Validation process and critical review of a 3D subsurface model: a test on the Trento Platform (Southern Alps, NE Italy)

3D geological modelling aids prediction of subsurface geology and finds application both in industry and academic research. Although a range of different approaches (e.g., implicit vs. explicit) exists, geological modelling entails interpolating (often sparse) outcrop and borehole observations regarding lithological composition and geometry of a geological structure and quantifying residual uncertainty. Although most often appealing, a 3D geological model must be in first place robust, and its robustness validated against observations. It is widely recognized that robust 3D geological models are valuable tools for tackling uncertainty in a range of industry projects, from ore mining to tunnelling and nuclear waste management, and can help promoting geological knowledge in the general public. In Italy, the Istituto Superiore per la Protezione e la Ricerca Ambientale (ISPRA) encourages the production of 3D geological models as part of the national 1:50.000 scale geological mapping project (CARG). Newly acquired geological mapping data from an area of ca. 42 km2 belonging to the 1:50.000 scale geological map sheet 081 “Rovereto” (Monte Pasubio, Trento Platform, NE Italy) were used to test an iterative explicit modelling workflow, which include multiple stages of validation. 3D modelling was accomplished using Move (PE Ltd) and the following workflow: 1) high-detail 1:10.000 geological field and remote sensing mapping; 2) digitalization of the geological map in a GIS system using a cm-resolution LiDAR DEM; 3) construction of a geological cross-sections grid; 4) 3D kriging interpolation of the obtained 2D horizons and faults to reconstruct the 3D distribution of each geological unit recognized in the area. According to the field mapping, the Jurassic geological units cropping out in the study area exhibit flat geometries with a regional N-NW strike attitude, display constant thickness and are affected by the NW-SE trending sinistral strike-slip faults of the Schio-Vicenza system. The resulting 3D model was critically evaluated and iteratively improved using the following validation process: a) check of the surfaces interpolation with the topography (should corresponds with the mapped geological limits); b) creation of thickness maps (should be coherent with field observations); c) edit of the geological cross sections if needed; d) restart the modelling from step 4 (kriging interpolation). This study demonstrates how validation by a detailed geological interpretation lead to the production of new and more reliable 3D models with an integrative and iterative process, especially needed when data are collected with an irregularly scattered distribution.