Retroarc Jurassic burial and exhumation of Barrovian metamorphic rocks dated by monazite petrochronology, Funeral Mountains, California

Many studies of metamorphic rocks are focused on determining pressure-temperature-time (P-T-t) paths, as these have implications for the tectonic history. Determination of P-T paths requires establishing the ways in which conditions changed during the metamorphic history. Evidence that may be interpreted to record changing conditions includes the following: growth-related chemical zoning in minerals (e.g., Spear and Selverstone, 1983; Gaidies et al., 2008), growth discontinuities on element maps of individual crystals that document partial resorption followed by regrowth (e.g., Kelly et al., 2015; Craddock Affinati et al., 2020), corona textures (e.g., Baldwin et al., 2007; Tateishi et al., 2004), symplectites (e.g., Belyanin et al., 2012; Harley, 1998; Peterman and Grove, 2010), and resorbed (corroded) rims (e.g., Page et al., 2007). Thermodynamic approaches may be used to calculate P-T paths from chemical growth zoning preserved in garnet (e.g., Moynihan and Pattison et al., 2013; Spear and Selverstone, 1983; Gaidies et al., 2008), whereas other portions of a path may be inferred from the locations in P-T space of reactions that are interpreted to have been responsible for mineral growth or resorption (e.g., Hallett and Spear, 2013; Cruz-Uribe et al., 2015; Kelly et al., 2015; Craddock Affinati et al., 2020).

Dating of P-T paths may be done by directly dating minerals that grew along the path such as garnet (e.g., Wells et al., 2012; Cruz-Uribe et al., 2015; Kelly et al., 2015; Craddock Affinati et al., 2020), or by dating inclusions that constrain the age of the reaction that grew the host mineral. Another approach utilizes the common accessory mineral monazite, which is known to grow or recrystallize throughout medium and high grades of metamorphism in pelitic schist (e.g., Janots et al., 2008; Pyle and Spear, 2003; Spear and Pyle, 2002; Hallett and Spear, 2015). This approach takes advantage of the ability of monazite to serve as a monitor of changes in matrix reservoir abundances of rare earth elements (REEs), and the fact that small grains or domains within grains (<8 µm) can be dated and analyzed for trace elements simultaneously (Kylander-Clark et al., 2013). Because garnet tends to concentrate heavy rare earth elements (HREEs), the availability of these elements in the matrix reservoir increases or decreases depending on whether garnet is growing or being consumed. Consequently, age-correlative changes in monazite HREE abundances in garnet-bearing metamorphic rocks have been interpreted to date or constrain the age of garnet-forming or garnet-consuming reactions (e.g., Stevens et al., 2015; Soucy La Roche et al., 2016).

In this chapter, we present new age and trace-element data for monazite and xenotime that constrain the timing of different segments of the P-T-t path reported in the study of Craddock Affinati et al. (2020) for the Chloride Cliff area in the Funeral Mountains, located in southeastern California in Death Valley National Park (Figs. 1A and 1B). The P-T path was interpreted in Craddock Affinati et al. (2020) to be evidence for southeast-directed Jurassic thrusting in the retroarc during early Sevier orogenesis. The new timing constraints on the P-T-t path add to the current understanding of regional tectonics in the retroarc region of the Jurassic plate margin.

The configuration of plates and arcs along the southwestern North American plate margin during the Triassic and Jurassic is uncertain and debated (e.g., Schweickert and Cowan, 1975; Dickinson and Lawton, 2001; Wakabayashi et al., 2010; Ernst, 2011; Yonkee and Weil, 2015; Schweickert, 2015; Saleeby and Dunne, 2015). However, it is generally agreed that the Permian to Triassic margin likely consisted of several separate arc-trench systems that produced magmatism (e.g., Coney and Evenchick, 1994; DeCelles, 2004; Paterson and Ducea, 2015; Yonkee and Weil, 2015; Cecil et al., 2018). The development of a single east-dipping subduction zone (Franciscan subduction) occurred around 165 Ma (Wakabayashi et al., 2010), or from 175 to 165 Ma (Yonkee and Weil, 2015), or by 180 Ma (Mulcahy et al., 2018).

The age of Franciscan subduction has been inferred from the dating of high-grade metamorphic rocks in the Franciscan Complex, which have been interpreted to have undergone subduction metamorphism, and the dating of igneous rocks in the Coast Range ophiolite. The Franciscan Complex has yielded U-Pb zircon ages of 176 to 120 Ma from inclusions in garnet (Mulcahy et al., 2018), Lu-Hf and Sm-Nd ages on garnet and lawsonite (Anczkiewicz et al., 2004; Mulcahy et al., 2014), and ^40Ar/³⁹Ar crystallization ages on muscovite and hornblende (Wakabayashi and Dumitru, 2007; Mulcahy et al., 2014). The oldest ages of Franciscan Complex metamorphism suggest that east-dipping subduction was ongoing by 176 Ma (Mulcahy et al., 2018). The formation of the Coast Range ophiolite is dated most reliably by U-Pb zircon in plagiogranites at 168–161 Ma (Mattinson and Hopson, 2008; Hopson et al., 1981). Other U-Pb zircon ages from the Coast Range ophiolite have been reported, ranging from 172 to 156 Ma (Shervais et al., 2005, and references therein). The overlap in ages from the Coast Range ophiolite and the majority of ages from the Franciscan Complex suggests that the Coast Range ophiolite formed above the east-dipping plate in the forearc (Shervais et al., 2005). The Coast Range ophiolite could have also formed in a pull-apart basin outboard from the margin and then been subsequently accreted onto the continent (e.g., Schweickert, 2015).

Terrane accretion occurred along the western side of the Jurassic magmatic arc, specifically within the western Sierra Nevada metamorphic belt (Fig. 1C). Metamorphism within the belt is dated from 174 to 165 Ma and likely dates the amalgamation and accretion of arcs onto the continent (Hacker, 1993). However, sediments overlying the arc and the metamorphic terrane indicate that accretion to the plate margin was complete by Callovian time (165–161 Ma; Edelman and Sharp, 1989). A more recent study concluded that the earliest deformation in the western Sierra Nevada metamorphic belt and its accretion occurred at 177–170 Ma (S₁ dated at 176 Ma in the Calaveras Complex; Sharp, 1988), but that accretion of additional terranes continued through 143 Ma (Schweickert, 2015).

