GNGTS 2013 - Atti del 32° Convegno Nazionale

electrodes. 3 tensiometers with log recorder were installed at 20, 30 and 40 cm depth in the subsoil. Methods. Our 3D micro ERT configuration consists of 24 or 48 electrodes buried in depth, installed on boreholes, plus 24 steel electrodes on the surface, Electrodes are round steel electrodes spaced 0.1 m, with all wire connection sited inside the PVC casing. This allows their installation by percussion avoiding invasive pre-drill with a very efficient contact resistance between electrodes and soil. For all the measurements we adopted a dipole-dipole skip-zero acquisition. For the inversion process we adopted a 3D mesh with detailed triangular prisms elements of 0.025 m side for the root zone box, while larger mesh cells were used for the background. All reciprocal measures, that show the statistical operator RSD (Relative Standard Deviation) exceeding the 5%, were removed from the dataset. The reciprocal procedure leads to different datasets, so to have comparable results the inversions are performed only on the same quadripoles that are present in all datasets. For the inversion we adopted the codes R3T (Binley et al. , 2011), a 3D current flows solution for finite element triangular prisms mesh. In order to enhance the time-lapse monitoring we inverted the data as a ratio of electrical resistances at a specific time respect to the resistance values at the background measurement (in our case the time step before the irrigation): (1) where R i is the electrical resistance at the i-th time-step, R 0 is the electrical resistance at the background measure and R hom is the electrical resistance for an homogenous space of 100 Ω∙m. The inversion of resistance values can clearly enhances the variation of electrical resistivity during the time-lapse experiment. The results of the inversion is a map of the percentage variation of resistivity respect to the background values: values equal to 100 Ω∙m show unchanged resistivity during time, while values less or more than 100 Ω∙m show a decreasing or an increasing resistivity, respectively. Results. The Val di Non case studies. We performed 3 irrigation tests in 3 different seasonal conditions and with different type of irrigation. In August 2011 we performed a drip irrigation (2.4 lt/h) experiment under the dry conditions of the summer season. In May 2012 we performed a widespread irrigation experiment (198 lt/h) in a wet conditions during the highest spring dynamic growing season period of the plants, and in November 2012 we repeated the experiment in very wet condition. For the first experiment micro enhanced the conductive areas are restricted on the 2 drip irrigation points. The wet areas progressively expand during irrigation; after the end of the irrigation the shallow wet areas slowly start to dry (Fig. 1a). The second and third irrigation experiments were conducted in different climatic conditions, during the top of the growing season plant. The site soil this time was very wet due to huge amount of rain during the previous month. The hot sunny days of the experiment (26°), assured a large amount of evapotranspiration. We used an irrigation sprinkler system in order to obtain a more uniform and widespread irrigation as possible. We performed also Time Domain Reflectometer (TDR) measurements during and after the irrigation to estimate soil-water content ( θ ). Results (Fig. 1b) show how, at the depth where are expected the main suction zone linked to root plant activity (De Silva, 1999), there is a clear increasing of resistivity due to the root uptake. Extremely interesting is the shape of the drier zone since position of the root activity is not uniformly widespread on the horizontal plane, but it is confined in the area of the parallelepiped that is directly linked with the location of the drip irrigation system. Results of the 3d experiment (Fig. 1c) show also an higher resistivity zone at greater depth after 3 h from irrigation start. Since resistivity properties are not effect only of water saturation but also of salinity, we believe the observed phenomenon can be related to the replacement of the resident 93 GNGTS 2013 S essione 3.2