Seismic Tomography Experiment at Italy's Stromboli Volcano

Eos, Vol. 89, No. 30, 22 July 2008 VOLUME 89 NUMBER 30 22 JULY 2008 EOS, TRANSACTIONS, AMERICAN GEOPHYSICAL UNION PAGES 269–276 Seismic Tomography Experiment at Italy’s Stromboli Volcano Multibeam bathymetry data were also collected before the OBS positioning. During the experiment, the air gun signals were displayed on real-time monitors of the permanent stations. This allowed for adjustments of the planned shot geometry to obtain the best signal- to- noise ratios from the air gun pulses. PAGES 269–270 Stromboli Island, located in the southern Tyrrhenian Sea, is the emerged part (about 900 meters above sea level) of an approximately 3-kilometer-high stratovolcano. Its persistent Strombolian activity, documented for more than 2000 years, is sometimes interrupted by lava effusions or major explosions. Despite the number of recently published geophysical studies aimed at clarifying the volcano’s eruption dynamics, the spatial extent and geometrical characteristics of its plumbing system remain poorly understood. In fact, knowledge of the inner structure and the zones of magma storage is limited to the upper few hundred meters of the volcanic edifice [Chouet et al., 2003; Mattia et al., 2004], and P and S wave velocity models are available only in restricted areas [Petrosino et al., 2002]. Over the past few years, Stromboli has experienced an increase in its activity, which started with the eruptive crisis of December 2002 to July 2003. The observed variations in Stromboli’s eruptive behavior, and the occurrence in December 2002 of a tsunami and in April 2003 of a paroxysmal vulcaniantype event with strong explosions and lava flows [Bonaccorso et al., 2003], have focused attention on the need to determine more suitable internal structural and velocity models of the volcano. To meet this need, a seismic tomography experiment through active seismics using air gun sources was done. A preliminary two-dimensional (2-D) tomographic study is presented here. institutions. Researchers on board the R/V Urania of the Italian National Council of Research (CNR), which was equipped with a battery of four 210- cubic- inch generated injection air guns (GI guns), fired more than 1500 offshore shots along profiles and rings around the volcano [Marsella et al., 2007]. The shots—scheduled in 2- minute intervals corresponding to shot spacings of 250 meters—were recorded by 33 inland digital three-component seismic stations (13 permanent and 20 temporary) and 10 ocean- bottom seismometers (OBS) (Figure 1). The OBS deployment allowed for the first ever exploration of the Stromboli submarine edifice. Data and Preliminary Results A large proportion of the recorded data showed a satisfactory signal- to- noise ratio, although high-level background noise related to the continuous explosive activity of the volcano was recorded. Moreover, the seismic ray transmission was strongly affected by both the heterogeneous shallow crust and the geological setting of the volcanic edifice. The shots made in the north-northeast (NNE) and SSW sectors provided the best signals (see inset in Figure 1 for example). Seismic Tomography Experiment From 25 November to 2 December 2006, the first active seismic tomography experiment at Stromboli volcano was carried out with the cooperation of four Italian research BY M. CASTELLANO,V. AUGUSTI, W. DE CESARE, P. FAVALI, F. FRUGONI, C. MONTUORI, T. SGROI, P. DE GORI, A. GOVONI, M. MORETTI, D. PATANÈ, O. COCINA, L. ZUCCARELLO, E. MARSELLA, G. AIELLO,V. DI FIORE, M. LIGI, G. BORTOLUZZI,V. FERRANTE, E. MARCHETTI, G. LACANNA, AND G. ULIVIERI Fig. 1. The recording system and the shot lines for the experiment. The inland black and white dots indicate the temporary and permanent seismic stations, respectively. The ocean-bottom seismometer deployment (large black dots) is also shown. Some recordings of shot L02-S0020 are shown in the inset. The signals are filtered in the 5- to 20-hertz band. Eos, Vol. 89, No. 30, 22 July 2008 Therefore, we chose to process NNE-SSW and its orthogonal ENE-WSW profiles (approximately 180 shots) to obtain a preliminary 2-D P wave velocity model and a first tomographic image of the