Primitive magmas in convergent margins and at oceanic spreading ridges : evidence from early formed phenocryst phases and their melt inclusions

\\(The\\) \\(Heating\\) \\(Stage,\\) \\(and\\) \\(Applications\\) \\(to\\) \\(Studies\\) \\(of\\) \\(Basalt\\) \\(Crystallization\\) Melt inclusions in early formed minerals in mantle-derived magmas can give vital information on the crystallization history of the magma. By carefully applying the microthermometric (heating stage) technique of homogenizing the inclusions under visual control, and taking into account the possible effects of processes such as Fe\\(^{2+}\\) - Mg re-equilibration and H\\(_2\\) diffusion, both the crystallization temperature and the melt composition for that particular inclusion can be accurately estimated. The behaviour of melt inclusions upon heating in each magma suite studied must be tested via a series of kinetic experiments to estimate optimum heating rates; only after these have been done can the technique be applied with confidence. After the experiments, results are further tested and checked using established mineral-melt equilibria before they can be accepted as giving the true crystallization temperature and melt composition. At the Australian-Antarctic Discordance (AAD) on the Southeast Indian Ridge spreading centre, a site proposed to be underlain by abnormally cold mantle, early formed phenocryst phases are as primitive as those found at
ormal\" segments of the oceanic spreading systems. Results from heating stage experiments give crystallization temperature of 1240°C for olivine Fo\\(_{89.7}\\) which is the same temperature as obtained from olivine of the same composition from MORB at normal spreading segments. Similarly the calculated parental melt composition in equilibrium with the most magnesian olivine (Fo\\(_{90.4}\\)) is similar to the parental melt calculated for other MORB suites (Dtnitriev et al. 1985; Sobolev et al. 1989). This suggests that there is unlikely to be any significant temperature difference between the upper mantle beneath the AAD and normal sub-spreading ridge upper mantle. The difference in glass compositions (e.g. generally higher K\\(^2\\)O contents) between the AAD and adjacent areas is most pronounced in the most evolved samples and the geochernical trend defined by these glasses cannot be explained by crystal fractionation involving olivine plagioclase or clinopyroxene. This difference is therefore likely to be caused by mixing of a normal MORB magma with a magma with low TiO\\(_2\\) and FeO* and high SiO\\(_2\\) Na\\(_2\\)O and K\\(_2\\)O contents This mixing occurs early in the history of the magma suite as these variations recorded from quenched pillow glasses are also present in melt inclusions in early formed olivine phenocrysts. \\(Magma\\) \\(Genesis\\) \\(in\\) \\(the\\) \\(North\\) \\(Fiji\\) \\(Basin\\) \\(and\\) \\(Hunter\\) \\(Ridge\\) At the southernmost part of the North Fiji Basin in an area around the mainly submarine Hunter Ridge glass geochemistry of dredged rocks reflects the complexity of an area where a back-arc basin spreading ridge is transecting a youthful intra-oceanic island arc. Magma types identified from the glass geochemistry include boninites island arc tholeiites back-arc basin basalts incompatible element enriched-basalts and mixed magmas and rhyolites. Each magma group has olivine phenocrysts and spinel inclusions in olivine of distinct composition. Although most magma groups have similar CaO contents CaO in olivine varies significantly between each group being lowest in the boninites and highest in the back-arc basin basalts and the incompatible element enriched basalts with the island arc tholeiites intermediate; this is likely to reflect the significantly different silica activity in melts from the different magma suites. A new classification scheme is proposed for all available basalt glass analyses from the North Fiji Basin back-arc basin using K\\(_2\\)O/TiO\\(_2\\) K\\(_2\\)O/Al\\(_2\\)O\\(_3\\) and TiO\\(_2\\)/AlO\\(_3\\); this scheme effectively distinguishes between the N-MORB-like glasses and two groups of incompatible element-enriched back-arc basin basalt glasses. Two distinct enrichment trends are evident suggesting mixing of N-MORB with two types of enriched magmas with different K\\(_2\\)O/TiO\\(_2\\). One of these is likely to be related to ocean island basalts from the Rotuma-Samoan lineament whereas the other could represent influences from E-MORB-like magmas believed from dredged seamount evidence to be present in the South Fiji Basin. New whole-rock analyses of back-arc basin basalts from the North Fiji Basin show that the basalts are strongly depleted in the light rare earth elements and Nb but enriched in Rb Ba and K suggesting that they are derived from a source more refractory than the N-MORE source with the addition of a slab-derived fluid. Melt inclusion studies of a single North Fiji Basin basalt sample indicate a simple evolutionary history from a single parental melt by fractionation of olivine only. Andesites from dredge station 108 on the Hunter Ridge are distinct in composition from the other magma groups defined from this region during this study and show strong mineralogical evidence for magma mixing. Olivine phenocrysts show bimodal compositional distribution ranging from Fo93_82 and both olivine and clinopyroxene show complex zoning patterns; furthermore two main groups of Cr-spinel inclusions occur in olivines. The most magnesian olivine crystallized at approximately 1350°C from a melt with approximately 17% MgO. Significant variations are evident in compositions of melt inclusions indicating mixing between two or more magma types. Melt inclusion studies together with whole-rock major and trace element analyses suggest that these mixed andesites evolved through open system fractionation during which frequent magma mixing of cogenetic refractory island arc tholeiitic magmas and fractionation of olivine and clinopyroxene buffered the major element composition but caused an increase in the incompatible minor elements and separated for example Y from Zr. Addition of back-arc basin basalt magmas to this system when the southward propagating spreading centre in the North Fiji Basin transected the Hunter Ridge provided this additional component to the mixing as is evident from compositions of Cr-spinel inclusions in olivines and trace element contents of the andesites. \\(Primitive\\) \\(Basaltic\\) \\(Magmas\\) \\(in\\) \\(Iceland\\) Whole-rock compositions of the most magnesian tholeiites from Iceland are too magnesian to be in equilibrium with their most magnesian olivines suggesting that they must have accumulated some olivine. Tholeiites with olivines > Fo\\(_{90}\\) are all depleted in the light rare earth elements whereas the only light rare earth element enriched sample has no olivines more magnesian than Fo\\(_{84.6}\\). Cr-spinel inclusions are also different in the enriched tholeiite and no evidence for mixing between the enriched and depleted tholeiites was found. A single sample from the Krafla system in northern Iceland contains two populations of Cr-spinel inclusions in olivines due to either mixing between the depleted Icelandic tholeiites and more typical N-MORB or because of assimilation of clinopyroxene and plagioclase from entrained gabbroic xenoliths. A review of published melt inclusion studies on phenocrysts in Icelandic tholeiites (Gurenko et al. 1988b; Hansteen 1991) reveals that the calculated primary melts of Gurenko et al. (1988b) and a parental melt calculated herein from a composition from Hansteen (1991) are both too rich in diopside to be in equilibrium with a mantle residue. This is likely to be caused by assimilation of clinopyroxene since both these studies were performed on samples containing abundant gabbroic xenoliths and resorbed clinopyroxene xenocrysts."