Petrology and geochemistry of the Valle de Santiago lower-crust xenoliths: Young tectonothermal processes beneath the central Trans-Mexican volcanic belt

The origin and evolution of magmatic and other geothermal systems in subduction zones along continental convergent margins are the subject of intense research and controversy because of the central role these complex processes play in the evolution of the continental crust and the ways in which volatiles or other elements are recycled through the mantle (McCulloch, 1993; Rudnick, 1995; Fyfe, 1997; Kerrick and Connolly, 2001). Indeed, crustal contamination of mantle-derived magmas also has been a recurrent but elusive problem for the geochemical and petrologic balance of magmatism along continental arcs. Paradigmatic systems such as the Andes in South America and the cordilleran igneous belts of western North America, including specific regions of Mexico, have been intensely studied by isotope (Sr, Nd, Pb, Os, and O) and trace-element geochemistry (e.g., DePaolo, 1981; James, 1982; Thorpe et al., 1984; Hildreth and Moorbath, 1988; Kempton and Harmon, 1992; Housh and McDowell, 2005), or by the analysis of deep-seated crustal xenoliths (e.g., Cigolini and Kudo, 1987; Roberts and Ruiz, 1989; Cameron et al., 1992; Taylor and McLennan, 1995), and yet no general agreement exists about the relative proportion of assimilation versus fractional crystallization (AFC), or other petrogenetic mechanisms such as partial melting and igneous mixing, for the final integration of the magmas extruded at the surface. The east-west–trending Trans-Mexican volcanic belt is a continental magmatic arc active since the Miocene (Ferrari et al., 2000; Gómez-Tuena et al., 2007a) that overlies the young Cocos and Rivera oceanic plates subducting under the continental crust (40–50 km thick according to geophysical data) of southern and central Mexico. Because of the oblique position of the Trans-Mexican volcanic belt relative to the dominant NW-trending tectonic structures of Mexico and the corresponding trench, the crustal underpinnings of the arc are composed of at least four exposed tectonostratigraphic terranes (Campa and Coney, 1983; Sedlock et al., 1993) (Fig. 1), the ages of which vary from Mesoproterozoic–Neoproterozoic (Maya and Zapoteco terranes to the east), through Paleozoic (Mixteco terrane), and Mesozoic (Guerrero terrane to the west). The quest to understand how this continental crust has influenced the composition and evolution of the Trans-Mexican volcanic belt magmas in the last 20 m.y. is also a major subject of debate, and many rather contrasting views have been advanced, including those that confer a fundamental importance to crustal contamination (e.g., Kudo et al., 1985; McBirney et al., 1987; Verma, 1999, 2000, 2001; Verma and Hasenaka, 2004; Chesley et al., 2002; Cebriá et al., 2011), and others that minimize that role (e.g., Luhr and Carmichael, 1985; Wallace and Carmichael, 1999; Martínez-Serrano et al., 2004; Gómez-Tuena et al., 2006, 2008, 2011, 2013; Straub et al., 2011). However, relevant data on the structure and composition of the crystalline basement (granitic and metamorphic rocks) beneath the Valle de Santiago volcanic field based on geological outcrops (e.g., Centeno-García, 2005), gravity and seismicity (Campillo et al., 1996; Molina-Garza and Urrutia-Fucugauchi, 1993; Urrutia-Fucugauchi and Flores-Ruiz, 1996), and rare deep-seated xenoliths or megacrysts (Righter and Carmichael, 1993; Aguirre-Díaz et al., 2002; Urrutia-Fucugauchi and Uribe-Cifuentes, 1999; Ortega-Gutiérrez et al., 2008a) are rather inconclusive, in contrast to the abundant and well-studied lower crust and mantle xenoliths that occur in many places of northern Mexico (Nimz et al., 1986; Ruiz et al., 1988; Hayob et al., 1989; Pier et al., 1989; Rudnick and Cameron, 1991; Schaaf et al., 1994; Aranda-Gómez and Luhr, 1996).

Although most gravity models (Molina-Garza and Urrutia-Fucugauchi, 1993; Urrutia-Fucugauchi and Flores-Ruiz, 1996) assume isostatic compensation along the Trans-Mexican volcanic belt, as well as standard crustal densities for a layered crust formed by an upper part of sedimentary and volcanic rocks, a middle part roughly composed of granitic rocks and low-grade metamorphic rocks, and a mafic granulite-facies lower crust, these assumptions are not always justified given the ages and lithotectonic diversity that characterize the Mexican crust. Thus, substantial uncertainties still persist about the real position and constitution of the Moho and the lower crust under critical areas of the Trans-Mexican volcanic belt, such as the Michoacán-Guanajuato volcanic field, where the xenolith-bearing maars of Valle de Santiago are located. In fact, the crust sampled by xenoliths beneath the Trans-Mexican volcanic belt should differ from that located to the north, because the arc lies astride several tectonostratigraphic terranes, many of which may be actually truncated across the province. Therefore, the use of the lower crust of northern Mexico to assess the petrogenetic role of that crust in the magmatic evolution of the Trans-Mexican volcanic belt (e.g., Verma, 2000) may not be valid—hence the importance of characterizing the real lower crust under the arc using the direct evidence provided by deep-seated xenoliths.

The main aim of this work is to describe and discuss in detail the petrologic and chemical significance of the only place in the Trans-Mexican volcanic belt where abundant, large, and diverse granulite-facies, lower-crust xenoliths have been found. Although some of the xenoliths have been already studied (Urrutia-Fucugauchi and Uribe-Cifuentes, 1999; Uribe-Cifuentes, 2006), our work is much more detailed and reports new types of xenoliths of great relevance for the petrogenetic characterization of the volcanic field and age of the underlying crust, including members of the charnockitic series and metanorthosites. We also provide detailed mineral and rock geochemical data representative of the granulitic xenoliths, calculate the approximate pressure-temperature (P-T) conditions for their crystallization, and for the first time report the crystallization age of one of these xenoliths based on U-Pb zircon systematics. On these bases, we propose a new model for the constitution, age, and evolution of the lower crust beneath Valle de Santiago in the central sector of the Trans-Mexican volcanic belt.

The maars of the Valle de Santiago volcanic field are emplaced in a distinctive marginal part of the Michoacán-Guanajuato volcanic field in the west-central sector of the Trans-Mexican volcanic belt (Fig. 1). These spectacular maars have attracted geoscientists for more than a century, who have studied them from the simplest morphological aspects (e.g., Ordóñez, 1900, 1906) to their environmental attributes (Park, 2005; Armienta et al., 2008; Kienel et al., 2009; Aranda-Gómez et al., 2013) and their petrologic, geochronological, geochemical, and volcanological features (cf. Cano-Cruz and Carrasco-Núñez, 2009). Xenoliths, on the other hand, have remained poorly studied, despite the importance of this locality and their bearing on central geological aspects such as the age of its crustal source or sources. The existence of granulitic xenoliths was first reported by Uribe-Cifuentes and Urrutia-Fucugauchi (1992) and subsequently studied in more detail by Urrutia-Fucugauchi and Uribe-Cifuentes (1999) and Uribe-Cifuentes (2006).

