Olivine thermometry and source constraints for mantle fragments in the Navajo Volcanic Field, Colorado Plateau, southwest United States: Implications for the mantle wedge
 The evolution of the mantle wedge below the Colorado Plateau during low-angle subduction has been investigated by analysis of fragments from the Navajo Volcanic Field (NVF), most from serpentinized ultramafic microbreccia (SUM) diatremes. In most SUM-hosted olivine, concentrations of Al and V are < 1 ppm, and Cr, Ca, and Na concentrations also are unusually low: most temperatures from three olivine thermometers (Al, Cr, and V-based) are in the range 530°C to 650°C. The temperatures are consistent with the presence of chlorite in some inclusions, and they support the utility of olivine thermometry for diverse mineral assemblages in cool peridotite. Most pressures calculated for discrete diopside grains in SUM correspond to depths in the range 80 to 120 km. Diopside is relatively high in Li (~3.5 ppm), and two of five diopside grains have high Sr/Nd. SUM diatreme sources are inferred to be serpentine-rich mélange tectonically eroded from the forearc, transported above the Farallon slab, and incorporated into the lithosphere of the Plateau. Garnet peridotite xenoliths from minettes of the NVF record deeper depths in the range 120 to 150 km. These garnet peridotites also may be from forearc mantle emplaced during Farallon subduction. Calculated temperatures preclude the possibility that asthenosphere was in contact with that lithosphere at or near 150 km depth for tens of m.y. before NVF formation. Structures observed in seismic images of mantle to depths of 150 km below the central Colorado Plateau may be related to Farallon subduction, not inherited from Proterozoic lithosphere.
- Trace-element olivine thermometry works for diverse assemblages in cool mantle
- The central Colorado Plateau lithosphere to 150 km was mostly cold at 25 m.y
- Eroded forearc rock was emplaced into that lithosphere by Farallon subduction
 Effects of low-angle subduction on mantle lithosphere have been investigated by analysis of mantle fragments from the Colorado Plateau. Most samples were hosted by diatremes of serpentinized ultramafic microbreccia (SUM), an unusual rock type of the Navajo Volcanic Field (NVF). A segment of the Farallon slab was subducted at a low angle below the Plateau during the Laramide orogeny, as illustrated by Saleeby  (Figure 1). The time period in which the low-angle segment rolled back or collapsed is not well-defined, but a common interpretation is that the slab was removed from below the Plateau about 30 to 20 m.y. ago [Levander et al., 2011]. Some of the SUM diatremes were emplaced at 30 Ma and others at 25 Ma [Roden et al., 1979; Smith et al., 2004].
 The SUM diatremes were emplaced as gas-solid mixes [McGetchin and Silver, 1970]. Although the serpentine-rich matrices of the diatremes have been referred to as kimberlite, there is no evidence for a magmatic component in the erupting mix [Gavasci and Kerr, 1968; Roden, 1981], and the eruptions were probably triggered by intrusion of minette magma into hydrated mantle [Smith and Levy, 1976]. Many inclusions in SUM diatremes are of obvious crustal rock types, but eclogite and peridotite are also present, as are discrete grains of pyrope, forsterite, and Cr-diopside. Garnet peridotite is very rare, and some of the forsterite and diopside grains are from sources unlike most or all represented by rock inclusions. Because some of the inclusions have been interpreted as fragments from the eruption sources [Smith, 2010], and none represent foreign fragments transported by magma, the term “xenolith” has been avoided in referring to mantle fragments in the SUM diatremes.
 Conflicting histories of peridotite and eclogite inclusions within the SUM diatremes have been proposed, and no consensus has been apparent. Lawsonite eclogites were suggested by Helmstaedt and Doig  to be fragments of a subducted slab: that hypothesis was disputed by Wendlandt et al.  and Smith et al.  but maintained by Usui et al. . Peridotite inclusions were proposed by Mercier  to be slices of oceanic ophiolites underthrust beneath the Plateau. Some of the peridotites contain hydrous minerals such as titanian clinohumite, chlorite, and tremolite, and Smith  suggested that these rocks represent Proterozoic mantle lithosphere, sheared in deep fault zones and hydrated by water from an underlying slab. However, Smith  later summarized evidence that the characteristic eclogites and some of the other mantle inclusions represent continental mantle tectonically eroded from near the trench and dragged about 700 km during subduction to below the Plateau, an idea suggested but not preferred for the eclogites by Helmstaedt and Schulze .
 Interactions between the Farallon slab and the overlying mantle remain poorly understood, and some interpretations based on seismic data depend upon the state of the lithosphere following those interactions. Interpretations of the maximum lithosphere thickness below the NVF now range from about 120 km to 160 km [Liu et al., 2011; Wilson et al., 2010]. The Moho below the central Colorado Plateau appears to be gradational, one suggested cause being Proterozoic eclogites, and subcrustal seismic discontinuities have been suggested to represent possible Proterozoic suture zones [Wilson et al., 2005, 2010]. As summarized by Jones et al. , some asthenosphere may have remained above most or all the slab, and small-scale mantle convection may have occurred during the Laramide orogeny below the Colorado Mineral Belt (COMB): an alignment of small intrusions in the COMB extends into the northeast margin of the NVF (Figure 1).
 Mantle fragments have been analyzed to test some of these ideas. Trace element analyses by LA-ICP-MS are emphasized. Attention is given to thermobarometry, and compositions of olivine are particularly useful, because of the context provided by the comprehensive study of trace elements in olivine by De Hoog et al.  and by the olivine thermometers that they developed. The samples include discrete grains of forsterite and chrome diopside from the SUM diatremes. Olivine in five SUM-hosted, two minette-hosted, and four basalt-hosted peridotite inclusions also has been analyzed. The five SUM-hosted inclusions span the range from spinel peridotite with no apparent mantle hydration to peridotite with chlorite and titanian clinohumite [Smith, 1979, 2010]. The two minette-hosted rocks are garnet peridotites from The Thumb [Ehrenberg, 1978, 1982a, 1982b], and the four basalt-hosted rocks are spinel peridotites from the Grand Canyon volcanic field [Smith and Riter, 1997] (Figure 1a). All these rock samples had been analyzed by other techniques in previous studies.
1.1 Analytical Procedures
 Analytical procedures summarized here are described in detail in the Analysis Procedures included in the Supporting Information.1 Trace elements were measured in three sessions of laser ablation quadrupole inductively-coupled plasma mass spectrometry (LA-ICP-MS). These analyses were made with a 193 nm New Wave FX193 excimer laser system, a 150-µm spot, and a large-format laser cell. Ablation products were analyzed with an Agilent 7500ce quadrupole mass spectrometer. NIST SRM 612 was the primary standard. BCR-2G was used as a secondary standard in all three sessions, and NIST SRM 614 was used in two of three. Data were processed with the Iolite software package [Paton et al., 2011], and representative limits of detection calculated for individual olivine analyses by that software are in Table 1a. Table S1 contains a summary of results for the secondary standards. New electron probe analyses were acquired with a JEOL JXA-8200 at 15 kV, a 35 nA beam current, and a minimum spot.
|Discrete grains||Peridotite inclusionsa|
|Green Knobs||Green Knobs|
|L Ab||L B||L C||L D||L E|
|Other datac||1, 2, 3, 4||1, 2, 3, 4||2, 3, 4||1, 2, 4, 12||5|
|100*Fe/(Fe + Mg)||7.81||8.15||7.57||8.04||9.48||10.61||8.57||10.30||9.60||8.00|
- na, not analyzed; nd, not detected. Representative limits of detection (ppm) for LA-ICP-MS data: Li, 0.007; Na, 1.5; Al, 0.04; Ca, 15; Ti, 0.1; V, 0.015; Cr, 0.13; Sr, 0.001; Y, 0.001; Zr, 0.003; Nb, 0.001; Yb, 0.003; Lu, 0.001
- a Assemblage information in Table 3
- b Contains sparse lamellae of diopside and titanian clinohumite
- c 1. Smith and Levy ; 2. Smith ; 3. Roden et al. ; 4. Roden and Shimizu ; 5. Smith ; 6. Smith and Riter ; 7. Riter ; 8. Ehrenberg [1982a]; 9. Ehrenberg [1982b]; 10. Smith et al. ; 11. Roden and Shimizu ; 12. Condie et al. ; 13. This paper.
