Structural and thermochronological constraints ... - Wiley Online Library

TECTONICS, VOL. 27, TC4012, doi:10.1029/2006TC002066, 2008

Structural and thermochronological constraints to the evolution of the West Antarctic Rift System in central Victoria Land Fabrizio Storti,1 Maria Laura Balestrieri,2 Fabrizio Balsamo,1 and Federico Rossetti1 Received 26 October 2007; revised 22 February 2008; accepted 10 March 2008; published 2 August 2008.

[1] Uncertainty still persists on the structure and evolution of the West Antarctic Rift System, which is one of the largest extensional provinces on Earth. In this paper, we present results of a combined structural and apatite fission track study of the western shoulder of the rift system, in the western coastal area of the Ross Sea between the Reeves and the Mawson glaciers. Structural data indicate that the onshore fault pattern is dominated by N–S striking right-lateral strike-slip to transtensional fault systems and their related subsidiary fault populations in the damage zones. Thermochronological data support a Cenozoic age for these faults and, in particular, the onset of the oblique rifting event at about 50 Ma, in Eocene times. Apatite fission track analyses also suggest the possible occurrence of an older cooling/exhumation event in Cretaceous times. When integrated with the geological and geophysical data sets available in the literature, our structural and thermochronological data provide further support to the interpretation of Cenozoic oblique rifting in the West Antarctic Rift System as induced by the transfer of dextral shear from the mid oceanic ridge in the Southern Ocean, into the Antarctic Plate interior. Citation: Storti, F., M. L. Balestrieri, F. Balsamo, and F. Rossetti (2008), Structural and thermochronological constraints to the evolution of the West Antarctic Rift System in central Victoria Land, Tectonics, 27, TC4012, doi:10.1029/2006TC002066.

1. Introduction [2] The West Antarctic Rift System (Figure 1) is one of the largest extensional provinces on Earth. It encompasses the Ross Sea to the west, the area under the Ross Ice Shelf to the south, and part of the area under the West Antarctic Ice Sheet to the east [e.g., Fitzgerald, 2002]. This rift system is considered active today because of (1) the presence of active volcanoes [Kyle and Cole, 1974; LeMasurier and Rex, 1991; Wo¨rner, 1999; Rocchi et al., 2002] and the unusually high heat flow along the western shoulder [Decker and Bucher, 1982; Blackman et al., 1987; Della Vedova et al., 1997], (2) the evidence of onshore [McKelvey 1 Dipartimento di Scienze Geologiche, Universita` ‘‘Roma Tre’’, Rome, Italy. 2 CNR, Institute of Earth Sciences and Earth Resources, Pisa, Italy.

