Mixed Layer lateral eddy fluxes mediated by air-sea interaction

The published study

The modulation of air-sea heat fluxes by geostrophic eddies due to the stirring of temperature at the sea surface is discussed and quantified. It is argued that the damping of eddy temperature variance by such air-sea fluxes enhances the dissipation of surface temperature fields. Depending on the timescale of damping relative to that of the eddying motions, surface eddy diffusivities can be significantly enhanced over interior values. The issues are explored and quantified in a controlled setting by driving a tracer field, a proxy for sea surface temperature, with surface altimetric observations in the Antarctic Circumpolar Current (ACC) of the Southern Ocean. A new, tracer-based diagnostic of eddy diffusivity is introduced which is related to the Nakamura effective diffusivity. Using this, the surface lateral eddy diffusivities associated with (i) eddy stirring and small-scale mixing and (ii) surface damping by air-sea interaction is quantified. In the ACC, a diffusivity associated with surface damping of a comparable magnitude to that associated with eddy stirring (~500m2/s) is found. In frontal regions prevalent in the ACC, an augmentation of surface lateral eddy diffusivities of this magnitude is equivalent to an air-sea flux of 100W/m2 acting over a mixed layer depth of 100m – a very significant effect. Finally, the implications of our results for other tracer fields such as salinity, dissolved gases and chlorophyll are discussed. Different tracers are found to have surface eddy diffusivities that differ significantly in magnitude.


To perform this analysis I used the cycle1 of a 1/8 global MITgcm simulation:
2000-2005 integration,
1/8-deg horizontal grid spacing,
50 vertical levels,
forced by NCEP reanalysis-1 surface atmospheric state and bulk formulae,
surface relaxation to montly Levitus SSS with a relaxation time constant of 44.5 days,
with full dynamic thermodynamic sea-ice model
Its sweet name is: lle.nb.01 and datas are available at MIT/CSAIL here:
More informations can be found here: http://data.guillaumemaze.org/ecco2-mitgcm
Binary list: http://data.guillaumemaze.org/ecco2-mitgcm#TOC-Cycle-1:-ll2.nb.01-:-2000--2005
Netcdf list: http://data.guillaumemaze.org/ecco2-mitgcm#TOC-Cycle-1:-ll2.nb.01-:-2000--2005-bro

How to compute stuff for Eddy diffusivity with 1/8 simulations ?

How to get basic input fields ?

It is assumed that all fields are already in netcdf, one file per time step, one folder per time step

  • Compute daily fields from 6-hourly for fields like Qnet, SST, ... with:


  • Compute monthly means from daily fields for fields and cross-fields like: Qnet.^2, Qnet*SST, ... with:




RQ: The following script allows to extract 6-hourly netcdf files from original binaries on the csail network and to send them to Beagle


How to get eddy fields ?

  • Compute long time means of 2D fields like [Q^2], [Q*SST], ... with:


  • Compute EOFs of coarsed resolutions fields with:


  • Compute the seasonal cycle eddy field like [Qsc*SSTsc] ... on the original 1/8 resolution with:


  • Compute eddy fields like [Q'SST'] ... with:


  • Also create a single file (mat and cdf) with all surface climatologies with:


How to get air-sea induced eddy diffusivity ?

  • Compute long time mean of 3D potential temperature with:


  • Compute horizontal gradient of 3D theta with:


  • Compute mean horizontal gradient of theta in the MLD (even if supposed to be vertically homogeneous) with:


  • Compute eddy diffusivity with: