Long-period Events Research Articles

Earthquake characteristics before eruptions of Japan's Ontake volcano in 2007 and 2014

AbstractWe detect earthquakes associated with Japan's Ontake eruptions in 2007 and 2014 using the match‐and‐locate method. Thirty‐seven times more events (4949) than the Japan Meteorological Agency (JMA) catalog (133) and 30 times more events (1880) than the JMA catalog (62) are detected during the eruptions of 2007 and 2014, respectively. The detected earthquakes are further classified into three different types: long‐period (LP) events, LP‐associated events, and volcano‐tectonic (VT) earthquakes. Both LP and LP‐associated events are observed in the two eruptions. We suggest that the observed intense seismicity before both eruptions indicates the rewaking of Ontake volcano and might be used to predict its impending eruption in the long term. The temporal persistence and spatial concentration of LP and LP‐associated events may be used to predict the eventual eruption and eruption location, and the maximum magnitude of the LP events may be used as an indicator of eruption magnitude.

Pressurization and depressurization phases inside the plumbing system of Mount Etna volcano: Evidence from a multiparametric approach

AbstractDuring 2013 Mount Etna volcano experienced intense eruptive activity at the summit craters, foremost at the New Southeast Crater and to a minor degree at the Voragine and Bocca Nuova (BN), which took place in two cycles, February–April and September–December. In this work, we mainly focus on the period between these cycles, applying a multiparametric approach. The period from the end of April to 5 September showed a gradual increase in the amplitude of long‐period (LP) events and volcanic tremor, a slight inflation testified by both tilt and GPS data, and a CO2 flux increase. Such variations were interpreted as due to a gradual pressurization of the plumbing system, from the shallowest part, where LP and volcanic tremor are located, down to about 3–9 km below sea level, pressure source depths obtained by both geodetic and CO2 data. On 5 September, at the same time as a large explosion at BN, we observed an instantaneous variation of the aforementioned signals (decrease in amplitude of LP events and volcanic tremor, slight deflation, and CO2 flux decrease) and the activation of a new infrasonic source located at BN. In the light of it, the BN explosion probably caused the instantaneous end of the pressurization, and the opening of a new vent at BN, that has become a new steady source of infrasonic events. This apparently slight change in the plumbing system also led to the gradual resumption of activity at the New Southeast Crater, culminating with the second lava fountain cycle of 2013.

Preparatory and precursory processes leading up to the 2014 phreatic eruption of Mount Ontake, Japan

We analyzed seismicity linked to the 2014 phreatic eruption of Mount Ontake, Japan, on 27 September 2014. We first relocated shallow volcano tectonic (VT) earthquakes and long-period (LP) events from August to September 2014. By applying a matched-filter technique to continuous waveforms using these relocated earthquakes, we detected numerous additional micro-earthquakes beneath the craters. The relocated VT earthquakes aligned on a near-vertical plane oriented NNW–SSE, suggesting they occurred around a conduit related to the intrusion of magmatic–hydrothermal fluids into the craters. The frequency of VT earthquakes gradually increased from 6 September 2014 and reached a peak on 11 September 2014. After the peak, seismicity levels remained elevated until the eruption. b-values gradually increased from 1.2 to 1.7 from 11 to 16 September 2014 then declined gradually and dropped to 0.8 just before the eruption. During the 10-min period immediately preceding the phreatic eruption, VT earthquakes migrated in the up-dip direction as well as laterally along the NNW–SSE feature. The migrating seismicity coincided with an accelerated increase of pre-eruptive tremor amplitude and with an anomalous tiltmeter signal that indicated summit upheaval. Therefore, the migrating seismicity suggests that the vertical conduit was filled with pressurized fluids, which rapidly propagated to the surface during the final 10 min before the eruption.

