The Swakane Gneiss, interpreted to represent sedimentary strata metamorphosed at 8–12 kbar, is the deepest exposed crustal levels within the exhumed North Cascades continental magmatic arc, yet the nature and age of its protolith and the mechanism by which it was transported to deep-crustal levels remains unclear. Zircons from 11 paragneiss and schist samples were analyzed for U-Pb age and Hf-isotope composition in order to investigate the tectonic history of the Swakane Gneiss from protolith deposition to metamorphism within the North Cascades arc. Zircons interpreted to have crystallized in situ during metamorphism and/or melt-crystallization within the Swakane Gneiss at depth have ca. 74–66 Ma ages. Detrital-zircon age and Hf-isotope characteristics demonstrate provenance shifts that correlate with maximum depositional ages of ca. 93–81 Ma. Samples deposited between ca. 93 and 88 Ma have dominantly Mesozoic age peaks with initial εHf values between depleted mantle and chondritic uniform reservoir (CHUR), whereas ca. 86–81 Ma sample show the addition of distinct Proterozoic populations (ca. 1380 and 1800–1600 Ma) and Late Cretaceous zircons with unradiogenic Hf-isotope compositions. Similar detrital-zircon age and Hf-isotope patterns are observed in several Upper Cretaceous forearc and accretionary wedge units between southern California and Alaska along the North American continental margin. The connection between the Swakane Gneiss and these potential protoliths located outboard of Cordilleran arc systems indicate burial by either underplating of accretionary-wedge sediments or underthrusting of forearc sediments. Therefore, the protolith and incorporation history for the Swakane Gneiss is likely similar to those of deep crustal metasedimentary units elsewhere in the North Cascades (i.e., the Skagit Gneiss Complex) and to the south along the continental margin (i.e., the Pelona-Orocopia-Rand schists and Schist of Sierra de Salinas). These observations suggest that burial of sediment outboard of continental magmatic arc systems may be a major mechanism for the transfer of sediment to the deep levels of continental arcs.


Metasedimentary rocks exposed in the exhumed roots of continental magmatic arcs provide evidence of the vertical and horizontal transfer of supracrustal material at convergent margins (e.g., Grove et al., 2003; Matzel et al., 2004; Ducea et al., 2009; Cristofolini et al., 2012; Chin et al., 2013). Geochronologic studies reveal information about the path these rocks followed from deposition to burial and peak metamorphism to exhumation, thus tracking the history of the arc system and potentially the continental margin (Jacobson et al., 2011; Gatewood and Stowell, 2012; Chapman et al., 2016).

Although intact metasedimentary units are rarely exposed within exhumed segments of the middle to lower crust of continental magmatic arcs, there is widespread isotopic and trace-element evidence that suggests significant supracrustal materials are present at depth and interact with mantle melts to form arc magmas (Plank and Langmuir, 1993; Ducea and Barton, 2007). In addition, some continental magmatic arcs, such as the North Cascades in Washington, contain mid-crustal metasedimentary rocks that are strongly deformed and locally migmatitic (Misch, 1966; Tabor et al., 1989). The presence of metasedimentary rocks in the deep levels of continental arcs is proposed to have wide-ranging effects, including creating strength contrasts (Miller and Paterson, 2001) and fueling pulses of high-flux magmatism (Ducea and Barton, 2007; DeCelles et al., 2009). For example, metasedimentary rocks that occur in a large range of lithospheric levels, from the mid-crust to upper mantle, may experience conditions favorable to produce significant volumes of melt, which may potentially contribute to pluton bodies and/or localize weakness in the crust (Whitney et al., 2004; Rosenberg and Handy, 2005; Behn et al., 2011).

Given these effects, documenting how metasedimentary rocks are incorporated into arc systems is important for understanding how new crust is created in continental arc settings and for the evolution of the arc system. Sediment may reach mid-crustal to upper-mantle depths through several mechanisms. This may include underthrusting of material from the forearc and retroarc sides of the arc system (Ducea, 2001; Barth et al., 2003; DeCelles et al., 2009). Additionally, accretionary-wedge sediments or eroded upper-plate material may be attached to the down-going slab and either underplated or relaminated at the base of the arc crust (Scholl and von Huene, 2007; Hacker et al., 2011). Thus, connecting metasedimentary rocks with protoliths that originated as forearc, backarc, or accretionary wedge units can provide insight into the mechanisms that are likely responsible for sediment incorporation. As sedimentary provenance information is largely erased for sediments that were assimilated into arc magmas, studies of the intact metasedimentary units in the deep levels of arc systems can be used to differentiate among processes that transfer sediment to depth.

The tectonic history of a metasedimentary unit within a continental magmatic arc can be examined in the Swakane Gneiss, which is compositionally homogeneous quartz–biotite paragneiss metamorphosed to upper-amphibolite facies (Cater, 1982; Tabor et al., 1987). The Swakane Gneiss currently occurs at the deepest exposed crustal levels in the North Cascades continental magmatic arc (Valley et al., 2003). Previous studies have interpreted this unit to have an Upper Cretaceous psammitic protolith (Matzel et al., 2004; Gatewood and Stowell, 2012) that was buried to mid-crustal depths at rates of either ∼7 km/m.y. (Matzel et al., 2004) or ∼1–3 km/m.y. (Gatewood and Stowell, 2012). Overall, the Swakane Gneiss is anomalous in structural setting, protolith age, and tectonic history within the North Cascades. The accreted oceanic supracrustal and oceanic arc-related clastic units that form the framework for the North Cascades arc are characterized by mostly Triassic to Lower Cretaceous protolith ages and Mesozoic detrital-zircon age peaks (Tabor et al., 1989; Miller et al., 1994; Gordon et al., 2017; Sauer et al., 2017b). In contrast, the provenance of the Swakane Gneiss protolith includes abundant Mesozoic and Proterozoic detrital zircons with a continental origin (Matzel et al., 2004; Gatewood and Stowell, 2012).

In this study we determined U-Pb ages and Hf-isotope compositions of zircons from 11 samples of the Swakane Gneiss. Analyses of both detrital zircons and zircons that (re)crystallized while the Swakane resided at up to 40 km depth in the active, arc crust are used to understand the sedimentary provenance and depositional age of the Swakane protolith and the timing and rate of protolith burial. The detrital-zircon characteristics of the Swakane protolith provide insight into the mechanism of sediment burial through comparison with potential protolith units. The results are compared to other examples of metasedimentary rocks in the deep levels of North American Cordillera arc segments and suggest that the transfer of upper-crustal material from the forearc side of the arc system is a major mechanism to introduce sediment into the deep levels of continental arc systems.


