* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download improvement of regional seismic hazard assessment considering
2008 Sichuan earthquake wikipedia , lookup
Kashiwazaki-Kariwa Nuclear Power Plant wikipedia , lookup
April 2015 Nepal earthquake wikipedia , lookup
1880 Luzon earthquakes wikipedia , lookup
2010 Pichilemu earthquake wikipedia , lookup
1570 Ferrara earthquake wikipedia , lookup
2009 L'Aquila earthquake wikipedia , lookup
1906 San Francisco earthquake wikipedia , lookup
2009–18 Oklahoma earthquake swarms wikipedia , lookup
1992 Cape Mendocino earthquakes wikipedia , lookup
Earthquake engineering wikipedia , lookup
UNIVERSITY OF PATRAS SCHOOL OF NATURAL SCIENCES DEPARTMENT OF GEOLOGY SEISMOLOGICAL LABORATORY Master Thesis in Engineering Seismology IMPROVEMENT OF REGIONAL SEISMIC HAZARD ASSESSMENT CONSIDERING ACTIVE FAULTS By ALEXANDROS D. TSIPIANITIS Environmental Engineer, Technical University of Crete, 2013 Submitted in partial fulfillment of the requirements for the degree of Master of Science in Applied, Environmental Geology & Geophysics Supervisor: Dr. Efthimios Sokos Referee: Dr. Akis Tselentis Referee: Dr. Ioannis Koukouvelas Patras, 2015 Page intentionally left blank i AUTHOR’S DECLARATION I hereby declare that the work presented in this dissertation has been my independent work and has been performed during the course of my Master of Science studies at the Seismological Laboratory, University of Patras. All contributions drawn from external sources have been acknowledged with the reference to the literature. Alexandros D. Tsipianitis ii ACKNOWLEDGEMENTS First and foremost, I would like to express my deepest gratitude to my supervisor, Dr. Efthimios Sokos, for his continuous support of my M.Sc. study and research, for his patience, motivation and immense knowledge. He helped me significantly to develop my background in the interesting field of Engineering Seismology. Besides my supervisor, I would like to thank the co-advisor of my master thesis, Dr. Laurentiu Danciu, Post-Doctoral researcher of ETH, Zurich, for his excellent guidance and support of my overall research progress. I would also like to thank the members of the examination committee, Dr. Akis Tselentis and Dr. Ioannis Koukouvelas, for their suggestions, remarks and insightful comments. My sincere thanks goes to the staff of the Seismological Laboratory of University of Patras, Dr. Paraskevas Paraskevopoulos and the Ph.D. candidate, Mr. Dimitrios Giannopoulos, for their assistance and cooperation. They provided me an excellent atmosphere for doing research. I am also grateful to Dr. Konstantinos Nikolakopoulos for his assistance considering the GIS part of my dissertation. Last but not the least, I would like to thank my family and my friends for their continuous support throughout my studies. Alexandros D. Tsipianitis Patras, April 2015 iii ABSTRACT Seismic hazard assessment is a required procedure to assist effective designing of structures located in seismically active regions. Traditionally, in a seismically active region as Greece, the seismic hazard evaluation was based primarily on the historical seismicity, and to lesser extent based on the consideration of the geological information. The importance of the geological information in seismic hazard assessment is significant, for the reason that earthquakes occur on faults. This approach also covers areas with few instrumental recordings. Mapping, analyzing and modeling are needed for faults investigation. In the present dissertation, we examined the seismic hazard for the cities of Patras, Aigion and Korinthos, considering the seismically active faults. The active faults considered in this investigation consists of 148 active faults, for which a minimum amount of information was available (i.e. length, maximum magnitude, slip rate, etc.). For some critical parameters, e.g. slip rate, if an estimate could not be found in the literature it was calculated based on empirical laws. Specifically, the slip rate for each fault was resulted from the division of total displacement with the stratigraphic age. Two different approaches (historical seismicity, length of faults) were followed for the estimation of total displacement for each fault. A distribution of slip rates was made because uncertainties are considered. The resulted slip rates were converted into seismic activity. Thus, we were able to construct a complete database for our research. Epistemic uncertainties were accounted at both seismic source models as well as at the ground motion via a logic tree framework resulted in two different calculation procedures (including or not the b value uncertainty). The seismic hazard model was implemented following the OpenQuake open standards – NRML, and the seismic hazard computation was performed for the region of interest. The seismic hazard was quantified in terms of seismic hazard maps, hazard curves and uniform hazard spectra for the region of interest. Different intensity measure types were considered, Peak Ground Acceleration, Spectral Acceleration at two fundamental periods 0.1 and 1.0 sec. Finally, the results of this thesis were compared with the Greek Seismic Code and other seismic hazard estimations for the investigation region. iv THESIS ORGANIZATION First chapter depicts an overview of the seismic hazard methodology, with a focus on the description of the general framework and highlights of the main features. Further, the region of investigation is introduced and an overview of the existing studies considering seismic hazard assessments in the regions of Europe, Greece and Patras is provided. Second chapter describes in greater details the probabilistic framework for ground motion evaluation. The theoretical aspects are illustrated together with the key elements (e.g. uncertainty, hazard curves, earthquake models, empirical relations) with a focus on their mathematical definition. Chapter three provides an overview of the software used: the OpenQuake hazard engine. Herein, the focus is the theory, the main concepts, the structure and critical parameters, e.g. logic tree types, GMPEs, hazard calculators. Fourth chapter describes the procedures adopted for building the seismic hazard model. All active faults database used in the present dissertation is described. Approaches and empirical relations are presented for the estimation of total displacement. The definition and evaluation of slip rates are also provided. Additionally, the conversion of slip rates into activity and an implementation of magnitude-frequency distribution are presented. The seismic sources and GMPE logic trees are provided. Chapter five contains the output of the seismic hazard evaluation. Hazard maps, hazard curves and uniform hazard spectra for the region of Corinth Gulf and the cities of Patras, Aigion and Korinthos are illustrated and commented. Finally, in chapter six comparisons with previous ground motion estimates are presented. Additionally, a comparison with the Greek Seismic Code is provided. Also, the summary, conclusions and remarks are presented herein. v Contents Acknowledgements ................................................................................................................................. iii Abstract ................................................................................................................................................... iv Thesis organization................................................................................................................................... v Contents .............................................................................................................................................. … vi 1. Introduction .......................................................................................................................................1 1.1 The importance of seismic hazard analysis ...................................................................................1 1.2 Seismic hazard ...............................................................................................................................1 1.3 The importance of geology and neotectonics ...............................................................................3 1.4 The study area ...............................................................................................................................4 1.5 Previous researches .......................................................................................................................5 1.5.1 Europe .................................................................................................................................5 1.5.2 Greece .................................................................................................................................7 1.5.3 Patras................................................................................................................................ 11 2. Probabilistic Seismic Hazard Assessment (PSHA) .......................................................................... 12 2.1 Introduction ............................................................................................................................... 12 2.2 Difference between DSHA & PSHA ............................................................................................. 13 2.3 Characterization of seismic sources ........................................................................................... 13 2.3.1 Source types ..................................................................................................................... 13 2.3.1.1 Area sources ................................................................................................................. 13 2.3.1.2 Fault sources ................................................................................................................ 13 2.3.2 Estimation of rupture dimensios ...................................................................................... 14 2.4 Spatial uncertainty ...................................................................................................................... 14 2.5 Relations of magnitude recurrence ............................................................................................ 16 2.5.1 Distribution of magnitude ................................................................................................ 17 2.5.1.1 Truncated exponential model ...................................................................................... 17 2.5.1.2 Characteristic earthquake models ............................................................................... 18 2.5.1.3 Composite model ......................................................................................................... 19 2.6 Relations of empirical scaling of magnitude vs. fault area ......................................................... 20 2.7 Activity rates ............................................................................................................................... 20 2.8 Earthquake occurrences with time ............................................................................................. 23 2.8.1 Memory-less model.......................................................................................................... 23 2.8.2 Models with memory ....................................................................................................... 24 2.8.2.1 Renewal models ........................................................................................................... 24 2.8.2.2 Markov & semi-Markov models ................................................................................... 28 2.8.2.3 Slip predictable model.................................................................................................. 29 2.8.2.4 Time predictable model ............................................................................................... 30 2.9 Ground motion estimation ......................................................................................................... 30 2.9.1 Parameters of ground motion .......................................................................................... 31 2.9.1.1 Amplitude ..................................................................................................................... 31 2.9.1.2 Frequency content ....................................................................................................... 31 2.9.1.3 Duration........................................................................................................................ 32 vi 2.9.2 Empirical ground motion relations................................................................................... 32 2.9.2.1 Factors affecting attenuation ....................................................................................... 36 2.10 Hazard curves ........................................................................................................................... 38 2.10.1 Hazard disaggregation ...................................................................................................... 39 2.11 Uncertainty ............................................................................................................................... 40 2.11.1 Epistemic uncertainty ....................................................................................................... 40 2.11.2 Logic trees ........................................................................................................................ 40 2.11.3 Aleatory variability ........................................................................................................... 40 3. OpenQuake ..................................................................................................................................... 41 3.1 Introduction ................................................................................................................................ 41 3.2 OpenQuake-Hazard .................................................................................................................... 42 3.2.1 Main concepts .................................................................................................................. 43 3.3 Workflows of calculation ............................................................................................................ 43 3.3.1 Classical Probabilistic Seismic Hazard Analysis (cPSHA) .................................................. 44 3.4 Description of input .................................................................................................................... 44 3.5 Typologies of seismic sources ..................................................................................................... 45 3.5.1 Description of seismic sources typologies........................................................................ 45 3.5.1.1 Simple fault sources ..................................................................................................... 46 3.6 Description of logic trees ............................................................................................................ 46 3.7 The PSHA Input Model (PSHAim) ............................................................................................... 48 3.7.1 The seismic sources system.............................................................................................. 48 3.7.1.1 Logic tree of seismic sources ........................................................................................ 48 3.7.1.2 Supported branch set typologies ................................................................................. 49 3.7.2 The system of ground motion .......................................................................................... 49 3.7.2.1 The logic tree of ground motion .................................................................................. 50 3.8 Calculation settings ..................................................................................................................... 50 3.9 The Logic Tree Processor (LTP) .................................................................................................. 51 3.9.1 The logic tree Monte Carlo sampler ................................................................................. 51 3.9.1.1 The sampling of seismic source logic tree .................................................................... 51 3.9.1.2 The sampling of ground motion logic tree ................................................................... 51 3.10 The earthquake rupture forecast calculator ............................................................................ 52 3.10.1 ERF creation-fault sources case........................................................................................ 52 3.11 Calculators of seismic hazard analysis ..................................................................................... 52 3.11.1 cPSHA calculator............................................................................................................... 53 3.11.1.1 Calculation of PSHA - Considering a negligible contribution from a sequence of ruptures in occurrence t ............................................................................................... 53 3.11.1.2 Calculation of PSHA – Accounting for contributions from a sequence of ruptures in occurrence t ................................................................................................................. 54 4. Description of methodology ........................................................................................................... 55 4.1 Introduction ................................................................................................................................ 55 4.2 The Greek Database of Seismogenic Sources (GreDaSS) ........................................................... 56 4.2.1 Introduction...................................................................................................................... 56 4.2.2 Types of seismogenic sources .......................................................................................... 57 vii 4.2.3 Properties of seismogenic sources ................................................................................... 58 4.2.4 Parameters of seismogenic sources ................................................................................. 61 4.2.4.1 Individual Seismogenic Sources (ISSs) ......................................................................... 61 4.2.4.2 Composite Seismogenic Sources (CSSs) ...................................................................... 62 4.3 Application of GIS ....................................................................................................................... 62 4.4 Earthquake scaling laws .............................................................................................................. 65 4.4.1 Wells & Coppersmith (1994) ........................................................................................... 65 4.4.1.1 Displacement per event (MD) Vs. Magnitude (M) ...................................................... 65 4.4.1.2 Maximum displacement (MD) Vs. Rupture length (SRL) ............................................. 66 4.4.1.3 Rupture width (RW) Vs. Magnitude (M) ...................................................................... 66 4.4.2 Pavlides & Caputo (2004) ................................................................................................ 66 4.5 Estimation of slip rate - Approaches........................................................................................... 66 4.5.1 Approach 1 – Historical seismicity ................................................................................... 67 4.5.2 Approach 2 – Length of faults .......................................................................................... 68 4.6 Estimation of minimum & maximum fault depth ....................................................................... 69 4.7 Fault characterization ................................................................................................................. 69 4.7.1 Slip rate evaluation........................................................................................................... 69 4.7.2 Conversion of slip rates into seismic activity ................................................................... 70 4.7.3 Magnitude-Frequency Distribution (MFD) ....................................................................... 71 4.8 Model implementation ............................................................................................................... 72 4.9 Configuration .............................................................................................................................. 74 5. Results ............................................................................................................................................. 75 5.1 Model A: mean b-value (no-uncertainty) ................................................................................... 75 5.1.1 Hazard maps of Corinth Gulf ............................................................................................ 75 5.1.2 Hazard curves of Patras .................................................................................................... 77 5.1.3 Hazard curves of Aigion .................................................................................................... 78 5.1.4 Hazard curves of Korinthos .............................................................................................. 79 5.1.5 Uniform hazard spectra.................................................................................................... 80 5.2 Model B: including b-value uncertainty...................................................................................... 82 5.2.1 Hazard maps of Corinth Gulf ............................................................................................ 82 5.2.2 Hazard curves of Patras .................................................................................................... 84 5.2.3 Hazard curves of Aigion .................................................................................................... 85 5.2.4 Hazard curves of Korinthos .............................................................................................. 86 5.2.5 Uniform hazard spectra.................................................................................................... 87 5.3 Comparison ................................................................................................................................. 88 5.3.1 Difference between 10% probability of exceedance for mean PGA values between Run #1 And Run #2 .................................................................................................................. 88 5.3.2 Difference between 2% probability of exceedance for mean PGA values between Run #1 And Run #2 ....................................................................................................................... 88 5.4 Comparisons with the Greek Seismic Code ............................................................................... 89 5.5 Comparisons with previous studies ........................................................................................... 91 6. Summary and conclusions .............................................................................................................. 95 6.1 Summary ..................................................................................................................................... 95 viii 6.2 Results ......................................................................................................................................... 96 Appendix................................................................................................................................................ 97 References ........................................................................................................................................... 111 ix CHAPTER 1 INTRODUCTION 1.1 The importance of seismic hazard analysis Many regions around the globe are prone to be affected by earthquakes. The threat to human activities is something that cannot be omitted, so this triggers a more careful structure design (Kramer 1996; Koukouvelas et al., 2010). Therefore, an earthquake-resistant building design has the aim to produce a structure which can sustain a sufficient level of ground motion, without presenting excessive damages (Kramer, 1996; Stein & Wysession, 2003; Baker, 2008). Generally, the construction of fully earthquake-resistant structures is generally impossible (Komodromos, 2012). For the reasons mentioned above, the seismic hazard analysis (SHA) plays a critical role to the quantitative estimation of the design seismic load, which is related with the seismicity of the study area, the level of structure‟s vulnerability and the danger that incurs to humans, which are mainly exposed to the seismic events (Pavlides, 2003; Pitilakis, 2010). The application of seismic hazard analysis is separated in two categories, which are mostly implemented for the description of earthquake ground motions (Kramer, 1996; Gupta, 2002; Pavlides, 2003; Orhan et al., 2007). The first category, defined as “deterministic method” or DSHA (Deterministic Seismic Hazard Analysis), is applied by using a historical seismic event that occurred in the past or a specific seismic fault that is seismically active and it has completely identified spatial and geometric parameters. The second category, defined as “probabilistic method” or PSHA (Probabilistic Seismic Hazard Analysis), takes into account the direct uncertainties relevant to the seismic magnitude and the time that of occurrence, using a strict mathematical way (Kramer, 1996; Koukouvelas et al., 2010; Pitilakis, 2010). 1.2 Seismic hazard The estimation of hazard caused by seismic events is one of the main purposes of earthquake prediction, especially referred to the realm of long-term prediction (Scholz, 1990). Generally, 1 CHAPTER 1 – INTRODUCTION macro or microzoning maps of a site are some relative applications (Gupta, 2002). Seismic hazard is defined as “the probability of a certain ground motion parameter to exceed a given value, for a specific period of time” (Tselentis, 1997; Papazachos et al., 2005; Godinho, 2007; Tsompanakis et al., 2008; Koukouvelas et al., 2010; Pitilakis, 2010; Koutromanos & Spyrakos, 2010). The ground motion parameter can be expressed through the seismic strain or the logarithm of ground acceleration and the time period can be considered as a year or the lifetime of a conventional building (i.e. 50 years) (Papazachos et al., 2005). Figure 1.1: Example of seismic hazard plot – PGA (Peak Ground Acceleration) vs. Annual frequency (Koutromanos & Spyrakos, 2010). Generally, seismic hazard depends on: the seismicity of the study area, the source-target distance, the local site conditions. The local site conditions (Fig. 1.2) can affect in significant extent the surface ground motion considering the following ways (Sanchez-Sesma, 1986; Papazachos et al., 2005; Psarropoulos & Tsompanakis, 2011): 1. The amplification (or the de-amplification, for the case of soft soils and earthquakes of large magnitude) of ground motion. 2. The extension of seismic duration. 3. The change of frequency spectrum. 4. The spatial variability of the ground response. 