Journal of Petrology Advance Access originally published online on August 12, 2008
Journal of Petrology 2008 49(9):1589-1618; doi:10.1093/petrology/egn039
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Volatiles in High-K Magmas from the Western Trans-Mexican Volcanic Belt: Evidence for Fluid Fluxing and Extreme Enrichment of the Mantle Wedge by Subduction Processes
1Department of Geological Sciences, University of Oregon, Eugene, OR 97403-1272, USA
2Department of Geosciences, Oregon State University, Corvallis, OR 97331-5506, USA
RECEIVED OCTOBER 1, 2007; ACCEPTED JULY 9, 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Primitive, high-K minettes and basanites erupted during the Pleistocene from cinder cones on the flanks of the Colima Volcanic Complex in the western Trans-Mexican Volcanic Belt. Melt inclusions in olivine (Fo89–92) from tephra at these cones reveal that both magma types are oxidized and volatile rich, with high H2O (
6·2 wt%), CO2 (5300 ppm), S (6700 ppm), Cl (2300 ppm), and F (8100 ppm) contents. A nearby calc-alkaline basaltic andesite cinder cone with more evolved composition (Fo78–80 olivine) has melt inclusions with similarly high H2O (5·5 wt%) but much lower CO2, S, and Cl compared with the potassic magmas. Melt inclusions from each cone have highly variable H2O and CO2, corresponding to crystallization pressures of < 100 bars to 
7 kbar. This indicates that olivine crystallized from variably degassed melts over a wide range of depths extending from the lower crust (>25 km depth) to very shallow levels. The H2O and CO2 variations cannot be explained by simple degassing models but instead requiring more complex, open-system processes or possibly reflect disequilibrium degassing. Trace element variations in the melt inclusions suggest that phlogopite and garnet were residual minerals during melting in the mantle source, and the presence of garnet suggests an origin in asthenospheric rather than lithospheric mantle. Decompression melting of phlogopite–garnet peridotite cannot produce the high H2O contents of the potassic magmas, and thus the presence of fluids during melting is required. Trace element modeling of a mantle source (intermediate in composition between enriched mid-ocean ridge basalt and ocean island basalt sources) that is fluxed with an H2O-rich fluid or hydrous melt from the subducting slab can reproduce most of the trace element characteristics of the potassic melts, demonstrating that they are clearly linked with subduction processes. Formation of the potassic magmas probably involved slab rollback, trenchward migration of the arc into the region above metasomatically enriched forearc mantle, and heating of this veined and fluid-fluxed mantle as a result of upwelling of hot mantle through a tear between the subducted Cocos and Rivera plates. KEY WORDS: melt inclusion; metasomatism; subduction zone; phlogopite; potassic magmas
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The coeval eruption of dominant calc-alkaline magmas and subordinate potassic magmas in subduction zones has been documented in a number of arcs worldwide, including the northern Mariana arc (Bloomer et al., 1989
), the Aeolian Islands (Ellam et al., 1988), the SW Japan arc (Tatsumi & Koyaguchi, 1989), the Sunda arc (Nicholls & Whitford, 1983), Kamchatka (Koloskov et al., 1999), and the western Trans-Mexican Volcanic Belt (TMVB; Luhr et al., 1989). The occurrence of potassic magmas in arc environments has been correlated with tectonic extension (e.g. Morrison, 1980; Nicholls & Whitford, 1983; Luhr et al., 1985) and this is the case for the western TMVB, where most of the potassic lavas have erupted within grabens (Allan & Carmichael, 1984; Luhr et al., 1989; Wallace & Carmichael, 1989; Lange & Carmichael, 1991). However, there is debate as to whether the extensional tectonics causes melting of the underlying mantle or if extensional faults simply create ascent paths for the small volume melts (Luhr et al., 1989; Carmichael et al., 1996; Hochstaedter et al., 1996). Despite considerable progress, we still have a relatively poor understanding of the mantle source of potassic magmas in subduction zones and how they relate to the volumetrically dominant calc-alkaline magmas.
Previous studies on potassic rocks in the TMVB suggest that the magmas formed by melting of phlogopite–clinopyroxene-rich veins within metasomatized mantle peridotite resulting in primitive, low-degree partial melts that are highly enriched in incompatible elements (Luhr et al., 1989; Carmichael et al., 1996; Luhr, 1997). However, there is still uncertainty as to whether these magmas are the result of decompression melting or melting caused by the addition of fluids from the subducted slab (fluid-fluxing), and where in the mantle (lithosphere vs asthenosphere) these veins are located (Luhr et al., 1989; Hochstaedter et al., 1996; Luhr, 1997; Ferrari et al., 2001; Hesse & Grove, 2003). To determine the pre-eruptive volatile content and oxidation state of the primitive potassic magmas as well as gain insight into their mantle source mineralogy, composition and melting processes, we have analyzed major, trace and volatile elements (H2O, CO2, S, Cl and F) in melt inclusions hosted in Fo-rich olivine phenocrysts from four potassic and one calc-alkaline cinder cone in the Colima graben of the western TMVB.
|
TECTONIC SETTING |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The TMVB is an east–west-trending arc active since the Miocene and related to subduction of the Cocos and Rivera plates along the Middle-American Trench (Fig. 1). Subduction of the young (
10 Ma) Rivera microplate is associated with volcanism in the western TMVB (Fig. 1) (Nixon, 1982). The Jalisco block is structurally defined by the Tepic–Zacoalco rift to the north, the Colima graben to the east, and the Middle-American Trench to the south. The Colima Volcanic Complex is located in the Colima graben, 175 km from the Middle-American Trench. The Colima graben has experienced significant faulting (>1· 5 km vertical offset; Allan, 1986), and crustal thickness beneath this region is 25–30 km (Urrutia-Fucugauchi, 1986; Urrutia-Fucugauchi & Flores-Ruiz, 1996).
|
The three stratovolcanoes that define the Colima Volcanic Complex are Volcan Cantaro (farthest north), Volcan Nevado de Colima (central) and the currently active Volcan Fuego de Colima (farthest south). Volcanism progressed southward within the Colima graben during the Quaternary (Luhr & Carmichael, 1981
). The volcanic front in the western TMVB as a whole migrated trenchward from the late Miocene to the Quaternary (Ferrari et al., 2001; L. Ferrari, personal communication, 2008), although the geological evidence for this trenchward migration and the subduction relationship of the Miocene volcanism has been disputed (Frey et al., 2007). Migration of the volcanic front may have been caused by steepening of the dip (slab rollback) of the Rivera Plate (Ferrari et al., 2001). Recent results from seismic tomography (Grand et al., 2007; Yang et al., in preparation) show a steep dip angle on the Rivera Plate (70°) and a gap between the Rivera and Cocos plates through which hot mantle may be flowing upwards, as suggested previously by Ferrari et al. (2001). |
POTASSIC MAGMAS IN THE WESTERN TMVB |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Small volume eruptions of potassic magma have occurred in a number of regions along the volcanic front of the western TMVB from
5 to 0·06 Ma (Allan & Carmichael, 1984; Wallace & Carmichael, 1989, 1992; Lange & Carmichael, 1991; Righter & Carmichael, 1992; Carmichael et al., 1996, 2006; Righter & Rosas-Elguera, 2001). Eleven cinder and lava cones erupted on the floor of the Colima graben during the late Pleistocene, NE and NW of volcanoes Nevado de Colima and Fuego de Colima (Fig. 1), and nine of these cones include high-K2O, low-SiO2 (nepheline-normative) lavas and scoria (Luhr & Carmichael, 1981; Carmichael et al., 2006). These cones are compositionally distinct (Fig. 2) from both the medium-K andesites that form Volcan Fuego de Colima and the other stratovolcanoes (e.g. Tequila, Ceboruco) to the NW and the medium-K basalt to andesite cones that dominate the Michoacan–Guanajuato Volcanic Field (MGVF; see Fig. 1) to the east (Hasenaka & Carmichael, 1985; Luhr, 1992).
|
Luhr & Carmichael (1981
) classified lava and scoria from the nine potassic cinder cones as minettes (phenocrysts of olivine, augite and phlogopite), basanites (phenocrysts of olivine and augite ± plagioclase), and leucite basanites (groundmass leucite ± phlogopite), and they noted that this last type was transitional between the first two. The minettes and basanites are similar in major element composition, but the basanites have slightly higher Al2O3 and CaO and slightly lower Na2O and P2O5 compared with the minettes. The remaining two cinder cones are calc-alkaline in composition. One is a basaltic andesite and the other is a primitive, high-alumina basalt (CAB in Fig. 2); both contain phenocrysts of olivine and plagioclase.
The nine potassic cinder cones erupted a total of 1·3 km3 of magma mostly between 240 and 60 ka (Carmichael et al., 2006). The two calc-alkaline cinder cones erupted <0·003 km3 of magma each and are much older (basaltic andesite, 1· 2 Ma; high-alumina basalt, 0·5 Ma). These dates indicate that the potassic magmatism coincided with eruption of calc-alkaline andesite at Nevado de Colima and the initial building stage of Volcan Fuego de Colima (Carmichael et al., 2006).
Potassic lavas erupted elsewhere within the Colima graben from 4·6 to 0·6 Ma and are almost exclusively confined to the graben faults (Allan & Carmichael, 1984). These lamprophyres (silica-undersaturated magmas with either amphibole or phlogopite phenocrysts) are mostly older and more evolved equivalents of the Colima cinder cone minettes, and they are spatially and temporally associated with calc-alkaline volcanic rocks. Other volcanic rocks exposed in the graben walls include high-K andesites. In the neighbouring Jalisco block (Fig. 1), a diverse array of potassic magmas has erupted coevally with calc-alkaline magmas for the past 5 Ma, including minettes, leucitites, and lower-K2O absarokites and hornblende trachybasalts (Luhr et al., 1989; Wallace & Carmichael, 1989, 1992; Lange & Carmichael, 1991; Carmichael et al., 1996; Luhr, 1997; Righter & Rosas-Elguera, 2001).
|
SAMPLES AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Volcanic rocks from the five cinder cones that we studied have been previously described and analyzed by Luhr & Carmichael (1981
) and Carmichael et al. (2006). Two of the cinder cones are minettes (Carpintero Norte and La Erita; we classify the latter as a minette because of the presence of phlogopite microphenocrysts), two are basanites [Apaxtepec and Cerro Colorado; the latter was referred to as Cerro Cuauhatemoc by Luhr & Carmichael (1981) and Carmichael et al. (2006)], and one is a calc-alkaline basaltic andesite (Cerro Usmajac). All the cinder cones that we sampled have been quarried, which allowed access to the interiors of the cones and exposed relatively extensive stratigraphic sections of the eruptive products. Care was taken to sample the least altered or oxidized tephra from fine-grained layers. Sampling of the finest material (ash to small lapilli) is advantageous because olivine crystals have been naturally separated from the glass and therefore no crushing is necessary.
Major and trace elements in bulk tephra samples (Table 1) from three of the cinder cones were analyzed by X-ray fluorescence (XRF) and inductively coupled plasma-mass spectrometry (ICP-MS) in the GeoAnalytical Laboratory at Washington State University. Detailed discussion of analytical techniques and sample preparation has been given by Knaack et al. (1994) and Johnson et al. (1999).
|
Olivine crystals were handpicked from tephra samples using a binocular microscope and were placed in immersion liquid (refractive index 1· 678) to select crystals with melt inclusions for analysis. The crystals with the largest and best preserved inclusions were separated, and morphological and textural characteristics were noted before sectioning. To prepare samples for quantitative infrared (IR) spectroscopic analysis, doubly polished wafers were prepared from melt inclusion-bearing crystals so that both faces of the wafer intersected the inclusion.
H2O and CO2 contents of the melt inclusions were determined using Fourier transform IR spectroscopy (FTIR) at the University of Oregon (Table 2). We used the quantitative procedures and band assignments described by Dixon et al. (1995). The IR aperture size varied from 10 to 100 µm in width depending on the size of the inclusion. The thickness of the wafers varied from 10 to 80 µm and was determined by mounting the wafers on a needle and immersing them in refractive index oil so they could be viewed under a microscope at 100 x magnification. The wafers were tilted so that they could be viewed edgewise, and this allowed for a cross-section view of the inclusion and measurement of its thickness using the objective reticle. Using this method, we measured the thickness of the inclusion precisely where the FTIR measurement was performed and the error in our thickness measurements varied from ±1 µm to ±5 µm depending on the geometry of the inclusion and wafer.
|
Quantitative measurements of dissolved total H2O and carbonate (CO32–) were determined using the Beer–Lambert law. Total dissolved H2O was measured using the intensity of the broad, asymmetric band centered at 3570 cm–1, which corresponds to the fundamental OH-stretching vibration. Total H2O concentrations were calculated from measured peak heights using an absorption coefficient of 63 ± 3 L/mol cm (P. Dobson et al., unpublished data, cited by Dixon et al., 1995
). Most absorbance values at the 3570 cm–1 band were <1, but for inclusions with values >1, peak shapes indicated that the IR detector was not saturated.
