Sunday 17 September - Session 1

A catalogue of earthquakes which have produced macroseismic effects in the Etnean area during 1832-1998, is presented. The need of a catalogue for this region derives from the observation that the traditional parametric catalogues do not characterise the seismicity in detail, for instance in terms of seismic sequences or seismotectonic analyses. The quantity and the quality of available information allow us to investigate these aspects. Many earthquakes have been associated with seismogenic structures indeed, so that it is possible to investigate fault behaviour in time. The catalogue has been compiled by the analysis of information reported in contemporaneous sources (scientific papers, bulletins, newspapers and archive documents) and the intensities assessed using the EMS-98 macroseismic scale. A total number of 1735 events is listed in the catalogue.

Focal parameters such as the epicentral location, both maximum and epicentral intensity, macroseismic magnitude and, when possible, seismogenic fault and coseismic surface faulting, are defined for each event. Available instrumental magnitude and depth values are also reported. Tests on determination of the epicentral coordinates and some examples of possible use and application of the catalogue dataset, are shown.

Between August 1995 and February 1996 several lava fountains episodes occurred at the North-East Crater of Mount Etna volcano. The related seismic data were simultaneously recorded up to 15 stations. Aim of this work is the analysis of records associated to the November 27, 1995 eruption. We have computed the spectra from these records and identified spectral peaks relating to the source. The Chouet's fluid-filled-crack model has been tested, by using an algorithm requiring the assumption of some physical parameters, such as the sound velocities a in the fluid and a in the rock the densities r_m of the fluid and r_s of the rock, the bulk modulus b of the fluid and the rigidity m of the rock.

Assuming the values from literature, i.e. a=3.4 km/s, b/m=0.1, a/a=1.7 and r_m/r_s=1.04 for a purely liquid fluid and a/a=2.3 and r_m/r_s=0.56 for a foam fluid, we have obtained numerous families of source models, with lengths ranging from 400 to 3000 m for a fully liquid system, and from 500 m to 2200 m for a foam system.

In a second step, the thermodynamic simulation has evidenced the strong dependence of these parameters with pressure P. In detail,
i) for pressure P in the range 0.3-0.7 kbar, it was obtained a purely liquid system with r_m=2.5 g/cm³ and a=2 km/s;
ii) for P ranging 0.1-0.25 kbar, a foam was obtained, with r_m in the interval 0.9 g/cm³ to 2.26 g/cm³, with a ranging from 1.59 km/s to 1.87 km/s;
iii) for P less than 0.1 kbar, an aerosol was obtained with r_m=0.41 g/cm³ and a=1.4 km/s.

Assuming these values and two different sound velocities a in the solid rock (3.4 km/s and 1.8 km/s), we have obtained about twenty source models well fitting all the attributed spectral peaks. The possible source lengths range from 500 to 2500 m, widths from 100 to 900 m, thickness ranging from 0.1 to 5. The lack of more constrains appears as the major problem for the application of the Chouet's model devoted to a very precise knowledge of the internal plumbing system of Mt. Etna volcano.

During the first six months of the year 2000, Bocca Nuova ejected hundreds of steam rings per day. Observations of the ejections have shown that the vent ejecting them was almost perfectly circular in shape and with a diameter of about 5 meters. Measures of the size of the steam rings were possible thanks to the observation of the projected shadow, and resulted close to 200 m. The radius of the core (i.e. of the thoroidal ring) ranged between 1 and 20 m. Videoclips of the rings showed that the core was fastly rotating along its thoroidal axis at rates between 0.5 and 3 turns per sec. We apply the Norbury-Fraenkel model of vortex rings (developed to fit laboratory experiment results on cm sized smoke rings) to the observed rings.

This model allows us to obtain the translation velocity, the circulation, the impulse and the total kinetic energy of fluid rings. From the circulation we get the turns per sec, in agreement with the observed one. So we can conclude that this model can well be extended to sizes many orders of magnitude larger than those for which it was developed. The piston-like mechanism of ring& generation, on the contrary, cannot operate on Etna.

