* Fixed sign error in thickness diffuse work

[marshallward]

This patch fixes a bug in the sign of the tendency of the work by
thickness diffusion in the top layer along the u-points.

This resolves a variance in model runs after a 90-degree rotation, and
results are now consistent with the work_v calculation.

This patch will change answers for any runs using work-based thickness
diffusion.

[codecov-io]

Codecov Report

Merging #1050 into dev/gfdl will decrease coverage by <.01%.
The diff coverage is 62.5%.

@@             Coverage Diff              @@
##           dev/gfdl    #1050      +/-   ##
============================================
- Coverage     45.06%   45.06%   -0.01%     
============================================
  Files           212      212              
  Lines         62624    62629       +5     
============================================
+ Hits          28224    28226       +2     
- Misses        34400    34403       +3

Impacted Files	Coverage Δ
...arameterizations/lateral/MOM_thickness_diffuse.F90	63.08% <62.5%> (-0.14%)	⬇️
...rc/parameterizations/vertical/MOM_tidal_mixing.F90	9.81% <0%> (-0.07%)	⬇️
src/user/RGC_initialization.F90	0% <0%> (ø)	⬆️
src/user/ISOMIP_initialization.F90	0% <0%> (ø)	⬆️
src/user/DOME_initialization.F90	0% <0%> (ø)	⬆️
src/user/DOME2d_initialization.F90	0% <0%> (ø)	⬆️
src/user/dumbbell_initialization.F90	0% <0%> (ø)	⬆️
src/user/dense_water_initialization.F90	0% <0%> (ø)	⬆️
src/user/BFB_initialization.F90	0% <0%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9676443...b598a2d. Read the comment docs.

[marshallward]

Via @adcroft, I need to add a flag to reproduce the old behavior before merging this.

[StephenGriffies]

Does this bug fix resolve the issues raised by @gustavo-marques with negative GM work term?

[adcroft]

We expect so but don't know yet.

[adcroft]

Pending test: https://gitlab.gfdl.noaa.gov/ogrp/MOM6/pipelines/9596

[gustavo-marques]

This patch does not fix NOAA-GFDL#965. GMwork is still positive at high latitudes, see figure below. This can be reproduced by the following changes in MOM6-examples/ice_ocean_SIS2/OM4_05:

1. cp diag_table.MOM diag_table
2. add "ocean_model", "GMwork", "GMwork", "ocean_daily", "all", "mean", "none",2 into diag_table
3. add #override USE_GM_WORK_BUG=False into MOM_override
4. set days = 10 in input.nml

[adcroft]

Is that the same color scale as for the figures in #965 ? If so then it does appear to be better.

[gustavo-marques]

Is that the same color scale as for the figures in #965 ? If so then it does appear to be better.

No, it is not the same scale. Below is a comparison between before (A) and after (B) this fix. The maximum value is larger after the fix but it does seem to reduce locations with positive values.

Mininum= -0.17472036 Maximum= 0.0049198894
Mininum= -0.1757154 Maximum= 0.009520339
Mininum= -0.013858554 Maximum= 0.009868649