19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 299

Boll.Soc.Geol.It. (Ital.J.Geosci.), Vol. 127, No. 2 (2008), pp. 299-316, 8 figs., 3 tabs.

Queste bozze, corrette e accompagnate dall’allegato preventivo firmato e dal buono d’ordine, debbono essere restituite immediatamente alla Segreteria della Società Geologica Italiana c/o Dipartimento di Scienze della Terra Piazzale Aldo Moro, 5 – 00185 ROMA

Active deformation in Southern Italy, Sicily and southern Sardinia from GPS velocities of the Peri-Tyrrhenian Geodetic Array (PTGA) LUIGI FERRANTI (1), JOHN S. OLDOW (2), BRUNO D’ARGENIO (3), RAIMONDO CATALANO (4), DAVID LEWIS (2), ENNIO MARSELLA (3), GIUSEPPE AVELLONE (4), LAURA MASCHIO (1), GERARDO PAPPONE (5), FABRIZIO PEPE (4) & ATTILIO SULLI (4)

ABSTRACT Campaign measurement (1995-2000) of Global Positioning System (GPS) site velocities in southern Italy, Central Mediterranean area, document differential displacements within the orogens rimming the Tyrrhenian Sea. Within the Southern Apennines, GPS velocities define two laterally juxtaposed belts of deformation, with transpression in the east and transtension in the west. In the east, ~8 mm/yr convergence between northern Murge-Gargano block and the International GPS System (IGS) site MATE is partitioned across ~east-west striking right-lateral faults, consistent with seismicity and with the offshore geological record. To the south, in northern Calabria, site velocities relative to MATE indicate transpression at ~5 mm/yr, not recorded by seismicity but consistent with the on-land and offshore geological record. In contrast, site velocities along the Tyrrhenian Sea coast to the west diverge from MATE at 2-3.5 mm/yr, and are consistent with the crustal extension documented by seismicity and fault slip studies. The transpressional belt is tracked southward across the Ionian Sea by oblique convergence of central Sicily sites (2-6 mm/yr) relative to the IGS site NOTO in the Hyblean block. North-western Sicily sites display clockwise rotation, a pattern reflected in the geological and paleomagnetic record. Ssignificant horizontal motion accompanies rotation in north-western Sicily and is probably accommodated by west-northwest – east-southeast and northeast-southwest-striking right- and left-oblique faults, respectively, consistent with a regional ~north-northwest – south-southeast trending shortening axis. The ~east-west striking belt of contractional earthquakes observed offshore northern Sicily is consistent with up to ~10 mm/yr geodetic convergence between Sicily and Sardinia. Southern Sardinia sites exhibits differential velocities relative to the IGS site CAGL, suggesting internal deformation which is not recorded by seismicity and might signal incipient fragmentation of the Sardinia margin in response to relative convergence with Sicily.

KEY WORDS: GPS velocities, active orogenic deformation, transpression and transtension fronts, Southern Italy, Sicily, southern Sardinia. RIASSUNTO Deformazione attiva in Italia meridionale, Sicilia e Sardegna meridionale da velocità GPS della Rete Geodetica Peri-Tirrenica (PTGA). Misure di velocità GPS effettuate in tre campagne (1995, 1997 e 2000) su siti distribuiti nell’area peri-tirrenica (Sicilia nord-occiden-

(1) Dipartimento di Scienze della Terra, Università degli Studi di Napoli «Federico II», Largo San Marcellino, 10 - Napoli, Italy. E-mail: [email protected] (2) Department of Geological Sciences, University of Idaho, Moscow, Idaho, US. (3) Istituto per l’Ambiente Marino costiero, CNR, Napoli, Italy. (4) Dipartimento di Geologia e Geodesia, Università di Palermo, Italy. (5) Dipartimento di Scienze per l’Ambiente, Università Parthenope, Napoli, Italy.

tale e sud-orientale, Sardegna meridionale, Campania occidentale, Puglia settentrionale e Calabria settentrionale) documentano la deformazione orogenetica attiva con significativi movimenti differenziali dei blocchi tettonici in cui tali siti sono localizzati. In un sistema di riferimento incentrato sulla placca europea, tutti i siti in Italia meridionale e Sicilia sono caratterizzati da rapidi movimenti verso i quadranti settentrionali, ma con significativa variabilità interna che sottolinea la presenza di distinti domini di deformazione orogenetica. Per apprezzare meglio la deformazione attiva all’interno dei vari segmenti orogenetici e confrontare le velocità GPS con i dati di deformazione supra-crostale (meccanismi focali di terremoti, breakouts in pozzo, analisi cinematica delle faglie), le velocità dei siti sono state riferite rispetto a stazioni permanenti (MATE, Matera in Italia meridionale; NOTO, Noto in Sicilia; e CAGL, Cagliari in Sardegna) della rete IGS (International GPS Service). In Italia meridionale, le velocità GPS definiscono due fasce di deformazione affiancate, con transpressione ad est sul margine adriatico-ionico di Puglia, Lucania e Calabria, e transtensione a ovest sul margine tirrenico di Campania e Calabria. Nel settore orientale, viene riscontrata una convergenza (~8 mm/a) relativa tra un blocco composto da Murge settentrionali e Gargano meridionale con il sito IGS di Matera. La convergenza viene ripartita attraverso fasce di deformazione orientate est-ovest e con dislocazione trascorrente destra, in maniera consistente con la sismicità. Un sistema di faglie sepolte nella bassa valle dell’Ofanto assorbe almeno ~5 mm/a di movimento differenziale tra Murge e Gargano; più a nord la faglia di Mattinata separa la parte settentrionale e quella meridionale del Gargano con movimento trascorrente destro di ~2 mm/a. La convergenza obliqua tra le Murge e il blocco di MATE viene presumibilmente assorbita dal margine occidentale murgiano. Verso ovest, la transizione tra i domini di transpressione e di transtensione è marcata da velocità longitudinali rispetto a MATE in Daunia (Puglia occidentale). Sul margine tirrenico, l’Appennino campano è caratterizzato da velocità divergenti rispetto a MATE crescenti verso la costa da ~2 a ~3,5 mm/a, che documentano estensione attiva, normale all’asse della catena, ben testimoniata dalla sismicità e da studi su faglie attive. In Calabria settentrionale, la transizione tra le due fasce di deformazione avviene a ovest dei massicci del Pollino e della Sila. Rispetto a MATE, le velocità dei siti nel settore meridionale del Pollino e nella Sila indicano transpressione rispettivamente a ~5 e ~2 mm/a, in accordo con recenti studi sulla deformazione di terrazzi fluviali e costieri e profili sismici nell’offshore ionico. Sul bordo meridionale del Pollino, la velocità residua di due siti a cavallo della faglia del Pollino-Castrovillari suggerisce ~2 mm/a di estensione obliqua sinistra. Nel massiccio della Sila, un movimento differenziale (~2 mm/a) sinistro tra due siti avviene a cavallo delle faglie di Cecita e dei Laghi. In Sicilia, la fascia di transpressione è suggerita dalla convergenza obliqua (2-6 mm/a) di siti nella parte centrale dell’isola rispetto al sito IGS di NOTO, con deformazione transpressiva presente sul fronte ibleo (~3 mm/a). Le velocità GPS dei siti PTGA in Sicilia nordoccidentale descrivono un movimento di rotazione oraria rispetto a NOTO, un fenomeno consistente con le evidenze geologiche e paleomagnetiche e con velocità GPS pubblicate nel settore orientale del margine tirrenico della Sicilia. I movimenti differenziali tra i siti sono probabilmente accomodati da scorrimento su faglie trascorrenti destre orientate ovest/nordovest-est/sudest e sinistre orientate nordest-sudovest, in risposta ad una asse di raccorciamento regionale orientato ~nord-nordovest/sud-sudest. Questa ricostruzione

