Дисертації з теми "Earthquake generator"

La France métropolitaine est éloignée des principales limites de plaques tectoniques. De ce fait, l'origine de la sismicité est plus difficile à comprendre que dans les régions actives localisées le long des limites de plaques. De plus, les données sismiques disponibles (séismes, accumulation des contraintes dans le sol, failles, mécanismes au foyer, etc.) y sont moins nombreuses. Ces observations sont deux des principaux freins rencontrés lorsque l'on souhaite caractériser l'aléa sismique sur un territoire à sismicité faible à modérée. L'approche proposée dans cette thèse s'appuie sur un générateur stochastique produisant des séismes synthétiques en combinant des observations de terrain et des connaissances théoriques sur la sismicité du territoire d'étude. Ce générateur de séismes se divise en trois parties : (i) le tirage temporel des séismes principaux, suivi du (ii) tirage de leur localisation conditionné par la magnitude et enfin de (iii) la production de répliques associées aux séismes principaux. Le tirage temporel est réalisé à l'aide d'une distribution fréquence – magnitude (DFM) calculée sur l'intégralité du territoire d'étude via la méthode non paramétrique des temps de latence. Ce calcul à l'échelle du territoire entier permet d'augmenter le nombre de données, de diminuer les périodes de retour et donc d'estimer les fréquences de chaque magnitude directement à partir des données. Une attention particulière a été portée à la méthode des temps de latence afin de pallier la faible quantité de données observées notamment pour les fortes magnitudes. Le tirage spatial des séismes principaux nécessite, quant à lui, une régionalisation et une densité spatiale caractérisant la sismicité. La régionalisation permet de limiter spatialement la magnitude maximale autorisée des séismes synthétiques. La localisation d'un séisme synthétique d'une magnitude donnée est tirée dans la densité spatiale uniquement au sein des régions autorisant cette magnitude. Les répliques sont ensuite générées autour des séismes principaux à partir de la loi de Bath et de la distribution proportion – magnitude des répliques. L'aléa associé à chaque séisme (principal et réplique) généré est calculé à l'aide de modèles de prédiction des mouvements du sol pondérés. La pondération est dépendante de la magnitude du séisme et de la distance séisme – site. Cette pondération s'effectue en utilisant la base de données européenne d'accélérations du sol RESORCE. Enfin, l'observation en un site donné des aléas produits par les séismes synthétiques sur un million d'années permet d'estimer les probabilités annuelles de dépassement de seuils d'aléa donnés
French mainland seismicity is considered as low to moderate due to its remoteness from tectonic plate boundaries. A first consequence is that the origins of its seismicity are harder to understand than in active regions close to tectonic plate boundaries. Another consequence is the lack of available data (earthquakes recorded but also strain rates, active faults...). These two observation make difficult to estimate seismic hazard in low-to-moderate seismic areas. The proposed approach is to generate synthetic earthquakes by combining observation and theoretical knowledge on the seismicity of the studied territory. This generator is based on a three-step scheme: (i) the temporal draw of main shocks, (ii) their spatial draw conditioned by magnitude and finally (iii) the generation of aftershocks they produce. The temporal step needs a recurrence rate. Past seismicity of the whole studied area is analysed thanks to the non-parametric inter event time method in order to obtain this wished recurrence rate. Computing the recurrence rate at the whole territory scale allows to keep the maximum quantity of data, reduce the return periods and so estimate main shock frequencies directly form observed data for each magnitude. An implementation has been developed to overcome the accuracy fall of the inter event time method observed when data are sparse. The spatial step needs a regionalization and a spatial density representing seismicity. The regionalization allows maximum magnitude limitation in space: each region is characterized by an allowed maximal magnitude. Location of a synthetic earthquake with a given magnitude is drawn in the spatial density only within regions that allow this magnitude. Aftershocks are generated around main shocks thanks to the Bath law and the proportion – magnitude distribution of aftershocks. The seismic hazard produced by each of the generated earthquakes (main shocks and aftershocks) is computed thanks to a set of weighted Ground-Motion Prediction Equations. The weights are obtained as function of magnitude and distance thanks to The European ground-motion database RESORCE. Finally, from direct observation of the seismic hazard produced by synthetic earthquakes over one million years, annual probabilities of exceedance can be calculated with ease

