Journal of Petrology Volume 42 Number 12 Pages 2333-2361 2001
© Oxford University Press 2001
Alkaline Lavas in the Volcanic Front of the Western Mexican Volcanic Belt: Geology and Petrology of the Ayutla and Tapalpa Volcanic Fields
KEVIN RIGHTER1,* and
JOSÉ ROSAS-ELGUERA2
1LUNAR AND PLANETARY LABORATORY, UNIVERSITY OF ARIZONA, TUCSON, AZ 85721, USA
2DEPARTMENT OF GEOLOGY, UNIVERSIDAD DE GUADALAJARA, AV. REVOLUCION NO. 1500, GUADALAJARA, JALISCO, 44840, MEXICO
Received
March 6, 2000;
Revised typescript accepted
June 15, 2001
 |
ABSTRACT
|
|---|
The Plio-Quaternary Ayutla and Tapalpa volcanic fields in the
volcanic front of the western Mexican Volcanic Belt (WMVB) contain
a wide variety of alkaline volcanic rocks, rather than only
calc-alkaline rocks as found in many continental arcs. There
are three principal rock series in this region: an intraplate
alkaline series (alkali basalts and hawaiites), a potassic series
(lamprophyres and trachylavas), and a calc-alkaline series.
Phlogopite-clinopyroxenite and hornblende-gabbro cumulate xenoliths
from an augite minette lava flow have orthocumulate textures.
The phlogopite-clinopyroxenite xenoliths also contain apatite
and titanomagnetite and probably formed by accumulation of minerals
fractionated from an augite minette more primitive than the
host. The intraplate alkaline series is probably generated by
decompression melting of asthenospheric mantle as a result of
corner flow in the mantle wedge beneath the arc. Alkaline magmas
may be common in the WMVB as a result of prior metasomatism
(during Tertiary Sierra Madre Occidental magmatism) of the Mexican
sub-arc mantle. Generation of the more evolved andesites and
dacites of the calc-alkaline series is due to either combined
assimilation and fractional crystallization (AFC) or magma mixing.
The preponderance of alkaline and hydrous lavas in this region
demonstrates that these lava types are the norm, rather than
the exception in western Mexico, and occur in regions that are
not necessarily associated with active rifting.
KEY WORDS: arc basalt; subduction; alkali basalt; minette; hawaiite; metasomatism
|
INTRODUCTION
|
|---|
In a widely accepted model of subduction zones, fluid-fluxed
mantle peridotite begins to melt at a depth of
100 km, and some
of this melt rises to form the volcanic front of the arc (e.g.
Davies & Stevenson, 1992
; Arculus, 1994
). As a result, magma
erupted at the volcanic front can provide information about
conditions in the sub-arc mantle and the nature of the fluids
and melts being released and generated there. The great diversity
of mafic volcanic rocks produced in the western Mexican Volcanic
Belt (WMVB) (e.g. Allan
et al., 1991
) is due in part to different
amounts and types of fluids fluxing the mantle wedge. However,
the majority of the diverse basic magmas have erupted within
tectonic depressions, making it unclear whether the unusually
diverse compositions are an intrinsic characteristic of the
subduction magmatism or have been amplified by extension. Lavas
from the volcanic front of the WMVB have been studied in the
Tepic, San Sebastian, Los Volcanes, and Mascota regions in the
north (Nixon
et al., 1987
; Verma & Nelson, 1989
; Righter
et al., 1995
; Carmichael
et al., 1996
) and also in the Sayula
and Colima regions in the south (Luhr & Carmichael, 1981
;
Allan & Carmichael, 1984
).
In this paper we focus on the relatively uncharacterized Tapalpa and Ayutla region of the volcanic front, which is apparently unaffected by extension (Fig. 1). The Plio-Quaternary geology, petrography, mineral compositional data, and major and trace element analyses of lavas from this intermediate region are used to obtain a comprehensive overview of volcanism in the WMVB and to better understand its relation to subduction and tectonic processes.
|
GEOLOGY OF THE TECOLOTLAN, AYUTLA AND TAPALPA REGION
|
|---|
Regional geological and tectonic setting
The Mexican Volcanic Belt (MVB) is oriented roughly east–west
and formed in response to subduction of the Cocos Plate beneath
the North American Plate along the Middle America Trench (MAT)
(Burbach
et al., 1984
). The western MVB (WMVB) is formed by
the subduction of the oceanic Rivera Plate beneath the North
American Plate; the subduction zone dips at an angle of 45°
(Pardo & Suarez, 1993
, 1995
). The top of the Rivera Plate
has been imaged by teleseismic studies at a depth of
100 km
beneath the volcanic front (Pardo & Suarez, 1995
; Fig.
1).
Within this part of the volcanic arc are two large-scale extensional
features, the Colima Rift and the Tepic–Zacoalco Rift
on the North American Plate (Luhr
et al., 1985
; Barrier
et al.,
1990
; Allan
et al., 1991
; Fig.
1), which together with the Middle
America Trench define a basement structural block, called the
Jalisco block. The Jalisco block is thus at the junction of
the MVB and the Gulf Extensional Province.
