WORKSHOP

The Colfiorito earthquake 1997-2007: ten years on

Rome, Italy

8-10 October, 2007

Istituto Nazionale di Geofisica e Vulcanologia Università di Roma TRE

EXTENDED ABSTRACTS

Scientific and organizing committee: Enzo Boschi

Franco Barberi Salvatore Barba

Lucia Luzi Paola Montone

Salvatore Stramondo To cite: The Colfiorito earthquake 1997-2007: ten years on. Extended abstracts. 8-10/10/2007, Rome, Italy, edited by: Barba et al., 2007, http://hdl.handle.net/2122/2533.

ii

This book of extendend abstracts has been deposited on the Earth-Prints Repository (http://www.earth-prints.org) under the Creative Commons Attribution-Share Alike 3.0 Unported License (license details at http://creativecommons.org/licenses/by-sa/3.0/)

Sponsored by: Istituto Nazionale di Geofisica e Vulcanologia Università di Roma TRE IPT srl

iii

Table of Contents

iv Contacts 1 The 1997-98 Umbria-Marche (central Italy) seismic sequence: lessons and open questions, by

Massimo Cocco 3 The contribute of structural geology and regional tectonics to the definition of large-scale

seismotectonic provinces and individual seismogenic sources: application to peninsular Italy and Sicily, by Giusy Lavecchia and GEOSIS-LAB researchers

7 Surface fractures associated with the 1997 Umbria-Marche earthquake sequence: prospects from a decade of surveys, analyses, and debates, by Francesca R. Cinti

10 Reflection and earthquake seismology: a first step towards integration in the Colfiorito area, by Alfredo Mazzotti

14 Seismic reflection data in the Umbria-Marche active area: limits and capabilities to unravel the subsurface structure, by Francesco Mirabella

16 Fault zone characterization as a challenge for em prospecting methods in the last decade: the Colfiorito case-history, by Agata Siniscalchi

21 Geochemical evidences of fluid involvement in the Northern Apennine seismogenesis, by Giovanni Chiodini

25 The 1997-98 Umbria-Marche earthquake: “geological” vs. “seismological” faults, by Massimiliano R. Barchi

31 Developments of seismic monitoring in Italy from 1997 to 2007, by Alessandro Amato 35 Can 3D and 4D seismic tomography help understanding the 1997 Umbria Marche seismic sequence?

by Claudio Chiarabba, Pasquale De Gori, and Davide Piccinini 37 Contribution of seismic moment tensors to seismic deformation and stress field studies in Italy, by

Andrea Morelli and Silvia Pondrelli 41 Spatio-Temporal Modeling of the Umbria-Marche 1997 Seismic Sequence, by Warner Marzocchi 45 The coseismic ground deformations of the 1997 Umbria-Marche earthquakes: a lesson for the

development of new GPS networks, by Marco Anzidei 49 SAR Interferometry for the analysis of crustal deformation in Italy: from the Umbria-Marche co-

seismic displacement field to mapping the strain accumulation on active faults, by Stefano Salvi 53 Extended sources of the main events of the Umbria-Marche (1997) seismic sequence inverted from

geophysical data, by Maria Elina Belardinelli 57 Macroseismic survey of the 1997-1998 Umbria-Marche earthquakes: from practice to practice, by

Romano Camassi 61 Codes, models, and reality: reductionism vs. holism in 10 years of microzonation studies in the

Umbria-Marche region, by Marco Mucciarelli 63 Site effect studies: what we can learn from Colfiorito experience, by Giuliano Milana and Antonio

Rovelli 67 Ground motion shaking scenarios for the 1997 Colfiorito earthquake, by Antonio Emolo, Aldo Zollo,

Vincenzo Convertito, Francesca Pacor, Gianlorenzo Franceschina, Giovanna Cultrera, and Massimo Cocco

70 From the “seismic classification” to the design ground motion parameters for the Italian building code, by Massimiliano Stucchi, Carlo Meletti, Fabio Sabetta, and Dario Slejko

75 Vulnerability evaluations for seismic risk assessment, by Mauro Dolce

iv

Contacts

Amato, Alessandro, Istituto Nazionale di Geofisica e Vulcanologia, Centro Nazionale Terremoti, amato@ingv.it

Anzidei, Marco, Istituto Nazionale di Geofisica e Vulcanologia. Centro Nazionale Terremoti, Via di Vigna Murata, 605 Roma, anzidei@ingv.it

Barba, Salvatore, Istituto Nazionale di Geofisica e Vulcanologia, Roma, barba@ingv.it Barberi, Franco, Università di Roma TRE, barberi@uniroma3.it Barchi, Massimiliano R., Dipartimento di Scienze della Terra - Università di Perugia - Gruppo di Geologia

Strutturale e Geofisica, mbarchi@unipg.it Belardinelli, Maria Elina, Università di Bologna, Dipartimento di Fisica, Viale Berti-Pichat 8,

40127 Bologna, mariaelina.belardinelli@unibo.it Boschi, Enzo, Istituto Nazionale di Geofisica e Vulcanologia, Roma, presidente@ingv.it Camassi, Romano, Isituto Nazionale di Geofisica e Vulcanologia, Bologna, camassi@bo.ingv.it Chiarabba, Claudio, Istituto Nazionale di Geofisica e Vulcanologia, chiarabba@ingv.it Chiodini, Giovanni, Istituto Nazionale di Geofisica e Vulcanologia,, Sezione di Napoli,

chiodini@ov.ingv.it Cinti, Francesca R., Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, Roma,

cinti@ingv.it Cocco, Massimo, Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, Roma,

cocco@ingv.it Dolce, Mauro, Dipartimento della Protezione Civile, Ufficio Valutazione, prevenzione e mitigazione del

rischio sismico e attività ed opere post-emergenza, Via Vitorchiano 4, Roma, mauro.dolce@protezionecivile.it

Emolo, Antonio, Università di Napoli “Federico II”, Dipartimento di Scienze Fisiche, Napoli, antonio.emolo@na.infn.it

Lavecchia, Giusy, Università di Chieti “G. d’Annunzio”, Dipartimento di Scienze della Terra, Laboratorio di Geodinamica e Sismogenesi, Campus Universitario, 66013 Chieti Scalo, Italia, glavecchia@unich.it

Luzi, Lucia, Istituto Nazionale di Geofisica e Vulcanologia, Milano, luzi@mi.ingv.it Marzocchi, Warner, Istituto Nazionale di Geofisica e Vulcanologia, sezione di Bologna,

marzocchi@ingv.it Mazzotti, Alfredo, Università di Pisa, Dipartimento di Scienze della Terra, Via S.Maria 53, 56123 Pisa,

mazzotti@dst.unipi.it Milana, Giuliano and Rovelli, Antonio, Istituto Nazionale di Geofisica e Vulcanologia, Roma,

milana@ingv.it Mirabella, Francesco, Università di Perugia, Dip. Scienze della Terra, Geologia Strutturale e Geofisica,

Piazza Univ. 1, 06100 Perugia (Italy), mirabell@unipg.it Montone, Paola, Istituto Nazionale di Geofisica e Vulcanologia, Roma, montone@ingv.it Morelli, Andrea and Pondrelli, Silvia, Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna,

Via Donato Creti 12, 40128 Bologna, morelli@bo.ingv.it, pondrelli@bo.ingv.it Mucciarelli, Marco, Università della Basilicata, Dipartimento di Strutture, Geotecnica e Geologia

Applicata, Potenza, Italy, Salvi, Stefano, Istituto Nazionale di Geofisica e Vulcanologia, Centro Nazionale Terremoti, salvi@ingv.it Siniscalchi, Agata, Dipartimento di Geologia e Geofisica, Università di Bari, sinia@geo.uniba.it Stramondo, Salvatore, Istituto Nazionale di Geofisica e Vulcanologia, Roma, stramondo@ingv.it Stucchi, Massimiliano, Istituto Nazionale di Geofisica e Vulcanologia, Milano, stucchi@mi.ingv.it

1

The 1997-98 Umbria-Marche (central Italy) seismic sequence: lessons and open questions

Massimo Cocco Istituto Nazionale di Geofisica e Vulcanologia

The earthquakes that struck Umbria-Marche in 1997 and 1998 (also known as the Colfiorito seismic sequence) have allowed the collection of an unprecedented quantity of good quality data for normal faulting which have promoted numerous studies aimed at investigating the seismogenic processes as well as the anatomy of the complex fault system activated during the sequence. Seismological data (hypocentral locations, seismograms and strong motion data), geological observations (geometry and mechanisms of the pre-existing fault systems) and geodetic measurements (GPS and InSAR data) have been simultaneously available for the first time for an Italian earthquake, whose analyses have allowed getting a detailed knowledge of source processes. This sequence is certainly one of the most important examples of normal faulting earthquakes within the international scientific literature. In this presentation the available multidisciplinary observations will be discussed together with the numerical modelling attempts in order to image the architecture and the mechanical behaviour of this complex fault system.

The 1997 Umbria-Marche seismic sequence consists of six moderate magnitude earthquakes (5<Mw<6) and thousands of aftershocks that in 40 days activated a ~45 km long, NW-trending, fault system. These seismic events ruptured shallow (1–8 km) normal fault segments, dipping to the SW. All the main shocks and most of the aftershocks show similar focal mechanisms. The sequence is characterized by an evident migration of seismicity along the strike of the main faults from NW to SE and the progressive activation of adjacent and nearby parallel segments. Seismicity is mostly located on the hanging-wall of main faults and it is absent in the foot-wall. The two largest events of the sequence (Mw 5.7 and 6.0) struck the Colfiorito area on September 26th (within 9 hours from each other) and ruptured two normal faults with opposite rupture directivity. Seismicity was mainly concentrated in this area until the beginning of October, when two other main shocks occurred on October 3rd and 6th (Mw 5.2 and 5.4, respectively). Starting from October 4th seismicity began to migrate towards SE where two other main shocks occurred on October 12th and 14th (Mw 5.2 and 5.6, respectively). The seismogenic volume is elongated in the Apenninic direction and seismicity occurred until November 1997 along the whole fault system. On April 3rd 1998 a Mw 5.1 normal faulting earthquake struck the area of Gualdo Tadino, further extending the seismogenic area to the North. On October 16th 1997 a strike slip aftershock (Mw 4.3) ruptured a nearly NS shallow structure inherited by previous tectonics. The analyses of seismicity pattern and local tectonic setting from structural geology suggest that NS inherited compressional structures were reactivated during the 1997-98 seismic sequence with strike slip earthquakes and the inversion of slip direction. This finding is consistent with coseismic stress changes caused by the normal faulting main shocks.

The geometry of each segment is quite simple and consists of planar faults gently dipping toward SW with an average dip of 40°–45°. The fault planes maintain a constant dip through the entire seismogenic volume, down to 8 km depth. Seismological observations do not show any evidence for listric faults. The observed fault segmentation has been attributed to the lateral heterogeneity of the upper crust: pre-existing thrusts inherited from Neogene's compressional tectonic intersect the active normal faults and control their maximum length. The NS lateral terminations of these compressional structures have been mapped at the surface as right-lateral minor transfer faults, which are not consistent with the present day active stress field. The stress tensor obtained by inverting the focal mechanisms of the six largest magnitude events of the 1997 sequence is in agreement with the tectonic stress active in the inner chain of the Apennine, revealing a clear NE-trending extension direction.

Each main shock had its own aftershock sequence with a temporal decay well described by an Omori-like power law. Nevertheless, the earthquake sequence shows a particular temporal evolution of seismicity with clear fluctuations of the rate of earthquake production during the sequence not only associated with the occurrence of a main shock. Moreover, the earthquake sequence has also shown unequivocal evidence of interaction among the different segments of fault system and the reactivation of

2

portions of faults inherited by the previous tectonic regime. Indeed, the spatial pattern of seismicity has been interpreted in terms of elastic stress changes, but the presence of normal faulting aftershocks on the hanging-wall of the main fault planes and the extraordinary migration of seismicity have been interpreted in terms of fluid flow and pore pressure relaxation.

Several features make the lessons emerging from this sequence particularly interesting: (1) extremely accurate earthquake locations (obtained by double-difference and cross-correlation techniques providing errors of the order of tens of meters) allow us to infer the main features of the spatial and temporal patterns of seismicity ; (2) seismicity migrated toward the SE-direction at nearly 1 km/day progressively activating the main fault planes before the occurrence of their impending mainshocks; (3) the analysis of coseismic stress changes shows that these normal faults interact with each other, but it cannot explain the presence of seismicity in the hanging wall of the main faults as well as the temporal migration of seismic events; (4) there are evidence of the presence of fluids in this area and it is likely to expect that they played a major role in earthquake triggering; (5) there is unequivocal evidence of fault reactivation: stress changes caused by normal faults reactivated inherited shallow strike-slip structures with slip inversion; (6) the heterogeneous distribution of slip on the fault planes has been constrained only for the three largest magnitude main shocks (the two events of September 26th and the October 14th Sellano earthquake) and, although the small dimension of the rupture planes, there are striking evidence of rupture directivity which strongly affected the observed ground shaking.

Despite the advanced state of knowledge, several open questions still exist that is useful to discuss for better understanding the mechanical behaviour of a complex system of normal faults. One certainly concerns the occurrence of the October 3rd main shock. It is impossible to prove if this event is located in the hanging-wall or on the same fault plane that ruptured during the September 26th main shock. In this latter case, it is important to point out that it nucleated within a high slip patch of the September 26th shock. A candidate mechanism to explain the repeated dislocation of the same fault portion during two distinct events is the fluid pressurization of the fault zone: the increase of pore-pressure driven by the previous dislocation can contribute to reload the fault patch in a relatively short time. Other evidence suggests that seismicity on the hanging wall of the main faults can be driven by a pressure pulse originated co-seismically from a deep source of trapped high-pressure fluids propagating into the damage region created at the time of the earthquake. Both these interpretations can explain the occurrence of the October 3rd earthquake in the stress shadow zone caused by previous main shocks. However, the role played by fluids in the earthquake generation process has only been inferred from numerical simulations and theoretical modeling. Another debated issue concerns the interpretation of coseismic surface ruptures: despite the shallow depth of mainshocks’ hypocenters and fault planes, the largest magnitude earthquakes did not produce clear evidence of a fault scarp at the earth surface.

Finally, the repeated occurrence of similar magnitude main shocks in a relatively short time has created a huge impact to society and complicated the management of the emergency activities. Despite the relevant scientific progresses in the understanding of earthquake mechanics and interaction, a useful forecasting of seismicity evolution for concrete real time applications is far to be achieved. This is one of the priorities of seismological and geophysical investigations and an important challenge for future research.

3

The contribute of structural geology and regional tectonics to the definition of large-scale seismotectonic provinces and individual seismogenic sources: application to peninsular

Italy and Sicily.

Giusy Lavecchia and GEOSIS-LAB researchers (Paolo Boncio, Francesco Brozzetti, Rita De Nardis, Deborah Di Naccio, Bruno Pace, Francesco

Visini) Università di Chieti “G. d’Annunzio”

The essential contribute of structural geology in the identification and parameterization of individual seismogenic source is now largely recognised. When both structural and seismological good-quality data are available, such as in the case of the Colfiorito 1997 sequence, the 3D fault geometry and segmentation pattern, the fracture propagation pattern, the kinematics and the associated stress tensor can be recognised and described in detail (Chiaraluce et al., 2005 and reference therein). The identification of the 3D geometry of historical and/or silent sources, which is fundamental for the prediction of the maximum associate earthquakes, is evidently more difficult. It may strongly benefit by the combined help of structural geology and regional tectonics, which may allow constraining a long term geologic and kinematic context for the present deformation field and associated seismicity.

I will review some key-cases, sometimes controversial, mainly concerning the active tectonics and the seismogenesis of 1) the Apennine extensional belt and 2) the Apennine-Maghrebian thrust front.

1) The close integration of geological and seismological data has allowed to identify a seismogenic extensional fault system which extends for about 500 km across northern Italy from Lunigiana, in northern Tuscany, to Val di Sangro in southern Abruzzo (e.g. Lavecchia et al., 1994; Cello et al., 1997; Barchi et al., 1999; Barchi et al., 2000; Argnani et al., 2003; Boncio et al., 2004; DISS3.0 and references therein). The fault system, which mainly consists of SW-dipping normal faults developed within the hanging wall of an east-dipping low-angle detachment system, has been named Etrurian Fault System (EFS, Boncio et al., 2000, 2004; Brozzetti et al., 2007). Our approach to the identification and parameterisation of the associated 3D individual seismogenic sources has been structural-seismotectonic: we have integrated surface geology data (trace of active faults, i.e. 2D features) with seismicity and subsurface geological–geophysical data (3D approach). A fundamental step has been to fix constraints on the thickness of the seismogenic layer and deep geometry of faults. When available, we have used constraints from the depth distribution of aftershock zones and background seismicity (e.g. Deschamps et al., 1984; Haessler et al., 1988; Amato et al., 1998; Barchi et al., 1999; Cattaneo et al., 2000; Boncio & Lavecchia, 2000; De Luca et al., 2000; Collettini & Barchi, 2002; Argnani et al., 2003; Piccinini et al., 2003; Boncio et al., 2004; Tinari D.P., 2007), but we have also considered information on the structural style of the extensional deformation at crustal scale (mainly from seismic reflection data), as well as on the strength and behaviour (brittle versus plastic) of the crust by rheological profiling.

Geological observations have allowed us to define a segmentation model consisting of major fault structures separated by first-order (kilometric scale) structural-geometric complexities considered as likely barriers to the propagation of major earthquake ruptures. We have named “seismogenic boxes” the plan projection of such major active structures and for each of them evaluated the maximum magnitude of the expected earthquake (Mmax). From northern Tuscany to southern Abruzzi we have identified and parameterized more than 30 boxes which are responsible or capable of experiencing earthquakes with M ≥ 5.5. These boxes lie within the boundary of the EFS extensional seismogenic province, whose background and moderate seismicity (M~<5.5) has also been considered from a seismic hazard assessment point of view (Pace et al., 2006).

Whereas the important seismogenic role of the EFS system and of its antithetic faults is now generally recognised in the literature, the imprinting of the Plio-Pleistocene extensional processes on the present architecture and deformation pattern of the southern Apennines is generally considered subordinate

4

and limited to regional-scale uplift processes. Based on detailed field work and on revision of the CROP04 seismic line and of the borehole stratigraphy, we have identified a regional extensional belt consisting of an east-dipping regional fault system and of a number of syn- and antithetical splays branching at increasing depth from west to east and rejuvenating in the same sense. Evidence of the east-dipping detachment can be traced down to a depth of 4-5 sec TWT (above 12 km depth) with a total displacement exceeding 3 km (Brozzetti and Salvatore, in prep.). The southern Apennine extensional fault system is therefore comparable to the northern Apennine EFS, in terms of dimensions and of contribute to extension. These results, which must be integrated with available seismological data, are evidently fundamental in order to both identifying likely individual seismogenic source and to defining the boundary of the extensional seismogenic province in southern Italy.

2) During Mio-Pliocene times, the development of intra-mountains extensional structures has been coeval and nearly co-axial to that of fold-and-thrust structures at the outer front of the compressional belt (Lavecchia et al., 1994). Although in the geological literature it is questioned if this front is still active (e.g. Monaco et al., 1996; Vannoli et al., 2004; Neri et al., 2005; Catalano et al., 2007) or not (e.g. Cinque et al., 1993; Patacca and Scandone, 2004), the use of integrated structural-seismological approach allows to demonstrate that the kinematic pair extension-compression is still active, at least in correspondence of two crust-scale portions of the outer basal thrust of Apennine-Maghrebian compressional system, located in the eastern Marche-Adriatic region and in mainland and central southern Sicily (Lavecchia et al., 2003, 2007). We have named them as Adriatic Basal Thrust (ABT) and Sicilian Basal Thrust (SBT). The ABT and the SBT are the outermost, and still active, structures of the inward-dipping, mainly in-sequence thrusts which progressively dislocate, shorten and thicken the Adriatic and Pelagian crust. A complete analysis of the historical (Working Group CPTI, 2004) and instrumental earthquakes (Monachesi et al., 2004; Castello et al., 2005; Neri et al. 2005) within rock volumes above the westward-dipping ABT and the northward-dipping SBT has allowed defining the boundaries of the overlying seismotectonic provinces. Each province has been subdivided into two sub-provinces of relatively homogeneous deformation, one shallow and associated with reverse, oblique and strike-slip shearing mainly along the upper crust segment (0-10 km) of the basal thrust, and one deep, associated with deformation of the middle-to-lower crust segment (10 to 25 km). The level of seismic activity within both provinces is moderate, but relevant and structurally-controlled, thus opening new perspectives in terms of seismogenic zoning and related seismic hazard assessment for the two large portions of the Italian territory.

In conclusion, the integration of regional tectonics with 3D structural and seismological data may strongly help to clarify the geometry and distribution of the active state of strain and to define the boundary of homogenous seismogenic provinces. In the case of regions extending partly offshore, as the Adriatic and Sicilian thrust fronts, the available seismological and geological data set is less well constrained and different possible seismotectonic models can be defined; consequently, a joint interpretation of data from different sources, together with the comparison of the same type of data from different but equivalent areas, is necessary to build a more constrained and coherent picture. We are well aware of the uncertainty that may affect the mixed use of geological and seismological data, but the multidisciplinary approach to the seismotectonics may help to minimize this uncertainty.

References Amato, A., Azzara, R., Chiarabba, C., Cimini, G. B., Cocco, M., Di Bona, M., Margheriti, L., Mazza, S.,

Mele, F., Selvaggi, G., Basili, A., Boschi, E., Courboulex, F., Deschamps, A., Gaffet, S., Bittarelli, G., Chiaraluce, L., Piccinini, D., and Ripepe, M., (1998), The 1997 Umbria-Marche, Italy, earthquake sequence; a first look at the main shocks and aftershocks. Geophys. Res. Lett. 25, 2861--2864.

Argnani, A., Barbacini, G., Bernini, M., Camurri, F., Ghielmi, M., Papani, G., Rizzini, F., Rogledi, S. and Torelli L., 2003. Gravity tectonics driven by Quaternary uplift in the Northern Apennines: insights from the La Spezia-Reggio Emilia geo-transect. Quaternay International 101-102, 13-26.

Barchi, M., Galadini, F., Lavecchia, G., Messina, P., Michetti, A.M., Peruzza, L., Pizzi, A., Tondi, E., and Vittori, E. (eds.), (2000), Sintesi delle conoscenze sulle faglie attive in Italia Centrale: parametrizzazione ai fini della carattrizzazione della pericolosità sismica. Roma, CNR-Gruppo Nazionale per la Difesa dai Terremoti, 62 pp.

Barchi, M.R., Paolacci, S., Pauselli, C., Pialli, G., Merlini, S., 1999. Geometria delle deformazioni estensionali recenti nel bacino dell’Alta Val Tiberina fra S. Giustino Umbro e Perugia: Evidenze geofisiche e considerazioni geologiche. Boll. Soc. Geol. It. 118, 617-625.

5

Boncio P. & Lavecchia G. (2000) – A geological model for the Colfiorito earthquakes (September-October 1997, Central Italy). Journ. Seismology, 4, 345-356.

Boncio, P., Brozzetti, F., and Lavecchia, G., (2000), Architecture and seismotectonics of a regional low-angle normal fault zone in central Italy. Tectonics, 19, 1038-1055.

Boncio, P., Lavecchia, G., and Pace, G., (2004), Defining a model of 3-D seismogenic sources for Seismic Hazard Assessment applications: The case of Central Apennines. Journ. Seism., 8, 407-425.

Brozzetti, F., and Salvatore, A,. (in prep.), Comparing active extension across the northern and southern Apennines. To be submitted to Terra Nova.

Brozzetti, F., Boncio, P., Lavecchia ,G., and Pace, B. (2007), Present activity and seismogenic potential of a low-angle normal fault system in central Italy (Città di Castello-Sansepolcro area): constraints from surface geology, seismic reflection data and seismicity. Submitted to Tectonophysics

Castello, B., Selvaggi, G., Chiarabba, C., Amato A., (2005), CSI Catalogo della sismicità italiana 1981-2002, version 1.0. INGV-CNT, Rome, http://www.ingv.it/CSI/.

Catalano, S., Torrisi, S., De Guidi, G., Grasso, G., Lanzafame, G., Romagnoli., G., Tortorici, G., Tortorici, L., (2007), Sistema a pieghe tardo- quaternarie nell’area di Catania: un esempio di fronte orogenico attivo. Rendiconti della Società Geologica Italiana 4, 181-183

Cattaneo, M., Augliera, P., De Luca, G., Gorini, A., Govoni, A., Marcucci, S., Michelini, A., Monachesi, G., Spallarossa, D., Trojani, L., and XGUMS, (2000), The 1997 Umbria-Marche (Italy) earthquake sequence: analysis of the data recorded by the local and temporary networks. J. Seismology 4, 401-414.

Cello, G., Mazzoli, S., Tondi, E. and Turco, E., (1997), Active tectonics in the central Apennines and possible implications for seismic hazard analysis in peninsular Italy. Tectonophysics 272, 43-68.

Chiaraluce, L, Barchi, M, Collettini, C, Mirabella, F, (2005), Connecting seismically active normal faults with Quaternary geological structures in a complex extensional environment: The Colfiorito 1997 case history (northern Apennines, Italy). Tectonics, 94, 99-116

Cinque, A., Patacca, E., Scandone, P., Tozzi, M., (1993), Quaternary kinematic evolution of the Southern Apenines. Relationships between surface geological features and deep lithospheric structures, Ann. Geofis. 36 829, 249-260

Collettini, C., Barchi, M., 2002. A low-angle normal fault in the Umbria region (Central Italy): a mechanical model from the related microseismicity. Tectonophysics 359, 97-115.

De Luca, G., Scarpa, R., Filippi, L., Gorini, A., Marcucci, S., Marsan, P., Milana, G., and Zambonelli, E., (2000), A detailed analysis of two seismic sequences in Abruzzo, Central Apennines, Italy. J. Seismology 4, 1-21.

Deschamps, A., Iannaccone, G., and Scarpa, R., (1984), The Umbrian earthquake (Italy) of 19 September 1979. Annales Geophysicae 2, 29-36.

Diss 3.0- Basili, R., G. Valensise, P. Vannoli, P. Burrato, U. Fracassi, S. Mariano, and M. M. Tiberti (2007). The Database of Individual Seismogenic Sources (DISS), version 3: summarizing 20 years of research on Italy’s earthquake geology. Tectonophysics, in press.

Haessler, H., Gaulon, R., Rivera, L., Console, R., Frogneux, M., Gasparini, G., Martel, L., Patau, G., Siciliano, M., and Cisternas, A., (1988), The Perugia (Italy) earthquake of 29, April 1984: a microearthquake survey. Bull. Seism. Soc. Am. 78, 1948-1964.

Lavecchia, G., Boncio, P. and Creati, N. (2003), A lithospheric-scale seismogenic thrust in Central Italy. Journ. Geodynamics, 36, 79-94.

Lavecchia, G., Brozzetti, F., Barchi, M., Keller, J.V.A. and Menichetti, M., (1994) - Seismotectonic zoning in east-central Italy deduced from an analysis of the Neogene to Present deformations and related stress fields. Bull. Soc. Geol. Am., 106, 1107-1120

Lavecchia, G., Brozzetti, F., Barchi, M., Menichetti, M., and Keller, J.V.A, (1994), Seismotectonic zoning in east-central Italy deduced from an analysis of the Neogene to present deformations and related stress fields. Geol. Soc. Am. Bull., 106, 1107-1120.

