Constraining the Passive to Active Margin Tectonics of the Internal Central Apennines: Insights from Biostratigraphy, Structural, and Seismic Analysis

The polyphase structural evolution of a sector of the internal Central Apennines, where the significance of pelagic deposits atop neritic carbonate platform and active margin sediments has been long debated, is here documented. The results of a new geological survey in the Volsci Range, supported by new stratigraphic constraints from the syn-orogenic deposits, are integrated with the analysis of 2D seismic reflection lines and available wells in the adjacent Latin Valley. Late Cretaceous syn-sedimentary faults are documented and interpreted as steps linking a carbonate platform to the adjacent pelagic basin, located to the west. During Tortonian time, the pelagic deposits were squeezed off and juxtaposed as mélange units on top of the carbonate platform. Subsurface data highlighted stacked thrust sheets that were first involved into an initial in-sequence propagation with top-to-the-ENE, synchronous to late Tortonian foredeep to wedge-top sedimentation. We distinguish up to four groups of thrust faults that occurred during in-sequence shortening (thrusts 1–3; about 55–60 km) and backthrusting (thrust 4). During Pliocene to recent times, the area has been uplifted and subsequently extended by normal faults cross-cutting the accretionary wedge. Beside regional interest, our findings bear implications on the kinematic evolution of an orogenic wedge affected by far-traveled units.


Introduction
Carbonate platforms are a type of passive margin sedimentary succession that can be commonly involved in the thrust-sheet imbrication of an orogenic wedge [1][2][3]. During in-sequence ongoing deformation, the wedge propagates by incorporating new portions of the foreland, which is commonly made up of crystalline basement, clastic and/or carbonatic successions, and overriding foredeep/foreland clastics with variable thickness and composition [4][5][6]. The so formed fold-and-thrust belt, incorporating distinctive tectonostratigraphic units, is the combined product of inherited syn-sedimentary structures and orogenic dynamics [7,8]. Thus, the wedge-related deformation style may strongly depend on the stratigraphic architecture and in particular on the presence and depth of décollement  [30,38,44], showing the active margin units and the Meso-Cenozoic passive margin units. The shortening time is in italic. (b) Crustal cross-section (modified after the work in [45]). Deep well location is taken from in [23].

The Central Apennines
The Apennines ( Figure 1) are a ~1500 km long accretionary wedge made of different pre-orogenic and syn-orogenic units accreted together during the progressive E/NE-ward migration of leading-edge frontal thrusts and associated active margin units deposited  [30,38,44], showing the active margin units and the Meso-Cenozoic passive margin units. The shortening time is in italic. (b) Crustal cross-section (modified after the work in [45]). Deep well location is taken from in [23].

The Central Apennines
The Apennines (Figure 1) are a~1500 km long accretionary wedge made of different pre-orogenic and syn-orogenic units accreted together during the progressive E/NE-ward migration of leading-edge frontal thrusts and associated active margin units deposited within foredeep and wedge-top basins (see, e.g., in [46][47][48][49][50]. From Miocene time, the Apennine foreland became progressively involved in pre-thrusting bulging, uplift, and erosion resulting from the wedge migration [51][52][53][54][55][56]. Since Tortonian time (~11 Ma), the west-directed subduction of the Adriatic slab drove the development of the accretionary wedge now exposed in the central sector of the Apennine belt [49,54]. Subsequently, the fold-and-thrust belt underwent severe crustal stretching, related to back-arc extension that progressively migrated from the Sardinian margin to the axial part of the central  [64]. On the right, tectonic context and stratigraphy of basin deposits is reported from the literature (see Appendix A) and original data at representative localities. Localities from the Ernici unit are highlighted by vertical gray stripes. Below, the geological map of the study area with the studied locations and their respective numbers.  [64]. On the right, tectonic context and stratigraphy of basin deposits is reported from the literature (see Appendix A) and original data at representative localities. Localities from the Ernici unit are highlighted by vertical gray stripes. Below, the geological map of the study area with the studied locations and their respective numbers.

The Volsci Range and the Latin Valley
The VR is traditionally subdivided into major mountain groups, i.e., West Lepini, East Lepini, Ausoni, and West and East Aurunci Mts (Figure 2), that are separated by major valleys or mountain passes. More to the SW, the Mount Massico structural high occurs. These groups share a similar tectonic and stratigraphic evolution. The VR is mostly composed of passive margin Mesozoic neritic carbonates belonging to the Latium and Abruzzi platform or Apennine carbonate platform (see, e.g., in [79][80][81]). The Mesozoic dominant facies are representative of inner to rim carbonate platform environments (see, e.g., in [63,64,82,83]).
A compilation of the Mesozoic lithostratigraphic units cropping out in the Lepini sectors is presented in Figure 2. The Upper and Lower Volsci thrust sheets differ from the Upper Ernici unit on the basis of the Cenozoic stratigraphy. Of note, the VR succession generally bears a thin and incomplete succession of Paleocene to Miocene deposits [84] atop late Cretaceous formations of different ages, possibly due to progressive drowning of some sectors of the platform during Late Cretaceous time [63]. On the other hand, in the Latin Valley, the Ernici unit is thicker and also contains Eocene to early Tortonian foreland units and late Tortonian to earliest Pliocene active margin siliciclastic formations (see in [63] and the references therein).
Seismic interpretation studies in the Latin Valley, carried out by AGIP and other companies (www.videpi.com) (accessed on 20 January 2021), trace top-platform seismic horizons that allowed us to locally outline a fold-and-thrust structure [85]. According to the authors of [64,86], the VR front propagation affected the Latin Valley foredeep deposits that were doubled or even triplicated [45]. Upper and lower units in the Volsci Range and in the Ernici units of the Latin Valley were thus distinguished. As also shown in the cross sections in [64], thrusting involved the Cretaceous carbonates of the Ernici unit together with upper Tortonian foredeep sediments of the Frosinone Formation [63,64]. Finally, out-of-sequence thrusting during and after the Messinian salinity crisis was documented in [77,87], possibly related to backthrusting, like at Carpineto Romano [88]. The thrust front does not crop out, but according to the most recent reconstructions, it is offset by normal faults [45,86]. At least from Middle Pliocene time, the study area experienced regional uplift, accompanied by subaerial exposure and consequent diffuse erosional processes that generated erosional surfaces, now found at different elevations [63].
According to the authors of [89], just north of VR the uplift rate increased during the last 2.4 Myr. In the VR, no such detail was reached yet. However, early to late Pleistocene slope, river, and lacustrine paralic and continental deposits were mapped within depressions bounded by high-angle NW-and NE-striking normal faults that dissected the fold-and-thrust fabric. Further, E-striking transtensional faults contribute to generate middle Pleistocene wrench zones and basins between the Latin Valley and the Pontina Plain. Syn-to post-tectonic upper Pliocene-middle Pleistocene continental successions are preserved in the Middle Latin Valley, the Pontina Plain, and locally in the VR intermontane depressions [64]. Further, during late Pliocene to possibly Holocene times, the fold-and-thrust belt was progressively cross-cut by a system of conjugate synthetic and antithetic normal faults determining the formation of the coastal plain and intra-mountain depressions [64,90,91]. The VR hosts volcanic terrains of Pleistocene age from both nearby volcanic districts and local eruptive centers belonging to the Volsci Volcanic Field (VVF; Figure 1 [64,92]).

