Multiple sediment incorporation events in a continental magmatic arc: Insight from the metasedimentary rocks of the northern North Cascades, Washington (USA) Open Access

Sediment that is incorporated into the mid to lower crust of continental magmatic arcs can significantly modify the composition and rheology of arc crust (e.g., Miller and Paterson, 2001; Hacker et al., 2011; Chin et al., 2013). As sediment is metamorphosed at (upper) amphibolite-facies conditions, hydrous minerals within the sediment will undergo dehydration reactions that may result in partial melting (e.g., Thompson, 1982; Clemens and Vielzeuf, 1987; Spear et al., 1999). Partial melting of metasedimentary rocks may also be driven by fluid infiltration from local crystallizing plutons (e.g., Whitney, 1992b; Rubatto et al., 2009; Brown, 2010; Wang et al., 2019). Metasedimentary rocks and associated partial melts are rheologically weak compared to solidified arc-related plutons and orthogneisses (e.g., Miller and Paterson, 2001; Rosenberg and Handy, 2005). Thus, the introduction of sediment into an active arc provides rheological heterogeneity within the arc crust, and strain is commonly partitioned into the weaker metasedimentary layers (Miller and Paterson, 2001; Klepeis et al., 2004). Furthermore, partial melts of metasedimentary rocks may contribute to increased periods of magmatism (e.g., Ducea and Barton, 2007; DeCelles et al., 2009; DeCelles and Graham, 2015) and to a more felsic magma component. Therefore, they may help to form andesitic-composition continental crust (e.g., Hildreth and Moorbath, 1988; Rudnick, 1995; Ducea and Barton, 2007; Reubi and Blundy, 2009; Hacker et al., 2011).

Sediment may be incorporated into continental magmatic arcs via multiple mechanisms that include underthrusting of forearc, accretionary wedge, and/or backarc material and/or via relamination or underplating of sediments off of or from the subducting slab (e.g., Haxel et al., 2002; Matzel et al., 2004; Currie et al. 2007; Ducea and Barton, 2007; DeCelles et al., 2009; Behn et al., 2011). Throughout the history of an arc, sediment may be incorporated multiple times, and the mechanism of sediment incorporation may change. For example, Pearson et al. (2017) described two discrete sediment incorporation events that formed the metasedimentary rocks of the Central Gneiss Complex within the Coast Mountains batholith, which include underthrusting of backarc sediments and, to a lesser extent, underplating of accretionary wedge sediments. Similarly, underthrusting of backarc sediments and/or the downward transport of sedimentary rocks and underplating of forearc and/or accretionary wedge sediments have both been interpreted for the Sierra Nevada batholith (Grove et al., 2003; Saleeby, 2003; Chin et al., 2013; Cao et al., 2015). More clarification is needed regarding these sediment incorporation events within arcs, the peak metamorphic conditions achieved by the sediment, and how sediment incorporation relates to the timing of high-flux magmatism and rheologic changes within the arc.

The Cretaceous–Eocene North Cascades continental magmatic arc in Washington, USA, exposes a significant amount of metasupracrustal rocks within the exhumed crystalline core (Figs. 1–2; e.g., Misch, 1966). Multiple sediment incorporation mechanisms have been invoked to explain the source of the metasedimentary units (e.g., accretionary wedge underthrusting, forearc underthrusting and underplating, and/or backarc incorporation within a fault zone; Miller, 1994; Matzel et al., 2004; Gordon et al., 2010b; Sauerx et al., 2017a, 2017b, 2019). Furthermore, the North Cascades arc is characterized by three high-magma flux events throughout the history of the arc (e.g., Miller et al., 2009b), and Sauer et al. (2017b) suggested that sediment incorporation from the forearc overlaps with the middle high-flux event. Thus, the North Cascades provides an opportunity to investigate multiple sediment incorporation events relative to high-flux events within the lifetime of an arc. Previous analytical and field work has identified multiple metasedimentary units that are closely spatially associated within the northern crystalline core of the exhumed North Cascades arc. However, the order in which the protoliths of these metasedimentary rocks were incorporated, the differences in their emplacement depths, and the contact relationships between the different metasedimentary groups remain unclear. Here we combine geologic mapping, geochemistry, detrital zircon geochronology, thermometry, and quartz-in-garnet elastic geobarometry (QuiG) to further differentiate the metasedimentary groups and to interpret their incorporation history relative to the overall evolution of the magmatic arc.

The North Cascades Range is an exhumed Cretaceous–Eocene continental magmatic arc that preserves a record of metamorphism, partial melting, and sediment incorporation during ongoing magmatism. It is the southernmost portion of the Coast Mountains batholith, which stretches from Alaska to Washington (Fig. 1A; e.g., Misch, 1966; Monger et al., 1982). Major arc magmatism began by ca. 96 Ma and occurred mainly in three high-magma flux events at ca. 96–84 Ma, ca. 78–59 Ma, and ca. 50–45 Ma (see summary in Miller et al., 2009a, 2009b, 2016). Upper amphibolite facies metamorphism and partial melting of the metasedimentary rocks was coeval with the middle high-magma flux event (Gordon et al., 2010a, 2010b; Sauer et al., 2017b).

The North Cascades crystalline core exposes these upper amphibolite facies metasedimentary rocks, orthogneisses, and plutons (Misch, 1966; Tabor et al., 1989; Whitney, 1992a; Valley et al., 2003; Miller et al., 2016). The crystalline core is bounded by Late Cretaceous–Eocene dextral strike-slip faults (Fig. 1B): the Straight Creek–Fraser fault to the west and the Ross Lake fault zone to the east. Restoring ca. 110–160 km dextral slip on the Straight Creek–Fraser fault (e.g., Misch, 1966; Tabor et al., 1989; Umhoefer and Miller, 1996; Monger and Brown, 2016) places the crystalline core directly east of ca. 110 Ma and 160–145 Ma plutons of the Coast Mountains batholith (Gehrels et al., 2009; Cecil et al., 2018). Overall, deformation within the crystalline core of the North Cascades was dominantly dextral transpressional until ca. 57–55 Ma. At this time, deformation switched to dominantly transtensional (Miller and Bowring, 1990).

Protoliths of some of the metasedimentary rocks of the crystalline core were accreted to the margin prior to arc magmatism, while others were incorporated into the arc during active magmatism. The Cascade River Schist and Napeequa Schist are both accreted units that have undergone metamorphism and deformation (Fig. 1). The Cascade River Schist consists mainly of Triassic to Late Cretaceous plagioclase–mica schists, metaconglomerates, and metavolcanics (Misch, 1966; Tabor et al., 2003; Gordon et al., 2017; Sauer et al., 2017b) while the Napeequa Schist is a metamorphosed, mid-Triassic, basalt- and chert-rich oceanic terrane (Cater and Crowder, 1967; Tabor et al., 1989; Miller et al., 1993; Brown et al., 1994; Sauer et al., 2017b; Gordon et al., 2017). Both units are host rocks for the arc (Tabor et al., 1989; Brown et al., 1994).

The metasedimentary units thought to be incorporated into the magmatic arc include the Swakane Biotite Gneiss, Skagit Gneiss Complex (subsequently called the Skagit Gneiss), and metamorphosed Methow basin strata (subsequently called meta-Methow). The Swakane Biotite Gneiss is found in the southern North Cascades and contains lithologically homogenous metasammites; these metasedimentary rocks were likely formed from one sedimentary protolith (Sauer et al., 2018). The Skagit Gneiss is a deeply exhumed metamorphic unit that consists of abundant orthogneisses and metasedimentary rocks (Figs. 1–2; Misch, 1966; Tabor et al., 1989; Whitney, 1992a; Tabor et al., 2003; Haugerud and Tabor, 2009); protolith sediments of the Skagit Gneiss were interpreted to be underthrusted from west of the arc into the North Cascades arc (Sauer et al., 2017b). Adjacent to the eastern boundary of the Skagit Gneiss are the metasedimentary rocks of the meta-Methow, which previous studies described as metamorphosed sediments from east of the arc (Kriens and Wernicke, 1990; Miller et al., 1994; Gordon et al., 2010b).

In the northern Skagit Gneiss, the Skagit rocks are structurally intermixed with the meta-Methow and portions of the Napeequa Schist (e.g., Tabor et al., 2003). Additionally, previous studies have suggested that the Skagit Gneiss consists of multiple metasedimentary units (Sauer et al., 2017b). To better understand the incorporation and metamorphic history of multiple sediment packages in the northern North Cascades magmatic arc, this study focuses on the interrelation of the Skagit Gneiss, meta-Methow, and Napeequa Schist metasedimentary units of the northern North Cascades. These units are described further below.

The Skagit Gneiss consists of tonalitic–granodioritic orthogneisses with lesser amounts of metasedimentary rocks (mapped as “banded gneiss”; Tabor et al., 2003), amphibolites, and ultramafic rocks (Fig. 2; Misch 1966). Tonalite and lesser granodiorite protoliths of the Skagit orthogneiss were emplaced between 89 Ma and 45 Ma, and the largest volume was intruded between 73 Ma and 59 Ma (Miller et al., 2016, and references therein). The Skagit Gneiss was metamorphosed to peak or near-peak metamorphic conditions (650–725 °C and 8–10 kbar; conventional thermobarometry) between 77 Ma and 65 Ma (Whitney, 1992a; Gordon et al., 2010a, 2010b; Sauer et al., 2017b). Migmatization of metasedimentary rocks and orthogneisses occurred between 68 Ma and 47 Ma (Gordon et al., 2010a). Foliation within the Skagit Gneiss generally strikes north–northwest, and lineations generally plunge shallowly southeast (Fig. 2; Wintzer, 2012; Miller et al., 2016). Map-scale, northwest-trending, gently plunging, upright to inclined folds are present throughout the crystalline core, including within the Skagit Gneiss (Fig. 2; Misch, 1966; Miller et al., 2006, 2016; Wintzer, 2012; this study). The P-T path, combined with coeval melt crystallization, ductile deformation, and cooling, suggest that the Skagit Gneiss underwent rapid, near-isothermal decompression and exhumation between 50 Ma and 45 Ma (Miller and Bowring, 1990; Whitney, 1992a; Haugerud et al., 1991; Wernicke and Getty, 1997; Tabor et al., 2003; Gordon et al., 2010b).

