Journal of Petrology Advance Access originally published online on March 18, 2005
Journal of Petrology 2005 46(8):1543-1563; doi:10.1093/petrology/egi025
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Petrogenesis of Mafic Inclusions in Rhyolitic Lavas from Narugo Volcano, Northeastern Japan
DEPARTMENT OF EARTH AND ENVIRONMENTAL SCIENCES, YAMAGATA UNIVERSITY, 4-12 KOJIRAKAWA-MACHI 1-CHOME, YAMAGATA 990-8560, JAPAN
RECEIVED MARCH 1, 2003; ACCEPTED FEBRUARY 11, 2005
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONMafic inclusions present in the rhyolitic lavas of Narugo volcano, Japan, are vesiculated andesites with diktytaxitic textures mainly composed of quenched acicular plagioclase, pyroxenes, and interstitial glass. When the mafic magma was incorporated into the silica-rich host magma, the cores of pyroxenes and plagioclase began to crystallize (>1000°C) in a boundary layer between the mafic and felsic magmas. Phenocryst rim compositions and interstitial glass compositions (average 78 wt % SiO2) in the mafic inclusions are the same as those of the phenocrysts and groundmass glass in the host rhyolite. This suggests that the host felsic melt infiltrated into the incompletely solidified mafic inclusion, and that the interstitial melt composition in the inclusions became close to that of the host melt (c. 850°C). Infiltration was enhanced by the vesiculation of the mafic magma. Finally, hybridized and density-reduced portions of the mafic magma floated up from the boundary layer into the host rhyolite. We conclude that the ascent of mafic magma triggered the eruption of the host rhyolitic magma.
KEY WORDS: mafic inclusion; stratified magma chamber; magma mixing; mingling; Narugo volcano; Japan
|
INTRODUCTION |
|---|
Fine-grained mafic inclusions are common in andesitic to rhyolitic rocks from subduction-related tectonic settings (e.g. Eichelberger, 1980
; Bacon, 1986; Koyaguchi, 1986a). These inclusions appear as clots or blobs ranging from a few millimeters to over 1 m in diameter. Many researchers have concluded, on the basis of features such as ellipsoidal shapes, chilled textures preserved in groundmass minerals and high vesicularity, that these inclusions are the quenched products of mafic magma in contact with a cooler and more silicic host magma (e.g. Eichelberger, 1980; Bacon, 1986). The process that produces such incompletely mixed magmas is commonly called ‘magma mingling’ and is distinguished from processes that produce well-mixed, blended magma, called ‘hybridization’ (e.g. Murphy et al., 2000). Practically, when magma mingling is the dominant process, small-scale mixing (hybridization) can also take place, especially around the boundary area between the end-member magmas (Bacon, 1986).
Among the many mechanisms proposed to create mingled or hybrid magmas, the following three mechanisms are the most likely to result in the formation of mafic inclusions: (1) forcible injection of mafic magma into a cooler, felsic magma chamber (Campbell & Turner, 1989; Pallister et al., 1992; Nakamura, 1995a); (2) turbulent mixing of contrasting magmas during eruption of zoned magma chambers (Koyaguchi, 1985; Blake & Ivey, 1986; Cioni et al., 1995); (3) flotation of vesiculated mafic magma up to the top of a more silica-rich magma body (Eichelberger, 1980). Recently, some studies have suggested that the mechanism might change temporally with evolution of the magma chamber system (Feeley et al., 1998; Murphy et al., 2000). Murphy et al. (2000) proposed a scenario in which mafic inclusions are originally emplaced as fragments of disrupted dykes that intrude into a highly crystalline, lower-temperature, magma body; subsequently, as the temperature and mobility of the host magma increases, the mafic magma floats as ‘blobs’. From these studies, it appears that the mechanisms that form mafic inclusions can be varied, depending on specific circumstances.
In Narugo volcano, 1–2% mafic inclusions are present in most of the rhyolitic lavas; these are usually round in form, up to 70 cm in diameter, and vesicular. The inclusions are andesitic in composition (55–60 wt % SiO2), with acicular plagioclase and pyroxene microlites, and interstitial glass in the groundmass; some inclusions have a minor amount of phenocryst quartz. Groundmass glasses in both the inclusions and the host lavas are well preserved. We have carefully examined these glasses in both the mafic inclusions and the hosts, as well as the textures, mineral compositions and whole-rock compositions. We describe the solidification sequences accompanying the formation of the mafic inclusions, and provide constraints for the timing of solidification of the inclusions based on glass compositions as well as other petrological data.
|
ERUPTIVE SEQUENCE OF NARUGO VOLCANO |
|---|
Narugo volcano, situated in the central part of the Nasu volcanic zone (Figs 1 and 2) is one of 17 active volcanoes in the northeastern Japan arc. According to Sakaguchi & Yamada (1988)
and Soda & Yagi (1991), the activity of Narugo began with eruptions of pyroclastic flows: the Nizaka pyroclastic flow (c. 73 ka) and the Yanagizawa pyroclastic flow (c. 45 ka). A large caldera, c. 5 km in diameter, was formed by these paroxysmal eruptions. Lake deposits found in the caldera indicate that water ponded in the caldera after c. 45 ka. Afterwards, lava flows and lava domes were formed in the inner part of the caldera. These lavas can be divided into three units: Ogatake–Kurumigatake lava dome, Matsugamine lava, and Toyagamori lava. The Toyagamori lava is younger than 11 830 ± 555 years BP based on a 14C date of a buried tree (Omoto, 1993). Craters observed on the lava surface were formed by phreatic eruptions. One of these is regarded as a historical event (AD 837) (Murayama, 1978).
|
|
|
SAMPLES AND ANALYTICAL PROCEDURES |
|---|
Samples were collected from the lava flows and lava domes of the three units of Narugo volcano.
