Probabilistic sediment characterization using Chirp sub-bottom profiler and Portable Free-Fall Penetrometer measurements

Reliable seabed characterization is essential for offshore engineering and marine science; however, it is often constrained by the trade-off between the continuity and accuracy of survey datasets. This study introduces a probabilistic framework to integrate data from a Chirp sub-bottom profiler and a Portable Free Fall Penetrometer (PFFP) for rapid sediment characterization. The methodology first derives probabilistic sediment classifications from each instrument—via geophysical inversion for Chirp and a machine learning model for the PFFP. The framework then uses Gaussian Process Regression (GPR) in a logit-transformed space to spatially model these probabilistic outputs. The resulting posterior distributions from each instrument model are formally combined using Bayesian methods. The framework produces a single, unified sediment profile with significantly reduced and quantified uncertainty. Validated in two distinct field case studies, the approach effectively uses the continuous Chirp data to improve spatial interpolation between PFFP measurements. This integrated analysis framework leads to a more accurate and robust characterization of the seabed than either method can achieve alone.