Re−Os Isotope and PGE Abundance Systematics of Coast Range Ophiolite Peridotites and Chromitite, California: Insights into Fore-Arc Magmatic Processes

Abstract

We report platinum-group element (PGE) and Re concentrations, and Re−Os isotopic data for peridotites and podiform chromitite from the mid-Jurassic Coast Range ophiolite (CRO), California. Our aim is to provide insights into the formation and evolution of the CRO in a fore-arc tectonic setting. The CRO peridotites are divided into two groups: abyssal and supra-subduction zone (SSZ). They have Ir-group PGE concentrations similar to estimates for the primitive mantle and nearly chondritic relative abundances [(Os/Ir)N ≈ 1.1]. Abyssal-type peridotites have slightly subchondritic Pd-group PGE (PPGE)−Re abundances and flat chondrite-normalized patterns, whereas the SSZ-type ones are depleted overall with highly fractionated PPGE−Re patterns. The CRO peridotites have 187Os/188Os values of 0.1188 to 0.1315 (γOs = −8.3 to 1.4) and 187Re/188Os ranging from 0.022 to 0.413. The oxygen fugacity based on the V/Yb ratios of the CRO peridotites is equivalent to the fayalite−magnetite−quartz buffer. The abyssal-type peridotites are residues after ≤5% melting of the primitive upper mantle and represent a remnant of oceanic lithosphere trapped in an SSZ setting but before it was re-melted or modified by subduction processes. The abyssal-type peridotites yield an aluminachron model age of ~1.5 Ga, implying that the CRO mantle had experienced episode(s) of melt extraction before the CRO crust was formed. The SSZ-type peridotites are refractory residues after ~5% to 15% melting. Extraction of fore-arc basalts generated mainly by decompression melting resulted in the SSZ-type peridotites. The chromitite has 187Os/188Os value of 0.1250 (γOs = −3.5) and PGE−Re patterns complementary to that of boninite, indicating a genetic link to fore-arc magmatism.Ophiolites are sections of the Earth’s oceanic crust and the underlying upper mantle that have been tectonically emplaced into continental margins, providing important insights into the processes of plate tectonics, the composition of the oceanic crust and mantle, and the dynamics of Earth’s interior. Ophiolites are also valuable as ore deposits hosting precious metals, including platinum-group elements (PGEs), ferrous metals (Cr, Mn, and Ti), and base metals (Co, Cu, and Ni). The oceanic crust preserved in ophiolites may form in any tectonic setting during the evolution of ocean basins, from the mid-ocean ridge to subduction initiation and final closure [1]. The Coast Range ophiolite (CRO) is a mid-Jurassic (~172 to 161 Ma) ophiolite terrane in central California, extending over 700 km from Elder Creek at its northernmost segment extent to Point Sal at its southernmost terminus [2-5] (Figure 1). Petrologic and geochemical data indicate its formation in a supra-subduction zone (SSZ) fore-arc setting, probably above the east-dipping proto-Franciscan subduction zone [2, 6, 7]. Initiation of the subduction is considered to have possibly started along a large-offset transform fault zone, when an exotic or fringing is collided with North America [5, 6, 8]. The arc-like geochemical characteristics of pyroclastic, volcaniclastic, and hypabyssal materials of ophiolitic crust provide important evidence for the SSZ setting [4, 9, 10]. The possibility of the CRO having formed in a distal mid-ocean ridge or back-arc settings has also been proposed [11]. However, a paleo-equatorial mid-ocean spreading origin for the CRO peridotites is ruled out due to the SSZ-like highly depleted geochemical characteristics observed in several suites of CRO peridotites [2, 7, 12] and paleomagnetic evidence for limited north–south movement [13, 14]. A back-arc setting has remained plausible because back-arc extension and magmatism would be contemporaneous with arc volcanism, and it is difficult to geochemically distinguish arc-related magmas produced in fore-arc, intra-arc, or back-arc settings [15]. Since peridotites from back-arc basin are generally similar to abyssal peridotites [16], however, such a possibility has also been excluded [2]. Meanwhile, in the SSZ setting, the melt extraction process (e.g. fore-arc basalt or boninitic) to produce the current CRO peridotites has remained unidentified.The Re−Os isotope system and PGE abundances can play a pivotal role in understanding petrogenetic processes in peridotites, such as mantle melting and fluid/melt-rock interaction [17]. In upper mantle peridotites, PGEs are dominated by base-metal sulfides (BMSs) due to their highly chalcophile characteristics [18, 19]. Based on melting temperature (Tm), the PGEs are classified into two groups: Ir-group (IPGE: Os, Ir, and Ru, Tm > 2000°C) and Pd-group (PPGE: Pt, Pd, and Rh, Tm < 2000°C) [20]. During partial melting, the IPGE group resides in residual phases (e.g. sulfides, alloys, or oxides) and thus behaves compatibly relative to the PPGE group [21]. Re behaves as a mildly incompatible element during mantle melting, which results in residual peridotites with much lower