Copyright 2008 by the American Geophysical Union. 0278-7407/08/2006TC002066

et al., 1991; Lanzafame and Villari, 1991; Stackebrandt, 1994; Jones, 1996] and offshore [Spezie et al., 1993; Barrett et al., 1995; Salvini et al., 1998; Rossetti et al., 2006a] Late Pliocene to Quaternary faults, and (3) the preferential clustering of earthquakes in this region of the Antarctic plate [Behrendt et al., 1996; Cattaneo et al., 2001; Bannister and Kennett, 2002; Reading, 2002, 2006] that, despite significant potential bias arising from the sparse and uneven distribution of Antarctic seismic recorders [Reading, 2006], suggests ongoing tectonic activity along the continental strands of the Balleny and Tasman intraplate strikeslip deformation belts starting from the midoceanic ridge in the Southern Ocean [Storti et al., 2007]. [3] Uncertainty still persists on the timing, architecture, and evolution of the West Antarctic Rift System because of the limited number of available geological and geophysical studies, compared to the Basin and Range province of the western United States and to the East African Rift, which are of similar size [Fitzgerald et al., 1986; LeMasurier, 1990; Tessensohn and Wo¨rner, 1991]. In the following, the main debated issues are illustrated. [4] 1. The amount of total extension accommodated within this stretched sector of the Antarctic plate is still uncertain and indirect estimates span from 120– 250 km [Busetti et al., 1999], to 200– 500 km [Grindley and Oliver, 1983], 255 –350 km [Fitzgerald et al., 1986; Kamp and Fitzgerald, 1987; Behrendt and Cooper, 1991], 350 – 450 km [Davey and Brancolini, 1995], 480 – 500 km [Trey et al., 1999], up to about 450– 1800 km from paleomagnetic studies [DiVenere et al., 1994] that, however, are hampered by large uncertainties on rotations between East Antarctica and the West Antarctic microplate assembly, and by the lack of correlative mid-Cretaceous poles [Luyendyk et al., 1996; Fitzgerald, 2002]. [5] 2. The tectonic architecture accommodating such crustal stretching is not well known and still under debate, also because of the complexity of the extensional deformational history [Fitzgerald, 2002]. Both pure shear [Stern and Ten Brink, 1989; van der Beek et al., 1994; Busetti et al., 1999] and simple shear models of orthogonal extension [Fitzgerald et al., 1986; Fitzgerald and Baldwin, 1997; Wo¨rner, 1999; Luyendyk et al., 2001; Fitzgerald, 2002; Siddoway et al., 2004; Davey and De Santis, 2006] have been proposed, as well as right-lateral transtension [Storey and Nell, 1988; Wilson, 1995; Salvini et al., 1997; Hamilton et al., 2001]. In the last decade, a significant structural data set has been collected in the Ross Sea region, i.e., Victoria Land and the Ross Sea, showing the dominant role of rightlateral strike-slip and transtensional faulting during Late Cenozoic times [Wilson, 1995; Salvini et al., 1997; Storti et

TC4012

1 of 21

TC4012

STORTI ET AL.: EVOLUTION OF THE WARS IN VICTORIA LAND

TC4012

Figure 1. Satellite-derived free-air gravity map of the study region (gravity grid from McAdoo and Laxon [1997] and Sandwell and Smith [1997]) showing the trace of major fracture zones and their onshore extensions [after Storti et al., 2007]. The approximate location of the Adare Basin [Davey et al., 2006] is also shown in darker grey. The boxed area indicates the location of Figure 2. al., 2001; Rossetti et al., 2000, 2002, 2003; Storti and Rossetti, 2002]. Despite these new evidences, the assumption of orthogonal rifting still characterizes recent models of the Cenozoic evolution in the West Antarctic Rift System [e.g., Busetti et al., 1999; Trey et al., 1999; Karner et al., 2005; Davey and De Santis, 2006; Davey et al., 2006]. [6] 3. The timing of rifting in this region of Antarctica has been classically interpreted in terms of two major pulses following initial breakup of Gondwana in the Jurassic [Schmidt and Rowley, 1986; Stump and Fitzgerald, 1992; Elliot, 1992; Storey, 1996]: the first one, in the Cretaceous, caused major crustal thinning in the entire West Antarctic Rift System while the second event, Cenozoic in age, was mostly confined in the western Ross Sea [e.g., Cooper et al., 1987, 1991; Behrendt et al., 1991a; Tessensohn and Wo¨rner, 1991; Davey and Brancolini, 1995; Behrendt, 1999; Karner et al., 2005]. Available apatite fission track data indicate that the first denudation pulse following the about 180 Ma Jurassic magmatic event [e.g., Elliot and Fleming, 2004], occurred in Late Cretaceous time both in southern Victoria Land [Fitzgerald et al., 2000; Fitzgerald, 2002] and in northern Victoria Land [Balestrieri et al., 1997, 1999; Balestrieri and Bigazzi, 2001]. Early Cretaceous denudation