The Swakane Gneiss is exposed in the North Cascades Range, which is the southernmost extension of the Coast Plutonic Complex that stretches >1500 km from Alaska to Washington (Fig. 1A). The Coast Plutonic Complex represents an ancient continental magmatic arc that was active from ca. 155 to 45 Ma (e.g., Gehrels et al., 2009). Arc magmatism was focused in the North Cascades from ca. 96 to 45 Ma (Mattinson, 1972; Haugerud et al., 1991; Brown and Walker, 1993; Miller et al., 2009), and a combination of magma emplacement and/or terrane accretion resulted in a crustal thickness of >55 km by the Late Cretaceous (Miller and Paterson, 2001).

The crystalline core of the North Cascades is exposed between two major dextral strike-slip structures: the Eocene Straight Creek fault to the west and the Late Cretaceous(?)–Eocene Ross Lake fault zone to the east (Misch, 1966; Miller and Bowring, 1990; Haugerud et al., 1991; Miller, 1994; Umhoefer and Miller, 1996) (Fig. 1B). The post-metamorphic, high-angle Entiat fault divides the crystalline core into the Wenatchee and Chelan blocks. The Swakane Gneiss is exposed in both blocks along with thick, arc-derived clastic rocks of the Cascade River Schist (Tabor et al., 2002) and the oceanic accretionary Napeequa Schist (Tabor et al., 1987, 2002) (Figs. 1B and 2). The Nason terrane is limited to the Wenatchee block and consists primarily of psammitic and pelitic schist. The Skagit Gneiss Complex, which is composed of tonalitic orthogneiss with lesser metasupracrustal rocks (Misch, 1966), is limited to the Chelan block (Figs. 1B and 2). Plutons with crystallization ages between ca. 96 and 84 Ma intrude both blocks, but the Chelan block also records younger (ca. 76–45 Ma) magmatism (Tabor et al., 1989; Haugerud et al., 1991; Miller et al., 2016).

The Swakane Gneiss occupies the deepest exposed structural level in the North Cascades and preserves peak pressures and temperatures of 8–12 kbar and 640–750 °C (Valley et al., 2003; Gatewood and Stowell, 2017). The Swakane consists of quartz + plagioclase + biotite ± garnet ± muscovite gneiss (∼90%) with rare amphibolite, metapelite, metaperidotite, and quartzite (Cater, 1982; Tabor et al., 1989). In both the Chelan and Wenatchee blocks, the Swakane Gneiss is in fault contact with the structurally overlying Napeequa Schist, and the base of the gneiss is not exposed. Foliations strike generally northwest, and lineations plunge shallowly (Crowder et al., 1966; Cater and Crowder, 1967; Miller et al., 2006). In the Chelan block, the Dinkelman décollement separates the Swakane Gneiss from the Napeequa rocks, and the Swakane forms a north-trending, gently plunging antiform that exposes 1.1 km of structural thickness (Paterson et al., 2004) (Fig. 2). Peak pressure estimates for the Swakane Gneiss in the Chelan block range from 9 to 12 kbars (Valley et al., 2003; Gatewood and Stowell, 2017). In the Wenatchee block, the Swakane Gneiss crops out along the western side of the Entiat fault and, to the south, as a horst-like body in the Eocene Chumstick Basin (Fig. 2). Peak pressure estimates are slightly lower than in the Chelan block, and range from 8 to 10 kbars (Valley et al., 2003; Gatewood and Stowell, 2017). The western contact with the overlying Napeequa complex is an unnamed fault. A thin swath of rocks mapped as the Holden unit of the Cascade River Schist separates the Swakane Gneiss from the Entiat fault (Cater and Crowder, 1967). The northernmost exposure of the Swakane in the Wenatchee block is separated from the main Swakane Gneiss body by the Miocene Cloudy Pass batholith (Tabor et al., 2002) (Fig. 2).

Early studies identified the enigmatic nature of the Swakane Gneiss in comparison to the dominantly Mesozoic Cascades core based on Middle to Late Proterozoic whole-rock Nd model ages and the presence of Proterozoic zircons (Mattinson, 1972; Rasbury and Walker, 1992). Matzel et al. (2004) analyzed zircons using isotope-dilution thermal ionization mass spectrometry (ID-TIMS) and interpreted a ca. 72 Ma maximum depositional age for the protolith of the Swakane Gneiss. Coupled with a ca. 68 Ma ID-TIMS leucogranite crystallization age, these results indicate rapid burial of the Swakane protolith within ∼4 m.y. (Matzel et al., 2004). Gatewood and Stowell (2012) used Sm-Nd garnet and U-Pb zircon geochronology to evaluate the Swakane burial history. Their results indicated that garnet crystallized between ca. 74 and 66 Ma (Sm-Nd isochron dates) and metamorphic zircon populations crystallized between 75 and 63 Ma (U-Pb zircon dates, laser ablation–inductively coupled plasma–mass spectrometry, LA-ICP-MS). They interpreted ca. 91 Ma dates to correspond to the youngest population of detrital zircons and ca. 81 Ma dates to the result of mixing between the detrital and metamorphic populations. Based on these results, they interpreted that burial of the sedimentary protolith of the Swakane Gneiss occurred between ca. 91 and 75 Ma. Matzel et al. (2004) and Gatewood and Stowell (2012) also identified Proterozoic detrital-zircon populations between ca. 1400–1300 and 1800–1600 Ma.

Other Metasupracrustal Components of the North Cascades

In both the Chelan and Wenatchee blocks, the Napeequa Schist is structurally above—and separated by—a fault from the Swakane Gneiss (Figs. 1B and 2). The Napeequa Schist is dominantly composed of amphibolite and quartzite (metachert), with lesser biotite schist, ultramafite, marble, and calc-silicate rock. It is interpreted to represent an oceanic accretionary complex (Cater, 1982; Tabor et al., 1987, 1989, 2002). The protolith age of the Napeequa Schist is poorly understood, but is likely Paleozoic–Jurassic, based on correlation with the Mississippian to Jurassic Bridge River–Hozameen terrane (Misch, 1966; Miller et al., 1993) and Triassic protolith ages for amphibolites (metabasalts) (Sauer et al., 2017b). The Napeequa Schist records similar pressures and temperatures in both the Wenatchee and Chelan blocks (9–11 kbar and 640–740 °C; Valley et al., 2003) and was likely at depth during the ca. 92 Ma intrusion of the 7–10 kbar Tenpeak pluton (Dawes, 1993; Matzel et al., 2006).

The Napeequa Schist is complexly intercalated with the dominantly clastic Cascade River Schist (Figs. 1B and 2). The Cascade River Schist consists of plagioclase–mica schist, metaconglomerate, and amphibole-rich schist, with lesser silicic schist, marble, and amphibolite (Tabor et al., 1989; 2002) that is spatially associated with the Triassic Marblemount–Dumbell plutonic belt. The Cascade River Schist also includes the Holden unit and Twentyfive Mile unit of Miller et al. (1994) and Tabor et al. (1987), respectively. The protolith of the Cascade River Schist is interpreted to be a thick pile of arc-related clastic rocks with lesser arc volcanic rocks (Tabor et al., 2002). Based on its association with the Marblemount–Dumbell plutonic belt, the Cascade River–Holden unit has been interpreted as a Triassic arc assemblage (i.e., Tabor et al., 1989), but recent geochronology of metasedimentary and metavolcanic units indicates that a portion of the unit is Cretaceous in age (Gordon et al., 2017; Sauer et al., 2017b).

