Bayesian Inversion of Heterogeneous Media: Introducing the Next Generation of Integrated Studies for Offshore Site Investigations

Abstract This paper introduces a new methodology for addressing the probabilistic determination of geomorphological features and of the material properties, which allows for its uncertainty quantification and the spatial correlation across parameters when information from different geophysical methods is merged. The proposed approach departs from traditional stratigraphic characterizations advancing the possibility of establishing a systematic and reproducible approach that can reduce the subjectiveness and the unnecessary costs related to over conservative estimates for hazard assessment and structural design. A synthetic case study for a 1D heterogeneous media is formulated to illustrate the use of the probabilistic calibration method and of Tarantola's operators, when applied to two geophysical techniques used at the same site. Results show probability distributions of the location of the transition between layers, of the material properties, and of the collaborative operators for merging information content from different geophysical methods.

Bayesian Inversion of Heterogeneous Media: Introducing the Next Generation of Integrated Studies for Offshore Site Investigations Conference Paper