Concentration.pfsb files report only Zeroes

37 views
Skip to first unread message

Allan Jones

unread,
Sep 2, 2025, 5:40:32 PMSep 2
to ParFlow
Hi everyone,

After completing the short course at MODFLOW & MORE in June 2024, our research group at the Illinois State Water Survey used ParFlow to determine road surface flooding due to rising groundwater elevations. Our research group would like to start using ParFlow for contaminant transport analyses.

We started by creating a very simple contamination model. Currently, we are using ParFlow version 3.13.0 in Docker Desktop. We are creating our preprocess and post processes scripts by using PyCharm.

We are trying to create a fully saturated, simple model that includes contaminant transport. As such, we are using the “Impes” solver. We found a few examples related to contamination model, such as example_single.py.

I have attached our simple model (simple_contamination_model.ipynb). We have run ParFlow successfully. We are getting out.concen.pfsb files and the 'out.log' file indicates convergence. We also recently updated the pftools package to implement the read_pfsb() function, per the ParFlow 3.14.1 release.

Problem:
The model is reporting concentrations of 0.0 for all cells and dump intervals, regardless of the initial concentrations implemented in the model with the <model>.PhaseConcen.<phase>.<contaminant_name>. ... entries.

We have reviewed Parflow release notes and issues looking for solutions or tips. We have checked the <model_name>.out.text file. Unfortunately, ParFlow is not solving our contamination parameters, printing out the following: "Node 0: Initial concentration volume for phase 0, contaminant 0 = 0.000000".

Questions:
  1. Why is our ParFlow model not reporting contamination concentrations other than zero? Any example of ParFlow contamination model would be appreciated!

  2. Does this have anything to do with the ParFlow version? If so, we have been using the Docker image from george135 (recommended in a recent short course). How should we update our ParFlow version in the Docker container?

  3. Any other thoughts on this topic would be greatly appreciated!

Many thanks, in advance, for your help.
- Allan

simple_contamination_model.ipynb

Allan Jones

unread,
Sep 5, 2025, 9:13:48 PMSep 5
to ParFlow
Wow. What timing!? We sent this question Friday last week; the question just populated in the user group now, and we just figured this out.

Solution:
The contamination values are not returned for cells along an edge of the computational domain. As such, the user cannot have a contaminant source region along an computational domain edge, or within a model that has has a computational domain where either NX, NY, or NZ = 1.

This was tricky to figure out. Thankfully, we revisited the example_single.py file and bumped up the Solver.MaxIter to have the model actually converge and we found the area designated as a contaminant source through review of the .pfsb files.

Thanks all,
Allan

Reed M. Maxwell

unread,
Sep 6, 2025, 8:06:06 AMSep 6
to Allan Jones, ParFlow
Allan-  I’m really glad to hear that you are using ParFlow and exploring these capabilities!  Thanks for posting to the list (both your question and solution).  One thing we are reviving is an old reactive transport branch that would do the things you want w/in Solver Richards

We have a little funding to get this merged into the main project branch this year, hopefully we can provide you with some updates along the way.  Thanks again

Reed

Reed M. Maxwell, Ph.D.
William and Edna Macaleer Professor

Department of Civil and Environmental Engineering

High Meadows Environmental Institute

Director, Integrated GroundWater Modeling Center

Princeton University

maxwell.princeton.edu

igwmc.princeton.edu

 

--
You received this message because you are subscribed to the Google Groups "ParFlow" group.
To unsubscribe from this group and stop receiving emails from it, send an email to parflow+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/parflow/504a374d-dcb6-4b30-979c-c947c86edd54n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages