Abstract

Differences in U-Pb metamorphic monazite ages in the northwestern Thor-Odin culmination of the Monashee complex, southern Canadian Cordillera, are explained in the context of the NNW-trending subvertical transcurrent Paleogene Victor Creek fault. Similar faults are present throughout the Canadian Cordillera. We demonstrate their potential importance in the interpretation of the history of Cordilleran deformation and metamorphism.

A pervasive transposition foliation (ST) is present throughout the Thor-Odin culmination as a result of Cordilleran and possibly earlier deformation. A pre-ST (or early ST) foliation is preserved as aligned inclusion trails in porphyroblasts such as garnet and kyanite. Monazite U-(Th-)Pb isotope dilution–thermal ionization and secondary ion microprobe mass spectrometry (ID-TIMS, SIMS) ages are used to relate monazite growth to pre- and syn-ST fabrics and associated metamorphism. The ages of both pre- and syn-ST fabrics, and the gap between pre- and syn-ST ages decrease toward the east, in an apparently continuous manner. While monazite west of the Victor Creek fault is latest Cretaceous to earliest Eocene in age, monazite east of the Victor Creek fault is exclusively Eocene.

Correlation of rock types across the faults is difficult because the same rock units are repeated many times on either side. However, distinctly different retained ages of metamorphism, and previously recognized differences in structures and detrital zircon signatures across the fault indicate 5–60 km offset along the fault. Similar trends occur across other faults along the western Monashee complex, and faults here and elsewhere in the Canadian Cordillera may have similar geological significance.

INTRODUCTION

The variation in ages of deformation and metamorphism in and around gneiss domes can be complex and has led to multiple, commonly contested interpretations. For example, in the Monashee complex in southeastern British Columbia, there is a progression from Late Cretaceous ages of deformation and metamorphism at high structural levels to Eocene ages at low structural levels (Carr, 1992; Parrish, 1995; Crowley and Parrish, 1999; Gibson et al., 1999). Parrish (1995) explained the downward younging of ages in terms of downward heat transfer from a hot overthrusted Selkirk allochthon on top of the Monashee complex. Gibson et al. (1999) and Crowley et al. (2001) further developed the model and included northeasterly directed telescoping of rocks by folding and thrusting. Kuiper et al. (2006), in contrast, explained the downward younging of ages by upward movement of rocks out of a midcrustal (channel) flow zone, where monazite and zircon grew and record ages at progressively later times. Despite the broad variety in interpretations as outlined here, none of them includes the effects of Paleogene faults that transect the Monashee complex (e.g., Johnson and Brown, 1996; Kruse and Williams, 2005) and the Canadian Cordillera in general (e.g., Gabrielse et al., 2006; Nelson et al., 2013). In this contribution, we consider the effects of one such fault, the Paleogene Victor Creek fault of the western Monashee complex, to explain at least part of the complex distribution of ages of deformation and metamorphism. This fault is known to at least locally display discontinuity in geology on either side of the fault (Kruse and Williams, 2005) and in detrital zircon signatures (Kuiper et al., 2014), and some offset has been inferred from this (see following). In this contribution, we investigate the differences in metamorphic history across the fault. We conducted U-(Th-)Pb monazite analysis on four samples, two east and two west of the Victor Creek fault, in order to test whether (1) their ages would show a progressive younging from higher to lower structural level, as is recognized regionally, and (2) whether ages change significantly across the Victor Creek fault. We discuss our results in the context of new and existing microstructural and petrographic constraints.

GEOLOGICAL BACKGROUND

The Monashee complex of the southern Omineca belt in southeastern British Columbia is a dome-shaped exposure of high-grade ancestral North American rocks (Fig. 1). It exposes some of the structurally deepest rocks of the Shuswap complex, which is a metamorphic complex bounded by the Eocene Okanagan Valley–Eagle River and Columbia River normal fault systems (Read and Brown, 1981; Lane, 1984; Parrish et al., 1988; Johnson and Brown, 1996; Johnson, 2006; cf. Fig. 1). In the Monashee complex, a Paleoproterozoic to Paleozoic cover of quartzite, marble, calc-silicate gneiss, pelitic schist, and paragneiss (Scammell and Brown, 1990; Parrish, 1995, and references therein; Crowley, 1997: Kuiper et al., 2014) overlies a Paleoproterozoic basement consisting of orthogneiss and paragneiss (Armstrong et al., 1991, and references therein; Parkinson, 1991; Crowley, 1999). The Monashee complex has two structural culminations: Frenchman Cap in the north and Thor-Odin in the south (Fig. 1). Rocks of the Selkirk domain (cf. Kuiper et al., 2014) west of the Monashee complex structurally overlie the Monashee complex. They consist of Neoproterozoic and younger quartzite, marble, paragneiss, and orthogneiss (Parrish, 1995, and references therein).

