Exploring the Dynamic Interactions Between the Southern San Andreas Fault and a Normal Fault Under the Salton Sea
(JGR Abstract) We investigate the dynamic interactions between the Southern San Andreas Fault (SSAF) and a proximal normal fault (NF) beneath the Salton Sea in southern California. The NF, positioned near the SSAF terminus at Bombay Beach, exhibits 11–15 displacement events across 14 stratigraphic sequences, with a range of 0.2–1.4 m of vertical offset since ∼2–3 ka. Notably, four of these events may align temporally with SSAF earthquakes, raising questions about the possible interplay between the two faults. Utilizing dynamic rupture models, we analyze the coseismic interactions between the SSAF and NF, addressing under what conditions the SSAF induces slip on the NF. Our findings reveal that a suite of SSAF ruptures, particularly those propagating from north to south, can trigger slip on the normal fault and replicate observed vertical offsets. If the SSAF extends beneath the Salton Sea, earthquakes originating south of the NF intersection are less likely to trigger normal fault slip, although we cannot exclude this possibility. Some SSAF ruptures do not trigger discernible slip on the NF, rendering such events undetectable in the stratigraphic record. Our research contributes toward discussions regarding the seismic hazard in southern California, shedding light on the interplay between the SSAF and NF.
Direct Dynamic Triggering Scenarios of the Southern San Andreas Fault by Moderate Magnitude Cross Fault Earthquakes in the Brawley Seismic Zone, CA
(SRL Abstract) Intersections between small faults and larger faults are ubiquitous throughout the world, including the strike-slip San Andreas system in Southern California. In particular, orthogonal intersections may exist in the Brawley Seismic Zone (BSZ) in the Salton Sea region between small left-lateral strike-slip faults and the main Southern San Andreas Fault (SSAF). This area often experiences earthquake swarms, and poses the question of whether moderate earthquakes on these left-lateral cross-faults may propagate to the nearby SSAF, triggering a large, damaging event. To address this question, we present a collection of dynamic rupture scenarios describing the interaction of a representative cross-fault intersecting the highly prestressed SSAF in the BSZ. Our models span a variety of cross-fault earthquake rupture scenarios that vary in magnitude (~Mw 5.2 to 6.1), rupture depth, location, and directivity to test their potential to trigger the SSAF. We use our models to investigate how the above parameters play an interconnected role in developing ruptures that might trigger the SSAF. Our results highlight that adjacency to the SSAF and shallow rupture enhance the ability of moderate-size cross-fault earthquakes to pass on the SSAF. We also show that earthquakes starting at the opposite edge of the cross-fault from the intersection are less likely to trigger the SSAF unless they propagate over almost the entirety of the cross-fault area. Our experiments provide for the first time a benchmark of comparison and insights into rupture parameters that might control the initiation of a significant SSAF event. They may also give insight into the general interactions of small faults with larger intersecting faults, such as in the case of the recent 2023 Kahramanmaraş, Turkey, Earthquake.
(Plain Language Summary from GRL ) Characterizing earthquake stress drops is important for both understanding earthquake processes as well as assessing seismic hazards. Estimating stress drops for earthquakes often involves a non-dimensional parameter C, which characterizes the effective elastic stiffness of the faulting system. In this study, we investigate how interactions between the Earth's surface and the fault affect the theoretical C value, which has not been systematically studied. We find that C decreases with a smaller earthquake depth, and its depth-dependence exceeds the general “rule of thumb.” Seismological studies for stress drop commonly assume C is a constant. With the SRCMOD earthquake source model catalog (Mai & Thingbaijam, 2014, https://doi.org/10.1785/0119990126; Thingbaijam et al., 2017, https://doi.org/10.1785/0120150291),we show that it is necessary to consider the depth-dependence of C when measuring stress drops for large earthquakes; otherwise, with a “constant C” assumption, the estimated stress drops would appear to be magnitude-dependent. Since C factor plays an important role in many theories that relate earthquake sources to their ground motion, our results here may help improve our current seismic and tsunami hazard assessments.