2 Contexts and Compositions of Minerals
 The population of discrete olivine grains in the SUM diatremes is unusually Fe-poor for Proterozoic and younger mantle. In particular, it is distinctly less iron-rich than most olivine in abyssal peridotite and in xenoliths from the adjacent Basin and Range and the Rio Grande Rift provinces (Figure 2). The discrete grains also are Fe-poor compared to the olivine typical of spinel peridotite inclusions from Green Knobs and from other Colorado Plateau occurrences, but the populations overlap. The compositions of grains from SUM plotted in Figure 2a represent the Green Knobs diatreme in the southern part of the NVF and the Moses Rock diatreme in the northern part. The most magnesian grains are all from the Moses Rock diatreme.
 The five discrete olivine grains chosen for analysis are from the SUM diatreme at Green Knobs [Smith and Levy, 1976]. The grains have fe values (100*Fe/(Fe + Mg)) in the range 7.6 to 9.5 (Figure 2a). Compositions of four of the five cluster about fe of 8, typical of the grains in the SUM diatremes. Four of the five, including the most Fe-rich, are free from optically visible lamellae, but one grain (fe = 7.8) contains sparse thin lamellae of diopside, identified by electron probe analysis, and of titanian clinohumite, identified by optical properties. The identification is consistent with the observations of Mosenfelder et al. , as they identified diopside as lamellae in a discrete olivine grain from Green Knobs and diopside plus titanian clinohumite as lamellae in a grain from the nearby Buell Park SUM diatreme.
 The 5 peridotite inclusions from Green Knobs chosen for olivine analysis have diverse mineralogies. Two of the five rocks are spinel peridotites with little to no evidence of mantle hydration: N16-GN lacks hydrous minerals of probable mantle origin and N53-GN contains only a trace of pargasite. The other three have prominent hydrous minerals: N17-GN contains chlorite in clots that partly replaced spinel; N23-GN contains titanian clinohumite and clots of chlorite plus tremolite in pseudomorphs after garnet; N178-GN contains chlorite, forsterite, diopside, enstatite, chrome spinel, and trace phlogopite.
2.1.1 Trace Element Concentrations in Olivine
 Average olivine analyses are in Tables 1a and 1b. Each discrete grain was analyzed by LA-ICP-MS at 5 to 9 spots. Grains appeared homogeneous. Cr, Al, V, Ca, and Na are the trace elements with the most distinctive concentrations in olivine of both discrete grains and peridotite inclusions from Green Knobs. For instance, V is less than 1 ppm in all Green Knobs fragments, and Al is less than 1 ppm in all but one. No values for Ca, V, Al, and Cr in olivine of xenoliths from other occurrences are as low as any of the values for Green Knobs (Figures 3, 4). Ca, V, and Cr concentrations in olivine of all five Green Knobs rocks are as low or lower than values in the discrete grains. Na is similarly depleted in olivine in the SUM-hosted grains and inclusions, as values are near or below the representative detection limit of 1.5 ppm, in contrast to the values of 20 to 400 ppm Na typical of the xenoliths analyzed by De Hoog et al. . The few comparable values for Na, for Cr and Al (Figure 3), and for Ca and V (Figure 4) represent orogenic peridotite masses. In contrast, concentrations in olivine of the four xenoliths from the Grand Canyon field and the two from The Thumb are within the higher ranges typical for mantle peridotite xenoliths.
|Spinel peridotite||Garnet peridotite|
|Basalt, Grand Canyon Field||Minette, The Thumb|
|Other dataa||6, 7||7||7||7||8, 9, 10, 11||8, 10, 11|
|100*Fe/(Fe + Mg)||8.67||7.83||9.72||9.92||11.19||8.76|
- a References listed in Table 1a.
 Concentrations of less temperature-sensitive trace elements in the Green Knobs olivine samples are similar to those in many other peridotites (Figure S1). Li is an element of particular interest in the mantle wedge above the Farallon slab, because Li has been used to trace fluid flow during subduction [Paquin et al., 2004; Savov et al., 2007]. However, concentrations of Li in all analyzed Colorado Plateau olivine fall in the range 1.2 to 2.8 ppm and are positively correlated with Fe, a typical range and correlation for olivine in other tectonic settings and host rocks. Four of the five discrete grains have Nb concentrations greater than 0.05 ppm, values that exceed those of olivine in most spinel-facies peridotite but are typical in kimberlite-hosted garnet peridotite. In all SUM-hosted samples, Ti and Zr concentrations are in the lower parts of ranges typical for peridotite olivine.
 Five discrete diopside grains from the SUM diatreme at Green Knobs have been analyzed, together with clinopyroxene in one SUM-hosted peridotite inclusion. All five diopside grains from Green Knobs have Na2O and Al2O3 contents similar to those of clinopyroxene in kimberlite-hosted garnet peridotite xenoliths (Figure 5). A diopside grain from the Garnet Ridge SUM diatreme [Skogby et al., 1990] also plots with the garnet peridotite group, as do 11 of the 34 clinopyroxene grains from the SUM diatreme at Moses Rock [McGetchin, 1968]. The Garnet Ridge grain and 8 of these 11 from Moses Rock are more sodic than the Green Knobs grains, perhaps indicating that the northern SUM diatremes tap garnet peridotite sources slightly different from those sampled at Green Knobs. Many of the 34 clinopyroxene grains from the SUM diatreme at Moses Rock analyzed by McGetchin  plot in the spinel-peridotite population, and so the discrete grains analyzed here represent only a subset of the clinopyroxene grains in the SUM diatremes.
|M A||M B||M C||M D||M E|
|Other dataa||1, 2, 3, 4|
|100*Fe/[Fe + Mg]||4.96||5.15||5.01||6.14||4.57||7.20|
 Pyroxenes in peridotites from Green Knobs have compositions distinct from those of the discrete grains on the Na2O-Al2O3 plot in Figure 5. Rock N23-GN and one other discussed by Roden and Shimizu  contain chlorite plus amphibole clusters inferred to have replaced garnet, and pyroxenes of these two rocks plot with the population from garnet peridotite at The Thumb. Pyroxenes of four spinel peridotites plot with those of basalt-hosted spinel peridotites. Diopside in the antigorite-peridotite inclusions and the chlorite-peridotite inclusion from Green Knobs described by Smith  is less sodic and typically less aluminous than the discrete grains.
