Development and Evolution of the Rattlesnake Creek Terrane, Klamath Mountains, Northern California

Abstract

The basement of the Rattlesnake Creek terrane (RCT) in the Klamath Mountains is a mélange of metamorphosed sedimentary and igneous blocks. Recent work shows that the overlying RCT cover sequence has a North American provenance but formed after accretion to the continental margin, so it is unclear if the basement mélange formed exotic or endemic to North America. This study presents petrography and zircon geochronology from RCT metasedimentary blocks and crosscutting intrusions. The southernmost RCT preserves both Early Jurassic and Middle-Late Jurassic cover sequence deposits and records continental clasts and 33% pre-Mesozoic zircons at ~201 Ma, effectively none at ~191 Ma, and 79–90% from 168 to 163 Ma. During active magmatism 207–193 Ma, the RCT was receiving continental sediment, inconsistent with a distant intraoceanic arc. We interpret that the RCT subduction zone formed proximal to North America in the Late Triassic and that there was a sediment pathway to the RCT at ~201 Ma. During Middle to Late Jurassic rifting and subsequent Nevadan compression, the cover sequences were dismembered and incorporated into the mélange by tectonic and sedimentary processes. The age and provenance of metasedimentary deposits in the RCT is inconsistent with west-dipping subduction models in the Klamath Mountains region.

1. Introduction

The Klamath Mountains province (Figure 1) straddles the California–Oregon border and is composed of a series of thrust-bounded tectonostratigraphic terranes. It is also the largest exposure of pre-Cenozoic rocks on the western continental margin of North America [1], making it a key location to study Mesozoic terrane tectonics. W. P. Irwin introduced the term “terrane” in the Klamath Mountains, as a fault-bounded tract of rocks with similar geologic features, such as stratigraphic formations, intrusive rocks, mineral deposits, and tectonic history, which differ from adjacent tracts [1,2]. Terranes are now recognized around the world and are critical to reconstructing plate tectonic histories. Yet the ages, correlations, and tectonic settings of many of these original tectonostratigraphic terranes in the Klamath Mountains are still debated.

One of the most enigmatic terranes is the Rattlesnake Creek terrane (RCT), which has been interpreted as forming marginal to North America in the Late Triassic to Early Jurassic before Middle Jurassic accretion to the continental margin [4,5,6,7] or as an east-facing exotic archipelago which formed far from North America and accreted by the Middle Jurassic [8,9] or in the Late Jurassic [10,11]. The RCT basement is a tectonic mélange with blocks of mostly ocean plate stratigraphy [4,5,12], but the cover sequence contains arc-related lavas and terrigenous detritus linked to North America [4,7,13]. Previous work constrained the deposition of the RCT cover sequence to Late Triassic to Early Jurassic based on a ~198 Ma crosscutting intrusion and fossils from coherent and dismembered deposits interpreted as the cover sequence [4]. However, recent detrital zircon analyses of clastic rocks in the cover sequence have maximum depositional ages (MDAs) 170–161 Ma and abundant Paleozoic-Archean ages [13]. Since at least part of the cover sequence is Middle-Late Jurassic, it is now less clear if the 207–193 Ma RCT intrusive suite [4] formed as an intraoceanic or marginal arc in the Late Triassic to Early Jurassic. Here, we present new petrographic data and 1245 new detrital zircon U-Pb ages from metasedimentary rocks in the mélange and crosscutting intrusions. Our data provide additional evidence of an Early Jurassic cover sequence with continental input, indicating proximity of the RCT to North America at ~201 Ma, and document Middle-Late Jurassic cover sequence farther west than previously mapped in the southern RCT.

2. Geologic Background

Terranes in the Klamath Mountains province (Figure 1) are regionally east-dipping but locally folded and cut by high angle faults and intruded by ultramafic to silicic plutons that range from early Paleozoic to early Cretaceous [1]. Cretaceous rocks of the Hornbrook Formation and Great Valley Group as well as Tertiary volcanic rocks of the Western Cascade Group unconformably overlie the Klamath Mountains [1]. Irwin [2] defined four major lithotectonic belts: the Eastern Klamath terrane, the Central Metamorphic belt, the Western Paleozoic and Triassic belt (WPzTrB), and the Western Klamath terrane (Figure 1). These belts are generally younger from east to west, right side up, and separated by east-dipping, west-verging thrust faults [14]. The WPzTrB is the largest of the four major belts and from east to west includes the Stuart Fork/Fort Jones terrane, the North Fork terrane (NFT), the Eastern Hayfork terrane (EHT), the Western Hayfork terrane (WHT), and the Rattlesnake Creek terrane (RCT). To the east are younger rocks of the Western Jurassic belt: the Josephine ophiolite, the Rogue–Chetco complex, and the Galice Formation.

2.1. The North Fork, Eastern Hayfork, and Western Hayfork Terranes

The NFT is an assemblage of internally folded ophiolitic rocks overlain by oceanic volcanic sedimentary rocks (Figure 1) [15]. Age control data include Permian to Middle Jurassic fossils in limestone and chert [16], a conservative 310–265 Ma zircon U-Pb age from a plagiogranite [15], a ~200 Ma ⁴⁰Ar/³⁹Ar hornblende age on a gabbro cutting diabase [17], and 225–180 Ma U-Pb detrital zircon ages from argillites in the central and southern NFT [16,18]. Some argillite detrital zircon ages of the NFT are similar to the early Jurassic Eastern Klamath arc and older Klamath terranes, indicating endemic formation [14,19,20].

The EHT is a mélange and broken formation of weakly crystallized and intensely tectonized metachert with fine-grained argillite as well as tectonic blocks of greenstone, blueschist, and small amounts of marble, quartzite, and siliceous schist [14,21]. Limestone and chert fossils range from Early Pennsylvanian to Late Triassic [16]. Matrix argillites with whole-rock ⁴⁰Ar/³⁹Ar at 263 Ma [22], continental zircon ages [18,23], and detrital zircon MDAs at ~272 Ma, ~211 Ma, and ~201 Ma suggest growth by progressive subduction accretion [18]. Detrital zircon U-Pb ages in exotic metasandstone blocks are dominantly Paleoproterozoic and Archean, consistent with the North American Cordillera [21], and mica and amphibole ⁴⁰Ar/³⁹Ar ages in micaceous quartzite blocks are 360–346 Ma [22].

The WHT is a Middle Jurassic arc cover sequence, and K-Ar and ⁴⁰Ar/³⁹Ar ages of magmatic amphiboles are 177–171 Ma with on K-Ar age at 168 Ma [24,25]. At ~170 Ma, the WHT was intruded by the Ironside Mountain Batholith [24], so the youngest WHT amphibole ages may represent cooling ages rather than true depositional ages. The presence of rare quartz- and K-feldspar-rich arenite is consistent with a minor component of sediment input from the North American craton above an east-dipping subduction zone [26], and the WHT has been interpreted as originally depositional on the EHT and RCT [12,24,25].

2.2. Rattlesnake Creek Terrane

The RCT can be divided into a basement mélange ± serpentinite-matrix and peridotite massifs and a cover sequence of volcanic and sedimentary rocks [4]. Many gabbroic to quartz dioritic plutonic bodies, referred to here as the RCT intrusive suite, intrude both the basement assemblage and rocks interpreted as the cover sequence of the RCT, and the intrusions and cover sequence are locally incorporated into the basement mélange [4].

2.2.1. Basement Mélange

There are significant differences in the structure and metamorphism of the RCT basement mélange in the southern and central Klamath Mountains. The southern RCT is a block-in-matrix mélange with a serpentinite matrix and has been metamorphosed to subgreenschist grade with well-preserved protolith texture with limited recrystallization [4]. Abundant fragments of harzburgite and minor dunite increase in serpentinization into the matrix and are likely its source [4]. Blocks include felsic to ultramafic plutonic bodies, greenstone, amphibolite, pillow basalt, chert, limestone, metavolcanic, and volcaniclastic rocks [4,5].

In the central Klamath Mountains, the RCT (an area of which has been referred to as the Marble Mountain terrane after Blake et al. [27]) is metamorphosed to amphibolite and even granulite facies [12,28,29,30]. Blocks of metasedimentary, metavolcanic, and metamorphosed ultramafic rocks are juxtaposed in a block-on-block mélange with little matrix [5,12]. Metamorphic assemblages in ultramafic, pelitic, and calcareous blocks record a wide range of metamorphic conditions at 500–800 °C and 3.5–9 kbar [5,12,29,30,31,32]. Mixed conditions are consistent with high-grade blocks dismembered into the mélange followed by regional low- to medium-grade metamorphism [12]. Metabasalt blocks in the basement mélange are typically 45–55% SiO₂ [4,33,34], and elemental analyses are consistent with MORB and within-plate basalt (WPB) signatures, indicating oceanic floor and seamounts [4], although the Marble Mountains also contain some minor arc-related calc-alkaline and island arc tholeiite (IAT) metabasites [5].

Fossils in the basement mélange have been mostly identified as Triassic to latest Middle Jurassic with a few Carboniferous and Permian fossil sites [16], and ammonites originally interpreted as Devonian [35] are now interpreted as Late Triassic [36]. Based on Nd isotopes in argillaceous blocks, the RCT basement mélange incorporates Precambrian sediment that may have been transported by wind, but the depositional age of the blocks and the timing of their incorporation into the mélange is unknown [7]. Zircon U-Pb analyses from a para-amphibolite block in the RCT near the Wooley Creek batholith yielded few ages 212–154 Ma [23].

2.2.2. Cover Sequence

The cover sequence was defined by Wright and Wyld [4] in the southern RCT (Figure 1) and divided into the Salt Creek assemblage and the Dubakella Mountain assemblage, but other metavolcanic and metasedimentary sequences farther north may be equivalent [7,13,37,38]. The Salt Creek assemblage consists of basalt, chert, argillite, and clastic sedimentary rocks, consistent with deposition in a shallow marine environment [4]. The overlying Dubakella Mountain assemblage is divided into the augite porphyry and quartz porphyry volcanic members [4]. This assemblage was deposited near an active volcanic system and has been crosscut by intrusions, one of which is a quartz diorite with a U-Pb zircon age of 198 Ma and interpreted as part of the RCT intrusive suite [4]. Interlayered chert beds and blocks that Wright and Wyld [4] interpret to be dismembered equivalents of the cover sequence yield Late Triassic to Early Jurassic radiolarians [4,39,40].

Irwin et al. [41] mapped the area of Wright and Wyld’s [4] cover sequence as mafic volcanic-dominated sections of the basement mélange, scattered clastic rocks, limestone blocks, minor amphibolite blocks, and bodies of intermediate to silicic volcanic rocks >2 km across. Many details of the mapping do not agree between Wright and Wyld [4] and Irwin et al. [41], including the extents of lithologies and interpretations of basement mélange or volcanic or sedimentary sequences. However, in the northwest corner of each map, both studies map an area of volcanic rocks which are locally cut by intrusions interpreted by both as equivalent to the RCT intrusive suite [4,41]. This is the same location as the quartz diorite dated to 198 Ma described above [4].

The RCT cover sequence has been linked to North America based on the presence of quartzose and metamorphic clasts [4], isotopic similarity of argillites to major river systems [7], and abundance of pre-Mesozoic detrital zircon ages [13]. LaMaskin et al. [13] recently documented detrital zircon U-Pb MDAs from the cover sequence at ~170, 162, and 161 Ma, much younger than the previously mapped crosscutting relationships and fossil ages suggest [4,41]. Field relationships of the cover sequence need to be reexamined to determine the extent and continuity of the cover sequence compared to volcanic rocks of oceanic origin and if there are multiple phases of cover sequence deposition.

2.2.3. Intrusive Suite

The RCT intrusive suite is composed of gabbroic to quartz dioritic intrusions, ranging from sizes of <1 km² to sills and plutonic masses of >10 km² [42,43] which intrude the basement mélange [4,41] and cover sequence [4]. Zircon U-Pb ages of the intrusive suite are 207–193 Ma from quartz diorite, but more mafic plutonic bodies did not yield zircon and are undated [4]. The Saddle Gulch pluton has a hornblende K-Ar age of 189 Ma [40], but no zircon has been dated. Geochemistry of the more mafic components of the RCT intrusive suite is consistent with MORB and IAT [4]. The RCT intrusive suite both cuts the mélange and is incorporated into it, indicating multiple episodes of mélange incorporation [4].

2.2.4. Previous Tectonic Interpretations for the RCT

Two models for the construction of the WPzTrB have been proposed. The first is an exotic model in the Late Triassic for the RCT which interprets a succession of terranes in the western Klamath Mountains (RCT, WHT, and Western Klamath terrane) as an exotic archipelago resulting from west-dipping, intra-oceanic subduction [8,9,10,11,44,45,46,47,48]. These models are derived from geologic relationships of the terranes of the Sierra Nevada foothills and California Coast Ranges [44,45,49,50,51,52,53] and suggest that the mechanism responsible for the Jurassic deformation in the Sierra Nevada foothills is the collision and accretion of the exotic, intra-oceanic Western Jurassic belt and Coast Range ophiolite [45,47,48,54]. This interpretation suggests a Late Jurassic-Cretaceous suture runs along the entire western margin of North America [10,11]. In contrast, endemic models suggest that most WPzTrB terranes formed proximal to the North American margin [7,13,14,21,24,28,37,55,56,57,58,59,60] and that Late Jurassic deformation was driven by re-accretion of an endemic rifted arc rather than an exotic arc [13,24,37,55,56,57,60,61,62,63,64,65,66,67]. Some authors suggest coeval models but different along-strike subduction configurations for the Klamath Mountains and Sierra Nevada foothills [13,47,48].

Wright and Wyld [4] proposed that the RCT began forming in an oceanic fracture zone outboard from the North American continent based on the serpentinization of peridotite, the abundance of oceanic blocks, the mid-ocean ridge basalt and within-plate basalt geochemistry of blocks, and the lack of a continental signature. This fracture zone was reactivated as a subduction zone, driving 207–193 Ma RCT intrusive suite magmatism with MORB and IAT geochemical signatures and deposition of the volcanic and sedimentary sequences of the RCT cover sequence [4]. The 177–168 Ma WHT is a volcanic arc cover sequence that has been interpreted as originally depositional on the RCT and Eastern Hayfork terrane and has only minor potential intrusive equivalents exposed [26]. There is an apparent ~15 m.y. gap between the end of the RCT intrusive suite and the initiation of the WHT arc, and igneous rocks 193–177 Ma are largely unknown in the Klamath Mountains, possibly due to a phase of highly oblique convergence [14]. At 170 Ma, the Ironside Mountain batholith cut the Wilson Point thrust between the WHT and EHT, and the intrusion is sheared along the Salt Creek thrust where the WHT is thrust over the RCT, indicating some slip occurred after intrusion [68]. Folds, faults, and penetrative foliation in both the basement and cover sequence of the southern RCT indicate that a phase of regional northeast-southwest shortening likely occurred in the Middle Jurassic [4]. At least part of the cover sequence is 170–161 Ma and was sourced from Klamath terranes farther east as well as other sources in North America [13]. Magmatism moved farther north into the Wooley Creek suite which was active 167–156 Ma [69].

