Abstract

Studies of plutons indicate that they are the result of a complex interplay of magmatic processes occurring during magma generation, ascent, and emplacement. A critical tool for deciphering these processes is high-precision geochronology, which can help determine the timing and rates of magmatism in the crust. We conducted a field and U-Pb geochronological study of the Cretaceous Black Peak intrusive complex in the North Cascades of Washington State to investigate magmatism at a detailed scale and to refine estimates of plutonic construction rates. High-precision chemical abrasion–thermal ionization mass spectrometry (CA-TIMS) U-Pb geochronology was carried out on 31 samples from five mapped intrusive phases. Field relations in the Black Peak intrusive complex show intrusive contacts that vary from sharp to gradational. Whole-rock Sm/Nd, zircon oxygen isotopes, and zircon trace elements were obtained on subsets of representative samples. The U-Pb geochronology from the Black Peak intrusive complex documents batholith intrusion over 4.5 m.y. and suggests that magmatism was semicontinuous for a minimum of 3.5 m.y. Individual samples display age dispersion in single-zircon dates that ranges from ∼105 yr to several 106 yr, with a general increase in the age range for younger samples. Whole-rock εNd and zircon δ18O for all Black Peak intrusive complex samples indicate that magmas were derived from mantle and crustal sources and that all magmas were isotopically homogenized prior to zircon saturation. Ti-in-zircon temperatures from zircon cores are generally above calculated zircon saturation temperatures, which suggests that most Black Peak intrusive complex magmas were zircon undersaturated in the melt source region. A range of thicknesses was considered, and a thickness of ∼10 km for the Black Peak intrusive complex gives an average intrusion rate of ∼1.1 ×10–3 km3/yr, which is high enough to sustain a magma reservoir in the shallow crust. The field evidence and long overall duration of intrusion are incompatible with the entire Black Peak intrusive complex being molten at any one time, but the larger, more compositionally homogeneous domains in the Black Peak intrusive complex are likely the solidified remnants of mushy magma bodies with ∼105 yr durations. These data suggest that the Black Peak intrusive complex may have remained “mushy” for long periods of time (105 yr) and may indicate that the spread in dates within individual samples is best interpreted as either antecrystic recycling and/or protracted autocrystic growth.

INTRODUCTION

The composite nature of plutons has been recognized by numerous studies over many decades (Read, 1948; Hamilton and Myers, 1967; Pitcher and Berger, 1972; Hill et al., 1985; Hutton, 1992; Coleman et al., 1995; Vigneresse and Bouchez, 1997; Cruden, 1998; Wiebe and Collins, 1998; Brown and McClelland, 2000; Miller and Paterson, 2001; Miller and Miller, 2002; Mahan et al., 2003; Glazner et al., 2004; Walker et al., 2007; Morgan et al., 2008; Rocchi et al., 2010; de Saint Blanquat et al., 2011; Miller et al., 2011; Brown, 2013). In the last decade, advances in geochronologic techniques, particularly the development of high-precision, single-zircon thermal ionization mass spectrometry (TIMS) U-Pb geochronology, have produced data that suggest that most plutons are constructed incrementally over time scales varying from several 104 yr to several 106 yr (e.g., Coleman et al., 2004; Matzel et al., 2006; Michel et al., 2008; Burgess and Miller, 2008; Schaltegger et al., 2009; Davis et al., 2011; Memeti et al., 2010; Leuthold et al., 2012; Schoene et al., 2012; Frazer et al., 2014; Barboni et al., 2015). These studies build on a foundation of research that has examined nearly every aspect of the magmatic process, including magma generation, transport, emplacement, fractionation, and mixing (e.g., Hildreth and Moorbath, 1988; Huppert and Sparks, 1988; Clemens and Mawer, 1992; Petford et al., 1994; Miller and Paterson, 1999; Bergantz, 2000; Dufek and Bergantz, 2005). With the addition of unprecedented temporal resolution afforded by modern U-Pb geochronology, effort has now been put into modeling these processes, focusing especially on the thermal viability of large magma chambers in Earth’s crust, and whether plutons represent the solidified remnants of these magma chambers (Annen et al., 2006; Annen, 2009; Karlstrom et al., 2009; Schöpa and Annen, 2013; Gelman et al., 2013; de Silva and Gregg, 2014).

All of these modeling studies depend critically on estimating magmatic recharge/intrusion rates over the lifetime of a plutonic system, but the intrusion rates are frequently based on sparse geochronologic data; in many cases, one or two samples are taken to represent tens to hundreds of square kilometers of pluton. These estimates have important implications for the interpretation of how large intrusive bodies are assembled, as models of pluton construction suggest that, given high enough intrusion rates, these regions could become large, dynamic magma chambers (Jellinek and DePaolo, 2003; Karlstrom et al., 2010; Gelman et al., 2013; Schöpa and Annen, 2013; Menand et al., 2015) similar to those observed in the southern Andes (Singer et al., 2014) and Altiplano (de Silva and Gosnold, 2007; Ward et al., 2014) and hypothesized in the Southern Rocky Mountain volcanic field (Lipman and Bachmann, 2015). Despite evidence for magmatic “pulses” in igneous systems (de Saint Blanquat et al., 2011; de Silva et al., 2015), most geochronologic data sets from plutons are not extensive enough or of high enough resolution to resolve brief (<100 k.y.) high-flux periods, or unambiguously capture the time frame of magma storage in regions where large plutons form.

We used high-precision chemical abrasion (CA)–TIMS U-Pb geochronology of zircon from the ca. 90 Ma Black Peak intrusive complex in the North Cascades of Washington State to illuminate a detailed time scale for construction of an intrusive complex. This study combines extensive field work, high-precision geochronology, in situ zircon trace element and isotopic analyses, and whole-rock Sm/Nd isotopic analyses to produce a thorough characterization of the complex over its ∼4.5 m.y. assembly. Our results suggest that the Black Peak intrusive complex is composed of many intrusions that were emplaced relatively continuously over the lifetime of the complex at a minimum intrusion rate of ∼1.1 × 10–4 km3/yr, comparable to rates from the well-studied Mount Stuart batholith (Matzel et al., 2006). Based on the timing of zircon crystallization in different samples, the Black Peak intrusive complex appears to have gone through a period of nascent magmatism, entered a peak thermal period, and waned until final cessation of intrusion. During the peak thermal period, parts of the complex may have existed as a “mushy” system on 105 yr time scales. This data set provides the most detailed geochronologic picture to date of intrusive magmatism in a Cordilleran arc plutonic complex.

