Corresponding author: A. Turtù, Dipartimento di Ingegneria e Geologia, Università “G. D'Annunzio” di Chieti-Pescara, Via dei Vestini 31, 66100, Chieti, Italy. (firstname.lastname@example.org)
 We report on a paleomagnetic study of the southern sector of the Olevano-Antrodoco-Sibillini (OAS) thrust front, which corresponds to the southern limb of the Northern Apennines (Italy) orogenic salient. A lively debate has developed regarding the oroclinal/progressive-arc versus non-rotational nature of the OAS, which has been alternatively interpreted as a dextral strike-slip fault, dextral transpressive fault, or frontal to oblique ramp that reactivated pre-existing Jurassic normal faults. Here, we document the paleomagnetism, integrated with biostratigraphic and structural data, of 52 new sites from both the OAS hanging wall and footwall. On the basis of 39 retained sites, we find a peculiar pattern of tectonic rotations along the OAS thrust that evidences four rotational domains. The thrust footwall is characterized by a southern domain that undergoes an approximately 30° counterclockwise rotation with respect to the stable foreland, and an approximately non-rotated domain. The data from the hanging wall indicate the occurrence of a dextral strike-slip component along the southern sector of the OAS thrust supported by a strong clockwise rotation close to the NE-SW lateral ramp, which rapidly fades 1 km from the thrust front. A slight but significant CW rotation observed in the remaining sites from the hanging wall confirms the progressive nature of the OAS, and its structural position as the southern limb of the Northern Apennines salient. Our detailed paleomagnetic study is crucial in discriminating between progressive-arc- and strike-slip-related components in the main curved orogenic front of the Northern Apennines.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Paleomagnetism is the only tool that can quantify the presence and amount of vertical axis rotations; thus, it is often used to unravel the kinematics of both curved belts and strike-slip systems. Paleomagnetism, integrated with structural analysis, constrains kinematic models and thus determines how, when, and why orogens acquired curvature.
 The Northern Apennines fold and thrust belt has been widely investigated over the last 40 years using paleomagnetic techniques. Progressively available paleomagnetic data caused changes in the way authors interpreted the kinematics and geodynamic evolution of this region. In early paleomagnetic studies, the Italian peninsula was interpreted as a single block rotated approximately 40° counterclockwise (CCW) since the late Cretaceous [Lowrie and Alvarez, 1975]. Later on, the variation in the amount of paleomagnetic rotations was related to thrust emplacement [Vandenberg et al., 1978] and to the kinematics of allochthonous units that were deformed during the Neogene [Channell, 1992; Van der Voo, 1993]. Since then, this tectonic interpretation has been confirmed by a large amount of data collected from Mesozoic and Tertiary pelagic deposits [e.g., Speranza et al., 1997; Cifelli and Mattei, 2010, and references therein]. The individual thrust fronts record variable vertical-axis rotations during the middle Miocene-Pliocene time interval, with different senses and amounts of rotation along strike of the arc (Figure 1a). Paleomagnetic rotations vary from CCW in the northern sector of the Northern Apennines Arc [Muttoni et al., 1998] to clockwise (CW) in the central sector, ahead of the Sibillini thrust front [Speranza et al., 1997]. The analysis of vertical axis rotations from Mesozoic to Pleistocene sediments in the external part of the Northern Apennines Arc allowed constraining of the timing of curvature of the belt. The Arc acquired its curvature mostly from the early Miocene to early Pliocene, with rotations that can be attributed to thrust emplacement [Dela Pierre et al., 1992; Speranza et al., 1997; Sagnotti et al., 2000].
 Considering the temporal relationships between thrusting and vertical axis rotations [Weil and Sussman, 2004], curved belts have been divided into (1) primary (or non-rotational arcs) [Eldredge et al., 1985; Marshak, 1988; Zweigel, 1998; Hindle and Burkhard, 1999]; (2) oroclines (or rotational arcs), when linear belts acquire curvature during tectonic deformation [Carey, 1955; Marshak, 1988; Weil et al., 2000]; and (3) progressive arcs that acquire their curvature or a portion of their curvature during progressive deformation or during a subsequent deformation phase [Merle and Brun, 1984; Sussman et al., 2004]. Rotations may be variably distributed along these arcs. In primary arcs, no rotations are recorded around vertical axes; thus, paleomagnetic declinations are parallel along the entire arc and are not influenced by structural setting. Both in oroclines and progressive arcs, paleomagnetic declinations vary along the arc. In curved oroclines, a systematic pattern of rotations is a function of the distance along the length of the arc and is characterized by maximum vertical-axis rotations in the end limbs. Local and unsystematic patterns are recorded from progressive arcs because vertical axis rotations occur during thrust activity and do not vary as a regular function of the structural trend.