Dike complexes in the Preston Peak complex [64], China Peak complex [64], Onion Camp complex [59], outer Condrey Mountain schist [70] document extension as early as ca. 173–170 Ma. Intra-arc extension formed the 164–162 Ma Josephine ophiolite and rifted apart the RCT [57,58,59,71]. The 163–157 Ma Rogue–Chetco arc complex [67,72] developed on the western edge of the rift basin, and the 160–150 Ma Galice Formation was deposited in the basin during Josephine ophiolite formation and later compression [13,66,67]. During the Nevadan orogeny, the basin collapsed, and the RCT was thrust westward over the Western Jurassic belt along the Orleans fault at 153–151 Ma, causing low-grade regional metamorphism and partial melting of amphibolite in the RCT basement [37,56,57,73,74].

As detailed above, all WPzTrB terranes incorporate sediment from North America, supporting endemic formation models, except the RCT basement mélange which contains argillite blocks with eolian continental sediment which could have been far-traveled [7]. There is a clear correlation between the continent and the RCT cover sequence, but recent detrital zircon ages suggest this may be post accretion of the RCT to the continental margin, and there is no constraint on the timing of any continental input to the basement mélange.

3. Geology of the Study Area

We studied two areas of the RCT: the Marble Mountain area in the central Klamath Mountains (Figure 2) and the Rattlesnake Creek area in the southern RCT (Figure 3). Both areas are dominated by the RCT basement mélange and peridotite massifs. Almost all blocks are on the order of 10 s to 100 s of m in both areas. Field photos from the Marble Mountains area are shown in Figure 4A–E and photos from the Rattlesnake Creek area are shown in Figure 4F–I. Sample information is presented in

Table 1. Sample locations, contexts, and ages.

Sample Name	Location			Rock Type	Setting	Age (Ma)	Age Interpretation
	Latitude	Longitude	Study Area
7.13.20.2KM	41.61793°	−123.1366°	MM	Gneiss	Basement block	>176	MDA or Pb loss?
8.06.211DU	41.62076°	−123.163409°	MM	Gabbro	Dike	>164	Intrusion, Pb loss
8.04.212DU	41.62076°	−123.163409°	MM	Diorite	Dike	>164	Intrusion, Pb loss
RC23026	40.408989°	−123.295809°	RC	Metavolcaniclastic	EJ Cover sequence	201	MDA
8.5.3.22DU	40.41148°	−123.29727°	RC	Metachert	Olistolith	191	MDA
8.5.1.22DU	40.39742°	−123.12273°	RC	Sandstone	M-LJ Cover sequence	168	MDA
8.5.2.22DU	40.40637°	−123.30276°	RC	Pebbly sandstone	M-LJ Cover sequence	168	MDA
RC23029	40.404979°	−123.29669°	RC	Sandstone	M-LJ Cover sequence	167	MDA
RC23023B	40.409152°	−123.300718°	RC	Pebbly sandstone	M-LJ cover sequence olistolith	163	MDA
8.2.3.22DU	40.39128°	−123.31113°	RC	Micaceous quartzite	Olistolith	-	-

EJ—Early Jurassic; MDA—maximum depositional age; M-LJ—Middle to Late Jurassic; MM—Marble Mountain area; RC—Rattlesnake Creek area.

.

3.1. Marble Mountain Area

The Marble Mountain area in the north (Figure 2) has high-grade metamorphic blocks with little matrix. Block lithologies include ultramafic to metaultramafic, amphibolite, and metasedimentary. Rare block boundary exposures commonly have a mm- to cm-scale layer of serpentinite, but outcrops dominated by serpentinite are rare. The mélange is cut by gabbroic and dioritic intrusions with irregular boundaries. Foliation is weak or absent in most intrusive rocks, but they are locally lineated. Gabbro dike sample 8.06.211DU is primarily composed of anorthite with ~15% amphiboles, and lineated diorite dike sample 8.04.212DU is ~50% amphibole and ~50% anorthite and albite.

Ultramafic and metaultramafic outcrops are blocky and nonfoliated, and serpentinite veins crosscut some blocks in a mesh texture (Figure 4A,B). Olivine and orthopyroxene alteration ranges from thin crosscutting veins of serpentine mesh to almost complete replacement by serpentine and talc ± anthophyllite with associated magnetite. Previous work in the area has interpreted that anthophyllite represents a high-T metamorphic event before mélange dismemberment [12].

Metasedimentary and amphibolite blocks are strongly foliated and even gneissic with cm- to mm-scale banding. Amphibolite blocks are composed of amphibole and altered plagioclase. Clinopyroxene is locally present, and quartz is rare or absent, and these blocks are interpreted as metabasites. Metasedimentary blocks include schist, phyllite, slate, and minor quartzite and marble. Schist blocks are dominated by recrystallized quartz and banded with variously plagioclase, amphibole, biotite, chlorite, muscovite, epidote minerals, and minor calcite, titanite, garnet, and potassium feldspar (Figure 4C). Plagioclase and amphibole are abundant in some samples and scarce in others, and amphiboles are often dark green in outcrop. Minor marble is foliated with dismembered quartz ribbons with minor muscovite and may represent interlayered limestone and chert protoliths (Figure 4D). Sample 7.13.20.2KM is a gneiss block with banded light layers with abundant quartz and plagioclase, dark layers with abundant hornblende and biotite, and muscovite and epidote group minerals throughout (Figure 4E). Given the prevalence of gneissic quartz-rich metasedimentary blocks, we tentatively interpret this sample as metasedimentary.

3.2. Rattlesnake Creek Area

The Rattlesnake Creek area in the south (Figure 1 and Figure 3) is composed of subgreenschist facies blocks in a serpentinite matrix and sections of deformed stratigraphy. Block lithologies include metaultramafic rocks, metabasite, metachert, sandstone, phyllite, and marble. The 193 Ma Bear Wallow pluton [43] cuts the basement mélange on the west side of the Rattlesnake Creek area, and the 159–150 Ma Galice Formation of the Western Jurassic belt [13,66] is exposed in the southwestern and southeastern corners (Figure 3). The serpentinite matrix is often poorly exposed but is commonly found on the edges of blocks in float. Deformation fabrics and calcite veins crosscutting the serpentinite matrix rarely have a preferred orientation.

Metavolcanic blocks are dominantly weakly foliated to massive metabasite in serpentinite matrix. Metabasite blocks are composed of a plagioclase lath groundmass with amphibole, pervasive chlorite alteration, and opaque minerals. Some blocks have calcite amygdules, and secondary epidote minerals are associated with fractures. Metachert blocks are mostly gray but sometimes red and are exposed in serpentinite matrix or as larger blocks with no matrix contacts exposed. In the northeast of the Rattlesnake Creek area, a depositional relationship between chert and marble is exposed.