Several retroarc contractional belts were active during or shortly after Franciscan subduction initiation (Fig. 1C). These include (from west to east) the East Sierran thrust system (188–140 Ma; Dunne and Walker, 2004), the southern Sevier fold-and-thrust belt (Middle Jurassic to mid-Cretaceous; DeCelles, 2004; Giallorenzo et al., 2018; Craddock Affinati et al., 2020), the Luning-Fencemaker fold-and-thrust belt (Early Jurassic–Late Jurassic; Wyld et al., 2003), and the central Nevada thrust belt (Taylor et al., 2000). The Death Valley region, including the Funeral Mountains, was located on the east side of the northern Sierra Nevada range in the Jurassic, taking into account movement along the Cretaceous dextral strike-slip Mojave–Snow Late fault (Fig. 1C; Schweickert and Lahren, 1990; Wyld and Wright, 2001; Dickinson, 2006).

In the Funeral Mountains, a classic Barrovian metamorphic terrane is exposed in the footwall of a major low-angle normal fault, the Boundary Canyon detachment (Labotka, 1980; Hoisch and Simpson, 1993). Rocks in the footwall comprise a Proterozoic sedimentary sequence, which includes, from oldest to youngest, the Crystal Spring Formation, Horse Thief Springs Formation, Beck Spring Dolomite, Kingston Peak Formation, Noonday Dolomite, Johnnie Formation, and Stirling Quartzite (Fig. 1A). Within the footwall, the metamorphic grade increases from upper greenschist facies in the central part of the range (Indian Pass) to upper amphibolite facies in the northwest part of the range (Monarch Canyon) over a distance of ~20 km (Fig. 1A; Labotka, 1980; Hoisch and Simpson, 1993). The Chloride Cliff area, where samples for the present study were collected, is located about halfway between these two areas and is metamorphosed to middle amphibolite facies (Fig. 1A). The age of the metamorphism was determined by Lu-Hf garnet dating in the Indian Pass and Chloride Cliff areas. In the Indian Pass area, an age of 158.2 ± 2.6 Ma was obtained from a Johnnie Formation metapelite (Hoisch et al., 2014), and from the Chloride Cliff area, ages of 167.3 ± 0.72 and 165.1 ± 9.2 were obtained from a Johnnie Formation metapelite and a Horse Thief Springs Formation garnet amphibolite, respectively (Craddock Affinati et al., 2020). The ages were interpreted to record garnet growth during burial that resulted from southeast-directed thrusting along the Funeral thrust (Hoisch et al., 2014; Craddock Affinati et al., 2020). Evidence consistent with this interpretation is the preservation of top-to-the-southeast shear fabrics in the Chloride Cliff area, including shear bands and mantled porphyroclasts in schist and asymmetric strained clasts in the Kingston Peak Formation diamictite (Craig, 2019; Craig et al., 2018). Metamorphism of younger age is suggested for the highest-grade part of the range in Monarch Canyon, where a pelitic schist yielded a Th-Pb monazite age of 91.5 ± 1.4 Ma (Mattinson et al., 2007).

Craddock Affinati et al. (2020) presented a P-T path from the Chloride Cliff area. The P-T path was determined by thermodynamically simulating chemical growth zoning in garnet. The simulations, which were conducted on 12 garnet crystals from six samples of metapelites from the Proterozoic Kingston Peak and Johnnie Formations, yielded a well-defined path consisting of a nearly isothermal pressure increase from 4.2 kbar and 550 °C to 6.5 kbar and 575 °C, followed by a prograde decrease in pressure to 5.1 kbar and 575 °C. The path was produced by compositing paths generated on individual garnet crystals, which were similar and strongly overlapped. A Lu-Hf garnet age of 167.3 ± 0.72 Ma dated the early pressure-increase portion of the path (Craddock Affinati et al., 2020). A continuation of the exhumation part of the path was inferred from the presence of a discontinuity in chemical zoning near the rims of garnets from the Kingston Peak Formation. The discontinuity was inferred to have formed during exhumation from the prograde crossing of a staurolite-producing, garnet-consuming reaction (reaction 1 of Craddock Affinati et al., 2020) followed by the retrograde recrossing of the same reaction, causing regrowth.

Exhumation of the Funeral Mountains occurred over a protracted period of time beginning in the Late Jurassic. The earliest exhumation is documented at Indian Pass with cooling through muscovite Ar closure (~425 °C; Harrison et al., 2009) at 152.6 ± 1.4 Ma and 146 ± 1.1 Ma following prograde garnet growth at 158.2 ± 2.6 Ma at ~525 °C and ~5 kbar (Hoisch et al., 2014). Craddock Affinati et al. (2020) interpreted that exhumation of this timing also affected the Chloride Cliff area, corresponding to the portion of the P-T path where it transitions from burial to exhumation during garnet growth. Exhumation is also documented at Chloride Cliff by cooling through hornblende Ar closure (~500 °C; McDougall and Harrison, 1988) at 112.9 ± 0.8 Ma (Hoisch and Simpson, 1993) and muscovite Ar closure from 86.0 ± 0.6 Ma to 67.4 ± 0.5 Ma (Sauer, 2012; Wells et al., 2019). Final exhumation of the Funeral Mountains is recorded by fission-track ages on apatite, zircon, and titanite from Monarch Canyon (Holm and Dokka, 1991) and on apatite from Chloride Cliff (Hoisch and Simpson, 1993), which indicate that Monarch Canyon cooled through ~285 °C by ca. 11 Ma and both Monarch Canyon and Chloride Cliff cooled through ~120 °C by ca. 6 Ma (Holm and Dokka, 1991; Hoisch and Simpson, 1993). This timing is confirmed by (U-Th)/He cooling ages obtained from zircon collected from the central through northern Funeral Mountains (Beyene et al., 2010).

Top-to-the-northwest folds and shear fabrics (shear bands, mica fish, asymmetric porphyroclasts, and grain shape fabrics) documented in Monarch Canyon and in a canyon incised into the Keane Wonder fault have been interpreted to be related to exhumation (Hoisch and Simpson, 1993; Applegate et al., 1992). In Monarch Canyon, this fabric was dated as Late Cretaceous (72–70 Ma) by dating crosscutting pre- and postkinematic granitic dikes (Applegate et al., 1992). Top-to-the-northwest displacement along the Boundary Canyon detachment exhumed the footwall in the late Miocene, resulting in rapid cooling (Hoisch and Simpson, 1993; Holm and Dokka, 1991). The Late Cretaceous shear fabrics may represent strain related to an earlier phase of movement along the Boundary Canyon detachment or a related precursor structure. The Funeral thrust was proposed to have been reactivated to become the presently exposed Boundary Canyon detachment (Hoisch et al., 2014; Craddock Affinati et al., 2020).

