The multiply deformed foreland fold-thrust belt of the Balkan orogen, northern Bulgaria

The Balkan orogen within Bulgaria and northern Greece is the eastern continuation of the European Mesozoic and Cenozoic Alpide folded chain. Within Bulgaria, northern Greece, and parts of Serbia and Macedonia to the west, the orogen consists of topographic units that are in part a reflection of their underlying geology (Figs. 1–3; e.g., see Boyanov et al., 1989; Dabovski et al., 2002; Zagorchev et al., 2009). The Balkan orogen is bounded on the north by the Moesian plain, which is underlain by a basement of Precambrian and Paleozoic rocks that were largely undeformed during alpine time, and by the Vardar zone (sensu lato) on the west and south (Schmid et al., 2008) and its continuation to the east in the Intrapontide zone, which contains the remnants of a complex assemblage of Mesozoic and early Cenozoic oceanic fragments (Fig. 2; e.g., Okay et al., 2001a, 2001b; Robertson et al., 2004). The main part of the Balkan orogen within Bulgaria consists of a northern belt of folded and thrusted Mesozoic and Cenozoic strata that forms the external fold-thrust belt of late Mesozoic and early Cenozoic age and a southern part that consists of deformed igneous and metamorphic rocks overprinted by Cenozoic extensional basins (Figs. 2 and 3). The fold-thrust belt is exposed within the Forebalkan, Stara Planina, and the northern part of the Sredna Gora topographic and tectonic units and is the main focus of this paper (Fig. 3). Although Stara Planina means the “old mountains,” it is a misnomer, as it is flanked on the south by the late Cenozoic to active extensional Sub-Balkan graben system, which is mostly responsible for its elevation (Tzankov et al., 1996; Roy et al., 1996). Within the Sredna Gora to the south are the remnants of a Late Cretaceous backarc or intraarc basin that has also been deformed; the structures are well exposed in the west, but are largely covered by younger strata to the east. Although these rocks are folded and thrusted, this belt of structures is generally not considered to be part of the external fold-thrust belt; however, we regard it as part of external fold-thrust belt (discussed herein). Farther south the rocks within the Rhodope Mountains and their continuation into Greece to the south consist of a complex of Precambrian(?), Paleozoic, Mesozoic, and early Cenozoic rocks metamorphosed and intruded by abundant magmatic rocks of Mesozoic and Cenozoic age (for a review see Burg, 2012). They were deformed principally during Mesozoic and early Cenozoic shortening events, and beginning in about middle to early-late Eocene time (Burchfiel et al., 2003), extensional tectonism became the dominant mode of deformation within central and southern Bulgaria with the formation of numerous extensional basins accompanied by abundant magmatic activity (Ivanov, 1988; Bonev and Beccaletto, 2007; Burchfiel et al., 2008).

The first description of the tectonics of Bulgaria was by Cvijic (1904) in his studies on the structure of the Balkan Peninsula. From south to north Cvijic (1904) distinguished the Rhodope Massif, Transitional zone, Balkan system, and Balkan plate (Moesian Platform in modern terminology). St. Bončev (1936) introduced the Balkanides as an orogenic system and subdivided it into three main units, from north to south, the Forebalkan, Balkan, and Srednogorie; he assigned the Moesian Platform to the north to the foreland of the orogen. The Srednogorie was accepted as the oldest tectonic element and in the strictest sense did not belong to the Balkanides. The boundaries between the units within the orogen were traced and interpreted in different ways based on different criteria (e.g., see Yovchev, 1971; Bonchev, 1986; Gocev, 1986; Ivanov, 1988; Dabovski et al., 2002). The differences in these interpretations reflect the complexity and unusual character of the orogen.

Useful published works that contain extended and detailed description of the structure include the tectonics of Bulgaria (Yovchev, 1971; geological maps of Bulgaria at scales of 1:100,000 [Cheshitev, 1990–1995] and 1:500,000 [Cheshitev and Kancev, 1989]). Our reconstructions concerning Bulgarian territory are based on these works as well as our own field work and field trips led by numerous Bulgarian geologists to many parts of the orogen.

The main focus of this paper is the fold-thrust belt that is exposed within the Forebalkan, Stara Planina, and part of the Sredna Gora topographic and structural units (Fig. 3). This fold-thrust belt has the unusual characteristic of having been deformed in at least three superposed different events of different ages, and deformation did not migrate from its internal to external parts, as in many fold-thrust belts. In addition, it is necessary to try to relate these three deformational events to orogen-wide processes. For at least the two older deformations (late-Early Cretaceous to early-Late Cretaceous and latest Cretaceous–early Paleogene) the dynamics are related to activity along the western and southern plate boundary occupied by the Vardar zone, but for the younger deformational event (middle Eocene to middle Miocene) the dynamics are related to activity within the Carpathian orogen to the north.

The Mesozoic and Cenozoic foreland fold-thrust belt of the Balkan orogen is best exposed within the Forebalkan and Stara Planina units and to a lesser degree within the western part of the Sredna Gora unit (Fig. 3). Rocks forming the fold-thrust belt are of Mesozoic and Cenozoic age and were deposited along the southern part of the Moesian shelf, the southern margin of the European continental crust. The fold-thrust belt also involves rocks deformed, metamorphosed, and intruded by plutons in pre-Mesozoic time that formed the crust below the Mesozoic Moesian shelf strata. Triassic to Middle Jurassic shallow-water deposits form the base of the Moesian Alpine succession and unconformably overlie pre-Mesozoic basement rocks; both sequences are exposed within the foreland fold-thrust belt and to the south in central Bulgaria (Fig. 4).

The pre-alpine basement rocks assigned to the Moesian shelf form the exposed parts of the Moesian craton in the foreland fold-thrust belt in Bulgaria and demonstrate that their overlying Triassic and Jurassic strata were deposited on continental crust (Fig. 4). Some Precambrian metamorphic rocks are present, but most of the basement consists of a thick succession of metasedimentary and locally metavolcanic rocks of late Precambrian to Carboniferous age. They were intruded by large bodies of middle to late Paleozoic igneous rocks and were metamorphosed from low to high grade. Locally they are unconformably overlain by Late Carboniferous–Permian terrigenous sedimentary rocks and nonmarine volcanic and volcaniclastic rocks that have been intruded by their plutonic equivalents (locally true for the volcanics but not for the sedimentary rocks). We regard all these rocks to be basement and parts of the Moesian craton before development of a passive margin during Triassic–Early Jurassic time.

The metamorphic rocks exposed south of the southern limit of known Mesozoic Moesian shelf strata probably mostly belong to the Moesian continental basement and possibly contain some metamorphosed strata of the Meosian shelf, but there are some rocks that may have been accreted in Mesozoic time (Fig. 4). Rocks of the Rhodope Mountains underwent extensive metamorphism and complex deformation in Mesozoic (and possibly Paleozoic) time, but the regional relations of these events are still being elucidated (e.g., see De Wet et al., 1989; Kilias et al., 1999; Zagorchev, 2001; Burg, 2012). Some units contain evidence for ultrahigh-pressure metamorphism, and diamonds have been reported from some of the rocks (Kostopoulos et al., 2000; Mposkos and Krohe, 2006; Bauer et al., 2007; Wawrzenitz and Mposkos, 1997; Liati et al., 2002). Some have suggested that the rocks in the southern part of Rhodope Mountains are an assemblage of continental fragments accreted at various times during the Mesozoic, the last accretion having occurred in Late Cretaceous time (e.g., Turpaud and Reischmann, 2010; Burg, 2012); however, we interpret most of the pre-Mesozoic rocks in Bulgaria as part of the Moesian continental basement by Mesozoic time. However, the width of the Moesian shelf in the south remains uncertain because of the magnitude of accretionary shortening and Cenozoic extension.

Nonmarine lower Triassic rocks transgress over the Permian and older basement rocks along an unconformity widely exposed in northwest and central Bulgaria (Fig. 4). Shallow-water marine conditions were established in late-Early Triassic time and extended through several transgressive-regressive cycles into Late Triassic time (Tronkov et al., 1965). Extension began in the Early Triassic and continued into the Jurassic; extensional structures are clearly evident in the subsurface of the main part of the foreland fold-thrust belt (see seismic section in Georgiev et al., 2001; cross sections in Vangelov et al., 2013). The faults show mainly down-to-the-south displacement with thickening of Late Triassic and Early Jurassic sedimentary units into the faults (Tronkov, 1963; Atanasov and Bokov, 1983). The Triassic to Early Jurassic shallow-water clastic and carbonate rocks are considered the earliest deposits of a south-facing passive margin in Bulgaria and began the Alpine period.

The position of the southern or western shelf edge for this continental margin is poorly known. In western Bulgaria is a small area of deep-water Jurassic strata (the Trekljano Group; Zagorchev and Tikhomirova, 1986; Fig. 5) that are above older shallow-water strata of the Moesian shelf and can be interpreted to be transitional into the oceanic rocks of the Vardar zone to the west (see following). All other rocks that are regarded as the deep-water equivalents to the shallow-water rocks of northwest and central Bulgaria are the allochthonous rocks in the Strandja Mountains of southeastern Bulgaria and the thin sequences of chert, slate, and fine-grained clastic rocks in northern Greece (the Circum-Rhodope belt; Figs. 4 and 5). The original tectonic positions of these allochthonous deep-water Mesozoic rocks are generally interpreted to have been south of the Moesian shelf and south of the metamorphosed Mesozoic rocks, and the basement that underlies them that is also considered to be part of the Moesian shelf (Georgiev et al., 2001; Okay et al., 2001a, 2001b). These autochthonous Mesozoic strata are poorly dated, but range from Early Triassic to Middle Jurassic (Chatalov, 1990), and are the southernmost strata of the Moesian shelf exposed in southern Bulgaria (Fig. 4). Thus, while it is clear that the Triassic and Early Jurassic rocks of northwest and central Bulgaria were deposited during extension, it is not clear how far these rocks were from the shelf edge of the Moesian continental margin to the south and west. Triassic deformation is known from rocks in the Strandja autochthon, but its nature is unclear (Figs. 4 and 5; Sakar unit). Passive margin sedimentation continued during most of the Early and Middle Jurassic and facies changes between different shallow-water environments can be determined throughout the area except in the southernmost part, where Jurassic rocks are rarely exposed or metamorphosed (Sunal et al., 2011).

Beginning in the late Kimmeridgian and extending into the Hauterivian two areas within the Moesian shelf subsided rapidly and several kilometers of flysch were deposited in the Nis and Trojan flysch troughs, which were later incorporated into the foreland fold-thrust belt (Fig. 6). Along the southeastern margin of these flysch troughs is a distinctive narrow belt of lower Triassic to Jurassic shallow-water and deep-water strata, the Kotel zone, characterized by a thick upper Triassic (Carnian to Rheatian) unit of black flysch that continues upward into a thick Middle Jurassic (Aalenian to Bathonian) sequence black shale with numerous blocks of Triassic carbonates (Georgiev et al., 2001; Tchoumatchenco et al., 2004), some of exotic stratigraphy. The interpretation of the Kotel zone rocks has been controversial. They have stratigraphic similarities to the allochthonous rocks in Strandja with which they are often correlated (Georgiev et al., 2001). The Kotel sequence is in thrust contact with more normal Moesian shelf deposits to the north, but are separated from the rocks in Strandja by a wide belt of younger rocks (Figs. 5 and 6). The original tectonic position of the Kotel zone has been subject to several different interpretations that are critical to the tectonics of the Balkan orogen. Gočev (1986) interpreted the Kotel zone as deposited south of the Moesian shelf and part of a far north-traveled Strandja allochthon. Other interpretations were proposed: Tchoumatchenko and Cernjvaska (1990) suggested a turbidite succession fringing the southern margin of the Moesian Platform; Bonchev (1983, 1986) suggested a Moesian marginal depression or a synkinematic mélange thrust northward over the Moesian Platform during the Mid-Cretaceous deformation; and Kanchev et al. (1995) proposed a rift basin within the Moesian shelf because some of the olistoliths are similar to rocks of the Moesian shelf. In the interpretations that the Kotel zone is far traveled, the rocks can be regarded as part of the Circum-Rhodope units. The interpretation of these rocks remains unresolved, but is important in the tectonic history of Balkan orogen (see following).

Another flysch succession, the Kraina subunit in northern Bulgaria (and extending into Romania), contains Berriasian to Barremian Sinaja flysch and currently forms a narrow belt of outcrops near the Serbian and Romanian borders (Figs. 6 and 7). These areas of subsidence and flysch deposition are now separated by a large and structurally disrupted carbonate platform. The tectonic position of the Sinaja flysch (and Kraina subunit) is uncertain and has been interpreted as either a far-traveled allochthonous flysch succession, like the Ceahlau unit of Romania (Sandulescu, l975) formed along the western continental margin of the Moesian Platform, or a flysch trough developed within basins that extend into the Moesian passive margin from the north at the northern end of a peninsula that terminates the Moesian unit within the Southern Carpathians (shown in fig. 9 of Fügenschuh and Schmid, 2005). In figure 9 of Fügenschuh and Schmid (2005), this peninsula is shown as extending into the Southern Carpathians as the Danubian unit (Fig. 6). The relations between the Kraina flysch unit and the Nis-Trojan flysch basins to the south within Bulgaria remain uncertain (see following). Okay et al. (2001a) and Georgiev et al. (2001) interpreted the flysch troughs to be a foredeep for the Strandja allochthon, but there is little evidence that the allochthon extended into western Bulgaria south of the troughs, and they may also be interpreted as inverted rift basins that developed into basins related to the first major deformation of the foreland fold-thrust belt in mid-Cretaceous time.

The now-separated flysch troughs appear to have extended across northern Bulgaria to the Black Sea; their age and sedimentary faces are considered to be the continuation of the oceanic Severin and Ceahlau units of the Carpathians. There is no evidence in Bulgaria that they contained an oceanic crust, unlike the rocks in the Carpathians; however, the relations of these flysch troughs to sedimentary environments that may have been deposited on oceanic crust within the Southern Carpathians remain unexplored.

The first deformation event within the Balkan fold-thrust belt was late-Early Cretaceous to early-Late Cretaceous, and deformed the flysch units within Nis-Trojan flysch basins, emplaced the Kraina unit and the Severin allochthon, their probable northward continuation in the Southern Carpathians, and strata broadly adjacent to these flysch basins (Figs. 7 and 8). This event is well documented and was often assigned to the Austoalpine phase of tectonism in the literature [tectonic phases, a concept often associated with the works of Hans Stille (1924) and further developed by Alexander Tollmann (1963, 1968), divided the geological record into short periods of deformation, or phases, separated by longer periods of general quiescence]. Although the terminology is now rarely used, it should be abandoned; our use of event applies to a period of time bounded by unconformities, and deformation within this period of time may be diachronous. The deformation is most closely dated by unconformities in the area of Vratsa and Lukovit, where the hiatus at the unconformity ranges from early Albian to Turonian (Tzankov et al., 1995; Figs. 7 and 8). The hiatus at the unconformity broadens to the east to include the Hauterivian to Cenomanian in the Tvarditsa area, the Hauterivian to Turonian in the area southwest of Etropole, and the Hauterivian to Santonian in the area of Gabrovo and Kalofer. At Tvarditsa there is a second very well dated late Santonian–early Campanian folding. The distribution of the unconformity at the base of the lower-upper Cretaceous rocks that places an upper limit on this deformational period is shown in Figure 7.

The region deformed during this time period is characterized by generally east-west–trending structures that extend through the foreland fold-thrust belt from the eastern part of the area through Elena to Vratsa in the west. Deformation also involved rocks farther south within the Stara Planina unit from Tvarditsa, and south of Etropole, through Svoge and into Serbia (Fig. 7). In the Trojan basin the structural vergence is generally to the north and Permian rocks are present in the cores of some of the folds, but no older rocks are exposed, suggesting that basement rocks are generally not involved in the deformation. Cleavage was formed in these rocks, particularly in the southern parts of the Nis and Trojan troughs. Farther south basement rocks were involved along the southern margin of the fold-thrust belt in tectonic elements south of Trojan in the areas of Tvarditsa, Kalofer, Etropole, south of Etropole, and Svoge. The younger of the two deformations present in the Tvarditsa area formed recumbent folds with overturned limbs of several kilometers. Areas northwest and west of Vratsa were not affected by deformation during this event (Fig. 7).

In the eastern part of the area north of Tvarditsa, the westernmost and earliest structures in the Kotel zone (Fig. 7), deep-water basinal facies of Triassic and Jurassic rocks, are unconformably overlapped by late Albian strata (Tchoumatchenko and Cernjavska, l989). They do not extend farther west than Tvarditsa. In some regional tectonic schemes these rocks are interpreted to be far traveled and represent the northern leading edge of thrust sheets now preserved in Strandja and derived from the south of the southern Meosian continental shelf edge (Gočev, 1986).

In northwesternmost Bulgaria, the Kraina unit was deformed during this first event. The lower Cretaceous Sinaja flysch, correlated with the Ceahlau and Severin units of the eastern and Southern Carpathians, respectively (see following), that makes up the Kraina subunit is strongly deformed and cleaved. The structures in northwest Bulgaria are poorly dated because they are overlain by Neogene sedimentary rocks, were overthrust by the Belogradchik anticlinorium in middle Eocene time, and have obviously been reworked by younger deformation (see following). Assigning the deformation of the Kraina unit during this first event is based on regional relations with the Carpathians in Romania (Sandulescu, l975; Fügenschuh and Schmid, 2005). If the Sinaja flysch is correlative with the units in the Southern and eastern Carpathians, it may be far traveled from the west and would be emplaced during this first event; however, newer interpretations may limit its displacement (see following).

In western Bulgaria and extending through eastern Serbia and eastern Macedonia are structures related to the Morava thrust belt (Zagorchev, 1996; Fig. 7). The Morava thrust belt is characterized by northeast-directed thrust faults that involve metamorphic rocks as well as Triassic to Kimmeridgian–Valanginian strata, the uppermost Jurassic and lower Cretaceous strata being characterized by turbidites (Treklyano Group, Fig. 7; Nachev and Nikolov, 1968; Zagorchev and Tikhomirova, 1986) representing a western deep-water assemblage in the western part of the Moesian shelf. The thrusts are overlapped by late Eocene strata, thus the timing based on stratigraphy is poorly known. Recent thermochronological data presented by Kounov et al. (2010) indicated that the thrusting occurred during late-Early Cretaceous time, between 139 and 112 Ma, a little older than the dated folding within the Forebalkan zone; however, the uncertainties are such that the deformation within the two areas could be mostly contemporaneous.