In the southern part of the Rattlesnake Creek area, a large body of metasedimentary rocks lacks outcrops of serpentinite. Most of the outcrop in this area is slate or phyllite with minor interlayered metasandstone and metachert. Marble is in apparent relict depositional contact with quartzite, and both have interlayered phyllite. Micaceous quartzite sample 8.2.3.22DU is dominantly quartz with minor chlorite, detrital muscovite, and feldspar (Figure 5A). Quartz grains are roughly equal proportions of mono- and polycrystalline quartz. Although the outcrops are strongly foliated and often folded (Figure 4F), there are no clear block and matrix relationships, suggesting this may be a large olistostromal block or part of a depositional cover sequence.

In the central part of the Rattlesnake Creek area, there is another concentration of metaclastic, metachert, and metavolcanic rocks. As in the metaclastic body in the southern part, this body is dominated by shale/slate interlayered with dismembered metasandstone and metachert. Samples RC23023B and 8.5.3.22DU (Figure 3) were collected from a block-in-matrix mélange where blocks of chert, shale, sandstone +/− shale, and metavolcanics are in a shale matrix (Figure 4G). Chert blocks have bedding or relict bedding, but one in the shale matrix mélange near sample RC23023B is brecciated chert with local metavolcanic lithics and chlorite in fractures. Samples 8.5.1.22DU, 8.5.2.22DU, RC23023B, and RC23029 were collected from pebbly and coarse sandstone beds and blocks with angular to subrounded clasts of quartz, plagioclase, and rock fragments including chert, volcanic clasts of various textures, mudstone, siltstone, and a range of schists including muscovite, chlorite, and/or epidote group minerals (Figure 5B,C). Sample RC23026 is from an outcrop of coarse-grained, massive, chloritized, volcaniclastic rock. Lithic clasts include volcanic clasts, metavolcanic clasts, and minor chert, quartzite, and mudstone (Figure 5D–F). It is in close proximity to a dioritic/gabbroic intrusion and more fine-grained sandstone and shale, although contacts between the lithologies are not exposed and no serpentinite was observed at this outcrop.

Most contacts between the metaclastic rocks and serpentinite of the basement mélange are covered (Figure 4H), but they are mostly not parallel to the metaclastic layering, more consistent with a tectonic contact than a depositional one. This is further supported by strong deformation and boudinage in some outcrops of the metcalstic rocks (Figure 4I) and more minor folding in others (Figure 4F). In the northeastern corner of the central metaclastic body, dark shale interbedded with fine-grained sandstone contains a layer of serpentinite ~2 m thick with a strong preferred orientation roughly parallel to the surrounding bedding and cut by calcite veins. This relationship suggests a tectonic shear zone where the metaclastic section was being dismembered along a bedding plane.

4. Methods

4.1. Point Counting

Modal abundances were determined for all six clastic and volcaniclastic samples that yielded zircon. Gneiss sample 7.13.20.2KM is interpreted as metasedimentary but too deformed for sedimentary point counting. We counted 450 grains in thin section from each sample using the Gazzi-Dickinson point counting method [77,78,79]. Recalculated abundances are presented in Table S1 in the Supplementary Material, and results are plotted on a ternary diagram with the tectonic fields of Dickinson et al. [80] in Figure 6.

4.2. Zircon U-Pb Geochronology

Zircons were separated by standard methods at the Arizona LaserChron Center at the University of Arizona. Ten samples yielded sufficient zircon for analysis. Backscattered electron imaging (BSE) was used to avoid inclusions in the zircons. Data were collected by laser ablation multi-collector inductively coupled plasma mass spectrometry (LA-MC-ICP-MS) [81] and reduced using the Matlab tool AgeCalc (v. 73a) [82].

The U-Pb data were filtered in two steps. First, commonly used filters were applied to data at 2σ error including both internal and external uncertainty, with discordance representing the mismatch between the ²³⁸U/²⁰⁶Pb and ²⁰⁶Pb/²⁰⁷Pb ages. The ²⁰⁴Pb limit was 600 counts per second, the ²³⁸U/²⁰⁶Pb and ²⁰⁶Pb/²⁰⁷Pb age uncertainty limit was 10%, the ²⁰⁶Pb/²³⁸U over-dispersion factor cutoff 0.6, and the ²⁰⁶Pb/²⁰⁷Pb OD factor was cutoff at 1. For analyses with ²³⁸U/²⁰⁶Pb age >700 Ma, the discordance limit was 20% and the reverse discordance limit was 5%. Discordance filters were not applied to younger ²³⁸U/²⁰⁶Pb ages because ²⁰⁶Pb/²⁰⁷Pb age errors increase with younger ages. Analyses rejected using the criteria from this first step are shown in black on concordia diagrams in Figure 7.

Some discordant analyses rejected in the first step of filtering fall along linear trends below concordia, indicating Pb loss, but ²³⁸U/²⁰⁶Pb and ²⁰⁶Pb/²⁰⁷Pb concordance is not useful for Phanerozoic ages because of the large error and low accuracy of ²⁰⁶Pb/²⁰⁷Pb ages. To create a second filter that would apply to all ages, we calculated non-iterative probability (NIPr) ages, minimum segmented discordance, and concordance classes [83]. This method uses all three sets of ratios (²⁰⁶Pb/²³⁸U, ²⁰⁷Pb/²³⁵U, ²⁰⁶Pb/²⁰⁷Pb) to evaluate concordance over the whole range of ages. In the second filter, analyses with concordance class 5 or higher were rejected and are shown in red in Figure 7, and the remaining accepted analyses that pass both filtering steps are shown in blue. Although there can still be hidden Pb loss [84], clearly perturbed analyses are removed for the whole range of ages with the combination of filtering steps. Full geochronology results are presented in Table S1 in the Supplementary Material. All concordia diagrams were made with IsoplotR (v. 6.8) [85].

MDAs were calculated for the Rattlesnake Creek area samples and gneiss sample 7.13.20.2KM from the Marble Mountains area which may be a paragneiss. Table 2 presents the youngest single grain (YSG), the weighted mean of the youngest cluster of three or more grains overlapping in 2σ error (YGC3+ (2σ)), the youngest statistical population (YSP) [86], and the maximum likelihood age (MLA) [87]. Weighted mean (WM) ages were calculated for igneous samples 8.06.211DU and 8.04.211DU (Table 2). All sample ages were calculated using IsoplotR (v. 6.8) [85], and plots are presented in Figure S1 in the Supplementary Material.

Samples were visually examined using cumulative distribution plots (CDPs), probability density plots (PDPs), kernel density estimates (KDEs), and histograms made in detritalPy (v. 1.4) [88]. Additionally, multi-dimensional scaling (MDS) analysis provides a visual representation of the similarity of samples with each other and potential sources or correlatives. We use the Kolmogorov–Smirnow (K-S) Test D distance as the similarity between KDEs. Using three dimensions, the metric stress value of 0.070336 from the Shepard plot represents a fair to good fit [89]. MDS plots were made in DZmds (v. 1.11) [90].