2 CHAPTER 1 – INTRODUCTION Figure 1.2: Main seismic actions (Tsompanakis & Psarropoulos, 2012). The arguments mentioned above cannot be neglected for cases such as the seismic design of high-risk structures (e.g. hospitals, nuclear power plants, dams), seismic risk assessment and microzonation studies (Esteva, 1977; Ruiz, 1977; Gupta, 2002; Klugel, 2008; Koutromanos & Spyrakos, 2010). 1.3 The importance of geology and neotectonics The estimation of seismic hazard for an area demands the specification and mapping of all the possible seismic sources, and the active faults that can trigger capable seismic tremors (Green et al., 1994; Pitilakis, 2010). The seismic source definition and the history of the seismicity of a region are very important parameters. The identification, the definition and the mapping of the seismic sources is based on the synthesis and analysis of a database, whose main characteristics are the following (Pitilakis, 2010): the historical seismicity of the study area, the information of instrumental recordings, the geological study of the area, the information related to neotectonics, the information from paleoseismological investigations (Fig. 1.3). 3 CHAPTER 1 – INTRODUCTION Figure 1.3: Paleoseismological investigation of the Eliki fault, Gulf of Corinth, Greece (Koukouvelas et al., 2000). 1.4 The study area The study area of this dissertation is the Corinth Gulf (CG) which contains the city of Patras, Aigion & Korinthos (Fig. 1.4). All of them are located in the north part of Peloponnese coast. Corinth Gulf is a very seismic prone area characterized by a high rate of deformation rates (Pantosti et al., 2004). The CG‟s length is approximately 115 km and its width ranges from 10 to 30 km (Stefatos et al., 2002). This region includes many normal onshore & offshore active faults that have played an important role to the geomorphological changes of the shorelines and landscapes (Koukouvelas et al., 2005). The most recent damaging seismic events were the 1981 earthquake sequence of Corinth and the 1995 earthquake of Aigion (Pantosti et al., 2004). Figure 1.4: The Corinth Gulf including the active faults from the database. 4 CHAPTER 1 – INTRODUCTION 1.5 Previous researches 1.5.1 Europe In this subchapter, some case studies on seismic hazard estimation are presented. Generally, many seismic hazard assessments have been carried out for the continent of Europe (ChungHan, 2011). It is worth mentioning the most important investigations: In the framework of Global Seismic Hazard Assessment Program (GSHAP, Fig. 1.5), a study was done for Europe and the Mediterranean region (Grunthal et al., 1999a,b; ChungHan, 2011). Figure 1.5: PGA (horizontal) seismic hazard map for an occurrence rate of 10% within 50 yearsGSHAP for the Mediterranean region (Grunthal et al., 1999b). Project SESAME (Seismotectonic & Seismic Hazard Assessment of the Mediterranean basin, Fig. 1.6), extended for entire Europe (Jimenez et al., 2003; Chung-Han, 2011). 5 CHAPTER 1 – INTRODUCTION Figure 1.6: ESC-SESAME hazard map for the European & Mediterranean region (Jimenez et al., 2003, www.ija.csic.es). Project SHARE (Seismic Hazard Harmonization in Europe, Fig. 1.7), which is the most updated assessment until now. A probabilistic approach was used and three interpretations of earthquake rates have been applied in the current project (Giardini et al., 2013): 1. The historical seismicity of moderate to large seismic events. A SHARE European Earthquake Catalog (SHEEC) was compiled, which contains a combination of 30377 seismic events in the period 1000-2007, with Mw 3.5. 2. The European Database of Seismogenic Faults (EDSF) includes an amount of 1128 active faults with a total length of 64000 km and models related to three subduction zones. 3. The deformation rates of earth‟s crust, as studied by GPSs (Global Positioning Systems. 6 CHAPTER 1 – INTRODUCTION Figure 1.7: European seismic hazard map for PGA expected to be exceeded with a 10% probability in 50 years-Application of OpenQuake (Giardini et al., 2013, www.share-eu.org). 1.5.2 Greece Greece presents an extremely high level of seismicity, thus a lot of scientific reports dedicated to the seismic hazard analysis of this territory and the surrounding regions exist. The main studies concerning the SHA of Greece are presented below. The Greek Seismic Code (EAK 2003). Figure 1.8: The unified seismic hazard zonation of Greece, return period of 475 years (EAK, 2003). 7 CHAPTER 1 – INTRODUCTION Tsapanos et al. (2004). All seismological observations and historical instrumental recordings have been considered for this SHA. For the reason that the attenuation law was related to shallow seismic events, only the shallow shocks were taken into account in this case. Figure 1.9: Probabilistic seismic hazard map of Greece and surrounding regions for PGA values. Return period of 475 years (10% probability in 50 years) (Tsapanos et al., 2004). Danciu et al. (2007). This hazard map (Fig. 1.10) has been generated by applying well known engineering parameters. The ground motion parameters investigated in this report have been applied through the use of the attenuation equations of Danciu & Tselentis (2007). These relationships are mainly based on strong ground motion data of Greek seismic events. Figure 1.10: Seismic hazard map of Greece for PGA values and probability of 10% in 50 years. Case of ideal bedrock soil condition (Danciu et al., 2007). 8 CHAPTER 1 – INTRODUCTION Tselentis & Danciu (2010). In this study, a PSHA for Greece has been implemented including some significant engineering parameters (PGA, PGV, Arias intensity, cumulative absolute velocity) for a lower acceleration value of 0.05g. The hazard map (Fig. 1.11) has been estimated for a return period of 475 years. Figure 1.11: Probabilistic seismic hazard map (PGA), according to Tselentis & Danciu (2010). Vamvakaris (2010). The computation of the maximum expected PGA values was achieved by making various comparisons related to the choice of the suitable attenuation relationships. For each type of hypocental depth (low, intermediate, high) different equations have been applied. Figure 1.12: Values of maximum expected PGA for seven return periods (Vamvakaris, 2010). 9 CHAPTER 1 – INTRODUCTION Segkou (2010). The methodology followed in this dissertation for the PSHA of Greece (Fig. 1.13) is based on the survey and appraisal of the respective previously generated hazard maps in global scale. The PSHA is based on the evaluation of different seismic source models identified by seismological, geological and geophysical observations, in order to be suitable to the requirements of Greek region. Specifically, different processes were applied for the estimation of total expected ground motion: - The linear seismic source model, which is based on the identification of active faults through geographical, seismological and geological criteria (Papazachos et al., 2001) and associated to the seismic hazard due to shallow earthquakes. - The random seismicity model, based on the analysis of shallow earthquakes seismicity catalogue. This model corresponds to the estimation of seismic hazard related to earthquakes with magnitude of 5 to 6.5 R. - A seismic source model aiming to describe seismicity associated with the subduction zone (this seismic source model is called by Segkou as “uniform basement zone”). Figure 1.13: Seismic hazard map (PGA) for rock basement. Average return period of 475 years (Segkou, 2010). Koravos (2011). A SHA for shallow earthquakes of the Greek territory was made by applying the Ebel-Kafka method (Fig. 1.14). This method uses synthetic catalogues computed with the Monte Carlo simulation. For the estimation of seismic hazard, the Ebel-Kafka code was modified for the purposes of the attenuation relationship suitable to the Greek area. The attenuation equation 10 CHAPTER 1 – INTRODUCTION used for the PGA computation of shallow shocks was taken from Skarlatoudis et al. (2003), because it contains seismicity data from Greece. Figure 1.14: Illustration of the maximum PGA estimation considering shallow earthquakes for 1000 years seismicity data. The probability of exceedance is 10% (Koravos, 2011). 1.5.3 Patras Sokos (1998) The seismic hazard estimation for the city of Patras (Fig. 1.15) was carried out using the SEISRISK III software. This program has the ability to estimate the maximum level of ground motion depended on the attenuation relationship considering a certain probability of exceedance for a specific time period. The seismic sources that were used in this application were these proposed by Papazachos (1990), Papazachos & Papaioannou (1997) and for the seismic hazard assessment of Rio-Antirio Bridge. Three different definitions for the seismic sources were made for the research of seismic hazard dependency on the seismic sources. Figure 1.15: Acceleration curves for the city of Patras with 90% probability of exceedance for the next 50 years (Sokos, 1998) 11 CHAPTER 2 PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 2.1 Introduction As inferred by Cornell (1968) and Baker (2008), the Probabilistic Seismic Hazard Analysis (PSHA) contains two representative features, the event (how, where, when) and the resulting ground motion (frequency, amplitude, duration). These characteristics provide a methodology relative to the quantitative representation of the relationship associated with the probabilities of occurrence, the potential seismogenic sources and ground motion parameters. “PSHA computes how often a specified level of ground motion will be exceeded at the site of interest” (Godinho, 2007; Ross, 2011). The resulting information is presented by the form of return period or annual rate of exceedance. Thus, seismic hazard computations provided by PSHA that can be implemented for seismic risk assessment. Therefore, engineers possess an extremely useful tool concerning the seismic resistance of a building (Godinho, 2007; Ross, 2011). According to Reiter (1990), PSHA can be divided into four steps: 1. The first step is referred to the identification and characterization of seismic sources. This step is similar to the first step of DSHA (Deterministic Seismic Hazard Assessment), with the difference that there should be a characterization of the probability distribution of the potential rupture locations within the source. 2. Secondly, there should be a characterization of the seismicity or the distribution of earthquake occurrence. The aim of a recurrence relationship is the specification of an average rate, at which a seismic event of some size will occur. Its use is related to the characterization of the seismicity of each seismogenic source. 12 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 3. In this step, the use of predictive equations should be linked with the produced ground motion at the area by seismic events of any possible size that occurred at any potential point in each seismic zone. 4. Finally, a combination between the uncertainties in earthquake size, location and ground motion parameter prediction is made, in order to obtain the probability of exceedance of ground motion parameter during a specific period of time. 2.2 Difference between DSHA & PSHA Before the development of PSHA, the compilation of many seismic hazard assessments was under the perspective of a deterministic view, using scenarios of location and magnitude for each source in order to evaluate the ground motion design (Abrahamson, 2006; Baker, 2008). It can be stated that PSHA is an assessment which is composed of an infinite number of DSHAs, taking into account all possible seismogenic sources and scenarios of distance and magnitude (Godinho, 2007; Koukouvelas et al., 2010). 2.3 Characterization of seismic sources In this section, there is a description of the rate at which earthquakes of given dimensions and magnitudes take place in a specific location. First of all, the potential sources are identified and their dimension parameters are modeled. This requires the definition of source type and the estimation of source dimensions (Godinho, 2007; Baker, 2008; Koutromanos & Spyrakos, 2010). 2.3.1 Source types 2.3.1.1 Area sources Some seismic faults which have inadequate geological data can be modeled as area sources, based on data related to their historical seismicity. Therefore, an assumption was made that seismic zones have unique source properties in time and space. Additionally, the use of area sources is preferred at the modeling of “background zones” of seismic areas, for the purpose of the occurrence of seismic events away from known mapped active faults (Abrahamson, 2006; Baker, 2008). 2.3.1.2 Fault sources The identification and definition of the location of seismic faults is feasible, when adequate geological data is available. Despite their linear source modeling, many fault source models have multi-planar characteristics and there is an assumption for the ruptures, which implies that they are distributed over the entire fault plane (Abrahamson, 2006). 13 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 2.3.2 Estimation of rupture dimensions The fault rupture dimensions can be estimated through the following two ways (Wells & Coppersmith, 1994; Henry & Das, 2001): based on the size of fault rupture plane, or based on the size of the aftershock zone. The measurement of length of fault expression on the free surface and the estimation of the seismogenic zone, are some actions required for the estimation of fault rupture. The distinction between primary and secondary source rupture is very important for the estimation of fault rupture length. The primary source is mainly associated with the tectonic rupture, which is the fault rupture plane that intersects the ground surface. On the other hand, the secondary rupture is related to fractures caused by initial rupture effects, such as landslides, ground shaking or ruptures from earthquakes which were triggered on nearby active faults (Wells & Coppersmith, 1994; Godinho, 2007). The corner frequency fc of source spectra for large events (obtained from ground motion recordings) plays an important role concerning the estimation of rupture dimensions (Molnar et al., 1973; Beresnev, 2002). The determination of the subsurface rupture length, as indicated by the spatial pattern of aftershocks, is the second method associated with the estimation of fault‟s dimensions. The determination of rupture width can also be done through this way. Studies have shown the reliability of this method, but it is known that there are factors which contribute to its uncertainty (Godinho, 2007). According to Henry & Das (2001), in the case that time period after the main seismic event is small, the aftershock territory provides reliable estimates of rupture dimensions. 2.4 Spatial uncertainty The tectonic processes play a significant role concerning the dimensions of earthquake sources (Fig. 2.1). Earthquakes generated in zones that are too small (i.e. seismic events caused by the activity of volcanoes) are characterized as point sources. The consideration of two-dimensional (2-D) areal sources can be taken into account in the case that earthquakes can occur at several different locations and a good definition of the fault planes exists. Threedimensional (3-D) volumetric sources can be considered when there are areas where (Kramer, 1996): there is an obvious extension of the faulting, so the separation of individual fault is not possible, 14 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) there is a poor definition of earthquake mechanisms. In order to compile a seismic hazard assessment, the source zones should present a similarity to the real seismogenic source. This depends on the dimensions of the source, the study area and the completeness of source data (Kramer, 1996). It is assumed that the distribution of earthquakes usually takes place within a specific source area. Ground motion parameters are expressed by some predictive relationships in terms of some measure of source-to-site distance, so the description of spatial uncertainty should be with respect to the suitable parameter of distance. A probability density function can describe this uncertainty (Kramer, 1996). Considering the point source (Fig. 2.1a), the distance, there is an assumption that the probability that , is presented as . Therefore, is to be 1 and the probability that is to be zero. In the case of linear source (Fig. 2.1b), the probability that occurs between and is similar to the probability that an occurrence of a seismic event takes place on a small section of the fault between and , so (Kramer, 1996): () ( ) ( ) where: ( ), ( ) probability density functions for the variables and . Figure 2.1: Geometries of source zones: (a) short fault – point source, (b) shallow fault – linear source, (c) 3-D source zone (Kramer, 1996). 15 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) Figure 2.2: Source-to-site distance variations for different source zone dimensions (Kramer, 1996). ( ) () ( ) For the assumption of the uniform distribution of the earthquakes over the length of the fault, () . Since the probability density function of has the following form (Kramer, 1996): ( ) ( ) √ The evaluation of ( ) by numerical rather than analytical processes is a more straightforward way for the case of having source zones with complex geometries. 2.5 Relations of magnitude recurrence The expression of the seismicity of a source is associated with a magnitude recurrence relation, with the premise that the dimensions of the source are well-defined and a suitable magnitude scale selected. The characterization of magnitude occurrence equations is referred to the activity rate of seismogenic sources and a function which describes the magnitude distribution. The integration of magnitude distribution density function and the scale considering the activity rate are the principal elements for the computation of a recurrence relation, as the following (Godinho, 2007): ∫ ( ) ( ) 16 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) where: : the average rate of earthquakes with magnitude greater than or equal to a magnitude M, : a specified magnitude, : source‟s activity rate, ( ): magnitude distribution density function. 2.5.1 Distribution of magnitude The definition of randomness in the number of relative number of large, intermediate and small sized seismic events occurring in a given source, can be done through a probability density function. There are two model types used for the representation of magnitude distributions (Godinho, 2007): 1. The truncated exponential model. 2. The characteristic earthquake model. Studied by Youngs & Coppersmith (1985), the characteristic model is more suitable for the characterization of individual active faults. There are seismicity models that use a hybrid approach, i.e. truncated exponential model for small-to-moderate seismicity and characteristic model for large magnitudes. The resulting difference in seismic hazard between the two models depends of fault-to-site distance and acceleration level, thus, on the SHA also (Godinho, 2007). 2.5.1.1 Truncated exponential model This model, based on Gutenberg-Richter magnitude recurrence relation (Gutenberg-Richter, 1956), is described through the following equation: ( ) where: : the a-value, which represents the source activity rate, : the b-value, which represents the relative likehood of earthquakes with different magnitudes (values between 0.8-1.0). In addition, there is an alternative form of the truncated exponential model: ( ) ( ) 17 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) where: and ( ) It is obvious that earthquake magnitudes present an exponential distribution. So, the mean recurrence rate of small magnitude earthquakes is a lot larger than that of large-sized earthquakes (Godinho, 2007). Despite the fact that the application of standard Gutenberg-Richter recurrence relation has to do with an infinite range of magnitudes, the application of bounds at minimum and maximum values of magnitude is very common because there is a connection between seismic sources and the capacity for producing maximum magnitude Mmax (Godinho, 2007). From the viewpoint of engineers, earthquakes of very small magnitudes, which do not cause some type of damage to buildings, are not being taken into account (Abrahamson, 2006). The following probability density function, which uses the minimum (Mmin) and maximum (Mmax) values, is presented through an equation and a graph: ( ) ( ( ) ) ( ) Figure 2.3: Magnitude probability distribution function – truncated exponential model (Godinho, 2007). 2.5.1.2 Characteristic earthquake models These types of models are based on the hypothesis that individual faults have the tendency to generate same size, or representative earthquakes (Schwarz & Coppersmith, 1985). According to Godinho (2007), prior to 1980‟s the magnitude associated with the characteristic earthquake was based on the assumption that some fraction of total fault length would rupture (i.e. ¼ of total fault‟s length) (Abrahamson, 2006). Nowadays, the prevailing theory states the 18 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) separation of active fault into segments, which can be used as boundaries of rupture geometry (Abrahamson, 2006). The characteristic earthquake model includes a type named as model of “maximum magnitude” (Godinho, 2007). This form is not applicable to smaller-to-intermediate events. The basic idea refers to the assumption of Abrahamson (2006), which supports that all seismic energy is derived from characteristic earthquakes. According to Figure 2.4, this model can be used only for a narrow range of magnitudes. Figure 2.4: Magnitude probability density function – truncated normal model (Godinho, 2007). 2.5.1.3 Composite model Previous investigations have applied a combination of the characteristic and truncated exponential model, for the accommodation of distribution related to large magnitude earthquakes (Youngs & Coppersmith, 1985). Therefore, the modeling of characteristic earthquake behavior is allowed, without other magnitude events being excluded. The magnitude density function concerning this model (Fig. 2.5) presents an exponential distribution with some magnitude, M, and a uniform distribution of given width, which is centered on the mean characteristic magnitude. Additionally, an extra constraint in order to define the relative amplitudes of two distributions is required (Godinho, 2007). As noted by Youngs & Coppersmith (1985), the relative amount of the released seismic moment through small magnitude events and characteristic earthquakes are represented by this constraint. This model is based on empirical data. 19 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) Figure 2.5: Magnitude probability density function – composite characteristic & exponential model (Godinho, 2007). 2.6 Relations of empirical scaling of magnitude vs. fault area Models of magnitude distribution, like those presented in the previous subchapter, have some limits between minimum and maximum magnitude values. The minimum level of energy release expected to cause damage to buildings is represented by the minimum magnitudes (Abrahamson, 2006). On the other hand, maximum magnitudes refer to stress drop and fault geometry. Specifically, the stress drop is a parameter which describes the distribution of seismic moment release in time and space (Godinho, 2007). Below, there is a table (Table 2.1) that presents some scaling relations between rupture dimension and magnitude (Godinho, 2006): Wells & Coppersmith (1994) All fault types Wells & Coppersmith (1994) Strike-slip Wells & Coppersmith (1994) Reverse Ellsworth (2001) Strike-slip for A>500km2 Somerville et al. (1999) All fault types ( ) ( ) ( ) ( ) ( ) Table 2.1: Magnitude (M)-area (A) scaling equations (Godinho, 2007). 2.7 Activity rates While relative earthquake rate at several magnitudes is provided by magnitude distribution models for the complete representation of source seismicity through a recurrence relation, there is a requirement of activity rate (Godinho, 2007). According to Godinho (2007), activity rate is the rate of earthquakes above a minimum magnitude. The activity rate of a seismic source can be defined through the following two approaches: 20 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 1. Seismicity There is a possibility of estimating the activity rates which are based on recordings from earthquake catalogues. This is applicable to seismically active areas where there is availability of significant historical data. When the exponential distribution is fitted to the historical data, the computation of seismicity parameters (b-value in Gutenberg-Richter‟s relation, activity rate) can be retrieved by using a regression analysis (maximum likelihood method) (Godinho, 2007). In the case of being based on earthquake catalogues, in order to provide data related to earthquake occurrence, it must be noted that there is a dependence of the accuracy of the estimated activity rate with catalogues‟ reliability. Thus, there must be a completeness and adequacy study of the earthquake data but also an exclusion of the aftershocks and foreshocks from the study (dependent events) (Abrahamson, 2006; Godinho, 2007). 2. Geological information-slip rate Slip rate can be useful to the estimation of activity rates for other earthquake models (characteristic earthquake model). This is feasible when there is adequacy of historical data for the estimation of activity rates (Youngs & Coppersmith, 1985). The advantage of this method is its application, because it covers seismic areas with few recordings related to earthquake occurrence (Godinho, 2007). It also provides further information concerning the recurrence that allows an improved computation of mean earthquake frequency (Youngs & Coppersmith, 1985). A reliable estimate of slip rate must be based both on historical and geological data (Godinho, 2007). Youngs & Coppersmith (1985) have made some hypotheses concerning the estimations of these parameters: The consideration of all observed slip as seismic slip, which can be assumed as an effect of creep. Short term fluctuations are not considered, because slip rate represents an average value. Slip rates at seismogenic depths and along the entire fault length are assumed to be represented by all surface measurements. The computation of activity rate is achieved by balancing the long term accumulation of seismic moment with is long term release (Godinho, 2007). According to Aki (1979), the rate of moment build up is expressed through this relation: 21 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) ̅ ̅̇ ( ) where: ̅̇ : the slip rate (cm/year), : the fault rupture area, : the shear modulus. If a scaling relation is used for the definition of fault‟s characteristic magnitude, ( ) ( ) The amount of moment released by an individual characteristic earthquake can be expressed by using a moment-magnitude relation. ( ( ) ( ) ) ( ) The product of the moment release per characteristic earthquake and earthquake occurrence rate ( ) equals the total rate of moment release. ̇ ( ) If the rate of moment release is equated with the rate of moment build-up, the direct estimation of activity rate is the next step. ̇ ̇ ( ̇ ( ̇ ⁄ ) ⁄ ) ( ) 22 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 2.8 Earthquake occurrences with time When the computation of recurrence rate of a given magnitude seismic event has been made, the next step is the conversion of this rate into a probability of earthquake occurrence (Godinho, 2007). A hypothesis concerning the earthquake occurrence with time is required, especially if a “memory” or “memory-less” pattern is followed by a process of earthquake occurrence (Godinho, 2007). For a better understanding of the physical process of earthquake occurrence, the theory of elastic rebound will be described. First introduced by Reid (1911) and also presented by Kramer (1996), the theory refers that “the occurrence of earthquakes is a product of the successive build-up and release of strain energy in the rock adjacent to faults”. The setup of strain energy is an outcome of the movement of earth‟s tectonic plates. This movement causes shear stresses increased on fault planes, which are considered as plates‟ boundaries (Godinho, 2007). In the case that shear stresses reach the maximum shear strength of rock, there is failure and release of the accumulated strain energy. A strong rock will rupture rapidly and the cause will be the sudden release of energy in the form of earthquake (Kramer, 1996). 