Situé dans le contexte d’une subduction rapide, (vitesse de convergence de 6-8 cm/an), l’Equateur est particulièrement exposé au risque sismique. La composante de la sismicité qui est directement attribuée au processus de subduction est depuis quelques années l’objet de nombreuses études, mais la sismicité “indirecte” due aux processus de déformation dans la plaque supérieure reste peu étudiée. Pourtant, les populations sont directement exposées à cette séismicité intracontinentale, qui reste dangereuse en raison de sa faible profondeur et distance aux zones densément peuplées.Le premier objectif de cette thèse est l’étude de la cinématique régionale, dominée par la présence d’un grand système de failles continentales décrochantes qui s’étend depuis la marge active (Golfe de Guayaquil) jusqu’à la frontière Colombienne. De nombreuses failles associées au processus d’extrusion du Bloc Nord Andin montrent une trace géomorphologique très nette mais ne sont pourtant pas cartées car les accès terrain ne permettent pas de conclure sur leur activité. L’analyse des photos aériennes, d’images satellite et de MNT à différentes échelles nous a permis d’établir tout d’abord un catalogue tectonique des structures majeures, homogène et systématique, puis de déterminer la cinématique de la plupart des structures actives en Equateur à l’échelle Quaternaire. De plus, nous avons intégré l’information sismologique (sismicité et mécanismes au foyer) et géodésique disponible à partir des réseaux nationaux de l’Equateur. Ceci nous a permis d’apporter des contraintes cinématiques sur le fonctionnement de ces structures et comparer les processus de déformation instantanés et cumulés à l’échelle du Quaternaire. Le système tectonique majeur d’Equateur Chingual-Cosanga-Pallatanga-Puná (CCPP) montre une vitesse moyenne de 7 à 8 mm/an par rapport à Sud Amérique. Les structures qui le définissent sont principalement transpressives dextres, avec une distribution nette de la composante décrochant sur les failles de direction NE-SW, et de la composante inverse sur les structures plutôt NS. Ce système tectonique majeur permet l’extrusion vers le nord du Bloc Nord-Andin. Notre interprétation implique l’existence et la définition d’un nouveau microbloc, isolé sur le flanc Ouest de la structure CCPP, qui est limité à l’ouest par les systèmes chevauchants de Quito et Latacunga, et à l’est par les failles du système Chingual Cosanga, depuis environ 3 Ma.Dans un deuxième temps, l’analyse des marqueurs morphologiques de la déformation et de l’évolution des réseaux de drainage associée à un travail de détail sur le terrain et de datations nous ont permis d’étudier cette deuxième région particulière des Andes d’Equateur. Le système de chevauchements de Quito est formé par des failles inverses aveugles qui produisent en surface une série de plis en échelon à vergence Ouest dans les dépôts volcaniques Quaternaire. Cette cinématique est confirmée par l’analyse de la sismicité superficielle et locale, et les mécanismes au foyer issus du réseau national RENSIG. Les données GPS montrent un taux de raccourcissement EW de 4 mm/an, accommodé sur cette structure compressive particulièrement active à l’échelle régionale. En profondeur le système de faille de Quito se connecte probablement vers l’ouest à l’ancienne suture, qui marque la limite des terrains océaniques accrétés au continent. L’analyse du réseau de drainage souligne de plus la propagation du système de failles de Quito vers le nord et vers l’Est, depuis le Quaternaire récent.Pour finir, à partir de ces nouveaux modèles locaux et régionaux de déformation Quaternaire pour l’Equateur, nous avons caractérisé et défini 19 nouvelles sources sismogéniques pour la croûte continentale. Cet apport permettra d’intégrer la tectonique active dans les futurs calculs d’aléa sismique pour aider à la définition du risque sur le territoire Equatorien. / Located in the Northern Andes along the active subduction zone of the Nazca plate beneath the South American continent, Ecuador is highly exposed to seismic risk. Moreover, the upper plate is actively deforming and faults responsible for crustal earthquakes are poorly known, showing the need to take them into account in modern assessments of seismic hazard.Our first objective is then the study of regional active faults and their kinematics, in Ecuador. Systematic analysis of air photos and satellite imagery, as well as geomorphic evidences gathered at different scales along these structures permitted us to obtain a regional tectonic map. Each observation was taken in account if also confirmed on digital terrain models (DTM) and field data. Finally this first step permitted to establish a first up-to-date and homogeneous catalog of major tectonic structures, active at the Quaternary time scale, consistent over the entire territory. We also determined the sense of motion of these active faults in Ecuador. Additionally, we integrate the seismic (instrumental and historical seismicity together with focal mechanisms for the higher magnitude events) and geodetic data available from national, global networks and field work to derive consistent kinematics models. Finally, the kinematics of each segment is compared to the instantaneous and cumulated Quaternary deformation.We are thus able to document a major tectonic system in Ecuador: the