Regionally persistent vein sets cut Early Jurassic through late Paleogene(?) strata throughout a study area >2000 km2 in the lower Cook Inlet forearc basin of Alaska. Using field, aerial, and GIS–based studies, we document vein orientations, group them into four dominant sets, and present relative timing observations to demonstrate their development during regional faulting and folding of Cook Inlet basin strata. All veins were restored about regional folds by first removing the bedding dip, and then rotating the bedding strike into parallelism with the regional structural trend (038°). The most dominant vein set strikes ~310°, orthogonal to the regional structural trend, and is present in all strata in the field area. The other sets strike 210°, 360°, and 260°. Fold-test results show that variations in the vein set orientations throughout the study area are correlated with the changes in bedding attitudes that define regional folds, indicating that the veins formed progressively with the folds. We document abutting and crosscutting relations between sets, and present a new ca. 52 Ma 40Ar/39Ar age of a dike that parallels the dominant set (310°) and is crosscut by others. Based on field relations, relative timing constraints, and the fold-test results, we suggest a sequence of vein development and its relationship to fold growth within the Bruin Bay fault system during Paleogene deformation along the southern margin of Alaska. Our results may serve as a case study for linking vein development to tectonic events in other ancient and modern forearc basins.


Sediment filling forearc basins provides a record of subsidence and exhumation driven by convergent margin tectonics. The tectonic processes during forearc basin evolution can be interpreted by examining the style and chronology of preserved structures that deform such basins. This is particularly true for long-lived continental subduction margins, which can undergo deformational events such as accretion, subduction erosion, and/or subduction of spreading ridges. Structural studies in forearc settings often focus on major fault and fold systems that deform the basins and their margins (e.g., Dickinson and Seely, 1979; Haeussler et al., 2000; Trop et al., 2005). Less well studied at these sites is the evolution of fracture systems that are associated with development of local and regional structures and that record changes in the regional state of stress with time. Note that fractures in this study are either joints (open fractures) or veins (filled fractures). Henceforth we distinguish fractures as either joints or veins, as the cohesion of a crack controls the mechanical behavior and permeability of the bedrock in which it occurs.

Southern Alaska comprises a series of accreted terranes (Fig. 1). In the Cook Inlet forearc basin (Figs. 1 and 2) large-scale structures deforming the basin principally occur offshore, and few publicly available subsurface data sets exist for their comprehensive structural analysis. Exposed for at least 100 km in the Iniskin-Tuxedni region of the Cook Inlet is a systematic regional network of fractures that are well expressed approximately at the magmatic arc-forearc basin boundary that is partly defined by the Bruin Bay fault system (BBFS; Figs. 13). The fractures cut forearc strata of Jurassic through late Paleogene(?) (possibly into Neogene) age, and therefore may record multiple episodes of forearc deformation driven by several documented tectonic events along the margin.

Fractures have long been known to be associated with folds and faults, their orientations recording the sequential strain history of a region during evolving deformation (e.g., Price, 1966; Reches, 1976; Bergbauer and Pollard, 2004; Ahmadhadi et al., 2008; Pastor-Galán et al., 2011; Lacombe et al., 2011; Weil and Yonkee, 2012). Such studies tend to focus on inferring sequential stress regimes for individual fracture sets based on their orientations and relative ages determined through stratigraphic controls and crosscutting and/or abutting relationships (e.g., Engelder and Geiser, 1980; Hancock, 1985; Engelder, 1987). Pervasive mesoscopic structures (cleavage, joint sets, paleomagnetic fabrics) have been quantitatively related to progressive stages of deformation, namely in the development of folds and faults (Fischer and Wilkerson, 2000; Yonkee and Weil, 2010a, 2010b; Weil and Yonkee, 2012; Li et al., 2013). Thus, mesostructures (~0.1–10 m) can now be seen as evolving alongside regional structures during progressive tectonic activity.

Within the Cook Inlet basin (CIB) study area, Detterman and Hartsock (1966) reported two regional joint sets, one striking 305° and the other striking 225°, roughly orthogonal to each other, and speculated that the joints may be associated with the local northeast-trending folds. More complex joint and vein networks were later identified on the Insikin Peninsula by Gillis et al. (2013) and Rosenthal et al. (2015), who noted that quartz and calcite cement is preserved on the joint faces and that they are parallel to pervasive vein sets that are also cemented with quartz and calcite, as well as clay; they all suggested that the joints and veins may be important with respect to CIB reservoir potential. Oil seeps along regional joint sets have long been noted (Detterman and Hartsock, 1966). Helmold (2013) and LePain et al. (2013) suggested that fracture networks in the Jurassic strata of the CIB might have contributed to the migration of hydrocarbons from their Jurassic source rocks to Cenozoic sandstone reservoirs, and that the Jurassic strata may be a productive fractured reservoir (see also Magoon and Claypool, 1981; Fisher and Magoon, 1982).

In this study we document orientations, distributions, and relative ages of several regionally prominent joint and vein sets (Fig. 4) through field and remote sensing-based observations. We present the first detailed investigation of the pervasive joint and vein networks that formed within Jurassic to late Paleogene(?) strata of the lower CIB, and interpret the results with respect to the growth of the local folds and the tectonic history of the BBFS and the CIB. Field observations focus on outcrop-scale vein and joint patterns at targeted sites in the Iliamna and Iniskin regions, an area spanning from southwestern Cook Inlet near Augustine Volcano, north to Tuxedni Bay (Figs. 1 and 2). Remote-sensing observations expand the study to cover joint patterns over the full Iniskin and Iliamna regions outlined in Figure 3. While the absolute timing of joint and vein development is challenging to constrain, we propose that most veins developed during Paleogene reorganization of subduction dynamics that resulted in regional folding and faulting and local rotation of the veins during progressive deformation and fold growth (e.g., Haeussler et al., 2003a). Joints developed along the existing vein sets, probably reflecting exhumation and unloading of the study area. We present a conceptual model in which veins formed in conjunction with fold and fault development within southern Alaska’s diverse Cenozoic tectonic framework.

Geologic Setting

The CIB is a northeast-trending forearc basin in southern Alaska (Fig. 1). Along its northwestern side, it is bounded by the Jurassic Talkeetna arc and modern Aleutian Arc, and to the southeast are the Aleutian Trench and Mesozoic accretionary prism (the Chugach terrane; Fig. 1). The system continues to develop above the Aleutian Trench and northwestward-subducting Pacific plate. The forearc elements are a product of continuous subduction since latest Triassic time and have been modified by several regional tectonic events during that period. These elements include (1) two apparent periods of subduction erosion, one in the Early to Middle Jurassic and the other in the mid-Cretaceous (Amato et al., 2013); (2) three episodes of terrane accretion (see Fig. 1 for modern-day locations) since Jurassic time: Wrangellia (Plafker et al., 1989; Trop et al., 2005; Trop and Ridgway, 2007), Chugach (Bol and Roeske, 1993), and Yakutat (Plafker et al., 1978, 1994); (3) subduction of a spreading center in late Paleocene to early Eocene time (e.g., Bradley et al., 2000), and possibly an earlier one in the Early Cretaceous (Pavlis et al., 1988); and (4) postulated oroclinal bending of the upper plate resulting in pronounced deflection of the continental margin in the latest Cretaceous–Paleogene (Coe et al., 1985). In addition to these, dextral strike-slip faulting has transported the forearc basin as much as 110 km southwestward with respect to the exhumed magmatic arc (Grantz, 1966; Pavlis et al., 1988; Trop et al., 2005). Most studies of the forearc basin focus on the geology of the better -studied upper CIB and Matanuska Valley. We focus here on categorizing the joint and vein sets of the lower CIB and interpreting their relative timing in order to determine how the events listed here influenced deformation of the lower CIB strata. Despite the complex and often unresolved tectonic history of the forearc as a whole, much of the CIB is relatively less deformed than the surrounding terranes (i.e., the Chugach–Prince William and Wrangellia composite terranes), and may preserve early records of deformation that have not been obscured by intense structural overprinting associated with the Yakutat microplate collision that defines the recent geologic history of central Alaska. This study focuses on the lower portion of the CIB as defined by LePain et al. (2013; see also Fig. 1). We establish the relative timing of vein development with respect to regional structures within lower CIB strata and interpret our results within the published tectonic framework to further understand the modes and mechanisms of upper crustal deformation in the CIB.

