Sheeted intrusive complexes represent unequivocal examples of incremental intrusive activity in plutonic systems and thus provide an opportunity to constrain processes associated with incremental magma emplacement. We present field observations, U-Pb zircon data, and whole-rock and mineral chemistry from the sheeted intrusive complex of the ca. 92 Ma Tenpeak pluton located in the North Cascades, Washington. Samples collected from an ∼750-m transect of the sheeted complex, including at least 58 individual sheets, provide an opportunity to understand the time scales, magma sources, and magmatic processes responsible for the generation of incrementally emplaced magma bodies. High-precision chemical abrasion-thermal ionization mass spectrometry (CA-TIMS) geochronology show relatively rapid construction of the complex over an interval of 94,000 ± 62,500 yr (95% confidence). With a volume estimate of ∼17 km3, this is equivalent to a magma emplacement rate of 1.1–5.4 × 10–4 km3 yr–1. In general, thinner sheets are found closer to the pluton margin and transition to thicker sheets in the interior. Although magmas of the sheeted complex show compositional and mineralogical similarities with the nearby voluminous Schaefer Lake tonalite phase of the Tenpeak pluton, U-Pb zircon ages from individual sheets suggest that formation of the sheeted complex occurred as a waning phase of the Schaefer Lake emplacement. Individual sheets also show relatively simple zircon populations consistent with low overall magma volumes and relatively rapid cooling following emplacement. We suggest that the sheeted complex resulted from localization of magma intrusions along anisotropies in the shear zone. In addition, the thermal boundary formed with the adjacent meta-supracrustal wall rocks likely facilitated rapid cooling of sheets and localized subsequent intrusive events. The formation of the range of sheet compositions requires at least three different parental magmas. The observed range of sheet compositions can be produced by mixing between a single mafic parental magma (SiO2 ∼50 wt%) and an array of felsic magma compositions (SiO2 from 59 to 67 wt%). Mineral populations within mafic samples suggest that the felsic parental magmas either had low crystallinity prior to mixing and/or substantial crystallization occurred after mixing. Textural and field evidence, along with mineral chemistry, suggest that final mixing between the mafic parent and felsic array occurred late, almost immediately prior to emplacement.
There is widespread consensus that many large plutons are constructed incrementally over time scales of several million years (e.g., McNulty et al., 1996; Matthews et al., 1999; Miller and Paterson, 2001a; Coleman et al., 2004; Hildreth, 2004; Žák and Paterson, 2005; Matzel et al., 2006; Lipman, 2007; Paterson et al., 2011; Shea et al., 2016). This idea is supported by a variety of field and geochronologic evidence and challenges earlier assumptions that plutons were once large, mobile magma bodies (e.g., Buddington, 1959). Incremental emplacement has important consequences regarding the thermal budget, intrusion rates, and the crustal response to magma emplacement. However, in larger plutons the evidence for incremental processes is also often cryptic, relying on high-precision geochronology and subtle field relations (e.g., Coleman et al., 2004; Glazner et al., 2004; Matzel et al., 2006; Miller et al., 2009; Miller et al., 2011a). As a result, there remains uncertainty regarding the rate of emplacement relative to the thermal history of the pluton, the frequency and volumes of individual increments, the petrological relationship and degree of consanguinity between incrementally emplaced magmas, and the ultimate origin of the magmas involved.
One plutonic environment that preserves unambiguous evidence for incremental emplacement of magmas and that allows some of these issues to be investigated in more detail is sheeted intrusive complexes (e.g., Pitcher and Berger, 1972; Hutton, 1992; McNulty et al., 1996; Paterson and Miller, 1998; Brown and McClelland, 2000; Lafrance and John, 2001; Miller and Paterson, 2001a; Mahan et al., 2003; Pons et al., 2006; Walker et al., 2007; Barbey et al., 2008; Žák et al., 2009). These consist of multiple tabular or sheet-shaped magmatic bodies, generally emplaced subparallel to each other, and with individual sheets that are typically lithologically distinct and separated by well-defined igneous contacts. Sheeted complexes are documented in a range of different plutonic contexts, and most commonly occur along margins of granitoid plutons, although they may also occur within pluton interiors and along internal contacts. Detailed examination of sheeted complexes offers several advantages for elucidating the processes of incremental emplacement and pluton formation. Sheeted complexes can preserve individual magma batches that may be obscured by mixing and other processes within the more homogeneous facies of plutons that lack discrete internal contacts, and can thus be used to (1) constrain field and contact relationships between successively emplaced magma batches, (2) probe the petrological and mineralogical relationships between different magmas that contribute to a pluton, and (3) establish the time scales required for a sheeted complex to form, and the timing relationship between sheeted complexes and more homogeneous plutonic facies.
Despite these advantages, the processes that lead to the formation of sheeted intrusions remain controversial, and a wide range of models have been proposed in the literature to explain the formation of magmatic sheets and sheeted complexes. These include emplacement into active shear zones or regional faults (e.g., Hutton, 1992); formation during the early stages of pluton emplacement prior to establishment of a localized magma chamber (Hanson and Glazner, 1995; Glazner et al., 2004); magma wedging, i.e., emplacement parallel to host-rock foliation or other sheets associated with displacement of the host rock (Miller and Paterson, 2001a); emplacement as layered intrusions associated with sinking of a pluton floor (Wiebe et al., 2004); or intrusion controlled by fractures or viscosity variations associated with differences in crystallinity (Bergantz, 2000; Žák et al., 2009; Miller et al., 2009).
To date there have also been relatively few detailed studies of complexes within the large tonalite-granodiorite batholiths that occur in crustal arc settings (e.g., Wiebe et al., 2004; Barbey et al., 2008; Pupier et al., 2008), and even fewer studies looking at sheeted complexes in deep-crustal sections (e.g., Miller et al., 2009). To help rectify this problem and focus on the processes that produce sheeted intrusive complexes in deep-crustal settings, we report a detailed field, petrographic, geochemical (whole-rock and mineral geochemistry), and high-precision U-Pb geochronology for a sheeted complex within the deep-crustal (700–980 MPa) Tenpeak pluton of the North Cascades (Washington State, USA) (Fig. 1). The pluton lacks strong metamorphic and structural overprinting as a whole, is well exposed and well studied, both in terms of field mapping, chemistry, and geochronology (e.g., Crowder et al., 1966; Cater and Crowder, 1967; Cater, 1982; Dawes, 1993; Miller and Paterson, 1999; Matzel et al., 2006, Miller et al., 2009). As a result, the Tenpeak pluton provides an important opportunity to study an incrementally built sheeted intrusive complex within the deep crust.
GEOLOGIC SETTING OF THE TENPEAK PLUTON
The 92.3–89.7 Ma Tenpeak pluton is part of the Cretaceous–Paleogene continental magmatic arc of the North Cascades, which is the southern extension of the >1200-km-long Coast Belt. Arc magmatism developed in the Late Cretaceous (ca. 96 Ma to 72 Ma) during southwest-northeast contraction and crustal thickening of the Coast Belt during subduction (e.g., Rubin et al., 1990; McGroder, 1991). These arc plutons were emplaced into amphibolite facies, previously accreted oceanic and arc terranes (Tabor et al., 1989) at a range of paleodepths (<5 km to ≥35 km; e.g., Brown and Walker, 1993; Dawes, 1993; Miller and Paterson, 2001b; Miller et al., 2009). Most of the deep-crustal plutons intruded into the Napeequa unit, consisting of amphibolite, quartzite, marble, metaperidotite, and biotite schist (Cater and Crowder, 1967; Miller et al., 2009), at ≥25 km depth; this is based on the presence of magmatic epidote (>600 MPa) (Zen, 1988) and Al-in-hornblende barometry (Brown and Walker, 1993; Dawes, 1993; Miller et al., 2000). These intrusions range in age from 96 to 84 Ma and form a variety of shapes, from thinly sheeted to broadly elliptical plutons that alternate between sheets and relatively homogeneous, irregularly shaped bodies (Miller et al., 2009).
Some of the deepest intrusions in the North Cascades, including the Tenpeak pluton, were emplaced into a 10-km-wide, orogen-parallel belt between the White River shear zone (WRSZ) and Entiat fault, a Paleogene high-angle fault located east of the map area. The WRSZ is between the Napeequa unit and the Chiwaukum Schist (Fig. 1), a metapelitic and metapsammitic schist with some amphibolite, metaperidotite, and marble (Plummer, 1980; Tabor et al., 1989). This northeast-dipping reverse shear zone is defined by mylonitic schists, protomylonitic Tenpeak tonalites, and the rotation of lineations to downdip orientations (VanDiver, 1967; Tabor et al., 1987; Magloughlin, 1993; Miller et al., 2003, 2006; Raszewski and Magloughlin, 2004). The latest activity on the shear zone occurred after 91 Ma (Umhoefer and Miller, 1996) and is characterized by synkinematic, retrograde greenschist facies metamorphism in the schist (VanDiver, 1967; Tabor et al., 1987; Magloughlin, 1993; Miller et al., 2006). Along the south and southwest margin of the pluton, there is a structural aureole with downward ductile flow of host rocks that partially overlaps with the shear zone (Miller and Paterson, 1999). The early history of the shear zone is less certain. It has been interpreted as a pre-Tenpeak terrane boundary, which may have been reactivated during pluton emplacement (Umhoefer and Miller, 1996), or an inverted unconformity (Brown and Dragovich, 2003) with some lower temperature deformation following intrusion (Raszewski and Magloughlin, 2004; Miller et al., 2009).
