| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 43 | Number 12 | Pages 2279-2303 | 2002
© Oxford University Press 2002
Mixed Magmas, Mush Chambers and Eruption Triggers: Evidence from Zoned Clinopyroxene Phenocrysts in Andesitic Scoria from the 1995 Eruptions of Ruapehu Volcano, New Zealand
1DIVISION OF EARTH AND PLANETARY SCIENCES, GRADUATE SCHOOL OF SCIENCES, HOKKAIDO UNIVERSITY, N10 W8 KITA-KU, SAPPORO, 060-0810, JAPAN
2DEPARTMENT OF EARTH SCIENCES, HOKKAIDO UNIVERSITY OF EDUCATION, ASAHIKAWA, JAPAN
3INSTITUTE OF GEOLOGICAL AND NUCLEAR SCIENCES, WAIRAKEI RESEARCH CENTRE, TAUPO, NEW ZEALAND
Received August 22, 2000; Revised typescript accepted May 23, 2002
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGENERAL PETROGRAPHY AND MINERAL...SAMPLES AND EXPERIMENTAL...TYPES OF ZONING PROFILE...DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Juvenile ejecta from the September and October 1995 eruptions of Ruapehu volcano, New Zealand, indicate that mixing occurred between relatively higher- and lower-temperature (high-T and low-T) andesitic magmas. Compositional zonations in clinopyroxene phenocrysts provide direct evidence for a pre-eruption crystal–melt mush chamber containing low-T magma, and elucidate the processes of magma mixing and eruption, following the injection of high-T magma. Many phenocrysts with Fe-rich cores derived from low-T magma have extremely reverse zoned mantles around slightly resorbed cores. Mg-value [100Mg/(Mg + Fe)] increases from 65–70 to
85 over a short width (<20 µm) at the inner edge of the mantle, usually accompanied by an abrupt change of Cr2O3 and Wo contents. After the first extreme reverse zoning, high Mg-value in the crystal mantle usually continues over several tens to 200 µm in width, with intermittent reversals, and then Mg-value decreases towards the outer rim. The high Mg-value (85) is the same as the core compositions of phenocrysts derived from the high-T magma, suggesting that these extremely zoned phenocrysts were surrounded immediately by invading high-T magma, as opposed to mixed melt. This does not indicate a simple melt-mixing process between crystal-rich magmas, but is interpreted as evidence for mixing in a mush chamber. It is suggested that high-T magma was injected into a mushy low-T magma, then the denser and hotter high-T magma displaced the lighter and cooler interstitial melt in the mush. In addition, the high-T magma could thermally erode crystals in the mush network. Hence, eroded crystals from the mush were captured by the high-T magma. Repeated reverse zonations in the crystal mantles, and variable mantle widths, indicate that injection of high-T magma intermittently continued, both to form and to enlarge magma pockets in the mush. In the process, melt mixing occurred owing to convection and/or conjunction and coalescence of the magma pockets. The mixing between the magma pockets and interstitial melt of the mush gave rise to normal zoning in the mantles of contained phenocrysts. Zoning in the outermost rims of pyroxenes differs between the September and October scorias. Almost all of the October phenocrysts, but few of the September phenocrysts, have thin (several micrometres) reverse zoned rims. Therefore we suggest that the October eruptions may have been triggered by intensive injection just before the eruption, whereas the September eruptions may have been triggered by either smaller-scale injections or coalescence of magma pockets. These differences in presumed triggering process are consistent with geophysical observations before and during the 1995 eruptions. KEY WORDS: magma mixing; magma chamber; mush chamber; compositional mineral zonation; Ruapehu volcano
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL PETROGRAPHY AND MINERAL...SAMPLES AND EXPERIMENTAL...TYPES OF ZONING PROFILE...DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Compositions of phenocryst minerals and their compositional zoning in volcanic rocks record magmatic processes in subvolcanic magma chambers before and during eruption. On the basis of the hypothesis that core compositions of phenocrysts reflect original melt compositions and magma temperature, petrological analysis of a range of phenocrysts can provide data that elucidate magma-mixing processes (Eichelberger, 1975; Sakuyama, 1979, 1981). On the other hand, it has been proposed that some minerals represent in situ crystallization in a chamber, and are mechanical additions to the erupting magma, not ‘true’ phenocrysts (suspended crystals in melt) (Fujimaki, 1982; Tait, 1988; de Silva, 1989; Nakada et al., 1994). Recently, several studies have focused on compositional zoning of phenocrysts and revealed, in more detail, magma-mixing processes (Nakamura, 1995; Hattori & Sato, 1996; Umino & Horio, 1998) or fractionation processes in a closed-system magma chamber solidifying from the margins inwards (Singer et al., 1995; Brophy et al., 1996; Kuritani, 1998). Similarly detailed petrological analysis of eruptive products from observed and monitored eruptive episodes ought to provide strong constraints on the structure of magma chambers and magmatic processes, linked to actual eruptions. However, there appear to be few such studies on record, and they deal with lava effusion (Paricutin volcano: McBirney et al., 1987; Kilauea volcano: Garcia et al., 2000), dome formation (Unzen volcano: Nakamura, 1995), and large silicic eruptions (Pinatubo volcano: Pallister et al., 1992; Hattori & Sato, 1996). These examples could not cover variable volcanic eruptions and various types of eruptive magmas.
Ruapehu volcano is one of five, large, dominantly andesitic, composite volcanoes (Hackett & Houghton, 1989) in the Tongariro Volcanic Centre, at the southern end of the Taupo Volcanic Zone, New Zealand (Fig. 1). Numerous small phreatic and phreatomagmatic explosions have occurred through the summit crater, occupied by Crater Lake. At least eight larger discrete eruptions or eruptive sequences (1945, 1966, 1969, 1971, 1975, 1977, 1995 and 1996) have produced juvenile magma. Individual eruptions are small (erupted magma volume <0·1 km3) and characterized by intermittent phreatomagmatic and strombolian phases (Houghton et al., 1987; Nairn & Ruapehu Surveillance Group, 1996).
|
During September and October 1995, and June and July 1996, Ruapehu volcano erupted intermittently. The activity was monitored geochemically and geophysically (Nairn et al., 1996; Bryan & Sherburn, 1999), and dated samples were collected from three major eruptions. Nakagawa et al. (1999) showed whole-rock and matrix glass chemistry, and core and rim compositions of phenocryst minerals of these samples, and documented their petrological variations. They concluded that these juvenile materials were probably derived from similar but distinct, mixed magma batches. They also proposed the hypothesis that the batches were formed by repeated injections of several types of relatively high-temperature magmas (ranging from 1000 to 1200°C) into a shallower low-temperature (1000°C) crystal mush chamber, resulting in the formation of several discrete magma pockets in a shallower magma storage system beneath the volcano.