The maars are located well behind the arc front at the northernmost limit of the central sector of the Trans-Mexican volcanic belt in the Michoacán-Guanajuato volcanic field (Figs. 1 and 2), facing the N-trending Tzitzio contractional structure of the Lower Mesozoic basement (Martini et al., 2009). The Michoacán-Guanajuato volcanic field is probably the largest monogenetic active volcanic complex on Earth (Hasenaka and Carmichael, 1987; Hasenaka, 1994), as it extends for ∼40,000 km² and contains over 1000 Quaternary monogenetic volcanoes (cones, flows, shield volcanoes, and maars), but only one polygenetic stratovolcano, the extinct Tancítaro stratocone (cf. Ownby et al., 2011). The Valle de Santiago volcanic field lies 360–370 km away from the Acapulco Trench, where the Cocos plate subducts beneath the continental plate of southern Mexico at an angle of ∼25°, indicated by position of the arc 200 km from the trench and clearly defined by the historical eruption of the Jorullo monogenetic volcano (Luhr and Carmichael, 1985).

In addition to Quaternary maars and lacustrine-fluvial deposits, the Valle de Santiago area consists of Pliocene to Holocene volcanic units, including abundant cinder and lava cones, shield volcanoes, and eroded composite volcanoes (cf. Urrutia-Fucugauchi and Uribe-Cifuentes, 1999). There are over 13 maars in the area (Fig. 2), nine of which define a NNW-trending line controlled by structures in the basement, such as a conspicuous geophysical lineament that follows a W101.25° orientation, and they are interpreted as an expression of the abrupt thinning of the crust lying west of the lineament (Ortega-Gutiérrez et al., 2008a).

The Rincón de Parangueo maar, the main object of this study, is located at 20°25′48″N, 101°15′00″W, in the state of Guanajuato, central Mexico, ∼1700 m above sea level; it forms a crater with external dimensions of 1.99 × 1.71 km, which includes an almost dry crater lake with a diameter of 1.18 km filled by silts, pyroclastics, evaporates, and stromatolites. The highest point of the northern rim is at 2012 m above sea level, whereas the lake bottom lies at 1685 m above sea level, i.e., having a maximum interior relief of 327 m. The alluvial-lacustrine plain outside the crater to the south is ∼66 m below the borders.

The nine principal maars of the Valle de Santiago volcanic field (locally called “hoyas”) are relatively young volcanoes dated at 1.1 Ma to 73 ka, with an age of 0.4 Ma for Rincón de Parangueo maar (Murphy, 1986). The maars extend 31 km from Lake Yuriria in the south to just north of the city of Valle de Santiago, and they vary in diameter from nearly 2 km (Rincón de Parangueo maar, La Cíntora, Hoya de Álvarez, and Hoya de Estrada) to a few hundred meters (Yuriria and La Alberca). The Rincón de Parangueo maar forms the southernmost volcanic structure of a cluster with three coalescing maars of about the same dimension, and probably age, as suggested by their similar and superposed geomorphologic profiles (Fig. 2, inset).

With the exception of the smallest structures, La Alberca and Yuriria, xenoliths were found in the eight maars visited at Valle de Santiago volcanic field. However, because Rincón de Parangueo displays the largest diversity and abundance of deep-seated xenoliths, this study is mostly based on samples taken from that maar, and complemented by petrologic and chemical data from La Cíntora and Álvarez maars, where most xenoliths have partially to totally preserved the original gabbroic texture. Figure 3 shows representative xenoliths and the transporting basalt. The hosting basalts at Rincón de Parangueo and La Cíntora maars are highly vesicular, megacrystic, plagioclase-olivine porphyries (Fig. 3A), which are described in detail next.

Lithologies, minerals, and textures of the Valle de Santiago xenoliths are not homogeneous and include igneous and granulitic metamorphic rocks, as well as hydrous high-grade assemblages. However, with the exception of a unique biotite-spinel-fayalite-plagioclase igneous xenolith, the rest are essentially mafic orthogranulites, gabbros, and metanorthosites, most of which show incipient to pervasive textures defining equal-angle triple junctions and anhydrous reaction rims, as well as tectonic fabrics that are characteristic of high-grade orogenic metamorphism. Table 1 shows a summary of xenolith mineralogy covering the whole range of lithologies found in the studied xenoliths. A noteworthy finding is the first observation in the region of two charnockitic xenoliths (faersundite RP19 and enderbite RP20), probably sourced in the deep granitic crust beneath the volcanic field.

Basalts containing xenoliths show macroscopic highly vesicular (∼40% pore) and porphyritic (∼20% phenocrysts) to aphanitic textures (Fig. 3A) with abundant olivine and plagioclase phenocrysts. Plagioclase phenocrysts are up to 4.5 mm in size, extremely fragmented, and corroded by the magma, suggesting a complex origin. Clinopyroxene phenocrysts are up to 1.8 mm in size, but it is mainly present in the matrix and as glomerocrysts. The largest olivine phenocrysts attain ∼2.5 mm in size and show faint zoning toward less forsteritic rims; they are fresh and often contain euhedral inclusions of chromite, whereas matrix olivines lack chromite inclusions and are intensely iddingsitized. Orthopyroxene is weakly pleochroic in shades of pale green to pale brown, similar to some orthopyroxene in the granulite xenoliths, indicating a xenocrystic character, as suggested by thin rims of clinopyroxene formed in contact with the magma. Deformed quartzite and two-pyroxene granulite and cumulitic gabbro microxenoliths as well as glomerocrysts of clinopyroxene-orthopyroxene are commonly enclosed in the basaltic matrix. Basalt-xenolith contacts show no thermal or chemical effects, indicating rapid transport to the surface.

Two main types of crustal xenoliths (5–10 cm long, 0.5–2.5 mm grain size) were collected: gabbroic (Fig. 3F) at Hoya de Alvarez and La Cíntora maars, and metamorphic mafic and felsic granulites (Figs. 3B and 3E) at Rincón de Parangueo and La Cíntora maars. This work deals mostly with the granulites of Rincón de Parangueo. Cataclastic and foliated textures (e.g., RP-9 and RP-11, respectively) in the metamorphic xenoliths intersect at high angles to the magmatic microbanding in some of the granulite xenoliths (see Figs. 3C and 3E), indicating solid-state ductile deformation at high temperatures after magmatic crystallization.

Gabbroic igneous xenoliths are essentially troctolites and norites consisting of various combinations of olivine, pyroxene, plagioclase, and rarely an opaque primary phase. The formation of abundant cumulate anorthosite and troctolite, as well as textural evidence of early crystallization of plagioclase from this magma (euhedral inclusions in olivine and clinopyroxene), indicates low water contents during their deep emplacement. Gabbroic xenoliths at La Cíntora maar (LC-1) preserve almost intact igneous cumulitic and poikilitic textures (Figs. 4E and 4F) progressively crystallizing green spinel, olivine, clinopyroxene, and plagioclase. Late titaniferous amphiboles, clinopyroxene, iron-rich olivine, titanomagnetite, and glass are associated with younger and shallower magmatic events. Spinel occurs as isolated euhedral crystals in the gabbroic fabric and as bands arranged along one or two crystallographic planes of the plagioclase (Fig. 4A). This later texture may be interpreted as exsolution of the spinel from a cumulitic, high-pressure and high-temperature plagioclase of the gabbroic rock (e.g., Wass, 1973; Wilkinson, 1975). Magnetite, clinopyroxene, and iron-rich epidote occur in veins and pockets, whereas biotite in the felsic xenoliths forms elongate laths up to 1.8 mm in size with an intense, black to bright yellow-brown pleochroism.

