| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology Volume 42 Number 3 Pages 555-626 2001
© Oxford University Press 2001
Eruptive Stratigraphy of the Tatara–San Pedro Complex, 36°S, Southern Volcanic Zone, Chilean Andes: Reconstruction Method and Implications for Magma Evolution at Long-lived Arc Volcanic Centers
1SECTION DES SCIENCES DE LA TERRE, UNIVERSITÉ DE GENÈVE, 13, RUE DES MARAÎCHERS, 1211 GENÈVE 4, SWITZERLAND
2DEPARTMENT OF GEOSCIENCES, UNIVERSITY OF MASSACHUSETTS, AMHERST, MA 01003-5820, USA
3US GEOLOGICAL SURVEY (MS 913), DENVER FEDERAL CENTER, DENVER, CO 80225, USA
Received July 1, 1999; Revised typescript accepted July 21, 2000
 |
ABSTRACT |
|---|
TOP
ABSTRACT
CONTENTSThe Quaternary Tatara–San Pedro volcanic complex (36°S, Chilean Andes) comprises eight or more unconformity-bound volcanic sequences, representing variably preserved erosional remnants of volcanic centers generated during
930 ky of activity. The internal eruptive histories of several dominantly mafic to intermediate sequences have been reconstructed, on the basis of correlations of whole-rock major and trace element chemistry of flows between multiple sampled sections, but with critical contributions from photogrammetric, geochronologic, and paleomagnetic data. Many groups of flows representing discrete eruptive events define internal variation trends that reflect extrusion of heterogeneous or rapidly evolving magma batches from conduit–reservoir systems in which open-system processes typically played a large role. Long-term progressive evolution trends are extremely rare and the magma compositions of successive eruptive events rarely lie on precisely the same differentiation trend, even where they have evolved from similar parent magmas by similar processes. These observations are not consistent with magma differentiation in large long-lived reservoirs, but they may be accommodated by diverse interactions between newly arrived magma inputs and multiple resident pockets of evolved magma and/or crystal mush residing in conduit-dominated subvolcanic reservoirs. Without constraints provided by the reconstructed stratigraphic relations, the framework for petrologic modeling would be far different. A well-established eruptive stratigraphy may provide independent constraints on the petrologic processes involved in magma evolution—simply on the basis of the specific order in which diverse, broadly cogenetic magmas have been erupted. The Tatara–San Pedro complex includes lavas ranging from primitive basalt to high-SiO2 rhyolite, and although the dominant erupted magma type was basaltic andesite (52–55 wt % SiO2) each sequence is characterized by unique proportions of mafic, intermediate, and silicic eruptive products. Intermediate lava compositions also record different evolution paths, both within and between sequences. No systematic long-term pattern is evident from comparisons at the level of sequences. The considerable diversity of mafic and evolved magmas of the Tatara–San Pedro complex bears on interpretations of regional geochemical trends. The variable role of open-system processes in shaping the compositions of evolved Tatara–San Pedro complex magmas, and even some basaltic magmas, leads to the conclusion that addressing problems such as arc magma genesis and elemental fluxes through subduction zones on the basis of averaged or regressed reconnaissance geochemical datasets is a tenuous exercise. Such compositional indices are highly instructive for identifying broad regional trends and first-order problems, but they should be used with extreme caution in attempts to quantify processes and magma sources, including crustal components, implicated in these trends. KEY WORDS: Andean volcanism; Tatara–San Pedro complex; magmatic differentiation; volcanic stratigraphy; petrologic modeling
|
CONTENTS |
|---|
|
|
INTRODUCTION |
|---|
Placing the eruptive record of a prehistoric volcanic center into a well-calibrated temporal framework is essential to meaningful modeling of the origin and evolution of its magmas. The first steps in reconstructing the eruptive history of an edifice are to define appropriate stratigraphic units and then to establish their relative ages. High-precision geochronologic studies, with superposition relations as an internal check on the reliability of direct or indirect dating techniques, may provide sufficient temporal resolution to permit estimates of (1) the ages and durations of eruptive phases, (2) the lengths of time gaps between them, and (3) long-term volumetric eruption rates. In addition, they may eliminate or support correlations among similar but physically separate units (e.g. Bacon, 1983
; Hildreth & Lanphere, 1994; Druitt et al., 2000). Whereas these tenets are self-evident, establishing a high-resolution stratigraphic framework for the entire history of any long-lived arc volcanic complex is a difficult and resource-intensive exercise.
Apart from uncertainties associated with geochronological measurements, investigations of large volcanic centers are intrinsically limited in terms of attainable stratigraphic control, as a result of incomplete preservation of eruptive products and limited exposures. The fraction of a volcano’s eruptive products that is accessible for sampling depends on natural exposures that penetrate the flanks and summit region of the edifice, and this fraction will be different for every volcano as a complex function of the non-uniform rates (temporally and spatially) of aggradation (growth and burial) and degradation (collapse, mass wasting, and erosive excavation of valleys), and of the original geometry and distribution of eruptive units. Even where the flanks of a large central-vent volcano may be sampled in deep valleys, on caldera walls, or on sector collapse scarps, individual vertical sections will inevitably be under-representative of the overall history of a volcanic center because of the consequences of edifice geometry. The following factors may play important roles: (1) the often eccentric positions of point-source and fissure vents relative to a volcano’s summit frequently result in correspondingly asymmetrical distributions of both effusive and pyroclastic eruption products; (2) lavas characterized by relatively high viscosities and/or low effusion rates may be restricted to vent-proximal locations (e.g. Rhodes, 1996); (3) lava flows are confined to topographic depressions generated either by volcanic construction or by degradation. The degree to which eruptive products accumulate asymmetrically will be much higher if distal flank vents are important and this may lead to greater overall stratigraphic and petrologic uncertainty. Thus, regardless of whether or not a significant fraction of an edifice has been removed, some or even most of the center’s eruptive products will be absent from any given exposed section. The distributions of recent lavas erupted from active volcanoes, which are commonly highly divergent in flow direction during short periods of time, suggest that typically 101–103 eruptions might be unrepresented between successive flows exposed in a particular vertical section depending on eruption frequency and vent position (e.g. Albarède et al., 1997, p. 174). Edifice degradation may amplify the tendency for biased local records by selectively removing: (1) stratigraphic units emplaced immediately before eruptive hiatuses, (2) deposits on the pole-facing flank, where glaciation is the main erosion mechanism, and (3) unconsolidated pyroclastic deposits, particularly those of minor volume (Hackett & Houghton, 1989). Vent-proximal units, such as silicic domes, may be susceptible to short preservation times because of a tendency for explosive auto-destruction and the consequences of episodic summit crater formation. In light of these hindrances, how may petrologists devise strategies for the study of prehistoric arc volcanoes that provide a representative record of a volcano’s magmatic evolution in combination with the relative and absolute temporal constraints required for meaningful modeling of this evolution?
The Tatara–San Pedro Project (Dungan et al., 1992) was undertaken, in part, to respond to this question. In the course of mapping and characterizing the well-exposed Tatara–San Pedro complex (TSPC: 930 ka to late Holocene; Fig. 1), the scientific team sampled 30 flow-by-flow vertical sections in eight deep glacial valleys around the complex (Fig. 2; Table 1). The chemistry and petrography of samples from 23 of these sections have been used as the basis for correlations, in conjunction with supporting paleomagnetic, photogrammetric, and geochronologic data, to determine the relative eruptive orders of flows in sequences of the Tatara–San Pedro complex that are amenable to such reconstructions, and for which we have adequate stratigraphic control. Volcán Tatara (100–60 ka), the upper Placeta San Pedro sequence (234 ka), and an older edifice remnant represented by the Estero Molino sequence (620–495 ka) are discussed in some detail. Three large edifices that are not treated here are the Holocene Volcán San Pedro, Volcán Pellado, and the lavas of Cordón El Guadal (Feeley & Dungan, 1996; Feeley et al., 1998).
|
, 
, •) are repeated in Fig.
|
|
|
|
|
|
|
The strengths, weaknesses, and benefits of this approach are evaluated with respect to the general problem of reconstructing eruptive histories and refining temporal resolution within lava sequences at Quaternary arc volcanic centers. These results offer encouragement concerning the stratigraphic and chronological control that may be achieved for such edifices, but they also sound a loud cautionary note with respect to the pitfalls of petrologic modeling on the basis of poorly constrained stratigraphic relations.
|
PREVIOUS AND CURRENT WORK |
|---|
The Tatara–San Pedro complex (TSPC; 36°S, Southern Volcanic Zone, Chilean Andes) is a large arc-front Quaternary volcanic center that was first identified by González & Vergara (1962)
. Subsequent reconnaissance mapping during the first stage of this study (1984–1986), and regional studies by Muñoz & Niemeyer (1984), led to the recognition of multiple edifices of variable age. The application of field mapping, photogrammetry, K–Ar and 40Ar/39Ar dating, and paleomagnetic data to the elucidation of the broad evolution of the TSPC has allowed us to define its overall eruptive and erosional history, to demonstrate that vent positions migrated though time, and to identify multiple unconformity-bound volcanic sequences (Singer et al., 1997). Volcanic sequences are the first-order stratigraphic unit referred to throughout this paper. Each sequence is inferred to represent a separate volcanic edifice. Consecutive sequences are usually separated in time by lengthy lacunae, indicating that a continuous record has not been preserved. Sequences older than 150–200 ka are erosional remnants of volcanic constructs that have been greatly reduced in volume, mainly by glaciation. The volcanic stratigraphic nomenclature employed in this paper generally follows that of Singer et al. (1997), but a few modifications are proposed in this paper. A modified geologic map of the TSPC, with a stratigraphic column and age data, is included in Electronic Appendix I (available on the Journal of Petrology Web site at http://www.petrology.oupjournals.org) as Fig. I-1.
Petrologic studies up to now provide evidence for a range of mafic parent magmas and multiple differentiation processes (Davidson et al., 1987, 1988; Ferguson et al., 1992; Singer et al., 1995, 1997; Feeley & Dungan, 1996; Feeley et al., 1998). Sampling for paleomagnetic studies has been conducted at a total of 243 sites within the TSPC, of which 30 within the Quebrada Turbia sequence record the Bruhnes–Matuyama magnetic polarity transition (Brown et al., 1994; Singer & Pringle, 1996).
|
GEOLOGIC SETTING AND PETROLOGIC CONTEXT |
|---|
Previous investigations of individual Andean Southern Volcanic Zone (SVZ) volcanoes (e.g. Calbuco, Hickey-Vargas et al., 1995
; López-Escobar et al., 1995; Villarrica–Lanin, Hickey-Vargas et al., 1989; Puyehue–Cordon Caulle, Gerlach et al., 1988; Mocho–Choshuenco, McMillan et al., 1989; Sollipulli, Murphy & Brewer, 1994; Gilbert et al., 1996; Lonquimay, Moreno & Gardeweg, 1989; Antuco, López-Escobar et al., 1981; Nevado de Longaví, Gardeweg, 1981; Laguna del Maule–Puelche, Frey et al., 1984; Hildreth et al., 1999; Calabozos caldera, Hildreth et al., 1984; Grunder, 1987; Grunder & Mahood, 1988; Quizapu, Hildreth & Drake, 1992; Planchon–Peteroa–Azufre, Tormey et al., 1995) combined with regional compositional profiles on the basis of multi-center geochemical traverses along and across the arc (e.g. López-Escobar et al., 1977, 1992, 1993; Déruelle, 1982; Harmon & Hoefs, 1984; Hickey et al., 1984, 1986; López-Escobar, 1984; Stern et al., 1984, 1991; Notsu et al., 1986; Muñoz & Stern, 1987, 1989; Futa & Stern, 1988; Hildreth & Moorbath, 1988, 1991; Stern, 1988, 1989, 1991a, 1991b; Davidson, 1991; Tormey et al., 1991) provide the petrologic context of this study. This body of literature addresses issues central to an understanding of continental arc volcanism: (1) What are the contributions of asthenospheric and subcontinental lithospheric mantle sources to the petrogenesis of the parental basaltic magmas? (2) Do variations in the age of the subducted plate, thickness of arc-trench sediment, and/or vigor of subduction erosion play roles in along-arc variability? (3) How do differentiation mechanisms and crustal contributions to the petrogenesis of evolved magmas vary as functions of crustal thickness, density, structure, age, and lithologic character along the arc?
The purpose of this paper is not to respond directly to the questions listed above, but to refine the geologic framework of the TSPC so that ultimately the regional trends may be assessed in light of a process-oriented evaluation of the contributions of diverse mantle and crustal sources at a single long-lived center. In this context, the setting of the TSPC is briefly reviewed. A modified version of fig. 1 of Hildreth & Moorbath (1988) is reproduced as Fig. 1 of this paper to illustrate the arc segmentation scheme of Wood & Nelson (1988). The narrow northern chain (33–34·2°S; Tupungato–Maipo segment) is located entirely on exceptionally thick and relatively old crust along the Andean crest. To the south (34·7–36°S; Palomo–Tatara segment), the arc broadens and a N20°E chain of frontal arc centers lies to the west of the continental divide. The southernmost chain generally lies near the topographic front of the Cordillera, which becomes increasingly less well defined to the south, and it extends from Nevado de Longaví to Osorno and Calbuco (not shown in Fig. 1), or beyond (36·2–42°S; Longaví–Osorno segment).
The TSPC is the southernmost volcanic center of an arc segment that is intermediate in terms of geography, Bouguer gravity (hence apparent crustal thickness; see Hildreth & Moorbath, 1988), and some geochemical characteristics between the highly contrasting Tupungato–Maipo and Longaví–Osorno segments. Although this part of the arc could be considered petrologically transitional (Tormey et al., 1991), Wood & Nelson (1988) argued that the volcanoes of each segment correspond to separate petrologic populations, and this point of view is reinforced by new data from the TSPC. One example of the importance of segmentation of the SVZ lies in the observation that many evolved TSPC magmas more closely resemble those of volcanoes to the north, which were constructed on thicker crust, than those to the south, which lie on crust that does not appear to be substantially thinner than the crust at 36°S.
Pre-Pliocene rocks exposed beneath the frontal arc between 37 and 35°S (referred to as basement) are late Mesozoic–Tertiary sedimentary and volcanic rocks and Tertiary granitoids. The TSPC overlies a deformed Tertiary sequence of dominantly volcanic and volcaniclastic rocks that are intruded by two granitic plutons (Huemul and Cerro Risco Bayo), both of which have been dated at 6·5 Ma by 40Ar/39Ar (Nelson et al., 1999). Apart from a suite of partially melted granitic xenoliths present in a pyroclastic deposit on Volcán Pellado, the vast majority of xenoliths in TSPC lavas are mafic plutonic rocks (troctolite, gabbro, and norite; Costa, 2000).
|
METHOD OF STRATIGRAPHIC RECONSTRUCTIONS |
|---|
Physiography and sampling
Along the northwest margin of Placeta San Pedro, a large plateau on the northwest flank of the TSPC (Figs 2–5), lavas of the pre-Volcán Tatara sequences, which flowed mainly from east to west, banked against a north–south basement ridge (Cordón Las Yeguas; CLY in Fig. 2). The confinement of these units within a perched basin contributed to partial preservation during multiple episodes of glacial valley incision. None the less, the overlying north flank of Volcán Tatara has been largely stripped away by ice flowing to the west and north (Figs 2 and 5), exposing underlying units. Pre-Volcán Tatara sequences are absent on the south flank of the complex, with the exception of Cordón El Guadal lavas. The southwest flank of Volcán Tatara thins to the west against Cordón Los Ñirales (CLÑ in Fig. 2), the southward continuation of Cordón Las Yeguas (CLY in Fig. 2). Important consequences of these paleotopographic accidents are that thick sections of pre-Volcán Tatara sequences on the northwest flank are dissected by Quebrada Turbia (QT) and other drainages incised into Placeta San Pedro (PSP), and that much of the history of Volcán Tatara is exposed in Estero San Pedro (ESP), Estero Molino (EM) and Quebrada Turbia.
|
Many of the early volcanic sequences of the TSPC (>200 ka), plus the northern flank of Volcán Tatara (100–60 ka), crop out on and around the margins of Placeta San Pedro. The Muñoz, Quebrada Turbia, Estero Molino, and lower and upper Placeta San Pedro sequences have been sampled in 12 flow-by-flow sections located in four valleys (see Fig. 2 and Table 1 for section locations and acronyms employed in the text), comprising 250 analyzed samples. Quebrada Turbia (south to north; QTW12, QTW14, QTW10, QTW11), which heads just below the eroded vent region of Volcán Tatara, marks the eastern limit of Placeta San Pedro. From south to north, the east to west drainages of Estero Molino (EML) and Quebrada Castillo (QCSE, QCNE, LV, QC98), and the south to north canyon of Estero San Pedro del Norte (ESPN) are incised into its western margin. Additional samples from these early sequences (CLL, QTE, and assorted localities) were collected from Cordón Los Lunes, the ridge separating Quebrada Turbia and Cajón Muñoz (CM).
The northwest flank of Volcán Tatara overlies the Estero Molino and Placeta San Pedro sequences on southern Placeta San Pedro (Figs 3–5), but is in contact with basement between Estero Molino and Rio de la Puente (Fig. 6). The west flank of Volcán Tatara is divided into northern and southern domains by the pyroclastic and intrusive near-vent facies exposed on eastern Cordón Tatara, in the divide between Estero San Pedro and Quebrada Turbia (VF; Fig. 3), across which stratigraphic units cannot be readily traced physically; the north slope of Cordón Tatara is covered almost entirely by talus (Figs 3–4) and the south face is an inaccessible cliff (Fig. 6). Approximately 280 Volcán Tatara samples have been analyzed from 13 sections, including: (1) to the NNW in the upper walls of Quebrada Turbia (QTW12), Estero Molino (EMU1–3), in an adjacent valley (BP), and in the largely moraine- and snow-covered region between Estero Molino and Quebrada Turbia (EMU4); (2) to the SSW in six sections within Estero San Pedro (ESPW2–ESPW3–ESPW4, ESPE1–ESPE2, and TR); (3) on the east wall in upper (UEP7) and lower (EPW5, EPW6) Estero Pellado. Proximal and distal sections in Estero San Pedro (ESPW3 and ESPW4) correspond to sections 1 and 2 of Ferguson et al. (1992, fig. 4).
Lava samples were collected almost exclusively on a flow-by-flow basis from continuous sections exposed on steep canyon walls. The vast majority of samples are from very fresh non-vesicular flow interiors lacking any macroscopic or microscopic evidence of alteration such as secondary phases in vesicles or veins, groundmass oxidation, or even iddingsitized olivine. Variably but mildly altered samples that constitute exceptions to these rules were collected from the lower 10 lavas of the QTW12 and ESPE1 sections, where groundwater flow has been concentrated at the interface between permeable lavas and relatively impermeable basement. Although neither the internal stratigraphic relations nor the petrology–geochemistry of Cordón El Guadal (Feeley & Dungan, 1996; Singer et al., 1997; Feeley et al., 1998), Volcán Pellado (immediately preceded Volcán Tatara), and Volcán San Pedro (Holocene) are discussed in detail, their compositions are shown in Figs 7 and 8 for comparative purposes, as these edifices are volumetrically and petrologically significant.
|
|
|
|
Chemical stratigraphy: whole-rock X-ray fluorescence analyses
Analytical data
The geochemical data presented here were obtained at the University of Massachusetts following the methods of Rhodes (1988). Information concerning the precision of major and trace element analyses in this laboratory can be found in table 2 of Rhodes (1996). Major elements plus Nb, Zr, Sr, Rb, Ba, Y, Ni, Cr, V, Ce, and Pb concentrations are reported in Table 2 for selected samples and for average chemical compositions of flow packages [point (1) below]. Total iron is reported as Fe2O3*. Samples that have been analyzed for Cs, Sc, Cr, Co, La, Ce, Nd, Sm, Eu, Tb, Yb, Lu, Hf, Ta, and Th by instrumental neutron activation analysis (INAA) at MIT (F. A. Frey & M. A. Dungan, unpublished data, 2000) are marked by an asterisk in Table 2. Specific references to analyses in Table 2 throughout the text refer to the index number at the head of each column. Whole-rock analyses, including INAA data, are provided in spreadsheet format in Electronic Appendix I. The entire TSPC dataset is arranged by sampled section in TSPC.xls, whereas reconstructed sequences are presented in stratigraphic order in QTS.xls, EMS.xls, UPSPS.xls, LTAT.xls, and UTAT.xls. The TSPC comprises mainly lavas ranging from medium-K basalt and basaltic andesite to high-K dacite and rhyolite. The following ranges of wt % SiO2 are used to define rock names: basalt <52, basaltic andesite 52–56, andesite 56–63, dacite 63–70, rhyolite >70, high-SiO2 rhyolite 72–76.
Reconstruction criteria
Internal reconstructions of multiple volcanic sequences of the TSPC have been rendered feasible by the presence of widespread chemically distinctive packages of lavas [point (1) below] that are correlative between multiple sampled sections on the basis of chemistry and petrography. Although some packages were identified as physically separate but correlative units in the field, the approach of tying flow-by-flow sampling traverses to images of canyon walls and establishing correlations among them is more efficient and potentially more certain than field-mapping of such units, which are commonly only subtly different in hand specimen, along canyon walls and from valley to valley. Sample traverse selection is a key element in the success of this approach. Our experience favors acquisition and photogrammetric analysis of canyon-wall images before sampling, to maximize the utility of each traverse. For example, subtle or even marked unconformities between sequences and packages that are evident from a distance are rarely so obvious when one is working on a steep rock face. We have found that representing stratigraphic subdivisions at the level of packages on a 1:25,000 geologic map would be virtually impossible without first compiling the integrated information from stereo images of canyon walls.
The reconstruction technique has been formulated on the basis of the following reasoning, criteria, and limitations:
(1) All sampled sections include groups of consecutive lava flows with closely similar [see points (3) and (4)] or virtually identical whole-rock chemistry and petrography that are unquestionably distinct from underlying and overlying flows. Where such groups of cogenetic flows may be correlated among multiple sections on the basis of their chemistry, or where a single distinctive flow unit is demonstrably laterally extensive, these lavas have been assembled as informal stratigraphic units herein referred to as flow packages (generally <1 km3), which are taken to represent discrete eruptive events. By analogy with historic observations of stratocone activity, eruptive events are inferred to represent continuous or intermittent effusion of magma from a reservoir–conduit system over periods ranging from weeks to decades. In cases where two adjacent groups of flows have similar chemical signatures, but they are distinguished by a small, abrupt shift in composition, separate packages also have been defined and their mutual affinities are noted. Where consecutive packages ± non-correlative flows [point (2)] are apparently closely related petrologically (e.g. to a particular parent magma type), the ensemble is tentatively referred to as the record of an eruptive episode, implying successive eruptive events during a relatively restricted but undefined period; such inferences are subject to further evaluation.
(2) Conversely, some magma compositions were sampled at only one locality. As all sampled sections are incomplete in the sense that a significant proportion of the eruptive products of the relevant center(s) are not present, the stratigraphic positions of these unique flows usually are imperfectly defined with respect to certain other flows or flow packages in other sections, which are, none the less, constrained to be close in time. These non-correlative flows (NC in Table 2) are discussed, as they are equally important with respect to magmatic evolution, but for the purpose of establishing a consistent and meaningful nomenclature they have been accorded package status only in special circumstances. There is potential for circular reasoning in generating stratigraphic reconstructions on the basis of chemistry, in the sense that there is a logical but subjective tendency to ‘resolve’ uncertainties by placing apparently cogenetic flows in stratigraphic proximity where a lack of constraints permits multiple interpretations. Despite the high density of sampling in this study, and abundant supporting geochronologic and paleomagnetic data, such dilemmas do arise [point (5)], and they are noted accordingly.
(3) Natural variability within the same lava flow or package [e.g. arising from zoned or heterogeneous magma reservoirs, eruption associated with incomplete mixing of two or more magmas, heterogeneous distribution of phenocrysts or fragments of cognate cumulate at the centimeter scale, or dispersed fragments of crustal xenoliths—point (4)] frequently renders reported lava compositions from single packages non-identical (i.e. outside analytical uncertainties). Decisions about intra-section flow groupings, and their correlative status among sampled sections, have been made on the basis of an empirical inspection of chemical data and thin sections, rather than on statistical criteria, because small variations do not negate fundamental petrogenetic relations among multiple lavas generated during the same eruptive event when the data are viewed from the perspective of co-variations among multiple elements and elemental ratios complemented by petrographic observations.
