fail to run real with lightning_option

[elka_wrf]

Dear All,
I try to run WRF V4 with lightning options but when i run real.exe it fails.
Please help me how to fix this problem, the error in output execution is :

taskid: 0 hostname: n106
module_io_quilt_old.F 2931 T
------ ERROR while reading namelist physics ------
Maybe here?: lightning_option = 2, 2,
Maybe here?: iccg_method = 1, 1,
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE: LINE: 11618
ERRORS while reading one or more namelists from namelist.input.


I had try a many options but it doesn't work. I run WRF in power AIX cluster

 

[kwerner]

Hi,
I don't really see anything blatantly wrong with your namelist. The only thing I notice is that you have iccg_method in there twice. The first occurrence, you have it set to 1, 1, and then later have it set to 2, 2. Typically this does not cause a problem, and the model will just take the value of the last occurrence, but perhaps there is something in the code that causes this to be problematic. Can you try deleting the one you don't want and see if that makes any difference?

 

[elka_wrf]

Thanks for your reply,

i had remove iccg_method = 2,2,, but i had still the same problem. i had also tried to desactivate wrf_chem=0 and also take other option of lightning_option (1,1, or 11,11) but the same error persist.
the surprsising error is when i take lightning_option = 0, 0, i get the same error :
taskid: 0 hostname: n061
module_io_quilt_old.F 2931 T
------ ERROR while reading namelist physics ------
Maybe here?: lightning_option = 0, 0,
Maybe here?: iccg_method = 1, 1,
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE: LINE: 11618
ERRORS while reading one or more namelists from namelist.input.
-------------------------------------------
Communication statistics of task 0 is associated with task key: 1262403034_0


when i desactivate lightning_option, the WRF run normally.

 

[kwerner]

Hi,
I ran a test using the namelist you originally attached. Even so, I am not able to repeat the error you are seeing. It runs without any problems.
1) Did you make any modifications to the code?
2) Which version of WRF V4 are you using (e.g., V4.1, V4.0.3, etc.)
3) Are you certain that the namelist file you attached is the one the wrf program is using? Can you attach your latest namelist.input file, along with your namelist.output file?
4) Please also attach the full error log.
5) When you say that you disable the lightning option, what exactly do you mean by that? Does that mean you are removing that line from the namelist?

 

[elka_wrf]

Thanks for your reply,

I use WRF V4.0 without any modification in the code. When i use all option of lightning, the real.exe don't run. i had
You find the namelist.input that i use in attached file. for the namelist.output it's empty.
-rw-r--r-- 1 wrfuser001 dsri 0 Dec 31 09:34 namelist.output

for the real.exe the error log is in error.log and out.log.
As i mentionned i run wrf in aix cluster (for real.exe it executed on NBPROC=24)

When i disabled some parameters of the lightning option in namelist the real.exe run normaly but the wrf.exe stop after the first step (00h). the namelist for that is attached (goodnamelist.input and goodnamelist.out.input (name changed juste to uploding it)).

for the last question, when i disable the lightning option the real.exe and wrf.exe run normally that means i dont use the lightning option in the namelist (i removed the line).

juste for information $WRF_CHEM=1 (chem is activated)

Thanks

 

[kwerner]

Hi,

You mentioned in this most recent post that chemistry is turned on. But you mentioned before that you tried turning it off. When you compiled without chemistry, were you still getting this problem? I have been unable to recreate the problem when I test with your namelist. real and wrf run without any problems.

 

[elka_wrf]

Hi

Thanks for reply
I get the same error for both WRF_CHEM=1 or 0.

 

[kwerner]

Okay, thanks. Do you mind attaching a couple of your met_em* files for each domain. Just one or two time periods will do. I want to see if I can repeat the problem with your files. If the files are too large to attach, take a look at the home page of this forum for information on uploading large files. Thanks!

 

[elka_wrf]

Thanks,
i had attached a couple of met_em* files for each domain ( met_em.d01.2019-08-31_15:00:00.nc & met_em.d02.2019-08-31_15:00:00.nc). as they say at home page. if there is any probleme i can upload it again.

Also, i try many others options, you fined attached namelist.input. For this namelist option, the real run successfully ( just for two option of lightning : lightning_option and lightning_dt) if i add any other option of lightning the real stop in the first with namelist.output empty.
but for this namelist.input, when i execute wrf.exe i get this message error (file attached rsl.error.0000.txt) :
"module_lightning_driver.f90", line 490: 1525-002 The I/O statement on the internal file cannot be completed because the end of the internal file was reached. The program will stop.
Communication statistics of task 0 is associated with task key: 2043796658_0

Thanks for all time and help.

 

[kwerner]

Hi,
I've tested both of your namelist (with and without the additional lightning options). They both run real.exe without any problems. I can't test the wrf.exe run, as I only have 1 time period of your met_em* files; however, the original problem, and the problem you need to get past is the real.exe problem. As I'm not able to recreate this problem, using your input files, your namelist, and the same version of WRF you are using, I'm running out of ideas. This may sound crazy, but as a last resort, can you try to recompile V4.0 from scratch in a new directory? Make sure you are getting the pristine WRF-supported version (https://www2.mmm.ucar.edu/wrf/users/download/get_source.html), and not just copying it from a version of your system. Once you get it compiled, please try to run it again, using the same namelist and met_em* files and see if somehow this fixes things - it's unlikely, but it's worth a try.

