei-stocens gravels

Lale Cretaceous-Miocene sequence

Weakly cleaved greywacke

Semi schist, strongly Foliated {TZIIB)

Schist, waakly segregaled (TZlll)

Schist, sfrpngly segregated

Schist,

Faull-intarnsdactive

t \ t Faull • <nfefred

1 0 km

Fig. 6 Interpretation of upper crustal structure based on seismic reflection observations, velocity modelling, and geological constraints. The various shades of schist and greywacke representthe varying degrees of metamorphism. The thick black lines represent inferred active faults; the dashed black lines represent inferred inactive faults. A, Structure derived solely by Cenozoicprocesses. The faults to the southeast are interpreted to be backthrusted off a deeper detachment fault. The anticlinal feature on the northwestern end is inferred to be a regional scale fold inwhich the northwestern limb is comprised of active westerly dipping faults. B, Structures from the Mesozoic are preserved and reactivated during Cenozoic deformation. The explanation forthe antiform is the same as described above, but the Mesozoic structures on the southeastern limb have been tilted by folding.

%

§

Qo

9f

<

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island 33

known southeast side up, reverse displacement. The dipdirection is at odds with the expected orientation of lateCenozoic faults southeast of the Alpine Fault (Davey et al.1998), which commonly dip to the northwest and aredownthrown toward the southeast (Cox & Findlay 1995).In the interpretation of Fig. 6A, the Irishman Creek Fault isshown as a backthrust from the Tekapo River Fault (i.e., ashallow backthrust from a backthrust off the Alpine Fault,similar to faults in the cartoon interpretation of Norris et al.(1990)). This interpretation implies significant slippartitioning from the Tekapo River Fault onto the IrishmanCreek Fault and probably out-of-section displacement, sincethere is no major change in geometry of the Tekapo RiverFault at its intersection with the Irishman Creek Fault.

The northwestern side of the seismic line crosses themouth of the Jollie valley, which is a major, 20 km northeast-trending topographical lineament. The valley has the sameorientation as recently mapped faults farther along-strike,and was inferred from geological mapping to be faultcontrolled, although no fault has been observed owing tothick Quaternary sediment cover (Fig. 1) (S. Cox, GNSunpubl. data; Spörli et al. 1974). Other supporting evidencefor a major fault includes: (1) a significant change incharacter of the contemporary strain field across the valley,with an increase in shortening rates toward the west (Beavanet al. 1999); (2) marked steps in gravity data adjacent to thevalley (Chetwin 1998), indicating shallowing of basementnorthwest of the valley mouth; and (3) partially or fullyannealed apatite fission tracks in basement rocks northwestof the valley (Tippett & Kamp 1993a,b), indicative of adifference in exhumation northwest and southeast of thevalley. A steeply dipping, dominantly strike-slip fault isdrawn in the vicinity of the Jollie valley and is interpretedto reflect the change from a transitional tectonic region ofrelatively low contemporary strains into much higher strainsand corresponding uplift of the inboard (western or central)zone.

The antiformal geometry of reflectors and discontinuitiesat the northwestern end of the seismic line is inferred to bea regional-scale fold (Fig. 4, 6A). Its northwestern limb isinterpreted to comprise active, westward dipping faultswhich breach the surface and are equivalents of the GreatGroove Fault and Main Divide Fault Zones exposed nearby(Lillie & Gunn 1964; Spörli 1979; Cox & Findlay 1995;Cox et al. 1997). At depth, active faults in the northwesternlimb of the antiform are shown to detach and delaminatedifferent layers of semischist and schist, similar topostmetamorphic faults now uplifted and exposed in AlpineSchist (Craw et al. 1987, 1994; Cox et al. 1997). Thesoutheastern limb of the antiform is interpreted to compriseolder Mesozoic contacts and faults that have been steepenedinto an easterly dip by differential uplift. Some of these mayhave been low-angle shear zones/detachment faults activein the outboard part of the orogen, but are now being foldedas they are being shifted closer to the Alpine Fault.

The trend of the antiform hinge cannot be determinedfrom the seismic section, but can be inferred by correlationbetween its axial position and a low seismic velocity zoneidentified by passive seismic experiment (Eberhart-Phillips& Bannister 2002). The low-velocity zone trends 040,oblique to the Alpine Fault, and the antiform is thought tohave a similar trend. The 040 trend is near-orthogonal tothe direction of maximum geodetic shortening (Beavan et

al. 1999) and is consistent with the notion of an activelygrowing fold.

Interpretation 2: Mesozoic architecture with Cenozoicoverprint

Bedding in the Torlesse Terrane defines a distinct arc acrossthe South Island, consistent with distributed dextral bendingsoutheast of the Alpine Fault (Fig. 1) (Spörli 1979). Surficialrocks at the southeastern end of the seismic line appear tobe essentially unaffected by this reorientation and probablyrepresent the original Mesozoic structural pattern. The natureof deep Mesozoic structure in this region, however, is poorlyknown but some inferences can be made on the basis ofstructure and structural history in Otago.