(4) Diverse trends with respect to eruptive order are recognized within flow packages comprising basaltic andesitic to andesitic lavas. In some cases, consecutive flows display no significant chemical variations, or no systematic variations as a function of stratigraphic position. More commonly, they are marked by progressive changes such that the early and late flows are significantly different; flow-to-flow changes are typically on the order of analytical uncertainties, without major internal discontinuities in either major or trace elements. Up-section trends toward more evolved and less evolved magmas are observed, and reversals of these internal trends are occasionally present. In all cases where internally variable packages have been identified, intra-package chemical variations are of a lesser magnitude than those that distinguish such a package from underlying and overlying packages or non-correlative flows.
(5) Eruptive episodes are rarely characterized by progressive trends toward greater degrees of evolution from a single parent magma, implying that the petrologic differences between many adjacent flow packages reflect discrete parent magma batches and unique evolution paths. If this inference is correct, duplication of the major and trace element characteristics of one magma batch by another that differs significantly in time should be rare, because multiple variables are involved. There are, none the less, examples of compositionally similar lavas that are unequivocally separated by intervening units; these can normally be distinguished because the concentrations of one or a few elements (and elemental ratios) are different between them. There is, in addition, one example of lavas from two different volcanic sequences that are nearly indistinguishable for major and trace elements. These occurrences may be explained as fortuitous repetitions of parent magma compositions and evolutionary histories, but they are impossible to rationalize as the persistence of a mafic magma body that remained unchanged during >105 years. This observation emphasizes the care required in utilizing chemical criteria in the reconstruction of eruptive histories in the absence of independent constraints derived from physical stratigraphy, petrography, paleomagnetic data, and/or geochronology; that is, chemical similarity, or even virtual identity, is not a guarantee of stratigraphic equivalence.
Revised 40Ar/39Ar chronology
A major emphasis of this paper is applying the reconstruction method outlined above to the Estero Molino sequence and to Volcán Tatara: both are thick, widespread and complex. New 40Ar/39Ar ages have been acquired from these sequences to clarify stratigraphic relations where uncertainties or conflicts existed as a result of K–Ar determinations characterized by (1) low radiogenic argon yields (<5%) and consequently low precision, (2) discordant groundmass–whole-rock ages (groundmass younger), and/or (3) out-of-sequence ages. It is probable that argon loss and xenocrystic argon contamination both affected the K–Ar measurements. Argon loss from glass-bearing groundmass, leading to disturbed Ar-release spectra and anomalously young ages, was encountered during dating of some groundmass separates. These results will be presented in detail in another paper, but the revised conclusions concerning the ages of the dated sequences are summarized here (Fig. 9).
|
The Estero Molino sequence is subdivided on the basis of chemical and 40Ar/39Ar age constraints into lower (
620–600 ± 20 ka), middle (579·3 ± 10·5 ka), and upper (493·7 ± 10·8 ka) lavas. Thus, the upper age limit of the Estero Molino sequence has been revised from <320 ka to 495 ka, and the age of the intra-Estero Molino sequence lacuna (middle–upper EMS) is now constrained to be much shorter than previously estimated. K–Ar ages previously determined for Volcán Tatara lavas range from 90 ± 19 ka to 19 ± 13 ka (Singer et al., 1997). New 40Ar/39Ar age determinations for much of lower and upper Tatara yield concordant plateau and inverse isochron ages in the range of 100–60 ka with variable uncertainties (typically ±10–20 ky on inverse isochron ages), although a few samples from basal lower Tatara units have yielded more complicated argon-release spectra that tentatively suggest ages as old as 100–130 ka. Thus, Volcán Tatara is older than was inferred from K–Ar measurements.
Petrologic and geochemical overview
Detailed petrologic studies with the objectives of identifying and quantifying processes operative during different stages in the complex’s evolution (e.g. Singer et al., 1995; Feeley & Dungan, 1996; Feeley et al., 1998), are currently in progress on the basis of the chemical data presented here combined with additional trace element and isotopic analyses plus petrographic and mineral chemical data. Delving deeply into petrologic interpretations of comprehensively characterized samples is not the main goal of this paper, but preliminary inferences derived from graphical treatments of whole-rock chemistry, which require additional testing, are discussed to (1) provide context for the temporal trends revealed by stratigraphic reconstructions and (2) illustrate the implications of these temporal trends for modeling investigations. These preliminary results are offered largely as working hypotheses with the goal of demonstrating the impact of high-density, stratigraphically controlled sampling on the identification of (1) multiple components implicated in the diversity of broadly cogenetic magmas, and (2) the processes that were most probably involved in the immediately pre-eruptive evolution of these magmas.
On the basis of previous work it has been suggested that the diverse lavas of the TSPC manifest evidence for a range of basaltic parent magma compositions and for multiple differentiation mechanisms. Eruptive products include the following: (1) heterogeneous mingled andesitic lavas generated by mafic–silicic magma interaction, such as those of Cordón El Guadal (Feeley & Dungan, 1996), the lower Placeta San Pedro sequence (probably correlative with lavas of Cordón El Guadal), and a large late Holocene eruption of Volcán San Pedro (Singer et al., 1995); (2) basaltic andesitic and andesitic lavas with phenocryst assemblages approaching equilibrium that were derived primarily by fractional crystallization from mafic parent magmas (Ferguson et al., 1992); (3) blended hybrids characterized by disequilibrium phenocryst assemblages; (4) intermediate magma produced when mafic parent magmas became contaminated by interaction with the crust concurrently with fractional crystallization (Davidson et al., 1987, 1988).
The following section (Figs 7 and 8) is a general comparison of lava compositions at the level of sequences, which is designed to introduce the following observations: (1) each sequence is distinct from all others; (2) there is no discernible temporal progression defined by lava compositions or differentiation mechanisms at the level of sequences; (3) most of the mafic and intermediate composition lavas of the TSPC display evidence for some type of open-system evolution; (4) the TSPC is substantially different, in terms of differentiation mechanisms, from most volcanoes of the Longaví–Osorno segment of the SVZ. For comparative purposes, data from the well-studied basalt to rhyolite suite of Volcán Puyehue–Cordón Caulle (Gerlach et al., 1988) are shown in several plots, as are fractional crystallization trends derived from these data. Features such as nearly constant ratios of some incompatible elements from basalt to highly differentiated rhyolite and extremely high FeO*/MgO in dacites and rhyolites led Gerlach et al. (1988) to conclude that nearly closed-system fractional crystallization was responsible for the spectrum from basalt to rhyolite, although some andesitic magmas formed by mixing of mafic and silicic end-members. Although we do not argue that closed- vs open-system differentiation is the only distinction between TSPC magmas and those at volcanoes to the south, we suggest that the Puyehue–Cordón Caulle lavas represent a limiting case with respect to the SVZ that serve as a useful reference.
Compatible elements: closed vs open systems
The compatible elements Mg, Ni, and Cr are higher in intermediate-composition hybrid magmas (i.e. at a given SiO2; Fig. 7a–c) than in magmas generated by closed-system fractional crystallization under similar P and fO2 conditions. In detail, the graphical trajectories of mafic to silicic differentiation trends defined by these elements may vary in accord with the following: (1) the fractionating mineralogy, which is a function of several factors (e.g. Grove & Kinzler, 1986); (2) the compositions of magmatic end-members in the case of magma mixing; (3) end-member compositions plus the ratio Ma/Mc (DePaolo, 1981) in the case of assimilation–fractional crystallization (AFC). The trends defined by different sequences of the TSPC (Fig. 7a–c) are compared with a reference basalt–rhyolite mixing curve, and with an inferred fractionation-dominated trend that is based empirically on data from Volcán Puyehue–Cordón Caulle (Gerlach et al., 1988). Hybrid andesites from Puyehue fall well to the right of the FC curve.
Mg-numbers of TSPC mafic andesitic magmas with 55–56·5 wt % SiO2 vary between 61 and 40 (5·3–2·6 wt % MgO) and Cr varies sympathetically by more than an order of magnitude (120–5 ppm): such divergent trends require multiple differentiation mechanisms. Low-mg-number basaltic andesitic to andesitic magmas are generally characterized by a closer approach to textural equilibrium (plag + pyx ± oliv), and are provisionally interpreted as the products of fractionation-dominated differentiation. High-mg-number andesitic magmas are texturally variable. Many have disequilibrium textures, multiple populations of plagioclase phenocrysts, and olivine plus two pyroxenes, whereas others contain olivine phenocrysts without pyroxene. These distinctions may reflect mixing among diverse end-members, evolution by AFC, or polybaric evolution involving multiple processes, which can only be addressed by geochemical modeling integrated with petrographic and mineral chemistry constraints.
LILE variations and crustal contributions
Rare primitive, uncontaminated basalts of the Southern Volcanic Zone and other arcs have high K/Rb (1000–750; Fig. 7d–f). The high K/Rb, Sr/Rb, Ba/Rb, and Ba/Th of Rb- and Th-poor basaltic magmas can be lowered by 25–50% relative to primitive values by only 2–10% contamination involving highly Rb- and Th-rich crustal components. Consequently, suites of mafic magmas that are not necessarily heavily contaminated in volumetric terms may be characterized by dramatic decreases in these ratios vs SiO2 or Rb. In contrast, these ratios are not changed drastically by closed-system fractional crystallization of an anhydrous mineral assemblage, particularly in the mafic range (e.g. Davidson et al., 1987, 1988). K/Rb in rhyolitic and dacitic magmas of the TSPC is 280–250 (lower than in comparable magmas from Puyehue; Fig. 7e), and some basement granitoids have still lower ratios, and in part, very high Rb contents (insets in Figs 7d and 8a). Thus, assimilation and mixing components with the potential for lowering K/Rb and other such ratios in mafic magmas were available. As the vast majority of mafic to intermediate TSPC magmas have K/Rb <450 (many <350), open-system behavior during some stage of their evolution is required. As SVZ basaltic magmas rarely have K/Rb in the primitive range, modification of primitive magmas during ascent by interaction with the crust appears to have been nearly ubiquitous (Hildreth & Moorbath, 1988), consistent with thermal–chemical–mechanical modeling (e.g. Huppert & Sparks, 1985; Reiners et al., 1995; Edwards & Russell, 1998) and diverse observations from various tectonic settings (e.g. McBirney et al., 1987; Philpotts & Asher, 1993; Kerr et al., 1995; Luhr et al., 1995; Davidson, 1996).
Relative enrichment of Y and HREE
Among the important distinctions between the northern and southern SVZ is the behavior of Y and the heavy rare earth elements (HREE) during progressive evolution from basalt to rhyolite. Some high-Y dacitic and rhyolitic lavas of the southern SVZ (Longaví–Osorno segment, Fig. 1) that are characterized by unusually high FeO*/MgO, iron-rich anhydrous ferromagnesian silicate mineralogy, and strongly decreasing Sr concentrations with increasing differentiation (e.g. Volcán Puyehue–Cordon Caulle; Gerlach et al., 1988) also have nearly the same La/Yb, Th/Yb, Ba/Y (Fig. 8c, inset), and Rb/Y (Fig. 8a–c) as associated mafic lavas. These observations are consistent with production of these intermediate to silicic magmas largely by closed-system fractional crystallization from mafic parent magmas in the upper crust as proposed by Gerlach et al. (1988). In contrast, Hildreth & Moorbath (1988), and Tormey et al. (1991) postulated suppression of Y–HREE enrichments in evolved rocks of the northern SVZ relative to closed-system fractionation trends (particularly the Tupungato–Maipo segment, Fig. 1) as a consequence of higher pressure differentiation in open systems characterized by incorporation of crustal components generated by partial melting wherein garnet remained in the source. Hildreth & Moorbath (1988) interpreted a correlation between the suppression of Y–HREE enrichments (high Ce/Yb) and inferred crustal thickness as an indication that assimilation occurs at increasingly greater depths in increasingly thicker crust (i.e. primarily near the base of the crust). As silicic magmas from 36°S northward commonly contain hornblende and those south of 36°S do not, differentiation of wet parent magmas (i.e. high cpx/plag and hornblende/plag) may have contributed to increases of light REE (LREE) over Y–HREE at the TSPC and at other centers in the northern SVZ, but this mechanism cannot explain the much higher concentrations of LILE in intermediate to silicic magmas of the northern SVZ.
In contrast to the strong Y-enrichment trends observed at Puyehue (inset, Fig. 8d), evolved magmas of the TSPC display multiple, divergent trends ranging from no Y enrichment with increasing SiO2 (i.e. the Los Lunes Rhyolite, Muñoz sequence, has lower Y and Yb than most basalts) to enrichment trends that are less pronounced than the most strongly enriched magmas of the Longaví–Osorno segment of the southern SVZ (inset, Fig. 8f). The tendency for increasing Ba/Y and Rb/Y with increasing SiO2 in some sequences is broadly interpreted as a signature of open-system behavior (i.e. involvement of high-Ba/Y and high-Rb/Y crustal end-members) at the TSPC. Low-mg-number andesitic magmas (mainly Volcán Tatara) display the greatest Y–HREE enrichments in combination with low Rb/Y and Ba/Y, thereby approaching the Puyehue trend. This diversity is impossible to interpret in terms of a single differentiation mechanism, or as a function of one crustal variable. A range of differentiation trends linked to multiple depths, assimilated crustal components, and processes is implied.
Limited isotopic variations
The tendency of arc magmas of the southern SVZ to display minor variations in isotopic ratios from basalt to rhyolite, caused in large part by low crust–magma isotopic contrast, is well established (Hickey et al., 1986; Davidson et al., 1987, 1988; Gerlach et al., 1988; Tormey et al., 1995). This has led to the tradition, which is followed here, of developing arguments that depend on trace element ratios as the basis for discussions of open-system vs closed-system magma evolution. None the less, an incomplete isotopic survey (87Sr/86Sr) of mafic to silicic magmas of the TSPC (J. P. Davidson & M. A. Dungan, unpublished data, 2000) has thus far defined a range of low ratios for primitive magmas (0·70379–0·70398) and generally higher ratios in mafic to silicic magmas that may have crustal contributions. In particular, the voluminous Muñoz dacite (0·704168) and Los Lunes rhyolite (0·704249) of the early Muñoz sequence, and rhyolite ESPE3.3 of Volcán Tatara (0·704825), have higher isotopic ratios than any basaltic values. As the Muñoz sequence silicic magmas and ESPE.3 have extreme trace element signatures that render unlikely a derivation by closed-system fractional crystallization from basaltic magmas, there is a strong prima facie case that at least some silicic magmas were generated by partial melting of the crust (possibly including late Tertiary, or even Quaternary, granitoids). As these magma compositions serve rather well as contaminants for many intermediate composition magmas, it is likely that such crustally derived melts were incorporated into hybrids. The mechanisms and locations of melting and assimilation, and the specific compositions of assimilated or mixed components at various stages of the evolution of diverse TSPC magmas, are the topics of subsequent work.
Diversity among parent magmas
Some of the most primitive and least contaminated basaltic magmas thus far described from a Quaternary Andean volcano have been recognized at the TSPC. Although this paper does not treat questions related to mantle source heterogeneity or basalt generation, we note that basaltic magma compositions at the TSPC display variable trace and major element signatures and we refer to temporal changes in parent magma character as a source of magmatic diversity. Although elemental co-variations are not perfectly systematic, basaltic magmas with low TiO2 (<0·9 wt %) and P2O5 (<0·15 wt %) are generally characterized by: (1) low LREE and La/Yb; (2) low Nb and Zr, high Ba/Nb (>100), and low Zr/Y (3·5–5) and Nb/Y (0·10–0.15); plus (3) high Al2O3 (>18 wt %) and Sr (>650 ppm) in combination with high Sr/Y (>45) and Sr/Rb (>70), and low Ba/Sr (0·3). In addition, there is a range of basaltic magmas with higher TiO2 (0·9–1·1 wt %) and P2O5 (0·2–0·3 wt %) that have higher Nb/Y and Zr/Y, but lower Ba/Nb and K/P. The most robust ratio with respect to parent magma discrimination is Nb/Y, which varies from 0·10 to 0·45 among basaltic magmas, and which displays little variation during closed-system evolution and only minor variations among lavas that appear, on the basis of criteria noted above, to be related by open-system processes.
In subsequent treatments of individual sequences, emphasis is placed on comparisons of lavas within sequences on the basis of the major and trace element variation diagrams that most clearly illustrate, in each specific case, distinctions among parent magmas and associated differentiation trends. Data from some sequences are also presented as a function of relative eruptive order as documentation of the general observation that progressive evolution from a single parental magma batch is generally not the cause of compositional changes from one eruptive event to the next, even where quasi-linear arrays on certain variation diagrams suggest a high degree of commonality among closely related, variably evolved magmas.
|
PRE-ESTERO MOLINO SEQUENCES |
|---|
Muñoz sequence (
930–825 ka)The Muñoz sequence (Figs 7 and 8) is a compositionally bi-modal suite that is dominated volumetrically by two major silicic units, the basal Muñoz Dacite (plag + hbl + biot + ox; Table 2a, index numbers 1 and 2) and the younger Los Lunes Rhyolite (plag + biot + ox + zirc; Table 2a, 8 and 9). These units are separated in time by
100 ky, and by oxygen isotope stage 22 (860 ka), and both comprise multiple flow units with slightly variable compositions. Intercalated between these large silicic units are the volumetrically minor mafic to intermediate Sin Nombre lavas and laharic breccias (reversed magnetic polarity; figs 3–4 of Singer et al., 1997). Mafic basaltic andesitic lavas exposed between the Muñoz Dacite and Los Lunes Rhyolite on the east wall of Quebrada Turbia (QTE) are diverse (Table 2a, 3–6), including (1) a relatively MgO-rich basaltic andesitic lava (52·3 wt % SiO2; mg-number 59), and (2) three basaltic andesitic lavas (52·2–54·5 wt % SiO2) with variable trace element ratios and much lower mg-numbers (46–42). Relatively mafic compositions with such a strong fractional crystallization imprint are unusual elsewhere in the early preserved remnants of the TSPC. Comparably low mg-numbers at the same SiO2 do not occur again until lower Volcán Tatara (e.g. packages 
and 
; Table 2h and i, 155–163).
|
|
In contrast, a single basaltic andesitic lava flow on the west wall of Quebrada Turbia with reversed magnetic polarity (QTW10.1; 55·4 wt % SiO2; Table 2a, 7), underlying a horizon defined by reworked blocks of Los Lunes Rhyolite, possesses an extreme chemical signature characterized by exceptionally to moderately low TiO2, Na2O, P2O5, Fe2O3*, Y, Nb, and Zr, plus moderately high K2O, Ni, Cr, and MgO, in comparison with other TSPC lavas with 55–56 wt % SiO2 (mg-number 61; K/Ti 3·6; K/P 26; Na2O/K2O 2). Such a magma cannot be derived by fractional crystallization of any basaltic composition thus far analyzed from the Southern Volcanic Zone, but would be consistent with mixing between basalt and a silicic magma (in proportions 5:1) similar to the Muñoz Dacite or Los Lunes Rhyolite; e.g. the P2O5 and TiO2 contents of this hybrid are so low, and the K/P is so high, that mixing of basalt with a silicic magma extemely depleted in TiO2 and P2O5 is the only viable model. This is the only lava that we have identified from the TSPC that can be modeled simply as mixture of basalt and a highly silicic magma, but it serves as a limiting reference composition. Apparently, all other open-system scenarios at the TSPC involved less primitive and/or less silicic components.
Quebrada Turbia sequence (QTS: 785–771 ka)
The paleomagnetic polarities and ages of QTS lavas are presented elsewhere (Brown et al., 1994; Singer & Pringle, 1996; Singer et al., 1997); most of these flows are characterized by transitional virtual geomagnetic pole positions as a result of eruption during the Matuyama–Bruhnes polarity reversal. This unit forms thick exposures within lower Quebrada Turbia (Figs 3 and 9), but is not present elsewhere. The erosion surface underlying the QTS probably formed during glaciation that culminated with oxygen isotope stage 20 (785 ka), and the upper erosion surface probably represents cumulative erosion associated with peak ice advances corresponding to oxygen isotope stages 18 and 16 (710 and 620 ka).
The QTS (33 samples; Figs 10 and 11) is divided informally into (1) a thin, heterogeneous lower group of five basaltic andesitic to andesitic lavas (QTW10.2–6; 52·7–58·5 wt % SiO2; Table 2a, 10–12) preserved in a paleodepression, and (2) a more extensive, locally thick upper package of basaltic andesitic lavas (QTW10.7–11 plus QTW11.1–23; 54·7–55·7 wt % SiO2; Table 2a, 13–19, 21) wherein early-erupted magmas are slightly less evolved (4·1–4·4 wt % MgO; Table 2a, 13–15) than late-erupted compositions (4·45–4·8 wt % MgO; Table 2a, 16 and 17). The three stratigraphically highest preserved flows of this sequence (QTW11.21–23) record relatively large fluctuations in composition (Fig. 11; Table 2a, 18–20), and a return to normal virtual geomagnetic poles. For the entire QTS, correlations between SiO2 and certain trace element ratios (K/Rb 345–260, Rb/Y 1·6–3·2; Fig. 7) indicate open-system evolution, but mg-numbers are sufficiently low (56–44) to suggest a more important role for crystal fractionation than is indicated for QTW10.1 (Muñoz sequence).
|
|
Although the total sample suite generally defines quasi-coherent trends on variation diagrams (e.g. mg-number, Nb/Y, Ba/Sr), the eruptive order of these six variably evolved magma compositions is essentially random (Fig. 11). The absence of a regular temporal progression leads to the conclusion that each compositional group within the QTS represents a separate sub-cycle of magma generation and differentiation. If the preserved lavas are representative of the evolution of the magmatic system, several temporal scenarios are possible. Assuming protracted activity over 5 x 103 to 1x104 years for eruption of the QTS (Singer & Pringle, 1996), relatively large swings between eruptions of basaltic and andesitic magmas would be implied during early stages, and then the system settled on steady-state behavior in which an intermediate composition was repeatedly erupted over a relatively long period. Alternatively, the last 20–25 lava flows of the upper QTS may represent rapid effusion of a large quantity of magma over a short period of time that fortuitously fell immediately before and during the termination of the Bruhnes–Matuyama geomagnetic transition.
In any case, the QTS is the earliest example of several in the TSPC wherein lavas that would appear to define a coherent differentiation trend in the absence of stratigraphic control actually comprise multiple, distinct differentiation events wherein lavas evolved repeatedly from similar parent magmas by similar processes. Our currently favored working hypothesis for the origin of these composite pseudo-trends is that they represent eruptive episodes during which there was recurrent magma emplacement, differentiation, recharge, mixing, and eruption.
|
ESTERO MOLINO SEQUENCE (EMS) |
|---|
Like most sequences exposed in the canyon walls incised into Placeta San Pedro, the three units (lower, middle, and upper) of the EMS appear to be distal flank remnants (WNW) of once much larger volcanic centers rooted in a dike swarm that is centered several kilometers to the east of Placeta San Pedro (Singer et al., 1997
). Although the amount of material removed by erosion is unconstrained, the preserved volumes probably constitute a small fraction of the original edifice(s). Our collection (132 samples) is representative of the Estero Molino units preserved on Placeta San Pedro, but the extent to which these lavas fall short of representing the full duration and total chemical variability of the original edifice(s) is unknown.
The EMS has been divided into three contrasting eruptive periods characterized by diverse parental magmas and distinct internal compositional patterns. None defines a long-term temporal progression from mafic to more evolved compositions. Most evolved lavas and some basaltic units display evidence of open-system differentiation. Age constraints suggest the following: (1) a total duration of 120 ky, but episodic activity and limited preservation; (2) the lower and middle EMS lavas are close in age; (3) the upper EMS (Laguna Verde and Laguna Azul lavas) followed 50–100 ky later.
Lower Estero Molino sequence (620–600 ka)
Lower EMS lavas, which postdate the peak glaciation marked by oxygen isotope stage 16 (620 ka), are present in proximal Quebrada Turbia (QTW12.1–21) and in Estero Molino (EML.0–5), filling a broad erosional depression cut into Quebrada Turbia lavas and basement (Figs 3 and 9). The east–west axis of this depression (centered in geographic upper Quebrada Turbia) projects, and shallows, to the west into geographic Estero Molino; i.e. distal lower EMS flows are trapped between ancient basement highs flanking modern Estero Molino (Fig. 5). Although lower EMS lavas in Quebrada Turbia superficially resemble some flows in the upper and lower Quebrada Turbia sequence, and they lie at about the same elevations at opposite ends of a valley wall (Fig. 9), correlations are eliminated on several grounds, including differences in trace element chemistry, phenocryst mineralogy, magnetic polarity, and stratigraphic order.
