Search

Understanding of lithosphere dynamics, tectonic stress localization, earthquake occurrences, and seismic hazards has significantly advanced during the last decades. Meanwhile, despite the major advancements in geophysical sciences, yet we do not see a decline in earthquake disaster impacts and losses. Although earthquake disasters are mainly associated with significant vulnerability of society, comprehensive seismic hazards assessments and earthquake forecasting could contribute to preventive measures aimed to reduce impacts of earthquakes. Modeling of lithosphere dynamics and earthquake simulations coupled with a seismic hazard analysis can provide a better assessment of potential ground shaking due to earthquakes. Here we present a block-and-fault dynamics (BAFD) model, which simulates earthquakes due to lithosphere dynamics and allows for studying the influence of fault network properties and regional movements on seismic patterns. The model's performance is analyzed in terms of reproduction of basic features of the observed seismicity such as the frequency-magnitude relationship, clustering of earthquakes, occurrences of large events, fault slip rates, and earthquake mechanisms. Several studies related to the application of the BAFD model to the following earthquake-prone regions are reviewed: the southeastern Carpathians, Caucasus, the western India, Tibet-Himalaya, and the Sunda Arc. We examine then a new approach to seismic hazard analysis, which is based on instrumentally recorded, historical and BAFD-simulated earthquakes, and analyze how earthquake modeling can assist in hazard assessment. Finally, we discuss perspectives in modeling of earthquake occurrences and improvements in hazard assessment.

On 22 November, 2014, an MS 6.3 earthquake occurred in Kangding County, China. Focal mechanism solution shows that the two nodal planes were 235°/82°/− 173° and 144°/83°/− 8° and the focal depth was 9 km. Seismic slip of the Kangding earthquake was bilateral with about 0.5 m maximum slip. The rupture zone was confined to depths ranging from 5 to 15 km and laterally extended along the slip and strike directions by about 10 and 12 km, respectively. Most of the seismic moment was released in the first 5 s of the rupture, resulting in an earthquake magnitude of MW 6.01. In contrast, a slip model obtained by interferometric synthetic aperture radar (InSAR) data indicates that the rupture zone was longer than that determined from the seismic data and the earthquake magnitude should be about MW 6.2. Although accounting for the contribution of the MS 5.8 aftershock and the other small aftershocks that occurred during the InSAR observations period, the total moment estimated based on the seismic slip model was significantly smaller than that obtained from the InSAR data. Based on our analysis, we found that the inconsistency between the results determined from the seismic data and the InSAR data may be caused by the decrease in the shear modulus at shallow depths, the noise in the InSAR data, and the occurrence of some afterslips in the northwest region of the fault zone. The seismic slip of this earthquake was too small to release the accumulated energy within the entire Xianshuihe fault. We also found that the Coulomb stress in the northwest zone of the Kangding–Daofu seismic gap increased as a result of the historical, 2008 MS 8.0 Wenchuan and the 2014 MS 6.3 Kangding earthquakes, suggesting that this area is expected to be a high seismic hazard region for the future.

Strong ground shaking during the 2008 Mw 7.9 Wenchuan earthquake, which occurred in the foothills of the Longmen Shan in southwestern China, resulted in widespread coseismic deformation features, such as liquefaction and water ejection. We present a systematic survey of the soil liquefaction and water ejection features caused by this major earthquake. The majority of liquefaction sites occurred along major alluvial fan-building rivers, where the water table was a few meters below the surface. While there is no clear correlation between water fountain height and peak ground acceleration, ~ 58% of liquefaction sites are located 20–35 km from the Beichuan fault. Clusters of sites with anomalously high (> 2 m) water ejections are located near the surface projection of the Range Front blind thrust and its splay faults. The density of anomalously high eruptions within the window surrounding the Range Front blind thrust is 3–6 times greater than outside this region. Our results suggest that geologic structures may play a role in augmenting liquefaction intensity and shaking-related seismic hazards in sedimentary basins. We speculate that the mechanism could be the amplification of shaking by fault zone structures. Alternatively, faults may act as pathways with increased vertical permeability, allowing fluids from deeper confined aquifers to migrate to and enhance liquefaction of the upper soil, as proposed previously by Wang (2007). Liquefaction associated with the Wenchuan earthquake thus demonstrates the importance of considering geologic structures other than the primary earthquake-producing fault in seismic hazard evaluation and earthquake resistance design in areas with similar geologic and hydrological settings.