Stratigraphic Review and New Paleontological Determinations
The lithostratigraphic architecture of the Meso-Cenozoic carbonate platform succession has been reviewed, following the scheme in [45], and it has been integrated with a stratigraphic chart that compares eighteen different key localities representative of preorogenic passive margin to syn-orogenic foreland basin lithostratigraphic units throughout the study area ( Figure 2). Erosive submarine and karst-related unconformities are re-Geosciences 2021, 11, 160 7 of 50 ported to support the regional review of the syn-orogenic evolution, also constrained by the absolute ages provided in [43] for the Massico Mt ridge. The overall stratigraphic setting allowed us to correlate diachronous events among different structural units from the Volsci Range and Latin Valley. Lithologies not constrained by biomarkers are traced by a question mark, whereas lithologic and biostratigraphic information coming from the review of the existing literature is resumed in the table of Appendix A. We have harmonized the stratigraphic information published in the 1:100,000 maps (i.e., Latina, Frosinone, and Alatri; https://www.isprambiente.gov.it/) (accessed on 20 January 2021), and in the more recent and detailed 1:50,000 maps (i.e., Anagni, Ceccano, and Velletri; https://www.isprambiente.gov.it/) (accessed on 20 January 2021) as well as and in other papers (i.e., in [72,84,[86][87][88]93,94] and, using the stratigraphic nomenclature after that in [64], then grouped the deposits into the broader informal lithostratigraphic subdivision of Figure 2. New biostratigraphic information was acquired by studying Upper Cretaceous-Miocene to early Pliocene samples collected from fifteen localities at Colle Cantocchio, Gorga, Gavignano, Carpineto Romano, Caccume Mt., and Siserno Mt. (Figure 2). Further sampling through the Latin Valley at Morolo, Ferentino, and Frosinone localities was performed in order to determine facies and fossil content of syn-orogenic deposits. Hard rock samples have been prepared for thin sections analysis, which provided thirty-three new age determinations. Further, we collected seventeen samples for nannoplankton using samples prepared under smear slide technique, and following the procedures described in [95]. We observed the nannoplankton content through the polarized light microscope Zeiss Axioscop equipped with an ×100 oil immersion objective lens. We performed a qualitative evaluation of the assemblages on all the samples, but only twelve of them proved to be fossiliferous, while five other ones are barren or poorly fossiliferous. Important time maker nannoplankton taxa were identified up to species level, as presented in Supplementary Material. We base our time determination on the micro-biostratigraphic frames in [82,[96][97][98] for the shallow-water carbonate assemblage and the biostratigraphic scale in [99][100][101] for the nannoplankton.

Structural Analysis
A new structural-geological survey of the carbonate and siliciclastic succession integrates previous work of the Geological Survey of Italy (ISPRA) (i.e., in [64,102,103] and the references therein). The resulting new geological map is built also considering a specific review of the 1:50,000 geological sheets "Anagni" and "Ceccano" in order to avoid lithostratigraphic synonymy (see Appendix A) [64,103].
Bedding attitude was retrieved from existing map sheet tables at the scale 1:25,000 on a stripe of about one kilometer to each side of the main cross section (Figure 3). In order to constrain fault kinematics, field measurements of faults, fractures, and slicken-fibers were collected at key localities and plotted by means of TectonicsFP software [104] with lowerhemisphere projections and rose diagrams. In particular, at each locality eigen vectors are calculated from the bedding and are indicative of the orientation of the axes of deformation, where the gray circles are representative of the plane between the principal and minimal eigenvector. In general, an eigenvector is a vector which gets stretched, but not rotated, when operated on by the matrix. Considering that eigenvectors have corresponding eigenvalues, the amount of squeezing or stretching (the strain) is called the eigenvalue. Eigenvectors from key localities are reported in Table S1 (Supplementary material).  3,5,8,11,12, and 13 are provided by Pentex Limited Italia (see Acknowledgments). Wells 6, 7, 9, and 10 are reported in [105]. Full lines are stratigraphic correlations within the same structural unItal. Black dashed lines are uncertain stratigraphic and tectonic correlations (blue lines). Regional cross section AB ( Figure 6) and the detailed structural maps of the Figures 6 and 9 are also shown.

Borehole Data from the Latin Valley
Composite well log data from the exploration and production of hydrocarbon activity were used to calibrate the seismic lines ( Figure 3). Fifteen wells were drilled through the syn-orogenic lithologies, and they provide insights on late Miocene siliciclastic deposits. Four wells are from a public database (www.videpi.com) (accessed on 20 January 2021), the others were extrapolated from the literature [64,106,107] or confidential reports provided by Pentex Italia Ltd. The stratigraphic calibration of the seismic profiles was performed by using (i) the Frosinone 1 well, which is located within a relatively dense network of seismic lines and drilled at total depth of 684 m, reaching the Orbulina Marl Fm at 526 m and the CBZ at 551 m, while the Cretaceous carbonate platform top was encountered at 620 m, and (ii) the Anagni 1 well, which encountered mesozoic platform carbonates between 47 and 162 m and reached again the carbonate top at 862 m after having crossed a thick siliciclastic succession (Figures 2 and 3). Three wells were characterized by velocity data that allowed us to calibrate seismic data and/or calculate average and interval velocity for the identified macro-units. Where velocity logs were not available, an average interval velocity based on our calculations was applied to fit with the correspondent lithology and reflector detected on seismic profile. In few cases, velocity logs were available for a direct local time-depth chart; in the other cases, average velocity obtained by the analysis of the available logs and from literature was used. These two velocity laws were used to depth-convert the two-way-time interpretation on seismic dataset, in order to define thickness and depth of the main top interpreted horizons to set the geological cross section ( Figure 14). Biostratigraphic data are available only for a few key wells (i.e., Paliano 1, Gavignano 1, Anagni 1, Frosinone 1, Liri 1, and Farnese 1) and have been anchored using the regional scale in [96].

Seismic Dataset
The structural setting of the Latin Valley presented in this study largely relies on thirty-eight 2D seismic reflection profiles irregularly arranged (map view Figure 3b). In the north, some seismic lines gather around the Gavignano 1 and the Anagni 1 wells, while in the south they occur together with different wells (Figure 3). The seismic sections originate from different acquisition campaigns carried out in the 1980s and 1990s for the exploration of hydrocarbons by AGIP and recently by Sovereign and Pentex. Most seismic lines are part of a public dataset (ViDEPI Project. Available online: https://www.videpi.com accessed on: 20 January 2021. This public network has been integrated by a few other seismic lines from different surveys, to better constrain the structural setting of the Latin Valley. The interpreted seismic dataset was a stack version. Public data were in raster format, so we produced segy files for each raster seismic line in order to be able to import all the dataset into the interpretation software (OpendTect). This was achieved using Kogeo© 2.7, a free and open software for 2D/3D seismic data analysis that allows to create a geo referenced seg-y file from a scanned seismic image (http://www.kogeo.de/index.htm) (accessed on 20 January 2021). Seismic quality is good to poor, probably due to a lack of reprocessing and therefore interpretation may be inaccurate in some points. In those cases, we have integrated the outcropping geological information to reconstruct a geological model along the seismic profile, identifying when possible the main reflectors.
The most evident reflectors are the unconformities at the top of the upper Cretaceous carbonates (Figure 4), and of the Orbulina Marl Fm. (UAM; Figure 2). To calibrate and detect the main reflectors/markers in the Latin Valley, a synthetic seismogram was created for the Anagni-1 well ( Figure 4) by focusing on the following formation discontinuities (from the bottom to the top): at the top of the Cretaceous limestones (UK), at the top of the Bryozoa and Lithothamnium limestone (CBZ), and at the base of the Frosinone Formation (FFS). For the interpretation of the seismic profiles, we identified the top-CBZ as the key reflector with the strongest acoustic impedance contrast observed over the entire Latin Valley. This often corresponds to the UAM lithostratigraphic unit (Figure 2), which at the basin scale corresponds with one of the most used reflectors that tie wells with seismic lines [108][109][110]. Miocene and Cretaceous near-top reflectors are well recognizable because of the characteristic geometry and energy picks that are stronger than the adjacent reflectors. In particular, the marly layers reflect most of the down-going seismic energy, obscuring the siliciclastic sequence or the underthrusted carbonate units. Despite the limited thickness of UAM, this reflector was followed also on the poorer quality seismic lines.
(from the bottom to the top): at the top of the Cretaceous limestones (UK), at the top of the Bryozoa and Lithothamnium limestone (CBZ), and at the base of the Frosinone Formation (FFS). For the interpretation of the seismic profiles, we identified the top-CBZ as the key reflector with the strongest acoustic impedance contrast observed over the entire Latin Valley. This often corresponds to the UAM lithostratigraphic unit (Figure 2), which at the basin scale corresponds with one of the most used reflectors that tie wells with seismic lines [108−110]. Miocene and Cretaceous near-top reflectors are well recognizable because of the characteristic geometry and energy picks that are stronger than the adjacent reflectors. In particular, the marly layers reflect most of the down-going seismic energy, obscuring the siliciclastic sequence or the underthrusted carbonate units. Despite the limited thickness of UAM, this reflector was followed also on the poorer quality seismic lines. . Simplified stratigraphic column of the Anagni 1 borehole with velocity log, synthetic seismogram, and an extract of a seismic section passing by the well (location in Figure 3). The seismic marker horizons and additional stratigraphic horizons interpreted in this study are also shown. . Simplified stratigraphic column of the Anagni 1 borehole with velocity log, synthetic seismogram, and an extract of a seismic section passing by the well (location in Figure 3). The seismic marker horizons and additional stratigraphic horizons interpreted in this study are also shown.