The Tapalpa region (Fig. 2) is located on the western shoulder of the Northern Colima graben. The basement comprises a Late Jurassic to Cretaceous volcanic and sedimentary sequence. The oldest exposed unit of this sequence is a distinctively greenish to violet colored andesite breccia with clasts 0·02–0·50 m in diameter. The bottom of this unit is not exposed, but it is at least 800 m thick. The volcanic breccia is exposed from the Los Volcanes region (Fig. 1) to some 75 km eastward of the Colima Volcanic Complex. Because of its broad distribution, we refer to this unit as the Jalisco breccia. The Jalisco breccia is mantled by the Alberca Formation (Berriasian–Hauterivian; Pimentel, 1980), a volcano-sedimentary sequence with two members: an upper member that consists of black limestone and shale, altered rhyolitic tuffs, some intercalated andesitic lava flows and brownish sandstone, and a lower fossiliferous member (e.g. Neocomites sp., Taraistes Off Neolonese Cantu Leopoldina sp., Late Valanginian–Early Hauterivian; Alarcón, 1983) that is composed of black shale, limestone, and tuffs.
South of Tapalpa (19·780°N, 103·802°W), Petroleos Mexicanos drilled the Tapalpa-1 oil-well, which reaches 389 m below sea level (1754 m depth; Grajales-Nishimura & López-Infanzón, 1983). This well cut 330 m of Alberca Formation and 1246 m of andesitic rocks intruded by granodiorite. The top of the Late Jurassic–Cretaceous sequence is a 79 Ma rhyolitic ash-flow tuff (Rosas-Elguera et al., 1997). Both of these units are intruded by late Cretaceous granitic rocks (69 Ma; Allan, 1986).
The Ayutla volcanic area (Fig. 2) lies at the volcanic front of the WMVB (Fig. 1), at the center of the Jalisco block. The basement in this region is also a Late Jurassic to Cretaceous sequence containing the Jalisco breccia, which, in this case, rests on a detrital succession of red sandstone and conglomerates. Cretaceous silicic ash-flow tuffs are the most abundant volcanic rocks in the Ayutla area, as well as in the Jalisco block. A silicic ash flow a few kilometers west of Ayutla was dated at 70·6 Ma (Righter et al., 1995) and is affected by strike-slip faulting that could correspond to the Cacoma system of Maillol et al. (1997). In the eastern part of the region, near Tecolotlan, is a Late Miocene tectonic depression filled with a succession of detrital sediments composed of three units. The oldest unit is a reddish conglomerate that changes to sandstone transitionally towards the top; cobbles in the conglomerate are silicic volcanic rocks and are smaller than 0·05 m. The middle unit is a white lacustrine mudstone, <4 m in thickness. The top unit is a conglomerate, which during detailed mapping was difficult to distinguish from the Quaternary deposits. However, the absence of mafic clasts in the infill conglomerates is in contradistinction to their presence in Plio-Quaternary units. As a result, the lack of mafic clasts in the infill sediments, together with a K–Ar date (see below), suggest a Late Miocene age for this basin.
Faulting in the western Mexican Volcanic Belt
Late Miocene to Quaternary deformation in the WMVB is concentrated along the boundaries of the Jalisco block: in the Tepic–Zacoalco and Colima rifts (Righter et al., 1995; Rosas-Elguera et al., 1996; Ferrari & Rosas-Elguera, 2000). According to our fault slip analysis, the southern boundaries of the Amatlan, Ameca, and Zacoalco half-graben structures are related to early Pliocene to Quaternary NNE extension (Figs
1 and 3). Using alignments of volcanic vents, NNE extension was proposed for the main volcanic chains along the Tepic–Zacoalco rift, in agreement with fault-slip analysis (Suter, 1991; Rosas-Elguera et al., 1997; Ferrari & Rosas-Elguera, 2000).
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 3. (a) Regional structural map of the western Mexican Volcanic Belt; (b) stereograms of the structural stations, Schmidt projection, lower hemisphere. Main extensional structures along the northern boundary of the Jalisco block: PS, Plan de Barrancas–Santa Rosa graben; AC, AM, and ZA, Amatlan de Canas, Ameca, and Zacoalco half-grabens, respectively. Double arrows are the minimum horizontal stress orientations after Rosas-Elguera et al. (1993, 1996) and Ferrari & Rosas-Elguera (2000). Cacoma and San Sebastian systems after Maillol et al. (1997).  , towns. (1), (2), and (3) in (a) are the structural stations. The stress tensor was calculated according to the method of Angelier (1990). The principal stress directions (S1, S2, and S3) and the tensor shape (phi) are given for each site.
|
|
To the south of the Amatlan half-graben, faulting is related to strike-slip motion along the San Sebastian and Cacoma systems (Fig. 3), which are thought to have formed in response to NW motion of the Jalisco block (Maillol et al., 1997). Our preliminary structural study shows evidence for strike-slip faulting only in the underlying Jurassic to Cretaceous basement (Fig. 3). The Neogene volcanic rocks are not affected by faulting (see Fig. 2a and b); furthermore, that the volcanic vents of the Tapalpa and Ayutla regions are not aligned is additional evidence that faulting ceased in the Neogene. Similar observations can be made for the rest of the volcanic front (Wallace & Carmichael, 1992; Carmichael et al., 1996). Regional geological mapping (Consejo de Recursos de Minerales, 1999; Ferrari et al., 2000) shows that none of the Pliocene to Quaternary volcanic rocks inside the Jalisco block are displaced by faults, yet lava flows of the same age at the edges of the Jalisco block are commonly displaced. We speculate that deformation within the Jalisco block, related to regional NNE extension, is partitioned as (1) reactivation of older (pre-Pliocene) faults within the Jalisco block (e.g. the Cacoma–San Sebastian system), and (2) the emplacement of dikes for the Plio-Quaternary volcanism that produced the western front of the MVB (e.g. Ayutla, Tapalpa, and Mascota volcanic fields).
