| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 44 | Number 8 | Pages 1455-1475 | 2003
© Oxford University Press 2003
Experimental Constraints on the Conditions of Formation of Highly Calcic Plagioclase Microlites at the Soufrire Hills Volcano, Montserrat
1 DEPARTMENT OF EARTH SCIENCES, BRISTOL UNIVERSITY, BRISTOL BS8 1RJ, UK
2 SCHOOL OF ENVIRONMENTAL SCIENCE, UNIVERSITY OF EAST ANGLIA, NORWICH NR4 7TJ, UK
3 DIPARTMENTO DE SCIENZE DELLA TERRA, UNIVERSITA DI CAMERINO, 62032 CAMERINO, ITALY
* Corresponding author. Present address: Department of Earth Sciences, Bristol University, Bristol BS8 1RJ, UK. E-mail: S.Couch{at}mail.com
RECEIVED NOVEMBER 2, 2002; ACCEPTED JANUARY 24, 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
High-pressure and -temperature experiments on a bulk-rock composition representative of the groundmass of the Soufrière Hills Volcano andesite have allowed the phase equilibria of the system to be determined; these are then compared with the natural samples. Experimental conditions varied from 825 to 1100°C and from 5 to 225 MPa; the main phases observed were clinopyroxene, crystalline silica, amphibole and plagioclase. A relationship between plagioclase microlite size and anorthite content is identified in samples of the natural andesite. Large crystals (>60 µm2 in area) have cores of An60–75, whereas small crystals (<60 µm2 in area) have cores of An40–60. Experimental results show that if the magma is heated to >950°C the high-anorthite microlite crystals can form at magma chamber pressures without any need for a change in bulk composition. It is proposed that convective self-mixing occurs within the magma chamber. Geothermometry of coexisting plagioclase–amphibole pairs confirms the complex crystallization history of the natural samples. Analysis of natural glass samples has identified compositional variations that can be related to the crystallinity of the sample and also the groundmass plagioclase composition. Rapidly erupted pumice samples have high glass contents, lower SiO2 glass compositions and plagioclase microlites that are large in size (>60 µm2) and have a high anorthite content (>An60). Slowly erupted dome samples are highly crystalline and contain numerous plagioclase microlites of variable size and composition.
KEY WORDS: glass evolution; experiment; Montserrat; plagioclase; self-mixing
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The magma erupted at Soufrière Hills Volcano, Montserrat, is a typical orogenic arc andesite, which is crystal rich and has a diverse range of mineral textures indicative of a complex history of magmatic evolution. The petrology and geochemistry of the andesite have been described by Devine et al. (1998a) and Murphy et al. (2000). Those workers proposed that a variety of disequilibrium features can be accounted for by recent reheating of the andesitic magma by basaltic intrusions.
Several observations are consistent with recent reheating. Many plagioclase phenocrysts have resorption surfaces, and calcic overgrowth rims (Fig. 1a and b). Plagioclase microphenocrysts generally have more calcic compositions than the phenocryst cores. Orthopyroxene phenocrysts have thin overgrowth rims with compositions that indicate high crystallization temperatures (up to 970°C) in comparison with estimates for core compositions (840–870°C) (Barclay et al., 1998; Murphy et al., 2000). Fe–Ti oxide pairs show evidence for heating (Devine et al., 1998a) and there are amphibole phenocrysts replaced by intergrowths of pyroxene, plagioclase and titanomagnetite (Devine et al., 1998b; Murphy et al., 2000). The andesite contains a small proportion (
1 wt %) of mafic inclusions with abundant acicular amphibole and plagioclase with a diktytaxitic texture characteristic of rapid crystallization. These observations led Murphy et al. (2000) to propose that the andesite has been heated and remobilized by intrusion of hydrous basalt magma into the chamber. Other observations are consistent with an open-system replenishment by basalt; notably, the increase in discharge rate with time in the 1995–1998 phase of dome growth (Sparks et al., 1998) and SO2 fluxes higher than can easily be reconciled by degassing of the andesitic magma (Young et al., 1998; Edmonds et al., 2001). Effects of decompression during magma ascent have been recognized, including microlite growth (Sparks et al., 2000; Fig. 1c) and the formation of amphibole reaction rims (Devine et al., 1998b; Fig. 1d). Experimental constraints on magmatic conditions (Barclay et al., 1998), in particular the stability of amphibole and apparent coexistence of quartz and amphibole, suggest that the overall magma chamber conditions did not exceed 830°C, with water pressures <130 MPa. Water contents of 4–5 wt % in melt inclusions indicate water pressures of 110–140 MPa (Barclay et al., 1998). Murphy et al. (2000) accounted for the overall homogeneity of the bulk magma throughout the eruption and the complex microscale textural and mineralogical variation as a consequence of convective stirring in a chamber heated from below and cooled by assimilation of wallrocks of similar igneous material. This concept received further support from Harford & Sparks (2001), who observed that some amphibole phenocrysts have heavy 
D isotopic compositions (–6 to +30
), which indicate exchange with hydrothermal meteoric water at temperatures below the solidus. These isotopic heterogeneities could be preserved only if these altered materials had been recently assimilated. Harford & Sparks (2001) estimated residence times of the altered amphiboles as only several years. Zellmer et al. (2003) estimated residence times for plagioclase phenocrysts at magmatic temperatures of a few decades to a few centuries based on the disequilibrium distribution of Sr in complexly zoned crystals.
|
In this paper new experimental results are presented, together with further geochemical data and new textural observations, to develop a more complete understanding of andesitic magma petrogenesis at Montserrat. Groundmass microlites are compared with plagioclase generated in the experiments, and coexisting amphibole–plagioclase pairs have been analysed to estimate temperatures and pressures of equilibration. Although the concept of open-system basalt replenishment is retained, the new experimental data provide additional constraints that help to explain previously ambiguous features, in particular the calcic compositions of plagioclase overgrowth rims and microphenocrysts. The textural and mineralogical characteristics of the andesite are attributed to a process of self-quenching (Couch et al., 2001), in which andesite at the base of the chamber is heated to high temperatures and is then cooled as it is mixed into the chamber interior by convection. Subsequent magmatic evolution can be explained by a polybaric crystallization with late-stage crystallization as a result of degassing during magma ascent (Blundy & Cashman, 2001). The possibility that the magma may be heated during ascent as a result of latent heat release from decompression-induced crystallization is considered. Together these results and interpretations indicate a complex open-system and polybaric crystallization history of the Soufrière Hills magma.