Previous detrital zircon work has shown that the protolith of the Skagit Gneiss may be heterogeneous. Sauer et al. (2017b) showed that detrital zircon ages from Skagit Gneiss metasedimentary rocks display two distinct patterns: some samples reveal a range of Proterozoic to Late Cretaceous ages (Group 1), whereas the second group of samples has Late Triassic to Late Cretaceous dates and lacks Proterozoic grains (Group 2). Both groups have many Late Cretaceous to Paleocene dates that are interpreted to be zircon grains that grew during metamorphism. The detrital zircon signatures are most similar to sediment found within the forearc or accretionary wedge to the North Cascades arc (Sauer et al., 2017b).

Adjacent to the Skagit Gneiss metasedimentary rocks are the meta-Methow metasedimentary rocks, which are lithologically different than the Skagit Gneiss (Miller et al., 1994) but were interpreted to have reached similar near-peak metamorphic conditions (Gordon et al., 2010b). The meta-Methow rocks are exposed in the Ross Lake fault zone between the Skagit Gneiss and the low-grade to non-metamorphosed Methow strata (see below) east of the fault zone. This unit consists of interlayered metasandstones, biotite schist, metapelites, and metaconglomerates that are intruded by metamorphosed hornblende porphyries (Kriens and Wernicke, 1990; Miller et al., 1994; Gordon et al., 2010b). Quartzite (metachert), granitic, and felsic to intermediate volcanic clasts in the metaconglomerate have been described as lithologically similar to conglomerates of the Methow terrane sediments (Miller et al., 1994). Metapelites of the meta-Methow yield near peak metamorphic temperatures and pressures of 650–700 °C and 8–10 kbar, respectively, using conventional thermobarometry (Gordon et al., 2010b). A transpressional step-over zone of the Ross Lake fault zone is interpreted to have buried these backarc sediments to deep arc depths (Gordon et al., 2010b). Similar to the Skagit Gneiss, the meta-Methow rocks were likely exhumed along a near-isothermal decompression path (Gordon et al., 2010b).

Structurally intermixed with both the Skagit Gneiss and meta-Methow metasedimentary rocks are the amphibolite-facies rocks of the Napeequa Schist (Gordon et al., 2010b). The Napeequa Schist consists dominantly of siliceous biotite schist, quartzite (metachert), and amphibolite with lesser amounts of marble and ultramafic rocks (Tabor et al., 1989; Valley et al., 2003; this study). The Napeequa Schist has been correlated with the non-metamorphosed Mississippian to Jurassic Bridge River–Hozameen terrane to the northeast of the Ross Lake fault zone (Tabor et al., 1989; Miller et al., 1994).

Low-grade to non-metamorphosed rocks that represent the backarc, forearc, and accretionary wedge of the North Cascades arc system are found on either side of the fault-bounded crystalline core (Fig. 1B; Misch, 1966). These sediments include the backarc Methow terrane, the Nanaimo Group and northwest Cascades thrust system located in the forearc, and the western mélange belt (WMB in Fig. 1B), which is representative of accretionary wedge sediments. The sediments found within these basins and within the wedge are potential protoliths of the metasedimentary units that were incorporated into the crystalline core (Sauer et al., 2017a, 2017b, 2018).

The Methow terrane (also called Methow basin and Methow sequence) is a Jurassic–Cretaceous sedimentary sequence east of the crystalline core (Fig. 1B). This sedimentary sequence is interpreted to have formed in a forearc basin (Tennyson and Cole, 1978; Mahoney, 1993; Surpless et al., 2014) but was translated to a backarc position (e.g., Miller, 1994) by ca. 91 Ma, when it was intruded by the Black Peak pluton (Fig. 1B; Shea et al., 2016). The terrane is dominated by deltaic and shallow marine sandstones, siltstones, and conglomerates (Tennyson and Cole, 1978; Mahoney, 1993) that unconformably overlie greenstone of the Lower Triassic Spider Peak Formation (Ray, 1986; Surpless et al., 2014). Some Cretaceous formations in the Methow terrane may have been sourced from the Mississippian–Jurassic Bridge River-Hozameen terrane, just west of the Methow terrane (Tennyson and Cole, 1978; Trexler and Bourgeois, 1985; Miller et al., 1994). Conglomerates are found in many formations within the Methow terrane. Those of the lower formations contain east-derived granitoid cobbles, while the conglomerates of the upper formations are rich in west-derived chert clasts (Tennyson and Cole, 1978; Trexler and Bourgeois, 1985; Surpless et al., 2014). By late Albian–Turonian time (ca. 105–90 Ma), the Methow terrane underwent contraction that led to folding and faulting, which occurred during intrabasinal magmatism (McGroder, 1989; Surpless et al., 2014).

The northwest Cascades thrust system, located in the forearc, consists of Paleozoic to Cretaceous variably metamorphosed rocks, which include island arcs, marginal basins, and oceanic seafloor that are bounded by major thrust faults (Tabor et al., 1989; Brown, 2012). The basal thrust is bracketed between 114 and 97 Ma, while the structurally highest thrust is younger than ca. 87 Ma (Brandon et al., 1988). Most units of the northwest Cascades thrust system experienced low-temperature, high-pressure metamorphism (Brown, 1987; Brandon et al., 1988). The northwest Cascades thrust system is interpreted to have accreted south of its present position and was subsequently metamorphosed, exhumed, and translated to its current latitude (Brown, 2012).

Another forearc unit is the Nanaimo Group, which is part of a Late Cretaceous–Paleocene forearc sedimentary basin west of the northwest Cascades thrust system (Fig. 1B; Mustard, 1994; Haggart et al., 2005; Englert et al., 2020). The Upper Cretaceous Nanaimo Group was folded and faulted by Eocene contraction related to the accretion of the Siletzia large igneous province (Mustard, 1994; Johnston and Acton, 2003). The lower units are alluvial and coastal marine arkosic sandstones and conglomerates with lesser mudstones (Mustard, 1994; Coutts et al., 2020). This sediment was dominantly locally sourced from the northwest Cascades thrust system and Coast Mountains batholith (Brandon et al., 1988; Mustard, 1994; Brown, 2012; Matthews et al., 2017; Coutts et al., 2020). Higher formations of the Nanaimo Group consist of locally sourced, deep marine gravity flows that exhibit discontinuous interbedding of arkosic sandstone, mudstone, siltstone, and conglomerate (Mustard, 1994; Coutts et al., 2020). By ca. 84 Ma, cratonic sediment containing Proterozoic zircon grains was deposited in these higher formations (Mustard, 1994; Matthews et al., 2017; Sauer et al., 2017a; Coutts et al., 2020).

The western mélange belt is separated from the northwest Cascades thrust system by the Darrington-Devils Mountain fault zone (Fig. 1B; Tabor, 1994). The western mélange belt is a Jurassic–Late Cretaceous accretionary complex that consists predominantly of an argillite and greywacke matrix with lenticular blocks of metagabbro, chert, marble, metadiabase, sandstone, and rare serpentinite (Frizzell et al., 1987; Tabor et al., 1989). These rocks were tectonically mixed and underwent low-grade metamorphism (Tabor et al., 1989). Similar to the Nanaimo Group, detrital zircon grains of the western mélange belt have a cratonic, Proterozoic component after ca. 80 Ma (Sauer et al., 2017a).

Previous studies have suggested ~700–3000 km of northward translation of arc-related terranes and sedimentary units along the paleomargin of North America (i.e., the “Baja-BC” or the “Mojave-BC” hypotheses; Irving et al., 1985; Umhoefer, 1987; Umhoefer and Blakey, 2006; Wyld et al., 2006; Rusmore et al., 2013; Sauer et al., 2019). Representative forearc and accretionary wedge units of the Jurassic–Eocene magmatic arcs along the entire North American paleomargin (e.g., Peninsular Ranges, Sierra Nevada, Idaho batholith, and Coast Mountains batholith) have detrital zircon signatures similar to those of the forearc and accretionary wedge units adjacent to the crystalline core and to those of the Skagit and Swakane Biotite Gneisses of the crystalline core (e.g., Sauer et al., 2017b, 2018, 2019). Detrital zircon data from many of these sedimentary deposits and the Skagit and Swakane Biotite Gneisses show a switch in sedimentary source at ca. 84–80 Ma to include Proterozoic zircon populations (e.g., Jacobson et al., 2011; Dumitru et al., 2016; Matthews et al., 2017; Sauer et al., 2017a, 2019).

The overall changes in lithologies and detrital zircon signatures through time in these different sedimentary units adjacent to the North Cascades crystalline core are important for identifying the source of multiple metasedimentary units within the Late Cretaceous–Eocene North Cascades arc.

Field work and sample collection targeted metasedimentary rocks exposed within the north-central to northeastern portion of the Skagit Gneiss (Fig. 2). This study area, ~10 × 20 km, included a portion of the Skagit Gneiss where Group 1 and Group 2 metasedimentary rocks were previously identified as well as where a portion of the Napeequa Schist and Methow terrane are interpreted to have been incorporated into and metamorphosed within the arc (meta-Methow, Group 3, Fig. 2; Kriens and Wernicke, 1990; Haugerud and Tabor, 2009; Gordon et al., 2010b; Sauer et al., 2017b). New samples as well as some samples collected by Gordon et al. (2010b) and Sauer et al. (2017b) were analyzed with a combination of geochronologic, geochemical, and thermobarometric techniques (Table 1); sample locations are shown in Figure 2.

Zircon grains from six paragneiss samples from the Skagit Gneiss and from two metapelite and two conglomerate samples from the meta-Methow were extracted using standard mineral-separation techniques. The entire zircon population from each sample was poured and mounted in epoxy to avoid bias in age populations. In addition, orthogneiss clasts were separated from one of the metaconglomerate samples, and the clasts were crushed to compare their detrital zircon signature to that of the bulk rock. Cathodoluminescence (CL) images of the zircon grains were collected using the JEOL JSM-7100FT field emission scanning electron microscope (FE-SEM) at the University of Nevada, Reno, USA (UNR). The CL images were used to identify potential core, mantle, and/or rim growth zones and to guide placement of the laser-ablation measurements.

