In this paper, we present a significant update of the Italian present-day stress data compilation not only to improve the knowledge on the tectonic setting of the region or to constrain future geodynamic models, but also to understand the mechanics of processes linked to faulting and earthquakes. In this paper, we have analysed, revised and collected new contemporary stress data from borehole breakouts and we have assembled earthquake and fault data. In total, 206 new quality-ranked entries complete the definition of the horizontal stress orientation and tectonic regime in some areas, and bring new information mainly in Sicily and along the Apenninic belt. Now the global Italian data set consists of 715 data points, including 499 of A–C quality, representing an increase of 37 percent compared to the previous compilation. The alignment of horizontal stresses measured in some regions, closely matches the ~N–S first-order stress field orientation of ongoing relative crustal motions between Eurasia and Africa plates. The Apenninic belt shows a diffuse extensional stress regime indicating a ~NE–SW direction of extension, that we interpret as related to a second-order stress field. The horizontal stress rotations observed in peculiar areas reflect a complex interaction between first-order stress field and local effects revealing the importance of the tectonic structure orientations. In particular, in Sicily the new data delineate a more complete tectonic picture evidencing adjacent areas characterized by distinct stress regime: northern offshore of Sicily and in the Hyblean plateau the alignment of horizontal stresses is consistent with the crustal motions, whereas different directions have been observed along the belt and foredeep.
The World Stress Map Project compilation and several single papers have clearly demonstrated the existence of first-order stress field (plate-scale) controlled by plate boundary forces, and second-order stress field (regional) controlled by major intraplate stress sources (Müller et al. 1992; Zoback 1992; Hillis & Reynolds 2000; Tingay et al. 2010b). In areas where high data density is present, a third-order stress field (local) can also be recognized linked to the presence of minor features (i.e. active faults, local inclusions, detachment horizons or density contrasts; Bell 1996; Yale 2003; Tingay et al. 2005). Such features may explain stress rotations that can be found not only between two adjacent wells but also in the same well along depth (e.g. Mariucci et al. 2002; Carminati et al. 2010) as observed when fractures, active faults or mechanically weak zones are crossed by, or close to, a borehole (e.g. Bell et al. 1992; Barton & Zoback 1994; Bell 1996; Pierdominici et al. 2005, 2011; Tingay et al. 2011).
The comprehension of the present-day stress state in tectonically dynamic regions is significant to understand the mechanics of various geological/geophysical processes including faulting and earthquakes (Zoback 1992; Sassi & Faure 1996; Suppe 2007). The determination of in situ stress is also one of the most important scientific objectives of deep drilling projects because very often the state of stress in the first 5–10 km of the crust is not revealed by any other data. No deep scientific boreholes have been drilled in Italy to date, but opportunity exists to analyse data from numerous petroleum industry boreholes (e.g. ENI S.p.A.) from throughout the country.
In this paper, we present the most complete and updated contemporary stress map of Italy with particular attention to new data from Sicily, an important region both for past large earthquakes and for oil and gas exploitation. We then depict the pattern of crustal stress emphasizing areas both where we recognize a first- or second-order present-day stress field and where we measure local variations recording the third-order stress field.
The data set contains 715 stress indicators (499 best-quality, 101 low-quality and 115 discarded data) including: borehole breakouts in deep wells, single earthquake focal mechanisms, formal inversions of focal mechanisms, faults and overcoring data (Fig. 1). A quality ranking between A and E is assigned to each stress data, with A being the highest quality. Following the most up-to-date ranking scheme (Heidbach et al. 2010), A-quality indicates that the stress orientation is accurate within ±15°, B-quality within ±20°, C-quality within ±25° and D-quality within ±40°. When data do not provide sufficient information or have standard deviations greater than 40°, the data are discarded (E-quality). We use A-, B- and C-quality stress indicators for analysing first-order stress patterns (Heidbach et al. 2010) although we also consider D-quality data (actually only breakouts) to define second- or third-order stress field (Montone & Mariucci 1999; Pierdominici et al. 2005; Mariucci et al. 2008, 2010) as observed in other studies in the world (Tingay et al. 2010b).
In the location map (Fig. 2), we report all the new data (see also Tables 1 and 2) to give a general overview of their distribution. Stress information derived from breakouts is widespread along the Italian territory particularly along the Po Plain–Adriatic foredeep, along the southern belt and foredeep and in Sicily; the Apenninic belt is well covered by several moderate magnitude earthquakes and some seismic sequences as well as Sicily and specific areas of northern Italy; finally, fault data usable for stress analysis are located along the central-southern Apennines.