Dissolved carbonate was measured from the absorbance value of the bands at 1515 and 1430 cm–1, which correspond to antisymmetric stretching of distorted carbonate groups (Dixon et al., 1995). Because the shape of the background in the region of the carbonate doublet is complex, it is necessary to subtract a carbonate-free reference spectrum to obtain a flat background. A peak-fitting program that fits the sample spectrum with a straight line, a devolatilized spectrum, a pure 1630 cm–1 band for molecular H2O, and a pure carbonate doublet was used (unpublished program by S. Newman). Based on fitting of replicate spectra for all inclusions, one standard deviation uncertainties in the carbonate peak heights are typically ± 10%. The molar absorption coefficient of carbonate in basaltic glass is compositionally dependent. We calculated an absorption coefficient for each inclusion using the equation of Dixon & Pan (1995). Room temperature glass densities for the inclusions were calculated using the method described by Luhr (2001).
Major and volatile (S, Cl, F) element compositions of the melt inclusions, sulfur K
wavelength shifts and major element compositions of the groundmass glasses, olivine host crystals, and Cr-spinel inclusions in olivine were determined using a Cameca SX-100 electron microprobe at the University of Oregon (Tables 2 and 3). Details of the analytical conditions and standards that were used are given in the legend to Tables 2 and 3. Details of the method used for measuring the sulfur K wavelengths have been given by Wallace & Carmichael (1994). Three melt inclusions from each cone were analyzed. Prolonged exposure to the beam causes an apparent increase in sulfur oxidation state for glasses with a significant proportion of S as S2– (Wallace & Carmichael, 1994), and this effect is greatest in glasses with low S contents (Rowe et al., 2007). For highly oxidized glasses, exposure to the electron beam (or to intense X-rays during X-ray absorption near edge structure (XANES) spectroscopy) can cause partial reduction that converts S6+ to S4+ (Wilke et al., 2008). To minimize these effects, we moved the sample relative to the beam every 20 s during the scan. Peak positions were converted to sulfur speciation (Table 2) using the linear relationship S6+/Stotal = (
SKunknown/SKanhydrite) x 100 defined by Carroll & Rutherford (1988).
Trace element compositions of at least three melt inclusions from each cinder cone were measured by laser ablation inductively-coupled plasma mass spectrometry (LA-ICP-MS) in the W. M. Keck Collaboratory for Plasma Mass Spectrometry at Oregon State University using a DUV 193 nm ArF Excimer laser and a VG ExCell quadrupole ICP-MS system. A general outline of the analytical techniques used for this instrument has been given by Kent et al. (2004), and analytical details for our samples are described in the legend to Table 2.
Melt inclusion compositions were corrected for both post-entrapment crystallization of olivine along the walls of the inclusion and Fe loss by diffusion between the inclusion and the host crystal (Danyushevsky et al., 2000). The post-entrapment crystallization correction involved adding 0·1% increments of equilibrium olivine back into each melt inclusion until the inclusion composition was in equilibrium with the olivine host. The exchange coefficient [KD = (Mg/Fe2+)melt/(Mg/Fe2+)olivine] was calculated for each melt inclusion using the equation of Toplis (2005), which relates KD to temperature, pressure, melt composition (including H2O content) and olivine composition (Table 2). Correction for Fe loss involved adding Fe2+ back into each melt inclusion until the FeOT of the inclusion was restored to the MgO vs FeOT trend of whole-rock and groundmass glass data from each cone (see Danyushevsky et al., 2000, for details). All major, trace and volatile element data (Table 2) described in the text and shown in figures are corrected values. Analyzed (uncorrected) values for all inclusions are reported in Electronic Appendix 1 (available for downloading at http://www.petrology.oxfordjournals.org).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Corrected compositions of melt inclusions and olivine host crystals
Olivine hosts for the melt inclusions have core compositions that are similar in both basanites (Fo88–91) and minettes (Fo89–92). The MgO contents of both basanite and minette melt inclusions vary from 7·3 to 11· 1 wt%. Although there is considerable overlap in the major element compositions of the melt inclusions, the minettes on average have higher SiO2 and K2O and lower Al2O3, CaO and FeO than the basanites for a given MgO content (Figs 2 and 3). Within each group, Al2O3 shows a negative correlation with MgO whereas CaO/Al2O3 shows a positive correlation. These variations are consistent with olivine + augite as the major fractionating phases. In the minette melt inclusions, CaO/Al2O3 changes more rapidly with decreasing MgO than it does in the basanite inclusions, suggesting a greater proportion of augite fractionation in the minette melts. The pattern of increasing K2O with decreasing MgO for inclusions from the La Erita minette is consistent with this interpretation, whereas the Apaxtepec basanite inclusions show little to no increase in K2O over the same range in MgO. Differences in composition between the melt inclusions and whole-rock samples (Fig. 3) probably reflect the effects of accumulation of olivine plus included chromite. Whole-rock samples show strong positive correlation of modal olivine with MgO, Ni, and Cr (Luhr & Carmichael, 1981
).
|
Compared with the potassic samples, melt inclusions from the calc-alkaline basaltic andesite (Cerro Usmajac) have higher SiO2 and Al2O3, and lower MgO, CaO, K2O and TiO2. These inclusions are trapped in more Fe-rich olivine (Fo78–80) than in the postassic samples. Luhr & Carmichael (1981
) noted that whole-rock samples from Cerro Usmajac were enriched in TiO2, Na2O, K2O, and P2O5 compared with the trends for calc-alkaline lavas from Volcan Colima and another nearby basaltic cinder cone (Tezontal). We also find higher concentrations of these elements as well as Nb and Ta in the Cerro Usmajac melt inclusions than are typical of calc-alkaline lavas from either the Colima graben or the MGVF to the east (Fig. 1; Johnson, 2008). In particular, the much higher K2O content of Cerro Usmajac inclusions (1· 5–2· 4 wt%) compared with other calc-alkaline lavas with similar MgO contents (e.g. mostly 0·5–1·2 wt% K2O in the MGVF) makes Cerro Usmajac transitional to the potassic group, whereas the higher Nb and Ta make it transitional to the intraplate alkaline lavas that have erupted elsewhere in the TMVB (Luhr, 1997).
The potassic melt inclusions are strongly enriched in highly and moderately incompatible trace elements compared with the basaltic andesite and the primitive calc-alkaline basalts from the MGVF (Fig. 4). However, the potassic melts have the relative depletion in Nb and Ta that is a typical characteristic of subduction-related magmas (Luhr et al., 1989). Within the potassic group, melt inclusions from the minettes are generally more enriched in incompatible elements than are those in basanites. Enrichments in the different elements are strongly correlated (e.g. Ba, Rb, Zr, Ce show strong positive correlations with Sr). In contrast, heavy REE (HREE) concentrations are similar in potassic and calc-alkaline melt inclusions, and Yb shows no corrrelation with Sr.
|
Based on whole-rock trace element data, Luhr & Carmichael (1981
) found a positive correlation between P2O5 and most incompatible elements, and their data showed that minettes were more enriched than basanites. Our data show a much greater amount of overlap in the P2O5 contents of basanite and minette melt inclusions, although the basanites do have lower average P2O5. Luhr & Carmichael (1981) also noted that samples from Apaxtepec had anomalously high concentrations of Ti, Zr, and Hf compared with the other basanites. Our melt inclusion data show similar anomalously high values of Ti, Zr, and also Rb in Apaxtepec melt inclusions compared with inclusions from the other basanite cone (Cerro Colorado). Amongst compatible trace elements, melt inclusion Sc shows strong postive correlation with CaO/Al2O3, consistent with the importance of augite fractionation discussed above.
The CaO/Al2O3 ratios of most melt inclusions are similar, within analytical uncertainty, to whole-rock analyses of bulk tephra from their respective cones (Fig. 3; note overlap between whole-rock and melt inclusion CaO/Al2O3 values, but whole-rock data are shifted to higher MgO because of olivine addition). Both CaO and Al2O3 are incompatible in olivine, but Al2O3 diffuses much more slowly in silicate melts than CaO. Thus, the similarity of the melt inclusion and whole-rock CaO/Al2O3 ratios suggests that the melt inclusion compositions are representative of the bulk melt from which the olivine crystallized and were not affected by disequilibrium processes (boundary-layer concentration build-up) during entrapment (Faure & Schiano, 2005). Similarly, the ratio Cl/P2O5 is useful because both are incompatible in olivine but P2O5 is a very slowly diffusing component in silicate melts (Baker, 2008). We find no systematic variation of Cl/P2O5 with inclusion size, and for all but one of our studied cones, Cl/P2O5 varies less than 10% relative within each suite of melt inclusions, suggesting that the inclusions were not affected by boundary-layer enrichment processes (Baker, 2008). It should be noted that use of the Cl/P2O5 ratio is predicated on there being no degassing of Cl, which we show to be the case (see below). The only cinder cone where the Cl/P2O5 variability exceeds 10% relative is La Erita (20%). We note that if inclusion LE-02-04 is removed from the dataset, the variability drops to 10%; this inclusion has a significantly more evolved composition compared with the rest of the suite (Table 2).
Volatile contents
For the basanites, H2O contents vary from 0·4 to 6·2 wt% for Apaxtepec melt inclusions and from 0·4 to 3·4 wt% for Cerro Colorado (Fig. 5). The CO2 contents vary from below detection (50–100 ppm depending on inclusion thickness) to 5300 ppm and show strong linear correlation with H2O. Water contents in the minettes vary from 0·9 to 5·5 wt% in La Erita melt inclusions and from 0·5 to 1· 6 wt% in Carpintero Norte. The CO2 contents of the minette melt inclusions range from below detection to 3000 ppm, and there is much more scatter in the data compared with the basanite melt inclusions. The basaltic andesite melt inclusions have H2O contents from 1· 0 to 5·5 wt% and CO2 contents from below detection to 800 ppm. Compared with the H2O and CO2 contents of the basaltic andesite (Fig. 5) and the primitive calc-alkaline magmas from the MGVF (Johnson, 2008), the potassic magmas generally have higher CO2 for a given H2O content. The maximum CO2 contents of the potassic melt inclusions are much higher than for mafic calc-alkaline magmas from arcs worldwide (Wallace, 2005) and are amongst the highest values reported for olivine-hosted melt inclusions. Maximum H2O contents for the potassic melt inclusions are similar to those of the calc-alkaline inclusions, and these values (5–6·2 wt% H2O) lie near the higher end of values for mafic calc-alkaline magmas worldwide (Wallace, 2005). Our data show that both the basanites and minettes have initially high H2O contents, suggesting that the absence of phlogopite in the basanites is caused by some other factor, such as their lower K2O. However, we note that the basanite melt inclusions have higher K2O than is necessary for H2O-saturated melts to crystallize phlogopite in experimental studies (Esperança & Holloway, 1987; Righter and Carmichael, 1996), and both basanite and minette melt inclusions have similar CO2/H2O, so it is unclear what factors caused the absence of phlogopite in the basanites.
|
compositional parameter for the average melt inclusion composition, and then found the corresponding SiO2 value based on the Sulfur contents are relatively high in both basanite and minette melt inclusions, with most values between 0·2 and 0·67 wt% (Fig. 6). As with H2O and CO2, there are significant variations in S contents for inclusions from each cone, and overall there is a positive correlation between S and H2O, although with considerable scatter. Chlorine contents are also similar in both basanite and minette melt inclusions, with average values of 0·18 wt% (basanite) and 0·19 wt% (minette). Compared with the S data, there is less scatter, and Cl remains relatively constant (or even slightly increases) with decreasing H2O (Fig. 6). Fluorine contents average 0·45 wt% in the basanites and 0·64 wt% in the minettes, and F contents remain high even at low values of H2O. The ranges of F content for minette and basanite melt inclusions do not overlap as much as for S and Cl. For a given H2O content, the Colima potassic magmas are enriched in S, Cl, and F compared with both the basaltic andesite cinder cone Cerro Usmajac and the primitive calc-alkaline cinder cones of the MGVF (Fig. 6).
|
The S6+/Stotal values for the minettes range from 0·63 to 0·91, and for the basanites from 0·61 to 0·86. For the basaltic andesite, the range is 0·78–0·94. Because oxidized S is more soluble than reduced S (e.g. Luhr, 1990
), the predominance of oxidized S in all samples is consistent with the relatively high total S contents of the melt inclusions. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Pressures and temperatures of olivine crystallization
Pressures of olivine crystallization can be inferred from melt inclusions by using the pressure-dependent solubilities of H2O and CO2 in silicate melts and assuming that the melts were saturated with an H2O–CO2 vapor phase (Table 2 and Fig. 5). This assumption is predicated on the wide range of H2O and CO2 contents for inclusions from a given cone, which shows that the melts were affected by variable degassing before crystallization and therefore must have been vapor saturated. For the basanite melt inclusions, trapping pressures are in the range of 1–7·5 kbar (Apaxtepec) and 0·02–2·7 kbar (Cerro Colorado). The minette cones have melt inclusions that record pressures of 0·1–6·2 kbar (La Erita) and 0·03–0·5 kbar (Carpintero Norte). The accuracy of these pressure estimates is difficult to assess because most of the calibration data for the VolatileCalc program are from relatively low pressures (Newman & Lowenstern, 2002
). A comparison of pressures for the minette and basanite inclusions calculated using both VolatileCalc and the method of Papale et al. (2006) suggests that the VolatileCalc pressures are accurate to ± 250 bars for pressures 1 kbar and ±800 bars for pressures of 1–8 kbar.
Our pressure estimates are minimum values because they do not take into account CO2 lost into shrinkage bubbles as a result of post-entrapment thermal contraction of melt and inclusion crystallization. Smaller vapor bubbles (10 vol.% of the inclusion) are probably shrinkage bubbles formed by differential thermal contraction of the melt and host crystal, whereas larger bubbles are likely to be primary vapor bubbles trapped together with melt. In both cases, some of the original dissolved CO2 in the inclusion will have been lost to the vapor bubble during post-entrapment cooling. The amount of CO2 lost to the bubble is difficult to quantify (e.g. Anderson & Brown, 1993), but original trapping pressures for some inclusions could be as much as 1–2 kbar higher than the pressures we have estimated.
Based on the pressures estimated above, the range in crystallization depths for Apaxtepec and La Erita extends from 
25 km to near the surface (assuming an average crustal density of 2800 kg/m3; Christensen & Mooney, 1995). Cerro Usmajac, the basaltic andesite cinder cone, has melt inclusions that record olivine crystallization pressures of 0·1–4·1 kbar (0·4–14 km depth). Overall, the melt inclusion data suggest that olivine crystallized over a wide range of depths, probably in response to decompression and water loss during ascent (e.g. Sisson & Layne, 1993; Roggensack, 2001; Johnson et al., 2008).
Righter & Carmichael (1996) made estimates of the crystallization pressures of phlogopite in the Colima minettes based on partitioning of Ba and Ti between phlogopite and melt. They calculated values from 4 to 31 kbar, but most values are between 4 and 10 kbar, a range that is higher than but partially overlapping with our pressure estimates for olivine-hosted melt inclusions. Most of this discrepancy is probably caused by the relatively large uncertainty of the phlogopite geobarometer (± 4 kbar standard error), but, as noted above, our data are minimum values for trapping pressures because we did not include the CO2 lost to bubbles in our pressure estimates. Additionally, the melt-inclusion-bearing olivines may not be representative of the olivine population as a whole, either because of selective preservation (higher pressure inclusions ruptured during ascent) or because conditions at higher pressures were not conducive to melt inclusion formation.
To obtain estimates of the trapping temperatures of the melt inclusions we have used the olivine–melt geothermometer of Sugawara [2000; equation (6a), which explicitly includes the effects of dissolved H2O] and the corrected melt inclusion compositions (Table 2). We found that this thermometer works well for hydrous potassic melts based on a comparison with temperatures from an experimental study of minette magmas (Righter & Carmichael, 1996). With one exception, the potassic melt inclusions record crystallization temperatures in the range of 1150–1228°C (Fig. 7). Nearly all of the inclusions appear to lie at temperatures above the liquidus and the phlogopite stability field for H2O-saturated minette melts, underscoring the fact that H2O-undersaturated phase relations must be considered when interpreting data for CO2-bearing melt inclusions (Johnson et al., 2008). For example, it has been shown experimentally that CO2 can enhance the stability of hydrous phases such as amphibole (Eggler, 1972). The most H2O-rich and highest pressure inclusions were trapped at 6·2–7·5 kbar pressure and apparent temperatures of 1150–1162°C. The rest of the inclusions were trapped at 5 kbar pressure and record a range of temperatures that corresponds to the range of dissolved H2O contents. Given the uncertainty in the geothermometer (±30°C) and the additional uncertainty created by our olivine-addition and Fe-loss corrections to the melt inclusions, it is difficult to assess whether the apparent overall increase in temperature with decreasing pressure is real or an artefact.
|
Degassing during ascent
The potassic melt inclusions are characterized by high CO2 for a given H2O content compared with calc-alkaline basalts and basaltic andesites in the Colima graben and nearby MGVF (Fig. 5). Equilibrium closed- or open-system degassing models for ascending and decompressing melts fail to reproduce the trends observed; in particular, the linearity of the degassing trend of Apaxtepec (Fig. 8). Similar but smaller discrepancies between melt inclusion H2O–CO2 data and simple degassing models have been seen in other systems (e.g. Atlas et al., 2006
; Spilliaert et al., 2006a; Johnson et al., 2008). A possible cause of the discrepancy is that during ascent melts equilibrated with more CO2-rich vapor fluxing through the plumbing system from below, where it was released by magma degassing deeper in the system (Rust et al., 2004; Wallace et al., 2005; Spilliaert et al., 2006a; Johnson et al., 2008). Isopleths of constant vapor composition (50–80 mol% CO2) simulate the effect of large amounts of gas flushing through the system such that the CO2/H2O ratio of the melt is partially buffered during decompression and degassing (Fig. 8). For both basanites and minettes, the vapor isopleths bracket many of the data, suggesting that CO2-rich vapor fluxing may be an important process. However, the gas fluxing hypothesis does not seem to adequately explain the remarkably linear correlation between H2O and CO2 for Apaxtepec.
|
An alternative possibility is that CO2 solubility in the basanite melt of Apaxtepec is significantly higher than predicted by the model of Dixon (1997
), which relates CO2 solubility to major element composition. This might happen, for example, if high dissolved H2O caused an increase in CO2 solubility (e.g. Mysen, 1976) relative to anhydrous melts, on which the Dixon (1997) model is mostly based. If CO2 had similar solubility to H2O, then degassing would result in fairly constant CO2/H2O. However, calc-alkaline cones from the MGVF with similarly high H2O do not show such linear behavior, suggesting that any potential effect of water on CO2 solubility would be insufficient to cause the correlated data. Another possibility is that the magmas interacted with carbonate wallrocks as they ascended through the crust. Late Jurassic to early Tertiary limestone is present on the floor of the Colima graben surrounding Apaxtepec (Allan, 1986) and could be extensive at depth, and interaction with the CaCO3 might have buffered the CO2/H2O of melts as they ascended through the upper crust. However, there is no evidence from Sr isotope data that such interaction has taken place (Luhr, 1997) although the high concentrations of Sr and other trace elements in the basanites might have masked any assimilation signature. A third possibility is that non-equilibrium degassing during rapid ascent causes melts to retain CO2 because it diffuses towards bubbles more slowly than H2O (Gonnermann & Manga, 2005). Unfortunately, there are no experimental data on diffusivity of CO2 in hydrous potassic mafic melts to evaluate this possibility (Baker et al., 2005). Regardless of the specific degassing processes involved, the wide range of H2O and CO2 in the melt inclusions indicates that the potassic melts were vapor saturated during ascent through the crust, degassed as they ascended, and crystallized olivine over a wide range of depths.
We have also considered whether post-entrapment diffusive loss of H2 or H2O through the melt inclusion host crystals may have lowered the inclusion H2O contents. Loss of H2 is probably limited to 1 wt % H2O by redox effects (Danyushevsky et al., 2002) and should cause strong oxidation of the inclusions (e.g. Mathez, 1984; Rowe et al., 2007), which we do not observe in the S K data. Diffusive loss of molecular H2O (Portnyagin et al., 2008) is not limited by redox effects, but it is unclear how such loss would cause the highly correlated H2O and CO2 observed for Apaxtepec. We also contend that diffusive loss of large amounts (several wt %) of H2O should cause crystallization of abundant daughter minerals in the inclusions, which we do not observe. Thus, although we cannot rule out some diffusive loss of H species from the inclusions, we conclude that such loss is not the primary process responsible for the large range of H2O contents and that this range primarily reflects variable degassing before inclusion entrapment.
Sulfur shows evidence of degassing as well, as there is a weak, positive correlation between S and H2O for the basanite and minette melt inclusions (Fig. 6). Similar weak positive correlations are seen in plots of S vs CO2 and S vs trapping pressure (not shown). The large degree of scatter in the data indicates that either there is a range of primary S concentrations for potassic melts from a given cinder cone or degassing involves more complex, poorly understood processes as discussed above for the H2O and CO2 data. Assuming that degassing can be described by a Rayleigh fractionation process, the S vs H2O variations for minette and basanite melt inclusions are consistent with DSvapor/melt values of 5–20 (Fig. 6). These values are lower than estimates for some subduction-related mafic magmas—D = 34 for Fuego volcano in Guatemala (Sisson & Layne, 1993) and D = 40 for Stromboli in Italy (Métrich et al., 2001)—but are similar to values for Etna (Spilliaert et al., 2006b) and Paricutin and Jorullo in the MGVF (Johnson, 2008). The lower DSfluid/melt values may partly be the result of higher pressure degassing recorded by the melt inclusions, because higher pressure favors retention of S in the melt (Johnson, 2008). Lower DSfluid/melt values may also be caused by higher S solubility in high fO2 magmas (see next section), which makes S partition less strongly into the vapor phase during degassing.
In contrast to S, Cl does not appear to degas in the potassic magmas (calculated DClfluid/melt 
1). For comparison, values of DClfluid/melt estimated for other volcanoes based on melt inclusion data are 2·8 for Fuego (Sisson & Layne, 1993), 1 for Etna (Spilliaert et al., 2006b), and 1 for Paricutin and Jorullo (Johnson, 2008). As with Cl, F in the potassic magmas does not appear to degas as H2O is lost from melt (Fig. 6). The behavior of both Cl and F is consistent with the relatively high solubility of these components in silicate melts at low pressure that has been found experimentally (Carroll & Webster, 1994).
Oxygen fugacity
Oxygen fugacities can be estimated from S6+/Stotal ratios in melt inclusions using calibrations based on experimental glasses (Wallace & Carmichael, 1994; Matthews et al., 1999; Jugo et al., 2005). To evaluate the applicability of these calibrations to potassic glass compositions, we obtained independent estimates of the oxygen fugacity of the potassic magmas using a method based on the Fe3+/Fetotal ratios of Cr-spinel in equilibrium with olivine (Table 3; Ballhaus et al., 1991). Results of the Cr-spinel fO2 calculations for Apaxtepec, La Erita and Carpintero Norte range from NNO + 0·8 to NNO + 1· 7 (where NNO is the nickel–nickel oxide buffer) and are plotted along with individual melt inclusion S6+/Stotal values for each cinder cone in Fig. 9a. The data fall mostly within error of the Wallace & Carmichael (1994) and Jugo et al. (2007; see Fig. 9 caption) curves relating S6+/Stotal to fO2, suggesting that these calibrations are appropriate for potassic glasses.
|
S) in single melt inclusions vs relative oxygen fugacity expressed as 
uncertainties in the peak position of the S K
S ratios as in (a).
|
Average fO2 values for each cone calculated from melt inclusion S6+/Stotal values are in a narrow range from NNO + 0·9 to NNO + 1· 3 for the potassic magmas and NNO + 1· 2 for the basaltic andesite (Fig. 9b). Compared with estimates from Fe3+/Fe2+ ratios in whole-rock samples (Carmichael et al., 2006
), our estimates of fO2 show excellent agreement for Apaxtepec and the Cerro Usmajac basaltic andesite, but for the other potassic cones our values are lower by 2·0–2·5 log units. Our results based on Fe3+/Fetotal in Cr-spinels and S6+/Stotal in melt inclusions thus suggest that some minette and basanite whole-rock samples may have been affected by oxidation. Our results are surprising because other methods, such as Fe3+/FeOT in phlogopite, are consistent with relatively high fO2 in at least some minette magmas (Feldstein et al., 1996: Righter & Carmichael, 1996). Clearly, more intercalibration of the various techniques for determining fO2 in phlogopite-bearing magmas is warranted.
