Regional tectonic framework

The peri-Mediterranean orogenic belts, including the Apennines, evolved within the overall framework of Africa–Eurasia major plate convergence in Late Cretaceous to Quaternary times (e.g. Mazzoli & Helman 1994), accompanied by the motion of minor blocks such as Corsica–Sardinia (of European origin) and Adria (of African affinity). These mountain belts are characterized by markedly curved geometries with crustal or lithospheric extension affecting the concave side of the arcs and, in the case of the Apennines, the axial zone of the chain (Ascione et al. 2013, and references therein).

The structure of the Apennines fold and thrust belt in the Marche region, involving a Mesozoic to Neogene sedimentary succession and the underlying basement, has been widely documented by field surveys and by the interpretation of seismic reflection profiles tied with deep well logs (e.g. Coward et al. 1999; Mazzoli et al. 2005, and references therein). Stratigraphic work on synorogenic deposits unravelled a clear foreland-ward (i.e. northeastward) progression of thrusting during the Neogene (Fig. 2a and b). Along the Adriatic coast and offshore Marche, the analysis of growth strata associated with thrust-related fold amplification reveals that their youngest age post-dates the Early–Middle Pliocene, whereas post-Early Pleistocene deposits offshore (i.e. at the thrust front) appear not to be involved in thrusting (Coward et al. 1999; Di Bucci & Mazzoli 2002; Tozer et al. 2006; Fig. 2c and d).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Subsurface data (location of profiles is shown in Fig. 1); (a) and (b) are from Tozer et al. 2006; (c) and (d) are from Di Bucci & Mazzoli 2002). (a) Line drawing of a seismic reflection profile across the coastal area of southern Marche and adjacent offshore. Open triangles represent hydrocarbon exploration wells; the main thrusts are indicated as T3 and T5. TWTT, two-way travel time. (b) Deformation chronology for the two thrust-related anticlines (Bellante and Costiera) imaged in (a). Biostratigraphic ranges are shown for key Pliocene and Pleistocene Foraminifera in the Marche Apennines and Adriatic Sea. FCO, first common occurrence; FO, first occurrence; G., Globorotalia; LCO, last common occurrence; LO, last occurrence; mP, middle Pliocene. (c) Detail of seismic line across the thrust front offshore northern Marche, and (d) interpretation. (Note planar, horizontal reflectors ascribed to the Middle–Upper Pleistocene (white), overlying slightly deformed reflectors ascribed to a sandy facies (light grey) and more deformed reflectors ascribed to a clayey facies (dark grey) of the Lower Pleistocene.)

Seismotectonics of the Marche region and the 2013 offshore earthquakes

The active tectonic behaviour of the Marche region has been intensely investigated (e.g. Di Bucci et al. 2003; Vannoli et al. 2004). The well-documented active NE–SW extension characterizing the axial zone of the Apennine chain is well established also in this region (Macchiavelli et al. 2012, and references therein). On the other hand, although all workers agree on the occurrence of active deformation in the outer zone of the fold and thrust belt, the size of active structures, and consequently the maximum expected earthquake magnitude, and their kinematics are still a matter of debate. Furthermore, based on the seismological constraints available before the 2013 earthquakes, the active tectonic behaviour of the outer Apennine sector could not be straightforwardly interpreted in terms of a simple geodynamic model (e.g. Macchiavelli et al. 2012).

The seismicity of the Marche region predating the 2013 earthquakes, as recorded by the INGV seismic network over the past 30 years (1983–2012; http://bollettinosismico.rm.ingv.it/, http://csi.rm.ingv.it/), is characterized by low- to medium-magnitude (M_w ≤ 5.0), dominantly shallow events that tend to cluster along the Apennine belt and adjacent foothills. Deep earthquakes (located at depths locally in excess of 40 km) are few and scattered. On the other hand, no earthquake with magnitude M_w > 5 occurred in the Adriatic offshore in recent times (e.g. Macchiavelli et al. 2012, and references therein). Owing to the low level of seismicity in the area, geoscientists over the years have tried to analyse in detail even moderate seismic activity and to obtain focal mechanisms from low- to medium-magnitude earthquakes. The analysis carried out by Macchiavelli et al. (2012) showed the occurrence of an area dominated by a horizontal maximum compression along the Adriatic coast and adjacent offshore of the Marche region. This domain, including both thrusting and strike-slip tectonic regimes, is characterized by variable orientations of maximum compression, which outline a complex tectonic behaviour. In fact, previous to the 2013 earthquakes, both NW–SE and NE–SW compression were recorded, suggesting that the complex, inherited fault structure characterizing the investigated area (Coward et al. 1999) could trigger significant local variations of the stress field resolved by seismological analysis.

