Replies: 7 comments 4 replies
-
|
@mdtoy This is probably more of a question for the fv3atm repo folks, since the host model is the one controlling output and clearing diagnostic buckets. If you search the ufs-weather-model repo for both 'fhzero' and 'nszero' (its integer iteration counterpart), you'll see how it is used in CCPP_driver.F90 (which is actually part of the FV3 host model code). In that file, it is used to call the diag_phys_zero and diag_rad_zero subroutines in GFS_typedefs.F90. You can see which variable buckets get cleared in there. 'fhzero' is also used in FV3GFS_io.F90, so perhaps its use in there is what is leading to your unexpected changes? Regardless, this is definitely an FV3 thing, so if you want a better answer, it would probably be best to seek support from FV3 folks. -Grant |
Beta Was this translation helpful? Give feedback.
-
|
Thanks, Grant. I'll go ahead and enter this as an issue in the https://github.com/NOAA-EMC/fv3atm repo. |
Beta Was this translation helpful? Give feedback.
-
|
Actually, change of plan. Unfortunately, it looks like this is something that got broken in the "ugwpv1_gsldrag" scheme. I just tested the sensitivity of the instantaneous fields (dynf*nc) to changes in "fhzero" with the FV3_GFS_v16 suite and there is no issue. I'll need to find the bug. |
Beta Was this translation helpful? Give feedback.
-
|
@grantfirl and @junwang-noaa I think you are both correct. The diag fields are calculated in each of the CCPP physics schemes, but what gets done with them, i.e., converting the cumulative fields into averaged tendencies over the period "fhzero" and bucket emptying seems to be done somewhere in fv3atm. I've looked at the tendency of zonal wind (dudt) due to physics runs with different values of fhzero, and they are identical until the forecast time fhzero is reached. Then the solutions diverge. This is even with all of the diag flags, i.e., ldiag, etc. set to false. |
Beta Was this translation helpful? Give feedback.
-
|
@mdtoy Just to recap where we are here. It seems you had found a bug - were you able to fix that? Are you saying that even with the bug fix the forecast of instantaneous zonal wind is sensitive to the value of fhzero? This is unexpected to me. |
Beta Was this translation helpful? Give feedback.
-
|
For historical reasons fhzero should always be set to 6 hours for global
applications. Forecasters are accustomed to this definition. We cannot
change this definition in products. While for the regional model, I
believe the operational setting is fhzero=3.
Fanglin
…On Thu, May 6, 2021 at 11:56 AM grantfirl ***@***.***> wrote:
@mdtoy <https://github.com/mdtoy> This is probably more of a question for
the fv3atm repo folks, since the host model is the one controlling output
and clearing diagnostic buckets. If you search the ufs-weather-model repo
for both 'fhzero' and 'nszero' (its integer iteration counterpart), you'll
see how it is used in CCPP_driver.F90 (which is actually part of the FV3
host model code). In that file, it is used to call the diag_phys_zero and
diag_rad_zero subroutines in GFS_typedefs.F90. You can see which variable
buckets get cleared in there. 'fhzero' is also used in FV3GFS_io.F90, so
perhaps its use in there is what is leading to your unexpected changes?
Regardless, this is definitely an FV3 thing, so if you want a better
answer, it would probably be best to seek support from FV3 folks.
-Grant
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#649 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKY5N2KCC4PY2Q7GEXSCSSDTMK32VANCNFSM44FSLFAQ>
.
--
*Fanglin Yang, Ph.D.*
*Chief, Model Physics Group*
*Modeling and Data Assimilation Branch*
*NOAA/NWS/NCEP Environmental Modeling Center*
*https://www.emc.ncep.noaa.gov/gmb/wx24fy/fyang/
<https://www.emc.ncep.noaa.gov/gmb/wx24fy/fyang/>*
|
Beta Was this translation helpful? Give feedback.
-
|
It is correct that changing fhzero should only alter the values of
accumulated fields, not the results of the instantaneous fields.
On Thu, May 6, 2021 at 12:02 PM Fanglin Yang ***@***.***>
wrote:
… For historical reasons fhzero should always be set to 6 hours for global
applications. Forecasters are accustomed to this definition. We cannot
change this definition in products. While for the regional model, I
believe the operational setting is fhzero=3.
Fanglin
On Thu, May 6, 2021 at 11:56 AM grantfirl ***@***.***> wrote:
> @mdtoy <https://github.com/mdtoy> This is probably more of a question
for
> the fv3atm repo folks, since the host model is the one controlling output
> and clearing diagnostic buckets. If you search the ufs-weather-model repo
> for both 'fhzero' and 'nszero' (its integer iteration counterpart),
you'll
> see how it is used in CCPP_driver.F90 (which is actually part of the FV3
> host model code). In that file, it is used to call the diag_phys_zero and
> diag_rad_zero subroutines in GFS_typedefs.F90. You can see which variable
> buckets get cleared in there. 'fhzero' is also used in FV3GFS_io.F90, so
> perhaps its use in there is what is leading to your unexpected changes?