A brittle failure model for long‐period seismic events recorded at Turrialba Volcano, Costa Rica

AbstractA temporary seismic network, consisting of 23 broadband and six short‐period stations, was installed in a dense network at Turrialba Volcano, Costa Rica, between 8 March and 4 May 2011. During this time 513 long‐period (LP) events were observed. Due to their pulse‐like waveforms, the hypothesis that the events are generated by a slow‐failure mechanism, based on a recent new model by Bean et al. (2014), is tested. A significant number (107) of the LPs are jointly inverted for their source locations and mechanisms, using full‐waveform moment tensor inversion. The locations are mostly shallow, with depths < 800 m below the active Southwest Crater. The results of the decompositions of the obtained moment tensor solutions show complex source mechanisms, composed of high proportions of isotropic and low, but seemingly significant, proportions of compensated linear vector dipole and double‐couple components. It is demonstrated that this can be explained as mode I tensile fracturing with a strong shear component. The source mechanism is further investigated by exploring scaling laws within the data. The LPs recorded follow relationships very similar to those of conventional earthquakes, exhibiting frequency‐magnitude and corner frequency versus magnitude relationships that can be explained by brittle failure. All of these observations indicate that a slow‐failure source model can successfully describe the generation of short‐duration LP events at Turrialba Volcano.

Body Wave Separation in the Time-Frequency Domain

Separation of a seismogram into its individual constitutive phases (Pand S-wave arrivals, surface waves, etc.) is a long-standing problem. In this letter, we use a high-resolution time-frequency transform to achieve this and reconstruct their individual waveforms in the time domain. The procedure is illustrated using microseismic events recorded during a hydraulic fracturing treatment. The synchrosqueezing transform is an extension of the continuous wavelet transform combined with frequency reassignment. Its high-resolution time-frequency decompositions allow for separation and identification of Pand S-waves with subtly different frequency contents that would not be recoverable using short-term Fourier transforms due to its smearing in the frequency domain. It is an invertible transform, thus allowing for signal reconstruction in the time domain after signal separation. The same approach is applicable to other seismic signals such as resonance frequencies and long-period events and offers promising new possibilities for enhanced signal interpretation in terms of underlying physical processes.

Laboratory simulations of fluid/gas induced micro-earthquakes: application to volcano seismology

Understanding different seismic signals recorded in active volcanic regions allows geoscientists to derive insight into the processes that generate them. A key type is known as Low Frequency or Long Period (LP) event, generally understood to be generated by different fluid types resonating in cracks and faults. The physical mechanisms of these signals have been linked to either resonance/turbulence within fluids, or as a result of fluids ‘sloshing’ due to a mixture of gas and fluid being present in the system. Less well understood, however, is the effect of the fluid type (phase) on the measured signal. To explore this, we designed an experiment in which we generated a precisely controlled liquid to gas transition in a closed system by inducing rapid decompression of fluid-filled fault zones in a sample of basalt from Mt. Etna Volcano, Italy. We find that fluid phase transition is accompanied by a marked frequency shift in the accompanying microseismic dataset that can be compared to volcano seismic data. Moreover, our induced seismic activity occurs at pressure conditions equivalent to hydrostatic depths of 200 to 750 meters. This is consistent with recently measured dominant frequencies of LP events and with numerous models.

Array analysis of the seismic wavefield of long‐period events and volcanic tremor at Arenal volcano, Costa Rica

AbstractWe use wavefield decomposition methods (time domain cross correlation and frequency domain multiple‐signal classification) to analyze seismic data recorded by a dense, small‐aperture array located 2 km West of Arenal volcano, Costa Rica, and operated during 2.5 days. The recorded wavefield is dominated by harmonic tremor and includes also spasmodic tremor and long‐period (LP) events. We find that the initial stages of LP events are characterized by three different wave arrivals. These arrivals propagate with similar back azimuths pointing to the volcano summit (∼80°N) and increasing apparent slowness of 0.4, 1.1, and 1.7 s/km. Spasmodic tremors cannot be regarded as coherent signals. On the contrary, harmonic tremors are highly coherent, characterized by the stability of the apparent slowness vector estimates. Apparent slowness lays in the range 1–2 s/km. Back azimuths point in the general direction of the volcano but with a large variability (40–120°N). Nevertheless, there are long‐term variations and evidences of multiple simultaneous components in the harmonic tremor wavefield. These observations suggest that LP events and tremor are generated in a shallow source area near the volcano summit, although they do not share exactly the same source region or source processes. The tremor source is located in the shallowest part of the plumbing system, beneath the lava crust. This dynamic region is subject to complex fluctuations of the physical conditions. Degassing events at different locations of this region might generate variable seismic radiation patterns. The effects of topography and heterogeneous shallow structure of the volcano may amplify these variations and produce the wide directional span observed for volcanic tremor. On the other hand, the LP source seems to be more repeatable. LP events are likely triggered by fragmentation of the fluid flow in a slightly deeper portion of the volcanic conduits.

