Hypocenter Distribution Research Articles

The role of fluids in earthquake swarms in northeastern Noto Peninsula, central Japan: insights from source mechanisms

AbstractA prolonged earthquake swarm has persisted since June 2018 in northeastern Noto Peninsula (central Japan), with activity focused into distinct southern, western, northern, and eastern clusters. To explore the role of fluids in the occurrence of this swarm, we analyzed the focal mechanisms of the earthquakes occurring from 1 January 2018 to 30 November 2022 and performed stress tensor inversions. The western, northern, and eastern clusters were dominated by a reverse fault-type mechanism with a horizontal P-axis oriented NW–SE. One of the nodal planes of those mechanisms aligns closely with the precisely relocated hypocenter distribution. The stress fields in these three clusters, as determined by stress tensor inversion, have maximum principal stresses oriented horizontally in the NW–SE direction and minimum principal stresses oriented vertically, aligning with the regional stress field. From these focal mechanisms and this stress field, we derived small misfit angles and large slip tendencies. These findings suggest that, in these three clusters, fluids diffused into faults largely aligned with the regional stress field, resulting in earthquakes with compatible focal mechanisms. Conversely, normal and strike-slip fault-type focal mechanisms, which are unfavorable to the regional stress field, dominated in the southern cluster. The estimated stress fields, deviating from the regional stress field, have maximum principal stresses closer to vertical and minimum principal stress oriented horizontally in the ENE–WSW direction. For earthquakes deeper than 15 km, this local stress field results in relatively larger misfit angles and lower slip tendency than the other clusters. These results suggest that those earthquakes in the southern cluster occurred on misoriented fault planes due to elevated pore fluid pressures. These findings provide strong evidence of the ascent of high-pore-pressure fluids from depth in the southern cluster and their subsequent diffusion into a southeast-dipping fault zone. Graphical Abstract

Triggering the 2022 eruption of Mauna Loa

Distinguishing periods of intermittent unrest from the run-up to eruption is a major challenge at volcanoes around the globe. Comparing multidisciplinary monitoring data with mineral chemistry that records the physical and spatio-temporal evolution of magmas fundamentally advances our ability to forecast eruptions. The recent eruption of Mauna Loa, Earth’s largest active volcano, provides a unique opportunity to differentiate unrest from run-up and improve forecasting of future eruptions. After decades of intermittent seismic and geodetic activity over 38 years of repose, Mauna Loa began erupting on 27 November 2022. Here we present a multidisciplinary synthesis that tracks the spatio-temporal evolution of precursory activity by integrating mineral and melt chemistry, fluid inclusion barometry, numerical modeling of mineral zoning, syn-eruptive gas plume measurements, the distribution and frequency of earthquake hypocenters, seismic velocity changes, and ground deformation. These diverse data indicate that the eruption occurred following a 2-month period of sustained magma intrusion from depths of 3–5 km up to 1–2 km beneath the summit caldera, providing a new model of the plumbing system at this very high threat volcano. Careful correlation of both the geochemistry and instrumental monitoring data improves our ability to distinguish unrest from the run-up to eruption by providing deeper understanding of the both the monitoring data and the magmatic system—an approach that could be applied at other volcanic systems worldwide.

Quasi-real-time earthquake relocation and monitoring in the northeastern Noto Peninsula

The seismicity rate markedly increased in the northeastern Noto Peninsula of Ishikawa Prefecture around the end of 2020, with an Mw6.2 event on 5 May, 2023, followed by many aftershocks. Previous earthquake relocation studies have detected upward migration of microearthquakes via multiple faults and clusters, suggesting the involvement of crustal fluids in this sequence. Since some active faults exist near the source region, there was concern that this sequence could lead to a larger earthquake; it became a reality with the Mw7.5 earthquake on 1 January 2024. This study aims to develop an algorithm to precisely relocate microearthquake hypocenters in quasi-real time for better monitoring. A fine view of seismicity requires relative relocation methods such as the double-difference (DD) method with numerous arrival time difference data precisely measured by the waveform correlation analysis. However, the standard DD method has the disadvantage of huge computational costs when data increase, making it unsuitable for real-time monitoring in such situations. We developed a quasi-real-time algorithm that relocates only new earthquakes using the DD method each time a new time window of data is added. The major improvement is that our method incorporates a traditional simple relative relocation and preserves constraints between different time windows; the relative locations of new events are constrained from reference events that were already relocated in the previous time windows. We tested a daily relocation algorithm on 11,546 events from 19 June, 2022, to 31 May, 2023, in the Noto Peninsula earthquake sequence. We found that our modification substantially reduced artificial hypocenter offsets between different time windows and succeeded in resolving the fine fault structures from the cloud-like distribution of initial hypocenters. If we do not impose constraints between different windows, the relocated hypocenters are scattered and do not show fine planar structures. Our algorithm greatly reduces the computational cost, allowing for quasi-real-time earthquake relocation and monitoring. We hope this algorithm will help monitor the spatio-temporal distribution of future earthquake sequences.Graphical