 Relying on the interpretations of different paleomagnetic data sets, the orocline hypothesis for the Northern Apennines curve-shaped belt has been alternatively accepted [Channell et al., 1978; Eldredge et al., 1985; Speranza et al., 1997] or rejected [Van der Voo and Channel, 1980; Lowrie and Hirt, 1986; Hirt and Lowrie, 1988]. Recently, a synthesis of available paleomagnetic data indicates a significant correlation between changes in declination and orogenic trend and allows for the interpretation of the Northern Apennines as a progressive arc [Cifelli and Mattei, 2010].
 The causes for orogenic curvature can be attributed to the crustal and/or lithospheric scale factors. The paleogeography of the colliding bodies (both the hanging-wall and footwall shapes) may influence the curvature of an arcuate belt [Weil and Sussman, 2004], depending upon the geometry and strength of the detachment horizon, the configuration of the sedimentary basin, strike-parallel variation in lithology and sedimentary thickness, and buttressing [Mitra, 1997; Lickorish et al., 1999; Macedo and Marshak, 1999; Paulsen and Marshak, 1999; Weil et al., 2010]. Larger-scale influences can be related to crustal-scale wrench-faults, indenters, or the geometry of the lithosphere [Marshak, 1988; Marshak et al., 1992; Cunningham, 1993; Ghisetti et al., 2009; Jolivet and Faccenna, 2000; Platt et al., 2003; Hall et al., 2004]. The bending of the Northern Apennines is driven by the rollback of the subducting lithospheric slab [Lucente and Speranza, 2001; Rosenbaum and Lister, 2004], as confirmed by seismic evidence [Chiarabba et al., 2009], or by mid-upper crustal folding, based on gravity and seismic reflection data [Patacca et al., 2008; Billi and Tiberti, 2009].
 Tectonic rotations accompanied by vertical axes rotations are also commonly observed in many strike-slip systems; thus, paleomagnetism combined with structural geology is a powerful tool to detect the presence of a strike-slip shear component [Luyendyk et al., 1985; Lamb, 1987; Wells and Heller, 1988; Molnar, 1992; Sonder et al., 1994]. Tectonic deformation in strike-slip systems is recorded within fault-bounded blocks that can have different sizes, shapes, mechanical behaviors, and distributions. Rotations documented along transform or transcurrent shear zones on the kilometer scale [Luyendyk et al., 1985; Wells and Heller, 1988; Laj et al., 1989] are expected to be clockwise in regions characterized by dextral shear and counterclockwise in regions characterized by sinistral shear. At the regional or local scale, the kinematics of transcurrent deformation is more complex and can be driven by two main mechanisms: pervasive continuum deformation or rotation of blocks [e.g., Sonder et al., 1994]. In rigid blocks, several mechanisms have been recognized: slip along undeformed crustal blocks; rigid block rotations with the same sense as the slip in the shear zone [Ron et al., 1984; McKenzie and Jackson, 1983]; generalized rotations in the opposite sense with respect to the main strike-slip movement [Garfunkel, 1989]; and small block rotations with variable, but uniform sense [Beck et al., 1986], or systematic rotation [Walcott, 1984].
 A detailed paleomagnetic sampling, done at the appropriate scale and spatial resolution, has been demonstrated to be a fundamental tool in understanding the evolution of minor structures [Satolli et al., 2005; Weil et al., 2010; Yonkee and Weil, 2010]. In fact, despite the huge amount of data collected from the Apennines over the last decades, the kinematics of some structures is still unclear. In particular, the lack of data is evident in the southern part of the Northern Apennines curve-shaped belt, along the Olevano-Antrodoco-Sibillini (OAS) thrust. The role of this fault has long been debated in the literature, particularly concerning the presence of a dextral strike-slip component of shear and its contribution to the observed curvature. The aim of this study is to clarify the nature and kinematics of the southern sector of the OAS thrust front by using the results of new paleomagnetic investigations, integrated with geological and structural data.
2 Geological Background
 The Apennines are a fold-and-thrust belt (Figure 1b) that has continued to develop since the late Oligocene as a consequence of the eastward roll-back of the Adriatic-Ionian slab [Malinverno and Ryan, 1986]. The orogenesis affected Triassic to Miocene sedimentary successions belonging to pelagic and platform domains of the Adria Mesozoic paleomargin [Ben Avraham et al., 1990; Ciarapica and Passeri, 2002; Lentini et al., 2002; Boccaletti et al., 2005; Patacca and Scandone, 2007]. During Neogene-Quaternary times, the Apennines were characterized by the coeval occurrence of normal and thrust faults along the western and eastern belt margins, respectively [Elter et al., 1975].
