Abstract

The Bruin Bay fault system defines the northwestern tectonic boundary of the Cook Inlet forearc basin for ∼450 km along the southern Alaskan forearc. The age, origin, and tectonic significance of the fault system are not well understood. We present field observations and a population of minor fault slip data (n = 296) collected within the Bruin Bay fault system from the Iniskin-Tuxedni region of the Cook Inlet. The minor faults cut Triassic–Paleogene strata and are subdivided into two kinematically distinct populations. Population A (n = 233, 79%) includes strike-slip and reverse faults that altogether record subhorizontal, southeast-trending tectonic shortening and bulk sinistral transpression. Population B (n = 63, 21%) includes strike-slip faults that are compatible with subhorizontal, northeast-trending shortening and southeast-trending extension and are younger. Gently deformed mafic and felsic dikes that intrude cataclasite within the Bruin Bay fault zone at two localities yield late Paleogene biotite (ca. 37 Ma), whole-rock (ca. 33 Ma), and plagioclase (ca. 31 Ma) 40Ar/39Ar ages. The ages of deformed strata and crosscutting dikes indicate that sinistral transpression (population A) occurred during the Paleogene prior to ca. 37 Ma, but some deformation persisted through at least the early Oligocene. Results place the Bruin Bay fault system in the Paleogene tectonic context of southern Alaska. We discuss several competing hypotheses to interpret the tectonic evolution of the fault system. We suggest that the majority of the Paleogene deformation likely occurred during either a spreading ridge subduction event or accretion of the Chugach–Prince William terrane to the southern Alaskan margin.

INTRODUCTION

Convergence along ocean-continent subduction margins shapes the geology of the overriding plate for hundreds of kilometers inboard of the trench (e.g., the American Cordillera). Accretionary prisms, forearc basins, and volcanic arcs are geologic features characteristic of ocean-continent subduction margins that are commonly separated from each other by large faults systems (e.g., Border Ranges fault system, Alaska; Pavlis and Roeske, 2007; Wilson et al., 2012; Sagaing fault system, Myanmar; Vigny et al., 2003; Wang et al., 2014; the megasplay fault, Nankai margin, Japan; Gulick et al., 2010; Sumatran fault, Indonesia; Sieh and Natawidjaja, 2000). Their tectonic development varies widely and is strongly dependent on the plate kinematics of the convergent margin, sediment flux at the trench, configuration of the subducting plate (e.g., buoyancy or roughness), and/or arrival of allochthonous terranes (Dickinson and Seely, 1979; Kopp, 2013; Noda, 2016). Long-lived subduction margins can lead to the reactivation of preexisting structures (Holdsworth et al., 1997) and the development of fault systems with slip and deformation histories that record both the initial tectonic setting and their reactivation under different stress regimes (Cembrano et al., 1996; Murphy et al., 1999). During plate convergence, these structures can also facilitate the transfer of slip deep into continental interiors (Storti et al., 2003). The transpressive nature of these long-lived structural zones is recorded in their magmatic and deformation history, which can be constrained through structural, petrological, geochronological and thermochronology analysis (e.g., Stewart et al., 1999; Benowitz et al., 2011; Niemi et al., 2013). Therefore, the composite geologic history of long-lived structures can help to constrain plate boundary conditions over significant amounts of time.

The Cook Inlet Basin (CIB) in south-central Alaska is a >25,000 km2 forearc basin that records Jurassic–modern tectonic evolution of the southern Alaskan subduction margin (Fig. 1; see reviews by Trop and Ridgway, 2007; LePain et al., 2013). The CIB is part of the Wrangellia composite terrane (Fig. 1), which is located between the Mesozoic to Cenozoic accretionary prism (Chugach–Prince William terrane) and interior Alaska. The CIB is bound on either side by regional long-lived tectonic boundaries known as the Bruin Bay (Magoon et al., 1976; Wilson et al., 2012) and Border Ranges fault systems (e.g., Pavlis and Roeske, 2007).

The late Mesozoic–early Cenozoic tectonic setting of south-central Alaskan subduction margin is widely debated. For example, some studies conclude that many of the geologic features characteristic of the Alaskan forearc (i.e., Peninsular and Chugach–Prince William terranes; Fig. 1) were strongly affected by the Paleogene subduction of a trench-ridge-trench triple junction in a paleogeographic setting that resembled the modern margin (e.g., Bradley et al., 2000; Haeussler et al., 2003a, 2003b). In contrast, other work indicates that much of this record (i.e., the Chugach–Prince William terrane) was likely translated a long distance (as much as 3000 km) along the North American margin to its present location between the Late Cretaceous and Eocene (e.g., Plumley et al., 1983; Bol et al., 1992; Garver and Davidson, 2015; Day et al., 2016). Despite extensive work on the Mesozoic–Cenozoic tectonic evolution of southern Alaska (e.g., Little, 1990, 1992; Pavlis and Sisson, 2003a, 2003b; Roeske et al., 2003; Ridgway et al., 2007; Trop, 2008; Finzel et al., 2016) there is still much to learn and additional studies are needed to better elucidate the tectonic history of this difficult to access and long-lived continental forearc.

This study focuses on the kinematic evolution, relative timing, and tectonic significance of brittle deformation that occurred within an understudied fault system in the CIB, known as the Bruin Bay fault system (BBFS). The BBFS generally separates Mesozoic–Cenozoic clastic rocks of the CIB from the exhumed roots of a Jurassic–Cretaceous volcanic arc (the Talkeetna arc; Fig. 1). We present field observations and kinematic analyses from 296 minor fault surfaces that are exposed within the BBFS as well as three new 40Ar/39Ar ages from two dikes that intrude the main fault zone of the BBFS. The data describe the bulk kinematics of deformation in the BBFS and define two kinematically distinct fault populations. Results indicate that the majority of the deformation within the BBFS records Paleogene sinistral transpression. We discuss several interpretations of the BBFS in the context of previously proposed tectonic models for early Cenozoic plate reorganization near the southern Alaskan margin.

GEOLOGICAL SETTING

General Tectonic Background of the Cook Inlet Basin and BBFS

The Mesozoic stratigraphic record within the CIB reflects the Triassic–Jurassic development of the oceanic Talkeetna arc and accretion of the arc as part of the Wrangellia composite terrane to North America during the Late Jurassic–Cretaceous (Fig. 1; Plafker et al., 1989, 1994; Clift et al., 2005a, 2005b; Trop et al., 2003; Trop and Ridgway, 2007; Hampton et al., 2010). Outboard of the arc, the CIB records sedimentation in a forearc setting from the Middle Jurassic to Cenozoic (e.g., Trop and Ridgway, 2007; LePain et al., 2013). The Cenozoic geologic record within the CIB reflects several tectonic events that shaped southern Alaska: (1) Paleocene–Eocene subduction of a spreading ridge (Bradley et al., 2000; e.g., Haeussler et al., 2003a, 2003b; Pavlis and Sisson, 2003b; Benowitz et al., 2012); (2) Paleogene dextral translation and accretion of the Chugach–Prince William terrane (Cowan et al., 1997; Butler et al., 2001; Cowan, 2003; Garver and Davidson, 2015; Day et al., 2016); (3) proposed oroclinal bending of southern Alaska (Hillhouse and Coe, 1994; Glen, 2004); and (4) the Oligocene to present flat-slab subduction collision of the Yakutat microplate (Plafker et al., 1978; Eberhart-Phillips et al., 2006; Finzel et al., 2011b; Benowitz et al., 2014; Lease et al., 2016). Although Mesozoic–Cenozoic deformation inflicted on the forearc is well described for the upper CIB (Bruhn and Pavlis, 1981; Little, 1990, 1992; Pavlis and Sisson, 2003a, 2003b; Bruhn and Haeussler, 2006; Pavlis and Roeske, 2007) and accretionary prism (Kusky et al., 1997; Haeussler et al., 2003a, 2003b, Pavlis and Roeske, 2007; Wallace, 2013), the tectonic significance of the BBFS and associated deformation in the lower CIB is not well understood.

The BBFS extends for ∼450 km from the upper Alaska Peninsula to its terminus near the upper Cook Inlet where it is likely truncated by the Castle Mountain fault (Fig. 1; Trop et al., 2005). For most of its length, the fault system separates Jurassic plutonic and volcanic rocks of the Talkeetna arc from Middle Jurassic to Neogene sedimentary rocks of the CIB (e.g., Wilson et al., 2012; Fig. 1). Provenance studies conclude that the ∼3000-m-thick succession of middle–upper Jurassic volcanogenic mudstone and sandstone in the CIB record the uplift and denudation of the Talkeetna arc in the hanging wall of the of the BBF during an inferred episode of Middle Jurassic reverse slip (LePain et al., 2011; Wartes et al., 2013; Trop et al., 2005). Other studies determined that oblique left-reverse (this study; Rosenthal et al., 2017), and left-normal (Decker et al., 2008) motion occurred along different parts of the BBFS during Paleogene to Neogene deformation. Prior work in the CIB suggests that the BBFS may record a tectonic history that spans from the Middle Jurassic to the Neogene.