The earliest deformation and metamorphism occurred in the Paleoproterozoic and has been recognized in some of the deepest basement rocks of the Frenchman Cap culmination (Armstrong et al., 1991; Crowley, 1999; Crowley et al., 2008). Later penetrative Cordilleran deformation and metamorphism affected rocks of the Monashee complex and the Selkirk domain. A penetrative transposition fabric comprises the oldest structure resulting from Cordilleran deformation (Johnston et al., 2000; Williams and Jiang, 2005; Kuiper et al., 2006; Spalla et al., 2011). It formed while the Monashee complex was part of a regional NE-directed subhorizontal shear zone, possibly part of a channel-flow zone (Williams and Jiang, 2005; Kuiper et al., 2006). The transposition fabric consists of at least two early generations of NE-verging, SW-dipping tight to isoclinal recumbent folds, overprinted by a generation of more open, predominantly NE-verging folds. In adjacent rocks of the Selkirk domain, these latest folds verge SW. Folds and lithological units are dismembered by transposition. Spalla et al. (2011) referred to the transposition deformation as DT and the transposition foliation as ST. Transposition fabrics are overprinted by gentle upright folds (DT+1 of Spalla et al., 2011). The latest deformational event (DT+2 of Spalla et al., 2011) is characterized by shear bands in and around the Monashee complex, including a zone of concentrated extension, the westerly dipping Greenbush shear zone, along the NW flank of the Thor-Odin culmination (Fig. 2; Johnston et al., 2000), and by later brittle normal and/or dextral faults (Kruse and Williams, 2005). Undeformed (sub)vertical Eocene pegmatite (e.g., Johnston et al., 2000) and lamprophyre dikes (e.g., Lane, 1984; Adams et al., 2005) intruded during extension.

Metamorphism, related to transposition, reached upper-amphibolite- to lower-granulite-facies conditions in the Monashee basement and cover, and lower- to upper-amphibolite facies at higher structural levels (Journeay, 1986; Johnston et al., 2000; Spalla et al., 2011; Zanoni et al., 2014). Monazite, zircon, titanite, and xenotime ages vary between ca. 78 Ma at the highest structural levels (Gibson et al., 1999) and 49 Ma at the deepest structural levels (Crowley and Parrish, 1999) in the Frenchman Cap culmination, and between 58 Ma at the highest structural levels and 51 Ma at low structural levels in the southern Thor-Odin culmination (Carr, 1992; Vanderhaeghe et al., 1999; Hinchey et al., 2006). They have been interpreted as indicating a younging of structures and thermal peak of metamorphism toward deeper structural levels in the Monashee complex. U-Pb data from the northwestern Thor-Odin culmination may fit the same pattern (cf. Johnston et al., 2000; Kuiper, 2003) and were further investigated here.

The Monashee cover sequence of the northwestern Thor-Odin culmination was metamorphosed under upper-amphibolite-facies conditions, with transposition occurring under estimated temperatures of ∼635–760 °C and highly variable estimated average depths between ∼15 and ∼30 km (Spalla et al., 2011; all estimates from samples southeast of Greenbush Lake; Fig. 2). No significant differences in pressure and temperature estimates were recognized across the Victor Creek fault (Spalla et al., 2011). The age of transposition is around ca. 54.5 ± 1.0 Ma east of the Victor Creek fault and younger than 56.5 ± 1.5 Ma west of it (Johnston et al., 2000; Spalla et al., 2011). At Joss Mountain, in the Selkirk domain immediately west of the northwestern Thor-Odin culmination (Fig. 2), transposition occurred at ca. 93 Ma (Johnston et al., 2000) under average temperatures of 640–700 °C and average depths of 21–29 km (Zanoni et al., 2014). The pre-ST to syn-ST path shows variable results from prograde to retrograde in the northwestern Thor-Odin culmination (Spalla et al., 2011) and was probably prograde at Joss Mountain (Zanoni et al., 2014). In both areas, transposition was followed by near-isothermal decompression and exhumation. Transposition at Joss Mountain ended earlier and exhumation started earlier than in the Monashee complex (Zanoni et al., 2014).

Victor Creek Fault

The Victor Creek fault is a steeply dipping, NNW-trending fault. It is marked by one of a set of northerly trending, slightly sinuous lineaments (set 1 of Kruse and Williams, 2005). The Victor Creek fault lineament extends from the central, western Frenchman Cap culmination to south of the Thor-Odin culmination. Outcrop is generally excellent between the lineaments but poor along them, although there are occasional breccias along the Victor Creek fault and other faults. Small mesoscopic faults, oriented the same as the Victor Creek fault and of the same relative age, have piercing points that indicate that they are transcurrent (as is also suggested by their steep attitude) and dextral (set 1 of Kruse and Williams, 2005). A set of westerly trending lineaments (set 3 of Kruse and Williams, 2005) can be shown, in the same way, to represent transcurrent faults, but a lack of piercing points makes it impossible to determine the sense of movement. However, in view of their orientation relative to the Victor Creek fault, they are interpreted as conjugate to the latter and therefore as sinistral. One other set of mesoscopic faults occurs (set 2 of Kruse and Williams, 2005). They dip easterly and occasionally westerly, are normal, and are not recognized as lineaments. The three sets of faults are the only post-ST mesoscopic faults recognized, and set 1 is the most common (Kruse and Williams, 2005). It forms prominent lineaments at the regional scale.