The Soufrière Hills eruption and its magmatic products
The Soufrière Hills eruption began in July 1995 after 3 years of precursory seismicity. The eruption has thus far been characterized by four main phases (Robertson et al., 2000): a first phase of earthquakes and phreatic explosions (July–November 1995); a sustained phase of dome growth (November 1995–March 1998) with eruption of 0·3 km3 (dense rock equivalent) of andesite magma; a third phase of residual volcanic activity (March 1998–November 1999) with ash-venting, numerous minor explosions and occasional pyroclastic flows caused by dome collapse but no dome growth; and the current fourth phase of dome growth (November 1999 to the time of writing, December 2002) with production of a further 0·22 km3 of magma (information provided by the Montserrat Volcano Observatory), accompanied by dome collapse pyroclastic flows and some explosive eruptions. Two main kinds of eruptive product have been identified: (1) pieces of dome rock sampled by pyroclastic flows formed by dome collapse; (2) pumice clasts erupted in Vulcanian explosions. In general, pumices have higher glass contents than dome rocks, although the latter have highly variable glass contents. These variations have been attributed to fluctuations in magma discharge rate and residence time in the dome (Sparks et al., 2000).
Petrological background
The petrology and geochemistry of the andesite have been documented continuously throughout the eruption (Devine et al., 1998a; Murphy et al., 2000). There are no systematic temporal variations in bulk magma composition or phase assemblages. There are minor variations, however, of bulk composition, with SiO2 varying between 58·4 and 62·4 wt % (Murphy et al., 2000). Layering on a centimetre scale is common in lava blocks, but the different coloured layers do not show any compositional contrasts outside analytical error. Rocks vary from dome lavas with low vesicle contents (5–15% typically) to highly vesicular pumices. There is no correlation of vesicularity with bulk-rock chemistry. The andesite is crystalline, with 45–55 wt % phenocrysts (>300 µm in length) and 10–15 wt % microphenocrysts (300–100 µm in length) by modal analysis and variable proportions of microlites (<100 µm in length). The distinction between microphenocrysts and microlites is arbitrary, and is defined by the minimum crystal size that can be point-counted accurately by optical microscope (Murphy et al., 2000). The andesite contains phenocrysts and microphenocrysts of plagioclase (35–40 wt %), and also amphibole (6–10 wt %), orthopyroxene, titanomagnetite and minor quartz. Clinopyroxene (<0·5 wt %) is present only as microphenocrysts. Quartz displays either resorbed boundaries or clinopyroxene reaction rims.
The plagioclase phenocrysts are compositionally and texturally complex. There are sodic phenocrysts (An48–58) with oscillatory zoning, reversely zoned sodic phenocrysts and microphenocrysts (cores of An48–58, rims of An65–80), reversely zoned dusty sieve-textured crystals (rims of An70–88) and rare highly calcic (An89–93) phenocrysts (Murphy et al., 2000). Zellmer et al. (2003) have also reported the presence of calcic cores (up to An80) in a few sodic phenocrysts. Nomarski interference images and detailed compositional profiles indicate complex growth histories for many individual crystals with oscillatory zoning and multiple thin zones of more calcic plagioclase (Zellmer et al., 2003). The microphenocrysts have cores of An48–80 and some crystals have narrow more albitic rims (An55–60).
The groundmass is composed of the same phases as the phenocrysts and microphenocrysts, except that amphibole is absent. The glass varies in proportion (5–35 wt %), and its composition is high-Si rhyolite (76–79 wt % SiO2). Mafic inclusions (1 wt %) are widespread throughout the andesite.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Experimental procedure
The variability in the groundmass crystallinity of the Soufrière Hills andesite is of particular interest as it can provide insights into late-stage crystallization processes. To determine the phase equilibria of the groundmass, a series of experiments were carried out as a function of pressure and temperature using a representative groundmass composition (aMon6a). This was estimated by averaging 200-rastered electron microprobe analyses of the dome rock sample MVO34 (Barclay et al., 1998; Table 1), collected in February 1996. K2O can be used to indicate the degree of crystallization represented by the groundmass, as K is nearly incompatible in the crystallizing mineral assemblage in the Soufrière Hills magma. K2O constitutes only around 0·1–0·4 wt % of plagioclase and 0·18 wt % of amphibole crystals (Murphy et al., 2000). The groundmass composition (aMon6a) represents
35 wt % of the bulk composition from mass balance calculations.
|
A glass with the aMon6a composition was prepared by A.-M. Lejeune. Reagent grade oxides or carbonates were dried before weighing, ground in an agate mortar, slowly decarbonated and heated to 1820 K in a platinum crucible in an electric muffle furnace. After four successive quenches, grindings and fusions the melt was finally quenched. Finely ground starting material (0·05 g) was placed inside 2·5 mm diameter Ag75Pd25 capsules, with sufficient H2O to ensure water saturation at the P–T conditions. After welding they were heated briefly to check they were sealed. Lower-temperature experiments (800–925°C) were carried out in rapid-quench cold-seal pressure vessels (Carroll & Blank, 1997), with an H2O pressurizing fluid at pressures of 5–225 MPa. Rouse (2000) determined the oxygen fugacity of the cold-seal set-up to be NNO + 1·3 (where NNO is nickel–nickel oxide) based on coexisting NiO–NiPd alloys (Taylor et al., 1992). This compares with estimated fO2 of
NNO + 1 for the Soufrière Hills andesite (Devine et al., 1998a). Experiments at higher pressure and temperature were carried out in a rapid quench TZM pressure vessel, which uses argon as its pressurizing medium. Comparison of calculated Fe3+ in pyroxenes grown in the TZM and cold-seal experiments suggests a slightly higher fO2, not exceeding 1–2 fO2 units, in the TZM experiments. Temperature was measured in both set-ups by K-type (chromel–alumel) thermocouples, precise to ±5°C with accuracy checked by B-type (Pt94Rh6–Pt70Rh30) thermocouples, calibrated at the melting point of gold (Gardner et al., 2000). Pressure was measured by either Nova Swiss static pressure gauges, precise to ±2·5 MPa, or a pressure transducer, precise to <1 MPa. The experiments were left for 1–14 days to equilibrate, and were then rapidly quenched. To test the approach to equilibrium reversal experiments were performed, where a sample was taken to higher pressure (50 MPa above desired pressure) for 5 h, before being adjusted to the desired P–T conditions and left for an appropriate time-period. Several experiments close to the plagioclase liquidus were seeded with 2 wt % of finely ground labradorite (An50). Excess H2O was present in all capsules at the end of the experiment.
Analytical techniques
Polished thin sections of experimental and natural samples were analysed on a Cameca Camebax Micro electron microprobe with SamX software and PAP correction procedure for crystal phases at 15 kV accelerating voltage, and a JEOL JXA8600 electron microprobe with LEMAS link software and XAF correction procedure for glass at 15 kV accelerating voltage. The JEOL system was used for glass analyses as it has four spectrometers; this allowed analysis to be completed quickly, before significant alkali loss occurred. Crystals were analysed at a 10 nA beam current with a focused beam, and glass with a 2 nA rastered beam, again to minimize alkali mobility.