Figure 7. Wetherill concordia diagrams with 2σ ellipses for zircon samples from the RCT. Black ellipses were rejected by the standard filtering described in the Methods section. Red ellipses were rejected by concordance class 5–7. Blue ellipses are the remaining data with concordance class 1–4. Plots were made with IsoplotR [85].

Figure 7. Wetherill concordia diagrams with 2σ ellipses for zircon samples from the RCT. Black ellipses were rejected by the standard filtering described in the Methods section. Red ellipses were rejected by concordance class 5–7. Blue ellipses are the remaining data with concordance class 1–4. Plots were made with IsoplotR [85].

Table 2. Maximum depositional ages, weighted mean ages, and zircon age abundances.

Sample Name	YSG	YGC3+ (2σ)			YSP			MLA	n	%
	Age ± 2σ (Ma)	Age ± 2σ (Ma)	MSWD	n	Age ± 2σ (Ma)	MSWD	n	Age ± 2σ (Ma)		Mz	Pz	PtZ	A	pre-Mz
7.13.20.2KM	156.0 ± 12.6	175.2 ± 1.3	0.86	41	175.8 ± 1.2	1.1	44	176.0 ± 1.6	79	97.5	0.0	2.5	0.0	2.5
RC23026	184.7 ± 17.0	200.9 ± 4.2	1.4	12	199.8 ± 4.3	1.3	10	201.1 ± 4.6	18	66.7	5.6	27.8	0.0	33.3
8.5.3.22DU	179.5 ± 7.5	189.27 ± 0.71	0.62	97	191.01 ± 0.61	1	133	191.61 ± 0.67	185	99.5	0.0	0.5	0.0	0.5
8.5.1.22DU	163.9 ± 4.3	168.6 ± 1.7	1.4	9	167.7 ± 1.9	1.1	7	168.6 ± 3.2	193	15.0	19.2	56.5	9.3	85.0
8.5.2.22DU	164.9 ± 5.1	168.7 ± 1.8	1.5	9	168.1 ± 1.8	1.1	8	168 ± 3	189	12.7	29.6	49.2	8.5	87.3
RC23029	164.7 ± 4.1	167.7 ± 0.8	2.1	9	166.6 ± 1.1	1.2	6	167.4 ± 1.5	204	9.8	17.2	67.2	5.9	90.2
RC23023B	160.6 ± 4.6	164.4 ± 1.4	1.4	9	162.3 ± 1.9	1.1	5	164.1 ± 2.3	135	21.5	23.0	55.6	0.0	78.5
8.2.3.22DU	467.4 ± 17.9	-	-	-	-	-	-	-	158	1.3	0.6	89.2	8.9	98.7
Sample Name		Weighted mean							n	%
		Age ± 2σ (Ma)	MSWD	n						Mz	Pz	PtZ	A	pre-Mz
8.06.211DU		164.09 ± 0.64	4.4	60					61	100.0	0.0	0.0	0.0	0.0
8.04.212DU		164.3 ± 1.0	6.6	20					23	91.3	0.0	8.7	0.0	8.7

A—Archean; MLA: maximum likelihood age [87]; MSWD—mean square of weighted deviates; Mz—Mesozoic; n—number of analyses; PtZ—Proterozoic; Pz—Paleozoic; YGC3+ (2σ)—youngest grain cluster of three or more ages overlapping within 2σ; YSG—youngest single grain; YSP—youngest statistical population [86].

Table 2. Maximum depositional ages, weighted mean ages, and zircon age abundances.

Sample Name	YSG	YGC3+ (2σ)			YSP			MLA	n	%
	Age ± 2σ (Ma)	Age ± 2σ (Ma)	MSWD	n	Age ± 2σ (Ma)	MSWD	n	Age ± 2σ (Ma)		Mz	Pz	PtZ	A	pre-Mz
7.13.20.2KM	156.0 ± 12.6	175.2 ± 1.3	0.86	41	175.8 ± 1.2	1.1	44	176.0 ± 1.6	79	97.5	0.0	2.5	0.0	2.5
RC23026	184.7 ± 17.0	200.9 ± 4.2	1.4	12	199.8 ± 4.3	1.3	10	201.1 ± 4.6	18	66.7	5.6	27.8	0.0	33.3
8.5.3.22DU	179.5 ± 7.5	189.27 ± 0.71	0.62	97	191.01 ± 0.61	1	133	191.61 ± 0.67	185	99.5	0.0	0.5	0.0	0.5
8.5.1.22DU	163.9 ± 4.3	168.6 ± 1.7	1.4	9	167.7 ± 1.9	1.1	7	168.6 ± 3.2	193	15.0	19.2	56.5	9.3	85.0
8.5.2.22DU	164.9 ± 5.1	168.7 ± 1.8	1.5	9	168.1 ± 1.8	1.1	8	168 ± 3	189	12.7	29.6	49.2	8.5	87.3
RC23029	164.7 ± 4.1	167.7 ± 0.8	2.1	9	166.6 ± 1.1	1.2	6	167.4 ± 1.5	204	9.8	17.2	67.2	5.9	90.2
RC23023B	160.6 ± 4.6	164.4 ± 1.4	1.4	9	162.3 ± 1.9	1.1	5	164.1 ± 2.3	135	21.5	23.0	55.6	0.0	78.5
8.2.3.22DU	467.4 ± 17.9	-	-	-	-	-	-	-	158	1.3	0.6	89.2	8.9	98.7
Sample Name		Weighted mean							n	%
		Age ± 2σ (Ma)	MSWD	n						Mz	Pz	PtZ	A	pre-Mz
8.06.211DU		164.09 ± 0.64	4.4	60					61	100.0	0.0	0.0	0.0	0.0
8.04.212DU		164.3 ± 1.0	6.6	20					23	91.3	0.0	8.7	0.0	8.7

A—Archean; MLA: maximum likelihood age [87]; MSWD—mean square of weighted deviates; Mz—Mesozoic; n—number of analyses; PtZ—Proterozoic; Pz—Paleozoic; YGC3+ (2σ)—youngest grain cluster of three or more ages overlapping within 2σ; YSG—youngest single grain; YSP—youngest statistical population [86].

5. Results

Three samples from the Marble Mountain area and seven from the Rattlesnake Creek area yielded sufficient zircon for analysis (Figure 1 and Figure 3), producing 1245 new ages from the RCT. Here, we report and discuss weighted means ages for igneous samples and both MLA and YSP MDAs for detrital samples. All errors are reported at 2σ.

5.1. Marble Mountain Area

Three samples from the Marble Mountain area in the north yielded zircon (Figure 1 and Figure 8). Sample 7.13.20.2KM (YSP: 175.8 ± 1.2 Ma, MLA: 176.0 ± 1.6 Ma) is a gneiss with a shoulder peak at 194 Ma and only two grains (582 Ma and 1180 Ma) >250 Ma. Sample 8.06.211DU (WM: 164.09 ± 0.64 Ma) is a gabbro dike, and all but one outlier grain are included in the 164 Ma weighted mean age. Sample 8.04.212DU (WM: 164.3 ± 1.0 Ma) is a lineated diorite dike with four older ages at 199, 247, 1293, and 1315 Ma.

