SIGSEGV, segmentation fault occurred

Ulysse Pasquier

unread,

Oct 15, 2020, 12:31:07 AM10/15/20

to wrf-hydro_users

Dear all,

I am facing an issue running WRF-Hydro in standalone mode. I am getting a "SIGSEGV, segmentation fault occurred" which I understand from the FAQ on the website is a common error.

Keeping my run as simple as possible (turning off all hydro outputs) I have tried different domain configurations, for example:

- 300m geogrid, 300m terrain routing. DXRT = 300. AGGFACTRT = 1.

- 300m geogrid, 30m terrain routing. DXRT = 30. AGGFACTRT = 10.

After visually exploring the geogrid domain and outputs from the GIS Pre-Processing tool, it seems the two are perfectly nested in each case.

I should note that I don't believe this to be a permission or memory issue as the Standalone Croton New York v5.1.1 Test Case runs without a problem.

I would appreciate any suggestions!

Many thanks,

Ulysse Pasquier

Katelyn FitzGerald

unread,

Oct 15, 2020, 3:59:20 PM10/15/20

to wrf-hydro_users, Ulysse Pasquier

Hi Ulysse,

Is your domain substantially larger than the Croton example case? If so and depending on the specs of your machine, memory could potentially still be an issue.

It'd also be helpful to know a bit more about where the model failed (e.g. if anything was written to the diag* or log files).

Ulysse Pasquier

unread,

Oct 15, 2020, 7:40:22 PM10/15/20

to wrf-hydro_users, Katelyn FitzGerald, Ulysse Pasquier

Hi Katelyn,

Thanks a lot for your response. My domain is indeed larger: 99x99 at 300m resolution, compared to the Croton's case 15x16 at 1km.

I've just tried to match exactly the example case's domain configuration (same size, with 1km resolution geogrid and 250m terrain routing) and still receive the same error (with DXRT = 250.0 and AGGFACTRT = 4):

wrfinput_flnm: './DOMAIN/wrfinput_d01.nc'

xstart,xend, ystart, yend 1 15 1 16

wrfinput_flnm: './DOMAIN/wrfinput_d01.nc'

xstart,xend, ystart, yend 1 15 1 16

forrtl: severe (174): SIGSEGV, segmentation fault occurred

Image PC Routine Line Source

wrf_hydro_NoahMP. 00000000202C8433 for__signal_handl Unknown Unknown

libpthread-2.26.s 00002AAAACBAD2D0 Unknown Unknown Unknown

libnetcdf_intel.s 00002AAAAB1973A6 ncuriparse Unknown Unknown

libnetcdf_intel.s 00002AAAAB195379 NC_urlmodel Unknown Unknown

libnetcdf_intel.s 00002AAAAB18A7C9 NC_open Unknown Unknown

libnetcdf_intel.s 00002AAAAB18A6F7 nc_open Unknown Unknown

libnetcdff_intel. 00002AAAAACE6896 nf_open_ Unknown Unknown

libnetcdff_intel. 00002AAAAAD1C359 netcdf_mp_nf90_op Unknown Unknown

wrf_hydro_NoahMP. 00000000200127EF Unknown Unknown Unknown

wrf_hydro_NoahMP. 00000000200B0E6E Unknown Unknown Unknown

wrf_hydro_NoahMP. 000000002000E2C6 Unknown Unknown Unknown

wrf_hydro_NoahMP. 000000002000E252 Unknown Unknown Unknown

libc-2.26.so 00002AAAACDDD34A __libc_start_main Unknown Unknown

wrf_hydro_NoahMP. 000000002000E16A Unknown Unknown Unknown

The model seems to fail very early as the diag* file is not created. I have compiled WRF-Hydro with export HYDRO_D=1 and do get a diag file with the example case.

I have tried to compare the netcdf files (for example with ncdump -h) for my run and that of the example case but am struggling to find a difference that may be causing this.

I am using:

- GIS pre-processor version 5.1.1 to create the terrain routing input files

- the create_wrfinput.R script for wrfinput

Thanks again,

Ulysse

Katelyn FitzGerald

unread,

Oct 15, 2020, 9:31:23 PM10/15/20

to wrf-hydro_users, Ulysse Pasquier, Katelyn FitzGerald

It looks like it's failing in the initialization.

If you haven't already, I would try a run without spatial_soil active (not sure if you're using this) and with idealized forcing data just to eliminate some potential points of failure.

Ulysse Pasquier

unread,

Oct 16, 2020, 1:51:36 AM10/16/20

to Katelyn FitzGerald, wrf-hydro_users

Hi Katelyn,

Removing spatial_soil did the trick! I didn't realize it couldn't be on while no spatially distributed parameters were provided.

Thank you very much for your help,

Ulysse

catalina andruchow colombo

unread,

Mar 2, 2023, 4:38:43 PM3/2/23

to wrf-hydro_users, Ulysse Pasquier, wrf-hydro_users, Katelyn FitzGerald

Hi everyone,

I am Catalina Andruchow Colombo and I'm doing my bachelor thesis using the WRF-Hydro model version 5.2.0 in uncoupled mode. Now I´m trying to run the simulation with the example test case in the Croton river but I got the same error of Ulysse. I first tried tuning off the spatial_soil parameter and got the same problem, then I kept the spatial_soil=0 but added the idealized forcing data and still got the same error. Also I modified the option of HYDRO_D in the setEnvar.sh file and again got the same error.

I have no idea what caused that problem, so any suggestion will be appreciated.

Best regards,

Catalina.

Orzu Odinaev

unread,

Mar 4, 2023, 2:51:12 AM3/4/23

to wrf-hydro_users, catalina andruchow colombo, Ulysse Pasquier, wrf-hydro_users, Katelyn FitzGerald

Hey Catelina,

I faced the same error and for my case the problem was with accessing a memory location that it wasn't allowed to or had limited access. So what I did is I added to my bashrc file (~/.bashrc) the line ulimit -s unlimited, then source your bashrc (source ~/.bashrc) so that it rereads the functions in your environment and then run the model.

Hope it helped,

Orzu

catalina andruchow colombo

unread,

Mar 4, 2023, 2:52:03 PM3/4/23

to Orzu Odinaev, wrf-hydro_users, Ulysse Pasquier, Katelyn FitzGerald

Hi Orzu,

Thanks for your reply

I just tried that with the bashrc file but I still got the same error.

Best regards,

Catalina

Orzu Odinaev

unread,

Mar 6, 2023, 12:36:58 AM3/6/23

to wrf-hydro_users, catalina andruchow colombo, wrf-hydro_users, Ulysse Pasquier, Katelyn FitzGerald, Orzu Odinaev

Hey Cataline,

seems like you didn't properly create your input files. Please refer to the WRF-Hydro FAQs (https://ral.ucar.edu/projects/wrf_hydro/faqs) where you can find more information on that issue.