Such a crystal-rich chamber, which consists of crystals (25–55 vol. %) and interstitial melt, has been proposed in a magma storage system beneath a mid-ocean ridge (e.g. Sinton & Detrick, 1992), and for arc volcanoes (e.g. Nakamura, 1995). However, there was actually little direct evidence for the shallower mushy chamber in the eruptions of Ruapehu volcano, excepting perhaps the porphyritic nature of the erupted andesites, and the presence of gabbroic inclusions (Nakagawa et al., 1999). In addition, Nakagawa et al. (1999) were not been able to interpret the whole sequence of the 1995 eruptive activity, which includes the following features (Bryan & Sherburn, 1999): (1) precursor activity (compositional changes and rising temperature of Crater Lake water, increasing numbers of volcanic earthquakes and intensity of tremor) began in April 1995, accompanied by steam eruptions; (2) the eruptions occurred in two episodes, 18–25 September and 7–31 October, separated by a dormant period. Precursor activity before each eruption episode differed. The September eruptions were preceded by tremor without significant volcanic earthquakes, but both occurred during the eruption and continued at much lower level in the quiescent period from 26 September until 7 October. On 7 October, a series of larger volcanic earthquakes occurred preceding the October eruption episode.
Compositional zonations in phenocryst minerals may possibly record the magmatic processes during the above precursor activity in magma chambers and during eruptions. Although Nakagawa et al. (1999) discussed compositional difference in core and rim of phenocryst minerals, they did not investigate in detail the zonation in each phenocryst. Here, we present new data for zoning profiles of many clinopyroxene phenocrysts in andesitic scoria from three major dated eruptions in 1995, to focus on the following two problems. First, does zoning provide strong evidence for the presence of a crystal mush chamber, and the processes occurring in it during precursor activity and eruption? Second, as geophysical monitoring has revealed differences in precursor activity between the September and October eruptions, were there differences in magmatic processes between the two eruption episodes, which are reflected in the crystallization of phenocrysts?
|
GENERAL PETROGRAPHY AND MINERAL CHEMISTRY OF THE 1995 ERUPTIVE SCORIAS |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL PETROGRAPHY AND MINERAL...SAMPLES AND EXPERIMENTAL...TYPES OF ZONING PROFILE...DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The juvenile magmas of the 1995 eruptions were all andesites, with similar chemistry of SiO2 57·7–58·5%, MgO 5·5–4·5% and K2O 1·3–1·4% (water-free 100% normalized value) (Table 1). Gamble et al. (1999) reported that dacitic ejecta (SiO2 >60%) were also present during initial eruptions (18 and 23 September 1995), but Wood et al. (1998) suggested that the dacite was a product of interaction (mixing) between juvenile magmas and hydrothermally altered materials comprising Crater Lake sediment. All the juvenile andesites are petrographically similar (modal volume of phenocrysts 25–37%), containing plagioclase, orthopyroxene and augite phenocrysts (Table 1) in a glassy to hyalopilitic matrix with groundmass plagioclase and pyroxenes. There are no oxide minerals either as phenocryst or as groundmass phases in the andesites. Matrix glass compositions are dacitic (Table 2), showing relatively wide compositional variations of SiO2 62·5–70%, MgO 3·2–0·9% and K2O 1·8–3·4% (water-free 100% normalized value). The phenocryst and groundmass minerals show disequilibrium features, interpreted as evidence of magma mixing by Nakagawa et al. (1999), as follows:
|
|
- wide compositional distributions in phenocryst cores (Figs 2 and 3);
- pyroxene phenocrysts with iron-rich cores have more magnesian rims, whereas those with more magnesian cores show normal zonation toward iron-rich rims;
- groundmass plagioclase is not less calcic than phenocryst plagioclase (Fig. 3).
|
|
These minerals cannot have crystallized from a single magma batch by simple cooling. In addition, least-squares mixing calculations suggest that variations in whole-rock and matrix glass chemistry cannot result from fractionation and/or accumulation of the phenocryst minerals. Thus these observations indicate that the 1995 andesites are mixing products of compositionally different magmas with distinct magmatic temperatures.
Also, whole-rock and matrix glass chemistries show linear relationships in oxide–oxide variation diagrams, being consistent with mixing products between two end-member magmas, low-temperature (silicic) and high-temperature (mafic) magmas, designated low-T and high-T (Nakagawa et al., 1999). The original phenocryst minerals in each magma are as follows. The high-T end-member magmas have magnesian pyroxenes with Mg-number 85 in clinopyroxene and Mg-number 82 in orthopyroxene [where Mg-number is 100Mg/(Mg + Fe)], and calcic plagioclase (An 80), whereas the low-T magmas have less magnesian pyroxenes (Mg-number 70 in clinopyroxene and 67 in orthopyroxene) and less calcic plagioclase (An 60). If mixing between the two end-member magmas occurred, the above disequilibrium features in mineral chemistry can be explained.
Although all the andesites have similar petrographic and chemical features, the September and October samples can be distinguished by their whole-rock chemistry, matrix glass chemistry and mineral chemistry (Nakagawa et al., 1999). Magmatic temperatures of two end-member magmas were estimated by pyroxene geothermometry (Lindsley, 1983). Although estimated magmatic temperatures of three samples of low-T end-member magmas are all 1000°C, the high-T end-member magmas show differences between September (1200°C) and October (1100°C) samples (Nakagawa et al., 1999). The data are inconsistent with successive discharge from a single mixed or zoned magma chamber, and suggest that each eruption episode evacuated distinct and separated magma pockets. Nakagawa et al. (1999) suggested that repeated injections of the high-T magmas into a low-T mushy magma storage system could have resulted in the separated magma pockets. This proposed process is consistent both with petrological variations among these dated andesites and the occurrence of gabbroic crystal aggregate in the 1995 andesites.
|
SAMPLES AND EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL PETROGRAPHY AND MINERAL...SAMPLES AND EXPERIMENTAL...TYPES OF ZONING PROFILE...DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In this study, we focus on compositional zoning of clinopyroxene phenocrysts of three samples (Table 1) from separate 1995 eruptions (18 and 23 September and 11 October: see Nakagawa et al., 1999). Whole-rock and matrix glass chemistry, and modal analysis are shown in Tables 1 and 2. Both plagioclase and orthopyroxene phenocrysts show complex zoning. Plagioclase compositions are known to reflect not only magmatic chemistry and temperature, but also magma volatile contents (Housh & Luhr, 1991; Hattori & Sato, 1996), so zoned phenocrysts may record degassing processes during ascent and eruption, as well as magma-mixing processes. In contrast, pyroxenes are largely unaffected by degassing. Some orthopyroxene phenocrysts are, however, surrounded by clinopyroxene overgrowth, indicating they were not always in contact with melt. On the other hand, no reaction relations between clinopyroxenes and melt were seen, so it can be assumed that clinopyroxene zoning records the continuing magmatic processes before and during eruption.