Determination of displacement on macroscopic faults is not generally possible in the Monashee complex. ST is a penetrative feature of the entire complex, and the original rock units are highly fragmented and repeated at all scales, so that it is generally impossible to use lithostratigraphy in order to determine displacement. However, a component of the movement on the Victor Creek fault north of Greenbush Lake can be determined by comparing structure across the fault (Kruse and Williams, 2005). On the west side, rocks including the cover sequence are exposed and are planar. Rocks including the same units occur on the east side, where they are folded into penetrative, kilometer-size folds, trending perpendicular to the fault. The minimum movement required to eliminate the visible mismatch is 1370 m (Kruse and Williams, 2005), but that assumes dip-slip motion only, and the fault is believed to be transcurrent (see previous). Assuming transcurrent movement, Kruse and Williams (2005) suggested a minimum of 5 km displacement. They also pointed to a group of folds on the west side of the Victor Creek fault lineament in the Frenchman Cap culmination that could match those east of the lineament at Greenbush Lake (Figs. 1 and 2). This correlation would require a 55–60 km displacement, but further detailed work is needed to explore this possibility. These folds in the Frenchman Cap culmination are truncated to the west by the Ratchford Creek–Perry River normal fault (Fig. 1) as defined in Kruse and Williams (2005; not to be confused with the Ratchford Creek thrust fault of Journeay, 1986, 1992), which is part of the same fault system (Kruse and Williams, 2005). The Ratchford Creek–Perry River normal fault may extend farther north than interpreted by Journeay (1986, 1992), along the Ratchford Creek lineament (dashed in Fig. 1), along which no geological data are available (Journeay, 1986, 1992).

The geological significance of the Victor Creek fault has been alluded to earlier by Lorencak et al. (2001), Kuiper et al. (2014), and (Toraman et al., 2014). Kuiper et al. (2014) noted that detrital zircon in Neoproterozoic and younger metasedimentary units west of the Victor Creek fault (samples 30 and 54; same as used in this study) yielded fewer Archean and Paleoproterozoic ages and more Mesoproterozoic ages than those east of the Victor Creek fault (samples 70, 92, and 97; same as used in this study), possibly because they were deposited in different regions and later juxtaposed by the fault. Data by Gilley (1999) and Van Rooyen et al. (2011) show the same trends across the Victor Creek fault, and data by Crowley (1997) show the same trends across the Ratchford Creek–Perry River normal fault (Kuiper et al., 2014). Lorencak et al. (2001) investigated the low-temperature cooling history of part of the Shuswap complex, of which the Monashee complex is a part. On the west side of the Victor Creek fault, zircon fission-track ages were 49.0 ± 4.4 Ma to 44.3 ± 4.0 Ma, and apatite fission-track ages were 48.5 ± 3.4 Ma to 42.8 ± 4.6 Ma, and on the east side, ages were 53.9 ± 5.6 Ma to 37.5 ± 5.0 Ma and 43.7 ± 4.0 Ma to 27.7 ± 3.4 Ma, respectively (Lorencak et al., 2001). The data suggest that the latest movement along the Victor Creek fault occurred sometime after ca. 45 Ma (Lorencak et al., 2001). However, new analyses by Toraman et al. (2014) resulted in additional apatite fission-track ages as young as 19.5 ± 2.6 Ma on the west side and 14.0 ± 4.0 Ma on the east side of the Victor Creek fault, which are within error of each other. Hornblende, muscovite, biotite, and K-feldspar 40Ar/39Ar ages do not show any specific differences or trends between the west and east sides of the Victor Creek fault (Johnson, 1994; Spark, 2001; Vanderhaeghe et al., 2003).

RESULTS

The four samples analyzed were collected from a part of the Monashee cover sequence (Fig. 2) that is interpreted as latest Neoproterozoic or younger, and at least in part Devonian, based on the presence of sparse latest Proterozoic or earliest Cambrian zircon grains and one Devonian (396 Ma) zircon grain in those samples (Kuiper et al., 2014). Samples lie along a transect from lowest to highest structural level, generally from east to west, in the following order: 30, 54, 70, and 92/97. Samples 30, 54, and 70 lie in the general area studied by Spalla et al. (2011; see previous). Samples 30 and 54 lie east of the Victor Creek fault, and samples 70 and 92/97 came from west of the Victor Creek fault.

For all samples, field and thin section descriptions are followed by U-Pb single monazite isotope dilution–thermal ionization mass spectrometry (ID-TIMS) results. For samples 91 and 97, zircon ID-TIMS data for single and multigrain fractions are included because both samples contain predominantly metamorphic zircon (as opposed to predominantly detrital zircon in the other samples; cf. Kuiper et al., 2014). ID-TIMS monazite analyses showed a potential for multiple monazite ages per sample, possibly also within individual grains. The probability of averaging discrete age growth domains through bulk grain dissolution was further investigated using secondary ion mass spectrometry (SIMS; sensitive high-resolution ion microprobe [SHRIMP II], see Appendix A) analysis on two samples, which were selected for their locations on either side of the Victor Creek fault and because of their variety in mineralogy. Subsequently, more detailed automated scanning electron microscopy (QEMSCAN) and backscattered electron (BSE) analysis was carried out on those samples in order to better understand the relationships between monazite and metamorphic assemblages and structural fabrics. Analytical methods for ID-TIMS, SHRIMP II, and QEMSCAN are provided in Appendix A. All ages in the text are reported at the 2σ level. For clarity, SIMS data in concordia diagrams are reported at the 1σ level, which minimizes overlap of error ellipses.

Quartzite East of the Victor Creek Fault (Sample 30)

Sample 30 was taken from a quartzite exposure (Figs. 2 and 3) that can be traced tens of kilometers south of the sample location (cf. Reesor and Moore, 1971; Spark, 2001). At the sample location site, the quartzite is ∼10 m thick, but elsewhere in the Monashee complex, quartzites have thicknesses up to several hundred meters (Read and Klepacki, 1981). The quartzite sample is medium grained and contains quartz, biotite, sillimanite, K-feldspar (minerals given in order of decreasing content), accessory titanite (especially in sillimanite-rich layers), zircon, apatite, monazite, rutile, ilmenite, and sphalerite. Locally, tourmaline and sillimanite knots are present. ST is parallel to the contact between the quartzite and adjacent rocks and dips moderately in a westerly direction (Figs. 2 and 3). A lineation, commonly defined by sillimanite, plunges SW.