 

[elka_wrf]

Hi,
Thanks for the reply,
Now i try to compile WRF-4.1.3, and test lightning option.
I want just to tell you, that i had encoutered some problem in compilation and i had add some change (in this version WRF-4.1.3 and also WRF-4 that i had the problem to runing lightning_option). most of the problems are related to the difference betwen compiler XLF and gf90 (principal is in commentaries).

But, i had change a value of a parameter to overcome the problem of compilation.

the error in compilation was :
"module_mosaic2_driver.f90", line 449.50: 1513-041 (S) Arguments of the wrong type were specified for the INTRINSIC procedure "max".
"module_mosaic2_driver.f90", line 2682.13: 1516-023 (S) Subscript is out of bounds.
"module_mosaic2_driver.f90", line 2687.13: 1516-023 (S) Subscript is out of bounds.
"module_mosaic2_driver.f90", line 2692.13: 1516-023 (S) Subscript is out of bounds.
"module_mosaic2_driver.f90", line 2697.13: 1516-023 (S) Subscript is out of bounds.
"module_mosaic2_driver.f90", line 2703.13: 1516-023 (S) Subscript is out of bounds.
"module_mosaic2_driver.f90", line 2704.13: 1516-023 (S) Subscript is out of bounds.
"module_mosaic2_driver.f90", line 2728.13: 1516-023 (S) Subscript is out of bounds.
"module_mosaic2_driver.f90", line 2729.13: 1516-023 (S) Subscript is out of bounds.
** module_mosaic2_driver === End of Compilation 1 ===
1501-511 Compilation failed for file module_mosaic2_driver.f90.


When i search, the problem was about xcut_atype_md1 that is defined in ./chem/module_data_mosaic_asecthp.f90 as :
xcut_atype_md1(0:maxd_atype_md1) with integer, parameter :: maxd_atype_md1 = 1 that means fixed size equal to 1

but in module_mosaic2_driver we had :

xcut_atype_md1(0:2) and xcut_atype_md1(0:3) and xcut_atype_md1(0:4) that's not fixed size.

to reslove that i change maxd_atype_md1=4. I do the same thing for maxd_atype_md2=4

I don't know it that solution will be not correct for the whole model real.exe and wrf.exe. I haven't checked.

 

[kwerner]

Thanks for letting me know. Please let me know when you get through real/wrf to let me know if you are able to run it. Thanks!

 

[elka_wrf]

Hy,

The compilation of WRF-4.1.3 was failed.

I had used the same configure.wrf as in the WRF4.0 (that sucess).

Iget many error that i had solved but the last i don't know what i can do.

you find attached the logfile of compilation and configure.wrf that i use if you have a time to look at it.

I get the real.exe & tc.exe but not the ndown.exe and wrf.exe.

The problem is in frame/module_dm.f90. there is a many change in module_dm.f90 between WRF4.1.3 and WRF4.0 i don't had any idea on how can i solve this.

Any help will be welcome.

Thnaks

 

[kwerner]

There are no actual errors listed in your compile log. The errors we are looking for are spelled with a capital 'E' (Error). This typically happens when you don't send the standard error and output to the log file. You do this using the "&>" in the compile command. Use the following commands from the WRF/ directory:

1) ./clean -a
2) ./configure
3) ./compile em_real >& log.compile

 

[elka_wrf]

Thanks for your reply.
and thank you so much for all

I had recompile the model as you had tell me in your steps.
1) ./clean -a
2) ./configure
3) ./compile em_real >& log.compile

i had the same errors like :

ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_nesting_reset
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_bcast_msgs
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_to_child_info
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_to_child_msg
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_from_parent_info
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_from_parent_msg
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_from_child_info
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_from_child_msg
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_merge_msgs
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_to_parent_info
ld: 0711-317 ERROR: Undefined symbol: .rsl_lite_to_parent_msg
ld: 0711-317 ERROR: Undefined symbol: .__module_input_tracer_NMOD_initialize_tracer
ld: 0711-317 ERROR: Undefined symbol: .__module_mp_morr_two_moment_aero_NMOD_morr_two_moment_init_aero
ld: 0711-317 ERROR: Undefined symbol: .flush
ld: 0711-317 ERROR: Undefined symbol: &&N&module_data_mosaic_asect
ld: 0711-317 ERROR: Undefined symbol: .chem_init
ld: 0711-317 ERROR: Undefined symbol: .__module_gocart_aerosols_NMOD_sum_pm_gocart
ld: 0711-317 ERROR: Undefined symbol: .__module_aerosols_sorgam_NMOD_sum_pm_sorgam
ld: 0711-317 ERROR: Undefined symbol: .__module_aerosols_soa_vbs_NMOD_sum_pm_soa_vbs
ld: 0711-317 ERROR: Undefined symbol: .__module_mosaic_sumpm_NMOD_sum_pm_mosaic
ld: 0711-317 ERROR: Undefined symbol: .__module_mp_morr_two_moment_aero_NMOD_mp_morr_two_moment_aero
ld: 0711-317 ERROR: Undefined symbol: .__module_input_chem_data_NMOD_flow_dep_bdy_chem
ld: 0711-317 ERROR: Undefined symbol: .__module_input_tracer_NMOD_flow_dep_bdy_tracer
ld: 0711-317 ERROR: Undefined symbol: .chem_driver