Summary of Stratigraphy

The CIB is filled by as much as 18.2 km (60,000 ft) of deep-marine to fluvial deposits, about half of which record the Mesozoic birth, death, and gradual exhumation of the Talkeetna arc.

Mesozoic strata in the study area include the Jurassic Talkeetna Formation, the Tuxedni Group, the Chinitna and Naknek Formations (Detterman and Hartsock, 1966), and the Cretaceous Saddle Mountain section (informal name after Magoon et al., 1980; LePain et al., 2012; Figs. 3 and 5). The Early Jurassic Talkeetna Formation, composed primarily of lava flows, volcanic breccias, agglomerates, tuffs, conglomerates, sandstones, and shales, formed the carapace of the south-facing Talkeetna oceanic arc (Clift et al., 2005a, 2005b; LePain et al., 2013; Bull, 2014, 2015).

The lower to middle Jurassic Tuxedni Group is divided by two unconformity-bound sequences (LePain et al., 2011, 2016; LePain and Stanley, 2015), the lower of which includes the Red Glacier Formation (deep-marine, organic-rich shale), Gaikema Sandstone, Fitz Creek Siltstone, and Cynthia Falls Sandstone, and the upper of which includes the Bowser Formation (fossiliferous sandstone) and Twist Creek Siltstone. The Red Glacier Formation is the likely oil-source rock for the Cook Inlet petroleum system (Magoon and Claypool, 1981; LePain et al., 2013; Stanley et al., 2013; Helmold et al., 2016). The Tuxedni Group records Early Jurassic denudation of the Talkeetna arc carapace (Detterman and Hartsock, 1966; Plafker et al., 1989; Trop et al., 2005).

The Middle Jurassic Chinitna Formation uncomformably overlies the Tuxedni Group and contains the gray Tonnie and Paveloff Siltstone Members (Detterman and Hartsock, 1966; Herriott and Wartes, 2014a, 2014b; Herriott et al., 2016a). The Chinitna Formation represents a continental shelf setting and deepening of the forearc basin (LePain et al., 2013).

Unconformably overlying the Chinitna Formation is the upper Jurassic Naknek Formation. The basal Chisik Conglomerate Member fills deep-water canyons in the underlying Chinitna Formation (Detterman and Hartsock, 1966; Herriott et al., 2017). The overlying Snug Harbor Siltstone Member is primarily thin bedded and fossiliferous (Wartes et al., 2013), and the top unit, the Pomeroy Arkose Member, is an arkosic sandstone (Detterman and Hartsock, 1966; Herriott et al., 2017). The Naknek Formation represents submarine fans that record exhumation of the plutonic roots of the Jurassic Talkeetna arc (Detterman and Hartsock, 1966; Trop et al., 2005; Wartes et al., 2013; Herriot et al., 2017).

The Late Cretaceous Saddle Mountain section (Maastrichtian, informal name after Magoon et al., 1980; LePain et al., 2012) unconformably overlies the Naknek Formation in two observed localities in the study area (Magoon et al., 1980; Gillis, 2016). This formation is composed of fluvial sandstones, conglomerates, and some minor coals and represents further erosion of the Talkeetna arc (Magoon et al., 1980).

Cenozoic strata on the western shores of the lower Cook Inlet near Chinitna and Tuxedni Bays (Fig. 1) were originally mapped (Magoon et al., 1976) as the Paleocene West Foreland Formation, a coarse conglomerate that locally unconformably overlies the Pomeroy Arkose Member of the Late Jurassic Naknek Formation and the Late Cretaceous Saddle Mountain section (Figs. 3 and 5). The age and classification of this unit are debated. Qualitative facies interpretations by Magoon et al. (1976) led to a West Foreland Formation assignment; however, earlier biostratigraphic ages based on macrofossils derived from the strata in the study area yielded a late Oligocene–middle Miocene age (Seldovian, after Wolfe et al., 1965). The macrofossil age is notably younger than middle Eocene depositional ages constrained (using multiple methods) from sandstones in the West Foreland Formation strata 150 km to the northeast of the study area (Gillis et al., 2016). On the basis of the biostratigraphic age in the study area and facies comparisons with the Eocene West Foreland Formation northeast of the study area, for the purpose of this paper we forego naming the unit and instead refer to the Cenozoic section in the study area as late Paleogene(?) strata (Pcg in Fig. 3). One site of particular importance to this study (JR45, Figs. 2 and 4) is located at the top of the preserved Paleogene(?) section and contains the youngest part of the stratigraphy discussed herein. We acknowledge that the age of the Cenozoic rocks may also extend into the Miocene, especially at site JR45.

Summary of CIB Structures

Early mapping within the CIB laid the framework for modern structural interpretations in the area (Martin and Katz, 1912; Kirschner and Minard, 1949; Hartsock, 1954; Detterman and Hartsock, 1966; Fisher and Magoon, 1978; Fisher et al., 1987a; Detterman and Reed, 1980; Detterman et al., 1996). In the upper part of the CIB, including Matanuska Valley, faults and folds that deform the Paleogene–Neogene sequence commonly root into the underlying Mesozoic basement rocks and have been attributed to the collision of the Yakutat microplate (Fig. 1; Nokleberg et al., 1994; Bruhn and Haeussler, 2006; LePain et al., 2013). The posited western edge of the Yakutat slab (Fig. 1; Eberhart-Phillips et al., 2006) is ~300 km east of the lower CIB study area discussed here.

The CIB is bounded by two regional fault systems. To the northwest, the BBFS generally separates the CIB from the magmatic arc (Fig. 1; Burk, 1965; Detterman and Hartsock, 1966; Detterman and Reed, 1980; Gillis et al., 2014). To the southeast, the Border Ranges fault system separates the CIB from the McHugh Complex of the Chugach terrane (Fig. 1; Pavlis and Sisson, 2003; Pavlis and Roeske, 2007). The fault-cored anticlines that occur throughout Mesozoic and Cenozoic strata of the lower CIB have northeast-trending axes that are subparallel to the two major fault systems, and north-northeast–trending axes in the upper CIB (Fig. 2; Burk, 1965; Magoon and Dow, 1994; Plafker et al., 1994; Nokleberg et al., 1994; Swenson, 1997; Haeussler et al., 2000; Bruhn and Haeussler, 2006; Rouse and Houseknecht, 2012; LePain et al., 2013).

