2004 Denver Annual Meeting (November 7–10, 2004)

Paper No. 8
Presentation Time: 10:05 AM

REINTERPRETATION OF REACTION PROGRESS AS A RECORD OF THE GEOMETRY OF FLUID FLOW DURING METAMORPHISM: WAITS RIVER FORMATION, EAST-CENTRAL VERMONT


PENNISTON-DORLAND, Sarah C. and FERRY, John M., Department of Earth and Planetary Sciences, Johns Hopkins Univ, Baltimore, MD 21218, sarahpd@jhu.edu

The spatial distribution of progress (ξ) of the infiltration-driven reaction Ms + 3Ank + 2Qtz=Bt + An + 2Cal + 4CO2 records the geometry of reactive fluid flow through marls during regional metamorphism at P≈8 kbar and T≈550°C. Marl is the sole lithology at the study site. Variations in ξ were measured perpendicular to layering at mm to m scales. Along 9- and 16-cm traverses ξ varies between 0 and 1.0±0.05 (1σ) mol/L in layers 0.5-4 cm thick. Similar variations in ξ exist among samples 0 to 3.1 m apart along two other traverses 5 and 35 m long. The variations traditionally would be interpreted in terms of reactive fluid flow in cm-wide layer-parallel channels. If this explanation is correct, variations in fluid composition should occur on the same scale. Along both cm-scale traverses, however, XCO2 and δ18OCal (proxy for δ18OFluid) are statistically consistent with single values whose best estimates are the weighted mean (±95% confidence interval): XCO2=0.130±0.003 (MSWD=0.6); δ18OCal=18.9±0.1‰ VSMOW (MSWD=2.3) for the 9 cm traverse and XCO2=0.180±0.008 (MSWD=0.5); δ18OCal=19.2±0.1‰ (MSWD=2.1) for the 16 cm traverse. Along the longer traverses some samples record statistically significant differences in XCO2 and δ18OCal over distances as small as 0.4 m, while other samples record uniform fluid composition over distances up to 4.2 m. Simple models demonstrate that cm-scale variations in ξ instead result from (a) layer-by-layer variations in the composition and amount of mineral reactants prior to reaction (initial Xan=0.00-0.37 and Fe/(Fe+Mg)Ank=0.1-0.4; initial Ank=1.5-3.0, Pl=0.2-2.0, Ms=0.4-1.2, and Cal=0-18 mol/L) and (b) cross-layer homogenization of fluid composition during subsequent infiltration and reaction. Significant m-scale cross-layer variations in XCO2 and δ18OCal rule out fluid flow across layers and point to layer-parallel flow in channels in some cases ≤1.4 m wide. Because there are no local sources of H2O, reactive fluid must be external. In general, time-integrated fluid flux and the geometry of fluid flow in rocks with mineral reactants and products that are solid solutions can never be determined from spatial variations in ξ at the scale over which fluid composition is homogenized during reaction. Data from these traverses place this distance between 15 cm and 4.2 m for Barrovian metamorphism.