add and test FV3_RRFS_v1beta by siwei-noaa · Pull Request #270 · ufs-community/regional_workflow

siwei-noaa · 2020-08-17T22:51:28Z

DESCRIPTION OF CHANGES:

The SDF FV3_RRFS_v1beta has been merged into the ufs-weather-model. The regional workflow does not have capability/configurations for running this new SDF. Such capability was added and tested in this request.

It should be noted that two topographic files are required to run FV3_RRFS_v1beta as "drag_suite" scheme is used. So far (Aug 17, 2020), no scripts for automatic creating the required files, and they are copied from a temporary directory.

TESTS CONDUCTED:

The capability for running SDF FV3_RRFS_v1beta has been added in the regional workflow.
The FV3_RRFS_v1beta test was performed on NOAA Hera with GSD_HRRR 25km pre-defined CONUS domain.

CONTRIBUTORS:

Dr. Michael Toy (NOAA/GSL) provided the required topographic files for using "drag_suite" scheme.

The capability to run RRFS_v1beta has been tested and added to the regional workflow. As drag_suite scheme was used in the SDF, additional topo files are required to run this suite. No scripts are available for creating these topo files automatically so far (Aug 17, 2020), the required files are put in a temporary dir (/scratch2/BMC/gsd-fv3-test/she/orog) for now.

mkavulich · 2020-08-18T19:27:48Z

@gsketefian Would it be appropriate to add this suite to one of the existing end-to-end test suites so that it gets regularly tested?

JulieSchramm · 2020-08-18T19:35:35Z

This suite or the FV3_GFS_v16beta needs to be the out-of-the-box case.

…

On Tue, Aug 18, 2020 at 1:28 PM Michael Kavulich ***@***.***> wrote: @gsketefian <https://github.com/gsketefian> Would it be appropriate to add this suite to one of the existing end-to-end test suites so that it gets regularly tested? — You are receiving this because your review was requested. Reply to this email directly, view it on GitHub <https://github.com/NOAA-EMC/regional_workflow/pull/270#issuecomment-675669390>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA3WNU22V7RJUQPAQDATNKDSBLI4HANCNFSM4QCLBFGQ> .

mkavulich

The file config.sh should not be committed to the repository; please remove it from version control in your branch.

mkavulich · 2020-08-18T19:54:48Z

    JPgrid_WIDE_HALO_WIDTH=6

-    DT_ATMOS="300"
+    DT_ATMOS="40"


We should not change the default timestep, this will change results for people unexpectedly. If you need a different timestep you can overwrite it in config.sh.

@mkavulich I suggested Siwei commit this change. The default timestep is too long, and will fail for this pre-defined domain every time. A dt_atmos of 40 is what we should be using as the default for all pre-defined domains, since they all use the same vertical coordinate structure, and that is currently the limiting factor for stability.

If that's the case then the change should definitely be noted in the PR message and broadcast more widely than just in this PR; it will change results for people who do not specify DT_ATMOS in their config.sh

Thank you. I revoked the change on DT_ATMOS, and will be on a separate PR.

mkavulich · 2020-08-18T19:57:42Z

 raw_orog_fp="${raw_dir}/${raw_orog_fn}"
 mv_vrfy "${raw_orog_fp_orig}" "${raw_orog_fp}"

+# copy two files for drag_suite to the orog directory (temporary)


If they are necessary for running the case, these files should be added to the standard static/fix files on each official platform. We should not point to personal directories that could change unexpectedly.

Thank you. We are waiting on Mike's script to replace those files, and will update this when it is ready.

JeffBeck-NOAA · 2020-08-18T20:15:51Z

The file config.sh should not be committed to the repository; please remove it from version control in your branch.

@siwei-noaa When you issue a "git add" for files in your local branch, be careful of "git add -A", as that will add everything that has changed, and may not be what you want to commit. The config.sh file is one example, since it's a user-defined file used for specific experiment configuration.

siwei-noaa · 2020-08-18T23:07:32Z

The file config.sh should not be committed to the repository; please remove it from version control in your branch.

The file has been removed.

gsketefian · 2020-08-20T22:27:32Z

@gsketefian Would it be appropriate to add this suite to one of the existing end-to-end test suites so that it gets regularly tested?

Yes, whenever a new suite is added, a new WE2E test configuration should be added to test it regularly, especially an important one like this one

@siwei-noaa Please create a new WE2E test for this scheme as follows. Start with the file regional_workflow/tests/config.regional_003.sh
Copy it to a new file named regional_workflow/tests/config.FV3_RRFS_v1beta.sh and then in that new file, change
CCPP_PHYS_SUITE="FV3_GSD_v0"
to
CCPP_PHYS_SUITE="FV3_RRFS_v1"
Then test it (on hera) as follows:
a) cd regional_workflow/tests
b) echo "FV3_RRFS_v1beta" > my_expts.txt
c) Make sure your python environment is loaded:
module use -a /contrib/miniconda3/modulefiles
module load miniconda3
conda activate regional_workflow
d) Run this new WE2E test:
./run_experiments.sh expts_file="my_expts.txt" machine="hera" account="gsd-fv3"
Here, I'm assuming you usually run under the gsd-fv3 account; if not, change it to your usual account.
This should generate a new experiment under expt_dirs (in a subdirectory named FV3_RRFS_v1beta) and automatically relaunch the workflow until it succeeds or fails.

