Geochemical Evidence of Water-Fluxed Crustal Melting in the Northern Colombian Andes

Abstract

Abstract The compositions of crustal magmas are powerful tools for understanding the formation and differentiation of continents. However, the geochemical fingerprints that distinguish the two dominant mechanisms of crustal melting, namely dehydration and water-fluxed melting, are still controversial. To provide new insights into this problem, we discuss the petrogenesis of the Paipa Volcanic Complex (PVC), an isolated Quaternary volcanic field in the Colombian Eastern Cordillera. The PVC is characterized by peraluminous trondhjemite-like rhyolites with exceptionally high Na2O contents (~6 wt %), super-chondritic Nb/Ta (~27), elevated Sr/Y ratios (~120), spoon-shaped REE patterns, and enriched isotopic compositions that overlap with those of the local basement. They also exhibit high pre-eruptive H2O contents (~up to 9.5 wt %) and abundant Paleozoic zircon inheritances. We demonstrate that these characteristics are inconsistent with a process of intra-crustal differentiation from a mafic or intermediate mantle-derived precursor. Instead, we propose that the origin of the PVC is best explained by melting the local (meta)sedimentary basement under H2O-saturated conditions, at middle-crustal pressures (~1.3 GPa) and relatively low temperatures (~690–740°C), following the complete breakdown of plagioclase and biotite, and the formation of reactive peritectic amphibole. This scenario differs from the high-temperature dehydration melting conditions that have been widely proposed for the Andes and globally, which result in the production of water undersaturated magmas in equilibrium with anhydrous lithologies rich in plagioclase and/or garnet. Accordingly, we speculate that an external H2O flux was ultimately sourced from a buoyant, cold, and hydrated mantle wedge that was extensively metasomatized by fluids derived from the Nazca and Caribbean flat-slab fronts. These conditions depressed the asthenospheric mantle potential temperature, likely inhibiting mantle melting. In turn, they facilitated the infiltration and ascent of mantle-derived H2O through pre-existing crustal faults and shear zones. Our results indicate that water-fluxed melting could be a plausible mechanism for generating crustal magmas in orogenic regions where the availability of free H2O has been difficult to confirm.