has been revealed by apatite fission track data in the central Transantarctic Mountains further south [Fitzgerald, 1994; Fitzgerald and Stump, 1997]. Apatite fission track thermochronology also indicates that, despite some additional complexity [e.g., Fitzgerald et al., 2006], the onset of the major phase of denudation in the Transantarctic Mountains generally began in Victoria Land at about 50 Ma [Gleadow and Fitzgerald, 1987; Fitzgerald and Gleadow, 1988; Fitzgerald, 1992; Balestrieri et al., 1997, 1999; Balestrieri and Bigazzi, 2001; Fitzgerald, 2002; Lisker, 2002; Rossetti et al., 2003; Fitzgerald et al., 2006]. Partitioning of total extension between these pulses is still under debate. Several lines of evidence indicate that most of total extension occurred prior to the breakup between the Campbell Plateau of New Zealand and Marie Byrd Land, at about 80 Ma [e.g., Lawver and Gahagan, 1994; Davey and Brancolini, 1995; Mukasa and Dalzier, 2000; Luyendyk et al., 2001; Siddoway et al., 2004]. Gravimetric [Karner et al., 2005] and kinematic [van der Beek et al., 1994; Busetti et al., 1999] modeling as well as geological and geophysical data from the eastern shoulder of the rift, in Marie Byrd Land [Luyendyk et al., 2001], support large Cretaceous extension. On the other hand, extrapolation from the Adare Trough

2 of 21

TC4012


southward, would indicate that a significant amount of extension (about 180 km) occurred in Eocene-Oligocene times [Cande et al., 2000; Cande and Stock, 2004], mostly in the kinematically linked Northern Basin and Victoria Land Basin [Davey et al., 2006]. Results from the Cape Roberts drilling project, if extrapolated to the western Ross Sea, would support a Eocene-Oligocene onset of rifting in the Victoria Land Basin [Cape Roberts Science Team, 2000; Hamilton et al., 2001; Davey and De Santis, 2006; Fielding et al., 2006]. [7] 4. The geodynamic setting of the West Antarctic Rift System has been commonly interpreted as a passive margin environment produced either by plate fragmentation [Cooper et al., 1987; Tessensohn and Wo¨rner, 1991; Behrendt et al., 1991a; Busetti et al., 1999; Trey et al., 1999; Cande et al., 2000; Luyendyk et al., 2001; Karner et al., 2005], or by the overriding by East Antarctica of an anomalously hot astenosphere [Smith and Drewry, 1984]. Inconsistencies among the structure of the West Antarctic Rift System, the distribution of rifting, and the great elevation of the mountains do not support simple orthogonal rifting and to reconcile them a plateau collapse scenario, with the Transantarctic Mountains as the plateau edge, has been recently proposed [Bialas et al., 2007]. The evidence that horizontal motions played a first order role on the evolution of this rift system, particularly in its western region, led Tessensohn [1994], Salvini et al. [1997], and Rossetti et al. [2003, 2006a] to interpret the Cenozoic evolution of the West Antarctic Rift System as a passive margin sheared during the Cenozoic. This unusual intraplate horizontal shear has been transferring into the rift system by the strike-slip reactivation in the last 50 Ma of the major fracture zones in the Southern Ocean [Storti et al., 2007] (Figure 1). [8] 5. Because of such significant uncertainty in terms of internal architecture, tectonic evolution, and geodynamic setting, the West Antarctic Rift System has been commonly used as a preferred site for minimizing misfits in global plate circuit reconstructions. Overlap problems between northern Victoria Land and the South Tasman Rise characterize plate tectonic reconstructions of Antarctica with respect to Australia [e.g., Sproll and Dietz, 1969; Weissel and Hayes, 1972] and have been tentatively solved by invoking the occurrence of intraplate deformations in the West Antarctic Rift System [e.g., Royer and Rollett, 1997; Tikku and Cande, 1999; Cande et al., 2000; Norvick and Smith, 2001; Steinberger et al., 2004]. The kinematic architecture and timing of these intraplate deformations are controversial and include roughly E– W extension in Eocene-Oligocene time [Cande et al., 2000] and dextral transtensional rifting older than about 50 Ma [Steinberger et al., 2004]. [9] A better understanding of the structure and development of the West Antarctic Rift System is thus necessary to reduce uncertainty in its evolution and, eventually, in the global plate tectonic circuit. Quite a large structural and apatite fission track data cover at the regional scale is available for the northern part of Victoria Land [Fitzgerald and Gleadow, 1988; Redfield, 1994; Balestrieri et al., 1997,

