Re:

["David L Parkhurst" <dlpark@xxxxxxxx>]

> I'm trying to execute a transport modelling of 1D-flowtube, containing an
ammonium contamination, that is flushed by pristine background water
(contains oxygen, but no ammonium).
Ammonium is retarded by cation exchange, which I can succesfully simulate.
But now I wanted to incorporate kinetic ammonium oxidation (inflowing
solution contains oxygen). I defined rates and kinetics in Phreeqc, but the
simulation allways ends after 7 transport steps.

> Is it possible that this is due to the fact that the time to travel a 1m
cell (=1 shift)  is high (8.3333... days= 720 000s, because the effective
groundwater velocity = 0.11875m/d)?
I tried step_divide > 1 (till 1000), but this didn't seem to help.

> Are these problems typically occurring when dealing with kinetics? have
you
any suggestions to resolve this problem or did I make some error in my
kinetic formulation?

Your formulation looks fine. The problem is in the numerical method and
redox. My guess is that the problem is at the edge of the front where N
concentrations are small. There is no dissolved O2 or H2. The redox is
defined by NO3/NO2/N2, which have concentrations 1e-10 or less. The solver
has a hard time with this, but usually has some strategies to keep going.

There are a few options you can try:

(1) Adding SOLUTION_SPECIES
      H2O + .01e- = H2O-.01
      log_k -9

This gives a little redox buffer under most conditions and makes the solver
happier. It should affect concentrations very little and not more than
1e-9. I think this is the best solution.

(2) Your problem runs on my machine if you use KNOBS; -tol 1e-16. I'm a
little worried that this could cause other numerical problems.

(3) You can delete N2 and NO2- (and N2(g)) from the data base. I think this
will alleviate the redox problem, but you lose some of the nitrogen redox
states. You may consider whether you want to form N2 or not anyway.

Finally, I would prefer the -formula for the KINETICS reation to be
      -formula Amm -1 NH3 +1

The -formula only considers elements added or removed from solution, the
charge on species is completely ignored. There is a tendancy to want to
write some of the aqueous equilibria or electron acceptor into the formula
(including the oxygen in the reaction) but it is not a good idea to force
oxygen to be the electron acceptor. It is better to simply put in the "NH3"
and let thermodynamics decide where it should end up and what should accept
the electrons. The N must end up as N2, NO2, or NO3 because these are the
only N species, and the program will reduce whatever is thermodynamically
favored (O2 in this case).

David


David Parkhurst (dlpark@xxxxxxxx)
U.S. Geological Survey
Box 25046, MS 413
Denver Federal Center
Denver, CO 80225

Project web page: https://wwwbrr.cr.usgs.gov/projects/GWC_coupled