ALBERT

Description: We have developed a network optimization method for regional-scale microseismic monitoring networks and applied it to optimize the densification of the existing seismic network in northeastern Switzerland. The new network will build the backbone of a 10-yr study on the neotectonic activity of this area that will help to better constrain the seismic hazard imposed on nuclear power plants and waste repository sites. This task defined the requirements regarding location precision (0.5 km in epicentre and 2 km in source depth) and detection capability [magnitude of completeness M c  = 1.0 ( M L )]. The goal of the optimization was to find the geometry and size of the network that met these requirements. Existing stations in Switzerland, Germany and Austria were considered in the optimization procedure. We based the optimization on the simulated annealing approach proposed by Hardt & Scherbaum, which aims to minimize the volume of the error ellipsoid of the linearized earthquake location problem ( D -criterion). We have extended their algorithm to: calculate traveltimes of seismic body waves using a finite difference ray tracer and the 3-D velocity model of Switzerland, calculate seismic body-wave amplitudes at arbitrary stations assuming the Brune source model and using scaling and attenuation relations recently derived for Switzerland, and estimate the noise level at arbitrary locations within Switzerland using a first-order ambient seismic noise model based on 14 land-use classes defined by the EU-project CORINE and open GIS data. We calculated optimized geometries for networks with 10–35 added stations and tested the stability of the optimization result by repeated runs with changing initial conditions. Further, we estimated the attainable magnitude of completeness ( M c ) for the different sized optimal networks using the Bayesian Magnitude of Completeness (BMC) method introduced by Mignan et al. The algorithm developed in this study is also applicable to smaller optimization problems, for example, small local monitoring networks. Possible applications are volcano monitoring, the surveillance of induced seismicity associated with geotechnical operations and many more. Our algorithm is especially useful to optimize networks in populated areas with heterogeneous noise conditions and if complex velocity structures or existing stations have to be considered.

Description: We describe the setup of testing regions for the China Earthquake Forecast Testing Center and provide preliminary forecast results in the scope of the Collaboratory for the Study of Earthquake Predictability (CSEP) project. We investigate the spatiotemporal variations of the completeness magnitude M c by using the frequency-magnitude distribution of the China Earthquake Networks Center (CENC) catalog. We find three periods of significantly different M c histories: (I) 1 January 1970–30 September 2001, (II) 1 October 2001–30 September 2008, and (III) 1 October 2008–31 August 2011. M c mapping provides median values , 2.2, and 1.6 for the three periods of time, respectively, showing the improvement in catalog completeness over time. We recommend using data from periods II and III to define a baseline long enough for retrospective forecast testing. Small magnitude events from period I should be used with caution due to important fluctuations in completeness. For period III, coordinates of all national and regional seismic stations are available, and we therefore apply the Bayesian magnitude of completeness (BMC) technique, mapping M c continuously by using prior information on the relationship between M c and the density of seismic stations. We define four potential testing/collection areas for CSEP-China: (A) All China, (B) North–South Seismic Belt (NSSB), (C) North and West Xinjiang Seismic Region, and (D) North China Seismic Region. In the current phase of CSEP-China, only the NSSB (region B) is considered. To demonstrate the type of earthquake predictability experiment that will be performed in the Chinese Testing Center, we present a series of retrospective forecast experiments with TripleS, a smoothed seismicity model. Online Material: The CENC earthquake catalog (1 January 1970–31 August 2011, restricted to magnitudes M ≥3.0) as well as completeness magnitude M c ( x , y ) spatial grids.