The White River shear zone separates the Nason terrane from the overlying Napeequa Schist in the Wenatchee block (Figs. 1B and 2). The Nason terrane includes the Chiwaukum Schist and Nason Ridge Migmatitic Gneiss (Tabor et al., 1987; 2002). The Chiwaukum Schist is dominantly metapsammitic gneiss and schist; amphibolite and metaperidotite are minor components, but interspersed throughout the unit, and marble occurs locally (Tabor et al., 1987; 2002). The Nason Ridge Migmatitic Gneiss has been extensively intruded by magmas and is thus thought to be a deeper equivalent to the Chiwaukum Schist (Tabor et al., 1987, 2002; Paterson et al., 1994; Stowell et al., 2007). The detrital-zircon signature of the Chiwaukum Schist is characterized by Early Cretaceous and Jurassic age peaks with some Precambrian dates (Brown and Gehrels, 2007; Paterson, 2014). The Early Cretaceous protoliths of the metasupracrustal units are interpreted to have been assembled in an accretionary complex (Paterson et al., 1994). The Chiwaukum Schist records metamorphic pressures and temperatures of 3–9 kbars and 540–700 °C (Brown and Walker, 1993; Stowell and Tinkham, 2003; Stowell et al., 2007).

The Skagit Gneiss Complex (referred to hereafter as the Skagit Gneiss) is located in the northern to central parts of the Chelan block and records pressures and temperatures of 8–10 kbars and 650–725 °C (Whitney, 1992; Gordon et al., 2010a) (Fig. 1B). The Skagit Gneiss consists dominantly of tonalitic orthogneiss with lesser amphibolite and biotite gneiss and rare ultramafite, calc-silicate rock, and metapelite (Misch, 1966; 1968). Deformed plutons of the Skagit Gneiss intrude Napeequa Schist lithologies in the northern Chelan block, and Skagit Gneiss metasupracrustal rocks exist as rafts and lenses within voluminous orthogneiss (Misch, 1966; Tabor et al., 1989, 2003). The protoliths of metasedimentary rocks throughout the Skagit Gneiss have maximum depositional ages between ca. 134 and 79 Ma (Sauer et al., 2017b).

Relevant Units Outside of the Crystalline Core

Low-grade to non-metamorphosed units that flank the North Cascades crystalline core may potentially represent analogues to the protolith of the Swakane Gneiss. The Straight Creek fault separates low-grade to non-metamorphosed rocks of the northwest Cascades system, western mélange belt, and the Nanaimo Group from the crystalline core (Fig. 1B). The northwest Cascades system is composed of a stack of nappes that contain diverse Jurassic and Lower Cretaceous oceanic rocks, volcanic arc fragments, and clastic rocks (Misch, 1966; Brandon et al., 1988; Tabor et al., 1989, 2002; Brown and Gehrels, 2007). The western mélange belt is interpreted to represent a Cretaceous(?) accretionary complex (Jett and Heller, 1988). The Nanaimo Group is a thick pile of Turonian–Maastrichtian marine and nonmarine sediments thought to have been deposited in a forearc basin to the Coast Plutonic Complex (Mustard, 1994; Mustard et al., 1995; Matthews et al., 2017).

To the east of the crystalline core, rocks of the Methow terrane are separated from high-grade rocks by the Ross Lake fault zone (Fig. 1B). The Methow block consists of Jurassic–Cretaceous sediments deposited unconformably on Triassic oceanic rocks. Cretaceous Methow terrane sediments are interpreted to have been deposited in a forearc basin (DeGraaff-Surpless et al., 2003), but were juxtaposed with the eastern boundary of the crystalline core by the ca. 91 Ma intrusion of the Black Peak batholith (Miller, 1994; Shea et al., 2016). The Methow terrane is bound on its east side by the Pasayten fault from the Early Jurassic and older volcanics and metasedimentary rocks of Quesnellia and the Early Cretaceous Okanogan Range Batholith (Hurlow, 1993) (Fig. 1B).


Zircon was separated from eleven metasedimentary samples from the Swakane Gneiss using standard rock-crushing and mineral-separation techniques at the University of Nevada, Reno. Nine of the eleven samples were collected in the Wenatchee block as previous studies have focused on characterizing the Swakane Gneiss in the Chelan block. For samples that were migmatitic, only material from the melanosome was crushed. The zircon separate was poured on adhesive tape to avoid picking bias and subsequently mounted in a 1-inch epoxy round. The mounts were polished to expose the approximate center of each zircon. Zircons were imaged using a cathodoluminescence (CL) detector housed on a JEOL JSM-7100FT field-emission scanning electron microscope (SEM) at the University of Nevada, Reno to identify growth patterns and internal textures within each zircon and thus provide textural context for laser-ablation analyses (Fig. 3). Zircons were first analyzed for U-Pb isotopes using a Nu Instruments Plasma HR multi-collector ICP-MS coupled to a 193 nm Photon Machines excimer laser at the University of California at Santa Barbara. Analytical specifications and data-reduction methods followed the procedure described in Sauer et al. (2017b). Zircon age populations were identified using (1) for Mesozoic dates, the weighted-mean 238U/206Pb age of >3 analyses that overlapped on concordia plots and yielded a mean square of weighted deviates (MSWD) of ∼1; or (2) for Proterozoic zircons, the weighted-mean 207Pb/206Pb age of >3 analyses with an MSWD of ∼1. The U-Pb results are summarized below, in Table 1, and in kernel density estimate and histogram plots (Figs. 4, 5, and 6). Data are also reported in Data Repository Table DR11.

Representative zircons from significant age populations were subsequently analyzed for Hf isotopes at the Radiogenic Isotope and Geochronology Laboratory at Washington State University using a New Wave UP 213 Nd-YAG (213 nm) laser and a ThermoFinnigan Neptune multi-collector ICP-MS. The Hf-isotope data collection and reduction followed the procedures described in Sauer et al. (2017b), which were modeled after Fisher et al. (2011, 2014). The CHUR Lu-Hf values of Bouvier et al. (2008) and the 176Lu decay constant determination of Söderlund et al. (2004) were used to calculate εHf values and initial 176Hf/177Hf ratios. The U-Pb and Hf-isotope results for major age populations from each sample are summarized in Table 1. In addition, the results are shown in Figures 5 and 6 and Data Repository Table DR2 (see footnote 1).


