Morphological Component Analyses of Volcanic Ash from Kikai-Akahoya Tephra: Possible Evidence for Magma-Water Interaction in a Lar ge Scale Volcanic Eruption

Tadashi Kudo, Hiroaki Sato*, and Keiko Suzuki-Kamata Department of Earth and Planetary Science Faculty of Science and Graduate School of Science and Technology Kobe University, Kobe 657, Japan phone: 078-803-0567, fax: 078-803-0490 e-mail: hsato@kobe-u.ac.jp

Abstract
We propose morphological component analyses of volcanic ash in large scale pyroclastic eruptions through classifing the volcanic ash fragments in terms of vesicularity and average vesicle size. The lateral variation of th e component analyses on the Kikai-Akahoya eruption products verify the usefu lness of the component analyses showing the poorly vesiculated ash, represen ting the quenching effect in the column, are more abundant in the pyroclasti c flow deposit compared with the co-ignimbrite ash fall deposit. It is pla usible that the poorly vesiculated ash fragments were quenched by interactio n of magma and surface water during vesiculation, and such fragments conform ed cooler lower pyroclastic flows, and vesicular fragments representing less er magma-water interaction conformed upper expanding pyroclastic flow which subsequently elutriated to form co-ignimbrite column

1. INTRODUCTION

Large explosive eruptions generate vesiculated volcanic ash. Some o f these eruptions are affected by the interaction of magma and external wate r. Heiken and Wohletz(1981) and Wohletz and Heiken (1985) described and dis cussed the morphology of volcanic ash formation in various volcanic deposits , and noted that blocky ash fragments are typical for phreato-magmatic erupt ions, whereas, well vesiculated ash fragments are common constitutent of mag matic explosions. Machida and Arai (1993) also decribed and classified volca nic ash into blocky, bubble-wall, and pumiceous types based on abundant obse rvation of the morphology of ash deposits in Japan and surrounding areas. On the other hand, vesiculation process in depressurizing magmas have been inve stigated in terms of numerical (Toramaru, 1995) and artificial experiments ( Hurwitz and Navon, 1995). These studies suggest that number density of vesic les in the ejecta depends on viscosity of magmas and depressurization rate, and also on the presence or absence of microlite crystas in the magma. In t he present work, we propose to classify the ash fragments in terms of vesicu larity and average vesicle size, and conducted component analyses on the ash deposits of the Kikai-Akahoya eruption, and discuss the genetic implications of the component analyses in respect of magma-water interaction.

Kikai-Akahoya eruption, 6300 y.b.p., produced ca. 200 km3 of pyroclastic deposits, and formed a caldera of ca. 12 kilometers across to the south of t he Kyushu Island, western Japan (Fig.1, 2). The caldera is now under 100-5 00 meters water depth, and it is envisaged that magma-sea water interaction occurred to some degree during the eruption. The crater diameter versus eje cta volume relationship also indicates that the Kikai-Akahoya erutpion had l arger kinetic energy of explosion to thermal energy than normal magmatic exp losions, suggesting possible effect of magma-water interaction (Sato and Tan iguchi, 1997) . The stratigraphy of the Kikai-Akahoya tephra deposit is, fr om lower to upper, Funakura Plinian fall deposit (20 km3), Funakura pyroclas tic flow deposit (less than 1 km3), Koya-Takeshima pyroclastic flow deposit (50 km3), and co-ignimbrite Akahoya ash fall deposit (>100 km3). Among thes e deposit, only Funakura pyroclastic flow deposit is mildly welded. Fig. 3 shows the columnar section of the Kikai-Akahoya deposit at Takeshima (calder a rim), Mizunoshiri (60 km form the center of caldera), and Aso (360 km). Fi g. 3 also shows the position of the sample we analyzed. We carried out grai n size analyses on the samples (Fig. 4). It is noted that the Mdf-sf relati ons of the Koya-Taksehima pyroclastic flow and Akahoya co-ignimbrite ash at Mizunoshiri section plot at the transition of pyroclastic fall and flow depo sit defined by Walker (1971). The Koya-Takeshima pyroclastic flow is one of the low-aspect ratio ignimbrite (Ui, 1973), and the Mizunoshiri section repr esents the distal facis of the deposit. The sieved ash samples in the size classes of 1-2 mm, 0.25-0.5 mm, and 0.063-0.125mm were mounted in epoxy resi n, and made into polished thin sections to carry out morphological component analyses.