Monazite and xenotime are common accessory minerals in pelitic schist (Spear and Pyle, 2002). Both minerals readily accept U and Th into their structures during crystallization but tend to reject Pb (e.g., Franz et al., 1996). Consequently, most Pb found in these minerals is produced by the radiogenic decay of U and Th, resulting in the potential for these minerals to yield precise ages (e.g., Kylander-Clark et al., 2013). Slow rates of Pb diffusion leading to high closure temperatures have been determined for both monazite and xenotime (>900 °C; Cherniak et al., 2004; Cherniak, 2006). Thus, ages determined for these minerals will represent crystallization ages in most metamorphic rocks, including the rocks in the present study.

Monazite and xenotime may occur as detrital grains in low-grade metapelites metamorphosed at temperatures up to 420–450 °C (e.g., Spear and Pyle, 2002; Janots et al., 2008). At low grades, xenotime can also crystallize as a result of monazite breakdown at the biotite and/or chloritoid isograd (see Janots et al., 2008, their eq. 4). However, metamorphic monazite has been identified in greenschist-facies rocks (350–550 °C; Spear and Pyle, 2002). Above 420–450 °C, monazite and xenotime detrital grains, if present, will recrystallize, and/or new grains will form from metamorphic reactions (Janots et al., 2008).

Ages determined on monazite grains that occur as inclusions in garnet constrain the age for that portion of garnet to be younger than the inclusion. HREE concentrations in monazite can indicate whether garnet was growing before, during, or after the crystallization of monazite, because garnet crystallization tends to deplete the matrix of HREEs as it grows (Spear and Pyle, 2002). For example, if monazite becomes more depleted in HREEs over time, it could indicate that garnet was growing contemporaneously with monazite. An increase in (Ho/Gd)_N ratios (HREEs over middle rare earth elements [MREEs]) with decreasing age would indicate progressive depletion of HREE compared to light rare earth elements (LREEs) in the matrix reservoir, suggesting that monazite was growing at the same time as garnet and/or xenotime, which, like garnet, also tends to preferentially incorporate HREEs (e.g., Stevens et al., 2015; Engi, 2017). The ratio Yb/Dy (HREEs over MREEs), when compared across multiple age populations, may also be utilized in the same way. Higher Yb/Dy values indicate relative enrichment of HREEs (e.g., Briggs and Cottle, 2018). In addition, abundances of REEs and Y within garnets may indicate whether xenotime and monazite were breaking down, crystallizing, or stable during garnet growth (Spear and Pyle, 2002). Xenotime coexisting with garnet buffers the Y concentration in garnet to a limiting value that is sensitive to temperature (Pyle and Spear, 2000). Thus, high-Y cores or annuli in garnet may reflect garnet growth during periods when xenotime was stable in the rock. In addition, Y annuli may form when garnet undergoes partial resorption followed by regrowth. Resorption releases Y hosted in the garnet rim into the matrix surrounding the grain. When garnet regrows, Y may be reincorporated into the regrown garnet rim at a higher concentration (Pyle and Spear, 1999).

In this study, we report data collected from five samples from the Chloride Cliff area: SCFM315-12C and SSFM307-8D are samples of high-alumina (Al) metapelites from the Neoproterozoic Johnnie Formation, and SCFM315-17A, SSFM307-7G, and SSFM307-7H are samples of low-Al metapelites from the Neoproterozoic Kingston Peak Formation (Fig. S1¹; Craddock Affinati et al., 2020). All samples comprise the mineral assemblage garnet + muscovite + quartz + plagioclase + staurolite + biotite ± chlorite. Garnets are partially resorbed and range in diameter from 1.0 to 5.1 mm. Element maps (Mn, Ca, and Y) show that the garnets preserve primary growth zoning (Fig. S1; Craddock Affinati et al., 2020). In the Kingston Peak garnets, a discontinuity in the garnet zoning profiles marks a hiatus in the garnet growth history. An increase in Y occurs at the discontinuity with higher concentrations in the rims, coinciding with a decrease in Ca (Fig. 2; Fig. S1). An increase in Mn also occurs across the discontinuity; this was interpreted as a step increase that was subsequently modified by diffusional relaxation (Craddock Affinati et al., 2020).

The discontinuity was interpreted to be the result of crossing a prograde garnet-consuming, staurolite-producing reaction, followed by retrograde regrowth when the same reaction was crossed in the retrograde sense upon cooling (Pyle and Spear, 1999; Craddock Affinati et al., 2020). All samples contained an abundance of monazite grains, the majority of which occurred in the matrix. Inclusions in garnet were analyzed from sample SSFM307-7G (one grain, M9) and sample SSFM307-7H (three grains, M1, M6, M9). SSFM307-7H was the only sample containing xenotime, which was found in the matrix. Monazite grains ranged in size from 450 to 15 μm, and xenotime grain sizes were 60 μm (X1) and 30 μm (X2). No allanite was found in these samples. The samples are unoriented; however, the preserved fabrics were likely formed during top-to-the-southeast–directed shear (Fig. 3), consistent with findings from oriented samples collected from the same area (Craig, 2019). Monazite and xenotime grains displayed no obvious relationships that would permit the ages determined in this study to constrain the age of the fabric.

Craddock Affinati et al. (2020) reported P-T paths that were determined on 12 garnet grains from the same samples analyzed here and sample SCFM314-1MW (from the Johnnie Formation), which contained no monazite or xenotime. P-T paths were calculated using the MATLAB script of Moynihan and Pattinson (2013), which utilizes the G-minimization codes of de Capitani and Petrakakis (2010). See Craddock Affinati et al. (2020) for a detailed description of the method and associated uncertainties. P-T paths from samples of the Johnnie Formation (SCFM314-1MW garnets G1 and G2, SCFM315-12C garnets G1 and G3, and SSFM307-8D garnet G2) document an increase in P and T starting at ~4.2 kbar and ~575 °C and ending at ~6.5 kbar and ~585 °C. P-T paths from the Kingston Peak Formation (SCFM315-17A garnet G3, SSFM307-7G garnets G1, G3, G4, and G5, and SSFM307-7H garnets G1 and G5) begin with a segment of steep pressure increase that overlaps the high–P-T part of the path captured in the Johnnie Formation garnets and continues to increase in temperature to ~590 °C during a decrease in pressure to ~5.1 kbar. The path was interpreted to continue along the decompression trajectory, whereupon it crossed a garnet-consuming and staurolite-producing reaction in a prograde sense, resulting in partial garnet consumption, and then crossed again in a retrograde sense during cooling and decompression, which resulted in garnet regrowth (Craddock Affinati et al., 2020). In the modeling, points analyzed within the regrown rims did not yield paths but plotted as scatter in P-T space. Craddock Affinati et al. (2020) attributed this to disequilibrium associated with changes in the bulk composition localized around garnet rims due to the preceding garnet-consumption reaction. Overall, a clockwise P-T path was documented and interpreted to represent burial by thrusting followed by exhumation either by erosion or fault-related exhumation, or a combination (Craddock Affinati et al., 2020).