Zircon U-Th-Pb isotopic compositions were measured on either a Nu Plasma 3-D multi-collector-inductively coupled plasma-mass spectrometer (MC-ICP-MS) with a 193 nm excimer Cetac/Photon Machines G1 laser at the University of California-Santa Barbara, USA (UCSB) using the laser-ablation split-stream (LASS) technique or an Agilent 7700 quadrupole ICP-MS with a 193 nm excimer RESOlution m-50 laser at UNR; the trace elements that were collected on an Agilent 7700X/S quadrupole ICP-MS at UCSB are not discussed here as they are not relevant to the study. Zircon cores were measured at both UNR and UCSB with a laser spot size of 25 µm, while large (>25 µm) mantle and rims were only measured at UCSB (Table S3¹). The only exceptions are the clasts from the one metaconglomerate sample; the spot size was 15 µm for these measurements, and they were completed at UCSB. Unknown measurements were bracketed every 8–10 measurements with 91500 (Wiedenbeck et al., 1995) as the calibration reference material and GJ1 (Jackson et al., 2004) or Plešovice (Sláma et al., 2008) as the validation reference material at UCSB and UNR, respectively. Data were reduced using Iolite v3 or v4 (Paton et al., 2010). All uncertainties are reported with 2σ absolute uncertainty. Detailed detrital zircon geochronology measurement methods can be found in Tables S1–S2.

A >10% discordance filter was applied to zircon measurements > ca. 400 Ma to eliminate grains affected by the inheritance of older material and Pb loss (e.g., Gehrels, 2012). The ²⁰⁶Pb/²³⁸U dates are reported except for grains >1.0 Ga, for which ²⁰⁷Pb/²⁰⁶Pb dates are reported (cf. Gehrels, 2012). Zircon core and rim results are reported in Table S3 and shown in kernel density estimate plots (KDEs; Vermeesch, 2012) in Figures 3–4. The KDEs include core, mantle, and rim measurements from single grains if they were measured (Table S3).

To better understand the protolith history and the time between deposition and metamorphism, maximum depositional ages (MDAs) were determined. To calculate MDAs, the zircon dates had to be interpreted as either detrital or metamorphic. This was completed by using a series of criteria. Rim and mantle measurements that had U/Th values >10 are suggestive of (re)crystallization due to metamorphism, an influx of fluids, and/or anatexis (e.g., Schaltegger et al., 1999; Rubatto and Gebauer, 2000; Sauer et al., 2017b) and were, therefore, considered zircon (re)crystallized during metamorphism. Core measurements, with U/Th values <10 that did not overlap with measurements of metamorphic rims/mantles, were considered to be detrital. For cores that did overlap with the measurement of rims/mantles that were interpreted to have grown during metamorphism, CL images were examined for textures indicative of metamorphic growth (e.g., resorbed boundaries, ghost zoning, and/or irregular concentric zoning) or a detrital origin (e.g., subrounded cores, truncated magmatic zoning, and/or metamict domains) (Corfu et al., 2003).This same criteria was used for cores that had U/Th values >10, for rim and mantle measurements that had U/Th values <10, and for rim and mantle measurements that had dates overlapping with core dates that were interpreted as detrital ages. Maximum depositional ages were calculated for dates that were interpreted as detrital from each sample by the Maximum Likelihood Age (MLA) method of Vermeesch (2021) using IsoplotR (Vermeesch, 2018).

Whole rock major- and trace-element geochemistry was measured for 15 samples. Samples chosen for whole rock geochemistry were not weathered and were collected a significant distance away from fractures and leucosomes to limit any compositional changes due to metasomatism or partial melting. Samples were either measured by Activation Laboratories Ltd. in Ontario, Canada, or at Pomona College following the procedure from Lackey et al. (2012). Detailed whole rock major- and trace-element geochemistry methodology can be found in Text S1 in the Supplemental Material (see ^{footnote 1}). Trace-elements were normalized to chondritic uniform reservoir (CHUR) and are displayed in rare earth element (REE) spider diagrams (McDonough and Sun, 1995). The whole rock major- and trace-element geochemistry of the samples is reported in Table S4 (see ^{footnote 1}).

Chemical compositions of garnet + biotite + plagioclase ± amphibole ± cordierite ± staurolite ± muscovite ± rutile were collected from 18 samples, including seven of the samples where zircon was dated (Table 1), using a JEOL JXA-8530FPlus electron probe microanalyzer (EPMA) at the University of Minnesota, USA (UMN) or a Cameca SX-100 electron microprobe at the University of California, Davis, USA (UC Davis). EPMA conditions and the correction methods used for the individual phases are summarized in Table S5 (see ^{footnote 1}), and the measurement method for rutile is described in Text S2 in the Supplemental Material as well. Tables S6–S12 provide representative chemical compositions for each phase in each sample.

Metamorphic temperatures were calculated using a combination of THERMOCALC average P-T software (e.g., Powell and Holland, 1988; Holland and Powell, 1998; Holland and Powell, 2011), garnet-biotite thermometry (Bhattacharya et al., 1992), Ti-in-biotite thermometry (Henry et al., 2005), and Zr-in-rutile thermometry (e.g., Zack et al., 2004; Watson et al., 2006; Ferry and Watson, 2007; Tomkins et al., 2007). For THERMOCALC average P-T and garnet-biotite thermometry, garnet cores were paired with matrix mineral cores, and garnet rims were paired with minerals directly adjacent to the garnet. An of 1 was assumed for all THERMOCALC calculations given the presence of hydrous phases (e.g., biotite, amphibole, and muscovite) in all the samples. The Zr-in-rutile temperatures were calculated using the calibration of Ferry and Watson (2007) because of the thermodynamical basis for the thermometer. Ferry and Watson (2007) thermodynamically model Zr concentrations linearly with T^–1 based on the solubility of Zr in rutile reaction versus Zack et al. (2004), who form an empirical regression line between Zr concentrations and T. A pressure correction was not applied because the rutile likely formed at <1 GPa (Tomkins et al., 2007). The samples all contain quartz, so an of 1 is used for the Zr-in-rutile calibration. The reported uncertainty is two standard deviations for each sample unless otherwise stated.

The temperature results used in the discussion section are predominantly based on the calculated Zr-in-rutile temperatures, which are not likely to reset at higher temperatures (e.g., Ewing et al., 2013). In addition, relative temperature differences among samples can be evaluated because the uncertainty is the same for all measurements (Watson et al., 2006). THERMOCALC Average P–T gives anomalous temperatures with high uncertainties for eight samples (NC-778, NC-792, SK14–11A, SK19–55, SK19–74B, SK19–123, SK19–142, and SK19–146) due to the limited mineral assemblage in the samples; therefore, they are not used in the final P–T interpretations.

Nineteen samples were measured for quartz-in-garnet (QuiG) barometry (e.g., Enami et al., 2007; Angel et al., 2014; Ashley et al., 2014); this includes the majority of the samples measured by EPMA described above plus three samples from Gordon et al. (2010b) for which EPMA data had already been collected (Table 1). Thirty- or 60-µm-thick, microprobe-quality polished sections were used. Suitable quartz inclusions were identified using an optical microscope. The criteria included: 3–10 µm in diameter for 30 µm sections or 3–12 µm in diameter for 60 µm sections; aspect ratios less than 2:1; and 2–3 radii away from both the top and bottom of the thin section or from cracks and other inclusions. Unpolarized Raman spectra were collected using a JY Horiba LabRam HR800 microscope with a 514.57 nm (100 mW source laser power) argon laser at the Virginia Tech Raman Laboratory over three collection periods following the measurement methods of Ashley et al. (2016), which are summarized here. Measurements were collected using a 100× objective, a confocal aperture of 400 μm, a 150 μm slit width, and 1800 lines/mm grating. The spectra were centered on 360 cm^–1 with a total spectral range of ~73.8–633.1 cm^–1. This allowed for simultaneous collection of the main quartz peaks (ca. 127 cm^–1, 206 cm^–1, and 464 cm^–1 at ambient conditions) along with three Ar plasma lines (at 116.04 cm^–1, 266.29 cm^–1, and 520.630 cm^–1).

The peak positions of the ca. 127 cm^–1, 206 cm^–1, and 464 cm^–1 quartz peaks were fit using the PeakFit v. 4.12 software of SYSTAT Software Inc. The peak positions of the ca. 127 cm^–1, 206 cm^–1, and 464 cm^-1 quartz peaks were fitted using the asymmetric Pearson IV peak shape following Schmidt and Ziemann (2000). Ar plasma lines were fitted using the symmetric Gaussian+Lorentzian area method. Calculated waveshifts for all quartz inclusions can be found in Table S13 (see ^{footnote 1}).

Residual pressure of the quartz inclusion (P_inc) was determined in two ways. An approach assuming hydrostatic pressure was used in which only the measured shifts and uncertainties in the 464 cm⁻¹ quartz peak were considered (). The polynomial regression used in this approach was after Ashley et al. (2016), which was based on Schmidt and Ziemann (2000). The second approach involved calculating the mean stress from the imposed strain on the quartz inclusions determined from the stRAinMAN computer program (Angel et al., 2019) and the measured shifts of the three quartz peaks (; Angel et al., 2019; Bonazzi et al., 2019).

Entrapment conditions were found using elastic modeling in the Windows GUI program EosFit7-Pinc for quartz inclusions in pure almandine, grossular, pyrope, and spessartine endmembers of garnet (Angel et al., 2017). The EosFit7-Pinc computes entrapment conditions of elastically isotropic, spherical inclusions using pressure-volume-temperature equations of state for both the host and the inclusion without assuming linear elasticity of the host-inclusion system (Angel et al., 2017). The final pressure of entrapment ( and ) was calculated using a simplified ideal mixing model of the four endmember components of garnet. Finally, an average from numerous quartz inclusions was determined for the final pressure of entrapment for each sample.

It is important to note that in a recent study by Cesare et al. (2021), the authors found that inclusions in garnet present in high-temperature granulites (equilibrated at temperatures >750 °C) were modified as a result of thermally induced grain boundary diffusion. With progressive heating, inclusion morphologies were found to change from irregularly shaped forms to dodecahedron or icositetrahedron geometries. This modification will impact the application of the QuiG geobarometer for high-temperature samples, and the resultant inclusion stress and strain will reflect post-entrapment modification. For samples from this study that equilibrated at these higher temperatures (>700 °C), it is possible that the conditions preserved by the quartz inclusions reflect a point along the cooling path (or until no further modification occurs) rather than initial entrapment conditions. Thus, samples that yield lower temperatures (<700 °C) may record garnet growth conditions along the prograde path, while higher temperature (>700 °C) samples may reflect post-modification along the retrograde path. This process is also dependent on rate; therefore, rapid, near-peak heating (e.g., Ague and Baxter, 2007; Spear et al., 2012; Viete and Lister, 2017) may have a nominal impact on inclusion resetting. As such, the samples described below that achieved temperatures >700 °C may have pressure estimates indicative of a retrograde path rather than recording entrapment conditions along the prograde path.

