Monitoring real faults towards their critical state

Project Information

MONIFAULTS

Grant agreement ID: 802777

DOI

10.3030/802777

Project closed

EC signature date 2 October 2018

Start date 1 January 2019

End date 31 December 2023

Funded under

• EXCELLENT SCIENCE - European Research Council (ERC)

Total cost

€ 1 393 174,00

EU contribution

€ 1 393 174,00

Coordinated by

UNIVERSITA DEGLI STUDI DI PADOVA
Italy

Periodic Reporting for period 4 - MONIFAULTS (Monitoring real faults towards their critical state)

Reporting period: 2022-10-01 to 2023-12-31

Summary of the context and overall objectives of the project

This project aims to characterize the state of stress and physical properties of the seismogenic part of the Earth's crust using seismological data, focusing on periods close to large earthquakes and during aseismic deformations. The results enhance understanding of the physical processes in faults near significant earthquakes, providing new insights for improved modeling and predictive systems.

Primary objectives include providing a comprehensive view of the state of stress, deriving other physical properties of the crust near major faults, and illuminating the physical processes occurring close to large earthquakes and during aseismic deformation. Methodologies developed are applied to faults along the central Apennines in Italy, where significant earthquakes pose a major hazard. This well-instrumented but insufficiently studied region offers potential for new and diverse observations about the physical processes in continental normal faults.

The project has significantly advanced the detailed study of seismicity by integrating seismological data with new observables related to rock changes at depth in response to external forces. By characterizing the state of stress and physical properties of the seismogenic part of the Earth's crust, particularly during periods close to large earthquakes and during aseismic deformations, we have gained crucial insights into the physical processes in faults.

The methodologies were specifically applied to the faults along the central Apennines in Italy, a region of significant seismic hazard. Results from this well-instrumented region provide valuable data illuminating the physical processes near major faults, both during and between seismic events. This comprehensive approach furthers our knowledge of stress states and crustal properties near major faults, offering a pathway to better predictive capabilities for earthquake hazards.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The expected working pace was strongly influenced by the COVID-19. Most of the PhD students and post-docs began their work just a few months before the pandemic struck. During this critical initial phase of research, they found themselves in a new country, confined to their homes, with only video interactions with me and among themselves. This situation significantly affected their morale and limited opportunities for discussing how to advance and perform practical analyses. Despite these challenges, we focused on the development and preliminary application of methodologies to advance the project during the first part:

Development of Coherence Analysis and Clustering: We implemented and successfully applied coherence analysis and clustering of resuming features with unsupervised machine learning to the region of the L'Aquila 2009 magnitude 6.1 earthquake (Shi et al., 2020, JGR).

Algorithm for Seismological Data Analysis: We developed an algorithm to extend the analysis of seismological data related to the fault generating the L'Aquila earthquake, enabling the reconstruction of seismological observations from 1990 to 2021. This 31-year dataset will help infer the state of stress in the crust before and after the 2009 earthquake. Methodology development issues were addressed, and results have been published (Majstorovic et al., 2021, 2023).

High-Resolution Seismic Catalog: We created a high-resolution catalog of seismic signals for the six months preceding the L'Aquila earthquake to track stress evolution leading up to the main earthquake (Cabrera et al., JGR, 2022, Cabrera & Poli, GRL, 2023).

Study of Nucleation of Small Magnitude Earthquakes: Using the methodology from (3), we studied the nucleation of a small magnitude 4 normal fault earthquake, highlighting the complex nature of pre- and post-seismic processes (Sanchez-Reyes, 2020, SRL).

High-Resolution Seismic Catalog for Alto Tiberina Fault: We completed an unprecedented high-resolution seismic catalog containing over 400,000 events spanning four years of data in the Alto Tiberina low-angle normal fault. We are beginning to use this information to derive insights about physical processes and the state of stress in this fault system (Essing et al., JGR, 2022, 2024).

Characterization of Velocity Variations: We characterized velocity variations in the region of the L'Aquila 2009 earthquakes using ambient seismic noise. This study focused on the susceptibility of velocity changes to different stress forcing (e.g. periodic deformation, earthquakes), providing important information about the layering of physical properties in the crust for this fault system (Poli et al., 2020, JGR). The same method was applied to the seismic region in southern Apennines (Mikhael et al., 2024).

This project has made substantial progress in understanding the dynamics of the Earth's crust in seismically active regions. By focusing on the central Apennines, a region of both high seismic risk and significant potential for new discoveries, we have developed methodologies that offer deeper insights into fault mechanics and stress states. This research represents a crucial step towards a more effective understanding of the earthquake cycle in slowly deforming regions. The results obtained have been disseminated through scientific publications, conferences, and collaboration with other research institutions, ensuring that the findings are utilized to advance the field of seismology and earthquake hazard mitigation. The research of this project also brought to the new definition of the preparation of earthquakes published in Martinez-Garcon & Poli (2024).

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

For the first time we demonstrated the possibility of isolating different parts of the seismic cycle using unsupervised machine learning applied to continuously recorded seismological data (She et al., 2020).

We obtained a detailed picture of the strain response of crustal rocks in the L'Aquila earthquake region, highlighting a strong stress susceptibility for rocks at the depth of the 2009 earthquake's nucleation (Poli et al., 2020).

We developed a method that requires only a single seismic station to obtain information about near-fault seismicity. This method provided a long-term observation of seismic events in the L'Aquila region over 30 years, tracking the time-dependent evolution of seismicity in response to dynamic and quasi-dynamic strain evolution (Majstorovic et al., 2020). Additionally, we explored machine learning algorithms and visualized the black box processing in convolutional neural networks.

We tracked stress evolution before the L'Aquila earthquake (Cabrera et al., 2021).

Interpreting convolutional neural network decisions for earthquake detection using feature map visualization, backward optimization, and layer-wise relevance propagation methods (Majstorović et al., 2023).

Highlighting the struggled rupture initiation of the Mw 6.1 2009 L'Aquila earthquake (Cabrera et al., 2023).

Observing enhanced tidal sensitivity of seismicity before the 2019 magnitude 7.1 Ridgecrest, California earthquake (Beaucé et al., 2023).

Unraveling earthquake clusters in the 2014 Alto Tiberina earthquake swarm via unsupervised learning (Essing et al., 2023).

Documenting non-linear seismic velocity variations during a seismic swarm in the Alto Tiberina low-angle normal fault from ambient noise correlation measurements (Mikhael et al., 2023).

Demonstrating that cascade and pre-slip models oversimplify the complexity of earthquake preparation in nature (Martínez-Garzón et al., 2024).

Summary of unsupervised learning of feature for earthquake cycle exploration