COMPLEX RESISTIVITY AND SPECTRAL INDUCED POLARIZATION TECHNIQUES FOR ENVIRONMENTAL APPLICATIONS

The potential and limits of the complex resistivity and the Spectral Induced Polarization techniques for environmental applications are investigated starting from the controlled laboratory conditions. The solution of the 2D forward problem for the complex resistivity is achieved on cylindrical-shaped models by a transformed formulation and using the Complete Electrode Model to define boundary conditions. The finite element algorithm, implemented in Matlab to solve numerically these equations, has proven to be accurate and stable through a comparison with the 3D full formulation and synthetic simulations. Accordingly, this formulation can be used to produce 2D reconstruction on cylindrical laboratory devices provided with one horizontal mid-section of electrodes (e.g. lysimeters). The whole geophysical process (data processing, forward modelling and inversion) has been included within EIDORS through an interactive Matlab algorithm with numerous new function added. Simulations with Gaussian noise, performed on heterogeneous models with anomalies included, provide low-error inverted models whereas anomalies are well-detected both for shape and position. EIDORS software could act as a base code for tomographic inversion of time-domain and frequency-domain data, both for laboratory and for in field investigations, because of its high-flexibility and reliability now reached by forward and inversion routines, with the primary aim of to share data and promote collaboration between groups working these fields. A new experimental device was presented to perform 2D and 3D measurement on cylindrical physical models in order to simulate typical subsoil configurations. The cylindrical symmetry avoids to have boundary effects of prismatic tanks and allows us to provide horizontal cross-section within the different layers. Preliminary water and sand models have highlighted the main problems due to saturation, electrodes, array and frequency range. The experimental AC instrument, after the calibration, provides reliable amplitude and phase dataset with stable and repeatable measurements. Measurements are not affected by the inductive coupling for frequency up to 5-10 Hz. Physical models built up in the laboratory have concerned the preparation of a typical sand-clay-sand multilayer model, simulating a shallow aquifer by saturation of the top layer. The HFE 7100, having physical properties similar to the chlorinated solvents, was spilled from the top to simulate a point source spilling, that is for example a failure of a little storage tank. Laboratory experiments have confirmed the reliability of the algorithm proposed. After HFE spilling, impedance amplitude increases progressively as expected, taking into account the percentage ratio between the global volume of HFE spilled and the sand volume. On the other hand the phase increases less evidently, since they are probably more sensible to the particular working conditions. Evidences from SIP phase spectra seems to be also convincing as we detect a single peak for frequency of about 1-2 Hz. These results confirm that complex resistivity tomography can play an important for contaminant detection, stressing once more the importance to define codified criteria for acquiring CR ad SIP in field data