Electron probe analysis of hydrous glasses can underestimate Na and overestimate Si as a result of alkali mobility. Studies by Harford (2000) and Steen (unpublished data, 2000) on the optimum analysing conditions for hydrous glasses have found that a 2 nA rastered beam at least 5 µm in width, with early analysis of Na, Si and Al, has an estimated sodium loss of <5%, consistent with the findings of Morgan & London (1996). Wherever possible the raster width was >5 µm. Morgan & London (1996) found that if the beam current is at least 10 nA, water-rich glasses (>5 wt % H2O) suffer significant alkali loss. Harford (2000) studied glasses with <1–1·7 wt % H2O and Steen (unpublished data, 2000) studied glasses with 3–9 wt % H2O. The influence of water content on Na loss was negligible at the conditions of a 2 nA current with a 5 µm rastered beam.
Calibration used known standards (albite for Na and Si; Al2O3 for Al; sanidine for K; andradite for Ca; St. John's olivine for Mg and Si; SrTiO3 for Ti; Fe for Fe) and then was checked by analysis of secondary standards (KK1 and Kn18). Typically, 5–10 plagioclase analyses were carried out for each experiment, and about five analyses for all other phases.
|
EXPERIMENTAL RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The main experimental results are summarized in Table 2. Representative scanning electron microscopy (SEM) images of experimental charges are shown in Fig. 2. There is a systematic variation in the morphology of the plagioclase crystals. At high P–T there are relatively few large crystals with tabular morphology. At lower P–T conditions the number of crystals increases, and the morphology tends towards skeletal and dendritic at very low pressures and temperatures.
|
|
Crystal phases
Figure 3 shows the phase diagram for the equilibrium experiments. The stability of clinopyroxene, crystalline silica, amphibole and plagioclase is dependent on the PH2O and temperature. Amphibole is restricted to temperatures below 875°C and water pressures greater than 100 MPa. Crystalline silica is found at low water pressures and variable temperatures. Spinel (titanomagnetite) is stable throughout the experimental conditions. Plagioclase is stable until high pressures and temperatures are attained. Clinopyroxene is stable in all experiments except at very high temperatures. The experimental phases are the same as the natural mineral assemblage, except for the absence of orthopyroxene, which is observed only in the natural groundmass.
|
Plagioclase can be affected by sluggish growth kinetics, especially just below the plagioclase-in reaction. At these conditions plagioclase failed to nucleate unless seeded, when narrow rims of high-An plagioclase were identified on the labradorite (An50) seed (Fig. 2g). In experiments where no rims were seen, or where the seed was resorbed, the plagioclase-in reaction is interpreted to have been exceeded (Fig. 2h). Figure 4 shows how the anorthite content of the plagioclase crystals changes with pressure and temperature. For the experiments that were seeded, the rims were analysed to determine the stable plagioclase composition. The anorthite content of plagioclase increases with increasing water pressure and rising temperature.
|
Glass
Glass compositions are rhyolitic and range from 71 to 79 wt % SiO2 (normalized to 100% anhydrous; Table 3). Using the mass balance calculation program PETMIX (Wright & Doherty, 1970), and assuming that K2O is incompatible, glass proportions are estimated to vary from 100 to 10 wt % dependent on experimental conditions (Table 2 and Fig. 5). There is a negative relationship between An content of plagioclase and K2O content of the glass (Fig. 6). High An contents can be reproduced at high- temperature and -pressure conditions, with high proportions of glass and thus low K2O contents. The seeded experiments that are within the stable plagioclase field differ slightly from the general trend, as they have elevated An contents relative to K2O content.
|
|
|
Blundy & Cashman (2001) have applied an approach first developed by Tuttle & Bowen (1958) for using the glass compositions to constrain the maximum water pressure of equilibrium of a melt, based on the ternary system of quartz–albite–orthoclase. The glass data (Fig. 7) show a linear trend from the starting composition away from albite, tracking towards quartz with decreasing experimental pressure and/or falling temperature. When silica saturation is attained the experimental glass compositions follow the silica–feldspar cotectic towards orthoclase enrichment. Several of the experimental glasses that have a crystalline silica phase do not plot at the expected cotectic pressure. This may reflect the problems with analysing glass in strongly crystalline samples, or may indicate that the projection scheme does not permit accurate estimation of pressure.
|
|
ANALYSIS OF NATURAL SAMPLES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Several Montserrat samples have been analysed to compare compositional trends in the plagioclase microlites and matrix glasses with experimental results to make inferences about crystallization conditions.
Plagioclase microlites
Between 20 and 35 microlites were analysed from each of 14 samples, and their dimensions (length and width) recorded. Figure 8 shows the data plotted as area of crystal against An content. The samples have been divided into three groups. The first and second groups are both dome samples collected from pyroclastic flow deposits related to dome collapse. The third group consists of pumice clasts erupted by Vulcanian explosions when magma discharge rates were high (Sparks et al., 1998; Druitt et al., 2002). No direct dome samples have been collected in the eruption and so it is not possible to know unambiguously whether a particular sample was erupted during periods of high or low magma discharge rate, and how long it was resident in the dome before collapse occurred. Samples, however, have been divided into those with a highly crystalline groundmass and those with a glassy groundmass. In general, dome samples from early in the eruption are much more variable in groundmass crystallinity than later in the eruption when discharge rates were generally higher (Sparks et al., 1998), but still very variable. The glassy dome samples are inferred to represent lava erupted at high magma discharge rate and with short residence time in the dome before collapse. Many plagioclase crystals have a narrow (<5 µm) rim that is compositionally distinct (Fig. 8). Generally, the microlites have a tabular morphology with two dimensions being approximately equal and the third dimension significantly longer. Crystals with a similar length to width dimension were not included in the data as they are likely to represent a plagioclase cut end-on rather than lengthwise.
|
All of the natural samples have a wide range of plagioclase compositions (An40–75). However, there are clear differences between the types of sample:
- pumices have a narrower range of An values, with all crystals being more calcic than An50. The highly crystalline dome samples have a much wider range of plagioclase compositions, which vary from An>75 to An<40.
- The pumices lack the small (<60 µm2) crystals that are numerous in most dome samples.
- Rims on the pumice plagioclase crystals are only slightly less anorthitic than the cores, and never less than An50. The larger crystals in the crystalline dome samples have rims, which generally have much lower An contents than their cores. These rims have similar An contents to the smallest crystals found in the same sample.
- In pumice and in dome samples, the anorthite content increases with crystal size.
- The glassy dome samples have intermediate characteristics between the pumice and crystalline dome samples. There are small crystals, but their An content does not extend to the low values observed in the crystalline dome samples. There are more sodic overgrowth rims on larger microlites, with a limited compositional range, analogous to the small crystals.
- Two samples are from prehistoric eruptions. MVO37 (3·6 ka pumice) and MVO25 (350 a dome lava) show similar relationships to the recent eruptive products.