Chingual-Cosanga-Pallatanga-Puná fault system, showing a relative velocity of 7-8 mm/yr. with respect to South America. The deformation is characterized mainly by a combination of dextral NE-SW transpressive faults and reverse NS faults. This system accommodates the northward tectonic extrusion of the North Andean Block. Our interpretation implies the existence and definition of a new micro block, isolated on the western flank of the CCPP structure, bounded on the west by the thrust fault system of Quito and Latacunga and to the east by the Chingual-Cosanga faults, probably active over the past 3 million years.In a second step, the analysis of the evolution of the drainage system and its interaction with the active tectonic deformation, together with focused sampling and dating was applied to the study of a particular region of the Andes of Ecuador: the Latacunga and Quito micro-block together with the Quito faults system. The Quito reverse faults system consists of blind thrust faults that outcrop at the surface as a series of en-echelon folds, dipping west and affecting Quaternary volcanic deposits. This kinematics is further confirmed by the analysis of surface and local seismicity and focal mechanisms provided mainly by the national network RENSIG. Available GPS data show a EW convergence at a rate of 4 mm/yr, which is accommodated on the Quito fault system, and particularly active at a regional level. At depth the Quito fault system is probably connected to the west to the old suture, which marks the limit of oceanic terranes accreted to the continent. The analysis of the drainage system has allowed us to highlight the propagation of Quito fault system eastward and northward during the Quaternary.To conclude from the new local and regional models of Quaternary deformation in Ecuador, we have characterized and defined new seismic sources for the continental crust. Each source corresponds to a characteristic magnitude value and deformation model. This contribution integrates active tectonics as should be done in future seismic hazard calculations to help better quantity seismic hazard on Ecuadorian territory.

The central and eastern United States is subject to seismic hazards from both natural and induced earthquakes, as evidenced by the 1811-1812 New Madrid earthquake sequence, consisting of at least three magnitude 7 and greater earthquakes, and by four magnitude 5 and greater induced earthquakes in Oklahoma since 2011. To mitigate seismic hazards, both earthquake sources and their effects need to be characterized. Ground motion site response can cause additional damage to susceptible infrastructure and buildings. Recent studies indicate that Vs30, one of the primary site-response predictors used in current engineering practice, is not reliable. To investigate site response in the New Madrid Seismic Zone, ratios of surface-to-bedrock amplitude spectra, TFT, from S-wave recordings at the two deep vertical seismic arrays in the sediment-filled upper Mississippi Embayment (i.e., VSAP and CUSSO) were calculated. The mean TFT curves were compared with theoretical transfer functions; the results were comparable, indicating that TFT estimates of the empirical, linear SH-wave site responses at these sites. The suitability of surface S-wave horizontal-to-vertical spectral ratios, H/V, for estimating the empirical site transfer function was also evaluated. The results indicate that mean S-wave H/V curves are similar to TFT at low frequencies (less than the fifth natural frequencies) at both CUSSO and VSAP. SH-wave fundamental frequency, f0, and fundamental-mode amplification, A0, were evaluated as alternatives to the Vs30 proxy to estimate primary linear site-response characteristics at VSAP, CUSSO, and nine other seismic stations in the CEUS. In addition, calculated f0 and A0 were compared with the first peaks of S-wave H/V spectral ratios. The f0 and A0 were found to approximate the 1-D linear, viscoelastic, fundamental-mode responses at most stations. Also, S-wave H/V from weak-motion earthquakes can be used to measure f0. However, S-wave H/V does not reliably estimate A0 in the project area. S-wave H/V observations reveal site response within the frequency band of engineering interest from deeper, unmodeled geological structures. Because damaging or felt earthquakes induced by hydraulic fracturing and wastewater disposal have occurred in the CEUS, characterizing background seismicity prior to new large-scale subsurface fluid injection is important to identify cases of and the potential for induced seismicity. The Rogersville Shale in the Rome Trough of eastern Kentucky is being tested for unconventional oil and gas potential; production of this shale requires hydraulic fracturing, which has been linked to induced seismicity elsewhere in the CEUS. To characterize natural seismicity and to monitor induced seismicity during testing, a temporary seismic network was deployed in the Rome Trough near the locations of new, Rogersville Shale oil and gas test wells. Using the real-time recordings of this network and those of other regional seismic stations, three years of local seismicity were cataloged. Only three earthquakes occurred in the Rome Trough of eastern Kentucky, none of which was associated with the deep Rogersville Shale test wells that were stimulated during the time the network was in operation.