2.2.1 Trace Element Concentrations in Diopside
 Trace-element compositions of the diopside grains (Table 2) are unlike those of clinopyroxene in the spinel peridotite inclusions from SUM diatremes but similar to those in some garnet peridotite. Clinopyroxene in four of the five studied Green Knobs inclusions has been analyzed by SIMS and other techniques [Roden and Shimizu, 1993; Roden et al., 1990], and clinopyroxene in one of these rocks (N53-GN) was also analyzed here. Clinopyroxene in SUM-hosted spinel peridotite N53-GN has low abundances of light LREE, as do many of spinel peridotites from SUM diatremes [Roden and Shimizu, 1993], and the discrete grains have distinctively higher Ce and Nd abundances (Figure 6a). Compositions of all five discrete grains are unusually magnesian, and four of the five have fe (100*Fe/(Fe + Mg)) in the range 4.6 to 5.2, comparable to the least Fe-rich diopsides in garnet peridotite xenoliths from cratonic mantle: they are compared to those of similarly magnesian diopside (fe in the range 4.5 to 4.7) of three kimberlite-hosted xenoliths from Siberia [Ionov et al., 2010] and to diopside of a minette-hosted garnet peridotite [Roden and Shimizu, 2000] in Figure 6b. Abundances of REE and many other trace elements in diopside of these three cratonic xenoliths are similar to those in the discrete grains from Green Knobs and to those in diopside of one xenolith from The Thumb.
 Concentrations and ratios of several trace elements in the discrete diopside grains are unusual (Figure 7). Sr/Nd in clinopyroxene provides insight into possible metasomatic histories, and normalized values of Sr/Nd for two of the five grains are significantly higher than clinopyroxene in most of the comparison group, except for that in spinel peridotite xenoliths from other Colorado Plateau localities. In contrast, clinopyroxene of the minette-hosted peridotite inclusions analyzed by Roden and Shimizu  does not have unusually high Sr/Nd. The discrete diopside grains also are high in Li, as is diopside in SUM-hosted spinel peridotite N53-GN. The three magnesian diopside samples from the Siberian craton lack that Li enrichment (Figure 6b).
 Thermobarometric reconstructions of geotherms in the mantle wedge can constrain possible histories of low-angle Farallon subduction, because those geotherms must have changed during the subduction [Helmstaedt and Schulze, 1991]. One difficulty in using SUM-hosted peridotite inclusions to calculate such constraints is that pyroxenes in many of these rocks have complex exsolution textures and are not homogeneous. Pyroxene-based temperatures fall in the range 830°C to 1000°C for spinel peridotite inclusions from Green Knobs studied here, but Fe-Mg olivine-spinel equilibria for some of the same rocks yield temperatures near and below 700°C [Smith and Levy, 1976; Roden and Shimizu, 1993]. SUM-hosted eclogites, chlorite-bearing pyroxenites, and discrete garnet grains also record temperatures near and below 700°C [e.g., Smith et al., 2004; Helmstaedt and Schulze, 1979; Wang et al., 1999]. Compositions of olivine both in discrete grains and in rocks and of diopside in discrete grains provide new opportunities to clarify temperature histories of their sources.
3.1 Olivine Thermometry
 De Hoog et al.  documented that concentrations of Al, Cr, V, Sc, Ca, and Na in olivine depend primarily upon temperature, and they presented empirical thermometers for Al-in-olivine, Cr-in-olivine, and Ca-in-olivine in garnet peridotite. They found the most satisfactory results for Al-olivine and the least satisfactory for Ca-in-olivine. Although the thermometers were calibrated with data for garnet peridotites, they found results for spinel peridotites generally accurate but with more scatter. Application of their Al-in-olivine and Cr-in-olivine thermometers to the xenoliths from the Grand Canyon field and The Thumb yields temperatures that agree well with results calculated using pyroxene-based thermometry (Figure 8a). An additional empirical thermometer for olivine was calibrated based on a fit to temperature and ln(V), where V is ppm vanadium in olivine and temperatures are based on the Ca-in-orthopyroxene thermometer of Brey and Kohler  as corrected by Nimis and Grutter . The resulting V-in-olivine thermometer is: T (°C) = -5549.9/(ln(V) - 5.5976) -273. For temperatures below 1100°C, V-in-olivine and Al-in-olivine temperatures are well-correlated for kimberlite-hosted and basalt-hosted peridotites (Figure 8b). For the two garnet peridotite xenoliths from The Thumb, however, the V-in-olivine thermometer yields temperatures distinctly higher than the compared techniques. Perhaps a pressure correction would improve the accuracy of the V-in-olivine thermometer at high temperatures. Nonetheless, for temperatures below 1100°C, the V-in-olivine thermometer is useful for comparisons with the Al-in-olivine and Cr-in-olivine thermometers of De Hoog et al. , and these comparisons may provide insights into peridotite histories.
 Application of all three olivine thermometers to the SUM-hosted samples yields unusually cool mantle temperatures (Figures 8b and 9 and Table 3). Most are in the range 530°C to 650°C. Wan et al.  provided a thermometer based on the partitioning of Al between olivine and spinel: for the three Green Knobs peridotites for which spinel analyses are also available, the Wan et al.  thermometer yields results in the 530°C to 600°C range. Results of the four methods are in good agreement, although the V-in-olivine thermometer yields slightly higher temperatures for the discrete grains. Three of these peridotite inclusions contain chlorite inferred to be of mantle origin by Smith [1979, 2000], but one contains no hydrous minerals except for post-emplacement serpentine (Table 3).
|Descriptiond||GPa||T (°C)||T (°C)||T (°C)||T (°C)|
|N70-GN L A||Grain with sparse lamellae of diopside and titanian clinohumite||3||596||565||638|
|N70-GN L B||Grain||3||559||534||572|
|N70-GN L C||Grain||3||581||555||704|
|N70-GN L D||Grain||3||568||536||643|
|N70-GN L E||Grain||3||576||551||609|
|N17-GN||Spinel peridotite with chlorite, amphibole||2||533||491||509|
|N23-GN||Peridotite with chlorite, amphibole, and titanian clinohumite in pseudomorphs of garnet||3||555||520||597|
|N53-GN||Spinel peridotite with trace amphibole||2||558||539||585||549|
|N178-GN||Chlorite peridotite with Cr-rich spinel||3||588||592||550||596|
3.2 Thermobarometry of Discrete Diopside Grains in SUM
 Temperatures and pressures for the five diopside grains from Green Knobs (Table 2) and the one from Garnet Ridge [Skogby et al., 1990] fall in the ranges 590°C to 700°C and 2.4 to 3.0 GPa (Figure 10); corresponding depths are about 80 to 100 km. These P-T values and others for diopside grains with compositions like those in garnet peridotite (Figure 5) were calculated with the thermobarometer of Nimis and Taylor . Temperatures for 10 of 11 appropriate grains from the Moses Rock diatreme analyzed by McGetchin  fall in ranges 660°C to 830°C, and pressures of all but one are less than about 3.5 GPa. Grutter  found that pressures calculated with that thermobarometer are less precise than those derived from orthopyroxene-based barometry, and the scatter of the T-P points partly may record that low precision. The McGetchin  analysis of the one Moses Rock grain with a significantly higher pressure (4.5 GPa) may have been affected by undetected mineral intergrowths or an analytical problem, and 3.5 GPa and the corresponding depth of about 120 km are considered the maximum values recorded by the discrete grains.
