The preindustrial atmospheric14CO2 latitudinal gradient as related to exchanges among atmospheric, oceanic, and terrestrial reservoirs

Abstract

A three‐dimensional global tracer transport model (derived from a general circulation model) simulates geographic variations in atmospheric Δ14C in response to oceanic boundary conditions. Regional atmosphere‐ocean 14CO2 fluxes are controlled by regional wind‐dependent gas exchange coefficients (E), air‐sea pCO2 differences (ΔpCO2) and oceanic 14C/12C deficiencies. We find that various preindustrial oceanic scenarios reconstructed from reasonable sets of such air‐sea variables all produce model latitudinal gradients in atmospheric Δ14CO2 (the “NS Δ14C”) significantly greater than the measured preindustrial NS Δ14C of +4.4 ± 0.5‰ (45°N to 45°S), as estimated from pre‐twentieth century tree rings. The NS Δ14C is insensitive to regional ΔpCO2 but is strongly contingent on the chosen values for surface ocean Δ14C and E of southern high latitudes (>50°S). The simultaneous seasonality in sea ice extent and high‐latitude wind speeds may in part explain the discrepancy between modeled and measured preindustrial NS Δ14C. Terrestrial 14C sinks with longer turnover times (such as peatlands) also potentially may help to reduce the NS Δ14C. With our best estimates for preindustrial surface ocean Δ14C, ΔpCO2, and E values for other regions, we find a “best fit” ocean scenario in which the preindustrial southern surface ocean Δ14C is −95 ± 10 ‰, its ΔpCO2 is 0 ± 20 ppmv, and its E is 0.088 ± 0.010 M m−2 yr−1 μatm−1 (about 20% reduced from the widely accepted value of 0.110). An independent estimate of +6.5 ± 0.5 kg yr−1 for the net preindustrial oceanic 14C uptake is an important constraint on global mean and Antarctic E.