Farther south in northern Greece on the Chakadiki Peninsula are an imbricated assemblage of basement rocks and Triassic and Jurassic rocks with both shallow-water and deep-water sections that are mixed together by faults and assigned to the Circum-Rhodope unit (Fig. 7; Papanikolaou, 2009). There are mafic and ultramafic rocks within the faulted assemblage that were emplaced eastward in Late Jurassic to Early Cretaceous time. Similar rocks are present in small areas at the south end of the peninsula as well as in limited outcrops in eastern Greece and in the Strandja allochthon that are also assigned to the Circum-Rhodope unit (Fig. 7).

In southeastern Bulgaria the low-grade Triassic and Jurassic metasedimentary clastic sequences of the Strandja thrust sheets were emplaced to the north or northeast and are unconformably overlain by Cenomanian strata (Chatalov, 1990). These rocks represent a deep-water facies and were originally deposited south of the Moesian shelf on which they lie. The underlying basement is overlain by metamorphosed marble and metaclastic rocks assigned a Triassic to Middle Jurassic age. The Strandja allochthon is also assigned to the Circum-Rhodope unit; thus the unit appears to be around the southern margin of the Rhodope Mountains and continues in rocks along the southeastern margin of the Vardar zone. The thrusting of the Strandja rocks suggests that they may have formed by obduction of deep-water rocks from the margin of, or from within, oceanic areas that were around the southern extent of the Moesian shelf (Papanikolaou, 2009; Bonev and Stampfli, 2011). The thrust faults were emplaced on the Moesian shelf and may have extended as far north as the Kotel zone in the eastern part of the Forebalkan–Stara Planina area (Gočev, 1986; Figs. 6 and 7). Dabovski and Savov (1988) and Okay et al. (2001a) presented evidence that the deformation and metamorphism of the Strandja rocks occurred in Late Jurassic time (Sunal et al., 2011). The unconformably overlying strata are Cenomanian, and Okay et al. (2001a) assigned the deformation a Late Jurasssic–Early Cretaceous age. Georgiev et al. (2001) regarded the deformation as mid-Cretaceous age. It is not difficult to interpret the deformation within the foreland fold-thrust belt to be related to thrusting along the continent-ocean boundary between the Moesian shelf and the Vardar oceanic realm that formed the Morava and Circum-Rhodope tectonic units. The deformation within the fold-thrust belt of central Bulgaria was partly localized within the Nis-Trojan flysch basins, but also affected rocks beyond the margins of the basins. The deformation extended west into the area east of Vratsa, as shown by the unconformity there, but farther west there is no evidence for folding of this age, except in the Kraina and southern Carpathians areas. In the area between Tetevan and Vratsa the deep Trojan flysch basin grades into shelf deposits and the deformation extended beyond the western limit of the basin to the east of Vratsa. Farther south the deformation extended to east of Sofia into the northern part of the Sredna Gora unit (Fig. 7).

Although there is no evidence for the westward continuation of these folds and thrust faults in the western part of the Forebalkan zone, there was deformation within the Kraina unit in the northwestern part of Bulgaria during the mid-Cretaceous event. The Kraina unit contains Jurassic and Early Cretaceous flysch that is correlated with the Sinaja flysch that occurs in the Severin allochthon in the Southern Carpathians (Fig. 6). The Severin allochthon tectonically overlies the Danubian unit, which consists of a metamorphic basement and an overlying Triassic–Early Cretaceous shallow-water section that is interpreted to be the western continuation of the Moesian continental crust of the Balkan foreland. Both the Danubian and Severin units are overlain by the Getic-Supragetic thrust faults, that consist mainly of a metamorphic basement overlain by a shallow-water Mesozoic section (Fig. 6). The Getic-Supragetic unit forms a huge half-window in the southern part of the Southern Carpathians and probably tectonically overlapped both the Danubian and Severin units, but was modified by normal faults in its eastern part in mid-Cenozoic time (see Fügenschuh and Schmid, 2005). The thrusting is from west to east and occurred in two events: the first was mid-Cretaceous thrusting of the Getic-Supragetic above the Severin oceanic unit, sealed by Albian–Cenomanian strata, followed by a second west to east thrusting of the Getic-Supragetic-Severin complex above the Danubian unit, sealed by Maastrichtian strata, forming the large half-window in the Southern Carpathians (Codarcea, 1940; Burchfiel, 1976; Sandulescu, 1984, 1994). We consider the first event to be the northwestern continuation of the mid-Cretaceous event within northwestern Bulgaria, but the structures are greatly disrupted by younger structures and traces of the oldest structures are difficult to correlate in detail (see following). A third structural event took place in the late Paleocene and later, and is related to the third deformation in the Forebalkan zone of Bulgaria (considered in more detail in the following).

Between the end of the Early Cretaceous deformational period and the next younger deformational period of latest Cretaceous–earliest Cenozoic age, broad areas of northwest and north-central Bulgaria were covered by both shallow- and deep-water strata. However, sedimentary rocks of this age are not preserved above the folded rocks in the central part of the Nis and Trojan troughs, e.g., in regions around Teteven and Trojan (Figs. 7 and 9) where Permian rocks are exposed in the cores of folds, thus the age of the structures in this area must be inferred by projecting the structures into this area from surrounding areas where the younger rocks are preserved and can be dated.

West of the folds east of Vratsa, deposition of shallow-marine calcareous rocks with local terrigenous interbeds took place paraconformably above lower Cretaceous rocks. Farther west subsidence was more rapid with the deposition of the Kula flysch in northernmost Bulgaria and extending to eastern Serbia and southern Romania (Figs. 7 and 9).

South of most of the foreland fold-thrust belt, during Late Cretaceous time, thick sections of volcanic and volcaniclastic rocks and flysch were deposited within an intraarc or backarc extensional setting (Fig. 9; Boccaletti et al., 1974; von Quadt et al., 2005; Burg, 2012), where they locally overlie Mesozoic sedimentary rocks and older basement rocks deformed in the earlier deformational event. The main axis of igneous activity was south of the Stara Planina, mostly within the Sredna Gora unit, and formed a magmatic arc developed above a north-dipping subduction zone (Boccaletti et al., 1974) located in northern Greece (present-day position) and continuing northward into the Vardar zone of Macedonia and Serbia. Georgiev et al. (2012) studied these rocks in the eastern part of the Balkan orogen; they showed convincingly their intraarc and backarc position and demonstrated that they were part of a magmatic belt formed between 92 and 78 Ma. The northern margin of the backarc-intraarc basin is not well defined because transitions between volcaniclastic rocks and the shallow-water strata farther north are not preserved. In the easternmost part of the area, beginning near Gabrovo and extending to the Black Sea, is a thick sequence of largely nonvolcanic terrigenous sediments, much of it flysch (referred to as the Luda Kamchia trough) ranging in age from Late Cretaceous to middle Eocene. Near the Black Sea there are some sediments of Eocene–Oligocene age, but they are overlain structurally by strata of the Luda Kamchia trough. The upper limit of the volcanic succession is pre–late Campanian. These strata become more volcanic rich to the south and probably represent a more northern part of the backarc–intraarc basin; they are separated by younger thrust faults and were deposited a considerable distance from contemporaneous Late Cretaceous shallow-water sediments of the Moesian Platform to the north (Fig. 9). Their thickness and lack of transition to the shallow-water nonvolcanic rocks farther north suggests that the backarc-intraarc basin had a sharply defined northern border formed by extension superimposed on the region affected by the earlier shortening deformation. It has been suggested, but is difficult to prove, that the northern margin of the backarc-intraarc basin may have been locally or regionally exposed dry land (Nachev, 1993). The extensional setting for these volcanic rocks forms a major break between the first and second convergent events within the fold-thrust belt. It also has been suggested that the extension is related to the extensional opening of the Black Sea and to the formation of the West Black Sea fault, a right-slip fault that is along the eastern continental margin of Bulgaria (Okay et al., 1994). East of this fault the Istanbul zone moved southward to be emplaced in the western part of the northern Pontides by early Eocene time (Okay et al., 1994; Fig. 9). The easternmost part of the Srednogorie zone to the south has a very thick volcanic succession, as thick as 4 km, with abundant pillow lavas of high potassium composition, suggesting an initial backarc rifting (Boccaletti et al., 1978; Georgiev et al., 2001) that did not develop oceanic crust.

The second major period of deformation to affect northwest and north-central Bulgaria took place in latest Cretaceous and early Paleogene time (Fig. 10). The broadest range for this deformation is Maastrichtian to Thanetian time and the evaluation of the hiatus that dates this period of deformation is shown schematically (Figs. 8 and 10). It is most narrowly dated near Gabrovo, where the hiatus between deformed and undeformed rocks is within the upper Paleocene, and a second deformation occurs a little later very close to the Paleocene-Eocene boundary. The hiatus broadens to the west; near Lukovit and Vratsa it ranges from late Paleocene to early Eocene and from early Paleocene to early Eocene, respectively. The hiatus also broadens to the south where it ranges from early Paleocene to early Eocene near Kalofer and southwest of Etropole and to latest Cretaceous to early Eocene near Tvarditsa (Figs. 8 and 10).

