Andreas Steinberg
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Andreas Steinberg has a total of 99 co-authors. Here are the top co-authors by number of shared publications:
- Peter Gaebler
BGR Federal Institute for Geosciences and Natural Resources, Stilleweg 2, 30655, Hannover, Germany
11 shared publications - Hannes Vasyura-Bathke
University of Potsdam
10 shared publications - Henriette Sudhaus
Institute of Geosciences Christian‐Albrechts‐University Kiel Germany
8 shared publications Henriette Sudhaus
Institute of Geosciences, Christian Albrechts University Kiel , 24118 Kiel, Germany
7 shared publications- Christoph Pilger
BGR, Hannover, Germany
6 shared publications - Lars Ceranna
BGR Hannover Germany
6 shared publications
Latest Publications
- Open Access
- 14 March 2026
- Published by Copernicus GmbH
Explosive yield estimates for shallow water explosions from moment magnitudes
The explosive yield- seismic magnitude relation of shallow submarine explosions are not well confined. Local agencies often use local seismic magnitude, such as the traditional richter scale, which are often not calibrated for submarine environments. The importance of an estimated explosive yield in TnT equivalent becomes obvious when security concerns arise. After the North Stream events a number of very differing magnitude were presented by several seismological surveys and therefore the related yield estimates varied a lot. This lead to derived estimates ranging from tens of kg to hundreds of kg TnT equivalent explosive used in the North Stream explosions, giving different plausible scenarios for potential perpetrators.We present an relatively simple and fast approach to use the comparison of recorded and forward modelled envelope and cepstral information to derive the moment magnitude of several large and small submarine explosion in the Baltic sea. Moment magnitudes are more robust in comparison to local magnitudes. We asses the performance of this approach by relating the moment magnitudes to yield estimates from known explosive events. We use events from the Baltic sea, including events from offshore Bornholm from September 2025 with around 400kg TnT yield, recorded at local and regional distances up to 500km.Infrasound recordings of stations in Germany are in good correlation with the seismic recordings, showing the possibility of combined energy release estimates. The strong importance of the source depth and shallow submarine geology is highlighted by the modelling results, providing still a large uncertainty range for unknown sources. We do find in general a good agreement between the estimated yield and actual yield for the known sources.
- 14 March 2026
- Published by Copernicus GmbH
Open AccessExplosive yield estimates for shallow water explosions from moment magnitudes
The explosive yield- seismic magnitude relation of shallow submarine explosions are not well confined. Local agencies often use local seismic magnitude, such as the traditional richter scale, which are often not calibrated for submarine environments. The importance of an estimated explosive yield in TnT equivalent becomes obvious when security concerns arise. After the North Stream events a number of very differing magnitude were presented by several seismological surveys and therefore the related yield estimates varied a lot. This lead to derived estimates ranging from tens of kg to hundreds of kg TnT equivalent explosive used in the North Stream explosions, giving different plausible scenarios for potential perpetrators.We present an relatively simple and fast approach to use the comparison of recorded and forward modelled envelope and cepstral information to derive the moment magnitude of several large and small submarine explosion in the Baltic sea. Moment magnitudes are more robust in comparison to local magnitudes. We asses the performance of this approach by relating the moment magnitudes to yield estimates from known explosive events. We use events from the Baltic sea, including events from offshore Bornholm from September 2025 with around 400kg TnT yield, recorded at local and regional distances up to 500km.Infrasound recordings of stations in Germany are in good correlation with the seismic recordings, showing the possibility of combined energy release estimates. The strong importance of the source depth and shallow submarine geology is highlighted by the modelling results, providing still a large uncertainty range for unknown sources. We do find in general a good agreement between the estimated yield and actual yield for the known sources.
