Earthquake swarms and complex seismic sequences driven by transient forcing in tectonic and volcanic regions 

Earthquake swarms are characterized by a complex temporal evolution and a delayed occurrence of the largest magnitude event. In addition, seismicity often manifests with intense foreshock activity or develops in more complex sequences where doublets or triplets of large comparable magnitude earthquakes occur. The difference between earthquake swarms and these complex sequences is subtle and usually flagged as such only a posteriori. This complexity derives from aseismic transient forcing acting on top of the long-term tectonic loading: pressurization of crustal fluids, slow-slip and creeping events, and at volcanoes, magmatic processes (i.e. dike and sill intrusions or magma degassing). From an observational standpoint, these complex sequences in volcanic and tectonic regions share many similarities: seismicity rate fluctuations, earthquakes migration, and activation of large seismogenic volume despite the usual small seismic moment released. The underlying mechanisms are local increases of the pore-pressure, loading/stressing rate due to aseismic processes (creeping, slow slip events), magma-induced stress changes, earthquake-earthquake interaction via static stress transfer or a combination of those. Yet, the physics behind such processes and the ultimate reasons for the occurrence of swarm-like rather than mainshock-aftershocks sequences, is still far beyond a full understanding.

This session aims at putting together studies of swarms and complex seismic sequences driven by aseismic transients in order to enhance our insights on the physics of such processes. Contributions focusing on the characterization of these sequences in terms of spatial and temporal evolution, scaling properties, and insight on the triggering physical processes are welcome. Multidisciplinary studies using observation complementary to seismological data, such as fluid geochemistry, deformation, and geology are also welcome, as well as laboratory and numerical modeling simulating the mechanical condition yielding to swarm-like and complex seismic sequences.

Co-organized by GMPV9/NH4/TS4
Convener: Luigi Passarelli | Co-conveners: Simone Cesca, Federica LanzaECSECS, Francesco Maccaferri, Maria MesimeriECSECS
vPICO presentations
| Thu, 29 Apr, 14:15–17:00 (CEST)

vPICO presentations: Thu, 29 Apr

Chairpersons: Luigi Passarelli, Maria Mesimeri, Francesco Maccaferri
Louis De Barros, Pierre Dublanchet, Frédéric Cappa, and Anne Deschamps

Fluid induced earthquake sequences generally appear as expanding swarms activating a particular fault. Such swarms are generally interpreted as fluid diffusion, which ignores the possibility of static, dynamic or aseismic triggering, and the existence of rapid migration. Here, we study the temporal evolution of a seismic swarm that occurred over a 10-day period in October 2015 in the extensional rift of the Corinth Gulf (Greece) using high-resolution earthquakes relocations. The seismicity radially migrates on a normal fault at a fluid diffusion velocity (~125 m/day). However, this migration occurs intermittently, with periods of fast expansion (2-to-10 km/day) during short seismic bursts alternating with quiescent periods. Moreover, the growing phases of the swarm illuminate a high number of repeaters. Therefore, we propose a new model to explain the combination of multiple driving processes for such swarms.  Fluid up flow in the fault may induce aseismic slip episodes, separated by phases of fluid pressure build-up. The stress perturbation due to aseismic slip may activate small asperities in the fault that produce bursts of seismicity during the most intense phase of the swarm. We then validated this model through hydro-mechanical modeling, where earthquakes consist in the failure of asperities on a creeping fault infiltrated by fluid. For that, we couple rate‐and‐state friction, non‐linear diffusivity and elasticity along a 1D interface. This model reproduces the dual migration speeds observed in real swarms. We show that migration speeds increase linearly with the mean pressurization, and are not dependent on the hydraulic diffusivity, as traditionally suggested.

How to cite: De Barros, L., Dublanchet, P., Cappa, F., and Deschamps, A.: Dual seismic migration velocities reveal imbricated fluid diffusion and aseismic slip In a Corinth Gulf  swarm (Greece), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1073,, 2021.

Tomáš Fischer, Sebastian Hainzl, Josef Vlček, and Ali Salama

Migration of hypocenters is a common attribute of induced injection seismicity and of earthquake swarms, which distinguishes them from aftershock sequences. Spreading of the triggering front is often examined by fitting the time dependence of hypocenter distances from the origin by the pore pressure diffusion model. The earthquake migration patterns however often exhibit not only spreading envelopes, but also fast-growing streaks embedded in the overall migration trends. We review the observed migration patterns and show that in the case of self-driven seismicity, where the new ruptures are triggered at the edge of previous ruptures, it is more suitable to examine the cluster growth as a function of the event index instead of time, which often discloses a continuous linear growth during time periods which appeared strongly discontinuous in the coordinate-time plots.

We propose a model that relates the speed of seismicity spreading to the average rupture area and the effective magnitude of the hypocenter cluster. Application of the model to selected linearly growing clusters of the 2008 West Bohemia swarm gives almost linear increase of the measured total rupture area with the event index, which fits the proposed model. This is confirmed by a self-similar scaling of the average rupture area with the effective magnitude for stress drops ranging from 0.1 to 1 MPa. The relatively small stress drop level indicates the presence of voids along the fault plane and a possible role of aseismic deformation. Further application of the model to seismic swarms from different areas confirms its validity and potential for distinguishing fluid-triggered seismicity.

How to cite: Fischer, T., Hainzl, S., Vlček, J., and Salama, A.: What drives the growth of earthquake clusters?, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2930,, 2021.

Tom Winder and Robert S White

Intense swarms of microearthquakes have been detected in the rift zone of Central Iceland since the 1970s, but the cause of their clear swarm-like nature remains enigmatic. We use the QuakeMigrate earthquake detection and location software1 to produce a highly complete catalogue of microseismicity from 2007-2020, using data from a dense local seismic network. Automatic hypocentre locations have been refined using waveform cross-correlation and double-difference relocation, and tightly constrained focal mechanisms have been obtained by manual analysis of a subset of events.

The resulting high-resolution earthquake catalogue reveals a network of conjugate strike-slip faults, oriented to accommodate plate-boundary extension. Sharply defined fault planes imaged by the microearthquake hypocentres range from 1-10 km in length, and are found between 1 and 8 km b.s.l., with their orientations closely matching the fault plane geometry inferred from the fault plane solutions. Seismicity within individual swarms displays a systematic migration of hypocentres at velocities of ~ 1 km/day. In the majority of swarms we also observe clusters of identical repeating events, providing evidence for re-loading of brittle asperities.

