incorrectly formulated salinity transport

[ecoon]

I believe that salinity transport isn't correctly formulated.

The EOS for saltwater defines molar_density_liquid presuming that it is "total mols" not "mols of water". Then, it assumes C is in mols Na / total mols (e.g. mole fraction) and not molar ratio, e.g. mols Na / mols H2O.

The combination of these two make transport well posed -- C * n_liq is in fact equal to mols of Na, so the transport conservation equation is fine (though a user would have to know that the initial and boundary data must be provided in mole fraction, not in mol ratio).

However, this means that water conservation is incorrect.

ATS's Richards equation should conserve mols of H2O. However, water content's evaluator is given by:

WC = sat * porosity * cell_volume * molar_density_liquid.

If molar_density_liquid is "total mols" per volume, then water content is in units of "total mols", not "mols of H2O."

I believe the correct fix is to update the EOS for saltwater to be:

1. MassDensity as is, but changed to expect C to come in as mols NaCL / mols H2O (mole ratio).
2. MolarDensity is just mols H2O / m^3, not total mols / m^3.

[dasvyat]

It should the same units as in any other transport. If we define C as mols component/mols H2O in any other transport we should assume in salt intrusion the same. All previous iterations of trasnsport we assumed that "total mols" = mols H2O

[ecoon]

I think the salinity transport tests are still failing, but now only due to normal differences, not bug differences. @dasvyat , please take a look (on master) and let me know what you think -- we can update the gold files if you are happy.