Illuminating the Hierarchical Segmentation of Faults Through an Unsupervised Learning Approach Applied to Clouds of Earthquake Hypocenters

AbstractWe propose a workflow for the recognition of the hierarchical segmentation of faults through earthquake hypocenter clustering without prior information. Our approach combines density‐based clustering algorithms (DBSCAN and OPTICS), and principal component analysis (PCA). Given a spatial distribution of earthquake hypocenters, DBSCAN identifies first‐order clusters, representing regions with the highest density of connected seismic events. Within each first‐order cluster, OPTICS further identifies nested higher‐order clusters, providing information on their number and size. PCA analysis is applied to first‐ and higher‐order clusters to evaluate eigenvalues, allowing discrimination between seismicity associated with planar features and distributed seismicity that remains uncategorized. The identified planes are then geometrically characterized in terms of their location and orientation in the space, length, and height. This automated procedure operates within two spatial scales: the largest scale corresponds to the longest pattern of approximately equally dense earthquake clouds, while the smallest scale relates to earthquake location errors. By applying PCA analysis, a planar feature outputted from a first‐order cluster can be interpreted as a fault surface while planes outputted after OPTICS can be interpreted as fault segments comprised within the fault surface. The evenness between the orientation of illuminated fault surfaces and fault segments, and that of the nodal planes of earthquake focal mechanisms calculated along the same faults, corroborates this interpretation. Our workflow has been successfully applied to earthquake hypocenter distributions from various seismically active areas (Italy, Taiwan, and California) associated with faults exhibiting diverse kinematics.

Linking the spatiotemporal distribution of static stress drops to source faults in a fluid-driven earthquake swarm, northeastern Noto Peninsula, central Japan

We investigated stress drops during an earthquake swarm in northeastern Noto Peninsula, central Japan, which is characterized by ongoing seismic activity in four clusters. We focused on the spatiotemporal distribution of the static stress drop and its relationship with the source faults of the earthquake swarm. Employing the empirical Green’s function method, we estimated static stress drops for 90 earthquakes of MJMA 3.0–5.4. We obtained logarithmic mean stress drops of 13 MPa and 19 MPa from P-wave and S-wave analyses, respectively, which were typical values for crustal earthquakes. We comprehensively analyzed the spatiotemporal distribution of static stress drops in the northern cluster due to the abundance of available data and clarity of fault structures there. We observed larger static stress drops for earthquakes along shallow portions of the source faults, as defined by the hypocentral distribution during a given period. Conversely, we observed smaller static stress drops for earthquakes at medial parts along the faults. These results suggest higher fault strength at shallower parts along the faults and reduced fault strength at medial parts. We attribute the high fault strength at shallow parts to low pore fluid pressure after only limited fluid diffusion near the fault terminus. In contrast, we attribute the reduction in fault strength at medial parts to high pore fluid pressure within the fault following penetration by migrating fluids.Graphical

Not trench-parallel but trench-normal source fault of the 1994 Hokkaido Toho-oki earthquake as revealed by the aftershock relocation using HypoDD

In the southwestern area of the Kurile Islands subduction zone, three large earthquakes including the Mw8.2 1969 and the Mw8.3 1994 Hokkaido Toho-oki earthquakes and the Mw7.8 1975 tsunami earthquake occurred in almost the same location. The three main shocks and their aftershocks were re-determined simultaneously using a double-difference earthquake location method, HypoDD. As a result, the 1969 and the 1975 main shocks and their aftershocks were concentrated near the upper surface of the subducting Pacific plate, indicating that both events were plate boundary megathrust earthquakes as shown by previous studies. On the other hand, the 1994 main shock and its aftershocks were distributed within the Pacific plate indicating that the 1994 event was an intraslab event as reported by previous studies. The result that interplate earthquakes and intraslab earthquakes could be appropriately distinguished is strong evidence that the accuracy of hypocenter determination is sufficiently high. The centroid moment tensor solution of the 1994 main shock has two nodal planes which are parallel or normal to the Kurile Trench. According to the high-resolution hypocenter distribution obtained in the present study, the aftershocks appear to locate on the nodal plane normal to the trench, suggesting that the source fault ruptured by the 1994 main shock was not parallel but normal to the Kurile Trench. Moreover, we found that the Coulomb failure stress change (ΔCFS) produced by the trench-normal fault plane is more likely to trigger the largest aftershock than ΔCFS produced by the trench-parallel fault plane.Graphical abstract

