Unresolved questions about infiltration-driven regional metamorphism include the driving force for fluid-rock reactions (disequilibrium or gradient flow) and the geometry of flow (layer-parallel or cross-layer). We addressed these questions by examining a 9 x 4.5 cm slab of finely-laminated impure carbonate rock from the Waits River Formation.

The sample contains Qtz, Ms, Bt, Pl, Cal, Ank, Chl, and accessory minerals. Low-grade equivalents lack Bt and Chl, which formed at 525-550°C and 7.8 kbar by:

3Ank + Ms + 2Qtz=> Bt + An + 2Cal + 4CO₂ ;

8Ank + Ms + 3Qtz + 4H₂O=> Bt + Chl + 8Cal + 8CO₂.

Mineral modes and compositions were measured on traverses across layers at intervals of 0.25 to 0.35 mm, with about 2,000 points counted for each distinct layer. Reactants and products of both reactions were observed in all layers. Significant variations in modes of biotite (0 to 15.6%) were observed on the millimeter scale. Calculated equilibrium fluid XCO₂ is constant within error across all layers (0.14 ± 0.01).

The driving force for reaction was evaluated by attempting to reproduce observed reaction progress in a one-dimensional model of disequilibrium flow, with distinct layers having the same thickness and protolith mode as the sample. The model predicts a reaction front with products but no reactants upstream and few or no products downstream. The spatial distribution of reaction progress observed in the rock is not reproduced, regardless of flow geometry, nor is the widespread occurrence of both reactants and products. A gradient flow mechanism for reaction is preferred.

Traditionally, variations in reaction progress by layer have been interpreted as evidence for heterogeneous layer-parallel flow. Constant XCO₂ suggests homogeneous uniform flow and equilibration of all layers over a distance of 9 cm by diffusion. Under these conditions, variation in reaction progress can be explained by the amount, composition, and extent of reaction of plagioclase in the protolith for each layer. Cm-scale variations in reaction progress do not necessarily constrain flow geometry. The question of flow geometry will be addressed by examination of multiple samples at progressively larger scales of separation.