Does carbon 13 track anthropogenic CO2 in the southern ocean?

A multiparametric linear regression technique was used for waters in the Southern Ocean to estimate the change in both δ13CDIC (Δδ13CDIC) and dissolved inorganic carbon (DIC) (ΔDIC) between 1978 and 1998, due to the accumulation of anthropogenic CO2. The observed decrease in Δ13CDIC at the surface, the Suess effect, was −0.015 ± 0.003‰ yr−1 in the Sub‐Antarctic Zone and −0.005 ± 0.003‰ yr−1 in the Antarctic Zone, similar to values reported for the southern Indian Ocean [Gruber et al; 1999; Sonnerup et al., 2000]. To compare the Δδ13CDIC with ΔDIC, we used the ratio of these two anomalies (ΔRC=–Δδ13CDIC/ΔDIC‰ (μmol; kg−1)−1). Along the section, ΔRC ranged from 0.015 ± 0.005 at 42°S to 0.007 ± 0.005‰ (μmol; kg−1)−1 at 54°S. The spatial variability in ΔRC in the Southern Ocean reflects different timescales for processes controlling the uptake of 13C from those controlling the uptake of 12C and indicates that Δδ13CDIC decouples from ΔDIC poleward of the Sub‐Antarctic Zone. The variations of ΔRC along the section suggest that the δ13CDIC anomaly is not a good predictor of the anthropogenic CO2 inventory in the Southern Ocean. Some methods for determining anthropogenic CO2 uptake both on the global and regional scale assume the penetration depths of Δδ13CDIC to be the same as ΔDIC, which implies a constant value for ΔRC in the ocean at ∼0.016‰ (μmol; kg−1)−1 [Heimann and Maier‐Reimer, 1996; Ortiz et al., 2000; Bauch et al., 2000]. The use of a constant ΔRC and the observed Δδ13CDIC to estimate anthropogenic CO2 could lead to an underestimate in the inventory of anthropogenic CO2 for the Southern Ocean by ∼50%.