Stratigraphic Review
The stratigraphy of the study area is schematically reported from the literature in Figure 2, where the lithostratigraphic units are anchored to the exposed sections at each of the eighteen localities presented in the map. The basics of the different tectonic units are exposed in Section 2.2. A new set of ages is proposed for the succession cropping out at the northern Volsci Range, as shown in the next section.
The Upper Campanian to Eocene carbonate platform succession that rest on the Hippuritid and Radiolitid limestone is generally missing [64], possibly due to a widespread depositional hiatus, although it locally crops out (e.g., at Gorga [103,111]). Note that the shallow-water Spirolina limestone (lower to mid-Eocene [112]), which crops out only in rare patches comprised between two unconformities-probably related to emersion events-was found at Gorga [112], Ferentino, and Castelforte ( Figure 2; see also in [96,113], while it was recognized in well logs of Paliano 1 DIR and Farnese 001dir ( Figure 3). In the Volsci Range, the Bryozoa and Lithothamnium limestone (CBZ) was dated as middle Miocene (see, e.g., in [64,87]). However, our data from the Volsci Range show that at least the CBZ base is early Miocene in age (see Section 4.1.2). Locally in the Volsci Range (e.g., Carpineto Romano, Figure 2), the CBZ lithotype is reported to occur within and beneath the allochthonous sub-Ligurian units [72], that can be compared with the Falvaterra Chaotic complex in [63].
Overall, the Falvaterra Chaotic complex is an ensemble of Paleocene to middle Miocene lithoclasts (from dm to decametric) wrapped within a matrix, whose best age constraints were provided mostly from the outcrops of Colle Cavallaro [114]. The basal contact of the Chaotic complex, although tectonically overprinted [63], is often marked by a ferruginous-limonitic veneer that occurs as a calcareous-detrital iron-oxide cruston. Differently from the classical carbonate hardgrounds, that are surfaces of synsedimentarily cemented carbonate layers that have been exposed on the seafloor under an extremely low sedimentation rate, the crustons of the Volsci Range could be either of karstic origin and/or the product of fluids involved into thrust faulting. Near Formia these crustons occur on top of peritidal limestones with benthic foraminifera (redetermined after the work in [63]) including Spirolina sp. [115], which can be possibly attributed to the early Eocene [111]. In particular, the foraminifera shown in [63] (their Figure 4) appear closer to some shallow-water discorbidae rather than planktonic forams. However, this need to be verified with new determinations. Our data constrain the top platform units providing new insights on the correlation, envisaged in [63], between these crustons and the Upper Cretaceous-lower Miocene succession preserved in the Chaotic complex (cf. Section 5.1 on the basis of the new stratigraphic constraints presented in Section 4.1.2.). Concerning the stratigraphic evidence from the Paleogene-early Aquitanian pelagic terms atop (Figure 2), they are mostly represented by Scaglia lithotypes (e.g., Formia and Spigno Saturnia, Figure 2). These lithotypes also crop out beneath the thrust south of Carpineto Romano, and beneath the Caccume Mt. and Colle Cavallaro klippen ( Figure 2). Further, Scaglia sensu latu lithotypes were found as blocks of various dimensions wrapped in clayey matrix together with: early-middle Miocene lithoclasts (Figure 2; Appendix A), upper Serravallian cherty marl, and massive to laminated arcosic greywackes with mica [103]. The latter resulted sterile at the Caccume Mt. [84]. Lithologies of clasts involved into the Chaotic complex belong to a wide chronostratigraphic interval (i.e., Paleogene-Serravallian pro parte; Figure 2). More to the south, beneath the Vele Mt. thrust, siliciclastic marly deposits, mapped as Chaotic complex equivalent units, occur. Our data provide age constraints for the northern Volsci Range, see Section 4.1.2, and provide insights on the stratigraphic development of the sedimentary succession later deformed as Chaotic complex.
In the Latin Valley, the Frosinone Fm. was homogeneously attributed to late Tortonian time, while on the northeastern edge of the valley the base seems to be younger (i.e., uppermost Tortonian [87]). The upper part of the Frosinone Fm. unit bears olistoliths and olistostromes [115], from Mesozoic platform and Chaotic complex equivalent lithologies. They are reported at Sgurgola [35] and in the Torre Ausente Valley [64,116], although not as nicely cropping out as at the Massico Mt. [37].
Well data show a highly variable facies pattern of the siliciclastic units that include carbonate intercalations and thick marly successions with minor to rare sandstone horizons (Gavignano 1; Anagni 1; Frosinone 01; Farnese 001 wells; Figure 3). Due to tectonic juxtaposition, these successions may appear repeated at least twice and thus also reaching a total thickness of about 1.8 to 2.5 km at Gavignano and Liri and Farnese wells. Single thrust-bounded siliciclastic units are up to some 0.7 km thick.
In particular, the Gavignano 1 well hits four repeated siliciclastic-marly sequences bounded by thrust faults juxtaposing older terrains above younger ones. The uppermost unit is constituted of Upper Cretaceous (UK) limestones (cf. Anagni 1 well). The deeper fault-bounded units are about 600-900 m thick. Their siliciclastic sequence is defined by different lithofacies associations including alternations of sandstone, marl, and limestone. By correlating the wells providing detailed biostratigraphic information (e.g., Paliano, Gavignano, and Frosinone), we have correlated similar lithostratigraphic units, thus providing a formation identification. Biostratigraphic data from wells do not report Messinian taxa. Thus, we consider the Messinian Monte San Giovanni Campano unit (MVP) following the work in [63] and composed of wedge-top clastics [87], including other formally defined lithostratigraphic units (i.e., Torrice Sandstone Fm, Figure 2). Despite this lack of subsurface biostratigraphic information, its occurrence at depth cannot be excluded. Further, the correlation among conglomerates bearing exotic clasts of granitoids (SBG) is not clear as not supported by resolutive available stratigraphic information. However, their occurrence is of regional relevance as they could be representative of the transition from late orogenic [117] to backarc settings (i.e., Formia; Figure 2).  (Figure 2; Appendix B), the early Miocene CBZ limestone was found disconformable on the Jurassic-Cretaceous limestone, which is marked by a hardground (structural details in Section 4.2.1).

New Stratigraphic Constraints
Atop the Meso-Cenozoic carbonate units, the Chaotic complex occurs as a mélange that contains both native and exotic blocks, the latter being Cretaceous to Miocene basinal to distal ramp deposits that are coeval with the in situ formerly described proximal succession ( Figure 5). Both block types are internally folded. South of Carpineto Romano (Figure 3), the deformed platform blocks involved within the Chaotic complex are stratigraphically comparable with the encrusted carbonates that are preserved at the top of the Lower Volsci Unit (cf. Figure 2). In particular, within the Chaotic complex, we have mapped several lenses of Cenomanian to early Campanian limestones covered by middle Campanian karstic breccias and ferruginous to limonitic cruston ( Figure 5; structural details in Section 4.2.2).  Differently from the native blocks, the Scaglia-type pelagic to hemipelagic limestones (with rare planktonic foraminifera and iron oxides) occur as exotic inclusions. In this category, at Carpineto Romano and Caccume Mt. (Figures 2 and 5; Appendix B), we have found CBZ blocks of early Miocene age represented by red dots (iron oxide spherules) glauconitic calcarenite associated with micaceous intercalations and chert ( Figure 5). Minor lenses of hemipelagic middle Miocene marl and sandstone occur as well. Overall, the blocks are wrapped within a sandy-clayey matrix that is alternated with shales, foliated brownish marl, greenish arenaceous beds with exotic lithic, and coarse-grained micro-conglomerate with carbonatic and crystalline elements.
The matrix of the Chaotic complex at the base of the Caccume and Siserno mounts, includes Paleocene-Eocene, Oligocene-early Miocene, middle Miocene, and perhaps also late Tortonian-Messinian nannofossil assemblages (Appendix B). A similar wide span of ages was obtained from the shaly units of Colle Cantocchio (Figure 2), where Mesozoic to Tortonian nannoplankton reworked specimens were found beneath a major thrust (Appendix B; see also Section 4.2.1).
In the Latin Valley, the nannoplankton from the Frosinone Fm. can be referred, although rare or hardly diagnostic, to late Tortonian time. Wedge-top conglomerate deposits were studied at two key localities. At Gavignano (Figure 3), folded calcareous conglomerate occurs atop karstified Cenomanian limestones that according to the well data are juxtaposed on arenaceous deposits (cf. Figure 3). The clasts of mixed origin are from the Upper Cretaceous carbonates (i.e., Coniacian-Campanian and Albian-Cenomanian; see also Farinacci, 1965) and from the Tortonian Orbulina Marl Fm. The embedding matrix is made of abundant quartz grains along with reworked Amphistegina and Elphidium that make it possible to refer the whole Gavignano clastic deposit to the MVP unItal. In particular, the fining upward series with rare sandy matrix at the base (LEP10L) are dated to the latest Tortonian-earliest Zanclean and the clay marl at the top (LEP10M) to the Messinian. Thus, we consider this topmost constrain as indicative of the Messinian age of the MVP unit in the Latin Valley.
Within the eastern Lepini backbone, the conglomerates of Gorga are composed of pebbles and rounded blocks of reworked conglomerates whose clayey matrix and a bioturbated marly pebble were investigated. The age of these samples is late Tortonian for the marly pebble due to the presence of the coccolithophore Discoaster surculus, and top Tortonian-earliest Zanclean for the clay matrix bearing the marker Amaurolithus primus.