In the study area, the BBFS strikes northeast to north-northeast and separates Jurassic plutonic and volcanic rocks of the Talkeetna arc in the hanging wall from the Mesozoic and Cenozoic sediments of the CIB in the footwall (Fig. 3; Detterman and Hartsock, 1966; Detterman and Reed, 1980). Recent work (Betka et al., 2017) documents that the BBFS dominantly records Paleogene sinistral transpression. In the study area (Fig. 3), Triassic to late Paleogene(?) strata are deformed by generally northeast-striking (Bruin Bay fault) and northwest-striking (cross faults) faults that are part of the BBFS (Fig. 3; Betka et al., 2017). On the Iniskin Peninsula, Middle Jurassic strata are folded to form an anticline-syncline pair known as the Fitz Creek anticline and Tonnie syncline (Fig. 3). These folds are gently doubly plunging to the northeast and southwest. North of Chinitna Bay in the footwall of the fault, Triassic(?) to Eocene strata are gently tilted, forming a southeast-dipping limb of a broad monocline.


Field Methods

We divide the field area (>2000 km2) into two regions (Fig. 3): (1) the Iniskin region south of Chinitna Bay contains the Fitz Creek anticline and Tonnie syncline, and (2) north of Chinitna Bay, the Iliamna region contains a monocline that dips gently to moderately southeast. To determine the orientations and map the distribution of joint and vein sets throughout the study area (Fig. 4), we visited 58 field stations (Figs. 3 and 4) using helicopter support. Stations were generally selected where wave-washed shoreline or formerly glaciated outcrops preserve >100 m2 of bedrock. At each station, we recorded the number of joint and vein sets present; the attitude, type of cement preserved, and mode of fracturing for each; and crosscutting and abutting relationships between sets. The orientation of bedding, formation and member name, stratigraphic age, and rock type were also noted at each station.

Subvertical joint faces preserve quartz and calcite cement and are parallel to the pervasive vein sets that are also cemented with quartz and calcite as well as clay. Thus, we deduced that the joint sets simply record opening along preexisting veins during exhumation and unloading of the study area and hereafter, for the purpose of this paper, we refer to the subvertical joints as part of the vein sets. Numerous bedding-parallel joints were also observed; however, they are not cemented and we did not observe evidence of flexural slip on bedding planes. We infer that bedding-parallel joints are also unloading joints, but have no genetic relationship with the vein sets that they crosscut. Thus, bedding-parallel joints are excluded from our analysis and are not discussed further herein.

We grouped the veins into sets on the basis of strike and dip, such that each set consists of veins with subparallel orientations at each station. An average of ~30 vein strike and dip measurements were taken at each station to create a large data set for statistical analyses (cf. Engelder and Geiser, 1980). We compared vein sets throughout the field area and grouped them (Figs. 46) by orientation analysis (discussed in the following). Cumulative frequency is a scale-independent method for determining the intensity of a vein set given a specified vein-size threshold that is determined by the vein widths (Ortega et al., 2006). To determine the vein intensity (number per unit length) of veins in each set, we measured the cumulative frequency of veins along scan lines normal to each vein set at several outcrops (Rosenthal, 2016).

Fold Tests, Structural Trend, and Statistical Determination of Vein Sets

To determine if the vein-set orientations identified at each outcrop persist regionally, and to test their relationship with the local folds, we compared vein set orientations after they were unfolded (Figs. 4 and 6). Veins were rotated by first removing the local bedding dip at each station (a horizontal-axis rotation; results shown in Figs. 4 and 6), and then rotating the local bedding strike into parallelism with the regional structure trend (a vertical-axis rotation, defined in the following; results are shown in Figs. 5 and 6). The K parameter was used to evaluate the degree of clustering during each step of rotation (cf. Branellec et al., 2015) for each vein set using Fisher statistics (Fisher et al., 1987b). K is a dispersion or concentration factor defined by:
where R is the resultant vector determined by summing the direction cosines of each measurement, and N is the number of measurements. The vector mean and apical angle of the 95% confidence cone (α95) for the distribution of vein poles in each set were also compared (Table 1). Both the rotations and statistical analyses were conducted using Stereonet 9.8 (Allmendinger et al., 2013; Cardozo and Allmendinger, 2013; Fig. 6).

Strike Test

We applied a strike test to evaluate the correlation between minor vertical-axis rotations of bedding and vein sets imparted during growth of the folds in the study area. A strike test, also referred to as the orocline test, is commonly used to evaluate orogen-scale curvature in map view by determining the correlation between changes in regional structural trend from a reference trend for the orogen (So - Sr) and changes in paleomagnetic or deformation fabric directions (e.g., fracture or vein set strikes, strain markers, layer-parallel shortening fabrics) relative to a reference (Fo - Fr) (e.g., Schwartz and Van der Voo, 1983; Yonkee and Weil, 2010a). Yonkee and Weil (2010b) presented a method for applying a weighted least-squares regression analysis to the strike test to determine a best-fit slope, confidence intervals, and the goodness of fit for the correlation. In this paper we apply the weighted least-squares strike test of Yonkee and Weil (2010b) at a more local scale to determine if variation in vein set strike (Fo - Fr, when Fr is the set mean strike) between data stations correlates with local changes in bedding strike (So - Sr, when Sr is the regional structural trend) caused by folding of the strata within the study area (cf. Yonkee and Weil, 2010a; Li et al., 2013).

The slope of the regression line helps to inform the kinematic evolution and relative timing of vein formation with respect to folding; its value gives the percent correlation between vein set strikes and map-scale curvature of bedding strikes that developed during folding, assuming no initial curvature of the vein sets. The following possibilities arise: (1) A regression slope close to zero if Fo - Fr is small, because there is little regional variation in vein set orientations throughout the study area. This implies that the vein sets were not affected by folding, and thus formed later. (2) A slope close to one indicates that both veins and bedding underwent a similar magnitude of local map-view rotation during folding and implies that the veins formed prior to the folds. (3) An intermediate slope indicates that the veins formed progressively with the folds. These three scenarios, respectively, are analogous to the classical primary arc, secondary arc, and progressive arc kinematic models of curved orogenic belts (e.g., Yonkee and Weil, 2010b). Here we apply the same concepts to evaluate correlations between local map-view rotations of the veins and bedding that occurred during the growth of the folds in the study area. Although the K test (discussed herein) determines whether the vein sets were folded, the slope of the weighted least-squares strike test better informs the degree of correlation between folding of bedding and vein sets in the study area.

Structural Trend

The regional structural trend (Sr = 038° ± 1°) of the study area was determined using a pi diagram of poles to bedding attitudes taken on the limbs of the folds, as well as map-trace trends of fold axes and the Bruin Bay fault that were measured in 3 km increments (Fig. 6A). The strikes of the bedding attitudes and trends of the fold axes and fault traces are normally distributed and centered on the pi axis (038°/0°; Fig. 6A). Uncertainty of Sr is given by et= st /√n, which is 1° when st is the standard deviation (st ~18°) of the distribution of bedding strikes and trends of fold and fault traces.

Weighted Regression, and Vein Set and Site Uncertainties

In this study the uncertainty of the regional structural trend (Sr) is small (et = ~1°), so the best-fit slope (m), intercept (b), and weighted misfit (χ2) were calculated using Equations 1 and 2 of Yonkee and Weil (2010b), which are not dependent on individual site uncertainties in structural trend. The goodness of fit of the model, Q, was determined by the right-tailed probability of the χ2 distribution evaluated for n-2 degrees of freedom; Q ≥ 0.1 is considered an acceptable fit (Yonkee and Weil, 2010b, and references therein).