In the last few years, a large effort has been dedicated to the compilation of the Database of Individual Seismogenic Sources (DISS) in Italy (Basili et al. 2008; DISS Working Group 2010). According to the DISS, seismogenic sources, both individual and ‘composite’, the latter encompassing several individual seismogenic sources, consist in the study area of SW-dipping thrust faults roughly parallel to the coastline (Fig. 1). The ‘composite seismogenic source’ labelled ITCS008 (Conero onshore) has a maximum assigned Mw of 5.8, whereas that labelled ITCS031 (Conero offshore) has a maximum assigned Mw of 5.9. This latter ‘composite seismogenic source’ (Conero offshore) has been indicated as being responsible for the 2013 earthquakes in the first reports to the media in Italy.

Most of the epicentres of the 2013 earthquakes, including the main shock, cluster along a WSW–ENE-trending orientation, close to the strike of one of the two nodal planes shown by the focal mechanisms for the three largest strike-slip events. Available locations show hypocentral depths in the range of 0.4–10.7 km (mostly 3.5–10.7 km). In a SW–NE-oriented vertical section, roughly perpendicular to the strike of the thrust faults indicated in the DISS as the major seismogenic sources in area, the hypocentres do not show any particular trend in depth distribution, being roughly uniformly distributed along the section (Fig. 3).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

SW–NE vertical section (location shown in Fig. 1b) showing earthquake hypocentres projected onto a plane normal to the thrust faults identified as ‘composite seismogenic structures’ in the DISS. These are displayed using the parameters defined in that catalogue (ITCS008: depth 3–6.5 km, dipping 45° to the SW; ITCS031: depth 2–7 km, dipping 40° to the SW; Basili et al. 2008; DISS Working Group 2010). Basement geometry after Fantoni & Franciosi (2010).

Discussion and concluding remarks

Detailed seismic reflection interpretation coupled with high-resolution biostratigraphy indicates that thrust-related fold growth in the offshore sector of the thrust belt, in the position affected by the recent earthquakes, occurred mainly in Pliocene times (being sealed by lowermost Pleistocene sediments; Fig. 2a and b) and that even the thrust front farther to the NE is sealed by undeformed sediments of the past 1 Ma (Di Bucci & Mazzoli 2002; Di Bucci et al. 2003; Fig. 2c and d). Although such evidence does not rule out the possible reactivation of Pliocene thrust faults within the present stress field, the associated late Quaternary cumulative displacement would be below seismic resolution. This confirms that limited or no thrust reactivation has occurred during the late Quaternary. Nevertheless, in recent years several studies suggested that blind thrust faults, accompanied by anticlinal growth, characterize the active tectonics of the Marche coastal area and the adjacent Adriatic offshore (e.g. Vannoli et al. 2004). According to this view, blind thrust faults would be responsible for the occurrence of the largest earthquakes in the Adriatic coastal area. Following this line of reasoning, NW–SE-striking, SW-dipping active thrusts have been considered as the only seismogenic sources to be included in the DISS for the Marche offshore (Basili et al. 2008; DISS Working Group 2010; Fig. 1b). The seismological evidence provided by INGV preliminary data on the 2013 Marche offshore earthquakes is consistent with a horizontal, NE–SW-trending active maximum compression, as envisaged by the DISS compilers. However, even for this theoretically favourably oriented compression, only minor reverse faulting is suggested by the fault plane solutions (Fig. 1) and, in any case, the unique (Mw = 4.2) thrust-related event did not involve rupturing along any of the ‘composite seismogenic sources’ envisaged in the DISS (Fig. 3). The epicentre of this earthquake is located along the alignment defined by the largest strike-slip events (Fig. 1b), thus suggesting that reverse fault rupturing occurred within the rock volume surrounding a major wrench zone, probably as a subsidiary process associated with overall strike-slip deformation. Furthermore, earthquake epicentres located NE of the leading edge of the DISS outer thrust fault (Conero offshore; Fig. 1b) are located in an area (the Adriatic monocline in Fig. 2a) where no NE-vergent thrusts or folds are present, as clearly shown by geological sections based on a large subsurface dataset acquired by the oil industry in the Conero offshore area (Fantoni & Franciosi 2010).