Structural Analysis of the Volsci Range
In this section, we document the field data used to reconstruct a geological cross section across the northern Volsci Range. The Western Lepini Mounts essentially consist of a 3 km thick Jurassic to Cretaceous carbonates dipping to (E)NE, whose local variations are shown in the stereoplots from 1 to 6 in Figure 6. The Neogene lithostratigraphic units atop are locally preserved beneath a few klippen structures that we document in detail in the next paragraphs. In the map and in the cross section of Figure 6, two areas are highlighted and described in detail as they preserve novel insights about pre-orogenic and syn-orogenic tectonics, which are presented from the oldest to the youngest event.
Near the western edge of the Western Lepini Mounts, a detailed survey performed at Colle Cantocchio allowed us to update the previous work by providing details on the stratigraphic contacts and fault kinematics ( Figure 7). In particular, we integrate the data from in [93] by describing the pre-orogenic contacts and the low-angle fault juxtaposing Cretaceous rocks onto the Orbulina Marl Fm. As we can see from the panoramic view and cross section ( Such an inherited tectono-stratigraphic setting is preserved at the footwall of a thrust, whose hanging wall consists of a one-hundred-meter-thick pile of Upper Cretaceous (earlymid Campanian) limestone, and whose base constitutes the roof of a cave. The cave is defined by an iron oxide-rich striated principal slip surface. In the hanging wall, cataclastic bands are crosscut by minor mirror-like faults. As constrained by nannoplankton analysis on samples from the fault core, both clasts and matrix (see Appendix B) are representative of different levels of a basinal sedimentary succession. The cataclasite also includes fragments of calcite mineralizations. The internal fabric is marked by the occurrence of slip surfaces associated with transpressive S/C structures indicating top-to-the-NE thrusting. Overall, the thrust seems to cut up-section although bounded and possibly tilted by later normal faults. The NW edge of the cave is bounded by a NE-striking normal fault with a displacement in the order of 20-40 of meters (red line in Figure 7h). At the top of the hill, the overall structure is topped by transgressive polygenic marine breccia composed by Miocene and Cretaceous calcareous clasts with a reddish cement and calcareous matrix, possibly crosscut by a SW-dipping normal fault with a displacement in the order of 150 m.     At Occhio di bue locality (plot-5), a block of middle Miocene limestones and marls with chert topped by light green clay of late Serravallian age (c.f.,  is affected by S/C structures indicating top-to-the NE shear. Coherently, at the contact with the Cenomanian limestone on top, 1-2 m of foliated proto-cataclasite bands are topped by (E)NE verging folds (plot-6; Figure 8). In the same plot, top-to-the-NE striated bedding is reported as it crops out more to the north at the top of the same lithon. While bedding is folded around N-to NNW-striking axes (cf. stereoplots 7-8; Figure 6), northeast of a major backthrust it is folded around NW-striking axes of folds (stereoplots 9-10).
As the Chaotic complex is concerned, field data from the Eastern Lepini Mounts highlight the top-to-the-ENE juxtaposition of the Upper Volsci unit above the Chaotic complex (i.e., Caccume Mt., Siserno Mt.), which in the Volsci Range is preserved in a few klippen atop the Lower Volsci Unit, whereas in the Latin Valley it is found on top of the Frosinone Formation (Figures 9a and 10a). At the Caccume Mt., we report structural information from the juxtaposition of folded Cenomanian Lower Cretaceous limestone on the Chaotic complex. The regional folding affecting the Lower Volsci Unit defines a well-marked NW-striking open fold while the Upper Volsci unit of the Caccume Mt. displays rather dipping beds folded around an NNW-striking axis. The basal contact of the Chaotic complex is marked by thrust grooves and ferruginous faint slicken lines along the crustons, while at the top of the Chaotic complex, S/C and C' structures display topto-(E)NE shearing. Cross-cutting field relationships show that thrust grooves are further cross-cut by high-angle en-échelon shear zones and normal faults.
with inferred right-lateral kinematics (Figure 9). More to the east (plot-18), the N-S trending flank of the salient is associated with transpressive S/C structures in Cretaceous limestones (plot 19). Overall, the fold-and-thrust fabric is cross-cut by NE-dipping normal faults at the northeastern VR edge. As it is downfaulted, the thrust front does not outcrop further north. In the VR, a salient has been mapped between Morolo and Patrica ( Figure  9), its most external point being characterized by the outcrop of Jurassic limestones. Upper Cretaceous units occur as klippe above the imbricated Chaotic complex juxtaposed to the foredeep deposits of the Frosinone Fm. later crossed by oxides-rich (D2+3) en-echelon fractures and later NW-striking oxides-free and cemented veins; 41°34′46″ N/13°13′60″ E). (e) Upper thrust juxtaposing the Cenomanian neritic limestone over the Chaotic complex (41°34′15.00″ N/13°13′55.13″ E), which, as shown as the sampling site of LEP67 on a lithotype that in panel (f), is affected by top-to-the-(E)NE S/C structures.
At the southern edge of the studied area of the Latin Valley (Figure 10a), the Chaotic complex was mapped as juxtaposed on the Frosinone Fm., and it reaches its maximum thickness west of the Siserno Mt. (about 250 m).

The Volsci Range Thrust Front and the Latin Valley Structures
The geometries of the frontal part of the Volsci Range and Latin Valley are shown from the SW to the NE (stereoplots 11-15, Figure 9). The thrust front between the Ernici and Lower Volsci units occurs as a series of imbricates of overturned Cretaceous to CBZ layers (i.e., NW of Morolo; Figure 9). New data allowed us to recognize a salient at the front of the Eastern Lepini Mounts. This structure is accompanied by a change in the fold trend from NW to W (plots 12 and 13; Figure 9) and by transpressive top-to-the-NE kinematics. The frontal part is defined by a large-scale anticline in the west and a syncline in the east (Figure 9). The two folds are separated by a series of NNW-striking tear faults with inferred right-lateral kinematics (Figure 9). More to the east (plot-18), the N-S trending flank of the salient is associated with transpressive S/C structures in Cretaceous limestones (plot 19). Overall, the fold-and-thrust fabric is cross-cut by NE-dipping normal faults at the northeastern VR edge. As it is downfaulted, the thrust front does not outcrop further north. In the VR, a salient has been mapped between Morolo and Patrica ( The channelized facies is made of arenaceous-pelitic associations with sets of thin pelitic-arenaceous and marly beds intercalated in thick massive arenaceous-pelitic layers. Southwest of Ferentino, paleocurrents are marked by a NW-SE direction, whereas the Frosinone formation is internally deformed and displays verticalized to overturned successions (Figures 9 and 10). There, the facies consists of an arenaceous association of amalgamated massive beds with arenaceous-pelitic and pelitic-arenaceous sets. As shown on the map (Figure 3), north of Sgurgola and north of the Siserno Mt., an anticline with upper Cretaceous and CBZ limestone belonging to the Ernici Unit emerges from the Latin Valley siliciclastics, which are locally bioturbated. In the syncline between this ridge and the Volsci Range, pelitic facies of the Frosinone Fm. occur.
At Gavignano (Figure 10f), the MVP Messinian calcareous conglomerate occurring on top of the Upper Volsci Unit overthrusting the Frosinone Formation is folded along an NNW-striking axis and is near vertical in places. In the most calcareous layers, pressure solution seams and veins crosscut the pebbles as typical of load-driven compaction.

