NIED seismic moment tensor catalogue for regional earthquakes around Japan: quality test and application

We have examined the quality of the National Research Institute for Earth Science and Disaster Prevention (NIED) seismic moment tensor (MT) catalogue obtained using a regional broadband seismic network (FREESIA). First, we examined using synthetic waveforms the robustness of the solutions with regard to data noise as well as to errors in the velocity structure and focal location. Then, to estimate the reliability, robustness and validity of the catalogue, we compared it with the Harvard centroid moment tensor (CMT) catalogue as well as the Japan Meteorological Agency (JMA) focal mechanism catalogue. We found out that the NIED catalogue is consistent with Harvard and JMA catalogues within the uncertainty of 0.1 in moment magnitude, 10 km in depth, and 15° in direction of the stress axes. The NIED MT catalogue succeeded in reducing to 3.5 the lower limit of moment magnitude above which the moment tensor could be reliably estimated. Finally, we estimated the stress tensors in several different regions by using the NIED MT catalogue. This enables us to elucidate the stress/deformation field in and around the Japanese islands to understand the mode of deformation and applied stress. Moreover, we identified a region of abnormal stress in a swarm area from stress tensor estimates.     

Increased correlation range of seismicity before large events manifested by earthquake chains

“Earthquake chains” are clusters of moderate-size earthquakes which extend over large distances and are formed by statistically rare pairs of events that are close in space and time (“neighbors”). Earthquake chains are supposed to be precursors of large earthquakes with lead times of a few months. Here we substantiate this hypothesis by mass testing it using a random earthquake catalog. Also, we study stability under variation of parameters and some properties of the chains. We found two invariant parameters: they characterize the spatial and energy scales of earthquake correlation. Both parameters of the chains show good correlation with the magnitudes of the earthquakes they precede. Earthquake chains are known as the first stage of the earthquake prediction algorithm reverse tracing of precursors (RTP) now tested in forward prediction. A discussion of the complete RTP algorithm is outside the scope of this paper, but the results presented here are important to substantiate the RTP approach.     

Integration and magnitude homogenization of the Egyptian earthquake catalogue

The aim of the present work is to compile and update a catalogue of the instrumentally recorded earthquakes in Egypt, with uniform and homogeneous source parameters as required for the analysis of seismicity and seismic hazard assessment. This in turn requires a detailed analysis and comparison of the properties of different available sources, including the distribution of events with time, the magnitude completeness, and the scaling relations between different kinds of magnitude reported by different agencies. The observational data cover the time interval 1900–2004 and an area between 22°–33.5° N and 25°–36° E. The linear regressions between various magnitude types have been evaluated for different magnitude ranges. Using the best linear relationship determined for each available pair of magnitudes, as well as those identified between the magnitudes and the seismic moment, we convert the different magnitude types into moment magnitudes M _W, through a multi-step conversion process. Analysis of the catalogue completeness, based on the M _W thus estimated, allows us to identify two different time intervals with homogeneous properties. The first one (1900–1984) appears to be complete for M _W ≥ 4.5, while the second one (1985–2004) can be considered complete for magnitudes M _W ≥ 3.     

Space-time distribution patterns of destructive earthquakes in the inner belt of central Japan: Activity intervals and locations of earthquakes

Based on a block structure model of the inner belt of central Japan, an examination was conducted of the space-time distribution patterns of destructiv magnitudes M 6.4 or greater (M =Japan Meteorological Agency Scale). The distribution patterns revealed a periodicity in earthquake activit seismic gaps. Major NW—SE trending left-lateral active faults divide the inner belt of central Japan into four blocks, 20–80 km wide. The occurrenc A.D. with M ≥ 6.4, which have caused significant damage, were documented in the inner belt of central Japan. The epicenters of these earthquakes close to the block boundaries.

Using the relationship between the magnitude of earthquakes which occurred in the Japanese Islands and the active length of faults that generated them, movement is calculated for each historical earthquake. Space—time distributions of earthquakes were obtained from the calculated lengths, the latitud of generation. When an active period begins, a portion or segment of the block boundary creates an earthquake, which in turn appears on the ground surf active period ends when the block boundary generates earthquakes over the entire length of the block boundary without overlapping.

Five seismic gaps with fault lengths of 20 km or longer can be found in the inner belt of central Japan. It is predicted that the gaps will generate ea magnitudes of 7.0. These data are of significance for estimating a regional earthquake risk over central Japan in the design of large earthquake resist

The time sequences of earthquakes on the block boundaries reveal a similar tendency, with alternating active periods with seismic activity and quiet pe activity. The inner belt of central Japan is now in the last stage of an active period. The next active period is predicted to occur around 2500 A.D.     

