Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Bug in weightgen for C48 grids #1005

Open
ClaraDraper-NOAA opened this issue Dec 6, 2024 · 10 comments
Open

Bug in weightgen for C48 grids #1005

ClaraDraper-NOAA opened this issue Dec 6, 2024 · 10 comments
Assignees
Labels
bug Something isn't working

Comments

@ClaraDraper-NOAA
Copy link
Contributor

ClaraDraper-NOAA commented Dec 6, 2024

The dimension for C48 Gaussian files is incorrect.

Here, 2x48 is 96, not 94.

jdim = 94 ! cres * 2

We need to change jdim to 96 and jdimp to 98. Same for the char variables.

@GeorgeGayno-NOAA GeorgeGayno-NOAA self-assigned this Dec 6, 2024
@GeorgeGayno-NOAA GeorgeGayno-NOAA added the bug Something isn't working label Dec 6, 2024
@GeorgeGayno-NOAA GeorgeGayno-NOAA changed the title Bug in weightgen for C49 grids Bug in weightgen for C48 grids Dec 10, 2024
@GeorgeGayno-NOAA
Copy link
Collaborator

@ClaraDraper-NOAA - I think the '94' dimension may be intentional. C48 is being mapped to the old spectral GFS T62 gaussian grid, which is defined here: https://www.nco.ncep.noaa.gov/pmb/docs/on388/tableb.html#GRID98. Note the 192x94 dimension.

@ClaraDraper-NOAA
Copy link
Contributor Author

Interesting! Do you know if all Gaussian C48 output is expected to be on the T62 grid (as in, the GFS output history files, and the GSI increments/DA)? The issue I have at the moment is that the enkf is outputting increments files for C48 that are 192x96, and the latitudes in that file do not include the poles (conflicting with the above). So, I'm wondering if:
i) we should be expecting the GSI to do the DA/output the increments on the T62 grid you linked to (in which case I can look into changing it there) ,or
ii) The DA/increments are on the expected grid, and I'm misusing UFS_UTILS/weight_gen by expecting it to match the GSI output.

@GeorgeGayno-NOAA
Copy link
Collaborator

I would have thought that everyone would use the standard T62 grid.

I hate to update a grid definition that others may be using. If the grids for the various processes can't be unified, I would be happy to add a new option (call it "C48da"?) with the dimensions you need.

@ClaraDraper-NOAA
Copy link
Contributor Author

Let me look into what's happening on the DA side, before we decide how to handle it. It may be as simple as the lats in the output file aren't right.

Do you want to close this issue for now?

@GeorgeGayno-NOAA
Copy link
Collaborator

Let's keep it open for now.

I am looking through the 'fix' directory, I see some T94 files with dimensions of 192x96. For example: /work2/noaa/da/ggayno/save/UFS_UTILS/fix/am/global_gaussian_latitudes.t94.192.96.txt.

@ClaraDraper-NOAA
Copy link
Contributor Author

Sounds good. I should be able to get to it before the end of this week.

@ClaraDraper-NOAA
Copy link
Contributor Author

I see what you mean - both the T94 and T64 options are there.

For C48, the forecast history files are output at T94, and the DA is done at that resolution. I need to do more digging to understand how the dimensions are set though.

@ClaraDraper-NOAA
Copy link
Contributor Author

Summarizing our discussion: the code is OK as is, but would be made clearer by changing this line:

jdim              = 94   ! cres * 2

to something like:

jdim              = 94   ! NOTE: default is T62 grid for C48

@GeorgeGayno-NOAA
Copy link
Collaborator

Summarizing our discussion: the code is OK as is, but would be made clearer by changing this line:

jdim              = 94   ! cres * 2

to something like:

jdim              = 94   ! NOTE: default is T62 grid for C48

Some of the other gaussian grids correspond to NCEP standard grids. That would be worth noting as well. Let me do some quick research.

@GeorgeGayno-NOAA
Copy link
Collaborator

Here is what I found that would be worth noting in comments:

  • C48 corresponds to T62 gaussian (NCEP Grib1 grid number 98)
  • C128 corresponds to T170 gaussian (NCEP Grib1 grid number 170)
  • C192 corresponds to T254 gaussian (NCEP Grib1 grid number 127)
  • C768 corresponds to T1534 gaussian (Does not have a Grib1 number, but was the operational GFS v14 physics grid)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

2 participants