TC4012

1999; Balestrieri and Bigazzi, 2001; Storti et al., 2001; Rossetti et al., 2002, 2003; Storti and Rossetti, 2002] and for southern Victoria Land [Gleadow and Fitzgerald, 1987; Fitzgerald and Gleadow, 1988; Fitzgerald, 1992; Wilson, 1995, 1999]. Conversely, much less information was available on the Ross Sea shoulder basin-boundary fault system in the central sector of Victoria Land. For this purpose, we performed a structural and thermochronological study in the area between the Reeves Glacier to the north, and the Mawson Glacier to the south (Figures 2 and 3). Our results provide new constraints to the kinematics of the basin boundary fault system along the rift shoulder, on the timing of its formation and, when integrated with the available geological, geophysical and chronological data sets, to the evolution of rifting in the Ross Sea region, in the northern half of the West Antarctic Rift System.

2. Geological Overview of the Ross Sea Region [10] The tectonic picture of the Ross Sea Region (Figure 2) includes dominant NW – SE onshore trends, produced during the Early Paleozoic Ross Orogeny [e.g., Stump, 1992], and dominant N – S offshore trends developed during Mesozoic extension followed by Cenozoic transcurrent and transtensional deformations [e.g., Cooper et al., 1987; Salvini et al., 1997]. The magnetic grain in the area [e.g., Behrendt et al., 1991b; Ferraccioli and Bozzo, 1999, 2003] mimics the tectonic one. 2.1. Onshore Stratigraphy [11] The exhumed roots of the Early Paleozoic Ross Orogen [e.g., Stump, 1992, and references therein] and associated Cambrian-Ordovician Granite Harbour Intrusives [Gunn and Warren, 1962; Armienti et al., 1990] are widely exposed in Victoria Land, where three major exotic terranes are classically recognized [e.g., Bradshaw et al., 1985; Kleindschmidt and Tessensohn, 1987] (Figure 2). The Upper Paleozoic to Triassic clastic rocks of the Beacon Supergroup [e.g., Barrett et al., 1972; Wolfe and Barrett, 1995] unconformably overly the Early Paleozoic terrane assembly south of the Leap Year Fault (Figure 2). They are in turn intruded and overlain by the Jurassic sills of the Ferrar Dolerite and flood basalts of the Kirkpatrik Basalt, respectively [Grindley, 1963; Elliot et al., 1986]. No evidence for the existence of the sub-Beacon erosional surface and/or the overlying Mesozoic rocks has been found to the NE of the Leap Year Fault [GANOVEX Team, 1987; Van der Wateren et al., 1999]. An elongated, roughly N – S trending belt of Cenozoic effusive and intrusive rocks of the McMurdo Volcanic Group [Kyle and Cole, 1974; LeMasurier, 1990] and Meander Intrusive Group [Tonarini et al., 1997], respectively, occurs along the coastal region of Victoria Land and in the western Ross Sea (Figure 2). The ages of these rocks span from about 48 Ma to Quaternary [Tonarini et al., 1997; Armienti and Baroni, 1999; Wo¨rner, 1999; Rocchi et al., 2002], being the Mt. Ritmann, Mt. Melbourne, and Mt. Erebus active volcanoes part of this magmatic belt.

3 of 21

TC4012


Figure 2. (a) Tectonic sketch map of the Ross Sea region [after Salvini et al., 1997; Salvini and Storti, 1999] showing location of the study area. The three major terranes identified in north Victoria Land are indicated. The Wilson Terrane consists of abundant granitic rocks of the Cambrian-Ordovician Granite Harbour Intrusives [Gunn and Warren, 1962; Armienti et al., 1990], with remnants of low- to high-grade metamorphic rocks [e.g., Stump, 1992, and references therein]. The Bowers Terrane is a narrow and elongated belt of Cambrian low grade metavolcanic and metavolcanoclastic rocks [e.g., Weaver et al., 1984]. The Robertson Bay Terrane mostly consists of slightly metamorphic Cambrian to Ordovician clastic rocks [e.g., Burrett and Findlay, 1984; GANOVEX Team, 1987; Rossetti et al., 2006b]. Location of deep wells described in the text is indicated. (b) Schematic geologic section across the Ross Sea [after Busetti et al., 1999].

