Search

Automated procedures for seismic arrival-time picking on large and real-time seismological waveform datasets are critical for many seismological tasks. Recent, high-performance, automated arrival-time pickers mainly use deep-neural-networks applied to nearly raw, seismogram waveforms as input data. However, there is a long history in earthquake seismology of rule-based, automated arrival detection and picking algorithms that efficiently exploit variations in amplitude, frequency and polarization of seismogram waveforms.Here we use this classical, seismological domain-knowledge to transform raw seismogram waveforms into input features for a deep-learning picker. We preprocess 3-component, broadband seismograms into 3-component characteristic functions of a multi-band picker (FilterPicker), plus the instantaneous modulus and inclination of the waveforms. We use these five time-series as input instead of the 3-component, raw seismograms to extend the deep-neural-network picker PhaseNet within the SeisBench platform. We compare the original, purely data-driven PhaseNet and our extended, domain-knowledge PhaseNet (DKPN), using identical training and validation datasets, with application to in- and cross-domain testing datasets.We find that the explicit information targeting arrival-time detection and picking introduced by the domain-knowledge processing enables DKPN to be trained with smaller datasets than PhaseNet. Relative to PhaseNet, DKPN shows improved performance and stability for P picking and slightly improved S picking, especially for cross-domain application. With increasing training dataset size PhaseNet performance generally improves and converges to that of DKPN, except for cross-domain P picking. The results suggest that DKPN primarily needs to learn pick characterization, while PhaseNet additionally requires learning the more difficult task of arrival detection., The 28th IUGG General Assembly (IUGG2023) (Berlin 2023)

We present the status of ShakeMap-EU, an initiative initially proposed in 2018 to: (i) provide an integrated archive of ShakeMaps at the European level built on EPOS Seismology (www.epos-eu.org/tcs/seismology) services & data products and modern community software; (ii) serve as a backup to authoritative ShakeMap implementations; (iii) deliver ShakeMaps for regions where no local capability is yet available.ShakeMap-EU is accessible at the shakemapeu.ingv.it web portal since 2020, and is jointly governed by participating institutions through voluntary contributions and EC-funded projects. It has become a reliable European seismological service, integrating authoritative models and workflows consistently. The system is based on: (a) the latest version of ShakeMap®; (b) the earthquake information delivered by the EMSC (www.emsc-csem.org); (c) the earthquake shaking data distributed by ORFEUS (orfeus-eu.org/data/strong); (d) the ground motion models adopted within EFEHR (www.efehr.org) for mapping seismic hazard across Europe; (d) the official ShakeMap configurations of some of the most hazardous countries in Europe. Configuration of, and input to the system are managed via a GitHub repository that allows automatic/manual triggering and interaction by authorized users. ShakeMap-EU provides a collaboration framework and laboratory for seismological agencies to address the challenges posed by the heterogeneity of ground-shaking mapping strategies across Europe and the need to promote homogenization and best practices in this domain. ShakeMap-EU is used in research projects as the test platform for novel international collaborative research: among recent examples are the ongoing enhancements towards an evolutionary hazard information system including real-time seismicity characterization and information on earthquake-induced phenomena., The 28th IUGG General Assembly (IUGG2023) (Berlin 2023)

When an earthquake occurs, there are many remote sensors that record data from the event: waveforms recorded by seismic instruments are used to locate the hypocentre and magnitude of the earthquake and to assess the strong ground shaking at the surface of the earth; people and buildings feel the shaking and observe or record the impact that it has on them. All of these data can be used to rapidly estimate, in the minutes, hours or days following an earthquake, the impact that the earthquake has on the surrounding people, buildings and infrastructure. In the first few minutes following an earthquake, a simple qualitative assessment (e.g., no impact, minor impact, major impact) is often sufficient to understand the magnitude of the event, and this is referred to as Rapid Impact Assessment In the following hours, however, it becomes important to understand the impact in terms of quantitative losses (e.g., number of collapsed buildings, number of fatalities or homeless people, direct economic loss) and this estimation is referred to as Rapid Loss Assessment. This fast assessment of the impact of the earthquake provides first order estimates of the losses which can be continually updated as more information and data arrive from the remote sensors.

In this work it is illustrated how the urgent computing (UC) capabilities in the tsunami natural hazard framework are strengthening the monitoring and analysis functions of the European Emergency Response Coordination Centre (ERCC) and its Situational Awareness Sector (SAS) by helping to design the multi-hazard advice service at global level and on a 24/7 operational basis. In this context, the ARISTOTLE-eENHSP project (All Risk Integrated System TOwards Trans-boundary hoListic Early-warning - enhanced European Natural Hazards Scientific Partnership) has been designed to offer a flexible and scalable system that can provide new hazard-related services to the ERCC.