upper part of the volcanic structure. Although the selected data constitute only 12% of the entire data set, the sampled volume covers the volcanic structure up to 2 kilometers below sea level (Figure 2). The data have been inverted for the P wave velocity (Vp) structure by using the Simulps13q algorithm [Eberhart-Phillips and Reyners, 1997], considering a 3-D grid of nodes spaced from 0.5 to 2 kilometers deep, beneath the central part of the volcano. A total of 961 P observations were used to invert for 491 velocity nodes using a damped least squares inversion iterative procedure. A damping value of 5 was selected by the trade-off curve, as the best compromise between model complexity and residuals variance reduction. After three iteration steps, a variance improvement of 44% was obtained, reaching a final root-mean square (RMS) of 0.11802 seconds. These preliminary results show a P wave velocity of about 1.8–2.8 kilometers per second for the exposed part of the volcano. Slower velocity zones (1.8–2.0 kilometers per second) were located around the summit craters. The P wave velocity increases with depth, varying from 2.8–3.0 kilometers per second at sea level to 4.0–4.5 kilometers per second descending to about 1.5 kilometers below sea level. Relatively high velocity zones were visible both in the inner part of the volcanic structure, at about 1.0 kilometers below sea level (∼3.8 kilometers per second), and in the last 200–300 meters of the edifice, in correspondence with the volcanic conduit (2.6–2.8 kilometers per second) (Figures 3a and 3b). The low P wave velocities observed in the upper part and along the flanks of the edifice are in agreement with volcanological and petrological information about lavas and shallow incoherent deposits at Stromboli [Apuani et al., 2005]. The relatively high velocity zones could suggest the presence of intrusive bodies related to the plumbing system. The reliability of the Vp model has been verified by the analysis of the complete resolution matrix (RM) and by synthetic tests. We used the spread function (SF) as defined by Michelini and McEvilly [1991] to quantify how the averaging vector is picked around the diagonal element of the RM. The tested synthetic model consists of an axial high-Vp anomaly (Figures 3c and 3d), with Vp perturbations of 10% with respect to the starting model, located where the two profiles intersect (see Figure 2). The inversion was carried out as in the real case. The 80% and 50% of synthetic anomaly were recovered on the N8 and N115 profiles, respectively (Figures 3e and 3f). The analysis of RM and synthetic tests showed that SF = 1.5 may be assumed as a threshold of acceptable resolution. Fig. 2. Raypaths and sampled area obtained with the selected 180 shots. Update and Conclusions Acknowledgments Beginning on 27 February 2007, Stromboli volcano again underwent an eruption. On 15 March 2007, a major explosion occurred, heralded about 2 days prior by significant variations in high-rate (1 hertz) GPS spectral power densities and the occurrence of shallow volcanotectonic earthquakes [Patanè et al., 2007]. Although still preliminary, the tomographic images obtained provide a first innovative view of the P velocity structure of Stromboli volcano, which will be very useful for the precise localization of the seismicity affecting this volcano. The processing of about 50,000 P wave first arrival times is in progress to provide a fuller picture of the Stromboli structure, which could better highlight the shallow part of the volcano’s crustal plumbing system. We believe that the complete results of this project will increase the knowledge of the Stromboli structure and will lead to an improvement in understanding both the phenomena and the behavior of this volcano, for scientific and civil defense purposes. The authors are grateful to A. Tibaldi, C. Chiarabba, M. Martini, L. D’Auria, M. Ripepe, G. D’Anna, F. Bianco, and the Stromboli Tomography Working Group for their support. The experiment has been funded by the Italian Dipartimento per la Protezione Civile (DPC-INGV Agreement 2004-2006, Projects V2-03 and V2-13). References Apuani, T., C. Corazzato, A. Cancelli, and A. Tibaldi (2005), Stability