In observatory seismology, the effective automatic processing of seismograms is a time-consuming task. A contemporary approach for seismogram processing is based on the Deep Neural Network formalism, which has been successfully applied in many fields. Here, we present a 4D network, based on U-net architecture, that simultaneously processes seismograms from an entire network. We also interpret Acoustic Emission data based on a laboratory loading experiment. The obtained data was a very good testing set, similar to real seismograms. Our Neural network is designed to detect multiple events. Input data are created by augmentation from previously interpreted single events. The advantage of the approach is that the positions of (multiple) events are exactly known, thus, the efficiency of detection can be evaluated. Even if the method reaches an average efficiency of only around 30% for the onset of individual tracks, average efficiency for the detection of double events was approximately 97% for a maximum target, with a prediction difference of 20 samples. Such is the main benefit of simultaneous network signal processing.

Laboratory loading experiments, in general, are used to estimate the mechanical properties of rocks. The application of acoustic emission (AE), a proxy for very small-scale earthquakes, offers insights into rock fracturing, especially when advanced seismological methods are applied. The uniaxial compressive test, the most conventional test within the scope of laboratory rock mechanics, estimates strength, elastic constants, or microcracking thresholds. Obviously, rock structure and grain size heterogeneity are the crucial parameters controlling these mechanical properties. Here, four granites with a different structure were uniaxially loaded, while AE was monitored, in order to observe the manner of fracturing. AE data were utilized at three different quality levels (AE recognition; AE localization; and the AE source mechanism). However, our focus was to study AE source mechanisms via the application of a shear-tensile crack source model. During a uniaxial compressive test, common (structurally independent) AE features led to the general characterization of microfracturing. On the other hand, sample related differences in AE results are discussed with respect to the role of granite heterogeneity. The comprehensiveness of the AE application demonstrated the validity of various levels of AE experiments.

We have developed a Recurrent Neural Network (RNN)-based phase picker for data obtained from a local seismic monitoring array specifically designated for induced seismicity analysis. The proposed algorithm was rigorously tested using real-world data from a network encompassing nine three-component stations. The algorithm is designed for multiple monitoring of repeated injection within the permanent array. For such an array, the RNN is initially trained on a foundational dataset, enabling the trained algorithm to accurately identify other induced events even if they occur in different regions of the array. Our RNN-based phase picker achieved an accuracy exceeding 80% for arrival time picking when compared to precise manual picking techniques. However, the event locations (based on the arrival picking) had to be further constrained to avoid false arrival picks. By utilizing these refined arrival times, we were able to locate seismic events and assess their magnitudes. The magnitudes of events processed automatically exhibited a discrepancy of up to 0.3 when juxtaposed with those derived from manual processing. Importantly, the efficacy of our results remains consistent irrespective of the specific training dataset employed, provided that the dataset originates from within the network.

We present an algorithm for seismic event discrimination and event approximate location based on multi-station seismograms. A deep learning approach was applied using a two-step algorithm: (i) signal onsets were identified in individual tracks based on the use of long-short-term memory neural network layers; (ii) if a sufficient number of onsets were reliably identified, a preliminary location was determined. We adopted a "reverse location approach " where the time sense of a seismogram is reverted and the origin time is predicted using a neural network approach based on previously determined onsets. Successful location or origin time prediction also served as a feedback for confirming previous onset identification.& nbsp;The method was tested using a data set of Acoustic Emission generated from the uniaxial loading of a Westerly Granite specimen. Accuracy of the method was better than 97%. Discriminated events were automatically located and their seismic moment tensor was determined. Both types of results were in good agreement with the baseline data set.& nbsp;With respect to the particular nature of processed data, we provide a demo code which shows examples presented in the article. In addition, a detailed description of the algorithm, including the control parameter values, is provided in the text. Based on this information the method can be applied on any data.

