Long-term observation of volcanic tremor on Stromboli volcano (Italy): A synopsis

The features of seismic activity on Stromboli are discussed and compared in terms of their relationship with the main changes of volcanic activity from 1990 to 1993.We considered a statistical approach for our data analysis. Cluster analysis was used to seek out classes of spectra which might characterize the condition of the volcanic system. The classes we have found provide insights into a scenario which evolves through different phases of volcanic activity, from paroxysms to low activity. We show that episodes of lava effusion and lava fountaining are heralded by variations in the spectral features of tremor after a preparation time. This result highlights the importance of tremor, and reveals that long-term observations are key to examine slow modifications in a volcanic system such as Stromboli, characterized by open conduits, and persistent explosive activity.     

Microtremors and volcanic explosions at Stromboli — Part 2

A seismic survey was carried out at Stromboli volcano during August 1973. Statistical and spectral analyses of volcanic tremor and explosion quakes were performed. The statistical analysis has shown that the value of them coefficient in Ishimoto & Iida’s relation is high and that the hourly frequencies of events are fairly constant. The spectral analysis has shown a similarity in shape between tremors and explosion quakes. These events have the dominant frequency of 5 Hz at the craters area.     

Volcanic tremor at Mt. Etna: Inferences on magma dynamics during effusive and explosive activity

During 1999, the volcanic activity at Mt. Etna was both explosive and effusive at the summit craters: Strombolian activity, lava fountains and lava flows affected different areas of the volcano, involving three of the four summit craters. Results from analysis of the 1999 volcanic tremor features are shown at two different time scales. First, the long-term time variation of the features of the volcanic tremor (including spectral and polarization parameters), during the entire year, was compared with the evolution of the eruptive activity. This approach demonstrated the good agreement between tremor data and observed eruptive activity; the activation of different tremor sources was suggested. Then, a more refined analysis of the volcanic tremor, recorded during 14 lava fountain eruptions, was performed. In particular, a shift of the dominant frequencies towards lower values was noted which corresponds with increasing explosive activity. Similar behaviour in the frequency content has already been observed in other explosive eruptions at Mt. Etna as well as on other volcanoes. This behaviour has been explained in terms of either an increase in the tremor source dimension or a decrease in the sound speed in the magma within the conduit. These results confirm that the volcanic tremor is a powerful tool for better understanding the physical processes controlling explosive eruptions at Mt. Etna volcano.     

On the spectral peaks of volcanic tremor at Stromboli

Seismic activity at Stromboli Volcano is characterized by a variety of signals, emanating from three vents. For a long time, the northwest vent has been in constant activity. Periodically, large explosions occur and material is ejected beyond the crater walls. These large explosions are accompanied by sonic and infrasonic pressure waves in the atmosphere, and explosion quakes. Apart from large explosions, there is constant activity in the form of continuous gas bursts which are related to low infrasonic pulses in the atmosphere and volcanic tremor. We assume that volcanic tremor and low pressure infrasonics are generated by gas bubbles inside the volcanic conduit, and accordingly, we compute synthetic tremor by modeling the source function as a pressure variation in a spherical cavity that propagates through a finely layered medium, by means of Haskell's formalism. To simulate a tremor, we superpose in time domain a large number of such pulses of varying amplitudes and time delays, according to the observed infrasonic series. In addition to the spectral similarity, the observed and synthetic tremor display the same autocorrelation and Hurst exponents, implying similar long-term correlation. We present strong evidence in favour of an interpretation of the spectral peaks of the volcanic tremor at Stromboli in terms of resonances of the layered structure, hence, as a path effect rather than a source effect.     