3. MORPHOLOGICAL CLASSIFICATION OF THE VOLCANIC ASH

The glassy ash grains were morphologically classified into seven types, i .e., type-A: bubble-wall glass shard, type-B: bubble-wall glass shard with l arge vesicles, type-C: low vesicularity-large vesicle glass, type-D: low ves icularity-small vesicle glass, type-E: high vesicularity-elongated vesicle g lass, type-F: low vesicularity-elongated vesicle glass, and type-G: coalesce d and deformed vesicle glass. Average vesicle size was arbitrarily chosen at 40 micron meters to distinguish large and small vesicles. Vesicularity w as also classified into high and low at the porosity of 60 %. For bubble wa ll glass shards, vesicle size is defined by the apparent diameter of a circl e in contact with the most concave boundary, and the vesicularity is estimat ed by the areal ratio of the void forming the vesicle and the half thickness of the wall multiplied by the length. Fig. 5 shows the typical ash fragment s in a digitized photomicrographs. These fragments are analyzed in detail, and the obtained vesicularity and vesicle size are noted in the figure capti on. Fig. 6 presents representative compositional diagram of the morphologic al component of the volcanic ash for grain size fraction of 0.25-0.50 mm. I t is noted that average vesicle size is roughly correlated with vesicularity for most of the sample. The left hand side diagrams of Fig. 6 represents sa mples from Takeshima, at the caldera rim. The lower plinian fall deposit (2 3a, 23b) are more enriched in type G (pumice with coalesced and deformed ves icles) compared with the pyroclastic flow deposit (23f, 23g). It is also no ted that bubble wall glass constitute only 20-30 % of the fragments in the d eposits of Takeshima. Compared with the samples of Takeshima, the Mizunoshi ri samples are much depleted in the low vesicularity glass fragments (type-C , -D). The Akahoya ash are enriched in bubble-wall glass fragments (type-A, -B). The Akahoya ash from the Aso area has the highest content of bubble-wa ll type glass fragments (87%).

4. VARIATIONS OF THE MORPHOLOGICAL COMPONENT ANALYSES

The morphological component analyses show variations against grain size, stratigraphic position and type of pyroclastic deposit, and also distance fr om the caldera area. Morphological component analyses is dependent on the grain size fraction of the sample ananlyzed. For Funakura plinian fall deposit, larger fraction (1 -2 mm) are mostly composed of type-B, -E and -G grains, the well vesiculated pumices, and lack the poorly vesiculated fragments. There are no distinct c ompositional difference for samples of different grain sizes in Koya-Takeshi ma pyroclastic flow deposit.

At Takeshima, stratigraphic variation of the morphological component an alyses on 0.25-0.5 mm size fraction is not conspicuous in terms of both vesi cularity and average vesicle size (Fig. 6). However, at Mizunoshiri, 60 k m from the source area, type-A and B (bubble-wall glass shards) are more enr iched in the upper pyroclastic flows and Akahoya co-ignimbrite ash. They do not show marked difference in the content of poorly-vesiculated fragments (F ig. 6). Lateral variation of the morphological component analyses is not d istinctive for the Funakura plinian fall deposit, whereas, for the Koya-T akeshima pyroclastic flow deposit, distant sample contains more bubble-wal l glass shards (type-A and -B), and lesser amount of poorly vesiculated glas s grains (type-C, -D, -F). Compared with the pyroclastic flow deposit, the co-ignimbrite Akahoya ash contains more abundant bubble-wall type glass frag ments, and distant sample at Aso contains the highest proportion of such ves icular and large vesicle glass grains (Fig. 7), suggesting differences of th e mode of ash formation, segregation during transportation, and settling.

5. GENETIC IMPLICATIONS

Most of the Plinian eruption produces bubble wall glasses. However, pres ence of poorly vesiculated glass grains in the deposit suggest possible quen ching effect by magma-water interaction through drastically lowering the t emperature of the ash laden cloud in the vent. The cooler volcanic ash colum n, including abun-dant poorly vesiculated ash, tends to collapse, generating pyroclastic flows (Koyaguchi and Woods, 1996).

Fig. 8 is a cartoon illustrating the mode of eruption in the later phase of the Kikai-Akahoya eruption. Earlier plinian fall deposit is rather unif orm in terms of the morphological component analyses of the ash grain, conta ining both vesiculated and poorly-vesiculated ash grains, which suggest thor ough mixing of fragments in the eruption column. In the later phase of erup tion as depicted in Fig. 8, pyroclastic uprush in the vent area encounters s ea water of 100-500 m depth, and the central part of the erupting column may penetrate the sea water with little magma-water interaction, whereas the mar ginal part of the erupting column interact with sea water. It is envisaged that magma/water ratio varies from place to place in the marginal part of th e erupting column, and generation of abundant water vapor cause much intense bulging of the eruption column. However, strong magma-water interaction dec rease the temperature of the eruption column, subsequently inducing column collapse and pyroclastic flows may occur. Cooler part of the column is dens er and conform a lower pyroclastic flows, which contains abundant quenched p oorly vesiculated glass fragments. Hotter part of the column are lighter an d overlies the cooler pyroclastic flow, and after elutriating denser fragmen ts during transportation, it will turn to be buoyant to form co-ignimbrite a sh columns. In this scenario, we assumed that poorly vesiculated glassy fra gments represent quenching effect of magmas by magma-water interaction in th e vent area. This idea is plausible but not proven at present, and further examination in respect of numerical modeling of vesiculation processes in la rge-scale plinian eruption should be taken into account.

