Convenzione-Quadro tra Dipartimento della Protezione Civile e Istituto Nazionale di Geofisica e Vulcanologia Triennio 2007-09 Convegno annuale dei progetti sismologici
Sala convegni del Rettorato - Università Roma Tre, Via Ostiense 159, Roma 19-21 Ottobre 2009
Abstracts
1
Session L’Aquila Earthquake Sequence
Strong Ground Motion Characteristics from the 6 April 2009 L’Aquila earthquake, Italy.
1Aybige Akinci, 1Luca Malagnini and 2Fabio Sabetta
1) Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Roma- Italy 2) Dipartimento della Protezione Civile, Via Vitorchiano 4, 00189 Roma, Italy
Abstract The April 6, 2009 earthquake occurred at 03:33 a.m. local time, magnitude Mw=6.3, in the Abruzzo region (Central Italy), close to the city of L’Aquila (67000 inhabitants), located at about 5 km northeast to the epicenter. The main shock has been recorded by fifty-eight accelerometric stations the highest number of digital recordings ever obtained in Italy for a single earthquake; this makes the Abruzzo event one of the best recorded earthquakes caused by a normal fault mechanism. Very high values of peak ground acceleration were recorded near L’Aquila town centre (5 stations at zero distance from the fault) with a station reaching at 0.63g value. The earthquake caused severe losses (299 victims and1500 injured) and damages (about 18000 unusable buildings) in the epicentral area. In this study, we aim discussing the issues related to the earthquake engineering applications and implications of the recorded ground motion. Therefore, we analyze the ground motion characteristics of both the main shock in terms of peak ground acceleration (PGA), peak ground velocity (PGV) and pseudo-acceleration response spectra (5% of damping ratio). In particular, we compare the 5% damped pseudo-acceleration response spectra with the EC8 design spectrum and the new Italian building code (NTC08). In order to understand the ground motion characteristics of the L’Aquila earthquake better, we also study the site response of the strong motion stations that recorded the Aquila seismic sequence. The spectral ratio and H/V procedures were applied to all the events for which the strong-motion recordings are available.
Evaluation of the local site effects in the middle Aterno valley
Ameri G.1, Augliera P.1, Azzara R.M.1, Bergamaschi F.1, Bordoni P.1, Cara F.1, Cogliano R.1, Cultrera G.1, D’Alema E.1, Di Giacomo D.4, Di Giulio G.1, Fodarella A.1, Franceschina G.1, Gallipoli M.R.2, Harabaglia P.3, Ladina C.1, Lovati. S.1, Luzi L.1, Marzorati S.1, Massa M.1, Milana G.1, Mucciarelli M.3, Pacor F.1, Parolai S.4, Picozzi M.4, Pilz M.4, Puglia R.1, Pucillo S.1, Riccio G.1
1Istituto Nazionale di Geofisica e Vulcanologia; Sezioni di Milano-Pavia, Roma1, Centro Nazionale Terremoti; [email protected], [email protected] 2IMAA-CNR, C. da S. Loja, Tito Scalo (PZ) 3DiSGG-Università della Basilicata, V.le Ateneo Lucano, Potenza 4Helmholtz Centre Potsdam GFZ German Research Centre for Geosciences, Telegrafenberg, 14473 Potsdam, Germania
The day after the April 6th, 2009, L’Aquila earthquake (Mw 6.3) and for more than one month, about 30 sites have been monitored with accelerometers and velocimeters to study the local site effects in the middle Aterno valley. The installation was carried out by several research institutions, including Istituto Nazionale di Geofisica e Vulcanologia (Italy), Università della Basilicata (Italy) and GFZ-Potsdam (Potsdam, Germany). The stations have been installed nearby or inside the most damaged villages, where the macroseismic intensity reached the IX/X MCS degree (i.e. Onna, Paganica, Poggio Picenze, etc.) and where coseismic surface effects have been observed. The seismic stations have been installed in free-field or inside buildings. In particular, the accelerometric sensors have been used to record the largest aftershocks and to evaluate the possible presence of nonlinear effects. The huge amount of collected data allows the estimation of the local amplification phenomena through empirical techniques, such as the horizontal to vertical spectral ratios for a single station (HVSR) or the spectral ratio between the horizontal (or vertical) components of the selected site and a reference site (SSR). The preliminary results of this study indicate that the villages located in the middle Aterno valley suffered relevant local amplifications at frequencies generally higher than 1 Hz, so that the large observed damage can be in part ascribed to local site effects.
Characteristics of strong ground motions recorded during the April 2009 L’Aquila (central Italy) seismic sequence
G. Ameri1, M.Massa1, D. Bindi1, E. D'Alema1, A. Gorini2, L. Luzi1, S. Marzorati1, F.Pacor1, R. Paolucci3, R.Puglia1 and C.Smerzini4
1INGV Milano – Pavia, Milan, [email protected] 2Dipartimento della Protezione Civile - Ufficio Valutazione, Prevenzione e Mitigazione del Rischio Sismico (SISM), Rome 3 Department of Structural Engineering, Politecnico di Milano, Milan, Italy 4 PhD student, Rose School, Pavia, Italy
The 6 April 2009 L’Aquila MW 6.3 earthquake and its aftershocks yielded the most extensive set of strong-motion data in the near-source region yet obtained in Italy. The mainshock was recorded by 56 strong-motion stations belonging to the RAN, with 19 of these located within 50 km of the surface projection of the fault. The available data set is composed of more than 300 three-components strong-motion records from Mw ≥ 4 events recorded by RAN and INGV MI-PV stations, with about 90 records within 50 km of the corresponding epicenters. The strong ground motions from the mainshock show a clear dependence on azimuth, that can be attributed both to source effects (i.e., directivity effects) and to different attenuation properties of seismic waves at crustal scale (as suggested by the peak acceleration maps from the two strongest aftershocks). These records contribute to fill important gaps in the magnitude-distance-style of faulting distributions of global and regional data sets used to derive ground motion prediction equations. The near-fault ground motions (RJB=0) are generally underestimated by GMPEs, while at larger distances the opposite trend is observed, especially for high- frequency ground motion parameters (i.e., PGA). The residuals calculated for the whole sequence respect to different GMPEs show a negative bias indicating that the recorded ground motions are on average smaller than the ones predicted the empirical models. Preliminary analysis of near-fault records from the mainshock shows that peak motion varies significantly for stations within 5 km from the epicenter. The PGA ranges from 327 cm/s2 to more than 1 g (AQM saturated station). A specific baseline correction procedure was applied to these records in order to recover permanent displacements, that were found to be consistent with results based on GPS measurements. A comparison of the observed acceleration response spectra with recently proposed design spectra for the town of L’Aquila shows that the near-fault motion generally exceeded the no-collapse limit state design spectra both for horizontal and vertical components.
Instrumental investigations after the 6 April 2009 L’Aquila earthquake
Ameri G. 1, Augliera P. 1, D’Alema E. 1, Franceschina G. 1, Galadini F. 1, Ladina C. 2, Lovati S. 1, Luzi L. 1, Maistrello M. 1, Marzorati S. 1, Massa M. 1, Piccarreda D. 1, Puglia R. 1
1.Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Milano-Pavia, Via Bassini, 15, 20133 Milano, [email protected] 2.Istituto Nazionale di Geofisica e Vulcanologia, Centro Nazionale Terremoti, Via di Vigna Murata 605, 00143 Roma
This work synthesizes the activities conducted in the field by The Milano-Pavia Department of INGV in the three months following the Mw 6.3 event occurred on 6 April 2009 at L’Aquila town (central Italy). The work has been carried out in three phases: 1) deployment of several strong-motion stations in order to record the strongest aftershocks in the near-field, in the days which followed the main event; 2) detection of local site effects in the framework of the C.A.S.E. project, to investigate the seismic response of the areas selected for the construction of the emergency residences for homeless people; 3) geophysical investigations as a contribution to the seismic microzoning for the post-emergency reconstruction in the municipalities with the highest degrees of damage. Phases 2 and 3 had the aim of individuating the soil resonance frequencies through the empirical analysis of the ground motion, such as ambient noise and weak motion, recorded by temporary seismic stations. The stations have been installed in sites with different lithologic and geomorphologic conditions, and a huge data set has been collected to investigate site effects. The instruments have worked from 7 April to 13 July 2009; recording 469 aftershocks in the local magnitude range 2.5 - 5.3. The analysis carried out for phases 2 and 3 was based on the spectral ratio techniques, using a single station (Nakamura technique, and Horizontal to Vertical Spectral Ratio of weak motions) or a reference station (Standard Spectral Ratio). In the three phases the aftershocks have been recorded in 30 sites, corresponding to 25 localities inside the province of L’Aquila; using 11 seismic stations, 9 velocimetric sensors and 8 accelerometers. Finally 70 ambient noise measurements have been executed.
The April 6th 2009 L'Aquila earthquake: the peculiar view of the Gran Sasso geodetic interferometers.
Amoruso A.1, Crescentini L.2
1.Dipartimento di Fisica, Università di Salerno, Via Ponte Don Melillo, Fisciano (SA), [email protected] 2.Dipartimento di Fisica, Università di Salerno, Via Ponte Don Melillo, Fisciano (SA), [email protected] email
Two laser interferometers for geophysical purposes is operating at 1400 m depth beneath the Gran Sasso massif since a few years. The interferometers measure changes in length of two 90-m-long baselines. One baseline is approximately perpendicular to the local direction of the Apennines, and the other is approximately parallel to it. The presence of low-pass filters in the electronics precluded recording coseismic offsets due to the main shock and major aftershocks of the April 6th L'Aquila earthquake. Nevertheless, the interferometers did not lose optical alignment and produced very clear recordings of postseismic deformation with uncommon time resolution (few milliseconds) and exceptionally high signal-to-noise ratio. Here we show results from the analysis of the initial few-day-long deformation history. Deformation after about 1.5 days is fully consistent with afterslip on a stationary region of the fault responsible for the ordinary earthquake. The preceding few-hour-long transient (whose seismic moment history is quasi-exponential) is fully consistent with 1D diffusive slip propagation along the shallower part of the earthquake causative fault. The propagation path ends where later afterslip probably occurred. It is known that faults can sustain ruptures over a wide range of time scales (seconds to years), but the underlying physical mechanism is not yet clear. Slow fault slips (slow earthquakes) were firstly observed by high-sensitivity borehole dilatometers in Japan and USA, but the first observation of a swarm of slow earthquakes was performed by the Gran Sasso geodetic interferometers in 1997. Their probable source was few kms South of the interferometers. At that time. we suggested that a diffusive slip propagation (similar to heat transfer across a slab) could explain the scaling law between the seismic moment and rise time of those slow events. Recordings after the 2009 L'Aquila earthquake give the first direct evidence of the role and details of diffusive slow rupture propagation.
The April 6, 2009, Mw 6.3 L’Aquila earthquake (central Italy): pre - earthquake, co- and post- seismic deformation from GPS measurements
Anzidei M. 1,Cannelli V1., Devoti R. 1, Galvani A. 1, Esposito A. 1, Melini D. 1, Pietrantonio G. 1, Riguzzi F. 1, Sepe V. 1, Serpelloni E. 1
1 Istituto Nazionale di Geofisica e Vulcanologia – Centro Nazionale Terremoti
After the 1997 Umbria-Marche seismic sequence, a closely spaced GPS network (Central Apennine Geodetic Network - CAGeoNet) was set up across the central Apennines and Abruzzi region. This network was repeatedly surveyed to increase the spatial resolution of crustal strain measurements provided by the available Continuous GPS (CGPS) stations operating in this seismically active region. Position time series of permanent and non permanent GPS stations for the 1999-2007 time span have evidenced slow and purely elastic strain rates (~10-8 yr-1) around the L’Aquila basin. Higher values are observed only in the northernmost areas bordering the Umbria-Marche Apennine region, characterized by recent moderate seismicity. On April 6, 2009, the city of L’Aquila was struck by a Mw 6.3 earthquake that killed 307 people, causing severe destruction and ground cracks in a wide area around the epicenter. Four days before the main shock we have integrated the existing CGPS network with five GPS stations of the CAGeoNet bordering the L’Aquila basin, so we have been able to measure the entire earthquake sequence, also through high frequency GPS observations. We detect maximum horizontal and vertical coseismic surface displacements at the geodetic stations of 10.39±0.45 cm and -15.64±1.55 cm, respectively. We use geodetic data and seismological parameters to estimate the source geometry. Our best fitting fault model is a 13x15.7 km2 rectangular fault, SW dipping at 55.3±1.8°, consistent with the position of observed surface ruptures. The estimated uniform slip (495±29 mm) corresponds to the seismological observation of a 6.3 moment magnitude. Moreover, high frequency GPS data, collected at a set of stations located in the near field, provided the first recording of coseismic ground motion at 1Hz and 10Hz. The GPS data collected continuously during the seismic sequence at about 40 stations, allowed to estimate the post seismic deformation in a wide region, providing high resolution recordings of a post seismic signal across a normal fault.
The Mw 6.3, 2009 L’Aquila earthquake: source, path and site effects from spectral analysis of strong motion data
D. Bindi, F. Pacor, L. Luzi, M. Massa and G. Ameri
INGV Milano – Pavia, via Bassini 15, 20133 Milano, [email protected]
The strong motion data of April 6, 2009 L’Aquila (Central Italy) earthquake (Mw=6.3) and of 12 aftershocks (4.1≤Mw≤5.6) recorded by 56 stations of the Italian strong motion network are spectrally analyzed to estimate the source parameters, the seismic attenuation, and the site amplification effects. The obtained source spectra for S-wave have stress drop values ranging from 2.4 to 16.8 MPa, being the stress drop of the main shock equal to 9.2MPa. The spectral curves describing the attenuation with distance show the presence of shoulders and bumps, mainly around 50 and 150km, as consequence of significant reflected and refracted arrivals from crustal interfaces. The attenuation in the first 50 km is well described by a quality factor equal to Q(f)=59f 0.56 obtained by fixing the geometrical spreading exponent to 1. Finally, the horizontal-to-vertical spectral ratio provides unreliable estimates of local site effects for those stations showing large amplifications over the vertical component of motion.
EDURISK III – Un progetto formativo sul rischio fra pianificazione e emergenza
Camassi R. 1, Peruzza L.
1.Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna, Bologna, [email protected] 2.Istituto Nazionale di Oceanografia e Geofisica Sperimentale, Sgonico Trieste, [email protected]
Con la presente Convenzione DPC-INGV 2007-09, il progetto EDURISK è arrivato al terzo ciclo di progettazione pluriennale, e presumibilmente al suo epilogo. Le risorse finanziarie sono in progressiva contrazione, e profonde incertezze da alcuni anni attraversano la scuola, rendendo complesso, se non del tutto impossibile intraprendere qualsiasi iniziativa formativa non curricolare; vi è la tendenza ricorrente da parte delle Istituzioni coinvolte a privilegiare iniziative in base alla loro visibilità e non tanto alla loro efficacia; tutto ciò rende difficile ipotizzare che una iniziativa così impegnativa e complessa come un progetto formativo sul rischio possa trovare in futuro un sostegno convinto da parte degli interlocutori più diretti (Ministero dell’Istruzione e DPC). A prescindere dagli esiti in prospettiva che avrà questo progetto, è comunque possibile oggi fare un bilancio di questa esperienza, che si è confrontata negli anni con situazioni molto diverse, fino ad affrontare direttamente una situazione di drammatica emergenza. Presenteremo brevemente quindi quanto realizzato nel corso degli ultimi anni, e in particolare a L’Aquila prima e dopo il terremoto del 6 Aprile 2009, per cercare di rispondere seriamente ad una domanda cruciale: la comunità scientifica (sismologica, in questo caso) ha fatto davvero tutto quello che era in suo potere per stimolare scelte orientate alla riduzione del rischio? Un breve spazio, lasciato aperto al dibattito, sarà dedicato a rispondere ad un’altra domanda: cosa possiamo fare oggi noi, cittadini ricercatori, per far sì che i risultati migliori delle nostre ricerche possano contribuire a cambiare le cose?
Detection of surface movements during the Abruzzo 2009 earthquake by GNSS techniques. (Special session L’Aquila Earthquake) Alessandro Caporali1
1.Department of Geosciences, University of Padova [email protected]
The April 6, 2009, Mw=6.3 earthquake in Abruzzo (Central Italy) is associated with displacements on the surface which are well visible in the time series of permanent GPS stations within few tens of km from the epicentre. By combining the observed coseismic 3D displacements with an elastic dislocation model we obtain estimates for the position and orientation of the fault plane, and the slip vector at depth. The results are compared with similar results obtained by GPS, but a partially different set of stations, InSAR with Envisat and Cosmos SkyMed data, and classical seismological techniques of hypocenter location and fault plane solution, finding in most cases an excellent agreement. The analysis of the GPS time series has additional information not yet exploited: several signals are visible in the time series but are not associated with seismic events. We investigate a number of these features at high time resolution (30 sec) using the kinematic mode of the Bernese software. Thus the Abruzzo earthquake offers an interesting example of how space borne techniques (InSAR, GNSS) can monitor creep and transient deformation leading to and following brittle failure, complementing the data obtained by classical seismological techniques.
Coseismic and initial postseismic slip of the 2009 Mw 6.3 l’Aquila earthquake, Italy, from GPS measurements
D. Cheloni1, N. D’Agostino1, E. D’Anastasio1, A. Avallone1, S. Mantenuto2, R. Giuliani3, M. Mattone3, S. Calcaterra4, P. Gambino4, D. Dominici5, F. Radicioni6 and F. Fastellini6
1 INGV, Centro Nazionale Terremoti, Via Vigna Murata 165, 00143 Roma, Italy 2 Leica GeosystemS, Roma Italy 3 ISPRA, Roma Italy 4 DPC-SISM, Roma, Italy 5 Dip. Arch. Urb. Fac. Ing.,Università degli Studi dell’Aquila 6 Dip. Ing. Civ. Amb, Fac. Ing.,Università degli Studi di Perugia e-mail: [email protected]
Here we report the preliminary results of GPS data inversions for coseismic and initial afterslip distributions of the Mw 6.3 2009 April 6th L’Aquila earthquake. Coseismic displacements of continuous and survey-style GPS sites, show that the earthquake ruptured a planar SW-dipping normal fault with ~0.6 m average slip and a seismic moment of 3.5x1018 Nm. Geodetic data agree with the seismological and geological information pointing out the Paganica fault, as the causative structure of the main shock. Inversions for uniform and variable slip point out that the fault ruptured with a marked southeastward directivity and that the main coseismic asperity probably ended downdip of the Paganica village at a depth of ~1km in agreement with the small (1-10 cm) observed surface breaks. Time-dependent postseismic displacements have been modelled with a logarithmic function with a characteristic time of 2.7 days. The comparison between coseismic slip and postseismic displacements for the first twenty days after the main shock, shows that afterslip occurred at the edges of the main coseismic asperity with a maximum slip of ~10 cm and an equivalent seismic moment of 3.9x1017 Nm. The activation of the Paganica fault, spatially intermediate between the previously recognized main active fault systems, suggest that strain accumulation in the central Apeninnes may be simultaneously active on distinct parallel fault systems.
The 2009 L'Aquila seismic sequence (Central Italy): fault system geometry and Kinematics.
Lauro Chiaraluce1, Luisa Valorso1, Raffaele Di Stefano1, Davide Piccinini2.
1 Centro Nazionale Terremoti – Istituto Nazionale di Geofisica e Vulcanologia; Via di Vigna Murata 605 – Roma ([email protected]; [email protected]; [email protected]) 2 Sismologia e Tettonofisica – Istituto Nazionale di Geofisica e Vulcanologia;Via di Vigna Murata 605 – Roma ([email protected])
1. Centro Nazionale Terremoti, Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy. 2. Sismologia e Tettonofisica, Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy.
On April 6 (01:32 UTC) 2009 a destructive Mw 6.3 earthquake struck the Abruzzi region in Central Italy. We use aftershocks distribution to investigate the geometry of the activated fault system and to report on spatio-temporal seismicity pattern and kinematics of the whole seismic sequence. Seismic data were recorded at both permanent stations of the Centralized Italian National Seismic Network and 45 temporary stations installed in the epicentral area. To manage such a large amount of earthquakes, we implemented a semi-automatic procedure able to identify local earthquakes and to provide consistently weighted P- and S-wave arrival times. This procedure yields consistent earthquake detection and high- quality arrival times data for hundreds of events per day. The accurate location for thousands of aftershocks defines a complex, 40 km long, NW-trending normal fault system, with seismicity nucleating within the upper 12 km of the crust. We show the geometry of two major SW- dipping normal faults that form a right lateral en-echelon system. The main fault activated by the 6th of April earthquake is 20 km-long, NW-trending and about 50° SW-dipping and is located below the city of L'Aquila. To the north, we find a second fault, activated on the 9th of April by a MW 5.4 earthquake, that is about 12-km-long and shows a dip angle of about 40° with hypocenters mainly located in the 6 to 10 km depth range.
The Rupture History of the 2009 L’Aquila earthquake by non-linear joint inversion of strong motion and GPS data
Cirella A.1, A. Piatanesi1, M. Cocco1, E. Tinti1, L. Scognamiglio1, A. Michelini1, A. Lomax2
1Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy 2ALomax Scientific, Mouans-Sartoux, France.
The 2009 L’Aquila earthquake (Mw 6.3) occurred in the Central Apennines (Italy) on April 6th at the 01:32 UTC and caused nearly 300 casualties and heavy damages in the L’Aquila town and in several villages nearby. The main shock ruptured a normal fault striking along the Apennine axis and dipping at nearly 50° to the SW. Most of the aftershocks are also associated with normal faulting, which is consistent with the present-day tectonic setting of this sector of the Apennines. The identification of the fault geometry of the L’Aquila main shock relies on the aftershock pattern, the SAR interferometric data, the GPS displacements as well as the induced surface breakages. The earthquake allowed for the collection of an excellent data set for normal faulting events. In this study, we have imaged the rupture history of the L’Aquila earthquake using a nonlinear joint inversion of strong motion and GPS data. Our results show that slip distribution on the fault is quite heterogeneous and characterized by a small shallow slip patch located up-dip the hypocenter and a large and deeper patch located southeastward. The rupture velocity is larger in the up-dip than in the along-strike direction. This difference can be partially accounted by the local crustal structure, which is characterized by a high velocity layer above the hypocenter (9.46 km) and a lower velocity below. The latter velocity seems to have affected the along strike propagation since the largest slip patch is located at depths between 9 and 14 km. The imaged slip distribution correlates well with the on-fault aftershock pattern as well as with mapped surface breakages. The rupture history is also consistent with the large PGA values recorded at L’Aquila that is located right above the hypocenter.
UR5 UNI-BAS: Preliminary study on dynamic response of accelerometer housings and their influence on strong-motion data recordings.
R. Ditommaso, M. Mucciarelli
Department of Structures, Geotechnics, Applied Geology, University of Basilicata, Potenza, Italy, [email protected], [email protected]
The time-histories recorded at stations placed within or near buildings could be not representative of the true free-field ground motions: these could be contaminated by the presence of the buildings. These effects are known for years for large structures, but little attentions have been paid for small structures. Most of the accelerometric stations of Accelerometric National Network (RAN) are located within the electric grid sub-stations housing. The main types of housings are masonry, reinforced concrete and pre-cast concrete housing. Generally, they may have one or two floors. The housings were built during different periods, so with different codes and different construction details, but they are united by having a relatively small size. Usually volume is smaller than the volume of a small building for residential purposes. Recent studies about the influence of the dynamic response of those housings on the accelerometric recordings using the rotational HVSR technique showed that this could be a problem, revealing a strong contamination of recordings by the dynamic response of the housing. Starting from these interesting results, for the station located in Potenza (PTZ Station Code), in order to understand the influence of small structures on accelerometric recordings, after the L'Aquila Earthquake (6th April 2009), we analyzed the main recordings and performed the dynamic characterization of some housings (and the related pillars fixed into the ground on which accelerometer is located). All considered housings are located close to the epicenter: Bussi (BSS code), San Demetrio nè Vestini (SDM), Scafa (SCF) and Cittaducale (CTD). Possible influence of the housing in the high-frequency range is analysed and discussed.
Shakemaps of the L'Aquila main shock
Faenza L. 1, Lauciani V. 1, Michelini A.1 1.Istituto Nazionale di Geofisica e Vulcanologia, CNT, Via di Vigna Murata 605, 00143, Roma, Italy, [email protected], [email protected], [email protected] This work addresses the determination of the shakemap of the l’Aquila, M6.3 April 6, 2009, main shock. Since 2006 and as part of national projects funded by the Italian Civil Protection and by the EU SAFER project, INGV has been determining shakemaps for M3.0+ using the USGS-ShakeMap software package and a fully automatic procedure, based on manually revised location and magnitude. This work summarizes how the shakemaps of the main shocks have been obtained. Focus of the presentation is on the importance that the data and the extent of the finite fault have in the determination of faithful ground motion maps. For the L'Aquila main shock, we have found that the data alone are not sufficient to replicate the observed ground motion in parts of the strongly affected areas. In particular, since the station coverage toward the SE where the earthquake rupture propagated is scantier, prompt availability of a rupture fault model would have been important to better describe the level of strong ground motion throughout the affected area. We present an overview of the performance of the INGV real time system during the L’Aquila main shock - the first time that INGV provides real time information to Civil Protection during a seismic crisis. Finally, we show a comparison between the intensities determined from the strong ground motion and those obtained from the macroseismic survey.
The April 2009 L'Aquila earthquake as seen by a 250 km distant accelerometer array
N. Maercklin1, A. Zollo1, A. Orefice1, G. Festa1, A. Emolo1, R. De Matteis2, B. Delouis3, A. Bobbio4
1 Department of Physics, University of Naples Federico II, Naples, Italy ([email protected], [email protected], [email protected], [email protected], [email protected]) 2 Department of Geology and Environment, University of Sannio, Benevento, Italy ([email protected]) 3 Geosciences Azur, University of Nice Sophia Antipolis, Nice, France ([email protected]) 4 Osservatorio Vesuviano, INGV, Naples, Italy ([email protected])
The Mw 6.3 L'Aquila earthquake on April 6, 2009 has been recorded by the Irpinia Seismic Network (ISNet) about 250 km SE of the epicenter. The ISNet array has an aperture of about 80 km and consists of 25 stations, each equipped with a 3C CMG-5T accelerometer. Waveforms from 19 stations could be used to estimate rupture geometry, event magnitude, and moment tensor. Standard array methods applied to low-pass filtered waveforms provided backazimuth and slowness of the incoming waves for subsequent event location. To image the rupture geometry we implemented a modified beamforming technique that back-projects the recorded direct P-wave amplitudes into the earthquake source region. A NW-SE striking rupture of 17 km length is imaged, propagating with an average velocity up to 3 km/s. We inverted P- and S-wave displacement spectra to determine the main source parameters, assuming an omega-square spectral model and a constant-Q attenuation function. The estimated seismic moment (2.1 x 1018 Nm) is consistent for P- and S-waves and yields a maximum Mw of 6.1. Relatively high P- and S-corner frequencies (0.6 and 0.4 Hz), due to directivity effects, lead to an underestimation of the rupture length and a high static stress drop value. However, an apparent stress of about 1.5 Mpa is measured from the radiated seismic energy, which is less sensitive to directivity effects. We determine the moment tensor solution for the earthquake by waveform modelling. One approach uses a point source approximation and a grid search over a set of trial source positions to identify the optimal centroid position, time, and moment tensor. In another approach the rupture is represented by a finite 1D source model, approximated by a summation over point sources along the fault strike. The focal mechanism and the linear seismic moment distribution along the strike are inverted simultaneously by a grid search combined with simulated annealing. We find a centroid depth of about 5 km and a prevalently normal fault plane solution with a dominant directivity toward SE. Our studies demonstrate that the use of array techniques and a dense accelerometer network can provide quick and robust estimations of source parameters of moderate-size earthquakes located outside the network. Focal mechanisms of 171 events of the Aquila sequence ( M ≥ 3), and computation of site effects at the stations of the INGV National Seismic Network in Central Apennines
Luca Malagnini1, Robert B. Herrmann2, and Irene Munafo’1 and Aybige Akinci1 1Istituto Nazionale di Geofisica e Vulcanologia 2Department of Earth and Atmospheric Sciences of Saint Louis University
Between October 1st, 2008, and September 24th, 2009, the National Seismic Network recorded 171 earthquakes with 3.0 Real-time forecasting following a damaging earthquake Marzocchi W.1,Lombardi A.M.1 1Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Roma, Italy, [email protected] We describe the results of the first prospective, real-time earthquake forecast experiment made during a seismic emergency. A Mw6.3 earthquake struck the city of L'Aquila, Italy on April 6, 2009, causing hundreds of deaths and vast damage. Immediately following this event, we began producing daily earthquake forecasts for the region, and we provided these forecasts to Civil Protection -- the agency responsible for managing the emergency. The forecasts are based on a stochastic model that combines the Gutenberg-Richter distribution of earthquake magnitudes and power-law decay in space and time of triggered earthquakes. The results from the first month following the L'Aquila earthquake exhibit a good fit between forecasts and observations, indicating that accurate earthquake forecasting is now a realistic goal. Our experience with this experiment demonstrates an urgent need for a connection between probabilistic forecasts and decision-making in order to establish -- before crises -- quantitative and transparent protocols for decision support. The role of structural geomorphological mapping in seismic hazard prevention: results from DPC Projects after the L’Aquila earthquake: Miccadei E.1, Cavinato G.P.2, Farbollini P.3, Piacentini T.1, 1.Università degli Studi “G. d’Annunzio” di Chieti Pescara, Via dei Vestini 31, 66100 Chieti scalo (CH) - ITALY, [email protected] 2. IGAG - Istituto di Geologia Ambientale e Geoingegneria, CNR, Rome, Italy, P.le Aldo Moro, 5 - 00185 Rome - ITALY, [email protected] X.Università degli Studi di Camerino, Via Gentile III Da Varano 62032 Camerino - ITALY, [email protected] April 6th 2009 at 3:32 a.m., after some three month of low magnitude tremor, a strong earthquake (MI Richter 5,8) hits the L’Aquila area, with epicenter close to the Paganica village, at the base of the Gran Sasso ridge, and ipocenter at 8,8 km. The event is characterized by a extensional movement along NW-SE fault planes with SW-NE extension (INGV, 2009; USGS, 2009). The earthquake has involved an area more than 30 km long, between L’Aquila and the Aterno river valley: L’Aquila, Paganica, Tempera, Onna, and other villages are heavily damaged or completely destroyed. In the epicenter area (particularly Paganica, Tempera and Onna, where the buildings are heavily damaged) the occurrence of geomorphologic effects has been surveyed, such as rock fall, debris landslides, small landslide reactivations, very shallow fissures and fractures with variable strike on different continental deposits, fissures on road pavement, fluvial embankment local falls, soil liquefaction, and even small snow avalanches. All these features occurred in almost saturation condition in the Quaternary continental deposits fill of the L’Aquila basin. The surface effects observed does not agree with the expected effects for the earthquake magnitude according to the intensity scale ESI 2007. The present work wants emphasize the role of detail geomorphologic survey (from 1:10.000 to 1:5.000 scale) and Quaternary continental deposits analysis in the L’Aquila basin, as base of a correct seismic prevention where historical seismicity and technical rules does not succeed in reducing seismic risk. Appraisal of the hypocentral location of the L'Aquila main shock Michelini A.1, Faenza, L. 1, Lomax, A.2 1.Istituto Nazionale di Geofisica e Vulcanologia, via di vigna murata, 605, 00143 Roma, Italy, [email protected], [email protected] 2. ALomax Scientific, Mouans-Sartoux, France. [email protected] This work addresses the location of the hypocenter of the l’Aquila, M6.3 April 6, 2009, main shock. A thorough analysis of the earthquake focus and its uncertainties is carried out using both a linearized standard, and a grid search methodology for earthquake location. To assess the robustness of the location, different velocity models and independently picked phase data sets (2) are tested. All the resulting solutions are comprised within a radius of less than 1 km horizontally whereas depth depends on the velocity model used. The solutions obtained using only the 30 closest stations and a locally calibrated velocity model resulting from the application of the Velest procedure feature final RMS residual values close to 0.07 s or less. The depth of the hypocenter is close to 8.8 km with a largest semi-axis of 0.7 km for the1-σ ellipsoid . No significant change in the location (~150 m and ~200 m horizontally and vertically, respectively) is found when using the two data sets both adopting the 30 closest stations. The location difference, however, increases when additional stations are included. The initial energy release at the hypocenter is comparable to that of a M4 foreshock, which occurred a few days earlier. The first burst of larger seismic energy is picked on some of the available data - it is found to occur about 0.8-1.0 s from OT, it locates up-dip (~2 km) from the initial nucleation and it is shifted the E-NE. This location is consistent with a ~50o SW dipping plane for the main rupture. Continuous waveform data stream analysis: Detection and location of the L'Aquila earthquake sequence Michelini A.1, Maggi, A.2 1.Istituto Nazionale di Geofisica e Vulcanologia, via di vigna murata, 605, 00143 Roma, Italy, [email protected], [email protected] 2. Ecole et Observatoire des Sciences de la Terre, 5 rue Rene Descartes, F-67084 Strasbourg Cedex, France, [email protected] We present a continuous waveform earthquake location technique which does not rely on wave onset phase picking or on event phase association. The technique adopts and further develops a methodology introduced by Withers (1999) based on the cross- correlation between ''Green's functions'' within a predefined distance range, and stream- like, continuously recorded data. At each time step, earthquake occurrence is tested through the sum of the cross-correlation values on a grid of potential location points. In real time processing and in absence of station latencies, the technique, when applied at regional and local scales, can provide locations within 10-30 seconds from origin time depending on the pre-assigned distance of the Green's functions and station coverage. In addition the technique can be used off-line to analyse previously recorded data. The methodology is applied to data recorded by the permanent Italian National Seismic Network operated by INGV. The resulting epicentral locations are generally found to match closely those determined using manual picks and standard location programs. In addition we detect a number of events that were missed by the standard manual location procedure. The technique shows potential for unattended and rapid seismic source detection/location in seismic monitoring centers. References Withers, M.; Aster, R. & Young, C. (1999), 'An automated local and regional seismic event detection and location system using waveform correlation', Bulletin of the Seismological Society of America 89(3), 657--669. Seismic site response in L’Aquila city center and in suburban areas. Milana G.1, Azzara R.M.2, Bergamaschi F. 2, Bertrand E. 3, Bordoni P. 1, Cara F. 1, Cogliano R. 4, Cultrera G. 1, Di Giulio G. 1, Duval A. 3, Fodarella A. 4, Marcucci S. 5, Pucillo S. 4, Régnier J. 3, Riccio G. 4. 1 Istituto Nazionale di Geofisica e Vulcanologia Sezione ROMA1,Via di Vigna Murata 605 - 00143 Roma, Italy (email: [email protected], [email protected], [email protected], [email protected], [email protected]) 2 Istituto Nazionale di Geofisica e Vulcanologia Osservatorio Sismologico Arezzo, Via Uguccione della Faggiuola 3 - 52100 Arezzo (email: [email protected], [email protected]) 3 LRPC de Nice, CETE Mediterranee, Nice Cedex 4, France (email: [email protected], anne-marie.duval@developpement- durable.gouv.fr, [email protected]) 4 Istituto Nazionale di Geofisica e Vulcanologia Sede Irpina, 83035 Grottaminarda (Av) (email: [email protected], [email protected], [email protected], [email protected]) 5Dipartimento della Protezione Civile, Via Vitorchiano 2 - 00189 Roma (email: [email protected]) Following the 6th of April 2009 Mw 6.3 L’Aquila earthquake, many scientific institutions were involved by Italian Civil Defense (DPC) to perform microzonation studies in the epicentral area. The INGV-RM1 group was in charge of seismic monitoring of downtown L’Aquila and its western suburbs extending towards NW up to Cansatessa. The extension of the areas and the peculiarity of L’Aquila old city center required an important effort both in terms of number of instruments and operators for the installation, maintenance and data collection and analysis. The working group included researchers from INGV, Ceté Mediterranee (Nice, France) and DPC. The installation started on May 27 in L’Aquila old town area with the deployment of 22 seismic instruments (12 from INGV, 7 from Cete Mediterranee and 3 from DPC). In the following days 3 extra instruments from INGV were deployed. Due to the seismicity rate it was possible to record a high number of events and to move some stations in different locations. The monitoring network was removed on July 3 except for two stations that are still running. Overall we monitored 37 different sites recordings more than 300 events with a good signal to noise ratio to be used for spectral ratio analysis. Some microtremor measurements were performed to extend the analysis to more than 50 sites all around the area. Data were analyzed using both the standard spectral ratio technique on earthquakes and seismic noise, and different reference sites were selected. The first results obtained confirm the presence of large amplification effects both in L’Aquila historic center and in the suburban areas. Downtown L’Aquila is characterized by a strong low frequency (around 0.6 Hz) amplification on horizontal components and a clear effect also on the vertical component with a frequency shifted towards higher frequencies (1-1.5 Hz). The suburban areas shows strong amplification at frequencies ranging from 2 to 4-5 Hz. The observation can be related to the available geological and geophysical data collected as a part of microzonation activities. The obtained results can also be used to define the site response of AQK and AQU strong motion stations in the framework of DPC-S4 project. These stations recorded all the main events of the sequence and are close to some of the monitored sites. Contributions from Project S4 to the investigations on the L’Aquila earthquake Pacor F.1 Paolucci R2. and S4 RUs 1Istituto Nazionale di Geofisica e Vulcanologia, Milano,Italy, [email protected] 2Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] The occurrence of the L’Aquila earthquake, has promoted within Project S4 several activities, stimulating some reflections on the way the Project itself should move on to comply with the requirements coming from DPC and from the national and international scientific and technical community as well. As a first major contribution, strictly related to the main objective of the Project, a great deal of efforts was spent to publish in ITACA, within a short amount of time from the earthquake occurrence, the strong motion data from the mainshock and the 12 largest events (M > 4) of the sequence. About 900 waveforms were included, both in the uncorrected and corrected version, recorded by more than 65 digital stations. After publication of records, the ITACA web site performed well when facing a dramatic increase of the number of accesses and downloads, around 150-200 visits/day, mainly from Italy but also from worldwide, with a positive fall-out in terms of the national and international visibility of ITACA. Besides, all the RUs of Project S4 were directly involved in the operational activities in the aftermath of the earthquake. This presentation aims at summarizing such activities, namely: RU1 and RU3 contributed to analyse the characteristics of different strong-motion parameters as function of distance, azimuth and site conditions. Furthermore RU1 analyzed the strong-motion data set with the aim of evaluating source, path and site spectral parameters. RU2 and RU6 carried out a geological survey at the accelerometric stations recording the Abruzzo main event. On the base of litological map and other available geophysical and geological information, a site classification of these stations was proposed, based on the EC8 classes. RU4, RU8 and RU7 applied active and passive surface wave methods for site characterization at accelerometric stations located in epicentral area, such as the AQP and AQA permanent stations, and Bazzano and Onna sites, where recording stations were installed after the mainshock. Insight on the response site at AQK station, located close to L'Aquila downtown, have been provided by the monitoring activity carried out by RU2 in the framework of microzonation study of L’Aquila town. Finally RU8 together with RU5 installed several strong motion arrays to monitor the seismic response of buildings during the seismic sequence. Furthermore, RU5 performed noise measurements at ENEL cabins where accelerometric station were installed to be compared with strong motion data in order to evaluate the interaction with hosting structures. Overview of the GFZ Task Force activities in the Atorno valley Picozzi M.1, Parolai S. 2, Bindi D.3, Pilz M.4, Di Giacomo D.5, Milkereit C.6, Sobiesiak M.7 1) Helmholtzzentrum Potsdam – German Research Center for Geosciences, Sek. 2.1 Telegrafenberg, 14473 Potsdam, Germany, [email protected] 2) Helmholtzzentrum Potsdam – German Research Center for Geosciences, Sek. 2.1 Telegrafenberg, 14473 Potsdam, Germany, [email protected] 3) Istituto Nazionale di Geofisica e Vulcanologia, Via Bassini 15, 20133 Milano, Italy 4) Helmholtzzentrum Potsdam – German Research Center for Geosciences, Sek. 2.1 Telegrafenberg, 14473 Potsdam, Germany, [email protected] Universität Potsdam, Institut für Geowissenschaften, Karl-Liebknecht-Str. 24, 14476 Potsdam, Germany 5) Helmholtzzentrum Potsdam – German Research Center for Geosciences, Sek. 2.1 Telegrafenberg, 14473 Potsdam, Germany, [email protected] Universität Potsdam, Institut für Geowissenschaften, Karl-Liebknecht-Str. 24, 14476 Potsdam, Germany 6) Helmholtzzentrum Potsdam – German Research Center for Geosciences, Sek. 2.1 Telegrafenberg, 14473 Potsdam, Germany, [email protected] 7) Helmholtzzentrum Potsdam – German Research Center for Geosciences, Sek. 2.1 Telegrafenberg, 14473 Potsdam, Germany, [email protected] On April 6, 2009, a Mw 6.3 normal faulting event struck the Abruzzo region close to the city of L´Aquila, central Italy, causing many casualties and serious damage. Soon after the main shock, a dense network of 7 velocimetric and 11 accelerometric stations was installed in the epicentral area to the south-east of the city of L´Aquila by the German Task Force for earthquakes with the aim of studying possible site effects and the response of the buildings. Additionally, we performed 2D array measurements nearby the village of Onna. These measurements were carried out using 17 seismic stations. The aperture of the array was of the order of a few hundred meters. We present an overview of the activities with special focus on site effect investigation and building monitoring. To this regard, we selected several events including also the strong aftershock occurring on April 9 (Mw 5.4). In general, significant site effects are found for some settlements in the middle portion of the Aterno valley, located both on recent soft alluvial soil and on mid-Pleistocene silts. The high ground-motion amplifications partially explain the observed macroseismic intensity values (IX on MCS scale). Nearby villages, only a few kilometers apart, show lower amplification effects in agreement with the observed level of damage (VI to VII MCS). For the building of the municipality in Navelli, monitored in cooperation with the University of Basilicata, clear variation of resonance frequency was observed during the strongest aftershocks. Seismic moment tensors of the April 2009, L'Aquila (Central Italy), earthquake sequence Silvia Pondrelli1, Simone Salimbeni1, Andrea Morelli1, Goran Ekstrom4, Marco Olivieri2, Enzo Boschi3. 