This data archive contains the training, validation and test datasets for ConvNetQuake_INGV as presented in this article: Lomax, A., Michelini, A., & Jozinović, D. (2019). An Investigation of Rapid Earthquake Characterization Using Single‐Station Waveforms and a Convolutional Neural Network. Seismological Research Letters, 90(2A), 517–529. https://doi.org/10.1785/0220180311 The training [validation] datasets consist of 15, 200 [1773] event and 10, 724 [1198] noise three- component waveforms and associated metadata from MedNet stations using events from 2010 to 2018 at 0°–180° with lower magnitude limits set as a function of event epicentral distance. The test datasets consists of 1003 event and 621 noise three-component waveforms and associated metadata and an extended set of 4074 test events from 2007 to 2009, selected otherwise with the same criteria as for the training and validation datasets.

In the age of data driven science the re-use of data and the compilation of existing data from monitoring infrastructures has become an integral part of research. For the sake of transparency and reproducibility of research it is crucial to be able to unambiguously identify data that were used as the basis of a publication. Globally unique and resolvable, persistent digital identifiers (PID) for digital data sets are an important tool to achieve this goal enabling unambiguous links between published research results and their underlying data. In addition, this unambiguous identification allows citation of data. Proven and community based examples are the usage of GenBank identifiers in the biological literature or the data citation method by using DOIs (digital object identifiers) already used widely in the scholarly literature. Identification of discrete digital objects is simple and citation can be formatted in analogy to citing literature. The identification of still ongoing, open time series does not seem to fit this pattern. A major prerequisite for the proper use of PIDs within data citations is the persistence of both, identifiers as well as the integrity of the associated data set. This poses questions when PIDs are to be used for unfinished data sets or open time series data. Such data is typically generated within research infrastructures during long lasting experiments such as satellite missions, environmental monitoring campaigns, or in permanent installations such as natural hazard detection and early warning systems (e.g., seismic traces acquired by field stations). Open time series data are often used in research during ongoing experiments and potentially published earlier than the underlying data set has been closed and is publicly released. It is therefore important to enable the scientific community to properly cite these data in their publications. Yet what is the meaning of “persistence” of data in ongoing time series? How does it relate to versioning? What is the granularity of a time series? In this publication we discuss and compare solutions currently used in some major European research infrastructures and propose transparent solutions which allow the citation of time series data using PIDs.

International audience; We retrieve seismic velocity variations within the Earth's crust in the region of L'Aquila (central Italy) by analyzing cross-correlations of more than two years of continuous seismic records. The studied period includes the April 6, 2009, Mw 6.1 L'Aquila earthquake. We observe a decrease of seismic velocities as a result of the earthquake's main shock. After performing the analysis in different frequency bands between 0.1 and 1 Hz, we conclude that the velocity variations are strongest at relatively high frequencies (0.5-1 Hz) suggesting that they are mostly related to the damage in the shallow soft layers resulting from the co-seismic shaking.

The Italian Tsunami Alert Center based at the Istituto Nazionale di Geofisica e Vulcanologia (CAT-INGV) has been monitoring the Mediterranean seismicity in the past 8 yr to get fast and reliable information for seismically induced tsunami warnings. CAT-INGV is a tsunami service provider in charge of monitoring the seismicity of the Mediterranean Sea and of alerting Intergovernmental Oceanographic Commission (IOC)/UNESCO subscriber Member States and the Italian Department of Civil Protection of a potentially impending tsunami, in the framework of the Tsunami Warning and Mitigation System in the North-eastern Atlantic, the Mediterranean and connected seas (NEAMTWS). CAT-INGV started operating in 2013 and became operational in October 2016. Here, after describing the NEAMTWS in the framework of the global effort coordinated by IOC/UNESCO, we focus on the tsunami hazard in the Mediterranean Sea. We then describe CAT-INGV mandate, functioning, and operational procedures. Furthermore, the article discusses the lessons learned from past events occurring in the Mediterranean Sea, such as the Kos-Bodrum in 2017 (Mw 6.6) and the Samos-Izmir in 2020 (Mw 7.0) earthquakes, which generated moderately damaging tsunamis. Based on these lessons, we discuss some potential improvements for the CAT-INGV and the NEAMTWS, including better seismic and sea level instrumental coverage. We emphasize the need for tsunami risk awareness raising, better preparation, and full implementation of the tsunami warning “last-mile” to foster the creation of a more integrated, interoperable, and sustainable risk reduction framework. If we aim to be better prepared for the next tsunami, these important challenges should be prioritized in the agenda of the IOC/UNESCO Member States and the European Commission.

International audience; Machine learning is becoming increasingly important in scientific and technological progress, due to its ability to create models that describe complex data and generalize well. The wealth of publicly-available seismic data nowadays requires automated, fast, and reliable tools to carry out a multitude of tasks, such as the detection of small, local earthquakes in areas characterized by sparsity of receivers. A similar application of machine learning, however, should be built on a large amount of labeled seismograms, which is neither immediate to obtain nor to compile. In this study we present a large dataset of seismograms recorded along the vertical, north, and east components of 1487 broad-band or very broad-band receivers distributed worldwide; this includes 629,095 3component seismograms generated by 304,878 local earthquakes and labeled as EQ, and 615,847 ones labeled as noise (AN). Application of machine learning to this dataset shows that a simple Convolutional Neural Network of 67,939 parameters allows discriminating between earthquakes and noise single-station recordings, even if applied in regions not represented in the training set. Achieving an accuracy of 96.7, 95.3, and 93.2% on training, validation, and test set, respectively, we prove that the large variety of geological and tectonic settings covered by our data supports the generalization capabilities of the algorithm, and makes it applicable to real-time detection of local events. We make the database publicly available, intending to provide the seismological and broader scientific community with a benchmark for time-series to be used as a testing ground in signal processing.

The goals of the DFT Working Group were:  Moving the discussion in the data community towards an agreed, upon suite of terms relevant to RDA groups with a focus on basic core model of related terms along with some basic principles that will harmonize the data organization solutions.  Fostering an effective RDA community culture by converging on essential terminology arising from agreed upon reference models. Based on a variety of data models and use cases presented by experts coming from different disciplines and about 120 interviews and interactions with different scientists and scientific departments, the DFT WG has composed a number of simple definitions for digital data in a registered1 domain based on group conceptualization and synthesis.

The evolution of High-Performance Computing (HPC) platforms enables the design and execution of progressively larger and more complex workflow applications in these systems. The complexity comes not only from the number of elements that compose the workflows but also from the type of computations they perform. While traditional HPC workflows target simulations and modelling of physical phenomena, current needs require in addition data analytics (DA) and artificial intelligence (AI) tasks. However, the development of these workflows is hampered by the lack of proper programming models and environments that support the integration of HPC, DA, and AI, as well as the lack of tools to easily deploy and execute the workflows in HPC systems. To progress in this direction, this paper presents use cases where complex workflows are required and investigates the main issues to be addressed for the HPC/DA/AI convergence. Based on this study, the paper identifies the challenges of a new workflow platform to manage complex workflows. Finally, it proposes a development approach for such a workflow platform addressing these challenges in two directions: first, by defining a software stack that provides the functionalities to manage these complex workflows; and second, by proposing the HPC Workflow as a Service (HPCWaaS) paradigm, which leverages the software stack to facilitate the reusability of complex workflows in federated HPC infrastructures. Proposals presented in this work are subject to study and development as part of the EuroHPC eFlows4HPC project., This work has received funding from the European High-Performance Computing Joint Undertaking (JU) under grant agreement No 955558. The JU receives support from the European Union’s Horizon 2020 research and innovation programme and Spain, Germany, France, Italy, Poland, Switzerland and Norway. In Spain, it has received complementary funding from MCIN/AEI/10.13039/501100011033, Spain and the European Union NextGenerationEU/PRTR (contracts PCI2021-121957, PCI2021-121931, PCI2021-121944, and PCI2021-121927). In Germany, it has received complementary funding from the German Federal Ministry of Education and Research (contracts 16HPC016K, 6GPC016K, 16HPC017 and 16HPC018). In France, it has received financial support from Caisse des dépôts et consignations (CDC) under the action PIA ADEIP (project Calculateurs). In Italy, it has been preliminary approved for complimentary funding by Ministero dello Sviluppo Economico (MiSE) (ref. project prop. 2659). In Norway, it has received complementary funding from the Norwegian Research Council, Norway under project number 323825. In Switzerland, it has been preliminary approved for complimentary funding by the State Secretariat for Education, Research, and Innovation (SERI), Norway. In Poland, it is partially supported by the National Centre for Research and Development under decision DWM/EuroHPCJU/4/2021. The authors also acknowledge financial support by MCIN/AEI /10.13039/501100011033, Spain through the “Severo Ochoa Programme for Centres of Excellence in R&D” under Grant CEX2018-000797-S, the Spanish Government, Spain (contract PID2019-107255 GB) and by Generalitat de Catalunya, Spain (contract 2017-SGR-01414). Anna Queralt is a Serra Húnter Fellow., With funding from the Spanish government through the ‘Severo Ochoa Centre of Excellence’ accreditation (CEX2018-000797-S).