Silicic volcanism and plutonism in the IBM arc

Transcription

1 NSF-IFREE MARGINS Subduction Factory Workshop Hawaii, 8-12 September 2002 MARGINS Web Site: Silicic volcanism and plutonism in the IBM arc Yoshihiko Tamura IFREE, JAMSTEC, Yokosuka , JAPAN Telephone: Fax: INTRODUCTION Although dacitic and rhyolitic material is uncommon at the southern Izu-Bonin and Mariana arcs (Bloomer et al., 1989), modern magmatism at the northern Izu-Bonin arc is bimodal with basalt and rhyolite predominating (Tamura & Tatsumi, 2002). The origin of rhyolite in oceanic arcs is a matter of considerable interest. We suggest that this rhyolite is a partial melt of calc-alkaline andesite occurring at depth within the oceanic island arc crust. An original calc-alkaline andesite magma is likely to be water-saturated and will therefore solidify in the crust, forming an andesite source region at depth, that could be reheated and remobilised by influxes of basalt. BIMODAL VOLCANISM One thousand and eleven chemical analyses of samples from 17 Quaternary volcanoes of the Izu- Bonin arc (30 N ~ 35 N) (Fig. 1) were reviewed to estimate the relative proportions of erupted magmas (Tamura & Tatsumi, 2002). Basalt and basic andesite (<57 wt % SiO 2 ) are clearly the predominant eruptive products of the Izu-Bonin arc, but rhyolite (>70 wt % SiO 2 ) also forms a major mode (Fig. 2). The latter peak would be emphasised even more strongly if pumices from submarine caldera-forming eruptions, dispersed far from their source volcanoes, were taken into consideration. For example, more than half of sediment layers drilled in the Sumisu Rift during ODP Leg 126, Sites 790 and 791 (Fig. 1), were made up of thick layers of two-pyroxene rhyolite pumice derived from nearby arc volcanoes (Nishimura et al., 1992). Although the eruptive products are volumetrically bimodal, magmas range from basalt, through andesite and dacite, to rhyolite. Turbidites, which have been delivered to the forearc largely through submarine canyons, provide a more complete record of arc volcanism than arc lavas, and are more voluminous and proximal than ashes (Gill et al., 1994). Hiscott & Gill (1992) and Gill et al. (1994) characterised the turbidite geochemistry of the Izu-Bonin arc by using 271 samples of volcaniclastic sand and sandstone collected from cores at the six ODP Leg 126 sites (787, 788, 790, 791, 792 and 793). These turbidites are andesitic on average (~60 wt % SiO 2 ) but are bimodal in detail (Gill et al., 1994). A large amount of silicic magma production, comparable with the volumes of basalt and basic andesite produced in the arc, is required to yield an average SiO 2 content of 60 wt %.

2 RELATIVELY DRY BASALT Basalt magmas on the volcanic front of the Izu-Bonin arc are either anhydrous or contain very little water (see references in Tamura & Tatsumi (2002)). Lower H 2 O contents in basaltic melts will result in higher melt density so that plagioclase crystals will be likely to float rather than sink. There is, for example, petrological evidence that plagioclase phenocrysts have accumulated in the upper parts of magma chambers beneath Izu-Oshima volcano where plagioclase phenocryst contents in lavas vary between 0 and 20 % by volume and mafic phenocrysts are rare, yet all the lavas show similar groundmass (melt) compositions. For plagioclase to float in a basaltic liquid, the H 2 O content must be less than 0 7 %. Aramaki & Fujii (1988) demonstrated that an average Izu-Oshima basalt can be generated by fractionation of a primary olivine basalt by removing ~50 wt. % crystals. Consequently, <0 4 % H 2 O would have been present in such primary basalts. WET CALC-ALKALINE ANDESITE Arculus and Bloomfield (1992) studied ashes recovered from ODP Leg 125 (Sites 782, 784 and 786) from the Izu-Bonin arc. These ashes are made up of volcanic glass, strictly representing liquid compositions. Figure 3 shows Miyashiro diagrams (FeO*/MgO vs. SiO 2 ) of the Izu-Bonin Arc ashes (Arculus & Bloomfield, 1992). Firstly, liquid compositions range from basalt to rhyolite. Secondly, tholeiitic andesites (FeO*/MgO > SiO ) exist in a state of liquid, but there are no analyses to indicate that calc-alkaline andesites (FeO*/MgO < SiO ) are erupted in a liquid state. Similarly, glasses of Mariana Trough fallout tephra contrast with the contemporaneous basaltic to dacitic lavas of the Mariana arc volcanoes (Straub, 1995); all compositions of glasses with SiO 2 < 57 % plot in the tholeiitic field, whereas FeO*/MgO vs SiO 2 variation in the Mariana arc volcanoes shows a complete chemical gradation from the tholeiitic to calc-alkaline fields (Straub, 1995). One possible explanation for the curious lack of calc-alkaline andesitic melts in the Izu- Bonin and Mariana arcs is that such liquids were originally water-saturated. Water-saturated liquidi have negative slopes in P-T space and this results the crystallization of these magmas before they can be erupted at the surface. For example, a water-saturated andesitic composition has a liquidus temperature of 970 C at 5 kb, but the liquidus temperature rises to 1200 C at 1 atm (Green, 1982). GENESIS OF RHYOLITE Hydrous basalt and/or andesite are likely source rocks from which rhyolites could be produced by partial melting. It is commonly accepted that rhyolites form by the melting of hydrous basaltic rocks in the crust (Beard, 1995), but basaltic magmas along the volcanic front of the Izu-Bonin arc are inferred to be almost anhydrous, and fractional crystallization would be inevitable within their crustal magma chambers. In other words, these basalt magmas cannot solidify within the crust without undergoing significant differentiation. On the other hand, the absence of calc-alkaline andesite liquids along the Izu-Bonin arc (Fig. 3) and the Mariana volcanic arc (Straub, 1995) suggests that liquids with these compositions could be water-saturated, causing them to solidify at depth and never erupt to the surface. Thus, the partial melting of solidified calc-alkaline andesites rather than basalts might play an important role in producing rhyolite. Rhyolites of the Izu-Bonin arc possess major element compositions similar to those of melts produced by dehydration melting of calc-alkaline andesite at low pressure (< 7kb) (Beard & Lofgren, 1991). The fundamental role of basalt magma in the genesis of rhyolite would therefore be to provide heat to masses of solidified calc-alkaline andesite at depth.