Older parts of the Torlesse Terrane were sutured to theCaples Terrane in an accretionary prism during the Jurassic,with resulting metamorphism forming the Otago Schist(Mortimer 1993). Southwest-directed nappe-sheets andrecumbent ductile folds are now exposed within the deeperTorlesse-derived portion of Otago Schist, whereastransitional ductile-brittle collisional structures are exposedin shallower parts of the pile (Roser & Cooper 1990; Craw& Norris 1991; Mortimer 1993). Metamorphic fabricsoverprint at least some earlier generations of steeplyplunging folds (Spörli 1979). Uplift and cooling of OtagoSchist also began in the Jurassic (Adams et al. 1985) andcontinued until the Late Cretaceous c. 100 Ma, while muchof the younger Torlesse of the Marlborough region was stillbeing deposited and accreted. Uplift appears to have beenassociated with extensional faults, many of which aremineralised or have evidence of multiple reactivation (Mutch& Wilson 1952; Bishop 1974; Craw & Norris 1991; Anguset al. 1997; Turnbull 2001).

We assumed that Torlesse rocks in the Mackenzie Basinarea underwent a similar structural history to those alongthe nearby flanks of the Otago Schist. They are likely tocontain numerous brittle or brittle-ductile, shallow levelstructures that accommodated Jurassic collisional shorteningand Cretaceous extension. Many of the major structures mayhave been broadly parallel to the orogen (cf. Group 1 faults,Bishop 1974), possibly with an overall sense of vergencesimilar to nappe folds deeper in the pile (Roser & Cooper1990), and/or a transition to oppositely verging folds andfaults in the toe of the Torlesse accretionary prism.Reorientation of any such structures by Cenozoic bendingadjacent to the present plate boundary would result ineastward-dipping features, such as those interpretedsoutheast of Landslip Creek in Fig. 6B.

The Irishman Creek Fault dips to the southeast (Fig. 3).Explaining this orientation in terms of formation in thepresent tectonic framework (Interpretation 1 above) requirespositing a backthrust off a backthrust. An alternativeexplanation is that the Irishman Creek Fault has had a muchmore protracted reactivation history (cf. Bishop 1974) andis similar to faults exposed in mountain ranges nearby thatdisplay a range of different fault rocks indicative of faultreactivation. In this interpretation (Fig. 6B), the IrishmanCreek Fault is considered to have formed with a northweststrike, dipping northeast, during either juxtaposition of theTorlesse Terrane into the Otago schist pile or duringpostmetamorphic extensional uplift, then rotated into itspresent orientation by Cenozoic regional bending. In thisscenario, the present southeast side up displacement is only

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

34 New Zealand Journal of Geology and Geophysics, 2003, Vol. 46

Alpine Fault

3-1

Generalisedtopography Une Location Canterbury Plains

40 60 SO IQO

Distance SE of Alplrw Faurft (km)

120

Fig. 7 The upper crustal interpretation placed in a regional context derived from Stern et al. (1997), Davey et al. (1998), and Kleffmanet al. (1998). The shaded region represents a zone of high conductivity (Wannamaker et al. 2002).

a minor reactivation from late Cenozoic shortening. Othersoutheast-dipping faults are interpreted on the southeast sideof the seismic line through southeast-dipping seismicreflections and discontinuities (Fig. 2B,C). They are shownwith variable senses of displacement and activity in orderto reflect the potential for protracted geological histories.The explanation of the existence of the Jollie valley faultremains the same as the interpretation above.

The antiformal structure on the northwestern side of theseismic line (Fig. 4) is similar to Interpretation 1 above. Bothinterpretations attribute westward-dipping faults on the foldsnorthwestern limb to be related to Main Divide Fault Zonebackthrusts from the Alpine Fault (Cox & Findlay 1995).Faults on the southeastern limb, however, are shown asMesozoic structures that have been tilted by folding. Theage of the antiform, and whether or not it is related to thepresent tectonic setting or an older structure, is not clear.Correlation with a local low seismic velocity zone (Eberhart-Phillips & Bannister 2002) suggests the antiform strikes 040,agreeing well with geodetic shortening directions andformation by contemporary deformation. However, thisdirection is also parallel to the strike of regional schistosityin Alpine Schist that contains only a weak neotectonicoverprint (Little et al. 2002), which may be a counterargument of an actively growing fold of the first inter-pretation. The Alpine Schist fabric is probably Mesozoic,with mounting evidence for mid-late Cretaceous recrystal-lisation, although an early Miocene age cannot be entirelyruled out (Grapes 1995; Walker & Mortimer 1999; Mortimer2000; Vry et al. 2000; Little et al. 2002). Folds of similarwavelength and orientation to the antiform exposed in OtagoSchist are known to be associated with formation of theMoonlight Fault (e.g., Earnslaw Synform, ShotoverAntiform—Craw 1985; Turnbull 2001), having formedduring late Tertiary wrench tectonics. In this interpretation,the age of the antiform is considered only to be older thanthe present regime of oblique transpression and uplift.

Comparisons with other geological and geophysicaldataInterpretations of seismic data from previous experimentsshow that crustal thickness varies across the South Islandfrom c. 27 km under the east coast of Canterbury to c. 45km just east of the Southern Alps Main Divide (Stern et al.1997; Davey et al. 1998; Kleffman et al. 1998). Mid-uppercrustal p-wave velocities vary from c. 5.0 km/s near thesurface to c. 6.2 km/s at c. 12 km depth; the lower crust hasvelocities varying from 6.7 to 7.1 km/s. Figure 7 places thesecond interpretation into the above context.

A study of small (mostly <M5) earthquakes in theSouthern Alps region indicates that elastic strain is beingreleased seismically down to 10-12 km depth (Leitner et al.2001). Earthquakes record a relatively uniform stress fieldwith a maximum shortening direction of 110-120°, similarto geodetic observations and plate motions (Beavan et al.1999; Leitner et al. 2001). The base of the seismogenic zoneis relatively uniform at 12 ± 2 km over large parts of theSouth Island, but shallows by 3-4 km under the high alpsregion. The antiform interpreted in the data beneath Tasmanvalley (Fig. 6A,B) displaces schists and semi-schistsvertically by c. 1-1.5 s TWT (i.e., c. 2-5 km), correspondingclosely with the position and degree to which theseismogenic zone shallows.

3D velocity (Vp and Vp/Vs) modelling by inversion ofactive and passive source seismic data has been used togenerate a 2D cross-section and depth slices across the SouthIsland (Eberhart-Phillips & Bannister 2002). At depths to c.14 km there is a relatively low-velocity region centred 28km southeast of the Alpine Fault, below the Tasman valley,which corresponds to the position of the interpreted antiform.Horizontal depth slices suggest the low-velocity "anomaly"has a 040° trend weakly discordant to the Alpine Fault andis continuous for >100 km. The low velocities are interpretedto reflect an actively deforming zone in which increasedcrack-density and high fluid pressures locally reduce the

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island 35

seismic velocity. Further evidence for fluids at depth can befound in a magnetotelluric transect c. 60 km northeast ofthe SIGHT98 seismic line. A low-conductivity anomaly hasbeen delineated at mid-lower crustal depths, thought toarise from prograde metamorphism within the thickeningcrustal root (Wannamaker et al. 2002). The position ofthe anomaly extends from a few kilometres southeast ofthe Alpine Fault to the northwestern third of the seismictransect (Fig. 7).

DISCUSSION

Two end-member interpretations have been presented whichfit geophysical observations to geological boundaryconditions and thus provide cross-sectional models of theshallow crust beneath the Southern Alps. The models differin the extent to which deformation southeast of the AlpineFault plate boundary is dominated by newly developed orinherited structures. The degree to which each interpretationmay be valid is not known, and it is expected the structuralarchitecture of the region will involve some combination ofthe two end members. Dislocation modelling and long-termdatasets of geodetic strain or earthquakes may provide usefultests of the tectonic scenarios presented, particularly if thisdata can be linked to fault-slip studies and observations oflonger term geological strains.

A regional-scale antiform beneath the Tasman valley isdrawn in both interpretations, located in the area wherehorizontal motion of Pacific plate material is expected tochange as rocks are translated up the Alpine Fault ramp.The antiform marks an area where significant uplift of AlpineSchist and thermally annealed greywackes are expected,coinciding with elevated contemporary geodetic strains.Depending on its age, which is uncertain, the presence ofthe antiform has implications for style of uplift anddeformation of the shallow crust, the relationship betweentectonics and topography, and the location of elastic strainrelease and hence potential seismic hazard.

The geometry of the antiform, although known indirectlyin terms of TWT rather than depth, seems to imply c. 12%horizontal shortening over 15 km (i.e., 1.8 km). At currentelastic GPS strain rates of 0.15-0.25 µstrain/yr, corres-ponding to rates of 2.25-5 mm/yr over the 15 km distance,the antiform would have taken between 360 000 and 800000 yr to "grow" if it was entirely late Cenozoic in age.Interestingly, the area of greatest geometric uplift coincideswith the Tasman valley where there was major glacialerosion. The question arises as to whether or not erosionmay have played a significant role in exhumation anddeformation east of the Main Divide, as suggested for thewest (e.g., Koons 1990, 1995; Norris & Cooper 1995). Ifthe calculated age of 360-800 ka for the antiform is correct,then erosion during and after the Nemona or WaimaungaGlacial Advances (Suggate 1990) may have initiated fold-growth and instigated local structural reorganisation of theshallow South Island crust. A west-dipping thrust duplexexposed immediately west of the Main Divide (Cox et al.1997), that postdates 880 ka (minimum) quartz-adularia-calcite veins (Teagle et al. 1998), would have been coincidentwith antiform growth and is presumably the exposedequivalent of larger scale antiform shortening.

The seismic experiment did not provide any newinformation on the nature of the greywacke-schist transition

at depth beneath the South Island. Primary cleavage andbedding in the Torlesse rocks are steep throughout the SouthIsland, whereas overprinting metamorphic fabrics in Otagoand Alpine Schist appear to be subhorizontal and steep,respectively. The interpretations presented in this study aredominated by low-lying lithological contacts, possiblyreflecting the difficulty of seismic surveys to resolve steepfeatures, but also due to the boundary conditions impartedon the interpretation by schist exhumed nearby (Table 1).The interpretations do not resolve the nature of the transitionbetween steep Torlesse bedding and low-angle schistosity;for example, whether or not it is a gradational overprintingrelationship (as suggested by Bishop 1974) or marked byfaults (Craw 1998).

