Coupling Urban Canopy Model as a subLSM for NoahMP.4.0.1 #1329

jyoo8 · 2023-05-01T18:58:08Z

jyoo8
May 1, 2023

Hi,
I have prepared Urban Canopy Model as a subLSM for NoahMP.4.0.1 as Crocus model and Snowmodel in the LIS.
I think I have set up everything for the coupling but I am getting a Segmentation fault error as soon as I execute the LIS retrospective run.

It seems that the error occurs when the LIS runs through LIS_sublsm_FTable.c file.
Specifically, at the function of void FTN(sublsminit)(char *j,int *kk, int len).

The section of the code in the LIS_sublsm_FTable.c file looks as below:

//BOP
// !ROUTINE: sublsminit
// \label{sublsminit}
//
// !INTERFACE:
void FTN(sublsminit)(char *j,int *kk, int len)
//
//
// !DESCRIPTION:
// Invokes the routine from the registry to perform
// land surface model initialization
//
// \begin{description}
// \item[j]
// name of the SUBLSM
// \item[kk]
// index of the SUBLSM
// \end{description}
//EOP
{

struct sublsminitnode* current;
int found ;

current = sublsminit_table;

while(strcmp(current->name,j)!=0){
current = current->next;

if(current==NULL) {
  printf("****************Error****************************\n"); 
  printf("init routine for SUBLSM %s is not defined\n",j); 
  printf("program will seg fault.....\n"); 
  printf("****************Error****************************\n"); 
}

}
current->func(kk);

}

I noticed that "current" in my case is "null" even after the line "current = sublsminit_table;" and the LIS crashed even before going into the while(strcmp(current->name,j)!=0){} section.

I think this is related with pnode in the void FTN(registersublsminit)(char *j, void (func)(int),int len) but my LIS retrospective run didn't go through the void FTN(registersublsminit)(char *j, void (func)(int),int len) section.

Also, I do have LIS_sublsm_pluginMod.F90 for Urban (Canopy Model) under plugins directory as below:

#if ( defined SM_Urban )
call registersublsminit(trim(LIS_UrbanId)//char(0),Urban_ini)
call registersublsmsetup(trim(LIS_UrbanId)//char(0),Urban_setup)
call registersublsmf2t(trim(LIS_UrbanId)//"+"//&
trim(LIS_retroId)//char(0),Urban_f2t)
call registersublsmf2t(trim(LIS_UrbanId)//"+"//&
trim(LIS_agrmetrunId)//char(0),Urban_f2t)
call registersublsmrun(trim(LIS_UrbanId)//char(0),Urban_main)
call registersublsmrestart(trim(LIS_UrbanId)//char(0),Urban_readrst)
call registersublsmdynsetup(trim(LIS_UrbanId)//char(0),Urban_dynsetup)
call registersublsmwrst(trim(LIS_UrbanId)//char(0),Urban_writerst)
call registersublsmfinalize(trim(LIS_UrbanId)//char(0),Urban_finalize)

!wirings between NoahMP and URBAN
call registersublsmsetLSMimport(trim(LIS_UrbanId)//char(0),&
Urban_setLSMimport)

#if ( defined SM_NOAHMP_4_0_1 )
call registerlsm2sublsmgetexport(trim(LIS_noahmp401Id)//"+"//&
trim(LIS_UrbanId)//char(0),NoahMP401_getURBANexport)
call registerlsmsetsublsmimport(trim(LIS_noahmp401Id)//char(0),&
NoahMP401_setURBANimport)
call registersublsm2lsmgetexport(trim(LIS_noahmp401Id)//"+"//&
trim(LIS_UrbanId)//char(0),Urban_getLSMexport)
#endif
#endif

Any help and comment will be highly appreciated.
Thank you.

dmocko · 2023-05-01T19:16:02Z

dmocko
May 1, 2023
Collaborator

Hi @jyoo8

Quick question - did you add the proper lines in your lis/make/default.cfg file to add this subLSM?

1 reply

jyoo8 May 1, 2023
Author

Hi @dmocko,

Here is the modification that I added in the lis/make/default.cfg file

[Urban]
enabled: False
macro: SM_Urban
path: surfacemodels/land/subLSM/Urban,
surfacemodels/land/subLSM/Urban/phys,
surfacemodels/land/subLSM/Urban/cplLSM

Thank you!

dmocko · 2023-05-01T19:25:55Z

dmocko
May 1, 2023
Collaborator

That "enabled" should be set to "True" - or your LISF/lis/make/user.cfg has to have a line: "Urban: On".

1 reply

jyoo8 May 1, 2023
Author

Hi @dmocko,
I made changes in the file. But still I don't get the dependency files generated. Do you know where I should make changes too?
Thank you.

dmocko · 2023-05-01T20:09:23Z

dmocko
May 1, 2023
Collaborator

I assume that you ran "gmake realclean" in your LISF/lis/make directory, and then re-compiled?