For a selection of swarms our high resolution seismic observations are complemented by GPS and InSAR measurements, allowing us to place constraints on the amount of fault slip. Comparing this, and the area of the fault plane activated in the swarm, to the seismic moment release reveals a significant contribution of aseismic slip, or very low effective stress drop. Analysis of swarms within this fault network triggered by the 2014 Bárðarbunga-Holuhraun dike intrusion provides further constraint on the amplitude of the stress cycle.

We combine our observations with comparisons to numerical & laboratory modelling studies, observed swarm scaling properties and knowledge of the geological and permeability structure of the Icelandic crust to determine the nature of the transient forcing driving these exceptionally well-recorded tectonic earthquake swarms.


1: Tom Winder, Conor Bacon, Jonathan D. Smith, Thomas S. Hudson, Julian Drew, & Robert S. White. (2021, January 15). QuakeMigrate v1.0.0 (Version v1.0.0). Zenodo.

How to cite: Winder, T. and White, R. S.: Slowly migrating tectonic microearthquake swarms in the Icelandic Rift Zone: driven by pore-pressure or aseismic slip transients?, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15544,, 2021.

Rebecca O. Salvage and David W. Eaton

The global pandemic of COVID-19 furnished an opportunity to study seismicity in the Kiskatinaw area of British Columbia, noted for hydraulic-fracturing induced seismicity, during a period of anthropogenic quiescence. A total of 389 events were detected from April to August 2020, encompassing a period with no hydraulic-fracturing operations during a government-imposed lockdown. During this time period, observed seismicity had a maximum magnitude of ML 1.2 and lacked temporal clustering that is often characteristic of hydraulic-fracturing induced sequences. Instead, seismicity was persistent over the lockdown period, similar to swarm-like seismicity with no apparent foreshock-aftershock type sequences. Hypocenters occurred within a corridor orientated NW-SE, just as seismicity had done in previous years in the area, with focal depths near the target Montney formation or shallower (<2.5 km). Based on the Gutenberg-Richter relationship, we estimate that a maximum of 21% of the detected events during lockdown may be attributable to natural seismicity, with a further 8% possibly due to dynamic triggering of seismicity from teleseismic events. The remaining ~70% cannot be attributed to direct pore pressure increases induced by fluid injection, and therefore is inferred to represent latent seismicity i.e. seismicity that occurs after an unusually long delay following primary activation processes, with no obvious triggering mechanism. We can exclude pore-pressure diffusion from the most recent fluid injection, as is there is no clear pattern of temporal or spatial seismicity migration. If elevated pore pressure from previous injections became trapped in the subsurface, this could explain the localization of seismicity within an operational corridor, but it does not explain the latency of seismicity on a timescale of months. However, aseismic creep on weak surfaces such as faults, in response to tectonic stresses, in addition to trapped elevation pore-pressure could play a role in stress re-loading to sustain the observed pattern of seismicity.

How to cite: Salvage, R. O. and Eaton, D. W.: Latent seismicity driven by aseismic creep and enhanced pore-fluid pressure in NE British Columbia, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6421,, 2021.

Georgios Michas, Vasilis Kapetanidis, George Kaviris, and Filippos Vallianatos

Earthquake diffusion is frequently observed in the spatiotemporal evolution of seismic clusters and regional seismicity, a characteristic that is attributed to a triggering mechanism, such as fluid flow, aseismic creep and/or stress transfer effects. In this work, we study the earthquake diffusion properties in the Western Gulf of Corinth (central Greece), an area that presents high extension rates, moderate to large magnitude earthquakes, intense microseismicity and frequent seismic swarms. We focus on the period 2013–2014 that is characterized by intense background microseismic activity along with significant seismic sequences. More specifically, the latter include the 2013 Helike swarm, the 2014 seismic sequence between Nafpaktos and Psathopyrgos, which culminated with an Mw 4.9 event on 21 September 2014, as well as moderate magnitude events that were followed by aftershock sequences. In the herein analysis, we employ a relocated earthquake catalogue of ~9000 events which delineates the activated areas during the study period in high-resolution. We consider the most significant seismic sequences and calculate their respective spatial correlation histograms and the evolution of the mean squared distance of the hypocenters with time, in order to study the earthquake diffusion rates and possible variations that might be related to the triggering mechanisms of seismicity. Our results demonstrate a weak earthquake diffusion process, analogous to subdiffusion within a stochastic framework, for the seismic sequences under consideration, providing further evidence for slow earthquake diffusion in regional and global seismicity. In addition, the earthquake diffusion rates exhibit variations that can be associated with the triggering mechanism. In particular, seismic sequences which are related with pore-fluid pressure diffusion present considerably higher diffusion rates than mainshock/aftershock sequences or the background activity. Such results may provide novel constraints on the triggering mechanisms of clustered seismic activity based on the study of the earthquake diffusion rates. 


We would like to thank the personnel of the Hellenic Unified Seismological Network ( and the Corinth Rift Laboratory Network ( for the installation and operation of the stations used in the current article. The present research is co-financed by Greece and the European Union (European Social Fund- ESF) through the Operational Programme «Human Resources Development, Education and Lifelong Learning 2014-2020» in the context of the project “The role of fluids in the seismicity of the Western Gulf of Corinth (Greece)” (MIS 5048127).

How to cite: Michas, G., Kapetanidis, V., Kaviris, G., and Vallianatos, F.: Variations of the earthquake diffusion rates in the Western Gulf of Corinth (Greece), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5282,, 2021.

Simone Cesca, Carla Valenzuela Malebrán, José Ángel López-Comino, Timothy Davis, Carlos Tassara, Onno Oncken, and Torsten Dahm

A complex seismic sequence took place in 2014 at the Juan Fernández microplate, a small microplate located between Pacific, Nazca and Antarctica plates. Despite the remoteness of the study region and the lack of local data, we were able to resolve earthquake source parameters and to reconstruct the complex seismic sequence, by using modern waveform-based seismological techniques. The sequence started with an exceptional Mw 7.1-6.7 thrust – strike slip earthquake doublet, the first subevent being the largest earthquake ever recorded in the region and one of the few rare thrust earthquakes in a region otherwise characterized by normal faulting and strike slip earthquakes. The joint analysis of seismicity and focal mechanisms suggest the activation of E-W and NE-SW faults or of an internal curved pseudofault, which is formed in response to the microplate rotation, with alternation of thrust and strike-slip earthquakes. Seismicity migrated Northward in its final phase, towards the microplate edge, where a second doublet with uneven focal mechanisms occurred. The sequence rupture kinematics is well explained by Coulomb stress changes imparted by the first subevent. Our analysis show that compressional stresses, which have been mapped at the northern boundary of the microplate, but never accompanied by large thrust earthquakes, can be accommodated by the rare occurrence of large, impulsive, shallow thrust earthquakes, with a considerable tsunamigenic potential.

