warning: cannot find file noahmptable.tb

[mkavulich]

I am attempting to bring in the latest UFS weather model hash to the UFS SRW application. However, I am getting this error:

 Reading in o3data from global_o3prdlos.f77
       oz_coeff =            6
        latsozp =           71
         levozp =           53
         timeoz =           12
An error occurred in ccpp_physics_init
An error occured in GFS_phys_time_vary_init: warning: cannot find file noahmptable.tb

FATAL from PE    80: Call to CCPP physics_init step failed

An error occurred in ccpp_physics_init
An error occured in GFS_phys_time_vary_init: warning: cannot find file noahmptable.tb

FATAL from PE    78: Call to CCPP physics_init step failed

An error occurred in ccpp_physics_init
An error occured in GFS_phys_time_vary_init: warning: cannot find file noahmptable.tb

FATAL from PE    95: Call to CCPP physics_init step failed

etc. etc.

I have a few questions about this:

1. Why am I receiving this error even when using the FV3_HRRR CCPP suite, which uses RUC lsm, not noahmp?
2. How should I be referencing this file at runtime? As of Land surface upgrades for HR2 #1777, there are multiple copies of noahmptable.tbl, one in ufs-weather-model/tests/parm and the other in the CCPP submodule (ufs-weather-model/FV3/ccpp/physics/physics/). It seems as if documentation was not added as part of this PR. Is this a file I should be linking from one of those locations at runtime? Why is it treated differently from other fix files?

[mkavulich]

@barlage I was told you might be able to offer some insight here. They also suggested Cenlin He but I don't know their github handle.

[barlage]

@mkavulich I would also suggest @HelinWei-NOAA since it's his PR that added this. First, to address some of your questions, the motivation for doing this is that we are transitioning to using a submodule for noahmp and that submodule will also contain the model parameters, see e.g., https://github.com/NCAR/noahmp/tree/master/parameters; part of this PR #1777 is a step in that direction.

If you are looking for an example, I suggest looking in global-workflow here or here, specifically the changes to sorc/link_workflow.sh. Note the file being accessed is FV3/ccpp/physics/physics/noahmptable.tbl.

[mkavulich]

@barlage Thanks for the answer. Since this work is still in flux, it seems like a quick and dirty temporary solution (linking the file from ufs-weather-model/tests/parm at runtime) is the way to go until the work settles down into its final form.

This still leaves me with question 1 unanswered: should this table be linked for all runs, even if we are not using NoahMP? The weather model runs seem to be failing if I don't link the table, but this doesn't make sense to me. Is this a bug, or is the table used for processes other than the NoahMP land surface model?

@HelinWei-NOAA Can you offer any insight?

[barlage]

@mkavulich, as @HelinWei-NOAA says we will fix this read requirement in an future land PR unless we can squeeze it into someone else's PR earlier. To be clear, this file should not need to be linked, but in the current code it does need to be.

I agree with your quick and dirty solution. My only suggestion would be to link FV3/ccpp/physics/physics/noahmptable.tbl. They are technically the same file, but really there should be only one file and it should come from the main tracked ccpp physics version. Like I mentioned earlier, the tracked version will eventually be in the noahmp repo so this ccpp version will be the one moved to the repo.

[HelinWei-NOAA]

@mkavulich This table should not be linked if we are not using NoahMP. Some initializations of NoahMP are still done in GFS_phys_time_vary.fv3.F90. We didn't limit open that table only for NoahMP case. We will fix this issue when we have a chance by moving all NoahMP initializations to noahmpdrv.F90.

[mkavulich]

Thanks @barlage and @HelinWei-NOAA for your feedback. I will implement a temporary solution in the SRW repo linking the table from the CCPP physics submodule, and will look out for upcoming changes to correct the issues mentioned above.