Lavas of the lower Estero Molino sequence range in composition from mafic basaltic andesite to high-SiO2 dacite (Figs 10 and 12). They are subdivided into: (1) a lower high-Nb/Y group of andesitic lavas present only in Quebrada Turbia; (2) a middle low-Nb/Y group of basaltic andesitic to dacitic magmas (dacitic pyroclastic unit), including lavas from the EML section; (3) a very tentatively defined upper group that includes andesitic and dacitic lavas in Quebrada Turbia (QTW12.20–21) plus basaltic andesitic lavas that crop out to the west (EML.5 and ESPN.1).
|
Flows QTW12.0–1 (55–55·3 wt % SiO2; Table 2b, 22) and QTW12.2–9 (57·5–60 wt % SiO2; Table 2b, 23–27) are characterized by higher Nb/Y, TiO2, P2O5, Nb, Zr, and Ce, and lower mg-numbers (Fig. 12), than overlying lavas of similar SiO2. The younger low-Nb flows (QTW12.10–19 and EML.0–4; Table 2b, 28–36) comprise alternations of basaltic andesitic and andesitic lavas punctuated by a dacitic pyroclastic unit (Table 2b, 33). Although two basaltic andesitic lavas of the EML section (.0 and .4) resemble lavas QTW12.10–12, the similarities are general, rather than sufficient to warrant a correlation (Table 2b, 29 vs Table 2b, 34). The more evolved andesitic lavas of the high-Nb/Y and low-Nb/Y groups define divergent trends, but in both cases they display evidence of open-system differentiation: (1) mg-numbers that are not significantly lower than those of more mafic associated lavas; (2) strong correlations between increasing SiO2 and certain incompatible element ratios (e.g. K/P, Ba/Y, Rb/Y, Rb/Zr); (3) phenocryst disequilibrium.
|
The boundary between the lower and middle EMS lavas is not defined on the basis of geochronology. Consequently, four lavas that make up a chemical group distinct from the underlying lavas and from packages in the overlying MEMS have been assigned, somewhat arbitrarily, to the LEMS and to a co-magmatic group; i.e. two evolved lavas in the QTW12 section (.20 and .21; Table 2b, 37 and 38) might be related to a pair of correlative basaltic andesitic flows (EML.5 and ESPN.1; Table 2b, 39 and 40), or to similar parental magmas (Fig. 12). These four samples define linear major and trace element trends that are commonly offset or divergent with respect to the underlying groups (e.g. Sr/Y, Ba/Y). On some plots these flows appear to display parent magma affinities with QTW12.0–9 (K/P, Rb/Zr), but on others they are indistinguishable from the trend defined by QTW12.10–19 (Nb/Y, Zr/Y, Ba/Y). Among the diverse compositional trends within the entire EMS, the more evolved lavas of this group display the most obviously fractionation-dominated evolution and one of the strongest Y–HREE enrichment trends observed in the TSPC (Fig. 8e). This is in contrast to the underlying lavas, particularly the low-Nb/Y suite, which display evidence for open-system evolution involving a silicic, probably crustal, component with high K/P and Rb/Zr.
Of some importance to an assessment of the reconstruction approach employed here is the fact that EML.5 and ESPN.1 are nearly indistinguishable compositionally from upper Quebrada Turbia basaltic andesitic lavas (Fig. 11; Table 2b, 21). Multiple splits of these samples have been processed and reanalyzed to confirm these characteristics. On the basis of bracketing ages in the range of 620–570 ka from underlying and overlying units, the uniformly normal geomagnetic pole orientations throughout the EML section, and geometric considerations, a correlation of lower EMS lavas EML.5–ESPN.1 with upper Quebrada Turbia lavas is ruled out despite their virtual chemical equivalence. This case is a prime example of the danger of relying solely on lava chemistry as a correlation criterion.
Middle (579 ka) and upper (495 ka) Estero Molino sequence
Contact relations and lacunae
Upper Estero Molino lavas are more laterally widespread than the lower or middle EMS (Figs 3 and 5), which filled topographic depressions following the peak ice advance of oxygen isotope stage 16 (620 ka). A distal basement high that projects to the west from Quebrada Turbia (Fig. 9) intervenes between flows sampled in Quebrada Castillo (LV–QCNE–QCSE–QC98) and those at the ESPN section. With the exception of ESPN.1, no flows of the lower or middle EMS have been sampled north of this paleodivide (Fig. 13). In addition, the middle EMS is much thicker within the paleodepression that coincides in position with present-day Quebrada Castillo than it is in the counterpart immediately to the south (i.e. Estero Molino; Fig. 5).
|
|
|
The presence of volcano-sedimentary units (‘sed’ in Figs 3 and 9), pronounced chemical shifts in lava composition, and 40Ar/39Ar dating have been used to define boundaries between the lower, middle, and upper EMS units. In Quebrada Turbia the volcaniclastic unit underlying the middle EMS laps northward (down-dip, distally) from the lower EMS (QTW12) onto older QTS lavas, further west in Quebrada Castillo lower EMS lavas are absent, and there is a subtle erosion surface between the lower and middle EMS in Estero Molino. Although these indications of erosion between the lower and middle EMS are not generally accompanied by the large angular unconformities observed between sequences, it appears none the less that the upper boundary of the lower EMS is a truncation surface. Comparable relations at the contact between the middle and upper EMS units also suggest that erosion, probably related to peak ice volumes during oxygen isotope stage 14 (
505 ka), may have truncated the middle EMS before emplacement of the upper EMS. The inference of relatively minor erosional events during Estero Molino volcanism may be reconciled with the marine climate record, which suggests that the 195 ky between the major ice advances marked by oxygen isotope stages 16 and 12 was a relatively stable period characterized by comparatively modest ice volumes. The upper bounding surface of the EMS reflects repeated erosion related to glacial maxima corresponding to oxygen isotope stages 12 (425 ka), 10 (335 ka) and 8 (240 ka); multiple units are in contact with upper EMS flows and contact relations change rapidly over short distances (Figs 3–5, 9 and 13).
Middle Estero Molino sequence: stratigraphy
The middle EMS lavas (Table 2c, 41–57), which are bracketed above and below by sedimentary units (Figs 3 and 9), are widespread, but each package (B–C–D–F) and non-correlative flow (underlying—QTW12.22; overlying—QCNE.1) has an eccentric distribution (Fig. 13). Packages B and F are laterally extensive in Quebrada Turbia (C and D may be present, but were not recognized), and only package B is present in Estero Molino (EML.6; Figs 5 and 13). The middle EMS is thin at both the medial (QTW10.12, package F—only one flow) and proximal (QTW12.22–24, packages B and F) Quebrada Turbia sections (Fig. 3), but is much thicker in the QTW14 section (QTW14.1–4 = B; QTW14.5–9 = F), which was sampled expressly to collect the middle EMS. Only within Quebrada Castillo, where packages C and D are exposed on the canyon floor (linking sections LV–QCNE–QCSE), are all four packages recognized in stratigraphic succession. These relations reflect ponding in a paleovalley (Fig. 5).
QTW12.22 and package B. An evolved basaltic lava (QTW12.22; Table 2c, 41) is the stratigraphically lowest lava assigned to the middle EMS. This assignment is predicated on chemical affinities rather than geochronology, as it resembles overlying lavas of package B (e.g. Ba/Nb 105 ± 10, Nb/Y 0·16 ± 0·02, Zr/Y 5·5 ± 0·3. This basalt is plagioclase-accumulative with minor olivine phenocrysts. It is notable in that it has the lowest Rb (8·2 ppm) and Cs (0·19 ppm), and the highest K/Rb (575), for a TSPC lava with 5·5 wt % MgO, suggesting that it is weakly contaminated despite being rather evolved. The basaltic andesitic lavas of package B (54·4 wt % SiO2; Table 2c, 42–47; plag + aug + minor oliv) are the most widespread in the middle EMS.
Packages C and D. Package C basaltic lavas (Table 2c, 48) are characterized by abundant coarse glomerophyric aggregates of olivine and plagioclase, high MgO, Ni, and Cr, and low Sr, Al2O3, and CaO compared with QTW12.22 and the younger package F basalts. A unique aspect of their chemical signature is exceptionally low V in relatively primitive compositions (155 ppm vs 175–215 in other basalts of the TSPC). Other distinguishing characteristics in comparison with the other basaltic compositions of the middle EMS include relatively high P2O5, Nb, Ba, and Zr, plus high Zr/Y, Nb/Y, Ba/Y, Ba/V, and Ba/Sr. The two analyses of package D basaltic andesitic lavas are different from each other, although similar affinities are evident and they are petrographically similar except for more olivine in MgO-rich lava QC98.4 (Table 2c, 49 and 50). Package D lavas share many characteristics with the package C basaltic lavas, including phenocryst mineralogy, low Sr, similarities among trace element ratios such as Ba/Nb, Ba/Y, Zr/Y, and Nb/Y, and relatively high P2O5, Zr, and Nb.
Package F.These olivine-phyric basalts are unique in the Estero Molino sequence in having relatively high Sr (>600 ppm) and Al2O3 (>18·4 wt %), but low plagioclase abundances, in combination with low P2O5, TiO2, Y, Nb, and Zr (high K/P, Sr/Y, Sr/Rb, and Ba/Nb plus low Ba/Sr and Zr/Y; Table 2, 51–56). Broadly similar characteristics reappear later, however, in the basaltic Upper Placeta San Pedro sequence (234 ka), and in some late-Tatara basaltic andesitic lavas (package 
3). These lavas are moderately evolved (Fo85 plus low Cr and Ni compared with package C and QCNE.1), but their low K2O and Rb contents and high K/Rb are suggestive of minimal crustal contamination. Large ranges in some trace element ratios (K/P, Sr/Rb, K/Rb = 614–966, Rb = 5–10 ppm) in lavas that do not display comparable major element variations, or ranges in trace element ratios less readily disturbed by assimilation (Zr/Y = 3·4–3·9 and weakly correlated with K/Rb), are apparently due to minor but variable degrees of contamination (0·5–2 %) by a high-Rb (150–250 ppm), low-K/Rb (250–150) component.
QCNE.1. The 40Ar/39Ar age of this primitive basaltic lava indicates that it belongs in the middle EMS. Although the bulk lava chemistry reflects 1·5% olivine accumulation, its forsteritic phenocrysts (Fo88), low Rb, K2O and SiO2, and an absence of textural evidence for hybridization, suggest limited differentiation and minimal contamination or mixing during ascent. It is distinguished from package F basalts by the presence of coarse glomerophyric aggregates of olivine and from the package C basalts by an absence of plagioclase phenocrysts. The chemistry of QCNE. one more closely resembles that of package C than package F, but it is distinct from both.
Middle EMS: summary
A high proportion of primitive basaltic magma compositions sets the middle EMS apart from all other units of the TSPC (Figs 10 and 14a–d). The middle EMS (Fig. 12) includes three primitive, nearly uncontaminated basaltic compositions (packages C and F plus QCNE.1), and an evolved but not significantly contaminated basaltic lava (QTW12.22). These basaltic lavas are rather diverse for both major elements (Fig. 10) and trace elements (Fig. 14a–d), reflecting different mantle-derived chemical signatures and distinct differentiation histories. The two intervals of basaltic andesitic lavas in the middle EMS are plausibly related to mafic parental magmas not unlike the immediately underlying basaltic lavas on the basis of indices such as Nb/Y (QTW12.22 
B?; C D?). If so, the associated differentiation paths do not reflect closed-system fractional crystallization.
|
Subdividing the upper EMS: Laguna Verde and Laguna Azul lavas
The upper EMS comprises two sub-groups of unresolvable age (Fig. 13). The lower sub-group, the Laguna Verde lavas (Figs 10 and 14e–h), is marked by compositional complexity and an absence of thick, voluminous packages: it includes a high proportion of non-correlative flows (EML.7–11, QTW12.25?) plus four thin but laterally extensive packages (H–I–J–K). The Laguna Verde lavas are thicker on western Placeta San Pedro, being nearly absent from more proximal sections in Quebrada Turbia. The maximum thickness, which is centered in Estero Molino, is shifted to the south relative to the maximum thickness of the underlying middle EMS, which is coincident with Quebrada Castillo. The overlying Laguna Azul lavas (packages M–Q–R–S–U–W; Figs 13 and 15–18), are chemically more coherent than the Laguna Verde lavas, they include two thick, laterally extensive packages (Q and W), few non-correlative flows (LV.10 and LV.12–13), and the aggregate is tabular in form with little relief on the basal contact—total thickness and preservation are variable as a result of truncation by subsequent erosion (Fig. 9).
|
|
Laguna Verde lavas (upper EMS): stratigraphy
Non-correlative flows. The stratigraphically lowest lavas of this unit comprise flows EML.7–11 (Table 2c and d, 59–62), consisting of a basaltic lava (EML.7), a mafic basaltic andesitic lava (EML.8), and three evolved basaltic andesitic lavas (EML.9–11). A single high-MgO (olivine-accumulative), but high-SiO2 basaltic flow in upper Quebrada Turbia (QTW12.25; Table 2c, 58) is placed with the Laguna Verde lavas but this assignment is almost entirely arbitrary, as it has a composition unlike any other TSPC basaltic lava (e.g. 400 ppm Sr). Although QTW12.25 overlies package F (middle EMS) and underlies package W (Laguna Azul lavas, upper EMS), its stratigraphic position is otherwise unconstrained (e.g. it could belong to the middle EMS), as it is the only flow sampled within Quebrada Turbia that could be a Laguna Verde lava.
|
Packages H–I–J–K. Package H (LV.4 and EML.12; Table 2d, 63 and 64) consists of low-Nb/Y basaltic andesitic lavas (
52·3 wt %) with lower MgO, Ni, and Cr (mg-number 55) than most evolved middle and upper EMS lavas with 53–56 wt % SiO2 (compare EML.11), suggesting that fractional crystallization has played an important role. Package H is exceptional with respect to other mafic Laguna Verde lavas in that it manifests higher Sr concentrations (640–650 ppm) than diverse associated basaltic lavas (Sr = 400–580 ppm). Two high-MgO basaltic lavas (package I: EML.13 and ESPN.2; Table 2d, 65 and 66) and the overlying basaltic andesites (package J: EML.14 and ESPN.3–4; Table 2d, 67 and 68) are assigned package status. Their consecutive occurrence at widely separated sections probably reflects near-source divergence of lava tongues into two topographic depressions. The relatively high-SiO2 and high-MgO basaltic lavas of package I (Fo83; olivine-accumulative) contain resorbed plagioclase phenocrysts suggestive of a hybrid origin. The fourth package (K) consists of olivine–plagioclase phyric basaltic andesitic lavas (EML.15 and LV.5B; Table 2d, 69 and 70) with a trace element signature (shared by package I), which in the case of many trace element ratios represents extreme values for the TSPC (Zr/Y 10·3; Ba/Y 28) when they are compared with rocks of the same SiO2 (55·5 wt %).
Laguna Verde lavas: summary
The Laguna Verde lavas are notable in that most packages and non-correlative flows are not related to stratigraphically adjacent units. Even in cases where multiple flows or packages have similar chemical signatures, they commonly are not in consecutive stratigraphic order (e.g. packages H–J and I–K; Fig. 14e–h). The preserved flows representing the Laguna Verde interval on Placeta San Pedro may record incompletely the overall magmatic history of the corresponding edifice because many intervening eruptions are absent from the sampled sections. Alternatively, there may have been multiple, concurrently active magmatic systems. In any case, there is little basis for constructing a model for the temporal evolution of this diverse assemblage of lavas. Although the overlying Laguna Azul lavas span about the same range in SiO2 as the Laguna Verde lavas, this subsequent phase of activity records a more coherent magmatic evolution history. Unlike the primitive, uncontaminated basaltic lavas of the middle EMS, the Laguna Verde basalts are hybrid magmas (QTW12.25, EML.7, package I), with high SiO2 (51–51·7 wt %).
Laguna Azul lavas (upper EMS): stratigraphy
The Laguna Azul lavas are much more laterally extensive than the underlying units of the EMS, and this unit is dominated by thick flow packages. Temporal patterns defined by variations in magma chemistry are constrained by high-density sampling and by robust correlations among different sampled sections, which are supported independently by data from 28 paleomagnetic sites. On the basis of paleomagnetic data from multiple sections (L. L. Brown, personal communication, 2000) the Laguna Azul lavas are divided into lower and upper eruptive episodes: (1) M + Q + LV.10–11 + R + LV.13, during which magmas generally became less evolved up-section (e.g. decreasing SiO2); (2) S + U + W, during which lavas became, for the most part, more evolved up-section (Fig. 16). Sites in package Q flows have average inclinations and declinations (n = 10: I = –38·9°, D = 359·5°) indistinguishable from those of LV.10–13 (n = 4: I = –37·6°, D = 6·2°). Packages U (n = 3: I = –45·9°, D = 339·9°) and W (n = 9: I = –49·1°, D = 346°) are indistinguishable but have magnetic directions that are different from those of the underlying episode. Interbedded sediment separates package Q from the base of S + U + W in Quebrada Turbia (Fig. 9).
Laguna Azul lavas (upper EMS): lower eruptive episode
Packages M and Q. Package M, the basal package of the Laguna Azul lavas, comprises one widespread lava or a small number of nearly compositionally and petrographically identical lavas (ESPN.5, LV.5A, QCNE.2; Table 2d, 71–73), which are distinct from the underlying and superficially similar package K (Laguna Verde lavas). The phenocryst mineralogy of package M includes olivine plus two pyroxenes (pyroxene phenocrysts are absent from package K), which display reaction rims that are inferred to be the consequences of mixing of mafic and intermediate magmas. Although MgO is not exceptionally high, the low Fe2O3* of these lavas results in elevated mg-numbers (56·5; Fig. 17a) for lavas of 55·5–56 wt % SiO2.
Package Q (Table 2d and e, 74–81) is present in southern (EML.16A-21; QCSE.1–8) and central (QCNE.3–5; LV.6–9; QTW10.13–17) Placeta San Pedro (Fig. 13). It is a thick assemblage of basaltic andesitic lavas (54·3–55 wt % SiO2), which become slightly less evolved up-section. Lava QTW10.13 is indistinguishable in chemistry from the average composition of package Q, but QTW10.14 is an outlier with respect to many compositional indices, and is petrographically distinct. QTW10.17 defines a compositional extreme on many plots, and it is the most evolved composition with affinities to package Q, even though it occupies a high stratigraphic position; i.e. contrary to the general up-section trend. The fact that flows comprising QTW10.13–17 are bounded above and below by prominent clastic units suggests that both QTW10.14 and QTW10.17 are physically within package Q, although QTW10.14 may have evolved in a separate magmatic system. Package Q lavas contain olivine, augite, and orthopyroxene phenocrysts, but textural disequilibrium is less evident than in package M. The mutual affinities of packages M and Q are indicated by similarities of many incompatible trace element ratios, but major elements and compatible trace elements do not permit a simple petrogenetic relationship between these two compositional subtypes. For example, although package Q lavas have lower SiO2 than package M, package Q lavas also have lower Cr and lower mg-numbers (higher Fe2O3* in package Q).
|
Package R and associated lavas. Within the interval between packages Q and S–U (LV.10–13; Table 2e, 82–86; Figs 13 and 16), there is a heterogeneous assemblage of flows that are generally less evolved than package Q, at least in terms of major elements. One flow (LV.11) that is distinctly different petrographically from all other Laguna Azul lavas is intercalated with three flows (LV.10 and LV.12–13) that chemically resemble the thick underlying package Q (LV.12 is indistinguishable from the average composition of Q). As LV.11 is chemically and petrographically indistinguishable from EML.22 (three samples), which directly overlies package Q (EML.16–21), EML.22 and LV.11 have been grouped as package R (flows overlying EML.22 in Estero Molino have not been analyzed, but petrographically distinct flows representing the S–U–W triplet extend southward into Estero Molino).
Laguna Azul lavas (upper EMS): upper eruptive episode
Packages S and U. Neither of these thin packages is as widespread as package Q or the overlying package W. The relative stratigraphic order of S and U is not known as they are not present in the same sampled section. Package S is found only within the upper Quebrada Castillo drainage (QCSE.9–10, QCNE.6–9; Table 2e, 89–92). Package U is similarly thin and restricted, but it has a more northwesterly distribution (LV.14, QTW10.18–19; Table 2e, 87 and 88). Package U is less evolved than package S (51·8 wt % SiO2 vs 52·5 wt %), but these two groups of mafic basaltic andesitic lavas display close chemical affinities (alphabetical order reversed in Table 2 to emphasize uncertain relative ages). Olivine is far more abundant than augite and no orthopyroxene has been identified. Plagioclase phenocrysts in both packages tend to be unusually elongate and they have distinctive marginal resorption zones.
Package W. Package W (Table 2e, 93–100) is the highest preserved unit in the upper EMS, and it is defined as 15 related flows (52·1–54 wt % SiO2) sampled in upper Quebrada Castillo (QCSE.11–12) and Quebrada Turbia (QTW12.26–29, QTW10.20–28). Although there are some internal compositional excursions (e.g. SiO2, mg-number, Ba, Sr) and flow QTW10.27 is an outlier for many concentrations and ratios (Fig. 16), the coherence of this unit is demonstrated by nearly constant or smoothly varying trace element concentrations (e.g. Y, Zr, Rb) and ratios (e.g. Sr/Y, Ba/Rb, Sr/Rb, Zr/Y).
Temporal evolution of the Laguna Azul lavas (upper EMS): implications
Laguna Azul magma compositions define narrow linear arrays with respect to many major element concentrations and ratios vs SiO2 (Fig. 15a–c), implying that the cotectic relations governing multiple saturation during magma differentiation did not change appreciably from one eruptive event to the next. If this inference is correct, then factors such as depth of magma differentiation and/or water content of parental magmas were not highly variable. The relative constancy of Nb/Y (Fig. 17c), compared with the larger ranges for this ratio among basaltic lavas in other stratigraphic units of the Estero Molino sequence, is an indication that parental magma composition did not change dramatically either during or between the two short-lived volcanic episodes required by the paleomagnetic data. However, few trace element concentrations or incompatible element ratios (Rb, Ba, Sr, Zr, Y, K/Ti, K/P, Na2O/K2O, Sr/Rb, Sr/Y, Ba/Rb; Figs 15–17) define trends with respect to SiO2 that are as linear or as narrow as are those for most major elements or Nb/Y. In particular, there are relatively large ranges for most incompatible element concentrations in magmas with between 53 and 54·5 wt % SiO2.
A partial explanation for this apparent dichotomy is evident when these data are viewed in stratigraphic context (Figs 16 and 17). The aggregate major and trace element arrays on Harker diagrams are seen to be pseudo-trends defined by the products of multiple, heterogeneous eruptive events rather than by successive eruptions along a single, continuous magma evolution path. Although this observation or conclusion applies to many groups of lavas discussed in this paper (e.g. Quebrada Turbia sequence), the Laguna Azul lavas are a complex but especially well-documented example that is instructive with respect to the impact of a well-established eruptive order on the formulation of petrologic models. Consequently, compositional relations among these lavas are illustrated and discussed in more detail than are those of the other stratigraphic units of the Estero Molino sequence.
Mixing versus fractionation-dominated evolution? At the first-order level, pseudo-trends such as those illustrated in Fig. 15 could arise from two end-member differentiation scenarios: (1) each eruptive event comprises variably evolved end products of similar but physically and temporally separate liquid lines of descent; or (2) each eruptive event is the consequence of mixing broadly similar magmatic components in variable proportions. Whereas crystal fractionation undoubtedly played a role in the evolution of these magmas, the overall trends of the data arrays cannot be the direct consequence of fractionation-dominated differentiation. Decreases in MgO, Cr, and Ni with respect to increasing SiO2 are far too limited to be consistent with closed-system fractionation (initially plag + oliv, then + aug + oxide). Moreover, greater than two-fold increases in highly incompatible elements such as Rb (Fig. 15e) are much too large to be compatible with the 35–40% fractionation needed to evolve from 51·5 to 56 wt % SiO2. The most evolved magmas, in terms of SiO2, are the package M andesitic lavas, which display the clearest petrographic evidence of hybridization (oliv + 2 pyx) and which have Cr and mg-numbers nearly as high as the most mafic compositions in the suite (Figs 15–17). These basic constraints require that mixing of mafic magmas with more evolved, probably andesitic end-members was important in determining the compositions of erupted magmas.
Divergent open-system evolution paths and multi-component mixing. Compositional variations within packages and distinctions among these internal trends from package to package, particularly with respect to trace element ratios, demonstrate that the lavas of each package represent magma genesis and differentiation events separate from subsequent and preceding events (Figs 16 and 17). Some flow packages that overlap with each other in certain variation diagrams are typically well separated in other plots (e.g. W and R), and apparently coherent trends defined by groups of stratigraphically associated packages on some diagrams break apart into divergent sub-trends on other diagrams (e.g. S–U–W). A general explanation for this diversity is that during each magmatic event the mafic and evolved components that mixed before eruption acquired unique trace element signatures as a result of some combination of fractionation, assimilation, and mixing that implicated some combination of components different from those that characterized other periods of activity. Whereas these differences are not enormous, they are sufficient to render each package distinct from the others. Although we cannot propose specific explanations grounded in detailed modeling for these distinctions at this time, a few examples are discussed briefly to verify the assertion that the major element arrays in Fig. 15 create an illusion of simplicity that is refuted by trace element signatures.