Whole-rock major element and selected trace element analyses were made on fused disks by X-ray fluorescence spectrometry (XRF) (Rigaku Rix2000) at Yamagata University. The glass disks were prepared using the method of Yamada et al. (1995). The calibration method followed that of Yamada et al. (1995) for major and trace elements. The standards used in the analyses are the Geological Survey of Japan (GSJ) igneous rock series. Analytical uncertainties for XRF trace elements are <5% for Nb, Zr, Y, Sr, Rb, and Ni; <10% for V and Cr; 5–15% for Ba. The range of uncertainties for a single element is based on the concentration range observed in the standards.
Mineral analyses were carried out with a JEOL 8600SS electron microprobe at Yamagata University using a wavelength-dispersive technique. Operating conditions were 15 kV accelerating voltage, 10 nA (plagioclase and glasses) to 20 nA (pyroxenes) beam current, and 8–20 s counting time for each element. Glasses were analyzed using a 10 µm defocused beam. All analyses were corrected using the oxide ZAF method.
|
SAMPLE DESCRIPTIONS AND PETROGRAPHY |
|---|
The petrographic characteristics of rocks from Narugo volcano are summarized in Table 1. Sixty-two samples were studied using a photomicroscope, and point counting to determine mineral modes was performed for 40 samples. Most samples are porphyritic rhyolites with
15–30 vol. % phenocrysts; some are weakly vesiculated. Two lithological types of mafic inclusion can be recognized within the rhyolites. One is gray to dark gray and the other is reddened. These mafic inclusions are usually in rounded form and up to c. 70 cm in diameter; they are moderately vesiculated, and occur in most of the lavas. Mafic inclusions that are included in gray to dark gray hosts are gray to whitish gray in color, whereas those that are reddened are included in reddened hosts. Some of the dark gray rhyolites from the Toyagamori lava are highly glassy in nature and weakly vesiculated.
|
Silicic hosts
The hosts to the inclusions are rhyolitic in composition (70·5–75 wt % SiO2). These samples have porphyritic textures with
15 to 30 vol. % phenocrysts. Photomicroscope images of the representative gray and reddened hosts are shown in Fig. 3a and b. The phenocryst assemblage consists of clinopyroxene, orthopyroxene, plagioclase, quartz, and Fe–Ti oxides. Clinopyroxene and orthopyroxene phenocrysts (
1 mm), and plagioclase phenocrysts (3 mm) are subhedral to euhedral in shape. Most of the plagioclase phenocrysts are clear, but some plagioclase phenocrysts have a dusty zone (e.g. Tsuchiyama, 1985), or patchy zoning (Vance, 1965). Both types show oscillatory zoning, but the range of compositions between zones is usually within 10 mol % anorthite component. Quartz (2 mm) phenocrysts always have a resorbed margin. The modal amounts of phenocrystic clinopyroxene, orthopyroxene, and Fe–Ti oxides are similar in each of the three geological units. However, the modes of phenocrystic plagioclase and quartz are variable between units. Phenocrystic plagioclase ranges from 17 to 25% in the Ogatake–Kurumigatake lava dome, and from 11 to 16% in the Matsugamine and Toyagamori lavas. Quartz ranges from 2 to 3% in the Matsugamine lava, from 1 to 2% in the Ogatake–Kurumigatake lava dome, and is <0·5% in the Toyagamori lava. The groundmass is mainly composed of glass with a hyalopilitic texture. Tiny Fe–Ti oxide grains are scattered throughout glass. The amount of Fe–Ti oxides is greater in the reddened type than in the gray type. Minor amounts of acicular to prismatic clinopyroxene, orthopyroxene and plagioclase can be seen. In the groundmass of some samples, perlitic textures or flow structure are observed. The glasses are clear in the gray type and are usually reddish to brownish in the reddened type.
