PhreeqcRM and version 3.2.0

[dlparkhurst]

Version 3.2.0 of PHREEQC programs--PHREEQC, PhreeqcI, IPhreeqc, and PhreeqcRM--have been released at http://wwwbrr.cr.usgs.gov/projects/GWC_coupled/phreeqc.

PhreeqcRM is new software to be used as the reaction engine for reactive transport modeling. A manuscript has been accepted by Advances in Water Resources (http://dx.doi.org/10.1016/j.advwatres.2015.06.001) that describes the use of the reaction module and its application. The module can be used with transport codes written in C, Fortran, or C++. Distribution files for compilation of the module library for Windows and Linux are at the web site. Details are in the PhreeqcRM README file at the web site.

Several of the PHREEQC databases have been updated. Tony Appelo has added silica to pitzer.dat; a new version of sit.dat is included; and a new low temperature (subzero) database, frezchem.dat, has been contributed by Jonathon Toner.

The numerical methods for SIT and Pitzer calculations has been improved by simplify many of the loops used in the calculatons, and by optimization of the calculation of the nonsymmetric mixing parameters, with thanks to Wouter Falkena and coworkers on the MoReS team.

The implementation of the IPhreeqc module for Fortran has changed. IPhreeqc is now a Fortran "module", which means it can be invoked with a USE statement (USE IPhreeqc). The switch to the new usage requires removing former include statements, adding the use statements, and including a new Fortran file in your project. Details are in the IPhreeqc README file at the website.

Additional bug fixes are described in the Release Notes on the web site.

[Pak]

If it is possible I would like to ask some questions about the approach for each reaction cell used in PhreeqcRM.

lc = S x n x RV

where lc is the water content, S is the liquid saturation, n porosity and RV is the representative volume.

From the PhreeqcRM article it is possible to transfer the following Data (If I am not wrong):

1)   From Transport to PhreeqcRM: SetTemperature, SetPressure, SetSaturation, SetPorosity, SetDensity.

2)   From PhreeqcRM to transport: GetSolutionVolume, GetDensity, and GetSaturation.

Therefore, I assume that the volumetric water content from transport to reactions can be modified by variations in saturation, and saturation from reaction to transport. My question would be:

Is the representative volume, which is chosen by (or default 1) the user, supposed to be constant during the whole simulation?

Different Concentration units are presented in PhreeqcRM, one of them is mol/l. If iPhreeqc gives as output mol/kgw, the way to get the term mol/l would be to multiply by the mass of water and divide by the solution volume obtained from the Phreeqc output?


Thank you

[dlparkhurst]

The representative volume is assumed constant for the whole simulation.

PhreeqcRM converts to mol/L as described in the documentation of the PhreeqcRM methods (ftp://brrftp.cr.usgs.gov/pub/charlton/phreeqcrm/html/index.html):
"To convert from moles of element in a cell to mol/L, the number of moles of an element is divided by the solution volume resulting in mol/L."

If you are using PHREEQC Basic functions, you could either use TOT("Na")*TOT("water")/SOLN_VOL or TOTMOL("Na")/SOLN_VOL.

[Pak]

Conceptually, is the representative volume an interface between the void volume of the transport simulator and the batch reaction of phreeqc?

[dlparkhurst]

The representative volume for PhreeqcRM is a volume of porous medium, one liter by default. It includes rock, water, and possibly unsaturated void volume.

[Pak]

I would like a naive question about the transport of water as a component. In PhreeqcRM you have the possibility to transport water as a component, I wonder if it should be transport as a component, namely taking into account diffusion (relative velocity) or has a phase where the diffusion does not exist.

[dlparkhurst]

The most accurate approach to modeling diffusion requires consideration of the individual aqueous species and allows the species to diffuse at different rates while maintaining charge balance. PhreeqcRM has an option to retrieve the concentrations and activity coefficients for individual species. I think logically, water should diffuse along its chemical potential gradient.

If you are using total concentrations and H2O for transport, you are really modeling dispersion, which attempts to account for the many different flow paths used by the fluid moving through the system. All aqueous species have the same diffusion coefficient, which usually becomes a small contribution to the dispersion coefficient (tensor).

People will argue, probably correctly, that waters do not actually mix in a flow system; water moves in many flow lines, but the only mixing is from diffusion.  Mixing actually occurs only when you sample. Dispersion is simply a convenient representation for spreading concentration fronts. Also, check out the papers of Berkowitz, who proposes a different approach to the dispersive process.