In this work, we successfully identified seismic events (observations of earthquakes) in seismograms using a Convolutional Neural Network (CNN). In accordance with past (analogue) seismogram interpretations, we did not treat digital seismograms as a time series, as per the general method, but, rather, converted them into time snaps of continuous data flow. Multichannel seismograms were represented with a time frequency domain in the form of multilayer images, with each signal channel forming one image layer. Images were then exposed to CNN (composed of three convolutional blocks whose architecture design was justified using Bayesian optimization). To improve reliability, we evaluated the posterior type function (PTP) as a combination of the probabilities of all of the considered classes of signal types (five in our case) which increased robustness of the identification. For data, we used records of acoustic emission (AE) events. The events were generated during laboratory loading experiments originally performed to study material/rock properties. As known, AE events may be studied in the same manner as natural earthquakes and treated in other ways as laboratory earthquake models. AE events are less complex compared to natural earthquakes where many of the physical parameters are known or may be controlled. Based on our results, we concluded that the successful identification of AE events is a necessary step prior to applying the proposed methodology for identifying natural earthquakes in seismograms.

Nearly half a century has passed since the first broadband seismometer was installed by A. Plešinger and his team. The principles behind this instrument, as well as its features and new possibilities regarding broadband data mining, have been reported within the scientific literature. This text is intended as a commemoration of this pioneering work and also serves as a historical reminder of the social and political circumstances surrounding the time period of broadband seismometer introduction that are not mentioned in original works but that played an important role during broadband seismometer development.

Correction| December 16, 2020 Erratum to the KHC Seismic Station: The Birthplace of Broadband Seismology Petr Kolář Petr Kolář * 1Institute of Geophysics, Czech Academy of Sciences, Prague, Czech Republic *Corresponding author: kolar@ig.cas.cz Search for other works by this author on: GSW Google Scholar Seismological Research Letters (2021) 92 (2A): 1202–1204. https://doi.org/10.1785/0220200412 Article history first online: 16 Dec 2020 Cite View This Citation Add to Citation Manager Share Icon Share Facebook Twitter LinkedIn Email Tools Icon Tools Get Permissions Search Site Citation Petr Kolář; Erratum to the KHC Seismic Station: The Birthplace of Broadband Seismology. Seismological Research Letters 2020;; 92 (2A): 1202–1204. doi: https://doi.org/10.1785/0220200412 Download citation file: Ris (Zotero) Refmanager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search nav search search input Search input auto suggest search filter All ContentBy SocietySeismological Research Letters Search Advanced Search In Kolář (2020), we described the history of very broadband seismology (VBB) with special regard to the outstanding role of the seismic station Kašperské Hory (KHC, Czech Republic). A recent message from a reader mentioned that a comparison of VBB output and output recorded by a standard method at the time would be beneficial. We fully agree with this comment. As such, two examples have now been added to demonstrate the advantages of the VBB system. Both of the examples were obtained from Plešinger and Horálek (1976), in which the new VBB system was introduced for the first... You do not currently have access to this article.

During inversion for the source mechanisms of laboratory acoustic emission events, relatively high misfit values (expressed as the RMS of the inverted equations) were observed. Our experiment was performed on Westerly Granite. A processed set of data consisting of 2405 acoustic emission events was used and a semi-homogeneous velocity model was considered. A correction for sensor radiation patterns and individual sensor constants was taken into account, and an acausal attenuation model was assumed. Source mechanisms were inverted for the moment tensor. The application of a more sophisticated medium model improved inversion quality only for some events. Introducing the source directivity, a standard approach for earthquakes with magnitudes larger than approximately 4, increased the number of successfully inverted events. Directivity was introduced using a Haskell source model; optionally unilateral and bilateral versions of the source were considered. Lower values of RMS for the Haskell source model were considered to justify the directivity approach. This formalism enables us to select the preferable fault from the two nodal planes within the fault plane solution. The rupture directions were observed to tend to the dip direction of the preferred fault. They were found to be preferably subparallel to slip directions for acoustic emissions with a significant DC component. The source time functions retrieved from the seismograms are in agreement with finite source theory.

