SCEC Award Number 22141 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Dynamic models of near-surface faulting and off-fault deformation in the 1971 San Fernando Earthquake
Investigator(s)
Name Organization
David Oglesby University of California, Riverside Elyse Gaudreau University of Victoria Gareth Funning University of California, Riverside
Other Participants James Hollingsworth, ISTerre, Université Grenoble-Alpes, France
Edwin Nissen, School of Earth and Ocean Sciences, University of Victoria, Canada
Jordan Cortez, Department of Earth and Planetary Sciences, University of California, Riverside
SCEC Priorities 3g, 4a, 2e SCEC Groups FARM
Report Due Date 03/15/2023 Date Report Submitted 12/05/2023
Project Abstract
 The 1971 San Fernando earthquake produced a rich pattern of ground motion and deformation that continues to be explored over 50 years later. The fault geometry has two main thrust segments (Sylmar and Tujunga). The Sylmar segment appears to have a steep surficial dip of ~55°, which cuts through young, soft sediments near the surface, while the Tujunga segment has a shallower near-surface dip of ~25°, and cuts through somewhat older and stiffer near-surface materials. Both segments appear to converge at depth to a single surface with a dip of ~40°.
A key observation in this event is the large amount of off-fault deformation (OFD) measured in recent analyses by Gaudreau et al. (2023). Vertical OFD was found to be larger as a percentage of total offset on the Sylmar segment. We use 2D dynamic finite element models to investigate how fault geometry and Drucker-Prager plasticity might have affected vertical OFD near the Sylmar and Tujunga segments. We find that there is more vertical OFD in Tujunga than Sylmar, but the OFD in Sylmar is closer to the fault trace, where it would be more likely to be resolved observationally. In this way, we reproduce the observed larger vertical OFD in Sylmar. The results also could pose a testable hypothesis that there may be more OFD in Tujunga, but located farther from the fault trace on the hanging wall.
Intellectual Merit The surface rupture of the San Fernando earthquake consists of two main fault segments offset by ∼1300 m: the western, W-striking Sylmar segment, which cuts through the San Fernando Valley, and the WNW-striking Tujunga segment that runs along the base of the San Gabriel Mountains to the east. Field surveys suggest that the near-surface portion of the fault ruptured along bedding planes, with the Sylmar segment having a steeper dip (∼55°; Weber, 1975) than that of the Tujunga segment (∼25°; Barrows, 1975; Kamb et al., 1971; Sharp, 1975). These dips are likely confined to a depth of ∼2 km, and connect with the main fault below, which has a dip of ∼40° based on aftershock relocations and seismic reflection data (e.g., Carena & Suppe, 2002; Fuis et al., 2003; Mori et al., 1995). The Tujunga segment cuts through an old, stiff sedimentary package, while Sylmar cuts through younger, more compliant sediments.
Three-dimensional coseismic surface offset measurements of the 1971 San Fernando earthquake calculated from optical image correlation of historical aerial photographs and DEM differencing indicate that most of the offset occurred in the fault’s damage zone outside of the inner, high-strain fault core (Gaudreau, Hollingsworth, Nissen and Funning 2023). In addition, the vertical component of OFD constitutes a higher percentage of the total offset on the Sylmar segment (81%) than on the Tujunga segment (64%), with such deformation typically concentrated on the hanging wall side of the fault.
In this project, we used 2D dynamic rupture modeling (Barall 2009) to investigate the source of this off-fault deformation, with a particular focus on why there was more vertical OFD in Sylmar than in Tujunga. In particular, we compared models with and without Drucker-Prager Plasticity (Hughes and Taylor, 1978; Simo and Hughes, 1998; Zienkiewicz and Taylor, 2000) to model the generation of damage in the bulk and near the Earth’s surface. As a proxy for a full 3D model, we perform two 2D models, one for each fault segment (Figure 1), and compare the slip and surface deformation patterns. The modeling parameters are in Figure 1 and Table 1.
We find that Drucker-Prager plasticity results in reduced near-surface slip for both the Tujunga and Sylmar models. This reduced near-surface slip, coupled with off-fault damage, leads to a significant amount of off-fault vertical surface deformation in both models (Figure 2). However, there is more obvious vertical OFD in Tujunga than in Sylmar, contrary to observations. We note that the OFD peak in the Tujunga model is significantly farther from the surface fault trace, though. When we zoom in to the scale of the observed OFD data, (Figure 3), we find that Sylmar has the more obvious vertical OFD, in agreement with observations.
Our tentative conclusions are the following:
The partitioning of surface displacement between fault slip and OFD is strongly determined by the bulk friction (material property) in the upper sedimentary layers. Plastic behavior can lead to slip dying out near the surface, with the spatial pattern of off-fault deformation being largely determined by the near-surface fault dip (fault geometry). Our results could help explain the measured spatial pattern of vertical OFD in the San Fernando earthquake. The results also could pose a testable hypothesis that there may be more OFD in Tujunga, but located farther from the fault trace on the hanging wall.
Two important caveats should be mentioned. First, the amount of OFD is sensitive to poorly constrained plastic parameters. Second, the stress drop in our model is almost certainly unrealistically high, leading to much greater slip and surface displacements than are measured in the field. We have recently implemented a fluid overpressure model that brings the absolute displacements down to more reasonable levels. Once the current models are perfected, we will start to set up our 3D models for our next funded SCEC work.
Broader Impacts Our results have implications for the behavior of faults in the near-surface volume, with consequent implications for near-source ground motion. In particular, it may be possible to predict the location of peak off-fault displacement if the near-surface fault geometry is well constrained. This result could have implications for zoning and engineering in Southern California and beyond.
This work has supported the education of Lupita Bravo, a graduate student researcher from a background historically underrepresented in earthquake science and graduate education in general.
Exemplary Figure Figure 3. Comparison of surface deformation for elastic and plastic models with low bulk friction in the top 2 km in Sylmar, with the axes zoomed into the 2 km closest to the fault trace, where the observational data in Gaudreau et al. (2023) are gathered. The large peak in vertical OFD in Tujunga is not observed this close to the fault, while the high near-fault vertical OFD in Sylmar is clear. This effect may help explain the higher vertical OFT for Sylmar in Gaudreau et al. (2023): perhaps the vertical OFD overall is larger in Tujunga, but isn’t resolvable by the current dataset.