GEOLOGIC SETTING

The crystalline core of the North Cascades (Cascades core) is the southernmost extension of the 1500-km-long Coast plutonic complex, which stretches from southeast Alaska to northern Washington (Rubin et al., 1990). The Cascades core records a history of mid-Cretaceous accretion, deformation, metamorphism, and plutonism (Misch, 1966). Eocene transtension and continued plutonism then preceded initiation of the modern Cascades arc around 40 Ma (Tabor et al., 1989; Miller and Bowring, 1990; Haugerud et al., 1991; Paterson et al., 2004; Gordon et al., 2010; Eddy et al., 2016). The core is bounded by the Straight Creek fault to the west, the Ross Lake fault zone (Hozameen, Foggy Dew, and Ross Lake faults, and the Gabriel Peak tectonic zone) to the east, and the Windy Pass thrust to the south (Fig. 1; Misch, 1966; Miller, 1985).

Geology of the Black Peak Intrusive Complex

The Black Peak intrusive complex (previously referred to as the Black Peak quartz diorite or the Black Peak batholith; Adams, 1964; Misch, 1966; Miller, 1987; Dragovich et al., 1997; Miller et al., 2009a) is exposed as an elliptical-shaped (map view) body emplaced along the eastern edge of Cascades core (Figs. 1 and 2A). The Black Peak intrusive complex is deformed by the Gabriel Peak tectonic zone on the west and intrudes the Triassic Twisp Valley schist and Cretaceous strata of the Methow Basin (specifically, the Midnight Peak Formation and the Pasayten Group) on the east (Barksdale, 1975; Miller and Bowring, 1990). The complex is deformed by ductile shear zones that are interpreted to have formed in a contractional step-over between the Twisp River fault and the Gabriel Peak tectonic zone, both of which are part of the Ross Lake fault zone (Miller and Bowring, 1990; Miller, 1994).

Al-in-hornblende barometry on samples from the Black Peak intrusive complex indicates that the eastern margin of the complex crystallized at 1–3 kbar, which is compatible with the low grade of host rocks of the Midnight Peak Formation and Pasayten Group (D.L. Whitney and R.B. Miller, personal observation from four samples; their analyses conducted in 1990–1992 and first cited in 1993). Amphibolite-facies rocks in contact with the western margin of the complex record pressures of 6–8 kbar (Miller et al., 1993). However, the western margin of the Black Peak intrusive complex is the Gabriel Peak tectonic zone, placing the complex in fault contact with these amphibolite-facies rocks. There is no evidence for higher pressures (e.g., magmatic epidote) along the western margin, nor is there clear evidence that the Black Peak intrusive complex intruded these higher-pressure rocks.

Rocks of the Black Peak intrusive complex are dominated by tonalite and granodiorite, but diorite, gabbro, and rare granite are also present (Adams, 1964; Miller, 1994). Intrusive relationships and previous geochronology suggest a temporal progression from mafic to felsic (Adams, 1964; Dragovich et al., 1997). Based on previous mapping and our field work, the complex can be subdivided into four texturally and compositionally distinct units: the Crescent Mountain unit, the Stiletto Mountain unit, the Reynolds Peak unit, and the War Creek unit (Figs. 2A–2C; Miller, 1987). Another region, the Louis Lake heterogeneous zone, is also separated out based on its unique magmatic features, discussed later herein (Fig. 2C).

Two types of contacts are delineated in the following unit descriptions, external contacts (between defined units) and internal contacts (within defined units). External contacts are determined by significant changes in color index, texture, and/or composition. These contacts, where observed, range from sharp to sheeted. Sheeted contacts consist of tabular rafts of the older unit within the younger unit. These rafts generally decrease in size and frequency with distance from the older unit. Internal contacts are often more subtle and are evidenced by changes in color index, hornblende-to-biotite ratio, grain size, and composition. These contacts are often difficult to observe directly, vary from gradational to sharp, and generally cannot be traced for more than 10 m.

The Crescent Mountain unit is a relatively small intrusion exposed over an area of ∼10 km2 along the eastern margin of the Black Peak intrusive complex (Fig. 2A). The unit is in sharp contact with the Twisp Valley schist (Fig. S11). Small (<10 cm across) xenoliths of host rock are present close to the contact, and thin (centimeter-scale) sheets of the Crescent Mountain unit intrude the host rock along foliation. The Crescent Mountain unit is mostly coarse- to fine-grained hornblende diorite with some hornblendite, gabbro, and biotite tonalite. Contacts between these diverse lithologies are rarely observable, but they vary from sharp to gradational where observed. Crosscutting relationships indicate that the Crescent Mountain unit is the oldest unit in the Black Peak intrusive complex (Adams, 1964; Dragovich et al., 1997).

The Stiletto Mountain unit is the most voluminous phase within the Black Peak intrusive complex (Figs. 2A and 2B) and is characterized by compositions that range from diorite to granite, although the majority of the unit is tonalitic. Like the Crescent Mountain unit, observable contacts between lithologies are rare, but they vary from sharp to gradational. To the west, the Stiletto Mountain unit is truncated and deformed by the Gabriel Peak tectonic zone. To the east, the unit intrudes the Crescent Mountain unit, the Twisp Valley schist, Winthrop Formation, and the Midnight Peak Formation (Figs. 1 and 2A). Intrusive relationships between the Stiletto Mountain and Crescent Mountain units vary from sharp to mingled (Figs. S2 and S3 [see footnote 1]). Contacts between the Stiletto Mountain unit and the Twisp Valley schist, Winthrop Formation, and Midnight Peak Formation are sharp and characterized by many centimeter- to meter-scale dikes of tonalite intruding and breaking off host-rock material (Fig. S4 [see footnote 1]).