Re/Os ratios than the fertile precursor [22]. The Re−Os isotopic system thus has great potential for dating melt depletion events in peridotites [23]. The high concentration of Os in the melting residue renders the Os isotopic compositions resistant to subsequent metasomatic processes. Os and Ir are not significantly fractionated during anhydrous partial melting [20, 21]. However, Os can be partitioned more than Ir into Cl-rich oxidizing slab-derived hydrous fluids or melts [24-26].Understanding the origin of the CRO is essential for unraveling the Mesozoic evolution of Cordilleran continental margin, and its development has implications for the origin of other continental margin ophiolites and provides a point of comparison with modern intra-oceanic fore-arcs (e.g. the Izu−Bonin system; [27, 28]). PGE compositions of fore-arc mantle have been rarely reported so far, yet we have indicated their potential for providing some profound insights into melting processes. This study presents PGE abundance and Re−Os isotope data for CRO peridotites and a podiform chromitite. Our aim is to constrain the formation and evolution of the CRO in a fore-arc regime and to evaluate its significance in the broader realm of ophiolite formation.The CRO is a sequence of dismembered oceanic rocks in fault contact with the underlying late Mesozoic−early Paleogene Franciscan Complex and is overlain by Upper Jurassic strata of the Great Valley sequence [3, 4, 29]. Crustal rocks (gabbro, diorite, basalt, and andesite) are the most abundant lithologies in the ophiolite though serpentinized peridotite tectonite is also widely distributed, and in some areas, it is the predominant lithology [30-32]. Igneous rocks of the ophiolite are dominantly tholeiitic basalts and basaltic andesites with arc affinities, with less common boninite, andesite, and primitive olivine–clinopyroxene–phyric basalts [3, 4]. CRO localities with extensive crustal sections in the northern Sacramento Valley (Elder Creek, Stonyford) and Diablo Range (Del Puerto Canyon, Llanada) are characterized by common felsic volcanic rocks (andesite, “keratophyre”) and shallow plutonic rocks comprising diorite, tonalite, and trondhjemite with calc-alkaline affinities [9, 33-35]. The felsic calc-alkaline series rocks overlie or cross-cut the older arc tholeiite series rocks and are in turn overlain or cross-cut by late MORB-like lavas and dikes [4, 35]. The latter magmatism is considered to have been caused by ridge collision, resulting in termination of ophiolite formation due to the change in relative plate motion and shallow subduction of young, buoyant crust [4].The CRO ranges in age from ~172 to ~161 Ma, based on U−Pb zircon ages from plagiogranite and quartz diorites associated with the ophiolite and on Ar−Ar ages of basaltic glass at Stonyford [3, 5, 36]. High-resolution U−Pb zircon chemical abrasion ages define a tighter age range of 161.2 to 167.9 Ma [3, 36]. The youngest age (~161.2 ± 0.1 Ma) is for felsic dikes in the Del Puerto ophiolite; the oldest age is from Elder Creek (~167.9 ± 0.3 Ma). Ages for most CRO locales cluster tightly at ~165 Ma, including the Ar−Ar ages on volcanic glass (164.2 ± 0.7 Ma), suggesting rapid onset of subduction and a short formation interval of circa 7 million years.Eight peridotites and one chromitite were obtained for this study, purposely sampling outcrops previously shown to exhibit the different crustal-rock associations described earlier. Chromitite occurs as irregular pods and lenses in dunite. Sample locations are shown in Figure 1, which include Grey Eagle Mine near Chrome (Red Mountain), Black Diamond Ridge (north of the Stonyford volcanic complex), Little Stony Creek (south of the Stonyford volcanic complex), Del Puerto Canyon, and the Burro Mountain.Based on mineral chemistry, previous studies [2, 12] divided the CRO peridotites into two groups: abyssal and SSZ peridotites. Abyssal peridotites are spinel lherzolites characterized by high-Al spinels (Cr# = ~15) and relatively high Al, Ti, Na, and rare earth element (REE) abundances in pyroxene, whereas the SSZ peridotites are refractory spinel harzburgites represented by high-Cr spinels (Cr# = ~40−73) and extremely low Al, Ti, Na, and REE abundances in pyroxene; dunite and orthopyroxenites from the SSZ locales have the highest Cr#s (74−77). The abyssal-type peridotites are considered to represent remnant oceanic lithosphere trapped in a SSZ setting but before being modified by subduction processes [6]. The SSZ-type peridotites are interpreted to represent partial melting in the mantle wedge above a subduction zone to form basalt (peridotites with Cr#s = ~40−55) or boninite (harzburgites and dunites with Cr# = 70−76); the latter group includes chromitites (Cr#s = 76−79) and orthopyroxenites (Cr# = 74) that may represent boninite cumulates.Three abyssal peridotite samples from Black Diamond Ridge and five SSZ peridotite samples from Little Stony Creek, Del Puerto Canyon, and Burro Mountain were obtained for this study. Also included is one chromitite from the Grey Eagle Mine, and in this