Volcanic tremor at Pavlof Volcano,Alaska, October 1973–April 1986

In thirteen years (1973–1986) of seismic monitoring of Pavlof Volcano, 488 episodes of volcanic tremor have been recorded, only 26 of which have been previously described in the literature. This paper tabulates and describes all the tremor episodes and reports on the results of all analyses to date. Pavlof tremor durations range from 2 minutes to greater than 1 week; episodes accompanying magmatic eruptions have durations greater than 1 hour, and sustained amplitudes of greater than 6 mmP-P (=54 nanometers at 1.5 Hz) on station PVV, 8.5 km from the vent. Digital data provide much better amplitude resolution than helicorders do. Helicorders, however, provide continuous coverage, whereas digital data are intermittent. Correlations of tremor with visual eruption observations shows that tremor amplitudes are roughly correlated with heights of lava fountains, but the correlation of tremor amplitudes with plume heights is more problematic. Fast Fourier Transform (FFT) spectra show that Pavlof tremor is quite statinary for the entire time period, 1973–1983. All principal spectral peaks lie between 0.8 and 3.0 Hz, and may be caused by resonance of magma and gas, and resonance of the volcanic pile. Preliminary analysis of 2-and 3-component data shows thatP, S, PL, and Rayleigh waves may be present in Pavlof volcanic tremor. Other waveforms can be misidentified as tremor, most commonly those caused by storms orS-waves of regional earthquakes. A strategy is proposed to distinguish tremor from noise using automatic seismic data acquisition and analysis systems. Pavlof's volcanic tremor is briefly compared with a preliminary sample of over 1100 cases of tremor from 84 volcanoes worldwide. Finally, several recommendations for monitoring and reporting volcanic tremor are discussed.     

Analysis of sustained long-period activity at Etna Volcano, Italy

Following the installation of a broadband network on Mt. Etna, sustained Long-Period (LP) activity was recorded accompanying a period of total quiescence and the subsequent onset of the 2004–2005 effusive episode. From about 56000 events detected by an automatic classification procedure, we analyse a subset of about 3000 signals spanning the December 17th, 2003–September 25th, 2004, time interval. LP spectra are characterised by several, unevenly-spaced narrow peaks spanning the 0.5–10 Hz frequency band. These peaks are common to all the recording sites of the network, and different from those associated with tremor signals. Throughout the analysed time interval, LP spectra and waveforms maintain significant similarity, thus indicating the involvement of a non-destructive source process that we interpret in terms of the resonance of a fluid-filled buried cavity. Polarisation analysis indicates radiation from a non-isotropic source involving large amounts of shear. Concurrently with LP signals, recordings from the summit station also depict Very-Long-Period (VLP) pulses whose rectilinear motion points to a region located beneath the summit craters at depths ranging between 800 and 1100 m beneath the surface. Based on a refined repicking of similar waveforms, we obtain robust locations for a selected subset of the most energetic LP events from probabilistic inversion of travel-times calculated for a 3D heterogenous structure. LP sources cluster in a narrow volume located beneath the summit craters, and extending to a maximum depth of ≈ 800 m beneath the surface. No causal relationships are observed between LP, VLP and tremor activities and the onset of the 2004–2005 lava effusions, thus indicating that magmatic overpressure played a limited role in triggering this eruption. These data represent the very first observation of LP and VLP activity at Etna during non-eruptive periods, and open the way to the quantitative modelling of the geometry and dynamics of the shallow plumbing system.     

Seismic activity accompanying the 1974 eruption of Mt. Etna

On January 30, 1974, an explosive eruption began on the western side of Etna. The activity evolved into two eruptive periods (January 30–February 17 and March 11–29). Two spatter cones (Mount De Fiore I and Mount De Fiore II) were formed at a height of about 1650 m a.s.l. and a distance of 6 km from the summit area. The effusive activity was very irregular with viscous lava flows of modest length.A seismic network of four stations was established around the upper part of the volcano on February 3. Moreover additional mobile stations were set up at several different sites in order to obtain more detailed informations on epicenter locations and spectral content of volcanic tremor.The volcanic activity is discussed in relation to the distribution of epicenters and the time-space distribution of the spectral characteristics of volcanic earthquakes and tremor. The characteristics of the seismic activity suggest that the flank eruption of Mount Etna was probably feed by a lateral branch of the main conduit yielding the activity at the Central Crater.     

Spectral features of long-period volcanic signals: Klyuchevskoi Volcano

This study is concerned with spectral features in the long-period volcanic signals observed on Klyuchevskoi Volcano during the period 2000–2007. The results were obtained from digital records made at seven seismic stations communicating via radio telemetry; they were deployed on the slope of and close to Klyuchevskoi. It is shown that the spectral content of long-period volcanic signals changed toward lower frequencies during the period from mid 2003 to the end of 2007. The change in the spectra is here related to the volcano’s activity, which increased at the end of 2003 and was accompanied by numerous explosions and the outflow of lava from the crater.     

