Usually treated as independent of concentration. May be affected by velocity, and other physical parameters.

Dear Azade. The dispersion coefficient does not depend on concentration, but on porosity (in case of soil) or on mixing (in case of a water mass). Dispersion and diffusion are expressed on the same unit : length*length/time.

It does occur. I haven't encountered it in my work.

See, for example,

Article A Novel Technique to Determine Concentration-Dependent Solve...

The usual web searches will find more instances.

Diffusion is related to dispersion. See

Technical Report Three Methods for Solution of Concentration Dependent Diffus...

Dear, Olivier Radakovitch

Thank you for your valuable reply.

Dear, Jeremy Dudley

The suggested papers are very useful, thank you for your kindness.

Dear Azade,

The dispersion coefficient is classically defined independent of actual contaminant or solute concentrations. Hence, Fick's Law, which is used to relate the dispersive flux to concentration gradient is a linear one. However several researchers reported that the relationship deviates from linearity under high concentration gradient conditions. I agree with that observation. You can see a list of such reports and proposed alternative models in my paper titled "Multispecies hydrodynamic dispersion under high concentration gradients."