Vein sets were classified at each site on the basis of orientation and field relations. Because the strike test only addresses vertical-axis rotations in the horizontal plane, vein sets were first rotated by restoring the local bedding dip to horizontal (back-tilted). The mean orientation of each set was determined using Fisher statistics (Fisher et al., 1987b). Individual site uncertainties for each vein set include measurement errors and natural dispersion within the sets (σm), structural noise (σn), and uncertainty in the restoration path assumed (σp) when unfolding the deformation fabric in a complex structural setting (e.g., plunging or conical folds). The total site uncertainty (σy) is defined by (from Yonkee and Weil, 2010b):

Measurement error and dispersion (σm) were determined at each site using the apical angle of the 95% confidence cone (α95) for the mean vector of the distribution of poles to veins for each set (Fisher et al., 1987b). Structural noise (σn) is an approximation of random block rotations and strain refraction between sites that may introduce variation in site deformation fabric directions in addition to the effects of regional folding (Yonkee and Weil, 2010b). We approximated structural noise by measuring axial-planar cleavage and cleavage-bedding intersection lineations along the trace of the Fitz Creek anticline on the Iniskin Peninsula, and thus defined σn = α95 for the distribution of cleavage strikes and intersection lineation trends. Uncertainty in the restoration path (i.e., the assumed unfolding method, σp) was assigned a value of 0° because the folds in the study area are generally cylindrical where the data were collected (Figs. 3B and 6A). Yonkee and Weil (2010b) demonstrated that the order of vertical- and horizontal-axis rotation is inconsequential for non-plunging cylindrical folds.

Aerial Photography and Satellite Imagery

In this study we used both high-resolution satellite imagery and aerial photography taken from helicopter to determine the regional extent of the vein and joint sets and to evaluate possible kinematic relationships between sets (i.e., conjugate orientations or Riedel shears). At some field stations we collected continuous high-resolution aerial photographs of the outcrops with a Garmin VIRB from a helicopter. Stills were cropped from the videos and mosaicked with Agisoft (cf. Tavani et al., 2014). The resulting high-resolution photomosaics were imported into ESRI ArcMap 10.2, oriented from field sketches, and georeferenced. Vein sets were interpreted from the georeferenced photomosaics, and a simple programming script was used to calculate the trends of lineaments (Rosenthal, 2016).

To zoom out to smaller scales, we interpreted regional lineaments that are parallel to the vein sets by mapping lineaments on SPOT (Satellite Pour l’Observation de la Terre) 2.5 m resolution satellite images in ESRI ArcMap 10.2, and confirmed some of the SPOT-based observations by field observations in the study area. We measured 2693 lineaments from SPOT images in the Iniskin and Iliamna regions. We infer that the lineaments are unloading joints that formed parallel to the preexisting vein sets, like those observed in outcrop (Fig. 7A).

Relative Timing Constraints

Stratigraphic, Abutting, and Crosscutting Relations

The maximum age of each vein set is bracketed by the age of the youngest strata in which it occurs (Fig. 5). Because multiple vein sets cut all of the strata within the study area, we attempted to determine their relative timing by recording abutting and crosscutting relationships between sets at each station. To minimize bias and collect a large number of observations, we set a scan line parallel to veins from each of the four sets. Using a hand lens, we recorded the number of times one of the other sets either abutted (via T or Y terminations; e.g., Hancock, 1985; Pollard and Aydin, 1988) or crosscut (younger vein fill truncates old vein fill) the scan-line vein. Younger veins may abut against existing veins if the existing vein is open (i.e., only partially cemented), forming a mechanical discontinuity (e.g., Kulander et al., 1979; Grout and Verbeek, 1983). This rule may not hold if the older vein is well cemented, in which case a younger vein may crosscut an existing one. Either way, the observations provide relative timing constraints for each set. However, older veins may open again at a later time (e.g., Ramsay, 1980; Cox and Etheridge, 1983; Laubach et al., 2004), or during progressive deformation (e.g., Bergbauer and Pollard, 2004; Hanks et al., 2006; Sanz et al., 2008; Lacombe et al., 2011; Bons et al., 2012), which could result in conflicting relative timing relationships between sets. We tabulated 608 relative timing observations.

Whole-Rock 40Ar/39Ar Geochronology

In one location, a basaltic dike is preserved that is parallel to one vein set and crosscut by two others. We determined a phenocryst-free whole-rock 40Ar/39Ar age of the dike at the Geochronology facility at University of Alaska Fairbanks. The basalt was crushed, sieved, washed, and hand-picked for phenocryst-free rock chips. The sample was irradiated at McMaster University in Ontario, Canada. Step heating followed the technique of Benowitz et al. (2014, and references therein). Error determination followed McDougall and Harrison (1999). (See Supplemental Item1 for detailed laboratory methods, data tables, and associated figures.)


Overview and Definition of Vein Sets

Vein strikes are widely distributed and mostly steeply dipping throughout the study area (Figs. 48), with maxima near 220 and 40 reflecting a dominance of ~310°-striking subvertical veins that are orthogonal to the regional structural trend. Submaxima occur with subhorizontal vein poles that trend north and south, as well as southeast and northwest, suggesting regionally consistent sets (Figs. 48).

Regionally, the veins occur in four sets (Figs. 46) on the basis of unfolded orientations. All of the veins restore to subvertical orientations, and dispersion decreases, when local bedding dip is removed. Dispersion within each set is further decreased after bedding strike is restored (Figs. 6B–6D), and four statistically distinguishable sets emerge that are defined by their average strikes: 360°, 310°, 260°, and 210° (Figs. 5B and 6). The 360° and 260° sets show the greatest decrease in dispersion (K increases by 7.6 and 5.5, respectively) after back-tilting and strike-correcting bedding, whereas K for the 310° and 210° sets only increases by 3.2 and 1, respectively (Figs. 6C, 6D). Hereafter, we refer to these average attitudes of the vein sets.

Vein Sets and Vein Stratigraphy

There are two dominant sets at most data stations (Figs. 4, 7, and 8), one approximately orogen-normal with an average strike of ~310° (53/56 stations), and one approximately orogen-parallel with an average strike of ~210° (30/56 stations). Two subsets oblique to these dominant sets were less common: an ~260° set (28/56 stations), and an ~360° set (28/56 stations). All of the veins measured throughout the study area have apertures (thicknesses) ranging from ~0.1 to 10 mm and are cemented by calcite, quartz, and/or clay.

310° set. Veins of the 310° set (Figs. 4, 7, and 8) are the most pervasive in the study area and occur in all of the stratigraphic units (Fig. 5). They are typically well developed, regularly spaced, and generally >3 m in length. The 310° set includes both mode I (tensile) and mode II (shear) veins. Where 310° veins record shear, they most commonly display centimeter-scale right-lateral separation of bedding and/or other vein sets (Figs. 7 and 8). Typically, mode II 310° veins form high-density clusters adjacent to mesoscopic northwest-striking right-lateral faults (Figs. 7A–7C; i.e., cross-faults in Betka et al., 2017). The mean intensity for 310° veins with apertures of ≥0.2 mm is ~6.1 m–1, the 1st and 3rd quartiles are ~2.3–8.5 m–1, respectively, and the maximum intensity is ~21 m–1 (Fig. 8F).