5.2. Rattlesnake Creek Area

In the Rattlesnake Creek area of the southernmost RCT, seven sedimentary, metasedimentary, and metavolcaniclastic samples yielded zircon (Figure 1, Figure 3 and Figure 8). The oldest sample is metavolcaniclastic sample RC23026 (YSP: 199.8 ± 4.3 Ma, MLA: 201.1 ± 4.6 Ma) from the southern RCT which yielded 18 grain ages. The main peak at ~201 Ma comprises 67% of the ages, but 33% are pre-Mesozoic and as old as 1874 Ma. Sample 8.5.3.22DU (YSP: 191.01 ± 0.61 Ma, MLA: 191.61 ± 0.67 Ma) is a metachert in the southern RCT with only one outlier grain at 1836 Ma.

Sample 8.5.1.22DU (YSP: 167.7 ± 1.9 Ma, MLA: 168.6 ± 3.2 Ma) is a sandstone with a main peak at 174 Ma, a shoulder peak at 188 Ma, and older peaks at 251, ~1075, ~1400, and ~1775 Ma. Ages are 15% Mesozoic, 19% Paleozoic, 57% Proterozoic, and 9% Archean. Sample 8.5.2.22DU (YSP: 168.1 ± 1.8 Ma, MLA: 168 ± 3 Ma) is a pebbly sandstone with a main peak at 168 Ma with a tail to 200 Ma and older peaks at 379 and ~1850 Ma. Ages are 12% Mesozoic, 29% Paleozoic, 51% Proterozoic, and 8% Archean. Sample RC23029 (YSP: 166.6 ± 1.1 Ma, MLA: 167.4 ± 1.5 Ma) is a sandstone with a main peak at 169 Ma and older peaks at 255, 361, and 1760 Ma. Ages are 10% Mesozoic, 17% Paleozoic, 67% Proterozoic, and 6% Archean. Sample RC23023B (YSP: 162.3 ± 1.9 Ma, MLA: 164.1 ± 2.3 Ma) is a sandstone with a main peak at 167 Ma and an older peak at 253 Ma. Ages are 21% Mesozoic, 23% Paleozoic, 56% Proterozoic, and 0% Archean.

Finally, sample 8.2.3.22DU is a micaceous quartzite with age peaks at 1096, 1050, 1235, 1300, 1440, 1733, 1746, and 2674 Ma. There are four scattered young ages at 76, 78, 467, and 655 Ma. The Cretaceous ages are younger than tectonic events in the WPzTrB and are likely a result of contamination during mineral separation. Excluding these two young ages, ages are 0% Mesozoic, <1% Paleozoic, 91% Proterozoic, and 8% Archean. The lack of a consistent young population and the large proportion of Proterozoic and Archean zircons indicates that an MDA calculation would not reflect the age of incorporation into the RCT.

6. Discussion

6.1. RCT Age and Provenance

Detrital zircon ages from the RCT indicate that there are significant differences in time and space. All but one sedimentary and metasedimentary samples collected from the RCT which yielded sufficient zircon are from the southernmost RCT. Metavolcaniclastic sample RC23026 has the oldest MDA at ~201 Ma (Table 2), coeval with the quartz diorites (207–193 Ma) in the RCT intrusive suite [4]. Since this sample is volcaniclastic and the RCT intrusive suite arc magmatism was active as young as 193 Ma, we interpret that it deposited during active volcanism and that the MDA is close to the true depositional age. Although this sample only yielded 18 ages, six of them (33%) are scattered between 250 and 1874 Ma, consistent with a continental source, but there are not enough ages to positively link with any particular continent or region. A continental source component is further supported by the presence of sedimentary and quartz-rich metamorphic clasts (Figure 6). Because contacts with other lithologies are not exposed, it is unclear if sample RC23026 is a block in the serpentinite matrix or part of the central metaclastic body in the Rattlesnake Creek area. The abundance of volcanic lithic clasts and the 201 Ma MDA suggest deposition of an Early Jurassic cover sequence on the RCT basement mélange proximal to the RCT intrusive suite. Pre-Mesozoic zircon ages are interpreted to signify a sediment pathway from a continental source to the RCT at the time of active RCT magmatism. Metachert sample 8.5.3.22DU has an MDA of ~191 Ma (Table 2), younger than documented ages in the RCT suite. Only one of the 185 ages is older than this age peak, so a continental source is not required. Zircon grains in chert are more likely to be sourced from volcanic airfall, so the MDA is likely close to the true depositional age. The metachert is a block from an interpreted olistostromal deposit with minor shale matrix, so the original depositional context is unclear. Samples RC23026 and 8.5.3.22DU both plot on the left side of the MDS plot with the RCT intrusive suite (Figure 9).

Gneiss sample 7.13.20.2KM in the Marble Mountain area (Figure 2) has an MDA of ~176 Ma (Table 2), similar to the oldest hornblende ages in the WHT [24,25], and a shoulder peak at ~194 Ma, consistent with the youngest ages of the RCT intrusive suite [4]. This gneiss is interpreted as metasedimentary based on the abundance of quartz in gneissic blocks, but more work is needed to distinguish between an orthogneiss and a paragneiss. One possible interpretation of the ages for sample 7.13.20.2KM is that the gneiss protolith was deposited on the RCT, incorporating zircon from the active WHT arc and the eroding RCT intrusive suite. However, the concordia plots (Figure 7) show that ages are spread for all three samples from the Marble Mountain area, potentially indicating Pb loss during one or more later phase of metamorphism [29]. In this case, the shoulder peak at 194 Ma may better represent the age of the protolith which may have been part of an Early Jurassic RCT cover sequence. Also in the Marble Mountain area, gabbro dike sample 8.06.211DU and lineated diorite dike sample 8.04.212DU have weighted mean ages of ~164 Ma, overlapping with the extensive 167–156 Ma Wooley Creek intrusive suite [69]. However, the MSWD values (4.4 and 6.6) indicate the zircons in our dike samples may also be affected by Pb loss to an extent, so the true intrusion age may be >164 Ma. All three samples from the Marble Mountains area plot in the lower center of the MDS plot, although gneiss sample 7.13.20.2KM plots closer to the RCT intrusive suite (Figure 9).

In the Rattlesnake Creek area to the south (Figure 3), sandstone samples 8.5.1.22DU, 8.5.2.22DU, RC23029, and RC23023B have MDAs 168–163 Ma and 79–90% pre-Mesozoic grains consistent with a North American provenance (Table 2). Because magmatism was active throughout the Middle Jurassic and volcanic lithic clasts are abundant in the sandstones, we interpret that the MDAs closely approximate the true depositional ages. These samples had good zircon yield (n = 135–204) and were deposited shortly after WHT and Ironside Mountain batholith magmatism and thus after accretion of the RCT to North America. Sample RC23023B (MDA~163 Ma) is a pebbly sandstone block in a shale matrix that is not strongly tectonized (Figure 4F), more consistent with sedimentary mélange formation by mass wasting than tectonic incorporation. Detrital zircon ages agree with those LaMaskin et al. [13] documented in the cover sequence ~15 km to the east (Figure 1 and Figure 8) and plot closely together on the right side of the MDS plot (Figure 9). We interpret that our Middle Jurassic sandstone samples represent a dismembered western extent of the Middle to earliest Late Jurassic cover sequence.

