The need for testing various combinations of radiation levels and physics levels

[climbfuji]

The current ccpp-physics code, which inherited the GFS physics from IPD, uses different levels for radiation variables than for physics variables. The radiation variables can also have extra layers on top.

For the example of the UFS host model, the radiation variables are allocated to vertical dimension

adjusted_vertical_layer_dimension_for_radiation = Model%levr + LTP

while the physics variables are allocated to vertical dimension

vertical_layer_dimension = Model%levs

or

vertical_interface_dimension = Model%levs+1

The problem is that not all of these combinations are tested. Specifically, in the UFS, LTP is hardcoded to zero. All configurations except one use Model%levr = Model%levs. The exception is the deep atmosphere / multigases test that uses Model%levr < Model%levs, but this test doesn't run in DEBUG mode as of now.

I heard from @matusmartini that NEPTUNE uses Model%levr = Model%levs and LTP > 0. @matusmartini please confirm.

I think we need to come up with a strategy to test all of the possible combinations or explicitly disallow them using a few checks in the init routines. Since LTP is hardcoded to zero in the UFS, and since testing with UFS is slow and not needed for checking the basic functionality, I would suggest looking for other options (SCM?) to create a complete test coverage of the combinations that we want to support.

[climbfuji]

@matusmartini @grantfirl @ligiabernardet @yangfanglin @SMoorthi-emc @gthompsnWRF FYI

[matusmartini]

Thank you @climbfuji The option to run with LTP=1 was very useful for us earlier this year as it removed substantial cold bias at model top. However, this option, in theory, should not be necessary for our newly-implemented physics grid, which we are using now. It is still necessary for native vertical grid as the vertical spacing greatly varies within spectral element resulting in relatively very thin vertical layer at model top causing large cooling from longwave. It would be nice to keep this capability available, but obviously since we completely switched to using our new vertical grid for physics it is not as important for us as before. I will discuss this with our physics group and can follow up next week.

[yangfanglin]

@climbfuji <https://github.com/climbfuji> We are planning to add the deep atmosphere IDEA physics to CCPP in the commings months. It is important to have the option to run the "traditional" tradition only up to certain layers increase of the entire atmospheric column.

…

On Fri, Nov 19, 2021 at 11:43 AM Matus Martini ***@***.***> wrote: Thank you @climbfuji <https://github.com/climbfuji> The option to run with LTP=1 was very useful for us earlier this year as it removed substantial cold bias at model top. However, this option, in theory, should not be necessary for our newly-implemented physics grid, which we are using now. It is still necessary for native vertical grid as the vertical spacing greatly varies within spectral element resulting in relatively very thin vertical layer at model top causing large cooling from longwave. It would be nice to keep this capability available, but obviously since we completely switched to using our new vertical grid for physics it is not as important for us as before. I will discuss this with our physics group and can follow up next week. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#784 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKY5N2O44WTHW5VRR6QFFMTUMZ5D5ANCNFSM5IMMDU5Q> . 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>.

-- *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/>*