MINERAL-MELT EQUILIBRIUM MODELS VS BULK-ROCK GEOCHEMISTRY: WHICH TELLS THE TRUE STORY OF CORDILLERAN ARC FORMATION?

Modeled melts in equilibrium with individual minerals in plutonic rocks in cordilleran arcs reveal significant chemical discrepancies from encompassing bulk-rocks. These discrepancies raise an important question: Does bulk -rock data represent original magma composition or artifacts of crystal accumulation and melt loss? We examine this problem in the context of Sr/Y, a common proxy for crustal thickness and garnet involvement, in two Cretaceous convergent margin arcs, the Peninsular Ranges Batholith (PRB) in S. California and the Median Batholith in Zealandia. Both arcs preserve belts of low-Sr/Y (<40) and high-Sr/Y (>40) plutons. While the low-Sr/Y belts have been interpreted as products of normal arc magmatism, prior studies suggest that high-Sr/Y rocks represent remelting of crust in the garnet stability field, potentially generating a dense, mafic arc root (i.e., arclogite). Bulk-rock studies of the high-Sr/Y rocks in the PRB and Zealandia support this interpretation, noting depletions in HREEs and a lack of Eu-anomalies.

We conducted 1,207 geochemical analyses across 88 magmatic amphiboles in 25 samples (18 in Zealandia, 7 in the PRB), modeled the melts in equilibrium with these analyses, and compared them to bulk rock XRF data from the same samples. Amphiboles are mostly magnesiohornblende and all regions have average Si (apfu) identical within error (Zealandia: 6.84 ± 0.3; PRB: 7.08 ± 0.3). Molar Mg# (100*Mg/(Mg+Fe)) ranges from 40.1-74.7 in Zealandia and 40.7-52.4 in the PRB. Pressures from Al-in-hornblende barometry range from 0.2-6.1 kbar in Zealandia and 0.5-6.2 kbar in the PRB. Analyses within the calibration range of published melt models produce melts much more silicic (~65-85+ wt% SiO₂) than their bulk-rock counterparts (~53-70 wt% SiO₂). Modeled melts in equilibrium with our analyses display prominent negative Eu-anomalies, flat to spoon-shaped HREE patterns, average Dy/Yb < 3, and 68.7% (830/1,207) have Sr/Y < 40, consistent with the crystallization of plagioclase and amphibole. We suggest that the Sr/Y signature in bulk rocks is unrelated to melt compositions, but instead is a product of mineral accumulation. Our results indicate that high Sr/Y bulk-rocks are not representative of melts and elevated Sr/Y does not necessarily indicate the presence of a garnet-bearing root.