Plio-Quaternary volcanism
Young volcanism in western Mexico takes three dominant forms. The first type consists of the large central andesitic volcanoes, such as Volcans Tequila, Ceboruco, and San Juan (between Tepic and Guadalajara), that define an overall SE–NW trend of the western Mexican Volcanic Belt (Fig. 1). The second type comprises smaller volume lava and pyroclastic flows and cones in the volcanic front of western Mexico, which are either alkaline or calc-alkaline in composition—many are lamprophyric, with hydrous phenocrysts (Verma & Nelson, 1989; Wallace & Carmichael, 1989, 1992; Lange & Carmichael, 1990, 1991; Righter et al., 1995). Alkali basalt, hawaiite and mugearite shield volcanoes, plateaux, and lines of cinder cones are the third type and are interspersed among the central and smaller volume volcanoes (Nelson & Carmichael, 1984; Nixon et al., 1987; Verma & Nelson, 1989; Righter & Carmichael, 1992; Righter et al., 1995). It has been proposed that the hydrous magmas from the WMVB, together with those from Jorullo (Luhr & Carmichael, 1985) and the Colima area (Luhr & Carmichael, 1980, 1981; Allan & Carmichael, 1984), define a hydrous volcanic front in western Mexico. The latter two types of magmatism are also found in the Ayutla (Fig. 2a) and Tapalpa (Fig. 2b) regions, located between the Colima Volcanic Complex and the Mascota–Talpa–Atenguillo–Los Volcanes areas (Fig. 1). With this characterization of the Ayutla and Tapalpa volcanic regions, coverage of the entire western volcanic front of the MVB is complete.
The alkaline and calc-alkaline volcanism in the Tapalpa and Ayutla regions (Fig. 2) is Pliocene in age. Allan (1986) reported a K–Ar age of 3·29 Ma for a phlogopite–hornblende lamprophyre lava flow in the Tapalpa area (Fig. 2b). Pliocene ages (4·4 Ma) for andesitic rocks were obtained for an area located some 10 km NNE of Tapalpa (Rosas-Elguera et al., 1997). We obtained a K–Ar age of 4·5 ± 0·2 Ma for a trachyandesite (sample 504; Table 1). In comparison, the oldest lava flows dated in the Los Volcanes region to the north (Fig. 1) are hornblende trachybasalt (3·43 Ma) and olivine minette (2·94 Ma; Wallace & Carmichael, 1992). For the San Sebastian, Mascota–Talpa, and Colima regions (Fig. 1), lavas are the youngest of the volcanic front (<1·5 Ma, Lange & Carmichael, 1991; Hooper, 1995; Carmichael et al., 1996). Thus, the age of volcanism in the Ayutla–Tapalpa region appears to be intermediate between the younger Mascota–Talpa and Colima regions (Fig. 1; Carmichael et al., 1996) and the contemporaneous to older northern Colima graben, where some of the volcanism is as old as 5 Ma [e.g. 5·17 Ma hornblende andesite of Allan (1986)].
|
PETROGRAPHY, MINERALOGY AND COMPOSITION OF PLIO-QUATERNARY LAVAS
|
|---|
Modal analyses for the Ayutla and Tapalpa lavas are based on
700–1400 points counted for each thin section (Table
1).
The small size of some of the cumulate xenoliths prevented a
larger number of counts (Table
4). Major, minor and trace element
analyses were obtained by wavelength-dispersive X-ray fluorescence
techniques (Tables
2 and
3), using a Rigaku 3370 X-ray Spectrometer
at Washington State University (Johnson
et al., 1999
). Samples
for fused bead analysis were prepared by mixing 3·5 g
of sample with 7·0 g Li
2B
4O
7 flux, and then fusing the
mixture into glass disks. Standards used in these analyses were
a set of USGS (US Geological Survey) standards (PCC-1, BCR-1,
BIR-1, DNC-1, W-2, AGV-1, GSP-1, G-2, and STM-1). Loss on ignition
measurements (Table
1) were performed by heating
2 g of rock
powder in a preheated ceramic crucible to 1000°C for 2 h
and determining the resulting weight loss. Phenocryst and groundmass
mineral phases were analyzed on a CAMECA SX50 electron microprobe,
using an accelerating voltage of 15 kV and beam currents of
10–30 nA (on MgO). Standards include both natural (albite,
diopside, potassium feldspar, fayalite, rhodonite, apatite,
chromite) and synthetic (Mn, Ni, V, TiO
2) materials. PAP
–
–Z
corrections were used in the data reduction (Pouchou & Pichoir,
1991
). Microprobe analyses of groundmass, Fe–Ti oxides,
olivine, pyroxene, amphibole, phlogopite, analcime, and feldspars
are presented in Tables
5–12
.
