Reconstructions of deltaic environments from Holocene palynological records in the Volga delta,northern Caspian Sea

Structural and geographical setting

The Caspian Sea is the world’s largest inland body of water containing over 40% of the lacustrine waters on the planet (Kaplin, 1995). The Caspian Sea receives freshwater input from a number of rivers of which the Volga, flowing into the north-western Caspian Sea, contributes by far the largest amount, over 80%. Salinity of the surface waters is lowest (c. 3‰ or less) close to the Volga outflow, increasing to around 13.5‰ in the central and southern basins (Kosarev and Yablonskaya, 1994). The surface of the Caspian Sea is currently about 26.5 m below global sea level (Arpe et al., 2012; Naderi Beni et al., 2013a) and not connected directly to the world’s oceans at the present time, although connections are known to have occurred in the recent geological past. A marine connection existed at the end of the Pliocene to beginning of the Pleistocene (Akchagylian stage) between c. 3.2 and 2 million years ago (Kosarev and Yablonskaya, 1994; Van Baak et al., 2013), and connections were re-established at various times throughout the Pleistocene (Badertscher et al., 2011; Forte and Cowgill, 2013; Rychagov, 1977; Tudryn et al., 2012). Most recently, high Caspian Sea levels, with connections to the Black Sea via the Manych Channel have occurred in the late Pleistocene (Khvalynian stage) as a result of increased runoff from glacial melt-waters after the ‘Last Glacial Maximum’ according to Chepalyga (2007), Svitoch (2010) and Yanina (2012).

The Caspian Sea, including the Volga delta region, is influenced by tectonics, regional topography and climate. During the Pliocene, the palaeo-Volga delta was situated about 500 km to the south of its present-day position (Kroonenberg et al., 2005; Reynolds et al., 1998) in response to the initial subsidence of the south Caspian basin and uplift of the Caucasus and Kopet Dag ranges (Allen et al., 2002). The middle Caspian basin (Derbent Depression) subsided at a later time (Dumont, 1998) and is separated from the south Caspian basin by the Apsheron Sill, an area of tectonic uplift (Allen et al., 2002, 2003) where present-day water depths vary between 80 and 200 m (Ferronsky et al., 1999; Kosarev and Yablonskaya, 1994). During the Pleistocene, at least within the last 700 kyr, extensive fluvio-deltaic sediments were deposited to the north of the present-day Caspian Sea coast, in response to major sea level changes, and these sediments formed the ‘wandering Volga delta’ (Kroonenberg et al., 1997). This Pleistocene delta complex exhibits much greater extremes of sedimentation than seen in the Holocene delta (Sidorchuk et al., 2009).

Many attempts have been made to reconstruct Holocene Caspian Sea levels (e.g. Chepalyga, 2007; Rychagov, 1997; Svitoch, 2009; Varushchenko et al., 1987). Most authors agree that there have been several transgressive/regressive cycles in the Caspian Sea during the Holocene, but there is no agreed model or consensus as to the precise timing and naming of the various Holocene Caspian Sea level events, due partly to a lack of uniformity in the dating methods and interpretations. The level of the Caspian Sea fell to about 80 m below its present level (Kroonenberg et al., 1997; Mayev, 1994; Rychagov et al., 2010) during the Mangyshlak stage at the beginning of the Holocene. The exposed shelfal areas were subject to aeolian processes, resulting in the formation of extensive east-west trending dune fields, the Baer hills. These underlie the Holocene delta (Kroonenberg et al., 2005; Overeem et al., 2003) and are visible to the west of the delta at the present time. They typically reach 20–25 m in height and often occur in association with shallow freshwater to saline lakes (Losev et al., 1998). Subsequently, it is generally agreed that Caspian Sea levels were variable, but mostly high, in the early to mid-Holocene with maximum levels of around −24 to −20 m below present-day global sea level (c. 3–7 m above present-day levels) between c. 9000 and c. 5000 years ago (Chepalyga, 2007; Rychagov, 1997). Further highstands then occurred, with Caspian Sea levels reaching −25 to −19 m below present-day global sea level (Naderi Beni et al., 2013a, 2013b) within the last few hundred years, following a late Holocene lowstand event known regionally as the Derbent. Overeem et al. (2003) showed how the very recent sediments and vegetation in the Damchik region have evolved over a period of (mainly) sea level fall since 1935, based on aerial photographs and vegetation maps. Between 1935 and 1977, Caspian Sea levels fell by c. 3 m, from c. −26 to c. −29 m (Kosarev and Yablonskaya, 1994) leading to an outbuilding of the delta by about 10 km (Alekseevskiy et al., 2010). Deposition was then fairly stable during a period of Caspian Sea level rise of around 2.7 m between 1978 and 1995; a minor fall has since been recorded and present-day levels fluctuate at around −26.5 m (Arpe et al., 2012; Naderi Beni et al., 2013a).