Trace-element and isotopic data were collected on monazite and xenotime in situ from the same polished sections that were used in the study of Craddock Affinati et al. (2020) to determine P-T paths. Data were collected using a laser-ablated aerosol that was split into two mass spectrometers, one to collect isotopic data for U-Th-Pb dating (a multicollector–inductively coupled plasma–mass spectrometer [ICP-MS]) and one to collect trace-element data (Y, REE, and other elements; a quadrupole mass spectrometer) (Kylander-Clark et al., 2013). The goal in collecting these data was to determine the ages of crystallization of monazite and xenotime and to interpret the timing of their growth relative to garnet. Backscattered secondary electron (BSE) images were created for each polished section on a JEOL JSM-6480LV scanning electron microscope at Northern Arizona University. Mineral grains displaying high Z values (relatively white) were analyzed using an Oxford Inca X-Sight energy-dispersive system (EDS) to identify monazite and xenotime grains (Fig. 3). Mass spectrometry data were collected at the University of California, Santa Barbara. Data acquisition and reduction were done as described in Kylander-Clark et al. (2013).

Isotopic data were collected with a Nu Instruments Plasma HR-ES multicollector ICP-MS with 12 Faraday cups and four low-mass ion counters. Trace-element data were collected with an Agilent 7700x quadrupole mass spectrometer. Both machines were connected to a Photon Machines Excite 193 nm laser-ablation system. These instruments were used to collect data for samples SCFM315-12C, SCFM315-17A, SSFM307-7G, and SSFM307-7H. For sample SSFM307-8D, we used a Nu Plasma 3D multicollector ICP-MS with 16 Faraday cups, 5 low-mass Daly detectors, and one high-mass Daly detector, along with the same Agilent 7700x quadrupole mass spectrometer, which were connected to the same Photon Machines Excite 193 nm laser. Concentrations of P, Ca, Sr, Lu, Th, U, and REEs were collected on the quadrupole mass spectrometer, and ²³⁸U, ²⁰⁶Pb, ²⁰⁷Pb, ²⁰⁸Pb, ²⁰⁴Pb, and ²³²Th data were collected on the multicollector ICP-MS.

The primary U-Pb standard used was 44069. The primary trace-element standard used was STERN. FC1M (monazite), FC1X (xenotime), and TREB were used as secondary standards to ensure the ages reported were within 2% of accepted values (Kylander-Clark et al., 2013). For both monazite and xenotime analyses, standards were analyzed five times each at the start of a new session and once each in between every five unknown analyses. Dwell time of the laser for each analysis was 11 s. The laser fluence was 1 J/cm².

Analysis spot sizes for both monazite and xenotime were 8 μm, and each grain was analyzed with 1–11 spots, depending on the size of the grain. In all, data were collected for 192 monazite spots, distributed among 5–9 grains for each of the five samples. Seven monazite grains were analyzed in SCFM215-12C, six grains were analyzed in SSFM307-8D, five grains were analyzed in SCFM315-17A, eight grains were analyzed in SSFM307-7G, and nine grains were analyzed in SSFM307-7H. Data were also collected for six xenotime spots on two grains from sample SSFM307-7H.

In situ laser-ablation chemical analysis of garnets was carried out using the Photon Machines Excite 193 nm laser and Agilent 7700x quadrupole mass spectrometer at the University of California, Santa Barbara. Analyses were done on the same garnets analyzed for P-T path modeling, following the same traverses along which microprobe data were collected (Fig. S1; Craddock-Affinati et al., 2020). Concentrations of Mg, Al, P, Ca, Ti, Cr, Mn, Fe, Sr, Zr, Lu, U, and REEs were collected with a spot size of 20 μm. Standards used for garnet analysis were BHVO and NIST612, and these were analyzed five times each at the start of a new session and once each in between every five unknown analyses (Kylander-Clark et al., 2013). Dwell time of the laser for each analysis was 11 s. Garnet traverses ranged from 14 to 82 spots, depending on the size of the garnet. Garnet G5 from SSFM307-7H was not available for analysis due to work in an earlier study that destroyed a portion of it when another garnet from the same sample was extracted for mounting into a 1 in. (2.5 cm) round pellet.

Elements maps of U, Th, and Y were produced for monazite grains from samples SCFM315-12C (M6) and 7H (M3 and M5) using the Cameca SX50 electron microprobe at Northern Arizona University. Data were collected with wave-dispersive spectrometers, a 1 μm spot, 200 nA beam current, and 15 kV accelerating voltage.

Mn, Ca, and Y element maps of all garnets used for P-T path modeling were presented in Craddock Affinati et al. (2020). New Mn and Ca element maps were produced in the current study for garnets with analyzed monazite inclusions from samples SSFM307-7G and SSFM307-7H at Northern Arizona University on a Zeiss Supra 40VP variable-pressure field emission–scanning electron microscope equipped with a Thermo Scientific UltraDry energy-dispersive X-ray spectrometer. Run conditions were 20 kV accelerating voltage and a pixel size of 16.34–20.23 µm with magnification from 21× to 26×, depending on the size of the garnet and monazite inclusion. New element maps of Mn and Ca were produced in order to identify where the monazite inclusions were hosted in relation to chemical zoning in garnet, and by inference, to the P-T path.