sample, SSZ-type high-Cr spinel is the main constituent mineral [2]. Our samples represent five distinct locales over a geographically wide area, and previous studies [2, 6, 12] have shown their internal similarity (except for the Black Diamond Ridge lherzolites) suggesting that they are related petrologically and tectonically to one another. The abyssal peridotites represent a remnant of large-offset transform oceanic lithosphere, and we have proposed elsewhere that proto-Franciscan subduction was initiated along this fracture zone [6]. Modeling of REE shows that the abyssal peridotites are residues after ~3% dry melting of depleted MORB mantle (DMM) source in the garnet stability field (followed by conversion to spinel lherzolite at shallower depth), and the SSZ peridotites are formed by ~15%−20% further melting in the spinel stability field, possibly under hydrous conditions [12, 37].The procedures for sample preparation closely followed those described in detail by Puchtel et al. [38]. Hand specimens ~200−300 g in weight collected from surface outcrops were cut into 1−2 cm thick slabs using a diamond saw to remove any signs of alteration. The slabs were abraded on all sides using SiC sandpaper to remove saw marks, rinsed in deionized water, dried, and crushed in an alumina-faced jaw crusher. A 50 g aliquot of crushed sample was preground in a shatter box armed with alumina grinding container and puck and then reground to a flour-grade powder in an alumina-faced disk mill; the resultant powder aliquots were used for the geochemical analyses.Whole-rock major and trace element concentrations were determined using a lithium metaborate/tetraborate fusion and an inductively coupled plasma (FUS-ICP) instrument and an ICP mass spectrometer (ICP-MS), respectively, at Actlabs, Ontario, Canada. The U.S. Geological Suervey (USGS) standards (DNC-1, SY-4, and BIR-1a) were analyzed together with the unknown samples. The precision was within 5% for major elements and within 10% for most trace elements. The results are given in online supplementary Table S1.To obtain the Re−Os isotopic and highly siderophile element (HSE) abundance data, the analytical protocols detailed in Puchtel et al. [38] were followed. Approximately 1.5 g of whole-rock peridotite and 300 mg of chromitite powder, 6 mL of triple-distilled concentrated HNO3, 3 mL of triple-distilled concentrated HCl, and appropriate amounts of mixed 185Re−190Os and PGE (99Ru, 105Pd, 191Ir, and 194Pt) spikes were sealed in double internally cleaned, chilled 25 mL Pyrex™ borosilicate Carius tubes and heated to 270°C for 96 hours. Osmium was extracted from the acid solution by CCl4 solvent extraction [39], back-extracted into concentrated HBr, and purified via microdistillation [40]. Ru, Pd, Re, Ir, and Pt were separated and purified using anion-exchange chromatography following a protocol modified after [41].Osmium isotopic measurements were done via negative thermal ionization mass spectrometry [42]. All samples were analyzed using a secondary electron multiplier detector of a ThermoFisher Triton mass spectrometer at the Isotope Geochemistry Laboratory, University of Maryland, College Park. The in-run precision of measured 187Os/188Os ratios for all samples was between 0.03% and 0.06% relative. The 187Os/188Os ratio of 500 pg loads of the in-house Johnson–Matthey Os standard measured during the 2-year period leading up to the current analytical sessions averaged 0.11377 ± 10 (2 SD, N = 64). This average 187Os/188Os value is within the uncertainty of the average 187Os/188Os = 0.1137950 ± 18 measured for the Johnson-Matthey Os standard on the Faraday cups of the IGL Triton [43]; as such, no instrumental mass-bias corrections were made. The 2SD uncertainty obtained characterizes the external precision of the isotopic analyses (0.09%), which was used to estimate the true uncertainty on the measured 187Os/188Os ratios for each individual sample in this study.The measurements of Ru, Pd, Re, Ir, and Pt were performed at the Plasma Laboratory, University of Maryland, College Park, on Faraday cups of a ThermoFisher Neptune Plus ICP-MS in static mode using 1013 Ω amplifiers for all masses of interest. Instrumental isotopic mass fractionation was monitored and corrected for by interspersing samples and standards. The external precision of the analyses was estimated, based on standard measurements performed during the period of the analytical campaign, to be 185Re/187Re, 99Ru/101Ru = 0.3%, 191Ir/193Ir = 0.2%, and 194Pt/196Pt, 105Pd/106Pd = 0.10% relative (2SD). The accuracy of the data was assessed by comparing the results for the reference materials IAG MUH-1 (Austrian harzburgite), IAG OKUM (ultramafic komatiite), and NRC TDB-1 (Diabase PGE Rock Material) obtained at the IGL with the reference values. Concentrations of all HSE and Os isotopic compositions obtained at the IGL are within the uncertainties of the certified reference values [43].The average total analytical blank (TAB) measured during the present analytical campaign was (in pg): Ru 2.1, Pd 34, Re 1.1, Os 0.74, Ir 0.14, and