Hydrography and climate

The Volga catchment drains an area of approximately 1.38 million km², with a total length of the main stream of 3700 km (Sidorchuk et al., 2009). Much of the catchment falls within the temperate continental climate zone (Kosarev, 2005), but extends into humid mid-latitudes (Leroy et al., 2007). The catchment as a whole extends from approximately 46° N to 62° N and from 32° E to 60° E, and thus drains an area with considerable diversity in climate and topography. In the Volga catchment area, mean annual temperature ranges from 3°C in the north to 9°C in the south, with mean annual precipitation ranging from 750 mm in the north to 150 mm in the south (Sidorchuk et al., 2009; Figure 1b). In the delta region, average monthly temperatures vary from 26°C in July to −8°C in January (Knystautas, 1987), occasionally reaching as high as c. 45°C in summer and as low as c. −40°C in winter, with rainfall fairly evenly distributed through the year at between 200 and 250 mm (Losev et al., 1998). Much of the shallow north Caspian area freezes in winter with an average of 120–140 days of ice cover and typical ice thicknesses of 20–30 cm in the Volga mouth region (Kosarev, 2005); ice-gouge features several kilometres in length and up to 200 m wide occur on the sea bed in the northern Caspian Sea (Ogorodov and Arkhipov, 2010). Discharge from the Volga between 1880 and 1980 averaged 241 km³/yr, with approximately 60% of the total discharge occurring during the snow melt and floodwater period of April to June (Kosarev and Yablonskaya, 1994), although significant variations from this average do occur (Kroonenberg et al., 1997; Sal’nikov et al., 2010). Evaporation is most pronounced in the summer and may reach 175–200 mm/month within the delta (Losev et al., 1998) and has a major influence on Caspian Sea level (Arpe et al., 2012). Winds in the northern part of the Caspian Sea are predominantly from the east, and less frequently from the west and north-west (Kosarev, 2005). Strong winds are relatively infrequent, but storm surges do occur, especially in the shallow waters in the north Caspian Sea region (Ivkina, 2010) and periodically in the delta region, with wave heights up to 1.5 m, reaching up to 3–4 m in extreme cases (Kosarev, 2005). For example, storm surges of 2–3 m above normal occurred in 1952 and caused flooding as far north as the city of Astrakhan (Kroonenberg et al., 1997). True tides are absent in the Caspian Sea (Naderi Beni et al., 2013a), although ‘seiches’ do occur (Kosarev and Yablonskaya, 1994) resulting in minor ‘tidal’ ranges of 10 cm or less in the coastal regions (Spidchenko, 1973).