Lavecchia, G., Ferrarini, F., de Nardis, R., Visini, F., and Barbano, M.S., (2007), Active thrusting as a possibile seismogenic source in Sicily (Southern Italy): some insights from integrated structural-kinematic and seismological data. Tectonophysics, in press

Monachesi, G., Frapiccini, M., Troiani, L., Melonaro, P., (2004), RSM On Line, il DB della Rete Simometrica Marchigiana. Regione Marche-Servizio Protezione Civile-Istituto Nazionale di Geofisica e Vulcanologia. Ancona. http://protezionecivile.regione.marche.it/dbrsm

Monaco, C., Mazzoli, S., Tortorici, L., (1996), Active thrust tectonics in western Sicily (Southern Italy): the 1968 Belice earthquake sequence. Terra Nova 8, 372-381.

6

Neri, G., Barberi, G., Oliva, G., Orecchio, B., (2005). Spatial variation of seismogenic stress orientations in Sicily, South Italy. Physics of the Earth and Planetary Interiors 148, 175-191

Pace, B., Peruzza, L., Lavecchia, G., and Boncio P., (2006), Layered seismogenic source model and probabilistic seismic-hazard analyses in central Italy. Bull. Seism. Soc. Am., 96, 107-132.

Patacca, E., Scandone, P., (2004), The Plio-Pleistocene thrust belt-foredeep system in the Southern Apennines and Sicily (Southern Apenninic Arc, Italy ). In Crescenti U., D'offizi S., Merlini S. & Sacchi R. Eds., Geology of Italy, Spec. Vol. Ital. Geol. Soc. IGC 32 Florence 2004, 93-129

Piccinini, D., Cattaneo, M., Chiarabba, C., Chiaraluce, L., De Martin, M., Di Bona, M., Moretti, M., Selvaggi, G., Augliera, P., Spallarossa, D., Ferretti, G., Michelini, A., Govoni, A., Di Bartolomeo, P., Romanelli, M., Fabbri, J., 2003. A microseismic study in a low seismicity area of Italy: the Città di Castello 2000-2001 experiment. Annals of Geophysics 46 (6), 1315-1324.

Tinari, D.P., (2007). Analisi simotettonica in lunigiana-garfagnana (appennino settentrionale). analisi della sismicita’ strumentale e confronto con geometrie superficiali e profonde delle strutture tettoniche. PhD Thesis, Università degli Studi di Chieti (Italy).

Vannoli, P., Basili, R., Valensise, G., (2004), New geomorphic evidence for anticlinal growth driven by blind-thrust faulting along the northern Marche coastal belt (central Italy). Journal of Seismology 8, 297-312

Working Group CPTI, 2004. Catalogo Parametrico dei Terremoti Italiani, versione 2004 (CPTI04), INGV Bologna, available on-line at: http://emidius.mi.ingv.it/CPTI.

7

Surface fractures associated with the 1997 Umbria-Marche earthquake sequence: prospects from a decade of surveys,

analyses, and debates

Francesca R. Cinti Istituto Nazionale di Geofisica e Vulcanologia

The Umbria-Marche sequence represents the first case in Italy for which the surface effects of moderate seismic events were systematically investigated and documented. Researchers from different Institutions were involved on field reconnaissance within the wide struck area and mapped the fractures at the ground associated to the major events [e.g. Basili et al., 1998; Cello et al., 1998; Cinti et al., 1999]. The coseismic features were mainly expressed by prevailing NW-SE trending, discontinuous, linear open fractures, as well as fracture swarms, locally with centimetric vertical displacements, affecting both rocks and loose deposits, buildings, paved and unpaved roads (Figure 1). Several ruptures in the epicentral areas of the 26th September mainshocks occurred near and along the high angle normal faults mapped prior the seismic crisis [e.g. Calamita and Pizzi, 1992; Cello et al., 1997] and recognized to have played an important role in the deepening of the Colfiorito and Costa-Cesi basins. Instead, no clear geomorphic evidence of long-term cumulative tectonic displacement were found along the set of prominent ruptures associated to the 14th October mainshock in the area of Sellano village.

Because of the earthquakes size, the deformation was faint making the detection more difficult and arising debates on the analysis and on its seismotectonic interpretation (surface projection of the seismogenic fault, triggered slip on secondary extensional faults, earthquake-induced sliding of debris). Different models of connection between surface offsets and displacements at depth were proposed integrating the geological observations with other geophysical data [e.g. Boncio and Lavecchia, 2000; Salvi et al., 2000; Hernandez et al., 2004; Chiaraluce et al., 2005]. However, during the past ten years a general consensus was reached by most of the geologists concerning the information that these ruptures (either direct or indirect expression of the deep dislocation) provide on the seismogenic structure at depth.

In this light, the surface investigation of the 1997 earthquakes marks a positive step for investigating the past activity of the seismogenic structure below the Colfiorito and Sellano areas. Paleosismological studies along the 1997 surface features (Pantosti et al., 1999) offer promise for determining the frequency and size of earthquakes originated along the deep structure of the region: both estimates are critical for seismic hazard assessment. In this perspective, up to now different hypotheses have been developed on the major statement: is the 1997 earthquake the expected maximum earthquake size for this sector of the Umbria-Marche? (e.g. Galli and Galadini, 1999; Messina et al., 2002). Moreover, the 1997 case can be extended in other zones characterized by moderate historical earthquakes for which the present seismicity does not trace any earthquake fault.

Finally, we learned the correct way to get prepared for field survey during seismic emergency, particularly for sequence that lasts so long, and that once again is critical to attend "Learning from the Past, Planning for the Future”.

References Basili, R, Bosi, C., Bosi, V., Galadini, F., Galli, P., Meghraoui, M., Messina, P., Moro, M. and Sposato, A.,

(1998). The Colfiorito earthquake sequence of September-October 1997. Surface breaks and seismotectonic implications for the central Apennines (Italy). J. of Earthquake Engineering, 102(2), pp. 291-302.

Boncio, P. and Lavecchia, G., (2000). A geological model for the Colfiorito earthquakes (September-October 1997, central Italy). J. of Seismology, 4, 345-356.

Calamita, F. and Pizzi, A., (1992). Tettonica quaternaria nella dorsale appenninica umbro-marchigiana e bacini intrappenninici associati. Studi Geologici Camerti. Spec. Vol. 92/1, 17-25.

Cello, G., Mazzoli, S., Tondi, E. and Turco, R., (1997). Active tectonics in the central Apennines and possibile implications for seismic hazard analysis in peninsula Italy. Tectonophysics 272,43-68.

8

Cello, G., Deiana, G., Mangano, P., Mazzoli, S., Tondi, E., Ferreli, L., Maschio, L., Michetti, A. M., Serva, L. and Vittori, E., (1998). Evidence for surface faulting during the September 26, 1997, Colfiorito (Central Italy) earthquakes. J. of Earthquake Engineering, 102(2), 303-324.

Chiaraluce, L., Barchi, M., Collettini, C., Mirabella, F. and Pucci, S., (2005). Connecting seismically active normal faults with quaternary geological structures in a complex extensional environment: The Colfiorito 1997 case history (northern Apennines, Italy). Tectonics, 24, doi:10.1029/2004TC001627.

Cinti, F.R., Cucci, L., Marra, F. and Montone, P., (1999). The 1997 Umbria-Marche (Italy) earthquake sequence: relationship between ground deformation and seismogenic structure. Geophys. Res. Lett. 26(7), pp. 895-898.

Galli, P. and Galadini, F., (1999). Seismotectonic framework of the 1997-98 Umbria-Marche (Central Italy) earthquakes, Seism. Res. Lett., 70(4), pp. 404-414.

Hernandez, B., Cocco, M., Cotton, F., Stramondo, S., Scotti, O., Courboulex, F. and Campillo, M., (2004). Rupture history of the 1997 Umbria-Marche (central Italy) mainshocks from the inversion of GPS, DInSAR and near field strong motion data. Ann. Geophys., 47, 4, 1355-1376.

Messina, P., Galadini, F., Galli, P. and Sposato, A., (2002). Quaternary basin evolution and present tectonic regime in the area of the 1997-1998 Umbria-Marche seismic sequence (central Italy). Geomorph., 42, 97-116.

Pantosti, D, De Martini P. M., Galli, P., Galadini, F., Messina, P., Moro, M. and Sposato A., (2000). Studi paleosismologici attraverso la rottura superficiale prodotta dal terremoto del 14 ottobre 1997 (Umbria-Marche). GNGTS – Atti del 18° Convegno Nazionale, Roma 1999.

Salvi, S., Stramondo, S., Cocco, M., Tesauro, M., Hunstad, I. and other nine authors, (2000). Modeling of coseismic displacements resulting from SAR interferometry and GPS measurements during the 1997 Umbria-Marche seismic sequenze. J. Seism., 4, 479-499.

9

Figure 1. Sketch including instrumental earthquakes of the 1997 seismic sequence, pre-existing faults, and azimuthal diagrams of the ground ruptures observed around Colfiorito, Costa-Cesi, Rasenna-Mevale and Sellano [modified from Cinti et al., 2000]. Photos by Cinti et. al, 2000 and courtesy of Emanuele Tondi.

10

Reflection and earthquake seismology: a first step towards integration in the Colfiorito area.

Alfredo Mazzotti Università di Pisa

Earthquake and reflection seismology are different branches of the same science that due to many reasons have undergone separate evolutions. Reflection seismology has been mostly developed and applied by large geophysical contractors and oil companies and the related research activities have been carried out by research consortia involving both academia and industry. Earthquake seismology, the initial founder science, has been dealt with by government and research institutions and by separate research groups in the universities. While reflection seismology takes into account backscattered wave phases, earthquake seismology mainly deals with transmitted pulses. This has led to different acquisition, different processing and in general, to a separation between the two scientific communities.

There is however solid evidence that the integration of the two disciplines may be of mutual benefit. The Colfiorito case history shows how the depth images produced by means of an industry-standard reflection survey help in the understanding of the subsurface structures and fault geometry, bringing further constraints and information to the earthquake seismologists [Ciaccio et al. 2005, Stucchi et. al. 2006]. The other way round, from earthquake to reflection seismology, it is also true. Think about the use of surface waves, a few years ago exclusive domain of earthquake seismologists, that nowadays are thoroughly used by exploration geophysicists.

Different degrees of integration are indeed possible due to the common nature of propagation of seismic waves. The Colfiorito and some other examples (Yokokura, 1999, Wang, 2002, Hunter et al. 2007), that I briefly discuss from the exploration seismologists point of view, represent the first level of possible integration. That of using reflection seismic images to get the geometry of the buried geological and possibly identify the seismogenic faults and estimate the time of the events. However, as the Colfiorito case shows, the achievement of such a goal may be not easy. In fact geophysicists can be confronted by old reflection data, acquired for rather different exploration purposes, with a generally low S/N ratio. Thus, an accurate assessment of the noise problems contaminating the data and the application of up to date processing sequences are mandatory. The following two figures [Stucchi et al. 2002] clearly show the significant increase in the quality of the final reflection seismic image obtained through a reprocessing tailored to overcome the many problems and noises affecting the original data.

In case new seismic reflection data should be acquired for seismological purposes, specific field planning studies should precede the data acquisition. Processing of the reflection data, particularly when the objective is the production of well focused depth images, is again a crucial step.

Although much remains to be done even at this basic level of integration, especially in the sense of a wider application, I sustain that other subjects would allow deeper degrees of integration. One of such fields is the estimation of the (P-wave) velocity field, a fundamental information for either disciplines yet only approximately known. As an exercise we have tried, with colleagues from INGV, to just compare the velocity field derived from reflection tomography along a limited segment of a profile with the velocity field derived from earthquake tomography in the Colfiorito area [Amato et al. 1998, Chiarabba and Amato, 2003]. Since earthquake tomography produces a tridimensional velocity field, it has been necessary to project the 3D data onto the 2D reflection seismic profile. The results are shown in the two figures here below.

Although some general characteristics are shared by the two velocity fields, differences occur as high as 500 m/s. Why is it so? Indeed, we may find a number of reasons in the acquisition, in the processing and even in the way we extract and interpolate the data. The point, however, is that it should not be so. Thus there is room for improvement. A way to improve is by truly integrating the two tomographic approaches, that is to develop a joint earthquake-reflection tomography.

11

Figure 1. Line C before and after reprocessing.

Another field of possible integration is the bandwidth integration. Reflection seismic (at least the conventional one) is band-limited from about 8 Hz to few hundreds Hz at most. Earthquake seismic ranges from fractions of Hz to a few Hz. Developing methods of spectral composition of the two data (acquired in corresponding locations) would be of interest.

Finally, a convergence between the two disciplines will also be brought by ongoing technological developments. Earthquake seismology is progressively adopting denser spatial sampling, while reflection seismology is adopting multicomponent sensors. The greater amount of data will make it necessary quality control and data handling procedures that nowadays are commonly used in reflection seismology. On the other end, reflection seismology is moving towards continuous recording and time lapse acquisition that with the advent of digital MEMS type sensors will bring to the recording of either active and passive seismic signals.

12

Figure 2. Comparison of velocity models from reflection and earthquake tomographies.

13

References Amato, A., Azzara, R., Chiarabba, C., Cimini, G. B., Cocco, M., Di Bona, M.L., Margheriti, S., Mazza, F.,

Mele, G., Selvaggi, et al., (1998). The 1997 Umbria-Marche, Italy, earthquake sequence: A first look at the main shocks and aftershocks. Geophysical Research Letters, 25, pp. 2861–2864.

Chiarabba, C., and Amato, A., (2003). Vp and Vp/Vs images in the Mw 6.0 Colfiorito fault region (central Italy): A contribution to the understanding of seismotectonic and seismogenic processes. Journal of Geophysical Research, 108, B5.

Ciaccio, M.G., Barchi, M.R., Chiarabba, C., Mirabella, F., Stucchi, E., (2005) Seismological, geological and geophysical constraints for the Gualdo Tadino fault, Umbria–Marche Apennines (central Italy. Tectonophysics, 406, pp. 233–247.

Hunter, J.A., Motazedian, D., Pugin, A., Kahesci-Banab, K., (2007). Near surface geophysical techniques for mapping soft soil earthquake shaking response within the City of Ottawa, Canada. Society of Exploration Geophysiscists, Exp. Abs. of the 2007 Annual Meeting, San Antonio, pp. 1118-1122.

Stucchi, E., Mirabella, F., Barchi, M., Merlini,S., Zanzi, L., Mazzotti, A., (2002) Characterisation of seismogenic faults by means of reflection seismic surveys. A case study in the Colfiorito area: 1997 Umbria-Marche earthquakes. European Association of Geoscientists and Engineers, Exp. Abs. of the 64th Conference & Exhibition — Florence, Italy, 27 - 30 May 2002, pp. E48.

Stucchi, E., Mirabella, F., and Ciaccio, M. G., (2006). Comparison between reprocessed seismic profiles: Seismologic and geologic data — A case study of the Colfiorito earthquake area. Geophysics, 71, 2, pp. B29–B40.

Wang, C.Y., (2002). Detection of a recent earthquake fault by the shallow reflection seismic method. Geophysics, 67, 5, pp. . 1465–1473.

Yokokura, T., (1999). Seismic investigation of an active fault off Kobe: Another disaster in the making?. The Leading Edge, 18, pp.1417–1421.

14

Seismic reflection data in the Umbria-Marche active area: limits and capabilities to unravel the subsurface structure

Francesco Mirabella Università di Perugia

Seismic data in tectonically active areas provide a constrain to draw the subsurface setting of the geological structures. The detail of the information mainly depends on i) the quality of the seismic data and ii) the geological complexity of the subsurface. The seismic data quality mainly depends on the acquisition tecniques, i.e. energy and density grid. The geological environment and the deformation history controls the structural complexity of a region.

In the Umbria-Marche Apennines (central Italy) of set seismic sections and some deep wells were acquired by Eni in the '80 for oil purposes. These data have been mede available in the last years for scientific studies and provide subsurface information down to depths of about 4 s (twt), corresponding to about 10-12 km of depth.

The same area is characterised by widespread seismicity [Chiarabba et al., 2005] occurring with both moderate magnitude events and microseismicity [Amato et al., 1998; Boschi et al., 2000; Deschamps et al., 1984; Haessler et al., 1998; Chiaraluce et al., 2007]. Most seismicity and moderate events occur on normal faulting in the upper 12 km of the crust, at similar depths as those investigated by the seismic sections. Hence in this area the seismic sections can be used as an indipendent data to be compared with seismicity data.

The seismic data are not uniformely distributed. A quite dense grid is available in the area between the Tiber Valley and the Gubbio Valley (Gubbio data set). These data were acquired mainly by explosives and are of good quality. In the area of Colfiorito (Colfiorito data set), three sections of lower quality are available, being acquired by Vibroseis tecniques. Besides, the Gubbio data set crosses mainly turbiditic rocks at the surface which allow a good penetration of the seismic signal, whilst in the Colfiorito area, the outcropping rocks mainly consist of highly deformed carbonatic rocks which severely hamper the signal penetration at depth.

In the area, also three further regional Vibroseis sections are available which reach the Adriatic sea to the East and the recently acquired deep crustal NVR Crop-03 section which crosses the Italian peninsula from West to East [Pialli et al., 1998]. The industrial seismic data were acquired for oil targets and with the above described technologies, affecting the resolution of the data.

Still with these limitations, seismic data, integrated with other data sources like surface geology and seismological data can provide hints on: active faults geometry (attitude and dips), the relationships between synthetic and antithetic faults, the reconstruction of fault surfaces, the lithologies involved in the seismic ruptures and on the fault growth history.

The data-set of Gubbio crosses the Gubbio normal fault, a 22 km long Quaternary normal faul dipping towards SW which many authors retain activated during the 1984 Gubbio earthquake (Ms=5.3). Previous interpretations of the subsurface setting of the area, mainly based on the subsurface extrapolation of the surface geology [Minelli and Menichetti, 1990] suggested that the Gubbio fault was a master fault to which the Tiber Valley normal fault (AltoTiberina fault - AtF) was antithetic. Later on, when the seismic data in the area were made available and after the interpretation of the Crop-03 NVR section, it was recognised that in fact the AtF is a crustal scale low-angle detachment extending for more than 55 km, to which the Gubbio fault is antithetic [Barchi et al., 1998; Boncio et al., 2000]. More recently, the intepretation of the whole set of seismic sections in the Gubbio area allowed to draw the fault offset distribution (about 3000 m maximum) accumulated during a complex polyphase tectonic history and to describe in detail its geometry [Mirabella et al., 2004] as listric.

The Colfiorito data-set is of a lower quality and the original data were reprocessed [Stucchi et al., 2006] in order to improve the imaging of both very shallow and deep reflectors. After the re-processing, the nature of the rocks exposed at the surface along the trace influences the quality of the seismic profiles. The data do not show a clear image of the subsurface attitude of the Quaternary normal faults cropping out

15

at the surface probably due to the low offsets of these faults (of the order of 600m Mirabella et al., 2005). Nevertheless, the reprocessed sections show a good image of the basement, mainly constituted of phyllites of Upper Paleozoic age which is a low-velocity-layer (LVL) within the upper crust stratigraphy. The recognition of the depth of this reflector is of primary importance as it provides a calibration to extrapolate the surface geology at depth and to draw cross-sections. In the Colfiorito area, the reprocessed seismic data suggest a depth of the basement of about 10-12 km [Stucchi et al., 2006], different from previous works which envisaged a basement depth ranging from 5-6 km [Calamita et al, 2000] to 6-8 km [Boncio and Lavecchia, 2000; Boncio et al., 2000] to about 14 km [Bally et al., 1986].

Comparing the subsurface geology with datailed earthquakes locations [Chiaraluce et al., 2003] a lithological control of the seismogenic layer thickness of the upper crust operated by this LVL is suggested [Mirabella et al., 2007]. Besides, the mainshocks locations suggest that they occur within the evaporites, a 2000 km thick succession of dolomites and anhydrites which has traditionally been considered the “weak” horizon of the Umbria-Marche stratigraphy. More studies would be needed in order to characterise the mechanical (and possibly seismogenic) behavior of such relatively strong sedimentary rocks (e.g. platform carbonates, anhydrites and dolomites).

Seismic data are a powerfull and indipendent tool to investigate the subsurface also in active regions. The use of such data should be sponsored by both research institutes and academy in the effort of a better understanding of the earthquake geology.

References Amato, A., Azzara, R., Chiarabba, C., Cimini, G., Cocco, M., Di Bona, M., Margheriti, L., Mazza, S.,

Mele, F., Selvaggi, G., Basili, A., Boschi, E., Courboulex, F., Deschamps, A., Gaffet, S., Bittarelli, G., Chiaraluce, L., Piccinini, G. and Ripepe, M. (1998). The 1997 Umbria-Marche, Italy earthquake sequence: a first look at the main shocks and aftershocks. Geophysical Research Letters, 25:2861-2864.

Barchi, M., De Feyter, A., Magnani, M., Minelli, G., Pialli, G. and Sotera, B. (1998). Extensional tectonics in the Northern Apennines (Italy): evidence from the CROP03 deep seismic reflection line. della Societá Geologica Italiana, 52:52-538.

Boschi, E., Guidoboni, E., Ferrari, G., Mariotti, D., Valensise, G. and Gasperini, P. (2000).Catalogue of strong italian earthquakes from 461 b.c. to 1997. di Geofisica,43:609- 68.

Chiaraluce, L., Ellsworth, W., Chiarabba, C. and Cocco, M. (2003). Imaging the complexity of an active complex normal fault system: the 1997 Colfiorito (central Italy) case study. Journal of Geophysical Research, 108 (B6):2294, doi:10.1029/2002JB002166.

Ciaccio, M., Barchi, M., Chiarabba, C., Mirabella, F. and Stucchi, E. (2005). Seismological, geological and geophysical constraints for the Gualdo Tadino fault, Umbria-Marche Apennines (Central taly). Tectonophysics, 406(3/4):233-247.

Haessler, H., Gaulon, R., Rivera, L., Console, R., Frogneux, M., Gasparini, G., Martel, L., Patau, G., Siciliano, M. and Cisternas, A. (1988). The Perugia (Italy) earthquake of 29 april 1984: a microearthquake survey. Bulletin of the Seismological Society of America, 78(6):1948-1964.

Mirabella, F., Boccali, V. and Barchi, M. (2005). Segmentation and interaction of normal faults within the Colfiorito fault system (Central Italy). In D. Gapais, J. Brun and P. Cobbold, editors. Deformation mechanisms, rheology and tectonics: from minerals to the lithosphere. Geological Society Special Publication 243, London.

Mirabella, F. and Pucci, S. (2002). Integration of geological and geophysical data along a section crossing the region of the 1997-98 Umbria-Marche earthquake (Italy). Bollettino della Societá Geologica Italiana, Vol. Spec. 1:891-900.

Pialli, G., Barchi, M. and Minelli, G. (1998). Results of the crop03 deep seismic reflection profile. Memorie della Societá Geologica Italiana, 52:654 pp.

16

Fault zone characterization as a challenge for em prospecting methods in the last decade:

the Colfiorito case-history.

Agata Siniscalchi Università di Bari

Summary Over the last decade electromagnetic (EM) measurements have provided new constraints on the

upper-crustal structure of the major fault zones in the world, both when they act as conduit and as a barrier, due to strong sensitivity of resistivity to fluids circulation and mineralization.

On the track of an high impact magnetotelluric (MT) study performed across the San Andreas Fault, high resolution EM data were collected in the Colfiorito epicentral area along profiles crossing some main fault lineaments. Being the study focussed both on shallow that on intermediate resistivity distribution in the brittle upper-crust, a MT profile was integrated by several electrical resistivity tomographies (ERT). These last were successful in locating faults even where the structures are buried by a wide covering of Quaternary deposits and in the recognition of different electrical signatures of the faults.

MT resistivity model crossing Mt. Prefoglio normal Fault clearly imaged the typical thrust structures of the area and a high conductive zone spatially related to the fault. Seismicity seems to be located outside such conductive area. The behaviour of such conductive zone suggests a fluidised and altered zone incapable of supporting significant stress internally.

Indeed, very recent development in this research field in the world obtained a significant improvement both for a physical than for a lithologic model of fault areas by quantitative joint inversion and interpretation of seismic and electrical parameters deriving by collocated joint experiments. Some examples of these approach will be discussed.

1. Introduction The significant role of fluids in earthquake rupture processes is now largely accepted. Since the

presence of an electrically conducting fluid, such as saline pore water, strongly influences the overall conductivity of the crustal rocks, the application of electrical and electromagnetic methods in fault zone studies is increasing notably. These methods can be divided into those that require the generation of artificial sources, like Direct Current (DC) methods, and those using natural sources, like magnetotellurics (MT). Detection of fault area anomalies is experienced even with the usually adopted large separations between MT soundings (generally some kilometres), but the true extent of the fault area and the correct resistivity contrast with the host rocks can be mis-estimated. This is the challenge for em methods in order to contribute to fault characterization. The sensitivity of surface geophysical surveys decreases with increasing depth. These factors are especially true for MT, since imaging to greater depths requires low frequencies with long wavelengths (skin depths); resistive rocks permit higher resolution in depth. [Eberhart-Phillips et al., 1995], on the basis of synthetic study too, pointed out that proper resistivity models can be obtained by a careful experimental design with MT data being recorded with dense site spacing and over a broad frequency range.

Soon after [Unsworth et al., 1997] performed high resolution MT profiles across two segments of the San Andreas Fault which exhibits very different patterns of seismicity: at Parkfield and in Carrizo Plain. In both surveys, electric fields were sampled continuously, with 100 m long dipoles laid end-to-end across the fault. Narrow zones of low resistivity in the core of the fault was clearly detected. That study demonstrated the feasibility of MT to give significant contributions in the fault zones characterization, expecially for identification of fluid trapped in the damage zones; different amount of fluids and structural differences imaged between the two segments were inferred. Further studies across the San Andreas Fault had shown that seismicity was restricted to regions outside the fault zone conductor (FZC), which were

17

interpreted to reflect fluidised and altered zones incapable of supporting significant stress internally [Bedrosian et al., 2004].

During the last decade, several study all over the world were performed on this topic, as tested by the increasing number of review papers which tryed to individuate common patterns of the resistivity responses over faults of the same type or similar tectonic setting or age [Unsworth and Bedrosian, 2004; Wannamaker et al., 2004; Ritter et al., 2005].

2. Colfiorito em study In the framework of a GNDT 2000-2002 project, high resolution EM data were collected in the

epicentral area of the earthquake sequence, which struck the Umbria-Marche region during 1997-’98, along profiles crossing some main fault lineaments. Being the study focussed both on shallow that on intermediate resistivity distribution in the brittle upper-crust, an MT profile was integrated by several dipole-dipole electrical resistivity tomographies (ERT), one of which along the same profile of the MT survey. These last measurements allows to obtain an accurate shallow resistivity imaging and can solve the inherent trade off in MT between shallow structure and the deeper response of the fault zone: shallow structure must be accurately determined if deeper structure is to be reliably imaged. In figure 1 is shown the location map of the measurements herein discussed.

Figure 1. Survey location map: MS profiles in blu; MT station sites in cyan (redrawn after [Chiaraluce et al., 2005].)

2.1 Data analysis Multicore cables and an automatic acquisition system was adopted for ERT measurements and the

dipole-dipole array was chosen as it enhances the horizontal resolution. A resolution test on synthetic models suggested to use the combined inversion of a dataset consisting of a large scale tomography and a target-dependent multi-resolution measurements, the multi-scale approach, in order to allow an increasing in the target (fault) resolution also at depth. This is made possible by the greater number of receivers that sense the resistivity changes in a fixed cell, thus better constraining the inversion process.

The MT profile was performed with a 100 m station spacing, where possible, and MT transfer function estimated in the frequencies ranging from 0.01 to 250 Hz. Since man-made electromagnetic noise sources, present in the largest part of the Italian areas, could bias the MT tensor estimate, a remote reference station was always used, while each station consisted of a number of channels ranging from 6 to 10. Such measuring configuration allowed to perform and compare apparent resistivity estimates obtained in different processing schemes: robust single station (SS), remote reference (RR) and multiple reference robust multivariate-errors-in variable (RMEV) [Egbert, 1997], observing significant differences among them in the highest frequency range (5-250 Hz). In particular, the RMEV method allowed to strongly

18

reduce the bias respect to the SS and RR methods. Moreover we adopted and compared different techniques in order to eliminate the power-line 50Hz noise and its harmonics, which influenced every geophysical field measurement.

