Observation et simulation du mouvement sismique en champ proche sur des milieux urbains – E-CITY

1. NEAR-FAULT SIMULATION CONSIDERING FAULT MECHANICS

The understanding of earthquake dynamics has progressed, especially thanks to the near-field seismological data and geodetic observations of the recent crustal earthquakes such as the 2016 Central Italy sequences, 2016 Kaikoura earthquake in New Zealand, 2017 Kumamoto in Japan, 2019 Ridgecrest in California, and others (Urata et al., 2017; Ulrich et al., 2019; Zhang et al., 2020). However, the near field ground motions are still difficult to understand (reproduce) without comprehending the mechanics behind, in particular, for our interest of this project, (1) initial state and boundary conditions, (2) nonlinearity in the source area (other than the site conditions) and (3) fault geometry and complex seismic wave radiation. We propose to improve these elements in physics-based simulations (PBS) both from the theoretical advances and the analysis of the observational data. We keep a macroscopic vision for seismic hazard application, without always solving the microscopic mechanisms of the phenomena.

2. TEMPORAL CHANGES OF MATERIAL PROPERTIES FROM EARTHQUAKE DATA

One of the main issues is the assessment of seismic response to low/high amplitude ground shaking. We need to identify why the geological material behaves differently for weak and strong ground motion, specially its in-situ spatial and temporal variability. A main issue is the use of Vs30 (mean shear wave velocity in the first 30 m) to characterize site response. We know that 30 m is not enough to portray all the complexities of sites located in deep basins, and whether we expect linear or nonlinear site effects using this single value. We measure the frequency-dependent velocity changes, which give an idea of the depth where nonlinear material behavior takes place. We take advantage of the relatively large number of KiK-net and KNET stations in Japan to compute the related velocity changes during earthquake activity.

3. CHARACTERIZATION OF THE URBAN GROUND MOTION FOR SEISMIC RISK ANALYSIS

The properties of locally resonant meta-materials (in our case, buildings being the urban meta-material) are governed by interferences between incident and scattered waves, which can lead to hybrid wavefields (Lemoult et al., 2013); a similar process expected in site-city interaction. At first order, the seismic behavior of structures is controlled by their resonant frequency and damping ratio. Under strong ground shaking, the co-seismic variation of resonance frequencies is linked to possible structural damage, a key issue to estimate the vulnerability of structure. Tracking the variation of damping under fast-dynamic loading (in addition to the resonant frequency) is an essential challenge to understanding the seismic vulnerability of structures, especially in the near-fault condition. We then propose the analysis of the contribution of structure motion on ground motion as urban meta-material.

WP1 concentrates in the development of computing the stresses in the media before the dynamic triggering of the earthquake. This drives how the earthquake begins by finding the initial conditions that the system holds.

WP2 implemented several robust signal processing methods to extract the coseismic linear and nonlinear response of the media. These methods are general to detect temporal and frequency changes due to cyclic loads, in particular earthquakes. They have been used to detect and quantify the nonlinear signature of the seismic stations in the whole country of Japan.

WP3 results provide a framework to quantify urban-like attenuation and scattering from ambient noise, directly relevant to seismic hazard assessment in cities. Results also emphasize the importance of instrumenting buildings in weak-to-moderate seismic regions to: (1) develop and calibrate realistic models of structural response; (2) improve understanding of uncertainties in risk assessment; and (3) investigate physical processes activated during seismic loading.

WP1. Numerical simulations of the dynamic rupture need to know the initial stresses of the crust. With the new method we can now compute better estimates of the rupture of seismic scenarios. Something that still needs to be done.

WP2. We are able to track temporal changes in the shallow crust produced by cyclic loads. This means the variability of the strength of the material, its shear modulus. Yet, we still need to identify and quantify the changes produced in the material damping. This is the next step for a fully characterization of the nonlinear response of the media.

WP3. Future work will apply and scale this approach to urban areas, integrating it with numerical models and earthquake observations to better constrain site–city interaction effects on ground motion. Future work will also include population-based classification of structural health by combining vibration- and earthquake-based indicators of these physical processes.

Ce projet va au-delà de la pratique actuelle d'évaluation du risque sismique, en intégrant les progrès récents réalisés dans le traitement des données et la modélisation numérique. Les principales questions scientifiques portent sur l’effet: (1) de la variabilité spatiale du mouvement du sol en champ proche sur la réponse des sols et des structures; (2) de l’endommagement et le changement des propriétés du milieu; et (3) de la contribution de groupes de bâtiments à la variabilité spatiale du mouvement du sol en milieu urbain. Nous étudions les tremblements de terre de Kumamoto (Japon) en 2016 et de Ridgecrest (Californie) en 2019 compte tenu du grand nombre d'enregistrements à proximité de la zone épicentrale. Comme exemple d'application, nous ciblons l'évaluation du risque sismique d’une partie de la ville de Quito, en Équateur, située dans un bassin construit sur une faille inverse active. Ce projet répond aux besoins du risque sismique à faible probabilité dans zones urbaines proches aux failles actives.