Overview of Stratigraphy in the Study Area

In the study area (Figs. 1 and 2), the main strand of the BBF dips moderately to steeply west-northwest and generally divides the crystalline rocks of the Talkeetna arc from clastic rocks of the CIB. In the footwall of the fault system, Middle–Late Jurassic marine strata include a >1500-m-thick succession of volcanogenic shale, siltstone, and sandstone that form the Tuxedni Group (Aalenian–Bathonian; Detterman and Hartsock, 1966; Stanley et al., 2013; LePain et al., 2013). Conformably overlying the Tuxedni Group is a >700-m-thick succession of lithic-rich sandy siltstone and fine-grained sandstone known as the Chinitna Formation (upper Bathonian–Callovian; Imlay, 1953; Detterman and Hartsock, 1966; Herriott and Wartes, 2014). The Naknek Formation unconformably overlies the Chinitna Formation and consists of >1250 m of lithic and arkosic sandstone, siltstone, and conglomerate (Oxfordian–lower Kimmeridgian; Detterman and Hartsock, 1966; Wartes et al., 2013; LePain et al., 2013; Herriott et al., 2017). North of Chinitna Bay (Fig. 2), Maastrichtian nonmarine strata including conglomerate, sandstone, mudstone, and coal that are part of the Saddle Mountain section (83 m thick) crop out and unconformably overlie the Naknek Formation (Magoon et al., 1980; LePain et al., 2012; Gillis, 2016).Late Paleogene(?) or early Miocene(?) nonmarine conglomerate and sandstone overlie the Saddle Mountain section (thickness unknown). The Cenozoic unit was interpreted by Magoon et al. (1976, 1980) as part of the Tertiary West Foreland Formation; however, biostratigraphic ages from Wolfe et al. (1965) place it in the Seldovian, which spans into the Oligocene and Miocene. For the purpose of this paper we forego naming the unit and instead refer it as late Paleogene(?) strata, and we acknowledge that it may span into the early Miocene.

Rocks exposed in the hanging wall of the BBFS include upper Triassic–lower Jurassic carbonate deposits that are interbedded with mafic volcanic and volcaniclastic deposits and together define the Kamishak Formation (Detterman and Reed, 1980; Wang et al., 1988; Detterman et al., 1996; Blodgett, 2008; Whalen and Beatty, 2008). The Talkeetna Formation (lower Jurassic; Martin, 1926) overlies the Kamishak Formation and is a thick (>2 km) succession of mafic to intermediate composition lavas, sills, volcanic breccia, and volcaniclastic sandstone to pebble breccia conglomerate that largely represent submarine volcanism (e.g., Clift et al., 2005b; Bull, 2014, 2015). The depositional contact between the Kamishak and Talkeetna Formations is gradational (Detterman et al., 1996; Palfy et al., 1999; Whalen and Beatty, 2008). Detterman and Hartsock (1966) divided the Talkeetna Formation into three members, which were later reduced to two informal subunits by Bull (2014, 2015) and are herein referred to as the lower and upper Talkeetna Formation (i.e., Fig. 2). Below and intruding the Talkeetna Formation, a suite of Jurassic quartz monzonite and quartz diorite defines the plutonic roots of the Talkeetna arc (Detterman and Hartsock, 1966; Clift et al., 2005a, 2005b). In the northern part of the study area the BBF cuts across the Talkeetna Formation, and thus the Talkeetna Formation occurs in the footwall of the fault near Tuxedni Bay (Fig. 2).

Overview of the Structure in the Study Area

Hartsock (1954) published the first map of the BBF on the Iniskin Peninsula. Detterman and Hartsock (1966) and Detterman and Reed (1980) inferred that the fault has accommodated 19–65 km of sinistral displacement since the Early Jurassic on the basis of poorly constrained stratigraphic piercing points. Detterman and Hartsock (1966) mapped a system of steeply dipping northeast-striking faults in the hanging wall of the BBF (Fig. 2) that locally juxtapose the Talkeetna and Kamishak Formations. Detterman and Reed (1980) postulated that the steeply dipping faults are part of the BBFS. Reconnaissance work (Gillis et al., 2013) focused on a well-exposed segment of the BBF southwest of the Iniskin Peninsula and indicated a component of top-to-the-southeast reverse motion along a northwest-dipping (40°–50°) fault plane.

On the Iniskin Peninsula in the footwall of the BBF (Fig. 2), the Tuxedni Group, and Chinitna and Naknek Formations are folded by upright open folds that trend northeast and form a regional syncline-anticline pair with wavelengths of ∼5 km. The folds occur between overlapping fault segments within the BBFS (i.e., Bruin Bay and Fitz Creek faults; Fig. 2; Wartes et al., 2016) that record apparent top-to-the-southeast reverse stratigraphic separation (Fig. 2). Map patterns indicate that the folds plunge shallowly southwest and northeast, and that they decrease in amplitude down-plunge in both directions (Fig. 2). Mapping campaigns undertaken by the Alaska Division of Geological and Geophysical Surveys (Gillis et al., 2014; Herriott, 2016) reveal several other map-scale faults in the study area that are both subparallel and oblique to the main segments of the BBFS (Fig. 2). To better determine the fault slip kinematics and tectonic significance of brittle deformation within the BBFS, we describe several exposures of the BBFS and present a new fault kinematic data set of 296 faults from the Iniskin-Tuxedni region of the CIB (Figs. 1 and 2).

METHODS

The study area was divided into seven domains (I–VII in Fig. 2) that represent parts of the BBFS with unique structural relationships or regions containing secondary structures that formed in the footwall or hanging wall of the BBF. Minor fault slip data were collected within each domain and integrated with field observations and map patterns to interpret the bulk kinematics of the BBFS.

Fault Slip Analyses

Fault slip data were collected by measuring the attitude and slip lineations of minor fault surfaces (n = 296). The sense of slip for each fault was determined using kinematic indicators including Riedel shears, fault surface asperities, preferred orientations of associated tensile or sigmoidal veins, and other common methods (e.g., Petit, 1987; Figs. 3A–3C). The quality of the kinematic indicators was ranked on a three point scale (A–C) and considered when interpreting the data. Faults that preserved either measureable offset markers or multiple corroborating kinematic indicators were given an A rank to indicate high confidence in the sense of slip (e.g., Fig. 3A). If offset markers were absent but at least one quality set of kinematic indicators was preserved (e.g., Fig. 3B) the faults were given a B rank. Faults with poor or equivocal kinematic indicators were given a C rank (e.g., Fig. 3C). If a slip lineation was not preserved, the fault was excluded from the kinematic analysis. Histograms showing the distribution of fault quality rankings for each domain are presented with the results (Figs. 4 and 610).

Kinematic axes (i.e., incremental shortening, P; extension, T; and null, B) were calculated for each datum and plotted on a stereogram using the software FaultKin (Marrett and Allmendinger, 1990). Shortening (P) and extension (T) axes for an increment of movement on a fault are contained within the plane defined by the slip lineation and pole to the fault surface (i.e., the movement plane) and bisect the angle between them. The sense of slip determines the shortening and extension axis (Fig. 3D). The null axis (B) is the pole to the movement plane (i.e., M pole of Arthaud [1969]) and represents the axis of zero strain for an increment of motion on a fault, assuming plane strain deformation (Fig. 3D). Populations of fault slip data and kinematic axes were analyzed to interpret the bulk kinematics of each structural domain.

Kinematically compatible, scale-invariant faults should form subsets that are defined by subparallel kinematic axes (Marrett and Allmendinger, 1990). We delineate subsets of minor faults within each domain by the attitudes of the fault surfaces, striations, and the sense of slip. The subsets were grouped to form fault populations with subparallel P and T axes. To determine the principal shortening and extension axes for fault subsets within each domain, directional maxima (e1—extension, e2—intermediate, e3—shortening) for clusters of P and T axes were calculated using the linked Bingham statistics function of FaultKin (Fig. 3E; Marrett and Allmendinger, 1990; Allmendinger et al., 2012). The faults are assumed to be kinematically scale invariant (Marrett and Allmendinger, 1992; Turcotte, 1986); this is qualitatively supported by fault populations in this study with gouge thicknesses that span four orders of magnitude (10–3–101 m).

B axis distributions for fault subsets are also a useful descriptor of bulk deformation and were considered separately. Unlike P and T axes, the B axis orientation is independent of both the sense of slip and the assumption about the magnitude of the angle between the strain axes and the fault plane. The distribution of B axes can also be used to determine the orientations of the principal strain axes of a fault population (X, extension; >Y, intermediate; >Z, shortening). In plane strain deformation Y = 0, in triaxial deformation Y ≠ 0 (Arthaud, 1969; Bruhn and Pavlis, 1981). The absolute magnitude of X, Y, and Z cannot be determined unless the slip magnitude is known for all of the faults in the population; however, the orientations of X, Y, and Z can be inferred from the distribution of kinematic axes. B axis distributions that form a point maxima on a stereogram reflect plane strain (Fig. 3E) and the orientation of the point maxima represents the intermediate principal strain axis (Y) of the population. If B axes form a girdle, then the faults record triaxial deformation and the girdle contains the Y axis as well as either X or Z. The pole to girdle defines the odd axis (either X or Z), which is determined by the orientations and sense of slip of the faults in the population (Fig. 3E). Principal strain axes for kinematically compatible subsets of faults in each domain were determined using P and T axes as well as B axis distributions. Results were compared to evaluate the bulk kinematics of the BBFS. For large fault sets (n > 25), 95% confidence cones (α95) were calculated for principal strain axes using Bingham statistics (Allmendinger et al., 2013; Cardozo and Allmendinger, 2013).

RESULTS

Field Observations and Fault Kinematic Results from the BBFS

In the following, field relationships and fault kinematic data are presented separately for each of seven structural domains in the study area. The discussion generally progresses from south to north. Field relationships from an additional outcrop of the BBF located outside of the study area (Contact Point, Fig. 1) are also presented, but this locality is not part of the kinematic analysis.

Domain I: Ursus Head

Ursus Head is a prominent ∼2-km-wide, south-facing seacliff located southwest of Iniskin Bay (Figs. 2 and 4). The BBF is well exposed at Ursus Head and dips northwest, juxtaposing the Kamishak Formation above the Naknek Formation. In the footwall, bedding of the Naknek Formation is folded into an overturned to the southeast, tight syncline that trends northeast (Figs. 4A, 4B). In the hanging wall, the Kamishak Formation consists of silicic and intermediate composition volcanic rocks and overlying thinly bedded limestone. Several low-angle thrust faults (subparallel to bedding) imbricate and thicken the limestone (Figs. 4B, 5A, and 5B). Felsic and mafic dikes (zircon U-Pb ages are ca. 207 Ma and ca. 213 Ma for two specimens; Gillis et al., 2017) oriented approximately orthogonal to bedding are offset by the thrusts and indicate a top-to-the-southeast sense of slip. Asymmetric minor folds in bedding are commonly breached by small, northwest-dipping thrust faults with several centimeters of displacement (Fig. 5A). The minor folds and thrusts confirm the top-to-the-southeast sense of slip that is inferred by the separation of the dikes. The contact between the limestone and underlying volcanic rocks is sheared and forms a local detachment horizon. Bedding within the basal ∼20 m of the limestone is folded by a train of small (<10 m) detachment folds that are exposed continuously across the outcrop (pink shading in Fig. 4B) and probably reflect the competency contrast between the limestone and underlying volcanic rocks. Four thrust horses (locations I–IV in Figs. 4B and 5B) imbricate the limestone and are rooted in the detachment horizon.

East of the limestone detachment, the BBF crops out in a gulley at the top of a talus cone (Fig. 5C; location in Fig. 4B). Here the fault zone is ∼2 m thick and defined by at least two imbricated thrust surfaces (Fig. 5C). Subvertical cleavage records flattening within the fault zone and confirms a top-to-the-southeast reverse sense of slip based on the asymmetry of the cleavage with respect to the fault. In the footwall, bedding within the Naknek Formation is vertical and truncated by the fault (Fig. 5C). A second outcrop of the BBF is located near the shoreline, downdip from the location of Figure 5C (Fig. 5D; location in Fig. 4B). Here, an ∼10-m-wide fault zone contains altered cataclasite and preserves minor fault surfaces that indicate top-to-the-southeast reverse slip. A small basaltic dikelet intrudes the cataclasite and is only gently deformed (Fig. 5D). A sample of the dikelet was collected for 40Ar/39Ar thermochronology to help constrain the timing of deformation (sample 14JR25; see 40Ar/39Ar thermochronology section). West of the BBF, a steeply dipping reverse fault crops out and truncates the limestone detachment horizon, uplifting stratigraphically lower volcanic rocks in the hanging wall above the limestone (location V in Figs. 4B and 5B). On the basis of similarities in dip, sense of slip, and crosscutting relationships, we infer that the BBF is kinematically similar to the reverse fault in its hanging wall, and thus postdates the low-angle thrusts and the detachment horizon.

Several subvertical, north-northwest–striking faults crosscut the reverse and thrust faults. The subvertical faults are best exposed near the top of the outcrop, where they clearly crosscut the limestone detachment horizon. The apparent sense of displacement is most commonly east-side down (Figs. 4A, 4B). The easternmost of these faults is inferred to crosscut the BBF where it is covered by talus. East of the subvertical fault, the BBF plane projects downdip to a position that is below the exposure of the BBF presented in Figure 5C (Figs. 4B and 5E). Thus, the outcrop in Figure 5C is located within the upthrown block west of the subvertical fault.

Fault Kinematic Data from Domain I

Minor thrust and strike-slip faults with at least several centimeters of displacement are ubiquitous throughout domain I (e.g., Figs. 5A, 5F). Minor thrust faults in the hanging wall of the BBF strike northeast, dip northwest, and record subhorizontal, southeast-trending contraction (P axes) and subvertical extension (T axes; Figs. 4C and 5A–5C). Two subsets of subvertical strike-slip faults with subhorizontal, southeast-trending P axes also occur. Left-lateral faults strike north-northeast and right-lateral faults (e.g., Fig. 5F) strike northwest. The thrust and strike-slip faults were grouped to form a population of minor faults that record subhorizontal southeast-trending shortening and subvertical extension (set A, Fig. 4C). B axes for set A faults form a subvertical, northeast-trending girdle in the XY plane. The Z axis plunges shallowly northwest and is subparallel to e3 (Fig. 4C). Set A faults are compatible with bulk subhorizontal, southeast-trending contraction and reverse slip on the BBF at Ursus Head (Figs. 4A–4C).

An additional subset of minor faults that are incompatible with set A is present and consists of subvertical, north-northwest–striking right-lateral faults with subhorizontal southwest-trending P axes and southeast-trending T axes (set B, Fig. 4D). The outcrop-scale north-northwest–striking subvertical faults that offset the BBF and limestone detachment (e.g., top of Fig. 4B and Fig. 5E) are subparallel to right-lateral strike-slip faults in set B. Thus, we deduce that the outcrop-scale faults are also right lateral, part of set B, and that set B is younger than set A. This is consistent with the apparent east-side-down offset of the northwest-dipping structures and strata that the subvertical faults displace. Set B faults record bulk southwest-trending shortening and southeast-trending extension parallel and perpendicular to, respectively, the map trace of the BBF (Fig. 4D).

Structural relationships at Ursus Head indicate that deformation occurred during at least two separate events. Set A records southeast-trending contraction on both low-angle thrust faults and high-angle reverse faults, including the BBF. Set B is defined by subvertical, right-lateral strike-slip faults that crosscut faults of set A, including the main plane of the BBF, and record southeast-trending extension.

Domain II: The Right Arm Strand of the BBFS

On the Iniskin Peninsula, the BBFS is defined by several fault strands that accommodate both thrust and strike-slip motion (Right Arm, domain II; Iniskin Bay, domain III; Fitz Creek, domain IV; Fig. 2). The apparent stratigraphic separation across the Right Arm strand of the BBF (domain II) decreases between Ursus Head and the Iniskin Peninsula. At Right Arm, the BBF juxtaposes the upper member of the Talkeetna Formation (northwest of the fault) above the Tuxedni Group (southeast of the fault) and is inferred to dip steeply northwest, similar to the dip at Ursus Head. Strata in the footwall of the fault are folded, forming an upright syncline-anticline pair known as the Tonnie syncline and Fitz Creek anticline (Fig. 2). Deformed rocks from the Right Arm strand of the BBFS intermittently crop out in a few locations in a low vegetated stream bed on the Iniskin Peninsula.

Fault Kinematic Results from Domain II

Fault slip data from domain II were collected from the Right Arm strand of the BBFS between Iniskin Bay and Chinitna Bay (Figs. 2 and 6) as well as several subsidiary faults that occur in adjacent hanging wall and footwall outcrops. Minor thrust faults dip southeast and record southeast-trending shortening and subvertical extension (Fig. 6A). Left-lateral faults with southeast-trending P axes form two clusters (Fig. 6A). Those that strike north-northeast are parallel to the map trace of the BBF, and the north-northwest–striking left-lateral faults are parallel to the expected orientation of Riedel shears (R) within the BBFS. Right-lateral faults with southeast-trending P axes are subvertical and strike northwest, forming a subset that is consistent with conjugate Riedel shears (R’). Altogether, the thrust, left-, and right-lateral faults record subhorizontal, southeast-trending shortening and subvertical to northeast-trending extension (set A; Fig. 6A). B axes from set A form a northeast-trending girdle in the XY plane; the Z axis plunges shallowly southeast and is within error of e3 (Fig. 6A). The Y axis is subvertical. The bulk kinematics of set A faults indicate oblique left-reverse slip within the BBFS in domain II.

