GNGTS 2017 - 36° Convegno Nazionale

624 GNGTS 2017 S essione 3.2 m towards SW, was interpreted as an unconformity created by erosion and sedimentation processes due to floods of Tiber river. A deeper interface (~0.8 m) is visible with the same SW-dipping trend which, according to the actual θ v data (Fig. 3b), was interpretable as the deposit with higher θ v . The θ v values calculated using both Topp’s and calibrated models are between 0.042- 0.045 (first 0.40 m of depth) for all the ε r values by hyperbola fitting method. CMP gather provided more detailed information, with θ v values ranging between 0.065 - 0.107, from topmost soil down to 0.8 m depth. The use of calibrated model in the laboratory gives θ v values ranging from 0.061 to 0.12 for the same depth. We highlight how an integrated methodology sensitive to WC increase is useful to obtain qualitative and quantitative data, like the subsurface geometries and reliable (calibrated) θ v values. Moreover, such results allow us to better describe the infiltration processes in the unsaturated zone. In fact, we interpret the clear variation of WC in depth as the downward movement of a slug of infiltrated water (Fetter, 2001) related to a summer rainstorm occurred about one week before the field survey. Conclusions. We provide the first reference values of ε r and WC for two typicalsandysoilswidelyoutcroppingin central Italy, through amultidisciplinary approach encompassing laboratory and field scale experiments. The results obtained for the materials and site conditions here investigated, confirm that the application of Topp’s equation leads to an overall WC underestimation, in particular for θ v > 0.17, regardless of the model used for WC estimation. Experimental GPR data collected on soil S B along the Tiber river were used to provide tridimensional information and a velocity model down to 1 m depth, necessary to estimate ε r values and θ v at the field scale. As expected, due to dry condition, the studied sands showed low WC on the shallow layers, whilst higher WC more in depth. The reconstructed 3D model of the site encompasses several interfaces, among those, a clear deeper reflection (about 0.80 m depth) slightly deepening toward SW, was interpreted as representing a higher θ v level by using the whole data at disposal. The multidisciplinary approach used in this study, was confirmed to be a valid tool for punctual and areal WC estimation, providing subsurface geometries and information on the water pattern of the unsaturated zone. We therefore strongly recommend the use of integrated analyses for an accurate WC estimation of sandy soils of different nature and different moisture conditions. Fig. 3 - Summary of data integration (sampling and GPR): a) sketch of the survey site and sampling locations; b) results of sampling water content analysis, showing a θv increment at a depth of about 0.80 m; c) pseudo-3D GPR volume showing a 3D model allowing to better interpret the subsurface (Y axis scaled x20, Zx4 exaggeration).