5 replies

jyoo8 May 1, 2023
Author

I am compiling the LISF within NU-WRF using the compilation command, "./build.sh -o debug,cleanfirst lis" which should be equivalent to the "gmake realclean", I guess. Thank you.

jyoo8 May 3, 2023
Author

Hi @dmocko,
Actually, "make realclean" was necessary before I continued to compile. Thank you.

Also, I made a few change in the default.cfg file for my Urban subLSM model as below:
[Urban]
enabled: True
macro: SM_Urban
path: surfacemodels/land/subLSM/urban,
surfacemodels/land/subLSM/urban/cplLSM
dependent_comps: NoahMP.4.0.1
NoahMP.4.0.1 path: surfacemodels/land/noahmp.4.0.1,
surfacemodels/land/noahmp.4.0.1/phys

Here in particular, I added lines for "dependent_comps:" and "NoahMP.4.0.1 path:".
I did it because the Urban Canopy Model uses Urban related module_sf_ xxxxxxxx.F90 files that exist under surfacemodels/land/noahmp.4.0.1/phys directory.

I am not sure if these lines are correct to implement as such.
But dependency files for Urban have been generated now.

Regardless, when I compile the LIS again I got error that looks like highly relevant as below:

error: can't open file module_sf_noahmpdrv_401.o for write
compilation aborted for ../surfacemodels/land/noahmp.4.0.1/phys/module_sf_noahmpdrv_401.F90 (code 1)
Makefile:151: recipe for target 'module_sf_noahmpdrv_401.o' failed
make: *** [module_sf_noahmpdrv_401.o] Error 1
make: *** Waiting for unfinished jobs....

Any comments will be highly appreciated.
Thank you!

Currently I am not sure which

jvgeiger May 3, 2023
Maintainer

Hello @jyoo8,

These lines

dependent_comps: NoahMP.4.0.1
NoahMP.4.0.1 path: surfacemodels/land/noahmp.4.0.1,
surfacemodels/land/noahmp.4.0.1/phys

are not needed in your default.cfg file. dependent_comps is used for setting features like irrigation, DA, or OPTUE. The dependency generator will find NoahMP 4.0.1.

Have you tried compiling just LIS - meaning not compiling LIS as part of NU-WRF?

Thanks,
Jim

jyoo8 May 3, 2023
Author

Hi @jvgeiger,
Thank you very much for your notes!
I have not tried compiling just LIS yet but I can do that too.
Thank you.

jyoo8 May 3, 2023
Author

Hi @jvgeiger,
I removed and compiled it again.
Now I am getting error with noahmp.4.0.1/noahmp_driver_401.F90 which has not been modified at all for the Urban model.
Part of error messages looks as shown below:
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(768): error #6633: The type of the actual argument differs from the type of the dummy argument. [ISNOWINOUT]
isnowinout , tvinout , tginout , caniceinout , canliqinout , eahinout , & ! in/out Noah MP only
-------^
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(768): error #6633: The type of the actual argument differs from the type of the dummy argument. [TGINOUT]
isnowinout , tvinout , tginout , caniceinout , canliqinout , eahinout , & ! in/out Noah MP only
-------------------------------------^
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(774): error #6633: The type of the actual argument differs from the type of the dummy argument. [PGSINOUT]
pgsinout , & ! in/out Noah MP only
-------^
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(777): error #6633: The type of the actual argument differs from the type of the dummy argument. [T2MVOUT]
t2mvout , t2mbout , q2mvout , q2mbout , relsmcout, & ! out Noah MP only
-------^
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(786): error #6633: The type of the actual argument differs from the type of the dummy argument. [PARAMETERS]
parameters, &
-------^
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(787): error #6633: The type of the actual argument differs from the type of the dummy argument. [FLDSTOIN]
rivstoin,fldstoin,fldfrcin, &
----------------^
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(791): error #6633: The type of the actual argument differs from the type of the dummy argument. [IDS]
ids,ide, jds,jde, kds,kde, &
-------^
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(791): error #6633: The type of the actual argument differs from the type of the dummy argument. [IDE]
ids,ide, jds,jde, kds,kde, &
-----------^
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(747): error #6631: A non-optional actual argument must be present when invoking a procedure with an explicit interface. [KTS]
call noahmplsm_401 (LIS_rc%udef, & ! in : LIS undefined value (David Mocko)

......

compilation aborted for ../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90 (code 1)
Makefile:151: recipe for target 'noahmp_driver_401.o' failed
make: *** [noahmp_driver_401.o] Error 1
make: *** Waiting for unfinished jobs....

jvgeiger · 2023-05-05T18:41:09Z

jvgeiger
May 5, 2023
Maintainer

Where is your working directory on discover? I will take a closer look.

1 reply

jyoo8 May 5, 2023
Author

Hi @jvgeiger,
I created a git repo at /discover/nobackup/jyoo8/NU-WRF_dev/urban_cloned .
Thank you very much for your time!

jvgeiger · 2023-05-05T21:53:57Z

jvgeiger
May 5, 2023
Maintainer

Hello @jyoo8,

The code that you gave me was missing the [Urban] section in the make/default.cfg file. I added it. Note that you may have a typo. You specified

path: surfacemodels/land/subLSM/Urban,

but the path is

surfacemodels/land/subLSM/urban

Uppercase U vs lower case u. I corrected the default.cfg file.