Studies of late Quaternary slip on the Alpine Faultaccount for 75% of the strike-slip plate motion, and variableamounts of the dip-slip motion (Norris & Cooper 2001).Importantly, significant proportions of the plate motion arenot accounted for and probably occur in the Southern Alps.It may imply potentially higher seismic risk to the easternSouth Island than previously considered (Norris & Cooper2001). The challenge is now to test and confirm whether ornot the inferred active faults can account for the amount ofmissing plate motion.

SUMMARY

The SIGHT98 experiment did not image any major regional-scale reflections or discontinuities in the uppermost 5 sTWT. Instead, more subtle smaller scale features arepresent, such as well-defined reflections that are laterallycontinuous for only 2-3 km, and strong angular discon-tinuities in reflections which can be mapped for up to 10km. Reflections and discontinuities in the data are thoughtto mark faults, bedding, or schistosity, and/or lithologicalcontacts.

Seismic reflection data provided a means by which astandard geological cross-section, drawn with relativeconfidence to depths of c. 2 km, can be extended to depthsof c. 12-15 km. Velocity models provided an independentcheck on the distribution of cover sediments in the uppermost3 km.

A series of strong, southeast dipping reflections coincidewith the trace of the Irishman Creek Fault, corroboratingprevious geological and geophysical evidence that this activestructure dips southeast. Slow velocities <3.6 km/s extendto depths of almost 2 km on the northwestern side of thefault, suggesting between 800 and 1200 m of LateCretaceous-Miocene sediments and 500 m of Pliocene-Pleistocene gravel are preserved in the footwall. Abasementhigh occurs in the hanging wall, with greywacke reachingto within 300 m of the surface, buried beneath Pliocene-Pleistocene cover and possibly a thin sliver of LateCretaceous-Tertiary sequence.

Oppositely dipping reflections and discontinuities in theseismic data define a previously unrecognised regional-scaleantiform beneath the Tasman valley and Mount Cook. Theantiform has a horizontal half-wavelength of 15 km, a crestat 1.2 s TWT (c. 3 km depth), and is observed to depths ofat least 3.5s(10 ± 2 km) within the brittle seismogenic crust.The antiform hinge is inferred to trend northeast (040) bycorrelation with a low seismic velocity zone that appears tobe coincident with the structure (Eberhart-Phillips &

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

36 New Zealand Journal of Geology and Geophysics, 2003, Vol. 46

Bannister 2002). The geometry/orientation of the antiformis consistent with growth during late Cenozoic oblique-convergence, causing uplift of the Southern Alps where thePacific plate becomes bent across the ramp of the AlpineFault, but an early Miocene or Jurassic age cannot be ruledout.

Two end-member interpretative cross-sections of theseismogenic crust of the central South Island have beenpresented in this paper. Interpretation 1 fits observations toa model in which the structure is dominated by Cenozoicdeformation, with little influence of pre-existing features inthe geology. It is analogous to sandbox models and numericalexperiments, which predict different structural regions withinthe Southern Alps that can now be defined by contemporarygeodetic strain rates (Beavan & Haines 2001). Interpretation2 fits observations to a model that accounts for thewidespread geologic evidence of fault reactivation, multipledeformation, and prolonged geologic history in Torlessegreywackes. Both interpretations draw extensively ongeological observations in rocks exposed adjacent to theseismic line, and both can satisfactorily account forgeophysical and geological observations to requirements ofthe geological and other geophysical constraints. The majordifference between the two interpretations is the degree towhich active structures are inferred to link into a basaldetachment and high-strain zones at depth. Independenttesting is now required to determine which interpretation ismost realistic.

ACKNOWLEDGMENTS

We acknowledge the skill and efforts of the SIGHT teams whocollected the data, and the numerous landowners, Department ofConservation, and other people and organisations who contributedto the fieldwork and aided this large-scale experiment. We thankStuart Henrys, Nick Mortimer, James Cull, and an anonymousreferee for thoughtful reviews that considerably improved thispaper. This project was funded by the United States NationalScience Fund (EAR 9418530) and the New Zealand Foundationfor Research, Science and Technology (C05811).

REFERENCES

Adams, C. J.; Bishop, D. G.; Gabites, J. E. 1985: Potassium-argonstudies of a low grade progressively metamorphosedgreywacke sequence, Danseys Pass, South Island, NewZealand. Journal of the Geological Society of London 142:339-349.

Adams, J. 1980a: Contemporary uplift and erosion of the SouthernAlps, New Zealand: Summary. Geological Society ofAmerica Bulletin part I, 91: 2-4.

Adams, J. 1980b: Contemporary uplift and erosion of the SouthernAlps, New Zealand. Geological Society of America Bulletinpart II,91: 1-114.

Angus, P. V. M.; de Ronde, C. E. J.; Scott, J. G. 1997: Explorationalong the Hyde Macraes Shear. Proceedings of the 1997Minerals and Mining Conference. Wellington, Ministry ofCommerce. Pp. 151-157.

Barrell, D. J. A.; Forsyth, P. J.; McSaveney, M. J. 1996: Quaternarygeology of the Rangitata fan, Canterbury Plains, NewZealand. Institute of Geological & Nuclear Sciences Report96/23. 67 p.

Batt, G. E.; Braun, J. 1997: On thermomechanical evolution ofcompressional orogens. Geophysical Journal International128: 364-382.

Batt, G. E.; Braun, J. 1999: The tectonic evolution of the SouthernAlps, New Zealand: insights from fully thermally coupleddynamical modelling. Geophysical Journal International136: 403–420.

Beaumont, C.; Kamp, P. J. J.; Hamilton, J.; Fullsack, P. 1996: Thecontinental collision zone, South Island, New Zealand:comparison of geodynamical models and observations.Journal of Geophysical Research 101: 3333-3359.

Beavan, J.; Haines, J. 2001: Contemporary horizontal velocity andstrain rate fields of the Pacific-Australian plate boundaryzone through New Zealand. Journal of GeophysicalResearch 106 B1: 741-770.

Beavan, J.; Moore, M.; Pearson, C.; Henderson, M.; Parsons, B.;Bourne, S.; England, P.; Walcott, D.; Blick, G.; Darby, D.;Hodgkinson, K. 1999: Crustal deformation during 1994-1998 due to oblique continental collision in the centralSouthern Alps, New Zealand, and implications for seismicpotential of the Alpine fault. Journal of GeophysicalResearch 104, B11: 25233-25255.

Becker, J. A.; Craw, D. 2000: Structure and neotectonics on theMain Divide, upper Wilberforce valley, mid Canterbury,New Zealand. New Zealand Journal of Geology andGeophysics 43: 217-228.

Bishop, D. G. 1974: Stratigraphic, structural and metamorphicrelationships in the Dansey Pass area, Otago, New Zealand.New Zealand Journal of Geology and Geophysics 17:301-335.

Black, B. V. N. 2000: Late terrestrial strata of North Otago andSouth Canterbury, and their genetic relationships.Unpublished MSc thesis, Geology Department, Universityof Otago, Dunedin, New Zealand.

Blick, G. H.; Read, S. A. L.; Hall, P. T. 1989: Deformationmonitoring of the Ostler fault zone, South Island, NewZealand. Tectonophysics 167: 329-339.

Braun, J.; Beaumont, C. 1995: Three-dimensional numericalexperiments of strain partitioning at oblique plateboundaries: implications for contrasting tectonic styles inthe southern Coast Ranges, California and central SouthIsland, New Zealand. Journal of Geophysical Research100, B9: 18059-18074.

Chetwin, E. 1998: Active faulting and subsurface structure in theMacKenzie Basin: an investigation using gravity andseismic modelling. Unpublished BSc (Hons) thesis,Victoria University of Wellington, Wellington, NewZealand.

Cox, S. C. 1995: Geological transect across the Southern Alps ofNew Zealand. Institute of Geological & Nuclear SciencesReport 95/30. 29 p.

Cox, S. C.; Findlay, R. H. 1995: The Main Divide Fault Zone andits role in formation of the Southern Alps, New Zealand.New Zealand Journal of Geology and Geophysics 38:489–499.

Cox, S. C.; Craw, D; Chamberlain, C. P. 1997: Structure and fluidmigration in a late Cenozoic duplex system forming theMain Divide in the central Southern Alps, New Zealand.New Zealand Journal of Geology and Geophysics 40:359-374.

Craw, D. 1985: Structure of the schists in the Mt Aspiring region,northwestern Otago. New Zealand Journal of Geology andGeophysics 30: 273-286.

Craw, D. 1998: Structural boundaries and biotite and garnet"isograds" in Otago and Alpine Schists, New Zealand.Journal of Metamorphic Geology 16: 395–402.

Craw, D.; Norris, R. J. 1991: Metamorphogenic Au-W veins andregional tectonics: mineralisation throughout the uplifthistory of the Haast Schist, New Zealand. New ZealandJournal of Geology and Geophysics 34: 373-383.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island 37

Craw, D.; Rattenbury, M. S.; Johnstone, R. D. 1987: Structuralgeology and vein mineralisation in the Callery Riverheadwaters, Southern Alps, New Zealand. New ZealandJournal of Geology and Geophysics 30: 273-286.

Craw, D.; Rattenbury, M. S.; Johnstone, R. D. 1994: Structureswithin greenschist facies Alpine Schist, central SouthernAlps, New Zealand. New Zealand Journal of Geology andGeophysics 37: 101-111.

Davey, F. J.; Henyey, T.; Kleffman, S.; Melhuish, A.; Okaya, D.;Stern, T. A.; Woodward, D. J. and others 1995: Crustalreflections from the Alpine Fault Zone, South Island, NewZealand. New Zealand Journal of Geology and Geophysics38: 601-604.

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

Eberhart-Phillips, D.; Bannister, S. 2002: 3-D Vp and Vp/Vs inthe Southern Alps region of New Zealand from inversionof local earthquake and active source data. Journal ofGeophysical Research 107, 10.1029/2001JB 000567.

Enlow, R. L.; Koons, P. O. 1998: Critical wedges in threedimensions: analytical expressions from Mohr-Coulombconstrained perturbation analysis. Journal of GeophysicalResearch 103, B3: 4897–4914.

Field, B. D.; Browne, G. H. 1986: Lithostratigraphy of Cretaceousand Tertiary rocks, Southern Canterbury, New Zealand.New Zealand Geological Survey Record 14. 55 p.

Field, B. D.; Browne, G. H.; Davy, B.; Herzer, R. H.; Hoskins,R. H.; Raine, J. I.; Wilson, G. J.; Sewell, R. J.; Smale, D.;Watters, W. A. 1989: Cretaceous and Cenozoic sedimentarybasins and geological evolution of the Canterbury region,South Island, New Zealand. New Zealand GeologicalSurvey Basin Studies 2.

