Feedback mechanism in the oceanic carbon cycle

Abstract

In this thesis, I designed and implemented a simple atmosphere-ocean coupled carbon cycle model which can be used as a tool to uncover the mechanisms of the interaction between the dynamics of the atmosphere-ocean system and the oceanic reservoir of CO 2 on the 101 to 103 years time scale. The atmosphere-ocean coupled model is originally developed by Marotzke (20,21), and the biogeochemical model is developed by Follows(personal communication). The atmosphere-ocean-carbon model makes the atmosphere-ocean dynamics and the carbon cycle fully interactive, and results in two stationary states characterized by two distinct patterns of the thermohaline circulation. The temperature driven, high latitudes sinking mode showed significantly lower atmospheric pCO2 than the salinity-driven, low latitudes sinking mode. The atmosphere-ocean dynamics dominates the system behavior of the model. The carbon cycle weakly feedbacks on the atmosphere-ocean system through the radiation balance. The model reveals two feedback mechanisms, the global warming feedback and the thermohaline pCO 2 feedback. The thermohaline pCO2 feedback has three sub-components, which are the biological pump feedback, the outgassing feedback and the DIC exporting feedback. The numerical experiments estimate the relative importance among them. The system becomes less stable when all the feedback mechanism is introduced. The model could be used to understand some basic mechanism of the situations similar to the anthropogenic global warming. The stability analysis is applied to evaluate the model runs. The current rate of 7 GTC yr - 1 can induce the spontaneous shutdown of thermohaline circulation after 550 years of constant emission. The stability of the thermohaline circulation rapidly decreases even before the system stops the thermohaline circulation. The model parameterized surface alkalinity as a simple function of sea surface salinity or as a constant, rather than solving the alkalinity cycle explicitly. The system is sensitive to the parameterization, in which different assumptions on alkalinity lead to different results both analytically and numerically.