Depth interval of the entire data set is between 0 and 40 km including faults, borehole breakouts and earthquake data. The results are reported in terms of minimum horizontal stress (Shmin) both in map (Fig. 3) and in the tables. Concerning breakouts their preferred orientations correspond to Shmin; because all the considered faults are normal faults, we assume the Shmin direction as perpendicular to the fault strike, assuming an Andersonian faulting and stress model. Because our seismological data set is mainly characterized by dip-slip faulting mechanisms, the use of principal axis directions (P-, T- and N-axes) can provide a reliable description of the regional stress orientation, consistent with the World Stress Map assumptions (Zoback 1992).
Data from new borehole breakout analysis
Borehole breakouts are stress-induced enlargements of the wellbore cross-section that occur when a well is drilled in rocks in an anisotropic stress field (Bell & Gough 1979; Zoback et al. 1985). They develop in two opposite zones of the borehole wall and are aligned along the direction of the minimum horizontal stress. In this work, we determine breakouts by using ‘four-arm calliper’ tool following the main criteria reported in Plumb & Hickman (1985): that is, the tool rotation ceases along the breakout zone; one of the two callipers is longer than the other one, approximately equal to the bit-size; the hole deviation is less than 15° and the azimuth of breakout is different (by more than 10°) from the hole azimuth. Usually, the breakout zones are plotted as a rose diagram, with bars proportional to cumulative breakout lengths. We apply the circular statistics of Mardia (1972) to compute, for each borehole, the mean breakout orientation and its standard deviation (95 percent confidence interval) weighed by length of breakout intervals. The WSM quality ranking system (Heidbach et al. 2010) used to classify the breakout orientation of each well, from A to E, takes into account the number of breakout zones, the total length of breakouts and the variability of the measurements (standard deviation).
Fig. 2 shows the location of the wells where new breakout analyses have been performed (black and grey dots). We have analysed 20 new wells in Sicily; 3 new wells located in northern Italy and along the Adriatic margin; revised all the 46 data previously used to constrain finite element models (Barba et al. 2010); entered 1 well crossing the M= 6.9 Irpinia fault (Pierdominici et al. 2011); included 2 wells recently analysed after the L’Aquila 2009 earthquake (Mariucci et al. 2010) and finally, we added 2 revised wells located in central Apennines (Mariucci et al. 2008). In total, we have 74 new entries, 36 of which have reliable A, B and C quality; the minimum depth of breakout data is 450 m and the maximum depth is 5585 m (Table 1).
Data from other sources
In our data set we include also other stress indicators as focal mechanism of earthquakes, formal inversions from diffuse minor seismicity and fault data to define the tectonic regime, explore deeper depth intervals and expand the study areas.
Earthquake focal mechanisms
The previous compilation by Montone et al. (2004) collected data up to 2003 June. Now we have considered the period from 2003 July to 2010 December and we have replaced previous Quick Regional CMT (http://autorcmt.bo.ingv.it/quicks.html) with the new computations (Pondrelli et al. 2004, 2007, 2010, 2011; RCMT European–Mediterranean Catalog, http://www.bo.ingv.it/RCMT). We take into account CMT-like solutions of earthquakes with M≥ 4 and crustal depth within the first 40 km (Table 2). The systematic error that can be associated to principal axis directions (P, T and N) from CMT-like solutions is ±14° (Helffrich 1997). Moreover, although focal plane solution principal axes could not be indicative of stress axes, the possible differences between Shmin derived from P-, T- and N-axes and Shmin from slip vectors lie within the error of the attributed quality as shown in Montone et al. 2004.
Taking into account the systematic error of CMT-like solutions and the range of stress orientations that would be consistent with each focal mechanism and that the orientation of the P-, N- and T-axes could deviate from the principal stress orientations (McKenzie 1969), the World Stress Map project (Heidbach et al. 2010) assigns C-quality to focal mechanism data (stress orientations within ±25°). Furthermore, in this paper, the use of principal axis directions (P-, T- and N-axes) is reasonable because we apply this data set to provide regional stress orientations and also we can be confident of other stress indicators.
Concerning the quality value, after the previous compilation (Montone et al. 2004), the criteria suggested by the World Stress Map Project have changed. In the last release (), the quality ranking scheme is more stringent and, especially, all single focal mechanisms have to be considered C quality (stress orientation range ±25°) even if with high magnitude, then all the earthquakes with B quality of the previous compilation have now C quality.