Trial of earthquake prediction in Japan and a statistical test of time-shift

Earthquake prediction was practiced in Japan to examine the hypothesis that “a pair of earthquakes with similar magnitudes may be a signal of an impending larger earthquake”. In the present study, predictions were announced with expected probabilities of 20–30% (rank A) or 10–20% (rank B). In 2001–2002, excepting the Ogasawara region, 26 and 6 cases among 61 and 30 predictions of ranks A and B, respectively, were successful. Based on a statistical test of time-shift, i.e., one-year shift in this paper, and averaged activity in 1990–1999, the success rate of 43% for rank A was shown to be greater than that expected by chance with a confidence level more than 99%. The success rate of 20% for rank B gave a corresponding confidence level of only about 40%, suggesting that the predictions of rank B were not confident in this period. According to the results, a statistical test of time-shift was found to be useful to evaluate the significance of prediction methods of this type.     

Induced earthquakes in the Izu peninsula by the Izu-Hanto-Oki earthquake of 1974, Japan

Recently small earthquakes in the Izu Peninsula, central Japan, occurred in a region where differential strain, or shear strain on the nodal planes, may have been enhanced by the Izu-Hanto-oki earthquake of 1974 (M = 6.9 after JMA). It is suggested that the seismic ctivity was induced by the redistribution of strain accompanying the Izu-Hanto-oki earthquake. The activity from August, 1975, may have also been affected by an abnormal uplift in the northeastern part of the peninsula. Based on plausible models, the uplift caused the accumulation of differential strain in the focal region of the subsequent earthquakes. Quantitatively, this change of crustal strain was of the order of 10⁻⁶; it is ten times as much as the average annual accumulation. Consequently, the sudden or rapid change of strain was likely to have played an essential role in the subsequent seismic activity. This effect could be one of the factors which trigger a shallow intra-plate earthquake.     

A statistical procedure for the analysis of seismotectonically induced hydrochemical signals: A case study from the Eastern Carpathians, Romania

Changes in the stress field of an aquifer system induced by seismotectonic activity may change the mixing ratio of groundwaters with different compositions in a well, leading to hydrochemical signals which in principle could be related to discrete earthquake events. Due to the complexity of the interactions and the multitude of involved factors the identification of such relationships is a difficult task. In this study we present an empiric statistical approach suitable to analyse if there is an interdependency between changes in the chemical composition of monitoring wells and the regional seismotectonic activity of a considered area. To allow a rigorous comparison with hydrochemistry the regional earthquake time series was aggregated into an univariate time series. This was realized by expressing each earthquake in form of a parameter “e”, taking into consideration both energetic (magnitude of a seismic event) and spatial parameters (position of epi/hypocentrum relative to the monitoring site). The earthquake and the hydrochemical time-series were synchronised aggregating the e-parameters into “earthquake activity” functions E, which takes into account the time of sampling relative to the earthquakes which occurred in the considered area. For the definition of the aggregation functions a variety of different “e” parameters were considered. The set of earthquake functions E was grouped by means of factor analysis to select a limited number of significant and representative earthquake functions E to be used further on in the relation analysis with the multivariate hydrochemical data set. From the hydrochemical data a restricted number of hydrochemical factors were extracted. Factor scores allow to represent and analyse the variation of the hydrochemical factors as a function of time. Finally, regression analysis was used to detect those hydrochemical factors which significantly correlate with the aggregated earthquake functions.This methodological approach was tested with a hydrochemical data set collected from a deep well monitored for two years in the seismically active Vrancea region, Romania. Three of the hydrochemical factors were found to correlate significantly with the considered earthquake activities. A screening with different time combinations revealed that correlations are strongest when the cumulative seismicity over several weeks was considered. The case study also showed that the character of the interdependency depends sometimes on the geometrical distribution of the earthquake foci. By using aggregated earthquake information it was possible to detect interrelationships which couldn't have been identified by analysing only relations between single geochemical signals and single earthquake events. Further on, the approach allows to determine the influence of different seismotectonic patterns on the hydrochemical composition of the sampled well. The method is suitable to be used as a decision instrument in assessing if a monitoring site is suitable or not to be included in a monitoring net within a complex earthquake prediction strategy.     