Domain II also contains a subset of minor faults that are not compatible with set A. Subvertical, north-northwest–striking right-lateral faults (Fig. 6B) cut the upper member of the Talkeetna Formation in the hanging wall of the BBF near Right Arm (Fig. 3A; location shown in Fig. 2). Two northwest-striking, left-lateral faults with southwest-trending P axes were also observed in domain II. The incompatible faults (set B; Fig. 6B) have shallowly plunging, southwest-trending P axes and southeast-trending T axes (Fig. 6B). The B axes define a steeply northwest-plunging point maxima (Y) that is consistent with dominantly strike-slip deformation. One shallowly plunging, southeast-trending B axis partly defines a girdle in the XY plane. The Z axis is subhorizontal and trends southwest. Set B faults from domain II record bulk subhorizontal, southwest-trending shortening and southeast-trending extension.

Domain III: The Iniskin Bay Strand of the BBFS

North of Right Arm (Fig. 2), the Iniskin Bay strand of the BBFS crosscuts strata of the Talkeetna and Kamishak Formations (Figs. 2, 7A, and 7B; Detterman and Hartsock, 1966). The fault zone dips steeply west-northwest, strikes north-northeast, and is inferred to be a hanging-wall splay that forms part of the BBFS (Fig. 2; Detterman and Read, 1980). The fault juxtaposes marble (Kamishak Formation?), Jurassic hypabyssal felsic intrusive rocks, and the Talkeetna Formation in the hanging wall against the Talkeetna Formation in the footwall (Figs. 2, 7A, and 7B). In the hanging wall, the northwestern contact between the marble and the Talkeetna Formation is moderately deformed. The marble displays a well-defined layering that is subvertical and probably reflects original bedding (foreground of Fig. 7A). A decameter-wide fault zone that contains cataclasite defines the southeastern contact between the marble and Talkeetna Formation (Fig. 7A; example fault in Fig. 3C). At the northern limit of exposure in domain III, the fault zone truncates the northwestern contact between the marble and Talkeetna Formation, cutting out the marble and juxtaposing Talkeetna Formation against itself (Fig. 7B).

Fault Kinematic Results from Domain III

Minor fault slip data from domain III include right- and left-lateral strike slip faults as well as thrust faults with subhorizontal, east- to southeast-trending P axes (Fig. 7C). Right-lateral faults are subvertical and strike east-northeast. Left-lateral faults are also subvertical and strike northwest to northeast. Thrust faults strike northeast and dip steeply southeast. Altogether, strike-slip and thrust faults in domain III reflect southeast to east-southeast–trending subhorizontal shortening, subvertical thickening, and southwest-trending subhorizontal extension (set A, Fig. 7C). B axes from set A faults form a broad, northeast-striking girdle in the XY plane. The Z axis is subhorizontal, trends southeast, and is within error of e3 (Fig. 7C). Set A faults from domain III record triaxial deformation that is characterized by southeast-trending contraction, subvertical thickening, and subhorizontal northeast-trending extension, indicating bulk flattening of the subvertical fault zone.

A small subset of faults in domain III (set B, n = 6; Fig. 7D) were excluded from set A. Set B consists of west-northwest–striking left-lateral faults and east-northeast–striking normal faults. Left-lateral faults in set B cannot be unequivocally distinguished from set A based on the orientation of their kinematic axes; however, we classify them as set B because they are parallel to left-lateral faults from set B in domains I and II (Figs. 4D and 6B) as well as elsewhere in this study (see following). Similarly, the normal faults were grouped with set B because they have southeast-trending T axes that are compatible with set B faults elsewhere. Altogether, set B faults from domain III record south-southeast–trending subhorizontal extension and both subvertical and east-trending subhorizontal shortening.

Domain IV: The Fitz Creek Strand of the BBFS and the Tilted Hills on the Iniskin Peninsula

From Iniskin Bay to Tuxedni Bay, strata of the Chinitna and Naknek Formations dip gently to steeply southeast and form a prominent homocline referred to as the Tilted Hills (Figs. 2 and 8A; Detterman and Hartsock, 1966). On the Iniskin Peninsula, the Tilted Hills form the forelimb of the Fitz Creek anticline, which is cored by a steeply northwest-dipping fault (Fitz Creek fault; Fig. 2). Domain IV includes outcrops from the Tonnie syncline, Fitz Creek anticline, Fitz Creek fault, and the Tilted Hills on the Iniskin Peninsula (Fig. 2). Several northwest-striking subvertical faults crosscut the folds and southeast-dipping strata of the Tilted Hills and form prominent lineaments in aerial imagery (Fig. 2 and cross-faults in background of Fig. 8A; Rosenthal et al., 2017). The cross-faults are mostly right lateral, except for a few cases that also preserve left-lateral kinematic indicators, and they are oblique to the northeast strike of the BBF. We infer that the cross-faults developed during the tilting of the strata and are likely analogous to tear faults that form during the growth of an anticline (e.g., Benesh et al., 2014).

Fault Kinematic Results from Domain IV

Minor faults from domain IV include subvertical right- and left-lateral strike-slip faults, as well as east-southeast– and west-northwest–dipping thrust faults (Fig. 8B). The thrust and left-lateral faults were measured along the Fitz Creek strand where it crops out in the core of the Fitz Creek anticline (Fig. 2). They strike northeast, parallel to the map trace of Fitz Creek fault, and record south- to south-southeast–trending subhorizontal shortening that is compatible with oblique left-reverse slip on the Fitz Creek fault. Two left-lateral faults strike northwest, consistent with the expected orientation of Riedel shears (R) within the Fitz Creek fault zone (Fig. 8B). The right-lateral faults represent cross-faults in domain IV. They are subvertical, strike west-northwest, and have subhorizontal, southeast-trending P axes. Altogether, fault slip data from the Fitz Creek strand and cross-faults define a set (set A) that records triaxial deformation characterized by subhorizontal southeast-trending shortening, northeast-trending extension, and subvertical thickening. The B axes form two submaxima that together define a northeast-striking girdle in the XY plane. The Z axis is subhorizontal and trends northwest. Although e3 is not within error of Z, the α95 confidence cones are only separated by ∼2° (Fig. 8B). Either method indicates south- to southeast-trending subhorizontal shortening that is compatible with the map-scale structures in domain IV, including the folds, right-lateral cross-faults, and oblique left-reverse slip on the Fitz Creek fault.

Domain IV also includes a subset of faults that are not compatible with set A (set B; Fig. 8C). Set B faults consist of subvertical, left-lateral faults that strike northwest and right-lateral faults that strike north-northwest. Set B faults have west-southwest–trending P axes and north-northwest–trending T axes (Fig. 8C). Set B faults in domain IV preserve a slip sense that is opposite of set A for subparallel fault surfaces. For example, northwest-striking strike-slip faults are right lateral in set A and left lateral in set B. The B axes form a subvertical point maxima (Y) that is consistent with strike slip. Altogether, set B faults record subhorizontal east-northeast–trending shortening and north-northwest–trending extension.

Domain V: The Tilted Hills North of Chinitna Bay and Horn Mountain

North of the Iniskin Peninsula, the Right Arm strand of the BBFS either bends or steps to the right below Chinitna Bay (Fig. 2). Near the north shore of Chinitna Bay, the BBF strikes north-northeast, dips steeply northwest, and juxtaposes the upper member of the Talkeetna Formation in the hanging wall against the Naknek Formation in the footwall. Strata in the footwall dip gently southeast and define the Tilted Hills homocline north of Chinitna Bay (foreground of Fig. 8A). In the hanging wall, the Talkeetna Formation dips moderately south. In both the hanging wall and footwall of the BBF fault, right-lateral northwest-striking cross-faults cut the tilted strata (Fig. 2; Rosenthal et al., 2017). Domain V includes outcrops from the Tilted Hills north of Chinitna Bay and from Horn Mountain (Fig. 2).

Fault Kinematic Data from Domain V

Minor fault data from both the hanging wall and footwall of the BBFS north of Chinitna Bay are similar, so they are discussed together. Right-lateral faults from domain V represent cross-faults north of Chinitna Bay. They are subvertical, strike northwest, and have subhorizontal, south-southeast–trending P axes (Fig. 8D). Left-lateral faults strike north-northeast in an orientation that is compatible with Riedel shears (R) to the BBF (Fig. 8D). The right- and left-lateral faults form a set (set A) that records subhorizontal, south-southeast–trending shortening, and east-northeast–trending extension. B axes for a set A faults form a subvertical point maxima (Y). Altogether, set A faults are consistent with bulk oblique left-reverse slip within BBFS north of Chinitna Bay (Fig. 8D; see also domain VI).

Domain V also contains a subset of minor faults that are incompatible with set A (set B; Fig. 8E). Set B faults include north-northwest–striking right-lateral and oblique right-reverse faults, as well as west-northwest–striking left-lateral and oblique left-reverse faults. Set B faults have subhorizontal, northeast-trending P axes and moderately northwest- and southeast-plunging T axes. The B axes form a northwest-striking girdle in the XY plane. The Z axis plunges shallowly northeast and is subparallel to e3 (Fig. 8E).

Domain VI: The BBF between Chinitna Bay and the Johnson River

In domain VI, the BBF strikes north-northeast and cuts across the Talkeetna Formation near the Johnson River. The BFF juxtaposes plutonic rocks of the Talkeetna arc in the hanging wall against volcanic rocks of the Talkeetna Formation in the footwall (Fig. 2). The fault zone is well exposed in several high passes that crop out between glacial valleys. Between the Johnson River and Lateral Glacier (Figs. 2, 9A, and B9B), the fault strikes north and branches to form two thrust splay faults that dip moderately to the west. The lower splay (1, Fig. 9A) uplifts lavas and volcanic breccia of the lower Talkeetna Formation above well-bedded lavas and volcaniclastic deposits that define the upper Talkeetna Formation. In the footwall, strata of the upper Talkeetna Formation are folded into an east-verging overturned syncline (Fig. 9A). The upper splay (2, Fig. 9A) juxtaposes Jurassic quartz diorite in the hanging wall above the lower Talkeetna Formation in the footwall. The upper splay contains an ∼5-m-thick zone of cataclasite that forms a distinct red-orange weathering band, probably resulting from oxidation of iron-sulfide minerals in the fault zone (Figs. 9A, 9B). The stratigraphic separation across both splays of the fault zone indicates a component of top-to-the-east reverse motion; however, slickenlines with moderate to shallow rakes are preserved on minor fault surfaces (locations 3 and 4 in Fig. 9B) and indicate that the sense of slip was oblique left reverse. The upper fault splay was observed at a second location (5 in Fig. 9A) where several northeast-striking fault surfaces also record oblique left-reverse slip.

Near Red Glacier, ∼10 km south from the Johnson River locality (Figs. 2 and 9C), the BBF strikes north-northeast and uplifts the lower member of the Talkeetna Formation in the hanging wall above the upper member of the Talkeetna Formation in the footwall. Strata in the footwall are folded, forming an upright syncline. In the hanging wall, the lower member of the Talkeetna Formation is also folded and forms an east-vergent anticline. Fault striations preserved on minor fault surfaces in the footwall of the fault (location 6 in Fig. 9C) have shallow plunges and record oblique left-reverse slip, similar to the outcrops near the Johnson River.

Fault Kinematic Data from Domain VI

Minor faults from domain VI were measured along the BBF between Chinitna Bay and the Johnson River (Fig. 2); they are predominantly strike slip. Left-lateral faults are subvertical and form three subsets with southeast-trending P axes (Fig. 9D). North-northeast–striking left-lateral faults are subparallel to the map trace of the BBF. North-northwest–striking and northeast-striking left-lateral faults form Riedel (R) and P-shear geometries, respectively, that are consistent with left-lateral slip on the BBF. Most right-lateral faults strike northwest, are subvertical, and form R’ shears that are antithetic to the BBF. Several shallowly to moderately dipping faults also record oblique right-reverse slip. Altogether, the faults define a set that records bulk southeast-trending subhorizontal shortening and northeast-trending extension (set A, Fig. 8D). B axes from set A are steeply to moderately plunging and form a poorly defined northeast-trending girdle in the XY plane. The Z axis is subparallel to e3 (Fig. 8D) and the Y axis is subvertical. Set A faults from domain VI are compatible with triaxial deformation that is characterized by bulk oblique left-reverse slip within the BBFS.

Domain VI also preserves a subset of strike-slip and normal faults that are incompatible with set A (set B, Fig. 9E). Right-lateral faults are subvertical and strike north-northwest. One shallowly southwest-dipping right-lateral fault also occurs, but the striation is subhorizontal, indicating that it is also strike slip. Normal faults strike east to northeast and have northwest-trending T axes that are subparallel to T axes from the right-lateral faults, so the normal faults were also grouped with set B (Fig. 9E). Set B faults record subhorizontal, southwest-trending shortening and southeast-trending extension. The normal faults record a component subvertical flattening that is also compatible with the southeast-trending T axes that are characteristic of set B throughout the study area (Figs. 7D and 9E).

Domain VII: The BBFS at Open Creek Pass

North of the Johnson River, the BBFS strikes north and dips steeply west or is subvertical and juxtaposes Jurassic intrusive rocks (west) against the lower member of Talkeetna Formation (east). Near Open Creek (Fig. 2), the BBF cuts across the plutonic rocks of the Talkeetna arc; it is unclear where the fault continues near the head of Tuxedni Bay. The fault zone is well exposed at a high pass above the headwaters of Open Creek (Figs. 2 and 10A), where it is subvertical and >250 m thick, separating Jurassic quartz diorite from the lower member of the Talkeetna Formation (Fig. 10A). Within the fault zone, lenses of limestone (Kamishak Formation?; Detterman and Hartsock, 1966) are tectonically interleaved with volcanic and volcaniclastic rocks of the lower Talkeetna Formation, indicating that the fault zone cuts the contact between the two rock formations. Cataclasite and tectonic breccia (Figs. 10B, 10C) are pervasive throughout the fault zone (subvertical layering in Fig. 10A) and contain numerous subvertical, minor strike-slip faults (e.g., Figs. 10B, 10C).

Fault Kinematic Data from Domain VII

Minor faults in domain VII are predominantly strike slip. Left-lateral faults form three subsets with southeast-trending P axes. North-northeast–striking left-lateral faults are parallel to the map trace of the BBF at this location. North-northwest–striking and northeast-striking left-lateral faults form Reidel (R) and P-shear geometries, respectively, that are also compatible with left-lateral slip on the BBF (Fig. 10D). Right-lateral faults are subvertical, strike west to northwest, and are consistent with the expected orientation of R’ shears (Fig. 10D). The right- and left-lateral faults define a set (set A; Fig. 10D) that records subhorizontal southeast-trending shortening and northeast-trending extension. B axes from set A form a subvertical point maxima (Y) that is compatible with strike slip. Altogether, the kinematic axes from domain VII record bulk left-lateral shear within the BBFS.

A small subset (n = 5) of right-lateral faults that strike north-northwest also occurs in domain VII and is incompatible with set A (set B, Fig. 10E). The faults record subhorizontal northeast-trending shortening and southeast-trending extension. They are parallel to right-lateral set B faults that were observed throughout the study area.

BBF at Contact Point

A large coastal outcrop of the BBF is located at the southwestern end of Contact Point, ∼37 km southwest of Ursus Head (Fig. 11; location CP in Fig. 1). Here the BBF dips ∼40° northwest and juxtaposes the Talkeetna Formation in the hanging wall above the Naknek Formation in the footwall. The hanging-wall strata are pervasively deformed and dip moderately northwest. In the footwall, the Naknek Formation is folded and forms an overturned northeast-trending syncline that verges southeast, indicating a reverse sense of slip for the fault that is similar to the structural relationships exposed at Ursus Head (Figs. 4 and 11).

Two felsic dikes intrude the fault plane (Fig. 11). The lower dike is ∼3–5 m thick, light gray, and porphyritic with plagioclase and hornblende phenocrysts as large as 4 mm. The upper dike is ∼6–7 m thick, pinkish-tan, and porphyritic with plagioclase (<4 mm) and biotite phenocrysts (<2.5 mm). The contact between the two dikes is defined by an ∼1.2-m-thick fault zone that contains cataclasite. The cataclasite displays a weak foliation that is parallel to the BBF plane. Although the contact between the dikes is faulted, the upper dike preserves emplacement cooling joints with 15–20 cm spacing (Fig. 11B) and only the upper ∼20 cm of the lower dike is deformed. The upper dike appears to be offset by a steeply dipping splay fault in the hanging wall of the BBF (black star in Fig. 11A). The field relationships indicate that the dikes were emplaced after most of the deformation in the fault zone had occurred because they are only moderately deformed with respect to the surrounding cataclastic host rock. Moreover, the lower dike contains small sandstone xenoliths from the Naknek Formation, signifying that the dike was likely emplaced after the fault juxtaposed the Talkeetna Formation above the Naknek Formation. A sample of the upper dike was collected for 40Ar/39Ar thermochronology to help constrain the timing of deformation within the BBFS (sample 12BG109A, Fig. 11B; see next section). Fault kinematic data were not collected at Contact Point due to limited time visiting this location.

40Ar/39Ar Thermochronology

We collected two samples for 40Ar/39Ar thermochronologic analyses at the University of Alaska Fairbanks. Sample 14JR25 is from an ∼10-cm-wide gently deformed basalt dikelet that intrudes cataclasite within the Bruin Bay fault zone at Ursus Head (Fig. 5D). Sample 12BG109A is from a silicic dike that intrudes the Bruin Bay fault zone near Contact Point and is moderately deformed (Fig. 11). The samples were irradiated at McMaster University (Hamilton, Ontario, Canada). Step-heating followed the techniques in Benowitz et al. (2014, and references therein). System error determination followed McDougall and Harrison (1999) (see summary of 40Ar/39Ar analysis in the Supplemental Files1).

Basalt Dike from Ursus Head, 14JR25

We determined a whole-rock 40Ar/39Ar age of the Ursus Head basalt dike (sample 14JR25). The basalt was crushed, sieved, washed, and hand-picked for phenocryst-free rock chips. A homogeneous whole-rock separate from the basalt dikelet (Fig. 12A, sample 14JR25) was analyzed. The integrated age (33.6 ± 0.9 Ma) and the plateau age (33.1 ± 0.8 Ma) and the isochron age (33.3 ± 0.8) are all within error of each other. We prefer the plateau age of 33.1 ± 0.8 Ma for sample 14JR25 because of the large error on the isochron-derived initial 40Ar/36Ar related to the generally homogeneous radiogenic content of the release (Fig. 12A; Supplemental Files [see footnote 1]).