THERMOCALC average P-T pressures were calculated for samples with sufficient mineral assemblages. However, THERMOCALC average P-T pressures were more variable in comparison to QuiG pressures. As described for the THERMOCALC temperatures, this is likely due to the limited mineral assemblage in many of the samples. Therefore, QuiG pressures were used in the final P-T interpretations.

Overall, detrital zircon grains from the Skagit Gneiss metasedimentary rocks lack crystal faces and are rounded to subrounded with a few euhedral grains. Many grains have distinct core and rim growth with mainly oscillatory zoning and minor sector zoning in the cores. When rims are present, they usually include a thicker, CL-dark rim that may be mottled and/or a thin, CL-bright rim (Figs. 3A and S1).

Sample SK19–32 from the southern side of Ruby Mountain is a garnet–aluminosilicate metapelite that yielded 49 zircon grains; the aluminosilicate is very fine-grained, which makes it difficult to determine if it is sillimanite or kyanite. The zircon measurements form modes at ca. 423 Ma, 245 Ma, and 150 Ma (Fig. 3B) with an MDA of 94 ± 4 Ma. Five measurements from four additional grains display concordant Proterozoic dates between 2276 Ma and 1531 Ma. Most grains have low U/Th ratios of <10; two ca. 65 Ma zircon rims have much higher U/Th ratios of 37 and 109.

Sixty-six zircon grains were measured from a garnet–sillimanite metapelite (sample SK19–55) from Red Mountain. The zircon measurements form a predominant mode at ca. 67 Ma with smaller modes at ca. 143 Ma and 103 Ma and one Proterozoic zircon grain at ca. 728 Ma (Fig. 3C). Many of the 76–54 Ma zircon measurements have U/Th ratios that are >10. A 78 ± 2 Ma MDA was determined for this sample.

Sample SK19–112 from near Colonial Glacier is a garnet–staurolite paragneiss that yielded 120 zircon grains. The zircon measurements have a range of Paleozoic–Eocene dates, with ca. 354 Ma, 242 Ma, 148 Ma, and 71 Ma modes (Fig. 3D) and an MDA of 91 ± 3 Ma. Many of the grains that yield dates in the youngest population (mode ca. 71 Ma) also yield high (>10) U/Th ratios. This sample also contains scattered, concordant Proterozoic zircon grains between 1583 Ma and 559 Ma.

A garnet paragneiss, sample SK19–159, was collected west of Red Mountain. This sample yielded 98 zircon grains. The zircon measurements form two modes: ca. 142 Ma and 68 Ma. There are a few scattered concordant dates between 359 Ma and 274 Ma and one discordant Proterozoic date (Fig. 3E). Many zircon measurements between 91 Ma and 62 Ma have high U/Th ratios (>10). Sample SK19–159 has an MDA of 81 ± 2 Ma.

Sample SK06–1C is a garnet–sillimanite metapelite from near Gorge Lake that yielded 183 zircon grains (Fig. 2). The zircon measurements form modes at ca. 140 Ma and 69 Ma similar to sample SK19–159. In addition, the zircon measurements form modes at ca. 958 Ma, 448 Ma, and 240 Ma (Fig. 3F). Most 83–49 Ma zircon grains have U/Th ratios that are >10. An MDA of 85 ± 2 Ma was determined for this sample.

More than 150 zircon grains were extracted from a garnet–cummingtonite paragneiss (sample SK19–123) from southern Diablo Lake. Zircon measurements form larger modes at ca. 147 Ma and 122 Ma and smaller modes at ca. 205 Ma and 60 Ma (Fig. 3G). The majority of the zircon grains have U/Th ratios that are <10, and there are three rim measurements between 71 Ma and 60 Ma with very high U/Th ratios (>100). This sample has an MDA of 82 ± 2 Ma.

Overall, five of the Skagit Gneiss samples yield a combination of Phanerozoic and Proterozoic zircon grains and are thus classified as Group 1 metasedimentary rocks (after Sauer et al., 2017b). In comparison, sample SK19–123 from southern Diablo Lake is the only Skagit sample studied that did not yield Proterozoic zircon grains. It also contains more Early Cretaceous (ca. 140–120 Ma) zircon grains than the other samples. Thus, SK19–123 is classified as a Group 2 metasedimentary rock (after Sauer et al., 2017b). The detrital zircon measurements of these samples are in addition to the Group 1 and Group 2 samples of Sauer et al. (2017b; Table 1).

Meta-Methow metapelites generally contained fewer zircon grains than the Skagit Gneiss metasedimentary samples. However, the meta-Methow metaconglomerates contained greater than 150 zircon grains. Overall, the zircon grains from meta-Methow samples are dominantly euhedral to sub-rounded (lack crystal faces) and have oscillatory zoning and rare sector zoning in the cores; when rims are present, they are usually dark with oscillatory zoning (Figs. 4F and S1).

Sample SK19–09B is an andalusite–muscovite–cordierite metapelite from Elijah Ridge that yielded 47 zircon grains. The zircon measurements form one mode at ca. 239 Ma. Two additional zircon grains have concordant Proterozoic dates of ca. 1326 Ma and 717 Ma (Fig. 4A). Most of the zircon grains have low (<10) U/Th ratios. A 106 ± 2 Ma MDA was determined for this sample.

Zircon grains were measured from a stretched-pebble metaconglomerate (sample SK19–07) from Elijah Ridge. The clasts are orthogneiss and metachert; these clasts were too small and deformed to isolate from the bulk rock. The measurements from 148 grains form two main modes at ca. 165 Ma and 143 Ma with a smaller ca. 207 Ma mode (Fig. 4B). The majority of the zircon grains have U/Th ratios that are <10, and there is no obvious rim growth on the grains. Sample SK19–07 has a 114 ± 2 Ma MDA.

Sample SK19–15C, from Gabriel Creek, is a garnet–andalusite–staurolite–cordierite metapelite. This sample yielded 24 zircon grains that form modes at ca. 215 Ma, 157 Ma, and 117 Ma (Fig. 4C). Both core and rim measurements have U/Th ratios of <10. This sample has an MDA of 93 ± 2 Ma.

A stretched-pebble metaconglomerate from Ruby Creek (sample SK19–20) yielded a total of 167 zircon grains. The clasts are orthogneiss, metachert, and biotite schist. Seventy-six zircon grains were measured from a crushed sample of the entire metaconglomerate and form modes at ca. 231 Ma, 150 Ma, and 121 Ma (Fig. 4D). These zircon grains have low U/Th ratios of <10 and an MDA of 116 ± 2 Ma. Ninety-one zircon grains from orthogneiss clasts were separated from the matrix and crushed. These zircon measurements form modes at ca. 154 Ma and 49 Ma and have a 100 ± 2 Ma MDA (Fig. 4E); three ca. 49 Ma zircon grains have U/Th ratios >10.

These four samples are included within a third group (Group 3) based on their locations, mineral assemblages, and distinct detrital zircon signatures as compared to those of the Skagit Gneiss samples.

The detrital zircon measurements combined with previous work helped to define which metasedimentary rocks belong to Group 1 versus Group 2 in the Skagit Gneiss and a third group of metasedimentary rocks (Group 3) that is separate from the Skagit Gneiss. The three metasedimentary groups are exposed in different parts of the study area. Group 1 samples are exposed in a general U-shape on a geologic map, and Group 2 samples are located mostly within the U-shape of the Group 1 samples (Fig. 2). The map locations of Group 1 and Group 2 samples approximate the general antiformal shape of the Skagit Gneiss proposed in previous studies (e.g., Misch, 1966; Tabor et al., 1989; Miller et al., 2016) if the antiform was gently plunging south-southeast (Wintzer, 2012; this study). When projected along-strike onto cross-sections across the Skagit Gneiss, Group 2 samples are in the center of the antiform and outer edges of synforms in the more northern (A–A’) cross-section, while the projected Group 1 samples are found in the center of the western synform and on the outer edges of the antiform in the more southern (B–B’) cross-section (Fig. 5). Thus, considering the south-southeastern plunge of the main antiform, Group 2 metasedimentary rocks are in a lower structural position than Group 1 metasedimentary rocks. In comparison, Group 3 metasedimentary rocks and the Napeequa Schist crop out mainly east of the Skagit Gneiss.

There are similarities and differences in the outcrop appearances of the three metasedimentary groups. Both Group 1 and Group 2 Skagit Gneiss metasedimentary rocks are generally biotite-rich schist, metapelite, or biotite gneiss that contain rare amphibole-rich layers; Group 2 metasedimentary rocks are also associated with rare calc-silicate and ultramafic rocks. Both Group 1 and Group 2 outcrops contain variable amounts of leucocratic material (<5–50 vol %), which forms small, patchy segregations (Fig. 6A) to stromatic, foliation (sub)-parallel layers that are ~1–30 cm thick (Fig. 6B) and occasionally folded and/or boudinaged (Fig. 6C). Both Group 1 and Group 2 metasedimentary rocks are also intruded by centimeter- to meter-thick felsic dikes with sharp boundaries. Group 3 metasedimentary rocks are biotite- or amphibole-rich schist and metaconglomerate. The metaconglomerate in Group 3 contains mainly metamorphosed chert (quartzite) clasts with rare plutonic (now orthogneiss) clasts; both clast types have aspect ratios of ~2–7 (Fig. 6D). Group 3 metasedimentary rocks do not contain leucosomes.

The mica-rich metasedimentary rocks have similar mineralogy. The typical assemblage includes biotite + plagioclase + quartz ± garnet ± amphibole (hornblende or cummingtonite) ± aluminosilicate ± staurolite ± cordierite with accessory graphite + rutile + zircon + apatite ± tourmaline (Table 1). Sillimanite is locally present in Group 1 and 2 samples, while andalusite is the stable aluminosilicate in Group 3 samples. Sillimanite is generally parallel to foliation. In comparison, andalusite in the Elijah Ridge sample cross-cuts the foliation (Gordon et al., 2010b; this study), while andalusite in the Gabriel Creek sample wraps around garnet porphyroblasts. In cordierite-bearing samples, the cordierite forms coronas around the garnets and the aluminosilicates (e.g., Whitney et al., 1999; Gordon et al., 2010b; this study).