We propose that the small Bocca Nuova vent ejects drops of steam that develop into ring via the Norbury-Fraenkel model. However, the thermal energy of the steam drop is much larger than the total kinetic energy computed by means of the model. We use this surplus of energy to inflate the small steam drop to the observed huge rings via an adiabatic expansion. In this way, we get a perfect agreement with the observed expansion rate and times of drop ejection.

The South East crater of Mt. Etna has shown since late summer of 1999 a persistent activity with periodic paroxysmal episodes characterised by strombolian activity and lava fountaining. Each eruptive episode has been marked by a significant increase in the tremor energy, which, as usually, was observed both in the time and frequency domains.

Present study aims to show the spectral features of volcanic tremor recorded at the stations of TDF and SLN, located at the top and at intermediate elevation on the volcano, respectively.

Each eruptive event was characterised by a gradual increase of the tremor amplitude at its beginning and a more rapid decrement of amplitude in its final stage. The maximum amplitude was usually greater than ten times the usual tremor amplitude. The average duration of each eruptive episode was ranging between 45 and 90 minutes.

Spectral features of tremors, sampled both during paroxysmal and quite periods, do not show significant differences except a pronounced increment of the overall spectral amplitude during the eruptive episodes.

The recurrence of such short eruptions was investigated, using a statistical approach, in order to look for a possible periodicity in the occurrence rate. Preliminary findings encourage pursuing in the used methodological approach, in order to achieve the recognition of significant precursors of Etnean eruptive activity.

Sunday 17 September - Session 2

The assessment of the risk connected to volcanic eruptions is a critical aspect when evaluating the safety of populated areas, either permanently or only temporarily, as it is the case for touristically attractive volcanoes. A stochastic approach has been developed for the analysis and simulation of data sampled at active volcanoes. This approach allows the detection of time correlation in the series, the statistical forecasts of volcanic events using Cox simulations and finally a volcanic tremor decomposition that can help identifying potential precursors preceding stronger eruptions.

The described stochastic approach can be applied at very different time scales, therefore it may result useful for volcanoes as different as dormant ones and constantly active ones.

In this presentation we concentrate mostly on the application of the method to data monitored at Stromboli volcano. Significant time correlation has been detected which leads one to describe Stromboli as a volcano with an exceptional memory of its recent past activity. Forecasting of the number of "normal" events for the next few days has been performed by Monte Carlo simulations. Finally, kriging of the time series derived from volcanic tremor intensity has enabled the extraction of time components, which could furnish additional monitoring variables for the forecast of paroxysmal phases at Stromboli. The use of different variables, either independently or together in a real multiparametric approach, is seen as a necessity in order to maximize the efficiency of such statistical monitoring tools.

The focus of investigation of external modulation of volcanoes has recently moved from earth-tides to weather. Stromboli is permanently active volcano in an extensional environment, which makes it ideal for studying long period and long term effects. A time series of 17 months has been provided and gaps in the data were filled using a long period approximation. Pressure, temperature, humidity and rain were investigated as modulators and their traces cross-correlated with the seismic traces to find any relationships.

A fresh influx of magma is thought to increase large events and local temperature through higher heatflow, while high tremor levels could be contributing to the local humidity.

A 10 day delay between rainfall and tremor provided an opportunity. to investigate tremor mechanisms. Water entering the magma chamber was ruled out with geochemical evidence, however, hydrology showed that 10 days was an acceptable time for rainwater to enter the system.

Thermal fracture is ruled out at shallow depths, but there is evidence of a hydrothermal system, which may repair micro-fractures. Fatigue micro-fracture is likely to be the dominant sub-critical growth mechanism, caused by regular volcanic activity. Internal growth of the volcano contributes to the stress field and hydrothermal system. Fracture induced by increased pore fluid pressure is the most likely source of the tremor.

Further geochemical and long-term seismic studies are needed to confirm these theories.

Short period seismic data have represented for a long time the major geophysical dataset available for the investigation of the processes going on inside a volcano with moderate permanent explosive activity such as Stromboli (Aeolian Islands, Italy). The installation of (arrays of) broadband seismic instruments has considerably widened our views but in order to go further with the understanding and with the modelling, the integration of seismic data with other signals is a strong necessity.