2.8.1 Memory-less model The assumption that earthquake process is memory-less is a basic feature of many PSHAs. This means that no memory of time, location and size of former events exists. It can be said that there is no dependence between the probability of an earthquake occurring in a given year and the elapsed time since the previous seismic event (Godinho, 2007). Therefore, an exponential distribution of earthquake recurrence intervals is characteristic of the Poisson process, which defines the occurrence of earthquakes (Godinho, 2007). ( ) ( ) ∫ ( ) ( ∫ ) ( ) where: : the recurrence rate, : time between events. 23 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) Figure 2.6: Probability density function of earthquake occurrence - exponential distribution model (Godinho, 2007). By using the probability theorem of Bayes, the expression of probability of an earthquake occurrence within years from former events is the following: [ ] [ ] [ ] ∫ ( ) ∫ ( ) ( ) ( ) ( ) ( ) where: : the elapsed time since the former seismic event, : the intermit time between events. The equation changes its form when there is evaluation of the probability expression using the cumulative distribution function, which is related to the assumption of Poisson: ( [ ) ( ] ) ( ) It can be noticed that the time which remains since the last earthquake ( ) does not exist anymore in the probability expression. This demonstrates the nature of “memory-less” model (Godinho, 2007). The hazard function of exponential distribution can be represented: ( ) ( ) ( ) ( ) 2.8.2 Models with memory 2.8.2.1 Renewal models A conventional way for the representation of earthquake occurrence with time is to assume it presents some periodicity (Godinho, 2007). In contrast with Poisson model, which supports the hypothesis that earthquake occurrence intervals are exponentially distributed, different 24 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) distributions are applied by renewal models that allow the increase of the probability of occurrence ( ) with elapsed time since the former earthquake (Cornell & Winterstein, 1988). Four types of typical distributions concerning the earthquake occurrence are examined: Lognormal, Brownian Time Passage, Weibull, Gamma. The main characteristics of most renewal model distributions are two statistical parameters, the covariance and the mean (Godinho, 2007). The first parameter is related to the measure of periodicity of earthquake recurrence intervals. The second parameter is associated with the average elapsed time between events (Cornel & Winterstein, 1988; Godinho, 2007). (a) Lognormal This distribution is one of the most ordinary distributions practically used: ( ) √ ( ( ) ) ( ) Figure 2.7: Probability density function of earthquake occurrence - lognormal distribution model (Godinho, 2007). It is worth to state that this type of mathematic distribution has some important parameters, such as the median ( ) and the standard deviation ( ). The relations which describe these parameters are the following (Godinho, 2007): 25 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) ̅ ( ( ) ) √ ( ) ( ) (b) Brownian Passage Time This category of distribution is also known as the Wald or Gaussian distribution. The basic parameters of Brownian Passage Time (BPT) are the mean recurrence interval ( ̅ ) and parameter, which represents the aperiodicity (Godinho, 2007). ( ) √ ̅ * ( ̅) + ̅ ( ) Figure 2.8: Probability density function of earthquake occurrence - BPT distribution model (Godinho, 2007). Examined by Matthews et al. (2002), the BPT distribution model is applied in the characterization of earthquake occurrence using a Brownian relaxation oscillator, which is represented by the state variable ( ). ( ) ( ) ( ) 26 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) Figure 2.9: Example of load state paths - Brownian relaxation oscillator (Matthews et al., 2002). (c) Weibull & Gamma These distributions have some similarities related to their general form and relation to the exponential density distribution. The constants and are associated with the variation and the mean distribution (Godinho, 2007): ( ) ( ( ) ( ) ( ) ) ( ) ( ) ( ) ( ) Figure 2.10: Probability density function of earthquake occurrence - Weibull distribution model (Godinho, 2007). 27 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) Figure 2.11: Probability density function of earthquake occurrence - Gamma distribution model (Godinho, 2007). 2.8.2.2 Markov & semi-Markov models Markov property is a main characteristic of many earthquake occurrence models, which are based on stochastic processes. Therefore, this transitional probability is conditional only on the present state. It is also independent of the process‟s state in the past (Patwardhan et al., 1980; Godinho, 2007). ( ) ( ) ( ) Figure 2.12: Schematic representation – semi Markov process (Patwardhan et al., 1980). Developed by Patwardhan et al. (1980) and also noted by Votsi et al. (2010), these models of earthquake occurrence apply this primary Markov property of one-step memory. The modeling of waiting time and size of successive earthquakes is allowed from the application of semi-Markov properties in earthquake occurrence models (Godinho, 2007). 28 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 2.8.2.3 Slip predictable model The dependence of future events on time of the last appearance is one conventional property of most earthquake occurrence memory models (Godinho, 2007). The magnitude of a successive earthquake, which is reflected by the amount of the released stress, consists of a function only of the time elapsed since the last earthquake. This is based on the hypothesis that stress accumulates at a stable rate for some time period and is independent of the former seismic event‟s magnitude (Kiremidjian & Anagnos, 1984). This shows the representation of a positive “forward” correlation between successive magnitudes and inter-arrival times, which are considered to be distributed in a random way (Godinho, 2007). Developed by Kiremidjian & Anagnos (1984), a schematic representation of the model is shown in Figure 2.13: Figure 2.13: Slip-predictable model: (a) time history of stress release and accumulation (b) relationship between time between seismic events and coseismic slip (c) sample path for the Markov renewal process (Kiremidjian & Anagnos, 1984). Below there is an illustration of the comparison between the Poisson and the slippredictable model. Figure 2.14: Comparison between Poisson and slip-predictable model (Kiremidjian & Anagnos, 1984). 29 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 2.8.2.4 Time predictable model Based on the hypothesis of time-predictable behavior, an alternative model has been developed while slip-predictable models use the time between events for the estimation of earthquake‟s magnitude (Godinho, 2007). In time-predictable models the information is provided by the magnitude of last earthquake. This means a correlation between earthquake size and intermit times (Godinho, 2007). Presenting many similarities to the slip-predictable model, Figure 2.15 is a schematic illustration of the corresponding time-predictable model: Figure 2.15: Time-predictable model: (a) time history of stress release and accumulation (b) relationship between time between seismic events and coseismic slip (c) sample path for the Markov renewal process (Kiremidjian & Anagnos, 1984). 2.9 Ground motion estimation As studied by Boore (2003), the application of ground motion estimation takes place in structure‟s design. This is feasible by using the existing building codes or the site-specific structures‟ design. Despite the efforts related to the gathering of more ground motion data in seismically active regions, it can be said that there are insufficient amount of data considering the empirical computation of design ground motions (Godinho, 2007). Therefore, many scientific projects have been devoted to the development of the estimation of ground motion parameters, which will be practical for structures‟ design based on the features of seismic sources, such as distance or magnitude (Godinho, 2007). 30 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 2.9.1 Parameters of ground motion 2.9.1.1 Amplitude Peak horizontal acceleration is a basic parameter which is used in the characterization of ground motion amplitude. Peak ground velocity, which is less sensitive to high frequencies, is applicable for the computation of structures‟ ground motions, which are vulnerable to frequencies of intermediate level (tall flexible structures) (Godinho, 2007). 2.9.1.2 Frequency content As defined by Godinho (2007), the way that ground motion amplitude is distributed amongst different frequencies is described by the frequency content. Its definition can be through different types of spectra and spectral parameters. Studied by Kramer (1996), a plot of Fourier amplitude represents a Fourier spectrum defined as the product of performing a Fourier time series‟ transformation. Immediate indications considering the ground motion‟s frequency content are given by the spectrum of Fourier (Godinho, 2007). The power spectrum is another type of spectrum which is used in the description of frequency content. It allows the computation of some statistical parameters used in stochastic methods for the development of ground motion estimation, with the premise that ground motion is characterized as a random process (Godinho, 2007). The maximum response of SDOF (Single Degree Of Freedom, Fig. 2.16) system containing a specific level of viscous damping (e.g. 5%) as a function of natural frequency is described by a response spectrum (Fig. 2.16, 2.17). It is commonly applicable to structural design and engineering purposes. The illustration of response spectrum is on tripartite logarithm scale, including in the same plot the parameters of velocity, acceleration response and peak displacement (Godinho, 2007). Figure 2.16: SDOF system (www.scielo.org.za). 31 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) Figure 2.17: Response spectrum (Godinho, 2007). 2.9.1.3 Duration The ground motion‟s duration is an important parameter related to the prevention of damage, which is caused by physical processes that are sensitive to the amount of load reversals (e.g. the degradation of stiffness and strength, the development of pore water pressuresliquefaction). There is also a correlation between the duration of ground motion and the length of rupture. Therefore, there is a proportion related to the parameters of an event‟s magnitude and the duration of ground motion. Specifically, when the size of an earthquake increases, the duration of the resulting ground motion increases too (Godinho, 2007). Through the bracketed duration, the duration can be defined as the time between the first and last exceedance of some threshold acceleration‟s value (e.g. 0.05g) (Bolt, 1969). The significant duration is an additional applicable parameter of duration, defined as the measure of time in which there is dissipation of a specified energy amount (Godinho, 2007). Another parameter, which is conventially used in determining liquefaction potential, is the equivalent number of ground motion‟s cycles, which consists an alternative expression of duration (Stewart et al., 2001). 2.9.2 Empirical ground motion relations A probability distribution function of a specific ground motion parameter (e.g. response spectra, peak acceleration) is a form that often characterizes the ground motions (Godinho, 2007). Equations named as attenuation relations or Ground Motion Prediction Equations (GMPE), which are derived through regression analysis of empirical data, determine some statistical moments such as standard deviation and median. These moments are based on 32 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) seismological parameters (source-to-site distance, magnitude). Table 2.2 presents some models for ground motion attenuation in active seismic areas: Magnitude Range Distant Range (km) Distance Measure Site Parameters Other Parameters 5.5-7.5 0-100 rjb 30m-Vs Fault type 4.7-8.1 3-60 rseism Soft rock, hard rock, depth to rock Fault type, hanging wall >4.7 0-100 r Soil/rock Fault type, hanging wall 4.0-8.0 0-100 r Soil/rock Fault type 4.6-7.4 1-100 r Rock only Fault type Atkison &Boore (1997) Campbell (1997, 2000, 2001) Abrahamson & Silva (1997) Sadigh et al.(1997) Idriss (1991, 1994) Table 2.2: Attenuation models for horizontal spectral acceleration in active fault areas (Godinho, 2007). The expression of the attenuation equation‟s general form is the following: ( ) ( ) ( ) ( ) ( ) ( ) where: : parameter of ground motion amplitude, : constants determined by regression analysis, : moment magnitude, : source to site distance (Fig. 2.18), : factor accounting for local site conditions, : factor accounting for fault type (e.g. reverse, strike-slip), : factor accounting for hanging-wall effects. The basis for most attenuation equations is expressed through a number of assumptions (Stewart et al., 2001): Uncertainty in ground motions The uncertainty or variability ( or ) in ground motion amplitudes and the mean ground motion ( ) are defined by attenuation relations. It is assumed that ground motion amplitudes 33 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) are lognormally distributed, so ( ) and ( ) consist the representations of mean and uncertainty. Magnitude dependence Moment magnitude and other magnitude scales are derived using the logarithm of peak ground motion parameters. Therefore, there is the hypothesis which supports that ( ) is proportional to the magnitude of the event ( ). Radiation damping The energy, which is released by a seismic fault during the occurrence of a seismic event, is radiated out through traveling body waves. When they travel away from the seismogenic source, there is a phenomenon called “radiation damping” which describes the reduction of wave amplitudes at a rate of ⁄ ( : source-to-site distance). Figure 2.18: Measures of source-to-site distance – ground motion attenuation models: (a) vertical faults, (b) dipping faults (Godinho, 2007). 34 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) Factors that affect attenuation Various factors associated to site and source characteristics affect the attenuation of ground motions. Therefore, a reference model is implemented in order to examine the influence on the attenuation of ground motions. The model introduced by Campbell & Bozorgnia (2003), consists of near-source horizontal and vertical ground motion attenuation relations for 5% damped pseudoacceleration response spectra and peak ground acceleration. ( ) √ ( ) ( ) ( ) ( ) ( ) It is observable that this model has a similar form to the equation presented above (2.27). Figure 2.19 presents two examples: M=7.5 and M=5.5 for Peak Spectral Acceleration (PSA) of 0.1 sec and Peak Ground Acceleration (PGA). Figure 2.19: Attenuation relations: (a) peak spectral acceleration, (b) peak ground acceleration (Campbell & Bozorgnia, 2003). 35 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 2.9.2.1 Factors affecting attenuation 1. Site conditions Many forms can represent the effects of local site conditions, starting from a simple constant till more complex functions (Godinho, 2007). There are some models applied for a simple soil/rock soil classification (Abrahamson & Silva, 1997; Sadigh et al., 1997), but others use more quantitative methods of classification, such as the 30m shear wave velocity (Atkinson & Boore, 1997). Generally, there is a hypothesis which supports that standard error in attenuation is unaffected by site conditions (Godinho, 2007). Figure 2.20: Peak spectral acceleration (damping=5%) using Campbell & Bozorgnia ground motion attenuation – effects of site conditions (Mw=7.0, rseis=10km, strike-slip fault) (Campbell & Bozorgnia, 2003). 2. Near-fault effects Many studies, such as Campbell & Bozorgnia (2003), have shown that near-fault effects on ground motion play a very important role. These surveys have concluded that there is a sensitivity of ground motion at near-source site to what is considered as “rupture directivity”. The long period energy of ground motion and the duration are affected by this parameter 36 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) (Godinho, 2007). The phenomenon which takes place when there is fault propagation towards the site is named “forward directivity”. Primarily, its effects are founded in the horizontal direction normal to fault rupture. Therefore, shock wave effects characterize the ground motion, which is associated with a short duration and large amplitudes at intermediate to long periods. On the other hand, a relatively low amplitude and long duration describes the ground motions, which are affected by backward directivity (Godinho, 2007). 3. Tectonic regime The tectonic region, in which the seismogenic sourced is located, is one of the most basic factors that affect the features of ground motion. For each subduction, stable continental and active region zones, there is a development of some attenuation relations. A development of a large proportion of attenuation equations is observed too, because of the specific amount of the available ground motion data (Godinho, 2007). There is not availability of very strong motion data for the case of stable continental areas. Therefore, for these areas the basis of attenuation relations refers to simulated motions instead of the available recordings (Atkinson & Boore, 1995-1997b; Toro et al., 1997). 4. Focal mechanism-fault type As studied by Boore (2003), ground motion parameters (frequency content, amplitude) are influenced by faulting mechanism. Strike slip faults can be used as a reference of attenuation relations and additional factors. A larger proportion of higher levels of frequency content for thrust and reverse active faults and higher mean ground motion are included in some observations of fault-type effects (Godinho, 2007). Figure 2.21: Peak spectral acceleration (damping=5%) using Campbell & Bozorgnia ground motion attenuation – effects of faulting mechanism (Mw=7.0, rseis=10km, firm soil) (Campbell & Bozorgnia, 2003). 37 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 5. Hanging wall effect Abrahamson & Somerville (1996) have concluded that sites which are located over the hanging wall of dipping faults present a considerable increase in ground motions. The experience (e.g. Northridge earthquake, 1994) has shown that this increase can be as much as 50% (Abrahamson & Silva, 1997). 2.10 Hazard curves The determination of the final seismic hazard can be done when distribution functions compute and characterize the ground motion estimates. The final step defines the frequency that a significant level of ground motion (peak ground acceleration, duration, displacement) will be exceeded at an area of interest (Godinho, 2007). The following equation describes the individual hazard of a single seismogenic source: ( ) ( ) ∫ ∫ ( ) ( ) ( ) ( ) ( ) where: : annual rate of events or return period, : level of ground motion, : specified level of ground motion to be exceeded, : magnitude, : distance, : number of standard deviation. The source-to-site distance, the ground motion and the probability density functions for magnitude are integrated over the above relation. The contribution of a single seismogenic source is reflected by the hazard expression mentioned above. In addition, a sum of total hazard contributions for each individual source is necessary, for the case of multiple seismic sources consideration (Godinho, 2007). ( ) ∑ ( ) ( ) 38 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) Then, the value of return period or annual rate of events must be converted into a form of probability. The likelihood that the ground motion will exceed the level at least once during a significant time interval is reflected by this probability (Godinho, 2007). Figure 2.22: Hazard curves for spectral period of 2 sec – individual source (McGuire, 2001). 2.10.1 Hazard disaggregation According to Bazzuro & Cornell (1999), disaggregation of hazard is a procedure that indicates the greatest contribution to the hazard. It is completed using a two-dimensional disaggregation into bins of different source-to-site distances and earthquake sizes. Then, Figure 2.23 represents the disaggregation of hazard corresponding to the total contribution for source hazard curves, shown in the previous graph (Fig. 2.22). Figure 2.23: Disaggregation of hazard for spectral period of 2 sec and ground motion level of 0.5g (McGuire, 2001). 39 CHAPTER 2 – PROBABILISTIC SEISMIC HAZARD ASSESSMENT (PSHA) 2.11 Uncertainty The definition and treatment of uncertainties are some important features of PSHAs. In the realm of structural system designing, the limitation of uncertainty is a very crucial and considerable factor (Tsompanakis et al., 2008). Two types of uncertainties are involved: epistemic uncertainty and aleatory variability (Godinho, 2007). 2.11.1 Epistemic uncertainty For the reason that is a product of limited knowledge and data, epistemic uncertainty is usually referred to as scientific uncertainty. Generally, this category of uncertainty can be reduced as more information becomes available and the use of alternative models is one of its characteristics (Godinho, 2007). 2.11.2 Logic trees Logic trees are a basic characteristic of PSHAs (Fig. 2.24). They are useful for the determination of design ground motions (Bommer & Scherbaum, 2013). The use of a logic tree is an ordinary way for handling the epistemic uncertainty related to the inputs to PSHA (Godinho, 2007; Bommer & Scherbaum, 2013). It provides some ways for the effective organization and assessment of the credibility of alternative models used in this uncertainty (Godinho, 2007). Logic trees have the form of separated branches, in which there are different types of uncertainties according to the choice of each researcher (Aiping & Xiaxin, 2013). Figure 2.24: Logic tree used in PSHAs (Godinho, 2007). 2.11.3 Aleatory variability The innate randomness in a process is the definition of aleatory variability. Generally, it is included in the calculations, specifically through the parameter of standard deviation and, therefore, it plays an important role considering the resulting hazard curve (Abrahamson & Bommer, 2005; Godinho, 2007). 40 CHAPTER 3 OPENQUAKE 3.1 Introduction OpenQuake (www.openquake.org) is a software used for the calculation of seismic hazard and risk, developed by the Global Earthquake Model (GEM) (Monelli et al., 2012; Silva et al., 2012; Crowley et al., 2013). Summer 2010 was the starting date of the application of OpenQuake, which derives from several GEM‟s projects (GEM Foundation, 2010) using a wide range of data related to hazard and risk (Danciu et al., 2010; Crowley et al., 2010a; Crowley et al., 2010b; Pagani et al., 2010; Crowley et al., 2011). Specifically, OpenQuake is a combination of Python and Java programming code. Their development was achieved by applying the most usual methods of an open source software improvement (open mailing lists, public repository, IRC channel) (Crowley et al., 2011). The released source code can be found on a free and accessible web based repository (www.github.com/gem). It must be mentioned that open source projects such as Celeryd, RabbitMQ and OpenSHA played a crucial role to the development of OpenQuake (Crowley et al., 2011). Therefore, the main characteristics of OpenQuake are the following (Monelli et al., 2012): The XML (eXtensible Markup Language) data schema is a basic feature. OpenQuake uses an alternative form of XML, defined as NRML („Natural hazard‟ Risk Markup Language). The description of a variety of data structures required for seismic hazard and risk assessment is feasible through this NRML formal. It is designed for evaluating seismic hazard models for various global areas and updated according to the special requirements of each regional seismic hazard/risk programs. The figure presented below (Fig. 3.1) is a schematic illustration of OpenQuake‟s structure and contains (Crowley et al., 2011): 41 CHAPTER 3 – OPENQUAKE 1. Purple boxes, which are the representation of the crucial modules of the hazard component. 2. White boxes, with main products estimated by the distinct modules. 3. Orange rectangles, which illustrate the essential input data. Figure 3.1: Openquake‟s schematic representation (Crowley et al., 2011). 3.2 OpenQuake-Hazard The basic definition of Probabilistic Seismic Hazard Analysis (PSHA) (see §2.1) has been rapidly developed over the years, and it has been more accurate because of the reduced degree of uncertainty (Crowley et al., 2011). This resulted from the improvement of instrumental seismology and the computing power of hardware. EQRISK (McGuire, 1976) and SEISRISK (Bender & Perkins, 1982, 1987) are programming codes which played an important role concerning the evolution of PSHA. Nowadays, many implementations of PSHA are more complex due to the challenges presented continuously. The location, the geographical scale and, generally, the differences of each studied case can affect the way of application. On the one hand, PSHA for specific sites and high-risk structures (e.g. nuclear plants) demand more detailed, complex inputs and a 42 CHAPTER 3 – OPENQUAKE more extensive characterization of the parameter of uncertainty (Crowley et al., 2011). On the other hand, PSHA for urban areas does not demand such complex data and input model (Crowley et al., 2011). 3.2.1 Main concepts OpenQuake follows the procedure presented below for the computation of probabilistic seismic hazard (Crowley et al., 2011): 1. The reading of the PSHA input model (e.g. the combination of the ground motion and seismic source system) and calculation options. The required information for the creation of one or many seismic source models can be found in the seismic source system. The epistemic uncertainties must be considered in such a calculation, thus the system contains the following tools (Crowley et al., 2011): One or many Initial Seismic Source Models. One logic tree, also called „seismic source logic tree‟. It describes the epistemic uncertainties associated with features and objects that characterize the Initial Seismic Source Models. The required information for the use of one or many ground motion models can be found in the ground motion system. The epistemic uncertainties must be taken into account. 2. The processing of logic tree structures in order to account for epistemic uncertainties, which are mainly connected with the seismogenic source and ground motion. Finally, ground motion and seismic source models are created. The necessary information for the creation of an ERF Earthquake Rupture Forecast (e.g. the seismicity occurrence probability model) without taking into account any epistemic uncertainty is contained into the seismic source model. The necessary data for the hazard computation using a seismic source model is included into a ground motion model. 