Advances in seismic monitoring at Deception Island volcano (Antarctica) since the International Polar Year

<p>Deception Island is an active volcano located in the south Shetland Islands, Antarctica. It constitutes a natural laboratory to test geophysical instruments in extreme conditions, since they have to endure not only the Antarctic climate but also the volcanic environment. Deception is one of the most visited places in Antarctica, both by scientists and tourists, which emphasize the importance of volcano monitoring. Seismic monitoring has been going on since 1986 during austral summer surveys. The recorded data include volcano-tectonic earthquakes, long-period events and volcanic tremor, among others. The level of seismicity ranges from quiet periods to seismic crises (e.g. 1992-1993, 1999). Our group has been involved in volcano monitoring at Deception Island since 1994. Based on this experience, in recent years we have made the most of the opportunities of the International Polar Year 2007-2008 to introduce advances in seismic monitoring along four lines: (1) the improvement of the seismic network installed for seismic monitoring during the summer surveys; (2) the development and improvement of seismic arrays for the detection and characterization of seismo-volcanic signals; (3) the design of automated event recognition tools, to simplify the process of data interpretation; and (4) the deployment of permanent seismic stations. These advances help us to obtain more data of better quality, and therefore to improve our interpretation of the seismo-volcanic activity at Deception Island, which is a crucial step in terms of hazards assessment.</p>

Fluid dynamics inside a “wet” volcano inferred from the complex frequencies of long-period (LP) events: An example from Papandayan volcano, West Java, Indonesia, during the 2011 seismic unrest

We present results of our study aimed at understanding the dynamics of fluids inside a “wet” volcano through the analysis of swarms of long-period (LP) events accompanying the 2011 seismic unrest at Papandayan volcano, West Java, Indonesia. Prior to this unrest, we measured an extremely high percentage (100 %) of CO2 in the ground at the summit crater, however with a very low value of SO2 flux (~6tons/day). Increase in volcanic activity was also observed from the records of a tiltmeter. A long-term inflation was followed by an abrupt deflation that took place concurrently with the swarms of LP events. Thereafter, swarms of local-tectonic (LT) and volcano-tectonic (VT) earthquakes started. We focus here on analyzing the LP events in the following manner. First, we estimate the source location of LP events by applying a 3-D non-linear hypocenter localization algorithm which includes topography. We then study the waveforms and spectral characteristics of LP events recorded at different stations and investigate whether or not these characteristics are due to source effects. Subsequently, we compute the oscillation frequencies (f) and the decay characteristics (Q factor) in the complex frequency domain of the coda part of the LP events by using the Sompi method which is based on a homogeneous auto-regressive (AR) equation. The rectangular fluid-filled crack model is used to estimate the physical processes related to the observed temporal variations in the complex frequencies. We divide the swarms of LP events into two intervals. The first interval occurred between June and July 2011 (48 LP events), while the second interval extended from September to October 2011 (36 LP events). The frequencies of LP events observed during these intervals range between 1.1 and 6.2Hz while the Q factors are widely scattered between 20 and 400. We estimate the compositions of fluids inside the crack during both intervals as either water foam (mixtures of water and H2O gas/steam) or misty gas (mixtures of water droplets and H2O gas/steam). We finally suggest that if an eruption were to have taken place following the 2011 unrest, it would have been in phreatic style rather than magmatic style. The results of our study therefore contribute to the effort in the prediction of the behavior of future eruptions, and to volcanic hazards assessment, and therefore to volcanic risk mitigation.

Distinguishing high surf from volcanic long‐period earthquakes

AbstractRepeating long‐period (LP) earthquakes are observed at active volcanoes worldwide and are typically attributed to unsteady pressure fluctuations associated with fluid migration through the volcanic plumbing system. Nonvolcanic sources of LP signals include ice movement and glacial outburst floods, and the waveform characteristics and frequency content of these events often make them difficult to distinguish from volcanic LP events. We analyze seismic and infrasound data from an LP swarm recorded at Pagan volcano on 12–14 October 2013 and compare the results to ocean wave data from a nearby buoy. We demonstrate that although the events show strong similarity to volcanic LP signals, the events are not volcanic but due to intense surf generated by a passing typhoon. Seismo‐acoustic methods allow for rapid distinction of volcanic LP signals from those generated by large surf and other sources, a critical task for volcano monitoring.

