Bugfix: Address an issue related to initialization of Noah-MP groundwater

[barlage]

TYPE: bug fix

KEYWORDS: Noah-MP, groundwater model

SOURCE: Michael Barlage (NCAR), Gonzalo Miguez-Macho (Spain)

DESCRIPTION OF CHANGES:

Address an issue related to initialization of the groundwater code:

When generic datasets were added in v3.9 there is an internal spin-up required. Previously there was no patch-to-patch communication needed; with the generic datasets, there is communication between patches at initialization. To do this, the groundwater_init subroutine call needed to be pulled out to start_em and also wrapped in a restart flag.

LIST OF MODIFIED FILES:

M Registry/registry.noahmp
M dyn_em/start_em.F
M phys/module_sf_noahmpdrv.F

TESTS CONDUCTED:

1. Regression test with WTF-4.06 (running)
2. Six month May-Oct 4km simulation over the central US

[davegill]

Approved - solves all problems

[davegill]

@barlage @dudhia @weiwangncar
Mike,
This is an EM halo. This needs to be in a EM_CORE ifdef. That means that this code does not work with NMM.
Dave

[davegill]

Mike,
So, can we assume that NoahMP is now EM only? If there is a need for NMM support, then we have some NMM changes (such as that halo) that need to go into their registry.

[davegill]

When the halo is inside of first_rk_step, that is an EM only scenario. Once the halo moves into the physics code, then it becomes visible to EM and NMM.

[barlage]

Dave, When Gonzalo added the groundwater code, I don't think he made it work with NMM, the other halo call (original code) in module_first_rk_step_part1 is: !gmm halo of wtd and riverflow for leafhydro #ifdef DM_PARALLEL IF ( config_flags%sf_surface_physics.eq.NOAHMPSCHEME ) THEN IF ( config_flags%opt_run.eq.5.and.mod(grid%itimestep,grid%STEPWTD).eq.0 ) THEN # include "HALO_EM_HYDRO_NOAHMP.inc" ENDIF ENDIF #endif

…

On Mon, Apr 30, 2018 at 1:04 PM, Dave Gill ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In phys/module_sf_noahmpdrv.F <#463 (comment)>: > NCOUNT=0 + DO NITER=1,500 + +#ifdef DM_PARALLEL +# include "HALO_EM_HYDRO_NOAHMP.inc" @barlage <https://github.com/barlage> @dudhia <https://github.com/dudhia> @weiwangncar <https://github.com/weiwangncar> Mike, This is an EM halo. This needs to be in a EM_CORE ifdef. That means that this code does not work with NMM. Dave — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#463 (review)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AMoMuQ3xxlqHAPlfWaQcXGwoPc91zAs7ks5tt2BCgaJpZM4TnUww> .

[barlage]

Dave - maybe we need to talk about this. If you want to remove the commit from this friendly, it's fine with me. We can get it into the next one.

…

On Mon, Apr 30, 2018 at 1:21 PM, Dave Gill ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In phys/module_sf_noahmpdrv.F <#463 (comment)>: > NCOUNT=0 + DO NITER=1,500 + +#ifdef DM_PARALLEL +# include "HALO_EM_HYDRO_NOAHMP.inc" When the halo is inside of first_rk_step, that is an EM only scenario. Once the halo moves into the physics code, then it becomes visible to EM and NMM. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#463 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AMoMuVwNDLaIyP0-AZwxJ7Cv_W03LW-Oks5tt2QlgaJpZM4TnUww> .

[davegill]

@barlage @dudhia @weiwangncar
Mike,

1. Put a wrapping EM-only ifdef around the em_halo call(s) that are in the physics directory.
2. In check_a_mundo, within a wrapping NMM_CORE ifdef, test for NoahMP + the ground water option. If these are both true, just throw an error that will eventually stop the code.

[barlage]

@davegill I'm compiling NMM now to see if the ifdefs I've added to module_sf_noahmpdrv are sufficient. I've also been laughing hysterically for the last few minutes while addressing your 2. since apparently Noah-MP does not run with NMM according to Line 384 in check_a_mundo. Good to know the time I've spent maintaining the compatibility with NMM over the last few years has been well spent. :)