Do major Neogene hiatuses in the Ciscaucasian semi-enclosed basin ( Eastern Paratethys , southwestern Russia ) record eustatic falls ?

Hiatuses in semi-enclosed basins can be caused by either eustatic falls or local tectonic uplifts. The Ciscaucasian basin is located in the south of European Russia. In the Neogene, it belonged to the Eastern Paratethys domain. On the basis of available stratigraphic data, four major hiatuses are traced in this basin as erosional surfaces or lengthy sedimentation breaks, namely the Tarkhanian, Middle/Upper Sarmatian, Sarmatian/Maeotian, and Kimmerian hiatuses. They are documented in most of the areas of the study basin. The three earlier hiatuses mark short-term and nearly isochronous, basinwide sedimentation breaks, whereas the latter hiatus is diachronous, embracing more than 2 myr. All reported hiatuses record the eustatic falls. Consequently, we argue that eustatic processes controlled sedimentation in the Ciscaucasian basin throughout the entire Neogene. This means the basin was connected to the open ocean throughout this period, with important consequences for our understanding of watermass history in the Mediterranean and Paratethyan basins further west and south.


Introduction
The Paratethys was a major palaeogeographical domain, consisting of a constellation of small sedimentary basins.During the Cenozoic, it stretched from the Alps in the west to the Caspian Sea in the east.The basins were isolated partially from the Mediterranean by the tectonic uplift associated with the Alpine Oro-geny, with important consequences for their watermass history and palaeoecology (RÖGL & STEININGER 1983;RÖGL 1996RÖGL , 1998RÖGL , 1999;;STEININGER & WES-SELY 1999;GOLONKA 2004;POPOV et al. 2006POPOV et al. , 2010;;KRIJGSMAN et al. 2010).Traditionally, the Paratethys is subdivided into three parts; Western, Central and Eastern (Fig. 1).Due to their peripheral connection with the World Ocean and with the Mediterranean (e.g., POPOV et al. 2006;KRIJGSMAN et al. 2010), Paratethyan basins offered an environment in which eustatic signals, particularly those encompassing the onset and duration of the so-called "Messinian Salinity Crisis", may be amplified within sedimentary successions.However, these successions may also have been modified by local tectonic activity.
Despite decades of research, the Eastern Paratethys has remained relatively poorly-known within the international audience and reviews of Paratethyan basin evolution have been overwhelmingly concerned with the Western and Central parts (SISSINGH 2001;BERGER et al. 2005;HARZHAUSER & MANDIC 2008;HARZHAU-SER et al. 2008;LIRER et al. 2009).Publications by RUBAN (2005), POPOV et al. (2006), and KRIJGSMAN et al. (2010) are amongst the minority of papers within the international scientific press specifically concerned with the Eastern sub-basins.Nevertheless, this region may provide some important clues to the under-standing of basinwide environmental changes during the Neogene.For example, so long as the connection between the Mediterranean and Eastern Paratethys remained open, the net precipitative flux in the Caucasian region, which receives up to 3,000 mm precipitation per year today, is likely to have been large enough to alter the degree of salinification of Mediterranean water.Consequently, before the Messinian Salinity Crisis can be understood mechanistically, it is critical that the presence/absence and magnitude of freshwater supply from the Eastern Paratethys to the Mediterranean is established (MEIJER & KRIJGSMAN 2005;KRIJGSMAN et al. 2010).
During the Neogene, the Earth experienced a series of glaciations and tectonic events (KENNETT 1977;ZACHOS et al. 2001;SMITH & PICKERING 2003;GOR-NITZ 2009), which resulted in a complicated chain of eustatic changes (HAQ et al. 1987;HAQ & AL-QAH-TANI 2005;MILLER et al. 2005;KOMINZ et al. 2008).Investigating evidence for the presence or absence of their signatures in the semi-enclosed Ciscaucasian basin therefore provides the empirical test for the connection of these basins to the World Ocean via the Mediterranean, and consequently the first direct indicator that these basins were a potential source of freshwater to the Mediterranean during the Messinian.This effort is in analogy to a previous study in Japan, where HIROKI (1995) and HIROKI & MATSUMOTO (1999, 2003) investigated eustatic signals within the Miocene sequence boundaries in central Honshu.Despite the complexity of the tectonic setting of Japan, which does drive some local differences between basins, a number of common surfaces were recognized as being traceable through the entire region, and positive shifts in δ 18 O that occurred synchronously with these relative sea level falls suggesting a relation to the phases of growth of the Antarctic ice sheet, and consequently to eustasy.The accumulation of deposits of the Pleistocene Atsumi Group, which occurred in the tectonically-active region, was also controlled by the eustatic fluctuations (HIROKI & KIMIYA 1990).
This paper is aimed at tracing major hiatuses in the Neogene sedimentary successions of the Ciscaucasian basin, which represents the central component basin of the Eastern Paratethys, with the intention of testing the degree of connectivity with the World Ocean.Eustasy and local tectonic activity are two important controls on the basinwide depositional settings (CATU-NEANU 2006).In the case of a basin positioned within the foreland of an active collisional zone, the local tectonics may reasonably be expected dominate the basin evolution.The main Caucasian orogeny started in the Greater Caucasu in the Paleogene, and accelerated from the mid-Sarmatian, i.e., early Tortonian (ERSHOV et al. 2003).SAINTOT et al. (2006) prescribes the Sarmatian tectonic pulse as the crucial event for the evolution of the entire region.This tectonic activity would definitely result in the development of major hiatuses spread across the entire Ciscaucasian basin or, at least in its southern areas.Should major changes in sedimentation coincide with major changes in eustatic sea level, this would confirm that the basin history is dominated by global (largely climatic) rather than local (largely tectonic) influences.If no coincidence can be shown, then the opposite conclusion may be drawn.The knowledge of hiatuses is therefore crucial to link regional sedimentation breaks with global environmental perturbations in this area.

Geologic setting
The Ciscaucasian basin is a typical foreland basin, which formed between the emergent Greater Caucasus in the south, which was probably rising during the period of interest of this study, and the stable Russian Platform in the north (ERSHOV et al. 2003;SAINTOT et al. 2006).As in the case of other Paratethyan basins, its origin and tectonic evolution were both related closely with the Alpine Orogeny (GO-LONKA 2004).In the Neogene, the Ciscaucasian basin was wide and had an asymmetrical profile, with its deepest part located close to the island of the Greater Caucasus (i.e. in the south).The Ciscaucasus basin was connected with the Euxinic basin in the west and the Caspian basin in the east (NEVESSKAJA et al. 1984;POPOV et al. 2006POPOV et al. , 2010;;Fig. 1).
The Neogene deposits vary in time and space within the Ciscaucasian basin.Sandstones, siltstones, and shales are dominating lithologies, whereas carbonates (including bioclastic limestones), conglomerates, diatomites, and other sedimentary rocks are also known.On the basis of lithology and facies, 17 areas are distinguished within this basin (NEVESSKAJA et al. 2004(NEVESSKAJA et al. , 2005; Fig. 2).Each area represents a peculiar Neogene sedimentary succession.The total thickness of Neogene deposits reaches up to 5700 m, and both short-term hiatuses (documented as erosional surfaces) and long-term hiatuses (represented by unconformities) occur within the succession (Fig. 3).The Neogene depositional environments in the Ciscaucasian basin did not remain constant.The position of the shoreline fluctuated significantly alongside the basin depth (Popov et al. 2010).Palaeoecological studies (ILYINA et al. 1976;NEVESSKAJA et al. 1984NEVESSKAJA et al. , 1986;;POPOV et al. 2006) also suggested significant changes in salinity of the Eastern Paratethys.Although these   (after NEVESSKAJA et al. 2004(after NEVESSKAJA et al. , 2005)).Data on the Rostov Dome are taken from RUBAN (2002RUBAN ( , 2005)).changes occurred cyclically, a general trend towards a decrease in salinity can be traced.Undoubtedly, this is linked with a more or less gradual isolation of the Eastern Paratethys from both the Mediterranean and the other Paratethyan counterparts.