Quantitative and qualitative assessment of the seismic risk to highway bridges is crucial in pre-earthquake planning, and post-earthquake response of transportation systems. Such assessments provide valuable knowledge about a number of principal effects of earthquakes such as traffic disruption of the overall highway system, impact on the regions' economy and post-earthquake response and recovery, and more recently serve as measures to quantify resilience. Unlike previous work, this study captures unique bridge design attributes specific to California bridge classes along with their evolution over three significant design eras, separated by the historic 1971 San Fernando and 1989 Loma Prieta earthquakes (these events affected changes in bridge seismic design philosophy). This research developed next-generation fragility curves for four multispan concrete bridge classes by synthesizing new knowledge and emerging modeling capabilities, and by closely coordinating new and ongoing national research initiatives with expertise from bridge designers. A multi-phase framework was developed for generating fragility curves, which provides decision makers with essential tools for emergency response, design, planning, policy support, and maximizing investments in bridge retrofit. This framework encompasses generational changes in bridge design and construction details. Parameterized high-fidelity three-dimensional nonlinear analytical models are developed for the portfolios of bridge classes within different design eras. These models incorporate a wide range of geometric and material uncertainties, and their responses are characterized under seismic loadings. Fragility curves were then developed considering the vulnerability of multiple components and thereby help to quantify the performance of highway bridge networks and to study the impact of seismic design principles on the performance within a bridge class. This not only leads to the development of fragility relations that are unique and better suited for bridges in California, but also leads to the creation of better bridge classes and sub-bins that have more consistent performance characteristics than those currently provided by the National Bridge Inventory. Another important feature of this research is associated with the development of damage state definitions and grouping of bridge components in a way that they have similar consequences in terms of repair and traffic implications following a seismic event. These definitions are in alignment with the California Department of Transportation's design and operational experience, thereby enabling better performance assessment, emergency response, and management in the aftermath of a seismic event. The fragility curves developed as a part of this research will be employed in ShakeCast, a web-based post-earthquake situational awareness application that automatically retrieves earthquake shaking data and generates potential damage assessment notifications for emergency managers and responders.
Errata added at request of advisor and approved by Graduate Office, March 15 2016.

A number of artificially generated earthquake time-histories (AGETH) fitting to a Eurocode 8 (EC8) response spectrum are randomly generated using SIMQKE software and the average generated spectrum compares well with the EC8 one. Two Finite element (FE) smeared crack models, named Multi-crack and Craft, are well validated against experimental data of concrete and Reinforced concrete (RC) structures under monotonic and cyclic loadings. They are then used in the analysis of RC bridge piers under the AGETH. Several techniques including Fourier analysis, normalised cumulative spectrum, energy dissipation, damage index as well as probability applications are applied to quantify the structural response and damage. Based on the convergence of the representative responses under different numbers of AGETH, a minimum representative number of AGETH from 6 to 11 may be sufficient depending on the confidence band width from the mean of all damage responses. Effects of several parameters of the earthquake and structure to the dynamic response and damage of the bridge pier are investigated. Throughout these parametric studies, several of the common circumstances that structural engineers face are addressed and the proposed number of artificial earthquake time-histories required for non-linear dynamic analysis is thereby validated.

Following the evolution of a damage avoidance design (DAD) frame system, with rocking beam-column joints, at the University of Canterbury, analytical studies are carried out to evaluate the performance of proposed structures, and verify the proposed design methodology. A probabilistic seismic risk assessment methodology is proposed, from which the expected annualised financial loss (EAL) of a structure can be calculated. EAL provides a consistent basis for comparison of DAD frame systems with state-of-practice ductile monolithic construction. Such comparison illustrates the superior performance of DAD frame systems. The proposed probabilistic seismic assessment methodology requires the response of the structure to be evaluated over a range of seismic intensities. This can be achieved by carrying out an incremental dynamic analysis, explicitly considering seismic randomness and uncertainty; or from a pushover analysis, and assuming an appropriate value of the dispersion. By combining this information with the seismic hazard, probabilistic response curves can be derived, which when combined with information about damage states for the particular structure, can be transformed into 'resilience curves'. Integration of information regarding the financial loss occurring due to each of the damage states, results in an estimate of EAL.

The Central Andean Plateau (CAP), as defined by elevations in excess of 3 km, extends over 1800 km along the active South American Cordilleran margin making it the second largest active orogenic plateau on Earth. The uplift history of this high Plateau, with an average elevation around 4 km above sea level, remains uncertain as paleoelevation studies along the CAP suggest a complex, nonuniform uplift history. As part of the Central Andean Uplift and the Geodynamics of High Topography (CAUGHT) project, we image the S wave velocity structure of the crust and upper mantle using surface waves measured from ambient noise and teleseismic earthquakes to investigate the upper mantle component of plateau uplift. We observe three main features in our S wave velocity model including (1) a positive velocity perturbation associated with the subducting Nazca slab; (2) a negative velocity perturbation below the sub-Andean crust that we interpret as anisotropic Brazilian cratonic lithosphere; and (3) a high-velocity feature in the mantle above the slab that extends along the length of the Altiplano from the base of the Moho to a depth of similar to 120 km. A strong spatial correlation exists between the lateral extent of this high-velocity feature and the relatively lower elevations of the Altiplano basin suggesting a potential relationship. Determining if this high-velocity feature represents a small lithospheric root or foundering of orogenic lithosphere requires more integration of observations, but either interpretation implies a strong geodynamic connection with the uppermost mantle and the current topography of the northern CAP.