The integrated analysis of preliminary epicentre distribution, hypocentre depth and fault plane solutions (Figs 1 and 3) suggests that coseismic slip occurred dominantly along a steep, ENE–WSW-striking left-lateral fault. The Mw = 4.9 main shock is located at what appears to be the western termination of a steep earthquake rupture (Fig. 1), thus suggesting that stress concentration occurred at a laterally propagating transcurrent fault tip. The envisaged active structure, as defined by the alignment of the three main strike-slip events shown in Figure 1b, is at a high angle to the thrust faults mapped in the DISS, cutting right across the alleged ‘composite seismogenic sources’ labelled ITCS008 (Conero onshore) and ITCS031 (Conero offshore). Taking into account the along-strike continuity of the thrust faults mapped in the DISS and the large number of seismic events extending to the NE, well beyond the upper tip line of the outer thrust (Conero offshore), it is clear that the main seismogenic structure responsible for the 2013 earthquakes cannot be interpreted as merely a transfer or tear fault forming part of a segmented thrust system. Rather, it represents a high-angle fault dissecting the thrust belt. Furthermore, hypocentre distribution indicates that most of the earthquakes, including the Mw = 4.9 main shock, were nucleated within the basement (Fig. 3). Therefore, earthquake nucleation occurs mainly in the crystalline upper crust, below the sedimentary cover and the envisaged lower boundary of the ‘composite seismogenic sources’ included in the DISS (Basili et al. 2008; DISS Working Group 2010; Fig. 3). That catalogue, placing the seismogenic fault segments at shallow depths, essentially within the sedimentary cover, also fails to provide a consistent picture of the dominant earthquake nucleation depth and embedding medium.

The 2013 Marche earthquakes confirm that the active tectonic behaviour of the Adriatic segment of the northern Apennines is different from that of the Po Plain segment, where thrust faults accommodate roughly similar SSW–NNE-oriented maximum horizontal compression (as testified by the 2012 northern Italy earthquakes; http://www.globalcmt.org/CMTsearch.html; Fig. 1a). Epicentre distribution (Fig. 1b) and hypocentral depths (Fig. 3) for the 2013 Marche earthquakes define a c. 25 km long, upper crustal seismogenic structure that we term the Conero offshore wrench fault (Fig. 1a). It is worth noting that active strike-slip faulting, showing a similar distribution of shocks at a high angle to the thrust belt, is recorded also by the Mw = 4.6, 1972 Ancona earthquake (Fig. 1b).

An enhanced seismotectonic model for the Adriatic sector of the outer northern Apennines should take into account the fundamental role played by such deep-seated, probably inherited and long-lived, steep crustal faults characterized by dominant strike-slip kinematics, similarly to the Gargano area to the south (e.g. Di Bucci & Mazzoli 2002), although with a different orientation of the maximum compressive stress axis. This poses new challenges to the mapping of seismogenic sources in the Adriatic sector of the outer northern Apennines and to the definition of their dimensional and kinematic parameters, which in turn has major implications for a correct assessment of earthquake hazard in the densely populated coastal area. In fact, the actual earthquake potential could differ substantially from that estimated for the, several tens of kilometres long, continuous thrust faults identified in the DISS as the only ‘composite seismogenic sources’ along the Adriatic offshore. On the other hand, the blind fault model envisaged in the DISS for thrust-type seismogenic sources is unlikely to be applicable for strike-slip faults along which earthquakes in the 0.4–10.7 km depth range are nucleated, based on available locations of earthquake hypocentres (Fig. 3). These faults could also offset the uppermost part of the sedimentary cover (although obviously earthquake nucleation is uncommon in the top 3 km). This may require a reassessment of the risks associated with, for example, installations on top of or below the seabed (oil or gas pipelines, etc.), besides stimulating a reappraisal and/or acquisition of relevant datasets such as seismic reflection profiles parallel to the strike of the thrust belt and across transverse seismogenic faults.