Findlay, R. H.; Spörli, K. B. 1984: Structural geology of the MountCook Range and Main Divide, Hooker valley region, NewZealand. New Zealand Journal of Geology and Geophysics27: 257-276.

Fox, A. N. 1987: The neotectonic history of the Lake Tekaporegion, McKenzie Basin, New Zealand. Unpublished MScthesis, lodged in the Library, University of Canterbury,Christchurch, New Zealand.

Gair, D. R. 1967: Sheet 20—Mt Cook. Geological map of NewZealand 1:250 000. Wellington, New Zealand. Departmentof Scientific and Industrial Research.

Gair, D. R. 1978: The late mobile phase: Tertiary stratigraphy ofMid and South Canterbury and North Otago. In: Suggate,R. P. ed. The geology of New Zealand. Vol. 2. Pp. 510-516.

Gallagher, K.; Brown, R.; Johnson, C. 1998: Fission track analysisand its application to geological problems. Annual Reviewsin Earth and Planetary Science 26: 519-572.

Garrick, R. A.; Hatherton, T. 1973: Seismic velocity studies in theSouthern Alps, New Zealand. New Zealand Journal ofGeology and Geophysics 16: 973-995.

Godfrey, N. J.; Christensen, N. I.; Okaya, D. A. 2000: Anisotropyof schists: contribution of crustal anisotropy to activesource seismic experiments and shear wave splittingobservations. Journal of Geophysical Research 105, B12:27991-28007.

Grapes, R. H. 1995: Uplift and exhumation of Alpine Schist inthe Franz Josef-Fox Glacier area of the Southern Alps,New Zealand. New Zealand Journal of Geology andGeophysics 38: 525-533.

Grapes, R. H.; Watanabe, T. 1992: Metamorphism and uplift ofAlpine Schist in the Franz Josef-Fox Glacier area of theSouthern Alps, New Zealand. Journal of MetamorphicGeology 10: 171-180.

Hicks, D. M. 1981: Deep-sea fan sediments in the Torlesse zone,Lake Ohau, South Canterbury, New Zealand. New ZealandJournal of Geology and Geophysics 2: 209-230.

Hole, J. A. 1992: Nonlinear high resolution three dimensionalseismic travel time tomography. Journal of GeophysicalResearch 97, B5: 6553-6562.

Hole, J. A.; Zelt, B. C. 1995: Three dimensional finite-differencereflection travel times. Geophysical Journal International121: 427–434.

Kleffmann, S.; Davey, F.; Melhuish, A.; Okaya, D.; Stern, T. 1998:Crustal structure in the central South Island, New Zealand,from the Lake Pukaki seismic experiment. New ZealandJournal of Geology and Geophysics 41: 39–49.

Koons, P. O. 1990: Two-sided orogen: collision and erosion fromthe sandbox to Southern Alps, New Zealand. Geology 18:679-682.

Koons, P. O. 1994: Three-dimensional critical wedges: tectonicsand topography in oblique collisional orogens. Journal ofGeophysical Research 99: 12301-12315.

Koons, P. O. 1995: Modelling the topographic evolution ofcollisional belts. Annual Reviews of Earth and PlanetaryScience 23: 375–408.

Koons, P. O.; Henderson, C. M. 1995: Geodetic analysis of modeloblique collision and comparison with the Southern Alpsof New Zealand. New Zealand Journal of Geology andGeophysics 38: 545-552.

Koons, P. O.; Craw, D.; Cox, S. C.; Upton, P.; Templeton, A. S.;Chamberlain, C. P. 1998: Fluid flow during active obliqueconvergence: a Southern Alps model from mechanical andgeochemical observations. Geology 26: 159-162.

Langdale, S.; Stern, T. 1998: Late Tertiary deformation inCannington Basin, South Canterbury, New Zealand:evidence from seismic and gravity data. New ZealandJournal of Geology and Geophysics 41: 247-257.

Lensen, G. J. 1975: Earth deformation studies in New Zealand.Tectonophysics 29: 541-551.

Leitner, B.; Eberhart-Phillips, D.; Anderson, H.; Nabelek, J. L.2001: A focused look at the Alpine Fault, New Zealand:seismicity, focal mechanisms and stress observations.Journal of Geophysical Research 106, B2: 2193-2220.

Lillie, A. R.; Gunn, B. M. 1964: Steeply plunging folds in theSealy Range, Southern Alps. New Zealand Journal ofGeology and Geophysics 7: 403–423.

Little, T. A.; Holcombe, R. J.; Ilg, B. R. 2002: Kinematics ofoblique continental collision inferred from ductilemicrostuctures and strain in mid-crustal Alpine schist,central South Island, New Zealand. Journal of StructuralGeology 24: 219-239.

MacKinnon, T. C. 1983: Origin of the Torlesse terrane and coevalrocks, South Island, New Zealand. Geological Society ofAmerica Bulletin 94: 967-985.

Maizels, J. 1989: Differentiation of late Pleistocene terrace outwashdeposits using geomorphic criteria: Tekapo valley, SouthIsland, New Zealand. New Zealand Journal of Geologyand Geophysics 32: 225-241.

