GNGTS 2015 - Atti del 34° Convegno Nazionale

the investigated area. In particular, the degree of lateral homogeneity of the considered model parameters is controlled by the strength of the constraints. If the expected lateral variability is small, a strong constraint will be applied; conversely if a large variation is expected, the strength of the constraint will be relaxed. For a reliable inversion auxiliary a-priori data are also fundamental to ensure that as much known information as possible is considered in the inversion process. Crucial information for waterborne surveys includes bathymetry and water resistivity, which describe the properties of the water column. By providing these constraints, the inversion procedure is focused on the deposits beneath the riverbed, thus allowing a more accurate delineation of the sediment’s electrical properties. The conceptual reference model on which the inversion process was based is a three layered medium. For each inversion it was possible to a-priori fix the thickness and the resistivity of the water column (first layer). The first layer thickness was a-priori known thanks to echo sounder measurements of the bathymetry conducted simultaneously to the electrical survey. The first layer resistivity was kept constant (43 Ω m) considering the low variation of the mean of the nearest potential dipoles. No constraints were set for the second layer (fluvial deposits) while the electrical resistivity of the third layer (Turin Hill marls) was fixed to the value 23 Ω m ) ± 20 Ω m (mean and standard deviation of the whole raw measurements dataset for the seventh dipole) in order to force the inversion to find a lateral continuity for the deepest deposits. An appropriate Matlab code was developed to implement the inversion, similar to the one described in Colombero et al. (2014). On the other hand, classical 2D inversion was carried out using Res2DInv software, in continuous resistivity profiling mode, fixing both the water resistivity and the bathymetry values (Loke and Lane, 2004). Preliminary results. In Fig. 3 the results of the inversion of the whole dataset are shown, both for the LCI approach and for the classical 2D tomography. To maintain a readable vertical scale the profile has been split in 1 km stretches. The depth of the riverbed varies from a minimum of 2 m to a maximum of 10 m. Below the water layer (blanked in all the sections) a thick layer of sediments with resistivity higher than water resistivity (43 Ω m) is found. The layer thickness is not homogeneous, showing irregular depressions and reliefs, but generally it seems to progressively slightly increase towards north. Resistivity values range from the water value up to 250 Ω m. Quite low values are shown in the first 2400 m of the survey line (43-90 Ω m), from this point to the end of the survey the resistivity increases. This first layer of sediments is characterized by the fluvial deposits (mainly silt, sand and gravel) of the Po River. The increase in resistivity from south to north can be linked to a local increase in granulometric size of the sediments or to the presence of more compacted or cemented horizons. Due to the three-layer model assumption, LCI results show for this horizon a unique mean value of around 150 Ω m in the last sections, while 2D tomography results distinguish an upper sub-layer with higher resistivity and a lower horizon with values comparable with the previous sections. The bottom part of each section is instead characterized by sediments with resistivity values (20-40 Ω m) lower than water resistivity, that could be likely related to the marls of Turin Hill marine sequence. The morphology of this horizon is quite undulating, suggesting strong erosional phenomena both along the main flow of the river and at the confluence of the lateral tributary streams. The depth of the marls ranges from a few meters in the southern sections to more than 15 m toward north. The geophysical results were compared with direct geological information, consisting of logs and water wells with stratigraphic information in the surroundings of the river banks. Unfortunately, not all the available direct surveys reached the marls, but where the investigation depth was higher, the data were found to be in good agreement (Fig. 3c). A strong electrical anomaly was detected near the first bridge of the survey, due to the pier foundation structure. 80 GNGTS 2015 S essione 3.2