Basic Conceptual Model for Artificial Groundwater Recharge

[bbardet]

Hello Everyone,

I am new to PHREEQC and am trying to teach myself the program. I am a Hydrogeologist for a water municipality and have relatively little Geochem experience. I am working on developing a few models that look at the interaction of injected water with native groundwater and native minerology during artificial groundwater recharge. An example of my approach thus far is below:

Code: [Select]

SOLUTION 1 Observed Groundwater
pH 7.85
pe 4
redox pe
temp    15
density     1
Units mg/l
Alkalinity 176.4 mg/l as CaCO3
As 0.006387
Ba 0.070111
Ca 57
Cl 50
F 0.3
Fe 0.467333
K 6.436
Mg 21.744
Na 34.6
S 75.68
Zn 0.044333
Al  0.05
Si 17
-water 1 # kg
END

Solution 2 Injection Water
pH 8.2
Temp 17
Alkalinity 50.3 mg/l as CaCO3
Units mol/l
Cl 0.00026514
S 0.000187126
Ca 0.000441617
Mg 6.16979E-05
Na 0.002005237
K 1.02297E-05
Si 0.000569241
END

EQUILIBRIUM_PHASES 1
Albite -0.63 10
Anorthite -1.69 10
Quartz 0.58 0
Muscovite 9.22 0
Vermiculite-Mg -0.08 10
Microcline 3.10 0
SiO2(a) -0.75 10
Phlogopite -3.52 10
#Cordierite -26.56 10
END

USE EQUILIBRIUM_PHASES 1
USE Solution 1
Save Solution 0
END

Mix 1
0 0.5
2 0.5
Save Solution 3
END

USE Solution 3
USE Equilibrium_Phases 1
SAVE SOLUTION 4
END
I this conceptual model I am only looking at mixing reactions between the two waters and any reactions with the native minerology. In the future I will other potential reactions (surface etc.). My basic workflow so far has been to take measured chemistry from the native groundwater and run it as a lone solution to determine the Saturation Indices of minerals that I have determined to be present in the system via XRD analysis. I then use Equilibrium Phases to add these minerals to the model with the resulting modeled SI values. This keeps the Native Groundwater solution from reacting with the minerology. I then add my Injection Water solution and mix that with the “Native Water” Solution and react in with the equilibrium minerals to create a final solution. In this case I used a simple 1:1 ratio for mixing but obviously this can change. They eventual goal is to bring this basic model into GMS flow model utilizing PHT3D to create a better understanding of what type of mixing/reactions occur during artificial recharge.
Is this approach valid? Please let me know if there are any errors with this approach or if there is a better way to get to my end goal. I appreciate any and all help!

Thank you,
Ben

[dlparkhurst]

I think it is good that you start by running some batch reaction mixing simulations to consider relevant reactions.

You can graduate to 1D transport with PHREEQC, where the dispersivities you set will allow you to mix with the native groundwater as your artificial recharge enters the column of native groundwater. An additional possibility is to reverse the direction of flow to simulate recovery.

You are on your own with GMS and PHT3D, but you might consider Phast4Windows, which has a beautiful interface. I understand the advantages of GMS/PHT3D, but with PHAST you gain support here.

[bbardet]

Thank you, I appreciate the quick response!

I will definitely look into PHAST more. The main reason I was considering the GMS/PHT3d approach was that we already had multiple GMS models built and calibrated for areas we are interested in. It doesn't seem like it would be too much extra work to recreate these models in PHAST however.

As for the batch reactions, does my current approach seem valid or is there something you would change? Typically my starting "known" values are that of the native groundwater chemistry, injection water chemistry and in some case the minerology as obtained by XRD. Using these starting points, is there a workflow that you would recommend? I appreciate your willingness to help me with this, this forum has been by far the best place to learn PHREEQC!

Thanks,

Ben

[dlparkhurst]

Just a few comments:

I would be careful fixing the saturation indices you do. Most of your aluminosilicates are probably not very reactive on the time scale of a storage recovery experiment and most Al analyses are uncertain. I wouldn't rely on specifying the saturation indicies of most of those minerals.

Specifying SI for both quartz and SiO2(a) is redundant.

In most cases, calcite is still the most reactive mineral, and I would be more concerned with it, and possible oxidation reactions with iron if your storage water is oxygenated.

You should define S(6) instead of S in solution, assuming sulfate is what was analyzed.

[bbardet]

Thank you for your comments, they are extremely helpful. A few other basic questions if that's alright:

1.When receiving XRD analysis results back, what is the best way to determine which minerals to be concerned with and which to essentially ignore?

2. Most of our injection water is aerobic and is being injected into an aquifer that is also aerobic. What is the best practice for determining these "other" reactions (surface complexation, oxidation etc) that I should be concerned about? Is this done via trial and error?

3. In general, if you were looking at modeling artificial injection what would your approach be? We are looking more and more at aquifer storage so any advice on approach to these systems would be incredibly helpful.

I know that these are probably basic questions in many senses but my lack of geochem knowledge makes these tougher. Is there any other resources such as books or courses that you recommend I look into to get a better understanding of these processes. I already have "Groundwater, Pollution and Geochemistry" by Appello but would appreciate any other recommendations you may have. We are experiencing more and more situations in which properly learning how to model these geochemical situations would be highly useful and I want to make sure that I start off on the right path. Thanks again for all you do to help novices like me get started, I really appreciate it.

Thanks agian,

Ben

[dlparkhurst]

I can't tell you which reactions are going to be important. It is much clearer once you have some experimental data.

I have participated on a couple ASR projects to develop PHAST models. Here are the references:

Geochemical effects of induced stream-water and artificial recharge on the Equus Beds Aquifer, South-Central Kansas, 1995-2004, 2007, Schmidt, Heather C. Ross, Ziegler, Andrew C., and Parkhurst, David L., US Geological Survey Scientific Investigations Report 2007-5025,  http://pubs.er.usgs.gov/publication/sir20075025.

Hydrologic and geochemical evaluation of aquifer storage recovery in the Santee Limestone/Black Mingo Aquifer, Charleston, South Carolina, 1998-2002, Matthew D. Petkewich, David L. Parkhurst, Kevin J. Conlon, Bruce G. Campbell, June E. Mirecki, 2004, US Geological Survey Scientific Investigations Report 2004-5046,
https://pubs.er.usgs.gov/publication/sir20045046.

[bbardet]

Thank you David,

I will definitely check out those references! I appreciate all your help with this and will continue to work on my model. Thanks again for being so available an willing to answer these questions

-Ben