A path-finder experiment was conducted in June 1999 recording simultaneous thermal, seismic and infrasonic measurements, and a more extensive field campaign was carried out in May 2000. The role of the gas has always been recognized as a major one in the dynamics of the conduit.

Analyses of data recorded in 1999 campaign indicate that periods of high and low rates of explosive activity can be related to the magnitude and frequency at which small gas bursts occur. These in turn can reflect the dynamics of the build-up and decay of foam layers and, possibly, offer an insight into the rate at which fresh, gas-rich magma supplies the shallow vents system. Variations in gas temperature, recurrence of gas bursts, and frequency of strombolian explosions indicate that magma-foam levels, or magma-gas supply, could change over minute-long periods as the system cycles between 5- to 40-minutes long periods characterized by different degassing rates.

These first results are being verified and extended by analyzing the longer datasets recorded in May 2000 campaign. In our model, foam layer generation and collapse at Stromboli's shallow system alternates between periods of low and high activity, which may be linked to the rise, and subsequent degassing and sinking, of discrete fresh magma/gas batches. This model seems to be confirmed by the good correlation between time delays between infrasonic and infrared onsets and temperature fluctuations at the vent.

Long period and hybrid events, seen at the Soufriere Hills Volcano, Montserrat, show dominant low frequency content suggesting the seismic wavefield is formed as a result of interface waves at the boundary between a fluid and a solid medium. This wavefield will depend on the impedance contrast between the two media and therefore the difference in seismic velocity. For a gas-charged magma, increasing pressure with depth reduces the volume of gas exsolved, increasing the seismic velocity. The seismic radiation pattern along the conduit can then be modelled. Where single events merge into tremor, gliding lines can sometimes be seen in the spectra and indicates either changes in the seismic parameters with time or varying triggering rates of the single events.

The differential equation describing the time dependence of bubble growth by diffusion is solved numerically for a variety of starting conditions: for a stationary magma column undergoing a decompression event, a column moving with constant velocity and also for an accelerating column, allowing for the expansion due to the increasing volume. The volume of gas is depth dependent, increasing with time as the bubbles grow and expand. It can be used to calculate the density, pressure and seismic velocity with depth and therefore how they evolve with time. The effect of parameters such as number density of bubbles and ascent rate to the rate of change of seismic velocity can be examined and compared with timescales given by the data.

Low-frequency seismic signals are believed to hold the key to the understanding of the internal stage of volcanic activity. In this contribution we try to link seismic observations directly to properties of the magma.

Modelling constraints are based on observations from Soufriere Hills volcano in Montserrat where low-frequency events occur often in cyclic swarms, are associated with tilt, and occasionally merge into harmonic tremor. Some episodes of tremor show shifting spectral lines, indicating the change of conduit/magma parameters in the order of minutes. The analysis of the spectral seismic signature could therefore provide a direct interpretation in terms of magma properties and excess pressure, which are crucial parameters for any assessment of volcanic activity.

We derive depth-dependent seismic velocity and density models for a gas-charged, visco-elastic fluid in a conduit by solving the differential equations governing degassing and diffusion, and the resulting pressure fluctuations in the magma. Based on that, we model the seismic wave field of low-frequency resonances in and around the conduit and include a possible feed-back system for a time-dependent triggering mechanism

No Abstract

Tuesday 19 September - Session 3

The seismic signals recorded around active volcanoes show a much greater variety than signals recorded in the vicinity of tectonic earthquakes. Wavefields of tectonic earthquakes are often recorded many wavelengths away from the source implying that the source can be represented as a point in space and a source time function which is very short. These geometrical considerations have allowed ray theory to be so successful for the analysis of local, regional and global earthquakes.

The problem is more complicated for active volcanoes: the recorded signals may be emergent, the sources may be of long duration, they may be located close to the summit of mountains with considerable topography, the signals may be recorded close to the actual sources, etc. All this implies that methods such as ray theory will not work well in such circumstances. What is needed are complete solutions to the solution of 3-D wave propagation problems including effects such as topography, attenuation and anisotropy.