I'm happy to talk on the phone if these instructions are not clear.

gsketefian · 2020-08-21T17:11:34Z

This suite or the FV3_GFS_v16beta needs to be the out-of-the-box case.
…
On Tue, Aug 18, 2020 at 1:28 PM Michael Kavulich @.***> wrote: @gsketefian https://github.com/gsketefian Would it be appropriate to add this suite to one of the existing end-to-end test suites so that it gets regularly tested? — You are receiving this because your review was requested. Reply to this email directly, view it on GitHub <#270 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA3WNU22V7RJUQPAQDATNKDSBLI4HANCNFSM4QCLBFGQ .

I agree -- probably should be this suite. But I suggest making that a separate PR.

JeffBeck-NOAA · 2020-08-21T19:08:37Z

@@ -244,6 +244,8 @@ Creating links in the FIXsar directory to the grid files..."
    fns=( \
 "C*${DOT_OR_USCORE}oro_data.tile${TILE_RGNL}.halo${NH0}.nc" \


@siwei-noaa Same comment as above for these lines. An "if [ "${CCPP_PHYS_SUITE}" = "FV3_RRFS_v1beta" ]" statement should be added for the oro_data_??.tile* files in this block. After the initial "if" statement, an "else" clause would have the original oro_data.tile* files, so that those would be defined for anything other than FV3_RRFS_v1beta.

In testing, will commit if it works.

… into develop Update local develop branch from upstream develop branch

Merege the latest upstream branch to my feature branch

JeffBeck-NOAA · 2020-08-22T03:13:14Z

 fi

+# two files for drag_suite scheme
+if [ "${CCPP_PHYS_SUITE}" = "FV3_RRFS_v1beta" ]; then


@siwei-noaa, one last bug in this block of code. All ${FIXsar} variables are now ${FIXLAM} in the community workflow. Once that change is in, I think we'll be ready to merge.

siwei-noaa added 2 commits August 17, 2020 21:36

RRFS_v1beta test finished

a2c6926

siwei-noaa requested review from BenjaminBlake-NOAA, JeffBeck-NOAA, JulieSchramm, RatkoVasic-NOAA, gsketefian and mkavulich August 17, 2020 22:51

siwei-noaa changed the title ~~add FV3_RRFS_vebeta test~~ add FV3_RRFS_v1beta test Aug 17, 2020

siwei-noaa changed the title ~~add FV3_RRFS_v1beta test~~ add and test FV3_RRFS_v1beta Aug 17, 2020

mkavulich requested changes Aug 18, 2020

View reviewed changes

siwei-noaa requested a review from mkavulich August 18, 2020 21:15

remove config.sh

26b866c

revoke change on DT_ATMOS

8f09ac1

JeffBeck-NOAA requested changes Aug 21, 2020

View reviewed changes

siwei-noaa added 4 commits August 21, 2020 21:24

add if condition for RRFS_v1beta

feb09a1

Merge branch 'develop' of https://github.com/NOAA-EMC/regional_workflow…

4b668a4

… into develop Update local develop branch from upstream develop branch

Merge branch 'develop' into feature/rrfs_v1beta_test

9194eba

Merege the latest upstream branch to my feature branch

bug fix for atmos_nthreads

9eb7f59

JeffBeck-NOAA requested changes Aug 22, 2020

View reviewed changes

FIXsar to FIXLAM

081dc3c

JeffBeck-NOAA merged commit 5d7fc1c into ufs-community:develop Aug 22, 2020

gsketefian mentioned this pull request Sep 1, 2020

WE2E test configuration for FV3_RRFS_v1beta #280

Merged

christinaholtNOAA mentioned this pull request Sep 3, 2020

Real-time components updated for dev1. NOAA-GSL/regional_workflow#2

Merged

		@@ -244,6 +244,8 @@ Creating links in the FIXsar directory to the grid files..."
		fns=( \
		"C*${DOT_OR_USCORE}oro_data.tile${TILE_RGNL}.halo${NH0}.nc" \

Conversation

siwei-noaa commented Aug 17, 2020

DESCRIPTION OF CHANGES:

TESTS CONDUCTED:

CONTRIBUTORS:

Uh oh!

mkavulich commented Aug 18, 2020

Uh oh!

JulieSchramm commented Aug 18, 2020 via email

Uh oh!

mkavulich left a comment

Choose a reason for hiding this comment

Uh oh!

mkavulich Aug 18, 2020

Choose a reason for hiding this comment

Uh oh!

JeffBeck-NOAA Aug 18, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mkavulich Aug 18, 2020

Choose a reason for hiding this comment

Uh oh!

siwei-noaa Aug 18, 2020

Choose a reason for hiding this comment

Uh oh!

mkavulich Aug 18, 2020

Choose a reason for hiding this comment

Uh oh!

siwei-noaa Aug 18, 2020

Choose a reason for hiding this comment

Uh oh!

JeffBeck-NOAA commented Aug 18, 2020

Uh oh!

siwei-noaa commented Aug 18, 2020

Uh oh!

gsketefian commented Aug 20, 2020

Uh oh!

gsketefian commented Aug 21, 2020

Uh oh!

Uh oh!

JeffBeck-NOAA Aug 21, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

siwei-noaa Aug 21, 2020

Choose a reason for hiding this comment

Uh oh!

JeffBeck-NOAA Aug 22, 2020

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

JeffBeck-NOAA Aug 18, 2020 •

edited

Loading

JeffBeck-NOAA Aug 21, 2020 •

edited

Loading