The Swakane Gneiss overall is a lithologically homogeneous body, and the majority of the metasedimentary samples studied here are representative and characterized by the major mineral assemblage of quartz + plagioclase + biotite ± garnet ± muscovite ± kyanite ± sillimanite. Accessory minerals include zircon + apatite ± titanite ± rutile ± graphite. Biotite and aluminosilicate (when present) define the foliation. In some samples, garnet, biotite, and muscovite are replaced by late hornblende, chlorite, and/or chloritoid. Given the homogeneity of the Swakane Gneiss, the U-Pb and Hf-isotope results of all samples are summarized together. A detailed description of results from each sample, in order of general geographic position from north to south in the Chelan and Wenatchee blocks, is available in the Data Repository (File DR1).

Cathodoluminescence images of zircons from the Swakane Gneiss metasedimentary rocks generally reveal distinct core, mantle, and rim zones. Zircon cores have a variety of CL patterns including oscillatory, flat (unzoned), sector, convolute, and ghost zoning (c.f. Corfu et al., 2003) (Fig. 3A). Mantle zones have dominantly oscillatory patterns (Fig. 3B), and the rims are usually unzoned and thin (∼1–5 µm), but can be as thick as ∼90 µm (Fig. 3C).

Overall, zircons from the Swakane Gneiss samples yield dates that are dominantly Mesozoic or Proterozoic (Figs. 4, 5, and 6). All samples are characterized by multiple Mesozoic peaks, including major Jurassic (ca. 160–140 Ma), Early Cretaceous (ca. 130–110 Ma), and Late Cretaceous (ca. 100–70 Ma) populations and minor Triassic peaks (ca. 250–200 Ma). Most samples, with the exception of garnet-biotite gneiss SK15-61A and garnet-kyanite metapelite SK14-28, also contain Proterozoic dates that are typically between ca. 1800–1600 and 1440–1360 Ma. Within samples containing Proterozoic zircon cores, the majority of the mantle zones have ca. 90–80 Ma dates (Figs. 3B, and 4). Zircons that yield dates younger than ca. 68 Ma (n = 7) are exclusively rim analyses and are usually unzoned or ghost zoned in CL. Dates between ca. 80 and 68 Ma include both rim (n = 45) and core (n = 145) analyses (Fig. 4). For Chelan block and some Wenatchee block samples, no correlations in zircon age and U/Th values are observed, and U/Th values are below ∼10 (i.e., KG-100B, SK15-31A, SK15-73, SK15-75) (Figs. 4, 5, 6, and 7). In the other Wenatchee block samples, U/Th values increase sharply up to ∼1000 for some core and rim analyses with dates younger than ca. 80 Ma (i.e., SK14-28, SK15-61A, SK14-43B, SK14-40C, SK14-19, SK16-01, SK14-32) (Figs. 4, 6, and 7).

The initial Hf-isotope compositions from the major zircon age populations reveal a number of noteworthy trends (Figs. 4, 5, and 6). In samples that lack Proterozoic populations (i.e., SK15-61A and SK14-28), nearly all εHfi values for Mesozoic dates plot between depleted-mantle and CHUR values. In all other samples, Late Cretaceous zircons have a wide range of εHfi values between depleted mantle and ancient crust (+13 to –22). Zircons from the Early Cretaceous and Jurassic populations have intermediate Hf-isotope compositions between depleted mantle and CHUR (εHfi = +13 to +1). In comparison, Triassic zircons dominantly have near-depleted mantle εHfi values. Hafnium-isotope results from the two Proterozoic populations at ca. 1.38 and 1.8–1.6 Ga have a similar range of εHfi values and mostly plot in a spread between depleted-mantle and CHUR values.


Three previous studies have investigated the tectonic history of the Swakane Gneiss (Matzel et al., 2004; Paterson et al., 2004; Gatewood and Stowell, 2012) and have interpreted different histories for the origin of the Swakane Gneiss protolith and the timing at which it was buried and metamorphosed. The following considers these previous interpretations and the relationship of the Swakane Gneiss to other components of the North Cascades crystalline core using the new detrital zircon U-Pb and Hf-isotope data presented here.

Timing of Metamorphism and In Situ Zircon Growth

Matzel et al. (2004) concluded that the youngest ID-TIMS dates (ca. 72 Ma) of oscillatory-zoned zircons represented the likely maximum depositional age (MDA) of the Swakane Gneiss protolith. In contrast, Gatewood and Stowell (2012) presented LA-ICPMS zircon core and rim dates that overlapped with their ca. 74–66 Ma Sm-Nd garnet isochron dates; therefore, they interpreted that analyses younger than ca. 75 Ma represent metamorphic ages. In the current study, zircons of in situ (those that grew during metamorphism and/or melt crystallization) versus detrital origin are differentiated based on core-rim age relationships, U/Th ratios, and CL-growth zoning. The observed patterns also provide insight into the specific mechanism by which the zircon formed (e.g., Chen et al. 2010).

In Situ versus Detrital Zircons

Zircons that form in situ during metamorphism and/or melt crystallization consist of either inherited grains that have been (partially) recrystallized, growth of entirely new grains, or new rim growth due to fluid-present (hydrothermal), melt-present (magmatic), and/or solid-state (metamorphic) reactions (Hoskin and Black, 2000; Möller et al., 2003; Harley and Kelly, 2007). Zircons from the Swakane Gneiss preserve complicated CL-zoning textures that likely relate to multiple zircon-crystallization events; grains show distinct core and rim relationships as well as a mantle zone in some grains (Fig. 3).

In addition to variable CL textures, zircons record a range of U/Th ratios. Some zircons younger than 80 Ma show a sharp increase in U/Th relative to zircons with older dates (Figs. 4 and 7). Uranium/Th ratios >10 (or Th/U < 0.1) are typically thought to be characteristic of metamorphic zircons (e.g., Rubatto, 2002) because competing minerals such as monazite and allanite can take up thorium during metamorphic zircon growth (Rubatto and Gebauer, 2000; Hermann, 2002). In addition, metamorphic fluids and partial melts commonly have a higher concentration of U versus Th; therefore, zircon crystallized from these fluids/melts will have a higher U/Th ratio (Rubatto, 2002). In contrast, other processes that form in situ zircon do not necessarily result in high U/Th values (Vavra et al., 1999; Möller et al., 2003). For example, U/Th values of in situ zircons that experienced solid-state recrystallization in the absence of a fluid record the U/Th ratio of the protolith zircon (Hoskin and Black, 2000; Möller et al., 2003; Chen et al., 2010). Thus, in high-grade rocks that experienced protracted metamorphism and varying degrees of solid-state recrystallization, it is common to observe a large spread of ages that relate to an open zircon system (Hoskin and Black, 2000). In this case, the youngest U-Pb age with the highest U/Th ratio can be interpreted to represent the age of complete recrystallization, whereas the older ages likely represent incomplete resetting of the U-Pb system (Hoskin and Black, 2000). In this study, the CL-zoning patterns and U/Th ratios were thus considered together to evaluate the origin of zircons in the Swakane Gneiss.