Seismic shaking and tectonic deformation during strong earthquakes can trigger widespread environmental effects. The severity and extent of a given effect relates to the characteristics of the causative earthquake and the intrinsic properties of the affected media. Documentation of earthquake environmental effects in well-instrumented, historical earthquakes can enable seismologic triggering thresholds to be estimated across a spectrum of geologic, topographic and hydrologic site conditions, and implemented into seismic hazard assessments, geotechnical engineering designs, palaeoseismic interpretations, and forecasts of the impacts of future earthquakes. The 2010–2011 Canterbury Earthquake Sequence (CES), including the moment magnitude (M w ) 7.1 Darfield earthquake and M w 6.2, 6.0, 5.9, and 5.8 aftershocks, occurred on a suite of previously unidentified, primarily blind, active faults in the eastern South Island of New Zealand. The CES is one of Earth's best recorded historical earthquake sequences. The location of the CES proximal to and beneath a major urban centre enabled rapid and detailed collection of vast amounts of field, geospatial, geotechnical, hydrologic, biologic, and seismologic data, and allowed incremental and cumulative environmental responses to seismic forcing to be documented throughout a protracted earthquake sequence. The CES caused multiple instances of tectonic surface deformation (≥ 3 events), surface manifestations of liquefaction (≥ 11 events), lateral spreading (≥ 6 events), rockfall (≥ 6 events), cliff collapse (≥ 3 events), subsidence (≥ 4 events), and hydrological (10s of events) and biological shifts (≥ 3 events). The terrestrial area affected by strong shaking (e.g. peak ground acceleration (PGA) ≥ 0.1–0.3 g), and the maximum distances between earthquake rupture and environmental response (R rup ), both generally increased with increased earthquake M w , but were also influenced by earthquake location and source characteristics. However, the severity of a given environmental response at any given site related predominantly to ground shaking characteristics (PGA, peak ground velocities) and site conditions (water table depth, soil type, geomorphic and topographic setting) rather than earthquake M w . In most cases, the most severe liquefaction, rockfall, cliff collapse, subsidence, flooding, tree damage, and biologic habitat changes were triggered by proximal, moderate magnitude (M w ≤ 6.2) earthquakes on blind faults. CES environmental effects will be incompletely preserved in the geologic record and variably diagnostic of spatial and temporal earthquake clustering. Liquefaction feeder dikes in areas of severe and recurrent liquefaction will provide the best preserved and potentially most diagnostic CES features. Rockfall talus deposits and boulders will be well preserved and potentially diagnostic of the strong intensity of CES shaking, but challenging to decipher in terms of single versus multiple events. Most other phenomena will be transient (e.g., distal groundwater responses), not uniquely diagnostic of earthquakes (e.g., flooding), or more ambiguous (e.g. biologic changes). Preliminary palaeoseismic investigations in the CES region indicate recurrence of liquefaction in susceptible sediments of ~ 100 to 300 yr, recurrence of severe rockfall event(s) of ca. 6000 to 8000 yr, and recurrence of surface rupturing on the largest CES source fault of ca. 20,000 to 30,000 yr. These data highlight the importance of utilising multiple proxy datasets in palaeoearthquake studies. The severity of environmental effects triggered during the strongest CES earthquakes was as great as or equivalent to any historic or prehistoric effects recorded in the geologic record. We suggest that the shaking caused by rupture of local blind faults in the CES comprised a ‘worst case’ seismic shaking scenario for parts of the Christchurch urban area. Moderate M w blind fault earthquakes may contribute the highest proportion of seismic hazard, be the most important drivers of landscape evolution, and dominate the palaeoseismic record in some locations on Earth, including locations distal from any identified active faults. A high scientific priority should be placed on improving the spatial extent and quality of ‘off-fault’ shaking records of strong earthquakes, particularly near major urban centres.

We present a new approach to assessment of regional seismic hazard, which accounts for observed (instrumentally recorded and historic) earthquakes, as well as for seismic events simulated for a significantly longer period of time than that of observations. We apply this approach to probabilistic seismic hazard analysis (PSHA) for the Tibet–Himalayan region. The large magnitude synthetic events, which are consistent with the geophysical and geodetic data, together with the observed earthquakes are employed for the Monte-Carlo PSHA. Earthquake scenarios for hazard assessment are generated stochastically to sample the magnitude and spatial distribution of seismicity, as well as the distribution of ground motion for each seismic event. The peak ground acceleration values, which are estimated for the return period of 475 yr, show that the hazard level associated with large events in the Tibet–Himalayan region significantly increases if the long record of simulated seismicity is considered in the PSHA. The magnitude and the source location of the 2008 Wenchuan M = 7.9 earthquake are among the range of those described by the seismic source model accepted in our analysis. We analyze the relationship between the ground motion data obtained in the earthquake's epicentral area and the obtained PSHA estimations using a deaggregation technique. The proposed approach provides a better understanding of ground shaking due to possible large-magnitude events and could be useful for risk assessment, earthquake engineering purposes, and emergency planning.

The temporal and spatial variations of the velocity structure 60 days before and after the March 24, 2011 Myanmar M7.2 earthquake in the Yunnan region are studied by using the ambient noise cross-correlation method with the continuous seismic waveform records. In addition, the response ratio of strain energy release is obtained using moderate and small earthquakes, and the Coulomb stress changes along different faults are calculated using the source parameters of the Myanmar earthquake. There are three regions where the velocities increased, the area from the Zemuhe–Xiaojiang fault zone to the Yuanmou–Lvzhi fault zone, the southern segment of the Honghe fault zone, and the area from the Nantinghe fault zone northward extending to the Dayingjiang fault zone. The seismic activity and the Coulomb stress of these regions are also enhanced. By contrast, there are several regions where velocities decreased, such as the area from the Nantinghe fault zone southward extending to the Simao region. The seismic activity and the Coulomb stress of these regions are decreased as well. The consistency of the variation indicates some relationship between the variation and the Myanmar earthquake.

