Comparison of recent Indian Ocean anthropogenic CO2 estimates with a historical approach

Abstract

This work compares the classic Chen and Millero [1979] approach for estimating anthropogenic CO2 from ocean carbon measurements with the more recent ΔC* technique used by Sabine et al. [1999] to estimate anthropogenic CO2 concentrations in the Indian Ocean. Application of the Chen technique to the WOCE/JGOFS Indian Ocean data set gives a total anthropogenic CO2 inventory that is essentially the same as the ΔC* inventory, but there are substantial differences in the distributions within the water column. Some of these differences result from details of the application of the techniques such as the choice of which equation to use for the preformed alkalinity concentration or the choice of stoichiometric ratio to use for the biological correction. More significant differences, however, result from two fundamental differences in the techniques. One fundamental difference between the two techniques is that changes in the properties of the subsurface waters are referenced to a single deep water value in the Chen approach instead of the multiple reference points from the isopycnal analysis used in the ΔC* approach. The second fundamental difference is in the estimation of the preindustrial TCO2 distribution. Many of the differences examined have counteracting effects that may result in a total anthropogenic CO2 inventory that is similar for the two techniques. However, this similarity does not imply that both approaches are right. Comparison with global carbon models and other measurement‐based techniques do not clearly demonstrate that one technique is better than another. However, given the additional constraints of the transient tracers and the isopycnal analysis, we believe that the ΔC* technique provides a more robust estimate.