Abstract

We present U-Pb geochronologic and Hf isotopic data from 29 plutonic samples within the Coast Mountain batholith, north-coastal British Columbia and southeast Alaska. Hf isotopic values do not correlate with age or variation in magmatic flux, but rather they increase systematically from west (εHf[t] = +2 to +5) to east (εHf[t] = +10 to +13) in response to changing country rock assemblages. By comparing our pluton Hf data with previously reported Nd-Sr and detrital zircon characteristics of associated country rocks, we identify three crustal domains in an area where crustal affinity is largely obscured by metamorphism and voluminous pluton intrusion: (1) a western domain, emplaced into continental-margin strata of the Banks Island assemblage; (2) a central domain, emplaced into the Alexander terrane; and (3) an eastern domain, underlain by the Stikine terrane and its inferred metamorphic equivalents. Between the interpreted Alexander and Stikine terranes, there is a zone of variable εHf(t) (+2 to +13) that coincides with the suture zone separating inboard (Stikine and Yukon-Tanana) from outboard (Alexander and associated) terranes. This variation in εHf(t) values apparently results from the structural imbrication of juvenile (Alexander and Stikine) and evolved (Yukon-Tanana) terranes along mid-Cretaceous thrust faults and the latest Cretaceous–early Tertiary Coast shear zone. Shifts in the Hf values of plutons across inferred terranes imply that they are separated at lower- to midcrustal levels by steep boundaries. Correlation between these Hf values and the isotopic character of exposed country rocks further implies the presence of those or similar rocks at magma-generation depths.

INTRODUCTION

The Coast Mountains batholith is a 1700-km-long belt of Jurassic through Tertiary plutonic rocks that extends along the length of coastal British Columbia, southeast Alaska, and southwestern Yukon (Fig. 1). These plutonic rocks are emplaced along the suture zone between two large arc-type fragments: the Alexander and Wrangellia terranes to the west, and the Stikine and associated terranes to the east. Although clearly related to plate convergence along the western margin of North America (e.g., Engebretson et al., 1985), it has long been suspected that Jurassic–Cretaceous collision of terranes along this suture played a significant role in the generation and exhumation of igneous rocks that make up the Coast Mountains batholith (e.g., Monger et al., 1982). This report uses Hf isotope data from plutonic rocks within and adjacent to the Coast Mountains batholith to investigate the petrogenesis of granitoids that make up this segment of the Coast Mountains batholith, and the architecture of the terranes that comprise the central-western Coast Mountains. The coupled Hf and geochronologic data shed new light on the nature of terranes at depth and the structural boundaries that separate them.

Much attention has been paid to the metasedimentary assemblages that occur as pendants within the widespread Jurassic through Eocene Coast Mountains batholith (e.g., Samson et al., 1991a; Jackson et al., 1991; Boghossian and Gehrels, 2000; Gareau and Woodsworth, 2000)(Fig. 2). Typically, these rocks are highly deformed and metamorphosed to amphibolite or even granulite grade (Hollister and Andronicos, 2000). Because of the voluminous nature of the plutonic rocks and the high grade of metamorphism, contacts are commonly obscured, and protolith determinations are difficult to make. This in turn has caused difficulties in correlating pendant rocks with terranes described for other parts of the Cordillera, and it has led to a limited understanding of the tectonostratigraphic relationships between the various crustal assemblages making up the central Coast Mountains batholith. In various parts of the batholith, however, the Nd and Sr isotopes of the metamorphic country rocks have been studied (Samson et al., 1990, 1991a, 1991b; Jackson et al., 1991; Patchett et al., 1998; Boghossian and Gehrels, 2000; Gareau and Woodsworth, 2000). These isotopic data, together with published U-Pb detrital zircon data and geologic observations, allow us to define the isotopic character of the various amalgamated terranes that make up the Coast Mountains. Hf isotopes of magmatic rocks that have interacted, even to a small degree, with a given terrane, should record signatures consistent with that region. Hf data from widely distributed plutons within the Coast Mountains batholith can therefore be used to identify the terranes with which the plutons interacted and/or partly assimilated.

Hafnium isotopes in magmatic zircons also act as probes for the chemical maturity of the lithosphere from which the melts were extracted. As such, the distribution of Hf isotopic signatures across the batholith can be used to glean important information about the tectonic construction of the terranes that make up the central Coast Mountains. Terranes are commonly conceptualized as discrete, coherent blocks of lithosphere that are accreted to, or transported laterally along, the margins of continents. In such a conceptual framework, terranes form “side-by-side” panels separated from adjacent terranes or continents by through-going, vertical structures. Alternatively, during their accretion to or collision with an existing margin, terranes can become imbricated by large-scale thrusts, making the boundaries between them more diffuse and less steep, as is the case with the mid- to Late Cretaceous fold-and-thrust system that developed along the boundary between the Insular (Wrangellia and Alexander) and Intermontane (Stikine and Yukon-Tanana) superterranes (Rubin et al., 1990). The thrust belt thickened the crust and effectively stacked rocks of the various existing terrane assemblages, laterally smearing them along thrusts. We use the arc-perpendicular distribution of Hf isotopes to evaluate the nature of the terrane boundaries and the degree to which the terranes have been structurally interleaved.

GEOLOGIC AND TECTONIC FRAMEWORK OF THE COAST MOUNTAINS BATHOLITH

Most of the igneous rocks that make up the Coast Mountains batholith are tonalitic and range in age from 160 Ma to 50 Ma (Gehrels et al., 2009). In general, the ages become younger progressively eastward, although there are also Jurassic ages in the easternmost part of the Coast Mountains batholith at the latitude of the study area. This is interpreted to result from sinistral strike-slip duplication of the Jurassic portion of the batholith during Early Cretaceous time (Gehrels et al., 2009).