Starting from late May 2012, the Emilia region (Northern Italy) was severely shaken by an intense seismic sequence, originated from a M L 5.9 earthquake on May 20th, at a hypocentral depth of 6.3 km, with thrust-type focal mechanism. In the following days, the seismic rate remained high, counting 50 M L ≥ 2.0 earthquakes a day, on average. Seismicity spreads along a 30 km east–west elongated area, in the Po river alluvial plain, in the nearby of the cities Ferrara and Modena. Nine days after the first shock, another destructive thrust-type earthquake (M L 5.8) hit the area to the west, causing further damage and fatalities. Aftershocks following this second destructive event extended along the same east-westerly trend for further 20 km to the west, thus illuminating an area of about 50 km in length, on the whole. After the first shock struck, on May 20th, a dense network of temporary seismic stations, in addition to the permanent ones, was deployed in the meizoseismal area, leading to a sensible improvement of the earthquake monitoring capability there. A combined dataset, including three-component seismic waveforms recorded by both permanent and temporary stations, has been analyzed in order to obtain an appropriate 1-D velocity model for earthquake location in the study area. Here we describe the main seismological characteristics of this seismic sequence and, relying on refined earthquakes location, we make inferences on the geometry of the thrust system responsible for the two strongest shocks.

For gaining insight into the preparatory process of the 1999 Chi-Chi, Taiwan earthquake and its related statistical feature, the Epidemic-Type Aftershock-Sequences model (ETAS model) was applied to the earthquake data in Taiwan region. By means of the ETAS analysis for Taiwanese earthquakes larger than a magnitude of 2.4, we found seismic quiescence over broader regions of Taiwan. It was also found that inland areas near the epicenter of the Chi-Chi earthquake showed seismic activation in the period from Jan. 1, 1998 to Sep. 20, 1999, which is just before the occurrence of the Chi-Chi earthquake. The assumption that this is due to precursory slip (stress drop) on the fault plane of the Chi-Chi earthquake is supported by previous researches such as a numerical simulation using rate- and state-dependent friction laws and the observation of abnormal change in crustal displacement for a station of Taiwan GPS network near southern edge of the source area of the Chi-Chi earthquake.

Stein et al. (2012) contends that the occurrence of some recent destructive earthquakes in areas mapped as low hazard in some probabilistic seismic hazard maps “illustrates limitations of earthquake hazard mapping.” This statement is fundamentally incorrect. These instances illustrate that short-term observations such as recent seismicity do not characterize the hazard from faults with long recurrence-time earthquakes. This is not an inherent limitation in the methodology of probabilistic seismic hazard assessment (PSHA). They repeat arguments made in their previous articles, one of them recent (Stein et al., 2011) and others on the New Madrid region of the central United States that were published and commented on several years ago (e.g., Frankel, 2003, 2004; Stein et al., 2003). Stein et al. (2012) makes three major points: 1) the inputs to seismic hazard maps should include long-term geologic data in addition to historical and instrumental seismicity, 2) uncertainties in hazard values should be estimated and communicated to users, and 3) seismic hazard maps need to be objectively tested. The first two points have long been recognized and practiced in the seismic hazard mapping community in the United States and elsewhere. The third point has also been the subject of substantial work (e.g., Stirling and Petersen, 2006). This comment points out problems

article i nfo Article history: We estimated the source rupture of the 2008 Wenchuan Earthquake (Ms 8), China, based on a back- projection of seismic waves to their source plane, using data from regional broadband arrays in Taiwan and northern Vietnam. Observations from these arrays, located at different azimuths, were processed to eval- uate the spatio-temporal rupture behavior of the fault. The seismic energy spot converted from windowed array waveforms was back-projected to the rupture plane of the earthquake, to image the instantaneous slip on the fault plane. Rupture processes were reconstructed based on the imaged time-dependent seismic energy radiating from the earthquake fault plane. The high station density of both arrays enabled a detailed examination of the rupture processes. The results indicate that the 2008 Wenchuan Earthquake had a rupture duration of approximately 100 s, with major asperities radiating energy at 50 s after the initiation of rupture. The location of asperities is in good agreement with sections of the earthquake fault along which major dam- age occurred, such as along the Yingxiu-Beichuan Fault. The radiated seismic energy shows a complex spatial distribution on the fault plane. Our analysis indicates that the rupture initiated at the epicenter and extended to the northeast for approximately 280 km. The average rupture velocity of this earthquake was approxi- mately 2.8 km/s.