6. CONCLUDING REMARKS

Followings are the major conclusions of the present analyses.
(1) Morphology of volcanic ash from the 6300 yr.b.p. Kikai-Akahoya eruption products are described in terms of vesicularity and average vesicle size, an d classified into seven types, which can be grouped as bubble-wall glass sha rds, poorly vesiculated glass, and finely-vesiculated pumice fragments.
(2) Stratigraphic and lateral variations of the morphological component anal yses of ash fragments in the deposit show marked upward increase of well ves iculated ash and decrease of poorly vesiculated glass fragments in the dista nt sections, which may be reconciled with heterogeneous quenching effect by external water on the eruption column, such that higher magma-water interact ion caused cooler and denser batch of the column, which subsequently flowed and deposited as pyroclastic flow deposit, whereas lesser magma-water intera ction formed hotter pyroclastics rich in vesicular and large vesicle size as h grains, which may be elutriated to form co-ignimbrite ash.
As a whole, morphological component analyses of volcanic ash in pyroclas tic deposits record processes acting during vesiculation and fragmentation o f magmas near the vent and may be requisite in understanding mechanisms of p yroclastic eruptions.

REFERENCES

Heiken, G. and Wohletz, K. (1985) Volcanic ash. Univ. of California Press, Berkeley, 246pp.
Hurwitz,S. and Navon, O. (1994) Bubble nucleation in rhyolitic melts: expe riments at high pressure, temperature, and water content. Earth Plan. Sci. Lett., 122, 267-280.
Koyaguchi, T. and Woods, A. (1996) On the formation of eruption columns foll owing explosive mixing of magma and surface-water. J. Geophys. Res., 101, 5 561-5574.
Machida, H. and Arai, F. (1992) Atlas of tephra in and around Japan. Tokyo Univ. Press, 276p.
Sato, H., and Taniguchi, H. (1997) Relationship between crater size and eje cta volume of recent magmatic and phreato-magmatic eruptions: implications for energy partitioning. Geophys. Res. Lett., 24, 205-208.
Toramaru, A. (1995) Numerical study of nucleation and growth of bubbles in viscous magmas. J. Geophys. Res., 100, 1913-1931.
Ui, T. (1973) Exceptionally far-reaching, thin pyroclastic flow in southern Kyushu, Japan. Bull. Volcanol. Soc. Japan, 18, 153-168.
Walker, G.P.L. (1971) Grain-size characteristics of pyroclastic deposits. J. Geol., 79, 696-714.
Wohletz,K.H.,(1983) Mechanisms of hydrovolcanic pyroclast formation: grain- size, scanning electron microscopy, and experimental results. J. Volcanol. Geotherm. Res., 17, 31-63.

Figure 1 Locality map, showing Kikai Caldera, Aso, and isopachs of the Kik ai-Akahoya co-ignimbrite ash fall deposit in the Japanese Island (Machida an d Arai, 1991).

Figure 2 Locality map, showing Kikai Caldera, Takeshima, Mizunoshiri, an d isopachs of the Funakura pumice fall deposit, and distribution of the Koya -Takeshima pyroclastic flow deposit.

Figure 3 Columnar section of the Kikai-Akahoya ash deposit in Takeshima, Mizunoshiri, and Aso localities (the thickness of the column, not to scale). The position of the samples are also shown.

Figure 4 Mdpfai-sigma pfai relationship for Kikai-Akahoya deposits.

Figure 5 Typical digitized images of the ash grains included in the Kikai -Akahoya deposit. A: type-A, vesicularity 92 %, vesicle size 360 mm, the s ame as follows, B: type B, 76%, 150 mm, C: type-C, 36 %, 50 mm, D: type-D, 2 5 %, 15 mm, E: type-E, 65 %, 10 mm, F: type-F, 54 %, 15 mm, G: type-G, 81 %, 85 mm.

Figure 6

Figure 7 Lateral variations of the morphological component analyses of th e Kikai-Akahoya deposit. The samples represents Takeshima, Mizunoshiri, and Aso localities.

Figure 8 A cartoon of the mode of eruption of the Kikai-Akahoya eruptio n. The collapsing eruption column are divided into cooler and denser basal pyroclastic flow and hotter and lighter upper pyroclastic surges which subse quently elutriate to form co-ignimbrite ash column.