Description: 〈span〉〈div〉Summary〈/div〉The complete part of the earthquake frequency-magnitude distribution, above the completeness magnitude 〈span〉mc〈/span〉, is well described by the Gutenberg-Richter law. On the other hand, incomplete data does not follow any specific law, since the shape of the frequency-magnitude distribution below max(〈span〉mc〈/span〉) is function of 〈span〉mc〈/span〉 heterogeneities that depend on the seismic network spatiotemporal configuration. This paper attempts to solve this problem by presenting an asymmetric Laplace mixture model, defined as the weighted sum of Laplace (or double exponential) distribution components of constant 〈span〉mc〈/span〉, where the inverse scale parameter of the exponential function is the detection parameter κ below 〈span〉mc〈/span〉, and the Gutenberg-Richter β-value above 〈span〉mc〈/span〉. Using a variant of the expectation maximization algorithm, the mixture model confirms the ontology proposed by Mignan [2012, 〈a href="https://doi.org/10.1029/2012JB009347"〉https://doi.org/10.1029/2012JB009347〈/a〉], which states that the shape of the earthquake frequency-magnitude distribution shifts from angular (in log-linear space) in a homogeneous space-time volume of constant 〈span〉mc〈/span〉 to rounded in a heterogeneous volume corresponding to the union of smaller homogeneous volumes. The performance of the proposed mixture model is analysed, with encouraging results obtained in simulations and in 8 real earthquake catalogues that represent different seismic network spatial configurations. We find that 〈span〉k〈/span〉 = κ/ln(10) ≈ 3 in most earthquake catalogues (compared to 〈span〉b〈/span〉 = β/ln(10) ≈ 1), suggesting a common detection capability of different seismic networks. Although simpler algorithms may be preferred on pragmatic grounds to estimate 〈span〉mc〈/span〉 and the 〈span〉b〈/span〉-value, other methods so far fail to model the angular distributions observed in homogeneous space-time volumes. Mixture modelling is a promising strategy to model the full earthquake magnitude range, hence potentially increasing seismicity data availability tenfold, since c. 90 per cent of earthquake catalogue events are below max(〈span〉mc〈/span〉).〈/span〉

Description: 〈span〉〈div〉ABSTRACT〈/div〉The historical record of earthquakes is a crucial data resource for seismic hazard analysis. In every region, the largest events are rare and difficult to parameterize. In cases for which such events are associated with the ruptures of mapped faults, defining the extent of possible fault ruptures is an important challenge that is often guided by historical precedent in which geologic, particularly paleoseismic, studies are limited. Counterfactual risk analysis recognizes that a historical event is just one realization of many possible alternatives: a fault rupture that happened in the past is just one of numerous ways in which seismic energy might have been dynamically released. Stochastic modeling of past earthquakes can provide additional insight into the complex geometry of multifault rupture. This counterfactual approach that is advocated here extends the effective time window of observation of the fault rupture process, well beyond the time span of earthquake catalogs. The basic concepts of counterfactual risk analysis are explained, followed by a seismological discussion of runaway earthquakes, illustrated in the case of the North Anatolian fault (NAF) in Turkey. This notable example demonstrates the practical utility of the counterfactual approach as a new supplementary tool for addressing one of the most difficult problems in probabilistic seismic hazard and risk assessment: selecting the multifault combinations to model as explicit seismic sources.〈/span〉

Description: 〈span〉〈div〉Abstract〈/div〉It is still debated whether earthquake occurrence can be described as a single process or whether foreshocks are different phenomena. If foreshocks behaved differently, this would suggest a change of physical processes in the mainshock preparation phase that would boost hopes of forecasting large earthquakes. Most research on foreshocks focuses on case studies or uses global datasets in which recordings of small earthquakes are incomplete and are thus neglected. We do comprehensive foreshock statistics on all mainshocks in a regional earthquake catalog that is complete above 〈strong〉M〈/strong〉 2.5. To detect possible differences between foreshocks and seismicity that follows a uniform triggering model (the epidemic‐type aftershock sequence [ETAS] model), we perform a null‐hypothesis test. We also estimate the size of the differences between observed and ETAS‐simulated foreshocks.We define different sets of foreshocks using two different methods, because there is no unique definition: a nearest‐neighbor declustering technique (〈a href="https://pubs.geoscienceworld.org/bssa#rf48"〉Zaliapin 〈span〉et al.〈/span〉, 2008〈/a〉) and a variety of space–time windows (e.g., 〈a href="https://pubs.geoscienceworld.org/bssa#rf2"〉Agnew and Jones, 1991〈/a〉). We use data from southern California, northern California, and Italy. For each region, we first search an appropriate null model: an ETAS model that describes aftershock numbers well. In southern California, we find an appropriate spatiotemporal model that is characterized by a large productivity parameter α. After performing a null‐hypothesis test for different mainshock and foreshock magnitudes, we find foreshock signals (p〈0.05) for all mainshocks sizes and independent of the foreshock’s lower magnitude threshold. Observed mainshocks have more foreshocks than the ETAS model predicts.〈/span〉