2.2 Results The combined inversion of the multi-scale surveys show to yield detailed structural information and

to infer some characteristics of the fault zone. Both l1 and l2 norm were tested for the inversion of the collected data sets. Due to the large resistivity contrasts characterizing the field and log data, the robust inversion was finally used being best suited to represent sharp discontinuities. ERTs models were supported by detailed geological surface data: analysis of some resistivity log data allowed a quantitative correlation between in situ resistivity distribution and geological units. By such procedure the electrical signatures of the crossed faults were differentiated. Minor faults show sharp lateral variations generated by a contact between different geological formations, while the major faults are also characterized by a decrease in resistivity values distributed on the fault damage zones [Diaferia et al., 2006]. The narrow core of a fault is instead beyond the resolution power of the electrical survey. In figure 2 the resistivity section for MS2 line (Fig. 1), obtained using the multi-scale technique, was compared with the corresponding geological section, constructed using the available detailed geological maps [Regione Umbria, 2002].

Figure 2. MS2 line. a) Geological cross section; b) recovered resistivity model with geological interpretation (rms=9.2). D: Debris; QD: Quaternary deposits; SRB: Scaglia Rossa/Bianca; MF: Marne a Fucoidi; MA: Maiolica; FDZ: Fault Damage Zone. Redrawn after [Diaferia et al., 2006].

MT data were founded consistent to a 2D resistivity modelling. Both TE and TM mode data were then included in the inversion process which accounted for topography too. The obtained MT resistivity section (profile in fig.1), crossing M. Prefoglio Fault is shown in figure 3. Good agreement was found between MS3 and MT shallow resistivity sections in the common investigation depth.

MT model images some conductive zones in its shallow part that seems well ascribable to the typical thrust structure of this area and clearly put in evidence in correspondence of the surface fault trace a conductive zone almost sub-vertical extending downward to the first km of depth. As experienced in the analyzed resistivity logs, within the calcareous units, fractures filled with interconnected fluids strongly enhances the electrical conductivity. Seismicity (aftershocks after [Chiaraluce et al., 2005]) seems to be located outside such conductive area. As argued by [Wannamaker et al., 2004], the conductive fault zones facilitate a concentration of movement, with seismic events potentially spawned at asperities in the broader fault zone.

19

Log

resi

stiv

ity (o

hm m

)

Mt. Prefoglio

0 1000 2000 3000 4000 5000 6000

distance (m)

-5000

-4000

-3000

-2000

-1000

0

1000

elev

atio

n (m

)

Mt. OrveColfiorito Plain

11.522.533.544.555.566.577.588.5

30000

10000

300

100

30

10

Figure 3. MT resistivity model obtained by inversion of both TE and TM mode; total rms=3.8. Aftershocks (black dots) are redrawn after [Chiaraluce et al., 2005]. Red arrow marks the fault.

3. Discussion on recent developments The shown examples illustrated the utility of the em methods in imaging fault zone structure within

the brittle crust. High resolution images are obtained if data are gathered with a dense site spacing, ideally with detailed areal coverage of the target region. The conductivity structures associated with fault zones in the brittle crust reflect a complex interplay between lithology, fluid supply, permeability, and fault geometry. This combination can best be untangled for an individual fault zone using additional information from structural geology and seismology.

Resistivity models can significantly contribute to the characterization of faults due to strong sensitivity of such parameter to fluids circulation and mineralization, but as outlined by many recent development in this research field a significant improvement both for a physical than for a lithologic model of fault areas is quantitative joint inversion and interpretation of seismic and electrical parameters. Seismic and MT methods are often favoured for crustal studies as they provide images of acoustic velocity and electrical resistivity, respectively, on similar scales and with comparable spatial resolution [Jones 1987]. By looking in tandem at velocity and resistivity we retain the strengths of each method, while lessening the susceptibility of our interpretation to their individual weaknesses. By analyzing data coming from a collocated joint seismic and MT experiment, [Bedrosian et al., 2007] developed a method based on the statistical correlation of physical properties in a joint parameter space. Such method is independent by theoretical or empirical relations linking electrical and seismic parameters. Regions of high correlation (classes) between resistivity and velocity can be mapped in depth section. The spatial distribution of these classes, and the boundaries between them, provide structural information not evident in the individual models. Some examples of these approach applied in the characterization of fault zones will be discussed.

4. References Bedrosian P. A., N. Maercklin, U. Weckmann, Y. Bartov, T. Ryberg and O.Ritter, (2007). Lithology-

derived structure classification from the joint interpretation of magnetotelluric and seismic models. Geophys. J. Int., 170, 737–748 doi: 10.1111/j.1365-246X.2007.03440.x.

Bedrosian P., M. Unsworth, G. Egbert and C. Thurber, (2004). Geophysical images of the creeping San Andreas Fault: Implications for the role of crustal fluids in the earthquake process, Tectonophysics, 358, doi:10.1016/j.tecto.2004.02.010.

Chiaraluce, L., Barchi, M.R., Collettini, C., Mirabella, F. and Pucci, S., (2005). Connecting seismically active normal faults with Quaternary geological structures in a complex extensional environment: The Colfiorito 1997 case history (northern Apennines, Italy), Tectonics, 24, TC1002, doi:10.1029/2004TC001627.

Diaferia I., M. Barchi, M. Loddo, D. Schiavone, and A. Siniscalchi, (2006). Detailed imaging of tectonic structures by multiscale Earth resistivity tomographies: The Colfiorito normal faults (central Italy), Geophys. Res. Lett., 33, L09305, doi:10.1029/2006GL025828.

20

Eberhart-Phillips, D., W. D. Stanley, B. D. Rodriguez, W. J. Lutter, (1995). Surface seismic and electrical methods to detect fluids related to faulting, J. Geophys. Res., 100(B7), 12919-12936, doi: 10.1029/94JB03256.

Egbert, G.D., (1997). Robust multiple-station magnetotelluric data processing. Geoph. J. Int., 130, 475-496.

Jones, A., (1987). MT and reflection: an essential combination, Geophys. J. R. Astron. Soc., 89, 7–18. Ritter, O., Hoffmann-Rothe, A., Bedrosian, P.A., Weckmann, U. and Haak, V., (2005). Electrical

conductivity images of active and fossil fault zones. In: High-Strain Zones: Structure and Physical Properties (D. Bruhn and L. Burlini, eds.), Geological Society of London Special Publications, London.

Wannamaker P.E., T. G. Caldwell, W. M. Doerner, and G. R. Jiracek, (2004). Fault zone fluids and seismicity in compressional and extensional environments inferred from electrical conductivity: the New Zealand Southern Alps and U. S. Great Basin, Earth Planets Space, 56, 1171–1176.

Unsworth M. and P. A. Bedrosian, (2004). On the geoelectric structure of major strike-slip faults and shear zones, Earth Planets Space, 56, 1177–1184.

Unsworth, M. J., Malin, P.E., Egbert, G.D., & Booker, J.R., (1997). Internal structure of the San Andreas fault at Parkfield, California, Geology, 25(4), 359-362.

21

Geochemical evidences of fluid involvement in the Northern Apennine seismogenesis

Giovanni Chiodini Istituto Nazionale di Geofisica e Vulcanologia

Among the techniques used for the search of earthquakes precursors, geochemistry has provided some high-quality signals, especially since the 1960’s, mainly as the result of instrumental developments (Toutain and Boubron 1999). Focusing interest on geochemical anomalies linked to seismo-tectonic activity is not an unexpected development, owing to the multiple evidence of a genetic link existing between fluid flows and faulting processes (Collettini et al., 2007 and references therein). Direct evidence of large fluxes of fluids into fault zones comes from oil exploration and from widespread recognition of fault-hosted vein complexes and hydrothermal alteration. Direct evidence of the effect of fluid pressure upon fault stability has come from earthquakes induced by direct injection of fluids down boreholes, as at Denver and Rangely in Colorado or in the KTB in Germany. High fluid pressure at depth play a major role in triggering earthquakes by reducing fault strength and potentially controlling the nucleation, arrest, and recurrence of earthquake rupture (Sibson, 1992, 2000). During the seismic cycle, crustal fluids can be trapped by low permeability mature fault zone seals or stratigraphic barriers. The development of fluid overpressures at the base of the fault zone during the interseismic period can help to facilitate fault slip. Once the seal is ruptured by an earthquake, permeability increases and fluids are redistributed from high to low pressure areas. The widespread development of crack-seal textures in mineral veins seems to be consistent with such behavior, as they suggest repeated cyclic build-ups in fluid pressure, hydrofracture development and fluid pressure release.

This weakening mechanism requires a continuous fluid supply near to the base of the brittle crust, a condition which has been verified in the Northern Apennine region. Regional-scale studies have shown that the western regions of Italy are affected by an intense CO2 degassing process which results at the surface in numerous CO2-rich gas emissions (INGV-DPC project V5, Fig. 1). The gas flow rates are very high; for example, the biggest gas emissions release amounts of CO2 similar to diffuse degassing from active volcanoes. Because of the relatively high CO2 solubility in water, the occurrence of gas emissions at the surface depends on the ratio of groundwater volumes circulating in the sub-surface relative to the amount of gas arriving from depth. Gas emissions at the surface are possible only when the average CO2 influx is higher than ~ 0.5 t d-1 km-2, 2 t d-1 km-2 and 5 t d-1 km-2 for the flysch, volcanic and carbonate aquifers respectively (Collettini et al., 2007). The CO2 flux threshold value is ten times higher for the carbonates cropping out in the eastern part of Central Italy, compared to the flysch and volcanic aquifers located in the western sector. Therefore, in the western sector, the rising CO2 can more easily reach the surface as a free gas phase generating the strong emissions observed at the surface. By contrast, the carbonate aquifers of the eastern sector can dissolve large amounts of deeply-derived CO2 preventing the formation of gas emissions. These findings allowed a map and an estimation of the CO2 flux to be made based on the isotopic composition of the carbon dissolved in the groundwaters (Chiodini et al. 2000; 2004).

The map shows the presence of two large regions (TRDS-Tuscan Roman Degassing Structure and CDS-Campanian Degassing Structure) where the CO2 flux is due to deeply derived CO2 (i.e. CO2 flux higher than 0.4 t d-1 km-2). A minimum of the total amount of deeply-derived CO2 released by these degassing structures was estimated to be ~20000 t d-1. Chiodini et al. (2004) have suggested that this huge amount of CO2 (~ 5-10% of the global CO2 output from volcanoes) was derived from the mantle overlying the Adriatic subduction zone which was metasomatised by crustal fluids originating from the down-going slab. This hypothesis is supported by the relatively low 3He/4He ratios of gas emissions in central Italy which fall in the same range as the He recorded from fluid inclusions trapped within olivine and pyroxene phenocrysts of the volcanites of central Italy (Martelli et al. 2004). This deep source of CO2 cause an important process of gas transfer trough the crust, where the reservoirs act as traps for the CO2 and contain as much gas as possible for their depth (Chiodini et al. 1995). The PCO2 of the deep well

22

fluids in TRDS are close to hydrostatic at shallow crustal levels (1-2 km) and approach lithostatic values at greater depths.

An intriguing coincidence among seismicity and the eastern boundary of the CO2 degassing areas is evident in Fig. 1. Earthquake epicenters concentrate in a belt that separates two distinct crustal domains: the Tyrrhenian hinterland, where both the TRDS and the CDS are located, and the Adriatic foreland where deeply derived CO2 is virtually absent. Following the information from the CROP03 profile, we hypothesize that CO2 from the mantle intrudes the deeper, now inactive, thrusts affecting the Moho and the upper crustal levels at the boundary between the Tyrrhenian hinterland and the Adriatic foreland. Here, at depths ranging from few kilometers to about 15 km b.s.l., the thrusts are dislocated by the active, low-angle normal faults, where the seismicity of Apennines is concentrated. In this zone, the arrangements of the deeper thrusts and low-angle normal faults describe outward verging structures which could act as traps in which CO2 may accumulate and generate overpressurized reservoirs. Direct measurements suggest the presence of overpressurized CO2 reservoirs at the boundary between the TRDS and the non-degassing zone: the wells of S. Donato and Pieve S. Stefano encountered a CO2 pressure of about 0.8 of the lithostatic weight. Furtehrmore the seismic crises that occurred at Colfiorito in 1997 have been interpreted as caused by pressure diffusion from deep overpressurized CO2 zones to shallow reservoirs at hydrostatic fluid pressure (Miller et al., 2004). Overpressurization in these deep structures, fed by mantle derived CO2, is in my opinion one of the primary trigger of the 1997-1998 Umbria Marche earthquakes.

Figure 1. Location of the CO2 rich gas emissions archived in GOOGAS (INGV-DPC V5 project), CO2 Earth degassing map from carbon mass balance of aquifers (Chiodini et al., 2004) and seismic activity in Central Italy.

If this overpressurizzation is one of the main causes of the Apennine earthquakes, a reasonable question concerns the fate of the fluids involved in the deep seismogenetic structures. Did the deeply derived CO2 reach the surface during the seismic events? Because I was not directly involved in the geochemical investigations performed during the 1997-1998 seismic crisis, I will refer to the results published by colleagues in international scientific journal. Geochemical investigations regarded cold groundwaters (Quattrocchi 1999; Bazzoffia et al., 2002), thermal springs (Quattrocchi 1999; Quattrocchi et al., 2000; Favara et al., 2001; Italiano et al. 2004; Caracausi et al., 2005), and gas emissions (Heinicke et al., 2000; Italiano et al., 2001; Caracausi et al., 2005). Many variations affected cold and thermal groundwaters of Umbria during the main seismic events (increase of turbidity, variation of the flow rates, variation of PH and of solute concentrations, variation of temperature, etc.). Most of these variations have been interpreted as consequence of crustal changes of permeability during the seismic crisis. Quattrocchi 1999 explicitly excludes the involvement of deep mantle derived, CO2 rich fluids and refers the observed anomalies to ‘a noteworthy enhanced fracture field” in the crust (and consequent mixing of different fluids). Heinike et al., 2000 again suggest that the variation observed in some gas vents can be interpreted as possible consequence of the crustal permeability changes induced by earthquake shaking, however they hypothesize ‘a general, probably earthquake induced, increase of the regional gas output’. The changes on

23

dissolved gas contents of few thermal waters suggested that crustal permeability modifications related to the seismogenic process appeared to be responsible for the observed anomalies and that that the seismic activity related with the beginning of the crises enhanced the release of CO2 on a regional scale (Italiano et al., 2004). Some deep mantle derived geochemical signs has been hypothesized investigating the He isotopes of three gas emission of the area systematically sampled during the 1997-1998 period (Italiano et al., 2001). On the base of these results seems that anomalies linked to the hypothesized overpressure in the deep seismogenetic structure were detected in some sites even if the interpretation of these anomalies is not easy and resulted some times controversial. This is not surprising because the Apennine is an area of large aquifers recharged by meteoric waters which tend to mask, to dilute any geochemical sign coming from depth.

As final remarks it is important to note that all the geochemical investigations here above discussed started after the beginning of the seismic crisis, i.e. they don’t constitute a suitable data set for investigating earthquakes precursors. In order to better interpret geochemical signals recorded during seismic crisis we need the development of permanent geochemical nets of sensors located in the most strategic sites (e.g. the hundreds gas emissions catalogued on the frame of the INGV-DPC V5 project). We need also robust conceptual models of the regional Earth degassing process and physical modeling of fluid transfer trough the crust.

References Bazzoffia A., Caracausi A., Innocenzi P., Italiano F., Martinelli G. 2002. Variazioni Geochimiche in acque

sotterranee fredde dell'Umbria connesse con la sequenza sismica del 1997-1998 Acque Sotterranee 78, 22-33.

Caracausi, A., Italiano, F., Martinelli, G., Paonita, A. & Rizzo, A., 2005.Long-term geochemical monitoring and extensive/compressive phenomena:case study of the Umbria Region (Italy), Ann.Geophys., 48, 43–53.

Chiodini G., F. Frondini and F. Ponziani. - 1995. Deep structures and carbon dioxide degassing in Central Italy. Geothermics, 24, pp. 81-94

Chiodini G. , Frondini F., Cardellini C., Parello F. and Peruzzi L. (2000). Rate of diffuse carbon dioxide earth degassing estimated from carbon balance of regional aquifers: the case of Central Apennine (Italy), J Geophys. Res., 105, 8423-8434

Chiodini G., Cardellini C., Amato A., Boschi E.,Caliro S., Frondini F., Ventura G. (2004) Carbon dioxide Earth degassing and seismogenesis in central and southern Italy. GRL, Vol. 31, No. 7, L07615 10.1029/2004GL019480

Collettini, C., C. Cardellini, G. Chiodini, N. De Paola, R. E. Holdsworth, and S. A. F. Smith (2007), Weakening due to fluid involvement in normal faulting: short- and long-term processes, J. Geol. Soc. (in press).

Favara, R., Italiano, F. & Martinelli, G., 2001. Earthquake induced chemical changes in the thermal waters of the Umbria region during the 1997–1998 seismic swarm, Terra Nova, 13, 227–233.

Heinicke, J., Italiano, F., Lapenna, V.,Martinelli, G. and Nuccio, P.M., 2000.Coseismic geochemical variations in some gas emissions of Umbria region (Central Italy). Phys. Chem. Earth, 25,289-293.

Italiano, F., Martinelli, G. & Nuccio, P.M., 2001. Anomalies of mantle derived helium during the 1997–1998 seismic swarm of Umbria-Marche,Italy, Geophys. Res. Lett., 28, 839–842.

Italiano, F., Martinelli, G. and Rizzo A.(2004) Geochemical evidence of seismogenic-induced anomalies in the dissolved gases of thermal waters: A case study of Umbria (Central Apennines, Italy) both during and after the 1997– 1998 seismic swarm. Geoch. Geophys. Geosyst. Q11001, doi:10.1029/2004GC000720

Martelli M., Nuccio P. M., Stuart F. M., Burgess R., Ellam R. M.,and Italiano F. (2004) Helium strontium isotopic constrains on mantle evolution beneath the Roman Comagmatic Prov., Italy. Earth Planet. Sci. Lett. 224, 295–308.

Miller, S. A., C. Collettini, L. Chiaraluce, M. Cocco, M. Barchi, and B. J. P. Kaus (2004), Aftershocks driven by a high pressure CO2 source at depth, Nature, 427, 724–727.

Quattrocchi, F., 1999, In search of evidences of deep fluid discharges and pore pressure evolution in the crust to explain the seismicity style of Umbria-Marche 1997–98 seismic sequence (Central Italy), Ann. Geof. 42(4), 609–636.

Quattrocchi, F., Pik, F., Pizzino, L., Guerra, M., Scarlato, P., Angelone, M., Barbieri, M., Conti, A., Marty, B., Sacchi, E., Zuppi, G.M., Lombardi, S. 2000. Geochemical changes at the Bagni di Triponzo

24

thermal spring during the Umbria-Marche 1997-1998 seismic sequence. J. of Seismology, 4 (4), pp. 567-587

Sibson, R. H. (1992), Fault-valve behavior and the hydrostatic-lithostatic fluid pressure interface, Earth Sci. Rev., 32, 141– 144.

Sibson, R. H. (2000), Fluid involvement in normal faulting, J. Geodyn., 29,449– 469. Toutain, J. P., and J. C. Baubron (1999), Gas geochemistry and seismotectonics: A review,

Tectonophysics, 304, 1–24.

25

The 1997-98 Umbria-Marche earthquake: “geological” vs. “seismological” faults

Massimiliano R. Barchi Università di Perugia

Introduction Before the 1997-98 earthquake, some major aspects of the regional seismotectonic scenario had been

already identified by the geological research: the extensional active stress regime affecting the region and the seismogenic role of SW-dipping normal faults were already been hypothesized by integrating the structural study of Quaternary faults with the distribution, geometry and kinematics of the instrumental and historical earthquakes [e.g. Lavecchia et al., 1994]. During the CROP03 project [Pialli et al., 1998], the interpretation of the seismic reflection profiles revealed the presence of an important, E-dipping detachment (Altotiberina Fault) in the area comprised between Sansepolcro, Gubbio and Perugia. However, these concepts were neither shared with other scientific components (seismologists and geophysicists), nor completely affirmed in the geological community.

The 1997-98 earthquake sequence, occurring in the central part of the Umbria Fault System, provided a huge amount of good quality seismological data, acquired by local networks soon after the beginning of the seismic sequence and processed with up-to date techniques. The earthquakes stimulated the acquisition of many other geological and geophysical data-sets (surface ruptures, structural analysis, geodetic data, geoelectrical and magnetic data, etc…). As a consequence, the 1997-98 sequence is possibly the best case-study in the world for extensional earthquakes, driving the possibility of building up an integrated model of an extensional system.

In particular, we are here interested in discussing the problem of the connection between the "seismological" faults (i.e. the short term, instantaneous ruptures generated during the earthquakes, representing the effect of a single event) and the "geological faults" (i.e. the long-term, active faults, exposed at the surface, that are the effects of multiple repeated events of faulting). In regions of low tectonic active extensional strain (≈0.05x10-6 yr-1), such as the Northern Apennines of Italy, where co-seismic ruptures at the surface are rarely observed and their tectonic significance is debated, the connection between geological and seismological faults is a matter of much discussion. The lack of earthquakes at depth shallower than 1 km contributes to the difficulty of integrating the two data sets.

Seismological faults In the Colfiorito and Sellano area the aftershock sequences, registered in September-October 1997,

are distributed in the depth interval 1-10 km, even if most of the seismicity is shallower than 8 km. Chiaraluce et al. [2003], assuming that aftershock distribution reflects the geometry of the major activated faults, define the 3D geometry of the activated fault system. The main peculiarities of these seismological faults can be summarised as follow:

- most faults dip towards the SW, trending about 140°N; - in most cases the trace of the seismic ruptures is almost planar, 35° to 45° dipping. - the focal mechanisms of the mainshocks show a similar dip (38° to 50°) and an almost pure dip-

slip (normal) kinematics. - faults did not rupture up to the surface, since most of the registered seismicity stops at about 1 km

b.s.l. Based on main shock and aftershock locations, focal mechanisms, geodetic data and strong motion

data, three major ruptured segments, aligned in a NW-SE-trending direction, were recognised. The length of these segments is 7-8 km.

In the Gualdo Tadino area [Ciaccio et al., 2005] the aftershocks registered in April 1998 identify a NW-SE-elongated, 8 km long ruptured fault segment, sub-parallel to the Colfiorito fault segments. The

26

aftershocks image a gently SW dipping fault (30° to 40°), confined between 3 and 8 km. In the central part of the fault, the upper part of the ruptured fault appears slightly steeper (about 37°) than the lower one (about 28°), the change of dip occurring at a depth of 5-6 km. The 30°-40° SW-dip of the plane is in agreement with the SW-dipping plane of the CMT solution of the mainshock.

Geological faults The pattern of the recent extensional structures within the epicentral area is quite complex: the

majority of the faults dip towards the SW, whilst antithetic, NE-dipping structures are relatively rare. Among these faults, the Colfiorito Fault System (CoFS) consists of a set of SW-dipping normal faults, grossly aligned in a ~N140° direction, which represent the geologically and geomorphologically prominent structures of the area. CoFS faults dips in the range 50°-80° and possess almost pure extensional kinematics. Considering their geometry, relative importance and the along-strike throw variations, the faults pertaining to the CoFS can be divided into four major segments, characterised by a comparable length (from 8 to 10 km) along their strike; from North to South: Mt. Pennino (F1), Mt. Le Scalette (F2), Cesi-S. Martino (F3) and Mt. Cavallo-Mt. Fema (F4).

Extensional tectonics is also expressed by a set of Pleistocene intramountain basins, located in the downthrown hanging-wall block of the CoFS. The still active Colfiorito swamp, in the central part of the studied area, represents a remnant of the lacustrine environment. The Lower-Middle Pleistocene fluvio-lacustrine deposits of the larger Colfiorito basin reach the thickness of at least 120 m at its depocentre.

No important Quaternary normal fault is exposed in the Gualdo Tadino area. Correspondingly, no Quaternary basin is present at the fault hangingwall, but only a lowered, almost flat area, following the western border of the Umbria-Marche ridge, where the Miocene bedrock is covered only by sparse and relatively thin (less than 15 m) outcrops of Quaternary gravels.

Connection between geological and seismological faults We hypothesise a strict connection between the geological faults and the seismological faults. In

this vision, the Quaternary tectono-sedimentary evolution and the present-day geologic and geomorphologic setting of the region is the result of repeated earthquakes, occurring on this NW-SE trending normal fault system.

In the Colfiorito area this hypothesis is supported by the following evidences [Chiaraluce et al., 2005]: 1) the epicenters map distribution strongly corresponds to the mapped Quaternary fault pattern; 2) both the lengths of the single geological and seismological faults (7-10 km) and of the whole activated system (40 km), are compatible; 3) the coseismic subdued region is located at the hanging-wall of the mapped faults and corresponds to the area characterised by the lower topography, suggesting a continuous subsidence of the hanging-wall blocks since Early Pleistocene; 4) in section view, the aftershocks alignments, that highlight the ruptured fault planes, point towards the surface into the major Quaternary faults; 5) the kinematics of the Quaternary and activated faults is very similar and the area has been affected since Early Pleistocene by a constant, NE-trending extensional stress field; 6) pre-existing N10°-trending transpressive faults acted as lateral barriers for both seismological and geological faults.

Seismic reflection data The seismic profiles, extensively used for oil exploration purposes, are a powerful tool for imaging

the subsurface structure. In the study of active faults, seismic profiles can help to assess: - the fault geometry at seismogenic depth and possibly its connection with the surface structure; - the tectono-sedimentary evolution of the faults and their long-term slip-rate; - the nature of the rocks involved in faulting, their lithology, geophysical and mechanical

parameters, through calibration by deep wells. In the Umbria region the best quality seismic reflection profiles are available in the area between

Città di Castello, Gubbio and Perugia, about 40 km NW of the Colfiorito epicentral area. The interpretation of these profiles effectively images the Gubbio normal fault [Mirabella et al., 2004] showing that:

- the fault dips 32°- 47° till about 4 km depth, and less than 20° at greater depth, where it inverts a pre-existing thrust;

27

- an antithetic E-dipping fault borders the western flank of the basin; - the maximum displacement is about 3000 m, decreasing along strike with a typical bell-shaped

profile; - the Gubbio fault is antithetic to a major, E-dipping, low angle normal fault, the AltoTiberina fault. In the epicentral area, contrasting interpretations of the subsurface setting have been proposed soon

after the earthquake. Three reprocessed seismic profiles were interpreted and the resulting structural setting was compared with the seismicity distribution, showing that: 1- the seismicity of the Umbria-Marche sequence is distributed within the Mesozoic sedimentary cover, above the top basement; 2- the mainshocks occur within the Triassic evaporites, consisting of alternated dolomites and anhydrites [Mirabella et al., 2007].

Early structural interpretations, based on 2-D geological sections, promoted the idea that the SW-dipping normal faults reactivated preexisting thrusts. Tomographic images partially supported this hypothesis. The structural setting observed at the surface and the 3-D subsurface reconstruction, based on seismic reflection data both contrast this hypothesis, showing that the arc-shaped geometry of the major basement thrusts is maintained at depth and that the thrusts are cross-cut by the 1997-98 seismicity (i.e. by the seismological faults). Therefore the faults responsible for the 1997-98 seismic sequence are not inherited from previous tectonic events.