The 2001 Bhuj earthquake took place in a region away from the active plate boundaries, which qualifies this earthquake to be a typical intraplate event that claimed a death toll of 20,000 people. The aftershock sequence of this earthquake is still continuing for the last eleven years, which makes this aftershock sequence as a unique intraplate earthquake sequence in the world. This sequence is being monitored by a close digital seismic network consisting of 5–12 three-component broadband seismographs and 10–20 accelerographs that led to a homogeneous earthquake catalog consisting of precise and accurate estimates of hypocentral and source parameters. This catalog enabled us to examine the evolution of this earthquake sequence. Our study reveals a few interconnected regularities in the rate of earthquake occurrence, b-values, fractal dimensions, stress drops, and the mean earthquake depth values. We observe that a slow decrease in background seismic activity and a number of bursts in seismic activity characterize the 2001 Bhuj earthquake sequence. We also notice that the background seismic activity is associated with a monotonic decrease in the mean earthquake depth values while the bursts in seismic activity are found to be associated with a decrease in fractal dimension, b-value and mean earthquake depth estimates, and an increase in stress drop values. We propose that these bursts in seismic activity are related to the episodes of break in deep fluid circulation in the direction of the Earth's surface. The effective background permeability of the mid-crust in the 2001 Bhuj earthquake region is evaluated to be about k ≈ 10 − 13 m 2 based on the estimates of changes in mean earthquake depth. Thus, we infer that seismogenesis of the 2001 Bhuj earthquake sequence is probably connected with active deep fluid circulation underlying the Kachchh seismic zone.

Satellite images and aerial-photography following the Wenchuan earthquake show that a broad area experienced severe landsliding in association with the main earthquake and aftershocks. A zone of high density landsliding (> 10% landslide surface area affected) stretches for ~ 240 km along the earthquake region, in proximity to the Yingxiu, Beichuan and Pennguan faults, which ruptured during the earthquake. The width of this severe landsliding zone varies along the strike, from ~ 25–30 km wide in the southwestern section of the earthquake region, to ~ 3–5 in the northeastern section. These variations coincide with differences in hillslope gradients, bedrock lithologies, co-seismic slip and ground shaking observed in these sections of the earthquake region. In general, there is good correlation between the magnitude and distribution of ground shaking experienced during the earthquake and mapped landslide density, though limited data and varying fault models prevent any in-depth comparison with high-resolution maps of ground shaking experienced. Landslides associated with Wenchuan earthquake highlight interesting aspects of erosion and tectonic evolution of the Longmen Shan. Short-term erosion rates in the Longmen Shan region measured from cosmogenic nuclides prior to the earthquake were ~ 0.2–0.3 mm/yr, reflecting the last 2000–3000 years — a time period shorter than the estimated recurrence interval of large earthquakes along the fault that ruptured, while long-term erosion rates measured by low-temperature thermochronology were higher, 0.5–0.7 mm/yr for the last 8–10 Ma. Integrating the total amount of erosion associated the earthquake into the short-term erosion rate estimates pushes the rate much higher. This suggests that the earthquake and associated landslide erosion was necessary for the landscape to ‘catch-up’ and balance the long-term tectonic growth of the Longmen Shan, particularly in the southwestern section of the earthquake region. Taken together, the earthquake and associated landslide erosion can be seen as the most recent expression of the rapid deformation/uplift and erosion that previous geomorphic evidence and geodynamic models had argued for along Longmen Shan front.

This paper presents the case history of an earthquake prediction, which was prepared by seismologists at the Far East Branch of the Russian Academy of Sciences and submitted to the Russian Ministry of Emergency Situations later on. The prediction, described here briefly, was confirmed with the occurrence of the 2 August 2007 M W 6.2 Nevelsk earthquake. The first symptoms of the large seismic event were recognized as early as 1997, within the so-called “seismic gap of the first kind” (Mogi, 1985). This is the area where large earthquakes are possible but have been absent for at least 100 years, as outlined from historical data along the western coasts of the Sakhalin, Hokkaido and Honshu Islands. The symptoms were related to the incipient “seismic gap of the second kind,” where shallow earthquakes with М ≥ 3.0 had disappeared. In December 2005, a long-term prediction of an earthquake with МS = 6.6 ± 0.6 was made when a “seismic gap of the second kind” (Mogi, 1985) became evident since the middle of 2003 in an area of 90 by 60 km. This prediction was to a large extent possible due to the local autonomous digital seismic network set up in 2001 in the southern region of Sakhalin Island. The prediction was accompanied by various anomalous phenomena in advance of the actual predicted event of 2 August 2007 that shocked the city of Nevelsk.

article A survey of published scientific literature was undertaken to identify and catalog observed earthquake pre- cursors. The earthquake precursors selected for analysis included electric and magnetic fields, gas emissions, groundwater level changes, temperature changes, surface deformations, and seismicity. For each of these precursors, the published scientific literature was searched to document the statistics of each reported earth- quake precursor (spatial extent, time, duration, amplitude, signal/noise ratio), to analyze dependence of the observable for each precursor on earthquake magnitude, and to explore proposed physical models to explain eachearthquakeprecursor.Somegeneralcharacteristics wereobservedfortheseprecursoryphenomena.First, the largest amplitude precursory anomalies tend to occur before the largest magnitude earthquakes. Also, the number of precursory anomalies tends to increase the closer in time to the occurrence of the earthquake. Finally,theprecursoryanomaliestendtooccurclosetotheeventualepicenterof theearthquake.Ingeneral,the physical models indicate that all of the precursory phenomena are related to deformation that occurs near the fault prior to the main earthquake. While the models provide plausible physical explanations for the pre- cursors, there are many free parameters in the models that are poorly resolved.

