The sulfur-isotopic composition of cenozoic seawater sulfate: Implicationsfor pyrite burial and atmospheric oxygen

The availability of high-temporal-resolution C- and S-isotope curves for th e Cenozoic permit for the first time modeling of the influence of the C and S cycles on the partial pressure of atmospheric O-2 on comparable time sca les. A simple isotope mass-balance model was used to calculate atmospheric O-2 levels from the burial rates of organic C and pyrite S. Burial rates we re derived from the C- and S-isotope records of seawater-dissolved inorgani c C and sulfate. Results indicate that in the early Cenozoic atmospheric O- 2 levels were about 16% higher than current levels. Extension of the model to Phanerozoic time scales yields atmospheric O-2 levels that are inconsist ent with geological evidence that suggests that the mass of atmospheric O-2 has not changed by more than a factor of two from the present atmospheric level since the Cambrian (Berkner and Marshall, 1974; Watson et al., 1978; Jones and Chaloner, 1991). These results indicate that either our knowledge of the parameters controll ing atmospheric O-2 is incomplete, or that the assumptions used in such mod els inadequately represent the complexity of the natural systems. Here we c ritically examine the assumptions inherent in isotope mass-balance models t o determine whether they may be the source of the model-data discord. A maj or problem with these models is the extreme sensitivity of the mass of atmo spheric O-2 to very small changes in the much larger masses of oxidized and reduced C and S reservoirs. For example, small variations in continental w eathering fluxes and the associated isotope ratios of river input have prof ound effects on calculated O-2 levels and need to be accounted for. Similar ly, variations in the isotopic composition of pyrite and organic C buried i n sediments, which are strongly influenced by changes in isotopic fractiona tion, dramatically influence calculated O-2 concentrations. Thus, constant fractionation factors should not be applied in such models. In addition, th e assumption that the isotopic composition of dissolved inorganic C is cont rolled only by the relative amounts of reduced and oxidized C buried in sed iments and their respective isotope ratios is questionable when relatively short time scales are considered. Isotope mass-balance models do not adequa tely encompass and simulate the actual processes being modeled because of t he simplifications and assumptions made. More "realistic" models are requir ed to achieve stabilization of atmospheric O-2 over geological time.