Country rocks of the batholith are generally mid- to high-grade metasedimentary assemblages derived from marine strata (Wheeler and McFeely, 1991). Due to the interpreted geologic setting of these protoliths, and supported by Nd-Sr (e.g., Samson et al., 1989) and detrital zircon (e.g., Gehrels and Boghossian, 2000) data, the country rocks are interpreted to have formed in settings ranging from juvenile volcanic arcs to pericratonic passive margins.

The lithologic and isotopic character of the main plutonic suites and their country rocks are described next and shown in Figures 1 and 2.

Western Portion of the Coast Mountains Batholith

The western portion of the Coast Mountains batholith is underlain by three distinct belts of plutonic rocks of Late Jurassic, Early Cretaceous, and mid-Cretaceous age. The ages of these bodies decrease systematically eastward (van der Heyden, 1989, 1992; Butler et al., 2006). Their composition also changes eastward, from predominantly quartz diorite on the west to mainly tonalite on the east. The emplacement depth of these bodies ranges from ∼10 km on the west to ∼25 km on the east (Butler et al., 2001, 2006), a change that is also reflected in the increasing metamorphic grade of the metasedimentary host rocks. A magmatic flux curve for the western portion of the Coast Mountains batholith, based on ages from these plutons, suggests high-flux periods from 160 to 140 Ma and 120 to 80 Ma, with little magmatism between 140 and 120 Ma (Gehrels et al., 2009, their Fig. 9).

Country rocks to these plutons (Fig. 1), from west to east, include the Wrangellia terrane, Banks Island assemblage, Alexander terrane, Gravina belt, and Yukon-Tanana terrane, as described in the following.

The Wrangellia terrane in the study area consists of Upper Paleozoic arc-type metavolcanic and metasedimentary rocks and Triassic rift-related(?) pillow basalts (Monger et al., 1992). These rocks presumably formed in a marine volcanic arc setting on the basis of their primitive Nd-Sr signature of correlative rocks on Vancouver Island (Samson et al., 1990).

The Banks Island assemblage consists mainly of highly folded quartzite interlayered with marble and subordinate pelitic schist (Gehrels and Boghossian, 2000). These rocks are intruded by a ca. 357 Ma orthogneiss and are crosscut by nondeformed dikes that are ca. 147 Ma in age (G. Gehrels, 2010, personal commun.). Detrital zircons in the quartz-rich strata yield dominant ages of 410–480 Ma, 1700–1850 Ma, 1940–2250 Ma, and 2620–2940 Ma. Metamorphic rocks of the Banks Island assemblage have evolved continental isotopic signatures, with initial Nd values ranging between +0.5 and −9.9, and relatively radiogenic Sr values ranging from 0.71178 to 0.71934 (Boghossian and Gehrels, 2000). These rocks are interpreted to have formed in a continental-margin environment because of the occurrence of interlayered metaclastic quartzite and marble and cratonal detrital zircon and Nd-Sr signatures. The occurrence of ca. 410–480 Ma detrital zircons, however, suggests possible connections with the Alexander terrane.

The Alexander terrane is composed of Neoproterozoic–Cambrian and Ordovician–Silurian meta-igneous and metasedimentary rocks that are interpreted to have formed in a marine volcanic arc (Gehrels and Saleeby, 1987). There is no sign of continental input in these assemblages in the geologic units (e.g., quartz-rich clastic strata; Gehrels and Saleeby, 1987), U-Pb geochronologic data (e.g., Precambrian inherited or detrital zircons; Gehrels et al., 1987, 1996), or Nd-Sr isotopes (Samson et al., 1989). Beginning in Early Devonian time, conglomeratic strata (referred to as the Karheen Formation) were shed from a source to the southwest (in present coordinates) that included rocks of 1120–2230 Ma age (Gehrels et al., 1996). Middle and Upper Paleozoic strata consist of shallow-marine clastic strata and carbonates with little sign of arc- or craton-derived detritus. The sequence is capped by a Triassic assemblage of rift-related(?) metavolcanic and metasedimentary rocks.

The Gravina belt consists of Upper Jurassic through Lower Cretaceous volcaniclastic turbidites and subordinate mafic and felsic metavolcanic rocks (Berg et al., 1972). These strata are interpreted to depositionally overlie the Alexander terrane to the west and the Yukon-Tanana terrane (described next) to the east (Gehrels, 2001). Detrital zircons in the metaclastic rocks record derivation from both the Alexander and Yukon-Tanana terranes, which suggests proximity of the two terranes by Late Jurassic time (Gehrels, 2001), as indicated on the basis of geologic relations examined by McClelland et al. (1992) and Saleeby (2000).

The Yukon-Tanana terrane consists of a Proterozoic–Lower Paleozoic assemblage of mainly quartz-rich metaclastic rocks (psammitic schist) and marble, a mid-Paleozoic assemblage of metavolcanic rocks, and Upper Paleozoic metasedimentary and metavolcanic rocks (Gehrels et al., 1992; Gehrels, 2001). In the study area, these rocks are locally referred to as the Scotia-Quaal assemblage, and they consist mainly of Devonian–Mississippian metavolcanic rocks and orthogneisses (Gareau, 1989). Although somewhat ambiguous, Nd-Sr data from the Scotia-Quaal are more evolved than those from the Wrangellia, Alexander, and Stikine terranes, and they indicate the presence of old continental crust (Gareau and Woodsworth, 2000).