Mantle source mineralogy for the potassic magmas
Insight into the mineralogy of the mantle source for the potassic magmas can be gained from variations in the abundances of certain elements (Fig. 4). Differences between the minettes and basanites in elements such as K, P, Ti, Nb and Ta are smaller than the differences in other elements (e.g. La, Pb). This suggests that K, P, Nb and Ta are less affected by the degree of partial melting or subduction-related enrichment processes than are other incompatible elements. The presence of metasomatic minerals with high partition coefficients for these elements would have the effect of buffering their concentrations during melting. The most likely candidates to host these elements in metasomatically enriched peridotite are phlogopite for K, apatite for P, and rutile and/or ilmenite for Ti, Nb and Ta (Haggerty, 1995; Grégoire et al., 2002).
A plot of La/K vs La (Fig. 10a) illustrates the control of residual phlogopite on K during partial melting (Feldstein & Lange, 1999). La is incompatible in all likely residual mantle phases, whereas saturation with phlogopite causes melts to have relatively high and constant K (Esperança & Holloway, 1987; Thibault et al., 1992; Righter & Carmichael, 1996; Elkins-Tanton & Grove, 2003). As the degree of partial melting increases (La decreases), buffering of melt K contents by phlogopite results in a decrease of the La/K ratio, which otherwise would remain constant during the melting of a peridotitic mineral assemblage. In addition, the K2O contents of the minette and basanite melt inclusions (3·5–7 wt%) are comparable with the values of experimental melts saturated with phlogopite (e.g. Esperança & Holloway, 1987; Righter & Carmichael, 1996). Based on this evidence, we conclude that the minette and basanite magmas were formed in equilibrium with residual phlogopite.
|
A similar plot of La/Ti vs La (not shown) suggests the presence of residual rutile or ilmenite. However, the TiO2 contents of the minette and basanite melt inclusions are relatively low compared with values required for rutile and ilmenite saturation (Ryerson & Watson, 1987
; Thy et al., 2006). No experimental phase equilibrium studies on potassic mafic magmas at crustal or upper mantle pressures have found rutile on the liquidus (Elkins-Tanton & Grove, 2003), and ilmenite has been found only in a few experiments with silica-rich minette melts at below liquidus conditions (57–59 wt % SiO2, 10–15 kbar, temperature 1100°C; Esperança & Holloway, 1987). Therefore, it seems unlikely that either rutile or ilmenite could have been a residual phase during partial melting. Instead, we suggest that mantle phlogopite, with its relatively high TiO2 content (Esperança & Holloway, 1987; Wilkinson & LeMaitre, 1987; Righter & Carmichael, 1996) is responsible for buffering of this element. However, phlogopite in equilibrium with Ne-normative melts does not have high enough partition coefficients for Nb and Ta (Foley et al., 1996) to explain the limited variations in these elements compared with other trace elements of similar incompatibility (Fig. 4), so this characteristic of the minette and basanite melts remains unexplained.
With regard to residual apatite, the P2O5 concentrations of the minette and basanite melt inclusions (mostly 1· 0–1· 5 wt%) are all substantially less than values required for apatite saturation (3–4 wt%) in Ne-normative melts at temperatures >1150°C (Green & Watson, 1982). The experimental study of Righter & Carmichael (1996) on minette melts found apatite saturation values 0·5 wt% P2O5 lower than those found by Green & Watson (1982) at temperatures of 1050–1100°C, and a smaller temperature dependence, possibly because of the presence of F. Even if the Righter & Carmichael (1996) values at 1050–1100°C are appropriate for the higher temperature (1160–1220°C) Colima minette and basanite magmas (which we consider unlikely), only about half of the minette melt inclusions would have P2O5 values high enough for apatite saturation, and the rest of the minettes and all of the basanite melt inclusions have lower values. Thus despite the limited variations of P2O5 compared with other trace elements of similar incompatibility (Fig. 4), we conclude that the minette and basanite magmas were not saturated with residual apatite during partial melting.
Strong enrichment of light REE (LREE) (La) relative to HREE (Yb) in the minettes and basanites (Fig. 10b) provides evidence for mantle melting within the garnet stability field (Luhr, 1997; Righter & Rosas-Elguera, 2001). A similar pattern was noted by Roden (1981) as evidence that minettes from the Colorado Plateau formed by melting of garnet peridotite, an interpretation that is consistent with the presence of garnet lherzolite xenoliths in the minette samples. As will be discussed in the next section, the presence of garnet during melting requires that these magmas formed at pressures 25 kbar within the mantle wedge (Esperança & Holloway, 1987; Elkins-Tanton & Grove, 2003; Conceição & Green, 2004).
The high concentrations of all volatiles (H2O, CO2, S, F, Cl) in the melt inclusions indicate a relatively volatile-rich mantle source for the potassic magmas. At the time of melting, some or all of the volatiles may have been stored in mineral phases in the mantle, and some may also have been introduced in a fluid phase or hydrous melt derived from the underlying subducting slab. Ultimately the two sources of volatiles are linked in that metasomatic minerals such as phlogopite probably formed by previous infiltration of fluids or hydrous melts from the slab (Luhr, 1997). As shown above, the Cl contents in the potassic magmas are roughly twice those of the calc-alkaline basalts, but the ratios of S/Cl and F/Cl are as much as five and three times higher, respectively, for the potassic magmas. This suggests that if fluids coming from the slab were the only source for these volatiles, then the fluid responsible for the Colima potassic magmas was more enriched in S and F than the fluid that generated the calc-alkaline basalts. A more likely alternative is that some S, F and perhaps Cl were stored in metasomatic minerals such as apatite (which must have been consumed during partial melting), phlogopite, and possibly a sulfate phase.
Pressure and temperature conditions during partial melting
The constraints on mantle mineralogy summarized above can be combined with experimental studies of phase equilibrium and partial melting to understand the melting conditions by which the minette and basanite magmas formed. Foley (1992a, 1992b) suggested that ultrapotassic magmas form by melting of phlogopite-clinopyroxenite veins in peridotite wallrock, and such veins probably form in the mantle above subducted slabs as rising fluids react with garnet and orthopyroxene to form phlgopite, diopside and spinel (Elkins-Tanton & Grove, 2003). Some experimental studies have also shown that potassic magmas can form by melting of garnet lherzolite that has been metasomatically enriched to include phlogopite ± pargasite (Conceição & Green, 2004). Shown in Fig. 11 are experimental phase relations for melting of a K-rich peridotite under H2O-undersaturated conditions. The phase diagram suggests that phlogopite is stable at the peridotite solidus in the pressure range 10–35 kbar (corresponding to 30–110 km depth) at temperatures of 1050–1250°C. If mantle phlogopite contains appreciable amounts of F, the upper thermal stability limit of phlogopite is shifted to higher temperatures (Foley et al., 1986) but low F/(F + OH) in phlogopite phenocrysts from the Colima minettes (Luhr & Carmichael, 1981) suggests a relatively minor role for F.
|
The La/Yb systematics of the minette and basanite melt inclusions indicate that garnet was a stable phase during partial melting (Fig. 10b). The experimental phase relations of Conceição & Green (2004
) show that garnet is stable at the solidus of phlogopite-bearing peridotite at pressures 25 kbar. At pressures above 25–30 kbar, pargasitic amphibole becomes unstable, and therefore the presence of garnet suggests that pargasite would not have been a stable phase during melting. Our estimated temperatures of 1150–1230°C based on olivine–melt thermometry and the pressures necessary for garnet stability are consistent with the experimental field for phlogopite–garnet peridotite + melt on the phase diagram (Fig. 11).
The inferred melting pressures are also consistent with experimentally determined melt compositions. Both direct melting studies of K-rich lherzolite and liquidus studies of K-rich magmas have determined that melts formed at pressures of 30 kbar from phlogopite lherzolite and phlogopite clinopyroxenite sources are olivine rich and silica undersaturated, ranging from nephelinite to olivine leucitite in composition (Elkins-Tanton & Grove, 2003; Conceição & Green, 2004). The degree of silica undersaturation increases with dissolved CO2 and F (Conceição & Green, 2004), and CO2 also plays a role in stabilizing orthopyroxene, which is otherwise commonly lacking in phase equilibrium studies on high-K compositions (Foley, 1992a, 1992b; Elkins-Tanton & Grove, 2003). We suggest that more H2O-rich sources form the moderately nepheline-normative minettes and basanites, and that more CO2- and/or F-rich (or deeper) sources produced the strongly silica-undersaturated kalsilite-bearing lavas exposed in the walls of the Colima graben (Allan & Carmichael, 1984) and the leucitites erupted to the west in the Jalisco block grabens (Wallace & Carmichael, 1989). In contrast, experimental melts formed by fluid-undersaturated melting of phlogopite lherzolite at lower pressures (10 kbar) are silica-oversaturated (Conceição & Green, 2004), and thus are unlike the Colima minettes and basanites.
Fluid flux vs decompression melting
A key question regarding the origin of the potassic magmas is whether they formed by decompression melting of a phlogopite-bearing source or whether fluxing with fluids or hydrous melts derived from the subducted slab was responsible for melting. Formation by decompression melting was suggested by Hochstaedter et al. (1996) based on the lower B/Be ratios of the potassic magmas compared with the calc-alkaline magmas. They attributed this to a lack of recent fluid addition from the slab. However, our melt inclusion H2O data, combined with the B data of Hochstaedter et al. (1996), show that the B/H2O ratios of the Colima potassic magmas overlap with primitive MGVF calc-alkaline magmas: B values average 4·2 ppm for calc-alkaline MGVF rocks with >6 wt% MgO and 4·3 ppm for the Colima potassic rocks (Hochstaedter et al., 1996). Similarly, maximum H2O contents are 6· 0 wt % for the calc-alkaline MGVF melt inclusions (Johnson, 2008) and 6·2 wt% for the potassic melt inclusions analyzed in this study. This suggests that a common, subduction-derived H2O-rich component is involved in formation of these diverse primitive magma types.
We examine the possible role of subduction-derived fluids in formation of the potassic magmas using a ternary plot of Na2O–K2O–H2O (Fig. 12). The figure shows the compositions of phlogopite, average normal mid-ocean ridge basalt (N-MORB) and enriched MORB (E-MORB), and empirically derived subduction zone fluid compositions for Mexico and other arcs. The data suggest that basaltic melts formed by fluid-flux melting of an N-MORB to E-MORB mantle source have relatively low K2O/H2O whereas melts formed by melting of phlogopite-bearing mantle will have much higher K2O/H2O. Consistent with the former, the primitive calc-alkaline melt inclusions from the MGVF and Cerro Usmajac plot within and near the field for mixtures of MORB and slab-derived H2O-rich components, suggesting that they are derived from N-MORB to E-MORB-type mantle fluxed with an H2O-rich component from the slab (e.g. Cervantes & Wallace, 2003). In contrast, the primitive minette and basanite compositions plot between the field defined by MORB and slab-derived H2O-rich components and the field for phlogopite-bearing mantle sources. This suggests a role for both phlogopite and fluxing with an H2O-rich component derived from the slab in the generation of the potassic magmas.
|
An alternative possibility is that phlogopite + pargasite in the mantle source could create the appropriate K2O/H2O in the potassic magmas (Fig. 12). However, the evidence above for presence of residual garnet suggests that pargasite would not be stable at the peridotite solidus (Fig. 11), and pargasite is not typically found on the liquidus of potassic compositions in phase equilibrium studies (Elkins-Tanton & Grove, 2003
). K-richterite has similar K2O/H2O to phlogopite, and therefore its presence would not change our interpretations based on the Na2O–K2O–H2O ternary. However, experimental studies suggest that K-richterite is not stable relative to phlogopite at pressures below 70–90 kbar (Kawamoto & Holloway, 1997; Trønnes, 2002), and, as with pargasite, it has not been found on the liquidus of potassic compositions at upper mantle pressures (Elkins-Tanton & Grove, 2003), although examples of phlogopite + K-richterite-bearing peridotite xenoliths do exist (Haggerty, 1995).