In the last decades, although the scientific community has attempted to explain a series of complex phenomena, ranging from natural hazards to physical conditions and economic crises, aspects of their generation process still escape our full understanding. The present thesis intends to promote our understanding of the spatiotemporal behavior and the generation mechanisms that govern large and strong earthquakes, employing a broad multidisciplinary perspective for the interpretation of catastrophic events. Two main questions are debated. The first question concentrates on “whether the generation process of an extreme event has more than one facets prior to its final appearance”. In the scientific study of earthquakes, attention is drawn to the predictive capability and monitoring of different precursory observations. Among them preseismic electromagnetic emissions have been also observed indicating that the science of earthquake prediction should be from the start multidisciplinary. Drawing on recently introduced models for earthquake dynamics, that address issues such as long-range correlations, self-affinity, complexity-organization and fractal structures, the present work endeavors to further penetrate on the analysis of preseismic electromagnetic emissions and elucidate their link with the generation process of large and strong earthquakes. A second question deals with “whether there is a unified approach for the study of catastrophic events”. This question implies the possibility for common statistical behavior of diverse extreme events and the potential for transferability of methods from the study of earthquake dynamics across other fields. On these grounds, the present work extends the focus of inquiry to the analysis of electroencephalogram recordings related to epileptic seizures, in the prospect to identify common mechanisms that may explain the nature and the generation process of both phenomena, and to open up different directions for future research. Finally, with a view to consider alternative ways of studying key theoretical principles associated with the generation process of catastrophic phenomena, a relevant framework based on proposed algorithms is presented, focusing on parameters such as: the energy of earthquakes, the mean and maximum magnitude of the sample, the probability that two samples may come from the same population. Such an attempt aims to contribute to the knowledge of natural phenomena, by extending the existing theory and models and providing a few more ways for their interpretation.

ABSTRACT SIMULATION OF SURFACE WAVES GENERATED BY A RAPID RISE OF A BLOCK AT THE SEA BOTTOM SENOL, Nalan M.Sc., Department of Civil Engineering, Supervisor: Assoc. Prof. Dr. ismail AYDIN July 2005, 74 Pages A mathematical model is developed for investigating time dependent surface deformations of a hydrostatic water volume, when it is subjected to a sudden partial rise of the sea bottom. In the model, 2-dimensional, compressible, and viscous Navier-Stokes equations are solved by Marker and Cell (MAC) method. Variable mesh size in both horizontal and vertical directions with a staggered grid arrangement is used. Limited compressibility model is utilized for pressure. Various computational tests are done for the selection of computational parameters of the model. It is found that the amplitude of surface waves generated by vertical displacements of the sea bottom depends on size and speed of bottom displacements.

Drawing upon 20 in-depth interviews with second generation Haitian young adults, I examined the ethnic identities and the involvement in ethnic organizations of the respondents. This study pays particular attention to how involvement in ethnic organizations influenced how the second generation Haitians believed the earthquake affected their identities and how they ultimately responded to the earthquake. Several of the findings revealed differences in how and why the respondents chose to ethnically identify such as Haitian, Haitian-American, black Haitian. The respondents' choice to join an ethnic organization was driven by different desires but the perceived influence of the organization on their ethnic identities resulted in an increase in cultural knowledge as well as an ability to stay rooted in the culture. However, the lack of participation on the part of some of the respondents was a choice dictated by conflicts of authenticity, time, and responsibilities. The comparison between involved and non-involved respondents in terms of their response to the earthquake revealed that involved respondents were more active in volunteer projects. Involvement in ethnic organizations influenced how the second generation Haitians perceived the earthquake affected their identities, and ethnic affirmation in terms of a desire to visit Haiti was expressed by involved respondents. The implications of this study revealed the importance of establishing ethnic organizations in middle and high schools in order to foster a sense of pride through knowledge at an earlier age.