4 of 21

TC4012

TC4012


Figure 3 5 of 21

TC4012

TC4012


2.2. Offshore Stratigraphy [12] Four major depocentres (Northern Basin, Victoria Land Basin, Central Trough and Eastern Basin) separated from each other by basement highs (Coulman High and Central High) underlie the Ross Sea [Cooper et al., 1987; Behrendt and Cooper, 1991; Davey and Brancolini, 1995] (Figure 2). The Victoria Land Basin, adjacent to the Transantarctic Mountains, is the westernmost and deepest one, with an estimated maximum sediment thickness of more than 12– 14 km in the Terror Rift, an elongated Cenozoic trough developed within the basin [Cooper and Davey, 1985; Cooper et al., 1987]. To the northeast, the Northern Basin projects into the Southern Oceans and reaches a maximum sediment thickness of about 5.5 km [Brancolini et al., 1995]. The Central Trough is formed by the envelope of three right-stepped depocentres, which are offset from south to north by the Aviator Fault and Lanterman Fault, the Tucker Fault and the Adare Fault, respectively (Figure 2). The estimated magnitude of these dextral offsets is of some tens of kilometres [Salvini et al., 1997]. Sediment thickness reaches about 6 km in the southern depocenter [Brancolini et al., 1995]. The Eastern Basin is the widest depocenter in the Ross Sea and shallows toward Marie Byrd Land, to the east. The deeper part of the depocenter (up to about 9 km of sediments [Brancolini et al., 1995]) occurs on the inner shelf, while the architecture of the outer shelf resembles the basin and range style of the western rift shoulder in Marie Byrd Land [Busetti et al., 1999]. [13] The depocentres in the Ross Sea are presumed to contain Cretaceous to Paleogene early rift sediments separated from the overlying Neogene ones by the widespread RSU6 unconformity [Hinz and Block, 1984; Cooper et al., 1987; Brancolini et al., 1995; De Santis et al., 1999; Luyendyk et al., 2001]. This unconformity is interpreted to separate Mesozoic from Cenozoic deformation events in the Ross Sea [e.g., Salvini et al., 1997]. The age of RSU6, however, is still uncertain. Hinz and Block [1984] and Hinz and Kristofferson [1987] gave to the RSU6 a Late Oligocene age, while Cooper et al. [1990] proposed an Eocene age. Busetti [1994], in a review paper, interpreted the age of the RSU6 unconformity as either Late Eocene (42 Ma) or Early Oligocene (30 Ma). On the other hand, the revised correlation of the seismic stratigraphy of the Victoria Land Basin with the data from the Cape Roberts wells led Davey et al. [2000] to propose an age of less than about 17 Ma. Such discrepancies reflect the scarcity of data constraining the timing of rifting in the Ross Sea and may suggest a diachrony of the RSU6 across the Ross Sea. [14] Few wells penetrated the sedimentary sequences in the Ross Sea, providing some chronological information (Figure 2). Deep Sea Drilling Projects (DSDP) wells 270 to 273 recovered glacio-marine sediments from early Late Oligocene (DSDP 270), Early Miocene (DSDP 272 and