|
Mafic inclusions
The mafic inclusions have a diktytaxitic texture and are andesitic in composition (55–60 wt % SiO2). Photomicrographs of representative gray and reddened inclusions are shown in Fig. 3c and d. Subspherical vesicles are abundant. The mafic inclusions only rarely have phenocrysts and these are solely of quartz. Quartz phenocrysts (
1·5 mm) occur only in the mafic inclusions within the Ogatake–Kurumigatake lava dome, and are subhedral in shape. These crystals are larger than the groundmass minerals but sometimes include some of the groundmass phases, indicating that some quartz crystallized contemporaneously with the groundmass minerals. The groundmass is composed of acicular minerals (clinopyroxene, orthopyroxene, ± olivine, and plagioclase), granular Fe–Ti oxides, and interstitial silica minerals and glass. Groundmass olivine is rare and found only in the Toyagamori lava. Some of the groundmass crystals have dendritic textures. Tiny acicular plagioclase and pyroxene grains and tiny granular Fe–Ti oxide grains are more frequently found in the reddened type than in the gray type. Relatively larger granular Fe–Ti oxides are found in the reddened type. The volume of interstitial glasses varies from sample to sample and is c. 10–15%. Usually vesicles can be found in glass of the inclusion. The volume percent of vesicles varies between samples from c. 10 to 25 vol. %. Glasses in the gray type of inclusion are usually clear and sometimes brownish. On the other hand, glasses in the reddened inclusions look dirty because most of the glass is speckled with tiny indistinguishable grains, most of which are probably magnetite. |
WHOLE-ROCK COMPOSITIONS |
|---|
Representative whole-rock analyses of host lavas and mafic inclusions are given in Table 2. The host lavas and mafic inclusions belong to the low-K calc-alkaline series according to the classification scheme of Gill (1981)
and to the medium-Fe suite of Arculus (2003) (Fig. 4).
|
|
Major oxide and trace element variation diagrams are shown in Figs 5 and 6. The silica contents of the host rocks are in the range 70·5–75 wt %. In most diagrams, the compositions of the host rocks define linear trends. MgO, FeO*, CaO, Al2O3, TiO2, MnO, P2O5, Sr and V decrease with increasing SiO2 content, whereas K2O, Na2O, Rb, Ba, Nb, Zr and Y increase. Data are scattered in the Cr diagram; Cr contents tend to decrease with increasing SiO2 content. Ni contents are low and show no systematic variation with respect to SiO2 content. The bulk-rock silica content of each geological unit is distinct: 70·5–71·5 wt % for the Ogatake–Kurumigatake lava dome; 73–75 wt % for the Toyagamori lava; 71·5–72·5 wt % for the Matsugamine lava.
The bulk-rock silica contents of the mafic inclusions are in the range 55–60 wt %. In most diagrams, the data for the mafic inclusions define linear trends. MgO, FeO*, CaO, Al2O3, TiO2, MnO, P2O5 and V decrease with increasing SiO2 content, whereas K2O, Na2O, Rb, Ba, Nb, Zr and Y increase. Sr contents slightly decrease with increasing silica content. Data are slightly scattered in the Cr and Ni diagrams. However, the Cr and Ni contents tend to decrease with increasing silica content. Extrapolations of the mafic inclusion trends to high-SiO2 contents do not coincide with the host rock trends in some diagrams (e.g. MgO, Na2O vs SiO2 diagrams in Fig. 5, and Cr, Ni vs SiO2 diagrams in Fig. 6).
|
|
|
MINERAL COMPOSITIONS |
|---|
Representative chemical compositions of clinopyroxene, orthopyroxene, plagioclase, olivine and Fe–Ti oxides in both hosts and mafic inclusions are given in Tables 3 and 4. The compositions of pyroxenes in both hosts and inclusions are illustrated in terms of their position in the pyroxene quadrilateral for each unit in Fig. 7.
|
|
|
The 100 x Mg/(Mg + Fe) (Mg-value) of most of the phenocryst orthopyroxene cores, as well as rim compositions, in the host rocks is between 58 and 64 in each sample. Orthopyroxene phenocrysts with higher Mg-value cores are rarely found, and those that show apparent reverse zoning in Mg-value are uncommon. Mg-values of most of the groundmass orthopyroxene in the host rocks are 58–64. However, minor proportions of phenocrysts with lower Mg-values (
56) are also present. The Wo contents of orthopyroxene phenocryst core, rim and groundmass are <3 mol %. Although small differences can be seen, the core, rim and groundmass orthopyroxene compositions of the host rocks are similar between units. In the mafic inclusions, the Mg-values of the cores compositions of the groundmass orthopyroxene are higher (65–78). The Mg-values of orthopyroxene rims are lower, ranging from 58 to 64; this is similar to the range for the phenocryst and groundmass orthopyroxene in the host rocks. The Wo content of groundmass cores is c. 3–4 mol %, and is higher than that of the rims (c. 2–3 mol %). The core and rim of groundmass orthopyroxene compositions in the mafic inclusions are similar between units as well as between gray and reddened types.
Clinopyroxene in the host lavas is augite; compositions are similar among phenocryst cores, rims and groundmass (Wo43En41). The Mg-values of the core compositions are around 70 (68–74). Although small differences are present, the core, rim and groundmass clinopyroxene compositions of the host rocks are similar between units.
Mg-values of clinopyroxene core and rim compositions in mafic inclusions are around 70 (65–75), which is similar to the phenocryst clinopyroxene composition of the host rocks. The Wo contents of rims tend to be higher (40–44 mol %) than those of cores (35–43 mol %). The rim compositions are similar to those of the clinopyroxene in the host rocks.