In terms of SiO2, the package R lavas (53·2–53·7 wt %) fall within the range of the package W lavas (52–54·2 wt %), and the two groups overlap for a number of other variables in Fig. 17. None the less, package R has lower mg-numbers, Cr, Ni, Rb, Rb/SiO2, and K/Ti, in combination with higher Sr, Al2O3, Na2O/K2O, Sr/Y, and Sr/Rb, than package W. These distinctions, particularly the different relative positions of these two groups in the mg-number–SiO2 and mg-number–Rb/SiO2 plots (Fig. 17a and b), imply that one or more of the mixing end-members implicated in the formation of package R magmas evolved in a system that was less open to crustal input and less affected by assimilation than the system in which the package W lavas evolved. Although the details are different, the same general relationships and inferences apply to package R vs the bracketing non-correlative lavas (LV.10 and LV.12–13) that more closely resemble the underlying package Q (Fig. 16).
Taken together, the Laguna Azul lavas define broad trends of increasing K/P and decreasing Na2O/K2O, Sr/Rb, Sr/Y, and Ba/Rb with respect to SiO2. In detail, packages W and Q have the largest spreads in SiO2, they define ‘micro-trends’ that are discordant with respect to these general trends, and the uppermost five samples from package W display a complex decoupling between some major and trace elements (Fig. 16). In particular, package W, anchored at the low-SiO2 end by the exceptionally high Rb/SiO2 sample QTW10.27 (Fig. 17b), is nearly constant throughout for K/P, Sr/Rb, Sr/Y, and Ba/Rb, in contrast to underlying package S, which defines steep orthogonal trends for Sr/Rb and Ba/Rb vs SiO2. The divergent trends defined by packages W and S can also be inferred from temporal variations in Rb, K/P, Sr/Rb, and Ba/Rb in Fig. 16.
Whole-rock compositional variations as a function of stratigraphic position in package Q generally define a weak up-section trend toward less evolved compositions, but in detail the pattern is irregular and not readily interpretable. For example, in the lower part of the unit there is a decrease in mg-number accompanied by increasing Ni, Ba, and Sr (Fig. 16). For most, but not all, compositional variables the vast majority of samples from package Q (and LV.12) plot in variation diagrams (Figs 17 and 18) as a tight cluster within a triangular field defined by the same three bounding compositions—EML.21, QTW10.16 and QTW10.17. The lava with the most evolved composition in terms of SiO2 and Cr (QTW10.17) is characterized by incompatible element ratios intermediate between those of the two least evolved lavas, one of which (EML.21) has the lowest Rb and Rb/SiO2 and the other (QTW10.16) has Rb and Rb/SiO2 as high as QTW10.17 (Figs 17 and 18). Thus, incompatible and compatible trace elements do not vary systematically as a function of major element trends. If package Q lavas are hybrid magmas, at least three end-members were involved: if EML.21, QTW10.16 and QTW10.17 are themselves hybrids, then the actual mixing end-members were still more disparate. Moreover, the package Q array does not lie consistently on a mixing line between the more evolved underlying package M compositions and the overlying slightly less evolved compositions of LV.10 and LV.13. To further complicate matters, QTW10.14, which lies stratigraphically within package Q, is an unrelated magma, and LV.12, which just as clearly lies physically outside package Q, falls compositionally near the center of the package Q cluster.
In summary, the idiosyncratic internal variations of several packages of the Laguna Azul lavas are comparable in character with the magma heterogeneity documented by Gamble et al. (1999) for recent eruptions of Ruapehu Volcano (NZ), wherein syn-eruptive magma mixing has been inferred and there are also considerable isotopic variations that most plausibly relate to diverse pre-mixing contamination histories. These complexities have been interpreted (1) in terms of the importance of mafic magma recharge as an eruption triggering mechanism, and (2) as an indication of a distributed subvolcanic conduit system lacking a large central reservoir. The Laguna Azul lavas appear to be an ancient example of a magmatic system that evolved in a similar manner. The prevalence of broadly analogous patterns of magma variation during previous (Quebrada Turbia) and subsequent (upper Placeta San Pedro and upper Volcán Tatara) episodes of activity at the TSPC suggests that such behavior may be typical of continental arc volcanoes during periods of dominantly mafic to intermediate activity.
|
PLACETA SAN PEDRO SEQUENCES |
|---|
Lower Placeta San Pedro sequence (LPSPS)
In Quebrada Castillo and Quebrada Turbia, and on Cordón Los Lunes, thick heterogeneous mingled flows of generally silicic andesitic composition fill depressions on the unconformity cut into the upper surfaces of the Estero Molino sequence, the Quebrada Turbia sequence, and Muñoz sequence lavas. Although these flows do not form a continuous unit, it is reasonable to suppose that they represent distal, valley-filling flows that issued from a common edifice, one that produced lavas markedly different from the Estero Molino sequence. As these lavas resemble the mingled andesitic to dacitic lavas of Cordón El Guadal in the southeast quadrant of the TSPC, which are intermediate in age between the Estero Molino and upper Placeta San Pedro sequences (Feeley & Dungan, 1996
; Singer et al., 1997), they are, most probably, far-traveled distal flows from the Cordón El Guadal edifice that was centered to the east of Cajón Muñoz. An andesitic lava exposed continuously along the northern wall of Quebrada Castillo (QCNE.10 and LV.15) is correlative with QTW12.30 (Quebrada Turbia) on the basis of chemistry (Table 2f, 101 and 102) and petrography. An isolated silicic dacitic lava (Table 2f, 104), surrounded by talus and moraine, occurs within the upper Estero Molino drainage near the unconformity between the EMS and Volcán Tatara, and near the southwestern wedge-out of the intervening UPSPS (Figs 3–5). This undated dacite lacks hydrous minerals despite its evolved composition, and because of its geometry and obscure contact relations it cannot be placed definitively in stratigraphic context. On the basis of chemical similarities with QTW11.24 (Table 2f, 103), it is tentatively designated as an LPSPS flow. This interpretation is supported by a new 40Ar/39Ar date of 421 ka on a nearby flow (EMU4.3–4; Table 2f, 105) that is similarly largely buried by snow, moraine, and lower Volcán Tatara lavas.
Upper Placeta San Pedro sequence (UPSPS)
Northern Placeta San Pedro and Cordón Los Lunes are capped by porphyritic basaltic lavas (49·9–52·3 wt % SiO2; Fig. 19) of the Upper Placeta San Pedro sequence (UPSPS). These flows are individually thick (5–12 m) and they form a substantial accumulation (minimum of 30 flows, top eroded, minimum original thickness >200 m more than 3 km from source) lacking evidence for an internal hiatus in activity. Paleomagnetic inclinations and declinations from 22 sites define an extremely tight grouping consistent with eruption during a short period (L. L. Brown, personal communication, 2000). Flow directions to the WNW are indicated by primary dips and the geometry of unusual ‘mega-pahoehoe’ structures, indicating that they are a flank remnant of a now largely destroyed edifice that was located to the east and south of geographic upper Quebrada Turbia. In all, 66 samples from seven sections have been analyzed. Supplementary material is presented in Fig. II-1a–j and Table II-1 in Electronic Appendix II.
|
Identifying specific UPSPS flows in multiple sampled sections on the basis of chemical correlation is tenuous, primarily because the flows are internally heterogeneous. However, early UPSPS lavas sampled in distal sections QTW11.26–38 and ESPN.6–11 filled a paleovalley and are characterized by relatively low K2O and Rb. These were, for the most part, emplaced before the flows sampled in QTW10.29–48 and in Quebrada Castillo (Figs 2 and 9), which generally have higher incompatible element abundances, but are significantly variable from flow to flow and within flows. Samples from Quebrada Castillo (QCSE.13, LV.16–21, QCNE.11–18) show a progression from one or two basal lavas with low K2O and incompatible element abundances to overlying flows typical of nearby QTW10.29–48. Three lavas (QTW11.35–36, ESPN.8) are distinguished by the lowest incompatible element abundances in the UPSPS and these are located at the boundary between the early low Rb–K2O group and the later more enriched variants.
All UPSPS flows contain abundant small olivine and plagioclase phenocrysts, and many flows contain up to 10% coarse xenocrystic fragments apparently derived by disaggregation of two different cumulate lithologies. Large rounded multi-grain aggregates of augite (2–6 mm) containing minor inclusions of plagioclase, oxide, and resorbed olivine, all of which are in reaction relationship with the host magma, appear to be fragments of clinopyroxenitic cumulates. Exceptionally large (3–10 mm) embayed olivine crystals with minor inclusions of plagioclase are present in many lavas, and in an intact open-textured troctolitic xenolith, wherein a large (>1 cm) embayed olivine–plagioclase intergrowth occurs in a matrix with grain size comparable with that of typical phenocrysts (<1 mm). Some of these olivines contain melt pockets in which mica has grown during post-eruptive cooling. Although there are few xenocrystic fragments that suggest a gradation between the troctolitic and clinopyroxenitic lithologies, it is possible that these are in fact cognate cumulates derived from different levels of the UPSPS conduit–reservoir system, particularly as both phenocrysts and at least some large olivines in high-mg-number lavas have about the same composition (Fo82–83). The modal proportions of plagioclase, augite and coarse olivine vary within single outcrops, and such variations are recorded in heterogeneous lavas with MgO, Cr, and Ni higher than in xenocryst-poor lavas (e.g. QCNE.13 and LV.17D; Table 2f, 107–108).
None of the UPSPS samples is likely to correspond perfectly to a liquid composition, but sequential flows QTW11.35–36 (Table 2f, 106) have minimal quantities of xenocrystic material, concentrations of MgO, Cr, and Ni intermediate between the extremes, the lowest abundances of incompatible elements, and extreme values for many elemental ratios (e.g. K/Rb 600; Na2O/K2O 5·5; K/P 7; Sr/Rb 85; Rb/Y 0·6; Nb/Y 0·12; Ba/Sr 30; Zr/Y 4·7). These may be representative of the parent magma composition for the entire suite. This inference is not easily tested, as the UPSPS basalts define broad and irregular polygonal fields in many element–element and element-ratio plots, in some cases with the putative parental composition at one apex. Although the more than three-fold enrichment in Rb is generally well correlated with K2O (1·9 times maximum enrichment), other incompatible trace elements such as Y, Ba, and Zr showing comparable maximum enrichments (1·3–1·8 times) are poorly correlated with Rb, as are many trace element ratios (e.g. Ba/Sr, Nb/Y, Zr/Y). To portray such complexities, the data are color-coded for a third compositional variable (e.g. [Rb] or [MgO]). Regardless of which elements or ratios are plotted, some samples that are nearly constant for the third variable are characterized by a range of values for the other two (Figs 19 and 20; Fig. II-1 in Electronic Appendix II).
|
If QTW11.35–36 were parental to the most Rb-rich lavas, an unrealistic minimum of 65% fractionation (bulk DRb = 0) would be required to generate the higher large ion lithophile element (LILE) concentrations in the evolved magmas. Such a model is untenable because the mg-numbers plus Cr, Ni, Sr, and Al2O3 concentrations of many of these Rb-rich samples are not significantly different from those of the likely parent magmas, yet such ratios as K/Rb vary over large ranges. Even for sample QTW11.33 (Table 2f), which has the lowest mg-number in the suite (54), moderately low Rb plus high K/Rb, and low Cr, the amount of olivine + plagioclase fractionation is severely limited by the modest differences in major elements. Comparisons of the UPSPS lavas with two contrasting suites of SVZ lavas that have evolved largely by fractional crystallization emphasize this point (Fig. 19d). Lavas along the low-P fractionation-dominated evolution trend from Volcán Puyehue (Gerlach et al., 1988), which have low Sr/Rb and high Rb comparable with high-Rb UPSPS basalts, are andesitic with 57 wt % SiO2 and low mg-numbers (38). The essentially aphyric high-Sr lavas of package 3 (upper Volcán Tatara; Table 2k, 205–210) are nearly identical to the high-Rb UPSPS lavas in terms of SiO2 and mg-number, but they have very different trace element systematics. Although the 3 lavas may be weakly contaminated, their high Sr/Rb is due in part to fractionation of a plagioclase-poor mineral assemblage. As shown by the shaded mixing regions in Figs 19 and 20, mixing between QTW11.36 and multiple evolved end-member magmas may have contributed to the combination of large ranges in trace element concentrations and ratios and a limited range of SiO2 and some other major elements.
|
Ten mixing curves have been calculated for illustration purposes using a wide range of mixing end-members, all from the TSPC (Table II-1, Electronic Appendix II). The black curve (Mix1) corresponds to mixtures of the inferred parental composition and a UPSPS basalt with high Rb (QTW10.38; Table 2f, 110), which (1) is a representative endpoint on the trend defined by the majority of the high-Rb, high-SiO2, and low-MgO lavas, and (2) itself presumably lies on a mixing line involving a still more evolved end-member. The other nine end-members were chosen from the entire TSPC dataset to most closely approximate the variations in UPSPS basalts. These range from dacitic to basaltic andesitic compositions, and among the latter, some evolved dominantly via fractional crystallization and others correspond to hybrid compositions. These curves terminate at 52·25 ± 0·10 wt % SiO2 for comparative purposes. Most of these mixing combinations result in MgO, CaO, Fe2O3, Al2O3, and Sr contents at 52·25 wt % SiO2 that deviate <±10% from those in Rb-rich UPSPS lavas, but Rb contents range from 16 to 29 ppm and Cr is highly variable (30–145 ppm); i.e. mixing trends involving fractionation-dominated evolved end-members are systematically lower in Cr, MgO, and Rb (Rb/SiO2), and have relatively high Sr/Rb at a given SiO2 in comparison with those involving hybrid end-members (Figs 19 and 20; Fig. II-1, Electronic Appendix II).
None of these nine end-members serves, for all elements and all elemental ratios, as the fictive end-member that is implied for the majority of the more evolved UPSPS lavas (i.e. a composition that lies along projections of Mix1). Consequently, the prevalent mixing component (or the prevalent sum of multiple end-members) is unlike any TSPC lava for which we have an analysis, particularly in terms of being excessively enriched in Rb with respect to other incompatible elements. The problem is illustrated by K/P–SiO2 and K/P–Rb systematics (Fig. 19e and f). All the mixing curves trend towards high SiO2 with respect to K/P, and although some reduction in SiO2 is evident for olivine-accumulative samples, a maximum of 8% olivine accumulation can only reduce Rb by 1–2 ppm. For example, samples with 12–22 ppm Rb at K/P 10 (Fig. 19f), which are characterized by weak mutual correlations between MgO and SiO2, require an even more diverse array of high-K/P mixing components than is shown. In addition, a small number of evolved samples (e.g. QTW11.33–34) seem to require mixing end-members with lower mg-numbers, Cr, and Rb, thus relatively high Sr/Rb and K/Rb, broadly corresponding to magmas that have evolved dominantly via fractional crystallization (Figs 19 and 20). If this inference is correct, multiple mixing end-members are required but their number and diversity are undetermined. Evolved UPSPS lavas with high K/Rb might record distinct evolution paths involving AFC rather than mixing, but AFC would appear not to be the dominant process in light of the presence of xenocrysts and possibly the sporadic accumulation of phenocrysts. Multiple evolved mixing components is a more likely explanation.
These observations raise the question of the degree to which the partly digested clinopyroxenitic or troctolitic xenocrystic material may have contributed to the incompatible trace element budgets of these lavas. As the troctolitic source may have been only partly solidified, interstitial melt may have been incorporated simultaneously along with xenocrysts, or mica and hornblende (high K/P and Rb) may have been present as reaction products as a consequence of open-system processes in the cumulate pile (magma or aqueous fluids), such as is observed in two suites of gabbroic xenoliths from Volcán San Pedro (Costa, 2000). However, poor correlations among certain elements and elemental ratios (e.g. high K/P, Nb/Y, and Ba/Sr at relatively low Rb) may also require multiple parental magmas.
A plausible physical model for the generation of the UPSPS hybrid magmas is the invasion of a partially solidified subvolcanic conduit–reservoir system by repeated pulses of moderately fractionated basaltic magma, not unlike the model proposed by Albarède & Tamagnan (1988) and Albarède et al. (1997) for some picrite lavas of Piton de la Fournaise volcano (Réunion). A similar combination of processes and components, which have masked mantle chemical and isotopic signatures, has been identified in Reykjanes Peninsula basaltic lavas (Gee et al., 1998). The UPSPS basaltic magmas incorporated lithologically diverse, partially solidified cumulates leading to hybrid magmas contaminated with variable amounts of refractory crystals plus evolved residual liquid and/or hydrous phases such as mica or hornblende (melted following entrainment). Mixing with other batches of evolved magma residing within the volcanic plumbing system almost certainly occurred. These complexities prevent us at present from quantitatively modeling the contamination–hybridization processes and end-members.
None the less, some concrete conclusions may be drawn concerning the impact of high-density, stratigraphically controlled sampling on the formulation of working hypotheses for such complicated suites of lavas. As only two of 66 samples are candidates for non-hybrid status, the utility of comprehensive sampling vis à vis the identification of parental mafic lavas that might carry an uncontaminated mantle geochemical signature is evident. Although the non-hybrid lavas (QTW11.35–36) are plausible parental magmas for most of this suite, a rigorous assessment of whether or not there was variability among mantle-derived magmas during UPSPS volcanism has been rendered difficult by the chemical and mineralogical overprint of crustal-level processes involving multiple components. It follows that basaltic continental arc magmas with 6–8 wt % MgO cannot be assumed a priori to faithfully preserve mantle geochemical signatures.
|
VOLCÁN TATARA |
|---|
Contact relations, lacunae, and preservation
Unlike the remnants of highly degraded older volcanoes that are exposed on Placeta San Pedro, Volcán Tatara is a relatively well-preserved edifice. For the purposes of petrologic sampling and stratigraphic control, the flank lavas of Volcán Tatara are well exposed in glacial valleys, particularly on the western flank. There are no samples from the high northern flank between Quebrada Turbia and the UEP7 section in upper Estero Pellado, and only a few stratigraphically uncontrolled samples exist from the distal southern flank (ESPE3). Volcán Tatara is divided into lower and upper lavas on the basis of an erosion surface that has been identified at high elevations on the west flank (Fig. 6).
The beheaded vent region of Volcán Tatara is located near the crest of a buried south-facing escarpment of pre-Tatara age which cuts down-section from TSPC units into basement. North of Cordón Tatara, which strikes approximately east–west through the western flank and vent region, the base of the Tatara sequence lies mainly on Estero Molino and Placeta San Pedro sequence lavas at elevations near 2500 m (Figs 3–5), although lower Tatara flows also filled a paleovalley that was coincident with present-day Quebrada Turbia (Fig. 3). An isolated thin remnant of early lower Tatara lavas (package 
2) located 8 km north of the Tatara vent on Placeta San Pedro indicates that a large volume has been removed from the north flank of Volcán Tatara (Figs 2 and 5). South of Cordón Tatara, in Estero San Pedro, basal Tatara lavas are in contact with basement, and this contact is as low as 2250 m proximally and as low as 1500 m distally (Figs 2 and 6). Sections ESPW2 and ESPW3 are located on the northern flank of this basement ridge and expose higher proportions of early lower Tatara flows than does the more distal section ESPW4, which contains a high proportion of interbedded volcaniclastic sediment. The western flank of Cordón El Guadal impeded the spread of lower Tatara flows to the east, but the absence of lower Tatara flows on the ESE flank also may be due to erosion before eruption of upper Tatara lavas, which are present in sections on the upper west wall of Estero Pellado and in Quebrada Honda (Fig. 2).
As pre-Tatara sequences preserved on Placeta San Pedro flowed to the west or northwest from centrally located source vents, it is logical to suppose that these constructs also produced southern flanks. The absence of such units around the southern periphery of the complex, except for the Cordón El Guadal lavas, indicates that the remnants of previous volcanic episodes have been removed down to basement, apparently by cumulative glacial erosion as a result of ice flowing to the SSW from topographic highs on the earlier Volcán Guadal and Volcán Pellado, and from the now eroded source(s) of the Estero Molino and the upper Placeta San Pedro sequences. This erosion was accomplished during and before the glacial maximum corresponding to oxygen isotope stage 6 (130 ka).
Comparisons with previous work
References to observations and interpretations of Ferguson et al. (1992), who studied Volcán Tatara on the basis of 50 chemical analyses, are made below for the purpose of contrasting the amount of additional information revealed by the more intensive sampling of this investigation. Ferguson et al. (1992, fig. 5) divided Volcán Tatara into lower and upper ‘units’ corresponding to two magmatic ‘cycles’ in which the dominantly mafic lavas of the lower unit were followed by a voluminous silicic eruption, the Tatara Dacite, and the mafic lavas of the second cycle were succeeded by intermediate to silicic magmas of Volcán San Pedro. Four compositional magma ‘types’ were defined on the basis of lavas with 52·5 wt % SiO2. Types I and II lavas occur in the lower unit (first cycle), and types III and IV are confined to the upper unit (second cycle). Types I and IV were suggested to be parental magmas to the Tatara Dacite and Holocene dacitic lavas, respectively.
The division of Volcán Tatara into lower and upper lavas in this study is in accord with Ferguson et al. (1992), but in addition to the recognition of multiple new magma compositions, the stratigraphic relations among and within correlatable packages (
–ß––
–µ–
–
–––
and 
–) plus associated non-correlative units have been refined (Figs 21–29). We note that for both upper and lower Tatara, stratigraphic relations are far better constrained by correlations among multiple sections for the youngest eruptive products than for earlier units: early lavas are spatially confined because they filled erosional depressions and later units partially covered the surfaces of constructional edifices.
|
|
|
|
|
|
|
|
|
Lower Volcán Tatara
Lower Volcán Tatara is subdivided into five volcanic episodes, each comprising multiple eruptive events, plus two silicic units (early Quebrada Turbia Dacite, late Tatara rhyolite) of somewhat uncertain stratigraphic position and petrologic affinity (Fig. 22). Four dominantly basaltic andesitic volcanic episodes (Alpha, Gamma, Epsilon, Theta) take the name of the package corresponding to the most mafic lavas (51·5–53·5 wt % SiO2) of the episode. The fifth episode comprises the voluminous Tatara Dacite plus minor associated lavas.
Mingled Quebrada Turbia Dacite (QTD)
Mingled silicic lavas containing inclusions of more mafic magma are characteristic of Cordón El Guadal (Feeley & Dungan, 1996) and the correlative lower Placeta San Pedro lavas, and of the late Holocene lava on the west flank of Volcán San Pedro (Singer et al., 1995), but such characteristics are generally absent from other dacitic lavas of the TSPC (e.g. Tatara Dacite; Singer et al., 1995). An exception to this general rule is an isolated remnant of dacitic lava that occupies the divide between Quebrada Turbia and the southeastern rim of Quebrada Castillo (QTD; Fig. 9). It overlies a thick clastic unit (probably moraine) that rests on upper Placeta San Pedro basalts and its 40Ar/39Ar age confirms that it is a Volcán Tatara unit. Although an age early in the history of Volcán Tatara is suggested by the absence of underlying Tatara lavas in a proximal location, the stratigraphic position of this unit is otherwise poorly constrained.
The Quebrada Turbia Dacite (Table 2i, 173; no relation to Quebrada Turbia sequence) is compositionally distinct from all other Volcán Tatara silicic lavas. This dacite is characterized by an overall composition resembling the Muñoz dacite and it has unusually low Y and HREE, which cannot be derived from mixing with the olivine-phyric mafic inclusions, as these contain higher abundances of these elements. The trace element signature of the only analyzed mafic inclusion closely resembles those of basalts from the early Alpha episode, particularly with respect to low Nb and Zr, thereby reinforcing the impression, based on equivocal geologic relations, that this is an early phase of activity.
Alpha
Mafic lavas (51·9–52·5 wt % SiO2), mainly confined to paleovalleys, are dispersed among sections at or near the base of Volcán Tatara. Stratigraphic relations do not constrain the relative ages of the four varieties of named Alpha basaltic compositions (0–1–2–3). The inferred order shown in Fig. 22 follows from the highly fallible geometric argument that basal-proximal lavas are likely to be older than basal-distal lavas. Thus, this case does not rigorously conform to the definition of package used elsewhere. These flows are grouped in this manner because even though the Alpha packages display some diversity among basaltic compositions, they are markedly distinct from overlying units of comparable SiO2 (e.g. 1–2–3) by virtue of lower or markedly lower concentrations of incompatible elements (e.g. TiO2, P2O5, K2O, Nb, Zr, Rb, Ba, and Y; Fig. 23) and low Nb/Y (Fig. 22).