 The outer front of the Northern Apennines exhibits a curved shape defined by NNW-SSE- and NNE-SSW-trending limbs to the north and south of the apical zone, respectively (Figure 1b). Folds show an axial trend parallel to the thrust plane direction in the hanging wall, whereas they trend NW-SE/NNW-SSE in the footwall [Coli, 1981; Salvini and Vittori, 1982; Koopman, 1983; Accordi et al., 1988; Calamita and Deiana, 1988; Corrado, 1995]. Shortening has been inferred to be maximum in the arc apex and to progressively decrease toward the endpoints [Pierantoni et al., 2005]. A strong influence of inversion tectonics has been documented in the Northern Apennines [e.g., Argnani and Gamberi, 1995; Tozer et al., 2002; Scisciani, 2009; Calamita et al., 2011]: during the Pliocene, NNE-SSW pre-existing normal faults were reactivated as oblique thrust ramps with fault-bend geometry, while NW-SE-oriented normal faults were passively displaced by the NNE-SSW oblique and the NW-SE frontal thrusts, with fault-propagation shortcut anticlines development [Calamita et al., 2012a, 2012b; Di Domenica et al., 2012].
 The southern part of the Northern Apennines' curved front (Figures 1 and 2) is represented by the so-called “Olevano-Antrodoco line” [Parotto and Praturlon, 1975]. This tectonic feature played an important role during the early Jurassic, separating the Umbro-Marchean pelagic domain and the Latium-Abruzzi carbonate platform [Castellarin et al., 1978]. This feature was reactivated during the Neogene, determining the over-thrusting of the Triassic to Miocene pelagic successions of the Umbro-Marchean domain over the Messinian foredeep sediments that rest on top of the Latium-Abruzzi carbonate platform. The Neogene kinematics of the OAS has been interpreted in a variety of ways in the literature, particularly concerning the existence of a dextral component of shear. Castellarin et al.  considered the Olevano-Antrodoco line as a dextral strike-slip and thrust fault that reactivated the Jurassic normal fault, while Salvini and Vittori  described it as a thrust fault within a multiphase tectonic evolution. Coli  and Lavecchia  interpreted such a feature as a dextral transpressive fault complicated by the interaction of several directions of movement between depositional paleodomains. Other authors [Koopman, 1983; Bally et al., 1986; Calamita et al., 1987; Calamita and Deiana, 1988; Tavarnelli et al., 2004; Finetti et al., 2005; Satolli and Calamita, 2008; Calamita et al., 2009; Di Domenica et al., 2012] interpreted the OAS as a frontal and oblique ramp complex that reactivated pre-existing Jurassic normal faults with different orientation.
 In the study area, the stratigraphic succession comprised (1) a hypothetical buried magnetic basement [AGIP SpA., 1981; Arisi Rota and Fichera, 1987] never exposed or drilled; (2) a Permian(?)-Triassic succession characterized by dolomites, evaporites, and continental deposits, documented only by well data and inferred by seismic line interpretation; and (3) outcrops of the Triassic to Miocene succession belonging to the Umbro-Marchean pelagic domain in the hanging wall of the OAS thrust, and to the Latium-Abruzzi carbonate platform, slope-to-basin and Messinian foredeep deposits in the southern part of the footwall (Figure 2).
 The lower part of the Umbro-Marchean succession evolved from shallow water carbonates (Calcare Massiccio Fm.; Lower Jurassic) to a pelagic domain characterized by Jurassic condensed, complete, and composite successions [Colacicchi et al., 1970; Centamore et al., 1971]. Differences in lithologies and thicknesses were horizontally leveled by the Maiolica Fm. (upper Titonian p.p.-lower Aptian p.p.), and deposition became uniform with the Marne a Fucoidi Fm. (Aptian p.p.-Albian p.p.). The Miocene hemipelagic succession (Bisciaro and Marne con Cerrogna Fms.) recorded the onset of foreland flexuration. Subsequently, the Messinian foredeep sediments (Laga Fm.) were deposited on both the Latium-Abruzzi carbonate platform and the Umbro-Marchean pelagic successions.
 Along the OAS thrust, the Mesozoic-lower Tertiary carbonate succession tectonically overlies the Miocene deposits of the Marne con Cerrogna Fm. The thrust trace is characterized by N50°-N60° to N-S trending segments. The hanging-wall anticline is characterized by an axial trend parallel to the thrust plane orientation. Close to Mt. Boragine, the OAS thrust is involved in the growth of NNW-SSE-trending footwall anticlines, which yield structural culminations of the thrust plane.
