Improving GPR data processing and interpretation on glaciers through attribute analysis

Ground Penetrating Radar (GPR) is often used to map the internal structure of glaciers. Due to the high resistivity of ice, the signal does not undergo strong attenuation phenomena and provides a representation of the ice body in the form of a radargram where ice is mostly electromagnetic transparent. Nevertheless, while travelling inside the glacier, the electromagnetic wave encounters several materials with different degrees of conductivity which cause the attenuation of the signal; this behavior is represented as a series of hyperbolas and scattering events more or less marked on the GPR display. However, the origin of these patterns is open to interpretation and not always obvious when analyzing the reflection amplitude alone. In order to better investigate the nature of the facies portrayed in the radargrams, we evaluated a series of attributes, which are mostly employed in seismic investigation and GPR surveys in the archaeological domain, in glaciology: more precisely on Indren and Rutor glaciers, located in the Aosta Valley Autonomous Region, Italy. Through the study of the results obtained from the simple GPR assessment, subsequently coupled with the attribute analysis focused on the instantaneous- and frequency-related signal properties, it was possible to better identify fractures and cavities excavated by water, a peculiar thermal effect in the proximity of the bedrock, different types of ice constituting the glacier, and other common features of the glacial exploration otherwise hardly detectable by carrying out a simple GPR survey. This research can be beneficial in identifying potentially dangerous elements, such as fractures and cavities filled with water, forming as a result of a warming climate, threatening the “good health” of the glacier and consequently the surrounding valleys.