Using the two-pyroxene thermometer of Lindsley (1983), the magmatic temperature deduced from core compositions of clinopyroxene and orthopyroxene phenocrysts in the host rhyolites is about 800–850°C (Fig. 7). Magmatic temperatures deduced from core compositions of clinopyroxene and orthopyroxene in mafic inclusions are higher than 1000°C, whereas those based on rim compositions are about 850°C, which is similar to the temperature of the host rhyolites (Fig. 7).
The Mg-values of groundmass olivine cores in the mafic inclusions of the Toyagamori lava are 69–73, and the rim compositions are 67–73. These values are similar to the Mg-values of the groundmass clinopyroxene cores. Thus groundmass olivine appears to be in equilibrium with the groundmass clinopyroxene core, as Obata et al. (1974) showed that the Mg-value of olivine is nearly equal to or slightly lower than that of clinopyroxene when these minerals coexist in equilibrium.
The plagioclase compositions of both host lavas and inclusions are shown in Fig. 8 as histograms of anorthite content. The compositional range of the phenocryst cores from the host rocks is An40–65 and the peak position is always around An45–50. Plagioclase phenocrysts with anorthite contents >An55 are rare, and are always of the clear type. The subordinate amounts of dusty-type plagioclases always have low An contents. The rim compositions are similar to those of the core composition, but all compositions are <An55. The compositional ranges and the peak positions are similar between units.
|
The groundmass plagioclases in the mafic inclusions exhibit strong normal compositional zoning, with core compositions of An65–90, and rim compositions of An40–60. The larger crystals tend to have higher anorthite contents.
Using the iron–titanium oxide thermobarometer (Stormer, 1983), the magmatic temperature and oxygen fugacity (fO2) indicated by coexisting magnetite–ulvöspinel and ilmenite–hematite solid solutions in the host rhyolites are calculated to be c. 850°C and fO2 slightly lower than the nickel–nickel oxide (NNO) buffer (Fig. 9). The data plot in the lower temperature area of the field of island arc calc-alkaline rocks of Takahashi et al. (1995). This temperature is consistent with that obtained by pyroxene thermometry.
|
|
COMPOSITION OF GROUNDMASS GLASSES |
|---|
Representative compositions of groundmass glasses are shown in Table 5. The SiO2 content of the groundmass glass ranges from 76 to 80 wt %. In the gray inclusions, the glass composition within the inclusions is very similar to that of their host. On the other hand, the groundmass glass composition of both reddened hosts and their inclusions is split into two groups. One group has c. 79–80 wt % SiO2, but lower K2O (c. 0·7–0·9 wt %), higher Na2O (c. 4·5–4·9 wt %), and higher CaO (c. 1·2–2·1 wt %) contents. The other type has c. 76 wt % SiO2, higher K2O (c. 5·0–7·8 wt %), lower Na2O (c. 2·0–3·0 wt %), and lower CaO (c. 0·1–0·5 wt %). Back-scattered electron images of representative occurrences of these two groups are shown in Fig. 10. The ratio of the high-SiO2 to the low-SiO2 type is approximately 8:2. The low-SiO2 glasses occur as isolated blebs in the high-SiO2 glasses or as colloidal skins of perlitic glass particles (Fig. 10). The average glass composition in the reddened type host and its inclusions is similar to the glass composition of the gray type host and inclusions. In most of the variation diagrams (e.g. Fig. 5), the compositions of the glass in both the gray type host and its inclusions plot towards the SiO2-rich extrapolation of the host-rock trends.
|
|
DISCUSSION |
|---|
Explanation of host rock trends
As described above, the chemical compositions of both phenocrysts and groundmass glass in the host rocks are similar between units, whereas the bulk chemical compositions and modal characteristics of each unit are distinct. These data suggest that differences in the bulk-rock compositions between units are controlled by the modal proportions of phenocrysts in each unit.
Many workers have suggested that magma mixing plays an important role in producing the compositional variations of volcanic rocks (e.g. Sakuyama, 1981; Feeley & Davidson, 1994; Ban & Yamamoto, 2002). Mafic inclusions, which provide evidence for magma mingling, are commonly observed in the Narugo lavas. When such mafic inclusions are incorporated, the composition of the host magma must be modified to some degree by mixing with the mafic magma. In the case of Narugo, however, the lavas lack petrographic evidence for magma mixing (hybridization) such as a disequilibrium phenocryst assemblages. Moreover, the features of the whole-rock chemical trends of the Narugo lavas are not consistent with a magma mixing process. If the compositional trends of the host rocks were created by differences in the degree of mixing between mafic magma and the host rhyolitic magma, the mafic inclusions should plot on the extension of the host rock trends to lower SiO2 contents. However, on some variation diagrams (e.g. Figs 5 and 6), the mafic inclusions do not plot on the extensions of the host lava trends. Thus, in the case of Narugo, we consider that magma mingling took place but that magma mixing was not important in producing variations in the bulk-rock chemical composition of the host rocks.
|
In Fig. 12, estimated bulk phenocryst compositions, based on the average phenocryst modal abundances (Table 1) for each unit and the mineral composition data reported in Tables 3 and 4, are shown. The estimated bulk phenocryst compositions lie on the low-SiO2 extrapolation of the host rock trends. Thus it is possible that the variations in whole-rock composition between units were caused by differences in the amount of the different phenocryst phases between the units.
Possible processes to produce the chemical trends among the mafic inclusions and the compositional characteristics of the glasses in both the mafic inclusions and their hosts
It seems clear that the particular type of mafic inclusion found in the Narugo lavas is the quenched product of mafic magma that intruded into cooler felsic magma (e.g. Eichelberger, 1980; Bacon, 1986; Koyaguchi, 1986a). Features such as the ellipsoidal shapes of the inclusions and the textural evidence for rapid crystallization of the groundmass minerals are further evidence of quenching (Eichelberger, 1980; Bacon, 1986).
Several processes can be considered that might have influenced the bulk-rock chemical characteristics of the mafic inclusions. One possibility is that their characteristics reflect mixing between the host magma and the intruding mafic magma. When considering two end-member mixing between mafic and felsic magmas, the mafic end-member can easily change its composition as a result of mixing with the felsic end-member (e.g. Kouchi & Sunagawa, 1985). In this case, the host lava composition should plot on the high-SiO2 extrapolation of the mafic inclusion trends in major and trace element variation diagrams (e.g. Figs 11 and 12). Although this is broadly correct for some major and trace elements, it is not the case in the MgO, Na2O, Cr, Ni vs SiO2 diagrams in Figs 6, 11 and 12.
|
|
The second possibility is that the compositional trends defined by the mafic inclusions result from in situ fractionation of the mafic magma, as suggested by Bacon (1986)
to explain the compositional variations of mafic inclusions in the silicic andesites of Mount Mazama. In this case, when the bulk composition of the mafic inclusions becomes more felsic, the residual glass compositions should also be more felsic. However, in the case of Narugo, the SiO2 contents of the residual glasses are similar in all the mafic inclusions, regardless of the bulk inclusion SiO2 content (Table 2).
The third possibility is that the inclusion trends are the result of interdiffusion of elements between two compositionally distinct magmas, as experimentally proposed by Baker (1991). In this case, alkalis diffuse much more rapidly during interdiffusion of silicate melts and can be decoupled from SiO2. However, in the case of Narugo volcano, such decoupling cannot be detected (Fig. 5).
An important feature of the mafic inclusions in the Narugo lavas is that the chemical compositions of the interstitial glasses are similar to those of the groundmass glasses of the host rocks. This is the key to explaining the chemical trends of the mafic inclusions, and strongly indicates that prior to the final stage of solidification of the mafic and felsic magmas, the residual melts of both the mafic inclusions and their hosts became chemically homogeneous. It is likely that these two melts infiltrated each other and consequently their compositions became similar. The amount of felsic melt that was involved in mixing and magma mingling may have been larger than that of the mafic melt, assumed because of the low proportion (1–2%) of the inclusions based on field observations. The bulk composition of the host lavas was not changed significantly during this process. This mechanism can explain the similarity in the chemical compositions of glasses within the mafic inclusions and their hosts. However, this process alone cannot explain the variation in compositions of the mafic inclusions, because the glass compositions do not plot on the high-SiO2 extrapolations of the mafic inclusion trends (e.g. Fig. 11).
Accordingly, a composite process must be considered to explain the chemical trends of the mafic inclusions. We conclude that before the final stage of solidification, the felsic melt infiltrated the incompletely solidified mafic inclusions, changing the bulk chemical composition of the mafic inclusion towards the groundmass glass composition in the host lava. The estimated chemical trends of this infiltration process and the inferred original variation trends of the mafic inclusions are illustrated in Fig. 12. The original variation may have been caused by variable degrees of differentiation, which took place prior to the magma mingling process.
Solidification sequence of minerals in the mafic inclusions and the environment in which the residual melts in the inclusion and the host lavas became homogeneous
Based on the above discussion, it is inferred that the mafic inclusions experienced chilling by a cooler host magma (Stage 1), and the composition of their residual interstitial melt became more felsic by the crystallization of cores of the acicular minerals (Stage 2). Prior to the final stage of solidification (Stage 3), the interstitial melt composition approached that of the silicic host magmas. A schematic representation of the creation of the mafic inclusions is shown in Fig. 13.