In all Swakane Gneiss samples, rims were observed on zircon cores (Fig. 3). Rims that were large enough to be analyzed have dates between ca. 79 and 66 Ma and were found on cores of all ages (Fig. 3). The unzoned ca. 78–66 Ma rims from two samples (SK14-28 and SK15-61A) have U concentrations similar to detrital zircons, but lower Th concentrations, resulting in U/Th values >10 (Figs. 5, 6, and 7A; Table 2). In all other samples, only ca. 70–68 Ma rims have U/Th values >10 (Fig. 7), whereas ca. 79–70 Ma rims mostly have variable U/Th values between ∼1 and 1000. Overall, based on the CL and U/Th characteristics, the ca. 79–66 Ma rims are interpreted to have formed in situ within the Swakane Gneiss. Given the scatter in U/Th values and variable CL patterns, the zircon rims with U/Th values either greater than or less than ∼10 likely formed due to different processes.

Mantle zones are observed in the majority of Swakane Gneiss samples and are dominantly ∼10–50 µm thick. The mantles are weakly oscillatory zoned, reveal dates between ca. 86 and 80 Ma, and are only found on Proterozoic cores. The U/Th ratios are variable (∼3–100) but typically less than observed for the ca. 79–70 Ma rims described above (Fig. 4). As the ca. 86–80 Ma mantles are not observed in all samples and are limited to Proterozoic cores, they are likely related to Late Cretaceous intrusions that assimilated Proterozoic crust and were subsequently eroded. Thus, the ca. 86–80 Ma mantles are interpreted as detrital, in contrast with the ca. 79–66 Ma in situ rims. The interpretation that the ca. 86–80 Ma mantle zones were detrital in origin disagrees with the conclusion of Gatewood and Stowell (2012) that the ca. 81 Ma zircon population was the result of mixed analyses.

Zircon cores with ages younger than ca. 80 Ma are within error of in situ rim ages and have highly variable CL textures and U/Th ratios. The majority of these cores have U and Th concentrations that overlap with detrital zircons in the sample, corresponding to low U/Th ratios (Fig. 7). They show a spectrum of textures, from oscillatory zoned to unzoned, and include features such as patchy-sector zoned overprints and clouded, ghost-CL textures (Fig. 3D). In comparison, the younger than ca. 80 Ma cores with high U/Th ratios have very weakly zoned to unzoned CL textures. The CL textures and U/Th ratios of younger than ca. 80 Ma core analyses overlap with other dates interpreted to represent in situ crystallization and indicate these zircons likely also record the timing of in situ crystallization or partial resetting of the U-Pb system.

Timing and Mechanisms of In Situ Zircon Crystallization

Based on the above observations, zircon cores and rims younger than ca. 80 Ma in Swakane Gneiss samples likely formed in situ. These zircons can be further divided into two groups based primarily on CL texture and U/Th ratios. Those in the first group have variable CL textures ca. 80–70 Ma, low U/Th cores with similar U and Th contents to detrital zircons (i.e., SK15-73, SK15-31A; Fig. 7). These zircons likely experienced solid-state recrystallization due to diffusion in the absence of a fluid (e.g., Hoskin and Black, 2000; Chen et al., 2010). In comparison, the second group is defined by dominantly unzoned, high U/Th ca. 80–68 cores and ca. 79–66 Ma rims that have very low Th but similar U contents in comparison to the detrital zircons (i.e., SK14-28, SK15-61A; Fig. 7). These in situ zircons and rim overgrowths likely grew in equilibrium with high-Th metamorphic minerals (i.e., garnet), in the presence of a fluid during metamorphism and/or during melt crystallization (e.g., Möller et al., 2003; Chen et al., 2010).

The in situ ages form a spectrum among the youngest interpreted detrital zircons from previous studies (ca. 72 Ma, Matzel et al., 2004; and ca. 91 Ma, Gatewood and Stowell, 2012), the garnet Sm-Nd isochron dates (74–66 Ma; Gatewood and Stowell, 2012), and the crystallization age of a leucogranite sheet (ca. 68 Ma; Matzel et al., 2004). A few samples studied here were collected from localities directly from or nearby to the previously analyzed samples (i.e., SK15-73, SK14-28, and SK15-75). In the Chelan block, rim ages (ca. 76–71 Ma) from sample SK15-73 overlap with a ca. 74 Ma Sm-Nd garnet date from a nearby sample (Gatewood and Stowell, 2012). In addition, Wenatchee block garnets from the same outcrop as garnet-kyanite schist SK14-28 have a ca. 66 Ma garnet date that is within error of the youngest zircon-rim date (ca. 67 Ma).

The Swakane Gneiss was previously interpreted to have resided in the mid to lower crust for likely ∼15 m.y., as bracketed by ca. 74–66 Ma garnet Sm-Nd dates and ca. 51–46 Ma Ar-Ar cooling ages (Matzel, 2004; Paterson et al., 2004; Gatewood and Stowell, 2012). Thus, open-system behavior (i.e., local solid-state recrystallization and/or new zircon/rim growth) likely occurred while the Swakane Gneiss occupied deep crustal levels and experienced temperatures between 640 and 750 °C (Valley et al., 2003; Gatewood and Stowell, 2017), resulting in the spread of ages between the likely youngest detrital-zircon populations (ca. 93–81 Ma) and the likely ca. 74–66 Ma age of metamorphism and/or melt crystallization.

Age and Provenance of the Swakane Gneiss Protolith

After removing in situ ages from the Swakane Gneiss samples, the weighted-mean average of the youngest three overlapping concordant analyses was used to approximate the maximum age of sediment deposition. The MDA determined from the youngest detrital-zircon dates in sedimentary rocks deposited in active tectonic settings (i.e., active continental magmatic arcs) are thought to closely match the actual depositional age of the sediment (Dickinson and Gehrels, 2009). Using this criteria, the Swakane Gneiss samples yielded MDAs between ca. 93 and 81 Ma (Table 2; Fig. 8).

Detrital-Zircon Characteristics and Potential Sources

The Swakane Gneiss contains mostly Mesozoic and Proterozoic detrital zircons (Matzel et al., 2004; Gatewood and Stowell, 2012; this study). Based on the MDAs, potential protoliths for the Swakane Gneiss include ca. 93–81 Ma sedimentary units in forearc, backarc, and accretionary-wedge positions to the North Cascades crystalline core and along nearly the entire western North American continental margin. The rocks of northwestern Washington and southern British Columbia are interpreted to have undergone between 700 and 1600 km of northward, margin-parallel translation (e.g., Wyld et al., 2006; Rusmore et al., 2013) and, therefore, they may be linked to rocks currently found between southern Oregon and northern Mexico. Thus, comparison of the detrital-zircon characteristics of the Swakane Gneiss with recent studies of sedimentary units within the North American Cordillera provide further context for understanding the origin of the protolith of the Swakane Gneiss (Jacobson et al., 2011; Sharman et al., 2015; Matthews et al., 2017; Sauer et al., 2017a).

