Mixed layer, bottom and barotropic 2-D fields from ESPC-D

[Alan Wallcraft]

The ESPC-D-V02 Global analysis replaces GOFS 3.5, but it does not provide the "_bottom" T,S,u,v or the following "sur" fields: qtot, emp, u_barotropic_velocity, v_barotropic_velocity, surface_boundary_layer_thickness, mixed_layer_thickness.

There is no way to recover surface_boundary_layer_thickness (from KPP), qtot, or emp, but we can approximate mixed_layer_thickness, u_barotropic_velocity, v_barotropic_velocity, and [TSuv]_bottom from the "3z" fields.

In HYCOM-tools see Add espc_mixlay and espc_bottom

I forgot to mention in the description of the programs that the input netCDF fields from ESPC-D can be a subset of lon and lat.  So you need all 40 levels but not necessarily the entire GLBy0.08 domain.

Alan.

[mark]

Alan,

Do you know if there are plans to add the mixed layer depth product ("_sur")  back into the mix whenever the next model update?  We have customers who really like this product and I am not sure we have the processing capability here to derive the estimated mixed layer depth as it was in the GLBy0.08 product.  Are there plans to update the model from the present ESPC-D-V02 any time soon?  Do you also have any idea whether they will 'fix' the strange eddies in the temperature data off the west coast of the U.S. and Canada that apparently are there due to bad buoy data going into the model?  Any info you can provide will be much appreciated.

Thanks,

Mark

[Alan Wallcraft]

Calculating MLD off-line is straight forward and relatively inexpensive, but it does require you to download all the 3z levels of T and S.  This is how we do it:

foreach f ( ${R}_${year}${mon}??12_M0000_t3z.nc )
  setenv A `echo $f | sed -e "s/_t3z.nc//g"`
#
  if (-f ${A}_t3z.nc && \
      -f ${A}_s3z.nc && \
      -f ${A}_u3z.nc && \
      -f ${A}_v3z.nc    ) then
    if (! -f ${A}_bot.nc) then
      setenv CDF_INST   "COAPS/FSU"
      setenv CDF_PUBLIC "yes"
      setenv CDF_HIST   "espc_bottom"
      setenv CDF_TAU    0
      setenv CDF001     ${A}_bot.nc
      /bin/rm -f $CDF001
      ${HT}/archive/src/espc_bottom <<E-o-D
${TT}/depth_${R}_${T}.nc
${A}_t3z.nc
${A}_s3z.nc
${A}_u3z.nc
${A}_v3z.nc
   0    'itest ' = grid point where detailed diagnostics are desired
   0    'jtest ' = grid point where detailed diagnostics are desired
 031    'iexpt ' = experiment number x10
   0.3  'tmljmq' = equiv. temp. jump across mixed-layer (degC, 0 no I/O)
E-o-D
      date
#     created ${A}_bot.nc

This is also calculating the velocity 2-D fields.

We are using above script fragment to form daily mean archives (for nesting) and it is running every day to form the previous days daily mean, see the "bot" files at /datasets/ESPC-D-V02/data/daily_netcdf/2025/

The Navy is unlikely to add MLD to its output unless one of its operational customers asks for it.

The following is from Joe Metzger on the next version:

We have plans to do a VTR for ESPC-D v2.1 by 1^st or 2^nd quarter of FY26. It will use HYCOM v2.3.01 and include a) rivers as a mass flux, b) a new rivers database developed by Cheryl Ann, c) include parameterized Stokes drift mixing (since we still only have one-way coupling to WW3), d) include SWOT altimeter assimilation utilizing adaptive scales that are defined by the resolution of the observations, i.e., take advantage of the finer space scales resolved by the wide-swath altimeter, e) include landfast sea ice in both hemispheres, and e) sea ice thickness assimilation. I’ve probably forgotten something but that is the main list of improvements on the ocean/sea ice side.

This is a substantial upgrade that should produce a measurable improvement.

A Validation Testing Report that demonstrates improvements over existing capabilities is the last step before entering OPTEST (i.e. before it runs every day in the operational queues), but the amount of time a product stays in OPTEST before becoming fully operational varies a lot.  We might be able to distribute the OPTEST in parallel with ESPC-D v2.  FY26Q2 ends March 31st, so at best this might be available a year from now.

The strange eddies were an NCODA issue that has been fixed. It is taking a while to spin down, but attached are two pictures. One is just after we started applying the correction and the other is after it’s been implemented a while. The checkerboard pattern is slightly visible in the more recent plot, but for less egregious.

Let me know if you think this is still an issue.

Alan.