京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第22259号 / 理博第4573号 / 新制||理||1657(附属図書館) / 京都大学大学院理学研究科地球惑星科学専攻 / (主査)教授 James Mori, 教授 久家 慶子, 教授 岩田 知孝 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DGAM

In many areas of the world, seismic hazards pose a great risk to both human and natural populations. In particular, earthquake-induced tsunamis are especially dangerous to many areas in the Pacific. The study and quantification of these seismic events can both help scientists better understand how these natural hazards occur and help at-risk populations make better preparations for these events. However, many events of interest occurred too long ago to be recorded by modern instruments, so data on these earthquakes are sparse and unreliable. To remedy this, a Bayesian method for reconstructing the source earthquakes for these historical tsunamis based on anecdotal data, called TsunamiBayes, has been developed and used to study historical events that occurred in 1852 and 1820. One drawback of this method is the computational cost to reconstruct posterior distributions on tsunami source parameters. In this work, we improve on the TsunamiBayes method by introducing higher-order MCMC methods, specifically the Hamiltonian Monte Carlo (HMC) method to increase sample acceptance rate and therefore reduce computation time. Unfortunately the exact gradient for this problem is not available, and so we make use of a surrogate gradient via a neural network fitted to the forward model. We examine the effects of this surrogate gradient HMC sampling method on the posterior distribution for an 1852 event in the Banda Sea, compare results to previous results collected usisng random walk, and note the benefits of the surrogate gradient in this context.

Along with being one of the most populated regions of the world, Indonesia has one of the highest natural disaster rates worldwide. One such natural disaster that Indonesia is particularly prone to are tsunamis. Tsunamis are primarily caused by earthquakes, volcanoes, landslides and debris flows. To effectively allocate resources and create emergency plans we need an understanding of the risk factors of the region. Understanding the source events of destructive tsunamis of the past are critical to understanding the these risk factors. We expand upon previous work focusing on earthquake-generated tsunamis to consider landslide-generated tsunamis. Using Bayesian inference and modern scientific computing we construct a posterior distribution of potential landslide sources based on anecdotal data of historically observed tsunamis. After collecting 30,000 samples we find a landslide source event provides a reasonable match to our anecdotal accounts. However, viable landslides may be on the edge of what is physically possible. Future work creating a coupled landslide-earthquake model may account for the weaknesses with having a solely landslide or earthquake source event.

This thesis presents the results of basically two separate studies. The first study involved identifying structural and earthquake parameters that influenced seismic structural collapse. The parameter study involved nonlinear dynamic analyses using single-degree-of-freedom (SDOF) bilinear models. Four parameters were associated with the SDOF models — the lateral stiffness, the post-yield stiffness ratio, the yield strength, and the stability ratio (P-Delta effects). Then, three parameters were associated with the ground motions — the records themselves, the lateral ground motion scales, and the vertical ground motion scales. From the parameter study, it was found that the post-yield stiffness ratio augmented by P-Delta effects (rp) in conjunction with the ductility demand was the best predictor of collapse. These two quantities include all four structural parameters and the seismic displacement demands. It was also discovered in the parameter study that vertical accelerations did not significantly influence lateral displacements unless a given combination of model and earthquake parameters was altered such that the model was on the verge of collapsing. The second study involved Incremental Dynamic Analysis (IDA) using bilinear SDOF models representative of low rise buildings in both the Western United States (WUS) and the Central and Eastern United States (CEUS). Models were created that represented three, five, seven, and nine story buildings. Five sites from both the WUS and CEUS were used. Four different damage measures were used to assess the performance of the buildings. The IDA study was primarily interested in the response of the structures between the earthquake intensities that have a 10 percent probability of occurring in 50 years (10/50) and 2 percent probability of occurring in 50 years (2/50). The results showed that all structures could be in danger of severe damage and possible collapse, depending on which damage measure and which earthquake was used. It is important to note, though, that the aforementioned is based on a damage-based collapse rule. The damage-based rule results were highly variable. Using an intensity-based collapse rule, proved to be more consistent. Due to the nature of the bilinear models, only those structures with negative