Ming Tang, Jiazhen Wang, Cin-Ty A. Lee

Garnet has been proposed to be an important fractionating phase during magmatic differentiation in thick volcanic arcs because garnet fractionation can reproduce the hallmarks of continental crust by driving residual magmas to higher silica, lower total iron, and higher iron oxidation states. Recently authors have measured the partition coefficients for Fe2+ and Fe3+ between garnet and melt and concluded that garnet crystallization has little influence on arc magma oxidation. However, these conclusions are based on conceptual errors in modeling the effects of garnet. First, their models intentionally limited the extent of fractionation to minimize the effect of garnet fractionation, inconsistent with existing experiments, including their own experiments. Second, Fe2+ partitioning was indirectly and incorrectly parameterized using Fe-Mg exchange coefficients, inconsistent with experimental partition coefficients. Application of these recent models leads to unrealistic Fe2+ partitioning behavior with progressive fractionation. When these models are corrected with proper mass balance and partitioning relationships, garnet as a driver of oxidation remains a viable and testable hypothesis.