The probable host-rock conditions during magma emplacement in the Tenpeak pluton are constrained by geobarometry and geothermometry of the Napeequa unit and Chiwaukum Schist. Peak conditions calculated for the Napeequa unit (640–740 °C and 860–980 MPa) and Chiwaukum Schist (540–700 °C and 700–890 MPa) are based on phase equilibria thermometers (e.g., Whitney et al., 1999; Valley et al., 2003; Miller et al., 2009). These pressures are consistent with barometric estimates of the Tenpeak pluton as described here.
The Tenpeak pluton is dominated by tonalite in the interior, and is surrounded by a narrow, discontinuous, heterogeneous zone of mingled and sheeted gabbro, tonalite, and hornblendite (e.g., Cater, 1982; Dawes, 1993; Miller and Paterson, 1999; Matzel et al., 2006). There are also many xenoliths (<5 cm to km scale) of Napeequa amphibolite, schist, and metaperidotite along the margins of the pluton, along internal contacts, and within the internally sheeted complexes (Miller et al., 2009). In map view, the pluton is broadly elliptical (aspect ratio of ∼5:1) with a northwest-trending protrusion, known as the White Mountain lobe, located to the northeast (Fig. 1). The Tenpeak pluton has a nearly vertical northeast contact but moderately to steeply inward-dipping southwest and south boundaries (Miller et al., 2009). The contact of the pluton outlines an asymmetrical wedge-shaped body that was extrapolated to a vertical extent of ≥9.5 km (Miller et al., 2009).
Geochemistry and Geochronology of the Tenpeak Intrusion
The Tenpeak intrusion can be subdivided into several units based on age, lithology, composition, and structure. The units are the (1) ca. 92.1 Ma mafic complex, (2) 92.3–91.8 Ma Schaefer Lake tonalite and associated sheeted complex (the subject of this study), (3) 91.3 Ma tonalitic gneiss, (4) 90.6 Ma northern mafic tonalite, (5) undated interlayered unit consisting of tonalite and rafts of the Napeequa unit, and (6) voluminous 89.7 Ma Indian Creek tonalite (Matzel et al., 2006). Previous thermal ionization mass spectrometry (TIMS) U-Pb zircon dating of the Tenpeak intrusive units indicate an intrusion period of ca. 92.3–89.7 Ma (Matzel et al., 2006). The geochronology also suggests that some Tenpeak units may not have sustained any sizeable magma reservoir with a melt-rich zone, or if one did exist, it was most likely ephemeral (Matzel et al., 2006).
The dominant tonalite in the interior of the pluton can be subdivided geochemically into the Schaefer Lake and Indian Creek subtypes (Miller et al., 2000). The Schaefer Lake compositional subtype is characterized by high MgO, Ni, and Cr concentrations, and the Indian Creek subtype is composed of isotopically evolved, calc-alkaline rocks (Miller et al., 2011b). Both subtypes are dominated by hornblende-biotite tonalite with titanite, magmatic epidote, and in some cases garnet. In contrast to the Indian Creek tonalite, the Schaefer Lake tonalite contains higher proportions of biotite relative to hornblende (Cater, 1982; Dawes, 1993). The focus of this study is the sheeted magmatic complex of the Schaefer Lake unit (Fig. 1), which we refer to as the sheeted complex herein.
A representative suite of 17 samples from the sheeted complex was selected for whole-rock major and trace element analyses (Supplemental Tables S1 and S21). Most samples originated from an ∼750 m measured transect with almost continuous exposure (Fig. 2; Supplemental Fig. S12; Table S3 [footnote 1]). Samples were analyzed at the Washington State University GeoAnalytical Laboratory using a Thermo Scientific ARL ADVANT’XP sequential X-ray fluorescence (XRF) spectrometer and Agilent 7700 inductively coupled plasma–mass spectrometer (ICP-MS). A detailed description of sample preparation and XRF analysis following a single low-dilution Li-tetraborate fused bead procedure was described in Johnson et al. (1999). Samples analyzed using ICP-MS follow procedures detailed in Lichte et al. (1987), Jarvis (1988), and Longerich et al. (1990). The precision for rare earth elements (REEs) and other trace elements are 5% and 10%, respectively. In cases where we have elements analyzed by both XRF and ICP-MS, the comparison between the techniques is generally good except for some elements of low abundance, which have lower accuracy with XRF (e.g., U, Nb). We have used ICP-MS data when available in plots and calculations.
In Situ Mineral Analysis
Plagioclase and amphibole from the sheeted complex (Supplemental Tables S4 and S5 [footnote 1]) were analyzed using electron microprobe analysis (EMPA) and laser ablation (LA) ICP-MS at Oregon State University. Major oxides (SiO2, TiO2, Al2O3, FeOT, MnO, MgO, CaO, Na2O, K2O), a few minor oxides (Cr2O3, SO2, P2O5), and select volatiles (F, Cl) were analyzed using a Cameca SX-100 electron microprobe with parameters summarized in Supplemental Table S6 (footnote 1). The analytical error (based on replicate analyses) is generally <4%, with the exception of TiO2 (15%) and K2O (28%) for plagioclase analyses. Errors associated with amphibole analyses are generally <3%, with the exception of TiO2 (13%). All EMPA analyses have a 2σ of < 0.8 wt%. On average, three to four analysis points were taken within each crystal from rim-to-core to capture major compositional changes. EMPA analyses with unacceptable totals (<98% or >102%) or inaccurate stoichiometry were discarded.
Trace element mineral analysis (Supplemental Tables S4 and S5 [footnote 1]) were measured with LA-ICP-MS using a Photon Machines Analyte G2 193 nm ArF fast Excimer Laser system and Thermo Scientific XSERIES 2 quadrupole ICP-MS systems. Instrument details, analysis conditions, and calibration standard uncertainties are summarized in Supplemental Table S6. The calibration and secondary standards were GSE-1G and BHVO-2G, respectively. The internal standard was 43Ca and elemental concentrations were calculated using CaO content measured by EMPA analysis. Accuracy (assessed by replicate analyses of BHVO-2G) is generally <13%, with the exception of Y, Nb, and Ta (<∼20%). LA-ICP-MS analysis locations were picked on or adjacent to those analyzed on the EMPA. Analyses were processed using an in-house LaserTRAM program, developed at Oregon State University, using Visual Basic in Microsoft Excel. The main functions of the program are summarized in Loewen and Kent (2012). During LaserTRAM processing, anomalous data that showed evidence of analysis of mineral inclusions and other instances where normalized intensities were not constant throughout analysis were eliminated.
Separation of heavy minerals for zircon dating followed standard crushing, magnetic separation, and heavy liquid techniques. Zircon crystals were picked in ethanol under a binocular microscope and were selected based on their morphology, color, clarity, and lack of inclusions. Zircon grains selected for U-Pb analysis were placed in quartz beakers and annealed in a muffle furnace at 900 ± 20 °C for 60 h. When possible, individual crystals were imaged using cathodoluminescence (CL) prior to dating. In some samples, the zircon were too small (<60 μm) to be both imaged and dated. In those cases, representative crystals were selected for imaging. The zircon were mounted in epoxy and polished to approximately half their original thickness. CL imaging of crystal mounts was obtained using the JEOL 733 Superprobe electron microprobe at the Massachusetts Institute of Technology (MIT). Image analysis was carried out with a 15 kV accelerating voltage and 10–30 nA beam current. All zircon images are grouped by sample number and are presented in Supplemental Fig. S2 (footnote 2). Imaged zircon crystals selected for analysis were plucked from epoxy mounts using fine-point tweezers.
All zircons selected for analysis were chemically abraded using a modified version of the technique of Mattinson (2005). Individual zircon grains were loaded into separate 300 μL Teflon FEP microcapsules, placed in a Parr vessel, and leached in ∼120 μL of 29 M HF for 12 h at 220 °C. Following the leach step, crystals were fluxed in HNO3 for 30 min and then sonicated for 45 min. After this step, crystals were rinsed 2 times in ultrapure water and fluxed in 6 M HCl for 30 min and sonicated for 45 min. Crystals were each rinsed 2 more times in ultrapure water and then loaded into individual microcapsules with ∼120 μL of 29 M HF and a mixed 205Pb-233U-235U spike (ET535). The zircon were dissolved at 220 °C for 48 h, dried to salts, and redissolved in ∼120 μL of 6 M HCl at 180 °C for at least 12 h. Using HCl-based anion exchange columns, modified from Krogh (1973), Pb and U were separated from each sample.