Backthrusts and Normal Faults
Backthrusts best crop out in the northwestern part of the VR, where their presence is highlighted by some pockets of Messinian-earliest Pliocene heterogeneous conglomerate ( Figure 11). Transpressive kinematics associated with a general top-to-the-(E)SE sense of thrusting was observed on the reverse faults along the Montelanico-Carpineto Backthrust. As typical of cannibalized wedge-top basins, blocks of conglomerates occur within a marly-conglomeratic matrix near Gorga ( Figure 11).
In Figure 11, we sketch the structural setting related to the backthrusts, which cross-cut and preserve the top-to-the-(E)NE Chaotic complex at the footwall of the Montelanico-Carpineto Backthrust. This major backthrust (i.e., Montelanico-Carpineto Backthrust) bounds the East Lepini structure, a large-scale anticline with its culmination at the Malaina Mt. (Figures 6 and 11). The backthrust is accompanied by recumbent folds and minor high-angle reverse faults. In the southwestern sectors of the VR (Figure 11), normal faults cross-cut older contractional structures. More to the SW, another high-angle backthrust was mapped west of Bassiano ( Figure 6). This structure allows the juxtaposition of the Jurassic and Early Cretaceous carbonate onto the upper Cretaceous and it is defined by transpressive kinematics (stereoplots in Figure 6).

Seismic Interpretation of the Latin Valley
By tracing the reflectors of the unconformable contact between the Meso-Cenozoic carbonates and the upper Miocene siliciclastic deposits on top (cf. Section 3.4), two major The Lower Ernici unit, apart from the distinctive near-top reflections, displays a variable amplitude and frequency with a discontinuous and chaotic pattern of reflectors that generally is characterized by noisy seismic facies. We exclude that this reflector is a coherent noise (multiple) as it can be followed over the entire study area and it displaces geometries that roughly differ from the above reflectors. Due to the scarce penetration of the seismic signal, this unit can be considered as the acoustic substratum of the area. No boreholes reached this unItal. By comparison due to our reconstruction of the thrust geometry, the top of the Lower Ernici seismic unit is possibly represented by the Meso-Cenozoic carbonates that crop out northeast of the Latin Valley ( Figure 2). Due to the above reported uncertainty, marks indicate the less-constrained portions of the interpreted cross sections.
Within the Latin Valley, minor thickness changes of the carbonate tectonic units occur. Due to the repetition of the top-CBZ reflector accompanied by underlying top-UK reflectors, we have recognized multiple repetitions of the Upper Ernici unit due the occurrence of several thrust faults. The Ripi I well [106]), although crossing a major thrust zone, shows no siliciclastic deposits under the Mesozoic carbonates, but rocks of the Orbulina Marl and CBZ formations.
To show the general structural trend of the research area, we present three representative seismic lines (Figure 12), constrained by field and borehole data, showing thrust sheets characterized by a general top-to-the-NE sense of shear. Major thrusts, although occurring in all of the seismic lines, are well evident but discontinuous in number and distribution from line to line. Four major groups of thrusts form before the occurrence of normal faulting (Figure 13). From the most internal to the outermost we describe them as (1)   displacement up to 1-2 km. In Figure 12, normal faults with appreciable offset were identified (labeled with number 5). NE-striking faults concentrate at the Latin Valley edges and do not clearly show in seismic lines. NW-striking faults bound Quaternary graben, where travertine, continental, and volcaniclastic deposits were cumulated. The normal fault trace in seismic lines was drawn when it is anchored to the outcrop evidence. In these cases, we have extended the minimum offset recognized at surface to the deeper structural levels.  The most prominent of this group of thrusts generates the outcrop of basal platform at the foothill of the VR Front. A few backthrusts were recognized at depth, with vertical displacement up to 1-2 km. In Figure 12, normal faults with appreciable offset were identified (labeled with number 5). NE-striking faults concentrate at the Latin Valley edges and do not clearly show in seismic lines. NW-striking faults bound Quaternary graben, where travertine, continental, and volcaniclastic deposits were cumulated. The normal fault trace in seismic lines was drawn when it is anchored to the outcrop evidence. In these cases, we have extended the minimum offset recognized at surface to the deeper structural levels.
The most distinctive unconformities occur at the top of the Mesozoic carbonate succession and above the Middle Miocene CBZ Fm., onlapped by late Serravallian-early Tortonian UAM horizons (Section 3 in Figures 12 and 13). At the borehole scale this contact may appear as a paraconformity but the discontinuous and variable thickness of both CBZ and UAM suggest that this is actually an unconformity with an irregular erosional surface. Three subunits, divided by two major unconformities, can be observed within the siliciclastics deposits and labeled as Lower Frosinone seismic subunit (FFS1), Upper Frosinone seismic subunit (FFS2), and Monte San Giovanni Campano seismic unit (MVP); the first two are made by the late Tortonian Frosinone Fm., while MVP is formed by the Messinian piggyback deposits (Monte San Giovanni Campano unit; see MVP in Figures 2 and 12).
The thickness of the syn-orogenic units varies depending on the fold-and-thrust belt structure, being the siliciclastic deposits thicker to the south and to the north (up to 0.600 sec) and thinner in the central part (usually limited to 0.180 sec). As shown in Figure 13, Subunit FFS1 is folded together with the underlying carbonates, showing a transparent seismic facies, while Subunit FFS2 is thicker in the syncline and thinner towards the anticline and it is possibly related to Thrust-2. In FFS2, minor internal unconformities, typical of syn-depositional antiforms in foredeep basins, are here expressed by lobatetype seismic facies. In detail, the antiformal-growth geometries are crestal erosional truncations and diverging/converging reflection patterns around the hinge of the anticlines. In the piggyback basins, the FFS2 is defined by well-reflecting horizons and is marked by an erosive unconformity that at Ceprano cross-cuts both FFS1 at anticline culminations ( Figure 13). This anticline is sealed by FFS2 and is formed on top of Thrust-3. As shown by the strike section in Figure 13, the thrust-and-fold geometry changes laterally as also reported for the Gavignano klippe more to the north.