Pt 127 (N = 9). For the peridotite samples, the average TAB constituted less than 0.1% for Os, Ir, and Ru; less than 1.0% for Pd; less than 4% for Re; and less than 6% for Pt of the total amount of element analyzed. The average TAB for the chromitite sample constituted less than 0.1% for Os, Ir, and Ru and 3.5%, 5.5%, and 16% for Re, Pt, and Pd, respectively, of the total element analyzed. We therefore cite the uncertainties on the concentrations of each element as 1/2 of the uncertainty from the TAB contribution, assuming 50% of variation in the TAB. The calculated uncertainty on the Re/Os ratio was propagated for each sample by combining the estimated uncertainties on the Re and Os abundances for each sample. The results are given in Table 1.The whole-rock major and trace element concentrations are provided in online supplementary Table S1. The samples are slightly to moderately serpentinized (18−61%; [2]) and have loss on ignition (LOI) values of 1.8% to 12.2%. The Mg number of CRO peridotites varies from 89.8 to 92.4. The CaO and Al2O3 range from 0.2 to 2.8 wt% and 0.3 to 2.7 wt%, respectively. The Ni, Co, and Cr contents range from 1820 to 2560 µg/g, 102 to 117 µg/g, and 1920 to 2900 µg/g, respectively. The V contents range from 13 to 64 µg/g. All samples show relative depletion in Cs, Rb, U, Th, and Sr. The heavy REEs have a wide range (Yb = 0.01−0.26 µg/g), but the light REEs have a relatively limited range of values (La = 0.06−0.14 µg/g), with (La/Yb)N ratios of 0.2 to 5.0 (online supplementary Table S1).The whole-rock PGE−Re concentrations of the CRO peridotites are 1.95−7.12 ng/g for Os, 2.27−4.16 ng/g for Ir, 4.52−11.59 ng/g for Ru, 1.20−10.89 ng/g for Pt, 1.21−10.15 ng/g for Pd, and 0.02−0.34 ng/g for Re (Table 1). These concentrations are within the ranges reported for mantle peridotites [17, 22]. The chromitite from Grey Eagle Mine has 96.9 ng/g for Os, 73.7 ng/g for Ir, 241.8 ng/g for Ru, 7.8 ng/g for Pt, 0.73 ng/g for Pd, and 0.10 ng/g for Re (Table 1). None of the PGE−Re concentrations exhibit any correlation with LOI values (not shown), supporting the notion that there are no secondary alteration effects. Chondrite-normalized PGE−Re abundances are shown in Figure 2. The CRO peridotites have IPGE concentrations similar to estimates for the primitive mantle and nearly chondritic relative abundances. Meanwhile, PPGE and Re concentrations span several orders of magnitude and are depleted overall relative to the primitive mantle. Abyssal-type CRO peridotites have slightly subchondritic PPGE−Re abundances and flat PPGE−Re patterns, whereas the SSZ-type CRO peridotites have highly fractionated PPGE−Re patterns. DP-18 has Ru–Pt–Pd concentrations higher than those observed in the primitive mantle (Figure 2(a)). 75BM-3 has higher Pt concentration and Pt/Pd ratio compared to the primitive mantle (Figure 2(a)). The chromitite is highly enriched in IPGE and strongly depleted in PPGE (Figure 2(b)).The CRO peridotites yield 187Os/188Os values between 0.1188 and 0.1315 (γOs = −8.3 to +1.4) and 187Re/188Os ranging from 0.022 to 0.413 (Table 1; Figure 3). The Grey Eagle Mine chromitite has 187Os/188Os of 0.1250 (γOs = −3.5) and 187Re/188Os of 0.005 (Table 1; Figure 3(a)). The Re−Os isotopic compositions do not show a correlation with LOI value (not shown), also underscoring the insignificance of secondary alteration effects. All CRO peridotites except DP-18 have 187Os/188Os ratios less radiogenic than the estimate of the primitive upper mantle (PUM = 0.1296; [44]). All samples including DP-18, however, have lower 187Re/188Os than the PUM (Figure 3(a)). Conventional Re−Os model ages (TMA; [23]) range from 7.9 Ga to the future, and Re-depletion model ages (TRD) range from 1.5 Ga to future age (Table 1). The possibility that suprachondritic Os isotopic composition of the DP-18 was induced by Os exchange with extremely radiogenic seawater of 187Os/188Os ≈ 1 [45] is excluded due to the fact that the sample has the highest 187Re/188Os ratio among the samples studied, and it has much higher Os concentration (4 ng/g) than seawater (8−10 fg/g; [46]).Representative major oxide concentrations recalculated on an anhydrous basis are shown in Figure 4. Also shown in the plots are fields for abyssal and SSZ peridotites [47, 48]. The Black Diamond Ridge peridotites plot within the field for abyssal peridotites, but the others are comparable to SSZ peridotites. In the MgO/SiO2−Al2O3/SiO2 space (Figure 5), the samples are within the range of terrestrial mantle array, indicating nearly isochemical serpentinization for the major elements. Only sample 75BM-2 had slightly higher MgO and lower SiO2, which could be due to excess olivine in this sample, as is evident in the modal proportion of olivine of up to 93.5% [2].Figure 6 shows chondrite-normalized REE patterns. Also shown in the plot are simple-model batch melting residues of DMM. The abyssal-type peridotites are depleted in light REE (LREE), which is consistent with their being melting residues [47]. In contrast, SSZ-type peridotites have U-shaped REE patterns with much lower