3 In Costa Rica, where another setting that contains no continental crust is made, the volcanic products are also chemically bimodal. Hannah et al. (2002) suggest that the silicic melts resulted from partial melting of relatively hot, evolved calc-alkaline rocks that were previously emplaced and ponded at the base of an over-thickened basaltic crust. CALC-ALKALINE PLUTONISM The genesis of rhyolite in the Izu-Bonin arc leads us to an interesting conclusion that a much larger amount of calc-alkaline andesites (3~5 times greater than the rhyolites) is concealed at depth within the oceanic arc crust. A detailed structural model across the Izu-Bonin arc along N (Suyehiro et al., 1996) indicates that the oceanic arc contains a middle crust with a P-wave velocity of ~6 km/s and occupying about 25 % of the crustal volume. Moreover, this middle crustal component is confined beneath the arc and is absent beneath the Shikoku back-arc Basin. The relatively low velocity (~6km/s) and its small increase with depth in the middle crust beneath the arc might be attributable to granitic rocks (Suyehiro et al., 1996). Based on the geological and petrological studies, Kawate & Arima (1998) suggested that the middle crust of the Izu-Bonin arc would be similar to the Miocene Tanzawa plutonic complex, central Japan, which is a tonalitic suite exposed at the northern end of the Izu-Bonin arc system (Fig. 1). The most voluminous intrusions in this suite comprise rocks with ~60 wt % SiO 2 (Kawate & Arima, 1998). It is thus possible that voluminous calc-alkaline andesites may have accumulated in the oceanic arc middle crust even though it is generally accepted that andesitic volcanism typifies continental arcs. TWO-STEP MODEL FOR GENESIS OF SILICIC MAGMAS Our contention is that mantle-derived hydrous magnesian andesite, not basalt magmas, may be parental to the calc-alkaline series rocks in the Shirahama Group (Tamura, 1994). Tamura (1994) developed this hypothesis based on the following interpretations: (a) a mantle diapir consisting of hydrous peridotite is formed in the lower part of the mantle wedge above the slab and ascends buoyantly through the mantle wedge, (b) this diapir is heated during ascent through the hot and dry mantle wedge (thorough hot fingers in Tamura et al. (2002)) and (c) finally the heated diapir, which still has a wet and cool interior and heated dry and hot rind, produces both magnesian andesite and basalt, respectively (Tamura, 1994). Hirose s (1997) experiments are consistent with this hypothesis. He showed that wet magnesian andesite magma (54.4 wt % SiO 2 and 6 wt % MgO on anhydrous basis) containing 6.3 wt % H 2 O is produced by melting of lherzolite KLB-1 with 1 wt % H 2 O at 1 GPa and a temperature of 1050 C. The same peridotite produces basalt (50.5 wt % SiO2 and 10.1 wt % MgO) at 1 GPa and a temperature of 1300 C under dry conditions (Hirose & Kushiro, 1993). The degrees of melting are 16 wt % and 12 wt %, respectively. Low H 2 O contents in pre-eruptive calc-alkaline andesite magmas, which are not saturated with water, are commonly highlighted as one of the major problems inherent in fractionation models from hydrous magnesian andesite, because H 2 O contents are too low to have been produced by crystal fractionation of H 2 O-rich mantle derived magmas (e.g. Tamura, 1995). Given solidification and remobilization of calc-alkaline magma, H 2 O contents can no longer be a constraint for genesis of calc-alkaline andesite. Water-saturated magmas solidify in the crust. Melting of such a solidified body produces water-deficient magmas; partial melting of largely solidified pluton cannot produce volatile-rich silicic magmas, because the volatile content of the constituent minerals is too low (Matthews et al., 1999).