How to cite: Cesca, S., Valenzuela Malebrán, C., López-Comino, J. Á., Davis, T., Tassara, C., Oncken, O., and Dahm, T.: Seismic doublets and a complex seismic sequence controlled by the rotation of the Juan Fernández microplate, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1119,, 2021.

Chairpersons: Luigi Passarelli, Federica Lanza, Simone Cesca
Marion Baques, Louis De Barros, Maxime Godano, Hervé Jomard, Clara Duverger, Francoise Courboulex, and Christophe Larroque

The Ubaye Region, where the city of Barcelonnette is settled, is the most seismically active region in the French Western Alps since at least two centuries. Seismicity in this area exhibits a dual behaviour, with mainshock-aftershock sequences alternating with abnormally high rate of seismicity associated with seismic swarms. Understanding processes triggering such a peculiar seismic behaviour is of primary importance in order to assess the seismic hazard in this region. The latest swarm activity started on February 26, 2012, with an earthquake of moment magnitude 4.2. It was followed two years later (on April 7, 2014) by a shock of magnitude Mw 4.8. From the first earthquake to the end of 2016, the seismic level has not returned to the background level and shares the same characteristics as a seismic swarm.

With the aim to discuss the seismogenic processes involved in the area, we focused on the two months following the 2014 mainshock (Mw=4.8). During this period, a dense temporary network (7 stations) was operating at a maximal distance of 10km from the epicentre area. We analysed this period starting with a double-difference relocation of ~ 6,000 earthquakes previously detected by template-matching. These hypocentres did not align on the fault plane of the 2014 mainshock, but on conjugated structures belonging to the 2-km wide damaged zone of the main fault plane and on remote structures with various orientations further away. We then computed 99 focal mechanisms from a joint inversion of P polarity and S/P ratio to clarify the geometry of the active structures. Many nodal planes are inconsistent with the structures deduced from the alignments of the earthquake locations. The stress-state orientation obtained from those focal mechanisms (σ1 trending N27°± 5°, plunging 50°± 9°, a σ2 trending N215°± 5°, plunging 40°± 9°, and a sub-horizontal σ3 trending N122°± 3°) is consistent with those previously calculated in the area (Fojtíková and Vavryčuk, 2018). Nevertheless, some structures are unfavourably oriented to slip within this stress-field, suggesting that additional processes are required to explain them. As the presence of fluids was highlighted for the 2003-2004 and the 2012-2015 crisis, we calculated the fluid pressure needed to trigger slip on the planes from the focal mechanisms using Cauchy's equation. We found that a median fluid-overpressure of ~20 MPa (range between 0 to 50 MPa) is needed to cause slip.  Although the origin of fluids and how they are pressurized at depth remains open. The fluid processes seem to be the most favourable additional processes and were also proposed to explain the 2003-2004 crisis.

How to cite: Baques, M., De Barros, L., Godano, M., Jomard, H., Duverger, C., Courboulex, F., and Larroque, C.: Complex behaviour highlighted by earthquake aftershock and swarm sequences in Ubaye Region (French Western Alps). , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5034,, 2021.

Riccardo Minetto, Agnès Hemlstetter, Philippe Guéguen, and Mickaël Langlais

We analyse the spatio-temporal variations of the seismicity recorded during the Maurienne swarm. The Maurienne swarm occurred between 2017 and 2018 in the French Alps in the central part of the external crystalline massif of Belledonne. This massif extends for more than 120km in N30 direction, it is bounded to the west by the wide topographic depression of the Isère valley and the Combe de Savoie, and it is crosscut by the Maurienne valley.  The location and the 3D shape of the seismic swarm are consistent with an outcroping N80 vertical fault zone. The seismic activity is interpreted as a result of the reactivation of this inherited vertical fault system. The largest event had a magnitude of 3.5.

We used a catalog of 58000 events that were detected using template-matching and relocated with a double-difference method.  
We show that the swarm is characterised by short-term (days) and long-term (months) migrations that may be related to the presence of fluids. 
We also observe that the b-value decreases with depth and we discuss how this variation may due to shallow fault systems whose geometry differs from the one of the main fault system. 
Part of the events occurred when only one station was active. This study shows that, by grouping earthquakes into groups of similar events (clusters), it is possible to study spatio-temporal variations in such conditions.

How to cite: Minetto, R., Hemlstetter, A., Guéguen, P., and Langlais, M.: Spatial and temporal variations of seismicity during the Maurienne swarm (French Alps): short- and long-term migrations and b-value dependence with depth, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7463,, 2021.

Konstantinos Michailos, N. Seth Carpenter, and György Hetényi

The Himalayan orogen, formed by the continental collision between the Indian and Eurasian plates, is a unique geological structure that has been extensively studied over the past few decades. These previous studies highlighted the occurrence of earthquakes in the orogen's roots beneath the central Himalayas. However, the characterization of these deep earthquakes remains limited. Here, we compiled a detailed, long-duration catalog, which we use to investigate the spatiotemporal characteristics of seismicity beneath the Himalayan orogen. 

To create this catalog, we collected all available continuous seismic data acquired during the last two decades in the central Himalayas region (i.e., 2001-2005). We applied a systematic, semi-automatic processing routine to obtain absolute earthquake locations using a 1-D velocity model. Using high-quality picks, ~8,000 preliminary earthquake locations have been determined, at least 1,000 of which have hypocentral depths >50 km. We plan to refine the preliminary locations and calculate local magnitudes for the intermediate-depth lithospheric earthquakes. Using this refined catalog, we will analyze the spatiotemporal evolution pattern and properties of the Himalayan deep seismicity. This analysis is expected to provide us with insights into the processes and mechanisms that control seismogenesis beneath the orogen. For example, is seismicity driven by earthquake stress transfer (mainshock-aftershock sequences), or is it caused by external processes like fluids or aseismic slip, or both?

How to cite: Michailos, K., Carpenter, N. S., and Hetényi, G.: Spatiotemporal evolution of deep seismicity beneath the central Himalayas, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9641,, 2021.