Figure 1. An integrated model of the Umbria-Marche seismogenic faults

An integrated model for the Umbria seismogenic faults The observations summarized in the previous paragraphs point to an integrated model is here

proposed (fig. 1), which can be extended to the entire SW-dipping UFS, from Sansepolcro to Norcia. Some peculiar features of this model are here summarized.

1. Fault geometry: fault dip changes with depth: - at the surface is 60°-70°; - from ∼ 2 km to ∼ 8 km is 35°-45°; - at depth greater than ∼ 8 km is < 20°. 2. Seismogenic layer: aftershocks occurr at depths raging 2-8 km, above the basement top.

28

3. The mainshocks are located close to the bottom of the seismogenic layer, within the Triassic Evaporites.

In the Northern part of UFS, the SW-dipping normal faults are antithetic to a major, NE dipping LANF, the Altotiberina fault [Boncio et al., 1998; 2000]. In this region, abundant microseismicity is released along ATF and in the hangingwall block as well. South of Gualdo Tadino, the presence of such a fault can be reasonably hypothesized, but no seismic reflection data is available, constraining its geometry at depth.

Next In the Umbria-Marche mountain ridge SW-NE extension is accommodated by both: NE-dipping

LANF (20°-30° dipping), characterised by both aseismic slip and microseismicity [Chiaraluce et al., 2007]; and SW-dipping normal faults (40°-50° dipping), generating moderate earthquakes (5.0 < M < 7.0). Can this model be extended to other portions of the Northern Apennines extensional belt ? (e.g. Etrurian Fault System, Boncio et al., 2000).

We have also seen that the seismicity related to the major earthquakes occurs within the Mesozoic sedimentary cover and do not penetrate the underlying basement; the mainshocks are located within the Triassic evaporites, consisting of alternated dolomites and anhydrites. The mechanical properties of these rocks (and of the heterogeneous rock assemblages which the underlying basement consists of) need to be investigated, in order to understand the mechanics of the normal faults earthquakes.

A better comprehension of earthquakes and fault mechanics in this region also address a better comprehension of very controversial subjects of general interest, such as:

- the factors influencing the structure and thickness of the seismogenic layer in continental regions;

- the mechanics and seismogenics role of LANFs; - the activity of seismogenic, non-planar or curviplanar normal faults; - the role of fluids in earthquakes nucleation and triggering [e.g. Miller et al., 2004].

References The citation of all the papers which the concepts exposed in this paper are referred to would have

exceeded the available space. I have only cited a few papers, where the main concepts of this abstract are reported in an expanded and more complete form, along with the necessary list of references.

Boncio, P., Brozzetti, F. and Lavecchia, G., (2000). Architecture and seismotectonics of a regional low-angle normal fault zone in central Italy. Tectonics, 19, 1038-1055.

Boncio, P., Brozzetti, F., Ponziani, F., Barchi, M. R., Lavecchia, G. and Pialli, G., (1998). Seismicity and extensional tectonics in the Northern Umbria-Marche Apennines. Memorie della Società Geologica Italiana, 52, 539-555.

Chiaraluce, L., Barchi, M.R., Collettini, C., Mirabella, F. and Pucci, S. (2005). Connecting seismically active normal faults with Quaternary geologicalstructures in a complex extensional environment: the Colfiorito 1997 casehistory (Northern Apennines, Italy). Tectonics (0278-7407), TC1002, 24, doi: 10.1029/2004TC001627.

Chiaraluce, L., Ellsworth, W. L., Chiarabba, C. and Cocco, M. (2003). Imaging the complexity of an active normal fault system: the 1997 Colfiorito (Central Italy) case study. Journal of Geophysical Research, Vol.. 108, No. B6, 2294, 10.1029/2002JB002166.

Ciaccio, M.G., Barchi, M.R., Mirabella, F. and Chiarabba, C. (2005). Seismological, geological and geophysical constraints for the Gualdo Tadino fault, Umbria-Marche Apennines (Central Italy). Tectonophysics (0040-1951), 406, 233-247

Lavecchia, G., Brozzetti, F. Barchi, M.R., Menichetti M. and Keller, J. V. A., (1994). Seismotectonic zoning in east-central Italy deduced from analysis of the Neogene to present deformations and related stress fields. Geol. Soc. of Am. Bull., 106, p. 1107-1120.

Miller, S.A., Collettini, C., Chiaraluce, L., Cocco, M., Barchi, M.R. and Kaus, B.J. (2004). Aftershocks driven by a high-pressure CO2 source at depth. Nature (0028-0836), February 19, 2004, 427, 724-727.

Mirabella, F., Barchi, M.R., Lupattelli, A., Stucchi E. and Ciaccio, M.G. (2007) - Insights on the seismogenic layer thickness from the upper crust structure of the Umbria-Marche Apennines (Central Italy) Submitted to Tectonics

29

Mirabella, F., Ciaccio, M.G., Barchi, M.R. and Merlini, S. (2004) The Gubbio normal fault (Central Italy): geometry, displacement distribution and tectonic evolution. Journal of Structural Geology (0191-8141), 26, 2233-2249

Pialli, G., Barchi, M. R. and Minelli, G. (1998). Results of the CROP03 deep seismic reflection profile. Memorie della Società Geologica Italiana, 52, 657 p.

30

31

Developments of seismic monitoring in Italy from 1997 to 2007

Alessandro Amato Istituto Nazionale di Geofisica e Vulcanologia

1. Introduction In this contribution I will summarize the huge efforts undertaken in the last 10 years by many of my

colleagues, mostly INGV researchers and technicians, to bring our monitoring system, including the National Seismic Network, MedNet, the portable network, the GPS networks, to the present, high quality standard.

The 1997 Umbria-Marche earthquake represents a milestone in the development of the monitoring system in Italy in the past decade. It was only in 2001, after the creation of INGV and the signature of the first 3-year agreement between INGV and the Department of Civil Protection (DPC), that the plan of an integrated monitoring system really takes off. However, the 1997 earthquake marks the awareness of the Italian seismologists that something had to be done to understand the earthquake process, and that a basic starting point is the availability of modern seismic and geodetic data. One main target was to fill the gap between permanent networks, developed during the ‘80’s and still operational in 1997, and the temporary networks that were already able to provide high-resolution images of the active faults, though not in real time [see Chiarabba, this volume]. An interesting point that came out after the long and complex analysis of data from the dense local network, is that we could predict the behaviour of fault activity (i.e., seismicity migration from one fault to another, activation of adjacent segments, etc.) during the sequence itself, if detailed data were available in real time.

2. The Italian National Monitoring System In 1997, the real-time National Seismic Network (Rete Sismica Nazionale, RSN hereinafter)

managed by INGV consisted of about 90 short period stations, with a sparse coverage and with low dynamic range. The main problems of such network were the low location accuracy in many regions, the time response, and the difficulty in determining correct magnitudes, due to the low dynamic range of the telephone lines used for data transmission. For small earthquakes, at least 5 minutes were needed to get hypocentre location and duration magnitudes (routinely used at that time), but even more time was necessary for M>4 earthquakes, for which the waveforms of the short-period network were mostly saturated. For moderate and large earthquakes, only data from broad band MedNet stations could be used for magnitude computation, but in 1997 they had dial-up connections and at least 15 to 30 minutes were needed to retrieve and analyze data.

In 2001 a strong improvement of the networks started, mainly the RSN, but also MedNet, the temporary seismic network, and finally, after 2004, the National GPS network (see http://ring.gm.ingv.it/). In this contribution, I will summarize the progress done with the permanent seismic networks, from the broad Mediterranean scale (MedNet) to the National scale (RSN), and finally to the regional networks. The local networks and data are described in other presentations at this meeting [Selvaggi, Chiarabba].

2.1 MedNet For moderate to large earthquakes in Italy and the surrounding regions, in order to have a good

coverage from different azimuthal directions and distances, it is important to have data available not only within Italy but also around it. For this reason, MedNet was improved with some new stations installed in various countries of the Mediterranean, and connecting in real-time all the stations, including some shared with our foreign partners. This guarantees broad band unsaturated seismic waveforms for all Mediterranean earthquakes and for teleseismic events in real time, useful for moment tensor and magnitude computations. Today, any M>4 earthquake occurring in the Mediterranean region is recorded by several stations at regional distance; the waveforms can be accessed immediately on our MedNet web site

32

(http://mednet.ingv.it). A recent example of data availability for a M5 earthquake in Greece is shown in Figure 1, where it is possible to see that data from many stations in Italy and from Malta, Morocco, Spain, France, Austria, Czech Republic, Romania, Bulgaria, Serbia, Albania, Greece, Turkey, Israel, are recorded and distributed in real time. The focal mechanism reported in the Figure was computed automatically after a few minutes and then reviewed by a seismologist. On-going developments include the set up of a Mediterranean warning system for potentially tsunamigenic earthquakes in the Mediterranean.

Figure 1. Real time data availability for a recent M5 earthquake (12/08/2007 11:20 UTC) on the MedNet web site. Data from MedNet stations and other networks are available, following EuroMediterranean agreements among many Institutes. Note the good azimuthal and distance coverage for such earthquake. The TDMT moment tensor is done automatically for earthquakes in the Mediterranean region and then reviewed by seismologists.

Figure 2. Number of seismic stations connected in real time to the INGV national monitoring system from 1988 to mid-2007. The strong impulse was given by the INGV-DPC first 3-year agreement, in 2001, and is still going on.

2.2 The National Seismic Network The National Seismic Network (RSN) has developed mainly after 2001 [Amato et al., 2006]. Both

the number of seismic stations and their quality increased significantly. Most of the new stations have a broad band seismometer, a strong motion sensor and a continuous GPS receiver, all connected in real time. Figure 2 shows the increase in the number of seismic stations connected in real time from 1988 to 2007.

33

The present configuration of the RSN is good enough to guarantee rapid and accurate locations in most regions of Italy. Some work is still going on to improve the coverage in a few areas. The Umbria-Marche region is well covered, with the exception of western Umbria. Figure 3 shows the comparison between the RSN status on 1997 and today.

Figure 3. Comparison between the RSN geometry in 1997 and in 2007. Besides the number of stations that increased dramatically, also the instrument quality has greatly improved. All the new stations have very broad band (red circles), broad band (green circles) or enlarged band seismometers (yellow circles). The blue circles, predominant until a few years ago, are the short period seismometers [after Cattaneo et al., in preparation].

The continuous data recorded at the network shown in Figure 3 flow in real time to the INGV acquisition system in Rome, where several procedures allow to rapidly recognize local and distant earthquakes, locate them, compute magnitudes and produce reports and maps for the Civil Protection (DPC). For earthquakes in Italy, the first locations are generally ready after 20 seconds; within the first 40-50 seconds, a preliminary estimate of magnitude (ML) is available; all the data are processed and a final more robust location is done within 2-3 minutes after the origin time. The present agreement between INGV and DPC states that the first information with location and magnitude must be sent within the first 120 seconds after any potentially felt earthquake in Italy. More detailed information follows in the following minutes. The accuracy of the preliminary locations is good in most regions of Italy. F. Mele (pers. comm., 2007) estimates that for M ≥ 3.0 about 70% of the earthquakes are automatically located within 5 km from the “true” location, and 90% of them within 10 km. In 2006, about 6000 earthquakes were located by the national network (approximately twice those of 2001), out of which about 600 are communicated to DPC within 2 minutes.

The data recorded today by the National network alone are good enough to correctly and rapidly locate earthquakes in most regions of Italy. However, a good estimate of the hypocentral depth is still limited to a few regions. This still limits the possibility of studying fault geometries and complexities with the National network data alone. In the next section, we will see how the integration with other regional and local networks helps in this direction.

Present and future trends in the rapid evaluation of earthquake activity and of its effects on the territory include i) the “ShakeMap” project, funded in the last 3 years by the DPC (project S4, see http://earthquake.rm.ingv.it/shakemap/shake/index.html), which allows to assess the predicted ground shaking after a few minutes by any earthquake in Italy and the surrounding regions; ii) the development of Early Warning systems, to have the earthquake information data available in target regions before the

34

maximum shaking occurs; iii) real-time connected OBS’s, extending what was recently developed for stand-alone recording (see http://www.ingv.it/primopiano/OBS_H_eng/OBS_H_eng.html) by the INGV-CNT Giblmanna Observatory.

2.3 The Regional Seismic Networks In the last few years, INGV has strengthened partnerships with national and international network

operators. In Italy, agreements with almost all the Institutions running seismic networks has brought to an improved integrated monitoring system. In the region where this was done (besides the INGV sections of Naples, Catania, Milan, and the Arezzo Observatory), namely northwestern Italy, in partnership with the University of Genova; central Italy, together with the Regione Marche, the Fondazione Prato Ricerche, the Osservatorio A. Bina in Perugia, the Gran Sasso INFN Laboratories, the Ximeniano Observatory in Florence, the National Electric Company ENEL) the improvement is significant, in terms of earthquake evaluation, rapidity, and data availability. On-going agreements include the INOGS, active in the Northeast, and the Università della Calabria.

As an example, Figure 4 shows a comparison between the seismic network in central Italy in 1997 and today. The improvement in the Marche-Abruzzi regions is dramatic, considering both the National Network and the two regional networks operated by INGV in the two regions. The Marche area is today one of the best monitored areas in Italy, thanks to the contribution of the Regione Marche, which funded and supports a regional network complementary to the National one. Cattaneo et al. [2007] have shown how the integrated data yields very accurate images of the active faults in the crust and in the uppermost mantle in this area.

Figure 4. Comparison between the National network in 1997 and the integrated networks in central Italy in 2007. The station spacing is now small enough to allow good locations also in depth [from Cattaneo et al., 2007]. Circles: National network (see figure 3 for colors); triangles: integrated regional networks.

References Amato, A., Badiali, L., Cattaneo, M., Delladio, A., Doumaz, F. and Mele, F.M. (2006). The real-time

earthquake monitoring system in Italy. Géosciences (BRGM), vol. 4, pp. 70-75. Cattaneo, M., Amato, A., De Luca, G. and Monachesi, G. (2007). Improved seismic data for an improved

knowledge of the seismogenesis in Central Italy. In preparation for Tectonophysics, special issue dedicated to G. Cello.

Acknowledgments The work reported in this presentation has been carried out by many researchers and technicians of the INGV -

Centro Nazionale Terremoti. In representation of them all, I wish to thank the colleagues who led the Laboratories and the Observatories during the last few years: M. Cattaneo, M. Di Bona, F. Mele, S. Mazza, G. Selvaggi, A. Michelini, G. D’Anna, L. Badiali, A. Delladio, C. Salvaterra and all the Seismology Lab, L. Mondiali. Thanks also to A. Basili, F. Doumaz, A. Bono who developed software for the new monitoring system; to G. Monachesi and G. De Luca, for their efforts on the regional networks. Finally, a special thought to Luciano Giovani, who passed this year after giving so much to the Institute and to all of us. The activities described in this presentation have been done with the support of the Italian “Dipartimento di Protezione Civile Nazionale”.

35

Can 3D and 4D seismic tomography help understanding the 1997 Umbria Marche seismic sequence?

Claudio Chiarabba, Pasquale De Gori, and Davide Piccinini Istituto Nazionale di Geofisica e Vulcanologia

Soon after the 26 September 1997 Mw 5.7 and 6.0 mainshocks, several digital seismic stations were installed over the Colfiorito epicentral region, building up a very dense temporary network. The digital stations recorded thousands of aftershocks over a 1-month long period. The availability of this huge high quality data, recorded since the beginning of the sequence, promotes the 1997 Umbria Marche as the best monitored normal faulting sequence until now. In this work we have merged all the high quality P- and S-wave arrival times from permanent and temporary stations. This new implemented data set allows us to improve and extend the model resolution with respect. to previous studies [Michelini et al. 2000, Chiarabba and Amato, 2003].

The main feature revealed by the seismic sequence is the earthquake migration along the fault system and the occurrence of several later mainshocks [Chiaraluce et al., 2003]. Such migration has been recently explained by fluid migration on adjacent fault segments [Antonioli et al., 2005]. In this work, we show how relocated earthquakes and 3D Vp, Vp/Vs, Qp, and Qp/Qs models of the uppermost crust help understand the on-going processes within the fault zone. While P-wave velocities are mostly controlled by lithology, Vp/Vs, Qp, and Qp/Qs can be used to infer the presence of fluids and the degree of saturation and fluid overpressure within the uppermost crust.

The new velocity and attenuation images have been computed selecting about 1600 aftershocks recorded by more than 60 seismic stations. A total of about 36,000 P- and S-wave arrivals and 15,000 t* values, computed by the decay of the P and S- waves velocity spectra, have been inverted to solve simultaneously for hypocentral and velocity and, independently, for attenuation parameters. We used the Simulpsq inversion code. The models are parameterized with a 3D grid of nodes, spaced 2km in x and y, and 1 km in depth.

Figure 1 shows vertical sections of Vp, Vp/Vs, Qp and Qp/Qs along the fault system. Beneath the Colfiorito region, we clearly define a structural high characterized by strongly positive Vp and Qp anomaly which inner bulk is composed by Triassic evaporitic rocks (Vp> 6.0 km/s, Qp> 200). Mainshocks and large aftershocks originated mostly at the top of the evaporites, within the Mesozoic limestone layer.

The joint interpretation of velocity and attenuation models constrain the fault characteristics at depth. We find that large aftershocks occurs on top of a high Vp, high Qp, low Vp/Vs evaporitic layer. High Vp/Vs and low Qp/Qs anomalies located above the evaporitic rocks suggest that over-pressured fluids propagate within the limestone layer. The observed earthquake migration and the occurrence of later mainshocks are explained as due to variation of fluid pressure on fault segments, inferred from velocity and attenuation models.

We also show preliminary results from repeated seismic tomography (4D in space and time). The huge high quality data set allows us to subdivide the sequence into three sub-epochs and compute Vp and Vp/Vs models for each epoch. We document that time resolved variations of Vp/Vs anomalies trace the propagation of rock fracturing and fluid overpressure along the fault system.

36

Figure 1. Vertical sections of velocity and attenuation models along the Colfiorito normal fault system. The brown lines indicate the resolved volumes. The three mainshocks and large aftershocks are indicated by large and small stars.

References Antonioli A., Piccinini D, Chiaraluce L, Cocco M. (2005). Fluid flow and seismicity pattern:Evidence

from the 1997 Umbria Marche (central Italy) seismic sequence, Geophys. Res. Lett., 32, doi:10.1029/2004GL022256.

Chiarabba C. and Amato A (2003). Vp and Vp/Vs images of the Colfiorito fault region (Central Italy): a contribute to understand seismotectonic and seismogenic processes, J. Geophys. Res., 108, 10.1029/2001JB001665.

Chiaraluce L., Chiarabba C., Cocco M., and Ellsworth W.L. (2003). Imaging the compexity of a normal fault system: The 1997 Colfiorito (Central Italy) case study, J. Geophys. Res., 108, 10.1029/2002JB00216.

Michelini, A., Spallarossa D., Cattaneo M., Govoni A., and Montanari A. (2000). The 1997 Umbria-Marche (Italy) earthquake sequenze: Tomographic images obtained from the GNDT-SSN temporary networks, J. Seismol., 4, 415-433.

37

Contribution of seismic moment tensors to seismic deformation and stress field studies in Italy

Andrea Morelli and Silvia Pondrelli Istituto Nazionale di Geofisica e Vulcanologia

The seismic sequence that hit Umbria and Marche during 1997 and 1998 was significant for a number of reasons, among which the fact that it marked in a dramatic way the evolution of analysis tools, and the development of geophysical knowledge of the region. In particular, since September 1997 we have a growing, coherent catalog of source moment tensors, that provides reliable information on the earthquake sources and the seismic strain field. This, together with other data, contributes to infer the regional stress field, and the large-scale geodynamic frame.

Regional Seismic Moment Tensors The Umbria-Marche earthquake sequence that hit central Italy during 1997 and 1998, marked the

beginning of the routine determination, in Italy, of seismic source parameters by long-period waveform fitting, that continues until today. This project has since broadened in its scopes, to reach the dimensions of the current European-Mediterranean RCMT catalog, extended backwards in time, and including both quick and revised solutions.

The Regional Centroid-Moment Tensor (RCMT) technique, first applied to to the events of the sequence starting in September 1997, uses a modification of the algorithm known as Centroid-Moment Tensor (Harvard/Global CMT), applied to modelling intermediate-period surface waves recorded at regional distance. Harvard/Global CMTs are routinely calculated (originally at Harvard, now at Columbia University) for M>5.5 at the global scale, and the CMT Catalog represents the best known and most reliable worldwide reference for seismic source parameters in the seismological community. The standard CMTs are calculated by fitting long-period body waves, together with long-period (T>135s) surface waves only for large magnitude events. For moderate-magnitude events, this signal generally has small signal-to-noise ratio, such that reliable inversion is not possible. On the other side, for shorter periods surface waves represent the most evident wave train at regional distance, even for moderate magnitude. These waves are highly influenced by lateral heterogeneity of the Earth structure, and their use, without an accurate account of wave speed variations, would produce results biased by unmodelled wavefield propagation effects. The approach followed, therefore, computes and fits sinthetic seismograms for Love and Rayleigh surface waves, with period between 35 and 60 seconds, using detailed phase velocity maps developed with high resolution at global scale. As Global CMTs, also surface-wave, Regional CMTs are result of a slight hypocenter relocation, both in depth and horizontal, during inversion. By resorting to data with period generally longer than the source duration, CMTs and RCMTs better represent the average source properties, and its location as a centroid rather than a 'true' hypocenter.

Regional CMTs were calculated, for the first time, for the Umbria-Marche seismic sequence, for more than 40 events occurred between fall 1997 and summer 1998, having magnitude Mw between 4.5 to 6.0 [Ekström et al., 1998; Morelli et al., 2000; Pondrelli et al., 2002]. As the seismic sequence was going on, and new solutions were constantly computed, a remarkable coherence among all mechanisms became quite striking. All moment tensors show exension along a NE-SW direction, confirmed by other studies made on the earthquakes. Only one event showed a strike-slip mechanism, but its differences from all the others of the sequence also extended to its larger focal depth, at about 50 km (Figure 1).

The situation has considerably evolved during this decade. Since 1997, source parameters for more than 600 earthquakes in the European and Mediterranean region have been computed, and more are being regularly included in a coerent Catalog [Pondrelli et al, 2002; 2004; 2007] also accessible on the worldwide web at www.ingv.it/seismoglo/RCMT. For each earthquake with magnitude between 4.5 and 5.5 (larger events can also be modelled, but are always analysed within the Global CMT project) in the Mediterranean and European region, a fast solution is computed using seismogram data available in near-real time (QuickRCMT), and published online (mednet.rm.ingv.it/quick_rcmt.php). These solutions are

38

then revised, when more data have become available through data centers (MEDNET, ORFEUS, IRIS). The final solutions are published in the literature and on the worldwide web. With time, the number of earthquake source mechanisms per year has increased, as the minimum magnitude threshold has decreased, in many cases down to about Mw=4.0. This is possible by virtue of the increased availability of broadband seismographic data at epicentral distance of a few hundred kilometers. Some of the source mechanisms of events of the Umbria-Marche earthquake sequence have been calculated with as few as 3 stations. Nowadays, instead, even a fast QRCMT generally results from analysis of about 10 stations, and final solutions are based on data from up to 30 stations. Of course, the geographic location of the event largely determines how good its data coverage will be, especially for smaller magnitudes that are not well recorded beyond few hundred kilometers. A uniform azimuthal coverage is actually the most critical factor.

The Italian subset of regional centroid-moment tensors has been extended in time, going back for as long as possible by the data requirements, so that it includes earthquakes from 1977 to present time [Pondrelli et al., 2006; www.ingv.it/seismoglo/RCMT/Italydataset].

Figure 1: Moment tensor solutions for the main events of the 1997-98 Colfiorito sequence [after Morelli et al., 2000]. Each solution is labeled with occurrence date (month and day). Focal sphere radii are proportional to moment magnitude.

Seismic deformation The use of moment tensor summation techniques allows the distributed deformation due to seismic

activity to be quantified. On the basis of this analysis one can evaluate the summary geometrical characters of seismic activity, integrated over time, and the importance of seismic motion relative to the overall deformation along the main active zones. This kind of study has been applied over all the entire Mediterranean since the 80’s, by virtue of the strong seismic activity characterizing the region, and of the consequentially quite large dataset available [Jackson and McKenzie, 1988]. The main features of the seismic strain were drawn in early papers, that mostly reported the deformation pattern at large scale along the main active tectonic boundaries [Pondrelli et al., 1995].

During the last decade many results have been added, giving a description of changes in the seismic strain pattern also for regions characterized by diffuse deformation and not only for zones related to large rigid plate motion [Selvaggi, 1998; Vannucci et al., 2004; Serpelloni et al., 2007]. The possibility to apply the moment tensor summation to smaller grids, compared to the large volumes over which seismic deformation was averaged in the older papers, is mainly due to the improvement of the dataset available. For this use, the numerous centroid-moment tensors made available in the last 10 years have shown to be a reliable source of information, even for regions of relatively moderate strain rate. Added to the EMMA database [Earthquake Mechanisms of the Mediterranean Area, Vannucci et al., 2004], they allowed the fine definition of styles of seismic deformation in Italy (Figure 2), and all around the Mediterranean, with also a

39

sketch of the depth distribution of the strain pattern. Where the dataset is rich enough, also a reconstruction of block modeling can be supported by seismic deformation determined at small scale and compared/added to geodetic deformation [Serpelloni et al., 2007]. This would have not have been possible 10 years ago.

Figure 2: Central Italy. Sum of moment tensors on a regular grid with mesh of half a degree, for earthquakes with depth < 50 km [after Vannucci et al., 2004].

The active stress field In the 90’s, data describing the state of stress of the crust have been gathered and used to draw and

study the active stress field and its relation with geological features. In the World Stress Map [Zoback, 1992] the active stress field for the Italian region was poorly constrained, due to the scarcity of data. For the same reasons, the interpolated map of the stress field in the Mediterranean [Rebaï et al., 1992] was particularly unreliable. Less than 10 years later, the active stress map of Italy [Montone et al., 1999; 2004] was the instrument to draw some of the most interesting, and sometimes still in discussion, features of the geodynamic framework of Italy. The active NE-SW extension along the Apennines shows to be focused along the inner belt, while stress data underline that in the external part of the chain a compressive regime is active. This rotation of stress directions, that is clearly present along the Northern Apennines, has strong implications in the different hypotheses on the process that produced the Apennine belt.

Figure 3: Smoothed Shmin orientation and stress regime of Central Italy. Legend: a) indicates Shmin directions from our data set and the World Stress Map 2003 release (colors mean different tectonic regimes: black is unknown, blue is compressional, green is strike-slip and red is extensional regime); b) indicates smoothed stress directions (colors as before); c) indicates direction of inferred regional stress regime (colors as before); and d and e indicate integrated seismogenic sources from geological/geophysical (yellow) and well constrained historical (grey) data [after Montone et al., 2004].

These results have been obtained thanks to the increasing availability of data reporting information on the state of the active stress in the crust. This contribution has been given by breakout data, made

40

available by oil companies (ENI-AGIP), by studies on fault data, and by the huge amount of new data produced by moment tensor computation. This allowed to have about 200 horizontal stress directions in the first version of the active stress map of Italy [Montone et al., 1999]. This number reached nearly 400 in the last version of the map [Montone et al., 2004], inferred from 61% of borehole breakouts, 38% of earthquake data and 1% of fault data, allowing to draw a reliable smoothed map of the stress field active along our peninsula (Figure 3).

References Ekström, G., Morelli, A., Boschi, E. and Dziewonski A.M., (1998). Moment tensor analysis of the central

Italy earthquake sequence of September-October 1997, Geophys. Res. Let., 25, 1971-1974. Jackson, J. and McKenzie, D., (1988). The relationship between plate motions and seismic moment

tensors, and the rates of active deformation in the Mediterranean and Middle-East. Geophys. J. Int. 93, 45–73.