Unit 0 is an intracanyon flow tentatively assigned to lower Volcán Tatara that was sampled at 10 localities for chemistry and paleomagnetic measurements (‘causeway flow’—south rim of Laguna Azul; Fig. 5). It is a heterogeneous olivine–plagioclase evolved basalt (51·9–52·2 wt % SiO2; 5·7–6·0 wt % MgO; Rb 10–17 ppm; Table 2f, 111–113). Two 1 lavas (52·3 wt % SiO2) are present at the bases of proximal sections in Estero San Pedro (ESPW3.2, ESPW4.1 = TAT158): both contain abundant plagioclase phenocrysts in accord with their high Al2O3 contents (>19·5 wt %) and sparse olivine microphenocrysts reflecting their relatively low MgO contents (<4·2 wt %). As 1 lavas (Table 2f, 114 and 115) have low mg-numbers (49) compared with many TSPC lavas of similar SiO2, they apparently represent the differentiation products, dominantly by fractional crystallization (low Rb/Y plus high Na2O/K2O and K/Rb), of a parental magma that was initially low in many incompatible elements. Although 0 and 1 are variably evolved (Fig. 24a), both are characterized by low K2O (0·75–0·85 wt %), Rb (10–17 ppm), Ba (250–310 ppm), Zr (85–100 ppm), Nb (3–3·5 ppm), and Ce (20–27 ppm).
Packages 2 and 3 are evolved basaltic lavas (51·3–52·4 wt % SiO2; MgO 5–6 wt %; Table 2f and g, 117–121) that are transitional in chemistry between the 0–1 flows and the incompatible element-enriched overlying basaltic andesitic units comprising the Gamma and Epsilon episodes. Two non-correlative lavas on the northern flank (2—EMU4.2, EMU4.5) are distinguished petrographically from distal, basal basaltic lavas in lower Estero San Pedro (3—ESPE1.1–3 and ESPE2.1–6) by abundant coarse plagioclase phenocrysts in the former, and their near absence in the olivine-phyric 3 lavas. Compared with 0–1 lavas, the 2–3 flows have higher concentrations of P2O5 and some LILE (K2O 0·9–1 wt %; Rb 20–25 ppm), LREE (Ce 30–40 ppm), and HFSE (Zr 110–120 ppm; Nb 4–4·5 ppm), but Ba and Y are not so markedly enriched. The youngest compositional variant of this mafic magmatic suite is ESPE1.4 (Fig. 21), the most magnesian basalt collected from Volcán Tatara (Table 2g, 122). This lava has characteristics, including low Nb contents comparable with 0–1 lavas, but it is unusual for having lower Sr.
|
Thus, in contrast to the single mafic magma composition (1) identified by Ferguson et al. (1992), sampling in this study suggests a minimum of five distinct basaltic lineages, all with less enriched incompatible element contents than the overlying Gamma lavas. Although the 2–3 lavas might be interpreted as defining a compositional transition toward the more incompatible element-enriched lavas of the Gamma episode, this trend is reversed by ESPE1.4. The Alpha episode appears to have terminated by the eruption of three newly discovered intermediate to silicic magma types (described below), which probably were derived during three separate events from parent magmas with compositional signatures similar to 3 basalts.
The basaltic lavas at the base of the ESPE1 section are overlain by a pair of hybrid andesitic lavas (ESPE1.5–6; 56·6 wt % SiO2; Table 2g, 123) characterized by high mg-number, K/P, Rb/Y, Zr/Y, Rb/Zr and Rb/SiO2, and low K/Rb (Table 2 and Fig. 24) plus strongly resorbed olivine phenocrysts mantled by unusually thick overgrowths of orthopyroxene. These andesites are not likely to be derived directly from the underlying basaltic magma (ESPE1.4) because they have much higher Nb/Y, nor can such a composition be parental to the overlying dacite (ESPE1.7; 65·6 wt % SiO2; Table 2g, 124) because of higher Rb/SiO2 in the andesite.
The stratigraphic position of a second cycle of andesitic flows overlying the dacite in section ESPE1 (ESPE1.8–13; Fig. 21) would be poorly defined, only on the basis of information from Estero San Pedro, because there are no lavas representing the ––µ– interval in the ESPE sections. However, a large valley-filling andesitic lava at the head of Quebrada Turbia (ArrQT.5; Fig. 9), which directly underlies basaltic andesitic lavas of package 3, is indistinguishable from ESPE1.8–13. These lavas are grouped as package ß(Table 2g, 125–127), the earliest example of correlative flows on the north and south flanks of Volcán Tatara. Package ß lavas may have been derived from an 3-like parent (i.e. Nb/Y 0·25), and they are too low in P2O5 and most incompatible trace elements (e.g. Nb, Ce) to be the fractionation products of a -like parental magma.
Gamma
Higher in the eruptive sequence are plagioclase-olivine phyric mafic lavas (52–56 wt % SiO2), with minor augite phenocrysts, characterized by high TiO2, P2O5, Ce, Nb, Zr, Y, and Ba, accompanied by moderately high concentrations of Rb and K2O (Table 2g and 2h, 128–143). These incompatible element-enriched magmas, corresponding to type I basaltic andesitic lavas of Ferguson et al. (1992; e.g. TAT128 = ESPW3.3–10; TAT132 = ESPW3.11–19; Fig. 21), plus previously unrecognized derivative mafic andesitic lavas, represent a major shift in chemistry relative to the Alpha episode. Gamma lavas have among the highest TiO2, P2O5, Zr, Nb, Zr, Ba, and Nb/Y (Figs 22, 23 and 24d) of any lavas in the TSPC (compared at the same SiO2) and are near the maximum for these variables with respect to all analyzed lavas of comparable composition from frontal arc volcanoes of the SVZ (e.g. Tormey et al., 1995). In this regard they tend to approach the compositions of lavas typical of cones located behind the frontal arc (e.g. Lanin and Quetrupillan, Hickey-Vargas et al., 1989; Laguna del Maule, Frey et al., 1984).
The early relatively mafic lavas of this episode comprise several groups of non-correlative flows, that commonly form locally thick multi-flow groups. Despite the absence of definitively correlative packages before the thin andesitic package , some constraints on relative order are known. These units are best represented at proximal section ESPW3, which can be tied in part to the more distal section ESPW2, and to section EMU1 on the north flank of the Cordón Tatara. In addition, mafic basaltic andesitic lavas (52·2–52·9 wt % SiO2) that are very similar chemically and petrographically, but not quite identical, to those in ESPW3.11–19 (1; Table 2g, 131–133) occur to the north on northern Placeta San Pedro (ESPN.12; 3; Table 2g, 134 and 135), and in section QTW12 (QTW12.31–33; 2; Table 2g; 136 and 137). The most mafic lava of this type, which underlies package lavas in section EMU1 (EMU1.1; Table 2g, 139), closely resembles 1 lavas.
Flows ESPW3.3–10 (53·5 wt % SiO2; Table 2g, 128–130) directly underlie the less evolved basaltic andesitic lavas of package 1, which they closely resemble. These lavas contain abundant plagioclase with highly variable zoning patterns and abundant olivine phenocrysts, but there is a modal change from lavas at the base (5·1 wt % MgO) that contain minor phenocrysts or xenocrysts of augite and orthopyroxene to slightly more mafic overlying flows (5·25 wt % MgO) that contain only a trace of pyroxene. ESPW2.3 (Fig. 21) is also more evolved (53·7 wt % SiO2; Table 2g, 140) than any lava in package 1. It contains a much higher proportion of pyroxene (opx and aug) than either package or , pyroxenes are commonly in reaction relationship (pyx–pyx, oliv–pyx), and plagioclase textures are also indicative of disequilibrium, suggesting that magma mixing played a role in its evolution. This sample is intermediate in composition between packages and , but it is also sufficiently distinct from both (chemistry and petrography) to suggest that it is not on the same liquid line of descent as either.
The evolved basaltic andesitic lavas of package (55·5 wt % SiO2; Table 2h, 141–143) include correlative flows from the north (EMU1.2–5) and south (ESPW3.20) flanks of the system. These lavas extend the type I compositional range recognized by Ferguson et al. (1992). They are sparsely porphyritic compared with packages 1–2–3, with lower oliv/aug and lower mg-numbers (47·5 vs 54). Phenocryst textures suggestive of magma mixing are less well developed than in ESPW2.3, and thin reaction rims of pyroxene on some olivine phenocrysts are interpreted as recording the onset of the peritectic reaction during progressive differentiation. Incompatible trace element ratios (e.g. Nb/Y; Fig. 22) are consistent with derivation of package andesites from a 1-like parent magma.
Epsilon
This volumetrically minor episode is inferred to comprise two closely related packages, and µ, of unknown relative age; (1) the basaltic andesitic lavas of package (53–54 wt % SiO2) occur in proximal sections (ESPW3.21 and ESPW2.4–7; Table 2h, 147–148) overlying flows of the Gamma episode, and (2) the newly recognized P2O5-rich, low-MgO andesitic lavas (60 wt % SiO2) of package µ are present only in distal southern sections (ESPE1.14–18, ESPE2.7–10; Table 2h, 144–146), where they are sandwiched between the superficially similar package ß andesites and the overlying package . Thus, the relative order shown for these two units in Figs 21 and 22 is entirely arbitrary, and even the coupling of package µ with package is somewhat speculative given the absence of packages , , , and from the ESPE sections. They are paired because of close similarities in trace element ratios, such as Zr/Y, Ba/Nb, K/Rb, and Nb/Y, for which both and µ are distinct from underlying and overlying units.
Package (plag + oliv ± aug) is equivalent to TAT134 of Ferguson et al. (1992), which they ascribed to hybridization of type I () and type II () magmas. This interpretation is now considered untenable on the basis of the clear stratigraphic progression from Alpha to Gamma volcanism, although it is correct that the trace element signature is intermediate in most respects between these two extremes. Like package ß, the package µ andesites (plag + 2pyx + FeTi-ox + ap) are characterized by low mg-numbers (39·5; e.g. equivalent to Muñoz dacite), low Cr, Ni, and V, and an absence of textural disequilibrium, indicating that fractional crystallization played an important role in their differentiation.
Theta
This episode begins and ends with volumetrically minor andesitic packages and non-correlative flows (Fig. 22), but it is dominated by basaltic andesitic lavas of package (53·8 wt % SiO2; Table 2h, 155–161; not sampled by Ferguson et al., 1992). It is probable that most of the inaccessible southern face of Cordón Tatara (Fig. 6) exposes lavas of this episode, and that most of these are package , which is very thick in proximal sections (ESPW2.9–23, EMU1.7–EMU3.10) but relatively thin in distal sections (ESPE1.19–26, ESPW4.2, QTW12.38).
The thin but widespread sequence of package andesitic lavas (56·5–57 wt % SiO2; Table 2h, 149–154) that underlies package corresponds to type IIA andesites of Ferguson et al. (1992; TAT135–136). Package is marked by low oliv/pyx (olivine occurs only as sparse microphenocrysts) and the appearance of magnetite microphenocrysts, consistent with low vanadium concentrations. A predominant role for fractional crystallization is also indicated by low Cr, Ni, and mg-numbers (40) for lavas of 56·9 wt % SiO2. Package is porphyritic with abundant plagioclase phenocrysts, high oliv/aug, and no opx, and it is distinct from most other TSPC lavas with 53·5–54 wt % SiO2 in having very low mg-numbers (41–45) and correspondingly low Cr and Ni (compare Sin Nombre; Table 2a, 3–6).
Overlying package are a series of progressively evolved andesitic lavas of the same chemical affinity. Three basaltic andesitic lava flows on the northern flank of Cordón Tatara comprise package (QTW12.39, EMU2.11–12; 54·2 wt % SiO2; Table 2i, 162 and 163). An overlying flow (EMU3.13) plus a distal flow in Estero San Pedro (ESPW4.4) are termed package (56·5 wt % SiO2; plag + 2pyx + oxides + minor ol; Table 2i, 164 and 165). Five non-correlative andesitic lava flows (QTW12.40–44; plag + 2pyx + oxides + ap; 58·5 wt % SiO2; Table 2i, 166 and 167) post-date package . These three groups are characterized by major and trace element ratios such as Na2O/K2O, K/P, Ba/Rb, Ba/Y, Ba/Nb, K/Rb, Nb/Y, Zr/Y, Rb/SiO2, and Rb/Zr that are similar to those of package lavas. No other series of mafic to andesitic magmas in the TSPC shows this degree of constancy among such ratios during successive eruptive events, and no other evolved magmas have such high Na2O/K2O in combination with such low mg-numbers at a given wt % SiO2 (Fig. 23a). None the less, in detail, these successive and increasingly evolved magmas appear to be end products of separate but similar differentiation paths rather than eruptions of magma along a single liquid line of descent.
Tatara Dacite and andesitic flows
Ferguson et al. (1992) proposed that the ultimate differentiation product of type IIA magmas () was a plagioclase–hornblende dacitic magma (Tatara Dacite of Singer et al., 1995; 68 wt % SiO2; Table 2i, 170 and 171). The presence of andesitic packages and between mafic package and the Tatara Dacite suggests that the latter may in fact be the culmination of a later differentiation episode, perhaps related to package or possibly to still younger parental magma. Potential candidates for such parental magmas are (1) a plagioclase plus two-pyroxene (no hornblende) andesitic lava (EMU3.14; 64·3 wt % SiO2; Table 2i, 169), from the same stratigraphic interval as the dacite (post-, north wall of Estero Molino on the opposite flank of a ridge that exposes >60 m of typical Tatara Dacite; Ferguson et al., 1992, fig. 5), and (2) distal non-correlative flow ESPE1.27 (62·3 wt % SiO2; Table 2i, 168), which lies at the appropriate stratigraphic position (post-, pre-upper Tatara). In detail, however, the andesitic lavas are unsuitable as parental compositions, as a result of higher concentrations of many incompatible elements (Nb, Zr, Y, Rb, Ba), as well as higher Rb/SiO2 and Rb/Y, and lower K/Rb than the Tatara Dacite, suggesting that the andesitic lavas evolved in a system that was more open to crustal input.
Tatara rhyolite
The only high-SiO2 rhyolite sampled from Volcán Tatara (ESPE3; Table 2i, 172—not shown in Figs 23 or 24a, c and d) is a stratigraphically unconstrained flow on the south flank. This inconspicuous unit is distinct from the Los Lunes rhyolite (Table 2a, 8) in having much higher Y and HREE, in keeping with the generally, but not uniformly, higher abundances of these elements in Volcán Tatara andesitic and dacitic magmas (Quebrada Turbia Dacite is the exception) compared with those of older sequences. The origin of this unit is particularly uncertain because of its poorly defined stratigraphic position and unusually high 87Sr/86Sr.
Lower Volcán Tatara: summary
The large and well-preserved lower Volcán Tatara may record a period of activity lasting 40–60 ky, which is almost certainly longer than the durations represented by remnants of highly reduced edifices discussed previously (e.g. Quebrada Turbia, lower–middle–upper Estero Molino sequences; 
15 ky?). Keeping this caveat in mind, there are a number of comparative points raised by the temporal evolution of this volcano, which in many ways is also very different from the overlying upper Volcán Tatara lavas described below.
None of these older, apparently less long-lived periods of activity exhibits the same high degree of diversity in parent magma compositions. Even the relatively varied middle Estero Molino sequence basalts do not span a range as large as that defined by the shift from 0–1 basalts to the mafic basaltic andesites. Restated, these two adjacent episodes are characterized by mafic magma compositions that span nearly the entire range of incompatible element concentrations and ratios recorded by preserved basaltic lavas during 930 ky of activity. Following the early chemical shift from packages 0–1 to the incompatible element-enriched lavas typified by packages 1–2–3, there is (1) a general, step-wise decline in incompatible element abundances, with the exception of increases in Y–HREE concentrations, and (2) a general tendency for fractional crystallization to play an increasing role during differentiation from basaltic to andesitic magmas.
For the purpose of monitoring post-Alpha temporal changes in lower Tatara parent magma chemistry, the early -like lavas of ESPW3.3–10 can be compared with the later package , and with the still younger package (Table 2g, h, and i, 130, 148, and 161, respectively), as these three basaltic andesitic magmatic suites have essentially identical SiO2 (53·6 ± 0·2 wt %). Except for TiO2 and Na2O, this comparison yields a regular progression as a function of decreasing age: Na2O/K2O, Al2O3, Fe2O3*, MnO, CaO and V increase as MgO, Cr, Ni, and mg-number decrease markedly (e.g. average mg-number 53·4, 49·8, and 42·8, respectively). Among incompatible major and trace elements K2O, P2O5, Nb, Zr, Sr, Pb, Ba, Ce, and Rb decrease (average Rb ppm 31·1, 26·4, and 22·4, respectively), but Y and the HREE increase (average Y ppm 18·6, 20·6, and 22·3, respectively). In detail, these compositional shifts are reflected in increasing Ba/Zr, Ba/Nb, Sr/Rb, Ba/Rb, and K/Rb (343, 377, and 407, respectively) combined with decreasing Sr/Y, Ba/Sr, Ba/Y, Nb/Y, Rb/Y (1·7, 1·3, and 1·0, respectively), and Zr/Y (8·4, 7·4, and 5·9, respectively).
Thus, a preliminary picture emerges in which post-Alpha lower Tatara lavas record three successive episodes, comprising multiple intrusion–differentiation–extrusion events, wherein erupted mafic magmas were characterized by progressive changes in composition (e.g. progressively lower P2O5, HFSE, and LREE; Nb/Y value of 0·48, 0·30, and 0·24, respectively). Each of these three parental magma types apparently gave rise to a spectrum of intermediate magmas before the next episode. Superimposed on changes in mafic magmas is a tendency for a progressively greater role of crystal fractionation (e.g. lower mg-number, Cr, and Ni at a given SiO2; Fig. 22 and Table 2) and a diminishing contribution from assimilation (e.g. lower K and Rb, but higher K/Rb; Fig. 22) with time. The increases in Y–HREE among lavas of comparable SiO2 during this same period may reflect lesser contributions of crustal contaminants with low Y–HREE and/or changes in parental magma composition related to a complex evolution pattern in the mantle source. These observations need to be evaluated in detail, but the perspective gained from a more complete roll-call of the stratigraphic units differs appreciably from that available to Ferguson et al. (1992). In particular, there is very little support for concurrently active magma reservoirs that were fed by contrasting mafic inputs.
The contributions of the new stratigraphic framework to an understanding of lower Volcán Tatara are, at this stage of our investigations, best stated in terms of the questions raised by these data. What factors were responsible for the semi-systematic but decidedly non-linear temporal progression defined by parent magma compositions? Why did fractional crystallization increasingly dominate over open-system differentiation following the Gamma episode? Might this in some way reflect long-term modification of the crustal column beneath the volcano? If this is a general pattern at long-lived continental arc volcanic centers, could sampling of only young eruptive products result in a biased assessment of the degree to which open vs closed system evolution has occurred? What factors are responsible for the preferential appearance of andesitic and dacitic magmas at or near the ends of most episodes?
Upper Volcán Tatara
The lavas of upper Volcán Tatara (103 samples) are essentially basaltic andesitic (52·3–56·5 wt % SiO2). Three internally compositionally variable stratigraphic units are recognized (Figs 25–29): (1) a group of relatively evolved, non-correlative flows (UEP7.1–10; 56·5–54·5 wt % SiO2; Table 2i, 176–179) on the northeast flank; (2) package (53·1–54·0 wt % SiO2; Table 2j, 181–188), which is confined largely to the southwest flank; (3) three closely related packages comprising the ‘Omega eruptive episode’ (1–2–3; 52·3–54·4 wt % SiO2; Table 2j and 2k, 189–210), which are present on both the west and east flanks of the cone, and which directly overlie both UEP7.1–10 and package . The three packages and package correspond respectively to types III and IV of Ferguson et al. (1992). Paleomagnetic data from seven sites in lavas (ESPE1) yield an average inclination (–40°) that is markedly different from the anomalously steep inclinations (average of 13 is –79·3°) of 1 and 2 flows in the EPW6 section (L. L. Brown, personal communication, 2000). As package has not been found in contact with the flows in UEP7.1–10, the relative eruptive order of the two pre- units is unknown (Fig. 25).
|
These three stratigraphic units correspond to three separate magmatic episodes, each with a complex history that merits detailed mineral chemistry studies (as yet not undertaken) integrated with more extensive geochemical modeling than can be elaborated here. None the less, the reconstructed internal stratigraphy of the three packages and a preliminary petrologic model for the Omega eruptive episode are presented below on the basis of whole-rock geochemistry. This discussion is supplemented with additional text, Fig. III-1 and Table III-1 in Electronic Appendix III. First, some basic constraints pertaining to the evolution of the other two suites are briefly outlined.
Package
Package overlaps with 1–2–3 in terms of SiO2, and their compositions are broadly similar for many other elements, but in contrast to the narrow, linear major elements trends defined by 1–2–3 in Fig. 26, package lavas define more scattered distributions and are in part higher in Na2O, Na2O/K2O, TiO2, Fe2O3*, and P2O5 (not shown), and in part lower in MgO and CaO, but higher in Al2O3 at a given SiO2. Although there are a few lavas within package with -like compositional signatures (e.g. Table 2j, 186), the majority of samples from package have high Y, Nb, Zr, Zr/Y, Ba/Sr, and Nb/Y compared with 1–2–3 (Fig. 28). Even though package flows overlap in Sr with package 1, package defines trends in Fig. 29 (Fig. III-1) that are divergent with respect to the packages. In addition, package flows have low mg-numbers and K/P (Fig. 28a and b), whereas the nominally more evolved flows of UEP7.1–10 have relatively high mg-numbers, overlapping with 1–2–3. Package flows are highly porphyritic, containing abundant, commonly equant (<2 mm) plagioclase phenocrysts that display a wide variety of zoning patterns and resorption textures. All the flows contain olivine microphenocrysts but augite is only sporadically present, and it is always far less abundant than olivine. In contrast to the packages and the flows in UEP7.1–10, package lavas do not define systematic variations in magma composition with respect to eruptive order (Fig. 27).
Ferguson et al. (1992) proposed that much of the diversity within the currently defined package was due to alternating eruptions from two separate, long-lived magma reservoirs that were supplied with two contrasting parent magma compositions. On the basis of our larger sample suite and better stratigraphic control, a more correct general statement would be that volcanism during the interval occupied by heterogeneous package reflects input from a diversity of basaltic magma compositions, including a minority that were in some ways similar to the distinctive compositional end-member that would subsequently dominate during the Omega eruptive episode, as well as a diversity of differentiation trends: there are no independent constraints that would enable us to tie specific contrasting magma compositions within package to physically separate conduit–vent systems. The age difference (30–40 ky) between upper Tatara lavas and Holocene Volcán San Pedro annuls the suggestion that San Pedro lavas are direct descendants of type IV () parent magmas, as proposed by Ferguson et al. (1992).
Non-correlative flows
The UEP7.1–10 lavas, which are compositionally distinct from the associated packages (Figs 26–29), are also texturally and mineralogically distinct in that they are characterized by an abundance of large glomeroporphyritic clots (mainly plag + aug) and gabbroic fragments, plus xenocrysts of orthopyroxene. Although these flows are relatively evolved in terms of SiO2, they contain Fo80 olivine phenocrysts with compositions more appropriate to basaltic magmas and they have relatively high mg-numbers, Cr, and Ni. The lavas of UEP7.1–10 define a significant but somewhat irregular up-section trend toward less evolved compositions (Fig. 27). Given the large ranges in some incompatible elements (e.g. Rb 34–52 ppm) in combination with a narrow range of mg-numbers, and evidence for mineralogical disequilibrium in these lavas, this range of compositions probably reflects mixing of basaltic and andesitic magmas in variable proportions.
The compositions of the UEP7.1–10 lavas give the appearance in Harker diagrams that they are extensions of the trends defined by 1–2–3 (Fig. 26, except Sr–SiO2), but this is not the case in terms of magmatic relations. Specifically, no magma composition from any of the packages serves as an appropriate mixing end-member for this array of hybrid magmas. Whereas none of the package lavas is an appropriate end-member in detail because of the systematically low mg-numbers in this unit, many compositional features of the UEP7.1–10 lavas would be consistent with involvement of a mafic mixing end-member with the typical trace element signature of some package magmas (i.e. higher Nb/Y, Ba/Sr, and Zr/Y, but lower Sr, Sr/Y, Sr/Rb, K/Ti, and K/P than 3 lavas; Figs 28 and 29). The completely separate question of whether or not lava compositions from UEP7.1–10 serve well as evolved mixing end-members for the 1–2–3 array is discussed below.
The Omega eruptive episode: insights into magmatic processes on the basis of variations in magma composition in the context of a high-resolution stratigraphic reconstruction
Lavas of the last preserved eruptive episode of Volcán Tatara were erupted from vents on both the west and east flanks. As there is considerable systematic up-section compositional variation, this unit is ideally suited for the reconstruction technique employed in this study. Although the three packages and the compositional variants within them are not distributed symmetrically, it has been possible to (1) identify abrupt compositional shifts between 1–2 and 2–3, and (2) propose plausible internal stratigraphic orders (particularly within 1 and 2) by combining relations from multiple sections. These constraints have been used to formulate a preliminary model for the petrologic evolution of this suite.
