Merged requested to merge dispersion_bycomp into master
It will be useful to allow different dispersion parameters for different components, even when they are in one phase. For instance, single-well chemical tracer tests are usually modeled in one phase (oil is immobile), but dispersion coefficients for different components are different.
See e.g. this article: "Dispersivity is taken to be a history match parameter and can be different for each tracer".
Fixes #1104 (closed)