Abstract
The Guntur-Gandapura Volcanic Complex is part of a cluster of active volcanoes forming a chain in the southern part of West Java, regionally belonging to the western part of the Sunda arc. The Complex, 60 km southeast of Bandung, rises from 700 meters above sea level with Masigit center (2,249 m) being the highest peak and Guntur cone (1,950 m) the most active and the youngest cone of the Complex. The Complex includes the products of fourteen main eruptive centers which have been identified.
Geologic evidence from the GGVC suggests that it is very young in age. Eight out of twenty samples analysed for K-Ar dating have usable ages ranging from 0.33±0.02 Ma (Gandapura center) to 0.07±0.03 Ma (Pasir Laku center). Indications related to the K-Ar dating point to the GGVC being younger than 1 Ma, and the initiation of volcanism at the Complex might have occurred no earlier than about 400,000 years ago.
Two suites, tholeiitic and calc-alkaline, are represented among the rocks of the Complex. In volume, calc-alkaline andesites are dominant, followed by tholeiitic basalts. Dacites are relatively rare. High-Mg basaltic andesites have the highest magnesium numbers (73-68) among the rocks in the Complex and thus are candidates for primary mantle-derived magmas. There is no definite relationship between the order of eruption and the tholeiitic or calc-alkaline affinities. Eruptions began with either high-Al low-K tholeiites or andesites at the centers which have both tholeiitic and calcalkaline rock types.
Texturally, tholeiitic lavas are porphyritic. Plagioclase (An₈₉-An₆₄) phenocrysts are abundant. Olivine and clinopyroxene are less abundant but very significant. Most of the tholeiitic rocks have a typically non-vesicular intergranular groundmass, and some of the rocks have a typically vesicular intersertal groundmass. Tholeiitic scoria of the Complex has a typical vesicular-hypocrystalline groundmass.
Texturally, calc-alkaline lavas are porphyritic with abundant plagioclase (An₈₇-An₄₄) and less abundant clinopyroxene, orthopyroxene, olivine, magnetite, and amphibole. Volume of phenocrysts is usually lower than the volume of the groundmass. High-Mg basaltic andesites are less porphyritic with less abundant plagioclase (An₈₀-An₄₅) compared with the 'normal' basaltic andesite (An₈₇-An₆₀).
The clinopyroxene from all analysed lavas includes a few clinopyroxenes which have a composition within the diopside field and relatively high Al, especially in the tholeiite lavas. Rims of pigeonite and orthopyroxene surrounding the olivine phenocrysts in tholeiite and basaltic andesite, respectively, suggest that there is a reaction relationship between olivine and liquid producing these Ca-poor pyroxenes. A reconnaissance study of the pyroxene geothermometry shows that if equilibrium were achieved, certain augites in both the high-Mg basaltic andesite and the normal basaltic andesite could have crystallised at temperatures as high as 1150° C.
Olivine phenocrysts of high-Mg basaltic andesites are normally zoned from Fo₈₈-₈₉ in the core to Fo₈₄-₈₇ in the rim. It suggests that the phenocryst cores formed by near-liquidus crystallisation with very little or no subsequent fractionation of crystals from the calculated equilibrium liquid, while the olivine phenocryst rims formed by removal of Mg and / or with little subsequent fractionation of crystals from such a liquid. Thus, the olivine of high-Mg basaltic andesites, is of a composition very similar to mantle olivine, and the addition of only 10% of Fo₈₈ or Fo₈₉ will give a total rock composition compatible with direct mantle derivation. On the other hand, the Mg/Mg+Fe²⁺ ratios of the observed tholeiite and normal basaltic andesite might result from cumulus origin from melts of more evolved composition, or partial re-equilibrium of the grains with late-stage liquids having lower Mg/Mg+Fe²⁺ ratios than the bulk rocks. So, it seems unlikely that either of the two olivine-bearing rocks represent unfractionated liquid compositions.
The calculated olivine (Fo₈₈) equilibrium temperature for the high-Mg basaltic andesite is about 1208° C.
The compositional ulvöspinel components in the titanomagnetite of the tholeiitic lavas are restricted (Usp₃₂-Usp₄₁). Titanomagnetites of basaltic andesites have a wider range of ulvöspinel components from Usp₂₅ to Usp₄₆. Titanomagnetites of andesites and high-Mg basaltic andesites are rather richer in ulvöspinel, 33-56% and 37-46%, respectively, than those of the tholeiite and of most normal basaltic andesites.
Magnesian hastingistic hornblende is found in some lavas from the Complex, especially the normal basaltic andesites.
