A signal-processing algorithm for the extraction of thin freshwater-ice thickness from short pulse radar data

A radar data processing algorithm is presented which can estimate freshwater ice thicknesses both greater than and less than the normal minimum resolvable thickness. The basis for the algorithm is a model of the geophysical system and the resulting phase signature of the transfer function. The algorithm attempts to isolate this signature and estimate the relevant parameters. In principle, the method can be used to detect layering in other dielectric media. The algorithm was based on a simple model for scattering from a thin-ice layer and refined using radar returns taken over a range of thin-ice thicknesses in a laboratory environment. The method was also applied to data obtained with two different radar frequencies and over a wide range of thickness and surface roughness. The results were compared to thickness estimates from classical deconvolution methods