Dates were calculated for monazite and xenotime using two isotopic systems, U-Pb and Th-Pb. A plot of ²⁰⁸Pb/²³²Th versus ²⁰⁶Pb/²³⁸U (e.g., Briggs and Cottle, 2018) shows that most of the analyses fall within error of concordia (Fig. S2). Concordance suggests that there is minimal contribution of unsupported ²⁰⁶Pb from the breakdown of ²³⁰Th, indicating that the calculated ²⁰⁶Pb/²³⁸U ages should be reliable (Tables S1 and S2). Six monazite analyses from sample SSFM307-8D were highly discordant. From these six analyses, we produced an isochron that yielded a lower-intercept age of 14.3 ± 1.9 Ma and an upper-intercept age of 4893 ± 180 Ma (Fig. S3). The calculated isochron propagated 2σ errors in the isotopic ratios (Table S2). The ages we present for these six analyses were calculated using the ²⁰⁷Pb age-correction option of Isoplot (Ludwig, 2012).

Xenotime ²⁰⁶Pb/²³⁸U dates ranged from 169 to 157 Ma (Fig. 4A; Table S2), where grain X1 recorded dates from 169 to 160 Ma (n = 4), and grain X2 recorded dates of 159.8 and 157.3 Ma. All analyses showed similar REE patterns, with overall enrichment compared to chondrite, but concentrations were internally depleted in LREEs and enriched in HREEs with a slightly negative Eu anomaly (Fig. 5). The REE patterns showed no systematic variation with age. The 2σ errors for xenotime dates are reported in Table S2 and ranged from ±7 to ±8 Ma.

Monazite dates comprised four distinct age populations defined by peaks at 170 Ma, 160 Ma, 76 Ma, and 15 Ma (Fig. 4B). Most grains (n = 169) belonged to the two older populations. We refer to the Jurassic populations as “170 Ma” and “160 Ma” due to the centralized peaks on these dates, with generalized date ranges of >167 Ma and 167–150 Ma. There is significant overlap of the two age populations. A smaller number of grains yielded younger dates of 77.0–75.2 Ma (n = 5) and 17.1–13.3 Ma (n = 6). The 2σ errors from monazite dates are reported in Table S1 and ranged from ±1 to ±10 Ma, with an average value of ±5.4 Ma.

Jurassic monazite grains in samples from the Johnnie Formation (SCFM315-12C and SSFM307-8D) yielded dates only in the 160 Ma population (168–156 Ma; Fig. 4B). Jurassic grains in both samples were enriched in REEs compared to chondrite but internally depleted in HREEs and enriched in LREEs, and they possessed strong negative Eu anomalies. Attempts to discern REE trends within individual samples by plotting Ho/Gd (HREE/MREE) versus age, as done for instance by Stevens et al. (2015), yielded statistically insignificant correlations with an R² value of 0.19107 from sample SCFM315-12C and an R² value of 0.07265 from sample SSFM307-8D (Fig. S4). However, REE patterns from analyses within single grains displayed age-correlated trends (see Discussion section; Table 1; Fig. 6; Fig. S5, A–M).

The Miocene grains, which were only found in sample SSFM307-8D (17–13 Ma), were also internally depleted in HREEs and enriched in LREEs and possessed comparatively shallow to positive Eu anomalies. Comparisons of REE ratios and Y concentrations (e.g., Soucy La Roche et al., 2016; Briggs and Cottle, 2018) from monazite grains in the Johnnie Formation showed that the Jurassic population is more enriched in Y and has a higher Yb/Dy ratio than the Miocene population. Miocene grains had smaller negative anomalies or positive anomalies (Eu/Eu* values of 0.74–1.18), whereas Jurassic grains all displayed comparatively large negative anomalies (0.18–0.23; Fig. 7B).

Monazite element maps from Jurassic grains in the Johnnie Formation showed low-Y and high-Th cores that increased and decreased in concentration, respectively, toward the rim, and no U pattern. The oldest age domains were typically found in the cores, with younger age domains at the rims (Fig. 8).

Monazite grains from samples of the Kingston Peak Formation (SCFM315-17A, SSFM307-7G, and SSFM307-7H) yielded dates of 176–116 Ma and 77.0−75.2 Ma (SSFM307-7G and SSFM307-7H). Most individual grains yielded dates that clustered around one of the 170 or 160 Ma peaks, while some grains yielded dates belonging to both populations (Fig. 4B; Fig. S5, N–JJ). Jurassic monazite domains (176–149 Ma) were overall enriched in REEs compared to chondrite but internally depleted in HREEs relative to LREEs and possessed moderately negative Eu anomalies (Fig. 6; Fig. S5, N–JJ). The 170 Ma population analyses possessed generally higher ratios of (Yb/Dy)_N and higher concentrations of Y than the 160 Ma population, especially for analyses with dates older than ca. 168 Ma (Fig. 7A). The (Ho/Gd)_N ratios showed no age-correlated trends within either the 170 Ma population or the 160 Ma population (Fig. S4). Grains with multiple monazite domains in the 170 Ma population displayed REE patterns having no trend with age (Fig. S5, Y, V, Z, GG, II). However, grains with multiple domains in the 160 Ma population displayed age-correlated trends in REEs (see Discussion section; also Table 1; Fig. 6; Fig. S5, N, O, P, Q, R, S, T, U, W, X, BB, CC, EE, FF). Five grains that possessed domains within the 170 Ma and 160 Ma populations also possessed younger domains that ranged in age from 146 to 116 Ma. These domains were more depleted in HREEs compared to the other analyses in the same grain (Figs. 6B and 6F; Fig. S5, N, R, U, EE, FF).

Monazite element maps from Jurassic grains in the Kingston Peak Formation samples showed generally patchy domains in Y and Th and essentially uniform U (Fig. 8). Dates tended to decrease toward rims on equant grains or toward the outermost points on elongated grains (Fig. 8).

Four monazite inclusions in garnet were dated. Monazite grain M9 from sample SSFM307-7G yielded dates of 165–163 Ma (n = 3). Monazite grain M1 from sample SSFM307-7H yielded dates of 170.9 to 159.1 Ma (n = 5), M6 yielded dates of 172 to 170 Ma (n = 5), and M9 provided one age of 143 Ma.

The Cretaceous population of monazite dates (77.0–75.2 Ma) comprised five determinations from two grains, M5 from SSFM307-7G (two analyzed spots) and M7 from SSFM307-7H (three analyzed spots). These grains did not contain Jurassic domains. REE patterns displayed shallower Eu anomalies with Eu/Eu* values of 0.78–0.91 compared to grains that yielded Jurassic dates (0.48–0.72), but no difference in (Yb/Dy)_N ratios or Y concentration (Fig. 7C).

When monazite points were aggregated for all samples, an age-correlated trend of HREE depletion was observed on a plot of (Yb/Dy)_N (HREE/MREE) versus Yb, which yielded dates of 176–160 Ma (Fig. 9A). However, age-correlated trends were not discernible for monazite points that yielded dates younger than 160 Ma (Fig. 9B).

Element maps of Ca were produced for garnets in samples SSFM307-7G and SSFM307-7H to evaluate chemical zoning where monazite inclusions were dated (Fig. 10). Element maps from other garnets in these same samples documented a discontinuity near the rim, where there was a sharply defined drop in Ca, a slightly diffuse increase in Mn concentration, and a sharply defined increase in Y concentration (Craddock Affinati et al., 2020). The discontinuity was interpreted to be a result of retrograde garnet regrowth following prograde consumption along the decompression path (Craddock Affinati et al., 2020). All dated monazite inclusions were located within the regrown rims (Fig. 10).

Rim-to-rim REE traverses for the Johnnie Formation garnets from SCFM315-12C and SSFM307-8D displayed core-to-rim depletions of HREEs and MREEs (Table S3; Fig. S6). LREE trends could not be evaluated due to LREE abundances that were below the detection limit (Table S3). Rim-to-rim traverse plots of Y for SCFM315-12C garnets displayed high concentration plateaus (~5500 μg/g) in the cores that sharply dropped to <100 μg/g at about half the radius (Fig. 2). The garnet from SSFM307-8D showed very low Y concentrations throughout (<85 μg/g) and asymmetric zoning.

Rim-to-rim REE traverses for the Kingston Peak Formation garnets displayed some scatter but generally displayed relatively high HREE abundances in the cores, decreases in HREEs outward from the cores, and high HREEs in the rims (Table S3; Fig. S6). For all Kingston Peak Formation garnets, Y profiles displayed spikes near the rims up to ~340 μg/g. The cores for all four garnets from sample SSFM307-7G displayed generally low values with a slight elevation of Y in the cores up to a maximum of ~75 μg/g, and SSFM307-7H G1 displayed a prominent Y spike in the core to ~770 μg/g (Fig. 2). The Y increases at the rims corresponded to features visible in Y element maps (Fig. S1; Craddock Affinati et al., 2020) and occurred within the postdiscontinuity segments of the garnet profiles.

The high Y abundances in garnet cores from samples SCFM315-12C and SSFM307-7H can be attributed to the breakdown of xenotime during early garnet growth (Lanzirotti, 1995). The lack of similar features in garnet cores from samples SSFM307-8D, SSFM307-7G, and SCFM315-17A suggests that xenotime in these samples either reacted out before garnet growth began or may not have existed at any time in the rock (Fig. 2). The one garnet analyzed from SSFM307-7H contained a Y spike in the core, but this had a magnitude that is too low (800 µg/g) to be consistent with Y saturation, which in SCFM315-12C is represented in the two garnets with Y core plateaus of ~5500 µg/g (Fig. 2). The Y spike in the core of garnet SSFM307-7H G1 may reflect either core growth occurring as Y saturation was waning, or a section through the garnet that intersected along the edge of a Y-saturated core. Evidence presented below from monazite ages and geochemistry indicates that garnet growth postdated the 170 ± 6 Ma date of the xenotime domain obtained from SSFM307-7H. The Y profiles in garnets from SSFM307-7H and SCFM315-12C suggest that Y saturation was lost shortly after garnet growth began. Thus, we interpret the xenotime as having been metastably preserved.

Within the 170 Ma population of monazite domains, there is no correlation of (Yb/Dy)_N with age (Fig. 7A), and also no age-correlated REE trends within individual grains (Figs. 6B, 6C, and 6F; Fig. S5, N, P, Q, R, S, AA, FF, GG, II; Table 1), which suggests that garnet was not crystallizing or breaking down when these monazite domains grew. The 170 Ma population analyses have higher and clustered (Yb/Dy)_N ratios and generally higher Y concentrations compared to the 160 Ma population. These observations suggest that garnet had not begun to crystallize by the time the monazite domains in the 170 Ma population grew but was crystallizing or had crystallized by the time the monazite domains in the 160 Ma population grew. This is consistent with the Lu-Hf garnet crystallization age of 167.3 ± 0.72 Ma reported for sample SSFM307-8D in Craddock Affinati et al. (2020).

In the Kingston Peak Formation, garnet started to crystallize at ~550 °C (Fig. 11A), indicating that the 170 Ma population of monazite domains and the 170 Ma xenotime domain must have crystallized below ~550 °C. Following Janots et al. (2008), we interpret that these domains crystallized from the recrystallization of detrital grains prior to garnet growth when temperatures reached 420–450 °C (Fig. 11).

A pattern of HREE depletion was documented in the whole population of data (176–160 Ma; Fig. 9A) and within individual monazite grains in two samples, SCFM315-17A and SSFM307-7H. Decreases in HREEs within the dated intervals were documented in SCFM315-17 monazite grains M3 (169 ± 5–161 ± 5 Ma), M4 (167 ± 6–162 ± 4 Ma), and M5 (167 ± 5–150 ± 5 Ma), and in SSFM307-7H M1(167 ± 7–159 ± 5 Ma), M2 (170 ± 6–159 ± 5 Ma), and M3 (163 ± 6–161 ± 5 Ma) (Fig. 6E; Fig. S5, O, P, Q, AA, BB, CC; Table 1). The decrease in HREEs suggests that garnet was growing through these intervals, depleting the matrix reservoir of HREEs available to growing monazite grains.

In sample SSFM307-7H, xenotime domains dated 163 ± 6–157 ± 7 Ma (n = 5) fall within the same date range as most monazite domains in this sample. REE patterns for all six xenotime domains (including the 170 ± 6 Ma domain) are identical (Fig. 6), which is inconsistent with the pattern of HREE decrease displayed in monazite grains. This suggests that the younger domains of the two analyzed xenotime grains were not in chemical communication with the matrix and formed from the recrystallization of older domains.

Domains within two monazite grains from Johnnie Formation sample SCFM315-12C displayed no REE trend, suggesting that these domains grew after garnet growth ended but before garnet resorption began (Fig. S5, B, C). These include SCFM315-12C M2 (165 ± 6–158 ± 5 Ma) and SCFM315-12C M3 (164 ± 7–161 ± 7 Ma). The ages postdate the 167.3 ± 0.72 Ma Lu-Hf garnet age obtained from Johnnie Formation sample SSFM307-8D (Craddock Affinati et al., 2020). This Lu-Hf age dates the early pressure-increase portion of the P-T path (Fig. 11B), and was interpreted by Craddock Affinati et al. (2020) to indicate that garnet growth ended earlier in the Johnnie Formation than in the Kingston Peak Formation. This portion of the P-T path was captured by garnet growth in both Johnnie Formation samples SSFM307-8D and SCFM315-12C.