rp values ever collapsed using an intensity-based collapse rule. Most of the WUS models had positive rp values and many of the CEUS models had negative rp values. While many of the CEUS structures had negative rp values, which made them prone to collapse, most of the CEUS structures analyzed did not collapse at the 2/50 intensity. The reason was that the periods of the CEUS models were much longer than the approximate periods that were required to determine the strengths. Consequently, the strength capacity of most of the CEUS models was much greater than the seismic strength demands. While many of the CEUS models did have sudden collapses due to the large negative rp values, the collapses happened at intensities that were generally much higher than the 2/50 event. In the IDA, it was also shown that vertical accelerations can significantly affect the ductility demands of a model with a negative rp post-yield stiffness ratio as the earthquake intensity approaches the collapse intensity. Since IDA is concerned with establishing collapse limit states, it seems that the most accurate collapse assessments would include vertical accelerations. / Master of Science

Liquefaction-induced lateral spread displacements cause severe damage to infrastructure, resulting in large economic losses in affected regions. Predicting lateral spread displacements is an important aspect in any seismic analysis and design, and many different methods have been developed to accurately estimate these displacements. However, the inherent uncertainty in predicting seismic events, including the extent of liquefaction and its effects, makes it difficult to accurately estimate lateral spread displacements. Current conventional methods of predicting lateral spread displacements do not completely account for uncertainty, unlike a performance-based earthquake engineering (PBEE) approach that accounts for the all inherent uncertainty in seismic design. The PBEE approach incorporates complex probability theory throughout all aspects of estimating liquefaction-induced lateral spread displacements. A new fully-probabilistic PBEE method, based on results from the cone penetration test (CPT), was created for estimating lateral spread displacements using two different liquefaction triggering procedures. To accommodate the complexity of all probabilistic calculations, a new seismic hazard analysis tool, CPTLiquefY, was developed. Calculated lateral spread displacements using the new fully-probabilistic method were compared to estimated displacements using conventional methods. These comparisons were performed across 20 different CPT profiles and 10 cities of varying seismicity. The results of this comparison show that the conventional procedures of estimating lateral spread displacements are sufficient for areas of low seismicity and for lower return periods. However, by not accounting for all uncertainties, the conventional methods under-predict lateral spread displacements in areas of higher seismicity. This is cause for concern as it indicates that engineers in industry using the conventional methods are likely under-designing structures to resist lateral spread displacements for larger seismic events.

Liquefaction-induced lateral spread displacements cause severe damage to infrastructure, resulting in large economic losses in affected regions. Predicting lateral spread displacements is an important aspect in any seismic analysis and design, and many different methods have been developed to accurately estimate these displacements. However, the inherent uncertainty in predicting seismic events, including the extent of liquefaction and its effects, makes it difficult to accurately estimate lateral spread displacements. Current conventional methods of predicting lateral spread displacements do not completely account for uncertainty, unlike a performance-based earthquake engineering (PBEE) approach that accounts for the all inherent uncertainty in seismic design. The PBEE approach incorporates complex probability theory throughout all aspects of estimating liquefaction-induced lateral spread displacements. A new fully-probabilistic PBEE method, based on results from the cone penetration test (CPT), was created for estimating lateral spread displacements using two different liquefaction triggering procedures. To accommodate the complexity of all probabilistic calculations, a new seismic hazard analysis tool, CPTLiquefY, was developed. Calculated lateral spread displacements using the new fully-probabilistic method were compared to estimated displacements using conventional methods. These comparisons were performed across 20 different CPT profiles and 10 cities of varying seismicity. The results of this comparison show that the conventional procedures of estimating lateral spread displacements are sufficient for areas of low seismicity and for lower return periods. However, by not accounting for all uncertainties, the conventional methods under-predict lateral spread displacements in areas of higher seismicity. This is cause for concern as it indicates that engineers in industry using the conventional methods are likely under-designing structures to resist lateral spread displacements for larger seismic events.