3. The hazard computation, taking into account as many seismic sources and ground motion models as needed for the adequate characterization of uncertainties. 4. The post-processing of the obtained results for distinct estimations and the calculation of simple mathematical statistics. 3.3 Workflows of calculation Various approaches are followed by the hazard component of OpenQuake-hazard, which computes seismic hazard analysis (SHA). There are three basic categories of analysis presented below (Crowley et al., 2011): 43 CHAPTER 3 – OPENQUAKE 1. Classical Probabilistic Seismic Hazard Analysis (cPSHA). This type calculates hazard curves and maps, considering the classical integration method (Cornell, 1968; McGuire, 1976) as mentioned by Field et al. (2003). 2. Event-Based Probabilistic Seismic Hazard Analysis (ePSHA), which calculates ground motion fields derived from stochastic event sets. 3. Deterministic Seismic Hazard Analysis (DSHA). It estimates ground motion fields from a single earthquake rupture event considering ground motion aleatory variability. For the purposes of this master dissertation, the Classical Probabilistic Seismic Hazard Analysis (cPSHA) is analyzed extensively in the next subchapter and used for the calculation of Corinth Gulf‟s hazard map. 3.3.1 Classical Probabilistic Seismic Hazard Analysis (cPSHA) Input data used for the cPSHA has a PSHA input model, which is provided with a set of calculation options. Then, the basic calculators applied for the analysis performance are (Fig. 3.1) (Crowley et al., 2011): Logic Tree Processor A seismic source model is created by the Logic Tree Processor (LTP), which takes the PSHA input model as an input data. Specifically, the seismic source model describes the activity rates and the geometry of each seismogenic source without any epistemic uncertainty. Then, a ground motion model is created by the LTP (Crowley et al., 2011). Earthquake Rupture Forecast Calculator (ERF) The ERF, which estimates the probability of occurrence over a specified time span for each earthquake rupture produced by the source model, uses the resulted seismic source model as an input (Crowley et al., 2011). cPSHA Calculator The ground motion model and the ERF are used by the cPSHA for the computation of hazard curves on each area specified in the calculation options (Crowley et al., 2011). 3.4 Description of input Two basic data blocks are discussed in this chapter, the PSHA input model and calculation settings. The accurate meaning of a PSHA input model (PSHAim) is taken from Crowley et al. (2011): “PSHAim defines the properties of the seismic sources of engineering interest 44 CHAPTER 3 – OPENQUAKE within the region considered in the analysis and the models capable to describe the properties of the shaking expected at the site”. Additionally, two main features are contained: the seismic source system and the ground motion system. Geometry, location, seismicity occurrence properties of active faults and probable epistemic uncertainties that affect this information are specified by the seismic source system. The details of ground motion forecast relationships adopted in the estimation and the associated epistemic uncertainties are described by the ground motion system (Crowley et al., 2011). Therefore, two forms of logic trees define the OpenQuake‟s PSHA input models. The seismic source logic tree, which describes the epistemic uncertainties related to the formation of the ERF, and the ground motion logic tree, which considers the uncertainties connected with the application of models able to forecast the expected ground motion at a region. When the epistemic uncertainties are inconsiderable, the logic tree structure has one branching level with only one branch (Crowley et al., 2011). 3.5 Typologies of seismic sources An amount of sources that belong to a measurable set of possible typologies is included in a usual OpenQuake input model (PSHAim). This software contains four seismic source categories; each of them has a limited number of parameters, which are indispensable for the specification of the geometry and seismicity occurrence. In the next subchapter a more extensive analysis of the source typologies supported by the OpenQuake software is provided (Crowley et al., 2011). 3.5.1 Description of seismic source typologies As mentioned above, four seismic source typologies are supported by OpenQuake (Pagani et al., 2010; Crowley et al., 2011): 1. Area source: the type with the most frequent use in regional and national PSHA models. 2. Grid source: for the reason that both area and grid sources model the distributed seismicity, this type can easily replace the area source category. 3. Simple fault source: the specification of a fault source in OpenQuake program becomes more fluent using the simple fault type, which is frequently used for the description of shallow active fault sources. It is also adopted for the purposes of the current master thesis. 45 CHAPTER 3 – OPENQUAKE 4. Complex fault source: this application is mostly related to the modeling of subduction interface sources with a complex geometry. The main hypotheses accepted in the definition of the above presented source typologies are the following (Suckale et al., 2005; Crowley et al., 2011): 1. The distribution of seismicity over the source is homogeneous (area & simple fault sources). 2. A Poissonian model is followed by seismicity temporal occurrence. 3. The frequency-magnitude distribution can be estimated to an evenly discretized distribution. 3.5.1.1 Simple fault sources The most applied source type for the modeling of faults is the “simple fault” category. The dimensions of the seismogenic source acquired by the projection of a trace or polyline along a dip direction are the meaning of the word “simple” (Crowley et al., 2011). Some interesting features of simple fault sources taken from Crowley et al. (2011) are: A fault trace in the form of a polyline. A rake angle, as specified by Aki & Richards (2002). A value of the dip angle, as specified by Aki & Richards (2002). A discrete frequency-magnitude distribution. A labeling which specifies if magnitude scaling equations are followed by the size of ruptures and a homogeneous distribution over the fault surface exists, or there is the acceptance of the assumption that the entire fault surface will always be ruptured by ruptures within a given magnitude range. 3.6 Description of logic trees Logic trees (Fig. 3.2 & 3.3) are a tool which purpose is to handle the epistemic uncertainties of models and parameters contained in a hazard analysis (Crowley et al., 2011). In our case, we used two types of logic trees. The first category contained the seismic source models with their adjusted weights. The second type of logic tree included additionally the b value uncertainty, which was adjusted in each seismic source model in order to attempt the reduction of the uncertainty parameter. 46 CHAPTER 3 – OPENQUAKE Figure 3.2: Example of branch set-epistemic uncertainties of faults dip angle (Crowley et al., 2011). Crowley et al. (2011) note three fundamental elements included in a logic tree: 1. Branching level. 2. Branch set (Fig. 3.3). 3. Branch. The distance of a given element from the start of the logic tree is expressed by the branching level. It can be said that each branching level is connected with a single type uncertainty, so the number of branching levels is proportional to its complexity (Crowley et al., 2011). An uncertainty model is described by a branch set, which contains various exclusive and exhaustive settings (Bommer & Scherbaum, 2008). Finally, a specific alternative in a set of branches is represented by a branch. Figure 3.3: Example of OpenQuake‟s logic tree structure (Crowley et al., 2011). 47 CHAPTER 3 – OPENQUAKE Figure 3.4: Logic tree data structure-individual branches, branch sets & branching levels (Crowley et al., 2011). 3.7 The PSHA Input Model (PSHAim) PSHAim includes (a) the data required for the definition of shape, position, activity rates and relative epistemic uncertainties of engineering importance seismogenic sources within a given data, and (b) the use of the ground motion models and related uncertainties for the estimation of PSHA. The seismic sources and the ground motion system are two corresponding objects contained in the PSHAim (Crowley et al., 2011). 3.7.1 The seismic sources system It consists of one or more initial seismic source models (list of seismic source data) and the seismic sources logic tree (Fig. 3.5). One or several seismogenic sources that account for distributed seismicity are usually included in a seismic source model (Crowley et al., 2011). Epistemic uncertainties related to the parameters applied for the characterization of the initial seismic source models are described by the seismic sources logic tree. During the application of this type of logic tree, the epistemic uncertainties related to all the parameters that characterize each source typology can be considered by the user (Crowley et al., 2011). 3.7.1.1 Logic tree of seismic sources This version of OpenQuake defines the seismic sources logic tree as following (Crowley et al., 2011): There is an assumption than one or more substitute initial seismic source models are described by the first branching level. Source parameter uncertainties are defined by subsequent branching levels. Each seismic source in a source model applies parameter uncertainties, which are assumed that are uncorrelated between various seismogenic sources. 48 CHAPTER 3 – OPENQUAKE Branching level can define one branch set. 3.7.1.2 Supported branch set typologies Only two built-in typologies of branch set are included in this version of OpenQuake. The next Figure 3.5 is the illustration of a source model logic tree, containing the settings available in the current version of this program (Crowley et al., 2011). Gutenberg-Richter b value uncertainties These uncertainties are depicted in Figure 3.5 as the branch set in the second branching level of the current seismic sources logic tree. An infinite amount of branches are contained in this branch set (Crowley et al., 2011). Figure 3.5: Seismic sources logic tree (Crowley et al., 2011). Gutenberg-Richter maximum magnitude uncertainties For this branch set, a value (positive or negative) can be specified by the user in order to be added to the Gutenberg-Richter maximum magnitude values (Crowley et al., 2011). 3.7.2 The system of ground motion The ground motion system is a blend of one or many logic trees, which are related with a particular tectonic area or a source group. The alternative ground motion models available for a specific source group are defined by each ground motion logic tree. Only hardcoded Ground Motion Prediction Equation (GMPE) are provided by the OpenQuake program (Fig. 3.6). An insufficiency of tools which allow the specification of new GMPEs by the user also exists (Crowley et al., 2011). 49 CHAPTER 3 – OPENQUAKE Figure 3.6: Ground Motion Prediction Equations (GMPEs) contained in OpenQuake and OpenSHA (Crowley et al., 2011). 3.7.2.1 The logic tree of ground motion The epistemic uncertainties associated to the ground motion models are represented by the ground motion logic tree (Crowley et al., 2011). The consideration of multiple GMPE logic trees, one for each tectonic area category taken into account in the source model, are supported by OpenQuake given that ground motion models are frequently associated to a specific tectonic area (Crowley et al., 2011). This version contains a GMPE logic tree permitted to have one branching level including one branch set, where a specific GMPE is linked to each individual branch. With these available options, epistemic uncertainties derived from different models can be considered, but this does not apply for the case of epistemic uncertainties inside each model (Crowley et al., 2011). 3.8 Calculation settings Calculation settings are an object that includes the data available for hazard estimation. Some relative basic elements are mentioned below (Crowley et al., 2011): The geographical coordinates of the study area, where the hazard computation is conducted and the site‟s soil condition (vs,30). The methodology followed for the hazard estimation (see §4.3). - cPSHA. - DSHA. - ePSHA. The typology of the expected results computed by the current version of OpenQuake: - Hazard maps. 50 CHAPTER 3 – OPENQUAKE - Hazard curves. 3.9 The Logic Tree Processor (LTP) In this section, the logic tree processor is presented analytically. LTP‟s purpose is the data processing in a PSHAim, which consists of a seismic source model creation derived from the seismic source logic tree (see §3.7.1.1) and ground motion model derived from the ground motion logic tree (see §3.7.2.1) (Crowley et al., 2011). 3.9.1 The logic tree Monte Carlo sampler The creation of a set of seismic source and ground motion interpretations, which represent the combinations permitted by the logic tree structure as defined by the user, is the main goal of a logic tree Monte Carlo sampler (LTMCS) (Crowley et al., 2011). The final results will reflect the uncertainty introduced by the lack of accurate parameter and model definition (Gupta, 2002; Crowley et al., 2011). 3.9.1.1 The sampling of seismic source logic tree The LTMCS creates a seismic source model processing all branching levels. In the first branching level, there is a selection of an initial seismic source model, with a probability equal to the weight of uncertainty (Crowley et al., 2011). For each branching level that follows, there is a start of a loop procedure over the seismogenic sources. Then, for each source there is a random selection of an epistemic uncertainty value (Crowley et al., 2011). 3.9.1.2 The sampling of ground motion logic tree The ground motion logic tree defines the multiple branch sets that include various ground motions models (Crowley et al., 2011). It follows a loop procedure over the various tectonic area categories, which are defined by the user. For each of them, there is a random selection of a GMPE considering their weights. A ground motion model for each tectonic area category, taken into account in the source model, will be included in the final sample set (Crowley et al., 2011). In addition, the methodology of the inverse transform method (Martinez & Martinez, 2002) is used for the sampling of epistemic weights. The method used for both the source model and ground motion logic trees, computes the inverse distribution of the epistemic weights and generate a uniform random value between 0 and 1.0 (Crowley et al., 2011).Then, an epistemic uncertainty model with a probability equal to the related weight is given (Crowley et al., 2011). 51 CHAPTER 3 – OPENQUAKE 3.10 The earthquake rupture forecast calculator The Earthquake Rupture Forecast (ERF) is a basic concept used in the OpenSHA framework (Field et al., 2003) and OpenQuake‟s hazard component (Crowley et al., 2011). The initial procedure of ERF‟s calculation includes a seismic source model, which is created by the LTP (Crowley et al., 2011). In the case of epistemic uncertainty‟s absence in the seismic source system, there is a one-to-one correspondence between the initial seismic source and the seismic source model applied in the hazard calculation (Crowley et al., 2011). Then, the LTP copies the data of seismic source model contained in the initial seismic source model and, finally, sources that produce seismicity in accordance with Poisson temporal occurrence model are supported by OpenQuake (Crowley et al., 2011). 3.10.1 ERF creation-fault sources case Two categories of fault sources are mainly supported by OpenQuake. Their differences are mostly associated to the dimensions of the fault surface. Shallow sources are modeled by fault sources with a simple geometry. On the contrary, subduction interface sources are modeled by fault sources which present a more complex geometry (Crowley et al., 2011). 3.11 Calculators of seismic hazard analysis Probabilistic seismic hazard computed by OpenQuake uses two methodologies: the classical method (cPSHA) and a method which is based on the generation of a stochastic event set (Crowley et al., 2011). The cPSHA methodology, which is used in OpenQuake, is the one mentioned by Field et al. (2003) and applied in the OpenSHA software. The specific structure mentioned above and also presented by Chiang et al. (1984), has the considerable feature of using probabilities during the calculation procedure instead of working with rates of occurrence (Bender & Perkins, 1987). The decoupling of the probability seismicity occurrence model creation is one benefit of the OpenSHA methodology (Crowley et al., 2011). The demonstration of Field et al. (2003) shows that this methodology is very stable by assuming negligible contributions to hazard derived from multiple occurrences. On the other hand, the method of stochastic event generation follows recent approaches in PSHA calculation (Musson, 2000). The basic benefits of the above mentioned approach are the following (Crowley et al., 2011): 52 CHAPTER 3 – OPENQUAKE 1. Hazard can be associated to an earthquake sequence. 2. The elements of ground motion remained on each studied area can be taken into account by considering the ground motion spatial correlation. 3.11.1 cPSHA calculator This way of calculation is the one considered as the most effective for the PSHA calculation results (hazard curves, hazard maps, hazard spectra), taking as input the elements presented below (Crowley et al., 2011): a ground motion model, an Earthquake Rupture Forecast (ERF). 3.11.1.1 Calculation of PSHA - Considering a negligible contribution from a sequence of ruptures in occurrence The PSHA calculation method which is available in OpenQuake is mainly applied in OpenSHA (Crowley et al., 2011). It is similar to the classical method considering the hypothesis of the negligibility of the hazard contribution derived from multiple ruptures (Field et al., 2003). The hazard estimation for a unique site ( ) and a single parameter of ground motion ( ) is performed through a repetitive process (Crowley et al., 2011). Then, the contributions are integrated. These are derived from the ruptures contained in the ERF and located at a distance from the site parameter, shorter than a minimum value of 200-300 km (Crowley et al., 2011). During each repetition procedure, a calculation of the probability of exceedance ( ) in time ( ) is made, by taking a rupture ( ) within source ( ). All these are described through the Equation 3.1, taken from Crowley et al. (2011): ( ) ( ) ( ) ( ) On the one hand, the product between the conditional probability of exceedance ( ) at site and the probability of occurrence in a time t, corresponds to the probability ( ). On the other hand the probability of occurrence linked to during the creation of ERF is defined by the symbol ( ). The next relationship (Equation 4.2, taken from Crowley et al., 2011) can be written in an alternative way by changing the corresponding magnitude and node within source ( ( ) ( ) ( ) to each rupture. ) ( ) 53 CHAPTER 3 – OPENQUAKE The product between the probability of on node ( occurrence exceedance and probability of magnitude ) corresponds to the probability of exceedance of in (Crowley et al., 2011). This is the interpretation of the above mentioned equation. Finally, the final hazard value located at a site ( ) will be acquired by merging the contributions derived from all of seismic sources taken into account during the process of ERF creation (Crowley et al., 2011). ( ) ∏ ( ( ) ( ) 3.11.1.2 Calculation of PSHA - Accounting for contributions from a sequence of ruptures in occurrence In some cases, the hypothesis of negligible contributions to the final hazard value derived from a sequence of ruptures is not valid. Therefore, in order to conduct more accurate hazard estimations, it is indispensable to consider any potential contribution, which is a product of ERF‟s sources (Crowley et al., 2011). Equation 3.4 is presented in order to account for repeated ruptures (Crowley et al., 2011): ( ) ∑( ) ) ( ( ) ( ) where: ( ): the definition of the probability of a least one exceedance of one or more ruptures occurring within source ( given ). Then, Equation 3.5 has the following form (Crowley et al., 2011): ( ) ∑ (∑( ( ) ) ( )) ( ) 54 CHAPTER 4 DESCRIPTION OF METHODOLOGY 4.1 Introduction Before the main part of this master thesis, i.e. the estimation of seismic hazard using the Openquake software, it is necessary to describe the methodology followed for the collection of the data related to active seismic faults of Greece. The identification of active faults is the first step of any seismic hazard assessment (Tselentis, 1997; Mohammadioun & Serva, 2001). The data which are used in this thesis are taken from two databases, the GreDaSS (Greek Database of Seismogenic Sources, www.gredass.unife.it) and fault database of the Institute of Geodynamics (National Observatory of Athens, www.gein.noa.gr) (IG-NOA) (Ganas et al., 2013). Additional information was taken from bibliographic sources, such as published papers and scientific books. The collected data refer to the active faults around the city of Patras (north Peloponnese, Greece) in a radius of approximately 200 km. It consist of some basic parameters, such as the code of each seismic fault (i.e. GR0785), the name, the minimum and maximum depth of the fault‟s surface, the strike, dip and rake, the slip rate, the maximum recorded magnitude, the location and, finally, the length and width. Pavlides et al. (2007) separated the faults into five categories, depending on their degree of activity: 1. Holocene active faults (confirmed displacement during the last 10,000 years & high values of slip rate). 2. Late Quaternary active faults (confirmed displacement during the last 40,000 years). 3. Quaternary active faults (confirmed displacement during the Quaternary & low to medium values of slip rate). 4. Capable faults of uncertain age, that can be possibly activated in the future. 5. Faults of uncertain activity, which are possibly inactive. 55 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Figure 4.1: Map of capable faults (Pavlides et al., 2007). 4.2 The Greek Database of Seismogenic Sources (GreDaSS) 4.2.1 Introduction In this chapter the database of GreDaSS is presented (Fig. 4.2). According to Sboras et al. (2014), the construction of this database is based on geological information and investigation techniques. As stated by Sboras et al. (2014), GreDaSS is a project which goals are: 1. The systematic collection of all available information related to neotectonics, active, capable faults and generally the seismogenic volumes. 2. The critical analysis of the collected data and the quantification of the basic seismogenic features of several sources including a related degree of uncertainty. 3. Provide a complete view of probable damaging active faults for a better effectiveness of SHA in Greece. 56 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Figure 4.2: The form of GreDaSS showing the ISSs & CSSs layers (www.gredass.unife.it). 4.2.2 Types of seismogenic sources Some fundamental elements of GreDaSS database are also presented, as inferred by Basili et al. (2008); Sboras et al. (2009). There are two basic categories of seismogenic sources, the “Individual Seismogenic Sources” (ISS) and the “Composite Seismogenic Sources” (CSS) (Fig. 4.2 & 4.3). “Individual Seismogenic Sources” (ISS) are derived from the synthesis of geological and geophysical data. These types of seismogenic sources include a complete set of geometric characteristics, such as strike, dip, length, width and depth, kinematic parameters (rake) and seismological-palaeoseismological features, such as the average displacement per event, the magnitude, the slip rate and the return period. Their use is referred to the deterministic seismic hazard assessment (DSHA), the calculation of earthquake scenarios and geodynamic research. “Composite Seismogenic Sources” (CSS) have the same initial features concerning the geometric and kinematic parameters, but the difference is about the looser definition and the combination of two or more individual sources. This type of seismogenic sources is not necessarily capable of a specific earthquake, but their possible activity can be detected from the existing data. Instead of the previous category, the CSSs have 57 CHAPTER 4 – DESCRIPTION OF METHODOLOGY a complete record of potential earthquake sources and accurate description. In conclusion, the CSS can be used for the preparation of regional PSHA and the study of geodynamic procedures. Figure 4.3: Schematic representation of an ISS (a) & CSS (b) seismogenic source (Sboras, 2011). The description is presented below, according to Basili et al. (2009). The depiction of the ISSs is associated with a rectangular (polygon) and a vector placed at the central part (Fig. 4.3a). The purpose of the rectangular is the representation of the vertical projection of fault plane on the ground surface. The top edge is associated with the fault trace, in the case that the fault is characterized as emergent. When the fault is blind, the section between the hypothetical continuation of ground surface and fault plane appears as a line parallel to the top edge. Finally, the slip vector of fault‟s motion is represented by a vector located in the rectangular. The CSSs (Fig. 4.3b) have a looser shape because of their capability of containing several fault segments (ISSs) and their incomplete data. The polygon includes two roughly parallel long sides (such as the ISSs), which correspond to the top and bottom edges of fault plane and two short lines parallel to the width. The top edge is represented with a thicker line and in the case of the fault reaches the surface, the scarps or fault traces are followed by the top edge. 4.2.3 Properties of seismogenic sources Further and useful information about a seismogenic source can be obtained by clicking on it. Then, a new web browser window will open, containing the information needed, separated 58 CHAPTER 4 – DESCRIPTION OF METHODOLOGY into four categories. This form is similar for both CSSs and ISSs. The information window contains the following metadata pages: i. “Source Info Summary”: the basic parameters are contained in this metadata page. These are the “General Information” (Code, Name, Compilers, Contributors, Latest update date), the “Parametric Information” (kinematic, geometric, seismotectonic information) and finally the “Associated Earthquake”, which is referred only to the ISSs (latest events, associated with a specific source). ii. “Commentary”: three sections are included: the “Comments” (contains helpful comments for a better description of the source, more details about the extra data, etc.), the “Open Questions” (contains whichever parameter remains doubtful) and the “Summary” (includes the information related to the source, which can be extracted from the available bibliography). iii. “Pictures”: pictures, figures, maps and images are enclosed in this metadata page. iv. “References”: this is the last page, which contains all the literature linked with the hosted source. Figure 4.4: Source Info Summary, example of Palaeochori ISS fault (Sboras, 2011). 59 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Figure 4.5: Commentary, example of Palaeochori ISS fault (Sboras, 2011). Figure 4.6: Pictures, example of Palaeochori ISS fault (Sboras, 2011). Figure 4.7: References, example of Palaeochori ISS fault (Sboras, 2011). 60 CHAPTER 4 – DESCRIPTION OF METHODOLOGY 4.2.4 Parameters of seismogenic sources For a better understanding of GreDaSS‟s environment, a definition and a qualitative description of the parametric fields is made starting with the ISSs and CSSs. We take into consideration the necessary precision and completeness. After Sboras (2011). 