TC4012

273) and Early Pliocene (DSDP 271) to the present, respectively [e.g., Brancolini et al., 1995]. Well MSSTS-1 cored through a Pliocene-Pleistocene glacial section into Late Oligocene glacial sediments [Barrett, 1986]. The base of the CIROS-1 core has been re-interpreted as early Late Oligocene [Wilson et al., 1998]. Core drilling in the small Cape Roberts Rift Basin [Hamilton et al., 2001] by the Cape Roberts Project [Cape Roberts Science Team, 1998, 1999, 2000] showed that the homoclinal section near the western margin of the Victoria Land Basin (Figure 2) ranges in age from latest Eocene (ca 34 Ma) to late Early Miocene (ca 17 Ma) and is unconformably overlain by a generally thin Plio-Pleistocene succession. 2.3. Cenozoic Tectonic Architecture [15] The Cenozoic tectonic architecture of the Ross Sea region is characterized by the reactivation of preexisting, inherited structural fabrics during later deformational events. In particular, the NW –SE and N – S trends of the Ross Orogen were reactivated during Mesozoic rifting between East and West Antarctica [e.g., Salvini and Storti, 1999]. The initial rift architecture in the Ross Sea was then overprinted by Cenozoic dextral transcurrence and transtension [Salvini et al., 1997]. The occurrence of multiple reactivation events eventually produced a complex 3-D distribution of Cenozoic fault geometry and kinematics (Figure 2). [ 16 ] Well-developed NW – SE striking, right-lateral strike-sip fault systems occur in northern Victoria Land, including the reactivation of the Lanterman and Leap Year faults [Salvini et al., 1997]. The cumulative right-lateral displacement associated with these fault systems splaying off the Tasman intraplate right-lateral strike-slip belt (Figure 1) has been estimated at about 300 km from the abrupt southeastward shift of the continental platform across this shear zone [Storti et al., 2007]. Large structural data sets have been collected on the Cenozoic architecture of the right-lateral strike-slip Lanterman Fault [Rossetti et al., 2002] and Priestley Fault [Storti et al., 2001]. Approaching their southeastern terminations, the major right-lateral strike-slip fault systems are abutted by the N – S striking basin-boundary fault systems in the western and central Ross Sea and transfer to them their residual dextral displacement in a crustal scale horsetail structure [Salvini et al., 1997] (Figure 2). This interpretation is questioned by Davey et al. [2006], who neglect the occurrence of significant Cenozoic transcurrent motions in the Ross Sea and Adare Basin to the north [Cande and Stock, 2006] (Figure 1) to support a persisting orthogonal extensional scenario where the offset basins are kinematically linked by NE – SW striking transfer fault systems. [17] Offshore stratigraphic record in seismic profiles clearly indicates that the earlier extensional deformations, when preserved, do not offset the RSU6 unconformity

Figure 3. Equal area stereographic projection (Schmidt net, lower hemisphere) of structural data collected in the study area. Field analysis sites listed in Appendix A are grouped into 12 field localities. The satellite image is from NASA (http:// earthobservatory.nasa.gov/Newsroom/NewImages/). See text for details. 6 of 21

TC4012


Figure 4 7 of 21

TC4012

TC4012


[Cooper et al., 1987; Salvini et al., 1997]. On the other hand, the transtensionally reactivated former extensional faults and the newly formed ones, as well as the offshore NW – SE segments of the major right-lateral strike-slip fault systems, cut across the RSU6 [Salvini et al., 1997]. In many cases, Cenozoic offshore faults have a bathymetric expression suggesting recent activity [e.g., Rossetti et al., 2006a]. In particular, high-resolution reflection seismic profiles cutting across the Cape Adare Fault (Figure 2) indicate the presence of well developed positive flower structures of modern age [Spezie et al., 1993; Storti et al., 2007]. Direct onshore age constraints to fault activity are provided by 40 Ar/39Ar radiometric dating of pseudotachylyte veins at the southern tip of the Priestley Fault, which gave ages of about 34 Ma [Di Vincenzo et al., 2004]. Furthermore, McMurdo dykes occurring close to the synkinematic ones exposed in the tip region of the Priestley Fault [Storti et al., 2001] were dated by Mu¨ller et al. [1990] at about 35 Ma. Preliminary GPS data [Negusini et al., 2005] suggest that deformation is still active today [Rossetti et al., 2006a].

