Residuals too large

[Karen Johannesson]

Dear All,

I am trying to figure out why an X1t script I have built will run with an initial SCM but not a nearly identical one, which returns the error  Solving for composition of fluid "inlet". Residuals too large.... The only difference between the two SCMs in that  FeOH_REE_Carrizo_2005.sdat has much lower weak and strong site concentrations than FeOH_REE_Carrizo_2005b.sdat. I think the very low surface site concentrations on the first SCM are probably not correct. Nevertheless, the models runs fine with "*_2005.sdat, but returns the residuals too large error when I use "*_2005b.sdat". I have tried to balance the simulation on Na+, Ca++, and Cl-, and have even #-out the "balance on Cl-" command in the script. I tried to turn off the balance by right clicking on the concentration units as suggested in a GWB forum post I read on line, but that function doesn't work on my computer. I run GWB on a Mac using Parallels, so maybe that is why the right clicking doesn't work. That said, right clicking on any of the other software I use through Parallels (SigmaPlot, etc.) works fine. 

I should mention that this "Chemical_Weathering_5o1_REE.x1t" is one I am testing at the moment, so there are lots of comments and commands I have commented out using the # symbol.  Those will eventually be cleaned up.

Best wishes,

Karen

 

Chemical_Weathering_5o1_REE.x1t thermo_REE_ionx.tdat FeOH_REE_Carrizo_2005.sdat FeOH_REE_Carrizo_2005b.sdat

[Karen Johannesson]

I was hoping someone at the forum might have some suggestions about how I can proceed. I posted this almost a month ago.

[Brian Farrell]

Hi Karen,

Sorry for losing track of this. You posted shortly before we left for the Goldschmidt Conference. I spent some time working with the reactive transport model you attached, but your input file is complicated, so I couldn’t get very far before I left. I lost track of it after returning to the office, and for that I apologize.

I'm taking a closer look and don't have a complete answer yet, but ultimately the 2- and 3-order magnitude differences in site densities assumed in your two datasets are not trivial changes. And these parameters aren't really intended to be adjusted independently of the binding constants. I'm working through this some more, but in the meantime I can answer your question about charge balancing and offer some other suggestions.

The charge balance selection isn't a Mac issue. I’m not sure where you found on the forum that you can right-click to specify charge balance settings, but if it was, that must have been stated in error. In React, you can set the charge balancing ion by left-clicking the unit pulldown for the desired ion and selecting “Balance species”. You can also disable charge balance by selecting “Do not balance”. X1t does not allow charge balancing to be disabled, unlike React. For more information, please see the “balance” command in the appropriate chapter of the GWB Command Reference.

And this won't fix your issue, but I recommend updating your surface datasets to make them more compatible with TEdit, which has many convenient quality control, analysis, and filtering capabilities. To do so, you should set your surface datasets to draw species from your custom thermo dataset. Opening TEdit and moving to the Header pane, find the "Thermo data:" field and click "Browse...", then navigate to your custom thermo dataset, then save the dataset in at least the "jan14" format. For even more functionality, choose the latest format available (i.e. "sep22" if you're using GWB 17).

Or, in a text editor, change the second line of the dataset from "Dataset format: dec99" to "Dataset format: jan14" add a line to the 7th row of the dataset "Thermo dataset: thermo_REE_ionx.tdat". Finally, add an "-end-" line. It should look like:

Dataset of surface reactions for gwb programs
Dataset format: jan14
Surface type: HFO
Model type: two-layer
Surface potential: variable    (specify a value in mV to set a constant potential model)
Surface capacitance: variable  (specify a value in F/m2 to set a constant capacitance model)
Thermo dataset: thermo_REE_ionx.tdat

-end-

Hope this helps,

Brian Farrell
Aqueous Solutions
 

[Karen Johannesson]

Thanks Brian!

I will update my datasets as per your suggestions. I still haven't switched to using the TEdit program. Too old, and too old school, I guess. But I'll look at this tomorrow and see if it makes any difference.

The reason why I tried to change the site densities was that I couldn't reproduce how my PhD student got those values, and I tried using the approach described by Dzombak and Morel (1990) for HFO. The problem is that although there is HFO in this aquifer, the SCM model we developed is for the bulk aquifer and not any specific mineral. So, the Dzombak and Morel approach doesn't exactly work. That said, the surface site densities I computed are of the same order of magnitude as those developed by Dzombak and Morel, whereas ours were so much lower. Hmm.

I will keep my eyes peeled on the forum awaiting any other advise you might have. Tomorrow, I will revisit my calculations and try again to figure out what my student did and how he arrived at those numbers.

Best wishes,

Karen

[Karen Johannesson]

Hi Brian,

I figured out my error, which was in the way I was converting the surface site concentrations from PHREEQC to GWB. For example, the weak site estimate was about 60-fold too high. I will fix this in database and give it another run.

Thanks again!

Karen

[Brian Farrell]

Hi Karen,

I'm glad to hear you discovered that error. I hope the updated database works better for you.

Best,
Brian