Individual monazite grains recorded HREE enrichment trends that suggest growth took place during a period of garnet consumption that followed the end of garnet growth. When garnet breaks down, it releases HREEs hosted within it into the matrix, causing HREE enrichment of the surrounding matrix. HREE enrichment trends were recorded in monazite grains SCFM315-17A M3 (159 ± 5–154 ± 5 Ma), SSFM307-7G M1 (162 ± 6–157 ± 6 Ma), SSFM307-7G M6 (Fig. 6D; 160 ± 5–157 ± 4 Ma), and SSFM307-7H M2 (159 ± 5–156 ± 4 Ma; Fig. 6E; Fig. S5, O, W, BB; Table 1). Data from SSFM307-7G M2 showed the highest HREE abundances at 156 ± 6 Ma. Data from SCFM315-12C showed the highest HREE abundances at similar ages: M1 at 159 ± 5 Ma, M4 at 158 ± 5 Ma, M5 at 162 ± 6 Ma, and M6 at 159 ± 7 Ma (Fig. 6A; Fig. S5, T, A, D, E, F; Table 1). The sequence of garnet growth followed by garnet breakdown was predicted to occur along the P-T path, which crosses a prograde garnet breakdown reaction after prograde garnet growth ends (Fig. 11).

REE trends within individual monazite grains provided evidence of garnet regrowth following prograde resorption. Depletion of HREEs in the matrix is predicted as garnet regrowth occurs. In sample SCFM315-17A, monazite grains M2 (Fig. 6B) and M6 displayed their lowest HREE abundances at 146 ± 5 Ma and 141 ± 4 Ma, respectively. HREE abundances decreased in the domains dated 157 ± 6–155 ± 7 Ma in monazite grain SSFM307-7G M1 (Fig. S5, S). In monazite grain SSFM307-7G M7, the youngest domain in this grain, at 149 ± 5 Ma, was also the most depleted in HREEs (Fig. S5, X). In sample SSFM307-7H, monazite grains M4 and M5 (Fig. 6F) showed a decrease in HREEs from 157 ± 4 to 145 ± 4 Ma and from 143 ± 5 to 131 ± 4 Ma, respectively. Monazite M3 from sample SSFM307-8D also showed a decrease in HREEs from 157 ± 2 to 156 ± 3 Ma (Fig. S5, J). The timing of garnet regrowth is constrained by a date of 143 ± 4 Ma obtained from a monazite inclusion (M9) within the regrown rim of a garnet in sample SSFM307-7H (Fig. S5, JJ). The decrease in HREEs documented in these grains and the inclusion age indicate that a second period of garnet growth occurred after garnet resorption. This interpretation is consistent with the P-T path, which predicts that a second phase of garnet growth occurred along the decompression part of the path due to crossing back over the garnet-consuming reaction in the retrograde sense (Fig. 11).

The monazite data, as interpreted above, provide constraints on the timing of different segments of the P-T path reported in Craddock Affinati et al. (2020) (Fig. 11). The monazite age data and geochemistry indicate that garnet began growing in the Johnnie Formation ca. 168 Ma, consistent with the previously reported Lu-Hf garnet age of 167.3 ± 0.72 Ma from sample SSFM307-8D (Craddock Affinati et al., 2020). The P-T path crossed the staurolite-producing, garnet-consuming reaction, referred to as reaction 1 in Craddock Affinati et al. (2020), along the prograde path ca. 160 Ma. In the Kingston Peak garnets, following peak metamorphism, garnet began to regrow as the rocks recrossed reaction 1 in the retrograde sense after ca. 157 Ma, and continued to grow through ca. 143 Ma. The inferred timing of the retrograde recrossing of reaction 1 between 157 and 143 Ma implies that cooling was occurring during this time, consistent with the timing of cooling through Ar closure in muscovite at 153–146 Ma in the lower-grade Indian Pass area ~10 km to the southeast (Hoisch et al., 2014).

The new monazite age data substantiate earlier findings of Middle Jurassic metamorphism in the Funeral Mountains (Hoisch et al., 2014; Craddock Affinati et al., 2020) while also providing new age constraints on the P-T path. The burial implied by the initial pressure-increase portion of the P-T path has been interpreted to reflect Jurassic retroarc crustal shortening (Hoisch et al., 2014; Craddock Affinati et al., 2020). Retroarc crustal shortening coincided with the initiation of Franciscan subduction (Mulcahy et al., 2018) and/or collisions of arc terrains along the continental margin (Dorsey and LaMaskin, 2007; Schweickert, 2015; Saleeby and Dunne, 2015). Burial in the Funeral Mountains was also contemporaneous with magmatic zircon crystallization ages of the Coast Range ophiolite, early metamorphic ages from the Franciscan Complex, and age estimates for earliest terrane accretion and deformation from the western Sierra Nevada metamorphic belt (e.g., Schweickert, 2015). The timing also coincides with the East Sierra thrust system, ages determined for the Wheeler Pass thrust and correlative thrusts in the southern Sevier fold-and-thrust belt, and the Luning-Fencemaker fold-and-thrust belt (Wyld et al., 2003; DeCelles, 2004; Dunne and Walker, 2004; Giallorenzo et al., 2018). The overlap in timing of these events suggests that the initiation of east-dipping Franciscan subduction in the Middle Jurassic (ca. 180 Ma) and possibly terrane accretion were concurrent with large-scale underthrusting of the Jurassic back-arc (Craddock Affinati et al., 2020). We interpret the monazite age obtained in this study (176 ± 5 Ma; Table S1) to represent growth during the early stages of thrust burial before garnet crystallized.

