Electromagnetic scattering properties in random media and its applications in snow remote sensing

This research investigates the wave propagation and scattering properties in dense random media through analytical analysis and numerical simulations. The applications are focus on the passive and active microwave remote sensing on terrestrial dry snow. In general, the dry snow is considered to be ice grains embedded in the air background and classified as a dense medium due to the appreciable fractional volume of the ice grains occupation. The research starts with the study of the snow structure. Two methods of snow structure generation are demonstrated, the sticky hard sphere model and Bi-continuous medium model. The correlation functions of both models are demonstrated. The hard sphere model is assumed the ice particles as non-penetrable spheres. Through the Monte Carlo simulation, the Numerical Maxwell Method of 3-Dimensional (NMM3D) simulation is performed on the sticky sphere model to explain the large cross-pol backscattering in the field experiments. In order to characterize more complicated snow structure, based on a continuous representation of interfaces between inhomogeneities within the medium, the Bi-continuous method is proposed. The microwave scattering properties are calculated by numerical method which is based on discrete dipole approximation (DDA) accelerated by FFT. For different configuration, the Bi-continuous model can give different frequency dependence, which is consistent with experiment data. Moreover, the numerical results are benchmarked with analytical Born approximation with the low contract among the permittivity and compare well with QCA in the real snow case. The large cross-polarization is also predicted by numerically solving the Maxwell's equations. The Bi-continuous model combined with the dense media radiative transfer theory has been validated through X-band to Ku-band field experiment data. The sensitivity tests of the rough surface effect are also included.;Finally, the retrieval algorithm is developed based on the physical forward model. The premier work on the retrieval algorithm confines in the passive remote sensing. The proposed algorithm retrieves two critical snow parameters, namely, the snow depth and snow grain size from two channels of brightness temperatures (19GHz/37GHz). The results are validated in the Western United states and Alaska.