Advances in the VT earthquakes locations by new structural models at Mt. Etna
obtained by using velocity and attenuation tomographies
Barberi G., Cocina O., Giampiccolo E., Patanè D., Zuccarello L.
Abstract
In the last decade, the high rate of eruptive events have offered good examples for quantitative
analysis based on seismic data. Several seismological researches have focused on Mt. Etna’s
structure by using different tomographic techniques and data sets. The overall findings provided
valuable insight into the behavior of the volcano as well as a significant contribution to modeling
its dynamics during eruptions. In this paper we discuss the state of the art of the studies performed
in recent years, together with assumptions and limitations. Thereafter, we present new velocity and
attenuation models obtained in the frame of the VOLUME Project by analysing the seismicity
preceding and accompanying the 2002-2003 flank eruption. The improvement in structural
resolution and the inference power brought about by the increased number and quality of seismic
stations and data are highlighted. Moreover, VP, VP/VS and QP are jointly interpreted giving
valuable constraints to the plumbing system of the volcano. Finally, we provide a better
description of the spatial distribution of the seismicity through accurate hypocentral locations
obtained by applying the NonLinLoc probabilistic approach and the new 3D velocity model.
All the studies carried out in the frame of the present project led to important and innovative
results other than providing valuable tools for future forecasting of Mt. Etna’s volcanic activity and
mitigation of volcanic risk.
Introduction
Mount Etna is one of the most active and powerful basaltic stratovolcano formed by several
superposed volcanic edifices. Like many basaltic volcanoes, Mt. Etna typically erupts in effusive to
weakly explosive styles, even though powerful explosive eruptions (sub-Plinian to Plinian events)
have been recognized during the Quaternary (Del Carlo et al., 2004).
Over the last 30 years Mt. Etna has had a high rate of eruptive events and therefore it constitutes
one of the most important natural laboratories for the understanding of eruptive processes and lava
uprising in basaltic-type volcanic environments. Seismic observations allowed us to carry out
detailed investigations on major aspects of seismicity. In particular, the most recent effusive
eruptions (e.g., 2001, 2002-2003, 2004, 2006, 2008) have offered good examples for quantitative
analysis based on seismic data. A repeatedly observed feature at Mt. Etna is that the summit
eruptions are not generally preceded by significant variations in the pattern of deformation and
volcano-tectonic seismicity (Patanè et al., 2004). On the other hand, lateral eruptions, such as the
two most recent 2001 and 2002-2003 eruptions, are forerun by a few days/hours of seismic crisis
just before the eruptive fissures open.
Seismicity is almost continuous in time and occurs over wide crustal volume ranging from
depth above the sea level, inside the volcanic edifice, to 30 km bsl in the western sector of the
volcano. Towards eastern sector of the volcano, the seismogenic depths are limited to the first 10
km of the crust. Among the seismic signals recorded in the area, two main classes are recognized
(Patanè et al., 2004): i) VT earthquakes which are tectonic-like events generated by regional
tectonic stresses and/or by local stresses deriving from magma migration in the earth’s crust and ii)
seismic signals related to the seismic manifestation of fluids dynamics (explosion quakes, LP
events, volcanic tremor).
In the frame of the VOLUME Project, aimed to enhance the knowledge of the structure and
plumbing system of the volcano, we performed tomographic studies both in velocity and attenuation
by analysing the seismicity which preceded and accompanied the 2002-2003 flank eruption (Fig. 1).
High resolution location techniques have also been applied to the same dataset of earthquakes, to
better investigate on the space time distribution of the seismicity. The 2002-2003 eruption was one
of the best documented lateral eruption at Mt. Etna. The eruption started in the night between
October 26 and 27, 2002 with a seismic swarm occurred in the central and upper part of the
volcano. Fissures on both the NE and S flanks were activated with a huge lava emission and
powerful explosive activity from the southern fracture field. A total of 874 M D ≥1 earthquakes
were recorded from the beginning of the seismic swarm until the end of the eruption. The swarm
decayed over about 2 weeks and most of the seismic energy was released during the first 4 days
(470 events of a total of 874, Mmax = 4.4). The eruption ended on January 28, 2003, after 94 days
(for detail see Barberi et al., 2004).
