Polarization Ray Tracing Calculation of Polarized Bidirectional Reflectance Distribution Function (pBRDF) of Microfaceted Surfaces to Investigate Multiple Reflection Effects

Abstract:

Remote sensing algorithms for aerosol retrieval rely on surface reflectance models for the extraction of path radiance of aerosol scattering in top of atmosphere measurements. A well-defined surface boundary condition is necessary due to the variability in the surface albedo and bidirectional reflectance distribution function. Polarization measurements can help constrain the surface model. Prior work features polarization measurements taken by Jet Propulsion Laboratory’s Ground-based Multiangle SpectroPolarimetric Imager (GroundMSPI). This work has shown that an analytical model that assumes singly reflected light from a rough surface comprised of microfacets sufficiently represents the polarized reflectance of natural surfaces (such as grass), but is less successful for manmade objects. For the linear Stokes parameters (I, Q, U), a single reflection of unpolarized light will result in a null U Stokes parameter relative to the scattering plane. However, some GroundMSPI measurements exhibit a non-zero U Stokes parameter. We show that multiple reflections may be a cause for this discrepancy by using a polarization ray trace (PRT) routine to calculate the polarized Bidirectional Reflectance Distribution Function (pBRDF) for a microfaceted surface. While the effect of multiple reflections, particularly for double reflections, is an order of magnitude smaller compared to single reflections, we show non-zero U Stokes parameters generated from multiple reflections. Furthermore, we have found that for illumination-view geometries with scattering angles less than ~45 degrees, Q and U parameters can have similar magnitude. We report on the magnitude of this effect and compare the PRT simulations to non-zero U measurements from GroundMSPI.