Volga delta sedimentation

The present-day Volga delta is a gently sloping ramp delta with a slope gradient of around 5 cm per km. Sediment load carried by the river consists mostly of fine sand, silt and clay. The delta bifurcates regularly because of the low gradient and fine grain size, resulting in more than 800 outlets into the Caspian Sea along a coastline of c. 200 km (Kroonenberg et al., 1997, 2005; Overeem et al., 2003; Figure 1c). The sedimentary history and ongoing sedimentary processes in the Volga delta are largely controlled by changes in the water level of the Caspian Sea, which, in turn, are driven by catchment climate and river outflow. The apex of the present delta is situated close to the city of Astrakhan and the Akhtuba tributary at an altitude of −22 m, with upper delta plain (−22 to −25 m) and lower delta plain (below −25 m) sediments deposited to the south (Rusakov, 1990). The upper delta plain region contains remnants of the Baer hills, and topographic lows between these hills are flooded by the Volga during peak discharge (Kroonenberg et al., 1997). The lower delta plain consists mainly of alluvial sediments deposited since the Caspian Sea lowering in the 1930s and contains the main Volga channels that were incised during the Derbent lowstand within the last 1000 or so years, with active progradation now occurring close to the delta front. Further seaward is an extensive submerged pro-delta (avandelta) region which extends some 80 km offshore in water depths of no more than about 4 m, and is formed mostly of eroded and re-deposited sediments (Kroonenberg et al., 1997).

The main sedimentary features and processes are described by Belevich (1956, 1960, 1963; Figure 2). During periods of sea level fall, the delta builds out (progradation) and mouth bars are formed at the delta front during times of peak river flooding. These mouth bars obstruct the water flow, becoming sub-aqueous levees, which then extend seawards. The inter-distributary bays (kultuk) are freshwater bodies (Belevich, 1956), with water depths usually of less than 1.5 m, containing freshwater and mostly floating vegetation (Kroonenberg et al., 1997). These kultuk are frequently connected to the avandelta region but may become isolated from open water during rapid progradation, and form delta lakes (ilmen). During periods of sea-level rise, deposition occurs in the more landward parts of the delta (aggradation) and the ilmen often become silted up. This pattern of sedimentation is likely to have occurred throughout the Holocene, at least subsequent to the early Holocene Mangyshlak regression (Kroonenberg et al., 2005), with alternating progradation and aggradation leading to a lateral shift in sedimentation and channel migration.

OPEN IN VIEWER

Figure 2.

Schematic summary of depositional features in the present-day Volga delta (after Belevich, 1956, 1963). Approximate location is shown in Figure 4a.

Vegetation

The Volga catchment encompasses a large variety of vegetation types, which influence the pollen types that are dispersed into the delta and Caspian Sea. The upper reaches of the Volga extend into mixed and broad-leaved forests of central and eastern areas of the East-European Plain where species of Quercus, Tilia, Ulmus, Acer, Fraxinus, Alnus, Betula and Pinus dominate now, and up to the southern limits of boreal coniferous (taiga) forest where Picea abies, P. obovata, Abies sibirica, Pinus sylvestris, Larix sukaczewii, Betula pubescens and B. pendula are common components (Figure 3). The river then flows southwards through steppe and semi-desert vegetation, characterized by Poaceae, Asteraceae, Artemisia and Amaranthaceae ¹ (Knystautas, 1987).

OPEN IN VIEWER

Figure 3.

Simplified vegetation map for the northern Caspian region, including the Volga catchment area (from various sources).

The Volga delta lies within the Volga–Akhtuba semi-arid province, which supports semi-desert vegetation, with frequent Artemisia and Poaceae. Limited forest vegetation dominated by Salix, Populus, Acer, Quercus, Fraxinus, Ulmus, Betula and Alnus grows in the floodplain areas (Bolikhovskaya and Kasimov, 2008; Nikolayev, 2007). The upper delta region, approximately to the north and north-east of Astrakhan, is now largely used for agriculture and grazing, with the natural vegetation consisting of dry steppe and semi-desert. The middle delta region occurs in a south-west to north-east belt approximately 25 km wide from north to south, and is located to the south and east of Astrakhan (Figure 1c). Part of this region is also used for agriculture and grazing, although large areas are flooded by the Volga for several months of the year, and are colonized by grassland and reedbeds. The natural (and some artificial) levees between the river channels are colonized by willows, mainly Salix alba and S. triandra (Losev et al., 1998) with other trees, including Fraxinus and Ulmus (Bolikhovskaya and Kasimov, 2008, 2010). When the Volga is confined to its river channels, the middle delta region is generally dry, and areas in between the river branches may support halophytic or xerophytic vegetation, including Amaranthaceae (e.g. Suaeda and Salicornia), Ephedraceae (Ephedra distachya) and Tamarix. The Baer hills region to the west supports mainly desert vegetation with Artemisia, Halocnemum, Suaeda and Calligonum leucocladum (Bolikhovskaya and Kasimov, 2010; Losev et al., 1998).

