Identifying the Crystal Graveyards Remaining After Large Silicic Eruptions

Abstract:

The accumulation of voluminous crystal-poor rhyolites from an upper crustal mush environment inherently necessitates the complementary formation of unerupted silicic cumulates. However, identification of such frozen cumulates remains controversial. This has motivated us to develop of a new geochemical model aimed at better constraining the behavior of trace elements in a magma reservoir concurrently tracking crystallization and imperfect segregation of melt. We use a numerical method to solve our model equations rather than seek analytical solutions, thereby relieving overly simplistic assumptions for the dependencies between partition coefficient or melt segregation rate as functions of crystallinity. Our model allows partition coefficient to vary depending on the crystallinizing mineralogy at any particular stage in magma cooling, as well as the ability to test different rates and efficiencies of crystal-melt segregation. We apply our model first to the Searchlight Pluton as a well-constrained case study, which allows us to quantitatively test existing interpretations of that pluton. Building on this, we broaden our model to better understand the relationship between volcanic and plutonic rocks utilizing the NAVDAT database. Our results produce unambiguous fractionation signatures for segregated melts, while those signatures are muted for their cumulate counterparts. These models suggest that some large granitiods may represent accumulations of crystals, having lost melt in some cases to volcanic eruptions or to higher level evolved plutonic units, although the trace element signature of this process is expected to be subtle.