 The pressures calculated for the diopside grains may not record source conditions at the time of SUM eruption for kinetic reasons. Pressures calculated with the thermobarometer of Nimis and Taylor  used here depend upon the distribution of Cr between garnet and clinopyroxene. Smith and Barron  measured steep gradients of Cr within about 100 µm of a mutual contact of diopside included in a discrete garnet grain from SUM, and they suggested that Cr would not equilibrate between garnet and diopside by diffusion alone during cooling below about 800°C. No compositional inhomogeneities were noted in the discrete diopside grains, but they are fragments of larger crystals. The homogeneity of the discrete grains from Green Knobs and the apparent absence of lamellae of low-Ca pyroxene are consistent with low-temperature recrystallization, but at unknown times before eruption. Hence, unless the discrete diopside grains equilibrated by recrystallization shortly before eruption, the temperatures below about 800°C and the corresponding pressures may record earlier histories, not conditions in the SUM source.
3.3 Thermobarometry of Garnet Peridotite Xenoliths in Minette
 The minettes of the NVF were emplaced in a time interval overlapping that of the SUM diatremes, and some contain garnet peridotite xenoliths. The most prolific source of these xenoliths has been the minette neck at The Thumb. Application of the thermobarometer of Nimis and Taylor  to the comprehensive analysis base of Ehrenberg  and the supplementary results of Smith et al.  yields an array of P-T points for The Thumb xenoliths, most in the ranges 1000°C to 1250°C and 3.6 to 4.4 GPa (Figure 10). The array is discordant to unperturbed mantle geotherms, consistent with the conclusion of Ehrenberg [1982a] that the xenoliths record perturbation of a geotherm by igneous intrusions accompanied by metasomatism.
 Thumb xenolith AO82 is of particular importance because of the well-defined interior-to-rim zonation of Ni in garnet, 26 to 61 ppm [Smith et al., 1991]: temperatures calculated from these Ni values are plotted as black squares in Figure 10. Two values are plotted for the garnet interior, 938°C and 850°C, based on the methods of Canil  and of Ryan et al. , respectively. The garnet rim Ni content of 61 ppm yields 1099°C with each method, in good agreement with other thermometers for the xenolith. For instance, clinopyroxene yields 1080°C, 3.8 GPa with the method of Nimis and Taylor , and clinopyroxene-orthopyroxene-garnet yields 1092°C, 3.9 GPa with the NG85-TA98 method recommended by Nimis and Grutter . The similar results are evidence that the compositions of the garnet rim, olivine, and pyroxene represent equilibrium. The agreement also supports the hypothesis that the Ni zonation in garnet records substantial heating, regardless of which Ni-in-garnet temperature difference from interior to rim is more correct -- 160°C or 250°C. The arrow in Figure 10 depicts the temperature increase.
 Garnet peridotite xenoliths from other minettes provide insight into the regional extent of the heating recorded by samples from The Thumb. Ehrenberg  analyzed minerals in one garnet peridotite from each of six other minette localities, but the compositions of clinopyroxene and garnet in two appear to be of altered material. The four apparently unaltered diopside compositions record pressures similar to those from The Thumb but temperatures similar to those calculated for interior garnet in Thumb xenolith AO82 (Figure 10). The four xenoliths are from widely separated localities (Figure 1b) ranging from Ship Rock, about 20 km northeast of The Thumb, to Black Rock Dike, about 95 km south of The Thumb. Hence, the relatively hot temperatures calculated for xenoliths from The Thumb were produced by a localized heating event that was not recorded by lithosphere at similar depths below these four other localities.
4.1 Implications of Calculated Temperatures and Pressures
 Olivine trace-element compositions are the best guide to temperatures in the source of the SUM diatremes. The Al, Cr, and V contents of olivine in discrete grains and peridotites in SUM and the derived temperatures are extraordinarily low for peridotite, and the most similar olivine compositions are in orogenic peridotite massifs (Figures 3, 4, 8). These temperatures are consistent with the presence of antigorite and chlorite of inferred mantle origin in some peridotite inclusions from Green Knobs [Smith, 2010]. These low temperatures are also consistent with the temperatures calculated for the five discrete diopside grains from Green Knobs based on the pyroxene thermobarometry of Nimis and Taylor ; four of the five diopside grains yield temperatures within the range 586 to 619°C, essentially the same as the range calculated for the five discrete olivine grains using the Al-in-olivine thermometer (568 to 596°C, Table 3). The agreement is evidence for the utility of these thermometry methods at mantle temperatures below those of most or all of the calibration data.
 No temperature gradients are recognizable in the mantle sampled by the Green Knobs diatreme, and peridotites and discrete olivine grains were erupted from sources at temperatures within the range 530°C to 650°C. The anhydrous spinel peridotite and chlorite-spinel peridotite must have crystallized at lower pressure than the protolith of the hydrated garnet peridotite, N23-GN, and at least some of the discrete olivine grains may be from garnet peridotite. However, the Al-in-olivine, Cr-in-olivine, and V-in-olivine temperatures for the SUM-hosted fragments are not obviously related to mineral assemblage or to inferred depth of source (Table 3). No evidence for a magmatic component has been recognized in the eruptive mix.
 Depths from which the SUM was erupted are less tightly constrained than temperatures, because diopside and garnet are less likely than olivine to record conditions close to the time of eruption. However, almost all diopside data are consistent with eruption from a source within the depth range 80 to 120 km. Most discrete pyrope grains record pressures in the range from about 1.8 to 3 GPa, corresponding to depths in the approximate range 70 to 100 km, and temperatures near or below 600°C [e.g., Wang et al., 1999; Griffin et al., 2004; Smith and Barron, 1991]. Source conditions within the ranges 500°C to 650°C and 80 to 120 km also are mostly consistent with calculated conditions for eclogite inclusions (Figure 10), and geochronology establishes that eclogite-facies assemblages of some of these rocks record conditions close to the time of eruption [Wendlandt et al., 1996; Usui et al., 2003; Smith et al., 2004]. Chlorite pyroxenites also are consistent with these source conditions [Helmstaedt and Schulze, 1979, 1991; Smith, 1995]. These consistencies establish that the depth range recorded by most diopside grains, 80 to 120 km, probably includes the source depths of the SUM eruptions.
 Temperatures recorded by peridotite fragments from all localities in the NVF except The Thumb are unusually cool for back arc Proterozoic mantle lithosphere. Inclusions in the SUM diatremes record conditions near those on the 40 mW/m2 geotherm of Carlson et al. , and the garnet peridotite xenoliths from the non-Thumb minettes scatter above and below the curve (Figure 10). The most comparable mantle xenolith temperatures are for kimberlite localities on stable cratons: most kimberlite-hosted xenoliths that record pressures below 4 GPa record temperatures within about 100°C of that geotherm [Grutter, 2009; Carlson et al., 2005]. At a pressure of 2 GPa, near the transition depth from spinel to garnet peridotite, the 40 mW/m2 geotherm is at about 550°C. In contrast, pyroxene thermometry yields temperatures in the range 800°C to 1100°C for spinel peridotite xenoliths from most wedge environments [Arai and Ishimaru, 2008], and mantle temperatures in many back arc environments are comparably hot for hundreds of km behind the arcs [Currie and Hyndman, 2006].
