GNGTS 2015 - Atti del 34° Convegno Nazionale

gravel. A longer profile allows also to reach a deeper investigation depth, thus extending the information provided. Unfortunately, the data at our disposal does not permit us to identify the embankment’s foundations, since they only show a conductive domain. Multichannel analysis of surface waves. Also the MASW data obtained well distinguish the reconstructed embankment from the natural levee. More in detail, the V s model of the natural part is characterized by rather low values slightly increasing with depth (from 150 to 260 m/s) that are likely related to the presence of clayey sand, in agreement with both the ERT results and the geotechnical soundings. On the contrary, the frequency-phase velocity diagrams of the reconstructed sector display large energy dispersion, since the propagation of superficial waves is probably strongly influenced by the heterogeneities within this part. Moreover, the stiffening due to the concrete septum is represented by the higher energy in correspondence to velocities close to 1000 m/s (especially for frequencies > 25 Hz). Surface Waves Analysis Wizard, Matlab, and SWAMI allowed the analysis of the MASW data. Ground penetrating radar and self-potential. The analysis of the GPR data was realized thanks to the REFLEX software and to the application of the empirical relation proposed by Topp et al. (1980). The results well agree with the ERT data, since G2 and T2 have higher average water content and lower average resistivity, respectively. It is important to underline that the water content distribution within the artificial part is likely irregular, with higher values in the cemented domain. Moreover, the signal to noise ratio is lower in G2 with respect to G1, thus agreeing, at least from a qualitative point of view, with more conductive materials. On the other hand, the drainage mattress compromised the SP data interpretation, since its position has been communicated only after the acquisitions and then confirmed by the ERT cross-embankment acquisitions. Thus, although the SP profiles show some voltage variations, they are likely due to changes in terms of materials within the mattress itself and not to infiltration phenomena. Conclusions. The characterization of natural and artificial river levee is a tricky challenge with important consequences on hydrogeological riskmitigation. Commonly applied techniques, although widely spread, provide only punctual data that are not adequate to properly describe the whole experimental domain. On the other hand, several geophysical methodologies allow non-invasive, cost-effective, and relatively faster soundings that can be combined in order to observe the same target under different points of view, as presented in this work. The joint use of ERT, MASW, GPR, and SP led to interesting results, from the identification of Fig. 3 – a) Lengthwise resistivity profile of the Frassine reconstructed embankment (dipole-dipole skip 0, 48 electrodes spaced 2 m). Water flow is from right to left. b) Cross-river resistivity profile. Picture facing upstream. All distances are expressed in meters. 48 GNGTS 2015 S essione 3.2