The controlled rocking heavy timber wall (CRHTW) is a high-performance structural solution that was first developed in New Zealand, mainly considering Laminated Veneer Lumber (LVL), to resist high seismic loads without sustaining structural damage. The wall responds in bending and shear to small lateral loads, and it rocks on its foundation in response to large seismic loads. In previous studies, rocking has been controlled by both energy dissipation elements and post-tensioning, and the latter returns the wall to its original position after a seismic event. The controlled rocking response avoids the need for structural repair after an earthquake, allowing for more rapid return to occupancy than in conventional structures. Whereas controlled rocking walls with supplemental energy dissipation have been studied before using LVL, this thesis proposes an adapted CRHTW in which the design and construction cost and complexity are reduced for low-to-moderate seismic hazard regions by removing supplemental energy dissipation and using cross-laminated timber (CLT) because of its positive economic and environmental potential in the North American market. Moreover, whereas previous research has focussed on direct displacement-based design procedures for CRHTWs, with limited consideration of force-based design parameters, this thesis focusses on force-based design procedures that are more common in practice. A design and analysis process is outlined for the adapted CRHTW, based on a similar methodology for controlled rocking steel braced frames. The design process includes a new proposal to minimize the design forces while still controlling peak drifts, and it also includes a new proposal for predicting the influence of the higher modes by referring to previous research on the capacity design of controlled rocking steel braced frames. Also, a numerical model is outlined, including both a baseline version and a lower-bound model based on comparison to experimental data. The numerical model is used for non-linear time-history analysis of a prototype design, confirming the expected performance of the adapted CRHTW, and the model is also used for incremental dynamic analyses of three-, six-, and nine-storey prototypes, which show a low probability of collapse. / Thesis / Master of Applied Science (MASc) / The controlled rocking heavy timber wall (CRHTW) is a high-performance structural solution that was developed to resist high seismic loads without sustaining structural damage. The wall responds in bending and shear to small lateral loads, and it rocks on its foundation in response to large seismic loads. In previous studies, rocking has been controlled by both energy dissipation elements and post-tensioning; the latter returns the wall to its original position after a seismic event. This controlled rocking behaviour mitigates structural damage and costly repairs. This thesis explores the value of an adapted CRHTW in which the design and construction costs and complexity are reduced for low-to-moderate seismic hazard regions by using post-tensioning but no supplemental energy dissipation. A design and analysis process is outlined; numerical analysis confirms the expected performance of the adapted CRHTW; and the system is shown to have a low probability of collapse.

Past earthquakes have demonstrated that Indian sub-continent is highly vulnerable to earthquake hazards. It has been estimated that about 59 percent of the land area of the Indian subcontinent has potential risk from moderate to severe earthquakes (NDMA, 2010). Major earthquakes in the last 20 years such as Khillari (30th September 1993), Jabalpur (22nd May 1997), Chamoli (29th March 1999) and Bhuj (26th January 2001) earthquakes have resulted in more than 23,000 deaths and extensive damage to infrastructure (NDMA, 2010). Although it is well known that the major earthquake hazard prone areas in India are the Himalayan region (inter-plate zone) and the north-east region, (subduction zone) the seismicity of Peninsular India cannot be underestimated. Many studies (Seeber et al., 1999; Rao, 2000; Gangrade & Arora, 2000) have proved that the seismicity of Peninsular India is significantly high and may lead to earthquakes of sizeable magnitude. This necessitates a seismic zonation for the country, as well as various regions in it. Seismic zonation is the first step towards an effective earthquake risk mitigation study. Seismic zonation is a process in which a large region is demarcated into small zones based on the levels of earthquake hazard. Seismic zonation is generally carried out at three different levels based on the aerial extent of the region, importance of site and the population. They are micro-level, meso-level and macro-level. The macro-level zonation is generally carried out for large landmass such as a state or a country. The earthquake hazard parameters used for macro-level zoning are generally evaluated with less reliability. The typical example of a macro-level zonation is the seismic zonation map of India prepared by BIS-1893 (2002), where the entire India is demarcated into four seismic zones based on past seismicity and tectonic conditions. Generally the macro-level seismic zonation is carried out based on peak horizontal acceleration (PHA) estimated at bedrock level without giving emphasis on the local soil conditions. Seismic zonation at the meso-level is carried out for cities and urban centers with a population greater than 5,00,000. The earthquake hazard parameters, for the meso-level zonation are evaluated with greater degree of reliability, compared to the macro-level zoning. The micro-level zonation is carried out for sites which host critical