These lavas are of some importance to petrologic interpretations of along-arc compositional variations in the SVZ because the essentially aphyric, last-erupted 3 end-member has a number of characteristics that set it apart from all mafic lavas at frontal arc volcanoes south of 36°S. Specifically, it is a magma with very high Sr (850–900 ppm), high Al2O3 (18·9 wt %), and relatively low Rb (12–13 ppm; Rb/Y 0·9; K/Rb 550–600) that lacks large plagioclase phenocrysts. As such, it closely approximates a liquid composition (see Arculus, 2000), and it probably has been only minimally contaminated. The latter inference is supported by the fact that 3 lavas have the lowest 87Sr/86Sr of any in the TSPC (0·70379; J. P. Davidson & M. A. Dungan, unpublished data, 2000). Low concentrations of Cr (35 ppm) and Ni (25 ppm) in combination with high Sr and Al2O3, plus unusually high Sr/Sc (40; Dungan et al., 2000), lead us to conclude that this magma type is the product of crystal fractionation-dominated evolution featuring a mineral assemblage with unusually low plag/(oliv + aug), as a result of either differentiation at high pressure (Grove & Kinzler, 1986) or fractionation of a water-rich magma (Gaetani et al., 1993).
Ferguson et al. (1992) recognized that the range of compositions in the currently defined Omega episode (1–2–3) cannot be the result of closed-system fractionation, and they proposed that this spectrum records invasion of a magma reservoir containing a hybrid andesitic magma by the 3 compositional end-member, mixing of these two magmas, and then progressive flushing of the resulting hybrid magma from the subvolcanic reservoir such that nearly pristine 3 magma was erupted at the end of the episode. Although generally accepting this hypothesis, we recognize a number of complexities on the basis of our larger sample set that require consideration of mixing between the 3 end-member and a diversity of resident components.
Petrography. Unlike the compositionally similar lavas of package (oliv + plag), the three packages contain small phenocrysts (<0·5 mm) and microphenocrysts of plagioclase, olivine, and augite. Both 1 and 2 also contain abundant larger xenocrysts of plagioclase, commonly with resorption textures, plus augite–olivine intergrowths characterized by many fine sub-round olivine inclusions in a larger augite grain or in multi-grain aggregates of augite. These resemble the xenocrystic clinopyroxene-rich clots in some upper Placeta San Pedro lavas that were discussed previously, and they may be recycled cognate phases ripped up from deeper zones of crystal accumulation. The higher, but irregular Cr and Ni contents of the nominally more evolved 1 and 2 lavas relative to 3 lavas probably reflect in part the presence of these crystal clots. We emphasize that lavas of package 3 are the least porphyritic and least texturally complex of the Omega episode, being distinguished from the vast majority of comparably mafic TSPC lavas by the presence of small euhedral phenocrysts of plagioclase, olivine, and augite, without xenocrysts, in an unusually plagioclase-rich groundmass. Perhaps because of different thermal histories, 2 lavas are characterized by fine-grained augite-rich groundmass textures that are markedly different from those of the 3 lavas.
Chemical stratigraphy. Subdivisions of the last preserved eruptive products of Volcán Tatara into three related packages have been identified on the basis of stratigraphically coherent chemical shifts from one package to the next (Fig. 27) that are accompanied by the mineralogical and textural changes described above. The internal sequencing of samples within 1 (Table 2j, 189–192 and 195) and 2 (Table 2j and k, 196–204) has been achieved by intercalating lavas from multiple sections on the basis of compositional similarities without violating the eruptive orders defined by individual sampled sections. The 1–2 boundary is defined in EPW5–EPW6, particularly by Sr concentration and related ratios, whereas the more subtle 2–3 boundary is present in EPW5 as well as sections on the western flank (EMU3 and TR).
Compositional variability within package 1 is subtle but complex. The flows that most resemble the overlying package 2 lavas occur near the top of package 1 in sections EPW5–EPW6, and these are characterized by relatively high concentrations of Sr (680–692 ppm) in combination with low, but not the lowest K2O, Rb, and Y. The lowest Sr concentrations (618–632 ppm) are in the middle of the package, and values intermediate between these extremes are typical of the basal flows (630–646 ppm). This bow-shaped pattern is mirrored by variations in CaO, SiO2, mg-number, Rb, and Y (Fig. 27).
Three flows with somewhat anomalous compositions at the base of the western TR section (Table 2j, 192–194) underlie 2 lavas and they have some compositional characteristics consistent with a stratigraphic position at the base of 2 (e.g. SiO2, CaO, mg-number). None the less, for many compositional variables TR.2 closely resembles 1 lavas on the eastern flank, and the two bracketing lavas have trace element signatures, particularly for Sr and trace element ratios involving Sr, that are intermediate between 1 and 2 (e.g. 765 ppm Sr in TR.1 vs 652 ± 35 ppm in 1 and 840 ± 15 ppm in 2); that is, these three flows define an alternation between 1 chemistry and magmas with compositions with some affinities to 2. Thus, despite the >100 ppm gap in Sr concentrations that separates packages 1 and 2, there are a few erupted lavas with transitional compositions. TR.1–3 are identified separately in Figs 26–29 (Fig. III-1 in Electronic Appendix III) but are not accorded a specific stratigraphic name.
The boundary between 2 and 3 is subtle for some elements, but it is petrologically meaningful. The latter is distinguished chemically from 2 by step-function changes in certain major element (Na2O/K2O) and trace element ratios (Sr/Y, Ba/Sr, Ba/Rb, Sr/Rb, K/Rb, and Rb/Y). Internal variations within 2 generally correspond to the overall up-section trends. Package 3, the essentially xenocryst-free and homogeneous compositional end-member of the 1–2–3 chemical progression, is characterized by high Al2O3, CaO, and Sr, in combination with low SiO2, K2O, Nb, Zr, Rb, Y, Ba, and Ce (Table 2k, 205–210).
Petrologic evolution. Unusual aspects of compositional variations among the three packages are the following: (1) TiO2 (0·97–1·00 wt %), Fe2O3* (8·2–8·5 wt %), and P2O5 (0·19–0·22 wt %) are nearly constant throughout a compositional progression in which K2O (0·9–1·2 wt %) and many incompatible trace elements vary systematically with stratigraphic position (e.g. Ba 360–245 ppm, Rb 33–13 ppm, Y 20·5–14 ppm); (2) mg-number generally increases up-section from 1 to 3 (inverse correlation with SiO2) but Cr concentrations display the reverse trend; (3) despite generally linear co-variations among other elements, Sr is characterized by a large step-function shift from 1 to 2–3. This combination of systematic variations is incompatible with a fractionation-dominated evolution trend, and the nominally more evolved lavas occur at the base of the section, not the top. Some of the compositional variability is due to differences in modal abundances of xenocrysts, but this is thought to be important in terms of the high and scattered concentrations of compatible trace elements such as Cr, Sc, and Ni, and small shifts in a few major elements and major element ratios such as MgO, Al2O3, and CaO/Al2O3 (Fig. III-1 in Electronic Appendix III), not as a major factor in the first-order up-section trends; that is, the xenocryst-free 3 lavas have the highest mg-numbers, CaO, and Al2O3. Consequently, internal compositional variations have been modeled as mixing relations wherein the average composition of 3 serves as the mafic, or replenished, end-member.
On the basis of the graphical relations in Fig. 26, one or more compositions in the UEP7.1–10 flows would seem to be candidates for the evolved mixing end-member(s). Although a relatively mafic lava from this group (UEP7.9) serves well as the evolved end-member for reproducing most compositional characteristics of 1, the UEP7.1–10 flows are eliminated in detail as mixing end-members because they have lower concentrations of Y and HREE than 1 lavas. Lacking a concrete example that suffices for all compositional variables, we can do no better than to suggest that the true mixing component was very similar to the most mafic magmas of the UEP7.1–10 spectrum, but that it was not sampled, if in fact it was erupted. As shown in Fig. 29 (orange curves), mixing proportions between 60% and 80% of the UEP7.9 end-member reproduce reasonably well the compositional spectrum of 1 lavas (Table III-1, Electronic Appendix III). The same mixing end-members in different proportions account for the 2 lavas characterized by relatively low Sr concentrations (810–825 ppm).
There are also some high-Sr lavas (840–860 ppm) at the base of package 2 on the western flank (e.g. EMU3.15–20 + TR.4–5), which are characterized by relatively low Sr/Rb and K/Rb because of anomalously high Rb. For most variables, these samples fall well off mixing curves involving end-members from UEP7.1–10, and they require mixing of 3–4% of an extremely evolved granitic component (black curve, Fig. 29, and Fig. III-1 in Electronic Appendix III) with exceptionally high Rb and SiO2 to reproduce this branch of the 2 compositions. The remainder of the 2 samples falls between these limiting cases. Mixing curves utilizing even slightly less Rb-rich end-member components fall much closer to the red curve involving the evolved UEP7.4 end-member, suggesting either assimilation of highly evolved granitic wall-rocks, or perhaps that there were pockets of partial melt derived from already evolved granitic compositions residing within the subvolcanic conduit–reservoir system during eruption of upper Volcán Tatara lavas, and that small quantities of such crustal melts were incorporated into mafic magmas. In this regard, we favor the suggestion of Grove et al. (1988) that assimilation may commonly be accompanied by decoupling of heat and mass transfer.
Summary remarks. An inescapable conclusion of this analysis is that multiple mixing components of diverse origins are required, in varying proportions, to account for the compositional variability exhibited by 1–2–3. The ‘anomalous’ compositions of TR.1–3 (high K/Rb but low K/P) bear no consistent relation to the mixing curves shown in Fig. 29. However these magmas formed, they cannot be explained entirely in terms of the end-members discussed above. This implies that more components than have been discussed here were involved. Thus, a dynamic model in which andesitic magma residing in a single reservoir was hybridized in varying proportions with a replenished mafic magma (Ferguson et al., 1992) is too simplified, particularly in light of the poorly defined source and entrainment mechanism responsible for incorporation of the augite–olivine intergrowths.
Distinctions among upper Tatara lavas have been emphasized for the purpose of developing a high-resolution stratigraphic reconstruction, but the total variability during this phase of activity is relatively limited. The range of implied parent magma compositions is narrow in comparison with those of lower Volcán Tatara or the Estero Molino sequence, and there are no sampled lavas with >56·5 wt % SiO2. Only the upper Placeta San Pedro (UPSPS) and middle Estero Molino sequences are compositionally more restricted. The UPSPS basaltic lavas are comparable with the basaltic andesitic Omega packages in the following respects: (1) both constitute compositionally variable assemblages of mafic lavas which are, in the majority, complex multi-component hybrids; (2) both are recharge-dominated open systems; (3) both are characterized by broadly similar high-Sr parental magmas with generally low abundances of most incompatible elements; (4) parental compositions are constrained on the basis of a small minority of flows that may be weakly contaminated or hybridized; (5) both contain olivine and augite xenocrysts (in come cases intergrown) that may represent recycling of cognate cumulates, suggesting that the parental basalts may have been water-rich magmas. Major distinctions between these two units are that (1) the internal chemical stratigraphy of the UPSPS is largely non-systematic in contrast to the regular step-wise variations defined by packages 1–2–3, and (2) the sources of mixing components rich in K and Rb appear to be rather different.
|
ASSESSMENT OF THE RECONSTRUCTION METHOD |
|---|
Composite stratigraphic reconstructions of parts of the TSPC have been rendered feasible by the presence of widespread chemically distinctive packages of lavas that are recognized in multiple sampled sections. High-density, flow-by-flow sampling at multiple localities is essential to the viability of such reconstructions because no two sampled sections are likely to be characterized by precisely the same succession of lavas: major ‘differences of omission’ are observed between sections located 1–2 km apart on walls of the same present-day valley. For example, the lower Estero Molino sequence is well represented in upper Quebrada Turbia but is hardly present elsewhere, and the middle Estero Molino sequence and Laguna Azul lavas of the upper Estero Molino sequence are also thicker and more continuous in Quebrada Turbia than to the west, but the intervening Laguna Verde lavas are present almost exclusively in the most westerly (distal) sampled sections (Fig. 13). Comprehensive sampling restricted to any single drainage would have failed to identify a significant fraction of the eruptive events of the Estero Molino sequence.
The absence of conflict between superposition relations in individual sections and interpretive correlations constructed on the basis of petrography and chemical equivalence is an indication of the robustness of the approach used here, but a certain fragility is also suggested by the occasional dependence of a reconstruction on the occurrence of one or two lava flows within a particular sampled section, which in turn can be tied into a more complete succession on the basis of chemical criteria (e.g. EML5–ESPN.1 at the top of the lower Estero Molino sequence; EML.6–QTW12.24–LV.1–3 at the base of the middle Estero Molino sequence). Although the relative ages of all the analyzed TSPC lavas are not known with absolute certainty, the reconstruction method employed in this study has enabled us to arrange the vast majority of sampled units in their correct relative order, and to identify clearly where uncertainties exist.
We acknowledge that it is highly unlikely that all the edifices of the TSPC were constructed entirely by eruptions from central vents. In cases where the chemistry of apparently volumetrically minor units intercalated with more voluminous units renders them petrologically unrelated to underlying and overlying lavas (e.g. QTW10.14–package Q, upper Estero Molino sequence; Fig. 18) it is reasonable to suspect that such magmas evolved in separate conduit–vent systems, or even that they are the products of different mantle sources. However, we stress that there remains no independent evidence that would directly link such ‘anomalous’ lavas to satellitic vents.
Stratigraphic relations revealed by closely spaced sample traverses in canyons cutting the flanks of Volcán Tatara, the least degraded edifice that we have sampled in detail, conform to the expected relationship wherein the oldest flows dominate proximal sections and younger units aggrade outward. The geometries and distributions of individual units diverge, equally predictably, from conical-layer configurations. Most units have eccentric distributions relative to underlying and overlying units, they display lateral thickness variations over short distances, and few appear to have blanketed an entire edifice. None the less, mafic lavas of the three Omega packages of upper Volcán Tatara were voluminous and laterally extensive. Collecting from vent-proximal as well as more distal flank exposures (including dikes) is, in principle, optimal for obtaining a highly representative sampling of a volcano’s eruptive products. In practice, the pyroclastic–intrusive vent facies of Volcán Tatara is largely inaccessible, and alteration and reworking of scoria deposits presumably would have posed problems for chemical stratigraphy. Collecting from valleys on Volcán Tatara’s flanks was the most practical approach, but we stress that some eruptive events surely were not sampled.
Although data coverage as nearly comprehensive as our geochemical–petrographic sampling is a practical impossibility with regard to paleomagnetic and geochronologic data, the identification of sequences and the stratigraphic reconstructions described here could not have been fully achieved without input from these methods. For example, an incorrect assignment of EML.5–ESPN.1 to the Quebrada Turbia sequence on the basis of chemical criteria, rather than to the Estero Molino sequence, would have been logical without bracketing ages and the observation of normal magnetic polarity in underlying lavas. This interpretation would have led in turn to an erroneous placement of EML.0–4. Conversely, a composite stratigraphic section of the Quebrada Turbia sequence, in which the lavas of QTW10.7–11 are correlated with chemically indistinguishable flows from QTW11.1–21, was used to obtain a record of the Matuyama–Bruhnes geomagnetic transition (Brown et al., 1994) and as the basis for dating the duration of the reversal transition (Singer & Pringle, 1996). Although sampling for paleomagnetic studies of critically important reversal transitions is commonly undertaken without establishing in detail the stratigraphic and volcanologic context of the units that record the reversal, the geochemistry-based approach outlined in this paper could be applied to the expansion and refinement of such records in single valleys or between valleys.
The large numbers of flows in some packages are a reflection of accumulation within paleovalleys. A corollary observation is that some modern valleys on the flanks of the TSPC are in part coincident with paleovalleys carved by preceding episodes of late Pleistocene glaciation. Five incidences of paleovalley-filling flows or pyroclastic deposits that now occur as thin remnant veneers on the walls of re-excavated valleys (e.g. lower Tatara flows in upper Quebrada Turbia; Fig. 3) suggest that present-day, dominantly north–south-trending valleys are at least second-generation features. Hildreth et al. (1984) documented a similar partial preservation pattern for regional ignimbrites of the nearby Calabozos caldera. Adding to the geometric and stratigraphic complexity at the TSPC is the fact that the trends of most paleovalleys on the northern flank of the complex (erosion corresponding to oxygen isotope stages 20–16–14–12) had azimuths (east–west or SE–NW; Figs 3, 5 and 9) that were at high angles to most modern drainages (north–south). The two modern valleys of Quebrada Castillo and Estero Molino also coincide with paleovalleys that controlled the distributions of flows of the lower and middle Estero Molino sequence on Placeta San Pedro; that is, the modern drainages are superimposed on paleotopographic lows generated by erosion that pre-dated 620 ka. Repeated valley excavation into the volcanic fill of pre-existing erosional depressions is probably a general pattern of landscape evolution in volcanic terrains (e.g. Hildreth et al., 1984; Lipman et al., 1996) that may be anticipated in attempts to establish the eruptive histories of multi-edifice volcanic complexes, particularly those in glaciated regions.
Given the extensive impact of erosion at the TSPC, it is no longer possible to assess how many eruptive events were not sampled because they are no longer present, but many units must have been lost. Silicic eruptive products extruded as summit domes, with or without pyroclastic aprons, or entirely as pyroclastic deposits, may be under-represented in our sample set because (1) summit areas appear to have been the most severely degraded parts of older edifices, and (2) pyroclastic deposits on volcano flanks have a much lower preservation potential than lavas (Hackett & Houghton, 1989). We made no attempt to sample units comprising reworked pyroclastic deposits. The main central edifices for pre-Volcán Pellado phases of activity have been degraded down to their intrusive roots and/or buried by younger units, and all these units are bounded by upper erosion surfaces, indicating that early units are preferentially preserved relative to later eruptive products of all but the youngest edifices. None the less, even the most recent edifices have been degraded; e.g. Volcán Tatara was beheaded (perhaps in part by sector collapse; Fig. 4) before the growth of Holocene Volcán San Pedro, and the eastern flank of Volcán San Pedro has been removed by late Holocene sector collapse.
|
IMPLICATIONS FOR PETROLOGIC MODELING STUDIES |
|---|
The lavas of the TSPC offer numerous opportunities for misinterpretation of stratigraphic and parent–daughter relations on the basis of apparently logical groupings of lavas with common geochemical characteristics, which are none the less not of the same eruptive episode, or even of the same sequence. Although petrologic modeling frequently requires the use of fictive magmatic components (i.e. ‘best fit’ parent magma or assimilant that cannot be linked independently to the magmas in question), there is a distinction between identifying them as such and assessing the uncertainties that come with this method versus constructing models of non-existent evolution trends among unrelated lavas that have been grouped erroneously because they display compositional similarities or geographic proximity. Conversely, a well-established eruptive stratigraphy may provide important constraints on the petrologic processes involved in magma evolution—simply on the basis of the specific order in which less and more evolved magmas have been erupted. Observations supporting this concept are discussed below.
Implications of intra-package compositional variations
Considerable progress has been made in understanding the origin of macroscopically mingled lavas in calc-alkaline systems (e.g. Eichelberger, 1980; Thompson & Dungan, 1985; Bacon, 1986; Grove et al., 1988; Singer et al., 1995; Feeley & Dungan, 1996; Clynne, 1999; Tepley et al., 1999), in part because they offer an obvious means of sampling the compositionally, thermally, and texturally distinct magmatic components implicated in the formation of these heterogeneous eruptive products. As mixing end-members become increasingly mafic, or increasingly similar, the probability of developing macroscopically visible heterogeneities decreases as the mechanical barriers to mixing diminish (Sparks & Marshall, 1986), and the motivation for obtaining multiple samples from single lavas becomes less obvious. Although the eruptive products of some volcanoes provide evidence for subvolcanic chambers in which homogenization of contrasting magmas appears to be the rule (e.g. Rhodes, 1983, 1988; Allan & Simkin, 2000), the results of this study demonstrate that it cannot be assumed a priori that a few samples from a package of macroscopically similar flows will convey all the compositional information recorded in the products of the corresponding eruptive event. High-density, flow-by-flow sampling at the TSPC shows that some individual mafic flows and many short-term eruptive events dominated by basaltic or basaltic andesitic magmas record significant internal variability in the absence of macroscopic evidence for heterogeneity, and similar petrologically useful internal variations have been recognized elsewhere (e.g. Dungan et al., 1986, p. 6013; Baker et al., 1991; Donnelly-Nolan et al., 1991).
Important constraints on petrologic processes may be obtained from high-density sampling of individual lavas or closely related suites of lavas. Such short-term diversity may record: (1) interactions between ascending mafic magmas and conduit wall-rocks; (2) the development of pre-eruptive heterogeneity or zoning following emplacement into shallow magma conduit–reservoir systems; (3) heterogeneities introduced by mixing events that trigger eruptions; or (4) some combination of the above. As increasing differentiation during an eruptive event at the TSPC is less common than the inverse, it is more likely that such variations record conduit–reservoir systems that were zoned downward to more mafic compositions, or that they reflect mafic magma replenishment close in time to eruption. The most robust example of the latter is the Omega episode of upper Volcán Tatara, which appears to record the ingress of the late-erupted 3 magmatic end-member into a conduit–reservoir system wherein the resident magmas were heterogeneous. Evidence for an analogous multi-component mixing scenario was obtained by high-density sampling of the monogenetic Burnt Lava by Grove et al. (1988).
Implications of intra-sequence compositional variations
Well-sampled volcanic sequences comprising multiple packages rarely define progressive evolution trends, although some packages may be grouped into eruptive episodes during which variably evolved magmas of apparently similar parentage were generated. Distinctions among different eruptive episodes within sequences in part reflect variability among parental magma compositions, but changes in magma supply rate and reorganization of subvolcanic conduit–reservoir systems between episodes may have resulted in changes in differentiation processes and/or differences in mixing or contamination components. In fact, even where adjacent packages or flows record similar parent magmas and differentiation histories there is usually evidence to suggest that such closely related packages represent magma evolution paths distinct from preceding and subsequent eruptive events; that is, eruptive episodes typically record several repetitions of similar magmatic histories rather than the continuing evolution of a single magma batch across the time represented by several eruptive events.
Problems related to a fragmentary eruptive record
As some, or even many intrusive events at large volcanic centers are not accompanied by eruptions, the virtually unattainable documentation of the complete eruptive record of a large prehistoric volcano would still not be perfectly representative of the total magmatic history (e.g. Dzurisin et al., 1984). An apparent lack of coherence during a reconstructed stratigraphic interval, such as among the Laguna Verde lavas of the upper Estero Molino sequence, may reflect the following: (1) rapid evolutionary changes in the magmatic system accurately represented by eruptive products; (2) low eruption frequency relative to differentiation and/or mafic magma recharge rates, leading to successions of lavas in which mutual petrogenetic relations are highly obscure; (3) an unrepresentative local record caused by geometric factors; or (4) some combination of these. With respect to the eruptive events preserved in a particular sampled section or sections of a volcanic center [point (3) above], it is probable that (3a) there will have been some eruptive products that were emplaced uniquely in other quadrants of the edifice, and (3b) some units emplaced on the same quadrant will not be represented at distal flank localities because they remained confined to more proximal positions. The net result of the uncertainties that are generated by these factors is that high-density sampling is the only means of identifying eruptive events and episodes that might be sufficiently complete to warrant modeling studies.
Summary remarks
A typical TSPC mafic to intermediate magma is the product of mixing, dynamically linked to eruption, of variably evolved and contaminated resident magma and/or complementary crystal-mush with a newly arrived, variably evolved mafic input, which may also have been variably contaminated. The temporal patterns that characterize most reconstructed intervals of the TSPC are not consistent with progressive evolution of magmas in large, long-lived reservoirs. The reconstructed patterns at the TSPC broadly correspond, however, to historic eruptive behavior and temporal variations in magma chemistry at Ruapehu and Ngauruhoe volcanoes (Taupo Volcanic Zone, New Zealand), which have been interpreted as evidence for magma evolution in conduit-dominated systems, specifically ‘a plexus of dikes and sills’, wherein intrusive and eruptive events are closely linked (Gamble et al., 1999; Hobden et al., 2000). A particularly well-documented example of this general type of behavior in another tectonic setting is the earliest phase of the continuing Puu Oo eruption of Kilauea Volcano (Garcia et al., 1989). In both examples it has been inferred that hybrid magmas were formed and erupted on the time-scale of days.