19_PT1-187 (299-316)

04-06-2008

7:34

300

Pagina 300

L. FERRANTI ET ALII

è consistente con le evidenze sismlogiche sia a terra che a mare. I terremoti compressivi lungo il margine settentrionale della Sicilia sono compatibili con la convergenza relativa tra Sicilia e Sardegna osservata nelle misure GPS (~10 mm/a). In Sardegna meridionale, sono presenti velocità differenziali rispetto al sito IGS di CAGL che documentano una apprezzabile deformazione interna. Tale deformazione è verosimilmente accumulata ai bordi o all’interno del bacino del Campidano, dove si localizza una zona di taglio destro ad orientazione nordovest-sudest. Le velocità riscontrate in Sardegna meridionale suggeriscono una incipiente frammentazione della parte meridionale del blocco sardo, considerato geologicamente stabile, in risposta alla convergenza relativa con la Sicilia

TERMINI CHIAVE: Velocità GPS, deformazione orogenetica attiva, fronti di transpressione e transtensione, Italia meridionale, Sicilia, Sardegna meridionale.

1. INTRODUCTION

Global Positioning System (GPS) geodetic observations have greatly contributed to the understanding of the displacement field within large subduction or collision systems (e.g. LARSON et alii, 1999; SELLA et alii, 2002). Space-based measurement of site velocities coupled to seismicity data is a fundamental supplement to long-term tectonic reconstructions based on the geological record, and helps validate the feasibility of kinematic models for plate boundary zones (e.g. NORABUENA et alii, 1998; AVÉ LALLEMANT & OLDOW, 2000; MCCAFFREY et alii, 2000; PAUL et alii, 2001). However, where the plate boundaries are diffuse and host relatively small lithospheric fragments, like in the Mediterranean Sea, geologic structures have a far greater complexity, which may be reflected in geodetic velocity fields (e.g. MCCLUSKI et alii, 2003; NOCQUET & CALAIS, 2003). The western-central Mediterranean Sea features small and arcuate-shaped orogens, young oceanic basins superimposed on older fold-thrust belts, and abruptly juxtaposed domains of contraction and extension (e.g. DEWEY et alii, 1989; GUEGUEN et alii, 1998; FACCENNA et alii, 2001). To what degree the deformation pattern established during Cenozoic deformation continues today is, however, not fully resolved. Within the region surrounding the southern Tyrrhenian Sea, the Peri-Tyrrhenian GPS Array, or PTGA (fig. 1), established by Italian (IAMC, CNR-Naples, Dipartimento di Scienze della Terra of Naples, Palermo, Siena and Cagliari Universities) and United States (University of Idaho, Moscow, Idaho) institutions in 1995, has the potential to provide critical insights to this issue. The velocity field derived from three campaigns spanning a 5-year interval (1995-2000) gives a clear outline of active deformation processes in the region evolved from destruction of the Apenninic and Siclian passive margins. Southern Apennines and Sicily sites exhibit displacements toward Europe but display a significant internal variability, indicative of fragmentation of the Sicilian and Adriatic block (OLDOW et alii, 2002). Whereas published geodetic studies have dealt with the regional scale pattern of crustal deformation across this diffuse plate boundary (ANZIDEI et alii, 2001; OLDOW et alii, 2002; HUNSTAD et alii, 2003; HOLLENSTEIN et alii, 2003; D’AGOSTINO & SELVAGGI, 2004; SERPELLONI et alii, 2005), the relation of surface velocities to inter-seismic strain accumulation rates at the scale of individual faults or fault arrays has

not been established. As a matter of fact, surface geodetic velocities measured during the inter-seismic period largely contain information about the motion of crustal blocks and elastic strain accumulation surrounding faults in the upper crust, but not the long-term behaviour of the bulk lithosphere. Combination of site velocities with local seismicity and structural studies, as discussed in this paper, indicates that, although broadly driven by convergence between Europe and Africa, the pattern of displacement in the peri-Tyrrhenian orogens is strongly influenced by intra-plate anisotropy and crustal deformation processes. 2. REGIONAL TECTONIC SETTING