Throughout the region immediately in the southwestern part and south of the foreland fold-thrust belt, there is a general lack of Paleogene strata (Fig. 11) and deformed upper Cretaceous rocks are overlain by Neogene rocks of late Miocene or even Quaternary age. Thus it is impossible to assign the deformation of the upper Cretaceous rocks to a narrowly defined event, although based on more regional data, folding and thrusting of this age was widespread (Fig. 10) and the distribution of structures of latest Cretaceous to late Paleocene or early Eocene age overprints structures of the early-Late Cretaceous deformation. The structures consist of east-west–trending folds and thrust faults, generally north vergent, from Varna in the east to Vratsa in the west. It probably continues farther west into the Southern Carpathians (see following), but the lack of Paleogene deposits in northwest Bulgaria does not permit the precise dating of structures other than pre–late Neogene. South of the fold-thrust belt evidence for deformation of this age is present near Kalofer and Tvarditsa, where there are Paleogene rocks and structures, but the lack of extensive Paleogene deposits makes it difficult to prove how widespread they are. The deformation of this time period was superimposed on the older structures that were developed in pre-Turonian rocks of the eastern and central foreland fold-thrust belt and regions to the south of it. In the central part of the foreland fold-thrust belt from Gabrovo to west of Teteven, there are no rocks younger than Early Cretaceous, and the relative effects of the two deformational events are very difficult to separate.

West and northwest of Vratsa the age of the structures is not clear. Paleogene rocks that would differentiate structures formed in this second event or younger time periods are largely missing. Neogene sedimentary rocks in the area west of Vratsa unconformably overlie deformed Jurassic or locally lower Cretaceous rocks, but it is impossible to separate events from the mid-Cretaceous to Paleogene. At three localities east and north of Rabisha (Fig. 11; see Filipov and Cheshitiv, 1992) Paleogene (middle Eocene) rocks overlie folded upper Cretaceous Carpathian-type strata dating the deformation as Late Cretaceous to middle Eocene. The middle Eocene is also folded, showing the effects of the third period of deformation. By tracing structures from east of Vratsa where they can be dated westward, it is possible to interpret much of the structure in northwestern Bulgaria to have formed during the middle Eocene, or perhaps in the earlier mid-Cretaceous. Early Paleogene and locally latest Cretaceous structures are present in the Southern Carpathians to the northwest and support the northwestward continuation of the fold-thrust belt in northwest Bulgaria (Codarcea, 1940; Burchfiel, 1976; Sandulescu, 1984, 1994).

Where the two unconformities that date the mid-Cretaceous and Late Cretaceous–early Paleogene events (Figs. 7 and 10) are close together, such as east of Vratsa, north of Teteven, near Gabrovo and north of Tvarditsa, the angular relations between beds at the unconformities are highly variable. East of Vratsa and north of Teteven the angular relations below both unconformities are moderate, a maximum of ∼30°–40°, suggesting the two events may be of about the same intensity; however, such relations are not very substantive. These areas are near the margins of the area deformed in mid-Cretaceous time. At Gabrovo rocks below the mid-Cretaceous unconformity are highly deformed, isoclinally folded, and contain a well-developed cleavage, whereas the Paleocene rocks above the unconformity are much less deformed and lack cleavage. Similar relations are present north of Tvarditsa and north of Kalofer. These relations yield only a preliminary picture of the relative importance of these two deformations, and much work needs to be done to unravel the relative strengths of the two deformations throughout the area.

On the regional scale, deformation occurring in the latest Cretaceous to early Paleogene time period probably is related to the events leading to the final closure of the Vardar ocean along the western and southern margins of the Moesian continental shelf (Burchfiel, l980; Roberson and Dixon, l984; Dercourt et al., l986, Şengör and Yilmaz, l981). Postcollisional convergence lasted into the middle Eocene, but the closure of the Vardar ocean is interpreted to be by northward subduction, in contrast to mid-Cretaceous subduction, which was southward (see following). Northward polarity of subduction best explains the geological relations at the time of the second period of deformation, although not all agree (for a discussion, see Burg, 2012).

To the east, the final displacement on the west Black Sea fault ended in early Eocene time with the emplacement of the Istanbul zone into the western Pontides (Okay et al., 1994; Fig. 9).

Following early Paleogene deformation, deposition of terrigenous strata with incursions of shallow-marine and brackish sediments took place during latest Paleocene to (locally) early-middle Eocene time (Zagorchev et al., 1989; Goranov et al., 1992; Figs. 8 and 11). These strata are preserved in only a few places, making it difficult to prove the age and extent of either the preceding early Paleogene structures or the succeeding middle Eocene and younger structures. Strata of this age are present within the fold-thrust belt southeast of Vratsa, south of Lukovit, near Veliko Tarnovo, east of Gabrovo, and in three small exposures near Rabisha (Fig. 11). Along the southern margin of the fold-thrust belt they are present north and west of Kalofer.

In the eastern part of the belt along the Black Sea coast and in the offshore strata are numerous unconformities within the early Paleogene strata, but there is no obvious break that separates the second and third periods of deformation (Stuart et al., 2011). However, the deformation in this area extends into the Oligocene, a time when the dynamic setting in the orogen changed between the second and third periods of deformation (see following).

Shortening deformation from middle Eocene to middle Miocene time is the third and final convergent deformation in the foreland fold-thrust belt (Fig. 12). Structures formed during that time period are commonly assigned to middle Eocene, Oligocene, or early Miocene deformations. However, dating of these structures is not well constrained in many places because there is a widespread hiatus in the stratigraphic units from latest-middle Eocene to middle (locally early) Miocene time (Figs. 8, 11, and 13), and middle Miocene sedimentary rocks are the oldest rocks that unconformably overlie the structures in northwest Bulgaria (Figs. 12 and 13); however, in easternmost Bulgaria and trending into the Black Sea to the east there are more complete sections that document deformation that extends through that time period (Stuart et al., 2011). Because the dating of the third period of deformation is poorly constrained, it may have been a protracted or diachronous event, or may have consisted of more than one period of deformation, but at present that cannot be resolved. Nevertheless, the deformation occurred in a tectonic setting different from the early two deformational events (see following).

From about late middle or late Eocene time to the present most of area south of the fold-thrust belt in Bulgaria was dominated by extensional tectonism (see following; Marchev et al., 2004; Bonev and Beccaletto, 2007; Burchfiel et al., 2008). Some evidence indicates that extension began during or between convergent deformational events in Paleogene time, but widespread extension became dominant in about early-late Eocene time, marking the end of major regional shortening within the Balkan orogen and most of Bulgaria; however, shortening within the fold-thrust belt continued and is related to a different tectonic setting from the earlier two events (see following). In a few places deformation can be constrained to be within the Eocene, but in other places deformation is younger; we discuss each deformation herein.

A distinct break between the second and third shortening events within the foreland fold-thrust belt is best constrained within the central part of the belt, poorly documented in the western part of the belt, and does not obviously appear to be present, as documented by Stuart et al. (2011), in the eastern part of the belt, where it trends into the Black Sea. Within the central part of the fold-thrust belt early and middle Eocene and younger sedimentary rocks that overlie structures of the second deformational event are everywhere deformed, proving a third deformational event, but the upper limit for the third deformational event or events is mostly provided by unconformably overlying middle and upper Miocene, and rarely lower Miocene, strata. Only in three areas can a strong, but circumstantial, case be made for middle Eocene deformation in the fold-thrust belt (Fig. 12). (1) South and west of Tvarditsa, deformed late middle and upper Eocene sedimentary rocks overlie structures that are dated as probable middle Eocene by projection from surrounding areas. (2) Likewise the thrust of crystalline rocks north and west of Kalofer overrides middle Eocene rocks and is considered to be older than the late-middle Eocene rocks that are weakly folded near Tvarditsa (Bonchev, 1978). However, constraints on the timing of these structures are poor. (3) Major north-vergent thrusts and folds are present south and east of Gabrovo that extend eastward into the Black Sea. They carry rocks as young as early-middle Eocene in their hanging walls, but concise upper limits on the deformation cannot be determined. Near the Black Sea coast these structures are thrust over Oligocene rocks (Dachev et al., 1988), suggesting that at least the easternmost part of the belt was active during the Oligocene. Much of the structure south and east of Gabrovo can interpreted as being part of this period of deformation (Fig. 12).

Recently published seismic and drilling data from the easternmost part of the fold-thrust belt where it extends into the Black Sea has documented thrusts and associated folds that involve early-middle Eocene rocks and are overlapped unconformably by late-middle Eocene rocks, constraining the age of deformation there as middle Eocene (Stuart et al., 2011; Fig. 14). The data from this area do not show a clear break between what we interpret to be the second and third periods of deformation, and the data of Stuart et al. (2011) show that deformation continued into the Oligocene in the Black Sea area. In the western Black Sea, the structures curve to the south and are beneath the western Black Sea continental margin, where they continue southeast and may continue into the northern margin of the Pontide belt of northern Turkey (Fig. 12). The relations of the structures to the west Black Sea fault and the Istanbul zone of the northern Pontides suggests that they are younger than the emplacement of Istanbul zone allochthon, interpreted to have been emplaced by the early Eocene (Okay et al., l994), and their continuation should be north of the allochthon along the southwestern Black Sea margin.

