Determination of precipitation amount - different database issue

[mstec@ichpw.pl]

I've found on this forum simple gypsum precipitation example (topic: http://phreeqcusers.org/index.php/topic,425.msg1259.html#msg1259) and I'm facing now different results when using different databases during simulation.
When using phreeqc.dat default database suphates in solution decrease and precipitated gypsum increases but when using miniteq.v4 database an extremum can be visible on the plots (see attached screenshots). No data in the source code is changed - only database during running the simulation.

I've found that the miniteq database contains additional phases like Halite, which influences the phases equilibrium.

Now my question is, how to decide which simulation is better (true?). Are there any hints?
Marcin

[dlparkhurst]

If you have moles of solutes, you should use pitzer.dat, which has a specific ion interaction aqueous model. Tony Appelo has optimized the parameters for many gypsum/anhydrite systems up to high temperature and pressure.

If you need to use and ion-association model, Tony has done optimization of gypsum systems with phreeqc.dat (and Amm.dat). However, the ion-association approach is limited to lower ionic strength. You can compare between phreeqc.dat and pitzer.dat.

[mstec@ichpw.pl]

Thank You