(Geosphere abstract) We use three-dimensional (3-D) dynamic finite-element models to investigate potential rupture paths of earthquakes propagating along faults through the western San Gorgonio Pass, a structurally complex region along the San Andreas fault system in southern California (USA). We focus on the right-lateral San Bernardino strand of the San Andreas fault system, the oblique thrust–right-lateral San Gorgonio Pass fault zone, and a portion of the right-lateral Garnet Hill strand of the San Andreas fault system. We use the 3-D finite-element method to model rupture propagation along a fault geometry that reflects current understanding of the local geometrical complexity and is consistent with long-term loading and observed surface deformation. We test three different types of pre-stress assumptions: (1) constant tractions (assuming pure right-lateral strike-slip motion on the San Bernardino and Garnet Hill strands and oblique thrust–right-lateral strike-slip motion on the San Gorgonio Pass fault zone), (2) a uniform regional stress regime, and (3) long-term (evolved) stress from quasi-static crustal deformation modeling. Our results imply that under the more realistic regional stress and evolved stress assumptions, throughgoing rupture propagation from the southeast to northwest (i.e., from the Garnet Hill to the San Bernardino strand) may be more likely than throughgoing rupture in the reverse direction (from the San Bernardino to the Garnet Hill strand). The results may have implications for the earthquake potential in the region as well as for ground motion in the Los Angeles Basin. The results also emphasize how fault geometry and stress patterns combine to influence rupture propagation on complex fault systems.
(GRL abstract) We use 3D dynamic rupture simulations to discover a previously un‐described effect of asymmetric topography on the earthquake process. With the Cajon Pass along the San Andreas Fault as an example, we find that asymmetric topography generates an alternating normal stress perturbation around the rupture front, near the free surface. When topography lies to the right of the propagating right‐lateral front, the normal stress perturbation is clamping ahead of the rupture front and unclamping behind. When topography alternates to the left, the perturbation reverses sign. The process is analogous to the normal stress variations on dip‐slip faults. While this effect does not strongly affect rupture propagation and slip in our current parametrization, it requires explanation and exploration. An understanding of the normal stress perturbation due to asymmetric topography will help prevent its misattribution to other sources and lead to a better understanding of the interplay of multiple processes during earthquakes.
Nucleation location (1) north of the Cajon Pass (see figure with map)
Nucleation location (2) south of the Cajon Pass (see figure with map)
Explanation of the topographic asymmetry effect using a synthetic topographic model. Normal stress changes for simulation with nucleation north of the Cajon Pass. Panels (a), (b) show normal stress changes at t = 10 s and t = 45 s respectively. Topography highlighted by dotted lines corresponds to mountains on the “near” (San Gabriels) and “far” side (San Bernardinos) of the fault viewer (see also Figure S2). (c) Shear and yield stress changes (Normal stress * 0.6) along a profile from -120 to 120 km at 50 m depth. Rupturable area extends from -100 to 100 km along strike. Dashed circles mark the area of topographic normal stress change near the rupture front. (d) Schematic representation of a right lateral strike slip fault with asymmetric topography to the right (1) and to the left (2) with respect to an observed moving with the rupture front.
Recent studies based on paleoseismological trenches have indicated that the two most recent events along the Southern San Andreas Fault (SAF) are indistinguishable (in time) from ruptures on Imperial Fault (IF) (Philibosian et 2011, Meltzner et al., 2015). Thus, it is very probable that these two major faults could rupture at the same time, a catastrophic scenario for Southern California. I’m using the 3D finite element code FaultMod (Barall, 2009) in order to run a series of experiments that will help me investigate the sensitivity of rupture propagation to nucleation location, and the partitioning of slip across the faults. The preliminary results indicate that if there are through-going fault structures between the main faults, rupture may propagate between the SAF and IF, and even across cross-faults connecting the two structures. Furthermore, I recently extended my research into new important questions that include the crucial issue of whether small “swarm-type” earthquakes may cascade into full-sized earthquakes on the San Andreas/Imperial Fault system. This investigation sheds new light on the implications for fault rupture, ground motion, and seismic hazard in the region. Results have been presented by at SCEC 2015, 2016 (poster) and SSA 2016 (oral) annual meeting.
Example of final slip model for dynamic rupture models with cross-faults. Map Data: Google, Image Landsat, Copernicus, INEGI.
(Based on Kyriakopoulos et al., 2019, JGR)
Map view (top) and 3D view (bottom) of rupture propagation across the Brawley Seismic Zone (BSZ). Map Data: Google, Image Landsat, Copernicus, INEGI.
