Petrological constrains on the coarse-grained, high-Mg basaltic enclaves of Capo Marargiu (Sardinia, Italy)

We present results from a textural and geochemical study conducted on calc-alkaline volcanic and hypoabyssal rocks from the Oligo-Miocene Capo Marargiu Volcanic District (CMVD; Sardinia, Italy). \nStratigraphic units of CMVD consist of lava domes, a pyroclastic breccia interbedded with lava flows, and dikes. The pyroclastic breccia is in lateral contact with a low crystallinity (∼15 vol.% phenocrysts), massive hypoabyssal body hosting decimetre-sized, coarse-grained enclaves with porphyritic textures (∼50 vol.% phenocrysts). \nThese crystal-rich enclaves and the host rock exhibit phase assemblages of clinopyroxene + plagioclase + amphibole + olivine + magnetite + low-Ca pyroxene, and plagioclase + clinopyroxene + magnetite + low-Ca pyroxene, respectively. Clinopyroxene phenocrysts (≤5 mm in size) in the crystal-rich enclaves show two compositionally distinct populations: type 1 clinopyroxenes are diopsides (Mg#85-90), whereas type 2 clinopyroxenes are augites (Mg#74-82). Plagioclase phenocrysts (≤1 mm in size) from crystal-rich enclaves and the host rock are normally zoned with An75-96 cores grading to An50-75 rims. The composition of amphibole phenocrysts (≤20 mm in size) is Mg-hastingsite. The mineral texture is poikilitic suggesting early crystallization of amphibole with respect to plagioclase. In fact, the primitive (∼Mg#76), high-T amphiboles include clinopyroxene, whereas the more evolved (∼Mg#65), low-T phenocrysts host plagioclase. Amphiboles are also surrounded by characteristic reaction coronas, consisting of tiny microlites (<5 μm in size) of clinopyroxene, low-Ca pyroxene, plagioclase, magnetite and ilmenite. Olivine occurs as phenocrysts (≤1 mm in size) with Fo79-87 cores surrounded by Fo66-79 rims. \nThe bulk-rocks of crystal-rich enclaves are high-Mg basalts (i.e., 10 wt.% MgO), whereas the host rock is a more differentiated basaltic andesite (i.e., 5 wt.% MgO). Major oxides and compatible trace element modelling suggest that the basaltic andesitic magma originates by crystal fractionation of olivine + clinopyroxene from a high-Mg basalt [1]. In turn, compatible trace elements in the high-Mg basalt are low (330 ppm Cr, 130 ppm Ni) relative to picritic arc magmas, as the result of crystal fractionation of olivine ± Cr-spinel from a primary magma at depth [1; 2]. \nThermobarometric calculations on phenocrysts from the crystal-rich enclaves in equilibrium with the high-Mg basalt yield pressures (600-400 MPa) and temperatures (1200-950 °C) consistent with phase diagrams derived by experiments conducted on primitive arc liquids [3]. Nevertheless, (i) the breakdown of the opacitic amphibole rim [4], (ii) the late appearance of plagioclase, and (iii) the correspondence between ∼Mg#65 natural amphiboles and mineral compositions experimentally-derived at 200 MPa [5], indicate that the crystal-rich enclaves experienced a decompression path started at higher crustal depths. In this view, the high-Mg basalt and the basaltic andesite represent two different regions of a chemically zoned magma chamber formed by crystal fractionation of a primary magma ponding at ∼500 MPa. Subsequently, buoyancy forces associated with density gradients caused upward migration of the basaltic andesite carrying portions of the adjacent high-Mg basalt to shallower crustal levels. \n \nReferences: \n[1] Yamamoto M (1988) Contrib Mineral Petrol 99:352-259 \n[2] Eggins SM (1993) Contrib Mineral Petrol 114:79-100 \n[3] Foden JD and Green DH (1992) Contrib Mineral Petrol 109:479-493 \n[4] Reagan MK et al. (1983) Bull Volcanol 49:415-434 \n[5] Sisson TW and Grove TL (1993) Contrib Mineral Petrol 113:143-166