Unusual signals recorded by ocean bottom seismometers in the flooded caldera of Deception Island volcano: volcanic gases or biological activity?

AbstractAn ocean bottom seismometer (OBS) network was deployed for 1 month at Deception Island volcano, Antarctica, in early 2005. Although only two volcano-tectonic and three long-period events were observed, the three OBSs located > 2 km apart inside the caldera detected over 3900 events that could not be attributed to known volcanic or hydrothermal sources. These events are found on one instrument at a time and occur in three types. Type 1 events resemble impulsive signals from biological organisms while type 2 and type 3 events resemble long-period seismicity. The largest number of events was observed in a region of volcanic resurgence and hydrothermal venting. All three types occur together suggesting a common cause and they show evidence for a diurnal distribution. The events are most likely to be due to aquatic animals striking the sensors, but a geological source is also possible. In the first case, these signals indicate the presence of a biological community confined to the caldera. In the second case, they imply widespread hydrothermal activity in Port Foster. Future OBS experiments should bury the seismometers, include a hydrophone, deploy instruments side-by-side, or include a video camera to distinguish between biological and geological events.

Long-period long-duration seismic events during hydraulic stimulation of shale and tight-gas reservoirs — Part 2: Location and mechanisms

Long-period long-duration (LPLD) seismic events that have been observed during hydraulic stimulation of shale-gas and tight-gas reservoirs appear to represent slow shear slip on relatively large faults. Within the limitations of the recording geometry, we determine the areas in the reservoirs where the events are located in two case studies in the Barnett shale. In one data set, LPLD events appear to occur in the region where the density of natural fractures as well as the fluid pressure during pumping were highest. In the other data set, the LPLD events are observed to occur between two wells and seem to establish a hydraulic connection between them. In both data sets, the LPLD events occur in areas with very few located microearthquakes. A combination of factors such as high fluid pressure and/or high clay content is potentially responsible for the slowly slipping faults. The LPLD events appear to be occurring only on faults large enough to produce a sequence of slow slip events. We suggest that these slowly slipping faults contribute appreciably to the stimulation of these extremely low-permeability reservoirs and hence mapping the distribution of faults and fractures and areas with rock properties that favor slow, sustained slip, can help in optimizing production.

Time-varying autoregressive model for spectral analysis of microseismic experiments and long-period volcanic events

S U M M A R Y Recent studies show that the frequency content of continuous passive recordings contains useful information for the study of hydraulic fracturing experiments as well as longstanding applications in volcano and global seismology. The short-time Fourier transform (STFT) is usually used to obtain the time–frequency representation of a seismic trace. Yet, this transform has two main disadvantages, namely its fixed time–frequency resolution and spectral leakage. Here, we describe two methods based on autoregressive (AR) models: the short-time autoregressive method (ST-AR) and the Kalman smoother (KS). These two methods allow for the AR coefficients to vary over time in order to follow time-varying frequency contents. The outcome of AR methods depends mainly on the number of AR coefficients. We use a robust approach to estimate the optimum order of the AR methods that best matches the spectral comparison between Fourier and AR spectra. Comparing the outcomes of the three methods on a synthetic signal, a long-period volcanic event, and microseismic data, we show that the STFT and both AR methods are able to track fast changes in frequency content. The STFT provides reasonable results even for noisy data using a simple and effective algorithm. The coefficients of the AR filter are defined at all time in the case of the KS. However, its better time resolution is slightly offset by a lower frequency resolution and its computational complexity. The ST-AR has a high spectral resolution and the lowest sensitivity to background noises, facilitating the identification of the various frequency components. The estimated AR coefficients can also be used to extract parts of the signal. The study of long-term phenomena, such as resonance frequencies, or transient events, such as long-period events, could help to gain further insight on reservoir deformation during hydraulic fracturing experiments as well as global or volcano seismological signals.

