Characteristics and Models for Carlin-Type Gold Deposits

Abstract Carlin-type gold deposits are restricted to a small part of the North American Cordillera, in northern Nevada and northwest Utah, and formed over a short interval of time (42–30 Ma) in the mid-Tertiary when the Yellowstone mantle plume is inferred to have been located below the subduction zone. They formed after a change in plate motions (43 Ma) at, or soon after, the onset of extension in an east-west-trending, subduction-related magmatic belt. The deposits do not show consistent spatial relationships to mid-Tertiary magmatic centers, rather, most are located along long-lived, deep crustal structures inherited from Late Proterozoic rifting and formation of a passive margin. These structures influenced subsequent patterns of sedimentation and deformation and localized multiple episodes of igneous and hydrothermal activity, many of which contain anomalous concentrations of gold. The mid-Tertiary surface topography was relatively flat and many systems were located below large shallow lakes. Most deposits are hosted in a Paleozoic miogeoclinal carbonate sequence that is either structurally overlain by a eugeoclinal siliciclastic sequence, the Roberts Mountains allochthon emplaced in Early Mississippian time, or stratigraphically overlain by a miogeoclinal siliciclastic sequence deposited in the resulting foredeep. These siliciclastic sequences are less permeable than underlying carbonate rocks and apparently caused fluids ascending along major structures to flow laterally into permeable and reactive rocks below them. In these areas, gold ore is localized at intersections of a complex array of structures with permeable and reactive strata. The common alteration, mineralogy, and geochemical signature of these deposits is a direct expression of the P, T, and composition of ore fluids. The deposits generally formed at depths of >2 km at temperatures of 250° to 150°C, from moderately acidic (pH ≈5), reduced fluids containing <6 wt percent NaCl equiv, <4 mole percent CO2, <0.4 mole percent CH4, and >0.01 mole percent H2S. The H2S concentration was critical because it suppressed the solubility of Fe, base metals, and Ag as chloride complexes and enhanced the solubility of Au and associated trace elements (e.g., As, Sb, Tl, and Hg) as sulfide complexes. Gold was transported as AuHS° and/or Au(HS)2−1 complexes. The main ore stage formed during cooling and neutralization of ore fluids by reactions with the host rocks. It is characterized by carbonate dissolution, argillization of silicates, sulfidation of ferroan minerals, and silicification of limestone. Gold occurs as submicron inclusions or solid solution in arsenian pyrite and precipitated as H2S was consumed by sulfidation of Fe released from ferroan minerals. The other common trace elements (e.g., Sb, Tl, Hg) also reside in arsenian pyrite. The ideal host rock consists of permeable ferroan carbonate that is completely dissolved and its contained iron completely sulfidized such that all that remains is gold-bearing arsenian pyrite. Accordingly, large tonnage, low-grade gold deposits (e.g., Gold Quarry) are in siliceous rocks with low carbonate and reactive iron contents, and small tonnage, high-grade gold deposits (e.g., Meikle) are in carbonate rocks with high concentrations of reactive iron. Late ore-stage quartz, calcite, orpiment, realgar, stibnite, and barite occur in open fractures and pores and their abundance varies tremendously from deposit to deposit. These minerals precipitated as the systems cooled and ore fluids mixed with local ground water. Boiling was generally not important. Isotopic data from different districts yield conflicting indications as to the source of ore fluids. Abundant stable isotope data (δD, δ18O, δ13C, δ34S) and limited radiogenic isotope data (Pb, Sr, Os) from the major trends and districts are consistent with models involving the circulation of meteoric water through sedimentary rocks. In contrast, δD, δ18O, and δ13C data from the Getchell trend suggest that gold was introduced by a deep-sourced fluid that was of metamorphic or magmatic origin. The apparent lack of mid-Tertiary intrusions in this district argues for a metamorphic fluid, although the characteristics of certain portions and stages of the deposits suggest there was a magmatic fluid component characterized by higher Cl, Fl, K, Fe, and Cs contents. N2/Ar/He ratios of fluid inclusions suggest there were inputs of mantle He. Carlin-type deposits do not fit neatly into any one of the models proposed for them. Although variably evolved meteoric water is present in all of them, they are deeper than low-sulfidation epithermal veins and there is little or no evidence of boiling. They are shallower than orogenic veins and metamorphic fluids have only been detected in one district. Magmatic models call upon concealed intrusions that are so far removed from the deposits that no coeval contact metamorphic rocks, breccia pipes, or zoned geochemical halos are recognized at current levels of exposure or drilling. If the numerous similarities among Carlin-type deposits reflect the presence of a common ore fluid, then only one of the fluids detected by isotopic methods can be the ore fluid and the others must be due to contamination. In this case, we find the metamorphic fluid model most attractive, because both Carlin-type and orogenic gold deposits form in broad thermal anomalies, are distributed along major crustal structures, form during a change in stress regime, have similar ages over wide areas, have monotonous geochemical signatures, and contain similar endowments of gold. If we rely on the best data available from each district, a variety of models is needed and the only common factor is the geologic setting. These considerations suggest that Carlin-type deposits are unique, or too complex, to neatly fit into any one of these models.