Montone, P., Amato, A. and Pondrelli, S., (1999). Active Stress Map of Italy, J. Geophys. Res., 104, 25595-25610.

Montone, P., Mariucci, M. T., Pondrelli, S. and Amato, A., (2004). An improved stress map for Italy and surrounding regions (central Mediterranean), J. Geophys. Res., 109, doi:10.1029/2003JB002703.

Morelli, A., Ekström, G. and Olivieri, M., (2000). Source properties of the 1997-1998 Central Italy earthquake sequence from inversion of long period and broadband seismograms, J. Seismol., 4, 365-375.

Pondrelli, S., Morelli, A. and Boschi, E., (1995). Seismic deformation in the Mediterranean area estimated by moment tensor summation. Geophys. J. Int., 122, 938-952.

Pondrelli, S., Morelli, A., Ekström, G., Mazza, S., Boschi, E. and Dziewonski, A.M., (2002). European-Mediterranean regional centroid-moment tensors: 1997-2000, Phys. Earth Planet. Int., 130, 71-101.

Pondrelli S., Morelli, A., and Ekström, G., (2004). European-Mediterranean Regional Centroid Moment Tensor catalog: solutions for years 2001-2002, Phys. Earth Planet. Int., 145, 1-4, 127-147.

Pondrelli, S., Salimbeni, S., Ekström, G., Morelli, A., Gasperini, P. and Vannucci, P., (2006). The Italian CMT dataset from 1977 to the present, Phys. Earth Planet. Int., 159/3-4, pp. 286-303.

Pondrelli, S., Salimbeni, S., Morelli, A., Ekström, G. and Boschi, E., (2007). European-Mediterranean Regional Centroid Moment Tensor catalog: solutions for years 2003-2004, Phys. Earth Planet., 164, 90-112.

Rebaï, S., Philip, H. and Taboada A., (1992). Modern tectonics stress field in the Mediterranean region: Evidence for variation in stress directions at different scales, Geophys. J. Int., 110, 106–140.

Selvaggi, G., (1998). Spatial distribution of horizontal seismic strain in the Apennines from historical earthquakes, Ann. Geofis, 44, 155-165.

Serpelloni E., Vannucci, G., Pondrelli, S., Argnani, A., Casula, G., Anzidei, M., Baldi, P. and Gasperini, P., (2007). Kinematics of the Western Africa-Eurasia plate boundary from focal mechanisms and GPS data, Geophys. J. Int., doi: 10.1111/j.1365-246X.2007.03367.x

Vannucci, G., Pondrelli, S., Argnani, A., Morelli, A., Gasperini, P. and Boschi, E., (2004). An atlas of Mediterranean seismicity, Ann. Geophys., 47, 247-306.

Zoback, M. L., (1992), First- and second-order patterns of stress in the lithosphere: The World Stress Map Project, J. Geophys. Res., 97, 11703–11728.

41

Spatio-Temporal Modeling of the Umbria-Marche 1997 Seismic Sequence

Warner Marzocchi Istituto Nazionale di Geofisica e Vulcanologia

The main goal of this work is to review the scientific researches carried out after the Umbria-Marche sequence related to the modeling of the spatio-temporal evolution of the sequence. Different physical and statistical aspects are taken into considerations, ranging from the modeling of faults interaction and the presence of fluids, to the stochastic description through Epidemic-Type Aftershock Sequence (ETAS) or more complex models. Finally, we discuss one topical point of the spatio-temporal modeling that has a paramount importance for society, the earthquake forecasting issue.

The seismic sequence occurred in a period of time dominated by the scientific concept of fault interaction via Coulomb Failure Function (CFF). Remarkably, the Umbria-Marche sequence represents a clear failure of this model to explain the whole or even the most relevant part of the spatial distribution of the sequence. In particular, Cocco et al. [2000] showed that CFF cannot explain the occurrence of some M 5.0+ events and, in general, of the high seismicity on the hanging wall of the fault where the “mainshock” occurred at the beginning of the sequence. This discrepancy was explained by introducing a significant influence of fluids flowing across the unclumped faults that would have favoured the occurrence of earthquakes [Miller et al., 2004; Antonioli et al., 2005]. This hypothesis is still matter of discussion and study, mostly because of the scarce direct evidence of fluid flowing across the earth. The only clear evidence is that Umbria-Marche region, as well as a large part of the italian territory, is characterized by a very high CO2 flux [Chiodini et al., 2004] that has been suggested to be related to the spatial distribution of the italian seismicity.

The CFF accounting for fluids can, at best, explain the spatial distribution of the earthquakes and, retrospectively, the succession of the largest earthquakes of the sequence. Anyway, in general, it cannot explain in a forward perspective the time evolution of any sequence, and it cannot provide any insights on the magnitude of the expected events. The temporal and size distributions are usually explained by stochastic modeling. In this respect, Console et al. [2003a] modeled the sequence through a classical ETAS model. The results show an impressive similarity between the time evolution of the real and expected seismicity. The only remarkable difference is that the expected number of events was a factor of three less than what really observed. This discrepancy can be due mostly to two different factors: 1) the use of a minimum magnitude for the ETAS modeling equal to the completeness magnitude can induce a similar bias [Sornette and Werner, 2005]; 2) a classical stationary ETAS model is not appropriate to explain the time-dependent effects of the fluids migration. In order to explain better the time evolution of the sequence, Lombardi et al. [2007, in prep.] have modeled the sequence through a nonstationary ETAS modeling. The difference with respect to the classical ETAS model is that the seismic background can vary with time, accommodating possible time variations in short-time interval induced by fluids migration. The results obtained are quite promising and explained better the temporal evolution of the Umbria-Marche seismic sequence.

As regards the size of the events, the Umbria-Marche sequence provided a further evidence against the well-known Bath law, that states that the maximum magnitude for an aftershock is 1.0-1.5 degree less that the mainshock. A formal statistical analysis carried out by Console et al. [2003b] clearly showed its inadequacy. This aspect has a pivotal importance in practice, because it suggests to not lower the social attention after a so-called mainshock, because aftershocks can have a comparable magnitude if not larger. In this respect, Umbria-Marche sequence represents a tutorial example.

The final point is related to earthquake forecasting that maybe represents the most relevant seismological contribution for society. One of the main goals of the retrospective analysis of the spatio-magnitude-temporal distribution(s) (like the Umbria-Marche seismic sequence) is to set up a reliable earthquake forecast model. Such a model has a two foremost prerogatives: 1) it allows any kind of hypothesis on earthquake generating processes to be scientifically tested through a real independent dataset

42

(i.e., the future events), and 2) it provides a basic ingredient for seismic hazard assessment, and clues for Civil Protection to set practical recipes for inhabitants of seismic areas during seismic crises. Before the occurrence of the Umbria-Marche sequence there was no explicit earthquake forecast map for this area. Boschi et al. [1995] published a probability map including this area but their forecast was made for magnitudes slightly larger than the ones observed (M 5.9+). Nevertheless, it is worth noting that the Umbria-Marche region has the sixth highest probability of occurrence in the interval 1995-2000 out of 20 regions. After the Umbria-Marche seismic sequence many works were devoted to improve our forecasting capability and some works supplied earthquake forecasts for this area. Basically, the attempts can be divided in two groups: 1) estimating probabilities on single seismic faults, 2) estimating probabilities on seismotectonic areas. In spite of anyone agrees that earthquakes occur on faults, the basic difference between the two approaches is linked to the existence of a complete seismogenic fault dataset. In fact, the first basic ingredient (not the only one) for a reliable earthquake forecast on single faults is the availability of a complete dataset [Marzocchi, 2007]. On the contrary, the use of seismotectonic areas is a way to accommodate such a possible epistemic uncertainty. Not quite surprisingly, the two approached led to almost opposite results. For example, Pace et al. [2006] estimated very low probabilities for the faults in Umbria-Marche region, while Faenza et al. [2003] and Cinti et al. [2004], analyzing seismotectonic areas, identified the Umbria-Marche region as one of the highest dangerous area in Italy (together with part of Calabria arc, part of Southern Apennines, and Friuli). Other attempts have been made [e.g., Romeo, 2005] and some are still in progress [i.e., in the frame of S2 INGV/DPC project, and of the European NERIES and SAFER projects], but the different approaches, philosophies, datasets (see above) and physical models used still produce results quite different. We argue that one of the major problems in achieving a general consensus on this important issue is due to the difficulty in testing and comparing the forecasting capability of different models; in practice, several models are not testable at all. We foresee in the next future more attempts to fill this gap and produce a class of models that are testable in a forward perspective.

References Antonioli, A., Piccinini, D., Chiaraluce, L., Cocco, M. (2005). Fluid flow and seismicity pattern: Evidence

from the 1997 Umbria-Marche (central Italy) seismic sequence, Geophys. Res. Lett., 32, L10311, doi:10.1029/2004GL022256.

Boschi, E., Gasperini, P., Mulargia, F. (1995). Forecasting where larger crustal earthquakes are likely to occur in Italy in the near future. Bull. Seismol. Soc. Am., 85, 1475 – 1482.

Chiodini, G., Cardellini C., Amato A., Boschi E., Caliro S., Frondini F., Ventura G. (2004). Carbon dioxide Earth degassing and seismogenesis in central and southern Italy. Geoph. Res. Lett., 31, doi:10.1029/2004GL019480.

Cinti F., Faenza, L., Marzocchi, W., Montone, P. (2004). Probability map of the next large earthquakes in Italy. Geochem. Geophys. Geosyst., 5, Q11003, doi:10.1029/2004GC000724.

Cocco, M., Nostro, C., Ekstrom, G. (2000). Static stress changes and fault interaction during the 1997 Umbria-Marche earthquake sequence. J. Seismol., 4, 501–516.

Console, R., Murru, M., Lombardi, A.M. (2003a). Refining earthquake clustering models. J. Geophys. Res., 108(B10), 2468, doi:10.1029/2002JB002130.

Console, R., Lombardi, A.M., Murru, M., Rhoades, D. (2003b), Bath’s law and the self-similarity of earthquakes, J. Geophys. Res., 108(B2), 2128, doi:10.1029/2001JB001651.

Faenza, L., Marzocchi, W., Boschi, E. (2003). A non-parametric hazard model to characterize the spatio-temporal occurrence of large earthquakes: An application to the Italian catalogue. Geophys. J. Int., 155(2), 521 – 531.

Lombardi, A.M., Marzocchi, W., Cocco, M. (2007). A stochastic model for earthquake occurrence accounting for fluid motion: The case of Umbria Marche 1997 sequence. In preparation.

Marzocchi, W. (2007). Comment on “Layered Seismogenic Source Model and Probabilistic Seismic Hazard Analyses in Central Italy”, by B. Pace, L. Peruzza, G. Lavecchia, and P. Boncio, Bull. Seismol. Soc. Am., 97, 1-3, doi: 10.1785/0120060192.

Miller, S.A:, Collettini C., Chiaraluce, L., Cocco, M., Barchi, M., Kaus, B.J.P. (2004). Aftershocks driven by a high-pressure CO2 source at depth. Nature, 427, 724-727.

Pace, B., Peruzza, L., Lavecchia, G., Boncio, P. (2006). Layered Seismogenic Source Model and Probabilistic Seismic-Hazard Analyses in Central Italy. Bull. Seismol. Soc. Am., 96, 107–132, doi: 10.1785/0120040231.

43

Romeo R.W. (2005). Earthquake Hazard in Italy, 2001–2030. Natural Hazards, 36, 383-405, DOI:10.1007/s11069-005-1939-1

Sornette D., Werner, M.J. (2005). Apparent clustering and apparent background earthquakes biased by undetected seismicity. J. Geophys. Res., 110, B09303, doi:10.1029/2005JB003621, 2005

44

45

The coseismic ground deformations of the 1997 Umbria-Marche earthquakes:

a lesson for the development of new GPS networks

Marco Anzidei Istituto Nazionale di Geofisica e Vulcanologia

After the occurrence of the two main shocks Mw=5.7 (00.33 GMT) and Mw=6.0 (09:40 GMT), on September 26, 1997, which caused severe damaged and ground cracks in a wide area of the Umbria Marche region [Tosi et al., 1999], the Istituto Nazionale di Geofisica in cooperation with the Istituto Geografico Militare Italiano, with the aim to detect the coseismic ground deformations, reoccupied 13 geodetic monuments placed across the epicentral area, belonging to the first order Italian GPS network IGM95 [Surace, 1993] and to the Tyrgeonet [Anzidei et al., 1995]. These stations consist in i) concrete pillars, ii) markers fixed on the ground or iii) markers placed on stable, small concrete buildings, undamaged by the coseismic ground shaking. Surveys were performed using advanced dual frequency Trimble 4000 SSE/SSI GPS receivers equipped with geodetic antennas placed on tripods as well as on pillars through IGM type mount [Anzidei et al., 1998; Anzidei et al., 1999]. The IGM95 network, which consists in about 1260 vertices located in the whole Italy [Surace, 1993; 1997], was planned for land surveys and cartography and specifically designed to be measured through GPS technique with the primary scope to provide a set of reference 3-D coordinates for each station, based on the World Geodetic System 1984 (WGS84). The average accuracy of this network was estimated at about 35 mm and 22 mm at 95% confidence level in the vertical and horizontal components, respectively [Surace, 1997].

Although data were too weak to provide significant results on the aseismic strain at the strain rates expected for the Apennine chain in the few years right after the completion of the network, the location of the geodetic monuments and the re-analysis of the field data, allowed the geophysical community to benefit of a GPS reference network for the estimation of ground coseismic displacements during the 1997 Umbria Marche earthquakes [Anzidei et al., 1997; 1998b; 1999; 2000] as well as recently occurred for the 2003 Molise earthquake [Giuliani et al., 2006]. This network was also used in the following years to provide the first estimates of the active extension across the central Apennines [D’Agostino et al., 2001], previously estimated through the measurement of historical triangulation network of the Italian Istituto Geografico Militare [Hunstad and England, 1999a; Hunstad et al., 2003] or by repeated GPS surveys across the central Apennines [Serpelloni et al., 2001].

The comparison between the 1995 coordinate set and those obtained during the surveys performed in the early days of October 1997 in the Umbria Marche earthquake epicentral area, shown maximum deformation values at the closest stations to the epicentres, up to 14.0±1.8 and 24.0±3.0 cm in the horizontal and vertical components, respectively. The availability of the IGM95 stations allowed translating geodetic data into relevant geophysical results. The estimation of postseismic coordinates at 13 vertices provided, for the first time in Italy, the evidence of significant displacements during a seismic sequence. These measured deformations have been used to identify the fault models responsible of the main shocks and to understand the seismic source mechanics [Hunstad et al., 1999b]. Moreover, the combination of GPS results with ERS-SAR differential interferograms as well as seismological parameters, provided the estimation of the geometry and slip distribution on the fault planes [Stramondo et al., 1999; Salvi et al., 2000].

The same actions were not applied to the October 14, 1997, Mw=5.6 Sellano earthquake, whose epicentre was located a few tenth km south of the previous earthquakes, due to a lack of available GPS vertices of the IGM95 network in the surroundings of the epicentral zone.

This fact, which prevented the estimation of coseismic deformations and seismic source modelling for this earthquake, clarified the need to set up tailor made GPS networks devoted to geophysical applications. Among the others, networks requirement were high accuracy monuments, suitable average spacing between stations with respect to significant active tectonic structures or faults, capability to capture the eventual coseismic and interseismic signal at the surface of the Earth’s crust at the scale of the expected

46

magnitudes and fault’s length [Johnson H.O., and F.K. Wyatt, 1994; Boschi et al., 1995; Valensise and Pantosti, 2001b; Wells and Coppersmith, 1994]. With this aim and the need to have a fiducial network, since 2000 the number of continuous monitoring stations were strongly increased in Italy, under national or international research programs and monitoring projects funded by the Italian Space Agency, INGV (under contracts of MIUR and Dipartimento della Protezione Civile) and other institutions (mainly Universities and Regioni), following the examples of USA [Zhang et al., 1997] and Japan [Miyazaki et al., 1994] which were establishing hundreds of stations since 1990 to measure the crustal dynamics and the coseismic displacements of their highly seismic regions [Bock et al., 1993; Tsuji et al., 1995 and references therein]. Increasingly GPS networks, whether regional or local, are the mainstay of deformation monitoring, especially over large areas and offer high accuracy and continuous observation as experimented for the case of the 2004 Sumatra earthquake [Boschi et al., 2006]. Dense regional networks, such as SCIGN (USA) demonstrate the value of such systems and the International GPS Service (IGS), has provided valuable scientific data and products to users since 1994.

In the meantime in Italy were also developed not permanent networks under specific projects, through the establishment of a large number of additional geodetic vertices in selected areas, such as the central Apennines [Anzidei et al., 2005] or Southeastern Sicily [Bonforte et al., 2002].

Presently, the gap has been quite filled and nowadays in the Italian region are active continuous and not permanent GPS networks useful for geophysics studies. To the former belongs the national GPS network RING managed by INGV, consisting of more than 100 continuously monitoring stations remote controlled mainly by satellite systems, with an average grid at about 50 km or less [Selvaggi et al., 2006], as well as the networks managed by the Italian Space Agency [Vespe et al., 2000], University of Bologna and Siena in Tuscany region [Cenni et al., 2006] and other institutions. To the latter, besides the IGM95 networks, more than 300 benchmarks belong to sub-regional or local network belonging to the INGV, distributed in selected seismic areas of the Italian region. The aim of these networks is to provide a detailed view of the current deformation during the seismic cycle in the Italian region into a unique reference system. Moreover, the combination of GPS and seismic observations help determine location and extent of co-seismic deformation. For this reason, during the last years have been deployed in Southern California the Integrated GPS Network (SCIGN) and the RING has followed this example in Italy.

First pioneering geodetic results at global and regional scale for the Mediterranean region are reported since [Smith et al., 1994; Anzidei et al. 1996; 1998a] and more recently in [Anzidei et al., 2001; 2005; Oldow et al., 2002; Hollenstein et al., 2003; Caporali et al 2003, Serpelloni et al., 2002; 2005; 2006; 2007; D’Agostino and Selvaggi, 2004; D’Agostino et al.,2005 and references therein], which all describe an active deformation of the Italian peninsula in the frame of the Mediterranean geodynamics, with variable strain rates and velocities, mainly related with the collision of the Eurasia-Africa plates and local complexities.

Here we show and discuss the development of the local, sub regional and regional GPS networks in the Italian region, as well as the improvements in data analysis, started since the early 1990 and dramatically increased after the 1997 Umbria Marche earthquakes, and the insights gained from this action [Caporali et al., 2003; D’Agostino and Selvaggi 2004; Anzidei et al., 2005, 2006; Selvaggi et al., 2006; Serpelloni et al., 2002; 2005; 2006; 2007; Bennet and Hreinsdóttir, 2007] which can be also integrated as Global Observing Strategy for the monitoring of our Environment from Earth and Space [Igos- Geohazard, 2006].

References Anzidei M., Baldi P., Casula G., Riguzzi F., Surace L., (1995). La rete Tyrgeonet. Supplemento al

Bollettino di geodesia a scienze affini, vol.LIV, n.2. Anzidei M., Baldi P., Casula G., Crespi M., Riguzzi F., (1996). Repeated GPS Surveys across the Ionian

Sea: evidence of crustal deformations. Geophysical Journal International, n.127, pp. 257-267 Anzidei M. Baldi P. Bonini C. Casula G. Gandolfi S. Riguzzi F., (1998a). GPS surveys across the Messina

Straits (Southern Italy) and comparison with terrestrial geodetic data. Journal of Geodynamics vol.25, n.2, pp.85-97, 1998

Anzidei M. Baldi P. Del Mese S. Galvani A. Guidi C. Hunstad I Pesci A., (1997). Studio preliminare del campo di spostamento e di deformazione prodotto dal terremoto di Colfiorito dall'analisi di dati geodetici. Pubblicazioni dell'ING, n.593, pp.38-41.

47

Anzidei M. Baldi P., Coticchia A., Del Mese S., Galvani A., Hunstad I., Pesci a., Surace L., Zanutta A., (1998b). Utilizzo della rete GPS IGM95 per lo studio delle deformazioni cosismiche dei terremoti umbro marchigiani del 26 settembre 1997. Bollettino di Geodesia e Scienze Affini, n.3.

Anzidei M., Baldi P., Galvani A., Pesci A., Hunstad I. and Boschi E., (1999). Coseismic displacement of the 27th september 1997 Umbria - Marche (Italy) earthquakes detected by GPS: campaigns and data. Annali di Geofisica, vol.42, n.4, 597-607.

Anzidei M., P. Baldi, G.Casula, A. Galvani, A. Pesci, Zanutta A., (2000). Spostamenti rilevati alle stazioni GPS della rete IGM95 poste nell’area dell’Appennino umbro-marchigiano dopo gli eventi sismici del settembre-ottobre 1997. Bollettino di Geodesia e Scienze Affini, anno LIX, n.3, 2000.

Anzidei M., Baldi P., Casula G., Galvani A., Mantovani E., Pesci A., Riguzzi F., Serpelloni E., (2001). Insights on present-day crustal motion in the Central Mediterranean area from GPS surveys. Geophysical Journal International, vol. 146, 98-110.

Anzidei M., P. Baldi, A. Pesci, A. Esposito, A. Galvani, F. Loddo, P. Cristofoletti, A. Massucci and Del Mese S., (2005). Geodetic deformation across the Central Apennines from GPS data in the time span 1999-2003. Annals of Geophysics, Vol. 48, N. 2, April 2005.

Anzidei M., Casula G., Galvani A., Riguzzi F., Pietrantonio G., Massucci A., Del Mese S., Loddo F., Pesci A., Esposito A., (2006). Le prime stazioni GPS permanenti INGV-CNT per il monitoraggio delle deformazioni crostali dell’area italiana. Quaderni di Geofisica n. 39.

Bennett R.A., Hreinsdóttir S., (2007). Constraints on vertical crustal motion for long baselines in the central Mediterranean region using continuous GPS. Earth and Pla. Sci. Lett. in press.

Bock Y., Agnew D.C., Fang P., Genrich J.F., Hager B.H., Herring A. T., Hudnut K.W., King R.W., Larsen S., Minster J.B., Stark K., Wdowinski S., Wyatt F., (1993). Detection of crustal deformation from the Landers earthquake sequenze using continuous geodetic measurements. Nature, vol. 361, 337-340-

Bonforte A., Anzidei M., Puglisi G., Mattia M., Campisi O., Casula G., Galvani A., Pesci A., Puglisi G., Gresta S., Baldi P., (2002). GPS surveys in the foreland-foredeep tectonic system of southeastern Sicily: first results (2002). Annals of Geophysics, vol 45, n.5, pp.673-682.

Boschi, E., Guidoboni, E., Ferrari, G., Valensise, G., Gasperini, P., (1995). Catalogo dei forti terremoti in Italia dal 461 a.C. al 1980, ING SGA, Bologna.

Boschi E., E. Casarotti, R. Devoti, D. Melini, A. Piersanti, G. Pietrantonio and Riguzzi F., (2006). Coseismic deformation induced by the Sumatra earthquake. Volume 42, Issues 1-3, 52-62.

Caporali A., Martin S., Massironi M., (2003). Average strain rate in the Italian crust inferred from a permanent GPS network – II. Strain rate versus seismicity and structural geology. Geoph. Jou. Int., 155, 254-268.

Cenni N., Baldi P., Ferrini M., Mantovani E., Viti M., D’Intinosante V., Babbucci D., Mantovani E., (2006). First results from GPS network in the Northern Apennines. Wegener meeting: Geodesy of the Mediterranean, Nice.

D’Agostino N., Giuliani R., Mattone M. and Bonci L., (2001). Active crustal extension in the central Apennines (Italy) inferred from GPS measurements in the interval 1994-1999. Geoph. Res. Lett., 28, 10, 2121-2124.

D’Agostino N., D. Cheloni, S. Mantenuto, G. Selvaggi, A. Michelini and Zuliani D., (2005). Strain accumulation in the southern Alps (NE Italy) and deformation at the northeastern boundary of Adria observed by CGPS measurements. Geoph. Res. Lett., Vol. 32, L19306, doi:10.1029/2005GL024266

D’Agostino N. and Selvaggi G., (2004). Crustal motion along the Eurasia-Nubia plate boundary in the Calabrian Arc and Sicily and active extension in the Messina Straits from GPS measurements. Jou. of Geoph. Res., Vol.109, B11402, doi:10.1029/2004JB002998.

Giuliani R., M. Anzidei, L. Bonci, S. Calcaterra, N. D’Agostino, M. Mattone, G. Pietrantonio, F. Riguzzi, G. Selvaggi, (2006). Co-seismic displacements associated to the Molise (Southern Italy) earthquake sequence of October - November 2002 inferred from GPS measurements. Tectonophysics, 2006,432, 21-35, 2006.

Johnson H.O., and F.K. Wyatt, (1994). Geodetic network design for fault mechanics studies. Manuscr. Geod. 19, 309-323.

Hollenstein C., Kahle H.G., Geiger A., Jenny S., Goes S., Giardini D., (2003). New GPS constraints on the Africa –Eurasia boundary zone in southern Italy. Geoph. Res. Lett., vol.30, No.18, 1935.

Hunstad I., England P., (1999). An upper bound on the rate of strain in Central Apennines, Italy, from triangulation measurements between 1869 and 1963. Earth and Pla. Sci. Lett., 169, 261-267.

Hunstad I., Anzidei M., Cocco M., Baldi P., Galvani a., Pesci A., (1999). Modelling coseismic displacements during the 1997 Umbria - Marche earthquake (Central Italy). Geophysical Journal International, 139, 283-295.

48

Hunstad I., Selvaggi G., D’Agostino N., England P., Clarke P., Pierozzi M., (2003). Geodetic strain in peninsular Italy between 1875 and 2001. Geoph. Res. Lett., 30, no.4, 1181.

Miyazaki S., Tsuji H., Hatanaka Y., Abe Y., Yoshimura A., Kamada K., Kobayashi K, Morishita H., Iimura Y., (1994). Establishment of the nationwide GPS array (Grapes) and its initial results on the crustal deformation of Japan. Bullettin of the Geographical Survey Inst. Vol. 42.

IGOS-geohazard Theme Report 2003, (2003). IGOS an Integrated Global Observing Strategy for the monitoring of our Environment from Earth and Space. http://dup.esrin.esa.it/igos-geohazards pp. 56.

Oldow, J.S., Ferranti, L., Lewis, D.S., Campbell, J.K., D’Argenio, B., Catalano, R., Pappone, G., Carmignani, L., Conti, P. and Aiken.L.V., (2002). Active fragmentation of Adria based on Global Positioning System velocities and regional seismicity. Geology, 30, 779–782.

Salvi S., S. Stramondo, M. Cocco, E. Sansosti, I. Hunstad, M. Anzidei, P. Briole, P. Baldi, M. Tesauro, E. Lanari, F. Doumaz, A. Galvani, A. Pesci (2000). Modelling coseismic displacement resulting from sar interferometry and gps measurements during the 1997 umbria – marche seismic sequence. Journal of Seismology,4, 479-499.

Selvaggi G. , Mattia M. , Avallone A. , D'Agostino N. , Abruzzese L. , Anzidei M. , Cantarero M. , Cardinale V. , Castagnozzi A. , Casula G. , Cerere G. , Cogliano R. , Criscuoli F. , D'Ambrosio C. , D'Anastasio E. , De Martino P. , Del Mese S. , De Luca G. , Devoti R. , et al., (2006). La Rete Integrata Nazionale GPS (RING) dell' INGV: una infrastruttura aperta per la ricerca scientifica. Atti della X Conferenza Nazionale dell’ASITA, Bolzano 14-17 novembre 2006.