1. Sezione di Bologna, Istituto Nazionale di Geofisica e Vulcanologia, Bologna, Italy. 2. CNT, INGV, Rome, Italy. 3. INGV, Rome, Italy. 4. Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY, USA. On April 6, 2009, the Central Apennines were hit by a Mw=6.3 earthquake. The region had been shaken since October 2008 by seismic activity that culminated in two foreshocks with Mw>4, one week and a few hours before the main shock. We computed seismic moment tensors for 26 events with Mw between 3.9 and 5.5, using the Regional Centroid Moment Tensor (RCMT) scheme. Most of these source parameters have been computed within one hour after the earthquake and rapidly revised successively. The focal mechanisms are all extensional, with a variable and sometimes significant strike-slip component. This geometry agrees with the NE-SW extensional deformation of the Apennines, known from previous seismic and geodetic observations. Events group into three clusters. Those located in the southern area have larger centroid depths and a wider distribution of T-axis directions. These differences suggest that towards south a different fault system was activated with respect to the SW-dipping normal faults beneath L'Aquila and more to the north. These considerations and other analyses have been possible for the rapid availability of these robust moment tensors. Fast Determination of Moment Tensors and Rupture History: Application to the April 6th 2009, L’Aquila Earthquake Laura Scognamiglio, Elisa Tinti, Alberto Michelini, Antonella Cirella, Massimo Cocco, and Alessio Piatanesi On April 6th 2009, a magnitude Mw=6.3 earthquake struck the Abruzzi region in central Italy. Despite its moderate size, the earthquake caused 293 fatalities and partially destroyed the city of L’Aquila and many villages in its surroundings. The main shock was preceded by an earthquake swarm, which started at the end of 2008. The largest earthquakes of the swarm occurred on 2009/03/30 (ML=4.1), and on 2009/04/05 (ML=3.9). To date, almost 7,000 aftershocks with ML>1.5 have been recorded by the INGV seismic network and three featured ML larger than 5.0. In this study, we present the results of the fast source parameters determination procedure adopted at the Istituto Nazionale di Geofisica e Vulcanologia (INGV) using the 2009 L’Aquila earthquake as a case study. The main task of this procedure is the fast calculation of source parameters within the first 24 hours after an earthquake. We apply a time domain moment tensor (TDMT) technique to compute the focal mechanisms of all the ML≥ 3.9 earthquakes by inverting broadband records of the Italian national seismic network. All events show normal faulting in agreement with the tectonic setting of the area. The preferred main shock moment tensor solution inferred is: strike 139°, dip 48°, rake - 77° and Mw= 6.1. The resulting moment magnitude seems to underestimate the value released by CMT (Mw 6.3) and we investigate this issue thoroughly using the regional scale MT-technique of C. J. Ammon et al. (1990) and R. Hermann and C. J. Ammon (2002). Using the main shock moment tensor to constrain the fault geometry, we invert the strong motion data provided by the Rete Accelerometrica Nazionale (RAN) and the MedNet station AQU to image the rupture history. The inferred model is representative of a rapid finite-fault solution to be used immediately after an earthquake to get a preliminary interpretation of ground shaking. The proposed rupture history highlights several relevant features. First, we have identified the SW dipping plane as the main shock rupture plane and the existence of rupture directivity associated with both the up-dip and SE along-strike propagation. Second, the inferred rupture velocity, constant all over the fault plane, is relatively low (2.2 km/s). Third, the rapid solution is able to identify the position on the fault plane where most of the energy is radiated, although the position and the amplitude of slip patches are not well constrained. Session Geological Input Progetto S1 - U.R 3.11: Cinematica recente ed attuale dei fronti nord-appenninici e dell’avampaese padano – Recent and present-day kinematics along the Northern Apennine arcs and triangle zone between Southern Alps and Emilian arc. Seno S. 1,Bonini L. 1, Burrato P. 2, Caputo R. 3, Di Bucci D. 4, Ahmad I. 1, Lucchesi S. 2, Toscani G. 1 1- Dipartimento di Scienze della Terra, Università di Pavia. Via Ferrata 1 - 27100 Pavia, Italy. 2- Istituto Nazionale di Geofisica e Vulcanologia. Via di Vigna Murata, 605 - 00143 Roma, Italy 3- Dipartimento di Scienze della Terra, Università degli Studi di Ferrara. Via Saragat, 1 – 44100 Ferrara, Italy 4- Dipartimento della Protezione Civile, Ufficio Rischio Sismico. Via Vitorchiano, 4 - 00189 Roma,Italy An overview of a study on recent and present-day kinematics along the Emilia and Ferrara arc of folds and the triangle zone between Southern Alps and Emilian Arc (Northern Apennines) will be presented, based on the integration of different geological datasets. Our work is aimed at studying how the deformation is partitioned among the different thrust fronts in order to (i) better understand the buried structural setting of the Po Plain and (ii) characterize the principal thrust faults in terms of seismotectonic parameters, paying particular attention to those which can be the source of future damaging earthquakes. Some regional reflection seismic sections have been considered; some of them have been interpreted, depth-converted (in two cases) and restored (in one case), the last one have been re-interpreted in the surficial reflectors (Plio-Pleistocene) in order define the structural features of the Northern Apennines fronts, of the central Po Plain buried structures and of the outermost Southern Alps fronts. They have been imported in a georeferred 3D model and published: new data have been homogenised to provide insights and a global view of the Po Plain subsurface architecture. To constrain the recent to present-day activity of the described fronts, over the restoration, we realized a detailed morphotectonic analysis, focused on the surface expression of the buried thrusts of the Emilian and Ferrara Arc. This analysis has been carried out by quantitatively comparing maps of terraces and associated Upper Pleistocene-Holocene deposits with drainage anomalies of the main rivers. The integration of these data with the regional cross sections, the reconstruction of buried, recent stratigraphic horizons (e.g. the main Upper Quaternary sedimentary bodies hosting the aquifers) and the structural maps of the major burid tectonic elements allow us to observe that some Holocene features of the mapped terraces and drainage pattern could be related to the mild but not negligible present-day activity of the buried thrust fronts. Finally, to reproduce in a simplified and scaled way the tectonic history of the surveyed structures, two different series of analogue models have been realized with two different setups. The first one aimed at reproducing the Emilian Arc fronts and their activity; the second one simulating two opposite verging thrust fronts with the aim of checking their possible mutual interference in recent times. The work presented is in progress and the results obtained up to now will be shown. Active faults and hystorical earthquakes in the Ionian Sea. Polonia A.1, Gasperini L.1, Mussoni P. 1,3, Riminucci F. 1,2, Torelli L. 3 1.ISMAR, CNR, Via Gobetti, 101, Bologna, [email protected] 2. Universita’ degli Studi di Bologna 3. Universita’ degli studi di Parma The Calabrian Arc (CA) is part of the most active seismic belt in Italy and it has been struck repeatedly by very strong historical earthquakes often associated with destructive tsunami. Although the regional architecture of the margin geometry has been described through the analysis of high penetration seismic data, some major questions remains unanswered: 1) is the Calabria subduction zone still active? 2) Where is the accretionary wedge outer deformation front located? 3) Which is the depth of main decollement level in the different domains of the subduction complex? 4) Which is the geometry and location of major active faults in the subduction complex? 5) What is the uplift and outward growth rates of the accretionary wedge? In order to address these issues we have analysed in detail the structure and the evolution of the external CA, through a multi-scale approach addressing tectonics, kinematics and mass balancing in the accretionary complexes as well as submarine earthquake geology. Geophysical data (MCS and CHIRP profiles) have been acquired in three key areas, selected in order to study active deformation across a) the outer deformation front of the accretionary wedge, b) the lateral western termination of the subdution complex at the foot of the Malta escarpement and c) the transition between the accretionary wedge and the backstop. Well targeted sediment samples have been collected in tectonically controlled sedimentary basins and across active fault in order to study the sedimentary structures that can be related to the seismic activity (sand injections, mass wasting deposits, homogenites, turbidites, etc.). Sediment cores from the Ionian abyssal plain (3800 m water depth) as well as other deep marine basins adjacent to inner deformation fronts have sampled turbidite sequences (at least 7 cycles including the Augias turbidite), which likely contain a record of the great earthquakes in the region. Radiocarbon ages will be obtained from monospecific planktonic foraminiferal samples above and beneath suspected seismic related deposits events, while Pb210 and Cs137 radiocarbon dating will be applied to selected cores to resolve seismites associated with some recent 19th and 20th century events. If we will be able to make correlations between the sedimentary record (of earthquake-triggered submarine landslides or turbidites) and the earthquake catalogue in Southern Italy, we will try to extend the paleoseismic catalog further back in time. Thus, it should be possible to determine which portions of the arc have experienced great earthquakes in the past, and whether the recurrence interval is regular or not, and how often catastrophic events occur in the different portions of the subduciton system. Geomorphic evidence and lateral variability of the Gondola Fault Zone in the south Adriatic Sea Ridente D.1, Di Bucci D.2, Trincardi F.3, Foglini F.3, Campiani E.3, Fracassi U.4, Valensise V4. 1IGAG-CNR, via Salaria km 29,300, Area della Ricerca di Roma 1, Roma, [email protected] 2Dipartimento della Protezione Civile, via Vitorchiano 4, Roma, 3 ISMAR-CNR, via Gobetti 101, Bologna 4INGV, via di Vigna Murata 605, Roma The Gondola Fault Zone is an inherited, poly-phased structure that dissects the southern Adriatic continental shelf and slope over a length of 140-150 km. Recent studies along the E-W striking portion of the Gondola Fault Zone dissecting the shelf area have unveiled clear evidence of activity during the Middle-Upper Pleistocene and Holocene, although with low slip rates (vertical slip rates up to 0.18 mm/a). Such evidence has been documented by means of very high resolution seismic data (Chirp Sonar and Sparker profiles) supported by diverse chronostratigraphic constraints that allowed detailed dating of the depositional sequences composing the middle-upper Pleistocene and Holocene succession. In some cases, active fault segments, in the order of 10-20 km length, rupture up to the Holocene seabed, locally determining a subtle fault-related morphology detectable on high resolution multibeam morpho-bathymetry. The geomorphic expression of the Gondola Fault is instead markedly evident along its NW-SE striking portion extending from the shelf edge to the base of slope, where the lower sedimentation and erosion rates prevented the obliteration of the tectonic relief. The analysis of the morpho- bathymetric relief along the NW-SE Gondola Fault, combined with evidence from seismic profiles, allowed the identification of all fault segments that offset the seafloor. Preliminary seismic correlation from sediment cores in the area indicate that Holocene sediments are clearly offset along this deformation, thus confirming that seafloor offsetting occurred during a recent interval; it is also clear that faulting offsets the sea floor at several locations and along segments greater than 4-5 km. This information is used for defining the distribution of active fault segments and the rates of deformation, in the light of attempting an evaluation of the state of activity, kinematics and possible seismogenic potential along the NW-SE portion of the structure. Finally, deformation along the Gondola Fault is likely in relation with significant slope failure affecting the shelf edge and slope in this sector of the Adriatic margin. Session Seismic Hazard A Time-Dependent Probabilistic Seismic Hazard Model For The Central Apennines, Italy 1Akinci, A., 2F. Galadini, 1D. Pantosti, 3M. Petersen, 1L. Malagnini and 3D. Perkins 1Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Rome, Italy, Tel.: +39 06 51860403, [email protected] 2, Istituto Nazionale di Geofisica e Vulcanologia, Sezione Milano-Pavia, Milano, Italy 3U. S. Geological Survey, MS 966, Box 25046, Denver, CO 80225 USA, e-mail: [email protected], [email protected] Abstract We constructed a PSHA model for the central Apennines based on the long-term recurrence behavior of active faults together with the spatial distribution of earthquake observed in historic time. We use the maximum magnitude model for the fault source together with a uniform probability of rupture along the fault (floating fault model) to mode fictitious faults to account for earthquakes that cannot be correlated with known geologi structural segmentation. The probabilities are obtained from time independent and tim dependent model characterized by Poissonian and a Brownian Passage Time recurrenc model with a aperiodicity ( ) of 0.3, 0.5 and 0.7. Earthquake hazard is quantified in terms o peak ground acceleration (PGA) and spectral accelerations (SA) for natural periods of 1. seconds. In order to determine the relative contribution from various sources at l’Aquila an Rome, we deaggregated the seismic hazard for SA1 and PGA, following the approac described by Harmsen and Frankel, (2001). We also selected three sites in whic deaggregation shows the changing role of the fault and background sources as get smaller. In general, time-dependent hazard is increased and the peaks appear to shift to th ESE of the Central Apennines with respect to the results of the Possionian source mode We observed that as Deaggregation results indicate that the hazard is highest near fault with the highest earthquakes rates as decreases (periodicity increases). This effect i strongest for the long-period (1 s) ground motions. RU02: Application of generalized GMPEs to PSHA and risk assessment Faccioli E.1, Villani M.2, Demartinos K.1, Vanini M.1, Stupazzini M.1 1 Department of Structural Engineering – Politecnico di Milano, Piazza L. da Vinci, 32 - 20133 Milano, [email protected] 2 Rose School, Via Ferrata, 1 - Pavia The datasets of most Ground Motion Predictive Equations (GMPEs) tend to be still insufficient in the near field of the source, despite a significant increase in recent years. To overcome this limit, explicit modelling of wave propagation at the site of interest can be implemented in the classical PSHA. With sufficient numerical simulations in which the fault rupture parameters are varied , suitable probability distributions for the shaking parameters and their first statistical moments (at each receiver) are estimated and used as an alternative to the GMPEs. Herein, an application to the Sulmona (Central Italy) sediment filled basin is described. Using the GEOELSE code, 3D deterministic simulations in which the hypocenter location was randomly chosen among approximately 200 mesh nodes inside the fault plane and the slip distribution among 28 possible patterns (from real fault inversions). The 3D numerical simulations have a band limited resolution (f < 2 Hz) and, hence, were integrated with an appropriate high frequency content beyond 2 Hz, in order to achieve synthetic motions over physically meaningful a wide frequency range, as required for the damage assessment applications. The high frequency portion was simulated using a stochastic point-source method applied to finite sources, as recently proposed (Atkinson and Boore, 2006; Motazedian and Atkinson, 2005). Ground motion contributions by subfaults, calculated by the stochastic approach, are suitably summed, accounting for fault dimensions and rupture process. The resulting acceleration time series are used in a probabilistic framework to account for the variability of the fault parameters, aimed at assessing seismic risk for classes of buildings. Seismic risk is expressed in terms of a probabilistic measure of damage, quantified through uncertain engineering criteria by means of the capacity spectrum method. To this end, a consistent probabilistic formulation has been developed and implemented. The formulation makes use of the numerically simulated records as a set of random realizations to calculate the desired probability through Monte-Carlo approximation. Double Branching model to forecast the next moderate-strong earthquakes in Italy Marzocchi W.1,Lombardi A.M.1 1Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Roma, Italy, [email protected] We propose a new forecasting model, named Double Branching, for large earthquakes in Italy. The model proposed is time-dependent, since it assumes that each earthquake can generate, or is correlated to, other earthquakes, through different physical mechanisms. By applying this model to the Italian historical seismicity of the last centuries we find that the time-dependent features are comparable to what observed in other applications for different time-space-magnitude window. Moreover, the comparison of the forecasting capability of the Double Branching model and of a spatially-variable stationary Poisson process (used in Italy for seismic hazard assessment) shows that the former has significantly better performances. Finally we provide some probability forecasting maps for different temporal windows. Remarkably, the model can be tested in a forward perspective, which is the most straightforward way to evaluate the reliability of any forecasting model. Moreover it is suitable for computing earthquake occurrence probability in real time and to take part to international initiatives as CSEP laboratory. CRISIS2008: an enhanced tool for advanced (but practical) Probabilistic Seismic Hazard Analysis Ordaz M.1, Meletti C.2, Martinelli F.2, D’Amico V.2, Venegas F.J.1, 1.Instituto de Ingenieria, UNAM, [email protected] 2.Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Milano-Pavia, [email protected] We present CRISIS2008, a suite of applications (desktop- and web-based), designed to perform probabilistic seismic hazard analysis (PSHA) that is advanced in some aspects, but preserves characteristics that make it useful for practical uses. CRISIS2008 preserves the general logic of all its CRISIS ancestors, but presents new features that are described in detail. Some of the most relevant are the following: Geometry: besides the old area, line and multi-point sources, the new version includes a multi-point source, designed to accommodate numerous sources (tens of thousands, if required), closely related with the new seismicity model that has been developed. Seismicity: besides the old Gutenberg-Richter and Characteristic Earthquake models, CRISIS2008 presents a generalized non-Poisson model, in which, for every point source, probabilities of occurrence of 1, 2,…, n events are explicitly given for various time-frames. Ground motion models: along with the old non-parametric ground motion models that have always been part of CRISIS, this new version includes a larger set of parametric, well- known ground motion models, and a new generalized model, intended to describe, in probabilistic terms, the ground motions computed by arbitrarily complex estimation models. Also, an abstract interface has been designed in order to make it very simple to add new ground motion models, without having to recompile the desktop or the web applications. General: due to the possible combination of Poisson and non-Poisson sources, hazard results are now given in terms of probabilities of exceedance of intensity values in a given time frame, and not in terms of intensity exceedance rates; results for disaggregated hazard can be obtained for M-R-epsilon combinations using a redesigned, friendlier interface; logic-tree capabilities have been improved. Web version: it is developed in order to provide some facilities, such us: the sharing of data between users (e.g. GMPEs implemented by a user, accessible to the others); the use of the last updated version; the independency from the operating system used by the application; the elaborations executed at the server site, so that the user computer is not busy in heavy computations. While CRISIS2008 preserves the open-source nature of all its ancestors, all major classes have been redesigned to facilitate their use outside our own development environment. A proof of this (a weak proof, perhaps) is that both our web and desktop applications use exactly the same classes. During the meeting in the PC area it will be possible to see the application running. Fault-based earthquake rupture forecast in Central Italy: lessons learned with the L’Aquila Mw 6.3 event Peruzza L.1, Pace B.2, Visini F.2 1.Istituto Nazionale di Oceanografia e Geofisica Sperimentale, Sgonico Trieste, [email protected] 2.Università “G. d’Annunzio” Chieti-Pescara, Chieti Scalo, [email protected], [email protected] In October 2008, Italy was made a testing region within the framework of the Collaboratory for the Study of Earthquake Predictability (CSEP) (http://eu.cseptesting.org/Meeting20081027). At that time, we decided to submit the LAyered Seismogenic Source model in Central Italy (LASSCI), developed in the early of 2000 and published in 2006, for validation. In accordance with the first deadline, we handled the original seismicity rates used for the time-dependent seismic hazard map in order to satisfy the rules and format given for mid-term earthquake forecast (5 and 10 years). When it became clear that the deadline was going to be postponed, we decided to go through the calculations and analyses, the bias of some months’ delay being negligible on our results; they hold on a desktop folder for a very short time, becoming obsolete before to be submitted. On April the 6th, at 3:32 a.m. local time, an Mw6.3 (Ml5.8) earthquake struck L’Aquila and the neighboring area, causing 306 casualties and about 60,000 displaced persons. While some of us carried out field trips to map surface faulting and shake effects, others started a retrospective analysis on the reasons why such a Cassandra’s Prophecy. Here we were able to show that the causative fault was the one having the highest probability of a characteristic earthquake in the mid-term, according to a renewal model conditioned to the time elapsed since the last event. Good fault-based definition of seismogenic sources, simplistic physically-motivated models of earthquake occurrence, and basic error propagation techniques were the ingredients for an empirically validated earthquake forecast. The occurrence of such a disaster does not significantly decrease the seismic hazard in the Central Apennines that in the past experienced long lasting seismic crises. The aggregate probability in the next 5 years that a M>6.0 earthquake will occur on at least one of the nearby known sources decreases from 9% to 6% after the earthquake, but it still remains a peak value along this sector of the Central Apennines. Besides integrated approaches to develop more sophisticated physically-based earthquake models, which the whole scientific community is working on, effective strategies of seismic risk reduction have to be promptly undertaken. Seismic hazard assessment: analysis of different hazard functions and forecast of the time-to-the-next event Rotondi R.1, Varini E.2, Betrò B.3, Barba S.4, Basili R.5 1.C.N.R. – I.M.A.T.I., Via Bassini 15, Milano, [email protected] 2.C.N.R. – I.M.A.T.I., Via Bassini 15, Milano, [email protected] 3.C.N.R. – I.M.A.T.I., Via Bassini 15, Milano, [email protected] 4.I.N.G.V. – sez. Roma, Via di Vigna Murata 605, Roma, [email protected] 5.I.N.G.V. – sez. Roma, Via di Vigna Murata 605, Roma, [email protected] Let us classify the stochastic models used in seismic hazard assessment in two main groups according to their hazard function: those with stepwise increasing hazard and those with non monotonic (at least in theory) hazard. We are aware that models unifying these two trends would be better but the simple combination of two models, one from each class, is not sufficient (Schoenberg, Bolt, BSSA (2000)) and more structured proposals are still under study (Varini, Statistical Methods and Applications (2008)). Here we exam in detail models from both the classes and compare their performance in terms of fitting and of simulated probability distribution of the time-to-the-next event. As for the first class, that is self-correcting models, first we analyze the stress release model and then we propose a new version of this model based on the comparison between expected and observed seismic slip; we denote this model by slip model. The models have been applied to the seismogenic areas of DISS 3.0.2 subdivided into eight tectonically homogeneous macro-regions. The earthquakes associated to each area have Mw ≥ 5.3 and are drawn from the CPTI04 catalog. An important role is played by the concept of area in the expression which links seismic moment and slip; different possible choices have been considered: the area computed through average length and width of the structure drawn from DISS, the sum of the area of all the seismogenic areas included in the macro-region, or the rupture area corresponding to the moment magnitude of the event according to the formula by Wells and Coppersmith (1994). Taking into account that CPTI04 catalog covers the period up to the end of 2002, we could validate the forecast times by the events occurred after that date; promising agreements have been obtained. Introduction of seismic source directivity on hazard map. Spagnuolo Elena1,2, Herrero André1, Cultrera Giovanna1, Spallarossa Daniele2 1 Istituto Nazionale di Geofisica e Vulcanologia, Sezione Tettonofisica e Sismologia, Roma. 2 Università degli Studi di Genova, Dip.Te.Ris., Genova. The seismic hazard maps are mainly influenced by the uncertainty associated to the ground motion predictive equation (GMPE). Longer is the time of interest, larger is the influence. This uncertainty incorporates the aleatory uncertainty of the shaking, which is increased also by the ergodicity assumption of the GMPEs and the epistemic uncertainty. For instance, the two standard source parameters, i.e. the magnitude and the distance, are not enough to describe the seismic source influence on the commonly used intensity measures (e.g. PGA, SD at 10sec., etc.). A way to reduce the uncertainty is to refine the seismic source description inside the GMPEs (e.g. NGA project). However, the reduction of the variance is not directly linked to the number of source parameters used in the equation, because many of them are not independent, and the ergodicity tends to blur the results. It is more efficient to group the parameters, linked by a physical effect (e.g. the radiation pattern, faulting style, etc.). In the framework of the DPC project S2, we have studied the introduction of the directivity in the GMPEs and its effect on hazard map. For this purpose, we have used two different corrective factors proposed respectively by Somerville et al. (1997) and by Spudich and Chiou (2008). The differences between the two factors in terms of formulation, source parameters involved and application field are important. The first factor depends on parameters and comes from theoretical deduction. It is simple to use, generic but does not introduce a reduction of the variance. The second one includes many source parameters and it is a hybrid factor, which functional formulation is deduced from the theory but at the same time is scaled on data. This procedure allows to obtain an estimation of a variance reduction, but its applicability is then narrowed to a specific GMPE (i.e. Abrahamson and Silva, 2008). The classic hazard equation is then adapted to increase the number of source parameters (i.e. to increase the number of integrals) and to take into account the corrective factors for directivity. We present the comparisons of hazard maps depending on the directivity factor and on the probability density functions of the fault strike and of the rupture laterality. Session Rapid Processing, Sharing, and Dissemination of seismological data after large earthquakes INGV Seismic Waveform Distribution and Archive Integration. Lauciani V., Mandiello A.G., Marcocci C., Pintore S., Mazza S. Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Roma, [email protected] The present, continuously increasing, number of seismic stations is posing a serious challenge to data centers regarding the management of data flows and station information. CNT is following the approach of a distributed data archiving, participating to the European Integrated Data Archive (EIDA), as proposed and realized in the European Project NERIES. Data are collected and archived in a Storage Area Network with a SEED Data Structure, while station information is organized in a MySQL database. Using the ArcLink protocol (by the Geofon Group, GFZ) station information is integrated in one virtual repository. Similarly, seismic data are shared in a common data-bank. This approach results in a highly scalable archive, in which a strong backbone is provided by the major Institutions, but where also minor participants can play their role. Each data producer can manage data and station information from its privileged workplace, whereas the technicalities behind the scene are hidden to the data users, who can access the same integrated data set contacting one data center only. As partners of EIDA we have given access to the European Archive through our web pages at http://eida.rm.ingv.it, where data from a huge number of global, regional, national and local broad band stations are available. At the same time, non-restricted data provided by projects and cooperations are available in the same page, as well as data from national and international temporary deployments. Automatic data quality control is at an experimental stage. Tools for non-interactive data extractions are under development. Integrating data from temporary networks into INGV Data archive. Mandiello A.G., Govoni A., Di Stefano R., Moretti M., Chiaraluce L., Lauciani V., Pintore S., Baccheschi P., Margheriti L., Mazza S. and Selvaggi G. Centro Nazionale Terremoti, Istituto Nazionale di Geofisica e Vulcanologia To manage and to share the continuous seismic recordings to every seismologist and to all international data centers INGV has chosen to distribute data in the international standard SEED format (Standard for the Exchange of Earthquake Data ). According to the format, waveforms are accompanied by all the information needed to use them, i.e. site and instrument information. The S5 project stimulate the integration of different source of data in this archive; in fact to achieve this goal we have archived in SEED only data volumes (miniseed) all the waveforms data that we gather from permanent INGV networks (Rome, Naples and Catania) and from temporary deployments (in case of experiments or following a seismic emergencies). The core of INGV acquisition system is based on the Seedlink protocol implemented in the SeisComP free software package (Geofon, GFZ). The Seismological Communication Processor (SeisComP) is a widely adopted software for seismological data acquisition and real-time data exchange over internet. Since the beginning it includes all the tools to gather, distrubute and archive waveform data and has proven to be a reliable and robust implementation. The SeedLink server handles each incoming data format with a plugin. This highly specialized program knows all the details of incoming data format, handles format control and data conversion and 'talks' to the SeedLink server using a small set of well documented library functions. This architecture allows an easy implementation of the needed plugins by the users or by the instrument producers (Nanometrics, RefTek) while the main source code of the server is maintained and developed by the same software team. ArcLink (SeedLink companion protocol available in the latest distribution of SeisComP 3.0) complements SeedLink by providing access to archived data and station database. ArcLink is based on TCP and uses simple ASCII coded commands for requesting all waveforms and metadata available in a time windows. Moreover ArcLink supports distributed archives. With this feature the data bank has a highly scalable architecture. At present the ArcLink server is under test and can be interrogated through a web page (http://hgp2.int.ingv.it/arclink/query?). Real time and off-line data form L’Aquila emergency deployment are available to all INGV researchers and non INGV ones, according to the data policy adopted for each data set. Data control and formatting procedures originally written for processing the Messina 1908-2008 experiment data set have been included in routine operations. ISIDe, the Italian Seismic Data Base. Mele F M, Basili A, Bono A, Lauciani V, Marcocci C, Marchetti A, Moro R, Pintore S, Quintiliani M, Scognamiglio L, Mazza S INGV, via di Vigna Murata 605, 00143, Roma, Italy, [email protected] ISIDe, the Italian Seismic Instrumental and parametric DatabasE, includes today more than 38000 earthquakes occurred in Italy and surrounding seas since April 16 2005, ranging from magnitude ML 0.2 to ML 5.9. The recent L'Aquila seismic sequence contributes with more than 12000 preliminary locations. ISIDe provides the best revised information about the Italian very recent seismicity, as soon as it is available, and includes: 1. Revised Quasi-Real-Time Locations, 2. the Seismic Bulletin (since April 16th 2005), 3. the Seismic Bulletin (before April 16th 2005). All the ML are computed from synthetic Wood-Anderson records obtained from horizontal broad band and very-broad band registrations. The minimum magnitude of completeness is ML 1.8, as computed from the cumulated Gutenberg-Richter distribution. Although ISIDe is spanning only the last four years and half of Italian seismicity, it is unprecedented in Italy for completeness and homogeneity and represents an optimum test data set to verify small and moderate seismicity pattern models. ISIDe is available at http://iside.rm.ingv.it/. Implementation of tsunami early warning using earthquake rupture duration Michelini A.1, Lauciani, L. 1, Lomax, A.2 1.Istituto Nazionale di Geofisica e Vulcanologia, via di vigna murata, 605, 00143 Roma, Italy, [email protected], [email protected] 2. ALomax Scientific, Mouans-Sartoux, France. [email protected] We have implemented the Duration-Exceedance algorithm (Lomax and Michelini, 2009) in the INGV seismic center. This algorithm can provide effective tsunami early warning for coastlines near a tsunamigenic earthquake requires with notifications within 5-15 minutes. The algorithm is based on the observation that tsunamigenic earthquakes have an apparent rupture duration, T0, greater than about 50 s. T0 gives more information on tsunami importance than moment magnitude, Mw, and we show application of a procedure which exploits seismograms recorded near an earthquake to rapidly determine if T0 is likely to exceed T=50 or 100 s. We show that this procedure can be completed within 3-10 min after the earthquake occurs, depending on station density, and that it correctly identifies most recent earthquakes which produced large or devastating tsunamis. This identification forms a complement to initial estimates of the location, depth and magnitude of an earthquake to improve the reliability of tsunami early warning, and, in some cases, may make possible such warning. Accounting for rupture directivity in ShakeMap: the case of the 2008 Miyagi earthquake Spagnuolo Elena1,2, Faenza Licia3, Cultrera Giovanna1, Michelini Alberto3, Herrero André1, Angela Saraò4 1 Istituto Nazionale di Geofisica e Vulcanologia, Sezione Tettonofisica e Sismologia, Roma. 2 Università degli Studi di Genova, Dip.Te.Ris., Genova 3 Istituto Nazionale di Geofisica e Vulcanologia, CNT, Roma. 4 Istituto Nazionale di Oceanografia e di Geofisica Sperimentale, Trieste ShakeMap may be considered as an interpolation operator based on ground motion empirical predictive equation (GMPE). If the station network is very dense, the weight of the interpolation is weak and the ShakeMap is mainly driven by the observations. However, only few regions, such as Japan and Taiwan, have a good coverage of observation points. For the other regions, the choice of the GMPE can affect strongly the goodness of the ground shaking estimation. However, the used GMPEs depend only on the magnitude of the event and the distance between the hypocenter and the observation point. When a ShakeMap is determined without the introduction of the causative fault, the geometry of the resulting interpolated ground motion is circular (or circular-like depending on the site effect correction), centered at the event location. Insertion of the causative fault and use of the minimum distance from the fault projection can correct the simple circular dependency above but this may not be enough to reproduce faithfully the ground motion close to the fault of a large earthquake, since directivity effects, deriving from source unilateral rupture, are not accounted. In general, directivity may influence the ground motion through a large frequency band and at large distances. The new generation of GMPEs (e.g, Spudich and Chiou, 2008), or the corrective factor proposed by Somerville et al. (1997), account for directivity and may improve the ShakeMap result when a few basic seismic source parameters become available (strike, rupture laterality, etc.). In order to test the behavior of these new equations to the purpose of determining more accurate ShakeMaps, we have chosen the 2008 Miyagi earthquake in Japan, which displayed a strong directivity toward the South (Cirella et al. 2009) and it was recorded by the dense K-net and Kik-net seismic networks. The test conducted in the project DPC-S3 consists of computing a reference ShakeMap, using all the data available. The second step studies the degradation of the ShakeMap when the number of PGM observation points is reduced. Thirdly, we introduce directivity within ShakeMap through the GMPEs modified with the Spudich and Chiou (2008) correction factor, and we quantify the improvement of the ground shaking estimation. The Earthquake Early Warning System in Southern Italy: Methods and Technological Development A.Zollo (1), G.Iannaccone(2), L.Elia(4), G.Festa(1), M. Lancieri(3), C.Martino(4) and C.Satriano(4) (1) University of Naples Federico II, Department of Physics (2)Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano (3) Ecole Normale Superieure, Paris, France (4) AMRA scarl (Centro di Competenza in Analisi e Monitoraggio del Rischio Ambientale) An earthquake early warning system (EEWS) is a real-time, modern information infrastructure which is able to provide rapid notification of the potential damaging effects of an impending earthquake. The final aim is to predict the expected damage and losses at critical infrastructures, tens of seconds before the occurrence of the strong ground shaking produced by the earthquake. The operability and reliability of an EEWS thus depends on the fast telemetry and processing of data from dense instrument arrays deployed in the source region of the event of concern (regional EEWS) or surrounding the target site (on-site EEWS). For a regional EEWS, the relevant source parameters (event location and magnitude) are estimated from the early portion of recorded signals in the near-source region. They are used to predict, with a quantified confidence, a ground motion intensity measure at a distant site where a target structure of interest is located. In the framework of several research projects financed by the Campania Region, European Union (SAFER) and RELUIS, a prototype system for seismic early and post-event warning is being developed and tested, based on a dense, wide dynamic, seismic network deployed in the southern Apenninic belt region (ISNet, Irpinia Seismic Network). Robust and effective algorithms for the real time and evolutionary event location have been developed to provide in few seconds the information about the hypocenter location. On the other hand, the feasibility of a real-time earthquake magnitude determination using the information carried out in the initial portion of recorded P- and S-signals is tested by analyzing the strong motion records from the European and Japanese data banks. The southern Italy EEWS implements a probabilistic, evolutionary approach for the real time location and magnitude estimation based on a general strategy, where the computation starts a few seconds after the data from a small number of recording stations are available, and the source parameters estimates along with their uncertainties are progressively updated with time. The system testing, validation and performance is illustrated through the off-line application to synthetic seismograms originated by an extended fault rupture process and to real strong motion records from the Euro-Mediterranean and Japanese data banks. Session Strong Ground Motion Assessing regional and site-dependent variability of ground motions for ShakeMap implementation in Italy. Bragato P.L.1 1. OGS, Dip. CRS, Via Treviso, 55, 33100 Cussignacco (UD), [email protected] This study investigates the theoretical improvement of the ground-motion models used for ShakeMap implementation in Italy, obtainable by accounting for regional differences of wave propagation and for site effects. The analysis considers 922 recordings of 116 earthquakes with local magnitude ML in the range 2.7-4.5, collected from 137 stations at epicentral distances up to 100 km. Alternative partitions of the available stations into separated zones are considered. For each zone in a partition, an empirical ground-motion model is estimated, which include station correction terms. A genetic algorithm is used in order to minimize the standard deviation of the residuals log(observed)-log(predicted) over the set of the possible partitions. For the logarithm of the peak ground acceleration (PGA), it is estimated that site effects contribute to about 30% of the overall standard deviation, while regional differences contribute with no more than 4%. Furthermore, the estimated station correction terms are weakly correlated with the amplification factors based on shear-wave velocity in the upper 30 m (Vs30) used in ShakeMap to model site effects. Similar results apply to peak ground velocity (PGV) and spectral acceleration response (PSA) at 0.3, 1, and 3 sec. These results suggest that parametrizations of site effects alternative to those based on Vs30 should be considered, while the contribution of any regionalization of the ground-motion models is marginal. Large effects of Moho reflections (SmS) in the Po Plain and implications for ground motion prediction in northern Italy. Bragato P.L.1, Augliera P.2, D’Alema E.2, Marzorati S.2, Massa M.2, Sugan M.1 1.OGS, Dip. CRS, Via Treviso, 55, 33100 Cussignacco (UD), [email protected], [email protected] 2.INGV - Sezione di Milano, via Bassini 15, 20133 Milano, Italy, 55, [email protected], [email protected], [email protected], [email protected] Recent moderate earthquakes occurred in northern Italy have evidenced how currently available ground-motion predictive equations largely underestimate shaking in the Po plain and surroundings at distances greater than 70 km. A data set of PGA data available for the area for weak and moderate earthquakes has been collected in order to investigate on the subject. Non-parametric regression analysis of PGA as a function of magnitude and hypocentral distance show that PGA is systematically unattenuated for distances between 70 and 130 km. According to the literature, this effect can be explained by strong reflection at the Moho. Such behavior appears to be independent on frequency and has been observed also for the 1976 Friuli earthquake (Mw 6.4), so that it can be reasonably generalized to strong earthquakes. Further investigation is required for assessing the correlation of the observed strong reflection with both the morphology of the Moho (e.g., its inclination and depth) and the structure of the crust. Regression analysis of MCS Intensity and ground-motion parameters in Italy and its application in ShakeMap Faenza L.1 , Michelini A.1 1.Istituto Nazionale di Geofisica e Vulcanologia, CNT, Via di Vigna Murata 605, 00143, Roma, Italy, [email protected], [email protected] In Italy, the Mercalli-Cancani-Sieberg, MCS, is the intensity scale in use to describe the level of earthquake