Transitional gabbroic to granulitic and anorthositic xenoliths (e.g., RP2-3, RP-9, RP-10, RP-17, RP-18) show a composite texture of metamorphic-type granoblastic texture with triple junctions between grains and igneous cumulitic features; interestingly, some of these xenoliths, including the anorthosites, preserve igneous olivine in the assemblage, whereas it is rare in the fully developed granulites, suggesting that magmatic olivine and orthopyroxene reacted with plagioclase to produce the composite granulite facies Pl-Opx-Cpx-Spl assemblage (as in Table 1; mineral abbreviations hereafter follow Whitney and Evans, 2010). Interstitial glasses (discussed in the following) are common in the mafic xenoliths, where many of the crystal-crystal contacts are occupied by brown glass and quenched, partially crystallized magma. Where melt penetrated the plagioclase mosaic, a distinctive microbrecciated texture developed, probably caused by overpressure of the volatile-rich infiltrated magma. Older glass occurs as spherical inclusions in olivine, orthopyroxene, clinopyroxene, and plagioclase, and although this glass was not analyzed, it may have originated from the basaltic magma parental to the cumulitic anorthosite. One- and two-pyroxene mafic granulites (samples RP-2-11, RP-16, and RP-1, RP-11, and RP-15, respectively) are the most abundant xenolith types at Rincón de Parangueo, accompanied by olivine and spinel-bearing granulites (samples (RP-5, RP-13, RP-17, MA-1, and RP-3, RP-7), as well as by orthopyroxene-spinel metanorthosites (RP-9). More rarely, charnockitic xenoliths (RP-19 and RP-20) containing abundant zircons served to date the igneous protolith (see following) of this lower-crust lithology. Mafic granulites are fine (0.5 mm) to moderate grained (1–2.5 mm), often defining diffuse compositional bands. Minor brown hornblende, kaersutite, and secondary opaque oxides may be present in some of the transitional mafic xenoliths (e.g., RP2-3). Zircon (20–50 µm long) was found as euhedral and elongate inclusions within plagioclase in only one of the anorthositic xenoliths but was not separated for dating. Nickeliferous pyrrhotite occurs in trace amounts as inclusions in olivine, suggesting an igneous origin, whereas interstitial and veined magnetite is always of secondary origin. Garnet, or its possible pseudomorphs, on the other hand, is absent from the studied assemblages.

The charnockitic xenolith (RP-19) consists of the equilibrium assemblage Pl-Opx-Qz-Bt-Kfs-Ilm-Zrn-Ap and accessory rutile and monazite. Glass sometimes associated with pyroxenes, magnetite, and myrmeckite appears to be a product of decompression melting involving biotite, because it occasionally developed narrow rims of glass-orthopyroxene-titanomagnetite inside plagioclase. This rock in fact preserves clear magmatic textures such as euhedral inclusions of biotite within euhedral plagioclase, bipyramidal quartz, elongate (12/1) euhedral zircons and aggregates of apatite-ilmenite.

Polished thin sections representative of the mafic granulites, after petrographic characterization, were carbon-coated and analyzed by wavelength-dispersive spectroscopy using a JEOL l -A 8000R microprobe equipped with five spectrometers at the Laboratorio Universitario de Petrología (LUP), Universidad Nacional Autónoma de México (UNAM), using natural and synthetic phases as standards. Current and voltage used were 10 nA and 20 kV, respectively, with counting times of 20 s for silicates and a focused beam, except for the analyses of glasses, for which counting times were reduced to 5 s, and the beam was defocused to 10 µm in order to minimize sodium loss.

Major- and trace-element analyses, as well as Sr, Pb, and Nd isotopic analyses, were performed on two representative mafic granulites, six basaltic samples from the Valle de Santiago volcanic field, and two plutonic rocks of La Huacana sequence (Table 2). Major elements were determined by X-ray fluorescence spectrometry using a Siemens SRS-3000 instrument at the Laboratorio Universitario de Geoquímica Isotópica (LUGIS) of UNAM, using procedures of Lozano-Santa Cruz and Bernal (2005). Trace-element data were obtained by inductively coupled plasma–mass spectrometry (ICP-MS) using a Thermo Series XII instrument at the Centro de Geociencias (CGEO) of UNAM, following the sample preparation and measurement procedures described by Mori et al. (2007). Sr, Nd, and Pb isotopic ratios were measured by thermal ionization mass spectrometry (TIMS) at LUGIS using a Finnigan MAT 262 system equipped with eight Faraday cups, following sample preparation and measurement procedures described by Schaaf et al. (2005) and Mori et al. (2009).

U-Pb laser-ablation (LA) ICP-MS geochronology analyses were performed on zircons separated from the charnockitic xenolith RP19, employing a Resonetics M050 excimer laser coupled to a Thermo Xseries quadrupole ICP-MS, according to the methodology described by Solari et al. (2010). A spot size of 34 mm was chosen for laser analysis, and the Pléšovice standard zircon (ca. 337 Ma; Sláma et al., 2008) was employed as an age standard. A secondary standard zircon 91500 (Wiedenbeck et al., 1995) was measured as unknown, to check data reproducibility. During the analytical session, 18 analyses of 91500 standard zircon were performed, obtaining a concordant age of 1063.2 ± 1.1 Ma, which is well in agreement with the accepted age of ca. 1065.4 ± 0.6 Ma (Wiedenbeck et al., 1995). The analyzed raw data were reduced, and error propagated, employing Iolite (Paton et al., 2010). The ²⁰⁷Pb/²⁰⁶Pb ratios, ages, and errors were calculated using the VisualAge data reduction scheme of Petrus and Kamber (2012).

