Journal compilation C © 2007 Blackwell Munksgaard

Since anthropogenic CO2 concentrations in the ocean cannot be measured, it is very difficult to assess the accuracy of the various estimates. Until now, only comparisons among the various approaches and hypotheses have been used to estimate the uncertainties of the results. Here we use three measured anthropogenic tracers (CFC-11, 14C and 3H) to assess the relevance of three estimates of anthropogenic CO2 distributions based upon very different hypotheses. In order to focus this work on the correlations among tracers and estimates of anthropogenic CO2 concentrations in the ocean, we chose as an example the data set from the WOCE I1 cruise (Indian Ocean; 1995), since it included data from the three tracers as well as data of the carbonate/CO2 properties. This choice further allows us to use the published results of anthropogenic CO2 concentrations using both the C ∗ and MIX approaches. Using four properties (total dissolved inorganic carbon, total alkalinity, dissolved oxygen and potential temperature) of this data set, we also estimated the distribution of anthropogenic CO2 using the recent TrOCA approach. The results of correlations of anthropogenic CO2 concentrations with the anthropogenic tracers 3H, CFC-11 and 14C, indicate that these correlations are significantly higher when anthropogenic CO2 is estimated using either the MIX or the TrOCA approaches than using the C ∗ approach. Based upon these results and the easiness to use the simple TrOCA approach we propose to use this method to unravel the distribution of anthropogenic carbon in the Ocean. 1