4 Figure 4 shows the two-step model for genesis of calc-alkaline magmas. The primary step is anhydrous or hydrous melting of mantle peridotite, which produces basalt or magnesian andesite, respectively. Tholeiitic series rocks are produced by fractionation of basalt (Fig. 4b), but water-saturated magnesian andesite and/or calc-alkaline series rocks, which are produced by fractionation of magnesian andesite, will solidify in the crust (Fig. 4b). Thus, calc-alkaline magmas would not appear on the surface without reheating and the second-stage melting by hot and dry basalt magmas, which are produced in the same mantle diapir (Tamura, 1994, Fig. 4c). Tamura et al. (2000) presented evidence for supercooling of arc basalt at Daisen volcano, Japan; such a process would be complementary to remelting. REFERENCES Sources of analytical data of the Quaternary volcanoes of the Izu-Bonin arc are listed in Tamura & Tatsumi (2002). Aramaki, S. & Fujii, T. (1988). Petrological and geological model of the eruption of Izu-Oshima volcano. Bulletin of the Volcanological Society of Japan 33, S297-S306(in Japanese with English abstract and figure captions). Arculus, R. J. & Bloomfield, A. L. (1992). Major element geochemistry of ashes from Sites 782, 784 and 786 in the Bonin forearc. In: Fryer, P. et al. (eds) Proceedings of the Ocean Drilling Program, Scientific Results 125, Beard, J. S. (1995). Experimental, geological, and geochemical constraints on the origin of low-k silicic magmas in oceanic arcs. Journal of Geophysical Research 100, Beard, J. S. & Lofgren, G. E. (1991). Dehydration melting and water-saturated melting of basaltic and andesitic greenstones and amphibolites at 1, 3, and 6.9 kb. Journal of Petrology 32, Bloomer, S. H., Stern, R. J. & Smoot, N. C. (1989). Physical volcanology of the submarine Mariana and Volcano Arcs. Bulletin of Volcanology 51, Gill, J. B., Hiscott, R. N. & Vidal, Ph. (1994). Turbidite geochemistry and evolution of the Izu-Bonin arc and continents. Lithos 33, Green, T. H. (1982). Anatexis of mafic crust and high pressure crystallization of andesite. In: Thorpe, R. S. (ed) Andesites. John Wiley & Sons. Hannah, R. S., Vogel, T. A., Patino, L. C., Alvarado, G. E., Perez, W. & Smith D. R. (2002). Origin of silicic volcanic rocks in Central Costa Rica: a study of a chemically variable ashflow sheet in the Tiribí Tuff. Bulletin of Volcanology 64, Hirose, K. (1997). Melting experiments on lherzolite KLB-1 under hydrous conditions and generation of high-magnesian andesitic melts. Geology 25, Hirose, K. & Kushiro, I. (1993). Partial melting of dry peridotites at high pressures: determination of compositions of melts segregated from peridotite using aggregates of