U-Pb geochronological data by LA-ICP-MS were obtained from 60 zircons belonging to a charnockitic xenolith (Opx-Pl-Kfs-Qz-Bt) recovered at Rincón de Parangueo maar. They show prismatic to stubby morphologies, are colorless to slightly yellow, and are up to 220 mm in size. The resulting ages are in general concordance within the assigned error (Table 3). Apart from one analyzed zircon, which is Paleozoic, straddling the concordia curve at ca. 500 Ma, all the others range in age from ca. 25 to ca. 75 Ma (Fig. 5A). In a probability density plot (Fig. 5B), the data show the most pronounced peak at 67 Ma, followed by a second at 45 Ma (Eocene). A few concordant zircons have younger ages at 39, 33, and 26 Ma (Fig. 5A; Table 3). The Th/U ratios are quite variable (0.04–1.16; Table 3) but generally above the upper limit of 0.07–0.1, which is taken as a proxy to suggest a possible metamorphic origin (e.g., Rubatto, 2002; Hoskin and Schaltegger, 2003). Moreover, the few analyzed rare earth elements (REEs) show a positive Ce anomaly and an appreciable negative Eu anomaly, as well as a fractionated heavy (H) REE pattern (Lu_N/Dy_N > 3; Table 4; Fig. 5B, inset). The only four analyses yielding Th/U 0.04–0.06 range in age from ca. 50 to ca. 68 Ma (Table 3). All these data suggest that zircon ages in fact reflect igneous crystallization, and our interpretation is that the most abundant, euhedral zircons with a mean age of 67 Ma indicate the apparent crystallization age of the igneous protolith. Those zircon analyses, which straddle the concordia between such an igneous age and ca. 25 Ma, are interpreted as the product of limited Pb diffusion of smaller zircons under crustal conditions during the subsequent granulite-facies metamorphism. Our interpretation is also plausible considering that Late Cretaceous granites are known in central-western Mexico, whereas those of early Eocene age are only locally inferred by means of cooling ages (e.g., Schaaf et al., 1995). Also, at the moment, none of the ca. 45 Ma known granites (e.g., Valencia et al., 2009) shows evident contamination with Late Cretaceous zircons. A further consideration is that the age of the youngest concordant zircon of 25.5 Ma (Th/U ratio of 0.12) constitutes the minimum age of granulite-facies metamorphism. Because the charnockite occurs together with mafic granulite xenoliths (Ol-Pl-Opx ± Cpx ± Spl) recrystallized at ∼9 kbar and ∼1000 °C, similar or lower P-T granulite-facies conditions are assumed for the metamorphism of the RP19 felsic orthogneiss.

In order to constrain the P-T conditions and evolution of the crust where xenoliths originated, as well as their possible interaction with the ascending basaltic magma, chemical analyses of representative mafic and anorthosite xenoliths were performed on all major phases, including glasses of samples RP9 (metanorthosite) and RP2 (mafic granulite) at Rincón de Parangueo and LC1 at La Cíntora maars.

This phase is present in all xenoliths studied in thin section. Cores and rims of plagioclase were analyzed in the mafic granulitic xenolith RP2 (Table 5), the composition of which is remarkably homogeneous and shows no zoning, as confirmed by cursory analyses across single grains. This plagioclase shows a variable and less calcic composition (An_56.8–71.5) compared with plagioclase from the anorthosite (An_75.7–77). The very low potassium in the latter (Or_0.7–0.8) is consistent with its postulated cumulitic origin. Plagioclase in RP2-3 mafic hydrous xenolith (Table 5), on the other hand, is quite variable in composition (An_48–70) but with similar potassium contents (Or_1.4–3.5) to that in the mafic anhydrous granulite. Optical determinations in other xenoliths also yielded compositions between labradorite and bytownite.

Orthopyroxene is the most common mafic phase in the granulitic and anorthositic xenoliths (Table 6) but is absent in the kaersutite metagabbro. At Rincón de Parangueo, orthopyroxene in sample RP2 shows a distinctly higher and homogeneous Mg# of 79–80 (En_76.0–77.8), compared that of the anorthosite (RP9), and Al₂O₃ up to 5.84 wt%, whereas the composition (En₇₁) and calcium oxide of the metanorthosite orthopyroxene are distinctly lower than in the mafic granulites. Its relatively high Al₂O₃ content (4.01–5.60 wt%) and distinctive pleochroism in shades of green and yellow are typical features of granulite-facies orthopyroxene; titanium occurs in minor quantities (0.15–0.36 wt% TiO₂). No compositional zoning was detected, but grains inside clinopyroxene or in contact with olivine or spinel are the most magnesian. All orthopyroxene plots in the low-Ca enstatite range of the pyroxene quadrilateral.

Clinopyroxene in Rincón de Parangueo mafic anhydrous xenoliths (Table 6) ranges in Mg# from 78 to 79, and its composition plots in the augite field of the pyroxene quadrilateral, whereas in the metanorthosite, the Mg# range is distinctly lower (69–71), but with higher titanium and sodium contents. Both types show high alumina (6.66–7.91 wt% Al₂O₃).

This oxide is ubiquitous in all mafic granulites and metanorthosites, as well as in the gabbroic xenoliths. In the granulites (RP2), it occurs in the matrix or as aligned euhedral inclusions in plagioclase (Fig. 4A), as well as intergrown with clinopyroxene (Fig. 4B). Its composition (Tables 7 and 8) is not homogeneous, with Mg# varying in the range of 58–60 (pleonaste) in the mafic xenolith RP2 to 37–38 (hercynite) in the anorthositic xenolith (RP9) of Rincón de Parangueo. Titanium oxide is distinctly higher in the latter. The Mg# of spinel in the granulite xenolith from La Cíntora varies from 41 to 55. Calcium, manganese, and zinc occur in very small amounts, whereas chromium oxide varies from 0.02 to 3.64 wt%. Exsolved and matrix spinels in RP2 show similar composition. Spinels from the RP2 mafic granulite and RP9 contain substantial chromium oxide (2.27–2.66 wt%), whereas most spinels in RP2 xenolith do not contain chromium and show a much lower Mg# of 37–38. The spinel xenocryst in the basaltic matrix shows a distinct composition (Table 7) compared with those in the gabbroic granulites (Table 8). It is rich in titanium (2.54 wt% TiO₂) and also yielded a low Mg# (41).

Phenocrysts and microlites of olivine from the host basalt at La Cíntora maar and grains of olivine from a mafic granulite xenolith inside the same basalt (Table 7) were analyzed for petrogenetic comparisons. The composition of the igneous (phenocrysts and microlites) olivine from the host basalt is similar in forsterite (Fo_70–74) to that from the granulite (Fo_{74 75}), although calcium seems to be higher in the former. Nickel and calcium are in all cases negligible. Olivine in the hydrous gabbroic granulite (RP2-3; Table 9) is present in two different generations: olivine (I), a forsteritic phase (Fo_75–77) in equilibrium with plagioclase and clinopyroxene, and olivine (II), as late rims of less forsteritic olivine (Fo₆₀) crystallized from a younger iron-rich gabbroic melt (see Fig. 6). Pyrrhotite with 2.23 wt% NiO is present as aligned inclusions in some olivine and plagioclase grains of the metanorthosite (RP9) and is probably an early magmatic phase.

Hornblende occurs in minor amounts as a late hydrous mineral in the metanorthosite xenolith (Table 10). The high Ti contents (avg. 3.92 and up to 4.06 wt%) corroborate a high crystallization temperature, whereas the elevated alumina (avg. 13.55 and up to 14.0 wt%) indicates relatively high pressures of crystallization (e.g., Ernst and Liou, 1998). The Mg# is 65, which is slightly higher than that of the kaersutite in the hydrous granulite and chemically corresponds to a titanian magnesio-hastingsite. Na and K almost saturate the A position of the structural formula.

Kaersutite in RP2-3 (Table 10) occurs always replacing the original clinopyroxene and olivine, accompanied by Fe-Ti oxides such as ilmenite and titanomagnetite. Titanium content of the kaersutites may be up to 5.16 wt%, accompanied by Na and K saturating the A position of the structural formula. The Mg# varies little from 58 to 62, and chromium in all amphiboles is negligible.