3 Structural Analysis
 Structural data were collected to analyze the local trends of the main structures, especially in the hanging wall of the OAS thrust (Figure 3). A mesostructural trend was calculated by measuring minor fold hinges, cleavage-bedding lineation, and bedding. Deformation is intense close to the thrust (sites OAL22, 29, 30, 31, 32, 33, 34) from the Scaglia Fm. and is characterized by recumbent to overturned mesofolds, with the fold axis oriented along approximately N60° (slightly decreasing to N40°-N30° between sites OAL30-OAL29) and moderately plunging toward the NE. Vertical strata, associated with mesofolding, show a NE strike parallel to the pressure-solution cleavage orientation; the lineation of the intersection between the cleavage and bedding confirms the approximately N60° fold axis orientation in the sampling area. Extension veins with calcite filling show a cross-cutting relationship with cleavage, indicating that the contemporaneous development of such structures is linked to pressure-solution processes. A few kilometers from the OAS thrust, the deformation is less intense (sites OAL23-28), as indicated by the development of moderately inclined mesofolds trending from N-S to approximately N10° (Figure 3b) and by the absence of pressure-solution cleavage. Similar structural trends characterize the site OAL46, as indicated by the presence of N10°-N20°-trending mesofolds (Figure 3d). Structural trends abruptly change between sites OAL39 and OAL45: minor fold hinges, vertical beds, and intersection between pressure-solution cleavage and bedding indicate a N150° to N170° trend (Figure 3c).
 In this area, the OAS thrust front is characterized by a shear zone with S-tectonites, which develop a pervasive foliation in Tertiary marls and shales (Marne con Cerrogna and Scaglia Cinerea Fms.) and a large spaced foliation in the marly/calcareous Scaglia Rossa Fm. Foliations are parallel to the thrust plane and are characterized by calcite shear veins, with a N60°-70° slip vector, compatible with the kinematics of the NW-SE-trending OAS thrust (Figure 3e).
 Both the footwall and the hanging wall of the OAS thrust are characterized by the presence of normal faults. The SW-dipping normal fault of Mount Boragine juxtaposes the Calcare Massiccio Fm. and the Scaglia Rossa/Scaglia Cinerea Fms. The fault has a downthrow of approximately 800 m and is displaced by the OAS thrust plane. The WSW-dipping normal fault in the backlimb of the Bacugno footwall anticline juxtaposes the Scaglia Rossa and Scaglia Cinerea Fms. and has a maximum downthrow of approximately 300 m. This normal fault plane is identified by a shear zone with an S-C fabric developed within the Scaglia Cinerea and Marne con Cerrogna Fms. The stratigraphic separation along the OAS thrust trace suggests a down-section trajectory of the thrust plane through pre-existing SW-dipping normal faults. A splay of the OAS transversal thrust displaced the carbonate succession belonging to its hanging-wall block, which involved pre-orogenic normal faults, and created Mount Nocella [Calamita et al., 2012b; Di Domenica et al., 2012].
4 Paleomagnetic Sampling and Analyses
 A detailed paleomagnetic investigation (Figure 2) was carried out in 2010 (from January to October) along the southern sector of the OAS thrust front. We sampled 52 sites, predominantly in upper Cretaceous to Eocene pelagic or slope-to-basin limestones belonging to the Umbro-Marchean basinal domain. Sites were located both in the OAS footwall and hanging-wall units and at variable distances from the thrust fault to detect the occurrence of strike-slip shear component along the OAS thrust front. In the hanging wall, sampling was designed to follow two transects perpendicular to two segments oriented roughly NE-SW and NNW-SSE. Sites were collected at different distances from the thrust front, in order to discriminate rotations related to regional-scale progressive-arc or local strike-slip shear. Most of the sites were sampled in the thin-bedded Scaglia Fm., with ages ranging from the middle-late Cretaceous (19 sites) to the Paleocene-Eocene (19); six sites (OAL24-OAL28 and OAL47) were collected from the Berriasian-Barremian Maiolica Fm., two sites (OAL37 and OAL52) were gathered from the Aptian-Albian Marne a Fucoidi Fm., two sites (OAL36 and OAL38) were collected from the Corniola Fm. (Pliensbachian), and four sites were collected from the Messinian Laga Fm. (OAL18-OAL21). Detailed age assignments of the sample ages are reported in the auxiliary material.1
 We collected 11–17 oriented cores (14 on average) from each site using a petrol-powered portable drill, spread laterally and vertically in the outcrop to average out the secular variation of the geomagnetic field. A total of 703 cylindrical samples (25 mm in diameter) were gathered. Cores were oriented in situ by a magnetic compass, corrected to account for the magnetic declination at the OAS area during 2010 (2°, according to Istituto Nazionale di Geofisica e Vulcanologia (INGV) ).
 The cores were cut into standard cylindrical specimens, and their natural remanent magnetization (NRM) was measured using a 2GDC-SQUID cryogenic magnetometer in the magnetically shielded room of INGV in Rome. All limestone samples were progressively demagnetized by stepwise thermal heating in 15 steps between 20 °C and 680 °C. Messinian clay samples were demagnetized in 16 steps in the 20–500 °C range. Demagnetization data were plotted on both orthogonal demagnetization diagrams [Zijderveld, 1967] and on equal-area projections (Figure 4). Remanence magnetization components were isolated by principal component analysis [Kirschvink, 1980] (or remagnetization circles when overlapping spectra components were observed) using the Paleomac software [Cognè, 2003].
