Characterization of the subsurface of arctic glaciers using polarimetric satellite SAR data

Abstract

Nowadays, the monitoring of glaciers and ice sheets is of fundamental importance for the scientific community, being these a primary indicators of climate change. The emergence of airborne and satellite remote sensing techniques has led to unprecedented improvements of glaciological observations in term of spatial coverage, temporal and spatial resolution. In this context, Synthetic Aperture Radars (SARs) represent a powerful tool. Thanks to their high resolution, widecoverage, and the capability to operate independently from the sunlight and in nearly all-weather conditions, SARs have become the most used techniques, replacing the more traditional Monitoring methods, such as in-situ measurements. SARs employ microwave frequencies which are able to penetrate into dry snow and ice, potentially allowing to infer Information about subsurface layers. In this thesis, particular attention has been given to olarimetric SAR (PolSAR), which represent an extension of conventional SAR. PolSAR exploits the polarization diversity of electromagnetic waves to gain additional Information about the observed scenario. For instance, considering a glacier scenario, PolSAR measurements are very sensitive to the different type of scatterers found below the ice surface (ice layers, ice lenses, pipes, firn). This makes PolSAR a suitable tool for the retrieval of several glaciological parameters (such as glacier facies extent, firn line, etc.), which cannot be obtained by other remote sensing techniques. Moreover, the deeper penetration and lower sensitivity to surface roughness qualify L-band as the most promising frequency for glacier zones mapping. Therefore, the primary objective of this thesis is to investigate the sensitivity of a set of polarimetric descriptors to glacier zones by using L-band space-borne data. The analysis has the aim of improving glacier facies characterization, which is extremely valuable for supporting mass balance estimation, a key parameter for the monitoring of glacier dynamics. Moreover an existing physical model, based on [16], is adopted and tested for the first time on satellite data, to retrieve large-scale maps of firn thickness. The investigation is based on 2 L-band datasets: the former acquired by the ALOSPALSAR sensor over West Greenland during the winter 2006/2007; the latter obtained by the ALOS-2 PALSAR-2 sensor over the Austfonna ice-cap during the winter 2016/2017. Validation is carried out by comparing the obtained results with available ground penetrating radar measurements, automatic weather station (AWS) data and existing literature. As a main result, this work confirms the potential of polarimetric SAR to identify the different ice zones and the respective subsurface scenarios. As well, the applicability of the adopted model to space-borne data for the firn (thickness) mapping on large scale has been successfully tested over both study areas.