Coherency matrix description for statistical linear regression of Stokes parameters in rough small-particle scattering

A statistical analysis of the Stokes parameters in light scattering by randomly rough small particles shows a linear regression law between the squares of the first two components I_s and Q_s of the Stokes vector. While the coefficients of this linear regression contain physical characteristics of the particles, they cannot be directly interpreted in terms of the degree of polarization of the scattered field. We propose an interpretation of this relationship between the Stokes parameters on the basis of the general coherence-density matrix formalism. The link between the statistical regression results and the polarization properties of the stochastic scattered components of the field is established through the coherency matrix elements.