Coupled δ18O-δ17O and 87Sr/86Sr isotope compositions suggest a radiogenic and 18O-enriched magma source for Neapolitan volcanoes (Southern Italy)

Abstract

The origin of large variations in stable and radiogenic isotope compositions of magmas erupted from the Neapolitan volcanoes, including Somma-Vesuvius and Campi Flegrei (Southern Italy), has always been contentious. Indeed, the role and relative importance of sediment subduction versus crustal assimilation to explain the chemical and isotopic variations of the erupted magmas remain unclear. Isotopic disequilibrium between minerals and their host indicate that bulk rock analyses are incapable of constraining the isotopic composition of the source. Therefore, we use isotopic (87Sr/86Sr, 18O/16O and 17O/16O) data on separated minerals (feldspar, clinopyroxene and olivine phenocrysts) from pyroclastic successions and lava flows of the Neapolitan volcanic area (Phlegrean Volcanic District and Somma-Vesuvius complex) to better constrain magmatic oxygen and strontium isotope ratios. Magmatic values recalculated from δ18O of olivine and clinopyroxene phenocrysts range from typical mantle values of 5.2‰ to almost 9‰ relative to SMOW. These compositions are very different from those of typical mantle sources. In order to assess the degree of magma evolution from which these minerals formed, the Mg# of clinopyroxene and olivine were converted into host melt Mg#, resulting in the range 44 to 76. Simple assimilation of silicic crustal rocks is difficult to reconcile with the mafic nature of these estimated host magma compositions. This indicates that some mafic, mantle-derived magmas, having unusually heavy oxygen isotope (up to ~9‰) and high Sr isotope compositions (0.7050 to 0.7085), must exist. Crustal assimilation of carbonates can be excluded by the lack of a link between isotopes and major and trace element signatures. Assimilation of either Hercynian-like crust or altered pyroclastic rocks, however, cannot be ruled out completely. However, assimilation of partial melts from a Hercynian-like crust would have to be ~12% and ~21% to explain the heavy oxygen isotope values of Campi Flegrei and Somma-Vesuvius, respectively. Such degrees of assimilation are unlikely since the magmas are either too mafic or too alkaline to be consistent with such high proportion of crustal components. Other, less mafic Campi Flegrei, Somma-Vesuvius and Ischia magmas with more typical mantle oxygen isotopes, have been possibly generated in a mantle source affected by minor contamination by pelagic and carbonate sediments and subsequent assimilation of Hercynian-like crust.Sr-O mixing models indicate that such magmas were derived from a mantle source that was contaminated by up to 10% of a 1:1 mixture of pelagic and carbonate sediments. These findings together demonstrate the highly complex and varied magma sources in the Campania volcanic district. Triple oxygen isotope variations (Δ17O), based on measurements of both 18O/16O and 17O/16O ratios in potential assimilants are in agreement with these conclusions.