The compositional zoning of clinopyroxene phenocrysts was investigated by optical microscope and scanning electron microscopy [SEM; back-scattered electron imaging (BEI)]. Mineral chemistry was determined using the JEOL 8600 superprobe at Hokkaido University of Education, Asahikawa, and for BEI the JEOL 8800 system at Hokkaido University was used. Beam current was 2 x 10-8 A on the Faraday cup. Counting time for Cr2O3 was 60 s (peak) and 30 s (background), and for other elements, 20 s (peak) and 5 s (background). Zoning profiles were obtained by step-scans (interval 1–4 µm), with full elemental analysis made at each point. ZAF corrections were applied to all analyses.
|
TYPES OF ZONING PROFILE FOR CLINOPYROXENE PHENOCRYSTS |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL PETROGRAPHY AND MINERAL...SAMPLES AND EXPERIMENTAL...TYPES OF ZONING PROFILE...DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Mg-values of clinopyroxene phenocryst cores are similar in the three samples, ranging from 65 to 85. Comparing core and rim compositions, phenocrysts with relatively iron-rich cores usually show reverse zoning, whereas those with relatively magnesian cores have normal zoning (Nakagawa et al., 1999). However, the zoning profiles are variable and complex. Clinopyroxene phenocrysts with relatively Fe-rich cores (Mg-number <
70) can be divided into three types (L-1, L-2 and L-3), and those with Mg-rich cores (Mg-number >80) into two types (H-1 and H-2) (Fig. 4). Representative chemical compositions of each type of clinopyroxene phenocryst are listed in Table 3.
|
|
|
Phenocrysts with iron-rich cores
Type L-1 has a resorbed core surrounded by a clear, reversely zoned mantle [fig. 5 of Nakagawa et al. (1999) and Fig. 5]. The mantles are characterized by an abrupt increase in Mg-number (>80), resulting in values nearly as high as the core compositions of those clinopyroxene phenocrysts from high-T end-member magma (Fig. 6). Optically, type L-1 phenocrysts are easily identified because the large Mg-number difference results in a greenish core but colourless mantle. Although mantle thickness and zoning pattern are variable, even in a single sample (Fig. 7), several reversals may be recorded in the mantle (Fig. 6). Mantle widths mostly range from several micrometres to 200–300 µm, though many type L-1 phenocrysts have only thin mantles (<50 µm) (Fig. 8). Although the thickness and zoning profile of the mantle of type L-1 phenocrysts are variable (Figs 7 and 8), we can recognize the following similarities:
|
|
|
- all cores show, to a varying extent, a degree of resorption (Fig. 5);
- at the innermost mantle, Mg-number increases rapidly to >80 over 20 µm (zone A);
- high-Mg-number (>80) continues across several tens or hundreds of micrometres width (zone B);
- beyond zone B, Mg-number decreases gradually toward the rim (zone C);
- within zones B and C, there are several reversals of Mg-number and Cr2O3 content (Fig. 6);
- the outermost rim of some type L-1 phenocrysts shows reverse zoning (Fig. 6);
- although the zoning profile of each mantle is complex, the Mg-number is always higher than that of the outer margin of the core (Figs 6 and 7).
Type L-2 and L-3 phenocrysts do not show the extreme reverse zoning seen in type L-1 phenocrysts (Fig. 9), and being greenish without a colourless mantle, are easily distinguished from type L-1. Type L-2 phenocrysts have only a thin rim (less than 10–20 µm) with subdued reverse zoning; Mg-number of the outermost rim is usually about 75–70, and does not exceed 80. Zoning of type L-3 is either indistinct or shows normal zoning toward the rim.
|
Phenocrysts with magnesian cores
Clinopyroxene phenocrysts with relatively magnesian cores (Mg-number >80) documented by Nakagawa et al. (1999), show two types of zoning, H-1 and H-2 (Fig. 10). These magnesian cores are colourless and are easily distinguished from the greenish iron-rich phenocrysts. Type H-1 is characterized by a normally zoned rim, whereas in type H-2, the inner part of the rim is normally zoned, and the outer part reversely zoned, but not exceeding Mg-number 80. Although zoning profiles of phenocrysts with magnesian cores are variable, comparing Mg-number between core and rim, it can be concluded that these phenocrysts are normally zoned.
|
Transitional types
Transitional (type T) clinopyroxenes occur only in the October sample (Ru-N102) (Fig. 11). In these phenocrysts, Mg-number passes through a peak in the outer part of the core, which is jacketed by a thin reversely zoned rim.
|
Proportion of zoning types
The proportions of each phenocryst type have been determined by microscope counting, BEI and step-scan analysis (Table 4). Crystals with Fe-rich cores (L-1, -2, and -3) account for >80% of the clinopyroxenes in all samples, and type L-1 are dominant throughout the 1995 eruptive sequence. Type L-3 and H-1 phenocrysts are moderately abundant in the September samples, but rare or absent in the October sample, which has more type H-2 phenocrysts. Coexistence of the various types of phenocrysts in a single sample indicates the complex injection and storage history of the shallow magma system, with phenocrysts in small pockets at different stages of development being swept together and ejected during each eruptive episode.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL PETROGRAPHY AND MINERAL...SAMPLES AND EXPERIMENTAL...TYPES OF ZONING PROFILE...DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Clinopyroxene–melt equilibria
Clinopyroxene–melt equilibria are investigated using Fe–Mg exchange between matrix glass and core compositions of clinopyroxene phenocrysts (Fig. 12). FeO/MgO ratios of matrix glass of three dated scoria show wide variations, indicating the heterogeneity in the glass. The phenocrysts with magnesian core (Mg-number >80) are not in equilibrium with the glass. On the other hand, although the phenocrysts with iron-rich cores (Mg-number <70) can be in equilibrium with relatively iron-rich glass in the 23 September scoria, those in both the 18 September and 11 October scoria would not be also in equilibrium with matrix glass. These suggest that the phenocrysts with magnesian and iron-rich cores could not crystallize from the same magma. Considering compositional zonations of these clinopyroxene phenocrysts, it can be suggested that normally zoned phenocrysts with Mg-rich cores crystallized from more mafic melt (high-T magma), and that those with iron-rich cores from silicic melt of 23 September scoria or a more silicic one (low-T magma) than that of both 18 September and 11 October. In addition, the matrix glass is a mingling product between mafic melt in the high-T magma and silicic melt in the low-T magma. This is consistent with the magma-mixing process proposed by Nakagawa et al. (1999).
|
Estimation of magma-mixing ratios and presumed zoning of phenocryst minerals
It is important to estimate magma-mixing ratios so that observed zoning profiles can be interpreted correctly. Sakuyama (1979, 1981) suggested that the phenocrysts from each end-member magma can form similar zoned rims, whose compositions are determined by the composition and temperature of the mixed magma. Kouchi & Sunagawa (1985) determined experimentally that mixed melts change their chemical compositions slowly as the mixing process moves gradually towards a final chemical composition corresponding to the ratio of the two end melts.
Figure 2 shows histograms of the chemical compositions of outermost rims of phenocryst minerals, which could indicate the chemical compositions and magmatic temperature of the mixed melts. Although rim compositions were determined at 3–10 µm from their edges by Nakagawa et al. (1999), those in Fig. 2 have been remeasured by step-scan analysis so that rim compositions are known more precisely. Average rim Mg-number is about 70–75 for clinopyroxene, and 67–72 for orthopyroxene, except in the 18 September sample, which shows wider variations, possibly as a natural consequence of incomplete mixing, or perhaps because the precise rim position had not been located accurately during analysis. The An content of groundmass plagioclase is similar to or slightly An-rich compared with phenocryst plagioclase from the low-T end-member magma (Fig. 3).