Both Pb and U were analyzed by TIMS on the MIT VG Sector 54 multicollector mass spectrometer or the MIT Isotopx X62 multicollector mass spectrometer. Both Pb and U were loaded together onto degassed single zone refined Re filaments with a silica gel–H3PO4 mixture (Gerstenberger and Haase, 1997). Lead was measured by peak-hopping on a single Daly detector. Uranium was measured in static Faraday mode. Isotope ratios of U and Pb were corrected for mass fractionation during analysis using the ET535 tracer solution. Data acquisition and reduction were conducted using the Tripoli and U-Pb Redux software packages (Bowring et al., 2011; McLean et al., 2011).
U-Pb data from individual zircon grains are reported at the 95% confidence levels in Table 1. U-Pb zircon data from these samples typically yield clusters of dates on or near concordia. Final solidification ages, determined by taking the weighted mean of the youngest dates within a sample that overlap within uncertainty, are interpreted using the 206Pb/238U dates, which give the highest possible precision for rocks of this general age range. The 206Pb/238U dates are calculated with the decay constants of Jaffey et al. (1971) and the present-day 238U/235U ratio recommended by Hiess et al. (2012); errors were calculated using the algorithms of McLean et al. (2011). All dates are corrected for initial 230Th disequilibrium using a whole-rock Th/U ratio of 2.9 ± 0.15, calculated from the average whole-rock Th/U from these samples. Altering this value by 50% changes the 206Pb/238U date by <10 k.y. We are primarily concerned with duration of magmatism; the errors of the weighted mean dates are internal errors based on analytical uncertainties only, including counting statistics, spike subtraction, and blank subtraction. When comparing our dates with those from another laboratory that uses a different spike, it is necessary to include tracer uncertainty. Comparing our dates with those from a different decay system would also require including both tracer and decay constant uncertainties.
Field Description of the Sheeted Complex
A section through the sheeted complex was mapped along a single well-exposed drainage that extends approximately normal to the WRSZ and to the dominant strike of individual sheets, supplemented with observations elsewhere (Fig. 1). Magmatic foliations within the complex typically strike to the northwest and dip moderately to steeply toward the northeast, mimicking the overall foliation trends seen elsewhere in the Tenpeak pluton. Along the drainage, the mean foliation strike was 316° with a dip of 36°NE (Fig. S3 [footnote 2]). However, in other areas of the sheeted complex, the foliation can be steeper (e.g., 60°–89°, Miller et al., 2009). A variety of dikes crosscut the sheeted complex, including strongly foliated, fine-grained dikes and some similarly strongly foliated rocks that could be rafts of Napeequa metavolcanic rocks, muscovite-bearing felsic dikes, hybridized dikes, and other igneous dikes (Fig. S3 [footnote 2]).
The results of our field observations are summarized in a lithostratigraphic section (Fig. 2). A detailed version of the lithostratigraphic section, with field descriptions, is presented in Supplemental Figure S1 (footnote 2) and Table S3 (footnote 1). This cross section covers a structural thickness of ∼750 m, with distance from the base of the transect, which we propose is close to the margin, assuming a northeast contact, with the WRSZ. The actual contact between the sheeted complex and Chiwaukum Schist is obscured by the WRSZ, so although we did not start mapping from the contact, we believe that we were close to the contact. We also did not reach the upper contact between the sheeted complex and Schaefer Lake tonalite phase, but the homogeneity and thickness of the upper sheets suggest that we were most likely near the gradational contact.
Intrusive sheets (49–66 wt% SiO2) were documented and categorized in the field based on color index (CI), grain size, mineralogy, internal features (e.g., enclaves) and the nature of sheet contacts (Fig. 3). The CI (ranging from ∼20–85) and mineralogy were used to divide the sheets into low-CI (≤49) and high-CI (≥50) groups (Fig. 2). Similar sheet groupings are also reflected in the whole-rock chemistry, which is discussed further herein. Several sheets show considerable internal variations in CI, and we labeled these hybrid sheets (Figs. 2 and 3A). A comparison between CI estimated in the field, CI determined from point count modal data (Table 2), and whole-rock SiO2 shows that CI values estimated in the field are a reliable tool for characterizing different sheets, thus giving us an efficient way to constrain the lithology and composition of individual sheets. Strong correlations exist between the estimated CI from field observations and the true CI calculated from modal data of our limited sample suite (R2 ∼ 0.84, p < 0.01). The correlation between field-estimated CI and composition is less strong (e.g., SiO2, R2 ∼ 0.68, p < 0.01), but is still significant at the 99% level, and the overall root mean square error suggests that CI in the field can be used to estimate silica content within at least ∼5 wt% SiO2. Thus our distinctions based on CI are broadly equivalent to SiO2 values: ≥∼59 wt% (low CI) and ≤∼58 wt% (high CI).
The composition and thickness of individual intrusive sheets vary throughout the lithostratigraphic section, and our cross section allows us to document and quantify relationships between sheet composition, thickness, and distance from the contact (Fig. 4). In the measured section, we documented 58 individual sheets. Sheets range in thickness from ∼15 cm to several hundred meters (∼302 m), although most have thicknesses of 1–10 m. In general, although with considerable scatter, thinner sheets occur near the base of the section and then transition to thicker sheets higher in the lithostratigraphic section (Figs. 2 and 4A). The thickest sheet documented is also the farthest from the pluton margin, located at the top of the section. For all intrusive sheets, the correlation between thickness and distance from the section bottom is significant at the 99% confidence level (p < 0.01; Fig. 4A), and persists at this high confidence level even if the single largest sheet is disregarded. There is also no clear relationship between composition and location in the measured section; sheets of low-CI and high-CI compositions appear throughout the first 300 m of the measured section (Fig. 4B), although there is a suggestion that higher CI sheets occur closer to the contact with the WRSZ. Both high- and low-CI rocks occur as thin sheets, many examples of both having thicknesses of <1 m (Fig. 4C).
Contact Styles and the Internal Structure of Sheets
A range of different contact styles was observed between sheets, and there is also a range of internal igneous features preserved within individual sheets. Typically, sheet contacts were parallel to each other and to the inferred orientation of the pluton margin throughout the section. However, significant deviations occurred locally (e.g., interfingering sheets, Fig. 3B). Overall, contacts range from sharply intrusive to gradational over distances of several centimeters (3–5 cm, Fig. 3C). There are some instances where the contact appears irregular and interdigitated or where the contact between individual sheets is offset by an apparent magmatic fault (Fig. 3D). We classify magmatic faults as those that show no brittle deformation or increased recrystallization, and where the apparent fault plane appears injected with melt. For all sheets, there was no indication of quenching or chilling at contacts. At the contact of some high- and low-CI sheets, we observed irregular, plagioclase-rich zones that were traceable for several centimeters into the adjacent high-CI sheets (Figs. 3A, 3E), which we interpret as crystal exchange between two sheets. Crystal exchange suggests that both sheets were sufficiently crystal poor to be mobile at the time of intrusion. In most cases, the orientation of crystals in these zones is at high angles to the magmatic foliation of the adjacent sheets. Within the interior of several high-CI sheets, we also saw irregular aggregates of plagioclase phenocrysts (Fig. 3F).
Enclaves are also pervasive in sheets, and vary in composition from hornblendite to quartz diorite. Commonly, there are high-CI enclaves within low-CI sheets. Most of these enclaves are highly elongate (average aspect ratio ∼6:1), and these are generally oriented with the long dimension parallel or subparallel to sheet contacts and magmatic foliation. We observed multiple distinct high-CI enclave populations that range from medium to coarse grained and exhibit textural and compositional differences within a single intrusive sheet. Some enclaves also have wispy edges (Fig. 3C). Evidence of crystal exchange, similar to that seen between sheet contacts, is also present at the margin of some enclaves (Figs. 3A, 3E). Xenoliths of deformed meta-supracrustal rocks of the Napeequa unit, the host lithology for most of the Tenpeak pluton, also occur throughout the section.
Several sheets have more complex internal textures and compositions. We refer to these as hybrid (Fig. 3A). Hybrid sheets occur predominantly toward the middle of the measured section, and consist of a broad range of compositionally heterogeneous and banded lithologies within an individual sheet. Although the external contacts of the hybrid sheets are similar to those for other sheets described here, hybrid sheets are internally heterogeneous, with a range of CIs (e.g., samples AKCP-112AF and –112AM). Within hybrid sheets, different CI magmas are juxtaposed to form heterogeneous banding and discrete, laterally discontinuous layers. These layers range in thickness from 1 to 10 cm, and often occur with numerous enclaves. Banded sheets also show internal contacts that range from discrete to diffuse.
The mineralogy of sheets within the complex is dominated by plagioclase and amphibole, with lesser biotite, quartz, titanite, zircon, and epidote (Table 2), in approximately decreasing order of modal abundance. Because alkali feldspars are rare to absent in these rocks, we refer to all feldspars as plagioclase herein. Based on modal mineralogy, determined by point counts (Table 2), the composition of individual sheets ranges from gabbro to tonalite with localized regions of hornblende-rich diorite. Lower CI tonalites are dominated by plagioclase and quartz, and tend to have lower amphibole/biotite ratios. In contrast, higher CI tonalites contain a higher overall percentage of amphibole and higher amphibole/biotite ratios. Solid-state deformation varies throughout, but is observed to be more prevalent in the low-CI sheets that are structurally higher in section. Deformation and alteration are evidenced by recrystallized quartz aggregates, bent and chloritized biotite, and corroded rims of major mineral phases.