An overview of tomographic studies at Mt. Etna
On a volcano, images of the seismic velocity and attenuation structure provide direct constraints
on the distribution of magma, thermal anomalies, and hydrothermal alteration . Moreover, the
spatial variation of the P-wave velocity to S-wave velocity ratio (VP/VS) in a volcanic area is
considered a useful potential diagnostic tool to discriminate between low P-velocity anomalies due
to partially molten rock and lowering of P velocity associated with fractured, vapor-dominated
hydrothermal systems. However, the interpretation of low- and high- velocity anomalies and/or
velocity ratio (VP/VS) becomes complicate due to the presence of fluids, cracks, and gas inside a
volcano. Different techniques for different datasets have been applied on Mount Etna to define the
velocity structure of the volcano, with the aim to locate magma stationing volumes. A common
feature of all the tomography studies which investigated the volcano at a large scale Chiarabba et
al., 2004 and references therein) is the presence of a high velocity body in the central part of the
volcano, extending towards SE. This volume, has been recognised between 1 km a.s.l. and 10 km
b.s.l and is generally interpreted as a main solidified body. The most recent tomography studies
have better detailed the shallower velocity structure of the volcano (Patanè et al., 2002, Patane et
al., 2006) and better refined the geometry of the high velocity body. This was allowed by the
higher number of stations operating during that period on Mt Etna. Moreover, the availability of a
good number of three-component stations allowed the precise recognition of a higher number of S
picking permitting to calculate also the VP/VS tomographic images.
To gain a full understanding of the physical processes involved in active volcanoes, the velocity
tomography models have to be jointly interpreted with attenuation analysis. In particular, the
attenuation of body waves is very sensitive to the thermal state of the crustal volume travelled by
seismic waves and to the saturation of rocks with fluids and partial melts (e.g., Sato and Sacks,
1990 and references therein).
Following the first studies on attenuation carried out at Mt. Etna using the Q-coda method (see
Patanè and Giampiccolo, 2004 and references therein), different approaches were used to estimate
the quality factor of body waves in the spectral (e.g. Patanè et al., 1994; Giampiccolo et al., 2007)
or time domain (de Lorenzo et al., 2006). These studies have pointed out significant variations of
QP and QS as a function of source depth and source-station azimuths. In particular, a drop of QP
values typically marks data issuing from earthquakes located at depth less than 5 km supporting the
idea that the shallow volcanic layers are characterized by high attenuation (Patanè et al., 1994). On
the other hand, as typical of volcanic environments, strong azimuthal variations of QP suggest the
presence of local heterogeneities and/or fluid-filled crack volumes (de Lorenzo et al., 2006;
Giampiccolo et al., 2007). Finally, a weak frequency dependence of Q P and QS according to the
power law Q=Q0fn, with n<0.5, was found in the central part of the volcano (Giampiccolo et al.,
2007) interpreted as due to the presence of magma bodies and partial melting of rocks.
A first P-wave attenuation tomography study was carried out at Mt. Etna, down to 15 km depth,
by De Gori et al. (2005). The main features of the obtained 3D QP model are the strong low- QP
anomalies found at 0 and 3 km depth, interpreted as the regions where the magma was stored at
shallow depth before the recent eruptions, and the small low- QP magmatic conduit in the central
and western part of the volcano at 6–12 km depth. In the deepest part of the model, a clear high- QP,
high-Vp anomaly represents the remnant of the old intrusions that fed the past activity of volcano.
Results from this first tomography have been refined in the shallow layers (around 2 km) of the
volcanic edifice by analyzing shallow earthquakes recorded during the 2001 eruption (MartìnezArevalo et al., 2005). In particular, a region of very low QP values (between 10 and 30), located in
the same place where the 2001 dike emplaced was found and was interpreted as the effect of fluid
intrusion (magma rich in gas) in the uppermost part of the Etna volcano, feeding the 2001 eruption.