Lorsqu'un séisme se produit sous l'océan, les ondes sismiques se transforment à l'interface croûte/océan en ondes acoustiques basses fréquences, appelées ondes T, qui peuvent se propager dans la colonne d'eau sur de très grandes distances. Ces ondes sont d'une grande importance pour la surveillance de l'activité sismique et volcanique sous-marine car elles comblent un manque d'information dans les données sismologiques terrestres. La localisation de la source acoustique peut être déduite par trilatération à partir des temps d'arrivée de l'onde T sur plusieurs hydrophones et une magnitude acoustique peut être dérivée des niveaux reçus. Toutefois, les informations obtenues à partir des signaux des ondes T restent incomplètes car les mécanismes de génération des ondes T, leur mode de propagation, et l'importance des effets 3D dans le processus sont mal connus. Pour traiter ces questions, cette thèse utilise un modèle analytique de la génération et de la propagation des ondes T et un code spectral aux éléments finis (SPECFEM) capable d'effectuer des simulations en forme d'onde complète des ondes sismiques dans la croûte terrestre et des ondes acoustiques dans l'océan. Le modèle analytique développé dans cette thèse décrit des ondes T qui se propagent sous forme de modes de Rayleigh ; ce modèle prédit également le spectre des ondes PN et SN, qui sont les précurseurs des ondes T dans les enregistrements d’événements de forte magnitude. Dans une configuration 2D avec un fond plat et un océan uniforme, les modes théoriques (solution analytique) et simulés (SPECFEM) d'ondes T et les spectres d'ondes PN et SN montrent un très bon accord. Le modèle numérique peut donc être appliqué avec confiance à des configurations pour lesquelles un modèle analytique ne peut pas être simplement dérivé s: une interface croûte/océan avec un mont sous-marin, avec ou sans canal SOFAR dans l'océan, et un fond marin plat avec une rugosité à courte longueur d'onde. Ces simulations mettent en évidence les conditions nécessaires à la génération d’ondes T énergétiques et confirment le rôle prédominant des modes de Rayleigh dans leur propagation. Les résultats du modèle avec un fond marin rugueux sont très comparables aux données hydroacoustiques d'un séisme majeur survenu sous une plaine abyssale. La version 3D de SPECFEM permet enfin d'étudier l'importance des effets 3D dans la génération des ondes T. À distance, les amplitudes et les temps d'arrivée diffèrent selon qu'un événement sismique se produit sous un mont sous-marin ou sous une dorsale. Par conséquent, les localisations de la source à partir de temps d'arrivée peuvent être biaisées par des effets topographiques en 3D à proximité de la région de l'épicentre
When an earthquake occurs below the ocean, seismic waves convert at the crust/ocean or solid/fluid interface into low-frequency acoustic waves, known as T-waves, which can propagate in the water column over very long distances.These waves are of great importance for monitoring the submarine seismic and volcanic activity, as they fill a gap of information in the land-based seismological records. The localization of the acoustic source can be deduced by trilateration from T-wave arrival-times at several hydrophones and an acoustic magnitude can be derived from the received levels.Beyond that, the information obtained from T-wave signals remains incomplete since the mechanisms for generating Twaves, the way they propagate and the importance of 3D-effects in the process are poorly known.To investigate these issues, this thesis uses a spectral finite element code (SPECFEM) able to perform full wave-form simulations of both seismic waves in the crust and acoustic waves in the ocean. The analytical model developed in this thesis describes propagating T-waves as Rayleigh modes; this model also predicts the spectrum of PN- and SN-waves, which are precursors of T-waves in records from high magnitude events. In a 2D configuration with a flat bottom and a uniform ocean, the theoretical (analytical solution) and simulated (SPECFEM) T-wave modes and PN- and SN-wave spectra show a very good agreement. The numerical model is thus applied with confidence to configurations for which an analytical model cannot be simply derived: a crust/ocean interface with a seamount, with or without a SOFAR channel in the ocean, and a flat seabed with a short-wavelength roughness. These simulations highlight the conditions necessary for the generation of energetic T-waves and confirm the predominant role of Rayleigh modes in their propagation. The outputs from the model with a rough sea-bottom compare well with hydroacoustic records from a major earthquake occurring below an abyssal plain. A 3D version of SPECFEM is finally used to investigate the importance of 3D-effects in the generation of T-waves. At a distance, different amplitudes and arrival-times are obtained depending on whether a seismic event occurs below a seamount or below a ridge. Hence, source localizations from arrival-times can be biased by 3D topographic effects in the vicinity of the epicenter region

This thesis presents validation and application of advanced soil constitutive models in cases of seismic loading conditions. Firstly, results of three advanced soil constitutive models are compared with examples of shear stack experimental data for free field response in dry sand for shear and compression wave propagation. Higher harmonic generation in acceleration records, observed in experimental works, is shown to be possibly the result of soil nonlinearity and fast elastic unloading waves. This finding is shown to have high importance on structural response, real earthquake records and reliability of conventionally employed numerical tools. Finally, short study of free field response in saturated soil reveals similar findings on higher harmonic generation. Secondly, two advanced soil constitutive models are used, and their performance is assessed based on examples of experimental data on piles in dry sand in order to validate the ability of the constitutive models to simulate seismic soil-structure interaction. The validation includes various experimental configurations and input motions. The discussion on the results focuses on constitutive and numerical modelling aspects. Some improvements in the formulations of the models are suggested based on the detailed investigation. Finally, the application of one of the advanced soil constitutive models is shown in regard to temporary natural frequency wandering observed in structures subjected to earthquakes. Results show that pore pressure generated during seismic events causes changes in soil stiffness, thus affecting the natural frequency of the structure during and just after the seismic event. Parametric studies present how soil permeability, soil density, input motion or a type of structure may affect the structural natural frequency and time for its return to the initial value. In addition, a time history with an aftershock is analysed to investigate the difference in structural response during the earthquake and the aftershock.

