GNGTS 2021 - Atti del 39° Convegno Nazionale

457 GNGTS 2021 S essione 3.2 GROUND PENETRATING RADAR ATTRIBUTE ANALYSIS FOR BEDROCK DETECTION: A CASE STUDY ON THE EASTERN GRAN ZEBRÙ GLACIER (CENTRAL ITALIAN ALPS) I. Santin 1 , E. Forte 1 , M. Guglielmin 2 1 Department of Mathematics and Geosciences, University of Trieste, Trieste, Italy 2 Department of Theoretical and Applied Sciences, Insubria University, Varese, Italy Introduction The Eastern Gran Zebrù (EGZ) glacier is an alpine glacier, located in the Italian Central Alps, Italy. Since 2019, it has been monitored in the frame of the project Glacier CC funded by Stelvio National Park. The monitoring is performed through remote sensing techniques to study its volumetric and areal evolution and with Ground Penetrating Radar survey to estimate its volume and reconstruct the morphology of the underlying bedrock. Geophysical methods are widely applied in glaciology, andespeciallyGPRbecauseof its highefficiency and resolution. Non-invasivegeophysical techniques as GPR can in fact provide a detailed imaging of the internal structure of glaciers and characterize different frozen materials. Providing quantitative indications about glaciers’ total volume requires a detailed reconstruction of the ice bottom morphology ( Colombero et al., 2019 ). Generally, on GPR data, ice is displayed as an electromagnetic (EM) transparent unit and the ice-bedrock contact appears as an high-amplitude reflection due to the high contrast of EM impedance between the ice and the bedrock. However, scattering phenomena and internal reflections are often identified within the ice. These events contribute to the overall attenuation of the EM signal, limiting the penetration depth often concealing the ice-bedrock contact. In many alpine glaciers, the boundary between ice and bedrock is not always clear and easy to identify in the entire GPR dataset ( Forte et al., 2015 ), as happens also in the EGZ glacier. As a matter of fact, a highly scattered area, located especially in the middle of the glacier, hides almost entirely the continuity of the ice-bedrock contact, making challenging and speculative its interpretation (FIG. 1). Such a diffuse scattering area within a glacier can be associated to the presence of liquid water due to a warmer temperature of ice ( Bælum and Benn, 2011 ) or to mixed ice, debris, with a small or even negligible amount of water ( Carturan et al., 2013; Forte et al., 2020) . Without discussing the causes of the concealment of the ice-bedrock contact, in this work we focused on the reconstruction of the bedrock morphology of the EGZ glacier, which assumes first the correct bedrock detection on GPR profiles. Therefore, we provided a multi-attribute analysis to support the ice-bedrock contact interpretation. As final results estimations of total area, volume, and ice thickness of the glacier are given. Methods A total of 8.6 km GPR profiles were collected in 2019 and 2020 with a Malå Geoscience ProEX GPR system, equipped with 250 MHz shielded antennas and dragged manually on the glacier surface. In order to improve imaging and favour the interpretation, the GPR dataset was processed with a standard processing flow including drift removal, bandpass filtering, background removal, exponential amplitude recovery, topographic correction, 2D Stolt f-k migration and depth conversion. Regarding the last two, the velocity analysis was conducted analysing diffraction hyperbolas detectable on GPR profiles and obtaining values close to 17 cm ns -1 with only local variations. Despite few artefacts and very localized overmigration effects, Stolt migration was successful and produced quite realistic results. During data interpretation, the main focus was the identification of the ice-bedrock contact, not always apparent. Therefore, we applied integrated GPR attributes analysis as support for data interpretation. GPR attributes are conceptually similar