GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 3.2 711 of profiles were selected for the interpretation, from a total survey length of 88 km. The EM velocity was estimated both for the snow cover and for the ice thickness: it was equal to 21.2 cm/ns and to 17 cm/ns, respectively. After the depth conversion, the glacier bed surface was reconstructed in both datasets (Fig. 1). In order to create a regular grid and preserve the original thickness values at their own location, a Kriging algorithm was applied and the same glacier border was set by joining information from Crepaz et al., 2013 and Carton et al., 2017. For the 2015 dataset the snow-ice horizon was also reconstructed thanks to the overall high quality and resolution of data, which allowed to clearly interpret and pick the boundaries between frozen units along all the selected profiles. Results and discussions. By picked data interpolation and gridding, we obtained 2004 and 2014 ice thickness maps (Fig. 2). The different spatial density of the two datasets forced us to exclude some areas from the comparison. The higher 2015 data coverage allowed to better reconstruct the ice thickness, avoiding unrealistic values as obtained in the X-area in 2004 map (Fig. 2a). We further estimated the mean ice thickness in 2004, which was equal to 18.0 m. In 2014 we inferred an average ice thickness equal to 12.9 m, appointing to a general ice thickness decrease equal, on average, to about 5 m. In the northern (lowermost) area the ice thickness variation determined the bedrock outcropping (black dashed zone in Fig. 2b). Such result agrees with independent conclusions inferred from aerial photographs by Carton et al., 2017. The maximum ice thickness decrease, close to about 25 m, is observed in the eastern part of the glacier and in a small area towards west (Fig. 2c). By combining the ice thickness maps with the interpreted glacier bed, we calculated the whole ice volume and the total amount of water stored inside the glacier (i.e. water equivalent - w.e.) in 2004 and 2014. A remarkable change of the area and, even more relevant, a change of the glacier volume, were observed. While the area covered by ice declined by about 22%, the volume, and the w.e., decreased by about 30% (from 16.2 m to 11.5 m). As a matter of fact, by comparing Fig. 2a and 2b, it can be noted that while in 2004 the Marmolada Glacier was mainly a single ice body, in 2014 its fragmentation into three almost separated ice bodies was evident, as highlighted in Fig. 2b. Indeed, during the 10-year analysed period, the glacier was interested by large changes in morphology and hypsometry, which, in function of the climate pattern of the area, proved the split of the Marmolada Glacier in smaller glaciers that will soon likely become glacial ice patches (Santin et al., 2019). Regarding the glacier bed, it was reconstructed independently from both GPR datasets giving comparable results thus demonstrating the efficiency of both ground- and helicopter-based GPR surveys. During the interpretation of the 2015 GPR dataset, two different EM ice facies were recognized. The first was mostly transparent to EM signal while the other was characterized by diffraction hyperbolas and scattering phenomena. In order to better highlight the different electromagnetic responses of the ice and to possibly interpret such remarkable spatial variations, the distributions of the dominant frequency and of sweetness attributes were analysed. Fig. 3 shows that attributes can better image zones with different EM signature. In detail, below the high scattering zones a low frequency shadow was detected (Fig. 3b), suggesting the higher low-pass filtering effect of frozen materials in such areas. On the other hand, sweetness helps detection and boundary definition of zones with different physical characteristics (Fig. 3c). It is interesting to notice that the high amplitude scattering starts some metres below the snow- ice contact, while it continues down to the glacier bed. Such result suggests the correlation of EM transparent zones with cold ice, while highly diffractive portions can be related to the presence of warm ice containing some free water (Pettersson et al., 2004). An hypothesis for the coexistence of warm and cold ice in a temperate glacier is that they are related to different seasonal factors. In winter, the snow cover insulates the glacier from possible heat exchanges with the atmosphere and this effect becomes gradually greater up to 80-100 cm of snow thickness when the material below the snowpack is completely insulated (Guglielmin, 2004). Snow redistribution over the ground surface eventually has opposite effects, cooling down