This thesis presents validation and application of advanced soil constitutive models in cases of seismic loading conditions. Firstly, results of three advanced soil constitutive models are compared with examples of shear stack experimental data for free field response in dry sand for shear and compression wave propagation. Higher harmonic generation in acceleration records, observed in experimental works, is shown to be possibly the result of soil nonlinearity and fast elastic unloading waves. This finding is shown to have high importance on structural response, real earthquake records and reliability of conventionally employed numerical tools. Finally, short study of free field response in saturated soil reveals similar findings on higher harmonic generation. Secondly, two advanced soil constitutive models are used, and their performance is assessed based on examples of experimental data on piles in dry sand in order to validate the ability of the constitutive models to simulate seismic soil-structure interaction. The validation includes various experimental configurations and input motions. The discussion on the results focuses on constitutive and numerical modelling aspects. Some improvements in the formulations of the models are suggested based on the detailed investigation. Finally, the application of one of the advanced soil constitutive models is shown in regard to temporary natural frequency wandering observed in structures subjected to earthquakes. Results show that pore pressure generated during seismic events causes changes in soil stiffness, thus affecting the natural frequency of the structure during and just after the seismic event. Parametric studies present how soil permeability, soil density, input motion or a type of structure may affect the structural natural frequency and time for its return to the initial value. In addition, a time history with an aftershock is analysed to investigate the difference in structural response during the earthquake and the aftershock.

碩士
國立臺灣大學
土木工程學研究所
103
Taiwan is located on the junction of Eurasian Plate and Philippine Sea Plate, so earthquakes occur frequently, thus cause multiple casualties and the loss of property of the people. Seismic performance assessment of buildings has become one effective tool to evaluate seismic risk of buildings. This study studies and improves the procedures for next-generation seismic performance assessment developed in the ATC-58 project in the Unites States, termed the ATC-58 procedures. The ATC-58 procedures provide two levels of risk evaluation using nonlinear dynamic and linear static analyses, respectively, and this study focuses on the improvement of the procedures using linear static analysis, termed the simplified analysis herein. In the simplified analysis, the structural responses of a building are computed using linear static analysis and then multiplied by correction factors to predict the linear/nonlinear responses of the building subjected to ground motions with different intensity. The formulas and parameters of correction factor were developed using regression analysis on the responses of thirteen prototype buildings subjected to ground motions with a wide range of intensity. This study proposes a new formula for calculating the correction factor, and improves the accuracy of the results of the simplified procedures. To compare the effectiveness of the correction factors recommended by the ATC-58 project and this study, this research performed a series of nonlinear response-history analyses using a three-story, a nine-story and a twenty-story moment resisting frames. The responses were compared with the predictions of the ATC-58 project and this study. For peak floor velocity, the formula of this study better predict the responses and reduce the variability in the residuals. The correction factors for both the ATC-58 procedures and this study fail to generate acceptable results for the peak floor acceleration (PFA) when the buildings are in the elastic range. This study recommends to use the modal superposition analysis to predict PFA in the elastic range.

