SCEC Award Number 25156 View PDF
Proposal Category Individual Research Project (Single Investigator / Institution)
Proposal Title Adjoint ruptures with slip-weakening friction for dynamic source inversions and sensitivity analysis
Investigator(s)
Name Organization
Eric Dunham Stanford University Wenqiang Zhang Stanford University
SCEC Milestones C1,2,3-1, C1-2, C2-2, C3-1 SCEC Groups FARM, RC, GM
Report Due Date 03/15/2026 Date Report Submitted 03/25/2026
Project Abstract
 Dynamic rupture models are used to understand rupture propagation and arrest, predict ground motion, and interpret seismic and geodetic data. Dynamic source inversions provide valuable constraints on frictional parameters and initial stresses but remain challenging to perform. Modelers either hand-tune parameters to find a single best-fitting model or use computationally expensive Bayesian inversion methods. In this project, we seek to automate and accelerate dynamic source inversions by using the adjoint method to calculate the gradient of an objective/misfit function with respect to an arbitrary number of model parameters. The gradient can also be used as a sensitivity kernel to quantify how uncertainty in model parameters propagates through the source model to uncertainty in the rupture history and ground motion. The project builds on our recent work deriving the adjoint method for rate-state friction and performing a proof-of-concept example using a 2D antiplane shear code. In this SCEC project, we extended the theory of adjoint rupture dynamics from rate-state to slip-weakening friction and implement both laws in the open-source, parallelized 3D discontinuous Galerkin dynamic rupture code drdg3d, written by postdoc Wenqiang Zhang. We combined this with L-BFGS to perform dynamic rupture inversions.
Intellectual Merit Dynamic rupture models are used to understand rupture propagation and arrest, predict ground motion, and interpret seismic and geodetic data. Dynamic source inversions provide valuable constraints on frictional parameters and initial stresses but remain challenging to perform. In this project, we extended the adjoint method for gradient-based optimization and inversion to dynamic rupture problems with slip-weakening friction. We also implemented the method in an open-source 3D code. This will allow the community to automatically perform dynamic source inversions, which will help us learn about fault friction and stresses, and the conditions under which natural earthquakes occur.
Broader Impacts The project trained a postdoctoral fellow. It also provided the earthquake science community with a powerful new tool that can be used to study earthquake rupture dynamics and ground motion.
Project Participants Wenqiang Zhang, Rikuto Fukushima, Eric M. Dunham - Stanford University
Exemplary Figure Figure 2
Linked Publications

Add missing publication or edit citation shown. Enter the SCEC project ID to link publication.