of a collapsing volcano (Stromboli, Italy): Limit equilibrium analysis and numerical modelling, J.Volcanol. Geotherm. Res., 144(1-4), 191–210. Bonaccorso, A., S. Calvari, G. Garfì, L. Lodato, and D. Patanè (2003), Dynamics of the December 2002 flank failure and tsunami at Stromboli volcano inferred by volcanological and geophysical observations, Geophys. Res. Lett., 30(18), 1941, doi:10.1029/2003GL017702. Chouet, B., P. Dawson, T. Ohminato, M. Martini, G. Saccorotti, F. Giudicepietro, G. De Luca, G. Milana, and R. Scarpa (2003), Source mechanisms of explosions at Stromboli Volcano, Italy, determined from moment-tensor inversions of very-long-period data, J. Geophys. Res., 108(B1), 2019, doi:10.1029/2002JB001919. Eos, Vol. 89, No. 30, 22 July 2008 Eberhart-Phillips, D., and M. Reyners (1997), Continental subduction and three-dimensional crustal structure: The northern South Island, New Zealand, J. Geophys. Res., 102, 11,843–11,861. Marsella, E., et al. (2007), The Stromboli geophysical experiment: Preliminary report on wide angle refraction seismics and morphobathymetry of Stromboli island (southern Tyrrhenian sea, Italy) based on integrated offshore-onshore data acquisition (Cruise STR06 R/V Urania), Tech. Rep. 102, 85 pp., Ist. di Sci. Mar., Bologna, Italy. (Available at http://projects.bo.ismar.cnr.it/MEDITERRANEAN/ STROMBOLI/STR06_REP) Mattia, M., M. Rossi, F. Guglielmino, M. Aloisi, and Y. Bock (2004), The shallow plumbing system of Stromboli Island as imaged from 1 Hz instantaneous GPS positions, Geophys. Res. Lett., 31, L24610, doi:10.1029/2004GL021281. Michelini, A., and T.V. McEvilly (1991), Seismological studies at Parkfield: I. Simultaneous inversion for velocity structure and hypocentres using cubic B-splines parameterization, Bull. Seismol. Soc. Am., 81, 524–552. Patanè, D., M. Mattia, G. Di Grazia, F. Cannavò, E. Giampiccolo, C. Musumeci, P. Montalto, and E. Boschi (2007), Insights into the dynamic processes of the 2007 Stromboli eruption and possible meteorological influences on the magmatic system, Geophys. Res. Lett., 34, L22309, doi:10.1029/2007GL031730. Petrosino, S., P. Cusano, G. Saccorotti, and E. Del Pezzo (2002), Seismic attenuation and shallow velocity structures at Stromboli volcano, Italy, Bull. Seismol. Soc. Am., 92, 1102–1116. Author Information Mario Castellano, Vincenzo Augusti, and Walter De Cesare, Istituto Nazionale di Geofisica e Vulcanologia (INGV), Sezione di Napoli, Osservatorio Vesuviano, Naples, Italy; E-mail: [email protected]; Paolo Favali, Francesco Frugoni, Caterina Montuori, and Tiziana Sgroi, INGV, Sezione di Roma-2, Rome, Italy; Pasquale De Gori, Aladino Govoni, and Milena Moretti, INGV, Sezione di Roma–Centro Nazionale Terremoti, Rome, Italy; Domenico Patanè, Ornella Cocina, and Luciano Zuccarello, INGV, Sezione di Catania, Catania, Italy; Ennio Marsella, Gemma Aiello, and Vincenzo Di Fiore, Istituto per l’Ambiente Marino Costiero–Consiglio Nazionale delle Ricerche, Naples, Italy; Marco Ligi, Giovanni Bortoluzzi, and Valentina Ferrante, Istituto di Scienze Marine–Consiglio Nazionale delle Ricerche, Bologna, Italy; and Emanuele Marchetti, Giorgio Lacanna, and Giacomo Ulivieri, Dipartimento di Scienze della Terra, Università di Firenze, Florence, Italy Fig. 3. (a and b) Preliminary two-dimensional P wave velocity model along the NNE-SSW and ENEWSW profiles. (c and d) Synthetic Vp model along the ENE-WSW and NNE-SSW profiles. (e and f) Reproduced synthetic anomalies after the inversion.The bold black lines—starting at about 1 kilometer below sea level in Figures 3a, 3b, 3e, and 3f—indicate the well-resolved regions of the model with speed function values ≤ 1.5. Original color image appears at the back of this volume. Eos, 2 22 January 2007 Eos,Vol. Vol.88, 89,No. No.1, 30, July 2008 Page XXX Fig. 3. (a and b) Preliminary two-dimensional P wave velocity model along the NNE-SSW and ENE-WSW profiles. (c and d) Synthetic Vp model along the ENE-WSW and NNE-SSW profiles. (e and f) Reproduced synthetic anomalies after the inversion. The bold black lines—starting at about 1 kilometer below sea level in Figures 3a, 3b, 3e, and 3f—indicate the well-resolved regions of the model with speed function values ≤ 1.5. Page 270