Our data on the H2O contents of primitive minette and basanite melts provide a quantitative constraint on the relative roles of phlogopite and a subduction-derived H2O-rich component in magma formation. The amount of H2O released during melting of phlogopite-peridotite is controlled by the amount of phlogopite consumed per unit of melt produced and the amount of H2O initially present in the phlogopite. Using experimental data for liquidus phase relationships of potassic magmas (Esperança & Holloway, 1987) and the method of Baker & Stolper (1994) we determined the following melting reaction stoichiometry for phlogopite peridotite at 10 kbar and the iron-wüstite-graphite (IWG) buffer:
|
|
Assuming mantle phlogopite contains a maximum of 4–5 wt % H2O based on experimental studies and xenoliths (Sweeney et al., 1993; note that substitution of F, Cl, and oxy-phlogopite would reduce the amount of OH present; Haggerty, 1995; Righter et al., 2002), then the maximum amount of H2O in a partial melt would be 1· 9–2·3 wt% based on this melting reaction stoichiometry. This is substantially less than the values of 3·8–6·2 wt% H2O for the most primitive minette and basanite melts based on the melt inclusion data. Therefore, it appears that melting of phlogopite peridotite in the absence of a hydrous fluid or melt phase cannot produce the high H2O contents of the potassic melts unless the ratio of phlogopite consumed to melt produced is a factor of two or more greater than indicated by the experiments of Esperança & Holloway (1987). As a limiting case, we also note that both congruent melting of pure phlogopite or incongruent melting (in which the solid phases produced contain no K2O or H2O) will produce melts with K2O/H2O ratios that are higher than those of the potassic magmas (Fig. 12). This further supports our contention that hydrous fluids or melts must be involved in melting to form the minette and basanite magmas.
Trace element model of flux melting
Based on the above evidence, we developed a simple quantitative model to test whether fluid-flux melting of a phlogopite-peridotite mantle source can give rise to the trace element abundances of the potassic magmas. In previous studies where such models have been created, the standard approach has been to assume a depleted MORB mantle source and use the compositions of volcanic rocks or glasses to constrain the composition of the H2O-rich subduction component transferred from the slab to the mantle wedge (Stolper & Newman, 1994; Grove et al., 2002; Cervantes & Wallace, 2003; Eiler et al., 2005; Portnyagin et al., 2007). The estimated compositions for the H2O-rich components in different arcs deduced in these studies show many similar features. Here we take a slightly different approach in that we use the average composition of H2O-rich components constrained by data for primitive, medium-K calc-alkaline melt inclusions from the MGVF, just to the east of the Colima graben (Johnson, 2008). Our goal with this approach is to test whether H2O-rich components of similar composition can form both the potassic and calc-alkaline magmas in the TMVB.
A key aspect of the model is the choice of composition for the mantle wedge before enrichment with an H2O-rich component derived from the slab. Numerous studies have shown that the mantle beneath the TMVB is heterogeneous (e.g. Verma & Nelson, 1989; Luhr, 1997; Wallace & Carmichael, 1999; Ferrari et al., 2001), and the use of ‘conservative’ or relatively fluid-immobile elements (e.g. Pearce & Peate, 1995) can be useful for discerning mantle heterogeneity that is unrelated to or pre-dates hydrous enrichment. On the basis of Nb/Y and Ti/Y ratios, the volumetrically dominant, medium-K calc-alkaline magmas in the western TMVB are generated from mantle sources that have variable compositions but overall are similar to an E-MORB mantle source (Fig. 13a). The Colima potassic magmas fall within the range for the calc-alkaline samples but lie at the enriched end of the array, with compositions that are intermediate between E-MORB and ocean island basalt (OIB). Within the western TMVB, there are also intraplate alkaline basalts with OIB-like trace element characteristics (no Nb–Ta depletion) that probably form as a result of decompression melting driven by corner flow (Luhr, 1997) or mantle upwelling through a slab tear beneath the rear-arc (Ferrari, 2004). Data for trace elements (Wallace & Carmichael, 1999; Siebe et al. 2004), volatiles (Cervantes & Wallace, 2003; Johnson, 2008), and radiogenic isotopes (Luhr, 1997) are consistent with the hypothesis that the E-MORB-mantle source regions beneath the TMVB form as a result of previous partial melting and melt extraction from more enriched (OIB) mantle sources, and then become overprinted with enrichments in large ion lithophile elements, LREE and H2O from the dehydrating slab (Fig. 13b). The Cerro Usmajac basaltic andesite is an intermediate case in that it has relatively high Nb/Ta, like the intraplate alkaline rocks, but also shows enrichment in Pb and Sr, like calc-alkaline magmas in the TMVB (Fig. 4). Such intermediate cases have also been found in the MGVF and Sierra Chichinautzin regions to the east and point to the heterogeneity of the subarc mantle and variable enrichment from subduction-derived fluids (Cervantes & Wallace, 2003; Johnson, 2008).
|
In our model, we assume that the H2O-rich component derived from the slab was added to a garnet lherzolite mantle wedge in two stages. The first created metasomatically enriched lherzolite containing phlogopite (or veins of phlogopite + clinopyroxene) but caused no partial melting, and the second was assumed to result in flux melting. The initial mantle source had trace element abundances intermediate between those of an E-MORB mantle source and an OIB source based on the Nb/Y systematics (Fig. 13a). We used the average composition for the H2O-rich subduction component beneath the MGVF (Johnson, 2008
), and we iteratively varied the modal mineralogy of the metasomatized lherzolite source, the amount of the H2O-rich component added to the wedge, and the degree of partial melting to find the range of closest fits to the average basanite and minette melt inclusion compositions. Because of uncertainties in (1) the relationship between source H2O content and degree of melting, (2) the partitioning of H2O between phlogopite lherzolite and partial melt, and (3) the melting reaction stoichiometry for phlogopite–garnet lherzolite, we used the modal batch melting equation for all calculations and we did not attempt to calculate the H2O contents of the partial melts. The distribution coefficients (D) used in the model are reported in Table 4. Models results are shown in Fig. 14.
|
|
The best fitting fluid-flux melting models had a residual mantle mineral assemblage of 50% olivine, 31% orthopyroxene, 10% clinopyroxene, 8% garnet, and 1% phlogopite, but the results are not very sensitive to variations in the first three major phases. Larger amounts of residual phlogopite caused Rb and Ba concentrations to be too low. The total amount of the H2O-rich component added in the two stages was 7 wt %, and the extents of partial melting were 5–15%. Despite the simplifying assumptions involved, these models demonstrate that fluxing of a mantle source intermediate between E-MORB and OIB with an H2O-rich component from the slab is a viable hypothesis for formation of the Colima potassic magmas. The models also show that the H2O-rich components responsible for formation of medium-K calc-alkaline magmas in the TMVB are virtually identical to the components required for formation of the minettes and basanites, as is the composition of their mantle sources before subduction-related enrichment (i.e. based on Nb/Y). The main discrepancy between the model melts and the Colima potassic magmas is that Sr is typically too high in the models and Ba is slightly low. This could easily be reconciled by invoking an H2O-rich component with lower Sr and slightly higher Ba than the values derived from calc-alkaline magmas in the MGVF. Also, the model melts show more variability in Nb and Ta than is observed. This could be explained by a few per cent of either residual rutile or ilmenite, although for reasons discussed above, we have concluded that these phases were not likely to have been residual during melting.
Formation of potassic and calc-alkaline magmas beneath the Colima graben
We have shown that the volatile contents of the potassic magmas require addition of an H2O-rich component from the subducted slab and that the trace element variations require the presence of garnet and metasomatic phlogopite in the mantle source. Based on this evidence and the results of trace element modeling, we interpret the potassic magmas to have formed by fluid fluxing and melting of a metasomatized mantle wedge that initially had a composition intermediate between that of E-MORB and OIB mantle sources. The presence of garnet in the source requires pressures (25 kbar) that are probably too high for the minettes and basanites to be created by melting of metasomatically enriched mantle lithosphere and instead require an origin in the asthenosphere. The current depth of the subducted Rivera and Cocos plates beneath the Colima graben is poorly constrained, but beneath the Jalisco block to the west of Colima, the Rivera Plate is 140 km (40 kbar) beneath the volcanic front (Grand et al., 2007; Yang et al., in preparation), along which potassic magmas have also erupted. Below 150 km depth, seismic tomography shows a slab tear between the Cocos and Rivera plates located beneath the region just west of the Colima graben (Grand et al., 2007; Yang et al., in preparation). Although the depth to the base of the lithosphere in this region is not constrained, the much shallower dip angle on the Rivera Plate during the middle to late Miocene (Ferrari et al., 2001; Gomez-Tuena et al., 2007) and the possibility of flat slab subduction (V. C. Manea, personal communication, 2008), as currently occurs trenchward of the TMVB near Mexico City, suggest that a thick lithospheric mantle root was probably not present beneath the Colima graben and Jalisco block during the Pliocene and Holocene. Results of experimental petrological studies also suggest a depth to the base of the lithospheric mantle of 60 km beneath the volcanic front in the Jalisco block (Hesse & Grove, 2003).
The eruptive volume of the potassic magmas in the Colima graben is trivial in comparison with calc-alkaline andesites of the Colima Volcanic Complex, which evolved from more primitive calc-alkaline basaltic magmas (Luhr, 1997). Interestingly, the potassic and calc-alkaline magmas have overlapping Sr, Nd and Pb isotope ratios (Luhr et al., 1989) and similarities in key trace element ratios in primitive samples of both magma types (e.g. Ba/Ce, Hochstaedter et al., 1996). Based on these similarities and the melting pressures required by residual garnet, Luhr (1997) proposed that melting of phlogopite-bearing veins in the asthenospheric mantle wedge formed small volumes of potassic magma and that dilution of the potassic component by greater amounts of peridotite melting from wallrock around the veins and exhaustion of residual phlogopite generates the calc-alkaline magmas (Luhr, 1997).
In addition to the Colima graben, similar potassic magmas have erupted elsewhere in the Jalisco block to the west (Fig. 1). Most of these occur in smaller extensional grabens, but some erupted in regions that do not show evidence of extension (Righter & Rosas-Elguera, 2001). It has been suggested (Lange & Carmichael, 1991; Righter & Rosas-Elguera, 2001) that the metasomatic enrichment of the mantle responsible for forming phlogopite veins occurred during earlier subduction-related magmatism in the Jalisco block, represented by widespread Cretaceous ignimbrites (65–92 Ma; see Righter et al., 1995, and summary by Frey et al., 2007) and granitic plutons (75–100 Ma; Schaaf et al., 1995). Our results suggest, however, that H2O-rich fluid was present at the time of melting and that the metasomatic veins were formed by fluids similar in composition to those responsible for recent calc-alkaline magmatism in the MGVF to the east of the Colima graben (Johnson, 2008). Thus we consider it less likely that the enriched veins formed during Cretaceous magmatism and instead link them to the active, modern subduction system, in agreement with the interpretation of Luhr (1997). This interpretation is also consistent with the similarity of 87Sr/86Sr ratios in the calc-alkaline and potassic magmas (Luhr, 1997). Phlogopite-bearing veins would have much higher Rb/Sr than peridotitic mantle so that 60–100 Myr of isotopic aging of these veins would cause elevated 87Sr/86Sr (Gomez-Tuena et al., 2007), although the presence of clinopyroxene (with high partition coefficient for Sr) with phlogopite in the veins would slow the isotopic aging effect (Schmidt et al., 1999).
A key question regarding the origin of the potassic magmas is to what extent they form as a result of normal subduction processes as opposed to lithospheric extension or other factors. Luhr (1997) favored the former interpretation and suggested that the role of lithospheric extension was to create pathways that made it easier for small volume potassic melts to leave the mantle and ascend through the crust. The pressures of partial melting required by the presence of garnet place the source region in the deeper parts of the asthenospheric mantle wedge, closer to the subducted slab, rather than in the lithospheric mantle or uppermost asthenosphere, so an origin by decompression melting caused by extension seems less likely. Our evidence for the presence of subduction-derived fluids seems to further rule out the hypothesis that mantle melting to form the potassic magmas was caused primarily by lithospheric extension because such fluids are not expected to reach the uppermost mantle (e.g. Cagnioncle et al., 2007). Finally, the amount of extension in the Colima graben (<10% over the last 5 Myr) is more than an order of magnitude less than is found in typical continental rifting environments (Ferrari et al., 2001, 2003) and is insufficient to cause decompression melting in the absence of fluids from the subducting slab (e.g. Harry & Leeman, 1995).