installations such as nuclear power plants (NPPs). As the NPPs are considered as very sensitive structures, the earthquake parameters, for the micro-level zonation of the NPP sites are estimated with a highest degree of reliability. The local soil conditions and site effects are properly counted for carrying out the micro as well as the meso-level zonation. Several researchers have carried out meso-level zonation considering effects of all major earthquake hazards such as PHA, site amplification, liquefaction (Mohanty et al., 2007; Nath et al., 2008; Sitharam & Anbazhagan, 2008 etc.) Even though the above definitions and descriptions are available for various levels of zonation, the key issue lies in the adoption of the suitable one for a given region. There are only a few guidelines available regarding the use of a particular level of zonation for a given study area. Based on the recommendation of the disaster management authority, the government of India has initiated the seismic zonation of all major cities in India. As it is evident that large resources are required in order to carry out seismic site characterization and site effect estimation, both the micro and meso-level zonations cannot be carried out for all these cities. Hence there is a need to propose appropriate guidelines to define the suitability of each level zonation for various re-gions in the country. Moreover there are many methodologies available for site characterization and estimation of site effects such as site amplification and liquefaction. The appropriateness of these methodologies for various levels of seismic zonations also needs to be assessed in order to optimize use of resources for seismic zonation. Hence in the present study, appropriate techniques for site characterization and earthquake hazard estimation for regions at different scale levels were determined. Using the appropriate techniques, the seismic zonation was carried out both at the micro and macro-level, incorporating all major earthquake hazards. The state of Karnataka and the Kalpakkam NPP site were chosen for the macro and micro−level seismic zonation in this study. Kalpakkam NPP site is situated in Tamil Nadu, India, 70 kilometres south of Chennai city. The NPP site covers an area of 3000 acres. The site is situated along the Eastern coastal belt of India known as Coromandel coast with Bay of Bengal on the east side. The NPP site host major facilities such as Indira Gandhi Centre for Atomic Research (IGCAR), Madras Atomic Power Station (MAPS), Fast Reactor Fuel Reprocessing (FRFC) Plant, Fast Breeder Test Reactor (FBTR), Prototype Fast Breeder Reactor (PFBR) etc. The state Karnataka lies in the southern part of India, covering an area of 1,91,791 km2, thus approximately constituting 5.83% of the total geographical area of India. Both the study areas lie in the Indian Peninsular which is identified as one of the most prominent and largest Precambrian shield region of the world. The first and foremost step towards the seismic zonation is to prepare a homogenised earthquake catalogue. All the earthquake events within 300 km radius from the boundary of two study areas were collected from various national and international agencies. The earthquake events thus obtained were found to be in different magnitude scales and hence all these events were converted to the moment magnitude scale. A declustering procedure was applied to the earthquake catalogue of the two study area in order to remove aftershocks, foreshocks and dependent events. The completeness analysis was carried out and the seismicity parameters for the two study areas were evaluated based on the complete part of earthquake catalogues. The next major step toward the estimation of earthquake hazard and seismic zonation is the identification and mapping of the earthquake sources. Three source models, mainly; 1) linear source model, 2) point source model and 3) areal source model were used in the present study for characterizing earthquake sources in the two study areas. All the linear sources (faults and lineaments) within 300 km radius from the boundary of two study areas were identified and mapped from SEISAT (2000). In addition to SEISAT (2000), some lineaments were also mapped from the works of Ganesha Raj & Nijagunappa (2004). These lineaments and faults were mapped and georeferenced in a GIS platform on which earthquake events were then super-imposed to give seismotectonic atlas. Seismotectonic atlas was prepared for both the study areas. The point source model (Costa et al. 1993; Panza et al. 1999) and areal source model (Frankel, 1995) were also adopted in this work. Deterministic and probabilistic seismic hazard analysis was found to be appropriated for micro, meso and macro-level zonations. Hence in the present study, the seismic hazard at bedrock level, both at the micro and macro-level were evaluated using the deterministic as well as the probabilistic methodologies. In order to address the epistemic uncertainties in source models and attenuation relations, a logic tree methodology was incorporated with the deterministic and probabilistic approaches. As the deterministic seismic hazard analysis (DSHA) considers only the critical scenario, knowing the maximum magnitude that can occur at a source and the shortest distance between that source and the site and the peak horizontal acceleration (PHA) at that site is estimated using the frequency dependent attenuation relation. Both for the micro as well as the macro-level, the DSHA was carried out, considering grid sizes of 0.001◦ × 0.001◦ and 0.05◦ × 0.05◦respectively. A MATLAB program was developed to evaluate PHA at the center of each of these grid points. The