 The cool temperatures recorded by Plateau mantle fragments are anticipated consequences of low-angle subduction [Helmstaedt and Schulze, 1991], and they constrain possible histories of lithosphere-asthenosphere interactions. For instance, the trend of intrusions along the Colorado Mineral Belt (COMB) terminates at the Carrizo Mountains laccolith, intruded at about 70 Ma in what is now the margin of the NVF [Semken and McIntosh, 1997] (Figure 1). Jones et al.  have suggested that the intrusions of the Colorado Mineral Belt (COMB) formed from small-scale upwelling within the asthenosphere and that some asthenosphere may have remained above most of the Farallon slab during low-angle subduction. However, lithosphere cools only slowly by conduction. Spencer  presented numerical solutions for conductive cooling in models for low-angle subduction and subduction erosion below the Plateau, but a simpler calculation illustrates the problem. If the thermal diffusivity is a plausible 10-6 m2/s [Lachenbruch and Sass, 1977], conductive cooling would be negligible at a distance of 100 km above the slab contact for times shorter than 20 m.y., and substantial cooling due to low-angle subduction would require a much longer time. Hence, if the cold Plateau lithosphere at 25 Ma was an in-place residue of Proterozoic mantle, asthenosphere is unlikely to have been present at a depth of 150 km below the NVF for long after COMB formation at 70 Ma.
4.2 Complex Histories of Colorado Plateau Mantle Lithosphere
 Emplacement of forearc mantle into the Plateau lithosphere is consistent with the cool geotherm. Bird  suggested that all continental lithosphere below a detachment horizon in the lower crust was displaced by Farallon subduction from below the Basin and Range and the Colorado Plateau provinces and emplaced further to the east. However, geochronologic studies of peridotite inclusions establish that old continental mantle lithosphere remained below parts of the Basin and Range and Colorado Plateau at the time of NVF emplacement [e.g. Roden et al., 1990; Lee et al., 2001]. Nonetheless, displacement of parts of the mantle wedge during Farallon subduction has been documented. Evidence includes removal of mantle that must have accompanied emplacement of the “southern California schist” terrane [Saleeby, 2003]. More evidence in southern California is provided by spinel peridotite xenoliths, some of which appear to be fragments of old continental mantle and others to be from the Farallon slab [Luffi et al., 2009]. These areas lie to the southwest of the NVF (Figure 1a). Helmstaedt and Schulze  suggested that eclogites in the SUM diatremes might represent continental mantle eroded in the southwest and emplaced below the NVF, although they preferred the Helmstaedt and Doig  hypothesis that the eclogites are fragments of the slab. However, Smith  summarized evidence that the characteristic eclogites, the garnetites, some of the peridotites, and perhaps many of the discrete grains are fragments of forearc mantle transported in a serpentine-rich mélange. Transport of tectonically eroded cool forearc mantle and its emplacement below the Plateau is a possible cause of the unusually cool lithosphere.
4.2.1 Histories Consistent with SUM-hosted Discrete Grains
 Compositions and mineral associations provide additional tools to evaluate possible histories for the source of the SUM diatremes. Some discrete pyrope grains have compositions appropriate for garnet peridotite and also contain inclusions of chlorite [Hunter and Smith, 1981], strong evidence that those garnets grew from cool, wet peridotite protoliths such as those in forearc environments. Olivine compositions are consistent with a forearc environment, because they are more magnesian than in typical xenoliths from the southwestern United States and from most other non-craton regions, and similarly magnesian olivine is found in oceanic forearcs [Ishii et al., 1992] (Figure 2). Olivine grains record temperatures unusually low for mantle xenoliths, and the most similar concentrations of highly temperature-sensitive trace elements in olivine are found in orogenic peridotites (Figures 3, 4, 8). Diopside is relatively Li-rich (Figure 7). High Li is one signature of subduction metasomatism in forearc mantle [Savov et al., 2007] and one not present in otherwise similar diopside from cratonic xenoliths (Figure 6b). One comparable data set for Li is that of Paquin et al.  for the Alpe Arami peridotite; they concluded that diopside had been enriched in lithium carried into the mantle wedge by slab-released aqueous solutions. The discrete olivine grains have typical Li contents, not relatively high values as does diopside (Figure S1); comparable differences in relative Li enrichment have been attributed to disequilibrium and to effects of coupled substitutions on trace elements in olivine [Mallmann et al., 2009; Yakob et al., 2012]. Sr/Nd values for two of the five diopside grains are significantly higher than for clinopyroxene in most of the comparison group, except for that in spinel peridotite xenoliths from other Colorado Plateau localities (Figure 7a). High Sr/Nd may be of particular value as a recorder of subduction-related metasomatism [Lee, 2005], and the exceptionally high values in clinopyroxene in some xenoliths from the Grand Canyon field have been attributed to introduction of Sr from the Farallon slab [Alibert, 1994; Riter, 1999]. Finally, discrete grains of forsterite, diopside, and pyrope are unusually water-rich [Mosenfelder et al., 2006; Skogby et al., 1990; Wang et al., 1999]. These characteristics are best explained by erosion of forearc mantle wedge, transport of fragments in a serpentine-rich mélange along the slab-wedge interface, and incorporation into the Plateau lithospheric mantle. Processes in the Mariana forearc described by Savov et al.  may be analogous to some recorded by source rocks of the SUM diatremes.
4.2.2 Possible Source Histories of the Minette-hosted Garnet Peridotite Xenoliths
 The interpretation that some of the SUM-hosted mantle fragments represent forearc mantle emplaced during Farallon subduction has implications for the minette-hosted garnet peridotite, because the SUM-hosted fragments record lower pressures than almost all those garnet peridotite xenoliths (Figure 10). Smith  discussed two possible scenarios consistent with that evidence. One was that the SUM-hosted mantle fragments are from serpentine-rich mélange intruded up into the Plateau lithosphere through the mantle represented by the garnet peridotite xenoliths. The second was that the minette-hosted garnet peridotite xenoliths are from Proterozoic lithosphere of the Plateau that was displaced during low-angle subduction but then flowed back when the slab was removed. A third possibility is that the source of these minette-hosted garnet peridotites was a slice of forearc or arc mantle transported above the slab and emplaced below the serpentine-rich mélange. Examples of such underplated lithosphere have been described by Canil et al.  and by Snyder .
 Evidence for the history of the minette-hosted garnet peridotites is inconclusive, however, perhaps partly because many of the xenoliths were overprinted by metasomatism shortly before minette ascent [Ehrenberg, 1982a; Alibert, 1994]. Nonetheless, olivine and diopside in some of the xenoliths retain evidence of melt depletion similar to that in the discrete grains (Figures 2, 6b). Isotope ratios of garnet peridotites from The Thumb do not differentiate between origins as lithosphere initially formed below the Plateau or as lithosphere emplaced from the southwest during Farallon subduction. Values of 187Os/186Os are consistent with Proterozoic depletion near 1.6 Ga [Lee et al., 2001], an age near both that of initial formation of Plateau crust [Wendlandt et al., 1993] and that of a major magmatic episode in the Mojave Province to the southwest [Barth et al., 2009].
4.3 Implications for Interpretations of Seismic Data and for Minette Sources
 The minimum thickness of the lithosphere sampled by mantle xenoliths in the NVF is constrained by the maximum pressures they record, about 4.5 GPa (Figure 10). The 4.5 GPa pressure corresponds to a depth of about 150 km, and that depth defines a minimum thickness for the lithosphere sampled by the NVF minettes. Liu et al.  evaluated seismic data and concluded that the depth to the lithosphere-asthenosphere boundary below the Colorado Plateau ranges from about 80 to 120 km; if the seismic interpretation is correct, at least 30 km of lithosphere below the NVF must have been removed or converted in the last 25 m.y. Reid et al.  used trace-element data to place the minette source at greater than or equal to 90 km. The xenolith data provide a firmer constraint. Magmas of the mafic minettes of the NVF must have formed at a depth greater than about 150 km and at temperatures greater than those recorded by the xenoliths at The Thumb, so in excess of 1250°C; melt temperatures greater than 1200°C are consistent with experiments on melting relationships applicable to NVF minettes [Esperanca and Holloway, 1987].