Three distinct rock series are represented in the Ayutla and
Tapalpa regions: intraplate alkaline (IA) with compositional
similarities to ocean-island basalt (OIB), a calc-alkaline suite
(CA), and a potassic series that includes hydrous magmas such
as lamprophyres, and relatively dry magmas such as shoshonites,
absarokites and other trachylavas. We use the term intraplate
alkaline in place of OIB so as to emphasize their continental
setting. These series are distinguished using major element
characteristics, but they have distinct trace element characteristics
as well.
A total alkalis and SiO2 diagram (Fig. 4) highlights the three rock series. The IA suite is a sodic series that is distinguished by having Na2O – 2 > K2O, and includes alkali basalt, hawaiite, mugearite, benmoreite and trachyte; the sodic character of the hawaiites and mugearites is reflected in their sodic groundmass feldspar (Muir & Tilley, 1961; Nelson & Carmichael, 1984; Righter & Carmichael, 1992). The potassic suites are more variable and include lamprophyres with amphibole and phlogopite phenocrysts, as well as absarokites, shoshonites and other trachylavas, with only olivine and augite phenocrysts. The lamprophyres are defined according to Williams et al. (1982): for example, minettes contain biotite phenocrysts and potassium feldspar is confined to the groundmass. Some of the alkaline rocks are not particularly sodic, but are nonetheless alkaline and plot in the trachybasalt, basaltic trachyandesite and trachyandesite fields—we here refer to these collectively as ‘trachylavas’ (see also Fig. 6). Calc-alkaline suite plots in the subalkaline field of total alkalis vs silica (LeBas et al., 1986). Finally, a number of criteria were used to identify primitive lavas—those that have undergone little change since they formed in the mantle. Such criteria include Ni >250 ppm, Cr >300 ppm, MgO >7 wt %, and mg-number >0·65.
Intraplate alkaline (alkali basalt and hawaiite)
Several lavas near Ayutla (AY-509, -510) contain sparse phenocrysts of olivine (Fig. 5a), groundmass andesine microphenocrysts, normative andesine (An37–40), have a differentiation index (DI) between 30 and 45 (Table 3), and K2O < Na2O – 2, indicating that these lavas belong to the alkali basalt–hawaiite series (e.g. Muir & Tilley, 1961; MacDonald, 1968; Thompson, 1972; Nelson & Carmichael, 1984; LeBas et al. 1986). These lavas plot just above the alkaline–subalkaline line of MacDonald & Katsura (1964) in an alkali vs SiO2 diagram (Fig. 4), and are primitive as indicated by their high MgO (9·84–10·08 wt %), Ni (224–230 ppm) and Cr (370 ppm) abundances (Table 3). An olivine-bearing lava collected near Tecolotlan along Highway 70 is hawaiite, with 51·8 wt % SiO2 and 5·7 wt % MgO (AY-501; Table 3).
View larger version (167K):
[in this window]
[in a new window]
|
Fig. 5. (a) Photomicrograph of primitive alkali basalt AY-510, showing olivine phenocrysts in a trachytic groundmass with plagioclase laths. (b) Photomicrograph of augite minette AY-506, with augite and phlogopite phenocrysts in a sanidine-bearing groundmass.
|
|
Potassic series
The minette suite from the WMVB is K-rich and stands out clearly from the other alkaline and calc-alkaline series in a plot of K2O vs MgO (Fig. 6). An augite minette lava that contains augite, phlogopite and large apatite phenocrysts (Fig. 5b) is present just SW of Tenamextlan (AY-506). The low MgO content of this sample (Table 2 and Fig. 6) indicates that it is an evolved member of the augite minette suite. Similar lavas are present in the Mascota, Los Volcanes and San Sebastian regions (Lange & Carmichael, 1991; Wallace & Carmichael, 1992; Carmichael et al., 1996). Hornblende gabbro and phlogopite clinopyroxenite xenolith inclusions are abundant within this augite minette (see section below). Another augite minette flow (AY-507) occurs just to the east of AY-506, and has a very similar bulk composition and petrography, but lacks xenoliths. In the Tapalpa region, there are two different minettes: a primitive olivine minette with 8·6 wt % MgO (TAP-04) and an evolved minette with phlogopite phenocrysts and analcime microphenocrysts (TAP-08; Table 1 and Fig. 7b). East of the town of Tapalpa is a phlogopite hornblende lamprophyre dated by Allan (1986) at 3·29 Ma (Fig. 2b).
