The ocean carbon cycle plays a fundamental role in the Earth’s climate system, on decadal to multi-millennial timescales. Of the carbon held in the ocean, the atmosphere, and the terrestrial biosphere combined, more than 90% resides in the ocean. Carbon enters the surface ocean through air-sea gas exchange and from terrestrial sources. It is transported to the deep ocean with the ocean circulation and through the so-called biological pump, where carbon is taken up in the surface ocean by photosynthetic organisms that fall down and decompose at depth. This thesis contributes to the understanding of the processes involved in ocean carbon storage and transport. It examines how these processes respond to model perturbations, and how this response influences our attempts to simulate glacial-interglacial fluctuations in atmospheric carbon dioxide (CO 2 ). The thesis investigates the response of the simulated ocean carbon storage, and distribution of the isotopic tracer δ 13 C, to changes in physical and biological parameters. In the included studies, we use observational as well as proxy records of oceanic properties to evaluate our model simulations. In addition, we use a climate model to interpret proxy evidence of glacial-interglacial changes in ocean δ 13 C. By using a separation framework, we identify the origin of the carbon in the model ocean, and attribute observed changes to the processes involved. The results indicate a strong link between ocean carbon storage and the strength of the global ocean overturning circulation. Stronger circulation leads to less carbon storage through a weakening of the biological pump, and through reduced solubility due to an increase in global ocean average temperature. In simulations of glacial climate, we find that biological adaptability to the surrounding nutrient conditions, through a flexible carbon-to-phosphorus ratio (C/P) in ocean photosynthesis, increases the ocean carbon storage compared to simulations where fixed C/P is applied. The biological flexibility improves the model’s ability to reproduce glacial atmospheric CO 2 . In line with previous research, we find freshwater input to the North Atlantic to be an important factor for reproducing glacial proxy records. The ensemble of simulations that achieve a good representation of glacial-interglacial δ 13 C indicates a deglacial whole-ocean change in δ 13 C of 0.28 ± 0.06‰. The thesis underlines the importance of the initial state, and the choice of model parameterisations, for the outcome of model ensemble, and intercomparison studies. Finally, it proposes a new method for estimation of ocean carbon transport, and attribution of this transport to different water masses and carbon system processes.