Enhancement of the velocity and attenuation models
Recently, Patanè et al., (2002) found an anomalous low VP/VS volume (ca. 1.6), located where
the magma intruded during the 2001 eruption. In literature low VP/VS values are usually associated
with molten material wealthy in gas or fluids in supercritical state (Sanders et al., 1995). Therefore,
the authors related this anomaly to the presence of magma rich in gas feeding the 2001 eruption
characterized by an elevate explosive content.
On the basis of this interesting result we propose, in the frame of VOLUME project, to enhance
the velocity structure in the shallower portion of the crust under Mt. Etna by studying the 20022003 eruption which showed a higher explosive activity with respect to the 2001 one.
A total of 712 well constrained earthquakes recorded from the end of 2001 eruption to the end
2002-2003 one were analyzed. In particular, 8587 P and 2293 S-wave arrivals were inverted to
model a 2 x 2 x 1 grid by using the SIMULPS-14 software (Thurber, 1993). New threedimensional VP and VP/VS models have been computed in the first 6 km of the crust under Mt.
Etna. The new 3D velocity structure better defines the shape and geometry of the upper portion of
the high-VP volume, already recognized by previous tomography studies. Moreover, thanks to the
high resolved VP/VS structure, an anomalous volumes with very low VP/VS ratio (values as small
as 1.64) was well recognized during the eruptive period and located in the same place of the 20022003 dike intrusions, as also modelled by geodetic data (Fig. 2) . We also investigated on the time
evolution of the velocity structure on the volcano starting from the end of the 2001 eruption and
using repeated three-dimensional tomography technique (4D tomography, Patanè et al., 2006). The
results showed that the anomalous volume related to the presence of magma rich in gas is weakly
present also some months before Therefore, the application of 4D tomography allowed to have
some indication on the beginning of the recharging phase, after the 2001 eruption, which lead to
the 2002-2003 one.
With the aim to better constrain the results of the new 3D velocity model and to enhance the
existing attenuation structure at Mt. Etna, especially in the shallow part of the crust, we computed
a new 3D QP model by using a selected data set (about 300) of shallow events with depth 5 km.
We followed the approach outlined by Rietbrock (2001) according to which the inversion
process can be divided into two main steps: i) the determination of the t* operator from P-wave
spectral slopes; ii) the spatial inversion of the whole t* data set.
A total of 2152 t* from 202 events were inverted for the 3D QP structure, by using the Thurber
(1993) code modified for attenuation by Rietbrock (2001). The medium was parameterized with the
same irregular 3D grid of nodes and velocity values reported in Patanè et al. (2006), assuming an
initial QP equal to 100. The retrieved pattern of 3D QP is reported in Fig. 3.
The most important result of this tomographic study (Fig. 3) is the NS trending high attenuation
anomaly (QP values between 30 and 50) located in correspondence with the fracture system active
during the 2002-2003 eruption and in the same place as the 2002-2003 dike intrusion modelled by
Patanè et al. (2005). A second low QP volume is found at 0-2 km depth beneath the northeastern
flank of the volcano. This low QP anomaly may be related both to the presence of soft and heated
rock volumes in the northeastern Rift and to the high degree of cracking, along the Pernicana fault
(e.g., Azzaro et al., 2001). Beneath the Valle del Bove, a clear high QP anomaly (between 100 and
170) is observed. This volume coincides with the region of high QP and high VP evidenced in
almost all of the attenuation and velocity tomographic studies (e.g., Patanè et al., 2002; Chiarabba et
al., 2004; Martinez-Arévalo et al., 2005; Patanè et al., 2006), interpreted as an intrusive body
composed by high density non-erupted material, almost entirely solidified. The high QP anomaly
beneath the eastern and southeastern flanks of the volcano is almost located in the same place where
strong high VP/VS volumes are observed during the eruptive period (Patanè et al., 2006). In these
regions, crack density increases because of the intense fracturing occurring both during the preeruptive and eruptive periods. Patanè et al. (2006) hypothesized that the high VP/VS anomalies are
related to the rapid migration of fluids from the intrusion zone into the cracked regions beneath the
flanks. Since attenuation is sensitive to saturation (e.g. Nur, 1987), the joint interpretation of V P,
VP/VS and QP models yields higher constraints on the degree of saturation in rocks. In fact,
following Nur (1987), we interpret the high VP/VS and high QP volume as fully saturated rocks.
