LAUSR

Abstract Volcanism, aridity, and endorheism converge in the central zone of the Andean Cordillera (Bolivia, Chile, and Argentina, between latitudes 19° S and 27° S). The Gorbea and Ignorado basins are pristine Andean sites in which the transfer of saline compounds from endogenous to exogenous environments occurs. In this area the regional volcanic rocks display strong argillic alteration, with Al and Fe (alunite, jarosite) and Ca (gypsum/anhydrite) sulfates. Native sulfur is also present in paleosolfataras. The Gorbea salt flat is covered by a discontinuous layer of selenitic gypsum of varying thickness (maximum 2 m). The discontinuity of the layer as well as its variable thickness is due to the fact that the original bed has been partially destroyed mainly by dissolution but also by deflation. Saline pond brines (Cl-SO4-Na [-Mg]) are strongly acidic reaching pH values lower than 2. The high temperature processes that caused the hydrothermal alteration in the Gorbea and Ignorado basins occurred in the Miocene (14 Ma) coinciding with a wet period that reached 9 Ma. Subsequently, the weather up to 120 ka was predominantly hyper-arid with a less arid interval between 6 and 3 Ma and the epithermal sulfates were recycled in saline lakes mainly in the Late Pleistocene wet period (120 to 11.7 ka). Evolution into the current salt flat occurred in the mid-early Holocene (11.7 to 4 ka), through a salt lake that first evolved into a ‘salina’ environment that gave rise to a selenitic gypsum layer (6.4 ka), and later to the final dryness. The highest values of δ34SVCDT and δ18OVSMOW found in the selenitic gypsum layer (+20‰ and + 28‰, respectively) show that the recycling was locally produced, mainly from isotopically heavier hypogene sulfates. The δ18OVSMOW (and less clearly, δ34SVCDT) values are higher in the basal part of the gypsum crust (about +27‰), which suggests an initial hypogene source that decreased towards the top due to mixing with supergene sulfate inputs. Bacterial activity, although catalyzing the supergene reactions, does not appear to have had a significant influence on the isotopic composition of sulfates. The crystallization water of the gypsum is isotopically lighter in the basal part of the selenitic layer (δ18OVSMOW ≈ +7‰), which indicates that the brines of the saline lake were still poorly evolved. These brines evolved to heavier (approximately +13‰) towards the top of the layer and towards the center of the salt flat, before the final drying. The partial destruction of the selenitic gypsum layer occurred during more recent wet periods over the last 4 ka that have been identified in wetlands and lakes in the Central Andean area. Isotopic data (δ34SVCDT and δ18OVSMOW) clearly suggest that sulfates in the saline compounds and evaporites from the entire Central Andean arid area have mostly a thermal origin with contributions from atmospheric deposition and locally, near the Pacific coast, marine aerosols. In addition, the great difference in altitude (approximately 3000 m) between the Altiplano and the lands located to the west, up to the Pacific coast, generated a constant flow of groundwater containing saline compounds that gave rise, to the salt flats of the basins located in lower topographic areas (Atacama, Punta Negra, Hilaricos, Soledad, Tamarugal, Salar de Pintados, Salar Grande) throughout the Cenozoic. Such salt flats have lower δ34SVCDT and δ18OVSMOW values mostly for two reasons: the secular mixing with atmospheric sulfate, and the isotope fractionation related to repetitive dissolution (or leaking)—migration—precipitation along the hydraulic gradient, a process that occurred throughout a large part of the Cenozoic. The last process also explains that the evaporites of some of these salt flats (e.g., Salar de Atacama, Salar Grande) display very high Cl/SO4 ratios.