Volcanic Tremor at Mt. Etna,Italy, Preceding and Accompanying the Eruption of July – August, 2001

The July 17 – August 9, 2001 flank eruption of Mt. Etna was preceded and accompanied by remarkable changes in volcanic tremor. Based on the records of stations belonging to the permanent seismic network deployed on the volcano, we analyze amplitude and frequency content of the seismic signal. We find considerable changes in the volcanic tremor which mark the transition to different styles of eruptive activity, e.g., lava fountains, phreatomagmatic activity, Strombolian explosions. In particular, the frequency content of the signal decreases from 5 Hz to 3 Hz at our reference station ETF during episodes of lava fountains, and further decreases at about 2 Hz throughout phases of intense lava emission. The frequency content and the ratios of the signal amplitude allow us to distinguish three seismic sources, i.e., the peripheral dike which fed the eruption, the reservoir which fed the lava fountains, and the central conduit. Based on the analysis of the amplitude decay of the signal, we highlight the migration of the dike from a depth of ca. 5 km to about 1 km between July 10 and 12. After the onset of the effusive phase, the distribution of the amplitude decay at our stations can be interpreted as the overall result of sources located within the first half kilometer from the surface. Although on a qualitative basis, our findings shed some light on the complex feeding system of Mt. Etna, and integrate other volcanological and geophysical studies which tackle the problem of magma replenishment for the July–August, 2001 flank eruption. We conclude that volcanic tremor is fundamental in monitoring Mt. Etna, not only as a marker of the different sources which act within the volcano edifice, but also of the diverse styles of eruptive activity. An erratum to this article is available at .     

Analysis of seismic patterns observed at Nevado del Ruiz volcano, Colombia during August – September 1985

The seismic activity of the Nevado del Ruiz volcano was monitored during August–September 1985 using a three-component portable seismograph station placed on the upper part of the volcano. The objective was to investigate the frequency content of the seismic signals and the possible sources of the volcanic tremor. The seismicity showed a wide spectrum of signals, especially at the beginning of September. Some relevant patterns from the collected records, which have been analyzed by spectrum analysis, are presented. For the purpose of analysis, the records have been divided into several categories such as long-period events, tremor, cyclic tremor episodes, and strong seismic activity on September 8, 1985.The origin of the seismic signals must be considered in relation to the dynamical and acoustical properties of fluids and the shape and dimensions of the volcano's conduits.The main results of the present experiment and analysis show that the sources of the seismic signals are within the volcanic edifice. The signal characteristics indicate that the sources lie in fluid-phase interactions rather than in brittle fracturing of solid components.     

Regional stress field in the area of Stromboli (Italy): insights into structural data and crustal tectonic earthquakes

Crustal tectonic seismicity on Stromboli is here discussed along with the regional tectonic and geodynamic context of the Aeolian archipelago, Italy. The aim of this paper is to contribute to the reduction of the crucial gap in information concerning the tectonic activity originating at the volcano Stromboli in the last decade. It is indeed well known that Strombolian eruptions, related to the persistent state of activity on this volcano, allow a fast collection of seismic data which are mainly represented by explosion quakes and tremor. However, only some sporadic information concerns tectonic earthquakes due to their low occurrence rate. A catalog of 150 crustal tectonic earthquakes recorded at Stromboli from 1985 to 1996 is presented here. The analyses of the seismograms were carried out on the basis of the records of the station STR, located in the southwestern part of the island. The magnitude considered ranged from 1.5 to 3.7. Hypocentral location indicates foci at less than 12 km, although for the majority of the earthquakes, which are not recorded at the seismic stations deployed on the other islands of the Aeolian archipelago, a shallower (within a few kilometers) depth can be hypothesized. The strain release associated with the 150 tectonic earthquakes of our catalog was also evaluated. The low levels of occurrence and seismic energy release reached throughout the ca. twelve years analyzed here confirm the minor role played by processes of brittle failure of rocks. This aspect of seismic activity is nevertheless important as it can provide a contribution to the knowledge of the dynamics of this volcano. The focal mechanisms available, integrated with data from the literature, show the T-axes oriented about NE–SW and NW–SE. These findings are consistent with the deformative and tectonic characteristics of this volcanic edifice and of the Aeolian archipelago, which shows the same two directions of extension. This is in agreement with subduction, which is very old or which passively evolves towards the final stages, and is controlled by the transform activity of the Aeolian–Tindari–Giardini fault system.     