 Seismic images of the mantle lithosphere of the central Plateau show several unusual features. The Moho is poorly defined, and it has been interpreted as perhaps gradational [Wilson et al., 2010], or perhaps duplicated due to delamination of a wedge of lower crust plus mantle [Levander et al., 2011]. Complications due to serpentinite diapirism into the uppermost mantle and lower crust should also be considered as an explanation, consistent with the suggestion of Ehrenberg and Griffin . Similarly, a shallow-dipping discontinuity at depths of 65 to 85 km beneath the Plateau was imaged by Wilson et al. , and they suggested that the discontinuity is relict from Proterozoic lithospheric formation. Instead, the discontinuity may record tectonic underplating of the serpentine-rich mélange during Farallon subduction.
 Compositions of olivine and diopside establish new constraints for the sources of the SUM diatremes. Concentrations of Al, Cr, Ca, V, and Na in SUM-hosted olivine are remarkably low, and comparable concentrations have been reported only for olivine in orogenic peridotite masses emplaced in the crust. Temperatures calculated by the Al-in-olivine and Cr-in-olivine thermometers of De Hoog et al.  and a new V-in-olivine thermometer are similar, almost all in the range 530°C to 650°C. Despite the variety of assemblages in peridotite inclusions at Green Knobs, olivine does not record systematically different temperatures within the peridotite population or between rocks and discrete grains. Olivine temperatures in rocks are not correlated with the presence or absence of hydrous minerals of mantle origin, and the calculated temperature for a peridotite with chlorite pseudomorphs of garnet is similar to those for spinel peridotites. The calculated temperatures are consistent with stability of antigorite and chlorite peridotite, and the results confirm the power of olivine thermometry for a variety of peridotite assemblages.
 Thermobarometry of discrete grains of diopside yields warmer temperatures that cluster in the range 590°C to 830°C; depths corresponding to calculated pressures cluster in the range from about 80 to 120 km. These diopside fragments may record earlier history, not conditions of the SUM source at the time of eruption. Nonetheless, evidence from a variety of fragments is consistent with eruption sources within the depth interval 80 to 120 km. The olivine temperatures are the best guide to conditions in the source of the Green Knobs diatreme, and no evidence was observed for thermal gradients in that source or for a magmatic component in the eruptive mix.
 The Colorado Plateau mantle fragments in the NVF may be samples of at least three sources with distinct histories. Typical spinel peridotite inclusions are probably from Proterozoic lithosphere of the Plateau [Roden et al., 1990]. Compositions of many of the SUM-hosted discrete grains of olivine and diopside are unusual for Proterozoic mantle. Major elements record depletion similar to that in some kimberlite-hosted cratonic xenoliths. Some discrete olivine grains are Nb-rich, as is olivine in kimberlite-hosted xenoliths. Discrete diopside grains have high Li contents similar to those in some diopside in orogenic peridotites attributed to metasomatism by slab-derived fluids, and some have high Sr/Nd. Compositions of these discrete grains are consistent with the suggestion by Smith  that they represent forearc mantle tectonically eroded by Farallon subduction and added to Colorado Plateau lithosphere in a serpentine-rich mélange, together with lawsonite eclogite and garnetite. At the time of SUM eruption, that mélange was at shallower depth than the source of garnet peridotite xenoliths in minettes of the Navajo Volcanic Field. The deeper source of those minette-hosted xenoliths may be a slice of forearc lithosphere underplated below the mélange source of the SUM-hosted discrete grains during low-angle Farallon subduction.
 The temperature and structure of the lithosphere to a depth of about 150 km and at about 25 Ma is established by SUM-hosted and minette-hosted xenoliths. That lithosphere was unusually cool for Proterozoic mantle, except for local volumes heated by magmatism precursory to minette eruption. If the lithosphere at 150-km depth was intact from the Proterozoic at 25 Ma, the low temperatures preclude contact with the asthenosphere at or near that depth during the Laramide orogeny. Alternatively, the source of garnet peridotite in minette may have been in a slice or slices of mantle lithosphere emplaced below the Plateau during Farallon subduction. Regardless, the temperatures and pressures recorded by mantle fragments preclude genesis of NVF minettes at depths shallower than 150 km. Seismic discontinuities in the mantle lithosphere of the central Colorado Plateau may represent features formed during Farallon subduction, not Proterozoic features.
 The LA-ICP-MS data were acquired with the invaluable assistance of Nathaniel R. Miller. The manuscript was improved by helpful comments on an early draft by Jaime D. Barnes and Michael F. Roden and on the submitted draft by reviewers William L. Griffin and Cin-Ty A. Lee. The research was supported by the Department of Geological Sciences and the Jackson School of Geosciences, both of The University of Texas at Austin.
All Supporting Information may be found in the online version of this article.
|Analysis_procedures.docxWord 2007 document , 132.1 KB||Analysis procedures.docx: Discusses analytical procedures in detail with references|
|Table_S1.docxWord 2007 document , 112.6 KB||Table S1.docx: compares analyses of secondary standards to reference values organized in columns below. The Mass column contains measured masses. BCR-2G and NIST614 are international reference standards used for trace element analyses. The dates of analysis are presented, together with the ratio (average measured composition)/(reference value) for each element.|
|Figure_S1.epsPS document, 393.5 KB||Figure S1.eps : Plots of olivine analyses and comparative data|
|Figure_S1_explanation.docxWord 2007 document , 102.5 KB||Figure S1 explanation.docx : Caption for Figure S1 and references for the comparison data plotted|
Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
1994), Peridotite xenoliths from western Grand Canyon and The Thumb: A probe into the subcontinental mantle of the Colorado Plateau, J. Geophys. Res., 99, 21605–21620.
2008), Insights into petrological characteristics of the lithosphere of mantle wedge beneath arcs through peridotite xenoliths: a review, J. Petrol., 49, 665–695.
2009), Assembling and disassembling California: A zircon and monazite geochronologic framework for Proterozoic crustal evolution in southern California, J. Geol., 117, 221–239.
1988), Formation of the Rocky Mountains, western United States: A continuum computer model, Science, 239, 1501–1507.
1990), Geothermobarometry in four-phase Lherzolites II. New thermobarometers, and practical assessment of existing thermobarometers, J. Petrol., 31, 1353–1378.
1999), The Ni-in-garnet geothermometer: calibration at natural abundances, Contrib. Mineral. Petrol., 136, 240–246.
2003), Lithospheric roots beneath western Laurentia: the geochemical signal in mantle garnets, Can. J. Earth Sci., 40, 1027–1051.
2005), Physical, chemical, and chronological characteristics of continental mantle, Rev. Geophys., 43, RG1001/2005, doi:10.1029/2004RG000156.
2004), Distribution of high field strength and rare earth elements in mantle and lower crustal xenoliths from the southwestern United States: The role of grain-boundary phases, Geochim. Cosmochim. Acta, 68, 3919–3942.