If the rim compositions of pyroxene phenocrysts and core compositions of groundmass plagioclase reflect mixed melt compositions, a low mixing ratio (high-T/low-T magma) is suggested, with the dominant melt being low-T type. This is consistent with the observed dacitic composition (SiO2 62·5–70%) of the matrix glass (Nakagawa et al., 1999), reflecting only a small component of high-T magma in the mix. If low-T melt did dominate the mixing process, the clinopyroxene phenocrysts originally from the low-T magma should have reversely zoned thin mantles in which Mg-number (Mg-number 70–75) is slightly higher than in the core (Mg-number 70). Conversely, the phenocrysts with a magnesian core (Mg-number 85) from the high-T melt should have a normally zoned mantle with similar Mg-number to the mantle of phenocrysts from the low-T magma.
Type L-2 and H-1 phenocrysts are consistent with this mixing scheme, having zoned rims in which Mg-number is 75, but the dominant type L-1 phenocrysts have extremely reverse-zoned mantles, with up to Mg-number 80, as is found in cores of phenocrysts from high-T magma. This latter pattern is inconsistent with a low-T dominant melt-mixing ratio.
Origin of type L-1 phenocrysts
Zoning profile of type L-1 phenocrysts
The abrupt increase of Mg-number into zone A in type L-1 phenocrysts (Fig. 13) suggests that they were surrounded, not by mixed melt, but by high-T melt, based on the fact that the Mg-number is the same as that of the core in phenocrysts derived from the high-T magma. In addition, the width of zone B indicates that these phenocrysts grew for some time. The widely accepted model for magma mixing, in which melt-rich magmas with suspended phenocrysts are mixed, cannot explain the presence of abundant type L-1 phenocrysts. Thus it is important to know how and when zones A and B form, as they provide a key to understanding the magmatic processes during the 1995 Ruapehu eruptions.
|
Difference in zonation between the core and mantle
High Mg-number and Cr content are both good indicators of the addition of mafic, high-T melt in the mixing process, and Figs 6 and 14 show that the two parameters correlate in both cores and mantles of type L-1 phenocrysts. Mantle Cr2O3 contents correlate well with Mg-number, indicating that both zonations reflect magmatic compositions. Cores usually have low Cr2O3 (<0·2%) and Mg-number (<75), overlapping with mantle values, although some cores (e.g. 87CPX and 83CPX) have high Cr2O3 contents (up to 0·9%) at lower Mg-number levels (79–68) than in mantle with similar Cr2O3 contents. The anomalous cores of 87CPX in Ru-069, and of 83CPX in Ru-N102, show clear and spike-like reverse zonations of Cr2O3 that do not correlate with Mg-number (Fig. 6).
|
The distinctive zoning profiles of Cr2O3 content and Mg-number in some type L-1 cores may be caused by differences in the diffusion coefficients of Fe–Mg and Cr in clinopyroxene. The interdiffusion coefficient of Cr in clinopyroxene has not been reported yet, but Sautter & Harte (1990) have calculated the interdiffusion coefficient of Al in clinopyroxene to be in the range 10-16-10-20 cm2/s at 1200°C. This value is much slower than that of Ca–Mg in clinopyroxene at 1200°C (10-15 cm2/s) recorded by Brady & McCallister (1983). Considering ionic charge, it seems likely that the interdiffusion coefficient of Cr may be similar to that of Al, and much smaller than that of Fe–Mg.
The effects of annealing are apparent in some cores in which Mg-value and Cr2O3 content do not follow each other closely (87CPX of Ru-069, and 83CPX of Ru-N102 in Fig. 6). Thus, after a zoned core had formed, if the crystal was subjected to a long period of annealing, the elements Fe, Mg and Ca (which have higher diffusion coefficients than Cr) would have dispersed more easily, so smoothing out their peaks in the original zoning profile relative to Cr. However, similar zoning profiles for Mg-number and Cr are more common (Fig. 6), indicating that the crystals were not subjected to extended annealing after the mantles had formed. Possibly, the mantles with their correlated zoning may represent just a short period of time leading up to, and during the 1995 eruption episode.
Time duration of the mantle growth
As mentioned above, widths of mantles range from less than 10 to 300 µm, and many are <50 µm (Fig. 8). They may have been growing during the whole 1995 episode, including the 6 month period of precursor activity from April 1995 to when the actual eruptions commenced.
Growth and nucleation rates in shallow volcanic systems (chambers) have been estimated from natural samples by the CSD method (crystal size distribution; Marsh, 1988), and have been shown to depend on melt composition. Thus, the growth rates of plagioclase in Mt. St. Helens dacite have been calculated to be 10-11-10-12 cm/s (Cashman, 1988), whereas those in Hawaiian basalts are faster at 10-10-10-11 cm/s (Cashman & Marsh, 1988; Cashman, 1993).
Growth rates of olivine in Hawaiian basalts are 10-9 and 10-10 cm/s (Mangan, 1990), and were higher than those of plagioclase, but in contrast, Armienti et al. (1994) concluded that the average growth rate of olivine in 1991–1993 Mt. Etna hawaiites (2·4 x 10-9 cm/s) was slower than that of plagioclase (7·1 x 10-9 cm/s). Growth rates of clinopyroxene have not been discussed extensively, although Armienti et al. (1994) investigated CSD for clinopyroxene and oxide minerals in the same samples, and considering the slope of the clinopyroxene CSD plot (population density vs crystal size), the growth rates for clinopyroxenes can be estimated to be intermediate between those of plagioclase and olivine in the same sample.
Melt compositions of 1995 Ruapehu andesites (SiO2 62–70%) are less silicic than those of Mt. St. Helens dacite (SiO2 72–80%), suggesting that growth rates of plagioclase in the 1995 Ruapehu andesite would be slightly faster, possibly of the order of 10-11 cm/s in the magma chamber, and similar to clinopyroxene growth rates. In fact, the magnesian mantles grew from high-T magma in a mixed system. During mixing, high-T magma would be quenched, and the degree of undercooling would be higher than during normal crystallization in a chamber. An increased degree of undercooling causes the growth rate of minerals to increase by two or three orders of magnitude. Thus, Cashman (1993) revealed that undercooling increases the growth rate of plagioclase over the range from 10-8 to 10-11 cm/s. In the microphenocryst crystallization stage of a Mauna Loa basalt lava flow, Crisps et al. (1994) determined growth rates of plagioclase to be about 10-7-10-9 cm/s, and in a sample that had both plagioclase and clinopyroxene, growth rates of clinopyroxene (5x10-9 cm/s) were slightly slower than those of plagioclase [(6–9) x10-9 cm/s]. Therefore, if the growth rate of clinopyroxene mantles of type L-1 phenocrysts is assumed to be about 10-10 cm/s, many mantles (width <50 µm) must have been growing for less than 2 years before eruption, and considering the results of Cashman (1993) and Crisp et al. (1994), the growth rates might have been faster than 10-10 cm/s; in this case, mantles of 50 µm thickness could have formed in as short a time as several months.