Primary relations between these assemblages are difficult to document because of younger deformation, pluton intrusion, and/or lack of exposure. For example, the contact between Wrangellia and the Banks Island assemblage is everywhere under water or intruded by Mesozoic plutons, the contact between the Banks Island assemblage and Alexander terrane is a major sinistral strike-slip fault (Kitkatla shear zone) of Early Cretaceous age, and the Yukon-Tanana terrane and eastern Gravina belt are highly imbricated along west-vergent thrust faults of mid-Cretaceous age (Chardon et al., 1999; Gehrels et al., 2009).

Axial Portion of the Coast Mountains Batholith

Axial portions of the Coast Mountains are underlain primarily by Paleocene tonalitic sills, large bodies of Eocene granodiorite, and high-grade metasedimentary rocks. Emplacement and ductile deformation of the Paleocene tonalitic sills occurred during east-side-up motion along the Coast shear zone, which can be traced for most of the length of the Coast Mountains (Gehrels et al., 2009). Elsewhere, the Coast shear zone reveals evidence for older dextral motion (Gehrels, 2000) and younger east-side-down motion (Rusmore et al., 2001; Hollister and Andronicos, 2006). Barometric studies of tonalitic sills in southeast Alaska indicate emplacement depths of ∼15–20 km (Hollister et al., 1987; Stowell and Crawford, 2000; Rusmore et al., 2005).

East of the sills and Coast shear zone, there are large plutons of homogeneous granodiorite that are primarily of Eocene age. These plutons are generally nondeformed, but locally they were emplaced along an east-side-down normal fault and associated ductile shear zone referred to as the Shames mylonite zone (Heah, 1990, 1991; Andronicos et al., 2003) or eastern boundary detachment (Rusmore et al., 2005).

Metasedimentary assemblages within axial portions of the Coast Mountains batholith, which are commonly referred to as the Central Gneiss complex, consist mainly of pelitic and psammitic schist with subordinate quartzite, marble, and calc-silicate gneiss (undivided metamorphic rocks in Figs. 1 and 2). These rocks are generally sillimanite grade, locally with sillimanite replacing kyanite and/or staurolite (Stowell and Crawford, 2000; Hollister and Andronicos, 2000; Rusmore et al., 2005). The tectonic affinity of these metasedimentary rocks in the study area is uncertain. To the north, in southeast Alaska, detrital zircon and Nd-Sr analyses record derivation primarily from continental source regions (Samson et al., 1990; Gehrels et al., 1992). Regional correlations and northward continuity suggest that these rocks belong to the Yukon-Tanana terrane. In the study area, however, quartzites and marbles that are characteristic of the Yukon-Tanana terrane are rare, and some workers have suggested correlations with strata of the Stikine terrane (Hill, 1985).

Eastern Portion of the Coast Mountains Batholith

The eastern portion of the Coast Mountains batholith consists of Jurassic through Eocene plutons that intrude low-grade Upper Paleozoic through Tertiary sedimentary and volcanic rocks of the Stikine terrane and overlying strata of the Bowser Basin (Wheeler and McFeely, 1991; Haggart et al., 2006a, 2006b, 2007; Mahoney et al., 2007a, 2007b, 2007c, 2007d, 2007e, 2009). The Jurassic–Cretaceous history of this portion of the batholith is somewhat different from that of the western portion because magmatism did not migrate eastward, and it was apparently continuous through Early Cretaceous time (Gehrels et al., 2009; Mahoney et al., 2009).

The Stikine terrane consists largely of widespread Triassic and Jurassic arc-type assemblages blanketed by Jurassic–Cretaceous marine strata of the Bowser Basin (Monger et al., 1992). Locally, arc-type volcanic and sedimentary assemblages as old as Devonian are exposed. Available Nd-Sr data from these rocks suggest formation in a juvenile arc setting with little continental influence (Samson et al., 1989). The only exception to this is the occurrence of inherited zircons of Precambrian age in Early Jurassic plutons, which may reflect the presence of Precambrian basement in some portions of the terrane (Thorkelson et al., 1995). Alternatively, it has been suggested that the Stikine terrane rests depositionally on rocks of the Yukon-Tanana terrane (e.g., McClelland et al., 1992; Jackson et al., 1991), which carries predominantly Precambrian detrital zircons.

ANALYTICAL STRATEGY AND METHODS

Zircons from 29 individual plutonic samples were separated, picked, and mounted, along with appropriate U-Pb and Hf standards, in 2.5 cm epoxy mounts. Epoxy mounts were imaged in plain light with a binocular microscope, and grain maps were produced. U-Pb analysis was first performed on individual zircon grains from each sample via ablation of 40-μm-diameter pits using methodology described next. After all U-Pb analyses for a sample were completed, Hf isotope measurements were made via ablation on top of the preexisting U-Pb pits; the following sections describe the details of these analyses. This analytical technique allows for the measurement of both U-Pb and Hf in the same part of the zircon crystal, which is important for two reasons: (1) Crystallization age of the analyzed zircons is required for calculating initial 176Hf/177Hf; and (2) zircons with complex growth histories may record multiple ages, such that it is necessary to collect Hf and U-Pb data from the same zircon domain. Because the data are acquired sequentially, and not simultaneously, it is possible that the successive pits drilled for U-Pb and Hf measurement are sampling different crystal depth domains (Kemp et al., 2009). This is likely not the case with zircons from the Coast Mountains batholith samples, however, given the characteristic homogeneity of analyzed crystals, as determined by cathodoluminescence imaging and age mapping of large grains, and the fact that we observed no marked changes in Hf isotopic ratios with time during data acquisition.