New data are broadly spread along the Italian peninsula (125 new entries), with two areas remarkably improved: Sicily on- and offshore and the Abruzzo region.
The 2009 L’Aquila seismic sequence has improved our data set with: (i) 25 moderate-sized earthquakes showing normal faulting mechanisms with T-axis NE-oriented (Table 2; no. 47–61 and 63–74; Pondrelli et al. 2010); (ii) NW-oriented surface breaks, not reported in the map and (iii) two deep wells providing a ~ENE Shmin orientation (Table 1; no. 7, 9; Mariucci et al. 2010).
Stress orientation determined from inversions of P-, T- and N-axes of diffuse seismicity has also been updated with the inclusion of three recent data located in Sicily, in the Hyblean Plateau (Musumeci et al. 2005). The formal inversions refer to seismic events from 1994 to 2002 with magnitude ML 1.0–4.6 and depth interval between 5 and 26 km; IBL1, IBL2 and IBL3 inversions include 22, 23 and 25 events, with an average misfit of 4.5°, 6.0° and 5.7°, respectively (Fig. 2).
As already described in Montone et al. (2004), we do not include faults for which focal mechanisms are available. For this reason we do not enter the 2009 L’Aquila earthquake surface breaks. Whereas, we have considered four new data (Table 3) relative to Melandro–Pergola fault (Moro et al. 2007), Aquae Iuliae fault (Galli & Naso 2009), southern Vallo di Diano fault (Villani & Pierdominici 2010) and San Pio fault (Di Bucci et al. 2011), although the activity of this latter is still under debate.
The geodynamic setting of the Italian region is characterized by a complex interaction of different processes (Fig. 2), mainly related to the continental collision between Africa and Eurasia plates and the subduction of Adria microplate beneath the Alps (to the north), Dinarides (to the east) and Apennines (to the west). Although Adria is recognized as one of the main features in the central Mediterranean region, its geometry and kinematics are not well defined yet. In fact, according to Anderson & Jackson (1987), Adria would be an independent microplate moving with a counter-clockwise rotation with respect to Eurasia around an Eulerian pole located in northern Italy: this would explain the tectonic extension along the Apenninic belt and the compression in the Eastern Alps and Dinarides. Recently D’ Agostino et al. (2008), on the basis of a new geodetic model, separate Adria in two main blocks, with two different poles of rotations. The boundary between the blocks would be along the Gargano–Dubrovnik tectonic structure (Westaway 1990; Calais et al. 2002; Oldow et al. 2002; Battaglia et al. 2004), where a diffuse seismicity is recorded. In accordance with this model, the two microplates could explain almost completely the complex tectonic regime along the peninsula. In this context, the Adriatic slab beneath the backarc Tyrrhenian basin, that has caused the migration towards east of extension–compression system along the Apenninic thrust and fold belt (Malinverno & Ryan 1986; Royden et al. 1987; Patacca & Scandone 1989; Frepoli & Amato 1997), would assume a less important role with respect to the Neogene evolution of this area, except for the northern Apennines and Calabria arc. In fact, as inferred from deep and intermediate earthquakes and seismic tomography (Selvaggi & Amato 1992; Chiarabba et al. 2005; Cimini & Marchetti 2006; Chiarabba et al. 2008), the subduction is still active in the southern Tyrrhenian Sea and, possibly, beneath the northern Apennines; whereas, along the central-southern Apennines most of the authors believe that the subduction has ended (e.g. Meletti et al. 2000; Di Stefano et al. 2009).
In this framework, we distinguish a first-order present-day stress field, associated to Eurasia–Africa Plate motion, in northern Italy (Friuli, Po Plain and Adriatic margin) and in Sicily; a second-order (characterized by a minor wave length) stress field along the extensional Apenninic belt (from Tuscany to Calabria) related to Adria microplate and third-order effects widely spread along the coastal Tyrrhenian margin or localized close to some tectonic structures (i.e. the Irpinia fault, Pierdominici et al. 2011).
Present-day stress data, mainly from seismicity, depict a compressive area from the eastern Alps throughout the Po Plain–Adriatic foredeep with two main stress orientations recording the ~N–S and the ~NE–SW oblique compression (Fig. 3). In detail, fault plane solutions in Friuli region and subordinately along the Po Plain foredeep are consistent with the Eurasia–Africa convergence, namely the northward push of Adria microplate beneath Europe. Along the Po Plain and the southern Alps front, breakout results confirm Shmin~E–W oriented, approximately parallel to the Alpine topographic front, with several localized stress rotations also with depth, related to a sequence of active minor structural arcs (Montone & Mariucci 1999; Selvaggi et al. 2001; Burrato et al. 2003; Pierdominici et al. 2005; Carminati et al. 2010). According to Reinecker et al. (2010), the same pattern, following the belt trend, is observed in the Molasse Basin (Switzerland and German Alps), confirming the first-order stress field and the role of the deep structure of the Alpine belt in stress rotations.