We compile a comprehensive list of hypocentral locations and source parameters for earthquakes occurred in the Taiwan region. The collected database is used to construct 3D patterns of seismic deformation that are associated with various tectonic processes such as subduction and collision between the Philippine Sea plate and Eurasia. Specifically, we divide the entire Taiwan region into a 3D grid of 10km interval and the amount of seismic moment released at each point is the sum of all individual events in the vicinity. In case when the physical dimension of an earthquake is larger than 10km, the corresponding seismic moment is distributed across multiple grid points. This quantitative approach gives a better depiction on many first-order features in the region. The most intriguing one is that a significant deficit in the total amount of released seismic moment can be clearly identified around the Chi-Chi source region before the big earthquake occurred in 1999. The deficit trough is then filled by the seismic moment of the Chi-Chi earthquake sequence. This deficit-then-fill pattern in a region's seismic moment distribution can be used as an indicator to pinpoint the locations of large earthquakes in the foreseeable future. Following this argument, we suggest that disastrous earthquakes with magnitudes comparable to that of the Chi-Chi earthquake are due for the two regions adjacent to the Chi-Chi source area (i.e., the Miaoli-Hsinchu domain to the north and the Chiayi domain to the south).

We describe a success story at the junction between South-Eastern Alps and external Dinarides that has led to an early deployment of GPS stations prior to the predicted July 12th 2004 moderate size Slovenia Krn Mountain earthquake. The success story consisted in a straightforward integration between a long-lasting lithosphere-scale rock mechanics experiment, along with GPS monitoring, leading to a physical model of stress evolution and tested earthquake prediction experiment using M8S, CN and RTP algorithms to point out the area of the impending earthquake. Within the alarmed area by the prediction algorithms, the lithosphere-scale rock mechanics experiment revealed that the location of the 2004 event falls within an area of stress shadow due to the recent 1998 Bovec earthquake, but is also very close to an area of increased stress due to the long-lasting effect of the 1511 event. The pre and post 2004 earthquake GPS data provided the following results: 1- the Krn Mountain earthquake magnitude has to be increased from MW 5.2 to 5.5, therefore doubling the fault slip in order to provide a better fit to the near-field displacements. Accordingly the RTP 2004 alarm in Northern Dinarides can be considered a successful prediction now that the magnitude is inside the prediction range. 2- the existence of an important amount of aseismic deformation related to such a moderate size earthquake and the feasibility of monitoring these transients; 3- the evidence of a resolved acceleration of the strain rates one year prior to the earthquake; 4- the robustness of the Bayesian approach in detecting discontinuities in the times series, their magnitude and statistical significance. The discontinuities or jumps in the time series can correspond to coseismic deformation or time-dependent deformation such as creeping, slow motion, strain acceleration and transients in general; 5- when integrated with tested earthquake prediction algorithms, the capability to forecast earthquakes can be extended to the scale of the active fault systems.

We in this study have calculated the standard normal deviate Z-value to investigate the variations in seismicity patterns in the Taiwan region before and after the Chi-Chi earthquake. We have found that the areas with relatively high seismicity in the eastern Taiwan became abnormally quiet before the Chi-Chi earthquake while the area in the central Taiwan with relatively low seismicity showed unusually active. Such a spatially changing pattern in seismicity strikingly demonstrates the phenomenon of “seismic reversal,” and we here thus present a complete, representative cycle of “seismic reversal” embedding in the changes of seismicity patterns before and after the Chi-Chi earthquake.

We introduce an empirical method of near real-time, near-field magnitude determination for large (M>6.5) crustal earthquakes. Time integration over the strong shaking duration on the absolute values of the acceleration records is carried out for nearby stations surrounding many large earthquake sources in Taiwan. The integrated quantity, here denoted as total effective shaking, is used in a regression process to derive an empirical relationship for a quick Mw determination useful for a reliable real-time operation in earthquake rapid reporting and earthquake early warning systems (EWSs).

Generally the seismic hazard of an area of interest is considered independent of time. However, its seismic risk or vulnerability, respectively, increases with the population and developing state of economy of the area. Therefore, many areas of moderate seismic hazard gain increasing importance with respect to seismic hazard and risk analysis. However, these areas mostly have a weak earthquake database, i.e., they are characterised by relative low seismicity and uncertain information concerning historical earthquakes. In a case study for Eastern Thuringia (Germany), acting as example for similar places in the world, seismic hazard is estimated using the probabilistic approach. Because of the lack of earthquakes occurring in the recent past, mainly historical earthquakes have to be used. But for these the actual earthquake sources or active faults, needed for the analysis, are imprecisely known. Therefore, the earthquake locations are represented by areal sources, a common practice. The definition of these sources is performed carefully, because their geometrical shape and size (apart from the earthquake occurrence model) influence the results significantly. Using analysis tools such as density maps of earthquake epicentres, seismic strain and energy release support this. Oversizing of areal sources leads to underestimation of seismic hazard and should therefore be avoided. Large location errors of historical earthquakes on the other hand are represented by several alternative areal sources with final superimposition of the different results. In a very similar way information known from macroseismic observations interpreted as source rather than as site effects are taken into account in order to achieve a seismic hazard assessment as realistic as possible. In very local cases the meaning of source effects exceeds those of site effects very likely. The influence of attenuation parameter variations on the result of estimated local seismic hazard is relatively low. Generally, the results obtained by the seismic hazard assessment coincide well with macroseismic observations from the thoroughly investigated largest earthquake in the region.