Waveform inversion of shallow repetitive long period events at Villarrica Volcano, Chile

[1] Villarrica volcano is presently characterized by an active lava lake in the summit crater, with mild strombolian style eruptions of basaltic-andesitic products. The lava lake acts as a permeable barrier where gas-rich magma releases volatiles and then descends through the process of conduit convection to a deep storage reservoir. We identified thousands of repetitive long-period (LP) seismic events within a three-year time period using a matched filter; each was discovered to be associated with an infrasonic emission, and hence outgassing from the lava lake. Using this catalog of events for timing, we stacked signals over the three-year deployment from 2010-2012 to produce representative seismic and infrasonic signatures, and performed a full-waveform seismic inversion using nine stations from the network. The data are well represented by a single horizontal force source-time function, oriented N75°E and dipping 7°W from horizontal, with a maximum amplitude of 34.4 MN. The source is stable in time, and can be explained by a force couple system produced by a bursting bubble in the lava lake and the responsive drag force of the lava sloshing to fill in the void produced by the gas emission in an asymmetrical crater. Path effects lengthen the coda of the events with increasing distance from the source, resulting in emergent tremor-like signals at distant seismometers. Since the occurrence of LP events correlates with tremor intensity, we suggest much of the tremor results from the superposition of discrete LPs.

Source process of long‐period seismic events at Taal volcano, Philippines: Vapor transportation and condensation in a shallow hydrothermal fissure

We analyzed observations of a swarm of more than 40,000 long‐period (LP) seismic events at Taal volcano, Philippines, in 2010–2011. The event waveforms are strongly correlated to each other, consistent with a fixed source location, and begin with a dilatational first motion. They have a peak frequency around 0.8 Hz and a quality factor Q of 6. Waveform inversion of the events pointed to a tensile crack source dipping 30°–60° at a shallow (100–200 m) depth. A simulation using a fluid‐filled crack model indicated that the complex frequencies of the waveforms are explained by the fundamental longitudinal mode resonance of a vapor‐filled crack 188 m long. A satellite thermal infrared image acquired during the swarm period suggests that the LP events were not accompanied by surface gas releases. We considered a vapor transportation model in which vapor exsolved from magma and rose in a fissure extending to the LP source. This model yielded estimates that 105–107 m3 of magma was involved in the LP swarm and that the temperature of vapor in the LP source crack was around 600 K. We modeled a triggering mechanism of the crack resonance based on sudden condensation of vapor at the crack tip in a cold aquifer. This model explained observed characteristics of the events including the dilatational first motion, the total volumetric change, and the fixed source location.

Moment tensor inversion for the source location and mechanism of long period (LP) seismic events from 2009 at Turrialba volcano, Costa Rica

Long-period (LP) seismic events were recorded during the temporary installation of a broadband seismic network of 13 stations from March to September 2009 on Turrialba volcano, Costa Rica. Over 6000 LPs were extracted using a modified STA/LTA method and a family consisting of 435 similar LP events has been identified. For the first time at Turrialba volcano, full-waveform moment tensor inversion is performed to jointly determine the location and source mechanism of the events. The LPs in the family are likely to be caused by crack mechanisms dipping towards the southwest at angles of approximately 10 to 20°, located at shallow depths (<800m) below the active Southwest and Central craters. As the locations are so shallow, the most probable causes of crack mechanisms are hydrothermal fluids resonating within or “pulsing” through a crack. The waveforms observed at the summit stations suggest a “pulsing” mechanism, but source resonance with a high degree of damping is also possible.

Magnitude scale for LP events: a quantification scheme for volcanic quakes

SUMMARY The peculiar source characteristics of long-period seismic events (time persistency of the source, low-frequency peaks in the source spectrum, absence of high-frequency radiation) prevent the formation of a definite high-frequency coda in the seismograms. In contrast, this is well formed in volcano‐tectonic quakes. For this reason, the widely used duration magnitude scale that is based on the proportionality between the energy and the coda duration cannot be used for long-period estimation. In observatory practice, the long-period magnitude is sometimes estimated using the same duration magnitude scale, leading to confusing results. In this report, we show a new method to estimate the magnitude of long-period events that generally occur for volcanoes, with some application examples from data for Mt Etna (Italy), Colima Volcano (Mexico) and Campi Flegrei (Italy).

