Seismic waves propagating through the Earth sample its structure, carry information about its fabrics and physical characteristics and record its present-day state and evolution. In the past, several velocity discontinuities within the radial Earth, which separate its fundamental regions, were retrieved. The lower mantle-core boundary was named as Gutenberg discontinuity in recognition of the Gutenberg’s discovery of the Earth’s core in 1913. This discontinuity relates to the abrupt decrease in P-velocity and diminishing of S-waves in the liquid core. In present-day terminology, the Gutenberg discontinuity is associated with the bottom of the D’’ layer. An area of low velocities in the Earth’s upper mantle denoted as G-discontinuity, has related to Gutenberg’s name until now. The low velocity zone exists just below the oceanic lithosphere, and its characteristics are often used globally in studies of lithosphere thickness in the view of modern plate tectonics. Gutenberg’s Seismicity of the Earth (1941) became a major influence in later scientists’ efforts to describe the theory of plate tectonics. The accuracy and validity of the Earth models depend on data quality and coverage, i.e., earthquake foci - seismic station ray distribution within the Earth volume studied. Small-sized to large-scale international passive seismic experiments, operated during several recent decades, recorded an unprecedented huge amount of high-quality data, which along with new techniques and computational facilities represent a big step forward in our knowledge of the Earth’s structure. However, many questions still remain unanswered and require further research. Current close international cooperation among seismologists involved in the experiments follow the spirit of Beno Gutenberg’s action as a driving force behind the acceptance of seismology as an international science of earthquake detection and the Earth studies.

We present models of the European lithosphere derived from the propagation of body waves, shear-wave splitting and radial and azimuthal anisotropy of surface waves, including ambient noise. Data for individual studies has been collected from international seismological databases (ISC, EIDA) and from several passive experiments we have organized or participated in. Initial isotropic models are upgraded into anisotropic ones, following the fundamental condition that seismic anisotropy is a 3D phenomenon and thus it has to be evaluated in 3D to get more realistic images of the Earth. We invert/interpret jointly anisotropic parameters of independent observables (directional variations of P-wave travel times, shear-wave splitting parameters) which leads to 3D self-consistent anisotropic models of the continental lithosphere with tilted symmetry axes and characteristic domain-like structure. The individual domains at size from several tenths to several hundreds of kilometers are often sharply bounded and of different thicknesses. We interpret the often sharply bounded domains with systematically oriented dipping fabrics in the continental mantle lithosphere by successive subductions of ancient oceanic plates and their accretions enlarging primordial continent cores. Consequent continental break-ups and assemblages of wandering micro-plates preserve fossil anisotropic fabrics and create patchwork structures of the present-day continents. Supporting arguments for such model exist in petrological and geochemical studies (Babuska and Plomerova, 2020).

How to cite: Plomerová, J.: Seismic images of the continental lithosphere, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9250, https://doi.org/10.5194/egusphere-egu24-9250, 2024.

80% of earthquakes occur underwater, so ocean-bottom seismometers (OBS) are crucial for improving our understanding of earthquake source mechanics along unexplored offshore faults, and fillling key gaps in our images of the deeper solid Earth. Even for ocean islands and island arcs, land stations alone struggle to image underlying structures. Broadband OBSs have been through many design iterations, but many OBS deployments now yield high data recovery rates (>90%). 

Even though my first OBS deployment experience left me feeling seasick, I have since continued to seismically explore the oceans, taking part in several OBS projects. In this talk, I will focus on my recent results from experiments across the Atlantic Ocean. Compared to the faster-spreading and subducting Pacific lithosphere, the less well-studied Atlantic offers a key endmember for refining our knowledge of global tectonics and associated hazards.

In the Lesser Antilles subduction zone, subducting Atlantic lithosphere is heterogeneously hydrated. Local earthquakes recorded by OBSs (VoiLA experiment), allowed me to image seismic attenuation to map fluid and melt pathways through the slab and mantle wedge, showing how slab fluids precondition melt generation and volcanism in arc settings. In the mid-Atlantic, long transform faults can host large M~7 earthquakes in ultra-wide (20-30 km thick) fault zones, allowing a uniquely macro-scale view of how damage zones control seismogenesis. In 2017, OBSs (PI-LAB experiment) recorded a nearby Mw 7.1 earthquake on the Romanche transform fault, triggering detailed teleseismic analysis that show back-propagating rupture fronts, which have since been seen during the 2023 M7.8 Türkiye earthquake. More recently, I analysed a seismic swarm and dyke intrusion in the Azores, which lies on a diffuse transtensional plate boundary. Here, a temporary OBS network (UPFLOW project) installed around the uniquely narrow island of São Jorge yields high-resolution seismicity locations that shed light on magma inflow and drainage along pre-existing faults.

Overall, OBS experiments yield fascinating results, but these results come from vast team efforts, particularly from ship crews and OBS technicians, that often go uncredited. We need to work harder to ensure the long-term sustainability of data from these expensive, often publicly-funded projects, with OBS-specific data preprocessing complications a partial barrier to this.

How to cite: Hicks, S. P.: Uncovering the tectonic secrets of the Atlantic with broadband ocean-bottom seismology, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14636, https://doi.org/10.5194/egusphere-egu24-14636, 2024.

Georgia is situated in the Caucasus region, which is one of the most seismically active regions in the Alpine-Himalayan collision belt. Analysis of the historical and instrumental seismology of this region shows that it is still of moderate seismicity. The seismicity of the area reflects the general tectonics of the region.

Recently, number of seismic stations and earthquake records in Georgia significantly increased. Thus, we can run more detailed studies regarding ground motion prediction. 

Ground motion prediction equations (GMPEs) relate ground motion intensity measures to variables describing earthquake source, path, and site effects. In this study ground motion prediction equations are obtained by classical, statistical way, regression analysis. Also, new data and new features such as local soil conditions, fault types, etc. were considered for analysis. In the study models are obtained for PGA (horizontal and vertical), 5%-damped pseudo-absolute-acceleration spectra (SA) are described for periods between 0.01 s and 10 s (for both vertical and horizontal components).

Next stage was to assess the standard deviation and its minimization. Fuzzy Analysis gives a possibility of making optimal decision when available data is insufficient and cannot represent real situation. In our case it is quite difficult to explain all physical processes related to earthquakes. However, it is very important to consider all processes during the hazard assessment. Also, during GMPE assessment it is very difficult to consider site effect very precisely because available data is still insufficient. In this case usage of Fuzzy Analysis is the best solution. We constructed membership functions based on shear wave velocity measurements for each site class. Site classifications were done according to Eurocode8. At the end a significant reduction of uncertainties (~30-40%) was observed.

How to cite: Jorjiashvili, N., Shengelia, I., Godoladze, T., Gunia, I., and Akubardia, D.: Dealing with uncertainties related to ground motion prediction models for Georgia, Caucasus Region., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-169, https://doi.org/10.5194/egusphere-egu24-169, 2024.

For decades, the seismological community has debated the scaling relationships of earthquake sources. The debate centers around whether the scaled energy (E_R/M₀) remains uniform across all magnitudes, indicating self-similarity, or if there is an increase in scaled energy with seismic moment, M₀. To contribute to this discussion, we analyzed coda derived source displacement spectra of 303 local earthquakes that occurred in and around the segments of the North Anatolian Fault Zone (NAFZ) within the Sea of Marmara. Our database includes digital waveform recordings of the events that were occurred between 2018 and 2020 (2.5≤ M_L ≤5.7 within a radius of 200 km) and were recorded at 49 seismic stations operated by the Kandilli Observatory and Earthquake Research Institute (KOERI) in the study area. We employed a joint inversion technique to optimize source-, path-, and site-specific factors simultaneously. This was achieved by comparing the observed coda envelope with its physically derived representative synthetic coda envelope based on Radiative Transfer Theory. Our inversion process, conducted across various frequency bands, enabled us to make reliable coda-based seismic moment (M₀) and moment magnitude estimates (M_w-coda) consistent with local catalogue magnitudes. The variation of the scaled energy (E_R/M₀) calculated from the total seismic radiated energy (E_R) using coda-derived source displacement spectra for each event tends to increase with seismic moment across most magnitude ranges. This indicates that the crustal earthquakes with M_w-coda 2.5 and M_w-coda 5.7 in this laterally heterogeneous region are likely to follow non-self-similarity. Our findings imply different rupture dynamics working for large earthquakes than small ones and relatively more efficient seismic energy radiation for larger earthquakes along the northwestern part of the NAFZ.

How to cite: Özkan, B., Eken, T., Gaebler, P., and Taymaz, T.: Implications for Non-Self Similar Energy and Moment Scaling of Small-to-Moderate Earthquakes Along the NAFZ: Source Displacement Spectra Derived from Coda Waves, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-749, https://doi.org/10.5194/egusphere-egu24-749, 2024.

The purpose of the study is to describe a geodynamic process in the study area using its focal mechanism and stress field inversion to characterize precise events along the study area, the rupture zones of the South Hangay Fault System (SHFS). This fault system was activated by four earthquakes which are occurred along the Bayanbulag fault (2012/10/03, Mw=4.7) and Bayankhongor fault (2013/01/05, Mw=4.2, & Mw=4.2; 2013/11/25, Mw=3.9). These earthquakes are the strongest in the fault zone.

From the Mongolian National Data Center's database, it has chosen 2228 occurrences (0.1ML5.4) from the Handay Experiments, which used 72 broadband seismometers to cover Hangay Dome. Using HypoDD with a double-difference technique, its seismic station density provides us with precise hypocenter location along the fault system. Among these events, 47 focal mechanism solutions were determined using the first-motion polarity of the P wave from the experimental seismic networks of Mongolia. Then, we classified the determined focal mechanism parameters. According to classification, three main cluster zones are related to the Bayanbulag (BB), Bayankhongor North (BHN), and Bayankhonor South (BHS) fault zones along the rupture area of the South Hangay Fault System. 

Furthermore, we determined the stress fields, stress regime, and the horizontal maximum (SHmax), and minimum (Shmin) stress orientations for all three zones.  

We concluded that the whole SHFS is a left-lateral strike-slip fault with normal and reverse components, NE-SW shortening, and corresponding NW-SE extension. Its compression orientation in the NE-SW direction is the same as the azimuth direction of the India-Asia collision.

We hope that this stress inversion results can be a useful tool for geodynamic and seismotectonic analysis of this part of Mongolia and it will give a better understanding of different stress regimes.

How to cite: Dashdondog, M., Chimed, O., Meltzer, A., and Dorjsuren, A.: The seismic source parameters of the South Hangay Fault System in Central Mongolia, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1310, https://doi.org/10.5194/egusphere-egu24-1310, 2024.

The seismic moment M₀, the associated moment magnitude M_w, and the corner frequency f_c are essential parameters for earthquake studies and seismic risk management. In the context of stable continental regions (SCRs), remote from active plate boundaries, the assessment of these parameters is made difficult by the low energy release associated with each earthquake and the low density of seismological networks.

In the north west of France, the Armorican Massif and its surroundings are part of a SCR, where the densification of the seismological network, completed in 2019, now allows for a reassessment of the regional seismicity. Though characterized by very small strain rates, the region currently displays a high rate of low-to-moderate earthquakes (up to a few M_w lower or equal to 4.0 – 5.0). For such small earthquakes, these assessments are particularly sensitive to the signal-to-noise ratio, to the seismic structure of the region, to its attenuation properties, and to the azimuthal distribution of the regional network with respect to the focal mechanisms.

We attempt to determine the M₀, and the f_c, of 106 earthquakes, detected in northwestern France between 2015 and 2023, with local magnitudes M_L ranging from 2.0 to 5.3, using spectral methods. We obtained a linear relationship between M_L and M_w for M_w ranging from 1.5 to 5.0. Our analysis also highlighted the importance of the frequency dependence of attenuation on the assessment of the f_c. This study will show the relation between M₀ and f_c in the region.

How to cite: Alloncle, M., Mocquet, A., and Bonnin, M.: Earthquake source characterization in stable continental regions: Application to the Armorican Massif, France, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1533, https://doi.org/10.5194/egusphere-egu24-1533, 2024.

Seismic urgent computing aims at assessing the potential impact of earthquakes through rapid simulation-based ground-shaking forecasts. However, uncertainty quantification remains a significant challenge in this domain.

While current practice accounts for the uncertainty arising from Ground Motion Models (GMMs), it neglects the uncertainty about the source model, which is only known approximately in the first minutes after an earthquake. Addressing this issue involves propagating earthquake source uncertainty from a multi-scenarios ensemble that captures source variability to ground motion predictions. In principle, this could be accomplished with 3D modelling of seismic wave propagation for multiple earthquake sources. However, full ensemble simulation is unfeasible under emergency conditions with strict time constraints.

Here we present ProbShakemap, a Python toolbox which generates multi-scenario ensembles and delivers ensemble-based forecasts for urgent source uncertainty quantification. It implements GMMs to efficiently propagate source uncertainty from the ensemble of scenarios to ground motion predictions at a set of points, while also accounting for model uncertainty (by accommodating multiple GMMs, if available) along with their intrinsic uncertainty. Notably, ProbShakemap does not rely on any recorded data, and only requires the following event-specific information: latitude, longitude, magnitude and time. ProbShakemap incorporates functionalities from two open-source toolboxes routinely implemented in seismic hazard and risk analyses: the USGS ShakeMap software and the OpenQuake-engine.

We quantitatively test ProbShakemap against past earthquakes, illustrating its capability to rapidly quantify earthquake source uncertainty.

How to cite: Stallone, A., Selva, J., Cordrie, L., Faenza, L., and Michelini, A.: ProbShakemap: a Python toolbox for urgent earthquake source uncertainty quantification, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1985, https://doi.org/10.5194/egusphere-egu24-1985, 2024.

Georgia is located in the Caucasus between The black and the Caspian seas and is surrounded by the Greater and Lesser Caucasus. Among the seismic areas of Georgia, the volcanic upland of Javakheti situated in the south of Georgia is notable for its high level of seismicity where three large earthquakes with M6 occurred in the last century. The main goal of the study is to investigate the attenuation properties of the lithosphere in the region using a hundred and twenty local earthquakes in 2008-2020  recorded at five seismic stations equipped with broadband Guralp CMG40T and Trillium 40  seismometers. Earthquake magnitudes varied from 1.5 to 4.1; epicentral distances and depth were smaller than 60 km and 19 km, respectively. The quality factors of coda waves Q_cand direct P, S waves Q_p,and Qs have been estimated using the single back-scattering model and the extended coda normalization methods, respectively. Wennerberg’s approach has been used to estimate intrinsic Q_i and scattering Q_s attenuation parameters. The Q values were fitted to a  power-law, of form Q(f)= Q₀ (f)ⁿ, where Q₀ is the quality factor at 1Hz and n is the frequency relation parameter, which depends on the heterogeneity of the medium. The obtained values of Q_c, Q_p, Q_s, Q_i, andQ_sc show the frequency-dependent character in the frequency range of 1.5-24 Hz and are expressed as:

Q_c = (47.6±3.8)^{(1.034±0.048)},  Q_p = (17.4±2.3)𝑓^{(1.100±0.033)}, Q_s = (28.8±3.3)𝑓^{(1.048±0.039)}, 

Q_i = (62 ± 4) f ^{(0.969±0.052)},  Q_sc = (177 ± 6) f ^{(0.932±0.051)}

The calculated attenuation parameters characterize the entire earth's crust under the Javakheti plateau and the surrounding area. The observed Q_c and Q_i values are almost identical at different central frequencies and both of them are less than Q_sc. This means that the effect of intrinsic attenuation is dominated by scattering attenuation. Comparison of our results for similar lapse times to those obtained in other tectonic and seismic active regions show that the Q values and their frequency-dependent relationships are in an interval of values of tectonically active and highly heterogeneous regions. The results obtained will be useful for source parameter estimation, ground motion prediction, and hazard assessment of the study regions.

How to cite: Shengelia, I., Jorjiashvili, N., Godoladze, T., and Buzaladze, A.: Qc, Qp, Qs, Qi, and Qsc attenuation parameters in the southern part of Georgia, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2254, https://doi.org/10.5194/egusphere-egu24-2254, 2024.

Based on the catalog of 225 earthquakes with magnitudes 5 or above, the catalog of earthquake sequences, and the focal mechanism of the historical earthquakes in Sichuan-Yunnan region from 1966 to 2021, and referring to the previous research and practice on the estimation of the tendency of the aftershock activity, 10 sample datasets for the judging features of the earthquake sequence types have been constructed. According to the earthquake sequences types—swarm type, mainshock-aftershock type, as well isolated type—three labels have been made. After processing the imbalanced state and the missing state of the feature parameters, a decision tree model was used to study and analyze the importance of feature parameters. The results showed that there were differences in the importance categories of the feature parameters in different periods. As the sequence data increased, sequence type judgment relied more on dynamic sequence data; the parameters related to the main shocks’ focal mechanism and the main shocks’ parameters have a high contribution rate to the sequence classification, while the contribution rate of sequence parameters is extremely low. In overall, the results provided by the model are consistent with the actual empirical prediction methods. The above results can provide some ideas for the preliminary screening, exclusion, and selection of the complex and numerous feature parameters.

How to cite: Zhao, X., Su, Y., and Meng, L.: Research on the Importance of Feature Parameters in Seismic Sequence Type Determination Based on Decision Tree, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2853, https://doi.org/10.5194/egusphere-egu24-2853, 2024.

Through the utilization of P-Alert network data from Taiwan, this study endeavors to estimate earthquake magnitude (Mcaa) using the cumulative absolute absement (CAA) methodology across varying window lengths after the arrival of P-wave. It is differentiated that even the proximity of the nearest 12 stations to the epicenter results in robust magnitude estimations. Notably, the standard deviation between the estimated Mcaa and the moment magnitude (Mw) using 12 stations decreases with the increase in window length and is found minimum for 5s window length. For 3s window the variation between Mcaa and Mw is found ±0.385, whereas, for 5s window it is ±0.313. Consequently, the estimation of Mcaa remains reliable. The magnitude Mpd is alternatively deduced from Pd, utilizing the closest 12 stations situated near the epicenter. The standard deviation of the order of ±0.412 is observed between the estimated Mpd and Mw for 3s window, whereas for 5s window it is ±0.281. A difference is observed using Mpdand Mcaafor comparison with Mw. The standard deviation error decreases for Mcaaand Mpd with increase in window length. While Mpd performs better under a 5s window scenario, it tends to underestimate the magnitude of an earthquake with a magnitude of Mw 7.0. On the other hand, CAA surpasses Pd in magnitude estimation, though with a slightly higher standard deviation compared to Mcaa. As a result, Mcaa is considered a more reliable magnitude indicator.

How to cite: Wu, Y.-M.: Cumulative absolute absement for magnitude determination in earthquake early warning system using low-cost sensors, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2969, https://doi.org/10.5194/egusphere-egu24-2969, 2024.

The Holocene Okavango Rift Zone (ORZ) marks the southern terminus of the Western Branch of the East African Rift System. Detailed knowledge of the crustal and lithospheric mantle structure of the ORZ is essential to decipher the rifting mechanism and nature of the lithosphere of this incipient continental rift. A 3-D shear wave velocity model from the surface to 160-km depth is constructed by jointly inverting the Rayleigh wave phase velocity dispersion and receiver function data through a non-linear Bayesian Monte-Carlo algorism. The crustal thickness estimates from our new velocity model generally agree with previous receiver function investigations of the region. The crust beneath the ORZ is thinned compared with the cratonic regions to both sides of the rift, suggesting a certain degree of continental stretching. Our velocity model also reveals two low velocity anomalies in the crust and upper mantle beneath the incipient rift, respectively. The shallow low velocity anomaly is mainly confined in the upper and middle crust, and the deeper low velocity anomaly extends from the Moho to at least 160 km depth, with a high-velocity lower crust (~4.0 km/s) in between. Although the two low velocity anomalies are probably both caused by rift-related decompression melting, the structural feature imaged suggests that they are generated separately and individually. Based on our observations, we propose that thermal upwelling and decompression partial melting in the upper mantle of the ORZ have a limited contribution to the stretching and thinning of the crust during the initiation of the continental rifting. The crustal rifting could be induced by an intra-plate relative motion between the South African block and the rest of the African continent along a pre-existing weak zone.

How to cite: Wang, T., Chen, L., Gao, S. S., Liu, K. H., Yu, Y., Yu, Z., and Wang, X.: Imaging of Crust and Lithospheric Mantle of the Incipient Okavango Rift Zone: Implications on the Rifting Mechanism, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3321, https://doi.org/10.5194/egusphere-egu24-3321, 2024.

 In the present study, seismic gaps are identified as periods with no occurrence of M ≥ 4.0 earthquake over dT ≥ 400 days. After examining all records in the Sichuan-Yunnan-Tibet-Qinghai junction area on the southeastern margin of the Tibetan Plateau in 1970−2022, a total of six M ≥ 4.0 seismic gaps were identified. Spatio-temporal images of the seismic gaps had similar characteristics and demonstrated spatial overlapping and statistical significance. The quiet periods of the six seismic gaps included 419−777 days (approximately 580 days on average). The semi-major-axis and semi-minor-axis lengths were in the 880−1050 km (approximately 987 km on average) and 500−570 km (about 533 km on average) ranges, respectively. Case analysis results revealed that the images of M ≥ 4.0 seismic gaps were of high significance in predicting M ≥ 6.7 strong earthquakes in the region, and they could be used as a predictive index on a time scale of about 1-0.5 year or less.

How to cite: Su, Y. and Zhao, X.: Characteristics and predictive significance of spatio-temporal space images of M ≥ 4.0 seismic gaps on the southeastern margin of the Tibetan Plateau, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3496, https://doi.org/10.5194/egusphere-egu24-3496, 2024.

The collision of the Indian and Eurasian plates has resulted in high-altitude Tibetan Plateau with active seismicity. In this study, we apply the seismic box tomography to the southern Tibetan Plateau, aiming to obtain a self-consistent and high-resolution (10−20 km) model of the crust and upper mantle beneath the region, including density as well as bulk and shear moduli without a priori constraints, which provides us with crucial constraints on the compositional and thermal structures of a highly deformed lithosphere in southern Tibetan Plateau.

In order to obtain the seismic tomographic model, we perform full-waveform inversion of teleseismic (30°−90°) surface- and body-wave waveforms recorded by the Hi-CLIMB network, a densely distributed (5−10 km station spacing) N-S oriented linear seismic array deployed during 2002 and 2005. In our iterative hierarchical inversion workflow, we calculate the sensitivity kernels based on the adjoint method and the model is updated by the L-BFGS algorithm. Data covariance matrices are introduced to control the data quality and objective weighting functions for different seismic events. We will present our preliminary results of the on-going study with comparison to existing models.

How to cite: Zhu, Q., Fuji, N., and Zhao, L.: Full-waveform tomography for the lithospheric structure of southern Tibetan Plateau, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4004, https://doi.org/10.5194/egusphere-egu24-4004, 2024.

We used the liquid-based R2 rotational seismometer in addition to several arrays of translation velocity seismometers on a 12-floor building in the National Cheng Kung University campus to evaluate the dynamic responses of the structure. During the observation period in August-October 2023, we encountered a moderate M 5.6 earthquake sequence 61 km north of the campus and one moderate typhoon passing through this 49-m-long and 12-m-wide building. By examining these data, we investigate the natural frequency and the rotation behavior of the long-strip-shaped building. Both the time-frequency and Fast Fourier Transform analyses of the microtremor and earthquake data show two dominant frequencies of ~1.2 Hz and 1.8 Hz occurred in the horizontal directions. The translation velocity and rotation rate are more significant in the transverse, short-axis direction and at the location away from the elevator of the building. The translation velocity array and rotational seismometer show rotations around the horizontal and vertical axes during the M 5.6 earthquake. The results of two natural frequencies and the corresponding rotational motions are most likely related to the asymmetric design of the building, which resulted in the non-rigid behavior of the structure. These findings may provide insights into improvements that could enhance the building’s resilience to seismic or typhoon events.

How to cite: Rau, R.-J., Wu, C.-F., Chen, Y.-C., Wu, H.-Y., and Lin, C.-J.: Dynamic responses of a building derived from microtremor and seismic signals, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4895, https://doi.org/10.5194/egusphere-egu24-4895, 2024.

In the continuation of the work carried out over the period 1962-2009 as part of the SI-Hex project, work is on going to revise the seismic catalog of mainland France from 2010 to 2018. This time period is characterized by both an upgrade of short-period stations with broadband stations and a major deployment of new broadband stations as part of the Résif-Epos research infrastructure (now called Epos-France), significantly increasing the amount of detected and processed events.

This catalog will benefit from our advances in the use of new artificial intelligence tools, such as PhaseNet, a deep learning automatic picking method, as well as in the development of a deep learning method for discrimination between earthquakes, quarry blasts and explosions.

This catalog will be built from those of the national observation service BCSF-Renass, CEA/LDG and regional seismological observatories (Isterre, OCA, OMP). The earthquake picks from these catalogs will be supplemented by those automatically obtained by deep learning on all the waveforms from the Epos-France (formerly Résif-Epos) stations daily used by BCSF-Renass (as part of its mission to monitor seismicity in mainland France) including stations from neighboring countries (GB, LU, BE, DE, CH, IT, ES), as well as those from temporary network stations (AlpArray, CifAlps2). 

The process workflow includes several steps. The first one consists in a clustering of picks close in time to reduce the amount of picks to process; duplicated picks are removed and priority is given to the manual ones. The second step is the association of seismic phases to create events, by combining the HDBSCAN algorithm - to merge picks close in time and space - with the PyOcto one - to discard picks that did not follow typical travel-time curves. The third step consists in event location using NonLinLoc algorithm with several regional models chosen based on the prior location obtained from PyOcto. At the last step, a moment magnitude Mw is computed (when possible) from waveform spectral fitting using a modified version of SourceSpec. To compute robust magnitudes in particular for low magnitude events, we include magnitude station corrections computed from statistics on magnitude differences between event and stations. Finally, events information (ie. origins, magnitudes) coming from the various catalogs are integrated into the multi-origin catalog according to the QuakeML standard, with the preferred location being the new one computed on the third step.

This catalog currently under revision will represent an update of seismicity in France over the period 2010-2018. Preliminary results show that it will incorporate a significantly increased number of low-magnitude events, detected thanks to the inclusion of picks from artificial intelligence tools. Event labeling is consolidated using our deep learning discrimination algorithm, and a Mw magnitude is calculated for each event using waveforms.

How to cite: Grunberg, M. and Lambotte, S.: A new workflow for revising the seismicity catalog for mainland France, covering the period 2010-2018, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5100, https://doi.org/10.5194/egusphere-egu24-5100, 2024.

We report on the temporal variations of the near-surface (< 500 m) seismic structure (Vp, Vs, and Vs anisotropy) of Taiwan using the empirical Green’s functions of body waves between vertical station pairs at 60 borehole sites. In our previous work, the obtained near-surface anisotropy are categorized into stress-aligned anisotropy (SAA) and orogeny parallel anisotropy (OPA). Since all the major geological units of Taiwan are well sampled by borehole arrays, and drilling data for 52 sites are available, we were able to find that OPA is typically stronger than SAA, SAA strength is generally higher in sedimentary rocks, igneous rocks, and gravel sediments compared to fine-grained sediments, and OPA is more pronounced in foliated metamorphic rocks than in dipping sedimentary strata. In this study, we aim to address the following specific questions with the obtained results: (1) How do the temporal variations of near-surface seismic properties in different geological units of Taiwan correlate with seismic activity or nearby earthquake events? (2) Are there distinct patterns in the temporal variations of anisotropy strength based on the specific geological composition? (3) Do sites characterized by OPA exhibit different temporal variations in response to seismic activity compared to sites dominated by SAA?

How to cite: Chen, H.-C., Gung, Y., Lee, H.-Y., and Chen, L.-W.: On the temporal variations of near-surface seismic structure of Taiwan and its geological inferences, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5890, https://doi.org/10.5194/egusphere-egu24-5890, 2024.

The repeating earthquake (RE) ruptures a single fault patch repeatedly and generates highly similar waveforms. The RE is often observed in the area of subduction zones (Uchida and Matsuzawa, 2013; Yu et al., 2013; Uchida, 2019). However, even in the intraplate region, the RE has been found in the ruptured fault zones (Li et al., 2007; Li et al., 2011; Bisrat et al., 2012). We searched for REs around the epicenter of the 2008 M_L 3.6 Gyerong earthquake that occurred in Mount Gyeryong, the Korean Peninsula, located in a stable intraplate region. In the study area, 48 earthquakes (M_L 0.4–3.6) were reported during 2002–2022, while we found 50 earthquakes during 2018–2022 using a template matching. We located the events based on the Hypoellipse (Lahr, 1999), and also refined the hypocenters using the double difference method (hypoDD; Waldhauser and Ellsworth, 2000) to obtain the high-resolution fault geometry. It is found that the epicenters exhibit a linear alignment of the fault striking along WNW-ESE consistent with one of the strikes of the M_L 3.6 event which has a strike-slip focal mechanism with a strike of 108° or 198°, a dip of 83° or 88°, and a rake of -2° or -173°, which indicates that the M_L 3.6 earthquake occurred with a left-lateral fault slip. We estimated the rupture directivity of the M_L 3.6 event from the apparent source time functions obtained by the empirical Green’s function approach. A vast number of microearthquakes including aftershocks of the M_L 3.6 event occurred in the rupture direction (i.e. the east-southeast of the epicenter of the M_L 3.6 event). We identified REs based on the waveform similarity (cross-correlation coefficient > 0.95) and their locations (co-location) to distinguish them from neighboring earthquakes. We found that the REs occurred within the rupture radius of the M_L 3.6 event. Upon categorizing these REs according to their family duration, we identified three swarm-type families that occurred in 2007, 2009, and 2010, along with a continuous-type family spanning from 2011 to 2019. These observations demonstrate the close relationship between the REs and the M_L 3.6, specifically highlighting the fault’s rupture and healing process.

How to cite: Jung, S., Sheen, D.-H., Cho, C.-S., and Kim, K.: Observation of intraplate repeating earthquakes within the fault zone of the 2008 ML 3.6 earthquake, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7189, https://doi.org/10.5194/egusphere-egu24-7189, 2024.

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.

How to cite: Steinberg, A., Gestermann, N., Ceranna, L., Hartmann, G., Lund, B., Dunham, E., Hupe, P., Voss, P., Larsen, T., Dahl-Jensen, T., Köhler, A., Schweitzer, J., Pilger, C., Plenefisch, T., Stammler, K., Donner, S., Gaebler, P., and Wiedle, C.: Source analysis of the 2022 Nord Stream and 2023 Balticconnector underwater explosions, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7660, https://doi.org/10.5194/egusphere-egu24-7660, 2024.

Six-degree-of-freedom (6-DoF) measurements, which combine rotational sensors and seismometers, provide a comprehensive dataset that allows seismologists to determine the back azimuth of a potentially moving source from a single-point measurement. Our investigation focused on tracking the movement of a vibroseis truck operating from 20 November 2019, 11:00 UTC, to 21 November 2019, 14:00 UTC. Using 480 sweep signals, each lasting 15 seconds and covering a wide frequency range from 7 to 120 Hz, we measured at 160 different locations. Back azimuths for each sweep were derived from the 6-DoF data, and root mean squares were calculated for each component. This procedure was repeated for five additional rotational sensors of the same type.
During the first day, the north component of all sensors recorded larger amplitude signals than the East and Vertical, indicating the dominance of SV (shear-vertical) wave energy. Subsequently, we observed gradually increasing amplitudes on the east component, which was consistent with the direction of the moving vibroseis truck. Although the dominant wave type recorded was SV, and the method of comparing horizontal rotation rates was used to calculate the back azimuth, we observed a relatively decreasing accuracy of direction estimates as the truck moved away from the sensors due to increased scattering. To fully understand the reason for this, we investigated the specific fingerprint of each wave type in the wave field. Our results suggest that direction estimates should be made using only the portion of the wavefield containing SV-type waves when using this method, and then the moving source should be tracked accordingly. This approach provides insight into the trajectory of the truck and improves our understanding of the seismic signals generated during the experiment.

How to cite: Izgi, G., Eibl, E. P. S., Krüger, F., and Bernauer, F.: Tracking a Vibroseis Truck and Investigating the Wavefield using 6 Rotational Sensors in Fürstenfeldbruck, Germany , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8482, https://doi.org/10.5194/egusphere-egu24-8482, 2024.

Apart from classical earthquake monitoring, seismological data can also be used to detect explosions in near-real-time on both regional and global scales. We demonstrate how seismic and infrasound data can provide more comprehensive and objective information about conflict-related explosions or suspicious events that might be the result of targeted attacks. We can identify the underwater explosions at the Nord Stream pipeline infrastructure in the Baltic Sea in September 2022. Cross-correlation analysis allowed us to identify sub-events several seconds apart which can be associate with specific locations along the pipelines. Furthermore, we detect a signal at the Finish seismic array in October 2023 which may be associated with the damage along the Balticconnector. The other example is from Ukraine, where we present the ability to automatically identify and locate ground explosions related to the Russia-Ukraine conflict with data from the Malin array (AKASG). Between February and November 2022, we observe more than 1,200 explosions from the Kyiv, Zhytomyr, and Chernihiv provinces. Both seismic and infrasound detections can be used to verify and improve accurate reporting of military attacks and help to provide an unprecedented view of an active conflict zone. We analyze events with a variety of seismo-acoustic signatures and significant differences in explosive yield. These can be associated with various types of military attacks, including artillery shelling, cruise missile attacks, airstrikes, or the destruction of the Kakhovka dam NE of Cherson.

How to cite: Goertz-Allmann, B., Dando, B. D. E., Koehler, A., Brissaud, Q., Schweitzer, J., and Kværna, T.: Near-real time detection of conflict-related explosions or suspicious events using seismological data , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8735, https://doi.org/10.5194/egusphere-egu24-8735, 2024.

On June 16, 2023 at 16h38 UTC, a moderate earthquake of magnitude M_W=4.9 stroke western France south of Niort city, near the small village of La Laigne (Charente Maritime). The shaking has been widely felt in the whole NW France and macroseismic intensity (EMS98) of VII was reached at the epicenter. Such an event is relatively rare in continental France and represents the second largest event in the western France in the last century. The epicentral region is located at the northern termination of the Aquitaine basin where 300 m of Mesozoic sediments covers the variscan basement. The focal mechanism obtained from waveform inversion corresponds to a pure dextral strike-slip motion or a pure senestrial strike-slip motion along a EW or NS striking fault plane, respectively.

The fault that ruptured on June 16 is not known. To gain insight on its characteristics, teams of Nantes (Osuna and LPG), of Strasbourg (EOST and ITES) and of the CEA deployed between June 17 and June 22, 2023 for approximately one month, a network of 3-components stations composed of 12 MEMS accelerometers, 104 five hertz geophones and 5 broadband velocimeters in a 40 by 30 km region around the epicenter, with a station inter-distance of approximately 4 km.

We present is this study the first results derived from this unique experiment. In particular, we show that the aftershock sequence (more than 600 events recorded) highlights a planar rupture zone of about 5.4 km², trending NS and strongly dipping to the East (75°), located between 2 and 5 km depth. Site effect analysis allows us to better understand large ground motion distributions over the area and their link with macroseismic intensities. The installed array also allows us to infer a preliminary 3D V_S model of the region. We show the extent to which a dense temporary network is mandatory for studying the fine structure of the fault plane in a region where previous knowledge of active geological structures is limited.

How to cite: Bonnin, M., Alloncle, M., Bes de Berc, M., Beucler, É., Fligiel, D., Grunberg, M., Hourcade, C., Perrin, C., Sèbe, O., Vergne, J., and Zigone, D.: Aftershock sequence and source characteristics of the June 16, 2023 MW=4.9 La Laigne earthquake, western France, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9479, https://doi.org/10.5194/egusphere-egu24-9479, 2024.

Seismic waves excited by human activity frequently mask signals due to tectonic processes, and are therefore discarded as nuisance.  Seismic noise-field analysis is, however, a powerful tool for characterizing anthropogenic activities. Here, I apply this analysis to examine seismic precursors to the October 7 Hamas attack on Israel. The precursory activity in Gaza included massive mobilization which took place in the hours leading to the attack, and was  documented on multiple media outlets. Favourable conditions, which arise due to a temporary lack of anthropogenic activity in Israel, allow remote seismic stations to record signals due to Gaza vehicle traffic. I use these seismograms in order to identify anomalous ground-motions, associate them with pre-attack mobilization, and precisely determine their location. By applying array analysis to three seismic stations located tens-of-kilometers from the Gaza strip, I was able to obtain valuable information on the Hamas attack plans. This suggests that embedding seismic noise-field analysis into decision-making protocols could enhance preparedness, thus providing an opportunity to blunt terrorist attacks and reduce the number of casualties.

How to cite: Inbal, A.: Seismic Precursor for the October 7th Terrorist Attack?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9598, https://doi.org/10.5194/egusphere-egu24-9598, 2024.

In the early morning of 23 February 1887, the ‘Ligurian earthquake’, a devastating seismic event currently estimated at MW 6.3-7.2, shook the towns of the Italian and French Riviera. It is the most devastating earthquake known in this region: it is thought to have claimed at least six hundred lives, displaced twenty thousand people, and destroyed many historic buildings and houses. As a result of the event, a tsunami with a maximum run-up of two meters near Imperia (Italy) also occurred and the record of the tide gauge in the port of Genoa (Italy) has long been considered the only existing record of the event.

However, we found that the 1887 earthquake was also recorded by historical magnetometers in the UK and France. These instruments were used to measure variations in geomagnetic field strength, but were also able to record seismic waves, which were essentially a simple ‘disturbance’. Almost uninterrupted records of this type of variometric data are held by the British Geological Survey (BGS). Traces recorded at Greenwich, Kew, and Falmouth magnetic observatories, which clearly show waveforms related to the event, were used. The Bureau Central de Magnetisme Terrestre (BCMT) also keeps magnetograms: in particular, we used the recordings of the instrument at Le Parc de Saint-Maur (Saint-Maur-des-Fossés, Paris).

The waveforms were digitized and processed according to the theory of Eleman (1966), which describes the response of a classical declinometer and/or a horizontal force instrument to harmonic ground displacement, and according to the work of Krüger et al. (2018).

The location of the epicenter and the magnitude of this historical earthquake are difficult to characterize with high accuracy, and the focal mechanism of the fault responsible for the event remains controversial to this day. We present the preliminary results of our research, which is focused on the revaluation of the Ligurian earthquake in terms of magnitude and focal mechanism. This would lead for the first time to a definition of magnitude on an instrumental basis for this important seismic event, whose macroseismic intensity is usually assessed based on studies conducted immediately after the event to determine the damage it had caused.

How to cite: Tarchini, G., Spallarossa, D., Parolai, S., Sandron, D., and Saraò, A.: Reassessment of the historical earthquake of 23 February 1887 in Liguria (north-western Mediterranean) on the basis of magnetogram recordings, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11960, https://doi.org/10.5194/egusphere-egu24-11960, 2024.

Earthquake early warning systems depend on the prompt, reliable arrival of seismic data at network data centers. Network operators invest significant resources into the design, installation and operation of real-time acquisition systems to ensure maximum data completeness and minimum data latency, to allow EEW processing modules to detect events and issue warnings as quickly as possible.

These mission-critical acquisition systems must perform before, during and after earthquakes, as main shocks are frequently preceded by foreshocks and followed by aftershocks, which are often just as dangerous. As such, a key consideration in the design of these systems is the impact that large earthquakes may have. Seismic data is generally encoded using Steim compression, which is a first difference algorithm. During large events the differences between samples grow, requiring more bits to record and, thus, increasing the data volume. This results in a surge in the throughput required during large events. System designers and network operators must be fully aware of this effect and plan for it accordingly.

This study expands on existing work to further characterize the impact of large events on seismic data compression and the corresponding spikes in throughput which must be supported by real-time acquisition systems. The study will examine the relationship between compression and various factors, including station magnitude, hypocentral distance, sample encoding technique, packet size, sample rate and system sensitivity.

How to cite: Laporte, M., Perlin, M., Jusko, M., and Easton, D.: Seismic Data Compression and Telemetry Bandwidth Considerations for Earthquake Early Warning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12248, https://doi.org/10.5194/egusphere-egu24-12248, 2024.

High Agri Valley is an intermontane basin of the axial portion of Southern Apennines (Southern Italy), characterized by a very strong seismogenic potential. Indeed, a  Mw=7.0  earthquake occurred in 1857. Currently, the seismic networks managed by ENI Oil Company and INGV, installed in the area, continuously record a low-magnitude natural seismicity. Furthermore, two anthropogenic earthquake clusters are documented in two distinct sectors of the valley, located NE and SW of the artificial Pertusillo lake, respectively. The first cluster is related to the fluid-induced microseismic swarms caused by the injection, through the Costa Molina 2 well, of the wastewater produced by the exploitation of the Val d’Agri oilfield. The second cluster is due to a protracted reservoir induced seismicity (RIS) affected by the combined effects of the water table oscillations of the Pertusillo lake, the regional tectonics and likely the poroelastic/elastic stress due to aquifers in the carbonate rocks.

In this study we investigated the attenuation properties of the High Agri Valley crust by the estimation of the S-coda waves Qc^-1, as it is widely recognized the role of fluids on this parameter. We selected a dataset of about 1800 events acquired from 2001 to 2015 by the two above mentioned seismic networks, with local magnitude (ML) ranging from 0 to 3.3. We estimated the attenuation of the target area by means of a linear regression analysis of the amplitude decay curves of the envelopes of the seismograms; these were filtered in the frequency ranges centered on 1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16 Hz. The Qc estimates were performed by using different time windows for the envelope fitting, starting from the time T₁ to the time T_L (the lapse time). In detail, we adopted, as T_1, 1.0*Ts, 1.5*Ts and 2.0*Ts (being Ts the S wave arrival time), and as T_L 10s, 15s, 20s, 25s and 30 s from the event origin time. Only the components for which the condition T₁<T_L<T₂ was fulfilled were considered for the linear regression, being T₂ the end-time of the coda envelope; the latter was automatedly found by a proper methodology implemented in this study.

The obtained results show the increase of Qc as a function of f at all the considered T_L. Compared with other tectonic regions worldwide, in the High Agri valley the Q_c(f) is very low: the Q₀, that is the Q_c at 1 Hz, ranges between 8 and 57. At greater frequencies, the highest estimated Qc value is lower than 400. These evidences could be interpreted as the effect of fluids in the investigated crust, thus providing a further hint on their possible role in the seismicity of the area. A complete characterization of seismic attenuation of the High Agri Valley will require further investigations, that is the separation of scattering and intrinsic contributions in the total attenuation and a 3D imaging: indeed, the latter could highlight possible overlapping between spatial attenuation anomalies and seismicity distribution in the investigated area.

How to cite: Serlenga, V., Lucente, S., de Lorenzo, S., Del Pezzo, E., Filippucci, M., Ninivaggi, T., Stabile, T. A., and Tallarico, A.: Preliminary estimation of attenuation properties in the High Agri Valley (Southern Apennines, Italy) by the coda attenuation method, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12588, https://doi.org/10.5194/egusphere-egu24-12588, 2024.

Geophysical monitoring requires the highest level of performance and reliability from purpose-built and tightly integrated instrumentation and infrastructure. Parallel and separate efforts between different scientific disciplines seen in the past came at the expense of duplicated infrastructure, telemetry and power subsystems, and even land use permits. This duplication increases costs, ultimately limiting station counts and reducing “the reach” of monitoring networks. Recent ambitions to combine multi-disciplinary geophysical applications into streamlined deployments led to initiatives such as the European Plate Observing System (EPOS) and the recent amalgamation of the SAGE and GAGE programs in the United States.

Modern seismic dataloggers, such as the Nanometrics Centaur, support a wide range of seismo-acoustic sensor interfaces and sensor types while maintaining ultra-low power consumption, precise timing, and reliable data transport with automatic back-fill features over flexible telemetry mediums. These properties transformed the Centaur’s capabilities to act as a highly versatile foundation in multi-disciplinary geophysical station deployments. 

Despite initially being designed as a high-performance data recorder for seismic applications, Centaur’s applicability has evolved to include data collection for the infrasonic, geodetic, magnetic, and meteorological domains. This triggered the development and addition of purpose-built features to support multi-disciplinary use cases with the same proven performance and reliability of a Centaur seismic station.

Both existing and planned capabilities that enable reliable and efficient multi-disciplinary science are discussed.

How to cite: Perlin, M., Trerice, N., Somerville, T., and Jusko, M.: Seismic Network Station Infrastructure as the Basis for Multi-Disciplinary Geophysical Stations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12990, https://doi.org/10.5194/egusphere-egu24-12990, 2024.

Size distribution of earthquakes, characterized by the power law decay (b-value), sometime displays the major earthquakes before their occurrence. The b-value reflect state of stress and proximity of fault failure condition according to previous studies. However, the causes are difficult to separate each other. This study proposes an additional indicator reflecting the proximity. Seismic moment release in a volume by small earthquake indicates inelastic strain. The efficiency of inelastic strain on stress loaded medium exhibits proximity to strength of the medium based on Mohr diagram and Coulomb failure condition. Thus, we adopt the efficiency as the indicator. We examine b-value and the efficiency variation in pre- and post- seismic activity of the 2016 Kumamoto earthquake sequence. Weighted b-value distribution by the efficiency captured the initiation point of the Kumamoto earthquake. The result suggests utilizing both b-value and the efficiency contribute to improving earthquake alerts and disaster mitigation.

How to cite: Matsumoto, S.: Inelastic strain efficiency of small earthquakes as an indicator for proximity of the 2016 Kumamoto earthquake (M7.3), EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13440, https://doi.org/10.5194/egusphere-egu24-13440, 2024.

Seismicity records during the 1990’s reveals that large inland earthquakes tend to concentrate near several active volcanoes in the central part of northeastern (NE) Japan (Hasegawa & Yamamoto, 1994 Tectonophysics). The inland seismicity around active volcanoes could be related to localized zones of high strain contraction detected by the GNSS measurements in 1997–2001 (Miura et al., 2004 JGR). Several studies speculated that the localized strain contraction is caused by inelastic deformation of weak lithosphere beneath the active volcanoes (Hasegawa et al., 2004 J. Seismol. Soc.). Such weak lithosphere (i.e., low-viscosity zone or LVZ) is inferred from high heat-flow observations (Tanaka et al., 2004 EPS), lithospheric strength simulation (Shibazaki et al., 2016 GRL; Muto, 2011 Tectonophysics) and seismic-velocity tomography (Hasegawa et al., 2005 JGR). However, because of complex interplay between elastic and inelastic processes during steady-state (i.e., interseismic) crustal deformation, the physical mechanism related to inelastic deformation is still poorly understood.

When the M_w9.0 2011 Tohoku-oki earthquake occurred, strong surface deformations were observed locally near the active volcanoes (Takada & Fukushima, 2013 Nat. Geosci.) and continued for several years after the mainshock (Muto et al., 2016 GRL). Past studies (e.g., Sun et al., 2014 Nature) advocated that the earthquake-related inelastic processes such as viscoelastic mantle relaxation dominates the crustal deformations in the postseismic period. In the present study, we identified localized strain contractions near the active volcanoes by extracting the short-wavelength strain rate (Meneses-Gutierrez & Sagiya, 2016 EPSL) from the GNSS observations during 2012–2014. We explained these localized strain contraction by building three-dimensional rheological models of small-scale LVZs beneath five active volcanoes of NE Japan. We simulated the volumetric deformation of viscoelastic LVZs using power-law Burgers rheology, which previously succeeded to explain the large-scale postseismic deformation of the 2011 Tohoku-oki earthquake (Agata et al., 2019 Nat. Commun.; Muto et al., 2019 Sci. Adv.; Dhar et al., 2022 GJI). The power-law Burgers rheology represents the bi-phasic nature of rock deformations (rapid transient with subsequent steady state) and power-law dependency of strain rate to evolving stress (proxy of dominating dislocation creep in high-stress mantle condition) (Muto et al., 2019 Sci. Adv. and references therein).

We found that the localized strain contraction near the active volcanoes can be explained by small-scale LVZs which have narrow tops of 10–20 km and wide roots of 60–100 km width. Our results conclude the minimum depths of the tops and roots of LVZs are 15 km and 40 km, respectively. The geometries of the LVZs vary (e.g., upright conic or inclined shape) from volcano to volcano. The effective viscosities of the LVZs are in the order of 10¹⁷ Pa·s immediately after the earthquake and increases to the order of 10¹⁸ Pa·s over the 3 years of postseismic deformation. Our results agree with the results of several past studies (Ohzono et al., 2012 EPS; Hu et al., 2014 EPS; Muto et al., 2016 GRL) who investigated the lithospheric rheology near Mt. Naruko using the postseismic surface displacements of the 2011 Tohoku-oki and 2008 Iwate-Nairiku earthquakes.

How to cite: Dhar, S., Takada, Y., and Muto, J.: Rheology of weak lithosphere beneath active volcanoes of NE Japan: Insights from postseismic deformation of 2011 Tohoku-oki earthquake, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13903, https://doi.org/10.5194/egusphere-egu24-13903, 2024.