Multiple ruptures for Long Valley microearthquakes: a link to volcanic tremor?

Despite several episodes of ground deformation and intense seismic activity starting in 1978, the Long Valley, California, volcanic area has not produced clearly recognized volcanic tremor. Instead, a variety of atypical microearthquakes have been recorded during these episodes, including events dominated by low-frequency (long-period) or mixed high and low-frequency (hybrid) signals. During a 1997 episode, a number of unusual microearthquakes occurred within a temporary 40-station seismic network surrounding the Casa Diablo area, allowing the events to be precisely located and analyzed as a function of azimuth, offset, and source characteristics. Eight prime examples lie within two, 7 km-deep clusters of seismicity separated by about 1 km, with four events in each cluster. Empirical Green's function deconvolution shows that these events are composed of two to three sub-events, the sub-events consisting of ordinary (single rupture, double-couple) microearthquakes. The delay times between the sub-events are constant within each cluster, equaling 0.092 s in one and 0.06 s in the other. Events from other clusters show similar delays. The signal interference produced by the closely spaced sub-events gives rise to modulated, delay-dependent source spectra. The regularity of the delays suggests that the sub-events are being triggered by a fixed length and/or time scale process, an example being the length/inflation rate of a magmatic or hydrothermal flow structure. With continued action of the triggering process, the sub-events could proliferate and evolve into observable volcanic tremors at Mammoth.     

Stromboli and its 1975 eruption

On November 4, 1975 in the evening, an eruption took place at Mt. Stromboli. On the following day lava flowed on the Sciara del Fuoco downward to the sea, accompanied by an intense explosive activity at the crater plane. Direct observations on the volcanic activity were carried out since November 6 while a seismic survey was made from Nov. 7 to 12. The total volume of the lava outpoured during this period of activity that lasted 21 days, was estimated to be about 10⁴ m³. This paper reports the results of direct observations, and of the petrological, radioactive disequilibria and seismic activity studies performed for this eruption. The eruption was preceded by an insignificant change of seismic activity, which was monitored by a seismic station located about 2 km East of the crater. A shallow seismicity was strietly related to crater explosions accompanying the eruptive phenomenon. Radioactive disequilibria showed a lack of disequilibrium between²²⁸Ra and²³²Th explainable in terms of a fast rising of magma in the conduit. Chemical analyses of lava samples and deep seismic sounding data indicate a correspondence between the depth (10–15 km) at which crystallization pressure of phenocrysts occurs and a low velocity laver.     

Volcanic tremor at Mt. Etna: State-of-the-art and perspectives

Volcanic tremor on Etna seems to have its origin within the main magma feeding system. On the basis of both spectral analyses at two permanent seismic stations and periodical measurements along the slopes of the volcano, two distinct sources are proposed. The former, characterized by low frequency contents (f<1.5 Hz), is located in a 2 km deep flat magma chamber, whereas the latter source seems to be linked to the upper part of the active vents.Turbulent motions in the magma-gas mixture, induced by escaping gases within the conduits, is one proposed cause of volcanic tremor on Etna (Seidl et al., 1981).From spectral analyses we propose approximate models of the feeding system of the main summit craters.Time variations of tremor energy were also investigated, and no regular patterns have been observed for the studied eruptions.More systematic information seems to be needed for a better knowledge of both the source model and location, and correlation between tremor features and volcanic activity.     

