Ou Wang

@jm-c I was able to get identical results using this PR and master for ECCO V4r5 setup that has ice-shelves.

Removing the two files (pkg/ctrl/ctrl_map_ini_ecco.F, pkg/seaice/seaice_cost_areasst.F) looks fine to me.

@mjlosch I did 3-month adjoint runs of latest ECCO V4r5 setup using this branch and also the master branch. There is no issue with compilation and the results of forward...

I can try to push the combined package. The file shelfice_thermodynamics.F is quite different from the one on the main branch. I'll probably have to rename shelfice_thermodynamics.F in the combined...

@mjlosch When @ifenty and I developed the combined package, we were able to reproduce results generated using the main branch code, based on an idealized configuration. I made a pull...

@hugke729 and @mjlosch I believe the sign of eps3 in the icefront package of the main branch is correct. We should not put a minus sign before eps3. However, there...

@hugke729 Thanks for the confirmation.

@mjlosch and @timothyas The latest commit now follows your suggestions: ctrlDir has to be added to filenames before calling active_file.F and active_file_ad.F. This is indeed a cleaner implementation. Another change...

@ifenty This is just a measure of the effect of incorrect albedo computations in this branch. I compared a pair of 3-year runs using this branch and the master branch....

@mjlosch Thank you for implementing this. As I mentioned in Monday's meeting, @ifenty and I are also implementing something to solve the albedo computation. See https://github.com/owang01/MITgcm/tree/tices_reset_in_seaice_growth_adx. A three-year forward run...