A large crustal earthquake (Mw=7.5) struck the Noto Peninsula, central Japan, at 16:10 (JST = UT + 9 hours) on New Year's Day, 2024. The main-shock rupture extended ~150 km in length, which covered the source regions of intense swarm activity in the northeastern tip of the peninsula [Amezawa et al., 2023] as well as the previous large crustal earthquakes such as the 2007 (Mw=6.7) and 2023 (Mw=6.3) events. The aftershock distribution of the 2024 event provides fundamental information for understanding the rapture process of the main shock and seismotectonics in the Noto peninsula. Therefore, we relocated the earthquake hypocenters that occurred immediately after the 2024 event by considering the three-dimensional velocity structure [Matsubara et al., 2022]. In the relocation, we applied the method proposed by Shiina and Kano [2022] to the arrival time data on the earthquake catalog compiled by the Japan Meteorological Agency. The applied method utilized the Markov Chain Monte Carlo technique, allowing us to evaluate uncertainty in hypocenter locations. Thus, we can discuss the distributions of the crustal earthquakes in and around the source area of the 2024 event, taking account of the spatial variations in uncertainty in the hypocenters. For example, some aftershocks occurred offshore, indicating that estimation accuracy in that area may get worse due to limited station coverage compared with the inland area. As the result of the relocation considering the three-dimensional structure, the depth of these offshore events was shifted about 5 km shallower. These hypocenters suggested that the aftershocks of the 2024 event occurred mainly between the ground surface and the depth of 15 km.

How to cite: Shiina, T., Horikawa, H., and Imanishi, K.: Relocations of earthquake hypocenters in and around the source area of the 2024 Mw 7.5 Noto Peninsula earthquake, Japan, by Bayesian inference, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14903, https://doi.org/10.5194/egusphere-egu24-14903, 2024.

In this study, we document unusual and recurring events that transpired within one hour on November 17, 2021. These incidents were identified through a seismometer array deployed in the Yilan area and Turtle Island, northeastern Taiwan. Preceding this series of events, a shallow submarine volcano near Turtle Island emitted sulfur smoke from October 28, 2021, lasting until November 22, 2021. This eruption was marked by a significant release of white sulfur smoke from the sea near Turtle Island. It reached a height exceeding 3 meters and extended over 100 meters into the air, making it the most substantial eruption of the year. At first, we proposed that the giant bubble could be generated during the submarine eruption and expanded through the water and into the atmosphere; the collapse of this bubble was considered a potential source of the recurring events. However, a grid-search method utilizing the arrival times of seismic stations indicates that the source location is close to the seacoast of Yilan, still dozens of kilometers away from Turtle Island. Upon further analysis of the seismic waveforms, it was observed that the propagation velocity is close to the speed of sound and only detected by surface stations, not by shallow-hole stations. This suggests that the source likely produced signals that couple well with the atmosphere rather than the solid Earth. The waveforms exhibit high consistency between different events at the same station, indicating that the sources occurred at the exact location several times within one hour. The possibility of an aircraft-induced shock wave was considered but needs more investigation. Trustworthy sources and their mechanisms remain to be clarified, and additional data, such as infrasound and pressure data, will be essential for a more comprehensive understanding shortly.

How to cite: Ku, C.-S., Huang, B.-S., and Lin, C.-H.: Repeating events detection in northeastern Taiwan using a broadband seismometer array, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15253, https://doi.org/10.5194/egusphere-egu24-15253, 2024.

When in 1997 we started to compute Regional CMT for seismic events in the Euro-Mediterranean region, we could not expect to create a Catalog that can well describe the seismicity, the seismotectonics of this really active region with such complex characteristics. The RCMT Catalog includes more than 3200 seismic moment tensors for earthquakes with a magnitude starting from 4.5, but for the Italian region also down to M 4.0, for the time span 1997 to 2023. All RCMTs are available on the web, on dedicated pages, with the possibility to select the preferred dataset choosing intervals for time, geography, magnitude, depth and quality factors (http://rcmt2.bo.ingv.it/searchRCMT.html). In the first years the RCMT computation was based only on the modelling of intermediate-period surface waves. After 2002, it has been possible to invert also for body waves, an improvement that for the RCMT computation has been important for events with M greater than 5.0. The homogeneity of the dataset given by the continuous use of the same algorithm is an added value that has been underlined by several comparisons with other regional catalogs. The lower magnitude threshold applied in the Euro-Mediterranean region produces a dataset three times more numerous with respect to what is available with the Global CMT data only. In 1997 RCMTs were the only seismic moment tensors available for earthquakes with M lower than 5.0 in the Euro-Mediterranean region. Later, several regional and local focal mechanisms have been computed, with different methods and for different sub-regions. At present, on average three or more regional solutions appear on the web after a M4.5 earthquake hits the Mediterranean. However, RCMT Catalog is the one with the longer time interval covered by homogeneous data. Today, the Catalog is continuously updated with a few months of delay between definitive and quick solutions, that are however available on the RCMT web pages up to the time the revised solution is ready.

How to cite: Pondrelli, S.: The European Mediterranean RCMT Catalog: more than 25 years of activity and data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15437, https://doi.org/10.5194/egusphere-egu24-15437, 2024.

On January 1st 2024, a Mw 7.6 earthquake shook the Noto Peninsula on the Sea of Japan coast of Central Japan causing over 202 casualties and >100 missing (at the time of submission). The quake follows a period of intensifying seismic activity starting in 2020. The Mw 6.3 Oku-Noto earthquake of May 5 2023 was the previous largest event of the sequence. The Jan. 1 2024 Noto Peninsula earthquake significantly impacted the Peninsula. A large number of landslides and rockfalls dissected the road network. Liquefaction damaged infrastructure up to 150 km away from the epicenter. Meter-scale coseismic uplift modified the northern shoreline with displacement of the coastline by up to 200 m seaward discernible on SAR and aerial image data. At the time of abstract submission (Jan. 10 2024) we only have limited preliminary observations. It appears that the Noto Earthquake ruptured the same or adjacent fault to the May 5 2023 Mw 6.5 earthquake and was in the vicinity of the March 25 2007 Mw 6.9 Noto earthquake. Coseismic displacement measured geodetically shows uplift of up to +3–4 m (SAR) in the northwest of the peninsula (Wajima-shi), and +1.06 m (GPS) in the main town of Wajima-shi. The uplift magnitude decreases gradually to the SE. The uplift is near zero (SAR) or -0.3 m (GPS) on Noto Island (Nanao-shi) 30 km to the south of the town of Wajima. Surface deformation goes back to near zero (GPS) a further 20 km to the south.

The coseismic deformation pattern broadly reflects the deformation recorded in the Noto landscape. Long-term moderate rock uplift in the north gives way to a complex history of long-term slow uplift around Noto Island that likely includes sustained episodes of subsidence, highlighted by its sinuous “drowned” coastline. Along the western shore (Shika-machi), marine terraces presumed to be 120 ka (last Interglacial) show a gradient in elevation also decreasing to the south. In the north, the newly emerged platform does not have a higher marine terrace counterpart. This may reflect the relationship between high wave power and moderate rock uplift resulting in the long-term retreat of the coastline and erosion of any terrace. The Noto Peninsula also holds widespread evidence of drainage reorganization that would reflect varying boundary conditions, in particular rock uplift, in deeper time beyond 100’s ka. The similarities between recent landscape morphology and coseismic displacement suggest that the Jan. 1 2024 rupture fits a recent pattern of crustal strain in Noto Peninsula (at least up to 100 ka). Earlier deformation pattern (>100’s ka) likely happened along different faults and/or at different rates as reflected by the transient drainage network.

By conference time, we will present field observations collected after the rescue and emergency work is completed.

How to cite: Malatesta, L. C., Sueoka, S., Kataoka, K. S., Komatsu, T., Tsukamoto, S., Bruhat, L., and Olive, J.-A.: Geology and geomorphology of the Jan 1st 2024 Mw 7.6 Noto Peninsula Earthquake: observations and context., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15494, https://doi.org/10.5194/egusphere-egu24-15494, 2024.

Numerous episodes of volcanic unrest have taken place at Vulcano island (Italy) since its last Vulcanian eruption occurred 133 years ago. Decades-long seismic monitoring has documented some of them. We have collected and examined all available seismic data recorded since 1985, most of which were in analog format and/or dispersed in old repositories. We were able to compile catalogs where three different types of seismic events are considered according to their location and magnitude: events in the Fossa Crater, in the Lipari-Vulcano complex, and earthquakes with M>2.5. The primary goal of this data collection was to identify the main features of seismic activity on and around the island in the 36 years preceding the last volcano unrest, which began in mid-September 2021 with a high occurrence frequency of Very Long Period (VLP) events. Our review of the past seismic activity allows us to contextualize the newly recorded anomalous variations. Furthermore, we sought the connection with the structural framework of the region.

The duration of the episodes of volcanic unrest in 1985 and 1988 was relatively short (lasting just a few months) when compared to the recent one, which ended in December 2023. The source of the seismic events during those past unrests was mainly close to the reference station (now IVCR) with hypocenters mostly beneath the island at shallow crustal depths (up to 5 km below sea level). Their magnitude remained low (<2.5) during both the episodes (i.e., 1985, and 1988).

Overall, the seismicity recorded in and around the island has reached a maximum value of M4.6 both in the 36 years preceding and during the 2021 unrest. Some preliminary insights can be drawn by comparing the seismicity occurred during past and recent unrest episodes: i) the peculiarly long duration of the most recent unrest, and ii) the importance of broadband equipment, which documented the substantial contribution of VLP seismicity during the 2021-2023 episode.

How to cite: Falsaperla, S., Barreca, G., Cocina, O., and Spampinato, S.: Insights into the 36 Years of Seismic Activity at Vulcano Island, Italy, preceding the Volcanic Unrest in 2021, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15942, https://doi.org/10.5194/egusphere-egu24-15942, 2024.

In this study, we have evaluated the scattering nature of the crust beneath the Bhutan Himalaya, located in the eastern part of the Himalayan arc. We have analysed high-quality waveforms of 566 events having magnitude (M_L) below 6, recorded by broadband stations of the GANSSER network operated by the Swiss Seismological Service at ETH Zurich from Jan, 2013 to Nov, 2014. We have investigated the peak delay time (t_pd), defined as the time interval between the initial S-wave appearance and the peak amplitude of its envelope, for the frequency ranges of 4–8, 6–12, 8–16 and 12–24 Hz. Initially, we have analysed frequency-dependent nature of t_pd at 9 stations (BHE01, BHE09, BHE13, BHN02, BHN06, BHN11, BHW01, BHW10 and BHW16). The observed values of B_freq, which indicates the frequency dependence of the peak delay time, show mostly low positive values up to 0.3. It shows that t_pd is independent of frequency which may be associated with the relative proportions of large as well as small scale heterogeneities present in the medium_. At BHE09, B_freq exhibits a negative value, which might be attributed to the limited sampling of high-frequency signals that capture small portions of the subsurface along their paths. The crust beneath BHE09 experiences reduced scattering, probably due to the absence of a strongly attenuating body in the subsurface. Furthermore, we aim to extend this study for all the stations and to compare the frequency-dependent nature of T_5%-75% (time interval between 5% and 75% of the total integrated power value) and the t_pd for the study region.

How to cite: Bala, S., Dutta, A., and Singh, C.: Frequency-dependent delay time analysis for Bhutan Himalaya, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17485, https://doi.org/10.5194/egusphere-egu24-17485, 2024.

The North Anatolian Fault Zone in Turkey spans 1400 km, passing through densely populated areas, including Düzce, which experienced the destructive Mw 7.2 event in 1999 that caused more than 700 lives. On 23 November 2022, for the first time in over 20 years, a moderate Mw 6.1 earthquake struck the city and surrounding area. Despite its moderate magnitude, the event caused unexpectedly severe damage to numerous buildings, as reported by local institutions (Disaster and Emergency Management Presidency; AFAD). Recognizing the potential impact of near-field effects such as ground motion pulses and directivity effects, which are known to increase damage in the vicinity of the fault, we investigate these phenomena using the AFAD-Turkish Accelerometric Database. Our analysis delves into the spatial distribution of ground motion intensities, revealing higher peak ground velocities in certain azimuthal ranges than predicted by existing ground motion models. Surprisingly, our findings challenge outcomes derived from previous studies, suggesting that impulsive ground motions associated with directivity effects mainly occur on the fault-normal component of large-magnitude events. In contrast, our examination of near-fault recordings indicates a concentration of velocity pulses, primarily on the fault-parallel component, and thus questions the widely established understandings of earlier studies.

How to cite: Türker, E., Yen, M.-H., Pilz, M., and Cotton, F.: Importance of Pulse-Like Ground Motions and Directivity Effects in Moderate Earthquakes based on the 23 November 2022, Mw 6.1 Gölyaka-Düzce Earthquake (Turkey)., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17715, https://doi.org/10.5194/egusphere-egu24-17715, 2024.

We re-compute the coefficients of the intensity prediction equation (IPE) in Italy using the data of the DBMI15 V2.0 intensity database and the instrumental and combined (instrumental plus macroseismic) magnitudes reported by the CPTI15 V2.0 catalog. We follow the same procedure described in a previous article, consisting of a first step in which the attenuation of intensity I with respect to the distance D from macroseismic hypocenter is referred to the expected intensity at the epicenter I_Eand a second step in which I_Eis related to the instrumental magnitude M_i, the combined magnitude M_c, the epicentral intensity I₀ and the maximum intensity I_max, using error-in-variable (EIV) regression methods. 

The main methodological difference with respect to the original article concerns the estimation of the uncertainty of I_Eto be used for EIV regressions, which is empirically derived from the standard deviation of regression between I_E and M_iand also used for the regressions of I_Ewith M_c, I₀ and I_max. 

In summary, the new IPE determined from DBMI15 V2.0 is

                                        𝐼=𝐼_𝐸−0.0081(𝐷−ℎ)−1.072[ln(𝐷)−ln(ℎ)]

 where 𝐷=√(𝑅²+ℎ²), h=4.49 km and I_Ecan be calculated from the intensity data distribution of the earthquake. If the intensity data distribution is not available, I_Ecan be calculated from the following relationships

                                        𝐼_𝐸=−2.578+1.867𝑀_𝑤

                                                      I_E=I₀

How to cite: Lolli, B., Gasperini, P., and Vannucci, G.: Recalibration of the intensity prediction equation in Italy using the Macroseismic Dataset DBMI15 V2.0, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18530, https://doi.org/10.5194/egusphere-egu24-18530, 2024.

The OTRIONS seismic network (University of Bari Aldo Moro, 2013, FDSN code OT) is a local network installed in the Apulia region (Southern Italy) with the aim of monitoring the seismicity of the Gargano area (Northern Apulia) and the Salento area (Southern Apulia). OT network is managed by the University of Bari Aldo Moro (UniBa) and by the National Institute of Geophysics and Volcanology (INGV). It started to operate in 2013 and in 2019 the recording stations migrated to EIDA (all details can be found in Filippucci et al., 2021a). In 2021 a first database was collected, with the event detection achieved both manually and automatically with SeisComP3 (Helmholtz-Centre Potsdam), and was released (Filippucci et al., 2021a; Filippucci et al., 2021b).

Now, after ten years of operations, we focus on the microseismicity of the Gargano area with the aim of collecting a new seismic database for the period from April 2013 to December 2022, by using a recently acquired software, CASP (Complete Automatic Seismic Processor), for the automatic detection, picking and location of seismic events (Scafidi et al., 2019). The CASP software is installed on a remote server implemented by RECAS-Bari, the computational infrastructure of INFN and UniBa.

Through an appropriate parameter setting, we adapted CASP and NonLinLoc (Lomax et al., 2000) to the Gargano area and to the seismic stations available, both OT and INGV. We used the 1D velocity model of Gargano (de Lorenzo et al., 2017).

The recorded seismic events were organized in two catalogs: the first one is the automatic catalog, obtained from the automatic locations of CASP; the second one is the manual catalog, obtained through a manual revision of P and S waves arrival times. To evaluate the reliability of CASP, a comparison between the automatic and manual catalog was performed.

From a comparison of the manual catalog with the already released catalog of the Gargano seismicity (Filippucci et al., 2021b), the number of events detected by CASP increased a lot. Furthermore, the results show that the choice of the CASP parameters allows us to lower the minimum magnitude threshold of the recorded microseismicity in the Gargano area. Preliminary analysis of the earthquakes foci shows that the seismicity pattern retrace, substancially, the same discussed in the work of Miccolis et al., 2021.

How to cite: Ferreri, A. P., Romeo, A., Giannuzzi, R., Cecere, G., de Lorenzo, S., Falco, L., Filippucci, M., Michele, M., Selvaggi, G., and Tallarico, A.: The new earthquake catalog of the Gargano (Southern Italy) OTRIONS seismic network., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18752, https://doi.org/10.5194/egusphere-egu24-18752, 2024.

Seismic networks monitor seismic activities across the globe, recording distinctive events within specific geographical and temporal frames. Whether old or new, each seismic record preserves valuable information, with its extraction relying mainly on the sophistication of the method. This study presents the implementation of an advanced earthquake detection workflow on a relatively old dataset, the Bhutan Pilot Experiment. This temporary five-station seismic network in Eastern Himalaya comprised a set of Broadband sensors deployed for 14 months from January 2002 to March 2003. However, outdated methodologies have limited the analysis of the recorded data, resulting in the reporting of only 175 local microearthquakes in this area. In this study, we reprocess the data using the recently introduced deep-scan Integrated Pair-Input deep learning and Migration Location workflow [1] to detect and locate local earthquakes. The IPIML employs the well-known Earthquake Transformer (EqT) model as its core function for initial phase picking, followed by a pair-input Siamese EQTransformer (S-EqT) to further mitigate the false negative rate using a pair-wise model. The S-EqT step demonstrated an approximately 40% increase in average detected phases compared to the standard EqT model. The detected phases are associated using the Rapid Earthquake Association and Location (REAL) method through grid searching, providing a preliminary list of detected events. This list encompasses 2458 detected events, several times larger than the previously reported catalog for this dataset. These events primarily cluster in central and eastern Bhutan, particularly along the Golpara lineament, a recognized strike-slip fault. The subsequent phase of this study involves precisely locating these events through the implementation of the Migration Location (MIL) method.

References
[1] H. Mohammadigheymasi et al., "IPIML: A Deep-Scan Earthquake Detection and Location Workflow Integrating Pair-Input Deep Learning Model and Migration Location Method," in IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1-9, 2023, Art no. 5914109, doi: 10.1109/TGRS.2023.3293914.

How to cite: Khurshid, Z., Mohammadigheymasi, H., Gao, D., Liu, J., and Mousavi, S. M.: Deep Scanning of the Bhutan Eastern Himalaya Seismic Dataset for Local Earthquakes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20056, https://doi.org/10.5194/egusphere-egu24-20056, 2024.

Modern ground motion prediction equations used in probabilistic seismic hazard assessment studies are now almost exclusively expressed as a function of moment magnitude M_W. Yet, earthquake catalogues produced by seismic observatories often provide local magnitude M_L only. It is for instance the case for the Laboratoire de Détection Géophysique (LDG) catalogue recently published by Duverger et al. (2021) for mainland France. A conversion relationship was proposed by Cara et al. (2015) from M_L (LDG) to M_W. It appears however that this relationship does not result in a good fit when compared to recently compiled M_W values for France and neighboring areas (Laurendeau et al., 2020). In this work, we propose a new conversion relationship based on the inversion of M_L-M_W couples for events present in both the LDG catalogue and the compilation of Laurendeau et al. (2021). In order to avoid the choice of an arbitrary number of segments to model the M_W=f(M_L) relationship, the inverse problem is solved in a Bayesian framework by means of the reversible jump Markov chain Monte Carlo (rj-McMC) algorithm (Green, 1995; Bodin et al., 2012). The number of segments, as well as their respective slopes and intercepts, are jointly invert for. As moment magnitude uncertainty is not known, it is also considered as an unknown, while the M_L uncertainties provided in the LDG catalogue are fully accounted for by random sampling during the McMC process. We observe a geographical dependence of the differences between the available M_W values and those obtained from calculation so a location-dependent term is also modeled. This allows to account for the regional attenuation variations that can affect M_L estimates. The new conversion law is then applied to the full LDG catalogue and its impact on seismic hazard assessment is explored.

How to cite: Arroucau, P., Duverger, C., Traversa, P., Daniel, G., Mayor, J., and Mazet-Roux, G.: Local to moment earthquake magnitude conversion in mainland France and implications for seismic hazard assessment, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20889, https://doi.org/10.5194/egusphere-egu24-20889, 2024.

A destructive M7.6 earthquake occurred on January 1st, 2024, at shallow depths along the northern coast of Noto Peninsula on the back-arc side of Central Japan. The earthquake rupture started from an area where an intensive seismic swarm has lasted for more than 3 years (from December 2020). The seismic swarm consisted of numerous small planar faults dipping toward the southeast. In May 2023, an M6.5 event, that was the largest event before the M7.6 rupture, emanated from the swarm area toward shallow depths, resulting in the subsequent increase in the seismicity in the swarm area (Kato 2024 GRL). Then, the seismicity had gradually decayed to a level before the 2023 M6.5 event. Here we have explored the seismic and geodetic data to revel the nucleation process of the M7.6 event. Approximately two weeks before the M7.6 event, seismic activity exhibited a weak localization around the point of rupture initiation. After that, a foreshock sequence commenced roughly one hour before the occurrence of the M7.6 event, concentrated in proximity to the epicenter (within a 1-kilometer epicentral distance). The tightly clustered foreshock sequence consisted of around 20 events, including an M5.5 event 4 minutes prior and an M3 class event 1 second before the onset of M7.6 event. The M7.6 rupture nucleated from the deep side of one of planar clusters that were dominantly dipping toward the southeast direction. The growth process of the rupture in the M7.6 event is characterized by a complicated nature.

How to cite: Kato, A. and Nishimura, T.: Foreshock sequence prior to the 2024 M7.6 Noto-Hanto earthquake, Japan, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22522, https://doi.org/10.5194/egusphere-egu24-22522, 2024.

The tsunami generated by the Mw7.6 earthquake of Noto Peninsula, Japan left widespread impact. We analyzed multi-modal information and data to elucidate its impact.

We modeled the tsunami propagation and inundation with multiple tsunami source models based on GNSS-based crustal movement and tsunami waveform data to understand its propagation and inundation characteristics. The model results are verified by using post-tsunami field survey data. Preliminary tsunami modeling results implied that severe tsunami impacts were around Noto Peninsula (Shika to Nanao). Through the visualization of tsunami propagation model, we found that the remarkable tsunami refraction around the continental shelf of Noto Peninsula were responsible for high tsunamis in Suzu City. This distinctive sea bottom topography also affected the directivity of tsunami energy toward the Japan sea coasts, especially Joetsu city, Nigata Prefecture. Tsunami in Toyama bay had long duration of oscillation caused by multiple-reflection. The leading (negative) tsunami wave could not be explained by fault rupture and this implied the possibility of submarine landslides.

The post-tsunami field survey teams at Suzu City preliminarily found tsunami run-ups of 3 m or higher with flow depths of 2.5m or higher. Inside the tsunami inundation zone around Noto Peninsula, we found at least 648 houses out of 3398 were destroyed by both the strong ground motion and tsunami.

The cellphone-based population data (Mobile Spatial Statistics) were used to analyze the exposed population in the aftermath of the event. The hourly population estimates with 500m spatial resolution in the coastal communities implied how people reacted and were affected. Approximately 2500 population increase were found in the areas above 10 m after the major tsunami warning was issued.

How to cite: Koshimura, S., Adriano, B., Mizutani, A., Mas, E., Ohta, Y., Nagata, S., Takeda, Y., Vescovo, R., Wiguna, S., Abe, T., and Suzuki, T.: The Impact of the 2024 Noto Peninsula Earthquake Tsunami, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22523, https://doi.org/10.5194/egusphere-egu24-22523, 2024.

This presentation will report preliminary results of multifaceted analyses for the geomorphological aspects of the Mw 7.5 earthquake struck northern tip of the Noto Peninsula, Japan, at 16:10JST on January 1, 2024. The earthquake caused significant uplift of the northern coastal areas of the peninsula, accompanying a tsunami observed widely in the surrounding coastline, along with extensive tectonic deformations observed inland. Spatial extent of the crustal movements accords generally with the relief structures and distribution of marine terraces in the Noto Peninsula, implying the long-term tectonic forcing on the landscape evolution in this region. Numerous coseismic landslides occurred in steep mountainous terrains, which yield vast volume of sediment from hillslopes into fluvial channels. Inventory mapping revealed the localized distribution of the landslides, regulated most probably by geologic and topographic conditions. Areal density of the landslides can be explained by coupled factors of lithological susceptibility of the hillslopes to the seismic shaking and amplification of ground motion at the hilltops.

How to cite: Matsushi, Y.: Geomorphological consequences of the 2024 Noto Peninsula Earthquake: tectonic deformations, coseismic landslides, and their implications, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22535, https://doi.org/10.5194/egusphere-egu24-22535, 2024.

Since November 30, 2020, an intense earthquake swarm with over 22,000 M≥1 earthquakes and transient deformation have been continuously observed in the Noto Peninsula, central Japan, which is a non-volcanic/geothermal area far from major plate boundaries. During the earthquake sequence, M_w6.2 and M_w7.5 earthquakes occurred on May 5, 2023, and January 1, 2024, respectively. We report the transient and coseismic deformation related to the earthquake sequence by a combined analysis of multiple Global Navigation Satellite System (GNSS) observation networks, including one operated by a private sector company (SoftBank Corp.), relocated earthquake hypocenters, and tectonic settings. The start of the transient deformation coincides with a burst-type activity of small earthquakes in late 2020. A total displacement pattern in the first two years shows horizontal inflation and uplift of up to ~60 mm around the source of the earthquake swarm. The overall deformation rate gradually decreased with time except for the coseismic displacement of the M_w 6.2 earthquake and its postseismic displacement. On January 1, 2024, the coseismic horizontal and vertical displacements reached ~2 m at several GNSS sites. The pattern of the postseismic displacement for the first three weeks is similar to that of the coseismic displacement, though spatial decay of the postseismic displacement from the epicentral area is much gentler than that of the coseismic displacement. Viscoelastic relaxation of the mantle and/or lower crust is probably an important factor in explaining the observed deformation. In order to explain the transient deformation before the M_w6.2 and M_w7.5 earthquakes, we assumed a southeast-dipping fault plane based on the observed seismicity and regional tectonics and estimated the distribution of both reverse-slip and tensile components on the fault plane. In the first three months, a significant tensile component with a small slip component was estimated around a depth of ~15 km. The estimated volumetric increase is ~1.4 x 10⁷ m³. Over the next 15 months, the observed deformation was well reproduced by shear-tensile sources, which represent an aseismic reverse-type slip and the opening of the southeast-dipping fault zone at a depth of 14–16 km. These slips and openings of the fault are estimated mainly at the down-dip extension of the intense earthquakes. We suggest that the upwelling fluid spread at a depth of ~16 km through an existing shallow-dipping permeable fault zone and then diffused into the fault zone, triggering a long-lasting sub-meter aseismic slip below the seismogenic depth. The aseismic slip further triggered intense earthquake swarms including the M_w6.2 and M_w7.5 earthquakes at the updip.

Acknowledgments: We are grateful to SoftBank Corp., ALES Corp., and GSI for providing us with GNSS data.

How to cite: Nishimura, T., Hiramatsu, Y., and Ohta, Y.: Deformation of the 2020-2024 Noto Peninsula earthquake sequence revealed by combined analysis of multiple GNSS observation networks in central Japan, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22539, https://doi.org/10.5194/egusphere-egu24-22539, 2024.

The Noto Peninsula earthquake (M_J7.6, M_W7.5) caused extensive damage to buildings and infrastructure in the Noto Peninsula located in the northern part of Ishikawa prefecture, Japan. The hypocenter was within the area of the earthquake swarm that started in 2020. However, the source fault bilaterally ruptured over a length of 150 km beyond this area. The main residential areas in Wajima, Suzu, and Anamizu are located almost above the western segment of the reverse fault. The geographical features of the Noto Peninsula pose significant challenges for aid and support, particularly due to embankment and soil failures that caused main road closure or limited access. This has led to increased traffic on the few accessible routes, further delaying the arrival of support. The situation has hindered the restoration of essential services such as water and sewage systems and has slowed down the process of demolishing buildings deemed dangerous.

Valuable strong motions were observed during this event. The maximum Peak Ground Acceleration (PGA) in the horizontal component reaching 2.78g was recorded at the K-NET ISK006 station, a location known for significant site amplification around 0.2s. This value aligns with the dominant period in the Spectral Acceleration (Sa), thus the extreme PGA was probably due to the enhanced short-period component in the shallow soil amplification. In addition, K-NET ISK002 and ISK005 recorded large PGVs of 1.31 m/s and 1.59 m/s, respectively, and observed the remarkable Sa with 1.3g and 2.2g at T=1.0s, respectively, which are similar to the damage-prone record in the 1995 Kobe earthquake (JR Takatori record).

In the main residential areas of Anamizu and Wajima, two seismic stations are operated. One is located on the stiff soil ground, and the other is located in zones where residential damage was most severe. In both Anamizu and Wajima, the records at the damage site were amplified in the periods of 1-4 s, suggesting that the residential damage is related to the site amplification. Since the spectral ratio of the weak motions shows the amplification at periods of less than 1s, the major reason for the amplification at periods of 1 to 4 seconds during the main event is due to the nonlinear response of the soil ground.

How to cite: Goto, H., Nakatsuji, A., Huang, D., and Montoya-Noguera, S.: Ground motions and geotechnical aspects of the Noto Peninsula earthquake, Japan, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22540, https://doi.org/10.5194/egusphere-egu24-22540, 2024.

The Niigata (Echigo) Plain facing the Sea of Japan is located downstream of two large rivers (the Shinano-gawa River and the Agano-gawa River), and has three sand dune ridges which formed along the coastal areas during the Holocene. Niigata city, with a population of ~770,000, lies in the lower catchment of the alluvial-coastal system. Despite the city being approximately 160 km away from the epicenter of the January 1st 2024 Mw 7.6 Noto Peninsula Earthquake, extensive damage to houses, buildings, and infrastructure occurred throughout Niigata city due to pervasive liquefaction (resulting from the earthquake) in the coastal and lowland areas.

Our field investigation focuses on the Nishi-ku (west ward) of the city, where much of the liquefaction-induced building damage (~ 700 houses at the time of submission of the abstract) is concentrated. Although our “ground truth” fieldwork is still ongoing, we have manually mapped the distribution of damaged houses/buildings, road deformation, sand boiling (sand volcanoes), cracks, slides, groundwater springs and other related phenomena onto map sheets, before then digitising these data using GIS.  The distribution of damage is concordant with geomorphology—such as the Holocene sand dunes (and associated landforms) and buried meander loop of the Shinano-gawa River—as well as with subsurface geology (e.g. the location of the water table). Some damage areas are coincident with artificially modified landforms.

Liquefaction conspicuously occurred on natural (i.e. not artificially modified) gentle slopes of the Holocene coastal sand dunes and interdune swale/lowland. In particular, ground was liquefied in the lower parts of the landward slope of the sand dune (formed ~1800­–900 years ago) which has a lateral extension of ~7 km at the elevation of ~0–3 m above sea level. Sandy subsurface geology and high groundwater level of the Holocene sand dune, together with the force of gravity on the slopes, were probable contributors to liquefaction.

Evidence for liquefaction —including damage to houses—was observed in modern residential areas developed above the buried meander loops of the Shinano-gawa River, which have been historically filled in artificially with sandy material. Damage was also noted in houses built upon an artificially buried pond. However, there was no liquefaction on the natural levee along the abandoned meander loops where relatively old settlements are present.

Similar liquefaction occurred in Niigata city on the sand dune slopes and associated lowlands at the time of the M 7.5 Niigata Earthquake in 1964; the epicenter was in the Sea of Japan, approximately 60 km from the city.  Despite the Noto Peninsula Earthquake occurring remotely from Niigata, the aftermath of the earthquake indicates that certain geomorphologic and geological factors, coupled with particular seismic conditions, can result in repeated liquefaction. 

The field observation is still ongoing after the earthquake. Therefore this abstract is based on tentative results and analysis of our investigation so far. Further information on liquefaction related to the geomorphology and subsurface geology in this area will be available by the time of the 2024 EGU General Assembly.

How to cite: Kataoka, K., Urabe, A., Nishii, R., Matsumoto, T., Niiya, H., Watanabe, N., Kawashima, K., Watabe, S., Takashimizu, Y., Fujibayashi, N., and Miyabuchi, Y.: Extensive liquefaction and building damage on the Niigata Plain due to the 1 January 2024 Noto Peninsula Earthquake: Geomorphological and geological aspects and land-use in coastal and lowland areas, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22541, https://doi.org/10.5194/egusphere-egu24-22541, 2024.

Seismic hazard evaluation at critical infrastructures, such as nuclear power plant, urges deeper understanding on the fundamental physics that govern the initiation, propagation and termination of damaging earthquakes. The 2024 moment magnitude (M_w) 7.5 Noto Peninsula earthquake produced great hazards and exhibited complex rupture process. We derive high-resolution 3D surface deformation of the event using dense space geodetic observations, which reveal two major deformation zones separated by ~40 km along the coast of the Peninsula. Two large (>10m) shallow slip asperities with over 10 MPa stress drop on the thrust faults explain excellently the geodetic observations. A calibrated back-projection using teleseismic array high-frequency data shows that the rupture was stagnant around the hypocentre for ~20s before it propagated bilaterally at the speed of 3.4 km/s and 2.8 km/s towards southwest and northeast, respectively. The slow start of the rupture coincides with the seismic swarm surged since 2020 due to lower crust fluid supply, suggesting low normal stress (high pore fluid pressure) at the bottom edge of the seismogenic zone slowed down the initial rupture. The first major asperity of the rupture was accompanied with intense high frequency seismic radiation, and such radiation is even stronger from the largest asperity located at the southern edge of the Peninsula where the Peak-Ground-Acceleration (PGA) exceeding 2.6G at a site that is less than 40km away from the nuclear power plant. Large stress accumulation together with rough fault geometry and/or friction are likely responsible for the exceedingly large high-frequency radiation, which is mostly responsible for devasting damages.

How to cite: Luo, H., Ma, Z., Zeng, H., and Wei, S.: The 2024 Mw 7.5 Noto Earthquake, shallow rupture with a stagnant initiation in a fluid-rich immature fault zone, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22563, https://doi.org/10.5194/egusphere-egu24-22563, 2024.

The propagation of rupture at the interface between a layer and a half-space, arising from their relative slipping, is commonly observed in geophysics and in technological materials. Fault ruptures develop as a consequence of ongoing aseismic interfacial movement caused by geological forces; and accumulation of stress over time ultimately leads to a sudden stress release resulting in seismic phenomena. In mathematical terms, it is a highly nonlinear and multiscale phenomenon that necessitates the accurate solution of elastodynamic equations and interfacial fault friction across an extensive domain over an extended time period. The most numerically efficient algorithm for simulating this process is the Boundary Integral Equation Method (BIEM). It computes field quantities at the fracture plane and reduces domain dimensionality of the problem by one, but its current applicability is mostly limited to planar interfaces and to unbounded geometries. BIEM computes elastodynamic fields at the fracture plane from a space-time traction history by utilizing suitable convolution kernels (dependent on the geometry and the material/bi-material characteristics across the interface). Applying BIEM to a geometry involving a layer over a half space presents significant challenges, particularly due to the difficulty in analytically deriving convolution kernels for finite geometry and in-plane deformations. Here, we develop an integrated approach, combining the traction formulation of BIEM (Ranjith, 2015) with the Finite Difference Method (FDM), to overcome these challenges. A hybrid FDM-BIEM has previously been used by Hajarolasvadi and Elbanna (2017) using the velocity formulation of BIEM (Geubelle and Rice, 1995) for unbounded geometries. In the present work, a fourth-order staggered-grid FDM is employed to model dynamic rupturing in a layer. The layer is adjoined by an elastic half-space. The elastodynamic equations in the half-space are handled using a BIEM. The FDM region incorporates a slip-weakening friction law at the fault interface using a thick fault zone model consisting of two grid rows (Madariaga, Olsen and Archuleta, 1998). By applying traction boundary conditions, it computes updated velocities in the domain, and passes them to BIEM. BIEM, in turn, responds with updated tractions at the boundary, utilizing convolution kernels derived by Ranjith (2015). We utilized the hybrid FDM-BIEM numerical algorithm to study the effect of finite layer thickness on rupture propagation, considering both homogeneous and bi-material fault interfaces. The results suggest that the present method has the capacity to effectively deal with a wide array of problems in finite domains often countered in both geophysics and technological domains.

How to cite: Painuly, A. and Kunnath, R.: Dynamic Slip Rupture at a Layer-Half-space Interface: A Hybrid Finite Difference and Spectral Boundary Integral Equation Method for Numerical Simulation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4263, https://doi.org/10.5194/egusphere-egu24-4263, 2024.

The Eastern Betic Shear Zone (EBSZ) experiences slow seismic deformation that leads to relatively low seismicity rates. Due to this, historical records underscore the substantial impact that earthquakes have had on local communities. The dearth of comprehensive data on moderate to large seismic events in this area, limits the accurate generation of seismic hazard and risk maps, posing a significant challenge for seismic risk planning. A way to address these limitations is leveraging physics-based earthquake simulations in the Southeast Iberian Peninsula. These simulations first require integrating paleoseismic data, models of fault distribution –such as the Quaternary-Active Faults Database of Iberia, seismic source characterizations and historical seismic catalogs, to construct an Earthquake Rupture Forecast (ERF), where likelihood of each fault rupture is weighted by an occurrence probability. Our study focuses on developing physics-based rupture scenarios and shake maps using CyberShake. CyberShake is designed to perform physics-based probabilistic seismic hazard assessments (PB-PSHA) by simulating a vast set of synthetic ground-motion time histories from kinematic rupture scenarios on the ERF three-dimensional finite-fault array. Originally tailored for PB-PSHA studies in Southern California by the SCEC (Southern California Earthquake Center); this research represents the first CyberShake application for Southeast Spain. The resulting shake maps represent an alternative basis for updating regional probabilistic seismic hazard maps and also could support crucial decision-making processes following a local earthquake, offering valuable insights for effective response strategies.

How to cite: Zamora, N., Monterrubio, M., Rojas, O., Blanco, R., Bhihe, C., and de la Puente, J.: Improving Seismic Hazard Assessment in Southeast Spain through CyberShake: A Physics-Based Approach, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4424, https://doi.org/10.5194/egusphere-egu24-4424, 2024.

The Central Italy area close to the town of Amatrice was hit by a seismic sequence that started with a Mw 6.2 mainshock and lasted more than 1 year, with the highest event being the Mw 6.5 earthquake in Norcia. Reliable prediction of ground motion is difficult due to the limited data available particularly in the near-source; for this reason, we need realistic simulations of near-source broadband ground motion for seismic hazard assessment. Such simulations should be accurate and computationally efficient. In this work, we performed physics-based simulations to investigate ground motion variability for the Amatrice and Norcia earthquakes. Using the Frequency-Wavenumber (FK) technique we generated broadband ground motion time histories up to 10 Hz for both earthquakes. We exploited accurate source rupture models and various sets of Green’s functions generated with 1D velocity models obtained by slightly modifying the 1D velocity model of the Central Apennine Area proposed by Hermann et al. (2011). First, we employed the Graves and Pitarka (2016) technique to generate kinematic rupture models. Then, FK Green's functions are computed using the propagator matrix method proposed by Zhu and Rivera (2002). Using the RotD50 SA goodness of fit (GoF) between the recorded and simulated ground motion, we conducted 1D velocity model sensitivity analysis. Overall, the simulated time histories match well the recorded ground motion. We found that the 1D velocity model of the Central Apennine Area, modified for the inclusion of thin near-surface sedimentary layers, performed better than the other 1D velocity models considered in the GOF analysis. Our ground motion simulations suggest that the FK-based simulation approach can effectively reproduce the recorded ground motion in the frequency range of 0-10 Hz. Consequently, this approach holds promise for the seismic hazard assessment in Central Italy, enabling significant computer time savings compared to more complex methodologies that involve 3D wave propagation modeling.

How to cite: Artale Harris, P., Pitarka, A., and Akinci, A.: Broadband Ground Motions Simulations for M≥6.0 Earthquakes in the 2016/2017 Central Italy Seismic Sequence through a 1D Frequency-Wavenumber Approach: a Velocity Models Sensitivity Analysis , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5129, https://doi.org/10.5194/egusphere-egu24-5129, 2024.

In this study we investigate the behavior of seismic waves in a high-scattering medium using numerical simulations of the full wavefied based on the Spectral Element Method solutions of the wave equation. The simulated 3D elastic medium was designed to have Laplacian correlated heterogeneity, creating a realistic representation of the complexities present in natural seismic environments. We conducted analyses on three distinct cases, each characterized by increasing levels of heterogeneity fluctuation, ranging from 10% to 25% standard deviation.
We checked the consistency between theoretical results and simulations with regard to the value of the mean free path, the asymptotic behavior at long times and the partitioning of energy between compressional and shear modes. Excellent agreements were obtained, indicating the reliability of the numerical models of coda waves used here.  Our analyses are made both at depth and at the free surface, allowing us to compare the behavior of seismic waves under varying conditions. Additionally, we validated our findings by conducting independent numerical simulations of wave energy densities that used Monte Carlo methods to solve the Radiative Transfer Equation, thus corroborating the robustness and accuracy of our results for long lapse times.
We show that under specific conditions, existing simulation codes can effectively replicate wave propagation in a highly scattered medium. This implies that a greater part of the waveform, namely the late envelops, could be employed in inversion processes, thus opening up new possibilities in the realm of inversion studies. Furthermore, we used these simulations to investigate the behavior of the wavefield and its gradient, exploring the information that can be extracted from their evolution over time to improve characterization of environmental heterogeneity.

How to cite: Bracale, M., Margerin, L., Brossier, R., and Campillo, M.: Scattered wave and coda reliability in 3D elastic seismic simulation: new insights for the advancement of inversion studies., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5475, https://doi.org/10.5194/egusphere-egu24-5475, 2024.

Through the progression of seismic activity, natural fault zones undergo a complex evolution characterized by the accumulation of damage and the formation of gouge within the fault core across multiple scales. Even though this is believed to be among the key factors affecting the evolution of fault seismicity over time, a deep understanding of the mechanisms at play is still missing. In this study, we explore the role of gouge production in the self-organization process of loaded rough faults, focusing on the evolving dynamics of earthquake nucleation, recurrence and moment partitioning during the seismic cycle. We model the stress and sliding dependence of gouge evolution by linearly coupling Archard's wear law with rate-and-state friction through the critical slip distance ( Dc ). Including this new formulation in 2D quasi-dynamic, elastic simulations of rough faults, we can reproduce the effects of spatially and temporally heterogeneous gouge evolution. Following the build-up of gouge over many cycles, we observe a progressive transition from cascade-driven to creep-dominated nucleation processes, marked by an increase of precursory slow slip and foreshock activity. A clear shift in the moment partitioning from faster to slower slip rates becomes evident as heterogeneity grows larger, followed by a reduction of the total cumulative moment released. Finally, the recurrence interval is observed to grow initially, then drop abruptly and become more unpredictable as the amplitude of Dc continues to rise. Incorporating a new formulation of gouge production in earthquake cycles simulations, this work sheds light on the role of gouge accumulation in the maturation process of natural faults, offering critical insights for seismic risk assessment and mitigation.

How to cite: Castellano, M., Milanese, E., Cattania, C., and Kammer, D.: The influence of gouge formation on seismicity and fault slip behavior. , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6202, https://doi.org/10.5194/egusphere-egu24-6202, 2024.

In a three-dimensional Cartesian coordinate system, the deformation of the medium around a particle includes strain, translation, and rotation. Rotational motion is an important aspect of current seismological research. Seismologists have recognized the importance of rotational motion in dynamic response and damage of structures caused by certain earthquakes, through investigations into earthquake damage. In rapid earthquake intensity reports, it is essential to not only consider factors such as earthquake location, source depth, magnitude, and fault rupture model, but also to emphasize the analysis of the amplification effect of shallow media. We discuss the attenuation characteristics and difference between seismic translational and rotational components by medium viscoelasticity through two-dimensional numerical simulation, analyze the amplification effect of shallow viscoelastic low-velocity layer on ground motion by the reference site spectral ratio (RSSR), and discuss the difference of the amplification caused by different low-velocity layer factors. The results show that the seismic primary frequency decreases more with increasing viscoelasticity, and the energy of rotational component attenuates more significantly than that of translational component. The elastic low-velocity layer amplifies high-frequency signals of body waves greater than the viscoelastic low-velocity layer, especially in rotational component. When shallow low-velocity layers consist of multilayered sediments compared to a single sediment, the amplification of surface wave is stronger, particularly in rotation. We follow the research method used for seismic translation to discuss the amplification effect of shallow viscoelastic medium on seismic rotation, which is important for performance-based seismic design and earthquake damage analysis.

How to cite: Li, W., Zhang, Y., and Wang, Y.: Attenuation and amplification effects of seismic translational and rotational components in shallow media, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7150, https://doi.org/10.5194/egusphere-egu24-7150, 2024.

We focus on the main rupture process of the Mw7.8 Februrary 6^th 2023 01:17 UTC Pazarcık, Turkey, propagating to the south-west direction where more than ten acceleration stations recorded the ground motions within a distance of a few kilometers from the fault. On one hand, we estimate the frictional parameters directly from the waveforms of the acceleration records. Several stations are sufficiently close enough to characterize the cohesive zone length, and the estimated critical displacement (Dc) ranges from 90 cm to 150 cm. On the other hand, we carry out the dynamic rupture simulations along the constructed non-planar fault and also simulate the ground motions in the surrounding, using boundary integral equation and finite difference methods.  Upon the constructed standard model, we prepare different models of Dc distribution both along dip and strike. Our numerical simulations show that a longer Dc is necessary in the shallowest depth (2-3 km depth) than in the deep seismogenic zone. The observed ground motion pattern in terms of PGV (Peak Ground Velocity) shows a strong correlation with the estimated strike-variated Dc and the rupture process controlled by the fault geometry.

How to cite: Aochi, H. and Cruz-Atienza, V.: Characterization of shallow fault parameters from the near-field ground motion data and non-planar dynamic rupture simulations for the Mw7.8 February 6th Pazarcık, Turkey, earthquake, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7928, https://doi.org/10.5194/egusphere-egu24-7928, 2024.

As part of the Southern Alps Long Skinny Array (SALSA) project, ~35+ seismometers have been deployed with 10–12 km spacing along a 450 km-long   section of the Alpine Fault. SALSA is focused on determining the ground motions likely to be produced by a future Alpine Fault earthquake. This project is addressing three principal objectives: (1) Determine the Alpine Fault’s subsurface geometry, present-day slip rates, and spatial variations in how tectonic stresses are currently accumulating on the fault, (2) Estimate the ground shaking that would be recorded at seismometers throughout central and southern New Zealand by localised slip at different points on the Alpine Fault, focusing on the synthesis of long-period Green's functions  representing accurate path effects between sources distributed along the fault and population centres throughout the South Island, and (3) Calculate the ground shaking hazard from geologically informed earthquake rupture scenarios. In this presentation we will address the influence of inferred Alpine Fault source models derived from empirical data as well as current knowledge of the fault geological and geophysical parameters on regional seismic hazard.

How to cite: Holden, C., Townend, J., Chamberlain, C., Warren-Smith, E., Juarez-Garfias, C., Pita-Sllim, O., Van Wijk, K., and Denolle, M.: Inferred source models for Alpine Fault Earthquake Scenarios and influence on seismic hazard., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8160, https://doi.org/10.5194/egusphere-egu24-8160, 2024.

Earthquakes that nucleate at depths shallower than a few km are very rare but pose high near-fault hazard despite moderate magnitudes. Some very shallow earthquakes have been associated with surface mass removal processes, both natural (e.g. glacier melting) and anthropogenic. A notable recent case is the November 11 2019 Mw 4.9 Le Teil, France earthquake. It called strong public attention because of its very shallow depth (slip shallower than 2 km), very strong ground motion (>1 g) affecting the near-fault population, and proximity to nuclear power plants. It has been proposed that this earthquake could have been triggered by mass removal from a large cement quarry located close to the epicenter. Indeed, the estimated Coulomb stress change induced by the quarry activity on the fault is of several 100 kPa. Here, we further evaluate the mechanical viability of the quarry-triggering hypothesis through 3D earthquake cycle simulations.