210° set. Veins of the 210° set are the second-most dominant, and also occur in all stratigraphic units within the study area (Fig. 5). They are subparallel to the regional structural trend, regularly spaced, well developed, and typically >2 m long. They commonly occur with and are orthogonal to veins of the 310° set, and veins of the two sets crosscut and abut each other. The 210° veins are most commonly mode I (Fig. 7E), but also include mode II veins. Where the veins are mode II, they are sinistral (cf. northeast-striking faults in Betka et al., 2017). The mean intensity for 210° veins with apertures of ≥0.2 mm is 4.1 m–1, the 1st and 3rd quartiles are 2.6 and 5.5 m–1, respectively, and the maximum intensity is 8.3 m–1 (Fig. 8F).

260° set. Veins of the 260° set are common in Mesozoic strata (Figs. 5, 7, and 8), but occur with less intensity in late Paleogene(?) strata than those of the 310° and 210° sets (Figs. 8C, 8D). The 260° veins do not occur in the youngest part of the stratigraphic section that was observed (site JR45, Oligocene–Miocene?; Figs. 2 and 5). The 260° veins are typically <2 m in length and were most commonly observed as mode I veins. The mean vein intensity for 260° veins with apertures of ≥0.2 mm is 4.8 m–1, the 1st and 3rd quartiles are 2.8 and 6.4 m–1, respectively, and the maximum vein intensity is 8.7 m–1 (Fig. 8F).

360° set. Veins of the 360° set also occur with less intensity than those of the 310° and 210° sets in the late Paleogene(?) section (Figs. 8C, 8D), but they are abundant in Mesozoic strata (Figs. 5, 6, and 8). The 360° veins are also absent from the youngest part of the stratigraphic section (site JR45, Oligocene–Miocene?; Figs. 2 and 5). Generally, 360° set veins are <2 m long and are both mode I and II. Mode II veins in the 360° set preserve millimeter- to centimeter-scale left-lateral separation. At many stations, 360° veins with left-lateral separation were observed to occur with veins of the 310° set that show right-lateral separation (Fig. 8A). The mean vein intensity for 360° veins with apertures of ≥0.2 mm is 5.2 m–1, the 1st and 3rd quartiles range from 3.6 to 6.2 m–1, respectively, and the maximum vein intensity is 8.9 m–1 (Fig. 8F).

Strike Test

The weighted least-squares regression for each of the sets was calculated using Equations 1 and 2 of Yonkee and Weil (2010b). Values of σm range from ~3° to 20° in this study (Table 1). Because we did not have adequate data to estimate σn at every site, we extrapolated the determination of σn from the Fitz Creek anticline (σn ~6°; Fig. 9A) to all of the sites in the study area (see discussion in Yonkee and Weil, 2010b). The strike-test slopes for all of the vein sets demonstrate positive correlation between deviation of the local bedding strike from the regional trend (So - Sr, Sr = 038°; Fig. 6A) and the orientation of local vein strikes from the set reference strike (Fo - Fr, Fr = set mean strike; Figs. 9B–9E). All cases yield an acceptable goodness-of-fit parameter (Q ≥ 0.1; Yonkee and Weil, 2010b), and the residuals are uncorrelated and normally distributed. The slope for the 310° vein set = 0.66 (±0.09) (n = 54), for the 210° vein set = 0.61 (±0.13) (n = 34), for the 260° vein set = 0.75 (±0.12) (n = 30), and for the 360° vein set = 0.79 (±0.12) (n = 32) (Figs. 9B–9E). The results indicate variable degrees of correlation between the vein set strikes and change in bedding strike throughout the study area. Note that Equations 1 and 2 of Yonkee and Weil (2010b) used here do not account for individual site uncertainties in structural trend; this is acceptable because the total error (et) in structural trend is small (et = 1°, Fig. 6A). If instead we use a bedding-strike measurement error of ~5°, then Equations 4 and 5 of Yonkee and Weil (2010b) (which account for individual site uncertainties in structural trends) apply, and the slope values increase by ~0.05–0.08 and errors by ~0.01. These differences are inconsequential for the purpose of this study.

Relative Timing Constraints

Crosscutting and Abutting Relations

All of the vein sets in the study area were observed to crosscut and abut one another, with veins of all sets appearing older or younger in approximately equal numbers of observations (Fig. 10). The only consistent trend is that veins of the 310° set appear to be younger than veins of all other sets more frequently than they are older. The 310° veins commonly occur with 210° veins where both sets were observed to crosscut and abut each other (Figs. 7, 8, and 10). Similarly, veins of the 310° set share mutual crosscutting relationships with veins from both the 260° and 360° sets. Altogether, these observations indicate that all of the sets formed together. However, the observation that the 310° and 210° sets occur in the youngest strata (Site JR45, Oligocene–Miocene?), where the 260° and 360° sets do not, indicates that the 310° and 210° sets persisted longer than the other two.

Whole-Rock 40Ar/39Ar Geochronology

A basaltic dike at field site JR06 intruded parallel to a densely clustered array of 310° set veins. The dike has several internal chilled margins parallel to veins of the 310° set. The dike is crosscut by veins of the 260° and 210° sets, and also truncates members of those sets (Fig. 11). These observations suggest that the dike was emplaced during the development of at least three of the vein sets.

We collected sample 14JR21 from the dike for 40Ar/39Ar geochronology. A phenocryst-free whole-rock separate from the dike yielded an isochron age of 49.1 ± 1.6 Ma, a plateau age of 52.0 ± 0.9 Ma (Fig. 11), and an integrated age of 51.7 ± 0.7 Ma. Based on the isochron regression to initial 40Ar/36Ar, there is no evidence of significant inherited 40Ar. Overlap in plateau and isochron ages (e.g., Benowitz et al., 2011), and their flat age spectra, further indicates no alteration. In addition, lack of other mid-Paleogene igneous activity in the area (i.e., this was an isolated dike) precludes thermal resetting. The ca. 49 Ma isochron age is rejected because the final step in its regression has an anomalous Ca/K ratio and age, indicating a change in composition. The plateau and integrated ages therefore place the age of the dike as ca. 52 Ma.

Joint Sets from Aerial Imagery Measurements

We traced 2693 lineaments from 5-m-resolution SPOT images within the study area (>2000 km2; Fig. 12). Lineament trends were compared with the original, uncorrected field measurements to determine if they are parallel to the vein sets on a regional scale (Fig. 12). Lineaments subparallel to all vein sets were identified in the SPOT imagery throughout the study area. Those parallel to the 310° set are the most common, and lineaments subparallel to the 210° set form an additional common subset. Lineaments subparallel to the 260° and 360° vein sets are also present in the SPOT imagery throughout the study area, although they are much less common than the other sets. Aerial images confirm that the sets defined on the basis of outcrop-scale observations are regionally persistent; thus, we interpret the lineaments as joints that formed parallel to the vein sets observed in outcrop (e.g., Fig. 7A).


Development and Timing of the Vein Sets

We present an interpretation of vein formation in the lower CIB based on a synthesis of field observations, relative timing constraints, and fold-test results. Veins from all four sets observed in this study disrupt strata that range in age from Early Jurassic (Talkeetna Formation) to late Paleogene(?) (Fig. 5). The 310° set has the highest range in intensity (Fig. 8F), and the remaining veins sets have lower ranges of vein intensity that are similar to each other in Mesozoic through late Paleogene(?) strata (Figs. 5 and 8). Crosscutting and abutting observations indicate that the formation of all of the vein sets hosted in Mesozoic through late Paleogene(?) strata at least partly overlapped in time because all of the sets were observed to crosscut and/or abut one another in approximately equal numbers of observations (Fig. 10). In the youngest part of the stratigraphy visited in this study (site JR45, Oligocene–Miocene? strata; Figs. 35) only veins of the 310° and 210° sets were observed, indicating that the 260° and 360° sets stopped forming by the end of the Paleogene while the 310° and 210° sets persisted into Oligocene–Miocene(?) time.

The vein stratigraphy and crosscutting and/or abutting relations are also consistent with geochronologic constraints. The ca. 52 Ma basaltic dike (station JR06, Fig. 3) exhibits multiple internal chilled margins parallel to veins of the 310° set (Fig. 11), suggesting several episodes of intrusion during the opening of 310° veins. Moreover, because the dike is crosscut by some veins of the 260° and 210° sets, but also truncates other members of those sets, we infer that the dike was emplaced during the progressive formation of the 210° and 260° sets, as well as the 310° set. This inference is compatible with the observation that all of the vein sets crosscut and/or abut one another and suggests a mid-Paleogene age for formation of all the vein sets.

On the basis of the vein stratigraphy and relative timing constraints, as well as the ca. 52 Ma age of the dike, we infer that formation of all of the vein sets began and that veins were likely pervasive by the early Eocene. We suggest that vein development was progressive, as indicated by the mutual crosscutting and/or abutting relationships between sets and coeval truncation and crosscutting of the ca. 52 Ma dike. The vein stratigraphy and relative timing constraints suggest that the tectonic events leading to the formation of vein sets occurred throughout the mid- to late Paleogene and possibly into the Miocene.

Interpretation of Vein Sets with Respect to Folding and the BBFS in the Iniskin-Tuxedni Region

We incorporate the fold-test results with the field and GIS observations to interpret the evolution of the four vein sets with respect to regional structures in the study area. Both the K-parameter and the strike-test results (Figs. 6 and 9) show that the orientations of all of the vein sets are positively correlated with folded bedding. The weighted-regression strike-test slopes between ~0.6 and ~0.8 (Fig. 9) for all of the vein sets are consistent with their progressive development during the growth of local folds, namely the Fitz Creek anticline on the Iniskin Peninsula and the southeast-dipping monocline north of Chinitna Bay (Fig. 3). The four vein sets form a classically described pattern for fold-related jointing, such that the 310° and 210° sets are orthogonal and parallel to the regional fold axis, respectively, and the 260° and 360° sets are ~50° oblique to the fold axis (e.g., Price, 1966; Reches, 1976; Cooper, 1992; Bergbauer and Pollard, 2004; Ahmadhadi et al., 2008). Fischer and Wilkerson (2000) demonstrate that joint trends form parallel to axes of maximum curvature of a folded surface because the joints open parallel to the maximum instantaneous stretch. The 210° veins are dominantly mode I and are subparallel to the regional fold axis, and thus are interpreted to record extension normal to the fold axis caused by bending of the strata. Similarly, the 260° and 360° sets are interpreted to reflect cross sets that formed parallel to the axes of maximum curvature imposed during the lateral growth of the folds (e.g., Fischer and Wilkerson, 2000). The 310° set is orthogonal to the fold axis and is also commonly mode I, suggesting that mode I 310° veins formed parallel to a northwest-trending principal shortening direction that is compatible with the northeast-trending folds (Fig. 13). Although the classical fold-related joint patterns match the mean orientations of the four vein sets in this study, additional interpretation is required to account for the fold-test results and the occurrence of mode II veins in both the 310° and 360° sets.

The slope of the strike test yields the degree of correlation between vertical-axis rotations of the vein set and bedding strikes imparted during folding, and the K-parameter test describes the change in dispersion within each set after removing bedding dip and restoring bedding strike. Both the 360° and 260° sets have higher strike-test slopes (~0.79 and ~0.75, respectively; Fig. 9) than the 310° and 210° sets (~0.65 and ~0.61, respectively; Fig. 9), indicating that the former are more strongly correlated with change in bedding orientations throughout the folds than the latter. Supporting this inference, the K-parameter test indicates that dispersion within the 360° and 260° sets decreases more after both horizontal- and vertical-axis rotations to restore local bedding orientations than dispersion within the 310° and 210° sets (Figs. 6 and 9). We infer that the 310° and 210° sets are less correlated with bedding orientations because they began forming later during the growth of the folds than the 360° and 260° sets, and possibly outlasted folding. This interpretation is consistent with the observation that both the 310° and 210° sets occur at site JR45 in the youngest part of the stratigraphic section (Oligocene–Miocene? strata; Fig. 5), whereas the 360° and 260° sets do not. Altogether, the fold-test results and crosscutting and/or abutting relationships indicate that all of the vein sets developed progressively with the folds. However, the 360° and 260° sets likely began forming earlier during fold growth because they are better correlated with change in bedding orientations than are the 310° and 210° sets.

Within the study area the Mesozoic and Paleogene(?) strata are concordant with each other and folded together (cf. Gillis et al., 2016), indicating that folding and vein formation must have occurred during the Paleogene at the earliest. The 360° and 260° sets stopped forming by the end of the Paleogene; the 310° and 210° sets persisted into Oligocene–Miocene(?) time. We postulate that the maximum horizontal principal compressive stress may have been north trending and compatible with the 360° and 260° sets at the onset of folding, and then rotated clockwise during the growth of the folds to a northwest trend compatible with the 310° and 210° sets during mid- to late Paleogene fold growth. However, the relative timing of the sets inferred from the strike-test slopes is not required to explain the regional pattern of the vein sets (i.e., Fischer and Wilkerson, 2000). Although we form our final interpretations of vein formation relative to folding on the basis of unfolding, which decreases dispersion, our interpretations do not depend on unfolding because the vein sets are regionally discernable in the unrotated data (Figs. 6B, 6C), supporting the interpretation that the veins formed at the same time as folding.

The occurrence of mode II veins alongside mode I veins within the 310° and 360° sets (Figs. 7 and 8) possibly also supports the hypothesis that principal stress orientations rotated during folding, to account for both opening and shear modes observed in subparallel veins within each set. Mode II veins in the 310° and 360° sets commonly occur together and display right- and left-lateral separation, respectively (Figs. 7 and 8). The angle between the pairs of veins at a given site is typically <60° (shown in Fig. 8A), indicating that the vein pairs do not form true conjugate sets, consistent with the observation that the veins crosscut and/or abut one another and thus did not form at the same time everywhere. Nevertheless, the apparent left- and right-lateral separations observed from the veins of the 360° and 310° sets, respectively (Fig. 8A), are compatible with subhorizontal northwest-trending shortening. Because the intersection line between all of the vein sets (with bedding dip removed) is subvertical and the regional fold axis trends northeast and is subhorizontal, we deduce that the left- and right-lateral separations associated with mode II veins of the 360° and 310° sets record northwest-trending subhorizontal shortening that is also compatible with the regional folds (Fig. 13). In some places, left-lateral separation was observed along veins of the 210° set (Betka et al., 2017), also consistent with the northwest-trending subhorizontal shortening. Considering that the crosscutting and abutting relationships and fold-test results indicate that the vein sets formed progressively with the folds, and the range in strike for each set (shaded regions in Fig. 13A), we infer that occurrence of both mode I and II veins in the 310° and 360° sets reflects either (1) the orientation of individual members of each set, in that veins subparallel to a northwest-trending principal shortening direction are mode I and those that are oblique are mode II (Fig. 13A); (2) the influence of locally induced strains and/or small rotations of the principal compression direction, which may have occurred during the growth of the folds (e.g., Fischer and Wilkerson, 2000) and may also reflect local shearing that accommodated variation in bed thickness or competency as the stratigraphic section was folded; or (3) a combination of these. In Figure 13, we show our preferred interpretation of the development of the vein sets with respect to the regional folds.