Although the dynamic behavior of single faults is well studied, the rupture propagation across non-planar complex fault geometries remains a matter of much discussion. This specific topic is of great importance since the ability of rupture to overcome mechanical obstacles and propagate over long distances separates small from large events. I recently started working on the development of dynamic rupture models, which allow me to investigate better the physics of fault interaction. The M7.2 2010 El Mayor-Cucapah, Mexico earthquake is a multi-fault event, possibly similar to a future large event along one of the main faults in southern California (SoCAL). For that reason, the study of this event helps us to shed light on the mechanical behaviour of SoCAL faults as either individual segments or part of a more complex system. I developed a series of finite elements models useful for dynamic rupture simulations and InSAR based geodetic models that give us insights into both the kinematics and physics of the event. Results from these work are presented in Kyriakopoulos et al., 2017, JGR.
(Figure 2 in Kyriakopoulos et al., 2017, JGR)
Mesh of the finite element model and fault geometry. (a) View of entire finite element mesh. (b) Close‐up view of the mesh around the fault interface. (c) Fault segments used to generate the underlying continuous fault interface. MapData: Landsat/Copernicus,SIO,NOAA, U.S. Navy,NGS,GEBCO,LDEO-Columbia,NSF
*Note: my figure below was selected for the cover of the Journal of Geophysical Research, Solid Earth, Vol.122, issue 12, Dec2017 (JGR Cover)
(Figure 4 in Kyriakopoulos et al., 2017, JGR)
Dynamic rupture simulation, at time = 2, 12, 20, and 40 s. The subplots (A, B, C and D) include panels with the time evolution of (starting from the top left and moving clockwise) fault slip (m), normal stress (MPa), shear stress (MPa), and fault slip rate (m/s). The view is approximately from the NE to the SW. MapData: Landsat/Copernicus,SIO,NOAA, U.S. Navy,NGS,GEBCO,LDEO-Columbia,NSF
In my experience, the study and analysis of subduction zones presents two (or at least two) main challenges. First, the seismogenic segment of the system is partially or entirely located offshore, limiting the effectiveness of land-based geodetic observations and resulting in a dramatic decrease of the data resolving power. Second, the subduction zone environment is a complex physical system that includes strong rheological transitions and evolves through different styles of tectonic deformation (e.g. coseismic rupture, creep along the interface at depth), making it very difficult to represent even with numerical methods and computer simulations. For the above reasons, basic parameters of the system are not currently well understood. However, recent advancements and new data from offshore geodetic sites allow us to take a step forward in our knowledge.
The subduction zones are among the most complex tectonic environments on our planet. Knowing how these systems work is fundamental for our society, because the majority of the seismic moment (~90%) in our planet is released through large earthquakes (megathrust events) occurring in the shallower part of these boundaries. The Nicoya peninsula in Costa Rica, is a unique case as far as regards the possibility to constraint the evolution of fault slip across the different phases of the seismic cycle using modern geodetic methods. The "uniqueness" of this place is...Another great opportunity offered by the tectonic setting around Nicoya is the
(Figure 7 in Kyriakopoulos et al., 2015, JGR)
Regional physiography and surface projection of the slab model with 4 km contours. The offshore bathymetric surface reveals the transition from a “rougher” Quepos plateau to “smoother” seafloor offshore Nicoya and Nicaragua, where Cocos Nazca Spreading Center crust (CNS‐1, CNS‐2) transitions to East Pacific Rise (EPR) generated crust.
(Figure 1 in Kyriakopoulos et al., 2016, JGR)
Three-dimensional view of the Nicoya Peninsula in Costa Rica and the subduction of the the Cocos plate beneath the Caribbean plate. A transparency window reveals the topography of the slab surface (gray mesh with white grid lines) beneath the Nicoya Peninsula. Continuous and campaign GPS stations used in the inversions for locking and coseismic slip are shown with yellow and red triangles, respectively.
(Figure 4 in Kyriakopoulos et al., 2015, JGR)
New 3‐D slab interface model, with resolution and accuracy maps. (a) The resolution is determined to be the radius of the maximum seismicity window used for determining the slab surface (minimum fixed at 5 km). (b) The model accuracy (standard deviation) is shown from 50 randomly resampled bootstrap runs of the smoothed model interface.