We consider a dipping fault governed by rate-and-state friction, with velocity-weakening steady-state behavior, and a realistic mass removal history constrained by analyses of aerial optical images across ~180 years of quarry activity. To account for uncertainties about the recurrence time of natural earthquakes on the fault and the timing of the previous natural event, we consider mass-removal loads starting at different times relative to the simulated natural earthquake cycle. Our simulations show that realistic mass removal rates can advance the failure time by thousands of years. Simulations with a constant mass-removal rate but same cumulative removed mass at 180 years produce a similar triggering timing. This indicates that the induced clock advance mostly depends on the cumulative load, rather than on its rate. The dependence on loading rate manifests through the following mechanism: the model with constant rate can trigger slow slip events instead of regular earthquakes, which postpones the next regular earthquake by a long time, whereas the model with realistic loading history (and higher load rates) always triggers regular earthquakes. The quarry's proximity to the fault and the frictional heterogeneity on the fault also play important roles. For example, clock advance is higher if the quarry location is closer to the edge of the velocity-weakening zone and lower in the middle. Also, the model with the classical rate-and-state model shows negligible impact if the quarry is at the top of a steady-state behavior zone and far away from the velocity weakening zone.

While these models confirm the possibility that mass removal can trigger shallow earthquakes on velocity-weakening faults, we will also report on additional simulations that examine whether such triggering can occur on a fault with velocity-strengthening behavior at shallow depth or it requires a more sophisticated fault rheology, such as friction with a transition from velocity-strengthening to velocity-weakening at increasing slip rate. These modeling efforts will be further constrained by ongoing laboratory experiments on representative materials of the fault that ruptured in the Le Teil earthquake.

How to cite: Sopaci, E. and Ampuero, J. P.: Triggering of very shallow earthquakes by surface mass removal processes - case study of the 2019 Mw4.9 Le Teil, France earthquake , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9827, https://doi.org/10.5194/egusphere-egu24-9827, 2024.

The Mw 7.8 and Mw 7.5 doublets of the 2023 Turkey seismic sequence show strong velocity pulses that may have caused extensive damage to buildings and structures. We analyze the velocity pulses from the empirical data (both raw and processed) to understand the causes of these for these doublets. The analysis includes a comparison with the velocity pulses from the synthetic data of the Jia et al. (2023) dynamic rupture simulation available for the first Mw 7.8 event and an analysis of the variability of large instrumented earthquakes of the last 30 years. We characterize the properties of the ground motion pulses (e.g., period, velocity, and orientation) using the algorithm proposed by Shahi and Baker (2014). The identified pulses in the synthetic data show the main characteristics of the pulses (periods, PGV). However, the pulse properties in the synthetic data show less variability than the natural variability found in the empirical data, particularly a random behavior in the pulse orientation. The results then indicate that the pulse characteristics in the near-fault regions of large-magnitude earthquakes exhibit a significant variability and that this variability is similar to the one observed in past large earthquakes. This pronounced variability can be attributed to various factors, including directivity effects and site effects. This suggests that the full complexity of earthquake rupture processes and site configurations should be taken into account to be able to capture the high variability in pulse properties.

How to cite: Yen, M.-H., Türker, E., and Cotton, F.: An analysis of strong velocity pulses from the empirical data and dynamic rupture simulations of the 2023 Kahramanmaras earthquake doublets, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9916, https://doi.org/10.5194/egusphere-egu24-9916, 2024.

Deep-focus earthquakes occur at depths of 300-700 km below the surface where brittle failure is unlikely due to the large lithostatic pressure. Instead, they require a ductile localization mechanism that can significantly reduce rock strength and create highly localized shear zones. The feedback loop of shear heating, temperature-dependent viscosity and localization is called thermal runaway and has been linked to deep-focus earthquakes.

We present one- and two-dimensional (1D and 2D) numerical, thermomechanical models that investigate the occurrence, nucleation and temporal evolution of thermal runaway in a simple shear setting. The models are characterized by a visco-elastic rheology where viscous creep is accommodated with a composite rheology of diffusion and dislocation creep as well as low-temperature plasticity. We utilize the pseudo-transient iterative method in combination with a viscosity regularization and adaptive time stepping to solve this nonlinear system of equations and avoid resolution dependencies.

Varying eight input parameters, we observe two distinct types of behavior. After elastic loading, models either release stress over hundreds to thousands of years, accompanied by low slip velocities and moderate temperature increase, or they release stress within seconds to minutes while slip velocity and temperature increase drastically – Thermal runaway occurs. With nondimensional scaling analysis, we unite the eight different input parameters into two nondimensional numbers that allow inferring the behavior. The ratio t_r/t_d describes the competition between heat generation by viscous dissipation and heat loss due to thermal diffusion whereas the ratio U_el/U_th compares the elastic and thermal energy density before stress relaxation.

2D experiments show that thermal runaway allows highly localized ductile ruptures to nucleate at small heterogeneities and propagate like brittle fractures. The ruptures accelerate during propagation and reach the highest velocities when two tips link up. Rupture trajectories are usually parallel to the direction of background deformation but bend in the vicinity of other ruptures to allow for a link up. The results demonstrate that thermal runaway can create highly localized, propagating shear zones that reach slip velocities in line with slow earthquakes at upper mantle and transition zone conditions.

How to cite: Spang, A., Thielmann, M., and Kiss, D.: Thermal runaway as driving mechanism of deep earthquakes – Constraints from numerical modeling, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10403, https://doi.org/10.5194/egusphere-egu24-10403, 2024.

We present a quantum computing concept for 1-D elastic wave propagation in heterogeneous media with two components: a theoretical formulation and an implementation on a real quantum computer. The method rests on a finite-difference approximation, followed by a transformation of the discrete elastic wave equation to the Schrödinger equation, which can be simulated directly on a gate-based quantum computer. An implementation on an error-free quantum simulator verifies our approach and forms the basis of numerical experiments with small problems on an actual quantum computer. As the presented approach promises exponential speedup compared to classical numerical wave propagation methods, it has the potential to significantly push the limits of global full-waveform inversion, particularly maximum feasible frequencies, on future quantum computers.

How to cite: Schade, M., Bösch, C., Hapla, V., and Fichtner, A.: A quantum computing concept for 1-D elastic wave simulation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10632, https://doi.org/10.5194/egusphere-egu24-10632, 2024.

Natural faults exhibit complex geometry. In this study, we model cycles of earthquake ruptures on non-planar faults governed by a friction formulation that combines rate and state friction for low slip velocity and enhanced weakening friction in the form of flash heating for high slip velocity, both consistent with rock friction experiments. The numerical method allows non-matching meshes across the fault, continuously updates the fault geometry, and employs variable time steps with quasi-static and fully dynamic time integration schemes during slow and fast deformation stages, respectively. To prevent the development of large stresses on the fault, the model also accounts for fault wear and inelastic off-fault deformation. We investigate the effect of macro-scale roughness on the fault slip behavior and rupture dynamics in terms of event magnitude, stress drop, and rupture style and speed. We analyze the relationship between the fault geometry, stresses from the preceding earthquakes, and rupture characteristics.

The simulation results show a significant increase in event variability with roughness levels, with both small partial ruptures and ruptures significantly more intense than those on planar faults. The planar faults host a sequence of earthquakes that rupture the entire fault, exhibiting similar magnitudes and stress drops. The substantial reduction in friction enables the ruptures to propagate under a low background shear-to-normal stress ratio as self-healing slip pulses, with a sub-Rayleigh rupture speed. Faults with low roughness levels generally show a similar pattern. Prior to some events, the stress ratio along the fault slightly increases, leading to ruptures with secondary slip pulses and larger magnitudes. As roughness increases, stresses become more heterogeneous, resulting in a more complex sequence of ruptures, some of which arrest at restraining bends with a low stress ratio. However, stress accumulation and slip deficit during these partial ruptures result in high stress ratios on the unruptured fault segments. These are eventually released by large events of crack-like ruptures with supershear propagation speed and stress drops and slip significantly larger than a typical event on a planar fault. Therefore, while fault roughness can cause rupture arrest, consistent with previous studies, it can also substantially increase earthquake magnitudes. This factor should be accounted for in earthquake hazard assessments.

How to cite: Tal, Y.: Rupture Dynamics and Characteristics During Earthquake Cycles on Nonplanar Faults with Strongly Rate-Weakening Friction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11379, https://doi.org/10.5194/egusphere-egu24-11379, 2024.

The Statewide Southern California Earthquake Center (SCEC) has designed and implemented CyberShake (CS), a high-performance computing (HPC) workflow to undertake Physics-Based Probabilistic Seismic Hazard Analysis (PB-PSHA) in California (CA). Here, we have ported CS from CA to the South Iceland Seismic Zone (SISZ) and the Reykjanes Peninsula Oblique Rift (RPOR), which experience sinistral transform motion and pose a very high earthquake risk to about 2/3 of the Iceland population. We consider a realization of the 3D SISZ-RPOR fault system, where fault areas are estimated from event magnitude through a scaling law (Mai & Beroza, 2017),  that fits maximum fault extents observed from slip inversions and surface mappings. The magnitude variability across the modeling region (~63.8°- 64.1°N, ~20°-23°W) is Mw 5-7. In this work, we employ CS to model ~2100 kinematic earthquake ruptures and quantify the resulting ground motion (GM) in terms of Pseudo Spectral Acceleration (PSA) intensity measures. An important computational milestone is the software development of an open-source in-house workflow manager at the Marenostrum Supercomputer that replaces the one used in CA by SCEC based on Pegasus and HTCondor. This new workflow manager handles input data (fault-plane geometries, rupture magnitudes, surface stations for GM recording and hazard studies), orchestrates the execution of CS components, and stores results (particle velocity seismograms and hazard curves). Among these components, the Graves-Pitarka (GP) kinematic rupture generator is used to produce finite-fault source descriptions characterized by a few large asperities. The other important component is the open-source fourth-order finite-difference staggered-grid AWP-ODC earthquake simulation code that allows for reciprocity and efficiently simulates rupture and seismic wave propagation in 3D heterogeneous Earth models. CS uses an adjoint computational procedure in which simulations of wave propagation are performed using a polarized delta source to compute the Strain Green Tensors (SGTs) at each fault point. The convolution of SGTs with GP ruptures yields particle-velocity seismograms at each station. SGT time histories are memory demanding, but the adjoint calculations are completely independent and therefore embarrassing parallel, making CS a highly efficient earthquake simulation tool. In this study, SGTs are constructed using a source frequency range of 0-1.0 Hz, generating ground motion synthetics resolved up to 0.5 Hz. CS rotation-invariant PSA values (3 and 5 sec periods) computed from our study show a good agreement with updated Bayesian ground motion prediction equations (Kowsari et al, 2022). This study is a first step towards a PB-PSHA in the SISZ-RPOR region and to routinely apply Cybershake outside of California.

REFERENCES:

Mai, M., & Beroza, G. Source scaling properties from finite-fault-rupture models. Bulletin of the Seismological Society of America, 90(3), 604-615, 2000.

Kowsari, M., Sonnemann, T., Halldorsson, B., Hrafnkelsson, B., Snæbjörnsson, J. &  Jonsson, S. Bayesian inference of empirical ground motion models to pseudo-spectral accelerations of South Iceland Seismic Zone earthquakes based on informative priors. Soil Dynamics and Earthquake Engineering, 132, 106075, 2020.

How to cite: Rojas, O., Bayat, F., Monterrubio-Velasco, M., Abril, C., Callaghan, S., Rodríguez, J. E., Kowsari, M., Halldórsson, B., Olsen, K., Gabriel, A.-A., and de la Puente, J.: CyberShake simulation of strike-slip earthquakes on the Southwest Iceland transform zone, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13834, https://doi.org/10.5194/egusphere-egu24-13834, 2024.

Friction plays a crucial role in rupture dynamics and yet its precise nature remains elusive. Consequently, a friction law must be assumed to model rupture. Commonly used constitutive laws for modeling friction include slip-weakening laws which are characterized by a drop from static to dynamic frictional strength. Within this framework, the prevailing understanding asserts that the frictional behaviour is solely controlled by the fracture energy - the area beneath the frictional strength versus the cumulated slip curve. In particular, it is claimed that the curve's shape itself has no influence on the system's response. Here we perform fully dynamic rupture simulations to challenge prevailing beliefs by demonstrating that the constitutive law shape exerts an intimate control over rupture profiles. For a consistent fracture energy but varying constitutive law shapes, the velocity profile is different: each abrupt slope transition leads to the localization of a distinct velocity peak. For example, in the case of a dual slip-weakening law featuring two different slopes, the rupture exhibits two distinct velocity peaks. This distinction significantly influences how a rupture responds to a stress barrier. These results are derived through two separate numerical schemes (spectral boundary integral and finite element methods) ensuring their independence from the computational approach employed.

How to cite: Ferry, R. and Molinari, J.-F.: Unveiling the influence of slip-weakening laws' shapes on rupture dynamics: beyond fracture energy in controlling rupture profiles, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15557, https://doi.org/10.5194/egusphere-egu24-15557, 2024.

Earthquakes are associated with the propagation of a dynamic rupture, which radiates elastic energy through seismic waves. The generation of seismic radiation is related to dynamic weakening of shear stress and stress drop. In modeling dynamic ruptures, shear stress evolution is commonly imposed through a constitutive law, such as the widely adopted slip weakening laws. According to these constitutive laws, shear stress evolves as a function of slip in each point of the rupturing fault, prescribing strength excess, stress drop and dynamic weakening.

Here, we compare two well-known slip weakening laws: namely, the classic Ida’s (1972) and the Ohnaka’s (1996) slip weakening laws. The former prescribes that fault stress increases from the initial stress to the peak stress with zero slip and then linearly decreases from the peak value to a residual value over a slip-distance Dc (dynamic weakening). The latter assumes that the initial stress hardening phase occurs over a non-negligible slip-distance Da and that shear stress decrease from the peak value is not linear. The Ohnaka’s law was validated with numerous laboratory experiments. The evolution of shear stress with slip allows the estimate of the breakdown work Wb, i.e. the excess of work above a minimum stress level with slip from 0 to Dc.

We collected data from high-velocity friction experiments to quantify yield, peak and residual stresses, Da and Dc distances for bare-rock samples of Carrara Marble and Gabbro deformed under various experimental conditions (room humidity, vacuum, pressurized fluids) and normal stress (from 5 to 40 MPa). The ratio Da/Dc is much lower for Carrara marble (0.015) than for Gabbro (0.12). We implemented the Ohnaka’s constitutive law in a 2D finite difference code for spontaneous dynamic ruptures characterized by a fault in a homogeneous elastic material.  We perform simulations using the two different slip weakening laws. We kept constant Dc, and we compared the results of the simulations in terms of rupture style, rupture velocity, breakdown work, and cohesive zone size. As expected both laws yield crack-like ruptures. Moreover, Ohnaka’s law in comparison to the linear slip weakening law produces:

• rupture velocity ~2 % higher;
• breakdown work (Wb) up to 60 % lower. Moreover, dividing the breakdown work into the energy dissipated between the yield stress and the peak stress over the slip-distance Da (Wb_a), we notice that Wb_a can reach up to the 30% of the total Wb in case of Gabbro (Da/Dc = 0.12).
• a cohesive zone size (defined as the portion of the fault in which the slip velocity is higher than zero and the stress is higher than its residual value) up to 50% larger.

Therefore, Ohnaka’s law generates more energetic ruptures (i.e. faster rupture velocity and peak slip-rate) despite having a larger cohesive zone due to the lower breakdown energy dissipated during rupture propagation. We discuss our results in terms of the difference between breakdown highlighting the implications on dynamic rupture propagation and earthquake energy budget. We emphasize that common interpretations of energy dissipated during rupture propagation are model-dependent.

How to cite: Cornelio, C., Murphy, S., Spagnuolo, E., Nielsen, S., and Cocco, M.: Differences in breakdown work and fracture energy in slip weakening constitutive laws, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16751, https://doi.org/10.5194/egusphere-egu24-16751, 2024.

How to cite: Chao, S.-H., Huang, J.-Y., Hsu, C.-C., Lin, C.-M., Sung, C.-H., and Kuo, C.-H.: Taiwan Non-Ergodic Ground Motion Prediction Equations for Spectral Accelerations and Instantaneous Powers, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18840, https://doi.org/10.5194/egusphere-egu24-18840, 2024.

Tectonic deformation plays a crucial role in shaping the Earth's surface, with strain localization resulting in the formation of shear zones in depth and faults on the surface. These structures accommodate a significant portion of the displacement between tectonic plates. While long-term deformation can be approximated as continuous visco-plastic processes, earthquakes involve cycles of stress loading and unloading that trigger rapid and catastrophic elasto-plastic deformation. Earthquake dynamic rupture models offer valuable insights into studying and comprehending earthquakes. However, these models heavily rely on initial conditions that are often challenging to obtain solely from observations. Particularly, a mechanically self-consistent prestress state loading a fault prior a seismic event and 3D fault geometry, especially in depth, are commonly poorly constrained. Nonetheless, the prestress state and the fault geometry significantly impact earthquakes initiate, propagate, and arrest and the associate radiation of seismic waves and ground shaking.

To address the lack of information on stress and fault geometry, one promising approach is to use long-term geodynamic numerical simulations. In this study, we employ pTatin3D, a visco-plastic finite element software, to simulate the evolution of strike-slip deformation in 3D over geological time scales. To ensure a physically consistent long-term model, the fault geometry is not prescribed but solved for based on the lithospheric rheology and tectonic plate velocities. However, the geodynamics model describes faults as continuous volumetric fields of finite deformation and strain-rate, rendering them 3D objects, while earthquake dynamic rupture models typically represent faults as 2D interfaces.

In this study, we outline a new and versatile method to link 3D geodynamic simulations to rupture dynamics earthquake and seismic wave propagation modelling. We first extract 3D volumetric shear zones from the geodynamic model and automatically convert them into surface representations. Next, we generate meshes including these as faults for dynamic rupture models. Finally, we showcase 3D dynamic rupture models utilizing the stress states and faults self-consistently as derived from the long-term geodynamic model as initial conditions.

How to cite: Jourdon, A., Hayek, N., May, D., and Gabriel, A.-A.: One-way linking of 3D long-term geodynamic models and short-term earthquake dynamic rupture models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19294, https://doi.org/10.5194/egusphere-egu24-19294, 2024.

Campi Flegrei is a volcanic region in Southern Italy of great interest for volcanic risk due to the presence of a potentially dangerous caldera collapse structure in a very densely populated area. In historical times, the Campi Flegrei area experienced explosive eruptions (the most recent – the Monte Nuovo eruption, 1538 CE), and in recent times (from 1969 to 1972 and from 1982 to 1984) critical seismic activity and bradyseism crises. Since 2020 the increase of seismicity related with the acceleration of ground uplift is a matter of the scientific and civil protection debate, given the vulnerability of the urban settlements under the effect of the volcanic phenomena. The recent bradyseism crisis climaxed in September 2023, with a high number of seismic events per month (1000 events per month) and a maximum magnitude (Md) of 4.2— the strongest event recorded in the last forty years. In general, predicting attenuation law in volcanic areas poses a significant challenge due to the limited availability of strong motion records, the predominance of lower magnitude events, and the distinct characteristics of waveforms compared to tectonic earthquakes. Moreover, additional challenges arise from the potential anisotropic behavior of the area, which could lead to high seismic impact for specific directions of seismic wave propagation. This makes it difficult to establish predictive models for ground motion, hindering the development of reliable risk scenarios and the effective implementation of civil protection measures. Since September 2023, the Civil Protection Department started improving the station coverage of the accelerometric network (RAN, Rete Accelerometrica Nazionale) by installing 3 new seismic stations along coastal areas and around Pisciarelli locality. The accelerometric data, recorded from the 18th of September 2018 to the 4th of October 2023 by 12 accelerometric stations of the RAN, fill a gap of information and represent an important contribution in adding new constraints to ground motion characterization. Specifically, we analyzed 3771 three-component records whose 186 exhibit a magnitude exceeding 3.5. We derived the engineering interest parameters (e.g., Peak Ground Acceleration, PGA; Peak Ground Velocity, PGV; Housner Intensities, HI; Arias Intensities, AI; significant duration, Td; Spectral accelerations) and compared them with the available ground motion prediction equations defined in the tectonic and volcanic areas in Italy and abroad. For the two events >= 3.8 we perform a comprehensive analysis. Our results unveil a trend similar to that predicted in the ground motion prediction equations in the near field but with a steeper attenuation recorded beyond approximately 5 km of distance. Furthermore, a relevant result is the existence of elevated peaks in PGA (Peak Ground Acceleration) at considerable distances also for low magnitude values underscoring the potential existence of preferential directions in propagation. These findings are crucial for understanding the region's seismic impact and enhancing risk assessment and civil protection strategies in this densely populated volcanic area.

How to cite: Pandolfi, C., Talone, D., Lavecchia, G., Costa, G., Pazzi, V., Fornasari, S. F., Filippi, L., Zambonelli, E., Ammirati, A., Sirignano, S., Akinci, A., and de Nardis, R.: A first look at the ground motion characteristics unveiled by accelerometric data: the case of Campi Flegrei area (Italy), EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19464, https://doi.org/10.5194/egusphere-egu24-19464, 2024.

Discrimination of underground explosions from naturally occurring earthquakes and other anthropogenic sources is one of the fundamental challenges of nuclear explosion monitoring. In an operational setting, the number of events that can be thoroughly investigated by analysts is limited by available resources. The capability to rapidly screen out events that can be robustly identified as not being explosions is, therefore, of great potential benefit. Nevertheless, possible mis-classification of explosions as earthquakes currently limits the use of screening methods for verification of test-ban treaties. Moment tensors provide a physics-based classification tool for the characterisation of different seismic sources and have enabled the advent of new techniques for discriminating between earthquakes and explosions. Following normalisation and projection of their six-degree vectors onto the hypersphere, existing screening approaches use spherically symmetric metrics to determine whether any new moment tensor may have been an explosion. Here, we show that populations of moment tensors for both earthquakes and explosions are anisotropically distributed on the hypersphere. Distributions possessing elliptical symmetry, such as the scaled von Mises-Fisher distribution, therefore provide a better description of these populations than the existing spherically symmetric models. We describe a method that uses these elliptical distributions in combination with a Bayesian classifier to achieve successful classification rates of 99% for explosions and 98% for earthquakes using existing catalogues of events from the western United States. Application of the method to the 2006–2017 nuclear tests in the Democratic People's Republic of Korea yields 100% identification rates. The approach provides a means to rapidly assess the likelihood of an event being an explosion and can be built into monitoring workflows that rely on simultaneously assessing multiple different discrimination metrics.

How to cite: Hoggard, M., Scealy, J., and Delbridge, B.: Improved classification of explosive moment tensors using elliptical distribution functions on the hypersphere, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2978, https://doi.org/10.5194/egusphere-egu24-2978, 2024.

Seismic ground motions in the near-fault region produce strong pulses in the velocity-time history, resulting in severe damage to structures. To accurately and effectively monitor these ground motion signals with strong pulses, Shock-Waveform (SW) method is introduced to quantitatively extract the largest velocity pulse from a given ground motion. SW method is an energy-based and adaptive signal analysis method, which has proven capability of analyzing different physical and engineering signals initiated by sudden actions. It is suitable to identify pulse components in the signal with low error and high efficiency. Three variables are proposed to classify ground motions, which is combined with the Principal Component Analysis (PCA) for data dimensionality reduction and subsequent analysis. In addition, an optimum classification standard on pulse-like and non-pulse-like ground motion is established. To avoid the subjective judgement induced by manual selection, unsupervised machine learning classification method and Support Vector Machine (SVM) are used successively to find the decision boundary. In this study, about 100 pulse-like ground motions with large-velocity pulses are identified from approximately 1000 near-fault ground motion recorded in PEER Next Generation Attenuation-West2 database. It shows that most of the pulse-like ground motions are caused by the directivity effect. Based on the proposed classification approach, new models are developed to forecast the possibility of a single pulse, multi-pulses, and pulse period for a given earthquake event. 

How to cite: Wang, J. and Li, Q.: Identification of strong-velocity pulses in seismic ground motion signals, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3453, https://doi.org/10.5194/egusphere-egu24-3453, 2024.

Distributed acoustic sensing (DAS) is proving to be an effective technology for seismological applications. Its success is due to the ability to deploy DAS instrumentation on the existing ever-growing telecommunication fiber networks across the globe. However, the benefits of DAS are hindered by the sheer volume of data commonly recorded from single-instrument deployments, which can easily reach tens of TBs. Additionally, since DAS measures along fiber strain, new data analysis paradigms are necessary to exhaustively exploit all the information contained within these large datasets. 

We showcase successful applications of DAS experiments using existing fiber cables located in different scenarios, from volcanic systems to densely populated urban environments. To harness the information within these novel datasets, we combine machine-learning tools with efficient algorithms running on high-performance computing architectures. For example, we showcase how the arrival times obtained from PhaseNet-DAS can provide real-time earthquake detection and localization, allowing for the inclusion of DAS data within earthquake early warning systems. Moreover, we demonstrate the capability of integrating real-time streamed DAS channels within seismic network operations. Our processing paradigm is proving to be an effective ground for discoveries and for creating the next generation of seismic monitoring frameworks.

How to cite: Biondi, E., Li, J., Saunders, J., Husker, A., and Zhan, Z.: The frontiers of distributed acoustic sensing for seismological applications, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4186, https://doi.org/10.5194/egusphere-egu24-4186, 2024.

The current limited knowledge about Earth system prevents deterministic earthquake prediction. This will probably continue for the foreseeable future. However, the improved capability of identifying reliable precursory phenomena can allow geoscientists to comprehend if the monitored system is evolving toward an unstable state. Among the premonitory phenomena preceding earthquakes, foreshocks represent the most promising candidate. Physically, two hand-member mechanisms have been proposed to interpret foreshocks. The first considers the failing of populations of small patches of fault that eventually but not necessarily become large earthquakes whereas the second assumes that foreshocks are a part of the nucleation process which ultimately leads to the mainshock. The prompt identification of foreshocks with respect to background seismicity is an issue and the task is worsened when dealing with low-magnitude earthquakes. However, the use of Artificial Intelligence (AI) can help real-time seismology to effectively discriminate precursory signals.

In the present study, we propose a deep learning method named PreD-Net (Precursor Detection Network) to address the precursory signal identification of induced earthquakes through the analysis of several statistical features. PreD-Net has been trained on data related to two induced seismicity areas, namely The Geysers, located in California, USA, and Hengill in Iceland. Notably, the network shows a suitable model generalization, providing considerable results on samples that were excluded from the training dataset of all the sites. The performed tests on related samples of induced relatively large events demonstrate the possibility of setting up a real-time warning strategy to be used to avoid adverse consequences during field operations.

This work is supported by project D.I.R.E.C.T.I.O.N.S. - Deep learning aIded foReshock deteCTIOn Of iNduced mainShocks, project code: P20229KB4F - - Next Generation EU (PRIN-PNRR 2022, CUP D53D23022800001)

How to cite: Convertito, V., Giampaolo, F., Amoroso, O., and Piccialli, F.: Deep learning forecasting of induced earthquakes through the analysis of precursory signals, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4268, https://doi.org/10.5194/egusphere-egu24-4268, 2024.

GEObit Instruments are proud to announce the market release of a new low-cost and low -latency seismic accelerograph, the GEO-T200, for earthquake monitoring, early waring applications, and structural monitoring. The device mainly consists of two sections, the triaxial sensor sensor and the digitizer. The architecture and the hardware is based on the GEObit GEOtiny platform.

The sensing elements are based in a re-designed previous generation GEObit force balance acceleration sensor unit [1], providing very high dynamic range 160+dB, and wide bandwidth, flat response DC to 260Hz. The acceleration range is user configurable and can be set between +/-4g to +/-0.5g but other ranges are also available upon request.

The digitizer is based on a 24bit ADC and provides high effective dynamic range 140dB, high sampling rate up to 4000sps, integrates seedlink server and the earthworm chain. The device is based on a locally running open-source components ported on ARM Linux board. It is able to apply local signal processing and trigger detection based on multiple schemes (amplitude, STA/LTA etc.) through open-source components ported from the Earthworm toolchain and transmit pick times over MQTT with ultra-low latency based signaling for trigger event distribution supporting multiple centralized or distributed schemes.

It Supports ethernet port and Wi-Fi. Also supports continuous data stream, triggered data stream (level, LTA/LTA, both) or both.

The device is housed into a small cylindrical enclosure, aluminum made, IP68 with dimensions 120mm diameter and 143mm height. Three leveling legs are provided along with a central bolt for proper mounting of the device. An bright OLED lcd screen reports the user about the instrument operation and state of health. The SOH stream is also transmitted in real time over TCP.

References:

[1]: Design, Modeling, and Evaluation of a Class-A Triaxial Force-Balance Accelerometer of Linear Based Geometry” N. Germenis, G. Dimitrakakis, E. Sokos, and P. Nikolakopoulos Seismol. Res. Lett. 93, 2138–2146, doi: 10.1785/0220210102

How to cite: Germenis, N.: A new low latency and low-cost force-balance accelerograph for earthquake and structural monitoring and for early waring applications., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5299, https://doi.org/10.5194/egusphere-egu24-5299, 2024.

GeoSphere Austria (GA, formerly ZAMG), the Italian National Institute of Oceanography and Applied Geophysics (OGS) and the Slovenian Environment Agency (ARSO) are the agencies dedicated to real-time seismological monitoring of Austria, north-eastern Italy and Slovenia, in cooperation with the respective civil protection authorities. In 2014, GA (then ZAMG), OGS and ARSO founded the “Central and Eastern Europe Earthquake Research Network” (CE3RN, http://www.ce3rn.eu/) to 1) formally establish the cross-border network, 2) define the rules of conduct for the management, improvement, extension and expansion of the network, 3) improve seismological research in the region and 4) support civil protection activities. As part of CE3RN, GA, OGS and ARSO have adoptd the “Antelope” software package for collecting, archiving, analysing and sharing seismological data.
In 2022, the international AdriaArray experiment was launched, following on from the previously successful AlpArray experiment. AdriaArray is a multinational effort to map the Adriatic plate and its active margins in the central Mediterranean with a dense regional array of seismic stations to understand the causes of active tectonics and volcanic fields in the region. GA, OGS and ARSO are actively involved in the AdriaArray experiment by providing data from their seismic monitoring networks and - in the case of OGS - also by installing and managing dedicated seismic stations. As part of the AdriaArray experiment, several additional seismic stations have been set up in Austria and north-eastern Italy. It is therefore to be expected that the additional seismic stations installed will improve the earthquake localization capabilities of GA, OGS and ARSO. This certainly applies to Austria and north-eastern Italy, but also to Slovenia, as a large part of its seismicity lies on the border with Italy.
The GOAT-CASE experiment aims to quantify the improvement in earthquake localization capability across the entire area. The underlying methodology is to locate earthquakes also using the additional seismic stations and to compare the results. The workload for the detections is distributed among the three partners, while the mapping is done centrally. An attempt will be made to use artificial intelligence to detect earthquakes and compare the results with the standard routines of the agencies.
The AdriaArray experiment is planned for a duration of 3 years starting around mid-2022. In this presentation we will illustrate the results of the first year of the experiment, from 01/07/2022 to 30/06/2023.

How to cite: Pesaresi, D., Horn, N., and Pahor, J.: GA-OGS-ARSO Transfrontier CE3RN AdriaArray Seismicity Experiment (GOAT-CASE): results of the first year of data collection and analysis, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5459, https://doi.org/10.5194/egusphere-egu24-5459, 2024.

Distributed Acoustic Sensing (DAS) is an emerging data acquisition technology that utilises an optical fiber to measure dynamic strain along its axis. Composed by an optical fiber and an interrogator unit (IU), the system emits laser pulses into the fiber and detects phase shifts in the backscattered light, converting them into strain or strain rate measurements. DAS is becoming popular in many seismological applications and, in particular, for logistically challenging environments such as offshore areas, boreholes, glaciers, and volcanic settings, where deploying conventional monitoring is challenging. Spatial and temporal sampling of DAS systems is much higher than traditional seismological instruments, offering a detailed picture of the recorded seismic wave field. This high spatial and temporal sampling of DAS systems results in massive data generation, especially over extended acquisition periods. For instance, a single day's data collected with a 1 km fiber, featuring inter-channel distances of approximately 1m and a temporal sampling rate of 0.5 ms, can easily reach 2 TB. This highlights the need for efficient data analysis procedures in Distributed Acoustic Sensing (DAS) with methods that are both computationally fast and capable of exploiting the extensive information embedded in such data. As DAS data acquisition experiments are still few in numbers, generating and using synthetic data becomes essential for evaluating performance across diverse DAS acquisition geometries and testing new data analysis techniques. Despite the constant growth of DAS systems, there is a lack of standard modelling and analysis tools that can be used within routine procedures. To address this issues, we formulated a versatile workflow designed to generate synthetic DAS data based on the convolutional model. A central component of this workflow is a travel-time calculator based on the solution of the Eikonal equation, accommodating various data acquisition geometries, including scenarios involving optical fibers deployed in deep boreholes—whether vertical or oblique. Synthetic DAS seismograms are subsequently generated by using the computed travel times, for both P and S phases, with the convolutional model. These seismograms contain several information, such as the radiation pattern of the source and the directivity of the fiber, with the possibility of selecting an arbitrary wavelet. While DAS synthetics computed using the convolutional model may be less realistic than those generated with methods like the reflectivity or the spectral element method, their computational speed is much higher. This efficiency becomes particularly crucial when dealing with the generation of extensive DAS synthetic datasets. The synthetic generation workflow can be used for 1) testing new seismic event detection and location methods for DAS data and 2) training machine learning models. Lastly, this work includes a comparative analysis of synthetics obtained through our workflow against those generated using the spectral element method, followed by an application with a waveform-based DAS event detector.

How to cite: Rapagnani, G., Gaviano, S., Pecci, D., Carelli, G., Saccorotti, G., and Grigoli, F.: A workflow for synthetic DAS data generation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5936, https://doi.org/10.5194/egusphere-egu24-5936, 2024.

How to cite: Trabattoni, A., Baillet, M., van den Ende, M., Vernet, C., and Rivet, D.: Accounting for shallow sedimentary layers for accurate earthquake localization using submarine Distributed Acoustic Sensing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6310, https://doi.org/10.5194/egusphere-egu24-6310, 2024.

Machine Learning (ML) applications in geoscience are growing exponentially, particularly in the field of seismology. ML has significantly impacted traditional seismological observatory tasks, such as phase picking and association, earthquake detection and location, and magnitude estimation. However, despite promising results, ML-based classical workflows still face challenges in analyzing microseismic data

Leveraging recent advances in Deep Learning (DL) methods, we present HEIMDALL: a grapH-based sEIsMic Detector And Locator for microseismicity. This tool utilizes an attention-based, spatiotemporal graph-neural network for seismic event detection and employs a waveform-stacking approach for event location, using output probability functions over a dense regular grid.

We applied HEIMDALL to a one-month dataset (December 2018) from the publicly available Hengill Geothermal Field in Iceland, collected during the COSEIMIQ project (active from December 2018 to August 2021). This dataset is ideal for testing seismic event detection and location algorithms due to its high seismicity rate (over 12,000 events in about two years) and the presence of burst sequences with very short interevent times (e.g., less than 5 seconds).

We assessed the methodology's performance by comparing our catalog with those obtained by two recent DL methods and one manually compiled by ISOR for the same period. The DL algorithms we considered are: (i) MALMI, a waveform-based location algorithm, and (ii) the recent GENIE graph-neural-network algorithm. For GENIE, we conducted a full repicking of continuous waveforms using the PhaseNet picking algorithm and subsequent retraining of its model to adapt it to the new seismic network.

Finally, we highlight the pros and cons of each method and discuss potential improvements for microseismic event detection and location, with a particular focus on induced seismicity monitoring operations at EGS sites.

How to cite: Bagagli, M., Grigoli, F., and Bacciu, D.: HEIMDALL: a grapH-based sEIsMic Detector And Locator for microseismicity, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6552, https://doi.org/10.5194/egusphere-egu24-6552, 2024.

On 6 February 2023 at 01:17 UTC, the Mw 7.8 Pazarcık earthquake struck south-eastern Türkiye and Syria along the East Anatolian Fault Zone (EAFZ), in the province of Kahramanmaraş. The Mw 7.6 Elbistan earthquake occurred about 9 hours later, with an epicenter located about 95 km north-northeast of the Mw 7.8 quake. The combination of these two shocks produced a devastating effect with nearly 55,000 confirmed deaths and about 1.5 million people left homeless.

In this work, we describe the Complete Automatic Seismic Processor (CASP) procedure that has been implemented to develop a large and comprehensive data set consisting of about 63,000 events of magnitude greater than 2.0, that occurred in south-eastern Türkiye between January 2019 and June 2023. The starting catalogue contains about 3.8 million waveforms recorded by 262 velocimetric and accelerometric instruments (network codes KO, TK and TU). The earthquakes were located using the Non-Linear Location technique (NLLOC) with a regional 1-D velocity model, based on the precise picking of P- and S-wave arrivals provided by the RSNI-Picker2 implemented in CASP. After several quality controls, the final high quality catalogue contains 8,475 well-located earthquakes, with a significant difference in depth with respect to the AFAD catalogue.

We present the spatio-temporal distribution of earthquakes before and after the two mainshocks, as well as the distribution of strong-motion parameters, such as peak ground acceleration (PGA), peak ground velocity (PGV), and Fourier amplitude spectra (FAS). Furthermore, preliminary results on earthquake source parameters obtained by spectral decomposition applied separately to background and clustered seismicity are also discussed.

The compiled data set can serve as a basis for studying seismic sequences during seismic crises and identifying the preparatory phase of strong earthquakes in geologically active areas.

How to cite: Colavitti, L., Tarchini, G., Spallarossa, D., Scafidi, D., Picozzi, M., Iaccarino, A. G., Bindi, D., Martínez-Garzón, P., Cotton, F., and Zaccarelli, R.: The advantages of Standardization and Data Sharing: a swift compilation of a high-quality data set for seismological studies in the East Anatolian Fault Zone, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7863, https://doi.org/10.5194/egusphere-egu24-7863, 2024.

Earthquakes are acknowledged as a potent force of nature that can cause substantial harm to populations and result in widespread damage. Therefore, having a seismic public alerting system is crucial for swiftly broadcasting warnings to the public and relevant risk agencies in the event of an earthquake. The system will send instantaneous notifications to users, allowing them to quickly implement protective measures for risk agencies, as well as offer feedback on individuals’ situations during the earthquake. In this regard, this study aims to build a wrapper for near-real-time earthquake monitoring. Our development includes four steps: (1) improvement of earthquake detection using PhaseNet (Zhu & Beroza, 2018) with PhasePApy (Chen & Holland, 2016) and the Rapid Earthquake Association and Location (REAL, Zhang et al., 2019) for picks association, (2) earthquake location refinement using the HYPOINVERSE program (Klein, 2002), (3) event classification with the CNN classification method, and (4) rapid earthquake notification through email and a locally designed application called SeismicBox2 for smartphones that include earthquake information and USGS shakemap. We conduct testing and validation of the system using earthquake data from Madagascar (archive and near-realtime)

How to cite: Rakotoarisoa, A. T. and Razafindrakoto, H. N. T.: Recent Advances in Earthquake Monitoring in Madagascar, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8614, https://doi.org/10.5194/egusphere-egu24-8614, 2024.

Real-time seismological applications are now crucial for the monitoring and surveillance of active volcanoes, as they are useful tools for the early detection of volcanic unrest. In open-vent active volcanoes,  Very Long Period (VLP) seismicity, typically associated with mild and persistent explosive activity, is of crucial importance for volcano monitoring, as its variations in occurrence rate and magnitude may prelude a phase of unrest.  Here we show a new method for the automatic real-time detection and characterization of  VLP seismicity at Stromboli active volcano (Italy).

The detection algorithm is based on the Three-Component Amplitude (TCA) obtained from waveform polarization and spectral analysis of the continuous recording, providing time of detection,  azimuth,  incidence,  amplitude, and frequency of the detected VLP events. The VLP amplitudes derived at all stations of the monitoring network, provided as peak-to-peak amplitudes and mean square amplitudes, are also used to perform an automatic localization of VLP source.

VLP detections and characterizations derived from our automatic detection algorithm are compared with detection derived from manual and automatic inspections of the seismic record and with VLP time histories from available published VLP datasets.

From this comparison, it turns out that the VLP detection time series produced by the automatic algorithm tracks fluctuations in the  VLP activity well,  as manually detected by the operators over a  ~20-year period, thus allowing us to include it into the real-time processing framework operating at Stromboli for volcano surveillance.

How to cite: Gammaldi, S., Delle Donne, D., Cantiello, P., Bobbio, A., De Cesare, W., Peluso, R., and Orazi, M.: Automatic detection and characterization of Very Long-period seismic events for volcanic monitoring applications., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9715, https://doi.org/10.5194/egusphere-egu24-9715, 2024.

The island of Vulcano gives its name to the so-called “Vulcanian eruptions”, an eruptive style with strong explosive characteristics and observed there for the first time. The last eruptive activity occurred between 1888 and 1890. Starting from mid-September 2021, an unrest, marked by relevant variations in geochemical and geophysical parameters, affected the island. Here, we analyze the seismic signals recorded from the onset of the unrest until December 2022. An increasing number of Very Long Period events was detected from September 2021 onwards, enhancing concerns linked to other measured anomalies, such as increasing CO₂ emissions and fumarole temperatures. Numerous types of signals were generally recorded on the island, partly caused by various man-made sources, such as the close-by passage of ships, dropping anchors, etc. The large variety of the seismic signals made standard amplitude-based monitoring techniques, such as RSAM, questionable. We therefore focused on creating an inventory of the recorded signals exploiting unsupervised machine learning techniques, namely Self-Organizing Maps and Cluster Analysis. We were able to identify various classes of seismic events related to volcanic dynamics and to distinguish exogenous signals, such as anthropic noise. This allowed us to visualize the development of signal characteristics efficiently. This classification can help build an effective alert tool to automatically identify different types of seismic signals, useful for surveillance purposes. Furthermore, it is a preparative step for other studies, such as event location and source process modeling.

How to cite: Langer, H., Falsaperla, S., Ferrari, F., and Spampinato, S.: Creating an Inventory of Seismic Signals at Vulcano Island, Italy, using Unsupervised Learning Techniques, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9764, https://doi.org/10.5194/egusphere-egu24-9764, 2024.

PrESENCE ANR project (2022-2025) focuses on seismic hazards induced by deep geothermal operations in northern Alsace, France, and their associated societal perception. Seismological observations are obtained using a large number of low cost internet-connected equipment (Raspberry Shake seismic station and associated open access data). The breakthrough strategy of the project relies on the deployment of the stations in residences or administrative buildings of non-seismologist volunteer citizens or authorities. The aim is to use those stations to densify the french permanent seismic network, and to improve the detection and location of seismic events, in particularly small ones. Our presentation will be focused on the Soultz-sous-Forêts and Rittershoffen areas (northern Alsace, France), which are sites of deep geothermal operations. 

The topology of the seismological network was determined by the location of permanent stations, from Epos-France permanent network (4) and public stations belonging to geothermal operators (2), the number of low-cost stations (35) to be deployed in the region, the location of deep geothermal power plants (Soultz and Rittershoffen) and the location of volunteer citizen hosts. Volunteer citizens were selected initially by word of mouth, then by a call for applications (through social networks, flyers, local newspapers). Twenty-one stations are currently (end of 2023) hosted in the area. About ten additional stations are planned to be deployed early 2024 in the area.

Based on our past experience in deploying similar networks in other contexts and regions (Mayotte, Vosges massif, Mulhouse, etc.), we have consolidated the installation of these stations to ensure reliable data acquisition and, in particular, to achieve better data completeness (acquisition directly at the station using the Seedlink protocol via a VPN, hardware watchdog). We use Ansible (an open source IT automation platform) to facilitate the deployment of Raspberry Shake stations configuration and management tasks, ensuring rapid and consistent production deployment.

The workflow for building the seismicity catalog benefits from our advances in the use of new artificial intelligence tools, such as PhaseNet, a deep learning automatic picking method, as well as in the development of a deep learning method for discrimination between earthquakes, quarry blasts and explosions. Our tests over the year 2023 show that even if the stations are installed in urban areas (and therefore in a noisy environment), the network is able to automatically detect and locate many small induced earthquakes, including around 250 with a high level of confidence, compared with the ten detected or so by the standard procedure of BCSF-Renass, the French National Observation Service.

How to cite: Turlure, M., Grunberg, M., Engels, F., Jund, H., Schlupp, A., and Schmittbuhl, J.: Contribution of SeismoCitizen Raspberry Shake dense network in monitoring induced seismicity in northern Alsace (France), EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10651, https://doi.org/10.5194/egusphere-egu24-10651, 2024.

Distributed Acoustic Sensing (DAS) is of critical value for the offshore expansion of seismological networks. The work presented here is part of the 5-years ERC ABYSS project, which aims at building a permanent seafloor seismic observatory leveraging offshore telecommunication cables along the central coast of Chile. 

In preparation for this project, a first experiment named POST was conducted from October to December 2021 on a submarine fiber-optic cable connecting the city of Concón to La Serena. DAS data were recorded continuously for 38 days over a distance of 150 km, constituting more than 37,500 virtual sensors sampled at 125 Hz. We develop a workflow to detect more than 3500 local, regional and teleseismic events with local magnitudes down to ML = 0.5, automatically processing over 72 TB of data. We show that applying those methods to DAS data combined with data from the national onland seismic network greatly increases the accuracy of the earthquake hypocenter localizations. As a first step, we perform automatic seismic phase arrival picking using PhaseNet pretrained on conventional seismological stations, followed by phase association with GaMMA. We then apply a correction of the phase picks to account for shallow sedimentary layers and invert for the event hypocenter with VELEST. Finally, we estimate a local magnitude based on peak ground displacements.  

The ABYSS project near-real time data collection started the 30th of September 2023 using three DAS units to sense two offshore telecommunications cables connecting the cities of Concón to La Serena and La Serena to Caldera. The DAS data covers over 500 km of cable, comprising 30,000 virtual sensors sampled at 62.5 Hz. These data are synchronized once a day with a storage server located in France, the volume of which is anticipated to reach an estimated 608 TB by the end of the project. By applying our workflow, tested and validated on the POST experiment, to our daily data, we are able to process data in near-real time to build a catalog that will span 5 years, and that will be used as a reference for subsequent studies within the framework of the ABYSS project. Furthermore, the size of our catalog, enriched with numerous offshore events is a significant improvement over the existing regional catalogs, which may aid future studies of the Chilean margin subduction zone seismicity. 

How to cite: Baillet, M., Trabattoni, A., van den Ende, M., Vernet, C., and Rivet, D.: A workflow for building an automatic earthquake catalog from near-real time DAS data recorded on offshore telecommunications cable in central Chile., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10668, https://doi.org/10.5194/egusphere-egu24-10668, 2024.

Distributed Acoustic Sensing (DAS) has emerged as a powerful tool in seismological applications, transforming fiber-optic cables into dense arrays of geophones that can continuously sample seismic wavefields across several kilometers. DAS data acquisition presents a versatile approach, utilizing either ad hoc installations with specific cables or leveraging existing telecommunication optical fiber-network infrastructure. Its adaptability makes DAS particularly advantageous for seismic monitoring in logistically challenging environments like volcanoes or offshore areas, where traditional seismometers may face limitations.

Conventional seismological techniques struggle to effectively process DAS data due to its unique characteristics—typically, wavefields are sampled at 1 m spacing with frequencies exceeding 1 kHz. As a result, this technology provides a detailed mapping of the seismic wavefield across the length of the fiber, and it also generates a significant amount of data compared to the sparse seismometer installations. In order to efficiently analyze these data, we introduced HECTOR, a waveform-based detection method designed specifically for DAS data (Porras et al. 2024).

In this study, we investigate the capabilities of HECTOR following preprocessing of DAS data using various characteristic functions (CF). We explore non-negative functions, including Short Term Average to Long Term Average (STALTA), Energy, and Envelope, whose peculiarity is to preserve noise. Conversely, zero-mean characteristic functions such as Short Term Average to Long Term Average derivative (STALTA derivative), Kurtosis, and Kurtosis derivative enhance signals and mitigate noise. Our objective is to assess HECTOR's performance when analyzing preprocessed data compared to raw data.

To validate our findings, we initially test the detector on synthetic data. These simulations encompass diverse optical fiber geometries, source configurations, and locations. Subsequently, we apply the algorithm to real data collected in two distinct scenarios. The first scenario involves the FORGE experiment situated in Utah, US, which entails a borehole installation of 1 km optical fiber deployed above a geothermal reservoir characterized by induced seismic activity. The second scenario involves a 90 km horizontal optical fiber deployed in the Pyrenees region. The area is characterized by natural earthquake activity with magnitudes (2.01≥ML≥0.02), alongside anthropic events due to quarry blasts. 

Our evaluation focuses on quantifying the enhancement in HECTOR's performance following the application of CFs compared to analyzing raw data.

Through this comprehensive exploration, we aim to advance the understanding of DAS data processing, demonstrating the efficacy of HECTOR across diverse scenarios. 

We would like to thank TotalEnergies for sharing this data set with us as well as Febus Optic for providing the DAS interrogator used for the data acquisition.

References: 

A Semblance-based Microseismic Event Detector for DAS Data.

• Porras, D. Pecci, G. Bocchini, S. Gaviano, M. De Solda, K. Tuinstra, F. Lanza, A. Tognarelli, E. Stucchi, F. Grigoli. Geophysical Journal International (GJI) 2024 (Accepted)

How to cite: Gaviano, S., Rapagnani, G., Pecci, D., Porras, J., Rebel, E., and Grigoli, F.: Exploring the application of Characteristic Functions on DAS data and their influence in event detection performance., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10893, https://doi.org/10.5194/egusphere-egu24-10893, 2024.

Subduction zones are among the most active tectonic areas on the planet. Their primary characteristic is the enormous amount of stress accumulated at the interface between the subducting oceanic plate and the overriding plate. The release of this stress is accommodated by a wide range of behaviours, ranging from aseismic slip (slip at speeds too slow to radiate seismic energy), through the spectrum of slow slip and tremor, to seismic slip capable of generating major earthquakes. The main investigative tools for subduction zones to map out this range of behaviour, and to assess the coupling properties of the subduction interface, involve the direct observation of ground movements through geodesy (either terrestrial or satellite-based) or through local seismic surveillance using near-field instrumentation, all of which are logistically complex, and typically only feasible on land.

Utilizing the recent expansion of seismic arrays in continental regions, we propose an alternative approach for the study of subduction zones that bypasses the aforementioned limitations through the use of teleseismic waves—recorded at a distance between 30º and 90º from the epicenter—based on the identification of the presence (or absence) of highly reflective layers at the megathrust interface. Previous studies using local seismic data have observed the presence of highly reflective layers, characterized by strong impedance contrasts, located at the megathrust interface, capable of producing a reflection in the wavefield that results into the presence of precursors of depth phases. Since impedance contrasts in the solid Earth are linked to variations in the elastic properties of the medium, reflectivity offers a window into the rheology of the plate interface. Understanding the reasons behind such strong impedance contrasts, their potential variability over time and space, could pave the way for understanding why the degree of coupling of subduction interfaces varies, whether it is related to transient processes, or if it is stable over time.

Here, we present an automated waveform processing approach designed to detect such reflections in remote seismic data, and illustrate this with a test region from the Central America subduction zone.  We analyse waveforms produced by seismic events with magnitudes ranging from 4.5 to 5.5 occurring at different times and recorded by small aperture seismic arrays. Our observations in Central America prove to be an excellent tool for studying the coupling properties of the megathrust interface. This work represents a first attempt, with the ultimate goal of mapping subduction zones and their coupling properties, even in currently inaccessible submarine areas, allowing for a better understanding of the seismic risk that subduction zones represent.

How to cite: Rappisi, F., Craig, T., and Rost, S.: Unveiling coupling properties of subduction zones with novel telesismic waveform approaches, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12443, https://doi.org/10.5194/egusphere-egu24-12443, 2024.

Focal depth of earthquakes is essential for studies of seismogenic processes and seismic hazards. Surface waves are usually the strongest seismic phases at local and regional distances, and its excitation is sensitive to source depth. We observe that the optimal period (the period corresponding to the maximum amplitude) of Rayleigh waves at local distances shows an almost linear correlation with focal depth, based on which we propose a method for resolving the focal depth of local earthquakes. We propose an automated data processing workflow, and applications to earthquakes in diverse tectonic settings demonstrate that reliable focal depth with uncertainty of 1~2 km can be determined even with one or a few seismic stations. Then, we use the Longmenshan region as a case study to systematically assess the impact of the 3D velocity model on the results through forward simulation. A total of 191 events at depths ranging from 5 to 20 km are simulated. The standard deviation between the focal depths determined by this method and the input values is approximately 1.5 km, with 95% events having errors within 2 times the standard deviation. This indicates that the method exhibits good applicability even in regions with complex velocity structures, and highlights the applicability of the method in scenarios characterized by sparse network coverage or historical events.

How to cite: He, X., Zhang, P., Ni, S., Wu, W., Chu, R., Wang, Y., and Zheng, K.: Automatic determination of focal depth with the optimal period of Rayleigh wave amplitude spectra and uncertainty assessment in 3D velocity model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13591, https://doi.org/10.5194/egusphere-egu24-13591, 2024.

The electrical network frequency (ENF) of the alternating current operated on the power grid is a well-known source of noise in digital recordings. The noise (i.e., signal) is widespread and appears not just in close proximity to high-voltage power lines, but also in instruments simply connected to the mains powers grid. This omnipresent, anthropogenic signal is generally perceived as a nuisance in the processing of geophysical data. Research has therefore been mainly focused on its elimination from data, while its benefits have gone largely unexplored. It is shown that mHz fluctuations in the nominal ENF (50 - 60Hz) induced by variations in power usage can be accurately extracted from geophysical data. This information represents a persistent time-calibration signal that is coherent between instruments over national scales. Cross-correlation of reliable reference ENF data published by electrical grid operators with estimated ENF data from geophysical recordings allows timing errors to be resolved at the 1s level. Furthermore, it is shown that a polarization analysis of particle motion at the ENF may assist in the detection of instrument orientation anomalies at the surface. Furthermore, it is explored whether this method can be applied to determine orientations of geophones inside seismic boreholes.

How to cite: Koymans, M., de Zeeuw-van Dalfsen, E., Evers, L., and Assink, J.: Passive Assessment of Geophysical Instruments Performance using Electrical Network Frequency Analysis, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15414, https://doi.org/10.5194/egusphere-egu24-15414, 2024.

Accurate earthquake localization is essential for advancing seismic processing and understanding geological structures. In this study, we explore the application of relative relocation methods—HypoDD (HD), GrowClust (GC), and Master Event (ME)—to refine event locations and analyze their implications beyond specific fault structure determination. While the primary focus is not exclusively on geological structures, the outcomes also serve broader purposes, contributing to critical aspects of seismic processing.
Our investigation employs a dataset from the REYKJANET seismic network located on the Reykjanes Peninsula in Iceland. The comparative assessment of these methods reveals significantly focused clusters in contrast to absolute event locations. Notably, individual event locations exhibit variations dependent on the chosen relocation method.
Furthermore, it is essential to note that Master Event (ME) is a program developed for event localization, offering the unique capability of sequential use. This feature proves valuable, especially in dynamic geological settings, such as the Reykjanes Peninsula in Iceland, where volcanic eruptions occur.
Additionally, we delve into the influence of control parameters for HD, GC, and ME on final location results, aiming to optimize these parameters while considering computational and memory demands. This research contributes to a comprehensive understanding of relative localization methods, emphasizing their broader applications and significance in seismic event analysis within the REYKJANET network.

How to cite: Konrádová, D., Doubravová, J., Růžek, B., and Horálek, J.: Relative methods of localization and their differences in results on the REYKJANET seismic network in Iceland, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17429, https://doi.org/10.5194/egusphere-egu24-17429, 2024.

Faults zones in the Earth’s crust alter permeability architecture relative to country rock and can function as fluid conduits. Documented cases of long-distance earthquake interactions suggest that pore-pressure gradients resulting from conduit flow can activate seismic slip where receiver faults might be sensitive to external forcing. When external stress forcing can be quantified, for example, in the form of ground motions that can be converted to stress, it provides an opportunity to measure the stress perturbation required to nucleate slip in cases where fault activation is triggered.  

In this study, we investigate the stress state of faults in the Lower Rhine Embayment (LRE), western Germany. We do so by quantifying the occurrence of remote dynamic triggering by transient stresses imparted by passing waves of distant mainshocks. The LRE hosts a system of normal faults with mean estimated slip rates of 0.1 mm/yr and moderate seismicity. We use the continuous Bensberg catalog starting in 1990 to estimate the statistical significance of seismicity rate changes surrounding teleseismic mainshocks identified as triggering candidates. We identify 21 teleseismic mainshocks with M_L > 7 (1990 – 2015) and M_L > 6 (2016 – present) that generate a theoretical peak-ground velocity (PGV) >0.02 cm/s within the study area. Two mainshocks associate with statistically significant seismicity-rate increases following the passing of their surface waves: the 1992 Roermond, and the 2021 M8.2 Chignik, Alaska earthquakes. Both mainshocks generated PGV values > 0.017 cm/s at 30s and have back-azimuths that are roughly parallel to the dominant strike of LRE faults. We observe a migrating sequence of earthquakes in the 10 days following the Roermond earthquake, where roughly half occur outside of the classical aftershock zone of ~2-3 fault lengths. We infer dynamic triggering to play a role in the generation of the migrating sequence, as migration outpaces diffusion time scales assuming realistic crustal diffusivity values of up to 3 m²/s. The July 2021 Alaska earthquake likely triggered a sequence of ~16 locatable earthquakes. The observed surface PGV values of the Alaska and Roermond earthquakes correspond to peak dynamic stress values of 1.4 kPa and < 30 kPa, respectively. Thus, stress values at the hypocentral depth of the triggered sequence of ~16 events inferred from 30s Rayleigh waves of the Alaska earthquake would correspond to 50-66% of the observed surface value.

Using remote dynamic triggering as a stress-meter to estimate stress thresholds that can potentially activate faults has important implications for earthquake physics, as well as for society. The LRE is being targeted for geothermal energy production. Prior work documents a series of 14 earthquakes of M_w > 5.0 since the 14th century, including the 1992 M_w 5.3 Roermond earthquake. Therefore, quantifying the triggerability of faults at a future energy production site prior to operation should be a key step in assessing the potential for fault reactivation.

How to cite: Roth, M. P. and Harrington, R. M.: Using remote dynamic earthquake triggering as a stress-meter of the Lower Rhine Embayment fault system in western Germany, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20910, https://doi.org/10.5194/egusphere-egu24-20910, 2024.

Distributed acoustic sensing (DAS), a technology that exhibits great potential for subsurface monitoring and imaging, has been regarded as a preeminent instrument for vibration measurements. In light of the tremendous amount of seismic data, numerous channels, and elevated noise levels, it becomes imperative to suggest an appropriate denoise procedure that is compatible with DAS data. In this regard, unsupervised deep learning with data clustering generally exhibits superior performance in facilitating the efficient analysis of sizable unlabeled data sets devoid of human bias. In addition, the clustering method is capable of detecting seismic waves, microseismic turbulence, and even unidentified new types of negligible seismic events, in contrast to a number of conventional denoising techniques. While current approaches reliant on f-k analysis remain valuable, they fail to fully exploit the information present in the wavefield due to their inability to identify the characteristic moveout observed in seismic data. In order to denoise DAS data more effectively, we investigate the capacity of the curvelet transform to extend existing deep scattering network methodologies. In this paper, we propose a novel clustering approach for the denoise processing of DAS data that utilises the Gaussian Mixture Model (GMM), curvelet transform, and unsupervised deep learning. 

The DAS data are initially subjected to the curvelet transform in order to derive the curvelet coefficients at various scales and orientations, which can be regarded as the first layer of extracted features. Following this, a deeper layer of features is obtained by applying the curvelet transform to the coefficients in the first layer. The aforementioned process continues in this manner until the depth of the layer satisfies the algorithm-determined expectation. By concatenating the curvelet coefficients from each layer, the original DAS data's features are generated. Afterwards, the signal is reduced to two dimensions using principal component analysis (PCA), which simplifies its interpretation by projecting the high-dimensional features onto two principal components, which facilitates the clustering of the features by GMM for achieving the final clustered results.

This methodology operates without the need for labels of DAS data and is highly appropriate for managing the substantial quantity and numerous channels of DAS. We used a variety of approaches, such as Bayesian information criteria and silhouette analysis, to determine the optimal number of clusters in GMM and evaluate the algorithm's clustering performance. We demonstrate the method on downhole data acquired during stimulation of the Utah FORGE enhanced geothermal system, and the results appear quite satisfactory, indicating that it can be utilised effectively to denoise DAS signals.

How to cite: Li, B., de Ridder, S., and Nowacki, A.: Clustering distributed acoustic sensing signals via curvelet transform and unsupervised deep learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4154, https://doi.org/10.5194/egusphere-egu24-4154, 2024.

Studying earthquake prediction is an interesting academic area. Earthquakes are classified as natural disasters that have the potential to inflict significant damage.  The magnitude of an earthquake provides substantial benefits, both immediately after the event and in the future, for risk assessment and mitigation. This study employs the Random Forest algorithm to predict the magnitude of earthquakes occurring on the Matano Fault in Sulawesi, Indonesia. The prediction is derived from the historical seismic data collected between 1923 and 2023, obtained from the BMKG and USGS Catalogs. The area of interest is situated along the Matano Fault in Sulawesi, Indonesia, with coordinates ranging from 2.99°S to 1.66°S and from 120.50°W to 122.47°W. The dataset comprises six attributes and is split into training and testing sets at a ratio of 70% and 30%, respectively. The variables employed in this investigation encompass origin time, latitude, longitude, magnitude, and magnitude type.  The investigation yields an RMSE score of 0.1929. Overall, the prediction model has outstanding performance, with a high degree of accuracy in predicting values that quite match the actual values. Additionally, the Root Mean Square Error (RMSE) value is 0.1929, indicating a low level of error in the predictions. this work attempts to propose an alternative approach, characterized by a straightforward and practical technique, to solve problems in the geophysics field, for instance in the area of earthquake prediction.

How to cite: Madona, , Rosid, M. S., Handoko, D., Riama, N. F., and Saiful, D.: Implementation of the Random Forest Algorithm for Magnitude Prediction in the Matano Fault Indonesia , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4807, https://doi.org/10.5194/egusphere-egu24-4807, 2024.

Responding to challenges from increasing seismic data, our study leverages deep learning to enhance seismic data processing's automation and efficiency. Recognizing Taiwan's unique geological structure, we have developed deep learning models using data from dense seismic arrays since 2018. We have integrated the Transformer model with GAN training techniques for phase picking. Our latest system, SeisBlue, has evolved from phase picking and earthquake location to include magnitude and focal mechanism estimation, primarily using SeisPolar, a CNN model for P-wave polarity classification, crucial for focal mechanism analysis. Additionally, our redesign of the seismic monitoring process emphasizes data pipelines and integrates software engineering technologies, including hardware, system environment, database, data pipelines, model version control, task monitoring, data visualization, and Web UI interaction. The model shows high proficiency in identifying P-wave polarity and deciphering focal mechanisms, with an accuracy of 85%, and precision and recall rates for three categories [positive, negative, undecidable] at [87%, 77%, 53%] and [84%, 80%, 54%], respectively. It notably achieves about 70% Kagan angle under 40 degrees for focal mechanism analysis. This semi-automated workflow, from data processing to phase picking, earthquake location, magnitude determination, focal mechanism estimation and Web UI, significantly boosts seismic monitoring's efficiency and accuracy. It facilitates quicker and more meaningful engagement for researchers in subsequent studies, marking a notable advancement in seismic monitoring and deep learning application.

How to cite: Chang, I.-H., Huang, C.-M., and Kuo-Chen, H.: SeisPolar: Seismic Wave Polarity Module for the SeisBlue Deep Learning Seismology Platform, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5031, https://doi.org/10.5194/egusphere-egu24-5031, 2024.

Earthquake forecasting is a highly complex and challenging task in seismology ultimately aiming to save human lives and infrastructures. In recent years, Machine Learning (ML) methods have demonstrated progressive achievements in earthquake processing and even labquake forecasting. Developing a more general and accurate ML model for more complex and/or limited datasets is obtained by refining the ‘ML models’ and/or enriching the ‘input data’. In this study, we present an event-based approach to enrich the input data by extracting spatio-temporal seismo-mechanical features that are dependent on the origin time and location of each event. Accordingly, we define and analyze a variety of features such as: (a) immediate features, defined as the features which benefit from very short characteristics of the considered event in time and space, (b) time-space features, based on the subsets of acoustic emission (AE) catalog constrained by time and space distance from the considered event, and (c) family features, extracted from topological characteristics of the clustered (family) events extracted from clustering analysis in different time windows. We use AE catalogs recorded by tri-axial stick-slip experiments on rough fault samples to compute event-based features. Then, a random forest classifier is applied to forecast the occurrence of a large magnitude event (M_AE>3.5) in the next time window. Results show that to obtain a more accurate forecasting model, one needs to separate background and clustered activities. Based on our results, the classification accuracy when the entire catalog data is used reaches 73.2%, however, it shows a remarkable improvement for separated background and clustered populations with an accuracy of 82.1% and 89.0%, respectively. Feature importance analysis reveals that not only AE-rate, seismic energy and b-value are important, but also family features developed from a topological tree decomposition play a crucial role for labquake forecasting.

How to cite: Karimpouli, S., Kwiatek, G., Martínez-Garzón, P., Dresen, G., and Bohnhoff, M.: Event-based features: An improved feature extraction approach to enrich machine learning based labquake forecasting, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5044, https://doi.org/10.5194/egusphere-egu24-5044, 2024.

One of the main tasks in seismology is the source characterization after an earthquake, in particular the estimate of the orientation of the fault on which an earthquake occurs and the direction of the slip. Currently, most seismological observatories compute moment tensor solutions for earthquakes above a certain magnitude threshold, but, for small to moderate earthquakes (i.e. aftershock sequences), or for large but close in time events, focal mechanism by first arrival polarities are often the only source information available (Sarao et al., 2021).

Focal mechanisms are important to better define the activated faults, to help in understanding the seismotectonic process, to improve the predicted ground shakings for early warning, the tsunami alert and the seismic hazard assessment. For these purposes, it becomes essential to produce and disseminate an estimate of the earthquake source parameters even for small events. Recently, machine learning techniques have gained significant attention and usage in various fields, including seismology where these algorithms have emerged as powerful tools in providing new insight into the earthquakes data analysis such as the prediction of the seismic wave's first arrivals polarities which can be used to compute focal mechanisms.

We present here a workflow developed to obtain earthquake focal mechanisms starting from the first p-wave polarities estimated through the method proposed by Ross et al (2018).

Our procedure consists of two stages: in the first stage, we use a combination of the available INGV web services (Bono et al., 2021) and the ObsPy functions to download the earthquake hypocentral location. We recover the waveforms recorded by the stations in the 0 -120 km distance range, and we create an input file with the appropriate information required for the prediction of the polarities for each waveform. We then use the convolutional neural network (CNN) proposed in Ross et al (2018) to obtain the polarities for each waveform, which can be UP, DOWN, or UNKNOWN. The second stage of the developed procedure aims to use the polarities that have been predicted to determine the focal mechanisms of the selected earthquakes. To do this, we use a modern Python implementation of HASH code (originally proposed in Fortran by Hardebeck et al. 2002, 2003) called SKASH (Skoumal et al. submitted). Finally, we present an application of this procedure to the September 2023, Marradi (Central Italy), seismic sequence that has been characterized by a magnitude Mw 4.9 mainshock followed by over 70 aftershocks in the magnitude range 2 - 3.4. Here, we focused on the estimation of the focal mechanism for events down to M 2.0. The application of the presented workflow permits to gain useful information about the kinematics of the earthquakes in the sequence, obtaining thus a more precise characterization of the activated structures.

How to cite: Tavani, F., Artale Harris, P., Scognamiglio, L., and Meier, M.-A.: Toward a Polarity Focal Mechanism Estimation via Deep Learning for small to moderate Italian earthquakes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5148, https://doi.org/10.5194/egusphere-egu24-5148, 2024.

Dynamic source inversion of earthquakes consists of inferring frictional parameters and initial stress on a fault consistent with co-seismic seismological and geodetic data and dynamic earthquake rupture models. In a Bayesian inversion approach, the nonlinear relationship between model parameters and data (e.g. seismograms) requires a computationally demanding Monte Carlo (MC) approach. As the computational cost of the MC method grows exponentially with the number of parameters, dynamic inversion of a large earthquake, involving hundreds to thousands parameters, shows problems with convergence and sampling. We introduce a novel multi-stage approach to dynamic inversions. We divide the earthquake rupture into several successive temporal (e.g. 0-10 s, 10-20 s, …) and spatial stages (e.g., 100 km, 200 km, …). As each stage requires only a limited number of independent model parameters, their inversion converges relatively fast. Stages are interdependent: earlier stage inversion results are a prior for a later stage inversion. Our main advancement is the use of Generative Adversarial Networks (GAN) to transfer the prior information between inversion stages, inspired by Patel and Oberai (2019). GAN are a class of machine learning algorithms originally used for generating images similar to the training dataset. Their unsupervised training is based on a contest between a generator that generates new samples and a critic that discriminates between training and generator’s images. The resulting generator should generate synthetic images/samples with noise in a low-dimensional latent space as an input. We train GANs on samples of dynamic parameters from an earlier stage of the inversion and use the GAN to suggest the dynamic parameters in a later stage of inversion. We show a proof of concept dynamic inversion of a synthetic benchmark, comparing performance of direct MC dynamic inversion with parallel tempering with our GAN approach. We efficiently handle large ruptures by adopting a 2.5D approximation that solves for source properties averaged across the rupture depth. The 2.5D modeling approach accounts for the 3D effect of the finite rupture depth while keeping the computational cost the same as in 2D dynamic rupture simulations. Additionally we show current results on the dynamic inversion of 2023 Mw 7.8 Kahramanmaraş, Turkey, earthquake.

How to cite: Premus, J. and Ampuero, J.-P.: Dynamic earthquake source inversion with GAN priors, with application to the 2023 Mw 7.8 Kahramanmaraş, Turkey earthquake, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6072, https://doi.org/10.5194/egusphere-egu24-6072, 2024.

Clustering techniques facilitate the exploration of extensive seismogram datasets, uncovering a variety of distinct seismic signatures. This study employs deep scattering networks (Seydoux et al. 2020), a novel approach in deep convolutional neural networks using fixed wavelet filters, to analyze continuous multichannel seismic time-series data spanning four months before the 2019 Ridgecrest earthquake sequence in California. By extracting robust physical features known as scattering coefficients and disentangling them via independent component analysis, we cluster different seismic signals, including those from foreshock events and anthropogenic noises. We investigate the variability of intracluster (dispersion within each cluster) and examine how it correlates with waveform properties and feature space. The methodology allows us to measure this variability, either through distance to cluster centroids or 2D manifold mapping. Our findings reveal distinct patterns in the occurrence rate, daily frequency, and waveform characteristics of these clusters, providing new insights into the behavior of seismic events versus anthropogenic noises.

How to cite: Esfahani, R., Campillo, M., Seydoux, L., Mouaoued, S., and Wang, Q.-Y.: Detailed clustering of continuous seismic waveforms with deep scattering networks: a case study on the Ridgecrest earthquake sequence, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6371, https://doi.org/10.5194/egusphere-egu24-6371, 2024.

As the amount of seismic data increasing drastically worldwide, there are ever-growing needs for high-performance automatic seismic data processing  methods and high-quality, standardized professional datasets. To address this issue, we recently updated the 'DiTing' dataset, one of the world's largest seismological AI datasets with ~2.7 million traces and corresponding labels,  with 1,089,920 three-component waveforms from 264,298 natural earthquakes in mainland China and adjacent areas, and 958,076 Pg, 780603 Sg, 152752 Pn, 25956 Sn earthquake phase arrival tags, in addition to 249,477 Pg, 41610 Pn first motion polarity tags from 2020 to 2023. We also collected 15375 non-natural earthquake waveforms in mainland China from 2009 to 2023 and a manually labeled noise dataset containing various typical noise signals from the China Seismological Network. With the support of the 'DiTing' dataset, we developed and trained several deep learning models referred as 'DiTingTools' for automatic seismic data processing. In the continuous waveform detection and evaluation of more than 1,000 stations over a year across China, 'DiTingTools' has achieved an average recall rate of 80% for event detection, mean square error ±0.2s for P phase picking, and ±0.4s for S, the average identification accuracy rate of Pg first motion polarity reached 86.7% (U) and 87.9% (D), and 75.1% (U) and 73.1% (D) for Pn first motion polarity, the average magnitude prediction error of a single station is mainly concentrated at ±0.5. The remarkable generalization capabilities of 'DiTingTools' were demonstrated through its application on the China Seismic Network. Specifically, 'DiTingPicker', a model within 'DiTingTools' designed for earthquake detection and phase picking, was employed to analyze the M 6.8  earthquake that struck Luding County, Sichuan Province, in 2022. This tool was instrumental in automatically processing data to examine the main shock and intricate fault structures of the aftershocks. The effectiveness of 'DiTingTools' in earthquake prevention and disaster reduction was further validated through these practical applications.

How to cite: Zhao, M., Xiao, Z., Zhang, B., Zhang, B., and Chen, S.: 'DiTing' and 'DiTingtools':a large multi-label dataset and algorithm set for intelligent seismic data processing established based on the China Seismological Network, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7119, https://doi.org/10.5194/egusphere-egu24-7119, 2024.

SeisBlue, a deep-learning-based earthquake monitoring system, is one of the solutions to deal with massive continuous waveform data and create earthquake catalogs. The SeisBlue workflow contains waveform data preprocessing, phase arrival detection by AI modules, phase associator, earthquake locating, earthquake catalog generation, and data visualization. The whole process can be done automatically and efficiently reduces the labor and time costs. In this study, SeisBlue is applied to three different regional seismic networks: the Formosa Array for the observation of magma chamber beneath the Tatun volcanic area, Taiwan (aperture ~80 km with 148 broadband stations and station spacing 5 km), the Chihshang seismic network (CSN) for monitoring micro-seismicity of Chihshang, Taiwan (aperture ~150 km with 14 broadband stations and station spacing 20 km), and the temporary dense nodal array for capturing the aftershock sequence of the 18 th Sep. 2022 Mw6.9 Chihshang earthquake, Taiwan (aperture ~70 km with 46 nodal stations and station spacing 3 km). The 2020 annual SeisBlue catalog of the Formosa Array contains 2,201 earthquakes, as background seismicity, compare to the 1,467 earthquakes listed in the standard catalog of the Central Weather Administration (CWA), Taiwan. The two-month SeisBlue catalog of the 2022 Mw6.9 Chihshang earthquake sequence, September to October, contains 14,276 earthquakes using the CSN dataset; however, the CWA standard catalog only lists 1,247 earthquakes during the same time period. By using waveform data of 18 th Sep. to 25 th Oct. 2022, SeisBlue detects 34,630 and 12,458 earthquakes extracted from the datasets of the dense nodal array and CSN, respectively. SeisBlue can effectively detects both background and aftershock seismicity and extracts small earthquakes via dense arrays.

Keywords: AI earthquake monitoring system, deep learning, AI earthquake catalog, SeisBlue, automatic waveform picking

How to cite: Pan, S.-Y., Sun, W.-F., Huang, C.-M., and Kuo-Chen, H.: Deep learning-based earthquake catalogs extracted from threebroadband/nodal seismic arrays with different apertures in Taiwan bySeisBlue, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7309, https://doi.org/10.5194/egusphere-egu24-7309, 2024.

In the past decade, the development of the Deep Neural Network formalism has emerged as a promising approach for addressing contemporary task in seismology, particularly in the effective and potentially automated processing of extensive datasets, such as seismograms. In this study, we introduce a 4D Neural Network (NN) based on the U-Net architecture, capable of simultaneously processing data from the entire seismic network. Our dataset comprises records/seismograms of Acoustic Emission (AE) events obtained during a laboratory loading experiment on a rock specimen. While AE event records share similarities with real seismograms, they exhibit simplifications in certain features.
To assess the capability of the proposed NN in handling complex data, including occurrences of multiple events observed during experiments, we generated double-event seismograms through the augmentation of unambiguous single-event seismograms. These augmented datasets were employed for training, validation, and testing of the NN. Despite the individual station detection rate being approximately 30%, the simultaneous processing of multiple stations significantly increased efficiency, achieving an overall detection rate of 97%.
In this work, we treat seismograms as "images," adopting an approach that proves to be fruitful. The simultaneous processing of seismograms, coupled with this image-based treatment, demonstrates high potential for reliable automatic interpretation of seismic data. This approach (possibly combined with other methodologies), holds promise for seismogram processing.

How to cite: Kolar, P., Petružálek, M., Šílený, J., and Lokajíček, T.: Double Acoustic Emission events detection using U-net Neural Network, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8233, https://doi.org/10.5194/egusphere-egu24-8233, 2024.

In seismology, accurately associating seismic phases to their respective events is crucial for constructing reliable seismicity catalogs. This study presents a comprehensive benchmark analysis of five seismic phase associators, including machine learning based solutions, employing synthetic datasets tailored to replicate the seismicity characteristics of real seismic data in a crustal and a subduction zone scenario.

The synthetic datasets were generated using the NonLinLoc raytracer, using real station distributions and velocity models and simulating a large range of seismic events across different depths. In order to generate sets of picks with quality and diversity similar to a real-world dataset, some modifications such as adjustments to arrival times simulating picking errors, selective station exclusion, incorporation of false picks, were included. Such a controlled environment allowed for the assessment of associator performance under a range of different conditions.

As part of project MILESTONE, we compared the performance of five state-of-the-art seismic phase associators (PhaseLink, GaMMA, REAL, GENIE, and PyOcto) across multiple scenarios, including low-noise environments, high-noise background activity, out-of-network events, and complex aftershock sequences. Each associator's accuracy in identifying and associating true events amidst noise picks and its ability to handle overlapping sets of arrival times from different events were rigorously evaluated.

Additionally, we conducted a systematic comparison of the advantages and disadvantages of each associator, attempting a fair and unbiased evaluation. This included assessing their processing times, a critical factor in operational seismology. Our findings reveal significant differences in the precision and robustness of these associators.

This benchmark study not only underscores the importance of robust phase association in seismological research but also paves the way for future enhancements in seismic data processing techniques. The insights gained from this analysis are expected to significantly contribute to the ongoing efforts in seismic monitoring and hazard assessment, particularly in the realm of machine learning applications.

How to cite: Puente Huerta, J. A., Münchmeyer, J., McBrearty, I., and Sippl, C.: Benchmarking seismic phase associators: Insights from synthetic scenarios, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8913, https://doi.org/10.5194/egusphere-egu24-8913, 2024.

Exploring large datasets of continuous seismic data is a challenging task. When targeting signals of interest with a good a priori knowledge on the signal's properties, it is possible to design a dedicated processing pipeline (earthquake detection, noise reduction, etc.). Many other sources can sign up in the data, with characteristics that differ from the targetted ones (changes in noise frequency, modulating signals, etc.). In this case, it is difficult to design a processing pipeline that will be robust to all the possible sources. In this work, we propose to use unsupervised learning to explore the data and reveal patterns in an interpretable way. We extract relevant features of continuous seismic data with a deep scattering network —a deep convolutional neural network with interpretable feature maps— and experiment various classical machine learning tools (clustering, dimensionality reduction, etc.) to reveal and interpret patterns in the data. We apply this method to various cases including to a decade of continuous data in the region of Guerrero, Mexico, and interpret the results in terms of seismicity and external datasets.

How to cite: Seydoux, L., Steinmann, R., Mouaoued, S., Esfahani, R., and Campillo, M.: Revealing and interpreting patterns from continuous seismic data with unsupervised learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8924, https://doi.org/10.5194/egusphere-egu24-8924, 2024.

Identifying earthquakes and selecting their arrival phases are essential tasks in earthquake analysis. As more seismic instruments become available, they produce vast amounts of seismic data. This necessitates the implementation of automated algorithms for efficiently processing earthquake sequences and for recognising numerous events that might go unnoticed with manual methods.

In this study, we employed the EQTransformer, trained on the INSTANCE dataset, and utilised PyOcto for phase association, focusing specifically on the Petrinja earthquake series. This series is particularly interesting for its initial phase, which was marked by a limited number of seismometers in the epicentral area during the onset of the sequence in late December 2020. This limitation was subsequently addressed by the swift deployment of five additional stations near the fault zone in early January 2021, followed by a gradual expansion of the seismic network to over 50 instruments.

Our analysis covers the Petrinja earthquake series from its onset on December 28, 2020, up to present, offering a complete and up-to-date view of the seismic activity as the seismic activity is still higher than in the interseismic period. We compare our findings from the machine learning-generated catalogue with a detailed manual catalogue. Focusing on the first week of the series, when the seismic network was sparse and there was a high frequency of overlapping earthquakes, we achieved a recall of 80% and a precision of 81% for events with local magnitude greater than 1.0. In contrast, for the subsequent six months of processed data, a period still characterised by a high frequency of earthquakes but with the fully expanded network, our recall improved dramatically to 95% with over 20,000 detected events. This comparison allows us to demonstrate the challenges, evolution, and effectiveness of automatic seismic monitoring throughout the earthquake sequence.

How to cite: Sindija, D., Mustac Brcic, M., Hetenyi, G., and Stipcevic, J.: An up-to-date seismic catalogue of the 2020 Mw6.4 Petrinja (Croatia) earthquake sequence using machine learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9120, https://doi.org/10.5194/egusphere-egu24-9120, 2024.

The use of P-wave receiver function and surface wave dispersion data is crucial in exploring the structure of the Earth's crust and upper mantle. Typically, to address the ambiguity resulting from using a single type of dataset for inversion, these two types of seismic data, which have different sensitivities to shear wave velocity structure, are jointly inverted to achieve a detailed velocity structure. However, methods that rely on a linearized iterative joint inversion approach depend on the initial model selection, while non-linear joint inversion frameworks based on model parameter space search are computationally intensive. To address these challenges, this study suggests employing a deep learning strategy for the joint inversion of P-wave receiver function and surface wave dispersion data. Two distinct neural networks are developed to extract features from the P-wave receiver function and surface wave dispersion data, and different loss functions are tested to train the proposed neural network. The proposed method has been applied to actual seismic data from South China, and the results are comparable to those obtained by jointly inverting body wave first travel-time, P-wave receiver function, and surface wave dispersion data.

How to cite: Hu, J.: Joint inversion of P-wave receiver function and surface wave dispersion data based on deep learning. , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9621, https://doi.org/10.5194/egusphere-egu24-9621, 2024.

Objectives and Scope:

Deep learning's efficacy in seismic interpretation, including denoising, horizon or fault detection, and lithology prediction, hinges on the quality of the training dataset. Acquiring high-quality seismic data is challenging due to confidentiality, and alternative approaches like using synthetic or augmented data often fail to adequately capture realistic wavefield variations, ambient noise, and complex multipathing effects such as multiples. We introduce an innovative seismic data augmentation method that incorporates realistic geostatistical features and synthetic multiples, enhancing the training and transferability of deep neural networks in multi seismic applications.

Methods and Procedures:

Our method comprises two primary steps: (1) Creating augmented impedance models from existing seismic images and well logs, and (2) Simulating seismic data from these models. The first step merges Image-Guided Interpolation (IGI, Hale et al., 2010) and Sequential Gaussian Simulation (SGS) to generate models that retain original structural features of the input seismic image and introduce random small-scale features aligned with the geostatistical properties of the input seismic data. The second step employs reflectivity forward modeling method (Kennett, 1984) to simulate both primary and multiple seismic data trace-by-trace. This approach, summing up infinite order internal multiples, effectively reproduces the full properties of reflection wavefields, which is a good approximation in areas without rapidly changing structures.

Results and Observations:

Our numerical tests validate the method's effectiveness. The IGI technique interpolates well log data into gridded velocity models, maintaining seismic horizons and smoothing fault features. The SGS method then generates stochastic velocity model implementations preserving the geostatistical distribution of the input seismic data. The resulting reflectivity forward modeling successfully distinguishes between multiples and primaries, facilitating the creation of nuanced training datasets and labels.

Further tests involve training two Transformer-based seismic fault detection neural networks: one with conventional data lacking multiples and another with our augmented data incorporating multiples. While both networks exhibit similar validation performance, their generalization capabilities differ markedly. The network trained with conventional data shows reduced accuracy and fault detection reliability on synthetic field data. In contrast, the network trained with our augmented data demonstrates better precision, accuracy, and recall on the same dataset.

Significance and Novelty:

Our approach generates augmented seismic data that retains the original seismic cubes' and well logs' geostatistical features and multiples, crucial for training deep learning models with high transferability for various seismological tasks. This method's novelty lies in its consideration of geostatistical characteristics, wavelet fluctuations, and multiples. The resulting data is more complex, varied, and realistic compared to conventional augmentation methods. Neural networks trained on this data exhibit enhanced transferability over those trained with traditional synthetic data incorporating only random noise. This advancement represents a significant leap in seismic data processing and interpretation, particularly for deep learning applications in geophysics.

How to cite: Zhou, T.: Seismic data augmentation with realistic geostatistical features and synthetic multiples for multi deep learning tasks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10219, https://doi.org/10.5194/egusphere-egu24-10219, 2024.

The exponential growth of seismological data and machine learning methods offer new perspectives for analysing the factors controlling seismic ground motions and predicting earthquake shaking for earthquake engineering. However, the first models (e.g. Derras et al., 2012) using "simple" neural networks to predict seismic motions did not convince the earthquake engineering community, which continued to use more conventional models. We analyse the weaknesses (from the perspective of engineering seismology) of this first generation of ML-based ground motion models and explain why this first generation did not provide sufficient added value compared to conventional models.  Based on this experience, we propose two evolutions and new methods that have advantages over conventional methods and therefore have greater potential.  A first class of models (e.g. Lilienkamp et al., 2022), based on a U-net neural network, predicts spatial variations in seismic motions (e.g. site effects in three-dimensional basins) by considering seismic motions in map form. A second class of approaches) combines AI methods (conditional generative adversarial networks,  Esfahani et al., 2023) and hybrid databases (observations and simulations selected for their complementarity) to train simulation models capable of generating not only a few parameters (e.g. PGA) describing ground motions, but the full acceleration time histories. We will discuss the potential advantages of this new generation of ML-based methods compared to conventional methods, but also the challenges (and proposed solutions) to best combine simulations and observations, and to calibrate both the best estimate and the variability of future ground motions.

How to cite: Cotton, F., Esfahani, R., and Lilienkamp, H.: Failures, successes and challenges of machine-learning-based engineering ground-motion models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12197, https://doi.org/10.5194/egusphere-egu24-12197, 2024.

Seismic waveforms of teleseismic earthquakes are highly complex since they are a superposition of numerous phases that correspond to different wave types and propagation paths. In addition, measured waveforms contain noise contributions from the surroundings of the measuring station. The regional distribution of seismological stations is often relatively sparse, in particular in regions with low seismic hazard such as Northern Germany. However, a detailed knowledge of the seismic wavefield generated by large earthquakes can be crucial for highly precise measurements or experiments that are carried out for instance in the field of particle physics, where seismic wavefields are considered noise. While synthetic waveforms for cataloged earthquakes can be computed for any point on the Earth’s surface, they are based on a highly simplified Earth model. As a first step towards the prediction of a dense seismic wavefield in a region with sparsely distributed stations, we propose to train a convolutional neural network (CNN) to predict measured waveforms of large earthquakes from their synthetic counterparts. For that purpose, we compute synthetic waveforms for numerous large earthquakes of the past years with the IRIS synthetics engine (Syngine) and use the corresponding actual measurements from stations in Northern Germany as labels. Subsequently, we test the performance of the trained neural network for events not part of the training data. The promising results suggest that the neural network is able to largely translate the synthetic waveforms to the more complex measured ones, indicating a means to overcome the lack of complexity of the Earth model underlying the synthetic waveform computation and paving the way for a large-scale prediction of the seismic wavefield generated by earthquakes.

How to cite: Bauer, A., Walda, J., and Hammer, C.: Deep learning prediction of measured earthquake waveforms from synthetic data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12357, https://doi.org/10.5194/egusphere-egu24-12357, 2024.

Machine learning-based algorithms are emerging in mining earthquake occurrences from continuous recordings, replacing some routine processes by human experts, e.g., phase picking and phase association. In this study, we explore the combination of phase picker and phase associator with challenging application scenarios: the complex seismogenic structure, wide study area (15 degrees of both longitude and latitude and a depth of 600 km), hundreds of stations, and intensive seismicity during the 2016 Mw7.8 Kaikōura earthquake that correlates with at least seven faults. The deep learning-based phase pickers usually follow the prototype of PhaseNet, which maps the phase arrivals into truncated Gaussian functions with a customized model. Recent studies have shown poor generalizability of the advanced models on data out of the training distribution. In this study, we argue that appropriate data augmentation enables the RED-PAN model, trained on the Taiwanese data, to generalize well on New Zealand data even under intense seismicity. We applied RED-PAN on year-long continuous recordings over 439 stations of the GeoNet during 2016 and 2017. RED-PAN produces approximately three million P-S pairs over the New Zealand-wide network, enabling the exploration of the advanced phase associators' robustness on local and regional scales and under intense seismicity, e.g., back-projection, GaMMA, and PyOcto. Finally, we developed a six-stage automatic pipeline producing a high-quality earthquake catalog: phase picking, phase association, 3-D absolute location by NonLinLoc, magnitude estimation, weighted template matching, and 3-D relative location by GrowClust. 

How to cite: Liao, W.-Y., Lee, E.-J., Manea, E., Aden, F., Fry, B., Kaiser, A., and Rau, R.-J.: Recipe For Regular Machine Learning-based Earthquake Cataloging: A Systematic Examination in New Zealand, from Local to Regional Scale, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14134, https://doi.org/10.5194/egusphere-egu24-14134, 2024.

Earthquake forecasting and prediction are going through achievements in short-term early warning systems, hazard assessment of natural and human-induced seismicity, and prediction of laboratory earthquakes.

In laboratory settings, frictional stick-slip events serve as an analog for the complete seismic cycle. These experiments have been pivotal in comprehending the initiation of failure and the dynamics of earthquake rupture. Additionally, lab earthquakes present optimal opportunities for the application of machine learning (ML) techniques, as they can be generated in long sequences and with variable seismic cycles under controlled conditions. Indeed, recent ML studies demonstrate the predictability of labquakes through acoustic emissions (AE). In particular, Time to Failure (TTF) (defined as the time remaining before the next main labquake and retrieved from recorded shear stress) has been predicted for the main lab-event considering simple AE features as the variance.

A step forward in the state of the art is the prediction of Time To Failure (TTF) by using raw AE waveforms. Here we use deep learning (DL) to predict not only the TTF of the mainshock with raw AE time series but also the TTF of all the labquakes, foreshocks or aftershocks, above a certain amplitude. This is a great finding for several reasons, mainly: 1) we can predict TTF by using traces that don’t contain EQ (but only noise); 2) we can improve our knowledge of seismic cycle predicting also TTF of foreshocks and aftershocks.

This work is promising and opens new opportunities for the study of natural earthquakes just by analyzing the continuous raw seismogram. In general laboratory data studies underscore the significance of subtle deformation signals and intricate patterns emanating from slipping and/or locked faults before major earthquakes. Insights gained from laboratory experiments, coupled with the exponential growth in seismic data recordings worldwide, are diving into a new era of earthquake comprehension.

How to cite: Laurenti, L., Johnson, C., Tinti, E., Galasso, F., Johnson, P., and Marone, C.: Deep learning to predict time to failure of lab foreshocks and earthquakes from fault zone raw acoustic emissions, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14239, https://doi.org/10.5194/egusphere-egu24-14239, 2024.

Recent advances in scientific machine learning have led to major breakthroughs in predicting Partial Differential Equations’ solutions with deep learning. Neural operators, for instance, have been successfully applied to the elastic wave equation, which governs the propagation of seismic waves. They give rise to fast surrogate models of earthquake simulators that considerably reduce the computational costs of traditional numerical solvers.

We designed a Multiple-Input Fourier Neural Operator (MIFNO) and trained it on a database of 30,000 3D earthquake simulations. The inputs comprise a 3D heterogeneous geology and a point-wise source given by its position and its moment tensor coordinates. The outputs are velocity wavefields recorded at the surface of the propagation domain by a grid of virtual sensors. Once trained, the MIFNO predicts 6.4s of ground motion velocity on a domain of size 10km x 10km x 10km within a few milliseconds.

Our results show that the MIFNO can accurately predict surface wavefields for all earthquake sources and positions. Predictions are assessed in several frequency ranges to quantify the accuracy with respect to the well-known spectral bias (i.e. degradation of neural networks’ accuracy on small-scale features). Thanks to its efficiency, the MIFNO is also applied to a database of real geologies, allowing unprecedented uncertainty quantification analyses. This paves the way towards new seismic hazard assessment methods knowledgeable of geological and seismological uncertainties.

How to cite: Lehmann, F., Gatti, F., Bertin, M., and Clouteau, D.: A deep learning-based earthquake simulator: from source and geology to surface wavefields, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14438, https://doi.org/10.5194/egusphere-egu24-14438, 2024.

Continuous seismological datasets offer insights for the understanding of the dynamics of many geological structures (such as landslides, ice glaciers, and volcanoes) in relation to various forcings (meteorological, climatic, tectonic, anthropic) factors. Recently, the emergence of dense seismic station networks has provided opportunities to document these phenomena, but also introduced challenges for seismologists due to the vast amount of data generated, requiring more sophisticated and automated data analysis  techniques. To tackle this challenge, supervised machine learning demonstrates promising performance; however, it necessitates the creation of training catalogs, a process that is both time-consuming and subject to biases, including pre-detection of events and subjectivity in labeling. To address these biases, manage large data volumes and discover hidden signals in the datasets, we introduce a Self-Supervised Learning (SSL) approach for the unsupervised clustering of continuous seismic data. The method uses siamese deep neural networks to learn from the initial data. The SSL model works by increasing the similarity between pairs of images corresponding to several representations (seismic traces, spectrograms) of the seismic data. The images are positioned in a 512-dimensional space where possible similar events are grouped together. We then identify groups of events using clustering algorithms, either centroid-based or density-based. 

The processing technique is applied to two dense arrays of continuous seismological datasets acquired at the Marie-sur-Tinée landslide and the Pas-de-Chauvet rock glacier, both located in the South French Alps. Both datasets include over a month of continuous data from more than 50 stations. The processing technique is then applied to the continuous data streams from either a single station or from the whole station network. The clustering products show a high number of distinct clusters that could potentially be considered as produced by different types of sources. This includes the anticipated main types of seismicity observed in these contexts: earthquakes, rockfalls, natural and anthropogenic noises as well as potentially yet unknown sources. Our SSL-based clustering approach streamlines the exploration of large datasets, allowing more time for detailed analysis of the mechanisms and processes active in these geological structures.

How to cite: Rimpot, J., Hibert, C., Malet, J.-P., Forestier, G., and Weber, J.: Self-Supervised Learning Strategies for Clustering Continuous Seismic Data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15571, https://doi.org/10.5194/egusphere-egu24-15571, 2024.

The intelligible understanding of natural phenomena such as earthquakes is one of the main epistemic aims of science. Its very aims are shaped by technological discoveries that can change the cognitive fabric of a research field. Artificial intelligence, of which machine learning (ML) is one of the fundamental pillars, is the cutting-edge technology that promises the greatest scientific breakthroughs. In particular, great hopes are placed in ML
models as a source of inspiration for the formulation of new concepts or ideas, thanks to their ability to represent data at different levels of abstraction inaccessible to humans alone.
However, the opacity of ML models is a major obstacle to their explanatory potential. Although efforts have recently been made to develop ML interpretability methods that condense the complexity of ML models into human-understandable descriptions of how they work and make decisions, their epistemic success remains highly controversial. Because they are based on approximations of ML models, these methods can generate misleading explanations that are overfitted to human intuition and give an illusory sense of scientific understanding.
In this study, we address the question of how to limit the epistemic failure of ML models. To answer it, we use the example of an ML model trained to provide insights into how to better forecast newly emerging earthquakes associated with the expansion of hydrocarbon production in the Delaware Basin, West Texas. Through this example, we show that by changing our conception of explanation models derived from interpretability methods,
i.e. idealised scientific models rather than simple rationalisations, we open up the possibility of revealing promising hypotheses that would otherwise have been ignored. Analysis of our interpreted ML model unveiled a meaningful linear relationship between stress perturbation distribution values derived from ML decision rules and earthquake probability, which could be further explored to forecast induced seismicity in the basin and beyond. This observation also helped to validate the ML model for a subsequent causal approach to the factors underlying earthquakes.

How to cite: Renouard, A., Stafford, P., Goes, S., Whittaker, A., and Hicks, S.: How to Limit the Epistemic Failure of Machine Learning Models?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16930, https://doi.org/10.5194/egusphere-egu24-16930, 2024.

Fault zone properties can evolve significantly during the seismic cycle in response to stress changes, microcracking, and wall rock damage. Distinguishing subtle changes in seismic behavior prior to earthquakes, even in locations with dense seismic networks, is challenging. In our previous works, we applied Deep Learning (DL) techniques to assess alterations in elastic properties before and after large earthquakes. To do that, we used 10,000 seismic events that occurred in a volume around the October 30th 2016, Mw 6.5, Norcia earthquake (Italy), and trained a DL model to classify foreshocks, aftershocks, and time-to-failure (TTF), defined as the elapsed time from the mainshock. Our model exhibited outstanding accuracy, correctly identifying foreshocks and aftershocks with over 90% precision and achieving good results also in time-to-failure multi-class classification.

To build upon our initial findings and enhance our understanding, this follow-up investigation aims to thoroughly examine the model's performance across various parameters. First, we will investigate the influence of earthquake magnitude on our model, specifically assessing whether and to what extent the model's accuracy and reliability are maintained across varying minimum magnitude thresholds included in the catalog. This aspect is crucial to understand whether the model's predictive power remains consistent at different magnitudes of completeness. In terms of source location, our study will extend to evaluate the model's reliability by selectively excluding events from specific locations within the study area, and alternatively, by expanding the selection criteria. This approach allows us to discern the model's sensitivity to spatial variations and its ability to adapt to diverse seismic activity distributions. Furthermore, we’ll pay particular attention to the analysis of null-results. This involves meticulously analyzing cases where the model does not perform effectively, producing low-precision or inconclusive results. By carefully examining these scenarios, our goal is to further assess and confirm the high-performance results obtained from previous works.

Our results highlight the promising potential of DL techniques in capturing the details of earthquake preparatory processes, acknowledging that while complexities of machine learning models exist, ML models have the potential to open hidden avenues of future research.

How to cite: Paoletti, G., Laurenti, L., Tinti, E., Galasso, F., Collettini, C., and Marone, C.: Further investigations in Deep Learning for earthquake physics: Analyzing the role of magnitude and location in model performance, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17061, https://doi.org/10.5194/egusphere-egu24-17061, 2024.

A signal, coined PEGS for Prompt Elasto-Gravity Signal, was recently identified on seismograms preceding the seismic waves generated by very large earthquakes, opening promising applications for earthquake and tsunami early warning. Nevertheless, this signal is about 1,000,000 times smaller than seismic waves, making its use in operational warning systems very challenging. A Deep Learning algorithm, called PEGSNet, was later designed to estimate, as fast as possible, the magnitude of an ongoing large earthquake from PEGS recorded in real time. PEGSNet was applied to Japan and Chile and proved capable of tracking the magnitude of the M_w 9.1 Tohoku-oki and M_w 8.8 Maule earthquakes within a few minutes from the events origin times. Here, we apply this algorithm to a very well instrumented region: Alaska. We find that, applied to such a dense seismic network, the performance of PEGSNet is drastically improved, with robust performances obtained for earthquakes with magnitudes down to 7.8. The gain in resolution also allows us to estimate the focal mechanism of the events in real time, providing all the information required for tsunami warning within less than 3 minutes.

How to cite: Bletery, Q., Juhel, K., Licciardi, A., and Vallée, M.: Machine learning based rapid earthquake characterization using PEGS in Alaska, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18303, https://doi.org/10.5194/egusphere-egu24-18303, 2024.

In April 2023 a seismic survey was carried out in southern Skagerrak using a towed-streamer and airgun setup. The aim of the survey was investigating the suitability of the Jammerbugt structure for CO2 storage. A fiber-optic cable is co-located with the Skagerrak 4 high-voltage interconnector cable between Denmark and Norway. The fiber was crossed multiple times by the surveying ship. Relative strain was measured along a 80 km section of the fiber using Distributed Acoustic Sensing (DAS) during the active seismic survey. 

Seismic arrivals from the airgun shots were clearly recorded by the fiber. The DAS data also contains a large number of other signals caused by passing ships and wave interactions. Shot-gathers were extracted from the DAS data using the timing and location of airgun shots. These were subsequently processed and compared with traditional shot-gathers recorded by the towed-streamer. The DAS data contains distinguishable direct, refracted and surface wave arrivals from the airgun shots. Reflection hyperbolas are also observed in the DAS data at larger receiver-offsets, but only when the source is close to the fiber.

The comparison indicates that DAS is able to at least partially record the same wavefield from an active source as that recorded by hydrophones. Consequently the DAS data can be used for imaging and subsurface characterization.

The utilized DAS interrogator unit is owned by the danish transmission system operator, Energinet, who provided the DAS data for this study. Data processing is carried out using MatLab and Promax.

How to cite: Damsgård, J., Hansen, T., Voss, P., Hansen, H., Steffansen, S., Nørmark, E., and Fyhn, M.: Subsurface characterization using Distributed Acoustic Sensing (DAS) on an offshore fiber between Denmark and Norway, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-414, https://doi.org/10.5194/egusphere-egu24-414, 2024.

Ubiquitous acoustic gravity waves in the atmosphere lead to elastic deformations of the Earth’s surface via ambient barometric pressure variations at ground level. The induced ground deformations are composed of vertical and horizontal displacements as well as ground tilts or equivalently ground rotations around horizontal axes. To make inferences about background levels of rotational ground motions we exploit the fact that ground tilts are sensed by both suitably oriented gyroscopes, as well as horizontal component accelerometers through tilt coupled gravity.  Based on 20 years of data from the Global Seismic Network (GSN) we estimate coherence and admittance between ambient atmospheric pressure and horizontal acceleration from collocted sensors.

Since atmospheric acoustic gravity waves propagate too slowly to efficiently excite Rayleigh waves in the Earth, we attribute horizontal accelerations which are coherent with pressure to tilt coupled gravity. Based on this line of reasoning and by restricting the analysis to time windows with high coherence, we can estimate lower bounds of background tilt and background rotation rate for all GSN stations and for the GSN as a whole. We find that below 20mHz and in the least noisy time windows the  pressure induced background rotation rate is 30dB higher than similar estimates based on the assumption that the terrestrial noise floor for rotations around a horizontal axis is defined by Rayleigh wave motion.

A notable consequence of the above findings is that for frequencies below 20 mHz  atmospheric pressure induced ground tilts lead not only to the well established large difference between background noise levels for vertical and horizontal seismic accelerations, but also for rotations around vertical and horizontal axes.  We will present preliminary new rotational low noise models valid for frequencies below the band of the marine microseisms. The caveat for such models is that they are drived from inertial seismometers and not from gyroscopes. Data from the ROMY gyroscope are analyzed in a companion poster by Brotzer et al. in this same session SM3.3

How to cite: Widmer-Schnidrig, R. and Brotzer, A.: On the limit imposed by variable atmospheric pressure for the observation of small terrestrial rotations around horizontal axes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1797, https://doi.org/10.5194/egusphere-egu24-1797, 2024.

This work delivers an in-depth bibliometric analysis of Distributed Acoustic Sensing (DAS) research within the realm of geophysics, covering the period from 2012 to 2022 and drawing on data from the Web of Science. By employing bibliographic and structured network analysis methods, including the use of VOSviewer®, the study highlights the most influential scholars, leading institutions, and pivotal research contributions that have significantly shaped the field of DAS in geophysics. The research delves into key collaborative dynamics, unraveling them through co-authorship network analysis, and delves into thematic developments and trajectories via comprehensive co-citation and keyword co-occurrence network analyses. These analyses elucidate the most robust and prominent areas within DAS research. A critical insight gained from this study is the rise of 'Photonic Seismology' as an emerging interdisciplinary domain, exemplifying the fusion of photonic sensing techniques with seismic science. The paper also discuss certain limitations inherent in the study, and concludes with implications for future research.

How to cite: Cheng, F., Zhao, K., Zhao, L., and Ajo-Franklin, J.: Photonic Seismology: A New Decade of Distributed Acoustic Sensing in Geophysics from 2012 to 2022, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2514, https://doi.org/10.5194/egusphere-egu24-2514, 2024.

The technology of distributed fibre-optic sensors (DFOS) has developed over more than four decades, initially confined to temperature sensing, which remains a valuable tool. In the last 15 years, however, the addition of distributed vibration/acoustic sensing has vastly increased the use of DFOS in geophysical applications.

The combination of acoustic, temperature and static strain measurement has provided a deeper insight in the subterranean and subsea realms, ranging from hydrocarbon and geothermal energy production, earthquake monitoring to oceanography and glaciology. Spare or disused capacity on long-distance fibre-optic communications links has opened up many opportunities for sensing the environment, especially in oceanography. Techniques developed for oil and gas exploration and production have led to reliable methods for conveying optical fibres in the very hostile found in boreholes and this has extended the applications of DFOS to understanding tectonic movements and also to decarbonising the energy supply industry, e.g. in carbon sequestration and geothermal production.

The talk will provide a brief overview of the instrumentation used for DFOS and it will discuss some of the key applications in geophysics. It will also examine some of the untapped opportunities and how technological improvements might enable these to be realised.  

How to cite: Hartog, A.: Overview of distributed fibre-optic sensing in geophysical applications., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2827, https://doi.org/10.5194/egusphere-egu24-2827, 2024.

The reliable estimation of earthquake magnitude and stress drop are key in seismology. The novel technology of distributed acoustic sensing (DAS) holds great promise for source parameter inversion owing to the measurements' high spatial density. Here, I demonstrate the robustness of DAS for magnitude and stress drop estimation using the empirical Green's function deconvolution method. This method is applied to nine co-located earthquakes recorded in Israel following the 2023 Turkey earthquakes. Spectral ratios were stacked along the fiber, and fitted with a relative Boatwright source spectral model. Excellent fits were obtained even for similar sized earthquakes. Stable seismic moments and stress drops were calculated assuming the moment of one earthquake is known. DAS derived estimates were found to be more stable and reliable than those obtained using a dense accelerometer network. The results demonstrate the great potential of DAS for source studies.

How to cite: Lior, I.: Accurate Magnitude and Stress Drop Using the Spectral Ratios Method Applied to Distributed Acoustic Sensing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3214, https://doi.org/10.5194/egusphere-egu24-3214, 2024.

Distributed Acoustic Sensing (DAS) has emerged as a transformative technology in recent years, effectively converting optical fibers into dense seismic arrays. Numerous studies have demonstrated the widespread applications of DAS in seismology, including earthquake detection and subsurface structure imaging. In terms of earthquake source studies using DAS, the conventional approach for determining earthquake magnitudes primarily relies on maximum amplitude measurements. However, this approach faces limitations, such as unknown cable couplings and instrument responses, single-component sensing, complex source radiation patterns, and uncommon amplitude saturation behaviors. To overcome these challenges, we propose a novel method that calculates earthquake magnitudes based on coda waves using DAS. Utilizing a 10 km-long DAS array deployed in Ridgecrest, California, we derive coda wave energy decay to estimate source amplitude terms. Our findings reveal a strong linear correlation between these estimates and seismic magnitudes estimated using broadband seismic network. Furthermore, our study provides insights into the attenuation structure beneath the DAS array, aligning well with shallow velocity structures. This study not only advances our understanding of seismic source characterization using DAS, but also paves the way for more accurate earthquake magnitude estimation using DAS.

How to cite: Ye, P. and Wang, X.: Earthquake Coda Magnitude with Distributed Acoustic Sensing at Ridgecrest, California, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3331, https://doi.org/10.5194/egusphere-egu24-3331, 2024.

Distributed Acoustic Sensing (DAS) technology is currently used to monitor seismic activity, offering a unique spatially-dense representation of the along-the-cable strain wavefield. Traditional seismic networks typically rely on the timing of specific seismic phases to estimate source locations. In this context, DAS arrays may fail to provide accurate traveltimes because of spatially-heterogeneous waveforms. The motivations are (but not limited to) the directional sensitivity, the heterogeneous cable ground-coupling and the enhanced sensitivity to lateral variations in the medium elastic properties. The resulting fluctuations in signal-to-noise ratios of the dense DAS channels pose significant challenges in the automatic picking of body phases, e.g., P-wave Absolute Arrival Times (P-ARTs). Consequently, the complex distribution of the estimated traveltimes impacts the accuracy of event locations, especially if incorrect assumptions on error statistics (e.g., Normal distribution) are considered. In this study, we address this issue by exploiting the intrinsic DAS measurements' spatial density and testing selected P-wave Differential Arrival Times (P-DATs) for source location. We estimate P-DATs for all the possible DAS channel pairs by identifying the time delay corresponding to the peak of each cross-correlation function. Subsequently, we select P-DATs based on two criteria: interchannel distance and cross-correlation index value. This procedure is often employed to reduce the risk of mixing delay times from coherent and incoherent waveforms. As a first test, using a probabilistic inversion (Hamiltonian Monte Carlo method), we demonstrate how the selected P-DATs provide a better constraint on the event's azimuthal direction compared to P-ARTs. Then, as a second experiment, we move from a subjective selection of P-DATs. To do so, we test a fully-automated and data-driven covariance matrix weighting procedure, in a probabilistic inversion scheme. Specifically, we compute posterior probability distributions for both the physical parameters (event location) and hyperparameters related to data features (interchannel distance and cross-correlation index thresholds). In this scheme, the hyperparameters define each weight along the diagonal of the covariance matrix. These tests offer useful insights into the utilization of P-DATs for event location with DAS. Moreover, we provide an automatic approach to avoid subjective biases based on pre‐conceptions in the a-priori data selection.

How to cite: Bozzi, E., Piana Agostinetti, N., Saccorotti, G., Fichtner, A., Gebraad, L., Kiers, T., and Nishimura, T.: Differential arrival times for source location with DAS arrays: tests on data selection and automatic weighting procedure, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3466, https://doi.org/10.5194/egusphere-egu24-3466, 2024.

Optical fiber networks, commonly known for data communications, could be extended beyond their conventional use. In our research, we propose a groundbreaking method by leveraging these existing terrestrial optical networks as wide distributed array sensors for earthquakes early detection. This approach is centered around the use of light polarization changes within the fiber cable, analyzed through a robust machine learning model that provide early warning alerts upon Primary earthquake waves arrival that induce strain, and precede earthquake’s destructive Surface waves. Unlike previous methods such as Distributed Acoustic Sensing and Frequency Interferometric Techniques, our approach avoids the use of expensive and specialized hardware. We introduce a centralized smart grid system that exploits the network’s existing terrestrial infrastructure, yet ensure cost effective and high efficient network modeling for initiating emergency plans and earthquake countermeasures. Our initial studies started by conducting experimental tests on a deployed fiber ring in Turin, Italy, using commercial Intensity Modulated – Direct Detection transceivers and polarimeters as polarization sensing devices, yield in promising results. Additionally, we identified the Primary waves arrival for a real 4.9 magnitude earthquake struck in the Marradi region, central Italy, with a 98% accuracy rate. This achievement was the result of a python-based waveplate model empowered by machine learning algorithm.  

Basically, when an earthquake occurs, networks nodes communicates with a centralized optical network controller that detects alterations in the light’s state of polarization by leveraging a pre-trained machine learning model. Upon the model confirmation, the controller activates early warning system in accordance with a predetermined emergency response mechanism. Building up on these findings, our current objective is to explore the impact of earthquake depth on seismic wave characteristics and their influence on light’s polarization to further investigate the potential of this advanced smart grid methodology. We aim to analyze real ground motion waves generated by two distinct earthquakes with same magnitudes but different depths. This knowledge is crucial in refining our machine learning model, which in turn will refine model prediction capabilities. Our approach promises more efficient optical network, by transforming the network into long range seismic sensors.

How to cite: Awad, H., Usmani, F., Virgillito, E., Bratovich, R., Straullu, S., Proietti, R., Pastorelli, R., and Curri, V.: Smart Grid Optical Fiber Network for Earthquake Early Warnings, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4418, https://doi.org/10.5194/egusphere-egu24-4418, 2024.

In recent years, there has been an increasing interest in Distributed Acoustic Sensing (DAS) technology for microseismic monitoring, especially in operations involving borehole installations. Despite the widespread adoption of DAS systems in such contexts, many questions regarding the data quality of the recordings are still open. Is the DAS self-noise higher than traditional systems? How does the ambient noise recorded by a DAS system attenuate with the depth as observed with traditional geophones? It is known that various noise types, including optical, thermal, and mechanical noise coupled with the fiber, affect DAS data. Additionally, the noise frequency band often overlaps with the signal frequency band, making frequency filtering alone inadequate for denoising. Therefore, specialized noise reduction workflows, such as FK Filtering and SVD, are necessary. Mitigating the impact of noise on DAS data remains a primary challenge for the seismological and geophysical community. This study aims to examine and characterize the noise influencing DAS data collected in borehole installations, with a specific focus on the data recorded at the Frontier Observatory for Research in Geothermal Energy site in Utah, USA. We use Power Spectral Density analysis to assess depth-dependent noise reduction and its temporal variations. Furthermore, the depth dependence of the signal-to-noise ratio for various microseismic events is evaluated. Finally, a comparison is drawn with geophones data colocated with the fiber, offering a comprehensive exploration of the advantages and disadvantages of the two data acquisition technologies.

How to cite: Pecci, D., Cesca, S., Rapagnani, G., Gaviano, S., Bocchini, G. M., Carelli, G., Stucchi, E., Iannelli, R., and Grigoli, F.: Noise analysis of Distributed Acoustic Sensing (DAS) systems in borehole installations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5900, https://doi.org/10.5194/egusphere-egu24-5900, 2024.

The use of optical fibre cables to sense ground motion is one of the most researched topics in seismology at present day. By using the technique of Distributed Acoustic Sensing (DAS), a single fibre can be turned into thousands of seismic sensors, providing unprecedented spatial resolution. The instrument response of optical fibre cables, however, is largely unknown and difficult to separate from source, path, and directivity effects on seismic records, preventing us from using the information from the full seismic waveform.

Here we present a full-waveform simulation scheme developed to model the DAS instrument response using a particle-based Elastic Lattice Model (ELM-DAS). The scheme allows us to simulate a virtual cable embedded in the medium and made of a string of connected particles. By measuring the strain along these particles, we are able to replicate the axial strain natively measured by DAS as well as the effects of irregular cable geometries. The scheme allows us to easily simulate complex properties of the material around the cable (e.g., unconsolidated sediments, nonlinear materials) as well as different degrees of cable-ground coupling, both of which are believed to be the key factors controlling the DAS instrument response.

By simulating DAS cables in 2D, we observe that at the meter scale, realistic DAS materials, cable-ground coupling, and the presence of unconsolidated trench materials around it dramatically affect wave propagation, each change affecting the synthetic DAS record, with differences exceeding at times the magnitude of the recorded signal. By expanding the scheme to 3D, we can accurately include the effects of realistic, complex–and at times sub-wavelength—cable geometries and how they influence DAS records. Our observations show that cable coupling and local site effects have to be considered both when designing a DAS deployment and analysing its data when either true or along-cable relative amplitudes are considered.

How to cite: Celli, N. L., Bean, C. J., and O'Brien, G. S.: Full-waveform modelling of coupling and site effects for DAS cables, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6263, https://doi.org/10.5194/egusphere-egu24-6263, 2024.

Distributed Acoustic Sensing (DAS) has become a revolutionary observational technology for different geophysical applications. DAS, known for its high spatial resolution, environmental resilience, and ease of deployment, which make it a potential replacement to the traditional physical sensors that have been used for decades in seismology. The primary distinction between DAS and conventional seismic sensors lies in the fact that DAS inherently captures strain (or strain rate), in contrast to seismic instruments which record translational ground motions. However, the problem of strain directional sensitivity poses challenges for its direct use in standard seismic analysis. Therefore, several physics-based methods have been proposed to convert DAS strain to ground motion response (displacement, velocity, or acceleration). Efficient conversion of strain to ground motion using physics-based methods relies on accurate estimation of phase velocity along the DAS cable which is unavailable in most cases. To overcome this problem, we introduce a novel deep learning (DL) approach to convert high-resolution Distributed Acoustic Sensing (DAS) strain measurements into ground motion (GM).  The DL model employs a Bidirectional Long Short-Term Memory (BiLSTM) network. The model is trained and evaluated utilizing data from the PoroTomo project at Brady Hot Springs Geothermal Natural Laboratory. This dataset includes earthquake waveforms recorded by collocated DAS channels and geophones. The model’s performance is evaluated using the Root Mean Squared Error (RMSE) metric. It demonstrated an average RMSE of 0.41 for training and 0.95 for testing, indicating the model's efficacy in transforming DAS strain to particle velocity. The comparison results of predicted and original geophone waveforms further validated the model's accuracy within the relevant frequency range. This study marks a significant advancement in adapting high-resolution DAS strain data for use with conventional seismic data analysis techniques, thereby expanding the capabilities of seismic monitoring and interpretation.

How to cite: Al-Qadasi, B. and Bin Waheed, U.: Deep Learning Driven DAS Strain Conversion to Geophone Ground Motion, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6576, https://doi.org/10.5194/egusphere-egu24-6576, 2024.

Distributed Acoustic Sensing (DAS) along seafloor fiber optic cables offers high-density and wide-aperture seismic data close to seismic sources, at a lower cost than conventional cabled ocean bottom seismic networks. It is thus a very promising approach to develop offshore monitoring systems for hazard mitigation and to obtain deeper insights into earthquake mechanics. We introduce a workflow for back-projection earthquake rupture imaging based on ocean bottom DAS data off the Chilean coast, taking full advantage of DAS data features to greatly refine the resolution and accuracy of source parameter estimation of local earthquakes. 

The workflow includes a number of steps that improve the back-projection performance. To reduce the negative effects of wave scattering on waveform coherence, we apply spatial integration to convert DAS strains into displacements. We refine travel time accuracy through shallow-sediment time corrections. We apply array processing on multiple overlapping cable segments (sub-arrays) to get the apparent slowness. The information from all sub-arrays is used jointly to locate the earthquakes using a 1D local velocity model.

Through systematic synthetic tests, utilizing the 120-km-long cable configuration off the coast of Chile, we identified a ‘high-precision, high-resolution source region”, which is also less sensitive to uncertainties of the velocity structure. This region extends to about 80 km laterally from the cable and reaches depths of up to 15 km, a range likely attributable to optimal signal focusing from various angles and that can be extended by increasing the cable length. We apply our method to data of roughly 50 local earthquakes with magnitudes from 1.5 to 3. We consistently obtain sharp back-projection images with high spatial accuracy, within 1 to 4 km, for earthquakes occurring within this defined region. Such precision is comparable to the location uncertainties of the seismic catalog.

The true strength of our approach is its potential for imaging the rupture process of larger earthquakes. We apply our method to the synthetic waveforms of a magnitude 7 earthquake constructed from multiple empirical Green's functions. We demonstrate that strong coda waves do not compromise the precise detection and location of subsequent sub-sources, if we apply a travel time calibration. The rupture speeds and locations of sub-sources are accurately recovered, even for concurrent multiple sources. We are currently improving the calibration of travel times to increase the location accuracy and resolution. These include waveform alignment with static calibration, 3D velocity model travel time tables, and slowness bias measurements and calibrations for each source-subarray pair. Collectively, these methods will increase the resolution and accuracy of our method, along with more sophisticated back-propagation methods for individual arrays. Our work holds promise for the development of earthquake and tsunami early warning, provided that we can effectively address the issue of amplitude saturation of DAS data.

How to cite: Xie, Y., Ampuero, J.-P., van den Ende, M., Trabattoni, A., Baillet, M., and Rivet, D.: Towards back-projection earthquake rupture imaging with ocean bottom distributed acoustic sensing data , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6610, https://doi.org/10.5194/egusphere-egu24-6610, 2024.

How to cite: Turquet, A., Wuestefeld, A., Nyhammer, F. K., Nilsen, E., and Refsum, V.: Advances in Avalanche Monitoring in Norway: Insights from Distributed Acoustic Sensing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7603, https://doi.org/10.5194/egusphere-egu24-7603, 2024.

Ambient noise tomography Derived from Distributed Acoustic Sensing (DAS) deployed on existing telecommunication networks provides an opportunity to image the urban subsurface at local to regional scales and high resolution effectively with a small footprint. This capability can contribute to the assessment of the urban subsurface's potential for sustainable and safe utilization in countless applications, such as geothermal development of an area. However, extracting coherent seismic signals from the DAS ambient wavefield in urban environments remains a challenge. One obstacle is the presence of complex noise sources in urban environments, which may not be homogeneously distributed. Consequently, long-duration recordings are required to calculate high-quality virtual shot gathers, which entails significant time and computational cost.

In this study, we present the analysis of 15 days of passive DAS data recorded on a pre-existing fiber optic cable (dark fibers) running along an 11~km long major road in urban Berlin (Germany). We identify anthropogenic activities, mainly traffic noise from vehicles and trains, as the dominant seismic source and use it for ambient noise interferometry. To retrieve Virtual Shot Gathers (VSGs), we apply interferometric analysis based on the cross-correlation approach. Before stacking, we designed a selection scheme to carefully identify high-quality VSGs, which optimizes the resultant stacked VSG . Moreover, we modify the conventional ambient noise interferometry workflow by incorporating a coherence-based enhancement approach designed for wavefield data recorded with large-N arrays. We then conduct Multichannel Analysis of Surface Waves (MASW) to retrieve 1D shear-wave velocity models of the subsurface along consecutive portions of the array and validate them against local lithologic models. Finally, a 2D velocity model of the subsurface is obtained by concatenation of individual 1D velocity models from overlapping array subsections. The expansion into 2D requires an automatic identification of high-quality VSGs based on unsupervised learning, such as clustering, to exclude transient incoherent noise in the process of selective stacking.

The clustering results reveal distinct groups of VSGs that exhibit similar patterns. These distinct groups provide valuable insights into the temporal variations in human activities and allow a better understanding and interpretation of the recorded DAS ambient noise data. We find that recordings obtained predominantly during rush hour are viable for further processing and improve the accuracy of dispersion measurements, in particular for traffic-induced noise data. Moreover, the resulting 1D velocity models correspond well with available lithographic information. The modified workflow yields improved dispersion spectra, particularly in the low-frequency band (< 1 Hz) of the signal. This improvement leads to an increased investigation depth along with lower uncertainties in the inversion result. Additionally, these enhanced results were achieved using significantly less data than required using conventional processing schemes, thus opening the opportunity for reduced acquisition times and efforts.

How to cite: Ehsaninezhad, L., Wollin, C., Rodríguez Tribaldos, V., Schwarz, B., and Krawczyk, C.: Enhancing 1D and 2D passive seismic imaging of urban ambient noise DAS recordings, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7791, https://doi.org/10.5194/egusphere-egu24-7791, 2024.

Seismic measurements record the superposition of many seismic sources, with anthropogenic ones dominating frequencies above 1 Hz. While the anthropogenic seismic vibrations in urban areas are too small to influence daily human life, measurements in high precision physics experiments, such as those carried out at the Deutsche Elektronen-Synchrotron (DESY) particle accelerators in Hamburg can be negatively influenced by these vibrations. To gain insight into the seismic wavefield at DESY, distributed acoustic sensing measurements were started in the WAVE initiative (www.wave-hamburg.eu). 

The goal of this study is to utilize unsupervised machine learning tools to detect and identify different anthropogenic seismic noise sources. Two different approaches were tested: the seismic measurements are clustered using a temporal average of one second on time-frequency representations and a deep embedded clustering technique. For the first method, the clustering methods fuzzy-c-means, Gaussian mixture model (GMM), hierarchical clustering and hierarchical density-based spatial clustering of applications with noise (HDBSCAN) were used. The clustering performance of all methods was compared using car signals on a short DAS fiber section as our ground truth data. Furthermore, the usage of spectrograms and continuous wavelet transforms was compared on the ground truth data set, with the continuous wavelet transform giving better results.

In a next step, the best-performing clustering methods GMM and HDBSCAN of the temporal average and deep embedded clustering were applied to the entire 12 km fiber to cluster seismic noise sources. Based on the results, the respective advantages and disadvantages of the different approaches were determined. The study was concluded with a "recipe'' on how to approach unseen DAS data based on scientific objectives and physical properties of interest, paving the way for an optimized DAS data analysis. 

How to cite: Kiel, A., Hadziioannou, C., and Hammer, C.: Exploring Unsupervised Clustering of Seismic Noise Sources in Urban DAS Data: A Methodology Guide, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8486, https://doi.org/10.5194/egusphere-egu24-8486, 2024.

Of the roughly 50.000 mines that were deployed in Danish waters during the First and Second World Wars, the Royal Danish Navy estimates that 4.000 to 6.000 units remain unexploded. Naval mines are to this day regularly found by fishermen or during surveys related to offshore construction work and reported to the Royal Danish Navy who then undertakes their controlled detonation. Seismic and hydroacoustic signals from naval mine explosions have been recorded by distributed acoustic sensing (DAS) on subsea fiber optic cabling where the hydroacoustic waves are readily identified. We have developed a simple technique that uses inversion of the travel time of hydroacoustic signals to determine the location of explosions. The technique has also been tested on hydroacoustic waves from a marine air gun seismic survey that crosses a fiber cable in shallow water monitored by DAS. We present the inversion results in addition to the data processing and analysis.

How to cite: Jensen, E. F., Damsgård, J. F., Voss, P. H., Funck, T., and Hansen, T. M.: Locating mine explosions in shallow waters from hydroacoustic waves using DAS, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9254, https://doi.org/10.5194/egusphere-egu24-9254, 2024.

Distributed Acoustic Sensing (DAS) is establishing as a promising technique in Seismology. This novel system turns a fibre optic cable into a continuous single component array with very dense spatial sampling. Simplicity of installation and availability of telecommunication cables (dark fibres) make the technique very advantageous for investigating harsh environments like seafloors, sensing up to hundreds of kilometres of fibre with fine spatial resolution. Given the high potential, the technology has been successfully tested in recent years for several earthquake monitoring tasks, such as location, subsurface characterization, focal mechanism determination, tomography, or source back projection. 
The transferability of standard seismological tools to DAS data is straightforward when working with time picking, while analysis of the amplitude content of the signal demands further research. This is the case of earthquake source characterization, where standard approaches require conversion of strain rate data into more classical kinematic quantities (i.e. acceleration or velocity). In this work we develop a new formulation that allows to estimate source parameters without the need for conversion. We start from the description of the far-field strain radiation emitted from a circular seismic rupture, showing that the time integral of the strain is related to the Source Time Function. Using this quantity, we develop the spectral modelling allowing for frequency domain inversion of DAS data for estimation of moment magnitude and corner frequency. The formulation accounts for the unique azimuthal sensitivity of the cable in the radiation pattern average, and explicitly shows DAS enhanced sensitivity to slow scattered waves propagating beneath the fibre.
We validated the proposed approach on two case-studies, for events in local magnitude range  0.4 - 4.3, comparing the results with estimates from standard seismic instruments. Earthquakes recorded on a 150km long cable offshore the coast of central Chile during a 1-month DAS survey exhibit scale invariant stress drops, with an average of Δσ=(0.8±0.6)MPa. Also, moment magnitude estimates agree with results from standard seismic instruments. The analysis of small magnitude events (M_L<2.5) recorded on a 1km long fiber during a 5-month DAS survey in the Italian southern Appenines shows agreement of moment magnitude estimates when compared with local seismic network estimations. Nevertheless, site effects dominate the high frequency domain resulting in an apparent corner frequency around 5Hz and masking the actual event size. An appropriate modelling of site effects using a parametric EGF approach was thus adopted to estimate corner frequencies for the highest magnitude events in the catalogue. 
This study shows the possibility to work with raw DAS data to retrieve information about earthquake source and highlights the high potential of these systems in characterizing the seismic moment and the size of earthquakes in a wide magnitude range.

How to cite: Strumia, C., Trabattoni, A., Supino, M., Baillet, M., Rivet, D., and Festa, G.: A new formulation for source parameters estimation from DAS native strain data., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9611, https://doi.org/10.5194/egusphere-egu24-9611, 2024.

The fast rise of Distributed Fiber Optic Sensing (DFOS, also known as DAS) technology in seismology has enabled to reach new horizons in volcano monitoring for example by its ability to attain hardly accessible environment and its high spatial and temporal resolution. Such advantages are extremely valuable for observatories located on islands where the ocean complicates the installation of traditional seismic networks and would require deploying ocean bottom seismometers.

In this research, we bring DFOS to a well-studied eruption that occurred in 2021 at La Palma (Canary Islands) by using a dark fiber, an unused telecom optic fiber, joining the islands together. The cable was interrogated using an HDAS (from Aragon Photonics) operated by INVOLCAN producing a 50 km-long array reaching outward from the island in the sea.

By using a combination of traditional seismic preprocessing and array detection methods such as CovSeisNet¹, we recover low frequency signals across the entire fiber. These steps enable us to detect and locate episodes of tremor linked to the volcanic activity which we compare with complementary observables.

¹ https://covseisnet.gricad-pages.univ-grenoble-alpes.fr/covseisnet/

How to cite: Fontaine, O., Caudron, C., Lecocq, T., D'Auria, L., and Barrancos, J.: Tremor analysis on dense network using Distributed Fiber Optic Sensing at La Palma, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9645, https://doi.org/10.5194/egusphere-egu24-9645, 2024.

Our recent theory shows that the 6C ground motion (three-component translation and three-component rotation) of ambient seismic noise is capable of measuring the local seismic anisotropy using azimuth-dependent 6C-based cross-correlation functions. However, seasonal variations in ambient seismic noise result in large uncertainties in local velocity measurements due to inaccurate corrections in the azimuth of wave propagation. Here, we show that the time-dependent small azimuth variation of ambient seismic noise can be visualized using horizontal rotation-based cross-correlation functions, which can be applied to constrain the local seismic anisotropy of Rayleigh waves. We apply this approach to a small seismic array (deployed to retrieve the rotational motions of seismic ambient noise) of Pinon Flat Observatory in Southern California. The estimated anisotropy is compatible with results calculated based on azimuth-dependent 6C cross-correlation functions from multiple pairs of stations, demonstrating the applicability of the proposed method.

How to cite: Tang, L., Igel, H., and Montagner, J.-P.: Constraining 6C-observed seismic anisotropy from seasonal ambient seismic noise, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10033, https://doi.org/10.5194/egusphere-egu24-10033, 2024.

Optical fiber sensing represents a promising technology for seismological monitoring, leveraging the widespread deployment of optical networks, and representing an important opportunity for the development of early warning systems. While so far Distributed Acoustic Sensing (DAS) has been widely employed in geosciences, this technology shows some limitations, like a restricted working range, requirement of dedicated fibers and criticalities in the management of big datasets. We focus on an alternative technique, coherent interferometry relying on ultrastable lasers, which is characterized by high sensitivity, long range, and full compatibility with the existing telecommunication infrastructure. The method allows detecting perturbations induced by seismic events through the measurement of the phase accumulated by an optical signal along the fiber path. The best performances are obtained employing narrow-linewidth lasers inherited from metrological applications due to their high coherence. While the technique was initially demonstrated on subsea cables, its application to on-land fibers poses new challenges. Indeed, the phase measurement integrates all the perturbations occurring along the fiber: this means that anthropic activities, such as vehicle traffic, represent important noise sources that must be taken into account. 

We present the details of an in-field implementation over a commercial fiber deployed in a highly seismic region in central Italy and connecting two populated towns. The experimental setup employs self-heterodyne interferometry detection, utilizing a continuous wave laser stabilized to an optical cavity through the Pound-Drever-Hall technique. The laser operates within a single channel of the Dense Wavelength Division Multiplexing (DWDM) grid, sharing the fiber with standard internet services. We show the results of continuous observations performed over a period of two years. We demonstrate the detection of about one hundred earthquakes, distinguishing them from typical noise sources such as acoustic interference and infrastructure oscillations. The results include the detection of both local and distant earthquakes, demonstrating the robustness of the technique. This allowed us to characterize for the first time the sensitivity curve of the technique, described by the probability of the event detection as a function of its magnitude and epicenter distance. We also show the correlation between the source magnitude and signal spectral analysis.

In conclusion, we present an operational fiber-based earthquake observatory, highlighting the compatibility of coherent interferometry with the existing telecommunication infrastructures and its effectiveness in seismic monitoring. The results are promising for the development of scalable sensing networks utilizing the extensive optical fiber infrastructure already in place, which can conveniently integrate in real-time the data acquired with the existing networks of classical seismological sensors.

How to cite: Donadello, S., Clivati, C., Govoni, A., Margheriti, L., Vassallo, M., Brenda, D., Hovsepyan, M., Bertacco, E., Concas, R., Levi, F., Mura, A., Herrero, A., Carpentieri, F., and Calonico, D.: An Earthquake Observatory based on Coherent Interferometry over the Optical Fiber Network in Italy, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10120, https://doi.org/10.5194/egusphere-egu24-10120, 2024.

The power of distributed acoustic sensing (DAS) lies in its ability to sample deformation signals along an optical fiber at hundreds of locations with one interrogator only. While the interrogator is calibrated to record ‘fiber strain’, the properties of the cable and its coupling to the rock control the ‘strain transfer rate’ and hence how much of ‘rock strain’ is represented in the recorded signal.

We use DAS recordings carried out with a Febus A1-R interrogator in an underground installation colocated with an array of strainmeters in order to measure the ‘strain transfer rate’ in situ. A tight-buffered cable and a standard loose-tube telecommunication cable (running in parallel) are used, where a section of both cables covered by sand and sandbags is compared to a section, where cables are just unreeled on the floor.

Signals from the Mw 7.7 and Mw 7.6 earthquakes that took place on the East Anatolian Fault on February 6th 2023 allow us a proper comparison of signals in the frequency-band between 50 mHz and 0.2 Hz. At lower frequencies the DAS signal-to-noise ratio is insufficient. At higher frequencies the invar-wire strainmeters show a parasitic response to vertical ground motion. For frequencies up to 1 Hz we use seismometer recordings to estimate strain for an incoming plane wave, based on the ray parameter and in this way extend the bandwidth of the comparison. The ray parameter varies along the recording but is sufficiently well known and can be validated against the strainmeter recording.

The ‘strain transfer rate’ is largely independent of frequency in the band from 0.05 Hz to 1 Hz and varies between 0.15 and 0.55 depending on cable and installation type. The sandbags show no obvious effect and the tight-buffered cable generally provides a larger ‘strain transfer rate’. The noise background for ‘rock strain’ in the investigated band is found at about an rms-amplitude of 0.1 nstrain in 1/6 decade for the tight-buffered cable. This allows a detection of the marine microseisms at times of high microseism amplitude.

How to cite: Forbriger, T., Karamzadeh, N., Azzola, J., Widmer-Schnidrig, R., Gaucher, E., and Rietbrock, A.: On DAS-recorded strain amplitude, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10604, https://doi.org/10.5194/egusphere-egu24-10604, 2024.

The integration of state-of-the-art fibre optic sensing technologies with telecommunication systems has not yet been achieved. There are many challenges which need to be overcome to allow for a pan-continental research instrument, all of which requires international collaboration. Such international collaborations would allow for the creation of novel applications and research into Earth science, such as cetology and abiotic and biotic marine interactions, oceanography, seismology, volcanology, and soundscape monitoring, to name but a few.

Over the past 5 years, research teams from National seismic and oceanographic infrastructures, together with National Research and Education Networks (NRENs), and partners from universities, research institutes and industry have pioneered sensing techniques to use submarine optical telecommunication fibres to monitor the Earth and its systems.

Fibre sensing technology and collaborations created by developing these techniques have now reached a level where a new paradigm shift can occur. This presentation will discuss the SUBMERSE project (SUBMarinE cables for ReSearch and Exploration), which is creating and delivering a pilot research instrument which could serve as a blueprint for continuous monitoring upon many more existing submarine fibre optic cables in the future.

The SUBMERSE project, which started in May 2023, is a Horizon Europe-funded, 36-month long initiative which is investigating the combined acquisition of SOP (State-of-Polarisation) and DAS (Distributed Acoustic Sensing) data from live telecom cables, with the aim to then make that data available to researchers globally and F.A.I.R. The project team, consisting of 24 organisations, uses existing fibre-optic infrastructure deployed across multiple national research infrastructures to create a pan-European research instrument.

Our presentation will discuss the latest field deployments of DAS and SOP technologies across 5 geographic locations on 3 cable systems, which are spread across the European continent and Atlantic Ocean. It will offer a first glimpse of the effects of running a DAS in the L-and C- Bands on a live DWDM telecoms system, in combination with SOP, in a submarine and terrestrial environment.

The aim of the SUBMERSE project is to disseminate the data following FAIR principles through established community data centres. The main challenges we have faced relate to ensuring compliance to security restrictions and handling huge data quantities generated by DAS.  The approaches to down sampling, frequency filtering, and potentially time-and-space-gating will also be presented.  We will discuss the approaches taken for acquiring, streaming from remote sites, data staging, pre-processing and raw file retention. We will also highlight the tools and approaches that we have adopted to develop best practices for running such a multi-national, distributed, sensing instrument which must take these elements into account.

Our work has shown that a pragmatic approach, with collaboration at heart, is needed to address these challenges. Without a strong commitment and collaboration between research communities and research infrastructure providers, the potential to lose valuable research data is high. This risk can be mitigated by focusing on datasets which are valuable to communities and ensuring the long-term availability of those data sets in a F.A.I.R manner.

How to cite: Atherton, C. and the SUBMERSE Project: SUBMERSE: Exploring the planet with live submarine telecommunications cables., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10764, https://doi.org/10.5194/egusphere-egu24-10764, 2024.

The notable benefits of Distributed Acoustic Sensing (DAS) technology—high coverage, high resolution, low cost—have led to its widespread application in the geophysical domain for high-quality data recording. Among possible applications, the ability to interrogate telecommunication cables has enabled the detection of a variety of seismic-volcanic events in poorly instrumented environments, such as densely populated urban areas.

Nevertheless, the sensing of commercial fiber optic cables has to deal with the presence of anthropogenic noise that frequently corrupts the seismic signal. Indeed, vibrations induced directly or indirectly by anthropogenic activities significantly reduce the signal-to-noise ratio by masking target events.

Taking advantage of the high spatiotemporal resolution of the DAS data, a deep learning approach has been adopted for noise removal. The architecture of the neural network together with the training strategy have enabled the extraction and preservation of salient information while neglecting anthropogenic noise.

The validation on real low-frequency seismic events recorded during the 2021 Vulcano Island unrest  has provided encouraging results, demonstrating the potential of the proposed approach as a pre-processing step to facilitate subsequent DAS signal analysis.

How to cite: Allegra, M., Cannavo, F., Corsaro, M., Currenti, G., Jousset, P., Palazzo, S., Prestifilippo, M., and Spampinato, C.: Denoising DAS data in urban volcanic areas through a Deep Learning Approach, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10925, https://doi.org/10.5194/egusphere-egu24-10925, 2024.

Microseismic source processes can be closely monitored during hydraulic stimulations with optical fiber deployed behind borehole casing, using Distributed Acoustic Sensing (DAS). The Bedretto Underground Laboratory for the Geosciences and Geoenergies (BULGG) provides a test site at the scale of hundreds of meters (meso-scale), where multiple boreholes are instrumented with fibers around a stimulation well. This enables the characterization of source properties of induced seismicity thanks to the dense sampling of the wavefield close to the stimulated region.

In 2023 various stimulation activities in the BedrettoLab produced M<-1 events that were recorded on three fibers surrounding the stimulated region. The interrogated fibers are running through the stimulated seismicity zone, and surround the majority of the events. Some of these events are recorded with high coherency and signal-to-noise ratio, making them suitable for further source characterization, such as location and moment tensor inversion. These events were at the same time recorded with other co-located point sensors such as geophones and acoustic emission sensors, which enables comparison to other instruments.

In this work, we select and process a subset of events with clear DAS recordings, and invert for their location, source time and moment tensor components using an adjoint inversion method. This includes computing the full forward and adjoint wavefield and gradient using a spectral-element solver. Using the full waveforms to invert for these events greatly improves the resolution of the source estimates, allows for incorporation of the full velocity model, and only two simulations per iteration are required: a forward and adjoint simulation, and gradient computation. The receiver coverage of the focal spheres by the surrounding optical fibers is an excellent test bed for the method, and the simulated domain remains on the order of hundreds of meters, which means that the simulation can be pushed to high frequencies (>100 Hz). This study provides a step forward to monitoring microseismicity in hydraulic stimulations with fiber-optic measurements.

How to cite: Tuinstra, K., Lanza, F., Noe, S., Fichtner, A., Rinaldi, A. P., Edme, P., Rosskopf, M., Obermann, A., Hertrich, M., Maurer, H., Giardini, D., and Wiemer, S. and the Bedretto Team: Adjoint-Source Inversion of Microseismic Sources with DAS in Boreholes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11389, https://doi.org/10.5194/egusphere-egu24-11389, 2024.

Seismic interferometry applied to Distributed Acoustic Sensing (DAS) arrays is an increasingly common approach for subsurface investigations. In this study, we show that analysis of urban seismic noise acquired on a linear underwater DAS array can be used to track depth-dependent seismic velocity variations caused by groundwater level changes in an alluvial aquifer.

We apply our methodology to the Crépieux-Charmy wellfield, a strategic site for the water supply of the Lyon metropolitan area in France. We analyze 4 weeks of ambient noise recorded during a water infiltration experiment along a 200m DAS cable placed at the bottom of an infiltration basin.

Using ambient noise interferometry, we derived time-lapse phase velocity variations (dc(f)/c(f)) of Rayleigh waves and inverted them for depth-dependent shear wave velocity variations (dVs(z)/Vs(z)) in the first 50 m of the subsurface. The obtained seismic velocity changes appear to be associated with variations in water saturation and effective pore pressure for different compartments of the aquifer.

Our results suggest that DAS combined with noise-based passive monitoring provides a solution to track the dynamics of an alluvial aquifer and estimate hydrological parameters relevant for effective groundwater resource management.

How to cite: Nziengui Bâ, D., Mordret, A., Coutant, O., and Jestin, C.: Groundwater monitoring in an alluvial aquifer with an underwater DAS cable recording urban seismic noise: Application to the Crépieux-Charmy Wellfield in France, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12344, https://doi.org/10.5194/egusphere-egu24-12344, 2024.

A high-sensitive, large-scale ring laser gyroscope provides access to direct observations of local rotational ground motions. A tetrahedral configuration of ring laser gyroscopes, such as ROMY (ROtational Motions in seismologY), located in a Geophysical Observatory near Munich, Germany, enables to redundantly observe all three components of the rotation vector.
For seismic accelerations below 30 mHz, the separation of low noise background levels between vertical and horizontal component are well established and understood to result from local tilts driven by atmospheric pressure variations. The promise of multi-component rotational observations is that ideally they can be used to decontaminate a co-located horizontal component acceleration sensor from contributions of ground tilt. Moreover, knowing and understanding the background levels for vertical and horizontal rotational ground motions at long periods is essential as benchmarks for instrument development towards higher sensitivity.
We use several months of multi-component data of vertical and horizontal rotation rates by ROMY and a co-located atmospheric pressure sensor to derive the pressure compliance for both vertical and horizontal rotational motions. Focusing on frequencies below 20 mHz, we find that time windows with energetic weather patterns consistently lead to high coherence of atmospheric pressure and horizontal rotations, but only little coherence between the atmospheric pressure and vertical rotation.
We consider this as a first indication that atmospheric pressure induced ground tilts are detected by the ROMY horizontal components. Different effects of ambient atmospheric pressure changes on the optical gyroscope itself, such as cavity deformation, are discussed. A small aperture barometer array surrounding ROMY to detect lateral pressure gradients is currently being deployed to provide additional constraints on ground deformations from atmospheric pressure waves.

Here we focus on a detailed analysis of ROMY gyroscope data, while accelerometer data are analyzed in a companion poster by Widmer-Schnidrig et al. in this same session SM3.3.

How to cite: Brotzer, A., Widmer-Schnidrig, R., and Igel, H.: On the influence of ambient atmospheric pressure on multi-component, direct observations of rotational ground motion, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12679, https://doi.org/10.5194/egusphere-egu24-12679, 2024.

Monitoring of seismic activity around volcanoes has been conventionally performed using data from continuous seismic and deformation networks, which give real-time information on the status of a volcano at any time. In case of a volcanic crisis, the number of earthquakes often increases with time and conventional networks are completed by deployment of additional sensors, which allow for a better hazard assessment, e.g., by lowering the detection threshold and improving earthquake locations. The deployment of such additional sensors is labour intensive and may be dangerous due to increased volcanic hazard.

Existing fibre optic telecommunication cables can be used with distributed dynamic strain sensing interrogators to density and complement the monitoring network. It has been demonstrated that the usage of fibre optic sensing allows for a rapid response and the acquisition of crucial data describing a developing crisis (e.g., at Vulcano, Italy). However, fibre optic interrogators are rarely deployed as permanent interrogating systems, despite the capability of such systems for long-term monitoring as demonstrated during a 7 months continuous recording on the Reykjanes Peninsula, Iceland, in 2020. Instead, interrogators are usually deployed for limited time periods when the activity occurs, ideally before new activity starts. For example, we connected an iDAS interrogator on the telecommunication 16-km long cable running between the Reykjanes and Svartsengi (“Blue Lagoon”) geothermal power plants in 2015 for an initial test of 10 days, in 2020 for 7 months (GFZ rapid response to the seismic crisis and precursory activity to the 2021 and 2022 eruptions of Fagradalsfjall volcano), and in November 2023 (recording still on 10.01.2024) as a GFZ rapid response before the 18 december 2023 eruption.

In this work, we investigate the possibility to use repetitive campaign-based measurements of dynamic strain sensing performed in the course of multiple years on the Reykjanes Peninsula (2015; 2020; 2023-2024) and at Etna volcano (2018; 2021; 2022; 2023-2024). Analysing earthquakes and ambient noise, we search for differences and similarities in the strain-rate response between the different and disjunct recording periods. We report preliminary results.

How to cite: Jousset, P., Hersir, G., Currenti, G., Wollin, C., Díaz-Meza, S., Allegra, M., Maass, R., Prestifilippo, M., Gudnason, E., Napoli, R., and Krawczyk, C.:  Long-term monitoring of seismic and volcanic activity using distributed fibre optic sensing: examples in Iceland (2015-2024) and Italy (2018-2024)., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13091, https://doi.org/10.5194/egusphere-egu24-13091, 2024.

Distributed Acoustic Sensing (DAS) is a technology that enables continuous, realtime measurements along the entire length of a fiber optic cable. The low-frequency band of DAS can be used to analyze hydraulic fracture geometry and growth. In this study, the low-frequency strain waterfall plots with their corresponding pumping curves were analyzed to obtain information on fracture azimuth, propagation speed, number of fractures created in each stage, and re-stimulation of pre-existing fractures. We also use a simple geomechanical model, described in full detail in Ortega Perez and Van der Baan (Geophysics, 2024), to predict fracture growth rates while accounting for changes in treatment parameters. As expected, the hydraulic fractures principally propagate perpendicular to the treated well, that is, parallel to the direction of maximum horizontal stress. During many stages, multiple frac hits are visible indicating that multiple parallel fractures are created and/or re-opened. Secondary fractures deviate towards the heel of the well, likely due to the cumulative stress shadow caused by previous and current stages. The presence of heart-shaped tips reveals that some stress and/or material barrier is overcome by the hydraulic fracture. The lobes of the heart are best explained by the shear stresses at 45-degree angles from the fracture tip instead of the tensile stresses directly ahead of the tip. Antennas ahead of the fracture hits indicate the re-opening of pre-existing fractures. Tails in the waterfall plots provide information on the continued opening, closing, and interaction of the hydraulic fractures within the fracture domain and stage domain corridors. Analysis of the low-frequency DAS plots thus provides in-depth insights into the rock deformation and rock-fluid interaction processes occurring close to the observation well.

How to cite: van der Baan, M. and Ortega Perez, A.: Interpretation of low-frequency DAS data acquired during hydraulic fracturing treatments based on geomechanical models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13428, https://doi.org/10.5194/egusphere-egu24-13428, 2024.

Long-term Distributed Acoustic Sensing (DAS) using a submarine cable is being conducted in the Nankai Trough, Japan. The first long-term DAS observation shows that the apparent strain is observed along the relatively shallow water depth section (i.e., < water depth 1000 m) on the submarine cable, suggesting a periodic temperature fluctuation up to about 6 degrees Celsius associated with the ocean tide takes place in this region (Ide et al., 2021). However, it is difficult to discuss how the ambient temperature changes, or how the ambient temperature affects the submarine cabled DAS observation because of the lack of the in-situ observation. For this reason, we installed temperature sensors near the submarine cable by a Remotely Operation Vehicle (ROV). We did not discover the submarine cable at the seafloor, and therefore the detailed location of the temperature sensors, i.e., the accurate cable position could not be constraint so far. Simultaneous observations of the DAS and the long-term ambient temperature were conducted for a period from 16 August 2021 to 04 October 2021, i.e. about 50 days. A cross-correlation analysis between the DAS and the ambient temperature by dividing into the 5-day dataset has speculated the position of the temperature sensors to be 24.75 km of the submarine cable. Additionally, the strain coefficient of the optical fiber w.r.t. the temperature change has been determined to be 7 micro-strain per 1 degree Celsius, which is comparable to the previous experimental study (Zumberge et al., 2018). Finally, the temperature correction was performed, but the phase delay still remains, suggesting that the thermal measurement should be conducted beside the submarine cable.

How to cite: Matsumoto, H., Araki, E., and Yokobiki, T.: Effect of ambient temperature fluctuation on DAS observation with a submarine cable, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13660, https://doi.org/10.5194/egusphere-egu24-13660, 2024.

Since almost two decades, there is a fast and steady progress in understanding the rotational part of the seismic wavefield and exploring possible applications. These achievements are based on studies using simulated data, array-derived measurements, and direct measurements of large (M > 5), teleseismic earthquakes by ringlasers. Since only a couple of years, direct measurements of smaller (M < 3) earthquakes in the local distance range are also feasible. This was made possible due to new instrumentation developments such as portable rotation sensors.

From experience with translational measurements, seismology has developed a relatively simple description of the seismic wavefield, as long as the observation is recorded in the source far-field, and site-effects at the point of observation can be neglected by choosing an appropriate frequency range for the analysis.

Within the NonDC-BoVo project two portable rotational sensors have been installed in the Vogtland/West-Bohemia earthquake swarm region with the goal to incorporate the rotational waveform data into the inversion for seismic moment tensors. Both sensors are located in an epicentral distance of ~10 km to the center of the swarm activity. So far, we have recorded 197 events with ML ≥ 1 and 6 events with ML ≥ 2.5.

Although we are positively surprised how well we can record rotational ground motion of earthquakes with even very small magnitudes, we encountered challenges in the details of the waveform recordings. At the sensor location in Landwüst (D) we recorded events down to ML ~ 0.5 with good signal-to-noise ratio in a frequency range of 5 to 25 Hz. At the second location in Skalna (CZ) the signal-to-noise ratio is worse and we recorded earthquakes only with ML ≥ 1.5. Relocating the sensor to Wernitzgrün (D), about 25 km to the North of Skalna, did not improve the quality of the waveform recordings. Technical issues with the sensor can be ruled out for both locations.

Here, we want to present details of the challenges with the rotational ground motion data from these small and local earthquake recordings. First analyses hint to a much stronger effect of local site conditions onto rotational than translational ground motion data. In addition, with the above-mentioned setting, we have to consider the complexities of the near-field part of the wavefield as well. With our contribution we aim to add another aspect to the understanding of the rotational wavefield.

How to cite: Donner, S., Lehr, J., Krüger, F., Hoffmann, M., Hobiger, M., and Heimann, S.: Challenges of Rotational Ground Motion Measurements in the Local Distance Range, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14922, https://doi.org/10.5194/egusphere-egu24-14922, 2024.

For a decade, Distributed Acoustic Sensing (DAS) has become an exciting new tool for research in geophysics. However, the vast amount of data produced by hundreds of virtual sensors presents a challenge, and the lack of labeled data hampers the training of new automated classifiers based on machine learning methods from scratch. Automatic classifiers are essential tools for labeling observations across large, complex, and continuous datasets. They are commonly used in seismology to classify all the geophysical events recorded by seismic stations with a high degree of accuracy. Transfer learning is a machine learning technique where a model trained on one task is repurposed on a second related task. This method is particularly beneficial as it allows for the reuse of pre-existing, labeled datasets, significantly reducing the need for new data collection and annotation. This approach is a promising way to create efficient DAS classifiers without requiring a lot of resources. We built three different classifiers: the first classifies events based on their location, distinguishing local earthquakes from distant events; the second differentiates anthropogenic events from natural ones, specifically quarry blasts from earthquakes; and the third is capable of identifying all three types of events - local earthquakes, distant events, and anthropogenic quarry blasts - simultaneously. We used random forest models to train our classifiers using a labeled dataset of 14345 seismometer signals from the NOA network made for seismological research and nuclear monitoring. We studied the transferability of those classifiers to DAS data from a new NORSAR facility called NORFOX. This infrastructure consists of 8 km of cables located in Norway with a sampling frequency of 100Hz. Its multidirectional cable configuration makes it ideal for seismological surveys. We built a small test dataset of 543 signals using 20 channels equally distributed along the array. Our approach succeeded in reaching F1 score thresholds above 76%. We also showed, by fine-tuning feature selection according to their correlations between the seismic and DAS domains, that these results can be improved. In this way, we were able to increase the scores of our models from 5.89% to 11.36%. Several transfer learning approaches were also explored and discussed. Those encouraging results highlight the potential of a transfer learning approach to build new DAS classifiers. By leveraging historical seismometer catalogs, this approach facilitates the creation of DAS classifiers, thereby saving substantial time and avoiding the need for creating specialized datasets exclusively for DAS. Several strategies for improving score thresholds for future classifiers based on these transfer learning methods can be derived.

How to cite: Donnadille, M., Turquet, A., Hibert, C., and Richard, C.: Distributed Acoustic Sensing automated classifiers design via transfer learning for seismology, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15284, https://doi.org/10.5194/egusphere-egu24-15284, 2024.

Distributed Acoustic Sensing (DAS) has become increasingly popular due to its capacity of extremely high spatial and temporal strain sensing over long distances. Due to the long reach of the current fiber sensing technology, and that all the electronic components are placed on land, it holds great potential in marine applications. This presentation shows how DAS can be used to observe ocean surface waves inside Kongsfjord and in the open ocean West of Svalbard.

A well-known problem when monitoring ocean phenomena is the sensor undersampling in the world’s oceans. This problem limits observations of fundamental scientific questions, like the dynamics of the oceans (Sladen et al., 2019). Fiber optic sensing has already been used to observe numerous ocean phenomena, such as tides, OSGW (Ocean-Surface-Gravity-Wave), and internal waves (Lindsey et al., 2019, Ide et al, 2021, Williams et al., 2023).

A comprehensive DAS data set is being collected at the Centre for Geophysical Forecasting’s (CGF) research lab in Ny-Ålesund, Svalbard. Data has been collected continuously since June 2022 providing the possibility of investigating oceanographic signals over a long time period in different marine environments. In this presentation, we focus on how ocean surface wave signals are influenced by local wind conditions and how the signal changes characteristics as a function of wind speed and direction.

The results from analyzing this data will be presented, and it will be demonstrated how the local wind-induced waves interact with the background swell signal which hit West Svalbard from a South-West direction. The difference between the portion of the fiber located inside Kongsfjorden will be compared to the portions in open ocean. Additionally, the obtained results will be compared to previous studies from the area (e.g., Wojtysiak et al., 2018). Finally, we use the well-known dispersion relation for OSGW to compute the associated phase velocity along the whole stretch of the fiber cable.

How to cite: Rørstadbotnen, R. A. and Landrø, M.: On the variation of ocean surface waves with wind speed and direction: A case study from offshore Svalbard, Norway, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15616, https://doi.org/10.5194/egusphere-egu24-15616, 2024.

Distributed Acoustic Sensing (DAS) has effectively transformed traditional telecommunication fiber-optic cables into highly efficient, dense seismic arrays. In this paper, we perform cross-correlation and stacking on the signals recorded by DAS fiber arrays in a Zig-Zag pattern, which increases the spatial coverage and improves the ability to detect and analyze various seismic waves modes. We observed that the resulting signals include both Rayleigh and Love waves. Additionally, the propagation characteristics of the wave field exhibit a Zig-Zag distribution pattern, consistent with the spatial distribution of the fibers. The challenge arises from the need to distinguish and accurately interpret the overlapping signals of Rayleigh and Love waves, which have different propagation characteristics. To address these challenges, we propose an inversion method specifically tailored for the near surface imaging with DAS array data when the fiber-optic cables are not laid in straight lines. We also conducted an on-site experiment using Zig-Zag shaped DAS arrays with known subsurface velocity model. This experiment was designed to record ambient noise signals over a week and analyze the propagation characteristics of both Rayleigh and Love waves within the surface waves captured by the DAS system. The inversion results obtained from the analysis of the recorded data showed a high degree of consistency with the ground truth subsurface structure of the test area. This demonstrates the effectiveness of the proposed method in overcoming the limitations of traditional dispersion curve-based inversion techniques, particularly in the context of non-linear fiber layouts. This research provides strong support for the practical application of dark fibers in near surface imaging, demonstrating the potential of dark fibers.

How to cite: Wu, P., Gu, C., Zhong, Y., Yuan, Z., Chen, Z., and Kang, B.: Near Surface Imaging with Zig-Zag shaped DAS Arrays Based on Cross Correlation of Ambient Noise, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16151, https://doi.org/10.5194/egusphere-egu24-16151, 2024.

Distributed acoustic sensing (DAS) on submarine fiber optic cables will contribute to resilient societies by significantly enhancing environment and hazard monitoring. Recent studies have emphasized applications to earthquake and tsunami early warning, volcanic eruptive events observation, and characterizing climate change effects. Widespread deployment of DAS instrumentation on the existing cable networks traversing the coastal zones and oceans can vastly expand data coverage with real-time capabilities at low cost. 

The use of DAS in existing long-haul telecommunication systems has so far been limited since most DAS interrogators rely on optical wavelengths that interfere with the existing telecom traffic. Recent developments in DAS technology enable co-existence of telecom traffic with the sensing application. Lab and field investigations have demonstrated that there is no detrimental effect from the interrogation on the transmission line performance when the DAS instrument is configured to operate at sufficiently separated wavelength channels. Thus, it is possible to utilize all existing cable networks for sensing applications in unison with continued telecom transmission without sacrificing any capacity in the link and maintaining high-quality DAS measurements.

This study describes a DAS deployment on a submarine cable system with live telecom traffic. The interrogation scheme facilitates consistent high sensing sensitivity exceeding 120 km range. The data quality and interrogator performance are quantified, and the localization and characterization for a representative set of environmental and anthropic signal sources is shown. We will describe real-time processing and detection implementations that transform such sensing enabled submarine telecom networks into measurement arrays suitable for seismic monitoring. Moreover, the technical solutions for DAS installation with existing terminal equipment and the practical aspects of data sharing will be described.

How to cite: Morten, J. P., Wienecke, S., Waagaard, O. H., Brenne, J. K., and Rønnekleiv, E.: Sensing enabled submarine telecom networks for seismology, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16375, https://doi.org/10.5194/egusphere-egu24-16375, 2024.

Since 2005 the Campi Flegrei caldera (Southern Italy) has been experiencing a long-term unrest characterized by a recent increase in seismicity and ground uplift rate. In August and September 2023, the unrest showed a sudden intensification in the number and maximum magnitude of earthquakes, culminated with the occurrence of a Md 4.2 event. In an effort to strengthen the monitoring activity, two Distributed Acoustic Sensing (DAS) devices were connected from October 2023 to telecommunication fiber optic cables in the densely populated Campi Flegrei area. The DAS interrogators are set up inside two TELECOM central offices in the city center of Naples and in Bagnoli. Dynamic strain rate data are continuously acquired with a gauge length of 10 m at an average spatial sampling of 4 and 5 m. In this framework, an automated real-time workflow has been implemented comprising both DAS data downsampling and transferring to INGV data processing center. 

From the beginning of the DAS acquisition (19 October 2023), more than 300 seismic events have been recorded by the INGV local seismic network, including a Md 3.0 earthquake occurred on 23rd November 2023. Unknown cable installation conditions, poor coupling of the fiber optic cable with the ground, intense traffic and anthropogenic activities make the DAS data highly noisy and, hence, pose challenges for the application of traditional seismic data processing algorithms. In this study, we propose and apply AI based algorithms to improve and fasten earthquake detection and seismic phase picking on the large data volume associated with the high number of DAS channels.  In particular, the compared algorithms cover recently published state-of-the-art deep learning networks. We show preliminary results that demonstrate the ability of the synergy between DAS and AI to contribute to the rapid response to volcanic crises.

How to cite: Corsaro, M., Allegra, M., Cannavò, F., Currenti, G., Festa, G., Jousset, P., Pellegrino, D., Palazzo, S., Prestifilippo, M., Privitera, E., Pulvirenti, M., Ricciolino, P., Spampinato, C., Scotto di Uccio, F., and Tramelli, A.: Distributed Fiber Optic Sensing and Artificial Intelligence: preliminary results on the Campi Flegrei caldera unrest, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16414, https://doi.org/10.5194/egusphere-egu24-16414, 2024.

Urban sustainable development and improved resilience to geohazards requires an exhaustive understanding of the geological structure, physical properties and dynamics of the shallow subsurface underneath urbanized areas at the sub-kilometer scale. Yet, our current understanding of the urban subsurface is limited by our ability to image its structure and temporal variations at high resolution using classical geophysical approaches. Recently, the application of conventional ambient noise interferometry analysis to dynamic strain data recorded using Distributed Acoustic Sensing (DAS) deployed on unused telecommunication fiber-optic cables (dark fibers) has emerged as an attractive alternative for cost-efficient, regional scale (10’s of km) seismic imaging and monitoring at high spatial and temporal resolution. Still, its application to urban environments remains vastly underutilized. One of the most significant hurdles is the lack of adequate and efficient data exploration and processing tools to address and harness the unique challenges associated with DAS-based urban seismic noise recordings, which include the complexity of the noise field, unconventional array geometries and non-uniform coupling conditions, and increasingly massive data volumes.

The InDySE project (Interrogating the Dynamic Shallow Earth) aims at addressing these challenges to develop and validate the next-generation of subsurface imaging and monitoring platforms in urban areas based on the combination of existing fiber-optic networks and high-frequency infrastructure noise. The project comprises (1) developing high-performance computational tools, advanced processing workflows and machine learning approaches for efficient data exploration, selection and processing using existing datasets, (2) field experiments in target areas to retrieve high-resolution velocity models and monitor changes in seismic velocities, (3) integrating the resultant high-resolution seismic models with complementary datasets such as deformation maps derived from InSAR measurements. One of the selected study areas is the highly populated metropolitan area of Istanbul (Turkey), where understanding the structure, properties and dynamics of the shallow subsurface at high-resolution is critical for evaluating geohazard exposure. Among our objectives will be illuminating potential hidden faults underneath the city, obtaining high-resolution maps of geological materials and subsurface properties that can be translated into maps of local site response to large earthquakes, and tracking seismic velocity changes linked to hydrological dynamics that are known to be responsible for ground movements such as landsliding and subsidence. Ultimately, InDySE aims at developing efficient approaches for using dark fiber and ambient noise in urban subsurface investigations with implications in geohazard assessment.

How to cite: Rodríguez Tribaldos, V., Krawczyk, C., Ehsaninezhad, L., Martínez-Garzón, P., and Bohnhoff, M.: Towards Fiber-optics-based, Next-generation Observational Platforms for Investigating the Urban Subsurface: the InDySE Project, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17020, https://doi.org/10.5194/egusphere-egu24-17020, 2024.

Distributed Acoustic Sensing (DAS) exploits Rayleigh light backscattering to extract images of seismic wave propagation along a fiber optic in time and distance. The spatial distribution of virtual point sensors represents an opportunity to develop innovative methods for seismic event sources detection and identification. We develop in this study a method based on Machine Learning solutions for events classification.

This method relies on the development of features which translate the characteristics of the signals we observe into quantities that can be processed by machine learning algorithms to achieve the source classification. Three families of features investigating temporal and spatial characteristics and similarity of the signal are proposed, such as spatial and temporal analysis of the standard deviation, kurtosis or skewness of the signal or cross-correlation and dynamic time warping characterization and enables to quantify their individual contribution. Then we use a supervised machine learning model named XGBoost to perform classification based on these developed features. We tested this approach with a dataset recorded along a 91 km-long fiber optic deployed in the Pyrenees in France. The data acquisition has been achieved using a FEBUS A1-R DAS interrogator and with the support of TotalEnergies, from August 30 to September 20, 2022. During this period, 11 earthquakes and 6 quarry blasts have been recorded.

The trained model is validated using cross-validation techniques. Our Machine Learning processing chain successfully detect and classify 13 regional events from continuous background noise made by natural and anthropogenic activities. In particular, spatial features help to reduce the contribution of moving vehicles, whose presence is unavoidable along existing long-distance telecommunication fiber sections installed alongside roads.  In the continuity of this study, we investigate the potential of transfer learning from geophones deployed along the studied cable to DAS data or to another fiber optic cable installed in the same area.

How to cite: Huynh, C., Hibert, C., Jestin, C., Malet, J.-P., and Lanticq, V.: A complete feature set for classification of seismic sources with Distributed Acoustic Sensing (DAS) in the context of long-range monitoring, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17284, https://doi.org/10.5194/egusphere-egu24-17284, 2024.

FORGE (Frontier Observatory for Research in Geothermal Energy) is an underground field laboratory located in Utah, western United States, and sponsored by the US Department of Energy. The FORGE site is situated above a young, hot granitoid formation. The site has been used as a testbed for institutions interested in enhanced geothermal systems. As part of recent research work conducted at the site, a new deviated well was drilled adjacent to a previously stimulated well. Both wells were drilled into the granite volume. Circulation tests were performed over a period of approximately two days following completion of the second well. The circulation tests were monitored with two distributed optical sensing systems. One of the systems was recording in the new well and included measurements of distributed acoustics (DAS), temperature (DTS) and strain (DSS). The other system recorded from a preexisting well located about 500 m away and consisted of DAS measurements in a vertical cable.

Over the duration of the circulation tests, the two DAS systems detected on the order of 250 microseismic events in the close vicinity of the new well. A preliminary analysis of this group of events in terms of their hierarchical clustering allowed the selection of a small subset of them, which was used for the joint inversion of their absolute location and a layered velocity model for the site. The two-well DAS data of some of the events with absolute location was then used as a reference to estimate relative locations for the rest of the event’s catalogue. The estimated relative locations display a SW-NE alignment approximately perpendicular to the orientation of the new deviated well within the depth range of about 2300m to 2500m measured from the ground level. The location uncertainties more often show longer horizontal elongations suggesting a better depth constraint compared to the epicentral locations.

DTS data was analyzed in conjunction with slow strain data derived from the DAS data. Following the circulation tests, a ramp-like feature was observed beginning at approximate 1500m migrating upward throughout time in the slow strain data potentially indicating preferential fluid flow at this depth. Slow strain data is inextricably tied to temperature due to the thermal elongation of fiber, and as such, a thermal anomaly with similar characteristics is also observed at this depth. However, similar warming trends exist at various intervals throughout the well without the accompanied slow strain response indicating a truly anomalous interval that features unique DTS and slow strain response.

How to cite: Stanek, F., Vera Rodriguez, I., Wolpert, J., Podrasky, D., Stork, A., Chamarczuk, M., Becker, M., and Ajo-Franklin, J.: Analysis of DAS and slow strain measurements recorded during circulation tests at the FORGE geothermal underground laboratory, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17595, https://doi.org/10.5194/egusphere-egu24-17595, 2024.

Mt. Etna, the largest volcano in Europe, is known for its almost persistent activity and complex seismic wavefield, making it an attractive location for examining volcanic explosions and testing new instrumentation in seismology (e.g., rotational sensors, strain-meters, fiber optic sensing). In 2018, a study was conducted at the Pizzi Deneri (PDN) observatory, situated near Mt. Etna’s summit to understand new instrumentation responses to the local seismo-acoustic wavefield. During volcanic explosions the released energy is mainly partitioned into seismic waves traveling through the ground, and sound waves traveling through the atmosphere. To capture this phenomenon, a temporary multi-parameter network comprised of infrasound sensors, broad-band seismometers (BB) and a fiber optic cable buried within the local loosed granular medium (scoria layer). The fiber optic cable was connected to a Distributed Dynamic Strain Sensing (DDSS) interrogator.

At Etna, volcanic explosions were observed at co-located BB, infrasound and DDSS virtual sensors. A notable example(visible in both BB and DDSS data) is the successive occurrence of a 1-2 Hz seismic signal with a duration of ~4 seconds, followed by a ~2 Hz acoustic signal originating from the explosion, recorded at infrasound sensors. Unusually, simultaneous to the arrival of the acoustic signal observed at the infrasound sensors, DDSS and BB sensors record a signal with a frequency content of 15-20 Hz with a duration of ~2 seconds. We hypothesize that the 15-20 Hz signal is resulting from a non-linear ground response due to the air-to-ground coupling of the air pressure wave.

In order to better characterize this phenomenon, a second experiment was conducted in 2019 at PDN with a similar instrumentation as in 2018, but with a different spatial arragenment. During three months the infrasound sensors recorded each about 65000 volcanic explosions. In this work we analyze the respective ground responses of volcanic explosions observed in the DDSS records of the 2019 campaign. We observe similar phenomenon as in 2018 (non-linear ground response), nevertheless, not all explosion can trigger this response. We first characterize the explosion events from both infrasound and DDSS records, and then classify them using their waveform similarity. The preliminary results provide a broad characterization of the non-linear ground response phenomenon and an insight into the physical properties and processes that are necessary for a pressure wave to trigger a non-linear ground response. The outcomes of this work provide a better understanding of acoustic-to-ground energy coupling in volcanic environments and their potential to trigger other hazards.

How to cite: Diaz-Meza, S., Jousset, P., Currenti, G., Costes, L., and Krawczyk, C.: Understanding non-linear ground response with Distributed Dynamic Strain Sensing (DDSS) at Mt. Etna volcano, Sicily., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17644, https://doi.org/10.5194/egusphere-egu24-17644, 2024.

Pyroclastic flows are highly hazardous phenomena demanding precise detection, localization, and comprehensive characterization for effective volcanic risk management. During October and December 2022, the volcanic activity of Stromboli produced more than 60 pyroclastic flows. The flows propagated from the craters (~700 m a.s.l.) to the sea, resulting in tsunami waves ranging from centimetres to meters in height. These events coincided with an experiment involving distributed acoustic sensing (DAS) data acquisition with a dedicated 4-kilometer-long fibre-optic cable. A 3-component array consisting of 27 geophones and the multi-parameter monitoring network managed by the Laboratory of Experimental Geophysics (LGS) at the University of Florence were active simultaneously. We study two distinct pyroclastic flows of varying intensities. Using array processing techniques applied to DAS, seismic, and infrasonic measurements, we estimate back-azimuths that consistently track flows moving at velocities between 40 and 50 m/s from the craters to the shoreline. Validation of these measurements was accomplished through georeferenced images obtained by a visible camera, affirming their accuracy. These results demonstrate the effectiveness of the three datasets in monitoring pyroclastic flows and the need for multi-parametric observations for a better interpretation of volcanic phenomena.

How to cite: Métaxian, J.-P., Biagioli, F., Trabattoni, A., Stutzmann, E., Lacanna, G., Risica, G., Bernard, P., Capdeville, Y., Mangeney, A., Monteiller, V., Diana, G., Innocenti, L., and Ripepe, M.: Using Distributed Acoustic Sensing, Seismic and Infrasonic Observation to Track Pyroclastic Flows at Stromboli Volcano (Italy) , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17953, https://doi.org/10.5194/egusphere-egu24-17953, 2024.

T phases are acoustic waves that propagate in the low velocity zone of the oceanic sound channel that acts as a waveguide, the SOFAR channel. They are generated by earthquakes through the conversion of seismic energy at the solid-liquid interface, but the exact processes involved are still under debate.

Due to their low attenuation and slow propagation velocity, these arrivals are especially useful for the detection and characterisation of small earthquakes in marine basins, as they can improve the location of the event while their waveforms can yield information on source rupture.

In October 2023, a Distributed Acoustic Sensing (DAS) interrogator was installed on the GeoLab dark fibre in the Atlantic, starting at the Praia Formosa CLS, in Madeira Island, Portugal. The instrumentation of this cable is part of a project by ARDITI and the Oceanic Observatory of Madeira where oceanographic data recorded by buoys and autonomous vessels are combined with DAS data to obtain a global view of the underwater environment of Madeira Island in all its physical, chemical and biological aspects, including the characterisation of regional seismicity. This initiative is also linked to the SUBMERSE project, as the Madeira cable is a pilot site to establish continuous DAS monitoring along many more submarine fibre-optic cables.

On October 27th, a near-source (<40 km) M2.9 earthquake was recorded by the DAS interrogator along the entire cable. The epicentre of the earthquake was east of the Desertas Islands, southeast of Madeira. Besides the P and S phases, very clear T phases are also visible. The recorded T waves have strain values larger than those of P and S waves. However, multiple T phases are identifiable, suggesting different points of conversion or even possible reflections.

This work was funded by the Portuguese Fundação para a Ciência e a Tecnologia (FCT) I.P./MCTES through national funds (PIDDAC) – UIDB/50019/2020 (DOI: 10.54499/UIDB/50019/2020), UIDP/50019/2020 (DOI: 10.54499/UIDP/50019/2020) and LA/P/0068/2020 (DOI: 10.54499/LA/P/0068/2020), and by ERC project SUBMERSE, HORIZON-INFRA-2022-TECH-01-101095055.

How to cite: Schlaphorst, D., Loureiro, A., Matias, L., Custódio, S., Corela, C., and Caldeira, R.: Near-source T-wave observations in the North Atlantic using Distributed Acoustic Sensing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18170, https://doi.org/10.5194/egusphere-egu24-18170, 2024.

The Istanbul Natural Gas Distribution Company (İGDAŞ) has recently embarked on utilizing existing Fiber-Optic (F/O) cables to enhance disaster prevention and mitigation efforts in Istanbul. We are exploring the potential of a novel technology called F/O Distributed Acoustic Sensing (DAS) for earthquake early warning systems. The strategic placement of the F/O cable, which crosses the North Anatolian Fault in the Marmara Sea, presents a unique opportunity for monitoring seismic activity. While seismic stations exist around the Marmara Sea, the absence of online operating Ocean Bottom Seismometer (OBS) stations makes the F/O cable the only sensor positioned across the fault lines expected to rupture during a major earthquake.

The monitored F/O cable, originally intended for telecommunications, spans 60 kilometers in the Sea of Marmara. Over the past 7 months starting in June 2023, more than 160 earthquakes ranging from magnitudes 1.0 to 7.5 have been recorded through the F/O cable. Notably, the F/O DAS system successfully captured significant distant events, notably the February 6, 2023, M7.8 and M7.5 earthquakes in Kahramanmaraş. This initiative highlights the critical stages, obstacles, and best practices associated with deploying this technology. It underscores the importance of precise cable layout, optimal sensor density, range optimization, and the conduction of shaking table tests.

Shaking table experiments were conducted to compare noise levels across various sampling rates. By subjecting a Force-Balanced Accelerometer (FBA) and F/O cable to simulated seismic activity resembling the 1999 Sakarya Earthquake (M6.9) with sine signals at frequencies of 0.25 Hz, 0.5 Hz, 1.5 Hz, 2 Hz, and 3 Hz, observations revealed that reducing the sample rate to 200 sps significantly lowered the interrogator's instrumental noise compared to 2000 sps. Hence, a lower sample rate proved advantageous in achieving a better Signal-to-Noise Ratio (SNR).

Through the analysis of acoustic signal variations along the F/O cable, the DAS systems can accurately pinpoint and characterize earthquake events, facilitating timely warnings. F/O DAS technology boasts distinct advantages in earthquake detection due to its capacity to capture a broad spectrum of seismic signals, ranging from low-frequency tectonic shifts to high-frequency ground vibrations. The effectiveness of F/O DAS measurements relies on proper coupling, ensuring the efficient transfer of acoustic signals to the optical fiber, thereby ensuring precise detection and interpretation of seismic activity.

How to cite: Coşkun, Z., Özgür, H. G., Koç, B., Dal, M. E., Özkan, E., Erkorkmaz, T., Pamukçu, N. B., Şahin, R. C., Tunç, S., Kalafat, D., Pınar, A., Buldanlı, K., and Özener, H.: Earthquake Early Warning Systems based on Fiber Optic Distributed Acoustic Sensing in the Sea of Marmara, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18371, https://doi.org/10.5194/egusphere-egu24-18371, 2024.

Distributed Acoustic Sensing (DAS) is a novel technology for monitoring seismic waves that shake fiber optic cables (FOC) buried in the ground, at unprecedented spatial resolution as low as 20 cm. Such technology can be in principle applied to laboratory experiments aimed to reproduce landslide phenomena. Here we present a preliminary study of DAS measurements in a 2m x 1m sand-box, where piles of sand act as landslide analogues. Our preliminary results demonstrate the capability of DAS recordings in locating seismic events occurring in the sand-box, related to the growth of the sand piles. This study opens to the possibility of using DAS technology to monitor large-scale (1-10m) laboratory analogue experiments aimed to reproduce geophysical and tectonic processes.

How to cite: Fumagalli, M., Piana Agostinetti, N., and Crosta, G. B.: Distributed Acoustic Sensing measurements in a box: a case-study for laboratory landslide test, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19403, https://doi.org/10.5194/egusphere-egu24-19403, 2024.

Reliable well completion technologies are mandatory for the safe and sustainable use of subsurface reservoirs. Achieving subsurface well integrity requires displacing the entire drilling mud with homogenous cement. For this purpose, surface pump parameters (rate, density, pressure, and volume) are generally measured, yielding average values along the borehole. Here, distributed fiber-optic sensing offers new continuous monitoring options with dense spatial sampling within a borehole.

In the GFK-monitor project (https://gfk-monitor.de/en/), we investigated the primary cementation of an 874 m surface casing at a geothermal well in Munich, Germany. A 699 m long fiber optic cable, implemented by the Geothermie Allianz Bayern (GAB), was attached to the outside of the casing in the annulus between the casing and formation and cemented. This allowed for monitoring distributed dynamic strain rate (DDSS or DAS) throughout cement placement with a sampling rate of 1000 Hz and a spatial resolution of 1 m.

We analyzed the average vibration energy in the borehole by DDSS data using a root mean square approach in rolling windows for different frequency bands. Linear features with two different characteristics appear prominently in the frequency band between 0.2-0.3 Hz. One feature, which we named the “slow feature”, shows a varying slope resulting in velocities ranging from 2.5 to 6 m/s. In contrast, the accordingly called “fast feature” indicates a relatively constant velocity of around 6 m/s.

To better understand these features, a theoretical model was developed that simulates the rising velocity of fluids along the annulus. This model uses a cumulative approach and considers the cement pumping rate, borehole geometry, and timings from the daily reports. We assume, that the predicted minimum velocity is required to fill the whole breakout volume. This means that the faster the velocity the smaller the flow paths cross section.

A comparison of data and models reveals that the varying velocities of the "slow feature" correlate with velocities predicted by using the borehole geometry from the caliper log. The rather constant “fast feature” correlates instead with the predicted velocity of the model based on a uniform borehole geometry with the smallest possible radius, the drill diameter, which thus neglects all breakouts.

In summary, due to the nearly constant pumping rate during the monitoring campaign, we hypothesize that different rising velocities result from variations in the cross-sectional area of the flow path.  Based on our observations, the majority of the water spacer does not flush the borehole breakouts. With a good match to the minimum required velocity, these breakouts are filled by the first arriving cement. The following cement, with the same density, rises again on the fastest track. Thus, measuring the cement rising velocities with fiber optics and comparing them to the minimum required velocity from modeling might be a new tool to assess displacement efficiency in real-time.

How to cite: Hart, J., Polat, B., Schölderle, F., Ledig, T., Lipus, M. P., Wollin, C., Reinsch, T., and Krawzcyk, C.: A Fiber-Optic Approach for Cement Placement Monitoring of Deep Boreholes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19804, https://doi.org/10.5194/egusphere-egu24-19804, 2024.

Distributed Acoustic Sensing (DAS) enables cost-efficient retrieval of wavefield information alongside large linear infrastructure elements, such as pipelines, cables or railways. The massive datasets require automated approaches to detect and classify anomalies which can trigger further investigation through an operator, if necessary.

Within this work we exploit one week of DAS data from a fiber-optic cable co-located with a 50 km long railway line south of Trondheim, in the center of Norway. The data were acquired with a temporal sampling of 2 kHz and channel spacing of 4 m, resulting in 12500 channels. Treating the DAS time-space domain matrices like images we can make use of well-established techniques from the field of computer vision. We compute sliding RMS windows of 60 s and 1.5 km with 50 percent overlap and use them as input images for a Convolutional Neural Network. The network classifies events such as trains, cars, unknown events as well as different noise classes and artefacts. In order to thoroughly train the network, we labeled approximately 1000 RMS images per class and further applied a variety of data augmentation techniques to finally obtain about 5000 labeled images per class. Once trained, we can simply apply a forward pass through the network every 30 s for all the 1.5 km overlapping segments to obtain a live-classification of events along the entire railway line.

We present our workflow as well as initial results and discuss the potential of DAS for future railway monitoring and the challenges that we encounter. If successful, these methods can open up an opportunity to exploit a large amount of fibers co-located with railway lines enabling automatization of real-time railway monitoring.

Acknowledgements:

We acknowledge Bane NOR and Alcatel Submarine Networks for conducting the data acquisition for this project. This work is supported by the SFI Centre for Geophysical Forecasting under grant No. 309960.

How to cite: Growe, K., Tveit, A., Dong, H., Wienecke, S., Landrø, M., and Jacobsen, J.: Towards a real-time railway monitoring system based on Distributed Acoustic Sensing and a Convolutional Neural Net, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20360, https://doi.org/10.5194/egusphere-egu24-20360, 2024.

The Bransfield Basin, located between the South Shetland Island and the Antarctic Peninsula in Western Antarctica, is formed by slab rollback of the South Shetland Trench and complementary extensional. The region also has large and small active volcanoes like the Orca volcano. Various earthquakes have occurred due to tectonic movement and volcanic activities. Since the Mw 4.9 earthquake struck at a depth of 10 km in the vicinity of South Shetland Islands on 29 August 2020 (USGS), moderate earthquakes have constantly been reported adjacent to Bransfield Basin like an earthquake swarm.The Extreme Geoscience Team (EGG) at the Korea Polar Research Institute (KOPRI) has been operating a seismic station at the King Sejong station. Other countries have been running their seismic stations in the South Shetland Islands, such as the Antarctic Seismographic Argentinean Italian Network (ASAIN) and Chilean seismic stations, to monitor local earthquakes associated with back-arc spreading and submarine volcanic activities in this area. However, these seismic stations are located on the islands, and the inter-station distance is over 50 km. To make up for the poor coverage of seismic stations, we installed 5 OBSs in the earthquake swarm area between 2020 and 2021.

Using cross-correlation with onland seismic data and OBSs data, we conducted the template matching analysis and detected 67,952 events from August 29, 2020, to December 31, 2021. We relocated 2,313 epicenters of the seismic events by manually picking the first arrival of the P-wave. We found that the earthquake swarm’s epicenters were mainly concentrated in the central basin of the Bransfield Strait.

We categorize the characteristics of the earthquake near Bransfield Basin. We will contribute to understanding the dynamics and evolution of subduction zones in Bransfield Basin.

How to cite: Jung, J., Park, Y., Lee, W. S., and Park, S.: Seismicity in the Bransfield Strait, Antarctica, using temporal Ocean Bottom Seismographs data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1332, https://doi.org/10.5194/egusphere-egu24-1332, 2024.

surfQuake is a new software designed to overcome the workflow process that involves the estimation of seismic source parameters. The complete set of toolboxes inside surfQuake allows the user the complete automation of seismic phases arrival times estimation and events association, event locations, magnitudes and attenuation, slowness vector, and moment tensor inversion.

The software is programmed in Python 3 and offers the users the possibility of three programming levels. From the core library, which allows the user to integrate the core of surfQuake into his own scripts, to the Command Line Interface, which gives the user access to an upper layer that simplifies the use of the core. Finally, surfQuake core is wrapped by a Graphical User Interface (GUI) and connected to a SQL Lite database to store all the results. The user has direct access to query the database tables to extract important information through the object Relational Mapper SqlAlchemy directly managed from the GUI.

The software has been fully tested with the earthquake cluster that occurred during the eruption of La Palma in 2021-22. The source parameters resulting from the eruption and the basic statistics associated with them are displayed using the database toolbox. Additionally, we offer a web tutorial with the documentation of surfQuake and a set of usage examples for the three programming levels.

How to cite: Cabieces, R., Junqueira, T. C., and Relinque, J.: SurfQuake (SQ): A new Python toolbox for the workflow process of seismic sources, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2816, https://doi.org/10.5194/egusphere-egu24-2816, 2024.

Accurate timing corrections for seismic data recorded by ocean bottom seismometers (OBS) are essential for a wide range of applications. The synchronization of internal OBS clocks with Global Positioning System (GPS) is only possible prior and subsequently to deployment in the seafloor. Thus, untracked, possibly nonlinear clock errors in seismic data may accumulate over the deployment period. The measurement of the clocks offset from GPS at retrieval, referred to as `skew’, can be used to correct the data solely under the assumption of a uniform rate of clock drift throughout the whole deployment. We estimate OBS clock drift curves for 40 OBSs of the large-scale UPFLOW amphibian array in the Madeira-Azores-Canaries region that lasted ~14 months in 2021-22. We use the relative shift of daily Empirical Green Functions obtained from seismic ambient noise recorded by all available data channels to track clock error. We find that 95% of our OBS clock drift observations have a substantial non-linear component, notably in the first months of the deployment. Overall, our estimated time-dependent clock drift curves accurately predict the final skew, with an average difference of ~160 ms to GPS skew values obtained at retrieval. We test our skew curves by using them to correct examples of recordings of teleseismic earthquakes and of local-regional seismicity. Uncertainty analysis of the skew curves gives a mean skew error of ~110 ms, indicating the suitability of the corrected data for future seismological studies such as for seismic tomography, seismicity analysis and moment tensor inversions. A dedicated open-source Graphical User Interface toolbox in Python 3 has been developed to facilitate OBS clock synchronization using seismic ambient noise.

How to cite: Krueger, F., Cabieces, R., Harris, K., Ferreira, A., Tshekmistrenko, M., Hicks, S., Geissler, W., Hannemann, K., and Schmidt-Aursch, M.: Clock skew corrections of a large-aperture OBS array in the Atlantic Ocean reveal the need to include non-linear drift corrections, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4409, https://doi.org/10.5194/egusphere-egu24-4409, 2024.

             Constructing a seismic network in the oceans has been one of the most difficult challenges in seismology. Almost 70 % of the Earth's surface is covered by the ocean, so such a network is necessary for investigating the Earth's interior. Because the Pacific Ocean is the largest ocean basin, it was considered impractical to establish a large-scale dense seismic array.

              However, in 2014, Kawakatsu and colleagues proposed a new array concept (Kawakatsu et al., 2014). Deploying ~15+ BBOBS as an array unit for 1-2-year observation period, and repeating such observations in a leap-frog fashion for a decade or so, would enable us to cover a large portion of the Pacific Ocean. International collaboration was essential to successfully implement this strategy, and since 2015, the concept has been realized as "Pacific Array" via international collaboration between partners in Japan, U.S.A, EU, South Korea, Taiwan, and China.

              Since then, many array units have been funded, providing broadband data from across the Pacific. Oldest-1 (Japan-S. Korea), Oldest-2 (Japan-Taiwan), Young- and Old-ORCA, and OHANA (U.S.A) have all been completed; Galapagos and SaLOON (U.S. A.) are currently deployed; and HEB (Japan-Germany), GTJ, and EPIC (U. S. A.) will be deployed in the near future. Although each of these observations have their own scientific goals, they contribute to the larger international Pacific Array structure. We summarize recent progress of the Pacific Array and some of scientific result in each array.

How to cite: Isse, T., Kawakatsu, H., Lee, S.-M., Kuo, B.-Y., and Gaherty, J. B.: Pacific Array: international collaboration for large-scale array experiment in the Pacific basin, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4486, https://doi.org/10.5194/egusphere-egu24-4486, 2024.

The Zabargad Fracture Zone (ZFZ), located between 23.5^o N – 26^o N is the largest rift-axis offset in the Red Sea. The ZFZ is a fundamental tectonic element of ~100 km rift-axis offset that marks the transition between the northern and central Red Sea. Due to data scarcity, our understanding of the seismic activity and the potential presence of transform faults or non-transform offsets in the ZFZ area is limited. Local seismological datasets so far have been restricted to the onshore recordings, thus hampering the ability to study the ZFZ's role in the Red Sea’s tectonic evolution and to assess its potential seismic hazard associated with transform faults accommodating the rift-axis offset in the Red Sea.

To fill this data gap, we deployed 14 broadband Ocean Bottom Seismometers (OBS) and four land stations for 12 months within the ZFZ to collect continuous waveforms. The newly collected data span the period 11/2021 to 11/2022. First, we corrected recorded OBS-waveforms for clock drift using skew values measured at the time of the OBS recovery and time-shifts measured from ambient noise cross-correlations for stations whose skews could not be measured. Next, we applied a deep-learning-based algorithm to automatically detect earthquakes and pick P and S phases. We verified and, if necessary, manually corrected these phase picks; this approach identified more than 3500 local earthquakes.

To obtain reliable hypocentral locations, we inverted for an optimum crustal 1-D seismic velocity model of the ZFZ and station delays simultaneously. Using as reference a nearby land station installed on Precambrian bedrock, we obtained positive OBS station delays of up to 2 seconds. These delays correspond to late phase arrivals, most likely due to thick sedimentary and salt deposits. Our local magnitude (M_L) calculations show that OBS-station M_L values are up to 1.5 units larger than for the land stations. We find that station delays and M_L deviations are correlated, highlighting the importance to account for the variability of shallow geological and bathymetric properties for accurate earthquake location and magnitude estimation.

Our seismicity analysis reveals two major distinct spatial clusters of earthquakes, in the southern (23.95 ^o N – 24.53 ^o N) and northern (24.68 ^o N – 25.43^o N) parts of the ZFZ. Seismicity in the south higher seismicity rate and spatially concentrated than in the north. We also identified three seismic swarms in the south and one seismic swarm in the north lasted for one to three weeks each. Our preliminary analyses document the potential of this new dataset to address key questions on the seismotectonics and seismic activity of the ZFZ.

How to cite: Shiddiqi, H. A., Parisi, L., Cano, E. V., Mai, P. M., Augustin, N., and Jónsson, S.: Seismic Activity in the Zabargad Fracture Zone (Red Sea): Insights from an Amphibious Seismic Network, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4563, https://doi.org/10.5194/egusphere-egu24-4563, 2024.

In this study, we focus on developing a low-angle active leveling device of geophone for Ocean Bottom Seismometers (OBS). The deployment of OBS in varied underwater environments poses a significant challenge due to unpredictable terrains, which often affect the positioning and orientation of these instruments. In such scenarios, the role of a leveling device becomes crucial, especially for self-floating OBS. Most of the leveling devices are designed with a 360° range, resulting in high costs and limited versality, our approach is informed by previous OBS positional data, suggesting that a limited corrective angle is adequate for underwater conditions. Our innovative leveling mechanism resembles a joystick in its transmission system, topped with a platform to enhance versatility. The control rod below the platform is controlled by two DC motors combined with worm gears, adjusting the rod along the X and Y axes with an approximate corrective angle of 30 degrees. For seismic sensing, our system utilizes one vertical and two horizontal geophones, housed in a custom sensor box to ensure the three axes are orthogonal to each other. This design simplifies the complexity of parts, combining off-the-shelf and custom components, offering a significant size and cost advantage. We have also tested its performance, with a focus on eliminating signal interference from resonance frequencies and assessing the system's stability. This includes a comparative analysis with older leveling mechanisms.

How to cite: Yang, C.-H., Lin, C.-R., Gung, Y., Lin, F.-S., and Chang, K.-H.: The Design of a Low-Angle Active Leveling Device for Short-Period Ocean Bottom Seismometer, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4888, https://doi.org/10.5194/egusphere-egu24-4888, 2024.

Human activities in the marine environment coupled with the efficient propagation of acoustic waves, have significantly increased ocean noise levels. This rise poses a threat to marine biodiversity, especially impacting vocal-dependent marine mammals. Therefore, obtaining a more comprehensive understanding of the complex ocean soundscape, shaped by both natural and anthropogenic sources, is essential.

Ocean-bottom seismometer (OBS) datasets present an opportunity to explore the marine acoustic noise field. OBS hydrophones can be used to effectively study regional marine soundscapes. Additionally, the three-component OBS seismometers capture acoustic-to-seismic conversions from the water column, enabling the detection and localization of acoustic events.

Through the analysis of OBS hydrophone recordings from the NE Atlantic offshore Ireland, we develop a frequency-dependent empirical sound propagation model for the area. The model provides information on (1) the spectral amplitude decay of sound sources in the region and (2) the distance between OBSs and acoustic sources.

To construct the model, we consider ship noise from individual vessels recorded by OBSs at known distances from the source, using Automatic Identification System (AIS) ship-tracking data. As a ship moves away from its closest point of approach (CPA) to an OBS, the hydrophone records the evolution of the frequency-dependent ship's acoustic signal spectral amplitude with distance. This enables us to compute amplitude decay curves for narrow frequency bands from different vessel signals recorded at various OBSs. We apply a signal-to-noise ratio (SNR) quality control criterion and consider ships sailing along 'straight' paths at a constant speed (< 2 kn variation). After normalizing and scaling the CPAs of the different curves to a reference, we stack them for each frequency band and compute polynomial fits for the stacked amplitude decay curves. We analyse the number of vessel tracks needed to build a reliable empirical model. Numerical simulations are computed to further understand the acoustic propagation properties from our empirical model.

The goal is to investigate (1) how spectral amplitudes at different frequencies behave with distance from the sound sources, (2) the efficiency of frequency-dependent sound propagation in the ocean in regions with variable seafloor morphologies, and (3) the potential use of the empirical model to obtain distances to other sound sources with unknown locations, such as Baleen whales. To achieve this, we compute ratios of spectral amplitudes for pairs of different frequency bands using the fitted amplitude decay curves. We test the model's ability to locate other vessels with known AIS positions before using it to locate different sound sources with unknown positions. Locating a sound source employing the distance provided by the empirical model is valuable when OBSs are too far apart for 'traditional' multi-sensor methods. In such cases, the incoming direction of the sound source is determined by rotating the horizontal-component OBS seismic data to point to the source back-azimuth. Also, the amplitude decay information provided by the model has multiple applications in other areas where noise pollution may be a concern (e.g., marine infrastructure construction).