Overall, the results from the Swakane Gneiss suggest a major provenance shift in the mid-Late Cretaceous. The two samples with the oldest MDAs, SK14-28 (ca. 93 Ma) and SK15-61A (ca. 88 Ma), lack Proterozoic zircons, whereas samples with MDAs younger than ca. 86 Ma have a combination of Mesozoic and Proterozoic peaks. The Mesozoic zircons in all samples dominantly have suprachondritic initial Hf-isotope compositions, but in samples with MDAs younger than ca. 86 Ma, ca. 100–81 Ma zircon have a wide range of εHfi values (+13 to –20). The unradiogenic εHfi values suggest these younger sediments were partly eroded from Late Cretaceous arc-related plutons that assimilated Proterozoic crust. This is further supported by the presence of ca. 86–81 Ma mantles on Proterozoic cores (Figs. 3B and 4). The relative percentage of Proterozoic detrital zircons in each sample also increases with younging MDA. No clear pattern between MDA or detrital-zircon signature and structural position is observed. These combined datasets likely indicate the addition of a distinct Proterozoic and Late Cretaceous sediment component in strata deposited after ca. 86 Ma.

Upper Cretaceous forearc and accretionary-wedge units deposited outboard of Cordilleran arc systems and along the continental margin from northern Mexico to southern British Columbia are also characterized by the same provenance shift. Accretionary-wedge and forearc strata deposited before ca. 90–72 Ma outboard of Cordilleran arc systems are also dominantly characterized by Mesozoic age peaks with a lack of Proterozoic zircons (Fig. 9). The units deposited during this time include: (1) the Californian forearc and accretionary Franciscan Complex units, characterized by a combination of Jurassic (ca. 160–140 Ma), Early Cretaceous (ca. 135–100 Ma), and Late Cretaceous (ca. 100–80 Ma) age peaks (Jacobson et al., 2011; Sharman et al., 2015); (2) strata deposited before ca. 80–72 Ma from the Nanaimo Group are characterized by ca. 90 Ma and ca. 150 Ma peaks with the variable presence of a ca. 118 Ma population (Mustard et al., 2006; Matthews et al., 2017); and (3) two samples from the arkosic petrofacies of the western mélange belt with MDAs of ca. 87 and 96 Ma characterized by a similar pattern with dominant ca. 98–95 Ma and ca. 165–145 Ma peaks and a smaller ca. 120–100 Ma population (Dragovich et al., 2009; Brown, 2012).

In comparison, potential protolith units located to the east of the North Cascades arc have a very different detrital-zircon signature. The Cretaceous sedimentary rocks of the Methow terrane are interpreted to have been deposited in a forearc basin but were juxtaposed with the inboard, eastern margin of the North Cascades arc by ca. 91 Ma (Miller, 1994). Therefore, the Methow terrane represents an analogue for sediments that may have been buried beneath the North Cascades arc via retroarc underthrusting in the Late Cretaceous. However, the Methow terrane lacks the distinct Proterozoic peaks observed in the younger than ca. 86 Ma Swakane Gneiss samples. In addition, the detrital zircons from the Methow do not align with data from the two older Swakane Gneiss samples (Fig. 9). Jurassic strata yield unimodal age peaks and near-depleted mantle isotope compositions, whereas Cretaceous units yield MDAs older than ca. 90 Ma, Mesozoic bimodal U-Pb age peaks, and initial Hf-isotope compositions between depleted-mantle and CHUR (DeGraaff-Surpless et al., 2003; Sauer et al., 2017a). In addition, rocks of Quesnellia and the Okanogan Range Batholith, located to the east of the Methow terrane, do not include Late Cretaceous sediments and, therefore, are also not candidates for the Swakane Gneiss protolith (Gordon et al., 2017).

Overall, the detrital-zircon patterns suggest the Swakane Gneiss protolith was derived from either accretionary wedge or forearc units. Mesozoic zircons in the Swakane Gneiss, as well as in these other accretionary wedge and forearc deposits, have ages that form Late Cretaceous, Early Cretaceous, Jurassic, and minor Triassic peaks. The Triassic–Cretaceous zircons have multiple potential sources throughout the western United States and southern British Columbia (summarized in Sauer et al., 2017a). The radiogenic Mesozoic zircons in all samples of the Swakane Gneiss have similar peaks and Hf-isotope compositions to plutonic rocks of the southern and central Coast Plutonic Complex (c.f. Cecil et al., 2011; Homan et al., 2017). However, the change in provenance in the mid-Late Cretaceous (ca. 86 Ma) from dominantly a Mesozoic detrital-zircon signature to one that includes significant Proterozoic and unradiogenic Late Cretaceous populations demonstrates the addition of new sediment sources that are not observed in rocks of the North Cascades or Coast Plutonic Complex (Fig. 9). These results may indicate that a topographic barrier, previously interpreted to be a high-standing, mid-Cretaceous arc, was breached in the mid-Late Cretaceous allowing sediment from inboard terranes that would yield Proterozoic zircons to reach west of the arc (e.g., Sharman et al., 2015).

Proterozoic cores represent 29% of the analyzed zircons (413 out of 1433) in the Swakane Gneiss samples. Their ages form a sharp peak at ca. 1380 Ma and a broad population between ca. 1800 and 1600 Ma with a peak at ca. 1710 Ma. The combination of these two populations is recognized in Upper Cretaceous–Paleocene forearc and accretionary-wedge units from southern California to Alaska (e.g., Dumitru et al., 2016), and more specifically, in forearc and accretionary-wedge units to the Sierra Nevada batholith from southwestern Oregon to southern California (e.g., Hornbrook Basin, Great Valley Group, southern California forearc, and Franciscan Complex) in strata deposited between ca. 90 and 72 Ma (DeGraaff-Surpless et al., 2002; Surpless and Beverly, 2013; Sharman et al., 2015; Chapman et al., 2016; Dumitru et al., 2016) (Fig. 9). To the north in the Coast Plutonic Complex, the distinct Proterozoic zircons do not appear until: after ca. 72 Ma in the northern Nanaimo Group (Matthews et al., 2017); the middle Campanian (ca. 80–75 Ma) in the Protection Formation of the southern Nanaimo Group (Mustard et al., 2006; Mahoney et al., 2014); and after ca. 72 Ma in the youngest identified clastic portion of the accretionary western mélange belt (Sauer et al., 2017a).