Within the central Mediterranean, contractional and extensional belts evolved simultaneously during Tertiary convergence between Europe and Africa (DEWEY et alii, 1989). Rapid Neogene migration of deformation fronts across the western Adriatic and northern Sicilian margins resulted in growth of the Apennine and Sicily orogens to the east and south, and stretching of the hinterland in the Tyrrhenian Sea to the west (fig. 1). Within this setting, development of coeval contractional and extensional belts is thought to have been driven by crustal delamination and slab roll-back (e.g. ROYDEN et alii, 1987; GUEGUEN et alii, 1998; FACCENNA et alii, 2001; FERRANTI & OLDOW, 2005). The Tyrrhenian basin grew from Miocene to Quaternary across the suture zone of the Alpine and Apennines orogens (fig. 1). The basin progressively widened from northwest to southeast following the roll-back of the Adriatic-Ionian slab (MALINVERNO & RYAN, 1986; DOGLIONI 1991), and is mainly floored by thinned continental crust with two oceanic patches in its southern part at water depths in excess of 3 km (KASTENS et alii, 1988). The Sardinia-Corsica block to the west (fig. 1) is a detached and rotated sector of the Alpine foreland including a small fragment of the western Alps in Northern Corsica (ALVAREZ, 1972; MONTIGNY et alii, 1981; VIGLIOTTI & LANGENHEIM, 1995). Collision of the block with the western Adriatic margin occurred in the mid-Miocene (PATACCA et alii, 1990). Since then, Sardinia and Corsica have been affected by extensional tectonics related to Tyrrhenian Sea rifting (ASSORGIA et alii, 1997; CASULA et alii, 2001), and at present are thought to be stable. The Sicily fold and thrust system (CATALANO et alii, 2000) rotated clockwise up to 120° during Neogene spreading in the Tyrrhenian Sea (CHANNELL et alii, 1990). Pliocene-Quaternary southerly propagation of the thrust system (TORELLI et alii, 1998; LICKORISH et alii, 1999) was accompanied by extensional and strike-slip displacements in northern Sicily (OLDOW et alii, 1990). Extension and contraction episodes alternated in the northern Sicily margin, and offshore Tertiary half-grabens experienced a complex history of tectonic inversion (CATALANO et alii, 2000; PEPE et alii, 2005). The Hyblean foreland in southern Sicily (fig. 1) is underlain by Mesozoic-Cainozoic carbonates which exhibit broad wavelength folds and are transected by strike-slip faults (COGAN et alii, 1989; TORELLI et alii, 1998; MONACO et alii, 2003). The Southern Apennines underlie the southern part of mainland Italy (fig. 1) and experienced easterly shortening directed toward the Apulian region and coeval

19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 301

ACTIVE DEFORMATION IN THE PERI-TYRRHENIAN AREA FROM GPS VELOCITIES

301

Fig. 1 - Generalized tectonic map of the periTyrrhenian area showing the main contractional (teeth toward up-trhown side), extensional (ball toward downthrown side), and transcurrent structures (redrawn from BIGI et alii, 1992), the schematic boundary between foreland contractional and hinterland extensional domain (thick white dashed line), and the geometry of the PTGA. Geodetic site domains A to F are enclosed in regions with 3.8) dai cataloghi Harvard CMT [19762006] http://www.seismology.harvard.edu/CMTsearch.html) e Mednet RCMT [1997-2006] (http://mednet.ingv.it/events/QRCMT/Welcome.h tml; PONDRELLI et alii, 2002; 2004), e da GASPARINI et alii (1985), and ANDERSON & JACKSON (1987)]. Epicentri dei terremoti del periodo 1981-2002 dal database dell’Istituto Nazionale di Geofisica e Vulcanologia (INGV; CASTELLO et alii, 2002). La linea grigia tratteggiata marca il limite tra sismicità transpressiva e transtensiva.

pressional earthquakes, the southern part including the Murge and Salento blocks lacks significant seismicity (fig. 2). Notwithstanding, northeast-southwest convergence between southern Apulia and the contractional orogen in western Greece and Albania across the southern Adriatic Sea occurs at ~5 mm/yr (HOLLENSTEIN et alii, 2003; SERPELLONI et alii, 2005). Seismicity wanes south of Apulia until eastern Calabria, where strike-slip and thrust earthquakes are recorded along the continental margin and into the Ionian Sea (fig. 2). The transpressional deformation front is mapped on the sea-floor by seismic-reflection profiles (DOGLIONI et alii, 1999), and turns westerly and crosses southern Sicily north of the Hyblean foreland. In Sicily, fault plane solutions (NERI et alii, 2005) and borehole breakouts (RAGG et alii, 1999) indicate a mixture of thrusting and strike-slip motions with northwest-southeast compressive axes both in the thrust belt to the north and in the Hyblean foreland to the south (fig. 2). Offshore northern Sicily, an ~east-west trending belt of crustal compressional earthquakes with northwest-southeast P-axes and 10-20 km focal depths stretches toward the Sardinia Strait (fig. 2; ANDERSON & JACKSON, 1987; PONDRELLI et alii, 1995; 2002). Geodetically-determined shortening between Sicily and Sardinia is estimated at ~1.5-2.0 mm/yr (SERPELLONI et alii, 2005) and has been taken to accommodate a large fraction of EU-NU convergence (D’AGOSTINO & SELVAGGI, 2004). In contrast to the other segments of the peri-Tyrrhenian area, onland Sardinia lacks historical and instrumental seismicity (fig. 2; BOSCHI et alii, 2000).

domains encompasses long-established tectonic regions, and specifically the western extensional province, the Apenninic frontal thrust belt and northern Apulian margin, the southern Apennines-Calabrian arc boundary, the Sicilian frontal thrust system and Hyblean foreland, the Tyrrhenian margin of northwestern Sicily, and southern Sardinia (fig. 1). All the sites consist of steel monuments drilled and cemented into hard bedrock with the location of the site established to monitor differential motion between known or suspected tectonic blocks. Geological criteria have been pivotal in site selection, favouring locations on structurally coherent blocks and sufficiently away from active faults. Baselines within clusters are consistently lesser than 100 km (fig. 1), which removes positioning uncertainties associated with geographic differences in atmospheric conditions and their impact on microwave propagation. Baselines between clusters vary from about 100 to 650 km, and clusters are linked by simultaneously occupied sites. Thirty PTGA sites were occupied in 1995, of which 24 were recovered and occupied in 1997 and 2000. Sites not recovered were either destroyed or inaccessible. Data acquisition was made with three Leica SR299 receivers and two Trimble 4000 SSE receivers in 1995, six Leica SR399 receivers in 1997, and six Leica SR9500 receivers with choke-ring antennas in 2000. In the first campaign, sites were occupied for at least 10 to 12 hours on successive days. In the second and third campaign, sites were occupied continuously for 24 to 48 hours (tab. 1). 5. DATA PROCESSING AND ANALYSIS