The Louis Lake heterogeneous zone is composed of four texturally distinctive rock types that are exposed over ∼2 km2 (Fig. 2C). This area is unique within the intrusive complex due to evidence for extensive magmatic interaction (e.g., enclaves, schlieren, mingling) that is uncommon in the other units of the Black Peak intrusive complex. The Louis Lake heterogeneous zone is dominated by coarse-grained biotite tonalite and medium- to fine-grained biotite tonalite, which are present in subequal amounts. Interactions between these lithologies are complex, and contacts vary from gradational to mingled, sometimes within a distance of a few meters (Figs. S5–S7 [see footnote 1]). Enclaves, schlieren, and magmatic layering characterize many outcrops, regardless of lithology. Truncating both of these units, there is a medium-grained biotite tonalite that lacks enclaves, schlieren, or other evidence for magmatic interaction. Additional lithologies in the Louis Lake heterogeneous zone include fine-grained gray porphyritic biotite tonalite dikes and less common irregularly shaped bodies, and coarse-grained leucotonalite dikes. The porphyritic bodies are generally less than 3 m in width. Their contacts are commonly sharp. The leucotonalite dikes are <2 m thick and are present throughout the zone. These dikes crosscut all units in the Louis Lake heterogeneous zone.

The Reynolds Peak unit comprises the central region of the Black Peak intrusive complex, where it intrudes along the central axis of the Stiletto Mountain unit (Fig. 2A). Rocks in the Reynolds Peak unit are dominantly tonalitic and are identified in outcrop by the presence of large (0.25–0.5 cm) quartz grains, minor porphyritic plagioclase, and a low color index (<25). The eastern contact of the Reynolds Peak unit with the Stiletto Mountain unit–Louis Lake heterogeneous zone is mostly obscured, making it difficult to ascertain contact relations; a complete transition from the Stiletto Mountain unit to the Reynolds Peak unit occurs over <200 m. Where exposed, this contact consists of tabular rafts of Stiletto Mountain unit within the Reynolds Peak unit. The northernmost margin of the unit is characterized by small euhedral hornblende in addition to the usual biotite.

The youngest unit in the Black Peak intrusive complex is the War Creek unit, a heterogeneous assemblage of biotite tonalite and leucotonalite. Many outcrops are sheeted on the centimeter to meter scale, with layers distinguished by biotite content and grain size (Fig. S8 [see footnote 1]). The western margin of the unit is truncated and deformed by the ductile Gabriel Peak tectonic zone (Miller, 1994), modifying or obscuring magmatic features and contact relationships. The contact between the War Creek unit and the Reynolds Peak unit is marked by tabular rafts of the Reynolds Peak unit within the War Creek unit in a zone at least 1 km wide. These rafts range from centimeters to meters thick and are a minimum of 10 m long.

Foliation in the western Black Peak intrusive complex is dominated by solid-state fabrics generated by the 65–48 Ma NW-trending Gabriel Peak tectonic zone (Miller, 1994); foliation intensity increases westward into the fault zone and becomes mylonitic at the contact with the Skagit gneiss complex and associated metamorphic rocks. Foliation in the eastern Black Peak intrusive complex is magmatic and has a weak to moderate solid-state overprint, possibly associated with deformation along the Ross Lake fault zone. Magmatic foliation is generally parallel to sheet contacts. Foliation in the eastern two thirds of the complex appears to define a kilometer-scale, NW-SE–trending, doubly-plunging antiform that extends along the length of the complex (Fig. 2A). The fold-like structure is truncated in the north by the Golden Horn batholith, suggesting deformation that predates ca. 48 Ma.

RESULTS

Analytical methods for zircon U-Pb geochronology, zircon in situ geochemistry, Sm/Nd isotopic data, and zircon oxygen isotopic data are given in the Supplemental File (see footnote 1). Data tables for zircon in situ geochemistry, Sm/Nd isotopes, and zircon oxygen isotopes are given in Tables S1–S3, respectively (see footnote 1).

U-Pb Geochronology

High-precision U-Pb TIMS analyses were conducted on the Massachusetts Institute of Technology (MIT) VG Sector 54 multicollector mass spectrometer or the MIT Isotopx X62 multicollector mass spectrometer. For zircon U-Pb geochronology, multiple samples from each of the four main map units and the Louis Lake heterogeneous zone subunit were analyzed to elucidate the variability within and between units of the Black Peak intrusive complex. Samples were selected, as much as was practicable, to be representative of compositional variation and spatial distribution within each unit. Cathodoluminescence (CL) images of zircons indicate that some crystals are concentrically zoned from core to rim or show concentric zoning around a homogeneous core (Figs. 3A and 3B). However, other crystals have obviously discordant cores that likely indicate inheritance (Figs. 3C and 3D). Microsampling (physical separation and analysis of zircon tips and cores) of zircons without obvious cores suggests that most rims and cores from these “simple” crystals have indistinguishable ages within uncertainty. U-Pb data are summarized in Table 1 and presented in full in Table 2 for each unit. Detailed sample and zircon descriptions, CL images, and concordia diagrams are available in the Supplemental File (Figs. S9–S62 [see footnote 1]).

An important consideration in current high-precision U-Pb geochronology is the meaning of a zircon date from a single sample, something that is considered briefly here but discussed in greater detail later herein. In the simplest case, if a magma crystallizes all zircon at the same time (within uncertainty), one would expect all of the zircon dates to overlap within their analytical uncertainties and to give a mean square of weighted deviation (MSWD) around 1.0. However, in many cases, the scatter of dates from a single sample cannot be explained by analytical uncertainties alone, yielding weighted mean dates with considerable overdispersion (MSWD >> 1). The geologic interpretation of this overdispersion could have several sources; older dates could be due to protracted zircon crystallization, inheritance from the source region, magma mixing, or a combination of these factors. U-Pb zircon dates from Black Peak intrusive complex samples typically are on or slightly offset from concordia (Figs. S58–S62 [see footnote 1]). Discordant analyses, possibly due to Pb loss, are not discussed in detail. These analyses are shown in the data set for completeness but are not included in weighted means or considered in our conclusions (Figs. 4–8; Table 2).

When two or more dates from the youngest zircons in a sample overlap within error, the weighted mean of these dates is used to estimate the timing of final solidification of the sample. However, if there is no cluster of youngest analyses that give a low MSWD, the youngest single-crystal zircon date is used as the best estimate of the timing of final solidification. There is still considerable debate regarding the use of the youngest single-crystal date because these zircons may represent rare instances where chemical abrasion did not completely remove all parts of the grain that experienced Pb loss. However, these instances are generally uncommon and obvious (e.g., Schoene et al., 2010) and do not significantly affect our interpretations.

