Transport model in PHAST with contaminant sorption

[daliagrig]

Dear developers,
I have developed a model in PHAST for a contaminant transport in groundwater (porous media) and would like to take into account its sorption, however, the result is not as expected.
My media is a box (1000 m x 1 m x 12 m) of soil with different heads at the ends, thus the flow is from left to right. At x=300 m, on the top I define a region (100 m x 1 m x 0 m) where infiltration of contaminant (contaminated water) takes place. I would like to get the concentration of contaminant in water at x=500 m (z=-6 m).
The result is, that the concentration seems to increase or is almost the same for a long time period (1000 y) despite the concentration of contaminant in the infiltrating water decreases significantly already after 100 y (see attachment).

As soil density in PHAST is not required, it seems that the retardation of contaminant due to sorption should be defined in PHREEQC. The structure of the PHREEQC file for my problem is:

SOLUTION_MASTER_SPECIES
[Cont]   [Cont]   0.0   [Cont]   14
SOLUTION_SPECIES
[Cont] = [Cont]
   log_k     0

SURFACE_MASTER_SPECIES
      Surfsand      Surfsand     

SURFACE_SPECIES
Surfsand = Surfsand
    log_k     0
Surfsand + [Cont] = Surfsand[Cont]
    log_k     -98.65 # measured sorption coefficient is 0.0039 l/g or 0.0039*soil_density/porosity = 0.0039*1830/0.32 = 22.3 (kgw/kgw); log_k = log(22.3)-log(1E+100)

SOLUTION_SPREAD #for each time step in PHAST is assigned a different solution No., Initial solution is set to Solution 1, i.e. conc. is 0 mmol/l.
   [Cont]
 0.00E+00
 6.87E-21
 7.07E-17
 1.32E-14
 4.82E-13
 7.21E-12
 6.16E-11
 3.59E-10
 ...
 1.45E-04
 ...
 2.08E+00
 2.38E+00
 ...
 7.70E+00
 ...
 2.13E-01
 ...
 2.67E-05
 ...
 5.93E-08
 1.79E-09
 5.42E-11
 1.63E-12
 4.83E-14
 1.42E-15
 4.16E-17
 1.21E-18
 3.51E-20
 ...
 1.48E-30
     
SURFACE 1 Surfsand
    Surfsand    1e+100
    -no_edl

SELECTED_OUTPUT 1
    -file                 input.dummy.sel
    -high_precision       true
    -reset                false
    -molalities           [Cont] Surfsand[Cont]
   
END

There should be something wrong?

Dalia

[dlparkhurst]

Seems possible. You have accumulated contaminant in the solids and it is slowly being released?

If you change your Kd so that less contaminant is accumulated, you should see the concentrations decrease faster.

[daliagrig]

I expected to model such situation:
Initial conditions: no contaminat in the system.
1st time step: a certain amount of contaminant is added in a cell (through infiltration), partitioning between surface and water takes place according to defined Kd.
2nd time step: a certain amount of contaminant is transported from the cell to adjacent cells and additional amount of contaminant is added in a cell (as a result of infiltration or transport), repartitioning of contaminant on surface and in the water takes place to keep with defined Kd.
etc.
I don't want to accumulate contaminant on the solid, I would like only to keep the right partitioning between surface and water during transport. The problem is that from the results I suppose that the mass of contaminant in the system has increased (the integrated area of the input and results curves are different) and that should not happen...
:( Dalia

[dlparkhurst]

PHREEQC conserves mass.  I don't see how that would be a problem.

Again, I would experiment with the Kd, starting with virtually no sorption. You should see no accumulation on the solid and conservative transport through the system. Increasing Kd will begin to sorb on the solid and will take longer to be removed from the system. A very high Kd will cause the contaminant to sorb and then move extremely slowly.

To say you do not want contaminant to accumulate on the solid makes no sense to me; that is a characteristic of Kd. Check, but I believe R = 1 + Kd is the retardation factor, which give how much longer it takes the center of mass of a sorbing contaminant to move relative to  a conservative constituent.