4.2.4.1 Individual Seismogenic Sources (ISSs) Location (degrees): this parameter indicates the location of the fault on the map. Length (km): the length of the fault plane. Minimum depth (km): this parameter is associated with the vertical distance (depth) of fault‟s top edge from the ground surface or the sea floor. Maximum depth (km): the calculated depth of the bottom edge of fault plane from the surface. Width (km): the distance between the top and bottom edges of fault plane. Strike (degrees): it has a similar meaning to fault‟s strike. Values which belong to the eastern semicircle (0-180o) have a dip direction (plunge) inside the southern semicircle (90-270o). On the contrary, values which belong to the western semicircle (181-360o) have a dip direction inside the northern semicircle (271-90o). Dip (degrees): the dip-angle of fault plane. Rake (degrees): the measured counter-clockwise angle, formed between the slip vector and the strike. The rake‟s range is between 0o and 360o. Slip per event (m): the mean co-seismic displacement on the fault plane is represented by this parameter. It is usually suggested by the database software based on empirical laws, or it can be set directly. Slip rate (mm/a): the ratio between the displacement and the necessary time to produce it. Recurrence (years): the recurrence interval time between two characteristic seismic events. Magnitude (Mw): this is a representation of the magnitude produced by a specific seismic event, or the possible magnitude of the fault which is based on scaling laws. In the realm of magnitude, there is dependence between the fault‟s dimensions and slip per event. Last earthquake (years): the date or the time elapsed from the last seismic event is included in this part. 61 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Penultimate earthquake (years): generally, it is a rarely available information derived from paleoseismological studies and, sometimes, from historical references. Elapsed time (years): the time interval between the last known seismic event and the year 2000, which is used as a reference. 4.2.4.2 Composite Seismogenic Sources (CSSs) Minimum depth (km): description similar to the ISSs. Maximum depth (km): description similar to the ISSs. Strike (degrees): it has the same meaning with the ISSs. The only difference has to do with the requirement of a range of values. Dip (degrees): description similar to the ISSs. Rake (degrees): the definition is the same with the ISSs, but in this case a range of values is required. Slip rate (mm/a): description similar to the ISSs. Maximum magnitude (Mw): it is the representation of the potential seismic magnitude, or the maximum expected magnitude produced by the CSS. Approximate location (degrees): same definitions with the ISSs. It is the center of the source. Total length (km): similar to the ISSs. Total width (km): similar to the ISSs. Typical fault length (km): it is based on the maximum magnitude field and calculated from several scaling laws. Typical fault width (km): this parameter is based on the maximum magnitude field, the typical length and the dip angle range. It is derived from calculations between analytical and scaling laws. Typical fault slip (m): it has similar meaning to the former two parameters. Typical fault slip is defined as the average displacement per event, based on scaling laws. 4.3 Application of GIS The G.I.S. (Geographic Information Systems) software is used in order to create a complete database for the case study (investigation of the active faults, 200 km away of Patras, north Peloponnese, Greece). These data files considering the active faults are taken from the GreDass‟ and IG-NOA‟s database. 62 CHAPTER 4 – DESCRIPTION OF METHODOLOGY The database related to active faults is classified into three layers, according to the completeness of the data. The first layer (Fig. 4.8) includes the active faults with complete data, the second layer (Fig. 4.9) includes the faults with an intermediate level of data completeness (slip rate, length, max magnitude) and the last category (Fig. 4.10) includes the seismogenic sources with poor data completeness (only length). Additionally, after the merging of all shape files, the total faults database is illustrated by Figure 4.11. The attribute table (Fig. 4.12) of each layer presents the values of the parameters of active faults. Figure 4.8: 1st layer – complete level of data. Figure 4.9: 2nd layer – intermediate level of data 63 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Figure 4.10: 3rd layer – poor level of data. Figure 4.11: The total faults database. 64 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Figure 4.12: The attributes table. 4.4 Earthquake scaling laws In this subchapter the scaling laws used in this dissertation are presented. Generally, these equations can define various parameters, such as displacement, magnitude, rupture length, seismic moment, etc. (Billion, 2007). In addition, the validity of models on the mechanics of seismic rupture can be tested through these empirical relationships (Papazachos et al., 2004). 4.4.1 Wells & Coppersmith (1994) Displacement vs magnitude and Dmax (maximum displacement) vs length relationships, taken from Wells & Coppersmith (1994), were applied in this thesis for the estimation of slip rate. These scaling laws were compiled using a database of approximately 400 seismic events. Shallow focus, continental, intraplate or interplate earthquakes of magnitude greater than 4.5 are included in this data. On the contrary, there is an exclusion of seismic events that take place in subduction zones and oceanic labs (Wells & Coppersmith, 1994). In addition, the rupture width vs magnitude relationship is used for the case that the parameter of width is not available in the database. 4.4.1.1 Displacement per event (MD) Vs. Magnitude (M) Figure 4.13: Displacement per event Vs. Magnitude relationship for each type of faults‟ kinematics (Wells & Coppersmith, 1994). 65 CHAPTER 4 – DESCRIPTION OF METHODOLOGY 4.4.1.2 Maximum displacement (MD) Vs. Rupture length (SRL) Figure 4.14: Displacement per event Vs. Rupture length relationship for each type of faults‟ kinematics (Wells & Coppersmith, 1994). 4.4.1.3 Rupture width (RW) Vs. Magnitude (M) Figure 4.15: Rupture width Vs. Magnitude relationship for each type of faults‟ kinematics (Wells & Coppersmith, 1994). 4.4.2 Pavlides & Caputo (2004) Magnitude (Ms) vs maximum vertical displacement (MVD) and surface rupture length empirical equations are proposed for the Aegean region by Pavlides & Caputo (2004). The equation used for the purposes of this master thesis is the following magnitude versus maximum vertical displacement relationship: ( ( ) ( ) ⇒ ) ( ) 4.5 Estimation of the slip rate – Approaches The active faults‟ slip rate is one of the most crucial features of seismic hazard analysis. Except from the literature data, a new database is made in order to present a more comprehensive distribution for the decrease of parameter uncertainty. The equation that defines the annual slip rate is associated with the total displacement and the age of each fault (L. Danciu personal communication): ( ) 66 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Different approaches have been made for the estimation of the parameter of total displacement. The stratigraphic age of faults was taken from Kokkalas et al. (2006); Marnelis et al. (2007); Papanikolaou et al. (2007); Pechlivanidis (2012). 4.5.1 Approach 1 – Historical seismicity Historical seismicity method is based on data related to historical earthquakes, for which earthquake magnitude can be estimated. The magnitudes of historical earthquakes can be correlated with empirical relationships. Therefore, many conclusions can be extracted with respect to active faults which cause large earthquakes. However, it is widely known that in some regions the historical data are usually incomplete, so the accuracy of this method is sometimes limited (Koukouvelas et al., 2010). This approach is based on the assumption that the total displacement of a fault derives from the sum of the displacements caused by seismic events that occurred near it. This contains a degree of uncertainty, because in some areas the correlation between past seismic activity and known fault structures is impossible (Cornell, 1968). Concerning the seismic events, the Seismicity Catalog (550 B.C-2010 A.D) (Papazachos et al., 2000; Papazachos et al., 2010) for magnitude greater than 4.5R is used in order to possess a complete data. Figure 4.16: Historical seismicity of Greece - application in GIS. In the next step, a buffer zone of 5km around each fault is made in order to link seismic events most probably related to the fault with fault‟s displacement. The following illustration (Fig. 4.17) springs from the GIS software. All the events that fall within the buffer zone are consider to belong to different ruptures of the fault, their magnitude is used as an 67 CHAPTER 4 – DESCRIPTION OF METHODOLOGY input to scaling laws and a displacement value is calculated. The sum of the calculated displacements is used together with the age of the fault and eq. 4.2 for the computation of slip rate. Figure 4.17: Buffer zone of the Argostoli fault, Kefallonia - application in GIS. 4.5.2 Approach 2 – Length of faults There are many studies and reports relative to the relationship between maximum displacement and fault length, for the comprehension of fault geometry over many length scales (Kim & Sanderson, 2005). In this thesis, the following relationship is used (Fig. 4.13) (Wells & Coppersmith, 1994): ( ) ( ) ( ) where: : the maximum displacement (km), : surface rupture length (km), coefficients. Specifically, the maximum displacement is estimated by knowing the total length of each fault. This can be applied with the premise that the entire length of the fault ruptures during the occurrence of a seismic event, although observations have shown the opposite (Petersen et al., 2011). This can lead to the overestimation of seismic hazard. Then, by knowing the total displacement and the stratigraphic age of a fault, the slip rate can be estimated through the basic equation (4.2). 68 CHAPTER 4 – DESCRIPTION OF METHODOLOGY 4.6 Estimation of minimum & maximum fault depth For active faults that the values of minimum and maximum seismogenic depth are not available (i.e. IG-NOA database), some assumptions are made. Then, the maximum depth value is taken from the depth of Mohorovic (Moho) discontinuity (the boundary between Earth‟s crust and upper mantle) for each fault. The range of Moho depth in Greece is presented in the following map: Figure 4.18: Map of Moho depths in the Greek territory (Tsokas & Hansen, 1997; modified from Somieski, 2008). 4.7 Fault characterization 4.7.1 Slip rate evaluation Slip rate is the most crucial parameter of the present investigation. Uncertainties exist, so a slip rate distribution was made and each fault contained 9 slip rate estimates. As defined by Eq. 4.2, it derives from the ratio between total displacement and stratigraphic age of fault. For the estimation of total displacement, two approaches were considered, historical seismicity (see §4.5.1) and length of faults (see §4.5.2). Two empirical relations were used for this 69 CHAPTER 4 – DESCRIPTION OF METHODOLOGY occasion: displacement per event vs magnitude of Wells & Coppersmith (1994) and of Pavlides & Caputo (2004). The second approach used the surface rupture length vs displacement empirical relation of Wells & Coppersmith (1994) with the premise that the entire fault length ruptures during an earthquake (worst case scenario). Thus, for each fault three values of total displacement were estimated. The stratigraphic age of faults was derived from the database or from the available literature. For the reason that uncertainties exist, an average, an upper, and a lower value were assumed. Each of three values of total displacement was divided with three estimates of stratigraphic age. Therefore, for each fault 9 slip rate values were resulted. 4.7.2 Conversion of slip rates into seismic activity According to the methodology of Bungum (2007), fault seismicity derived from slip rates can be estimated using programs. The initial step of this methodology is the application of the following two relationships: the cumulative occurrence relationship of Chinnery & North (1975) (Eq. 4.4) and the total moment release rate equation of Brune (1968) (Eq. 4.5). ( ( ) ) ( ) ( ) where: N: the number of earthquakes equal to or above magnitude M, a: the absolute level of activity, b: the ratio between smaller and larger earthquakes, M, Mmax: earthquake magnitudes H: the Heaviside step function. ( ) where: : the total moment release rate, μ: the rigidity, S: the slip rate, A: the rupture area. Combining the above Eqs. (4.4)-(4.5), the relationship of Anderson & Luco (1983) that determines the number of events N for magnitudes 4-5 R is presented below: ( ) ( ̅ ̅ ̅ )( ) ̅( ) (( ̅ ) ) ( ) 70 CHAPTER 4 – DESCRIPTION OF METHODOLOGY where: ̅ ̅ ( ( )), ( ( )), ( )) ( √ ), , ( ): seismic moment for Ms=0, d: the magnitude scaling coefficient. 4.7.3 Magnitude-Frequency Distribution (MFD) Defined by Crowley et al. (2013), MFD consists of a function describing the rate of earthquakes per year, across all magnitudes (see §2.5). The double truncated GutenbergRichter distribution is frequently used in PSHAs (Crowley et al., 2013). Figure 4.19: The double truncated Gutenberg-Richter MFD (Crowley et al., 2013). As described in the previous paragraph, 9 slip rate estimates were calculated for each fault following two approaches. For each slip rate, the cumulative a value was estimated for magnitudes from 0-6.5, as can be seen from the following Figure 4.20: 71 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Figure 4.20: Cumulative a value vs vMagnitude chart. The consideration of two approaches provided a wide range of cumulative a-values forming a zone. Therefore, all distributions were taken into account in the evaluation considering uncertainties. 4.8 Model implementation The application of logic trees is an appropriate method of modeling uncertainty. Logic tree approach allows alternative models assigning in each of them a weighting factor that represents the probability of that model being correct (Kramer, 1996). In this thesis, two logic tree approaches are made. The initial run of the code was done using the nine source models and equal weights to each one of them (Fig. 4.21). The second run of the code included the b value uncertainty in the logic tree. The following logic tree of Figure 4.22 contains nine source models and additionally the b value uncertainty presented in three values (0.9, 1.0, 1.1) with equal weights to each of them. 72 CHAPTER 4 – DESCRIPTION OF METHODOLOGY Figure 4.21: Logic Tree without b value uncertainty. Figure 4.22: Logic Tree including b value uncertainty for each source file. 73 CHAPTER 4 – DESCRIPTION OF METHODOLOGY 4.9 Configuration After the generation of XML files (source model, logic trees, GMPE logic tree), the compilation of the configuration file (.ini) follows. The configuration file controls the input model definition and the parameters used in the calculation. First of all, the structure and basic parameters for seismic hazard are described. The next steps contain the specification of the area (i.e. polygon, distance, grid points, etc.) where hazard will be computed, the logic tree processing, the specification of the discretization level of the mesh that represents faults and the definition of local soil conditions (Crowley et al., 2011). Nine XML files are the seismic sources model of this implementation. The XML file of logic tree models the epistemic uncertainty related to seismic sources model and b-value (see ). The GMPE logic tree (Fig. 4.23) is an XML file that includes the following approaches considered for active shallow crust: Akkar & Bommer (2010), Cauzzi & Faccioli (2008), Chiou & Young (2008), Zhao et al. (2006). Figure 4.23: GMPE Logic Tree. 74 CHAPTER 5 RESULTS 5.1 Model A: mean b-value (no-uncertainty) 5.1.1 Hazard maps of Corinth Gulf After the first OpenQuake execution, the hazard maps were generated for PGA values, for Spectral Acceleration (SA) of 0.1 sec (referred to a three or four-storey building) and SA of 1.0 sec (referred to a multi-storey structure). Considering the probabilities of exceedance, the values that are used in this survey are the mean values, for 10% probability of exceedance (POE) in 50 years (return period of 475 years) and 2% POE in 50 years (return period of 2500 years). PGA – Return period of 475 years Figure 5.1: Hazard map of Corinth Gulf, 10% POE in 50 years. PGA – Return period of 2500 years Figure 5.2: Hazard map of Corinth Gulf, 2% POE in 50 years. 75 CHAPTER 5 – RESULTS SA 0.1 sec – Return period of 475 years Figure 5.3: Hazard map of Corinth Gulf, 10% POE in 50 years. SA 0.1 sec – Return period of 2500 years Figure 5.4: Hazard map of Corinth Gulf, 2% POE in 50 years. SA 1.0 sec – Return period of 475 years Figure 5.5: Hazard map of Corinth Gulf, 10% POE in 50 years. 76 CHAPTER 5 – RESULTS SA 1.0 sec – Return period of 2500 years Figure 5.6: Hazard map of Corinth Gulf, 2% POE in 50 years. 5.1.2 Hazard curves of Patras Additionally to hazard map calculation, hazard curves for Patras, Aigion and Korinthos were calculated for 10% probability of exceedance in 50 years and. According to Krinitzsky et al. (1990), different percentiles reflect the range of uncertainty given by the expert in various seismic-source characteristics. Figure 5.7: Hazard curves of Patras for PGA, 10% POE in 50 years. 77 CHAPTER 5 – RESULTS Figure 5.8: Hazard curves of Patras for SA 0.1 sec, 10% POE in 50 years. Figure 5.9: Hazard curves of Patras for SA 1.0 sec, 10% POE in 50 years. 5.1.3 Hazard curves of Aigion Figure 5.10: Hazard curves of Aigion for PGA, 10% POE in 50 years. 78 CHAPTER 5 – RESULTS Figure 5.11: Hazard curves of Aigion for SA 0.1 sec, 10% POE in 50 years. Figure 5.12: Hazard curves of Aigion for SA 1.0 sec, 10% POE in 50 years. 5.1.4 Hazard curves of Korinthos Figure 5.13: Hazard curves of Korinthos for PGA, 10% POE in 50 years. 79 CHAPTER 5 – RESULTS Figure 5.14: Hazard curves of Korinthos for SA 0.1 sec, 10% POE in 50 years. Figure 5.15: Hazard curves of Korinthos for SA 1.0 sec, 10% POE in 50 years. 5.1.5 Uniform hazard spectra Finally, uniform hazard spectra were calculated for 10% POE and for the same cities. Patras Figure 5.16: Uniform hazard spectra for Patras, 10% POE in 50 years. 80 CHAPTER 5 – RESULTS Aigion Figure 5.17: Uniform hazard spectra for Aigion, 10% POE in 50 years. Korinthos Figure 5.18: Uniform hazard spectra for Korinthos, 10% POE in 50 years. 81 CHAPTER 5 – RESULTS 5.2 Model B: including b-value uncertainty The second OpenQuake execution contains hazard maps generated for PGA values, for Spectral Acceleration (SA) of 0.1 sec (referred to a three or four-storey building) and SA of 1.0 sec (referred to a multi-storey structure). Considering the probabilities of exceedance, the values that are used in this survey are the mean values, 10% POE in 50 years (return period of 475 years) and 2% POE in 50 years (return period of 2500 years). 5.2.1 Hazard maps of Corinth Gulf PGA – Return period of 475 years Figure 5.19: Hazard map of Corinth Gulf, 10% POE in 50 years. PGA – Return period of 2500 years Figure 5.20: Hazard map of Corinth Gulf, 2% POE in 50 years. 82 CHAPTER 5 – RESULTS SA 0.1 sec – Return period of 475 years Figure 5.21: Hazard map of Corinth Gulf, 10% POE in 50 years. SA 0.1 sec – Return period of 2500 years Figure 5.22: Hazard map of Corinth Gulf, 2% POE in 50 years. SA 1.0 sec – Return period of 475 years Figure 5.23: Hazard map of Corinth Gulf, 10% POE in 50 years. 83 CHAPTER 5 – RESULTS SA 1.0 sec – Return period of 2500 years Figure 5.24: Hazard map of Corinth Gulf, 2% POE in 50 years. 5.2.2 Hazard curves of Patras Hazard curves for Patras, Aigion and Korinthos were calculated for 10% probability of exceedance in 50 years and different percentiles. Figure 5.25: Hazard curves of Patras for PGA, 10% POE in 50 years. Figure 5.26: Hazard curves of Patras for SA 0.1 sec, 10% POE in 50 years. 84 CHAPTER 5 – RESULTS Figure 5.27: Hazard curves of Patras for SA 1.0 sec, 10% POE in 50 years. 5.2.3 Hazard curves of Aigion Figure 5.28: Hazard curves of Aigion for PGA, 10% POE in 50 years. Figure 5.29: Hazard curves of Aigion for SA 0.1 sec, 10% POE in 50 years. 85 CHAPTER 5 – RESULTS Figure 5.30: Hazard curves of Aigion for SA 1.0 sec, 10% POE in 50 years. 5.2.4 Hazard curves of Korinthos Figure 5.31: Hazard curves of Korinthos for PGA, 10% POE in 50 years. Figure 5.32: Hazard curves of Korinthos for SA 0.1 sec, 10% POE in 50 years. 86 CHAPTER 5 – RESULTS Figure 5.33: Hazard curves of Korinthos for SA 1.0sec, 10% POE in 50 years. 5.2.5 Uniform Hazard Spectra Uniform hazard spectra were calculated for the same towns. These results are comparable to elastic design spectra of the Greek Seismic Code. Patras Figure 5.34: Uniform hazard spectra for Patras, 10% POE in 50 years. Aigion Figure 5.35: Uniform hazard spectra for Aigion, 10% POE in 50 years. 87 CHAPTER 5 – RESULTS Korinthos Figure 5.36: Uniform hazard spectra for Korinthos, 10% POE in 50 years. 5.3 Comparison In this chapter we examine the differences between the hazard calculations performed in this thesis and between the published results for the region of Corinth Gulf, Greece. The first hazard calculation procedure in this thesis didn‟t include the b value uncertainty (Run #1), in contrast with the second hazard calculation (Run #2) during which the b-value was varied by 0.1. 5.3.1 Difference between 10% probability of exceedance for mean PGA values between Run #1 and Run #2 Subtracting the acceleration values of both hazard calculations, the difference between Run #1 and Run #2 does not exceed the range ±0.1g. As it can be seen from the following maps, the b value uncertainty increased slightly the resulting hazard values. Figure 5.37: Difference map between Run #1 and Run #2 (Run#1 – Run#2) for mean PGA values, return period of 475 years. 88 CHAPTER 5 – RESULTS 5.3.2 Difference between 2% probability of exceedance for mean PGA values between Run #1 and Run #2 Figure 5.38: Difference map between Run #1 and Run #2 (Run#1 – Run#2) for mean PGA values, return period of 2500 years. Finally, the conclusion that is made shows that the variability of b-value is not significant when the return period is increased. The same applies to the other hazard calculations. 5.4 Comparisons with the Greek Seismic Code The first comparison is made between the New Hazard Map of Greece and the hazard results of Run #2, which is considered as the “worst case scenario”. Figure 5.39: Hazard maps for the comparison of PGA values. 10% probability of exceedance for the next 50 years (return period of 475 years). 89 CHAPTER 5 – RESULTS Although a direct comparison of the results is not easy (the hazard map of Greece is the result of a zonation thus cannot be compared with discrete values. Anyway, it can be observed that our results gave higher values than the corresponding 0.24g of Greek Seismic Code. Our highest values range between 0.5g and 0.6g. The Greek Seismic Code is considering the seismicity while here we considered individual faults. The uniform hazard spectra (UHS) of our results (Run #2) and of the Greek Seismic Code for Soil Class A (bedrock) were also compared. The uniform hazard spectrum of the Greek Seismic Code was adjusted to our study area, so the acceleration has the value 0.24g because it belongs to Seismic Hazard Zone II and the parameter γ (Importance Factor) has the value 1.00 because the research is referred to ordinary residential and office buildings, industrial buildings, hotels, etc. Figure 5.40: Comparison between Patras UHS & Greek Seismic Code. 10% probability of exceedance for the next 50 years. Figure 5.41: Comparison between Aigion UHS & Greek Seismic Code. 10% probability of exceedance for the next 50 years. 90 CHAPTER 5 – RESULTS Figure 5.42: Comparison between Korinthos UHS & Greek Seismic Code. 10% probability of exceedance for the next 50 years. The comparison between the second hazard calculation and the Greek Seismic Code leads us to the same conclusions. Mean UHS of Korinthos is below the standards of Greek Seismic Code and mean UHS for the cities of Patras and Aigion are upper the regulations. Thus, it is proposed that the Greek Seismic Code requires a new approach and methodology in order to be more precise principally for regions that present high levels of seismicity, such as the Corinth Gulf. 5.5 Comparisons with previous studies The aim of this subchapter is to compare our estimates with some previous studies. The results are checked in order to validate the approach that we have made. SHARE Figure 5.43: Hazard maps for the comparison of PGA values. 10% probability of exceedance (return period of 475 years). 91 CHAPTER 5 – RESULTS This project is a strong argument considering the checking of our results, because it consists of a combination of area sources model, seismotectonic characteristics, historical seismicity, fault sources and strain deformation rates. Comparing the values of PGA distribution for our survey and for SHARE project, it can be noticed that our estimates (0.35g0.55g approximately) agree with the corresponding approach of SHARE (0.40g-0.45g approximately) for a return period of 475 years. Tselentis & Danciu (2010) Figure 5.44: Hazard maps for the comparison of PGA values. 10% probability of exceedance (return period of 475 years). Tselentis & Danciu (2010) examined seismic hazard maps of Greece and of the surrounding regions including some significant engineering parameters (PGA, PGV, Arias intensity, cumulative absolute velocity). Considering the mean PGA estimates (0.4g-0.6g approximately) for Corinth Gulf of the above presented seismic hazard map, there is a very good correlation with our results (0.35g-0.55g approximately) for the same region. Segkou (2010) The seismic hazard estimation for the Greek territory was carried out following some various approaches relative to seismological, geological and geophysical observations. The linear source model, the random seismicity model of shallow earthquakes and a seismic source model of intermediate depth was applied for this implementation. 92 CHAPTER 5 – RESULTS Figure 5.45: Hazard maps for the comparison of PGA values. 10% probability of exceedance (return period of 475 years). For the reason that Segkou (2010) took into account both an approach that includes source models and historical seismicity, it can be observed that our results distribution (0.35g0.55g approximately) agrees in significant degree with these depicted in Figure 5.16 (0.30g0.45g approximately). Vamvakaris (2010) The estimation of maximum PGA values was made using attenuation relationships adjusted in each type of hypocentric depth (low, intermediate, high). Figure 5.46: Hazard maps for the comparison of PGA values. 10% probability of exceedance (return period of 475 years). 