Plagioclase as archive of magma ascent dynamics on “open conduit” volcanoes: The 2001–2006 eruptive period at Mt. Etna

Plagioclase is the most common phenocryst in all Etnean magmatic suites (~ 50% in volume), as well as in most lavas erupted worldwide. Its stability field is strongly dependent on the physico-chemical conditions of the melt and, consequently, it can be used as a tool to record the processes occurring within the feeding system. With this aim, a detailed textural and compositional study of plagioclase was performed on the products emitted during the 2001, 2002–2003, 2004–2005 and 2006 eruptions. Four distinct textures were recognized at the crystal cores: (1) clear and rounded (An 73–85 ), (2) dusty and rounded (An 73–85 ), (3) sieved (An 82–88 ) and (4) patchy (An 60–81 ), while two distinct textures are commonly observed at the crystal rim: (1) dusty (An 73–90 ) and (2) with melt inclusion alignments (An 70–76 ). Moreover all plagioclases present a thin (10–20 μm) outermost less calcic (An 53–76 ) rim. For each crystal a complex evolutionary path was reconstructed, and several growth and resorption episodes were identified. The f O 2 was estimated using Plag–Cpx/liquid equilibrium in order to calculate the Fe +3 /Fe 2+ ratio in the melt and, in turn, to reconstruct the primitive magma composition by adding a wehrlitic assemblage to the least evolved lava of the four eruptive episodes. MELTS modeling was then developed using this primary magma composition, as well as a trachybasaltic lava. Calculations were performed at variable pressures (400–50 MPa, step of 0.50 MPa) and H 2 O contents (3.5–0 wt.%, step 0.5 wt.%) in order to estimate the crystallization temperature of olivine, clinopyroxene, plagioclase and spinel, decreasing T from the liquidus down to 1000 °C at steps of 20 °C. P–T and water contents were also determined using geothermobarometers and plagioclase–melt hygrometers respectively, aiming at verifying the parameters used in the MELTS modeling. At this point plagioclase textural features and compositions were related to specific P–T– f O 2 –H 2 O conditions. Plagioclase stability models indicate that: (1) H 2 O strongly influences the plagioclase–melt equilibrium allowing the crystallizations of more calcic compositions only at shallow levels; (2) patchy cores form at high pressure (up to 350 MPa) and low water content (< 1.7 wt.%); (3) clear dissolved cores form at lower pressure (150 MPa) and higher water content (1.5–2.8 wt.%); (4) dusty rims form at even lower pressure straddling the H 2 O-saturation curve and, (5) melt alignments form during degassing. According to experimental works each of these textures can be related to a different process within the feeding system, such as multiple magma inputs (patchy core), volatile addition or increase in T (clear core), mixing (dusty rims) and rapid decompression and degassing (melt inclusion alignment at rims). These inferences were successfully compared with the eruptive evolution of each event as deduced from direct observations, and geophysical and petrological data. The overall picture shows that plagioclase crystallizes under polybaric conditions in a vertically extended and continuous feeding system in which at least two magma crystallization levels were identified. Plagioclase stability also indicates that a large variability in water content characterizes the magma within the feeding system.