The Zhangjiakou–Penglai seismotectonic zone (ZPSZ) lies in the northern part of North China and extends along the Zhangjiakou–Beijing–Tianjin–Bohai Bay–Penglai–Yellow Sea. It is about 900 km long and some 250 km wide in a northwest direction. The great Sanhe-Pinggu (MS=8.0) earthquake occurred on September 1679 and the Tangshan (MS=7.8) earthquake on July 1976 caused serious economic and life losses. According to some differences in crust structure and regional tectonic stress field, the ZPSZ is divided into western and eastern segment by the 117°E line for study on long-term seismic hazard analysis. An analysis of Gutenberg–Richter's empirical relation of earthquake-frequency and time process of historic and recent earthquakes along the eastern and western segments shows that the earthquake activity obeys a Poisson process, and these calculations indicate that the earthquake occurrence probability of MS=6.0–6.9 is 0.77–0.83 in the eastern segment and the earthquake occurrence probability of MS=7.0–7.9 is 0.78–0.80 in the western segment of the ZPSZ during a period from 2005 to 2015.

A probabilistic estimate of seismic hazard can be obtained from the spatial distribution, of earthquake sources, their frequency–magnitude distribution and the rate of attenuation of strong ground motion with distance. We calculate the earthquake perceptibility, i.e. the annual probability that a particular level of ground shaking will be generated by earthquakes of particular magnitude, by weighting frequency–magnitude data with the predicted felt area for a given level of ground shaking at a particular magnitude. This provides an earthquake selection criterion that can be used in the anti-seismic design of non-critical structures. We calculate the perceptibility, at a particular value of isoseismal intensity, peak ground acceleration and velocity, as a function of source magnitude and frequency for the broad Aegean area using local attenuation laws. We use frequency–magnitude distributions that were previously obtained by combining short-term catalogue data with tectonic moment rate data for 14 tectonic zones in Greece with sufficient earthquake data, and where contemporary strain rates are available from satellite data. Many of the zones show a ‘characteristic earthquake’ distribution with the most perceptible earthquake equal to the maximum magnitude earthquake, but a relatively flat perceptibility between magnitudes 6 and 7. The maximum perceptible magnitude is in the fastest-deforming region in the middle of the Aegean sea, and tends to be systematically low on the west in comparison to the east of the Aegean sea. The tectonic data strongly constrain the long-term recurrence rates and lead to low error estimates (±0.2) in the most perceptible magnitudes.

High-resolution seismic surveys, including P- and S-wave studies, have been conducted in an area of the Ottawa River valley located 80 km east of Ottawa (Canada). Based on dating of paleolandslides, the existence of paleoearthquake activity has been postulated in this area. The target zone for the seismic survey is characterized by surface disturbance and sediment deformation. P-wave seismic imaging was used to map the overburden–bedrock interface as well as to indicate reflecting boundaries within the overburden. The area of surface disturbance was found to overlie a buried bedrock basin, 8 km in diameter, infilled with a maximum thickness of 180 m of unconsolidated Quaternary sediments. Preliminary results of core logging show the presence of sand overlain by deformed fine sediments within the disturbed area. Shear-refraction studies reveal differences in the velocity–depth profiles between the disturbed area and the surrounding undisturbed areas. The shear-wave reflection method was used to produce a fundamental resonant period map for the area. Surface sediment disturbance was probably due to a combination of ground-motion amplification due to the basin (thick soft sediments) and the presence of water-saturated sand at depth.

The modal summation method to compute synthetic seismograms is applied to carry out a study of seismic zonation in Albania. From a seismotectonic synthesis and the available earthquake data, eight seismogenic zones are defined and for each of them, a characteristic focal mechanism is selected. A new seismic velocity model is used in the computations and is constructed on the basis of previous studies. Correlation relations between the maximum intensity values observed and the ground motion parameters obtained from this procedure are derived. The distribution of the calculated values of displacement, velocity and acceleration, will contribute to the quality of anti-seismic design practice in Albania and to the mitigation of earthquake consequences.

