Periodic Reporting for period 4 - FaultScan (Passive seismic scanning of the preparation phase of damaging earthquakes)
Période du rapport: 2023-12-01 au 2025-05-31
What is the problem/issue being addressed?
______________________________________
The quest for earthquake precursors is among the most vivid and active field of research in seismology. Still, our understanding of the preparation phase of earthquakes is quite limited and we are far from being able to detect and interpret robust, systematic early signs of a forthcoming large and damaging earthquake. As an example, in 2019, the magnitude 7.1 earthquake hit the Ridgecrest area in California and even though it was the largest in the United States in 20 years, the background seismicity preceding the mainshock didn’t show any signs of anomalous behavior. On more fundamental aspects, it is still debated if large earthquakes are predictable and would, for example, follow a process where slow fracture mechanisms would connect anomalously large sections of the fault before brutal rupture. On the contrary, some studies argue that large earthquakes are unpredictable and that their final size results from a stochastic cascade of smaller events without any previous connections between them. In the FaultScan project, we attempt to provide new insights to this problem by looking for signs indicating that large earthquakes rupturing over tens of kilometers are not the only consequence of random processes, but that they are preceded by a preparation phase that we will strive to detect.
Why is it important for society?
____________________________
After decades of progress led by fine seismicity analysis, satellite remote sensing (GPS, InSAR), and large-scale simulations, the quest for earthquake predictability has stalled. Here, we propose a disruptive approach to monitor active faults that is still in its early developments and thus capable of ground breaking discoveries. Even if this approach can only provide a small improvement in the diagnostic of a fault to produce a large earthquake, the relevance for society will be huge. Earthquakes are still among the deadliest natural disasters and mostly affect developing countries that do not have earthquake-resistant structures.
Moreover, the monitoring approach that we develop has the potential to be applied to other domains like volcano monitoring, C02 sequestration and geothermal production.
What are the overall objectives?
____________________________
The overall objective is to discover new signs of the preparation phase of earthquakes by using a groundbreaking seismic approach that relies on monitoring small stress perturbations in the vicinity of active faults. This technique uses seismic waves generated by powerful anthropogenic sources of noise like trains, car traffic or mining operations. We developed a special technique to extract seismic waves that dive deep down into fault zones at a few kilometers depth, making our approach highly sensitive to changes occurring on the fault. Apart from borehole instrumentation, this seismic approach is the only one that enables us to directly probe faults at depth.
______________________________________
The quest for earthquake precursors is among the most vivid and active field of research in seismology. Still, our understanding of the preparation phase of earthquakes is quite limited and we are far from being able to detect and interpret robust, systematic early signs of a forthcoming large and damaging earthquake. As an example, in 2019, the magnitude 7.1 earthquake hit the Ridgecrest area in California and even though it was the largest in the United States in 20 years, the background seismicity preceding the mainshock didn’t show any signs of anomalous behavior. On more fundamental aspects, it is still debated if large earthquakes are predictable and would, for example, follow a process where slow fracture mechanisms would connect anomalously large sections of the fault before brutal rupture. On the contrary, some studies argue that large earthquakes are unpredictable and that their final size results from a stochastic cascade of smaller events without any previous connections between them. In the FaultScan project, we attempt to provide new insights to this problem by looking for signs indicating that large earthquakes rupturing over tens of kilometers are not the only consequence of random processes, but that they are preceded by a preparation phase that we will strive to detect.
Why is it important for society?
____________________________
After decades of progress led by fine seismicity analysis, satellite remote sensing (GPS, InSAR), and large-scale simulations, the quest for earthquake predictability has stalled. Here, we propose a disruptive approach to monitor active faults that is still in its early developments and thus capable of ground breaking discoveries. Even if this approach can only provide a small improvement in the diagnostic of a fault to produce a large earthquake, the relevance for society will be huge. Earthquakes are still among the deadliest natural disasters and mostly affect developing countries that do not have earthquake-resistant structures.
Moreover, the monitoring approach that we develop has the potential to be applied to other domains like volcano monitoring, C02 sequestration and geothermal production.
What are the overall objectives?
____________________________
The overall objective is to discover new signs of the preparation phase of earthquakes by using a groundbreaking seismic approach that relies on monitoring small stress perturbations in the vicinity of active faults. This technique uses seismic waves generated by powerful anthropogenic sources of noise like trains, car traffic or mining operations. We developed a special technique to extract seismic waves that dive deep down into fault zones at a few kilometers depth, making our approach highly sensitive to changes occurring on the fault. Apart from borehole instrumentation, this seismic approach is the only one that enables us to directly probe faults at depth.