Unified scaling theory for distributions of temporal and spatial characteristics in seismology

The paper describes the unified scaling theory for distribution functions of temporal and spatial characteristics in seismology. It is based on the scaling of seismological characteristics calculated for various energy–spatial–temporal intervals. The common mathematical methods for the scaling of distribution functions are developed. The means to test possibility of such scaling are found as well. The relationship between the unified scaling theory and other present scaling approaches is determined. The theory is applied to two characteristics of different seismoactive regions. The first characteristic is the waiting time between earthquakes ΔT, the second one is a new space parameter ΔD_min, which is the minimum distance of a current seismic event to the nearest (in space) neighbor in an energy–spatial–temporal interval. The distribution of the characteristics ΔT and ΔD_min allows estimating the time interval to the next earthquake and the distance of the following earthquake from previous earthquakes. Thus, these characteristics are very important for seismic hazard estimations. Scaling of distributions functions is proven to be successful for ΔD_min in all energy–spatial–temporal intervals and for ΔT with variations of energy/magnitude range. The distribution function of ΔT for various time domains was stable in only 60% of the cases, and near to unstable for spatial variations.     

Effect of Aftershocks on Earthquake Hazard Estimation: An Example from the North Anatolian Fault Zone

In order to investigate the effect of aftershocks on earthquake hazard estimation, earthquake hazard parameters (_m, b and M_max) have been estimated by the maximum likelihood method from the main shocks catalogue and the raw earthquakes catalogue for the North Anatolian Fault Zone (NAFZ). The main shocks catalogue has been compiled from the raw earthquake catalogue by eliminating the aftershocks using the window method. The raw earthquake catalogue consisted of instrumentally detected earthquakes between 1900 and 1992, and historical earthquakes that occurred between 1000–1900. For the events of the mainshock catalogue the Poisson process is valid and for the raw earthquake catalogue it does not fit. The paper demonstrates differences in the hazard outputs if on one hand the main catalogues and on the other hand the raw catalogue is used. The maximum likelihood method which allows the use of the mixed earthquake catalogue containing incomplete (historical) and complete (instrumental) earthquake data is used to determine the earthquake hazard parameters. The maximum regional magnitude (M_max, the seismic activity rate (_m), the mean return period (R) and the b value of the magnitude-frequency relation have been estimated for the 24°–31° E, 31°–41° E, 41°–45° E sections of the North Anatolian Fault Zone from the raw earthquake catalogue and the main shocks catalogue. Our results indicate that inclusion of aftershocks changes the b value and the seismic activity rate _m depending on the proportion of aftershocks in a region while it does not significantly effect the value of the maximum regional magnitude since it is related to the maximum observed magnitude. These changes in the earthquake hazard parameters caused the return periods to be over- and underestimated for smaller and larger events, respectively.     

New observational information on the precursory accelerating and decelerating strain energy release

Recent reliable data are used to study the behavior of seismic activity before 46 strong shallow earthquakes (M ≥ 6.0), which correspond to five complete samples of mainshocks. These samples include 6 mainshocks (M = 6.0–7.1) that occurred in western Mediterranean since 1980, 17 mainshocks (M = 6.0–7.2) which occurred in the Aegean (Greece and surrounding area) since 1980, 5 mainshocks (M = 6.4–7.5) that occurred in Anatolia since 1980, 12 mainshocks (M = 6.0–7.3) that occurred in California since 1980 and 6 mainshocks (M = 7.0–8.3) that occurred in Japan since 1990. In all 46 cases, a similar precursory seismicity pattern is observed. Specifically, it is observed that accelerating Benioff strain (square root of seismic energy) release caused by preshocks occurs in a broad circular region (critical region), with a radius about eight times larger than the fault length of the mainshock, in agreement with results obtained by various research groups during the last two decades. However, in a much smaller circular region (seismogenic region), with a radius about four times the fault length, the corresponding preshock strain decelerates with the time to the mainshock. The time variation of the strain follows in both cases a power law but the exponent power is smaller than unit (m ¯ = 0.3) in the case of the accelerating preshock strain and larger than unit (m ¯ = 3.0) in the case of the decelerating preshock strain. Predictive properties of this “Decelerating In–Accelerating Out Strain” model are expressed by empirical relations. The possibility of using this model for intermediate-term earthquake prediction is discussed and the relative model uncertainties are estimated.     

Additional evidence on some relationship between Seismic Electric Signals (SES) and earthquake focal mechanism

Investigating the period 1983–1994 for western Greece, a possible correlation between the selectivity characteristics of the SES (seismic electric signals of the VAN method) and earthquake parameters has been reported by Uyeda et al. [Uyeda, S., Al-Damegh, K.S., Dologlou, E., Nagao, T., 1999. Some relationship between VAN seismic electric signals (SES) and earthquake parameters, Tectonophysics, 304, 41–55.]. They found that the earthquake source mechanism changed from largely strike-slip type to thrust type at the end of 1987, and this coincided with a shift in the SES sensitive site from Pirgos (PIR) to Ioannina (IOA) VAN station. Here, we report the results for the period January 1, 2002–July 25, 2004, during which the SES sensitive site of PIR became again active, after a 10-year period of “quiescence”. This activation was followed by strike slip earthquakes (on August 14, 2003 and March 17, 2004 with magnitude 6.4 and 6.5, respectively) in the Hellenic arc, which provides additional evidence on the correlation reported by Uyeda et al. The SES activities recorded at PIR have been discriminated from “artificial” noise by employing the natural time-domain analysis introduced recently.     