- Open Access
- 14 March 2026
- Published by Copernicus GmbH
Observation and analysis of core and secondary teleseismic phases at broadband stations in Germany using array methods
In its routine analysis the Federal Seismological Survey at BGR evaluates local seismic earthquakes in Germany as well as teleseismic events. For the analysis of teleseismic events we use the seismic stations in Germany and in some cases stations from surrounding countries. Because the stations are far from the events, we apply array methods (f-k analysis) to identify distinct seismic phases and locate events. We use waveforms of the densely spaced Gräfenberg array sites (GRF) with an interstation distance of 10 to 15 km or of the large aperture German Regional Seismic Network (GRSN) stations if coherent phase picking is possible.The majority of detected and evaluated phases are first-arriving phases, such as P phases and, at more distant epicentre distances, Pdiff and PKP phases. However, in the case of stronger events, we can also detect later phases, such as PP, PS, SS, PcP, ScS, SKS and others. These secondary phases are identified well via slowness, azimuth and travel time.Here, we show some striking examples with PKP phases as well as later phases.Events with PKP phases are numerous and mostly located within the subduction zones at Fiji and Tonga. Due to the epicentral distances of around 145 degrees the German stations are close to the caustic of PKP branches. Therefore, PKPdf, PKPbc, and PKPab often show strong amplitudes. In some cases all three branches as well as their corresponding depth phases are visible and can be picked.Some of the later phases, such as SKP, are rarely observed. We investigate to what extent the magnitude, the focal depth and also the focal mechanism are responsible for the amplitude of the phase and thus for their visibility at German stations. In order to determine the influence of the focal mechanism, we use the solutions of international agencies and calculate radiation coefficients for the respective phase in the direction of the stations under consideration. This approach may help to decide whether an observed depth phase is a pP or sP phase, for example, and thus enables a better depth determination.Another special feature are very late phases, such as PKKP. They run on the long path around the Earth to the station and have an azimuth shifted by 180 degrees to the short path. These phases may be misinterpreted as independent events. Here, we show an example.
- 14 March 2026
- Published by Copernicus GmbH
Open AccessObservation and analysis of core and secondary teleseismic phases at broadband stations in Germany using array methods
In its routine analysis the Federal Seismological Survey at BGR evaluates local seismic earthquakes in Germany as well as teleseismic events. For the analysis of teleseismic events we use the seismic stations in Germany and in some cases stations from surrounding countries. Because the stations are far from the events, we apply array methods (f-k analysis) to identify distinct seismic phases and locate events. We use waveforms of the densely spaced Gräfenberg array sites (GRF) with an interstation distance of 10 to 15 km or of the large aperture German Regional Seismic Network (GRSN) stations if coherent phase picking is possible.The majority of detected and evaluated phases are first-arriving phases, such as P phases and, at more distant epicentre distances, Pdiff and PKP phases. However, in the case of stronger events, we can also detect later phases, such as PP, PS, SS, PcP, ScS, SKS and others. These secondary phases are identified well via slowness, azimuth and travel time.Here, we show some striking examples with PKP phases as well as later phases.Events with PKP phases are numerous and mostly located within the subduction zones at Fiji and Tonga. Due to the epicentral distances of around 145 degrees the German stations are close to the caustic of PKP branches. Therefore, PKPdf, PKPbc, and PKPab often show strong amplitudes. In some cases all three branches as well as their corresponding depth phases are visible and can be picked.Some of the later phases, such as SKP, are rarely observed. We investigate to what extent the magnitude, the focal depth and also the focal mechanism are responsible for the amplitude of the phase and thus for their visibility at German stations. In order to determine the influence of the focal mechanism, we use the solutions of international agencies and calculate radiation coefficients for the respective phase in the direction of the stations under consideration. This approach may help to decide whether an observed depth phase is a pP or sP phase, for example, and thus enables a better depth determination.Another special feature are very late phases, such as PKKP. They run on the long path around the Earth to the station and have an azimuth shifted by 180 degrees to the short path. These phases may be misinterpreted as independent events. Here, we show an example.