Crescent Mountain Unit

Four samples from the Crescent Mountain unit were analyzed, including a fine-grained quartz diorite (PX10-13B), two hornblende gabbros (MAF-1, PX10-34B), and a biotite tonalite (PX10-54; Table 1; Fig. 4). Weighted mean and youngest crystal dates range from 91.755 ± 0.040 Ma (MAF-1) to 91.18 ± 0.12 Ma (PX10-13B), spanning ∼600 k.y. (Fig. 4). Crystallization ages in the Crescent Mountain unit correlate with a general compositional trend that becomes more felsic with time. Older zircons are rare; only one sample (PX10-54) has a zircon with an age (91.904 ± 0.083 Ma) that is resolvably (≈700 k.y.) older than the rest of the zircons in the sample (91.378 ± 0.36 Ma; Fig. 4).

Stiletto Mountain Unit

Thirteen samples from the Stiletto Mountain unit were analyzed, including two granodiorites (PX10-76, PX10-34A), six biotite tonalites (PX10-86, GP309-1, GP-322, K9, K16B, K55), three hornblende tonalites (K26, K42B, PX11-289), one granite (PX10-251), and one hornblende diorite (PX10-236). Our field work and geochronology suggest that samples from the eastern Stiletto Mountain unit have higher hornblende content and are generally older than those from the western part of the unit. Weighted mean ages in the eastern part of the Stiletto Mountain unit range from 90.345 ± 0.030 Ma (PX10-76) to 89.710 ± 0.043 Ma (K42B; Table 1; Figs. 5 and 6). Final solidification ages in the western part of the Stiletto Mountain unit range from 89.183 ± 0.067 Ma (PX10-86) to 88.81 ± 0.30 Ma (GP309-1; Table 1; Fig. 6). Older zircons are uncommon in samples from the eastern Stiletto Mountain unit and become more common in the western Stiletto Mountain unit. Zircon dates range up to 90.975 ± 0.064 Ma (PX10-34A), ∼700 k.y. older than the inferred age of the sample.

Louis Lake Heterogeneous Zone

Six samples from the Louis Lake heterogeneous zone were dated (Table 1). These samples (PX11-261A, PX10-209, PX11-263C, PX11-284, PX11-270, PX11-268) are all tonalitic. The oldest sample has a weighted mean age of 89.762 ± 0.033 Ma (PX11-261A), and the youngest has a single-crystal age of 89.017 ± 0.072 Ma (PX11-268; Table 1; Fig. 6). Most samples show evidence for scatter outside of analytical uncertainty; older zircons are common, ranging up to 90.875 ± 0.052 Ma, ∼1.7 m.y. older than the youngest date in the sample (PX11-270; Fig. 6).

Reynolds Peak Unit

Six samples from the Reynolds Peak unit were analyzed, including two biotite tonalites (PX10-103A, GP-158-10), two biotite leucotonalites (PX10-103B, SCP), one granodiorite (PX10-96), and one granite (PX10-221). The oldest sample has a weighted mean age of 88.467 ± 0.031 Ma (PX10-96), and the youngest has a weighted mean age of 87.765 ± 0.072 Ma (PX10-103A; Fig. 7; Table 1). Distinctly older zircons are common in all samples; individual dated zircons range up to 93.38 ± 0.12 Ma, ∼5 m.y. older than the weighted mean age of the sample (GP-158-10; Fig. 7). However, out of 60 analyses, only four are older than 90 Ma. Instead, most samples have dates that are dispersed from 1 to 2 m.y. older than the inferred solidification ages of the sample.

War Creek Unit

Two samples from the War Creek unit were analyzed, PX10-175 and WCG. Both samples are composed of homogeneous, fine-grained, strongly foliated biotite tonalite. Sample WCG has a single-crystal age of 87.474 ± 0.052 Ma, and a weighted mean of the youngest three analyses from sample PX10-175 gives an age of 86.862 ± 0.062 Ma. Older zircons are common in both War Creek unit samples, and individual dated zircons range up to 161.43 ± 0.11 Ma (WCG; Table 1; Fig. 8). Like the Reynolds Peak unit, zircons this old are relatively infrequent (only three total analyses older than 100 Ma out of 23 analyses). Instead, most samples have dates that are dispersed up to 3 m.y. older than the inferred solidification ages of the sample.

Zircon Trace Element Data

In situ titanium (Ti) and hafnium (Hf) concentrations of zircons from 11 samples (MAF, PX10-76, PX11-284, PX10-261A, PX10-34A, SCP, PX10-175, WCG, PX10-86, PX10-221, PX10-236) were acquired using the sensitive high-resolution ion microprobe–reverse geometry (SHRIMP-RG) located at the U.S. Geological Survey–Stanford Ion Probe Laboratory. Zircons from the Black Peak intrusive complex are characterized by relatively low Hf cores surrounded by higher Hf rims, indicating magmas became more fractionated with time (Fig. 9; Claiborne et al., 2010). In situ analyses of zircon cores from sample MAF-1 (gabbro) have the lowest overall Hf values, compatible with its more mafic composition. In situ zircon model temperatures from the Black Peak intrusive complex, based on the titanium (Ti)-in-zircon thermometer of Watson et al. (2006) and Ferry and Watson (2007), span a range of values from ∼900 °C to 700 °C (Fig. 9). Because rutile is not present in our samples, TiO2 activity is less than 1. However, titanite is present, which allows us to constrain the titanium activity between 0.6 and 0.9, and we calculated temperatures for these end members (Fig. 9; Hayden and Watson, 2007). SiO2 activity is estimated at 1 (Fig. 9; Ferry and Watson, 2007). Absolute temperatures are less important than relative temperature variations; most samples from the Black Peak intrusive complex record a 100–200 °C change in temperature from core to rim (Fig. 9). In general, zircons from the Stiletto Mountain unit have higher Ti concentrations than zircons from the Reynolds Peak unit or the War Creek unit. Similar to previous studies of other intrusions (e.g., Claiborne et al., 2010; Barth and Wooden, 2010), zircons from the Black Peak intrusive complex illustrate an inverse relationship between zircon model temperature and Hf concentration (Fig. 9).

Sm/Nd Isotopic Data

Sm/Nd whole-rock isotopic analyses were carried out on the MIT VG Sector 54 TIMS. In total, five samples from the Black Peak intrusive complex were analyzed: PX10-13B, PX10-148, PX10-251, PX10-236, and WCG (Fig. 10). All of these samples but PX10-148 were also dated using U-Pb zircon geochronology (see footnote 1). Samples were chosen as compositionally representative of each of the units within the Black Peak intrusive complex. Sm/Nd data and sample descriptions are also available in the Supplemental File (Table S2 [see footnote 1]). Two samples, MAF (Crescent Mountain unit) and SCP (Reynolds Peak unit), previously analyzed by Matzel et al. (2008) are also included for comparison. The εNd data are presented as εNd(t), which is calculated at the inferred solidification age of the sample.