heavy and middle REE (HREE and MREE) abundances than the Black Diamond Ridge peridotites. Strong depletion in the HREE and MREE indicates that the rocks are residues after high degree of partial melting. On the other hand, the observed enrichment in LREE requires secondary processes to have followed melt extraction. Postmelting refertilization by trace element–enriched melt or fluid of a refractory mantle in a SSZ setting can explain the elevated abundances of LREE and ultra-depletion in HREE and MREE of the samples [37, 47-50].The PGE in mantle rocks are mostly concentrated in Fe–Ni–Cu sulfides (BMS) and, in particular, monosulfide solid solution (Mss) [18, 51]. Re is not a true PGE but is often considered in conjunction with this group due to its chemical similarity and 187Os being the radioactive decay product of 187Re. The element Re exhibits siderophile behavior in metal-silicate systems and typically resides in sulfide phases in the mantle, though with some control by silicate phases [52, 53]. Depletion in PPGE−Re (Figure 2(a)) reconciles with the origin of the CRO peridotites as residues after partial melting and melt extraction. This fractionated PGE−Re patterns can be attributed to incongruent melting of BMS during mantle melting which produces Rh–Pt–Pd-enriched sulfide melt and leaves Os–Ir–Ru-enriched Mss in the residue [21].Chondrite-normalized modeled Pt/Ir and Pd/Ir ratios in residual peridotite after incongruent melting of the BMS are shown in Figure 7 for comparison. Assuming that Ir, Pd, and Pt are 100% concentrated in the BMS, their abundances were calculated using mass balance equations for nonmodal fractional melting of the BMS. The PUM is assumed to contain 150 µg/g for S, 3.5 ng/g for Ir, 7.1 ng/g for Pd, and 7.6 ng/g for Pt [54]. We have used in the calculations Mss/silicate melt partition coefficients after [20]: DIr = 3500, DPd = 370, and DPt = 360. The modeling shows that the (Pt/Ir)N and (Pd/Ir)N ratios decrease with increasing degree of partial melting (Figure 7). Except for the two samples grossly off the defined trend (DP-18 and 75BM-3), the CRO peridotites define a trend in remarkably good agreement with the modeled curve, despite uncertainties in the S contents of the primitive mantle and S solubility in partial melts. The abyssal-type CRO peridotites are residues after less than ~5% melting, and the SSZ-type CRO peridotites are residues after ~5% to 15% melting. This is consistent with the results modeled with REE (Figure 6), considering the fact that DMM represents 2%−3% melt removal from the primitive mantle [55]. Sample DP-18 has Ru–Pt–Pd concentrations higher than those observed in the primitive mantle (Figure 2(a)), probably indicating secondary refertilization processes resulting in precipitation of metasomatic Cu–Ni-rich sulfides from infiltrated silicate melts [18, 51]. Sulfidization reactions between S-rich fluids and olivine, or metals dissolved in volatile-rich alkaline melt may also result in precipitation of BMS [17, 18]. In that case, however, suprachondritic Os/Ir ratios are expected in the host peridotite due to enrichment in volatile–chalcophile element such as Os [17]. Since DP-18 has a chondritic Os/Ir ratio (Figures 2(a) and 8(a)), this possibility is excluded for this sample. Sample 75BM-3, on the other hand, has suprachondritic (Pt/Ir)N but subchondritic (Pd/Ir)N ratios (Figure 7). Also, 75BM-3 has higher Pt concentration than that in the primitive mantle (Figure 2(a)), which cannot be explained by melt removal, as Pt has slightly lower partition coefficients in sulfides than that of Pd [21]. The Pt excess observed in sample 75BM-3 may be due to nugget effects of discrete Pt-rich microphases, possibly produced during secondary melt percolation or serpentinization [51, 56-58].The 187Os/188Os ratios for the CRO peridotites display generally positive correlation with 187Re/188Os but do not define meaningful Re−Os isochrons (Figure 3(a)), likely indicating Re mobility in the samples. The data point distributions result in unrealistic TMA model ages with some sample sets yielding future ages while others produce ages older than the age of the Earth (Table 1). The open-system behavior of Re is widely observed in mantle peridotites [23, 46, 59, 60]. The TRD model age, which corresponds to the minimum estimates of the age of melt extraction and assumes no change in 187Os/188Os after time of melt extraction [23], can be an alternative dating method. However, it can provide a good approximation for the time of melting in highly refractory peridotites (e.g. SSZ type) but significantly underestimate for relatively fertile samples (e.g. abyssal type). While the Re−Os isotope data do not define an isochron, all samples possess 187Re/188Os ratios lower than the putative PUM (Figure 2(a)), suggesting that metasomatic addition of Re has not been significant. This is also substantiated by the samples’ suprachondritic (Pd/Re)N ratios (Figure 2(a)). For comparison, the reference isochron forced through PUM of 165 Ma, the CRO formation age, is shown in Figure 3(a). It should be