- Open Access
- 18 March 2025
- Published by Copernicus GmbH
Combined (DAS/Seismometers) Seismic Analysis of an Underwater Explosion in the Baltic Sea
In November 2024, an Ml ~1.5-2.0 underwater explosion occurred in Danish waters north of Bornholm Island. It was recorded by the Danish and Greenlandic national seismic networks as well as the Swedish national network. In addition, the event was captured by a nearby Distributed Acoustic Sensing (DAS) system deployed along a 120 km-long underwater fiber-optic cable. We investigated the event location by integrating data from these complementary recording systems, systematically assessing the trade-off between the number of DAS channels with respect to the number of permanent seismometers and the quality of the picks. Our results indicate horizontal uncertainties on the order of 2–3 km for the final event location. To further constrain the depth of the explosion and its yield, we conducted a spectral analysis of the seismograms, joined with a non-linear inversion of the P-waveforms at the closest station. The inferred source parameters are consistent with the known water depth and velocity at the explosion site, revealing that the event probably involved two distinct detonations located less than 10 m above the seafloor, each with an approximate 30 kg TNT equivalent yield. These findings highlight the advantages of combining conventional permanent seismic instrumentation with underwater DAS, thereby improving the detection and characterization of anthropogenic seismic sources and offering enhanced protection for critical submarine infrastructures.
- 18 March 2025
- Published by Copernicus GmbH
Open AccessCombined (DAS/Seismometers) Seismic Analysis of an Underwater Explosion in the Baltic Sea
In November 2024, an Ml ~1.5-2.0 underwater explosion occurred in Danish waters north of Bornholm Island. It was recorded by the Danish and Greenlandic national seismic networks as well as the Swedish national network. In addition, the event was captured by a nearby Distributed Acoustic Sensing (DAS) system deployed along a 120 km-long underwater fiber-optic cable. We investigated the event location by integrating data from these complementary recording systems, systematically assessing the trade-off between the number of DAS channels with respect to the number of permanent seismometers and the quality of the picks. Our results indicate horizontal uncertainties on the order of 2–3 km for the final event location. To further constrain the depth of the explosion and its yield, we conducted a spectral analysis of the seismograms, joined with a non-linear inversion of the P-waveforms at the closest station. The inferred source parameters are consistent with the known water depth and velocity at the explosion site, revealing that the event probably involved two distinct detonations located less than 10 m above the seafloor, each with an approximate 30 kg TNT equivalent yield. These findings highlight the advantages of combining conventional permanent seismic instrumentation with underwater DAS, thereby improving the detection and characterization of anthropogenic seismic sources and offering enhanced protection for critical submarine infrastructures.
- Open Access
- 18 March 2025
- Published by Copernicus GmbH
Deep learning based phase picking on seismological IMS stations
We present our work on training and the application of deep learning algorithms for the automated phase picking of body waves on the the IMS network. We train new IMS data based seismic phase pickers from both EQT and PhaseNet architectures. Phase picking is a necessary step before event localization and characterization and deep learning based models have been proven to perform well at this task. PhaseNet and EQTransformer are two prominent state-of-the-art phase picking algorithms that have been retrained on several different datasets.Waveform data from primary and auxiliary stations is used in the training and evaluation. For training we use good quality picks from REB events between 2013 until 2023. We evaluate the performance in comparison with unseen evaluation REB phase picks and manual phase picks. We compare the performance with applying other pre-trained phase pickers to the IMS data to determine if already pre-trained models can be used satisfactory out of the box for seismological IMS data. We also evaluate the generalization ability of the two IMS data trained models by applying them to other non IMS seismological stations of the German Regional Seismic Network (GRSN). We further train new phase pickers based on the PhaseNet architecture and a database of 20 years listed in the earthquake catalog of BGR. The models are trained and evaluated with manual phase picks of BGR analysts. We compare the performance of the newly trained models by also applying other pre-trained PhaseNet and EQTransformer based phase pickers on unseen data. We determine if existing pre-trained models can satisfactorily be used out of the box for phase picking on waveforms of the GRSN.