Serpelloni E., Anzidei M., Baldi P., Casula G., Galvani A., Pesci A. & Riguzzi F., (2002). Combination of permanent and non-permanent GPS networks for the evaluation of the strain-rate field in the central Mediterranean area. Bollettino di Geofisica Teorica ed Applicata, Vol.43, n.3-4, 195-219.

Serpelloni, E., Anzidei, M., Baldi, P., Casula, G. and Galvani, A., (2005). Crustal velocity and strain-rate fields in Italy and surrounding regions: new results from the analysis of permanent and non-permanent GPS networks. Geophysical Journal International 161(3), 861-880. doi:10.1111/j.1365-246X.2005.02618.

Serpelloni, E., Anzidei, M., Baldi, P., Casula, G. and Galvani, A., (2006). GPS measurements of active strains across the Apennines. Ann. Geoph., 49, n.1, 319-329.

Serpelloni, E., Vannucci, G., Pondrelli, S., Argnani, A., Casula, G., Anzidei, M., Baldi, P., Gasperini, P., (2007). Kinematics of the Western Africa - Eurasia Plate Boundary From Focal Mechanisms and GPS Dat., Geophys. J. Int. doi: 10.1111/j.1365-246X.2007.03367.x.

Smith D.E. Koenkiewicz R., Nerem R.S., Dunn P.J., Torrence M.H., Robbins J.W., Klosko S.M., Williamson R.G., Pavlis E.C., (1994). Contemporary global horizontal motion. Geophys. Jou. Int., 119, 511-520.

Stramondo S., Tesauro M., Briole P., Sansosti E., Salvi S., Lanari R., Anzidei M., Baldi P., Fornaro G., Avallone A., Buongiorno M.F., Franceschetti G., Boschi E., (1999). The September 26,1997 Central Italy earthquakes: coseismic surface displacement detected by sar interferometry and GPS, and fault modeling. Geophysical Research Letters, vol.26, n.7, pp.883-886 April, 1.

Surace L., (1993). Il progetto IGM95. Boll. Geod. Sci. Aff., 3, 220-230. Surace L., (1997). La nuova rete geodetica nazionale IGM95: risultati e prospettive di utilizzazione. Boll.

Geod. Sci. Aff., 3, 335-377. Tosi P., Tertulliani A., De Rubeis V. and Gasparini C., (1999). Preliminary results of a macroseismic

survey of the Colfiorito sequence (Central Italy). Phys. Chem. Earth., 24, 477-481. Tsuji H., Hatanaka Y., Sagiva T., Hashimoto M., (1995). Coseismic crustal deformation from the 1994

Hokkaido-Toho-Oki earthquake monitored by a nationwide continuous GPS array in Japan. Geoph. Res. Lett. 22, No.13, 1669-1672.

Valensise G. and Pantosti P., (2001b). Database of potential sources for earthquakes larger than M=5.5 in Italy. Ann. Geophys., 44 (supp. To n.4), pp.180 with CD-Rom.

Vespe F., Bianco G., Fermi M., Ferraro C., Nardi A., Sciarretta C., (2000). The Italian GPS fiducial network: services and products. Jou. of Geodyn., 30, 327-336.

Wells D. L. and Coppersmith K.J., (1994). New empirical relationships among Magnitude, rupture length, rupture width, rupture area, and surface displacement. Bull. Seism. Soc. of Am., Vol.84, n.4, 974-1002.

Zhang J., Bock Y., Johnson H., Fang P., Williams S., Genrich J., Wdowinski S. and Behr J., (1997). Southern California Permanent GPS Geodetic array: error analysis of daily positions estimates and site velocities. Jou. Geoph. Res., 102, n. B8, 18035-18055.

49

SAR Interferometry for the analysis of crustal deformation in Italy: from the Umbria-Marche co-seismic displacement field to mapping the strain accumulation on active faults

Stefano Salvi Istituto Nazionale di Geofisica e Vulcanologia

Differential SAR Interferometry (DInSAR), is a radar signal processing technique which uses two Synthetic Aperture Radar (SAR) images, to obtain a map of the ground deformation occurred between the image acquisition dates [Massonnet and Feigl, 1998]. If any deformation has occurred, a detailed picture of the ground displacement can be obtained, with spatial resolution up to few tens of meters and measurement accuracies up to few mm.

The DInSAR technique was first applied to earthquake analysis in 1993, to produce a spectacular map of the permanent ground displacement occurred during the 1992 Landers earthquake, California [Massonnet et al., 1993]. Since then, DInSAR has been extensively applied to the measurement of co-seismic deformation, often providing the only geodetic data to gain information on the dislocation geometry and kinematic. The most evident advantages of DInSAR over classical geodetic measurements and GPS, are the coverage of large areas at high resolution, the independence from established ground networks, which enhances usage flexibility, and the small cost of surveys. On the other hand it is impossible to use DInSAR where the ground is not ‘stable’ in the interferometric sense, e.g. in farmlands, densely vegetated or snow-covered areas.

In Italy the DInSAR technique was applied for the first time at INGV for the analysis of the co-seismic displacement of the Umbria Marche sequence, 1997 [Stramondo et al., 1999]. From ERS satellite data, several differential interferograms were calculated and clearly showed the pattern and magnitude of the surface ground displacement caused by the fault dislocation at depth (Figure 1).

The displacement maps reported a maximum ground subsidence of 28 cm in the Annifo plain and 14 cm in the Colfiorito plain, completely unnoticed in the field. The fringe pattern did not show the presence of evident surface ruptures. Along the SW margin of Mt. Pennino a high displacement gradient was observed, suggesting the possible presence of a surface rupture, but none was found on the ground.

Salvi et al. [2000] carried out a detailed elastic modeling of the displacement data, and were able to identify the dislocations responsible for the two Colfiorito earthquakes of September 26th and for the Sellano earthquake of October 14th. The models fitted most of the observed deformation, although there remained some minor patterns in the fringe distribution, not explained by the elastic dislocation. Seven years after, some of these patterns have been re-evaluated as being due to large gravitational mass movements triggered or reactivated by the seismic shaking [Moro et al., 2007].

The lessons learnt from the successful Umbria-Marche DInSAR analysis were many, and prompted further development of the technique both at INGV and at other italian research institutions. At INGV we started to invest more resources in the modeling procedures and post-processing of the results, and we also began to design an integrated data collection-analysis system able to generate the co-seismic displacement maps so rapidly to fulfill the emergency management needs (few days after the mainshock).

The fallout of this line of research is the present SIGRIS project, funded by ASI, in which a prototype of such a system will be developed in support of the Italian Civil Protection Department, under the scientific responsibility of the INGV Remote Sensing Laboratory.

From the 1990s to the present, the DInSAR technique evolved considerably, the most important advances being in the capability of measuring ground velocities with very high precisions (1 mm/yr and less), opening a wide range of new applications in tectonics and seismology. Two italian SAR research groups were at the forefront of this research, called multitemporal InSAR since it requires the availability of several tens of SAR images acquired over the same area.

The Permanent Scatterers technique was the first to be developed, by the Politecnico di Milano-DEI Department, and is now commercially exploited by a POLIMI spin-off company: Tele Rilevamento Europa srl (www.treuropa.com). Shortly after came the Small BASeline (SBAS) technique, developed by

50

IREA-CNR, Napoli (www.irea.cnr.it) in the framework of two ASI projects headed by INGV for the measurement of slow crustal deformation in Italy. The two techniques are now essentially mature, and several studies have shown their ability to measure ground velocities with accuracies up to 0.5 mm/yr over tens to hundreds of points for square kilometre [Ferretti et al., 2007; Casu et al. 2006].

Figure 1. Three of the co-seismic ERS differential interferograms showing the ground displacement occurred during the largest earthquakes of the Umbria Marche sequence.

INGV researchers are using these two and other multitemporal InSAR techniques developed worldwide [IPTA, www.gamma-rs.ch; ESD, Fornaro et al., 2007; etc.], for the study of crustal deformation at local and regional scale. The ground velocities measured at the Mt. Etna, Campi Flegrei-Vesuvius, and Colli Albani volcanic areas, provided important insights into the sources of volcano deformation [Lanari et al., 2004; Salvi et al., 2004]. In the urban areas of Rome, Guidonia and Bologna, cm/yr ground subsidence was measured at various scales and interpreted as due to different phenomena [Stramondo et al., 2007]. In the seismically active areas of Central Abruzzi and the Gargano Promontory, the modeling of the regional ground velocities allowed to investigate the mechanisms by which the tectonic strain accumulates along the active faults [Atzori et al, 2007]. A detailed analysis of local ground deformation in several areas of the Apennines, allowed to detect large, actively deforming gravitational mass movements whose catastrophic collapse hazard might be strictly related to the local seismic hazard [Stramondo et al., 2005].

All of these studies were made possible by the availability of a large archive of ERS SAR images, dating back to the 1992. The long-sighted policy of the ESA, which provided continued acquisition and archival of ERS images over the most tectonically active areas of the world, was the foundation over which the success of SAR interferometry was built. To fully exploit the large information content contained in the 1992-2000 ERS archive, INGV has started in 2007 a scientific initiative for the generation of a high resolution map of the ground velocity over the seismogenic areas of Italy: the VELISAR initiative (Figure 2). An important goal of such initiative is to disseminate the information on ground deformation retrieved by InSAR techniques, to the scientific community and to the public in general. This is accomplished through a dedicated web site: http://kharita.rm.ingv.it/Gmaps/vel/index.htm#homevel.

51

While the ERS mission demonstrated the importance of dense temporal acquisitions of SAR images, this may not be so simple to obtain with the new SAR missions like ENVISAT, ALOS, TerraSar X or the italian Cosmo Skymed. Their capability to provide several different modes of acquisition, although useful to favour new applications in different fields, is not the best choice for building the continuous archive needed by the multitemporal InSAR techniques. Indeed, the geophysical community worldwide is now aware that the full exploitation of the great potential of InSAR multitemporal techniques for the monitoring of crustal deformation, now requires a long term, dedicated SAR mission.

References Atzori S., I. Hunstad, C. Tolomei, S. Salvi , A.Ferretti, S. Cespa, (2007), Interseismic strain accumulation

in the Gargano Promontory, Central Italy, Geophysical Research Abstracts, 9, 7651, EGU Burgmann R., Hilley G., Ferretti A., and Novali F., (2006), Resolving vertical tectonics in the San

Francisco Bay Area from permanent scatterer InSAR and GPS analysis, Geology, 34, 3, 221, 224 Casu F., M. Manzo, R. Lanari, (2006), A quantitative assessment of the SBAS algorithm performance for

surface deformation retrieval from DInSAR data, Rem. Sens. of Environment, 102, 195-210 Ferretti A., G. Savio, R. Barzaghi, A. Borghi, S. Musazzi, F. Novali, C. Prati, and F. Rocca, (2007),

Submillimeter Accuracy of InSAR Time Series: Experimental Validation, IEEE Transactions on Geoscience and Remote Sensing, 45, 5, 1142-1153

Fornaro G., A. Pauciullo, and F. Serafino, (2007), Enhanced Spatial Differences (ESD): a new technique for DInSAR monitoring of deformation over wide areas, Geophysical Research Abstracts, 9, 10814, EGU

Lanari R., Berardino P., Borgstrom S., Del Gaudio C., De Martino P., Fornaro F., Guarino S., Ricciardi G.P., Sansosti E., Lundgren P., (2004), The use of IFSAR and classical geodetic techniques for caldera unrest episodes: application to the Campi Flegrei uplift event of 2000, J. of Volcanol. and Geoth. Res., 133, 247-280

Massonnet, D., Rossi, M., Carmona, C., Adragna, F., Peltzer, G., Feigl, K., Rabaute, T., (1993), The displacement field of the Landers earthquake mapped by radar interferometry, Nature, 364, 138-142.

Massonnet, D., and Feigl, K., (1998), Radar interferometry and its applications to the changes in the Earth's surface, Reviews of Geophysics, 36, 441-500.

Moro M., Saroli M., Salvi S., Stramondo S., Doumaz F., (2007), The relationship between seismic deformation and deep-seated gravitational movements during the 1997 Umbria-Marche (Central Italy) earthquakes, Geomorphology, 89, 297-307

Salvi S, Stramondo S., Cocco M., Tesauro M., Hunstad I., Anzidei M., Briole P., Baldi P., Sansosti E., Lanari R., Doumaz F., Pesci A. and A. Galvani, (2000), Modeling Coseismic Displacements resulting from SAR Interferometry and GPS measurements during the 1997 Umbria-Marche seismic sequence, Journal of Seismology, 4, 479-499

Salvi, S. Atzori, C. Tolomei , J. Allievi, A. Ferretti, F. Rocca, C. Prati, S. Stramondo, N. Feuillet , (2004), Inflation rate of the Colli Albani volcanic complex retrieved by the Permanent Scatterers SAR interferometry technique, Geophysical Research Letters, 31

S. Stramondo, M. Saroli, C. Tolomei, M. Moro, F. Doumaz, A. Pesci, F. Loddo, P. Baldi, E. Boschi, (2007), The subsidence in Bologna city (Italy) and in the Po Plain detected by InSAR SBAS technique, Geophysical Research Abstracts, 9, 7192, EGU

Stramondo S., Tesauro M., Briole P. Sansosti E., Salvi S., Lanari R., Anzidei M., Baldi P., Fornaro G., Avallone A., Buongiorno M.F., Franceschetti G., Boschi E., (1999), The September 26, 1997 Colfiorito, Italy, earthquakes: modeled coseismic surface displacement from SAR Interferometry and GPS, Geophys. Res. Lett., 26, 7, 883-886.

Stramondo S., M. Saroli, M. Moro, S. Atzori, C. Tolomei, S. Salvi and Riccardo Lanari, (2005), Monitoring long-term ground movements and Deep Seated Gravitational Slope Deformations by InSAR time series: cases studies in Italy, Proc. FRINGE 2005, ESA-ESRIN, Frascati, Rome, available at: http://earth.esa.int/fringe2005/proceedings/papers/814_stramondo.pdf

52

Figure 2. The web interface of the VELISAR initiative.

53

Extended sources of the main events of the Umbria-Marche (1997) seismic sequence inverted from geophysical data

Maria Elina Belardinelli Università di Bologna

Introduction The three largest shocks of the Umbria-Marche sequence occurred on September 26, 1997 at 00:33

(Event 1, MW=5.7) and 09:40 (Event 2, MW=6.0) in the Colfiorito area and on the October 14, 1997 at 15:23 (Event 3,MW=5.6) in the Sellano area (figure 1).

Figure 1. The epicentres (red stars) of three largest events in the Umbria-Marche (1997-1998) sequence.

Several studies of the sources of events 1-3 were enabled by the availability of geodetic and seismological data, such as coseismic permanent displacements recorded by GPS stations [e.g Hunstad et al., 1999; Lundgren and Stramondo, 2002; Belardinelli et al., 2003; Hernandez et al., 2004], DInSAR image pairs [e.g. Lundgren and Stramondo, 2002; Belardinelli et al., 2003; Santini et al., 2004; Hernandez et al., 2004; Crippa et al., 2006] repeated measures along a leveling line [e.g. De Martini et al., 2003], waveforms recorded by strong-motion accelerometers [e.g. Capuano et al., 2000; Hernandez et al., 2004] and by broad-band seismic stations at regional distance (less than 500 km) from the epicentral area [e.g. Pino and Mazza, 2000]. It is to be noted that no relevant coseismic deformation at GPS sites was observed for event 3 due to the sparse configuration of the GPS network in the Sellano area.

I briefly review these studies with regard to their approach and the inferred properties of the seismic source of events 1-3.

54

Methods and uncertainties The above mentioned studies estimated (or inverted) the vector of parameters m characterizing the

seismic source by minimizing a misfit or cost function which, following the notation of Belardinelli et al. [2003], can be expressed as

)()( mod1Tmod ddCdd −−= −ε (1) where C is the covariance matrix, whose elements depend on data uncertainties, d is the vector of data used and dmod is the vector containing the estimates of data. The latter depends on m through a model equation of the following kind

);( 0mod mmgd = (2)

where m0 is a vector of parameters fixed or assumed as known in the study. Generally these studies assumed the seismic sources to be described by dislocations in an elastic homogeneous half-space with rigidity values about 30 GPa, except for those studies considering strong-motion data [e.g. Capuano et al., 2000; Hernandez et al., 2004] where a layered half-space was considered with rigidity values smaller than 30 GPa in the first 7-8 km depth. Apart from Lundgren and Stramondo [2002] who inverted also for the direction of the slip vector in different points of the extended sources, most of the above mentioned studies assumed a homogenous rake angle and determined the slip amplitude distribution over the fault plane. The latter is constrained by the scalar seismic moment of each seismic event, a reference value of which was given by the moment tensor analysis of long period seismological data from the MedNet network [Ekström et al., 1998] (CMT solution hereinafter).

The studies of the extended sources of events 1-3 using only one kind of data were generally checked to be consistent with other observational evidences or inferences made in previous studies. For instance Santini et al. [2004] verified the smallness of the misfit with GPS data and a DInSAR interferogram not used to invert the slip distribution of events 1-3. In order to improve the resolution of the inversion obtained with a single data set, a joint inversion of different kinds of data can be performed. In a joint inversion the cost function is obtained by assigning a different weight to the different data (all stored in only one vector d) by means of weight diagonal matrix W

)()( mod1Tmod ddWCdd −−= −ε (3) For event 2, both DInSAR and GPS data were used by Belardinelli et al. [2003] to determine the

fault mechanism, the location of the hypocenter on the fault plane, the depth of the fault and a simplified slip distribution, largely inspired to previous findings [Hunstad et al., 1999]. Analogous joint inversion was performed by Hernandez et al. [2004] to infer the slip distribution of events 1-2. Lundgren and Stramondo [2002] made a joint inversion of GPS data, different DInSAR images and the total seismic moment (with respect to the CMT solution) to infer the slip vector distribution and the seismic moment of events 1-3. In these studies the weight matrix W in (3) was chosen in such a way that the weight of a GPS datum is about 5-10 times larger than the weight of a DInSAR datum (a point of the set used to discretize a DInSAR fringe of equal slant range displacement). In general, in order to limit the parameter space explored during the inversion, all the a priori constraints on the inverted parameters should be used, mainly in cases when the link between the data and the parameters (g in equation 2) is non linear and the inversion can be not univocal. To this aim, both Belardinelli et al. [2003] and Santini et al. [2004] determined a slip distribution fulfilling the constraint of a scalar seismic moment equal to the CMT solution.

The estimates of the inverted parameters are affected by uncertainties that can be related to those affecting the data used or the parameters fixed in the inversion, but are also intrinsic to the inversion method, taking into account the limited resolution of the data. An uncertainty of 10-20 % in the value of the inferred parameters was estimated by Capuano et al. [2000] and by Pino and Mazza [2000] by perturbing the parameters m around the best fit solution and verifying to obtain a vector dmod compatible with the reading errors of data. Similarly, Belardinelli et al. [2003] provided estimates of the uncertainty of each of the inferred parameters for event 2 by analyzing the distribution of inverted parameters reproducing on average the data within the measurement error. For events 1-2, Crippa et al. [2006] showed that the resolution of geodetic data gets worse with increasing the depth. In particular the details of the seismic source of event 1-2 inverted from geodetic data are scarcely resolved at depths larger than 3 km [Hernandez et al., 2004].

55

About the three seismic sources All these studies recognized the three sources as nearly NW-SE trending and SW dipping faults,

with a dominant normal component of slip [e.g. Capuano et al., 2000; Lundgren and Stramondo, 2002; Belardinelli et al., 2003] in agreement with the CMT solution. However the values assumed or inferred (by trial and error) for the strike of event 3 belong to a rather large interval 125°-160° [Santini et al., 2004; Lundgren and Stramondo, 2002; Capuano et al., 2000].

The hypocenter of events 2 (3) is located near the SE (NW) border of the fault. The rupture of events 2-3 propagated from the hypocenters with an average rupture velocity of about 2.6-3 km/s [Capuano et al., 2000; Pino and Mazza, 2000]. In most of the studies, the hypocenter of Event 1 is assumed to be located near the NW border of the fault suggesting a unilateral rupture; however Capuano et al. [2000] attributed a bilateral rupture to event 1. The source of event 2 has a fault length of about 12-15 km [e.g. Capuano et al., 2000; Pino and Mazza, 2000] and a small left-lateral component of slip [Lundgren and Stramondo, 2002, Belardinelli et al. 2003] in agreement with the CMT solution. Even if less resolved by GPS and DInSAR data, the inferred value of the dip angle of event 2 is anomalously shallow for a normal fault, around 40 degrees [e.g. Belardinelli et al., 2003; Hernandez et al., 2004; De Martini et al. 2003] in agreement with the CMT solution.

The inverted slip distributions for events 1-3 are heterogeneous and characterized by a patch of high slip localized near the hypocenter. For event 2, this patch has about 5 km extent along strike and it concentrates below about 2 km depth [Lundgren and Stramondo, 2002; Belardinelli et al., 2003; Santini et al., 2004; Hernandez et al. 2004; Crippa et al., 2006]. For event 2, another smaller patch of slip, located close to NW end of the fault, can be inferred [Pino and Mazza, 2000; Lundgren and Stramondo, 2002; Hernandez et al., 2004; De Martini et al., 2003; Crippa et al., 2006]. Outside and around the two high-slip patches, the distribution of aftershocks recorded in the first three months after event 2 shows a larger density [De Martini et al., 2003] suggesting the presence of an asperity.

For event 2, GPS data require the top edge of the fault to be shallower (0.04-0.7 km) [e.g. Lundgren and Stramondo, 2002; Hernandez et al. 2004] than using levelling data (2.8-4 km) [De Martini et al., 2003], DInSAR data (0.5-3 km) [e.g. Crippa et al., 2006; Lundgren and Stramondo, 2002] or strong-motion data (0.8-3.4 km) [Hernandez et al., 2004; Capuano et al., 2000]. On the average the top of the fault of event 2 is found to be shallower than that of events 1 and 3. If the layering of the crust is taken into account [Capuano et al., 2000; Hernandez et al., 2004], the top edges of the three faults tend to be deeper (2.5-3.5 km , 0.8-3.4 km, 2.4-4.8 km for events 1,2,3 respectively) than when half-space models are used (1-2 km, 0.04-3.4 km, 0.7-1.2 km, for events 1,2,3 respectively) [e.g. Lundgren and Stramondo, 2002; De Martini et al., 2003; Santini et al., 2004; Crippa et al., 2006], as expected from theoretical studies of dislocations in media with low rigidity layers near the surface [Savage, 1998].

References Belardinelli, M. E., Sandri, L. and Baldi, P., (2003). The major event of the 1997 Umbria-Marche (Italy)

sequence; what could we learn from DInSAR and GPS data? Geophys. J. Int., 153, 1, 242-252. Capuano, P., Zollo, A., Emolo, A., Marcucci S. and Milana, G., (2000). Rupture mechanism and source

parameters of Umbria-Marche mainshocks from strong motion data. J. Seismol., 4, 463–478. Crippa, B, Crosetto, M, Biescas, E, Troise, C., Pingue, F. and De Natale, G.,(2006). An advanced slip

model for the Umbria-Marche earthquake sequence: coseismic displacements observed by SAR interferometry and model inversion. Geophys. J. Int., 164 (1), 36-45.

De Martini, P. M., Pino, N. A., Valensise G. and Mazza, S. ,(2003). Geodetic and seismologic evidence for slip variability along a blind normal fault in the Umbria-Marche 1997-1998 earthquakes (central Italy). Geophys. J. Int., 155, 819-829.

Ekström, G., Morelli, A., Boschi, E. and Dziewonski, A. M., (1998). Moment tensor analysys of the Umbria-Marche earthquake sequence of September-October 1997. Geophys. Res. Lett., 25, 11, 1971-1974.

Hernandez, B., Cocco M., Cotton, F., Stramondo S., Scotti, O., Courboulex, F. and 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 Geophys., 47 (4), 1355-1376.

Hunstad, I., Anzidei, M., Cocco, M., Baldi, P., Galvani, A. and Pesci, A., (1999). Modelling coseismic displacements during the 1997 Umbria-Marche earthquake (central Italy). Geophys. J. Int., 139, 2, 283-295.

56

Lundgren, P. and Stramondo, S., (2002). Slip distribution of the 1997 Umbria-Marche earthquake sequence: Joint inversion of GPS and synthetic aperture radar interferometry data, J. Geophys. Res., 107(B11), 2316, doi:10.1029/2000JB000103.

Pino, N.A. and Mazza, S., (2000). The Umbria-Marche (central Italy) earthquakes: relation between rupture directivity and sequence evolution for the Mw >5 shocks. J. Seismol., 4, 451–461.

Santini, S., Baldi, P. Dragoni, M., Piombo, A., Salvi, S., Spada, G. and Stramondo, S., (2004). Monte Carlo Inversion of DInSAR Data for Dislocation Modeling: Application to the 1997 Umbria-Marche Seismic Sequence (Central Italy). Pure Appl. Geophys., 161, 817–838.

Savage, J.C., (1998). Displacement field for an edge dislocation in a layered half-space. J. Geophys. Res., 103(B2), 2439-2446.

57

Macroseismic survey of the 1997-1998 Umbria-Marche earthquakes: from practice to practice

Romano Camassi Isituto Nazionale di Geofisica e Vulcanologia

The seismic sequence of the Umbria-Marche Appenines started on September 4, 1997, at 00:07 AM (local time) with an earthquake located in the area of Colfiorito, which caused slight damage to a few villages of the Serravalle di Chienti (MC) municipality and was followed by some minor shocks over the next few days. Then, on September 26, a much stronger earthquake hit the same area at 00:33 AM (GMT), causing high damage to villages in the Serravalle di Chienti municipality and nearby; at 09:40 AM still another strong shock occurred a few kilometres north-west of the former, worsening the damage in Serravalle di Chienti, causing high damage in Foligno, Nocera Umbra and Fabriano, and lesser damage in many more Umbrian and Marchesan sites.

On October 3, there was a strong earthquake near Nocera Umbra, a little to the north of the previous ones and over the next few days several strong shocks occurred at various locations in a section of Apennines that runs between Sellano-Norcia (south) and Nocera Umbra-Gualdo Tadino (north). On October 6 there was one in the Casenove-Forcatura area, a few kilometres west of Colfiorito. On October 12, and again on October 14, it was the turn of the Sellano-Preci neighbourhood to suffer significant damage. From the end of October the events started to tail off and over the next few months the sequence would go on fitfully, but from September 4 to the end of October, the National Seismic Network had already recorded some 2,500 seismic events.

Figure 1. The collapses in the church in Annifo (29 September 1997).

On 26 March, 1998, there was a new strong shock in the Nocera Umbra-Gualdo Tadino area: with an hypocentral depth estimated at around 50 kilometres, it was felt over a large area and caused slight damage to a few Marchesan towns (Cingoli, Macerata, Tolentino). On April 3, the same area was affected again by a strong shock that worsened extant damage in the Nocera Umbra-Gualdo Tadino area and was felt over most of Umbria and the Marches. In the following months the seismic activity gradually decreased, apart from a few episodes which were particularly felt by the local population.

The Survey The macroseismic effects of the September 4 earthquake were routinely monitored by the GNDT

Research Unit of the Geophysical Observatory of Macerata, on behalf of the regional administration of the Marches. After the sequence’s main earthquakes, from September 26 onwards, a task-force of researchers started a macroseismic survey meant to effect a quick evaluation of the maximum effects and to define the areas of damage, on behalf of the national Department of Civil Protection (DPC). The involved researchers

58

belonged to the National Group for Protection against Earthquakes (GNDT), the National Institute of Geophysics (ING), and the National Seismic Service (SSN).

