Dynamic Rupture on Complex Faults

Complex Fault Systems

Natural fault systems can be quite complex, occasionally consisting of numerous intersecting, bending and disconnected fault strands. Many of the largest continental strike-slip earthquakes that have occurred in the 21st century have taken place on geometrically complex faults. Examples of such earthquakes include the 2002 Mw 7.9 Denali earthquake, the 2008 Mw 7.9 Wenchuan earthquake, the 2016 Mw 7.8 Kaikoura earthquake and the 2023 Mw 7.8 and Mw 7.6 earthquake doublet in Turkey. In addition to these geometric fault complexities, there can be variability in the topography, seismic velocity and the elastic/inelastic response of the region around the fault system. Understanding how these complexities influence earthquake rupture propagation is an essential component to properly assessing seismic hazard. In this research I utilize 3D dynamic rupture simulations to explore the physics of rupture along various complex systems. Below are a few of my ongoing and past projects.

Statewide California Earthquake Center Community Fault Model Version 7. Plesch et al., 2024

Dynamic Rupture Models of off-fault Damage: Elsinore Fault at Coyote Creek. (Ongoing)

Detailed field mapping of the Elsinore fault in Southern California has highlighted significant off-fault damage and rock pulverization near the Coyote Mountain segment. In this ongoing research I am utilizing dynamic rupture simulations to investigate the underlying physics that may explain these site-specific field observations. In particular how rapid strain release and coseismic volumetric stress changes in the surrounding rock could lead to inelastic off fault behavior.

Figures from the ongoing project. The panel on the left shows the mapped fault trace and slip offset measurements conducted primarily by Tom Rockwell and Ashley Griffith. The panel on the right is the final results from a dynamic rupture model of this fault section, showing the on-fault slip and the resultant accumulated inelastic strain at the Earth's surface.

Effects of Topography on Supershear Rupture. (Marschall & Douilly, 2026 BSSA)

Many faults are in close proximity to mountain ranges and it has been shown that this topography can introduce dynamic changes in normal stress during rupture. In this study we show that the presence of nearby mountains can suppress free surface induced supershear rupture, in essence causing the rupture to remain in the Sub-Rayleigh speed domain. The suppression of the induced supershear can actual promote through-going rupture at fault intersections. So with all initial on-fault conditions equal (stress, friction, etc.), the presence of nearby topography can actually lead to larger earthquakes on a complex fault system.

Figures Marschall & Douilly, 2026. Top left is a 2D schematic of the model setup , bottom left is a resultant model run showing bilateral supershear rupture for the Flat X model (no nearby mountains) and unilateral rupture in the Topography X model with topography. The right panel shows the slip-rate evolution and final slip of the two models, note that rupture only propagates to the secondary segments for the model with topography.

Effects of Dip Angle on Rupture along Branch Faults. (Marschall & Douilly, 2025 BSSA)

Faults can often intersect with each other, forming what is commonly referred to as a branch fault system. In this research we investigated how non-vertical fault segments can promote throughgoing rupture at branch fault intersections. We found that under certain initial stress conditions, a rupture nucleating on a secondary fault is more likely to continue onto the main fault strand if the secondary fault is shallowly dipping because of dip-slip induced stress interactions with the free surface. When we model the branch fault as a buried system, these free surface effects decrease and rupture remains solely on the secondary fault. This may be able to explain why some earthquakes, such as the 2002 M 7.9 Denali earthquake, are able to grow into large events.

Figure from Marschall & Douilly, 2025 in BSSA. The figure shows rupture evolution on dipping branch faults, the models are identical in geometry and stress, with the only difference being the bottom panel is buried by 1 km. The rupture is only able to propagate in the un-buried model that ruptures the free surface.

Page updated

Google Sites

Report abuse