GNGTS 2024 - Atti del 42° Convegno Nazionale

Session 3.2 GNGTS 2024 electrodes hammered between the rocks and coupled with sponges soaked in salt water (e.g. Figure 1A). Subsequently, the same ERT array was measured using the conductve textle sachets electrodes pushed between the boulders and moistened with salt water (e.g. Figure 1B). Figure 1. A) Example of measurements collected with steel-spike electrodes hammered between the boulders and coupled with sponges soaked in salt-water. B) Example of measurements collected with conductve textle sachets electrodes pushed between the boulders and wet with salt-water. Pictures realized during the comparatve test performed on the inactve rock glacier of Sadole. 6. Data processing To evaluate the reliability of the conductve textle sachets we compared the contact resistances, the injected electrical current, the measured apparent resistvites, the quadrupoles reciprocal error, and the inverted resistvity models obtained with the measurements collected with the diferent electrode’s approaches. A statstcal analysis was carried out following the approach proposed by Bast et al. (2015), using R (R-core-Team, 2022). Violin plots have been used for the data exploraton, which include both the density distributon of the data and notched box plots. The robust Spearman correlaton coefcient was calculated for each of the fve variables. To statstcally compare the results obtained with steel-spikes electrodes and conductve textle sachets electrodes, for each site and variable, we fted a robust linear regression model with an extended MM estmator (using the R package robustbase - Rousseeuw et al., 2009). Diagnostc regression plots were realized to verify that the model assumptons were met. Finally, the inverted resistvity models (inversion modelling performed with ResIPy – Blanchy et al. (2020)) are ploted and compared to evaluate the diferences.