Mildenhall, D. C. 2001: Pollen analysis of Pliocene-PleistoceneKowai Formation (Kurow Group), Mackenzie Basin, SouthCanterbury, New Zealand. New Zealand Journal ofGeology and Geophysics 44: 97-104.

Molnar, P.; Anderson, H. J.; Audoine, E.; Erberhart-Phillips, D.;Gledhill, K.; Klosko, E. R.; McEvilly, T. V.; Okaya, D.;Savage, M. K.; Stern, T.; Wu, F. T. 1999: Continuousdeformation versus faulting through the continentallithosphere of New Zealand. Science 286: 516-519.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

38 New Zealand Journal of Geology and Geophysics, 2003, Vol. 46

Mortimer, N. 1993: Jurassic tectonic history of Otago Schist, NewZealand. Tectonics 12: 237-244.

Mortimer, N. 2000: Metamorphic discontinuities in orogenic belts:example of the garnet-biotite-albite zone in Otago Schist,New Zealand. InternationalJournal of Earth Sciences 89:295-306.

Mortimer, N.; Tulloch, A. J.; Spark, R. N.; Walker, N. W.; Ladley,E.; Allibone, A.; Kimbrough, D. L. 1999: Overview of theMedian Batholith, New Zealand: a new interpretation ofthe geology of the Median Tectonic Zone and adjacentrocks. Journal of African Earth Sciences 29: 257-268.

Mutch, A. R.; Wilson, D. D. 1952: Reversal of movement of theTitri Fault. New Zealand Journal of Science andTechnology 22: 398–403.

Norris, R. J. 1979: A geometrical study of finite strain and bendingin the South Island. In: Walcott, R. I.; Cresswell, M. M.ed. The origin of the Southern Alps. Royal Society NewZealand Bulletin 18: 21-28.

Norris, R. J.; Cooper, A. F. 1995: Origin of small-scalesegmentation and transpressional thrusting along theAlpine fault, New Zealand. Geological Society of AmericaBulletin 107: 231-240.

Norris, R. J.; Cooper, A. F. 2001: Late Quaternary slip rates andslip partitioning on the Alpine Fault, New Zealand. Journalof Structural Geology 23: 507-520.

Norris, R. J.; Koons, P. O.; Cooper, A. F. 1990: The obliquely-convergent plate boundary in the South Island of NewZealand: implications for ancient collision zones. Journalof Structural Geology 12: 715-725.

Okaya, D.; Christensen, N.; Stanley, D.; Stern, T. 1995: Crustalanisotropy in the vicinity of the Alpine Fault Zone, SouthIsland, New Zealand. New Zealand Journal of Geologyand Geophysics 38: 579-583.

Oliver, P. J.; Keene, H. W. 1989: Sheet K36AC and part sheetK35—Mt Somers. Geological map of New Zealand 1:50000. Map (1 sheet) and notes (36 p). Wellington, NewZealand. Department of Scientific and Industrial Research.

Oliver, P. J.; Campbell, J. D.; Speden, I. G. 1982: The stratigraphyof the Torlesse rocks of the Mt Somers area (S81) Mid-Canterbury. Journal of the Royal Society of New Zealand12: 243-271.

Pearson, C. 1993: Rate of co-seismic strain release in the northernSouth Island, New Zealand. New Zealand Journal ofGeology and Geophysics 36: 161-166.

Pearson, C. 1994: Geodetic strain determinations from the Okaritoand Godley-Tekapo regions, central South Island, NewZealand. New Zealand Journal of Geology and Geophysics37: 309-318.

Pearson, C. F.; Beavan, J.; Darby, D. J.; Blick, G. H.; Walcott, R.1. 1995: Strain accumulation across the Australia-Pacificplate boundary in the central South Island, New Zealandfrom 1992 GPS and earlier terrestrial observations. Journalof Geophysical Research 100: 22071-22081.

Pickerill, R. A.; Irwin, J. 1983: Sedimentation in a deep glacier-fed lake—Lake Tekapo, New Zealand. Sedimentology 30:63-75.

Roser, B. P.; Cooper, A. F. 1990: Geochemistry and terraneaffiliation of Haast Schist from the western Southern Alps,New Zealand. New Zealand Journal of Geology andGeophysics 33: 1-10.

Simpson, G. D. H.; Cooper, A. F.; Norris, R. J. 1994: LateQuaternary evolution of the Alpine Fault Zone at Paringa,South Westland, New Zealand. New Zealand Journal ofGeology and Geophysics 37: 49-58.

Smith, E. G. C.; Stern, T.; O'Brien, B. 1995: A seismic velocityprofile across the central South Island, New Zealand, fromexplosion data. New Zealand Journal of Geology andGeophysics 38: 565-570.

Speight, R. 1963: Late Pleistocene historical geomorphology ofthe Lake Pukaki area, New Zealand. New Zealand Journalof Geology and Geophysics 6: 160-188.

Spörli, K. B. 1979: Structure in the South Island Torlesse in relationto the origin of the Southern Alps. In: Walcott, R. J.;Creswell, M. M. ed. The origin of the Southern Alps. RoyalSociety of New Zealand Bulletin 18: 99-104.