Maria Mesimeri, Kristine Pankow, Ben Baker, and J. Mark Hale

The Mineral Mountains are located in south-central Utah within the transition zone from the Basin and Range to Colorado Plateau physiographic provinces, near the Roosevelt Hot Springs. First evidence of swarm-like activity in the area was found in 1981, when a six-station temporary network detected a very energetic swarm of ~1,000 earthquakes. More recently, in mid-2016, a dense local broadband seismic network was installed around the Frontier Observatory for Research in Geothermal Energy (FORGE) in southcentral Utah, ~10 km west of the Mineral Mountains. Beginning in 2016, the University of Utah Seismograph Stations detected, located, and characterized 75 earthquakes beneath the Mineral Mountains. In this study, we build an enhanced earthquake catalog to confirm the episodic swarm-like nature of seismicity in the Mineral Mountains. We use the 75 cataloged earthquakes as templates and detect 1,000 earthquakes by applying a matched-filter technique. The augmented catalog reveals that seismicity in the Mineral Mountains occurs as short-lived earthquake swarms followed by periods of quiescence. Earthquake relocation of ~800 earthquakes shows that activity is concentrated in a <2 km long E-W striking narrow zone, ~4 km east of the Roosevelt hydrothermal system. Two fault orientations, both N-S and E-W parallel to the Opal Mound and Mag Lee faults, respectively, are observed after computing composite focal mechanisms of highly similar earthquakes. After examining the spatial and temporal patterns of the best recorded earthquake swarm in October 2019, we find that a complex mechanism of fluid diffusion and aseismic slip is responsible for the swarm evolution with migration velocities reaching 10 km/day. We hypothesize that these episodic swarms in the Mineral Mountains are primarily driven by migrating fluids that originate within the Roosevelt hydrothermal system.

How to cite: Mesimeri, M., Pankow, K., Baker, B., and Hale, J. M.: Episodic earthquake swarms in the Mineral Mountains, Utah driven by the Roosevelt hydrothermal system, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1227,, 2021.

Minkyung Son, Chang Soo Cho, Jin-Hyuck Choi, Jeong-Soo Jeon, and Yun Kyung Park

Low-magnitude earthquakes (maximum Mw: 3.2) were recorded from late April 2020 onward in the county of Haenam, southwestern South Korea. Moderate to strong earthquakes had not previously been documented in instrumental, historical, or geological records. We identified 226 hypocentres in this earthquake sequence from April 25 to May 11, 2020. The seismic front of this sequence migrated in a manner similar to a diffusing fluid, with a hydraulic diffusivity of 0.012 m2/s. This is the first observation of natural seismicity on the Korean Peninsula imitating fluid diffusion. We applied a cross-correlation approach to detect unrecorded events, and relocated the hypocentres of the 71 previously recorded and 155 newly detected events using data collected at permanent seismic stations; clear linearity was observed at the metre scale. Spatially, the hypocentres were distributed within a 0.3 km × 0.3 km fault plane at a depth of ~20 km, trending west-northwest–east-southeast with a dip of ~70° in the south-southwestern direction. The moment tensor solution of the largest event had a strike of 98°, dip of 65°, and rake of 7°, which correspond to the fault geometry of the relocated hypocentres. The hypocentres progressed toward the upper eastern edge of the lineament. The largest event occurred at a shallow region of the fault plane, in the direction of hypocentre migration. Together, these results showed that the migration sequence of the 2020 Haenam earthquake mimicked the flow of a diffusing fluid. The structural data indicate that a fault–fracture mesh geometry channelled fluid flow, supporting the concept of a “fluid-driven earthquake swarm” for the 2020 Haenam earthquake sequence. Regarding the final parts of the sequence, there appeared to be a second intrusion at the western end, and a permeability barrier at the eastern end, of the fault plane. The well-constrained hypocentre locations in our study provide essential data for future research, and our interpretations of hypocentre migration during this earthquake sequence may help to elucidate the mechanisms driving earthquake swarms under conditions of intraplate stress.

How to cite: Son, M., Cho, C. S., Choi, J.-H., Jeon, J.-S., and Park, Y. K.: The 2020 Haenam earthquake sequence: The first observation of a seismic front on the Korean Peninsula migrating in a manner similar to fluid diffusion, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-536,, 2021.

Ferdinando Napolitano, Ortensia Amoroso, Valeria Vitale, Anna Gervasi, Mario La Rocca, and Paolo Capuano

The 2010-2014 Pollino seismic sequence occurred in a well known seismic gap zone in Southern Italy. Although paleoseismological studies revealed the occurrence of at least two earthquakes of MW6.5-7 in the last 10,000 years, no earthquakes larger than M6 occurred in historical times. The sequence had a long duration and it was characterized by a variable seismic rate and by two mainshocks (ML4.3 and ML5.0) occurred in May and October 2012, two years after the beginning of the swarm. In the same area a slow slip event started three months before the ML5.0 earthquake and lasted for one year.

The aim of this work is the investigation of the elastic properties of the seismogenic volume and the presence of abundant fluids inferred from the study of attenuation. The role that fluids in highly fractured media play in triggering and driving the occurrence of earthquake swarms is believed very important, but yet to be understood clearly. In order to investigate the elastic properties of the medium, we performed a local P- and S-wave 3D tomographic image. We selected 870 earthquakes (ML1.8–5.0) occurred between 2010 and 2014 from the sequence and nearby within a volume of 100x120x25km3. We manually picked 9981 P and 6862 S arrivals recorded by 39 seismic stations. The picking consistency was estimated by modified Wadati diagram which also provided an estimate of VP/Vs equal to 1.786.

We applied a linearized, iterative delay-time inversion approach, which simultaneously inverts the first arrivals of direct waves for both velocity model parameters and earthquake locations. The dataset and the station distribution allow us to set a 5x5x1km3 grid for the inversion. We performed several numerical tests to estimate a reliable starting 1D P- and S-wave velocity model. A finer grid of 0.5x0.5x0.5km3 has been set to compute the theoretical arrival travel times at each station through a finite-difference solution of the eikonal equation. The model parameters have been inverted using LSQR method. The best regularization parameter of the inversion has been obtained from the trade-off curve between the model parameters and the data variances. The Derivative Weight Sum and the checkerboard tests have been performed to assess the resolved area of the map.

The preliminary results show a significant increase of VP and VS velocity at depth of about 6 km beneath Mt. Pollino. This interface likely corresponds to the top of the Apulian platform. A low VP, low VP/VS anomaly is found above the eastern cluster of seismicity, and a low VP, high VP/VS anomaly appear north and south-east of the sequence. The latter is spatially consistent with the fluid-rich volume suggested by the results of attenuation analysis. Further analyses will follow to provide more insights about this complex sequence and, in a broader view, about similar swarm-like sequences.

