DOES THERMAL STRUCTURE ADEQUATELY EXPLAIN MOST MASS TRANSFER AT DEPTH? AN ALTERNATIVE BASED ON GEOCHEMICAL MODELLING OF FRACTURE-WALLROCK INTERACTION

Ore deposit and hydrocarbon migration models in the uppermost crust appeal to thermal, topographic and mechanical causes for variations in hydraulic head that drive fluid flow and mass transfer. Because "basement" isotope and geochemical signatures are locally apparent in such upper crustal rocks and sediments, models have also been developed for the chemical interaction of basement and cover sequences, and some of these models appeal to thermally driven circulation of fluid into relatively low permeability basement rocks and back into the cover. At greater (metamorphic) depths, some researchers also envisage that fluid and mass transfer is driven primarily by thermally induced volume changes to the fluid. However, in order that thermal structure dominates hydraulic head gradients, permeabilities need to be high and fairly constant, and interconnection of porosity must occur over long distances, a difficult requirement in overpressured rocks. An alternative proposed here is that chemical interaction between wallrocks and fluids during fracturing cycles is adequate to explain most of the mass transfer observed in low permeability rocks, because visible chemical alteration in such rocks is typically dominated by veins and related alteration. The "reactor-style" geochemical modelling package HCh uses the Gibbs minimization method for titration, flow-through and more complex fluid-rock interactions for conditions up to 800°C and 500 MPa. The simulations demonstrate fluid-wallrock interaction followed by isolation of the modified fluid, pressure (or temperature) change, and mineral precipitation in a vein. Differential mass transfer is caused by the varying solubilities of minerals, with Ti- and Al-bearing minerals confined to wallrocks, others precipitating in veins. Using a variety of rocks and fluids representative of greenschist to amphibolite hydrothermal systems worldwide, simulations of pore pressure decreases of up to 10 MPa (similar to those measured during modern seismogenic events) produced good correlation with observed vein infill patterns. Both the field observations and models suggest that this method of differential mass transfer dominates most mid- to lower crustal (sub-solidus) metal transport.