[BUG/ISSUE] Seg. fault from gamap_mod.F90 running v13.0.0 with bpch diagnostics

[rgryan92]

GEOS-Chem v13.0.0 runs well with standard netcdf output but crashes with a segmentation fault when compiled with BPCH diagnostics turned on. This is the case using both GEOS-FP and MERRA-2 meteorology. No error reported in the GEOS-Chem log file, segmentation fault reported in the standard output:

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

Backtrace for this error:
#0  0x2af1df8eb3ff in ???
#1  0xdf0f24 in __hco_driver_mod_MOD_hco_final
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/HEMCO/src/Core/hco_driver_mod.F90:344
#2  0x5c669c in __hco_interface_gc_mod_MOD_hcoi_gc_final
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosCore/hco_interface_gc_mod.F90:1082
#3  0x519525 in __emissions_mod_MOD_emissions_final
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosCore/emissions_mod.F90:376
#4  0x45ced5 in cleanup_
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosCore/cleanup.F90:133
#5  0x10dd51b in __error_mod_MOD_geos_chem_stop
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosUtil/error_mod.F90:523
#6  0x10dede1 in __file_mod_MOD_ioerror
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosUtil/file_mod.F90:130
#7  0x9e4085 in write_tinfo
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosCore/gamap_mod.F90:545
#8  0x9e4ebc in create_tinfo
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosCore/gamap_mod.F90:489
#9  0x9e59e5 in __gamap_mod_MOD_do_gamap
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosCore/gamap_mod.F90:146
#10  0x58325e in __gc_environment_mod_MOD_gc_init_extra
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/GeosCore/gc_environment_mod.F90:820
#11  0x405ced in geos_chem
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/Interfaces/GCClassic/main.F90:2316
#12  0x40c57f in main
	at /home/ucfargr/gc_rgr/utnox/v1300_geosfp_spinup_025x03125/CodeDir/src/GEOS-Chem/Interfaces/GCClassic/main.F90:32
/var/opt/sge/node-d00a-041/job_scripts/6632279: line 52: 114800 Segmentation fault      (core dumped) ./gcclassic > log_gc_run_$run_num`

The error was tracked down by @eamarais : missing value in gamap_mod.F90 at line 542 corresponding to the number of molecules of carbon in previous model versions. This was stopping 'tracerinfo.dat' from writing out.

Line 542 changed from:

         ADJUSTL( NAME ), ADJUSTL( FNAME ), MWT, N, SCALE, TRIM( UNIT )

to this (add “, 1”):

         ADJUSTL( NAME ), ADJUSTL( FNAME ), MWT, 1, N, SCALE, TRIM( UNIT )

This workaround solved the problem but the way tracerinfo.dat is written may need updating for v13.0.0
logfiles_ryan.zip.zip

[yantosca]

Thanks for writing @rgryan92. This bug should have already been fixed in commits 1e481bb and 197e26e (dated Jan 15, 2021). Do you have these commits in your src/GEOS-Chem folder?

For 13.0.0 we ran several integration tests for simulations such as TOMAS and RRTMG that turn on the BPCH_DIAG option, and these all passed. But if this is still an issue, please let us know and we can take another look at it.

[rgryan92]

Hi @yantosca , thanks for that, I'll add these commits to my src/GEOS-Chem folder. Cheers

[yantosca]

Thanks @rgryan. I'll close out this issue for now. Feel free to submit a new issue if you are still having problems.