Rocks from the Black Peak intrusive complex exhibit a very limited range of εNd(t) values (from +6.3 ± 0.5 to +5.14 ± 0.12; Fig. 10). The Crescent Mountain unit exhibits modest variation in εNd(t), ranging between +6.3 ± 0.5 (MAF; Matzel et al., 2008) and +5.14 ± 0.12 (PX10-13B). The Stiletto Mountain unit shows no discernible change between three samples (+6.16 ± 0.83 to +5.89 ± 0.86; Fig. 10). The εNd(t) values for the Reynolds Peak unit (+5.4 ± 0.5) and the War Creek unit (+4.89 ± 0.81) both overlap within error with the Stiletto Mountain unit (Fig. 10).

Oxygen Isotopic Data

Oxygen isotopes were measured using the University of California–Los Angeles (UCLA) Cameca IMS-1270 from a subset of the zircons previously analyzed for trace elements using the SHRIMP-RG. These samples were selected to represent major intrusive units in the complex (MAF, PX10-76, PX10-86, PX10-221, PX10-175). In situ oxygen isotope data in zircon suggest a limited range of δ18O values in the Black Peak intrusive complex. Average δ18O values of zircons from all five analyzed samples vary between 6‰ and 7‰, i.e., above the mantle value of 5.3–5.5‰ ± 0.3‰ (Valley, 2003). These values suggest that magmas comprising the Black Peak intrusive complex must have been in part derived from partially melted crust, since they do not have pure mantle values. Our data show that the most mafic sample (MAF) has a slightly lower (<0.5‰) δ18O value than the felsic samples, but otherwise there are no discernible trends (Fig. 11).

DISCUSSION

Intrusive History

U-Pb zircon analyses from the Black Peak intrusive complex indicate that the complex was intruded over ∼4.5 m.y. and followed a broad compositional trend from older, mafic intrusions to younger, felsic intrusions. The relatively high density of samples from each unit allows us to describe a detailed intrusive history for the entire complex. Notably, units defined by Miller (1987) fall into unique age groups, indicating that each lithologic map body is also geochronologically distinct.

Determining whether a pluton represents “episodic” versus “continuous” intrusion is often a matter of interpretation and the time scale over which intrusion occurs (Samperton et al., 2015). In general, “episodic magmatism” implies periods of high intrusion rate separated by lulls that are often >1 m.y. long (e.g., Matzel et al., 2006; de Silva and Gosnold, 2007). In contrast, “continuous magmatism” implies a steady-state intrusion of material. While it is reasonably easy to propose episodic intrusion, particularly when geochronology is sparsely distributed, it is practically impossible to definitively prove that continuous magmatism occurred. In an attempt to establish whether the Black Peak intrusive complex was intruded episodically or continuously, we took the difference between the inferred solidification age of each sample and age of the next oldest sample (“age break”; Fig. 12). These differences give us the maximum amount of time between intrusions in the Black Peak intrusive complex (Fig. 12). With only two exceptions, the maximum temporal break between intrusions in the Black Peak intrusive complex is <400 k.y. We note that in areas of high sampling density (the eastern part of the Stiletto Mountain unit and the Crescent Mountain unit), this gap is <∼200 k.y. The largest breaks occur between intrusion of the Crescent Mountain unit and the Stiletto Mountain unit and between the two samples collected from the War Creek unit. These gaps may represent pauses in magmatism, but they are equally likely to be an artifact of the number of samples dated. Magmatism in the Black Peak intrusive complex does not appear to have been episodic (in the way that this term is normally used), but there is no conclusive evidence that magmatism was continuous. Instead, magmatism in the Black Peak intrusive complex is proposed to have been at least “semicontinuous,” which we characterize as magmatism that makes up the bulk (>60%–70%), but not necessarily the entirety, of the duration of intrusion.

The Crescent Mountain unit (91.7–91.1 Ma) is composed of multiple small (<0.5 km2), easily differentiated intrusions varying in composition from tonalite to gabbro. Textural and compositional variations, along with our geochronology, are interpreted as evidence that the unit was constructed from a series of small intrusions. U-Pb geochronology of samples from the Crescent Mountain unit suggests assembly of the unit occurred semicontinuously over a minimum of ∼700 k.y. (Fig. 12). Apparent gaps between samples may represent brief pauses in magmatism as long as 209 ± 45 k.y., but they are equally likely to be an artifact of the number of samples dated.

Following construction of the Crescent Mountain unit, there was a maximum hiatus in magmatism of 835 ± 120 k.y., the longest recorded quiescent period in the complex, preceding intrusion of the Stiletto Mountain unit (Fig. 12). The Stiletto Mountain unit (90.3–88.8 Ma) is composed of texturally variable tonalite and minor amounts of granite. Color index and hornblende-to-biotite ratio also vary within the unit. Textural and compositional variations, along with geochronologic data from the Stiletto Mountain unit, suggest incremental, semicontinuous construction over the 1.5 m.y. lifetime of the unit; apparent gaps in magmatism could have lasted as long as 278 ± 45 k.y. (Fig. 12).

Intrusion in the Louis Lake heterogeneous zone (89.7–89.0 Ma) overlapped with the final ∼700 k.y. of assembly in the Stiletto Mountain unit (Figs. 6 and 12; Tables 1 and 2). Within this small (<2 km2) region, there are at least four unique lithologies. Magmatic features, including schlieren, enclave swarms, gradational contacts, and layering, strongly suggest multiple small (<<0.5 km2) batches of magma passed through the region. The mingled relationship between compositionally distinct magmas in this zone, combined with our geochronology, suggests incremental, semicontinuous intrusion for a minimum of 700 k.y. As the only location within the Black Peak intrusive complex with extensive evidence for magmatic interactions and ages that overlap with most of the intrusive history of the western Stiletto Mountain unit, the Louis Lake heterogeneous zone is interpreted as a preserved magmatic transfer zone (e.g., Hasalová et al., 2011). This region may have served as a “conduit” for magma migrating from the mid-crust into the shallow crust, although further work is warranted to confirm this interpretation.