pointed out that the CRO peridotites yield much steeper array compared with the reference line in the plot, indicating its origination from a protolith that had undergone ancient melt depletion episode(s).Al2O3, CaO, or Yb are usually considered to exhibit degrees of incompatibility similar to that of Re during mantle melting but are less easily perturbed than Re during metasomatism [60, 61]. Therefore, Al2O3 content is often used as a proxy for Re/Os to estimate the mantle separation ages of peridotites. In the alumina 187Os/188Os peudo-isochron (aluminachron; Figure 3(b)), the abyssal-type and SSZ-type CRO peridotites show two independent correlations, implying temporally distinct melting episodes. The trend of abyssal-type CRO peridotites (solid circles) passes through the PUM in the aluminachron, suggesting that these rocks might have originated from a PUM-like source. Assuming that there has been a single or multiple melting events closely spaced in time, the 187Os/188Os ratio at ~0.7 wt% Al2O3 as representative of Re/Os = 0 in anhydrous melting [61, 62] on the aluminachron can be considered as the initial ratio, which corresponds to TRD of 1.50 ± 0.13 Ga for the abyssal-type CRO peridotites (Figure 3(b)). This implies that the CRO mantle had experienced episode(s) of melt extraction before the abyssal-type CRO crust was formed. While this time of lithospheric stabilization seems too old to be expected in abyssal peridotite, the age decoupling between abyssal or fore-arc peridotites and their overlying crust, and long-term preservation of refractory domains, including a parcel of buoyant arc mantle, in the asthenosphere have been reported in several cases [46, 49, 63-67]. The aluminachron age of ~1.5 Ga for the abyssal-type lherzolites corresponds roughly to the age of rifting in the Columbia supercontinent, separating Laurentia/Baltica/Siberia from Antarctica/North Australia/South America to form an ocean basin that would evolve into the proto-Pacific Ocean basin in the Phanerozoic [68]. The Os isotopic evidence for ancient melting can be explained by unradiogenic sulfides or platinum-group metals encapsulated in silicate or oxide hosts and protected from subsequent melting or diffusion [52, 59, 65].The SSZ-type CRO peridotites show a positive correlation with a much steeper slope in the aluminachron (open symbols on the diagram in Figure 3b). However, the most refractory SSZ-type sample 75BM-2 (represented by the lowest 187Os/188Os ratio and Al2O3 content) falls on the lower limit of the correlation for abyssal-type peridotites, indicating that the abyssal-type CRO peridotites might have been the source for the second melt extraction event. It is estimated that the second event(s) occurred at ~165 Ma, but accurate dating is difficult with currently available data. We have shown a regression line for Burro Mountain peridotites for comparison in the plot (Figure 3(b)). We observe that samples with and without metasomatic overprinting in the Burro Mountain and Del Puerto peridotites are collinear in the plot (Figure 3(b)), which suggests a possibility of coeval second-stage melting episode and melt percolation events.Volcanism in a near-trench setting during subduction initiation is represented by fore-arc basalts and boninites [28, 69-71]. For example, in the Izu−Bonin−Marianas (IBM) fore-arc, the former stratigraphically underlie the latter [28, 69-73]. Fore-arc basalts are tholeiitic with major element compositions at the depleted end of the MORB array [70, 73-75]. They are characterized by being lower in TiO2, P2O5, Zr, and LREEs than MORB, implying their derivation from mantle that is more refractory than MORB-source mantle [70, 76, 77]. The fore-arc magmas are considered to be generated by MORB-like mantle decompression melting during near-trench spreading with little or no mass transfer from the subducting plate [28, 69-72].Compared with fore-arc basalts, boninites have higher MgO and SiO2, lower TiO2, and relatively higher concentrations of fluid-mobile trace elements (e.g. K and Sr) [28, 71, 74, 78]. Boninites were generated later when the residual, highly depleted mantle (refractory harzburgite) after extraction of the fore-arc basalts melted at relatively shallow levels after being fluxed by fluids and melts from the subducting plate [27, 28, 69, 71, 79]. Boninitic lavas are located stratigraphically above arc tholeiite basalts in the CRO [4, 35].Vanadium is a moderately incompatible element and exists in multiple oxidation states (V3+, V4+, and V5+) in terrestrial magmas. Partition coefficient for V in mantle phases therefore show a strong dependence on oxygen fugacity, which decreases with increasing fO2 [80-82]. Vanadium concentration can thus provide information on oxygen fugacity. Figure 8 is a plot of V against Yb for all CRO peridotite samples, with modeled trends for fractional melting in spinel lherzolite facies at different oxygen fugacities. The V contents of the CRO peridotites show a positive correlation with Yb, indicating a dominant control by igneous processes rather than hydrothermal