Improvements in earthquake locations by using the new velocity model
The recent definition of 3D structural models, which more adequately represent the internal
structure of the volcano, have allowed a more accurate investigation and characterization of the
seismic signals recorded, as well as an accurate description of their space-time evolution.
Recent studies have shown substantial improvements in location precision for earthquakes when
event-clustering techniques and waveform cross-correlation are used to improve arrival time
estimates or determine high-precision relative arrival times. Gambino et al., (2004) already tested at
Mt. Etna, for the pre-eruptive 2002-2003 period, the double-difference approach by Waldhauser and
Ellsworth (2000) on 3D locations. The authors conclude that the use of a standard tomography has
the ability to improve the locations by a more reliable velocity model and that the HypoDD location
method, employing 3D hypocentral starting parameters, produces much more accurate relative
event locations. In fact, by comparing HypoDD and 3D locations the application of HypoDD code
on the 3D locations further collapses the remnant scattered seismicity, improving the spatial
clusters of earthquakes.
In order to obtain better constrained Mt. Etna’s VT events location we applied the probabilistic
non-linear earthquake location method (NonLinLoc) defined by Lomax et al. (2000) to relocate the
same dataset used for the tomographic inversions. In particular, we compared the non-linear
location results with those obtained by using 1D standard, traditional and linearized earthquake
location algorithms (Hypoellipse; Lahr, 1989), and those obtained by using a 3D linearized
inversion (SimulPS; Thurber, 1983).
NonLinLoc combines the probabilistic methodology by Tarantola and Valette (1982) with
efficient global sampling to obtain an estimate of the posterior probability density function (PDF) in
3D space, for seismic event location. The location program NonLinLoc can be used with any
available velocity model (1D, 2D, or 3D) and represented by a posteriori probability density
function (PDF) (Lomax and Curtis, 2001).
We used the Oct-Tree algorithm which is 100 times faster than the grid search algorithm. This
algorithm involves an initial global sampling of the misfit function on a coarse grid, followed by
repeated application of division of the grid cell with the highest location probability and the
evaluation of the misfit function in sub-cells.
All the events were relocated using the same parameters such as number of samples drawn from
PDF (1000) and initial grid size that is the one used for the 3D velocity inversion (Patanè et al.,
2006). The PDF includes information on uncertainty and resolution and represents a complete,
probabilistic solution to the location problem. In the comparison with the 1D and 3D method, the
Oct-tree shows how our relocated earthquakes are better constrained along a specific seismic
pattern mainly focused in the upper part of the eastern sector of the volcano. In particular, the
strongest earthquakes occur along well known structural alignments (the northern rim of Valle del
Bove; the Pernicana fault system; the Santa Venerina area) (Fig. 4).
Even if the maximum depth of sesmicity remains unchanged, with the NonLinLoc method most
of the shallow seismicity moves to greater depth, suggesting that the probabilistic earthquake
location offers more consistent and reliable focal depths than linearized earthquake location.
Finally, the NonLinLoc method provides, thanks to the new 3D velocity models more stable
solutions with a significant improving of the location quality (RMS and GAP values).
