Project

Back to overview

Development of Physics- and Statistics-Based Source Imaging and Ground Motion Modeling Methods Following a Significant Earthquake

English title Development of Physics- and Statistics-Based Source Imaging and Ground Motion Modeling Methods Following a Significant Earthquake
Applicant Dalguer Luis
Number 140459
Funding scheme Project funding
Research institution Institut für Geophysik ETH Zürich
Institution of higher education ETH Zurich - ETHZ
Main discipline Geophysics
Start/End 01.05.2012 - 30.04.2014
Approved amount 211'310.00
Show all

Keywords (6)

Ground motion modeling; Seismic hazard; Earthquake source inversion; Rupture dynamic; Earthquake source physics; Pseudo-dynamic source

Lay Summary (English)

Lead
Lay summary

During the recent destructive earthquakes such as the 2009 Mw 6.3 L’Aquila (Italy), 2010 Mw 7.1 Haiti, 2010 Mw 7.0 Darfield (New Zealand), 2011 Mw 6.3 Christchurch (New Zealand) and in particular the 2011 Mw 9.0 Tohoku earthquake (Japan), nature has reminded us that the current science, technology and methods still lack credibility in studying foreseen earthquakes and expected source and ground motion characteristics following moderate and large earthquakes. The main lesson that spreads after the occurrences of those recent earthquakes is that future research in earthquake seismology must be soundly based on PHYSICS. Reliable information from the earthquake source produces reliable information from the ground shaking features, and it is therefore vital for disaster management and risk mitigation in areas where a damaging earthquake is expected, such as in Switzerland, where we expect an earthquake of magnitude Mw ~ 6.5 or higher every 80 years. The next event of this class is expected in the Valais region within the next 40 years.

Useful practice to estimate earthquake source parameter following an earthquake is by means of kinematic source inversion of seismological and geodetic data. But the current stage of the kinematic source inversion procedure indicates that the kinematic models lack physical constraints. Therefore, the kinematic models obtained from the inversion is expected to have limitations for predicting source-dominated ground motion phenomena, especially when the model has to extrapolate beyond the range (in distance, rupture length, etc) of recorded strong motion data. In addition, they are often very ill-posed because of insufficient data, many erroneous factors can be involved in the inversion process such as Green’s function calculation, fault gridding, regularization, etc., which makes their solutions highly non-unique. In other words, multiple solutions can equally fit the data.

In the present project we propose to develop a novel procedure to develop kinematic source inversion, by invoking some of the spirit of physics-based forward dynamic rupture modeling, the so-called pseudo-dynamic source inversion. This new procedure will improve the finite source inversion by adopting more physics-based source inversion schemes, physically regularizing our model space with constraints inferred from rupture dynamics and previous source inversion results, instead of fully relying on insufficient data. We will explore the possibility that we can implement the dynamic constraints in the prior distribution in the Bayesian inversion in the form of covariance matrix. The new inversion procedure will also introduce the concept of earthquake source statistics, i.e., the source process is characterized with a set of random variables (i.e., random field) assigned to each type of kinematic source parameters. In this way the inverse problem can still be formulated in the kinematic domain but the model space can be regularized by both physical and observational constraints. Then the resulted pseudo-dynamic source model will be used to estimate ground motion parameters.

In this project we aspire to automate our new methodology to respond shortly after the occurrence of a significant earthquake and be able to provide preliminary descriptions of the source image and near source ground motion characteristics. Our product that uses observational and physical constraints is oriented to contribute to the improvement of our capability for earthquake damage assessment and seismic risk mitigation following an earthquake, and it will serve as the basis for future development of real-time and early warning systems, as such, this project reaches several fronts of the earthquake scientific and professional communities as well as government and general public, as it is situated at the interface of earthquake science, earthquake engineering, decision makers and media. 

Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Name Institute

Publications

Publication
Pseudo-dynamic source modelling with 1-point and 2-point statistics of earthquake source parameters
Song Seok Goo, Dalguer Luis A., Mai P. Martin (2014), Pseudo-dynamic source modelling with 1-point and 2-point statistics of earthquake source parameters, in GEOPHYSICAL JOURNAL INTERNATIONAL, 196(3), 1770-1786.
Importance of 1-point statistics in earthquake source modelling for ground motion simulation
Song S. and L.A. Dalguer (2013), Importance of 1-point statistics in earthquake source modelling for ground motion simulation, in Geophys. J. Int., 192, 1255-1270.

Collaboration

Group / person Country
Types of collaboration
The Southern California Earthquake Center (SCEC) United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
Swiss Seismological Service (SED), ETH-Zurich Switzerland (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Research Infrastructure
Université Joseph Fourier, Grenoble France (Europe)
- in-depth/constructive exchanges on approaches, methods or results

Abstract

During the recent destructive earthquakes such as the 2009 Mw 6.3 L’Aquila (Italy), 2010 Mw 7.1 Haiti, 2010 Mw 7.0 Darfield (New Zealand), 2011 Mw 6.3 Christchurch (New Zealand) and in particular the 2011 Mw 9.0 Tohoku earthquake (Japan), nature has reminded us that the current science, technology and methods still lack credibility in studying foreseen earthquakes and expected source and ground motion characteristics following moderate and large earthquakes. The main lesson that spreads after the occurrences of those recent earthquakes is that future research in earthquake seismology must be soundly based on PHYSICS. Reliable information from the earthquake source produces reliable information from the ground shaking features, and it is therefore vital for disaster management and risk mitigation in areas where a damaging earthquake is expected, such as in Switzerland, where we expect an earthquake of magnitude Mw ~ 6.5 or higher every 80 years. The next event of this class is expected in the Valais region within the next 40 years. Useful practice to estimate earthquake source parameter following an earthquake is by means of kinematic source inversion of seismological and geodetic data. But the current stage of the kinematic source inversion procedure indicates that the kinematic models lack physical constraints. Therefore, the kinematic models obtained from the inversion is expected to have limitations for predicting source-dominated ground motion phenomena, especially when the model has to extrapolate beyond the range (in distance, rupture length, etc) of recorded strong motion data. In addition, they are often very ill-posed because of insufficient data, many erroneous factors can be involved in the inversion process such as Green’s function calculation, fault gridding, regularization, etc., which makes their solutions highly non-unique. In other words, multiple solutions can equally fit the data.In the present project we propose to develop a novel procedure to develop kinematic source inversion, by invoking some of the spirit of physics-based forward dynamic rupture modeling, the so-called pseudo-dynamic source inversion. This new procedure will improve the finite source inversion by adopting more physics-based source inversion schemes, physically regularizing our model space with constraints inferred from rupture dynamics and previous source inversion results, instead of fully relying on insufficient data. We will explore the possibility that we can implement the dynamic constraints in the prior distribution in the Bayesian inversion in the form of covariance matrix. The new inversion procedure will also introduce the concept of earthquake source statistics, i.e., the source process is characterized with a set of random variables (i.e., random field) assigned to each type of kinematic source parameters. In this way the inverse problem can still be formulated in the kinematic domain but the model space can be regularized by both physical and observational constraints. Then the resulted pseudo-dynamic source model will be used to estimate ground motion parameters.In this project we aspire to automate our new methodology to respond shortly after the occurrence of a significant earthquake and be able to provide preliminary descriptions of the source image and near source ground motion characteristics. Our product that uses observational and physical constraints is oriented to contribute to the improvement of our capability for earthquake damage assessment and seismic risk mitigation following an earthquake, and it will serve as the basis for future development of real-time and early warning systems, as such, this project reaches several fronts of the earthquake scientific and professional communities as well as government and general public, as it is situated at the interface of earthquake science, earthquake engineering, decision makers and media.
-