How to cite: Napolitano, F., Amoroso, O., Vitale, V., Gervasi, A., La Rocca, M., and Capuano, P.: ​Elastic properties and fluid abundance in the source volume of the 2010-2014 Pollino seismic sequence from P and S wave tomography, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1682,, 2021.

Giuseppe Davide Chiappetta, Anna Gervasi, and Mario La Rocca

We studied two seismic swarms occurred recently in Calabria, one in the Mesima-valley and one near Albi. Earthquakes were located by manually picking P and S waves. A search for clusters of events characterized by similar waveform was done, then the relative location was performed for any clusters. The focal mechanism was computed for as many events as possible, comparing the observed seismograms with synthetic signals for events of M>2.8, and considering the polarity of P and S waves for smaller events. For very small earthquakes we tried an estimation of the focal mechanism by comparison of the few clear signals with the recordings of stronger events. This analysis is aimed at investigating whether the many earthquakes of a swarm are produced by the same fault or by faults characterized by different orientation.

The Mesima valley area was affected by a seismic swarm that begun with a M3.6 earthquake on May 26, 2019. More than 140 events of smaller magnitude occurred in the same area during the following month. The relative location shows a hypocenter distribution with depth between 16 and 19 km and elongated for about 2 km in the NE-SW direction. The seismogenetic volume estimated from the relative location is of about 12 km3. The focal mechanisms computed for the 9 strongest events of the swarm are very similar among them, indicating a dip-slip normal kinematics. The comparative observation of P-wave polarity suggests that the most events of this swarm were likely generated by the same fault. In fact, even very small earthquakes (M<1.5) for which we can't give a reliable estimate of the focal mechanism, are characterized by P wave of the same polarity of stronger events at the stations around the epicenter.

Albi seismic swarm is one of the most interesting occurred in the central-eastern part of Calabria during the last 10 years. It begun on January 16, 2020, with a M3.8 earthquake, followed by more than 120 events in a month, and many others later. Detailed analyses were performed on as many earthquakes as possible, including absolute location, search for clusters of similar events and their relative location, and the estimation of focal mechanism. Results clearly indicate that this swarm was generated by a much greater seismogenetic volume if compared with the Mesima valley swarm. In fact hypocenters are much more spread, forming a cloud in the 6-12 km depth range, with a volume of at least 30-40 km3, and without any clear shape or direction. The search for clusters gave many families of similar events. Events of different clusters show waveforms quite different among them. Sometimes earthquakes located very near to each other have opposite P-wave polarity at the same station. Focal mechanisms confirm the heterogeneity of this swarm. The only common feature is the normal kinematics, while strike and dip cover wide ranges of values. Therefore we conclude that this swarm was generated by many small faults with different directions, activated by an extensional stress field.

How to cite: Chiappetta, G. D., Gervasi, A., and La Rocca, M.: Comparison of two low magnitude seismic swarms in Calabria (Italy), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4532,, 2021.

David Essing, Piero Poli, and Glenn Cougoulat

The Alto Tiberina fault (ATF) system (Northern Apennines, Italy) is dominated by a low-angle normal fault with syn- and antithetic splay faults located in the hanging wall. Starting in August 2013 the hanging wall has been affected by a swarm-like sequence that lasted until the end of 2014. Within this period more than ~20k events are listed to have nucleated along the same fault segment with the largest events having magnitudes of ~Mw 3.9.

In this study we aim to constrain the physical forces driving the swarm-like sequence (e.g. pore pressure diffusion, transient slow slip) in this fault segment by combining a template matching approach with continuous seismic data from a borehole array deployed in the near field of the ATF. This array approach helps us to identify small events which are hidden in the background noise and usually undetected with conventional picking approaches.

We are able to extend the preexisting catalog by a factor > 5. The new detected events decrease the magnitude of completeness and the inter-event time resolution. We use the extended catalog to analyze the spatio-temporal evolution, scaling properties and statistical behavior to enhance insights on the physical forces driving this swarm like sequence.

How to cite: Essing, D., Poli, P., and Cougoulat, G.: Insights into a tectonic swarm-like seismic Sequence related to a Low Angle Normal Fault system from a Seismic Catalog enhanced by Template Matching, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10143,, 2021.

Luca Carbone, Rita de Nardis, Giusy Lavecchia, Laura Peruzza, Enrico Priolo, Adelaide Romano, and Alessandro Vuan


During the seismic sequence which followed the devastating L’Aquila 2009 earthquake, on 27 May 2009 the OGS (Istituto Nazionale di Oceanografia e di Geofisica Sperimentale) and the GeosisLab (Laboratorio di Geodinamica e Sismogenesi, Chieti-Pescara University) installed a temporary seismometric network around the Sulmona Basin, a high seismic risk area of Central Italy located right at SE of the epicentral one. This area of the central Apennines is generally characterized by low level seismicity organized in low energy clusters, but it experienced destructive earthquakes both in historical and in early instrumental time (Fucino 1915 =XI MCS, Majella 1706 =X-XI MCS, Barrea 1984 =VIII MCS).

From the 27 May 2009 to 22 November 2011, the temporary network provided a huge amount of continuous seismic recordings, and a seismic catalogue covering the first seven months of network operation (-1.5≤ML≤3.7, with a completeness magnitude of 1.1) and a spatial area that stretches from the Sulmona Basin to Marsica-Sora area. Aiming to enhance the detection of microearthquakes reported in this catalogue, we applied the matched-filter technique (MFT) to continuous waveforms properly integrated with data from permanent stations belonging to the national seismic network. Specifically, we used the open-source seismological package PyMPA to detect microseismicity from the cross-correlation of continuous data and templates. As templates we used only the best relocated events of the available seismic catalogue. Starting from 366 well located earthquakes we obtain a new seismic catalogue of 6084 new events (-2<ML<4) lowering the completeness magnitude to 0.2. To these new seismic locations, we applied a declustering method to separate background seismicity from clustered seismicity in the area. All the seismicity shows a bimodal behaviour in term of distribution of the nearest-neighbor distance/time with the presence of two statistically distinct earthquake populations. We focused the attention on two of these clusters (C1 and C2) that numerically represent the 60% of the catalogue. They consist in 2619 and 995 events, respectively, with magnitude -2.0<ML<3.6 and -0.5<ML<3.2 occurred in Marsica-Sora area. C1 shows the typical characteristics of a seismic swarm, without a clear mainshock, but with 8 more energetic events (3.0≤ML≤3.5); the temporal evolution is very articulated with a total duration of one month with different bursts of seismicity and characteristic time extension of approximately one week. C2 instead has a different space-time evolution and consists of different swarm-like seismic sequences more discontinuous in comparison with C1. These swarms are described in greater detail to investigate the influence of overpressurized fluids and their space-time distribution.