2006), The thermal structure of subduction zone back arcs, J. Geophys. Res., 111, B08404, doi:10.1029/2005JB004024.
2010), Trace-element geochemistry of mantle olivine and application to mantle petrogenesis and geothermobarometry, Chem. Geol., 270, 196–215.
1999), Cryptic metasomatism in the upper mantle beneath southeastern Austria: a laser ablation microprobe-ICP-MS study, Mineral. Petrol., 67, 143–161.
1978), Petrology of potassic volcanic rocks and ultramafic xenoliths from the Navajo volcanic field, New Mexico and Arizona, Ph.D. dissertation, Univ. Calif. Los Angeles.
1982a), Petrogenesis of garnet lherzolite and megacrystalline nodules from The Thumb, Navajo volcanic field, J. Petrol., 23, 507–547.
1982b), Rare earth geochemistry of garnet lherzolite and megacrystalline nodules from minette of the Colorado Plateau province, Earth Planet. Sci. Lett., 57, 191–210.
1979), Garnet granulite and associated xenoliths in minette and serpentinite diatremes of the Colorado Plateau, Geology, 7, 483–487.
1987), On the origin of some mica-lamprophyres: experimental evidence from a mafic minette, Contrib. Mineral. Petrol., 95, 207–216.
1968), Uranium emplacement at Garnet Ridge, Arizona, Econ. Geol., 63, 859–876.
2000), Trace element residence and partitioning in mantle xenoliths metasomatized by highly alkaline, silicate- and carbonate-rich melts (Kerguelen Islands, Indian Ocean), J. Petrol., 41, 477–509.
2003), Garnet lherzolites from the Kaapvaal craton (South Africa): Trace element evidence for a metasomatic history, J. Petrol., 44, 629–657.
2004), Lithosphere mapping beneath the North American plate, Lithos, 77, 873–892.
2009), Pyroxene xenocryst geotherms: Techniques and applications, Lithos, 112S, 1167–1178.
1975), Eclogite nodules from kimberlite pipes of the Colorado Plateau -- Samples of subducted Franciscan-type oceanic lithosphere, Phys. Chem. Earth, 9, 95–111.
1979), Garnet clinopyroxenite - chlorite eclogite transition in a xenolith from Moses Rock: further evidence for metamorphosed ophiolites under the Colorado Plateau, in The Mantle Sample: Inclusions in Kimberlite and Other Volcanics, edited by F. R. Boyd and
, pp. 357–365, AGU, Washington, D. C.
1991), Early to mid-Tertiary inverted metamorphic gradient under the Colorado Plateau: evidence from eclogite xenoliths in ultramafic microbreccias, Navajo volcanic field, J. Geophys. Res., 96, 13225–13235.
1980), Fertile and barren Al-Cr-Spinel harzburgites from the upper mantle: Ion and electron probe analyses of trace elements in olivine and orthopyroxene: Relation to lherzolites, Earth Planet. Sci. Lett., 50, 41–58.
1981), Garnet peridotite from Colorado Plateau ultramafic diatremes: hydrates, carbonates, and comparative geothermometry, Contrib. Mineral. Petrol., 76, 312–320.
2010), Composition of the lithospheric mantle in the Siberian Craton: New constraints from fresh peridotites in the Udachnaya - East kimberlite, J. Petrol., 51, 2177–2210, doi:10.1093/petrology/egq053.
1992), Petrological studies of peridotites from diapiric serpentinite seamounts in the Izu-Ogasawara-Mariana forearc, Leg 125, in Proceedings of the Ocean Drilling Program, Scientific Results, vol. 125, edited by P. Fryer et al., pp. 445–485, Texas A&M University, College Station, Texas.
2011), Hydrodynamic mechanism for the Laramide orogeny, Geosphere, 7, 183–201, doi:10.1130/GES00575.1.
2007), Depleted and enriched mantle processes under the Rio Grande rift: spinel peridotite xenoliths, Contrib. Mineral. Petrol., 154, 135–151, doi:10.1007/s00410-007-0183-y.
1997), Patterns in the hydrogen and trace element compositions of mantle olivines, Phys. Chem. Miner., 24, 385–395.
1977), Heat flow in the United States and the thermal regime of the crust, in The Earth's crust: Geophysical Monograph 20, edited by J. G. Heacock, pp. 626–675, AGU, Washington, DC.
2001), Preservation of ancient and fertile lithospheric mantle beneath the southwestern United States, Nature, 411, 69–73.
2005), Trace-element evidence for hydrous metasomatism at the base of the North American lithosphere and possible association with Laramide low angle subduction, J. Geol., 113, 673–685.
2000), A global geochemical database structure for rocks, Geochem. Geophys. Geosyst., 1, doi:10.1029/1999GC000026.
2011), Continuing Colorado plateau uplift by delamination-style convective lithospheric downwelling, Nature, 472, 461–465, doi:10.1038/nature10001.
2008), Water contents in mantle xenoliths from the Colorado Plateau and vicinity: Implications for the mantle rheology and hydration-induced thinning of continental lithosphere, J. Geophys. Res., 113, B09210, doi:10.1029/2007JB005540.
2011), Imaging crustal and upper mantle structure beneath te Colorado Plateau using finite frequency Rayleigh wave tomography, Geochem. Geophys. Geosyst., 12, Q07001, doi:10.1029/2011GC003611.
2009), Lithospheric mantle duplex beneath the central Mojave Desert revealed by xenoliths from Dish Hill, California, J. Geophys. Res., 114, B03202, doi:10.1029/2008JB005906.
2009), Fluid/mineral interaction in UHP garnet peridotite, Lithos, 107, 38–52, doi:10.1016/j.lithos.2008.07.006.
2009), Heterogeneous distribution of phosphorus in olivine from otherwise well-equilibrated spinel peridotite xenoliths and its implications for the mantle geochemistry of lithium, Contrib. Mineral. Petrol., 158, 485–504, doi:10.1007/s00410-009-0393-6.
1968), The Moses Rock dike: geology, petrology, and mode of emplacement of a kimberlite-bearing breccia dike, San Juan County, Utah, Ph. D. dissertation, Calif. Inst. Technology, Pasadena, California.
1970), Compositional relations in minerals from kimberlite and related rocks in the Moses Rock Dike, San Juan County, Utah, Am. Mineral., 55, 1738–1771.
1976), Single-pyroxene geothermometry and geobarometry, Am. Mineral., 61, 603–615.
2006), Hydrogen incorporation in natural mantle olivines, in Earth's Deep Water Cycle, Geophysical Monograph Series, vol. 168, edited by S. D. Jacobsen and S. van der Lee, pp. 45–56, AGU, Washington, DC.
2010), Internally consistent geothermometers for garnet peridotites and pyroxenites, Contrib. Mineral. Petrol., 159, 411–427, doi:10.1007/s00410-009-0455-9.
2000), Single clinopyroxene thermobarometry for garnet peridotites. Part 1. Calibration and testing of a Cr-in-Cpx barometer and an enstatite-in-Cpx thermometer, Contrib. Mineral. Petrol., 130, 541–554.
1998), Melting and metasomatism in the continental lithosphere: laser ablation ICPMS analysis of minerals in spinel lherzolites from eastern Australia, Contrib. Mineral. Petrol., 130, 240–255.
2004), Cosmochemical estimates of mantle composition, in The Mantle and Core, edited by R. W. Carlson, vol. 2, Treatise on Geochemistry, edited by H. D. Holland and K. K. Turekian, pp. 1–38, Elsevier-Pergamon, Oxford.
