← Back to Sessions

Probabilistic Multiphysics Inference for Permafrost Characterization and Earthquake Site Hazards Assessment

Description: 

Permafrost thaw negatively impacts ecosystems and infrastructure throughout northern regions, which are warming at greater than double the global average rate. Permafrost plays a crucial role for soil stiffness and shear strength, and its degradation is a source for increasing risks of natural hazards, including landslides, thaw subsidence, and earthquake-related hazards. The severities of these impacts remain poorly studied, and mitigation efforts require better subsurface characterization and monitoring of temporal permafrost changes, as well as improved constraints on permafrost mechanical (poro-elastic) properties. This work considers the potential of permafrost characterization with probabilistic multiphysics inference of non-invasive geophysical observations. Our analysis employs a probabilistic (Bayesian) framework for robust characterization of uncertainties in inferred permafrost properties. We demonstrate this analysis with synthetic audio-magnetotelluric, seismic surface-wave dispersion, and seismic horizontal-to-vertical ratio data under varying permafrost conditions. The complementary information contained in the different data types, and the probabilistic estimation of objective data weights, enables higher-resolution inference of shallow subsurface properties than can be achieved by independent inference from individual data types. This is particularly attractive for permafrost characterization and monitoring, which considers strong elastic and electrical resistive signatures. Finally, inferred shallow Earth structure under varying permafrost conditions are used to estimate parameters for typical seismic site hazard classification and assessment. Our initial results indicate climate-change likely has, and will, influence earthquake site hazards. Furthermore, our work represents an initial step towards improved permafrost characterization for seismic hazard assessment with robust uncertainty quantification while avoiding subjective inversion practices.

Session: Applications and Discoveries in Cryoseismology Across Spatial and Temporal Scales [Poster Session]

Type: Poster

Date: 5/2/2024

Presentation Time: 08:00 AM (local time)

Presenting Author: Jeremy

Student Presenter: No

Invited Presentation: 

Authors