U-Pb ID-TIMS data for monazite from the quartzite are shown in Figure 4A and Table 1. The analyzed monazite crystals are subrounded to euhedral, with indentations. This morphology suggests growth or modifications during solid-state metamorphism. Some of the monazite grains are zoned in BSE images, with cores (Fig. 4A, BSE image 1) and some patchy domains (Fig. 4A, BSE image 2). Reverse discordance displayed by fraction M4 is interpreted as a result of excess 206Pb. For that reason, we use 207Pb/235U ages herein. Fraction M3 is discordant, possibly due to presence of an older age component, an issue which may also affect fraction M2. The interpreted monazite age is taken to be the 54.2 ± 0.3 Ma 207Pb/235U date of the reversely discordant fraction M4, which is presumably least affected by mixing of age domains. If such domains were mixed, the data would tend to be pulled onto or below the concordia curve.

Pelitic Schist East of the Victor Creek Fault (Sample 54)

Sample 54 is a fine- to medium-grained pelitic schist (Figs. 2 and 3), which consists of biotite, sillimanite, K-feldspar, quartz, and garnet, with accessory zircon, apatite, monazite, and rutile. Retrograde chlorite locally replaces garnet and biotite, while white mica locally replaces biotite and sillimanite. Garnet poikiloblasts (∼0.5 cm) locally contain inclusion trails of biotite, quartz, and rutile that define a pre-ST foliation (Fig. 5). The schist matrix contains ST and a weak southwesterly plunging sillimanite lineation.

Monazite occurs as inclusions in garnet, locally aligned with the internal foliation, and is also intergrown with fibrous sillimanite and biotite in the foliated matrix that envelopes garnet (Fig. 5). Matrix monazite is commonly aligned with the transposition foliations and/or along cleavage planes in sillimanite and biotite. Sillimanite and biotite are replaced by white mica, especially along cleavage planes (Fig. 5). Sillimanite and white muscovite are embayed into the monazite (Fig. 5). Monazite inclusions in garnet show better-defined patchy zoning in BSE images than monazite in the matrix.

U-Pb ID-TIMS results of four single monazite grains are shown in Figure 4B and Table 1. In the BSE images, some of the monazite grains have cores and/or patchy zoning (Figs. 5 and 6). Two monazite grains analyzed, M1 and M2, are concordant, while grains M3 and M4 are normally discordant, most likely due to the presence of an older inherited component. Based on ID-TIMS alone, the 54.6 ± 0.6 Ma weighted average of 206Pb/238U ages of concordant fractions M1 and M2 (Table 1; Fig. 4) is interpreted as reflecting the dominant time of monazite growth. The 206Pb/238U age is favored here as opposed to the 207Pb/235U age, because the data are not reversely discordant, and the weighted average of 207Pb/235U ages also yields 54.6 ± 0.6 Ma. U-Th-Pb SIMS data are concordant or near-concordant. Spot ages calculated from 206Pb/238U ratios generally overlap and smear along concordia between ca. 55 and 49 Ma. However, 208Pb/232Th dates fall into two populations (Table 2; Fig. 6). We report 208Pb/232Th ages for all SIMS analyses, because they are more precise than 207Pb/235U-based measurements, and they do not have a potential for excess 206Pb (Th was not measured during ID-TIMS analysis, and 208Pb/232Th ages are therefore only reported for SIMS analyses). The weighted averages of 208Pb/232Th ages for the two populations are 55.8 ± 1.0 Ma (mean square of weighted deviates [MSWD] = 1.8; n = 7) and 51.6 ± 1.3 Ma (MSWD = 1.4; n = 5). The ID-TIMS age is concluded to represent a bulk average of the two age populations.

Pelitic Schist West of the Victor Creek Fault (Sample 70)

Sample 70 is a medium- to coarse-grained pelitic schist (Figs. 2 and 3) with coarse-grained (<2 cm) anhedral to subhedral garnet, and with biotite, sillimanite, plagioclase, K-feldspar, quartz, kyanite, corundum, cordierite, and relict staurolite, and accessory zircon, apatite, monazite, and rutile. Blue kyanite is visible in hand specimen, and thin sections reveal the presence of kyanite, locally overgrown by sillimanite. Corundum occurs in locations isolated from quartz. Staurolite is replaced by corundum and sillimanite or zoisite. Garnet contains inclusion trails of biotite, quartz, feldspar, rutile, apatite, and monazite, which preserve an older foliation, and of ilmenite and staurolite (Fig. 7). Some garnet rims are free of inclusions. Retrograde white mica and chlorite are present. A WSW-plunging sillimanite lineation occurs on ST.

Monazite occurs as inclusions in garnet and kyanite, and in matrix sillimanite and biotite (Fig. 7). Within garnet, monazite occurs in two textural locations: as part of the internal foliation formed by inclusion trails, and within inclusion-free rims (Fig. 7). Monazite and ilmenite in garnet embay rutile, all within an inclusion trail (Fig. 7). Monazite embayed by kyanite occurs as an internal inclusion trail in one kyanite grain (Fig. 7), along a cleavage plane that shows replacement by white mica, biotite, and chlorite (Fig. 7). All monazite grains imaged in thin section exhibit patchy zoning.

