Help Needed: WRF-Chem SIGSEGV: Segmentation Fault - Invalid memory reference

[Mariya Benny]

Dear Community,

I am encountering an issue with my WRF-Chem model simulation, which halts abruptly after producing output for 2 out of the 7 simulation days. The run stops on the 3rd day at 2 hours and 45 minutes with the error "SEGMENTATION FAULT: INVALID MEMORY REFERENCE." I am using WRF version 4.4.2 with chem_opt=10 (CBMZ Mosaic with aqueous option). The simulation is running on 1 node with 40 cores.

Even when I attempt to restart the run from the 3rd day, it stops again at the same point (2 hours and 45 minutes).

Here is the relevant portion of the rsl.error.0016:

calculate MEGAN emissions at ktau, gmtp, tmidh =         6090   2.00000000       2.74583340
photolysis_driver: called for domain            1
no aerosols initialization yet

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  0x7fc0f7c473ff in ???
#1  0x421ccc5 in ???
#2  0x422076c in ???
#3  0x422f398 in ???
#4  0x2f85915 in ???
#5  0x2f8f0dd in ???
#6  0x24b865b in ???
#7  0x231bf0c in ???
#8  0x49a661 in ???
#9  0x409fb1 in ???
#10  0x4095bc in ???
#11  0x7fc0f7c33554 in ???
#12  0x4095f3 in ???
#13  0xffffffffffffffff in ???

I have attached my namelist.input file, rsl.error.0000, and rsl.out.0000 for reference.

I would greatly appreciate any assistance or suggestions to resolve this issue.

Looking forward to your quick response.

Thank you.

Regards, 

Mariya Benny

M.S Ocean Engineering

Indian Institute of Technology, Madras.

oe24...@smail.iitm.ac.in

[Mariya Benny]

I am also attaching rsl.error.0016 file in which I got the segmentation fault error. 

[Gabriele Pfister]

Dear Mariya, 

we cannot help you with chemistry options other than the suite of MOZART schemes that we implemented. Chem_opt=10 is also a non-KPP option and probably has not been given any attention or been tested in a long time. 

As a general guidance, segmentation faults sometimes can happen when you have wrong values in your input files (e.g. negative emissions). I suggest checking your input files first. If these all look ok, then try running your simulation without any chemistry and adding stepwise chemistry and the different emissions. This might help you pinpoint down the problem. 

Gabi

--

================================
Gabriele Pfister

Atmospheric Chemistry Observations & Modeling (ACOM) Lab

NSF National Center for Atmospheric Research (NCAR)

Email: pfi...@ucar.edu

Work Phone: +1 303 497 2915

Web: https://staff.ucar.edu/users/pfister

[esros kesafos]

What I learnt from WRF Chem is this: your problem is definitely 99.9% related to your namelist
!

To unsubscribe from this group and stop receiving emails from it, send an email to wrf-chem+u...@ucar.edu.

--

€sros

[Mariya Benny]

Hi Gabi and Esros,

Thank you both for your inputs. I am looking into it. I gave a run by reducing my timestep and changed phot_opt from 3 to 2 (TUV to Fast J photolysis). It ran past the time where it stopped last time—but again, stopped on the 4th day at the 12th hour. 

Regards, 

Mariya