Relocation of Earthquakes Hypocenter Using the Double Difference Method in Bali and Its Surrounding Areas

A study on the relocation of earthquake hypocenters using the Double Difference method has been conducted in Bali and its surrounding areas. This research was carried out at the Center for Meteorology, Climatology, and Geophysics Region III Denpasar. The purpose of this study is to find out the distribution of earthquake hypocenters before and after the relocation and to know the continuity of earthquake hypocenters in the Bali region. The data used are arrival time data of 1,815 earthquake events from January 1, 2014 to December 31, 2023, and data from 26 earthquake recording stations. The relocation method applied is the Double Difference method, which is processed with ph2dt and HypoDD. The relocation that has been carried out shows a good hypocenter position which is characterized by a residual value of travel time that is close to 0. The distribution of the earthquake hypocenter in the Bali area before the relocation spread throughout the region and after the relocation, the position of the hypocenter was getting closer to the source of the earthquake, namely in the subduction zone under the Bali area as shown in the cross-section results. In addition, based on the results of the analysis of the hypocenter of the earthquake after relocation, it was found that the trend of the hypocenter continued toward the subduction subduction zone.

Hypocenter Distribution of Sinabung Volcano during January - December 2015

Hypocenter Distribution of Sinabung Volcano during January - December 2015

Hypocenter Distribution Analysis of Sinabung Volcano During January-May 2017

Hypocenter Distribution Analysis of Sinabung Volcano During January-May 2017

Seismotectonics of the Philippine and Taiwan Subduction Systems and Implications for Seismic Hazards

AbstractThe seismicity of the Philippines and Taiwan provides insight into the tectonics and seismic hazards of a region characterized by subduction and collision. We summarize the seismotectonics of the Philippines and Taiwan by documenting the distribution of hypocenters for earthquakes of magnitude M ≥ 4.6 and focal mechanisms for earthquakes of magnitude M ≥ 5.0 over ∼21 years. We quantify seismicity rates (earthquake frequency) and compare seismicity distributions with proposed tectonic and faulting models. 6,187 earthquakes of magnitude M ≥ 4.6 occurred between 1 January 2000 and 31 March 2021, 79% at depths <70 km and 70% having magnitudes M < 5.0. Approximately 88 earthquakes of magnitude M ≥ 5.0 occur per year, with 12 events of magnitude M ≥ 7.0 occurring since January 2000. Seismic activity decreases exponentially between 50 and 210 km depth at a rate ∼10% faster than the global average. Intermediate and deep earthquakes at depths >70 km trace the Wadati‐Benioff zones of subducting slabs, most of which are only seismically active to depths of ∼250 km. The distribution of earthquakes at depths >70 km is likely influenced by the subduction of young lithosphere, slab tearing, and phase boundary interactions between depths of 410 and 660 km. Shallow earthquakes at depths ≤70 km are generated by megathrust, crustal, and intraslab faulting. Crustal thrust and strike‐slip faulting are the most abundant and prevalent sources of damaging earthquakes. The Philippines and Taiwan are subject to high seismic risk, similar to nearby Indonesia.

Determination of Earthquake Depths Using Data of Cross-Sectional and Areal Deep Seismic Studies in Siberia