Micaceous quartzite sample 8.2.3.22DU has 99% pre-Mesozoic detrital zircon ages (Table 2). Notably absent is the 360–380 Ma age peak present in most of the young southern samples and the RCT cover sequence (Figure 8). Older age peaks match well with those in the four Middle-Late Jurassic cover sequence samples and the RCT cover sequence [13] (Figure 7 and Figure 8). Sample 8.2.3.22DU is likely an olistostromal block, but its ages are inconsistent with the large olistostromal blocks in the EHT mélange which are dominated by ages >1500 Ma [21] and must represent a different North American source. Fossil ages from limestone and marble blocks in the area are interpreted as Permian or early Mesozoic, Late Triassic or Early Jurassic, and undetermined [16] (localities 59 and 60). Since the depositional age of sample 8.2.3.22DU is undetermined, it is unclear if there are significant age outliers within the southern zone of siliciclastic material (Figure 3). Although the lithologies are similar to the Middle Jurassic central metaclastic body in the Rattlesnake Creek area, the southern metaclastic body is more metamorphosed, permissibly Early Jurassic, and could be further evidence of the early proximity of the RCT to North America. More age constraints are necessary to date the southern metaclastic body and its incorporation into the RCT.

Overall, siliciclastic rocks are a minor but informative component of the RCT. The distribution of metaclastic zircon ages could be an artifact of limited sampling, but data from this study and LaMaskin et al. [13] suggest that the RCT incorporated 33% continental zircon at ~201 Ma, effectively none at ~191 Ma, and 79–93% from 170 to 161 Ma. North Fork terrane sandstones and Eastern Hayfork terrane matrix sandstones to the east have similar age peaks to the Middle Jurassic cover sequence (Figure 8), although all samples have low zircon yield (n = 15–67) [18,20], making it challenging to correlate them. RCT formation overlaps with 225–180 Ma deposition in the North Fork terrane [14,20] and begins approximately coeval with the end of Eastern Hayfork terrane subduction accretion in the Late Triassic [16,18]. Interestingly, the Galice Formation is at the center of all MDS groups (Figure 9), consistent with the interpretation that the WPzTrB was one of the sources of Galice Formation zircon [13,66,67].

6.2. Subduction Vergence

Many tectonic models for the Early Jurassic margin of western North America include the Mezcalera plate subducting to the east beneath North America and to the west beneath an intraoceanic arc [8,10,11,45]. However, this does not seem to be consistent with the EHT and RCT subduction complexes. We might expect opposite vergence deformation fabrics or a zone of mixing as the two terranes came into contact. However, after ~200 Ma, there are no similarities in age or composition of blocks in the mélanges, and both terranes have generally east-dipping fabrics. The southern EHT in particular is characterized by large quartz-rich blocks with detrital zircon ages >1500 Ma [21], a pattern not documented anywhere in the RCT. We might also expect to see a larger proportion of material accreted into subduction complexes as the narrowing ocean filled with sediment shed from two approaching terranes, but the age of collision between the EHT and RCT is not well documented by siliciclastic or volcanic blocks in either mélange. If the RCT formed distal to North America, subduction accretion in the EHT would have continued well after initiation of the RCT intrusive suite as the ocean between the trenches closed. Instead, the minimal overlap in ages of accreted blocks and presence of continental input to the RCT at ~201 Ma is more consistent with the initiation of east-dipping subduction beneath the RCT marginal to North America. As a result, the Sigloch and Mihalynuk [10,11] Late Jurassic-Cretaceous suture model is not supported by the geology in the Klamath Mountains and further challenges the validity of this model along strike to the north and south along the Cordilleran margin [13,66,67].

6.3. The RCT Cover Sequence

In the Rattlesnake Creek area, we propose that there is evidence for two phases of cover sequence deposition, one in the Early Jurassic and one in the Middle Jurassic to earliest Late Jurassic. Volcanic and volcaniclastic rocks are only a small component of the Rattlesnake Creek area cover sequence in contrast with the mostly volcanic Salt Creek and Dubakella Mountain assemblages of the cover sequence defined by Wright and Wyld [4] ~15 km to the east. Volcaniclastic sample RC23026 has an MDA of ~201 Ma, consistent with Early Jurassic cover sequence deposition during RCT intrusive suite magmatism and the crosscutting relationships documented farther east [4,41]. Our four youngest samples agree well with the 170–161 Ma MDAs and older ages found by LaMaskin et al. [13] to the east. The age peaks at ca. 1.1, 1.4, and 1.75 Ga in the RCT cover sequence are a strong indicator of southwestern Laurentia provenance [91,92,93,94,95], and LaMaskin et al. [13] determined that the provenance of the RCT cover sequence is most consistent with a mixture of sediment from older Klamath and Sierra Nevada terranes, transcontinental sources, and southwestern US sources. Outcrop lithologies and our geochronology suggest that the cover sequence in the Rattlesnake Creek area is dominantly Middle-Late Jurassic with limited exposure of the Early Jurassic sequence.

Poorly exposed relationships between the basement mélange and sedimentary rocks in the Rattlesnake Creek area are generally not parallel to layering and appear to be tectonic rather than depositional. Samples 8.5.3.22DU, RC23023B, and 8.2.3.22DU are also interpreted as olistostromes. We propose that most original contacts between the cover sequences and the basement mélange have been disrupted by tectonic and sedimentary processes which we discuss more in Section 6.4 below. Because of this disruption, there is a possibility that the cover sequences were originally deposited elsewhere and later incorporated into the RCT mélange. However, sample RC23026 was likely deposited proximal to the RCT intrusive suite as discussed in Section 6.1 above, and there are no known correlatives to the Middle-Late Jurassic cover sequence deposits associated with any another Klamath terrane. We interpret that all cover sequence deposits were originally depositional on the RCT.

In the Marble Mountain area, metasedimentary blocks are abundant. Although these include marble and metachert which may represent blocks of ocean plate stratigraphy, there also blocks of schist and gneiss. Gneiss sample 7.13.20.2KM is the only basement block which yielded zircon, but it was likely affected by Pb loss during metamorphism as discussed in Section 6.1 above. If the protolith age is represented by the 194 Ma age peak, this may be evidence for Early Jurassic cover sequence in the central RCT. Other rocks such as the Lake Mountain metasediments of Barrows [38] may also be equivalent to the cover sequence, but more work is needed to better understand the ages and protoliths of metasedimentary blocks in the central RCT.

6.4. Mélange Formation in the RCT

The RCT has undergone multiple phases of mélange formation and dismemberment, but how blocks younger than the RCT intrusive suite (<193 Ma) were incorporated is unclear. Significant mélange dismemberment and mixing predates the RCT intrusive suite which crosscuts the basement mélange, and most mélange blocks are dismembered ocean plate stratigraphy. Initial mélange formation has been interpreted to have originated in an oceanic fracture zone which was reactivated as a subduction zone, forming the RCT intrusive suite [4]. These relationships require that the initial RCT mélange was in an upper plate position above a subduction zone in the Late Triassic to Early Jurassic. Blocks of the intrusive suite are also found within the basement mélange, so mélange formation must have continued after arc magmatism [4] perhaps as late as earliest Late Jurassic based on fossils [16] and zircon ages of the Middle-Late Jurassic cover sequence ([13] and this study).