2004), Li-Be-B systematics in the ultrahigh-pressure garnet peridotite from Alpe Arami (Central Swiss Alps): implications for slab-to-mantle wedge transfer, Earth Planet. Sci. Lett., 218, 507–519.
2011), Iolite: Freeware for the visualization and processing of mass spectrometric data, J. Anal. At. Spectrom., 26, 2508–2518.
2012), Melting under the Colorado Plateau, USA, Geology, 40, 387–390.
1999), Geochemical and tectonic evolution of the Colorado Plateau mantle lithosphere: Evidence from Grand Canyon mantle xenoliths, PhD dissertation, U. Texas Austin, Austin, Texas.
1981), Origin of coexisting minette and ultramafic breccia, Navajo Volcanic Field, Contrib. Mineral. Petrol., 77, 195–206.
1993), Ion microprobe analyses bearing on the composition of the upper mantle beneath the Basin and Range and Colorado Plateau provinces, J. Geophys. Res., 98, 14091–14108.
2000), Trace element abundances in mantle-derived minerals which bear on compositional complexities in the lithosphere of the Colorado Plateau, Chem. Geol., 165, 283–305.
1979), Age and extent of potassic volcanism on the Colorado Plateau, Contrib. Mineral. Petrol., 43, 279–284.
1990), Chemical constraints on lithosphere composition and evolution beneath the Colorado Plateau, J. Geophys. Res., 95, 2811–2831.
1996), Garnet geotherms: Pressure-temperature data from Cr-pyrope garnet xenocrysts in volcanic rocks, J. Geophys. Res., 101, 5611–5626.
2003), Segmentation of the Laramide slab -- evidence from the southern Sierra Nevada region, Geol. Soc. Am. Bull., 115, 655–668.
2007), Shallow slab fluid release across and along the Mariana arc-basin system: Insights from geochemistry of serpentinized peridotites from the Mariana fore arc. J. Geophys. Res., 112, B09205, doi:10.1029/2006JB004749.
2006), Melt- versus fluid-induced metasomatism in spinel to garnet wedge peridotites (Ulten Zone, eastern Italian Alps): clues from trace element and Li abundances, Contrib. Mineral. Petrol., 151, 372–394.
2000), The distribution of lithium in peridotitic and pyroxenitic mantle lithologies - an indicator of magmatic and metasomatic processes, Chem. Geol., 166, 47–64.
1997), 40Ar/39Ar age determinations for the Carrizo Mountains laccolith, Navajo Nation, Arizona, New Mexico Geol. Soc. Guidebook, 48th Field Conference, pp. 75–80.
1990), Hydroxide in pyroxene: Variations in the natural environment, Am. Mineral., 75, 764–774.
1979), Hydrous minerals and carbonates in peridotite inclusions from the Green Knobs and Buell Park kimberlitic diatremes on the Colorado Plateau, in The Mantle Sample: Inclusions in Kimberlites and Other Volcanics, edited by F. R. Boyd and H. O. A. Meyer, pp. 345–356, AGU, Washington, DC.
1995), Chlorite-rich ultramafic reaction zones in Colorado Plateau xenoliths: recorders of sub-Moho hydration, Contrib. Mineral. Petrol., 121, 185–200.
2000), Insights into the evolution of the uppermost continental mantle from xenolith localities on and near the Colorado Plateau and regional comparisons, J. Geophys. Res., 105, 16,769–16,781.
2010), Antigorite peridotite, metaserpentinite, and other inclusions within diatremes on the Colorado Plateau, SW USA: Implications for the mantle wedge during low-angle subduction, J. Petrol., 51, 1355–1379, doi:10.1093/petrology/egq022.
1991), Pyroxene-garnet equilibration during cooling in the mantle, Am. Mineral., 76, 1950–1963.
1976), Petrology of Green Knobs diatreme, New Mexico, and implications for the mantle below the Colorado Plateau, Earth Planet. Sci. Lett., 19, 107–125.
1997), Genesis and evolution of low-Al orthopyroxene in spinel peridotite xenoliths, Grand Canyon field, Arizona, USA, Contrib. Mineral. Petrol., 127, 391–404.
1991), Trace-element zonation in garnets from The Thumb: heating and melt infiltration below the Colorado Plateau, Contrib. Mineral. Petrol., 107, 60–79.
2004), Evolution of Navajo eclogites and hydration of the mantle wedge below the Colorado Plateau, southwestern United States, Geochem. Geophys. Geosyst., 5, doi:10.1029/2003GC000675.
2008), Stacked uppermost mantle layers within the Slave craton of NW Canada as defined by anisotropic seismic discontinuities, Tectonics, 27, TC4006, doi:10.1029/2007TC002132.
1994), A numerical assessment of slab strength during high- and low-angle subduction and implications for Laramide orogenesis, J. Geophys. Res., 99, 9227–9236.
1998), An experimental test of some geothermometers and geobarometer formulations for upper mantle peridotites with application to the thermobarometry of fertile lherzolite and garnet websterite, N. Jb. Miner. Abh., 172, 381–408.
2003), Fate of the subducted Farallon plate inferred from eclogite xenoliths in the Colorado Plateau, Geology, 31, 589–592.
2007), Trace element fractionation in deep subduction zones inferred from a lawsonite-eclogite xenolith from the Colorado Plateau, Chem. Geol., 239, 336–351.
2008), Experimental calibration of aluminum partitioning between olivine and spinel as a geothermometers, Am. Mineral., 93, 1142–1147.
1999), Mineral inclusions in pyrope crystals from Garnet Ridge, Arizona, USA: implications for processes in the upper mantle. Contrib. Mineral. Petrol., 135, 164–178.
1993), Nd and Sr isotope chronostratigraphy of Colorado Plateau lithosphere: implications for magmatic and tectonic underplating of the continental crust, Earth Planet. Sci. Lett., 116, 23–43.
1996), Thermal history of Colorado Plateau lithosphere from Sm-Nd mineral geochronology of xenoliths, Geol. Soc. Am. Bull., 108, 757–767.
2005), Imaging the seismic structure of the crust and upper mantle beneath the Great Plains, Rio Grande Rift, and Colorado Plateau using receiver functions, J. Geophys. Res., 110, B05306, doi:10.1029/2004JB003492.
2010), High-resolution receiver function imaging reveals Colorado Plateau lithospheric architecture and mantle-supported topography, Geophys. Res. Lett., L20313, doi:10.1029/2010GL044799.
2005), The effect of temperature on the equilibrium distribution of trace elements between clinopyroxene, orthopyroxene, olivine and spinel in upper mantle peridotite, Chem. Geol., 221, 65–101, doi:10.1016/j.chemgeo.2005.04.005.
2012), Lithium partitioning between olivine and diopside at upper mantle conditions: An experimental study, Earth Planet. Sci. Lett., 329–330, 11–21, doi:10.1016/j.epsl.2012.01.035.
2010), Geochemical and Sr-Nd isotopic characteristics and pressure-temperature estimates of mantle xenoliths from the French Massif Central: evidence from melting and multiple metasomatism by silicate-rich carbonatite and asthenospheric melts, Geol. Soc. London Spec. Pub., 337, 153–175.
2009), Fluid-metasomatized mantle beneath the Ouachita belt of southern Laurentia: Fate of lithospheric mantle in a continental orogenic belt, Lithosphere, 1, 370–383.