| Summary: | The global ocean is a net sink for atmospheric carbon dioxide (CO2) under present-day climate, as the result of a complex interplay of various drivers which vary spatially and temporally. Global ocean-biogeochemical models are tools to investigate these drivers and can be applied to assess projected changes in the ocean CO2 uptake under climate change. However, model-based estimates of the present-day ocean CO$_2$ sink show discrepancies between different models and compared to observation-based estimates, highlighting the uncertainty in the estimates. The work presented in this thesis addresses three themes regarding the discrepancies and uncertainty in the simulated CO2 flux; 1) the skin temperature effect; 2) the drivers of the CO2 flux in the Southern Ocean; and 3) the use of biogeochemical data assimilation to identify and correct model biases. The focus is on one ocean-biogeochemical model, consisting of a physical model (NEMO), coupled to a biogeochemical model (MEDUSA) and a sea ice model (CICE). NEMO-MEDUSA-CICE forms the ocean component of the UK Earth System Model (UKESM1), which contributed to the latest phase of the Coupled Model Intercomparison Project. The analysis comprises runs with the coupled UKESM1 from the ensemble of historical UKESM1 runs and ocean-only runs in which NEMO-MEDUSA-CICE is forced with observation-based atmospheric data. The skin temperature effect describes a temperature gradient, typically a decrease, in the top millimeter of the ocean. When accounted for in observation-based estimates, the skin temperature effect enhances the ocean CO2 uptake, increasing the discrepancy to model-based estimates. In global models, in contrast, the skin temperature effect is typically not included. Here, the impact of including the skin temperature in NEMO-MEDUSA-CICE on the global CO2 uptake is tested in a 35 year ocean-only simulation (1980-2014). Initially, accounting for this effect leads to a transient increase in CO2 uptake. The added CO2 uptake largely accumulates in the mixed layer, ...
|
|---|