Abstract ––Information on the distribution of earthquake hypocenters for many seismically active zones of Siberia remains insufficient, which is associated with sparse networks of seismological observations. The paper presents the results of determining the earthquake depths in several seismogenic areas in the Altai-Sayan region, Cisbaikalia, Transbaikalia, and Yakutia using the travel time of longitudinal refracted waves from the Mohorovičič discontinuity (Pn waves) from earthquakes and the recently obtained information about the deep structure of these regions. The depth determination algorithm is tested using data from aftershocks of large earthquakes occurring in Tuva in 2011 and 2012 (ML = 6.7 and 6.8), recorded both by the regional seismological network and by a local group of seismic stations. Different methods are applied to reveal that some aftershock depths have a close match, including those for the main shocks – the Tuva-1 and Tuva-2 earthquakes. Another good match of earthquake depths is obtained using Рn waves with materials from regional and detailed studies of the Baikal branch of the Geophysical Survey of the Russian Academy of Sciences in the Muyakan activation site in the Baikal rift zone. The resulting data complement the information on the hypocenter of the main shock and confirm the change of the Muyakan activation cluster from large depths to shallow ones since the onset of activation in 2014. New information on the earthquake depths using Рn waves is obtained in Yakutia along the border of the largest Eurasian and Okhotsk tectonic plates. It is revealed that they decrease to 6–12 km as compared to higher depths of 20–30 km in adjacent areas. The resulting series of new information on the distribution of earthquake hypocenters using Рn waves is extremely important mostly because it indicates the possibility of identifying and redefining earthquake hypocenters using previous seismological observations in seismically active zones of Siberia.

Interpretasi Mekanisme Erupsi Gunung Api Lokon berdasarkan Aktivitas Kegempaan Pra-Erupsi 4 – 12 September 2014

Prior to the eruption of Lokon Volcano on September 13 2014 at 03.09 WITA, there were 62 volcanic earthquake events recorded by seismometers since September 4 2014. This research aims to interpret the eruption mechanism of Lokon Volcano based on seismic activity in the pre-eruption period. 04 - 12 September 2014. The distribution of source points for volcanic earthquakes, both hypocenters, can be visualized using the seismic method, namely by processing digital seimogram data using Swarm 3.1.0 software, GAD (Geiger's method with Adaptive Damping) and Origin 8. Data used, sourced from three stations, namely, Empung (EMP), Wailan (WLN), Kinilow (KIN). Lokon volcano. The hypocenter distribution of volcanic earthquakes is concentrated at a depth of 0 - 1,500 m above sea level and it was found that the aseismic area which is assumed to be a shallow magma pocket area from the Lokon volcano is at a depth of 1,500 - 3,000 m above sea level. The interpretation results show that the eruption mechanism of Lokon Volcano originates from disturbances in the volcanic system which are thought to be the result of tectonic earthquakes in the subduction zone. This causes the magma to rise upwards to supply shallow magma pockets until it reaches maximum pressure and an eruption occurs.

Magma intrusion process during pre-magmatic period (2010−2013) of Sinabung volcano as revealed by seismicity of volcano-tectonic and hybrid earthquakes

Sinabung volcano in Indonesia reactivated, with a series of phreatic eruptions, in August and September 2010. These eruptions were followed by various types of magmatic eruption at the summit, including lava dome growth, pyroclastic density currents, lava flow, and vulcanian eruptions, starting in December 2013. Prior to the magmatic eruptions, Sinabung exhibited two sequences of phreatic eruption and an increase in the seismicity of volcanic earthquakes. This study clarifies the progress of magma intrusion during the pre-magmatic period based on the hypocenter distribution, waveform similarity, seismic moment, rupture length, and stress drop for volcano-tectonic (VT) and hybrid earthquakes. It was found that the hypocenters of VT earthquakes are distributed north and northwest of the summit at a depth range of 1–10 km below sea level. Deep seismicity (4–10 km) alternated with shallow seismicity (2–4 km). The last shallow seismicity, from July to mid-December 2013, was different from previous seismicity in terms of higher intensity, as determined from the seismic moment, a temporary increase in the rupture length, and the migration of hypocenters towards the summit. The seismicity of VT earthquakes was altered by a swarm of hybrid earthquakes in mid-December. The hybrid earthquakes were smaller, in terms of seismic moment (mostly <3 × 1010 Nm), than the VT earthquakes. Their hypocenters were concentrated in the shallowest depth range (−0.5 to 1.5 km below sea level) directly below the summit, their source process was repeatable (they could be grouped into six earthquake families), their dominant peak was in a low-frequency range (2.5–4.5 Hz), and they had a relatively low stress drop (<0.15 MPa). This suggests that the swarm of hybrid earthquakes was induced by a frequent repeated fracture of fluid-filled cracks due to the intrusion of magma up to directly below the summit. The transition of the earthquake family and the change in its source parameters, namely an increase in the stress drop prior to the appearance of a lava dome and then a slight decrease, may reflect a gradual change in internal pressure in the hypocentral zone through the magma intrusion process.