Which kind of disasters can affect human life and how can science help to reduce the consequence of tragedy? Indeed, there is a range of challenges including technological or man-made hazards and natural hazards. Here in this thesis, one of the very critical natural hazards which has a major impact on human living i.e., earthquake hazards, is investigated. Population growth and patterns of economic development are two important issues directly affected by an earthquake occurrence and induced impacts, leading to dramatic disaster situations. The main aim of this thesis is to discuss how science can help to reduce the effect of the earthquake on human life. To human knowledge, precise earthquake prediction i.e., specification of the time, location, and magnitude of future earthquakes are almost impossible. Moreover, earthquake prediction is sometimes distinguished from earthquake forecasting, which can be defined as the probabilistic assessment of general earthquake hazards, including the magnitude and frequency of damaging earthquakes in a given area over the years or decades. Both earthquake prediction and forecasting are also different from earthquake warning systems, in which the latter can provide a warning to neighboring regions that might be affected for an ongoing earthquake. The Earthquake Early Warning (EEW) systems rapidly provide in-advanced warnings of impending strong ground motion in real-time as soon as detection of the ongoing earthquakes and before the impact of the ground vibrations. The initial part of the primary waves which is typically low-amplitude ground motion waveform i.e., P-waves is normally used to estimate the potentially large-amplitude ground motion. Note that issuing and transmitting the alarms information using telecommunication is faster than seismic wave propagation speed, thus, the early warnings may arrive at a target site before the strong shaking itself, thereby providing invaluable time for both people and automated systems to take actions to mitigate earthquake-related injury and losses. These actions might range from complex automated procedures as stopping high-speed trains to simple procedures as warning people to get themselves to a safe location. History of implementation of the first EEW system backs to 1991 (SASMEX) in Mexico City [Espinosa-Aranda, et al., 2009]. Nowadays, there is a significant improvement of EEW systems throughout the world which are operating in many parts of the world to provide warnings at high seismic hazard regions. Japan Meteorological Agency (JMA) and ShakeAlert EEW systems are two important examples developed in Japan and the west coast of the United States, respectively. In addition, EEW systems are being tested in other countries as Italy, Taiwan, Romania, China, South Korea, Turkey, and Switzerland. EEW standard approaches estimate the location and magnitude of an earthquake, the key ingredients among the other parameters which are used in a ground motion prediction equation (GMPE) to calculate expected ground shaking. If the expected ground motion is greater than a manually specified threshold, the user is alerted. For instance, the JMA system provides alarms to subprefectures whenever ground motions are expected to exceed JMA intensity 4 within that subprefecture. The JMA system has released hundreds of alerts, including alerts sent to several million people during the 2011 M9.0 Tohoku earthquake [Fujinawa and Noda, 2013]. Although nowadays EEWS is one of the various important challenges in seismology and that a lot of scientific efforts have been done to develop it, there is still a long way to consider it as a consolidated technology. The physical theory behind EEWS is not fully clear and all parameters measured from early motion with rather non-negligible uncertainty are used to predict the final earthquake characteristics. The main assumption of most models, both the processes and algorithms, used in standard EEWS approaches and induced wave propagation are based on some simplifications to model the earthquake source and wave propagation. Standard approaches for the peak motion prediction in EEW methods are typically based on the point-source approximation and on simple empirical attenuation relationships, depending on the magnitude and hypocentral distance. On average, few portions of the P-waves, 3 seconds are used to the real-time computation of the event magnitude and location, which could be a problem for any estimation of large events in which have a complex rupture process over tens of seconds. Several efforts are done in the last decade to measure a rupture during its early stages. Here in this thesis, we mainly focus on filling this gap, developing the algorithms to measure rupture characteristics and then consider the refined extended source to generate the shake map. Therefore, the thesis results can open a new research topic in real-time ground-shaking prediction for ongoing seismic events. All these concepts can be considered all together to issue the alarm and they will trigger “the next generation of EEW systems”. In the framework of SERA infrastructure (Seismology and Earthquake Engineering Research Infrastructure Alliance for Europe, call INFRAIA-01-2016-2017), and JRA 6 (Joint Research Action, “Real-Time earthquake Shaking”), different methodologies are being developed and tested to generate evolutionary ground shaking maps by considering a rupture kinematic description and reliable finite-fault model. In this regard, we have refined and tested various methodologies to retrieve the earthquake source. Same as the standard EEW approaches, the initial P-wave signals will be explored to identify the best proxies for the rapid source characterization (moment, length and duration). Updated kinematic rupture models (space-time slip function) are inferred by the consideration of progressively enlarged P-wave time windows as they are available at the network probes. The final output is the time-varying predicted-ground motion at the Earth surface at sites of interest in a recurrent manner. For this purpose, we first, evaluate many possibilities and algorithms in the offline analysis of the initial P-wave signals. In particular, the time evolution of peak amplitude parameters will be used for the rapid prediction of the source magnitude and for estimating and then modeling the moment rate function. These estimates are used to build simplified kinematic source models. The rupture speed and the rise time are selected accounting for the medium elastic properties and the event magnitude. A single patch slip distribution is imposed: its extension and position with respect to the nucleation are controlled by the moment estimates and by preliminary directivity estimates, respectively. The convolution of these models with pre-computed Green’s functions provides complete wavefield synthetic seismograms and thus early estimates of the expected amplitude vibrations (PGA/PGV) at the EEW target sites. The alert decision scheme is thus defined upon the exceedance of a user-compliant PGA/PGV threshold by the predicted synthetic values. In addition, the inversion methodology will be implemented and tested on synthetic and real waveforms in off-line acquisition mode. A database of synthetic waveforms will be generated for a variety of case-studies (Ischia and Norcia earthquakes occurred in Italy). The off-line application will be checked, but the main objective is the development, the implementation and validation of efficient algorithms for the real-time signal processing, slip inversion and ground-shaking forecast that will improve the predictive performance of EEWS. The structure of the thesis is based on four main chapters as follows: The first chapter, as an introduction, describes concepts of the earthquake early warning system from standard approaches to those expected in the next-generation tools. The second chapter illustrates the new model to compute the earthquake source characteristics. In the third chapter, using the source model resulted from the previous chapter, the evolutionary ground shaking prediction considering the Norcia event as a case study is evaluated. Finally, the last chapter is about calculating the source mechanism and rupture model from the inversion of a near-Source record.

碩士
國立臺灣科技大學
營建工程系
97
This study proposed a mode for earthquake damage simulation of an area based on full numerical analysis. A detailed nonlinear analytical model is built for each building in the area to obtain its aseismic capability by numerical simulation. More detailed and realistic results are expected in the proposed mode. However, a simulated area usually consists of a large number of buildings. Their modeling and analysis are sure to be a labor-intensive and time-consuming work. This study adopted and integrated automatic modeling and distributed computing to realize the proposed mode. First, automatic modeling of all building in a simulated area is achieved by using a structural component model database, which collects various structural component models addressing various structural details. Since the handling of structural component modeling can be provided by querying the proposed database, the nonlinear analytical models of all buildings can be built automatically with each building data, such as the geometrical, topological and material properties, provided in a structural format. Afterward, since the numerical analyses of all building models are independent to each other, all the building models are analyzed in parallel by being distributed to a cluster of computer.