Plagioclase is the largest and most abundant phase in the felsic samples. The grains are generally subhedral to euhedral. Crystal size varies, but is typically 1–4 mm in length and 0.75–3 mm in width. Albite and polysynthetic twinning are common in most of the samples. Optical zoning is visible in most crystals, ranging from oscillatory to patchy and concentric. The width of individual zones varies from 1 to 4 µm. Some plagioclase crystals have antiperthitic textures. The antiperthitic lamellae are 2–5 µm long and 1 µm wide. Disequilibrium textures are very common for plagioclase grains with rounded and tattered crystal edges. These phenocrysts typically contain many mineral inclusions, in particular zircon and apatite; some also exhibit sieve texture and resorbed rims (Figs. 5A, 5B).
Amphibole is typically subhedral to anhedral, although there are a few sheets with euhedral amphibole grains. In general, amphiboles from the felsic samples are more euhedral and less altered than those from the high-CI sheets. Many of the rims of amphibole appear to be removed or partially resorbed. Some samples show multiple populations of amphibole: (1) equant crystals (30 µm by 20 µm), (2) large and prismatic crystals (50 µm by 12 µm), and (3) smaller crystals (10 µm by 5 µm). The last group appears to mantle larger amphibole phenocrysts. Amphiboles from the high-CI sheets occasionally form glomerocrysts with embayed rims and intersecting grains (Fig. 5C).
Biotite is generally euhedral to subhedral, although there are some that have been chloritized. Some grains enclose plagioclase and quartz grains (Figs. 5B, 5D). Secondary intergrown biotite grains are also observed. The grain size is typically 10–50 µm in length and 5–15 µm in width.
Most of the quartz is recrystallized and tends to be anhedral. Quartz typically appears as subgrains and/or recrystallized aggregates that are found interstitially between larger phenocrysts and is commonly associated with biotite grains (Figs. 5A, 5B).
Accessory Minerals (Apatite, Zircon, Titanite, Epidote)
The accessory minerals represent <∼3% of the modal percentage. Apatite and zircon are typically found as inclusions in plagioclase, and titanite is observed adjacent to amphibole and biotite. These minerals are usually euhedral to slightly subhedral. Most of the zircon show no evidence of xenocrystic cores or younger rims (for more detail, see discussion of U-Pb zircon geochronology). One sample (AKCP-17B) is especially enriched in titanite grains, most of which are subhedral and do not appear as inclusions. Magmatic epidote, which is common in the Tenpeak intrusion (Cater, 1982; Zen and Hammarstrom, 1984), was identified in several samples, appears subhedral to euhedral, and is associated with biotite, quartz, and hornblende.
A 1-m-thick hornblende-rich quartz diorite sheet (AKCP-121A) intruded into a 45-m-thick hybrid sheet located near the top of the lithostratigraphic section was also sampled. Although mineralogically similar to other sheets, this sample is significantly finer grained than the rest of the sheeted complex. The major phases include plagioclase (35%), amphibole (58%), biotite (4%), and quartz (1%). In thin section, there appears to be higher degrees of crystal intergrowth and less distinct boundaries between individual phases. In some instances, amphibole phenocrysts mantle plagioclase grains.
We selected a subset of samples for further geochemical study, including whole-rock compositions and mineral chemistry. These provide representative selections of the major lithologies present in the sheeted complex. Samples include a range of CI between 20 and 85, representative low- and high-CI portions of a single hybrid sheet (AKCP-112AF and AKCP-112AM), a single high-CI enclave (AKCP-111) from a low-CI sheet, and a single hornblende-rich, fine-grained diorite sheet (AKCP-121A).
Intrusive Sheet Chemistry
The sheeted complex is composed of metaluminous compositions that range from gabbro to tonalite (Fig. 6), but it is dominantly tonalite. Based on trends evident in bivariate diagrams versus SiO2 (Figs. 7 and 8), we divide our sheets into two different compositional groupings. One set of samples, which we designate felsic, have SiO2 contents ≥∼59 wt% and form well-defined linear arrays in bivariate plots of SiO2 versus major and trace elements, with decreasing FeO*, TiO2, Al2O3, P2O5, Na2O, Ni, Cr, Sr, Sc and increasing K2O, Ba, Pb, and REEs (Figs. 7 and 8; note that REEs are not shown) as SiO2 increases. In contrast, samples we designate as mafic have SiO2 between 50 and 58 wt%. While they also show decreasing TiO2, FeO*, Ni, Cr, and Sc with increasing SiO2, the mafic samples also show more complex trends for Al2O3, Na2O, P2O5, and Sr, with trends generally increasing with SiO2, but also showing significant scatter. For these latter elements, mafic and felsic samples together define bent or humped trends with inflections at ∼60 wt% SiO2. In addition, Al2O3 and Na2O in mafic samples define fan-shaped arrays, with greater variability at higher SiO2. Note that our subdivision of samples into mafic and felsic also broadly mirrors the field-based subdivision into low- and high-CI samples, with the low-CI (≤49) equivalent to felsic samples and high-CI (≥50) equivalent to mafic ones.
Two samples from the hybrid sheet were analyzed: (1) low-CI sample (AKCP-112AF) that plots along the felsic sample array and we associate with the felsic group, and (2) high-CI sample from the hybrid sheet (AKCP-112AM) that plots within the mafic samples and we consider part of the mafic group. The high-CI enclave sample (AKCP-111) also plots within the mafic sample group, although it has notably higher K2O and Ba and light REE enrichment (Figs. 7F, 8E, and 9A). The hornblende-rich quartz-diorite sheet (AKCP-121A) is characterized by low SiO2 (49.13 wt%), but the highest FeO (10.55 wt%) and TiO2 (1.44 wt%, Fig. 7) present in the sheeted complex samples. AKCP-121A is also distinct, with the lowest Ni, Cr, and Sr and highest Sc, Ga, and Y concentrations (Fig. 8).
Most of the samples from the sheeted complex show moderately steep REE patterns [Fig. 9A; (La/Yb)N = ∼5–11], with the exception of two mafic samples (AKCP-1 and AKCP-112AM) that both exhibit flatter REE patterns, and (La/Yb)N ratios of 3.66 and 4.03, respectively. All of the samples show a slight scoop-shaped heavy REE depletion. The mafic samples are more depleted in middle to heavy REEs. One felsic sample (AKCP-17B) is anomalous, with a highly fractionated REE pattern [(La/Yb)N = 42.33]. This sample also shows an unusually high abundance of titanite, which is likely to explain the steepened REE patterns. Ratios of La/Yb increase while Dy/Yb remains broadly constant with SiO2. The felsic sheets show a range of La concentrations (43–210 ppm). No significant europium anomalies (Eu/Eu*, where the subscript N denotes values normalized to chondrite following Sun and McDonough ), appear in these rocks (Eu/Eu* = 1.0–1.2). On mid-oceanic ridge basalt–normalized multielement diagrams, following Pearce and Parkinson (1993), all of the Tenpeak samples show enrichment in large ion lithophile elements (e.g., Rb and Ba; Fig. 9B) and most show minor depletions in Y and Yb.
Major and trace element compositions were determined in plagioclase and amphibole in a subset of samples. Within most of these individual crystals, chemical zoning is absent or difficult to identify due to loss or alteration of grain rims during deformation. Typically core and non-core analyses are similar, although in some samples there are subtle core-rim variations.
Overall, plagioclase phenocrysts show a range of anorthite (An) compositions (An5 to An60); the majority are between An20 and An40 (Fig. 10A). Some crystals are normally zoned with more albitic compositions toward the rim relative to the core. Other grains show little difference between rim and core compositions, possibly as a result of alteration or recrystallization of crystal rims during deformation.
Plagioclase compositions are broadly similar regardless of the host sample chemistry (Fig. 10A), although there is also a tendency for plagioclase from the most mafic samples to have slightly higher overall An contents. The diorite sheet is an exception, with the lowest whole-rock SiO2 content (∼49 wt%) and plagioclase with distinctly lower anorthite compositions (Fig. 10A). Strontium concentrations mirror trends in An content with slightly higher Sr in mafic samples relative to felsic host samples, and significantly lower Sr in the dioritic plagioclase (Fig. 10B). Incompatible elements, such as Ba, show an inverse relationship, with lower Ba in plagioclase hosted in mafic samples relative to those from felsic samples. Barium concentrations in plagioclase from the diorite are comparable to those from mafic samples (Fig. 10C). There are strong similarities in An, Sr, and to a lesser degree Ba among plagioclase from the different felsic host samples, while those from mafic hosts show more variation in compositions.
Following the nomenclature of Leake et al. (1997), as modified by Ridolfi et al. (2010), amphiboles from the sheeted complex range from magnesiohornblende to tschermakite (Fig. 11A). Both mafic and felsic samples are dominated by magnesiohornblende amphibole, in contrast to the diorite amphiboles that are almost exclusively tschermakite. The Mg# [Mg# = molar Mg/(Mg + Fe) considering all Fe as FeO] of amphibole (0.60–0.71, average 0.66) in mafic samples are broadly similar to amphibole in felsic samples (0.55–0.69), the latter having a slightly lower average of 0.61. In the hybrid sheet, amphiboles show no real distinction between mafic and felsic samples (Mg# 0.61–0.68). The diorite amphibole (Mg# 0.55–0.65) are broadly similar to the felsic amphibole though the diorites have the lowest Mg# average (0.59).