 Specific analyses for the magnetic mineralogy were not performed. However, the unblocking temperature spectra observed in the Scaglia Fm. carbonates during thermal cleaning indicate that the remanence is carried by minerals of the magnetite and titano-magnetite family in a large number of samples (Curie temperature close to 580 °C); for 12% of the samples, unblocking temperatures exceeding 620 °C indicate the presence of minor hematite in addition to magnetite. Unblocking temperatures of 300–400 °C suggest that the magnetic mineralogy of the Miocene clays (sites OAL18-21) is likely dominated by ferrimagnetic iron sulfides, as described by Speranza et al.  for sites from the same formation sampled farther east. In most samples from Corniola Fm. (sites OAL36 and 38), it was not possible to isolate ChRMs due to low blocking temperatures, low NRM intensities, and the occurrence of a strong viscous component overlapping the GAD field.
5 Remanent Magnetization Components
 Thermal demagnetization data reveal three magnetization components (Figure 4): (1) a normal-polarity low-blocking temperature component removed before 180 °C, (2) an intermediate-temperature component (ITC) isolated in 18 sites (15 Scaglia, 2 Corniola, and 1 Marne a Fucoidi sites; Figure S2) between 220 °C and 320–460 °C with both normal and reverse polarity, and (3) a characteristic remanent magnetization (ChRM) isolated between 300–340 °C and 580–620 °C in carbonates and in the 120–275 °C and 475–525 °C range in clay samples. Almost all samples are characterized by the presence of at least two of these magnetization components.
 The mean low-temperature component is defined by D = 358.6°, I = 58.4°, α95 = 2.3° in situ coordinates. The semi-angle of confidence overlaps with the Geocentric Axial Dipole (GAD) field direction expected at the sampling locality (D = 0°, I = 60.1°). Thus, we interpret the low-temperature component as being related to a viscous overprint acquired during the Brunhes polarity chron.
 The ChRMs (Table 1 and Figures 5 and S3) are well defined for 536 out of 703 demagnetized samples. For 31 samples, no ChRMs are apparent, and only remagnetization circles could be calculated. Site-mean directions were computed using either Fisher's  statistics, when ChRMs could be isolated for all samples from a site, or the McFadden and McElhinny  method, which combines direct observations with remagnetization circles. Site-mean paleomagnetic directions are defined for 42 out of 52 sites; four sites are characterized by an intensity of magnetization that is too low to be measured with the cryogenic magnetometer (OAL28, 37, 38, 47) and further six sites (OAL13, 21, 26, 27, 36 and 46) were discarded from tectonic considerations as characterized by α95 > 20°.
Table 1. Paleomagnetic Results From the Olevano-Antrodoco-Sibillini Thrust Fronta
aEpoch and stages (in Ma) for the Mesozoic and Tertiary are from the timescale of Gradstein et al. . Bedding is expressed in dip azimuth and dip values (r indicates overturned strata); n/N is the number of reliable samples with respect to the total number of studied samples at site; D and I are site-mean declination and inclination; k and α95 are statistical parameters after Fisher, ; R and F are the site-mean rotation and flattening values [Demarest, 1983] relative to coeval D and I African values expected at the study area (latitude 42°34′N, longitude 13°04′E). The African poles used are from Besse and Courtillot . Sites labeled with (b), characterized by α95 > 20°, were discarded; sites labeled with (c), characterized by an intensity of magnetization that is too low to be measured, were discarded; sites labeled with (d) were not used in the computation of the mean rotational domain; the ages of sites labeled with (e) were determined using map information and/or by comparing NRM intensities (see text and auxiliary material).