The structural style within the fold-thrust belt in northwest Bulgaria is different from the central and eastern areas (Fig. 14). It contains extensive involvement of pre-Mesozoic rocks, in contrast to the central area, where there is only limited involvement of the pre-Mesozoic rocks along its southern part and the eastern area, where only younger Mesozoic and Cenozoic rocks are exposed within the structures at the surface. This change in structural style may be related to the tectonic setting of the fold-thrust belt during late-middle Eocene to early Miocene time and is discussed in more detail in the following.

A more general argument can be proposed to suggest that middle Eocene deformation was widespread and the beginning of last major shortening event to affect the Balkan fold-thrust belt. Most Paleogene sedimentary sequences end with lower-middle Eocene strata; upper Eocene rocks, although rare, begin a sedimentary rock sequence that ranges from late Eocene to early Miocene and has a different environment of deposition than that of early-middle Eocene strata (Fig. 13). These younger rocks were deposited mainly in continental and brackish environments in their basal parts, locally contain coal, and in southern Bulgaria they contain abundant volcanic rocks. They are present mainly in southern Bulgaria and northern Greece, where they were deposited in grabens or half-grabens. Particularly in southern Bulgaria there is abundant evidence they were deposited during active extension (Bonev and Beccaletto, 2007; Burchfiel et al., 2008). These regional relations have been interpreted to suggest that the major shortening deformation may have begun in middle Eocene time but extended into Miocene time (see following).

Based on regional arguments structures in northwest and north-central Bulgaria are considered to belong to the middle Eocene deformation and may have extended into Oligocene time (Fig. 12). Lower Miocene rocks unconformably overlie deformed lower and middle Eocene rocks south of Lukovit, and the structures here are regarded as middle Eocene in age. Similar relations are present near Vratsa, but the two sequences are not in direct contact. The second folding of the upper Cretaceous Kula rocks in northwest Bulgaria, and the folding of Eocene rocks at Veliko Tarnovo, are also considered to have been deformed in middle Eocene time (Fig. 12), although in these latter areas they can only be dated as pre-Neogene. In northwest Bulgaria the large Belogradchik, Berkovica, and Svoge anticlinoria also are considered to have formed during middle Eocene time (Bonchev, 1971); however, in all these areas there is no evidence for the middle Eocene age of deformation, and only an upper limit of locally early Miocene or more broadly middle Miocene can be confirmed.

The intensity of middle Eocene deformation increases toward the south across the fold-thrust belt. Structures near Vratsa, Lukovit, and Veliko Tarnovo consist of generally east-west–trending open folds, whereas at and east of Gabrovo the rocks are more tightly folded and were involved in north-vergent thrusting. Northwest of Kalofer (Fig. 12) at least 15 km of northward thrusting of crystalline rocks took place; the basal part of the thrust sheet contains well-developed mylonites and the footwall rocks contain a well-developed cleavage. The three major anticlinoria in northwest Bulgaria are north vergent and their northern flanks are marked by thrusts carrying pre-Mesozoic basement rocks and locally Mesozoic rocks in their hanging walls and Mesozoic rocks and only locally Cenozoic rocks in their footwalls (Figs. 12 and 14). Rocks to the north and below the frontal thrusts are strongly folded and commonly overturned. The anticlinoria consist of smaller folds and associated thrust faults. The Belogradchik anticlinorium (BA, Fig. 12) consists of two large north-vergent folds, thrust along their northern flank, that plunge eastward beneath folded Mesozoic rocks unconformably overlapped by mostly unfolded Neogene sedimentary rocks; therefore the upper limit of their age can be assigned to only early or middle Miocene time. The Berkovica anticlinorium (BeA, Fig. 12) passes through a gentle axial depression south of Vratsa, where its Mesozoic cover is exposed, then reverses axial plunge and continues to the southeast to Etropole (Fig. 12). The Svoge anticlinorium (SoA, Fig. 12) has a gentle westward plunge west of Svoge and trends east from Svoge where it has been disrupted by Neogene normal faults. Because it carries upper Cretaceous volcanic rocks in both its hanging wall (west of Sofia) and footwall (near Etropole), and because its basement and pre–Late Cretaceous Mesozoic cover is part of the Moesian shallow shelf sequence, it is considered to be related to the two other anticlinoria to the north. These anticlinoria are considered to have formed in middle Eocene time, and probably are in the hanging walls of major southwest-dipping thrust faults; however, the timing of these structures is unknown with certainty. The structures are overlapped by middle Miocene rocks (Badenian) and in only two places do they involve Eocene rocks; in the footwall of a thrust at the north flank of the Belogradchik anticlinorium upper Eocene rocks are present northeast of Belogradchik (BA, Fig. 12), and ∼20 km southeast of Vratsa the eastward continuation of the Berkovica anticlinorium is structurally above lower-middle Eocene rocks. On the basis of these two localities the structures are assigned a post-Eocene and pre-middle Miocene age (Bonchev, 1971).

A major north-vergent thrust system that overrides strongly folded and overturned rocks forms a continuous structural zone along the southern margin of the fold-thrust belt and passes north of Kalofer (Fig. 12). Rocks in the hanging wall of this thrust zone carry pre-Mesozoic metamorphic basement rocks and in its footwall contain both Mesozoic sedimentary cover and their pre-Mesozoic metamorphic basement. In the footwall of this structure are not only Mesozoic strata, but also Paleogene rocks of middle Eocene age. This is the so-called Shipka overthrust or Stara Planina granite thrust. In the area between Tvarditsa and Gabrovo this structure was formed in early-Late Cretaceous time; it is overlapped unconformably by late-Late Cretaceous to early Eocene sedimentary rocks (Kanchev, 1962; Kanchev et al., 1995). The structure involves Cenomanian and Turonian strata thrusted few kilometers to the north. The thrust is covered by late Senonian sediments. Kanchev (1962) assumed that this was a local event affecting only the northern limbs of the Shipka (western part) and Tvarditsa (eastern part) anticlines. In a general way this area is between Gabrovo and Tvarditsa. All thrusts are above Paleogene rocks, so the time of last deformation is the same. Toward the west this structural zone was last deformed at a later time, probably late Eocene, because lower Eocene rocks are present below the north-vergent thrust, marking the continuation of this zone southeast of Vratsa. This thrust system therefore had a multiphase development.

In the area where this structural zone is south of most of the folds of the fold-thrust belt, south of Trojan where the foreland fold-thrust belt can be dated as mid-Cretaceous, diverse rocks, including the Late Cretaceous to early Eocene overlap assemblage mentioned here, the structures are overridden by the generally horizontal Shipka thrust north of Kalofer (SH in Fig. 12), where basement rocks make up most of the hanging wall. The thrust relationship is clear where it overrides Mesozoic and lower Eocene rocks. Toward the south the thrust is folded with a moderate (∼30°) south dip and it is displaced by normal faults that bound the north side of the late Cenozoic Sub-Balkan graben system (Tzankov et al., 1996; Fig. 15). Within the region south, west, and east of the Shipka allochthon are other thrust faults that carry only crystalline rocks in their hanging walls and locally have Mesozoic rocks in their footwalls. These thrust faults are interpreted as forming part of a major allochthon covering a large area to the south into the Sredna Gora unit (Kockel, 1927; Vulchanov, 1971), but it has been disrupted and covered by the strata of the Sub-Balkan graben system. The age of emplacement of this allochthon is interpreted to be post-middle Eocene, using the regional arguments presented here; however, its upper limit of deformation is poorly constrained.

Within the area of northwest and north-central Bulgaria, deformation that occurred from the end of the late Eocene to early or middle Miocene time is very difficult to prove except locally. Such structures would involve rocks deposited in late Eocene or Oligocene time; sedimentary rocks of this age are very rare and crop out only near Tvarditsa and southwest of Sofia (Fig. 13). Several other areas where shortening occurred during this time period occur. (1) Along the northernmost thrust carrying rocks of the Luda Kamchia sedimentary sequence east of Tvarditsa, upper Eocene rocks are present in the footwall (Fig. 13). (2) In the Bobov dol-Persink and Pernik basins in western Bulgaria, Oligocene rocks unconformably overlie weakly folded Eocene rocks and deformation is dated as early Oligocene. These rocks are folded and cut by thrust faults of Oligocene and/or early Miocene age (Zagorchev, 1998). However, the structural relations here were reinterpreted by Kounov et al. (2011) to suggest that if thrusting had occurred it was the result of restraining bends in strike faults; this suggestion is in support of the hypothesis we present here for the dynamic setting of this third event of deformation, i.e., that the thrust overlies proven Oligocene–latest-early Miocene strata, thus the deformation is early Miocene or latest Oligocene. (3) Rocks of Oligocene age in the Tikvesh-Ovchepole basin of Macedonia were folded in latest Oligocene to early Miocene time (Dumurdzanov et al., 2005). (4) In the Thrace Basin of northwest Turkey folds are dated as late Oligocene to earliest Miocene (Perincek, 1991). The latter two areas are outside the fold-thrust belt of Bulgaria and their folding may be mainly unrelated to deformation of the fold-thrust belt. Although the northern margin of the Luda Kamchia basin was deformed and thrusted in post-middle Eocene time (see preceding discussion concerning subsurface information along the Black Sea coast in Fig. 14), more southern parts of the basin continued to receive sediments into late Eocene and even Oligocene time, and they are deformed. The timing of this thrust is not well constrained.

