GNGTS 2015 - Atti del 34° Convegno Nazionale

The geophysical acquisition. A waterborne continuous electrical profile was acquired on the Po River, from the confluence of the Sangone River (south) to the very city centre of Turin (north, Murazzi del Po), for a total length of approximately 4850 m of acquisitions (Fig. 2). The survey line passes throughout four bridges that represented difficult survey points both for the acquisition operations and for the strong electrical anomalies due to pier foundations. An array of nine electrodes fixed on a floating cable (96 m) dragged by a small boat was used for the survey. The array has two current electrodes, in the cable part closest to the boat, followed by eight potential electrodes. The current electrodes are 32 m apart, while the seven couples of potential electrodes had exponentially increasing spacing, from 0.5 m to 32 m. The first potential electrode was 0.5 m from the farthest current electrode. In the continuous profiling set up, dipole-dipole array data are collected measuring voltage potential differences between subsequent couples of potential electrodes given the same current injecting dipole. The towed cable floated on the river surface thanks to plastic floaters fixed near the electrodes that were fully submerged. The cable was kept stretched by a raft fixed at its end. We used a multichannel georesistivimeter (Syscal Pro in Sysmar upgrade – Iris Instruments) which was able to simultaneously acquire the seven potential measurements. The resistivity meter and the end of the cable were connected to a GPS device, in order to accurately record the spatial position of the acquired data. The acquisition step is about 2 seconds which results, on average, in one vertical electric sounding every 4 m. To recover the water depth, in order to constrain the data inversion, we connected to the IRIS georesistivimeter, and fixed to the side of the boat, a 170 kHz Airmar DT800 echo sounder. On average we had a bathymetry measure every 1 m. We track the boat and the cable as in the following. Two different GNSS (Global Navigation Satellite Systems) instruments were used during the survey: the first one was a dual frequency multi-constellation receiver (Leica 1230+GNSS) that was installed on the main boat, while the second one was a single frequency cartographic receiver (Topcon GRS-1) installed on the raft in order to estimate the direction of the cable where all electrodes are settled. Both real-time and post-processing approaches were followed. The position of the boat was determined in real-time thanks to the Regione Piemonte CORSs (Continuous Operating Reference Stations) network, performing an NRTK (Network Real-Time Kinematic) positioning, obtaining an accuracy of solutions of about 2-4 cm. As far as the scow is concerned, we acquired only the raw data that we post-processed in a single base solution (considering a master station 6 km far from the test-site), thanks to a commercial software, reaching a sub-centimeter accuracy. We considered only positions with fixed phase ambiguities in order to obtain the best accuracy available today with GNSS instruments, at the end of the process we obtained 2 highly accurate positions every second. Data processing. Before the inversion a statistical analysis of the data was performed, in order to evaluate the homogeneity of water resistivity and data variability with depth. The first three potential dipoles with smallest reciprocal spacing mainly investigated the river water, which has a constant resistivity value of 43 Ω m over the whole travel path. On the other hand, the following four dipoles are expected to give information about the riverbed sediments. CVES data were inverted using both a classical 2D tomographic processing and a laterally constrained inversion (LCI) approach. The LCI was developed to invert CVES data acquired along a profile by Auken and Christiansen (2004). This approach is based on a pseudo-2D layered parameterization of the investigated geological medium: the inversion result is a set of 1D consecutive resistivity models, each one corresponding to a sounding, composing a pseudo- 2D section. All the VES soundings along a profile are inverted simultaneously by minimizing a common objective function, which contains all the acquired data, the available a-priori information and lateral constraints among consecutive models. Through the lateral constraints, information from one vertical electrical sounding are interconnected with the neighbouring ones, producing the final pseudo-2D section. The lateral constraints are chosen in a way to allow for pseudo-2D sections that are more or less homogeneous on the basis of the geological setting of GNGTS 2015 S essione 3.2 79