Biotite was only optically identified in the charnockitic granulites, but it was not analyzed. It is strongly pleochroic in reddish shades and is systematically associated with colored glass. It is not in equilibrium with orthopyroxene, which in part seems to be an incongruent melting product of the biotite. In contact with ilmenite, biotite apparently was also partially melted, because the resulting glass is rich in potassium and titanium (see Table 11). Zircon is present in the granulitic anorthosite RP9 as euhedral to rounded crystals 10–50 µm long, and a few contain inclusions of older euhedral zircons, a feature that is promising for the future dating of these xenoliths.

Ilmenite was the only oxide found in the metanorthosite (RP9), where it shows (Table 12) a considerable amount of magnesium (up to 5.67 wt% MgO) and is similar to ilmenites of gabbroic, layered complexes such as the Skaergaard of Greenland (Lovering and Wilson, 1968). The oxide forms a solid solution phase with the composition Ilm_80.3–89.7. Secondary, pure magnetite occurs in microveins in this rock. Fe-Ti oxides of the kaersutitic gabbroic granulite RP2-3 include coexisting ilmenite (Ilm_86–87) and titanomagnetite (Usp_56–58), as well as secondary magnetite. Alumina in the titanomagnetites is rather high, varying from 3.70 to 7.60 wt% Al₂O₃.

As mentioned already, glasses are present in many of the studied xenoliths, particularly in those richer in plagioclase. Up to three types of glasses are evident: (1) brown, oval inclusions in olivine, pyroxenes, and plagioclase, (2) intergranular between plagioclase grains (Ti-poor, colorless), and (3) in close contact with ilmenite or biotite (Ti-rich, deep brown). However, glasses were only analyzed (Table 11) in one of the xenoliths (RP-9), where their SiO₂ varied from 55 wt% to 60.4 wt%, thus precluding direct injection from the host basalt. All glasses except those in contact with ilmenite fall in the alkaline field of mugearites and benmoreites, with total alkalis at 6.31–8.32 wt%. They are peraluminous (Shand’s index of 1.24–1.39), although corundum is absent from the CIPW (Cross, Iddings, Pirsson, and Washington) norm. These late melts may be interpreted either as products of differentiation of an alkali basalt parental to the cumulate anorthosite, or derived from the decompression partial melting of an anorthositic xenolith, albeit with accessory biotite in order to explain the abundance of potassium (2–5 wt% K₂O) and high titanium (up to 3.5 wt% TiO₂) in those glasses next to ilmenite (Table 11, last three columns).

The igneous I event (gabbroic cumulates) included plagioclase, olivine, orthopyroxene, and clinopyroxene, as well as magmatic spinel, although the latter more commonly crystallized in symplectitic metamorphic aggregates of Cpx-Spl (Fig. 4D) in close relationship with plagioclase-orthopyroxene contacts, as well as broad rims around plagioclase. The original igneous assemblages with olivine and plagioclase of the troctolitic gabbros reacted (granulite-facies metamorphism) to form pyroxene and spinel by the general reaction 2Ol + An = 2Opx + Di + Spl, while the more specific reaction Ol + An = Di + Spl + Qz probably generated most of the Cpx-Spl symplectites. Although quartz rarely occurs as inclusions inside orthopyroxene, it may have been incorporated into a more sodic plagioclase or dissolved in a cryptic fluid phase. Zoning of plagioclase visible in thin section may reflect a change to a more sodic composition. The following equilibrium metamorphic assemblages were noticed in the mafic xenoliths: Pl-Opx-Cpx-FeTi oxides, Pl-Opx-Cpx-Ol, Pl-Opx-Cpx-Spl, and Pl-Opx-Cpx-Hbl–Fe-Ti oxides, all of which indicate attainment of granulite-facies conditions at low (Ol-Pl) to moderate (Cpx-Spl) pressures. Figure 7 illustrates the compatibility for the main six phases (plus Qz, Bt, and Kfs in the charnockitic samples) identified in the metamorphic xenoliths. Assemblages with olivine are present, but in most cases the rocks were converted to two-pyroxene-spinel granulites. Brown hornblende and kaersutite are late phases (igneous II) in certain gabbros and hydrous granulites (e.g., RP2-3) replacing orthopyroxene. Some of these metagabbroic xenoliths show the late access of a high-iron basaltic magma that infiltrated the granulites and partially replaced the original phases by the general assemblage kaersutite-ilmenite-augite-olivine-plagioclase-glass. The assemblage developed complex mosaic and coronitic textures indicative of lower-crustal metasomatic-magmatic processes associated with the Valle de Santiago volcanic field plumbing system. The origin of hornblende and apatite in this rock apparently involved a late metasomatic fluid that enriched it in water, potassium, and phosphorous. Where melt or supercritical fluids injected the xenoliths (e.g., RP-17), a distinctive set of narrow fissures (50 µm wide) and small quenched pockets occupied by micrometric albitic plagioclase and pistacitic epidote developed (igneous III), most probably during this late magmatic event, as suggested by the common occurrence of brown glass with quenched, highly oxidized products (magnetite, hematite, and iddingsite) along the mineral boundaries of granulites and gabbroic xenoliths. The glass was distributed preferentially along plagioclase grain boundaries in intimate association with hornblende, and in the quenched areas, it is strongly vesicular and oxidized to an irregular groundmass of magnetite and hematite. Figure 6 summarizes the proposed sequence of magmatic crystallization (four events) and the high-grade metamorphism thought to have occurred during the evolution of the gabbroic lower crust and the transporting magmas (igneous IV) that provided the xenoliths. However, the petrologic complexity of the late high-temperature igneous events and lack of adequate analytical data preclude a more systematic attention to the petrogenesis involving the glasses and iron-rich oxidized phases (fayalitic olivine, magnetite, kaersutite, biotite, and epidote).

Several conventional thermometers were applied for the granulitic xenoliths of Valle de Santiago (see Table 13). For RP2-3 Hbl-bearing xenoliths, the Holland and Blundy (1994) Pl-Hbl geothermometry yields temperatures of 1086–1094 °C, which are independently supported by Ti-in-amphibole geothermometry (e.g., Otten, 1984), which yields a temperature of 998 °C, and the Mg-Fe olivine-augite exchange thermometer of Loucks (1996), yielding temperatures in excess of 1000 °C for several Ol-Cpx pairs in the mafic granulite at assumed pressures of <10 kbar. Further support for metamorphic temperatures ∼1000 °C is drawn from the two-pyroxene geothermometer, which yields temperatures between 1100 °C and 1120 °C (QUILF program of Andersen et al., 1993) for the RP2 granulite pyroxenes; these are rather similar to the 1055 ± 6 °C obtained from coexisting Ilm-Mag pairs in the hydrous mafic granulite (RP2-3) using the ILMAT-2 excel worksheet thermometer of Lepage (2003).

