Antarctic Bottom Water in CMIP5 models : characteristics, formation, evolution

Observations suggest that the properties of Antarctic Bottom Water (AABW) are changing, causing significant steric sea level rise. Understanding the causes of these changes is critical for projections of future sea level, yet previous generations of climate models failed to represent AABW accurately. Present-day biases in AABW potential temperature, salinity and density are assessed for models from the Coupled Model Intercomparison Project phase 5 (CMIP5). \nCMIP5 models either have inaccurate bottom water properties in the present-day Southern Ocean or form AABW via the wrong process, open ocean deep convection in the subpolar gyres. Under climate change scenarios, open ocean deep convection is responsible \nfor bringing the warming signal to the Southern Ocean abyss. It is then advected equatorwards by AABW transport. In turn, the decrease in density associated with the warming results in a weakened density-driven AABW transport. The mean of 24 CMIP5 models projects a mean global steric sea level rise of 3.8 mm by 2100 for the abyssal 500 m, albeit with a large uncertainty due to the cross-model disagreement on bottom salinity changes. \nThe parameterisation of overflows does not show an improvement in AABW properties. \nSensitivity experiments are performed on the model HadGEM3. The trigger for deep convection in the Weddell Sea, a positive sea ice anomaly leading to anomalies in \nthe mixed layer depth, is identified. Varying three vertical mixing parameters modifies the original mixed layer anomaly, leading to a range of responses from arrested deep convection to deep convection over the entire Weddell Sea. In the arrested convection simulations, \nthe Antarctic Circumpolar Current strength is improved and the AABW properties and North Atlantic Deep Water formation are unchanged. These experiments indicate a \npossible way to stop Weddell Sea deep convection in models, to improve their Southern Ocean representation. \n