A recurrent criticism of the deterministic approaches for Seismic Hazard Analysis (SHA) is that they make use of a controlling earthquake (usually the maximum to have occurred within a given time and space domain) instead of the frequency of earthquake occurrence. We propose a deterministic method for SHA, in which, instead of a fixed controlling earthquake, the earthquake frequency is used. The method is based on the deterministic procedure, developed by Costa et al. [Costa et al., 1993. J. Appl. Geophys. 30, 149–160]. This procedure computes synthetic seismograms over a regular grid of receivers. The input for these estimations consists of available earthquake sources and structural models. The earthquake sources are modelled as follows: the seismogenic zones are divided into cells of dimension 0.2°×0.2° and a point source is placed in the centre of each cell. The magnitude of the maximum observed earthquake in the cell and the available focal mechanisms are then used to assign the seismic moment and the double-couple orientation to the source. In this study, the rate of earthquake occurrence is taken into account, using the recently computed [Molchan et al., 1997. Bull. Seismol. Soc. Am. 87, 1220–1229] Gutenberg–Richter parameters in each seismogenic zone in Italy. An annual probability of exceedance and an average return period are calculated for each event of a given magnitude. The average annual number of events producing ground motion exceeding a given value, and, therefore, an annual probability of exceedance, are then calculated at each receiver. The methodology allows us to estimate both the ground motion due to large and rare events as well as that due to small and frequent earthquakes, and the obtained results can be easily compared with the probabilistic studies (if available) for standard return periods. The approach for seismic hazard analysis proposed here has two outputs: a `deterministic' (i.e. fixed ground motion due to a fixed event), and a `probabilistic' (i.e. annual average number of events, or probability of ground motion exceedance over a given time period). The method we propose overcomes one of the most recurrent criticism of the deterministic approaches for SHA, and at the same time maintains their advantages: clear and realistic estimation of engineering parameters needed in the calculation of Seismic Risk. As an example, the `deterministic–probabilistic' approach has been applied to estimate the seismic hazard for the region of eastern Sicily in Italy.

The 1932 deep Fiji earthquake is studied from the point of view of comparing it to the recent large events in the transition zone. Spectral analysis and body-wave modeling suggest a moment of 3.4 × 10 27 dyn cm (within a multiplicative or divisive factor of 1.5), which makes it comparable to, but not significantly larger than, the great 1994 earthquake, 770 km to the north. Its seismic moment remains well below that of the largest deep shocks (1994 in Bolivia or 1970 in Colombia). Relocation efforts show that the earthquake took place on the front edge of the Wadati-Benioff Zone, in an area where the latter features a complex geometry involving significant warping. The mechanism of the earthquake involves down-dip compression and can be readily explained in terms of the prevailing large-scale stresses in the slabs.

Earthquake rate analysis explores the quantitative relationship between earthquake occurrence rates and geological deformation rates. One application is the estimation of tectonic deformation rates from seismicity, for comparison with tectonic models. Another application is the use of geological models to help quantify seismic hazards. Seismicity is described by a catalog of the largest earthquakes in the region. The contribution of small earthquakes towards causing geological deformation is insignificant, although the occurrence rate of small events may be predictable from the occurrence rate of large events by a b value relationship. The critical information about each earthquake is its seismic moment, or, better still, its seismic moment tensor. When moment is not available directly, it can be estimated from the earthquake magnitude, although this increases uncertainties considerably. The average annual moment from all the earthquakes in the catalog is the seismic moment rate. Uncertainties in seismic estimates of moment rate arise from the short time period of observations, from the magnitude-moment relationship, from the shape of the distribution function giving number of earthquakes as a function of magnitude and from uncertainties in the physical properties of the Earth near the source. Geological models for deformation rates are described by slip rates on individual faults or by regional strain rates. A regional strain rate can be derived as a limiting case where there are numerous individual faults. Such a strain rate is a particularly useful concept where information about each of the faults is limited. In the case of either faulting or regional strain, the average seismic moment rate is proportional to the deformation rate. Geological uncertainties arise from difficulties in estimating slip rates or strain rates, from fault creep and other aseismic deformation, from uncertainty in the part of the fault that fails during earthquakes and boundaries of the deforming region and from uncertainty about the Earth's physical properties. The seismic moment rate is the critical factor for relating the earthquake occurrence rate and the geological deformation rate. Using this intermediate parameter, one can infer earthquake occurrence rates from deformation rates, or infer deformation rates from earthquake occurrence rates. Generally, it is found that deformation rates and earthquake occurrence rates are consistent within uncertainties. Where both are well constrained, they are generally consistent within a factor of two or three.

A method for calculation of seismic risk, earlier developed for Fennoscandia, is applied to an area comprising the whole of Turkey, using a relatively homogeneous earthquake catalogue for the period 1913–1970. The Turkish area is divided into 40 regions, for each of which the seismic risk is given for the surface-wave magnitude range from 4.5 to 7.5.

The earthquake hazard in Israel is assessed on the basis of the Cornell-McGuire probabilistic approach using the available set of instrumental and historical regional earthquake data. A horizontal peak ground acceleration probability map is presented, based on a proposed seismicity model, earthquake source zonation and the application of the Joyner-Boore attenuation relationship. The present results differ from those previously obtained owing to the technique used, the incorporation of a greater amount of instrumental data and the use of a more appropriate attenuation relationship. For the first time regional earthquake hazard analysis incorporates the active Carmel and Farah faults; this alters the risk estimate and is important considering their proximity to densely populated and industrialized areas. An approximate sensitivity analysis is carried out for main model input parameters with inherent uncertainties.

