A study of dense medium scattering and its applications in sea ice research in ross island, antarctica

This paper summarizes the study of dense medium scattering and its applications towards sea ice research in Ross Island, Antarctica. Based on the Radiative Transfer Theory and earlier work by Tjuatja and Fung, an improvement to the phase matrix to account for the near field effects of electrically dense media was developed, called the Dense Medium Phase and Amplitude Correction Theory (DM PACT). An analysis on the scattering of discrete scatterers in an electrically dense medium was carried out and a backscattering model based on the iterative solution of Radiative Transfer Theory, incorporating both the DMPACT and the IEM model to characterize for rough surface was later developed. With the concern on climate change, the study shifted towards the application of dense medium scattering on sea ice and ice shelf. Using ground truth measurement data collected from Ross Island, Antarctica to compute the backscattering coefficient with the developed forward model, a good matching was obtained between the model predictions and the satellite measurements from RADARSAT. The model was then extended to incorporate multiple layers to take into account the effect of the snow layer on top of the sea ice. A good matching was obtained between the multi-layer model, the Matrix Doubling Method and the CEAREX measurements. From here, the research moved towards the development of inverse models for the purpose of sea ice thickness retrieval due to the parameter's importance in the study of climate change. For active remote sensing, the Radiative Transfer Inverse Scattering Model (RTISM), utilizing the RT-DMPACT forward model and Levenberg-Marquardt Optimization Algorithm was proposed. Initial results with the model showed good matching between measurement data and model predictions. For passive microwave remote sensing, two models were proposed for sea ice thickness retrieval: the training of neural networks using the RT-DMPACT forward model and the use of Genetic Algorithm for sea ice thickness retrieval. Both methods showed promise and the results were encouraging.