4. GEOMETRY OF THE PERI-TYRRHENIAN GPS ARRAY AND DATA ACQUISITION

The Peri-Tyrrhenian GPS Array (PTGA) consists of 24 GPS sites distributed in six clusters across southern mainland Italy, Sicily, and Sardinia (fig. 1). Location of

All observations were processed with BERNESE 4.2 from the Astronomicsches Institut Universität Bern (HUGENTOBLER et alii, 2001) operating on a HP-UNIX platform at the University of Idaho. BERNESE utilizes baseline double-difference processing in which common

19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 303

ACTIVE DEFORMATION IN THE PERI-TYRRHENIAN AREA FROM GPS VELOCITIES

error sources such as satellite ephemerides, atmospheric delays, and clock errors are eliminated (WELLS et alii, 1986). Ocean tide loading corrections were applied, carrier phase ambiguity was resolved using a quasi-ionosphere-free strategy, and cycle-slips were detected and repaired for each session. Daily observations were divided into 8 to 12 hour sessions for processing to improve assessment of positioning repeatability. Changes in receiver and antenna, antenna heights, and phase center corrections between campaigns are accounted by using calibration tables (HUGENTOBLER et alii, 2001). The PTGA observations were collected at 15 second epochs and processed together with 30 second data from continuously monitored IGS (International GPS Service) sites at Matera (MATE) in southern Italy, at Noto (NOTO) in southern Sicily, and at Cagliari (CAGL) in southern Sardinia. We used precise satellite ephemerides and earth rotation parameters available from the IGS, transformed into the ITRF00 frame using the trfsp3 program developed at the Natural Resources Canada Geodetic Survey Division (J. KOUBA, written communication, 2000). For each of the occupations, the coordinates of the European IGS reference sites were calibrated using ITRF00 coordinates and velocities referenced to the mean epoch of the campaign. ITRF00 velocities for the PTGA are presented in tab. 2. Time series for the 24 PTGA site coordinates show the stability of the ITRF00 velocity estimates for each site over the five year time span of observations (fig. 3). Each time series plot is normalized to the mean position of the site calculated in the ITRF00 reference for the 1997 campaign. Linear regressions were determined using positions determined for each ~8 hour session weighted by the formal error. Sixteen sites have R2 values ≥0.95 for both north and east components, three sites have one component that falls between 0.90 and 0.95, 3 sites with one component lying between 0.85 and 0.90, and in the worst case two sites have one component lying between 0.80 and 0.85. All of the sites in Southern Italy (A, B, and C) have R2 values greater than 0.93 for both components and show linear changes in position with time. In Sicily, the velocities are linear, but two sites, D1 and E1, show decreased stability in the north component with R2 values of 0.86 and 0.80, respectively. For Sardinia, linear velocities are also prevalent, but two sites (F1 and F5) have north components with R2 values of 0.88 and one (F2) has a north component with an R2 value of 0.81. We estimated velocity uncertainties using a propagation of coordinate uncertainty method. Contributions to velocity uncertainty from monument random walk (LANGBEIN & JOHNSON, 1997; DIXON et alii, 2000) are negligible because all PTGA sites are drilled and cemented into bedrock, and as such we assume the random walk component to the uncertainty budget is zero (SELLA et alii, 2002). Similarly, correlated noise is greatly reduced by the double-difference baseline-mode processing (MAO et alii, 1999) employed by BERNESE 4.2 where ambiguity resolution is achieved (ZUMBERGE et alii, 1997). Formal root mean square values for coordinate uncertainty calculated by BERNESE 4.2 were scaled by the ratio of the sigma of the coordinate group over the a posteriori sigma of unit weight to provide a more realistic estimate of position error (HUGENTOBLER et alii, 2001). The scaled coordinate uncertainties from each session solution were propagated to provide a very conservative

303

TABLE 1 PTGA sites and yearly observations. – Identificazione dei siti PTGA e ore di osservazione per campagna.

Station ID

Station name

Region

A1 A2 A3 A4 B1 B2 B3 B4 B5 C1 C2 C3 C4 D1 D2 D3 D5 E1 E3 F1 F2 F3 F4 F5

Tempa di Don Giovanni S. Giuliano M. Tobenna S. Maria del Granato M. La Serra M. Aquilone la Murgetta M. Serbarola Celle S. Vito M. la Serra Cozzi dell’Anticristo Luzzi Lago Cecita Pizzo Crastone Belmonte Mezzagno C.da le Rocche C.da Castagnola M. Judica Monumento Conservatore P.ta Ruinas Teulada M. Urpino M. Campulongu Baccu Scovas

Campania Campania Campania Campania Apulia Apulia Apulia Apulia Apulia Calabria Calabria Calabria Calabria Sicily Sicily Sicily Sicily Sicily Sicily Sardinia Sardinia Sardinia Sardinia Sardinia

yearly observation hours* Sept. Sept. June 95 97 00 46 26 24 12 23 11 23 15 22 24 45 26 12 9 42 18 17 19 10 23 16 12 32 55

22 20 24 24 23 47 20 23 23 22 21 40 22 25 25 21 18 24 9 24 43 20 24 24

48 48 23 24 24 24 25 24 50 25 24 24 24 22 20 23 24 22 44 25 24 48 23 24

estimate of the velocity error that was modified by a gain factor g, formulated by BROCKMANN (1996), that is related to the time span between occupations: g = (3k2 – 3k + 1)1/2 where k is the total span in years (BROCKMANN, 1996). The total time span for our three observations is 5 years. The PTGA results were aligned with Stable European Reference Frame (SERF) by inclusion of IGS sites BOR1, JOZE, PENC, POTS, and WTZR, all of which have ≤0.5 mm/yr velocity residuals with respect to one another (NOCQUET et alii, 2001; NOCQUET & CALAIS, 2003; SERPELLONI et alii, 2005). Solutions from all sessions were combined to produce a total velocity field (tab. 2) determined by constraining the velocities and coordinates of selected SERF sites to their ITRF00 values, then performing a 6-parameter Helmert transformation of the combined network to minimize residuals (HUGENTOBLER et alii, 2001). 6. RESULTS

