real.exe segmentation fault 11 - ERA5-pressure-level data

[okurum]

hello all,
I am very new to WRF and trying to use ERA5-pressure-level data to simulate a hurricane event.
Following the tutorial here https://dreambooker.site/2019/10/03/Initializing-the-WRF-model-with-ERA5-pressure-level/ .
I got the following files for pressure level and surf data for the period I am interested in.
ERA5-20180911-20180919-pl.grib
ERA5-20180911-20180919-surf.grib

- geogrid.exe runs fine and I can visualize my nested domains.
- ungrib.exe

below is what I am doing and I wonder if it's the correct way of doing it:

./link_grib.csh /ERA5-20180911-20180919-pl.grib
before running ungrib.exe, I edit the lines:
&ungrib
out_format = 'WPS',
prefix = 'PRES'
./ungrib.exe
rm GRIBFILE.AAA

./link_grib.csh /ERA5-20180911-20180919-surf.grib
before running ungrib.exe, I edit the lines:
&ungrib
out_format = 'WPS',
prefix = 'SURF'
./ungrib.exe

I end up with PRES* and SURF* for all timesteps. Then I run ./metgrid.exe with the &metgrid file prefixes as below (feeding both SURF and PRES file to metgrid):
&metgrid
fg_name = 'SURF','PRES'
io_form_metgrid = 2
/

I get all met_em files (all timesteps for d01 and first timestep for d02), link them to WRF/run folder.

Next, I am running real.exe with:
mpiexec -np 1 ./real.exe

and I get a segmentation fault 11 error - I do not see a spike in RAM usage so I don't think it's a "not enough RAM" issue.

I would really appreciate any advice on how to proceed. I attached bunch of files that can help.
Cheers

Please note that I repeated the same process with:
./link_grib.csh /ERA5-20180911-20180919-
where GRIBFILEAAA and GRIBFILEBBB were created and ran successfully with metgrid.exe - getting the same error with real.exe though.

 

[kwerner]

Hi,
It sounds like you're running the WPS process correctly, but you may not be using all the data you are supposed to. Take a look at this page to learn about using the invariant file.

As for the seg-fault, there are many different reasons that could be happening. Perhaps it could be related to the input data, or it could also be a disk space issue. Check to make sure you have enough space on your disk to write the output from running real.exe.

 

[okurum]

Hi Kelly,

thanks for your answer.

Just to help others who may have a similar problem, here is what I did (Specifically for ppl downloading ERA5 data directly from Copernicus):

1) When downloading the surface level data, do not include the variable "land_sea_mask" - it conflicts with WRF's land sea mask. ERA5 data hosted on NCAR's Research Data Archive (RDA) does not have "land_sea_mask" variable.
2) ./link_grib.csh ./WRF_Data/external/ERA5- (note that this folder has ERA5-xxxxx-surf.grib AND ERA5-xxxxx-pl.grib)
3) ln -sf ungrib/Variable_Tables/Vtable.ECMWF Vtable (Vtable.ECMWF is identical to Vtable.ERA-interim.pl)
4) ungrib.exe
5) metgrid.exe
in the namelist.wps, add:
&metgrid
constants_name = '/path_to/WPS/ERA5_INVARIANT',

6) real.exe then wrf.exe as usual.

Segmentation fault was probably irrelevant. Changed back to GNU compilers and it seems to be working fine so far. I will revisit the intel compilation later.

Cheers

 

[kwerner]

Hi,
Thank you so much for posting a solution to your issue!