After intrusion of the Stiletto Mountain unit–Louis Lake heterogeneous zone, there was a maximum hiatus in magmatism of 343 ± 78 k.y. before intrusion of the Reynolds Peak unit (88.4–87.7 Ma). The Reynolds Peak unit is composed of texturally and compositionally variable tonalite, granodiorite, and granite that show variations in the concentration of plagioclase phenocrysts and hornblende. These differences in composition and texture, along with geochronology from the unit, suggest that intrusion was incremental, semicontinuous, and lasted a minimum of ∼700 k.y.; apparent breaks in magmatism during this time may have lasted as long as 342 ± 82 k.y. (Figs. 2 and 12).

The War Creek unit (ca. 86.8 Ma) was intruded after a maximum magmatic hiatus of 291 ± 88 k.y. and represents the final magmatic episode in the Black Peak intrusive complex. The unit is composed of numerous sheets of deformed biotite tonalite separated by sharp contacts, indicating these rocks were rheologically strong when sequential magma intruded. Because of the heterogeneity and complex geochronology, it is difficult to say how long intrusion in the War Creek unit lasted. Our data suggest a minimum duration of intrusion of ∼600 k.y.

Dating Magmatic Zircons in Plutonic Samples

Many samples from the Black Peak intrusive complex show a dispersion of concordant zircon ages, a feature that has been observed in numerous other high-precision studies of plutons (e.g., Coleman et al., 2004; Matzel et al., 2006; Schaltegger et al., 2009; Memeti et al., 2010; Davis et al., 2011; Mills and Coleman, 2013; Barboni et al., 2013, 2015; Frazer et al., 2014; Samperton et al., 2015) and large crystal-rich intermediate ignimbrites (e.g., Schmitt et al., 2003; Bachmann et al., 2007; Folkes et al., 2011). As has already been noted, the meaning of this dispersion has important implications for understanding plutonism in the crust. Zircons in such a system could crystallize within a discrete pulse of magma (autocrysts), be recycled from earlier magmatic phases (antecrysts), or be inherited from host rocks that substantially predate magmatism (inherited or xenocrystic zircons; Miller et al., 2007; Cooper and Reid, 2008; Cooper, 2015).

Particularly because new pulses of magma can result in zircon growth on earlier-formed zircon crystals, the distinction between antecrysts and autocrysts is not always clear (Miller et al., 2007; Schaltegger et al., 2015). This distinction is further complicated by the nature of TIMS analyses, which requires the total dissolution of entire grains or grain fragments and the loss of information on the spatial variability in age within the crystal. Although secondary ion mass spectrometry (SIMS) U-Pb geochronology could help elucidate possible intercrystal age variations, the typical precision for rocks of this age would make this challenging unless the intercrystal age variation was much larger than that captured by the total spread of ages within samples. Novel, intracrystal microsampling techniques are being applied to deal with some of these complexities (e.g., Samperton et al., 2015). As noted previously, microsampling and U-Pb TIMS dating of crystal interiors and edges/tips did not show appreciable variations in age (outside of 2σ errors for individual zircon analyses). Additionally, if the magmas that produced the pulses that fed the Black Peak intrusive complex were formed by “remelting” (or assimilation) of only slightly older juvenile intrusions (e.g., Ratajeski et al., 2001; Annen, 2005), any zircons from these remelted/assimilated intrusions that survived anatexis and transit from source to emplacement level would be considered “inherited,” at least in the way that term is normally used. Despite the imperfections of this terminology, it has utility in conveying where zircons may have crystallized along the source–to–emplacement level continuum, and so it is used herein. The dispersion of single, concordant zircon dates within the Black Peak intrusive complex suggests the presence of autocrysts, recycled antecrysts, and true xenocrysts/inherited crystals (Figs. 4–8; Miller et al., 2007). Zircons that are older than 91.744 Ma, the nominal age of the oldest sample from the Black Peak intrusive complex (MAF-1), are considered xenocrystic or inherited.

The change in zircon population in samples from the Black Peak intrusive complex can be quantified by defining a value, Δ time, as the difference between the oldest zircon in a given sample younger than MAF-1 (the oldest sample from the Black Peak intrusive complex; 91.744 Ma), and the youngest U-Pb date in a particular sample, which is inferred to approximate the final solidification age for that sample. For sample MAF-1, Δ time is defined as the difference between the oldest and youngest U-Pb dates. These values give an estimate of the population of Black Peak intrusive complex–related magmatic zircons present in a sample (Fig. 13). Many samples with younger final solidification ages (younger than 90 Ma) show a notable increase in Δ time, even when the oldest antecrystic/autocrystic outliers are discarded (Fig. 13).

Zircon Age Dispersion and Crystallization in the Black Peak Intrusive Complex

Integration of the geochronologic, zircon trace element, and whole-rock and zircon isotopic data sets provides insight into how to interpret the spread of zircon ages observed in the Black Peak intrusive complex. Zircon trace element data from the Stiletto Mountain, Reynolds Peak, and War Creek units form linear down-temperature trends, and zircon cores with high-Ti/low-Hf concentrations are surrounded by low-Ti/high-Hf rims. These trends suggest that many zircons in each sample have similar time-temperature histories and crystallized from the same or very similar melt batches (Fig. 9). Samples with the greatest Δ time, which belong to the two youngest units in the Black Peak intrusive complex (Reynolds Peak unit, War Creek unit), have zircons with lower Ti-in-zircon model temperatures at a given Hf than the other units (e.g., Stiletto Mountain).

The increase in Δ time is unlikely to be caused by a larger component of zircon inherited from the melt source because, even with uncertainty in TiO2 activity, the Ti-in-zircon model temperatures (Fig. 9) are generally well above the zircon saturation temperatures for all of the Black Peak intrusive complex rocks (∼650–700 °C using the calibration of Boehnke et al., 2013). This suggests that magmas were zircon-undersaturated in the melt source (e.g., Harrison et al., 2007).