These estimated durations of the clinopyroxene mantle growth are consistent with the 6 months of precursor activity before the 1995 eruptions, and we believe that the mantles of type L-1 phenocrysts may have formed entirely during the precursor and eruptive phases.
Evidence for a mush chamber
Stagnation of injecting high-T magma in a magma chamber
It has been considered that a magma chamber consists of liquid magma (suspension zone) surrounded by a boundary layer, which varies from a rigid crust (crystallinity >50–55%) to a mush zone (crystallinity >25–50%) (Marsh, 1996). Generally, the floor boundary layer is much thicker than the roof sequence (Tait & Jaupart, 1996). A simplified magma chamber could be drawn as in Fig. 15a, after Barth et al. (1994) and Marsh (1996). In the crystal-poor, melt-rich layer, phenocrysts are suspended by convection, whereas the boundary layer consists mainly of a crystal network and/or chain with interstitial melt. It has been reported that some phenocrysts originate in boundary layers that form at the roof and wall of a crystallizing magma chamber (Tait, 1988; Nakada et al., 1994), and this process may explain the origin of type L-1 phenocrysts.
|
To understand the behaviour of high-T magma during injection into a high-level magma chamber beneath Ruapehu volcano, we have compared the calculated density difference between the magma and chamber (Fig. 15b) assuming that the interstitial melt in the chamber has the same composition as silicic matrix glass (No. 8 of sample Ru-069 in Table 2) from the 1995 juvenile andesites. The density of this melt represents a maximum possible density of the interstitial melt, because the analysed silicic glass compositions would represent already mixed melt compositions in the 1995 andesite. With regard to the injecting high-T magma, we assume two melt compositions: one is represented by the least siliceous of the 1995 andesites (Ru-N102 in Table 1), and the other is an older basaltic andesite (SiO2 54·2%, MgO 6·74% and K2O 0·68%) from Ruapehu (Graham & Hackett, 1987). The true density of the postulated high-T melt is likely to lie between the calculated densities of these analysed andesites. We do not consider basaltic magma as a possible high-T end-member magma, because olivine phenocrysts have not been found in the 1995 mixed andesites.
Injecting high-T magma could be expected to have intruded as a dyke through both the rigid, crusted floor of the magma chamber and any highly crystalline mush (55% vol. % crystals), because it has lower density than either (Fig. 15b). On the other hand, mush with a crystal content of 25 vol. % or less might be less dense than the injecting mafic magma (Fig. 15b). If so, the high-T magma could stagnate within the mush zone rather than passing through as a dyke. However, the less crystalline mush would still have a considerable yield strength and would inhibit large-scale convection (Marsh, 1989), preventing melt-mixing processes such as chamber fountaining (Campbell & Turner, 1988), convective entrainment (Snyder & Tait, 1996) or interfacial turbulent mixing (Cardoso & Woods, 1996). Instead, it is likely that crystals separated from the network of the mush would be surrounded immediately by the high-T magma.
Interaction between the high-T magma and mush zone
All the cores of type L-1 phenocryst show various degrees of resorption (Fig. 13), indicating they had been eroded thermally before mantle growth. Thus, it can be concluded that thermal erosion processes in a crystal mush occurred before the crystals were surrounded by high-T magma. Accounting for thermal erosion, the possible processes of high-T magma injection are as follows:
- injection of high-T magma could supply heat to the boundary layer, causing thermal erosion of crystal networks and/or chains in any crystal mush. The crystals would be separated from the networks or chains and surrounded by the high-T magma. At the same time, clusters of crystals could detach from the boundary layer by delamination (Marsh, 1996), and would form crystal clots or gabbroic inclusions in the high-T magma. Some of these crystal clots may then have been disaggregated as the temperature rose, to be captured by the high-T magma as apparent phenocrysts.
- In a mush zone with relatively lower crystal content, the interstitial melt could easily migrate, and the density difference between the invading high-T melt and the interstitial melt of the boundary layer would cause the lighter interstitial melt to be forced out of the crystal network and/or chains.
If the low-T magma chambers, before the injection of high-T magma, comprised mainly crystal-poor melt with only a thin boundary layer, phenocrysts with extremely zoned mantles, captured from the crystal-rich boundary layer, should not be so common in the mixed magma. In fact, about 50% of the phenocrysts in the 1995 andesites are type L-1 (Table 4), and this implies that the low-T magma storage system beneath Ruapehu volcano was crystal-rich, and can be regarded as a crystal-mush chamber, as deduced by Nakagawa et al. (1999). In conclusion, we consider the dominance of type L-1 phenocrysts in the 1995 andesite to be direct evidence for the presence of a mushy magma chamber within Ruapehu volcano.
Magma-mixing processes in a chamber
Here we discuss possible magmatic processes in a mushy chamber, which may be recorded in the zoning profiles formed after overgrowth of zone A in the mantle of type L-1 phenocrysts. In addition, we also discuss the origin of the other types of clinopyroxene phenocrysts.
Formation and growth of magma pockets
Magma pockets (dyke, sheet or lens) formed by the injection of high-T magma into a mushy chamber contain phenocrysts originally from both high-T magma and low-T mush (Fig. 16). During the formation of the magma pocket, the first reverse zonations on the low-T crystals are represented by zone A. As mentioned above, the melt-mixing process between the magma pockets (high-T magma) and interstitial low-T melt in the mushy chamber could not easily progress, and phenocrysts captured into the magma pockets were surrounded by high-T melt, resulting in the progressive growth of zone B. If the injection of the high-T magma did not occur frequently, these magma pockets would be cooled in the crystal mush chamber. However, repeated reverse zonations observed in mantles suggest that high-T magma injections continued intermittently (Nakagawa et al., 1999). In some cases, the repeated injections would have enlarged pre-existing magma pockets in the mush. Therefore, zone B would continue to grow because the temperature and composition of the pocket were maintained. At the same time, crystals newly captured from the mush would be incorporated into the enlarged pocket so that zone A would form on these crystals at the same time as oscillations were produced in zone B of previously incorporated crystals. Thus, although crystal mantle growth rates may have been similar for each phenocryst, the widths of the mantles and the zoning profiles of the type L-1 phenocrysts are variable (Figs 5 and 6).
|
Melt-mixing process
The normally zoned C zones around zone B (Fig. 13) are probably related to melt mixing between high-T magma pockets and the interstitial felsic melt in the low-T mushy magma (Fig. 16). Melt mixing could occur as a result of convection in magma pockets, conjunction and coalescence of the pockets, and/or discharge of the pockets from the mushy chamber. Large-scale convection in the mush is unlikely (Marsh, 1989; Barth et al., 1994), but small convection cells might have been generated in response to temperature gradients produced by repeated injection of new melt. Indeed, Barth et al. (1994) observed small convection cells in a cooling, crystal-rich (mush) lava lake, and if similar convection cells formed in the Ruapehu mush, melt mixing could have occurred more easily.
