GAINING INSIGHT INTO MAGMA CRYSTALLIZATION PROCESSES BY IMAGING SPATIAL DISTRIBUTIONS OF MINERALS AND MINERAL COMPOSITIONS USING COMPOSITION-CALIBRATED, GRAYSCALE-THRESHOLDED BACKSCATTERED ELECTRON IMAGES OF THIN SECTIONS

Rock textures and mineral compositions are critical data used to interpret magma crystallization. The regression model method described in a separate presentation by Willis et al. produces thresholded images of thin sections that can be used to: show sequences of mineral growth or intergrowth; estimate volume proportions of specific phases or mineral compositions; quantify crystal size or other features for specific composition ranges. Thermodynamic phase equilibria (e.g. MELTS algorithm, Ghiorso & Sack, CMP 1995; Asimow & Ghiorso, AmMin 1998) model phases and compositions that crystallize from a starting liquid with decreasing Temperature. We use regression model results to assign grayscale values to compositions predicted from MELTS and produce thresholded images of thin sections that correspond to crystallization steps; we see the spatial arrangements of mineral compositions that crystallized together.

We apply these methods to the Morgantown-Jacksonwald System (MJS), a smaller sibling of the Palisades Sill and Watchung Basalts in the Newark Basin (PA-NJ-NY). The 201 Ma gabbro-diabase intrusions and flood basalts are part of the Central Atlantic Magmatic Province emplaced during rifting of Pangaea. We combine thresholded BSE images with field relations, whole-rock and mineral compositions, and MELTS models. Orthopyroxene phenocrysts were brought into a MJS sill by the host magma; BSE images thresholded by composition show that, while Mg-rich cores are small (< 1 vol. %), the entire Opx crystals make up about 25 vol. %, consistent with incompatible trace element models of whole-rock concentrations. Thresholded BSE images of samples from different levels in the MJS show Plagioclase with different zoning patterns, as well as different proportions and distributions of late-stage phases. MELTs models can connect compositions to Temperature and liquid proportions and properties. Taken together, results suggest different crystallization conditions at different levels: more crystal sorting, compositional convection, and melt migration at 5-6 km depth; stronger intracrystalline zoning due to more rapid crystallization and small-scale fractionation from a greater proportion of trapped interstitial liquid at higher levels (1-2 km depth).