As initially planned, we purchased a pool of 470 seismic stations meant to be deployed on the field along the San Jacinto Fault right when the pandemic broke in spring 2020. Fortunately, we already had dense array data acquired in 2018 and 2019 in that region that we reprocessed together with permanent station data for the last 10 years in that same region of Anza in Southern California (see Image 1). We developed a new approach that consists in learning from the temporary dense array data and applying the result backward in time to the permanent stations. The results were unexpected and stunning: The high quality of our observations based on the seismic waves generated by the powerful freight train passages in the nearby Coachella valley allowed us to monitor the stress-state of the fault over 10 years. We found a clear anomalous transient in the beginning of 2014 that was not detected before by any other means. We interpreted this perturbation as the effect of a slow fault movement in a region that was thought to be completely locked. We further found that this type of silent movement would typically have the potential to trigger a large-scale earthquake (Sheng et al. 2022).
This monitoring method that we developed is referred to as passive ballistic-wave passive monitoring (Takano et al. 2020, Brenguier et al. 2020). It is based on detecting and characterizing the space and time distribution of specific strong anthropogenic sources that have the power to generate body-waves that travel down to a few kilometers depth close to the earthquake nucleation region. This method requires to deploy temporary dense seismic arrays in order to characterize these sources and assess the properties of the seismic wavefield that will be used for further monitoring. Following this stage, it is then possible to trace back stress perturbations on the fault using permanent seismic stations with the hypothesis that the sources of noise are stationary in time and this even before the installation of the dense array, allowing to study past earthquakes in a way that has never been done before (Sheng et al., 2022).
In April 2022, we deployed the first time ever long-term nodal seismic array with 300 stations recording seismic data for almost 3 years continuously at 10 km from the San Jacinto Fault. The analysis of the first data is fascinating. For example, we discovered a new powerful source of useful noise that seems to be associated with sedimentary basin resonance fed by the energy of vehicle traffic. We are investigating if this strong seismic energy could be used to perform noise-based monitoring of the fault. We also detected for the first time at this scale (a few km depth) changes in seismic velocities due to deformation of the Earth induced by solid earth tides. We find a very high sensitivity of seismic velocities to deformation induced by tides which we interpret as the effect of the high permeability and high fluid pressure in this region due to tectonic deformation and past earthquakes. This result will be instrumental when interpreting velocity changes associated with tectonic movements.
This monitoring method that we developed is referred to as passive ballistic-wave passive monitoring (Takano et al. 2020, Brenguier et al. 2020). It is based on detecting and characterizing the space and time distribution of specific strong anthropogenic sources that have the power to generate body-waves that travel down to a few kilometers depth close to the earthquake nucleation region. This method requires to deploy temporary dense seismic arrays in order to characterize these sources and assess the properties of the seismic wavefield that will be used for further monitoring. Following this stage, it is then possible to trace back stress perturbations on the fault using permanent seismic stations with the hypothesis that the sources of noise are stationary in time and this even before the installation of the dense array, allowing to study past earthquakes in a way that has never been done before (Sheng et al., 2022).
In April 2022, we deployed the first time ever long-term nodal seismic array with 300 stations recording seismic data for almost 3 years continuously at 10 km from the San Jacinto Fault. The analysis of the first data is fascinating. For example, we discovered a new powerful source of useful noise that seems to be associated with sedimentary basin resonance fed by the energy of vehicle traffic. We are investigating if this strong seismic energy could be used to perform noise-based monitoring of the fault. We also detected for the first time at this scale (a few km depth) changes in seismic velocities due to deformation of the Earth induced by solid earth tides. We find a very high sensitivity of seismic velocities to deformation induced by tides which we interpret as the effect of the high permeability and high fluid pressure in this region due to tectonic deformation and past earthquakes. This result will be instrumental when interpreting velocity changes associated with tectonic movements.
The monitoring method that we develop is considered groundbreaking by the community and was referenced to as an emerging topic by the Seismological Society of America (Pinzon-Rincon et al. 2021). It has the potential to make significant discoveries not only in earthquake studies, but also in volcanology and in the monitoring of C02 sequestration and geothermal production. Before the end of the project, we plan to obtain unprecedented monitoring observations along the San Jacinto Fault by deploying the first time ever long-term (3 years) dense array (300 stations). By deploying such a high number of stations for such a long time, we plan to detect smaller silent fault movements that would unveil the hidden mechanisms that prepare fault rupture. We also plan to study more in details the Ridgecrest region where a M7.1 earthquake stroke in 2019. This is a unique opportunity to monitor the preparation phase of a large earthquake that already happened thanks to the availability of permanent seismic stations in the area. Finally, we also plan to learn from a controlled fluid injection experiment in Wyoming where we deployed 150 stations. By controlling the stress perturbation generated by fluid injection, we plan to calibrate our stress-seismic velocity model which is still largely unknown.