ground shaking, and its subsequent effects on communities and on the built environment. This scale differs to some extent from the Mercalli Modified scale in use in other countries and adopted as standard within the USGS-ShakeMap procedure to predict intensities from observed instrumental data. We have assembled a new PGM/MCS-intensity data set from the Italian database of macroseismic information, DBMI04 and the Italian accelerometric database, ITACA. We have determined new regression relations between intensities and PGM parameters (acceleration and velocity). Since both PGM parameters and intensities suffer of consistent uncertainties, with a consequent large scatter of PGM data for each intensity unit, we have used the orthogonal distance regression technique. The new relations are IM C S = 1.68 ± 0.22 + 2.58 ± 0.14 log P GA, σ = 0.35 and IM C S = 5.11 ± 0.07 + 2.35 ± 0.09 log P GV , σ = 0.26. Tests designed to assess the robustness of the estimated coefficients have shown that single-line parameterizations for the regression are sufficient to model the data within the model uncertainties. The relations have been inserted in the Italian implementation of the USGS-ShakeMap to determine intensity maps from instrumental data and to determine PGM maps from the sole intensity values. Comparisons carried out for earthquakes where both kinds of data are available have shown the general effectiveness of the relations. Ground motion Prediction equations and S-Projects Pacor F.1 and Working Group S-Projects 1Istituto Nazionale di Geofisica e Vulcanologia, Milano,Italy, [email protected] The S-Projects are all involved in the development and/or evaluation of Ground motion Prediction Equations (GMPEs). A workshop was organized in June 2009 with the aim of discussing the following questions: − Do the S-Projects use different GMPEs because they are applied with different aims? − Does it make sense to seek the most appropriate GMPEs for the Italian territory? − Does it make sense to use consider regional GMPEs? A preliminary review of the GMPEs adopted in the S-Projects, ranging from global models (S2 and S3) to Italian (S4) and regional ones, was necessary in order to clarify the criteria for their selection and application in the framework of each Project. A preliminary comparison among different GMPEs adopted by different S-Projects was performed through their median estimations. Difficulties arise when a quantitative comparison is attempted considering models defined over different ranges of the explanatory variables and for different functional forms. Differences in the site classification have also to be accounted in the comparison. The occurrence of the 2009 Abruzzo earthquakes allowed us to test some GMPEs proposed in the Projects. Preliminary results based on the residual analysis show that both national and global models tend to overestimate the observed peak ground accelerations,, although Italian GMPEs underestimate the near-source PGA, especially for rock conditions and large magnitudes. This complex problem cannot, of course, be solved within the S-Projects and exhaustive studies should be devoted to this topic, as in the past and ongoing international projects (e.g., PEER-NGA project, 2008, SHARE and GEM projects). A product of the workshop held in June 2009 will be a critical review of the existing GMPEs used in the S-Projects, provided with their limits of applicability to have a clear frame of reference about the models that can be adopted in Italy for ground motion predictions. Near-fault Earthquake Ground Motion Simulations in complex geological configurations by a High Performance Spectral Element Code: GeoELSE Stupazzini M.1, Cauzzi C. 2, Faccioli E. 3, Paolucci R. 4, Smerzini C.5, Vanini M.6 1 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 2 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 3 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 4 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 5 Doctoral School of Earthquake Engineering and Engineering Seismology, ROSE School, IUSS Pavia, Italy, [email protected] 6 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] The coupling of near-fault conditions with complex geological configurations, such as deep alluvium filled basins or irregular topographic profiles, can generate distinctive features of earthquake ground motions. To deal with them, large-scale 3D numerical simulations of seismic wave propagation are appropriate, including the combined effects of source directivity, complex geomorphological conditions and a sufficiently detailed crustal model. This contribution illustrates selected applications of high performance computing by the Spectral Element Method (SEM) with the code GeoELSE (http://geoelse.stru.polimi.it). This can perform linear, non-linear viscoelastic and viscoplastic wave propagation analyses in large-scale Earth models, that include the seismic source, the propagation path, the local near-surface geology, and, if needed, the interaction with built structures . The implementation of GeoELSE in parallel computing architectures allows to resolve tens of million of degrees of freedom up to 2 Hz in reasonable computer times. After introducing the main theoretical features of the code and its implementation and performance in parallel architectures, some of the more representative results obtained by GeoELSE (involving 3D near-source simulations)are summarised, including: • The study case of the Grenoble alluvium filled valley (ESG06 international benchmark), which shows how the unfavourable interaction of fault rupture and complex geology may generate values of peak ground velocity in excess of 1 m/s even in low or moderate seismicity areas, and therefore increase considerably the seismic risk, especially in highly populated and industrially active regions, such as the Alpine valleys; • The 3D parametric analyses with respect to the source parameters on the Sulmona alluvium basin (Central Italy), carried out within the S2 Project (Task 4) to demonstrate how numerically simulated near-field ground motions can be used in conjunction with probabilistic tools for seismic hazard assessment, and adopted as well for estimating topographic amplification factors suitable for inclusion in GMPEs. Session Project S1 Analysis of the seismic potential in Italy for the evaluation of the seismic hazard S1-UR3.01: Regional geological and instrumental (InSAR) constraints to tectonic deformation and their associated uncertainties: implications in seismic hazard estimates. Akinci A.1, Barba S.1, Basili R.1, Burrato P.1, Carafa M.1, Fracassi U.1, Kastelic V.1, Megna A.1, Scognamiglio L.2, Stramondo S.2, Tiberti M. M.1, Tolomei C.2, Valensise G.1, Vannoli P.1 1. INGV-RM1, Via di Vigna Murata, 605, Roma, [email protected] 2. INGV-CNT, Via di Vigna Murata, 605, Roma This Research Unit (RU) is carrying out a multifaceted work that includes in-depth analysis of geological and InSAR data in the area of Crotone Peninsula (Calabrian Arc) and geomorphic/structural data in the Dinarides. In addition, the RU is carrying out studies to better the construction and validation of a seismogenic source model and seismic hazard calculations for the entire Italian territory. In this report we will illustrate a selection of the main results obtained in some of these studies. Crotone, long-term uplift. We analyzed satellite images and aerial photos of the area between the Squillace and Sibari Gulfs. The analysis, carried out at different scales, was much more detailed in the Crotone Peninsula where it led to detecting and mapping the most significant geomorphic features, such as scarps, terrace risers, terrace treads, stream network and catchments, and landslides. We also carried out several field campaigns during which we collected over 100 samples, including corals and mollusks from Late Pleistocene terraces, and Holocene markers (algal rims, bivalves, lithodomes) from raised shorelines. A number of these samples are being dated by means of U/Th and radiocarbon techniques. Crotone, short-term uplift. We selected and acquired the necessary dataset to fully image the area under study (roughly 100x100 km) and organized it for processing. The selected dataset is composed by: 37 ERS1-2 data for track 179, 41 ERS1-2 for 451, 38 Envisat data for track 451, and covers a time span from 1992 to present for the multitemporal analysis. Dinarides. We carried out a regional analysis based on structural geology and tectonic geomorphology of the Dinarides thrust belt along the coast of Croatia and Montenegro. Tectonic consistency and seismic moment balance of seismogenic sources. A three-step test was developed for a thorough tectonic validation of the DISS content against independent seismicity data. The DISS version used in these tests is the 3.1.0. Seismicity data includes instrumental data from the INGV Bulletin, historical data from the CPTI04, and focal mechanisms from EMMA. Uncertainties on probabilities of earthquake occurrence. This study aims at determining the extent to which the application of strain-derived slip rates decreases the variability of fault occurrence probability estimates obtained from geologic slip rates and historical earthquake occurrences. First results were obtained for the Individual Seismogenic Sources from DISS 3.0.4. The usage of the Composite Seismogenic Sources is still under development. S1 - RU 5.02: Reliability of the inversion of coseismic geodetic data in homogeneous and layered half-spaces . Amoruso A.1, Barba S.2, Crescentini L.3, Megna A.4 1.Dipartimento di Fisica Università di Salerno, Via Ponte Don Melillo Fisciano (SA), [email protected] 2.Istituto Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605 Roma, [email protected], Address, email 3.Dipartimento di Fisica Università di Salerno, Via Ponte Don Melillo Fisciano (SA), [email protected] 4.Istituto Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605 Roma, [email protected] In geodetic data inversions the medium is often represented as a homogeneous, isotropic, and elastic half-space or as a layered medium, and their elastic parameters are considered to be standard. The use of layered models neglects many of the real characteristics in the crust, as the lateral heterogeneities; thus computed displacements and observed data may differ more than hypothesized measurement errors, using a layered although standard model. We aim to quantify the effects of using standard layering when assessing source features from coseismic deformation data, in geological environments typical of the Apennines. For this purpose, we invert synthetic coseismic displacements at the surface - computed by a finite element model (MARC) for some source parameters and complex crustal structures - for a planar fault embedded in a standard layered medium. The blind inversions are performed using an ad-hoc improved version of the ANGELA code, previously developed by some of us in the framework of other projects. We will show the results of a few preliminary tests. RU5.01: The present-day behavior of the Castrovillari Fault Abdelkrim Aoudia1, Alessandra Borghi.2 1. Abdus Salam International Centre for Theoretical Physics, Strada Costiera 11, 34100, Trieste, [email protected] 2.OGS c/o Politecnico di Milano, p.zza Leonardo da Vinci 32, 20133 Milano, [email protected] A key issue in our understanding of the earthquake cycle and seismic hazard is the behaviour of an active fault during the interseismic phase. Locked and creeping faults represent two end-members of mechanical behaviours that are given two extreme rupturing hazard levels, that is, high and low, respectively. GPS data analyses have been carried out over the Pollino Range, an extensional environment within the Africa–Eurasia plate boundary, to disclose the behaviour of the long-lasting quiescent Castrovillari normal fault. A campaign GPS network of about 30 x 17 km composed by 10 sites and centered in the normal faulting area has been set up. Six measurement campaigns have been performed from 2003 up to 2009. The geophysical inversion of GPS velocity shows fast aseismic creeping at all depths of the fault plane, and this behavior is confirmed by DInSAR analysis. We have also analysed nine UNAVCO continuous GPS stations across the Calabrian Arc, looking for eventual transients and discontinuities in time series. Three methods have been tested: wavelet transform, the Bayesian approach and Blake and Zissermann variational model. Preliminary results will be presented. RU_6.01: In-land tsunami deposit studies (off-fault paleoseismology) for hazard assessment in Eastern Sicily (in collaboration with RU 6.02) Barbano M.S.1, De Martini P.M.2, Gerardi F.1, Pirrotta C.1, Smedile A.2, Pantosti D.2, Pinzi S.2, Del Carlo P. 3 1. Dipartimento di Scienze Geologiche, Università di Catania, Corso Italia 55, 95129 Catania, (RU 6.01) [email protected], [email protected], [email protected] 2. INGV, Sezione Sismologia e Tettonofisica, Via Vigna Murata 605, 00143 Roma, (RU 6.02) [email protected], [email protected], [email protected], [email protected] 3. INGV, Sezione di Pisa, Via della Faggiola 32, 56126 Pisa, (RU 6.02) [email protected] Geomorphological study of the coastal landscape by aerial-photographs analysis, satellite image interpretation and field surveys have been undertaken, with the aim to identify new areas potentially suitable for tsunami deposit findings in Eastern Sicily. We selected 3 coastal areas with low energy sedimentation where the tsunami deposits may sediment and be preserved: Fiumefreddo Oasis in the eastern flank of Mt Etna, Pantano Morghella and the Ispica lagoons in the southeastern coast of Sicily. Detailed field surveys and core campaigns have been carried out in order to identify anomalous sandy sheets deposited by tsunami waves. 33 cores dug until a distance of 1.2 km from the coastline in Pantano Morghella allowed us to delineate over an area of about 3 km2 the presence of a sandy layer found at about a meter of depth, in a clayey sequence. Sedimentologic/stratigraphic and paleontological analysis revealed that the sandy layer characteristics indicate a marine origin. Sandy samples contain lithic and microfauna association different with respect to the clayey samples collected above and below. These latter samples contain an oligothipic lagoonal assemblage of benthic foraminifera (Haynesina germanica and Ammonia tepida). On the contrary, the sandy samples show an assemblage made of several reworked foraminifera (both planktonic and benthonic) with many marine macro fossils fragments (corals, sea urchins, bryozoa and molluscs), few lagoonal specimens and contain a lithic component made of well-rounded yellowish mainly carbonatic clasts and shell fragments, revealing similarity to the local beach sand. Radiocarbon dating, performed on three samples collected just above, within and below the sandy anomalous layer, gave ages close in time, suggesting a sudden deposition due to a high energy event occurred in the interval 270-650 A.D. These preliminary results allow us to tentatively correlate the marine inundation with the 365 A.D. Crete tsunami. Further sample analysis and radiocarbon datings are in progress. At Fiumefreddo site, micropaleontological analyses have been done. In a dark grey clayey silt sequence rich in vegetal remains, volcanic sand and rare quartzs, a sandy layer was found at about 1.3 m depth in 2 cores, but samples are barren except for two plancktonic foraminifera found in a sample. Preliminary test cores in the Ispica lagoons (Pantani Cuba, Bruno and Ponterio) showed a silty stratigraphic sequence. However more core surveys and analysis will be carried out in the last two areas where the environments are promising for the recognition of high energy sediments of marine origin. Large boulders along the coast of the Vendicari Marine Reserve (south-eastern Sicily): storm and tsunami deposits M.S. Barbano, C. Pirrotta, F. Gerardi Dipartimento di Scienze Geologiche, Università di Catania, corso Italia 57, 95129 Catania, (RU 6.01) [email protected], [email protected], [email protected] We analyzed anomalous deposits of boulders occurring in Vendicari Marine reserve along the south-eastern Sicilian coast in order to verify if storm or tsunami waves were responsible for their detachment, transport and deposition. Three different GPS surveys (September 2006, July-September 2008 and February-April 2009) have been performed with the aim to directly observe the distance of each boulders with respect to the shoreline and if storms either deposited new boulders or removed them. A morphological analysis aiming to identify boulder shapes, measuring their volumes, elongation axis azimuth and their pre-transport setting was carried out. Boulders are made up by calcarenites with a specific weight of about 2,3 g/cm3; the weight of some boulders is up to about 24 tons. Most of the boulders, showing wide, flat rock pools and barnacles on their surface, have been carved out from the supratidal zone. Some boulders were carved out from the mid or sublittoral zone, showing widespread algal, biogenic encrustations (Vermetids and Serpulids) and Lithophaga holes. Sometimes fresh encrustations and Lithophaga shells, still preserved inside the holes, suggest a recent deposition. Thanks to the repeated field surveys, we observed that, after a strong winter storm occurred during January 2009, several boulders were moved, deposited and again removed by storm waves on the wave cut platform. Hydrodynamic equations jointly to statistical analysis of sea storms have been used to determine the extreme event, geological or meteorological, responsible for this singular accumulation. We computed the minimum wave height required to move each boulders, the decay equation and the maximum flooding for storm and tsunami waves of the Ionian coast, successively compared with the boulder distribution. The results, jointly with observations, show that the strongest storms were probably responsible for the current distribution of many boulders but the ones placed at distance >40-45 m from the coastline are likely deposited by tsunamis. Statistical analysis on boulder distribution along the coast has shown that they are distributed following three Gauss curves confirming that they are mainly deposited by three type of waves. Two kinds of waves are linked to the strongest storms, whereas the third one, which is probably due to tsunamis, has a longer period and therefore a lower attenuation. Radiocarbon dating performed on three boulders, having weight of about 15 t and sited at a distance >40 m from the shoreline, suggests that two of them were likely deposited by the 1693 tsunami, and one by a tsunami occurred after 650-930 AD. UR 02.02 – Structure of the Crust (Integrated 3D model) Part 01: Moho Depth in Italy and smooth 3D P-wave velocity model from the joint interpolation of Controlled Source Seismology and Receiver Functions data. Bianchi I.1, Di Stefano R.1, Piana Agostinetti N.1, Ciaccio MG.1, Amato A.1 1 Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, [email protected] The main goal of this Research Unit (RU) is to obtain a high definition integrated 3D seismic velocity model of the crust in Italy. Several studies revealed strong differences between the European, the Tyrrhenian, and the Adriatic crustal thickness beneath Italy and surrounding regions. Moreover there exists a strong mutual influence between lateral variations of crustal thickness and the geodynamic evolution of this region. This evidence makes a detailed map of the Moho a key element for structural and geophysical studies and to develop the reference 3D velocity model that is the aim of this RU. To retrieve a reliable map of crustal thicknesses, able to account for strong Moho undulations and overlapping, a large and dense high quality dataset of Moho depths and a stable interpolation method are necessary. We used an interpolation method able to determine the 3D topography and lateral continuity of seismic interfaces using information from different data-sets. Such method was developed by Waldhauser (1996) to interpolate data from Controlled Source Seismology (CSS) profiles, and later implemented by Di Stefano (2005) to include results from high resolution Local Earthquake Tomography (LET). During the first year of this project we further implemented the method to include important information’s on the Moho depth retrieved through Receiver Functions (RF) studies. We collected results from three resuming works (Piana Agostinetti & Amato 2009 for the whole peninsular stations, Lombardi et al 2008 for the Western Alps, and Geissler et al 2008 for the Eastern and Middle Alps) and used the Pg-Pn tomographic model by Di Stefano et al. (2009) to account for the P-wave velocities. The large amount of Receiver Functions observations and new data from CSS profiles that we introduced in this update helped to increase the overall reliability of our map and to better image regions less sampled before. We present the updated maps of the European, Adriaric and Tyrrhenian plates beneath Italy and surrounding regions and the related smooth 3D P-wave velocity model. We also compare our results with the latest and more complete version of the European Map of the Mohorovicic discontinuity. RU S1_1.01: Vertical Crustal Movement from Tide Gauges and Satellite Altimetry Braitenberg C.1, Grillo B. 2, Mariani P. 3, Tunini L. 4, Nagy I. 5 Dipartimento Scienze della Terra, Universita’ di Trieste, Via Weiss 1, 34100 Trieste [email protected], [email protected], [email protected] [email protected], [email protected], The goal of our study is to determine crustal movement rates along the Italian coast. The method consists in studying tide gauge and satellite altimetric data jointly. The sea level change measured by tide gauges is the sum of the sea surface height variation and the vertical crustal movement. Tide gauge data therefore contain the vertical crustal movement rates, but the problem consists in separating this signal from the sea level observations. Satellite altimetric observations give the geocentric sea surface height and can therefore in principle be used to determine the sea surface height rates and separate the sea surface from the crustal signal in tide gauges. Along the Italian coastline 26 tide gauge stations with 10-years of continuous data are available. We have explored the sea level data statistically, including a study of the space-time variation of rates, of the optimal way of calculating rates and of the correlation-coefficients between stations. The tidal sea level change rates have a relatively large variability, greatly ascribable to the short time interval of analysis (10 years), and are comprised between -5 and +9 mm/yr with an approximate uncertainty of 0.5 mm/yr. The satellite altimetric data pose a challenge to the joint analysis, due to very different time resolution (10 days versus hourly sampling of tide gauges), sparse spatial sampling of the Mediterranean due to track distances, and the difficulty in acquiring altimetric data in the vicinity of the coast. We have analyzed the satellites Topex/Poseidon and Jason1, the study of satellite ENVISAT is under way. We explore also the satellite altimetric data statistically, investigating the correlation matrices of the data and the time-space variation of the sea surface change rates. We find that the sea surface change rates derived from the altimeter have an even greater variability compared to the tide gauges for the same years, showing that changes far from the coast are greater than near the coast. We discuss the problem of how to use the altimeter data in those cases in which the tracks are far from the tide gauges, as is the case for the satellite Topex/Poseidon and the tectonically interesting tide gauge stations of Sicily and Calabria. An adequate space-time interpolation of the satellite data produces maps of spatial variations of sea surface change that are used in the study of the differential sea level rates of tide gauges. We show that geologically consistent crustal uplift rates are found for the highly seismic area of eastern Sicily and Calabria. RUS1_2.04 (2): Lithospheric imaging from tomographic and moment tensor non- linear inversion studies Brandmayr E.1, Panza G.1,2, Romanelli F.1 1. Dipartimento di Scienze della Terra-Università degli Studi di Trieste, Via Weiss 4 Trieste, [email protected] 2. The Abdus Salam International Centre for Theoretical Physics, Trieste In the description of the lithosphere-asthenosphere system of the Italic region use is made of: a) the database of the structural models for the lithosphere-asthenosphere system, obtained from surface wave tomography and non-linear inversion of dispersion curves, for 1°x1° cells in the whole Italic region (see RUS1_2.04), and b) the lithospheric thickness as defined according to the distribution of seismicity. The properties of earthquakes sources are studied, in order to obtain a constraint to the geodynamic modelling, using an advanced waveform inversion technique (INPAR) that allows the retrieving of the full earthquake moment tensor. INPAR allows the use of relatively short period waveforms, significantly improving the reliability of depth determination for shallow events, usually fixed by CMT and RCMT. This lead to a drastic relocation and change of mechanism of a major event of Umbria Marche 1997 seismic sequence, and provides new constraints to the geological setting and local stress field. The major results enlighten from the 3D structural model obtained from the structural inversion and from the distribution of the focal mechanisms obtained through INPAR, allow us to add new constraints to the geodynamics of the Appenninic subduction and retro-arc expansion of Thyrrenian basin, along the Crop-03 seismic experiment profile. In particular the structural model shows clear evidences of crustal doubling at the eastern margin of Corsica, probably a fragment of Alpine-Betic Orogen, involved in the subsequent expansion process. Moreover the presence of frequent seismicity at intermediate depths (20-50 km) near to the Apennines compressive front is a robust constraint to the location of the subducting slab. The data from 8 contiguous cells, along a section WSW-ENE oriented, are processed in order to highlight the consistency between geological interpretations and geophysical data. RUS1_2.04: Structure and thickness of the Italian lithosphere Brandmayr E.1, Zuri M.1, La Mura C.1, Panza G.1,2, Romanelli F.1 1. Dipartimento di Scienze della Terra-Università degli Studi di Trieste, Via Weiss 4 Trieste, [email protected] 2. The Abdus Salam International Centre for Theoretical Physics, Trieste The study of the structure and thickness of the lithosphere in Italy and its surroundings is important to the understanding of the geodynamic setting of the region; furthermore, the structural models for the crust and uppermost mantle (lithosphere-asthenosphere system) represent a major input for reliable geodynamic modelling, for the evaluation of seismic hazard using also deterministic methodologies, and for the determination of earthquake source mechanisms. An important goal is obtained with the determination of structural models for the lithosphere-asthenosphere system, obtained from surface wave tomography and non-linear inversion of dispersion curves, for 1°x1° cells in the whole Italic region. This was done through surface wave tomography using dispersion curves of Rayleigh fundamental mode along properly selected wave paths in the studied region. We used a non-linear inversion where the unknown independent parameters are S-wave velocities (Vs) and thickness of layers. All the solutions for each cell are processed with an optimization method with the aim to define a smooth 3D model of the lithosphere-asthenosphere system, in agreement with the concept of Occam razor. The search for representative solution is performed through three different algorithms. Local Smoothness Optimization (LSO), which searches iteratively within a set of cells the solution for which the lateral difference in Vs is minimized. LSO optimization has been accomplished for the whole Italian region and the resulting 3D structural model is presented. The LSO technique is powerful and fast, but strongly dependent from the starting cell. To avoid this dependence, two other methods were developed, GSO (Global Smoothness Optimization) and GFO (Global Flatness Optimization), which respectively search for the minimizing solution along a row of cells and through the whole study area. The refinement of cellular models by the intersection of the cited smoothing methods has already been accomplished for the Northern Italy and the Alpine region. As a final step, each of the resulting cellular models has been perturbed by (a) substituting the calculated Poissonian Vp with the values of Vp obtained from a proper processing of P wave tomography models and (b) fixing, within the original range of variability obtained with inversion, the values of Vs in such a way to obtain a medium as close as possibile to the Poissonian one. RU S1_1.02: Present day horizontal velocities of permanent GPS stations and the implied regional strain rate field Alessandro Caporali1 1.Department of Geosciences, University of Padova [email protected] We present a set of horizontal velocities in the Italian and surrounding areas computed by stacking normal equations of weekly network adjustments of permanent GPS stations starting 1999 and extending to September 2009. The individual normal equations have removable constraints, and the stacking is accomplished by imposing one set of constraints in position and velocity, according to the ITR2005 realization of the ITRS. Velocities computed out of an Eulerian, rigid body model of the Eurasian plate rotation are subtracted from the resulting velocities, and a picture of the implied kinematics is obtained in a rigorous way. The greatest increase in detail is obtained in NE Italy, where the inclusion of GNSS stations of the Province of Bolzano, the Veneto and Friuli VG regions enable the kinematics of the surface deformation to be detailed with unprecedented resolution. Considerable more information is obtained in Central Southern Italy, where several stations from RING of INGV and networks of Regional Governments again contribute to improve the resolution. To better constrain the boundaries of the national territory, we further present the combination of our network with a Central European Network CEGRN recently computed with identical approach and standards within the TopoEurope project, which has several stations in Austria and former Yugoslavia. The strain rate is computed with reference to this larger network. The evaluation is done at the center of the ZS9 seismic zones of INGV, with the intent to providing for each ZS an estimate of the zonal deformation taking place at present. Comparison of the strain rate eigenvectors with fault plane solutions PT axes projected to the horizontal give a quantitative estimate of the relation between the principal directions of seismic stress release and the principal directions of ground deformation. Late Quaternary tectonic activity in the Ferrara area inferred from stratigraphic architecture Caputo R.1, Stefani M..1, Bignardi S.1, Minarelli L.1, Nasser A-Z.1, Caputo R.1, Santarato G.1 1.Dept. Earth Sciences, Univ. Ferrara, via Saragat 1, 44100 Ferrara ([email protected]) Since Late Miocene, the Emilia-Romagna portion of the foredeep basin was progressively affected by compressional deformation, due to the propagation of the Apennines fold-and- thrust belt. The major tectonic structures within the basin have been recognised and are relatively well known, thanks to the widespread seismic survey performed for hydrocarbon exploration. More recently, a large amount of surface and shallow-subsurface information has been provided by the CARG project. Therefore, the area provides an opportunity to discuss the genetic relationship between tectonics, eustatic-paleoclimatic fluctuations, and depositional architecture. Activity on blind thrusts and fault-propagation folds caused repeated angular unconformities and impressive lateral variations in the Pliocene- Quaternary stratigraphy, causing thickness changes, from a few metres, close to the Apennines piedmont line, to more than 9 km, in fast subsiding depocenters (e.g. Lido di Savio). In the Ferrara region, the post-Miocene succession ranges from about 4 km, west of Sant’Agostino, to less than 200 m, on the Casaglia anticline, where Late Quaternary fluvial strata unconformably rest on Miocene marine marls. In this sector of the Po Plain, the tip-line of the northernmost thrust has been reconstructed north of the Po River and is associated with the growth of a large fold (Ferrara-Casaglia anticline), cross-cut by a complex splay of minor backthrusts and reverse faults. The thrust-anticline structure hosts an energy producing geothermal field, whose hydrogeological behaviour is largely influenced by the fracture pattern. The Apennines frontal thrust probably provided the seismic source for the earthquakes that severely damaged Ferrara, during the 1570 a.D. fall season, as documented by the structural damage still visible in many historic buildings (e.g. Cathedral). Surface deformation affected the fluvial drainage pattern and hence the depositional geometry. In the subsurface of the urban area, coarse grained synglacial sands are well developed. The depth of their top unconformity is largely variable, reflecting the ongoing deformation, from 30-40 m, in the southern syncline areas, to less than 10 m, on the Ferrara-Casaglia anticline. Where the synglacial sand reaches the shallower level, they can be amalgamated with, and cannibalised by, the Po River meanders (e.g. Bondeno-Sette Polesini area). Beneath the urban area, the surface rapidly rises northward, from 30 to 15 m, across what is possibly the near-surface expression of an active back-thrust structure. Beneath the southern portion of the present-day town, the upper Holocene succession is much thicker, reflecting higher sedimentation rates. Holocene depositional surfaces show some dip, induced by a spatially differentiated subsidence. The fluvial drainage pattern of the area can be reconstructed with a growing degree of accuracy, over the last 3 ka, recording the growing impact of human activity. The southern portion of diachronous splay of the Po delta distributary channels shows some parallelism with the buried tectonic structures. The large anthropogenic impact has however to be cautiously considered while investigating the tectonic induced anomalies of the drainage pattern. Further investigation on the role of the recent tectonic activity in the shaping of the fluvio-deltaic depositional architecture will likely support an improved estimation of the seismic hazard of the area. S1 - RU 5.03 : Strenght of the lithosphere and strain rate in the Northern Apennines Carafa M.M.C1, Barba S.2 1Istituto Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605 Roma, [email protected] 2Istituto Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605 Roma, [email protected] Deformation of the Northern Apennines lithosphere is studied here through numerical experiments. The integrated strength of lithosphere and the velocity field of the area are investigated by varying rheology and heat flow in appropriate ranges and using the finite element code SHELLS. For all the experiments, as boundary conditions, we imposed a relative compression between Nubia and Eurasia affects the Southern edge of the model, a NW-directed basal shear traction drives the lithosphere beneath the Apenninic chain, and all the other edges are fixed with respect to Eurasia. The testable predictions of any simulation experiment include: - the horizontal velocities computed in a locked-fault state, which are compared with interseismic GPS measurements; the RMS of residuals defines the GPS misfit; - the direction of the most-compressive horizontal strain rate computed in the long-term state, which is compared with the SHmax directions from the World Stress Map; the average angular difference between measured and predicted directions defines the SHmax misfit; - the tectonic regime (extension, compression, strike-slip) in the continuum, that is compared with the focal mechanisms of the regional database; the percentage of predictions not matching the observed focal mechanisms defines the regime misfit. Among all the sets of parameters tested, we discarded the model results showing any misfit larger than the 50th-percentile value of the misfit distribution. Among remaining experiments, we selected and averaged three sets of the best 30 model results: one for Umbria (west Norther Apennines), one for Marche (east Northern Apennines) and the third for the Adriatic Offshore. The best 30 model results were chosen according to the SHmax misfit in each area; their averages constitute our results and define our best model for Umbria, Marche, and the Adriatic Offshore. The results show a clear difference in the predicted rheology, strenght and strain rates for the 3 areas. We found that the integrated lithospheric strenght increases moving eastward from Umbria to Adriatic Offshore. The best rheological profile for Umbria is of the "creme brule" type, with integrated litospheric strength concentrated in the upper crust. Instead, the rheological profile of Marche and Offshore is characterized by a strong mantle and by the deepening of the brittle - ductile transition in the crust. The strain rate in Umbria is more than one order of magnitude greater than the other 2 areas. We conclude that the sharp change in the lithospheric strenght beneath the Apenninic chain has a major role in localizing the strain rate and has to be accounted for in strain-rates and slip-rates studies. S1 – UR 5.03 : Temperature measurements in Italy Carafa M.M.C1, Megna A.2, Bellani S.3, Barba S.4 1 Istituto Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605 Roma, [email protected] 2 Istituto Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605 Roma, [email protected] 3 Istituto di Geoscienze e Georisorse - CNR, Via G. Moruzzi 1 Pisa, [email protected] 4 Istituto Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605 Roma, [email protected] The mechanical behavior of the lithosphere largely depends on temperature, which controls the brittle-ductile transition depth and the cutoff depth of seismicity. We review here the temperature data and measures for Italy. We discuss the observed temperature at depth and present the isotherms, obtained interpolating existing data, for the seismogenic zones of Italy. Results show that isotherms on the Adriatic side are deeper than the Thyrrenian side. This feature is related with rheology and seismicity. Hot Tyrrhenian lithosphere is one order of magnitude weaker then the Adriatic one and can deform more easily in a ductile way. Deformation happens to be localized in the transition zone – the Apennines, where the well-known larger faults can release the strain seismically. Geotherms serve as an input to determine the maximum length along-dip of the seismic sources and help characterizing the seismic source zones. RU 02.01 – Update of the Italian Seismic Catalog (2003-2007): Catalogo della Sismicità Italiana CSI v2.0 Castello B.1, Di Stefano R.1, Chiarabba C.1 1 Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, [email protected] Earthquake catalogues are the basic tools that furnish parametric data for seismic hazard evaluation, studies on evolution of seismic sequences and earthquake occurrence. The INGV seismic network covers a large part of the italian region and it is complemented by several regional permanent network handled by other institutions. CSI results from the combination of INGV seismic bulletin with bulletins produced by other institutions. To update CSI from previous release to version 2.0 we collected seismic bulletins sent to INGV from, at present, 12 institutions managing permanent seismic networks during 2003- 2007. Procedures to convert different file formats to PHS format as input files of Hypoellipse program have been setup that also perform preliminary checks on possible errors. To correctly merge different seismic bulletins it is mandatory to have a strict control on phase associations. To do this, additional procedures to identify earthquakes external to the interest area and wrong associations of different earthquakes based on geographic control network associations and stations' residuals after event location have been produced. About 35,000 eartquakes, a mean value of ~6600 earthquakes per year, with more than 600.000 P-phase arrivals and more then 250,000 S-phase arrivals have been collected and located with Hypoellipse. To enhance final locations quality we applied a multiple location approach and then selected the best among several calculated hypocentres for each event. Location criteria are based on the use of two different weighting schemes for stations' distance combined with several reference regional 1D velocity models. Native Ml Magnitudes from 2003 to 2007 are retrieved and then associated to the corresponding event from INGV bulletin. When native Ml is not available, Ml based on regression law by Castello et al. 2007, is attributed. The update of CSI catalog, also due to the strong improvement of the INGV seismic network between 2005 and present, offers a more complete image of the Italian seismicity and a new important reference for further studies on the evolution of this region and earthquake occurrence. R.U. S1_3.02: Active kinematics and dynamics in NE Sicily: new geological constraints for the location of the seismogenic sources within the 932 Seismogenic Zone S. Catalano1, G. Romagnoli1, G. Tortorici1 & L. Tortorici1 1.Dipartimento di Scienze Geologiche – Università di Catania [email protected], [email protected], [email protected], [email protected] The seismicity of NE Sicily is to frame into the kinematics of a mobile crustal block, which is moving apart, towards the NE, from the rest of the Sicilian collision zone. Since 600 ka B.P., this mobile block has been also characterized by a 1.1 mm/a tectonic uplifting, higher than in the adjacent regions (<0.8 mm/a). This difference in the uplift-rate has been totally accommodated along a N100 oriented fault zone that, extending from Naso to Rocca Novara, cumulated a vertical displacement of about 200 m. Along the Tyrrhenian coast of the block, the tectonic uplifting has produced a flight of Late Quaternary (< 580 ka) marine terraces that extends almost undisturbed over all the major on-shore fault zones. In the Peloritani Ridge, the flight of terraces consists of distinct polycyclic abrasion surfaces (OIS 11; OIS 9 and OIS 5) that are bordered by N20-40 oriented normal faults. These are confined to the southwest by a roughly NW-SE transfer zone (Rodì Milici) that, composed of N-S and N130 oriented faults, can be prolonged to the discrete NW-SE oriented Salina- Lipari-Vulcano shear zone. The transfer zone and the Peloritani faults form the southwestern and southeastern boundaries of a Late Quaternary (<600 ka) tectonic depression, respectively. Along the Peloritani margin of this basin, the on-shore faults cumulated vertical displacements for discrete periods of the Late Quaternary, during which an active abrasion surface developed on their stable hangingwall, while the footwall was uplifting at the rate of the regional signal. This suggests a gravity-induced motion along the faults, accommodating the tectonic uplift. As consequence, the active boundaries of the basin have to be confined in the off-shore of the terraced regions, while the relict branches of the margins are uplifted in the on-shore. The kinematics of N20 oriented (pre-125 ka) fault segments indicate that fault planes have been activated by both a NW-oriented and a NE-oriented direction of extension. In addition, the post-220ka Eolian fault belt also displays kinematics which are related to a NE-oriented compression. The contribution of three different stress fields on the remobilizing the active structures, allow us to relate the apparently contrasting available seismological and geodetic information, to the deformation around the tectonic depression located between the Eolian Islands and Milazzo. In particular the major seismicity of the area (e.g. 1978; 1999) can be referred to distinct moments of remobilization of its southwestern border, from Salina to Rodì Milici. In addition, the low-magnitude seismicity suggests that the NW-oriented extension causes local reactivations spread on the whole NE-Sicily mobile block. Seismicity and Bouguer anomalies distribution along the CROP03 and CROP11 regional seismic reflection profiles Ciaccio M.G.1, Barchi M.R.2, Mirabella F.2, Pauselli C.2 1Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata, 605-00143, Roma, [email protected] 2Università di Perugia, Dip.to di Scienze della Terra, Piazza Università – 06123, Perugia We reconstructed the map of seismicity and Bouguer anomalies along the central part of the CROP03 -Umbria-Marche Apennines- and CROP11 -Lazio-Abruzzo Apennines- regional seismic reflection profiles and of the area included between them to verify the geometrical and mechanical relationships between the main decollement levels of the apenninic chain and the seismicity cut-off. We analyzed and correlated the distribution of seismicity (in map and depth), the thickness of the seismogenic layer and the resulting regional variation of the seismicity cut-off with the distribution of the Bouguer anomalies (further indicator of the crustal rheological properties). We tried to identify the lithological/structural boundaries and the analysis of the seismicity cut-off and the study of the distribution of the Bouguer anomalies to contribute to definition of the boundary between brittle and ductile behaviour (B/D transition) in the crust and the analysis of its horizontal and vertical variation. UR 02.02 – Structure of the Crust (Integrated 3D model) Part 02: Regional Seismic Tomography of Italy Ciaccio M.G.1, Di Stefano R.1, De Gori P.1, Chiarabba C.1 1 Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, [email protected] High resolution seismic tomography and a detailed map of the Moho beneath Italy and surrounding regions are key elements to enhance quality and resolution of information about the structure of the crust and its physical state and to get an advanced starting point to obtain new more detailed velocity models, both at regional and at local scale. Di Stefano et al. (2009) obtained a regional high resolution 3D P-wave velocity model, by inverting 165,000 P-wave arrivals, from 8,206 earthquakes, recorded between 1988 and 2002 at about 600 