The high-Mg basaltic andesites contain both 'microphenocrysts' and inclusions of Cr-rich spinel in olivine. A general absence of the Cr-spinel in the pyroxene or other phases suggests that its crystallisation may have been terminated, or at least restricted, by the time of appearance of the pyroxene phases in the system.
Glasses in some tholeiites and basaltic andesites contain micrometer diameter droplets of immiscible high RI, presumably high-Fe glass. The clear high-silica glasses closely resemble the end products of classical magmatic differentiation by crystal fractionation, but intermediate- and low-silica glasses are variable and similar to early silicate melt inclusions in olivines of lunar samples. Petrographic observations suggest that the low-silica glasses are of late origin. These glasses are much lower in iron and are more moderate in silica content than immiscible high-iron glasses from lunar rocks. They have not reached the immiscibility volume where split into silica-rich and iron-rich liquids would normally occur.
The Mg-numbers of the tholeiites are mostly too low to be compatible with direct derivation from water-saturated peridotite mantle source, and the trend between silica and Mg-number argues against derivation of the magmas by melting of oceanic basalt sources. The tholeiitic rocks show relatively unfractionated patterns of chondrite-normalised REE with flat patterns of LREE and slightly positive slopes for their HREE. The observed tholeiitic REE patterns and the calculated primary magmas are matched satisfactorily by melting models based on spinel-bearing peridotite sources at approximately twice or three times chondrite levels, provided that degress of partial melting are 10-18%. Fractionation of plagioclase and lesser amounts of olivine and titanomagnetite, with fractionation or addition of clinopyroxene, is capable of explaining the major element variation within the tholeiites of the Complex and can also account for a transition from these to a higher silica basaltic andesite.
The geochemistry of the calc-alkaline rocks shows a wide range of composition from 52 to 65% SiO₂, and is more variable in alumina than the tholeiites. The high-Mg basaltic andesites are distinctive in having lower Al₂O₃ than any other rock types in the Complex. The basaltic andesite and andesite are slightly and substantially more enriched, respectively, in their chondrite-normalised LREE. The chondrite-normalised REE patterns of dacites are within the upper range of the andesite patterns. The observed patterns for the Complex calc-alkaline high-Mg basaltic andesite are consistent with 10% melting of amphibole-bearing lherzolite at approximately 4 to 5 times chondrite levels. The andesites and dacites have steeper REE patterns than the basaltic andesites, and are parallel to model curves for 8% melting of amphibole-bearing lherzolite. Source for normal basaltic andesite, andesite, and dacite requires 3 to 4 times chondrite abundances. Fractionation is a possible mechanism in the calc-alkaline magma series to derive the more evolved rocks from the high-Mg basaltic andesite composition. However, the low-K tholeiites cannot be parental to basaltic andesite of the calc-alkaline series.
Spidergram analyses for the tholeiitic rocks show that approximately 80% of Ba, Th, and Rb, 60% of Ta and K, 59% and 55% of Sr and Ce, respectively, come from a subduction and/or other additional origin. There is no significant contribution from the subduction zone for Nb, Hf, Zr, Sm, Y, and Yb, and Ti tends to be depleted.
Spidergram analyses of the andesites-dacites indicate that large contributions of Th, Rb, Ba, K, Ce, and Ta are of subduction origin. Zr, Hf, and Sm show a small contribution of subduction components. There is no enrichment of P, Y, and Yb, and Ti shows much more depleted patterns than the other incompatible elements.
Pre-Guntur tholeiites are relatively low in ⁸⁷Sr/⁸⁶Sr (0.70397 and 0.70392), while the tholeiites of the Guntur cone are slightly higher (0.70417 and 0.70422). These values are higher than typical Indian MORB at an average of 0.70273. The calc-alkaline lavas are higher and have greater variability in ⁸⁷Sr/⁸⁶Sr ratios than the tholeiites. High-Mg basaltic andesites have ⁸⁷Sr/⁸⁶Sr ratios of 0.70436 and 0.70445. Three andesites have ⁸⁷Sr/⁸⁶Sr ratios 0.70433, 0.70456, and 0.70438, and a dacite has 0.70452. The tholeiites of the GGVC might have a similar mantle source to those of the eastern Indian oceanic-floor tholeiites modified by small involvement of the components from the subducted slab. The subducted slab components probably account for the high ⁸⁷Sr/⁸⁶Sr ratios of the more evolved rocks of the Complex.
A combination of petrogenetic processes is suggested involving dehydration of the subducted oceanic lithosphere, formation and drag-mechanism of hydrous upper mantle peridotite, enrichment in LIL elements derived from the subducted slab, fractionation, and some magma and crystal re-mixing.