How to cite: Gómez-García, C. and Bean, C. J.: Empirical insights into ocean acoustic wave propagation and source localization using ocean-bottom seismometers, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5502, https://doi.org/10.5194/egusphere-egu24-5502, 2024.

We present the analysis of more than 9000 hydro-acoustic earthquake records, recorded by a network of 50 instruments called Mobile Earthquake Recording in Marine Areas by Independent Divers (MERMAIDs), which are freely floating in the South Pacific Ocean. This network is part of the collaborative South Pacific Plume Imaging and Modeling (SPPIM) project. Our analysis focuses on evaluating how the MERMAID stations complement the conventional worldwide network of stations reporting to the ISC, comprising records of globally distributed seismic stations. In the context of routine earthquake (re)location, we evaluate the improvement of results of earthquake location estimates, particularly focusing on the Tonga-Kermadec subduction zone, where current distribution of seismic land stations is extremely sparse. MERMAIDs are often the closest “station” to earthquakes occurring in the Tonga subduction zone, and frequently fill significant station azimuthal gaps.

We matched the MERMAID records from June 2018 to December 2023 to more than 3000 earthquakes which are reported to the ISC by many international agencies worldwide, allowing us to construct the MERMAID catalogue, complimented with additional parametric data such as MERMAIDs locations, observed direct P-phase picks and travel time residuals with respect to ak135 model, as well as their corrections for bathymetry. These data were integrated into the existing ISC Bulletin, and relocated using the ISC hypocentre relocation algorithm ISCLoc. The results of this successful first integration of MERMAID data into ISC routines presents an important step towards routinely including MERMAID traveltime picks in ISC Bulletin.

How to cite: Namjesnik, D., Sigloch, K., Simon, J. D., Garth, T., Harris, J., Storchak, D., Simons, F. J., Bonnieux, S., Yu, Y., and Obayashi, M.: Re-location and integration into the ISC Bulletin of earthquakes recorded by hydro-acoustic MERMAID instruments freely floating in the South Pacific Ocean, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6052, https://doi.org/10.5194/egusphere-egu24-6052, 2024.

During the TYrrhenian Deep-sea Experiment (TYDE) 14 Ocean Bottom stations were deployed in the period December 2000-May 2001 on the southern Tyrrhenian seafloor around the Aeolian islands All the stations were equipped with Hydrophones (OBH) and six of them with 3-component geophones (OBS). This experiment represented one of the first and most important passive ocean bottom data acquisition ever realized in southern Tyrrhenian and the records were mainly used to perform receiver functions, modeling of the seismic velocity field, and detect and locate seismicity.

In this work we present a reprocessing of these old data using modern techniques of analysis to improve the quality of the signals and extract more information of what has been already published.

We show the workflow used for cleaning the records together with Machine Learning approaches to improve the detection of small events and the application of methods that were not common at that time, such as the Ambient Noise analyses.

Work supported by the “Pianeta Dinamico” call, 2023–2025 CAVEAT

How to cite: Calò, M., Gamez Lindoro, J. A., Di Luccio, F., and Bernal Manzanilla, K.: Reprocessing of the ocean bottom seismic data of the TYDE experiment using modern techniques., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6977, https://doi.org/10.5194/egusphere-egu24-6977, 2024.

A fruitful close collaboration between teams of researchers and engineers of Geoazur, a research laboratory specializing in seismic observation, and Osean, a dynamic company combining the talents of engineers in low-noise, low-power electronics specialized in underwater acoustics, gives this instrument a guarantee of robustness, reliability and innovation by capitalizing on all their knowledge in their respective fields. Halios has an autonomy of 18 months, dimensions of 1.1 x 1.1 x 1m and a total weight with ballast of 346 kg. Electrical power is supplied by lithium or alkaline batteries. Halios is made entirely from corrosion-resistant materials. The 4-panel, mortise-and-tenon structure naturally creates a well in which the seismometer is protected. This well is covered by two flaps that open when the OBS sinks into the water, and close when the OBS stops on the bottom. During descent, the seismometer is suspended in the well. Once the OBS is in place, the seismometer is released automatically or by an acoustic command to land on the ground without any mechanical contact with the structure. This type of coupling is ideal, and protects the sensor from underwater currents. Communication with the OBS is possible via Ethernet, Wi-fi or acoustics once in the water by non-specialists. An acoustic module makes it possible to retrieve a health report on demand, which groups together the essential parameters of the station, and to modify some of them once the OBS is in the water. Halios incorporates precise clock or a CSAC-type atomic clock, if time precision is required.The seismometer is a compact Nanometrics Trillium, integrated in a dedicated container which also houses an accelerometer. An hydrophone, a precision temperature sensor and an absolute pressure sensor complete the range of sensors integrated on Halios. An dedicated acoustic modem also enables partial data retrieval from the surface.To retrieve the OBS, an acoustic command activates a mechanical release which frees the ballast. Dynamic assistance from leaf springs accelerates the positive thrust provided by the OBS's high-density syntactic foam. This foam is injected into a PHD shell, which effectively protects the whole unit and, thanks to its compact shape, provides effective anti-shock protection. The ascent of the OBS into the water column can be monitored by the acoustic module.Once at the surface, the time difference between the OBS clock and the GPS datum is automatically calculated. The OBS sends its GPS coordinates to the ship by VHF to facilitate recovery. At night, an LED flashing light can also be used to pinpoint its position. Halios is equipped with an interface that enables it to be connected to a real-time cable network, making it a versatile OBS of the highest performance and innovation.

How to cite: Hello, Y., Rebour, C., Sigloch, K., Bonnieux, S., and Philippe, O.: HALIOS  an innovative and clever Broad Band OBS., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8185, https://doi.org/10.5194/egusphere-egu24-8185, 2024.

Cabled ocean bottom seismometer (OBS) solutions are financially and logistically challenging and autonomous OBS systems do not provide operators with seismic data until recovery. To address this issue Güralp has developed Aquarius, an ultra-low-power, free-fall OBS system, operational at any angle and with the ability to transmit seismic data in near-real-time from the seafloor without the use of cables.

The Güralp Aquarius incorporates seafloor-to-surface acoustic communication technology that allows state-of-health and noise performance interrogation during installation followed by retrieval of seismic data throughout the deployment period.

Omnidirectional broadband seismometer components allow the Aquarius to land and operate on steep slopes without requiring a gimbal mechanism that inherently introduces noise and failure modes. Raw data is recorded uncorrected for orientation to allow users to correct during post-processing.

These unique features allow the sensor to function on uneven seafloor as well as transmitting seismic data to the surface where the operator can use noise characteristics, location, and orientation data to determine if the landing site is suitable.

Intelligent battery design allows for typical 18-month deployments, with charging being possible alongside data transfer. This allows recharging, download and configuration simultaneously on the ship in between deployments.

Ease of configuration, deployment and recovery followed by simple data processing are all central themes to the Aquarius. The capital investment required to purchase OBS systems often means that the OBS instrument must be adaptable to a range of use-cases.

The Aquarius is in use in 6 different countries and is the unit of choice for the Canadian OBS pool, comprising 120 units for deployment around the globe. Aquarius units have been successfully deployed in the Mediterranean, North Atlantic as well as the North and South Pacific. Here, we focus on the mechanics of deployment and recovery for demonstrating to experienced and prospective principle investigators how this system simplifies and improves confidence in deployment.

How to cite: Clark, A., Lindsey, J., Hill, P., Foster, C., Restelli, F., Watkiss, N., Price, E., Nedimović, M., and Cairns, G.: Pushing Boundaries with Ocean Bottom Seismometers (OBS) with a Pool-Ready System: Güralp Aquarius, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10807, https://doi.org/10.5194/egusphere-egu24-10807, 2024.

Autonomous Ocean Bottom Seismometer (OBS) deployments have often involved a degree of “drop-and-hope” due to the inherent lack of seismic data communication during installation as well as waiting extended periods before data collection. Cabled solutions provide real-time data during and immediately after deployment, sometimes with opportunity to adjust the instrument before it is left to operate remotely. However, cabled solutions are inherently financially and logistically challenging both in terms of seismic hardware and arguably more significantly, deployment hardware (ships, ROVs, cables etc.). The geographical reach of these experiments is also often limited to within a few hundred kilometres of the coast. These constraints often mean cabled OBS are beyond the scope of most scientific bodies.

Güralp Systems, in collaboration with INGV, has successfully manufactured and demonstrated a method of reducing financial and logistical constraints and extending geographical range by utilising force-feedback seismic instrumentation in cabled OBS systems. The recent successful deployment of the InSEA Wet Demo SMART (Science Monitoring And Reliable Telecommunications) cable displays a world first in how science can partner with industry to achieve this.

SMART cables are primarily telecommunication cables that secondarily serve as hosts for scientific monitoring equipment. Commercial viability for these systems relies on the cable being laid as if the science element did not exist, thereby minimising additional deployment costs and reducing barriers to cooperation with cable laying companies. GSL and INGV deployed 3 seismometer-accelerometer pairs housed inline repeaters along the 21km cable length using standard cable-laying techniques to show proof of concept.

This pioneering installation using telecommunication cables marks a significant step towards drastically improving local knowledge of inaccessible oceanic regions as well as global azimuthal coverage for teleseismic events, all in real time.

How to cite: Restelli, F., O'Neill, J., Nicholson, B., Watkiss, N., Marinaro, G., Simeone, F., Embriaco, D., Price, E., Foster, C., and Clark, A.: Successful Deployment of a 21km SMART Cable with Force-Feedback Seismometer and Accelerometers in the Mediterranean Sea, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11062, https://doi.org/10.5194/egusphere-egu24-11062, 2024.

The Iceland hotspot and its interaction with the Mid-Atlantic Ridge have strongly influenced the structure and evolution of the North Atlantic region. The Iceland hotspot has been linked to the presence of the underlying Iceland plume, rising from the deep mantle below Iceland. Due to the sparsity of the seismic data coverage in the North Atlantic Ocean, the structure and dynamics of the underlying plume, as well as its influence on the surrounding North Atlantic region, remain a topic of debate. As part of the project SEA-SEIS (Structure, Evolution And Seismicity of the Irish offshore), a network of ocean bottom seismometers was deployed across a large part of the North Atlantic Ocean—from Ireland and Britain to near Iceland—in September 2018. Over a period of 19 months, they recorded seismic data on the seafloor. 14 OBSs were retrieved in May 2020, of which 12 successfully recorded 19 months of seismic data. The data was thoroughly preprocessed, including the reduction of compliance and tilt noise on the vertical-component waveforms of teleseismic events. These waveforms were individually inverted for surface, S- and multiple S-waves using the Automated Multimode Inversion (AMI), extracting structural information on the lithosphere and underlying mantle.

Here, we combine this new seafloor seismic data with a massive global dataset of waveform fits for almost 1.5 million seismograms, with the coverage maximised in the hemisphere around the North Atlantic, and compute a new seismic waveform tomography model of the North Atlantic upper mantle: NA24. The tomography capitalises on the improved data coverage and reveals both the S-wave velocity and the azimuthal anisotropy structure below the Iceland hotspot, the adjacent mid-ocean ridge, and the entire North Atlantic region at a new level of detail. It shows low seismic velocities below Iceland and the adjacent Reykjanes Ridge and Kolbeinsey Ridge down to ~260 km depth, below which they merge into one low-velocity body located west of Iceland at around 330 km depth. Around 410 km depth, the low-velocity body is present below eastern Greenland where it remains throughout the transition zone. This observation implies that the Iceland plume is located below eastern Greenland in the transition zone, moving eastward as it rises in the upper mantle. In the shallow upper mantle, a strong ridge-parallel azimuthal anisotropy structure is revealed along the full length of the Reykjanes Ridge. This implies a strong channelled horizontal flow from the Iceland plume towards the south below the ridge axis, while any sign of a radial outward flow of the plume is absent.

How to cite: de Laat, J. I., Lebedev, S., Bonadio, R., Thsekhmistrenko, M., and Celli, N. L. and the SEA-SEIS Team: Waveform tomography of the upper mantle beneath the North Atlantic region using regional seafloor and global data., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11074, https://doi.org/10.5194/egusphere-egu24-11074, 2024.

Since the introduction of the ocean bottom seismograph (OBS), it has played a pivotal role in the expansion of our comprehension of the Earth’s interior structure and dynamics. Typically deployed through free-fall from a ship, the position of the OBS on the seafloor and its orientation are initially unknown. When using airgun shots to locate and orient an instrument, a severe downside is the often-limited azimuthal distribution of shots around the OBS, possibly leading to large location and, subsequently, orientation errors. We introduce a novel and efficient method aimed at enhancing the accuracy of station locations in active-source seismic experiments. The approach uses airgun shots recorded as part of the experiment, and integrates both acoustic wave travel time information and waveform polarizations. The inverse problem of station location is formulated in terms of Bayesian inference. At each location of a search grid, the misfit of observed and theoretical water wave travel times and the clustering of polarization data are combined into one probabilistic formulation to map the relative likelihood of a station’s position and orientation. We demonstrate the practical utility of the method via application to the location and orientation of OBS deployed during a recent seismic experiment located across the Hawaiian Ridge. Thereafter, we will use the results to compute radial seismograms, extract the travel times of crustal and mantle S waves, and develop S-wave models of the oceanic lithosphere across the Hawaiian Ridge.

How to cite: Cryder, M., Dunn, R., and Xu, C.: Integrated Bayesian Approach for Station Location and Orientation in Marine Active-Source Seismic Shear Wave Studies, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11876, https://doi.org/10.5194/egusphere-egu24-11876, 2024.

The Ontong Java Plateau (OJP), formed around 120 Ma, represents a significant event in the Earth’s geologic history. It is the largest preserved Large Igneous Province (LIP) by volume on the Earth. The voluminous and rapid magmatism placed a thick 30-40 km crust on a normal oceanic lithosphere in submarine conditions. The lithosphere has since experienced low subsidence suggesting only modest thermal perturbation. Several hypotheses have been proposed to explain the unique emplacement history, including a hot mantle plume, meteorite impact, or passive mantle upwelling close to a fast-spreading ridge. In this study, we image the crust and upper mantle structure beneath the OJP using body waves and surface waves. A challenge of body wave imaging in submarine environments is the severe reverberations generated in the water column and seafloor sediments. Additionally, multiple reflections from shallow interfaces (e.g., Moho and intra-crustal layers) may interfere with deeper upper mantle phases, making interpretation ambiguous. In this work, we obtain high-resolution Ps-RFs at 17 OBS (ocean bottom seismometer) stations in the OJP region using a transdimensional hierarchical Bayesian deconvolution approach. We then take advantage of two recently developed techniques: (1) FADER (Fast Automated Detection and Elimination of Echoes and Reverberations), which uses autocorrelation and cepstral analysis to identify and remove water and sediment reverberations, and (2) CRISP-RF (Clean Receiver function Imaging with SParse Radon Filters), which uses sparse Radon transforms to eliminate multiples and incoherent noise. We anticipate applying our novel ideas to traditional seismic imaging will provide robust constraints on the discontinuities detected beneath the OJP, shedding new light on its lithospheric structure and origin.

How to cite: Zhang, Z. and Olugboji, T.: How did the Largest Oceanic Plateau modify the Normal Oceanic Lithosphere?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13081, https://doi.org/10.5194/egusphere-egu24-13081, 2024.

Ocean bottom seismometer data are strongly affected by noise from seafloor currents.  Techniques based on the transfer function between noise on the horizontal and vertical channels are commonly used to “clean” this noise from low-frequency (< 0.1 Hz) vertical channel data, but questions remain about the efficiency of this technique and its effect on seismological signals.  We present a method to “pre-clean” the vertical data through simple rotation of the seismometer to minimize the noise on the vertical channel.  The rotation is nearly identical to that indicated by the transfer function method, but has the advantage of losing no information.  Once this pre-cleaning is performed, transfer-function based cleaning can still be performed, generally resulting in lower noise levels than those obtained using the transfer function method alone and the remaining transfer function may provide information beyond that of the instrument’s mis-alignment with the gravitational field.  We will present examples from datasets around the world and discuss new information revealed.

How to cite: Crawford, W., Laske, G., Harmon, N., and Rychert, C.: Removing low frequency ocean bottom seismometer tilt noise using simple sensor reorientation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14905, https://doi.org/10.5194/egusphere-egu24-14905, 2024.

Upward mantle flow is key to understand global mantle geodynamics yet its imaging remains challenging due to potential associated low velocity contrasts and small lateral dimensions. In this study we use the UPFLOW ocean bottom seismometer (OBS) dataset to build images of radial anisotropy to constrain the patterns of mantle upwellings in the Azores-Madeira-Canary Islands region. We use the partitioned waveform inversion (PWI) method whereby non-linear waveform fitting of surface waves filtered between T~ 16 s and T~ 300 s is performed using a successive series of time-frequency windows in two stages. Firstly, the surface wave fundamental mode is extracted via phase match filtering and is used to obtain path average perturbations in shear radial anisotropy and isotropic shear wave velocity from a smoothed combination of the 1-D mantle model ak135 and the crustal model CRUST1.0. These perturbations establish a new initial model, which is subsequently used to estimate the path averaged radial anisotropy model that leads to the best fit between the observed trace and the synthetic waveform obtained by summing all overtones (up to n = 20). An iterative, regularized least squares inversion is used to invert for 3-D radially anisotropic mantle structure. Uncertainties are automatically quantified and used to interpret resolved seismic structures. Preliminary results are compared to previous tomographic models of the Atlantic region.

How to cite: Harris, K., Witek, M., Ferreira, A. M. G., Chang, S.-J., and Miranda, M.: Surface wave mantle anisotropy tomography of the Azores-Madeira-Canaries region using UPFLOW data: initial results, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16613, https://doi.org/10.5194/egusphere-egu24-16613, 2024.

The iReverb project explores technical methodologies employed in the UPFLOW project, which addresses the challenges associated with tidal-induced noise on Ocean Bottom Seismometers (OBS) deployed in the North Atlantic Ocean. OBS sensors, unlike those on land stations, are exposed to oceanic currents, whose coupling with the instruments induces reverberations in the recordings. This project showcases a method for exploration of reverberations of tidally-modulated current-induced noise across different OBS types around the Azores, Madeira and Canaries region. The overarching objective is to present a comprehensive technical framework to enhance our understanding of Ocean Bottom Circulation (OBC), providing insights for calibrating current models. A central aspect of this project is the proposed utilisation of machine learning (ML) algorithms for automating the mapping of resonances and obtaining a proxy for OBC.

Here, we present the iReverb methodology, involving manual curation of a pixel-level annotated spectrogram dataset, after this, a ML classifier is trained to identify features in each spectrogram (known as supervised semantic segmentation). To begin with, 15-minute spectrograms between 1-20 Hz are manually annotated using an open-source labelling tool. A deep Convolutional Neural Network (CNN), consisting of a ResNet Encoder, a UNet decoder, and a pixel-wise classification head, is then trained to classify features in each spectrogram using supervised learning. We explore the various factors, such as architecture changes to the CNN, that contribute to improving the model performance for annotating key features. We also discuss insights into attempted alternatives to standard supervised learning (semi-supervised and synthetic labels).

Finally, we discuss the results and evaluate the effectiveness of the ML/DL approach in mapping resonances, demonstrating its potential in utilising resonances as signals for tracking ocean currents (and whales). 

This project was funded by the UPFLOW project (ERC grant 101001601). This work was partially funded by FCT I.P./MCTES through PIDDAC (UIDB/50019/2020, UIDP/50019/2020 and LA/P/0068/2020).

How to cite: Corela, C., Saoulis, A., Tsekhmistrenko, M., Loureiro, A., Miranda, M., and Ferreira, A.: Decoding Ocean Depths: Machine Learning Insights into Tidal Dynamics with UPFLOW’s OBS Data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19546, https://doi.org/10.5194/egusphere-egu24-19546, 2024.

How to cite: Aminian, M.-A., Stutzmann, E., Crawford, W. C., and Montagner, J.-P.: Imaging of the West Indian Ocean Subsurface using Compliance from Ocean-Bottom Stations , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21685, https://doi.org/10.5194/egusphere-egu24-21685, 2024.

ORFEUS (Observatories and Research Facilities for European Seismology, http://orfeus-eu.org/) is a non-profit foundation that promotes seismology in the Euro-Mediterranean area and beyond through the collection, archival and distribution of seismic waveform data, metadata, and associated services and products. The data and services are collected or developed at national level by more than 60 contributing Institutions. They are further developed, integrated, standardised, homogenised and promoted through ORFEUS. Among the goals of ORFEUS are: (a) the development and coordination of waveform data products; (b) the coordination of a European data distribution system, and the support for seismic networks in managing digital seismic waveform data; (c) the encouragement of the adoption of best practices for seismic network operation, data quality control and data management; (d) the promotion of open access to seismic waveform data, products and services for the broader solid Earth science community. These goals are achieved through the development and maintenance of data services targeted to a broad community of seismological data users, ranging from earth scientists to earthquake engineering practitioners. Three Service Management Committees (SMCs) are consolidated within ORFEUS, devoted to managing, operating and developing (with the support of one or more Infrastructure Development Groups): (i) the European Integrated waveform Data Archive (EIDA; https://www.orfeus-eu.org/data/eida/); (ii) the European Strong-Motion databases (SM; https://www.orfeus-eu.org/data/strong/); the European mobile instrument pools (https://orfeus-eu.org/data/mobile/). Products and services for computational seismologists are also considered for integration in the ORFEUS domain. ORFEUS services currently provide access to the waveforms acquired by ~ 24,000 stations, including dense temporary experiments (e.g. AdriaArray; https://orfeus.readthedocs.io/en/latest/adria_array_main.html), with strong emphasis on open, high-quality data. Contributing to ORFEUS data archives means benefitting from long-term archival, state-of-the-art quality control, improved access, increased usage, and community participation. Access to data and products is ensured through state-of-the-art information and communication technologies, with strong emphasis on federated web services that considerably improve seamless user access to data gathered and/or distributed by the various ORFEUS institutions. Web services also facilitate the automation of downstream products. Particular attention is paid to adopting clear policies and licenses, and acknowledging the crucial role played by data providers, who are part of the ORFEUS community. There are significant efforts by ORFEUS participating institutions to enhance the existing services to tackle the challenges posed by Big Data, with emphasis on data quality, improved user experience, and implementation of strategies (e.g. Cloud) for scalability, high-volume data access and archival. ORFEUS actively encourages interoperability and integration of multidisciplinary datasets in seismological and Earth Science workflows. ORFEUS data and services are assessed and improved through the technical and scientific feedback of a User Advisory Group (UAG), which comprises selected European Earth scientists with expertise on a broad range of disciplines. All ORFEUS services are developed in coordination with EPOS and are largely integrated in the EPOS Data Access Portal (https://www.ics-c.epos-eu.org/). ORFEUS is one of the founding Parties and a fundamental pillar of EPOS Thematic Core Service (TCS) for Seismology. ORFEUS and its community are actively involved in EC projects (http://www.orfeus-eu.org/organization/projects/), notably Geo-INQUIRE (https://www.geo-inquire.eu/) and DT-GEO (https://dtgeo.eu/) in 2024.

How to cite: Cauzzi, C., Clinton, J., Crawford, W., Custódio, S., D'Amico, S., Evangelidis, C., Haberland, C., Kiratzi, A., Luzi, L., Kolínský, P., Roumelioti, Z., Schaeffer, J., Sigloch, K., Sleeman, R., and Strollo, A.: Status and Outlook of ORFEUS Data Services, Products and Activities to Coordinate Access to Seismic Waveform Data in the Euro-Mediterranean Region and Beyond, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3477, https://doi.org/10.5194/egusphere-egu24-3477, 2024.

The seismic hazard department and department of the Carpathian region seismicity of the Subbotin Institute of Geophysics of the National Academy of Science of Ukraine, with the support and collaboration of the U.S. Department of Energy, Lawrence Livermore National Laboratory, Michigan State University, and the EarthScope Consortium had done first steps to expand the Ukrainian National Seismic Network (UT) through the installation of permanent broadband seismic stations in the territory of Ukraine.

The main goal of the Seismic Network Expansion and Modernization in Ukraine is to improve regional network coverage by making high-quality data from new stations openly available to the global scientific community in real time. It will strengthen national seismic monitoring and earthquake response capabilities in Ukraine by upgrading and expanding the national networks with high-quality broadband seismometers and strong motion sensors. It will emphasize data quality, real-time data exchange, and network sustainability through training in best practices on station site selection, installation, data management, and network operation.

To maximize the effectiveness of investigating existing and new seismic sites, multiple factors must be considered during the initial selection, preparation, and installation of new seismic stations. One critical component during the site selection process of any seismic network is an assessment of the seismic noise level at potential sites. The capacity of any seismic station to detect earthquakes and record high quality waveforms will be determined by the signal and noise characteristics of the site. Proper site selection is strongly related to the network's region and can be a critical issue. Besides the earth’s natural background noise, there are other noise sources to consider like those related to infrastructure close by (roads, traffic, mining activity, tenants, etc.). The goal of this work is to conduct noise surveys that can be quickly deployed in order to efficiently evaluate potential sites for the installation permanent seismic stations.

An initial noise survey was conducted at two sites: one at an existing site LUBU (near Liubeshka village, Lviv district) and one at a new site SUGL (near Mala Uhol`ka village, Zakarpattia district). We analyzed, and report the data in the form of both time-history examples and standardized Probability Density Function noise plots. Seismic spectral analysis based on the calculation of Power Spectral Density distribution using a Probability Density Function by McNamara approach.

How to cite: Farfuliak, L., Amashukeli, T., Chiang, A., Mackey, K., Haniiev, O., Kuplovskyi, B., Aderhold, K., Burk, D., Prokopyshyn, V., and Petrenko, K.: Initiation of the Seismic Network Expansion and Modernization in Ukraine, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4414, https://doi.org/10.5194/egusphere-egu24-4414, 2024.

In recent years, significant attention has been devoted both to seismic data processing and data quality procedures. At the Italian scale, EIDA Italia (https://eida.ingv.it/it/getdata) and ISMDq (http://ismd.mi.ingv.it/quality.php) represent the web portals currently available for checking seismic data quality. In this work, we introduce the SDQ (Seismic Data Quality) project, a new open-source Python-based tool, designed for the automatic data quality check of sismo-accelerometric stations considering both selected earthquakes and continuous data streams. Regarding earthquake data, the quality of individual waveforms is assessed by initially comparing – at first - the ground motion parameters derived from co-located accelerometers and velocimeters. SDQ operates by using a simple external input file including the INGV event-id and both station and network codes. Event information, station metadata, and waveforms are obtained from FDSN (https://www.fdsn.org) web services (https://www.fdsn.org/webservices/). Each single waveform is assigned to a quality class ranging from A (high quality) to D (data to be rejected) based on time- and frequency-dependent algorithms. Classification thresholds were empirically obtained by combining visual signal inspection and statistical analysis considering 15.000 waveforms recorded in Italy from 2012 to 2023 by IV (National Seismic Network, https://www.fdsn.org/networks/detail/IV/) and MN (MedNet network, https://www.fdsn.org/networks/detail/MN/) sismo-accelerometric stations. Concerning continuous data streams, mini-seed recording signals are analyzed at each selected station to set empirical thresholds considering several data metrics (i.e. frequency-dependent Root Mean Square, RMS, and Power Spectral Density, PSD) and data availability information (i.e. % gap and data availability, sum of gaps, maximum gap etc.) to build a station-quality archive. Users can select and build target time histories for each network, station, data stream and single ground motion component related to the selected input data included in a local mini-seed archive representing the starting point of the procedure. SDQ finally provides summary tables for both earthquake and continuous data, collecting all relevant parameters for each processed waveform and data stream, along with explanatory text files (log and warning files), allowing the user to better evaluate the results. Although SDQ is currently under development, now it is freely available and  downloadable at  https://gitlab.rm.ingv.it/EIDA/quality/sdq

How to cite: Varchetta, F., Massa, M., Puglia, R., Danececk, P., Rao, S., Mandiello, A., and Piccinini, D.: SDQ (Seismic Data Quality): a Python project for seismo-accelerometric data quality check, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7986, https://doi.org/10.5194/egusphere-egu24-7986, 2024.

The Swedish National Seismic Network (SNSN) currently operates 67 permanent and 13 temporary broadband seismic stations. All stations transmit continuous realtime data to the data centre in Uppsala, and data streams of about 40 stations are automatically forwarded to subscribing institutes in the neighboring countries and to ORFEUS. In addition to the SNSN stations we receive realtime data from about 120 stations located in Norway, Finland, Denmark, Germany, Poland, the Baltic States, and Russia. 

SNSN processes the waveform data of this virtual network of about 200 stations using the SeisComp and Earthworm systems in parallel. Both systems are set up to be very sensitive in order to detect as small events as possible, which also increases the probability of generating spurious events. In order to screen out spurious events we generate a common bulletin which contains events that have been located by both systems independently. Our common bulletin is very reliable (no spurious events during the last 1.5 years), captures events down to about ML = 1 and contains almost all events with ML > 1.5 in Fennoscandia.

All events in the common bulletin are automatically classified by an artificial neural network as earthquakes, blasts or mining-induced events. The classifier has been developed in the framework of a PhD project, and was implemented into the SNSN processing queue during 2023 (Eggertsson et al, "Earthquake or Blast? Classification of Local-Distance Seismic Events in Sweden using Fully-Connected Neural Networks", accepted GJI 2024). It has been thoroughly tested, and, comparing the automatic classification with analyst-reviewed classification, we found a 97% match

Since December 2023, SNSN provides the automatic common bulletin as a simple webpage https://www.snsn.se/combullUTC/ - mainly for the general public and for quick reference. For the seismological community, SNSN has set up an automatic real-time forwarding of complete event parameters for all events with ML >= 2 to the European-Mediterranean Seismological Centre.

How to cite: Roth, M., Eggertsson, G., Schmidt, P., Shomali, H., Oskooi, B., and Lund, B.: The automatic seismic event bulletin of the Swedish National Seismic Network, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12258, https://doi.org/10.5194/egusphere-egu24-12258, 2024.

Large dense seismic networks popping up around the world in the last two decades enable studying the wave propagation and structure of the Earth with unprecedented details. Hundreds of broadband seismic stations spaced by tens of kilometers produce large amounts of data, which is usually processed by automatic routines. The data is no longer supervised by seismologists on a detailed level of every record as thousands of hours of data are handled at once. Ensuring the quality of data and accompanying metadata is nowadays a discipline by and of itself. Besides the classical techniques, which investigate the properties of data at a single station, large dense seismic networks allow for a multi-station approach to review the quality of the data. The diagnostic tools of multi-station methods are based on the detection of outlying stations/records among many others. Properties of the wavefield of wavelengths longer than the station spacing vary smoothly and hence comparing the measurement at neighboring stations allows for identifying anomalous behaviors. These methods work under the assumption that most of the (meta-) data is correct, and therefore a small number of outliers can be detected. Thus, not only do large dense seismic networks contribute to research, which is their primary goal, but thanks to the design of these networks, data quality can also be tested more precisely than before. We review both types of techniques, showing examples of the AlpArray experiment (2015 - 2022), and discussing the development of the approaches over the years to what is nowadays applied to the AdriaArray Seismic Network. AdriaArray experiment started in 2022 and encompasses now around 1000 permanent as well as over 430 temporary broadband stations. We show how the availability and retrievability of the data are checked, how amplitude and phase information from the ambient and deterministic wavefields are used to assess the correctness of metadata and how we represent the results of these tests in maps and tables. The purpose of these tests is aimed in both directions: towards the users, so that they are aware of potential issues, as well as towards the station operators so that they can be notified and asked to fix the detected problems.

How to cite: Kolínský, P., Vecsey, L., Stampa, J., Eckel, F., Belinić Topić, T., Working Group, T. A., and Seismology Group, T. A.: Data quality of large dense seismic networks – lessons learnt from AlpArray and application to AdriaArray, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12520, https://doi.org/10.5194/egusphere-egu24-12520, 2024.

MUDA (geophysical and geochemical MUltiparametric DAtabase) is a new infrastructure of the National Institute of Geophysics and Volcanology (INGV, www.ingv.it) serving geophysical and geochemical multiparametric data, designed and developped in the framework of Dynamic Planet -Working Earth project (https://progetti.ingv.it/it/pian-din).

MUDA is a dynamic and relational database based on MySQL (https://www.mysql.com) with a web interface realised in php (https://www.php.net) using a responsive design technique. The multi-parametric data are stored and organised using a table-structure able of correlating different types of data that allow possible future integration with new type of data acquired through both real-time and off-line transmission vectors.

MUDA collects information from different types of sensors, such as seismometers, accelerometers, hydrogeochemical sensors, sensors for measuring the flux of carbon dioxide on the ground (CO₂), sensors for detecting the concentration of Radon gas and weather stations with the aim of making possible correlations between seismic phenomena and variations in environmental parameters such as the level of groundwater as well as its temperature and electrical conductivity.

MUDA archives and publishes data of multiparametric stations belonging both to permanent (i.e. the National Seismic Network, RSN, https://www.fdsn.org/networks/detail/IV/) or temporary (e.g. PDnet, Massa et al., 2021, https://www.fdsn.org/networks/detail/ZO_2021/) INGV seismic networks, as well as data from a multi-parametric Salse di Nirano Reserve (MO) site in cooperation with the PD PROMUD 2023-2025 (Definition of a multidisciplinary monitoring PROtocol for MUD volcanoes) project and two additional multi-parametric sites installed in the inter-mountain basin of Norcia, as a part of the GEMME 2023-2025 project. Data from Radon stations belong to the INGV-IRON national network (Italian Radon Monitoring Network, https://www.ingv.it/en/monitoring-and-infrastructure/monitoring-networks/ingv-and-its-networks/iron).

MUDA daily publishes multi-parametric data updated to the previous day and offers the chence to view and download dynamic time series for all available data and for different periods, up to a maximum of 30 days. For longer periods, users can request data to muda@ingv.it.

MUDA is now published at http://muda.mi.ingv.it

How to cite: Massa, M., Ferrari, E., Rizzo, A., Lovati, S., and Luzi, L.: MUDA: the dynamic geophysical and geocehmical multiparametric database, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12895, https://doi.org/10.5194/egusphere-egu24-12895, 2024.

In the last 5 years, the seismological community has experienced an impressive growth in the novel Distributed Acoustic Sensing (DAS) technique, in terms of both the number of experiments and the generated data volume. DAS experiments can generate data at a much finer resolution in space and time, than is seen with standard acquisition techniques. This creates challenges not only for data centres regarding the data management, but also for users that need to access and process this data. The current seismological standards for data and metadata formats, as well as community services specifications, are not capable of handling these datasets in an effective way - not unexpected considering that data volumes and ‘station’ numbers are orders of magnitude larger than typical broad-band experiments.

Within the context of the “Geosphere INfrastructures for QUestions into Integrated REsearch” project (Geo-INQUIRE, https://www.geo-inquire.eu/), we have defined a roadmap to advance towards community standards for some of these aspects. The main objective of improving the FAIRness of these datasets was separated in 3 steps. First, we defined how to archive downsampled versions of the datasets in standard community formats (i.e. miniseed and StationXML). Second, we wanted to support the definition (and foster the adoption) of a new metadata standard for DAS experiments based on the outcome of the DAS Research Coordination Network group (DAS-RCN), an initiative led by US researchers. And finally, we wanted to work on the definition of a data format capable of providing fast processing on the data centre side, as well as being able to provide the data to the user to be processed elsewhere.

We worked with 3 datasets from the Global DAS Month (February 2023), acquired by INGV, ETH and GFZ. These datasets had been published and made available in different non-standard formats. We used these experiments as test cases to later apply this workflow to the datasets generated by the Transnational Access Calls of this project at a variety of Research Infrastructures across Europe (e.g. at Etna, Bedretto, Ligurian, Madeira, Irpinia, and others).

Regarding the data volume and lack of standardisation, we have improved “dastools”, a software package developed at GFZ, to read DAS data in proprietary formats from different manufacturers and convert it to standard miniseed. Downsampling in time and space it provides a reduced version ready to be archived in seismological data centres.

Regarding metadata formats, we included in “dastools” the support for the DAS-RCN proposal, discussed and agreed within the community during the last 3 years. We can generate a first draft version of the metadata based on the information available in the raw data of the experiment. We also added a converter to StationXML (still beta) in order to support each step of the archival of a downsampled version of the DAS data.

We plan to work soon on the definition of a data format for this type of experiment as it is a key part of our project.

In parallel, we’ve just started the development of a Seedlink plugin (real-time transmission) to be deployed and tested at interrogators.

How to cite: Quinteros, J., Currenti, G., Prestifilippo, M., Clinton, J., Edme, P., Wollin, C., Rivet, D., Murphy, S., Schaeffer, J., Pedersen, H., and Strollo, A.: Towards a community standard for DAS metadata: Latest advances within the Geo-INQUIRE project, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13131, https://doi.org/10.5194/egusphere-egu24-13131, 2024.

Seismic activity in Ukraine varies across regions with notable active zones in the Carpathians and Crimean-Black Sea segments. Southwest Ukraine is affected by the Vrancea seismic zone (Romania) and demands attention, alongside rare but potentially powerful seismic events in the stable Ukrainian shield. Mining and industrial activities also induce ground vibrations (Mw 2.5–4) in otherwise stable areas. Although major destructive earthquakes in Ukraine are infrequent, the cumulative impact of smaller seismic events can shape the geological and geophysical landscape of the region.

An effective seismic network is crucial for safety and research in Ukraine. Yet, the existing seismic network at the Institute of Geophysics of National Academy of Science of Ukraine faces numerous challenges, limiting its capacity to provide accurate seismic information. In addition, the Institute of Geophysics faces a demographic imbalance, with a critical shortage of younger scientists entering the field. This knowledge gap poses implications for the future of scientific research in Ukraine.

In response to these challenges, and considering the Russian invasion, we opted to distribute 28 budget Raspberry Shake Seismographs across schools and universities in Ukraine. Initially, these budget seismometers serve as a short-term solution for seismic data collection. Acknowledging the pros and cons of these stations in contrast to broadband sensors, it's noteworthy that their simplicity in installation, low cost, and near-real-time data transmission make them as a suitable option during the conflict in identifying and characterizing local and regional events.

This initiative also directly supports science teaching from middle to high school in Ukraine. Integrating seismometers into schools cultivates education based on real-time seismic records, familiarizing students with scientific data. The aim is to ignite students’ interest, nurturing a curiosity not only in seismology but also in science as a whole. This goal is accomplished not just through presentations and lessons but also through hands-on involvement, allowing students to take ownership by installing a seismometer and consistently monitoring its output. The analysis of time series seismic signals, regardless of their source – whether earthquakes or artificial noise – forms a fundamental component of any seismology-focused educational program. Currently, such initiatives are lacking in Ukrainian educational institutions.

Raspberry Shake 3D Seismographs have been installed at the Institute of Geophysics in Kyiv, Lviv Polytechnic National University, and Ivan Franko National University of Lviv. Two RS seismometers are at the Mykolaiv Water Hub to support the establishment of the Mykolaiv Innovation Lab in south Ukraine. In November 2023, educational materials for seismology at middle and high school levels in Ukraine were created with the assistance of the GFZ German Research Centre for Geosciences. This approach not only will help students develop practical skills but also provides a starting point for exploring numerical methods and coding at the university level.

How to cite: Amashukeli, T., Malatesta, L., Farfuliak, L., Haniiev, O., Kuplovskyi, B., Prokopyshyn, V., and Petrenko, K.: The prospects of the Ukrainian seismic network:  Introducing a Low-Cost Seismic Network Initiative in Ukraine, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13158, https://doi.org/10.5194/egusphere-egu24-13158, 2024.

The Republic of Armenia and the neighboring areas are located in the central part of theArabia-Eurasia collision zone, which is characterized by active seismicity.Over the past decade, the Institute of Geological Sciences (IGS) of NAS RA has developed an advanced and dense seismic network.To develop the local seismic network, the Institute of Geological Sciences in collaboration with the National Taiwan University and Institute of Earth Sciences, Academia Sinica (Taiwan) established 11 seismic stations with broadband seismometers and 6 cGPS (continuous) stations. Data from the existing seismic stations were collected, archived and treated, and main earthquake parameters were determined; the conducted works included catalogue maintenance, recalculation of the main earthquake data, calculation of seismic tomography with the input of recalculated and updated data, and adjustment of the three-dimensional (3D) velocity model for the area of the RA.

Starting from 2017, the Institute of Geological Sciences has participated in two partner ISTC projects: A-2334 Project (Transect) “The Uplift and Seismic Structure of the Greater Caucasus” and KR-2452 Project (SNECCA) “Seismic Network Expansion in the Caucasus and Central Asia” with support from the U.S. Department of Energy.  In the framework of the indicated projects, 32 seismic stations with broadband seismometers were established along with 8 seismic borehole stations with fully broadband seismometers (Real Time) and 8 strong motion sensors. These two projects are implemented through the Seismic Targeted Initiative of the International Science and Technology Center and the Science and Technology Center in Ukraine.The records collected from the eight (8) already operational permanent seismic stations of the indicated network are in real-time mode sent to the International Seismology Center of the IRIS (Incorporated Research Institutions for Seismology). To download and archive the database of seismic stations, a new computer server was acquired under the project, and all required configurations were made to provide for its uninterrupted operation. Seiscomp software set is applied to produce automatic solutions for earthquakes recorded within the region. Adjustments and recalculations of the automatic earthquake solutions are implemented to produce the resulting main earthquake parameters with uncertainties reduced to the rate as low as possible. The collected data from all existing stations are widely used to determine and to re-estimate the main parameters of earthquake occurring in Armenia and in surrounding territories. These data are included in local bulletins and catalogues.For strong motions, the on-going research includes processing of design accelerograms, preparation and analysis of hazard response spectra (RS), selection of the actually recorded (real) acceleration time-history, and spectral matching of the time-history to the hazard response spectrum with application of Seismosoft (Earthquake Engineering Software Solutions, SeismoMatch, SeismoSignal, SeismoSelect, SeismoSpec) software set.All the results obtained from the data seismic stations mentioned above are used for the purposes of scientific research and are summarized in articles.

How to cite: Babayan, G., Sahakyan, E., Babayan, H., Gevorgyan, M., and Sargsyan, L.: Development of the new seismic network of the IGS, Armenia, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14842, https://doi.org/10.5194/egusphere-egu24-14842, 2024.

Güralp Data Centre (GDC) interface offers ‘one click’ tools to configure instruments to stream data to a central (typically cloud based) server where it is saved in miniSEED in configurable folder structures. This application is particularly important for operators dealing with large volumes of seismic waveform data from regional/national networks.

Additionally, the data can be transmitted to downstream processors such as Earthworm or SeisComP for more advanced seismic monitoring and data analysis. GDC has a simple interface to set up and monitor the operation of the network and is easy to implement into existing systems and networks with minimal configuration as industry standard protocols are employed throughout.

An integrated VPN/Tunnel circumvents Network Address Translations (NATs) present in internet modems and ADSL connections, providing the facility to remotely update digitizer firmware and upload configuration files to multiple units simultaneously.

Long term latency monitoring, network outages and bandwidth usage are captured and displayed in a number of applets that further simplify maintenance of large networks. The GDC dashboard allows network managers to view data integrity over time so that latency performance can be monitored.

Trigger events from instruments can be recorded and displayed on a map as part of a range of features dedicated to EEW implementations. This information is conveyed using the open Common Alert Protocol (CAP). The CAP messages are created by individual station or sub-network triggers and contain important parameters such the on-site recorded PGA, PGV and PGD, providing the lowest possible latency for network early warning.

How to cite: Watkiss, N., Hill, P., Price, E., Foster, C., Clark, A., Restelli, F., and Lindsey, J.: Güralp Data Centre Software for Easy Mass Data Acquisition and Station Metadata Monitoring, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17386, https://doi.org/10.5194/egusphere-egu24-17386, 2024.

Seismic signals coupling to physical (e.g., temperature, pH, Eh, electrical conductivity, flow rate) and geochemical changes in ground and spring waters as well as variations in soil flux regimes (e.g., CO₂, CH₄, radon) represent a valuable tool to better understand the interaction between tectonics and crustal fluids dynamics (e.g., Italiano et al., 2001, 2004; Wang and Manga, 2021; Chiodini et al., 2020; Gori and Barberio, 2022 and references therein). Pre-, co- and post-seismic modifications are markers of the local tectonic stress acting in the crust and are extremely site-specific due to the local geological and lithological features besides being simultaneously influenced by other environmental conditions (e.g., meteorological and climatic). Therefore, local continuous monitoring of all the involved parameters is needed to delineate crustal fluids response to seismicity site by site.
Multiparametric stations have been set up in Italy starting from the end of 2021, placed on the major seismogenic structures, and widely distributed among the Alps, Apennines and Pianura Padana. They are equipped with: (i) sensors installed in water wells measuring water level, temperature, and electrical conductivity; (ii) meteorological sensors measuring atmospheric pressure, temperature, rain, humidity, wind speed and direction; (iii) seismic sensors providing accelerometric and velocimetric datasets; (iv) radon sensors; (v) CO₂ soil flux chamber. 
Data are transmitted in near real-time to an ad hoc developed dynamic relational database (MUDA-geophysical and geochemical MUltiparametric DAtabase) and displayed in a dedicated website (http://muda.mi.ingv.it). The built-in philosophy is to easily compare distinct parameters from the various sensors and possibly recognize cause-effect relationships among them. 
To our knowledge, our new multiparametric network is the first developed in Italy showing all these features.
A statistic approach is also applied to the time-series to investigate intra-annual and inter-annual trends and correlations among different parameters. Alternative methods (e.g., signal decomposition, spike detection) will be presented and discussed. 

References
-Chiodini G., Cardellini C., Di Luccio F., Selva J., Frondini F., Caliro S., Rosiello A., Beddini G., Ventura G., 2020: Correlation between tectonic CO2 Earth degassing and seismicity is revealed by a 10-year record in the Apennines, Italy. Science Advances, https://www.science.org/doi/10.1126/sciadv.abc2938
-Gori F., Barberio M.D., 2022: Hydrogeochemical changes before and during the 2019 Benevento seismic swarm in central-southern Italy. Journal of Hydrology, 604:127250
-Italiano F., Martinelli G., Nuccio P.M., 2001: Anomalies of mantle-derived helium during the 1997 – 1998 seismic swarm of Umbria-Marche, Italy. Geophysical Research Letters, 28(5):839-842
-Italiano F., Martinelli G., Rizzo A., 2004: Geochemical evidence of seismogenic-induced anomalies in the dissolved gases of thermal waters: A case study of Umbria (Central Apennines, Italy) both during and after the 1997–1998 seismic swarm. Geochemistry, Geophysics, Geosystems, 5:11, doi:10.1029/2004GC000720
-Wang C.-Y., Manga M., 2021: Water and Earthquakes. Lecture Notes in Earth System Sciences, Springer Cham, 387 pp., https://doi.org/10.1007/978-3-030-64308-9 

How to cite: Ferrari, E., Massa, M., Rizzo, A. L., Lovati, S., and Di Michele, F.: The Italian multiparametric network for detection and monitoring of earthquake-related crustal fluids alterations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17749, https://doi.org/10.5194/egusphere-egu24-17749, 2024.

We report on establishing the Finnish mobile seismic instrument pool that is owned and operated by seven Finnish academic and research institutions. The pool infrastructure is funded by the Research Council of Finland, through the FLEX-EPOS project and under the FIN-EPOS umbrella. It is financing the build-up stage and started in 2021 with an end in 2024. By then the seismic instrumentation is anticipated to include 46 Güralp broadband seismometers, 5 Güralp accelerometers, and 1197 and 70 Geospace and SmartSolo self-contained geophone units, respectively. It is making this probably the largest coherent mobile seismic instrument pool in Europe in the public sector. The pool supports domestic and international collaborative projects of temporary deployments to enhance data-driven subsurface and environmental applications. Those deployments are for active or passive experiments and can last a few days up to a few years. The acquisition of such pool is motivated and facilitated by the advent of efficient data storage and transmission and powerful computing systems; progress in the understanding of the seismic wavefield coupled with the development of new types of analysis techniques and algorithms; and the manufacturing of sensitive, affordable data-dense sensor systems. Despite these game-changing and promising developments, the access to many seismic sensors for large-N deployments is not pervasive. Even in developed countries, it is challenging for a single institution to acquire and maintain a sufficiently large mobile pool of instruments and ensure sustainable data production and distribution. Our report on the equipment, facilities, ownership, and governance structure, project management, and data systems is essential background information for the access to and utilization of the pool instruments, and the interaction with the support community. A discussion about best practices for establishing and maintaining a mobile pool infrastructure can benefit from our experience of building such an extensive public seismic infrastructure from the ground up, and it provides relevant information for communities considering similar research infrastructure projects. Among the many challenges and opportunities associated with establishing effective pool management and operations, we highlight the lack of a coherent community protocol to store, find, disseminate, and analyze the associated large datasets. The Finnish mobile seismic instrument pool actively engages with ORFEUS/EIDA and the Geo-INQUIRE project to contribute to developing community solutions for data discovery and accessibility.