You can get all the error in the file log.compile.

i compile the WRF by submtting it like a job so the standard error and output is passed to the log file (like compileWRF4.1.eo.1802244) we use LoadLevler on AIX system.

 

[kwerner]

Unfortunately that log.compile file is still not useful. The errors that would help me would be errors with a capital 'E' (Error), and since you run this as a batch script, it's not giving me the right information. While I do see the errors you listed, they may or may not be the original cause of the problem. I would suggest perhaps trying to follow this compile tutorial to get WRF compiled correctly: https://www2.mmm.ucar.edu/wrf/OnLineTutorial/compilation_tutorial.php
Because you are using an AIX system, there may be some slight differences, but you should ensure that you are able to pass the system environment tests, compiler/library tests, and that you have all of the correct libraries/compilers in your paths when you build. You need to make sure that the libraries and NetCDF are all built with the exact compiler (and version) that you are planning to use to build WRF.

 

[elka_wrf]

Hy, Sorry to be so late.
the WRF4.1.3 was compiled.
when i run it with lightning_option :


&physics
mp_physics = 6, 6,
!cu_physics = 95, 95,
ra_lw_physics = 4, 4,
ra_sw_physics = 4, 4,
radt = 10, 10,
sf_sfclay_physics = 1, 1,
sf_surface_physics = 2, 2,
bl_pbl_physics = 1, 1,
bldt = 0, 0,
cu_physics = 5, 0,
cudt = 0, 0,
isfflx = 1,
ifsnow = 0,
icloud = 1,
surface_input_source = 1,
num_soil_layers = 4,
!sf_urban_physics = 0, 0,
do_radar_ref = 1,
lightning_option = 1, 1,

the real.exe run successfully, but for the wrf.exe i got this error :

"module_lightning_driver.f90", line 490: 1525-002 The I/O statement on the internal file cannot be completed because the end of the internal file was reached. The program will stop.
Communication statistics of task 0 is associated with task key: 820495876_0


When i activate other option of lightning :
&physics
mp_physics = 6, 6,
!cu_physics = 95, 95,
ra_lw_physics = 4, 4,
ra_sw_physics = 4, 4,
radt = 10, 10,
sf_sfclay_physics = 1, 1,
sf_surface_physics = 2, 2,
bl_pbl_physics = 1, 1,
bldt = 0, 0,
cu_physics = 5, 0,
cudt = 0, 0,
isfflx = 1,
ifsnow = 0,
icloud = 1,
surface_input_source = 1,
num_soil_layers = 4,
!sf_urban_physics = 0, 0,
do_radar_ref = 1,
lightning_option = 1, 1,
iccg_method = 1, 1,
flashrate_factor = 1.0, 0.9,
cellcount_method = 0, 0,
cldtop_adjustment = 2, 2,
iccg_prescribed_num = 0.,
iccg_prescribed_den = 1.,

I got this error when i execute real.exe :


module_io_quilt_old.F 2931 T
------ ERROR while reading namelist physics ------
Maybe here?: lightning_option = 1, 1,
Maybe here?: iccg_method = 1, 1,
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE: LINE: 11720
ERRORS while reading one or more namelists from namelist.input.

 

[kwerner]

Hi,
I am still unable to repeat your problem. With your exact namelist settings, and your met_em* files, real.exe runs without any problem. For this particular new build (V4.1.3) did you just build basic WRF, or chemistry? Can you attach your file "module_lightning_driver.f90"? Thanks!

 

[elka_wrf]

Hi
I build WRF4.1 with chemistry WRF_CHEM=1.
The program module_lightning_driver.f90 is the same for the version WRF4 and WRF4.1.

So i get the same error.

I want just to clarify that when i use in the namelist.input the only option that works is :
lightning_option = 2, 2,

when i add other optios (like iccg_method = 1, 1, or others) i get the error :

module_io_quilt_old.F 2931 T
------ ERROR while reading namelist physics ------
Maybe here?: lightning_option = 1, 1,
Maybe here?: iccg_method = 1, 1,
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE: LINE: 11720
ERRORS while reading one or more namelists from namelist.input.

 

[zxdawn]

That's related to the update of registry.em_shared_collection.
See: https://github.com/wrf-model/WRF/issues/1021

Current temporary solution is moving these to &chem:

 lightning_start_seconds            = 600,
 iccg_method                        = 1,
 flashrate_factor                   = 1.0,

Hope they will fix it soon.