The spread in age observed in the Black Peak intrusive complex is likely due to incorporation of zircons that crystallized in earlier intrusive pulses (i.e., antecrysts; Miller et al., 2007; Cooper and Reid, 2008; Schaltegger et al., 2015), or to an increase in the duration of autocrystic zircon crystallization in the magma bodies. Although it is not possible to pinpoint which dated zircons are autocrysts, the overall increase in Δ time and shift toward lower Ti concentrations (lower Ti-in-zircon temperatures) in younger units of the Black Peak intrusive complex suggest an evolution of the Black Peak intrusive complex from nascent magmatism (Crescent Mountain unit) to waxing through a peak thermal period (Stiletto Mountain unit–Louis Lake heterogeneous zone) to ultimately waning of the system (Reynolds Peak–War Creek units). During the thermal peak, Black Peak intrusive complex magmatism may have sustained a mushy, emplacement level (mid- to shallow crust) system, resulting in increasingly protracted autocryst growth. However, the general shift to lower Ti-in-zircon model temperatures requires that zircon crystallized in cooler magmas and/or mush. As intrusion of the Stiletto Mountain unit and Louis Lake heterogeneous zone subsided, the overall cooler magmatic environment would have favored greater antecryst recycling and incorporation of true xenocrysts because the magmas would have been less capable of dissolving earlier-formed zircon (Reynolds Peak and especially War Creek units; Fig. 12).

Whole-rock εNd and zircon δ18O data from representative samples from each unit generally support the view that the zircons crystallized in a common magma system, as the data show no resolvable variation (i.e., outside of uncertainty; Figs. 10 and 11). This suggests that: (1) all Black Peak intrusive complex samples are a mixture of mantle-derived melt and crustal material, and (2) the zircons crystallized in magmas that were isotopically well homogenized prior to zircon saturation.

The relatively high, positive εNd and overall low δ18O values observed suggest a strong mantle signal in the magmas that fed the Black Peak intrusive complex. Using simple binary mixing models based on oxygen isotopes, it is possible to estimate the relative proportions of crustal and mantle components, assuming δ18O in zircon approximates the δ18O value in the magma from which it crystallized. Oxygen isotopic measurements for local crustal components (e.g., Twisp Valley schist) are not available. We used an altered oceanic crust value of 15‰ (Lackey et al., 2005), since it likely represents the most extreme crustal value for any plausible assimilants. Using a mantle δ18O = 5.3‰ (Valley, 2003), we estimate that crustal melt may have constituted between 10% (MAF-1) and 20% (PX10-86) in Black Peak intrusive complex magmas. This is consistent with the Sm/Nd isotopic data, which also require a large, juvenile, depleted mantle component. However, if the assimilant was instead igneous or meta-igneous crust, then the proportion of crustal component could be appreciably higher. Importantly, the values recorded by the zircon are compatible with at least some modest crustal input, and the overall oxygen isotopic homogeneity of zircon suggests that crustal assimilation and isotopic homogenization (e.g., a melting, assimilation, storage, and homogenization [MASH] process) happened prior to appreciable zircon crystallization.

Intrusion Rates in the Black Peak Intrusive Complex and the Viability of Magma-Mush Bodies

The geochronologic data suggest that early magmatism in the Black Peak intrusive complex may have been short-lived (Crescent Mountain unit), but this was followed by relatively continuous magmatism for at least 3.5 m.y. (Stiletto Mountain, Reynolds Peak, and War Creek units). Within this time interval, the field, geochronologic, and zircon geochemical data may suggest the growth and development of shorter-lived magma bodies of appreciable size, perhaps corresponding in some cases to the mapped units. This is particularly the case for the Stiletto Mountain and Reynolds Peak units, which have large areas that are relatively homogeneous and/or where internal contacts are very subtle and gradational. The viability of long-lived, mid- to shallow-crustal magma bodies that may have developed during construction of the Black Peak intrusive complex depends critically on magma intrusion rate (e.g., Schöpa and Annen, 2013; Gelman et al., 2013; Menand et al., 2015), which is difficult to accurately determine in plutons.

It is possible to estimate minimum intrusion rates for the Black Peak intrusive complex from the exposed area and the topographic relief of each unit to estimate volume and the geochronologic constraints. The area of each unit in the Black Peak intrusive complex was estimated using geographic information system (GIS) software. The topographic relief in the Black Peak intrusive complex reaches 1.4 km and establishes a lower bound on the thickness for the units, allowing a calculation of minimum volumes (Table 3). The time-averaged rate of intrusion was then calculated by dividing the estimated volumes by the difference between the crystallization age for the oldest and the youngest samples from the Black Peak intrusive complex and from each of the major units (Table 3). This yielded a minimum time-averaged rate of intrusion of 1.1 × 10–4 km3/yr for the Black Peak intrusive complex; estimates for individual map units range from 10–5 to 10–4 km3/yr (Table 3). The range of intrusion rates required to sustain an eruptible, melt-rich magma system in the shallow crust has been modeled to be on the order of 10–3–10–2 km3/yr (Gelman et al., 2013; Menand et al., 2015). Thus, sustaining large magma chambers within the Black Peak intrusive complex over long periods of time would appear prohibitive.

Looking in more detail at the geochronology, this time-averaged rate of intrusion may underestimate the actual intrusion rate when somewhat smaller spatial domains are considered. Many Black Peak intrusive complex samples show evidence for prolonged and overlapping zircon crystallization (several 105 yr), and in at least one area where the sampling density is high (eastern part of the Stiletto Mountain unit; Fig. 2B), the spread in solidification ages is well within uncertainty for samples taken over a several-kilometer length scale. The samples collected in this area (Fig. 2B) cross subtle variations in mode or mineralogy and/or texture on the hundreds of meters scale, which may represent either smaller injections into, and/or differential cooling of a tonalitic crystal mush (e.g., Miller et al., 2011) that is at least several cubic kilometers, but could be considerably larger. Based on the complete overlap of U-Pb dates in the eastern part of the Stiletto Mountain unit, it is possible that many tens of cubic kilometers to perhaps as much as 102 km3 could have been maintained in a mushy state for up to several 105 yr.

The time-averaged rate of intrusion necessarily gives a minimum estimate because constraints on pluton thickness (and therefore intrusive volume) are often poor, the volume of material removed by erosion is also often unknown, and the volume of possible cogenetic and coeval volcanic rocks is also often unknown. An intrusive thickness >>1 km for the Black Peak intrusive complex is not implausible. Vertical exposure for other Cretaceous North Cascades plutons is 1–2.5 km (comparable to Black Peak intrusive complex), which would be a minimum thickness, given that internal and external contacts are commonly steep (Miller et al., 2009a). It would be a remarkable coincidence for the Black Peak intrusive complex to be truncated just below the deepest exposure level. Estimates of pluton thickness for larger batholiths are >25 km (e.g., Ducea, 2001; Lipman and Bachmann, 2015). An intrusive thickness of 10 km for the Black Peak intrusive complex increases the time-averaged intrusion rate to 1.1 × 10–3 km3/yr, which would be sufficient to sustain a long-lived mush body in the upper crust (Gelman et al., 2013; Menand et al., 2015; Lee et al., 2015), especially considering the unaccounted-for volume from erosion and possible coeval volcanic rocks.