These Proterozoic peaks have been attributed to either southwestern Laurentia (Matzel et al., 2004; Jacobson et al., 2011; Sharman et al., 2015; Garver and Davidson, 2015; Matthews et al., 2017) or northwestern Laurentia (Mahoney et al., 1999; Dumitru et al., 2016). Both regions also have potential sources of unradiogenic ca. 100–81 Ma zircons: sediment eroded from the Idaho batholith that intrudes the edge of northwestern Laurentia (Gaschnig et al., 2011) and Late Cretaceous plutons that intrude the Mojave terrane in southwestern Laurentia (Barth et al., 2016; Fisher et al., 2017), respectively. The initial Hf-isotope signature and age peaks of the Proterozoic zircons from sediment sources in southwestern and northwestern Laurentia cover a similar range and are not diagnostic (Sauer et al., 2017a). However, sediment derived from northwestern Laurentian sources would likely contain ca. 2700–2400 Ma dates, as these Archean dates are present in the Belt Supergroup (Stewart et al., 2010). As the Swakane Gneiss samples do not yield significant Archean zircons, we interpret that the Proterozoic zircons in the Swakane Gneiss most closely match Mesozoic and Proterozoic sediment sources in southwestern Laurentia, but a northwestern Laurentian provenance cannot be ruled out.

The protolith of the Swakane Gneiss has been previously compared to the Upper Cretaceous Nanaimo Group based on similar depositional age and detrital-zircon signature (Matzel et al., 2004). Overall, the Nanaimo Group contains similar Triassic–Cretaceous and Proterozoic age populations to the Swakane Gneiss; however, the majority of published age data shows that the Nanaimo strata that contain abundant Proterozoic zircons have MDAs between ca. 80 and 72 Ma (e.g., Mahoney et al., 1999; Matthews et al., 2017), which postdate the ca. 93–81 Ma Swakane Gneiss protolith. The ca. 96–87 Ma portions of the western mélange belt have similar peaks to the ca. 93–88 Ma Swakane Gneiss samples but no ca. 87–72 Ma clastic portions of the western mélange belt have been identified to evaluate their similarity to the rest of the Swakane Gneiss. In comparison, two Proterozoic peaks are observed in Franciscan Complex and Californian forearc units as old as ca. 90 Ma (Chapman et al., 2016). In addition, the Pelona-Orocopia-Rand schists, Schist of Sierra de Salinas, and San Emigdio Schist (collectively referred to as the PORS), have the same detrital-zircon age patterns as Californian forearc and accretionary wedge units (Jacobson et al., 2011; Chapman, 2016) and the Swakane Gneiss. Thus, the age signature observed in the Swakane Gneiss is not diagnostic of a particular unit in the North American Cordillera but instead strongly connects the Swakane Gneiss protolith to Upper Cretaceous forearc and accretionary wedge units that are either currently located to the south along the continental margin (e.g., the Franciscan Complex, California forearc, PORS) or were likely deposited and subsequently translated north from between the latitude of southern Oregon and southern California (e.g., Nanaimo Group and western mélange belt).

Relationship to Other Components of the Crystalline Core

The results from the Swakane Gneiss provide insight into its relationship with other metasupracrustal bodies exposed through the crystalline core. The Skagit Gneiss is the other main large body at deeply exposed crustal levels within the North Cascades arc (Whitney, 1992). The Skagit metasedimentary rocks have MDAs mostly between ca. 134 and 96 Ma, with the exception of one sample with a ca. 79 Ma MDA (Sauer et al., 2017b). Thus, the protoliths of most Skagit metasediments are older and do not overlap with the age range of the Swakane Gneiss. However, the one ca. 79 Ma Skagit sample has nearly an identical detrital-zircon age distribution and Hf-isotope signature to the Swakane Gneiss samples deposited at ca. 81 Ma (Fig. 9). In addition, the Skagit protoliths reached depth between ca. 74 and 65 Ma (Gordon et al., 2010b; Sauer et al., 2017b), which is coeval with the timing of metamorphism in the Swakane Gneiss (Matzel et al., 2004; Gatewood and Stowell, 2012; this study). This indicates the youngest rocks of the Swakane Gneiss and the Skagit Gneiss likely share a similar protolith and burial histories.

In addition to mostly older protolith ages for the Skagit Gneiss metasedimentary rocks, there are some important differences in lithology and structural setting between the Skagit and Swakane Gneisses. The Skagit Gneiss metasediments are extensively intruded by arc plutons and exist as rafts or lenses within voluminous orthogneiss, whereas no pre-Miocene plutons intrude the Swakane Gneiss. The Skagit Gneiss metasediments are also migmatitic and have undergone extensive metamorphic differentiation and partial melting resulting in more micaceous and plagioclase-rich lithologies than the Swakane Gneiss. In addition, the Swakane Gneiss is in fault contact with all other older units (e.g., the Napeequa Schist), whereas the protolith plutons of the Skagit Gneiss orthogneisses intrude the surrounding metasupracrustal units such as the Napeequa Schist.

The Dinkelman décollement separates the Swakane Gneiss and Napeequa Schist in the Chelan block and records top-to-the-north shear (Fig. 2). The latest motion was extensional, and is only loosely bracketed to between ca. 73 and 45 Ma (Paterson et al., 2004). The Swakane Gneiss and Napeequa Schist in the Wenatchee block are separated by a likely analogous structure with similar, but less well-defined, kinematics (Miller et al., 2006). Some evidence is preserved in the Napeequa Schist that these structures were originally SW-vergent thrusts that were later reactivated as extensional structures; however, no evidence of this is observed in the Swakane Gneiss (Paterson et al., 2004). The Napeequa Schist and Swakane Gneiss on either side of these bounding structures record similar P-T conditions (Valley et al., 2003), but likely reached these peak conditions at different times (ca. 90 Ma versus 72–65 Ma, respectively) (Paterson et al., 2004). The Napeequa Schist in the Wenatchee block was at depth when it was intruded by the ∼7–10 kbar, ca. 92 Ma Tenpeak pluton (Dawes, 1993; Matzel et al., 2006). The Tenpeak intrusion was coeval with the MDA of the oldest identified sediments that formed part of the Swakane Gneiss protolith. The timing of the juxtaposition of the Swakane Gneiss and Napeequa Schist is uncertain, but it is younger than the protolith age of the Swakane Gneiss (ca. 81 Ma) and older than ca. 48 Ma motion on the Entiat fault that offsets the contact (Tabor et al., 1989).

The Nason terrane, consisting of the Nason Ridge Migmatitic Gneiss and Chiwaukum Schist, has a Lower Cretaceous protolith age (ca. 145–120 Ma; Brown and Gehrels, 2007; Paterson, 2014). The likely accretionary complex protolith was intruded by the ca. 96–91 Ma Mount Stuart batholith (Matzel et al., 2006). Cooling in the Nason terrane occurred soon after metamorphism, as indicated by the closely tied ca. 88–86 Ma garnet Sm-Nd dates (Stowell et al., 2007) and ca. 85–75 Ma biotite-cooling ages (Matzel, 2004). The cooling ages also indicate that the exhumation of the Nason terrane coincided with the deposition and burial of the Swakane Gneiss. Overall, the results from the Swakane Gneiss together with observations from the Nason terrane, Napeequa Schist, and Skagit Gneiss highlight that supracrustal rocks were incorporated into the crystalline core to a variety of crustal levels during two main Late Cretaceous episodes (i.e., 96–90 Ma and 74–66 Ma).