epistemic uncertainties in source models and attenuation relations have been addressed using a logic tree approach (Bommer et al., 2005). A typical logic tree consists of a series of nodes to which several models with different weightages are assigned. Allotment of these weightages to different branch depends upon the degree of uncertainties in the model, and its accuracy. However the sum of all weightages of different branches at a particular node must be unity. Two types of seismic sources are employed in DSHA and they are linear and smoothed point sources. Since both the types of sources were of equal importance, equal weightages were assigned to each of them. The focal depth in the present study was taken as 15 km. The attenuation properties of the region were modelled using three attenuation relations, Viz. Campbell & Bozorgnia (2003), Atkinson & Boore (2006) and Raghu Kanth & Iyengar (2007). The attenuation relation proposed by Raghu Kanth & Iyengar (2007) was given higher weightage of 0.4 since it was devel-oped for the Indian peninsular region. The attenuation relations by Atkinson & Boore (2006) and Campbell & Bozorgnia (2003) which were developed for Eastern North American shield region, shared equal weightages of 0.3. Maps showing spatial variation of PHA value at bedrock level, for both micro and macro-level are presented. Response spectra at the rock level for important location in the two study areas were evaluated for 8 different periods of oscillations, and the results are presented in this thesis. Probabilistic seismic hazard analysis (PSHA) incorporating logic tree approach was per-formed for both micro as well as macro-level considering similar grid sizes as in DSHA. Two types of seismic sources considered in the PSHA are linear sources and smoothed gridded areal sources (Frankel, 1995) with equal weightage distribution in the logic tree structure. Smoothed gridded areal sources can also account the scattered earthquake events. The hypocentral distance was calculated by considering a focal depth of 15 km, as in the case of DSHA method. A MAT-LAB program was developed for PSHA. The same attenuation relations employed in DSHA were used in PSHA as well with the same weightage allotment in logic tree structure. Considering all major uncertainties, a uniform hazard response spectrum (UHRS), showing the variation of PHA values with the mean annual rate of exceedance (MARE), was evaluated for each grid point. From the uniform hazard response spectrum, the PHA corresponding to any return period can be evaluated. Maps showing the spatial variation of PHA value at bedrock level, corresponding to 475 year and 2500 year return periods for both micro and macro-level are presented. Response spectra at the rock level for important location in two study areas were evaluated for eight different periods of oscillations, and the results are presented in this thesis. In order to assess various earthquake hazards like ground motion amplification and soil liquefaction, a thorough understanding of geotechnical properties of the top overburden soil mass is essential. As these earthquake hazards strongly depend on the geotechnical properties of the soil, site characterization based on these properties will provide a better picture of these hazards. In the present study, seismic site characterization was carried both at the micro and macro-level using average shear wave velocity for top 30 m overburden (Vs30). At the micro-level, the shear wave velocity profile at major locations was evaluated using multichannel analysis of surface waves (MASW) tests. MASW is an indirect geophysical method used in geotechnical investigations and near surface soil characterization based on the dispersion characteristics of surface waves (Park et al., 1999). The MASW test setup consists of 24-channel geophones of 4.5 Hz capacity. A 40 kg propelled energy generator (PEG) was used for generating surface wave. Based on the recordings of geophones, the dispersion characteristics of surface waves were evaluated in terms of a dispersion curve. The shear wave velocity (Vs) profile at a particular location was determined by performing inversion analysis (Xia et al., 1999). After the evaluation of V s profile at all major locations, the site characterization at the micro-level was carried out as per NEHRP (BSSC, 2003) and IBC (2009) recommendations. Maps showing the spatial distribution of various site classes at the micro-level are presented in this thesis. Standard penetration tests were also carried out in the site as part of subsurface investigation and in this study a new correlation between V s and corrected SPT-N values was also developed. Apart from carrying out site characterization, low strain soil stiffness profile was evaluated based on SPT and MASW data. In this work, seismic site characterization at the macro-level was also carried out. As it is not physically and economically viable to carry out geotechnical and geophysical testing for such a large area, like the Karnataka state, the seismic site characterization was carried out based on topographic slope maps. Wald & Allen (2007) has reported that the topographic slope is a perfect indicator of site conditions. Based on the correlation studies carried out for different regions, Wald & Allen (2007) has proposed slope ranges corresponding to each site class. In this study, the topographic map for the entire state of Karnataka was derived from ASTER Global Digital Elevation Model GDEM. This thesis also presents a comparison study between the Vs30map generated from topographic slope data and Vs30map developed using geophysical field tests, for Bangalore and Chennai. Based on this study, it is