Stratification‐dependent mixing may increase sensitivity of a wind‐driven Atlantic overturning to surface freshwater flux

Abstract

[1] Stratification-dependent mixing is employed in a coupled climate model of intermediate complexity with a 3-dimensional ocean component. Oceanic vertical diffusivity is calculated as k $ N Àa , where N is the local buoyancy frequency. The sensitivity of the Atlantic meridional overturning circulation (AMOC) to freshwater forcing is tested for exponents 0 a 2 by first slowly increasing, then decreasing the freshwater forcing over the North Atlantic, keeping the model close to equilibrium. The surface fresh anomaly imposed between 20°N and 50°N in the Atlantic reaches the deep ocean by vertical diffusion, and by AMOC advection via the northern convection sites. The fresh anomaly leads to enhanced stratification and thereby reduces vertical mixing stronger for higher values of a. Consequently, the freshwater anomaly reaches the northern deep water formation regions less diluted, and reduces the AMOC more strongly compared to lower values of a. Our findings indicate that modeled changes in the AMOC depend critically on the details of the mixing parameterization employed in the model.

Salinity stratification of the Mediterranean Sea during the Messinian crisis: A first model analysis

“…To create significant stratification during the MSC we found that κ needs to be at the lower end of this range. Although corresponding values for diffusivity are not known it has been pointed out that in marginal settings (Bryden et al, 2014) or when strong water column stability is present (e.g., Marzeion and Levermann, 2009), mixing may indeed be minor. If κ is less than 10 −5 m 2 /s, even larger w and q or lower f values than those reported at the start of this section would allow for significant stratification (see supplementary Fig.…”

The impact of meltwater discharge from the Greenland ice sheet on the Atlantic nutrient supply to the northwest European shelf

“…2.3). Reasons for the uncertainty in simulated AMOC 182 M. Mathis and U. Mikolajewicz: The impact of meltwater discharge from the Greenland ice sheet responses have been linked to the grid resolution of the ocean model (Swingedouw et al, 2013;Weijer et al, 2012;Böning et al, 2016) and the implemented mixing schemes and convection parameterizations (Nilsson et al, 2003;Yu et al, 2008;Marzeion and Levermann, 2009), affecting the simulated export of the added freshwater from the SPG region to adjacent ocean basins. These uncertainties related to GIS meltwater discharge have to be assessed in addition to the uncertainties in the atmospherically driven AMOC weakening and stratification strengthening deduced from multi-model ensembles.…”

“…A schematic of the ocean grid and coupling domain is shown, e.g., in M19. Simulated physical and biogeochemical conditions on the NWES have been evaluated in various studies employing MPIOM/HAMOCC in coupled and uncoupled modes (Gröger et al, 2013;Mathis et al, 2015Mathis et al, , 2018Mathis et al, , 2019Pätsch et al, 2017;Hátún et al, 2017).…”

The impact of melt water discharge from the Greenland ice sheet on the Atlantic nutrient supply to the Northwest European Shelf

“…Besides, in global circulation models (GCMs) the simulated AMOC weakening varies between 10-60% after 100 years for a typical freshwater hosing experiment of 0.1 Sv to the northern North Atlantic (Stouffer et al, 2006;Yu et al, 2016;Swingedouw et al, 2015). Reasons for the uncertainty in simulated AMOC responses have been linked to the grid resolution of the ocean model (Swingedouw et al, 2013;Weijer et al, 2012;Böning et al, 2016) and the implemented mixing schemes and convection parametrizations (Nilsson et al, 2003;Yu et al, 2008;Marzeion and Levermann, 2009), affecting the simulated export of the added freshwater from the SPG region to adjacent ocean basins. These uncertainties related to GIS meltwater discharge have to be assessed in addition to the uncertainties in the atmospherically driven AMOC weakening and stratification strengthening deduced from multi-model ensembles.…”