Geostatistical interpolators for the estimation of the geometry of anthropogenic deposits in Rome (Italy) and related physical-mechanical characterization with implications on geohazard assessment.

The role played by the anthropogenic deposits with poor geotechnical characteristics to induce seismic amplification and its influence on hydrogeological conditions, due to its permeability characteristics, has been demonstrated by numerous studies conducted in urban environments. The historical center of Rome (Italy) is characterized by the presence of high thickness of anthropogenic deposits, a stratigraphic unit that accumulated during the centuries of urban development. The modeled reconstruction of this unit could provide a tool for the elaboration of preliminary maps used for urban planning by public institutions. In this work, the reconstruction of the basal surface of the anthropogenic deposits was carried out using lithostratigraphical and geotechnical data recorded in 1355 boreholes drilled within the historical center of Rome. Deterministic and geostatistical estimators have been then used to predict the values at non-sampled locations to construct different maps of the basal surface. Grid statistics allowed the comparison of the obtained maps to select the best spatial predictor. Mono and multivariate geostatistical intepolators [ordinary kriging, kriging with external drift (KED) and cokriging] have been selected and discussed, and KED was the best interpolator for the used dataset. The btained surface was then used to calculate a thickness map by subtracting it to the Digital Terrain Model. Furthermore, results of some mechanical characteristics of the anthropogenic deposits are presented and discussed in the light of the reconstructed geometry. The knowledge of both geometry (i.e., basal surface and thickness) and mechanical characteristics of these deposits could provide useful information for the evaluation of geological hazards in urban areas.