The three metasedimentary groups show an overall inverse relationship between SiO₂ with FeO, MgO, CaO, MnO, and Al₂O₃ (Fig. 7). As expected, the increase in SiO₂ composition generally corresponds to an increase in the modal volume amount of quartz and feldspar and a decrease in biotite and/or amphibole. Group 1 and Group 2 metasedimentary rocks have similar chemistry. They have comparable SiO₂ (58–67 wt%), FeO (~4–8 wt%), MgO (~1–5 wt%), MnO (~0.05–0.15 wt%), and Al₂O₃ (~15–18 wt%); however, there is more variation among Group 1 metasedimentary rocks. In addition, Group 1 samples tend to have higher K₂O (>2 wt%) and lower CaO (<~4 wt%) than Group 2 (K₂O < 2 wt%; CaO > 3.5 wt%).

Group 3 metapelite samples have chemical compositions that reflect their mineralogy; bulk rock chemistry was not determined from Group 3 metaconglomerate samples (Fig. 7). Group 3 metapelite SK19–15C from Gabriel Creek has similar major element proportions as Group 1 and Group 2 metasedimentary rocks but contains the highest SiO₂ (68.7 wt%) and lacks plagioclase. Sample SK19–09B from Elijah Ridge (Group 3) has higher FeO (12.3 wt%), Al₂O₃ (20.5 wt%), and K₂O (2.8 wt%) and lower SiO₂ (52.9 wt%) than the other metasedimentary samples. The high Al is reflected in the andalusite, cordierite, and tourmaline accessory mineralogy of this sample. The high K is likely related to muscovite in the mineral assemblage, which other samples lack. The higher Fe content is likely from the Fe-rich biotite; Fe-rich biotite is observed in nearby float sample SK19–09A (see below; Table S8). Metapelite SK19–74C has high FeO (10.7 wt%) and MnO (0.39 wt%) due to the numerous garnets (~15–20% of the sample) that are generally Fe-rich with distinctly Mn-rich cores (see below).

All three metasedimentary groups have similar REE concentrations. The samples are enriched in REEs as compared to CHUR (Fig. 8), and many have patterns similar to those of upper continental crust (UCC) but are slightly more depleted in light rare earth elements (LREEs). However, there are a few exceptions: Group 1 sample NC-775, from western Gorge Lake, has a much lower concentration of heavy rare earth elements (HREEs), and Group 3 samples SK19–09B from Elijah Ridge and SK19–74C from western slopes of Beebe Mountain are more enriched in HREEs than UCC (Fig. 8).

Major and accessory phases from all three metasedimentary groups have similar compositions. Garnets are typically porphyroblastic, and most crystals show evidence of resorption and have late cracks (Fig. 9). Garnets range from <0.5 mm (Group 1 sample SK19–159) up to 5 mm in diameter (Group 1 sample SK19–142). Group 2 sample SK15–10B contains skeletal garnets, whereas Group 3 metapelite samples generally contain euhedral garnets. Garnets in all groups contain quartz, biotite, plagioclase, rutile, and graphite inclusions; a staurolite inclusion occurs in Group 1 sample SK19–112; rutile inclusions are crystallographically oriented. Many garnets within Group 1 and Group 2 metasedimentary rocks also contain fluid(?) inclusion trails (e.g., Whitney, 1992b). Group 2 samples SK14–13 and SK19–146 have garnets with inclusion-rich cores and rims that lack inclusions. Group 3 samples SK19–74B and SK19–74C from western Beebe Mountain have garnets with the opposite patterns: inclusion-poor cores and inclusion-rich rims; both of these samples lack zircon grains. Furthermore, Group 3 sample SK19–09B, from Elijah Ridge, has spherical clusters of intergrown biotite, muscovite, quartz, and tourmaline that form a pseudomorphic texture that is likely after garnet.

Most garnets in Group 1 and 2 are almandine-rich (Alm_60–80Prp_10–25Grs_07–20Sps_01–07) with weak bell-shaped Mn and Mg zoning (Fig. 9A; Table S6). Garnets in some Group 3 metasedimentary rocks have pronounced zoning. Garnets from Elijah Ridge sample SK19–09A have an almandine-rich and grossular-poor core with a discontinuous increase in grossular content, discontinuous decrease in almandine content, and continuous decrease in spessartine content from core to rim; the pyrope content remains relatively homogenous (core: Alm₈₅Prp₀₄Grs₀₃Sps₀₈; rim: Alm₇₃Prp₀₇Grs₂₀Sps₀₀; Fig. 9B; Table S6). Similar Elijah Ridge garnet zoning has been described in Gordon et al. (2010b). The samples from western Beebe Mountain (SK19–74B,C) have garnets with notably higher spessartine and lower almandine compositions in the core than the rim (core: Alm₅₆Prp₀₅Grs₂₃Sps₁₆; rim: Alm₇₂Prp₁₄Grs₁₃Sps₀₁; Fig. 9C; Table S6). Sample SK19–15C from Gabriel Creek is the only Group 3 metasedimentary rock that has weakly zoned garnets. Garnets are more almandine-rich and slightly more grossular-rich than garnets from the Skagit Gneiss (core: Alm₇₃Prp₀₆Grs₁₇Sps₀₄; rim: Alm₇₆Prp₀₈Grs₁₆Sps₀₁; Table S6).

Plagioclase, biotite, amphibole, staurolite, and cordierite crystals in all samples are unzoned. Plagioclase compositions range from An_25–50Ab_50–75Or_00–01 among samples with no clear textural patterns (Table S7). The one exception is in SK15–10B from Colonial Glacier, which contains plagioclase with a higher anorthite composition near garnets (An_50–65Ab_35–50Or_00–01; Table S7). Biotite generally has a Mg# of 0.4–0.6; however, biotite from Elijah Ridge (SK19–09A) is Fe-rich with a lower Mg# of 0.25–0.30 (Table S8). Titanium content in biotite is variable among samples. In Group 1, matrix biotite has Ti of 0.21–0.42 atoms per formula unit (apfu). In comparison, biotite adjacent to garnets has Ti of 0.06–0.34 apfu. Group 2 samples have matrix biotite with Ti of 0.14–0.33 apfu, and biotite adjacent to garnets has Ti of 0.06–0.32 apfu. Group 3 samples have Ti of 0.17–0.35 apfu, and biotite adjacent to garnets have Ti of 0.04–0.32 apfu (Table S8). Amphibole in Skagit Gneiss samples has either cummingtonite compositions (Group 1 sample SK19–146 from Diablo Dam and Group 2 sample SK19–123 from southern Diablo Lake) or magnesio-hornblende compositions (Group 2 samples SK14–11A from eastern Gorge Lake and SK15–10B from Colonial Glacier; Table S9). Cordierite from Group 3 samples is Mg-rich (Mg# of 0.62–0.64 for SK19–74C from western Beebe Mountain and Mg# of 0.54–0.57 for SK19–15C from Gabriel Creek; Table S10). The staurolite inclusion in the garnet from sample SK19–112 from near Colonial Glacier is Fe-rich (Mg# of 0.26; Table S11). In comparison, matrix staurolite from Group 3 sample SK19–15C has less Mg (Mg# of 0.18; Table S11).

In addition to samples from the three metasedimentary groups, P–T conditions were determined for one Napeequa Schist sample (SK08273) near the Ross Lake fault zone. This sample has garnet compositions of Alm₈₅Prp₀₄Grs₀₃Sps₀₈, plagioclase compositions of An₃₃Ab₆₇Or₀₀, and pargasitic hornblende (see Gordon et al., 2010a, for thorough mineral chemistry of SK08273).

Garnet-biotite, Ti-in-biotite, and Zr-in-rutile temperatures are generally all within uncertainty of each other for each sample (Table S14; see ^{footnote 1}). Temperatures obtained from garnet cores and rims paired with matrix mineral cores and adjacent minerals, respectively, yield similar temperatures for unzoned garnets. However, for samples with zoned garnets (SK19–09A from Elijah Ridge and SK19–74B and SK19–74C from western Beebe Mountain), there is a ~80–100 °C increase in temperature from core to rim using the garnet-biotite thermometer.

Overall, Group 1 samples yield temperatures in the range of 640–800 °C (Fig. 10A; Tables 1 and S14). In comparison, Group 2 samples give a more narrow temperature range of 660–720 °C (Fig. 10B; Tables 1 and S14). Samples from Group 3 reveal slightly lower temperatures of 570–700 °C (Fig. 10C; Tables 1 and S14).

Entrapped quartz inclusions measured for QuiG pressures were generally spherical (Figs. 11A–11C), which limited any complications related to stress concentrations at points and corners of the inclusions. Furthermore, the morphologies of the quartz inclusions do not appear to have undergone high-T, post-entrapment shape modifications (Figs. 11A–11C). The quartz inclusions display variations in their Raman spectra. Samples from Group 1 and Group 2 metasedimentary rocks and Group 3 samples SK19–74B and SK19–74C from western Beebe Mountain generally have tensional quartz inclusions, in which there is a negative change in quartz peak positions (Figs. 11A, 11B, and 11D). In contrast, all Group 3 samples except for samples SK19–74B and SK19–74C have compressional quartz inclusions, in which there is a positive change in quartz peak positions (Figs. 11C–11D).

QuiG entrapment pressures calculated using the two approaches (see Methods section) provide pressures that are within uncertainty of each other (Tables 1 and S15; see ^{footnote 1}). The pressures are ~0.3–1.3 kbar lower than pressures for compressional inclusions. In comparison, the pressures are generally ~0.1–0.7 kbar higher than pressures for tensional inclusions, though there are three exceptions that have pressures that are ~0.3–0.4 kbar lower than pressures (SK15–10B from Colonial Glacier, NC-778 from Ross Lake Dam, and SK19–74C from western Beebe Mountain). The differences between pressures and pressures for tensional inclusions are statistically indistinguishable and geologically insignificant.