The spatial pattern of the epicentral distribution in the northwest Chubu district of Honshu Island, Japan changed drastically after the occurrence of the 1948 Fukui earthquake ( M = 7.1 ). The change can be explained by invoking that the WNW-ESE trending seismic belt which contains the focal region of the Fukui earthquake, was activated, and that the activity along another seismic belt trending NE-SW which had been conspicuous before the Fukui earthquake, disappeared. A notable feature is that large earthquakes occurred successively along the newly activated seismic belt migrating to the east. The 1984 Nagano-ken Seibu earthquake ( M = 6.8 ) occurred on the extension of that migrational activity. The focal region of the Nagano-ken Seibu earthquake is situated at a distinct inflection point in the seismic zone which was activated after the Fukui earthquake. The spatial pattern of the seismic zone is very similar to the distribution of Quaternary volcanoes in that region, and the same pattern is also recognized in the Bouguer anomaly. These features suggest that the seismic belt is a weak zone in the crust corresponding to the boundary of the tectonic blocks. The focal region of the Nagano-ken Seibu earthquake had been very quiet over a period of 10 years, with the seismic activity subsequently resuming several years before the Nagano-ken Seibu earthquake presenting two active centers in the seismic gap, near one of which a large destructive earthquake had occurred in 1906. It seems that the occurrence of the Nagano-ken Seibu earthquake could have been predicted if the characteristics of the seismic activity after the Fukui earthquake and the tectonic features in the region had been correctly interpreted.

In view of the potential importance of long-period ground motion in the design of large structures, near-field ground displacement is computed by the elastic dislocation theory for several earthquake fault models. The validity of such computations is confirmed by comparing the computed seismogram with the observed long-period seismogram of the 1923 Kanto earthquake. The ground motions are computed for three hypothetical earthquakes, a hypothetical Kanto earthquake, Tokai earthquake and Nemuro-Oki earthquake. The location and the nature of the faulting of these earthquakes are predicted by plate tectonics and precise earthquake mechanism studies. Major conclusions are: Tokyo may suffer, in the hypothetical Kanto earthquake, ground motions about half as large as those experienced in the 1923 Kanto earthquake; Hamamatsu, a large city on the Tokai coast, may experience in the hypothetical Tokai earthquake ground motions which are as large as, or even larger than, those experienced in the epicentral area of the 1923 Kanto earthquake; the hypothetical Nemuro-Oki earthquake may cause ground motions as large as those experienced in the 1968 Tokachi-Oki earthquake on the coastal cities in Hokkaido.

Three types of crustal movements in seismic regions can be distinguished: (α) slow movement evidenced between the spurts of seismic activity; (β) accelerated movement which is evidenced for 10–15 years before some earthquakes; (γ) rapid movement which is connected with the earthquake itself and is continued up to the dying down of seismic activity. The compilation of the maps of recent crustal movements in seismic regions would be a help for the prediction of earthquake places. The systematic observations for the crustal movements would be useful for the earthquake time prediction.

If a group of foreshocks can be distinguished from ordinary seismic activities, it would contribute to earthquake prediction. In the case of two Japanese earthquakes, which were preceded by immediate foreshocks, the coefficient “b” in the magnitude versus frequency equation was significantly smaller than that of the aftershocks or of the ordinary seismic activity in the region. The same characteristics were also found for the Chilean earthquake of 1960. Very large earthquakes can also take place without any immediate foreshocks. The great Kanto earthquake of 1923 was not preceded by any immediate foreshocks. However, the seismic activity in the general region became abnormally high from one year prior to the earthquake. Whether large earthquakes in a specific region are preceded by foreshocks or not, or, if not, what kind of foregoing activity is noticeable may probably vary from one seismic region to another. Therefore, detailed seismic data over a long period of time will be necessary to know the regional characteristic of foreshocks or foregoing activity. To achieve earthquake prediction, the physical process of accumulation and release of seismic energy must eventually be made known as well as statistical data.

A major earthquake occurred in the Straits of Messina in Southern Italy on December 28, 1908. The shock became known in literature as the “Messina earthquake”. The present paper deals mainly with a re-interpretation of the source mechanism of this shock using modern interpretation techniques. The following parameters are discussed: Epicentre, hypocentre, origin-time, macroseismic intensity, body and surface wave magnitudes, fault-plane solution, seismic moments derived from body and surface waves, residual displacements and tsunami values. The parameters are compared with the tectonic setting of Southern Italy. As a result of the large magnitude (M = 7) and the 30 km long fault extension of this shock, the Messina earthquake reflects the large scale tectonics of Southern Italy more significantly than any other earthquake from which seismograms exist. The focal mechanism studies show a good agreement between the orientation of the shear-plane of the earthquake and the tectonic features of a fault-line which is known from geological data, the so-called Comiso-Catania-Messina-S. Eufemia fault. The rupture of the shallow earthquake was associated with normal faulting. According to the normal faulting mechanism, the earthquake resulted from tensional forces acting approximately perpendicular to the strike direction of the Straits of Messina. Geodetic measurements which were performed shortly before and after the earthquake indicate an apparent asymmetry in the volumes of uplift and subsidence. It is suggestive that there exists a strong net subsidence of material displaced with the earthquake faulting. The normal faulting mechanism of the Messina earthquake is in tectonic accordance with the uprise of olivine basalt in the volcanism observed at Etna. However, the normal faulting is contrary to the postulation of a subducting oceanic plate giving rise to the Calabrian-Sicilian arc-like feature, as this hypothesis would require mainly a compressional zone in the region of the Straits of Messina.