Speaker: Prof. Brandan Meade
Affiliation: Harvard
Date: Tuesday, September 29, 2026
Time: 3:30 PM
Location: 3853 Slichter Hall
Speaker Website: https://eps.harvard.edu/research-group/meade-group/
Abstract
Large earthquakes are infrequent while motion across fault systems is persistent. With the modern volume of geodetic data, the problem is no longer data scarcity but inference robustness: what can and cannot be resolved about kinematics on a three-dimensional fault network, and with what uncertainty. Here we describe a method for probabilistic imaging of fault system activity that achieves resolving power through the combination of dense geodetic observations, non-linear coupling constraints, and Hamiltonian Monte Carlo sampling. Applied to the California fault system, we infer heterogeneous interseismic coupling behavior with no general spatial relationship to microseismicity. These estimates of decadal scale fault system state are combined with a statistical seismicity model (without the concepts of stress or friction) to generate synthetic earthquake sequences that provide realizations of coseismic slip partitioning.
Speaker Bio
TBA
Research Expertise
Geodesy/earthquake cycle