Though operating in different ways (GNDT and SSN jointly, ING on its own), the involved teams were well aware of the need for cooperation: they set up a common coordinating unit and a centre for the exchange of information. Thus, working against the clock and in forbidding conditions, they carried out the tremendous task of providing the DPC and the scientific community with an enormous set of data in a very short time. It is to be noted that this was the first occasion in which the Internet was massively and efficaciously used by Italian researchers as a fast communication and data-storage tool, to the huge acclaim of all concerned parties (http://emidius.mi.ingv.it/GNDT/T19970926/home.html).

From its very start the survey campaign proved extremely taxing and complex, both on account of the huge extension of the area that must be covered (the entire territories of the Regions Umbria and Marches), and of the still ongoing seismic activity. Each new strong shocks caused the level and distribution of damage to alter, and the survey had to be partially repeated, to cover the whole high and medium damage area and large stretches of the lesser damage area. A vast amount of photographic documentation was collected, together with local evidence derived from interviews to expert witnesses (firemen, municipal officers, surveyors, etc.). The results of the macroseismic survey were integrated with those of the Macroseismic Questionnaire routinely distributed by ING in the large area affected by the main earthquakes. The resulting data did mainly contribute to the definition of the effects in non-damaged areas.

The evaluation of effects was carried out according to the macroseismic Mercalli-Cancani-Sieberg (MCS) scale. Once the first emergency was over, a trial of the new European Macroseismic Scale (EMS-98) [Grünthal, 1998] was also carried out. This was of great interest as a text of the new scale, whose special requirements (detailed information on building typologies to be found in each locality, and survey of damage distribution for each building typology) would have been hard to reconcile with those of a speedy survey [Stucchi, 1997].

Figure 2. Damages in Sellano (October 1997).

The currently available total estimates of the macroseismic effects of the 1997 sequence are mainly based on the results of the specific survey in the area [Camassi et al., 1997; Gasparini et al., 1997; Molin et al., 1997; Monachesi et al., 1997; Stucchi, 1997; Camassi et al., 1998, Camassi, 2002], on the critical analyses of the daily survey diaries, and on the analyses of the buildings safety assessment data.

The Damage The serious and diffuse damage area was mainly restricted to the Provinces of Perugia and Macerata,

mostly in the municipalities of Foligno (PG), Nocera Umbra (PG), Sellano (PG), Serravalle di Chienti (MC) and Visso (MC), and the nearby municipalities of Cerreto di Spoleto (PG), Fiuminata (MC), Monte Cavallo, (MC) Preci (PG) and Valtopina (PG). Moderate damage occurred in many areas distributed throughout the Regions of Marche and Umbria, with isolated episodes also outside these regions.

The most serious damage, with many partial and total collapses, occurred in limited mountain stretches of Serravalle di Chienti, Foligno, Nocera Umbra and Sellano. The involved settlements were mainly made up of vulnerable building typologies: stone and masonry buildings, two or three floors high, usually with wooden floors that were not anchored to the walls; in some cases such buildings were in bad repair or only occasionally lived-in.

In the more southern areas involved in the damage (mainly in the area of Sellano), various buildings had been undergoing modifications for seismic improvement, particularly after the earthquake of Valnerina in 1979. Such modifications mainly involved substituting wooden floors and roofs with cemented

59

materials. In many cases, such improvements were not accompanied by measures to reinforce the walls and the connections between the vertical and horizontal structures. In these situations, there was particularly serious damage, caused mainly by the excessive overloading of the structures, which was not compensated for by improvements in the walls (a typical example is the serious damage to the church in Annifo).

In this area there were also to a lesser extent buildings in reinforced concrete, both of recent construction (and therefore built according to the seismic regulations) and of less recent construction. In general, these buildings resisted the earthquakes well.

There were many situations in which there were large irregularities in the distribution of damage (examples being the cases of Cesi Alto and Cesi Basso, and of Forcella and Collecurti); these irregularities can certainly be put down to individual situations relating to the vulnerability of the buildings involved. However, there were numerous cases in which the variability of the damage was put down to the various local geological and geomorphic conditions. In some cases, there was evident colocalisation of subsidence phenomena, originating from pre-existing damage or due to the amplification of the effects of the earthquakes (as seen by the cases of Forfi and the various built-up areas that make up Annifo).

Lerning from Practice These earthquakes, the most important seismic event since those of Irpinia in 1980, represented

dramatic circumstances for the population involved, and together provided a unique occasion for the Italian scientific community and for the national civil protection system for the measuring of their respective abilities for the understanding and management of the event.

Figure 3. Damages in Arvello (30 September 1997).

In this sense, the event put the abilities of the scientific community under heavy testing, to efficiently and timely monitor the impact of the earthquakes on the area. The long seismic period found a scientific community in strong transformation, and not ready, perhaps, to confront such an important emergency. It is enough to remember, to take an example, that even now there does not exist a unique structure with the specific responsibility for the monitoring of damage, with adequate resources and organisation. Nevertheless, over a period of a few hours, it was the operators themselves who constructed the coordination and invested all of the available energies in providing an invaluable service within the national community.

As has already been discussed, the monitoring of the damage was extremely problematic. First of all, the general impossibility of distinguishing the effects of the single earthquakes within the main sequence that was activated on 26 September needs to be stressed. Between 26 September, 1997, and the end of October, there were 27 earthquakes with magnitude greater than 4.0, while there were a few thousand events with lower magnitude; these are what determined the continual variations in the effects in the individual areas. Many areas were affected many times, without it always being possible to clearly distinguish the increments in the damage.

60

From Practice to Practice Many problems, and therefore extremely complex, but this situation has allowed the Italian

seismological community to learn many things. Also macroseismology (including historical seismology) has knowingly confronted important methodological problems, from which it should have learnt in these past ten years (and from which maybe it has not).

Those who are involved in macroseismology are confronted, and sometimes with difficulties, by the procedures for the assigning of macroseismic intensity, on the use of the scale. This critical point is buried in its own complexity both due to the presence of great differences in damage over short distances, and for the impossibility of distinguishing the effects of earthquakes that follow closely in both space and time, within such a complex sequence. In the practice of today, nevertheless, no lesson can consider itself totally learnt, either in relation to recent earthquakes or those historical.

In 1997, for the first time, there was an experimental laboratory for the EMS-98 scale that has given very interesting results, and from that moment, although still with some difficulties, the new European macroseismic scale has become used ever more frequently also in Italy. Only the use of this scale appears to make it possible to reduce the heterogeneity of the processes.

From an organisational point of view, in a more general framework of the reorganisation of the structure responsible for protection from earthquakes, a few advances have been made. The QUEST working group has been active for a few years, which is actually nothing more than the formalisation of that mobilisation of researchers and technicians that was activated on a voluntary basis those ten years ago.

References Camassi, R. (2002). La crisi simica del 1997. I dati del rilievo macrosismico del terremoto Umbria-Marche

1997. In: Il patrimonio culturale dall'emergenza sismica del 1997 al piano di ripristino recupero e restauro. il caso delle Marche (Canti M. and Polichetti M.L., eds.), pp. 31-38.

Camassi, R., Cova, E., Ercolani, E., Monachesi, G. and Postpischl, L., (1997). Earthquakes of September and October 1997 in Umbria-Marche (Central Italy): Virtual exploration of the epicentral zone. Internet: http://emidius.mi.ingv.it/GNDT/UMEQ_VR/index_i.html

Camassi, R., Galli, P., Molin, D., Monachesi, G. and Morelli, G. (eds.), (1998). Rilievo macrosismico preliminare del terremoto umbro-marchigiano di settembre-ottobre 1997. Ingegneria Sismica, XIV, 3, 22-26.

Gasparini, C., Anzidei, M., Conte, S., Del Mese, S., De Rubeis, V., Maramai, A., Massucci, A., Riguzzi, F., Tertulliani, A., Vannucci, C. and Vecchi, M., (1997). Indagine macrosismica per la sequenza sismica Umbro-Marchigiana settembre - ottobre 1997. In: Studi preliminari sulla sequenza sismica dell’Appennino Umbro-Marchigiano del settembre-ottobre 1997 (Boschi E and Cocco M., eds.), Pubbl. ING n. 593, 42-44.

Grünthal, G. (ed.), (1998). European Macroseismic Scale 1998 (EMS-98). European Seismological Commission, Working Group Macroseismic Scales. Luxembourg.

Molin, D., Soddu, P., Gasparini, C., Tertulliani, A., Camassi, R., Monachesi, G. and Stucchi, M., (1997). Unificazione delle valutazioni degli effetti macrosismici dei terremoti umbro-marchigiani del settembre/ottobre 97. Rapporto Protezione Civile.

Monachesi, G., Camassi, R. and Molin, D. (eds.), (1997). Earthquakes of September-October 1997 in Umbria-Marche (Central Italy): Macroseismic survey after the earthquakes of 26 September 1997 and following days.

Internet: http://emidius.mi.ingv.it/GNDT/T19970926_eng/T19970926agg.html Stucchi, M. (ed.) (1997). Earthquakes of September-October 1997 in Umbria-Marche (Central Italy):

Macroseismic data in terms of EMS-92 scale. Internet: http://emidius.mi.ingv.it/GNDT/T19970926_eng/EMS_intro.html

61

Codes, models, and reality: reductionism vs. holism in 10 years of microzonation studies

in the Umbria-Marche region.

Marco Mucciarelli Università della Basilicata

In the 10 years elapsed from the Umbria-Marche earthquake, several microzonation studies were carried out in the two regions. In the immediate aftermath of the event, the focus was on the epicentral area and toward emergency intervention and reconstruction plans. In the following years, regional and national projects were aimed to transfer the lessons learned to other towns (among the others Spoleto, Cerreto di Spoleto, Città di Castello, Gubbio, Fabriano, Cagli, Treia, Serra de' Conti, Offida, Senigallia).

Usually, those two kinds of microzonation studies are referred to as “simplified” and “detailed”. The difference is more subtle, and leads to the question if a microzonation study can be tackled

following a reductionist approach, i.e. leaving different experts taking care of a limited subject (geology, geophysics, seismology, geotechnics, structural engineering). The impression looking back to 10 years of studies is that an holistic approach would be more appropriate to describe a system (structure-soil-bedrock) that is non-linear, inhomogeneous, and presenting feedback among its components.

A second problem that emerges is the link with codes and practitioners. During the past 10 years the seismic code has been changed and a new version is on arrival. A leitmotiv often repeated is that the codes must be simple otherwise the practitioners could have problem in applying them. One may wonder if any of us would oppose the diffusion of medical ultrasonography or Computerised Tomography Scan on the basis that old-fashioned doctors are familiar just with their wooden stethoscopes. A wrong diagnosis is never good, not only if the illness is worst than assumed, but also in case of over-conservatism, causing unnecessary expenses and re-routing limited resources where they are not needed. Moreover, the last proposed version of the code is based on a parameter (Vs30) that is discussed in the same country where it was first adopted and that proved to be a poor proxy of seismic amplification, and introduces a parameter (acclivity) that appears to be a second- or third order problem with respect to others that are completely disregarded (e.g., 2-d site effects).

This last category of site effects is a good example of the problem that we have when dealing with models: after the Mexico City experience, any basin is considered able to generate 2-d effects, and this is demonstrated on the basis of simplified models. Real basin could not produce 2-d effects due to lateral heterogeneity, as demonstrated by the Parkway basins in New Zealand or the Bovec basin in Slovenia.

But why are we so fond with simplified model and try to apply them everywhere? For the simple reason that our knowledge about distribution and amplitude of site effects is biased by selective under-sampling. In other words, we act as the drunk man that look for its lost keys under a street lamp, not because he is sure that he lost them there, but because there is the light.

From 1997 to 2004 years, our group conducted two kinds of HVSR campaigns: microzonation studies and post earthquake surveys. The aim was different: in the first case, we sampled all the municipalities or sites within the study area without a priori selection; in the second case, we performed measurements aimed at investigating possible resonance phenomena between the fundamental frequency of the soil and the one of damaged buildings. Thus in this second case we performed a selection driven by the observed damage. The existence of hidden selection criteria in a sample may introduce unwanted bias in the outcome. We performed all the measurement using the same instrumentation and processing technique. With a database of 540 HVRS measurements, we compared three samples: 407 measurements from microzonation studies in different Italian regions (Umbria, Marche, Basilicata, Molise), 79 from post- earthquake surveys (Umbria-Marche, Slovenia, Calabro-Lucano, Izmit) and 54 from our latest post-earthquake survey (Molise). In this last case, we performed measurements not only at the most damaged sites, but also more uniformly in the investigated area. The variable we have considered for our statistics is the higher HVSR value in the frequency range 0.5-10 Hz. It is clear (Fig. 1) that the data from microzonation studies and post-quake survey have similar distributions but very different parametrisation.

62

This means that even if microtremors HVSR do not have an amplitude exactly matching amplification from other techniques, there is a monotonic correlation between HVSR and damage. Another observation derived from Fig. 1 is about the probability of observing high HVSR values at randomly sampled sites. The sampled data from microzonation studies has a lower probability than the one resulting from post-quake studies. This means that there are few sites with strong local amplification, and they are those that claim for attention when the damage is driving our researches. The distribution observed after the Molise 2002 event shows that a study conducted after an earthquake yield the same distribution of earthquake-independent studies, provided that the sampled location are not pre-selected but the whole area is investigated. Now the same pattern is emerging using SSR technique applied to earthquakes.

There is no clear cut conclusion from this list of question and problems, but to evaluate the virtues of an holistic approach we can find inspiration in another day of remembrance: this December it will be the 150th anniversary of the 1857 Val d'Agri event that led Robert Mallet to his famous survey and prompted him to write “First principles of observational seismology”. He was an engineer, stated the rules for seismology and wrote beautiful pages about local geology and site effects: disciplinary barriers and fences were introduced much later on, and we can ask ourselves if we had a real payback from that.

Figure1. Comparison between the CDFs of different sets of HVSRs, as a function of the largest value observed in the frequency range 0.5-10 Hz.

References Mucciarelli M. , M. R. Gallipoli (2004) The HVSR technique from microtremor to strong motion:

empirical and statistical considerations; 13th World Conference on Earthquake Engineering Vancouver, 2004 – CD-Rom, Paper No. 45

63

Site effect studies: what we can learn from Colfiorito experience

Giuliano Milana and Antonio Rovelli Istituto Nazionale di Geofisica e Vulcanologia

The role of site amplification became an important topic in ground motion estimation starting from Mexico earthquakes in 1985. Since that experience the studies about seismic effects appeared more frequently in scientific literature involving both theoretical and experimental issues.

In the ‘90s the occurrence of destructive earthquakes in developed countries increased the attention to this issue particularly with 1994 Northridge (California) [Gao et al. 1996, Graves et al. 1998] and 1995 Kobe (Japan) earthquakes [Pitarka et al. 1998].

The 98 stations temporary array installed in the Northridge area were able to record a long aftershocks sequence and to depict an area where important amplification effects occurred. These investigations represent some of the first attempts to interpret site effects using an approach that overtake the simple 1D modeling based on surface geotechnical data.

The strong local effects observed in Kobe earthquakes was modeled in terms of basin edge effects taking into consideration also source kinematics and directivity trying to construct models more complicated than 1D ones.

In Italy a very early site evaluation attempt was tried for Ancona earthquake (1972) where seismic response was simulated using both 1D and 2D approaches based on available geological and geotechnical information [Marcellini 1986]. For the 1976 Friuli earthquake amplification studies were performed in the Tarcento area with the support of “in situ” shear waves measurements. After Irpinia 1980 event some temporary strong motion station were installed in the epicentral area to detect differences in ground motion due to different geological, geotechnical and topographic condition. Also a wide set of geotechnical measurements were performed to classify the strong motion stations that recorded the event. For Val Comino 1984 earthquake some seismological study was performed, using spectral techniques, to justify an anomalous distribution in peak ground motion. In all the cited cases data were not exhaustive anyhow and only some partial indication was obtained.

The 1997 Colfiorito seismic sequence represents the first Italian significant event since developing of site response studies. After the two mainshocks of September 26th several activities aimed to site effects evaluation were started both in epicentral area and in villages and town that suffered severe damages. Many researchers, belonging to different scientific institutions and universities, deployed both strong and weak motion instruments in sites, sometimes very close to each other, characterized by different surface geology conditions. The long duration and the spatial evolution of the sequence allowed to investigate quite a big number of sites including the city of Fabriano, Nocera Umbra, Sellano, Mevale di Visso, Verchiano [Marcellini et al. 2000, Regione Umbria 2000, Cattaneo et al. 2000, Gaffet et al. 2000]. An other important product was the publication of an extensive microzoning campaign that involved the entire Umbria territory with an approach based on available geological and geotechnical information and a numerical modeling of site amplification using 1D or 2D approaches according to site typology.

The epicentral area of Colfiorito September 26th sequence was interested by very important site amplification effects including propagation in sedimentary basins, topography, and effects linked to highly fracturated rock.

The response of Colfiorito [Di Giulio et al. 2003] basin was investigated through seismic arrays (Figure 1). This tool provided information on both the incident wavefield and the vertical velocity structure. In particular array techniques allow to explain the long duration of 1-s resonance and the loose of coherence in terms of local reverberations. The data analysis also showed the difference between first vertically reflected S waves arrivals and later laterally scattered surface wave trains. Array ambient noise measurements [Di Giulio et al. 2006] allowed to obtain a shear waves velocity measurements based on the inversion of Rayleigh waves dispersion curves.

64

Topography effects were investigated along the hill where Nocera Umbra historical centre is built. The interest in Nocera Umbra site was raised by the extremely high peak ground acceleration recorded for the two mainshocks and for one of the strongest shocks (pga > 0.5 g), these acceleration values are actually the strongest recorded in Italy.

Figure 1. Colfiorito basin along with examples of recorded seismic signals.

Nocera Umbra hill was monitored with few seismic stations located along the axis of the ridge and along its slopes. Amplification effects due to topography were found, particularly in the top of the hill, but with spectral amplification not as big as required to reach the observed pga values. A very interesting feature of topography effect on ground motion seems to be the strong polarization induced on seismic signals, this feature is now under investigation through the numerical 2D modeling of the area.

The strong motion recording station of Nocera Umbra was classified as “rock” site. The extremely high ground motion recorded values, despite the stiff nature of the site (Vs reaches values as high as 1000 m/s at a depth of 8 meters), suggested the necessity of a detailed geological and seismological study of the area. As shown in Figure 2, a set of 6 seismic stations was deployed in a linear array configuration along a cross section of Topino river [Cultrera et al. 2003]. One of the new deployed station was very close to the strong motion recording site. Also in this case a strong difference in ground motion was identified at a very local scale and some detailed geological investigation was performed to better understand local site condition. The obtained cross section, (Figure 2), enlighten the presence of a fault zone in the area, characterized by highly fracturated rock. Ground motion results to be amplified in the fault zone in a broad (4-10 Hz) frequency band with amplification factors quite impressive. An other interesting feature is that amplification strongly depends on event azimuth. The interpretation of Nocera Umbra amplification was given in terms of trapped waves within the fault zone with an efficiency that strongly depends on source azimuth.

The site effects studies produced after Colfiorito earthquakes gave some very interesting results in several area and different surface geological conditions. Many lessons can be learned from these studies, and some of the applied techniques can also now be used before a destructive event in order to forecast amplification effects and to give indication to reduce their effects. Unfortunately there are still many cases where complexities in local geological setting and lack of information do not help in making such a kind of predictions.

65

Figure 2. Horizontal component seismograms of a subcrustal (Mw = 5.3) and two shallow earthquakes located south-southeast (event 31, Mw = 3.2) and north (event 75, Mw = 4.3) of the array projected onto a geologic section (trace A–B in Figure 2). Ground motion variations are significant at the scale of tens of meters along the profile.

The example of Molise (2002) events, with their dramatic effects in San Giuliano di Puglia, gives a clear idea of the efforts that still need to be done to better understand local site effects and to reduce their consequences on human life. For this reason site amplification studies are more and more frequent and play an important role in many scientific projects included those promoted by National Civil Defense Department.

66

References Cattaneo, M. and Marcellini, A. (editors) (2000). Terremoto dell'Umbria-Marche: - Analisi della sismicità

recente dell'Appennino umbro-marchigiano. - Microzonazione sismica di Nocera Umbra e Sellano, CNR-Gruppo Nazionale per la Difesa dai Terremoti - Roma, 228 pp.

Cultrera, G., Rovelli, A., Mele, G., Azzara, R.M., Caserta, A. and Marra, F. (2003). Azimuth-dependent amplification of weak and strong ground motions within a fault zone (Nocera Umbra, central Italy), J. Geophys. Res., 108 (B3), 2156.

Di Giulio G., Rovelli, A., Cara, F., Azzara, R.M., Marra, F., Basili, R., and Caserta, A. (2003). Long-duration asynchronous ground motions in the Colfiorito plain, central Italy, observed on a two-dimensional dense array, J. Geophys. Res., 108 (B10), 2486.

Di Giulio, G., Cornou, C., Ohrnberger, M., Wathelet, M., and A. (2006). Deriving wavefield characteristics and shear-velocity profiles from two-dimensional small-aperture arrays analysis of ambient vibrations in a small-size alluvial basin, Colfiorito, Italy, Bulletin of the Seismological Society of America, Vol. 96, 1915-1933.

Gaffet, S., Cultrera, G., Dietrich, M., Courboulex, F., Marra, F., Bouchon, M., Caserta, A., Cornou, C.,Daschamps, A., Glot, J.P, and Guiguet, R. (2000). A site effect study in the Verchiano valley during the 1997 Umbria-Marche (Central Italy) earthquakes, Journal of Seismology Vol. 4.

Gao, S., Liu, H., Davis, P.M., and Knopoff L. (1996). Localized amplification of seismic waves and correlation with damage due to Northridge earthquake, Bulletin of the Seismological Society of America, Vol. 86, S209-S230.

Graves, R.W., Pitarka A., and Sommerville P.G. (1998). Ground motion amplification in the Santa Monica area: effects of shallow basin edge structure, Bulletin of the Seismological Society of America, Vol. 88, 1224-1242.

Marcellini, A. (1986). Breve storia della MS in Italia, in: Progetto Finalizzato Geodinamica Monografie finali, Vol. 7, Elementi per una guida alle indagini di Microzonazione Sismica, a cura di Ezio Faccioli

Marcellini, A., and Tiberi, P. (editors) (2000). La Microzonazione Sismica di Fabriano, Biemmegraf Srl, Piediripa di Macerata

Pitarka, A., Irikura, K., Iwata, T., and Haruko S. (1998). Three-Dimensional Simulation of the Near-Fault Ground Motion for the 1995 Hyogo-ken Nanbu (Kobe), Japan, Earthquake, Bulletin of the Seismological Society of America, Vol. 88, 428-440.

Regione Umbria (2000). La Microzonazione Sismica Speditiva relativa ai terremoti del 1997-1998 in Umbria. Le situazioni morfostratigrafiche suscettibili di amplificazione sismica locale, Tipolito Visconti, Terni.

67

Ground motion shaking scenarios for the 1997 Colfiorito earthquake

Antonio Emolo (1), Aldo Zollo (1), Vincenzo Convertito (2), Francesca Pacor (2), Gianlorenzo Franceschina (2), Giovanna Cultrera (2), and Massimo Cocco (2)

(1) Università di Napoli “Federico II” (2) Istituto Nazionale di Geofisica e Vulcanologia

1. The statistical-deterministic approach In the recent years, two Italian research projects have been devoted to the simulation of ground

shaking scenarios in different areas. A large part of the activities has been performed in the Umbria region and was in particular related to the 1997 Colfiorito earthquake.

In general the statistical-deterministic approach was adopted for evaluating the scenarios for strong motion parameters (peak values, spectral ordinates, signal integral quantities, and so on) associated with the occurrence of a characteristic earthquake on a given fault.

This approach is based on the realistic occurrence of a single earthquake related to the fracture of an a priori well identified active fault. According to the characteristic earthquake model, an earthquake rupture can repeatedly occurs along the same fault (or fault system) with an almost constant geometry, mechanism and seismic moment, these parameters being mainly related to the direction and intensity of the large scale tectonic stress regime. These ideas are supported by numerous paleoseismic studies of active faults in different tectonic environments [e.g., Pantosti and Valensise, 1990]. On the other hand, each faulting process may not repeat the same style of nucleation, propagation and arrest during successive rupture episodes occurring along a given fault zone, depending these characteristics on the pre-fracturing conditions of rock strength and/or yielding stress along the fault zone. It is therefore assumed that the large scale source characteristics (i.e., fault size and position, focal mechanism and seismic moment) are a priori known as the result of previous geological, geophysical and historical seismicity investigations.

The variability of the rupture process is expected to produce variable strong ground motions at the earth surface, depending on the distribution of the kinematic parameters (final slip distribution, rupture velocity, slip duration …) along the faulting surface. In order to account for the possible variation of the source process from one rupture event to another, a large number of synthetic seismograms should be computed for different (and possible) rupture histories occurring along the characteristic fault selected, so to provide a representative set of strong motion records to be used for hazard estimation. By this strategy, the massive computation of synthetics for different possible rupture models does not provide a single earthquake scenario (as for the standard deterministic approach) but a set of possible scenarios whose variability substantially reflects the heterogeneity of the source process. The advantage of this approach is that the variability of the selected strong ground motion parameter at a given site can be described by the statistical quantities inferred from the large number of simulations available. The earthquake scenario can then be represented, for example, by a couple of maps, one describing the spatial distribution of the mean value of the considered ground motion parameter and the other representing the associated variability for example in terms of standard deviation.

2. Ground motion simulation techniques The Colfiorito earthquake was simulated using different numerical approaches that provide both

approximated and full wave field. We will discuss only the results from the approximated techniques Asymptotic Method (hereinafter, ASM) and Deterministic-Stochastic Method (hereinafter, DSM). Although these techniques are approximated, they are suitable to generate shaking scenarios near an extended fault where the direct S wave field is generally dominant in amplitude with respect to the reflected, converted and surface phases.

According to the ASM method [Zollo et al., 1997], the seismic wave field associated with the rupture of a finite fault (or fault system) is computed solving numerically the representation integral. Since

68

the technique is aimed at simulating the high frequency radiation, the Green function are computed by the ray theory in a flat layered propagation medium [Farra et al., 1986]. The fault is discretized in rectangular sub-faults whose size is chosen so as to reduce the numerical spatial aliasing. The space-time slip distribution on the fault is obtained assuming a ramp time function parameterized by the rupture time, the rise time and the final slip for each sub-fault. The final slip distribution on the fault is computed in accordance to the k-squared model [Herrero and Bernard, 1994].

The DSM technique [Pacor et al., 2005] starts from the simulation of an acceleration envelope modelled using the isochron formulation [Bernard and Madariaga, 1984] for a kinematic source process. The normalized envelope is then used to window a Gaussian white noise, thus reproducing the phase incoherency of high frequency acceleration. The Fourier spectrum of the resulting time series is then multiplied by a target spectrum defined by seismic moment of the target earthquake and contains source, attenuation and site terms. The target spectrum is site-position dependent because its corner frequency is computed as the inverse of the envelope duration as seen by the receiver. Thus the corresponding acceleration time series will contain information on the directivity effects as modelled by the isochron formulation.

3. Simulation facts and results Computations were performed for the M6, September 26, 1997, Colfiorito earthquake. The

geometry, mechanism and position of the causative fault were inferred from the inversion of near source accelerometric data [Capuano et al., 2000]. As explained before, there parameters were kept constant in all simulations. On the other hand, final slip distributions and rupture velocity were allowed to vary. Scenarios for bilateral and unilateral rupture propagations were considered accounting for different rupture nucleation point positions. Synthetic seismograms were computed for each source rupture process considered at a regular grid of 64 virtual receivers regularly distributed on an area on 60km x 60km around the fault.

