How reliable are dilution estimates in contaminated aquifers?

In my current monograph, I validated the dilution coefficient using two fully independent approaches:
- chloride concentrations as a conservative tracer
- solution density calculated in PHREEQC (Laliberté–Cooper model)

I then performed a robust statistical comparison of both dilution estimates (n ≈ 2000).
The result: no statistically significant difference between the two methods (Mann–Whitney test), with very close median values and negligible effect size.

❗What does this mean in practice?
It confirms that density-based calculations — often ignored in groundwater contamination studies — provide robust, independent validation of classical tracer-based approaches.
This is especially important for highly mineralized mine waters, where density-driven processes (e.g. buoyancy effects, Rayleigh–Taylor instability) can critically affect plume behavior.