Discussion
The tectono-stratigraphic analysis of field and subsurface data enabled us to define different thrust units, providing insights for a time-deformation analysis of one of the innermost portions of the Central Apennines. Hereby, we present a geological cross section, interpretative of the deep structures produced after the integration of field and subsurface structures (Figure 14), that includes pre-orogenic passive margin deposits, mélange units, foredeep, and wedge-top deposits. In the following, we discuss the main novel features of the geologic history that led to the development of the geological setting of Figure 14.
In the cross section, we correlate the Upper Volsci Unit remnants of the Colle Cantocchio, Carpineto Romano, and Caccume Mt klippen. Based on the mixed exotic-native composition of the blocks of the Chaotic complex, we recognize that they were overthrusted together with the Upper Volsci Unit on top of the Lower Volsci UnItal. As shown in the cross section, the Lower Volsci Unit of the Western Lepini Mounts is a monocline essentially composed of Jurassic to Cretaceous carbonates dipping to (E)NE, that together with the remnants of the upper units was further crossed by high-angle faults. In detail, the Montelanico-Carpineto backthrust, bounds the Eastern Lepini pop-up that is affected by small-scale folds and reverse faults, whose geometry suggests positive reactivation of pre-orogenic normal faults during shortening. The wedge-top pockets preserved by the backthrusts are infilled by MVP Messinian conglomerate that was deposited directly on the Lower Volsci Unit, when the Upper Volsci unit was already dismantled. Thrusts and folds are mostly evident in the Latin Valley (Figure 12), whose substrate has been reconstructed by applying a depth conversion on a structural model published in [45].
By studying the top of the Mesozoic carbonate platform both in the Lower Volsci Unit and in the blocks embedded in the Chaotic complex (Appendix B; Figure S1), we have reported the occurrence of an irregular surface at the top of the platform. Such a paleotopography was likely the result of Late Cretaceous syn-sedimentary tectonics. In such scenario, the most elevated structures might have been affected by karstism (possibly with the formation of ferruginous crustons) during the latest Cretaceous (see Section 4.1). The occurrence of a Late Cretaceous tectonics is supported by the lithostratigraphic unit we refer to the "Gorga bioclastic limestone and dolostone" upper Campanian to Maastrichtian in age, whose lateral change and abrupt facies shift points to syn-depositional tectonics (Figure 2). At Gorga (Figure 3), this unit is represented by about 250 m thick rock volume [112], that thins rapidly towards the west, whereas it lacks in the rest of the Volsci Range. In particular, as recognized at Caccume Mt. and near Carpineto Romano ( Figure 5), the unconformity occurring at the top of the platform is marked by a very thin younger breccia partially overprinted by a dolomitic and ferrougeneous cruston (cf. Figures 5 and 9), whose age and origin need to be further constrained.
The occurrence of a Late Cretaceous tectonics is supported by the lithostratigraphic unit we refer to the "Gorga bioclastic limestone and dolostone" upper Campanian to Maastrichtian in age, whose lateral change and abrupt facies shift points to syn-depositional tectonics (Figure 2). At Gorga (Figure 3), this unit is represented by about 250 m thick rock volume [112], that thins rapidly towards the west, whereas it lacks in the rest of the Volsci Range. In particular, as recognized at Caccume Mt. and near Carpineto Romano ( Figure  5), the unconformity occurring at the top of the platform is marked by a very thin younger breccia partially overprinted by a dolomitic and ferrougeneous cruston (cf. Figures 5 and  9), whose age and origin need to be further constrained. In the Apennine platform, the transition from the Upper Cretaceous carbonates to Paleocene-Eocene margin, slope, and Scaglia-type basin deposits was guided by a synchronous regional extension during Maastrichtian-Eocene time that affected both the Jurassic base-of-slope domains [30] and the demised sectors of the neritic platforms [118]. We recognize that the discordant stratigraphic contacts of Colle Cantocchio are due to the development of a submarine paleoescarpment, guided by normal faults down-stepping towards the WSW. The bluish hardground (highlighted by yellow dots Figure 15) can be interpreted as a submarine unconformity marking the onlap (escarpment contact) of the lower Miocene intraformational pebbly calcarenite on the Mesozoic carbonates. Similar facies have been reported elsewhere by the authors of [119] and are here interpreted as a diagenetic effect on the articulated inherited physiography of the previously unedited fault escarpment described in Figure 6. A simplified back-restoration of section C-D (Figure 7c) is attempted in Figure 15, where a fault step occurred to the south with an offset in the order of 700-1000 m due to the exposure of the Jurassic terrains and the downthrowing of the Cretaceous units in the hanging wall. The Semprevisa Fault can be still recognized laterally for over 10 km, although overprinted by later Pliocene-Quaternary tectonics, and possibly remarks at least part of this inherited structure. In our interpretation, as shown by the stratigraphic contacts, the Jurassic units of the southwestern slope of the Semprevisa Mt. were already exposed in early Miocene time ( Figure 15). As suggested by the clasts within the Chaotic complex, coeval basinal sedimentation occurred more to the WSW [120]. In particular, the recognition of Cretaceous-Paleogene Scaglia lithotypes and of distal early Miocene CBZ limestones in the exotic blocks of the Chaotic complex (see Figures 7,9 and 10) suggest that sedimentation occurred in a bypass slope setting during Paleogene-Neogene time. In particular, the Paleogene is recorded by a condensed to hem- In the Apennine platform, the transition from the Upper Cretaceous carbonates to Paleocene-Eocene margin, slope, and Scaglia-type basin deposits was guided by a synchronous regional extension during Maastrichtian-Eocene time that affected both the Jurassic base-of-slope domains [30] and the demised sectors of the neritic platforms [118]. We recognize that the discordant stratigraphic contacts of Colle Cantocchio are due to the development of a submarine paleoescarpment, guided by normal faults down-stepping towards the WSW. The bluish hardground (highlighted by yellow dots Figure 15) can be interpreted as a submarine unconformity marking the onlap (escarpment contact) of the lower Miocene intraformational pebbly calcarenite on the Mesozoic carbonates. Similar facies have been reported elsewhere by the authors of [119] and are here interpreted as a diagenetic effect on the articulated inherited physiography of the previously unedited fault escarpment described in Figure 6. A simplified back-restoration of section C-D (Figure 7c) is attempted in Figure 15, where a fault step occurred to the south with an offset in the order of 700-1000 m due to the exposure of the Jurassic terrains and the downthrowing of the Cretaceous units in the hanging wall. The Semprevisa Fault can be still recognized laterally for over 10 km, although overprinted by later Pliocene-Quaternary tectonics, and possibly remarks at least part of this inherited structure. In our interpretation, as shown by the stratigraphic contacts, the Jurassic units of the southwestern slope of the Semprevisa Mt. were already exposed in early Miocene time ( Figure 15). As suggested by the clasts within the Chaotic complex, coeval basinal sedimentation occurred more to the WSW [120]. In particular, the recognition of Cretaceous-Paleogene Scaglia lithotypes and of distal early Miocene CBZ limestones in the exotic blocks of the Chaotic complex (see Figures 7,9 and 10) suggest that sedimentation occurred in a bypass slope setting during Paleogene-Neogene time. In particular, the Paleogene is recorded by a condensed to hemipelagic sedimentation, evolving during the Miocene to mixed calcareous-siliciclastic turbidites with chert. The Orbulina Marl Fm. (Serravallian pp.) sealed the pre-orogenic early Miocene topography. The Colle Cantocchio pre-orogenic fault is a part of the normal fault system that produced the steps from the exposed Jurassic carbonates to the basin and is here proposed to be at least Eocene in age, although older ages cannot be excluded. Synthetizing, according to the new data, we propose a provenance of the Chaotic complex (i.e., including the exotic blocks) from a hemipelagic paleogeographic domain with slow depositional rates placed to the WSW of the present-day Volsci Range.
inherited top-platform physiography and thrust geometry. In our interpretation, this structure is an inherited depression at the top of the platform that was later crosscut by the Upper Volsci Thrust. At its southern tip, as demonstrated by Accordi [71], this thrust still occurs as it doubles of the upper Cretaceous units although not involving anymore the Chaotic complex, whereas, as shown on the map (Figure 6), at the northern of the Upper Volsci Thrust, the younger Montelanico-Carpineto backthrust cross-cut it ( Figure  11). The Upper Volsci Unit is mainly composed by Upper Cretaceous neritic carbonates (e.g., Carpineto Romano, Figure 8), implying that this unit detached essentially above the uppermost Lower Cretaceous Orbitolina Marl level during shortening. However, although rare, older Mesozoic rocks can also be found. A second detachment level, highlighted by subsurface data, corresponds with the Orbulina Marl Fm, which allowed the doubling. The chronological relationship between Thrust-1 (marking the overthrust of the Upper Volsci Unit on to the Upper Ernici unit) and Thrust-2 (between the Ernici Units of The ongoing research in the southern Volsci Range, is providing constraints for the determination of the age of the encrusted normal faults bounding the Formia plain and Spigno Saturnia areas, whose data from the literature are reinterpreted above (cf. Figure 2). A comparable syn-sedimentary setting, leading to the deposition of Scaglia deposits has been recorded nearby the VR [8,121] and documented at the western tip of the Volsci Range [122]. Of note, at Colle Cantocchio ( Figure 5), the early Miocene transgression over the Jurassic-Lower Cretaceous rocks occurred on a step of the escarpment, where there was no record of Paleogene basinal sedimentation. In alternative, this sector could be associated with renewed normal faulting activity along a pre-existing Cretaceous-Paleogenic normal fault, which may have further exposed the Mesozoic rocks with its reactivation and allowed the CBZ-UAM units to settle on top prior to the Tortonian onset of thrusting.

Chaotic Complex Emplacement and Thrust Propagation
To define the overthrusting towards the (E)NE of the Upper Volsci Unit and to understand the evolution of the Chaotic complex, we correlated the carbonate klippen by documenting the stratigraphic and structural elements of the syn-orogenic deposits. This correlation was initially proposed by Accordi [71], but inherited structures, thrust kinematics, and age of the syn-orogenic deposits needed to be better constrained. With the degree of allochthony and origin of the Chaotic complex being long debated [45,64,67,86,123], in this section we discuss the Chaotic complex origin and the role of the thrust propagation towards the foreland into the late Miocene wedge growth.
Starting from the southwest, the Colle Cantocchio cataclasite and shale preserved underneath the Upper Volsci Thrust can be interpreted as a thin Chaotic complex unit juxtaposed on the paleo escarpment setting (c.f. Section 5.1). In this frame, the inherited topography produced a ramp in the upper thrust during shortening. A comparable setting occurs more to the south at the Vele Mt. (Figure 2), where the siliciclastic deposits underneath the thrust could be correlated with the Chaotic complex sliver of Colle Cantocchio ( Figure 15). As commonly occurring in mélange complexes [124,125], the Chaotic complex formed at the expenses of the Lower Volsci Unit, whose inherited and articulated top was scraped off and grooved (see . The Chaotic complex is a combination of (i) autochthonous "native" and (ii) allochthonous "exotic" blocks ( Figure 15). The latter derive form a discontinuous series of Paleogene-Burdigalian pelagic deposits deposited more to the south and progressively mixed with lower Serravallian to upper Tortonian siliciclastic units bearing also crystalline clasts.
In particular, the matrix of the Chaotic complex shows the same composition of the embedded blocks, but it also shows the occurrence of late Tortonian-Messinian nannofossil assemblages, which may have deposited during the final stage of thrusting related to the Upper Volsci Thrust. Further, we are able to further narrow this time range to the late Tortonian, considering also the absence of Amaurolithus sp., typical marker of top Tortonian-Messinian. Provided that the overthrust of the pelagic elements of the Chaotic complex is due to the juxtaposition of the Upper Volsci Unit, which squeezed them out towards the foredeep, they must have originated from about the same distance reached by the Upper Volsci Thrust front (Thrust-1).
In this frame, the SE-ward termination of the Chaotic complex and the lens-like shape of the outcrop at Carpineto Romano ( Figure 6) provide an example of interaction between inherited top-platform physiography and thrust geometry. In our interpretation, this structure is an inherited depression at the top of the platform that was later crosscut by the Upper Volsci Thrust. At its southern tip, as demonstrated by Accordi [71], this thrust still occurs as it doubles of the upper Cretaceous units although not involving anymore the Chaotic complex, whereas, as shown on the map (Figure 6), at the northern of the Upper Volsci Thrust, the younger Montelanico-Carpineto backthrust cross-cut it ( Figure 11).
The Upper Volsci Unit is mainly composed by Upper Cretaceous neritic carbonates (e.g., Carpineto Romano, Figure 8), implying that this unit detached essentially above the uppermost Lower Cretaceous Orbitolina Marl level during shortening. However, although rare, older Mesozoic rocks can also be found. A second detachment level, highlighted by subsurface data, corresponds with the Orbulina Marl Fm, which allowed the doubling. The chronological relationship between Thrust-1 (marking the overthrust of the Upper Volsci Unit on to the Upper Ernici unit) and Thrust-2 (between the Ernici Units of the Latin Valley) is beneath the resolution of our data. However, provided their geometrical distribution, these thrusts are likely to represent a classical thrust propagation towards more external and lower structural levels through time (i.e., towards the foreland). The minimal shortening associated with Thrust-1 is of about 25-30 km, which corresponds with the approximated present-day distance between Colle Cantocchio and the frontal klippe along the ENE-directed Thrust-1; while Thrust-2 ranges about 20 to 25 km as shown by the thrust-2 structures in Figure 14. These amounts are comparable with the shortening estimated at the thrust fronts of the Gran Sasso Massif (>20 km [30]) and of the Apennine platform in the southern Apennines (>60 km [126]), while it is significatively lower than the translation that affected the Ligurian Accretionary Complex onto the foredeep units (> 100 km [127]). In this frame, the Tortonian southern Apennine platform thrusting [28] matches our thrust dynamics ( Figure 15). As also typical of the far-traveled Sicilian platform units [128], the thrust geometry is characterized by long flats (10-15 km) and thin thrustsheets, that in our case can be as thin as about 0.7 km near the front. This implies that the Orbitolina level and Orbulina Marl Fm preferred slip levels were very efficient in allowing far-traveled thrusting.
As shown by thickness and facies variations of the siliciclastic deposits of the Latin Valley, the Thrust-2 shortening stage was accompanied by syn-sedimentary folding of the deposits of the FFS2 seismic unit (Figure 13). In our interpretation, while the unconformable FFS1 contact with the CBZ limestone marks the flexuration of the foredeep, the unconformable contact associated with wedge shape and channelized FFS2 facies marks the growth of pop-up anticlines, thus being representative of wedge-top settings initially developed during Thrust-2.
The channelized facies may be, respectively, representative of syn-tectonic fringe and lobe deposits and of inner channelized sand bodies, while pelitic facies are rather typical of outer fans [129]. In particular, the observed syn-sedimentary folded channelized structures (Figure 10b), show that, the deposition of the Frosinone Fm. thus encompassed an increasing input (mostly during the FFS1 stage), later followed by a progressive channelization of turbidity flows onto the synclines during the FFS2 stage. As already suggested in [130] for the Latin Valley on the channelization of the foredeep to wedge-top sediments, the active margin possibly followed a comparable evolution similar to what elsewhere envisaged in the southern Apennines by Casciano et al. [131].
At the front of our study area, a transition between the mélange and the flysch units occurs. Based on published maps [64], wells, and seismic lines on the southwestern edge of the Latin Valley, we also confirm that the Chaotic complex is juxtaposed to the Frosinone Fm. of the upper Ernici unit (cf. Gavignano; Figures 10 and 13). For this feature, the authors of [132] proposed an olistostrome origin, while Centamore et al. (2007) proposed gravitational sliding of the Chaotic complex off the Volsci Upper UnItal. Further, this level can correlate with the mélange levels of the Massico Mt. [43,133].
To explain the abrupt thickening of the Chaotic complex east of the Caccume Mt. (Figure 10), we suggest that a growth structure was forming during the initial uplift of the Volsci Range front as testified by fault-propagation fold ( Figure 14) at the hanging wall of thicker FFS units with syn-sedimentary folds (Figures 10 and 12). This generated the glide of the Chaotic complex on top of the FFS units. Similar contexts were reconstructed for other mélange units at thrust fronts, where the remobilization of the formerly emplaced thrust sheets, allows the incorporation of the extrabasinal (exotic) lithologies within the foredeep [18,134]. An alternative possible explanation to allow the juxtaposition of the Upper Volsci unit onto the FFS units, would envisage thrusting to occur during the uppermost Tortonian-earliest Messinian.