Concluding remarks
The new velocity and attenuation models here obtained provided an adequate 3D structural
image of etnean crust and useful constraints on the magma intrusion and mechanical properties of
rocks and fluids during the 2002-2003 Mt. Etna pre-eruptive and eruptive period. Furthermore,
these new models allowed to enhance the clustering capability of etnean seismicity by applying also
the NonLinLoc method for hypocentral locations, suggesting that this approach could be used in
routinely earthquake locations on Mt. Etna for surveillance purposes.
Although we have already obtained very good results in these fields, further analysis are still in
progress. In particular, by the application of high-resolution double-difference seismic tomography
methods (see Montellier et al., this volume), we expect in future to obtain a more detailed velocity
structure at the spatial resolution of as little as a few hundred meter
References
Azzaro, R., Mattia M., Puglisi G., (2001), Fault creep and kinematics of the eastern segment of the
Pernicana fault (Mt. Etna, Sicily) derived from geodetic observations and their tectonic
significance, Tectonophysics, 333, 401 – 415.
Barberi, G., Cocina O., Maiolino V., Musumeci C., Privitera E., (2004), Insight into Mt. Etna (Italy)
kinematics during the 2002-2003 eruption as inferred by seismic stress and strain tensors,
Geophys. Res. Lett., 31, L21614, doi: 10.1029/2004GL020918.
Chiarabba, C., De Gori P. and Patanè D., (2004), The Mt. Etna plumbing system: The contribution
of seismic tomography. In: Mt. Etna volcano, American Geophysical Union, AGU (Geophysical
monograph series), 191-204.
De Gori, P., Chiarabba C. and Patanè D., (2005), QP structure of Mt. Etna: constraints for the
physics of the plumbing system, J. Geophys. Res., 110, B05303, 10.1029/2003JB002875.
De Lorenzo S., Filippucci M., Giampiccolo E., Patanè D., (2006), Intrinsic Qp at Mt. Etna from the
inversion of rise times of 2002 microearthquake sequence, Annali di Geofisica, 49(6), 12151234.
Del Carlo P., Vezzoli L., and Coltelli M., (2004), Last 100 ka tephrostratigrafic Record of Mount
Etna. In: Mt. Etna volcano, American Geophysical Union, AGU (Geophysical monograph
series), 77-89.
Gambino, S., Mostaccio A., Patanè D., Scarfì L., and Ursino A., (2004), High--precision locations
of the microseismicity preceding the 2002–2003 Mt. Etna eruption, J. Geophys. Res., VOL. 31,
L18604, doi:10.1029/2004GL020499.
Giampiccolo, E., D’amico S., Patanè D. and Gresta S., (2007), Attenuation and source parameters
of shallow microearthquakes at Mt. Etna volcano, Bull. Seism. Soc. Am., 97, 1B, 184-197.
Lahar, J.C., (1989). HYPOELLIPSE/Version 2.0: A computer Program for Determining Local
Earthquake Hypocentral Parameters, Magnitude and First Motion Patter, 92 pp., U.S. Geol.
Surv., Menlo Park,.
Lomax, A., J. Virieux, P. Volant and C. Berge, (2000), Probabilistic earthquake location in 3D and
layered models: Introduction of a Metropolis-Gibbs method and comparison with linear
locations. In Advances in Seismic Event Location, C.H. Thurber,, and N. Rabinowitz (eds.),
Kluwer, Amsterdam, 101-134.
Lomax, A. and A. Curtis, (2001). Fast, probabilistic earthquake location in 3D models using Octtree importance sampling. Geopys. Res. Abstr., Vol. 3, pp. 955.
Martınez-Arèvalo, C., Patanè D., Rietbrock A., and Ibanez J. M., (2005), The intrusive process
leading to the Mt. Etna 2001 flank eruption: Constraints from 3-D attenuation tomography,
Geophys. Res. Lett. 32, L21309, doi 101029/2005GL023736.
Nur, A., (1987), Seismic rock properties for reservoir descriptions and monitoring. In: G. Nolet,
Editor, Seismic Tomography, D. Reidel Publ. Co, Dordrecht, pp. 203–237.
