Quantifying key drivers of marine pyrite content and isotopic composition

Microbial sulfate reduction (MSR) and subsequent pyrite burial in marine sediments plays a crucial role in Earth’s long-term carbon and oxygen budgets; by reducing sulfate and producing alkalinity, this process effectively increases atmospheric O₂ and lowers CO₂ levels. Given that MSR exhibits a large and reduction-rate-dependent sulfur-isotope fractionation, changes in pyrite sulfur-isotope compositions (δ³⁴S values) through geologic time have long been interpreted to reflect global signals such as marine sulfate reduction rates, microbial community behavior, and the amount of sulfur buried as pyrite vs. evaporite minerals. However, recent research has demonstrated that marine MSR fractionation likely operates near an equilibrium limit regardless of reduction rate. These studies instead implicate local environmental and sedimentological factors as drivers of pyrite δ³⁴S values. Despite this advancement, the sensitivity of pyrite formation rate and δ³⁴S value to changes in environmental and sedimentological variables—both today and through geologic time—remains unconstrained due to the complex interactions between controlling variables.

To provide mechanistic and quantitative constraints, we developed and applied a non-dimensional diagenetic model that extracts the natural variables governing pyrite formation. Assuming equilibrium MSR isotope fractionation and using only locally measured or globally interpolated boundary values as inputs (i.e., no free parameters), our model accurately predicts all available modern observations (n = 216 cores) with an average root-mean square error of 0.3 wt % for pyrite content and 16.5 ‰ for δ³⁴S. Extrapolating this result, we estimate global pyrite burial flux to be ~1.3 × 10¹² mol FeS₂ yr⁻¹ with a weighted-average δ³⁴S value of ~-21 ‰ VCDT.This flux is statistically identical to independent estimates of total riverine sulfate input (i.e., pyrite-oxidation and evaporite-dissolution derived), indicating the sulfur cycle currently operates in steady state. However, calculated pyrite burial exceeds pyrite-oxidation derived inputs, suggesting net atmospheric O₂ release and CO₂ consumption by the sulfur cycle.

Mechanistically, we conclude that pyrite formation rate is highly sensitive to local reactive iron input, whereas δ³⁴S value is primarily controlled by organic carbon reactivity-to-sedimentation rate ratio (termed Da*, a modified Damköhler number) and organic carbon-to-sulfate ratio (termed Γ₀). In contrast to previous models, we show that pyrite δ³⁴S is largely insensitive to bioturbation due to counter-balancing impacts on Da* and Γ₀. Rather, when combined with a geologic pyrite δ³⁴S record, our interpretation requires an increase in Da* and decrease in Γ₀ since the Paleozoic, possibly driven by changing organic matter reactivity and sulfate concentration through geologic time.