Active Faults, Kinematics, and Seismotectonic Evolution during Tajogaite Eruption 2021 (La Palma, Canary Islands, Spain)

During the 2021 La Palma strombolian and fissure eruption, two faults were identified that controlled the spatial distribution of earthquake hypocenters and effusive eruptive vents. One of these faults has a NW-SE trend (Tazacorte Fault: TZF) and the other one shows an ENE-WSW trend (Mazo Fault: MZF). Previous works on fault structural analysis in La Palma indicated that the eruption zone was compatible with an extensional tectonic strain ellipsoid which activated normal-strike-slip directional faults at the confluence of TZF and MZF. These fractures were activated during the 2021 Tajogaite eruption, determining the NW-SE and WSW-ENE spatial distribution of vents. Both faults were mapped in real time during the volcanic eruption from fieldwork and remote sensing imagery (aerial drone images). We have collected more than 300 fracture data associated with the effusive vents and post-eruption seismic creep. Since the affected area was densely inhabited, most of these fractures affect houses and infrastructures. Some of the houses affected by the TZF were damaged 9 months after the eruption, although they were not damaged during the eruption. Surprisingly, these houses already had repairs made to the same fractures since 1980, giving information of previous fault creep movement. During the 2021 Tajogaite eruption, shallow seismicity was spatially related to both faults, suggesting a seismic behavior instead of the precedent creep movement. However, the lack of seismicity after the eruption indicates that the faults went back to creep aseismic behavior, similarly to 1980. The mapping and monitoring of these faults (TZF and MZF) is relevant bearing in mind that they have been active since 1980 and the post-eruptive phase of the 2021 volcanic eruption, which has to be included in the land use planning in areas affected by the volcanic eruption and creep movement. Furthermore, both faults could act as seismogenic sources triggering volcanic earthquakes with potential high macroseismic intensities and mass movements. The data presented here show the importance of having this type of study before the onset of the eruption, thus allowing a better interpretation of seismic data during volcanic unrest.

The 2022 Goesan earthquake of the moment magnitude 3.8 along the buried fault in the central Korean Peninsula

AbstractOn October 28, 2022, a moment magnitude (Mw) 3.8 earthquake occurred in Goesan, South Korea, typically characterized as a stable continental region. Herein, we analyze 42 earthquakes, including the Mw 3.8 earthquake, the largest foreshock (Mw 3.3), which preceded the mainshock by 17 s, and the largest aftershock (Mw 2.9). The primary aim of this study is to identify interactions among the seismic events. To this end, we utilized the permanent seismic networks with the closest station at 8.3 km from the epicenter, and the temporary network deployed eight hours after the mainshock’s occurrence. Relocation results delineate that the mainshock occurred at the southeastern tip of the hypocenter distribution of three foreshocks, trending west-northwest–east-southeast. The aftershocks form an overall spatially diffused seismic pattern that propagates toward both ends of the inferred lineament in the downdip direction. The rupture directivity of the mainshock, along with waveform similarity across the mainshock and foreshocks, confirms the inferred geometry, corresponding well with the focal mechanisms of the mainshock and the largest foreshock. We demonstrate that the change in Coulomb failure stress (ΔCFS) by the largest foreshock was positive where the mainshock occurred and that the mainshock generated ΔCFS capable of triggering the propagation of the aftershocks.

HYPOCENTER DISTRIBUTION ANALYSIS OF SINABUNG VOLCANO ERUPTION IN 2021 USING GEIGER'S WITH ADAPTIVE DAMPING (GAD) METHOD

Sinabung Volcano is located in the Karo Highlands, Karo Regency, North Sumatra, Indonesia, with a peak of 2460 meters above sea level. This volcano experienced an increase in volcanic activity in the April-October 2021 period, so it is necessary to analyse volcanic seismicity to determine the hypocenter distribution of the volcano. This study aims to determine the hypocenter distribution and classification of volcanic earthquake types based on seismogram recording data of Sinabung Volcano at four stations, namely LKW, BGR, MDD and SGR stations. The method used in this research is Geiger's method with adaptive damping (GAD) to determine the distribution of hypocenters. The results showed that the hypocenter distribution gathered below the crater centre with depths from 0.172 km to 7.139 km. The determination of the hypocenter distribution of volcanic earthquakes is expected to provide information on the depth of the earthquake to determine the movement of magma and the characteristics of events that occur during the pre-eruption process of Sinabung Volcano so that this data can be used for mitigation decision making when there is an increase in the activity of Sinabung Volcano.