Group 1 samples yield QuiG pressures in the range of 5.5–7.8 kbar; SK06–91 from Ruby Mountain yields higher kbar pressures of ~6.8–9.6 kbar (Fig. 10A). The larger pressure range for sample SK06–91 from Ruby Mountain results from the larger range in estimated temperatures from Gordon et al. (2010b) (Table 1). Similar to Group 1, Group 2 samples give QuiG pressures of 5.5–7.9 kbar (Fig. 10B). Samples from Group 3 reveal higher pressures of 8.7–10.5 kbar than those of the Skagit Gneiss samples (Fig. 10D). The only exceptions are the samples from western Beebe Mountain (SK19–74B,C) that yield lower pressures of ~5.5–8.3 kbar (Fig. 10C). The meta-Methow samples with cordierite coronas around garnets (SK19–15C, SK19–74B, and SK19–74C) yield THERMOCALC average P-T pressures of ~4–6 kbar using the garnet-cordierite barometer, which is similar to samples described in Whitney (1992a) and Gordon et al. (2010b).

Differentiating the northern North Cascades metasedimentary rocks using lithologies, detrital zircon signature measurements, and near peak metamorphic estimates allows for a synthesis of the potential protoliths of the metasedimentary rocks (e.g., were the sediments more likely sourced from units east or west of the arc?) with the history of metamorphism (e.g., timing, P-T conditions, and emplacement depths). Furthermore, evaluating the similarities and differences in near-peak P-T conditions, lithology, structure, and geochemistry, in light of the detrital zircon signatures, helps to determine if the protoliths of some of the metasedimentary rocks were incorporated together and how the sediment incorporation history is connected with other arc events including magmatism and regional transpressional/transtensional deformation regimes.

Most Skagit Gneiss Group 1 and Group 2 metasedimentary rocks from this study and Sauer et al. (2017b) contain prominent Late Cretaceous–Paleocene zircon modes between 74 Ma and 60 Ma (Figs. 4B–4F; Sauer et al., 2017b). These modes are formed from mainly rim and some core measurements with commonly elevated U/Th ratios that are >10. These measurements were interpreted as metamorphic (re)crystallization of zircon grains based on the criteria outlined in the methods section. The 74–60 Ma modes constrain the timing of metamorphism of the Skagit Gneiss Group 1 and 2 metasedimentary rocks at P–T conditions of 640–800 °C and 5.5–7.9 kbar. Gordon et al. (2010a) has similarly documented ca. 68–63 Ma metamorphic and migmatization ages of the Skagit Gneiss metasedimentary rocks and associated deformed leucosomes from the Gorge Lake outcrop (sample SK06–1C).

Most Group 3 meta-Methow samples lack or have very limited metamorphic zircon growth. Group 3 metaconglomerate SK19–20, from Ruby Creek, is the exception; it yielded a small zircon mode at ca. 49 Ma and a few scattered dates between 85 Ma and 56 Ma. The ca. 49 Ma dates are consistent with igneous ages from the adjacent Ruby Creek Heterogeneous Plutonic Belt (Misch, 1966; Miller et al., 2016), while the dates between 85 Ma and 56 Ma are interpreted as mixing ages between ca. 49 Ma and detrital ages since the isotopic ratios consistently changed throughout these measurements. Metamorphism to P–T conditions of 560–700 °C and 8.7–10.5 kbar of the meta-Methow Group 3 metasedimentary rocks likely occurred during post-ca. 65 Ma regional transpression on the Ross Lake fault zone (Gordon et al., 2010b; this study). The ca. 49 Ma dates are interpreted to represent minor contact metamorphism that is related to the emplacement of the Ruby Creek intrusions rather than a regional metamorphic event. More age data from the meta-Methow and Napeequa Schist in the Ross Lake fault zone are needed to further estimate the timing of the metamorphism recorded in these rocks.

To better compare the detrital zircon signatures of the metasedimentary rocks with potential protoliths, the zircon measurements related to metamorphism were removed from the KDEs in Figure 12 and are not discussed in the following section.

The most likely sources of the protoliths of the North Cascades metasedimentary rocks include the Nanaimo Group or the northwest Cascades thrust system within a forearc position, the western mélange belt in an accretionary wedge position, or the Methow terrane within a backarc position to the arc (e.g., Matzel et al., 2004; Sauer et al., 2017a, 2017b, 2018). All three metasedimentary groups, as well as the potential sediment source bodies, contain zircon modes between 160 Ma and 145 Ma (Fig. 12). The Coast Mountains batholith experienced a period of high magmatic flux between 160 Ma and 140 Ma (Gehrels et al., 2009). Given its proximity after restoring motion on the Straight Creek-Fraser fault, the Coast Mountains batholith is a likely source of the 160–145 Ma zircon grains observed throughout the sediment sources and the metasedimentary rocks. The remainder of the detrital zircon modes in each metasedimentary group suggests some differences in their protoliths. It is possible that the protoliths of the metasedimentary rocks are no longer exposed and/or are currently not adjacent to the exhumed North Cascades magmatic arc because of large strike-slip displacements. However, the most likely sediment sources are compared below with the detrital zircon signatures and whole-rock geochemistry observed in the northern North Cascades metasedimentary rocks.

The forearc northwest Cascades thrust system is a less likely source of the North Cascades metasedimentary rocks (Sauer et al., 2017b). Two nappes in the northwest Cascades thrust system contain detrital zircon grains with scattered 2800–800 Ma Proterozoic dates (Fig. 12E; Brown, 2012); however, the northwest Cascades thrust system has dominantly unimodal Late Jurassic detrital zircon modes (Brown and Gehrels, 2007; Sauer et al., 2017a). Furthermore, the northwest Cascades thrust system detrital grains are all older than ca. 105 Ma, while many of the Skagit Gneiss samples yield younger MDAs (94–78 Ma).

The Nanaimo Group forearc sediments are broadly divided into the Coniacian(?)-early Campanian (90–84 Ma) age lower Nanaimo Group and the Campanian–Paleocene (84–63 Ma) age upper Nanaimo Group. The lower Nanaimo Group has prominent modes at ca. 90 Ma and between 156 Ma and 146 Ma (Fig. 12F; Matthews et al., 2017; Huang et al., 2019; Coutts et al., 2020; Englert et al., 2020). The lowest portions of the lower Nanaimo Group locally contain ca. 360 Ma grains (Huang et al., 2019). The higher portions of the lower Nanaimo Group have ca. 400–300 Ma grains and rare Proterozoic grains that were interpreted to be sourced from the northwest Cascades thrust system nappes (Brown, 2012; Matthews et al., 2017; Coutts et al., 2020). In comparison, most of the upper Nanaimo Group reveals similar modes at ca. 90 Ma and between 156 Ma and 146 Ma with larger ca. 90 Ma modes than the lower Nanaimo Group (Fig. 12F; Matthews et al., 2017; Coutts et al., 2020). The upper Nanaimo Group has 15–57% Proterozoic grains with prominent ca. 1700 Ma and 1380 Ma modes potentially sourced from the Belt-Purcell Basin or Mojave-Sonoran region (Matthews et al., 2017; Coutts et al., 2020). Younger sediments of the upper Nanaimo Group have 80–70 Ma modes (Matthews et al., 2017; Coutts et al., 2020).

Detrital zircon signatures from both Skagit Gneiss metasedimentary groups have similarities to those of the upper and lower Nanaimo Group (Figs. 12D and 12F). A ca. 90 Ma mode is observed in the Group 1 sample from Ruby Mountain and a Group 2 sample from eastern Gorge Lake (Sauer et al., 2017b; Figs. 12A, 12B, and 12D). Parts of the lower Nanaimo Group have a greater number of 150–100 Ma zircon grains than the upper Nanaimo Group (Coutts et al., 2020); thus, the lower Nanaimo Group may be the protolith of the Skagit Gneiss Group 2 metasedimentary rocks. In addition, the 400–300 Ma grains of the lower Nanaimo Group and the lowest portions of the upper Nanaimo Group are similar to the 448–354 Ma mode in the Skagit Gneiss Group 1 metasedimentary rocks. The distinct ca. 1700 Ma and 1380 Ma modes in the upper Nanaimo Group are also observed in one Group 1 metasedimentary sample from Ruby Mountain (Sauer et al., 2017b). However, the ca. 958 Ma grains seen in the Skagit Gneiss Group 1 metasedimentary rock (sample SK06–1C) are absent in both the upper and lower Nanaimo Groups. The cutoff used for reporting ²⁰⁶Pb/²³⁸U dates verses ²⁰⁷Pb/²⁰⁶Pb dates (see Methods section; Gehrels, 2012) affects this age more significantly than the larger Triassic and younger populations. If the ²⁰⁷Pb/²⁰⁶Pb dates are used instead of the ²⁰⁶Pb/²³⁸U dates, the measurements yield 1200–1000 Ma dates, which are more similar to young ²⁰⁷Pb/²⁰⁶Pb dates of the ca. 1380 Ma modes seen in the upper Nanaimo Group. Overall, the detrital signatures of the lowest part of the upper Nanaimo Group and the lower Nanaimo Group are most similar to those of the Skagit Gneiss Group 1 and Group 2, respectively.

The western mélange belt within an accretionary wedge position shows similarities to the Nanaimo Group and thus to both of the Skagit Gneiss groups. The western mélange belt contains prominent bimodal modes at 110–74 Ma and 170–150 Ma (Fig. 12G; Dragovich et al., 2016; Sauer et al., 2017a). Similar to the Nanaimo Group, the western mélange belt shows a shift in detrital signatures for sediments deposited after ca. 80 Ma, as these sediments include scattered Proterozoic grains as old as ca. 2.1 Ga. (Sauer et al., 2017a). However, the western mélange belt lacks the 140–120 Ma dates found in Group 2 samples and the overall variability in the detrital zircon signatures of both Skagit Gneiss metasedimentary rocks (Fig. 12G; Dragovich et al., 2016; Sauer et al., 2017a).