Comparison between the Black Peak Intrusive Complex and the Mount Stuart Batholith

Similarities between the Black Peak intrusive complex and Mount Stuart batholith, the largest Cretaceous intrusive complex in the North Cascades, have been noted as far back as Misch (1966). The Mount Stuart batholith is a shallow-crustal (6–12 km) body intruded into the Cascade core between ca. 96.3 and 90.8 Ma, representing ∼5.5 m.y. of magmatism (Fig. 1; Matzel et al., 2006). The Mount Stuart batholith is divided into four major lithologies, ranging from granodiorite to gabbro (minor hornblendite). The similar emplacement levels, partially overlapping ages (Black Peak intrusive complex ca. 92–87 Ma), and similar duration of intrusion present an excellent opportunity to compare intrusive styles in the mid- to shallow crust of a single continental magmatic arc. Geochronology of 12 samples by Matzel et al. (2006) suggested intrusion during four major episodes: (1) 96.3–95.4 Ma, (2) 94.7–94.3 Ma, (3) 92.7–92.6 Ma, and (4) 90.9–90.8 Ma. Gaps between these episodes range from ∼0.7 m.y. to ∼1.7 m.y., i.e., significantly longer than those in the Black Peak intrusive complex (∼278–835 k.y.). Geochronologic data from the Mount Stuart batholith show no apparent relationship between sample composition and crystallization age; e.g., mafic (gabbro/diorite) samples range in age from ca. 95.9 Ma to 90.8 Ma.

Detailed geochronologic data sets in both the Mount Stuart batholith and the Black Peak intrusive complex allow a comparison of intrusion rates between the two bodies (Table 3). Values for the Mount Stuart batholith were taken from Matzel et al. (2006) and were estimated in a similar way using a minimum thickness of 2.5 km, yielding a minimum time-averaged rate of intrusion of 2.2 × 10–4 km3/yr. It is important to note that the analyses of Matzel et al. (2006) predated the widespread use of the chemical abrasion technique, and individual zircon analyses are less precise than those from the Black Peak intrusive complex. With this caveat in mind, the data show that the total magmatic intrusion rates in the Black Peak intrusive complex and the Mount Stuart are comparable within a factor of two (the approximate difference between their thicknesses; Table 3). However, further comparison of geochronologic data from the Mount Stuart batholith and the Black Peak intrusive complex indicates that the two complexes appear to differ significantly in their tempo of intrusion. Overall, the Black Peak intrusive complex shows evidence for relatively continuous intrusion over the lifetime of the complex, while the Mount Stuart batholith shows evidence for brief magmatic episodes separated by intrusive lulls (Matzel et al., 2006) or possibly periods of zircon undersaturation. The number of samples analyzed by Matzel et al. (2006) is less than half of the samples analyzed from the Black Peak intrusive complex, and a higher sampling density in the Mount Stuart batholith may show evidence for more continuous intrusion in the complex. The complications in comparing the Mount Stuart batholith and Black Peak intrusive complex data sets underscore that the time-averaged intrusion rate calculations taken from plutons having complex zircon age spectra may be misleading and perhaps are not particularly useful in understanding the overall time-temperature construction history of a large intrusive complex.

CONCLUSIONS

Construction of the Black Peak intrusive complex lasted ∼4.5 m.y., from ca. 91.5 Ma until ca. 87 Ma. Magmatism was semicontinuous from at least ca. 90 Ma to ca. 87 Ma and possibly throughout the entire lifetime of the complex. Whole-rock εNd and zircon δ18O data from the Black Peak intrusive complex show no resolvable differences outside of uncertainty, suggesting that all samples from the complex represent roughly similar mixtures of mantle-derived melt and crustal material. Time-averaged intrusion rate estimates for the complex are consistent with incremental growth but have large uncertainties. Where sampling density is high enough, the zircon data yield final solidification ages for kilometer-scale regions of the Black Peak intrusive complex (Stiletto Mountain unit) that are identical at uncertainties of 0.05%, but age dispersion in individual samples can spread over several 105 yr in some cases. Comparison of the zircon saturation temperatures with the temperatures recorded by Ti-in-zircon thermometry suggests (1) that the magmas were initially undersaturated in zircon; (2) that the Black Peak intrusive complex transitioned into a thermally mature system, where parts of the complex may have remained “mushy” for long periods of time (105 yr), resulting in an abundance of autocrystic zircons, and (3) that proliferation in xenocrystic and antecrystic zircons in younger samples is the result of magmas moving upward through an increasingly complex crustal intrusive column, cooling, and crystallizing zircon and incorporating earlier-formed zircon crystals.

A comparison of the Black Peak intrusive complex and the Mount Stuart batholith suggests that the overall rate of intrusion of magma into both bodies was relatively similar. However, in detail, it appears that magmatism in the Black Peak intrusive complex was more continuous than magmatism in the Mount Stuart batholith. This may reflect the smaller sample set from the Mount Stuart batholith or natural heterogeneity in intrusive style in the shallow crust.

Careful and thoughtful reviews by Editor Shan de Silva, Associate Editor Rita Economos, and two anonymous reviewers improved the manuscript immensely. This paper benefited from conversations with Seth Burgess, Tim Grove, Oli Jagoutz, Adam Kent, Noah McLean, Catherine Mottram, and Matt Rioux. Analytical assistance from Jahan Ramezani, Nilanjan Chatterjee, Frank Dudás, Rita Economos, Axel Schmitt, and Joe Wooden is gratefully acknowledged. Good-humored and able field assistance was provided by Christine Chan, Adam Bockelie, and Kyle Gilpin. We are grateful to the Methow Valley Ranger District and the North Cascades National Park for access. This research was supported by National Science Foundation grants EAR-0948388 to Bowring, EAR-0948685 to J. Miller and R. Miller, and EAR-1119358 to R. Miller.

1Supplemental File. Explanations of analytical techniques, detailed sample descriptions, field photographs, and CL images of zircons. Please visithttp://dx.doi.org/10.1130/GES01290.S1 or the full-text article onwww.gsapubs.org to view the Supplemental File.