add dry mass fixer for IAU #14
Conversation
junwang-noaa
requested review from
yangfanglin,
RussTreadon-NOAA and
junwang-noaa
|
Jeff's branch was tested on Venus. The iau_drymassfixer works as intended when compiling and running the model in 64-bit mode. Dry mass is not conserved when running the model in 32-bit mode. This non-conservation is not due to IAU. That said, with IAU on the change in dry mass increases with respect to a non-IAU run. Turning on the iau_drymassfixer in the 32-bit model improves dry mass stability. Since GFS v16 compiles and runs the model in 32-bit mode with IAU on, this change is required for the GFS v16 implementation. |
|
I communicated with Lucas Harris at GFDL regarding the loss of dry mass loss when the model is run at 32-bit precision. He confirmed it is indeed the case due to floating point round issue. The loss is minor compared to the loss caused by IAU. |
* revisions to agree with gitlab atmos_cubed_sphere * modified to include all diagnostic variables. no math edits * modified to include all missing subroutines and 'math edits' * removed temporary files * removed extra & in fv_diagonstics line 2277. * Delete DIFF_DEC30 * Delete DIFF_JAN02 * Delete get_diff.sh~ * reverts model/fv_cmp.F90 to null version; removes extraneous files ; removes commented lines after fixed lines
sync with dev/emc
mods to atmosphere.F90 to implement dry mass fixer for IAU in https://onlinelibrary.wiley.com/doi/full/10.1111/j.1600-0870.2007.00299.x
requires iau_drymassfixer jswhit2/fv3atm branch (for mods in GFS_typedefs.F90 to add iau_drymassfixer logical namelist var.
related pull requests:
fv3atm #61
ufs-weather-model #56