Th-Pb dating of monazite in xenoliths of low-temperature metamorphic eclogite facies rocks from diatremes of the Navajo volcanic field in the center of the Colorado Plateau (southwest United States) yields ages of ca. 28 Ma. Because monazite is not a primary phase in basic igneous or metamorphic rocks, but introduced during metasomatism, we suggest that the fluid responsible for monazite growth was derived from prograde metamorphic dehydration reactions in serpentinites and related rocks in the subducted Farallon plate. These fluids hydrated the overlying sub-plateau lithospheric mantle, consuming garnet (thus mobilizing rare earth elements) and lowering mantle density and increasing volume, contributing to the uplift of the Colorado Plateau in early Oligocene time.
The Colorado Plateau is a vast area in the southwestern United States of relatively undeformed Phanerozoic crust surrounded by regions of intense deformation and great topographic relief. Despite extensive study, the reasons for its uplift and the timing have not been fully resolved. Mechanisms proposed for the uplift include upwelling of hot asthenosphere, warming of the lithosphere, crustal thickening, removal of the lithospheric mantle, magmatic activity, and hydration of the lithosphere (e.g., Flowers, 2010, and references therein). Until now, the only certain constraint on the timing of uplift is that it must be post–Late Cretaceous, following deposition of upper Cretaceous marine sedimentary rocks.
Evidence supporting a role for hydration of the lithospheric mantle as a cause of uplift exists in mantle-derived ultramafic xenoliths from diatremes of serpentinized ultramafic microbreccia in the Navajo volcanic field near the center of the plateau (Fig. 1). Most of the peridotite and pyroxenite xenoliths contain hydrous minerals (e.g., chlorite, antigorite) that formed in the upper mantle, but are very unusual in suites of ultramafic xenoliths from elsewhere (e.g., McGetchin and Silver, 1970; Helmstaedt and Schulze, 1979, 1988; Smith, 1979, 2010). The timing of hydration, however, has not been well constrained.
Low-temperature, high-pressure metamorphic eclogites are also prominent members of the mantle-derived xenolith assemblage in these diatremes (e.g., Helmstaedt and Doig, 1975). We have obtained crystallization ages for monazite in several eclogite facies xenoliths from the Moses Rock diatreme in southeastern Utah. Because monazite is extremely rare in mafic rocks, and thus unexpected in the eclogite suite, we suggest that it was introduced metasomatically by the same fluids that caused the associated hydration of the sub-plateau lithospheric mantle in what we refer to as the “Great Hydration” event. The age of the monazite therefore should place constraints on the timing of the hydration of the upper mantle, and thus the uplift of the Colorado Plateau.
Monazite was identified as isolated anhedral to euhedral grains (50–500 μm) in three xenoliths from Moses Rock. One is a phengite-bearing eclogite. The other two samples are composite xenoliths consisting of a small region of essentially bimineralic eclogite in contact with Cr-omphacitite (monazite occurs in the latter assemblage).
The major and minor element composition of monazite, garnet, clinopyroxene, and phengite (see Tables DR1 and DR2 in the GSA Data Repository1) were determined using a JEOL JXA-8230 electron microprobe with wavelength dispersive methods and a combination of natural and synthetic standards at Queen’s University (Kingston, Canada). Methods and detailed results of the laser ablation–inductively coupled plasma–mass spectrometry dating of monazite are described in Table DR3; the results of X-ray mapping of the monazite are in Figure DR1.
RESULTS OF Th-Pb DATING OF MONAZITE
Results are summarized in Figure 2 and Table DR3, where averages of multiple analyses on the same spot (in the case of pulsed laser ablation) and multiple spots on the same grain (in the case of continuous ablation) are summarized. Pulsed and higher oxide analyses appear to bias ages upward. The data acquired using time-resolved analyses tuned to reduce oxide formation to a minimum appear to give the most precise and consistent ages, and are shown as filled symbols in Figure 2. The following mean ages were determined for the three samples under these conditions: sample 12-83-38, 29.5 ± 0.6 Ma (mean square of weighted deviation [MSWD], of 2.3 on 11 data); sample 12-83-33, 29.1 ± 0.4 (MSWD of 2.3 on 20 data); sample 12-97-05, 28.1 ± 0.2 (MSWD of 1.6 on 24 data).