93 CHAPTER 5 – RESULTS Comparing the hazard maps for a return period of 475 years it can be implied that the correlation of PGA distributions are quite good. The estimates of Vamvakaris (2010) range between 0.30g and 0.50g, while our results range between 0.35g and 0.55g. 94 CHAPTER 6 SUMMARY and CONCLUSIONS 6.1 Summary The dissertation examined the seismic hazard assessment for a seismic prone region, Corinth Gulf (north Peloponnese, Greece), considering the active faults that surround this area. Two fault databases were used, GreDaSS‟s and Institute of Geodynamics‟. Three source categories were defined, according to the level of data completeness. The first category included faults with adequate level of data (i.e. slip rate, dip, rake, etc.), the corresponding second category included intermediate amount of information (i.e. maximum magnitude, length) and the third category contained faults with poor level of data (i.e. length). The unknown values of critical parameters (i.e. displacement, maximum magnitude, length,) in the attributes table were estimated by the application of empirical laws. Nine different slip values per fault were calculated. A distribution of slip rates was made dividing the total displacement with the stratigraphic age of each fault after the assumption of two approaches, historical seismicity and fault length. Slip values were converted to seismic activity compiling some Matlab scripts (see Appendix). The hazard calculation of OpenQuake Engine was divided in two parts. The first part included the logic tree that contained the seismic sources model without the b value uncertainty. On the contrary, the second part considered the b value uncertainty in the calculation. Thus, a comparison of them was made. We used OpenQuake in order to compute hazard maps-hazard curves and uniform hazard spectra for PGA, SA (0.1sec & 1.0sec) and uniform hazard spectra for bedrock soil conditions. All of them are referred to return periods of 475 & 2500 years and compared with previous research. The GMPE‟s used in this study were the Akkar & Bommer (2010), Cauzzi & Faccioli (2008), Chiou & Young (2008) and Zhao et al. (2006) considered for the active shallow crust of Greece. 95 CHAPTER 6 – SUMMARY and CONCLUSIONS 6.2 Results The scope of this thesis was the estimation of seismic hazard of Corinth Gulf considering active faults for bedrock soil conditions. The OpenQuake engine, developed by GEM, was used for this purpose. It is a software that uses an innovative methodology for hazard calculation. The execution is performed using a configuration file and XML files (seismic sources model, logic tree, GMPE model). The epistemic uncertainty (i.e. slip rate, b-value) can also be modeled. The comparison of two hazard calculations drew the conclusion that b-value uncertainty did not reflect our estimates. The differences between Run #1 and Run #2 were smoothed when the return period was increased. The fault database needs more enhancement because there was a lack of information considering the slip rate estimates. Previous implementations considering seismic hazard assessment for PGA and return period of 475 years were compared with our study and showed that our results are correlated significantly with their corresponding estimates. The Greek Seismic Code needs a better and more detailed approach in order to be more precise, especially for seismic prone areas. The comparison of our uniform hazard spectra with the corresponding of Greek Seismic Code for the cities of Patras, Aigion and Korinthos showed that the hazard suggested by the Greek Seismic Code could be underestimated. 96 APPENDIX PROGRAMMING Openquake operates by using the NRML format, which is an alternative version of XML data schema. In order to create these files, some Matlab scripts and functions were compiled for the purposes of XML-file construction (source model, logic tree) and the conversion of slip rates to seismic activity, a necessary parameter for SHA. All of them are extensively presented in the Appendix. The aim of basic Matlab script is to introduce some fundamental parameters deduced from the ArcGIS Shape Files (.shp), give specific values to significant parameters (slip rate, aspect ratio) and create the appropriate XML files needed for the structure of the basic source model of Openquake. The purpose of this action is to produce several XML files that contain all faults for nine different slip rates. These nine slip rate values were derived from the application of : Displacement vs magnitude relationships of Wells & Coppersmith (1994) and Pavlides & Caputo (2004) in the historical seismicity approach. Displacement vs length scaling law of Wells & Coppersmith (1994) in the length of fault approach. Three estimates of fault age (minimum, medium and maximum stratigraphic age) Thus nine values of slip rate were calculated per fault i.e. three scaling laws and three fault ages. In addition, this script uses some features from the attributes table of active faults‟ shape files, such as the name, coordinates, dip and rake, which are parameters included in the XML files. Then, the seismic activity rate is estimated by using the methodology proposed by Bungum (2007). 97 APPENDIX – PROGRAMMING THE BASIC MATLAB SCRIPT *Original code was provided by Dr. Laurentiu Danciu function ok = write_simple_fault(filename_shp) %% Load fault source rShape=shaperead(filename_shp); names = fieldnames(rShape) for j=1:9 % loop over slip rates for i=1:length(rShape) % loop over faults %Fields of the attributes table of each .shp file code{i}=rShape(i).CODE; name{i}=rShape(i).NAME; longitude{i}=rShape(i).X; latitude{i}=rShape(i).Y; dip(i)=rShape(i).DIPP; upper_depth(i)=rShape(i).MINDEPTH; lower_depth(i)=rShape(i).MAXDEPTH; rake(i)=rShape(i).RAKE; maxmag(i)=rShape(i).MAXMAG; fault_length=rShape(i).LENGTH; % 20)) % % % % %aspect ratio_GreeDass_database if ((rake(i)==170) || (rake(i)==173) || (rake(i)==aspect_ratio(i)=4; else aspect_ratio(i)=1; end %aspect ratio_Geodynamic_Institute_database if (rake(i)==180) aspect_ratio(i)=4; else aspect_ratio(i)=1; end % prepare the name of the slip rate definition (1-9) slip_R = ['rShape(i,1).SLIPRT' num2str(j)]; slipRate = eval(slip_R)*0.1; % convert to cm/year 98 APPENDIX – PROGRAMMING % case slip = 0 if slipRate==0 slipRate =0.0001; end % get a_value a_cum_value_max(i)=calc_fsz_activity2(fault_length,upper_depth (i),lower_depth(i),dip(i),maxmag(i),slipRate); end % get values for XML file filenamexml=['SLIP_RATE_' num2str(j) '.xml']; ok = writefxml(code,name,longitude,latitude,dip,upper_depth,lower_d epth,aspect_ratio,maxmag,a_cum_value_max,rake,filenamexml) end end MATLAB FUNCTION FOR THE COMPUTATION OF CUMULATIVE A VALUE *Code provided by Dr. Laurentiu Danciu function [ a_cum_value_max ] = calc_fsz_activity2( length,upper_depth,lower_depth,dip,maxmag,slipRate) % % % % % % % % % % % % % % % % Script to compare total moment of seismicity with moment determined for fault parameters Incoming: fBvalue : b-value fS : Slip rate (mm/year) fD : Average slip (m) fLength : Fault length (m) fWidth : Fault width (m) fM00 : M0(0) for Ms = 0, c in logM0=c-dM fMmin : Minimum magnitude fBinM : magnitude binnning fMmax : Maximum magnitude fDmoment : d in logM0=c-dM fRigid : Rgidity in Pascal Model 2: Anderson and Luco Units are in CGS %% Calculate moment from faults 99 APPENDIX – PROGRAMMING % Parameters for Seismic Moment from Faults % b-value bVal = 1.00; % Rgidity in GPascals --> miu =30GPa % convert to dyne/cm2 (CGS units) --> or (N/m^2) (SI units) % convert shear modulus from Pa (N/m^2, kg/(m * s^2)) % to dyn/cm^2, 1 dyn = 1 g * cm/s^2 = 10^-5 N % 1 GPa = 10^9 kg/(m * s^2) = 10^12 g/(m * s^2) = 10^10 g/(cm *s^2) % = 10e10 dyn/cm^2 miu = 30 * 1.0e10; % this is dyne/cm^2 % d in logM0=c-dM dKanamori = 1.5; % c in logM0=c-dM cKanamori = 16.05; % : Fault length (km) -->cm (*1.0e05) fLength = length * 1.0e05; % Fault width (km) -->mm fWidth=(abs(upper_depth-lower_depth))/sind(dip)*1.0e05; % aspect ratio %aspectRatio = fLength/fWidth; % fWidth= 10 * 1.0e05; % Minimum magnitude Mmin = 0; % Maximum magnitude fMmax=maxmag; % magnitude binnning deltMFD = 0.1; % parameters for Recurrence Model % bbar value b_bar = bVal * log(10); % Magnitude scaling coefficient d_bar = dKanamori * log(10); % Fault slip-length ratio alpha = 1.0e-04; 100 APPENDIX – PROGRAMMING % Seismic Moment-Magnitude scaling for Mw=0, units are dyne/cm^2 momZero = 10^cKanamori; % beta coeficient Model 2: Anderson and Luco beta_numerator = alpha * momZero; beta_denominator = miu * fWidth; beta = sqrt(beta_numerator/beta_denominator); % MFD vMagnitude = Mmin:deltMFD:fMmax; %% Calculation factors for Recurrence MOdel No 2 fFac1 = (d_bar-b_bar) / d_bar; fFac2 = slipRate / beta; fFac3 = exp(b_bar * (fMmax - vMagnitude)); fFac4 = exp(-(d_bar / 2) * fMmax); %% Bungum Equations 7: Originally by Anderson and Luco, BSSA, 1983 vCumNumber = fFac1 * fFac2 * fFac3 * fFac4; vMagnitude = Mmin:deltMFD:fMmax; %% compute aGR-value % cumulative a_cum_value_max1 = log10(vCumNumber) + bVal * vMagnitude; a_cum_value_max=a_cum_value_max1(1,1) MATLAB FUNCTION FOR THE CONSTRUCTION OF XML FILES *Original code was provided by Dr. Laurentiu Danciu function ok = writefxml(code,name,longitude,latitude,dip,upper_depth,lower_d epth,aspect_ratio,maxmag,a_cum_value_max,rake,filename_xml) %construction of XML file docNode=com.mathworks.xml.XMLUtils.createDocument('nrml'); nrml=docNode.getDocumentElement; nrml.setAttribute('xmlns:gml','http://opengis.net/gml'); nrml.setAttribute('xmlns','http://openquake.org/xmlns/nrml/0.4 '); %write source model source_model_element=docNode.createElement('sourceModel'); 101 APPENDIX – PROGRAMMING source_model_element.setAttribute('name','Simple Fault Model'); nrml.appendChild(source_model_element); %write simple faults source element for q=1:numel(name) %use of 'for' loop - type all faults in one XML file simple_fault_source_element=docNode.createElement('simpleFault Source'); simple_fault_source_element.setAttribute('id',num2str(code{q}) ); %%% simple_fault_source_element.setAttribute('name',name{q}); %%% %tectonic region-Active Shallow Crust simple_fault_source_element.setAttribute('tectonicRegion','Act ive Shallow Crust'); source_model_element.appendChild(simple_fault_source_element); %add simple fault geometry simpleFaultGeometry_element=docNode.createElement('simpleFault Geometry'); simple_fault_source_element.appendChild(simpleFaultGeometry_el ement); gml_LineString=docNode.createElement('gml:LineString'); simpleFaultGeometry_element.appendChild(gml_LineString); %add the vertex list of each polyline in clock-or counter clock wise gml_posList=docNode.createElement('gml:posList'); gml_LineString.appendChild(gml_posList); llon=longitude{q} llat=latitude{q} for i=1:length(llon)-1 %coordinates of each fault gml_posList.appendChild(docNode.createTextNode([num2str(llon(i )) ' ' num2str(llat(i)) ' '])) end 102 APPENDIX – PROGRAMMING %add dip dip_element=docNode.createElement('dip'); simpleFaultGeometry_element.appendChild(dip_element); dip_element.appendChild(docNode.createTextNode(num2str(dip(q)) )); %%% %add upper seismogenic depth element upperSeismoDepth_element=docNode.createElement('upperSeismoDep th'); simpleFaultGeometry_element.appendChild(upperSeismoDepth_eleme nt); upperSeismoDepth_element.appendChild(docNode.createTextNode(nu m2str(upper_depth(q)))); %%% %add lower seismogenic depth element lowerSeismoDepth_element=docNode.createElement('lowerSeismoDep th'); simpleFaultGeometry_element.appendChild(lowerSeismoDepth_eleme nt); lowerSeismoDepth_element.appendChild(docNode.createTextNode(nu m2str(lower_depth(q)))); %%% %add magnitude scaling relationship magScaleRel_element=docNode.createElement('magScaleRel'); simple_fault_source_element.appendChild(magScaleRel_element); magScaleRel_element.appendChild(docNode.createTextNode('WC1994 ')); %add rupture aspect ratio ruptAspectratio_element=docNode.createElement('ruptAspectRatio '); simple_fault_source_element.appendChild(ruptAspectratio_elemen t); ruptAspectratio_element.appendChild(docNode.createTextNode(num 2str(aspect_ratio(q)))); %%% %%add truncGutenbergRichterMFD_element truncGutenbergRichterMFD_element = docNode.createElement('truncGutenbergRichterMFD'); truncGutenbergRichterMFD_element.setAttribute('aValue', num2str(a_cum_value_max(q))); truncGutenbergRichterMFD_element.setAttribute('bValue', num2str(1.0)); truncGutenbergRichterMFD_element.setAttribute('maxMag', num2str(maxmag(q))); %%% 103 APPENDIX – PROGRAMMING truncGutenbergRichterMFD_element.setAttribute('minMag', num2str(4.5)); simple_fault_source_element.appendChild(truncGutenbergRichterM FD_element); %add rake rake_element=docNode.createElement('rake'); simple_fault_source_element.appendChild(rake_element); rake_element.appendChild(docNode.createTextNode(num2str(rake(q )))); end xmlwrite(filename_xml,docNode); ok=1; MATLAB FUNCTION FOR THE CONSTRUCTION OF LOGIC TREE XML FILE *Original code was provided by Dr. Laurentiu Danciu function [kk]=logic_tree(filename_xml) %construction of logic tree XML file docNode=com.mathworks.xml.XMLUtils.createDocument('nrml'); nrml=docNode.getDocumentElement; nrml.setAttribute('xmlns:gml','http://opengis.net/gml'); nrml.setAttribute('xmlns','http://openquake.org/xmlns/nrml/0.4 '); logic_tree_element=docNode.createElement('logicTree'); logic_tree_element.setAttribute('logicTreeID','lt1'); nrml.appendChild(logic_tree_element); %1st branching level for the source models logic_tree_branching_level_element=docNode.createElement('logi cTreeBranchingLevel'); logic_tree_branching_level_element.setAttribute('branchingLeve lID','bl1'); logic_tree_element.appendChild(logic_tree_branching_level_elem ent); 104 APPENDIX – PROGRAMMING logic_tree_branch_set_element=docNode.createElement('logicTree BranchSet'); logic_tree_branch_set_element.setAttribute('uncertaintyType',' sourceModel'); logic_tree_branch_set_element.setAttribute('branchSetID','bs1' ); logic_tree_branching_level_element.appendChild(logic_tree_bran ch_set_element); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b1'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_1a.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.111))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b2'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_2b.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.111))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b3'); 105 APPENDIX – PROGRAMMING logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_3c.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.111))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b4'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_4d.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.111))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b5'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_5e.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); 106 APPENDIX – PROGRAMMING uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.112))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b6'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_6f.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.111))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b7'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_7g.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.111))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b8'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); 107 APPENDIX – PROGRAMMING uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_8h.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.111))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b9'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(' SLIP_RATE_9i.xml')); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.111))); %2nd branching level for b_value logic_tree_branching_level_element=docNode.createElement('logi cTreeBranchingLevel'); logic_tree_branching_level_element.setAttribute('branchingLeve lID','bl2'); logic_tree_element.appendChild(logic_tree_branching_level_elem ent); logic_tree_branch_set_element=docNode.createElement('logicTree BranchSet'); logic_tree_branch_set_element.setAttribute('uncertaintyType',' bGRRelative'); logic_tree_branch_set_element.setAttribute('branchSetID','bs21 '); logic_tree_branching_level_element.appendChild(logic_tree_bran ch_set_element); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b211'); 108 APPENDIX – PROGRAMMING logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(n um2str(0.9))); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.333))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b212'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(n um2str(1.0))); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.334))); logic_tree_branch_element=docNode.createElement('logicTreeBran ch'); logic_tree_branch_element.setAttribute('branchID','b213'); logic_tree_branch_set_element.appendChild(logic_tree_branch_el ement); uncertainty_model_element=docNode.createElement('uncertaintyMo del'); logic_tree_branch_element.appendChild(uncertainty_model_elemen t); uncertainty_model_element.appendChild(docNode.createTextNode(n um2str(1.1))); uncertainty_weight_element=docNode.createElement('uncertaintyW eight'); logic_tree_branch_element.appendChild(uncertainty_weight_eleme nt); 109 APPENDIX – PROGRAMMING uncertainty_weight_element.appendChild(docNode.createTextNode( num2str(0.333))); %type the logic tree XML file xmlwrite(filename_xml,docNode); type(filename_xml); end 110 REFERENCES Abrahamson, N.A. (2006). Notes on probabilistic seismic hazard analysis - an overview. Rose School, Pavia, Italy. Abrahamson, N.A., Bommer, J.J. (2005). Opinion paper: probability and uncertainty in seismic hazard analysis. Earthquake spectra, Vol. 21, no.2, pp. 1-5. Abrahamson, N.A., Silva, W. (2008). Summary of the Abrahamson & Silva NGA ground motion relations. Earthquake Spectra, 24(1):67-97. Abrahamson, N.A., Silva, W.J. (1997). Empirical response spectral attenuation relations for shallow crustal earthquakes. Seismological Research Letters, Vol. 68, no.1. Abrahamson, N.A., Somerville, P.G. (1996). Effects of the hanging wall and footwall on ground motion recorded during Northridge earthquake. Bulletin of the Seismological Society of America, Vol. 86, pp. S93-S99. Agalos, A., Papadimitriou, P., Makropoulos, K. (2007). Source parameters estimation from broadband regional seismograms for earthquakes in the Aegean region and the Gorda plate. Bulletin of Geological Society, 40, 1032-1044, Greece. Aiping, T., Xiaxin, T. (2013). Logic trees for probabilistic seismic hazard analysis in low seismological hazard zone. The 3rd Annual Trilateral of the Strategic Chinese-KoreanJapanese Cooperative Program: Seismic Hazard Assessment for the Next Generation Map, June 16-19, 2013, Sendai, Japan. Aki, K. (1966). Generation and propagation of G-waves from Niigara earthquake of June 16, 1964. Estimation of earthquake movement, released energy and stress-strain drop from G spectrum. Bulletin of the Earthquake Research Institute, 44, 23-88. Aki, K. (1979). Characterization of barriers on an earthquake fault. Journal of Geophysical Research, 84, issn: 0148-0227. Aki, K., Richards, P.G. (2002). Quantitative seismology. University Science Books, Sausalito, California. 111 REFERENCES Allen, T., Wald, D. (2010). Prediction of macroseismic intensities for global active crustal earthquakes. Journal of Seismology. Ambraseys, N. (2009). Earthquakes in the Mediterranean and Middle East: a multidisciplinary study of seismicity up to 1900. Cambridge University Press, New York, pp. 947. Ambraseys, N.N. (2001). Reassessment of earthquakes, 1900-1999, in the Eastern Mediterranean and the Middle East. Geophysical Journal International, 145, 471-485. Ambraseys, N.W., Jackson, J.A. (1990). Seismicity and associated strain of central Greece between 1890 and 1988. Geophysical Journal International, 101, 663-708. Anastasakis, G., Piper, D.J.W., Tziavos, C. (2007). Sedimentological response to netotectonics and sea-level change in a delta-fed, complex graben: gulf of Amvrakikos, western Greece. Marine Geology, 236, 27-44. Anderson, H., Jackson, J. (1987). Active tectonics of the Adriatic region. Geophysical Journal of the Royal Astronomical Society, 91, 937-983. Anderson, J.G., Luco, J.E. (1983). Consequences of slip rate constants on earthquake occurrence relations. Bulletin of the Seismological Society of America, 73, 471-496. Angelier, J. (1978). Tectonic evolution of the Hellenic arc since the late Miocene. Tectonophysics, 49, 23-36. Angelier, J., Lyberis, N., Le Pichon, X., Barrier, E., Huchon, P. (1982). The tectonic development of the Hellenic arc and the sea of Crete: a synthesis. Tectonophysics, 86, 159-196. Atkinson, G.A., Boore, D.M. (2006). Earthquake ground motion prediction equations for eastern north America. Bulletin of the Seismological Society of America, 96(6):21812205. Atkinson, G.M., Boore, D.M. (1997). Earthquake ground motion prediction equations for eastern north America. Bulletin of the Seismological Society of America, Vol. 93, no.1, pp. 314-331. Baker, C., Hatzfeld, D., Lyon-Caen, H., Papadimitriou, E., Rigo, A. (1997). Earthquake mechanisms of the Adriatic sea and western Greece. Geophysical Journal International, 131, 559-594. Baker, J. (2008). An introduction to Probabilistic Seismic Hazard Analysis (PSHA). Vol. 1.3. Basili, R., Kastelic, V., Valensise, G., and DISS Working Group 2009 (2009). DISS3 tutorial series: guidelines for compiling records of the database of individual seismogenic sources, version 3. INGV, Rapporti Technici, 108, 16pp. 112 REFERENCES Bazzuro, P., Cornell, A.C. (1999). Disaggregation of seismic hazard. Bulleting of the Seismological Society of America, Vol. 89, no.2, pp. 501-520. Bell, R.E., McNeill, L.C., Bull, J.M., Henstock, T.J., Collier, R.E.L., Leeder, M.R. (2009). Fault architecture, basin structure & evolution of the Gulf of Corinth Rift, central Greece. Basin Research, 21, 824-855. Bender, B., Perkins, D.M. (1982). SEISRISK II: a computer program for seismic hazard estimation. Open-file report 82-293, United States, Department of the Interiors, Geological Survey. Bender, B., Perkins, D.M. (1987). SEISRISK III: a computer program for seismic hazard estimation. Bulletin 1772, United States, Geological Survey. Benedetti, L., Finkel, R., Papanastassiou, D., King, G., Armijo, R. (2002). Post-glacial slip history of the Sparta fault (Greece) determined by 36Cl cosmogenic dating. Evidence for non-periodic earthquakes. Geophysical Research Letter, 29(8), 87(1-4). Benetatos, C., Dreger, D., Kiratzi, A. (2007). Complex and segmented rupture associated with the 14 August Mw 6.2 Lefkada, Ionian islands, earthquake. Bulletin of Seismological Society of America, 97(1B), 35-51. Benetatos, C., Kiratzi, A., Roumelioti, Z., Stavrakakis, G., Drakatos, G., Latoussakis, I. (2005). The 14 August 2003 Lefkada island (Greece) earthquake: focal mechanisms of the mainshock and of the aftershock sequence. Journal of Seismology, 9, 171-190. Beresnev, I.A. (2002). Source parameters observable from the corner frequency of earthquake spectra. Bulletin of the Seismological Society of America, Vol. 92, no.5, pp. 2047-2048. Billion, P. (2007). Earthquake slip-length scaling: new insights from an expanded data base. Master thesis. LGIT, Grenoble. Boccaletti, M., Caputo, R., Mountrakis, D., Pavlides, S., Zouros, N. (1997). Paleoseismicity of the Souli fault, Epirus, western Greece. Journal of Geodynamics, 24, 1-4, 117-127. Bolt, B.A. (1969). Duration of strong motion. Proceedings of the 4 th Conference on Earthquake Engineering, pp. 1304-1315. Santiago, Chile. Bommer, J., Scherbaum, F. (2008). The use and misuse of logic trees in probabilistic seismic hazard analysis. Earthquake Spectra: November 2008, Vol.24, no.4, pp.997-1009. Boore, D.M. (2003). Simulation of ground motion using the stochastic method. Pure & Applied Geophysics, Vol. 160, pp. 635-676. Boore, D.M., Atkinson, G.M. (2008). Ground-motion prediction equations for the average horizontal component of PGA, PGV, and 5%-damped PSA at spectral periods between 0.01 s and 10.0 s. Earthquake Spectra, 24(1):99-138. 113 REFERENCES Boore, D.M., Joyner, W.B., Fumal, T.E. (1997). Equations for estimating horizontal response spectra and peak acceleration from western north American earthquakes: a summary of recent work. Seismological Research Letters, 68(1):128-153. Brooks, M., Clews, J.E., Melis, N.S., Underhill, J.R. (1988). Structural development of neogene basins in western Greece. Basin Research, 1, 129-138. Brooks, M., Ferentinos, G. (1984). Tectonins and sedimentation in the Gulf of Corinth and the Zakynthos and Kefallinia channels, western Greece. Tectonophysics, 101, 25-54. Brune, J.N. (1968). Seismic moment, seismicity, and rate of slip along major fault zones. Journal of Geophysical Research, 73, 777-784. Bungum, H. (2007). Numerical modeling of fault activities. Computer & Geosciences, 33, 808-820. Burton, P.W., Melis, N.S., Brooks, M. (1995). Coseismic crustal deformation on a fault zone defined by microseismicity in the Pavliani area, central Greece. Geophysical Journal International, 123, 16-40. Campbell, K.W. (1997). Empirical near-source attenuation relationships for horizontal and vertical components of peak ground acceleration, peak ground velocity and pseudoabsolute acceleration response spectra. Seismological Research Letters, 68(1):154-179. Campbell, K.W., Bozorgnia, Y. (2003). Updated near-source ground motion (attenuation) relations for the horizontal and vertical components of peak ground acceleration and acceleration response spectra. Bulletin of the Seismological Society of America, Vol. 93, no.1, pp. 314-331. Campbell, K.W., Bozorgnia, Y. (2008). NGA ground motion model for the geometric mean horizontal component of PGA, PGV, PGD and ranging from 0.01 to 10 s. Earthquake Spectra, 24(1):139-171. Caputo, R. (1990). Geological and structural study of the recent and active brittle deformation of the neogene-quaternary basins of Thessaly (central Greece). Scientific Annals, 12, Aristotle University of Thessaloniki, Thessaloniki, Greece, pp 252. Caputo, R. (1995). Inference of a seismic map from geological data: Thessaly (central Greece) as a case of study. Annals of Geophysics, 38, 1, 1-19. Caputo, R. (1996). The active Nea Anchialos fault system (central Greece): comparison of geological, morphotectonic, archaeological and seismological data. Annals of Geophysics, 39, 3, 557-574. Caputo, R., Helly, B. (2005). The Holocene activity of the Rodia fault, central Greece. Journal of Geodynamics, 40, 153-169. 114 REFERENCES Caputo, R., Pavlides, S. (1991). Neotectonics and structural evolution of Thessaly (central Greece). Bulletin of Geological Society, XXV, 119-133, Greece. Caputo, R., Pavlides, S. (1993). Late cainozoic geodynamic evolution of Thessaly and surroundings (central-northern Greece). Tectonophysics, 223, 339-362. Caputo, R., Zouros, N. (1993). Examples of Alpide deformation from Epirus: local anomalies or need to re-evaluate the amount of shortening in the western Hellenides? Bulletin of Geological Society, XXVIII, 1, 315-326, Greece. Chiang, W.L., Guidi, G.A., Scoof, C.G., Shah, H.C. (1984). Computer programs for seismic hazard analysis-a user manual (STASHA). Report 62, The J.A. Blume Earthquake Engineering Center. Chinnery, M.A., North, R.G. (1975). The frequency of very large earthquakes. Science, 190, 1197-1198. Chiou, B.S.-J., Youngs, R.R. (2008). An NGA model for the average horizontal component of peak ground motion and response spectra. Earthquake Spectra, 24:173-215. Chouliaras, G. (2009). Seismicity anomalies prior to 8 June 2008, Mw=6.4 earthquake in Western Greece. Natural Hazards and Earth System Sciences, 9, 327-335. Chousianitis, K., Agalos, A., Papadimitriou, P., Lagios, E., Makropoulos, K. (2010). Source parameters of moderate and strong earthquakes in the broader area of Zakynthos island (W. Greece) from regional and teleseismic digital recordings. Bulletin of Geological Society, 43(4), 2005-2014, Greece. Chung-Han, C., Grunthal, G. (2011). Hybrid zoneless probabilistic seismic hazard assessment: test and first application to Europe and the Mediterranean. NERIES, JRA2, Deliverable 8. Collier, R.E.L., Pantosti, D., D‟ Addezio, G., De Martini, D.M., Masana, E., Sakellariou, D. (1998). Paleoseismicity of the 1981 Corinth earthquake fault: seismic contribution to extensional strain in central Greece and implications for seismic hazard. Journal of Geophysical Research, 103, 30, 001-30, 019. Cornell, C.A. (1968). Engineering seismic risk analysis. Bulletin of the Seismological Society of America, Vol. 58, pp. 1583-1606. Cornell, C.A., Winterstein, S. (1988). Temporal and magnitude dependence in earthquake recurrence models. Bulletin of the Seismological Society of America, Vol. 78, no.4, pp. 1522-1537. 