Silicic Dike from Contact Point, 12BG109A

We determined biotite and plagioclase 40Ar/39Ar ages from the silicic dike at Contact Point (Figs. 11B, 12B, and 12C; sample 12BG109A). The dike was crushed, sieved, washed, and hand-picked for 99% pure biotite and plagioclase separates. A homogeneous biotite separate (Fig. 12B) was analyzed. The integrated age (35.5 ± 0.9 Ma) and the plateau age (36.5 ± 0.8 Ma) and the isochron age (36.3 ± 1.1) are all within error of each other. We prefer the plateau age of 36.5 ± 0.8 Ma for sample 12BG109A because of the large error on the isochron-derived initial 40Ar/36Ar related to the generally homogeneous radiogenic content of the release (Supplemental Files [see footnote 1]).

A homogeneous plagioclase separate from this sample (Fig. 12C, sample 12BG109A) was also analyzed. The integrated age (31.6 ± 0.8 Ma), the plateau age (31.5 ± 0.7 Ma), and the isochron age (30.8 ± 0.7) are all within error of each other. We prefer the isochron age of 30.8 ± 0.7 Ma for sample 12BG109A because of evidence of excess 40Ar based on the isochron-derived initial 40Ar/36Ar of 300.6 ± 5.4 (Supplemental Files [see footnote 1]; Cassata et al., 2009; Gillis et al., 2017). Based on the difference in nominal closure temperatures for Ar loss between the biotite (350–400 °C; Reiners et al., 2005) and plagioclase (225–300 °C; Cassata et al., 2009) minerals, we calculate a minimum rock cooling rate of 9 °C/m.y. for this sample between ca. 36.5 Ma and ca. 30.8 Ma (Fig. 12D).

The determined rock cooling rate of 9 °C/m.y. between ca. 36.5 Ma and ca. 30.8 Ma for sample 12BG109A is likely not related to postemplacement thermal relaxation or reheating due to fault zone fluid flow because of the relatively slow and prolonged period (∼5.7 m.y.) of rock cooling. Without further regional thermochronologic constraints we cannot determine if this rock cooling event is related to a regional exhumation event or slip along the BBFS. Given the lack of documented auxiliary structures that could be responsible for this exhumation event, we prefer an interpretation of exhumation related to slip along the BBFS. This is consistent with dike emplacement ∼30 km to the northeast at Ursus Head. Overall, our constraints on magmatic and exhumation events along the BBFS indicate this fault was active to at least ca. 31 Ma with the possibility of a younger slip history.

DISCUSSION

In the following we synthesize the fault kinematic results, field observations, and thermochronology presented in this paper. We discuss several possible tectonic interpretations for the BBFS.

Synthesis of Fault Kinematic Results

To further investigate the bulk kinematics of the BBFS throughout the study area, we present an analysis of the fault slip data from all of the structural domains (cf. Betka and Gillis, 2014). The fault slip results are remarkably similar throughout the study area such that two kinematically distinct sets (A and B) are present in each domain. Population A (n = 233) includes 79% of the data set, forming the dominant population of minor faults within BBFS. Population B (n = 63) includes 21% of the data set. Figure 13 shows the combined fault kinematic results for populations A and B from all of the field sites in this study.

Population A consists of northeast-striking reverse faults, north-northwest– to northeast-striking left-lateral faults, and northwest- to southwest-striking right-lateral faults (Fig. 13A). Reverse faults dip steeply to moderately northwest or southeast, and striations rake between 90° and 45°. The strikes of the reverse faults are either parallel or slightly oblique (clockwise) to the map trace of the BBF, which is indicative of bulk oblique, left-reverse motion within the BBFS. The left-lateral faults are subvertical, have subhorizontal to moderately plunging striations, and form three subsets. North-northeast–striking left-lateral faults are parallel to the map trace of the BBF. We interpret the north-northwest–striking and northeast-striking left-lateral faults as Riedel (R) and P shears, respectively, that record left-lateral motion within the BBFS. Right-lateral faults mostly strike northwest and are interpreted as R’ shears that are also compatible with a left-lateral BBFS. The subset of right-lateral faults that strikes northeast are from domain III; thus, their orientation reflects the more northern strike of the Iniskin Bay strand of the BBFS fault in that location (Figs. 2 and 7).

The P and T axes from population A faults yield a subhorizontal southeast-trending principal shortening axis (e3 = 147, 06). The B axes from population A form a northeast-striking girdle in the XY plane. The bulk shortening axis is subhorizontal, trends southeast (Z = 138, 06 ± 12), and is within error of e3 (Fig. 13A). Altogether the kinematic results from population A indicate horizontal, southeast-trending shortening (∼143°), vertical thickening, and horizontal northeast-trending extension that is compatible with bulk oblique left-reverse slip within the BBFS. In the southern part of the study area (domains I and II) the BBF strikes northeast and accommodated mostly reverse slip. In contrast, in the northern part of the study area the BBFS strikes north-northeast to north and accommodated mostly left-lateral motion (domains VI and VII), indicating that slip was partitioned along differently oriented segments of the fault with respect to the bulk southeast-trending principal shortening axis of population A.

Population B consists of subvertical left-lateral faults that strike northwest, and right-lateral faults that strike north-northwest. Striations on population B faults are generally shallowly plunging. Oblique-normal faults (n = 6) from domains III and VI are also included in population B because they record southeast-trending subhorizontal extension; they are grouped with the respective left- or right-lateral strike slip faults in Figure 13B. P and T axes from population B yield a subhorizontal southwest-trending principal shortening axis (e3 = 227, 08), the intermediate strain axis is subvertical (e2), and the extension axis trends northwest (e1 = 317, 01). B axes form a subvertical point maxima (Y = 315, 80 ± 9; Fig. 13B). The strikes of left- and right-lateral subsets of population B appear to form a conjugate set that might be used to infer a southeast-trending principal shortening axis if only the B axes are considered; however, the sense of slip for each subset and the orientations of the P and T axes are not consistent with conjugate faults, underscoring the importance of considering all of the kinematic axes and ranking the quality of each fault measurement.

Instead, we deduce that population B faults either formed new faults and/or reactivated existing fault surfaces from population A. This reasoning is based on structural relationships from domains I–III. In domain I (Fig. 4) north-northeast–striking subvertical faults are parallel to right-lateral faults of population B and cut the BBF with an apparent sense of offset that is compatible with right-lateral slip. Thus, we infer that the subvertical faults are part of population B and that population B is younger than population A. Similarly, in domain II right-lateral faults of population B crosscut tilted strata of the Talkeetna Formation in the hanging wall of the BBF (e.g., Fig. 3A; location shown in Fig. 2). In domain III, left-lateral minor faults that crop out near Iniskin and Chinitna Bays are parallel to the northwest-striking right-lateral cross-faults from population A. The left-lateral faults of population B commonly occur adjacent to the cross-faults; however, explicit crosscutting relationships between population A and B faults were not observed in domain III. Population B records strike-slip faulting that is compatible with horizontal, southeast-trending extension (∼137°) within the BBFS and is likely younger than the deformation that formed population A. Although normal faults in population B record a minor component of vertical thinning, they are few (n = 6), and the B axes form a subvertical point maxima that indicates dominantly strike-slip, plane-strain deformation.

Summary of Age Constraints for Deformation within the BBFS

In the study area, Jurassic–late Paleogene(?) strata are folded in orientations that are sympathetic with southeast-trending tectonic shortening and minor faults of population A (cf. Rosenthal et al., 2017; Betka and Gillis, 2014). New mapping in Gillis (2016) indicates that bedding in Maastrichtian–Cenozoic deposits (i.e., the Saddle Mountain section and late Paleogene? strata) between Chinitna and Tuxedni Bays is concordant with the underlying Jurassic strata despite the unconformities that separate them. Because the folds in the study area trend northeast and are kinematically compatible with population A, we infer that the tectonic event that formed population A faults occurred during or after the deposition of the late Paleogene(?) unit (cf. Rosenthal et al., 2017).

Supporting this inference, at Ursus Head a gently deformed 33.3 ± 0.8 Ma basalt dikelet (sample 14JR25; Fig. 12A) intrudes cataclasite within the Bruin Bay fault zone that contains minor fault surfaces from population A (domain I, Figs. 4B, 4C, and 5D). Based on the observation that the dikelet (sample 14JR25) is less deformed than its cataclasite host, we infer that the majority of reverse slip within the BBFS at Ursus Head occurred prior to the Oligocene but continued until at least ca. 33 Ma. This is consistent with the 30.8 ± 0.7 Ma plagioclase age and 36.5 ± 0.8 Ma biotite age from a silicic dike (Figs. 11, 12B, and 12C; sample 12BG109A) that intrudes the BBF at Contact Point. The calculated minimum rock cooling age of ∼9 °C/m.y. from sample 12BG109A at Contact Point indicates a minimum hanging-wall uplift rate of ∼0.2–0.4 mm/yr between ca. 37 and 31 Ma assuming a 25–50 °C/km range of geothermal gradients and steady-state erosion (Fig. 12D). The dikes are moderately deformed in both localities, indicating that some deformation persisted after ca. 31 Ma. On the basis of crosscutting relations and cooling ages of the syntectonic dikes, as well as the ages of the folded strata in the study area, we conclude that the majority of slip (population A) within the Iniskin–Tuxedni–Contact Point segment of the BBFS occurred in the early-middle Paleogene, but continued until at least ca. 31 Ma. Field relationships from domains I and II indicate that population B faults are likely younger than population A faults (discussed herein), and thus population B probably represents late Paleogene to Neogene deformation.

