6. Methods
Tide Model Simulations
• OTIS (Oregon State Tidal Inversion
Software)
• M2, S2, O1, K1
• Di
ff
erent resolutions (up to 1/12°)
• Di
ff
erent Internal Tide (IT) Drag
Parameterizations
• Di
ff
erent LGM bathymetries (ICE5G,
ICE6G)
Climate Model Simulations
• Oregon State University version of
University of Victoria model (OSU-UVic)
• Tidal Mixing Parameterization (Jayne, St.
Laurent, Simmons)
• Model of Ocean Biogeochemistry &
Isotopes (MOBI) includes paleo tracers
δ13C, and radiocarbon
• Vary Southern Ocean buoyancy
fl
uxes
• Standard LGM Boundary Conditions (ice
sheets, greenhouse gas concentr., orbit)
8. =
Tidal Mixing Parameterization
Diapycnal Di
ff
usivity:
Subgrid-scale bathymetry:
Schmittner & Egbert (2014) Geosc. Mod. Devel.
Dissipation E
ffi
ciency = Fraction
of locally dissipated energy
Considers only changes in
locally dissipated energy, which
is only 1/3 of the total!
15. Conclusions
• Increased tidal mixing in LGM is robust result, but quantitatively depends on
reconstructed bathymetry (basin geometry; land ice extent)
• Increased di
ff
usivities increase AMOC & AABW
fl
ow rates
• AMOC strongly a
ff
ects isotopes
• E
ff
ect of increased mixing is more subtle but improves model-data
agreement
• All these results are conservative because they neglect changes in remotely
dissipated energy
• To Do: include remotely dissipated energy in future simulations e.g. by using Eden & Olbers
parameterization (IDEMIX)
16. Conclusions
• Feedback (climate <-> mixing) is probably small
• Even large increase in tidal dissipation (doubling - tripling) leads only to
modest increases in AMOC (1-3 Sv) and AABW (1-2 Sv)
• (Except when AMOC is close to threshold)
• E
ff
ects of changes in surface buoyancy
fl
uxes are much larger
• Paleo simulations should include changes in tidal mixing