115 REFERENCES Crowley, H., Cerisara, A., Jaiswal, K., Keller, N., Luco, N., Pagani, M., Porter, K., Silva, V., Wald, D., B., W. (2010a). GEM1 seismic risk report: part 2. GEM Technical Report 2010-5. GEM Foundation, Pavia, Italy. Crowley, H., Colombi, M., Crempien, J., Erduran, E., Lopez, M., Liu, H., Mayfield, M., Milanesi, M. (2010b). GEM1 seismic risk report: part 1. GEM Technical Report 20105. GEM Foundation, Pavia, Italy. Crowley, H., Monelli, D., Pagani, M., Silva, V., Weatherill, G. (2011). OpenQuake book. The GEM Foundation, Pavia, Italy. Crowley, H., Monelli, D., Pagani, M., Silva, V., Weatherill, G. (2013). OpenQuake engine user instruction manual. GEM Foundation, Pavia, Italy. Cundy, A.B., Kortekaas, S., Dewez, T., Stewart, I.S., Collins, P.E.F., Croudace, I.W., Maroukian, H., Papanastassiou, D., Gaki-Papanastassiou, P., Pavlopoulos, K., Dawson, A. (2000). Coastal wetlands as recorders of earthquake subsidence in the Aegean: a case study of the 1894 Gulf of Atalanti earthquakes, central Greece. Marine Geology, 170, 326. Danciu, L., Monelli, D., Pagani, M., Wiemer, S. (2010).GEM1 hazard: review of PSHA software. GEM Technical Report 2010-2. GEM Foundation, Pavia, Italy. Danciu, L., Sokos, E., Tselentis, G.-A. (2007). Probabilistic seismic hazard assessment in terms of engineering parameters in Greece. International Symposium on Seismic Risk Reduction, 26-27 April 2007, Bucharest, Romanian Academy. Danciu, L., Tselentis, G.-A. (2007). Engineering ground motion parameters attenuation relationships for Greece. Bulletin of the Seismological Society of America, 97, 1-22. Dewey, J.F., Sengor, M.C. (1979). Aegean and surrounding regions: complex multiplate and continuum tectonics in a convergent zone. Bulletin of Geological Society of America, 90, 84-92. Dologlou, E. (2009). Power law relationship between parameters of earthquakes and precursory electrical phenomena revisited. Natural Hazards and Earth System Sciences, 9, 17-23. Doutsos, T., Kokkalas, S. (2001). Stress and deformation patterns in the Aegean region. Journal of Structural Geology, 23, 455-472. Drakos, A.G., Stiros, S.C., Kiratzi, A.A. (2001). Fault parameters of the 1980 (Mw 6.5) Volos, central Greece earthquake from inversion of repeated leveling data. Bulletin of Seismological Society of America, 91, 6, 1673-1684. 116 REFERENCES Dziewonski, A.M., Ekstrom, G., Franzen, J.E., Woodhouse, J.H. (1988). Global seismicity of 1980: centroid-moment tensor solutions for 515 earthquakes. Physics of the Earth and Planetary Interiors, 50, 127-154. Dziewonski, A.M., Friedman, A., Woodhouse, J.H. (1983). Centroid-moment tensor solutions for January-March, 1983. Physics of the Earth and Planetary Interiors, 33, 71-75. Dziewonski, A.M.,Ekstrom, G., Maternovskaya, N.N. (1998). Centroid-moment tensor solutions for October-December, 1997. Physics of the Earth and Planetary Interiors, 109(3-4), 93-105. EAK (2003). Greek seismic code (in Greek). Earthquake planning & protection organization, ed. Athens-Greece, pp. 72-77 appendixes. Ekstrom, G., Dziewonski, A.M., Maternovskaya, N.N., Nettles, M. (2005). Global seismicity of 2003: centroid-moment-tensor solutions for 1087 earthquakes. Physics of the Earth and Planetary Interiors, 148, 327-351. Ekstrom, G., Nettles, M. (1997). Calibration of the HGLP seismograph network and centroidmoment tensor analysis of significant earthquakes of 1976. Physics of the Earth and Planetary Interiors, 101(3-4), 219-243. Ekstrom, G., Nettles, M., Dziewonski, A.M. (2012). The global CMT project 2004-2010: centroid-moment tensors for 13,017 earthquakes. Physics of the Earth and Planetary Interiors, 200-201, 1-9. Ekstron, G., England, P. (1989). Seismic strain rates in regions of distributed continental deformation. Journal of Geophysical Research, 94, 10231-10257. Elias, P., Briole, P., Sykioti, O. (2011). Synergy of SAR acquisitions for ground deformation monitoring by means of PSI, SBAS and DInSAR. Case study of western rift of Corinth, Greece. Geophysical Research Abstracts, Vol. 13, EGU2011-10938. Esteva, L. (1977). Microzoning: models and reality. Proceedings of the World Conference on Earthquake Engineering, 6th, New Delhi. Feng, L., Newman, A.V., Farmer, G.T., Psimoulis, P., Stiros, S.C. (2010). Energetic rupture, coseismic and post-seismic response of the 2008 Mw 6.4 Achaia-Elia earthquake in northwestern Peloponnese, Greece: an indicator of an immature transform fault zone. Geophysical Journal International, 183, 103-110. Field, E.H. (2000). A modified ground-motion attenuation relationship for southern California that accounts for detailed site classification and a basin-depth effect. Bulletin of the Seismological Society of America, 90(6B):S209-S221. 117 REFERENCES Field, E.H., Jordan, T.H., Cornell, C.A. (2003). OpenSHA - A developing community modeling environment for seismic hazard analysis. Seismological Research Letters, 74:406-419. Flotte, N., Sorel, D., Muller, C., Tensi, J.( 2005). Along strike changes in the structural evolution over a brittle detachment fault: example of the Pleistocene Corinth-Patras rift (Greece). Tectonophysics, 403, 77-94. Fokaefs, A., Papadopoulos, G.A. (2004). Historical earthquakes in the region of Lefkada island, Ionian sea – estimation of magnitudes from epicentral intensities. Bulletin of Geological Society, 36, 1389-1395, Greece. Galanakis, D. (1997). Neotectonic structure and stratigraphy of neogene-quaternary sediments of the Almyros-Pagasitikos, Pilion, Oreon-Trikeri and Maliakos basins. PhD thesis, Aristotle University of Thessaloniki, Greece. Galanakis, D., Pavlides, S., Mountrakis, D. (1998). Recent brittle tectonic in AlmyrosPagasitikos, Maliakos, N.Euboea and Pilio. Bulletin of Geological Society, 42(1), 263273, Greece. Gallovic, F., Zahradnik, J., Krizova, D., Plicka, V., Sokos, E., Serpetsidaki, A., Tselentis, G.A. (2009). From earthquake centroid to spatial-temporal rupture evolution: Mw 6.3 Movri mountain earthquake, June 8, 2008, Greece. Geophysical Research Letters, 36, L21310. Gambini, R., Tozzi, M. (1996). Tertiary geodynamic evolution of the southern Adria microplate. Terra Nova, 8(6), 593-602. Ganas, A., Oikonomou, A.I., Tsimi, Ch. (2013). NOAfaults: a digital database for active faults in Greece. Bulletin of the Geological Society of Greece, Vol. XLVII. Proceedings of the 13th International Congress, Chania, Greece. Ganas, A., Papoulia, I. (2000). High-resolution, digital mapping of the seismic hazard within the Gulf of Evia rift, central Greece, using normal fault segments as line sources. Natural Hazards, 22, 203-223. Ganas, A., Pavlides, S., Karastathis, V. (2005). DEM-based morphometry of range-front escarpments in Attica, central Greece, and its relation to fault slip rates. Geomorphology, 65, 301-319. Ganas, A., Roperts, G.P., Memou, T. (1998). Segment boundaries, the 1894 ruptures and strain patters along the Atalanti fault, central Greece. Journal of Geodynamics, 26(2-4), 461-486. 118 REFERENCES Ganas, A., Serpelloni, E., Drakatos, G., Kolligri, M., Adamis, I., Tsimi, C., Batsi, E. (2009). The Mw 6.4 SW-Achaia (western Greece) Earthquake of 8 June 2008: seismological, field, GPS observations, and stress modeling. Journal of Earthquake Engineering, 13, 1101-1124. Ganas, A., Sokos, E., Agalos, A., Leontakianakos, E., Pavlides, S. (2006). Coulomb stress triggering of earthquakes along the Atalanti fault, central Greece: two April 1894 M6+ events and stress change patterns. Tectonophysics, 420, 357-369. Ganas, A., Spina, V., Alexandropoulou, N., Oikonomou, A., Drakatos, G. (2007). The Corini active fault in southwestern Viotia region, central Greece: segmentation, stress analysis and extensional strain patterns. Bulletin of Geological Society, 37, Greece. Gawthorpe, R.L., Hardy, S., Ritchie, B. (2003). Numerical modeling of depositional sequences in half-graben rift basins. Sedimentology, 50, 168-105. Gawthorpe, R.L., Hurst, J.M. (1993). Transfer zones in extensional basins: their structural style and influence on drainage development and stratigraphy. Journal of Geological Society, 150, 1137-1152, London. GEM Foundation (2010). GEM1 executive summary. GEM Technical Report 2010-1. GEM Foundation, Pavia, Italy. Giardini, D., Woessner, J., Danciu, L., Cotton, F., Crowley, H., Grunthal, G., Pinho, R., Valensise, G., Akkar, S., Arvidsson, R., Basili, R., Cameelbeck, T., Campos-Costa, A., Douglas, J., Demircioglou, M.B., Erdik, M., Fonseca, J., Glavatovic, B., Lindholm, C., Makropoulos, K., Meletti, C., Musson, R., Pitilakis, K., Rovida, A., Sesetyan, K., Stromeyer, D., Stucchi, M. (2013). Seismic hazard harmonization in Europe (SHARE): online data resource, doi:10.12686. Godinho, J. (2007). Probabilistic seismic hazard analysis – an introduction to theoretical bases and applied methodology. Department of Civil Engineering, University of Patras, Greece. Goldsworth, M., Jackson, J. (2001). Migration of activity within normal fault systems: examples from the quaternary of mainland Greece. Journal of Structural Geology, 23, 489-506. Goldsworthy, M., Jackson, J. (2000). Active normal fault evolution in Greece revealed by geomorphology and drainage patterns. Journal of Geological Society, 157, 967-981, London. Goldsworthy, M., Jackson, J., Haines, J. (2009). The continuity of active fault system in Greece. Geophysical Journal International, 148, 596-618. 119 REFERENCES Green, R.A., Hall, W.J. (1994). An overview of selected seismic hazard analysis methodologies. University of Illinois at Urbana-Champaign, Urbana, Illinois. Grunthal, G. (1999a). GSHAP region 3 working group: seismic hazard assessment for central, north and northwest Europe: GSHAP region 3. Annali di Geofisica, 42(6), 999-1011. Grunthal, G., Bosse, C., Sellami, S., Mayer-Rosa, D., Giardini, D. (1996b). Compilation of the GSHAP regional seismic hazard for Europe, Africa and the Middle East. Annali di Geofysica, 42(6), 1215-1223. Gupta, I.D. (2002). The state of the art in seismic hazard analysis. ISET, Journal of Earthquake Technology, paper no.428, Vol.39, no.4, pp.311-346. Gutenberg, B., Richter, C.F. (1956). Earthquake magnitude, intensity energy and acceleration. Bulletin of the Seismological Society of America, Vol. 46, pp. 105-145. Hanks, T.C., Kanamori, H. (1979). A moment magnitude scale. Journal of Geophysical Research, 85(B5), 2348-2350. Haslinger, F., Kissling, E., Ansorge, J., Hatzfeld, D., Papadimitriou, E., Karakostas, V., Makropoulos, K., Kahle, H.-G., Peter, Y. (1999). 3D crustal structure from local earthquake tomography around the Gulf of Arta (Ionian region, NW Greece). Tectonophysics, 304, 201-218. Hatzfeld, D. (1999). The present-day tectonics of the Aegean as deduced from seismicity. Journal of Geological Society, Special Publications, 156, 415-426, London. Hatzfeld, D., Karakostas, V., Ziazia, M., Kassaras, I., Papadimitriou, E., Makropoulos, K., Voulgaris, N., Papaioannou, C. (2000). Microseismicity and faulting geometry in the gulf of Corinth (Greece). Geophysical Journal International, 141, 438-456. Hatzfeld, D., Kassaras, I., Panagiotopoulos, D., Amorese, D., Makropoulos, K., Karakaisis, G., Coutant, O. (1995). Microseismicity and strain pattern in northwestern Greece. Tectonics, 14, 773-785. Hatzfeld, D., Pedotti, G., Hatzidimitriou, P., Makropoulos, K. (1990). The strain pattern in the western Hellenic arc deduced from a microearthquake survey. Geophysical Journal International, 101, 181-202. Hatzfeld, D., Ziazia, M., Kementzetzidou, D., Hatzidimitriou, P., Panagiotopoulos, D., Makropoulos, K., Papadimitriou, P., Deschamps. A. (1999). Macroseismicity and focal mechanisms at the western termination of the north Anatolian fault and their implications for continental tectonics. Geophysical Journal International, 137, 891-908. 120 REFERENCES Henry, C., Das, S. (2001). Aftershock zones of large shallow earthquakes: fault dimensions, aftershock area expansion and scaling relations. Geophysical Journal International, Vol. 147, pp. 272-293. Hollenstein, C., Muller, M.D., Geigen, A., Kahle, H.-G. (2008). Crustal motion and deformation in Greece from a decade of GPS measurements, 1993_2003. Tectonophysics, 449, 17-40. IFP-IGRS. (1966). Etude geologique de l‟ Epire. Ed. Technip, pp. 36. IGME (Institute of Geological and Mining Research) (1989). Seismotectonic map of Greece. IGME, scale 1:500,000, Athens. Jackson, J. (1999). Fault death: a perspective from actively deforming regions. Journal of Structural Geology, 21, 1003-1010. Jackson, J., McKenzie, D. (1983). The geometrical evolution of normal fault systems. Journal of Structural Geology, 5(5), 471-482. Jackson, J.A., Gagnepain, J., Houseman, G., King, G.C.P., Papadimitriou, P., Soufleris, C., Virieux, J. (1982). Seismicity, normal faulting and the geomorphological development of the gulf of Corinth (Greece): the Corinth earthquakes of February and March 1981. Earth and Planetary Science Letters, 57, 377-397. Jackson, J.A., King, G., Vitza-Frintzi, C. (1982). The neotectonics of the Aegean: an alternative view. Earth and Planetary Science Letters, 61, 303-318. Jackson. J., McKenzie, D. (1999). A hectare of fresh striations on the Arkitsa fault, central Greece. Journal of Structural Geology, 21, 1-6. Jenkins, D.A.L. (1972). Structural development of western Greece. AAPG Bulletin, v. 56, p.128-149. Jimenez, M.-J., Giardini, D., Grunthal, G. (2003). The ESC-SESAME unified hazard model for the European-Mediterranean region. EMSC/CSEM Newsletter, 19, 2-4. Jolivet, L. (1993). Extension of thickened continental crust from brittle to ductile deformation: examples from Alpine Corsica and Aegean sea. Annals of Geophysics, 36(2), 139-153. Kamberis, E., Ioakim, Ch., Tsaila-Monopolis, St., Tsapralis, V. (1992). Geodynamic and paleogeographic evolution of western Peloponnesus (Greece) during the neogene. Paleontologia I Evolucio, v. 24-25, 363-376. Kanamori, H., Anderson, D.L. (1975). Theoretical basis of some empirical relations in seismology. Bulletion of Seismological Society of America, 65, 1073-1095. 121 REFERENCES Karakaisis, G.F., Papazachos, C.B., Scordilis, E.M. (2010). Seismic sources and main seismic faults in the Aegean and surrounding area. Bulletin of Geological Society, 43(4), 20262042. Greece. Karakostas, V.G., Papadimitriou, E.E., Karamanos, C.K., Kementzetzidou, D.A. (2010). Microseismicity and seismotectonics properties of the Lefkada-Kefalonia seismic zone. Bulletin of the Geological Society, 53(4), 2053-2063, Greece. Karakostas, V.G., Papadimitriou, E.E., Papazachos, C.B. (2004). Properties of the 2003 Lefkada, Ionian islands, Greece, earthquake seismic sequence and seismicity triggering. Bulletin of Seismological Society of America, 94(5), 1976-1981. Karastathis, V.K., Ganas, A., Makris, J., Papoulia, J., Dafnis, P., Gerolymatou, E., Drakatos, G. (2007). The application of shallow seismic techniques in the study of active faults: the Atalanti normal fault, central Greece. Journal of Applied Geophysics, 62, 215-233. Katopodi, I., Iosifidou, K. (2004). Impact of the Lefkada earthquake (14-08-2003) on marine works and coastal regions. 7th Pan-hellenic Geographical Conference, October 14-17, Mytilene, Greece, paper E1K230. Kementzetzidou, D. (1996). Etude sismotectonique du systeme Thessalie-iles Sporades (Grece central). PhD Thesis, Observatoire de Grenoble, 151 pp. Kim, Y.-S., Sanderson, D. (2005). The relationship between displacement and length of faults: a review. Earth-Science Reviews, 68, 317-334. King, G., Sturdy, D., Whitney, J. (1993). The landscape geometry and active tectonics of northwest Greece. Bulletin of Geological Society of America, 105, 137-161. King, G., Tselentis, A., Gomberg, J., Molnar, P., Roecker, S., Sivhal, H., Soufleris, C., Stock, J. (1983). Microearthquake seismicity and active tectonics of northwestern Greece. Earth and Planetary Science Letters, 66, 279-288. King, G.C.P., Ouyang, Z.X., Papadimitriou, P., Deschamps, A., Gagnepain, J., Houseman, G., Jackson, J.A., Soufleris, C., Virieux, J. (1985). The evolution of the gulf of Corinth (Greece): an aftershock study of the 1981 earthquakes. Geophysical Journal of the Royal Astronomical Society, 80, 677-693. Kiratzi, A., Benetatos, C., Roumelioti, Z. (2007). Distributed earthquake focal mechanisms in the Aegean sea. Bulletin of Geological Society, Greece. Kiratzi, A., Louvari, E. (2003). Focal mechanisms of shallow earthquakes in the Aegean sea and the surrounding lands determined by waveform modeling: a new database. Journal of Geodynamics, 36, 251-274. 122 REFERENCES Kiratzi, A.A., Karakaisis, G.F., Papadimitriou, E.E., Papazachos, B.C. (1985). Seismic source-parameter relations for earthquakes in Greece. Pure and Applied Geophysics, 123, 27-41. Kiratzi, A.A., Langston, C.A. (1991). Moment tensor inversion of the 1983 January 17 Kefallinia event of Ionian islands (Greece). Geophysical Journal International, 105, 529-535. Kiremidjian, A., Anagnos, T. (1984). Stochastic time-predictable model for earthquake occurrences. Bulletin of the Seismological Society of America, Vol. 74, no.6, pp. 25932611. Klugel, J.-U. (2008). Seismic hazard analysis - quo vadis? Earth-Science Reviews, 88, 1-32. Kokinou, E., Kamberis, E. (2009). The structure of the Kythira Antikythira strait, offshore SW Greece (35.7° 36.6°N). Journal of Geological Society of London, Special Publications, 311, 343-360. Kokinou, E., Kamberis, E., Vafidis, A., Monopolis, D, Ananiadis, G., Zelilidis, A. (2005). Deep seismic reflection data from offshore western Greece: a new crustal model for the Ionian sea. Journal of Petroleum Geology, 28(2), 185-202. Kokkalas, S., Koukouvelas, I., Pavlides, S., Chatzipetros, A. (2004). Evidence of paleoseismicity in Greece – some case studies for understanding active fault behavior. 5th International Symposium of Eastern Mediterranean Geology, April 14-20, 2004, Thessaloniki, Greece, Ref: S2-16. Kokkalas, S., Pavlides, S., Koukouvelas, I., Ganas, A., Stamatopoulos, L. (2007). Paleoseismicity of the Karapelli fault (eastern Corinth gulf): evidence for earthquake reccurence and fault behavior. Bulletin of Geological Society, 126(2), 387-395, Italy. Kokkalas, S., Pavlides, S., Koukouvelas, I., Ganas, A., Tsodoulos, I., Stamatopoulos, L., Gountromichou, C., Valkaniotis, S. (2007). Preliminary paleoseismological faults from Karapelli fault (central Greece): evidence of seismic events for the past 10.000 years. Hellenic Journal of Geosciences, 42, 65-74. Kokkalas, S., Xypolias, P., Koukouvelas, I., Doutsos, T. (2006). Postcollisional contractional and extensional deformation in the Aegean region. Geological Society of America, Special Paper 409, 97-123. Komodromos, P. (2012). Advanced structure analysis using computational methods (in Greek). Course notes. University of Cyprus, Nicosia. Konstantinou, K.I., Melis, N.S., Lee, S.-J., Evangelidis, C.P., Boukouras, K. (2009). Rupture process and aftershocks relocation of the 8 June 2008 Mw 6.4 earthquake in northwest 123 REFERENCES Peloponnese, western Greece. Bulletin of the Seismological Society of America, 99(6), 3374-3389. Koravos, G. (2011). A contribution to the study of seismic hazard estimation in the broader Aegean area based on design earthquakes. PhD thesis. Geophysical Laboratory, School of Geology, Aristotle University of Thessaloniki, Greece. Koukouvelas, I., Mpresiakas, A., Sokos, E., Doutsos, T. (1996). The tectonic setting and earthquake ground hazards of the 1993 Pyrgos earthquake, Peloponnese, Greece. Journal of the Geological Society, v. 153, p. 34-39. Koukouvelas, I.K., Katsonopoulou, D., Soter, S., Xypolias, P. (2005). Slip rates on the Helike fault, Gulf of Corinth, Greece: new evidence from geoarchaeology. Terra Nova, Vol.17, no.2, 158-164. Koukouvelas, I.K., Kokkalas, S., Xypolias, P. (2010). Surface deformation during the Mw 6.4 (8 June 2008) Movri mountain earthquake in the Peloponnese, and its implications for the seismotectonics of western Greece. International Geology Review, 52(2-3), 249268. Koukouvelas, I.K., Kokkalas, S.A., Zygouri, V.N. (2010). Geology & earthquakes (in Greek). Disigma Publications. Koukouvelas, I.K., Stamatopoulos, L., Katsonopoulou, D., Pavlides, S. (2001). A palaeoseismological and geoarchaeological investigation of the Eliki fault, Gulf of Corinth, Greece. Journal of Structural Geology, 23, 531-543. Koutromanos, I., Spyrakos, K. (2010). Earthquake-resistant buildings (in Greek). Course notes. Laboratory of Earthquake Engineering, School of Civil Engineering, NTUA, Athens. Kraft, J.C., Rapp, G.JR., Ascenbrenner, S.E. (1975). Late holocene paleogeography of the coastal plain of the Gulf of Messenia, Greece, and its relationships to arcaeological settings and coastal change. Bulletin of the Geological Society of America, 86, 11911208. Kramer, S.L. (1996). Geotechnical Earthquake Engineering. Prentice Hall International Series in Civil Engineering & Engineering Mechanics. Kreemer, C., Chamot-Rooke, N. (2004). Contemporary kinematics of the southern Aegean and the Mediterranean ridge. Geophysical Journal International, 157, 1377-1392. Krinitzsky, E.L., Slemmons, D.B. (1990). Neotectonics in Earthquake Evaluation. Geological Society of America. Reviews in Engineering Geology, Vol. VIII. 124 REFERENCES Ladas, I., Mariolakos, I., Fountoulis, I. (2004). Neotectonic deformation of eastern Pylia (SW Peloponnese, Greece). Bulletin of Geological Society, 36, 1652-1661, Greece. Laigle, M., Alfred Hirn, A., Sachpazi, M., Clement, C. (2002). Seismic coupling and structure of the Hellenic subduction zone in the Ionian islands region. Earth and Planetary Science Letters, 200, 243-253. Laigle, M., Hirn, A., Sapin, M., Lepine, J., Diaz, J., Gallart, J., Nicholich, R. (2000). Mount Etna dense array local earthquake P and S tomography and implications for volcanic plumbing. Journal of Geophysical Research, 105, 148-227. Leeder, M.R., Collier, R.E.L., Abdul Aziz, L.H., Trout, M., Ferentinos, G., Papatheodorou, G., Lyberis, E. (2002). Tectono-sedimentary processes along an active marine/lacustrine half-graben margin: Alkyonides gulf, E. gulf of Corinth, Greece. Basin Research, 14, 25-41. Leeder, M.R., Seger, M.J., Stark, C.P. (1991). Sedimentation and tectonic geomorphology adjacent to major active and inactive normal faults, southern Greece. Journal of Geological Society, 148, 331-343, London. Lekkas, E. (2001). The Athens earthquake (7 September 1999): intensity distribution and controlling effects. Engineering Geology, Vol.59, 297-311. Lekkas, E., Kranis, H., Voulgaris, N. (2007). The Sophades (Thessaly) earthquake revisited: morphotectonic analysis of the Ekkara fault system and seismic risk assessment of SW Thessaly. Geophysical Research Abstracts, Vol.9, 07897. Louvari, E., Kiratzi, A.A., Papazachos, B.C. (1999). The Cefalonia Transform Fault and its extension to western Lefkada island (Greece). Tectonophysics, 308, 223-236. Lyon-Caen, H., Armijo, R., Drakopoulos, J., Baskoutas, J., Delibassis, N., Gaulon, R., Kouskouna, V., Latoussakis, J., Makropoulos, K., Papadimitriou, P., Papanastassiou, D., Pedott, G. (1988). The 1986 Kalamata (south Peloponnese) earthquake: detailed study of a normal fault evidences for East-West extension in the Hellenic Arc. Journal of Geophysical Research, 93(B12), 14967-15000. Makris, J., Papoulia, J., Papanikolaou, D., Stavrakakis, G. (2001). Thinned continental crust below northern Evoikos Gulf, central Greece, detected from deep seismic soundings. Tectonophysics, 341, 225-236. Margaris, B., Papaioannou, C., Theodulidis, N., Savvaidis, A., Anastasiadis, A., Klimis, N., Makra, K., Demosthenous, M., Karakostas, C., Lekidis, V., Makarios, T., Salonikios, T., Sous, S. (2003). Preliminary observations on the August 14, 2003, Lefkada island (western Greece) earthquake. EERI Special Earthquake Report. 125 REFERENCES Mariolakos, I., Fountoulis, I., Logos, E., Lozios, S. (1989). Surface faulting caused by the Kalamata (Greece) earthquakes (13.9.86). Tectonophysics, 163, 197-203. Mariolakos, I., Fountoulis, I., Logos, S. (1989). Methods to study the torsional neotectonic deformation: the case of Kalamata area (SW Peloponnesus, Greece). In: Qingxuan C.(Ed), Regional crustal stability and geological hazards, proceedings of the IGCP project 250, 3, 15.21. Mariolakos, I., Fountoulis, I., Marcopoulou-Diacantoni, A., Mirkou, M.R. (1994). Some remarks on the kinematic evolution of Messinia province (SW Peloponnesus, Greece) during the pleistocene based on neotectonics, stratigraphic and palaeoecological observations. Munster. Forsch. Palaont., 76, 371-380. Mariolakos, I., Schneider, H., Fountoulis, I., Vouloumanos, N. (1992). Paleogeography, sedimentation and neotectonics implications at the Kambos depression and Kitries bay area (Messinia, Peloponnesus, Greece). Bulletin of Geological Society, 28(1), 397-413, Greece. Marnelis, F., Roussos, N., Rigakis, N., Karakitsios, V. (2007). Structural geology of the western Greece, fold and thrust belt. Energy earthquake and exhibition 2007, Athens, Greece. Martinez, W.L., Martinez, A.R. (2002). Computational statistics handbook with matlab. Chapman and Hall/CRC, Boca Raton, London, New York, Washington D.C. Matthews, M.V., Ellsworth, W.L., Reasenberg, P.A. (2002). A brownian model for recurrent earthquakes. Bulletin of the Seismological Society of America, Vol. 92, no.6, pp. 22332250. Mavroulis, S., Fountoulis, I., Lekkas, E. (2010). Environmental effects caused by the Andravida (08-06-2008, ML=6.5, NW Peloponnese, Greece) earthquake. 11th IAEG, New Zealand. McGuire, K.K. (1976). Fortran computer program for seismic risk analysis. Open-file report 67-76, United States, Department of the Interiors, Geological Survey. 102 pages. McGuire, R.K. (2001). Deterministic vs. probabilistic earthquake hazards and risks. Solid dynamics and earthquake engineering, Vol. 21, pp. 377-384. McKenzie, D. (1972). Active tectonics of the Mediterranean Region. Geophysical Journal of the Royal Astronomical Society, 30, 109-185. McNeill, L.C., Cotterill, C.J., Bull, J.M., Henstock, T.J., Bell, R., Stefatos, A. (2007). Geometry and slip rate of the Aigion fault, a young normal fault system in the western Gulf of Corinth. Geology, 35, 355-358. 126 REFERENCES Metaxas, C.P. (2008). Active faults hazard on offshore pipelines: case of the submarine gas pipeline route across the south Evoikos Gulf, central Greece. Earthquake Planning & Protection Organization, Greece. Mohammadioun, B., Serva, L. (2001). Stress drop, slip type, earthquake magnitude and seismic hazard. Bulletin of the Seismological Society of America, 91, 694-707. Molnar, P., Tucker, B.E., Brune, J.N. (1973). Corner frequencies of P and S waves and models of earthquake sources. Bulletin of the Seismological Society of America, Vol. 63, no.6, pp. 2091-2104. Monelli, P., Pagani, M., Weatherill, G., Silva, V., Crowley, H. (2012). The hazard component of OpenQuake: the calculation engine of the Global Earthquake Model. Proceedings of the 15th World Conference on Earthquake Engineering, Lisbon, Portugal. Morewood, N.C., Roberts, G.P. (1999). Lateral propagation of the surface trace of the south Alkyonides normal fault segment, central Greece: its impact on models of fault growth and displacement length relationships. Journal of Structural Geology, 21, 635-652. Moshou, A., Papadimitriou, P., Makropoulos, K. (2007). Teleseismic body wave inversion. Bulletin of Geological Society, 40, 1177-1186, Greece. Mountrakis, D., Kilias, A., Pavlides, S., Zouros, N., Spyropoulos, N., Tranos, M., Soulakelis, N. (1993). Field study of the southern Thessaly highly active fault zone. Proc. 2 nd Congr. Hellenic Geophysical Union, May 5-8, 1993, Florina, 603-614. Musson, R.M.W. (2000). The use of Monte Carlo simulations for seismic hazard assessment in the U.K. Annals of Geophysics, 43(1):1-9. Orhan, A., Seyrek, E., Tosun, H. (2007). A probabilistic approach for earthquake hazard assessment of the province of Eskisehir, Turkey. Natural Hazards and Earth System Sciences, 7, 607-614. Pagani, M., Monelli, D., Crowley, H., Danciu, L., Field, E.H., Wiemer, S., Giardini, D. (2010). GEM1 hazard: description of input models, calculation engine and main results. GEM Technical Report 2010-3. GEM Foundation, Pavia, Italy. Pantosti, D., Collier, R., D‟ Addezio, G., Masana, E., Sakellariou, D. (1996). Direct geological evidence for prior earthquakes on the 1981 Corinth gulf (central Greece). Geophysical Research Letters, 23(25), 3795-3798. Pantosti, D., De Martini, P.M., Koukouvelas, I., Stamatopoulos, L., Palyvos, N., Pucci, S., Lemeille, F., Pavlides, S. (2004). Palaeoseismological investigations of the Aigion fault (Gulf of Corinth, Greece). C.R. Geoscience, 336, 335-342. 