How to cite: Courbis, R., Hillers, G., Koivisto, E., Haapanala, P., Kukkonen, I., Ding, Y., Fordell, T., Heinonen, S., Junno, N., Juntunen, A., Komminaho, K., Kozlovskaya, E., Leveinen, J., Moisio, K., Näränen, J., Oksanen, T., Skyttä, P., Tanskanen, E., and Tiira, T.: The mobile Finnish Seismic Instrument Pool, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17890, https://doi.org/10.5194/egusphere-egu24-17890, 2024.

In the analysis of sub-annual climatological or geodetic time series such as tide gauges, precipitable water vapor, or GNSS vertical displacements time series but also temperatures or gases concentrations, seasonal cycles are often found to have a time-varying amplitude and phase.

These time series are usually modelled with a deterministic approach that includes trend, annual, and semi-annual periodic components having constant amplitude and phase-lag. This approach can potentially lead to inadequate interpretations, such as an overestimation of Global Navigation Satellite System (GNSS) station velocity, up to masking important geophysical phenomena that are related to the amplitude variability and are important for deriving trustworthy interpretation for climate change assessment.

We address that challenge by proposing a novel linear additive model called the fractional Sinusoidal Waveform process (fSWp), accounting for possible nonstationary cyclical long memory, a stochastic trend that can evolve over time and an additional serially correlated noise capturing the short-term variability. The model has a state space representation and makes use of the Kalman filter (KF). Suitable enhancements of the basic methodology enable handling data gaps, outliers, and offsets. We demonstrate our method using various climatological and geodetic time series to illustrate its potential to capture the time-varying stochastic seasonal signals.

How to cite: Kermarrec, G., Maddanu, F., Klos, A., and Proietti, T.: The fractional Sinusoidal wavefront Model (fSwp) for time series displaying persistent stationary cycles, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-24, https://doi.org/10.5194/egusphere-egu24-24, 2024.

On some maps of the first military survey of the Habsburg Empire, the upper direction of the sections does not face the cartographic north, but makes an angle of about 15° with it. This may be due to the fact that the sections were subsequently rotated to the magnetic north of the time. Basically, neither their projection nor their projection origin is known yet.

In my research, I am dealing with maps of Inner Austria, the Principality of Transylvania and Galicia (nowadays Poland and Ukraine), and I am trying to determine their projection origin. For this purpose, it is assumed, based on the archival documentation of the survey, that these are Cassini projection maps. My hypothesis is that they are Graz, Cluj Napoca or Alba Julia and Lviv. I also consider the position of Vienna in each case, since it was the main centre of the survey.

The angle of rotation was taken in part from the gufm1 historical magnetic model back to 1590 for the assumed starting points and year of mapping. In addition, as a theoretical case, I calculated the rotation angle of the map sections using coordinate geometry. I then calculated the longitude of the projection starting point for each case using univariate minimization. Since the method is invariant to latitude, it can only be determined from archival data.

Based on these, the starting point for Inner Austria from the rotation of the map was Vienna, which is not excluded by the archival sources, and since the baseline through Graz also started from there, it is partly logical. The map rotation for Galicia and Transylvania also confirmed the starting point of the hypothesis.  Since both Alba Julia and Cluj Napoca lie at about the same longitude, the method cannot make a difference there; and the archival data did not provide enough evidence. In comparison, the magnetic declination rotations yielded differences of about 1°, which may be due to an error in the magnetic model.

On this basis, I have given the assumed projections of the three maps with projection starting points, and developed a method for determining the projection starting points of the other rotated grid maps. The results suggest that there is a very high probability that the section network was rotated in the magnetic north direction, and thus provide a way to refine the magnetic declination data at that time.

With this method I managed to give new indirekt magnetic declinations data from Central-East Europe, which can help to improve the historical magnetic field models. The main reason for this is that we don’t have any measurement from that region.

Furthermore the difference beetwen the angle of the section north and the declination data from gufm1 always 0.8-1°. Maybe there are systematical data error at that region.

Supported by the ÚNKP-23-6 New National Excellence Program of the Ministry for Culture and Innovation from the source of the National Research, Development and Innovation Fund.

How to cite: Koszta, B. and Timár, G.: A possible cartographical data source for historical magnetic field improvement: The direction of the section north of the Habsburg first military survey, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-582, https://doi.org/10.5194/egusphere-egu24-582, 2024.

The alteration of traditional hydrological patterns due to global warming is leading to a modification of the hydrological cycle. This situation draws a complex scenario for the sustainable management of water resources. However, this issue offers a challenge for the development of innovative approaches that allow an in-depth capturing the logical temporal-dependence structure of these modifications to advance sustainable management of water resources, mainly through the reliable predictive models. In this context, Bayesian Causality (BC), addressed through Causal Reasoning (CR) and supported by a Bayesian Networks (BNs), called Bayesian Causal Reasoning (BCR) is a novel hydrological research area that can help identify those temporal interactions efficiently.

This contribution aims to assesses the BCR ability to discover the logical and non-trivial temporal-dependence structure of the hydrological series, as well as its predictability. For this, a BN that conceptually synthesizes the time series is defined, and where the conditional probability is propagated over the time throughout the BN through an innovative Dependence Mitigation Graph. This is done by coupling among an autoregressive parametric approach and causal model. The analytical ability of the BCR highlighted the logical temporal structure, latent in the time series, which defines the general behavior of the runoff. This logical structure allowed to quantify, through a dependence matrix which summarizes the strength of the temporal dependencies, the two temporal fractions that compose the runoff: one due to time (Temporally Conditioned Runoff) and one not (Temporally Non-conditioned Runoff). Based on this temporal conditionality, a predictive model is implemented for each temporal fraction, and its reliability is assessed from a double probabilistic and metrological perspective.

This methodological framework is applied to two Spanish unregulated sub-basins; Voltoya river belongs to Duero River Basin, and Mijares river, in the Jucar River Basin. Both cases with a clearly opposite temporal behavior, Voltoya independent and Mijares dependent, and with increasingly more problems associated with droughts.

The findings of this study may have important implications over the knowledge of temporal behavior of water resources of river basin and their adaptation. In addition, TCR and TNCR predictive models would allow advances in the optimal dimensioning of storage infrastructures (reservoirs), with relevant substantial economic/environmental savings. Also, a more sustainable management of river basins through more reliable control reservoirs’ operation is expected to be achieved. Finally, these results open new possibilities for developing predictive hydrological models within a BCR framework.

How to cite: Zazo, S., Molina, J. L., Patino-Alonso, C., and Espejo, F.: Predictive ability assessment of Bayesian Causal Reasoning (BCR) on runoff temporal series, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1988, https://doi.org/10.5194/egusphere-egu24-1988, 2024.

Climate-induced forest mortality poses an increasing threat worldwide, which calls for developing robust approaches to generate early warning signals of upcoming forest state change. This research explores the potential of satellite imagery, utilizing advanced spatio-temporal indicators and methodologies, to assess the state of forests preceding mortality events. Traditional approaches, such as techniques based on temporal analyses, are impacted by limitations related to window size selection and detrending methods, potentially leading to false alarms. To tackle these challenges, our study introduces two new approaches, namely the Spatial-Temporal Moran (STM) and Spatial-Temporal Geary (STG) approaches, both focusing on local spatial autocorrelation measures. These approaches can effectively address the shortcomings inherent in traditional methods. The research findings were assessed across three study sites within California national parks, and Kendall's tau was employed to quantify the significance of false and positive alarms. To facilitate the measurement of ecosystem state change, trend estimation, and identification of early warning signals, this study also provides "stew" R package. The implications of this research extend to various groups, such as ecologists, conservation practitioners, and policymakers, providing them with the means to address emerging environmental challenges in global forest ecosystems.

How to cite: Alibakhshi, S.: Spatial-Temporal Analysis of Forest Mortality, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3857, https://doi.org/10.5194/egusphere-egu24-3857, 2024.

Iram Parvez¹, Massimiliano Cannata², Giorgio Boni¹, Rossella Bovolenta¹ ,Eva Riccomagno³ , Bianca Federici¹

¹ Department of Civil, Chemical and Environmental Engineering (DICCA), Università degli Studi di Genova, Via Montallegro 1, 16145 Genoa, Italy (iram.parvez@edu.unige.it,bianca.federici@unige.it, giorgio.boni@unige.it, rossella.bovolenta@unige.it).

² Institute of Earth Sciences (IST), Department for Environment Constructions and Design (DACD), University of Applied Sciences and Arts of Southern Switzerland (SUPSI), CH-6952 Canobbio, Switzerland(massimiliano.cannata@supsi.ch).

³ Department of Mathematics, Università degli Studi di Genova, Via Dodecaneso 35, 16146 Genova, Italy(riccomag@dima.unige.it).

The deployment of hydrometeorological sensors significantly contributes to generating real-time big data. The quality and reliability of large datasets pose considerable challenges, as flawed analyses and decision-making processes can result. This research aims to address the issue of anomaly detection in real-time data by exploring machine learning models. Time-series data is collected from IstSOS - Sensor Observation Service, an open-source software that stores, collects and disseminates sensor data. The methodology consists of Gated Recurrent Units based on recurrent neural networks, along with corresponding prediction intervals, applied both to individual sensors and collectively across all temperature sensors within the Ticino region of Switzerland. Additionally, non-parametric methods like Bootstrap and Mean absolute deviation are employed instead of standard prediction intervals to tackle the non-normality of the data. The results indicate that Gated Recurrent Units based on recurrent neural networks, coupled with non-parametric forecast intervals, perform well in identifying erroneous data points. The application of the model on multivariate time series-sensor data establishes a pattern or baseline of normal behavior for the area (Ticino). When a new sensor is installed in the same region, the recognized pattern is used as a reference to identify outliers in the data gathered from the new sensor.

How to cite: Parvez, I.: Exploring Machine Learning Models to Detect Outliers in HydroMet Sensors, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4280, https://doi.org/10.5194/egusphere-egu24-4280, 2024.

For its immunity to post-formation geological modifications, zircon is widely utilized as chronological time capsule and provides critical time series data potential to unravel key events in Earth’s geological history, such as supercontinent cycles. Fourier analysis, which assumes stationary periodicity, has been applied to zircon-derived time series data to find the cyclicity of supercontinents, and wavelet analysis, which assumes non-stationary periodicity, corroborates the results of Fourier Analysis in addition to detecting finer-scale signals. Nonetheless, both methods still prognostically assume periodicity in the zircon-derived time-domain data. To stay away from the periodicity assumption and extract more objective information from zircon data, we opt for a Bayesian approach and treat zircon preservation as a composite stochastic process where the number of preserved zircon grains per magmatic event obeys logarithmic series distribution and the number of magmatic events during a geological time interval obeys Poisson distribution. An analytical solution was found to allow us to efficiently invert for the number and distribution(s) of changepoints hidden in the globally compiled zircon data, as well as for the zircon preservation potential (encoded as a model parameter) between two neighboring changepoints. If the distributions of changepoints temporally overlap with those of known supercontinents, then our results serve as an independent, mathematically robust test of the cyclicity of supercontinents. Moreover, our statistical approach inherently provides a sensitivity parameter the tuning of which allows to probe changepoints at various temporal resolution. The constructed Bayesian framework is thus of significant potential to detect other types of trend swings in Earth’s history, such as shift of geodynamic regimes, moving beyond cyclicity detection which limits the application of conventional Fourier/Wavelet analysis.

How to cite: Qian, H., Tian, M., and Zhang, N.: Unveiling Geological Patterns: Bayesian Exploration of Zircon-Derived Time Series Data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5268, https://doi.org/10.5194/egusphere-egu24-5268, 2024.

Semi-enclosed freshwater and brackish ecosystems, characterised by restricted water outflow and prolonged residence times, often accumulate nutrients, influencing their productivity and ecological dynamics. These ecosystems exhibit significant variations in bio-physical-chemical attributes, ecological importance, and susceptibility to human impacts. Untangling the complexities of their interactions remains challenging, necessitating a deeper understanding of effective management strategies adapted to their vulnerabilities. This research focuses on the bio-physical aspects, investigating the differential effects of spring and summer light on phytoplankton communities in semi-enclosed freshwater and brackish aquatic ecosystems.

Through extensive field sampling and comprehensive environmental parameter analysis, we explore how phytoplankton respond to varying light conditions in these distinct environments. Sampling campaigns were conducted at Müggelsee, a freshwater lake on Berlin's eastern edge, and Barther Bodden, a coastal lagoon northeast of Rostock on the German Baltic Sea coast, during the springs and summers of 2022 and 2023, respectively. Our analysis integrates environmental factors such as surface light intensity, diffuse attenuation coefficients, nutrient availability, water column dynamics, meteorological data, Chlorophyll-a concentration, and phytoplankton communities. Sampling encompassed multiple depths at continuous intervals lasting three days.

Preliminary findings underscore significant differences in seasonal light availability, with summer exhibiting extended periods of substantial light penetration. These variations seem to impact phytoplankton abundance and diversity uniquely in each ecosystem. While ongoing analyses are underway, early indications suggest distinct phytoplankton responses in terms of species composition and community structure, influenced by the changing light levels. In 2022 the clear water phase during spring indicated that bloom events have occurred under ice cover much earlier than spring, while in the summer there were weak and short-lived blooms of cyanobacteria. The relationship between nutrient availability and phytoplankton dynamics, however, remains uncertain according to our data.

This ongoing study contributes to understanding the role of light as a primary driver shaping phytoplankton community structures and dynamics in these environments.  Our research findings offer insights for refining predictive models, aiding in ecosystem-specific eutrophication management strategies, and supporting monitoring efforts of Harmful Algal Blooms.

How to cite: Kaharuddin, A. and Kaligatla, R.: Comparative Study of Spring and Summer Light Effects on Phytoplankton Communities in Semi-Enclosed Fresh- and Brackish Aquatic Ecosystems., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5733, https://doi.org/10.5194/egusphere-egu24-5733, 2024.

Forecasting the aurora and its location accurately is important to mitigate any potential harm to vital infrastructure like communications and electricity grid networks. Current auroral prediction models rely on our understanding of the interaction between the magnetosphere and the solar wind or geomagnetic indices. Both approaches do well in predicting but have limitations concerning forecasting (geomagnetic indices-based model) or because of the underlying assumptions driving the model (due to a simplification of the complex interaction). By applying machine learning algorithms to this problem, gaps in our understanding can be identified, investigated, and closed. Finding the important time scales for driving empirical models provides the necessary basis for our long-term goal of predicting the aurora using machine learning.

Periodicities of the Earth’s magnetic field have been extensively studied on a global scale or in regional case studies. Using a suite of different time series analysis techniques including frequency analysis and investigation of long-scale changes of the median/ mean, we examine the dominant periodicities of ground magnetic field measurements at selected locations. A selected number of stations from the SuperMAG network (Gjerloev, 2012), which is a global network of magnetometer stations across the world, are the focus of this investigation.

The periodicities retrieved from the different magnetic field components are compared to each other as well as to other locations. In the context of auroral predictions, an analysis of the dominating periodicities in the auroral boundary data derived from the IMAGE satellite (Chisham et al., 2022) provides a counterpart to the magnetic field periodicities.

Ultimately, we can constrain the length of time history sensible for forecasting.

How to cite: Gilmore, K. R., Bentley, S. N., and Smith, A. W.: Magnetospheric time history:  How much do we need for forecasting?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6065, https://doi.org/10.5194/egusphere-egu24-6065, 2024.

Dynamical systems can display a range of dynamical regimes (e.g. attraction to, fixed points, limit cycles, intermittency, chaotic behaviour) depending on the values of parameters in the system. In this work we demonstrate how non-parametric entropy estimation codes (in particular NPEET) based on the Kraskov method can be applied to find regime transitions in a 3D chaotic model (the Lorenz 1963 system) when varying the values of the parameters. These infromation-theory-based methods are simpler and cheaper to apply than more traditional metrics from dynamical systems (e.g. computation of Lyapunov exponents). The non-parametric nature of the method allows for handling long time series without a prohibitive computational burden. 

How to cite: Amezcua, J. and Chakraborty, N.: Using information-theory metrics to detect regime changes in dynamical systems, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6151, https://doi.org/10.5194/egusphere-egu24-6151, 2024.

Evaluation of scaling properties and fractal formalisms is one of the potential approaches for modelling complex series. Understanding the complexity and fractal characterization of drought index time series is essential for better preparedness against drought disasters. This study presents a novel visibility graph-based evaluation of fractal characterization of droughts of three meteorological subdivisions of India. In this method, the horizontal visibility graph (HVG) and Upside-down visibility graph (UDVG) are used for evaluating the network properties for different standardized precipitation index (SPI) series of 3, 6 and 12 month time scales representing short, medium and long term droughts. The relative magnitude of fractal estimates is controlled by the drought characteristics of wet-dry transitions. The estimates of degree distribution clearly deciphered the self-similar properties of droughts of all the subdivisions. For an insightful depiction of drought dynamics, the fractal exponents and spectrum are evaluated by the concurrent application of Sand Box Method (SBM) and Chhabra and Jenson Method (CJM). The analysis was performed for overall series along with the pre- and post-1976-77 Global climate shift scenarios. The complexity is more evident in short term drought series and UDVG formulations implied higher fractal exponents for different moment orders irrespective of drought type and locations considered in this study. Useful insights on the relationship between complex network and fractality are evolved from the study, which may help in improved drought forecasting. The visibility graph based fractality estimation evaluation is efficient in capturing drought and it has vast potential in the drought predictions in a changing environment.

Keywords:  Drought, Fractal, SPI, Visibility Graph

How to cite: Rajesh, S. M., Bahuleyan, M., Nair GR, A., and Sankaran, A.: Fractal complexity evaluation of meteorological droughts over three Indian subdivisions using visibility Graphs, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9367, https://doi.org/10.5194/egusphere-egu24-9367, 2024.

The Wavelet-Induced Mode Extraction procedure (WIME) [2] was developed drawing inspiration from Empirical Mode Decomposition. The concept involves decomposing the signal into modes, each presenting a characteristic frequency, using continuous wavelet transform. This method has yielded intriguing results in climatology [3,4]. However, the initial algorithm did not account for the potential existence of slight frequency fluctuations within a mode, which could impact the reconstruction of the original signal [4]. The new version (https://atoms.scilab.org/toolboxes/toolbox_WIME/0.1.0) now allows for the evolution of a mode in the space-frequency half-plane, thus considering the frequency evolution of a mode [2]. A natural application of this tool is in the analysis of Milankovitch cycles, where subtle changes have been observed throughout history. The method also refines the study of solar activity, highlighting the role of the "Solar Flip-Flop." Additionally, the examination of temperature time series confirms the existence of cycles around 2.5 years. It is now possible to attempt to correlate solar activity with this observed temperature cycle, as seen in speleothem records [1].

[1] Allan, M., Deliège, A., Verheyden, S., Nicolay S. and Fagel, N. Evidence for solar influence in a Holocene speleothem record, Quaternary Science Reviews, 2018.
[2] Deliège, A. and Nicolay, S., Extracting oscillating components from nonstationary time series: A wavelet-induced method, Physical Review. E, 2017.
[3] Nicolay, S., Mabille, G., Fettweis, X. and Erpicum, M., A statistical validation for the cycles found in air temperature data using a Morlet wavelet-based method, Nonlinear Processes in Geophysics, 2010.
[4] Nicolay, S., Mabille, G., Fettweis, X. and Erpicum, M., 30 and 43 months period cycles found in air temperature time series using the Morlet wavelet, Climate Dynamics, 2009.

How to cite: Faulx, E., Fettweis, X., Mabille, G., and Nicolay, S.: Wavelet-Induced Mode Extraction procedure: Application to climatic data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9537, https://doi.org/10.5194/egusphere-egu24-9537, 2024.

A wide range of geoprocesses manifest as observable events in a variety of contexts, including shifts in palaeoclimate regimes, evolutionary milestones, tectonic activities, and more. Many prominent research questions, such as synchronisation analysis or power spectrum estimation of discrete data, pose considerable challenges to linear tools. We present recent advances using a specific similarity measure for discrete data and the method of recurrence plots for different applications in the field of highly discrete event data. We illustrate their potential for palaeoclimate studies, particularly in detecting synchronisation between signals of discrete extreme events and continuous signals, estimating power spectra of spiky signals, and analysing data with irregular sampling.

How to cite: Marwan, N. and Braun, T.: New concepts on quantifying event data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10258, https://doi.org/10.5194/egusphere-egu24-10258, 2024.

Nowadays photovoltaic panels are becoming more affordable, efficient, and popular due to their low carbon footprint. PV panels can be installed in many places providing green energy to the local grid reducing energy cost and transmission losses. Since the PV production is highly dependent on the weather conditions, it is extremely important to estimate expected output in advance in order to maintain energy balance in the grid and provide enough time to schedule load distribution. The PV production output can be calculated by various statistical and machine learning prediction methods. In general, the more data available, the more precise predictions can be produced. This poses a problem for recently installed PV panels for which not enough data has been collected or the collected data are incomplete. 

A possible solution to the problem can be the application of an approach called Transfer Learning which has the inherent ability to effectively deal with missing or insufficient amounts of data. Basically, Transfer Learning is a machine learning approach which offers the capability of transferring knowledge acquired from the source domain (in our case a PV panel with a large amount of historical data) to different target domains (PV panels with very little collected historical data) to resolve related problems (provide reliable PV production predictions). 

In our study, we investigate the application, benefits and drawbacks of Transfer Learning for one day ahead PV production prediction. The model used in the study is based on complex neural network architecture, feature engineering and data selection. Moreover, we focus on the exploration of multiple approaches of adjusting weights in the target model retraining process which affect the minimum amount of training data required, final prediction accuracy and model’s overall robustness. Our models use historical meteorological forecasts from Deutscher Wetterdienst (DWD) and photovoltaic measurements from the project PVOutput which collects data from installed solar systems across the globe. Evaluation is performed on more than 100 installed PV panels in Central Europe.

How to cite: Lóderer, M., Sandanus, M., Pavlík, P., and Rozinajová, V.: Application of Transfer Learning techniques in one day ahead PV production prediction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10415, https://doi.org/10.5194/egusphere-egu24-10415, 2024.

In each magnetic observatory, the magnetic field is registered in local Cartesian coordinate systems associated with the geographic coordinates of the locations of these observatories. To observe extraterrestrial magnetic field sources, such as the interplanetary magnetic field or magnetic clouds, a method of joint processing of data from magnetic observatories of the international Intermagnet network was implemented. In this method, the constant component is removed from the observation results of individual observatories, their measurement data is converted into the ecliptic coordinate system, and the results obtained from all observatories are averaged after the coordinate transformation.

The first data on joint processing of measurement results from the international network of Intermagnet magnetic observatories in the period before the onset of magnetic storms of various types, during these storms and after their end are presented. There is a significant improvement in the signal-to-noise ratio after combining the measurement results from all observatories, which makes it possible to isolate weaker external magnetic fields. A change in the shape of magnetic field variations is shown, which can provide new knowledge about the mechanism of development of magnetic storms.

How to cite: Zhumabayev, B., Vassilyev, I., Mendakulov, Z., Fedulina, I., and Kapytin, V.: Results of joint processing of magnetic observatory data of international Intermagnet network in a unified coordinate system, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11897, https://doi.org/10.5194/egusphere-egu24-11897, 2024.

We introduce the CLEAN algorithm to identify narrowband Ultra Low Frequency (ULF) Pc5 plasma waves in Earth’s magnetosphere. The CLEAN method was first used for constructing 2D images in astronomical radio interferometry but has since been applied to a huge range of areas including adaptation for time series analysis. The algorithm performs a nonlinear deconvolution in the frequency domain (equivalent to a least-squares in the time domain) allowing for identification of multiple individual wave spectral peaks within the same power spectral density. The CLEAN method also produces real amplitudes instead of model fits to the peaks and retains phase information. We applied the method to GOES magnetometer data spanning 30 years to study the distribution of narrowband Pc5 ULF waves at geosynchronous orbit. We found close to 30,0000 wave events in each of the vector magnetic field components in field-aligned coordinates. We discuss wave occurrence and amplitudes distributed in local time and frequency. The distribution of the waves under different solar wind conditions are also presented. With some precautions, which are applicable to other event identification methods, the CLEAN technique can be utilized to detect wave events and its harmonics in the magnetosphere and beyond. We also discuss limitations of the method mainly the detection of unrealistic peaks due to aliasing and Gibbs phenomena.

How to cite: Inceoglu, F. and Loto'aniu, P.: Using the CLEAN Algorithm to Determine the Distribution of Ultra Low Frequency Waves at Geostationary Orbit, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12928, https://doi.org/10.5194/egusphere-egu24-12928, 2024.

Understanding volcanic activity through time series data analysis is crucial for uncovering the fundamental physical mechanisms governing this natural phenomenon.

In this study, we show the application of multifractal and fractal methodologies, along with statistical analysis, to investigate time series associated with volcanic activity. We aim to make use of these approaches to identify significant variations within the physical processes related to changes in volcanic activity. These methodologies offer the potential to identify pertinent changes preceding a high-energy explosion or a significant volcanic eruption.

In particular, we apply it to analyze two study cases. First, the evolution of the multifractal structure of volcanic emissions of low, moderate, and high energy explosions applied to Volcán de Colima (México years 2013-2015). The results contribute to obtaining quite evident signs of the immediacy of possible dangerous emissions of high energy, close to 8.0x10^8 J. Additionally, the evolution of the adapted Gutenberg-Richter seismic law to volcanic energy emissions contributes to confirm the results obtained using multifractal analysis. Secondly, we also studied the time series of the Gutenberg-Richter b-parameter of seismic activities associated with volcanic emissions in Iceland, Hawaii, and the Canary Islands, through the concept of Disparity (degree of irregularity), the fractal Hurst exponent, H, and several multifractal parameters. The results obtained should facilitate a better knowledge of the relationships between the activity of volcanic emissions and the corresponding related seismic activities.  

How to cite: Monterrubio-Velasco, M., Lana, X., Arámbula-Mendoza, R., and Zúñiga, R.: Applying Multifractal Theory and Statistical Techniques for High Energy Volcanic Explosion Detection and Seismic Activity Monitoring in Volcanic Time Series, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12938, https://doi.org/10.5194/egusphere-egu24-12938, 2024.

Spatially interpolated meteorological data products are widely used in the geosciences as well as disciplines like epidemiology, economics, and others. Recent work has examined methods for quantifying uncertainty in gridded estimates of near-surface air temperature that produce distributions rather than simply point estimates at each location. However, meteorological variables are correlated not only in space but in time, and sampling without accounting for temporal autocorrelation produces unrealistic time series and potentially underestimates cumulative errors. This work first examines how uncertainty in air temperature estimates varies in time, both seasonally and at shorter timescales. It then uses data-driven, spectral, and statistical methods to better characterize uncertainty in time series of estimated air temperature values. Methods for sampling that reproduce spatial and temporal autocorrelation are presented and evaluated. The results of this work are particularly relevant to domains like agricultural and ecology. Physical processes including evapotranspiration and primary production are sensitive to variables like near-surface air temperature, and errors in these important meteorological inputs accumulate in model outputs over time.

How to cite: Doherty, C. and Wang, W.: Characterizing Uncertainty in Spatially Interpolated Time Series of Near-Surface Air Temperature, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13593, https://doi.org/10.5194/egusphere-egu24-13593, 2024.

Vegetation phenology is the recurring of plant growth, including the cessation and resumption of growth, and plays a significant role in shaping terrestrial water, nutrient, and carbon cycles. Changes in temperature and precipitation have already induced phenological changes around the globe, and these trends are likely to continue or even accelerate. While warming has advanced spring arrival in many places, the effects on autumn phenology are less clear-cut, with evidence for earlier, delayed, or even unchanged end of the growing season (EOS). Meteorological droughts are intensifying in duration and frequency because of climate change. Droughts intricately impact changes in vegetation, contingent upon whether the ecosystem is limited by water or energy. These droughts have the potential to influence EOS changes. Despite this, the influence of drought on EOS remains largely unexplored. This study examined moisture’s role in controlling EOS by understanding the relationship between precipitation anomalies, vegetation’s sensitivity to precipitation (SPPT), and EOS. We also assess regional variations in responses to the impact of SPPT on EOS.

The study utilized multiple vegetation and water satellite products to examine the patterns of SPPT in drought and its impact on EOS across aridity gradients and vegetation types. By collectively evaluating diverse SPPTs from various satellite datasets, this work offers a comprehensive understanding and critical basis for assessing the impact of drought on EOS. We focused on the Northern Hemisphere from 2000 to 2020, employing robust statistical methods. This work found that, in many places, there was a stronger relationship between EOS and drought in areas with higher SPPT. Additionally, a non-linear negative relationship was identified between EOS and SPPT in drier regions, contracting with a non-linear positive relationship observed in wetter regions. These findings were consistent across a range of satellite-derived vegetation products. Our findings provide valuable insights into the effects of SPPT on EOS during drought, enhancing our understanding of vegetation responses to drought and its consequences on EOS and aiding in identifying drought-vulnerable areas.

How to cite: Choi, E. and Gray, J.: Understanding the role of vegetation responses to drought in regulating autumn senescence, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13879, https://doi.org/10.5194/egusphere-egu24-13879, 2024.

We present a machine-learning-based approach for predicting the geomagnetic main field changes, known as secular variation (SV), in a 5-year range for use for the 14^th generation of International Geomagnetic Reference Field (IGRF-14). The training and test datasets of the machine learning (ML) models are geomagnetic field snapshots derived from magnetic observatory hourly means, and CHAMP and Swarm-A satellite data (MCM Model; Ropp et al., 2020). The geomagnetic field data are not used as-is in the original time series but were differenced twice before training. Because SV is strongly influenced by the geodynamo process occurring in the Earth's outer core, challenges still persist despite efforts to model and forecast the realistic nonlinear behaviors (such as the geomagnetic jerks) of the geodynamo through data assimilation. We compare three physics-uninformed ML models, namely, the Autoregressive (AR) model, Vector Autoregressive (VAR) model, and Recurrent Neural Network (RNN) model, to represent the short-term temporal evolution of the geomagnetic main field on the Earth’s surface. The quality of 5-year predictions is tested by the hindcast results for the learning window from 2004.50 to 2014.25. These tests show that the forecast performance of our ML model is comparable with that of candidate models of IGRF-13 in terms of data misfits after the release epoch (Year 2014.75). It is found that all three ML models give 5-year prediction errors of less than 100nT, among which the RNN model shows a slightly better accuracy. They also suggest that Overfitting to the training data used is an undesirable machine learning behavior that occurs when the RNN model gives accurate reproduction of training data but not for forecasting targets.

How to cite: Sato, S. and Toh, H.: A machine-learning-based approach for predicting the geomagnetic secular variation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16981, https://doi.org/10.5194/egusphere-egu24-16981, 2024.

The quality of the seismic signal recorded at permanent and temporary stations is sometimes degraded, either abruptly or over time. The most likely cause is a high level of humidity, leading to corrosion of the connectors but environmental changes can also alter recording conditions in various frequency ranges and not necessarily for all three components in the same way. Assuming that the continuous seismic signal can be described by a normal distribution, we present a new approach to quantify the seismogram quality and to point out any time sample that deviates from this Gaussian assumption. To this end the notion of background Gaussian signal (BGS) to statistically describe a set of samples that follows a normal distribution. The discrete function obtained by sorting the samples in ascending order of amplitudes is compared to a modified probit function to retrieve the elements composing the BGS, and its statistical properties, mostly the Gaussian standard deviation, which can then differ from the classical standard deviation. Hence the ratio of both standard deviations directly quantifies the dominant gaussianity of the continuous signal and any variation reflects a statistical modification of the signal quality. We present examples showing daily variations in this ratio for stations known to have been affected by humidity, resulting in signal degradation. The theory developed can be used to detect subtle variations in the Gaussianity of the signal, but also to point out any samples that don't match the Gaussianity assumption, which can then be used for other seismological purposes, such as coda determination.

How to cite: Beucler, É., Bonnin, M., and Cuvier, A.: Introducing a new statistical theory to quantify the Gaussianity of the continuous seismic signal, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17344, https://doi.org/10.5194/egusphere-egu24-17344, 2024.

Climate events may induce abnormal ocean wave activities, that can be detected by seismic array on nearby coastlines. We collected long-term continuous array seismic data in the Groningen area and the coastal areas of the North Sea, conducted a comprehensive analysis to extract valuable climate information hidden within the ambient noise. Through long-term spectral analysis, we identified the frequency band ranging from approximately 0.2Hz, which appears to be associated with swell waves within the region, exhibiting a strong correlation with the significant wave height (SWH). Additionally, the wind waves with a frequency of approximately 0.4 Hz and gravity waves with periods exceeding 100 seconds were detected from the seismic ambient noise. We performed a correlation analysis between the ambient noise and various climatic indexes across different frequency bands. The results revealed a significant correlation between the North Atlantic Oscillation (NAO) Index and the ambient noise around 0.17Hz.

Subsequently, we extracted the annual variation curves of SWH frequency from ambient noise at each station around the North Sea and assembled them into a sparse spatial grid time series (SGTS). An empirical orthogonal function (EOF) analysis was conducted, and the Principal Component (PC) time series derived from the EOF analysis were subjected to a correlation analysis with the WAVEWATCH III (WW3) model simulation data, thereby confirming the wave patterns. Moreover, we conducted the spatial distribution study of SGTS. The spatial features revealed that the southern regions of the North Sea exhibit higher wind-wave energy components influenced by the Icelandic Low pressure system and topography, which explains the correlation between ambient noise in the region and the NAO index. Furthermore, spatial features disclosed a correlation between the first EOF mode of the North Sea ocean waves and the third mode of sea surface temperature anomalies. This research shows the potential of utilizing existing off-shore seismic monitoring systems to study global climate variation and physical oceanography.

How to cite: Zhong, Y., Gu, C., Fehler, M., Prieto, G., Wu, P., Yuan, Z., Chen, Z., and Kang, B.: Unveiling Climate-Induced Ocean Wave Activities Using Seismic Array Data in the North Sea Region, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17566, https://doi.org/10.5194/egusphere-egu24-17566, 2024.

Several machine learning algorithms and analytical techniques do not allow gaps or non-values in input data. Unfortunately, earth observation (EO) datasets, such as satellite images, are gravely affected by cloud contamination and sensor artifacts that create gaps in the time series of collected images. This limits the usage of several powerful techniques for modeling and analysis. To overcome these limitations, several works in literature propose different imputation methods to reconstruct the gappy time series of images, providing complete time-space datasets and enabling their usage as input for many techniques.

However, among the time-series reconstruction methods available in literature, only a few of them are publicly available (open source code), applicable without any external source of data, and suitable for application to petabyte (PB) sized dataset like the full Landsat archive. The few methods that match all these characteristics are usually quite trivial (e.g. linear interpolation) and, as a consequence, they often show poor performance in reconstructing the images. 

For this reason, we propose a new methodology for time series reconstruction designed to match all these requirements. Like some other methods in literature, the new method, named seasonally weighted average generalization (SWAG), works purely on the time dimension, reconstructing the images working on each time series of each pixel separately. In particular, the method uses a weighted average of the samples available in the original time series to reconstruct the missing values. Enforcing the annual seasonality of each band as a prior, for the reconstruction of each missing sample in the time series a higher weight is given to images that are collected exactly on integer multiples of a year. To avoid propagation of land cover changes in future or past images, higher weights are given to more recent images. Finally, to have a method that respects causality, only images from the past of each sample in the time series are used.

To have computational performance suitable for PB sized datasets the method has been implemented in C++ using a sequence of fast convolution methods and Hadamard products and divisions. The method has been applied to a bimonthly aggregated version of the global GLAD Landsat ARD-2 collection from 1997 to 2022, producing a 400 terabyte output dataset. The produced dataset will be used to generate maps for several biophysical parameters, such as Fraction of Absorbed Photosynthetically Active Radiation (FAPAR), normalized difference water index (NDWI) and bare soil fraction (BSF). The code is available as open source, and the result is fully reproducible.

References:

Potapov, Hansen, Kommareddy, Kommareddy, Turubanova, Pickens, ... & Ying  (2020). Landsat analysis ready data for global land cover and land cover change mapping. Remote Sensing, 12(3), 426.

Julien, & Sobrino (2019). Optimizing and comparing gap-filling techniques using simulated NDVI time series from remotely sensed global data. International Journal of Applied Earth Observation and Geoinformation, 76, 93-111.

Radeloff, Roy, Wulder, Anderson, Cook, Crawford, ... & Zhu (2024). Need and vision for global medium-resolution Landsat and Sentinel-2 data products. Remote Sensing of Environment, 300, 113918.

How to cite: Consoli, D., Parente, L., and Witjes, M.: A new methodology for time-series reconstruction of global scale historical Earth observation data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18061, https://doi.org/10.5194/egusphere-egu24-18061, 2024.

Climate variability gives rise to many different kinds of extreme impact events, including heat waves, crop failures, or wildfires. The frequency and magnitude of such events are changing under global warming. However, it is less known to what extent such events occur with some regularity, and whether this regularity is also changing as a result of climate change. Here, we present a novel method to systematically study the time-autocorrelation of these extreme impact events, that is, whether they occur with a certain regularity. In studies of climate change impacts, different types of events are often studied in isolation, but in reality they interact. We use ensembles of global biophysical impact simulations from the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) driven with climate models to assess current conditions and projections. The time series analysis is based on a discrete Fourier transformation that accounts for the stochastic fluctuations from the climate model. Our results show that some climate impacts, such as crop failure, indeed exhibit a dominant frequency of recurrence; and also, that these regularity patterns change over time due to anthropogenic climate forcing.

How to cite: Zantout, K., Frieler, K., and Schewe, J. and the ISIMIP team: The regularity of climate-related extreme events under global warming, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18197, https://doi.org/10.5194/egusphere-egu24-18197, 2024.

The availability of comprehensive aerial photography is limited to the mid-20th century, posing a challenge for quantitatively analyzing long-term surface changes in proglacial areas. This creates a gap of approximately 100 years, spanning the end of the Little Ice Age (LIA). Employing digital monoplotting and historical terrestrial images, our study reveals quantitative surface changes in a LIA lateral moraine section dating back to the second half of the 19th century, encompassing a total study period of 130 years (1890 to 2020). With the long-term analysis at the steep lateral moraines of Gepatschferner (Kauner Valley, Tyrol, Austria) we aimed to identify changes in vegetation development in context with morphodynamic processes and the changing climate.

In 1953, there was an expansion in the area covered by vegetation, notably encompassing scree communities, alpine grassland, and dwarf shrubs. However, the destabilization of the system after 1980, triggered by rising temperatures and the resulting thawing of permafrost, led to a decline in vegetation cover by 2020. Notably, our observations indicated that, in addition to morphodynamic processes, the overarching trends in temperature and precipitation exerted a substantial influence on vegetation development. Furthermore, areas with robust vegetation cover, once stabilised, were reactivated and subjected to erosion, possibly attributed to rising temperatures post-1980.

This study demonstrates the capability of historical terrestrial images to enhance the reconstruction of vegetation development in context with morphodynamics in high alpine environments within the context of climate change. However, it is important to note that long-term mapping of vegetation development through digital monoplotting has limitations, contingent on the accessibility and quality of historical terrestrial images, as well as the challenges posed by shadows in high alpine regions. Despite these limitations, this long-term approach offers fundamental data on vegetation development for future modelling efforts.

How to cite: Ramskogler, K., Altmann, M., Mikolka-Flöry, S., and Tasser, E.: Long-term vegetation development in context of morphodynamic processes since mid-19th century, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18210, https://doi.org/10.5194/egusphere-egu24-18210, 2024.

Discrimination of  geomagnetic quasi-periodic signals by using SSA Transform

• Palangio¹, L. Santarelli ¹

¹Istituto Nazionale di Geofisica e Vulcanologia L’Aquila

³Istituto Nazionale di Geofisica e Vulcanologia Roma

Correspondence to:  lucia.santarelli@ingv.it

Abstract

In this paper we present an application of  the SSA Transform to the detection and reconstruction of  very weak geomagnetic signals hidden in noise. In the SSA Transform  multiple subspaces are used for representing and reconstructing   signals and noise.  This analysis allows us to reconstruct, in the time domain, the different harmonic components contained in the original signal by using  ortogonal functions. The objective is to identificate the dominant  subspaces that can be attributed to the  signals and the subspaces that can be attributed to the noise,  assuming that all these  subspaces are orthogonal to each other, which implies that the  signals and noise  are independent of one another. The subspace of the signals is mapped simultaneously on several spaces with a lower dimension, favoring the dimensions that best discriminate the patterns. Each subspace of the signal space is used to encode different subsets of functions having common characteristics, such as  the same periodicities. The subspaces  identification was performed by using singular value decomposition (SVD) techniques,  known as  SVD-based identification methods  classified in a subspace-oriented scheme.The  quasi-periodic variations of geomagnetic field  has been investigated in the range of scale which span from 22 years to 8.9 days such as the  Sun’s polarity reversal cycle (22 years), sun-spot cycle (11 years), equinoctial effect (6 months), synodic rotation of the Sun (27 days) and its harmonics. The strength of these signals vary from fractions of a nT to tens of nT. Phase and frequency variability of these cycles has been evaluated from the range of variations in the geomagnetic field recorded at middle latitude place (covering roughly 4.5 sunspot cycles). Magnetic data recorded at L'Aquila Geomagnetic observatory (geographic coordinates: 42° 23’ N, 13° 19’E, geomagnetic coordinates: 36.3° N,87°.2 E, L-shell=1.6) are used from 1960 to 2009.

How to cite: Paolo Giovanni, P. and Lucia, S.: Discrimination of  geomagnetic quasi-periodic signals by using SSA Transform, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19601, https://doi.org/10.5194/egusphere-egu24-19601, 2024.

Over the past decades, especially since 2014, large quantities of Earth Observation (EO) data became available in high spatial and temporal resolution, thanks to ever-developing constellations (e.g.: Sentinel, Landsat) and open data policy. However, in the case of optical images, affected by cloud coverage and the spatially changing overlap of relative satellite orbits, creating temporally generalized and dense time series by using only measured data is challenging, especially when studying larger areas.

Several papers investigate the question of spatio-temporal gap filling and show different interpolation methods to calculate missing values corresponding to the measurements. In the past years more products and technologies have been constructed and published in this field, for example Copernicus HR-VPP Seasonal Trajectories (ST) product.  These generalized data structures are essential to the comparative analysis of different time periods or areas and improve the reliability of data analyzing methods such as Fourier transform or correlation. Temporally harmonized input data is also necessary in order to improve the results of Machine Learning classification algorithms such as Random Forest or Convolutional Neural Networks (CNN). These are among the most efficient methods to separate land cover categories like arable lands, forests, grasslands and built-up areas, or crop types within the arable category.

This study analyzes the efficiency of different interpolation methods on Sentinel-2 multispectral time series in the context of land cover classification with Machine Learning. We compare several types of interpolation e.g. linear, cubic and cubic-spline and also examine and optimize more advanced methods like Inverse Distance Weighted (IDW) and Radial Basis Function (RBF). We quantify the accuracy of each method by calculating mean square error between measured and interpolated data points. The role of interpolation of the input dataset in Deep Learning (CNN) is investigated by comparing Overall, Kappa and categorical accuracies of land cover maps created from only measured and interpolated time series. First results show that interpolation has a relevant positive effect on accuracy statistics. This method is also essential in taking a step towards constructing robust pretrained Deep Learning models, transferable between different time intervals and agro-ecological regions.

The research has been implemented with the support provided by the Ministry of Culture and Innovation of Hungary from the National Research, Development and Innovation Fund, financed under the KDP-2021 funding scheme.

Keywords: time series analysis, Machine Learning, interpolation, Sentinel

How to cite: Simon, M., Richter-Cserey, M., Pacskó, V., and Kristóf, D.: Temporal Interpolation of Sentinel-2 Multispectral Time Series in Context of Land Cover Classification with Machine Learning Algorithms, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22262, https://doi.org/10.5194/egusphere-egu24-22262, 2024.

Earthquake-like ruptures disrupt the frictional interface between contacting bodies and initiate frictional motion (stick-slip). The interfacial slip (motion) immediately resulting from a rupture during each stick-slip event is usually much smaller than the total slip recorded during the duration of the event. Slip after the onset of friction is generally attributed to the continuous motion of global ‘dynamic friction’. Here, we demonstrate that numerous hitherto invisible secondary ruptures are initiated immediately after each initial rupture by directly measuring the contact area and slip at the frictional interface. Each secondary rupture generates incremental slip that, when not resolved, may appear as steady sliding of the interface. Each slip increment is linked, via fracture mechanics, to corresponding variations of contact area and local strain. Cumulative interfacial slip can only be described if the effects of these secondary ruptures are taken into account. These weaker slip sequences can also be observed in bimaterial interfaces and exhibit strong directional effects. In addition, the seismic moments we estimate based on slip sequences are consistent with the Gutenberg-Richter (G-R) law. These results have important implications for our fundamental understanding of frictional motion and the important role of aftershocks within natural faults in generating earthquake-mediated slip/afterslip.

How to cite: Shi, S., Wang, M., Poles, Y., and Fineberg, J.: Frictional slip sequences in homogeneous and bimaterial interfaces, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2384, https://doi.org/10.5194/egusphere-egu24-2384, 2024.

In recent years, an intriguing feature of back-propagating rupture (BPR) has been reported during some earthquakes (Ide et al., 2011; Houston et al., 2011; Hicks et al., 2020; Okuwaki et al., 2021; Vallée et al., 2023). The occurrence of BPR challenges the classical interpretation of rupture propagation as a “forward” problem, while remaining less understood by the earthquake science community. Here, using fracture mechanics, we first argue that BPR is nothing but an intrinsic component of rupture propagation; however, its observability is usually masked by the superposition effect of interfering waves behind the primary, forward-propagating rupture front. We then suggest that perturbation to an otherwise smooth rupture process can break the superposition effect and hence can make BPR observable. To test our idea, we report results of mode-II rupture propagation from both numerical simulations and laboratory observations. By introducing a variety of perturbations (lateral variation in bulk or interfacial properties along the fault, singular stress drop, coalescence of two rupture fronts, etc.) during rupture propagation, we show that prominent phases of BPR indeed can be successfully excited. We further classify BPR into two modes: higher-order rupture or interface wave, depending on whether the already-ruptured fault is quickly healed and whether additional stress drop can be produced. Lastly, we propose several application potentials for BPR, such as constraining the velocity structure of fault zones, probing the mechanical state of faults, and studying the stability of perturbed slip along a homogeneous or bimaterial interface. Our study refines the understanding of the nature and complexity of rupture process, and can help improve the assessment of earthquake hazards.

How to cite: Ding, X., Xu, S., Fukuyama, E., and Yamashita, F.: Back-Propagating Rupture: Nature, Excitation, and Applications, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3912, https://doi.org/10.5194/egusphere-egu24-3912, 2024.

Modeling the seismic cycle requires multiple assumptions and parameters. Providing a quantitative assessment of the model behavior is pivotal for determining the degree of similarity between different scales and modeling strategies and for exploring dependencies with respect to selected parameters. Here we compare stick-slip ruptures nucleating spontaneously in scaled seismotectonic models (i.e., laboratory experiments capturing the first-order physics of the seismic cycle of subduction megathrusts) with slow earthquakes in nature. We rely on two non-dimensional parameters, namely the Ruina number (Ru) and system dimension (D) to quantify model behavior. Ru is proportional to the ratio of the asperity size to the critical nucleation size. Within the rate- and state friction framework, for velocity weakening asperities Ru controls the behavior of the system, which can be either periodic or not, and it can exhibit both slow and fast ruptures. D measures how complicated the system evolution is. D reveals how many variables are required to describe the seismic cycle because it tells us the minimum dimension needed to embed the observed dynamics. 

By coupling the Simulated Annealing algorithm and quasi-dynamic numerical simulations, we retrieve rate and state friction parameters characterizing single asperity models with different lateral extent of the velocity weakening patch. Similarly to slow earthquakes, we found optimal rate and state parameters indicative of low (< 4) Ru. We also found a direct proportionality between Ru and the lateral extent of the asperity. 

Next, we implement tools from non-linear time-series analysis and Extreme Value Theory to compute D from models of different sizes, materials, deformation rates and frictional configurations (single or twin asperities along strike). Our analysis supports the existence of a low dimensional attractor (D<5) describing the dynamics of scaled seismotectonic models. In particular, our models display D=3.0-4.2, which is remarkably similar to D=3.2 of slow earthquakes identified along the Cascadia subduction zone. Under the explored conditions, D appears more affected by the material behavior of the analog upper plate (i.e., gelatin vs. foam rubber) than by the lateral frictional segmentation of the megathrust.

Despite the different spatio-temporal scales, our results support a scenario where scaled seismotectonic models and slow earthquakes share similar dynamics.