Processes > Inverse modelling

Inverse modelling using historic, limited data

(1/1)

jmdriscoll:
Hello-

I'm working on a project that involves trying to use PHREEQC's inverse model between two dilute waters in a snowmelt-dominated system. I am attempting to use historic data, which limits available chemical parameters (Ca, Na, Mg, K, Cl, SO4, NO3, Si). I have been having some trouble, and hoping I can get some feedback to help me work out these issues or better define the limitations of using PHREEQC with these data.

I have not included pH as a factor, leaving the default set at 7 for both input solutions, and therefore there is no change of pH between input waters. Will PHREEQC perform a basic linear algebra solution for the addition/subtraction of solutes via weathering reactions? Would omission of pH data influence the phases in the output solutions? Do I need to put a -balance on pH to avoid having it affect my solutions?

I am also having some problems with uncertainty: I have to crank it up to 0.3 to get any solutions, and would, of course, prefer for that to be lower. Could this be due to my dilute values? Or my limited parameters and lack of charge balance?


An example of solutions 1 and 2:

SOLUTION_SPREAD
    -units    mmol/l
     Ca        Mg        Na         K                Cl            N(5)             S(6)        Si   
mMol/l    mMol/l    mMol/l    mMol/l    mMol/l    mMol/l    mMol/l    mMol/l              
 0.0201    0.0042    0.0102    0.0028    0.0022    0.0141    0.0113    0.0148
 0.0318    0.0071    0.0126    0.0048    0.0028     0.014    0.0205    0.0232

dlparkhurst:
The inverse technique automatically includes a charge balance equation, whether you do anything with a pH balance or not. Your solutions have about 25% charge imbalance, which will be taken into account in the models. You will need enough uncertainty to account for this charge imbalance. The concentrations of the solutions will be adjusted to charge balance, and it will show up as deltas in the inverse modeling output.

The deficiency is in anions, whereas, I would have thought the pH would be low, and you were ignoring a significant concentration of H+. Similarly, ignoring ammonium would also lead to a deficiency of cations. So I am not sure what the problem is with the data. Make sure you converted from equivalents to moles correctly. Otherwise, give some thought as to what is missing (or biased) in your analyses that accounts for the charge imbalance.

jmdriscoll:
Thank you for your response, David. Very helpful.

Navigation

[0] Message Index