Lineament patterns that are parallel to Paleogene-age vein sets and are documented at larger scales of observation with remote-sensing imagery suggest that the vein sets persist on a regional scale (>2000 km2; Fig. 12) and thus record an important regional Paleogene deformation within the BBFS and CIB (discussed in the following).

Recent kinematic analyses of minor fault populations and slip vectors measured in the study area (Betka et al., 2017) indicate a subhorizontal northwest-trending principal shortening direction within the Bruin Bay system (e3; Fig. 13A). Betka et al. (2017) conclude that the folds in the study area formed during northwest-trending contraction and left-lateral shear within the BBFS during mid-Paleogene deformation of the CIB. On the basis of the observations presented here and the results in Betka et al. (2017), we deduce that the vein sets, folding, and regional faulting within the BBFS are kinematically linked during mid-Paleogene deformation. Moreover, Bruhn and Haeussler (2006) concluded that steeply northwest-dipping faults in the CIB, such as the Bruin Bay fault, are oriented optimally for dextral transpression under the contemporary east-southeast–trending subhorizontal principal compressive stress direction (s1), as determined from analog modeling of fault-cored anticlines and earthquake focal mechanisms in the CIB (Flores and Doser, 2005). Bruhn and Haeussler (2006) presented a tectonic model that considers a temporally variable stress field where s1 rotates clockwise and trends southeast during periods of time when the Aleutian megathrust is locked and elastically loading; they postulated that both southeast-trending s1 and abnormally high fluid pressures are required to trigger rupture and reverse slip along steeply northwest-dipping faults of the CIB. We speculate that the veins in the lower CIB are hydrofractures that might have formed during episodes of elevated fluid pressure and northwest-southeast–trending contraction within the CIB associated with mid-Paleogene slip along the BBFS and growth of the Fitz Creek anticline. This hypothesis can be tested with compositional and textural analysis of the vein fill; this is beyond the scope of this study.

Tectonic Evolution of Veins and Regional Deformation in the CIB

Brief Overview of South-Central Alaska Margin Tectonics

We seek to place the development of veins associated with folds and faults of the lower CIB into a regional structural framework. Four main tectonic events are well recorded and documented along the southern Alaska margin. These include the Jurassic to Cretaceous collision and indentation of the Wrangellia composite terrane (Plafker et al., 1989, 1994); eastward migration of a trench-ridge-trench triple junction from middle Paleocene to early Eocene time (Bradley et al., 2000); northward and westward translation of the Chugach terrane from Late Cretaceous to early Eocene time (Plafker et al., 1994; Roeske et al., 2003; Day et al., 2016); and the predominantly Neogene to present subduction and collision of the Yakutat microplate (Haeussler, 2008). Less well understood or documented are the following other postulated events: intervals of subduction erosion (Clift et al., 2005a; Amato and Pavlis, 2010; Amato et al., 2013); Early Cretaceous subduction of an oceanic spreading center (Pavlis et al., 1988; Barnett et al., 1994); and oroclinal bending of the southern Alaskan margin (Grantz, 1966; Coe et al., 1985; Glen, 2004). The following is a brief summary of the tectonic events that, based on our results, might have influenced vein formation during Paleogene deformation in the CIB. We exclude events with no documented record in the Cook Inlet sector of the forearc basin, or that occurred prior to the Paleogene, because the veins are hosted in both latest Cretaceous (Maastrichtian) and Paleogene(?) strata. We therefore exclude the Mesozoic docking of the Wrangellia composite terrane (Plafker et al., 1989; Nokleberg et al., 1994; Trop et al., 2005; Trop and Ridgway, 2007; Hampton et al., 2007) and Early Cretaceous ridge subduction (Pavlis et al., 1988) as possible drivers of the deformation discussed in this study.

Timing of deformation within the study area overlaps with (1) postulated Paleogene subduction of a spreading ridge below the Iniskin-Tuxedni region (the hypothesized Kula-Resurrection spreading ridge; Haeussler et al., 2003b; Madsen et al., 2006), (2) Paleogene arrival of the Chugach–Prince William terrane to south-central Alaska after as much as thousands of kilometers of dextral translation along the Border Ranges fault system, and (3) far-field deformation of the CIB that is associated with the arrival and subduction of the Yakutat microplate (e.g., Finzel et al., 2011; see also discussion in Betka et al., 2017). Southern Alaska is the site of widespread deformation and magmatism throughout the Eocene (e.g., Little and Naeser, 1989; Burns et al., 1991, and references therein; Bradley et al., 2000). Thus it is difficult to distinguish whether deformation within the BBFS is related to spreading ridge subduction or recently postulated accretion of the Chugach–Prince William terrane, both of which are thought to have occurred during the Paleogene. Further complicating the tectonic setting, a postulated ~50° of counterclockwise oroclinal bending of the southern half of Alaska, thought to have occurred between ca. 65 and 50 Ma (e.g., Hillhouse and Coe, 1994), overlaps in time with the ridge subduction event and with northward translation of inboard terranes (e.g., Beck, 1980; Coney et al., 1980; Butler et al., 2001). Glen (2004) proposed a kinematic scenario wherein bending of the southern Alaskan orocline is partly accommodated by shortening on northeast-striking thrust fault splays off of northwest-striking dextral strike-slip faults. In the Cook Inlet region, this shortening may have been manifest in motion along the Border Ranges fault and BBFS, and thus is a possible explanation for Paleogene northwest-trending compressive stress recorded by vein patterns and fold orientations within the study area. This event may be genetically related to the subduction of a spreading ridge and/or arrival of the Chugach–Prince William terrane.

Perhaps supporting the ridge-subduction hypothesis, the ca. 52 Ma northwest-striking basaltic dike (Fig. 11) may reflect magmatism related to the subduction of the spreading ridge. Paleogene–Neogene basaltic dikes are reported north of the study area and throughout southern Alaska (e.g., Csejtey et al., 1978; Silberman and Grantz, 1984; Miller et al., 1996; Cole et al., 2006; Haeussler et al., 2009) that have been previously interpreted to record slab window magmatism (e.g., Benowitz et al., 2012). Zeumann and Hampel (2015) showed that strain patterns in the forearc region above a subducting spreading ridge vary as the ridge axis migrates along the margin. Trench-parallel displacements within the upper plate occur in the spreading direction and can affect the upper plate hundreds of kilometers inboard of the trench (Zeumann and Hampel, 2015; Lagabrielle et al., 2004; Ramos, 2005; Russo et al., 2010). We postulate that Eocene subduction and eastward migration of a spreading ridge could have imparted northwestward-directed shear tractions at the base of the crust above the trailing Kula plate, which drove northwest-directed contraction within the CIB (cf. Haeussler et al., 2003a; Betka et al., 2017). The Cocos Ridge and Chile Ridge subduction events may serve as analogs to the scenario we propose, in which veins develop alongside faulting (the BBFS) and folding during impact and subduction of a spreading ridge. At the Panama–Costa Rica impact site of the Cocos Ridge, forearc sediments similarly record rapid basin fill followed by deformation and uplift (Corrigan et al., 1990). Strata were displaced vertically and horizontally along two major faults, and structures within the sequence record minor, subhorizontal, margin-parallel extension and margin-perpendicular shortening (Corrigan et al., 1990). Corrigan et al. (1990) suggested in particular that the inner forearc underwent subhorizontal compression due to basal shear stresses resulting from impingement of the Cocos Ridge, in contrast to vertical movements in the outer forearc resulting from isostatic adjustment (see also Geist et al., 1993). We suggest that such a scenario is likely applicable to the CIB. However, shortening associated with the bending of the Alaskan orocline (e.g., Glen, 2004) and/or Paleogene juxtaposition of the Chugach–Prince William terrane against south-central Alaska could also explain the observed Paleogene deformation within the study area without a ridge subduction event (cf. Betka et al., 2017).