Five single monazite grains were dated by ID-TIMS (Table 1; Fig. 4C). The grains are rounded and have indentations. BSE images reveal irregular patchy zoning (Figs. 7 and 8). Grains M3, M4, and M5 are normally discordant, possibly as a result of inherited age components. Based on the ID-TIMS data only, the 63.6 ± 2.5 Ma 206Pb/238U date of near-concordant grain M2 is interpreted as reflecting the timing of the dominant growth of monazite in this sample. As in sample 54, the 206Pb/238U age is used here because it is generally more precise than the 207Pb/235U age, and the data are not reversely discordant. Spot analyses using SIMS (Table 2; Fig. 8) yielded 208Pb/232Th ratios that reveal two principal groupings with weighted average 208Pb/232Th ages of 65.5 ± 1.6 Ma (MSWD = 2.5; n = 9) and 55.5 ± 3.7 Ma (MSWD = 3.0; n = 3). We likewise conclude that the ID-TIMS age for this sample (M2) probably represents an average of these discrete age domains.

Paragneiss (Sample 97) and Leucosome in Paragneiss (Sample 92), West of the Victor Creek Fault

Sample 97 is a migmatitic semipelitic gneiss with quartz- and feldspar-bearing leucosomes (sample 92; see following). The samples were taken along the Trans-Canada Highway at Three Valley Lake (Fig. 2). Sample 97 consists of biotite, sillimanite, K-feldspar, quartz, inequant garnet, and retrograde white mica. Garnet has inclusions of biotite, quartz, feldspar, and fine-grained accessory minerals. Accessory minerals are zircon, apatite, monazite, ilmenite, and rutile. ST contains a W-plunging sillimanite lineation. Sample 92 is leucosome from the paragneiss (sample 97). It contains quartz, feldspar, biotite, garnet, and sillimanite, with accessory zircon, apatite, monazite, rutile, and ilmenite. Retrograde or late white mica, chlorite, and epidote locally replace biotite, garnet, and sillimanite. Garnet is coarse grained (<4 mm), euhedral, and inequant.

Zircon crystals from sample 97 (Table 1; Fig. 4D) are colorless, stubby (aspect ratios between 1:1 and 1:2), and subrounded with crystal facets. In BSE images, zircon crystals are zoned, preserving irregularly shaped cores, and overgrowths (Fig. 4D, BSE images 1 and 2). In some crystals, vague concentric zoning is visible (BSE image 1). Zircon crystals from sample 92 occur as colorless bipyramidal prisms with aspect ratios between 1:1.5 and 1:8. In BSE images, zircon crystals have core and rim zoning. Locally, concentric oscillatory zoning is visible in the main domain of the zircon (Fig. 4D, BSE image 3).

U-Pb ID-TIMS data from samples 92 (fractions indicated with small letters) and 97 (fractions indicated with capital letters) are considered together because they are from the same location and do not show significant differences. Zircon data form a discordant array of 206Pb/238U and 207Pb/235U ages between ca. 73 and ca. 100 Ma and 207Pb/206Pb ages between ca. 79 and ca. 230 Ma (Table 1; Fig. 4). Fraction F (three zircons) is considered concordant and has a 206Pb/238U date of 79.5 ± 0.6 Ma. However, considering the large spread in zircon data, the possibility of a mixing age cannot be excluded.

Monazite grains from sample 97 show core-rim zoning patterns in the BSE images (Fig. 4D), and those from sample 92 have patchy zoning (Fig. 4D). In our age interpretation, only clusters of reversely discordant U-Pb ID-TIMS monazite data, possibly supported by near-concordant data, are considered, because they are presumably least affected by potential mixing between younger and older age domains. Reverse discordance is interpreted as a result of excess 206Pb, and therefore 207Pb/235U ages are reported here. Two age populations exist. Three monazite grains (M1, M2, m2) have a weighted average 207Pb/235U age of 71.6 ± 1.2 Ma (MSWD = 1.6), and three others (m1, M5, M6) have a weighted average 207Pb/235U age of 58.6 ± 4.3 Ma (MSWD = 3.7; Table 1; Fig. 4). Fraction m3 is near-concordant at 75.7 ± 0.3 Ma (1.8% discordant; Table 1). It is a single fraction and not part of a population. The lack of excess 206Pb in m3 may indicate that it actually belongs to the slightly younger monazite population formed by M1, M2, and m2, with excess 206Pb, but with some older radiogenic Pb as is present in fractions M4 and m4. We consider the two weighted averages of 207Pb/235U monazite ages of 71.6 ± 1.2 Ma and 58.6 ± 4.3 Ma as most reliable and use those in our interpretation herein.

