The observed stability of Earth's climate over millions of years is thought to depend on the rate of carbon dioxide (CO2) release from the solid Earth being balanced by the rate of CO2 consumption by silicate weathering. During the Cenozoic era, spanning approximately the past 66 million years, the concurrent increases in the marine isotopic ratios of strontium, osmium and lithium suggest that extensive uplift of mountain ranges may have stimulated CO2 consumption by silicate weathering, but reconstructions of sea-floor spreading do not indicate a corresponding increase in CO2 inputs from volcanic degassing. The resulting imbalance would have depleted the atmosphere of all CO2 within a few million years. As a result, reconciling Cenozoic isotopic records with the need for mass balance in the long-term carbon cycle has been a major and unresolved challenge in geochemistry and Earth history. Here we show that enhanced sulphide oxidation coupled to carbonate dissolution can provide a transient source of CO2 to Earth's atmosphere that is relevant over geological timescales. Like drawdown by means of silicate weathering, this source is probably enhanced by tectonic uplift, and so may have contributed to the relative stability of the partial pressure of atmospheric CO2 during the Cenozoic. A variety of other hypotheses have been put forward to explain the 'Cenozoic isotope-weathering paradox', and the evolution of the carbon cycle probably depended on multiple processes. However, an important role for sulphide oxidation coupled to carbonate dissolution is consistent with records of radiogenic isotopes, atmospheric CO2 partial pressure and the evolution of the Cenozoic sulphur cycle, and could be accounted for by geologically reasonable changes in the global dioxygen cycle, suggesting that this CO2 source should be considered a potentially important but as yet generally unrecognized component of the long-term carbon cycle.

Samples of river bank sediments and rocks were collected from across the entire study 3 site in order to constrain the elemental and isotopic composition of different lithologic end-4 members. Sub-samples of the river bank sediments were separated using a riffle splitter and 5 powdered in a ball mill. Rock samples were disaggregated using an agate mortar and pestle 6 before being ground in a ball mill. Only measurements of Na, Ca, Mg, and Sr concentrations are discussed in this paper. The vessel. During the reaction, the digestion vessels were heated from below with a hotplate. 25The liberated H 2 S gas was passed through a condenser and bubbled through a solution of 26

The interplay between the hydrologic processes that supply, store, and route water in catchment systems and the chemical weathering reactions that add and remove solutes acts as an important control on chemical weathering fluxes. In this study, we use paired measurements of solute chemistry and runoff in four nested catchments that span the transition from the Andes Mountains to the Amazonian foreland floodplain in Peru in order to investigate the links between hydrology and weathering processes and to determine how these links change across a geomorphic gradient. All of the sites show variation in elemental concentrations and ratios with runoff consistent with hydrologically driven changes in lithologic sources, the extent of secondary mineral precipitation, and, potentially, fluid flow paths. In the Andean sites, solute concentrations are relatively constant despite large changes in runoff. This is in direct contrast to the foreland floodplain site, where solute concentrations are diluted as runoff increases. In the Andes-Amazon, the concentration-runoff behavior is correlated with the mean catchment slope angle, which suggests that erosional processes, by modulating the timescales over which weathering reactions occur within the critical zone, can be an underlying control on solute production and therefore on chemical weathering. Due to the co-variation between the geomorphic and hydrologic controls on chemical weathering, weathering fluxes in Andean sites are more sensitive to seasonal changes in runoff than in the foreland floodplain site.