View larger version (103K):
[in this window]
[in a new window]
|
Fig. 7. (a) Backscattered electron image of orthocumulate-textured phlogopite clinopyroxenite AY-506A with cumulus phlogopite, augite and apatite and intercumulus titanomagnetite. (b) Backscattered electron image of minette TAP-08 showing phlogopite phenocrysts and analcime microphenocrysts in a groundmass consisting of alkali feldspar, augite, apatite and titanomagnetite.
|
|
Trachylavas are common in the Ayutla and Tapalpa regions: four trachylava flows south of Atengo are highly alkaline and include an olivine-bearing basaltic trachyandesite (AY-505), an olivine–augite–amphibole-bearing basaltic trachyandesite (AY-508; Fig. 2a), and hornblende-bearing trachylavas (AY-504 and AY-500) with high alkali contents, from 6 to 8 wt % Na2O + K2O (Fig. 4). Sample AY-504 yielded a K–Ar whole-rock age of 4·6 Ma. In the Tapalpa area are two olivine–augite-bearing trachylavas (TAP-07 and TAP-12) and an olivine–augite–amphibole-bearing basaltic trachyandesite (TAP-15; Fig. 2b). Samples AY-500, -504, -508 and TAP-15 are lamprophyric because they contain hornblende phenocrysts, but here they are called trachylavas for simplicity of terminology.
Calc-alkaline series (andesite)
South of Tenamextlan, along Highway 80, is an andesite flow (AY-503), with olivine, plagioclase and amphibole phenocrysts. Calc-alkaline lavas are sparse in the Ayutla and Tapalpa regions, and the low alkali content of AY-503 (Fig. 4) makes this the only subalkaline sample from the entire region.
Cumulate xenoliths
Among the xenoliths we collected from the augite minette flow (AY-506) are three phlogopite clinopyroxenites with orthocumulate textures. They contain cumulus apatite, phlogopite, magnetite and post-cumulus potassium feldspar and augite (Table 4; Fig. 7a). A fourth phlogopite clinopyroxenite xenolith contains all the same phases, but is a mesocumulate with only post-cumulus augite. All four of these xenoliths contain a large amount of apatite, from 6 to 21 modal % (Table 4). A fifth gabbroic xenolith is distinct from the others in mineralogy, but is also an orthocumulate. It contains cumulus amphibole and post-cumulus plagioclase feldspar, with minor Fe–Ti oxide and augite.
|
INTENSIVE VARIABLES IN THE QUATERNARY LAVAS AND XENOLITHS
|
|---|
Calculation of the temperatures, pressures, water contents and
oxygen fugacities of the Ayutla and Tapalpa region lavas can
help to determine their formation conditions and origin. The
results for each suite of lavas are summarized below.
Intraplate alkaline series
The olivine and whole-rock compositions of AY-509 and -510 indicate that the former are phenocrysts and not xenocrysts; the Fo90 cores are the expected olivine composition to crystallize from the bulk composition assuming a ferric/ferrous ratio for an oxygen fugacity of NNO + 1 (where NNO is nickel–nickel oxide) (Snyder & Carmichael, 1992). The olivine–liquid compositions yield temperatures between 1070°C (hawaiite) and 1225°C (alkali basalt) (Table 13; olivine–liquid thermometer of Sisson & Grove, 1993b; ±11°C).
Liquids in equilibrium with olivine, orthopyroxene and clinopyroxene from 5 to 30 kbar can be represented on the CIPW molecular normative basalt tetrahedron, projected from diopside onto the base jadeite + Ca-Tschermak’s molecule (Jd + CaTs)–quartz (Qz)–olivine (Ol) following Falloon & Green (1988). Because the IA series in western Mexico contains very low water contents (Righter & Carmichael, 1992; Righter et al., 1995), these dry phase equilibria can be used to estimate pressures of equilibration. Primitive IA basaltic lavas (>7 wt % MgO) from western Mexico fall mainly about cotectics in the 8–12 kbar range (Fig. 8) in the (Jd + CaTs)–Qz–Ol system. The highest-pressure liquids from western Mexico are the alkali basalts AY-509 and -510, which fall between the 15 and 20 kbar cotectics (Fig. 8).
Potassic series
Phlogopite–groundmass and olivine–groundmass pairs for the olivine minette give temperatures from 1170 to 1190°C, consistent with its higher MgO, Ni and Cr contents (Table 13), and higher than the results for the augite minettes (1030 and 1140°C; Table 13) and the analcime-bearing minette (TAP-08; 1100°C). Phase relations in the diopside–sanidine–olivine–H2O (Diop–San–Ol) system indicate that many of the lamprophyres from the WMVB region were generated by melting of phlogopite- and augite-bearing mantle rocks, at pressures from 10 to 25 kbar (Wallace & Carmichael, 1989). This pressure range is also consistent with the BaO and TiO2 contents of phlogopite phenocrysts from minettes in the WMVB—Righter & Carmichael (1996) demonstrated that phlogopite in many of the lamprophyres grew at depths close to the base of the crust in western Mexico. This is also supported in the present study as the minettes from the Ayutla and Tapalpa region yield pressures of between 6 and 15 kbar (Table 13).
The augite minettes (AY-506) crystallized under oxidized conditions as indicated in the composition of iron–titanium oxide lamellae intergrowths found in a single phenocryst. High relative oxygen fugacities are calculated (
NNO = +4·3; Table 7) using the two-oxide thermometer of Sack & Ghiorso (1991). Recent work suggests that the two-oxide thermometers overestimate fO2 by perhaps an order of magnitude (Evans & Scaillet, 1997), but even considering this, the calculated fO2 values are high. Furthermore, reintegration of these lamellae into a single oxide composition (Table 7) yields ferropseudobrookite, a stable phase in experimental studies of minettes at oxidized conditions (NNO = 0 to NNO = +5; Righter & Carmichael, 1996).