Neogene lithostratigraphy of the Ciscaucasian basin is summarized by NEVESSKAJA et al. (2004NEVESSKAJA et al. ( , 2005)), who re-evaluated the available information, and defined or re-defined formations and groups.The ages of these lithostratigraphic units are established on the basis of bivalves, foraminifera, mammals, ostracods, calcareous nannoplankton, and other palaeontological data, and thus is based on published frameworks for the regional stages of the Eastern Paratethys (RÖGL 1996;STEININGER 1999;NEVESSKAJA et al. 2004NEVESSKAJA et al. , 2005;;POPOV et al. 2006;RUBAN 2009).Correlation of regional and global stages remains uncertain, however, because of poor biostratigraphic control of the correlation between the Eastern Paratethyan region fra-mework and global chronostratigraphy (see discussions in KRIJGSMAN et al. 2010).Detailed correlation between global chronostratigraphic stages fixed by Gobal Stratotype Sections and Points (GRADSTEIN et al. 2004;OGG et al. 2008) and the Neogene succession of the Eastern Paratethys remains an objective for further studies.Meanwhile, absolute dating of regional stage boundaries CHUMAKOV et al. (1992a, b) is the primary basis of correlation of existing regional and global stages (RUBAN 2005(RUBAN , 2009;;Fig. 4).New results obtained by KRIJGSMAN et al. (2010) facilitate this correlation significantly.

Materials and methods
We use the dataset compiled by NEVESSKAJA et al. (2004) as the basis for tracing major hiatuses in the Neogene sedimentary successions of the Ciscaucasian basin.For the territory of the Rostov Dome (area 14 on Figs. 2, 3), previous constraints by RUBAN & YANG (2004) and RUBAN (2002RUBAN ( , 2005) ) as well as results from new field investigations are used.Correlation between global and regional stages is based on the framework proposed by RUBAN (2009), which takes into account recent chronostratigraphical developments (STEININGER et al. 1997;CASTRADORI et al. 1998;RIO et al. 1998;HILGEN et al. 2000a,b;2003, 2005, 2006; VAN COUVERING et al. 2000;KUIPER 2003;BILLUPS et al. 2004;GRADSTEIN et al. 2004;KUIPER et al. 2005;HÜSING et al. 2007HÜSING et al. , 2010;;OGG et al. 2008), absolute dating of Upper Miocene regional stage boundaries (CHUMAKOV et al. 1992a, b), and earlier constraints by NEVESSKAJA et al. (2005).New results presented by KRIJGSMAN et al. (2010) are also accounted.Examples from the Swiss Molasse basin (BERGER et al., 2005), the Dacian basin (VASILIEV et al. 2004), and the Central Paratethys (LIRER et al. 2009), provide great confidence in the efficacy of our approach.
Regional hiatuses are considered major if they can be traced in most of the areas of the Ciscaucasian basin.The next step is comparison of regionally-documented major hiatuses with global eustatic falls.For this purpose, we used two widely-accepted eustatic curves.Although the compilation by MILLER et al. (2005) was updated by KOMINZ et al. (2008), these authors altered only the pre-Pliocene part of the dataset, so here we use the original data of MILLER et al. (2005).The second eustatic curve considered in this paper is that proposed by HAQ & AL-QAHTANI (2005), who updated the earlier constraints by HAQ et al. (1987).The correlation of basinwide major hiatuses and eustatic fluctuations is possible on the basis of the correlation between regional and global Neogene stages.
We assume that a coincidence of major basin-wide hiatuses and eustatic falls indicates a global sea-level control on regional sedimentation.Absence of this signal either indicates a lack of connectivity with the open ocean or complication derived from local tectonic activity.Basin subsidence larger than eustatic fall would prevent a hiatus from appearing, whereas uplift would produce additional hiatuses.Thus, finding a significant coincidence of hiatuses and eustatic falls is a good indication of a relatively stable tectonic regime and absence of significant activity within the given basin.

Tracing the major hiatuses
Four major hiatuses can be documented within the Neogene deposits of the Ciscaucasian basin (Fig. 4).The lowest encompasses the entire Tarkhanian regional stage, and affects the succession in 14 of the 17 areas of the basin (Fig. 3).However, this hiatus is diachronous, appearing in some areas as erosional surfaces at the bottom and/or the top of the Tarkhanian, whereas in other areas it embraces the entire stage.An increase in the number and extent of sedimentation breaks occurred in the Kotsakhurian and remained until the Karaganian, indicating that this major hiatus was a culmination of sedimentation disruption, which embraced 3 regional stages.Despite some diachroneity of this hiatus, the absolute timerange encompassed was not so extensive, around 0.5 Ma, because the absolute duration of the Tarkhanian stage was probably short (NEVESSKAJA et al. 2004(NEVESSKAJA et al. , 2005;;RUBAN 2009).The Tarkhanian hiatus corresponds to the Burdigalian/Langhian boundary of the global chronostratigraphic scale (RUBAN 2009; Fig. 4).