How to cite: Carbone, L., de Nardis, R., Lavecchia, G., Peruzza, L., Priolo, E., Romano, A., and Vuan, A.: Enahancing the detail on low-level seismicity and swarms in central-southern Italy by template matching, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16045,, 2021.

Katrin Hannemann, Matthias Ohrnberger, Nikolaus Lerbs, Dorina Domigall, Marius Isken, René Voigt, Daniel Vollmer, Ralf Bauz, Jakub Klicpera, Lutz Sonnabend, Michael Korn, Frank Krüger, Tomáš Fischer, and Torsten Dahm

Within the ICDP project “Drilling the Eger Rift”, we focus on the German-Czech border region West Bohemia/ Vogtland which is known for its earthquake swarms. These swarms are clusters of small magnitude (ML<4) earthquakes which are supposed to be linked to the rise of fluids with mainly mantle origin. We aim to improve the seismological observation of these small magnitude earthquakes and related processes especially at frequencies above 100 Hz by installing three dense small aperture 3D arrays. Each single 3D array will consist of a 400 m deep vertical array borehole installation and a small aperture (400 m) surface array.

The drill site S1 in Landwüst and its surroundings serve as pilot site for the first installation. The borehole chain consists of eight 3-component 10 Hz geophones and the continous recordings are sampled with 1000 Hz. In parallel, twelve surface stations are installed which are equiped with 4.5 Hz geophones. The data were recorded with 400 Hz sampling rate at most locations, but at some selected stations we additionally record data with 1000 Hz sampling rate being the desired sampling rate for the final array configuration. Due to the high sampling rates and the high frequency content of the recorded earthquake signals, local site conditions may lead to non-coherent recordings for different parts of the array which have a major influence on the overall array performance. However, preliminary results from broad band frequency wave number analysis (5-180 Hz) in a moving time window (0.2 s) with first test installation data also indicate that the coherency across the array site is still high enough to clearly identify P and S waves from local earthquakes.


In the period December 2020 – January 2021, an earthquake swarm took place with two activity clusters in Nový Kostel (Czech Republic) and Obertriebel/ Oelsnitz (Vogtland, Germany) about 20 km apart. This swarm was recorded by both borehole stations and surface stations in Landwüst. Preliminary results show that more than 14000 events can be identified at the borehole stations and that about 70-80% of these events are also observed at the surface stations. For small earthquakes, mainly the S wave can be identified, but also impulsive P waves are clearly visible at the surface stations. These high frequency waves (up to 230 Hz at the surface) show a good coherency across the surface array. At the borehole stations, we observe an even higher frequency content up to 300 Hz and more. We present recordings from selected events to analyse frequency content and coherency across the 3D array.

How to cite: Hannemann, K., Ohrnberger, M., Lerbs, N., Domigall, D., Isken, M., Voigt, R., Vollmer, D., Bauz, R., Klicpera, J., Sonnabend, L., Korn, M., Krüger, F., Fischer, T., and Dahm, T.: High frequency array observations of December 2020 swarm at surface and borehole stations at ICDP Eger Rift site Landwüst (Vogtland), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9561,, 2021.

Ali Salama and Tomas Fischer

During the West-Bohemia/Vogtland earthquake swarms thousands of events are detected within short periods of few days, whose preliminary location is provided by an automated procedure. The manually verified high quality catalog is provided with some delay and is usually less complete than the automatic one.

We developed a template matching procedure combined with differential time measurement and double difference location whose application in real time will allow to provide precise hypocentre locations for much larger data set than provided by the manual processing. Among others, the template matching approach includes flexible setting of the time difference between P and S waves which allows for event detection in a wider distance to the template’s hypocentre. This makes the size of the template dataset small enough to allow for efficient detection process.

Our application of the template matching approach is aimed at identifying repeated activation of some patches during the swarms and weak background activity in the intermediate periods. Detecting and analyzing the repeating earthquakes will help revealing the continuing background activity and identifying fault areas that are active permanently. This will point to the possible sources of fluids or aseismic creep that are supposed to play significant role in swarm generation.

How to cite: Salama, A. and Fischer, T.: Improving the detection capability of the West Bohemian network by template matching approach , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15429,, 2021.

Josef Vlcek, Roman Beránek, and Tomáš Fischer

In past decades, a significant effort was spent to find the origin of recurring earthquake swarms in West-Bohemia/Vogtland. Widespread understanding accepts that crustal fluids migration along the fault zones is responsible for earthquake triggering in this area. Recently, a new model was suggested, which tests the hypothesis whether the diffusion of hydraulically induced pore pressure could be a valid trigger mechanism. In this approach the precipitation signal was transformed by diffusion equation to the hypocenter depth and statistically compared with the earthquake occurrence in time and concluded that at least 19% of the seismicity could have been triggered by rain. 

In our study we apply a different approach to verify the validity of these results. We use two types of rain signal on the input which is compared with the time series of earthquake weekly rate for the past 25 years. To remove the strong episodic character of the swarm seismicity we use a declustered seismic catalog, which is characteristic by almost continuous seismic activity.

The rain signal is represented first by the precipitation data and second by the water level data in the Horka reservoir, which is located above the main focal zone of Nový Kostel. We test the possible relation to the earthquake swarm activity by cross correlating both the rain signal types and the seismicity rate. To amplify the possible seasonal periodicity of the data we stacked the explored time series data (precipitation, water level and seismic activity) according to their occurrence date in a single year. The results show that in any of the input data and seismicity do not correlate. 

In the next step, we tested the possible (annual) periodicity of the data in question by the singular spectral analysis (SSA), which is a sensitive method to identify possible periodic signals in the presence of noise. While the water level data showed a striking peak for the period of 1 year, any indication of annual periodicity was never found in the seismicity data. Accordingly, we conclude that our analysis has shown no influence of the precipitation or the water level fluctuations in the Horka dam to the earthquake swarm activity in West Bohemia/Vogtland.

How to cite: Vlcek, J., Beránek, R., and Fischer, T.: Are the recurring earthquake swarms in West-Bohemia rain triggered?, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14986,, 2021.