TEXTURAL RELATIONSHIPS OF MONAZITE

A summary of our monazite ages and ages of metamorphism from the literature is given in Figures 2 and 9. Two age populations are present in all of our pelitic rock samples, which may be related to pre- and syn-ST fabrics as follows. In sample 54 (east of the Victor Creek fault), the first generation of monazite is part of an internal foliation formed by inclusions in garnet. The internal foliation in garnet has been recognized by Spalla et al. (2011) as pre-ST, based on differences in assemblages and compositions of minerals included in garnet compared to ST-forming minerals. However, it is possible that the pre-ST assemblages and fabrics represent an early stage of transposition. We relate our first generation of monazite to their pre-ST assemblage (and name it as such), bearing in mind that “pre-ST” may represent early ST. The second generation of monazite is aligned with the transposition foliation, and it occurs in biotite and sillimanite that form part of the transposition foliation and lineation (Fig. 5). The fact that sillimanite forms embayments into monazite indicates that monazite was resorbed while sillimanite grew. Because monazite and sillimanite are compositionally different, it is most likely that other minerals were growing or breaking down at the same time and/or that fluids were involved. Biotite and white mica embayments may indicate further monazite resorption during their growth, but it is perhaps more likely that biotite and white mica simply replaced sillimanite. Monazite growth has previously been related to reactions involving aluminosilicate growth (Pyle and Spear, 2003; Foster et al., 2004; Gervais and Hynes, 2013) or breakdown (Pyle and Spear, 2003; Wing et al., 2003). The monazite growth reaction that resembles our assemblages best is reaction 5 of Pyle and Spear (2003): quartz + plagioclase + muscovite + monazite + xenotime + apatite = sillimanite + melt (see also reactions in Spalla et al., 2011; Zanoni et al., 2014). While our sample 54 does not show evidence for extensive melt, or early xenotime, it is possible that a similar reaction occurred. The fact that this generation of monazite is aligned with the transposition foliation, and certainly did not grow after sillimanite, biotite, and white mica, indicates that this generation of monazite is syn-ST. It is possible that some of these grains contain pre-ST cores. While we did not date monazite in situ, an intuitive correlation between two generations of monazite in thin section and two monazite ages leads to the interpretation that pre-ST monazite is 55.8 ± 1.0 Ma and syn-ST monazite is 51.6 ± 1.3 Ma. Some monazite growth may have occurred in a more continuous manner between the pre-ST and syn-ST populations, which explains the spread in ages between the two populations in Figure 6.

In sample 70, monazite in garnet cores (in inclusion trails forming an internal foliation) may be the same generation as monazite in relict kyanite, and monazite in generally inclusion-free garnet rims may be the same generation as that in matrix sillimanite and biotite (Fig. 7). Alternatively, up to four generations of monazite may exist. Monazite grains in garnet cores are interpreted as pre-ST for the same reasons as outlined for sample 54. Monazite and ilmenite in garnet can be interpreted as overgrowing rutile (Fig. 7), indicating decompression, consistent with results for the pre-ST to syn-ST path interpreted by Spalla et al. (2011) and the general decompression path recognized by Norlander et al. (2002) in the southern Thor-Odin culmination. Together with the garnet cores, kyanite has been recognized as part of the pre-ST assemblage (Spalla et al., 2011), making it likely that monazite in relict kyanite and monazite in garnet cores are of the same generation. Monazite embayment by kyanite indicates that monazite was resorbed while kyanite grew, similar to monazite resorption and sillimanite growth in sample 54. Relict kyanite was replaced by biotite and white mica, resulting in (apparent?) embayment of monazite by those minerals. We interpret monazite occurring in garnet cores and relict kyanite as having grown pre-ST.

Matrix monazite occurs in sillimanite and biotite, which make up the transposition foliation and lineation. Sillimanite and biotite are replaced by retrograde white mica and chlorite. Inclusion-free garnet rims have been interpreted by Spalla et al. (2011) as part of the syn-ST assemblage. We interpret monazite in garnet cores and in relict kyanite as pre-ST and 65.5 ± 1.6 Ma, our oldest monazite age population, and monazite in sillimanite, biotite, and garnet rims as syn-ST and 55.5 ± 3.7 Ma, our youngest age population. The syn-ST grains may contain pre-ST cores, and some monazite may have grown in a more continuous manner as in sample 54.

Samples 97 and 92 yielded 71.6 ± 1.2 Ma and 58.6 ± 4.3 Ma monazite ages, and discordant zircon data that are close to the concordia curve between ca. 100 Ma and ca. 80 Ma (Fig. 4; Table 1). We interpret the spread of zircon data as mixing ages between the younger ages of metamorphism reflected by the monazite data and older detrital zircon grains (Kuiper et al., 2014), or possibly mid-Cretaceous (or earlier?) metamorphic zircon growth (cf. Kuiper et al., 2006). U-Pb SIMS data from zircon grains and rims from a leucosome in the same paragneiss yielded concordant ages at ca. 74 Ma (Gilley, 1999). Six zircon fractions from an amphibolite boudin in the paragneiss plot along a discordia chord with a ca. 1.5–1.6 Ga upper intercept, interpreted as the protolith age, and a lower intercept near 73.4 ± 1.7 Ma, interpreted as an age of metamorphism (Parkinson, 1992). Multizircon ID-TIMS analyses by Wasteneys (1998, personal commun.) are consistent with our spread of zircon data between ca. 100 Ma and ca. 80 Ma, and the ca. 74–73 Ma zircon ages presented herein. Our 71.6 ± 1.2 Ma monazite age is within error of the ca. 74–73 Ma zircon ages. Samples 97 and 92 have similar mineralogy and fabrics as samples 70 and 54, with the difference that they are migmatitic. They contain garnet with inclusions, and sillimanite and biotite are part of the transposition fabric. While we do not have the same textural context of monazite in general as for samples 54 and 70, it is possible that the 71.6 ± 1.2 Ma age represents pre-ST and the 58.6 ± 4.3 Ma age represents syn-ST monazite growth, similar to the two monazite growth events in samples 54 and 70.