Evidence for gas influence on volcanic seismic signals recorded at Stromboli

The meaning of the large variety of seismic waveforms showing different frequencies associated with volcanic activity is still uncertain. No definitive model for source dynamics has been proposed yet. At present, seismic models explaining the origin of the spectral content of the seismic signal are mainly linked to geometrical features of the volcanic conduit and to resonant effects induced by pressure fluctuations in a fluid-filled conduit. Such models assume the physical system to be in a steady state over a long interval of time. At Stromboli, the seismicity produced by each of the three active vents is not stable and can suddenly change in time. Therefore, the application of stable resonator models appears to be unwarranted.On the basis of infrared image analyses, atmospheric pressure, and seismic wavefield signals, we show that different frequency contents and different explosive styles occur at Stromboli at the same time at different vents. Moreover, we give evidence to support the idea that seismicity and explosivity are both controlled by a variable gas flow regime released during explosions at the top of the magma column. We have recognized two main pressure regimes, controlling the explositivity at Stromboli: one is impulsive, short and possibly produced by a gas accumulation beneath a cap rock; the other has an intermittent regime feature, and lasts longer. We demonstrate how the first one generates a monochromatic low-frequency wavefield, while the second shows a high-frequency spectrum where the frequency content depends on pressure fluctuations and on pressure gradient. We suggest different pressure growth and gas flow in the magma as the common source for both explosive style and seismic wavefield.     

Echo-resonance and hydraulic perturbations in magma cavities: application to the volcanic tremor of Etna (Italy) in relation to its eruptive activity

A study is presented of spectral features of volcanic tremor recorded at Mount Etna (Sicily, Italy) following the methods of analysis suggested by the resonant scattering formalism of Gaunaurd and Überall (1978, 1979a, 1979b) and the model for hydraulic origin of Seidl et al. (1981). The periods investigated include summit and flank eruptions that occurred between 1984 and 1993. Recordings from a permanent station located near the top of the volcano were used, and the temporal patterns associated with (a) the average spacing ( ) between consecutive spectral peaks in the frequency range 1–6 Hz, (b) the spectral shape and (c) the overall spectral amplitude were analyzed. values are thought to depend on the physical properties of magma, such as its density, which, in turn, is controlled by the degree of gas exsolution. Variations in the spectral shape are tentatively attributed to changes in the geometrical scattering from the boundary of resonant conduits and magma batches. Finally, the overall amplitude at the station should essentially reflect the state of turbulence of magma within the superficial ascending path. A limit in the application of the resonant scattering formalism to the study of volcanic tremor is given by the fact that the fundamental modes and integer harmonics are difficult to identify in the frequency spectra, as tremor sources are likely within cavities of very complex geometry, rather than in spherical or cylindrical chambers, as expected by theory. This study gives evidence of some correlations between the analyzed temporal patterns and the major events in the volcanic activity, related to both lava flow and explosions at the summit vents. In particular, relatively high values of have been attained during the SE crater eruption of 1984, the complex eruptive phases of September–October 1989 and the 1991–1993 flank eruption, suggesting the presence of a relatively dense magma for all of these events. Conversely, very low values have been recorded in coincidence with the December 1985 activity and the paroxysmal explosions at the summit craters of early 1990, which are interpreted here as fed by fluid-vesiculated magma. Appreciable modifications in the spectral shape have been observed in relation to changes of the volcanic activity that probably preceded the opening and disactivation of shallow dykes or magma batches. Finally, the overall amplitude seems to be a sensitive indicator of the state of gas turbulence within the shallow conduits, as is suggested by the high values attained during phases of intense volcanic activity.     

Statistical analysis of intermittent volcanic tremor associated with the September 1989 summit explosive eruptions at Mount Etna, Sicily

The pattern of volcanic tremor accompanying the 1989 September eruption at the south-east summit crater of Mount Etna is studied. In specific, sixteen episodes of lava fountaining, which occurred in the first phase of the eruption, are analysed. Their periodic behaviour, also evidenced by autocorrelation, allows us to define the related tremor amplitude increases as intermittent volcanic tremor episodes. Focusing on the regular intermittent behaviour found for both lava fountains and intermittent volcanic tremors, we tried an a posteriori forecast using simple statistical methods based on linear regression and the Student’ t-test. We performed the retrospective statistical forecast, and found that several eruptions would have been successfully forecast. In order to focus on the source mechanism of tremor linked to lava fountains, we investigated the relationship between volcanic and seismic parameters. A mechanism based on a shallow magma batch ‘regularly’ refilled from depth is suggested.     