Glass
Electron microprobe analyses of matrix glasses from the natural samples are rhyolitic (76·5–81·0 wt % SiO2, normalized to 100% anhydrous; Table 4). The glass proportions are estimated to vary from 34 to 10 wt % (Table 4) by mass balance of K2O. These calculations assume an assemblage of 70 (±3) wt % plagioclase containing 0·2 (±0·03) wt % K2O, 7 (±0·5) wt % amphibole containing 0·18 (±0·01) wt % K2O and whole-rock K2O contents from X-ray fluorescence (XRF) values (±3% relative; 0·023 wt %). These calculations are consistent with estimates determined by image analysis (Murphy et al., 2000; Couch et al., 2003). Reliable analysis of some samples proved impossible because of lack of sufficiently large matrix glass areas. Glass estimates for some of these samples by image analysis (Murphy et al., 2000) indicate glass proportions as low as 5 vol. %.
|
There is considerable variation in the glass composition and groundmass crystallinity of the dome lava samples, whereas the pumices show less variation (Table 4). The variation between two blocks from the same pyroclastic flow deposit is as great as the variation between lava blocks from different pyroclastic flow deposits. For example, glass contents for lava blocks from the January 1997 deposits vary from
5 to 25 wt %. The variability is attributed to two factors: first, dome collapses sample regions in the dome with different degassing and cooling histories and hence different crystallinities; second, pyroclastic flows can erode blocks deposited from previous flows (Cole et al., 2002). Some samples show heterogeneity in major element glass composition, where there appear to be two `glasses', such as analyses MVO58K and MVO58Na (Table 4). Similar observations have been made for Mount St. Helens (Cashman, 1992). These heterogeneities have been interpreted as an intimate mixture of a K-rich true glass, and very finely crystalline intergrowths of feldspar and quartz (Blundy & Cashman, 2001). The results on Montserrat samples support this interpretation. The K-rich zones contain chlorine, whereas the Na-rich zones have no chlorine, consistent with the former zones comprising glass, and the latter zones comprising volatile-free microcrystalline regions (Harford, 2000).
The ‘melt’ compositions of the natural samples follow the same trend as the experiments, when plotted on a Qz–Or–Ab ternary (Fig. 9), with increasing Qz away from Ab until silica saturation and then following the silica–feldspar cotectic. Some glasses from the crystalline dome samples plot further along the silica–feldspar cotectic than the experimental glasses. Also plotted are the compositions of melt inclusions from quartz and plagioclase (Devine et al., 1998a). The plagioclase melt inclusion composition fits well within the overall melt trend. However, the quartz inclusion composition is relatively enriched in silica compared with both experimental and natural melt compositions. The plagioclase melt inclusion is the least evolved (75·1 wt % SiO2) composition of all the natural samples analysed, and the least evolved melt composition is that of an explosion pumice (76·5% SiO2). Selected glass compositions from natural samples have been plotted against the most sodic plagioclase microlite composition in the sample (Fig. 6). As the melt composition becomes more evolved (K2O increases), the plagioclase becomes more sodic.
|
Plagioclase–amphibole thermometry
Amphibole phenocrysts in the andesite contain numerous plagioclase inclusions (Fig. 1d). Using the amphibole–plagioclase geothermometer of Holland & Blundy (1994), the temperature of coexisting plagioclase and adjacent amphibole can be estimated. The thermometer is estimated to be accurate to 2
±75°C in the range 400–1000°C, although the precision between inclusions is likely to be considerably better; 2 ±30°C (J. Blundy, personal communication, 2002). Many inclusions were analysed in selected amphiboles from several samples to investigate differences in temperature within and between samples. For each pair the analyses of the two minerals were made as close to the contact as possible. The thermometer utilizes two equations:
 |
 |
For silica-saturated rocks both thermometers can be used; however, for silica-undersaturated rocks only thermometer B can be used. As the upper stability of quartz for the bulk rock has been determined as <830°C at 130 MPa (Barclay et al., 1998), it is possible to determine which thermometer to use depending on the temperature result returned by the thermometers. A water pressure of 130 MPa was assumed based upon likely magma chamber conditions (Barclay et al., 1998; Murphy et al., 2000).
Table 5 summarizes the results and Fig. 10a and b shows SEM images of selected amphibole crystals with estimated temperatures for individual plagioclase inclusions analysed. Estimated temperature varies both within and between crystals. Most crystals show unsystematic variation in temperatures from 790 to 860°C. There is no clear relationship between amphibole crystal size and the mean temperature.
|
|
Figure 11 shows histograms of all temperature estimates and the mean temperature for each amphibole crystal analysed. The temperatures have a wide range from 764 to 935°C. The mean temperature, based on all analyses, is 846°C (2
of ±54°C). This is consistent with estimates of magma chamber temperatures by independent methods, such as Fe–Ti oxides (Devine et al., 1998a) and orthopyroxenes (Murphy et al., 2000). Crystals analysed from MVO37, a 3·7 ka pumice, and a crystal from MVO1217 have lower mean temperatures of 820–830°C. Plagioclase inclusions commonly range from An50 to An64 (Fig. 12). The mean plagioclase composition, based on all analyses, is An57 (2 of ±An8). At water pressures of 130 MPa, the plagioclase compositions can be used to infer temperatures of between 825 and 980°C, based on our experimental results (Fig. 4). A few temperature estimates in Table 5 exceed 900°C.
|
|
|
ESTIMATES OF LATENT HEAT |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
During magma ascent, volatile loss raises the liquidus temperature of the melt, inducing crystallization, which in turn releases latent heat. If crystallization occurs sufficiently fast, this heat cannot be lost, and therefore the temperature of the system will rise. Microlites thought to be related to decompression form up to 32% of the Montserrat andesite samples, so this heating could be significant. The temperature rise can be estimated utilizing existing thermodynamic data (Robie et al., 1978). Latent heat L is calculated by
 |
Hm is enthalpy of melting and 
is the heat capacity at the temperature of interest. The Hm for pure anorthite is 81 000 J/mol, and 59 280 J/mol for pure albite. We assume an initial T 860°C, the for pure An is 327·67 J/mol per K, and for pure Ab 318·13 J/mol per K. Calculations of the change in magma temperature during ascent from an initial temperature of 860°C, assuming that plagioclase (An50) is crystallized, indicate that temperature rises of up to 45°C are possible. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Experimental starting composition
There is a question as to whether aMon6a is a representative starting material. aMon6a cannot be a liquid at the inferred magma chamber conditions (840–870°C, 130 MPa, Barclay et al., 1998); at the inferred water pressures, the liquidus of aMon6a is
1050°C and the plagioclase liquidus is 950°C (Fig. 3). Therefore what is of importance is the melt composition of experiments run close to magma chamber conditions (860°C, 130 MPa). Experiment sc10 was run at 875°C, 125 MPa and its glass composition plots on the Qz–Ab–Or ternary in the cluster of points that intersect the 100 MPa cotectic line in Fig. 9. This composition is very similar to the least evolved melt compositions observed in the pumiceous samples, and is slightly more evolved than the melt inclusion composition from a plagioclase phenocryst. Experiments at lower water pressures and temperatures than sc10 all fall along the same trend as the natural samples. Hence, the relevant comparison of natural and experimental data is for experiments run at inferred conditions for the magma chamber and during magma ascent.