The progress in the past few years in computer technology (in particular parallel processing) has allowed us to approach realistic frequency ranges for 3-D problems. The propects of TeraFlop calculations on TeraByte machines implies that soon the calculation of complex wave propagation problems in three dimensions will be feasible.

We will present first steps in preparing algorithms, which previously have been used in exploration, regional and global seismology for the problem of large scale problems of volcano seismology. This implies mainly (1) including topography and (2) parallelization and implementation on supercomputers. Two methods are investigated:

First, a staggered-grid finite-difference approach is used to solve the wave equation. Topography can be included by performing grid stretching along the z-direction according to local topography.
Second, we are performing tests with a novel approach to solving the wave equation on unstructured grids using explicit local differential operators based on the concepts of natural neighbors and voronoi cells. With these algorithms we want to investigate the scattering properties of volcanic regions and their topography, seismic signatures of tremors and pyroclastic flows.

We address the issue of locating the sources of volcanic tremor and emergent events by using methods that requires a limited amount of equipment. We propose to set up around the volcano a network of several small tripartite seismic antenna made of one component sensors.

At each array, the apparent velocity and azimuth of wave propagation are estimated on a sliding window by using the time delays between the sensors. Weighted histograms are then built up in order to select the dominant and stable directions of propagation. To determine the intersection of the directions obtained at the different arrays, a probability density function describes the global fit between the observed and the calculated azimuths.

The source position in the horizontal plane is taken at the point where the probability density is maximum. We present results of an experiment carried on at Arenal volcano with several small dense arrays of 3 to 8 seismometers. We test the precision and robustness of the method by varying the array configuration (number of sensors, distance between seismometers), by using different types of signal (explosions, LP events, harmonic tremor) and by exploring several methodological aspects. Notwithstanding some difficulties due to the wavefield complexity in the volcanic structure, some encouraging preliminary results are obtained.

No Abstract

Merapi volcano in Central Java belongs to the most active volcanoes in the world. Located at the subduction zone between the Eurasian and the Indo-Australian Plate, the dominant magmatic mechanism is andesitic, i.e. explosive. In 32 of 67 reported historical eruptions, pyroclastic flows took place. IAVCEI has classified Merapi as a high-risk volcano and included it into the list of 16 Decade Volcanoes.

In cooperation with the Volcanological Survey of Indonesia and other institutions in Indonesia and Germany, the GeoForschungsZentrum initiated an interdisciplinary monitoring program in 1994, supplementing a number of ongoing national and international activities on this volcano.

The MERAPI project (Mechanism Evaluation, Risk Assessment, Prediction Improvement) is supposed to contribute to the development of prediction and warning strategies on different time scales as well as early warning capability. Special emphasis has been put on continuously monitoring activity indicators such as seismicity, deformation and gas emanation.

The modeling and interpretation of the continuously monitored data are supported by information delivered by geophysical measurement campaigns. Active seismic profiling, microgravity measurements and geoelectric soundings as well as magnetotelluric and LOTEM measurements are contributing parameters of the internal structure of the edifice. All these information combined with geological investigations of magmatism and of the explosive history of the volcano form the basis for a better understanding of the complex processes prior to an eruption.

For the analysis of seismic events, especially for the separation of source, path and site effects seismic structure information is indispensable. Therefore, an active seismic experiment was carried out at Merapi volcano. 2.5 l mudguns, fired in 3 water basins 5 km apart from the summit, transmit a well defined seismic source signal. Three 3 km long seismic profiles were aligned radial around the volcano summit, each consisting of up to 30 3-component seismometers. First arrivals are vertically polarized, but their amplitudes fade with distance unusually. Correlated later phases can be explained by a 2D ray tracing model as refracted waves reflected on nearly vertical faults or weak zones. Some of this structures are equidistant to the summit and the nearest one, modeled with a velocity reduction of 10%, seems to be the transition boundary of the inner hydrothermal zone. However, the unusually long and complex wave trains cannot be sufficiently explained by deterministic models. The main part of energy is shifted towards later arrival times due to strong multiple scattering. These coda waves are polarized horizontally, with no preferred azimuths and mainly incoherent along profiles.

Using the diffusion model we find that the attenuation coefficient by scattering is much larger than the intrinsic one and a mean free path is around 100 m. It means that only a few 100 m far from the source strong multiple scattering has to be taken into account.

The use of high dynamic range, three-component seismometers on volcanoes has improved our resolution of the tremor and volcanic events which make up a volcano's seismic wavefield. We have not yet, however, been able to solve the problem of the source process, as has been done for the double-couple shear fracture at an earthquake's hypocenter. There are two major impediments: The medium and the source.

First, before a seismogram or seismogram characteristic may be used for source modelling, the effect of the medium must be recognized and removed. Second, unlike the recordings for most earthquakes, thus far no model proposed for volcanic sources has uniquely explained the wavefield. The moment tensor for the forces at the source would be a valuable additional input for modelling.

At several volcanoes, events have been recorded with non-double couple mechanisms. They have variously been attributed both to shear fractures on non-planar features and to processes involving injection of magma and/or hydrothermal fluids. As both the quantity and quality of volcano seismic data sets improve, the possibility of determining moments and moment tensors relating both to volcanic events and even volcanic tremor offer the opportunity to learn more about the forces at work in volcanoes

Tuesday 19 September - Session 4

Seismic activity is a common component of volcanic eruptions. The majority of volcanic earthquakes are small; their magnitude do not exceed 2 or 3. At the same time, the significant (moderate-size and large) volcanic earthquakes may occur also. The study of 28 significant (Mw ³ 4.5) worldwide earthquakes associated with the volcanic eruptions of XX-th century allowed us to formulate the following characteristics of significant volcanic earthquakes:

(1) During the XX-th century, there were recorded only 3 large (Mw ³ 7.0) earthquakes directly associated with the volcanic eruptions in the world.
(2) The maximum magnitude Mw was estimated as 5.4 for the earthquakes related to central eruptions, 7.0 for the caldera collapse, 5.9 for the central eruptions accompanied by flank and (or) fissure eruptions, 7.2 for the flank and (or) fissure eruptions without central eruptions, and 5.9 for the submarine eruptions. For the seismicity recorded just before the eruption, maximum magnitude of earthquake was estimated as 7.2, for the seismicity recorded at the initial stage of eruption as 7.0, for the paroxysmal stage of eruption as 5.6, and for the final stage of eruption as 6.1.
(3) It was shown that the recurrence time of eruptions of the same type of the volcano associated with significant (Mw ³ 4.5) earthquakes was more than 100 years for 86% of studied events. The conditions for appearance of significant volcanic earthquakes are discussed.

Volcano seismology often deals with rather shallow seismic sources and seismic stations deployed in their near-field. The complex stratigraphy on volcanoes and near-field source effects have a strong impact on the seismic wavefield complicating the interpretation techniques that are usually employed in earthquake seismology. In addition, as most volcanoes have a pronounced topography, the interference of the seismic wavefield with the stress-free surface results in severe waveform perturbations which affect seismic interpretation methods.

In this study we deal predominantly with the surface effects, but take into account the impact of a typical volcano stratigraphy as well as near-field source effects. We derive a correction term for plane seismic waves and a plane free surface such that for smooth topographies the effect of the free surface can be totally removed. Seismo-volcanic sources radiate energy in a broad frequency range with a correspondingly wide range of different Fresnel zones.

A 2-D boundary element method is employed to study how the size of the Fresnel zone is dependent on source depth, dominant wavelength and topography in order to estimate the limits of the plane wave approximation. This approximation remains valid if the dominant wavelength does not exceed twice the source depth. Further aspects of this study concern the particle motion analysis to locate point sources and the influence of the stratigraphy on particle motions. Furthermore, the deployment strategy of seismic instruments on volcanoes, as well as the direct interpretation of the broadband waveforms in terms of pressure fluctuations in the volcanic plumbing system are discussed.

No Abstract.

No Abstract.

No Abstract.