|
The possible crystallization sequences and temperatures at each stage are shown as follows. In Stage 1, the cores of acicular groundmass plagioclase (An70–90), clinopyroxene (Wo35–43), and orthopyroxene with Mg-value 65–78 crystallized. The elongated crystals possibly formed a network structure, which later became the framework of a diktytaxitic texture (Fig. 10). The magmatic temperature when this framework was established is estimated to be higher than 1000°C based on pyroxene thermometry. Subsequently, the rims of the acicular groundmass plagioclase (<An60), clinopyroxene (Wo40–44), and orthopyroxene with Mg-value 58–64 crystallized from a more differentiated melt (Stage 2). The magmatic temperature is estimated to have decreased to c. 850°C on the basis of pyroxene thermometry. This temperature is similar to the crystallization temperature of pyroxene phenocrysts in the host rhyolites. The rim compositions of the groundmass plagioclase and pyroxene are similar to the phenocryst plagioclase and pyroxenes of the host rhyolites. Moreover, the groundmass glasses of both the mafic inclusions and the host rhyolites have similar chemical compositions. These observations suggest that from the initial chilling stage (Stage 1) to crystallization of groundmass minerals (Stage 2), the host rhyolitic magma infiltrated into the framework of the mafic inclusions, and became well mixed with the existing interstitial melt (Stage 3). The average SiO2 content of all groundmass glasses is high (78 ± 2 wt %). This means that the amount of host rhyolitic melt involved in mixing (on the interstitial melt scale) and magma mingling was much larger than the volume of residual melt in the mafic inclusions. Thus the mixed melts have a high SiO2 content. This estimation is supported by the field observation that the approximate proportion of the inclusions is 1–2%. A similarity between the chemical composition of groundmass glasses in both mafic inclusions and their host lavas was shown for silicic rocks in Dikii Greben' volcano (Kamchatka) by Bindeman & Bailey (1994)
. They used this observation to deduce the similarity in density of both magmas. However, they did not propose a mechanism for the similarity. On the other hand, Bacon (1986) reported chemical compositions of residual glasses in mafic inclusions that are different from those of the glass in the host Mt. Mazama andesite. The difference between these two examples may be in the relative amounts of felsic and mafic magmas involved in mixing and magma mingling.
Several ideas have been advanced to define where mafic and felsic magmas might contact and mingle in a magma feeding system. Eichelberger (1980) showed that exsolution of volatiles during crystallization of basaltic magma can cause a density inversion in a stratified basalt–rhyolite magma chamber, and that the vesiculated mafic magma can float up into the silicic magma to form mafic inclusions. Koyaguchi (1985) proposed that mixing and mingling can take place during magma ascent through a conduit. He showed that this largely depends on the difference in the mixing ratio of the end-member magmas, and whether the mafic magma becomes a mafic inclusion in a felsic host or the two magmas mix to become a compositionally homogeneous mixed magma (Koyaguchi, 1986a). Bacon (1986) suggested that the mechanism largely depends on the thermal and compositional contrasts between the two magmas, and that, in most cases, the mafic inclusions originate in a hybrid layer in a stratified basalt–rhyolite magma chamber. Such hybridization can take place in the boundary layer between the mafic and felsic magmas. This might be established during the injection and ponding of hot mafic magma at the base of a silicic magma chamber (e.g. Eichelberger, 1975; Koyaguchi, 1986b; Bacon & Druitt, 1988; Feeley & Davidson, 1994; Tomiya & Takahashi, 1995; Feeley & Dungan, 1996).
In the case of Narugo volcano, the following scenario is likely. Most inclusions lack olivine or Mg-rich pyroxene, which shows that the mafic magma was itself already differentiated, perhaps in a boundary layer within the stratified magma chamber. When the temperature dropped below the liquidus (Stage 1), acicular minerals began to crystallize and the composition of the residual melt evolved towards that of the felsic magma and mixed with the chamber magma, especially near the boundary (Stage 2). Volatile components may be continually supplied to the crystallizing boundary layer from deeper parts of the mafic magma layer. Although some of the volatiles might be released to the felsic layer along with the melts in the mafic layer, the volatile phases would be diluted in the larger felsic layer. Finally, the density of some parts of the boundary layer became lower than the overlying felsic magma and detached blobs (Eichelberger, 1980) or boudin-shaped waves (Koyaguchi, 1986a) of the mafic magma began to float up into the felsic magma. In the magma conduit during ascent to the surface, these mafic magma blobs were not completely solidified and thus may have become disrupted into smaller pieces. In the case of Narugo, it is deduced that the mafic inclusions did not solidify completely until the host magma solidified, because the glass compositions can be split into two types, high-K and low-K in both reddened type host lava and inclusions. One of the possibilities to produce the two kinds of glasses is a local difference of mineral assemblage and proportion of crystallized groundmass phases; however, such a difference cannot be observed. Another possibility is that some kind of liquid immiscibility may be induced at high oxidation states; however, we do not have sufficient data to prove this possibility. Regardless, the existence of high-K and low-K glasses in both reddened type host lava and inclusions suggests that the residual melts of the mafic inclusions solidified after the host magma was extruded at the Earth's surface.