seismic stations. To enhance the resolution of such model we extended the dataset by including a selection of earthquakes from the preliminary update of the CSI 2.0 (RU 02.01). About 130,000 P-wave arrival observations from ~9.000 earthquakes recorded in the period 2003-2006 have been added. Additional events have been strictly selected for location quality (azimuthal gap < 180°; horizontal error <= 2km; vertical error <= 4km; rms < 1s) and a number of P observations great than 8. We here present the updated tomographic P-wave velocity model of Italy. Preliminary results confirm the main structural features in the best resolved parts of the inverted volume and show a better resolution in some of the previously less resolved areas, due to both the larger number of inverted phases and the more even distribution of seismic stations. We also compare the 3D velocity model with the updated map of the crust-mantle boundary, as a test study to introduce the Moho topography as an a-priori information for the next step of the 3D tomographic inversion. In addition we present preliminary results of the “Software module 1” for re-gridding and integration of local tomographies with the regional tomographic model and of the “Software module 2” for best 1D model calculation from the 3D regional one. Uncertainty of geologic and geodetic observables computed via FEM simulation of Italy Conti L.1, Carafa M.M.C.2, Barba S. 3 1 INGV, Via di Vigna Murata 605 Roma, [email protected] 2 INGV, Via di Vigna Murata 605 Roma, [email protected] 3 INGV, Via di Vigna Murata 605 Roma, [email protected] We modelled slip rate and stress rate in Italy using the finite-element code SHELLS, including tractions, faults and realistic rheology in a two-layer grid (crust and lithospheric mantle) with laterally-varying thickness, heat flow and topography. Starting from several sets of boundary conditions and input parameters, a large number of models have been generated and for each one we have evaluated: velocities, strain rates, stresses, slip rates, etc.. The numerical error that affects the predicted quantities is often underestimated, as not all the sources of error can be effectively accounted for. To evaluate the reliability of numerical predictions, for each variable of interest we calculated a “misfit” parameter defined as the difference between the value estimated by the model and the experimental data. None single “best” model exists that simultaneously minimizes all the misfit differences. To reduce arbitrariness in choosing a single best model, we have defined a “synthetic index” (Isynth) as the weighted average of all available misfits normalised to the respective standard deviation. The differences between the experimental errors of data collected by different authors, and the limited extension of the geographical areas where some datasets have been collected, suggest to weight differently misfit associated with different datasets. The ensemble of the best 100 models with the lower Isynth values has been selected and the value of each variable of interest has been averaged on this ensemble. In principle all known dataset available in literature should be included in the comparison procedure, but, since the value of weights is non a priori known, the inclusion/exclusion of each dataset in the comparison procedure can influence the Isynth definition and consequently can also partially change the best 100 model ensemble. To evaluate this issue we have analysed the dependence of the Isynth definition from different choices of a threshold criterion (to select acceptable models), and from the inclusion of each single dataset one a time. We have observed that, although the intrinsic stability of the Isynth value for a given model can not be guaranteed, however the ensemble averaged value of observables, is stable (within the measurement errors) with respect to the inclusion/exclusion of each one dataset a time. For each observable the experimental error and the numerical uncertainty are of the same order of magnitude. The suggested approach gives slip rate and strain rate values compatible with all the observations available in literature. UR1.0.3 - Strain-rate and block modeling of dense GPS velocity fields D’Agostino N.1, Serpelloni E. 1, Avallone A.1, Cavaliere A.1, Cheloni D.1, D’Anastasio E.1, Mastrolembo B.1 1 Centro Nazionale Terremoti, INGV, Via Vigna Murata 605, 00143 Roma e-mail: [email protected] The rapid advance of space-geodesy techniques, and in particular of the GPS, is providing a large amount of information on crustal motion at scales ranging from plate motion (~103 km) to single-fault strain accumulation (~100 km). These data are triggering a rapid improvement in the understanding of different processes such as the sudden release of coseismic deformation during earthquakes to the slow deformation of the continents and require new methodological tools to extract information from a given area. In this work we propose the preliminary results of two innovative approaches to analyse and interpret the dense GPS velocity field of the Italian territory and surrounding regions. These complementary approaches increase the definition of the contemporary crustal deformation in terms of diffuse strain rates and slip rates on faults. In the first approach (Haines and Holt, 1993; Kreemer and Hammond, 2007) we model the crustal horizontal strain rate field under the assumption that the crust deforms as a continuum. In the second approach (McCaffrey, 2007; Meade and Hager, 2005) we divide the region in fault- bounded blocks and solve for the rotation of the blocks and the magnitude of the style of slip on the bounding faults. One advantage of the continuum strain rate modeling approach is that no knowledge of the location and geometries of blocks and faults is needed, and a smooth estimate of the deformation field is provided. The drawaback of the continuum approach is that many processes including earthquakes, may be constrained to faults. We analyze data from more than 900 Continuous GPS stations in the Euro- Mediterranean region by means of the GAMIT, GIPSY and QOCA software. We estimate velocities for more than 600 CGPS stations with an observation time span larger than 2.5 years and used the horizontal velocity field and (1) an elastic block modeling and (2) continuum strain rate approaches to study the first order features of micro-plates kinematics and block-bounding faults slip-rates in the central Mediterranean region. RU T.02 3D geological modeling supporting seismic hazard assessment Chiara D’Ambrogi1, Mariapia Congi1, Marco Pantaloni1 1Servizio Geologico d’Italia - ISPRA, Via Curtatone, 3, [email protected], mariapia.congi@ isprambiente.it, marco.pantaloni@ isprambiente.it The three-dimensional geological modeling techniques represent an essential tool to understand and analyse crustal and sub-crustal structures and a fundamental support to the seismic hazard assessment. The 3D modeling softwares are able to process and display great amount of data, both geological and geophysical; through a full space analysis, they enable the integration and comparison of multi-scale multiple parameters useful in the definition of the seismic hazard (i.e. DISS database, seismic sections, earthquake locations, crustal structures). Moreover they allow to verify the model consistency related to the available data and to rebuild or update the model owing to information updating. The objectives of the Research Unit are to support the processing of data collected within the S1 Project, and to share and disseminate the 3D elaborations deriving from those data, pairing the construction of the Geodatabase, and supporting the models validation. First of all the seismogenic sources (individual and composite) availables in DISS (version 3.1.0) are modeled in three-dimension, using the interpolation tools of 3D modelling software. The result is a 3D imagery of seismogenic structures for the entire Italian region. This activity is carried out in collaboration with the RU 3.12. The 3D representation of DISS is integrated with the CSI 1.1 (1981-2002) and CSI 2.0 - preliminary version (2003-2007) where earthquake location parameters and related magnitude are stored. These data are supplied by the RU 2.01. These two datasets are coupled with a preliminary 3D elaboration of the structural model for the lithosphere-asthenosphere system (1°x1° cell), deriving from the activity of RU 2.04. Volumes characterized by the thickness (h), Vs, Vp and Density values assigned to the corresponding layer in the original dataset and describing the main characteristics of the lithosphere-asthenosphere system are built. The upper portion of this model will be compared with the detailed 3D crustal model for the entire Italian region (RU 2.02) and with some detailed model for specific area (e.g. RU 3.14). The comparison within these datasets will highlight the existing, both geometrical and parametrical, inconsistencies in the defined models supporting the definition of a more comprehensive model. Finally the 3D modeled data and bi-dimensional surfaces representative of specific parameters variation (i.e. heat-flow, gravity anomalies, strain rate), stored in both Geodatabase and 3D space, can be easily retrieved, and used to perform analyses and comparison with other digital geological datasets, supporting the specific activities of Civil Protection Department. RU_6.02: Combined in-land (off-fault paleoseismology) and off-shore (marine paleoseismology) tsunami deposits studies, for hazard assessment in Eastern Sicily and Southern Calabria De Martini P.M.1, Polonia A.3, Barbano M.S.2, Smedile A.1, Pantosti D.1, Gasperini L.3, Bellucci L.3, Pinzi S.1, Gerardi F.2, P. Del Carlo4, Pirrotta C.2 1. INGV, Sezione Sismologia e Tettonofisica, Via Vigna Murata 605, 00143 Roma, (RU 6.02) [email protected], [email protected], [email protected], [email protected] 2. Dipartimento di Scienze Geologiche, Università di Catania, Corso Italia 55, 95129 Catania, (RU 6.01) [email protected], [email protected], [email protected] 3. CNR-ISMAR Sezione Geologia Marina, Via Gobetti, 101 40129 Bologna, (RU 3.08) [email protected], [email protected], [email protected] 4. INGV, Sezione di Pisa, Via della Faggiola 32, 56126 Pisa, (RU 6.02) [email protected] In the frame of paleo- and historical tsunami study in Eastern Sicily and Southern Calabria, the RU activity focused on two subjects: 1) in-land tsunami deposits identification, characterization and dating (in collaboration with RU 6.01); 2) investigation of the geological signature of tsunami occurrence in the submarine environment: analysis of any subtle anomaly in the sediments (fauna assemblages, physical properties, etc.) in cores collected offshore (in collaboration with RU 3.08). As for the offshore research, we sampled two cores: the first one in the Augusta bay (110 cm at 70 m water depth) and the second offshore Crotone (150 cm at 50 m water depth). Our idea is to replay the experience done with the core MS-06 collected offshore Augusta during the previous INGV-DPC agreement. Core MS-06 was collected at 72 m bsl and it comprises 6.7 m of homogeneous, fine-grained mud interrupted at ~3 m below the top by a tephra. Several laboratory analyses (X-Ray imaging, sedimentological, micropaleontological, physical properties) were performed. Eleven anomalous concentrations of displaced epiphytic foraminifera were found. These 11 layers are often characterized by anomalous high amount of very fine-fine sand, shells and vegetal remains as well as low values in the magnetic susceptibility log. Because those epiphytic foraminifera are quite common in infralittoral zones on vegetated substrates, but not expected at 70 m bsl, we advanced the hypothesis that these anomalous concentrations might be related to tsunami waves causing substantial uprooting and seaward displacement of Posidonia. The time constraints (14C,210Pb and 137Cs and tephrochronology) yielded an age of 4500 yrs for the core. Dating results suggest that five of the eleven concentrations are embedded in age intervals that encompass the date of major tsunamis that hit eastern Sicily (1908, 1693, 1169) or the broad Eastern Mediterranean (Crete in AD 365 and Santorini at ~3500 BP). As for the in-land research of tsunami deposits in Southern Calabria we investigated the S. Eufemia plain (3 sites) and the Sibari plain (5 sites) through detailed field surveys and core campaigns. Preliminary results of the sedimentologic/stratigraphic and paleontological analyses do not indicate any evidence for deposits possibly related to tsunami waves. We should mention that all the coastal lagoons and low lands investigated present little amount of fine sediments, usually in the uppermost m of the cores. This setting could be influenced by significant vertical movements of the coastal areas related to local and/or regional geodynamics. Analysis of anisotropic macroseismic fields De Rubeis V.1, Sbarra P.1, Tosi P.1 and Ferrari C. 1Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, [email protected] A macroseismic survey system, based on a web questionnaire, is running at INGV since 2007. Reported effects are statistically analyzed to extrapolate Mercalli-Cancani-Sieberg and European Macroseismic Scale intensities. The final result is the definition of the intensity degrees, averaged for each town or village, with the evaluation of the associated uncertainty. Maps of macroseismic intensity are displayed on-line in almost real time and continuously updated in case of availability of new data. Since 2007 more than 106000 questionnaires were compiled providing more than 1000 intensity maps. We received even questionnaires of felt effects in Italy from some Greek events with high magnitude. Web intensity collection gives a high spatial density of data in case of medium-large magnitude events: this feature allows the definition of interpolated macroseismic fields, with evaluation of noise component. In several cases resulting macroseismic fields are not isotropic. The agreement with ground motion parameters indicate that the anisotropic pattern is real. The study of these fields shows that the cause is to be searched in both geological setting of the territory and source mechanisms. DPC-S1, UR1.04 Geodetic deformation field in the Italian area Resp. R. Devoti Team: M. Anzidei, A. Cavaliere, A. Esposito, E. Flammini, A. Galvani, G. Pietrantonio, A. Pisani, F. Riguzzi, E. Sepe, E. Serpelloni The Italian peninsula area is characterized by a complex set of independent or partially independent crustal blocks, in the framework of the slowly converging African and Eurasian plates. In this region different GPS networks coexist and may be useful to solve the present tectonic processes at different scales. The most promising is the RING network (Real time Integrated National GPS network) a large network of more than 130 permanent GPS sites, nationally distributed, and conceived especially for the crustal monitoring of a large seismogenetic area. A number of local GPS networks (FREDNET, IREALP, ASI and other governmental and private networks) are also useful to improve the spatial resolution of the kinematic field whereas the EUREF sites constitute, as usual, the backbone for the reference system realization of the chosen area. The RU-1.04 “Geodetic deformation field in the Italian area” archives all the available GPS data in the Italian area for at least the last 10 years and regularly process all the data at two different analysis centers at the INGV (Roma-CNT and Bologna-CNT) producing daily coordinates for all the GPS sites. The data are reduced with two different processing softwares (GAMIT and BERNESE), using only slightly different approaches. Over 400 permanent sites are routinely processed forming the input to build independent and combined coordinate time series. We show an example of obtained individual site time- series and a set of combined time-series used to test our combination strategy. The combination test covers one year of GPS data of a cluster of 40 sites, in which each daily solution from the two softwares are combined into a consensus solution that represents the average network position obtained with the full covariance matrix. The combined time- series should in principle minimize the software dependent error sources in the time-series and represents, in some sense, the optimized solution of the GPS position time evolution. The two preliminary velocity solutions (GAMIT and BERNESE) are compared in their planar components giving an overall RMS difference of only a few fractions of mm per year in the worst case. Occasionally large deviations at certain sites may reveal problems in the data reduction phase of one or both the analysis centers and therefore this comparison/validation phase is really crucial to clean out the analysis bugs. The feedback from the combination process may trigger a re-analysis of the GPS data that iterates until all the individual network solutions converge towards a common ‘consensus’ solution. A Near Real-Time Tool for High Rate GPS and GPS-Seismology Resp. R. Devoti Serpelloni E. 1, Devoti R. 2, Cavaliere A. 3 1 Istituto Nazionale di Geofisica e Vulcanologia – Centro Nazionale Terremoti, Bologna, [email protected] 2 Istituto Nazionale di Geofisica e Vulcanologia – Centro Nazionale Terremoti, Roma, [email protected] 3 Istituto Nazionale di Geofisica e Vulcanologia – Sezione di Bologna, Bologna, [email protected] Four days before the 6th April M5.8 l’Aquila main-shock, a few GPS receivers recording at 10Hz and 1Hz sampling rates have been set up by INGV in the area affected by the seismic swarm in place by mid-January 2009. These data allow us to measure for the first time in Italy the dynamic co-seismic displacements with periods ranging from fractions of seconds to several minutes and the full time spectra of the surface co-seismic deformation with GPS instruments. We use TRACK, the kinematic module of the GAMIT/GLOBK software package, to perform epoch-by-epoch solutions of GPS raw data and obtain 3D time series of surface displacements. TRACK uses floating point LC (L3) observations between pairs of stations and the Mebourne-Wubena Wide Lane combination, with ionospheric constraints, to determine integer ambiguities at each epoch. It requires a fixed station and one, or more, kinematic stations. Usually, the static station is chosen to be far enough from the epicentral area not to be affected by the co-seismic displacements. So far no automatic processing engine exists for the TRACK module, so we built a new shell script, which take full advantage of the routinely used Linux CPU-cluster to analyze 30 seconds GPS data. The new tool allows to automatically process pairs of stations (i.e., a network) and getting raw time series of several stations simultaneously (depending on the number of cluster nodes available) in a few seconds or minutes, depending on the length of session analyzed. Our new processing tool allow the use of broadcasted, ultra-rapid (containing predictions), rapid and final IGS orbits, thus making quasi-real time processing possible, and actually limited by the access to raw high-rate GPS data remote archives. Since that the only two stations recording 10Hz data in the L’Aquila region, are located close to the main shock epicenter and no data were available at other sites in Italy, we built a new tool to generate a virtual far field reference station acquiring 10Hz data by interpolating the available 1Hz RINEX data. The interpolated sites permit to properly solve the epoch-by-epoch position of the epicentral sites with the TRACK module. High frequency GPS data are severely affected by multipath noise, which can reach the same magnitude of the co-seismic displacements, and need to be removed consistently. For this reason, we investigate the effect of time and space-wise filters (sidereal and common mode filters) and set up a Matlab tool to perform time and spatial filtering on the raw time series produced by our processing tool. Project S1_UR_3.03 - Active tectonics in Calabria. Part I: Structural and seismic data: preliminary results. Faccenna C. 1, Neri G. 2, Billi A. 1, Minelli L. 3, Olivetti V. 1, Orecchio B. 4, and Presti D. 2 1. Università degli Studi Roma Tre, L.go S. L. Murialdo 1, 00146 Roma, Italy, [email protected] 2. Università degli Studi di Messina, Sant'Agata - Salita Sperone, 31, 98166 Messina, Italy, [email protected] 3. Università degli Studi di Roma “La Sapienza”, P.le Aldo Moro, 5, 00185 Roma, Italy, [email protected] 4. Università della Calabria, Ponte P. Bucci, cubo 12b, 87036 Arcavacata (CS), Italy, [email protected] Calabria and eastern Sicily are seismically-active regions, where several destructive earthquakes (M ≥ 7) were recorded during historical times. Despite the documented inheritance of destruction and death, the source (i.e., the causative faults) of most destructive earthquakes in these regions is still unknown or poorly constrained. Here, we have focused our attention on some seismically active zones of north-eastern Sicily and Calabria to find possible (seismically-) active faults or, at least, to constrain some seismotectonic features of these zones. We have mainly used seismological and field structural methods coupled with and corroborated by other geological and geophysical evidence, where available. In particular, structural analysis has been carried out over the main recent faults to extract the recent slip direction. We measured kinematic indicators, as slickensides, striae, grooves and, only locally fibers, along both main and subsidiary fault planes of the main fault zones. Data are implemented and compared with previous studies carried out using the same criteria. Along this fault we carried out quantitative analyses of stream and river profile, and other geomorphological features. Seismological analysis has been carried out by relocating earthquakes recorded during the last twenty years using new tomographic models. Non linear method (Bayloc) is used to define new and reliable plane solutions, T and P axis orientation and to define the seismic temporal sequence. We focus our attention on four key areas: 1) The Fault zone between Tindari and the Aeolian Islands, 2) the seismic zones between Mt Etna and Cefalù, 3) the region of Le Serre-Capo Vaticano, 4) Crati region and southern Sila massif. Our preliminary results show a good agreement between fault geometry and Bayloc distribution and between T axis and fault kinematics along the arc. In detail, this integrated analysis show some complex features where future detailed investigations will be carried out. Fore example, in region (3) a more detailed analysis is necessary to better understand the role of Capo vaticano EW bearing fault system; in region (4) a specific studies is necessary to understand the location of the seismic sequences October 2001 and their bearing with the southern Sila structures. Project S1_UR_3.03 - Active tectonics in Calabria. Part II: Structural and reflections seismic data: preliminary results. Faccenna C. 1, Pepe F. 2, Minelli L. 3, and Olivetti V. 1 1. Università degli Studi Roma Tre, L.go S. L. Murialdo 1, 00146 Roma, Italy, [email protected] 2. Università degli Studi di Palermo, Via Archirafi, 20/22, 90123 Palermo, Italy, [email protected] 3. Università degli Studi di Roma “La Sapienza”, P.le Aldo Moro, 5, 00185 Roma, Italy, [email protected] We carried out structural analysis over the main recent faults to extract the architecture and the recent slip direction along the Calabrian margin from Sila to northern Sicily. We measured kinematic indicators, as slickensides, striae, grooves and, only locally fibers, along both main and subsidiary fault planes of the main fault zones. To obtain a significant statistical orientation, we collected several measurements for each fault zone. Overall, we collected more than 200 fault kinematics. Data are implemented and compared with previous studies carried out using the same criteria. Along this fault we carried out quantitative analyse of stream and river profile, and other geomorphological features. Our result shows a shift in the stretching direction moving north of the Catanzaro basin within the Sila massif and Crati basin, where extensional direction trends NNE-SSW or NE-SW. South of the Catanzaro through, in the Serre massif the extensional trend changes and appears dominated by a well defined NW-SE stretching direction, as previously ( Tortorici et al., 1996). In this view, the Capo Vaticano E-W fault system represent an accommodation zone. West of the Messina through, in northern Sicily, the most recent fault system slightly turns again into a more EW stretching direction. We analyze fault pattern off-shore Calabria using the unpublished Sister-D1 multi-channel seismic line (Bertotti et al., 1999) that runs parallel to the W Calabrian coast. This profile is particularly relevant to analyze the possible source of the September, 8, 1905 earthquakes that is probably localized between Capo Vaticano and the Eolian Island (e.g., Michelini et al., 1995; Galli et al., 2009). Our preliminary result show the possible putative fault structure that border to the south the off-shore prosecution of the Capo Vaticano uplifted block . RU3.04: Holocene tectonic displacements along the Calabrian arc coastline: constraints on unknown Holocene earthquakes and offshore seismogenic structures. Ferranti L.1, Monaco C.2, Antonioli F. 3, 1 Dipartimento di Scienze della Terra, Università Federico II, Largo S. Marcellino 10, Naples, Italy, [email protected] 2 Dipartimento di Scienze Geologiche, Università di Catania, Corso Italia 55, Catania, Italy, [email protected] 3 ENEA, Via Anguillarese, Casaccia, Rome, Italy. The study of new outcrops integrated with a revision of published data provided an unprecedented understanding of the Holocene deformation occurred along the Calabrian arc coastline (Calabria and eastern Sicily). A total of ~75 vertical tectonic displacement rates coming from~15 different sectors of the Calabrian arc were stipulated based on radiometric ages or accurate archeological dating of Late Holocene (<7 ka BP) outcrops, and in few instances early Holocene (>7-10 ka BP) boreholes or submerged caves markers. Our compilation allows to grasp the spatial extent of the uplifting region with a remarkable spatial control in southern Calabria, northeastern and southeastern Sicily. In four locales, based on the finding of superposed shorelines, we distinguished between a regional (deep-seated) vs. local (upper-crustal) component of uplift, the latter one yielding constraints on unknown Holocene earthquakes. At least 3 co-seismic events were already established in the Taormina (at 5, 3.2, and <3-2 ka) and Scilla (at 5.1 (?), 3.5 and 1.9 ka) regions, facing the opposite sides of the Messina Strait (De Guidi et al., 2003; Ferranti et al., 2007). These events caused vertical tectonic displacements of between 1.5 to 0.8 m. Our research established ancient earthquakes also at Milazzo on the northern shore of Sicily, with two events occurred between 3.8-1.6 and <1.4 ka, and having vertical component of displacement of ~1 m and ~0.5-0.8 m, respectively. On the Tyrrhenian coast of central Calabria, at the southern side of Capo Vaticano, mapping of superposed shorelines yielded two episodes of co-seismic uplifts between 5.5-3.9 ka (~0.70 m) and after 3.9 ka (~1 m). Scattered observations suggest also Holocene co-seismic events in NE Calabria (coasts of Pollino and Sibari plain). Sources for these earthquakes are poorly established, because they reside largely offshore. In the case of Capo Vaticano, located across a prominent normal fault systems, our data suggest that the Coccorino strand is locked or inactive, whereas the Nicotera strand may be active, in agreement with seismicity data. Offshore normal faults were already proposed as the cause for the co-seismic events detected at Taormina and Scilla (De Guidi et al., 2003; Ferranti et al., 2007). In NE Calabria, transpressional faults mapped both onshore and offshore could account for folding of ~600 ka marine terraces. Analysis of seismic reflection profiles and numerical modelling indicate that about 20% of the total uplift of these terraces can be ascribed to fault-related folding and transpressional fault ruptures. RECURSIVE GPS VELOCITY FIELD Flammini E.1, Devoti R.1 1 Istituto Nazionale di Geofisica e Vulcanologia,[email protected], [email protected] We implemented an iterative method to compute an estimate of the velocity field of a GPS network by adding recursively the station coordinates. The velocity field is obtained as the solution of the restricted least square problem. At each epoch ti for which we have observations xi with covariance matrix Σi, we estimate, via a least square approximation, the velocity field of the entire network upgrading the apriori velocity field calculated at the epoch ti-1. In this way we obtain an updated velocity field for each epoch leading to a velocity time series that may be useful for a rough estimation of new site velocities and for the study of the station velocity stability. To validate this method, we compare the recursive velocity field of a network of 20 GPS sites, with the velocity field obtained fitting all observations at once in a standard way. The test solutions, obtained with the Bernese software, cover a time span of 2 years (2007- 2008) and the two estimated velocity fields agree at the level of 0.05 mm/yr. Locating and sizing historical earthquakes by the attenuation of macroseismic intensity with distance: A new version (4.0) of Boxer code. Gasperini P. 1, Vannucci G. 2, Tripone D. 2 and Boschi E. 1,2 1. Dipartimento di Fisica, Università di Bologna 2. Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna About ten years after the first release of Boxer code we implemented new methods of epicentral location and magnitude estimation from macroseismic data as well as the consistent evaluation of parameter uncertainties. The improved code allows to locating epicenters in the sea or in uninhabited areas by maximizing the likelihood function of an attenuation equation with observed intensity data. It also allows, in the most favorable cases, to estimating source depth and local attenuation coefficients. The code computes errors in two ways: i) by formal methods, like the inversion of the Hessian of the log- likelihood function at the maximum or the standard deviation of parameter mean, and ii) by bootstrap simulations. We tested the performance of the old and new methods by a comparison with reliable instrumental data, finding a rather good agreement for epicentral location and magnitude but not for hypocentral depth. We also found a good correspondence between fault-trace orientations computed by macroseismic data with focal mechanisms of earthquakes with Mw≥5.7. The new version of Boxer code, which includes a user-friendly interface for data selection, parameter setting and graphical post- processing, will be freely downloadable from the web soon. UR 4.01: Seismogenic sources of major earthquakes of the Maiella and Abruzzo foothill area (central Italy): constraints from macroseismic field simulations and regional seismotectonic data. Lavecchia G.1, de Nardis R.12 1. Laboratorio di Geodinamica e Sismogenesi, Dipartimento di Scienze della Terra, U d’A Chieti, Campus Universitario, 66013 Chieti Scalo, Italy, [email protected] 2. Dipartimento della Protezione Civile, Via Vitorchiano4 00195 Roma, Italy [email protected] We apply a grid search-based inversion scheme to evaluate the source parameters of the destructive 1706 (Io=X-XI MCS), 1933 (Io=IX MCS) and 1881 (Io=VIII MCS) Maiella earthquakes. To constrain the inversion procedure and to obtain more reliable solutions, we have preliminarily perform: 1. the identification of all the plausible seismogenic source models, selected on the basis of integrated seismological, geological and geophysical data; 2. the measurement of the ambient noise in 37 sites around the Maiella Massif for the estimation of local response using the Nakamura method. The geophysical and geological information allows us to identify 5 alternative seismogenic source models (Abruzzo Citeriore Basal Thrust ABT, Maiella Basal Thrust MBT, Morrone- Porrara esxtensional fault system MPEF, Caramanico normal fault CNF, E-W strike-slip discontinuities EWF) and to define, for each one of them, a 3-D grid representing the space of variability of the hypocentre and of seismogenic fault plane parameters. The results obtained from the microtremor measurements show H/V curves with a sharp pick at the fundamental frequency, indicating strong impedance contrasts at depth. These results have been qualitatively correlated with local geology and available geotechnical data and opportunely considered in the inversion procedure (de Nardis et al., 2009). The numerical modeling has been carried out by means of a stochastic approach for finite faults (Beresnev and Atkinson 1998 and references therein). For each nucleation point of the search-grid of each source model, we have generated synthetic ground motion waveforms at all the macroseimic sites around the considered earthquake (1706, 1933, 1881). Consequently, we have performed a grid search generating 2970 scenarios for the ABT model, 1458 for the MBT model, 1890 for the MPEF model, 81 for the CNF and 1280 and 900 for two selected EWF. We have converted calculated peak ground acceleration into intensities using empirical relationships and performed the inversion minimizing the data distance between calculated and observed intensity values in terms of L1. Because the conversion from peak ground acceleration to intensity is a very crucial point and can affect the final solution, we have used and compared three empirical relations: the one proposed by Trifunac and Brady (1975), the new ones proposed by Faccioli and Cauzzi (2006) calibrated on Italian data and the Faenza and Michelini (2008) relations calibrated on the same data set and obtained with orthogonal distance regression. Moreover, we simulated the destructive 1706 Maiella earthquake assuming two different values of the equivalent moment magnitude, derived from the Working Group CPTI 2004 (Mw 6.6) and the Working Group CPTI08 (Mw 6.8) catalogues. We discuss the results obtained from the global and the partial variance and compare them with those which may be obtained by applying to the same earthquakes a relationship which does not depend on any modelling assumptions and only considers the macroseismic epicentral location. Beresnev, I. A. and Atkinson, G. M., 1998, FINSIM - a FORTRAN program for simulating stochastic acceleration time histories from finite faults, Seism. Res. Lett. 69, 27-32. de Nardis R., Lavecchia G., Pace B., Visini F., 2009. Exploring the best seismotectonic solution from macroseismic data: the 1781 Cagli earthquake case study. Geophysical Journal International, Accepted. Faccioli, E. and Cauzzi, C., 2006. Macroseismic intensities for seismic scenarios estimated from instrumentally based correlations, Proc. First European Conference on Earthquake Engineering and Seismology, paper number 569. Faenza, L. and Michelini, A., 2008. Shakemap per eventi storici: il terremoto di Messina- Reggio Calabria del 1908’. Extended abstract Miscellanea INGV, available on http://portale.ingv.it/servizi-e-risorse/archivio-congressi/convegni-2008/messina-1908- 2008/Convegno19082008/ Trifunac, M. D. & Brady, A. G., 1975. On the correlation of seismic intensity scales with the peaks of recorded ground motion, Bull. Seism. Soc. Am. 65, 139-162. Working group CPTI, 2004. Catalogo Parametrico dei Terremoti Italiani, (CPTI04). INGV, Bologna. http://emidius.mi.ingv.it/CPTI/. Working group CPTI, 2008. Catalogo Parametrico dei Terremoti Italiani, (CPTI08). INGV, Bologna. http://emidius.mi.ingv.it/CPTI/. RU 3.06: Update of present-day stress field data in Italy: the results of new borehole breakout analyses Mariucci M.T.1, Montone P.1, Pierdominici S.1 1.Istituto Nazionale di Geofisica e Vulcanologia, Sezione Sismologia e Tettonofisica, Via di Vigna Murata 605 00143 Roma, [email protected], [email protected], [email protected] Borehole breakouts are considered one of the most reliable indicators of present-day stress field, providing information about the orientation of the main axes of the stress tensor. In order to increase the dataset of present-day stress orientations in Italy, we have collected 62 new borehole logs and performed breakout analysis using the “standard high resolution dipmeter tool” data. Here we present the results of the analysis of the first year project and the preliminary results of the second year in terms of horizontal stress orientations (Shmin or Shmax). Concerning their locations, 46 wells are in southern Italy (7 along the Latium Tyrrhenian margin, 17 along the belt between Campania and Basilicata, 17 along the Apennine foredeep, and 5 in northern Calabria), 12 wells in eastern Sicily, 3 in the central-northern Adriatic offshore and 1 in the Po Plain. In detail, the borehole depths range between a minimum of 1296m and a maximum of 5543m, whereas the dipmeter logs available for the analysis reach the maximum depth of 5541 m. The breakout zone total length recognized over a single well varies from about 25 m to about 1100 m and the errors associated to the average Shmin orientation in a borehole, reported as standard deviation, range between 8° and 35°. Following the World Stress Map ranking system (from the best A to E quality), 40 wells have been classified B, C or D, and 22 wells have been discarded (E quality). We have split the study area in different sectors according to homogeneous regional and local stress orientation trends. In each sector, we have analyzed the new data also all together obtaining a mean trend of the area. Then we have averaged the mean orientations of all the wells, weighted by well quality, both considering only the new data and including all the published data of each sector. In general, the new Shmin orientations confirm the results from previous works, as for instance the NE-SW trend in the southern Apennine belt or the WNW-ESE trend in the Sicily foredeep, but also highlight a large local stress perturbations in some areas, as the Tyrrhenian margin. Part of the new dataset has been used by the researchers of the UR 5.03 for an integrated crustal model. S1 - RU 5.03 : Numerical modeling of the seismogenic structures Megna A.1, Carafa M.M.C.1, Barba S.1 1 Istituto Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605 Roma, [email protected] Slip and strain rates are fundamental quantities to study earthquake occurrence and their knowledge at the scale of few kilometers is an important tool for the seismic hazard studies. The goal of this RU is to reduce the uncertainty associated with the strain and slip- rates previously known for Italy. Imposing several boundary conditions, specified rheologies, and densities, the results of the finite-element models were compared with borehole breakout stress orientations, GPS velocities, tectonic regime, vertical velocities and seismicity of the areas. Results are strain rate maps and slip rates for seismogenetic sources. In the northern Apennines we found that strain rate is localized in Toscana-Umbria area, which is undergoing extension. The Marche area and the Adriatic offshore is in compression with slip rates values for seismogenetic sources of 0.1-0.4 mm/yr. As regards Mattinata fault system we found slip rates of 0.3-0.7 mm/yr. Such a fault system can be described as an element of structural weakness with a 0.4-0.5 effective fault coefficient, lower with respect to normal values (0.6). For the seismic hazard we think that an important role seems to be played by the structural heterogeneity along the fault plane. We also built a model at local scale to valuate if a more detailed crustal structure, i.e. non- stationary heat flow, non-uniform rheology, and the presence of detachments, is sufficient to obtain realistic results. In this first phase we have chosen to build a model introducing a free-slip fault, and a thin layer as possible detachment under the dislocation. Furthermore, we have imposed different velocities along the edges of the model in order to simulate an extensional regime and a probable compression zone on a lateral edge. The results show that the length of the thin detachment influences the strain rate: in particular, when the detachment is about ¾ of total length the strain rate is maximum in the range where the detachment is present. RU3.07 - Active compressive tectonics and paleoseismic evidence along the Southern Alps between Garda and Maggiore Lake: Holocene surface faulting at the Monte Netto and Borgo Vico sites Michetti A.M.1, Livio, F.A.1, Berlusconi, A.1, Sileo G.1, Chunga K.1, Serva L. 2, Vittori E. 2, 1Università degli Studi dell’Insubria. Via Valleggio, 11 – Como – Italy. [email protected] 2ISPRA, Via V. Brancati 48, 00144, Roma A belt of alpine fault-related folds, deforming the Plio-Pleistocene Po Plain basin infilling, has been mapped based on a revision of commercial seismic reflection data. A common characteristic is the Quaternary migration of active tectonic deformation from the more external south-vergent forethrusts to an inner position, driven by high angle backthrusts (e.g., Sileo et al., 2007, Livio et al., 2009). In order to characterize the relations between Quaternary capable faults and earthquake potential in the study area, based on geological and geomorphological field survey we selected two sites where recent surface deformation is well expressed, i.e., the Monte Netto of Capriano del Colle (Brescia) and Borgo Vico (Como) sites Monte Netto is an isolated hill located ca. 10 kn S of Brescia, whose relief is the culmination of a fault-related fold driven by a north-vergent alpine backthrust (e.g., Desio, 1965). Quarry works have exposed a fluvioglacial and loessic sequence bent in two decametric folds and faulted by extrados normal structures. OSL and AMS radiocarbon dating of the polycyclic soils developed on the loessic parent material, provides evidence of 3 paleo-earthquakes, whose bracketing ages span between Late Pleistocene and Early Holocene. The observed paleoseismic features (secondary surface faulting and liquefaction) are consistent with the environmental effects of an earthquake similar to the December 25th, 1222 (Io IX MCS) Brescia event, whose epicentral area includes the Monte Netto site. The BorgoVico site is located near the city of Como, along the trace of a well known alpine structure, the Monte Olimpino Backthrust (e.g., Bernoulli et al., 1989). The exposed sequence at this site is composed by the Villa Olmo Conglomerate, belonging to the Chiasso Formation of Early Oligocene age (e.g. Gelati et al., 1988), thrust over a Late Quaternary fluvioglacial and lacustrine sequence. The fault plane dips N210/65, coinciding with the Monte Olimpino Backthrust trace. Drag folds affect the laminae of the lacustrine deposits along the fault contact, demonstrating the reverse kinematics of this fault. Palaeoseismological analyses are in process in order to identify the coseismic vs. fault creep nature of the Latest Glacial to Holocene displacement. Our case studies shows how detailed field observations can provide valuable geological and geomorphological evidence of neotectonic activity by detailed mapping of capable (sensu IAEA, 2002) faults. This basic geological information has certainly to be considered a vital element for the evaluation and mitigation of seismic hazards in terms of both ground shaking and surface faulting, as the reactivation of the Paganica Fault during the recent Abruzzo seismic sequence has clearly demonstrated. Since the historical catalogue shows a low seismicity for the study area, which is not verified by the geological evidence, we argue that the seismic potential of the Southern Alps in Lombardia should be carefully re- evaluated, taking advantage of the comprehensive database of existing subsurface geological information, coupled with the systematic paleoseismological analysis of all the mapped Quaternary tectonic structures of the region. RU4.02TaskB: Source parameters of significant italian historical earthquakes within the SISMOS project framework. Palombo B.1, Bernardi F.2, Ferrari G.1, Hunstad I. 1. 1.Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata, 605, Rome, [email protected] 2.Istituto Nazionale di Geofisica e Vulcanologia, Via D. Creti, 12, Bologna, [email protected] We have analyzed 13 out of the selected 19 events of the Italian seismic history. We have paid particular attention to the Adriatic coast earthquakes, 1908 Calabro Messinese and the 1917 Sansepolcro earthquakes. The historical bulletins, the instrumental parameters and the paper seismograms for all the 13 events have been recovered. The relocation of the events for which we have the complete set of data (the P and S arrival times and the paper records to read the missing arrival times) was carried out. We have vectorialized seismograms for the recovered components of the following earthquakes: Calabria 1905, Calabro Messinese 1908, Monterchi-Citerna 1917, Senigallia 1930, Adriatic coast 1934,1938, 1962. For the earthquakes occurred during war periods, is very difficult to find the bulletins, the paper seismograms and the instrument parameters. However, for the 1916 Adriatic earthquake we have enough bulletins to estimate the instrumental location. The Sismos team has developed a new methodology to compute the moment tensor and the moment magnitude. This method has the advantage to be suitable for historical dataset, but require long time computation even using the new multi-dualcore computers. Using this technique we have calculated the moment tensor and magnitude for those earthquakes for which we have both instrument parameters and vectorialized seismograms. RU6.04. The Messina Straits 1908 tsunami: tide-gage, run-up and the causative source Alessio Piatanesi, Fabrizio Romano, Stefano Lorito Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy After a whole century, the causative source of the tsunami remains one of the most discussed issues about the 1908 earthquake. Recently, two papers reached substantially opposite conclusions and have raised the debate again. The authors of the first paper hypothesize that the origin of the tsunami should be related to a large submarine landslide; conversely, the second paper demonstrates that the run-up distribution along the Sicilian and Calabrian coast is much more compatible with a tsunami generated by a seismic fault than by a submarine landslide. A very recently paper also proposed a submarine landslide, tough significantly smaller than the former, as the main source of the tsunami. We dwell into this problem by focusing on new constraints posed by additional tsunami data, namely four tide-gage waveforms recorded at Malta, Naples, Ischia and Civitavecchia, never used so far. We use these data along with run-up data to investigate the causative source of the tsunami. We discuss several scenarios, by means of forward and inverse modeling. We conclude that proposed seismic sources are not sufficient to explain the observed tsunami, both in the near- and far-field, and additional seismic slip and/or a submarine landslide is needed. Dense Italian velocity field: strategy of analysis and preliminary strain rate map Pietrantonio G 1, Pisani A.R 1, Riguzzi F.1 1Istituto Nazionale di Geofisica e Vulcanologia, CNT, via di Vigna Murata, 605 – 00143 Roma, [email protected] Starting from 2004, the Istituto Nazionale di Geofisica e Vulcanologia (INGV) set up a dense permanent GPS network, called RING (Rete Integrata Nazionale GPS) in Italy in order to improve the knowledge of the geodynamics and tectonic processes. Today RING includes more than 130 sites, mostly built on tripod or concrete pillar monumentation and capable to observe the GPS signal at 1 Hz in real time. A subset of GPS receivers are collocated with seismometers and accelerometers and constitute a valuable infrastructure to monitor real time seismic events and to study tectonic processes of the Mediterranean area. Due to the complexity of the tectonic behaviour in the area, the INGV also arranged the data analysis of all the permanent sites available in the Italian region: RING and other sites belonging to different networks (ASI, EUREF, FREDNET, IREALP, ITALPOS and also Abruzzo, Friuli Venezia Giulia, Puglia, Sardegna, Umbria, Veneto regional networks) are included in the routinely performed analysis. In this work we present time series and velocity field obtained with BERNESE software ver. 5.0, analyzing the complete data set from 1998 to 2009 of about 400 GPS permanent stations. The processing is performed adopting a distributed session approach, with more than 10 clusters, sharing common stations, each of them consisting of about 40 stations. Daily loosely constrained solutions are routinely produced for each cluster and the velocity fields are obtained by stacking the daily normal equations rigorously. Moreover we show a preliminary Italian strain rate map, obtained adopting the algorithm described in Shen et al., 1996. An algorithm for the search of homogeneous strain rate fields Pietrantonio G.1, Pisani A.R.1 1Istituto Nazionale di Geofisica e Vulcanologia, CNT, via di Vigna Murata, 605 – 00143 Roma, [email protected] The aim of this paper is to describe the theoretical fundamentals and the main features of a software suitably implemented to estimate the strain rate tensor from continuous GPS data. Current softwares developed for geophysical applications generally estimate or compute bi-dimensional strain, since this is the most requested use. On the contrary, this software allows for a three-dimensional estimate of the strain rate tensor. It accounts for all the significant GPS velocities and estimates the strain rate components by the least squares method starting from the hypothesis of one homogeneous strain rate field. An initial field has to be defined by at least 4 sites which pass the chi-squared test on the strain rate homogeneity. The developed algorithm automatically searches for sites belonging to this initial homogeneous field, starting from the site nearest to the barycentre of the first 4 sites and proceeding until a user-defined limit distance. Each time a site is added, the homogeneity of the whole field is suitably tested by a number of statistic tests. In this work the algorithm has been also applied to some areas of geophysical interest. Terrestrial laser scanner technique: 3D reconstruction of San Lorenzo boulder (southern-eastern Sicily, southern Italy) Pignatelli C.1, Gerardi F.2, Marsico A.1, Piscitelli A.1, Pirrotta C.2, Barbano M. S.2, Mastronuzzi G.3 1Dipartimento di Geologia e Geofisica, Università di Bari, Via Orabona, 4 – 70125 Bari, email: [email protected], [email protected], [email protected] 2Dipartimento di Scienze Geologiche, Università di Catania, Corso Italia, 55 – 95129 Catania, email: [email protected], [email protected], [email protected] 3LaGAT-Ta - Laboratorio Gis geo-Ambientale e di Telerilevamento, Università di Bari, Via A. De Gasperi - Paolo IV – 74100 Taranto, email: [email protected] A boulders field is located in southernmost part of Sicily along the Ionian Sea and it constitutes the emerged part of the Hyblean foreland domain of the Apennine orogen. The area is surrounded by seismogenetic coastal zones which have produced large tsunamis in historical times. San Lorenzo is an old farm surrounded by a summer village placed on the south of Vendicari Reserve, about 30 Km south to the Siracusa city. The coastal belt is characterized by flat landscape with beaches, gently sloping rocky coasts and low cliffs weakly incised by a poorly developed hydrographical network and tectonically controlled. Between the rocky coasts, actual beaches and marshes are present. Adlittoral zone is characterized by elevate roughness and a convex profile. Here, surface micro-topography shows small karstic landforms, shaped on the sub-aerially exposed calcareous sandstones; they are represented mainly by potholes, which became increasingly deeper and wider toward the coastline. In the spray zone potholes are coalescent, giving place to pinnacle-like forms. Calcarenitic bedrock presents very long fractures that become wider toward to the coastline. Several of these fractures are parallel to shoreline; some of these are also oriented to NE. The investigated area is characterized by a boulders field testifying a huge wave impact. In particular, in San Lorenzo area, a large boulder of about 100 tons is placed close to spray zone. This block was probably removed from the near submerged rock where the detaching niche is still visible. Resent sea storms have scattered inland several boulders as testified by the presence of well preserved biological encrustation. In order to distinguish the probable mechanism of transport of these boulders a 3D model was built using terrestrial laser scanner technique. Using hydrodynamic equations it is possible to calculate the minimum wave height (for tsunami and for sea storms) able to transport boulders inland. Each boulder was surveyed with Leica Scan Station 2 using detailed scan density (horizontal and vertical spacing of 1 mm). During the field work scans from different position were carried out in order to obtain a complete coverage of surveyed area. In the data processing, the point clouds of each scan are linked together trough tie points in order to obtain a virtual reconstruction of boulders. In order to reconstruct the surface morphology of each boulder, the point clouds were elaborated and interpolated using RapidformTM Software; this software allows to calculate surface and volume of 3D model. Inland tsunami deposits study (of fault paleoseismology) for assessment in Ionian Apulia and Eastern Sicily. Pignatelli C.1, Scicchitano G.2, Piscitelli A.1, Spampinato C. R.2, Monaco C.2, Mastronuzzi G.3 1Dipartimento di Geologia e Geofisica, Università di Bari, Via Orabona, 4 – 70125 Bari, email: [email protected], [email protected] 2Dipartimento di Scienze Geologiche, Università di Catania, Corso Italia, 55 – 95129 Catania, email: [email protected], [email protected], [email protected] 3LaGAT-Ta - Laboratorio Gis geo-Ambientale e di Telerilevamento, Università di Bari, Via A. De Gasperi - Paolo IV – 74100 Taranto, email: [email protected] Recent studies highlight as tsunami waves are able to detach megaclasts weighting up to hundreds tons from intertidal/adlittoral zone and to transport them onshore at different distances and altitudes above sea level. When boulders are present along the rocky coasts it is possible to evaluate the intensity of past catastrophic waves. Terrestrial Laser Scanner TLS surveys have been performed in order to obtain the 3D reconstruction of boulders. Post-processing techniques, applied to TLS point clouds, have been used in order to obtain the dimensions, the shape, the volume of each boulder. By means of hydrodynamic equations, minimum wave height necessary to entrain and transport boulders landward, have been calculated. This calculated catastrophic wave can be used to the estimation of the coastal flooding. The inland penetration limit of a tsunami wave is influenced by coastal roughness expressed by Manning’s coefficient. The Manning number is an empirically derived coefficient, which is dependent from many factors, including surface roughness and sinuosity; it is calculated by empirical formula for open channel flow, or free-surface flow driven by gravity. The value of Manning number was estimated directly measuring the roughness of a surface by means of Terrestrial Laser Scanner (TLS). This kind of surveys generate cloud points useful to obtain digital elevation model DEM and standard deviation of surface micro-topography. Applying calculated tsunami wave heights it is possible to obtain a specific Manning number for each surface and material typology, in different scenarios. This new methodology has been tested in two coastal sectors of Ionian Sea, where evidence of past tsunami have been recognised: Torre Squillace in Apulia region and La Maddalena peninsula in Eastern Sicily. THE GEODATABASE OF S1 UNIT RESEARCH DELIVERABLES Maurizio Pignone (INGV, sede Irpinia) Unità di Coordinamento S1 One of the activities of S1 Project Coordination Unit is the organization within a geographic database of all unit research (RU) deliverables. The goal is to provide the Department of Civil Protection (DPC) a georeferenced dataset ready for applications and processing for the study of Italian seismic hazard. The wide variety of data produced by the research units has required the development of Guidelines for the data delivery to define four different types: text files, tabular formats, geographic formats (GIS), cartographic representations or digital maps. For each research unit It has also been developed two sheets about deliverables. The first sheet, called "general", is partly precompiled and focuses on the way each research unit will deliver geographic data (dataset). Also in this sheet it is possible to choose the collaboration with the T.02 Research Unit for 3D modeling of datasets. The second sheet is detailed for a single dataset and contains the relevant information such as the general description, the geographic coordinate system, the extension of the dataset, the distribution and usability. In collaboration with the Department of Civil Protection the information enclosed in this second sheet has been coded according to the standard ISO 19115 metadata and international standards of the European Directive INSPIRE. The application of metadata in the management of the S1 project deliverables allows all participants to insert information about their datasets and share information with a common language. The geographic database has been developed in a ESRI GEODATABASE, the data model of ArcGIS 9.x. The GEODATABASE is an object oriented data model for representing geographic information using standard relational database technology. It allows to share different types of geographic data within a single container. Datasets are stored in the GEODATABASE as feature class, feature dataset, raster dataset, non-spatial tables and relationship classes RU3.09-Task1. Quaternary tectonics of the 1968 Belice Earthquake area, Sicily. Pucci S.1, Avellone G.2 , Smedile A.1 1. INGV, Via di Vigna Murata, 605, 00194 Roma, [email protected] 2. Università degli Studi di Palermo, Via Archirafi, 22, 90123 Palermo, [email protected] A preliminary study for the characterization of the surface deformation induced by active tectonics and analysis of its main parameters was performed in the Belice area. Aerial photo interpretation has been performed and the recognized geomorphic and geologic features has been digitized and implemented in the GIS environment. The significant geomorphic features related to the setting of depositional and tectonic landforms, such as paleosurfaces, scarps, slope changes, crests, peaks, landslides and tectonic lineaments were also reported. A geological/geomorphological field survey was focused on the collection of detailed information on the continental deposits. The mapping was conducted to grasp the spatial distribution and the paleoenvironmental significance of the deposits potentially associated to tectonic activity, along with the reconstruction of fault systems and gravitational processes. As a result, the character of the continental deposition and its evolution, along with the recognition of a set of marine and fluvial terraces, identify the occurrence of regional uplift. The marine Plio-Quaternary sediments, occurring at different altitudes are widely exposed along the foreland slope, from the prominent S and SW-verging folds involving the Mesozoic Saccense Domain to the present-day shoreline. Such deposits show a prograding deposition that indicates a progressive, relative lowering of the sea level. On these Pleistocene deposits, successive sea-level changes formed a sequence of marine terraces, the lower of which is probably of Tyrrhenian age. The Belice River, flowing South-West, down-cut the Monte Magaggiaro anticline and the Plio-Pleistocene marine deposits, where we have recognized a flight of fluvial terraces, some of which links and correlate to the marine terraces. Since the Belice River crosses orthogonally both compressive and extensional structures, it offer the interesting possibility to reconstruct the paleo-longitudinal profiles of its course. With this aim, we have reconstructed the flight of inset terrace surfaces of the valley, as useful proxy to detect the short-term signal of deformation related to local structures and to characterize their kinematic. A detailed (1:10.000 scale) geomorphological map of fluvial terraces along the Belice River was carried out, as well as the marine terraces in the foreland area. The correlation and dating (by means of radiocarbon and OSL methods) of these Late Pleistocene-Holocene fluvial terraces was undertaken. At the same time, to detect the long-term signal of regional uplift and local active structures, dating of the Middle-Late Pleistocene paleosurfaces and marine terraces by means of biostratigraphical analysis and U/Th methods, is in progress. Evaluation of Local Amplification Effects: Some Results from the Maiella Area Romano M. A.1, De Nardis R.2, Lavecchia G.3, Milana G.4, Pace B.3 1Istituto Nazionale di Oceanografia e Geofisica Sperimentale, Borgo Grotta Gigante 42/C, 34010 Sgonico, Trieste (Italy), [email protected] 2Dipartimento della Protezione Civile, Via Vitorchiano 4, 00195 Roma, Italy [email protected] 3Laboratorio di Geodinamica e Sismogenesi, Dipartimento di Scienze della Terra, U d’A Chieti, Campus Universitario, 66013 Chieti Scalo, Italy, [email protected] e [email protected] 4Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143 Roma, Italy [email protected] The 1706 Maiella earthquake is one of the most strong events characterizing the seismic history of Abruzzo, but yet several doubts exist about its seismogenic source. The macroseismic field of this event, with I0=IX-X MCS, is characterised by a singular pattern of intensities. The highest values (Imax=X-XI MCS) correspond to the northern and southern terminations of the Maiella anticline; the Santo Spirito hermitage has I=VI-VII MCS, although it is the nearest site at the macroseismic epicentre; the intensity does not diminish regularly with distance. Also others earthquakes with different epicentres, like the 1984 Val di Sangro one, have similar anomalies. In order to evaluate the local seismic response of those sites where the earthquake had caused severe damages (I≥IX), several investigation were carried out. Geophysical, geotechnical and geologic data have been collected. A field survey of ambient seismic noise has been made for all sites, both reprocessing data acquired previously and recording new data. Nakamura’s method has been applied in order to obtain H/V spectral ratio curve for every site. These curves represent at first order the transfer functions of sites, providing their fundamental frequency of resonance, but not their amplification factor. A large number of H/V curves obtained shows a sharp pick at the fundamental frequency, indicating strong impedance contrasts at depth. These results have been roughly correlated with the available geotechnical data and the geological condition surrounding the sites. A 2D numerical analysis of the seismic local response working with boundary elements has been made only for Manoppello and Palena localities. After constructing a 2D geo-lithological model associated with a seismic velocity model for every site, the interaction between the seismic radiation associated with a plane wave and the superficial geology has been simulated. Of all synthetic seismograms obtained, the twenty signals corresponding to the villages and that one corresponding to the reference site on outcropping bedrock have been chosen along the two section in order to calculate average transfer functions. By comparing the H/V spectral ratio curves of seismic noise with the transfer functions calculated, a good correspondence can be found in the estimated fundamental frequencies of resonance in both cases. Because Serramonacesca has a geo-lithological model and an H/V spectral ratio curve of seismic noise very similar to those of Manoppello, located near it, it can suppose that seismic amplification effects characterize a region wider than single sites. S1 UR 3.05 - New stratigraphic and structural evidences from the Lake Garda area Scardia G.1, Galadini F.1, Marzorati S.1, Massa M.1, Monegato G.2, Pini R.3, Rogledi S.4 1. INGV – Sezione di Milano-Pavia, [email protected] 2. Università degli Studi di Padova, [email protected] 3. CNR-IDPA, [email protected] 4. ENI E&P New investigations have been focused on Plio-Pleistocene mainly coarse-grained deposits from Lake Garda area, in order to unravel the recent evolution and the present-day tectonic setting of this sector by means of field data coupled with geophysical subsurface data. In the framework of this study, a geologic field mapping of key-sites was performed and new chronologic constraints were provided by calcareous nannoplankton and pollen biostratigraphy, and paleomagnetism. On the western bank of Lake Garda, new biostratigraphic data from the Mt. San Bartolomeo succession allowed to refer marine clays uplifted at over 500 m a.s.l. to the calcareous nannoplankton biozones NN14 and NN15 (4.13–3.85 Ma, Zanclean), pointing to a minimum uplift rate of ~0.15 mm/yr since the middle Pliocene (Piacenzian). Along the eastern bank of the Chiese River a Pliocene conglomerate crops out at Muscoline, crossed by several ENE/WSW-trending joints. The Muscoline conglomerate forms a small hill surrounded by the most external Garda moraine ridge, whose age has been paleomagnetically constrained to the Subchron late Matuyama (0.99–0.78 Ma). The spatial relationship between the Muscoline conglomerate and surrounding moraines suggests that the former was already uplifted before the onset of the major Pleistocene glaciations according to a NE-SW (Giudicarie) tectonic structure. The Muscoline uplift phase can be likely constrained between the Late Pliocene and the early Pleistocene. Investigations in the Sirmione peninsula were performed on the Cretaceous bedrock and continental conglomerates, doubtfully ascribed to the Messinian. The new gathered data allowed to divide the Sirmione conglomerates in two main superimposed units, different in petrographic composition and sedimentology. The lower body consists of poorly-sorted, matrix-supported, coarse-grained gravels, with blocks exceeding the 0.5 m size; sand and gravel petrographic analyses point to a provenance from the Adige catchment, with porphyries locally reaching the 80% of the whole petrographic composition in the pebble fraction. The upper body is totally characterized by well-sorted limestone and dolostone pebbles, pointing to a drainage confined in the Southern Alps, likely from a former Lagarina Valley. The geologic survey allowed to detect several clayey silt layers interbedded in the conglomerates, sampled for pollen analysis. The whole conglomeratic body is crossed by two sets of vertical joints, displaying NW/SE and NE/SW trends, respectively. A set of vertical NE/SW faults, separated by fault rocks, crosses the Sirmione conglomerate with a strike-slip motion suggested by striations and slickensided grooves. Preliminary pollen analysis of samples from the lower conglomeratic body yielded associations typical of Pleistocene cold stages, rich in conifers, birch and xerophytes, likely constraining the observed faulting to the middle Pleistocene or later. Evidences for NW/SE- and NE/SW-trending joints and strike-slip faults have been observed also in the Cretaceous bedrock of the Sirmione peninsula, where these structural elements overprint a previous compressive NE/SW trending (Giudicarie) deformation. The new above- mentioned stratigraphic and structural data, coupled with the available Fault Plane Solutions, suggest that in the Lake Garda area compressive and strike-slip elements co- exist and likely interplays, with the latter overprinting and segmenting the former previously-dominant Giudicarie overthrusts. RU3.10: Seismotectonic characterization of the central-northern Apennines accretionary prism Scrocca D.1, Livani M.2, Bigi S.3, Carminati E.4 1 CNR – IGAG, Roma, [email protected] 2 CNR – IGAG, Roma, [email protected] 3 Dip. Sc. Terra – Univ. Sapienza Roma, [email protected] 4 Dip. Sc. Terra – Univ. Sapienza Roma, [email protected] The main goal of the UR 3.10 is to investigate the seismotectonic characteristics of the accretionary prism in the central and northern Apennines, with a special attention to the emiliano-romagnolo-ferrarese, marchigiano-abruzzese e central Adriatic sectors, whose tectonic activity is quite well documented During the I phase of the project the following activities have been carried out. The available data have been organized within a geographic information system (ArcGIS 9.1) made up by the following data. • Seismic reflection profiles available in the study area, i.e., public multi-channel seismic reflection profiles belonging to the so-called “Zone B”, CROP profiles, other public profiles made available by the VI.D.E.P.I. project (http://www.videpi.com); • well data; • field data in the marchigiano-abruzzese sector; • published geological and geophysical data (maps, seismic profiles, and geological cross- sections). In the central Adriatic Sea zone the seismic reflection data, originally available as paper copies have been transformed in digital segy files: Both seismic and well data (formation tops and time-depth curves) have been also loaded in Kingdom software (developed by Seismic Micro-Technology, Inc.) which is an advanced software devoted to seismic and well data interpretation. Based on the integrated interpretation of the available data the following results have been worked out. • In the Po Plain, several seismogenic sources have been already identified (e.g., ITSA050 - Poggio Rusco-Migliarino, ITSA051 - Novi-Poggio Renatico, ITGG107 – Mirandola). An improved definition of the geometry at depth of some of these thrusts has been carried out. • In the central Adriatic off-shore, the seismic profiles has been interpreted to map, in two- ways times, the main seismic horizons (e.g., Top Messiniano, Top Scaglia calcarea, Top Fucoidi) and to define the geometry and the phases of tectonic activity of the recognized thrusts . • In the marchigiano-abruzzese on-shore area, the interpretation of the available seismic profiles has been integrated with the results of field surveys. In this way, the geometry (two ways time maps) of the main horizons belonging to both the pre- and syn-orogenis successions (top Fucoidi, top lower Pliocene, and top upper Pliocene) has been reconstructed. • Beneath the Po Plain, all the accessible seismic profiles and geological cross-sections have collected. Based on the analysis of this dataset, the geometry of the basal detachment surface of the Northern Apennines accretionary prism has been analysed. RU S.03: Assessment of the earthquake potential of the Italian seismic sources Slejko D.1, Petrini R. 2, Riggio A. 3, Slejko F.F. 4, Santulin M.5 1. O.G.S., Borgo Grotta Gigante 42c, 34010 Sgonico (TS); [email protected] 2 DST, Universita' di Trieste, via Weiss 8, 34127 Trieste; [email protected] 3 O.G.S., Borgo Grotta Gigante 42c, 34010 Sgonico (TS); [email protected] 4 DST, Universita' di Trieste, via Weiss 8, 34127 Trieste; [email protected] 5 O.G.S., Borgo Grotta Gigante 42c, 34010 Sgonico (TS); msantulin@inogs Two research themes have been developed: 1) the use of the regional strain rate as basis for assessing the probability of earthquake occurrences in Italy, and 2) the possible relations between variations of the local strain and geochemical anomalies in fluids in the Friuli region. 1) We have developed new approaches to calculate 30-year probabilities for the occurrence of moderate-to-large earthquakes in Italy. Geodetic techniques and finite- element modeling are used to separately calculate the expected seismicity rates inside seismogenic areas. The earthquake probabilities obtained from the two approaches show similarities in most parts of Italy, the largest probabilities being found in the southern Apennines. 2) The strain increase in time, deducted a posteriori from the earthquake parameters in the Friuli region, has been compared with the continuous measurement of radon from soil degassing in a test site and the radon and chloride changes observed in springs along localized faults. The study area encompasses the easternmost sector of the Southern Alps. The sedimentary sequences include gypsum-bearing evaporite layers of Permian and Triassic age, which contain scattered grains of chloride-bearing minerals and represent a possible source for the partition of Cl- ions into circulating waters during the water-rock interaction process. In addition, the occurrence of U-bearing sandstones in the sedimentary sequence constitutes an additional reservoir for radon (222Rn). The soil degassing data of CO2 and radon in soil have been collected on the study area over a grid of about 150 points. The recurrence interval of earthquakes has been computed from the local seismicity and calibrated with the regional strain rate observed from GPS data. From the parameters of every earthquake, which occurred in the broad study region, from its recurrence interval, and the regional strain rate value, we have derived the strain time history at the hypocenter and then at a particular test site. During a period of strain, the permeability along the fault may rise, owing to the release of Rn-rich fluids and Cl- ions from pores into solution, yielding geochemical transients in the aquifer. A close association between radon activity from soils and strain changes is observed. In addition, anomalous geochemical signals for radon and chloride in waters have been detected. Modeling the temporal variations in chloride concentrations on the basis of the double-layer theory and Donnan equilibrium seems to satisfactorily simulate the experimental data. RU3.14: Preliminary results on the seismogenic structures of the Ligurian Sea from 3 D inversion and seismicity Solarino S.1, Eva E.1, 1.Istituto Nazionale di Geofisica e Vulcanologia, CNT, c/o DipTeRis, Viale Benedetto XV, 5, 16132 Genova, [email protected]; [email protected] The active part of the GROSMARIN experiment (seismic shots by a large airgun array (8300 ci) onboard R/V “l' Atalante” at a rate of 1 source per minute, recorded by 20 OBS (Ocean Bottom Seismometers) and 20 land stations) was followed by a 5-month passive recording of the seismicity of the Ligurian Sea and coast. The position and number of instruments changed during the experiment: in fact 6 OBS’s were removed right after the active experiment and a few stations on land experienced failures, making necessary small adjustments in their positions. However, the size of the array remained basically unchanged throughout the experiment and was potentially able to ensure a fair constraint on about 80 earthquakes that occurred in the larger area during the passive recording period. Unfortunately the phase picking showed for some OBS’s data a very low signal to noise ratio or problems with the timing system, and the final capability of the network proved to be dramatically decreased by these technical problems. The available data were then merged with a database compiled from the national and regional Italian and French seismic networks and relative to previous time periods. Nevertheless, the contribution of the data collected during the GROSMARIN experiment is significant with respect to the pre-existing dataset, as the plots of the ray coverage show. The comprehensive database has then been used for a local earthquake tomography (LET) with the SIMULPS (Thurber, 1983) technique, a process of iterative simultaneous inversion for 3D velocity structure and hypocenter parameters using travel time residuals from local earthquakes. The preliminary images of both P-wave velocities and Vp/Vs ratio show the relationships between the crustal structure of the basin and the occurrence of earthquakes. The addition of the data of the OBS stations has a considerable influence on the constraint of the depth for the events located at sea. The tomographic reconstructions and cross-sections confirm an upwelling of the Moho in the Ligurian Sea (up to 20 km depth) with an unprecedented level of detail, due to both the availability of rays crossing the study volume from the south and the extremely dense grid of the tomographic model. They also show the existence of a Vp/Vs anomaly the existence and cause of which are still under debate. The next step will be the computation of focal mechanisms and the relocation of seismic events with high-precision technique in order to speculate about the relationships between mapped faults and their seismic activity. In particular, the results of these investigations should be able to distinguish whether the seismicity has to be attributed to the inverse rather than to the strike-slip faults. Apart from the obvious consequences on seismic hazard, the attribution may also influence the potential of tsunami occurrence on the Ligurian coast. RU6.05: Updating of the Italian Tsunami Catalogue and reconstruction of the effects of the 28 December 1908 Messina Straits tsunami Tinti S.1, Gallazzi S.1, Manucci A.1, Armigliato A.1, Pagnoni G. 1, Tonini R. 1, Zaniboni F.1 1.Università di Bologna, Dipartimento dI Fisica, Settore Geofisica, Viale Carlo Berti Pichat, 8 – 40127 Bologna, [email protected] The research activity of RU 6.05 in the frame of the Project S1 is being focused on the search for new primary historical sources containing information on the tsunamis that hit central and southern Calabria, as well as eastern Sicily. Moreover, a re-examination of already known sources has been performed with the goal of improving our knowledge on selected historical events. This part of the work, which will result in an update of the Italian Tsunami Catalogue, is made in co-operation with RU6.01 (responsible Prof. Serafina Barbano). The event that has received major attention in the analysis is the 28 December 1908 tsunami, both because of its relevance in the Italian earthquake and tsunami history and because of the special interest associated with the 100-year anniversary of the event. The re-examination of the studied published in the last decades and of numerous available historical sources has led to determining the way the tsunami attacked the Calabria and Sicily coasts site by site. A GIS database has been setup, containing geo-referenced maps with overlaid information on earthquake-tsunami delay time, first tsunami arrival polarity, run-up height, inundation depth. The maps help in gaining a general view of the tsunami effects and characteristics at a regional scale, and at the same time they allow to inspect the variability of the effects themselves at very local level. The database is being integrated in the general GIS database that is one of the main results of the European Project called TRANSFER (Tsunami Risk And Strategies For the European Region), co- ordinated by the Department of Physics of the University of Bologna. As regards the Messina Straits area, in the TRANSFER GIS database it is possible to compare the historical data for the 1908 tsunami with the results of the numerical simulations which have been performed by different partners in TRANSFER and which are expressed in term of inundation maps and tsunami risk analysis. RU3.12: Scientific update and technological development of DISS. Vannoli P.1 and DISS Working Group1,* 1.Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy, [email protected] *: Barba S., Basili R., Burrato P., Fracassi U., Kastelic V., Tiberti M.M., Valensise G. and Vannoli P. The goal of the UR 3.12 is to increase and update the content of the Database of Individual Seismogenic Sources (DISS), and to improve its functions and structure. The last version of DISS (version 3.0.4) was released in October 2007. During this Project, the activities of this UR have been devoted to the scientific and technological development of DISS. These improvements and new data have been implemented as the new DISS release, version 3.1.0, published online in June 2009. This UR focused on furthering the characterization of the already existing seismogenic sources, both Individual Sources and Composite Sources. The UR implemented and completed background information and support data, images, and texts. New significant bibliographic data that augment information about DISS sources have been included. A new layer containing information about debated seismogenic sources has been included, concerning tectonic structures believed by some authors to be seismogenic and not yet included in DISS. This new layer contains an original critical review of the available data. We focused on the identification and characterization of six new Composite Sources in the eastern Adriatic area. These sources describe sections of the external Dinaric thrust system that affect the eastern side of the Adriatic foreland. New thematic maps concerning new representations of seismogenic sources have been included. An extensive technological development of DISS has been carried out, on both the online interface and standalone version. The latter has been substantially improved. The layer previously known as “surface faults” has been modified to include also fold axes, named “Active Faults and Folds” and will contain data about Late Pleistocene-Holocene features with documented activity. The map symbol of Composite Sources has been improved as to mark clearly the upper edge of the fault, either day-lighting or blind. A new tool to allow for multiple-version handling of the database core has been developed. We implemented most of the above and other minor improvements in the online version. Data access and database navigation has been made easier via (a) the restructuring of auxiliary information associated with the sources, (b) a new organization of current layers, and (c) the replacement of the poorly readable ones. The DISS web portal has been restyled and the W3C (html 4.01, css) validation was obtained for all pages (http://diss.rm.ingv.it/diss). The overall look of the web site now complies with most of the requirements of the Italian Law n. 4, January 9, 2004. A new web-GIS engine, based on open-source software, is been developed and will be made available online by the end of the Project. A number of data conversion tools were developed to generate customized data files that comply with several input requirements for experiments carried out by Research Units of S1 and S2 Projects. Earthquake Mechanisms of the Mediterranean Area (EMMA) version 3.x: an improved tool for characterizing the tectonic deformation styles in the Mediterranean. Vannucci G. 1, Imprescia P. 2 , Gasperini. P. 3 1 Istituto Nazionale di Geofisica e Vulcanologia (sezione di Bologna), via Donato Creti 12, 40128 Bologna, Italy. e-mail: [email protected] 2 Dipartimento Scienze Geologiche, Università di Catania, Corso Italia, 55, I-95129 Catania (Italy), e-mail: [email protected] 3 Università degli studi di Bologna, Dipartimento di Fisica, V.le Berti Pichat 8, 40127 Bologna, Italy. e-mail: [email protected] EMMA database is continuously improving by the addition of further focal mechanisms found in the literature. Presently, the EMMA database version 3.x includes about 12700 focal solutions (about twice of previous official release 2.2). As in previous versions we uniform the different formats and notations of the data available from different sources and try to solve misprints, inaccuracies and inconsistencies that might make the data unusable for other investigations. By an automatic procedure, based on several criteria, we choose the most “representative” (best or preferred) solution when more than one is available for the same earthquake. The database allows to making selections and to export data files suitable to be handled by graphic software and user written procedures. The new version is still based on MS-ACCESS, but a reduced web version is currently in preparation. We improve the available parameters concerning hypocentral and magnitude data found on the original papers with those reported by regional and local catalogs and bulletins. EMMA was already used in the past and will be (hopefully) useful in the future to better characterize the tectonic deformation styles (e.g. by moment tensors sum within given areas or over regular geographical grids) particularly in areas of the European region where seismicity is moderate and only few CMT solutions are available. UR 3.13 - MAXIMUM OBSERVABLE SHAKING (MOS) MAPS OF ITALY Gaetano Zonno1, Gemma Musacchio1, Fabrizio Meroni1, Roberto Basili2, Walter Imperatori3 and P. Martin Mai3 1Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Milano - Pavia 2 Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Roma1 3 Institute of Geophysics (ETH), Zurich, Switzerland The main goal of UR 3.13 is to establish a work flow for a multi-layer map that includes the seismicity of Italy in terms of Maximum Observable Shaking (MOS), and the near-field/far- field boundaries (NF/FF) with respect to the major seismogenic faults mapped within the DISS database. Here we will discuss only the procedure to derive the MOS-map of Italy. Our approach merges updated knowledge on the Italian regional tectonic setting and on the Source Zone (SZ) definition and broadband scenario-like calculation of expected maximum shaking on a given area. For a given SZ, broadband ground shaking is computed for a rupture model derived from a Maximum Credible Earthquake (MCE) and its associated Typical Fault (TF). Amplitude spectra for deterministic Low Frequency and stochastic High Frequency waveforms are reconciled at intermediate frequency, where their domain of validity overlaps, to derive broadband synthetics and compute the associated shaking. As the MCE and TF float along the SZ, broadband ground motion is computed at each point surrounding the given fault and the maximum among observable shaking according to that scenario is plotted on the MOS map. So far the procedure was entirely successfully tested on the Macro Region MR4 (central-northern Apennine), while more detailed analysis is done on the MCE and TF suggested for the Colfiorito earthquake. Here our broadband ground motion scenario shows, besides a complex pattern of variation, a southwestern area of high PGA values, at about 20 km distance from the fault, likely associated to with the properties of the spatio-temporal complexity of the rupture process. For the purpose of the project a complete new map of SZ and MCE is under compilation, grouping seismogenic sources according to Mw and faulting mechanisms. This goal can be achieved most efficiently by targeted numerical simulations that cover the parameter range of interest (in terms of magnitude and distance etc) and consider a large suite earthquake rupture scenarios. Session Project S2 Development of a dynamical model for seismic hazard assessment at national scale RU06 - Testing PSHA outcomes. Albarello D.1, Mucciarelli M.2 1. Dip.Sci. della Terra - Università degli Studi di Siena, Via Laterina, 8, 53100 Siena, [email protected] 2. DiSGG - Università degli Studi della Basilicata, V.le dell'Ateneo Lucano 10, 85100 Potenza [email protected] Several probabilistic procedures are presently available for seismic hazard assessment (PSHA). Furthermore, since the parameterization these computational models could be controversial, these result in a number of possible outcomes all generally compatible with available observations and supported by plausible physical models of the seismic process. In general, these outcomes are combined in the frame of logic-tree approaches. These require the scoring of such models that is generally performed on the basis of ex-ante evaluations based on the reliability of the elements concurring in the hazard estimates within each computational model. This approach appears unsatisfactory since, e.g., the reliability of each model cannot depend on the concurring element only, but mainly on the combination of these ones. A different approach is here developed that is based on the ex- post evaluation of PSHA models. This is based on the basic idea that each computational model should be evaluated as a whole, by comparing the relevant probabilistic “forecasts” for a fixed exposure times with empirical evidence relative to seismic occurrences during some selected control periods of dimension comparable with the relevant exposure time by taking into account the inherent probabilistic character of hazard estimates. Likelihood and Counting methods have been considered to this purpose along with alternative approaches based on numerical simulations. Another approach that will be shortly outlined is based on the comparison of standard PSHA outcomes with alternative estimates deduced from a statistical analysis of available macroseismic observations (site seismic histories). This comparison can concern the relevant hazard curves, the reference level of ground shaking (i.e, the one corresponding to any fixed probability level) , or the results of deaggregation analyses. Some results obtained by the application of these testing procedures will be shortly outlined. RU02: summary of activities in S2 project Faccioli E.1, Cauzzi C.2, Demartinos K.3, Garavaglia E.4, Petrini L.5, Vanini M.6, Villani M.7, Smerzini C.8, Stupazzini M.9 1 Department of Structural Engineering – Politecnico di Milano, [email protected] 2 Department of Structural Engineering – Politecnico di Milano, [email protected] 3 Department of Structural Engineering – Politecnico di Milano, [email protected] 4 Department of Structural Engineering – Politecnico di Milano, [email protected] 5 Department of Structural Engineering – Politecnico di Milano, [email protected] 6 Department of Structural Engineering – Politecnico di Milano, [email protected] 7 ROSE School, Pavia, [email protected] 8 ROSE School, Pavia, [email protected] 9 Department of Structural Engineering – Politecnico di Milano, [email protected] RU2, in addition to closely cooperating with its leader (E. Faccioli) as S2 Project coordinator, carries the primary responsibility in the development of Tasks T3 (ground motion attenuation and site effects) and T4 (near-field numerical simulations), and plays a significant role in Tasks T1 (development of the new seismic hazard assessment tool CRISIS++), T5 (probabilistic risk assessment) and T6 (Model validation). The main results obtained by RU2 at the present stage of the Project are selectively highlighted in this poster. For T1, exploratory hazard evaluations on the Sulmona (Central Italy) scenario area are described, aimed at testing the feasibility of “generalized” attenuation descriptions within CRISIS++, through the first two moments of the sample of simulated ground motion parameters at each receiver. For T3, the main contents of Deliverable D3.1, illustrating the choice of GMPEs to be introduced as preferable options in the new code (CRISIS++) are illustrated, as well as the advancement achieved on including (through modified Ground Motion Prediction Equations, or GMPEs) topographic amplification effects in regional hazard analyses. The results of 3D deterministic ground motion simulations (with the resulting variability), performed for the Sulmona test area through an in-house developed parallel spectral element code (GeoELSE), are presented for T4. The same simulation results, integrated with a suitable high frequency contribution, are used in T5 (as elastic response spectra) as input for assessing seismic damage to the Sulmona building stock. This is achieved through a direct probabilistic formulation based on the capacity spectrum method, specifically developed within the Project framework. Finally, for T6, the impact of using probability-based preference criteria to rank the performance of different GMPEs in SHA on a simple source-site configuration is discussed. Mechanical models for large scale damage assessment, including independent verification of their effectiveness and uncertainty estimation Lagomarsino S., Università di Genova Earthquake loss estimation procedures require data capture on the vulnerability of both building and exposed elements. The damage assessment needs referring to reliable vulnerability models which are able to establish a correlation between hazard and structural damage. To this aim, mechanical models, which describe the overall inelastic response of the structure by mean of a force – displacement curve (“capacity” curve) seem particularly attractive. In fact, it presents the following main advantages: to employ the results of sophisticated hazard analyses (by using the seismic input in the spectral form which may be compared with the capacity curve by adopting the non linear static procedures as tool of seismic verification); to keep explicitly into account the different parameters which concur to define the structural response. These reasons lead to the choice to adopt in the S2 Project mechanical models which are explicitly based on geometrical and mechanical parameters. Moreover this choice allows us to take explicitly into account the statistical variability of the parameters leading to the definition of a stochastic model of vulnerability. The attention has been focused on masonry and reinforced concrete structural types which represent the most important classes in the Italian territorial. Starting from some models proposed in the literature (the model proposed in Cattari et al. (2004) and Pagnini et al. (2008) for masonry buildings and that of Crowley et al. (2004) for the r.c. ones), some modifications to these original proposals have been introduced, in particular in order to take into account some peculiarities of existing buildings and to overcame some drawbacks emerged from preliminary applications. With reference to the uncertainty estimation, it is useful highlighting that the procedure proposed allows us to take into account and distinctly quantify the uncertainties associated both to the “capacity” (related to the parameters which the mechanical models are founded on, the damage states and the model error, respectively) and of the “demand” . It is worth pointing out that, in case of assessment at territorial scale, the minimum unit of analysis is represented by a stock of buildings with homogeneous behaviour and not by a single building: thus a fundamental step of the proposed procedure consists of processing the available data in order to properly aggregate them and of establishing suitable correlations with the parameters which the interpretative models adopted are founded on. In particular a classification based on data ISTAT(which represent a database homogeneous at national scale) is proposed. Finally the application to some case studies (Sulmona and L’Aquila) is carried out (this task represents the main activity of the second year of the project). S2-RU5: Earthquake rate (ER) models for Italy Marzocchi W.1, Akinci A.1, Console R.1, Falcone G.1, Lombardi A.M.1 and Murru M1. 