|Distribution of carbon isotopes in the glacial ocean: A model study|In: Paleoceanography. American Geophysical Union: Washington, DC. ISSN 0883-8305, more
A series of simulations are conducted with a global climate model of intermediate complexity (MoBidiC). This model includes ocean circulation dynamics, including the carbon cycle, coupled to a zonally averaged atmosphere. Oceanic distributions of nutrients, apparent oxygen utilization, radiocarbon, and carbon 13 are discussed for the preindustrial era (validation) as well as three states of the glacial ocean, termed interstadial (very active formation of deep water in the North Atlantic Ocean), stadial (moderate convection and important flow of Antarctic Deep Water in the Atlantic), and Heinrich (no formation of North Atlantic Deep Water). The stadial and interstadial states are stable. The Heinrich state is forced by a continuous freshwater discharge into the North Atlantic. The model exhibits significant changes in the isotopic composition of the ocean between the three modeled glacial states. Results for d13C tend to be in qualitative agreement with paleoceanographic data, except that the model fails at representing the strong depletion in d13C in the Southern Ocean. The Heinrich Atlantic Ocean is older than the stadial ocean at all depths (up to 1500 years). The “interstadial” ocean has younger deep water and older intermediate water than the “stadial.” It is recognized that the simulated changes in intermediate water age are less reliable because of the structure of the model. Color tracers are used to show that changes in the isotopic composition of Atlantic bottom water are mainly related to a redistribution of water masses. A simple method is tested, by which it is possible to reconstruct the North Atlantic water flow from the zonal profiles of salinity and ?14C. Finally, dividing artificially the gas exchange rate in the Southern Ocean by four results in a 0.4‰ decrease in the d13C of Antarctic Bottom Water. Changes in new production are, comparatively, less effective at altering the d13C ratio.