alteration. It is worth noting that most CRO peridotites fall along the fayalite−magnetite−quartz (FMQ) buffer, regardless of whether they are abyssal or SSZ types. The Black Diamond Ridge peridotites fall within the range of abyssal peridotites, and SSZ-type CRO peridotites overlap with the abyssal peridotite field, but they also extend to lower concentrations beyond that field; one sample (75BM-2) appears to somewhat more oxidized than the others. The oxygen fugacity of the MORB source is at or just above the FMQ buffer, whereas subarc mantle is generally more oxidized than ambient mantle, including a substantial proportion of rocks with fO2 > FMQ+1 [83-85]. Oxidation of the mantle is attributed to the infiltration of oxidized fluids or silicic melts derived from the subduction slab, where sulfate could be the most plausible oxidant [85, 86]. This observation indicates that the CRO peridotites had limited interaction with oxidizing fluids in the fore-arc setting due to the relatively short residence time above a dehydrating slab, and fore-arc basalts are generated mainly by decompression melting [83, 87, 88]. This finding supports the claim of Birner et al. [49] that the fore-arc mantle is not pervasively oxidized relative to mid-ocean ridge mantle.Supra-chondritic Os isotope ratios (up to 187Os/188Os ~0.157) are commonly found in arc peridotite xenoliths [24, 25, 89], which is attributed to inputs of radiogenic Os from a subducted oceanic crust and sediments. Osmium can be mobile during slab dehydration or melting processes, and its mobility in a fluid or a melt increases with increasing oxygen fugacity and salinity [26, 90]. The CRO peridotites studied, however, are characterized by subchondritic Os isotopic compositions (Figure 3). In a fore-arc region, slab dehydration and fluid salinity are relatively low in cold subduction and high in hot subduction [91]. Oxygen fugacity equivalent to the FMQ buffer in a fore-arc setting (Figure 8) and possibly cold subduction as evidenced by slight to moderate serpentinization (18−61%; [2]) observed in the CRO peridotites would have resulted in Os migration being limited in this setting. Unradiogenic 187Os/188Os ratios are also observed in IBM fore-arc peridotites [66]. This is independently supported by predominantly chondritic Os/Ir ratios observed in the CRO peridotites studied (Figure 9(a)). Also note that the Os isotopic ratios do not show correlation with Os content of the peridotites (Figure 9(b)), ruling out any mixing processes. Sample 75BM-2 has slightly suprachondritic Os/Ir ratio (Figure 9(a)) and Os and Ru enrichment relative to Ir (Figure 2(a)). Meanwhile, 75BM-2 has the lowest 187Os/188Os ratio among the CRO peridotites (Figure 9(b)), indicating that the elevated Os–Ru abundances could be nugget effect (i.e. heterogeneous distribution of laurite or Ru–Os alloy) rather than secondary fluid infiltration.This finding, however, is not in line with the results of a previous study of serpentinized peridotites from Point Sal, the southernmost CRO remnant, carried out by Snortum and Day [7]. The Point Sal peridotites record significant melt depletion (>20%), which resulted in exhaustion of sulfides in the source during melting, leading to loss of Os (open diamonds in Figure 9). Meanwhile, the Point Sal peridotites are characterized by radiogenic 187Os/188Os (Figure 9(b)) and fractionated Os/Ir ratios (Figure 9(a)), as well as suprachondritic (Pd/Ir)N (~3−25) or (Pt/Ir)N (~2−6) ratios, and enrichments in Ba, U, and Sr [7]. Also note that the major element compositions of the Point Sal peridotites are less systematic than ours, and some samples (PS1703) are highly enriched in Al2O3 and CaO (Figures 4 and 5, and online supplementary Figures S1 and S2) but depleted in Sc (not shown). Taken together with the high LOI values (14−18 wt%) of the Point Sal samples, this is probably an effect of hydrothermal alteration and secondary gabbroic melt(s) infiltration. That is, the Point Sal peridotites could be a mantle section associated with Cl-rich oxidizing fluid-driven melt depletion and refertilization processes in a fore-arc setting [7].Grey Eagle Mine chromitites have a chemical composition with boninitic affinity [2]. Podiform chromitite is considered to be the combined product of crystal fractionation in the early stages and melt-harzburgite reaction in an open system as primitive hydrous melts migrate through the upper mantle in a SSZ setting [92]. Dissolution of orthopyroxene in harzburgite by reaction with hydrous melts can produce dunite coupled with crystallization of high-Cr# chromite. The chromite in boninites could thus be the products of primitive melt crystallization in the crust or of incongruent melting of orthopyroxene in the uppermost mantle, which is carried to the surface by entrapment in the melts [93]. The studied chromitite is highly enriched in IPGE and strongly depleted in PPGE (Figure 2(b)). Ru shows more compatible behavior than Os and Ir in chromite as observed in the previous studies [94, 95]. The IPGE-chromite association could be due to the presence