Age structure and inferences on the source of Neogene and Quaternary mafic volcanic rocks from the Southern Puna Plateau (NW Argentina).

RISSE, A., TRUMBULL, R.B., COIRA, B., KAY, S.M., VAN DEN BOGAARD, P.

Potsdam, Germany.

Simposio; 19 th Colloquium on Latin American Geociences.; 2005

IGCP 508 Project-IUGS-UNESCO

The appearance of young mafic volcanism up to 150 km behind the active arc and beneath the world’s second-largest continental plateau offers an opportunity to investigate scenarios for magma formation with respect to mantle dynamics in this unique geodynamic setting. Current hypotheses, based on geophysical, tectonic and geochemical evidence, and supported by numerical modelling, envision asthenospheric influx due to lithospheric delamination and/or changes in slab geometry. Volcanism in the Puna Plateau has been well-studied and despite important and locally detailed studies of the young mafic centers, more information on their age structure and magma sources is needed to address the underlying causes of magmatism. Our study comprised geochronology and geochemistry of basalts and basaltic andesites selected to cover the geographic and compositional range of the mafic volcanism. Laser 40Ar/39Ar dates from 20 centers were obtained and the results improve the geographic coverage of age information, provide better constraints on the youngest volcanism, and allow better assessment of compositional variations with time. The new ages, combined with earlier data, show that apart from sporadic older events, mafic magmatism in the study area started at ca. 8 Ma and lasted into the Pleistocene. The youngest dated centers (<1 Ma) are concentrated in the south and their compositions show enrichment in some incompatible trace elements (La, Ta, Nb, Zr) relative to the older centers as well as lower 87Sr/86Sr and higher 143Nd/144Nd ratios. The Sr-Nd isotope ratios correlate closely with SiO2, demonstrating clear and variable influence of crustal contamination. The compositional signatures of samples younger than 1 Ma suggest less contact with the crustal column during ascent, which may relate to an inferred change from a compressional to transtensional kinematic regime at about 2 Ma. The ubiquitous crustal contamination and lack of mantle xenoliths in these volcanic rocks severely limits the ability to characterize the mantle source. We can partly circumvent this problem by analysis of olivine-hosted melt inclusions and clinopyroxene phenocrysts from samples with the least evidence for crustal influence; and by focussing on trace element ratios most likely to reflect the source. For example, whole-rock HFSE/HREE ratios (Hf/Sm or Zr/Yb) vary only slightly over a spread of 87Sr/86Sr (0.7055 up to 0.7075) and are thus insensitive to crustal contamination. The Zr/Yb ratios in melts calculated from clinopyroxene LA-ICPMS data are used with batch melting models of a depleted mantle source to explore possible melting scenarios. Assumption of a depleted mantle source is supported by data from late Cretaceous basanites and mantle xenoliths from the region. The melting models match observed Zr/Yb values for 5-10% melting of a garnet-bearing source and this suggests a source region below ca. 80 km. Furthermore, based on the different mineral-melt partitioning for Zr/Yb in eclogitic (Ca-rich) versus peridotitic garnets, we can infer that involvement of eclogite was not important in the magma source. Water concentrations in melt inclusions analysed by confocal laser-Raman spectroscopy are 0.04 – 0.08 wt %, lower than average MORB. Thus, the source mantle was apparently dry and the cause of melting can be attributed to asthenospheric upwelling as predicted by the delamination and slab steepening scenarios.