Zuni-Bandera volcanism, Rio Grande, USA – melt formation in garnet- and spinel- facies mantle straddling the asthenosphere-lithosphere boundary.
 
*Timothy J. Peters, Martin Menzies, Matthew Thirlwall, and Philip R. Kyle
The following text is published in Lithos (2007): Copyright Elsevier (2007)

The Zuni-Bandera Volcanic Field:

Introduction:

-Located in west central New Mexico (Fig. 1), the Zuni-Bandera Volcanic Field (ZBVF) is a sequence of mixed tholeiite and alkali basalt lava flows, spatter ramparts, scoria cones, small shields, maars, and collapse pits erupted through Proterozoic continental crust over the last 1.5 Ma (Anders et al., 1981; Maxwell, 1986; Baldridge et al., 1991; Menzies et al., 1991). 

-The ZBVF is located in the transitional zone between the actively extending terrains of the Rio Grande rift and the Basin and Range province, and the relatively stable southern edge of the Colorado Plateau (Fig. 1). Across this region the asthenosphere-lithosphere boundary varies significantly in depth from 45-55 km beneath the Rio Grande rift, to 120-150 km beneath the Colorado Plateau (West et al., 2004).
 
-Over the last 4.5 Ma, volcanic activity associated with the Rio Grande rift has shifted, to some degree, from the rift to the Jemez Lineament (Fig. 1), an 800 km NE-trending crustal discontinuity that cuts across the rift. Seismic tomography has identified regions interpreted as evidence for the presence of melt, and may suggest that the lineament acts as a route towards the surface for magma transport beneath these volcanic fields (Spence and Gross, 1990; Menzies et al., 1991). More recent teleseismic studies by West et al. (2004) across the Rio Grande rift revealed the absence of a deep-mantle low-velocity anomaly beneath the region of extension. This supports the interpretation of lithospheric extension and magmatism as results of stresses localised within the lithosphere and small-scale mantle convection, rather than deep-mantle upwelling due to the presence of an active plume/thermal anomaly beneath the rift (West et al., 2004).

Figure 1. General tectonic and volcanic map of Northern New Mexico, USA. Tectonic features include the extensional Rio Grande rift and Basin and Range, and stable lithospheric regions of the Great Plains and Colorado Plateau. Volcanic fields (VF) can be observed to lie along with the Jemez lineament. The Zuni Bandera Volcanic Field resides on the western flank of the Rio Grand rift in a transitional region with the Colorado Plateau. After Goff and Grigsby (1982).
ŠElsevier (2007)
LANDSAT image of the Zuni-Bandera Volcanic Field. Zuni-Bandera Volcanic Field (click link to field guide)

Mantle source for the ZBVF basalts:


Menzies and Kyle (1990) and Menzies et al. (1991) proposed a model where by the chemistry of the ZBVF magmas was explained by the interaction of three mantle sources: asthenosphere, lower lithosphere, and a mantle plume. The consequence of this model was that alkali basalts were believed to have originated within the relatively isotopically homogeneous convecting asthenospheric mantle (87Sr/86Sr ratio <0.703), whereas the tholeiitic basalts were extracted from the isotopically heterogeneous lithospheric mantle or a plume-contaminated asthenospheric mantle (87Sr/86Sr ratio >0.706). With no evidence to support a deep mantle plume-like mantle source, heterogeneity is likely to be sourced in the lithospheric mantle.

Rare Earth Elements (REE) modelling and mantle-facies partial-melting: 


-Semi-quantitative partial melting models using REE ratios can be calculated to constrain the extent and depth of partial  melting, along with the facies mineralogy and chemistry, using the methodologies of Thirlwall et al. (1994) and Shaw et al. (2003), and partition coefficients from McKenzie and O’Nions (1991). 

-With partial melting of either spinel- or garnet- facies peridotite, the Light REE (e.g. La) will be enriched in the melt to produce La/Yb variations with variable degrees of partial melting. Enrichment in the Middle REE (e.g. Dy) relative to Heavy REE (e.g. Yb) occurs during melting only where garnet is a residual phase. The HREE (Yb) are preferentially retained by garnet (highDYb ~4·0–15) relative to the MREE (Dy). Thus, increasing MREE/HREE (Dy/Yb) ratios develop with decreasing degrees of partial melting within the garnet-facies, producing large differences between source and melt ratios.  

-In contrast, melting of a spinel peridotite will result in a minimal variation in MREE/HREE (Dy/Yb) ratios with variable melting because DDyCpx is near-equal to DYbCpx (McKenzie and O’Nions, 1991), even in near sub-solidus clinopyroxene (Blundy et al., 1998), and the melt and source ratios will be comparable. 

-The relative contribution of each facies can be identified using a plot of (Dy/Yb)N vs. (La/Yb)N. On such plots (Fig. 8ab) the denominator for each ratio remains the same (Yb) and mixing between melts from each facies will produce linear arrays (Hanson and Langmuir, 1978) with varying phase contributions and degrees of partial melting (Thirlwall et al., 1994). 

-Figure displays a plot of (Dy/Yb)N against (La/Yb)N for the ZBVF, along with modelled trajectories for non-modal fractional melts of garnet- and spinel- peridotite sources. The source is a primitive mantle composition (Sun and McDonough, 1989) as a crude approximation to begin modelling.


Conclusions:

-Alkali basalt samples display depleted 143Nd/144Nd and 87Sr/86Sr ratios relative to Bulk Earth (143Nd/144Nd = 0.51264 and 87Sr/86Sr ~ 0.7045) supporting melting from a LREE-depleted source, although not as depleted as MORB source.

-Tholeiite basalts straddle the Bulk Earth 143Nd/144Nd ratio and plot above the Bulk Earth 87Sr/86Sr, suggesting both LREE-enriched and -depleted sources were involved during melting. 

-Involvement of a LREE and isotopically enriched mantle source could have arisen through melting at the base of the lithosphere, which has been isolated from mantle convection as to produce aged isotopic values [87Sr/86Sr <0.706 (Menzies et al, 1991)], and has undergone metasomatic alteration, specifically from garnet-facies silicate fluids.

-To model the production of isotopically enriched tholeiite basalts as partial melts from a LREE-enriched source, starting with a primitive mantle composition (Sun and McDonough, 1989), ~15% source enrichment of La with an increased (Dy/Yb)N of ~1.25 have been introduced into the model. The results indicate that the ZBVF samples define a trend that intersects the spinel-facies melting trajectory between <1% and the garnet-facies melting trajectory at <1% melt (~0.7% sample QBO 607) (Fig. 8b).
 
-This is an expected result and implies that mantle melting beneath the ZBVF occurred without a mantle thermal anomaly. Consequently the amount of shallow mantle involved in polybaric melting remains the same allowing both suites to be produced at similar degrees of melting.