concluded that topographic slopes can be used for developing Vs30maps for meso and macro-level with reasonable accuracy. The topographic map for macro-level was generated at a grid size of 0.05◦ × 0.05◦. Based on the value of slope at a particular grid point, the Vs30for that grid point was assigned as per Wald & Allen (2007). A similar procedure was repeated for all the grid points. Spatial variation of various seismic site classes for the macro-level has been presented in this work. The site amplification hazard was estimated for both micro and the macro-level. The assessment of site amplification is very important for shallow founded structures and other geotechnical structures like retaining walls and dams, floating piles and underground structures as the possible earthquake damages are mostly due to extensive shaking. The site amplification hazard at the micro-level was estimated using 1D equivalent linear ground response analyses. The earthquake motion required for carrying out ground response analysis was simulated from a target response spectrum. 1D equivalent linear analyses were performed using SHAKE 2000 software. Spatial variations of surface level PHA values, site amplification, predominant frequency throughout the study area are presented in this work. As it is not physically viable to assess site amplification hazard at the macro-level using the 1D ground response analysis, the surface level PHA value for the entire state of Karnataka was estimated using a non-linear site amplification technique pro-posed by Raghu Kanth & Iyengar (2007). Based on the site class in which particular grid belongs and bedrock level PHA value, the amplification for that grid point was evaluated using regression equations developed by Raghu Kanth & Iyengar (2007). The liquefaction hazard both at the micro and macro-level was evaluated and included in this thesis. The micro-level liquefaction hazard was estimated in terms of liquefaction potential index (LPI) based on SPTN values (Iwasaki et al., 1982). As the LPI was evaluated by integrating the factor of safety against liquefaction (FSL) at all depths, it can effectively represent the liquefaction susceptibility of the soil column. LPI at the micro-level was evaluated by both deterministic as well as the probabilistic approaches. In the deterministic approach, the FSLat a particular depth was evaluated as the ratio of the cyclic resistance of the soil layer to the cyclic stress induced by earth-quake motion. The cyclic stress was estimated as per Seed & Idriss (1971), while the cyclic soil resistance was characterised from the corrected SPT-N values as proposed by Idriss & Boulanger (2006). However in the probabilistic method, the mean annual rate of exceedance (MARE) of factor of safety against liquefaction at different depth was estimated using SPT field test data by considering all uncertainties. From the MARE curve, the FS L for 475 year and 2500 year return period were evaluated. Once FS L at different depth were evaluated, the LPI for the borehole is calculated by integrating FS L for all depths. The liquefaction hazard at the macro-level was estimated in terms of SPT and CPT values required to prevent liquefaction at 3 m depth, using a probabilistic approach. The probabilistic approach accounts the contribution of several magnitudes acceleration scenarios on the liquefaction potential at a given site. Based on the methodology proposed by Kramer & Mayfield (2007), SPT and CPT values required to resist liquefaction corresponding to return periods of 475 years and 2500 years were evaluated at the macro-level. It has been observed that the spatial distribution of intensity of each these hazard in a region is distinct from the other due to the predominant influence of local geological conditions rather than the source characteristics of the earthquake. Hence it’ll be difficult to assess risk and vulnerability of a region when these hazards are treated separately. Thus, all major earthquake hazards are to be integrated to an index number, which effectively represents the combined effect of all hazards. In the present study, all major earthquake hazards were integrated to a hazard index value, both at the micro as well as macro-level using the Analytical Hierarchy Process (AHP) proposed by Saaty (1980). Both micro and macro-level seismic zonation was performed based on the spatial distribution of hazard index value. This thesis also presents the assessment of earthquake induced landslides at the macro-level in the appendix. Landslide hazards are a major natural disaster that affects most of the hilly regions around the world. This is a first attempt of it kind to evaluate seismically induced landslide hazard at the macro-level in a quantitative manner. Landslide hazard was assessed based on Newmark’s method (Newmark, 1965). The Newmark’s model considers the slope at the verge of failure and is modelled as a rigid block sliding along an incline plane under the influence of a threshold acceleration. The value of threshold acceleration depends upon the static factor of safety and slope angle. At the macro-level, the slope map for the entire state of Karnataka was derived from ASTER GDEM, considering a grid size of 50 m × 50 m. The earthquake motion which induces driving force on the slope to destabilize it was evaluated for each grid point with slope value 10 degree and above using DSHA. Knowing the slope value and peak horizontal acceleration (PHA) at a grid point, the seismic landslide hazard in terms of static factor of safety required to resist landslide was evaluated using Newmark’s method. This procedure is repeated for all grid points, having slope value 10 degree and above.