The most studied locality within the lower delta and avandelta is the Damchik region of the Astrakhan Man and Biosphere Reserve (Lychagin et al., 1995), where the cores for this present study were collected. This area contains both permanent and weakly running rivers and streams (Figure 4a) as well as permanent lakes (ilmen), seasonally dry lakes and freshwater embayments (kultuk). The main river channels are bounded by natural levees that are typically 0.5–2 m above low water level (Labutina et al., 1995; Figure 4b). The flora and vegetation of the Damchik region have been described in detail by Baldina et al. (1999) and mapped by Labutina et al. (1995). In the northern region, vegetation consists of forests, shrublands and grasslands. The forests include willow forest (dominated by S. alba and S. triandra) and mixed forest that includes species of Salix as well as Populus (native) and Fraxinus (introduced). Shrublands are limited to stands of Tamarix (on saline soils) and Rubus and Carex (on levees). Short grasslands occur at slightly higher elevations, which are less inundated by floodwater and contain halophytic species such as Aeluropus, Suaeda and Salicornia. Tall grasslands include mainly reedbeds, dominated by Phragmites and Typha in the partially flooded areas. In the permanently flooded regions, Labutina et al. (1995) classified the vegetation into three main types: emergent, floating and submerged. The emergents include mainly Phragmites, Typha, Sparganium, Nelumbo and Salvinia; the floating vegetation is characterized by Trapa, Nymphoides, Nymphaea and Nuphar, with submerged associations dominated by Potamogeton, Ceratophyllum, Butomus and Vallisneria spiralis (Hydrocharitaceae). A notable feature is the presence of Nelumbo nucifera (‘lotus’), a species mostly found in tropical latitudes (Baldina et al., 1999); it is unclear if this species is native to the Volga delta, but it does occur frequently, forming dense stands in the avandelta region (Baldina and Labutina, 2010). A simplified map of the vegetation in the Damchik region is shown in Figure 4c, re-drawn from the original maps of Labutina et al. (1995) and Overeem et al. (2003).

OPEN IN VIEWER

Figure 4.

Maps of the Damchik region of the Astrakhan Man and Biosphere Reserve. (a) Hydrography, showing principal rivers, streams and lakes (after Lychagin et al., 1995); (b) geomorphology and physiography (after Lychagin et al., 1995); (c) vegetation (after Labutina et al., 1995; Overeem et al., 2003) and (d) location of seismic surveys, cores and surface samples.

