Removed parameter KPP_BEFORE_KAPPA_SHEAR since always true. #219
Conversation
Removed parameter KPP_BEFORE_KAPPA_SHEAR since always true. This cleans up some awkward code related to removing, then adding back, the shear mixing diffusivity. MOM6 implementation of KPP always matches the boundary layer diffusivity to a zero value of the interior diffusivity, even if there is a nonzero interior diffusivity. The reason for this approach is that matching to a nonzero interior diffusivity can produce some very spurious results. Because MOM6 always matches to zero interior diffusivity, we can remove the code related to earlier implementations where we matched to a non-zero interior value. Removing this code cleans up some logic nicely.
|
I see what you mean about changing the answers! Before your patch, setting To avoid associating the answer changes with this commit (which isn't the real cause of the answer changes), I've first changed the default to |
- In NOAA-GFDL/MOM6-examples#47 @StephenGriffies notes that answers significantly when removing the code that removed and then added back diffusivity due to the Jackson et al. scheme (kappa-shear). The answer changes have nothing to do with the removal of the code but is solely due to the round-off differences associated with adding and subtracting a number. To record this, I am changing the default to not make the Kappa-Shear adjustment so that the code change in #219 does not get associated with the answer changes.
- NOAA-GFDL/MOM6@8e6283a *Changed default for KPP_BEFORE_KAPPA_SHEAR - In #47 @StephenGriffies notes that answers significantly when removing the code that removed and then added back diffusivity due to the Jackson et al. scheme (kappa-shear). The answer changes have nothing to do with the removal of the code but is solely due to the round-off differences associated with adding and subtracting a number. To record this, I am changing the default to not make the Kappa-Shear adjustment so that the code change in NOAA-GFDL/MOM6#219 does not get associated with the answer changes.
|
Good call. Thanks for doing this. And I agree...we are much the better Steve On Fri, Sep 4, 2015 at 3:39 PM, Alistair Adcroft [email protected]
Dr. Stephen M. Griffies |
…date_16may2022 (*)Merge MOM6/main from 16 May 2022
updated makefile for use with current EMC build system
Removed parameter KPP_BEFORE_KAPPA_SHEAR since always true.
This cleans up some awkward code related to removing, then
adding back, the shear mixing diffusivity.
MOM6 implementation of KPP always matches the boundary
layer diffusivity to a zero value of the interior
diffusivity, even if there is a nonzero interior diffusivity.
The reason for this approach is that matching to a nonzero
interior diffusivity can produce some very spurious results.
Because MOM6 always matches to zero interior diffusivity,
we can remove the code related to earlier implementations where
we matched to a non-zero interior value. Removing this code
cleans up some logic nicely.