- 18 March 2025
- Published by Copernicus GmbH
Open AccessDeep learning based phase picking on seismological IMS stations
We present our work on training and the application of deep learning algorithms for the automated phase picking of body waves on the the IMS network. We train new IMS data based seismic phase pickers from both EQT and PhaseNet architectures. Phase picking is a necessary step before event localization and characterization and deep learning based models have been proven to perform well at this task. PhaseNet and EQTransformer are two prominent state-of-the-art phase picking algorithms that have been retrained on several different datasets.Waveform data from primary and auxiliary stations is used in the training and evaluation. For training we use good quality picks from REB events between 2013 until 2023. We evaluate the performance in comparison with unseen evaluation REB phase picks and manual phase picks. We compare the performance with applying other pre-trained phase pickers to the IMS data to determine if already pre-trained models can be used satisfactory out of the box for seismological IMS data. We also evaluate the generalization ability of the two IMS data trained models by applying them to other non IMS seismological stations of the German Regional Seismic Network (GRSN). We further train new phase pickers based on the PhaseNet architecture and a database of 20 years listed in the earthquake catalog of BGR. The models are trained and evaluated with manual phase picks of BGR analysts. We compare the performance of the newly trained models by also applying other pre-trained PhaseNet and EQTransformer based phase pickers on unseen data. We determine if existing pre-trained models can satisfactorily be used out of the box for phase picking on waveforms of the GRSN.
- Open Access
- 20 January 2025
- Published by Copernicus GmbH
Source analysis of the 2022 Nord Stream and 2023 Balticconnector underwater explosions
On 26 September 2022 two seismic events near the Danish island of Bornholm in the Baltic Sea were detected. The first event with a magnitude Mw 2.3 occurred at 00:03 UTC 40 km east-southeast of Bornholm. The determined location and the origin time of the event are consistent with data of the pressure decrease on one of the Nord Stream 2 pipelines. Another sequence of events occurred 17 hours later at 17:03 UTC around 60 km north-east of Bornholm with a maximum magnitude of Mw 2.7. It consists of three closely successive, but separable, single events. Using relative localisation methods and the gas pressure inside the pipeline recorded at the landing site in Germany, we can assign the epicentres of the three events to the locations of the leaks in the pipelines of Nord Stream 1 and 2.Based on comparable events in the region, which include both tectonic earthquakes and explosions, the explosive character of the investigated Nord Stream events can be verified. Infrasound signals associated with the destruction of the Nord Stream pipelines were recorded at two stations (I26DE in the Bavarian Forest and IKUDE near Kühlungsborn) in Germany. Particularly after the event sequence at 17:03 UTC, distinctive signals were registered whose characteristics indicate an explosive event with subsequent gas leakage at the surface.Our modelling of the sources shows that the measured seismic signals can sufficiently be explained by the instantaneous gas release. Synthetic seismograms for such a source and a subsurface model adapted for the study area show high consistency with the measured signals. Based on the released energy and the characteristics of the recorded waveforms, we conclude that the impulsive gas release from the burst gas pipes constitutes the dominant part of the signal source. The model places an upper limit of approximately 50 kg TNT equivalent on the yield of the chemical explosive component of the events, but we note that smaller yields may also be consistent with the data.We also carried out an analysis of the seismic signals of the event on the Balticconnector pipeline between Finland and Estonia on 8 October 2023 and found that again the instantaneous gas release can sufficiently explain the observed data. This supports a possible mechanical cause of the damage.
