Incorporate global VIIRS 30s vegetation type data #698
Comments
GeorgeGayno-NOAA
assigned GeorgeGayno-NOAA, barlage, yangfanglin and BenjaminBlake-NOAA
|
Copy the original data file (from @barlage) from Hera: to WCOSS2 (Cactus): |
|
Recovered a program from a defunct branch that will convert the raw file to a version readable by ufs_utils. The program, build script and run script are on WCOSS2 (Cactus) here: |
|
The program was run twice to create global and NH (5S to 90N) files. The files were placed on WCOSS2 (Cactus) here: |
|
@GeorgeGayno-NOAA George, could you please work with @KateFriedman-NOAA to add the data to the "fix" archive on all EMC supported platforms ? Thanks |
|
To test the new data, I recreated the Then, added the new vegetation files and linked to the other input surface files as needed: |
|
@GeorgeGayno-NOAA is there a naming convention for 30" data, e.g., vegetation_type.viirs.igbp.30s.nc, since this isn't technically 0.01deg? for future reference, did you use chop.f90 to add the attributes and lat/lon/corners info needed in the global surface type data file as well as the NA cutout? |
I can name them as '30s' is you like.
Yes, I added some attributes and corner point lat/lons. The latter are needed by the ESMF regridding. |
GeorgeGayno-NOAA
changed the title
|
The branch was tested on Cactus using 03c5790. A large, North American regional grid was created using the new NH/30s file. The grid driver script was modified as follows: The output model grid vegetation files were and examined using |
|
The above test was repeated using the new global 30s data. The grid driver script was modified as follows: The output model grid vegetation files were identical to those from the first test as expected. |
|
All |
|
Does anyone want to test my branch before the merge? I have the new data on WCOSS2. But I can set it up on another machine if needed. |
|
@GeorgeGayno-NOAA I pointed @BenjaminBlake-NOAA to your new dataset. I believe he will give it a test for the North American RRFS domain. |
@BenjaminBlake-NOAA I will need to help you set up your test. |
|
@GeorgeGayno-NOAA Sounds good, I was planning to use the SRW app workflow to regenerate the surface climo fix files. But if it makes more sense to use the UFS_UTILS driver scripts I'm happy to do that as well. |
|
All, let me know when you're ready for me to pull new fix files into the primary fix set. Thanks! :) |
Ben, test however you want. Just be sure to link to the new data ( |
Kate, there are three new files - The new files are on Cactus here: Wait until Ben is finished with his testing before doing anything. |
|
@GeorgeGayno-NOAA I created new surface climo fix files for the RRFS_NA_3km grid yesterday and am waiting to hear back from a colleague if the new files look good. |
|
Per recommendation of @barlage, I added the compression step for all new files:
This step was added to the |
|
New files must also be available for the community at: https://noaa-ufs-srw-pds.s3.amazonaws.com/index.html#fix/fix_sfc_climo/ |
|
@KateFriedman-NOAA There are three new vegetation files that need to be hosted on each machine in the
You can remove the "vegetation_type.viirs.igbp.conus.0.01.nc" file. But create a link with that name that points to the "vegetation_type.viirs.igbp.conus.30s.nc" file (for backwards compatibility). Thank you! |
|
I think we'd prefer not to make the symlink unless there is a compelling reason otherwise. Now that we are versioning fix directories in the workflow, the sfc_climo fix version will update at the same time as the scripts change the filename. Versions with the old filename would still point to the old sfc_climo fix directory that have the old file. |
|
@GeorgeGayno-NOAA I have copied the three new fix files into a new sfc_climo timestamp subfolder (20221017) and removed the requested file (without making a symlink). See here on Cactus: /lfs/h2/emc/global/noscrub/emc.global/FIX/fix/sfc_climo/20221017 If this looks good to you I'll rsync it to the other platforms and then update |
|
https://github.com/ufs-community/UFS_UTILS/blob/develop/fix/link_fixdirs.sh in UFS_UTILS should have the default fix version updated, so any stand-alone testing is also updated. |
Done at b8ae8c1. |
Perfect. Thanks.
The workflow does not create grids, correct? It points to the pre-existing data in the 'orog' directory, right? If so, your workflow is unaffected. |
|
@KateFriedman-NOAA As a heads up, there is another issue open (#702) that will add new sfc_climo files. You can also place these additional files in 20221017. Details to come. |
|
@KateFriedman-NOAA Please rsync the data to the other platforms we support. |
|
I have rsync'd the new 20221017 sfc_climo subfolder to the supported platforms. Will add the additional files when provided. Let me know when you're ready for global-workflow to change |
New high-resolution (30 sec.) North American and global files were added to the 'fixed' directory in a new version directory. All script comments and 'readthedocs' were updated. The link_fixdirs.sh script was updated to point at the latest version of the sfc_climo subdirectory. Fixes #698.
The default in UFS_UTILS |
To save memory and to speed up the creation of model grids, the original VIIRS 30 second vegetation data was upscaled to 0.03, 0.05 and 0.1-degree resolution. The upscaling was done by accumulating the raw data in each lower-res grid box, then picking the predominate category. A CONUS version of the original high-res data was also created. All files are found under
fix/sfc_climoTo support larger regional grids that cover North America, the regional group needs UFS_UTILS to work with the original high-resolution global dataset.
The text was updated successfully, but these errors were encountered: