Heterogeneous Initial Conditions

[Mickey Rush]

I am using PFLOTRAN GENERAL mode with an UNSTRUCTURED EXPLICIT grid. 

I need to prescribe spatially-varying initial PRESSURE and TEMPERATURE. 

To do this, I am having to project my initial pressures and temperatures onto a new grid to create a GRIDDED HDF5 file. This step is fairly memory and time-intensive and I'd rather eliminate if possible. It's a little goofy having to interpolate on my end, just so PFLOTRAN can interpolate again. 

Is there a way to simply prescribe initial conditions using a cell-indexed file? When I have tried this, I get the error: 

ERROR: Dataset "ip_dataset" from filename "initial_pressure.h5" of class "dataset_common_hdf5_type" is not supported in "PatchGetCouplerValueFromDataset".

[Hammond, Glenn E]

Mickey,

Unfortunately, there isn’t a straightforward way to use cell-indexed datasets with GENERAL mode. In contrast, the RICHARDS mode does allow for direct input of pressures.

My suggestion is to create an HDF5-formatted checkpoint file and overwrite the necessary values. However, please ensure that the primary dependent variables remain consistent with the system’s state as defined in the checkpoint file—this consistency is critical. What you’re attempting is quite advanced and requires expert-level knowledge.

Glenn

From: 'Mickey Rush' via pflotran-users <pflotra...@googlegroups.com>
Date: Friday, April 10, 2026 at 11:38 AM
To: pflotran-users <pflotra...@googlegroups.com>
Subject: [pflotran-users: 8753] Heterogeneous Initial Conditions

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

--
You received this message because you are subscribed to the Google Groups "pflotran-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pflotran-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/12f1aa89-7f6c-4c67-a638-099583865293n%40googlegroups.com.

[Mickey Rush]

Hi Glenn, 

Thanks for the quick response! I had wondered about trying the checkpoint file approach – I will give that a shot.