Terra rossa clays have long been thought to form by residual dissolution of limestone and/or by accumulation of detrital mud, ash, or dust on preexisting karst limestones. Here we present new petrographic and field evidence that the Bloomington, Indiana, terra rossa forms by replacement of limestone by authigenic clay at a moving metasomatic front, with the clay's major chemical elements, Fe, Al and Si, probably coming from dissolved dust. The replacement of clay for limestone, because it releases acid, which dissolves out new porosity at the front, should trigger a reactive-infiltration instability that would be predicted to "carve" the funnels and sinks that contain the terra rossa itself. Quantitative modeling of the chemical dynamics of terra rossa formation correctly predicts the formation of two adjacent zones in the reaction front, replacement and leaching, and considerable leaching porosity in the latter zone, as observed. Crucial in the model is that replacement takes place via crystallization stress-driven pressure solution of the host by the guest mineral. Numerical simulations produce predicted front widths and front velocities well within an order of magnitude of the observed front thickness (approximately 9 cm) and paleomagnetically-derived rates of terra rossa formation at Bloomington of at least 2 m/Ma. A separate quantitative model of terra rossa formation incorporating the alternative replacement mechanism of dissolution-precipitation preferred by most geochemists, also produces two reaction zones, but they come out in reverse spatial order to that observed in the field. The predicted volumes of kaolinite grown and calcite dissolved differ by three orders of magnitude, also contrary to the observation that the clay-for-limestone replacement preserves bulk mineral volume. This alternate model suggests that replacement cannot proceed by "dissolution-precipitation".

Fossil hydrothermal system associated with world class Carlin-type gold mineralization is characterized by the absence of a clear magmatic fluid source, relatively deep (>5 km) fluid flow, large δ¹⁸O shift (between 10 to 20‰), and remarkably high temperatures (>200°C) at shallower depths (< 2000 m). While the plumbing of this system varies, geochemical and isotopic data collected from the fossil hydrothermal systems suggest that fluid circulation along fault zones was comprised of variably exchanged Pleistocene meteoric water. We present a suite of two-dimensional local (40–50 km) scale hydrologic models to study the plumbing of the fossil hydrothermal systems of the Great Basin using multiple constraints, including temperature, δ¹⁸O of fluids and/or rocks and silica compositions. Our results suggest that the fossil hydrothermal systems were density driven. In the Carlin trend, the fault controlled fluid circulation extended deep (5 to 6 km) down into Paleozoic siliciclastic rocks, which afforded more mixing with isotopically enriched higher enthalpy fluids. Computed fission track ages along the Carlin trend included the convective effects, and ranged between 91.6 and 35.3 Ma. Older fission track ages occurred in zones of groundwater recharge, and the younger ages occurred in discharge areas. We found that either an amagmatic system with more permeable faults (10^-11 m²) or a magmatic system with less permeable faults (10 ^-13 m²) could account for the published isotopic and thermal data along the Carlin trend systems. However, both magmatic and amagmatic scenarios require the existence of deep, permeable faults to bring hot fluids to the near surface.