Patanè D., and Giampiccolo E., (2004), Faulting processes and earthquake source parameters at
Moubt Etna: state of the art and perspectives. In: Mt. Etna volcano, American Geophysical
Union, AGU (Geophysical monograph series), 167-189.
Patanè D., Cocina O., Falsaperla S., Privitera E., and Spampanato S., (2004), Mt. Etna volcano: a
seismological framework. In: Mt. Etna volcano, American Geophysical Union, AGU
(Geophysical monograph series), 147-165.
Patanè, D., Barberi G., Cocina O., De Gori P. and Chiarabba C., (2006), Time-resolved seismic
tomography detects magma intrusions at Mount Etna, Science, 313, 821-823.
Patanè, D., Chiarabba C., Cocina O., De Gori P., Moretti M. and Boschi E., (2002), Tomographic
images and 3D earthquakes locations of the seismic swarm preceding the 2001 Mt. Etna
eruption: Evidence for a dyke intrusion. Geophys. Res. Lett., 29, 10, 10.1029/2001GL014391.
Patanè, D., Ferrucci F. and Gresta S., (1994), Spectral features of microearthquakes in volcanic
areas: attenuation in the crust and amplitude response of the site at Mt. Etna, Italy, Bull. Seism.
Soc. Am., 84, 1842-1860.
Patanè, D., Mattia M. and Aloisi M., (2005), Shallow intrusive processes during 2002-2004 and
current volcanic activity at Mt. Etna, Geophys. Res. Lett., 32, L06302, 10.1029/2004GL021773.
Rietbrok, A., (2001), P-wave attenuation structure in the fault area of the 1995 Kobe earthquake, J.
Geophys. Res., 106, 4141-4154.
Sanders C. O., Ponko S. C., Nixon L. D., Schwartz E. A., (2005), Seismological evidence for
magmatic and hydrothermal structure in Long Valley caldera from local earthquake attenuation
and velocity tomography, Seismological evidence for magmatic and hydrothermal structure in
Long Valley caldera from local earthquake attenuation and velocity tomography J. Geophys.
Res. 100, 8311-8326.
Sato, H., and Sacks I. S., (1990). Magma generation in the upper mantle inferred from seismic
measurements in peridotite at high pressure and temperature. In Magma Transport and Storage,
edited by M. P. Ryan, pp. 277-292, John Wiley, New York.
Tarantola, A. and B. Vallette, (1982). Inverse problems: Quest for information, J. Geophys. Res.,
50, 159-170.
Thurber, C. H. (1983). Earthquake locations and three-dimentional crustal structure in the Coyote
Lake area, central California, J. Geophys. Res., 88, 8226-8236.
Thurber, C.H., (1993), Local earthquake tomography: Velocities and VP/VS theory. In: Seismic
Tomography, edited by H.M. Iyer, and K. Hirahara, 563-580,Chapman and Hall, New York.
Waldhauser, F., and Ellsworth W. L., (2000), A double-difference earthquake location algorithm:
Method and application to the northern Hayward fault, Bull. Seismol. Soc. Am., 90, 1353–1368.
CAPTIONS
Fig. 1. Map of Mt. Etna area reporting the seismic stations operating during the study periods. The
elevation contour lines (gray lines) at 1,000 m intervals are also illustrated. Gray circles are the
epicentral locations of the selected earthquakes recorded during the 2002-2003 eruption.
Fig.2. VP/VS models down to 4 km b.s.l.. In each layer, the white lines surround the well resolved
regions of the model with SF < 2. The gray lines are the elevation isolines (every 500 m).
Fig. 3. QP models obtained for the 2002-2003 dataset. In each layer, the white lines surround the
well resolved regions of the models with SF<1.5. The black dots represent the earthquakes located
with the 3D velocity model (Patanè et al., 2006).
Fig.4. Map and cross sections of Mt. Etna showing the relocated earthquakes occurred during the
2002-03 eruption using Oct-Tree sampling algorithm (grey dots) and the relative a posteriori
probability density function (PDF).