Perceptible earthquakes in the broad Aegean area

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.     

Tsunami magnitudes in Taiwan, the Philippines and Indonesia

The regional characteristics of tsunami magnitudes in the SE Asia region are discussed in relation to earthquake magnitudes during the period from 1960 to 1994. Tsunami magnitudes on the Imamura-Iida scale are investigated by the author's method (Hatori 1979, 1986) using the data of inundation heights near the source area and tide-gauge records observed in Japan. The magnitude values of the Taiwan tsunamis showed relatively to be small. On the contrary, the magnitudes of tsunamis in the vicinities of the Philippines and Indonesia exceed more than 1–2 grade (tsunami heights: 2–5 times) compared to earthquakes with similar size on the circum-Pacific zone. The relation between tsunami magnitude, m, and earthquake magnitude, M _s, is expressed as m = 2.66 M _s– 17.5 for these regions. For example, the magnitudes for the 1976 Mindanao tsunami (M _s= 7.8, 3702 deaths) and the 1992 Flores tsunami (M _s= 7.5, 1713 deaths) were determined to be m = 3 and m = 2.5, respectively. The focal depth of tsunamigenic earthquakes is shallower thand< 36 km, and the detectively of tsunamis is small for deep earthquakes being d > 40 km. For future tsunamis, it is indispensable to take precautions against shallow earthquakes having the magnitudes M _s> 6.5.     

Application of a modified pattern informatics method to forecasting the locations of future large earthquakes in the central Japan

We propose a modification of the Pattern Informatics (PI) method that has been developed for forecasting the locations of future large earthquakes. This forecast is based on analyzing the space–time patterns of past earthquakes to find possible locations where future large earthquakes are expected to occur. A characteristic of our modification is that the effect of errors in the locations of past earthquakes on the output forecast is reduced. We apply the modified and original methods to seismicity in the central part of Japan and compared the forecast performances. We also invoke the Relative Intensity (RI) of seismic activity and randomized catalogs to constitute null hypotheses. We do statistical tests using the Molchan and Relative Operating Characteristic (ROC) diagrams and the log-likelihoods and show that the forecast for using the modified PI method is generally better than the competing original-PI forecast and the forecasts from the null hypotheses. Using the bootstrap technique with Monte-Carlo simulations, we further confirm that earthquake sequences simulated based on the modified-PI forecast can be statistically the same as the real earthquake sequence so that the forecast is acceptable. The main and innovative science in this paper is the modification of the PI method and the demonstration of its applicability, showing a considerable promise as an intermediate-term earthquake forecasting tool.     

Stress transfer in earthquakes, hazard estimation and ensemble forecasting: Inferences from numerical simulations

Observations indicate that earthquake faults occur in topologically complex, multi-scale networks driven by plate tectonic forces. We present realistic numerical simulations, involving data-mining, pattern recognition, theoretical analyses and ensemble forecasting techniques, to understand how the observable space–time earthquake patterns are related to the fundamentally inaccessible and unobservable dynamics. Numerical simulations can also help us to understand how the different scales involved in earthquake physics interact and influence the resulting dynamics. Our simulations indicate that elastic interactions (stress transfer) combined with the nonlinearity in the frictional failure threshold law lead to the self-organization of the statistical dynamics, producing 1) statistical distributions for magnitudes and frequencies of earthquakes that have characteristics similar to those possessed by the Gutenberg–Richter magnitude–frequency distributions observed in nature; and 2) clear examples of stress transfer among fault activity described by stress shadows, in which an earthquake on one group of faults reduces the Coulomb failure stress on other faults, thereby delaying activity on those faults. In this paper, we describe the current state of modeling and simulation efforts for Virtual California, a model for all the major active strike slip faults in California. Noting that the Working Group on California Earthquake Probabilities (WGCEP) uses statistical distributions to produce earthquake forecast probabilities, we demonstrate that Virtual California provides a powerful tool for testing the applicability and reliability of the WGCEP statistical methods. Furthermore, we show how the simulations can be used to develop statistical earthquake forecasting techniques that are complementary to the methods used by the WGCEP, but improve upon those methods in a number of important ways. In doing so, we distinguish between the “official” forecasts of the WGCEP, and the “research-quality” forecasts that we discuss here. Finally, we provide a brief discussion of future problems and issues related to the development of ensemble earthquake hazard estimation and forecasting techniques.     