−48.1 ± 10.5
−6.0 ± 7.3
−22.9 ± 9.0
9.0 ± 8.4
−50.3 ± 12.9
14.4 ± 11.2
−0.5 ± 14.5
11.8 ± 11.4
−29.2 ± 9.6
15.8 ± 8.3
−55.0 ± 16.0
12.6 ± 12.9
−1.3 ± 10.3
4.6 ± 7.7
−23.6 ± 17.4
0.5 ± 12.4
−18.9 ± 13.0
7.5 ± 11.1
−35.4 ± 14.4
15.1 ± 11.8
−23.3 ± 8.2
14.1 ± 6.9
Late Albian-Early Cenomanian
−8.6 ± 24.8
10.1 ± 27.4
−2.1 ± 15.3
29.0 ± 14.6
−60.2 ± 18.3
17.3 ± 15.2
−51.8 ± 21.9
6.7 ± 15.6
−50.5 ± 25.7
−0.6 ± 15.8
13.5 ± 15.0
29.2 ± 13.0
−11.0 ± 6.9
12.5 ± 4.8
−6.3 ± 13.9
7.9 ± 8.8
20.8 ± 15.8
17.9 ± 13.3
−22.4 ± 14.5
2.0 ± 10.8
45.9 ± 18.1
−0.5 ± 15.7
64.5 ± 23.6
−18.3 ± 14.0
8.0 ± 17.2
9.6 ± 14.7
9.0 ± 14.4
10.7 ± 12.4
55.1 ± 14.0
7.9 ± 11.8
28.6 ± 12.3
−1.9 ± 8.8
80.7 ± 26.3
−3.7 ± 15.0
88.7 ± 18.2
9.4 ± 15.0
−104.2 ± 18.9
19.1 ± 16.1
−1.6 ± 12.1
−7.3 ± 7.3
−32.4 ± 9.1
2.4 ± 7.1
17.9 ± 15.3
16.5 ± 12.7
−17.2 ± 9.8
−3.5 ± 6.3
13.7 ± 19.4
−2.5 ± 11.7
23.2 ± 8.2
12.5 ± 7.7
29.1 ± 23.3
−3.5 ± 26.3
11.1 ± 25.9
24.7 ± 28.7
−130.6 ± 14.1
18.8 ± 13.0
−86.8 ± 38.5
−14.6 ± 14.8
12.7 ± 15.3
5.8 ± 12.9
32.4 ± 12.3
13.1 ± 11.4
 The ITC shows two different behaviors. In Corniola Fm. (OAL36 and 38), Marne a Fucoidi (OAL37), and two Scaglia Fm. (OAL29 and OAL34) sites, it is D = 12.4°, I = 48.9°, α95 = 6.5° in in situ coordinates (11° from the GAD field direction). In 13 sites from Scaglia Fm. (OAL01, 02, 04, 07, 08, 30, 31, 32, 33, 40, 42, 43, 45), the ITCs show opposite polarity with respect to the relative ChRM components (e.g., as in Figure 4e), both in in situ and tilt-corrected coordinates. The reversal test [McFadden and McElhinny, 1990] performed in tilt-corrected coordinates at site scale is positive class C for site OAL07 (γ = 4.4°, γc = 16.5°) and indeterminate for sites OAL30 (γ = 17.3°, γc = 23.4°), OAL31 (γ = 22.8°, γc = 24.2°), and OAL33 (γ = 14.3°, γc = 32.2°). For the rest of the sites, the test is negative, probably due to an overlap in the unblocking temperature spectra of the ChRM and low-temperature components. Inclination-only fold test indicates that this ITC was acquired at 77.0% of unfolding (I = 50.1°, k = 7.2, with 95% confidence limit between 62.3% and 85.0%) [Enkin and Watson, 1996].
6 Age of the Characteristic Magnetization Components
 There are several lines of evidence that support the primary nature of the ChRMs. (i) The inclination-only fold test indicates that ChRMs were acquired when bedding was horizontal. As fold and reversal tests could not be performed because the sites have different ages and underwent different amounts of rotations (Table 1), we performed an inclination-only fold test [Enkin and Watson, 1996], on the retained sites (excluding samples from Maiolica and Laga Fms. due to the strong difference in age). Best unfolding is at 95.7% (k = 12.1; 95% confidence limit of 88.3% and 103.0% unfolding), indicating that ChRMs are acquired pre-folding. (ii) ChRMs show both normal and reverse polarities and are systematically far from the GAD field direction both in in situ and tilt-corrected directions (except for OAL18-20 sites, sampled in the Miocene Laga Fm.), suggesting the absence of magnetic overprints (Figure S3). (iii) The magnetic polarities gathered from the 42 sites and their ages were compared with the Cenozoic [Berggren et al., 1995] and Mesozoic [Gradstein et al., 2004] geomagnetic polarity timescales (Figure 6). Although most of the site ages straddle several polarity chrons, Marne a Fucoidi and older Scaglia sites (ranging from Aptian to Santonian) are found entirely within the long normal Cretaceous superchron (Figure 6). Consistently, all these sites show a normal polarity, supporting the primary nature of their ChRMs. (iv) Observed inclination values agree at first approximation with the expected inclinations for the Adriatic/African foreland [Besse and Courtillot, 2002]. Box plotting on flattening values was performed on all retained sites (save OAL35, OAL49, and OAL50 that were discarded on the basis of statistical consideration explained further in this study). Flattening is 7.7° ± 9.9° on average and 50% of samples are included between 13.8° and −0.2°. A significantly negative flattening value is only observed in the Maiolica site OAL25 (−18.3° ± 14.0°), which does not have a well-determined age. The slight observed flattening confirms previous evidence from other works. In fact, shallow paleomagnetic inclinations have been largely reported from the Apennines [e.g., Speranza et al., 1997; Satolli et al., 2005] and all over the Mediterranean region [Beck et al., 2001] and have been routinely attributed to the effects of shallowing and diagenesis that strongly affect sedimentary rocks [Deamer and Kodama, 1990; Bazhenov and Mikolaichuk, 2002].