Amphibole form compositionally distinct groups based on major and trace element chemistry (Figs. 11 and 12). Compositional differences are apparent between amphibole from mafic and felsic host samples, and between individual mafic and hybrid samples. This is clearly shown by molar Al/Si and Na2O compared to Mg# (Figs. 11B, 11C). Amphibole from the most mafic samples (AKCP-1) have higher Na2O and Al2O3 at a given Mg# (Figs. 11B, 11C). These amphibole grains are also enriched in Li and Cu but depleted in Nb, Sc, Zn, Y, and REEs (Fig. 12) at a given Mg#. Amphibole from the other mafic and hybrid samples (AKCP-112AF, AKCP-112AM, AKCP-115AF, AKCP-115AM, and AKCP-119) are defined by intermediate Al2O3 and Na2O values, and slightly higher La, Dy, and Zn relative to the first group. Amphibole from felsic samples have lower Al2O3 and Na2O, and those from diorite samples have distinctly higher overall Al2O3 and Na2O than amphibole from other samples (Figs. 11B, 11C).
These differences are also broadly reflected in abundances of REEs and other trace elements (Fig. 13A). Europium anomalies of amphibole also vary between sheets and appear to vary consistently with host-rock SiO2 composition (Fig. 13B). Europium anomalies in amphibole become progressively lower with increasing whole-rock SiO2 contents while mafic samples have Eu/Eu* values that are closer to 1.
U-Pb Zircon Geochronology
Nine samples were dated from the sheeted complex and are grouped here on the basis of magma composition. Results are summarized in Figure 14 and presented in Table 1. Details about individual samples, along with zircon images, are presented in Supplemental Figure S2 (footnote 2).
Modern U-Pb TIMS geochronology is capable of producing very precise dates from single zircon crystals. However, it is important to consider what zircon ages from a single sample actually represent. In the simplest case, if magma crystallizes all of the zircon at the same time (within uncertainty), it would be expected that zircon dates from this magma would overlap within analytical uncertainties and give a weighted mean with an MSWD (mean square of weighted deviates) of ∼1.0. However, the distribution of dates from a single sample often cannot be explained solely by analytical uncertainty; in these cases, weighted means exhibit considerable overdispersion (MSWD >> 1). This overdispersion could have several sources, including protracted zircon crystallization during slow cooling of plutonic samples, inheritance from host or source rocks, magma mixing, or a combination of these factors. In addition, some analyses may also represent rare instances where chemical abrasion did not completely remove all parts of the grain that underwent Pb loss.
The best interpretation for zircon data sets with MSWD >> 1 is a matter of debate (e.g., Miller et al., 2007). For this study we took an approach tailored to constraining the final age of zircon crystallization in each sample. To calculate an age for a given zircon population, we included all zircon dates that overlap within analytical uncertainty of the youngest zircon in a given sample (Table 1). The weighted mean of these dates is then used to estimate the timing of final solidification of the sample. Although this approach results in the exclusion of some zircon in our final calculated ages, we stress that if we disregard obviously inherited grains (e.g., AKCP-124; Fig. 14), making changes to which zircon grains are included in our weighted mean ages will not meaningfully affect the calculated ages or our interpretations.
Felsic Magma Sheets
Four low-CI sheets from the sheeted complex were selected for zircon geochronology, AKCP-2, CCE7B, CCE8D, and AKCP-124. In addition, we dated AKCP-114, which was taken from a low-CI ladder dike-like feature in a high-CI sheet (Fig. 3G). CL images show zircon crystals with homogeneous cores and oscillatory zoned rims (Supplemental Fig. S2 [footnote 2]). U-Pb TIMS analyses of several microsampled (physically separated tips and cores) zircon grains without obvious cores suggest that most rims and cores from these simple crystals do not have any statistically significant difference in age. The range of weighted mean ages for the low-CI sheets is between 91.929 ± 0.023 and 91.835 ± 0.030 Ma, with single-crystal dates ranging to 92.21 ± 0.14 Ma (Fig. 14; Table 1).
Mafic Magma Sheets
Two high-CI sheets were dated at the base of the lithostratigraphic section close to the interpreted pluton margin, CC1 and AKCP-1. CL images show generally elongate crystals without evidence of xenocrystic cores or younger rims (Supplemental Fig. S2 [footnote 2]). Some crystals have oscillatory zoned rims surrounding light colored cores. The range of weighted mean ages for the high-CI sheets is 91.906 ± 0.023–91.900 ± 0.041 Ma, with single-crystal dates ranging to 92.144 ± 0.085 Ma (Fig. 14; Table 1).
Dioritic Magma Sheet (AKCP-121A)
AKCP-121A is from a hornblende-rich diorite sheet, which appears to intrude into a 45-m-thick hybrid sheet near the top of the section. CL images of zircon grains from AKCP-121A exhibit oscillatory zoned rims surrounding light colored cores (Supplemental Fig. S2 [footnote 2]). Some U-Pb analyses of zircon from this sample show evidence for an inherited component; individual analyses range in age from 90.580 ± 0.15 to 90.277 ± 0.088 Ma. A weighted mean of the three youngest analyses (Fig. 14; Table 1) gives an age of 90.261 ± 0.038 Ma (n = 3, MSWD = 0.38).
Fine-Grained Tonalite Dike (CCE8A)
Sample CCE8A came from a fine-grained, ∼1-m-thick tonalite dike that crosscuts an ∼80-m-thick, plagioclase-porphyritic high-CI sheet. Minor assimilation of host rock can be observed at the margins of the dike. Oscillatory zoning is common in zircon from this sample and some crystals are bordered by thin rims (Supplemental Fig. S2 [footnote 2]). Analysis of 10 crystals gave a range of dates between 90.470 ± 0.11 and 90.277 ± 0.056 Ma. A weighted mean of the 5 youngest analyses (Fig. 14; Table 1) gives an age of 90.303 ± 0.030 Ma (n = 5, MSWD = 1.1).
Timing and Duration of Sheeted Complex Formation
The U-Pb zircon geochronology is consistent with the formation of the sheeted complex ca. 91.9 Ma. This age is close to, but younger than, ages from the adjacent Schaefer Lake tonalite (92.27 ± 0.09–92.07 ± 0.04 Ma; Figs. 1 and 15; Matzel et al., 2006). Matzel et al. (2006) recognized the similarity between the sheeted complex and the Schaefer Lake tonalite and mapped the former as a subunit of the Schaefer Lake unit of the Tenpeak pluton. The data of Matzel et al. (2006) also suggest that the intrusion of the Schaefer Lake occurred at broadly the same time as the White Mountain lobe farther to the north (92.22 ± 0.17 Ma; Matzel et al., 2006; Fig. 1). Our geochronology suggests that the sheeted complex may have formed during the waning stages or immediately after emplacement of the Schaefer Lake tonalite. It is unlikely that the sheeted complex represents a separate intrusion from the voluminous Schaefer Lake because of their compositional similarities and gradational contact. In the Tenpeak pluton, there is a notable hiatus after the Schaefer Lake phase and sheeted complex, until emplacement of the Indian Creek tonalite phase ca. 89 Ma (Fig. 15; Matzel et al., 2006). Excluding samples AKCP-121A (the fine-grained diorite) and CCE8A (a crosscutting tonalite dike), the majority of samples from the sheeted complex have indistinguishable ages within uncertainty (Fig. 14). A single sample (AKCP-2) collected from a low-CI sheet near the base of the section has a slightly younger age than the rest of the sheets, even in contrast to other sheets (AKCP-1 and CC-1) sampled near the contact.
To estimate the total time required for the formation of the sheeted complex, we have compared the ages of the oldest and youngest individual sheets (AKCP-114 and AKCP-2). The age difference between these is 94,000 yr. Using the standard deviations calculated from the individual zircon crystals in these samples (excluding zircon ages not used in the calculation of the age) and the appropriate t-statistic with 95% confidence interval for the total duration of formation of the sheeted complex is estimated to be 94,000 ± 62,500 yr. From the lithostratigraphic section, we observed a minimum of 58 distinct sheets (a minimum given gaps in the section that did not have outcrops). Over a period of 94,000 ± 62,500 yr, this corresponds to intrusion of a sheet at least every ∼500–2700 yr (on average).
It is also possible to calculate the minimum rate of intrusion of the sheeted complex using minimum volume estimates and our high-precision geochronology. Matzel et al. (2006) estimated a minimum volume for the sheeted complex of 20 km3, which is based on a total thickness of 1.5 km for the complex and vertical contacts; the latter is unlikely to be valid for the southern portion of the pluton. Our revised volume estimate for the sheeted complex is ∼17 km3, which assumes shallower contacts (∼36°, the average for sheet contacts, Fig. S3 [see footnote 2]), and a topographic thickness of at least 750 m. We assumed that the sheeted zone extended downward to a depth of 1.5 km, the approximate topographic relief of the pluton as a whole, and used its length in map view (∼8 km). The minimum intrusion rate, using the 17 km3 volume estimate, over 94,000 ± 62,500 yr is 1.1–5.4 × 10–4 km3 yr–1, with a median value of ∼1.8 × 10–4 km3 yr–1. Note, however, that the base of the sheeted zone to the southwest is not exposed in our transect, and the intrusive rate may be as much as twice as large.