Instrumentation

Reported U-Pb and Hf isotope data were collected using a New Wave 193 nm ArF laser ablation system coupled to a Nu Plasma HR inductively coupled plasma–mass spectrometer (ICP-MS) at the University of Arizona (for additional information, see http://sites.google.com/a/laserchron.org/laserchron/). Ablation was performed in a New Wave SuperCell™, and sample aerosol was transported with He carrier gas through Teflon-lined tubing, where it was mixed with Ar gas before introduction to the plasma torch. The multicollector (MC) ICP-MS utilizes 12 Faraday detectors equipped with 3 × 1011 Ω resistors and four discrete dynode ion counters, which remain fixed as beams are directed into them via an electrostatic zoom lens. For U-Pb analyses, U, Th, and Pb isotopes were measured simultaneously in Faraday collectors, with the exception of 204Pb, which was measured using an ion counter. For Hf analyses, masses 171 through 180 were all measured simultaneously in Faraday collectors. Pure Hf solutions and Hf solutions doped with various amounts of Yb and Lu were introduced in Ar carrier gas via a Nu DSN-100 desolvating nebulizer.

U-Pb Geochronology

Geochronologic analyses presented here were performed by laser ablation (LA) ICP-MS at the Arizona Laserchron Center using methods described by Gehrels et al. (2008). Laser ablation was done using a 40-μm-diameter spot and a pulse rate of 7 Hz. The laser was run in constant energy mode with output energy of 8 mJ/pulse, which corresponds to an energy density of ∼2 J/cm2 and an estimated excavation rate of 0.7 μm/s. The analytical routine consisted of a 15 s on-peak background measurement with the laser off, followed by 15 s of peak measurement, performed at 1 s integration times, with the laser firing. This results in an analysis pit of ∼15 μm depth.

The samples we used are a subset of those analyzed by Gehrels et al. (2009), and they generally can be characterized by simple, prismatic zircons with internal oscillatory zoning and rare inherited components or younger growth rims. For each sample, single pits were ablated on 20–30 individual zircons. In the case of samples 04GJP-09, 04GJP-13, and 05MT-135, which had zircons with rims and cores distinguishable in cathodoluminescence images and of variable age, multiple analyses (2–5 spots at 40 μm per spot) were performed on single crystals. Weighted mean ages were then determined for each component, and a magmatic age was assigned based on the interpreted igneous domain. Fractionation between U and Pb was accounted for by bracketing every five measurements with analysis of a Sri Lankan zircon standard of known age (see Gehrels et al., 2008). Corrections for the interference of mercury were made by monitoring 202Hg and using the natural ratio of 202Hg/204Hg to subtract the Hg contribution from mass 204. Corrections for common Pb were made by measuring 204Pb and assuming an initial Pb composition based on the Pb evolution model of Stacey and Kramers (1975). Uncertainties for reported 238U-206Pb ages are ∼1%–2% (2σ) and include both a systematic error (typically ∼1%–2%), and an error associated with the scatter and precision of a set of measurements for a given sample (∼1%, 2σ) (for details of error analysis, see Gehrels et al., 2009).

Hf Isotope Measurement

Interference Correction

Accurate in situ measurement of Hf isotopes in zircon is made difficult by the isobaric interferences of 176Yb and 176Lu on 176Hf, the correction of which has been discussed in detail (e.g., Griffin et al., 2002; Woodhead et al., 2004; Iizuka and Hirata, 2005; Hawkesworth and Kemp, 2006; Gerdes and Zeh, 2009; Wu et al., 2006; Kemp et al., 2009). Properly correcting for 176Yb (and to a lesser degree 176Lu) is critical given that 176Yb/176Hf of typical zircons is commonly between 10% and 30% and can be as much as 70%. The ratio of stable isotopes 179Hf/177Hf is used for mass bias corrections, and an exponential mass bias function is used in all calculations. Interference-free 173Yb and 171Yb were monitored during the Hf analysis to calculate Yb mass bias (βYb) and the contribution of Yb to the measurement of 176(Hf + Lu + Yb). Because the magnitude of the Yb correction is so great, small inaccuracies in the Yb mass bias can lead to large analytical errors (Woodhead et al., 2004). Unlike 179Hf/177Hf, the precision of the 173Yb/171Yb measurement, and consequently the accuracy of βYb, is dependent upon the Yb signal intensity (Fig. 3). At 171Yb intensities of less than 0.015 V, it becomes very difficult to reliably estimate βYb, and for those analyses, Hf mass bias (βHf) was used to correct 176Yb/171Yb. Unfortunately, Chu et al. (2002) and Woodhead et al. (2004) have shown that Hf and Yb exhibit slightly different fractionation behavior, which we also observed to be true (Fig. 3C). So, although it is not ideal to use Hf fractionation factors to correct for Yb mass bias, low-Yb zircons require relatively minor correction, and as such it is possible to use βHf without introducing large errors to the corrected 176Hf/177Hf ratio. The scatter introduced by the interference correction is not included in the final error attached to 176Hf/177Hf values, but it is believed to be a relatively minor contribution to the quoted uncertainty. If there were a source of significant, unaccounted error, we would expect a given set of measurements to be overly dispersed. This is not the case, as evidenced by a mean square weighted deviation (MSWD) of less than one for all reported sample data (see GSA Supplementary Data1).