Structural Evolution of the Iniskin-Tuxedni Region and CIB

Fault kinematic results and map patterns presented in this study indicate that the map-scale folds in the Iniskin-Tuxedni region are part of the BBFS (cf. Rosenthal et al., 2017). Northeast-trending map-scale folds on the Iniskin Peninsula (i.e., Fitz-Creek and Mt. Eleanor anticlines, Tonnie syncline; Fig. 2) accommodate southeast-trending shortening that is compatible with population A faults. The amplitude of each fold dissipates along strike to the northeast and southwest while the apparent reverse displacement along the BBF increases in each direction. For example, the apparent stratigraphic separation along the Right Arm strand is small near the head of Chinitna Bay where the BBF juxtaposes the upper Talkeetna Formation against the lower member of the Tuxedni Group and may have zero vertical displacement (Fig. 2; see cross sections). In contrast, southwest of the Iniskin Peninsula at Ursus Head (domain I), the Right Arm strand of the BBFS separates the Kamishak Formation in the hanging wall from the Naknek Formation in the footwall (Figs. 2 and 4). Extrapolating the stratigraphic thicknesses from the Iniskin Peninsula requires ∼5 km of throw at Ursus Head, indicating that throw along the Right Arm strand of the BBF increases southeastward from Chinitna Bay to Ursus Head.

North of Chinitna Bay, the BBF fault juxtaposes the upper member of the Talkeetna Formation in the hanging wall against the lower member of the Naknek Formation in the footwall (Wartes et al., 2016). Here there is a maximum of ∼3 km of throw because some of the apparent stratigraphic separation resulted from left-lateral displacement (i.e., domain VI, set A; Fig. 9D). Along strike to the southeast, the Fitz Creek strand on the Iniskin Peninsula accommodates small vertical displacement (<100 m; Fig. 2). The map trace of the Fitz Creek fault suggests that it is continuous with the main segment of the BBF north of Chinitna Bay, and thus the fault increases in throw northward across Chinitna Bay (Fig. 2; Wartes et al., 2016). Altogether, the BBFS between Ursus Head and Red Glacier forms an en echelon thrust transfer zone if shortening on the BBF was transferred into the folds on the Iniskin Peninsula where the Fitz Creek and Right Arm strands overlap (i.e., Dahlstrom, 1970, figure 26 therein; Wartes et al., 2016; contractional step-over of Christie-Blick and Biddle, 1985).

Alternatively, it is also plausible that the Right Arm strand of the BBFS jogs to the right below Chinitna Bay to connect with the main strand of the BBFS north of Chinitna Bay (Fig. 2; Detterman and Hartsock, 1966; Wartes et al., 2016). In this case, the oblique-reverse segment of the BBF that uplifts the Talkeetna Formation may cut a buried detachment beneath the Tonnie syncline, Fitz Creek anticline, and Fitz Creek fault. The latter interpretation is consistent with geologic relations at Ursus Head, where a bedding parallel detachment horizon in the Kamishak Formation is truncated by a high-angle oblique-reverse fault (Figs. 4B and 5B). For the purpose of this paper, we favor neither the en echelon transfer zone model nor thin- to thick-skinned faulting model over the other; kinematic modeling of the faults and folds could be undertaken to distinguish between the two hypotheses. Nevertheless, we conclude that folding of Mesozoic to late Paleogene(?) strata occurred during southeast-directed contraction of the BBFS that is compatible with population A faults.

Reflection seismic interpretations and industry well data from the CIB southeast of the study area indicate that Jurassic–Miocene strata throughout the CIB are deformed by northeast-trending structures (Haeussler and Saltus, 2011; Fisher et al., 2013). The northeast-trending folds and faults exposed within the BBFS are subparallel to the fault-cored anticlines throughout the lower CIB (Wilson et al., 2012; LePain et al., 2013; Fig. 1). Thus, we infer that the deformation along the BBFS occurred during regional Paleogene to Neogene contraction of the CIB and surrounding regions. Strike-slip faulting that is compatible with southeast-trending extension of the BBFS (population B) occurred later and probably reflects the Neogene tectonic setting of the southern Alaskan margin (discussed in the following).

Proposed Tectonic Models for the BBFS

There are several proposed tectonic events that formed the southern Alaskan margin during the Mesozoic and Cenozoic, including the Mesozoic accretion of Wrangellia (e.g., Plafker, et al., 1994; Trop and Ridgway, 2007); Paleocene–Eocene subduction and migration of a spreading ridge (Bradley et al., 2000; Haeussler et al., 2003a, 2003b); the Late Cretaceous–Eocene accretion and dextral translation of the Chugach–Prince William terrane (Garver and Davidson, 2105; Day et al., 2016); and the possible Paleogene development of the Alaskan orocline (Stone, 1980; Hillhouse and Coe, 1994; Glen, 2004). We showed here that the majority of the deformation within the BBFS occurred during the early-middle Paleogene. In the following, we discuss the BBFS in the context of several competing hypotheses that are pertinent to the tectonic development of south-central Alaska. We emphasize that the following discussion is intended to provoke future research; we cannot confidently settle on one particular interpretation without additional understanding of the Cenozoic tectonic evolution of southern Alaska.

Late Jurassic–Early Cretaceous Terrane Accretion Hypothesis

Prior studies suggest that reverse slip within the BBFS helped drive Late Jurassic uplift and exhumation of the plutonic roots of the Talkeetna arc in the hanging wall of the fault (Fig. 14A; Trop et al., 2005; Wartes et al., 2013) during the assembly and accretion of the Wrangellia composite terrane (Fig. 14A; Clift et al., 2005a). Only population A includes reverse faults within the BBFS that could uplift the Talkeetna arc. Although we cannot rule out the possibility that population A faults record both Late Jurassic and Paleogene deformation with similar fault slip kinematics, the results of this study indicate that the majority of the deformation within the BBFS is Paleogene (cf. Rosenthal et al., 2017). Moreover, it is possible that the Late Jurassic uplift and exhumation of the Talkeetna arc was accomplished by other processes, such as thermal and mechanical effects associated with the emplacement of a large-volume batholith (Kimbrough et al., 2001), or the transition from subduction to erosion to accretion during the collision of the Talkeetna arc with Wrangellia (Clift et al., 2005a). Thus, we deduce that the rapid denudation of the Talkeetna arc and deposition of the Naknek Formation in the Jurassic did not require reverse slip within the BBFS, although we cannot unequivocally rule out the possibility.

Spreading Ridge Subduction Hypothesis

It is widely hypothesized that early Paleogene subduction of a spreading ridge–transform system affected the entire forearc region of southern Alaska (e.g., Bradley et al., 2000; Haeussler et al., 2003a, 2003b). The spreading ridge migrated from west to east along the Aleutian megathrust, forming a trench-ridge-transform triple junction (Fig. 14B; Bradley et al., 2000, 2003; Sisson et al., 2003; Madsen et al., 2006). Subduction of the buoyant ridge caused both uplift of the CIB to a nonmarine environment (Fig. 14B; Trop et al., 2003; Trop and Ridgway, 2007; Finzel et al., 2011b, 2016) and polyphase deformation of Chugach terrane (Haeussler et al., 2003a; Pavlis and Sisson, 2003b). In Benowitz et al. (2012), rapid Paleocene–Eocene rock cooling and exhumation of the Tordrillo Mountains, north of the study area, was reported, and the cooling was attributed to thermal effects of the subducting slab window. Based on similarities in age of Paleogene folding associated with the BBFS in the Iniskin-Tuxedni region as well as the subparallel orientation of numerous northeast-trending folds throughout the lower CIB, we propose that deformation associated with the spreading ridge subduction persisted landward of the accretionary complex and affected the entire forearc region, including the BBFS.

Conceptual models of forearc deformation during spreading ridge subduction (Figs. 14B and 15A; Kusky et al., 1997; Haeussler et al., 2003a) suggest that upper plate faults accommodate transpression where the leading plate and buoyant ridge axis converge with the forearc and move obliquely along it. When the spreading ridge and trailing plate passes below, the forearc undergoes transtension with a sense of shear that is opposite of that imposed by the leading plate (Fig. 15A). These constructs are similar to the results of three-dimensional finite element models that predict that the lateral migration of a ridge axis below a forearc temporally changes the state of stress in the upper plate when it is susceptible to shear tractions from the trailing plate (Fig. 15A; Zeumann and Hampel, 2015; cf. Russo et al., 2010a, 2010b). The effect of ridge subduction on a cordilleran forearc may be gleaned from a modern example in the Andes. At the Chile Rise trench-ridge-trench triple junction, topographic, thermal, and tectonic effects of several episodes of Cenozoic ridge subduction occur for hundreds of kilometers inboard from the trench (Lagabrielle et al., 2004; Ramos, 2005; Russo et al., 2010a, 2010b). Similarly, the slip kinematics of faults in the upper plate changed during the lateral migration of the Chile Rise (Lagabrielle et al., 2004), suggesting that fault reactivation during spreading ridge subduction and migration may be characteristic of the long-term deformation of forearcs.