We stress that quantitative constraints on (1) the depth, volume, and geometry of subvolcanic conduit–reservoirs, and (2) the timing of intrusive events relative to differentiation, mixing, and eruption may be obtained only from a combination of petrologic studies and geophysical monitoring of active volcanic systems. Interpretations of pre-eruptive magma dynamics in ancient systems are commonly dependent on analogies with well-constrained modern examples. None the less, it is worth noting the probability that some variations in whole-rock magma chemistry, and much of the complexity in phenocryst zoning patterns and textures that are characteristic of continental arc magmas, are due to a combination of polybaric differentiation and magma–magma interactions during ascent through 25–60 km of crust. We particularly adhere to the description by Hildreth & Moorbath (1988) of crustal-scale subvolcanic feeders as ‘magma percolation columns’ wherein new inputs from depth preferentially follow previously used conduits, leading inevitably to the formation of hybrid magmas. The high surface to volume ratios of dike-like feeders, compared with the equant chambers that feature so prominently in much petrologic thinking, lead intrinsically to relatively rapid cooling, hence efficient differentiation. Moreover, dynamic interactions between ascending magmas and their wall-rocks may favor assimilation relative to large shallow chambers (e.g. Huppert & Sparks, 1985; Philpotts & Asher, 1993). Although our investigations of ‘short-term’ temporal variations among mafic to intermediate magmas at the TSPC do not rigorously constrain the nature of the associated subvolcanic conduit–reservoir systems or the depth(s) of magma evolution, they mitigate against the routine establishment of large magma reservoirs.
Petrologic studies of ancient, or even young prehistoric volcanic activity will not benefit in the foreseeable future from the temporal resolution that may be developed by the quasi-continuous observation and sampling that is possible in the case of contemporary eruptions. The value of such investigations was established by results obtained during the compositionally zoned Parícutin eruption (Wilcox, 1954; McBirney et al., 1987), and a host of important insights have been derived from more recent monitoring of erupting systems in diverse settings [e.g. Garcia et al. (1989, 1992, 1996) and Pallister et al. (1992), to cite but a few examples]. None the less, as emphasized by O’Hara (1998), a comprehensive investigation of the surface exposures of an active volcano could limit the resulting sample suite to a few percent of the edifice’s volume and temporal evolution. In light of the growing body of isotopic and trace element data demonstrating (1) changes in mantle sources for active oceanic volcanoes on the time-scale of years to decades (e.g. Rhodes & Hart, 1995; Pietruszka & Garcia, 1999), (2) invasions of one volcano’s conduit system by magmas bearing the distinctive signature of an adjacent center (e.g. Rhodes et al., 1989; Geist et al., 1999), and (3) short-term temporal changes in the degree of magma–crust interaction (e.g. Garcia et al., 1998; Francalanci et al., 1999), it is evident that progress in understanding the dynamics of magma genesis and evolution during the construction of large volcanoes requires judicious combinations of investigations targeting different time-scales of activity with high-density, stratigraphically controlled sampling.
|
MAGMATIC PROCESS RATES |
|---|
This paper focuses on sequences of the TSPC that are dominated volumetrically by mafic to intermediate-composition eruptive products. Most of these units are composed mainly of variably evolved hybrid lavas that seemingly record coherent differentiation trends on major-element variation diagrams, but that are not regular temporal progressions in the sense that eruptive orders are closer to random than they are to successions of increasingly evolved magmas. If we take the Laguna Azul lavas of the upper Estero Molino sequence (Figs 15–18) as an example of this kind of magmatic system, we may argue that just where a batch of magma falls along the aggregate pseudo-trend at the time of eruption depends on the differentiation rate versus the volume and timing of replenishment events, plus the dynamics of hybrid magma generation. If the magma differentiation rate is high compared with the supply rate of new magma batches with less evolved compositions, the average hybrid magma will be shifted toward a more evolved composition, and the evolved mixing end-members may be more silicic, relative to the contrasting case wherein the magma supply and ascent rates are high.
Virtually all of the groups of lavas that appear to be petrologically mutually related were erupted during sufficiently short periods of time that the absolute age differences among successive eruptive events are not currently resolvable by 40Ar/39Ar geochronology (see Singer & Pringle, 1996). Thus, although discussion of absolute rates of magmatic processes is out of the question for the TSPC, some qualitative insights into relative differences are accessible by virtue of the large data populations available for multiple volcanic sequences. Lavas of different edifices are compared on the basis of variations among the following: (1) the composition of the most primitive erupted basaltic magma; (2) the compositions of eruptive modes (i.e. volumetrically dominant eruptive products); (3) the average composition of each sequence. These relations are summarized in Table 3 (see Figs 10, 15, 19, 23 and 26 for supporting information) for several sequences (subdivided in the case of the Estero Molino sequence and Volcán Tatara on the basis of internal erosion surfaces and geochronology), which are primarily basaltic or basaltic andesitic. This highly simplified résumé utilizes the entire analytical dataset for each stratigraphic unit. Thus, even though they may be incomplete with respect to original volume relations, they are representative of the preserved portions of these edifices. Whereas SiO2 wt % is only a general indicator of magmatic differentiation given (1) the occurrence of fractionation-dominated evolved magmas as well as those shaped by open-system processes (i.e. tendency for higher SiO2 in the latter), and (2) the added complexity of phenocryst-accumulative lavas, there are correlations between the composition of the least evolved magma in each stratigraphic unit and variables that broadly indicate whether the prevalent erupted magmas were more or less differentiated.
|
The only two dominantly basaltic units (middle Estero Molino and upper Placeta San Pedro) are also the only two lacking andesitic or more evolved compositions. The high proportion of primitive and weakly contaminated but compositionally diverse basalts in the middle Estero Molino sequence is a unique case, perhaps with respect to the Quaternary Andean frontal arc, which may reflect unusually rapid egress of magmas to the surface, perhaps as a result of an extensional stress regime. The potentially voluminous basaltic lavas of the upper Placeta San Pedro sequence are more evolved than the most primitive from the middle Estero Molino sequence, but the inferred parental composition is not strongly contaminated. If the upper Placeta San Pedro basalts are a minor flank remnant of an originally much larger, symmetrical, but short-lived edifice, and the tentatively identified parental composition was as monotonous as it appears to have been (all speculative but plausible), a high mafic magma supply rate relative to other TSPC edifices is indicated.
For the remaining units, subtle increases in minimum wt % SiO2 generally correlate with increasing silica contents of the dominant eruptive modes and average compositions, and none has yielded a sample of primitive or uncontaminated basalt. The lower Estero Molino and Quebrada Turbia sequences preserve only a few lavas with <53 wt % SiO2 and these two units have the highest average SiO2 and the most evolved eruptive modes. True basalts (<52 wt % SiO2) are volumetrically minor components of Volcán Tatara and the upper Estero Molino sequence, the most primitive magmas are less evolved than those of the Quebrada Turbia or lower Estero Molino sequences, and the eruptive modes are also less evolved despite the presence of diverse silicic magmas in lower Volcán Tatara.
Thus, for dominantly mafic to intermediate volcanic centers of the TSPC the probability of a primitive basaltic magma reaching the surface was apparently inversely correlated with the proportion of increasingly evolved magma erupted from the system; that is, as eruptive modes shift to higher SiO2, erupted basaltic magmas become proportionally less abundant and the most primitive erupted magmas become more evolved. This relation may be rationalized in part as being due to the presence of increasingly evolved residual magmas with increasingly higher viscosities and lower densities in conduit systems, which would generally act as increasingly efficient impediments to the ascent of dense primitive magma. However, the fundamental relation must be the balance between the rates of magma supply and ascent versus the rate of differentiation, which appears to have varied considerably from sequence to sequence. This general model can be tested and refined by integrating results from detailed textural and mineral chemical investigations of constituent lavas with temporal whole-rock compositional variations so as to recover information concerning differentiation mechanisms and the compositions of end-members involved in hybrid magma formation. In the petrologically simpler realm of oceanic island volcanoes, a number of workers have integrated information from temporal variations in chemistry and mineralogy of evolved magmas, their associated vent and conduit geometries, and lithospheric structure to address the question of how depth of magma differentiation and mantle source region processes might be related to magma supply rate (e.g. Geist et al., 1998; Allan & Simkin, 2000).
In addition to the distinctions among dominantly mafic edifices, the two sequences characterized by high proportions of intermediate to silicic magma are markedly different from each other in the sense that the long-lived Cordón El Guadal–lower Placeta San Pedro magmatic system was dominated by repeated large eruptions of heterogeneous mingled dacitic to andesitic lavas (Feeley & Dungan, 1996), and the Muñoz sequence features two even larger eruptive episodes of dacite (930 ka) and high-SiO2 rhyolite (825 ka) that lack evidence of mafic–silicic magma interaction. The origin of these distinctions is not treated further in this section, but they are noted to emphasize the remarkable diversity among the various sequences of the same volcanic complex. The occurrence of such a large range of different styles of volcanism at the same long-lived volcanic complex raises the question of which factors are responsible for shifting the dominant eruptive modes at arc volcanoes from the prevalent case of mafic to intermediate lavas to dominantly silicic activity.
|
SOURCES OF SILICIC MAGMAS |
|---|
A variety of origins for dacitic to rhyolitic magmas are indicated by the diversity of magma compositions and their associations with less evolved magmas. The close correspondence between the trace element signatures of most dacitic magmas in the lower Estero Molino sequence (QTW12.ASH, QTW12.20–21) and lower Volcán Tatara (ESPE1.7, Tatara Dacite) and their stratigraphically associated mafic and intermediate lavas provides evidence that these silicic magmas are differentiation products of associated less evolved magmas, albeit with variable contributions from open-system processes. Important exceptions to this general observation are the high-silica Tatara Rhyolite (exceptionally high 87Sr/86Sr) and the mingled Quebrada Turbia Dacite of Volcán Tatara, which are characterized by large enrichments of LILE relative to Y–HREE. Similar trace element signatures and high 87Sr/86Sr are found in the Muñoz Dacite (
68·5 wt % SiO2) and high-silica Los Lunes Rhyolite (74·5 wt % SiO2) of the Muñoz sequence, which is unique in the TSPC in being composed almost exclusively of silicic magma. These two large effusive eruptions, separated by 100 ky, lack evidence for direct mafic–silicic magma interactions. As these magmas share chemical characteristics with basement granitoids, our working hypothesis is that they are either the products of partial melting of the deep or middle crust under conditions similar to those that produced the granites, or products of shallow melting of Tertiary granitoids. The long-lived Guadal–lower Placeta San Pedro volcanic center (510–350 ka) produced multiple voluminous, heterogeneously mingled units (dacite–andesite or dacite–basaltic andesite), indicating that shallow silicic magma chambers were established repeatedly and then intruded by new inputs of mafic magma. Feeley et al. (1998) suggested that the origin of the volumetrically dominant dacitic component in this case is remelting of Quaternary gabbroic intrusions. |
IMPLICATIONS FOR INTERPRETATIONS OF ALONG-STRIKE REGIONAL COMPOSITIONAL PATTERNS |
|---|
The SVZ literature includes multiple papers focused on along-arc variations in chemical and isotopic composition as functions of factors such as crustal thickness and age, characteristics of the slab and other subducted components, and mantle structure, which have been inferred (and disputed) to vary progressively along the arc. Such studies have been extremely useful for defining the regional framework and calling attention to outstanding problems. The TSPC was chosen for a detailed investigation in part because of its critical position at the inflection point where regional elevation begins to increase northward, and because it is the southernmost center at which important crustal contributions to evolved magmas are apparent (Davidson et al., 1987
, 1988).
As one simple illustration of how the TSPC affects the interpretation of regional data (Fig. 30), the total ranges of a few select elements and elemental ratios at the TSPC are compared at 57·5 wt % SiO2 with the regression values calculated by Hildreth & Moorbath (1988). The distribution patterns for several of these variables (e.g. FeO*/MgO and Ce/Yb) are sufficiently complex, because the total dataset reflects multiple divergent trends, that it is questionable whether a linear regression is an instructive means of characterizing the increases in such indices with respect to increasing SiO2. The high level of observed diversity is meaningful in the sense that volcanoes south of 36°S show little evidence of being as petrologically complex as the TSPC in terms of crustal contributions to their eruptive products, but it also may be an ‘inverse artifact’ in the sense that (1) the TSPC is by far the most comprehensively sampled volcanic center in the Andes, and (2) although other frontal arc volcanic centers of the SVZ may have been as long-lived as the TSPC, their early phases of activity are either not preserved or have rarely been characterized.
For the elements shown in Fig. 30, the range of compositions spanned by TSPC andesitic magmas is greater than that defined by the regression values from the entire Palomo–Tatara segment, and for some indices the TSPC data encompass most or all the regression values for the Tupungato–Maipo segment and the northern cones of the Longaví–Osorno segment. Exceptions are (1) higher K2O and Ba (Rb) at the TSPC than at cones to the south (Longaví–Osorno segment), where closed-system magma evolution apparently dominates, and (2) lower Ce/Yb than those in the exceptional Tupungato–Maipo segment to the north, where extreme suppression of Y–HREE enrichments in evolved lavas apparently reflects assimilation of high Ce/Yb and low-Yb (Y) crustal components (although note the occurrence of basaltic andesitic magmas with high Ce/Yb and SiO2 <56 wt % at the TSPC).
The TSPC data do not invalidate the major features of previously identified regional trends, which are not in dispute in broad terms, but they indicate the need for more high-density sampling along the arc, and they lead to the conclusion that the underlying causes of these trends need to be re-evaluated (1) from the standpoint of magmatic processes and the possibility that there are diverse sources for crustal components (locally and regionally), and (2) in the framework of a segmented arc (Fig. 1). Crustal contributions to evolved magmas along the SVZ are variable, but in many sequences of the TSPC (36°S) they appear to be more important than within the adjacent Longaví–Osorno segment (36·2–42°S). Conversely, the TSPC is comparable with nearby multi-edifice volcanic complexes of its own arc segment (Figs 1 and 30) such as Azul–Quizapu–Descabezado Grande (Hildreth & Moorbath, 1988; Hildreth & Drake, 1992) and Planchon–Peteroa–Azufre (Tormey et al., 1995). This abrupt discontinuity in the regional compositional pattern, apparently coinciding with a segment boundary, suggests that there is a need to consider other factors, in addition to crustal thickness variations parallel to the arc, that could affect magma genesis and evolution within and between contrasting arc segments. Among the possible additional factors are the following: (1) important changes in lithologic character of the crust along the axis of the arc; (2) regional and local variability in the state of stress in the crust (including short-term changes), which may affect the tendency for crust–magma interaction or depth of magma segregation; (3) offsets between arc segments that may be reflections of structural discontinuities in the subducting plate. The roles of such variables will be difficult to rigorously assess until (1) volcanic centers at the northern end of the Longaví–Osorno segment are better characterized, and (2) additional geophysical constraints on crustal–lithospheric thickness and structure, plus the geometry and thermal structure of the subducting plate and asthenospheric mantle wedge, are available.
In light of the diversity in parent magma compositions, the variable proportions of mafic, intermediate, and silicic magmas, and the diverse differentiation mechanisms recorded in the various preserved remnants of the long-lived TSPC, inferences about its long-term evolution or ‘characteristic’ compositional signature derived from a short temporal slice of its activity inevitably would misrepresent the true relations. As only a few TSPC basaltic magmas, in a restricted number of volcanic sequences, arrived at the surface in a weakly contaminated state, we suggest that extreme caution should be used in attempts to draw conclusions about source region compositional variables, mantle melting and magma segregation processes, or elemental fluxes through subduction zones on the basis of evolved basaltic magmas. Restated, we are in agreement with Hildreth & Moorbath (1988) that continental arc magmas are likely to be the final outcomes of polybaric, multi-stage, multi-process, multi-component magma evolution paths in which crustal-level processes frequently impose a compositional imprint with the potential to obscure the original mantle signature.
|
ACKNOWLEDGEMENTS |
|---|
We extend thanks to the many participants in the Tatara–San Pedro Project, without whom this paper would not have been possible: Jon Davidson, Kurt Ferguson, Gabriel Sanchez, Leo Lopez-Escobar, Mike Colucci, Russ Harmon, Wes Hildreth, Bob Drake, Fred Frey, Mike Rhodes, Laurie Brown, Jim Pickens, Brad Singer, Yann Vincze, Laura Webb, Andrea Marzoli, Todd Feeley, Jorge Lobato, Kathy Ervin, Lyn Gaultieri, Steve Nelson, Paul Carrera, and Fidel Costa. A.W. is especially grateful for counsel and assistance in the University of Massachusetts XRF laboratory provided by Mike Rhodes and Pete Dawson, and R.T. acknowledges the essential contributions of Jim Messerich during the preparation of photogrammetric projections. We appreciate the patience, technical ability, and creativity displayed by Jacques Metzger during the preparation, and multiple revisions, of the figures. Without the skilled logistical support provided by José Oviedo (1984–1986) and Francisco Espinoza (1992–1998), and various members of their families, field-work in this area would have been far more difficult, if not impossible.
M.A.D. acknowledges the financial support of the US National Science Foundation (EAR-90-19441, EAR-90-17467 and EAR-93-16544) and the Swiss Fonds National de Recherche Scientific (21-36509.92, 21-37449.93, 20-42124.94, 20-49730.96, 20-55852.98). Jon Davidson, Wes Hildreth, and Fred Frey have been especially encouraging and supportive colleagues during the 15 years that this project has been under way, and M.A.D. has benefited immensely from discussions with them about arc volcanism in general and the southern Andes in particular during this period. Laurie Brown has generously shared unpublished paleomagnetic data. Mike Holdaway provided critical moral and material support during several stages of this investigation, and without the original initiative taken jointly by Russ Harmon and Leo Lopez-Escobar this project would not have been conceived.
This manuscript has been improved substantially thanks to reviews by Wes Hildreth, Richard Price, Mike Clynne, Chris Nye, and especially Dennis Geist. We appreciate the effort that is required to rigorously review a manuscript of this size and complexity, and we have benefited from the constructive suggestions for improvement. Marge Wilson has graciously offered encouragement and understanding with regard to the publication of such an unusual manuscript, as well as some extremely insightful editorial suggestions for revising and restructuring of the original version.
|
FOOTNOTES |
|---|
*Corresponding author. Telephone: +41-22-702-6630. Fax: +41-22-320-5732. E-mail: michael.dungan{at}terre.unige.ch
Present address: Department of Geosciences, University of Iowa, Iowa City, IA 52242, USA.
Extended dataset can be found at http://www.petrology.oupjournals.org
|
REFERENCES |
|---|
Albarède, F. & Tamagnan, V. (1988). Modelling of the recent geochemical evolution of the Piton de la Fournaise Volcano, Réunion Island, 1931–1986. Journal of Petrology 29, 997–1030.
Albarède, F., Luais, B., Fitton, G., Semet, M., Kaminski, E., Upton, B. G. J., Bachèlery, P. & Cheminèe, J.-L. (1997). The geochemical regimes of Piton de la Fournaise Volcano (Réunion) during the last 530,000 years. Journal of Petrology 38, 171–201.
Allan, J. F. & Simkin, T. (2000). Fernandina Volcano’s evolved, well-mixed basalts: mineralogical and petrological constraints in the nature of the Galapagos plume. Journal of Geophysical Research 105, 6017–6041.
Arculus, R. J. (2000). Use and abuse of the term calcalkaline. In: Smith, I. E. M., Davidson, J. P., Gamble, J. A. & Price, R. C. (eds) State of the Arc 2000: Processes and Timescales. Wellington: Royal Society of New Zealand, pp. 6–9.
Bacon, C. R. (1983). Eruptive history of Mt. Mazama and Crater Lake caldera, Cascade Range, U.S.A. Journal of Volcanology and Geothermal Research 18, 57–116.
Bacon, C. R. (1986). Magmatic inclusions in silicic and intermediate volcanic rocks. Journal of Geophysical Research 91, 6091–6112.
Baker, M. B., Grove, T. L., Kinzler, R. J., Donnelly-Nolan, J. M. & Wandless, G. A. (1991). Origin of compositional zonation (high-alumina basalt to basaltic andesite) in the Giant Crater lava field, Medicine Lake Volcano, Northern California. Journal of Geophysical Research 96, 21819–21842.
Brown, L., Pickens, J. & Singer, B. (1994). Matuyama–Brunhes transition recorded in lava flows of the Chilean Andes: evidence for dipolar fields during reversals. Geology 22, 299–302.
Clynne, M. A. (1999). Complex magma mixing origin for multiple volcanic lithologies erupted in 1915, from Lassen Peak, California. Journal of Petrology 40, 105–132.
Costa, F. (2000). The petrology and geochemistry of diverse crustal xenoliths, Tatara–San Pedro Volcanic Complex, Chilean Andes. Terre et Environnement 19, 1–119.
Davidson, J. P. (1991). Comment on ‘Role of subduction erosion in the generation of Andean magmas’. Geology 19, 1054–1056.
Davidson, J. P. (1996). Deciphering mantle and crustal signatures in subduction zone magmatism. In: Bebout, G. E., Scholl, D. W., Kirby, S. H. & Platt, J. P. (eds) Subduction: Top to Bottom. Geophysical Monograph, American Geophysical Union 96, 251–262.
Davidson, J. P., Dungan, M. A., Ferguson, K. M. & Colucci, M. T. (1987). Crust–magma interactions and the evolution of arc magmas. The San Pedro–Pellado volcanic complex, southern Chilean Andes. Geology 15, 443–446.
Davidson, J. P., Ferguson, K. M., Colucci, M. T. & Dungan, M. A. (1988). The origin and evolution of magmas from the San Pedro–Pellado volcanic complex, S. Chile: multiple sources and open system evolution. Contributions to Mineralogy and Petrology 100, 429–445.
DePaolo, D. J. (1981). Trace element and isotopic effects of combined wallrock assimilation and fractional crystallization. Earth and Planetary Science Letters 53, 189–202.[Web of Science]
Déruelle, B. (1982). Petrology of the Plio-Quaternary volcanism of the south–central and meridional Andes. Journal of Volcanology and Geothermal Research 14, 77–124.
Donnelly-Nolan, J. M., Champion, D. E., Grove, T. L., Baker, M. B., Taggert, J. E., Jr & Bruggman, P. E. (1991). The Giant Crater lava field: geology and geochemistry of a compositionally zoned, high-alumina basalt to basaltic andesite eruption at Medicine Lake Volcano, California. Journal of Geophysical Research 96, 21843–21863.[Web of Science]
Druitt, T. H., Edwards, L., Mellors, R. M., Pyle, D. M., Sparks, R. S. J., Lanphere, M., Davies, M. & Barreiro, B. (2000). Santorini Volcano. Geological Society, London, Memoir 19, 165 pp.
Dungan, M. A., Lindstrom, M. M., McMillan, N. J., Moorbath, S., Hoefs, J. & Haskin, L. A. (1986). Open system magmatic evolution of the Taos Plateau volcanic field, northern New Mexico 1. The petrology and geochemistry of the Servilleta Basalt. Journal of Geophysical Research 91, 5999–6028.
Dungan, M. A., Singer, B., Brown, L., Rhodes, J. M. et al. (1992). The life history of an Andean volcano. EOS Transactions, American Geophysical Union 73, 406–407.
Dungan, M. A., Hildreth, W., Davidson, J. P. & Frey, F. A. (2000). An assessment of along-arc compositional patterns in the Andean Southern Volcanic Zone (SVZ: 33–41°S) from the perspective of the Quaternary Tatara–San Pedro complex (TSPC: 36°S). In: Smith, I. E. M., Davidson, J. P., Gamble, J. A. & Price, R. C. (eds) State of the Arc 2000: Processes and Timescales. Wellington: Royal Society of New Zealand, pp. 55–60.
Dzurisin, D., Koyanagi, R. Y. & English, T. T. (1984). Magma supply and storage at Kilauea volcano, Hawaii: 1956–1983. Journal of Volcanology and Geothermal Research 21, 177–206.[Web of Science]
Edwards, B. R. & Russell, J. K. (1998). Time scales of magmatic processes: new insights from dynamic models for magmatic assimilation. Geology 26, 1103–1106.
Eichelberger, J. C. (1980). Vesiculation of mafic magma during replenishment of a silicic magma reservoir. Nature 288, 446–450.
Feeley, T. C. & Dungan, M. A. (1996). Compositional and dynamic controls on mafic–silicic magma interactions at continental arc volcanoes: evidence from Cordón El Guadal, Tatara–San Pedro Complex, Chile. Journal of Petrology 37, 1547–1577.
Feeley, T. C., Dungan, M. A. & Frey, F. A. (1998). Geochemical constraints on the origin of mafic and silicic magmas at Cordón El Guadal, Tatara–San Pedro Complex, central Chile. Contributions to Mineralogy and Petrology 131, 393–411.