Coalescence of several magma pockets may occur in a mushy chamber (Takada, 1994), causing melt mixing, as was proposed for Unzen volcano (Sato, 1996). In such cases, convection and/or coalescence should be accompanied by the movement of many phenocrysts, which behave like a stirrer, and increase mixing efficiency, as suggested by Kouchi & Sunagawa (1985).
During melt mixing, overgrowths form on phenocrysts corresponding to the ratio of the melt components. Consequently, both zone C of the type L-1 crystals and the normal-zoned rims of type H-1 crystals were formed. In addition, during the mixing, some phenocrysts with Fe-rich cores (type L-2) would be captured from the mushy chamber and incorporated into the mixed melt, producing less extreme zonations compared with zone A of type L-1 crystals, and consistent with a simple mixing between melt-rich magmas.
Some type L-1 phenocrysts have a zone D, which is a reversely zoned outermost rim. The reverse-zoned outermost rim of other types of phenocrysts (types L-2 and H-2) probably grow at the same time, but these reverse zonations are not so extreme, indicating that the rims must be formed by melt mixing in the final stages just before and/or during eruption. Types L-2 and H-2 phenocrysts, which show reverse zoning at the outermost rim, are more abundant in the October sample, whereas type L-3 are dominant in the September samples. This suggests that the magmatic processes recorded by the outermost rim might be different between the September and October eruptive episodes.
Petrological comparison between the September and October eruptions
Difference between the September and October magmatic processes, just before and during eruptions, may be reflected in the nature of the zoning of the outermost rims of phenocrysts. Almost all the October clinopyroxene phenocrysts show reverse zoning at the outermost rims (Fig. 17), suggesting that the injection of high-T magma occurred intensively just before and during eruptions. The widths of these reverse zoned rims are <8 µm, and if the crystal growth rate of clinopyroxene is 10-10 cm/s during shallow magma storage, the rims had grown for <12 days before eruption. In fact, some part of the outermost rims probably formed during ascent of magma, in which case, crystal growth rate could have been faster than 10-10 cm/s, and the outermost rims may have grown only a few days before the 11 October eruptions. On the other hand, zoning at the outermost rims of the September clinopyroxene phenocrysts is not well established, suggesting only minor magma mixing just before and during the eruptions in September.
|
The petrological differences between the September and October samples are in good agreement with results of geophysical monitoring (Bryan & Sherburn, 1999). Before the 11 October eruption, a series of intense volcanic earthquakes had been monitored on 7 October, accompanied by phreatic eruptions. A large-scale injection of high-T magma several days before 11 October, as indicated by our petrological analysis, could correspond to that period of volcanic earthquakes. In contrast, volcanic tremor but no intense volcanic earthquakes had been detected for several days before the September eruptions, suggesting that injection of high-T magma just before eruption was of small scale or absent, also consistent with our petrological investigations (Fig. 17).
Sparks et al. (1977) suggested that magma injection into a magma chamber can trigger an eruption, and many examples have been suggested (e.g. Gourgaud et al., 1989; Pallister et al., 1992). In the case of the October eruptions of Ruapehu, we conclude that injections of the high-T magma were the trigger in the way proposed by Sparks et al. (1977), but our petrological observations and lack of precursor earthquakes suggest this may not have been so for the September eruptions. However, there could have been much smaller-scale injections, which were not detected seismically. Another possibility is that the eruptions were triggered by the coalescence of several magma pockets in a mushy chamber. We favour the latter possibility, and suggest that the volcanic tremor observed before the September eruption (Bryan & Sherburn, 1999) may have been generated as two or more small magma pockets combined and became unstable.
Initial injections of high-T magma into the mush zone may have occurred several months before September (Bryan & Sherburn, 1999; Nakagawa et al., 1999). However, following the discharge of considerable amounts of magma during the September eruptions, the depleted chambers needed intensive new injections of high-T magma to destabilize them before the onset of the October eruptions. Hence it is important to trace volcanic activity and to know the likely current status of the magma plumbing system when monitoring and predicting small-scale andesitic explosive eruptions of the type characterized by Ruapehu.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL PETROGRAPHY AND MINERAL...SAMPLES AND EXPERIMENTAL...TYPES OF ZONING PROFILE...DISCUSSIONCONCLUSIONSREFERENCES |
|---|
We focus on compositional zoning of clinopyroxene phenocrysts of three samples from separate 1995 eruptions (11 and 23 September and 11 October) of Ruapehu volcano, which are mixing products of higher-temperature (high-T) and lower-temperature (low-T) magmas. On the basis of core compositions and zoning profiles of clinopyroxene phenocrysts, the phenocrysts with Fe-rich cores (Mg-number <
70), which originally crystallized the low-T end-member magma, can be divided into three types, and those with Mg-rich cores (Mg-number >80), derived from high-T end-member magma, into two types. The most dominant type with Fe-rich cores is type L-1, which has a resorbed core surrounded by a clear, reversely zoned mantle. The mantles are characterized by an abrupt increase in Mg-number (>80) at the innermost part of the mantle, resulting in values nearly as high as the core compositions of those clinopyroxene phenocrysts from the high-T end-member magma. Then, high-Mg-number (>80) in the mantles continues across several tens or hundreds of micrometres width, and Mg-number decreases gradually toward the rim. The abrupt increases of Mg-number in the mantles suggest that the phenocrysts were surrounded, not by mixed melt, but by high-T melt. This could be explained not by the widely accepted model for magma mixing, in which magmas with suspended phenocrysts are mixed, but by mixing processes in a mush chamber. It is suggested that high-T magma was injected into a mushy low-T magma, then the denser and hotter high-T magma displaced the lighter and cooler interstitial melt in the mush. In addition, the high-T magma could thermally erode crystals in the mush network. Hence, eroded crystals from the mush were captured by the high-T magma. Zoning profiles of the mantles following the abrupt increase of Mg-number must record mixing processes, and suggest that high-T magma had intermittently continued to form and enlarge magma pockets in the mush. Considering the width of the mantles and possible growth rate of clinopyroxene, repeated injections of high-T magma may have occurred between 2 years and several months before the initial magmatic eruption. This is consistent with the results of monitoring, which indicate that obvious precursor activity had occurred from 5 months before the eruption. Repeated injection of high-T magma formed and enlarged magma pockets in the mush. In the process, melt mixing occurred owing to convection and/or conjunction and coalescence of the magma pockets. The mixing between the magma pockets and interstitial melt of the mush gave rise to both normal zoning in the mantles of contained phenocrysts and zoning of other types of phenocrysts. Zoning in the outermost rims of pyroxenes differs between the September and October scorias. Almost all of the October phenocrysts, but few of the September phenocrysts, have thin (several micrometres) reverse zoned rims. Therefore we suggest that the October eruptions may have been triggered by intensive injection just before the eruption, whereas the September eruptions may have been triggered either by smaller-scale injections or coalescence of magma pockets. These differences in presumed triggering process are consistent with geophysical observations before and during the 1995 eruptions. Although our speculation might be one of a number of possible interpretations, it is evident that analysis of compositional zoning in phenocrysts provides useful information for understanding complex magmatic processes during and before eruptions.