The lamprophyres are clearly water-rich magmas, as attested by the presence of hydrous phenocrysts. Recent experimental work has placed constraints on how much water is necessary to stabilize hydrous phenocrysts in these liquids. For example, Righter & Carmichael (1996) demonstrated that augite and olivine minette liquids require 3·0–5·5 wt % dissolved water at 1–2 kbar (saturated with H2O fluid), to stabilize phlogopite. Similarly, Moore & Carmichael (1998) demonstrated that basaltic trachyandesites require 4–6 wt % water to stabilize amphibole near the liquidus. It is important to emphasize that magmas containing augite and olivine phenocrysts may contain as much as 3 wt % H2O and still not saturate with hydrous phenocrysts.
Temperatures calculated for the trachylavas range from 970 to 1090°C (Table 13). Such wet and low-temperature magmas are consistent with the presence of amphibole and olivine phenocrysts, as the former are stable only at temperature <1050°C (Moore & Carmichael, 1998), and the latter has an expanded stability field in hydrous systems (Kushiro, 1969). The slightly higher temperatures in AY-504 may reflect a higher fluorine component in the hornblende (thus stabilizing it to higher temperatures), but the uncertainty inherent in two-pyroxene thermometry (±50°C) also allows AY-504 to be perched at the ‘hornblende-in’ line (e.g. Moore & Carmichael, 1998), without plagioclase.
Calc-alkaline series
Application of plagioclase–hornblende thermometry (Holland & Blundy, 1994) to AY-503 yields a temperature of 720°C (at 2 kbar; Table 13). It is well known that water will suppress plagioclase stability (e.g. Sisson & Grove, 1993a), thus the absence of plagioclase in many of the mafic rocks along the volcanic front of this region attests to high magmatic water contents. However, plagioclase is present in sample AY-503, and the plagioclase–liquid hygrometer calibrated by Housh & Luhr (1991) was used to calculate pre-eruptive water contents (using data from Tables
5 and 12) of 2·8–3·8 wt % (Table 12). These are similar to water contents of two-pyroxene andesites from the Nayarit region (Righter et al., 1995) with 3·0–3·4 wt % H2O, but lower than water contents of hornblende andesites with 3·7–5·7 wt % H2O, all calculated by the same method. This pattern has also been observed in andesites from the stratovolcanoes, Colima and Ceboruco (Luhr, 1992): hornblende-free Ceboruco flows have low water contents (0·5–2·2 wt %), whereas the hornblende-bearing Colima andesites have higher water contents (2·5–4·5 wt %).
Xenoliths
No common thermometers or barometers are represented in the assemblages in the phlogopite clinopyroxenites, but indirect evidence from phlogopite and augite compositions suggests that they formed during low-temperature fractionation of a more mafic or primitive augite minette liquid. Comparison of major (MgO, FeO) and minor (Al2O3, TiO2) oxides in both clinopyroxenes and phlogopites shows that the xenolith minerals overlap the range of compositions represented in augite and phlogopite phenocrysts in augite minettes from the San Sebastian, Mascota, Los Volcanes, Ayutla, Tapalpa and Colima areas (Fig. 9). Phlogopites with lower TiO2 (2–3 wt %) are from hotter, more primitive magmas, whereas those with higher (3–5 wt %) TiO2 were derived from fractionation at lower temperatures (from a more evolved magma; see Righter & Carmichael, 1996). The preponderance of high-TiO2 phlogopite in the cumulate xenoliths suggests that they formed at the lower temperatures more typical of evolved minette magmas (Fig. 9).
Equilibration temperatures in the hornblende gabbro have been calculated using two methods—the plagioclase–hornblende thermometer of Holland & Blundy (1994) and the Fe–Ti oxide thermometer of Sack & Ghiorso (1991). The former yields a temperature of 1041°C for a pressure of 2 kbar (XAb = 0·48 and Si p.f.u. = 6·15), whereas the latter yields a temperature of 1030°C and log fO2 = -10·06 (Table 5). These are typical temperatures for andesitic magmas and similar to temperatures calculated for other WMVB lavas (e.g. Righter et al., 1995). The T–fO2 value is 0·29 log fO2 units below the value defined by the NNO oxygen buffer (Fig. 10) and at the low end of the range of NNO values calculated for whole-rock basaltic andesites and andesites from many WMVB localities (e.g. Wallace & Carmichael, 1989; Lange & Carmichael, 1990; Righter et al., 1995). It is not clear whether this gabbro is related to the Plio-Quaternary magmatism or is part of an older intrusive complex, which has been observed nearby (e.g. Lange & Carmichael, 1991). The hornblende is more potassic than any hornblende phenocrysts from calc-alkaline rocks from the WMVB (e.g. Luhr & Carmichael, 1980; Allan & Carmichael, 1984; Lange & Carmichael, 1991; Wallace & Carmichael, 1992; Righter et al., 1995), but closer in composition to hornblendes from alkaline rocks from the region, suggesting a link to this series, if any.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 10. Top: histogram of NNO values for basic rocks in western Mexico. Oxygen fugacities for all samples are calculated using whole-rock Fe2O3, FeO and the approach of Kress & Carmichael (1991). Intraplate alkaline data are from Nelson & Carmichael (1984), Righter & Carmichael (1992) and Righter et al. (1995); CAB data from Luhr et al. (1989), Lange & Carmichael (1990), Wallace & Carmichael (1992) and Carmichael et al. (1996); potassic data from Luhr & Carmichael (1981), Allan & Carmichael (1984), Wallace & Carmichael (1989, 1992), Lange & Carmichael (1991), Righter et al. (1995) and Carmichael et al. (1996); and Tamayo Fracture Zone values are from Bender et al. (1984) and Christie et al. (1986). Bottom: values of NNO calculated using results of two-oxide thermometry (Sack & Ghiorso, 1991) on groundmass oxides. Ayutla and Tapalpa samples are those from Table 7, and samples from other WMVB localities are from the studies of Nelson & Carmichael (1984), Wallace & Carmichael (1989) (augite minette), Righter & Carmichael (1992), Wallace & Carmichael (1992) (trachylava) and Righter et al. (1995) (IA). All oxygen fugacities calculated relative to the nickel–nickel oxide buffer determined by Huebner & Sato (1970).
