Contrary to assumptions made in physics-based investigations, "pure substance" rock is generally a heterogeneous medium and exists as a conglomerate of arbitrary compositions and structures. A particularly important property of rock is that all of these components are "mixed" in a natural manner such that at any massif point, the composition and structure are identical within the limits of representative volume.
For this rock, the volume should be a representative one, provided its size is much larger than the dimensions of the separate rock components but allowed to remain within massif limits. In this case, we can consider the model of the same rock at the macrolevel and assume that it is uniform at this level.
The macrohomogeneous medium is characterized by its effective EM properties (conductivity, permittivity and magnetic permeability). Passing alternating current through such a medium leads to an accumulation of charges on the borders between its components and phases. For this reason, such media experience induced polarization, and as a result, the conductivity and permittivity become dependent on time and frequency. Similar conclusions can be repeated regarding to magnetic polarization.
Accounting for the medium’s polarization inevitably leads to an essential increase in the number of direct problem parameters. As a result, the EM sounding data inversion can become problematic. Separating the induction and polarizing fields can offer a possible solution for this difficult situation. In this case, previously developed methods of inversion can be used for the induction portion of the field. Recovering polarization portion of the field can give additional information regarding the nature of the polarizable object in question, which is very important for providing an accurate geological interpretation of geophysical data.