1 Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma 1 This activity aims to produce and provide a set of earthquake rate models and the probability of occurrence (M≥5.5) in the next 30-50 years to be used for the seismic hazard assessment in Italy. These models have also been submitted to the Collaboratory for the Study of Earthquake Predictability (CSEP) forecasting center in ETH, Zurich, to be tested for Italy. In this framework we produce several ER models based on seismicity and geological and geodetic database. For recurrence time of the seismic activity in the models we assume both time-dependent and time-independent probability functions. The first model uses gridded historical seismicity assuming that the future earthquakes will be clustered spatially near locations of historical mainshocks. In this model, spatially smoothed seismicity, the forecast seismicity rate is the sum of constant background seismicity and is assumed constant in time. The second forecast is based on the fusion of a statistical renewal model with a physical model considering fault interaction that in real circumstance can either increase or decrease the future earthquakes probability with respect to what is expected by a simple renewal model. We consider fault interaction by the computation of the co seismic static permanent Coulomb stress change caused by previous earthquakes on the investigated fault. The last model proposed is time-dependent and assumes that each earthquake can generate, or is correlated to, other earthquakes, through different physical mechanisms. Specifically it consists of a sequential application of two branching processes, in which any earthquake can trigger a family of later events on different space-time scales. The forecasting maps have been provided for the first year of activity. RU3-S2: Activities for the improvement of seismic hazard assessment in Italy Meletti C. 1, Galadini F. 1, Meroni F. 1, D’Amico V. 1, Gomez Capera A.A. 1, Martinelli F. 1, Sudati D. 1, Gori S. 1, Falcucci E. 1, Scardia G. 1, Messina P. 2, Sposato A. 2, Giaccio B.2 1.Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Milano-Pavia, [email protected] 2.Istituto di Geologia Ambientale e Geoingegneria, CNR, [email protected] The activities performed by RU3 (INGV Milano) cover 5 tasks out of 6 of the S2 project. The main effort has been devoted to the realization of the seismic hazard assessment tool, in a strict cooperation with UR1 (UNAM); based on the standalone version of Crisis software (under development by Unam), a web-based interface has been produced. With regard to the development of the presentation layer for the Crisis application, the application is based on a principal object, the HazardModel, which uses aggregates of other objects, each one describing elements of the project: sources, attenuation models, etc. Firstly the development required the identification of the objects used by the HazardModel and their classes. The HazardModel object contains groups of objects of each class. The Web presentation layer was realized mapping each group of objects to a Web page. This solution was required considering the need to minimize the interaction between the browser and the server, to minimize the user waiting times. The Web version of Crisis is running, but the access is limited to the project partners only. The contribution to task 2 is the set up of the same input elements of the Italian seismic hazard official map (MPS04) for Crisis code, to be used as a reference. Within task 3, consolidated models for attenuation of macroseismic intensity have been implemented in Crisis, in order to allow computation of seismic hazard maps also in terms of macroseismic intensity. Thus, a review of the most recent empirical models of intensity attenuation so far proposed for Italy has been performed. The contribution to task 4 is devoted to the definition of the 3D geological model of the Sulmona area, so that it can be used as reference model in the definition of non parametric attenuation model. Geological analyses performed along the Mt. Morrone fault system allowed to define a rake of the fault system of about 260°, a slip rate ranging between 0.4±0.07 and 0.8±0.09 mm/yr and a length of the structure of about 23 km. This suggests a maximum expected magnitude of an earthquake originated along this fault ranging between 6.6 and 6.7. Further active normal fault systems, potentially responsible for large magnitude earthquakes (M 6.5-7), affect the sector surrounding the Sulmona basin area. This UR contributed also to the activities of Task 6, in particular to deliverable 6.2 coordinated by UR6-UNISI. RU-S2/4: Activity Report 2008-09 UNICH-OGS Peruzza L.1, Pace B.2, Visini F.2 1.Istituto Nazionale di Oceanografia e Geofisica Sperimentale, Sgonico Trieste, [email protected] 2.Università “G. d’Annunzio” Chieti-Pescara, Chieti Scalo, [email protected], [email protected] UR4 OGS-UNICH is involved in several Tasks, namely T1.4, T2.4, T4.1 and T6.1. Internal reports, presentations and electronic supplements, have been released during the 1st phase of the Project. Here, a short description of activities is given. Task 1.4 – Simulation with realistic and fictitious sources in Central Italy We performed some simulations of SHA with CRISIS 2007 (v. 5.4), using both “fictious” and real source geometries. The sensitivity tests have been done in order to better understand the behaviour of CRISIS software, using the parameters usually considered for probability maps based on individual faults and slip rate data. Code performances using the LASSCI seismicity rates computed for CSEP activities (i.e. annual rate for 0.1 magnitude bins, referred to gridded points each 0.1° - see Task 6.1), is scheduled with the last release of CRISIS for web. Task 2.4- ERF model based on BPT behaviour applied to DISS 3 seismogenic structures and LaSSCI model We implemented two Matematica® notebooks in order to obtain: a) an estimate of Mean_Recurrence_Time and Aperiodicity_Coefficient, to be used in BPT distribution functions, and b) its conditional probability, given an Elapsed_Time. The Earthquake Rupture Forecast of LaSSCI model (Layered Seismogenic Source model in Central Italy, from Pace et al., 2006; Peruzza et al., 2007) has been improved during this project, by introducing error propagation theory. Task 4.1- Scenario simulations The activity done for selecting some scenario test cases, both for the Project’ purposes or for answering peculiar requests of DPC, is widely documented by two internal reports, and a contribution to D4.1. The first report concerns the synthesis of Previous S2’ Project results, to enlighten the areas most prone to an impendent earthquake; the second one is a quick analysis of M>6.0 sources in five regions, selected by DPC. Finally the contribution to D4.1 is a state-of-the-art about parametrization of Sulmona seismogenic source and its hazard. Task 6.1- Definition of model to be tested - LASSCI In the frame of seismic hazard validation activities planned by the Project, we fixed up the LaSSCI model, in order to submit two middle-term forecasts of 5 and 10 years for a validation test by the CSEP international project, with formal statistical tests on the observed seismicity rates during the period 2009-2014 (see details in http://eu.cseptesting.org/Home). Session Project S3 Fast evaluation of parameters and effects of strong earthquakes in Italy and in the Mediterranean ShakeMap: accounting for finite fault trough synthetic waveforms Ameri G.1, Cirella A.2, Cultrera G.2, Herrero A.2, Pacor F.1, Piatanesi A.2, Saraò A.3 1 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy 2 Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy 3 Istituto Nazionale di Oceanografia e Geofisica Sperimentale, Trieste, Italy ShakeMap package uses empirical ground motion prediction equations (GMPEs) to estimate the ground motion where recorded data are not available. Recorded and estimated values are then interpolated in order to produce a shaking map associated to the considered event. Typically, at close distances from the fault, a relative small number of recordings are available and the ShakeMaps are largely based on values estimated with GMPEs. Anyway these models are rather simple and accounts only for average characteristics of ground motion. The aim of this work is to evaluate whether synthetic seismograms generated with finite-fault model can be used in place of (or as an integration to) GMPEs within the ShakeMap methodology. An advantage of using simulated motions is that source effects, as directivity due to the rupture propagation along the extended fault, can be accounted for. This is particularly appealing in Italy where the regional GMPEs are based on a few number of near-source records for moderate-to-large earthquakes, not reliable for estimating ground motion in the vicinity of the source. We chose the Mw 6.9 2008 Iwate-Miyagi (Japan) earthquake as a case study. This earthquake has been recorded by a very large number of stations and in this case the corresponding ShakeMap relies almost totally on the recorded ground motions. To validate the use of the synthetic data to integrate the Shake-map, the following steps were performed. First, we remove half of the stations and invert this reduced dataset to reconstruct a finite-fault source model. Then, we use the DSM technique and the source model obtained above to compute synthetic waveforms at the removed stations. The removed values are then substituted with synthetic values and the obtained maps are compared to the original ones (containing observed data only). RU S3/4 INGV Milano-Pavia: Implementation of the ShakeMap® System at the RAIS Augliera P., Ameri G., Bindi D., D’Alema E., Franceschina G., Ladina C., Lovati S., Luzi L., Massa M., Marzorati S., Pacor F. Istituto Nazionale di Geofisica e Vulcanologia, Via Bassini, 15, 20133 Milano, [email protected] This work summarizes the activities performed by the INGV MI-PV research unit (RU4) with reference to the planned work-packages (WP). In order to assure the availability of the highest number of strong-motion data and with the aim to provide reliable real-time ground shaking evaluations, the activity was devoted to update the data acquisition system of the RAIS (Strong Motion Network in Northern Italy; in italian: Rete Accelerometrica in Italia Settentrionale) network. At present, 7 out of the 21 stations of the network were updated by replacing the original acquisition system (with modem GSM transmission system) with a real-time acquisition system, based on digital recorders with TCP/IP or Wi-Fi link. All stations are equipped with Kinemetrics Episensor accelerometers. For the management of the real time stations the SEEDlink and Earthworm packages were adopted and the MiniSEED format was introduced for real-time data exchange (WP1.1). The data dissemination was achieved by making available on the website (http://rais.mi.ingv.it) the computed strong motion parameters, together with the amplification functions, for each recording station. Acceleration data are also made available on the Italian Accelerometric Archive (http://itaca.mi.ingv.it) (WP1.4). The ShakeMap program was successfully installed. Ground shaking maps of 3 events recorded by the National Seismic Network and used as reference solutions, were accurately reproduced. Moreover, in order to verify the quality of the procedures used for the Peak Ground Motion (PGM) evaluation, several strong motion parameters were calculated starting from the input waveforms and then successfully compared with the reference values (WP2.1). Four GMPEs were used to predict the ground shaking of December 23, 2008, Ml 5.1 and Ml 4.7 Parma earthquakes and to compare predicted PGAs with the peak values recorded by 33 stations of the RAN and RAIS networks. Comparisons were made considering both national and regional relationships (WP3.4). Part of the RU4 activity was devoted to investigate the high-frequency methodologies suitable to simulate synthetic seismograms to be used in place of GMPE within the ShakeMap methodology. The June 14, 2008, Mw=6.9, Miyagi (Japan) earthquake was selected as case study. Indeed, this event was recorded by a large number of K-NET and KIK-NET stations and produced a “reference ShakeMap” almost totally relying on the recorded ground motions. Starting from another, deliberately ill-constrained ShakeMap, obtained using only 8 stations of the network and considering the information on the finite- fault characteristic of this earthquake, ground-motion simulations were performed with different stochastic techniques. Ground motion parameters (PGA, PGV and PSA) were calculated for a dense configuration of sites in order to compare these “synthetic ShakeMaps” with the “reference ShakeMap” (WP 4.3). RU8: Calibration of soil amplification factors for real time ground motion scenarios in Italy and comparison with experimental data from L’Aquila sequence Barani S.1, De Ferrari R.1, Ferretti G.1, Spallarossa D.1 1. Dipartimento per lo Studio del Territorio e delle sue Risorse, University of Genoa, Viale Benedetto XV 5, 16132, Genoa, Italy, [email protected], [email protected], [email protected], [email protected] This work presents a study that deals with the calibration of soil amplification factors (or soil coefficients) to be used within the ShakeMap procedure (Wald et al., 2003) for generating real-time ground motion scenarios that account for the amplification effects produced by specific soil conditions. Nowadays, these scenarios are routinely produced in Italy by applying the soil coefficient calibrated for the US using the Borcherd method (Borcherd, 1994). Hence, it seems necessary to calibrate site amplification factors based on an Italian database of soil profiles. In this work the ground response of 100 soil profiles is studied through 1-dimensional (1D) equivalent linear numerical simulations. To this purpose, several real rock ground motion time histories are driven through models of the soil columns. Accelerograms were selected from both national and international ground motion databanks and then were grouped into different classes corresponding to different horizontal peak ground acceleration (PGAH) ranges. Soil amplification factors are calculated here using different definitions, either as the ratio of the spectral acceleration at the surface to the spectral acceleration at the rock outcrop or by dividing the (acceleration or pseudo-velocity) response spectrum intensity at the surface to the reference response spectrum intensity. For each PGA class, regression analyses are performed to derive empirical equations that relate the amplification factor to different soil parameters, such as the average shear wave velocity, VS,30, in the top 30m of a soil profile and the soil fundamental frequency, f0. Finally, the reliability of amplification factors from numerical analyses is verified through comparison with experimental earthquake data recorded during the recent L’Aquila sequence at Paganica, Tempera, and Bazzano. The Reference Site Method (RSM) has been used to derive both the H/Href amplification curve and the amplification factor at each recording site. RU DST-UNITS: Fast moment magnitude estimation and ShakeMaps computation in the Southeastern Alps. Costa G.1, Gallo A.1, Laprocina E.1, Moratto L.1, Suhadolc P.1 1Dipartimento di Scienze della Terra – Università di Trieste, via Weiss 1 - 34127 Trieste, [email protected] In the framework of Project INGV-DPC S3 (2007-2009), a stable and automatic method is implemented at DST to estimate in real time the seismic moment, moment magnitude and corner frequency from broad-band and accelerometric data. The procedure has two steps: the first one consists in interfacing with the Antelope system from where waveforms are retrieved. The real-time waveforms come from the transfrontier network which integrates the Friuli Venezia Giulia (DST, OGS), Austrian (ZAMG) and Slovenian (ARSO) networks. The second step consists in estimating the seismic moment and the corner frequency following Andrews method (1986). At the end the results are stored in a database table. The procedure is tested with the recordings of some strong earthquakes: Carnia 2002 (Mw=4.9), Bovec 2004 (Mw=5.1), Parma 2008 (Mw=5.4) and Aquila 2009 (Mw=6.1), some its aftershock and those of some minor events occurred in the SE Alps area, for which independent seismic moment and Mw estimates, obtained by waveform inversion, are available. The real-time procedure is running at DST and the results are visible on the web page www.dst.units.it/RAF06. The Mw computation together with the moment tensor estimation (TDMT) will also provide better input data for the real-time ShakeMap computation. At DST the “ShakeMap” software has been implemented for the SE Alps area to obtain a stable interface with “Antelope” acquisition system in order to extract the ground motion parameters from waveforms and to generate ShakeMaps within five minutes of the earthquake occurrence. The “ShakeMap” software has been regionally calibrated by adopting a specific near- surface geological classification, different GMPEs for weak and strong motion, and the relationship proposed by Faccioli and Cauzzi (2006) to compute the instrumental intensity. The model is validated by comparing the observed intensity maps data with the instrumental intensities derived from the related scenarios calculated for three important past seismic events in the studied area (Cansiglio 1936, Friuli 1976, Bovec 1998). The computation of the intensity misfit value gives satisfactory results. Finally, an automatic procedure has been developed to exchange in real-time the ground motion parameters with INGV and the other partners of the INGV-DPC S3 project. The results are posted automatically on the department web pages (www.dst.units.it/RAF06). RU DST-UNITS: The integrated RAN-RAF accelerometric network in Friuli Venezia Giulia. Costa G.1, Filippi L.2, Gallo A.1, Laprocina E.1, Moratto L.1, Suhadolc P.1, Zambonelli E.2, 1Dipartimento di Scienze della Terra – Università di Trieste, via Weiss 1 -34127 Trieste, [email protected] 2Presidenza del Consiglio dei Ministri - Dipartimento della Protezione Civile - Ufficio Valutazione, prevenzione e mitigazione del rischio sismico Since 2003 an agreement between the National Civil Defence (DPC) and the Regional Civil Defence (FVG-PC) has been signed in order to co-ordinate and integrate on the Friuli Venezia Giulia territory the National Accelerometric Network (RAN) and the Friuli Venezia Giulia Accelerometric Network (RAF). This co-operation will lead to the installation of new RAN accelerometric stations in the area and, in particular, will allow to automatically send RAN and RAF data to the DPC headquarters in Rome and to DST, through the FVG-PC data center. As FVG-PC consultant for the strong motion data, the DST collaborates to the sites selection and to the management of the new RAN stations. DST has recently performed detailed noise measurements in order to give a local geological classification of the recording sites. The RAF-RAN integrated network is configured to record accelerations at several important sites in the seismic area of Friuli Venezia Giulia and near its borders with Slovenia and Austria. The network geometry is built up taking into account also other seismological agencies and networks operating in Friuli Venezia Giulia (OGS) as well as in Austria (ZAMG) and Slovenia (ARSO). The data recorded in real-time, or quasi real-time, by the network, are collected and exchanged through the datacenters by the Antelope software. Antelope also analyzes and stores the collected waveforms in a database containing the signals recorded by this network and by the integrated transfrontier seismic network of the Southeastern Alps. An automatic procedure has been developed by DST to extract in real-time the ground motion parameters (PGA, PGV, PSA03, PSA10, PSA30, Arias and Housner intensities) and to exchange these data with other partners of the Project INGV-DPC S3. The ground motion parameters are used also to generate regional real time ShakeMaps, to compute regional weak motion GMPEs and possible ground-shaking scenarios. The collected waveforms are also used to study the physics of the seismic source, to estimate site effects and to perform microzonation analyses. At DST an accelerometric database with RAF data – as of today about 700 digital 3- component records, related to about 400 events - is maintained since 1993. . Preliminary assessment of site stratigraphic amplification for Shakemap processing. A. d’Onofrio1, L. Landolfi2, F. Silvestri3, G. Tropeano4 1 Università di Napoli Federico II [email protected] 2 Università di Napoli Federico II [email protected] 3 Università di Napoli Federico II [email protected] 4 Università della Calabria [email protected] The research activity carried out by RU 7 DIGA-UNINA in the first 18 months of the project was aimed to the definition of a simplified procedure to include non-linear site effects in Shakemap. Site stratigraphic amplification is represented by the amplification factor SS = as/ar . which can directly multiply the reference ground motion amplitude ar generated by Shakemap on a stiff and flat rock outcrop to obtain the surface motion as. A first set of stratigraphic amplification factors was identified in the seismic databases ITACA and SISMA, selecting those stations on deformable soil for which a record of the same event exists also at a nearby station (within 1 km) located on outcropping rock. Three sites have been identified, representative of as many classes specified by the National Technical Code (DM 14.I.2008): - class B: Forgaria Cornino (deformable) / San Rocco (rock) - class C: San Giuliano di Puglia: School (deformable) / Church (rock) - class D: Cesi Valle (deformable) /Monte (rock) A second set of experimental data has been obtained by analysing the accelerograms at 8 RAN station sites (from ITACA), where the subsoil is adequately characterized (layering, bedrock depth, shear wave velocity profile, non-linear soil behaviour). The recorded data have been de-convolved to the bedrock, allowing to back-figure the originally related seismic motion. All the sites belong to class B, except one that is classified as class C, according to DM 14.I.2008. A third set of data came from 1D numerical site response analyses performed on three virtual stratigraphic profiles, consistent with the ground classification criteria suggested by DM 14.I.2008, and subjected to 74 acceleration time histories recorded at those Italian sites classified as rock sites (class A). The experimental data were interpreted with the following best-fit power functions: -0.13 - class B soil: SS = 1.15 ag -0.21 - class C soil: SS = 1.01 ag -0.32 - class D soil: SS = 0.77 ag where ag is expressed in g. The effects of non-linearity are evident for each analyzed data set. The amplification factors corresponding to the lower values of the reference acceleration are higher than those provided by the NTC and by the relationship currently implemented in Shakemap (Borcherdt et al. 1994). Instead, in the higher ranges of acceleration, the amplification factors are: higher (class B), similar (class C) and lower (class D) than those provided by DM 14.I.2008 and by Borcherdt et al. (1994). References - Borcherdt, R. D. (1994). Estimates of site-dependent response spectra for design (methodology and justification), Earthquake Spectra, 10, 617-654. - D.M. 14.I.2008 " Norme Tecniche per le Costruzioni" - ITACA http://itaca.mi.ingv.it/ItacaNet/ RU S3/2_1 – DSF-UniNA: ShakeMap® Installation and Validation at DSF-UniNA and Automatic Data Exchange Emolo A.1, Caccavale M.1, Convertito V.2, Delouis B.3, De Matteis R.4, Di Crosta M.1, Festa G.1, Gallovic F.1,5, Lucca E.1, Satriano C.1, Stabile T.A.1, Zahradnik J.5, Zollo A.1 1. Dept. of Physics, University “Federico II”, Naples, Italy, [email protected] 2. INGV – Osservatorio Vesuviano, Naples, Italy, [email protected] 3. Géosciences Azur, University of Nice, Sophia Antipolis, France, [email protected] 4. DSGA, Sannio University, Benevento, Italy, [email protected] 5. Dept. of Geophysics, Charles University, Prague, Czech Republic, [email protected] The DSF-UniNA research unit manages the Irpinia Seismic Network (ISNet), deployed along the Campania-Lucania Apennines, which comprises 28 6-components, real-time telemetred stations. With a mean spacing of 10 km, ISNet is able to provide a denser coverage of ground motion measurements, integrating and complementing the National Seismic Network in the region. The computing infrastructure of ISNet is employed in the framework of the S3 project to: a) realize a backup system for national-level computation of shake maps; b) provide INGV, within a few minutes from the event detection, with the peak ground motion measured at the ISNet stations. The ShakeMap installation at DSF-UNINA is based on a Linux/Intel architecture, which comprises an HP ProLiant DL160 server and the free, open source Linux distribution Debian 5.0. ShakeMap 3.2 has been installed from the INGV custom distribution, which includes three example earthquakes. The installation has been validated from tests conducted on these earthquakes, comparing our results with reference maps. The interface between our system and the INGV servers is realized through an automatic procedure, which retrieves peak measurements from National Seismic Network, computes shaking maps and uploads the results to the INGV ftp server. If an earthquake is recorded at the ISNet as well, the procedure will also upload the ground motion measurements at the ISNet stations. We illustrate through a synthetic example how the integration of the ISNet stations with the National Seismic Network can improve the accuracy of shake maps for the Campania- Lucania Apennines. RU S3/2_2 – DSF-UniNA: Research Activities in the Frame of the S3 Project Emolo A.1, Caccavale M.1, Convertito V.2, Delouis B.3, De Matteis R.4, Di Crosta M.1, Festa G.1, Gallovic F.1,5, Lucca E.1, Satriano C.1, Stabile T.A.1, Zahradnik J.5, Zollo A.1 1. Dept. of Physics, University “Federico II”, Naples, Italy, [email protected] 2. INGV – Osservatorio Vesuviano, Naples, Italy, [email protected] 3. Géosciences Azur, University of Nice, Sophia Antipolis, France, [email protected] 4. DSGA, Sannio University, Benevento, Italy, [email protected] 5. Dept. of Geophysics, Charles University, Prague, Czech Republic, [email protected] The DSF-UniNA research unit has performed activities concerning the moment tensor estimation, the ground motion simulation, the kinematic source inversion, and the determination of ground motion predictive equations. The moment tensor solution is determined by modelling the strong-motion waveforms using two different approaches. The former one uses the point source approximation and performs a grid search over a set of trial source positions and time shifts in order to identify the optimal centroid position, time and moment tensor through a minimization of the residual errors. In the second method the rupture is represented by a finite 1D source model. Source finiteness is approximated by a summation over point sources aligned along fault strike. The focal mechanism and the linear seismic moment distribution along the strike of the fault are inverted at the same time using a fast and optimized grid search combined with a simulated annealing algorithm. Concerning the seismic wavefield simulation, it is numerically modeled using three different algorithms. The first, based on the asymptotic ray-theory approximation, rapidly computes high frequency seismograms including direct and reflected waves from 1D velocity models. The second code uses an hybrid approach, with the low frequency seismograms computed by a discrete wave number method and enriched by stochastic high frequency modeling. Finally, a high-order spectral element method code is used for the full 3D numerical modeling of the wavefield. The kinematic inversion is aimed at determining the rupture direction, the final slip distribution on the fault plane and the propagation velocity of the rupture. The methodology is based on representation integral, in the form proposed by Burridge and Knopoff (1964). It is solved by a finite elements technique that uses a Delaunay’s fault plane triangulation. The slip is parameterized through a 2D Gaussian overlapping functions, and the inverse problem is solved using the Neighbourhood algorithm with a L1 norm. Finally, we estimated a GMPE for low-magnitude earthquakes (M< 4.0) in the Campania- Lucania region in southern Apennines (Italy. The model concerned peak ground acceleration (PGA) and velocity (PGV) and has been retrieved on a data-set of about 160 earthquakes recorded by the Irpinia Seismic Network (ISNet) (Iannaccone et al., 2009) in the last four years. High Frequency Attenuation of S-waves in Eastern Alps Gentili S.1, Franceschina G.2 1. Istituto Nazionale di Oceanografia e Geofisica Sperimentale, Centro Ricerche Sismologiche, Udine, ITALY, [email protected] 2. Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Milano-Pavia, Milano, ITALY, [email protected] Anderson and Hough (1984) proposed an empirical model for the shape of acceleration spectra obtained with strong-motion data from local earthquakes, observing that the logarithm of the spectrum exhibits a linear trend over the corner frequency of the recorded event. A similar trend, quantified by the spectral decay parameter k, can be observed also for weak-motion data when the recording bandwidth is sufficiently large. Although, at least for some datasets, an event dependence can not be excluded for k, the spectral decay was generally interpreted in terms of propagation effects considering both station- and distance-dependent contributions. Given a set of recording stations and assuming a simplified stratigraphy, consisting of a single layer with relatively strong attenuation properties overlying deeper less attenuating materials, the station-dependent contribution would be related to the specific propagation path beneath the site. Conversely, the distance-dependent part of the k parameter, roughly common to all of the station-to-site paths, would be associated to the deeper propagation. Anderson and Hough (1984) proposed a linear dependence of k on the epicentral distance R: k=k0+k1*R, where k0 is the station-dependent term and k1 is the slope, common to all the stations. Anderson (1991) proposed a more general approach, imposing that the distance-dependent term is a smooth function of R, close to zero for 0 distance. A numerical method can be used to evaluate k for a finite number of distances Ri by the joint inversion of all the estimated values of k. For each distance, the kj(Ri) estimated values are obtained by a linear regression of the S-wave acceleration spectra of all earthquakes recorded at all stations with epicentral distance Ri. In this poster, we make the comparison of the two methods using the data of 11 short- period seismic stations of the North Eastern Italy network managed by the department “Centro Ricerche Sismologiche” of the “Istituto Nazionale di Oceanografia e di Geofisica Sperimentale”. The analysis is performed on 302 earthquakes (1263 3D traces) recorded in the period 1994-2007. The obtained values of k0, approximately ranging from 0.015 and 0.055 s, are compatible with the results obtained with a smaller dataset by Franceschina et al. (2006). The estimated slope, k1 = 0.00013 s/km, is consistent with high-frequency Q values ranging from 2200 and 2560, when S-wave velocities between 3.0 and 3.5 km/s are hypothesized for the deeper part of the crust. In addition, the site-dependent term of k is proportional to the mean of the Nakamura’s ratio on a large frequency band, obtained independently for the same stations by Bragato and Slejko (2005) and Barnaba et al. (2008). Testing of a Prototype Tsunami Early Warning System for the Coasts of Italy Lorito S., A. Piatanesi and F. Romano Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy In the framework of the agreement between Italian Civil Protection and INGV (DPC-S3 project), we are studying the feasibility of a Tsunami Early Warning System (TEWS) for the coasts of Italy. The performances of such a prototype TEWS have been evaluated. The underlying concept is borrowed from the Japan Meteorological Agency’s TEWS. This tsunami warning system is in fact based on a set of elementary submarine earthquake sources. In case of an earthquake, the wave heights generated by each elementary source must be interpolated to evaluate in real-time if and where a significant tsunami could have been generated, in order to launch an alert for the most exposed coastlines. A series of synthetic tests have been conducted in order to assess the optimal resolution of the parameter’s space sampling, and to construct a sufficiently accurate database of elementary sources. The results of the synthetic tests are presented here. Moreover, the Mw=6.9 May 2003 Boumerdes- Zemmouri (Algeria) earthquake, which generated a moderate tsunami causing significant damages particularly at Balearic Islands (Spain), can be used as a case-study for the validation of the database. Moment rate spectra and site correction from Italian National Network Mercuri A.1, Malagnini L.1, Mayeda K.2, Akinci A.1 1 Isitituto Nazionale di Geofisica e Vulcanologia 2 Weston Geophysical Corporation, Lexington, Massachusetts. In the context of the S3 project for the fast evaluation of parameters and effects of strong earthquakes in Italy and in the Mediterranean, the WP3.2 deliverable is dedicated to determine the absolute site responses at the broad band stations of the Italian National Network. This task requires the knowledge of the source and attenuation terms according to the method proposed by Malagnini et al. (2004). Here we present the absolute moment rate spectra calculated for more then 600 Italian events obtained applying the coda method developed by Mayeda et al. (2003). Specifically, we provide the moment-rate spectra of 170 events with Ml>=3.0, recorded during the sequence of the April 6 2009 L'Aquila mainshock, and the moment-rate spectra calculated for 439 earthquakes recorded between 2005 and 2009 in NE Italy and Sicily. The mentioned, absolute moment-rate spectra, together with the regional attenuation functions found in the literature, may be used for the determination of the absolute site terms relative to all the recording stations. Rapid information (intensity) and KF scenario in quasi-real time; L’Aquila, 2009 and Frignano, 2008, two comparisons. Pettenati F.1, Sirovich L.1, Sandron D.1 1.Istituto Nazionale di Oceanografia e di Geofisica Sperimentale – OGS, Borgo Grotta Gigante 42c, 34016 Sgonico Trieste, [email protected] The QUEST teams of the Italian INGV and DPC (coordinated by P. Galli and R. Camassi) produced 223 preliminary MCS intensity estimations for the earthquake of L’Aquila, and 141 for that of Frignano (Dec. 23, 2008). These data sets are the empirical benchmarks for measuring the quality of any synthetic simulations. Thanks to this precious reference, we could test the qualities of three kinds of fast intensity scenarios of both earthquakes: two versions of ShakeMap, released by USGS and INGV, and our KF parametric scenarios. We stress that the KF scenario, originally intended for preventive purposes, does not use co-seismic recordings and, also for this reason, it does not compete with fast-information algorithms. Our task in the INGV-DPC S3 project is to compare fast information and KF results to understand whether the KF formula (Sirovich, 1996) could substitute the isotropic GMPEs inside ShakeMap in areas with few instrumental data. We calculated our scenario for Frignano after the QUEST data were available, and distributed it within our research unit; the focal parametres available within the 24 hours from the event were used, however. Mean intensities and standard deviations were calculated following our procedure (Sirovich and Pettenati, 2009, Soil Dyn. Earthq. Eng., 29, 1113-1122; Sirovich et al. 2009, Seism.Res.Letters, 80, 6, in press). The KF scenario of L’Aquila was calculated in the morning of April 6, 2009 with the source parametres by Mednet, USGS and INGV. In our parametric calculation, their differences were assumed as uncertainties. The resulting KF scenario was well balanced with the QUEST evidence, but out of centre (like both ShakeMaps were) mostly because of the directivity (not yet understood at the time) and of a small error in the first epicentral location. In both cases studied, the number of recording instruments was insufficient and for L’Aquila also source effects were crucial. For these reasons, in this case all attempts of reliable fast information failed (ours included). On the other side, after some weeks, and knowing the source better, our parametric KF scenario succeeded to fit both regional patterns of intensity observed in the field. The two cases studied suggest that neither ShakeMap® nor our parametric KF technique are able to garantee rapid and enough reliable information on damage for immediate post-event civil protection purposes. Rather, our hazard-planning oriented technique got two more validations and, thus, the free KF formula seems promising for improving existing rapid algorithms, but to obtain information for public consumption. RU S3_06: Fast evaluation of parameters and effects of strong earthquakes: a synthesis of first year results Saraò A.1, Barnaba C.1, Bragato P.L.1, Gentili S.1, Pettenati F.1, Sandron D.1, Sirovich L.1 1. Istituto Nazionale di Oceanografia e Geofisica Sperimentale, OGS, Trieste, ITALY, [email protected] We briefly list here the main results obtained by the OGS RU until now. All the topics will be detailed and discussed in further presentations given by the authors. As first, we activated the real time exchange of waveforms and metadata between OGS and INGV. We installed the ShakeMap tuned for Italy by INGV, and we made efforts for ensuring the coherency with the maps of ground-motions produced at other Italian data centers for the same earthquake. Also the real time computation of moment tensor has been set up. At now, moment tensor is automatically activated for ML ≥ 3.7 earthquakes occurred in NE Italy and surroundings. A web site to publish shake-maps and moment tensors has been realized within this project and will be soon on-line. For validating Shakemap results we assessed the impact of regional and site characteristics on the uncertainty of the available ground-motion predictive equations. Using the data of the national seismic network managed by INGV, we showed that site conditions contribute to about 30% of such uncertainty, while regional characteristics contribute for less than 4%. Major differences are located in north-eastern Italy, an area that is also affected by strong Moho reflection, which significantly enhances ground motion at hypocentral distances between 70 and 130 km. In order to improve ShakeMap results: 1) we started the site characterization of OGS stations installed in the Friuli Venezia Giulia region performing noise analysis and collecting station site information; 2) we investigated the effects of the finite source in shake-maps and, jointly with other S3 RUs, we studied the 2008 Miyagi earthquake Mw=6.9, chosen as target event for its optimal data coverage; 3) we applied the parametric KF scenario of intensity (Sirovich and Pettenati, 2009) to understand whether the KF formula could be used inside ShakeMap as a new type of GMPE (which includes source effects) mostly in areas with few instrumental data. The comparison between KF maps and the available intensity shakemaps for three earthquakes (2008 Jul. 29, Los Angeles earthquake, M=5.4; 2008 Dec. 23 Frignano earthquake, M=5.3 and the 2009 Apr. 6, L’Aquila Mw=6.3 event) reveals that the KF formula is promising for improving existing rapid algorithms. ShakeMap®, Community Internet Intensity Map, and a-posteriori KF intensity scenario of the M5.4 July 29, 2008 Earthquake in South Los Angeles; a comparison. Sirovich L.1, Pettenati F.1, Sandron D.1 1.Istituto Nazionale di Oceanografia e di Geofisica Sperimentale – OGS, Borgo Grotta Gigante 42c, 34016 Sgonico Trieste, [email protected] This presentation bases in part on Sirovich et al. (Seism.Res.Letters, 80, 6, 2009). We make two comparisons using the ShakeMap® intensities (USGS site, July 31, 2008). First, we compare them with the Community Internet Intensity Map of USGS (CIIM; also known as DYFI questionnaires). Second, we compare the mentioned ShakeMap® with our a- posteriori KF parametric scenario (calculated a few hours after the event following Sirovich and Pettenati, 2009, Soil Dyn. Earthq. Eng., 29, 1113-1122). Note that the traditional MMI “Felt Reports” are no longer available in the States; so, the quality of our intensity estimations and of ShakeMap® is not measurable. It is shown, however, that -in this case- the CIIM map can be taken as empirical benchmark for validation and the ShakeMap® is particularly reliable because of the large number of the recording instruments (mostly, west of the epicentre). A low-intensity zone was pointed out on July 29, 2009 by both CIIM and ShakeMap® downtown Los Angeles (in the Downey-Rowland-Hacienda area, west of the epicentre). Our interpretation has shown that: i) during this earthquake, important site effects only occurred far from the epicentre, in the western part of the San Fernando Basin; ii) the aforementioned relatively low intensities downtown Los Angeles do not imply permanent relative safety there, because they were caused principally by the near-field radiation from the source. Note that during the reviewing process of our paper, an M=4 earthquake with different mechanism occurred downtown Los Angeles, and its ShakeMap® was rather uniform in the meizoseismal area, thus confirming low site effects, if any. We also test the chance of KF of improving fast information after moderate earthquakes (an interesting task mostly in regions with insufficient instrumental coverage, where an isotropic, predictive, empirical equation is not satisfactory), which is our task in the DPC-INGV S3 project. In fact, our KF formula can also be thought of as an attenuation law that includes the source finiteness, the directivity and the radiation patterns of the horizontal components of the S body-waves (in this, we follow a suggestion by Molchan et al., 2004; Pageoph., 161, 8, 1725-1247). This is the second validation of our KF parametric scenario for hazard planning purposes. Finally, since the implementation of ShakeMap® is currently underway also in Italy (e.g., Michelini et al., 2008; Moratto et al., 2009), we hope that our work is of interest also to our Country, for general information to the public. UR9: Development of real-time algorithms for the detection of tsunami signals on sea-level records: testing and application to PMEL/NOAA (USA) and ISPRA (Italy) data Tinti S.1, Bressan L.1, Zaniboni F. 1 1.Università di Bologna, Dipartimento dI Fisica, Settore Geofisica, Viale Carlo Berti Pichat, 8 – 40127 Bologna, [email protected] In the framework of the Project S3, UR9 is focusing on the development and testing of real-time algorithms for the detection of tsunami signals on tide-gauge records, the basic idea being that of distinguishing the tsunami onset from the background signal. The procedure is conceives to run continuously: at each time step, a control function is computed which compares a parameter characterizing the previous background signal with another parameter representing the last minutes of the real signal. If the control function exceeds a given threshold a condition of tsunami alert is generated. The algorithm is intended to be applicable to both open-ocean sea-level signals (like the ones recorded by DART buoys) and coastal tide-gauge records. In the latter case, the challenge is to have the algorithm allowing to adapt the control function behavior to the characteristics of the wave oscillations at the specific site where the instrument is installed. Hence the algorithm must be flexible enough to work with parameters chosen on a case-by-case basis. UR9 is applying and testing the algorithm on data provided by PMEL/NOAA in Seattle and relative to DART buoys and coastal tide gauges, and also on data provided by ISPRA (ex APAT) relative to the Italian sea-level network and in particular to the stations of Porto Empedocle, Lampedusa and Trieste. The ISPRA data has not been received in the framework of S3 but ISPRA allowed UR9 to use the data in the project. We present the results of the application of the real-time algorithm on the two sets of data (PMEL/NOAA and ISPRA). In the first case, the algorithm is tested versus the occurrence of real tsunamis in the signals. As regards the Italian harbor data, that do not contain any tsunami occurrence, UR9 determined the typical oscillation periods for the harbor of Trieste for which data were sampled every minute. For Porto Empedocle and Lampedusa, the sampling rate of 10 minutes unfortunately do not allow for the determination of harbor eigenfrequencies, but it was anyhow possible to characterize interesting phenomena as the “marrobbio”, which is typical of the basin delimited by Sicily and Tunisia and that consists in rapid fluctuations of the sea level with amplitudes up to 1-2 m. Relationship between “Hai sentito il terremoto” web based macroseismic intensity and ground motion data Tosi P.1, De Rubeis V.1, Ferrari C.1, Sbarra P.1 1Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, [email protected] “Hai sentito il terremoto” is an online questionnaire, reachable at the address www.haisentitoilterremoto.it. that collects information voluntarily given by common people. The effects recorded in the questionnaire are statistically analyzed to extrapolate Mercalli-Cancani-Sieberg and European Macroseismic Scale intensities. The final result is the definition of the intensity degrees, averaged for each town or village, with the evaluation of the associated uncertainty. Maps of macroseismic intensity are displayed on- line in almost real time and continuously updated in case of availability of new data. The analysis of correlation between peak ground velocity (PGV) and the peak ground acceleration (PGA) with intensity was deeply studied in a large amount of works. The main difficulties in such analysis stems from both inadequate number of accelerometric stations and small number of evaluated intensity close to the former. The web questionnaire data allow us to overcome these difficulties. Our intensity data, in fact, are dense in space allowing a reliable interpolation of intensities where accelerometric stations are located. Moreover, intensity values cover even the lowest degrees extending the range in which correlations are usually defined. Our aim is providing a more general and detailed study of the relationships between intensity and ground motion that takes into account the whole spatial variability of the seismic field considering both low and high intensity degrees. A detailed analysis is shown with regard to data of L’Aquila earthquake on 6 April 2009 01:32 UTC. Session Project S4 Italian Strong Motion Database RU07 – Seismic characterization of RAN stations on rock mass and stiff soil Albarello D.1, Hailemikael S.4, Lunedei E.5 , Pileggi D.3 , Scarascia-Mugnozza G.2 1. Dip.Sci. della Terra - Università degli Studi di Siena, Via Laterina, 8, 53100 Siena, [email protected] 2. Dip.Sci. della Terra - Università degli Studi di Roma “La Sapienza”, P.le Aldo Moro 5, Roma, [email protected] 3. Dip.Sci. della Terra - Università degli Studi di Siena, Via Laterina, 8, 53100 Siena, [email protected] 4. Dip.Sci. della Terra - Università degli Studi di Roma “La Sapienza”, P.le Aldo Moro 5, Roma, [email protected] 5. Dip.Sci. della Terra - Università degli Studi di Siena, Via Laterina, 8, 53100 Siena, [email protected] The research activity focuses on the development and field application of fast and inexpensive experimental procedures for the seismic characterization of subsoil in correspondence of RAN stations located on stiff soil or rock mass. These stations play a major role in the definition of reference ground motion for seismic microzoning studies and the development of empirical attenuation relationships to be applied in PSHA studies and seismic scenarios modeling. A basic problem of these stations is that, in most cases, their location has not been preceded by any statisfactory geological/geophysical characterization of the local subsoil. This implies that in many cases, the stiff character of the underground and the eventual lack of significant stratigraphic resonance phenomena is inferred on the basis of very rough considerations. Thus, the development of experimental low-cost procedures to be extensively applied for the seismic classification of all the sites of this kind is mandatory. To this purpose, the RU is developing a multi-disciplinary approach based on the comparison of geophysical measurements and geomechanical analyses. As concerns geophysics, ambient vibrations techniques have been considered that are based on single station (HVSR) and array configurations. As concerns geomechanics, a number of parameters has been identified that allows the characterization of the rock subsoil in terms of jointing conditions (Jv) and average size of rock blocks (Ib). The basic idea is to find a link between the parameterizations deduced by geophysical and geomechanical data to identify configurations that are responsible of alteration of the expected ground motion (lateral heterogeneities, weakness zones related to rock mass weathering and/or faulting alteration of the rock masses, etc.) by taking into account the geological context of the considered area. On this purpose, possible effects of lateral variations in the subsoil mechanical properties have been of major concern. To check the reliability of this methodology, a set of 11 RAN stations have been selected after fast geological/geophysical surveys the have considered a larger set of RAN sites (15). In order to select the sites, several aspects have been considered: particular seismological interest (e.g., the sites located in the area of the recent L’Aquila earthquake), the presence of outcropping rock mass or stiff soil below the station,etc.. The resulting 11 sites are located in Calabria (VBM, VBV, MRM, SPS), Latium (MTC) and Abruzzo (CPS, AQG, PSC, SCN and AQP). Preliminary results concerning the considered sites will be presented and discussed in order to outline most problematic aspects. RU4-S4: Multimodal Inversion of Surface Waves for the Near Surface Characterization of National Accelerometer Network Stations P. Bergamo1, S. Foti1, M. Maraschini1, K. Tokeshi2. 