127 REFERENCES Pantosti, D., De Martini, P.M., Papanastassiou, D., Palyvos, N., Lemeille, F., Stavrakakis, G. (2001). A reappraisal of the 1894 Atalanti earthquake surface ruptures, central Greece. Bulletin of Seismological Society of America, 91(4), 760-780. Papadimitriou, E.E., Karakostas, V.G. (2003). Episodic occurrence of strong (Mw>6.2) earthquakes in Thessalia area (central Greece). Earth and Planetary Science Letters, 215, 395-409. Papadimitriou, P., Chousianitis, K., Agalos, A., Moshou, A., Lagios, E., Makropoulos, K. (2012). The spatially extended 2006 April Zakynthos (Ionias islands, Greece) seismic sequence and evidence for stress transfer. Geophysical Journal International, 190, 10251040. Papadimitriou, P., Kaviris, G., Makropoulos, K. (2006). The Mw=6.3 2003 Lefkada earthquake (Greece) and induced stress transfer changes. Tectonophysics, 423, 73-82. Papadopoulos, G.A., Daskalaki, E., Fokaefs, A., Giraleas, N. (2010). Tsunami hazard in the eastern Mediterranean sea: strong earthquakes and tsunamis in the west Hellenic arc and trench system. Journal of Earthquake and Tsunami, 4(3), 145-179. Papadopoulos, G.A., Fokaefs, A., Orfanogiannaki, K. (2004). Off-fault aftershock activity: the case of the 14 August 2003 Lefkada earthquake (Mw 6.2). 5th International Symposium on Eastern Mediterranean Geology, April 14-20, 2004, Thessaloniki, Greece, Ref: T5-27. Papadopoulos, G.A., Karastathis, V., Kontoes, C., Charalampakis, M., Fokaefs, A., Papoutsis, I. (2010). Crustal deformation associates with east Mediterranean strike slip earthquakes: The 8 June 2008 Movri (NW Peloponnese), Greece, earthquake (Mw 6.4). Tectonophysics, 492, 201-212. Papadopoulos, G.A., Karastathis, V.K., Ganas, A., Pavlides, S., Fokaefs, A., Orfanogiannaki, K. (2003). The Lefkada, Ionian sea (Greece), shock (Mw 6.2) of 14 August 2003: evidence for the characteristic earthquake from seismicity and ground failures. Earth Planets Space, 55, 713-718. Papanastassiou, D. (1999). Seismic hazard assessment in the area of Mystras-Sparta, south Peloponnesus, Greece, based on local seismotectonic, seismic, geologic information and on different models of rupture propagation. Natural Hazards, 18, 237-251. Papanastassiou, D., Gaki-Papanastassiou, K., Maroukian, H. (2004). Recognition of past earthquakes along the Sparta fault, Peloponnesus, south Greece. 5th International Symposium on Eastern Mediterranean Geology, 14-20 April 2004, Ref: T5-30, Thessaloniki, Greece. 128 REFERENCES Papanastassiou, D., Gaki-Papanastassiou, K., Maroukian, H. (2005). Recognition of past earthquakes along the Sparta fault (Peloponnesus, southern Greece) during the Holocene, by combining results of different dating techniques. Journal of Geodynamics, 40, 189-199. Papanastassiou, D., Stavrakakis, G., Makris, G., Makaris, D. (2001). Recent micro-earthquake activity at northern Evoikos gulf, central Greece. Bulletin of Geological Society, 34(4), 1567-1572, Greece. Papanikolaou, D., Fountoulis, I., Metaxas, Ch. (2007). Active faults, deformation rates and quaternary paleogeography at Kyparissiakos gulf (SW Greece) deduced from onshore and offshore data. Quaternary International, 171-172, 14-30. Papanikolaou, D., Lykousis, V., Chronis, G., Pavlakis, P. (1988). A comparative study of neotectonic basins across the Hellenic arc: the Messiniakos, Argolikos, Saronikos and southern Evoikos gulfs. Basin Research, 1, 167-176. Papanikolaou, D., Metaxas, C., Chronis, G. (2001). Neotectonic structure of the Lakonikos gulf. Bulletin of Geological Society, 34(1), 297-302, Greece. Papastamatiou, D., Mouyiaris, N. (1986). The Sophades earthquake occurred on April 30th 1954 – field observations by Yannis Papastamatiou. Geological and Geophysical Research, Special Issue, 341-362. Papathanasiou, G., Pavlides, S., Christaras, V., Pitilakis, K. (2004). Liquefaction phenomena triggered by the earthquake of August 14, 2003, in the island of Lefkada. Bulletin of Geological Society, 36, 1976-1805, Greece. Papathanasiou, G., Valkaniotis, S., Chatzipetros, A., Neofotistos, P., Sboras, S., Pavlides, S. (2008). Liquefaction-induced ground disruption triggered by the earthquake of June 8, 2008, in NW Peloponnesus, Greece. 31st General Aseembly of the European Seismological Commision, 7-12 September, 2008, Hersonissos, Crete, Greece, Short Papers, 334-341. Papathanassiou, G., Pavlides, S., Ganas, A. (2005). The 2003 Lefkada earthquake: field observations and preliminary microzonation map based on liquefaction potential index for the town of Lefkada. Engineering Geology, 82, 12-31. Papazachos, B., Kiratzi, A., Karacostas, B., Panagiotopoulos, D., Scordilis, E., Mountrakis, D.M. (1988). Surface fault traces, fault plane solution and spatial distribution of the aftershocks of the September 13, 1986 earthquake of Kalamata (southern Greece). Pure and Applied Geophysics, 126(1), 55-68. 129 REFERENCES Papazachos, B., Kiratzi, A., Papadimitriou, E. (1991). Regional focal mechanisms for earthquakes in the Aegean sea. Pure and Applied Geophysics, 136(4), 405-420. Papazachos, B., Papazachou, C. (1997). The earthquakes of Greece (English edition). Editions Ziti, 304 pp., Thessaloniki. Papazachos, B., Papazachou, C. (2003). Oi seismoi tis Ellados (The earthquakes of Greece) (in Greek). Editions Ziti, 286 pp., Thessaloniki. Papazachos, B.C. (1999). Seismicity of the Aegean and surrounding area. Tectonophysics, 178, 287-308. Papazachos, B.C., Comninakis, P.E., Karakassis, G.F., Karakostas, B.G., Papaioannou, C.A., Papazachos, C.B., Scordilis, E.M. (2000). A catalogue of earthquakes in Greece and surrounding area for the period 550BC-1999. Publications Geophysical Lab, University of Thessaloniki, 1,333 pp. Papazachos, B.C., Comninakis, P.E., Scordilis, E.M. (1984). Properties of the FebruaryMarch 1981 seismic sequence in the Alkyonides gulf of central Greece. Annals of Geophysics, 2(5), 537-544. Papazachos, B.C., Comninakis, P.E., Scordilis, E.M., Karakaisis, G.F., Papazachos, C.B. (2010). A catalogue of earthquakes in the Mediterranean and surrounding area for the period 1901-2010. Publications of the Geophysics Laboratory, University of Thessaloniki. Papazachos, B.C., Hatzidimitriou, P.M., Karakaissis, G.F., Papazachos, C.B., Tsokas, G.N. (1993). Rupture zones and active crustal deformation in southern Thessalia, central Greece. Bull. Geofis. Teor. Applic., XXXV, 139, 363-374. Papazachos, B.C., Karakaisis, G.F., Chatzidimitriou, P.M. (2005). Introduction to seismology (in Greek). Ziti Publications. Papazachos, B.C., Mountrakis, D.M., Papazachos, C.B., Tranos, M.D., Karakaisis, G.F., Savvaidis, A.S. (2001). Faults that caused known strong earthquakes in Greece and surrounding areas from the 5th century B.C. until now (in Greek). 2nd Greek Conference on Earthquake Engineering & Engineering Seismology, T.E.E., Thessaloniki, 28-30 September, 1, 17-26. Papazachos, B.C., Panagiotopoulos, D.G., Tsapanos, T.M., Mountrakis, D.M., Dimopoulos, G.Ch. (1983). A study of the 1980 summer seismic sequence in the Magnesia region of central Greece. Geophysical Journal of the Royal Astronomical Society, 75, 155-168. Papazachos, B.C., Papaioannou, Ch.A., Papazachos, C.B., Savvaidis, A.S. (1997). Atlas of isoseismal maps for strong shallow earthquakes in Greece and surrounding area 130 REFERENCES (426BC-1995). University of Thessaloniki, Geophysical Laboratory, Publication No.4, 176 pp. Papazachos, B.C., Scordilis, E.M., Panagiotopoulos, D.G., Papazachos, C.B., Karakaisis, G.F. (2004). Global relations between seismic fault parameters and moment magnitude of earthquakes. Bulletin of the Geological Society of Greece, Vol.XXXVI. Proceedings of the 10th International Congress, Thessaloniki, April 2004. Papazachos, C., Papaioannou, Ch.A. (1997). The macroseismic field of the Balkan area. Journal of Seismology, 1, 181-201. Papoulia, J., Lykousis, V., Sakellariou, D. (1998). Neotectonic activity and seismic hazard in central Greece. Bolletino di Geofisica Teorica ed Applicata, Vol.39, N.2, pp. 113-124. Papoulia, J., Makris, J. (2004). Microseismicity and active deformation of Messinia, SW Greece. Journal of Seismology, 8, 439-451. Papoulia, J., Makris, J. (2010). Tectonic processes and crustal evolution on/offshore western Peloponnese derived from active and passive seismics. Bulletin of Geological Society, 43(1), 357-367, Greece. Papoulia, J., Makris, J., Drakopoulou, V. (2006). Local seismic array observations at north Evoikos, central Greece, delineate crustal deformation between the north Aegean trough and Corinthiakos rift. Tectonophysics, 423, 97-106. Papoulia, J., Stavrakakis, G., Papanikolaiou, D. (2001). Bayesian estimation of strong earthquakes in the inner Messiniakos fault zone, southern Greece, based on seismological and geological data. Journal of Seismology, 5, 233-242. Patwardhan, A.S., Kulkarni, R.B., Tocher, D. (1980). A semi-Markov model for characterizing recurrence of great earthquakes. Bulletin of the Seismological Society of America, Vol. 70, pp. 323-347. Pavlides, S. (1993). Active faulting in multi-fractured seismogenic areas; examples of Greece. Zeitschrift fuer Geomorphologie. N.F., 94, 57-72. Pavlides, S. (2003). Earthquake geology (in Greek). University Studio Press, Thessaloniki, Greece. Pavlides, S., Caputo, R. (2004). Magnitude versus faults‟ surface parameters: quantitative relationships from the Aegean region. Tectonophysics, 380, 159-188. Pavlides, S., Valkaniotis, S., Chatzipetros, A. (2007). Seismically capable faults in Greece and their use in seismic hazard assessment. 4th International Conference on Earthquake Geotechnical Engineering, paper no. 1609. Thessaloniki. 131 REFERENCES Pavlides, S.B., Papadopoulos, G.A., Ganas, A., Papathanassiou, G., Karastathis, V., Keramydas, D., Fokaefs, A. (2004). The 14 August 2003 Lefkada (Ionian sea) earthquake. 5th International Symposium on Eastern Mediterranean Geology, April 1420, 2004, Thessaloniki, Greece, Ref: T5-34. Pavlides, S.B., Valkaniots, S., Ganas, A., Keramydas, D., Sboras, S. (2004). The Atalanti active fault: re-evaluation using new geological data. Bulletin of Geological Society, 36, 1560-1567, Greece. Pechlivanidis, A. (2012). Ydrogeologikoi charactires tou ydatikou diamerismatos Attikis me emfasi sti lekani Kalyvion. Diploma thesis. School of Mining & Metallurgical Engineering, National Technical University of Athens, Greece. Perissoratis, C., Angelopoulos, I., Mitropoulos, D. (1991). Surficial sediment map of the Aegean sea floor. Pagasitikos sheet, scale: 1: 200.000, Ed. IGME, Athens. Perissoratis, C., Mettos, A., van Andel, T. (1988). Stratigraphy and structure of the south Evoikos gulf and the greater neogene basin. Bulletin of Geological Society, 23(1), 209221, Greece. Perissoratis, C., Mitropoulos, D., Angelopoulos, I. (1986). Marine geological research at the eastern Corinthiakos gulf. Geol. and Geoph. Research (IGME Publications), Special Issue, 381-401. Perissoratis, C., Papadopoulos, G. (1999). Sediment instability and slumping in the Aegean sea and the case history of the 1956 tsunami. Marine Geology, 161, 287-305. Perissoratis, C., van Andel, T.H. (1991). Sea-level changes and tectonics in the quaternary extensional basin of the south Evoikos gulf, Greece. Terra Nova, 3(3), 232-348. Perouse, E. (2013.) Cinematique et tectonique active de l‟ ouest de la Grece dans le cadre geodynamique de la Mediterranee central et orientale. These de doctorat. Universite Orsay-Paris Sud (Paris XI). Petersen, M., Dawson, T., Chen, R., Cao, T., Willis, C., Schwartz, D., Frankel, A. (2011). Fault displacement hazard for strike-slip faults. Bulletin of the Seismological Society of America, Vol.101, no.2, pp. 805-825. Piccardi, L. (2000). Active faulting at Delphi, Greece: seismotectonic remarks and a hypothesis for the geologic environment of a myth. Geology, 28, 651-654. Piper, D.J.W., Perissoratis, C. (2003). Quaternary neotectonics of the south Aegean arc. Marine Geology, 198, 259-288. 132 REFERENCES Pirli, M., Voulgaris, N., Chira, A., Makropoulos, K. (2007). The March 2004 Kalamata seismic sequence: a case of efficient seismicity monitoring in the area of Peloponnese, southern Greece, by the Tripoli seismic array. Journal of Seismology, 5, 233-242. Pitilakis, K. (2010). Geotechnical Earthquake Engineering. Editions Ziti, Thessaloniki. Pondrelli, S., Morelli, A., Ekstrom, G. (2004). European-mediterranean regional centroidmoment tensor catalog: solutions for years 2001-2002. Physics of the Earth and Planetary Interiors, 145(1-4), 127-147. Pondrelli, S., Morelli, A., Ekstrom, G., Mazza, S., Boschi, E., Dziewonski, A.M. (2002). European-mediterranean regional centroid-moment tensors: 1997-2000. Physics of the Earth and Planetary Interiors, 130, 71-101. Pondrelli, S., Salimbeni, S., Morelli, A., Ekstrom, G., Boschi, E. (2007). Europeanmediterranean regional centroid-moment tensor catalog: solutions for years 2003-2004. Physics of the Earth and Planetary Interiors, 164(1-2), 90-112. Pondrelli, S., Salimbeni, S., Morelli, A., Ekstrom, G., Postpischl, L., Vannucci, G., Boschi, E. (2011). European-mediterranean regional centroid-moment tensor catalog: solutions for years 2005-2008. Physics of the Earth and Planetary Interiors, 185, 74-81. Poscolieri, M., Lagios, E., Gregori, G.P., Paparo, G., Sakkas, V.A., Parcharidis, I., Marson, I., Soukis, K., Vassilakis, E., Angelucci, F., Vassilopoulou, S. (2006). Crustal stress and seismic activity in the Ionian archipelago as inferred by satellite and ground-based observations, Kefallinia, Greece. Journal of Geological Society, Special Publications, 261, 63-78, London. Poulimenos, G., Doutsos, T. (1996). Barriers on seismogenic faults in central Greece. Journal of Geodynamics, 22(1/2), 119-135. Poulos, S.E., Lykousis, V., Collins, M.B. (1995). Late quaternary evolution of Amvrakikos gulf, western Greece. Geo-Marine Letters, 15, 9-16. Poulos, S.E., Lykousis, V., Collins, M.B., Rohling, E.J., Pattiaratchi, C.B. (1999). Sedimentation processes in a tectonically active environment: the Kerkyra-Kefalonia submarine valley system (NE Ionian sea). Marine Geology, 160, 25-44. Psarropoulos, P. & Tsompanakis, Y. (2011). Facing complex problems of geotechnical earthquake engineering, according to EC8 (in Greek). ETAM conference. Application to Eurocode 8 for the design & evaluation of civil engineers‟ works, Athens. Reid, H.F. (1911). The elastic-rebound theory of earthquakes. Berkley, CA. University of California Press. 133 REFERENCES Reiter, L. (1990). Earthquake hazard analysis - issues and insights. Columbia University Press, New York, 254 pp. Roberts, G., Stewart, I. (1994). Uplift, deformation and fluid involvement within an active normal fault zone in the gulf of Corinth, Greece. Journal of Geological Society, 151, 531-541, London. Roberts, G.P., Ganas, A. (2000). Fault-slip directions in central and southern Greece measured from striated and corrugated fault planes: comparison with focal mechanism and geodetic data. Journal of Geophysical Research, 105(B10), 23443-23462. Roberts, S., Jackson, J. (1991). Active normal faulting in central Greece: an overview. Journal of Geological Society, Special Publications, 56, 125-142, London. Rondoyanni, T., Sakellariou, M., Baskoutas, J., Christodoulou, N. (2012). Evaluation of active faulting and earthquake secondary effects in Lefkada island, Ionian sea, Greece: an overview. Natural Hazards, 61, 843-860. Rondoyanni, Th., Sakellariou, M., Baskoutas, J., Christodoulou, N. (2012). Evaluation of active faulting and earthquake secondary effects in Lefkada island, Ionia sea, Greece: an overview. Natural Hazards, 61, 843-860. Ross, Z. (2011). Probabilistic fault displacement hazard analysis for reverse faults and surface rupture scale invariance. Master thesis. School of Civil & Environmental Engineering, California Polytechnic State University. Roumelioti, Z., Benetatos, C., Kiratzi, A. (2009). The 14 February 2008 earthquake (M 6.7) sequence offshore south Peloponnese (Greece): source models of the three strongest events. Tectonophysics, 471, 272-284. Roumelioti, Z., Kiratzi, A. (2008). The 14 October 2008 moderate magnitude sequence offshore NE Evia island (Aegean sea): fast time-domain moment tensors and shake map. Report submitted to EMSC, 6pp. Roumelioti, Z., Kiratzi, A., Benetatos, C. (2011). Time-domain moment tensors for shallow (h=40km) earthquakes in the broader Aegean sea for the years 2006 and 2007: the database of the Aristotle University of Thessaloniki. Journal of Geodynamics, 51, 179189. Ruiz, S.E. (1977). Influencia de las condiciones locales en las caracteristicas de los sismos. Instituto de Ingenieria, UNAM, 387, pp.65. Sachpazi, M., Hirn, A., Cl_ment, C., Haslinger, F., Laigle, M., Kissling, E., Charris, P., Hello, Y., L_pine, J.-C., Sapin, M., Ansorge, J. (2000). West Hellenic subduction and 134 REFERENCES Cephalonia transform: local earthquakes and plate transport and strain. Tectonophysics, 319, 301-319. Sadigh, K., Chang, C.Y., Egan, J.A., Makdisi, F., Youngs, R.R. (1997). Attenuation relations for shallow crustal earthquakes based on California strong motion data. Seismological Research Letters, Vol. 68, no.1, pp. 180-189. Sakellariou, D., Lykousis, V., Alexandri, S., Kaberi, H., Rousakis, G., Nomikou, P., Georgiou, P., Ballas, D. (2007). Faulting, seismic-stratigraphic architecture and late quaternary evolution of the gulf of Alkyonides basin-east gulf of Corinth, central Greece. Basin Research, 19, 273-295. Sanchez-Sesma, F.J. (1986). On the seismic response of alluvial valleys. Proceedings of the Summer School organized in Strasbourg, France, July 15-1 August, EMSC. Sboras, S. (2011). The Greek Database of Seismogenic Sources: seismotectonic implications for North Greece. Dottorato di ricerca in scienze della terra. Universita degli Studi di Ferrara. Sboras, S., Caputo, R., Pavlides, S., Chatzipetros, A., Papathanasiou, G., Valkaniotis, S. (2009). The Greek Database of Seismogenic faults (GreDass): state-of-the-art. GNGTS. Sboras, S., Ganas, A., Pavlides, S. (2010). Morphotectonic analysis of the neotectonic and active faults of Beotia (central Greece), using G.I.S. techniques. Bulletin of the Geological Society of Greece. Proceedings of the 12th International Congress, Patras. Scholz, C. (1990). The mechanics of earthquake and faulting. Cambridge University Press. Schwartz, D., Coppersmith, K. (1985). Fault behavior and characteristic earthquakes: examples from the Wasatch and San Andreas fault zones. Journal of Geophysical Research, Vol. 89, issue B7, pg. 5681-5698. Scordilis, E.M., Karakaisis, G.F., Karacostas, B.G., Panagiotopoulos, D.G., Comninakis, P.E., Papazachos, B.C. (1985). Evidence for transform faulting in the Ionian sea: the Cephalonia island earthquake sequence of 1983. Pure and Applied Geophysics, 123, 388-397. Segkou, M. (2010). Estimation of frequency-depended seismic hazard of Greece, using deterministic and probabilistic methodologies (in Greek). PhD thesis. Faculty of Geology & Geoenvironment, National & Kapodistrian University of Athens, Greece. Serpetsidaki, A., Sokos E., Tselentis, G.-A., Zahradnik, J. (2010). Seismic sequence near Zakynthos island, Greece, April 2006. Identification of the activated fault plane. Tectonophysics, 480, 23-32. 135 REFERENCES Silva, V., Crowley, H., Pagani, M., Monelli, D., Pinho, R (2012). Development and application of OpenQuake, an open source software for seismic risk assessment. Proceedings of the 15th World Conference on Earthquake Engineering, Lisbon, Portugal, paper n.4917. Skarlatoudis, A.A., Papazachos, B.C., Margaris, B.N., Theodulidis, N., Papaioannou, C., Kalogeras, I., Scordilis, E.M., Karakostas, V. (2003). Empirical peak ground motion predictive relations for shallow earthquakes in Greece. Bulletin of the Seismological Society of America, 93, 2591-2603. Skourtsos, E., Lekkas, S. (2011). Extensional tectonics in Mt Parnon (Peloponnesus, Greece). International Journal of Earth Sciences, 100, 1551-1567. Sokos, E. (1998). Synthesis of potential ground motions in the city of Patras with emphasis on local site conditions (in Greek). PhD Thesis, Seismological Laboratory, Department of Geology, University of Patras. Somieski, A.E. (2008). Astrogeodetic geoid and isostatic considerations in the North Aegean Sea, Greece. PhD thesis, No.17790. ETH, Zurich. Stefatos, A., Papatheodorou, G., Ferentinos, G., Leeder, M., Collier, R. (2002). Seismic reflection imaging of active offshore faults in the gulf of Corinth: their seismotectonic significance. Basin Research, 14, 487-502. Stein, S., Wysession, M. (2003). An introduction to seismology, earthquakes and earth structure. Blackwell Publishing. Stewart, I.S., Hancock, P.L. (1990). Brecciation and fracturing within neotectonic normal fault zones in the Aegean region. Journal of Geological Society, Special Publications, 54, 105-112, London. Stewart, I.S., Hancock, P.L. (1991). Scales of structural heterogeneity within neotectonic normal fault zones in the Aegean region. Journal of Structural Geology, 13(2), 191-204. Stewart, J.P., Chiou, D., Bray, J.D., Graves, R.W., Somerville, P.G., Abrahamson, N.A. (2001). Ground motion evaluation procedures for performance based design. Peer report 2001/09. Peer Center, College of Engineering, UC Berkley. Stiros, S., Chasapis, A. (1988). Geodetic constrains to the focal mechanism of the 1980, Almyros (central Greece) seismic sequence. Bulletin of Geological Society, XXXII (4), 275-281, Greece. Stiros, S.C., Arnold, M., Pirazzoli, P.A., Laborel, J., Laborel, F., Papageorgiou, S. (1992). Historical coseismic uplift on Euboea island, Greece. Earth and Planetary Science Letters, 108, 109-117. 136 REFERENCES Stiros, S.C., Psimoulis, P., Pitharouli, S. (2007). Geodetic constraints to the kinematics of the Kaparelli fault, reactivated during the 1981, gulf of Corinth earthquakes. Tectonophysics, 440, 105-119. Suckale, J., Grunthal, G., Regnier, M., Bosse, C. (2005). Probabilistic seismic hazard assessment for Vanuatu. Scientific technical report, ISSN 1610-0956. Toro, G.R., Abrahamson, N.A., Schneider, J.F. (1997). Model of strong ground motions from earthquakes in central and eastern north America: best estimates and uncertainty. Seismological Research Letters, Vol. 68, no.1, pp. 41-57. Tsapanos, T.M., Mantyniemi, P., Kijko, A. (2004). A probabilistic seismic hazard assessment in Greece and the surrounding region including site-specific considerations. Annals of Geophysics, 47(6), 1675-1688. Tselentis, A. (1997). Modern Seismology (in Greek). Editions Papasotiriou. Tselentis, G.-A., Danciu, L. (2010). Probabilistic seismic hazard assessment in Greece-part 1: engineering ground motion parameters. Natural Hazards and Earth System Sciences, 10, 1-15. Tselentis, G.-A., Makropoulos, K., Voulgaris, N. (1989). Cluster and spectral characteristics of the aftershock activity of the Kalamata, September 13, 1986 earthquake, south Greece. Tectonophysics, 169, 135-148. Tselentis, G.-A., Sokos, E., Martakis, N., Serpetsidaki, A. (2006). Seismicity and seismotectonics in Epirus, western Greece: results from a microearthquake survey. Bulletin of Seismological Society of America, 96, 5, 1706-1717. Tsodoulos, I.M., Koukouvelas, I.K., Pavlides, S. (2008). Tectonic geomorphology of the easternmost extension of the Gulf of Corinth (Beotia, central Greece). Tectonophysics, 453, 211-232. Tsokas, G.N., Hansen, R.O. (1997). Study of the crustal thickness and the subducting lithosphere in Greece from gravity data. Journal of Geophysical Research, 102:2058520597. Tsompanakis, Y., Lagaros, N.D., Stavroulakis, G.E. (2008). Soft computing techniques in parameter identification and probabilistic seismic analysis of structures. Advances in Engineering Software, 39, 612-624. Tsompanakis, Y., Psarropoulos, P.N. (2012). Dynamic soil-structure interaction: reality versus seismic norms. CST2012 & ECT2012 Lectures on Computational Methods for Engineering Science, Dubrovnik, Croatia, 4-7 September 2012. 137 REFERENCES Valkaniotis, S. (2009). Correlation between neotectonics structures and seismicity in the broader area of gulf of Corinth (central Greece). Unpublished PhD thesis (in Greek), Aristotle University of Thessaloniki, 241 pp. Vamvakaris, D. (2010). Contribution to the study of time-changing seismicity and seismic hazard (in Greek). PhD thesis. Faculty of Geology, Aristotle University of Thessaloniki, Greece. Van Andel, T.H., Perissoratis, C., Rondoyanni, T. (1993). Quaternary tectonics of the Argolikos gulf and adjacent basins, Greece. Journal of Geological Society, 150, 529539, London. Vannucci, G., Gasperini, P. (2003). A database of revised fault plane solutions for Italy and surrounding regions. Computer and Geoscienses, 29, 903-909. Vannucci, G., Gasperini, P. (2004). The new release of the database of earthquake mechanisms of the Mediterranean area (EMMA Version 2). Annals of Geophysics, supplement to Vol.47, 307-334. Votsi, I., Limnios, N., Tsaklidis, G., Papadimitriou, E. (2010). Semi-Markov models for seismic hazard assessment in certain areas of Greece. Bulletin of the Geological Society of Greece, Proceedings of the 12th International Congress, Patras, May 2010. Wells, D.L., Coppersmith, J.K. (1994). New empirical relationships among magnitude, rupture, length, rupture width, rupture area and surface displacement. Bulletin of Seismological Society of America, 84, 974-1002. Youngs, R.R., Coppersmith, K.J. (1985). Implications of fault slip rates and earthquake recurrence models for probabilistic seismic hazard estimates. Bulletin of the Seismological Society of America, Vol. 75, no.4, pp. 939-964. Zahradnik, J., Serpetsidaki, A., Sokos, E., Tselentis, G.-A. (2005). Iterative deconvolution of regional waveforms and a double-event interpretation of the 2003 Lefkada earthquake, Greece. Bulletin of Seismological Society of America, 95(1), 159-172. Zahradnik, J., Sokos, E., Tselentis, G.-A., Martakis, N. (2008). Non-double mechanism of moderate earthquakes near Zakynthos, Greece, April 2006; explanation in terms of complexity. Geophysical Prospecting, 56, 341-356. Zelilidis, A., Kontopoulos, N. (1999). Plio-pleistocene alluvial architecture in marginal extensional narrow sub-basins: examples from southwest Greece. Geological Magazine, 136(3), 241-262. Zhao, J.X., Zhang, J., Asano, A., Oyno, Y., Oouchi, T., Takahashi, T., Ogawa, H., Irikura, K., Thio, A.K., Somerville, P.G., Fukushima, Y., Fukushima, Y. (2006). Attenuation 138 REFERENCES relations of strong ground motion in Japan using site classification based on predominant period. Bulletin of the Seismological Society of America, 96:898-913. Zovoli, E., Konstantinidi, E., Koukouvelas, I.K. (2004). Tectonic geomorphology of escarpments: the cases of Kompotades and Nea Anchialos faults. Bulletin of Geological Society, XXXVI, 1716-1725, Greece. Zygouri, V., Verroios, S., Kokkalas, S., Xypolias, P., Koukouvelas, I.K. (2008). Scaling properties within the Gulf of Corinth, Greece; comparison between offshore & onshore active faults. Tectonophysics, 453, 193-210. 139