(Figure 5 in Kyriakopoulos et al., 2016, JGR)
Comparison between (a) two‐ and (b) three‐dimensional interface models of locking for trench‐normal and vertical interseismic deformation. Thin black contours show the interface geometries from Feng et al. [2012] and Kyriakopoulos et al. [2015] for the two‐ and three‐dimensional models, respectively.
(Figure 7 in Kyriakopoulos et al., 2016, JGR)
Comparison between (a) two‐ and (b) three‐dimensional interface models of locking for trench‐normal and vertical interseismic deformation. Thin black contours show the interface geometries from Feng et al. [2012] and Kyriakopoulos et al. [2015] for the two‐ and three‐dimensional models, respectively.
The M9 2011 Tohoku-oki earthquake has given us the unprecedented opportunity to integrate the study of this megathrust event with offshore geodetic sensors of movement [Kato et al., 2011], positioned right above the main rupture. Although only five in number, these offshore sensors have thrown new “light” on basic mechanical parameters governing the fault kinematics across the seismogenic zone. In order to take advantage of the new data I developed one of the most comprehensive existing 3D finite element model of the Japan subduction zone (Honshu Arc) that hosted the 2011 catastrophic event (Kyriakopoulos et al., 2013b). The model implements several geophysical features of the Honshu Arc including the main geological bodies with representative material properties (e.g. forearc, volcanic arc), non-planar fault interface, topography and realistic bathymetry of the oceanic seafloor. I combined land-based (700 GPS stations) and offshore (5 sensors) geodetic measurements with my model to estimate the coseismic slip distribution and investigate the sensitivity of the system to basic mechanical properties. From my results I found that, especially in the shallower part of the plate interface, the overall kinematics of the fault system is severely affected by the rigidity contrast between the overriding (hangwall) and underthrusting (footwall) plate. Specifically, the new model shows a variation of fault slip of up to 25% compared to a homogeneous model. My results illuminated the asymmetric slip across a fault that separates materials with contrasting elastic properties and demonstrated the sensitivity of inverse analyses of geodetic data to the simulated domain geometry and configuration of material properties.
(Figure 4 in Kyriakopoulos et al., 2013, JGR)
Coseismic slip based on the geodetic inversion of 700 on-land plus 5 offshore geodetic sites. (a, b) Slip distribution for the heterogeneous (JHET) model and the homogeneous (JHOT) model, respectively. (c) Differences between JHOT and JHET models (Diff = JHOT – JHET ).
(Figure 2 in Kyriakopoulos et al., 2013, JGR)
Finite element (FE) implementation of the Honshu Arc subduction zone, Japan. (a) Mechanical components: forearc (yellow), volcanic arc (dark green), backarc (light green), lithosphere (red), crust (blue), mantle (gray). The Young modulus and Poisson ratio values are fixed for each geologic entity. (b) A combination of topography and bathymetry (ETOPO1 project) describes the free surface of the model. The dimensions of the central sector of the model are 1300 km × 1300 km × 600 km. (c, d) 3-D section of the FE model illuminates the rigidity contrasts between the upper crust and the down-going slab. (e) Mesh grid of the model. For visual clarity we hide element edges in (a),(b),(c), and (d).
My work on seismic swarms is oriented to decipher the mechanisms that govern their occurrence and evolution in time (along known faults) through detailed observations of surface deformation. In order to do this, I use space geodetic deformation maps derived from InSAR. The key role of the observations is that they allow us to measure both the seismic and aseismic deformation of these sequences. I successfully applied this technique (using data from the European Space Agency satellite ENVISAT) combined with standard modelling methods and seismological observations to an earthquake swarm occurred in the Messinia prefecture (Jul-Oct, 2011), Southwest Greece (Kyriakopoulos et al., 2013a). From my results I was able to follow the monthly migration of the swarm (see panel A) and estimate the total geodetic moment released each month (panel B). I found that a significant amount of the total elastic energy was released from aseismic movement along the normal fault activated during the swarm. I concluded that the swarm was the “symptom” of a wider creep episode occurring in the region. The aseismic release of energy was probably a key factor for this sequence not culminating in a larger and potentially more dangerous event (as for example in the l’ Aquila case). Further investigation is required on the role of the aseismic component and the energetic budget of these sequences. Moreover, we want to know how measuring aseismic deformation with space techniques could lead us to a more informed evaluation of seismic swarm activity.