We propose a new hypothesis in which the heterogeneous slip kinematics of the BBFS are considered with respect to the early Paleogene ridge subduction event. Population A faults within the BBFS might have formed during subduction and eastward migration of the leading plate and buoyant ridge axis. Population B faults may record the reactivation of the fault system under southwest-trending compression driven by the spreading direction of the trailing plate (Figs. 14B and 15). Although the majority of deformation in the BBFS occurred during the early-middle Paleogene, we acknowledge that the exact timing of deformation within the BBFS and location of the forearc (see below) must be better constrained to confidently link the ridge subduction event to deformation within the CIB and BBFS.

Late Cretaceous–Cenozoic Terrane Accretion and Oroclinal Bending Hypothesis

In contrast with the ridge-subduction hypothesis, recent provenance studies of Maastrichtian–Paleocene flysch from the Chugach–Prince William terrane (part of the accretionary prism; Figs. 1 and 14B) indicate that the rocks originated mainly from the Coast Plutonic Complex of British Columbia (Amato et al., 2013; Garver and Davidson, 2015; Day et al., 2016) and were translated a long distance (as much as thousands of kilometers) to the southern Alaska margin. Arrival of the Chugach–Prince William terrane to south-central Alaska during the Paleogene would likely result in southeast-trending compression across the CIB due to either the initially curved margin of southern Alaska or postulated coeval counterclockwise rotation and development of the Alaskan orocline (e.g., Hillhouse and Coe, 1994; Glen, 2004), or both (Figs. 14B and 15B). We suggest that population A faults in the BBFS record deformation caused by the accretion of the Chugach–Prince William terrane to the southern Alaskan margin after the plate margin had reached its present northeast-trending orientation sometime in the late Paleogene (Figs. 14B and 15B). Population B faults may reflect Neogene deformation (see following). However, the Paleogene tectonic setting of southern Alaska and geometry of the plate margin remain widely debated.

Late Paleogene to Present Deformation of Southern Alaska

The onset of the northward translation and consequent subduction collision of the Yakutat terrane began in the late Paleogene (Bruns, 1983; Plafker et al., 1994) and caused Oligocene crustal shortening and inversion of Mesozoic sedimentary basins in south-central Alaska (Trop and Ridgway, 2007; Finzel et al., 2011b, 2016). The arrival and flat-slab subduction of the Yakutat plate may have contributed to post-Eocene contraction in the BBFS that is compatible with population A. This hypothesis is consistent with the observation that some slip on the BBF occurred after it was intruded by Oligocene dikes; however, most of the deformation within the BBFS occurred prior to the emplacement of the dikes. Thus, we conclude that the Yakutat subduction collision may have reactivated the BBFS but was probably not a major contributor to the deformation within it. Instead, we postulate that population B faults accommodated southeast-trending horizontal extension during late Cenozoic subsidence of the CIB (Fig. 15C; cf. Haeussler and Saltus, 2011; Finzel et al., 2011b). This hypothesis is similar to previously proposed interpretations that the northeast-striking Border Ranges fault system along the southeast margin of the CIB (Fig. 1) was reactivated by Neogene normal faulting (Pavlis and Bruhn, 1983; Haeussler and Saltus, 2011). We acknowledge that the data are limited and further investigation is needed to document and understand Neogene subsidence of the CIB.

Geodetic data indicate that south-central Alaska is currently rotating counterclockwise away from the Yakutat microplate and independently from the Bering plate of western Alaska (Cross and Freymueller, 2008; Finzel et al., 2015). The modeled continuous strain rate field from global positioning system data indicate that the Cook Inlet region is undergoing both southeast- and northeast-trending horizontal contraction between the Yakutat and Bering Sea blocks (Fig. 1; Finzel et al., 2011a, 2014, 2015). Population B faults are partly compatible with the northeast-trending shortening axes predicted by the strain-rate model for the CIB region (Finzel et al., 2015). Thus, we also suggest that population B may be the result of far-field compressional stress from the subduction collision of the Yakutat terrane (Figs. 14C and 15C). However, the southeast-trending extension axes from population B are contrary to the predictions of the geodetic model, so we consider this interpretation less favorably than the hypotheses discussed above. Moreover, crustal stress maps of the Cook Inlet region from earthquake focal mechanism inversion indicate that the lower Cook Inlet region is currently under subhorizontal southeast-trending compression while the upper Cook Inlet records east-trending compression (Ruppert, 2008), implying that the effects of Yakutat flat-slab subduction in the forearc region may be limited to the upper Cook Inlet (Bruhn and Haeussler, 2006). This is consistent with new apatite (U-Th)/He cooling ages from the Kenai Peninsula (Fig. 1) that show an abrupt transition between cooling ages from rocks overlying and to the south of the flat-slab segment. Apatite (U-Th)/He ages record ca. 10–2 Ma ages above the flat slab segment, whereas rocks from the southern part of the Kenai Peninsula do not record apatite-helium ages younger than ca. 10 Ma (Valentino et al., 2016). Therefore, ancient low-angle subduction of the Pacific plate (∼42° modern dip; Syracuse and Abers, 2006) may be responsible for our newly documented slip activity along the BBFS during the Oligocene–Neogene.

CONCLUSIONS

The BBFS of southern Alaska has a polyphase kinematic history but predominantly records Paleogene sinistral transpression. The fault system consists of two kinematically distinct populations of minor faults. Northeast-striking left-lateral and reverse faults and northwest-striking right-lateral faults define one population (A, n = 233, 79% of data set). Population A records southeast-trending subhorizontal shortening and bulk oblique left-reverse deformation within the BBFS. North-northwest–striking right-lateral and west-northwest–striking left-lateral faults define a second population that records subhorizontal northeast-trending shortening and southeast-trending extension (B, n = 63; 21% of data set). At one locality, population B faults crosscut the BBF; therefore, population B is inferred to be younger than population A.

Folds in the study area tilt Mesozoic–late Paleogene(?) strata and are compatible with southeast-trending contraction and population A within the BBFS. At two exposures of the BBF, gently deformed dikes intrude cataclasite within the fault zone and yield ca. 37–31 Ma 40Ar/39Ar ages. The dike ages and field relationships indicate that most of the deformation within the BBFS (population A) occurred during the early-middle Paleogene, but some deformation persisted after the early Oligocene emplacement of the dikes.

This study places the BBFS in the context of the widely debated Paleogene tectonic setting of south-central Alaska. The BBFS dominantly records Paleogene southeast-trending contraction that is compatible with northeast-trending folds of Mesozoic to Cenozoic strata throughout the CIB. We postulate that the main phase of sinistral transpression within the BBFS (population A) reflects either (1) Paleogene subduction and migration of a spreading ridge along the southern Alaskan margin, or (2) Paleogene accretion of the Chugach–Prince William terrane to southern Alaska. Other tectonic events possibly recorded by the BBFS include deformation during the Late Jurassic accretion of the Wrangellia composite terrane, Paleogene bending of the Alaskan orocline, the onset of flat-slab subduction of the Yakutat terrane, and Oligocene–present subduction of the Pacific plate.

ACKNOWLEDGMENTS

This project was supported by the U.S. Geological Survey National Cooperative Geologic Mapping Program under STATEMAP awards G13AC00157 and G15AC00199 and by the State of Alaska. We thank CIRI (Cook Inlet Region Inc.), the Seldovia Native Association, Tyonek Native Corporation, Ninilchik Native Association, Salamatof Native Association, Knik Tribal Council, and Chickaloon Village as well as Lake Clark National Park for land access permits. Bear Mountain Lodge and Snug Harbor Cannery provided excellent accommodations and logistical support during the field season. We thank our helicopter pilots, Roger Hinsdale and Merlin (‘Spanky’) Handley, from Pathfinder Aviation for safe transportation during the field seasons. Jacob Rosenthal, Rebecca Tsigonis, and Paul Wilcox were excellent field assistants. We thank Marwan Wartes, Dave LePain, Trystan Herriott, Kate Bull, Paul Decker, Ken Helmold, and Rick Stanely for productive discussions in the field that inspired this study. We thank Terry Pavlis and Richard Lease for very insightful reviews that improved the presentation of the data in this manuscript.

1Supplemental Files. Summary of 40Ar/39Ar analysis, 40Ar/39Ar step heating results (Table A1), and laser step heating data (Table A2). Please visit http://doi.org/10.1130/GES01464.S1 or the full-text article on www.gsapubs.org to view the Supplemental Files.

Supplementary data