Spörli, K. B.; Lillie, A. R. 1974: Geology of the TorlesseSupergroup—the northern Ben Ohau Range, Canterbury.New Zealand Journal of Geology and Geophysics 17:115-141.

Spörli, K. B.; Stanaway, K. J.; Ramsay, W. R. H. 1974: Geologyof the Torlesse Supergroup in Southern Liebig and BurnettRanges, Canterbury, New Zealand. Journal of the RoyalSociety of New Zealand 4: 177-192.

Stagpoole, V. 1997: A geophysical study of the Northern TaranakiBasin, New Zealand. Unpublished PhD thesis, VictoriaUniversity of Wellington, Wellington, New Zealand.

Stern, T. A. 1995: Gravity anomalies and crustal loading at andadjacent to the Alpine Fault, New Zealand. New ZealandJournal of Geology and Geophysics 38: 593-600.

Stern, T. A.; Wannamaker, P. E.; Eberhart-Phillips, D.; Okaya, D.;Davey, F. J.; Team, S. 1997: Mountain building and activedeformation studied in New Zealand. EOS 78 (32):329-336.

Suggate, R. P. 1990: Late Pliocene and Quaternary glaciations ofNew Zealand. Quaternary Science Reviews 9: 175-197.

Sutherland, R. 1995: The Australia-Pacific boundary and Cenozoicplate motions in the SW Pacific: some constraints fromGeosat data. Tectonics 14: 819-831.

Sutherland, R. 1996: Magnetic anomalies in the New Zealandregion, 1:4 000 000 version 1.0. Institute of Geological &Nuclear Sciences Geophysical Map 9.

Sutherland, R. 1999: Cenozoic bending of New Zealandbasement terranes and Alpine Fault displacement: a briefreview. New Zealand Journal of Geology and Geophysics42: 295-301.

Teagle, D. A. H.; Hall, C. M.; Cox, S. C.; Craw, D. 1998: Ar/Ardating and uplift rate of hydrothermal minerals in theSouthern Alps, New Zealand. In: Arehart, G. B.; Hulston,J. R. ed. Water-rock interaction. Rotterdam, Balkema. Pp.801-804.

Templeton, A. S.; Chamberlain, C. P.; Koons, P. O.; Craw, D. 1998:Stable isotopic evidence for mixing between metamorphicfluids and surface-derived waters during recent uplift ofthe Southern Alps, New Zealand. Earth and PlanetaryScience Letters 154: 73-92.

Tippett, J. M.; Kamp, P. J. J. 1993a: The role of faulting in rockuplift in the Southern Alps. New Zealand Journal ofGeology and Geophysics 36: 497-504.

Tippett, J. M.; Kamp, P. J. J. 1993b: Fission track analysis of thelate Cenozoic vertical kinematics of continental Pacificcrust, South Island, New Zealand. Journal of GeophysicalResearch 98: 16119-16148.

Turnbull, I. M. comp. 2001: Geology of the Wakatipu area. Instituteof Geological & Nuclear Sciences 1:250 000 GeologicalMap 18. 1 sheet.

Upton, P.; Caldwell, C. P.; Chamberlain, C. P.; Craw, D.; James,Z.; Jiracek, G. R.; Koons, P. O.; Wannamaker, P. E. 2000:Fluids in a backthrust regime (Southern Alps, NewZealand). Journal of Geochemical Exploration SpecialIssue: Proceedings of Geofluids III, 69: 517-521.

Van Dissen, R. J.; Hull, A. G.; Read, S. A. L. 1993: Timing oflarge Holocene earthquakes on the Ostler fault zone, NewZealand. Proceedings of the CRCM '93, Kobe, December6-11, 1993. Pp. 381-386.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14

Long et al.—Crustal structure, central South Island 39

Vry, J. K.; Maas, R.; Little, R. A.; Grapes, R. H.; Dixon, M. 2000:Diachronous (Cretaceous and Cenozoic) growth of zonedgarnets in Alpine Schist, Southern Alps, New Zealand.Geological Society of New Zealand MiscellaneousPublication 108A: 163.

Walcott, R. I. 1998: Modes of oblique compression: late Cenozoictectonics of the South Island, New Zealand. Reviews ofGeophysics 36: 1-26.

Walker, N. W.; Mortimer, N. 1999: Dating high-grade metamorphicmineral growth in the Haast Schist using U-Pb TIMS.Geological Society of New Zealand MiscellaneousPublication 107A: 167.

Wannamaker, P. E.; Jiracek, G. R.; Stodt, J. A.; Caldwell, T. G.;Gonzalez, V. M.; McKnight, J. D.; Porter, A. D. 2002: Fluidgeneration and pathways beneath an active compressionalorogen, the New Zealand Southern Alps, inferred frommagnetotelluric data. Journal of Geophysical Research107, 10.1029/2001JB000186.

Ward, C. J.; Spörli, K. B. 1979: Occurrence and possible origin ofexceptionally large steeply plunging folds in the Torlesseterrane, New Zealand. Journal of Geology 87: 187-193.

Wellman, H. W. 1979: An uplift map for the South Island of NewZealand and a model for uplift of the Southern Alps. In:Walcott, R. I.; Cresswell, M. M. ed. The origin of theSouthern Alps. Royal Society of New Zealand Bulletin 18:13-20.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

exas

Lib

rari

es]

at 0

3:45

26

Nov

embe

r 20

14