Investigating the source characteristics of long-period (LP) seismic events recorded on Piton de la Fournaise volcano, La Réunion

Magmatic and hydrothermal processes play a significant role in generating seismicity at active volcanoes. These signals can be recorded at the surface and can be used to obtain an insight into the volcano's internal dynamics. Long period (LP) events are of particular interest as they often accompany or precede volcanic eruptions, but they are still not well understood. Piton de la Fournaise volcano, La Réunion Island, is one of the most active volcanoes in the world however LP events are rarely recorded there. A seismic network of 20 broadband seismometers has been operational on Piton de la Fournaise volcano since November 2009. Between November 2009 and January 2011 the volcano erupted five times, but only 15 LP events were recorded. Three of these eruptions were preceded by LP events, and several LP events were recorded during an intrusive phase. A family of three repeating LP events exists within the dataset. In order to characterize these events we locate and perform moment tensor inversion on the LP family. The LP events are located within the summit crater at shallow depths (<200m below the surface). Inversions show that the source mechanism is best represented by a tensile crack with horizontal crack geometry. We also investigate the relationship between LP occurrence and eruptive characteristics (size of the eruption, deformation of the edifice, etc.), and we find that the events exist only during flank eruptions and can be generated by the activity of the hydrothermal system and/or by the deformation inside the crater.

Signs of magma ascent in LP and VLP seismic events and link to degassing: An example from the 2010 explosive eruption at Merapi volcano, Indonesia

The link between seismicity and degassing is investigated during the VEI 4 eruptions of Merapi volcano (Indonesia) in October and in early November 2010. Seismicity comprised a large number and variety of earthquakes including Volcano-Tectonic events, a sustained period of Long Period Seismicity (LPS), i.e., Long-Period events (LP), Very Long Period events (VLP) and tremor. LPS seismicity is ascribed to the excitation of fluid-filled cavity resonance and inertial displacement of fluids and magma. We investigate here LPS that occurred between 17 October and 4 November 2010 to obtain insights into the volcano eruption processes which preceded the paroxysmal phase of the eruption on 4–5 November. We proceed to the moment tensor inversion of a well-recorded large VLP event during the intrusion phase on 17 October 2010, i.e., before the first explosion on 26 October. By using two simplified models (crack and pipe), we find a shallow source for this VLP event at about 1km to the south of the summit and less than 1km below the surface. We analyze more than 90 LP events that occurred during the multi-phase eruption (29 October–4 November). We show that most of them have a dominant frequency in the range 0.2–4Hz. We could locate 48 of those LP events; at least 3 clusters of LP events occurred successively. We interpret these observations as generated by different fluid-filled containers in the summit area that were excited while magma rose. We also observe significant variations of the complex frequency during the course of the eruption. We discuss these changes in terms of a variable ratio of fluid to solid densities and/or by possible conduit geometry change and/or permeability of the conduit or the edifice and/or by resonance of different fluid-containers during the release of more than 0.4Tg of SO2 and large but unknown masses of other volcanic gases. Finally, we also discuss how the major explosions of the eruption were possibly triggered by passing waves resulting from regional tectonic earthquakes on 3 and 4 November.

Source amplitudes of volcano-seismic signals determined by the amplitude source location method as a quantitative measure of event size

The amplitude source location (ASL) method, which uses high-frequency amplitudes under the assumption of isotropic S-wave radiation, has been shown to be useful for locating the sources of various types of volcano-seismic signals. We tested the ASL method by using synthetic seismograms and examined the source amplitudes determined by this method for various types of volcano-seismic signals observed at different volcanoes. Our synthetic tests indicated that, although ASL results are not strongly influenced by velocity structure and noise, they do depend on site amplification factors at individual stations. We first applied the ASL method to volcano-tectonic (VT) earthquakes at Taal volcano, Philippines. Our ASL results for the largest VT earthquake showed that a frequency range of 7–12Hz and a Q value of 50 were appropriate for the source location determination. Using these values, we systematically estimated source locations and amplitudes of VT earthquakes at Taal. We next applied the ASL method to long-period events at Cotopaxi volcano and to explosions at Tungurahua volcano in Ecuador. We proposed a practical approach to minimize the effects of site amplifications among different volcano seismic networks, and compared the source amplitudes of these various volcano-seismic events with their seismic magnitudes. We found a proportional relation between seismic magnitude and the logarithm of the source amplitude. The ASL method can be used to determine source locations of small events for which onset measurements are difficult, and thus can estimate the sizes of events over a wider range of sizes compared with conventional hypocenter determination approaches. Previously, there has been no parameter widely used to quantify the sources of volcano-seismic signals. This study showed that the source amplitude determined by the ASL method may be a useful quantitative measure of volcano-seismic event size.