The Lu correction was done by monitoring 175Lu and using 176Lu/175Lu = 0.02653 (Patchett, 1983) and βYb, assuming that Lu behaves similarly to Yb. All corrections are performed on a line-by-line basis, and in all cases, Hf and Yb isotope data were normalized to 179Hf/177Hf = 0.72350 (Patchett and Tatsumoto, 1980) and 173Yb/171Yb = 1.132338 (Vervoort et al., 2004), respectively. A 176Lu decay constant of 1.876 × 10−11 (Scherer et al., 2001; Söderlund et al., 2004) was used in all calculations. Chondritic values of Bouvier et al. (2008) were adopted for the calculation of εHf values.

Hf Solution Analysis

Analyses of pure Hf solutions, as well as Hf solutions doped with variable amounts of Yb and Lu, were performed to test our ability to reliably correct for Yb and Lu interferences. Solution analyses were run in three blocks of 20 measurements, with additional background measurements being automatically performed between blocks. Backgrounds were measured using electrostatic analyzer deflection for 60 s at the start of the run, and measurements were integrated over 5 s. For 10 ppb solutions, total Hf beams of ∼5 V were achieved (this is the maximum possible with our 3 × 1011 Ω resistors). Solution data were collected during many analytical sessions over the course of this study, and Hf standard solution measurements were always made after instrument tuning and before acquisition of laser data. Repeated analysis of JMC 475 (n = 71) over the course of this study yielded a weighted mean of 176Hf/177Hf = 0.282159 ± 15, which was nearly identical to the accepted JMC 475 value of 0.282160 (Vervoort et al., 2004). No normalization of the data to JMC 475 was performed. Hafnium Spex solution, although not an ultrapure interlaboratory standard, was also analyzed (n = 50) and found to be isotopically the same as JMC 475 (Hf Spex 176Hf/177Hf = 0. 282159 ± 12). Hafnium Spex solutions doped with natural Yb and Lu produced corrected 176Hf/177Hf values similar to that of the pure solution, although scatter in the high (Yb + Lu)/Hf data was greater (Fig. 4). No statistical correlation was found between 176Hf/177Hf and 176(Lu + Yb)/176Hf, indicating that Yb and Lu interferences were adequately removed.

Hf Laser-Ablation Analysis

In situ Hf isotope data were acquired using a 40 μm beam centered directly on top of the pit previously excavated for U-Pb analysis. Laser run conditions were the same as those described for U-Pb geochronology. Under those conditions, total Hf beams ranged from 2 to 7 V for standard zircons. The in situ analytical routine began with a 40 s on-peak background measurement, followed by 60 s of laser ablation with a 1 s data integration time. This resulted in a laser pit that was ∼50 μm in depth, with ∼15 μm for the U-Pb analysis and 35 μm for the Hf analysis. All corrections were automatically calculated during the run on a line-by-line basis, and a 2σ filter was applied to each 60 measurement data block offline to remove outliers.

The zircon standards Mud Tank, Temora-2, FC-52 (compositionally similar to FC-1, from an anorthosite of the Duluth complex), 91500 (all described in Woodhead and Hergt, 2005), and Plesovice (Sláma et al., 2008) were analyzed. The results of repeated in situ analysis of these zircons during many analytical sessions over the course of roughly 6 mo are given in Figure 5 and Table 1. All zircon standards were added to each sample mount (3–4 samples per mount) and were analyzed between each set of unknowns in order to monitor laser stability and Hf ratio accuracy. In most cases, the long-term measured laser-ablation averages overlap (within error) the long-term solution values for those zircons, indicating that the previously described Lu and Yb interference correction method is also successful for laser analyses. A small discrepancy exists between the long-term 176Hf/177Hf laser average of FC-52 (0.282169 ± 10; 95% confidence; n = 74) and the long-term solution average of FC-1 (0.282184 ± 16; 2 standard error [S.E.]; n = 42) (Woodhead and Hergt, 2005). The cause of this difference is not clear, although it is likely not a function of the interference correction, as discussed later.

The five standards chosen have rare earth element (REE) concentrations ranging from REE-poor (Mud Tank) to REE-rich concentrations (Temora-2 and FC-52). Because of their high and variable REE content, Temora-2 and FC-52 are the most useful for testing the reliability of in situ Hf isotope data. Both have 176(Yb + Lu)/176Hf values that range from only a few percent to almost 50% (similar to the range that we observe in rocks from the Coast Mountains), and they are not correlated with 176Hf/177Hf. This provides further evidence that Yb and Lu are being properly subtracted from the laser 176Hf/177Hf calculations, as well as suggesting that any discrepancy in FC-52 Hf isotopic values is not related to REE interference correction. It is likely that the observed difference between the FC-52 long-term solution and laser averages is the result of isotopic heterogeneity in the Duluth anorthosite, given that the solution Hf data were generated from a different batch of zircon crystals than we analyzed.

U-Pb GEOCHRONOLOGY RESULTS

We present 29 new U-Pb zircon ages from two suites of plutonic rocks: (1) Jurassic–Eocene plutons of the central Coast Mountains batholith (n = 23), and (2) Ordovician–Early Devonian plutons clearly identified as part of the Alexander terrane in southeast Alaska (n = 6). Detailed geochronology data, including concordia and weighted mean plots, can be found in the GSA Data Repository (see footnote 1).

Coast Mountains batholith pluton ages range from ca. 151 to 53 Ma and represent nearly the entire time span of magmatism in the Coast Mountains. West of the Coast shear zone, ages decrease systematically from west to east, indicating eastward migration of magmatism at ∼1 km/m.y. across this area between 150 and 80 Ma. Jurassic to Early Cretaceous plutons are also found east of the Coast shear zone, which is ascribed to duplication of the Jurassic arc by sinistral displacement along strike-slip faults in the Early Cretaceous (Gehrels et al., 2009). Most samples from plutons east of the Coast shear zone record younger (Late Cretaceous to Eocene) ages and show no apparent migratory trends.