Sample 30 is characterized by a single age population for monazite at 54.2 ± 0.3 Ma. Its textural relationship is unknown. Kyanite in a 54.5 ± 1.0 Ma pegmatite (Johnston et al., 2000) ∼500 m from sample 30 was interpreted by Spalla et al. (2011) as being related to their pre-ST kyanite-bearing assemblage, suggesting that it intruded before (or at the onset of?) transposition. If true, monazite in sample 30 may represent pre-ST deformation and metamorphism.

SIGNIFICANCE OF PALEOGENE FAULTS: DISCUSSION AND CONCLUSION

Our ages are consistent with ages of metamorphism reported by Johnston et al. (2000) and Kuiper (2003), and all are summarized in Figures 2 and 9. Given our ages and the described textural relationships of monazite, three general trends can be recognized. First, generally, both our pre- and syn-ST ages can be interpreted as younging toward the east. These trends are consistent with previous interpretations of U-Pb ages in the Monashee complex in terms of younging of structures and the thermal peak of metamorphism toward deeper structural levels (cf. Carr, 1992; Parrish, 1995, and references therein; Crowley and Parrish, 1999; Gibson et al., 1999). However, the recognition of pre- and syn-ST monazite generations and an age gap between the two is new. A second trend is that the age gap between pre- and syn-ST monazite is progressively smaller toward the east. This age gap is clearly recognizable west of the Victor Creek fault, but it is almost unrecognizable to invisible east of the Victor Creek fault. The widening age gap toward the west (higher structural levels) suggests that there is more time between formation of the pre- and syn-ST assemblages toward the west. The significance of this age gap is unclear, as there is some variation in pressure and temperature estimates for pre- and syn-ST assemblages from various samples from the northwestern Thor-Odin culmination (Spalla et al., 2011), and because it is unclear whether pre-ST assemblages reflect a pretransposition deformation event, or simply an earlier part of the transposition. In any case, it appears that metamorphic conditions changed more slowly in the west than in the east before and during transposition. The third observed trend is that samples east of the Victor Creek fault yielded Eocene ages only, while samples west of the Victor Creek fault yielded latest Cretaceous (Maastrichtian) to Eocene ages. Thus, despite the apparent similarities in rock types, structures, and metamorphic assemblages across the Victor Creek fault, differences exist in ages of metamorphism.

The downward younging of ages of metamorphism has been explained through various models (see previous discussion herein), including downward heat transfer (Parrish, 1995), northeasterly directed telescoping of rocks with progressively younger ages of metamorphism to the east and now at deeper structural levels (Gibson et al., 1999; Crowley et al., 2001), and channel flow (Kuiper et al., 2006), but in these models, abrupt changes across late brittle faults are not discussed. Zanoni et al. (2014) explained their observation that transposition at Joss Mountain ended earlier and exhumation started earlier than in the Monashee complex by late- to postexhumation normal movement along the westerly dipping Greenbush shear zone (Johnston et al., 2000; Fig. 2). While that interpretation is valid for their observations, and could explain part of our observations in this contribution, it does not explain differences between sample 70 and samples 54 and 30, as they are all east of the Greenbush shear zone. In addition to the variation in ages of metamorphism across the Victor Creek fault, structures (Kruse and Williams, 2005) and detrital zircon populations (Kuiper et al., 2014), and to some extent apatite fission-track ages (Lorencak et al., 2001; Toraman et al., 2014), differ across the fault (see earlier herein).

The Victor Creek fault has previously been interpreted as a normal fault (Johnson and Brown, 1996; Lorencak et al., 2001). Known normal faults within the Thor-Odin culmination have a similar strike to the Victor Creek fault but have displacements measurable in meters rather than kilometers (cf. Kruse and Williams, 2005). The only known normal faults that may have kilometers of displacement are bounding faults such as the Okanagan–Eagle River and Columbia River faults (Read and Brown, 1981; Lane, 1984; Parrish et al., 1988; Johnson and Brown, 1996; Johnson, 2006) and they are not vertical like the Victor Creek fault. Our interpretation of a significant dextral component along the Victor Creek fault is based on the self-consistency argument (it has the same orientation, same breccia or chlorite fill, and same relative age as small-scale faults that are fully exposed and have piercing points) and is further consistent with Cretaceous–Eocene dextral motion along faults that exist throughout the Canadian Cordillera and display significant offsets up to ∼800 km (Gabrielse et al., 2006; Nelson et al., 2013). Significant movement on a transcurrent Victor Creek fault, with or without a dip-slip component (Johnson and Brown, 1996; Lorencak et al., 2001; Kruse and Williams, 2005), implies that rocks west of the Victor Creek fault formed and deformed farther south, and possibly at a different structural level than those east of the Victor Creek fault (cf. Johnson and Brown, 1996; Kruse and Williams, 2005; Kuiper et al., 2014). This could explain their older ages of deformation and metamorphism, as well as differences in structures and detrital zircon populations.

