GNGTS 2014 - Atti del 33° Convegno Nazionale

deposits and waters of the Vermigliana creek, whose resistivity have orders of magnitude of 1000 Ωm and 100 Ωm respectively. In order to explain the features of this domain, whose existence in the HZ has also been confirmed by two auxiliary ERT surveys, we hypothesized the presence of a high clay fraction coming from the glacial moraines and transported by the creek itself, which not only increases the electrical conductivity in the sub-riverbed, but also reduces its hydraulic conductivity. On the other hand, if we focus on the areas adjacent to the river, also referred to as “riparian zones”, it is possible to highlight a difference between the left and the right bank (Fig. 2), since the former has an average electrical conductivity slightly lower than the latter. Finally, albeit being outside our domain of interest (i.e., the area between the surface electrodes and the borehole electrodes), the resistivity cross-section in Fig. 2 also shows the bedrock, identified by the very high resistivity domain at a depth of 10 m below ground level. The second inversion technique is strongly related to the ERT time-lapsemonitoring currently in progress in the Vermigliana site and aims at highlighting how resistivity varies over time. As already described by Perri et. al (2012), this approach is based on the following equation: where R i is the transfer resistance measured at time t , R 0 is the background (i.e., time 0) transfer resistance and R hom is the transfer resistance for a homogenous resistivity distribution model; all these transfer resistance values are referred to the same quadripole. Once R is computed for every electrodes quadripole common to all the available datasets, the ERT data inversion with an error equal to 3% is performed, as already described above. The results consist of T–1 cross-sections (with T = total number of ERT acquisitions) displaying the resistivity variation over time expressed as percentage, with respect to the background survey (100% indicates no changes, higher values imply an increase in resistivity, and lower values are related to a decrease). Although the time-lapse monitoring began in July 2013, for this work we consider only the acquisitions carried out from May 5, 2014 onwards, because of the nature of the observed phenomena. In order to better analyse the outcome of this second inversion technique, it is useful to divide the domain of interest into three parts: sub-riverbed, left bank, and right bank.In the first part, between May 14 and July 16, 2014 an increase in resistivity variation takes place (up to 150%), followed by a rapid decrease to 100%. Such behaviour may be due to a glacial water pulse, which is poor in ions and therefore characterized by a high resistivity. On the other hand, the left bank is characterized by a constant decrease in resistivity variation, from values higher than 120% to values around 60% (on average). This variation is probably related to the presence of an effluent of two small lakes upstream, whose waters are presumably more conductive. Finally, the last part shows a slight increase in resistivity variation over time (values are always, on average, Fig. 2 – Example of resistivity cross- section resulting from the ERT survey conducted in Vermiglio on July 30, 2013. The cross-section is facing downstream. A low resistivity domain is located under the Vermigliana creek, while the riparian zones show, on average, higher resistivity values. The black dots represent the electrodes position both on the levee surface (24 stainless steel electrodes) and inside the perforation drilled under the Vermigliana creek (48 brass electrodes). 132 GNGTS 2014 S essione 3.2 € R = R i R 0 R hom