Along the Adriatic foredeep, close to the northern Apennine thrust front, data indicate NE–SW compression (Shmin~NW–SE oriented) consistent, as mentioned earlier, with a still active subducting or sinking slab at depth (Selvaggi & Amato 1992).
On the contrary, in the central part of the foredeep, the remarkable stress change (around latitude 43°N) is linked to a different tectonic regime: Shmin orientations quickly rotate from NW to NE (Fig. 3) evidencing a no longer active compression through the southern foredeep (around latitude 40°N). In this area the absence of intermediate/deep seismicity and clear seismic tomography images suppose that the Adriatic slab is interrupted (Amato et al. 1993; Lucente et al. 1999; Cimini & Marchetti 2006).
In the inner part of central Italy (Umbria region), the Apenninic belt is characterized by a diffuse extension NE oriented. In particular, the contemporary extension–compression pair evidenced also by our data is probably related to the passive slab retreat (e.g. Malinverno & Ryan 1986; Selvaggi & Amato 1992; Frepoli & Amato 1997; Lucente et al. 1999; Faccenna et al. 2001; Scrocca et al. 2003; Cimini & Marchetti 2006; Di Stefano et al. 2009).
Tectonic extension (NE oriented) is also well defined by the data set (seismicity, breakouts and faults) recently achieved in the Abruzzo region, after the destructive Mw 6.3, 2009 earthquake.
Towards south, NE–SW Shmin orientations are well represented and stable along the central-southern Apennines belt and foredeep where several earthquake focal mechanisms depict mainly an extensional and subordinately strike-slip regime, respectively. This section of Italy is characterized by the lack of intermediate earthquakes and also by a continuous high-velocity anomaly depicting the subducted slab. Today is not verified yet if the tomographic images are related to a detached slab or a less pronounced velocity/ density anomaly (Amato et al. 1993; De Gori et al. 2001).
Along the Calabria region crustal stress indicators show an extensional regime with a radial pattern of stress orientations always perpendicular to the main tectonic structures. At depth, the evidence of a subducting slab is very clear from both seismological and geological data, with a well-defined geometry characterized by a narrow, long, downdip compression slab of oceanic Ionian lithosphere (amongst many others, Piana Agostinetti & Amato 2009 and reference therein).
In the Gargano area, data from seismicity have largely increased with respect to the previous map, indicating NE-oriented Shmin mainly related to compressional and strike-slip tectonic regime (i.e. E–W right–lateral nodal planes, no. 81 and 82 in Table 2; Fig. 3). These data can strengthen the hypothesis of a tectonic lineament, ~ENE-oriented from Gargano to the Dinarides, possibly splitting the Adria microplate in two blocks (Westaway 1990; Calais et al. 2002; Oldow et al. 2002; Battaglia et al. 2004; D’Agostino et al. 2008). In central Adriatic, new data related to compressive earthquakes highlight another possible disengagement area of the Adria.
New breakout data along the Tyrrhenian margin (from Rome towards 40°N) show two different Shmin orientations (Fig. 3): a prevalent NE and subordinately NNW, revealing a poor constrained current stress field that we believe is linked to similar horizontal stress magnitudes. We associate this stress pattern to the presence of a local stress field characterized by a predominant vertical stress.