Magma chamber processes
Plagioclase–amphibole thermometry
The range of estimated temperatures (764–935°C) exceeds the precision range (±30°C, J. Blundy, personal communication, 2002), indicating that there are real variations recorded by the amphibole phenocrysts. The average range of most temperature estimates (840–870°C) agrees well with estimates based on other methods, suggesting that the absolute values of the temperature estimates are reasonable. The broad variation of temperatures around the estimated magma chamber conditions of 840–870°C and 130 MPa is consistent with the amphiboles predominantly forming at these conditions. The localized variation in temperatures together with variation in plagioclase inclusion compositions suggest that amphiboles and coexisting plagioclases grew in a fluctuating field of temperature and possibly PH2O. Taking the 2 value about the mean of 54°C for all data as reflecting these fluctuations suggests variations of the order of 100°C.
Several temperature estimates exceed 865°C, which is the experimentally determined upper stability limit of amphibole at 130 MPa water pressure (Barclay et al., 1998). The highest temperature estimated is 935°C, in a 350 a dome sample (MVO25). This temperature is in excess of the sum of 865°C and the precision (±30°C) and therefore either amphibole persists metastably above its stability limit or the amphibole formed at higher water pressures. Experimental work by V. Buckley (unpublished data, 2002) suggests that amphibole will break down within days if heated above its stability field at magma chamber pressures, limiting metastable amphibole preservation at these conditions to very short time-scales. Rutherford et al. (1998) experimentally determined that amphibole was stable in the Montserrat andesite to temperatures of 900°C at PH2O = 250 MPa. The Holland & Blundy geothermometers are dependent on pressure. However, the difference in calculated temperature of plagioclase–amphibole coexistence at 130 MPa and at 250 MPa is generally <5°C. For most of the inclusions, which have temperature estimates of >865°C, increasing water pressure would allow amphibole to be stable, based on the experimental phase boundaries determined by Rutherford et al. (1998).
The likely PH2O of each pair was estimated (Fig. 10c and d) using the temperature estimates from the geothermometers in conjunction with the experimentally determined contours of plagioclase anorthite content with pressure and temperature (Fig. 4). The experimental anorthite contours were derived from a more evolved composition than that which would have crystallized the amphibole phenocrysts. However, comparison of these experimental results with similar experiments performed using a bulk andesite composition (Rutherford et al., 1998) suggests similar anorthite contents at the same P–T conditions. The two amphibole crystals shown in Fig. 10c and d display an apparent range of estimated water pressures from 130 to 240 MPa. As with the temperature estimates, there is no systematic variation in estimated water pressure. The amphibole crystals could be interpreted to have formed over a large range of water pressures, corresponding to maximum depths of 10 km. The estimate of magma chamber conditions (130 MPa) may therefore constrain only the top of the chamber (Barclay et al., 1998). The observations of random inferred water pressures, with in some cases high estimated pressures at the crystal rim, could require a model in which the chamber is so large that it convects over a large depth range (5–10 km). Alternatively, the errors associated with these calculations may be substantial. Table 6 shows estimated pressure ranges for the crystals shown in Fig. 10. If the precision range of ±30°C is used, the range of estimated pressures for a given pair is limited, such that for crystals shown in Fig. 10c and d the difference between the highest pressure and lowest pressure estimate is significant (Table 6). These results suggest that some of the amphibole–plagioclase pairs formed at water pressures >130 MPa.
|
The overall consistency of the temperature ranges for each crystal implies that the crystallization of amphibole and plagioclase has remained in a similar thermal regime for considerable periods of time. The results also indicate that quartz and most of the amphibole crystals cannot be in equilibrium, as the amphiboles crystallized predominantly at temperatures in excess of the quartz liquidus (830°C).
Plagioclase microlites
Pyroxene geothermometry (Murphy et al., 2000) and Fe–Ti oxide geothermometry (Devine et al., 1998a) have been used to estimate the approximate temperature of the Montserrat andesite as 840–870°C. Barclay et al. (1998) determined that the water pressure of the magma chamber must be greater than 120 MPa, on the basis of H2O contents of melt inclusions. Hence if it is assumed that all the plagioclase microlites crystallized either in the magma chamber, or while ascending to the surface, the range of An values should be An55 (860°C, 130 MPa) to An<30 (860°C, <25 MPa) based on our experimental results. This expectation does not correspond to the plagioclase microlite compositions seen in natural samples, which range from An75 to An32. The low anorthite contents can be related to decompression and ascent, and will be discussed later. However, the high An values of the natural samples are difficult to reconcile with the proposed pressures and temperatures in the magma chamber.
The process of self-mixing (Couch et al., 2001) is invoked to explain the high anorthite contents observed in the plagioclase microlites. Mafic material is intruded into the magma chamber, ponding at the base. With time the mafic material heats the overlying adjacent andesite, forming a hot buoyant boundary layer. This layer is estimated to become unstable within tens of days. Hence plumes of hot andesite form; these rise through the cooler material, quench the hot magma and mix with adjacent andesite. By mixing packets of magma of the same composition but with different temperatures, the variation in thermal crystallization history on a local scale can be explained. Our experimental data indicate that parts of this boundary layer must have exceeded 970°C to precipitate >An70 microlites and overgrowth rims on phenocrysts by this process.
Magma mixing is a possible explanation for the heating and high-temperature crystallization observed in the magma. The textures of the mafic inclusions in the andesite show characteristic features of magmatic quenching of hydrous basalt against cooler andesite (Murphy et al., 2000). These inclusions could be a source of calcic microlites. However, the mafic inclusions contain abundant fine-grained acicular amphibole with an entirely different composition compared with the large amphibole phenocrysts in the andesite (Murphy et al., 2000). Mixing of this magma into the andesite should produce abundant acicular groundmass amphibole xenocrysts, but these are not observed in the andesite.
Further evidence for localized but pervasive heating comes from detailed study of Fe–Ti oxides by Devine et al. (1998a), which suggests that material was heated to 880°C. There are reverse zoned rims on orthopyroxene crystals, which have calculated temperatures of crystallization of 880–1050°C (Murphy et al., 2000). There is textural evidence of quartz crystals that have resorbed edges, which could be due to reheating above stable quartz conditions.
Another possible cause of the high anorthite plagioclase microlites is that they formed at water pressures >130 MPa. The current estimate of magma chamber water pressure constrains only the top of the chamber to 130 MPa. As discussed earlier when considering the geothermometry results of plagioclase–amphibole pairs, at least some of the crystallization may have taken place at greater water pressures. At greater pressure, more anorthite-rich plagioclase can crystallize at a given temperature (Fig. 4).
The relationship between plagioclase microlite size and anorthite content in the Montserrat samples can be explained in part by convective self-mixing. The cores of larger plagioclase microlites crystallized at elevated temperatures as reflected by their high anorthite content, along with the Ca-rich overgrowth rims on the large plagioclase phenocrysts. The amount of time available for crystallization at these conditions was not long [based on the experimental results of Couch et al. (2003) probably several days], allowing microlites and overgrowth rims to grow, but not phenocrysts. The medium-sized microlites, which have anorthite contents of An50, could have crystallized at the temperature–pressure conditions estimated for the bulk andesite (860°C, 130 MPa), after the heating event.
