Add RUC LSM option to land perturbations, add snow albedo and albedo to perturbed parameters #1
Conversation
background albedo components to the perturbations.
|
@DomHeinzeller Dom, Could you please add Clara Draper to the reviewer's list. Thank you! |
|
@ClaraDraper-NOAA FYI - once you accept the invite to collaborate on the GSL fork of stochastic_physics, I can add you as reviewer. Thanks! |
DomHeinzeller
changed the title
Add semis to lndp_apply_perts
…railing whitespaces, pass only one value lsoil from wrapper
More cleanup for Tanya
|
Are we done with the changes or not? I have started the regression tests already, changes. but need to redo (some of) them if we keep making changes. |
|
Dom, |
No problem, I will recreate our rt_ccpp_dev baseline and verify against it. |
|
Thank you, Dom. |
|
@climbfuji I looked more carefully at the print-outs from my latest run, and see that the values of surface parameters become unrealistically small. The output after time step 381: We must have broken something (maybe when we add emissivity?) because I did not see unrealistic values in my previous pre-PR testing. The old output is in: /scratch2/BMC/wrfruc/smirnova/fv3_rt/MANUAL_TEST_C768_2020090500/output_stoch_repro/out |
|
We didn't break anything, we fixed it. Because you didn't loop over the blocks after running the land surface perturbation code, only one block - if at all - got updated, all others remained unchanged. I will hold back until you adjusted the parameters so that the results look better. |
|
@climbfuji You are right, you fixed the bug which was masking this problem. My further testing showed that the unrealistic values of surface parameters appear because of applying perturbations at every time step, and it was not designed this way by Clara Draper who wrote the main code. I would need to consult with Clara on this. |
2. When lndp_each_step=.true. perform conversion from change/per hour to change per time step 3. Mimicking what is done in HRRR, apply soil moisture perturbations to RUC LSM at the beginning of the run when time step = 2.
|
@climbfuji Dom, in the last commit I have implemented some ideas after our conversation with Clara this morning. The code also has fixes for the unreasonable parameter values. Now the code is very close to final. I am still waiting for some modifications from Clara for roughness length perturbations. Then I think we'll be good to run regression tests. Thank you! |
|
Thanks, Tanya. Given that I had PRs on top of yours I will need to update my PRs first and see how it all looks like. Probably tomorrow?
… On Jan 7, 2021, at 2:51 PM, tanyasmirnova ***@***.***> wrote:
@climbfuji <https://github.com/climbfuji> Dom, in the last commit I have implemented some ideas after our conversation with Clara this morning. The code also has fixes for the unreasonable parameter values. Now the code is very close to final. I am still waiting for some modifications from Clara for roughness length perturbations. Then I think we'll be good to run regression tests. Thank you!
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AN7FF5DX724UZCMTOZ3H67LSYYUGLANCNFSM4VEQI66Q>.
|
|
Sounds good. Thank you! |
This PR adds RUC LSM to the land perturbation code and also adds snow albedo and background albedo components to the perturbed parameters.
Associated PRs:
#1
NOAA-GSL/ccpp-physics#71
NOAA-GSL/fv3atm#68 (contains NOAA-GSL/fv3atm#65)
NOAA-GSL/ufs-weather-model#57
See NOAA-GSL/ufs-weather-model#57 for regression testing.