3. Structural Data [18] Structural data were collected at 32 field analysis sites, grouped into 12 field localities. The position of each field site is provided in Appendix A. The collected data include the attitude of fault surfaces and of slickenlines associated with them (Figure 3). Shear sense criteria were provided by grooves, corrugations, and by subsidiary faults and fractures associated with the master slip surfaces [Petit, 1987] and, in few cases, by quartz shear fibres on slickensides (Figure 4). In many places, pseudotachylytes formed along slip surfaces (Figures 4b and 4c). Faulting frequently localized along the boundaries of mafic dykes (Figure 4d), which are abundant in the area [Rossetti et al., 2000]. Cumulative analysis of the total data set shows that faults are steeply dipping to near vertical (Figures 4e – 4h). Cumulative azimuth data in each field locality were analyzed by polymodal Gaussian distribution statistics to obtain average orientations of the occurring different fault types [Storti et al., 2006]. Results are schematically illustrated in Figure 5. A qualitative description of the orientation and kinematics of the most abundant fault sets recognized in the visited localities is provided below, while quantitative results of their statistical analysis are listed in Appendix B.

TC4012

[19] Moving along the southern flank of the Reeves Glacier and the Ross Sea coast north of the Drigalsky Ice Tongue, NW – SE (localities 1, 2, and 4 in Figure 3) and NNW – SSE striking (localities 3 and 4) right-lateral strikeslip faults occur and become the dominant structures at locality 2, where many of them show cataclastic cores. NNE – SSW right-lateral strike-slip faults are abundant at locality 1. Extensional faults and few reverse ones also occur and are oriented parallel to the right-lateral strike-slip ones. In particular, NNE – SSW extensional faults are abundant at locality 3 where, in many cases, they splay off from the tips of the NNW– SSE right-lateral strike-slip faults. Left-lateral strike-slip faults oriented NE – SW occur in all the four field localities (Figure 5). The two localities to the west (5 and 7 in Figure 3) are characterized by NE – SW to E – W striking left-lateral strike-slip faults. NNE – SSW striking right-lateral strike-slip faults are abundant at locality 5, with few extensional faults paralleling them (Figure 5). Data collected on both sites of the David Glacier (localities 6 and 8 in Figure 3) show the presence of N– S striking right-lateral strike-slip faults, which are best developed at locality 8 where cataclastic cores hosting pseudotachylyte veins are abundant. The same locality is characterized by the abundance of NNE – SSW striking right-lateral strikeslip faults with few reverse faults paralleling them. NE – SW striking left-lateral strike-slip faults occurs on both sides of the David Glacier and are more abundant at locality 6, while extensional faults occur on the southern side and have comparable orientation (Figure 5). The coastal area to the south (localities 9, 10, and 11 in Figure 3) is characterized by the presence of NNE –SSW to NE – SW (locality 11) right-lateral strike-slip faults. In many places, pseudotachylyte veins occur within these fault zones. NE – SW striking left-lateral strike-slip faults were also found at localities 9 and 10, while extensional faults with similar orientation occur in all the three localities (Figure 5). Finally, data collected on the northern side of the Mawson Glacier (locality 12 in Figure 3) indicate the presence of a segment of a major N– S oriented fault system exposed in a large deglaciated area (Figure 6). First order fault strands have metre- to decametre-thick cataclastic cores including ultracataclasites and pseudotachylytes, which impart them blackish tones that strongly contrast with the reddish-pink to whitish colors of the country rocks. Subsidiary faults in the damage zones are mostly NNW – SSE striking right-lateral strike-slip faults. In many cases, extensional faults splay

Figure 4. (a) Right-lateral strike-slip fault slickenside in granite at Starr Nunatak (locality 10 in Figure 3). The white arrow is parallel to the slickenlines and points to a pencil for scale. (b) Plan view of a pseudotachylyte-bearing subsidiary left-lateral strike-slip fault zone at Prior Island (locality 7 in Figure 3). (c) Plan view of a pseudotachylyte bearing left-lateral strike-slip subsidiary fault zone at Mt. Priestley (locality 7 in Figure 3). Note the small aplitic dyke offset by the fault. (d) Mafic dyke boundary reactivated as a right-lateral strike-slip fault zone at Tarn Flat (locality 3 in Figure 3). (e) Cumulative equal area stereographic projection (Schmidt net, lower hemisphere) of contoured right-lateral strike-slip fault poles and related slickenlines (black dots). (f) Cumulative equal area stereographic projection (Schmidt net, lower hemisphere) of contoured left-lateral strike-slip fault poles and related slickenlines (black dots). (g) Cumulative equal area stereographic projection (Schmidt net, lower hemisphere) of contoured extensional fault poles and related slickenlines (black dots). (h) Cumulative equal area stereographic projection (Schmidt net, lower hemisphere) of contoured reverse fault poles and related slickenlines (black dots). 8 of 21

TC4012


Figure 5 9 of 21

TC4012

TC4012


TC4012

Figure 6. (a) Oblique view from the helicopter of a N-S- striking right-lateral strike-slip fault zone near the Mawson Glacier (locality 12 in Figure 4). The fault zone in places offsets mafic dykes and, in adjacent places, is partially crosscut by similar dykes. (b) Field view from the field (looking from the north) of the same fault zone. The Mawson Glacier is in the background.

from the tip of the right-lateral strike-slip ones in horsetail arrays. Left-lateral strike-slip faults are also abundant and strike NE – SW, while extensional faults are oriented NNE – SSW (Figure 5). In places, near vertical basaltic dykes either cut across or are arrested against the same strike-slip fault

strand (Figure 6a). This suggests that magma injection there postdated or was overally coeval to faulting. [20] In summary, the investigated sector of the western Ross Sea shoulder is characterized by a brittle fault pattern that typically includes NNE – SSW striking right-lateral

Figure 5. (a) Average azimuthal trends of faults recognized in the 12 field locations (results are provided in Appendix B). The satellite image is from NASA (http://earthobservatory.nasa.gov/Newsroom/NewImages/). (b) Cumulative Gaussian distribution statistics of average azimuthal trends in Figure 5a, analyzed by fault kinematic type. Histograms and the corresponding Gaussian best fit curves are shown and indicate the good azimuthal clustering of strike-slip faults and the poor clustering of dip-slip ones, characterized by a much wider azimuthal variability in the different field sites. 10 of 21

TC4012


Figure 7. Sample locations and apatite fission track ages in the study area. The satellite image is from NASA (http:// earthobservatory.nasa.gov/Newsroom/NewImages/). See text for details. strike-slip and extensional subsidiary faults, and NE – SW striking left-lateral strike-slip subsidiary faults (Figure 5 and Appendix B). Three segments of major N – S striking rightlateral strike-slip fault systems are exposed at Tarn Flat, Cape Philippi, and Mawson Glacier, respectively. A major segment of a NW – SE oriented right-lateral strike-slip fault system is exposed at Mt. Gerlache, at the northern end of the study area. When observed, overprinting relationships indicate that, in many cases, right-lateral strike-slip motions postdated extensional ones. In some cases, this behavior is violated and opposite overprinting relationships occurs. Reverse faults are in many cases parallel to the strike-slip ones and arranged in positive flower structures [Harding, 1985], where both dip-slip and strike-slip striations mutually overprint.

4. Apatite Fission Track Analysis 4.1. Method and Sampling [21] Fission tracks are linear damage zones produced in apatite crystals by the radioactive decay of 238U. The density of fission tracks depends on the time over which

TC4012

tracks have accumulated. The length of confined fission tracks is proportional to the maximum temperature experienced by each track [e.g., Green et al., 1986]. Over geological time, fission tracks are partially retained in apatite at temperature between 60° and 120°C (Partial Annealing zone, PAZ) with a mean closure temperature at 110° ± 10°C [Green and Duddy, 1989]. The sampling strategy implying the collection of samples along vertical profiles has been extensively applied in the Transantartic Mountains. Exhumed paleo-PAZ has been identified in these profiles with recognition of break in slopes interpreted as the base of an exhumed paleo-PAZ [Gleadow and Fitzgerald, 1987; Fitzgerald and Gleadow, 1990]. The break in slope defines the transition from a time of relative thermal and tectonic stability to a time of rapid cooling due to exhumation. Samples above the break show confined track length distribution with significantly shortened mean length and broad standard deviation, the distribution being composed by tracks formed in the pre-exhumation PAZ and by a later set of post-exhumation tracks. Confined track length distributions below the break reflect rapid cooling (mean track length >14 mm, standard deviation