Geostatical modelling of hydrofacies in the suboil of Porto Marghera and the lagoon of Venice (Ne, Italy): implications for aquifer connectivity assessment

Progress towards improved modelling of porous groundwater reservoirs requires the use of large datasets of sedimentological and hydrogeological information from boreholes. Careful integration and analysis of these data can indeed allow for robust conceptual models, which are imperative for appropriately choose the geostatistical method to be used when simulating and/or interpolating reservoir properties in 3D. This study focuses on the 3D modelling of hydrofacies in a Late Pleistocene-Holocene alluvial aquifer by means of different geostatistical techniques. The case study is located in the inland of Venice (NE Italy) and consists of a 5 X 10 km area over which an impressive dataset of boreholes (5290, with average spacing of 150 m) is available in the shallow (35m) subsoil. The modelled subsoil consists of three main stratigraphic units which are, from older to younger: i) the pre-Late Glacial Maximum (LGM) pebble-dominated Montebelluna fan delta; ii) the LGM sand-dominated braided fan delta of the Brenta river, iii) the post-LGM alluvial deposits resting on top of a laterally extensive paleosoil key horizon and iv) the anthropogenic layer. Hydrofacies distribution was simulated using both pixel-based (e.g. sequential indicator simulation SISIM, Truncated Gaussian simulation TGS) and object-based methods routinely employed in modelling of groundwater and oil reservoirs. The modelling output includes voxel-by-voxel (cell size: 50.0 X 50.0 X 0.2 m) sets of equiprobable hydrofacies distributions allowing for a probabilistic approach to reservoir body characterization. The comparison of results across different modelling runs and approaches are discussed with reference to connectivity of the main sandbodies present in the subsoil and provisionally validated by means of both in-hole hydraulic tests and indirect evidences of acquifer transmissivity.