In the eastern continuation of the foreland, more Cenozoic strata are present as the fold-thrust belt plunges to the east into the Black Sea. Subsurface data show numerous unconformable relations within the middle Cenozoic strata, and while there is a major middle and late Eocene shortening with an unconformity at the base of the youngest upper Eocene strata, there is no obvious break between deformation of the second and third deformational events. However, important shortening took place in late Eocene time and deformation continued throughout the Oligocene (Stuart et al., 2011, Fig. 14). Middle and upper Eocene rocks are folded with Oligocene rocks, but to a lesser degree, and this late folding and thrusting is late Oligocene (Dachev et al., l988; Stuart et al., 2011; Fig. 14) and belongs to the third deformational event in the fold-thrust belt. Although the younger two events cannot be separated in eastern Bulgaria, the termination of the deformation is marked by extensional faulting in the unconformably overlying middle Miocene strata.

These deformational structures we consider to be related to a different dynamic system from the second event; however, the transition from the dynamic systems from the second to the third events remains unclear in the eastern part of the Bulgarian fold-thrust belt (see following).

From the interpreted distribution of middle-late Eocene to middle Miocene structures it is likely that much of the foreland fold-thrust belt was deformed in this time period. This is the last major deformation to affect northwestern Bulgaria and was superposed on early Paleogene structures; and, even though it is it poorly dated, it is clearly a younger event than the second deformational event of early Paleogene age. It is very likely that the region between Vratsa and Elena affected by a strong middle Cretaceous deformation and possibly weaker Paleogene deformation was deformed for the third time during middle Eocene to middle Miocene time. All three unconformity-bounded sequences of sedimentary rocks that show the effects of all three superposed deformations are present in the area near Vratsa and Lukovit.

During middle Miocene to Holocene time, southern Bulgaria was characterized by extensional deformation, whereas northern Bulgaria was characterized by a broad area of deposition with slow subsidence except in the northwest within the Moesian Platform, where more rapid subsidence occurred in the Lom depression (Fig. 15) and activity in the fold-thrust belt ceased. The present topography of Bulgaria developed during middle Miocene to Holocene time. North and south Bulgaria are separated by the east-west–trending Stara Planina Mountains through north-central Bulgaria (Fig. 1). The Stara Planina is bounded on its south side by the Sub-Balkan graben system formed by a system of generally east-west–trending, southwest-dipping normal faults with the exception of the Sofia graben, which has major normal faults on both sides (Fig. 15; Tzankov et al., 1996).

The middle Miocene (Badenian) and younger rocks of northern Bulgaria were deposited on the Moesian Platform and extend south into the northwestern and external part of the fold-thrust belt. They formed in a foredeep position, but are younger than the last major deformation within the fold-thrust belt, although younger shortening deformation extends into the Southern Carpathians (Codarcea, 1940; Sandulescu, 1984, 1994). Some normal faults with small displacement cut the late Cenozoic rocks. One of us (Nakov, 2009) suggested that some shortening deformation occurred in post-Chersonian time (between 10 and 8.9 Ma), but if so, it was very weak and unimportant in causing the distribution of late Cenozoic rocks. This weak shortening was established northeast of Vratsa, but its extent and role further to the west and east is unknown.

During the middle Miocene to Holocene, extensional tectonism was widespread throughout much of southern Bulgaria and its widespread effects reached the southern parts of the fold-thrust belt in late Miocene to Pliocene time (Burchfiel et al., 2000, 2008). The grabens now flanking the southern Stara Planina have mainly south-dipping normal faults on their north side and considerable down-to-the-north rotation of hanging-wall blocks (Fig. 15). The rotation on the graben fill is syntectonic and indicates that the faults are listric, the dip shallowing at depth to the south (Tzankov et al., 1996). An interpretation by Roy et al. (1996) indicates that displacement on the south-dipping normal faults unloaded the crust in their footwalls, causing it to rise. Such unloading forms the highest topography, with a steep south-sloping gradient in the footwall near the faults and a gentle northward slope north of the topographic high explaining the present-day topography. The Sofia graben, which has a major fault along its south side, is to the west of the Sub-Balkan graben system and also forms the south flank of the Stara Planina. Thus the present topography is the result of the two different types of Neogene sedimentary basins in northern Bulgaria; extensional grabens in the south and a broad northward-deepening basin in the north separated by a topographically asymmetric mountain range formed by footwall uplift. The faults of the Sub-Balkan graben system are active, and so the elevation of the Stara Planina probably continues today, indicating that the Stara Planina is one of the youngest topographic features in Bulgaria.

The normal faults of the Sub-Balkan graben system displace the south-dipping thrust faults of the foreland fold-thrust belt (Figs. 14 and 15) so that their southern continuations are not exposed, but are covered beneath the Sredna Gora lowlands. How far south the thrust faults extend remains unknown, but because pre-Mesozoic rocks are present within their hanging walls from central Bulgaria to the west, it indicates that Mesozoic strata in their footwalls must extend an unknown, but considerable, distance south of the Stara Planina.

Foreland fold-thrust belts in most mountain ranges often show deformation that progresses from the inner to outer part of the orogen, and are intimately related to the tectonic evolution of the entire orogen. In contrast, the foreland belt in the Balkan orogen was formed during three superposed shortening events that do not show the expected progressive development; the question of why this happened is an orogen-wide problem of the dynamics of the Balkan orogen.

The preceding discussion of the foreland fold-thrust belt is almost entirely from surface geology; few seismic lines are available, making an analysis of the geometry at depth and the magnitude of shortening structures uncertain. The vergence of the fold-thrust belt is generally northward for the Late Cretaceous and/or early Paleogene and middle Eocene to Miocene deformations as determined by the sense of overturning of folds and dips on thrust faults; for the mid-Cretaceous event it is less certain, but is most likely also north vergent. How all the structures of different ages interact at depth is unknown, but a simplistic analysis indicates that the Moesian Platform currently dips gently south to be below the Stara Planina unit; how much farther south it continues is unknown. That basement rocks are involved in the thrusting within the Stara Planina and in parts of the Sredna Gora units and some of youngest strata are present in the footwalls of the thrust faults suggest that the Mesozoic rocks continue farther south for at least tens of kilometers (Fig. 14). Tectonic overlaps of thrust hanging walls in the Gabrovo and Shipka areas are ∼15–20 km where the thrust faults are nearly subhorizontal and give a minimum displacement (central section, Fig. 14). Tens of kilometers of displacement on the foreland belt would suggest that at a minimum, much of the crust in the Sredna Gora unit is also allochthonous. Both of these thrust faults were emplaced during the youngest event and both thrusts are displaced by late Cenozoic normal faults.

The mid-Cretaceous folding within the foreland belt is within the Nis and Trojan flysch basins that formed prior to the shortening. The flysch basins shown in Figure 6 are in present-day positions, and they certainly moved northward relative to the foreland during the two younger deformational events. Their narrow geometry and thick flysch strata that grade to both flanks and along strike into shallow-water strata suggest that they were rifted basins inverted during the mid-Cretaceous deformation, but if they involved basement rocks in their formation, the locations of the rift basins were in the southern part of the foreland belt and their positions below the basal thrust of the foreland fold-thrust belt would now be even farther south. What caused rifting of this age is not clear, but the positions of the basins relative to exposed Mesozoic strata adjacent to the basins and late Paleozoic rocks within the folded basin strata indicate that they were within the Moesian shelf area. This is different from the flysch of the Severin unit in the Southern Carpathians of similar age to the northwest, regarded to be underlain by oceanic or very thin continental crust (Fig. 7). How these flysch basins relate to the oceanic area of the same age in the Carpathians is unclear and subject to considerable interpretation.

In their study of the Southern Carpathians, Fügenschuh and Schmid (2005) suggested in their figure 9E (but did not discuss in detail in their text), that the flysch of the Southern Carpathians (Severin flysch) formed two belts on either side of Moesian basement rocks that formed the basement for the Forebalkan fold-thrust belt in Bulgaria. The basement rocks would continue northward as the Danubian tectonic unit of the Southern Carpathians and end farther north, thus forming a peninsula. Fügenschuh and Schmid (2005, fig. 9E therein) suggested that the flysch units overlying oceanic crust pinch out southward and do not continue very far into Bulgaria. This interpretation is important because it suggests that the Moesian basement east of where the flysch pinches out is part of basement rocks for the entire shelf area of Bulgaria. Such an interpretation has merit, in that it could be suggested that the Nis-Trojan basins are the easternmost extent of the flysch basins, but were rifts extending east from the oceanic area of the Southern Carpathians into the Moesian continental crust? Thus the Balkan crust in northern and central Bulgaria was entirely continental crust (Fig. 4), and there is no good evidence that oceanic crust extended into Bulgaria. Such an interpretation needs further investigation because these connections between Bulgaria and Romania have been strongly disrupted by younger shortening and strike-slip deformation.

