Re-evaluation of Infiltration-driven Regional Metamorphism in Northern New England: New Transport Models with Solid Solution and Cross-layer Equilibration of Fluid Composition

Abstract

The spatial distribution of reactants and products of infiltration-driven decarbonation reactions can be a record of the geometry and amount of reactive fluid flow during regional metamorphism. In the past, these distributions have been interpreted assuming (1) minerals are fixed in composition and (2) single layers are chemically isolated. Because neither assumption is normally valid, new transport models were developed that specifically predict the spatial distribution of reactants and products of the well-studied biotite-forming reaction in marls from northern New England. The models consider (1) isothermal and isobaric (iso-P–T) flow, (2) horizontal flow in the direction of increasing T (up-T flow), and (3) vertical, upward flow in the direction of decreasing P and T (down-P–T flow). All models assume a medium composed of many thin layers that differ in the amounts and compositions of mineral solid solutions prior to reaction, that fluid flow is parallel to layering, and that fluid composition is homogenized across layering over a distance much greater than layer thickness. All models reproduce (1) the amounts of mineral reactants and products at outcrops in Maine and Vermont where there are extensive data, and (2) regional-scale observations of mineral reactants only at low grades and a region several kilometers wide at higher grade where mineral reactants and products coexist. Models of up-T flow and down-P–T flow are preferred because they additionally predict spatially widespread complete reaction at the highest grades. Results show that when reactants and products are solid solutions, down-P–T flow predicted by simple hydrodynamic models of regional metamorphic fluid flow is fully consistent with observed widespread spatial distributions of mineral reactants and products in metamorphic terrains. Results further demonstrate that layer-scale variations in reaction progress are better explained in terms of layer-by-layer variations in initial mineral abundances and compositions coupled with homogenization of fluid composition across layers than by channelized fluid flow focused into layers with elevated reaction progress. All models predict minimum time-integrated fluid fluxes of ∼103 mol fluid cm−2 rock, ∼1–2 orders of magnitude less than what has been estimated assuming minerals are fixed in composition.