Our calculated intrusion rate is similar to that calculated for the entire Tenpeak pluton assuming a minimum volume of 394 km3 (1.5 × 10–4 km3 yr–1 overall; Matzel et al., 2006). Even if the true volume of the sheeted complex is twice as large as our estimate, the intrusion rate would still be of the same magnitude to the entire Tenpeak pluton. In addition, it overlaps with that estimated for the ca. 96 Ma Mount Stuart batholith, a mid-crustal to shallow (6–12 km) pluton located ∼20 km south of the Tenpeak that has a minimum total volume of 1200 km3 (2.2–30 × 10–4 km3 yr–1 overall; Matzel et al., 2006), and the ca. 90 Ma Black Peak pluton (1.1–11 × 10–4 km3 yr–1; Shea et al., 2016), a shallow (3–10 km) pluton located ∼50 km to the north that has a minimum total volume of 365 km3. There are similarities between the observed intrusion rates for the sheeted complex and those estimated for larger composite plutons in the North Cascades, and with other continental batholiths (e.g., Miller et al., 2009; de Saint Blanquat et al., 2011). This suggests that construction of sheeted complexes (relative to a more homogeneous phase of the Tenpeak pluton) may not be simply related to gross differences in magma intrusion rates, although there are considerable uncertainties (e.g., Fowler, 1994; Miller and Paterson, 2001a; Walker et al., 2007; Žák et al., 2009).
Compositional Evolution and Diversity of Intrusive Sheets
We observe a variety of chemical trends in sheet compositions when major and trace element abundances are compared to SiO2 (Figs. 7 and 8). Elements that are compatible during fractionation of mafic and intermediate calc-alkaline magmas (e.g., TiO2, FeO*, Ni, Cr, and Sc) show decreasing contents across the range of SiO2 contents; this is evident in both mafic and felsic samples. Incompatible elements (K2O, Ba, and Pb) increase with SiO2 for both mafic and felsic samples. In addition, for some elements, mafic and felsic samples appear to be broadly colinear (e.g., TiO2), although in other cases there are subtle offsets between the trends (e.g., Cr, Ni, and Pb). Mafic and felsic magmas show distinctly different trends for some elements (Al2O3, Sr, P2O5, and Na2O). The clearest example of this is Al2O3 (Fig. 7D) where felsic samples form a well-defined negative trend with SiO2, whereas mafic samples (including the one mafic enclave analyzed) form a broadly fan-shaped array.
From studies of plutonic rocks, it is clear that a variety of processes could influence the compositions of individual sheets. In the following discussion we consider the potential roles of fractional crystallization, magma mixing, and compaction and liquid loss. The variations evident in major and trace element abundances defined by mafic and felsic samples resemble, at least superficially, those that are expected from fractional crystallization of a basaltic magma. In particular, inflections in Harker diagrams are often considered to reflect saturation of specific phases during ongoing crystal fractionation, with Al, Na and Sr (plagioclase) and P (apatite) commonly showing such behavior in crystallizing basaltic liquids (e.g., Wade et al., 2005; Lee and Bachmann, 2014). Despite this, a simple model where the full range of sheet compositions observed represents the liquid line of descent of a mafic parental magma appears unlikely. Evidence for this includes the following.
1.Instead of defining a single trend, Al2O3, Na2O, and to some extent P2O5 show fan-shaped arrays in mafic samples, suggesting greater variability of these elements at relatively high SiO2 contents. This is inconsistent with mafic samples representing a single liquid line of descent (Fig. 16).
2.The variations evident in sheet compositions are a poor match for experimental studies of liquids produced by progressive crystallization of a hydrous basalt parent at pressures similar to or higher than the emplacement pressures (∼700–900 MPa; see discussion herein; Figs. 16A–16C). Specifically, Al2O3 and Na2O contents in experimental liquids increase rapidly at low SiO2 from initial crystallization of an olivine + clinopyroxene assemblage. This increase is a common feature in experiments at pressures >∼500 MPa, despite differences in starting composition and experimental design (e.g., Blatter et al., 2013; Nandedkar et al., 2014). Both Al2O3 and Na2O decrease when plagioclase becomes a liquidus phase, and this occurs at lower SiO2 contents (∼50–52 wt%) than the inflection evident in our sheeted complex compositions (∼60 wt% SiO2).
3.The volume of low-CI sheets observed in the field is a poor match with expected proportions of felsic magma compositions produced from crystal fractionation. Experiments show that SiO2 contents in excess of ∼60 wt%, the lowest SiO2 in felsic sheets, require crystallization of 70%–80% of a hydrous basalt. Thus if the sheet compositions are produced by fractionation and emplaced into the sheeted complex in the same proportions at which they are produced, felsic magmas should be volumetrically minor with respect to the more mafic sample compositions. However, if we assume that sheet thickness in our profile is roughly proportional to sheet volume, then the volume of the sheeted complex is dominated by sheets with intermediate SiO2 of 60 ± 5 wt% (primarily because of one very thick sheet that occurs in the upper parts of the cross section). Sheets with more mafic or felsic compositions, <55 wt% and >65 wt%, respectively, appear to have broadly similar volumes, and compose subequal (∼20%) proportions of the observed section (Fig. 2).
4.Trace element variations in our samples also appear inconsistent with fractional crystallization of a single basaltic parental magma. For example, bivariate plots of compatible versus incompatible elements (e.g., Ni versus Ba; Fig. 17) show broadly linear trends for both mafic and felsic samples, rather than the curved trends expected for fractional crystallization.
We suggest that a more complex set of petrogenetic processes is required to explain the variation in sheet compositions, and we specifically see an important role for magma mixing in the generation of intrusive sheet compositions. One way to produce the range of mafic sample compositions is magma mixing involving at least three magmatic end members. For all elements, mafic samples primarily are within a triangle defined by the most mafic composition (AKCP-1) and the extent of the array defined by felsic samples (e.g., Fig. 16A). Thus, the range of mafic compositions we observe could be produced via mixing between a single mafic parent with a composition close to AKCP-1, the most mafic sheet composition, and the full range of felsic compositions. If this is the case, then the composition of each mafic sample is determined both by the composition of the felsic magma involved in mixing and the proportions of mafic and felsic parent, resulting in broadly fan-shaped compositional arrays evident in mafic compositions in Al2O3, P2O5, Na2O, and Sr (Figs. 7, 8, and 16). Modeling of this mixing process suggests that mafic magmas are the result of the mafic parental melt mixing with as much as ∼50% of magma from the felsic array (Fig. 16A).
Felsic samples define relatively tight linear arrays in major and trace element bivariate plots (e.g., Figs. 7 and 8), even when comparing elements with contrasting compatibility, such as Ni and Ba (Fig. 17). This observation is also consistent with magma mixing between two felsic parental end members with SiO2 contents of ∼60 and >67 wt% to produce the observed range of felsic compositions. The linear array observed between Ba and Ni is also strongly suggestive of mixing, because the considerable differences in partition coefficients for these elements should result in strongly curved trends from fractional crystallization (Fig. 17). Note that this model does not rule out fractional crystallization processes in producing the high-SiO2 end member for the felsic sample array from the lowest SiO2 felsic magma, and calculations suggest that it is possible to produce the highest SiO2 felsic magma compositions from the lowest SiO2 felsic compositions (i.e., AKCP-112AF or AKCP-115A) by removal of ∼40–50 wt% of an assemblage similar to that observed in the lowest SiO2 felsic sample. However, our data suggest that the ranges of felsic sample compositions are not being produced by progressive fractional crystallization.
In magmas or mushes with significant proportions of liquid present, removal of residual liquid during progressive crystallization and compaction can also result in substantial changes in bulk rock composition (e.g., McKenzie, 2011). In deep-crustal intrusions such as the Tenpeak pluton, emplaced magmas may retain a liquid fraction for long periods, potentially allowing for compaction and liquid loss to occur. The trends produced by such a process could be complex, with variable amounts of a liquid of different compositions removed from each sheet during progressive crystallization and compaction (e.g., Tegner et al., 2009; McKenzie, 2011). One method to test the overall importance of compaction is to compare modal mineral abundances and trace elements that are sensitive to loss of liquid. In Figure 16D we show Eu/Eu* ratios for both mafic and felsic samples compared to modal plagioclase in each sheet. If compaction was important, we would expect that Eu/Eu* and Sr contents to correlate positively with modal plagioclase, as their abundance would primarily be controlled by accumulation of plagioclase as liquid is lost, such that higher Eu/Eu* would occur in magmas that have retained greater proportions of plagioclase. However, no significant correlations are evident between modal plagioclase contents and Eu/Eu* ratios (Fig. 16D) or Sr contents (not shown). Modal amphibole contents also do not correlate significantly with most incompatible trace element abundances. From these observations we infer that if compaction and liquid loss occurred, they were not likely major processes in controlling compositional variations in individual sheets.