Description: Hainzl and Christophersen (2016) recently suggested that the approach used by Mignan (2016) to visualize the Omori law in a log–log plot using the complementary cumulative density function (CCDF) was misleading. They showed that one should use a temporal upper bound t max =max( t obs ), with t obs the occurrence time of a set of observed aftershocks, instead of t max -〉, as used by Mignan. They found that both the Omori law and stretched exponential function (SEF) are undistinguishable on a corrected CCDF log–log plot but that the Omori law is preferred based on maximum-likelihood estimations (MLEs). I first clarify the rationale for using the CCDF log–log plot for comparison of the SEF with a power law but verify that the comparison indeed becomes misleading when the data sequence is incomplete. However, I then show that MLEs obtained for the Omori law are ambiguous, because this function may win versus the SEF, even if the true relaxation process follows an SEF (subject to early data incompleteness). This debate highlights that statistics alone may not be enough to choose between the stretched exponential and the Omori law and that physical considerations should be added to the discussion.

Description: Assessing the completeness magnitude Mc of earthquake catalogs is an essential prerequisite for any seismicity analysis. We employ a simple model to compute Mc in space based on the proximity to seismic stations in a network. We show that a relationship of the form [IMG]/medium/1371eq1.gif" ALT="Formula "〉, with d the distance to the kth nearest seismic station, fits the observations well, k depending on the minimum number of stations being required to trigger an event declaration in a catalog. We then propose a new Mc mapping approach, the Bayesian magnitude of completeness (BMC) method, based on a two-step procedure: (1) a spatial resolution optimization to minimize spatial heterogeneities and uncertainties in Mc estimates and (2) a Bayesian approach that merges prior information about Mc based on the proximity to seismic stations with locally observed values weighted by their respective uncertainties. Contrary to the current Mc mapping procedures, the radius that defines which earthquakes to include in the local magnitude distribution is chosen according to an objective criterion, and there are no gaps in the spatial estimation of Mc. The method solely requires the coordinates of seismic stations. Here, we investigate the Taiwan Central Weather Bureau (CWB) seismic network and earthquake catalog over the period 1994-2010.

Description: Various seismicity patterns before large earthquakes have been reported in the literature. They include foreshocks (medium-term acceleration and short-term activation), quiescence, doughnut patterns and event migration. The existence of these precursory patterns is however debated. Here, we develop an approach based on the concept of stress accumulation to unify and categorize all claimed seismic precursors in a same physical framework. We first extend the Non-Critical Precursory Accelerating Seismicity Theory (N-C PAST), which already explains most precursors, to additionally include short-term activation. Theoretical results are then compared to the time series observed prior to the 2009 Mw = 6.3 L'Aquila, Italy, earthquake. We finally show that different precursory paths are possible before large earthquakes, with possible coupling of different patterns or non-occurrence of any. This is described by a logic tree defined from the combined probabilities of occurrence of the mainshock at a given stress state and of precursory silent slip on the fault. In the case of the L'Aquila earthquake, the observed precursory path is coupling of quiescence and accelerating seismic release, followed by activation. These results provide guidelines for future research on earthquake predictability.

Description: [1]  Understanding and forecasting earthquake occurrences is presumably linked to understanding the stress distribution in the earth's crust. This cannot be measured instrumentally with useful coverage. However, the size-distribution of earthquakes, quantified by the Gutenberg-Richter b -value, is possibly a proxy to differential stress conditions and could therewith act as a crude stress-meter wherever seismicity is observed. In this study, we improve the methodology of b -value imaging for application to a high-resolution 3-d analysis of a complex fault network. In particular, we develop a distance dependent sampling algorithm and introduce a linearity measure to restrict our output to those regions where the magnitude distribution strictly follows a power-law. We assess the catalog completeness along the fault traces using the Bayesian Magnitude of Completeness method and systematically image  b -values for 243 major fault segments in California. We identify and report b -value structures, revisiting previously published features, e.g. the Parkfield asperity, and documenting additional anomalies, e.g. along the San Andreas and Northridge faults. Combining local b -values withlocal earthquake productivity rates, we derive probability maps for the annual potential of one or more M6+ events as indicated by the microseismicity of the last three decades. We present a physical concept of how different stressing conditions along a fault surface may lead to b -value variation and explain non-linear frequency magnitude distributions. Detailed spatial b -value information and its physical interpretation can advance our understanding of earthquake occurrence and ideally lead to improved forecasting ability.