The Late Stages of Shortening
As observed in seismic lines (Figures 12 and 13), thrust-3 produced the doubling of the flat of the far-traveled Thrust-2, by involving deeper carbonates in the thrust ramps. We have also shown that in the area break back thrusting occurred [135] (Figure 16). As shown near Ceprano well (Figure 13), MVP wedge-top deposits that include calcareous pebbles from the CBZ unit [87] were directly deposited on Mesozoic carbonates deformed by an anticline. This contact is representative of a wedge with regional subsidence slower than local antiformal growth [136]. Nannoplankton determination finally allowed constraining the age of the folded conglomerates and atop marls of Gavignano, thus allowing a correlation with the MVP stratigraphic unit (Figure 10). This unit represents a folded Messinian thrust top deposit and this constraint attributes this late folding stage to late Messinian-earliest Pliocene time. As supported by subsurface data (Figures 3 and 13) the Gavignano klippe was involved into the renewed deformation of the VR front, which would correspond with the latest stage of thrusting and veining dated in [43] at the late Messinian on the Massico Mt. (cf. Figure 2). Those absolute constraints can be used to review the regional thrust kinematics. In this sense, the ages determined along the thrusts in areas more to the south can be compared to what provided in [114]. These authors have attributed a late Miocene-Pliocene age to the clayey matrix beneath the thrust at the front of the Siserno Mt. Similar to what reported for the Chaotic complex in this work (Appendix B), they have also reported that the exotic clasts are representative of a wide range of ages, from Late Cretaceous (including Scaglia Rossa pelagic limestone) to early-middle Miocene. The degree of fragmentation of microfauna embedded within the Chaotic complex [114] suggests active deposition during the late Miocene-Pliocene as well. Therefore, we can envisage a late involvement of Pliocene deposits into the reactivated thrust zones at the VR front. In this interpretation, the Chaotic complex was already exhumed likely after the strong erosion related to the Messinian salinity crisis [137][138][139][140], which also affected the Ernici Mts [77], implying reactivation in the rear [49].
shown near Ceprano well (Figure 13), MVP wedge-top deposits that include calcareous pebbles from the CBZ unit [87] were directly deposited on Mesozoic carbonates deformed by an anticline. This contact is representative of a wedge with regional subsidence slower than local antiformal growth [136]. Nannoplankton determination finally allowed constraining the age of the folded conglomerates and atop marls of Gavignano, thus allowing a correlation with the MVP stratigraphic unit (Figure 10). This unit represents a folded Messinian thrust top deposit and this constraint attributes this late folding stage to late Messinian-earliest Pliocene time. As supported by subsurface data (Figures 3 and 13) the Gavignano klippe was involved into the renewed deformation of the VR front, which would correspond with the latest stage of thrusting and veining dated in [43] at the late Messinian on the Massico Mt. (cf. Figure 2). Those absolute constraints can be used to review the regional thrust kinematics. In this sense, the ages determined along the thrusts in areas more to the south can be compared to what provided in [114]. These authors have attributed a late Miocene-Pliocene age to the clayey matrix beneath the thrust at the front of the Siserno Mt. Similar to what reported for the Chaotic complex in this work (Appendix B), they have also reported that the exotic clasts are representative of a wide range of ages, from Late Cretaceous (including Scaglia Rossa pelagic limestone) to early-middle Miocene. The degree of fragmentation of microfauna embedded within the Chaotic complex [114] suggests active deposition during the late Miocene-Pliocene as well. Therefore, we can envisage a late involvement of Pliocene deposits into the reactivated thrust zones at the VR front. In this interpretation, the Chaotic complex was already exhumed likely after the strong erosion related to the Messinian salinity crisis [137−140], which also affected the Ernici Mts [77], implying reactivation in the rear [49].  In this context, the late Messinian shortening event could be correlated with the late orogenic structures in the northern VR that are crossed by a series of SW-directed backthrusts ( Figure 11). In our interpretation, the SW-directed Montelanico-Carpineto backthrust cross-cuts the top-to-the (E)NE older Upper Volsci Thrust. Despite the lack of valuable data from the main lineament, minor thrusts show that top to the SW-backthrusting, could be accounted as partially reactivating the older fabric. Further, the fault strike of the backthrusts diverges about 20 • from the trend of the upper Volsci Thrust that is underthrusted beneath the Eastern Lepini pop-up ( Figures 5 and 11).
So far, scarce constraints of top-to-the-SW shear were found, although backthrusting is possibly localized more to the NE of the studied area of Figure 11. Our stratigraphic constraints (Appendix B) from the MVP conglomerate near Gorga, document Messinain Lago-Mare conglomerates that are produced after iterative cannibalization of older wedgetop deposits. The further occurrence of upper Messinian deposits in the Pian della Faggeta area (Figure 5), is a possible clue indicating depositional activity on top of the Volsci Range during the Messinian salinity crisis (5.96-5.33 Ma). During that time, the area was exposed to linear erosion followed by the deposition of sandy gravels that Centamore et al. (2010) dated at the early Pliocene (south of Castro dei Volsci; Figure 3). This implies that the major valleys were already formed before the latest orogenic compressional events affected both the VR and Latin Valley [117]. Field evidence in the rear (Figure 5), suggests the presence of a major backthrust with transpressive kinematics further south, possibly implying that a deeper backthrust affected the southwestern slope of the VR during the early Pliocene. At that time, the Apennines experienced renewed shortening with frontal thrusting accompanied by backthrusting and tilting toward the foreland to the northeast ( Figure 16).
During late orogenic deformation, thrust front migrated towards the outermost active margin units (Figure 1), and the inherited fold-and-thrust belt of the external Apennines was folded together with lower Pliocene syn-orogenic conglomerates (i.e., Rigopiano conglomerate [30,78]). Meanwhile, the previous in-sequence structure of the internal Apennines was truncated by triangle zones (Figure 12) and by more internal backthrusts (e.g., in the Volsci Range, Figures 11 and 14).
In our interpretation (Figure 16), the backthrusting roots at deeper levels, by following the dip of the basal detachment towards the backarc. In this sense, moving to the inner parts of the wedge, the inner wedge is remobilized, affecting a larger volume with respect to the external part. In the case of late orogenic deformation affecting only the sedimentary cover, shortening localizes within the weakest stratigraphic levels, possibly by reactivating the décollement of the older fore-thrusts [136,[141][142][143][144][145], while in the rear faulting tends to broaden and possibly involve also deeper structural levels.
Finally, Pleistocene to Holocene NW-and NE-trending normal faults deeply affected the fold-and-thrust belt structure. In particular, the almost constant NE-dip shown by the bedding planes of the studied carbonates might be interpreted as the result of the activity of the major NW-striking and SW-dipping listric normal faults bordering the Pontina Plain, which were also documented at depth [146].

Conclusions
This study contributes to constraining the timing of initiation and progressive development of platform-derived thrust sheets, mélange units, foreland, foredeep, and wedge-top sediments of the internal Central Apennines. The main phases of the evolution of the belt are as follows:

1.
Late Cretaceous extensional tectonics. The dismembering of the carbonate platform into shallower and deeper domains is constrained by the finding of crustons that may testify moments of subaerial exposure, characterizing the top of the Lower Volsci UnItal. Cave exploration and field mapping allowed us to recognize a previously unreported fault-controlled paleo-escarpment constituted by Cretaceous and Jurassic carbonates sealed by early Miocene deposits that were previously dated as middle Miocene. These units seal a hardground settling on a platform edge facing to the west, where basinal to bypass slow-rate sedimentation occurred till Burdigalian time.

2.
Tortonian Chaotic complex emplacement (Thrust-1) and foreland-directed (in-sequence) thrust propagation (Thrust-2). During the overthrusting of the Upper Volsci Unit, Paleogene to Neogene basinal deposits were squeezed off towards the Foredeep and juxtaposed as a mélange unit on top of the carbonate platform together with early to middle Miocene calcareous-cherty-siliciclastics. The Chaotic complex also bears highly deformed basinal exotic and native blocks of neritic carbonates, the latter being scrapped off by the overthrust of the embedding Chaotic complex, whose Paleogene-Miocene matrix includes up to Tortonian nannoplankton. Seismic analysis supported by well logs at the regional scale highlighted repeated carbonate thrust sheets that have first been involved into an initial in-sequence propagation towards the foreland to the ENE occurred during foredeep to wedge-top sedimentation. Late Pliocene to Holocene normal faulting. Post-shortening extension has determined NE-and NW-striking orthogonal normal faults or WNW-ESE-trending right-lateral transtensional faults. These faults may have locally intercepted pre-existing normal faults that had been passively transported within the thrust sheets.
Finally, our findings bear implications on platform derived thrust sheets associated with active margin successions and mélange units. The far-traveled thrust sheets, hereby documented both in the field and in the subsurface, constitute a key aspect for the development of the internal Apennines, whose degree of allochthony and role of inherited structures was long debated. Furthermore, at the light of our new interpretation, the deeper platform units could be a new focus for hydrocarbon accumulation and may provide targets for geothermal and/or hydrocarbon research in the area. Beside the regional geological aspects, this work bears implications on the modes of involvement of mélange units at the transition from passive margin to foreland basin systems. Data Availability Statement: Data used for seismic interpretation and model reconstruction can mainly be found in the public VIDEPI database (www.videpi.com) (accessed on 20 January 2021) and in the Pentex Ltd. database. Data available at the ENI data room were also viewed.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Appendix A
In the following, we report the Biostratigraphic and lithostratigraphic data of outcrops and stratigraphic units available from the literature related to syn-orogenic deposits shown in the representative stratigraphic logs of Figure 2 in the main manuscript. The formation labels are also related to Figure 2. in situ [37,115] Appendix B

Group of
In the following, we report the new Stratigraphic constraints and age determination of the samples collected from twenty-five different localities in the study area representative stratigraphic logs of Figure 2 in the main manuscript.