Ferguson, K. M., Dungan, M. A., Davidson, J. P. & Colucci, M. T. (1992). The Tatara–San Pedro volcano, 36°S Chile: a chemically variable, dominantly mafic magmatic system. Journal of Petrology 33, 1–43.
Francalanci, L., Tommasini, S., Conticelli, S. & Davies, G. R. (1999). Sr isotope evidence for short magma residence time for the 20th century activity at Stromboli volcano, Italy. Earth and Planetary Science Letters 167, 61–69.
Frey, F. A., Gerlach, D. C., Hickey, R. L., López-Escobar, L. & Munizaga-Villavicenzio, F. (1984). Petrogenesis of the Laguna del Maule volcanic complex, Chile (36°S). Contributions to Mineralogy and Petrology 88, 133–149.
Futa, K. & Stern, C. R. (1988). Sr and Nd isotopic and trace element compositions of Quaternary centers of the southern Andes. Earth and Planetary Science Letters 88, 253–262.
Gaetani, G. A., Grove, T. L. & Bryan, W. B. (1993). The influence of water on subduction-related igneous rocks. Nature 365, 332–334.
Gamble, J., Wood, C. P., Price, R. C., Smith, I. E. M., Stewart, R. B. & Waight, T. (1999). A fifty year perspective of magmatic evolution on Ruapehu volcano, New Zealand: verification of open system behavior in an arc volcano. Earth and Planetary Science Letters 170, 301–314.
Garcia, M. O., Ho, R. A., Rhodes, J. M. & Wolfe, E. W. (1989). Petrologic constraints on rift-zone processes. Results from episode 1 of the Puu Oo eruption of Kilauea volcano, Hawaii. Bulletin of Volcanology 52, 81–96.
Garcia, M. O., Rhodes, J. M., Wolfe, E. W., Ulrich, G. E. & Ho, R. A. (1992). Petrology of lavas from episodes 2–47 of the Puu Oo eruption of Kilauea Volcano: evaluation of magmatic processes. Bulletin of Volcanology 55, 1–16.
Garcia, M. O., Rhodes, J. M., Trusdell, F. A. & Pietruszka, A. J. (1996). Petrology of lavas from the Puu Oo eruption of Kilauea Volcano: III. The Kapaianaha episode (1986–1992). Bulletin of Volcanology, 58, 359–379.[Web of Science]
Garcia, M. O., Ito, E., Eiler, J. M. & Pietruszka, A. J. (1998). Crustal contamination of Kilauea Volcano magmas revealed by oxygen isotope analyses of glass and olivine from Puu Oo eruption lavas. Journal of Petrology 39, 803–817.
Gardeweg, M. (1981). El volcanismo cenozoico superior del area del Nevado de Longaví: una zona de transicíon en los Andes de Chile central. VIII Congreso Geológico Argentino, San Luis, Actas III, 221–240.
Gee, M. A. M., Thirlwall, M. F., Taylor, R. N., Lowry, D. & Murton, B. J. (1998). Crustal processes: major controls on Reykjanes Peninsula lava chemistry, SW Iceland. Journal of Petrology 39, 819–839.
Geist, D., Naumann, T. & Larson, P. (1998). Evolution of Galápagos magmas: mantle and crustal fractionation without assimilation. Journal of Petrology 39, 953–971.
Geist, D., White, W., Naumann, T. & Reynolds, R. (1999). Illegitimate magmas of the Galápagos: insights into mantle mixing and magma transport. Geology 27, 1103–1106.
Gerlach, D. C., Frey, F. A., Moreno-Roa, H. & López-Escobar, L. (1988). Recent volcanism in the Puyehue–Cordon Caulle region, Southern Andes, Chile (40·5°S): petrogenesis of evolved lavas. Journal of Petrology 29, 333–382.
Gilbert, J. S., Stasiuk, M. V., Lane, S. J., Adam, C. R., Murphy, M. D., Sparks, R. S. J. & Naranjo, J. A. (1996). Non-explosive, constructional evolution of the ice-filled caldera at Volcán Sollipulli. Bulletin of Volcanology 58, 67–83.[Web of Science]
González-Ferrán, O. & Vergara, M. (1962). Reconocimiento geológico de la Cordillera de los Andes entre los paralelos 35° y 38° latitud sur. Universidad de Chile, Facultad de Ciences Fisicas y Matematicas 24, 121 pp.
Grove, T. L. & Kinzler, R. J. (1986). Petrogenesis of andesites. Annual Review of Earth and Planetary Sciences 14, 417–454.[Web of Science]
Grove, T. L., Kinzler, R. J., Baker, M. B., Donnelly-Nolan, J. M. & Lesher, C. E. (1988). Assimilation of granite by basaltic magma at Burnt Lava flow, Medicine Lake volcano, northern California: decoupling of heat and mass transfer. Contributions to Mineralogy and Petrology 99, 320–343.
Grunder, A. L. (1987). Low 18O silicic volcanic rocks at the Calabozos Caldera Complex, Southern Andes. Contributions to Mineralogy and Petrology 95, 71–81.
Grunder, A. L. & Mahood, G. A. (1988). Physical and chemical models of zoned silicic magmas: the Loma Seca tuff and Calabozos caldera, southern Andes. Journal of Petrology 29, 831–867.
Hackett, W. R. & Houghton, B. F. (1989). A facies model for a Quaternary andesitic composite volcano: Ruapehu, New Zealand. Bulletin of Volcanology 51, 51–68.
Harmon, R. S. & Hoefs, J. (1984). Oxygen isotope ratios in Late Cenozoic Andean volcanics. In: Harmon, R. S. & Barreiro, B. A. (eds) Andean Magmatism: Chemical and Isotopic Constraints. Nantwich, UK: Shiva, pp. 9–20.
Hickey, R. L., Gerlach, D. & Frey, F. A. (1984). Geochemical variations in volcanic rocks from central–south Chile (33–42°S). In: Harmon, R. S. & Barreiro, B. A. (eds) Andean Magmatism: Chemical and Isotopic Constraints. Nantwich, UK: Shiva, pp. 72–95.
Hickey, R. L., Frey, F. A., Gerlach, D. C. & López-Escobar, L. (1986). Multiple sources for basaltic arc rocks from the Southern Volcanic Zone of the Andes (34°–41°S): trace element and isotopic evidence for contributions from subducted oceanic crust, mantle, and continental crust. Journal of Geophysical Research 91, 5963–5983.
Hickey-Vargas, R., Moreno-Roa, H., López-Escobar, L. & Frey, F. A. (1989). Geochemical variations in Andean basaltic and silicic lavas from the Villarrica–Lanin volcanic chain (39·5°S): an evaluation of source heterogeneity, fractional crystallization and crustal assimilation. Contributions to Mineralogy and Petrology 103, 361–386.
Hickey-Vargas, R., Abdollahi, M. J., Parada, M. A., López-Escobar, L. & Frey, F. A. (1995). Crustal xenoliths from Calbuco Volcano, Andean Southern Volcanic Zone: implications for crustal composition and magma–crust interaction. Contributions to Mineralogy and Petrology 119, 331–344.
Hildreth, W. & Drake, R. E. (1992). Volcán Quizapu, Chilean Andes. Contributions to Mineralogy and Petrology 54, 93–125.
Hildreth, W. & Lanphere, M. A. (1994). Potassium–argon geochronology of a basalt–andesite–dacite arc system: the Mount Adams volcanic field, Cascade Range of southern Washington. Geological Society of America Bulletin 106, 1413–1429.
Hildreth, W. & Moorbath, S. (1988). Crustal contributions to arc magmatism in the Andes of central Chile. Contributions to Mineralogy and Petrology 98, 455–489.[Web of Science]
Hildreth, W. & Moorbath, S. (1991). Reply to Comment on ‘Crustal contributions to arc magmatism in the Andes of Central Chile’ by W. Hildreth and S. Moorbath. Contributions to Mineralogy and Petrology 108, 247–252.
Hildreth, W., Grunder, A. L. & Drake, R. E. (1984). The Loma Seca tuff and Calabozos caldera: a major ash-flow and caldera complex in the southern Andes of central Chile. Geological Society of America Bulletin 95, 45–54.
Hildreth, W., Fierstein, J., Godoy, E., Drake, R. E. & Singer, B. (1999). The Puelche volcanic field: extensive Pleistocene rhyolite lava flows in the Andes of central Chile. Revista Geológica de Chile 26, 275–309.[Web of Science]
Hobden, B., Houghton, B. F., Davidson, J. P. & Weaver, S. D. (2000). Rapid growth and non-systematic magmatic evolution of an active andesitic cone: Ngauruhoe Volcano, New Zealand. In: Smith, I. E. M., Davidson, J. P., Gamble, J. A. & Price, R. C. (eds) State of the Arc 2000: Processes and Timescales. Wellington: Royal Society of New Zealand, pp. 109–112.
Huppert, H. H. & Sparks, R. S. J. (1985). Cooling and contamination of mafic and ultramafic magmas during ascent through continental crust. Earth and Planetary Science Letters 74, 371–386.[Web of Science]
Kerr, A. C., Kempton, P. D. & Thompson, R. N. (1995). Crustal assimilation during turbulent magma ascent (ATA); new isotopic evidence from the Mull Tertiary lava succession, N.W. Scotland. Contributions to Mineralogy and Petrology 119, 142–154.[Web of Science]
Lipman, P. W., Dungan, M. A., Brown, L. L. & Deino, A. (1996). Recurrent eruption and subsidence at the Platoro caldera complex, southeastern San Juan volcanic field, Colorado: new tales from old tuffs. Geological Society of America Bulletin 108, 1039–1055.
López-Escobar, L. (1984). Petrology and chemistry of volcanic rocks of the Southern Andes. In: Harmon, R. S. & Barreiro, B. A. (eds) Andean Magmatism: Chemical and Isotopic Constraints. Nantwich, UK: Shiva, pp. 47–71.
López-Escobar, L., Frey, F. A. & Vergara, M. (1977). Andesites and high-alumina basalts from the central–south Chile High Andes: geochemical evidence bearing on their petrogenesis. Contributions to Mineralogy and Petrology 63, 199–228.
López-Escobar, L., Vergara, M. & Frey, F. A. (1981). Petrology and geochemistry of lavas from Antuco volcano, a basaltic volcano of the Southern Andes. Journal of Volcanology and Geothermal Research 11, 329–352.[Web of Science]
López-Escobar, L., Parada, M. A., Moreno, H., Frey, F. A. & Hickey-Vargas, R. (1992). A contribution to the petrogenesis of Osorno and Calbuco volcanoes, Southern Andes (41°–41°30’S): comparative study. Geochemical Journal 19, 209–226.
López-Escobar, L., Kilian, R., Kempton, P. D. & Tagiri, M. (1993). Petrography and geochemistry of Quaternary rocks from the Southern Volcanic Zone of the Andes between 41°30’ and 46°00’S, Chile. Revista Geológica de Chile 20, 33–55.
López-Escobar, L., Parada, M. A., Hickey-Vargas, R., Frey, F. A., Kempton, P. D. & Moreno, H., (1995). Calbuco Volcano and minor eruptive centers distributed along the Liquiñe–Ofqui Fault Zone, Chile (41°–42°S): contrasting origin of andesitic and basaltic magma in the Southern Volcanic Zone of the Andes. Contributions to Mineralogy and Petrology 119, 345–361.
Luhr, J. F., Pier, J. G., Aranda-Gomez, J. J. & Podosek, F. A. (1995). Crustal contamination in early Basin-and-Range hawaiites of the Los Encinos Volcanic Field, central Mexico. Contributions to Mineralogy and Petrology 118, 321–339.
McBirney, A. R., Taylor, H. P. & Armstrong, R. L. (1987) Paricutin re-examined: a classic example of crustal assimilation in calc-alkaline magma. Contributions to Mineralogy and Petrology 95, 4–20.
McMillan, N. J., Harmon, R., López-Escobar, L. & Strong, D. F. (1989). Crustal sources involved in continental arc magmatism: a case study of Volcán Mocho–Choshuenco, southern Chile. Geology 17, 1152–1156.
Moreno, H. & Gardeweg, M. C. (1989). La erupcion en el complejo volcanico Lonquimay (Diciembre 1988–), Andes del Sur. Revista Geológica de Chile 16, 93–117.
Muñoz, B. & Niemeyer, R. H. (1984). Carta Geológica de Chile No. 64: Hoja Laguna del Maule. Santiago: Servicio National de Geológia y Minera de Chile.
Muñoz, J. & Stern, C. R. (1987). Transverse structural and petrochemical variations across the segment 38–39°S of the Quaternary volcanic belt of the southern South American continental margin. Journal of South American Earth Science 1, 147–162.
Muñoz, J. & Stern, C. R. (1989). Alkaline magmatism within the segment 38°–39°S of the Plio-Quaternary volcanic belt of the southern American continental margin. Journal of Geophysical Research 94, 4545–4560.
Murphy, M. D. & Brewer, T. S. (1994). Geochemistry of mafic magmas from Sollipulli Volcano, southern Chile: constraints on models of magmagenesis in the Southern Volcanic Zone of the Andes. Goldschmidt Conference Abstracts. Mineralogical Magazine 58A, 639–640.
Nelson, S. T., Davidson, J. P., Heizler, M. T. & Kowallis, B. J. (1999). Tertiary tectonic history of the southern Andes: the subvolcanic sequence to the Tatara–San Pedro volcanic complex, lat 36°S. Geological Society of America Bulletin 111, 1387–1404.
Notsu, K., López-Escobar, L. & Onuma, N. (1986). Along arc variation of Sr-isotope composition in volcanic rocks from the southern Andes (33°–55°S). Geochemical Journal 21, 307–313.
O’Hara, M. J. (1998). Volcanic plumbing and the space problem—thermal and geochemical consequences of large-scale assimilation in ocean island development. Journal of Petrology 39, 1077–1089.
Pallister, J. S., Hoblitt, R. P. & Reyes, A. G. (1992). A basalt trigger for the 1991 eruptions of Pinatubo volcano. Nature 356, 424–428.
Pietruszka, A. J. & Garcia, M. O. (1999). A rapid fluctuation in the mantle source and melting history of Kilauea Volcano inferred from the geochemistry of its summit lavas (1790–1982). Journal of Petrology 40, 1321–1342.
Philpotts, A. R. & Asher, P. M. (1993). Wallrock melting and reaction effects along the Higganum diabase dike in Connecticut: contamination of a continental flood basalt feeder. Journal of Petrology 34, 1029–1058.
Reiners, P. W., Nelson, B. K. & Ghiorso, M. S. (1995). Assimilation of felsic crust by basaltic magma: thermal limits and extents of crustal contamination of mantle-derived magmas. Geology 23, 563–566.
Rhodes, J. M. (1983). Homogeneity of lava flows: chemical data for historical Mauna Loa eruptions. Journal of Geophysical Research 88A, 869–879.
Rhodes, J. M. (1988). Geochemistry of the 1984 Mauna Loa eruption: implications for magma storage and supply. Journal of Geophysical Research 93, 4453–4466.
Rhodes, J. M. (1996). Geochemical stratigraphy of lava flows sampled by the Hawaiian Scientific Drilling Project. Journal of Geophysical Research 101, 11279–11746.
Rhodes, J. M. & Hart, S. R. (1995). Episodic trace element and isotopic variations in historical Mauna Loa lavas: implications for magma and plume dynamics. In: Rhodes, J. M. & Lockwood, J. P. (eds) Mauna Loa Revealed: Structure, Composition, and Hazards. Geophysical Monograph, American Geophysical Union 92, 263–288.
Rhodes, J. M., Wenz, K. P., Neal, J. W., Sparks, J. W. & Lockwood, J. P. (1989). Geochemical evidence for invasion of Kilauea’s plumbing system by Mauna Loa magma. Nature 337, 257–260.
Singer, B. S. & Pringle, M. S. (1996). Age and duration of the Matuyama–Bruhnes geomagnetic polarity reversal from 40Ar/39Ar incremental-heating analyses of lavas. Earth and Planetary Science Letters 139, 47–61.
Singer, B. S., Dungan, M. A. & Layne, G. D. (1995). Textures and Sr, Ba, Mg, Fe, K, and Ti compositional profiles in volcanic plagioclase: clues to the dynamics of calc-alkaline magmas. American Mineralogist 80, 776–798.[Abstract]
Singer, B. S., Thompson, R. A., Dungan, M. A., Feeley, T. C., Nelson, S. T., Pickens, J. C., Brown, L. L., Davidson, J. P. & Metzger, J. (1997). Volcanism and erosion during the past 930 ky at the Tatara–San Pedro complex, Chilean Andes. Geological Society of America Bulletin 109, 127–142.
Singer, B. S., Wijbrans, J. R., Nelson, S. T., Pringle, M. S., Feeley, T. C. & Dungan, M. A. (1998). Inherited argon in a Pleistocene andesite lava: 40Ar/39Ar incremental-heating and laser-fusion analyses of plagioclase. Geology 26, 427–430.
Sparks, R. S. J. & Marshall, L. A. (1986). Thermal and mechanical constraints on mixing between mafic and silicic magmas. Journal of Volcanology and Geothermal Research 29, 99–124.
Stern, C. R. (1988). Source region versus intra-crustal contamination in the petrogenesis of the Quaternary volcanic centers at the northern end (33–34°S) of the Southern Volcanic Zone of the Andes. Actas 5th Congreso Geológico Chileno, pp. 129–143.
Stern, C. R. (1989). Pliocene to present migration of the volcanic front, Andean Southern Volcanic Zone. Revista Geológica de Chile 16, 154–162.
Stern, C. R. (1991a). Role of subduction erosion in the generation of Andean magmas. Geology 19, 78–81.
Stern, C. R. (1991b). Comment on ‘Crustal contributions to arc magmatism in the Andes of Central Chile’ by W. Hildreth and S. Moorbath. Contributions to Mineralogy and Petrology 108, 241–246.
Stern, C. R., Futa, K., Muehlenbachs, K., Dobbs, F. M., Muñoz, J., Godoy, E. & Charrier, R. (1984). Sr, Nd, Pb, and O isotope compositions of late Cenozoic volcanics, northernmost SVZ (33°–34°S). In: Harmon, R. S. & Barreiro, B. A. (eds) Andean Magmatism: Chemical and Isotopic Constraints. Nantwich, UK: Shiva, pp. 96–105.
Stern, C. R., Frey, F. A., Futa, K., Zartman, R. E., Peng, Z. & López-Escobar, L. (1991). Trace-element and Sr, Nd, Pb, and O isotopic composition of Pliocene and Quaternary alkali basalts of the Patagonian plateau lavas of southernmost South America. Contributions to Mineralogy and Petrology 104, 294–308.
Tepley, F. J., III, Davidson, J. P. & Clynne, M. A. (1999). Magmatic interactions as recorded in plagioclase phenocrysts of Chaos Crags, Lassen Volcanic Center, California. Journal of Petrology 40, 787–806.
Thompson, R. A. & Dungan, M. A. (1985). The petrology of the Handkerchief Mesa mixed magma complex, San Juan Mountains, Colorado. Journal of Volcanology and Geothermal Research 26, 251–274.
Tormey, D. R., Hickey-Vargas, R., Frey, F. A. & López-Escober, L. (1991). Recent lavas from the Andean volcanic front (33 to 42°S); interpretations of along-arc compositional variations. In: Harmon, R. S. & Rapela, C. W. (eds) Andean Magmatism and its Tectonic Setting. Geological Society of America, Special Paper 265, 57–77.
Tormey, D. R., Frey, F. A. & López-Escobar, L. (1995). Geochemistry of the active Azufre–Planchon–Peteroa volcanic complex, Chile (35°15’S): evidence for multiple sources and processes in a cordilleran arc magmatic system. Journal of Petrology 36, 265–298.
Wilcox, R. E. (1954). Petrology of Paricutin Volcano, Mexico. US Geological Survey Bulletin 956-B, 165–279.
Wood, C. A. & Nelson, K. L. (1988) Inter-segment variation of magmatic compositions in the Andes of central Chile. EOS Transactions, American Geophysical Union 69(44), 1494.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. E. Schmidt and A. L. Grunder The evolution of North Sister: A volcano shaped by extension and ice in the central Oregon Cascade Arc Geological Society of America Bulletin, May 1, 2009; 121(5-6): 643 - 662. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Reubi and J. Blundy Assimilation of Plutonic Roots, Formation of High-K 'Exotic' Melt Inclusions and Genesis of Andesitic Magmas at Volcan De Colima, Mexico J. Petrology, December 24, 2008; (2008) egn066v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. S. Singer, B. R. Jicha, M. A. Harper, J. A. Naranjo, L. E. Lara, and H. Moreno-Roa Eruptive history, geochronology, and magmatic evolution of the Puyehue-Cordon Caulle volcanic complex, Chile Geological Society of America Bulletin, May 1, 2008; 120(5-6): 599 - 618. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. C. Feeley, M. A. Clynne, G. S. Winer, and W. C. Grice Oxygen Isotope Geochemistry of the Lassen Volcanic Center, California: Resolving Crustal and Mantle Contributions to Continental Arc Magmatism J. Petrology, May 1, 2008; 49(5): 971 - 997. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. RodrIguez, D. Selles, M. Dungan, C. Langmuir, and W. Leeman Adakitic Dacites Formed by Intracrustal Crystal Fractionation of Water-rich Parent Magmas at Nevado de Longavi Volcano (36{middle dot}2{degrees}S; Andean Southern Volcanic Zone, Central Chile) J. Petrology, November 1, 2007; 48(11): 2033 - 2061. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. L. A. Charlier, O. Bachmann, J. P. Davidson, M. A. Dungan, and D. J. Morgan The Upper Crustal Evolution of a Large Silicic Magma Body: Evidence from Crystal-scale Rb Sr Isotopic Heterogeneities in the Fish Canyon Magmatic System, Colorado J. Petrology, October 1, 2007; 48(10): 1875 - 1894. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. R. Bacon and M. A. Lanphere Eruptive history and geochronology of Mount Mazama and the Crater Lake region, Oregon Geological Society of America Bulletin, November 1, 2006; 118(11-12): 1331 - 1359. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. A. Ramos and S. M. Kay Overview of the tectonic evolution of the southern Central Andes of Mendoza and Neuquen (35{degrees}-39{degrees}S latitude) Geological Society of America Special Papers, January 1, 2006; 407(0): 1 - 17. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Kay, W. M. Burns, P. Copeland, and O. Mancilla Upper Cretaceous to Holocene magmatism and evidence for transient Miocene shallowing of the Andean subduction zone under the northern Neuquen Basin Geological Society of America Special Papers, January 1, 2006; 407(0): 19 - 60. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Costa and M. Dungan Short time scales of magmatic assimilation from diffusion modeling of multiple elements in olivine Geology, October 1, 2005; 33(10): 837 - 840. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. REUBI and I. A. NICHOLLS Structure and Dynamics of a Silicic Magmatic System Associated with Caldera-Forming Eruptions at Batur Volcanic Field, Bali, Indonesia J. Petrology, July 1, 2005; 46(7): 1367 - 1391. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Kay, E. Godoy, and A. Kurtz Episodic arc migration, crustal thickening, subduction erosion, and magmatism in the south-central Andes Geological Society of America Bulletin, January 1, 2005; 117(1-2): 67 - 88. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Dungan and J. Davidson Partial assimilative recycling of the mafic plutonic roots of arc volcanoes: An example from the Chilean Andes Geology, September 1, 2004; 32(9): 773 - 776. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. COSTA, B. SCAILLET, and M. PICHAVANT Petrological and Experimental Constraints on the Pre-eruption Conditions of Holocene Dacite from Volcan San Pedro (36{degrees}S, Chilean Andes) and the Importance of Sulphur in Silicic Subduction-related Magmas J. Petrology, April 1, 2004; 45(4): 855 - 881. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. M. Frey, R. A. Lange, C. M. Hall, and H. Delgado-Granados Magma eruption rates constrained by 40Ar/39Ar chronology and GIS for the Ceboruco-San Pedro volcanic field, western Mexico Geological Society of America Bulletin, March 1, 2004; 116(3-4): 259 - 276. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. COSTA and B. SINGER Evolution of Holocene Dacite and Compositionally Zoned Magma, Volcan San Pedro, Southern Volcanic Zone, Chile J. Petrology, August 1, 2002; 43(8): 1571 - 1593. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. C. FEELEY, M. A. COSCA, and C. R. LINDSAY Petrogenesis and Implications of Calc-Alkaline Cryptic Hybrid Magmas from Washburn Volcano, Absaroka Volcanic Province, USA J. Petrology, April 1, 2002; 43(4): 663 - 703. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. COSTA, M. A. DUNGAN, and B. S. SINGER Hornblende- and Phlogopite-Bearing Gabbroic Xenoliths from Volcan San Pedro (36{degrees}S), Chilean Andes: Evidence for Melt and Fluid Migration and Reactions in Subduction-Related Plutons J. Petrology, February 1, 2002; 43(2): 219 - 241. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||