|
|
|
DISCUSSION
|
|---|
Intraplate alkaline basalt in western Mexico
Intraplate alkaline basalts have been found all over the WMVB
(Nelson & Carmichael, 1984
; Verma & Nelson, 1989
; Wopat,
1990
; Righter & Carmichael, 1992
; Righter
et al., 1995
),
yet few are primitive and most have MgO <7 wt %. Four lines
of evidence suggest that the Ayutla alkali basalts (AY-509,
-510; Tables
1–3
) are truly primitive liquids, having
undergone little change since melting from the mantle. First,
this bulk composition would be in equilibrium with olivine of
Fo
90 composition (based on an olivine liquid
Kd of 0·32;
Snyder & Carmichael, 1992
)—identical to the olivine
phenocryst core compositions in both lavas (Table
6). Second,
the Cr and Ni contents of these lavas are high despite the absence
of spinel inclusions in the olivines, commonly found in other
olivine-bearing basic arc liquids. Third, Os isotopic values
measured for the alkali basalts are generally low, indicating
very little assimilation or crustal interaction (Chesley
et al., 1999
). Finally, these lavas plot along a higher pressure
(15–20 kbar) cotectic in the (Jd + CaTs)–Qz–Ol
ternary of the CIPW molecular normative basalt tetrahedron (Fig.
8).
Decompression melting of asthenosphere advected into the shallow mantle wedge by corner flow has been advocated as a mechanism for generation of IA magmas in western Mexico (Luhr, 1997). Such a scenario is consistent with both phase equilibria and trace element constraints. In a previous section it was shown that many of the IA basalts (and hawaiites) fall along moderate pressure (8–12 kbar) cotectics (Jd + CaTs)–Qz–Ol system. Also, in a general sense, the La/Yb ratios of basaltic liquids should reflect the presence or absence of garnet in the source region—high La/Yb suggests a garnet-bearing source, whereas low La/Yb suggests a garnet-free source. Western Mexican IA-type basalts have mid-ocean ridge basalt (MORB)-like, low La/Yb, and high Yb concentrations (Fig. 11) indicating a garnet-free or shallow source for the IA series (Luhr, 1997). In addition, the crust in western Mexico is thin (39–40 km; Urrutia-Fucugauchi & Hernan Flores-Ruiz, 1996), indicating the availability of mantle at shallow depth beneath the arc.
Origin of the wet alkaline volcanic front—metasomatism of sub-arc mantle during Sierra Madre Occidental magmatism
One of the most fundamental characteristics of volcanism in western Mexico is the preponderance and diversity of wet and dry alkaline basic magmas in the volcanic front. These rock types are uncommon in volcanic arcs and indicate involvement of unusual mantle sources. It is likely that the sub-arc mantle beneath western Mexico has been modally metasomatized over time, with subduction tectonics in operation from as far back as the Jurassic, and certainly on a greater scale during emplacement of the Sierra Madre Occidental (SMO; see Fig. 1 inset) volcanic province from 50 to 20 Ma (e.g. McDowell & Clabaugh, 1979; Aguirre-Diaz & McDowell, 1991; Nieto-Samaniego et al., 1999). Boron abundances in MVB rocks prompted Hochstaedter et al. (1996) to propose that Ba, Sr, and K in lamprophyres could be explained by melting of previously modally metasomatized mantle, whereas the B, U, and Cs concentrations in the lamprophyres, as well as in basalts in other series, are attributable to fluids derived from current subduction of the Rivera Plate. Similarly, Lange & Carmichael (1991) proposed that the oxidized lamprophyres along the volcanic front of western Mexico were produced by melting of veined metasomatized mantle peridotite that formed during subduction modification in the production of the SMO. Although there is no SMO present within the Jalisco block, there is a magmatic phase of the same age (52–45 Ma; Zimmerman et al., 1988; Lange & Carmichael, 1991) represented in the San Sebastian and Ameca regions, near the center of the Jalisco block (Fig. 1). Mexican peridotite xenoliths provide further support for prior subduction modification of the sub-arc mantle, because samples from localities closer to the paleotrench record higher oxygen fugacities than those up to 600 km away (Luhr & Aranda-Gomez, 1997). Such a metasomatized phlogopite-bearing peridotite would provide the necessary source to derive the olivine and augite minettes by small degrees of melting (Wallace & Carmichael, 1989; Carmichael et al., 1996).