abstract: Earth's topographic surface forms an interface across which the geodynamic and geomorphic engines interact. This interaction is best observed along crustal margins where topography is created by active faulting and sculpted by geomorphic processes. Crustal deformation manifests as earthquakes at centennial to millennial timescales. Given that nearly half of Earth's human population lives along active fault zones, a quantitative understanding of the mechanics of earthquakes and faulting is necessary to build accurate earthquake forecasts. My research relies on the quantitative documentation of the geomorphic expression of large earthquakes and the physical processes that control their spatiotemporal distributions. The first part of my research uses high-resolution topographic lidar data to quantitatively document the geomorphic expression of historic and prehistoric large earthquakes. Lidar data allow for enhanced visualization and reconstruction of structures and stratigraphy exposed by paleoseismic trenches. Lidar surveys of fault scarps formed by the 1992 Landers earthquake document the centimeter-scale erosional landforms developed by repeated winter storm-driven erosion. The second part of my research employs a quasi-static numerical earthquake simulator to explore the effects of fault roughness, friction, and structural complexities on earthquake-generated deformation. My experiments show that fault roughness plays a critical role in determining fault-to-fault rupture jumping probabilities. These results corroborate the accepted 3-5 km rupture jumping distance for smooth faults. However, my simulations show that the rupture jumping threshold distance is highly variable for rough faults due to heterogeneous elastic strain energies. Furthermore, fault roughness controls spatiotemporal variations in slip rates such that rough faults exhibit lower slip rates relative to their smooth counterparts. The central implication of these results lies in guiding the interpretation of paleoseismically derived slip rates that are used to form earthquake forecasts. The final part of my research evaluates a set of Earth science-themed lesson plans that I designed for elementary-level learning-disabled students. My findings show that a combination of concept delivery techniques is most effective for learning-disabled students and should incorporate interactive slide presentations, tactile manipulatives, teacher-assisted concept sketches, and student-led teaching to help learning-disabled students grasp Earth science concepts.
Dissertation/Thesis
Ph.D. Geological Sciences 2014

碩士
國立中央大學
水文所
96
On December 26, 2006, two ML=7.0 submarine earthquakes occurred offshore Ping-Tung, Taiwan. Though these two earthquakes have not caused any tsunami hazard, the tidal gauges near the epicenters did record the incidences of long water waves, similar to tsunami waves. There have not been many seismic records of large submarine earthquakes offshore southwestern Taiwan. As such, people living in the coastal area around Taiwan generally lack alertness for potential tsunami hazards caused by submarine earthquakes. In order to assess the potential tsunami hazards in this region, numerical simulations of the tsunami-wave propagation generated by the two Pingtung submarine earthquakes are conducted in this study. The initial tsunami wave height is assumed to be equal to the seafloor vertical displacement caused by the earthquake, and Harvard CMT source parameters are used to calculate the seafloor vertical displacement using elastic fault model. The water-level records of the available tidal gages are used to validate the simulation results. The effects of the source parameters (source depth, average dislocation, and fault-plane width) on the initial tsunami wave height and the subsequent propagation are also studied. The results indicates that the initial tsunami wave height increases from 0.09 m to 0.4 m and the maximum water level at Hou-Bi-Hu tidal-gauge station increases from 0.2 m 0.4 m, as the source depth decreases from 40 km to 10 km; the initial tsunami wave height increases from 0.02 m to 0.6 m and the maximum water level at Hou-Bi-Hu tidal-gauge station increases from 0.35 m to 1 m, as the average dislocation increases from 1 m to 3 m. The water depths in the southwest region offshore Taiwan are between 2,000 m and 3,000 m, which can result in a large propagation speed of the tsunami wave. Our simulation result indicates that the leading tsunami wave arrives at Heng-Chun coast within 17 min after the quake. This implies that if the same submarine earthquake with a larger magnitude occurs, it would certainly cause devastation to the southwestern coasts of Taiwan.

碩士
國立中央大學
水文所
96
Almost all devastating tsunamis were triggered by shallow earthquakes in sub-duction zones. We thus focus on the Manila sub-duction zone (or called Manila Trench), deemed as the most likely source region within the South China Sea (SCS), to assess the potential tsunami hazards in Taiwan. To cope with the changing strikes, we group the entire Manila Trench into six hypothetical fault segments, similar to those of Kirby (2006). The width and dislocation of each fault are then inferred from its length, according to the empirical relations of Papazachos et al. (2004). In addition, by summing all the six fault segments, we create a hypothetical 990 km long fault along the Manila Trench, whose width, dislocation and focal depth are determined based on the data of the three largest earthquakes occurred in the past one hundred year . Upon simulation of tsunami caused by earthquakes on hypothetical faults, results indicate that the heights of tsunami waves decreases as they propagate from the source region into South China Sea (SCS). However, a 60% amplification of wave height is observed along the southeast China offshore when the wave crossing the continental slope and shelf. For the 990 km long fault scenario, the tsunami wave height reaches about 5 m in southern Taiwan, Suao, and Ilan. It reaches about 10 m in southeast China and Vietnam. For the island nearest to the source region, Luzon island, the tsunami wave height can reach as high as 15 m. Simulation results of the worse case scenario also indicate that tsunami waves will arrive at the Luzon island 10 min after the earthquake and the southern Taiwan 20 min after. The more than 20min evacuation time for other regions suggests that the establishment of an effective tsunamis warning system is plausible. We also compare the simulation results of the 990 km long fault with the tsunami of the 2004 Sumatra earthquake. The epicenter distances from Sumatra to Sri Lanka and India are comparable to those from Manila to Vietnam. As the results, we deduce comparable tsunami arrival times and wave heights along Sri Lanka, India, and Vietnam. However, while the tsunami waves can propagate further into the open Indian Ocean for the Sumatra case, the relative closeness of SCS tends to trap tsunami waves, causing more damages for Manila case. Manila results of this study suggest that the establishment of SCS tsunami warning system is worthwhile and earthquakes in Manila sub-duction zone need to put into highly alert.

