Repacking-Driven Compaction in the Spirit Mountain Batholith, Southern Nevada

Abstract Mechanical phase separation in crustal reservoirs is one of several processes that influences chemical differentiation in magmatic bodies and the generation of potentially eruptible horizons of magmas. The rate of melt-crystal separation in magmas remains hard to constrain, notably because of the poor understanding of the processes that govern phase separation, such as crystal settling and mush compaction. To that end, it is necessary to identify natural laboratories that record mechanical phase separation. Here, we investigate the Spirit Mountain Batholith (SMB) for chemical and textural evidences of mechanical phase separation. We identify a near-linear unmixing (of crystals and pore-occupying melt) trend in major, minor, and trace element geochemistry defined by samples within a ca. 3-km transect at the base of the exposed batholith and pooled leucogranites near the top of the batholith. Plagioclase compositions are similar in the samples within the transect, despite a broad range in bulk SiO2 (62–74 wt % SiO2), suggesting the samples crystallized from the same parental magma and subsequently either lost or accumulated SiO2-rich interstitial melt. Trace element modeling and mass balance calculations fortify this finding. The minimum volumetric trapped melt fraction calculated based on trace element modeling is 0.3, consistent with maximum packing fractions inferred by applications of compaction models to recent mechanical phase separation experiments. We find that the trapped melt volume fraction decreases with depth in the transect and that melt loss is accompanied by the development of fabric perpendicular to the axis of paleo-gravity, consistent with repacking-driven compaction. The batholith is comprised of numerous repeated injections of sills over the course of approximately two million years, as indicated by previous researchers using zircon secondary-ion mass spectrometry age dates (Walker Jr et al., 2007). We suggest that these sills underwent some degree of crystallization prior to melt-crystal separation by crystal repacking. Forward modeling of repacking-accommodated melt extraction within a single sill suggests that a lens of pure melt 3.5 km thick (approximate thickness of leucocratic layer of granite in the SMB) can be extracted in approximately 30 ka. This estimate is likely an upper bound as the extracted melt was likely sourced from several sills, the exact number of which is unclear.