EVALUATING DISCREPANCIES BETWEEN BULK-ROCK AND SINGLE-MINERAL GEOCHEMICAL DATA IN THE MAGMATIC EVOLUTION OF CORDILLERAN ARC ROCKS

Bulk-rock geochemistry has long been used to understand the formation of plutonic rocks, but recent single-mineral studies of mafic silicates (e.g. amphibole and clinopyroxene) in the lower and middle crust of global arc systems have shown that equilibrium-melt models for these minerals often do not match bulk-rock chemistry. This raises significant questions about our current understanding of magma processes at convergent boundaries. We investigate shallowly emplaced plutonic rocks (3-7 kbar) with high Sr/Y values (>40) in bulk-rock data exhumed at paleo-convergent margins in New Zealand and S. California using laser-ablation mass-spectrometry of amphiboles from 25 samples (20 in New Zealand and 5 in S. California). Empirical correlations have been found between Sr/Y values and paleo-crustal thickness in these regions, and the Sr/Y proxy has been interpreted as evidence for garnet as a fractionating or residual phase deep in the arc. These locations are ideal to continue studying the mismatch between bulk-rock and single-mineral melt models.

We conducted over 1,200 microbeam analyses of igneous amphibole in the Separation Point Suite (SPS) in New Zealand and the La Posta Pluton (LPP) in S. California. In the SPS, 72% (650/906) of analyses are classified as magnesiohornblende and in the LPP 84% (252/301) are magnesiohornblende. Partition coefficient-based melt models show that 65% (591/906) in the SPS and 94% (284/301) in the LPP are in equilibrium with melts with low average Sr/Y (<40) and Dy/Dy* (x=0.5), average Dy_N/Yb_N = 3, and spoon-shaped heavy-rare-earth element (REE) patterns, unlike bulk-rock data from the same samples. Depletion of Dy/Dy* is an indicator of REE concavity, a feature related to amphibole control, and we do not see evidence of garnet control in Dy/Dy* vs Dy_N/Yb_N. Amphibole model crystallization temperatures range widely from 903-613 °C in the SPS and 895-630 °C in the LPP. Our data indicate that in these shallow-crustal arc regions, bulk-rock compositions are also likely not equivalent to parent melt compositions. This suggests that crystal accumulation, melt expulsion, and mixing processes are on-going at all levels in transcrustal magma systems and that the use of any bulk-rock data to approximate melt composition, pressure, or temperature in these regions should be carefully evaluated.