博士
國立中央大學
土木工程學系
105
Taiwan is one of the world’s most disaster-prone regions, as indicated in the World Bank’s report entitled “Natural Disaster Hotspots: A Global Risk Analysis.” In fact, every year Taiwan public and private organizations need to work together to prepare various training programs in hope of mitigating disaster impact. Design of an appropriate disaster drill, hence, plays an important role of enhancing the capability of their emergency response units. However, developing a reasonable scenario for the disaster drill is a time-consuming, error-prone task, and experienced drill designers, disaster management officers and/or first responders are required to join the development work. The level of a community’s disaster preparedness and readiness are also highly dependent on whether such work is performed well. In this research, an earthquake disaster and an area in the city of Taipei were selected for the demonstration of the proposed decision support model and its system, called EDSS (Earthquake Drills Generation and Simulation System). EDSS requires an input data set containing the impact assessments of an earthquake disaster, which can be obtained from Taiwan Earthquake Loss Estimation System (TELES) and are similar to the results of using HAZUS (Hazards in the U.S., a tool created by FEMA and the base version of TELES). EDSS utilizes Unity, a 3D serious game engine, to actually display the events of an earthquake disaster in accordance with the drill script, which is traditionally shown as plain text. In the designated area, real building geometry and simulated earthquake event are rendered inside Unity; hence, after the occurrence of a real earthquake, building damage information can be collected and entered into EDSS for further analysis, in order to show the remaining, disaster-related events not currently discovered by first responders. In model validation, senior drill designers, disaster management officers and first responders were invited to evaluate the effectiveness of EDSS, and the results show that EDSS has the potential of improving first responders’ awareness of field conditions as well as helps first responders better understand how a disaster evolves over time.

The strain-stress state generated by faulting or cracking and influenced by the strong heterogeneity of the internal earth structure precedes and accompanies volcanic and seismic activity. Particularly, volcanic eruptions are the culmination of long and complex geophysical processes and physical processes which involve the generation of magmas in the mantle or in the lower crust, its ascent to shallower levels, its storage and differentiation in shallow crustal chambers, and, finally, its eruption at the Earth’s surface. Instead, earthquakes are a frictional stick-slip instability arising along pre-existing faults within the brittle crust of the Earth. Long-term tectonic plate motion causes stress to accumulate around faults until the frictional strength of the fault is exceeded. The study of these processes has been traditionally carried out through different geological disciplines, such as petrology, structural geology, geochemistry or sedimentology. Nevertheless, during the last two decades, the development of physical of earth as well as the introduction of new powerful numerical techniques has progressively converted geophysics into a multidisciplinary science. Nowadays, scientists with very different background and expertises such as geologist, physicists, chemists, mathematicians and engineers work on geophysics. As any multidisciplinary field, it has been largely benefited from these collaborations. The different ways and procedures to face the study of volcanic and seismic phenomena do not exclude each other and should be regarded as complementary. Nowadays, numerical modeling in volcanology covers different pre-eruptive, eruptive and post-eruptive aspects of the general volcanic phenomena. Among these aspects, the pre-eruptive process, linked to the continuous monitoring, is of special interest because it contributes to evaluate the volcanic risk and it is crucial for hazard assessment, eruption prediction and risk mitigation at volcanic unrest. large faults. The knowledge of the actual activity state of these sites is not only an academic topic but it has crucial importance in terms of public security and eruption and earthquake forecast. However, numerical simulation of volcanic and seismic processes have been traditionally developed introducing several simplifications: homogeneous half-space, flat topography and elastic rheology. These simplified assumptions disregards effects caused by topography, presence of medium heterogeneity and anelastic rheology, while they could play an important role in Moreover, frictional sliding of a earthquake generates seismic waves that travel through the earth, causing major damage in places nearby to the modeling procedure This thesis presents mathematical modeling and numerical simulations of volcanic and seismic processes. The subject of major interest has been concerned on the developing of mathematical formulations to describe seismic and volcanic process. The interpretation of geophysical parameters requires numerical models and algorithms to define the optimal source parameters which justify observed variations. In this work we use the finite element method that allows the definition of real topography into the computational domain, medium heterogeneity inferred from seismic tomography study and the use of complex rheologies. Numerical forward method have been applied to obtain solutions of ground deformation expected during volcanic unrest and post-seismic phases, and an automated procedure for geodetic data inversion was proposed for evaluating slip distribution along surface rupture.
Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Catania
Unpublished
3.6. Fisica del vulcanismo
open