HR4 Gravity Wave Drag Update #207
Conversation
…wp.F90 module_sf_noahmplsm.F90 noahmpdrv.F90
Qingfu-Liu
requested review from
barlage,
cenlinhe,
grantfirl,
dustinswales,
mdtoyNOAA and
BoYang-NOAA
as code owners
|
@JongilHan66 are there namelist switches to go with this update? |
|
@lisa-bengtsson The namelist change related to workflow, I will contact the workflow code manager to change the repository |
|
@Qingfu-Liu I understand. I am wondering what they are, so I can test on my own. |
|
@lisa-bengtsson Can you take a look of the compile problem for suite suite_FV3_HRRR_c3.xml ? There is a conflict with the new code. |
|
@Qingfu-Liu can you merge the latest develop? there are noahmp changes that are not included in your branch. |
@barlage OK. I will merge the new change. |
…ule_sf_noahmp_glacier.F90
…ule_sf_noahmp_glacier.F90
|
New code is uploaded, but there is problem to compile the new code. The CCPP-physics repository has updated significantly, there is a conflict to merge the new code. Currently working on it. |
Code clean-up is on my to-do list. I think when the dust settles and we have a final version for GFSv17 would be the time to do it. |
Mike, as we discussed before we need to reorganize and remove some of the legacy code as well. There are too many versions and some duplicated namelist options. |
|
@Qingfu-Liu @mdtoyNOAA The plan is to combine #210 into this one? Does anyone need any help debugging/testing to move this along? |
@grantfirl The "alpha_fd" in #210 is already combined into the updated codes, so we don't need PR#210 any more. |
Thanks @JongilHan66 . I'll close #210, I just didn't see the "updated codes" reflected in this PR yet. I'm assuming that @Qingfu-Liu is working on it locally and hasn't pushed to GitHub yet (which is fine). |
I just able to compile with the new update from the GWD. I will commit the new changes soon |
|
All the code for HR4 GWD and Noahmp are just updated based on the new suggestions |
Hi @grantfirl I have uploaded all the code and only run tests for two suites. If you have time, could you please run all the suites tests. Thank you very much |
OK, I'll run them now. |
|
@Qingfu-Liu I've kicked off the full rt.conf on Hera. I'll help debug if there are any unexpected failures. |
|
@Qingfu-Liu Please merge in Qingfu-Liu#3 for some minor fixes. |
Random fixes for HR4 GWD update
@grantfirl I just merged Qingfu-Liu#3 to PR#207 |
@Qingfu-Liu I verified the merge. Looks good! |
@grantfirl Thanks |
grantfirl
dismissed
barlage’s stale review
Requested changes have been addressed.
|
The changes of the NoahMP model (which are not used for HR4-GWD-update tests) have been retracted. All the regression tests are passed on Hera |
| do k = kpblmax,km | ||
| if ((taud_ls(i,k)+taud_bl(i,k)).ne.0..and.prsl(i,k).le.pcutoff) then | ||
| denfac = min(ro(i,k)/pcutoff_den,1.) | ||
| dtfac(i,k) = min(dtfac(i,k),denfac*abs(velco(i,k) & |
There was a problem hiding this comment.
@JongilHan66 This is the line causing the latest trouble in ufs-community/ufs-weather-model#2290. The problem is that this look is going up to 'km' when velco is only defined up to km-1. Would it work to simply change the loop from kpblmax to km-1, or does this mess up something else?
There was a problem hiding this comment.
Let me know, and I can make the change and test that it works real quick.
There was a problem hiding this comment.
Let me know, and I can make the change and test that it works real quick.
@grantfirl I don't think the change causes a problem. Please change the loop from kpblmax to km-1.
BTW, in 'drag_suite.meta', 'vtype' and 'psl_gwd_dx_factor' were removed from the original meta file. Is it ok?
There was a problem hiding this comment.
@JongilHan66 Yes, they should have been removed because drag_suite_psl() is not actually exposed to the CCPP. It is only being called from within ugwpv1_gsldrag_run or unified_ugwp_run. Both the ugwpv1_gsldrag and unified_ugwp schemes have those two variables accessible to pass in to drag_suite_psl already.
There was a problem hiding this comment.
@JongilHan66 Yes, they should have been removed because
drag_suite_psl()is not actually exposed to the CCPP. It is only being called from withinugwpv1_gsldrag_runorunified_ugwp_run. Both the ugwpv1_gsldrag and unified_ugwp schemes have those two variables accessible to pass in to drag_suite_psl already.
@grantfirl I see. Thanks!!
…broutine; change upper bound of loop to prevent array bound excursion
The code of this PR is provided by Jongil Han.
This update is a combination of the gravity wave drag (GWD) versions from the NOAA/GSL and NOAA/PSL. The PR also update the Land Noahmp code.
The test results with this update can be seen in:
a) summer: https://www.emc.ncep.noaa.gov/gmb/jhan/vsdbw/hr3e50s
b) winter: https://www.emc.ncep.noaa.gov/gmb/jhan/vsdbw/hr3d11w
HR4GWD: same as HR3 but with the updated GWD
Compared to the HR2 and HR3, the HR4GWD shows a significant improvement especially in 500mb height AC and CONUS precipitation forecast skills. The HR4GWD reduces the cold and dry biases in the lower troposphere compared to the HR3. It also reduces the negative wind speed biases in the troposphere compared to the HR2.