Timing of solidification of the mafic inclusions
Pallister et al. (1992) estimated that in the case of the Pinatubo 1991 eruption, injection of basaltic magma from deeper levels into a shallower felsic magma chamber caused magma mixing and triggered the eruption. They considered that the preservation of a disequilibrium mineral assemblage suggested that magma mixing took place shortly before eruption. Murphy et al. (2000) suggested that the 1995–1999 eruption of the Soufrière Hills volcano, Montserrat, was triggered by a recent influx of hot mafic magma, based on petrological observations as well as data on seismicity, extrusion rate, and SO2 fluxes. However, these petrological observations do not provide direct constraints on the timing of mafic magma intrusion events beneath Pinatubo and the Soufrière Hills volcano. If the magmas mixed well and some phenocrysts were out of equilibrium with the surrounding melt, the residence time of these phenocrysts in the mixed (hybrid) magma could be estimated using geo-speedometers, such as the Ni content in olivine (Nakamura, 1995b), or the dissolved width of plagioclase (Nakamura & Shimakita, 1998). However, in the case of (complete) magma mingling, such disequilibrium phenocrysts cannot be found. Practically, when magma mingling is a dominant process, small-scale mixing can also take place especially along the boundary between the end-member magmas (Bacon, 1986). However, even in such cases, disequilibrium phenocrysts are rare. In these cases, it is difficult to use geo-speedometers to estimate the time lapse between initiation of the mingling event and a subsequent eruption.
In the case of Narugo, disequilibrium phenocrysts are rare, and the major evidence for magma mingling is solely the existence of the mafic inclusions. As already discussed, it is widely accepted that mafic inclusions are the products of quenched mafic magma. Thus it must take only a short time from the initial chilling to the complete solidification of the mafic magma to form the mafic inclusions. We conclude that the mafic inclusions and their host rhyolites completely solidified almost simultaneously after the eruption, based on the chemical compositions of the residual glass in both the reddened type of mafic inclusion and the host rhyolite. In other words, shortly after the mafic magma invaded the host rhyolitic magma chamber, these magmas mingled and erupted to the surface.
|
ACKNOWLEDGEMENTS |
|---|
We are grateful to R. J. Arculus and Marjorie Wilson for many constructive comments and suggestions on the manuscript. We express our thanks to Professors T. Yoshida at Tohoku University and Y. Yamaguchi at Shinshu University for helpful suggestions, Dr. R. W. Jordan at Yamagata University for correcting the English in this paper, and Professor H. Tanaka at Yamagata University for his continual support of this research. Technical advice on XRF analyses from Y. Yamada at Rigaku Co., Ltd. was very helpful. We also appreciate the financial support from the Japanese Ministry of Education.
* Corresponding author. Telephone: +81-23-628-4642. Fax: +81-23-628-4661. E-mail: ban{at}sci.kj.yamagata-u.ac.jp
|
REFERENCES |
|---|
Arculus, R. J. (2003). Use and abuse of the terms calcalkaline and calcalkalic. Journal of Petrology 44, 929–935.
Bacon, C. R. (1986). Magmatic inclusions in silicic and intermediate volcanic rocks. Journal of Geophysical Research 91, 6091–6112.
Bacon, C. R. & Druitt, T. H. (1988). Compositional evolution of the zoned calcalkaline magma chamber of Mount Mazama, Crater Lake, Oregon. Contributions to Mineralogy and Petrology 98, 224–256.[CrossRef][ISI]
Baker, D. R. (1991). Interdiffusion on hydrous dacitic and rhyolitic melts and the efficacy of rhyolite contamination of dacitic enclaves. Contributions to Mineralogy and Petrology 106, 462–473.[CrossRef][ISI]
Ban, M. & Yamamoto, T. (2002). Petrological study of Nasu-Chausudake Volcano (ca. 16 ka to present), northeastern Japan. Bulletin of Volcanology 64, 100–116.[CrossRef][ISI]
Bindeman, I. N. & Bailey, J. C. (1994). A model reverse differentiation at Dikii Greben' Volcano, Kamchatka: progressive basic magma vesiculation in a silicic magma chamber. Contributions to Mineralogy and Petrology 117, 263–278.[CrossRef][ISI]
Blake, S. & Ivey, G. N. (1986). Magma-mixing and the dynamics of withdrawal from stratified reservoirs. Journal of Volcanology and Geothermal Research 27, 153–178.[CrossRef][ISI]
Campbell, I. H. & Turner, J. S. (1989). Fountains in magma chambers. Journal of Petrology 30, 885–923.
Cioni, R., Civetta, L., Marianelli, P., Metrich, N., Santacroce, R. & Sbrana, A. (1995). Compositional layering and syn-eruptive mixing of a periodically refilled shallow magma chamber: the AD 79 plinian eruption of Vesuvius. Journal of Petrology 36, 739–776.
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.[Abstract]
Eichelberger, J. C. (1980). Vesiculation of mafic magma during replenishment of silicic magma reservoirs. Nature 288, 446–450.[CrossRef]
Feeley, T. C. & Davidson, J. P. (1994). Petrology of calc-alkaline lavas at Volcán Ollagüe and the origin of compositional diversity at central Andean stratovolcanoes. Journal of Petrology 35, 1295–1340.
Feeley, T. C. & Dungan, M. A. (1996). Compositional and dynamic controls on mafic–silicic magma interactions at continental arc volcanoes: evidence from Cordon El Guadal, Tatara–San Pedro Complex, Chile. Journal of Petrology 37, 1547–1577.