On the other hand, because of the general absence of garnet and quartz in all the mafic and anorthositic xenoliths, geobarometers could not be applied to determine the metamorphic pressures attained by the mafic granulites with adequate precision. Nonetheless, the stability of olivine and plagioclase in the system CMAS (Ca-Mg-Al-Si) to yield pyroxenes and spinel with increasing pressure, as documented in Valle de Santiago mafic xenoliths, can be used to constrain the pressure at which the granulitic xenoliths equilibrated if the temperature is given. In basaltic systems, the P-T stability field of olivine coexisting with calcic plagioclase against the spinel-pyroxene higher-pressure equivalent assemblage shown in Figure 8 (e.g., Kushiro and Yoder, 1966; Obata, 1976; Johnson and Essene, 1982) indicates, for temperatures ∼1000 °C, maximum pressures of ∼9 kbar by the time the reaction took place in the mafic lower crust beneath Valle de Santiago volcanic field. The PROBE-AMPH 3 program of Tindle and Webb (1994) and geobarometry by Schmidt (1992) yield very similar average values of 8 kbar. Pressures between 8 and 9.2 kbar were obtained by the CpxBar Excel sheet by Nimis (cf. Nimis, 1999) for the aluminous clinopyroxenes of the mafic granulite RP2 and 4.5–9.4 kbar for those in the metanorthosite RP9.

Volcanic rocks in the Valle de Santiago area have variable compositions, but most of them can be chemically classified as Na-alkaline hawaiites to mugearites (Fig. 9). These rocks, including the Rincón de Parangueo juvenile magma, usually display low MgO but higher TiO₂ and FeO contents, and follow a near tholeiitic fractionation trend that differs significantly from the typical calc-alkaline tendency displayed by most rocks from the Michoacán-Guanajuato volcanic field (Hasenaka and Carmichael, 1987; Verma and Hasenaka, 2004).

Two granulite xenoliths analyzed in this study (Table 2) can be classified as hypersthene-normative gabbro (VS1) and olivine-normative anorthositic gabbro (VS2), with major-element compositions that are within the range of those reported by Urrutia-Fucugauchi and Uribe-Cifuentes (1999). Xenoliths have lower TiO₂, MnO, Na₂O, K₂O, and P₂O₅ but higher MgO and CaO than the host lava and other TiO₂-rich mafic volcanic rocks from the Valle de Santiago area (Fig. 9). Their major-element compositions somehow overlap with the most mafic calc-alkaline volcanic rocks from the Michoacán-Guanajuato volcanic field, albeit xenoliths have significantly lower Na₂O and K₂O contents. These chemical features most likely indicate that the studied xenoliths represent mafic cumulates rather than true crystallized liquids.

Trace-element patterns of xenoliths are very different to those of TiO₂-rich mafic magmas and other more typical calc-alkaline volcanics from the Michoacán-Guanajuato volcanic field, or even compared to older intrusive rocks that comprise the igneous basement of the region (i.e., La Huacana plutons; Fig. 10). Calc-alkaline magmatism of the Michoacán-Guanajuato volcanic field, as well as older plutons, have trace-element compositions that are typical of an arc environment, displaying positive spikes in large ion lithophile elements and negative Nb-Ta anomalies. In contrast, the xenolith-bearing mafic magmas have a smoother trace-element pattern and negligible Nb-Ta negative anomalies, which that are instead very similar to those found in intraplate tectonic settings. The analyzed xenoliths show prominent positive spikes in elements that are considered compatible with feldspar, such as Ba, Sr, and Eu, but they do not display relative enrichments in elements like Rb, U, or Th with respect to Nb and Ta. Thus, and despite the fact that their trace-element compositions must be strongly influenced by their plagioclase-rich cumulate character, the trace-element compositions of the xenoliths do not preserve features that may hint at derivation from a typical arc magma, and instead suggest a more close relationship with intraplate-like basaltic magmas of the region.

Chondrite-normalized REE patterns of the xenoliths are slightly enriched in light (L) REEs, with a strong positive Eu anomaly and a small but distinguishable concave-upward pattern of the middle to heavy REEs (Fig. 11A). Absolute REE concentrations of the xenoliths are lower than the host magmas but they display subparallel patterns with opposite Eu anomalies. Interestingly, if the trace-element patterns of the xenoliths are compared to those of contemporaneous rhyolitic volcanism from the western Trans-Mexican volcanic belt (Gómez-Tuena et al., 2014), a very clear “mirror” image in feldspar-compatible elements becomes apparent. These rhyolites have been interpreted as being derivative liquids of intraplate-like mafic magmas similar to those of Valle de Santiago that experienced a small amount of contamination (Gómez-Tuena et al., 2014), and in which plagioclase was clearly a major fractionating phase. It thus seems possible that the studied xenoliths from Valle de Santiago may represent the plagioclase-rich cumulative remnants of fractionating intraplate-like basalts that ultimately can give rise to derivative rhyolites like those found in western Mexico.

The Sr-Nd isotopic compositions of the studied xenoliths are slightly more enriched than their host lavas and plot at the enriched end point of the compositional spectrum documented in volcanic rocks from the Michoacán-Guanajuato volcanic field (Verma and Hasenaka, 2004). They are also identical to the compositions of Paleogene basement plutons from the region (i.e., La Huacana; Fig. 12A). The Rincón de Parangueo xenoliths are nonetheless more depleted than other lower-crustal xenoliths collected in northern Mexico (Ruiz et al., 1988; Schaaf et al., 1994), and they are also significantly different from the xenolith reported by Urrutia-Fucugauchi and Uribe-Cifuentes (1999), which plots well outside the overall Mexican crust-mantle trend. The Pb isotopic compositions of the calc-alkaline rocks from the Michoacán-Guanajuato volcanic field follow a well-defined linear correlation bracketed between mid-ocean-ridge basalt (MORB) and Cocos plate sediments, but the Valle de Santiago xenoliths are slightly displaced to higher ²⁰⁶Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb ratios, with values that overlap the compositions of high-TiO₂ intraplate-like volcanoes and older plutons in the area (Figs. 12B and 12C).

Calculated Sm-Nd depleted mantle model ages (Sm-Nd_TDM) from our two studied xenoliths are 582 and 900 Ma, which are similar to the calculated Sm-Nd_TDM model ages of Paleogene Huacana plutons or even to the high-TiO₂ intraplate-like basalts from Valle de Santiago (Fig. 12D). This indicates that Sm-Nd_TDM model ages of the studied xenoliths cannot be directly associated to crustal residence times and likely represent an average age of mixed mantle and crustal sources.