GPS velocities for the PTGA sites are presented in a fixed European frame (fig. 4) and in three local frames (tab. 3) by fixing the IGS sites MATE, NOTO, and CAGL (figs. 5, 6, 7). The fixed European frame shows the differential displacement of the peri-Tyrrhenian sites with respect to stable Europe and provides insight into the relation between displacement within the orogen and motion between the European and African plates (OLDOW et alii, 2002; OLDOW & FERRANTI, 2006).

19_PT1-187 (299-316)

304

04-06-2008

7:34

Pagina 304

L. FERRANTI ET ALII

Fig. 3a. Fig. 3 - Time series of horizontal components of non-permanent GPS station position: a) sites A1 to B4; b) sites B5 to D3; c) sites D4 to F5; d) reference sites MATE, NOTO, CAGL. – Serie temporali delle componenti orizzontali della posizione di siti GPS non-permanenti: a) siti A1-B4; b) siti B5-D3; c) siti D4-F5; d) siti di riferimento MATE, NOTO, CAGL.

In a fixed European reference frame, displacements in the southern peri-Tyrrhenian belt are dominated by north to northeast GPS velocities ranging from 5 to 12 mm/yr (fig. 4). The velocity field defined by PTGA and IGS sites varies around the southern Tyrrhenian Sea, with greatest magnitude displacements localized in Sicily and southern Italy. The IGS site in Sardinia (CAGL) has a small to statistically insignificant east velocity with respect to stable Europe, but the PTGA network in southern Sardinia records velocities with variable orientations up to 5 mm/yr. In northwestern Sicily, velocities are north to northeast with magnitudes reaching 10 mm/yr. To the southeast near the Hyblean foreland, velocities of sites

straddling the contractional front are northerly and vary between 3 to 8 mm/yr. In southern Italy, GPS site velocities show large spatial variability and range from near zero along the Tyrrhenian coast to magnitudes up to 12 mm/yr with northeast and locally northwest azimuths in the Apennines and near the Adriatic coast. The results presented here are broadly consistent with the pattern of displacement recorded by other networks of permanent and campaign sites in the region (ANZIDEI et alii, 2001; HOLLENSTEIN et alii, 2003; SERPELLONI et alii, 2005). Although a joint velocity field has not been determined, the comparison of IGS site velocities (in a fixed European frame) for MATE, NOTO, and CAGL, calculated

19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 305

ACTIVE DEFORMATION IN THE PERI-TYRRHENIAN AREA FROM GPS VELOCITIES

305

Fig. 3b.

independently using different definitions of the stable European frame, shows that the PTGA results are in good agreement with the results of other studies (NOCQUET & CALAIS, 2003; SERPELLONI et alii, 2005). Using published and our estimates of velocity uncertainties at 95% confidence, the site velocities of MATE and NOTO calculated during the PTGA campaigns are statistically indistinguishable from the results for the same sites determined by SERPELLONI et alii (2005). For the site CAGL on Sardinia, the north components of the velocities are statistically equivalent but the east component of the PTGA velocity is 0.5 mm/yr greater. When compared to the results of NOCQUET & CALAIS (2003), PTGA velocities are typically within 0.5 mm/yr with the greatest deviation recorded being 1.0 mm/yr for the easterly component of CAGL. Overall, the PTGA velocities in the European frame clearly depict impingement of southern Italy and Sicily

into the European continental margin in northern Italy and its southern extension in Sardinia (OLDOW et alii, 2002; SERPELLONI et alii, 2005). Divergence between the PTGA velocities and the Nubia-Europe relative motion, which occurs south of Sicily at ~4 mm/yr along a ~northwest-southeast direction (fig. 4), suggests the existence of a distinct microplate in the southern peri-Tyrrhenian region. Nevertheless, when viewed in the stable European frame, the PTGA velocity field is difficult to reconcile with deformation recorded by active tectonic structures and by the belts of seismicity within the orogen, because incremental strain axes associated with earthquakes and active fault slip record differential motion between fault blocks. To aid comparison of earthquake focal mechanisms and the GPS displacements, we recomputed the GPS velocities in local reference frames. For southern

19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 306

306

L. FERRANTI ET ALII

Fig. 3c.

Italy, we fix on the IGS site MATE, for Sicily develop a velocity field fixed on NOTO, and for Sardinia present the velocities in a frame fixed on CAGL 6.1. ACTIVE DEFORMATION IN SOUTHERN ITALY Velocities and related uncertainty ellipses of southern Italy sites are shown in fig. 5a with MATE held fixed. Within this reference frame, a different velocity pattern is readily apparent between the Tyrrhenian and the Adriatic-Ionian margins of the orogen. Specifically, divergent and convergent velocities in the MATE reference frame are found on the Tyrrhenian and Apulian margin, respectively; the transition between the two domains is narrow and marked by orogen-parallel velocities (fig. 5a). When viewed relative to MATE, sites located on the Tyrrhenian margin of Campania move toward the southwest at rates which increase from ~2 to 3.5 mm/yr from

east to west (fig. 5a). Geodetic deformation is reflected in the tensile axis of extensional earthquakes, slip vector of active normal faults, and maximum elongation of borehole breakouts in the region (fig. 5b). Site A2 is located in the hanging-wall of the the Irpinia Fault (fig. 5b), which slipped during the Mw=6.9, 1980 earthquake (PANTOSTI et alii, 1993). The small divergence between site A2 and MATE at ~1-2 mm/yr (fig. 5a) suggests additional extensional strain may be accumulating east of the Irpinia fault. This finding is consistent with the existence of active faults east of this location, some of which have slipped during the 1980 and 1930 earthquakes (WESTAWAY, 1992; PANTOSTI et alii, 1993). The velocity increase between site A2 in the mountain chain and sites A3 and A4 along the Tyrrhenian Sea coast places a ~1-2 mm/yr bound on the extension accommodated across the belt of active faults along the Apennines axis (fig. 5b). Because of the little relative motion between

