Martin Losch

Hi @jm-c it turns out that adding vertical diffusion is not so easy, because the obcs-values are computed at the beginning of the time step when nothing of the 3D-viscosity...

@jm-c do you have any opinion about this? I might have a chance to work on this within the next days.

> Need a test for the diagnostics, what do folks recommend? Which experiments have TLM tests? The TLM experiments are almost the same as the ADM experiments: 1D_ocean_ice_column, global_ocean.cs32x15, OpenAD,...

@jrscott here are my comments about the sea ice configuration: - the CPP-flags are default (except for salt), and I am fine with that. You may want to define SEAICE_LSR_ZEBRA...

@jrscott > how strongly do you feel that .TRUE. here is better physics? FWIW I'm getting results on par with PIOMAS estimates with it .FALSE., much too much ice with...

Thanks for fixing the editorial comments. I like your suggestion: have the experiment test different parts of the code, but add instructions how to make the z and the p-coordinte...

I also think that the most useful filename convention for hand-written AD and TL code is to attach suffixes `_ad` and `_tl`. Not sure if I understand the comment about...

Correction: with `LSR_minIniGuess .GE. 2`, the computation of the a new initial guess for the LSR solver is based on **local** residuals of the free drift solution and the original...

A similar issue appeared [here](http://mailman.mitgcm.org/pipermail/mitgcm-support/2021-August/012978.html) and could be fixed with the above recipe.

@ifenty, @jrscott As far as I know, the model always copies i = 1,1+Olx to i = sNx+1,sNx+Olx (and similar for i=1-Olx,0 and j). In that sense, it always assumes...