Atmosphere Ocean Science Colloquium
Measuring mesoscale mixing to guide parameterization
Speaker: Shafer Smith, Courant/CAOS
Location: Warren Weaver Hall 1302
Date: Wednesday, September 29, 2021, 3:30 p.m.
Parameterizations of mesoscale eddies will remain an essential part of ocean climate models for the foreseeable future, motivating ongoing efforts to constrain and ``measure" eddy fluxes in both real and high-resolution modeled oceans. This has proved a surprisingly challenging task. Here I'll review two primary functions of current parameterizations --- the adiabatic relaxation of isopycnal surfaces, and the along-isopycnal diffusion of tracers --- explain why it's been difficult to constrain the fluxes they seek to parameterize in the same framework, and show that the method of multiple tracers overcomes these issues. The method used to measure fluxes in an eddy-resolving simulations set in a zonally-reentrant channel, forced by steady winds and surface buoyancy, with a meridional ridge that produces anisotropy and inhomogeneity in the flow.
The measured fluxes reveal the following: (1) in a significant fraction of the simulated flow, variance transport appears to be small, rationalizing local parameterization theory; (2) the mixing rate of tracers and QG potential vorticity (QGPV) are very similar; (3) the eddy QGPV flux is dominated by the vertical derivative of the buoyancy flux (the momentum and anisotropy terms are generally small); (4) as a result of the above, in most regions a QG-based theory that relates the mixing coefficients for adiabatic relaxation and isopycnal diffusion works well. This relationship is consistent with the fact that values of the adiabatic relaxation coefficient necessary to achieve a realistic stratification are about an order of magnitude smaller than observed values of isopycnal diffusion. Despite this, most ocean models set these coefficients equal to each other, resulting in serious errors in the representation of biogeochemical tracers like oxygen and carbon.