1.Politecnico di Torino - Distr, Corso duca degli Abruzzi 24, 10129, Torino, [email protected], [email protected], [email protected] 2. Politecnico di Torino, - Ditag, Corso duca degli Abruzzi 24, 10129, Torino, [email protected] This note reports on surface wave surveys conducted near some stations of the Italian National Accelerometer Network (RAN) to estimate the shear wave velocity profile of the subsoil. In particular, the surveys are located in Piemonte, Liguria and Sicilia. In most sites, shallow bedrock is expected on the basis of geological information. Where deemed necessary and logistically possible. active and passive data have been used in order to retrieve a broader frequency range of the experimental dispersion curve, and consequently, to increase the penetration depth of the test. The passive dispersion curves are evaluated by the beam-forming method using 2D arrays, and the active ones are obtained by an f-k analysis of the ground roll present in active acquisitions. Refraction seismic analysis was also performed on active data in order to retrieve information on the shallower part of the subsoil before the inversion of dispersion curves and to locate the depth of the water table when possible. The presence of sound contrasts between adjacent layers or inversions in the shear wave velocity profiles can affect the energy distribution on the experimental dispersion curve, which can be generated by modal superposition. For this reason the inversion was performed using a Monte Carlo multimodal code which allows inverting dispersion curves generated by the superposition of more modes, both in the low and in the high frequency range. The inversion method is based on the Haskell-Thomson matrix method, which calculates dispersion curves for a horizontally layered elastic subsoil model as the zeros in the velocity frequency domain of the determinant of a matrix which depends on the model parameter. The misfit associated to a given subsoil modes is the norm of the vector containing the determinant of the Haskell Thomson method evaluated on experimental points. The inversion algorithm is a Monte Carlo algorithm: for a set of subsoil profiles randomly generated between boundaries which depend on the experimental dispersion curves the misfit is evaluated and the best fitting profiles represent the solution of the inversion. This algorithm allows to include all the branches of the experimental dispersion curve, without the need of specifying to which mode of the theoretical dispersion curve each experimental data point belongs to, and thus, to avoid the problem of the choice of the initial model which affects deterministic inversion algorithms. Identification of ITACA sites with distinctive features in their seismic response based on analysis of strong motion data Bindi D.1, Di Alessandro C.2, Giorgetti S.3, Luzi L.4, Pacor F.5, Paolucci R.6, Rovelli A.7, Smerzini C.8 1 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] 2 Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy, [email protected] 3 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 4 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] 5 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] 6 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 7 Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy, [email protected] 8 Doctoral School of Earthquake Engineering and Engineering Seismology, ROSE School, IUSS Pavia, Italy, [email protected] We illustrate an empirical procedure to identify strong motion stations of the ITACA database showing seismic response features that cannot be easily explained by simplified classification schemes, so that they are likely dominated by more complex stratigraphic or topographic site effects, or by interaction effects of the recording station with the hosting or surrounding structures. In the framework of Project S4 (Task 4: identification of ITACA sites and records presenting distinctive features in the seismic response), the objective is to enrich ITACA with additional information regarding the identification of such “anomalous” stations, or even anomalous seismic events, in order to improve the rational use of the database. To this end, we identified those stations, or earthquake events, for which the recorded spectral ordinates tend to fall on average significantly above/below the standard deviation band of the attenuation relationship. The method consists of computing the residuals of the recorded 5% damped response spectral acceleration with respect to the predictions from two selected empirical attenuation relationships (Bindi et al., 2008; Di Alessandro et al., 2008) calibrated on the Italian records. The two approaches differ by the site classification adopted, one based on the Sabetta & Pugliese (1996) scheme and the other based on the predominant period of the horizontal-to-vertical response spectral ratio (Fukushima et al., 2007). Residuals are corrected for the inter-event variability, to reduce the specific contributions of the seismic event, and to better highlight the effect of the specific features of the site response of the station. Residuals for all the stations in the dataset are shown as a function of period. To provide synthetic measures of the observed dependence of residuals on period, the average values of residuals in 4 representative period ranges are calculated. The stations exceeding the threshold of 1.65σ over at least one of the selected period range for the considered site class are identified and discussed. References Bindi D.,·Luzi L., M. Massa and F. Pacor (2009). Horizontal and vertical ground motion prediction equations derived from the Italian Accelerometric Archive (ITACA), Bull Earthquake Eng, DOI 10.1007/s10518-009-9130-9, published on line. Di Alessandro, C., F. Bonilla, A. Rovelli, and O. Scotti (2008). Influence of site classification on computing empirical ground-motion prediction equations in Italy, AGU Fall Meeting, San Francisco, CA, USA, 15-19 December 2008, paper n. S12A-05. Fukushima Y., Bonilla L. F., Scotti O., Douglas J. (2007). Site classification using horizontal- to-vertical response spectral ratios and its impact when deriving empirical ground-motion prediction equations, J. Earthq Eng., 11, 712-724. Sabetta, F. and Pugliese, A. (1996) “Estimation of Response Spectra and Simulation of Nonstationary Earthquake Ground Motions,” Bull. Seismol. Soc. Am., 86(2), 337-352. A New Site Classification of the ITACA Stations Based on the Predominant Period of H/V Response Spectra Ratios Di Alessandro C.1, Rovelli A.2 1.Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy, [email protected] 2.Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy, [email protected] Following the previous work by Di Alessandro et al. (2008 and 2009) we propose a site classification method based on the predominant period of ground motion at the site. The site predominant period is identified from the average horizontal-to-vertical (H/V) spectral ratios of the 5%-damped response spectra of Italian earthquake records. We have selected 602 three-component analogue and digital recordings from 120 earthquakes recorded at 214 seismic stations within an hypocentral distance of 200 km. Selected events are in the Mw range of 4.0 to 6.8 and the focal depth ranges from 5 to 40 km. Whenever possible, we classified each site by assigning them to one of six predominant period classes (in the range 0.05 to 2 seconds) that we propose as a modification of the Zhao et al. (2006) procedure. We then investigate the impact of this classification scheme on empirical ground-motion prediction equations (GMPEs). We adopted the same functional form of Fukushima et al. (2007) and we computed a nonlinear period- dependent regression that allowed us to derive site coefficients using the proposed six predominant period classes. Our empirical site classification scheme based on strong- motion data provides the opportunity to explore whether we can decrease the misfit by improving the site characterization of the Italian data set. Comparison of our results with GMPEs based on a conventional site classification provides only a small reduction of overall standard deviation. However, our scheme allows to recognize well distinguished behavior of the proposed classes, both in terms of predicted spectral shape and relative amplification with respect to rock sites. Furthermore, the small values of the standard error of the mean amplification (around 10%) give us confidence in the predictive capability of our approach. As a conclusion, the use of H/V spectral ratios in site classification results in promising highlights on capturing the signature of typically flat frequency-response sites as well as deep and shallow soil profiles, characterized by long- and short-period resonance, respectively, with the advantage of a relatively quick and inexpensive method. A comparison with the conventional classification based on Vs30 criteria indicates that our approach is able to distinguish sites with different spectral response that would be not distinguished in terms of Vs30. The tools we propose are especially useful in the recognition of rock sites to be used as reference stations, through the innovative classification criterion of a flat, unitary level in the H/V response spectra ratio. The new ITACA monograph: main features and data compiling Di Capua G. 1, Lanzo G. 2, Peppoloni S.1, Scasserra G.2 1Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605 – 00143 Roma, [email protected]. 2Sapienza Università di Roma, Via A. Gramsci, 53 – 00197 Roma, [email protected] One goal of Task 2 (S4 project) was to develop a new monograph for the collection and/or the organization of geological, geomorphological, geomechanics, geotechnical, geophysical and seismological data on the stations of National Accelerometric Network (RAN), managed by the Civil Protection Department. The structure of the monograph consists in No 11 cognitive modules with boxes and filling spaces for the insertion of data and thematic maps. The information is summarized in a final form to allow a morphological, lithological and seismic site classification. This new tool replaces the old ITACA monograph and includes numerous fields of information previously not planned. Furthermore, the representation of geotechnical and geophysical data can be done in a standardized way. Finally, following the suggestion of the Evaluation Committee, the RU1 in Milan, participating in the activities of Task 2, has prepared a web page for online compilation of ITACA monographs, which is being tested and modified to allow in a short time the insertion of all new monographs compiled. In the first year of the S4 project No 16 forms on accelerometric stations that recorded the Irpinia earthquake (1980) have been compiled, using the new format. The compilation has made use of existing data, especially geotechnical, these latter being completely re- analyzed and presented in a standard format. Furher, the results of a new geotechnical investigation (boreholes and CH tests) conducted during the previous S6 seismological project were also used. No 10 geological detailed forms from ENEL, compiled for some of these stations, were also retrieved. Finally, the presence of landslides in the sites using the IFFI database (Italian Landslides Inventory) was verified and all the geotechnical and geophysical data available were re-analyzed. Following the April 6, 2009 L’Aquila earthquake, the compilation activities were immediately moved to the No 57 stations that recorded the mainshock (Mw=6.3). Among these there are also some new digital stations that were not included in the initial sample to be analyzed. Therefore, No 136 monographs on these stations (SOGIN monographs), containing information about the station location, a geological map at 1:25,000 scale and geological sections at 1:25.000 and 1:2.000 scales, were also retrieved. The No 57 forms were compiled and are being to be uploaded on the ITACA Database. Data collected and on site surveys have also allowed the preparation of a report on the RAN stations located in the city of L'Aquila (http://esse4.mi.ingv.it/images/stories/Classificazione_Sito_Stazioni_RAN_AQ.pdf), which summarizes the morphological and lithological characteristics of the stations and provides site classification according to the EC8-NTC2008 subsoil categories. S4-Task 3: Seismic Characterization of the Sites of the Italian Accelerometric Network Foti S.1, Parolai S.2, D. Albarello3, Milana G. 4, Mucciarelli M.5, Puglia R. 6 1Politecnico di Torino, corso Duca degli Abruzzi 24, Torino, Italy, [email protected] 2Deutsches GeoForschugsZenrtum GFZ, Telegrafenberg, Potsdam, Germany, [email protected] 3University of Siena, Via Laterina 8, Siena, Italy, [email protected] 4Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, Roma, [email protected] 5University of Basilicata, V.le dell'Ateneo Lucano 10, Potenza, Italy, [email protected] 6Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Milano-Pavia, via Bassini 15, Milan, Italy, [email protected] Task 3 of project S4 is devoted to improve the knowledge of the sites of the Italian Accelerometric Network (RAN), mainly by providing a reference shear wave velocity profile for each station. Surface wave tests have been selected as the primary tool for the characterization, due to their flexibility and cost effectiveness. We report on the ongoing activities and we show some of the already available S-wave velocity profiles. The selection of sites to be investigated was driven by several criteria. The attention was pointed in particular to stations that recorded interesting events in the past and to the recently installed digital stations, also trying to obtain a good coverage for the whole Italian territory. A preliminary feasibility study, based on the available information and on Google Earth imagine analysis, has been conducted to exclude sites with complicated topography and with no sufficient space for performing surface wave measurements. Finally, considering the geological characteristics of the selected sites (thin, i.e. meters/tens of meters of soft sediments, or thick, i.e. hundreds of meters of soft sediments) it was decided to assign them amongst the different teams by considering their expertise. In particular, sites with thin sedimentary covers have been assigned mainly to UR with large experience in active methods and with appropriate instruments (multi- channel acquisition systems with high frequency geophones) while the deep basins were assigned to teams with large experience in passive source methods and equipped with short period seismometers. In intermediate situations, a combination of active and passive methods will be used to guarantee adequate depth of exploration and good resolution at shallow depth. A benchmark test of different techniques was carried out in Bevagna. The seismic characterization of stiff-soil and rock-mass sites (behaving as seismic bedrock), represents a critical aspect for effective and proper location of seismic and accelerometric stations and for the analysis of seismic response by using the reference station approach. Selected rock sites have been thoroughly investigated to assess the effects of faulting, jointing and weathering with a combination of surface wave surveys, NHV measurements and classical geomechanical approaches. The experience gained by the RUs within the project was important for the preparation and release of a deliverable summarizing advantages and disadvantages of the used methods. REXELite, internet-based record selection on ITACA. Iervolino I.1, Galasso C.2, Spinelli A.3, Paolucci R.4, Pacor F.5 1. Dipartimento di Ingegneria Strutturale, Università degli Studi di Napoli Federico II, via Claudio 21, 80125, Napoli, [email protected] 2. Dipartimento di Ingegneria Strutturale, Università degli Studi di Napoli Federico II, via Claudio 21, 80125, Napoli, [email protected] 3. IMTEAM - team QUALITY, Via Sigismondi 40, 24018, Villa d´Almé, Bergamo, [email protected] 4. Dipartimento di. Ingegneria Strutturale. Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133, Milano, [email protected] 5. Istituto Nazionale di Geofisica e Vulcanologia, sezione di Milano, via Bassini 15, 20133, Milano, [email protected] REXELite is an internet version, operating on the ITalian ACcelerometric Archive (ITACA, http://itaca.mi.ingv.it/), of REXEL, a software developed for automatic selection of ground motion suites for code-based structural analysis, and freely available at the website of Rete dei Laboratori Universitari di Ingegneria Sismica (ReLUIS, http://www.reluis.it/) project. REXELite allows to search for combinations of seven 1- or 2-components strong motion records, compatible on average with a specified target spectrum. More specifically, REXELite: (1) automatically builds code spectra for any limit state according to Eurocode 8 and the new Italian building code spectra; (2) finds the set of seven ITACA records having the most similar spectral shape with respect to the code, and whose average also matches the target spectrum in a user-specified period range and with the desired tolerance. The records are selected according to specific features, in terms of magnitude, distance and soil conditions, selected by the user. The selected combination of accelerograms may include unscaled (original) or amplitude-scaled records and may be used for code- compliant non-linear time history analyses of structures. REXELite was developed, in cooperation with ReLUIS, within the INGV-DPC S4 project (http://esse4.mi.ingv.it) funded by the National Department of Civil Protection. Monitoring of a hill in central Italy to study possible topographical effects: the case of Narni (TR) ridge Lovati S.(1), Massa M.(1), D'Alema E.(1), Marzorati S.(1), Gori S.(1), Falcucci E.(1), Maistrello M.(1), Bakavoli M.(2), Pacor F.(1) and R. Paolucci(3) 1.Istituto Nazionale di Geofisica e Vulcanologia, Sezione Milano-Pavia, via Bassini, 15, 20133 Milano 2.International Institute of Earthquake Engineering and Seismology (IIEES), Tehran, Iran 3 .Politecnico di Milano, Piazza L. da Vinci, 32 - 20133 Milano corresponding: [email protected] In the framework of the S4 Project – Task4- a monitoring activity was planned with the aim to investigate possible site amplification due to topographic irregularity affecting strong- motion stations. After geological and seismological surveys, the Narni hill (TR), in the Umbria region (Central Italy) was selected. Here, the analog RAN station, installed on the top of the hill, recorded only the 16 December 2000, Ml 4.1 earthquake, occurred at hypocentral distance of 5.5 Km. Narni is a little village built on an elongated rocky ridge with NW-SE orientation, characterized by a vertical West side and by a moderate slope in the East one. In order to find suitable sites for installation, in February 2009 a microtremor survey was performed both at the top and at the bottom of the hill. On the basis of the results, obtained by Nakamura method (1989), 7 velocimetric stations (belonging both to INGV-MI-PV and INGV-CNT Department) were installed in 10 different sites from March 25 to September 7, 2009 . As reference stations, three sites (NRN2, NRN3, NRN5) installed at the base of the ridge were selected. To complete the network, one site was installed at the mid of the East side (NRN1), while the other stations (NRN4, NRN6, NRN7, NRN8, NRN9 and NR10) were located on the top of the ridge, some of these nearly the vertical slope of the West side. The velocimetric network recorded 702 earthquakes in the magnitude range from 1.5 to 5.8 and epicentral distances up to 100 km. Most of the recorded events belong to the 6th April 2009 L'Aquila sequence. Preliminary results based on horizontal to vertical spectral ratios (HVSR; Lermo and Chavez-Garcia, 1993) and standard spectral ratios (SSR; Borcherdt, 1970) analyses show amplification peaks up to 6 in frequency band between 2 and 4 Hz, at stations installed on the top of the ridge. In particular, directional amplification in the transverse direction respect to the orientation of the ridge can be observed through the computation of rotated HVSR. Similar results were already found on noise measurements, showing larger amplification peak at frequencies around 3 Hz, when the HVSR is computed respect to EW component. Characterization of Italian strong-motion recording sites in the perspective of a new soil classification Luzi L.1, Gallipoli M.R.2, Pacor F. 1, Mucciarelli M. 3 1 INGV-Mi, Via Bassini, Milano, [email protected] 2 IMAA-CNR, C.da S.Loja, Tito Scalo (PZ), [email protected] 3 DiSGG-University of Basilicata, V.le Ateneo Lucano, Potenza, [email protected] The characterization of near surface geology is fundamental for calculating realistic ground motion shaking scenarios and mitigating the effects of damaging earthquakes. It is common practice to classify soils on the base of the average shear wave velocity of the topmost 30m. This classification scheme provides a quantitative description of the soil characteristics, but it does not take into account other factors, such as velocity inversion, non-linear soil behaviour, bi-dimensional effects at basin edges or the effect of different degrees of rock fracturation, among others. This works is carried out in the framework of the project “Italian strong-motion database” funded by the Italian Civil Protection Department, whose aims include the evaluation of the seismic response of strong-motion recording stations and to establish a soil classification scheme for calculating new ground motion prediction equations for the Italian territory. We collected data for about 90 sites in Italy, where the Vs profiles and detailed geological data were available, together with a relevant number of recordings of moderate and strong earthquakes. We calculated the site response using empirical techniques applied to the seismic recordings and simplified theoretical methods, and then we classify stations according to geotechnical or geophysical parameters (average Vs over different depths, resonance frequency, amplification value, etc.) The performance of soil classification was tested from the deviation from the acceleration response spectra provided by national and international seismic provisions or from the residual analysis performed on ground motion prediction equations. Alternatives to the use of Vs30 have been explored and the results suggest that for a correct determination of the expected ground motion additional parameters can be used such as soil fundamental frequency. RU1 INGV MI-PV: Activities carried out in the framework of the project S4 Italian strong-motion database Lucia Luzi1 and RU1 working group 1INGV Milano – Pavia, via Bassini 15, 20133 Milano, [email protected] The research unit of Milano – Pavia collaborated to all tasks of the S4 project. In detail, Task1 regards the update of the ITACA database (http://itaca.mi.ingv.it), which consisted in the improvement of the web portal facilities and the storage of the strong-motion data relative to 2005-2007. Because of the relevance of the recent Mw 6.3 L’Aquila and Mw 5.4 Parma earthquakes, the priority was given to the storage of the accelerograms relative to these events. A feasibility study has been performed for the interaction between the ITACA database and the software REXEL (http://www.reluis.it/), a tool for the selection of strong- motion data based on average spectral compatibility, with a positive feedback. In Task2 the RU collaborated to the design of the new reports of recording stations and to the implementation of the software for the automatic generation of station reports. In the framework of Task3, the RU collaborated to the execution of the low cost geophysical surveys in Emila Romagna and in Abruzzo (Modena, Faenza, Novellara, Onna among other sites). The activity carried out within Task4 consisted in the individuation of recording sites with peculiar behavior. This activity was conducted through statistical data analyses, to identify the deviation from average values predicted by ground motion prediction equations (GMPE’s), and through GIS spatial analysis, in order to detect recording sites located on complex landforms or alluvial basins. Regarding the site monitoring, the RU collaborated to the installation of a temporary network to collect data to investigate the response site of Norcia plian (PG). An additional temporary network was installed by the RU staff in the village of Narni (TR), to detect possible topographic site effects. Finally, in the framework of Task5, a work on site classification has been carried out by the statistical analysis conducted on different geotechnical and geophysical parameters of the recording stations with relevant number of available observations. All the recording stations in the ITACA database have been classified according to the EC8 code and to the Sabetta and Pugliese (1996) criterion. The sensitivity of empirical ground motion models on the proposed classifications has been tested through the variability of the standard deviation of the proposed models. RU2 - INGV RM1: RU2 activity report . Milana G.1, Di Capua G. 1, Rovelli A. 1 1Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605 – 00143 Roma, (email: [email protected], [email protected], [email protected]). As part of Task 2 the activities of data collection and compilation of new monographs produced 136 forms on new digital accelerometric stations containing: station coordinates and elevation, photographic documentation and aerial photo of the site. Geological map at 1:25,000 scale, geological sections, respectively at 1:25,000 and 1:2,000 scale are also included. Following the April 6, 2009 L’Aquila earthquake, the activities of new monographs compilation have been redirected on the n. 57 stations of the Accelerometric National Network (RAN) which recorded the mainshock. Site classification of ITACA stations according to the EC8-NTC2008 subsoil classes has been revised and the new version (2.0) is going to be released with the next ITACA website upgrade. The ITACA stations are all georeferenced and have been incorporated into a GIS for the assignment of the subsoil category on the basis of the lito-technical map derived from the Geological Map of Italy at 1:100,000 scale. The UR2 planned and realized in Task 4 a seismic network in Fucino Basin. The activity is aimed at detecting amplification effects in the basin and to relate them to the geometrical and geotechnical properties of soils. The network was based on 18 seismic stations deployed between October and January 2009 and was operative until October 2009. In the recording period the seismicity was dominated by L’Aquila earthquake sequence that produced many hundreds record nowadays only partially processed. Many events generated by other seismic region in a 30-40 km distance range were recorded. A preliminary spectral ratios analysis performed both on earthquakes and on seismic noise data is presented. The contribution of UR2 to Task 5 consists in an attempt to formulate a new site classification for the Italian territory which could be independent of the knowledge of the site geological conditions and not based on the conventional Vs30 intervals. The innovative approach we introduce for ITACA is based on the estimate of a site predominant period T. For sites where strong ground motions are recorded, the predominant period is identified from the average horizontal-to-vertical spectral ratios of the 5%-damped response spectra of records. The method formulated in Task 5 is applied to the determination of sites with relevant amplification/deamplification performed in Task 4 Achievements of UR5 UNI-BAS in project S4 M. Mucciarelli Department of Structures, Geotechnics, Applied Geology, University of Basilicata, Potenza, Italy. Contact: [email protected] The RU5 was involved in the following tasks: 1.Update of ITACA data base(Task1) 2.Surface waves procedures for site investigations (Task 3) 3.Identification of anomalous recordings (interaction with buildings- Task 4) 4.Site classification (Task 5) As for the update of the data base, it was agreed to transfer the data of 20 stations of the DiSGG-IMAA network. For each station the R.U. prepared also the identification sheet with all the available data in the format agreed within the ITACA project. As for the surface wave procedures, the task of the R.U. was to perform surface wave measurements to be transferred to the R.U. at UniSI. Six measurements were already carried out and one more is under way. As for the identification of anomalous recordings, the scope of the R.U. at UniBas was to identify the presence of possible building-soil interaction analysing the recordings at the ITACA stations. We analysed 37 station of the RAN The analysis technique was rather simple. We took the average of the rotational HVSR looking for a specific pattern: the signature of the translational modes of a building should be made of two directional peaks (the first higher) separated by 90° (rotational modes make this pattern more complex). We first analysed the effect for station hosted inside a building. We found that the two-peaks signature is always present within ENEL substations, ranging from 6-8 Hz to 12-18 Hz depending on the building’s typology (masonry or pre-cast r.c.). We then considered the interaction with structures located at some distance, starting from dams that are known to affect the recordings as shown for the Italian case of Tolmezzo (Ambiesta Dam) by Barnaba et al. (2007). We observed the structure “footprint” in both the examined cases, the dams at Villetta Barrea and Fiastra (Monte Fiegni station). After the L'Aquila Earthquake (6th April 2009), we analyzed the main recordings and performed the dynamic characterization of some housings (and the related pillars fixed into the ground on which accelerometer is located). All considered housings are located close to the epicenter: Bussi (BSS code), San Demetrio nè Vestini (SDM), Scafa (SCF) and Cittaducale (CTD). Possible influence of the housing in the high-frequency range is analysed and discussed in a separate poster. As for the site classification activity, the R.U. interacted with the R.U. at INGV-MI to create a data base of stations (even outside ITACA) having both good quality recordings and down-hole profile reaching at least 30 m. The R.U. re-analysed the data in its possess, providing the data for 47 sites. The statistical analysis of the joint data base are reported in a separate poster. ITACA: The New Italian Strong-Motion Database Pacor F.1, Paolucci R.2 and Working Group ITACA 1.Istituto Nazionale di Geofisica e Vulcanologia, Milano,Italy, [email protected] 2. Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] The main objective of Project S4 is to make available through the Internet an updated and improved release of the Italian strong motion database (ITACA), originally developed within project S6, in the framework of the 2004-2006 DPC-INGV agreement. This work starts from the alphfa version of ITACA (Luzi et al., 2008), where 2182 3-component records from 1004 earthquakes where processed and included in the database, together with the earthquake metadata, the recording station information and reports on the available geological-geophysical information of 452 recording sites, corresponding to about 70% of the total. Most records come from the National Accelerometric Netowrk (RAN), operated by DPC, that is expected to include 511 digital stations by 2011, with a 20-30 km average distance in high-risk seismic areas The beta version of ITACA, which will reach its final stage by the end of the project, around mid-2010, will include several improvements and additional features, namely: - strong motion records from other local and/or temporary networks, and from recent seismic events, in primis the L'Aquila earthquake of Apr 6, 2009, and its main aftershocks; - updated reports, with an improved format, on the available geological/geophysical information for all recording stations, including average HVSR from microtremors and earthquakes where available, and the measured Vs profile at about 60 stations, corresponding to about 10% of the total; - identification of stations and records showing distinctive features, either due to geological/topographic irregularities or due to seismic source effects; - on-line tools for selection of spectrum-compatible records (Rexelite). All records were re-processed with respect to the alfa version, with a special care to preserve information about late-triggered events and to ensure compatibility of corrected records, i.e., velocity and displacement traces should be truly the first and second integral of the corrected acceleration signal. Finally, ITACA is expected to be soon integrated within other international strong motion networks, such as COSMOS and NERIES, in order to promote the dissemination and use worldwide of the Italian strong motion records. Problems and solutions for processing strong-motion records in the Italian ITACA database Paolucci R.1, F. Pacor2, R. Puglia3, G. Ameri4, C. Cauzzi5, L. Luzi6, M. Massa7 1 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 2 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] 3 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] 4 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] 5 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 6 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] 7 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] In the process of reviewing the signal processing techniques of the alpha version of the ITACA Italian strong motion database, we have faced several problems to improve the quality and the usability of corrected records. Namely: • to check the accuracy and reliable frequency range of corrected records and compare them with the corresponding records available within other international databases, such as the PEER and the European Strong Motion Database; • to ensure the compatibility of corrected accelerograms, so that the no further correction is required to obtain by single and double integration the velocity and displacement traces; • to avoid whenever possible the retention of zero-pads at the beginning of the processed record, that would be inevitably erased by the end-user of the accelerogram, typically when used for computer time-consuming non linear time-history analyses; • to devise a specific treatment for late-triggered records, typically on the S- phase, to provide meaningful and usable corrected waveforms from analog instruments that recorded most of the Italian earthquakes up to 1990.. A novel procedure for processing the ITACA strong motion records has been devised with the objective of providing a rationale solution to the previous problems, on one side, and, on the other side, to be robust and reliable enough to be effectively used to re-process all the ITACA records, including the most recent ones from the Parma (December 2008) and L'Aquila (April 2009) earthquakes. The procedure is presented here, together with several comparisons with the corresponding corrected records from the PEER and ESMD databases. Research Activity of RU3 (Politecnico di Milano) in Project S4 Paolucci R.1, Smerzini C.2, Cauzzi C.3, Figini R.4 , Stupazzini M.5 1 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 2 Doctoral School of Earthquake Engineering and Engineering Seismology, ROSE School, IUSS Pavia, Italy, [email protected] 3 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 4 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 5 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] Besides contributing to Project coordination, UR3 has given its scientific contributions to Tasks 1 and 4, especially as regards: - Updated processing of ITACA records Corrected records in the alpha version of ITACA were carefully re-examined and compared with the corresponding records available within other international databases, such as the PEER and the European Strong motion Database. A new procedure for strong motion record processing has been devised, carefully checked, and implemented in cooperation with researchers of INGV Milano. This procedure ensures the compatibility of corrected accelerograms, so that the no further correction is required to obtain by single and double integration the velocity and displacement traces. Special care has been paid to a specific treatment of the late-triggered records, typically on the S-phase, to provide meaningful and usable corrected records also for this case that is quite frequent in the Italian database. All ITACA records, including the most recent ones from the Parma and L'Aquila earthquakes have been re-processed according to the new procedure. - Processing of strong motion records in the near-fault of L'Aquila earthquake A notable effort has been made together with researchers of INGV Milano to make available in due time within ITACA the strong motion records of the Parma (December 2008) and L'Aquila (April 2009) seismic sequences. Particularly, the near-field L'Aquila records have been carefully processed in the long-period range and compared with the up- to-date ground motion prediction models. - Identification of ITACA stations with distinctive features in their seismic response A procedure was developed in the framework of Task 4 to identify which stations or earthquake events exhibit a distinctive trend so that the recorded peak values of ground motion deviate significantly from the estimations provided by the attenuation relationships. In particular the method consists of computing the residuals of the recorded 5% damped response spectral acceleration with respect to the predictions from two selected attenuation relationships calibrated on the Italian data. - 3D numerical analysis of the seismic response of Gubbio basin A considerable effort was devoted to perform 3D numerical simulations of the seismic response of the Gubbio sedimentary plain (Central Italy) during the main shock of 1997 Umbria-Marche earthquake sequence. The spectral element code GeoELSE (http://geoelse.stru.polimi.it) for seismic wave propagation analyses in 3D heterogeneous media was used. The simulated ground motion turns out to be in good agreement with the recordings at two representative accelerometric stations at Gubbio downtown (GBP) and inside the Gubbio plain (GBP), at least between 0 and 2 Hz. Furthermore, the comparison between the results obtained by both 1D and 3D numerical modelling clearly highlights the limitations of standard approaches based on vertical plane wave propagation in horizontally layered media, to simulate earthquake ground motion in complex geological structures. Key publications within Project S4 Ameri A., M. Massa, D. Bindi, E. D’Alema , A. Gorini, L. Luzi, S. Marzorati, F. Pacor, R. Paolucci, R. Puglia and C. Smerzini (2009). The 6 April 2009, Mw 6.3, L’Aquila (Central Italy) earthquake: strong-motion observations, accepted for publication in Seismol. Res. Letters. Ameri G., Bindi D., Cauzzi C., Chioccarelli E., D’Alema E., Iervolino I., Luzi L., Marzorati S., Massa M., Pacor F., Paolucci R., Puglia R. (2009). Strong ground motion recorded during the L'Aquila seismic sequence. Special issue of Progettazione Sismica, Accepted for publication. Paolucci R. Long-period earthquake ground motion: recent advances and observations from the April 6 2009, Mw6.3 L'Aquila earthquake, Italy. Invited lecture at the ACES Workshop on Performance-Based Earthquake Engineering, Corfù, Greece, 5-6 July 2009. Stupazzini M., Paolucci R. and H. Igel (2009) “Near-fault earthquake ground motion simulation in the Grenoble Valley by a high-performance spectral element code”. Bull. Seism. Soc. Am., Vol. 99(1): 286-301. RU8 GFZ: Overview of the scientific activities with particular focus on the shear wave velocity profile estimations at the RAN sites and on the temporary seismic network installation in Norcia. Parolai S.1, Picozzi M. 1, Di Giacomo D. 1, Pilz M. 1, Bindi D. 1,2, Strollo A.1 1.Deutsches GeoForschugsZenrtum GFZ, Telegrafenberg 14473 Potsdam Germany, [email protected] 2INGV Via Bassini 15, 20133, Milano, Italy The activities carried out by the RU8 GFZ include (1) the field measurements and the preliminary analysis of the array data (Task 3) collected at 10 sites of the RAN, (2) the installation and de-installation of a seismological network in the area of Norcia (Task 4) and a preliminary analysis of the data, and (3) the preliminary numerical simulations of the 3D seismic response of the Gubbio Basin (Task 4). The responsible of the RU8-GFZ Stefano Parolai also participate as expert consultant to the scientific discussions and meeting promoted within Task 5. The seismic noise measurements in array configurations were carried out at ten sites, three in Basilicata, two in Central Italy (Umbria and Abruzzo) and five in Emilia Romagna. Twelve to seventeen EDL stations equipped with 1Hz short period sensors were used for the measurements. The analysis of the data is ongoing in cooperation with RU1. Additionally, array measurements of seismic noise were also carried out in Onna (Abruzzo) and the S-wave velocity profile estimated. From 17 January 2009 to 25 May 2009 UR8 installed, together with UR1 a seismological network in the area of Norcia. This site was selected as an interesting basin where complicated ground motion amplifications due to 2 or 3D effects can take place. The network was composed by fifteen EDL 24 bit acquisition systems equipped with Mark-L- 4C-3D short period sensors. A large number of local events have been recorded including the mainshock (although clipped within the valley) and the aftershocks of the Abruzzo sequence. During the de-installation of the network in May single station noise measurements were carried out. The analysis of the data is ongoing. UR8 released to the project the data set of weak motion recordings collected between September and October 2006 in Val D’Agri. While the large amount seismic noise collected by the stations provided interesting results about seismic noise amplitude variation within the valley, the earthquake data analysis, carried out in close cooperation with UR5, highlighted the difficulties in obtaining robust site response results due to the paucity of earthquake data with good signal to noise ratio. UR8 also contributed to the numerical simulation of the Gubbio basin, in a close cooperation with UR3 (PoliMi) and UR1. The results highlighted the strong amplification of ground motion within the basin as well as an extension of the significant ground shaking due to wave diffracted and reflected at the edge of the valley. GIS-based topographic characterization of recording stations of ITACA Pessina V.1, Paolucci R.2, Fiorini E. 3, Giorgetti S. 4 1. Istituto Nazionale di Geofisica e Vulcanologia Dept of Milano-Pavia, Milano, Italy, [email protected] 2. Dept. of Structural Engineering, Politecnico di Milano, Italy, [email protected] 3. EUCENTRE, Pavia, Italy, [email protected] 4. Dept. of Structural Engineering, Politecnico di Milano, Italy, [email protected] Surface topography can significantly affect earthquake ground motions: amplification effects have been documented from seismic records, although to a limited extent with respect to the corresponding stratigraphy effects. Morphometric analyses of high resolution digital elevation models (DEM), with the support of Geographic Information Systems (GIS), have been implemented to provide a practical tool for the identification of critical topographic sites. Simple GIS functions are used to calculate slope parameters and to classify critical ranges of inclination, while the identification of ridges or height difference require more complex procedures. A method has been firstly developed to perform analyses at national or regional scale: this approach identifies the simultaneous presence of zones of potential topographic amplification and of critical elements, such as recording sites . Critical slope and ridge detection maps have been elaborated for the whole national territory, based on the 30x30 m resolution GDEM (ASTER instrument built by METI and NASA). Topographic amplification effects depend not only on simple morphologic parameters of the sites, such as average slope angle, width and height of the relief, but also on the type of relief (isolated cliff or ridge) and on the location of the site relative to the relief. Therefore, a deeper level of analysis has been tested to provide a more detailed landform classification (Topographic Position Index algorithm, Jenness (2006)). This procedure is very time consuming, therefore it is more easily applied to local scale investigations. Herein, we present an example of application of both procedures, with particular attention to the recording stations located on the Apennine mountains in Central Italy (Marche region). 