In Bulgaria the mid-Cretaceous deformational event caused shortening of the flysch basins and affected rocks beyond the basin margins. Major northeast-vergent shortening of similar age occurred along the eastern margin of the Vardar zone on the Morava thrust (Zagorchev and Tikhomirova, 1986; Zagorchev, 1986, 1996; Kounov et al., 2004, 2010; Schmid et al., 2008) and with the emplacement of the Strandja allochthon by northward thrusting in eastern Bulgaria from the eastern continuation of Vardar zone (Fig. 7; e.g., Papanikolaou, 2009). In the footwall of the Strandja thrust are metamorphic rocks of the Rhodope unit (sensu lato) that are probably the southernmost exposures of the Moesian basement, but the Mesozoic cover rocks are poorly known and dated and were metamorphosed and deformed in Cretaceous time (e.g., Bonev and Boccaletto, 2007; Bonev et al., 2006). In the footwall of the Morava thrust are Jurassic to lower Cretaceous strata of the Trekljano Group that unconformably overlie upper Triassic rocks of the Moesian shelf (Zagorchev and Tikhomirova, 1986). The basal unit of the group consists of polymictic sandstone and conglomerate of Early Jurassic age that grade upward into a sequence of back slates, siltstones, and radiolarites that range in age from Middle to Late Jurassic (Kimmeridgian); flysch deposits of middle Tithonian to Berriasian age unconformably overlie these. Both sections are in the footwall of the Morava thrust. These rocks are interpreted to be a deep-water succession transitional from the shallow-water strata to the north and east to the oceanic strata of the Vardar zone to the west (Zagorchev and Tikhomirova, 1986; Zagorchev, 1986). The position of the Morava thrust is thus interpreted to have originated from the western edge of the Moesian continental crust.

While the timing of the shortening within the foreland belt is reasonably well known on the basis of stratigraphy (from ca. 110 to 90 Ma; Fig. 8), the events within the Rhodope are less well known and their contemporaneity with shortening in the fold-thrust belt needs to be better constrained. Burg (2012) presented a detailed discussion of deformation within the Rhodope region during that time, and pointed out that the geology of the Rhodope is complex and remains controversial. Burg (2012) interpreted the Rhodope as a collage of thrusted units that are mainly south vergent, units he suggested are related to northward subduction of oceanic arc and thin continental crustal rocks. However, the Circum-Rhodope rocks contain mafic and ultamafic rocks that have been interpreted as remnants of oceanic crust (Papanikolaou, 2009). Following Papanikolaou (2009), we interpret the emplacement of the Circum-Rhodope allochthon to be the result of a mid-Cretaceous period of south- and southwest-vergent subduction within an oceanic setting (see also Bonev and Stampfli, 2011), a tectonic setting that remains controversial. Meinhold and Kostopoulos (2013) indicated that the Circum-Rhodope belt may be a complex of different units from different tectonic positions. The tectonic setting is further complicated by the interpretation of Naydenov et al. (2013) that the Maritsa fault zone (MFZ; Fig. 5) was a major right-slip shear zone during Late Jurassic–Early Cretaceous time; they suggested that the shortening within the southern Balkans was strongly obliquely convergent with north-vergent thrusting within the Strandja domain separated by right shear along the Maritsa shear zone from south-vergent thrust in the Rhodope domain to the south. The nature of the Jurassic to early-Late Cretaceous tectonism still remains poorly understood (see the discussion in Burg, 2012), but we interpret the complex Circum-Rhodope unit with its deep-water strata and local ophiolitic rocks to have been thrust onto the southern extent of the continental rocks of southern Bulgaria. It is somewhat irrelevant whether the rocks in the Kotel zone are the leading edge of the Circum-Rhodope unit, as they contain north-vergent structures of mid-Cretaceous age (Georgiev et al., 2001; Tchoumatchenco et al., 2004). We suggest that the position far to the north of the present position of the Strandza unit is because the Late Cretaceous and subsequent extension separated the Kotel zone from its original position.

In such a setting, the shortening within the Nis-Trojan basins is explained as a result of orogen-wide deformation related to activity along the southern margin of the Moesian shelf where it passed southwest or south beneath rocks along the eastern and northern margin of the Vardar zone. The fold-thrust belt would thus be synthetic to the thrusting and related to westward and southward subduction in an oceanic domain either within or north of the Vardar zone. The geometry of structures within the present-day Rhodope crust (as shown in, e.g., Fig. 7) has been greatly disrupted by younger Mesozoic deformation that is dominantly south-vergent thrusting and later superposed Cenozoic extension (Burg, 2012).

Between the mid-Cretaceous and latest Cretaceous–early Paleogene shortening events within the foreland belt is when extensional rifting occurred within the Sredna Gora area with the formation of a backarc or intraarc basin that probably extended eastward through the Black Sea area (Boccaletti et al., 1974; Burchfiel, 1980; von Quadt et al., 2005; Fig. 9). This and the abundance of magmatism of Late Cretaceous age are interpreted to be related to rollback along a north-dipping subduction zone beneath the Rhodope. This rift formed within the Moesian continental crust, nonvolcanic Mesozoic rocks of which are present within the foreland fold-thrust belt to the north and within the northern part of the Rhodope area to the south. The width of the volcanic basin is unknown, but transitions of strata to the north are rare and the lack of volcanic rocks to the north suggests that it had a sharp northern boundary and/or was separated from the foreland fold-thrust belt rocks by some significant distance. The fact that the rift was within or behind the volcanic arc indicates that the vergence of subduction within the Vardar zone had reversed, from southward during the mid-Cretaceous deformation to northward subduction during the Late Cretaceous magmatism and rifting (see also Bonev and Stampfli, 2011). The formation of the backarc extension resulted in a pause in shortening deformation that caused a break between the mid-Cretaceous and latest Cretaceous–Paleogene shortening.

The second period of deformation occurred within the foreland fold-thrust belt at the same time as the backarc or intraarc basin closed and metamorphism and deformation occurred within the Rhodope Mountains (Fig. 16). The timing of metamorphic and deformational events within Rhodope is becoming clearer (e.g., Bonev and Beccaletto, 2007; Bonev et al., 2006; Burg, 2012; Silke et al., 2010). Much of the deformation and metamorphism is latest Cretaceous to about middle Eocene and forms the core of the Balkan orogen at the time the second shortening deformation occurred within the foreland fold-thrust belt and during the closure of the backarc-intraarc basin. The Rhodope deformation is related to northward subduction within the Vardar zone along the south margin of the Balkan orogen (Burg, 2012); therefore, the foreland fold-thrust belt is a retroarc or antithetic belt and has a dynamic setting different from that of the earlier mid-Cretaceous deformation. There are suggestions that the Rhodope area was not a single continental unit at the time; it may have contained ophiolitic rocks (e.g., see Turpaud and Reischmann, 2010; Burg, 2012), so the details of this part of the orogen remain to be determined. However, it appears that the foreland fold-thrust belt formed the northern marginal tectonic element of the orogen that extends from the northward subduction within the Vardar zone that led to its closure by middle Eocene time, through a thickening Rhodope metamorphic core in the south to the Moesian Platform in the north.

Within the metamorphic core there is local(?) evidence for Maastrichtian–Paleocene extension in the eastern Rhodope unit near Krumovgrad, where coarse clastics appear to be back-rotated in the hanging wall of a south-dipping extensional detachment fault (Boyanov and Goranov, 2001; Bonev et al., 2006; Fig. 10). Although it is one of very few places where such evidence exists, it is unknown how widespread such Late Cretaceous–Paleocene extension may be. Timing of such extension would be contemporaneous with the second period of deformation within the foreland fold-thrust belt (see Fig. 8). If so, it remains to be determined how this extension fits within the broader tectonic setting of the orogen and to the retroarc setting of the fold-thrust belt, but extension contemporaneous with regional shortening has been recognized in many orogens (e.g., Burchfiel et al., 1992).

The folds and thrust faults of the second event continued to the northwest into the Southern Carpathians where Late Cretaceous structures are well known; however, their continuity has been greatly disrupted by Cenozoic strike-slip faulting and clockwise rotation (Fügenschuh and Schmid, 2005; see following discussion). Deformation of this period and the following period also disrupted the Late Jurassic–Early Cretaceous paleogeography and the mid-Cretaceous structures, so their continuity is not clear. However, there have been attempts to elucidate the trans-border correlations (Tchoumatchenco et al., 2011, and references therein; Kräutner and Krstić, 2003).

The third deformational event in the foreland thrust belt of middle Eocene to early Miocene time took place when the tectonic setting of southern Bulgaria and northern Greece became an area of regional extension that has lasted to the present, a major change from the preceding regional shortening environment. Thus the dynamic setting for shortening within the foreland fold-thrust belt of its third event is different from the two preceding shortening events. However, the timing of the change is complex and its transition is probably not contemporaneous or abrupt, and may have been protracted.

The third period of deformation within the fold-thrust belt began in the late-middle Eocene, but the upper time limit for this deformation is poorly constrained because overlying strata are not present across much of the belt. Where present the strata are locally middle Miocene or younger, except in eastern Bulgaria, where the structures trend into the Black Sea. Thus, whether some of the structures belong to this period of deformation remains uncertain. It is clear that this period of shortening deformation was contemporaneous with regional extension to the south, making the dynamic setting for this third period of deformation different from the two previous deformational periods. The tectonic setting is based on regional considerations that the southern part of the orogen in the Rhodope and parts of the Sredna Gora units became an area of regional extension that began in late-middle to late Eocene time and has dominated the region of southern Bulgaria until present (Burchfiel et al., 2008). However, because the timing of extension and the first shortening deformation between southern and northern Bulgaria, respectively, comes close to overlapping with the second and third periods of shortening, it suggests a transition between the dynamics of the second and third periods of deformation. The younger part of the third period of shortening deformation within the fold-thrust belt was in a tectonic setting that is contemporaneous with regional extension in the southern Balkans.