5 diamond. Earth and Planetary Science Letters 114, Hiscott, R. N. & Gill, J. B. (1992) Major and trace element geochemistry of Oligocene to Quaternary volcaniclastic sands and sandstones from the Izu-Bonin arc. In: Taylor, B., Fujioka, K. et al. (eds) Proceedings of the Ocean Drilling Program, Scientific Results 126, Kawate, S. & Arima, M. (1998). Petrogenesis of the Tanzawa plutonic complex, central Japan: Exposed felsic middle crust of the Izu-Bonin-Mariana arc. The Island Arc 7, Matthews, S. J., Sparks, R. S. J. & Gardeweg, M. C. (1999). The Piedras Grandes-Soncor eruptions, Lascar volcano, Chile; evolution of a zoned magma chamber in the central Andean upper crust. Journal of Petrology 40, Nishimura, A., Rodolfo, K. S., Koizumi, A., Gill, J. & Fujioka, K. (1992). Episodic deposition of Pliocene-Pleistocene pumice from the Izu-Bonin arc, leg 126. In: Taylor, B., Fujioka, K. et al. (eds) Proceedings of the Ocean Drilling Program, Scientific Results 126, Straub, S. M. (1995). Contrasting compositions of Mariana Trough fallout tephra and Mariana Island arc volcanics: a fractional crystallization link. Bulletin of Volcanology 57, Suyehiro, K., Takahashi, N., Ariie, Y., Yokoi, Y., Hino, R., Shinohara, M., Kanazawa, T., Hirata, N., Tokuyama, H. & Taira, A. (1996). Continental crust, crustal underplating, and low-q upper mantle beneath and oceanic island arc. Science 271, Tamura, Y. (1994). Genesis of island arc magmas by mantle-derived bimodal magmatism: evidence from the Shirahama Group, Japan. Journal of Petrology 35, Tamura, Y. (1995). Liquid lines of descent of island arc magmas and genesis of rhyolites: evidence from the Shirahama Group, Japan. Journal of Petrology 36, Tamura, Y. & Tatsumi, Y. (2002). Remelting of an andesitic crust as a possible origin for rhyolitic magma in oceanic arcs: an example from the Izu-Bonin arc. Journal of Petrology 43, Tamura, Y., Yuhara, M. & Ishii, T. (2000). Primary arc basalts from Daisen volcano, Japan: equilibrium crystal fractionation versus disequilibrium fractionation during supercooling. Journal of Petrology 41, Tamura, Y., Tatsumi, Y., Zhao, D., Kido, Y. & Shukuno, H. (2002). Hot fingers in the mantle wedge: new insights into magma genesis in subduction zone. Earth and Planetary Science Letters 197, Taylor, B. (1992). Rifting and the volcanic-tectonic evolution of the Izu-Bonin-Mariana arc. In: Taylor, B., Fujioka, K. et al. (eds) Proceedings of the Ocean Drilling Program, Scientific Results 126,

6 Fig. 1 Map showing the 11 Quaternary volcanoes, 8 Quaternary submarine caldera volcanoes with which this study is concerned and the location of the Mio-Pliocene Shirahama Group and the Miocene Tanzawa plutonic complex after Tamura & Tatsumi (2002). Numbered dots indicate sites drilled on the Philippine Sea Plate in the Izu-Bonin region during ODP Legs 125 and 126. The location map (lower left) shows the structure of the Izu-Bonin-Mariana arc system (Taylor, 1992). Double lines indicate spreading centres, active in the Mariana Trough and relic in the Shikoku and Parece Vela Basins. The Izu-Bonin, West Mariana and Mariana arcs are outlined by the 3-km bathymetric contour, and other basins and ridges are outlined by the 4-km contour. 6

7 Fig. 2 Volume-weighted histogram of rock type from 17 Quaternary volcanoes in the Izu- Bonin arc based on 1011 chemical analyses, showing a bimodal basalt-rhyolite profile. Fig. 3 Liquids comparable with calc-alkaline andesites (low FeO*/MgO andesites) are missing in the Izu-Bonin arc. Tholeiitic and calc-alkaline boundary (FeO*/MgO = SiO ) after Miyashiro (1974). Ashes from Sites 782, 784 and 786 if ODP Leg 125 (solid circles) plotted on FeO*/MgO vs. SiO 2 (Arculus & Bloomfield, 1992). Open circles are Quaternary Izu-Bonin volcanic rocks.

8 Fig. 4 Model for evolution of mantle-derived basalt and magnesian andesite in higher-level magma chambers (Tamura & Tatsumi, 2002). Previous calc-alkaline magma batches have partly solidified and are then remobilised and partially melted by later batches of basalt magmas in the same system. Numbered magma batches evolving from (a) to (c). (a) A diapir, which has wet and cool interior and dry and hot rind, produces wet and cool magnesian andesite and dry and hot basalt magmas, respectively (Tamura, 1994). (b) Tholeiitic series magmas are produced from dry basalt magmas, which are superheated by decompression and then cool and evolve through fractional crystallization (1 and 4). These dry magmas can erupt. Wet magnesian andesites magmas and their derivatives, however, become saturated and solidify within the crust (2, 3 and 5). New magma batches produced in the mantle ascend through the crust (6 to 10). (c) Hot basalt magmas (8 and 10) are emplaced beneath the frozen andesite magma bodies, which reheat and remobilize the andesite (3 and 5), triggering eruptions of calc-alkaline andesite, dacite and/or rhyolite. Some basalts (10) could show evidence of super-cooling (Tamura et al., 2000).