- 20 January 2025
- Published by Copernicus GmbH
Open AccessSource analysis of the 2022 Nord Stream and 2023 Balticconnector underwater explosions
On 26 September 2022 two seismic events near the Danish island of Bornholm in the Baltic Sea were detected. The first event with a magnitude Mw 2.3 occurred at 00:03 UTC 40 km east-southeast of Bornholm. The determined location and the origin time of the event are consistent with data of the pressure decrease on one of the Nord Stream 2 pipelines. Another sequence of events occurred 17 hours later at 17:03 UTC around 60 km north-east of Bornholm with a maximum magnitude of Mw 2.7. It consists of three closely successive, but separable, single events. Using relative localisation methods and the gas pressure inside the pipeline recorded at the landing site in Germany, we can assign the epicentres of the three events to the locations of the leaks in the pipelines of Nord Stream 1 and 2.Based on comparable events in the region, which include both tectonic earthquakes and explosions, the explosive character of the investigated Nord Stream events can be verified. Infrasound signals associated with the destruction of the Nord Stream pipelines were recorded at two stations (I26DE in the Bavarian Forest and IKUDE near Kühlungsborn) in Germany. Particularly after the event sequence at 17:03 UTC, distinctive signals were registered whose characteristics indicate an explosive event with subsequent gas leakage at the surface.Our modelling of the sources shows that the measured seismic signals can sufficiently be explained by the instantaneous gas release. Synthetic seismograms for such a source and a subsurface model adapted for the study area show high consistency with the measured signals. Based on the released energy and the characteristics of the recorded waveforms, we conclude that the impulsive gas release from the burst gas pipes constitutes the dominant part of the signal source. The model places an upper limit of approximately 50 kg TNT equivalent on the yield of the chemical explosive component of the events, but we note that smaller yields may also be consistent with the data.We also carried out an analysis of the seismic signals of the event on the Balticconnector pipeline between Finland and Estonia on 8 October 2023 and found that again the instantaneous gas release can sufficiently explain the observed data. This supports a possible mechanical cause of the damage.
- Open Access
- 31 December 2024
- Published by Springer Nature in Acta Geophysica
- Vol. 73 (1) , 577-592
Probabilistic seismic hazard assessment associated with induced seismicity at geothermal sites in the Upper Rhine Graben (Southern Germany)
The extraction of geothermal energy is associated with induced seismicity. Depending on the extraction parameters, such as injection pressure and volume, the induced seismicity is time-dependent. We investigate the case of the two geothermal power plants of Insheim and Landau, which are located in the Upper Rhine Graben in Southwest Germany. The induced seismicity was observed by a local seismic network consisting of a total of 19 stations, resulting in an earthquake catalog comprising 930 events for the Landau reservoir and 1985 events for the Insheim reservoir, both between 2012 and 2022. Based on this earthquake catalog, seismic source areas are defined for both reservoirs, and a probabilistic seismic hazard assessment (PSHA) is performed. Using temporal subsets of the earthquake catalog, PSHA can also be performed for shorter time ranges, resulting in larger expected PGV values in times of higher induced seismicity. The seismic velocity profiles obtained by site effect studies based on ambient seismic noise measurements highlight relatively large variations of the site effects on short scales in the area. The consequences of these lateral variations on the seismic hazard assessment are also discussed.
- 31 December 2024
- Published by Springer Nature in Acta Geophysica
- Vol. 73 (1) , 577-592
Open AccessProbabilistic seismic hazard assessment associated with induced seismicity at geothermal sites in the Upper Rhine Graben (Southern Germany)
The extraction of geothermal energy is associated with induced seismicity. Depending on the extraction parameters, such as injection pressure and volume, the induced seismicity is time-dependent. We investigate the case of the two geothermal power plants of Insheim and Landau, which are located in the Upper Rhine Graben in Southwest Germany. The induced seismicity was observed by a local seismic network consisting of a total of 19 stations, resulting in an earthquake catalog comprising 930 events for the Landau reservoir and 1985 events for the Insheim reservoir, both between 2012 and 2022. Based on this earthquake catalog, seismic source areas are defined for both reservoirs, and a probabilistic seismic hazard assessment (PSHA) is performed. Using temporal subsets of the earthquake catalog, PSHA can also be performed for shorter time ranges, resulting in larger expected PGV values in times of higher induced seismicity. The seismic velocity profiles obtained by site effect studies based on ambient seismic noise measurements highlight relatively large variations of the site effects on short scales in the area. The consequences of these lateral variations on the seismic hazard assessment are also discussed.