Petrological studies of Mount Pelée (Martel et al., 2000) have identified high anorthite content plagioclase microlites, which are also likely to have formed in the magma chamber during heating before eruption. It is therefore important to consider microlite crystallization as a process that will occur both in the magma chamber and during ascent.
Petrological and theoretical studies by Stewart & Fowler (2001) have suggested that plagioclase phenocryst overgrowth rims grew shortly before eruption, which agrees with this study. However, those workers proposed that the overgrowth rims formed during ascent and started to grow at a maximum depth of 900 m. The experimental results of this study show that this interpretation cannot be correct. Overgrowths with compositions of An<30 would be expected rather than the observed high An contents of overgrowth rims. Furthermore, if phenocryst overgrowth rim formation was an ascent-related process, we would expect to see a correlation between ascent rate and overgrowth rim thickness. Plagioclase phenocrysts from pumice samples would lack rims, as the rapid ascent would prevent growth (Couch et al., 2003), whereas dome samples, which ascended slowly, would have overgrowth rims. No such relationship is observed.
The similarity in plagioclase microlite sizes and compositions between recent samples and prehistoric samples (MVO37 and MVO25) suggests that processes such as self-mixing have been a long-term feature of the magma chamber conditions. This has implications for the likely eruptive behaviour in the future, implying that if the magma retains a similar composition, material will continue to ascend and erupt in the ways observed during the current eruption.
Ascent processes
Plagioclase
During ascent, decompression-induced crystallization can occur in andesite magmas, with plagioclase as the predominant crystallizing phase. The kinetics of plagioclase nucleation and growth are sluggish, such that there can be a delay between depressurization and water exsolution, and the onset of crystallization (see Couch et al., 2003). This delay can be used to understand differences in groundmass textures and compositions, and to infer variations in their likely ascent history.
Magma that ascends sufficiently rapidly experiences no crystallization during depressurization owing to the nucleation delay (Couch et al., 2003). Pumices have a groundmass containing large microlites (>100 µm2 crystal area) with high anorthite contents (>An50). Comparison with experimental results indicates that microlite groundmass crystallization took place at elevated PH2O, in the magma chamber, with enough time for large microlites to grow. Ascent must have been sufficiently rapid to prevent decompression-related crystallization. Dome samples contain a wide range of plagioclase microlite sizes and compositions. The large, high-anorthite crystals could have formed at elevated pressures in the magma chamber. The numerous smaller, sub-100 µm2 area crystals must have formed at lower pressures in the conduit during ascent, reflected in their lower anorthite contents (<An50). The most sodic plagioclase microlites analysed have a composition of An32, and this can be interpreted as the crystallization products of magma ascending to very shallow levels.
Glass
The similarity in melt composition between the pumice samples and the experimental glasses for the inferred magma chamber conditions provides evidence that the rapidly decompressed samples did not undergo significant crystallization during ascent. In some cases the melt compositions from the pumices are less evolved than the experimental magma chamber melt (Fig. 9). This is consistent with heating in the magma chamber, resulting in melting of crystals and generation of a less evolved melt.
The dome samples are variable in composition, from those similar to the experimental magma composition, to others that have crystallized a silica phase. This variation is due to the broad range of groundmass crystallinities reflecting decompression-induced crystallization. The most evolved compositions sit further along the silica–feldspar cotectic than any experimental compositions, indicating more extensive crystallization at either low temperatures or low pressures than the experimental conditions (<825°C or <25 MPa). The results indicate that the crystallization of quartz is confined to the shallowest parts of the conduit and dome itself, where the water pressure is very low. Cooling is not thought to be significant in the crystallization of the dome (Sparks et al., 2000), because conductive cooling affects only the outermost parts of the dome in a crust estimated to be only metres in thickness. Although it is possible that some samples come from the cooled crust of the dome, the volumetric proportion of crust in a major dome collapse is likely to be negligible. Crystallization of quartz is also inferred to be confined to samples formed during periods of low discharge or that have long residence time in the dome before a collapse. These concepts agree with observations by Baxter et al. (1999) that cristobalite is found in substantial amounts only in crystalline dome materials, and is absent in explosively erupted pumice and ash.
Relationship to eruptive conditions
The difference in composition and crystallinity observed between samples is interpreted as a consequence of differences in eruptive conditions and magma residence time at shallow levels. Pumice clasts are the most K2O-poor, least crystalline samples, erupted during explosive eruptions and some dome eruptions. During Vulcanian explosions newly arrived magma from the uppermost several hundred metres was probably evacuated in 1–2 min, giving insufficient time for crystallization. During such eruptions, magma extrusion rate is estimated at 9 m3/s (Druitt et al., 2002). For pumice, magma ascent from an estimated depth of >5 km (Barclay et al., 1998) to the surface, assuming a conduit area of 700 m2 (Druitt et al., 2002), would have taken 4 days. During the sub-Plinian eruption of 17 September 1996, when it is inferred that the majority of the conduit was evacuated (Robertson et al., 1998), magma ejected as pumice may have spent as little as an hour in transit to the surface.
Magmas erupted as part of the lava dome would have experienced a relatively slow ascent. These samples are more crystalline, with estimated glass contents of 10–31 wt %. Magma ascent from the chamber would have taken between 5 and 80 days for typical extrusion rates of 8–1 m3/s, assuming a conduit area of 700 m2 (Druitt et al., 2002). Furthermore, these samples would have spent an unknown time period in the lava dome (days to months). The amount of crystallization therefore reflects both ascent time and residence time in the dome before collapse.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
By combining experimental work with detailed study of natural samples, constraints can be placed on the conditions of crystallization of the groundmass. There are several conclusions:
- the composition of natural plagioclase microlites is dependent upon their size. Large microlites have higher anorthite contents than the smaller crystals. Dome samples display a wide range of microlite sizes, with pumice samples lacking small crystals (<60 µm2). In the dome samples, the large microlites have overgrowth rims of significantly less anorthitic plagioclase.
- Microlites in natural samples are similar to those determined by equilibrium experiments at the inferred conditions in the magma chamber.
- Geothermometry of plagioclase inclusions within amphibole phenocrysts agrees with previous estimates of magma chamber temperatures and supports the idea that there has been a complex and variable crystallization history.
- Latent heat released as a result of microlite crystallization can raise magma temperature significantly (by up to 45°C).
- Comparison of experimental and natural glasses indicates that much groundmass crystallization can occur as a result of decompression during magma ascent. Comparison of natural samples and experiments shows that crystalline silica bearing, highly crystalline dome lava was at equilibrium at pressures of
25–50 MPa, representing depths significantly less than 2 km.
- estimates of the pressure conditions for the coexistence of plagioclase inclusions within amphibole suggest a complex, polybaric crystallization history.
