dtc/hwrf-physics: HWRF RRTMG (based on #412)

[climbfuji]

This PR is based on #412 by @mzhangw, but for branch dtc/hwrf-physics instead of dtc/develop. It also contains several bug fixes and updates to the original PR.

The following commits were cherry-picked from #412 (all from @mzhangw):
9e9222a
4d9e68f
6f9fec9
57873f2
7e492ca
298d1ae
5404462

Commit 298d1ae, "remove the connection of iovrlw/iovrsw with physparam", was reverted by commit 28d1bc2, "Clean up HWRF RRTMG additions", since these changes were not required and introduced an unnecessary, large amount of modifications to the radiation codes.

Associated PRs:
#430
NCAR/fv3atm#38
NCAR/ufs-weather-model#36

For regression testing information, see NCAR/ufs-weather-model#36

[mzhangw]

looks good to me. Thanks for your hard work, Dom!

[climbfuji]

looks good to me. Thanks for your hard work, Dom!

Thanks for reviewing. I am trying to make one more improvement in the ufs-weather-model repo, consolidating the run scripts and namelist templates. These will be tested in the final round of regression tests against the new baseline. Otherwise I think these PRs look good now.

[ligiabernardet]

I am not clear on how one switches between the GFS and HWRF radiation schemes. Is there a new scheme that corresponds to the HWRF radiation sw/lw?

[mzhangw]

There are a lot new-added WRF/phys subroutines. But for users, HWRF RRTMG switch mimics how this is done in HWRF namelist option: icloud =3 iovr_lw =4 ivor_sw=4 will do the trick. In SDF file, the calling sequence of SW/LW need to be switched too

…

On Wed, Apr 8, 2020 at 3:05 PM ligiabernardet ***@***.***> wrote: I am not clear on how one switches between the GFS and HWRF radiation schemes. Is there a new scheme that corresponds to the HWRF radiation sw/lw? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#430 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AG7TW2UQRSEU6E3KRAQGIWDRLTRJZANCNFSM4MD3XKKQ> .

[climbfuji]

See difference in the regression tests: ufs-weather-model PR, file tests/tests/fv3_ccpp_regional_c768_FA_update_moist versus tests/tests/fv3_ccpp_regional_c768_FA_HWRF_PBL

…

On Apr 8, 2020, at 3:16 PM, mzhangw ***@***.***> wrote: There are a lot new-added WRF/phys subroutines. But for users, HWRF RRTMG switch mimics how this is done in HWRF namelist option: icloud =3 iovr_lw =4 ivor_sw=4 will do the trick. In SDF file, the calling sequence of SW/LW need to be switched too On Wed, Apr 8, 2020 at 3:05 PM ligiabernardet ***@***.***> wrote: > I am not clear on how one switches between the GFS and HWRF radiation > schemes. Is there a new scheme that corresponds to the HWRF radiation sw/lw? > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#430 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AG7TW2UQRSEU6E3KRAQGIWDRLTRJZANCNFSM4MD3XKKQ> > . > — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#430 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB5C2RPLQ6IRXXRFS2QGNMLRLTSSTANCNFSM4MD3XKKQ>.

[ligiabernardet]

iovr changes the algorithm for cloud overlap. icloud changes the algorithm for calculation of partial clouds. I understand that part. Now, regarding the actual radiative transfer model, are there no differences between HWRF and GFS? On Wed, Apr 8, 2020 at 3:18 PM Dom Heinzeller <[email protected]> wrote:

…

See difference in the regression tests: ufs-weather-model PR, file tests/tests/fv3_ccpp_regional_c768_FA_update_moist versus tests/tests/fv3_ccpp_regional_c768_FA_HWRF_PBL > On Apr 8, 2020, at 3:16 PM, mzhangw ***@***.***> wrote: > > > There are a lot new-added WRF/phys subroutines. But for users, > HWRF RRTMG switch mimics how this is done in HWRF namelist option: > icloud =3 > iovr_lw =4 > ivor_sw=4 > > will do the trick. In SDF file, the calling sequence of SW/LW need to be > switched too > > > On Wed, Apr 8, 2020 at 3:05 PM ligiabernardet ***@***.***> > wrote: > > > I am not clear on how one switches between the GFS and HWRF radiation > > schemes. Is there a new scheme that corresponds to the HWRF radiation sw/lw? > > > > — > > You are receiving this because you were mentioned. > > Reply to this email directly, view it on GitHub > > <#430 (comment) >, > > or unsubscribe > > < https://github.com/notifications/unsubscribe-auth/AG7TW2UQRSEU6E3KRAQGIWDRLTRJZANCNFSM4MD3XKKQ > > > . > > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub < #430 (comment)>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AB5C2RPLQ6IRXXRFS2QGNMLRLTSSTANCNFSM4MD3XKKQ >. > — You are receiving this because your review was requested. Reply to this email directly, view it on GitHub <#430 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AE7WQAQOY22FV2BTVS47Y73RLTS2ZANCNFSM4MD3XKKQ> .

[climbfuji]

iovr changes the algorithm for cloud overlap. icloud changes the algorithm for calculation of partial clouds. I understand that part. Now, regarding the actual radiative transfer model, are there no differences between HWRF and GFS? On Wed, Apr 8, 2020 at 3:18 PM Dom Heinzeller [email protected] wrote:
…

@mzhangw will be able to answer this question.

[mzhangw]

view the difference of radsw_main.f and radsw_main.F90, radlw_main.f and redlw_main.F90. A large part is related to the method to how to generate mcica clouds and feed into radiative transfer model, which I did not observe substantial difference from WRF. But It would be a good idea to consult with developers to confirm it.

…

On Apr 8, 2020, at 3:39 PM, Dom Heinzeller ***@***.***> wrote: iovr changes the algorithm for cloud overlap. icloud changes the algorithm for calculation of partial clouds. I understand that part. Now, regarding the actual radiative transfer model, are there no differences between HWRF and GFS? On Wed, Apr 8, 2020 at 3:18 PM Dom Heinzeller ***@***.*** ***@***.***> wrote: … <x-msg://142/#> @mzhangw <https://github.com/mzhangw> will be able to answer this question. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#430 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AG7TW2RHSNID7AMW44U5RGTRLTVJDANCNFSM4MD3XKKQ>.

[grantfirl]

I had a similar question as Ligia's. I was under the impression that we agreed to make a different version of the RRTMG schemes for the HWRF suite, but this work is merged in with the existing GFS version. Are we 100% certain that the underlying radiative transfer is IDENTICAL between HWRF and GFS? Are there included version numbers or some other clue that this is indeed the case? In other words, if one were to substitute the actual radiative transfer routines from WRF into the GFS and ran with them, would the result be bit-for-bit?

[grantfirl]

I don't see a problem from a coding point of view, so approved. The programmatic question may still be an issue (did we do what we agreed to do?).