|
ACKNOWLEDGEMENTS |
|---|
We thank T. Thordarsson and S. Sherburn for discussions about the volcanic activity and magmatic systems of Ruapehu volcano. Thanks are also extended to T. Kuwashima and H. Nomura for making thin sections. Reviews by J. G. Brophy, M. Jellinek, D. Geist, F. J. Tepley and an anonymous reviewer improved this manuscript. This study was financially supported by the New Zealand Foundation of Research, Science and Technology, and the Japanese Ministry of Education (No. 11640471).
|
FOOTNOTES |
|---|
*Corresponding author. E-mail: nakagawa{at}ep.sci.hokudai.ac.jp
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGENERAL PETROGRAPHY AND MINERAL...SAMPLES AND EXPERIMENTAL...TYPES OF ZONING PROFILE...DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Armienti, P., Pareschi, M. T., Innocenti, F. & Pompilio, M. (1994). Effects of magma storage and ascent on kinetics of crystal growth. The case of the 1991–93 Mt. Etna eruption. Contributions to Mineralogy and Petrology 115, 402–414.[Web of Science]
Barth, G. A., Kleinrock, M. C. & Helz, R. T. (1994). The magma body at Kilauea Iki lava lake: potential insights into mid-ocean ridge magma chambers. Journal of Geophysical Research 99, 7199–7217.
Berman, R. G. (1988). Internally-consistent thermodynamic data for minerals in the system Na2O–K2O–CaO–MgO–FeO–Fe2O3–Al2O3–SiO2–TiO2–H2O–CO2. Journal of Petrology 29, 445–522.
Brady, J. B. & McCallister, R. H. (1983). Diffusion data for clinopyroxenes from homogenization and self-diffusion experiments. American Mineralogist 68, 95–105.[Abstract]
Brophy, J. G., Dorais, M. J., Donnelly-Nolan, J. & Singer, B. S. (1996). Plagioclase zonation styles in hornblende gabbro inclusions from Little Glass Mountain, Medicine Lake volcano, California: implications for fractionation mechanisms and the formation of composition gaps. Contributions to Mineralogy and Petrology 126, 121–136.[Web of Science]
Bryan, C. J. & Sherburn, S. (1999). Seismicity associated with the 1995–96 eruptions of Ruapehu volcano, New Zealand: narrative and insights into physical processes. Journal of Volcanology and Geothermal Research 90, 1–18.[Web of Science]
Campbell, I. H. & Turner, J. S. (1988). Fountains in magma chambers. Journal of Petrology 30, 885–923.[Web of Science]
Cardoso, S. S. S. & Woods, A. W. (1996). Interfacial turbulent mixing in stratified magma reservoirs. Journal of Volcanology and Geothermal Research 73, 157–175.[Web of Science]
Cashman, K. V. (1988). Crystallization of Mount St. Helens 1980–1986 dacite: a quantitative textural approach. Bulletin of Volcanology 50, 194–209.[Web of Science]
Cashman, K. V. (1993). Relationship between plagioclase crystallization and cooling rate in basaltic melts. Contributions to Mineralogy and Petrology 113, 126–142.[Web of Science]
Cashman, K. V. & Marsh, B. D. (1988). Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallization II: Makaopuhi lava lake. Contributions to Mineralogy and Petrology 99, 292–305.[Web of Science]
Crisp, J., Cashman, K. V., Bonini, J. A., Hougen, S. B. & Pieri, D. C. (1994). Crystallization history of the 1984 Mauna Loa lava flow. Journal of Geophysical Research 99, 7177–7198.
de Silva, S. L. (1989). The origin and significance of crystal rich inclusions in pumices from two Chilean ignimbrites. Geological Magazine 126, 159–175.[Abstract]
Eichelberger, J. C. (1975). Origin of andesite and dacite: evidence of mixing at Glass Mountain in California and at other circum-Pacific volcanoes. Geological Society of America Bulletin 86, 1381–1391.
Fujimaki, H. (1982). Basalt produced by mechanical mixing of andesite magma and gabbroic fragments: Hakone volcano and adjacent areas, Central Japan. Journal of Volcanology and Geothermal Research 12, 111–132.[Web of Science]
Gamble, J. A., 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. Earth and Planetary Science Letters 170, 301–314.[Web of Science]
Garcia, M. O., Pietruszka, A. J., Rhodes, J. M. & Swanson, K. (2000). Magmatic processes during the prolonged Pu’u ’O’o eruption of Kilauea volcano. Journal of Petrology 41, 967–990.
Gourgand, A., Fichaut, M. & Joron, J.-L. (1989). Magmatology of Mt. Pelee (Martinique, F.W.I.). I: Magma mixing and triggering of the 1902 and 1929 Pelean nuées ardentes. Journal of Volcanology and Geothermal Research 38, 143–169.[Web of Science]
Graham, I. J. & Hackett, W. R. (1987). Petrology of calc-alkaline lavas from Ruapehu volcano and related vents, Taupo Volcanic Zone, New Zealand. Journal of Petrology 28, 531–567.
Grove, T. L. & Bryan, W. B. (1983). Fractionation of pyroxene-phyric MORB at low pressure: an experimental study. Contributions to Mineralogy and Petrology 84, 293–309.[Web of Science]
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.[Web of Science]
Hattori, K. & Sato, H. (1996). Magma evolution recorded in plagioclase zoning in 1991 Pinatubo eruption products. American Mineralogist 81, 982–994.[Abstract]
Houghton, B. F., Latter, J. H. & Hackett, W. R. (1987). Volcanic hazard assessment for Ruapehu composite volcano, Taupo Volcanic Zone, New Zealand. Bulletin of Volcanology 49, 737–751.
Housh, T. B. & Luhr, J. F. (1991). Plagioclase–melt equilibria in hydrous systems. American Mineralogist 76, 477–492.[Abstract]
Kouchi, A. & Sunagawa, I. (1985). A model for mixing basaltic and dacitic magmas as deduced from experimental data. Contributions to Mineralogy and Petrology 89, 17–23.[Web of Science]
Kuritani, T. (1998). Boundary layer crystallization in a basaltic magma chamber: evidence from Rishiri volcano, northern Japan. Journal of Petrology 39, 1619–1640.[Web of Science]
Lange, R. L. & Carmichael, I. S. E. (1990). Thermodynamic properties of silicate liquids with emphasis on density, thermal expansion, and compressibility. In: Nicholls, J. & Russell, J. K. (eds) Modern Methods of Igneous Petrology: Understanding Magmatic Processes. Mineralogical Society of America, Reviews in Mineralogy 24, 25–64.
Lindsley, D. H. (1983). Pyroxene thermometry. American Mineralogist 68, 477–493.[Abstract]
Mangan, M. T. (1990). Crystal size distribution systematics and the determination of magma storage times: the 1959 eruption of Kilauea volcano, Hawaii. Journal of Volcanology and Geothermal Research 44, 295–302.[Web of Science]
Marsh, B. D. (1988). Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallization I. Theory. Contributions to Mineralogy and Petrology 99, 277–291.[Web of Science]
Marsh, B. D. (1989). Magma chambers. Annual Review of Earth and Planetary Sciences 17, 439–474.
