Three years (2011-2013) of eruptive activity at Mt. Etna: working modes and timescales of the modern volcano plumbing system from micro-analytical studies of crystals

Tracking magma dynamics at active volcanoes by using the compositional and textural records of minerals is becoming an important aspect of modern volcanology to solve key issues, such as temporal relationships between magma recharge and eruption, duration of magma storage and final ascent upward to the surface. These issues are relevant at Mt. Etna, where the style of volcanic activity during the last years showed drastic variations in duration and intensity. The paroxysmal activity between 2011 and 2013 at the volcano summit has given the opportunity to investigate the chemical-physical processes driving the activity and to fix their spatial-temporal relationships into the plumbing system. An extensive compositional dataset of plagioclase and olivine crystals from lavas emitted between the 2011 and 2013 at Mt. Etna has been used to constrain modes and timescales of magma storage and transfer to the surface. Plagioclase crystals display either near-equilibrium or disequilibrium textures at the core and rim that indicate complex histories of magma crystallization under variable chemical and physical conditions. Crystals with different textures have been characterized for major (An), minor (Fe and Mg) and trace elements (Sr/Ba). Textural relationships at the plagioclase cores support the idea that crystals undergo variable decompression rates in the deep parts of the plumbing system. Fe and Mg zoning vs. anorthite in correspondence of plagioclase sieve textures at the rim suggest that processes of gas-flushing have a dominant role in triggering paroxysmal eruptions, determining a sudden intensification of the eruptive activity up to the fountaining phase. The Sr/Ba ratio in oscillatory-zoned plagioclase revealed the dominant presence into the plumbing system of high-Sr magma volumes coexisting with low-Sr magma batches that preserve a geochemical signature similar to that of magmas feeding the pre-1971 activity. Through Sr-diffusion modeling in plagioclase the maximum time of magma storage during the considered eruptive period has been evaluated. Timescales of crystal residence in the plumbing system are short (five years to three decades), suggesting limited magma storage and faster transfer dynamics to the surface. Chemical zoning of olivine crystals highlights processes of multi-step magma transfer and residence at different levels of the plumbing system. The upward migration of magmas occurred primarily through multiple episodes of injection and mixing between five compositionally distinct magmatic environments (Mi), whose P-TƒO2 characteristics and concentrations in dissolved volatiles were constrained by thermodynamic modeling on the basis of the forsterite contents found at the olivine cores. From a deepest reservoir, located at depth of ~600 MPa, the most primitive magma M00 (Fo84) moved along dominant pathways, intercepting the M0 (Fo80–82) at ~390 MPa and/or M1a (Fo78; 250 MPa), M1b (Fo75; ~140 MPa) and finally the shallow M2 (Fo70–73; ~40 MPa) storage zones. For some eruptive episodes, olivine zoning highlights a preferential route of transfer, connecting the M00 and M2 storage zones that facilitated the fast migration of primitive magma at shallow depth. Fe-Mg diffusion modeling on olivine normal and reverse zoning defines the timescales of magma transfer and storage across these magmatic environments, which vary from ~1 to 18 months, whereas intrusion and mixing by more basic magma into the shallowest reservoir occurred always within 5 months before the eruption. Relevance of this study mainly relies on the quantification of volcanic processes at depth that may have considerable consequences in development of unusual, high-energy eruptions at basaltic volcanoes, generally acknowledged for their weak to mild explosive activity.