Explosion quakes at Stromboli (Italy)

This paper reports the results of two seismic experiments aimed at determining the wave field of explosion quakes at Stromboli Island (Mediterranean Sea, Southern Italy). The typical Strombolian activity mostly consists of explosive phenomena causing pyroclastic, materials to be emitted together with jets of volcanic gases from one or more craters. Stromboli is an active volcano characterized by persistent seismic activity consisting of explosion quakes that are seismic events associated with the explosive volcanic phenomena. Explosion quakes are short lived seismic events occurring intermittently whose amplitude tends to decrease with distance from the vent. A distinctive feature of explosion quakes is the presence on seismograms of two, often clearly distinct, seismic phases. The first, low-frequency seismic phase (<2 Hz) is in fact usually followed by a high-frequency seismic phase (>3–4 Hz) after one second or more. The first seismic phase of explosion quakes has been shown to be characterized by a nearly radial linear polarization and by an apparent propagation velocity estimated at 600–800 m/s. The second phase is characterized by a more chaotic motion and a lower apparent propagation velocity of 150–450 m/s. The wavefield associated with the first low-frequency seismic phase appears to be generated by a resonating P-wave seismic source accompanying gas explosion and emission of pyroclastic materials. The wavefield associated with the second high-frequency seismic phase of explosion quakes appears to be mainly composed of scattered and converted waves due to the critical topography of the volcano.     

Automatic Recognition of Landslides Based on Neural Network Analysis of Seismic Signals: An Application to the Monitoring of Stromboli Volcano (Southern Italy)

In the last 9 years, the amount and the quality of geophysical and volcanological observations of Stromboli's' activity have undergone a marked increase. This new information highlighted that the landslides on the Sciara del Fuoco flank are tightly linked to the volcanic activity. Actually, at the beginning of the December 28, 2002, effusive eruption, the seismic monitoring network was less dense than now, and therefore it is not known if there was an increase in the landslide rate before the eruption. Despite this, it is known that a big landslide occurred 2 days after the beginning of the eruption which caused a tsunami (December 30, 2002). More recently, the effusive eruption in February 2007 was preceded by an increase in landslides on the Sciara del Fuoco flank, which were recorded by the seismological monitoring system that had been improved after the 2002–2003 crisis. These episodes led us to believe that monitoring the Sciara del Fuoco flank instability is an important topic, and that landslides might be significant short-term precursors of effusive eruptions at the Stromboli volcano. To automatically detect landslide signals, we have developed a specialized neural algorithm. This can distinguish between landslides and the other types of seismic signals usually recorded at the Stromboli volcano (i.e., explosion quakes and volcanic tremor). The discrimination results show an average performance of 98.67 %. According to the experience of the crisis of 2007, to identify changes that can be considered as precursors of effusive eruptions, we set up an automatic decision-making method based on the neural network responses. This method can operate on a continuous data stream. It calculates a landslide percentage index (LPI) that depends on the number of records that are classified by the net as landslides over a given time interval. We tested the method on February 27, 2007, including the beginning of the effusive phase. The index showed an increase as early as at 09:00 UTC on that day and reached its maximum value (100 %) at 12:00, about 40 min before the onset of the eruption. After the beginning of the effusive phase, the index remains high due to the blocks that roll down along the slope from the front of the lava flow. On the basis of these tests, we propose a decision-making method that is able to recognize a trend in the LPI similar to that of 2007 eruption, allowing the identification of precursors of effusive phases at the Stromboli volcano.     

Microtremors and volcanic explosions at Stromboli (Aeolian Islands)

Seismic observations were carried out at Stromboli from October 16^th to 24^th, 1972. One three-component seismic station was set up at Semaforo Nuovo, about 2 km on the E of the crater plain. Seismic observations showed the occurrence of microtremors, with variable amplitude but constant frequency content, and different types of explosion earthquakes. Spectral and vibration orbit analyses of microtremors and quakes were made in order to infer their nature and the depth of origin. Changes in microtremor amplitudes corresponding to different phases of volcanic activity indicate that their study may be useful in forecasting strongly explosive volcanic phases at Stromboli.