The presence of these alkaline magmas along the volcanic front in Mexico is thus linked to the prior geological and plate tectonic history of the subduction zone. This previously metasomatized mantle is tapped in western Mexico not only as a result of subduction, but also because of the prevalence of extension within that part of the arc, allowing magmas easier passage to the surface (e.g. Luhr, 1997). The presence of such suites in other arcs would thus depend upon whether at some point the sub-arc mantle had been metasomatized in some previous episode of magmatism, and whether there is intra-arc extension. The absence of lamprophyres in the Cascades arc, despite many similarities to western Mexico including intra-arc extension, may be a result of a very different previous history of the Cascades sub-arc mantle compared with that of western Mexico (Righter, 2000). Specifically, the Cascades sub-arc mantle may not have experienced long-term metasomatism, as did the mantle beneath the MVB. The presence of lamprophyres and other alkaline magmas in arcs can best be interpreted with full knowledge of the arc’s geological history in hand.
Calc-alkaline basalt series
Elevated alkaline earth (AE; Ba, Sr)/light rare earth element (LREE; La, Ce, Nd), and LREE/HFSE (high field strength elements; Zr, Nb, Ti) (Perfit et al., 1980; Pearce & Peate, 1995) compared with oceanic basalts is a nearly universal characteristic of calc-alkaline arc magmas. Primitive calc-alkaline basalt is mainly formed by melting of sub-arc mantle peridotite after infiltration by hydrous, incompatible-element-enriched fluids. For instance, Wallace & Carmichael (1999) showed that the Zr–Ti–Nb characteristics of primitive calc-alkaline lavas from across the Mexican Volcanic Belt can be produced by variable degrees of melting of depleted mantle, yet many of these same lavas have AE/HFSE ratios higher than expected from derivation solely from depleted mantle. The latter require additions of an AE- and LREE-enriched subduction-related component, perhaps similar to the ‘subduction component’ plotted in Fig. 12 [as calculated by Wallace & Carmichael (1999)]. Although crustal interaction is clearly important in more evolved rocks such as basaltic andesites and andesites [see next section and also Nelson et al. (1995)] the primitive basalts have 187Os/188Os values that are the same as the mantle, indicating little to no crustal input (e.g. Chesley et al., 1999; Lassiter & Luhr, 2001).
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 12. (a) Ba/La vs Ba, (b) La/Nb vs La and (c) Ba/Zr vs Zr for WMVB lavas. Open symbols are data from previous studies (see references below) and filled symbols are samples from the Ayutla–Tapalpa region. Also shown (large  ) are total crust (TC) and lower crust (LC) values from continental arcs (tabulated by Rudnick & Fountain, 1995), the composition of the sediment column on the Cocos Plate (Plank & Langmuir, 1999), and the composition of the H2O-rich subduction component of Wallace & Carmichael (1999). Mexico intraplate alkaline data are from Nelson & Carmichael (1984), Wopat (1990), Righter & Carmichael (1992), Righter et al. (1995), this study; CAB data from Luhr et al. (1989), Lange & Carmichael (1990), Wallace & Carmichael (1992), Carmichael et al. (1996), this study; potassic data from Luhr & Carmichael (1981), Allan & Carmichael (1984), Wallace & Carmichael (1989, 1992), Lange & Carmichael (1991), Righter et al. (1995), Carmichael et al. (1996), this study. Horizontal dashed lines are the depleted (bottom line) and primitive (top line) mantle ratios from Sun & McDonough (1989).
|
|
As opposed to the intraplate alkaline basalt series discussed in a previous section, the calc-alkaline basalts have low Yb values (Fig. 11) indicating initiation of mantle melting in the garnet stability field. This has also been documented in the Cascades (Conrey et al., 1997), and the Andes (Hildreth & Moorbath, 1988), where primitive basalts have <2 ppm Yb.
Sediment subduction and subduction erosion are important processes shaping tectonics and magmatism in convergent margins (e.g. von Huene & Scholl, 1991), and the size, shape, and distribution of sediment and forearc material is distinct within every arc. The Mexican forearc contains no sediment wedge, despite a thin sediment layer on the Rivera Plate (Timofeev et al., 1983; Couch et al., 1991). This indicates that there may be a significant amount of sediment subducted beneath western Mexico. Righter et al. (1995) demonstrated that the elevated Sr/Zr and Ba/La ratios of some alkalin
TOP
ABSTRACT
INTRODUCTION
) along the volcanic front of the western Mexican Volcanic Belt. Pliocene–Quaternary motions (thick arrows) of the Jalisco and Michoacan blocks relative to North America and their relation to coeval extensional deformation zones (Rosas-Elguera et al., 1996