The most significant increase in the data set is in Sicily both regarding breakouts and earthquakes (Fig. 4). In this case, to discuss the stress pattern of the region we have also plotted D-quality breakout data to highlight local stress field effects. In fact, as several authors have argued (e.g. Tingay et al. 2010a,b) D-quality data may best reflect local stress variations and sometimes may provide insight to larger scale stress fields. In particular along the northern offshore of Sicily (Kabylian–Calabrian thrust front), several earthquakes, with prevalent reverse focal mechanisms, describe a clear ~NNW–SSE current compression (Fig. 4). In southeastern Sicily (Fig. 4), breakout results along the Hyblean foreland are all consistent with ENE-oriented Shmin (Ragg et al. 1999; Montone et al. 2004). It is worth noting that although presently the Hyblean Plateau shows a low level of instrumental seismicity, in the past strong earthquakes have occurred with a magnitude up to 7.4 (http://emidius.mi.ingv.it/CPTI04). A detailed analysis of earthquakes (1.0 ≤ML≤ 4.6), recorded from 1994 to 2002 in southeastern Sicily (Musumeci et al. 2005), has evidenced NNW–SSE compression from predominately strike-slip focal mechanisms. From these data, the authors have computed three formal inversions of earthquakes, whose results are included in this paper (IBL1, 2 and 3 in Fig. 4), indicating a stress regime characterized by Shmin ENE oriented. Then, contemporary stress indicators along the active Kabylian–Calabrian thrust front and within the Hyblean foreland showing a consistent NNW–SSE maximum horizontal stress orientation are in agreement to global plate-motion studies (NUVEL-1, De Mets et al. 1990) that predict a ~340°N convergence between Africa and Eurasia. The localized compression in these two areas points out that the northward Africa push dominates in this part of Sicily.
Along the Gela thrust front (no longer active) breakouts describe its complex structure with several rotations that follow the curved trend of the front itself, almost always pointing perpendicularly to it (Fig. 4). Similar stress rotations have been also observed along active thrust fronts, for instance, offshore NW Borneo collisional margin (King et al. 2010) and along the fold and thrust mountain range in New Guinea (Hillis & Reynolds 2000).
Some normal faulting earthquakes showing a WNW–ESE extension suggest that eastern Sicily and southern Calabria (Fig. 4) are dominated by an incipient rifting (Catalano et al. 2008). Extension is commonly interpreted as the result of different directions and rates of motion between Sicily and Calabria (Hollenstein et al. 2003; D’Agostino & Selvaggi 2004).
We also highlight the occurrence of several strike-slip focal mechanisms (~E–W-oriented T-axis) between the Aeolian Arc and Etna volcano (Fig. 4) that can strengthen the hypothesis of a transfer zone in this area (Billi et al. 2010). Towards south, the NNW striking Malta escarpment (Bianca et al. 1999; Argnani & Bonazzi 2005 and reference therein) delineates the boundary between the Ionian Sea and the Hyblean foreland (Casero & Roure 1994,;Nicolich et al. 2000). Results from multichannel seismic data show that active extensional faults (NNW–SSE trending) are present only in the Malta escarpment northern portion (Argnani & Bonazzi 2005). Few data from earthquakes indicate in this area a predominant strike-slip component with ~NE-oriented T-axis supposing the coexistence of extensional and strike-slip deformation.
This paper presents the latest significant updating and complete collection of data on the present-day stress orientations in Italy. The achieved map (Fig. 3) can be employed by many users that not only work on this topic and/or related ones such as geophysical modelling, seismic hazard assessment, rock mechanics laboratory experiments, deep drillings, but also on oil and gas well production and construction of nuclear waste deposits.
In some areas of Italy (Sicily, Friuli and Po Plain), the alignment of horizontal stresses closely matches the ~N–S direction of ongoing crustal motions with respect to stable Europe. This result can be associated to the first-order stress field that drives the plate movement. Along the entire Apenninic belt, from north to south, a diffuse extensional stress regime is clearly showed by a large and broad data set indicating a NE–SW direction of extension, probably related to a second-order stress field.
The stress rotations observed in some areas (i.e. Po Plain minor arcs and Gela thrust front in Sicily) reflect a complex interaction between first-order stress field and local effects that perturb the large-scale regional stress, revealing the importance of the tectonic structure orientations. In particular, in Sicily the new data delineate a more complete tectonic picture highlighting adjacent areas characterized by distinct stress regimes. In this study, the use of stress data of all qualities has been essential for identifying small-scale stress variations. The small-scale changes in the stress orientations, and in some cases also in tectonic regime, over distances of a few tens of kilometres indicate complex t'ectonic processes and interactions amongst different stress field orders.
This work has been carried out in the framework of the following Projects: FIRB Project ‘Research and Development of New Technologies for Protection and Defense of Territory from Natural Risks’, W.P. C3 ‘Crustal Imaging in Italy’ coordinated by PM, funded by Italian Ministry of University and Research; INGV-DPC Project S1, coordinated by S. Barba and C. Doglioni, funded by Italian Presidenza del Consiglio dei Ministri – Dipartimento della Protezione Civile (DPC). ENI S.p.A. is thanked for providing borehole data. We would like to thank Alessandro Amato for a careful review of the paper and Silvia Pondrelli for the discussion on earthquake focal mechanisms. Many thanks are also due to two anonymous referees for detailed comments that largely improved the manuscript.