Although it is impossible to unequivocally distinguish one tectonic interpretation from another without better timing constraints and improved understanding of the Paleogene tectonic setting of southern Alaska, the results presented herein indicate that the majority of the deformation within the study area occurred during the Paleogene (Betka et al., 2017) and is thus attributed to early Cenozoic tectonic reorganization of the southern Alaska margin.

The Oligocene arrival and flat-slab subduction of the Yakutat terrane (e.g., Trop and Ridgway, 2007; Finzel et al., 2011) may have also contributed to deformation of the forearc region, but this event is difficult to pinpoint in the vein stratigraphy. Surface uplift and deformation associated with the subducting Yakutat slab is restricted to the upper CIB (e.g., Haeussler and Saltus, 2011). The Iniskin-Tuxedni area of the CIB is >300 km from the proposed (Eberhart-Phillips et al., 2006) and imaged (Kim et al., 2014) northwestern edge of the Yakutat plate (Fig. 1), within a region where the modern-day horizontal velocity field indicates <5 mm/yr northwestward motion (Finzel et al., 2011). In contrast, the northern CIB is moving at 2–3 times that rate. Similarly, the strain rate in the lower CIB is significantly lower than that in the upper CIB (Finzel et al., 2011). Consequently, deformation associated with Yakutat collision and subsequent rotation has been well documented in the upper CIB (e.g., Bruhn and Haeussler, 2006; Haeussler and Saltus, 2011). In contrast, detailed thermochronology in the lower CIB region around the BBFS indicates slow cooling interpreted as low exhumation rate at the time of Yakutat microplate collision (Gillis et al., 2014). It is possible that the Oligocene onset of flat-slab subduction may have imposed northwest-trending horizontal compressive stress throughout the forearc region that is compatible with the 310° and 210° vein sets in this study. However, the extent of the flat-slab segment in the Oligocene is not well known, and in the modern setting the lower part of the CIB does not appear to undergo deformation associated with the Yakutat plate (e.g., Ruppert, 2008). Normal subduction-related deformation associated with the Pacific plate shows that the lower and upper CIB remain under different stress conditions (see Fig. 2 for focal mechanisms). The upper CIB is undergoing compressive stresses that result in earthquakes along north-striking thrust faults, while lower CIB earthquakes are along northeast-striking left-lateral faults (focal mechanisms in Fig. 2 are derived from stress tensors defined in Ruppert, 2008), suggesting that Yakutat-related stresses differ between the upper and lower Cook Inlet regions.


Vein sets within the Iniskin-Tuxedni region of the CIB record mid- to late Paleogene deformation of the forearc strata, and formed during the progressive growth of regional folds. There are four vein sets in the study area that can be classified on the basis of their unfolded orientations, relative ages, and preserved opening modes. The sets are grouped according to their strikes: 360°, 260°, 310°, and 210°. The 310° set is the most pervasive and has an average vein intensity of 6.1 m–1 for veins with an aperture of ≥0.2 mm. The other vein sets are less pervasive. All four sets share mutual crosscutting and abutting relationships, indicating that they formed contemporaneously.

A 40Ar/39Ar whole-rock age of ca. 52 Ma from a basaltic dike that intrudes parallel to veins of the 310° set, and both truncates and is crosscut by members of the 210° and 260° sets, partially constrains the age of 310°, 210°, and 260° vein sets and is consistent with the age of mid- to late Paleogene(?) strata that host the vein sets. Two vein sets (360° and 260°) are absent from the youngest part of the stratigraphic section that was observed (site JR45; Oligocene–Miocene?), limiting the age of the 360° and 260° sets to the Paleogene.

Fold-test results confirm that the development of the vein sets occurred progressively during the growth of the Fitz Creek anticline and broad monocline north of Chinitna Bay. Veins of the 360° and 260° sets yield strike-test slopes of ~0.79 and ~0.75, respectively, while the 310° and 210° sets have lower slopes (~0.66, ~0.61). The former sets are better correlated with folded bedding orientations throughout the study area and thus formed earlier during folding. The vein sets occur in classically described patterns of fold-related joint sets. The 310° and 210° sets are normal and parallel to the axis of the Fitz Creek anticline, respectively, and the 260° and 360° sets strike ~50° to the fold-axis-normal set. The occurrence of subparallel mode I and II veins in each set suggests that small rotations of the stress field and/or local variation in fold-related strains occurred throughout folding. We conclude that the main phase of folding and vein development occurred together during mid- to late Paleogene(?) contraction within the BBFS.

Pervasive joint sets, identified both in outcrop (approximately tens of meters) and larger spatial scales (approximately tens of kilometers) using remote sensing imagery, formed along preexisting veins and indicate that the veins reflect regional deformation of CIB strata. We suggest that vein formation and related deformation within the BBFS in the study area reflect the subduction of a spreading ridge and/or arrival of the Chugach–Prince William terrane to southern Alaska. Far-field northwest-trending compressive stress from the Oligocene arrival of the Yakutat microplate may have caused vein formation in the younger, late Paleogene–Miocene(?) strata.


This project was supported by the U.S. Geological Survey National Cooperative Geologic Mapping Program under STATEMAP awards G13AC00157 and G15AC00199, as well as the State of Alaska Division of Geological and Geophysical Surveys (ADGGS) and a Graduate Student Research Grant provided by the Geological Society of America to J. Rosenthal. This paper is part of a master’s thesis completed by Rosenthal at the University of Alaska, Fairbanks. We thank CIRI (Cook Inlet Region, Inc.), and 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 and Merlin (“Spanky”) from Pathfinder Aviation for safe transportation during the field seasons. Rebekah Tsigonis and Paul Wilcox were excellent field assistants. Conversations with Marwan Wartes, Trystan Herriott, and Dave LePain of ADGGS, and Rick Stanley, Ken Helmold, and Paul Decker of the Alaska Division of Oil and Gas helped us better understand the local stratigraphy. Anupma Prakash helped guide the geographic information system study. We thank Arlo Weil for guidance on implementing the strike test. Jonathan Caine and Peter Haeussler provided thorough and constructive reviews that significantly improved this manuscript.

1Supplemental Item. Detailed laboratory methods, data tables, and associated figures. Please visit http://doi.org/10.1130/GES01435.S1 or the full-text article on www.gsapubs.org to view the Supplemental Item
Science Editor: Raymond M. Russo
Guest Associate Editor: James V. Jones
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.

Supplementary data