Résumé
Modeling the global ocean-atmosphere carbon dioxide system is becoming increasingly important to greenhouse gas policy. These models require initialization with realistic three-dimensional (3-D) oceanic carbon fields. This report presents an approach to establishing these initial conditions from an extensive global database of ocean carbon dioxide (CO2) system measurements and well-developed interpolation methods. These methods are limited to waters below the deepest mixed layer. The data used for these interpolations include the recent high-quality data sets from the World Ocean Circulation Experiment (WOCE), Joint Global Ocean Flux Study (JGOFS), and Ocean-Atmosphere Carbon Exchange Study (OACES) programs. The interpolation methodology employs correlation between CO2 system properties and other more widely measured properties: potential temperature, salinity, and apparent oxygen utilization. The correlations are computed for each profile, and the coefficients are interpolated to the 1А з 1А з 32 vertical-layer grid at a monthly temporal resolution. Finally, the gridded coefficients are applied to a global monthly climatology of ocean temperature, salinity, and oxygen to compute total CO2 (TCO2) and total alkalinity (TALK) for the 3-D grid. This approach offers advantages over spin up of a single profile in defining spatial variation in CO2 system properties because it reduces initialization time and provides a more accurate carbon field. The results provide an unprecedented "view" of the global distribution of TALK and TCO2 in the ocean. These results as well as those from the monthly mixed layer depths can be used in diagnostic and prognostic global ocean models.