There is no obvious difference between ages of cores and rims within single grains. Low-oxide analyses show that high-Th mantles in sample 12-83-38 give slightly older ages than lower Th cores and rims. The age of monazite in sample 12-97-05 appears to be resolvably younger than those in the other two sections by ∼1 m.y., but the overall degree of bias under different conditions suggests that the ages of all monazite may be the same. The weighted average of low-oxide analyses from the three sections is 28.4 ± 1.5 Ma.
Low-temperature lawsonite- and phengite-bearing eclogite xenoliths from the Navajo diatremes were interpreted in Helmstaedt and Doig (1975) as fragments of altered oceanic crust of the subducted Farallon plate, similar in origin to Franciscan low-temperature eclogites. In this model, these eclogitized metabasalts were transported in a flat subduction zone under the Colorado Plateau (Helmstaedt and Doig, 1975) overridden by the westward-moving North American plate following cessation of igneous activity in the Sierra Nevada Batholith at ca. 80 Ma.
U-Pb dating of zircon in the eclogites has yielded conflicting results. Usui et al. (2003) analyzed zircons from Garnet Ridge and Moses Rock in situ by secondary ion mass spectrometry and determined ages from 81 to 33 Ma; they concluded that the age data supported the subducted Farallon plate origin for the eclogites. Smith et al. (2004) used isotope dilution–thermal ionization mass spectrometry to analyze zircon fractions separated from each of three eclogites (from Moses Rock and Mule’s Ear). Each rock yielded fractions with analyses that plot on or very near concordia in the period from ca. 70 to 35 Ma, but each rock also yielded a discordant fraction with a 207Pb/206Pb age older than 1000 Ma. Smith (2010) interpreted these data as indicating that Proterozoic oceanic metabasalts were scraped from the overlying mantle wedge near the trench and dragged along by the flat-subducting Farallon plate, with renewed zircon growth continuing until the eclogites were erupted as xenoliths in the diatremes ca. 30 Ma (Helmstaedt and Doig, 1975). Both models, however, invoke a flat Farallon slab carrying oceanic metabasalts to the center of the Colorado Plateau by Paleogene time.
Other than containing monazite, the phengite-eclogite in our study resembles other low-temperature eclogite xenoliths from these pipes with respect to texture and mineral compositions. The other two monazite-bearing samples are unusual in that the monazite grains are in Cr-omphacitite (Schulze et al., 2014) portions of composite xenoliths with eclogite domains. As in the monazite-bearing eclogite, the eclogite portions of these two composite xenoliths are similar to the other eclogites from this suite. Equilibration conditions of the eclogite were determined using garnet-clinopyroxene-phengite equilibria (Krogh Ravna and Terry, 2004). A clinopyroxene-phengite pair enclosed in garnet yields 590 °C, 37.7 kbar, and the same assemblage outside the garnet yields 660 °C, 45.5 kbar, consistent with prograde metamorphic evolution of the eclogite. The eclogite domains in the two monazite-bearing Cr-omphacitites lack phengite but yield equilibration temperatures of 640–645 °C at an assumed pressure of 45 kbar.
Monazite has a very high U-Th-Pb closure temperature (>800 °C; Cherniak et al., 2004; Cherniak and Pyle, 2008) and original U-Pb and Th-U ages can be preserved through cycles of metamorphism. Primary crystallization ages have been retained even through granulite and ultrahigh-pressure metamorphic conditions (e.g., Williams et al., 1999), and later generations of monazite have overgrown existing cores without significant diffusive exchange. Although the monazite grains have complex zonation patterns suggestive of numerous episodes of growth and dissolution (Fig. DR1), all of the Th-Pb ages are identical within error. The episodic growth and dissolution of the monazite is suggestive of wide variation in the composition of the fluids responsible for these textures. The low equilibration temperatures estimated for the eclogites suggest that the ages determined for the monazite in this study are the ages of formation of the grains and not due to a late overprint or resetting. What is the significance of such young ages in a suite of eclogites that apparently grew zircons over tens of millions of years of eclogite facies metamorphism?