Traditionally in seismology and in acoustic emission (AE), full moment tensor (MT) is applied as a default model of the mechanism. We present an alternative - an application of shear tensile crack (STC) source model to AEs generated by uniaxial compression loading of Westerly granite. The advantages of STC over the conventional MT are as follows: (i) contrary to the MT, the STC is physical source since it describes straight and simple fracture modes anticipated inside a loaded sample, namely the shear-slip and both of opening and closing tensile cracks; and (ii) the STC is simpler because it is described by fewer parameters (five instead of six required for an unconstrained MT), an essential feature for stabilizing the inverse problem.Better suitability of STC over MT is demonstrated by three exemplary AEs (tensile, shear, and combined). The obtained results were confirmed using a statistical analysis of 1630 reliably determined source mechanisms. The STC, as compared to the MT, provides smaller confidence regions for orientation and even smaller regions for decomposition parameters. Thus, the STC solution appeared to be substantially more useful than the MT, namely for mechanisms with a high content of non-double-couple (non-DC) component, as it allowed better distinction between tensile and shear AEs.Grain scale cracks of tension and combined source type, localized within the middle-height circumferential portion of the specimen, dominated fracturing. Azimuthal distribution of fault planes was found to be approximately uniform for all three source types. The fault dip increased with increasing content of the non-DC component. The average values were 16°, 21°, and 26° for the tensile, combined and shear source types, respectively. The specimen failed by flaking in areas of a high AE activity. AE locations and failure mechanisms indicate a perfect confinement end-boundary conditions between the tested specimen and the loading platens.

Here, we present an experimental study of the fracturing of anisotropic migmatite with plane-parallel structure (foliation). Four specimens, with different dips of foliation, including subhorizontal (13°), subvertical (86°), and oblique (47°, 67°), were uniaxially loaded up to failure. Acoustic emission monitoring and ultrasonic sounding were applied for fracturing characterization. In case of subhorizontal and subvertical foliation, the tensile source type played an important role before reaching nucleation stress when shearing became dominant. The dominance of non-tension microcracking was characteristic for oblique foliation. The uniaxial loading of migmatite resulted in triaxial stress state regardless of the foliation dip. The minimum stress axis was in the subhorizontal direction of the foliation dip. The triaxial stress state caused the preferential orientation of induced shear and tension microcracks that had the same azimuthal orientation as the foliation. Preferential microcracking resulted in preferential orientation of the failure plane. Based on the anisotropic behavior of migmatite’s characteristic mechanical properties, as well as its mechanism of failure that is typical for anisotropic rocks, the obtained conclusions may be generalized for other types of metamorphic anisotropic rocks with a plane-parallel structure.

Set of events from West Bohemian 2008 seismic swarm with known source mechanisms is processed. The events or their slips respectively are clustered into two groups: (i) principal events with slip laying in the main fault plane and (ii) complementary events deviating from that plane. From those slips we constructed image of slip distribution (a new way of data/slip presentation) and from slip distribution and variations we hypothesized about foci zone properties. Namely, we propose that western block is more rigid and compact; the eastern block appears to be constituted from several sub-blocks which can interact with each other during the swarm course. Our hypothesis is supported by similar image constructed from relative rupture velocities, which we consider as independent data. The proposed structural model agrees with the existence of the different observed types of source mechanisms.

Here a shear‐tensile crack (STC) model is presented as suitable for acoustic emission (AE) events. Experimental data were obtained from a uniaxial compression test performed on a Westerly Granite specimen using a 14‐channel AE monitoring system. The advantages of the STC versus a traditional MT (moment tensor) approach are as follows: (i) it is a physical source, contrary to the MT, since the STC describes the straight and simple fracture modes anticipated inside a loaded sample, namely, the shear slip and both the opening and closing tensile cracks; and (ii) it is simpler because it is described by fewer parameters (five instead of the six required for an unconstrained MT), which is essential for solving the inverse problem. The presented STC procedure was tested on 38 AE events selected over a range of 50–98% for the uniaxial compressive strength. As compared to the MT model, the STC model displayed a similar fit for input data while providing far smaller confidence regions. The results indicate a more certain determination for the mechanism of orientation and improved reliability for the decomposition components. In addition, use of STC model allowed better distinction between tension and shear type for AE events, which may be crucial for recognizing an approaching failure. For our experiment, application of the STC model proved to be useful for recognizing the threshold of unstable microcracking and indicative for determining the failure plain orientation.

Measurement of acoustic emission (AE) during laboratory experiments – rock sample loading is an important tool for study of material mechanical properties as well as understanding of modes of its failure. For correct data processing it is necessary to calibrate used AE sensors; the calibration constants can depend on property of individual sensors, on their contact conditions, etc. As there can be a deformation induced charge in sensor contact condition, the calibration constants may vary with time. AE sources can be modeled (in the same way as natural earthquakes) as double couple sources with use of moment tensor (MT) formalism (3x3 symmetric tensor, i.e. 6 independent values, the problem is linear). We adopted method, which enable calibration of an individual station of a seismic network. The method is based on simultaneous determination of MTs and calibration constant(s). During the experiment there were repeatedly measured velocities of elastic waves by ultrasonic sounding, when AE sensors cyclically acted as sources. We processed this data and for each cycle determined calibration constant for all the sensors but one, which is supposed to be 1. The source is supposed to be of single force type, which further reduces number of determined parameters. The evaluation was performed cyclically for all the sensors, the final calibration constants are then the mean values.

