ORIGIN OF LEUCOSOME LENSES AND ASSOCIATED SELVAGES IN MIGMATITES: INTERPRETATIONS BASED ON PSEUDOSECTION MODELING OF THE NASON RIDGE MIGMATITIC GNEISS, WA

Thermodynamic modeling using pseudosections for rocks from the Nason Ridge Migmatitic Gneiss in the Cascades Core (WA) are used to investigate the origin of quartz-plagioclase lenses that locally occur with biotite-rich selvages. The migmatite has leucosomes with multiple origin, most of the leucosomes are interpreted to have formed by melt injection during and after the last deformation event. P-T paths, based on thermobarometry and P-T pseudosections, indicate peak metamorphism at 600-700 ºC, 6-9 kbar, with a gradient of increasing pressures toward the NE. Pseudosection topology indicates that kyanite and sillimanite are stable in most of the pressure and temperature range of interest at subsolidus conditions and that they are stable at suprasolidus conditions for the most aluminous-rich samples. Garnet is predicted to be stable in the entire P-T range considered. Observation of garnet in leucosomes and melanosomes and kyanite in leucosomes is consistent with modeling predictions of peritectic garnet growth and limited kyanite-sillimanite consumption during melting. Partial melting is predicted between 655 ºC (10 kbar) and 675 ºC (6 kbar). Textures, predicted metamorphic conditions, and MnNaCaKFeMgAlSiH pseudosections are compatible with partial melting and local segregation of melt into quartz-plagioclase leucosome lenses parallel to or at a low angle to schistosity. The most important melt compositional variations in predictions are tied to the modeled water content because water acts as flux for quartz and plagioclase melting, which produces a rapid increase of SiO₂ in the melt. Only slight discrepancies between melt compositional predictions and observed leucosome compositions are noticed for SiO₂, Al₂O₃, and Na₂O. However, significant discrepancies are evident for K₂O and CaO; these may be product of leucosome modification processes such as fractional crystallization with melt escape and/or back reactions between melt and country rock. An alternative explanation is that the melt activity model is inadequate for predicting melt compositions.