The absence of pre-Mesozoic outcrops along the entire Trans-Mexican volcanic belt represents a major problem for the petrogenetic interpretation of the province, particularly in regards to the lithotectonic nature and age of the underlying deep crust (e.g., Ortega-Gutiérrez et al., 2008a, 2008b). However, the Valle de Santiago volcanic field xenoliths sampled lower-crust mafic and felsic granulites and gabbroic xenoliths that might represent the buried pre-Mesozoic lower crust in the roots of the volcanic arc. Previously, a single “granulite-facies” xenolith was reported only 117 km to the east of Rincón de Parangueo maar, associated with an ignimbritic breccia of the Pliocene Amealco caldera (Aguirre-Díaz et al., 2002). However, unlike the Valle de Santiago volcanic field mafic granulitic xenoliths, which were carried up by alkaline basalts, the Amealco xenolith is hosted in an ignimbrite at the caldera rim; the xenolith is of tonalitic composition with the assemblage Pl-Qz-Opx-Cpx-Ilm-Ap-Zrn, and 20% interstitial glass of trachiandesitic to highly siliceous rhyolitic composition. Geochemical data show that the xenolith represents an arc-related igneous protolith with an isotopic composition that is slightly more enriched than the studied xenoliths, albeit with a Sm-Nd_TDM model age of 683 Ma, which is within the range of that documented for Valle de Santiago (Fig. 12D). These petrologic and geochemical characteristics indicate a relatively evolved magma, which probably underwent considerable assimilation of pre-Mesozoic crustal components. The upper-crustal pressures (<3.2 kbar) obtained for the Amealco xenolith on the basis of fluid inclusions rich in CO₂ (Aguirre-Díaz et al., 2002) were interpreted to reveal a granulite-facies Precambrian terrane underlying the central Trans-Mexican volcanic belt at depths of only 5–10 km. However, this conclusion is not compelling, because the Precambrian “age” is only a Nd depleted mantle age, which may be also interpreted as derived from the assimilation of Mesozoic sedimentary protoliths shed from nearby Precambrian crust (e.g., Oaxaquia). In fact, granitoids of Oligocene–Eocene age (Martini et al., 2009; Schaaf, 2013, personal commun.) are widely exposed, forming the crystalline basement of the Michoacán-Guanajuato volcanic field (Luhr and Carmichael, 1985; Verma and Hasenaka, 2004), and shallow granitic xenoliths occur at Quaternary Paricutín, Jorullo, and Arocutín monogenetic volcanoes within the Michoacán-Guanajuato volcanic field (Wilcox, 1954; Luhr and Carmichael, 1985; Corona-Chávez et al., 2006, respectively).

Urrutia-Fucugauchi and Uribe-Cifuentes (1999) published preliminary isotopic data on the Valle de Santiago xenoliths and reported a Sm-Nd_TDM model age of 1.5 Ga Ma, which is significantly older than the calculated model ages of our studied xenoliths (Fig. 12D). Exposures of crystalline rocks near the Valle de Santiago volcanic field include the Eocene La Huacana pluton (Table 14; Figs. 9, 10, and 11), with Rb-Sr whole-rock ages of ca. 42 Ma (Schaaf et al., 1995) and Ar-Ar plagioclase cooling dates of ca. 34 Ma (Martini et al., 2009), which indicate that shallow crystalline crust underlying the Michoacán-Guanajuato volcanic field has a young emplacement and cooling history. Moreover, Paricutín volcano (55%–56% SiO₂, δ¹⁸O = + 6.9‰–7.0‰, ⁸⁷Sr/⁸⁶Sr = 0.7038), only 148 km to the SW from the Valle de Santiago volcanic field maar, contains granitic xenoliths with similar silica and isotopic composition as the Huacana granitoids (McBirney et al., 1987), perhaps suggesting assimilation of more-evolved crust. In fact, AFC modeling of Paricutín lavas (Cebriá et al., 2011) estimates up to 40% of continental crust assimilation (granitic melts), with a plumbing system envisaged as a differentiated, density-stratified magma chamber feeding the volcano at depths below the granitic roots. These roots were tapped and partly assimilated by the post-1947 magma batches, including a crustal contaminant represented by the shallow granitic xenoliths. However, the total absence of paragneisses in the xenolith population of Valle de Santiago volcanic field maars, the latest Cretaceous crystallization age of the granulite felsic protoliths, the near absence of pre-Mesozoic zircon inheritance in the dated xenoliths, and the similar isotopic compositions between the Valle de Santiago xenoliths and La Huacana pluton argue against the existence of pristine Precambrian or Paleozoic crust beneath this part of the Trans-Mexican volcanic belt.

We thus conclude that Sm-Nd_TDM model ages provide little direct information on the actual age of crust-forming events, because they likely represent averages and mixtures between different mantle and crustal components. Nevertheless, and despite these limitations, the significantly lower Nd isotopic compositions found in apparently Precambrian xenoliths from northern Mexico confirm that the continental crust beneath Valle de Santiago and the rest of the Michoacán-Guanajuato volcanic field is probably of much younger age. Thus, Mesozoic igneous and sedimentary rocks belonging to the Guerrero terrane, which possibly include some recycled components of Precambrian and Paleozoic rocks, would better represent the basement beneath this part of the Trans-Mexican volcanic belt.

The overall geochemical and petrological evidence of the studied xenoliths is likely indicative of a complex interaction between recent influxes of mantle-derived melts and a preexisting continental crust. Such interaction has to be more complex than simple bulk assimilation coupled with fractional crystallization, as it likely involved mechanical hybridization and a selective transfer of certain chemical elements among different geologic components. Indeed, while the major- and trace-element compositions of mafic xenoliths appear consistent with a cumulate origin derived from fractionating intraplate-like basalts (Fig. 10), their slightly enriched isotopic compositions apparently require an additional input from a preexisting continental crust that must be compositionally identical to La Huacana pluton (Fig. 12). It is interesting to note, however, that the melt-rock interaction that occurs below Valle de Santiago appears to be completely isolated from the subduction environment, as more typical calc-alkaline rocks from the Michoacán-Guanajuato volcanic field evolved independently and incorporated magmatic components with different Pb isotopic compositions (Figs. 12B–12C). Thus, while granitic xenoliths are common in lavas all across the Michoacán-Guanajuato volcanic field, the processes of intraplate mafic influx, crystallization, crustal hybridization, and granulitization recorded in the xenoliths and magmas of Valle de Santiago do not necessarily represent widespread phenomena occurring in the entire Michoacán-Guanajuato volcanic field or the Trans-Mexican volcanic belt as a whole.

The hot nature of the lower crust under the Valle de Santiago volcanic field is evident from the mafic granulites described herein, which lack “cold” assemblages typical of cratonic or thickened lower crust, such as garnet-rutile, or clinopyroxene-garnet pairs, but instead abound in “hot,” anhydrous phases, including ubiquitous equilibrium assemblages of Pl-Ol, and Pl-Opx-Cpx-Spl. Thermal modeling of the crust in the Michoacán sector of the Mexican subduction zone (e.g., Currie et al., 2002; Manea et al., 2005) predicted temperatures above 800 °C at the base of a putative 40-km-thick crust. These conditions require an average geothermal gradient of ∼20 °C/km and consequently the stability of garnet in a mafic lower crust, and yet no garnets or their vestiges have been found in any of the xenoliths studied at Valle de Santiago maars. More recent studies on the thermal state of the present crust beneath the Valle de Santiago volcanic field (Manea and Manea, 2011) predict a Curie temperature (magnetite) of 585 °C located at 17–20 km below the area, and thus a geothermal gradient between 34 and 29 °C/km, which is more similar to that obtained here from the geothermometry of the Valle de Santiago xenoliths (∼33 °C/km). The mineral assemblages of all granulites from Valle de Santiago (Opx-Cpx-Pl-Ol-Spl) grade from metanorthosites to two-pyroxene granulites, and many xenoliths show rough mineralogical banding, indicating that this part of the lower crust may be represented by recrystallized layered gabbroic bodies, where fractional crystallization of plagioclase and olivine played a major role during the igneous event. In those xenoliths where spinel instead of olivine is present, the assemblage Cpx-Opx-Pl-Spl for similar temperatures would be stable up to ∼9 kbar, equivalent to a crustal depth of ∼33 km. Thus, the total interval of the lower crust sampled by these xenoliths may be as thick as 10 km, lying between 22 km (Ol-Pl) and 33 km (Cpx-Spl) deep. If the local crust is indeed 40 km thick (11 kbar), as suggested by gravity and other geophysical methods (Molina-Garza and Urrutia-Fucugauchi, 1993; Mazzarini et al., 2010), the Valle de Santiago xenoliths would not come from the base of the lower crust. Then where did the granulite xenoliths from Valle de Santiago volcanic field maars come from? Is there a tectonically uplifted granulite, anhydrous crust at middle depths below the Cenozoic volcanics? If so, what would be its age? Do the xenoliths represent Neogene–Quaternary basaltic cumulates emplaced near the base of the local lower crust and metamorphosed to granulite facies? Our model prefers this latter alternative, but definite answers to these questions must await further studies and dating of the mafic xenoliths, as well as the application of more precise thermobarometric determinations than the ones reported in this work.