of IPGE minerals (IPGM: laurite, erlichmanite, Os–Ir ± Ru alloys, and so on) in chromite [95] or IPGE incorporation in solid solution into the chromite structure [94, 96]. The IPGM could be one of the first phases to crystallize from cooling magma, or refractory residual phase during partial melting in the mantle, which were then entrapped during growth of chromite [97-99]. Available PGE−Re abundances of boninites are shown in Figure 2(b) for comparison. The boninites show PGE−Re patterns complementary to that of chromitite. This observation suggests that the boninitic melts might be generated during concomitant fractional crystallization of olivine and chromite [93, 97]. However, the number of samples analyzed is insufficient to generalize and needs to be validated with more data in the future. The Os isotope composition of the chromitite is within the range of the initial 187Os/188Os ratios of the CRO peridotites (Figure 3(a)), suggesting indistinguishable flux of radiogenic Os from the slab into the mantle source of the boninite.Geochemical data show that the CRO of California contains both SSZ and abyssal peridotites. The abyssal types represent a remnant of oceanic lithosphere trapped during subduction initiation along transform fault(s) (the proto-Franciscan subduction zone) in mid-Jurassic which produced the SSZ-type peridotites.Depletion in PPGE−Re and subchondritic 187Os/188Os ratios indicate that the CRO peridotites are residues after partial melting and melt extraction. Chondrite-normalized modeled HREE abundances, and Pt/Ir and Pd/Ir ratios indicate that the abyssal-type CRO peridotites are formed after less than ~5% melting and the SSZ-type ones after ~5% to 15% melting of the PUM.The aluminachron model age of ~1.5 Ga for the abyssal-type peridotites implies that the CRO mantle had experienced melt extraction event(s) much older than the formation age of the CRO lithosphere, which may reflect rifting of the Columbia supercontinent.Volcanism in a near-trench setting during subduction initiation is represented by fore-arc basalts and boninites. Unfractionated (Os/Ir)N ratios of ~1.1 and the fO2 values close to the FMQ buffer indicate that the SSZ-type peridotites are residues after mainly fore-arc basalt extraction with little or no mass transfer from the subducting plate. The abyssal-type peridotites are considered to be the source for the second melting event in a fore-arc regime. Meanwhile, the podiform chromitite has PGE−Re pattern complementary to that of boninites, suggesting a genetic link between the two by flux melting and crystallization in a previously depleted mantle as a result of fore-arc basalt magmatism.Geochemical data show that the CRO of California contains both SSZ and abyssal peridotites. The abyssal types represent a remnant of oceanic lithosphere trapped during subduction initiation along transform fault(s) (the proto-Franciscan subduction zone) in mid-Jurassic which produced the SSZ-type peridotites.Depletion in PPGE−Re and subchondritic 187Os/188Os ratios indicate that the CRO peridotites are residues after partial melting and melt extraction. Chondrite-normalized modeled HREE abundances, and Pt/Ir and Pd/Ir ratios indicate that the abyssal-type CRO peridotites are formed after less than ~5% melting and the SSZ-type ones after ~5% to 15% melting of the PUM.The aluminachron model age of ~1.5 Ga for the abyssal-type peridotites implies that the CRO mantle had experienced melt extraction event(s) much older than the formation age of the CRO lithosphere, which may reflect rifting of the Columbia supercontinent.Volcanism in a near-trench setting during subduction initiation is represented by fore-arc basalts and boninites. Unfractionated (Os/Ir)N ratios of ~1.1 and the fO2 values close to the FMQ buffer indicate that the SSZ-type peridotites are residues after mainly fore-arc basalt extraction with little or no mass transfer from the subducting plate. The abyssal-type peridotites are considered to be the source for the second melting event in a fore-arc regime. Meanwhile, the podiform chromitite has PGE−Re pattern complementary to that of boninites, suggesting a genetic link between the two by flux melting and crystallization in a previously depleted mantle as a result of fore-arc basalt magmatism.The data are provided in the article and the supplementary materials.The authors declare that they have no conflicts of interest.This study was supported by the National Research Foundation of Korea (NRF) funded by the Korea government (MSIT) (NRF-2022R1A2C1003508).We thank associate editor Chuan-Lin Zhang and two anonymous reviewers for their constructive comments on this article.The supplementary materials available include an Excel table and two figures. Table S1 provides major and trace element concentrations for spinel peridotites and chromitite from Coast Range ophiolite, California. Figure S1 shows the major element compositions for Point Sal peridotites, the southernmost Coast Range ophiolite remnant. Figure S2 shows the MgO/SiO2 versus Al2O3/SiO2 ratios for Point Sal peridotites.