Although monazite is a common accessory mineral in granitoids and low-Ca metapelites (e.g., Spear and Pyle, 2002), it is very rare in mafic and ultramafic rocks (e.g., Williams et al., 2007). In such unusual occurrence, its presence is attributed to introduction by alkaline magmas (e.g., Rudnick et al., 1993) or hydrothermal fluids (e.g., Gorton and Schandal, 1995). Because monazite is unexpected in our suite of eclogites and related ultramafic rocks, we suggest that it was introduced by the metasomatic fluids that were also responsible for the hydration in the other ultramafic xenoliths in the Navajo diatremes. This is consistent with the presence of hydrous mineral assemblages overprinting high-grade Precambrian metamorphic assemblages in xenoliths of crustal rocks from the Navajo diatremes (e.g., Broadhurst, 1986), one of which yielded many Th-Pb ages in monazite in the range 91–58 Ma (Butcher et al., 2013). Equilibration temperatures of Navajo eclogite xenoliths (and hydrated peridotite xenoliths; Smith, 2010) are too low to be consistent with magmatic infiltration, despite the fact that the elevated Th contents of the monazite grains are more similar to magmatic than hydrothermal monazite, as documented in crustal occurrences (Gorton and Schandal, 1995).
GREAT HYDRATION EVENT
Many studies of the ultramafic xenoliths from the serpentinized ultramafic microbreccia diatremes in the Navajo volcanic field have mentioned or discussed the presence and significance of hydrous minerals such as antigorite, chlorite, and pargasite that are unusual, or completely absent, from other suites of mantle xenoliths (e.g., McGetchin and Silver, 1970; Helmstaedt and Schulze, 1979, 1988; Smith, 1979, 1995, 2010). Although these and other studies have dealt with aspects of hydration of the sub-plateau lithospheric mantle, nowhere in the literature has the sheer magnitude of this hydration been emphasized. In the southern diatreme cluster most peridotite and pyroxenite xenoliths have hydrous minerals that formed in the mantle, from trace amounts to extensive replacement of the primary minerals (e.g., Smith, 1979, 2010). In the population of ultramafic xenoliths from the northern cluster of Navajo diatremes mantle hydration effects are so severe that peridotites containing fresh olivine are extremely rare. Almost all (former) peridotite xenoliths are completely altered to schists dominated by antigorite, amphibole, and chlorite. Even xenocrysts of olivine are uncommon in the northern diatremes. Cr-pyrope xenocrysts are common in all Navajo diatremes, however, probably due to liberation of incompletely chloritized relict garnet from severely hydrated garnet peridotite xenoliths during violent eruption of the diatremes. The sub-plateau lithospheric mantle, as represented by xenoliths from these two Navajo diatreme clusters, thus underwent massive hydration in what we refer to as the Great Hydration event.
Some of the xenoliths of hydrated peridotite and pyroxenite contain clusters of chlorite-dominated assemblages in which partially altered remnants of primary pyrope garnet remain; these are the products of garnet replacement (Helmstaedt and Schulze, 1979; Smith, 1995). One source of rare earth elements (REEs) contributing to monazite growth in crustal metapelites is suggested to be REEs mobilized during garnet breakdown (e.g., Spear and Pyle, 2002); therefore, the garnet pseudomorphs in the xenoliths may similarly be viewed as a source of REEs for monazite formation in the plateau xenoliths. Although garnet is typically enriched in heavy REEs, these elements would be preferentially incorporated into amphibole in hydrous systems (Tiepolo et al., 2000), leaving the light REEs available to form monazite from the hydrous fluids. The monazite crystallization ages determined by us are thus interpreted as dating the Great Hydration event under the plateau in Oligocene time.
Some xenoliths from the Navajo volcanic field have been shown to be the products of prograde metamorphic dehydration reactions in the Farallon plate, suggesting that slab dehydration was the source of the fluids that altered the overlying mantle lithosphere beneath the Colorado Plateau. For example, Smith (2010) showed a Cr-magnetite dunite xenolith to be the product of dehydration of serpentinite and suggested that the fluids from this dehydration process hydrated the overlying subcontinental mantle, supporting the model of Humphreys et al. (2003) for hydration as the reason for plateau uplift. In Schulze et al. (2014) eskolaite was documented as a product of dehydration of guyanaite in a Cr-omphacitite xenolith, strengthening the case for a dehydrating Farallon slab as a source for fluids to hydrate the sub-plateau lithospheric mantle. The overlap in equilibration conditions between the monazite-bearing eclogites and the dehydrated serpentinites and hydrated peridotites documented by Smith (2010), and shown in Figure 3, is consistent with the hydration event as contributing to the introduction of monazite into the eclogites.
