Implications of chemical and isotopic variation in neogene Puna plateau ignimbrites for central andean crustal evolution.

KAY, S.M. AND COIRA, B.

Vancouver, Canada

Conferencia; Goldscmidt Conference; 2008

Geochemical Society and the European Association of Geochemistry

Voluminous dacitic ignimbrites with ages less than 11 Ma erupted in a subduction setting at a compressional margin in the central Andes.  A compilation of some 600 published and new analyses allows a view of the spatial-temporal crustal evolution of these igimbrite complexes that erupted through the thickened crust of the Puna platea between 22¢XS and 27¢X`S.   Their variable chemistry reflects the  conditions that generate large crustal magma systems.  AFC modeling largely based on 87Sr/86Sr and 18O/16O ratios under reasonable thermal conditions permit the large volume ignimbrites (>500 km3) to be approximately 50-50 mixtures of mantle-wedge derived basalt and variable crustal contaminants. Erupted mafic magmas show the mantle-wedge derived basalt is isotopically enriched (0.7055; ?ÕNd -2), most likely through crustal recycling by forearc subduction erosion and delamination. An estimated volume of 8200 km3 for the erupted ignimbrites and a 90 km3/km/Ma arc magma production rate produces a plutonic/volcanic ratio of 4:1 and 5.4 km of new crust distributed under the plateau.  Crustal contaminants  in the ignimbrites need to be spatially and temporally variable with 87Sr/86Sr ratios ranging from about 0.715 to 0.735 at 300 to 125 ppm Sr and 18O/16O ratios from 10 to 14. Variable Al/(K+Na+Ca), Na/K and isotopic ratios require a crustal contaminant that is more sedimentary-like and radiogenic in the north than in the south.  Temporal trends to lower 87Sr/86Sr ratios and more metaluminous compositions in the north indicate an evolving crustal contaminant as new magmas enter the crust. Trends from higher to lower La/Yb (40 to 10) and Sm/Yb (7.5 to 2) ratios in the same regions indicate a changing role for garnet as a controlling residual phase.  Trace element evidence for residual garnet and a requirement for high Sr concentrations in the crustal contaminants indicate that crustal contamination occured by melting in the deep crust and that the contaminated melts accumulated in magma chambers like those seen in seismic images near 20 km depth.  Negative Eu anomalies superimposed on high pressure-type REE patterns and calculated bulk distribtuion cofficients are best explained by feldspar fractionation in mid-crustal magma chambers.