19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 307

ACTIVE DEFORMATION IN THE PERI-TYRRHENIAN AREA FROM GPS VELOCITIES

sites A1 and A2, however, displacement of site A4 must be largely accommodated west of site A1. No known active normal fault is mapped west of site A1, and aseismic slip might occur on faults which root at relatively shallow depth along old detachment surfaces between imbricate thrust sheets (fig. 5b), which are mapped in crustal seismic profiles (MAZZOTTI et alii, 2000). In northern Calabria, site C3 moves south-southeast away from MATE, and, in contrast, site C4 shows oblique convergence toward MATE at ~2 mm/yr. Whereas site C4 is located in a region dominated by transpressional earthquakes, extensional first-motions and active normal faults are found west of site C3 (fig. 5c). Thus, these two sites straddle the transition between the active extensional and transpressional belts in this sector of the orogen (figs. 5a). The ~2 mm/yr left-oblique differential motion between sites C3 and C4 occurs across the Lake and Cecita faults (fig. 5c), and is consistent with paleoseismological evidence of left strike-slip on these faults (GALLI & BOSI, 2003). Whereas the Lake fault is thought to have slipped during the M=6.7 1638 earthquake, an elapsed time of 1000 yr or more characterizes the Cecita fault (GALLI & BOSI, 2003). Because the two PTGA sites are located across the silent fault, strain accumulation reflected in GPS velocities suggests the potential of a near-future event. The boundary between transpressional and extensional belts is tracked further north by velocity of sites C1 and C2 which are moving toward MATE at ~4-5 mm/yr (fig. 5a). These two sites are located along-strike of a belt of diffuse transpressional faults mapped in seismic reflection profiles offshore the Ionian Sea (fig. 5c; DOGLIONI et alii, 1999). Along the coastal projection of the deformation front, Middle Pleistocene and younger marine terraces are uplifted and locally folded (BIANCA & CAPUTO, 2003; MAZZELLA et alii, 2006), and ~east-west striking folds are mapped in Middle Pleistocene sediments of the San Arcangelo basin north of the two sites

Fig. 4 - PTGA velocity field in a fixed European reference frame. Also shown are velocities of Nubia in European frame from SELLA et alii (2002) and MCCLUSKI et alii (2003). – Campo di velocità PTGA nel sistema di riferimento europeo. Sono anche mostrate le velocità di Nubia rispetto ad Europa (da SELLA et alii, 2002; e MCCLUSKI et alii, 2003).

307

Fig. 3d.

(fig. 5c; CASCIELLO et alii, 2000). Thus, motion of sites C1 and C2 relative to MATE is consistent with ongoing shortening in southern Lucania Apennines and offshore the Ionian Sea.

19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 308

308

L. FERRANTI ET ALII

TABLE 2 PTGA site velocities (mm/yr) in ITRF00 and in a stable European fixed frame. Uncertainties are presented as 1σ in north and east components. – Velocità dei siti PTGA (in mm/a) nei sistemi di riferimento ITRF00 e europeo. Le incertezze per le componenti N e E sono espresse come 1σ. ITRF00

Uncertainties

Central Europe

Site

Latitude

Longitude

VN

VE

σN

σE

VN

VE

A1 A2 A3 A4 B1 B2 B3 B4 B5 C1 C2 C3 C4 D1 D2 D3 D5 E1 E3 F1 F2 F3 F4 F5 CAGL MATE NOTO

40.47 40.69 40.71 40.45 41.78 41.58 41.16 41.14 41.33 39.85 39.89 39.44 39.37 38.02 38.06 37.90 37.72 37.50 36.96 39.54 38.95 39.22 39.14 39.51 39.14 40.69 36.88

15.35 15.48 14.84 15.05 15.60 15.77 16.04 15.35 15.18 16.09 16.09 16.34 16.53 12.86 13.37 12.80 13.27 14.65 14.69 8.60 8.81 9.14 9.51 9.52 8.97 16.70 14.99

18.70 17.28 15.65 14.22 17.55 17.12 18.44 15.93 21.86 21.11 19.62 14.60 17.28 20.56 17.56 20.51 18.69 16.91 20.76 15.39 19.77 11.50 19.62 12.52 14.33 17.85 18.50

21.35 21.85 20.73 22.05 23.01 21.29 16.22 25.13 21.29 21.97 23.18 21.26 21.57 22.45 21.71 19.64 21.77 22.50 19.89 20.76 20.69 21.53 22.63 21.09 21.26 23.35 20.39

0.35 0.41 0.45 0.45 0.49 0.49 0.45 0.45 0.55 0.60 0.45 0.51 0.53 0.45 0.35 0.66 1.01 0.45 0.78 0.45 0.47 0.45 0.53 0.37 0.19 0.27 0.27

0.35 0.49 0.53 0.41 0.78 0.70 0.35 0.27 0.41 0.49 0.39 0.41 0.57 0.45 0.43 0.66 1.31 0.45 0.80 0.53 0.47 0.41 0.70 0.33 0.19 0.19 0.19

5.07 3.13 2.10 0.97 4.72 4.10 4.90 3.64 8.69 9.42 7.69 2.55 5.37 6.08 4.02 7.03 5.28 3.44 7.19 –0.28 3.47 –4.39 3.25 –2.32 0.29 4.47 5.15

1.19 0.19 –0.87 1.56 1.11 –0.86 –6.28 2.75 –1.08 0.79 3.55 1.60 3.80 1.22 1.41 –1.76 7.59 1.44 –0.66 0.17 0.12 1.00 3.62 2.04 1.51 2.23 –0.93

