SCO2 + GIRT + SOURCE_SINK causing SIGABRT at global_aux.F90:383

[Ahmed, Syed H]

Hello PFLOTRAN developers,

I am a graduate student working on a reservoir-scale model of CO2 injection into formation brines for lithium recovery (advisor: Dr. Kyung Jae Lee). I am running PFLOTRAN development version on a university HPC cluster (Carya).

I am encountering a consistent SIGABRT crash whenever I combine SCO2 flow mode + GIRT reactive transport + SOURCE_SINK with any rate-based condition. The crash occurs at cleanup/deallocation after the simulation fails, not during initialization.

Full backtrace:

  #8   utility.F90:2038

  #9   global_aux_module_MOD_globalauxvarstrip       at global_aux.F90:383

  #10  global_aux_module_MOD_globalauxvararraydestroy at global_aux.F90:358

  #11  global_aux_module_MOD_globalauxdestroy         at global_aux.F90:426

  #12  auxiliary_module_MOD_auxdestroy                at auxiliary.F90:108

  #13  patch_module_MOD_patchdestroy                  at patch.F90:11875

Setup:

- MODE SCO2 with ISOTHERMAL_TEMPERATURE 50.0 C

- MODE GIRT reactive transport

- 1D column, 100 cells, 1m each

- Mineral kinetics: Calcite and Magnesite precipitation

- SOURCE_SINK with SCALED_MASS_RATE NEIGHBOR_PERM (also tried MASS_RATE)

- Single Dirichlet pressure BC at east face

- Chemistry: CO2(aq), HCO3-, CO3--, H+, Mg++, Ca++, Na+, Cl-, Li+

The same simulation runs successfully to 50 years with zero failures when SOURCE_SINK is removed and replaced with Dirichlet pressure BCs at both faces. So the SCO2+GIRT coupling and chemistry are correct — the issue is specifically triggered by the presence of SOURCE_SINK.

I have also tried:

- MASS_RATE

- SCALED_MASS_RATE VOLUME

- SCALED_MASS_RATE NEIGHBOR_PERM

- Moving SOURCE_SINK regions away from BC faces

- OSRT instead of GIRT transport mode

All produce the same crash at global_aux.F90:383.

Michael Hammond suggested this may be a transport constraint initialization issue and recommended reporting here for better error messaging. He also pointed me to the SCO2 paper input deck on Zenodo (doi: 10.5281/zenodo, Hammond et al. 2024) but I wanted to report this issue as it seems like the PETSc errors should produce more descriptive messaging.

I am attaching my input deck. Any help or workaround would be greatly appreciated.

Thank you,

Syed Hassan Ahmed

Graduate Student, University

Advisor: Dr. Kyung Jae Lee

[Hammond, Glenn E]

Syed,

Thank you for bringing this issue to my attention. I have been able to reproduce the problem and observe that rt_auxvar%aqueous%dtotal is null when accessed, which appears to be the root cause. However, I am still investigating why it is not being allocated as expected. I will continue working on a solution.

Glenn

From: pflotr...@googlegroups.com <pflotr...@googlegroups.com> on behalf of Ahmed, Syed H <sahm...@CougarNet.UH.EDU>
Date: Monday, March 23, 2026 at 12:47 PM
To: pflotr...@googlegroups.com <pflotr...@googlegroups.com>
Subject: [pflotran-dev: 6388] SCO2 + GIRT + SOURCE_SINK causing SIGABRT at global_aux.F90:383

You don't often get email from sahm...@cougarnet.uh.edu. Learn why this is important

Check twice before you click! This email originated from outside PNNL.

--
You received this message because you are subscribed to the Google Groups "pflotran-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pflotran-dev...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pflotran-dev/SA3PR06MB9540ADD9D50C8A66B92D24F7F64BA%40SA3PR06MB9540.namprd06.prod.outlook.com.