Incorporation of the Swakane Gneiss into the North Cascades Arc

The forearc and/or accretionary-wedge sediment source indicates that the Swakane Gneiss protolith likely originated outboard of the North Cascades arc, and, along with the arc, was between ∼700 and 1600 km south of its current latitude. Next, the Swakane Gneiss protolith was either buried along the subduction trench (underplating) or by thrust faults in the overriding continental plate (underthrusting) (Fig. 10). Underplating may involve subduction erosion of the overriding plate, which could imbricate forearc material into the accretionary wedge. Conversely, the accretionary wedge may be accreted onto the continent and, thus, could be buried by underthrusting in the upper plate. Either scenario may bury accretionary wedge and/or forearc material, and so other observations about the metamorphic history in addition to the protolith origin are required to understand the incorporation of the Swakane Gneiss in the North Cascades arc.

Burial of the Swakane Gneiss protolith may have been progressive and, thus, could have begun immediately after deposition of the oldest components (i.e., after ca. 93 Ma). However, it more likely began after the youngest recognized rocks were deposited (ca. 81 Ma) based on the similar ranges of in situ zircon ages in samples across the Swakane Gneiss and narrow range of MDAs (Fig. 8). The Swakane Gneiss protolith reached pressures of 8–12 kbars (Valley et al., 2003; Gatewood and Stowell, 2017) by ca. 74–66 Ma based on in situ zircon ages, crystallization ages for cross-cutting leucocratic sheets, and Sm-Nd garnet dates (Matzel et al., 2004; Gatewood and Stowell, 2012; this study). These time constraints correspond to burial to 40 km depth at a rate of between 3 and 6 km/m.y. after deposition. Cooling through hornblende- and biotite-closure temperatures occurred between ca. 51 and 47 Ma (Matzel, 2004; Paterson et al., 2004).

Underplated sediments elsewhere in the North American Cordillera, such as the PORS of southern California, have been interpreted to be the result of shallow-slab subduction and the removal of the sub-arc mantle (Grove et al., 2003; Saleeby, 2003). Underplating occurs at plate-tectonic rates; for example, the PORS schists reached 30–33 km depth in less than 3 m.y. (Grove et al., 2003), corresponding to a rate of 10–11 km/ m.y. Underplating has also been proposed for the deepest exposed paragneiss in the Central Gneiss Complex in British Columbia (Pearson et al., 2017) and the Condrey Mountain Schist in northern California (Saleeby and Harper, 1993).

Several lines of evidence suggest that underplating was an unlikely scenario for the Swakane Gneiss. Matzel et al. (2004) point out that the peak-temperatures and clockwise P-T paths preserved in the Swakane Gneiss (Valley et al., 2003) contrast with the metamorphic conditions and counter-clockwise paths suggested for southern Californian underplated schists (Jacobson, 1995; Kidder and Ducea, 2006). In addition, the North Cascades crust was postulated to be ∼55 km thick by ca. 90 Ma, and underplating the Swakane Gneiss protolith to ∼40 km depth would thus had to have involved significant removal and/or transfer of lower crustal material. There is also evidence for magmatism in the North Cascades between ca. 96 and 45 Ma (Miller et al., 2009), whereas underplating would likely have slowed or stopped magmatism.

Alternatively, it is possible that the Swakane Gneiss protolith was underplated. Even though there is magmatism throughout the North Cascades, as described above, there are no plutons that intrude the Swakane. In addition, some paleogeographic reconstructions indicate a southern Californian paleolatitude for the North Cascades (e.g., Kim and Komada, 2004; Umhoefer and Blakey, 2006). Therefore, the Swakane Gneiss protolith may be analogous to units to the south along the continental margin (e.g., the PORS), and could have been incorporated into the North Cascades arc at latitudes where flat-slab subduction was has been proposed (Saleeby, 2003).

Underthrusting of upper-crustal material from the forearc side of the arc is the likely alternative sediment-incorporation mechanism. It does not require displacement of the North Cascades lower crust and would likely preserve higher geothermal gradients, similar to those preserved in the Swakane Gneiss (Fig. 10). The structure along which the Swakane Gneiss protolith was buried has not been identified, but may be represented by the Dinkelman décollement. Underthrusting of sediments from the retroarc side of the subduction system has been interpreted in multiple regions of the North American Cordillera (e.g., Chin et al., 2013; Pearson et al., 2017). This study highlights a potential example of forearc underthrusting within a subduction system, which may be an important mechanism that introduces supracrustal material to depth within continental arcs.


The Swakane Gneiss protolith was likely deposited to the west of the North Cascades and other segments of Cordilleran arc systems, based on analogous detrital-zircon U-Pb and Hf-isotopes characteristics observed in forearc and accretionary-wedge units distributed along the continental margin from British Columbia to southern California. The U-Pb and Hf-isotope analyses and CL images of discrete rim, core, and mantle zones of Swakane Gneiss zircons are used to untangle ages that represent in situ growth from youngest detrital peaks. In situ zircon rims and cores record ca. 74–66 Ma ages, whereas the youngest detrital-zircon ages indicate protolith-depositional ages that range from ca. 93–81 Ma. Based on connections with accretionary wedge and forearc rocks, the Swakane Gneiss protolith was likely either underthrust or underplated to >40 km depth in the arc after ca. 81 Ma and before ca. 74–66 Ma. The detrital-zircon characteristics and timing of sediment incorporation indicate a link between the youngest metasedimentary units of the Skagit Gneiss Complex and the Swakane Gneiss.


We would like to thank Joel DesOrmeau for assistance in sample preparation and obtaining zircon-CL images at the University of Nevada, Reno SEM facility. Thank you to Andrew Kylander-Clark for help with U-Pb analyses at the University of California at Santa Barbara ICPMS laboratory and to Charles Knaack and Da Wang for help with Hf-isotope analyses at the Radiogenic Isotope Laboratory at Washington State University. This work was supported by National Science Foundation grants EAR-1419810 to S.M. Gordon and EAR-1419787 to R.B. Miller, U.S. Geological Survey EDMAP grant G13AC00124 to S.M. Gordon, and Geological Society of America student research awards to K.B. Sauer. Mihai Ducea, David Pearson, and editor Laurent Godin provided helpful reviews that improved this manuscript.

1GSA Data Repository Item 2018144, U-Pb and Hf-isotope data tables (Tables DR1 and DR2) and detailed results (File DR1), is available at http://www.geosociety.org/datarepository/2018, or on request from editing@geosociety.org.
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.