In addition to the subduction relationship implied by the presence of fluids at the time of melting, we suggest that several additional factors were important, and we formulate the following model for the origin of the potassic magmas in the western TMVB. First, we suggest that veins of phlogopite formed in both the forearc mantle wedge and deeper parts of the wedge at a time when the plate dipped more shallowly and the volcanic front was farther from the trench than it is today (Fig. 15a). Temperatures in the forearc and deeper portions of the wedge would have been low enough for phlogopite to be stable, and garnet would have been stable with phlogopite at 25 kbar pressure. Second, as the slab dip angle increased starting at 8·5 Ma (Ferrari et al., 2001), these phlogopite-bearing mantle regions were advected downwards and towards the trench by corner flow caused by slab rollback, and continued to receive fluid flux from the downgoing slab (Fig. 15b). Heating of this veined and fluid-fluxed mantle as a result of arc-parallel flow of hot mantle derived from both the slab tear and western Rivera Plate edge (Ferrari et al., 2001) caused melting to form the potassic magmas. Thus the increase in slab dip angle and trenchward migration of the volcanic front into the old forearc caused small volume potassic melts to form beneath the entire volcanic front of the TMVB, not just beneath regions undergoing extension. The compositions of the potassic melts ranged from moderately Ne-normative (basanites and minettes) to strongly Ne- and Lc-normative (leucitites, kalsilite-bearing ankaratrites) because partial melting occured relatively deep (25 kbar; Conceição & Green, 2004), in the asthenospheric wedge just above the subducted slab. The leucitites from the Jalisco block, in particular, require melting pressures of 3–4 GPa (Elkins-Tanton & Grove, 2003). Third, lithospheric extension along the Colima graben was associated with upwelling of hot mantle through a tear between the downgoing Cocos and Rivera plates (Grand et al., 2007; Yang et al., in preparation). Heating caused by this process augmented melting of the phlogopite-veined and fluid-fluxed asthenospheric mantle, and this caused the Colima graben to be a locus for eruptions of potassic magmas from 4·6 Ma until nearly the present, a longer time span than in any of the other Jalisco block regions with potassic magmatism. Fourth, the upwelling of hot mantle beneath the Colima graben has probably amplified the partial melting related to subduction processes in the asthenospheric wedge beneath the large Colima volcanoes, causing them to have larger volume and be closer to the trench than other stratovolcanoes in the western TMVB (Yang et al., in preparation). Finally, we agree with previous workers in suggesting that lithospheric extension also helped to provide pathways for the small volume potassic melts to make it through the mantle and crust.
|
In our model, it is a combination of trenchward migration of the arc into the older forearc region, upwelling of hot mantle through a tear between the Cocos and Rivera plates, arc-parallel flow of hot mantle within the wedge from both the slab tear and the western Rivera Plate edge, and extension across a broad region of the Jalisco block that is ultimately responsible for the formation and eruption of potassic magmas. This particular combination explains why such high-K magmas have not erupted elsewhere in the TMVB and are not common in arcs in general. To the east of the Colima graben in the MGVF, trenchward migration has occurred, but there has been less extension, and the region is located farther from the Cocos–Rivera slab tear or Rivera Plate edge. At the volcanic front in the MGVF, hornblende trachybasalts with higher K2O than calc-alkaline basalts have erupted, but no minettes or basanites similar to those in the Jalisco block or Colima graben have been found (Luhr, 1997
; Hasenaka & Carmichael, 1987). Even further to the east in the Sierra Chichinautzin volcanic field near Mexico City, trenchward migration and slab steepening have not occurred. Only two lavas with elevated K2O (andesitic trachybasalts with 2·4–3·2 wt% K2O) have been reported from this region (Wallace & Carmichael, 1999; Meriggi et al., 2008). |
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Olivine-hosted melt inclusions from basanite and minette cinder cones in the Colima graben reveal that both primitive magma types are volatile rich (
6·2 wt% H2O; 5300 ppm CO2) and relatively oxidized (NNO + 0·8 to NNO + 1·7). A more evolved basaltic andesite cinder cone (Cerro Usmajac) near the potassic cones has similar oxygen fugacity and high H2O (5·5 wt%), but has lower CO2, S, and Cl. The volatile concentrations in the basaltic andesite are comparable with those in primitive calc-alkaline melt inclusions from the MGVF to the east. Vapor saturation pressures recorded by H2O and CO2 in the melt inclusions suggest that olivine crystallized over a large range of depths (from >25 km to near the surface) during degassing of the melts. Water and CO2 variations with depth cannot be explained by simple degassing models but instead require more complex, open-system processes or possibly reflect disequilibrium. The melt inclusion data show that S was partially lost by degassing during ascent, but Cl and F remained dissolved in melt, presumably reflecting their relatively high solubility.
Modeling of trace element variations in the potassic melt inclusions indicates that phlogopite and garnet were residual phases in their mantle source during partial melting. The presence of garnet in the source requires pressures (25 kbar) that are probably too high for the minettes and basanites to be created by melting of metasomatically enriched mantle lithosphere and instead require an origin in the asthenosphere. The high S, Cl, and F of the potassic magmas relative to calc-alkaline compositions suggests additional metasomatic phases such as apatite and possibly sulfate, but these must have been consumed during partial melting. Decompression melting of phlogopite peridotite cannot produce the high H2O contents of the potassic magmas, and thus the presence of fluid or hydrous melt at the time of melting is required. On the basis of Nb/Y and Ti/Y ratios, the mantle source that was fluxed and melted to form the potassic magmas must have initially been intermediate between E-MORB and OIB mantle sources in composition. Trace element modeling of such a mantle source fluxed repeatedly with H2O-rich fluids or hydrous melts from the subducting slab (first to form metasomatic phases such as phlogopite, and second to provide fluid at the time of melting) can reproduce most of the trace element characteristics of the potassic melts. The coexistence of potassic and calc-alkaline magmas in the Colima graben from 4·6 to 0·06 Ma is probably related to a combination of factors including the steepening of the slab dip angle starting in the late Miocene, the initiation of a slab tear between the Cocos and Rivera plates beneath the Colima graben, heating by upwelling of hot mantle through the tear and consequent arc-parallel flow of hot mantle, and lithospheric thinning caused by extension. The extensional tectonics also allowed for easier ascent of the low volume potassic magmas through the crust.
|
SUPPLEMENTARY DATA |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Supplementary data for this paper are available at Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
This manuscript is dedicated to the memory of Jim Luhr in honor of his pioneering work on volcanism in the Trans-Mexican Volcanic Belt and the cinder cones of the Colima graben. We would like to thank Hugo Delgado Granados and Carlos Linares for their help in the field, Emily Johnson for help in the field and with sample analysis, and John Donovan for assistance with electron microprobe analysis. We greatly appreciate discussions with Pedro Jugo, Don Baker, Becky Lange, Luca Ferrari, and Steve Grand, and detailed reviews by Becky Lange, Kevin Righter and Richard Wysoczanski that led to numerous improvements in the final manuscript. This work was supported by funding from the National Science Foundation (grant EAR0309559 to P.W.).
*Corresponding author. Present address: Department of Earth Sciences, Simon Fraser University, Burnaby, B.C., Canada V5A 1S6. Telephone: 778-782-6780. Fax: 778-782-4198. E-mail: nvigouro{at}sfu.ca
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONTECTONIC SETTINGPOTASSIC MAGMAS IN THE...SAMPLES AND METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Allan JF. Geology of the northern Colima and Zacoalco grabens, southwest Mexico: Late-Cenozoic rifting in the Mexican Volcanic Belt. Geological Society of America Bulletin (1986) 97:473–485.
Allan JF, Carmichael ISE. Lamprophyric lavas in the Colima Graben, SW Mexico. Contributions to Mineralogy and Petrology (1984) 88:203–216.[CrossRef][Web of Science]
Anderson AT, Brown GG. CO2 contents and formation pressures of some Kilauean melt inclusions. American Mineralogist (1993) 78:794–803.[Abstract]
Atlas ZD, Dixon JE, Sen G, Finny M, Lillian A, Pozzo MD. Melt inclusions from Volcan Popocatepetl and Volcan de Colima, Mexico: Melt evolution due to vapor-saturated crystallization during ascent. Journal of Volcanology and Geothermal Research (2006) 153(3–4):221–240.[CrossRef][Web of Science]
Baker DR, Freda C, Brooker RA, Scarlato P. Volatile diffusion in silicate melts and its effects on melt inclusions. Annals of Geophysics (2005) 48:699–717.[Web of Science]
Baker MB, Stolper EM. Determining the composition of high-pressure mantle melts using diamond aggregates. Geochimica et Cosmochimica Acta (1994) 58(13):2811–2827.[CrossRef][Web of Science]
Baker RD. The fidelity of melt inclusions as records of melt composition. Contributions to Mineralogy and Petrology (2008) doi:10.1007/s00410-008-0291-3.
Ballhaus C, Berry RF, Green DH. High pressure experimental calibration of the olivine–orthopyroxene–spinel oxygen geobarometer: implications for the oxidation state of the upper mantle. Contributions to Mineralogy and Petrology (1991) 107:27–40.[CrossRef][Web of Science]
Bloomer SH, Stern RJ, Fisk E, Geschwind CH. Shoshonitic volcanism in the Northern Mariana arc. I. Mineralogic and major and trace element characteristics. Journal of Geophysical Research (1989) 94:4469–4496.
Borg LE, Clynne MA, Bullen TD. The variable role of slab-derived fluids in the generation of a suite of primitive calc-alkaline lavas from the southernmost Cascades, California. Canadian Mineralogist (1997) 35:425–452.[Web of Science]
Cagnioncle A-M, Parmentier EM, Elkins-Tanton LT. Effect of solid flow above a subducting slab on water distribution and melting at convergent plate boundaries. Journal of Geophysical Research (2007) 112. article no. B09402.
Carmichael ISE, Lange RA, Luhr JF. Quaternary minettes and associated volcanic rocks of Mascota, western Mexico: a consequence of plate extension above a subduction modified mantle wedge. Contributions to Mineralogy and Petrology (1996) 124:302–333.[CrossRef][Web of Science]
Carmichael ISE, Frey HM, Lange RA, Hall CM. The Pleistocene cinder cones surrounding Volcán Colima, Mexico re-visited: eruption ages and volumes, oxidation states, and sulfur content. Bulletin of Volcanology (2006) 68:407–419.[CrossRef][Web of Science]
Carroll MR, Rutherford MJ. Sulfur speciation in hydrous experimental glasses of varying oxidation state: Results from measured wavelength shifts of sulfur X-rays. American Mineralogist (1988) 73:845–849.[Abstract]
Carroll MR, Webster JD. Solubilities of sulfur, noble gases, nitrogen, chlorine, and fluorine in magmas. Volatiles in Magmas. Mineralogical Society of America, Reviews in Mineralogy—Carroll MR, Holloway JR, eds. (1994) 30:231–279.
Cervantes P, Wallace PJ. Role of H2O in subduction-zone magmatism: New insights from melt inclusions in high-Mg basalts from central Mexico. Geology (2003) 31:235–238.
Christensen MI, Mooney WD. Seismic velocity structure and composition of the continental crust—a global view. Journal of Geophysical Research—Solid Earth (1995) 100:9761–9788.[CrossRef]
Conceição RV, Green DH. Derivation of potassic (shoshonitic) magmas by decompression melting of phlogopite + pargasite lherzolite. Lithos (2004) 72:209–229.[CrossRef][Web of Science]
Danyushevsky LV, Della-Pasqua FN, Sokolov S. Re-equilibration of melt inclusions trapped by magnesian olivine phenocrysts from subduction-related magmas: petrological implications. Contributions to Mineralogy and Petrology (2000) 138:68–83.[CrossRef][Web of Science]
Danyushevsky LV, McNeill AW, Sobolev AV. Experimental and petrological studies of melt inclusions in phenocrysts from mantle-derived magmas: an overview of techniques, advantages and complications. Chemical Geology (2002) 183:5–24.[CrossRef][Web of Science]
Delgado-Granados H. Late Cenozoic tectonics offshore western Mexico and its relation to the structure and volcanic activity in the western Trans-Mexican Volcanic Belt. Geofísica Internacional (1993) 32:543–559.
Dixon JE. Degassing of alkalic basalts. American Mineralogist (1997) 82:368–378.[Abstract]
Dixon JE, Clague DA, Wallace P, Poreda R. Volatiles in alkali basalts from the North Arch Volcanic field, Hawaii: extensive degassing of deep submarine-erupted alkalic series lavas. Journal of Petrology (1997) 38:911–939.[CrossRef][Web of Science]
Dixon JE, Pan V. Determination of the molar absorptivity of dissolved carbonate in basanitic glass. American Mineralogist (1995) 80:1339–1342.[Abstract]
Dixon JE, Stolper EM, Holloway JR. An experimental study of H2O and carbon dioxide solubilities in mid-ocean ridge basaltic liquids. Part I: Calibration and solubility results. Journal of Petrology (1995) 36:1607–1631.
