search

ufs-community / UFS_UTILS

Utilities for the NCEP models.
Other
21 stars 104 forks source link

Add LAM DA utility codes to UFS-UTILS #415

Closed EricRogers-NOAA closed 1 year ago

EricRogers-NOAA commented 3 years ago

Three utility codes were written by Tom Black to insert the GSI analysis into the LAM 0-h boundary file, as part of the changes to address the model imbalances seen when first running the regional FV3.

create_expanded_restart_files_for_DA.F90 (runs before forecast model during LAMDA) prep_for_regional_DA.F90 (runs after forecast model during LAMDA) move_DA_update_data.F90 (runs after GSI analysis)

Suggested plan : create new ./ufs_utils/sorc/lamda_utils.fd with all three codes, use Tom's build script to create binaries

GeorgeGayno-NOAA commented 3 years ago

@EricRogers-NOAA UFS_UTILS uses a CMake build now. So you should not need Tom's build script. I can help with the build.

edwardhartnett commented 3 years ago

Also we need these codes to be documented with doxygen. We can help with the formatting.

We also need some tests for these codes. Do you have any?

EricRogers-NOAA commented 3 years ago

Ed, I assume you mean regression tests? No, we don't have any, but setting them up should not be a big deal. Currently they are run in the 3 km LAM DA parallel, I've also run them at C96 resolution (a while ago). At 3 km resolution two of the three codes runs fast, one runs slow (because it processes FV3 restart files) George, I checked out the UFS_UTILS develop branch and added the three new codes under "lamda_utils.fd" and I think I have the cmake build for them working. If you can access WCOSS now, please see /gpfs/dell6/emc/modeling/noscrub/Eric.Rogers/fv3lam_for_dellp3.5/sorc/regional_utils_latest.fd. I see executables for the three codes in the ./exec directory. I'll test them on Thursday (I'm taking tomorrow off) Thanks Eric

GeorgeGayno-NOAA commented 3 years ago

Ed, I assume you mean regression tests? No, we don't have any, but setting them up should not be a big deal. Currently they are run in the 3 km LAM DA parallel, I've also run them at C96 resolution (a while ago). At 3 km resolution two of the three codes runs fast, one runs slow (because it processes FV3 restart files) George, I checked out the UFS_UTILS develop branch and added the three new codes under "lamda_utils.fd" and I think I have the cmake build for them working. If you can access WCOSS now, please see /gpfs/dell6/emc/modeling/noscrub/Eric.Rogers/fv3lam_for_dellp3.5/sorc/regional_utils_latest.fd. I see executables for the three codes in the ./exec directory. I'll test them on Thursday (I'm taking tomorrow off) Thanks Eric

@EricRogers-NOAA Please create your own fork of UFS_UTILS, then create a branch to add your changes. See the wiki page for details: https://github.com/NOAA-EMC/UFS_UTILS/wiki/3.-Checking-out-code-and-making-changes

EricRogers-NOAA commented 2 years ago

Forked UFS_UTILS develop branch and created feature/lamda_utils branch

GeorgeGayno-NOAA commented 2 years ago

@EricRogers-NOAA The nccmp utility is available to compare netcdf files. That utility is loaded from the build scripts. Example: https://github.com/ufs-community/UFS_UTILS/blob/develop/modulefiles/build.hera.intel.lua#L60

Have your regression test script load the build module. Example: https://github.com/ufs-community/UFS_UTILS/blob/develop/reg_tests/chgres_cube/driver.hera.sh#L29

And here is how nccmp is invoked for chgres_cube tests. It gives a nice summary of any differences - https://github.com/ufs-community/UFS_UTILS/blob/develop/reg_tests/chgres_cube/c96.regional.sh#L62

EricRogers-NOAA commented 2 years ago

@GeorgeGayno-NOAA Thanks, I made the change to use nccmp in the lamda_utils regression test working script and it worked fine on all machines.

GeorgeGayno-NOAA commented 1 year ago

@EricRogers-NOAA is now retired. Will ask the regional group if this is still needed. @BenjaminBlake-NOAA do you know?

BenjaminBlake-NOAA commented 1 year ago

@GeorgeGayno-NOAA I am not sure if these utilities are needed (or will be needed) in RRFS, but I don't think so. The LAMDA real-time parallel has also been shut off for a month or two now. @MatthewPyle-NOAA or @ShunLiu-NOAA, do you think these three codes are needed for RRFS?

GeorgeGayno-NOAA commented 1 year ago

@GeorgeGayno-NOAA I am not sure if these utilities are needed (or will be needed) in RRFS, but I don't think so. The LAMDA real-time parallel has also been shut off for a month or two now. @MatthewPyle-NOAA or @ShunLiu-NOAA, do you think these three codes are needed for RRFS?

Thanks @BenjaminBlake-NOAA. Let me close this issue. We can always reopen if necessary.

MatthewPyle-NOAA commented 1 year ago

I'm 99% sure this functionality was replaced by something simpler at GSL.

ShunLiu-NOAA commented 1 year ago

Ben and Matt,

Yes, GSL developed a new utility to handle this.

Shun

On Tue, Apr 11, 2023 at 9:21 AM MatthewPyle-NOAA @.***> wrote:

I'm 99% sure this functionality was replaced by something simpler at GSL.

— Reply to this email directly, view it on GitHub https://github.com/ufs-community/UFS_UTILS/issues/415#issuecomment-1503342457, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACD5GZ7UIZKIN7J6G25JMYLXAVLGRANCNFSM4ZHH3MBA . You are receiving this because you were mentioned.Message ID: @.***>