3D Numerical simulations of the seismic Response of the Gubbio alluvial basin during the 26 September 1997 Mw6 Umbria-Marche Earthquake Smerzini C.1, Stupazzini M.2, Pilz M.3, Paolucci R.4, Pacor F.5, Parolai S.6 1 Doctoral School of Earthquake Engineering and Engineering Seismology, ROSE School, IUSS Pavia, Italy, [email protected] 2 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 3 Deutsches GeoForschungsZentrum, Potsdam, Germany, [email protected] 4 Department of Structural Engineering, Politecnico di Milano, Italy, [email protected] 5 Istituto Nazionale di Geofisica e Vulcanologia, Milano, Italy, [email protected] 6 Deutsches GeoForschungsZentrum, Potsdam, Germany, [email protected] In the framework of Project S4 - Task 4 (Identification of ITACA sites and records presenting distinctive features in the seismic response) a considerable effort was devoted to the set up of large-scale 3D numerical model including the Gubbio sedimentary plain (Central Italy) and the causative seismogenetic source of the main shock of 1997 Umbria- Marche earthquake sequence (MW6.0 26.09.1997, time 09.40). 3D numerical modelling of seismic wave propagation aims at reproducing the seismic response of sites characterized by distinctive features induced by complex geological configurations and topographical irregularities. To this end, both the spectral element code GeoELSE (Stupazzini et al., 2009) and a hybrid Seismic rays-Finite Difference code (Oprsal et al., 2002) for seismic wave propagation analyses in 3D heterogeneous media were used. The most recent results inherited from Project S3 of the previous DPC-INGV agreement about the 3D structure of the bedrock-soft deposits interface of the Gubbio plain were used to calibrate the numerical models. Fault rupture is modelled through a kinematic representation adopting the parameters and the slip distribution proposed by Hernandez et al. (2004). As a reasonable approximation, a simplified homogenous description of the dynamic properties of the Gubbio plain deposits was adopted, based on a linear gradient of the S- and P- wave velocity with depth. The combination of a realistic model of the sedimentary plain with a proper characterization of the seismogenic source allows us to obtain numerical ground motion time histories that are in good agreement with the recordings at two representative accelerometric stations at Gubbio downtown (GBP) and inside the Gubbio plain (GBP), at least in the range of frequencies between 0 and 2 Hz. For both different computational codes, Peak Ground Displacement and Velocity values are in reasonable agreement with the estimates given by the most recently developed empirical attenuation relationships. Significant long period amplification and lengthening phenomena are found to occur inside the basin at GBP. The comparison between the recorded spectral ratios GBP/GBB with those obtained by both 1D and 3D numerical simulations underlines the need of realistic 3D numerical modelling to predict the combined effects of radiation pattern, propagation path in irregular geological structures and complex site effects, that may be strongly underestimated or neglected at all by standard approaches based on 1D wave propagation. References Stupazzini M., Paolucci R. and H. Igel (2009) “Near-fault earthquake ground motion simulation in the Grenoble Valley by a high-performance spectral element code”. Bull. Seism. Soc. Am., Vol. 99(1): 286-301. Hernandez B, Cocco M, Cotton F, Stramondo S, Scotti O, Courboulex F Campillo M (2004) “Rupture history of the 1997 Umbria-Marche (Central Italy) main shocks from the inversion of GPS, DInSAR and near field strong motion data,” Annals of Geophysics 47: 1355-1376. Oprsâl, I., J. Brokesová, D. Fäh, D. Giardini (2002) “3D hybrid ray-FD and DWN-FD seismic modeling for simple models containing complex local structures”, Stud. Geophys. Geod. 46, 711-730 Session Project S5 High resolution multi-disciplinary monitoring of active fault test-sites areas in Italy Test site “L’Aquila” Task 4 S5 project: L’Aquila fault system. A test site to understand the physical processes of the earthquake preparation and generation. Amato A. 1, Margheriti L1, Pantosti D. 2,Cecere G. 1, Govoni A. 1, 3, De Gori P. 1, Improta L. 2, Bruno P. 2, Cinti F. 2, Pucci S. 2,Marzocchi W2, Zollo A. 4, Neri G. 5 and Task 4 working group 1Istituto Nazionale di Geofisica e Vulcanologia, Centro Nazionale Terremoti 2Istituto Nazionale di Geofisica e Vulcanologia, Sismologia e Tettonofisica 3 OGS - Istituto Nazionale di Oceanografia e di Geofisica Sperimentale, Trieste 4 Universita’ Federico II, Napoli 5 Universita’ di Messina The occurrence of a moderate size earthquake in a dense monitored seismic region as the recent Mw 6.3 L’Aquila earthquake, represents a unique opportunity to investigate the complexity of earthquake rupture processes and to refine our observation and modelling tools, which can be of great support for the emergency management during future seismic crises. In this task we propose to integrate and coordinate a series of activities, experiments and data collections taking L’Aquila as a test-site, in these studies are involved most of the INGV researchers and personal of other Institutions. The results derived from this task will be made available to the other seismological projects. The task is organized in 6 work packages: RU5: High Frequency GPS as a potential contribution for monitoring a seismogenic structure Avallone A.1, Marzario M.1, D’Anastasio E. 1, D’Agostino N. 1, Cecere G. 2, D’Ambrosio C. 2, Falco L. 2, Abruzzese L.2, Cardinale V.2, De Luca G.2, Memmolo A.2, Minichiello F.2 & Zarrilli L. 2 1Istituto Nazionale di Geofisica e Vulcanologia, Centro Nazionale Terremoti, via di Vigna Murata, 605, 00143, Roma ([email protected]) 2Istituto Nazionale di Geofisica e Vulcanologia, Centro Nazionale Terremoti, Osservatorio di Grottaminarda, contrada Ciavalone, 83035, Grottaminarda (AV). The aim of this RU is to investigate the feasibility of creating a quasi-real time procedure able to quickly detect co-seismic displacements and, thus, to contribute in providing a quick estimation of a seismic event magnitude. Nowadays, the development of continuous Global Positioning System (cGPS) networks and the technological improvements of the instrumentation provide the acquisition of GPS data at different frequency rates (from 0.033Hz up to 20Hz) to contribute to different research field in geophysics, such as the detection of the strain accumulation on the faults and the fault rupture process history, and to improve other disciplines, such as engineering studies. Thus, in the last 5 years, the INGV deployed a permanent and real-time monitoring GPS network (RING, http://ring.gm.ingv.it) all over Italy. This network consists in about 130 stations acquiring at 1Hz or 30s sampling rate and transmitting the data (epoch-by-epoch) in real-time. Thus, this network represents a GPS reference network to which the possible co-seismic displacement detection in Irpinia (target of this Project Test Site) can be referred. The high-rate data acquisition at the sites in Irpinia represents one of the topics of the project, because the quick estimation of the potential co-seismic dislocations at RING sites after an earthquake strongly depend on the latency of the acquired data. The other main aspect of this RU for the 1st year consisted in learning the appropriate analysis of the acquired high rate GPS data. The need to obtain quick determinations of the co-seismic displacements at the CGPS sites after an earthquake, constrained us to process GPS data with an epoch-by-epoch analysis strategy. The first analysis has been performed in post- processing, using precise orbits and high rate clocks provided by JPL and using the Gipsy/Oasis II software precise point positioning strategy in kinematic mode. The comparison with the previous results obtained with GPS data acquired during the 2004 Sumatra earthquake, allowed us to calibrate and to confirm our procedure. The analysis of epoch-by-epoch GPS signal has also been performed in order to minimize not-seismic contributions (such as multipath and common-mode). Finally, 10Hz GPS data collected at two sites located on the activated fault during the destructive April, 6th, 2009 Mw 6.3 L’Aquila earthquake, have been analyzed with the same procedure. Here, we will show the first results on the 10Hz dynamic and static displacements for such a moderate earthquake. Tectonic evolution of the Tiber Valley between Perugia and Città di Castello Barchi M.R.1, Melelli L.1, Pazzaglia F.1, Pucci S.2, Saccucci L.1 1.Università degli Studi di Perugia, Piazza Università 1, 06123 Perugia, [email protected], [email protected], [email protected], [email protected] 2.INGV, Via di Vigna Murata 605, 00194 Roma, [email protected] The reconstruction of the long-term (2 My) tectono-sedimentary evolution of the Tiber Basin is aimed to better understand the present-day tectonic setting, which is characterized by active SW-NE extension, as indicated by both geodetic and seismicity data. During the first year of the project we followed three main lines of activity. A- Field mapping of the continental deposits infilling the basin, aimed to the stratigraphic reconstruction of the continental succession infilling the basin and of the main faults driving the basin evolution. The actual stratigraphy of these deposits is still largely uncertain and further effort is needed to better constrain their age. Moreover, seismic profiles show that the basin succession is several hundred meters thick, comprising older deposits, not exposed at the surface. B- Aerial photo interpretation of the alluvial plain and surrounding areas, in order to identify late-Pleistocene-Holocene tectonic morphologies and deformations of geomorphological markers due to neotectonic activity (e.g. river network anomalies; offset paleosurfaces; escarpments alignments; etc.). C- Digital Elevation Model (DEM) analysis on selected areas, with the main goal of quantifying subsidence and uplift rates. We analysed the area between Città di Castello and Pierantonio, considering the river network with longitudinal profiles, concavity index, knick-points and fluvial terraces correlations. The topographic analysis also investigates the flat surfaces in order to identify significant clusters of altitude slope and aspect values. All the collected data and interpretations have been implemented in a GIS environment (GeoDB). From north to south, the High Tiber Valley can be divided into three sub-basins, separated by transversal thresholds. Our preliminary results indicate that the evolution of these basin show significant differences. In fact the northernmost Sansepolcro basin is characterized by the absence of fluvial terraces, while the Umbertide basin and the Ponte Pattoli basin, where the valley is narrower, show three orders of fluvial terraces. The central Umbertide basin is characterized by wide catchments, elongated on the transverse discontinuity; the passage to the adjacent sub-basins is marked by anomalous catchments, with a headwater moving back and nonconforming with the main threshold. These differences can be temptatively related to the interference between the local subsidence, mainly driven by the normal fault segments, and the regional uplift, occurring on a larger scale. In the second year of the project, we will also compare our surface data with the seismic images (WP 1.4), in order to produce an integrated model of the tectonic evolution of the basin. Shallow imaging of the Tiber valley basin by high-resolution seismic profiling: preliminary results Bruno P.P.1, Castiello A.1, De Rosa D.3, Improta L.2, Pierdominici S.2, Punzo M.1, Varriale F.1, Villani F.2, Mirabella F.3, Montone P.2, Chiaraluce L4 1 Ingv, Osservatorio di Grottaminarda, Grottaminarda, [email protected] 2 Ingv, Dip. Sismologia e Tettonofisica, Roma, [email protected] 3 Università di Perugia, Dip. Scienze della Terra, Perugia, [email protected] 4 Ingv, Centro Nazionale Terremoti, Roma, [email protected] During September 2009, an high-resolution active seismic survey was carried out in the framework of the INGV-DPC Project S5 (Task1 Test Site “Alto Tiberina”, UR1-WP2). The main goal of the survey was the shallow imaging of the Tiber valley Quaternary basin in order to identify possible recent dislocations associated to the Alto Tiberina Fault system within these deposits to be correlated with the surface structures. The investigated area is located about 10 km to the south of Umbertide and is crossed by commercial reflection profiles, which constrain the large-scale structure of the basin (700-1000 m thick). However, shallow imaging of the basin and of the fault systems (down to 200-300 m) is quite poor, mainly because of frequent acquisition gaps and static problems. In order to overcome this drawback of industry data and accurately image the internal architecture of the basin, we collected two high-resolution profiles. The profiles are 1.6 and 2.1 km long, strike W-E, and are tied to the western and eastern margin of the basin, respectively. We used a non-conventional dense, wide-aperture acquisition geometry. Sources, provided by a vibroseis (IVI-MINIVIB), were recorded by a 168-multichannel device. Source and receiver spacing was 10 and 5 m, respectively. Consequently, the aperture of the geophone spread was 835 m. This acquisition geometry allows the collection of both multifold, common-mid-point, reflection data and of dense refraction data in the extended offset range (maximum 835 m). Record sections display a good signal to noise ratio, clear first arrivals even for far offset traces, and evident high-frequency shallow and deep reflections (down to 0.5 s TWT). Data processing and inversion are on-going and will be completed at the end of the Project. Common-depth-point processing of reflection data will provide high-resolution stack sections across the basin. These reflectivity images will be complemented by high-resolution, multiscale Vp models of the basin infill obtained by first- arrival traveltime tomography. RU4: Focal Mechanisms Of Low-Magnitude Earthquakes In The Messina Straits Region D’Amico S.1, Orecchio B.2, Presti D.3, Zhu L.1, Herrmann R.1, Neri3, 4 G. 1 Department of Earth and Atmospheric Sciences, Saint Louis University, 3642 Lindell Boulevard, 63108 St. Louis, MO, USA ([email protected], [email protected] ,[email protected] ) 2 Dipartimento di Fisica, Università della Calabria, Ponte P. Bucci, Arcavacata di Rende (CS), Italy ([email protected]) 3 Dipartimento di Scienze della Terra, Università di Messina, Salita Sperone 31, Messina, Italy ([email protected], [email protected]) 4 Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, Roma, Italy The Cut and Paste method (CAP; Zhu and Helmerger, 1996) computes focal mechanisms of earthquakes in a wide magnitude range by waveform inversion of Pnl and surface wave segments. Its application to low-magnitude events in the Messina Straits has allowed us to significantly increase the number of earthquake focal mechanisms available in this area for understanding the local geodynamics. CAP has proven to be effective for analyzing earthquakes even with magnitudes between 2.5 and 4 (see e.g., Zhu et al., 2006) for which traditional P-wave focal mechanisms fail because of the lack of sufficient observations. We analyzed earthquakes with magnitudes between 3 and 4 occurring in the Messina Straits in the last few years, recorded by the broadband stations of the Italian National Seismic Network and by the Mediterranean Very Broadband Seismographic Network. The solutions show that most of earthquakes are due to normal faulting with T- axes aligned northwest-southeast and the remaining ones are strike-slip. This finding supports the most recent model of the local geodynamics (Billi et al., 2007; Neri et al., 2009) according to which the Messina Straits area is transitional between the extensional domain of southwestern Calabria (related to residual rollback of the only in-depth continuous segment of the South Italy subduction system) and the compressional domain of western-central Sicily associated with continental collision started after local detachment of the subducting lithosphere. Automatic Detection, and P- and S-wave Picking Algorithm: an application to the 2009 L'Aquila (Central Italy) earthquake sequence. Di Stefano1, F. Aldersons2, L. Chiaraluce1, R., D. Piccinini3 and L. Valoroso1. 1 Centro Nazionale Terremoti – Istituto Nazionale di Geofisica e Vulcanologia; Via di Vigna Murata 605 – Roma ([email protected]; [email protected]; [email protected]) 2 Scientific Consulting - Amman – Jordan ([email protected]) 3 Sismologia e Tettonofisica – Istituto Nazionale di Geofisica e Vulcanologia;Via di Vigna Murata 605 – Roma ([email protected]) In order to process the enormous amount of digital waveforms continuously recorded at permanent and temporary seismic stations as quickly as possible, we implemented a semi-automatic procedure in order to identify local earthquakes and to provide consistently-weighted P- and S-wave arrival times. The automatic picking system Mannekenpix (Aldersons, 2004), originally working on vertical component data, has been improved to tackle 3-component data. In order to increase the reliability of P-wave and S- wave picking, the system is now virtually capable of discriminating P-wave samples and S- wave samples, among noise samples. This Identification is performed by a C5 decision tree (Quinlan, 1993) derived from training data. Five groups of predicting variables are included: Energy, Polarization, Spectral Power, Skewness and Kurtosis. In addition, the SEDSL algorithm (Magotra et al., 1989) is also used as a predictor. The picking procedure requires a preliminary calibration derived from a reference subset of high-quality manual picks. After calibration, the picking system is statistically able to mimic the picking by a human analyst and to provide consistent uncertainty estimates translated into picking weights. We illustrate very satisfying results of the successful automatic procedure showing P- and S-phase automatic readings for the L’Aquila sequence. These readings are fully comparable to those of a good human analyst allowing high quality earthquake locations of many low-magnitude events in an extremely short space of time. Within a day of continuous recordings, we obtain around 2600 triggers, 75% of these are very high quality locatable events (mean RMS 0.1s) with at least 25 phase readings. References Aldersons, F. (2004) Toward a Three-Dimensional Crustal Structure of the Dead Sea region from Local Earthquake Tomography. Ph.D. thesis, Tel Aviv University. Magotra, N., Ahmed, N. and Chael, E. (1989) Single-Station Seismic Event Detection and Location IEEE Transactions on Geoscience and Remote Sensing, 27, 1, 15-23. Quinlan, J.R. (1993) C 4.5: Programs for machine learning. The Morgan Kaufmann Series in Machine Learning, San Mateo, CA. Building 1D reference velocity model of the Irpinia region (Southern Apennines): microearthquake locations and focal mechanisms Matrullo E.1, Pasquale G.2, Satriano C.3, De Matteis R.4, Zollo A.5 (1) Università di Napoli “Federico II”, RisscLab [email protected] (2) Università degli Studi del Sannio [email protected] (3) AMRA scarl (Napoli), RisscLab [email protected] (4) Università degli Studi del Sannio [email protected] (5) Università di Napoli “Federico II”, RisscLab [email protected] The Irpinia region is one of the most active seismic zones of Southern Apennines. Large destructive earthquakes occurred both in historical and recent times, the last of which occurred on 1980 is the strongest event (Ms = 6.9) of the past century in the Southern Apennines. Today, the seismic activity is characterized by low to moderate earthquakes, mainly concentrated on and at the border of along the 1980 earthquake fault system. We analyzed the instrumental seismicity of the Irpinia region (Southern Apennines), recorded by the ISNet (Irpinia Seismic Network, AMRA) and the nearby INGV (Istituto Nazionale di Geofisica e Vulcanologia) stations, during the period August 2005 – October 2008. We re-picked P- and S wave arrival times for a total of 5685 P- and 3118 S- phases of 640 earthquakes, in a range of local magnitudes between 0.1 and 3.2. A first evaluation of picking consistency has been performed analyzing the modified Wadati diagram, which also provides an estimate of Vp/Vs between 1.8 and 1.9. Then, the picking quality has been assessed by performing a preliminary location and looking, for each station, at outliers on the histograms of residuals. The 1D velocity model used for location has been constructed as a 3-layer gradient model, by visually averaging several models of the area available in literature. For all these preliminary analyses we used the NonLinLoc location code. To improve the quality of the hypocentral locations we computed a one-dimensional (1D) velocity model that approximates the seismic structure of the study area by inverting P- and S-wave arrival times of 230 well-constrained earthquakes, using the VELEST code. This “minimum 1D velocity model” is complemented by station corrections, which can be used to take into account of possible near-surface velocity heterogeneities beneath each station. We relocated the earthquakes with the new velocity model and station correction. Relocations are well constrained and generally improved. Finally, we computed fault plane solutions for the selected earthquakes previously located, using first motion polarities (FPFIT code). Focal mechanisms calculated in this work are in large part normal and strike-slip solutions, and their tensional axes (T-axes) have a generalized NE-SW orientation. Project S5, UR 3 WP 2.4: Crustal motion along the Calabro-Peloritano Arc as imaged by twelve years of measurements on a dense GPS network: Mario Mattia1, Mimmo Palano1, Valentina Bruno1, Flavio Cannavò1 1Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Catania, Piazza Roma 2, 95123 Catania. E-mail: [email protected] In this work, we show the results of 12 years of continuous and survey-mode GPS measurements carried out along the western part of the Calabro-Peloritano Arc, from 1996 until the more recent acquisitions in 2008. The results highlight that a NW–SE-oriented ~0.15 μstrain/yr extension across the Messina Strait and the Aeolian–Tindari–Letojanni fault system is active. Moreover, a N–S compressive strain-rate (~0.65 μstrain/yr) is acting across Vulcano and Lipari Islands coupled with an extensional strain-rate of ~0.15 μstrain/yr in the E–W direction. Finally, taking into account the observed horizontal velocity field, an analytical inversion was performed to obtain a reliable model of deformation of the investigated area. The main results are consistent both with focal mechanism solutions and the current structural setting of the investigated area. Upper crust structure and tectonic evolution of the Altotiberina normal fault (central Italy) F. Mirabella1, M. R. Barchi1, F. Brozzetti2, A. Lupattelli1 (1) Università di Perugia, Dipartimento di Scienze della Terra, Piazza Università 1, 06100, Perugia, Italia, [email protected], [email protected], [email protected] (2) Università degli studi G. D’Annunzio, Dipartimento di Scienze della Terra, via dei Vestini 31, 66013, Chieti, Italia, [email protected] In the last years a structural control exerted by the Altotiberina low-angle normal fault (ATF) on the seismicity of the Umbria region (central Italy) has been hypothesized. However, the evaporites/basement boundary (EBB) could also exert a lithological control on the seismogenic layer thickness of the region. The area has been widely studied in the last years and many research articles describe the ATF geometry and tectonic evolution. Starting from the already acquired knowledges, the work of this WP is aimed at a detailed reconstruction of the Altotiberina fault geometry, of its splays and of the most significant stratigraphic markers. Such work is necessary in an area where a consistent amount of investments is expected for the installation of monitoring networks and which is becoming a test-site for earthquake mechanics studies in Italy. In fact, despite the amount of work already made, there still exist a number of topics which need to be addressed: i) the ATF breakaway zone is not weel defined; ii) no sequentially balanced cross-sections have been produced; iii) the three dimensional reconstruction of the subsurface structures can be implemented. A further motivation to increase the precision in the definition of the ATF geometry at depth is given by an ongoing project proposal for ICDP aimed at drilling the ATF at seismogenic depths. Within the present project the ATF and the EEB have been studied by using a large amount of surface and subsurface data. About 40 seismic reflection profiles were interpreted, calibrated with six wells and a large quantity of geological maps and field surveys. On the basis of these data and with the aid of innovative software (GoCAD, 3D Move, 4D Move, TKS), the 3D subsurface setting of the area can be reconstructed in detail and can be compared with the distribution of the seismicity. In order to estimate the total amount of extension (between 9.9 km and 8.5 km) and the cumulative displacement of the ATF (about 7.5 km), three geological sections across the fault were balanced through a sequential restoration technique. The sequential restoration also shows the kinematic interaction between the ATF and the antithetic Gubbio fault and the influence of the staircase geometry of the ATF on the evolution of the Tiber basin. On the basis of the inferred extension values, the long-term slip-rate across the ATF is comprised between 2.4 and 2.7 mm/yr. RU03 S5 wp2.1-2.2: Deployment of an on-land, off-shore seismic network to buil up an integrated seismic data archive Moretti M.1, Govoni1,3 A., Mangano1 G., D’Anna G., Mandiello1 A.G., Margheriti1 L., Mazza1 S., Patanè2, D., Selvaggi1, G. and “Messina 1908 - 2008" Working Group 1Centro Nazionale Terremoti - Istituto Nazionale di Geofisica e Vulcanologia 2 Sezione di Catania - Istituto Nazionale di Geofisica e Vulcanologia 3 Centro Ricerche Sismologiche - Istituto Nazionale Oceanografia e Geofisica Sperimentale - Udine A passive seismic experiment to monitor the Messina Strait [Margheriti et al., 2008] and the whole Calabrian arc began in October 2007 and it is still in progress. Overall about 16 temporary seismic stations were installed on land around the the Strait of Messina [Moretti et al., 2008] and 5 ocean bottom seismometers (OBS) [D’Anna et al., 2008; 2009] to better monitor the area largely covered by the sea. The network incorporates and upgrades existing seismic stations (more than 20) of the permanent networks located in area. All the data gathered in the Messina 1908-2008 project together with all metadata will be archived in the same SEED data bank and will be accessible by the whole scientific community trough an ArcLink server.'s assemblage of a database and integration of innovative technologies could transform our understanding of the crust and mantle structure of the active tectonics and seismic hazards of the Strait of Messina [Mandiello et al., 2009, poster]. References - D'Anna, R., Mangano, G., D'Alessandro, A., D'Anna, R., Passafiume, G., Speciale, S., Selvaggi, G., Margheriti, L., Patanè, D., Luzio, D., and Calò, M. (2008) - "MESSINA 1908- 2008" - Progetto di ricerca integrato sul'area Calabro - Peloritana: la campagna OBS/, Convegno: “1908 — 2008: Scienza e Società a 100 anni dal grande Terremoto”, Reggio Calabria (Italy) | December 10—12 2008- - D'Anna G., G. Mangano, A. D'Alessandro, R. D'Anna, G. Passafiume, S. Speciale, S. Passarello (2009) - Progetto "Messina 1908 - 2008". Rapporto della campagna OBS nell'area Eoliana e dello Stretto di Messina, Rapporto Tecnico INGV, N. 98. - Mandiello A.G., Govoni A., Di Stefano R., Moretti M., Chiaraluce L., Lauciani V., Pintore S., Baccheschi P., Margheriti L., Mazza S. and Selvaggi G. (2009) - Complete Uniform and Available Data archive at INGV , First Annual Meeting, 19-21 October 2009. - Margheriti, L. and Messina 1908-2008 team (2008) - Understanding crust dynamics and subduction in southern Italy, Eos Trans. AGU, 89(25), 225-226. - Moretti M., A. Govoni L. Abruzzese, G. Aiesi, P. Baccheschi, F. Criscuoli, G. D'Anna, R. D'Anna, G. De Luca, D. Franceschi, L. Giovani, F.P. Lucente, L. Zuccarello, A.G. Mandiello, A. Basili, A. Bono, C. Castellano, F. Criscuoli, S. Mazza, O. Torrisi (2008) - "Messina 1908-2008" - Progetto Di Ricerca Integrato Sull'area Calabro - Peloritana: La Campagna Sismica, Convegno: “1908 — 2008: Scienza e Società a 100 anni dal grande Terremoto”, Reggio Calabria (Italy) | December 10-12 2008. Microearthquake source and attenuation parameters in southern Apennines from the non-linear inversion of P- and S-displacement spectra Orefice A.(1), Zollo A.(1) and Convertito V.(2) (1) Università di Napoli Federico II [email protected] [email protected] (2) Osservatorio Vesuviano, INGV, Napoli [email protected] The objective of this work is the estimation by using spectral techniques of the main attenuation and source’s parameters (seismic moment, source radius, seismic energy and stress release) of microearthquakes currently occurring along and around the segmented fault system which caused the 1980, Ms 6.9 Irpinia earthquake. For this aim a selected data-set of 339 earthquakes recorded between 2005 and 2008 at the Irpinia (ISNet, AMRA) (25 stations) and INGV (10 stations) seismic networks has been compiled and re- located using manually picked first-P and S-arrivals. The total number of analyzed three- component records is 15297, for events with local magnitude ranging between 0.2 and 3.2 The P- and S-displacement spectra from velocity and acceleration sensors have been inverted using the non-linear Levenberg-Marquardt least-square algorithm for curve fitting and assuming the Boatwright(1980)’ spectral model. The anelastic attenuation effect is accounted by a constant-Q , attenuation operator in the frequency range 0.5-40 Hz, since preliminary analyses have shown that the frequency dependent Q- models do not provide a significant misfit improvement. The site response function and refined source parameter estimations are obtained by an iterative procedure based on the computation of displacement spectra residuals and stack at each receiver site. At smaller seismic moments, there appears an abrupt slope change which is the evidence for a violation of the self-similar scaling at very small magnitudes (Mw < 1.5). The plot of static stress drop vs. seismic moment confirms the transition at Mo about 1e12 Nm, above which the stress drop is nearly constant with a value of 30 MPa. The radiated seismic energy is computed from the integral of the spectrum and it increases with seismic moment, but with a slope greater than expected from the Wyss(1970)’ theoretical relation for apparent stress release. This implies that the apparent stress drop increases with the seismic moment in the explored magnitude range, which is stable result even applying the correction for the finite instrument bandwidth. RU03 S5 wp2.3: Quasi real-time Crustal Anisotropy Estimates Piccinini D.1, Pastori M2, Margheriti L.3 , Zaccarelli L. 4,5,Bianco F. 5 1Istituto Nazionale di Geofisica e Vulcanologia, Sismologia e Tettonofisica 2 Universita’ degli studi di Perugia 3Istituto Nazionale di Geofisica e Vulcanologia, Centro Nazionale Terremoti 4 IPGP, Paris 5 Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano The study of seismic anisotropy in the crust and uppermost mantle help defining the deformation field of the medium sampled by the seismic waves. In particular the anisotropic parameters in the crust individuate the fracture field geometries connected with the active stress field, and might be sensible to variations of stress contributing to define, from a seismological point of view, the strain field of the region. The first step of our work was focused on building a semi-automatic procedure analysing near real time seismic data in order to estimate the deformation and fracture field of the crust. We compare two different semi-automatic codes (SPY and ANISOMAT) based on different techniques independently developed under MatLab by two research team (INGV-RM and INGV-NA). Independent test of the codes over a common dataset recorded during a field seismic survey in Val d’Agri basin (Southern Italy) lead us to preliminary results that show a general agreement between the two codes, although small differences mainly due to the different sensitivity of the codes to the S-wave picking accuracy. We modify the selection criteria of the events and some parameters of the analysis to get consistent results form the two codes implementing the code ANISOMAT+. The second step was to apply the code based on the cross-correlation technique over another dataset collected during a field survey in Central Italy (Città di Castello area), In this analysis to confirm the robustness of the semiautomatic code we developed, we compare it (ANISOMAT+) with an other code, SHEBA developed by Bristol University team. Preliminary results show in total similar fast orientation and delay time values but there is not agreementfor single measurements because the selection criteria of the event- station pairs is very different between ANISOMAT+ and SHEBA. Our next test will be to experiment the code on the Messina 1908-2008 data and hopefully also in a quasi real-time chain ( Alto Tiberina test site) in order to routinely process data streams from permanent seismic network and collect a long time series of anisotropy estimates. RU03 S5 wp2.3: Quasi real-time Crustal Anisotropy Estimates Piccinini D.1, Pastori M2, Margheriti L.3 , Zaccarelli L. 4,5,Bianco F. 5 1Istituto Nazionale di Geofisica e Vulcanologia, Sismologia e Tettonofisica 2 Universita’ degli studi di Perugia 3Istituto Nazionale di Geofisica e Vulcanologia, Centro Nazionale Terremoti 4 IPGP, Paris 5 Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano The study of seismic anisotropy in the crust and uppermost mantle help defining the deformation field of the medium sampled by the seismic waves. In particular the anisotropic parameters in the crust individuate the fracture field geometries connected with the active stress field, and might be sensible to variations of stress contributing to define, from a seismological point of view, the strain field of the region. The first step of our work was focused on building a semi-automatic procedure analysing near real time seismic data in order to estimate the deformation and fracture field of the crust. We compare two different semi-automatic codes (SPY and ANISOMAT) based on different techniques independently developed under MatLab by two research team (INGV-RM and INGV-NA). Independent test of the codes over a common dataset recorded during a field seismic survey in Val d’Agri basin (Southern Italy) lead us to preliminary results that show a general agreement between the two codes, although small differences mainly due to the different sensitivity of the codes to the S-wave picking accuracy. We modify the selection criteria of the events and some parameters of the analysis to get consistent results form the two codes implementing the code ANISOMAT+. The second step was to apply the code based on the cross-correlation technique over another dataset collected during a field survey in Central Italy (Città di Castello area), In this analysis to confirm the robustness of the semiautomatic code we developed, we compare it (ANISOMAT+) with an other code, SHEBA developed by Bristol University team. Preliminary results show in total similar fast orientation and delay time values but there is not agreementfor single measurements because the selection criteria of the event- station pairs is very different between ANISOMAT+ and SHEBA. Our next test will be to experiment the code on the Messina 1908-2008 data and hopefully also in a quasi real-time chain ( Alto Tiberina test site) in order to routinely process data streams from permanent seismic network and collect a long time series of anisotropy estimates. Strain Accumulation Across the Messina Straits and Kinematics of Sicily and Calabria From GPS Data and Dislocation Modeling Serpelloni E. 1, Bürgmann R. 2, Anzidei M. 3, Baldi P. 4, Mastrolembo B. 1,5 1 Istituto Nazionale di Geofisica e Vulcanologia – Centro Nazionale Terremoti, Bologna, [email protected] 2 Department of Earth and Planetary Science, University of California, Berkeley, California, USA, [email protected] 3 Istituto Nazionale di Geofisica e Vulcanologia – Centro Nazionale Terremoti, Roma, [email protected] 4 Dipartimento di Fisica, Università di Bologna, [email protected] 5 Istituto Nazionale di Geofisica e Vulcanologia – Sezione di Bologna, Bologna, [email protected] We use Global Positioning System (GPS) velocities and dislocation modeling to investigate the rate and nature of interseismic strain accumulation in the area affected by the 1908 Mw 7.1 Messina earthquake (southern Italy), constrained in the framework of the complex central Mediterranean micro-plates kinematics. Our data confirm a change in the velocity trends between Sicily and Calabria, moving from NNW-ward to NE-ward with respect to Eurasia, and details a fan-like pattern across the Messina Straits where maximum extensional strain-rates are ~65 nanostrain/yr. Extension normal to the coasts of Sicily is consistent with the presence of SW-NE trending normal faults. Half space dislocation models of the GPS velocities are used to infer the slip-rates and geometric fault parameters of the fault zone that ruptured in the Messina earthquake, obtaining optimal values of 3.5 and 1.6 mm/yr for the dip-slip and strike-slip components, respectively, along a 30° dipping normal fault, locked above 7.6 km depth. By developing a regional elastic block model that accounts for both crustal block rotations and strain loading at block- bounding faults, we show that the measured velocity gradient across the Straits may be significantly affected by the elastic strain contribution from other nearby faults. In particular, when considering the contribution of the possibly locked Calabrian subduction interface onto the observed velocity gradients in NE-Sicily and SW-Calabria, we find that this longer wavelength signal can be presently super-imposed on the observed velocity gradients in NE-Sicily and Calabria. The inferred slip rate on the Messina fault is impacted by elastic strain from Calabria; higher locking on the subduction thrust allow for substantially higher slip rates on the normal fault. Ambient noise levels and seismic detection threshold of ISNet network, Southern Italy Vassallo M.1 and Bobbio A.2 1 Analisi e Monitoraggio del Rischio Ambientale (AMRA) Scarl, via Nuova Agnano, 11- 80125 Napoli, Italy, [email protected] 2 Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano-sezione di Napoli, [email protected] We have characterized the background seismic noise in the Irpinia region (Southern Italy) by means of velocity Power Spectral Density (PSD) analysis. We selected a large number of one hour waveform segments since January 2008 to April 2009, from continuous data acquired by the recently installed Irpinia Seismic Network (ISNet). ISNet is a dense network displaced over an area of approximately 100 × 70 km2, composed by seismic stations equipped with a strong-motion accelerometer and a three-component seismometer (short period or broadband sensor). For each station we computed the velocity power spectral density useful for analyzing the diurnal and seasonal seismic noise variability in the area. Following the approach proposed by McNamara and Buland (2004), we performed a statistical analysis of all PSD curves by means of the noise Probability Density Functions (PDF) to estimate both the overall station quality and the level of earth noise at each site. In fact, using the PDF values, we have calculated for each frequency and for each component the percentile levels of PSD. Finally, in order to evaluate the spatial variability of the detection thresholds of network for local earthquakes, we compared the seismic noise amplitude levels at seismic stations with the theoretical spectral curves computed using the Brune (1970) source modeling. We computed the minimum Mw magnitude detectable by the network for each hypocenter point located at nodes of a regular grid covering the whole network. A detection is declared if the theoretical Brune spectra exceeds the 90° percentile noise levels at least 5 ISNet stations. The resulting map is useful for characterizing the performance of ISNet network and should be relevant to plan the best places of future siting seismic stations. RU S5/6 – DSF-UniNA: Research Activities in the Frame of the S5 Project Zollo A.1, Festa G.1, Maercklin N.1, Satriano C.2, Amoroso O.1, Bobbio A.3, Iannaccone G.3, Matrullo E.1, Orefice A.1, Stabile T.A.2, Vassallo M.2 1. Università di Napoli Federico II 2. RISSC-Lab AMRA scarl, Napoli 3. INGV, Osservatorio Vesuviano, Napoli The research activity of RU6 is targeted at the Irpinia test site, one of the most active seismic zones of the Southern Apennines. Large destructive earthquakes occurred both in historical and recent times in this region, which was struck in 1980 by the strongest event (Mw = 6.9) of the past century in the Southern Apennines. The main research facility of the area is the Irpinia Seismic Network (ISNet), managed by AMRA (Analysis and Monitoring of Environmental Risks). ISNet is a high-density, high- dynamics seismic and accelerometric network, built on innovative technological and methodological concepts, and focused on real-time data acquisition, processing and modeling. ISNet complements the INGV network in the region, making the Irpinia test site one of the highest instrumented seismic areas in Italy and an ideal site for experimenting new approaches for seismic monitoring and imaging of active fault systems. In the framework of S5 project, the research carried by RU6 is based on the real-time and off-line analysis of noise and microearthquake data collected by the ISNet and INGV networks. In particular, we are actively developing and experimenting new methodologies in the following research fields: • Seismic noise analysis and green functions: use of the random wavefield to retrieve images of the sub-soil through cross-correlation and stacking of continuous recorded signal; study of the ambient noise level and the seismic detection threshold for ISNet. • Refined estimates of micro-earthquake source parameters: retrieval of high resolution images of the fault system through the accurate determination of location, size and fault mechanisms in the magnitude range 1