Both the Nanaimo Group and western mélange belt contain mudstone and/or argillite (Frizzell et al., 1987; Mustard, 1994; Coutts et al., 2020), which are likely protoliths of the Al-rich compositions in the Skagit Gneiss metasedimentary rocks (Fig. 7). The higher K₂O and slightly enriched LREEs exhibited by the Proterozoic zircon-bearing Group 1 metasedimentary rocks may be related to cratonic sediments introduced to many of the forearc and accretionary wedge units exposed along western North America after ca. 80 Ma (Mustard, 1994; Matthews et al., 2017; Sauer et al., 2017a; Coutts et al., 2020). However, both the Nanaimo Group and western mélange belt have significant lithologies that are not well correlated to the Skagit Gneiss. The Nanaimo basin has substantial interbedded sandstones and conglomerates (Mustard, 1994), and the western mélange belt includes blocks of metagabbro, chert, marble, metadiabase, sandstone, and rare serpentinite (Frizzell et al., 1987; Tabor et al., 1989; Tabor, 1994). The metamorphosed equivalents of these lithologies are rare or nonexistent in the Skagit Gneiss (e.g., Misch, 1968; this study). Therefore, if either the Nanaimo Group or western mélange belt are protoliths of the Skagit Gneiss, a portion that locally did not include much of these lithologies was incorporated.

Previous results suggest the backarc Methow terrane is a poor fit for the detrital signatures observed in the Skagit samples (Sauer et al., 2017b). The Methow terrane shows prominent bimodal detrital zircon signatures with Early Cretaceous modes at ca. 115 Ma and Late Jurassic modes at ca. 160 Ma (Fig. 12H; DeGraaff-Surpless et al., 2003; Surpless et al., 2014; Sauer et al., 2017a). Scattered 1190–278 Ma modes are also seen in compiled detrital zircon results (Surpless et al., 2014); however, <1% of all grains in the Methow terrane yield Proterozoic dates. Thus, the Methow sediments cannot account for the Proterozoic zircon modes in the Skagit Gneiss Group 1 metasedimentary rocks, and they are not likely to be a protolith of the Skagit Gneiss Group 1 and 2 metasedimentary rocks.

The meta-Methow Group 3 metasedimentary rocks share similarities with the detrital zircon signature from the younger formations of the Methow terrane rather than the forearc or accretionary wedge sediments. In particular, the meta-Methow and Methow rocks commonly have bimodal signatures with modes at 160–150 Ma and 115 Ma (Figs. 12C, 12D, and 12H). Furthermore, the 116–93 Ma MDAs of the meta-Methow samples are very similar to the Albian (109–99 Ma) MDAs of the stratigraphically highest Cretaceous units of the Methow terrane (e.g., Harts Pass Formation and Winthrop Formation; Surpless et al., 2014). Interbedded siltstones in the Cretaceous units of the Methow terrane could be protoliths of the metapelite compositions of Gabriel Creek and western Beebe Mountain samples. Furthermore, the Harts Pass Formation and Virginia Ridge Formation (another stratigraphically high Cretaceous unit of the Methow terrane) both contain conglomerates that could be the unmetamorphosed correlatives to the Elijah Ridge and Ruby Creek metaconglomerates (e.g., Miller, 1994).

The unimodal ca. 239 Ma detrital zircon mode from the meta-Methow Group 3 metapelite SK19–09B from Elijah Ridge does not obviously match with any of the sediment sources adjacent to the North Cascades arc. Furthermore, the low SiO₂ content (52.9 wt%) and higher FeO content (12.3 wt%) in this sample is in stark contrast to the content of the other metasedimentary rocks and is indicative of a more primitive protolith composition. The Albian Virginian Ridge Formation of the Methow terrane is interpreted to be sourced from the oceanic Mississippian–Jurassic Bridge River–Hozameen terrane (Tennyson and Cole, 1978; Trexler and Bourgeois, 1985), which is the unmetamorphosed correlative to the Napeequa Schist (Monger, 1986; Miller et al., 1994). The Napeequa Schist has a 260–220 Ma mode (Sauer et al., 2017b) that matches well with the ca. 239 Ma mode of sample SK19–09B from Elijah Ridge. However, this sample has a few zircon dates >300 Ma, which may suggest a different source. Thus, the age mode and the composition of Elijah Ridge sample SK19–09B may be consistent with the Virginia Ridge Formation in the Methow terrane that currently does not have detrital zircon data (despite attempts to find zircon grains; e.g., Sauer et al., 2017a), the Bridge River–Hozameen terrane, or the Napeequa Schist. Overall, the meta-Methow Group 3 detrital zircon signatures match sources east of the arc more closely.

Previous studies have suggested that all units within the northern crystalline core of the North Cascades arc, including the Skagit Gneiss and the meta-Methow unit, reached similar temperatures and pressures (650–800 °C, 8–10 kbar; Whitney, 1992a; Gordon et al., 2010b). Field relationships and the new P–T data from Group 1 and Group 2 metasedimentary rocks of the Skagit Gneiss also suggest similar near-peak metamorphic conditions that overlap the P-T results from previous studies (640–800 °C, 5.5–7.9 kbar; Fig. 10). However, the higher-pressure range of ~7.0–7.9 kbar for half of the Group 2 samples indicates that at least part of Group 2 was structurally lower than Group 1, which mostly (in five out of seven samples) yields pressures less than 7.2 kbar. This is consistent with the observed map locations in which Group 2 samples are in the core of the south-southeast plunging antiform of the Skagit Gneiss, which is currently structurally below the Group 1 samples on the limbs of the antiform. The general outline of this antiform defined by Group 1 and Group 2 samples suggests that pre-metamorphic stratigraphy may, in part, be preserved, and that foliation may be sub-parallel to parallel to sedimentary layers.

Deformation in the North Cascades crystalline core included the development of a foliation and lineation, outcrop-scale isoclinal folding, and upright map-scale folding (Wintzer, 2012; Miller et al., 2016). Both the outcrop-scale isoclinal folds and parasitic folds to the map-scale fold likely resulted in a more complex interfingering than was captured by the sampling distribution of this study. Despite earlier deformation, there does appear to be a pattern in the distribution of Group 1 and Group 2 samples throughout the Skagit Gneiss that is consistent with the main antiform mapped throughout the crystalline core.

The current structural position of Group 2 metasedimentary rocks (which lack Proterozoic zircon) below Group 1 metasedimentary rocks (which contain Proterozoic zircon) mirrors the detrital signatures of the Nanaimo basin stratigraphy. As discussed above, possible forearc and accretionary protoliths along the paleomargin of North America reveal a switch in the sedimentary source at ca. 84–80 Ma to include Proterozoic zircon populations (e.g., Jacobson et al., 2011; Sharman et al., 2015; Dumitru et al., 2016; Matthews et al., 2017; Sauer et al., 2017a, 2019). This transition is observed in the Nanaimo basin stratigraphy (e.g., higher formations contain Proterozoic zircon), whereas the western mélange belt is tectonically mixed and thus lacks coherent stratigraphy (Tabor et al., 1989; Tabor, 1994). If forearc sediments were underthrust upright into the North Cascades arc and metamorphosed without broad-scale disruption of the stratigraphy, metasedimentary rocks containing Proterozoic zircon grains (e.g., Group 1) would be structurally higher than metasedimentary rocks lacking Proterozoic zircon grains (e.g., Group 2). This is consistent with current observations within the Skagit Gneiss metasedimentary rocks. Therefore, the apparent structural relationship between the two Skagit metasedimentary groups is compatible with a Nanaimo protolith of the Skagit Gneiss metasedimentary rocks.

Additionally, the new P–T data reveal that the meta-Methow records higher pressures (8.7–10.5 kbar versus 5.5–7.9 kbar) and lower temperatures (560–700 °C versus 640–800 °C) than the adjacent Skagit Gneiss (Fig. 10D). Similarly, the Napeequa Schist sample (SK08–273) near the Ross Lake fault zone shows near-peak pressures and temperatures similar to those reached by the meta-Methow (Fig. 10C). The overall higher temperatures recorded in the Skagit Gneiss are likely due to the intrusion of magma, which formed the abundant ca. 73–59 Ma orthogneisses that are contiguous with the Skagit Gneiss metasedimentary rocks (Haugerud et al., 1991; Miller et al., 2016). The emplacement of the magma would have generated additional heat at those crustal levels; there are no significant plutons/orthogneiss of this age near the fault-bounded meta-Methow. The pressure differences between the meta-Methow, Napeequa Schist, and the Skagit Gneiss suggest that they were incorporated and/or thermally equilibrated at different levels: ~31–40 km depth for the meta-Methow and Napeequa units versus ~19–30 km depth for the Skagit Gneiss (using a bulk density of 2.7–2.9 g/cm). Subsequently, the meta-Methow and Napeequa Schist must have been transferred upward to the same structural level as the Skagit Gneiss metasedimentary rocks given their current exposure adjacent to the Skagit Gneiss.

Previous mapping by Haugerud and Tabor (2009) interprets the contact between the Skagit Gneiss with the Napeequa Schist on Elijah Ridge and western Beebe Mountain to be a low-angle, extensional fault that is likely part of a transtensional step-over zone of the Ross Lake fault zone (Miller and Bowring, 1990; Gordon et al., 2010b). However, the Napeequa Schist and meta-Methow metasedimentary rocks, which have higher metamorphic pressures (8.7–10.5 kbar), are structurally above the Skagit Gneiss rocks, which have lower metamorphic pressures (5.5–7.9 kbar); this suggests that the contact may have been transpressional at some point in the tectonic history of this fault zone (Fig. 13). Miller and Bowring (1990) demonstrated an older component of reverse slip along portions of the dextral strike-slip Ross Lake fault zone and a younger component of extension during continued dextral movement in the fault zone. Thus, it is likely that the structure along the Elijah Ridge Skagit–Napeequa contact has also experienced both transpressional and transtensional activity. However, additional age data estimating the timing of metamorphism and exhumation are necessary to evaluate which structures were active during the burial and exhumation of the meta-Methow and Napeequa Schist in the Ross Lake fault zone.

By combining proposed protolith sources, structural relationships, and P–T conditions, a more detailed sediment incorporation history of the Skagit Gneiss and meta-Methow metasedimentary rocks can be evaluated. Within the North Cascades magmatic arc, major mid-Cretaceous contraction overlapped magmatism, which began at ca. 96 Ma (Tabor et al., 1989; Miller and Paterson, 2001; Miller et al., 2009a, 2009b). By ca. 84–80 Ma, the protolith of the Skagit Gneiss Group 1 metasedimentary rocks was inferred to have been deposited on top of the Skagit Gneiss Group 2 protolith in a forearc basin (Fig. 14A). A portion of these sediments were underthrust upright as a relatively coherent package into the active magmatic arc, and reached at least ~19–30 km depth, by 74–60 Ma as documented by metamorphic zircon rims (Fig. 14B; Sauer et al., 2017b; this study). This underthrusting occurred dominantly during regional transpression in the North Cascades, which started as early as ca. 73 Ma (Paterson et al., 2004). Sediments may have been partially imbricated and folded together during incorporation, causing some of the overlap in pressure data observed in Group 1 and 2 samples. Thus, there may be a more complicated field relationship between the two “groups” at a finer sampling scale. Underthrusting of the forearc sediments was coeval with a high magmatic flux event at 78–59 Ma (Miller et al., 2009a, 2016).