Our suggestion that magma mixing plays a major role in producing the compositional variations evident between different sheets requires that there were at least three distinct parental magmas in the genesis of the sheeted complex, one mafic and two felsic magma compositions (although the two felsic magmas may ultimately have been related by crystal fractionation). The presence of multiple magma types during formation of the sheeted complex is consistent with the understanding that magma genesis within the mantle and lower crust of subduction systems is capable of producing a range of magma compositions (e.g., Hildreth and Moorbath, 1988; Dufek and Bergantz, 2005; Annen et al., 2006). These magmas form via melting within the mantle wedge, followed by assimilation, and mixing and fractional crystallization with the lower crust. The lowest SiO2 mafic sample (e.g., AKCP-1) has compositions that overlap with those of primitive basaltic melts produced within modern subduction systems, such as the Cascade arc, and thus may represent magmas that are close in composition to those produced by melting within the mantle wedge during subduction. Moreover, at a given SiO2 content, mafic compositions have MgO contents that are at the high end of the range evident in modern Cascade volcanic rocks, consistent with the relatively high MgO contents of other North Cascade plutons (e.g., Mount Stuart; Anderson, 1992). Such high MgO may reflect the relatively mafic nature of the crustal protolith in this area, or may also result from processes occurring within the mantle wedge, such as greater contributions from hydrous melting of depleted harzburgite (Grove et al., 2005, Ruscitto et al., 2011).
Felsic magma compositions may also be produced in MASH-like (mixing, assimilation, storage, and hybridization) zones within the lower crust during extended episodes of magma injection, crystal fractionation, and hybridization with wall-rock lithologies (Hildreth and Moorbath, 1988; DeBari et al., 1998; Dufek and Bergantz, 2005; Annen et al., 2006). DeBari et al. (1998) suggested that the felsic end member of deep-crustal plutons in the North Cascades core, including the Tenpeak pluton, could be derived from ∼20% melting of a mafic garnet-amphibolite crustal source at 15–16 kbar. The felsic samples exhibit chemical trends broadly consistent with melting of garnet-bearing lithologies, such as high Sr/Y, La/Yb, and Sm/Yb ratios, and lower heavy REEs (Fig. 9A), which contrasts from the mafic parental end member.
Field Evidence for Magma Mixing
Our field observations also support the importance of magma mixing in generating intrusive sheet compositions, and provide some evidence for the timing of mixing processes. Although the bulk compositions of hybrid sheets are broadly intermediate, they are highly heterogeneous at the local scale, exhibiting macroscopic banding and a range of discrete and diffuse internal contact styles between magmas of contrasting CI (Fig. 3). Individual bands within a hybrid sheet are laterally discontinuous, and typically contain abundant enclaves. We have not made an exhaustive study of their geochemistry, but the two samples we report from one hybrid sheet (e.g., AKCP-112AF and AKCP-112AM, selected to represent the lowest and highest CI material available from this sheet, respectively) suggest that hybrid sheets represent incompletely mixed magmas. The hybrid samples appear to define a mixing line between a composition at the low SiO2 end of our felsic magma array, and the inferred mafic parental magma (Figs. 7 and 8; hybrid samples are indicated by a dash through the symbol). If this is the case then it also implies that at least some magma mixing between our mafic and felsic parental magmas may have occurred relatively late, almost immediately preceding magma emplacement, or at the site of emplacement. Macroscopic features such as the abundance of enclaves (Fig. 3) and evidence of crystal transfer, based on textural observations, between sheets are indicative of some magma interactions following emplacement and for the presence of a range of parental magma types. In addition, the existence of hybrid sheets and textural variations within individual sheets (e.g., Figs. 3A, 3F) suggests that at least some magma interactions occurred as an ongoing process during emplacement of individual sheets.
Mineral Chemistry of Mafic and Felsic Sheets
The major and trace element compositions of plagioclase and amphibole from selected sheets are shown in Figures 10–13. Both amphibole and plagioclase show considerable diversity in composition with respect to both major and trace element chemistry, with differences in composition evident within and between individual sheets.
The dominance of magma mixing processes for the sheeted complex inferred from whole-rock compositions suggests that we might be able to recognize different mineral populations within individual sheets that are derived from the different parental magmas involved. This situation is common in volcanic rocks produced by magma mixing where mineral compositions commonly reflect input from multiple parental magma sources (e.g., Eichelberger, 1978; Ginibre et al., 2002; Cooper and Reid, 2003; Kent et al., 2010; Kent, 2014).
In our samples from the sheeted complex we see differences in mineral compositions between sheets that are in our designated mafic and felsic sample groups (e.g., Figs. 10–13). For plagioclase these differences are relatively subtle; however, plagioclase from mafic samples extends to slightly higher An contents than plagioclase from felsic samples (with considerably overlap) and trend to lower Ba and higher Sr and Sr/Ba ratios (e.g., Fig. 10). The compositional differences in amphibole are more distinct, with amphibole from mafic samples having markedly higher Al/Si, Al2O3, Na2O, and lower MnO at a given Mg# (e.g., Fig. 11). Mafic sample amphibole also show trace element differences compared with those from felsic samples with higher REEs, Zn, Sc, Pb, Ba, and Y, and lower Eu/Eu*, Zr, and Li (e.g., Figs. 12 and 13).
One critical observation for all samples is that we do not see clear evidence for unique compositional populations of plagioclase and amphibole that might represent minerals derived from distinctly different parental magmas. In particular, in the mafic samples, we do not see evidence for contributions from crystals formed in felsic magmas, even though the whole-rock chemical compositions and field relations suggest that many sheets are the product of mixing between a single mafic parent and a range of felsic magma compositions. This is acutely clear in terms of amphibole compositions, where amphibole hosted in mafic and felsic samples have distinctly different compositions, and within a single sheet there are no bimodal crystal populations as might be expected from mixing processes (Figs. 11–13).
We propose two explanations for the lack of multiple distinct mineral populations in mafic sheet compositions. One possibility is low crystallinity of the felsic parental magma prior to mixing. Magma ascending adiabatically from the lower crust can attain a near-liquidus or superliquidus state due to differences in the slope of the liquidus and adiabat (Annen et al., 2006), and mixing between felsic magmas and hotter mafic magma can also result in reheating to near-liquidus or superliquidus conditions (e.g., Sparks and Marshall, 1986) and dissolution of any crystalline material present. If this is the case, then felsic magmas involved in mixing may supply few, if any, crystals that formed prior to mixing. Evidence to support this idea is evident in the single hybrid sheet for which we have mineral chemistry data. In this sheet, amphibole from both the mafic (AKCP-112AM, 51.39 wt% SiO2) and felsic (AKCP-112AF, 58.84 wt% SiO2) samples have trace element compositions that overlap with those of amphibole from other mafic samples (Fig. 12). This suggests that even in the relatively SiO2-rich portion of the hybrid sheet, amphibole is derived primarily from a mafic magma source and little, if any, derived from the felsic parental magma.
Alternatively, the lack of distinct mineral compositions in mixed magmas may also reflect greater extents of postemplacement crystallization. Compared to volcanic rocks, plutonic rocks undergo substantially greater extents of crystallization close to the solidus. Magmas are only mobile at crystallinities of 40%–50% or less (e.g., Bachmann and Bergantz, 2008), and thus a significant interval of crystallization in a plutonic rock occurs after emplacement, and the majority of minerals that form will have compositions that reflect the liquid compositions produced as the emplaced magma cools and crystallizes. Minerals that form close to the solidus, and after any mixing events that establish the bulk composition, may thus mask the presence of early-formed minerals. We see some evidence for this process in the REE compositions of amphibole, which have the characteristic humped patterns with a significant range of REE concentrations and in many cases show pronounced Eu/Eu* anomalies (Fig. 13A). Although there is some scatter, Eu/Eu* is correlated with abundances of REE (Fig. 13B), Sr, and other trace elements; this is consistent with amphibole recording progressive growth of a mineral assemblage dominated by plagioclase and amphibole. A similar trend is also observed in some volcanic amphibole (e.g., Thornber et al., 2008). In general, Eu/Eu* ratios in amphibole increase with decreasing Gd contents (Fig. 13B). There is a greater degree of scatter in Eu/Eu* at higher Gd content, which is also to be expected during crystallization close to the solidus, as at higher crystallinity the extent of equilibration between the remaining amounts of liquid is reduced. Therefore local differences in phase proportions, which are evident in field relations (e.g., Fig. 3F), can drive divergent liquid evolution (e.g., Kent et al., 2008). We also note that, despite the pronounced negative Eu/Eu* values in most amphibole in hybrid and felsic samples, all bulk-rock Eu/Eu* values are 1.0–1.2. This suggests that most amphibole growth occurred late, following mixing, emplacement, and initiation of plagioclase growth.