The transition from burial to exhumation is constrained by the data in this study to have occurred at 160–157 Ma. The timing of exhumation coincided with Middle and Late Jurassic extension documented in the central Mojave Desert (Schermer and Busby, 1994), and intrusion of the 148 Ma Late Jurassic Independence dike swarm in the eastern Sierra Nevada range and Mojave Desert within a regime of northeast-southwest extension and regional sinistral shear (Schermer and Busby, 1994; Saleeby and Dunne, 2015). Saleeby and Dunne (2015) estimated that the transition from a regime dominated by dextral transpression to a regime of transtensional sinistral shear took place ca. 157 Ma. The change led to the development of rift-related basins into which the McCoy Mountains Formation was deposited in southeastern California and southwestern Arizona (Spencer et al., 2011). This interpretation is supported by the geochemistry of lavas and sills within the lower part of the McCoy Mountains Formation (Spencer et al., 2011), and 152–150 Ma volcanic ash beds within the Morrison Formation in Utah, Colorado, and New Mexico (Christiansen et al., 2015). Possible explanations for Late Jurassic to Early Cretaceous exhumation in the Funeral Mountains include lower lithospheric delamination (e.g., as suggested for the Late Cretaceous by Wells et al., 2012), oversteepening of the orogenic wedge, which caused gravitationally induced normal faulting of overthickened crust (e.g., Platt, 1986), and/or erosion of topographically elevated thickened crust (e.g., England and Thompson, 1984). The implied decrease in net convergence ca. 157 Ma may have yet led to gravitationally induced normal faulting of overthickened crust.

Two monazite grains in this study yielded Late Cretaceous ages. Ages of 77.0 ± 2.5–75.2 ± 2.8 Ma (grains M5 in sample SSFM307-7G and M7 in sample SSFM307-7H) overlap with 86–67 Ma muscovite ⁴⁰Ar/³⁹Ar cooling ages from the same area (Sauer, 2012; Wells et al., 2019), indicating they grew during cooling through ~425 °C (Harrison et al., 2009). The cooling has been interpreted to have been associated with exhumation along ductile shear zones that were likely precursors to the Boundary Canyon detachment (Sauer, 2012; Wells et al., 2019). The 77.0 ± 2.5–75.2 ± 2.8 Ma monazite ages come from discrete grains, suggesting that they formed from the full recrystallization of older grains during ductile shearing or possibly due to a retrograde reaction that resulted in the crystallization of new grains (e.g., Bollinger and Janots, 2006). The 77.0 ± 2.5–75.2 ± 2.8 Ma grains also have small negative Eu anomalies compared to the larger negative Eu anomalies in the Jurassic grains. One possible explanation for a smaller Eu anomaly is crystallization during interaction with oxidizing hydrothermal fluids (Bence and Taylor, 1985).

In the Death Valley area, two events occurred around the same time as the growth of Miocene-age (17.1 ± 3.1–13.3 ± 1.7 Ma) monazite found in sample SSMF307-8D: (1) rapid slip along the Boundary Canyon detachment, which exhumed the northern Funeral Mountains (11–5 Ma), and (2) major volcanism of the Timber Mountain silicic caldera system, located east of Death Valley (14–11.5 Ma; Sawyer et al., 1994). The Miocene monazite grains are discrete grains that did not grow along the margins of grains already present in the rock, and they have distinctly different REE signatures than the Jurassic grains. This suggests that the Miocene grains formed from the recrystallization of older grains associated with slip along the Boundary Canyon detachment and/or associated with hydrothermal fluids possibly related to upper-crustal convective recirculation driven by magmatic heat. The small negative and positive Eu anomalies in the Miocene grains suggest that hydrothermal fluids were interacting with the rock when these grains crystallized (Fig. 7B; Fig. S5).

This study provides detailed timing constraints on a metamorphic P-T path determined for the Chloride Cliff area in the Funeral Mountains involving prograde garnet growth followed by prograde partial resorption and retrograde regrowth in five samples of pelitic schist (Craddock Affinati et al., 2020). Specifically, the timing of the portion of the P-T path corresponding to the first phase of garnet growth is constrained to have occurred after 168 Ma and before 160 Ma (Fig. 11). This is consistent with a previously determined Lu-Hf garnet age of 167.3 ± 0.72 Ma (Craddock Affinati et al., 2020) that dates the early pressure-increase portion of the path. The prograde crossing of the reaction that partially consumed garnet (reaction 1 of Craddock Affinati et al., 2020) occurred at 160–157 Ma. Crossing of this reaction was inferred by a continuation of the P-T path along the prograde pressure-decrease trajectory recorded by garnet growth zoning (Fig. 11) and by element maps (Y, Ca, and Mn) that display a ragged discontinuity near the rims, suggesting that garnet underwent partial consumption followed by regrowth (Craddock Affinati et al., 2020; see also Fig. 2; Fig. S1). Garnet regrowth occurred when reaction 1 was recrossed in the retrograde sense during cooling. Regrowth continued at least through ca. 143 Ma, which is the age of the youngest monazite inclusion found in the regrown portion of a garnet rim. Cooling during 157–143 Ma is consistent with cooling ages from the lower-grade Indian Pass area ~10 km to the southeast, where temperatures passed through muscovite Ar closure (~425 °C) at 153–146 Ma (Hoisch et al., 2014).

REE and Y data for monazite compiled for all samples displayed two general populations centered on ages of 160 Ma and 170 Ma. The 170 Ma population possesses higher Yb/Dy and Y abundances than the 160 Ma population, consistent with the 170 Ma population predating garnet growth and the 160 Ma population coinciding with or postdating garnet growth. However, further timing constraints on the garnet reaction history could not be discerned by examining the data set aggregated as a whole or aggregated by sample. An alternative approach of examining age-correlated trends within individual monazite grains was successful in providing additional details on the timing of the full garnet reaction history (Table 1). Within individual samples, age-correlated trends of decreasing or increasing HREE abundances within individual monazite grains recorded different portions of the history, while some grains showed no trends. This suggests that changes in REE abundances in the matrix reservoir due to garnet growth and resorption were not evenly distributed throughout the matrix of individual samples; that is, there was a lack of rock-wide equilibration with respect to changes in REEs. However, there was sufficient local equilibration around individual monazite grains to track changes in REE matrix reservoir abundances that occurred as a result of garnet growth or resorption.

We thank Andrew Kylander-Clark for his assistance with data collection and processing in the laser-ablation split-stream laboratory at University of California–Santa Barbara, James Wittke for his help using the microprobe at Northern Arizona University, and Aubrey Funke for her help using the scanning electron microscope and energy-dispersive system at Northern Arizona University. We also thank Alicia M. Cruz-Uribe, Chris Mattinson, Nicholas Hayman, and two anonymous reviewers for their helpful and detailed reviews of this manuscript. This work was supported by National Science Foundation grants EAR-1550154 to T.D. Hoisch and EAR-1550158 to M.L. Wells.