GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 3.3 ______ ___ GNGTS 2023 Results First, we show that by applying the method to a numerical dataset, we can recover the true resistivity model from the data. Then we apply the rescaling to the data of a modified model and show that we can retrieve a good approximation to the modified model. A simple layered system was generated, and the apparent resistivity measurements were simulated using the EMPYMOD routine (Werthmüller 2017), the simple layered system had the following characteristics Resistivity [Ohm*m] Thickness of the Layer [m] 450 500 2500 800 450 Half Space Table 1: Parameters of the resistivity model Using the process described in the method section, a polynomial expression able to correct the ∆ misfit between apparent resistivity and the resistivity measured was obtained. Once the apparent resistivities are rescaled using the polynomial expression the layered system is retrieved and is then compared with the true resistivity model. Figure 1 shows that the proposed method can retrieve a smooth version of the resistivity model using apparent measurements from an MT survey. One question that arises is if the polynomial expressions generated using the known model can be used to correct apparent measurements from nearby zones in which only apparent measurements have been conducted. We hence created a numerical model with a variation in the resistivity of the target layer and we used it to generate apparent resistivity data and then we applied the previously obtained rescaling function to the new data without introducing any a priori knowledge about the model parameters. In Figure 2 we show the results of the rescaling and compare them with the resistivity models. In blue the true model and the retrieved model used for retrieving the recalling function (as in Figure 1) and in yellow the modified true models (solid) and the retrieved models (dashed) by applying the rescaling function retrieved for the blue model. Figure 3 shows how the rescaling process works in a more complex scenario. Conclusions We showed that, if a local 1D resistivity model is known in an area where EM measurements are available, it can be used to compute a rescaling function that can then be applied directly to the data over larger areas to transform the apparent resistivity data directly into resistivity models