Feeley, T. C., Dungan, M. A. & Frey, F. A. (1998). Geochemical constraints on the origin of mafic and silicic magmas at Cordon El Guadal, Tatara–San Pedro Complex, central Chile. Contributions to Mineralogy and Petrology 131, 393–411.[CrossRef][ISI]
Gill, J. B. (1981). Orogenic Andesites and Plate Tectonics. New York: Springer, 390 pp.
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.[CrossRef][ISI]
Koyaguchi, T. (1985). Magma mixing in a conduit. Journal of Volcanology and Geothermal Research 25, 365–369.[CrossRef][ISI]
Koyaguchi, T. (1986a). Textural and compositional evidence for magma mixing and its mechanism, Abu volcano group, Southwestern Japan. Contributions to Mineralogy and Petrology 93, 33–45.[CrossRef][ISI]
Koyaguchi, T. (1986b). Evidence for two-stage mixing in magmatic inclusions and rhyolitic lava domes in Niijima Island, Japan. Journal of Volcanology and Geothermal Research 29, 71–98.[CrossRef][ISI]
Lindsley, D. H. (1983). Pyroxene thermometry. American Mineralogist 68, 477–493.[Abstract]
Miyashiro, A. (1974). Volcanic rock series in island arcs and active continental margins. American Journal of Science 274, 321–355.[Abstract]
Murayama, I. (1978). The Narugo volcano. In: Volcanoes in Japan (I). Tokyo: Daimeido, 239 pp. (in Japanese).
Murphy, M. D., Sparks, R. S. J., Barclay, J., Carroll, M. R. & Brewer, T. S. (2000). Remobilization of andesitic magma by intrusion of mafic magma at the Soufrière Hills volcano, Montserrat, West Indies. Journal of Petrology 41, 21–42.
Nakamura, M. (1995a). Continuous mixing of crystal mush and replenished magma in the ongoing Unzen eruption. Geology 23, 807–810.
Nakamura, M. (1995b). Residence time and crystallization history of nickeliferous olivine phenocrysts from the northern Yatsugatake volcanoes, Central Japan: application of a growth and diffusion model in the system Mg–Fe–Ni. Journal of Volcanology and Geothermal Research 66, 81–100.[CrossRef][ISI]
Nakamura, M. & Shimakita, S. (1998). Dissolution origin and syn-entrapment compositional change of melt inclusion in plagioclase. Earth and Planetary Science Letters 161, 119–133.[CrossRef][ISI]
Obata, M., Banno, S. & Mori, T. (1974). The iron–magnesium partitioning between naturally occurring coexisting olivine and Ca-rich clinopyroxene: an application of the simple mixture model to olivine solid solution. Bulletin de la Société Française de Minéralogie et de Cristallographie 97, 101–107.
Omoto, K. (1993). Radiocarbon ages of organic materials collected from Narugo basin, Miyagi Prefecture. Quaternary Research 32, 227–229 (in Japanese).
Pallister, J. S., Hoblitt, R. P. & Reyes, A. G. (1992). A basalt trigger for the 1991 eruptions of Pinatubo volcano? Nature 356, 426–428.[CrossRef]
Sakaguchi, K. & Yamada, E. (1988). ‘The Kitagawa Dacite’, pyroclastic flow deposits around the Onikobe caldera, northeast Japan. Report of the Geological Survey of Japan 268, 37–59 (in Japanese with English abstract).
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.
Soda, T. & Yagi, K. (1991). Quaternary tephra studies in the Tohoku district northeastern Honshu, Japan. Quaternary Research 30, 369–378 (in Japanese with English abstract).
Stormer, J. C. (1983). The effects of recalculation on estimates of temperature and oxygen fugacity from analyses of multicomponent iron–titanium oxides. American Mineralogist 66, 1189–1201.
Takahashi, M., Noguchi, T. & Tagiri, M. (1995). The REE composition of Miocene icelandite in Northeast Japan, and implication for the origin of icelandite magma. Memoirs of the Geological Society of Japan 44, 65–74 (in Japanese with English abstract).
Tomiya, A. & Takahashi, E. (1995). Reconstruction of an evolving magma chamber beneath Usu Volcano since the 1663 eruption. Journal of Petrology 36, 251–274.
Tsuchiyama, A. (1985). Dissolution kinetics of plagioclase in melt of the system diopside–albite–anorthite, and origin of dusty plagioclase in andesite. Contributions to Mineralogy and Petrology 89, 1–16.[CrossRef][ISI]
Umeda, K., Hayashi, S., Ban, M., Sasaki, M., Ohba, T. & Akaishi, K. (1999). Sequence of the volcanism and tectonics during the last 2·0 million years along the volcanic front in Tohoku district, NE Japan. Bulletin of the Volcanology Society of Japan 44, 233–249 (in Japanese with English abstract).
Vance, J. A. (1965). Zoning in plagioclase: patchy zoning. Journal of Geology 73, 636–651.
Yamada, Y., Kohno, H. & Murata, M. (1995). A low dilution fusion method for major and trace element analysis of geological samples. Advances in X-Ray Analysis 26, 33–44 (in Japanese with English abstract).
CiteULike 
Connotea 
Del.icio.us What's this?
| |||||||