The second major, but short-term hiatus, which modifies the succession in 13 of the 17 areas (Fig. 3), is observed within the Sarmatian regional stage (Fig. 4).This hiatus is a generally isochronous erosional surface with few exceptions.In the area 1, this surface appears to be diachronous (Fig. 3), whereas long-term hiatuses are registered in the areas 14, 16, 17, and, partly, in the area 15 (Fig. 3).According to data presented by NE-VESSKAJA et al. (2004), this hiatus marks the boundary between the Middle Sarmatian and the Upper Sarmatian, for which an absolute age was established by CHUMAKOV et al. (1992b) of 11.2 Ma, which lies just above the Serravallian/Tortonian boundary dated as 11.608 Ma (OGG et al., 2008;Fig. 4).
The third major short-term hiatus is established at the top of the Sarmatian regional stage (Fig. 4).It is traced in 15 of the 17 areas of the Ciscaucasian basin (Fig. 3), and it is marked by a slightly diachronous erosional surface which is sometimes embraced by lengthy hiatuses.Diachroneity is evident from the areas 6, 7, and 8, where erosional surfaces are traced below the upper boundary of the Sarmatian (NEVES-SKAJA et al., 2004; Fig. 3).The Sarmatian/Maeotian hiatus occurs within the middle interval of the Tortonian global stage (RUBAN; 2009; Fig. 4).
The last major hiatus is pronounced in both its duration and spatial extent (Fig. 4).It encompasses the entire Kimmerian regional stage.Its signatures (erosional surfaces and lengthy hiatuses) are found in 15 of the 17 areas in the basin (Fig. 3).As in the case of the Tarkhanian hiatus, the concentration of hiatuses in the sedimentary successions appears to pre-date the major break in sedimentation since the upper Pontian, and continues up to the upper Aktchagylian with a culmination in the Kimmerian.Plotted against the global chronostratigraphic scale, this major hiatus started in the late Messinian (the Messinian/Zanclean boundary has an age of 5.332 Ma;OGG et al. 2008) and ended in the early Piacenzian as one may judge by stage correlations attempted by CHUMAKOV et al. (1992b) and RUBAN (2009) and improved recently by KRIJGSMAN et al. (2010).The time span of this hiatus exceeded 2 myr.
We can thus distinguish two kinds of Neogene major hiatuses in the Ciscaucasian basin.The Tarkhani-an, Middle/Upper Sarmatian, and Sarmatian/Maeotian hiatuses, which were short-term and relatively isochronous, and the Kimmerian hiatus, which was long-term and diachronous.

Major hiatuses versus eustatic falls
Comparison of the timing of major Neogene hiatuses in the Ciscaucasian semi-enclosed basin with eustatic fluctuations through the same period (Fig. 4 It is important to question whether there were significant eustatic falls, which did not leave an imprint in the Neogene stratigraphic record of the Ciscaucasian Basin.The falls of such kind occurred in the late Aquitanian, the mid-Burdigalian, the mid-Langhian, and probably in the early Gelasian (Fig. 4).Weak or no local evidence of these falls can be found (Fig. 3).It should be noted, however, that a lack of correspon- dence between some eustatic falls and the stratigraphic architecture of the Ciscaucasian Basin does not disprove an eustatic control on the basinwide depositional setting.This is likely to reflect that tectonic conditions in the basin masked the eustatic signal during some time intervals (NEVESSKAJA et al. 1984).We conclude that there is evidence of persistant, if punctuated, eustatic control on sedimentary rearrangments in the Ciscaucasian Basin throughout the entire Neogene.