Marsh, B. D. (1996). Solidification fronts and magmatic evolution. Mineralogical Magazine 60, 5–40.[Abstract]
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.[Web of Science]
Nairn, I. A. & Ruapehu Surveillance Group (1996). Volcanic eruption at a New Zealand ski resort prompts reevaluation of hazards. EOS Transactions, American Geophysical Union 77, 189–191.
Nakada, S., Bacon, C. R. & Gartner, A. E. (1994). Origin of phenocrysts and compositional diversity in Pre-Mazama rhyolitic lavas, Crater Lake, Oregon. Journal of Petrology 35, 127–162.
Nakagawa, M., Wada, K., Thordarson, T., Wood, C. P. & Gamble, J. A. (1999). Petrologic investigations of the 1995 and 1996 eruptions of Ruapehu volcano, New Zealand: formation of discrete and small magma pockets and their intermittent discharge. Bulletin of Volcanology 61, 15–31.[Web of Science]
Nakamura, M. (1995). Continuous mixing of crystal mush and replenished magma in the ongoing Unzen eruption. Geology 23, 807–810.
Pallister, J. S., Hoblitt, R. P. & Reyes, A.G. (1992). A basalt trigger for the 1991 eruptions of Pinatubo volcano? Nature 356, 426–428.
Sack, R. O. & Ghiorso, M. O. (1994). Thermodynamics of multicomponent pyroxenes: III Calibration of Fe2+(Mg)-1, TiAl2(MgSi2)-1, TiFe23+(MgSi2)-1, AlFe3+(MgSi)-1, NaAl(CaMg)-1, Al2(MgSi)-1 and Ca(Mg)-1 exchange reactions between pyroxenes and silicate melts. Contributions to Mineralogy and Petrology 118, 271–296.[Web of Science]
Sack, R. O., Carmichael, I. S. E., Rivers, M. & Ghiorso, M. S. (1980). Ferric–ferrous equilibria in natural silicate liquids at 1 bar. Contributions to Mineralogy and Petrology 75, 369–376.
Sakuyama, M. (1979). Evidence of magma mixing: petrological study of Shirouma-Oike calc-alkaline andesite volcano, Japan. Journal of Volcanology and Geothermal Research 5, 179–208.[Web of Science]
Sakuyama, M. (1981). Petrological study of the Myoko and Kurohime volcanoes, Japan: crystallization sequence and evidence for magma mixing. Journal of Petrology 22, 553–583.
Sato, H. (1996). A model of the 1991 eruption of Unzen volcano inferred from petrographic texture of the ejecta. Memoirs of Geological Society of Japan 46, 115–125.
Sautter, V. & Harte, B. (1990). Diffusion gradients in an eclogite xenolith from the Roberts Victor kimberlite pipe: (2) kinetics and implications for petrogenesis. Contributions to Mineralogy and Petrology 105, 637–649.[Web of Science]
Silver, L. A., Ihinger, P. D. & Stolper, E. (1990). The influence of bulk composition on the speciation of water in silicate glasses. Contributions to Mineralogy and Petrology 104, 142–162.[Web of Science]
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 magma chambers. American Mineralogist 80, 776–798.[Abstract]
Sinton, J.M. & Detrick, R.S. (1992). Mid-ocean ridge magma chambers. Journal of Geophysical Research 97, 197–216.
Snyder, D. & Tait, S. R. (1996). Magma mixing by convective entrainment. Nature 379, 529–531.
Sparks, R. S. J., Sigurdsson, H. & Wilson, L. (1977). Magma mixing: a mechanism for triggering acid explosive eruptions. Nature 267, 315–318.
Tait, S. R. (1988). Sample from the crystallizing boundary layer of a zoned magma chamber. Contributions to Mineralogy and Petrology 100, 470–483.[Web of Science]
Tait, S. R. & Jaupart, C. (1996). The production of chemically stratified and adcumulate plutonic igneous rocks. Mineralogical Magazine 60, 99–114.[Abstract]
Takada, A. (1994). Development of a subvolcanic structure by the interaction of liquid-filled cracks. Journal of Volcanology and Geothermal Research 62, 207–224.
Umino, S. & Horio, A. (1998). Multistage magma mixing revealed in phenocrysts zoning of the Yunokuchi pumice, Akagi volcano, Japan. Journal of Petrology 39, 101–124.[Web of Science]
Wilson, C. J. N., Houghton, B. F., McWilliams, M. O., Lanphere, M. A., Weaver, S. D. & Briggs, R. M. (1995). Volcanic and structure evolution of Taupo Volcanic Zone, New Zealand: a review. Journal of Volcanology and Geothermal Research 68, 1–28.[Web of Science]
Wood, C. P., Gamble, J. A., Price, R. C., Smith, I. E. M. & Waight, T. (1998). Fifty years of magmatic evolution and eruption at Ruapehu volcano, New Zealand. 1998 Western Pacific Geophysical Meeting. Transactions of the American Geophysical Union 79, W122.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. J. Streck Mineral Textures and Zoning as Evidence for Open System Processes Reviews in Mineralogy and Geochemistry, January 1, 2008; 69(1): 595 - 622. [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]
|
|
|
|
 |
|
 S. J. Fowler, F. J. Spera, W. A. Bohrson, H. E. Belkin, and B. De Vivo Phase Equilibria Constraints on the Chemical and Physical Evolution of the Campanian Ignimbrite J. Petrology, March 1, 2007; 48(3): 459 - 493. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. DOWNES, J.-A. WARTHO, and B. J. GRIFFIN Magmatic Evolution and Ascent History of the Aries Micaceous Kimberlite, Central Kimberley Basin, Western Australia: Evidence from Zoned Phlogopite Phenocrysts, and UV Laser 40Ar/39Ar Analysis of Phlogopite-Biotite J. Petrology, September 1, 2006; 47(9): 1751 - 1783. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. TOMIYA and E. TAKAHASHI Evolution of the Magma Chamber beneath Usu Volcano since 1663: a Natural Laboratory for Observing Changing Phenocryst Compositions and Textures J. Petrology, December 1, 2005; 46(12): 2395 - 2426. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Di Muro, E. Bonaccorsi, and C. Principe Complex colour and chemical zoning of sodalite-group phases in a hauynophyre lava from Mt. Vulture, Italy Mineralogical Magazine, August 1, 2004; 68(4): 591 - 614. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Coombs, M. L. Coombs, and J. E. Gardner Reaction rim growth on olivine in silicic melts: Implications for magma mixing American Mineralogist, May 1, 2004; 89(5-6): 748 - 758. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. AMMA-MIYASAKA and M. NAKAGAWA Evolution of Deeper Basaltic and Shallower Andesitic Magmas during the AD 1469-1983 Eruptions of Miyake-Jima Volcano, Izu- Mariana Arc: Inferences from Temporal Variations of Mineral Compositions in Crystal-Clots J. Petrology, December 1, 2003; 44(12): 2113 - 2138. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||