It is difficult to completely evaluate the relative importance of these two alternatives and it is likely that both play an important role in dictating the mineral compositions present in individual sheets. We also note that reequilibration of minerals during postsolidus cooling is also unlikely to control the compositions of minerals in individual sheets (cf. Elburg, 1996) as we see compositional distinctions in elements, such as Al and the REEs that diffuse slowly in amphibole at near solidus and subsolidus temperatures. A general calculation based on available diffusivities suggests that the time scales for complete diffusional reequilibration for Sr, An content, and REEs in centimeter-size amphibole and plagioclase crystals at temperatures ≤700 °C are >100 m.y. (Brady and Cherniak, 2010).
Formation of the Tenpeak Sheeted Complex
Our data allow us to constrain the processes that lead to the formation of a well-exposed magmatic sheeted complex hosted in a deep-crustal arc pluton. One key observation is the timing relationship between the emplacement of the sheeted complex and the rest of the Schaefer Lake tonalite unit of the Tenpeak pluton. Previous workers have considered the sheeted complex as a subunit of the Schaefer Lake unit on the basis of compositional and mineralogical similarities and a shared gradational contact (e.g., Miller et al., 2000; Matzel et al., 2006; Miller et al., 2011b). Our zircon geochronology suggests that the sheeted complex postdated the emplacement of the Schaefer Lake tonalite by ≤150 k.y. This rules out a change in emplacement style during the nascent stages of pluton emplacement as a mechanism for forming the sheeted complex (e.g., Hanson and Glazner, 1995; Glazner et al., 2004), and instead suggests that formation of the sheeted complex may have occurred after or during the waning phase of emplacement of the Schaefer Lake tonalite. Despite this, the exact reason for the transition from the Schaefer Lake phase to the more heterogeneous sheeted complex is unclear. Given the relative imprecision of emplacement rate estimates, it is difficult to dismiss the possibility of somewhat lower transient emplacement rates associated with waning activity of the Schaefer Lake tonalite.
The transition from more voluminous and homogeneous Schaefer Lake tonalite to tonalitic and dioritic sheets may have been aided by thermal conditions along the WRSZ. The shear zone is located adjacent to the southwestern margin of the Tenpeak pluton (Fig. 1), juxtaposes the Chiwaukum Schist against the pluton, and probably deforms the southwestern portion of the sheeted complex to the south of our study area. Individual sheets and mineral fabrics within the complex are broadly subparallel to the WRSZ contact. The well-developed magmatic and solid-state fabrics in sheets suggest that sheet emplacement and/or cooling was concurrent with displacement in the shear zone. While deformation in the shear zone may have occurred during emplacement, the potential genetic relationship between the sheeted complex and shear zone is less transparent. The WRSZ is contractional, so dilational jogs, common in normal and strike-slip fault zones, are unlikely to have accommodated magmas (e.g., Brown, 1994). The rates of emplacement are also considerably faster than displacement rates in fault zones, a further argument that dilation in the shear zone is not a major material transfer process during intrusion (cf. Paterson and Tobisch, 1992).
However, it is still conceivable that anisotropies of the fault zones helped to localize magmatism along that southwestern margin. Construction of sheeted complexes is commonly associated with activity along fractures, faults, or shear zones (e.g., Pawley et al., 2002) by either direct emplacement or magma wedging (Miller and Paterson, 2001a). The process of magma wedging involves downward and lateral displacement of the host rock and/or older crystal-rich magmatic blocks (e.g., Hutton, 1992; Miller and Paterson, 2001a; Paterson et al., 2008). In the Tenpeak sheeted complex, the strongly foliated rafts of Napeequa meta-supracrustal rocks within the sheeted complex are consistent with magma wedging. Postemplacement crystallization may have also facilitated sheet emplacement by dictating the rheology of emplaced magmas and localizing subsequent intrusive events (e.g., Bergantz, 2000; Miller et al., 2011a).
Zircon systematics are also consistent with relatively quick cooling of individual sheets after emplacement and minimal interaction between emplaced sheets. The zircons from individual sheets record a significantly more limited range of ages compared to samples from the more homogeneous phases of the Tenpeak and other North Cascade plutons (e.g., Matzel et al., 2006; Shea et al., 2016), where age populations span a range of several million years or more (e.g., Shea et al., 2016). The simple zircon populations of the sheeted complex (Fig. 15) likely reflect (1) the small overall volume of the sheets relative to the adjacent voluminous Schaefer Lake tonalite and other larger phases of the Tenpeak pluton (e.g., Caricchi et al., 2014), and (2) relatively rapid cooling following emplacement, leaving zircon growth to occur within a narrow window between emplacement and cooling below the solidus. Rapid cooling requires that the wall rocks adjacent to the pluton are sufficiently cold to act as a thermal boundary. While peak metamorphic temperatures estimated for the Napeequa unit (640–740 °C; Whitney et al., 1999) are close to the granite solidus, it is likely that the host rock was cooler during emplacement of the sheeted complex and acted as a thermal boundary. The 40Ar-39Ar hornblende ages from the southwestern portion of the Tenpeak intrusion (Matzel, 2004) show rapid cooling upon emplacement suggesting host rock temperatures were below the hornblende closure (∼500-550°C). This suggests that the host-rock temperatures were cool enough to enable rapid cooling of the intrusion. There is some field evidence for magma interaction between emplaced sheets (e.g., crystal exchange and heterogeneous hybrid sheets; Figs. 3A, 3E) although major and trace element compositions of the amphibole and plagioclase suggest that this was not a widespread process.
The sheeted complex of the Schaefer Lake unit in the Tenpeak pluton has provided the opportunity to understand the time scales, mechanisms, and magma processes that contribute to incremental growth of plutons in the deep crust. Results from this study suggest that emplacement rates during formation of the sheeted complex are similar to many other larger and/or more homogeneous plutons and thus gross differences in emplacement rate do not appear to be a key driver for the generation of sheeted intrusions. Correlations between major fault zones and pluton emplacement have been well documented elsewhere (e.g., Berthé et al., 1979; Hutton, 1992; D’Lemos et al., 1992), although their genetic relationships are not always clear (e.g., Paterson and Schmidt, 1999). In the case of the Tenpeak sheeted complex, dilation in the WRSZ did not play a major role in host-rock transfer during magma emplacement, but anisotropies in the shear zone and/or a thermal boundary created with adjacent wall rocks may have localized the sheets.
(1)Detailed mapping of an ∼750-m-thick sheeted magma complex along the southwest margin of the Tenpeak pluton revealed ∼58 discernable intrusions of gabbro, heterogeneous tonalite, and quartz diorite. The sheeted complex is a subunit of the adjacent Schaefer Lake tonalite unit of the Tenpeak pluton, and is between the more homogeneous tonalite and the contact with host rocks in the WRSZ. The contacts of the sheets are largely parallel to the shear zone along the southwest margin of the pluton and with each other. Sheet thickness generally increases structurally upward away from the contact toward the more homogeneous Schaefer Lake rocks.
(2)The sheeted complex formed between 91.929 ± 0.023 and 91.835 ± 0.030 Ma, over an interval of 94,000 ± 62,500 yr (95% confidence interval). Emplacement rates are estimated as 1.1–5.4 × 10–4 km3 yr–1, within the range of estimated rates for the Tenpeak pluton overall, as well as other North Cascades plutons.
(3)Magmatic compositions appear to represent mixing between a single mafic parental magma and an array of felsic magma compositions. Production of the array of felsic magmas, formed via mixing of different crustal melts and/or fractionation, occurred relatively early and possibly at the site of magma generation within the lower crust. Mixing between the mafic parent and the array of felsic sheet compositions produced a diversity of mafic sheet compositions, and this mixing event appears to have happened relatively late, almost immediately prior to emplacement and/or at the site of emplacement.
(4)Mineral compositions show little evidence for magma mixing preserving diverse crystal populations derived from different parental magmas, suggesting that most crystal growth occurred after mixing had occurred, and/or that early minerals are difficult to recognize among higher proportions of minerals that formed after mixing and emplacement during cooling to the solidus.
(5)The intrusion of magmatic sheets appears to have been localized by fault anisotropies along the WRSZ, and sheet emplacement probably occurred via magma wedging. The WRSZ and adjacent Chiwaukum Schist also likely represented an important thermal boundary during sheet emplacement, resulting in relatively rapid cooling of individual sheets and relatively simple zircon age populations within each dated sheet.
This research was supported by National Science Foundation grants EAR-0948616 to Kent, EAR-0948388 to Bowring, EAR-0948685 to J. Miller and R. Miller, and EAR-1119358 to R. Miller. Analytical assistance was provided by Frank Tepley, Matthew Loewen, Alison Koleszar, and Dale Burns. David Burney helped with some of the data collection and we give a hearty toast to Aaron, the mule packer, for his courageous spirit during one of our epic field expeditions. This manuscript benefited from discussions with Scott Paterson, John Dilles, Anita Grunder, and other members of the VIPER (Volcanology, Igneous Petrology and Economic Geology) research group at Oregon State University. We acknowledge Tom Sisson, Erin Todd, Tamara Carley, and an anonymous reviewer for their constructive reviews. We also thank Geosphere Science Editor Shanaka de Silva and Associate Editor Rita Economos for their insightful feedback.