APPLICATION TO MODELS OF COLORADO PLATEAU UPLIFT
Numerous models have been presented to account for the causes and timing of uplift of the Colorado Plateau, but all models must include the passage of the flat slab of the Farallon plate beneath the plateau following cessation of Sierra Nevada arc magmatism in the Late Cretaceous. The Navajo xenoliths include samples interpreted to be from this slab, as well as from the overlying lithospheric mantle, and thus contain key geologic information relating to the Cenozoic history of the Colorado Plateau.
In the dynamic topography model of Liu and Gurnis (2010) the passage of the Farallon plate under the plateau initially caused subsidence in the region, followed by significant uplift (∼1.2 km) from the Cretaceous to the Eocene, in agreement with other studies (e.g., Flowers et al., 2008). An additional uplift of ∼700 m in the Oligocene was attributed to mantle upwelling (also see Reiter, 2014). The 28 Ma age for monazite growth in our samples is in agreement with the Oligocene uplift of the Liu and Gurnis (2010) and Reiter (2014) models. We argue that these ages represent the part of the Great Hydration event in which the peridotite in the mantle lithosphere of the central plateau was partially hydrated to antigorite, chlorite, and amphibole. The mantle upwelling of Liu and Gurnis (2010) would correspond to the expansion and rise of the hydrated mantle due to volume increase and density decrease, the timing of which we have documented.
Humphreys et al. (2003) proposed that hydration raised the Colorado Plateau throughout Laramide time (ca. 70–45 Ma). Although metamorphic growth of zircon was taking place in the eclogites carried by the Farallon slab during that time, our data suggest that monazite growth only occurred just prior to the eruption of the diatremes, which carried the mantle xenoliths to the surface. The slight age spread in our data suggests that hydration was a protracted process lasting ∼1–2 m.y. preceding eruption and also may have been a causal factor in diatreme formation. Hydration and growth of monazite may have continued in the mantle following diatreme eruption, but the xenolith record of geologic events ends at the time of eruption, and our data cannot be used to evaluate proposals that the timing of the uplift was much more recent, such as the suggestion by Sahagian et al. (2002) that most of the uplift of the Colorado Plateau occurred in the past 5 m.y.
Estimates of the magnitude of hydration (serpentinization) required for these two models (Liu and Gurnis, 2010; Sahagian et al., 2002) have been made, assuming a 40-km-thick forsteritic lithospheric mantle serpentinized to antigorite, with a 35% volume increase. The 700 m uplift could be accounted for by 5% serpentinization, a value probably similar to that observed in the peridotite xenoliths from the southern diatreme cluster. To account for the entire uplift of the plateau, ∼1.8 km since Cretaceous marine sediment deposition, the same mantle would require ∼13% serpentinization. This value is between the estimates for the degree of hydration for the two diatreme clusters, although closer to that of the southern cluster xenolith population. These values are only approximations, and would change with assumptions about lithosphere composition and thickness, but are considered to be a reasonable estimate. For our ca. 28 Ma age to be compatible with the model of Sahagian et al. (2002), however, the monazite ages would have to represent the onset of hydration that continued past the Oligocene, to the present time, although the xenolith record ends at the time of eruption.
The 232Th-208Pb system is an effective geochronometer for dating relatively young monazite in thin section in situations in which U-Pb dating is less effective. This provides a potentially powerful tool for determining precise ages for hydrothermal events that involve REE mobility, including mineralization and metamorphism, such as described here.
Using this method, we have determined a crystallization age of ca. 28 Ma for monazite in eclogite suite xenoliths with origins interpreted to be related to the flat subduction of the Farallon plate. The likely metasomatic origin of this monazite provides a strong link to the Great Hydration event in the sub-plateau lithospheric mantle that is well represented in the xenolith assemblages of the Navajo volcanic field. We suggest that the decrease in density and the increase in volume from this hydration played important roles in at least Oligocene, and possibly later, stages of plateau uplift.
We thank C. Bray, G. Kretchman, J. Adwent, and A. Dias for technical support, and D. Smith, W.G. Ernst, M.L. Williams, and R.L Rudnick for reviews. We thank the Natural Sciences and Engineering Research Council of Canada and the Office of the Vice President of Research at the University of Toronto, Mississauga, for financial support.