TABLE 3 PTGA velocities (mm/yr) calculated in local reference frames fixed on IGS reference sites MATE, NOTO, and CAGL. Uncertainties are listed in tab. 2. – Velocità dei siti PTGA (in mm/a) nei sistemi di riferimento locali centrati sui siti IGS di riferimento MATE, NOTO e CAGL. Le incertezze sono riportate nella tab. 2. MATE

NOTO

CAGL

Site

Latitude

Longitude

VN

VE

VN

VE

VN

VE

A1 A2 A3 A4 B1 B2 B3 B4 B5 C1 C2 C3 C4 D1 D2 D3 D5 E1 E3 F1 F2 F3 F4 F5 CAGL MATE NOTO

40.47 40.69 40.71 40.45 41.78 41.58 41.16 41.14 41.33 39.85 39.89 39.44 39.37 38.02 38.06 37.90 37.72 37.50 36.96 39.54 38.95 39.22 39.14 39.51 39.14 40.69 36.88

15.35 15.48 14.84 15.05 15.60 15.77 16.04 15.35 15.18 16.09 16.09 16.34 16.53 12.86 13.37 12.80 13.27 14.65 14.69 8.60 8.81 9.14 9.51 9.52 8.97 16.70 14.99

0.62 –1.32 –2.37 –3.50 0.31 –0.32 0.48 –0.80 4.25 4.97 3.24 –1.91 0.93 – – – – – – – – – – – – 0.00 –

–1.03 –2.03 –3.07 –0.65 –1.11 –3.08 –8.51 0.54 –3.29 –1.45 1.32 0.64 1.55 – – – – – – – – – – – – 0.00 –

– – – – – – – – – – – – – 0.73 –1.32 1.69 –0.06 –1.88 1.91 –5.60 –1.86 –9.72 –2.09 –7.66 –5.32 – 0.00

– – – – – – – – – – – – – 2.12 2.31 –0.86 8.47 2.29 0.17 1.33 1.29 2.15 4.75 3.17 2.18 – 0.00

– – – – – – – – – – – – – – – – – – – –0.55 3.18 –4.66 2.98 –2.58 0.00 – –

– – – – – – – – – – – – – – – – – – – –1.40 –1.45 0.55 2.08 0.50 0.00 – –

19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 309

ACTIVE DEFORMATION IN THE PERI-TYRRHENIAN AREA FROM GPS VELOCITIES

309

Fig. 5 - a) PTGA velocity field in the Southern Apennines in a fixed MATE reference frame. The dashed white line marks the boundary between convergent and divergent geodetic velocities relative to MATE; b) to d) active deformation in western Campania, northern Calabria and northern Apulia, respectively. PTGA velocities in MATE reference frame (fig. 5a). Focal mechanisms of crustal earthquakes as in fig. 2. Sh minimum orientation in borehole break-outs after AMATO & MONTONE (1997). Faults, dotted where buried/inferred (sources: PANTOSTI et alii, 1990; CINTI et alii, 1997; MONACO & TORTORICI, 2000; GALLI & BOSI, 2003; MASCHIO et alii, 2005): CF=Cecita Fault; IF=Irpinia Fault; LF=Lake Fault; MF=Mattinata Fault; MMF=Monti della Maddalena Fault; MuF=Murge border Fault; OF=Ofanto Fault; PCF=Pollino-Castrovillari Fault; VAF=Val d’Agri Fault; VCF=Valle del Crati Faults; VDF=Vallo di Diano Fault. Shown in fig. 5c the schematic fold axis developed in Middle to Late Pleistocene fluvial and marine terraces and deposits after (1) CASCIELLO et alii (2000), (2) BIANCA & CAPUTO (2003), and (3) MAZZELLA et alii (2007). SAB=Sant’Arcangelo Basin. – a) campo di velocità PTGA in Appennino meridionale nel sistema di riferimento centrato su MATE. La linea tratteggiata bianca segna il limite tra velocità geodetiche convergenti e divergenti rispetto a MATE; b, c, d) deformazione attiva in Campania occidentale, Calabria settentrionale e Puglia settentrionale. Velocità PTGA nel sistema di riferimento di MATE (fig. 5a). Meccanismi focali di terremoti crostali come in fig. 2. Orientazione dell’asse principale minimo di deformazione in pozzo da AMATO & MONTONE (1997). Faglie (puntinate quando sepolte o supposte (fonti: PANTOSTI & VALENSISE, 1990; CINTI et alii, 1997; MONACO & TORTORICI, 2000; GALLI & BOSI, 2003; MASCHIO et alii, 2005); CF=Faglia di Cecita Fault; IF=Faglia dell’Irpinia; LF=Faglia dei Laghi; MF=Faglia di Mattinata; MMF=Faglia dei Monti della Maddalena; MuF=Faglia del bordo delle Murge; OF=Faglia dell’Ofanto; PCF=Faglia Pollino-Castrovillari; VAF=Faglia della Val d’Agri; VCF=Faglie della Valle del Crati; VDF=Faglia del Vallo di Diano. Sono anche mostrate in fig. 5c le tracce assiali schematiche di pieghe in terrazzi e depositi fluviali e marini del Pleistocene medio-superiore, da (1) CASCIELLO et alii (2000), (2) BIANCA & CAPUTO (2003) e (3) MAZZELLA et alii (2007). SAB=Sant’Arcangelo Basin.

19_PT1-187 (299-316)

310

04-06-2008

7:34

Pagina 310

L. FERRANTI ET ALII

Fig. 6 - a) PTGA velocity field in Sicily in a fixed NOTO reference frame. Sh maximum orientation in borehole break-outs after RAGG et alii (1999). Faults, dotted where buried/inferred (sources: BIGI et alii, 1992): TF, Frontal thrust of the Sicilian-Maghrebian Chain; SRF, ScicliRagusa Fault; GTN=Gela thrust nappe; b) active deformation in northwestertn Sicily, PTGA velocities in NOTO reference frame (fig. 6a). Focal mechanisms of crustal earthquakes as in fig. 2, with the addition of some solutions from FREPOLI & AMATO (2000). Faults after CATALANO et alii (2005a,b). MKR=Monte Kumeta ridge; RBR=Rocca Busambra ridge; MG=Monte Galiello. – a) campo di velocità PTGA in Sicilia nel sistema di riferimento centrato su NOTO. Orientazione dell’asse principale massimo di deformazione in pozzo da RAGG et alii (1999). Faglie (puntinate quando sepolte o supposte) (fonti: BIGI et alii, 1992): TF, thrust frontale della catena siculomaghrebide; SRF, Faglia Scicli-Ragusa; GTN=Falda di Gela; b) deformazione attiva in Sicilia nordoccidentale. Velocità PTGA nel sistema di riferimento di NOTO (fig. 6a). Meccanismi focali di terremoti crostali come in fig. 2, con l’aggiunta di alcuni meccanismi da FREPOLI & AMATO (2000). Faglie da CATALANO et alii (2005a,b); MKR=Monte Kumeta Ridge; RBR=Rocca Busambra Ridge; MG=Monte Galiello.