Our favored interpretation is that the third period of deformation is related to a dynamic system that includes the northern part of the foreland and thrust belt in Bulgaria and the Southern and eastern Carpathians as these units were molded around the western part of the Moesian foreland crust. Tectonic analyses and paleomagnetic data indicate that the western part of the Southern Carpathians moved northward into the eastern Carpathians by at least latest Cretaceous time, shearing right laterally with respect to the Moesian crust to the east (Fig. 17). This crustal unit moved north and east, rotating clockwise, into the subduction zone within the eastern Carpathians and was molded around the western and northern part of the Moesian crust (Fügenschuh and Schmid, 2005; van Hinsbergen et al., 2008). During the early part of this motion, right shear along the western part of the Moesian crust was transferred into shortening within Bulgaria, molding rocks to the southern part of the Moesian crust. The transfer of right shear into shortening ended by early Miocene time, but the northward motion in the western part of the Southern Carpathians and Serbia continues to the present, manifested by localization of right shear forming the discrete Cerna-Jiu and Timok faults, which have a minimum of 50 km of right slip (see Kräutner and Krstić, 2003). The map of Kräutner and Krstić (2003) also showed geology indicating that units within the northwestern Bulgarian fold-thrust belt were rotated clockwise into the Southern Carpathians before they were displaced by the younger discrete faults.

The change in the regional tectonic setting during Paleogene time changed the nature of tectonic activity across the northern Balkan orogen. The change from a convergent to extensional orogen in southern Bulgaria and northeastern Greece is related to the closure of the Vardar ocean and a change in location of subduction from the Vardar zone to west and south into the Hellenides and south into southwestern Turkey (Papanikolaou, 2009). The beginning of extension within southern Balkan region began an evolving Cenozoic tectonic setting related to northward subduction and rollback in the Hellenides and western Turkey, usually interpreted to be a backarc extensional tectonic setting, a relation discussed by many (e.g., Le Pichon and Angelier, 1981; Jolivet and Brun, 2010; Jolivet et al., 2013) and not discussed further here.

North of the extensional region, shortening occurred within the foreland fold-thrust belt in Bulgaria and continued westward into the Southern Carpathians of Romania, where deformation involved strike-slip faulting, shortening, clockwise rotation, and orogen-parallel extension (Fügenschuh and Schmid, 2005; van Hinsbergen et al., 2008). The Cenozoic deformation of the arcuate Southern Carpathians mountain chain is the result of molding of deformed units against an arcuate foreland continental terrane as oceanic and/or thinned continental crust was subducted westward and rolled back in the eastern Carpathians, closing from north to south (Burchfiel, 1976, 1980; Fügenschuh and Schmid, 2005). The westward extension of the closure continues from the southern East Carpathian thrust belt into the Southern Carpathians; however, in northwest Bulgaria oceanic crust was lacking during the Cenozoic, and eastward or northward thrusting around the increasingly convex-west curved belt was intracratonic and associated with continued Cenozoic motion that developed discrete right-lateral strike-slip faults (Ratschbacher et al., 1993; Fügenschuh and Schmid, 2005; Fig. 17).

In early Cenozoic time convergence was east-west within and west of the Moesian Platform within continental crust in the Southern Carpathians, forming thrusts and folds that extend into northwest Bulgaria during the third crustal shortening event. By early Miocene time the transfer of right shear to shortening ceased in northwest Bulgaria. Kounov et al. (2011) interpreted the southern extension of the strike-slip faults to continue into western Bulgaria, where they may merge into transpressional steep thrust faults (see also Gerdjikov and Georgiev, 2006), faults described by Zagorchev (1993) as thrust faults of Paleogene age. The thrust faults and folds within northwestern Bulgaria appear to end against the north-striking Cerna-Jiu and Timok faults, but these are late Cenozoic faults and whether the thrust faults in northwest Bulgaria had continuations in the Southern Carpathians or were terminated or merged into early Cenozoic right shear zones remains to be investigated.

The transfer of strike slip to generally northward thrusting may explain the difference in structural style between structures in northwest Bulgaria and the remainder of the thrust belt to the east. It is unclear how far east our postulated transfer motion extends; it could extend across the entire fold-thrust belt, but with decreasing involvement of basement rocks and possibly displacement, or it could begin to grade into relative southward motion of a combined Moesian–Black Sea crust, causing minor shortening along its southern boundary. This problem is unresolved. Nevertheless, the tectonic setting for the third period of convergent deformation is clearly different from the previous two periods where stresses cannot be transmitted from the south through the South Balkan extensional system.

The northward motion along the strike-slip faults within the Moesian crust within the Southern Carpathians (the Danubian unit) also caused northeast extension along the eastern contact between the Getic and Danubian units, where a detachment fault is along their contact (Fig. 17; Fügenschuh and Schmid, 2005). Additional step-over transfer to extension farther south within western Bulgaria may also have contributed to the extensional detachment with the Osogovo Mountains (Kounov et al., 2004, 2010).

The timing of the change from orogen-wide compression across the Balkan orogen from the Vardar zone to the fold-thrust belt to the shortening within the fold-thrust belt that is contemporaneous with extension within the southern Balkans and related to Carpathian tectonics is not well established. The oldest thrusting within the third period of deformation within the fold-thrust belt is late-middle Eocene, and the final closure of the Vardar zone was at about the same time, thus it is unclear if this middle Eocene deformation is the last event related to retroarc antithetic deformation and older than the time when the regional extension began in the southern Balkan area. By late Eocene time extension was dominant in the southern Balkans. How, and when, the change from the retroarc to transfer tectonics occurred remains to be established; the transition may have been protracted and diachronous. Nevertheless, the third dynamic system that produced the youngest structures in the fold-thrust belt was active by late Eocene to early Miocene time.

By late Eocene time the shortening within the thrust belt of northern Bulgaria was contemporaneous with extension in the southern Balkans and separated the Balkan orogen into two different deformational provinces, a South Balkan extensional regime (Burchfiel et al., 2008) and a northern convergent area with the boundary approximately along the Sub-Balkan graben system (Tzankov et al., 1996; Figs. 12, 14, and 17). Thus, this youngest period of shortening in the Forebalkan fold-thrust belt is related to a tectonic setting different from the previous two deformational events.

The fold-thrust belt within Bulgaria formed in three separate events related to three different tectonic settings, thus making it a unique foreland fold-thrust belt. A mid-Cretaceous shortening event was localized along extensional flysch-filled troughs formed by rifting within the Moesian continental crust during latest Jurassic–Early Cretaceous time. Folds and thrust faults formed within and beyond the flysch troughs and were related to north- and northeast-vergent thrusting in southern part of the Moesian crust north of the Vardar ocean. We interpret the shortening of the fold-thrust belt to be synthetic with respect to the southward subduction in the Vardar zone (sensu lato). This shortening ceased when Late Cretaceous intraarc and/or backarc extension occurred through central Bulgaria in the Sredna Gora unit related to northward subduction and rollback of the subducted slab.

The second period of shortening within the fold-thrust belt developed during closure of the intraarc or backarc rifts during Late Cretaceous–Paleocene time, but the basal décollement below the belt probably extended south below the Sredna Gora. Deformation in the fold-thrust belt was part of orogen-wide shortening related to northward subduction within the Vardar oceanic zone, a reversal of subduction from the earlier shortening event: the fold-thrust belt formed in an antithetic, retroarc tectonic setting and was contemporaneous with metamorphism and shortening within the Rhodope area.

The third and final shortening event within the fold-thrust belt began in late-middle to late Eocene time, and may have continued to early Miocene time. We relate this event to transfer of transpressive right shear within northwestern Bulgaria, eastern Serbia, and the Southern Carpathians into shortening within the fold-thrust belt in northwest Bulgaria. The right shear continues to the present, but the transfer into shortening in northwest Bulgaria ceased in early Miocene time. Late Cenozoic right shear became localized within the Southern Carpathians, forming the Cerna-Jiu and Timok faults. The deformation is part of the complex Cenozoic deformations within the Southern Carpathians, where deformation is a consequence of northward and convergent motion causing molding of tectonic units around the Moesian continental crust. Late Eocene to early Miocene shortening in the fold-thrust belt was contemporaneous with and separated tectonically from regional extension in the southern part of the Balkan orogen, the former site of major shortening, metamorphism, and plutonism. The change from retroarc shortening to shortening related to molding of tectonic elements around the Moesian crust began in the middle Eocene; the change may have taken place during the earliest deformation in the third deformational event, but this is unknown due to poor timing constraints. Nevertheless, the third deformational event is related to a dynamic system distinct from the older two, and occurred when the orogen in Bulgaria was separated into two distinct parts: a northern part characterized by shortening and a southern part characterized by regional extension.

Our work in Bulgaria has taken place for more than two decades and has been supported by several different sources. The work began with support from Maxus Energy Corporation, but the company was sold in the l990s; their support was concurrent with that from National Science Foundation (NSF) grants EAR-9903021 and EAR-9628225. Final support was from a grant for the Medusa Project by NSF Continental Dynamics Program grant EAR-0409373. The Schlumberger Chair and the Department of Earth, Atmospheric and Planetary Sciences at the Massachusetts Institute of Technology supported the development of the manuscript and illustrations. This work could not have been done without the close cooperation and continued support of the Geological Institute of the Bulgarian Academy of Sciences (Sofia). We thank A.M.C. Şengör, W. Frisch, and Alexandre Kounov for reviews of the manuscript.