Ana L. Lordi, Maria C. Neves, and Susana Custódio

Hydrological loads can be either surface loads induced by precipitation, changes in water levels at crater volcanic lakes, or subsurface loads created by seasonal changes in groundwater levels. These may contribute to strain and stress transients that trigger small earthquake swarms at faults that are already near failure. This work focusses on how annual and multi-annual stress changes of hydrological origin may affect the generation of seismic sequences on several tectonic settings, such as the New Madrid Seismic Zone and the Azores. The New Madrid seismic Zone is used as a benchmark test study region, while the Azores has been chosen for its intense seismic activity of both tectonic and volcanic origin. The magnitude of the hydrologically derived variations in stress is small compared with the long-term tectonic stresses, so we look for seasonal and inter-annual modulations of the earthquake occurrence rate. This requires the manipulation of seismic catalogues and the use of statistical methods to check if the seasonal and inter-annual variations are statistically significant, and not the result of extreme climatic events. The impact of hydrologic loads on faults is addressed using high-quality time series of seismic sequences, rainfall and other loads produced by variations in water levels, methods of decomposition and reconstruction of geophysical time series (SSA and wavelet transform) to identify modes of oscillation, and correlation analysis to recognize common patterns in seismicity and water loads. The results provide the first assessment of cyclic variations in seismicity and its relationship with atmospheric disturbances and hydrologically-driven load in the Azores region, and contributes to improve our understanding of the physics of earthquake triggering processes. The authors would like to acknowledge the financial support FCT through project UIDB/50019/2020 – IDL. This is a contribution to the RESTLESS project PTDC/CTA-GEF/6674/2020.

How to cite: Lordi, A. L., Neves, M. C., and Custódio, S.: Investigating the relationship between seismic sequences and hydrological loading in the Azores, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2654,, 2021.

Vaibhav Vijay Ingale, Sara Bazin, and Jean-Yves Royer

In 2018, two earthquake swarms occurred along segments of the ultra-slow Southwest Indian Ridge (spreading rate: 14-15 mm/a). The first swarm is located at the spreading-ridge intersection with the Atlantis Fracture Zone and comprises 9 Mw > 5.0 events (GCMT catalogue) and about 227 lower magnitude events (ISC catalogue), spanning over 9 days (July 10-18). The second crisis is more of a cluster of events focusing near a discontinuity, 220km away from the Indian Triple Junction and comprises 6 Mw > 5 events (GCMT) and 87 lower magnitude events (ISC catalogue), spanning over 30 days (September 28 to October 27). All focal mechanisms (GCMT) indicate normal faulting for both swarms. These two swarms are examined using hydroacoustic records from the OHASISBIO network with 7 to 9 autonomous hydrophones moored on either side of Southwest Indian Ridge.

The first swarm initiates with a Mw=4.9 event (July 10 2018, 03h55) which triggers numerous events with an average of ~250 events per day for the first three days (July 10to 12), propagating in the NE direction. After this, the seismic activity ceases down along with a sparse distribution of events until another burst of activity initiating after July 15, lasting for 3 days and comprising of several high intensity events. Overall, this swarm includes ~1100 hydroacoustic events spanning over 13 days.

The second swarm, further east, starts with two events, Mw=5.5 and 5.6 (Sept. 28 2018, 6h21 and 7h06), followed by a few discrete events. After 3 days, a dense cluster of events initiates with a Mw=5.4 event (October 1st, 18h16) and lasts for 7 days (~415 events per day) and decreases till the end of October. Two additional sub-swarms occur on October 1st and on October 6, both propagating towards the NE. Several other high intensity events occur October 10, after which seismic activity propagates towards the SE and fades away until October 27. Overall, this swarm includes ~5000 hydroacoustic events spanning over 33 days.

The number of events per day is thus larger for the second swarm than for the first one. Also, event source levels are in average smaller in the second crisis than in the first one. Further analyses of these characteristics, along with the different geographical and time distribution of the ~6000 acoustic events (vs ~300 events in the land-based catalogues), provide insights on the onset and on the tectonic or magmatic origin of these two contrasting swarms.

How to cite: Ingale, V. V., Bazin, S., and Royer, J.-Y.: Hydroacoustic observations of Two Contrasted Seismic Swarms along the Southwest Indian Ridge in 2018 , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5131,, 2021.

Diana Roman, Federica Lanza, John Power, Cliff Thurber, and Thomas Hudson

We investigate the processes driving a significant earthquake swarm that occurred between June and December 2020 on Unalaska Island, Alaska, ~12 km southeast of the summit of Makushin Volcano. The swarm was energetic, with two M>4 events that were widely felt by the population in Dutch Harbor, ~ 15 km west of the epicenters. This is the strongest seismic activity ever recorded at Makushin since instrumental monitoring began in 1996. To date, no eruptive activity or other surface changes have been observed at the volcano in satellite views, webcam images, GPS or InSAR. Seismic swarms close to volcanoes are often associated with the onset of unrest that can lead to eruption. However, determining whether they reflect magmatic rather than tectonic stresses is challenging. Here, we integrate information from space-time patterns of the hypocenters of the swarm earthquakes with their double-couple fault-plane solutions (FPS). We relocate swarm events using double-difference relocation techniques and a 3D velocity model. We find that most of the events cluster into two perpendicular lineaments with NW-SE and SW-NE orientations, but no apparent migration in time towards a preferred fault. On the one hand, the lack of temporal migration (with both faults slipping concurrently), and FPS for M3+ events consistent with regional stresses, seem to indicate a tectonic driving process. On the other hand, FPS for the lower-magnitude earthquakes have 90°-rotated P-axes perpendicular to the regional principal stress orientation, providing strong evidence for dike inflation/magma intrusion. Coulomb stress modeling indicates that the rotated FPS are best explained by an inflating dike to the SW of the swarm epicenters, in a zone of long-term elevated seismicity. This complex overlapping of regional and magmatic stresses is also evident in the statistical analysis of the sequence, which started as a main-shock/aftershock sequence with the first event having the largest magnitude, and evolved into a swarm sequence indicative of a more pronounced role of magmatic processes.

How to cite: Roman, D., Lanza, F., Power, J., Thurber, C., and Hudson, T.: Complex magmatic-tectonic interactions during the 2020 Makushin Volcano, Alaska, earthquake swarm, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16252,, 2021.