7 Rotational Pattern Across the OAS Thrust Front
 To evaluate rotations related to the strike-slip and thrust sheet tectonics of the OAS, tilt-corrected ChRMs were compared to coeval directions expected for the Adriatic foreland. Rotation and flattening values were computed according to Demarest  using reference African paleopoles from Besse and Courtillot  because Adria is considered to have mirrored the African drift since at least Permian times [Channell, 1992; Van der Voo, 1993; Muttoni et al., 2001; Satolli et al., 2007, 2008]. The declination and inclination from each site were compared to the reference paleopole closer to the site mean age. To evaluate the error effect on the rotations given by the age assignment criterion, we also computed rotations using the youngest and oldest pole within the age range for each site (Figure 4). For most sites, the difference in rotation obtained by using the youngest and oldest reference paleopoles is included in the error bar of the chosen paleopole (e.g., for OAL45, the youngest and oldest paleopole are 110 Ma and 95 Ma, respectively, while the rotation was computed using the 105 Ma reference pole; these observations imply an error in the rotation of approximately 7°, which is included in the error bar). Thus, we assume that the estimated rotations are sufficiently accurate and reliable to be used for tectonic interpretations.
 Rotations are highly variable in sign and magnitude (Table 1), as expected in tectonically complex areas. We statistically investigated the relationships between rotations and structural trends in the OAS thrust front, by using structural data from this work (Figure 3) and from Calamita et al. [1987; 2012b], Piana , and Di Domenica et al. . Folds axes are characterized by a regional direction of approximately N160°-N180° in the OAS footwall, while in the hanging wall, they vary from N150°-N170° to N10° and N60°, following the same direction of the thrust front. We applied data clustering separately for OAS thrust footwall and hanging-wall data because of their provenance from different structural units. A structural trend versus site rotations plot (Figure 7a) shows a clear clustering for both footwall and hanging-wall sites, confirmed by a successful Z-test for data clustering (Z = −7.9096; the test is successful when Z < −1.96, which is the critical value of the Z statistic for clustering at a significance level of 95%; Swan and Sandilands ; Trauth ). A Gaussian mixture distribution [McLachlan and Peel, 2000] was applied to fit data into clusters. This method performs fitting through an iterative Expectation-Maximization (E-M) algorithm, which assigns posterior probabilities to every component density with respect to each observation, indicating that data points have some probability of belonging to each cluster. Results of such analysis individuate three domains, with 0.99 of probability at 95% of confidence. There are two clusters from the hanging wall: one encompasses a strong CW-rotated domain lying along the N60°-trending segment of the OAS thrust (D1 = 54.8° ± 5.5° CW from five sites) and the other one includes all the others sites (D2 = 15.1° ± 5.8° CW). There is evidence for a single CCW cluster in the footwall. On the basis of geological considerations, the footwall can be further divided in two domains, due to the presence of the outer thrust of Mount Gabbia Unit [Piana, 1995]: a CCW-rotated domain (D3 = −30.8° ± 4.7° from 16 sites in the Scaglia Fm.) and an approximately non-rotated domain (D4 = -5.2° ± 3.7° CCW for four sites).
 Sites OAL35, OAL49, and OAL50 are not part of any cluster. A box-plot visualization of each rotational domain confirms that OAL35, OAL49, and OAL50 are outliers (Figure 7b). Geological observations allow us to hypothesize that OAL35 is strongly CCW rotated as sampled in a thrust splay, while strong CCW rotations in OAL49 and 50 should be due to tectonic interference, even if no fault is apparent, probably due to coverage of continental deposits. On the basis of cluster analysis, these sites were excluded from any further consideration.
 The paleomagnetic and structural study of the southern sector of the OAS thrust allowed us to recognize four rotational domains with different signs and magnitudes (Figure 8). The hanging wall domain D2 shows a slight CW rotation, in agreement with oroclinal tests from the literature [Speranza et al., 1997; Cifelli and Mattei, 2010]. Rotations do not change moving toward the NNW-SSE-trending thrust front, while they increase toward the NE-SW-trending sector, becoming strong (D1) at a few hundred meters from the thrust. The strong increment in rotations suggests the presence of a dextral strike-slip component [Castellarin et al., 1978; Coli, 1981; Salvini and Vittori, 1982; Lavecchia, 1985; Corrado, 1995]. As a consequence, considering approximately N60°-70° slip vector, the NE-SW- and NNW-SSE-trending sectors are interpreted as oblique and frontal ramps of the OAS thrust, respectively [in agreement with Koopman, 1983; Calamita et al., 1987; Bally et al., 1986; Calamita and Deiana, 1988; Tavarnelli et al., 2004; Finetti et al., 2005; Satolli and Calamita, 2008; Calamita et al., 2009]. The strike-slip component of the OAS thrust is probably due to the oblique convergence (N60°-70°) into the frontal foredeep basin [Macedo and Marshak, 1999] and related to variations in lithology in the footwall [Weil et al., 2010]. A similar mechanism was predicted for the OAS thrust by Corrado  based on anisotropy of magnetic susceptibility (AMS) data.