Similar to the Victor Creek fault, other faults may be of importance in the reconstruction of the geological history of the Monashee complex. Syn-ST metamorphic monazite at Joss Mountain (Fig. 2) has been constrained at 93.0 ± 1.5 Ma (Johnston et al., 2000; Zanoni et al., 2014), i.e., significantly older than any of our ages. While there are various interpretations possible to explain the difference, it should be noted that the Three Valley–Joss Pass fault, another Paleogene brittle fault, separates the Joss Mountain area from our rocks of the northwestern Thor-Odin culmination. The fault may explain part of the difference in ages. A similar situation may exist in the Frenchman Cap culmination, where ages of metamorphism appear to get progressively younger from higher to lower structural levels. However, rocks east of the Ratchford Creek–Perry River normal fault (Fig. 1; Kruse and Williams, 2005) are ca. 58–49 Ma or latest Paleocene to Eocene in age, while older metamorphic ages up to ca. 78 Ma only exist west of the fault (Crowley and Parrish, 1999; Gibson et al., 1999; Crowley et al., 2001; Foster et al., 2002). While speculative, the Ratchford Creek–Perry River normal fault may in part be responsible for the difference in ages.

In summary, late brittle faults such as the Victor Creek fault affected the Monashee complex late in its tectonic history and are relevant in our geological interpretations. Cretaceous–Eocene dextral and normal faults exist throughout the Canadian Cordillera. While some of these faults are well recognized, others may appear as subtle as the Victor Creek fault. All of these faults are potentially significant in the reconstruction of the geological history.

We thank Sharon Carr for patiently teaching Kuiper the meticulous art of isotope dilution–thermal ionization mass spectrometry analysis. Katharina Pfaff and Jae Erickson assisted with automated scanning electron microscopy (QEMSCAN) imaging. Nigel Kelly is thanked for providing input on the QEMSCAN images and for initiating and assisting with associated backscattered-electron imaging. We thank Davide Zanoni and Nigel Kelly for commenting on earlier drafts of this manuscript. Dan Gibson and two anonymous reviewers provided constructive comments, which improved this manuscript significantly.

APPENDIX A—ANALYTICAL METHODS

U-Pb ID-TIMS geochronology was carried at Carleton University, Ottawa, Canada. Heavy minerals were recovered from ∼10–25 kg rock samples by standard crushing, grinding, Rogers Gold™ table methods, and heavy liquid separation. Heavy mineral concentrates containing the least magnetic zircon and monazite were isolated using conventional Frantz™ magnetic separation techniques. Minerals with the least amount of inclusions, alteration, and fractures, the best clarity, and lacking visible cores and overgrowths were selected for dating by handpicking under an optical microscope. All zircon fractions were air abraded with pyrite (Krogh, 1982) and washed in warm HNO3 to remove abrasion residues. Minerals selected for dating were spiked with a mixed 205Pb-233U-235U tracer (Parrish and Krogh, 1987) and dissolved in HF (zircon) or HCl (monazite) in Teflon™ microcapsules (Parrish, 1987). U and Pb were separated using conventional anion exchange column chemistry procedures (e.g., Parrish et al., 1987; Krogh, 1973; Roddick et al., 1987), and their isotopic compositions were measured on a Finnigan-MAT 261 variable multicollector mass spectrometer (Roddick et al., 1987). Procedural blanks were generally less than 5 pg for U and 6–60 pg for Pb. Common-Pb corrections were made using Pb compositions derived from Stacey and Kramers (1975), while decay constants used were those recommended by Steiger and Jäger (1977). Errors on isotope ratios were estimated using numerical error propagation (Roddick, 1987). All errors on ages are reported at the 2σ level. Discordia chords through data were calculated using a modified York (1969) regression (Parrish et al., 1987). Data were plotted using Isoplot/Ex v. 2.49 (Ludwig, 2001).

Monazite from samples 54 and 70 was analyzed on the SHRIMP II ion microprobe at the Geological Survey of Canada (GSC) in Ottawa, following methods described by Stern (1997) and Stern and Sanborn (1998). Representative mineral grains that had been separated for the ID-TIMS study were mounted in epoxy and polished to reveal their internal structures. Monazite BSE images were acquired using a Cambridge Instruments S360 scanning electron microscope (SEM) at the GSC in Ottawa.

For the ion microprobe (SHRIMP II) work, the imaged grain mount was repolished and coated with conductive high-purity Au (99.9999%). Spots were selected for analysis based upon the dimensions of apparent age domains evident through X-ray element mapping and BSE imaging. Spot sizes were ∼10 × 13 μm (sample 54 spots 1.2, 1.3, 3.3, 6.1 and sample 70 spots 2.2, 7.1) and 17 × 21 μm (all other spots). The Pb/U and Pb/Th ratios were calibrated using GSC monazite standards 2908, 3345, and 4170 (Stern, 1997). The 206Pb/238U dates were corrected for common Pb using the 204Pb method (Stern, 1997).

Automated scanning electron microscopy (QEMSCAN) was carried out at the Colorado School of Mines. Following carbon coating, the samples were loaded into the QEMSCAN instrument, and the analyses were initiated using the control program (iDiscover, FEI). Four energy-dispersive X-ray (EDX) spectrometers acquired spectra from each particle with a beam stepping interval (i.e., spacing between acquisition points) of 20 µm (overview scans) or 4 or 2 µm (for high-resolution scans), an accelerating voltage of 25 keV, and a beam current of 5 nA. Interactions between the beam and the sample were modeled through Monte Carlo simulation. The EDX spectra were compared with spectra held in a look-up table allowing an assignment to be made of a composition at each acquisition point. The assignment makes no distinction between mineral species and amorphous grains of similar composition. This procedure allows a compositional map of the particle to be generated. Results were output by the QEMSCAN software as a spreadsheet giving the area percent (area %) of each composition in the look-up table.