Jump to content
GMS, SMS, and WMS User Forum
Sean Czarniecki

Changes to MODFLOW-USG?

Recommended Posts

I hope someone can help me out here.  I finished a complex model back in February (I had used GMS 10.3.05).  I have been asked to provide model files to someone else who will be using the most recent version of GMS (currently 10.3.07).  I thought it was a good idea to do a test run using that latest version.  The model didn't work....and the window showing modeling progress was showing some extreme errors that I definitely hadn't seen before.  The model crashed after a few stress periods.

I thought that it might be the software change (I hadn't changed anything else).  Sure enough, I loaded the older version and it worked like it did before.  In the progress window, I can see that the MODFLOW-USG version is different.  10.3.05 used Version 1.3.00 from 2015, whereas 10.3.07 uses Version 1.4.00 from 10/27/2017.  Can anyone tell me what changes occurred that would result in my model not converging any more?  This is bad news to me.....

Share this post


Link to post
Share on other sites

I looked into the version notes on MODFLOW-USG and can't tell if anything they updated would cause my particular model to crash.  It would be sad if they corrected a bug which previously allowed my model to converge and now won't allow it.  Nothing jumps out as fitting their bug corrections.  Any chance that it is something related to merging the newer version into GMS?  I do note that the progress window used to show what unit the CLN Heads were supposed to write to ("CLN HEAD UNIT 767") after every time step.  It no longer does that, but that shouldn't have any impact on the convergence...I'm just noting a difference.

Share this post


Link to post
Share on other sites

I've looked a little further at the version notes and think I see what may be the problem.

GWF2BCF-LPF-U1 -- Removed the minimum saturation value.  Difficult

       problems behave better by allowing saturation to go to zero

       and letting the solver handle it. Revised the way CLN-GW flows

       are written to the budget file.

CLN2BASU1 -- Removed the minimum saturation value.

 

So something that caused problems for other people appears to have been changed…..and this is likely what now doesn’t allow my model to solve properly.  I just ran the model with no wells pumping and it converges.  The model had been crashing when it got to the period of zero recharge, which would be when wells may be in cells which could go dry due to pumping due to lack of water.  This is not good. 

 

If anyone has insight into if I'm interpreting the issue correctly, I would love to hear from you.  To the GMS developers, if you can confirm that there wouldn't be any other reason for my non-convergence (e.g., any chance that it is actually some coding issue related it input files generated for the new MODFLOW-USG?), I would appreciate it.  It would be great if there were an option to still keep the minimum saturation value for some models, but not for others.  Based on the release notes, it doesn't look like that is an option. 

Share this post


Link to post
Share on other sites

I'm running the model with some wells pumping (in areas that shouldn't go dry even in low recharge times).  Seems to work....but I wanted to provide a screen shot of the progress window.   You will see that the error values just don't make sense (in the older version, they were reasonable).....and a weird text item came up during SP 18, TS 1.  Doesn't make me feel too good....

Capture.JPG.db38c2832ae6f0bfa14d6bf86836cba7.JPG

Share this post


Link to post
Share on other sites

Thanks for the suggestion - I may need to go down that path as I try to get through the problem.  This is certainly frustrating, as I've done hundreds of runs which now don't work.  I'm checking to see which well fields are causing the solution to not converge (and am narrowing it down), but the bigger problem is that the solution is now totally different in areas without wells because of what appears to be that change in the minimum saturation value.....

Share this post


Link to post
Share on other sites

From the error numbers it looks like the model is failing to converge and crashing. If you could pass the model on to tech support we could take a look at the model files saved in 10.3.5 and 10.3.7 to make sure there aren't any differences. We could also pass the model on to the model developers to see if they have ideas.

Share this post


Link to post
Share on other sites

Hi Bill - I contacted the MODFLOW-USG developers and have sent them my native text files.  I'll send a set of GMS files to tech support.  Maybe you guys will see something.

Thanks!

Share this post


Link to post
Share on other sites

It may be a long shot as I dont have much details about the model itself, but there might have been some changes in the default values for the solver between MF versions.

Share this post


Link to post
Share on other sites
Quote

Thanks for the thought Michal - The MODFLOW-USG developers changed how the solver handles cell saturation, so in order to continue to use the model I have, it appears that I need to use the previous version.

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×