 In the footwall, the non-rotational domain D4 is separated from a CCW rotation domain, which was not previously documented, and that can be attributed to the outer thrust of Mount Gabbia Unit [Piana, 1995]. This thrust is also documented at depth from the Antrodoco 1 well (Figure 1) [AGIP, 1959, http://unmig.sviluppoeconomico.gov.it/videpi/pozzi/dettaglio.asp?cod=336], which intersected lower middle Cretaceous pelagic sediments below approximately 3 km of Triassic anhydrites and dolomites. Data from the OAS footwall are in agreement with paleomagnetic [Mattei et al., 1995] and AMS [Corrado, 1995] data recorded SW with respect to our study area, in the Latium-Abruzzi carbonate platform, indicating that it was involved in strong CCW rotations during the orogenesis.
 Based on our data integrated with previous data from the literature, we hypothesize three evolution phases for the Northern Apennines (Figure 9): (1) in the pre-orogenic stage, the area is not yet involved in the orogenesis and is characterized by the architecture of the Adria Mesozoic paleomargin; (2) in the early orogenic stage, thrust emplacement is strongly influenced by pre-existent normal faults (inversion tectonics) and leads to the development of a primary arc; and (3) in the late orogenic stage, tectonic rotations were induced in thrust sheets, accentuating the curvature of the belt. Thus, we can state that the Northern Apennines is a progressive arc, strongly influenced in its shape by the architecture of the Mesozoic Adria paleomargin [Calamita and Deiana, 1988; Tavarnelli et al., 2004; Satolli and Calamita, 2012], whose curvature was accentuated by the rotation of thrust nappes, CCW and CW along its northern and southern edges, respectively. In particular, the southern sector of the OAS thrust, which reactivated the Ancona-Anzio Jurassic normal fault [Castellarin et al., 1978; Di Domenica et al., 2012], is articulated in oblique and frontal thrust ramps. This transpressive thrust realized the kinematic decoupling between CW- and CCW-rotated domains in the hanging-wall and footwall blocks.
 A detailed paleomagnetic sampling was performed in the southern sector of the OAS thrust to unravel the influence of dextral strike-slip components during the thrust emplacement. Results of cluster analysis of paleomagnetic data and geological constrains allowed us to document a peculiar distribution of tectonic rotations (Figure 8): (1) a strongly rotated CW domain (D1 = 54.8° ± 5.5°) in the hanging wall, close to the NE-SW-trending segment of the thrust; (2) a less CW-rotated domain in the hanging wall (D2 = 15.1° ± 5.8°) that includes both the NNW-SSE-oriented thrust segment and sites far from the thrust; (3) a uniform CCW rotation (D3 = -30.8° ± 4.7°) in the Monte Gabbia area in the footwall; and (4) an approximately null rotation (D4 = −5.2° ± 3.7°) in the external footwall. The rotational difference between D3 (Mount Gabbia Unit) and D4 (external footwall) is supported by the presence of a WNW-ESE-trending thrust front. The rotational domain (D2) of approximately 15° in the hanging wall confirms the progressive nature of the Northern Apennines [Cifelli and Mattei, 2010], while a strike-slip component is documented only along the NE-SW-trending segment of the thrust by the strongly CW-rotated domain (D1). The strong CW-rotated D1 domain is related to a lateral or strongly oblique ramp development induced by the oblique convergence toward a foredeep basin characterized by variations in lithology, while weaker CW rotations at Mt. Boragine (D2 domain) are related to the presence of a frontal ramp.
 Our data confirm that the Northern Apennines is a progressive arc, whose shape has been strongly influenced by the architecture of the Mesozoic Adria paleomargin. Its curvature was accentuated by tectonic rotations that are enhanced close to transpressive structures. In the studied area, the strike-slip-driven CW rotations rapidly decrease moving away from the front (they are documented only up to 1 km away from the thrust plane). Our study confirms the importance of detailed paleomagnetic analyses in discriminating between oroclinal/progressive-arc- and strike-slip-related components in major orogenic arcs.
 We want to thank R. Van der Voo, A. B. Weil, and two anonymous reviewers for constructive comments that improved the quality of this paper. We also thank Editor T. Parsons and Associate Editor S. Gilder for additional comments. This work has been supported by FIRB-2008 (S. Satolli) and MIUR-ex 60% (F. Calamita) grants.