> Regardless, this is definitely an FV3 thing, so if you want a better
> answer, it would probably be best to seek support from FV3 folks.
>
> -Grant
>
> —
> You are receiving this because you are subscribed to this thread.
> Reply to this email directly, view it on GitHub
> <
#649 (comment)
>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AKY5N2KCC4PY2Q7GEXSCSSDTMK32VANCNFSM44FSLFAQ
>
> .
>
--
*Fanglin Yang, Ph.D.*
*Chief, Model Physics Group*
*Modeling and Data Assimilation Branch*
*NOAA/NWS/NCEP Environmental Modeling Center*
*https://www.emc.ncep.noaa.gov/gmb/wx24fy/fyang/
<https://www.emc.ncep.noaa.gov/gmb/wx24fy/fyang/>*
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#649 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AI7D6TOCMQB375E4BLWSOYDTMK4RXANCNFSM44FSLFAQ>
.
|
Beta Was this translation helpful? Give feedback.
-
|
If I understand correctly, all the related calculations for fields related
to fhzero are done in CCPP, I am wondering what we expect to do in fv3 repo.
…On Thu, May 6, 2021 at 12:53 PM mdtoy ***@***.***> wrote:
Thanks, Grant. I'll go ahead and enter this as an issue in the
https://github.com/NOAA-EMC/fv3atm repo.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#649 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AI7D6TK4OOQFFNTFLDNAJVLTMLCRZANCNFSM44FSLFAQ>
.
|
Beta Was this translation helpful? Give feedback.
-
|
@ligiabernadet I am still working on this bug. Am getting help from Sam Trahan. Yes, even with the previous bug fix, the intantaneous fields, wind, etc., are sensitive to the value of fhzero. It is an issue with the ugwpv1_gsldrag scheme — it doesn’t happen with other schemes like drag_suite and unified_ugwp.
… On May 12, 2021, at 2:44 PM, ligiabernardet ***@***.*** ***@***.***>> wrote:
@mdtoy <https://github.com/mdtoy> Just to recap where we are here. It seems you had found a bug - were you able to fix that? Are you saying that even with the bug fix the forecast of instantaneous zonal wind is sensitive to the value of fhzero? This is unexpected to me.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#649 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFMODLFAKDUI4SPMSBG45GLTNLSEXANCNFSM44FSLFAQ>.
|
Beta Was this translation helpful? Give feedback.
-
|
I found the bug. Please see ufs-community/ufs-weather-model#908. Since it changes the answer, someone who know GWD should take a look at the impact. |
Beta Was this translation helpful? Give feedback.
-
|
Wow! Great catch. That is a very subtle mistake. Thank you.
… On Nov 10, 2021, at 3:54 PM, Phil Pegion ***@***.***> wrote:
I found the bug. Please see ufs-community/ufs-weather-model#908 <ufs-community/ufs-weather-model#908>. Since it changes the answer, someone who know GWD should take a look at the impact.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#649 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFMODLDHKPWBQBCYOZNXMQTULLZ2HANCNFSM44FSLFAQ>.
Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
Beta Was this translation helpful? Give feedback.
-
I am coming up to speed on outputting additional diagnostics from the FV3GFS. I have a question about the "fhzero" variable with the CCPP standard name "hours_between_clearing_of_diagnostic_buckets". As an experiment to gain some understanding of the variable, I changed it from the default 6 hours to 1 hour. My understanding is that this should only change the period of time for averaging the diagnostic variables, however, varying this value changed the forecast variables in the "dynf" as well as "phyf" output files such as urgd, vgrd winds, etc. They changed quite substantially, certainly beyond round-off error, after a 6-hour forecast. So, unless my understanding on the effect of changing "fhzero" is incorrect, this seems to be a bug in the code, which may mean this should become an issue.
My run directories for these two experiments, where the only difference is the "fhzero" input.nml option, are on Hera at:
/scratch1/BMC/wrfruc/mtoy/git_local/ufs-weather-model/run_C96_127L/fhzero_6 and
/scratch1/BMC/wrfruc/mtoy/git_local/ufs-weather-model/run_C96_127L/fhzero_1
Here is a plot of the difference in zonal wind (ugrd in m/s) between a 6 hour forecast with fhzero=6 and fhzero=1 for tile 1:
dlta_u_6hrs_tile1_200mb.pdf
Beta Was this translation helpful? Give feedback.
All reactions