GNGTS 2018 - 37° Convegno Nazionale

GNGTS 2018 S essione 3.2 693 affect the measured data in laboratory studies. Therefore, we developed a method to assess such 3D effects on 2D ERT data measured along the river levees (Arosio et al., 2018). Moreover, we developed an analytical procedure to take into consideration the effect of buried electrodes. The presence of buried electrodes, not in contact with the air-soil surface, appears to modify the measured resistivity, lowering the real value in the shallow depth. The result is a data correction strategy to correct the effect of buried electrodes, transforming the measured data as it was performed with surface electrodes, and then to correct the geometry effect of each test implementing a forward modelling with Res3dmod and then obtaining a correction curve indicating the geometrical amplification of the resistivity of the soil (Arosio et al., 2018). Fig. 3a and Fig. 3b illustrate the final correction graphs obtained for T29 andT1.37. The graphs are calculated for the Wenner array with the unit electrode spacing used in the corresponding experiment. The correction graphs show how fundamental is to process 2D ERT data for 3D effects to remove artefacts in the analysis of the resistivity variations. It is clear that such an amplification is not correlated with geological aspects and can strongly affect the interpretation of the data. Fig 3c and Fig. 3d show the pseudosections of T29 and T1.37 corrected using their corresponding correction graphs. Fig.3.c now shows a reasonable range of resistivity values. The soil is mostly saturated even if the water level is not at a maximum height in the canal due to the high permeability of pure sand. TDR data also showed that shortly after adding water in the canal, the volumetric water content reached a maximum value of 35% and then remained constant at this value regardless the further increase of water level. Resistivity variations in Fig.3.c are attributed to compaction variations resulting from the manual construction of the levee model. Fig.3.d shows the situation after the drawdown procedure when the resistivity values are slightly higher. The sand has a fast response to water circulation and therefore, Fig. 3 - Correction graph for a) T29 and b) T1.37. Corrected pseudosections for c) T29 and d) T1.37.