Major hiatuses and orogeny
The available data (NEVESSKAJA et al. 2004(NEVESSKAJA et al. , 2005; Fig. 3) provide evidence that the four most significant hiatuses in the Ciscaucasian basin all coincide well with major eustatic falls (Fig. 4).Therefore, though intuitively it might be assumed that in a foreland semi-enclosed setting such as this tectonics would dominate over eustasy, this does not appear to be the case for the Ciscaucasian basin.In particular, we highlight the two major hiatuses reported from the Sarmatian, which were near-isochronous and short in duration, and so could not be produced by tectonic activity.The likely eustatic origin of these hiatuses provides a disproof of previous assumptions of a Sarmatian pulse or an acceleration in orogeny (ERSHOV et al. 2003;SAINTOT et al. 2006).
If even local tectonic activity in the Ciscaucasian Basin or in the neighbour Greater Caucasus explains a lack of regional signature of some eustatic falls (see above), this fact is not enough to hypothesize any significant tectonic pulses for at least two reasons.First, major regional hiatuses linked to global sea-level falls may be absent in only the case of increasing subsidence (e.g., this might have been the case during the pre-Tarkhanian interval), but not uplift.Second, there were eustatic falls with no major hiatuses in the Ciscaucasian Basin, but all major hiatuses have an appropriate eustatic explanation.

Other local hiatuses and local tectonics
We do not observe numerous local hiatuses in the lower-middle Miocene stratigraphic interval (Fig. 3).Their occurrence increases at the Kotsakhurian-Tchokrakian interval, which is linked to a series of prominent eustatic lowstands (Fig. 4).Many local (i.e., those registered in few areas only) Miocene hiatuses were short-term, and they are marked often by erosional surfaces or significant interruptions in the sedimentary record (NEVESSKAJA et al. 2004).These local events are unlikely to have been formed by tectonic pulses.There is some increase in both the quantity and the duration of local hiatuses in the latest Miocene and Pliocene (Fig. 3), but this coincides with the onset of higher frequency eustatic fluctuations (HAQ & AL-QAHTANI 2005;MILLER et al. 2005) linked to the strengthening of Antarctic glaciation and then an appearance of ice sheets in the Arctic (ZACHOS et al. 2001;GORNITZ 2009;Fig. 4).
The areas 1, 2, 3, 6, 7, 8, 9, 11, 12, and 13 located in the south of the basin, i.e., along the Greater Caucasus, are supposed to be most prone to tectonic influence (Fig. 2).However, these areas are not distinguished by a higher number of local hiatuses in comparison to other areas (Fig. 3).In contrast, areas located on the gentle northern slope of the Ciscaucasian basin (14-17 -see Fig. 2) are characterized by a higher number of local hiatuses, which is consistent with frequent interruption of sedimentation on the shallow basin periphery, where even small eustatically-driven fluctuations led to the emergence of large areas.
The clear regional signature of the global eustatic fluctuations in the Ciscaucasian basin implies a rather stable geodynamic regime, confirming an earlier assumption made by EFENDIYEVA & RUBAN (2009).Our results do not imply an absence of tectonic activity in the Greater Caucasus or its influences on sedimentation in the Ciscaucasian basin.In fact, tectonism might have been responsible for some local hiatuses.However, it seems that eustatic control prevailed over local tectonic control within the Ciscaucasian basin during the Neogene.Further structual, fission-track, and isotope studies will allow testing of the exact timing of deformation phases and uplifts in the Caucasian region.