Three pulses of intense magmatism are recorded on the Pacific margin of Mexico that occurred during Late Cretaceous–Cenozoic times: (1) latest Cretaceous–Paleocene (Laramidic), represented in southern Mexico by the coastal batholiths of Jalisco and Michoacán (Schaaf et al., 1995), (2) the Eocene–early Miocene plutons and ignimbrite flare-ups of the Sierra Madre Occidental (Ferrari et al., 2002), and (3) the Miocene–Holocene Trans-Mexican volcanic belt (Gómez-Tuena et al., 2007b). Thus, the studied granulite-gabbroic xenoliths could be of any age younger than the Late Cretaceous charnockitic xenoliths. Unfortunately, U-Pb geochronology performed on one felsic xenolith permits us only to propose the presence of considerable volumes of granitic crust formed during the Laramidic pulse beneath the Valle de Santiago volcanic field. Indeed, as the main cause for the origin of the mafic granulites and charnockites, the preferred model in this work (Fig. 13) envisages the massive access and stagnation in the lower crust of batches of Neogene and Quaternary basaltic magmas related to the Trans-Mexican volcanic belt. The 400 ka basalt that contains the xenoliths would be the youngest expression of the local magmatic event that included the maars of Valle de Santiago.

Although gravity (Molina-Garza and Urrutia-Fucugauchi, 1993; Urrutia-Fucugauchi and Flores-Ruiz, 1996) or theoretical models based on self-similar clustering of basaltic vents and their relationships to hydraulically connected fractures (Mazzarini et al., 2010) suggest a thickness of ∼40 km for the local crust (i.e., Moho pressures of ∼11 kbar), where garnet would be stable at temperatures <1000 °C, the apparent absence of that phase in the Valle de Santiago volcanic field xenoliths implies that, either the underlying source represents the base of a thinner (∼33 km) and hot (>1000 °C) crust, or that the xenoliths equilibrated up to 10 km above the Moho in a crust that is ∼40 km thick. Figure 13 shows our preferred crustal model, which can explain the nature and crustal evolution of the Valle de Santiago volcanic field magmas that brought the xenoliths to the surface. The model features the possible underlying presence of conduits rooted in former magmatic pools in the lower crust that eventually fed the maars. The transformation of the igneous plagioclase-olivine-pyroxene-spinel gabbros to the metamorphic clinopyroxene-spinel–bearing granulitic assemblage may be interpreted as the product of isobaric cooling of an original basaltic system emplaced in the deep crust at temperatures ∼1200 °C, rather than as the isothermal tectonic thickening of the local crust. Stagnation of mafic magmas in the lower crust has been proposed as the source for granulite-facies cumulate mafic xenoliths at La Olivina in northern Mexico (Cameron et al., 1992), for which pressures ≤7.2 kbar and temperatures of 850 °C were recorded. The higher pressures and temperatures estimated for the Rincón de Parangueo cumulitic metagabbroic granulites (∼8–9 kbar and 950–1050 °C) attest to a similar origin but at deeper levels in the lower crust. Although magmatic crustal thickening by crystal accumulation in the deep crust associated with intense volcanism may be a common process in hot continental arc systems, as modeled for eastern Nevada (Grunder, 1995), the absence of garnet and the textural stability of Ol-Pl in some xenoliths favor a relatively thin crust (∼33 km) underplated by basaltic magmas beneath the Valle de Santiago volcano field. The total absence of mantle xenoliths in fact may be explained if the xenolith uptake occurred above the Moho discontinuity as shown in Figure 13.

We conclude that mafic granulites, gabbros, and metanorthosite xenoliths unambiguously indicate a mafic lower-crustal source beneath Valle de Santiago maars. Mafic xenoliths that clearly preserve gabbroic and transitional metamorphic textures, such as euhedral crystals in the groundmass and within recrystallized mafic phases of the xenolith, strongly suggest that the precursor rocks of the mafic granulites were gabbroic igneous cumulates (anorthosite-troctolite-gabbro) derived from underplated ocean-island basalt–type basaltic magmas. The latest Cretaceous U-Pb zircon igneous age of the charnockitic xenolith and absence of older inheritance imply a magmatic event (Laramide) overprinted by a late Cenozoic granulite-facies metamorphism that affected the original gabbroic crust beneath the Valle de Santiago volcanic field. The lack of datable zircons in the mafic and anorthositic xenoliths unfortunately does not permit more precise age determinations of both the underplated gabbroic magmas and the high-grade metamorphism that transformed them into granulites. The intimate coexistence in the xenolith population of olivine-orthopyroxene gabbros and their gradual transformation to equivalent two-pyroxene–spinel mafic granulites strongly suggest a major tectonomagmatic event affecting the lower crust beneath the Valle de Santiago volcanic field in the late Cenozoic. Dating of the metabasites and their granulite-facies metamorphism is required to confirm the Neogene or Quaternary age of both crustal-forming events postulated in this work.

Financial support from PAPIIT-DGAPA (Programa de Apoyo a Proyectos de Investigación e Innovación Tenológica-Dirección General de Apoyo al Personal Académico) projects IN104706 to F. Ortega-Gutiérrez and IN107810 to Gómez-Tuena is gratefully acknowledged. This is also a contribution to project CB164454 kindly granted by CONACyT (Consejo Nacional de Ciencia y Tecnología) to F. Ortega-Gutiérrez. Carlos Ortega-Obregón, Ofelia Pérez-Arvizu (Centro de Geociencias, Universidad Nacional Autónoma de México [UNAM]), Gabriela Solís, Juan Julio Morales, and Teodoro Hernández-Treviño (Instituto de Geofísica, UNAM) are thanked for instrument maintenance and assistance during isotopic determinations. Special thanks are due to John Goodge, who did an exhaustive and critical revision of the original manuscript, suggesting substantial improvements on all aspects of this contribution.