Eggler DH. Amphibole stability in H2O-undersaturated calc-alkaline melts. Earth and Planetary Science Letters (1972) 15:28–34.[CrossRef][Web of Science]
Eiler JM, Carr MJ, Reagan M, Stolper EM. Oxygen isotope constraints on the sources of Central American arc lavas. Geochemistry, Geophysics, Geosystems (2005) 6. doi:10.1029/2004GC000804.
Elkins-Tanton LT, Grove TL. Evidence for deep melting of hydrous metasomatized mantle: Pliocene high-potassium magmas from the Sierra Nevadas. Journal of Geophysical Research—Solid Earth (2003) 108. article no. 2350.
Ellam RM, Menzies M, Hawkesworth CJ, Leeman WP, Rosi M, Serri G. The transition from calc-alkaline to potassic orogenic magmatism in the Aeolian Islands, Southern Italy. Bulletin of Volcanology (1988) 50:386–398.[CrossRef][Web of Science]
Esperança S, Holloway JR. On the origin of some mica-lamprophyres—experimental evidence from a mafic minette. Contributions to Mineralogy and Petrology (1987) 95(2):201–216.
Faure F, Schiano P. Experimental investigation of equilibration conditions during forsterite growth and melt inclusion formation. Earth and Planetary Science Letters (2005) 236(3–4):882–898.[CrossRef][Web of Science]
Feldstein SN, Lange RA. Pliocene potassic magmas from the Kings River region, Sierra Nevada, California: Evidence for melting of a subduction-modified mantle. Journal of Petrology (1999) 40:1301–1320.[CrossRef][Web of Science]
Feldstein SN, Lange RA, Vennemann T, O’Neil. Ferric–ferrous ratios, H2O contents and D/H ratios of phlogopite and biotite from lavas of different tectonic regimes. Contributions to Mineralogy and Petrology (1996) 126:51–66.[CrossRef][Web of Science]
Ferrari L. Slab detachment control on mafic volcanic pulse and mantle heterogeneity in central Mexico. Geology (2004) 32(1):77–80.
Ferrari L, Petrone CM, Francalanci L. Generation of oceanic-island basalt-type volcanism in the western Trans-Mexican volcanic belt by slab rollback, asthenosphere infiltration, and variable flux melting. Geology (2001) 29:507–510.
Ferrari L, Petrone CM, Francalanci L, Tagami T, Eguchi M, Conticelli S, Manetti P, Venegas-Salgado S. Geology of the San Pedro–Ceboruco graben, western Trans-Mexican Volcanic Belt. Revista Mexicana de Ciencias Geologicas (2003) 20:165–181.
Foley S. Petrological characterization of the source components of potassic magmas—geochemical and experimental constraints. Lithos (1992a) 28:187–204.[CrossRef][Web of Science]
Foley S. Vein-plus-wall-rock melting mechanisms of the lithosphere and the origin of potassic alkaline magmas. Lithos (1992b) 28:435–453.[CrossRef][Web of Science]
Foley SF, Taylor WR, Green DH. The effect of fluorine on phase relationships in the system KAlSiO4–Mg2SiO4–SiO2 and the solution mechanism of fluorine in silicate melts. Contributions to Mineralogy and Petrology (1986) 93:46–55.[CrossRef][Web of Science]
Foley SF, Jackson SE, Fryer BJ, Greenough JD, Jenner GA. Trace element partition coefficients for clinopyroxene and phlogopite in an alkaline lamprophyre from Newfoundland by LAM-ICP-MS. Geochimica et Cosmochimica Acta (1996) 60:629–638.[CrossRef][Web of Science]
Frey FA, Clague D, Mahoney JJ, Sinton JM. Volcanism at the edge of the Hawaiian plume: Petrogenesis of submarine alkalic lavas from the North Arch Volcanic Field. Journal of Petrology (2000) 41:667–691.
Frey HM, Lange RA, Hall CM, Delgado-Granados H, Carmichael ISE. A Pliocene ignimbrite flare-up along the Tepic–Zacoalco rift: Evidence for the initial stages of rifting between the Jalisco block (Mexico) and North America. Geological Society of America Bulletin (2007) 119:49–64.
Gomez-Tuena A, Langmuir CH, Goldstein SL, Straub SM, Ortega-Gutierrez F. Geochemical evidence for slab melting in the Trans-Mexican Volcanic Belt. Journal of Petrology (2007) 48:537–562.
Gonnermann HM, Manga M. Nonequilibrium magma degassing: Results from modeling of the ca. 1340 AD eruption of Mono Craters, California. Earth and Planetary Science Letters (2005) 238:1–16.[CrossRef][Web of Science]
Grand SP, Yang T, Wilson D, Guzman-Speziale M, Gomez-Gonzales J, Dominguez-Reyes T, Ni J. The seismic structure of the Rivera subduction zone. EOS Transactions, American Geophysical Union, Fall Meeting Abstracts (2007) T41. C-0702.
Green TH, Blundy JD, Adam J, Yaxley GM. SIMS determination of trace element partition coefficients between garnet, clinopyroxene and hydrous basaltic liquids at 2–7.5 GPa and 1080–1200°C. Lithos (2000) 53:165–187.[CrossRef][Web of Science]
Green TH, Watson EB. Crystallization of apatite in natural magmas under high-pressure, hydrous conditions, with particular reference to orogenic rock series. Contributions to Mineralogy and Petrology (1982) 79:96–105.[CrossRef][Web of Science]
Grégoire M, Bell DR, Le Roex AP. Trace element geochemistry of phlogopite-rich mafic mantle xenoliths: their classification and their relationship to phlogopite-bearing peridotite and kimberlites revisited. Contributions to Mineralogy and Petrology (2002) 142:603–625.[Web of Science]
Grove TL, Parman SW, Bowring SA, Price RC, Baker MB. The role of an H2O-rich fluid component in the generation of primitive basaltic andesites and andesites from the Mt. Shasta region, N California. Contributions to Mineralogy and Petrology (2002) 142(4):375–396.[Web of Science]
Grove TL, Chatterjee N, Parman SW, Medard E. The influence of H2O on mantle wedge melting. Earth and Planetary Science Letters (2006) 249(1–2):74–89.[CrossRef][Web of Science]
Haggerty SE. Upper mantle mineralogy. Journal of Geodynamics (1995) 20:331–364.[CrossRef][Web of Science]
Harry DL, Leeman WP. Partial melting of melt metasomatized subcontinental mantle and the magma source potential of the lower lithosphere. Journal of Geophysical Research—Solid Earth (1995) 100:10255–10269.[CrossRef]
Hasenaka T, Carmichael ISE. The cinder cones of Michoacán–Guanajuato, central Mexico: their age, volume and distribution, and magma discharge rate. Journal of Volcanology and Geothermal Research (1985) 25:105–124.[CrossRef][Web of Science]
Hasenaka T, Carmichael ISE. The cinder cones of Michoacán–Guanajuato, central Mexico: petrology and chemistry. Journal of Petrology (1987) 28:241–269.
Hesse M, Grove TL. Absarokites from the western Mexican Volcanic Belt: constraints on mantle wedge conditions. Contributions to Mineralogy and Petrology (2003) 146:10–27.[CrossRef][Web of Science]
Hochstaedter AG, Ryan JG, Luhr JF, Hasenaka T. On B/Be ratios in the Mexican Volcanic Belt. Geochimica et Cosmochimica Acta (1996) 60:613–628.[CrossRef][Web of Science]
Johnson DM, Hooper PR, Conrey RM. XRF analysis of rocks and minerals for major and trace elements on a single low dilution Li-tetraborate fused bead. Advances in X-ray Analysis (1999) 41:843–867.
Johnson ER. Volatiles in basaltic melts from Central Mexico: from subduction to eruption. (2008) University of Oregon. Ph.D. dissertation.
Johnson ER, Wallace PJ, Cashman KV, Delgado-Granados H, Kent AJR. Magmatic volatile contents and degassing-induced crystallization at Volcán Jorullo, Mexico: Implications for melt evolution and the plumbing systems of monogenetic volcanoes. Earth and Planetary Science Letters (2008) 269:477–486.[Web of Science]
Jugo PJ, Luth RW, Richards JP. Experimental data on the speciation of sulfur as a function of oxygen fugacity in basaltic melts. Geochimica et Cosmochimica Acta (2005) 69:497–503.[CrossRef][Web of Science]
Jugo PJ, Wilke M, Susini J. Disequilibrium experiments and micro-XANES analysis: Novel tools to unravel the speciation of sulfur in silicate melts. EOS Transactions, American Geophysical Union, Fall Meeting Abstracts (2007) V22. B-04.
Kawamoto T, Holloway JR. Melting temperature and partial melt chemistry of H2O-saturated mantle peridotite to 11 Gigapascals. Nature (1997) 276:240–243.
Kent AJR, Stolper EM, Francis D, Woodhead J, Frei R, Eiler J. Mantle heterogeneity during the formation of the North Atlantic Tertiary Province: Constraints from trace element and Sr–Nd–Os–O isotope systematics of Baffin Island picrites. Geochemistry, Geophysics, Geosystems (2004) 5:Q11004. doi:10.1029/2004GC000743.[CrossRef]
Knaack C, Cornelius SB, Hooper PR. Trace element analyses of rocks and minerals by ICP-MS. In: Technical Note, GeoAnalytical Lab (1994) Washington State University. http://www.wsu.edu/~geolab/note/icpms.html.
Koloskov AV, Flerov GB, Seliverstov VA, Dorendorf F, Churikova TG. Potassic volcanics of the Central Kamchatka and the Late Cretaceous–Paleogene Kuril–Kamchatka alkaline province. Petrology (1999) 7:527–543.[Web of Science]
Lange RA, Carmichael ISE. A potassic volcanic front in western Mexico: the lamprophyric and related lavas of San Sebastian. Geological Society of America Bulletin (1991) 103:928–940.
LaTourette T, Hervig RL, Holloway JR. Trace element partitioning between amphibole, phlogopite, and basanite melt. Earth and Planetary Science Letters (1995) 135:13–30.[CrossRef][Web of Science]
le Roux PJ, Shirey SB, Hauri EH, Perfit MR, Bender JF. The effects of variable sources, processes and contaminants on the composition of northern EPR MORB (8–10°N and 12–14°N): Evidence from volatiles (H2O, CO2, S) and halogens (F, Cl). Earth and Planetary Science Letters (2006) 251:209–231.[CrossRef][Web of Science]
Luhr JF. Experimental phase relations of water- and sulfur-saturated arc magmas and the 1982 eruptions of El Chichón volcano. Journal of Petrology (1990) 31:1071–1114.
Luhr JF. Slab-derived fluids and partial melting in subduction zones: insights from two contrasting Mexican volcanoes (Colima and Ceboruco). Journal of Volcanology and Geothermal Research (1992) 54:1–18.[CrossRef][Web of Science]
Luhr JF. Extensional tectonics and the diverse primitive volcanic rocks in the western Mexican Volcanic Belt. Canadian Mineralogist (1997) 35:473–500.[Web of Science]
Luhr JF. Glass inclusions and melt volatile contents at Parícutin Volcano, Mexico. Contributions to Mineralogy and Petrology (2001) 142:261–283.[Web of Science]
Luhr JF, Carmichael ISE. The Colima Volcanic Complex, Mexico. II. Late-Quaternary cinder cones. Contributions to Mineralogy and Petrology (1981) 76:127–147.[CrossRef][Web of Science]
Luhr JF, Nelson SA, Allan JF, Carmichael ISE. Active rifting in southwestern Mexico: manifestations of an incipient spreading-ridge jump. Geology (1985) 13:54–57.
Luhr JF, Allan JF, Carmichael ISE, Nelson SA, Hasenaka T. Primitive calc-alkaline and alkaline rock types from the western Mexican Volcanic Belt. Journal of Geophysical Research (1989) 94:4515–4530.
Mathez EA. Influence of degassing on oxidation states of basaltic magmas. Nature (1984) 310:371–375.[CrossRef][Web of Science]
Matthews SJ, Moncrieff DHD, Carroll MR. Empirical calibration of the sulphur valence oxygen barometer from natural and experimental glasses: method and applications. Mineralogical Magazine (1999) 63:421–431.[Abstract]
McDonough WF. Constraints on the composition of the continental lithospheric mantle. Earth and Planetary Science Letters (1990) 101:1–18.[CrossRef][Web of Science]
Meriggi L, Macías JL, Tommasini S, Capra L, Conticelli S. Heterogeneous magmas of the Quaternary Sierra Chichinautzin volcanic field (central Mexico): the role of an amphibole-bearing mantle and magmatic evolution processes. Revista Mexicana de Ciencias Geológicas (2008) 25:197–216.[Web of Science]
Métrich N, Bertagnini A, Landi P, Rosi M. Crystallization driven by decompression and water loss at Stromboli volcano (Aeolian Islands, Italy). Journal of Petrology (2001) 42:1471–1490.