We found that results are strongly dependent on the rupture velocity value assumed. In particular, the higher is the rupture velocity the higher are the peak values. Moreover, the DSM technique provides scenarios that are dominated by the directivity effect while for the ASM the spatial distribution of ground motion parameters is mainly determined by the radiation pattern. In any case, the variability of ground motion parameters obtained by both simulation techniques is comparable or smaller then the standard deviation associated with the empirical model UMA2005 [Bindi et al., 2006].

The ground motion variability observed for the Colfiorito earthquake is well reproduced when an unilateral rupture propagation toward the NW is considered with rupture velocity of about 2.7 km/s. However, the data can be modelled only in some particular frequency bands due to the relevant amplification associated with the site effects (e.g., the case of Nocera Umbra).

One of the main lesson learnt form the Cofiorito earthquake is that the extended fault effects (e.g., the rupture directivity) are important also in the case of moderate size earthquake.

4. Seismic hazard assessment: integrating probabilistic and deterministic approaches The characteristics of the Colfiorito earthquake were also used to introduce a new technique aimed

at integrating the probabilistic and deterministic approaches for seismic hazard assessment. This technique (Convertito et al., 2006) is based on the site-dependent evaluation of the probability of exceedance for the chosen strong motion parameter. The probability is obtained from the statistical analysis of the synthetic waveform database produced for a large number of possible rupture histories occurring on a characteristic earthquake fault, as described before. This integrated technique is aimed at overcoming some of the limitations of both purely probabilistic and purely deterministic approaches for seismic hazard evaluation. In particular it allows the time variable (in terms of return period and time of interest) to be accounted in the deterministic scenario studies and the source characteristics and effects (e.g., fault geometry, radiation pattern, directivity, …) to be accounted in the probabilistic technique. Due to the availability of synthetic waveforms, the analysis can be easily extended to any desired ground motion parameter. Moreover, the site effect can be taken into account if the site transfer function was available.

The method has been applied to the hazard evaluation in the Umbria-Marche region, assuming as threat a fault having the same geometry and mechanism as the Colfiorito earthquake.

69

5. References Bernard, P., and Madariaga, R. (1984). A new asymptotic method for the modelling of near field

accelerograms. Bull. Seism. Soc. Am., 74, 539-558. Bindi, D., Luzi, L., Pacor, F., Franceschina, G. and Castro, R.R. (2006). Ground-motion predictions from

empirical attenuation relationships versus recorded data: the case of the 1997-1998 Umbria-Marche, Central Italy, strong-motion data set. Bull. Seismol. Soc. Amer., 96, 984-1002.

Capuano, P., Zollo, A., Emolo, A., Marcucci, S. and Milana, G. (2000). Rupture mechanism and source parameters of the Umbria-Marche main shocks from strong motion data. J. Seism., 4, 436-478.

Convertito, V., Emolo, A. and Zollo, A. (2006). Seismic-hazard assessment for a characteristic earthquake scenario: an integrated probabilistic-deterministic method. Bull. Seism. Soc. Am., 96, 377-391.

Farra, V., Bernard, P. and Madariaga, R. (1986). Fast near source evaluation of strong motion for complex source models. In: Earthquake source mechanics (S.Das, J.Boatwright and C.H.Scholz Eds.), pp. 121-130. AGU Geophysical Monograph 37, Washington, D.C.

Herrero, A. and Bernard, P. (1994). A kinematic self-similar rupture process for earthquakes. Bull. Seism. Soc. Am., 84, 1216-1228.

Pacor, F., Cultrera, G., Mendez, A. and Cocco, M. (2005). Finite fault modelling of strong ground motions using a hybrid deterministic-stochastic approach. Bull. Seism. Soc. Am., 95, 225-240.

Pantosti, D. and Valensise, G., (1990). Faulting mechanism and complexity of the Novembre 23, 1980, Campania-Lucania earthquake, inferred from surface observations. J. Geophys. Res., 95, 15319-15341.

Zollo, A., Bobbio, A., Emolo, A., Herrero, A. and De Natale, G., (1997). Modelling of ground acceleration in the near source range: the case of 1976 Friuli earthquake (M6.5). J. Seism., 1, 305-319.

70

From the “seismic classification” to the design ground motion parameters for the Italian building code

Massimiliano Stucchi (1), Carlo Meletti (1), Fabio Sabetta (2), and Dario Slejko (3) (1) Istituto Nazionale di Geofisica e Vulcanologia; (2) Dipartimento della Protezione Civile

(3) Istituto Nazionale di Oceanografia e di Geofisica Sperimentale

At the time of the 1997, Umbria-Marche earthquake the Italian building code was still ruled by the law n.64 of 1974, and related updating. It was mainly based on a standard shape design spectrum, with one anchorage value for each of the three seismic zones. A large portion of the territory (more than 5100 Communes out of about 8100) was not subjected to any seismic provision (Fig. 1). The quoted law required the Communes to be associated to one of the seismic zones according to the system of the so-called seismic classification; in other words, according to expert judgement, without any criteria supplied by the law.

The map in Fig.1 was obtained between 1981 and 1984 by means of Decrees of the Ministry of Public Works, following the proposal of the CNR’s “Progetto Finalizzato Geodinamica” [CNR-PFG, 1980; WG coordinated by V. Petrini], released soon after the 1980, Irpinia and Basilicata earthquake. Such a proposal combined three parameters: i) a probabilistic, hazard one; ii) the distribution of maximum earthquake effects in the last millennium; iii) a risk one. It also underlined the need for further investigation in some areas. The only variation in the Decrees with respect to the proposal was the introduction of a third seismic category, applied only to the areas damaged by the 1980 event, including Naples, but not extended to other areas with similar seismic hazard values. After 1984 no strong events took place to provide a test for the seismic zones.

Figure 1. Seismic classification in 1984.

Figure 2. Map of the “Proposta 98”.

In 1997, the “Gruppo Nazionale per la Difesa dai Terremoti” made available an updated seismic hazard assessment [Slejko et al., 1998], based on new seismogenic source model, earthquake catalogue, attenuation law. This elaboration put also in evidence some gaps in the official seismic map. For the

71

initiative of the State Secretary for the Civil Protection, F. Barberi, the Commission for the Major Risks established in 1997 a Working Group charged of updating the map of the seismic classification. The WG, coordinated by C. Gavarini, released in 1998 a study based on the combination of three parameters: the i) Housner intensity for 95-years return period, spectral interval 0.1-0.5 sec; ii) the Housner intensity for 475- years return period, spectral interval 0.2-2 sec; iii) the maximum observed intensity in the last millennium.

The study and related map (Fig. 2), from here on “Proposta ‘98”, was approved by the Commission for the Major Risks, transmitted to the Ministry of Public Works and published [Gruppo di Lavoro, 1999], but never adopted. In the same period the system of seismic provisions became complicated as a consequence of the Decree n. 112 of 1998. According to it, the power on the building code was left to the central State, while the power for assessing the seismic zones was assigned to the Regions, although the criteria for such an assessment were left to the State. The High Council for Public Works established a Committee aimed at formulating the aforementioned criteria; in 1999 the criteria upon which the “Proposta ‘98” was based were submitted to the Regions. The main objections came about the third category, which had to be extended from 99 up to 1698 Communes.

While more elaborations appeared later on [Albarello et al., 2000; etc.], the official map did not change until the earthquake of Molise (2002), which showed in a dramatic way the gaps in that seismic zonation. Actually, some of the most damaged Communes, including San Giuliano di Puglia, did not belong to any seismic zone, while the “Proposta ‘98” assigned them to the 2nd category.

In the aftermath of the earthquake, in December 2002 the State Secretary to Prime Minister, G. Letta, established a Committee, coordinated by G.M. Calvi, charged of updating, as soon as possible, the entire seismic building code. The Committee worked very hard and in January 2003 released a corpus of elaborations compiled in the view of the Eurocode 8; this corpus was then officially adopted by the Government by means of the Prime Minister Ordinance 3274 “First elements about new classification criteria of the Italian territory and of building codes in seismic areas”, published on May 8, 2003 and then integrated by the P.M. Ordinance 3316 of October 2, 2003. The new code allowed to assess in a scientific way the seismic reference for the definition of the seismic zones. In particular, Ordinance 3274 foresaw 4 zones, defined by standard seismic hazard values on a very stiff soil (PGA with 10% exceeding probability in 50 years). A first and main consequence was that all the territory was then assigned to one of the 4 zones.

Figure 3. Seismic zones as from Ordinance 3274.

Figure 4. Communes to be declassified.

72

Ordinance 3274 also defined the criteria according to which the reference seismic hazard had to be assessed. Since no elaborations were found to match such criteria, a new elaboration was requested within one year and, as a first application, the map of the “Proposta ‘98”, read in terms of PGA and without the “declassifications” suggested by it (about 800 Communes, Fig. 4), was temporarily adopted. The Regions were allowed to shift the Communes by one category at maximum in this phase; most of them adopted it as it was, with the exception of Sicilia, Lazio, Basilicata and Trento Province, which introduced slight changes. However, the seismic provisions for category 4, the less hazardous one, were generally left on a waiting list.

Following the suggestion of the Commission for Major Risks, to fulfill the above mentioned request INGV established in July 2003 a Working Group (coordinated by M. Stucchi) devoted to the compilation, according to the newly defined criteria and within May 2004, of a reference seismic hazard map of the Italian territory. The new elaboration was due for the following, two main reasons: i) to complete the transition from the old building system to the new one (the Ordinance 3274 was still based on the classification system); ii) to assign to a seismic zone, in a stable way, the 800 Communes which had not been “declassified”, yet.

The Working Group, composed by INGV and non-INGV researchers, released a first version in November 2003, which was reviewed by a group of European experts. The final version, called MPS04 [Gruppo di Lavoro MPS, 2004; Fig. 5], was released in April 2004; it was based on the most updated input data and, for the first time in Italy, on a logic tree exploring epistemic alternatives of: i) catalogue completeness; ii) seismicity rates; iii) ground motion attenuation relationships. In such a way, the distribution of the 16th and 84th percentile were also evaluated as uncertainty estimates.

In May 2006 the PM Ordinance 3519 re-defined, without many changes, the criteria for the seismic hazard assessment already established by the Ordinance 3274. Further than that, it finally adopted MPS04 as the official reference map to be considered by the Regions and the relevant values, covering about 16,000 grid points, were made available to the public on a dedicated website (http://zonesismiche.mi.ingv.it).

Figure 5. The seismic hazard map MPS04.

Figure 6. Straightforward application of MPS04.

73

If adopted in a straightforward way, MPS04 would give rise to a seismic zonation with many differences with respect to the current one (Fig. 5). In particular: i) zone 1 would be reduced by about 150 Communes; ii) as for zone 2, the number of Communes would remain the same, although their distribution in space would vary considerably; iii) the number of Communes of zone 3 would pass from about 1600 to about 3600; iv) zone 4 would “loose” about 1700 Communes (from about 3400 to about 1700). However, to avoid that a minimal difference could automatically assign two Communes to two seismic zones and, therefore, to two fairly distant values of the design spectrum, Ordinance 3519 gave the Regions the power of using a “tolerance belt”, to be exerted according to mainly political criteria.

Two Regions, Toscana and Molise, made their Decrees in 2006; as a matter of fact, the concept of a centrally defined classification was close to an end, as varied scenarios could be proposed [Meletti et al., 2006]. But further changes were on the way. The availability of seismic hazard point values stimulated the possibility that such values be directly used for locally defining the design spectrum; after all, the building code already partially allowed it. The project INGV-DPC S1 started in 2005, with the goal of assisting DPC in “squeezing out” the maximum possible information from MPS04 data and logic tree.

The project, early coordinated by G.M. Calvi and M. Stucchi and later by C. Meletti, having G. Di Pasquale as DPC representative, evaluated the seismic hazard for 9 exceedance probabilities in 50 years and 10 spectral ordinates, from 0.1 to 2 s on a regular grid of about 5 km.

With such a wealth of available data to the public through a dedicated WebGis interface [http://esse1.mi.ingv.it] it was then clear that the design spectrum could be easily defined at each site from the probabilistic ground motion parameters, in a more precise way than the spectra linked to the four seismic zones. This direction was adopted by two Committees established by the High Council for Public Works, with DPC representatives, which worked from March 2006 to January 2007 and from February to July 2007. The output, to be published, is a new release of the building code in which the design spectrum is based on three parameters, including the standard PGA, derived from the aforementioned data of the S1 project. With respect to the seismic zones, the new code puts an end to the “seismic classification” for what the building provisions are concerned: the Communes will still be assigned by the Regions, for administrative purposes only, to one of the four zones on the basis of MPS04 data.

Figure 7. An example of the WebGis interface offering the seismic hazard data at the grid points.

Figure 8. Communes not assigned to a seismic zone up to 2002.

74

In conclusion the 1997, Umbria-Marche earthquake raised the concern on the seismic zonation; the 2002 Molise earthquake triggered a deep change in the seismic provisions, including the building code and the seismic hazard assessment. Unfortunately, the absence of seismic provisions in the areas not assigned in 2002 to any seismic category (Fig. 8; some of them should be now in zone 2!) is still far from being properly addressed, by the Regions and the central Government, after 5 years.

References Albarello D., Bosi V., Bramerini F., Lucantoni A., Naso G., Peruzza L., Rebez A., Sabetta F. e Slejko D.

(2000). Carte di pericolosità sismica del territorio nazionale. Quaderni di Geofisica, 12, Roma, 7 pp., CD-ROM, 4 allegati.

CNR-PFG (1980). Proposta di riclassificazione sismica del territorio nazionale. Pubbl. 361, Roma, 83 pp. Gruppo di Lavoro (1999). Proposta di riclassificazione sismica del territorio nazionale. Ing. Sism., 16, 1, 5-

14. Gruppo di Lavoro MPS (2004). Redazione della mappa di pericolosità sismica prevista dall’Ordinanza

PCM 3274 del 20 marzo 2003. Rapporto Conclusivo per il Dipartimento della Protezione Civile. INGV, Milano-Roma, aprile 2004, 65 pp. + 5 allegati.

Meletti C., Stucchi M., Boschi E. (2006). Dalla classificazione sismica del territorio nazionale alle zone sismiche secondo la nuova normativa sismica. In: Guzzoni D. (a cura di), Norme Tecniche per le costruzioni. Il Sole 24 Ore editore, Milano, 139-160.

Slejko D., Peruzza L. e Rebez A. (1998). The seismic hazard maps of Italy. Ann. Geofis., 41, 2, 183-214.

75

Vulnerability evaluations for seismic risk assessment

Mauro Dolce Dipartimento della Protezione Civile

Seismic risk is a public safety issue that requires appropriate risk management measures and means to protect citizens, properties, infrastructures and the built cultural heritage. The aim of a seismic risk analysis is to estimate the consequences of seismic events on a geographical area (a city, a region, a state or a nation) in a certain period of time. The effect to be predicted is the physical damage to buildings and other facilities, together with the consequences of the physical damage: the number and type of casualties, the potential economic losses, due to the direct cost of damage and to indirect economic impacts (loss of the productive capacity and business interruption), the loss of function in lifelines and critical facilities (such as hospital, fire stations, communication system, transportation networks, water supply, etc.) and also social, organizational and institutional impact. The results provided by seismic risk analysis are used in all the phases of the emergency cycle: before, during and after the event.

One of the main factors in loss assessment is the seismic vulnerability of buildings, i.e. the susceptibility of buildings to suffer a certain damage level if subjected to an earthquake. Since the building behaviour is not deterministic, the suffered damage is to be considered as a random variable. Hence building vulnerability is expressed as the probability of reaching or exceeding a specified damage level, given an intensity measure of the earthquake. Intuitively, the more the building is vulnerable, the higher will be the suffered damage, given the same seismic intensity. In operative terms a vulnerability method should correlate the seismic intensity to the physical damage suffered by the building, depending on the structural, geometric, and technological building features.

Despite the simple and straightforward definition, several methods for vulnerability assessment have been developed and proposed in recent years. Consequently, several have been the attempts to provide classification criteria for them [UNDP/UNIDO, 1985, Corsanego and Petrini, 1994; Dolce et al. 1994; Dolce, 1996, Di Pasquale et al., 2001, Calvi et al., 2006]. Almost all the classification criteria agree on the distinction between three types of vulnerability approaches, when reference is made to the genesis of the methods:

- empirical vulnerability methods, based on post-earthquake damage observations or laboratory tests; - mechanical vulnerability methods, based on structural mechanical models; - method based on expert judgment or Delphi methods. Each one of these approaches is differently characterized by positive features and limitations.

Observed vulnerability methods are based on statistics of past earthquake damage data and usually lead to the definition of DPM’s, i.e. Damage Probability Matrices [Withman, 1973; Braga et al., 1982; Rossetto and Elnashai, 2003], or directly to fragility curves [Coburn and Spence, 1992]. One drawback of the approach is that data are often limited to the building typologies and to the felt seismic intensities in the affected area. On the contrary mechanical models may be used to analyse any type of building for any seismic intensity. However they reflect the building response according to the model or to the seismic code and not to the real seismic behaviour, and without differentiating between partial and total collapses. The first systematic attempt to codify the seismic vulnerability of buildings completely from expert judgment was made by the Applied Technology Council, providing DPM for 78 engineering facility classes, 40 of which were relevant to buildings (ATC13, 1987). In this case drawbacks arise from the lack of any direct check or derivation with reality, apart from the experience cumulated by the expert. Hence recent approaches have also considered the possibility of hybrid techniques, when empirical observations are used to condition mechanical vulnerability evaluations [Kappos et al. 1996].

Another important difference among methods is the number of buildings the model refers to. This is strictly related to the accuracy of the building information required in the model. It is possible to express vulnerability just for one building. In this case a detailed building representation is required and an accurate prediction is expected. The data collection is quite cumbersome since the exact geometry and most of the structural details, as well as the material properties, are required to be known. For instance in

76

RC buildings the size and position of beams and columns, reinforcement details, concrete and steel strengths are required to make reliable models. This procedure is not obviously applicable when the response of a huge number of buildings is to be predicted. In this case, according to human resources, only the most important parameters can be collected. Hence models have to be, necessarily, more simplified and, consequently, less accurate. In this case similar buildings are often grouped in classes of buildings and the same model refers to all the buildings belonging to the same class. Hence within the same class, one building is not distinguishable from another.

This means that when performing loss analysis at different scale of analysis (single building, local or territorial scale) it is not possible to operate with the same vulnerability method. On the other hand, usually, a multi-level approach is introduced, where progressively more accurate methods are used as the number of buildings to be investigated decreases. As an example in Risk-UE project [Lagomarsino and Giovinazzi, 2006] two different approaches have been proposed: a I level approach to be used with macroseismic intensity hazard scenarios and a II level approach, mechanical-based, for PGA and response spectra hazard scenarios. Other examples may be found in [Goretti et al., 2007; Grant et al., 2007, Cosenza et al., 2005]. In the distinction between model accuracy and data required, the problem of uncertainty of the model has seldom be addressed. In other words, in practice models have been selected and data have been collected based on personal experience, without any study on the benefit of introducing or removing several parameters in the model with respect to the desired accuracy of the prediction.

Another important distinction has to be made between primary and secondary vulnerability [Di Pasquale and Goretti, 2001]. While primary vulnerability provides the probability of suffering a physical damage conditional to building type and seismic intensity, secondary vulnerability provides the probability of losses (economic, functional, etc) given the building type and the suffered damage. The convolution of the primary and the secondary vulnerability provides the total vulnerability that is the building susceptibility to suffer losses from earthquakes. Some models provide the only primary vulnerability, some others provide the total vulnerability. It is the case, for example, of the Petrini model [Guagenti and Petrini 1989], that provides the mean relative economic loss due to the earthquake.

Sometimes in loss assessment the model used for the primary and secondary vulnerability are provided by different authors. In this case it is important to check the internal consistency of the model. It has been shown [Di Pasquale and Goretti, 2001] that the secondary vulnerability may significantly differs when different measures of physical damage are considered: the mean damage to vertical bearing structures, the maximum damage to vertical bearing structures or a mean structural damage that takes into account a weighted sum of the damage suffered by each building structural component. Hence it happens that it is not possible to couple a primary vulnerability model with a secondary vulnerability model, if the damage measures are not the same. However, this simple and quite obvious concept has been often disregarded in applications.

Several other important differences among existing models regard the damage variable and the intensity variable. Both can be discrete or continuous. Example of discrete damage-discrete intensity models are the so-called DPMs [Braga et al., 1982], where damage ranges usually between 0 (no damage) to 5 (the total collapse), and intensity is expressed in terms of macroseismic intensity ranging fro V to XII. The numerous different macroseismic scales that have been used in the past to assess vulnerability (MCS, MM, MSK, EMS98) further complicate a comparison between different methods. An example of discrete damage-continuous intensity model is the one presented in [Rota et al., 2006] where the intensity measure is expressed in terms of PGA. A continuous damage-continuous intensity model is the Petrini model, where PGA is correlated with the relative economic cost, ranging fro 0 to 1. Other examples of intensity measures used in vulnerability models are PGV, Arias intensity [Sabetta et al., 1998], spectral accelerations [Goretti et al., 2006], spectral displacements [Calvi, 1999, Crowley et al., 2004], etc.

Damage is usual considered as a scalar quantity (that is damage to the whole structure, damage to the vertical bearing structure, etc.). In some analysis the damage is considered a vectorial quantity (damage to structural components and damage to non structural components). In this case, each quantity is usually derived independently from the other. In some cases, a joint probability function has been proposed to represent the building vulnerability, although this complicates considerably the analysis. In some cases even the secondary vulnerability has been expressed in term of a damage vector, providing the building unusability probability given the suffered structural and non structural damage [Di Pasquale and Goretti, 2001].

77

Recently a Macroseismic method has been derived [Lagomarsino and Giovinazzi, 2006] from the definitions provided by the EMS-98 macroseismic scale. The method provides DPM’s for building types, translating the definitions provided by the EMS-98 scale to identify the intensity degree from the damage suffered by buildings. This has been done by the use of the Fuzzy Set Theory and of the Probability Theory. The model has the positive feature of providing a unique approach suitable for all European regions, however it enforces the building types to behave according to the EMS98 scale and often this has been proved not to be the case [Maffei et al., 2006].

In conclusion since vulnerability deals with the problem of predicting the real damage to structural and non structural components in many building types when affected by an increasing intensity measures, it is inevitably an argument still open for research. What is reputed necessary in applications is that the parameters used to express hazard should be suitable for an efficient building response representation and that the number of the building type information should be consistent with the model accuracy and with the resources needed to collect data.

References ATC, 1985 Earthquake Damage Evaluation Data for California Applied Technology Council report ATC-

13 Redwood (CA) Braga, F., Dolce, M. and Liberatore, D., (1982), A statistical Study on Damaged Buildings and Ensuing

Re-view of the MSK-76 Scale, Proc. VIII European Conference on Earthquake Engineering, Athens, September 1982.

Calvi, G.M. (1999), A displacement-based approach for vulnerability evaluation of classes of buildings, Journal of Earthquake Engineering, Vol. 3, No. 3, pp. 411-438.

Calvi, G.M., Pinho, R., Magenes, G.,Bomber, J. J., Restrepo-Velez, L.F. and Crowley, H., (2006), The Development of Seismic Vulnerability Assessment Methodologies Over the Past 30 Years”, ISET Journal of Earthquake Technology, Vol. 43, No. 3.

Coburn, A. and Spence, R., (2002), Earthquake Protection (Second Edition), John Wiley and Sons Ltd., Chichester, England.

Corsanego, A. and Petrini, V., (1994), Evaluation criteria of seismic vulnerability of the existing building patrimony on the national territory, Ingegneria Sismica, Patron ed., Vol. 1, pp. 16-24.

Cosenza, E., Manfredi, G., Polese, M. and Verderame, G.M., (2005), A multi-level approach to the capacity assessment of existing RC buildings, Journal of Earthquake Engineering, Vol. 9, No.1, pp. 1-22.

Crowley, H., Pinho, R. and Bommer, J.J., (2004), A probabilistic displacement-based vulnerability assessment procedure for earthquake loss estimation, Bulletin of Earthquake Engineering, Vol. 2, No.2, 173-219.

Di Pasquale, G. and Goretti, A., (2001), Vulnerabilità Economica e Funzionale degli Edifici Residenziali Colpiti dai Recenti Eventi Sismici Italiani", X Congresso Nazionale "L'Ingegneria Sismica in Italia", Potenza-Matera, 9-13 September.

Di Pasquale, G., Goretti, A., Dolce, M. and Martinelli A., (2001), Confronto fra differenti modelli di vulnerabilità degli edifici, X Congresso Nazionale “L’ingegneria Sismica in Italia”, Potenza-Matera, 9-13 September.

Dolce, M., Zuccaro, G., Kappos, A. and Coburn, A., (1994), Report of the EAEE Working Group 3: Vulnerability and risk analysis, Proc. X European Conference on Earthquake Engineering, Vienna, Vol. 4, pp. 3049-3077.

Dolce, M., (1996), Seismic vulnerability evaluation and damage scenarios, US-Italian Workshop on Seismic Evaluation and Retrofit, December 12-13, 1996 Columbia University, New York City.

Goretti, A., Di Pasquale, G. and Giordano, F., (2006), La Vulnerabilità Sismica degli Edifici Residenziali: Recenti Proposte del SSN, Progetto Reluis Linea 10, Workshop su Vulnerabilità dell’Edilizia Ordinaria, Genoa, 11 July

Goretti, A., Naso, G., Giordano, F., Vulcano, A. and Mazza, F., (2007), Scenari Sismici a Scala Urbana. I Primi Risultati del Progetto Crotone, XII Congresso Nazionale “L’ingegneria Sismica in Italia”, Pisa, 10-14 June.

Grant, D., Bomber, J.J., Pinho, J., Calvi G.M., Goretti, A. and Meroni, F., (2007), A Prioritization Scheme for Seismic Intervention in School Buildings in Italy, Earthquake Spectra, Volume 23, Issue 2, pp. 291-314

78

Guagenti, E. and Petrini, V., (1989), Il caso delle vecchie costruzioni:verso una nuova legge danni intensità, IV Convegno Nazionale “L' Ingegneria Sismica in Italia”, Milano.

Kappos, A.J., Pitilakis, K. and Stylianidis, K.C., (1996), Cost-benefit analysis for the seismic rehabilitation of buildings in Thessaloniki, based on a hybrid method of vulnerability assessment, Proc. of the 5th International Conference on Seismic Zonation, Nice, France, Vol. 1, pp. 406-413.

Lagomarsino, S. and Giovinazzi, S., (2006), Macroseismic and Mechanical Models for the Vulnerability Assessment of Current Buildings, Bulletin of Earthquake Engineering, Volume 4, Number 4, November 2006, Special Issue “Risk EU Project” Guest Editors: R.Spence and B.Le Brun

Maffei, J., Bazzurro, P., Marrow, J. and Goretti, A., (2006), Recent Italian earthquakes: examination of structural vulnerability, damage and post-earthquake practices – case studies and comparisons to U.S. practice, Earthquake Engineering Research Institute.

Rossetto, T. and Elnashai, A., (2003), Derivation of vulnerability functions for European-type RC structures based on observational data, Engineering Structures, Vol. 25, pp. 1241-1263.

Rota, M., Penna, A. and Strobbia, A., (2006), Typological fragility curves from Italian earthquake damage data, Proc. of the 1st European Conference on Earthquake Engineering and Seismology, Geneva, paper no. 386.

Sabetta, F., Goretti, A. and Lucantoni, A., (1998), Empirical fragility curves from damage surveys and estimated strong ground motion, Proc. of the Eleventh European Conference on Earthquake Engineering, Paris.

UNDP/UNIDO, (1985), Project RER/79/015, Post-earthquake damage evaluation and Strength Assessment of Buildings under Seismic Condition, UNDP, Vienna, Vol. 4

Whitman, R. V., (1973), Damage Probability Matrices for Prototype Buildings, Massachusetts Institute of Technology, Department of Civil Engineering Research, Report R73-57, Cambridge, Massachusetts.