In addition to intrusive rocks of the central Coast Mountains batholith, Alexander terrane plutons from southeast Alaska were analyzed for the sake of comparing εHf values of the Alexander terrane with those of the plutons intruding Coast Mountains batholith crust of unknown affinity, as discussed in later sections. U-Pb ages of the Alexander plutons range from ca. 480 to 390 Ma and are consistent with the range of ages reported by Gehrels and Saleeby (1987).

Hf ISOTOPIC RESULTS

Hafnium isotopic compositions were measured in situ via LA-MC-ICP-MS directly on top of the spot previously excavated for U-Pb analysis, such that each Hf isotopic measurement is directly tied to a corresponding U-Pb age. Hf data, and related ages, are reported in Table 2 for the 29 plutonic samples discussed in the previous section. For each sample, between 15 and 55 individual spot measurements were made, and mean values are reported. Individual measurements and weighted mean plots of all Hf sample data can be found in the GSA Data Repository (see footnote 1).

Measurements from a given sample were highly reproducible, and uncertainties associated with the precision and scatter of a set of analyses are low (≤1 unit of εHf at the 2σ level). Measured 176Hf/177Hf values were corrected for the radiogenic in-growth of 176Hf, although that correction is small, given that zircon incorporates relatively little Lu (measured 176Lu/177Hf values range from 0.0005 to 0.0015). Corrections for the isobaric interference of 176Lu and 176Yb ranged from minor (∼5% change in the 176Hf/177Hf ratio) to large (∼35% change in the 176Hf/177Hf ratio). Good reproducibility of corrected 176Hf/177Hf values, however, inspires confidence that even major changes in isotope ratios were accurately accounted for. For example, sample 80JA11 yielded zircons that have up to 70% of their total mass 176 contributed from Yb and Lu, and yet the reproducibility of corrected 176Hf/177Hf values from those sample grains was excellent (Fig. 6). Plots showing reproducibility of the corrected 176Hf/177Hf as a function of Yb and Lu interference can be found in the GSA Data Repository (see footnote 1).

Initial εHf values from plutonic rocks of the Coast Mountains batholith range widely from +1 to +13, and values cluster between +9 and +13 for Paleozoic plutonic rocks of the Alaskan Alexander terrane. A general increase in εHf(t) is observed from west to east across the central Coast Mountains batholith, although nearly the entire range of Hf values is found in rocks located within and along the periphery of the Coast shear zone (see Fig. 2; Table 2).

INTERPRETATION OF U-Pb-Hf DATA AND IMPLICATIONS FOR THE CRUSTAL ARCHITECTURE OF THE COAST MOUNTAINS BATHOLITH

Hafnium isotopic signatures from intrusive rocks across the west-central Coast Mountains batholith are relatively juvenile, suggesting derivation of Coast Mountains batholith plutons from similarly juvenile mantle or crustal sources. In this respect, the Hf data presented here are consistent with the notion that the Coast Mountains batholith represents the growth of new crust in a continental arc system (e.g., Samson et al., 1989; Friedman et al., 1995). However, the range of measured εHf(t) values is great (+1 to +13), and in all cases, those values are lower than the depleted mantle array (εHf[500–0 Ma] values of +14 to +18; Vervoort and Blichert-Toft, 1999), indicating heterogeneity in magma source regions and/or interaction with more evolved crustal materials. Although absent in most samples, traces of inherited zircon were present in three of the plutons analyzed in this study, also indicating the incorporation of older, recycled crust into melts.

Because a considerable amount of age control exists for this portion of the Coast Mountains batholith, our data can be used to evaluate relations between petrogenesis and magmatic flux. Magmatism in the Coast Mountains batholith is interpreted to be strongly episodic, with distinct flare-up events at 160–140 Ma, 120–78 Ma, and 55–48 Ma, and a long-lived period of relative magmatic inactivity between 140 and 120 Ma (Gehrels et al., 2009). Our geochronologic data do not reflect that periodicity because we intentionally chose samples that were either known or inferred to have ages corresponding to both high- and low-flux events. Isotope pull-downs, or negative excursions in whole-rock initial εNd values, have been temporally correlated with magmatic flare-ups (Ducea and Barton, 2007; DeCelles et al., 2009), suggesting a link between isotopic signatures and periods of lithospheric thickening and crustal melt production. Our data show no clear negative excursions or any correlation between εHf(t) and U-Pb age and/or the timing of magmatic flux events (Fig. 7). This is probably attributable to the relatively small (n = 23) size of the Coast Mountains batholith U-Pb-Hf data set presented here. The range of isotopic values presented here likely records normal variation in the background arc–mantle wedge flux (Ducea and Barton, 2007), but it is nonetheless significant because these variations can be used as tracers of input from distinctive sources.

Linking Hf Plutonic Signatures and Country Rock Assemblages

A fundamental observation of our data is that εHf(t) increases from west to east; the lowest values were obtained for plutons along the western coast, and the highest values were obtained from plutons in the easternmost, inland areas (Fig. 2). Based on the distribution of Hf signatures, we discriminate between three distinct crustal domains into which plutons of the Coast Mountains batholith were emplaced: a western domain, characterized by relatively evolved Hf isotopic compositions, a central domain, with intermediate Hf compositions, and an eastern domain, characterized by juvenile Hf signatures. Because these variations coincide with variations in country rock assemblages, as well as previously reported Nd-Sr and detrital zircon data, we interpret those domains to be part of the Banks Island assemblage, the Alexander terrane, and the Stikine terrane, respectively (Fig. 8).