Transpression in the North Cascades was followed by transtension beginning at ca. 57–55 Ma (Miller and Bowring, 1990). The eastern-bounding Ross Lake fault zone records transpressional slip by ca. 65 Ma that transitioned to transtension after ca. 57 Ma (Miller and Bowring, 1990). As previously suggested, a post-ca. 65 Ma transpressional step-over zone of the Ross Lake fault zone incorporated a sliver of the protoliths of the meta-Methow and Napeequa Schist units to depths of ~31–40 km (Figs. 14C–14D; pressures of 8.7–10.5 kbar; Gordon et al., 2010b; this study). As described above, the pressure data on Elijah Ridge also suggest that the structure that separates the Skagit Gneiss and Napeequa Schist/meta-Methow may have had a thrust-sense. This indicates that a transpressional shear zone may have juxtaposed the Napeequa Schist/meta-Methow rocks with the Skagit Gneiss during an early stage of exhumation that is estimated at depths of ~19–30 km based on the pressures of 5.5–7.9 kbars of the Skagit Gneiss (Fig. 13).

Shortly after the North Cascades transitioned to a transtensional regime at ca. 57 Ma, the final high magmatic flux event occurred at ca. 50–45 Ma (Eddy et al., 2016b; Miller et al., 2016). The two largest plutons (Cooper Mountain and Golden Horn) intruded into and across the Ross Lake fault zone (Eddy et al., 2016a), and much smaller Eocene sheets intruded the Skagit Gneiss (Miller et al., 2016). Within the Skagit Gneiss, shallowly plunging, upright folding was coeval with the Eocene magmatism and affected both Group 1 and Group 2 metasedimentary rocks (Fig. 14E; Wintzer, 2012; Miller et al., 2016; Shea et al., 2016). Intrusion of the ca. 48 Ma Golden Horn batholith likely resulted in a Buchan metamorphic overprint of the nearby meta-Methow and Napeequa Schist (Misch, 1966). This is texturally supported by the late andalusite and the evidence for multiple garnet growth events within the Elijah Ridge sample. Shortly after this intrusion, the Skagit Gneiss, meta-Methow, and Napeequa Schist were exhumed together along a near-isothermal decompression path as evidenced by cordierite replacing garnet, sillimanite, and andalusite in samples within the Skagit Gneiss, Napeequa Schist, and meta-Methow, and the garnet pseudomorph in the sample from Elijah Ridge (Whitney, 1992a; Whitney et al., 1999; Gordon et al., 2010b). By ca. 50–45 Ma, the metasedimentary rocks had rapidly cooled through temperatures of ~300 °C (Engels et al., 1976; Wernicke and Getty, 1997; Tabor et al., 2003).

As described above, the North Cascades magmatic arc reveals three high magmatic flux events (Miller et al., 2009a, 2009b). Metamorphism of the Skagit metasedimentary rocks at ca. 77–65 Ma is coeval with the second high-flux magmatic event and migmatite crystallization in parts of the Skagit Gneiss (Gordon et al., 2010a; Sauer et al., 2017b; this study). Migmatites in the Skagit Gneiss were formed by both subsolidus segregation and fluid-present partial melting (e.g., Whitney, 1992b; Whitney and Irving, 1994). Estimated metamorphic conditions of the Skagit Gneiss metasedimentary rocks were insufficient for biotite breakdown (e.g., temperatures >800 °C; Clemens and Vielzeuf, 1987; Whitney, 1992a) as evidenced by the abundant biotite present within the metasedimentary rocks; instead, the water-rich conditions necessary to produce partial melting were more likely derived from pluton crystallization (Whitney, 1992b; Whitney and Irving, 1994). This migmatization is not interpreted to have generated large volumes of partial melt (Whitney and Irving, 1994) because it would be limited by the available fluid reservoir (e.g., Clemens and Vielzeuf, 1987; Brown, 2010). Therefore, the partial melts produced by the Skagit Gneiss metasedimentary rocks likely did not drive the high magmatic flux event. However, the emplacement of the magmas associated with the middle high-magma flux event would have added heat to the crust and likely contributed to the higher temperatures recorded in the Skagit Gneiss in comparison to those of the meta-Methow rocks and the Napeequa Schist within the Ross Lake fault zone.

In comparison to the Skagit Gneiss metasedimentary rocks, metamorphism of the meta-Methow metasedimentary rocks was coeval with the youngest portions of the second high magmatic flux event (Miller and Bowring, 1990; Miller, 1994). The ca. 65 Ma Oval Peak batholith is the only large magma unit emplaced near the meta-Methow during ca. 65–57 Ma transpression of the Ross Lake fault zone (Miller and Bowring, 1990); however, this batholith is considerably (~25 km) south of the Elijah Ridge meta-Methow and Napeequa rocks. Therefore, significant magmatism directly associated with the incorporation of the meta-Methow metasedimentary rocks is not currently recognized.

Overall, the incorporation of the metasedimentary rocks in the northern North Cascades crystalline core does not appear to have generated significant magmatism. However, the coeval magmatism and sediment incorporation events may suggest the opposite causal relationship.

The metamorphic temperatures and depth of burial indicate a paleo-geothermal gradient of ~28 °C/km for the Skagit Gneiss metasedimentary rocks (~700 °C at ~25 km depth) and ~18 °C/km for the meta-Methow metasedimentary rocks (~625 °C at ~35 km depth). Miller et al. (2009b) proposed a connection between advective heat from magmatism and rapid crustal thickening to explain the relatively low paleo-geothermal gradient recorded in the mid- to deep-crust of the North Cascades magmatic arc. The abundant magma intruded during the second high-volume magmatic event, within the Skagit Gneiss and other areas of the crystalline core, likely thermally weakened the arc crust temporarily. This thermal weakening of the crust could have facilitated the incorporation of the Skagit Gneiss and the meta-Methow protoliths during underthrusting and step-over transpressional faulting, respectively.

A combination of previous and new field relationships, detrital zircon signatures, P–T data, and geochemical data provide insights into two sediment incorporation events that affected the North Cascades magmatic arc. The combined data reveal different sediment sources, depths of burial, and overall mechanisms of sediment incorporation. First, the Skagit Gneiss metasedimentary rocks can be divided into two groups that reflect relatively undisturbed depositional layers of their forearc source. Sediment containing Proterozoic zircon grains (source of Group 1) was deposited over the protolith sediment lacking Proterozoic zircon grains (source of Group 2) starting at ca. 84 Ma in a forearc basin. The sedimentary package was underthrusted, upright and with minor mixing, into the active North Cascades arc by ca. 77–65 Ma to temperatures and pressures of 640–800 °C and 5.5–7.9 kbar (ca. 19–30 km depth).

Second, metasedimentary rocks exposed within the Ross Lake fault zone (Group 3) are interpreted to have been similar to backarc terrane sediments based on their lithology and detrital zircon signatures. The backarc sediments were incorporated along a transpressional step-over zone in the Ross Lake fault zone between ca. 65 Ma and 57 Ma. Metamorphism of the meta-Methow metasedimentary rocks occurred at higher pressures and lower temperatures (560–700 °C, 8.7–10.5 kbar; ca. 31–40 km depths) than those of the Skagit Gneiss metasedimentary rocks. After burial, the meta-Methow metasedimentary rocks were transferred upward in a transpressional fault zone to similar structural depths as the Skagit Gneiss.

Sediment incorporation was coeval with other processes that affect magmatic arc evolution such as regional transpression and periods of high magmatic fluxes. Regional transpression facilitated sediment incorporation by underthrusting forearc sediments into the arc and locally burying sediments from a backarc setting in a transpressional step-over zone. Furthermore, heat from the ca. 78–59 Ma magmatic flux likely weakened the arc crust, which would further facilitate sediment incorporation. In addition, these magmas, once crystallized, provided the fluids for partial melting of the metasedimentary rocks, which further weakened the arc crust.

Overall, the ubiquity of metasedimentary units in continental magmatic arcs (Miller and Snoke, 2009) suggests that sediment incorporation is significant in the overall evolution of arc crust. This and other studies have suggested that discrete sediment incorporation events occur multiple times throughout the history of continental magmatic arcs (e.g., North Cascades [Sauer et al., 2017b, 2018; this study], Central Gneiss Complex [Pearson et al., 2017], and Sierra Nevada batholith [Grove et al., 2003; Saleeby, 2003; Chin et al., 2013; Cao et al., 2015]). These sediment incorporation events are associated with major changes in the arc that include accretion, high strain rates, foreland thrusting, high magma addition rates, and/or transpression. (e.g., DeCelles et al., 2009; Cao et al., 2015; Sauer et al., 2017b). Thus, it is necessary to recognize the multiple periods of sediment incorporation to better understand and discern causal relationships among arc events in both ancient and/or active continental magmatic arcs.

We thank Joel DesOrmeau for assistance in sample preparation, obtaining mineral energy-dispersive detector and zircon-cathodoluminescence images, and U-Pb measurements at the University of Nevada, Reno. We also thank Andrew Kylander-Clark for help with U-Pb measurements at the University of California, Santa Barbara; Anette von der Handt of the University of Minnesota and Nick Botto of the University of California, Davis, for assistance with electron probe microanalyzer data collection; Jade Star Lackey and Jonathan Harris of Pomona College for help with whole-rock geochemical measurements; and Stephen Hanson for assisting in the field. We also thank William Matthews and an anonymous reviewer, Associate Editor Craig Jones, and Science Editor David Fastovsky for their helpful comments on this manuscript. This research was funded by National Science Foundation Grant EAR-1419810 to S.M. Gordon and EAR-1419787 to R.B. Miller, U.S. Geological Survey EDMAP grant G19AC00219 to S.M. Gordon, and a Geological Society of America Graduate Student Research grant to A.E.H. Hanson.