Multiarchive paleoseismic record of late Pleistocene and Holocene strong earthquakes in Switzerland

A multiarchive approach has been applied to the investigation of the late Pleistocene and Holocene record of strong earthquakes in Switzerland. The geological archives used for this study include active faults, lake deposits, slope instabilities, and caves. In the Basle area, eight trenches were opened across the Basle–Reinach fault, nearby rockfall deposits were systematically investigated, sediment cores were taken from two lakes, and nine caves were studied. In Central Switzerland, five lakes were investigated by means of high-resolution seismic lines and sediment cores. Furthermore, three caves were studied in Central Switzerland. Altogether, the investigations are based on more than 350 km of high-resolution reflection seismic lines, 450 m of core samples, 260 m of trenches, and 245 radiocarbon age determinations. The measured co-seismic displacements along the Basle–Reinach fault supply independent information for the magnitude of the AD 1356 Basle earthquake exclusively based on geological evidence. Deformation features related to three well-documented strong historic earthquake shocks were identified. Deformation features of the AD 1774 Altdorf and AD 1601 Unterwalden earthquakes can be used to calibrate paleoseismic evidence in Central Switzerland. Altogether, traces of 13 earthquakes could be found in the two study areas, all of them with magnitudes M_w  6 or greater. For the first time, the earthquake catalogue for Switzerland can be extended back beyond historic records, into the late Pleistocene, spanning 15,000 years.     

Magma genesis beneath Northeast Japan arc: A new perspective on subduction zone magmatism

It is being accepted that earthquakes in subducting slab are caused by dehydration reactions of hydrous minerals. In the context of this “dehydration embrittlement” hypothesis, we propose a new model to explain key features of subduction zone magmatism on the basis of hydrous phase relations in peridotite and basaltic systems determined by thermodynamic calculations and seismic structures of Northeast Japan arc revealed by latest seismic studies. The model predicts that partial melting of basaltic crust in the subducting slab is an inevitable consequence of subduction of hydrated oceanic lithosphere. Aqueous fluids released from the subducting slab also cause partial melting widely in mantle wedge from just above the subducting slab to just below overlying crust at volcanic front. Hydrous minerals in the mantle wedge are stable only in shallow (< 120 km) areas, and are absent in the layer that is dragged into deep mantle by the subducting slab. The position of volcanic front is not restricted by dehydration reactions in the subducting slab but is controlled by dynamics of mantle wedge flow, which governs the thermal structure and partial melting regime in the mantle wedge.     

Radiant flux as a guide to relative seismic efficiency

The parameters “radiant flux” (energy radiated per unit area of an earthquake fault in unit time) and “radiant flux per unit displacement” reflect the power dissipated on a fault during slip. Values for moderate-to-large earthquakes range over two orders of magnitude, implying considerable variations in seismic efficiency, even for events of similar magnitude occurring on faults of the same type.     

Maximum magnitudes in aftershock sequences in Taiwan

In this work, Båth’s Law, the b-value in Gutenberg–Richter Law (G–R Law) in the form of the 1/β relationship, and both the a- and b-values in the G–R Law were introduced in order to estimate maximum aftershock magnitudes of earthquake sequences in the Taiwan region. The averaged difference of magnitude between the mainshock and the maximum aftershock is 1.20, and is consistent with Båth’s Law, however, with a large uncertainty. The large uncertainty implies that the difference may result from a variable controlled by other factors, such as the aftershocks number of an earthquake sequence and magnitude threshold for mainshock. With 1/β, since 86% of the earthquake sequences with a M ⩾ 6.0 mainshock follow this relationship, the upper bound of the maximum magnitude can be estimated for an earthquake sequence with a large mainshock. The a- and b-values in the G–R Law was also considered by evaluating maximum aftershock magnitudes. As there are low residuals between the model and the observations, the results suggest that the G–R Law is a good index for maximum aftershock magnitude determinations. In order to evaluate the temporal decays of maximum aftershock magnitudes, modified Omori’s Law was introduced. Using the approaches mentioned above, the maximum magnitudes and the temporal evolution of an earthquake sequence could be modeled. Among them, the model of the G–R Law has the best fit with observations for most of earthquake sequences. It shows its feasibility. The results of this work may benefit seismic hazards mitigation in the form of rapid re-evaluations for short-term seismic hazards immediately following devastating earthquakes.