I get this error:

surfacemodels/land/subLSM/urban/Urban_driver.F90(853): error #6784: The number of actual arguments cannot be greater than the number of dummy arguments.   [NOAHMP_URBAN]

You are calling noahmp_urban with too many variables.

3 replies

jyoo8 May 5, 2023
Author

Hi @jvgeiger,
Thank you very much for your efforts, Jim.
I was able to reproduce the same error now with the newly cloned repository at /discover/nobackup/jyoo8/NU-WRF_dev/urban_cloned/nu-wrf-dev.

Interestingly, the original version of the cloned repository produced the previous error still:
-I../surfacemodels/land/rdhm.3.5.6/ -I../surfacemodels/land/rdhm.3.5.6/sac/ -I../surfacemodels/land/rdhm.3.5.6/frz/ -I../surfacemodels/land/rdhm.3.5.6/snow17/ -I../surfacemodels/land/rdhm.3.5.6/sachtet/ ../surfacemodels/glacier/noahmp.3.9.1.1/noahmpglacier3911_readcrd.F90
ifort: command line warning #10434: option '-nomixed_str_len_arg' use with underscore is deprecated; use '-nomixed-str-len-arg' instead
../surfacemodels/land/noahmp.4.0.1/noahmp_driver_401.F90(768): error #6633: The type of the actual argument differs from the type of the dummy argument. [ISNOWINOUT]
isnowinout , tvinout , tginout , caniceinout , canliqinout , eahinout , & ! in/out Noah MP only
-------^

Also, there was not the section for the Urban in the make/default.cfg in the the newly cloned repository!
Strange.

It seems that the original repository at /discover/nobackup/jyoo8/NU-WRF_dev/nu-wrf_urban202304/nu-wrf-dev may have some mysterious issue, which cannot be resolved easily.

Anyway, let me fix the issue in the Urban_driver.F90 for now.
Thank you very much for your time.
Have a nice weekend!

jyoo8 May 19, 2023
Author

Hi @jvgeiger,
I was able to compile the LIS along with Urban Canopy Model and I am testing the code now.
Currently running on the retrospective mode and it ran through reading in config file of the Urban Canopy Model section successfully.
Now I am getting error with ESMF_StateAdd for an exchange variable, lhField.
Interestingly, there were three other exchange variables before the lhField in which ESMF_FieldCreate and ESMF_StateAdd have been completed successfully.
I don't know why the lhField didn't pass the section.
Do you have any idea on where I need to make sure?
Thank you very much for your valuable inputs!

jyoo8 May 19, 2023
Author

Hi @jvgeiger,
I figured out that special characters (e.g., /, &, etc. ) cannot be used in the name parameters of ESMF_FieldCreate as well as duplications of the names among variables.
I was able to pass the section.
Thanks!

jvgeiger · 2023-06-12T20:06:31Z

jvgeiger
Jun 12, 2023
Maintainer

Great.

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Coupling Urban Canopy Model as a subLSM for NoahMP.4.0.1 #1329

{{title}}

Replies: 6 comments 11 replies

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

Select a reply

Coupling Urban Canopy Model as a subLSM for NoahMP.4.0.1 #1329

jyoo8 May 1, 2023

Replies: 6 comments · 11 replies

dmocko May 1, 2023 Collaborator

jyoo8 May 1, 2023 Author

dmocko May 1, 2023 Collaborator

jyoo8 May 1, 2023 Author

dmocko May 1, 2023 Collaborator

jyoo8 May 1, 2023 Author

jyoo8 May 3, 2023 Author

jvgeiger May 3, 2023 Maintainer

jyoo8 May 3, 2023 Author

jyoo8 May 3, 2023 Author

jvgeiger May 5, 2023 Maintainer

jyoo8 May 5, 2023 Author

jvgeiger May 5, 2023 Maintainer

jyoo8 May 5, 2023 Author

jyoo8 May 19, 2023 Author

jyoo8 May 19, 2023 Author

jvgeiger Jun 12, 2023 Maintainer

jyoo8
May 1, 2023

Replies: 6 comments 11 replies

dmocko
May 1, 2023
Collaborator

jyoo8 May 1, 2023
Author

dmocko
May 1, 2023
Collaborator

jyoo8 May 1, 2023
Author

dmocko
May 1, 2023
Collaborator

jyoo8 May 1, 2023
Author

jyoo8 May 3, 2023
Author

jvgeiger May 3, 2023
Maintainer

jyoo8 May 3, 2023
Author

jyoo8 May 3, 2023
Author

jvgeiger
May 5, 2023
Maintainer

jyoo8 May 5, 2023
Author

jvgeiger
May 5, 2023
Maintainer

jyoo8 May 5, 2023
Author

jyoo8 May 19, 2023
Author

jyoo8 May 19, 2023
Author

jvgeiger
Jun 12, 2023
Maintainer