Samples from the Banks Island domain have the most evolved εHf(t) values (+2.6 to +5.2), which are consistent with evolved continental Nd-Sr values from metamorphic country rocks of the west-central Coast Mountains batholith (Boghossian and Gehrels, 2000). Epsilon Nd values of those rocks are more negative (∼+0.5 to −9.9) than our reported zircon Hf values for younger intrusive rocks of the same area. This appears to be typical of Coast Mountains batholith plutons, which likely originated from juvenile mantle, but which are sensitive to small amounts of more evolved crustal input, such that recorded εHf(t) values represent mixing between arc-type melts and older, preexisting crust. Old continental crust has high Hf concentrations and highly negative εHf. For example, Middle Proterozoic crust of the western United States has modern εHf values ranging from −10 to −30; Early Proterozoic and Archean crust is even more negative (Vervoort and Patchett, 1996). Only small additions (a few percent) of isotopically depleted continental material would therefore be necessary to drive down the εHf of magmas, as has been pointed out for Nd-Sr systematics (e.g., Patchett and Bridgwater, 1984; Samson et al., 1990, 1991a).

The central domain, which is the southern continuation of the Alexander terrane, consists of a north-south–trending belt of rocks located between the Banks Island domain to the west and the Gravina belt and mid-Cretaceous thrust system to the east. Plutons making up this belt have εHf(t) values that range from +5.9 and +8.1, and they are more juvenile than those of the Banks Island terrane (Fig. 8). They are interpreted to belong to the Alexander terrane because of: (1) the occurrence of distinctive Ordovician–Silurian magmatic arc assemblage overlain by Devonian conglomeratic strata; (2) the juvenile nature of the Alexander terrane understood from Nd-Sr isotopes (Samson et al., 1989), and the lack of continental input in geologic units and U-Pb zircon populations (Gehrels and Saleeby, 1987; Gehrels et al., 1987, 1996); and (3) geographic position; the crust into which this part of the batholith was built forms the southern continuation of the main Alexander terrane to the north (see Fig. 1).

To test the assignment of these rocks to the Alexander terrane, we analyzed Hf isotopes from igneous rocks that are known to belong to the terrane in southern SE Alaska. These Paleozoic (ca. 480–390 Ma) intrusive rocks have juvenile εHf(t) values ranging from +8.6 to +12.9. The εHf values of the same plutons at 100 Ma, an age which approximates those of Alexander plutons in the central Coast Mountains batholith, yield highly consistent values of +2.5 to +7. Hafnium signatures of Cretaceous plutons intruding the inferred Alexander terrane (+5.9 to +8.1) are therefore likely recording either (1) direct melting of Paleozoic Alexander basement; or (2) melting of the mantle wedge followed by partial melting and assimilation of Alexander and/or Banks Island assemblages. It is also possible that the interpreted Alexander terrane of British Columbia represents a transitional zone between primitive Alexander of southeast Alaska and the more evolved Banks Island assemblage to the south. Connections between the northern Alexander terrane and Banks Island are consistent with the presence of ca. 480–410 Ma detrital zircons in Banks Island strata (Gehrels and Boghossian, 2000). Furthermore, Hf isotopic signatures within plutons of the southeast Alaskan Alexander terrane become more evolved to the southwest, suggesting a gradational relationship between the Alexander and Banks Island terranes (Fig. 8).

The eastern domain is characterized by juvenile εHf(t) values ranging from +10.2 to +15.1, and it is interpreted as Stikine terrane based on proximity with Stikine to the east, lack of a major tectonic boundary separating the two, the recognition by Hill (1985) that the central gneiss complex (eastern domain) contains fossiliferous marbles potentially correlative with Stikine strata, and primitive Nd-Sr values reported for Stikine rocks by Samson et al. (1989). There is no evidence for the input of older, reworked continental crust in any Stikine rocks in the study area, suggesting that, much like in the case of the Alexander terrane, they were produced either by the wholesale melting of juvenile Stikine lower crust or by direct melting of the mantle with variable contributions from terrane components.

Between the outboard Banks Island and Alexander terranes and the inboard Stikine terrane, there is a zone of structural deformation delineated by a regionally extensive belt of mid-Cretaceous thrust faults (e.g., Rubin et al., 1990) and by the Coast shear zone (e.g., Rusmore et al., 2005; Hollister and Andronicos, 2006), which is characterized by marked heterogeneity in Hf signatures (Fig. 8). Initial εHf values within this zone range from +1.5 to +11.6, and these are interpreted to represent the imbrication of juvenile Alexander and Stikine terranes with continental-margin rocks of the Yukon-Tanana terrane. This is consistent with the geologic evidence for protracted and large-scale displacement along these structures.

It is important to note that there is very little interaction of plutons with country rocks at the present level of exposure, which represents emplacement depths of ∼10–25 km (Butler et al., 2001, 2006). Furthermore, new geochemical data from the Coast Mountains batholith indicate that plutons were generated and emplaced at depths greater than 35 km (Girardi et al., 2008). The fact that the Hf isotopic signatures of the plutons appear to be tracking with the country rock assemblages therefore raises interesting questions about the nature of the terranes at depth and the structural boundaries that separate them. Lateral changes in Hf isotopes imply marked heterogeneity in the magma source regions; the correlation between Hf signatures and the known isotopic character of country rocks indicates the presence of those or similar assemblages at depth. The assertion that presently exposed metamorphic rocks could be present at melt-generation depths is corroborated by high δ18O values in quartz from the same plutons (Wetmore and Ducea, 2011).

Relationships between Crustal Terranes of the Coast Mountains Batholith

Discernible shifts in Hf isotopic values in plutons are interpreted to reflect boundaries between discrete crustal terranes (Fig. 9). This implies that structures controlling those boundaries act as through-going crustal barriers restricting magma migration and/or that the crustal boundaries have remained steeply dipping through time. For example, the thrusting and crustal thickening of mid-Cretaceous age along a regional belt in the axial Coast Mountains batholith (Rubin et al., 1990) have not affected the magma source regions in adjacent parts of the batholith. This is evidenced in pluton ages in the eastern domain, which range from 151 to 55 Ma, but which have no associated change in Hf behavior. This is also observed in individual samples with zircon inheritance. For example, sample MT05-135, an Eocene pluton located on the eastern periphery of the Coast shear zone, has inherited Jurassic and Early Cretaceous cores and magmatic rims with measured ages of 55 Ma (Fig. 10). There is very little change, however, in εHf(t) values across those zircon domains, despite the fact that a major period of deformation and shortening between 101 and 85 Ma occurred in the zone immediately adjacent to where this pluton was intruded (Rubin et al., 1990) (Fig. 10). Although thrusting has imbricated various terranes along the axial belt, the zone of imbrication is apparently thick skinned and restricted to a narrow region (∼20-km-wide swath). Thrusting may have been thin skinned and low angle along the margins of the belt at shallow crustal levels, e.g., above the present 10–25 km levels of exposure.

Many of the structural and/or stratigraphic relationships between the crustal terranes discussed here are unclear because the original contacts between those terranes have been commonly overprinted by more recent deformation. It remains unclear, for example, why the Banks Island continental margin–type assemblage is located outboard of the primitive Alexander terrane. Relationships among Alexander, Stikine, and Yukon-Tanana terranes are likewise enigmatic. East of the Coast shear zone, plutonic rocks record uniformly juvenile, Stikine-like εHf(t) signatures. This is different than relationships inferred to the north, where isotopically evolved rocks of the Yukon-Tanana terrane and associated metamorphic assemblages are observed wedged between primitive Alexander and Stikine terranes (Samson et al., 1991a). There appears to be an along-strike change, therefore, in the tectonostratigraphic relationship between the Yukon-Tanana terrane and Stikine terrane. One notable exception to this comes from an Eocene pluton (53 Ma) located at the eastern margin of outcropping batholithic rocks, which has an εHf(t) value of +2.0, i.e., distinctly more evolved than those from any other portion of the Coast Mountains batholith. It is possible that this signature reflects a component of evolved continental-margin strata of the Yukon-Tanana terrane that extends beneath the western Stikine terrane. Although the nature of the links between Stikine and Yukon-Tanana is unclear, they have been proposed on the basis of inherited Precambrian zircons in Jurassic Stikine plutons (Thorkelson et al., 1995), and evolved Nd isotopic characteristics of Upper Triassic Stikine strata (Jackson et al., 1991).

CONCLUSIONS

Plutons making up the west-central Coast Mountains batholith represent ∼100 m.y. of continental arc magmatism. In general, the juvenile character of Hf isotopic data from those plutonic rocks suggests the production of new continental crust derived primarily from mantle sources, with little recycling of Precambrian continental crust into arc-type melts. Substantial variation in εHf(t) values of plutons (+1 to +13), and the systematic spatial distribution of those values, however, suggests that Hf isotopes are tracing heterogeneities in source regions and that those heterogeneities are a function of the influence of different crustal terranes. From outboard to inboard, discrete crustal panels appear to be composed of the Banks Island assemblage, the Alexander terrane, and the Stikine terrane, with the imbrication of thin fragments of Yukon-Tanana terrane along mid-Cretaceous thrust faults and along the Coast shear zone, and the possibility of Yukon-Tanana terrane basement beneath Stikine strata in the easternmost part of the study area.

The juxtaposition of these crustal terranes requires complicated structural and/or stratigraphic relationships between the various terranes, particularly in the case of the Yukon-Tanana and Stikine terranes. If the Yukon-Tanana terrane to the north was indeed emplaced outboard of the Stikine terrane during a transpressive regime, as has been suggested by Samson et al. (1991a), the inferred structural imbrication of the Yukon-Tanana terrane near the Coast shear zone in our study area could perhaps represent the pinching out of the Yukon-Tanana terrane to the south. Structural emplacement of Yukon-Tanana terrane rocks outboard of the Stikine terrane along Cretaceous thrusts and left-lateral faults could explain an enigmatic quartzite cobble conglomerate located in the south-central part of our study area, west of the Coast shear zone, which, unlike other local lithologies, has Archean detrital zircon populations (Boghossian and Gehrels, 2000; Gehrels and Boghossian, 2000). The presence of an isotopically evolved pluton to the east, however, suggests a potential stratigraphic tie between the Stikine and the Yukon-Tanana terranes. That relationship remains cryptic due to the uniformity of juvenile Hf signatures in plutonic rocks presented here and juvenile Nd-Sr signatures of Stikine country rocks (Samson et al., 1989). Further isotopic work to the east and to the south of the study could help resolve the extent of Yukon-Tanana terrane influence in Coast Mountains batholith plutons.

This work was sponsored by National Science Foundation (NSF) award EAR-0309885 for support of the BATHOLITHS projects, and by EAR-0732436 for support of the Arizona LaserChron Center. The authors thank two anonymous reviewers, whose thoughtful and critical comments greatly improved the manuscript.

1GSA Data Repository Item 2011234, Weighted mean and concordia plots for all U-Pb and Hf data presented, is available at www.geosociety.org/pubs/ft2011.htm, or on request from editing@geosociety.org, Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140, USA.