In addition to the previously processed set of 91 events from year 2008 seismic swarm from West Bohemian region (Czech Rep.), we investigated another 368 events, i.e. 454 events in total (M L 0.1–3.8). We determined parameters of a finite source model by stopping-phases method, namely, source radius and rupture velocity. Newly, the methodology was amended by applying a correction for the effect of attenuation; this effect was found to have an ineligible influence on the results, especially for weaker events (M L  < 1.2). The results are presented both in numerical and in graphical form. From the determined source parameters, we again estimated stress drops of processed events—the results confirm a hypothesis about constant stress drop (with the mean value being 2.9 ± 1.3 MPa). Even if a newly implemented correction for the effect of attenuation decreases the stress drop variation (especially for weaker events), the possibility of partial saturation still remains open. We also present an attempt of a 3D structure model of foci volume and speculate about possible sub-vertical structure in the foci area.

Finite seismic source parameters were determined for a set of 91 selected events of the West Bohemia year 2008 earthquake swarm (M (L) from 0.6 to 3.7) using the stopping phases method. According to the theory, two stopping phases are generated along the source border where the rupture process terminates. These two phases form a Hilbert transform pair; it is also a criterion for their identification. Circular and elliptical source models were considered and their parameters were recovered using the differences in arrival times between the identified stopping phases. Generalization of the circular to elliptical model was found to be statistically significant only for a minority of the events; consequently, only circular source models were investigated in detail. Individual source parameter errors were estimated with the use of the jackknife method. Our results are in good agreement with a previously published theoretical formula relating source radius and magnitude and also with the relation derived in the year 2000 swarm. Our results also confirm rather well the general theoretical assumption about the constant stress drop (with median value of 2.4 MPa and with the majority of values ranging from 1 to 10 MPa).

In the previous work we discovered a possible correlation between seismic activity in West Bohemia (Czech Rep.) and electro-magnetic emission (hereafter EME) during earthquake swarm in 2008 year (frequency range of observation: 0.2 - 10 Hz). Since then the EME observation has been continuing for 5 years. We present new results of EME observation during two consecutive swarms (in years 2011 and 2013) - we have to conclude, that the statistical correlation observed during 2008 year swarm was observed only partly and weakly for 2011 year swarm and it was not confirmed for the 2013 year swarm. We are not able to explain observed fact (due to heuristic and statistical character of original hypotheses), it can be an accidental correlation (in 2008 year), but the effect could also vanish for weaker swarms (i.e. for 2011 year and especially for 2013 year), or it could be an effect of different medium property round the hypocenters (the seismic activity migrates to the north with time).

We study seismicity of the West Bohemia seismic swarm area. The area is covered by high quality seismic stations of WEBNET network. In seismic records of local events we can identify not only direct P- and S- wave arrivals but also other waves. Seismograms contain many reflected, refracted or split phases, which reflects complex geological as well as tectonic settings of the region. We try to identify these phases in seismograms and their origin at structural interfaces. Results of such investigation will improve our knowledge about upper crustal structure of the swarm area.

Parameters of finite seismic source model were determined for a set of 36 selected events of the West Bohemia 2000 earthquakes swarm (Ml from 1.7 to 3.0) using stopping phases method. Two stopping phases are generated along the source border where the rupture process terminates and these phases form Hilbert transform pair, which is also the criterion for their identification. Circular and elliptical source models were considered and corresponding source parameters were calculated by inverting interpreted stopping phases delays. As generalization of circular to elliptical model was found to be statistically insignificant, only results related to the circular source including error estimates are presented. Our results are in a good agreement with previously published theoretical formula concerning source radius and magnitude and also fairly well confirm general theoretical assumption about constant stress drop. The determined stress drop ranges between 1 - 10 MPa with the typical value of 2.4 MPa.