Estimating the contribution of thermal processes to generation of seismicity in the junction zone of the Chuya Depression and the Kyrgyz Ridge of the Northern Tien Shan

This paper presents the attempt to estimate the contribution of thermal processes in the Earth’s crust to the generation of seismicity by the example of the seismically active region of the Chuya depression and the Kyrgyz ridge of the Northern Tien Shan. For this purpose, we use the models of temperature, petrophysical characteristics, and elastic moduli constructed in the previous works. The silica content model based on seismic tomography data is used to build a lithotype model. The constructed thermal conductivity model is utilized, along with the temperature model, to create a depth model of heat flow density. Based on the density, elastic moduli, and temperature models, the shear and thermoelastic stress models in the rocks are constructed. Their comparison with the distribution of earthquake hypocenters suggests that on the scale of the seismically active zone of the Chuya depression and the Kyrgyz ridge of the Northern Tien Shan, seismicity is mainly determined by the thermomechanical effect resulting from the hot ascending flow of acid magma from the upper mantle beneath the Muyunkum–Narat median mass.

Double-Difference Method for Relocating the Hypocenter of Aftershock Earthquakes in Seririt Singaraja

A method that does not require a main earthquake (master event) that can be used simultaneously to relocate a very large number of earthquakes with wide hypocenter separation is called the double-difference method. A method used to relocate the aftershocks in Seririt Singaraja on November 14 2019 with coordinate positions 113.478 – 115.181 East Longitude and 8.357 – 7.894 South Latitude. The earthquake data used in this research was accumulated from 85 BMKG seismic stations. Data analysis uses cross-correlation time differences which can increase the accuracy of travel time between the receiving station and the earthquake, thereby reducing errors in calculations. The double difference method used to relocate the earthquake in the Seririt Singaraja area showed that there was a shift in the location of the earthquake hypocenter before and after it was relocated. Horizontally and vertically, the distribution of earthquake hypocenters before and after being relocated occurs when there is a collection of location shifts. The results of this research were able to relocate 152 aftershocks properly. The main earthquake after being relocated was at a depth of 17 km, while the distribution of aftershocks was at a depth of around 6-25 km, so that it can more accurately describe the position of the earthquake source and is able to show clearer and easier to interpret structural patterns.

Fault roughness controls injection-induced seismicity

Surface roughness ubiquitously prevails in natural faults across various length scales. Despite extensive studies highlighting the important role of fault geometry in the dynamics of tectonic earthquakes, whether and how fault roughness affects fluid-induced seismicity remains elusive. Here, we investigate the effects of fault geometry and stress heterogeneity on fluid-induced fault slip and associated seismicity characteristics using laboratory experiments and numerical modeling. We perform fluid injection experiments on quartz-rich sandstone samples containing either a smooth or a rough fault. We find that geometrical roughness slows down injection-induced fault slip and reduces macroscopic slip velocities and fault slip-weakening rates. Stress heterogeneity and roughness control hypocenter distribution, frequency-magnitude characteristics, and source mechanisms of injection-induced acoustic emissions (AEs) (analogous to natural seismicity). In contrast to smooth faults where injection-induced AEs are uniformly distributed, slip on rough faults produces spatially localized AEs with pronounced non-double-couple source mechanisms. We demonstrate that these clustered AEs occur around highly stressed asperities where induced local slip rates are higher, accompanied by lower Gutenberg-Richter b-values. Our findings suggest that real-time monitoring of induced microseismicity during fluid injection may allow identifying progressive localization of seismic activity and improve forecasting of runaway events.

Reappraisal of volcanic seismicity at the Kirishima volcano using machine learning

Volcanic earthquakes provide essential information for evaluating volcanic activity. Because volcanic earthquakes are often characterized by swarm-like features, conventional methods using manual picking require considerable time to construct seismic catalogs. In this study, using a machine learning framework and a trained model from a volcanic earthquake catalog, we obtained a detailed picture of volcanic earthquakes during the past 12 years at the Kirishima volcano, southwestern Japan. We detected ~ 6.2 times as many earthquakes as a conventional seismic catalog and obtained a high-resolution hypocenter distribution through waveform correlation analysis. Earthquake clusters were estimated below the craters, where magmatic or phreatic eruptions occurred in recent years. Increases in seismic activities, b values, and the number low-frequency earthquakes were detected before the eruptions. The process can be conducted in real time, and monitoring volcanic earthquakes through machine learning methods contributes to understanding the changes in volcanic activity and improving eruption predictions.Graphical