Seyyedmaalek Momeni
Seismic history of the North Tabriz fault (NTF), the main active fault of Northwestern Iran near Tabriz city, and its relation to the Sahand active Volcano (SND), the second high mountain of the NW Iran, and to the 11 August 2012 Ahar-Varzaghan earthquake doublet (Mw6.5&6.3) (AVD), is investigated. I infer that before AVD seismicity of the central segment of NTF close to SND was very low compared to its neighbor segments. Magmatic activities and thermal springs near central NTF close to Bostan-Abad city and low-velocity anomalies reported beneath SND toward NTF in tomography studies suggest that the existing heat due to SND magma chamber has increased the pore-fluid pressure that overcomes the effective normal stress on the central NTF, resulting in its creep behaviour. Two peaks of cumulative scalar seismic moments of earthquakes observed on both lobes of the creeping segment, confirming the strong difference in the deformation rate between these segments. On 2012, AVD struck in the 50 km North of NTF, in the same longitude range to SND and with the same right-lateral strike-slip mechanism to NTF, as a result of partial transfer of the right-lateral deformation of NW Iran toward the North of NTF on the Ahar-Varzaghan fault system. A cumulative aseismic slip equal to an Mw6.8 event is estimated for the creeping segment of NTF, posing half of the 7mmy-1 geodetic deformation has happened in the creep mode. This event has transferred a positive Coulomb stress field of >1 bar on the AVD and triggered them. Also, the western and eastern NTF segments received >4 bar of positive Coulomb stresses from the creeping segment and are probable nucleation locations for future earthquakes on NTF. The observed creep may be the reason for the NTF segmentation during the 1721AD M7.6 and 1780 AD M7.4 historical earthquakes.

How to cite: Momeni, S.: Triggering of the 2012 Ahar-Varzaghan earthquake doublet (Mw6.5&6.3) by the Sahand Volcano and North Tabriz fault (NW-Iran); Implications on the seismic hazard of Tabriz city, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-495,, 2021.

Gaspard Farge, Claude Jaupart, and Nikolaï Shapiro

Many subduction zones host intermittent, low-frequency, low-magnitude seismic activity emitted from the vicinity of the plates' interface. For instance, in Guerrero, Mexico, deep (30--50 km) low-frequency earthquakes (LFEs) occur in bursts, and migrate in cascades along the subduction interface. Those patterns are often attributed to episodic pulses of fluid pressure and slow slip that travel within the fault zone. However, the dynamic behavior of the permeable system in which fluid-pressure circulates remains a blindspot in most models of tremor generation, even as geological observations report pervasive imprint of strong, localized fluid pressure and permeability variations in its source region.

In order to analyze the role of such processes in generating tremor, we design a simple model of how fluid pressure and permeability can interact within the subduction interface, and generate realistic, tremor-like patterns. It is based on seismic source triggering and interaction in a permeable channel. The latter contains a number of low-permeability plugs acting as elementary fault-valves. In a mechanism akin to erosive burst documented in porous media, valve permeability abruptly opens and closes in response to the local fluid pressure. The brutal pressure transient and/or mechanical fracturing associated with valve opening acts as the seismic source of an LFE-like event. The strong fluid pressure transient that it triggers allows valves to interact constructively: as a valve breaks open, neighbor valves are more likely to break. This interaction therefore leads to cascades and migrations of synthetic seismicity along the model fault channel, that can synchronize into larger bursts of activity that migrate more slowly along the channel. In our model, valve activity draws patterns of that closely resemble tremor patterns in Guerrero and other subduction zones.

The input metamorphic fluid flux at the base of the channel exerts a key control on the occurence of and distribution of synthetic tremor in space and time. A weak input flux will not allow valves to open, conversely a strong flux will not allow them to close. In both cases, no activity will occur. However when the value of the fluid flux is intermediate, permanent regimes of sustained activity arise. Depending on its value, activity can be strongly time-clustered, quasi-periodic or random but constant in time.

Our model is based on a simple yet powerful and realistic description of the permeability and its dynamics in fault zones. It allows for new interpretations of low-frequency seismicity in terms of effective flux and fault-zone permeability, both for long-term regimes and finer scale, transient dynamics. Eventually, it could lead to deep enhancements of our understanding of fault-zone hydraulic processes and how they are coupled with fault-slip.

How to cite: Farge, G., Jaupart, C., and Shapiro, N.: Clogging and un-clogging of the subduction plumbing system may generate tremor-like patterns, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2124,, 2021.

Thoralf Dietrich and Eva P.S. Eibl

The 2014-2015 Holuhraun eruption was the largest eruption in Iceland in the last 230 years. After magma ascended below the Bárðarbunga volcano’s icecap, an about 2 week long lateral migration of earthquakes was observed; later interpreted as dike formation in 4km to 6km depth. An eruption started on 29th and 31st of August 2014 at Holuhraun. During dike formation and eruption a long-lasting seismic signal called tremor was recorded by seismometers in the area. Eruptive tremor emerged with the onset of the eruptions on 29th and 31st of August . Tremor sources were located and interpreted in the context of the fissure and the lava flow field. However, a complete geophysical model to explain these is missing. Our starting point is the model on tremor generation based on conduit wall vibrations exited by laminar flow (B. Julian 1994) to replicate the observed tremor signals. We performed a grid search and compare it with other models. In the range of rock parameter tolerance, we present implied characteristics of frequency and amplitude of the signals, if the Julian model were used as explanation for the tremor signals.

How to cite: Dietrich, T. and Eibl, E. P. S.: Analyzing the tremor of the Holuhraun eruption 2014-2015 using tremor modelling, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14520,, 2021.

Yu-Fang Hsu, Hsin-Hua Huang, and Ray Y. Chuang

Spatiotemporal evolution of earthquake clusters can give insights into fault geometry, triggering process, and potential interaction with fluid and heat. Taiwan is one of the most active orogenic belts with high deformation rate and complex crustal structures, so it is expected to observe seismicity driven by varying mechanisms among different geological processes. For investigating the tectonic complexity and the triggering processes of seismicity in Taiwan, a high-quality and robust catalog of earthquake clusters is critical. This study collected a long-term-effort earthquake catalog from the Central Weather Bureau from 1990/01 to 2018/06 and produced the earthquake cluster and background seismicity catalogs by four different declustering methods. Among which, the statistics-based nearest neighbor approach (NNA) performs most desirably for passing the Poisson process statistic tests while also remaining more events. We further classified the extracted earthquake clusters into the typical mainshock-aftershock (M-A) sequences and the swarms. Most of the M-A sequences are distributed near the Western Foothill. The asperity sizes, duration, and cluster event numbers all show positive correlations with mainshock magnitude. In contrast, the swarms are mainly distributed in the northern and southern Central Range and the northern Hualien regions. The lower correlation of the asperity sizes, duration, and swarm event numbers with the mainshock magnitude is showed in swarms. Moreover, we find that some of the swarm may be driven by fluid diffusion and spatial correlated with the high heat flow and spring regions.

How to cite: Hsu, Y.-F., Huang, H.-H., and Chuang, R. Y.: Earthquake Cluster Analysis with Nearest Neighbor Approach in Taiwan, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8629,, 2021.