Morrison GW. Characteristics and tectonic setting of the shoshonite rock association. Lithos (1980) 13:97–108.[CrossRef][Web of Science]
Mysen BO. Role of volatiles in silicate melts—solubility of carbon-dioxide and water in feldspar, pyroxene and feldspathoid melts to 30 kb and 1625 degrees C. American Journal of Science (1976) 276(8):969–996.
Newman S, Lowenstern JB. VolatileCalc: a silicate melt–H2O–CO2 solution model written in Visual Basic for Excel. Computers and Geosciences (2002) 28:597–604.[CrossRef]
Nicholls IA, Whitford DJ. Potassium-rich volcanic-rocks of the Muriah Complex, Java, Indonesia—products of multiple magma sources. Journal of Volcanology and Geothermal Research (1983) 18:337–359.[CrossRef][Web of Science]
Nixon GT. The relationship between Quaternary volcanism in central Mexico and the seismicity and structure of subducted ocean lithosphere. Geological Society of America Bulletin (1982) 93:514–523.
Paktunc D. ChrStr: Calculator of structural formula of spinel group minerals from electron microprobe analysis, version 5.0. In: Unpublished program (2006) Department of Earth Sciences, University of Ottawa, Ottawa, Ont.
Papale P, Moretti R, Barbato D. The compositional dependence of the saturation surface of H2O + CO2 fluids in silicate melts. Chemical Geology (2006) 229:78–95.[CrossRef][Web of Science]
Pearce JA, Peate DW. Tectonic implications of the composition of volcanic arc magmas. Annual Review of Earth and Planetary Sciences (1995) 23:251–285.[CrossRef][Web of Science]
Petrone CM, Francalanci L, Ferrari L, Schaaf P, Conticelli S. The San Pedro-Cerro Grande volcanic complex (Nayarit, Mexico): Inferences on volcanology and magma evolution. Neogene-Quaternary continental margin volcanism: A perspective from Mexico. Geological Society of America Special Paper—Siebe C, Macías JL, Aguirre-Díaz GJ, eds. (2006) 402:65–98. doi: 10.1130/2006.2402(03).[CrossRef]
Portnyagin M, Hoernle K, Plechov P, Mironov N, Khununaya S. Constraints on mantle melting and composition and nature of slab components in volcanic arcs from volatiles (H2O, S, Cl, F) and trace elements in melt inclusions from the Kamchatka Arc. Earth and Planetary Science Letters (2007) 255:53–69.[CrossRef][Web of Science]
Portnyagin M, Almeev R, Matveev S, Holtz F. Experimental evidence for rapid water exchange between melt inclusions in olivine and host magma. Earth and Planetary Science Letters (2008) 272:541–552.[CrossRef][Web of Science]
Righter K, Carmichael ISE. Hawaiites and related lavas in the Atenguillo graben, western Mexican Volcanic Belt. Geological Society of America Bulletin (1992) 104:1592–1607.
Righter K, Carmichael ISE. Phase equilibria of phlogopite lamprophyres from western Mexico: biotite–liquid equilibria and P–T estimates for biotite-bearing igneous rocks. Contributions to Mineralogy and Petrology (1996) 123:1–21.[CrossRef][Web of Science]
Righter K, Rosas-Elguera J. Alkaline lavas in the volcanic front of the western Mexican Volcanic Belt: geology and petrology of the Ayutla and Tapalpa volcanic fields. Journal of Petrology (2001) 42:2333–2361.
Righter K, Carmichael ISE, Becker TA, Renne PR. Pliocene–Quaternary volcanism and faulting at the intersection of the Gulf of California and the Mexican Volcanic Belt. Geological Society of America Bulletin (1995) 107:612–626.
Righter K, Dyar MD, Delaney JS, Vennemann TW, Hervig RL, King PL. Correlations of octahedral cations with OH–, O2–, Cl–, and F– in biotite from volcanic rocks and xenoliths. American Mineralogist (2002) 87:142–153.
Roden MF. Origin of coexisting minette and ultramafic breccia, Navajo Volcanic Field. Contributions to Mineralogy and Petrology (1981) 77:195–206.[CrossRef][Web of Science]
Roggensack K. Sizing up crystals and their melt inclusions: a new approach to crystallization studies. Earth and Planetary Science Letters (2001) 187:221–237.[CrossRef][Web of Science]
Roman DC, Cashman KV, Gardner KV, Wallace PJ, Donovan JJ. Storage and interaction of compositionally heterogeneous magmas from the 1986 eruption of Augustine Volcano, Alaska. Bulletin of Volcanology (2006) 68:240–254.[CrossRef][Web of Science]
Rowe MC, Kent AJR, Nielsen RL. Determination of sulfur speciation and oxidation state of olivine-hosted melt inclusions. Chemical Geology (2007) 236(3–4):303–322.[CrossRef][Web of Science]
Rust AC, Cashman KV, Wallace PJ. Magma degassing buffered by vapor flow through brecciated conduit margins. Geology (2004) 32:349–352.
Ryerson FJ, Watson EB. Rutile saturation in magmas—implications for Ti–Nb–Ta depletion in island-arc basalts. Earth and Planetary Science Letters (1987) 86:225–239.[CrossRef][Web of Science]
Schaaf P, Moran-Zenteno D, Hernandez-Bernal M, Solis-Pichardo G, Tolson G, Köhler H. Paleogene continental margin truncation in southwestern Mexico: Geochronological evidence. Tectonics (1995) 14:1339–1350.[CrossRef][Web of Science]
Schmidt KH, Bottassi P, Vannucci R, Mengel K. Trace element partitioning between phlogopite, clinopyroxene and leucite lamproite melt. Earth and Planetary Science Letters (1999) 168:287–299.[CrossRef][Web of Science]
Siebe C, Rodriguez-Lara V, Schaaf P, Abrams M. Geochemistry, Sr–Nd isotope composition, and tectonic setting of Holocene Pelado, Guespalapa and Chichinautzin scoria cones, south of Mexico City. Journal of Volcanology and Geothermal Research (2004) 130(3–4):197–226.[CrossRef][Web of Science]
Sisson TW, Layne GD. H2O in basalt and basaltic andesite glass inclusions from four subduction-related volcanoes. Earth and Planetary Science Letters (1993) 117:619–635.[CrossRef][Web of Science]
Spilliaert N, Allard P, Métrich N, Sobolev AV. Melt inclusion record of the conditions of ascent, degassing, and extrusion of volatile-rich alkali basalt during the powerful 2002 flank eruption of Mount Etna (Italy). Journal of Geophysical Research—Solid Earth (2006a) 111. doi:10.1029/2005JB003934.
Spilliaert N, Métrich N, Allard P. S–Cl–F degassing pattern of water-rich alkali basalt: modelling and relationship with eruption styles on Mount Etna volcano. Earth and Planetary Science Letters (2006b) 248:772–786.[CrossRef][Web of Science]
Stolper E, Newman S. The role of water in the petrogenesis of Mariana trough magmas. Earth and Planetary Science Letters (1994) 121:293–325.[CrossRef][Web of Science]
Sugawara T. Empirical relationships between temperature, pressure, and MgO content in olivine and pyroxene saturated liquid. Journal of Geophysical Research (2000) 105:8457–8472.[CrossRef]
Sun S-s, McDonough WF. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. Magmatism in the Ocean Basins. Geological Society, London, Special Publications—Saunders AD, Norry MJ, eds. (1989) 42:313–345.
Sweeney RJ, Thompson AB, Ulmer P. Phase relations of a natural MARID composition and implications for MARID genesis, lithospheric melting and mantle metasomatism. Contributions to Mineralogy and Petrology (1993) 115:225–241.[CrossRef][Web of Science]
Tatsumi Y, Koyaguchi T. An absarokite from a phlogopite–lherzolite source. Contributions to Mineralogy and Petrology (1989) 102(1):34–40.[CrossRef][Web of Science]
Thibault Y, Edgar AD, Lloyd FE. Experimental investigation of melts from carbonated phlogopite lherzolite: Implications for metasomatism in the continental lithospheric mantle. American Mineralogist (1992) 77:784–794.[Abstract]
Thy P, Lesher CE, Nielsen TFD, Brooks CK. Experimental constraints on the Skaergaard liquid line of descent. Lithos (2006) 92:154–180.[CrossRef][Web of Science]
Toplis MJ. The thermodynamics of iron and megnesium partitioning between olivine and liquid: criteria for assessing and predicting equilibrium in natural and experimental systems. Contributions to Mineralogy and Petrology (2005) 149:22–39.[CrossRef][Web of Science]
Trønnes RG. Stability range and decomposition of potassic richteite and phlogopite end members at 5–15 GPa. Mineralogy and Petrology (2002) 74:129–148.[CrossRef][Web of Science]
Urrutia-Fucugauchi J. Crustal thickness, heat flow, arc magmatism, and tectonics of Mexico—preliminary report. Geofísica Internacional (1986) 25:559–573.
Urrutia-Fucugauchi J, Flores-Ruiz JH. Bouguer gravity anomalies and regional crustal structure in Central Mexico. International Geology Review (1996) 38:176–194.
Verma SP, Nelson SA. Isotopic and trace-element constraints on the origin and evolution of alkaline and calc-alkaline magmas in the northwestern Mexican Volcanic Belt. Journal of Geophysical Research—Solid Earth and Planets (1989) 94(B4):4531–4544.[CrossRef]
Vigouroux N. A melt inclusion study of the potassic magmas of the Colima Volcanic Complex, Mexico. In: M.S. thesis (2006) University of Oregon, Eugene.
Wallace PJ. Volatiles in subduction zone magmas: concentrations and fluxes based on melt inclusion and volcanic gas data. Journal of Volcanology and Geothermal Research (2005) 140:217–240.[CrossRef][Web of Science]
Wallace PJ, Carmichael ISE. Minette lavas and associated leucitites from the western front of the Mexican Volcanic Belt: petrology, chemistry, and origin. Contributions to Mineralogy and Petrology (1989) 103:470–492.[CrossRef][Web of Science]
Wallace PJ, Carmichael ISE. Alkaline and calc-alkaline lavas near Los Volcanes, Jalisco, Mexico: geochemical diversity and its significance in volcanic arcs. Contributions to Mineralogy and Petrology (1992) 111:423–439.[CrossRef][Web of Science]
Wallace PJ, Carmichael ISE. S speciation in submarine basaltic glasses as determined by measurements of SK X-ray wavelength shifts. American Mineralogist (1994) 79:161–167.[Abstract]
Wallace PJ, Carmichael ISE. Quaternary volcanism near the Valley of Mexico: implications for subduction zone magmatism and the effects of crustal thickness variations on primitive magma compositions. Contributions to Mineralogy and Petrology (1999) 135(4):291–314.[CrossRef][Web of Science]
Wallace PJ, Johnson EJ, Delgado-Granados H. Open-system degassing, gas fluxing, and the growth and entrainment of olivine during small volume mafic eruptions: A case study of Volcan Jorullo, central Mexico. EOS Transactions, American Geophysical Union, Fall Meeting Abstracts (2005) (V12A-05).
Wilke M, Jugo P, Klimm K, Susini J, Botcharnikov R, Kohn SC, Janousch M. The origin of S4+ detected in silicate glasses by XANES. American Mineralogist (2008) 93:235–240.
Wilkinson JFG, LeMaitre RW. Upper mantle amphiboles and micas and TiO2, K2O, and P2O5 abundances and 100 Mg/(Mg + Fe2+) ratios of common basalts and andesites—implications for modal mantle metasomatism and undepleted mantle compositions. Journal of Petrology (1987) 28:37–73.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. R. Johnson, P. J. Wallace, H. Delgado Granados, V. C. Manea, A. J. R. Kent, I. N. Bindeman, and C. S. Donegan Subduction-related Volatile Recycling and Magma Generation beneath Central Mexico: Insights from Melt Inclusions, Oxygen Isotopes and Geodynamic Models J. Petrology, September 1, 2009; 50(9): 1729 - 1764. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. J.R. Kent Melt Inclusions in Basaltic and Related Volcanic Rocks Reviews in Mineralogy and Geochemistry, January 1, 2008; 69(1): 273 - 331. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||