Connections of the Eastern Paratethys
One further inference must be made when documenting the evident eustatic control on the Neogene sedimentation in the Ciscaucasus.It has already been hypothesized, particularly by RÖGL & STEININGER (1983), NEVESSKAJA et al. (1984), CHEPALYGA (1995), STEININGER & WESSELY (1999), POPOV et al. (2006), and KRIJGSMAN et al. (2010), that the Eastern Paratethys retained at least ephemeral connections with the World Ocean via the Mediterranean Sea or the Indian Ocean until the end of the Neogene.Our results confirm this was present during the majority of the period studied.It therefore becomes crucial to consider whether this connection was via an Indian Ocean corridor or through an Euxine basin corridor.Given our knowledge of the palaeogeography of the time, the latter seems more likely and this has significant consequences for our understanding of the Messinian Salinity Crisis in the Mediterranean.The modern net freshwater flux from the Black Sea into the Mediterranean reducesthe total net freshwater export from the basin by 10% (BETHOUX & GENTILI, 1999) and the presence/absence of this flux is one of the most important unresolved issues in quantitative asses-sment of the Miocene Mediterranean (MEIJER & KRIJGSMAN 2005;ROHLING et al. 2008;KRIJGSMAN et al. 2010).Incorporation of the net freshwater flux from the Ciscaucasian and Caspian basins, which is the implication of the basin connectivity described in this paper, could mean that the Euxinic net freshwater flux was an even more important parameters in determining late Neogene Mediterranean palaeoceanography than it is in the late Quaternary.This connection is well reflected in the close relationships between Late Messinian Lago-Mare faunas from the Mediterranean and Ciscaucasian basins (ESÜ 2007) caused by a westwards faunal invasion from the Paratethyan basins into the Lago-Mare basins.Early Messinian links between the two regions can also be demonstrated on the basis of cardiid bivalve faunas common to both southern Italy and the Ciscaucasian regions (PEDLEY et al. 2008) and indicates an earlier global eustatic fall which encouraged ecological "leakage" from the Paratethys into the semi-isolated Mediterranean basins.Compelling evidence for earlier global eustatic control influencing water exchange between the two interconnected regions is demonstrated by the Tarkhanian event which correlates precisely with a major Burdigalian/Langhian lowstand within the Mediterranean (GRASSO et al. 1994).
Further work on the location of the connections between the Eastern Paratethys and Mediterranean basins the watermass exchanges associated them should therefore be a priority for future research.

Conclusions
Four major Neogene hiatuses are traced in the Ciscaucasian semi-enclosed basin, which played a key role in the Eastern Paratethys domain.These include the Tarkhanian (Burdigalian/Langhian), Middle/Upper Sarmatian (lower Tortonian), Sarmatian/Maeotian (mid-Tortonian), and Kimmerian (late Messinian-early Piacenzian) hiatuses.The Ciscaucasian successions reflect well the eustatic falls recorded by global sea level datasets (HAQ & AL-QAHTANI 2005;MILLER et al. 2005).Eustatic control on basinwide sedimentation breaks persisted througout the Neogene, which suggests a relatively "calm" tectonic regime and rather stable connections of the Eastern Paratethys and the World Ocean.
Further studies should be aimed at a precise reconstruction of the Neogene transgressions/regressions and depth changes in the Ciscaucasian basin.These may then be compared with known eustatic fluctuations and the position of the corridor connecting the Caucasian region with the World Ocean.KOMINZ et al. (2008) pointed out a broad interregional comparison of data on sea-level changes as the most desirable tool to reveal the true eustatic changes, but it is equally true that identification of known eustatic signals can be critical in understanding the history of poorly-known regions.
) indicates a high degree of coincidence, especially with the HAQ & AL-QAHTANI (2005) dataset.The Tarkhanian (Langhian/Burdigalian) hiatus corresponds to the pronounced global sea-level fall documented by MILLER et al. (2005) and is similar in timing (post-dates by no more than 0.5 myr) to a fall indicated by HAQ & AL-QAHTANI (2005).Additionally, the noted regional hiatus coincides some inversion of eustatic trends.The Middle/Upper Sarmatian (lower Tortonian) hiatus coincides with the onset of a very abrupt and strong eustatic fall on the curve of MILLER et al. (2005) and again post-dates (by no more than 0.5 myr) the fall indicated by HAQ & AL-QAHTANI (2005).The Sarmatian/Maeotian (mid-Tortonian) hiatus corresponds well to the global sea-level fall documented by both alternative curves (HAQ & AL-QAHTANI 2005; MILLER et al. 2005).Finally, the Kimmerian (late Messinianearly Piacenzian) hiatus formed at a time marked by a strong eustatic fall registered as by HAQ & AL-QAHTANI (2005) and a period of strong eustatic variability in the MILLER et al. (2005) dataset.If growth and fluctuation of Antarctic and then both Antarctic and Arctic ice sheets (KENNETT 1977; ZACHOS et al. 2001; SMITH & PICKERING 2003; GORNITZ 2009) is presumed as a main control on the global Neogene sea-level changes (MILLER et al. 2005; KOMINZ et al. 2008; GORNITZ 2009), we need to hypothesize a direct influence of the global climate perturbations on the regional sedimentation in the Ciscaucasian basin, because all major hiatuses from there coincide well with the global eustatic falls.