On the other hand, the two sites are located across a major normal fault (Pollino-Castrovillari fault, fig. 5c) which had prehistoric activity (CINTI et alii, 1997). Although near the limit of detection given uncertainties, the small residual velocity between sites C1 and C2 suggests ongoing left transtensional displacement across the Pollino-Castrovillari fault at ~2 mm/yr, which is consistent with the Holocene (~1 mm/yr; CINTI et alii, 1997) and Pleistocene (3-4 mm/yr; COLELLA & CAPPADONA, 1988) rate estimates. The transpressional belt is tracked to the north by velocities of the Apulian margin sites, which show a significant component of oblique convergence and lateral motion with respect to MATE (fig. 5a). The transition between diverging and converging velocities is marked along the foothills of the Apennines by orogen-parallel motion of site B5 at ~5 mm/yr relative to MATE, consistent with a belt of strike-slip earthquakes recorded beneath the foothills and with the locus of strike-slip faults inferred in the subsurface (fig. 5d). The eastern Apulian sites in Gargano and northern Murge are moving westward relative to MATE (fig. 3) with velocities which taper toward the north from ~8 mm/yr (site B3 in Murge) to virtually zero (site B1 in Gargano). Rapid displacement of Murge reflected in site B3 motion is not recorded by seismicity (fig. 2), and might indicate substantial strain accumulation across the deep faults underlying the plateau escarpments (e.g. MuF, fig. 5d). The progressive northerly decrease in westward motion of sites B3, B2 and B1 relative to MATE suggests differential displacements along ~east-west trending strike-slip faults mapped or inferred across these sites. About 5 mm/yr differential displacement between B3 and B2 must be accommodated by structures buried beneath the Ofanto Valley (OF or sub-parallel faults, fig. 5d), which should be considered significant seismogenic sources (e.g. DISS Working Group, 2006).

Further to the north, differential motion between sites B1 and B2, straddling across the western part of the east-west striking Mattinata Fault (MF, fig. 5d), which dissects the Gargano block, yields an estimation of ~2 mm/yr maximum right-lateral slip. This finding matches the recent slip behaviour of the fault as suggested by paleoseismological studies (PICCARDI, 1998) and seismicity (fig. 2). 6.2. ACTIVE DEFORMATION IN SICILY GPS site velocities in the Sicily domains D and E are shown as residuals to the IGS site NOTO (tab. 3; fig. 6a). Motion of sites in southern Sicily suggests active displacement along the buried frontal thrust system (TF, fig. 6a). Site E1, located beyond the frontal thrust, exhibits a slightly oblique convergence with NOTO at ~3 mm/yr, in a direction consistent with the orientation of the shortening axis in borehole break-outs (fig. 6a). Within the Hyblean foreland, the large uncertainty on site E3 velocity does not allow inference on the kinematics of the Scicli-Ragusa fault, where geomorphologic studies (CATALANO et alii, 2006) suggest recent left-lateral motion superposed on a main right-lateral slip (MONACO et alii, 2003). Geodetic site velocities in north-western Sicily (D1, D2, D3 and D5) relative to the IGS site NOTO exhibit a complex horizontal velocity field (fig. 6a). Whereas site D5 displays a large eastward motion relative to NOTO, a clockwise block rotation is highlighted by motion of sites D1, D2 and D3. Lateral displacement of PTGA sites relative to NOTO might be accommodated by right- and left-lateral motion on ~northwest-southeast and northeast-southwest trending structures mapped or inferred between these structural domains (fig. 6b, CATALANO et alii, 2000; 2004; 2005a,b; AVELLONE & BARCHI, 2003). Relative westward motion between sites D1, D2 and D3 is

19_PT1-187 (299-316)

04-06-2008

7:34

Pagina 311

ACTIVE DEFORMATION IN THE PERI-TYRRHENIAN AREA FROM GPS VELOCITIES

311

Fig. 7 - a) PTGA velocity field in southern Sardinia in a fixed CAGL reference frame. Faults, dotted where buried/inferred (sources: CASULA et alii, 2001 and BIGI et alii, 1992, for the onland and offshore sector, respectively): FF=Fangario Fault; b) deformation field in southern Sardinia inferred from PTGA velocities and faults. The grey lines show the schematic trend and displacement of fault systems which bound velocity blocks. – a) Campo di velocità PTGA in Sardegna meridionale nel sistema di riferimento centrato su CAGL. Faglie, puntinate quando sepolte o supposte (fonti: CASULA et alii, 2001 e BIGI et alii, 1992 per le faglie a terra e a mare, rispettivamente): FF=Fangario Fault; b) campo di deformazione in Sardegna meridionale desunto dalle velocità PTGA e dalle faglie. Le linee nere e le freccette mostrano schematicamente l’orientazione e il senso di scorrimento di faglie che limitano blocchi a differente velocità.

consistent with right-oblique displacement on a diffuse belt of west-northwest – east-southeast striking faults mapped between Monte Galiello-Rocca Busambra and Monte Kumeta (fig. 6b) and further west along the Tyrrhenian coastline, where they displace upper Pleistocene marine terraces (MAUZ & RENDA, 1995). The geodetic velocity field and the geological displacement resolved on the faults in the region is consistent with focal mechanisms of moderate earthquakes in the region (fig. 6b). The focal solutions of the 5