- The mechanism of convective self-mixing might explain the high anorthite content of the plagioclase microlites (Couch et al., 2001). Hot andesite, reheated by the intrusion of mafic material at the base of the magma chamber, is quenched to form microlites and overgrowth rims against the cooler interior. The extra heat allows plagioclase with a much higher anorthite content to crystallize, explaining the large, An-rich microlites seen throughout the samples. Convective self-mixing does not require direct involvement of another more mafic magma to generate the complex textures commonly observed in porphyritic orogenic intermediate and silicic volcanic rocks, although mafic magma emplaced at the base of the magma chamber provides the likely heat source. The concept reconciles otherwise puzzling petrological features, and also supports the notion that magma chambers can convect vigorously.
- The more sodic overgrowth rims and small microlites (<60 µm2) are interpreted to be the result of decompression-induced crystallization. Magma ascent rate determines the microlite population both texturally and compositionally. Rapidly decompressed samples, such as pumices, had no time to crystallize during ascent, resulting in microlites with limited, An-rich compositions, and an absence of small crystals. More slowly decompressed samples had sufficient time for nucleation and growth of small An-poor crystals and the growth of thin rims on existing crystals.
- The similarity between prehistoric samples and recently erupted samples suggests that similar magma chamber and conduit processes have occurred at Soufrière Hills for at least 3·6 ka.
|
ACKNOWLEDGEMENTS |
|---|
Thanks are due to Jon Blundy and Jenni Barclay for useful discussions, and Richard Brooker for assistance with TZM experiments. Members of the Montserrat Volcano Observatory are acknowledged for providing the samples used in this study. Thorough reviews by Alison Pawley and Caroline Martel are appreciated and helped to improve the manuscript. S.C. and C.L.H. acknowledge NERC studentships, and R.S.J.S. an NERC professorship and MRC support from Gruppo Nazionale de Vulcanologia.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSEXPERIMENTAL RESULTSANALYSIS OF NATURAL SAMPLESESTIMATES OF LATENT HEATDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Barclay, J., Rutherford, M. J., Carroll, M. R., Murphy, M. D., Devine, J. D., Gardner, J. E. & Sparks, R. S. J. (1998). Experimental phase equilibria constraints on pre-eruptive storage conditions of the Soufrière Hills magma. Geophysical Research Letters 25, 3437–3440.[CrossRef][Web of Science]
Baxter, P. J., Bonadonna, C., Dupree, R., Hards, V. L., Kohn, S. C., Murphy, M. D., Nichols, A., Nicholson, R. A., Norton, G. E., Searl, A., Sparks, R. S. J. & Vickers, B. P. (1999). Cristobalite in volcanic ash of the Soufrière Hills Volcano, Montserrat, British West Indies. Science 283, 1142–1145.
Blundy, J. & Cashman, K. V. (2001). Magma ascent and crystallisation at Mount St Helens, 1980–1986. Contributions to Mineralogy and Petrology 140, 631–650.[Web of Science]
Carroll, M. R. & Blank, J. G. (1997). The solubility of H2O in phonolitic melts. American Mineralogist 82, 549–556.[Abstract]
Cashman, K. V. (1992). Groundmass crystallisation of Mount St Helens dacite, 1980–1986: a tool for interpreting shallow magmatic processes. Contributions to Mineralogy and Petrology 109, 431–449.[CrossRef][Web of Science]
Cole, P., Calder E. S., Sparks, R. S. J., Clarke, A. B., Druitt, T. H., Young, S. R., Herd, R. A., Harford, C. L. & Norton, G. E. (2002). Deposits from dome-collapse and fountain-collapse pyroclastic flows at Soufrière Hills Volcano, Montserrat. In: Druitt, T. H. & Kokelaar, B. P. (eds) The Eruption of the Soufrière Hills Volcano, Montserrat, from 1995 to 1999. Geological Society, London, Memoirs 21, 231–262.
Couch, S., Sparks, R. S. J. & Carroll, M. R. (2001). Mineral disequilibrium in lavas explained by convective self-mixing in open magma chambers. Nature 411, 1037–1039.[CrossRef][Medline]
Couch, S., Sparks, R. S. J. & Carroll, M. R. (2003). The kinetics of degassing-induced crystallization at Soufrière Hills Volcano, Montserrat. Journal of Petrology 44, 1477–1502.
Devine, J. D., Murphy, M. D., Rutherford, M. J., Barclay, J., Sparks, R. S. J., Carroll, M. R., Young, S. R. & Gardner, J. E. (1998a). Petrologic evidence for pre-eruptive pressure–temperature conditions, and recent reheating, of andesitic magma erupting at the Soufrière Hills volcano, Montserrat, WI. Geophysical Research Letters 25, 3669–3672.[CrossRef][Web of Science]
Devine, J. D., Rutherford, M. J. & Gardner, J. E. (1998b). Petrologic determination of ascent rates for the 1995–1997 Soufrière Hills volcano andesitic magma. Geophysical Research Letters 25, 3673–3676.
Druitt, T. H., Young, S. R., Baptie, B., Bonadonna, C., Calder, E. S., Clarke, A. B., Cole, P., Harford, C., Herd, R. A., Luckett, R., Ryan, G. & Voight, B. (2002). Episodes of repetitive Vulcanian explosions and fountain collapse at Soufrière Hills volcano, Montserrat. In: Druitt, T. H. & Kokelaar, B. P. (eds) The Eruption of Soufrière Hills Volcano, Montserrat, from 1995 to 1999. Geological Society, London, Memoirs 21, 281–306.
Edmonds, M., Pyle, D. M. & Oppenheimer, C. (2001). A model for degassing at the Soufrière Hills volcano, Montserrat, West Indies, based on geochemical data. Earth and Planetary Science Letters 186, 159–173.[CrossRef][Web of Science]
Gardner, J. E., Hilton, M. & Carroll, M. R. (2000). Bubble growth in highly viscous melts during continuous decompression from high pressure. Geochimica et Cosmochimica Acta 64, 1473–1483.[CrossRef][Web of Science]
Hammer, J. E., Cashman, K. V. & Voight, B. (2000). Magmatic processes revealed by textural and compositional trends in Merapi dome lavas. Journal of Volcanology and Geothermal Research 100, 165–192.[CrossRef][Web of Science]
Harford, C. L. (2000). The volcanic evolution of Montserrat. Ph.D. thesis, University of Bristol.
Harford, C. L. & Sparks, R. S. J. (2001). Recent remobilisation of shallow-level intrusions on Montserrat revealed by hydrogen isotope composition of amphibole. Earth and Planetary Science Letters 185, 285–297.[CrossRef][Web of Science]
Holland, T. H. & Blundy, J. (1994). Non-ideal interactions in calcic amphiboles and their bearing on amphibole–plagioclase thermometry. Contributions to Mineralogy and Petrology 116, 433–447.
Martel, C., Bourdier, J.-L., Pichavant, M. & Traineau, H. (2000) Textures, water content and degassing of silicic andesites from recent plinian and dome-forming eruptions at Mount Pelée volcano (Martinique, Lesser Antilles arc). Journal of Volcanology and Geothermal Research 96, 191–206.[CrossRef][Web of Science]
Morgan, G. B. & London, D. (1996). Optimising the electron microscope analysis of hydrous alkali aluminosilicate glasses. American Mineralogist 81, 1176–1185.[Abstract]
Murphy, M. D., Sparks, R. S. J., Barclay, J., Carroll, M. R. & Brewer, T. S. (2000). Remobilisation of andesite magma by intrusion of mafic magma at the Soufrière Hills volcano, Montserrat, West Indies. Journal of Petrology 41, 21–42.
Robertson, R. E., Cole, P., Sparks, R. S. J., Harford, C., Lejeune, A.-M., McGuire, W. J., Miller, A. D., Murphy, M. D., Norton, G. E., Stevens, N. S. & Young, S. R. (1998). The explosive eruption of Soufrière Hills volcano, Montserrat. Geophysical Research Letters 25, 3689–3692.
Robertson, R. E., Aspinall, W. P., Herd, R. A., Norton, G. E., Sparks, R. S. J. & Young, S. R. (2000). The 1995–1998 eruption of the Soufrière Hills volcano, Montserrat, WI. Philosophical Transactions of the Royal Society of London, Series A 358, 1619–1637.[CrossRef]
Robie, R. A., Hemingway, B. S. & Fisher, J. R. (1978). Thermodynamic properties of mineral and related substances at 298·15 K and 1 bar (105 Pascals) pressure and at higher temperatures. US Geological Survey Bulletin 1452, 1–453.
Rouse, P. J. (2000). Experimental phase equilibria of sodic phonolites from Montana Blanca, Tenerife. Ph.D. thesis, University of Bristol.
Rutherford, M. J., Devine, J. D. & Barclay, J. (1998). Changing conditions and ascent rates during the Soufrière Hills eruption on Montserrat. Geological Society of America Today 8, 1–7.
Sparks, R. S. J., Young, S. R., Barclay, J., Calder, E. S., Cole, P., Darroux, B., et al. (1998). Magma production and growth of the lava dome of the Soufrière Hills volcano, Montserrat, West Indies: November 1995 to December 1997. Geophysical Research Letters 25, 3421–3424.[CrossRef][Web of Science]
Sparks, R. S. J., Murphy, M. D., Lejeune, A.-M., Watts, R., Barclay, J. & Young, S. R. (2000). Control on the emplacement of the andesite lava dome of the Soufrière Hills volcano by degassing-induced crystallisation. Terra Nova 12, 14–20.[CrossRef][Web of Science]
Stewart, M. L. & Fowler, A. D. (2001). The nature and occurrence of discrete zoning in plagioclase from recently erupted andesitic volcanic rocks, Montserrat. Journal of Volcanology and Geothermal Research 106, 243–253.[CrossRef][Web of Science]
Taylor, J. R., Wall, V. J. & Pownceby, M. I. (1992). The calibration and application of accurate redox sensors. American Mineralogist 77, 284–295.[Abstract]
Tuttle, O. F. & Bowen, N. L. (1958). Origin of Granite in the Light of Experimental Studies in the System NaAlSi3O8–KAlSi3O8–SiO2–H2O. Geological Society of America, Memoir 74.
Wright, T. L. & Doherty, P. C. (1970). A linear programming and least squares computer method for solving petrologic problems. Geological Society of America Bulletin 81, 1995–2008.
Young, S. R., Sparks, R. S. J., Aspinall, W. P., Lynch, L., Miller, A. D., Robertson, R. E. & Shepherd, J. B. (1998). Overview of the eruption of Soufrière Hills volcano, Montserrat, 18 July to December 1997. Geophysical Research Letters 25, 3389–3392.[CrossRef][Web of Science]
Zellmer, G. F., Sparks, R. S. J., Hawkesworth, C. J. & Wiedenbeck, M. (2003). Magma emplacement and remobilization timescales beneath Montserrat: insights from Sr and Ba zonation in plagioclase phenocrysts. Journal of Petrology 44, 1413–1431.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Genareau, A. B. Clarke, and R. L. Hervig New insight into explosive volcanic eruptions: Connecting crystal-scale chemical changes with conduit-scale dynamics Geology, April 1, 2009; 37(4): 367 - 370. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Lange, H. M. Frey, and J. Hector A thermodynamic model for the plagioclase-liquid hygrometer/thermometer American Mineralogist, April 1, 2009; 94(4): 494 - 506. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Blundy and K. Cashman Petrologic Reconstruction of Magmatic System Variables and Processes Reviews in Mineralogy and Geochemistry, January 1, 2008; 69(1): 179 - 239. [Full Text] [PDF]
|
|
|
|
 |
|
 B. Charoy and P. Barbey Ferromagnesian silicate association in S-type granites: the Darongshan granitic complex (Guangxi, South China) Bulletin de la Societe Geologique de France, January 1, 2008; 179(1): 13 - 27. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Jerram and V. M. Martin Understanding crystal populations and their significance through the magma plumbing system Geological Society, London, Special Publications, January 1, 2008; 304(1): 133 - 148. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. G. Lensky, R. S. J. Sparks, O. Navon, and V. Lyakhovsky Cyclic activity at Soufriere Hills Volcano, Montserrat: degassing-induced pressurization and stick-slip extrusion Geological Society, London, Special Publications, January 1, 2008; 307(1): 169 - 188. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Genareau, R. Hervig, and A. Clarke Geochemical variations in late-stage growth of volcanic phenocrysts revealed by SIMS depth-profiling American Mineralogist, August 1, 2007; 92(8-9): 1374 - 1382. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Viccaro, C. Ferlito, and R. Cristofolini Amphibole crystallization in the Etnean feeding system: mineral chemistry and trace element partitioning between Mg-hastingsite and alkali basaltic melt European Journal of Mineralogy, July 1, 2007; 19(4): 499 - 511. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Martel, A. Radadi Ali, S. Poussineau, A. Gourgaud, and M. Pichavant Basalt-inherited microlites in silicic magmas: Evidence from Mount Pelee (Martinique, French West Indies) Geology, November 1, 2006; 34(11): 905 - 908. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. ANNEN, J. D. BLUNDY, and R. S. J. SPARKS The Genesis of Intermediate and Silicic Magmas in Deep Crustal Hot Zones J. Petrology, March 1, 2006; 47(3): 505 - 539. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. N. WILSON, S. BLAKE, B. L. A. CHARLIER, and A. N. SUTTON The 26{middle dot}5 ka Oruanui Eruption, Taupo Volcano, New Zealand: Development, Characteristics and Evacuation of a Large Rhyolitic Magma Body J. Petrology, January 1, 2006; 47(1): 35 - 69. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Couch and S. Couch Experimental investigation of crystallization kinetics in a haplogranite system American Mineralogist, October 1, 2003; 88(10): 1471 - 1485. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. COUCH, R. S. J. SPARKS, and M. R. CARROLL The Kinetics of Degassing-Induced Crystallization at Soufriere Hills Volcano, Montserrat J. Petrology, August 1, 2003; 44(8): 1477 - 1502. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||