6.4.1. Subduction Accretion

Suprasubduction mélanges often form by subduction accretion at the trench. In this setting, classic subduction zone high-pressure, low-temperature metamorphism might be expected, but blueschist has not been documented in any portion of the RCT. However, the EHT, which is often interpreted as a subduction complex [17,21,22,96,97], contains little blueschist, and there are only a few documented locations with high-temperature, low-pressure metamorphism in the subduction complex in the Yarlung suture zone between India and Asia [98,99,100]. Blueschist facies metamorphism is diagnostic of subduction accretion, but its absence does not necessarily preclude subduction accretion. That being said, convergence may have been oblique at times [14,56], and the magmatic gap between the RCT intrusive suite and Western Hayfork terrane magmatism may indicate a period of transpression. Oceanic blocks could have accreted with only minor underthrusting and/or strain could have been accommodated in the weak basement mélange, disrupting Early Jurassic intrusive and stratigraphic relationships in the RCT. While there is no concrete evidence for subduction accretion in the RCT, we caution against rejecting this mechanism of mélange formation based on metamorphic conditions alone, especially since much of the central RCT has experienced multiple episodes of metamorphism [12,28,29,60].

6.4.2. Siskiyou Thrusting

The Salt Creek thrust placed the WHT over the RCT before being intruded by the ~170 Ma Ironside Mountain Batholith [68] (Figure 1). This episode of thrusting could have contributed to dismemberment of the existing basement mélange, but it is unlikely that new blocks were incorporated during this phase of deformation. With the possible exception of gneiss sample 7.13.20.2KM in the Marble Mountain, no documented blocks in the RCT are consistent with incorporation of the WHT, and the Middle-Late Jurassic cover sequence postdates the Salt Creek thrust. However, Siskiyou thrusting also structurally buried much of the RCT, making it challenging to deposit and/or incorporate new material without significant exhumation.

6.4.3. Jurassic Rifting

Extension in the RCT began as early as ca. 173–170 Ma [59,64,70], and intra-arc extension formed the 164–162 Ma Josephine ophiolite [57,58,71]. The Onion Camp complex within the Western Jurassic belt has been interpreted as a rifted portion of the RCT and is associated with the Late Jurassic Fiddler Mountain olistostrome [59]. Rifting could have exhumed the basement mélange and created steep topographic slopes that promoted mass wasting. Olistostromal deposits in the Rattlesnake Creek area are older and more continental than the Fiddler Mountain olistostrome [59], consistent with the Rattlesnake Creek area being on the eastern edge of the rift basin, more proximal to the continent. We agree with the interpretation of LaMaskin et al. [13] of an extensional rift setting for deposition of the Middle-Late Jurassic cover sequence. However, the lack of clear depositional relationships between the cover sequence and the basement mélange and evidence of mixing m-scale blocks by sedimentary processes may indicate that much of the cover sequence and blocks with Middle Jurassic fossils were dismembered by rifting and mass wasting. Subsequent closure of the Josephine ophiolite basin and thrusting of the Nevadan orogeny may have further disrupted the Middle-Late Jurassic cover sequence.

6.5. Proposed Tectonic Model

The RCT basement mélange may have begun as an oceanic transform system which was reactivated as a subduction zone (Figure 10A), driving the 207–193 Ma RCT intrusive suite magmatism which crosscuts the basement mélange and a section interpreted as the RCT cover sequence [4]. At ~201 Ma, volcaniclastic sample RC23026 with metamorphic clasts and 33% continental detrital zircon ages was deposited in the southern RCT (Figure 1 and Figure 3), indicating a sediment pathway from North America while final subduction accretion was occurring in the EHT mélange (Figure 10B). This sample is consistent with an initial Early Jurassic phase of cover sequence deposition [4]. A ~191 Ma metachert incorporated into the southern RCT contains no evidence of continental input, potentially indicating disruption of the sediment pathway to the RCT after RCT intrusive suite magmatism (Figure 10B). A metasedimentary gneiss in the northern study area has an MDA of 176 Ma and a shoulder peak at ~194 Ma. This represents either incorporation of Western Hayfork terrane and RCT intrusive suite zircon or a block of the Early Jurassic cover sequence that experienced Pb loss during Middle-Late Jurassic metamorphism. Regardless, there are only two pre-Mesozoic ages, so a continental source is not required. Magmatic quiescence between the 207–193 Ma RCT intrusive suite and Western Hayfork magmatism may indicate a period of transpression in the Early Jurassic.

After accretion of the RCT to the continental margin, the 177–168 Ma WHT arc is interpreted to have formed on the amalgamated RCT and EHT [26], and the hinterland was uplifted along west-verging thrust faults during the Siskiyou contractional event, likely indicating the resumption of eastward subduction beneath the RCT along the western margin of North America (Figure 10C). The crosscutting 170 Ma Ironside Mountain batholith places a minimum constraint on the timing of slip along the Wilson Point thrust which thrust the EHT over the WHT and the Salt Creek thrust which thrust the WHT over the RCT [68] (Figure 10D). Rifting began as early as 173 Ma (Figure 10C) and culminated with the formation of the Josephine ophiolite at 164–162 Ma (Figure 10E). At 170–161 Ma, the Middle-Late Jurassic RCT cover sequence was deposited on top of the southern RCT basement mélange, potentially on basement exposed by rifting (Figure 10D,E). The nearby EHT and NFT likely contributed some sediment to the cover sequence (Figure 9), but significant Proterozoic and Archean ages (Table 2; Figure 8) suggest additional transcontinental and southwest US sources [13]. In the cover sequence ~15 km east of the Rattlesnake Creek area (Figure 1), Wright and Wyld [4] interpreted deposition in rift basins that were later inverted. More detailed mapping is needed to determine whether the rifting occurred during the Early Jurassic or Middle-Late Jurassic phase of the cover sequence deposition and how much the cover sequence was disrupted by later deformation. Dikes in the central RCT crosscut the basement mélange prior to 164 Ma.

At 164–162 Ma, the RCT rifted apart, forming the Josephine ophiolite [57,58,71] and dismembering the RCT cover sequence by mass wasting. Micaceous quartzite sample 8.2.3.22DU with 99% pre-Mesozoic ages is likely an olistostromal block, and the timing of incorporation into the RCT is unknown, but sample RC23023B from a sedimentary mélange indicates sedimentary mixing after 163 Ma. The Galice Formation was also deposited in this rift basin (Figure 10E) and as compression resumed during the Nevadan orogeny (Figure 10F). The RCT cover sequence may have been further deformed and dismembered during basin inversion during the Nevadan orogeny. Both generations of RCT cover sequence are now exposed as dismembered blocks in the southern RCT.

7. Conclusions

Blocks in the RCT basement mélange record continental input at ~201 Ma (33%) and effectively none at ~191 Ma. These data are inconsistent with models including the RCT as part of a large, continuous, west-dipping subduction zone or the western margin of a subducting Mezcalera plate. Instead, our preferred interpretation is that the RCT developed marginal to North America above an east-dipping subduction zone. In the southern RCT, we interpret a ~201 Ma volcaniclastic block with metamorphic clasts and pre-Mesozoic grains as part of Early Jurassic RCT cover sequence and a larger extent of dismembered clastic and metaclastic rocks equivalent to the Middle-Late Jurassic RCT cover sequence. Blocks <193 Ma were likely dismembered and incorporated into the basement mélange by tectonic and sedimentary processes during rifting of the 164–162 Ma Josephine ophiolite and later Nevadan compression. More work is needed to understand the multiple phases of deposition recorded in the RCT cover sequence and how the RCT may correlate along strike beyond the Klamath Mountains.