- Open Access
- 2 May 2024
- Published by Springer Nature in Pure and Applied Geophysics
- Vol. 182 (12) , 4977-4997
Deep Neural Networks Based Denoising of Regional Seismic Waveforms and Impact on Analysis of North Korean Nuclear Tests
We test a deep learning based denoising autoencoder algorithm on regional and teleseismic seismological and hydroacoustic datasets, which we compile from the International Monitoring System of the Comprehensive Nuclear-Test-Ban Treaty Organisation. We focus on stations which can be relevant to investigate North Korean nuclear tests. Denoising of waveform records using autoencoder techniques potentially enables improved signal detection and processing due to lowered signal-to-noise ratios. We train and compare the performance of several different denoising autoencoder models, for short- and long waveform periods, trained on the complete station network as well as on individual stations. We investigate if the denoised waveform signals are useful for seismic source analysis and if they can still be reliably used in downstream analysis for further inferences on the seismic source type, i.e. seismic moment tensor analysis. The declared North Korean nuclear tests are a suitable benchmark test set, as they have extensively been researched and their source type and location might be assumed known. Verification of the source type is of particular interest for potential nuclear tests under international law. We find that care needs to be taken using the denoised waveform data, as a slight bias is introduced in the seismic moment tensor analysis. However we also find promising results hinting at possible future use of the technique for standard analyses, as it improves the investigation of smaller events. Autoencoder based denoising techniques could be employed in future routine frameworks to increase earthquake catalog completeness and possibly aid in detecting smaller potential treaty relevant events.
- 2 May 2024
- Published by Springer Nature in Pure and Applied Geophysics
- Vol. 182 (12) , 4977-4997
Open AccessDeep Neural Networks Based Denoising of Regional Seismic Waveforms and Impact on Analysis of North Korean Nuclear Tests
We test a deep learning based denoising autoencoder algorithm on regional and teleseismic seismological and hydroacoustic datasets, which we compile from the International Monitoring System of the Comprehensive Nuclear-Test-Ban Treaty Organisation. We focus on stations which can be relevant to investigate North Korean nuclear tests. Denoising of waveform records using autoencoder techniques potentially enables improved signal detection and processing due to lowered signal-to-noise ratios. We train and compare the performance of several different denoising autoencoder models, for short- and long waveform periods, trained on the complete station network as well as on individual stations. We investigate if the denoised waveform signals are useful for seismic source analysis and if they can still be reliably used in downstream analysis for further inferences on the seismic source type, i.e. seismic moment tensor analysis. The declared North Korean nuclear tests are a suitable benchmark test set, as they have extensively been researched and their source type and location might be assumed known. Verification of the source type is of particular interest for potential nuclear tests under international law. We find that care needs to be taken using the denoised waveform data, as a slight bias is introduced in the seismic moment tensor analysis. However we also find promising results hinting at possible future use of the technique for standard analyses, as it improves the investigation of smaller events. Autoencoder based denoising techniques could be employed in future routine frameworks to increase earthquake catalog completeness and possibly aid in detecting smaller potential treaty relevant events.
- Open Access
- 12 January 2024
- Published by McGill University Library and Archives in Seismica
- Vol. 3 (1)
Discontinuous transtensional rupture during the Mw 7.2 1995 Gulf of Aqaba earthquake
The Gulf of Aqaba earthquake occurred on 22 November 1995 in the Northern Red Sea and is the largest instrumentally recorded earthquake in the region to date. The event was extensively studied during the initial years following its occurrence. However, it remained unclear which of the many faults in the gulf were activated during the earthquake. We present results from multi-array back projection that we use to inform Bayesian kinematic rupture models constrained by geodetic and teleseismic data. Our results indicate that most of the moment release was on the Aragonese fault via left-lateral strike slip and shallow normal faulting that may have been dynamically triggered by an early rupture phase on the Arnona fault. We also identified a predominantly normal fault-segment on the eastern shore of the gulf that was activated in the event. We dismiss the previously proposed hypothesis of a co-seismic sub-event on the western shore of the gulf and confirm that observed deformation can be rather attributed to post-seismic activity. In conclusion, the gulf shows many signs of active tectonic extension. Therefore, more events close to the shorelines are to be expected in the future and should be considered conducting infrastructure projects in the region.
- 12 January 2024
- Published by McGill University Library and Archives in Seismica
- Vol. 3 (1)
Open AccessDiscontinuous transtensional rupture during the Mw 7.2 1995 Gulf of Aqaba earthquake
The Gulf of Aqaba earthquake occurred on 22 November 1995 in the Northern Red Sea and is the largest instrumentally recorded earthquake in the region to date. The event was extensively studied during the initial years following its occurrence. However, it remained unclear which of the many faults in the gulf were activated during the earthquake. We present results from multi-array back projection that we use to inform Bayesian kinematic rupture models constrained by geodetic and teleseismic data. Our results indicate that most of the moment release was on the Aragonese fault via left-lateral strike slip and shallow normal faulting that may have been dynamically triggered by an early rupture phase on the Arnona fault. We also identified a predominantly normal fault-segment on the eastern shore of the gulf that was activated in the event. We dismiss the previously proposed hypothesis of a co-seismic sub-event on the western shore of the gulf and confirm that observed deformation can be rather attributed to post-seismic activity. In conclusion, the gulf shows many signs of active tectonic extension. Therefore, more events close to the shorelines are to be expected in the future and should be considered conducting infrastructure projects in the region.
- Open Access
- 6 November 2023
- Published by Springer Nature in Research Square
Neural Network Based Denoising of Regional Seismic Waveforms and Impact on Analysis of North Korean Nuclear Tests
We test a deep learning based denoising autoencoder algorithm on regional and teleseismic seismological and hydroacoustic datasets, which we compile from the International Monitoring System of the Comprehensive Nuclear-Test-Ban Treaty Organisation. We focus on stations which can be relevant to investigate North Korean nuclear tests. Denoising of waveform records using autoencoder techniques potentially enables improved signal detection and processing due to lowered signal-to-noise ratios. We train and compare the performance of several different denoising autoencoder models, for short- and long waveform periods, trained on the complete station network as well as on individual stations. We investigate if the denoised waveform signals are useful for seismic source analysis and if they can still be reliably used in downstream analysis for further inferences on the seismic source type, i.e. seismic moment tensor analysis. The declared North Korean nuclear tests are a suitable benchmark test set, as they have extensively been researched and their source type and location might be assumed known. Verification of the source type is of particular interest for potential nuclear tests under international law. We find that care needs to be taken using the denoised waveform data, as a slight bias is introduced in the seismic moment tensor analysis. However we also find promising results hinting at possible future use of the technique for standard analyses, as it improves the investigation of smaller events. Autoencoder based denoising techniques could be employed in future routine frameworks to increase earthquake catalog completeness and possibly aid in detecting smaller potential treaty relevant events.
- 6 November 2023
- Published by Springer Nature in Research Square
Open AccessNeural Network Based Denoising of Regional Seismic Waveforms and Impact on Analysis of North Korean Nuclear Tests
We test a deep learning based denoising autoencoder algorithm on regional and teleseismic seismological and hydroacoustic datasets, which we compile from the International Monitoring System of the Comprehensive Nuclear-Test-Ban Treaty Organisation. We focus on stations which can be relevant to investigate North Korean nuclear tests. Denoising of waveform records using autoencoder techniques potentially enables improved signal detection and processing due to lowered signal-to-noise ratios. We train and compare the performance of several different denoising autoencoder models, for short- and long waveform periods, trained on the complete station network as well as on individual stations. We investigate if the denoised waveform signals are useful for seismic source analysis and if they can still be reliably used in downstream analysis for further inferences on the seismic source type, i.e. seismic moment tensor analysis. The declared North Korean nuclear tests are a suitable benchmark test set, as they have extensively been researched and their source type and location might be assumed known. Verification of the source type is of particular interest for potential nuclear tests under international law. We find that care needs to be taken using the denoised waveform data, as a slight bias is introduced in the seismic moment tensor analysis. However we also find promising results hinting at possible future use of the technique for standard analyses, as it improves the investigation of smaller events. Autoencoder based denoising techniques could be employed in future routine frameworks to increase earthquake catalog completeness and possibly aid in detecting smaller potential treaty relevant events.
- Open Access
- 18 May 2023
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 235 (1) , 48-73
The January 2022 Hunga Volcano explosive eruption from the multitechnological perspective of CTBT monitoring
SUMMARY: The massive eruption of the Hunga Volcano on 15 January 2022 provides an ideal test case for reviewing established methods to discriminate and analyse source processes. Discriminating source mechanisms and identifying their origins is a key task when analysing suspicious events in the frame of the Comprehensive Nuclear-Test-Ban Treaty (CTBT). Earthquakes and explosions can be distinguished in some cases using well established methods such as inversion for the seismic moment tensor. In more complex cases the combination of analyses of the seismic, infrasonic and hydroacoustic waveform content can be of help. More challenging is the discrimination of the specific kind of explosive source such as a nuclear test and a volcano eruption based on the data from the three waveform technologies alone. Here, we apply standard techniques destined to analyse relevant events in the frame of the CTBT, that is all three waveform technologies (seismology, infrasound and hydroacoustic) and atmospheric transport modelling of radionuclides. We investigate the potential of standard analysis methods to discriminate a source and identify their possible weaknesses. We show that the methods applied here work very well to identify, investigate and discriminate an explosive event. During discrimination we could not only exclude a shear-source (i.e. earthquake) but also distinguish the volcanic explosion in contrast to a man-made explosion. However, some tasks remain difficult with the available methods. These tasks include the reliable estimation of the strength of a non-shear event and thereupon a yield estimation of a possibly CTBT relevant event. In addition to evaluating our methods, we could relate our results with specific phases of the eruption process providing a more detailed insight of what happened. Our investigations of the eruption details only provide a starting point for further in-depth analysis. However, they underline the importance of the Hunga eruption event for science. The huge amount of observations provide a unique opportunity for knowledge gain in several subdisciplines of the geosciences. In addition, although not being a nuclear test, it also provides a useful and important data set for further developing multitechnology analyses in the frame of the CTBT.
- 18 May 2023
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 235 (1) , 48-73
Open AccessThe January 2022 Hunga Volcano explosive eruption from the multitechnological perspective of CTBT monitoring
SUMMARY: The massive eruption of the Hunga Volcano on 15 January 2022 provides an ideal test case for reviewing established methods to discriminate and analyse source processes. Discriminating source mechanisms and identifying their origins is a key task when analysing suspicious events in the frame of the Comprehensive Nuclear-Test-Ban Treaty (CTBT). Earthquakes and explosions can be distinguished in some cases using well established methods such as inversion for the seismic moment tensor. In more complex cases the combination of analyses of the seismic, infrasonic and hydroacoustic waveform content can be of help. More challenging is the discrimination of the specific kind of explosive source such as a nuclear test and a volcano eruption based on the data from the three waveform technologies alone. Here, we apply standard techniques destined to analyse relevant events in the frame of the CTBT, that is all three waveform technologies (seismology, infrasound and hydroacoustic) and atmospheric transport modelling of radionuclides. We investigate the potential of standard analysis methods to discriminate a source and identify their possible weaknesses. We show that the methods applied here work very well to identify, investigate and discriminate an explosive event. During discrimination we could not only exclude a shear-source (i.e. earthquake) but also distinguish the volcanic explosion in contrast to a man-made explosion. However, some tasks remain difficult with the available methods. These tasks include the reliable estimation of the strength of a non-shear event and thereupon a yield estimation of a possibly CTBT relevant event. In addition to evaluating our methods, we could relate our results with specific phases of the eruption process providing a more detailed insight of what happened. Our investigations of the eruption details only provide a starting point for further in-depth analysis. However, they underline the importance of the Hunga eruption event for science. The huge amount of observations provide a unique opportunity for knowledge gain in several subdisciplines of the geosciences. In addition, although not being a nuclear test, it also provides a useful and important data set for further developing multitechnology analyses in the frame of the CTBT.