The two-component model coincidence: Evaluating the validity of marine dissolved organic radiocarbon as a stable-conservative tracer at Station M

This study examines the applicability of the two-component mixing model in one dimension (depth) to published depth profiles of DOC concentrations and Δ¹⁴C values from hydrographic Station M (34°50′N, 123°00′W). We have rederived the two-component model assuming DOC exhibits stable conservative tracer properties and compared it to a model that includes reactivity and radioactive decay. We used ¹⁴C^∗, the fractionation- and decay-corrected concentration of ¹⁴C atoms in DOC or DIC, as a new tool for evaluating the two-component model and carbon biogeochemistry at Station M. All relationships between DOC and other tracers predicted under this framework exhibited sharp discontinuities (“cusps”) at ~85–~250 m depth. Below the cusps, the best fit relationships to the depth profiles yielded negligible DOC reaction rates consistent with conservative mixing. These observations separate the water column into two distinct biogeochemical regimes, with the apparent dominance of production and consumption of more labile DOC constituents in the epipelagic, and physical mixing at depths below the cusp. Our results indicate DO¹⁴C^∗ vs. DI¹⁴C^∗ correlations are more precise tools than comparisons with Δ¹⁴C or salinity, for identifying regions with unique DOC biogeochemistry that are currently obscured by measurement uncertainties and natural variability in traditional two-component model analyses, such as Keeling Plots. Higher resolution observations are required to precisely locate the cusps and understand their relationship to the separation of more labile DOC in the surface from more conservative DOC at depth. These results do not refute the two-component model, but rather support it as a simplified approximation to the complex biogeochemical processes that vary throughout the water column. In this way, DOC concentrations and Δ¹⁴C values can be approximated as a semi-conservative tracer for select applications.