Caspian dinocysts

Dinoflagellate cysts (dinocysts) are an important palynological component in Pleistocene and Holocene sediments of the Caspian Sea region. Dinocysts occurring include forms originally described from the Black Sea such as Spiniferites cruciformis and Pyxidinopsis psilata (Wall and Dale, 1973; Wall et al., 1973), as well as Impagidinium caspienense and Caspidinium rugosum described from the Caspian Sea (Marret et al., 2004). The requirements and salinity tolerances of all these taxa are not entirely known but, in general, they are all characteristic of low salinity waters such as those present in the Caspian Sea at the present time. S. cruciformis is also known to occur in freshwater settings (Kouli et al., 2001; Leroy and Albay, 2010). All these dinocyst species can exhibit very varied morphologies (e.g. Leroy, 2010; Marret et al., 2004; Mudie et al., 2001), the exact significances of which are not yet fully understood but are most probably related to water temperature and salinity (Zonneveld et al., 2013). The euryhaline species Lingulodinium machaerophorum and the cosmopolitan genus Brigantedinium also occur. Process length in L. machaerophorum varies (Lewis and Hallett, 1997; Mertens et al., 2009), with mainly short-processed forms (<6 µm) occurring in the Caspian Sea at the present time (Mertens et al., 2012), thought to be a response to the relatively low salinity. Leroy et al. (2013b) suggested that L. machaerophorum abundances may have increased since c. 3200 cal. BP, possibly in response to rising sea surface temperatures.

Several dinocyst records are available from the offshore Caspian area (Leroy et al., 2007). Cores CP-18 and CP-21 (Figure 1a) are dominated throughout by I. caspienense, with slightly increased amounts of S. cruciformis and L. machaerophorum in the lower to mid-sections, whereas high numbers of P. psilata and S. cruciformis are present prior to c. 4000 cal. BP in core CP-14. Core TM (Leroy et al., 2013a) from the south-eastern coastal region also has abundant I. caspienense in the lower section of the core (10,640–7500 cal. BP) and varied frequencies of S. cruciformis, I. caspienense and L. machaerophorum in the mid to late Holocene intervals. In core GS05 in the offshore area, part of the early Holocene (equivalent to the Mangyshlak local stage) has an assemblage dominated by I. caspienense (Leroy et al., 2013c).

Cores and sampling

A detailed geophysical survey of the study area in the lower Volga delta was carried out in 2006, consisting of shallow seismic surveys along the main river channels (Figure 4d) in the Damchik region (Kroonenberg et al., 2008). Over 70 km of shallow seismic data were collected, focused on the uppermost 10 m of sediment, and the principal seismic facies types used to characterize sedimentation and stacking patterns (Hoogendoorn et al., 2010). A total of 25 shallow cores were collected, also in 2006, with core localities chosen based on the results of the seismic survey, to give representative coverage of all localities and sediment types from all accessible parts of the study area. Cores of up to c. 10 m in length (average 8.7 m) were taken using a piston corer and labelled as D-01 to D-25. Eight cores were chosen for detailed palynological analysis, giving a spread of localities from the proximal (northern) and distal (southern) parts of the study area, with additional spot samples examined for calcareous microfossils. Surface samples for modern pollen and calcareous microfossils, principally ostracods, were also collected from six localities (Figure 4d).

Radiocarbon dating

In all, 44 samples were selected for accelerated mass spectrometry (AMS) radiocarbon dating, with conventional ¹⁴C ages converted into calibrated years BP by Beta Analytic Ltd. using the IntCal09 database (Reimer et al., 2009). The dated material consisted of presumed in situ mollusc shells, and in a few cases ostracod shells and plant or woody material. For the purposes of this present study, no adjustment for reservoir or ‘hardwater’ effects has been made (i.e. ΔR = 0). The lower Volga delta regions, where the dated samples were obtained, are largely freshwater regions. This is substantiated by 42 out of the 44 samples studied having δ¹³C values of −5‰ or less. Radiocarbon ages shown in the text are in calibrated years BP (cal. BP), and the individual ages shown are the ‘intercept’ points with the IntCal09 curve (rounded to the nearest 10 years unless otherwise indicated).

Palynomorphs and calcareous microfossils

In total, 231 samples were analysed palynologically from cores D-07, D-08, D-09, D-17, D-18, D-21, D-22 and D-24, with samples being selected as representative of the lithological units present in each core. The sample coverage averages out at between two and three samples per metre, and is sufficient to give a good representation of the main sedimentary units. All core and surface samples were processed using cold HCl (20%) and cold HF (40%), followed by a second HCl treatment. Residues were then sieved through 10 µm mesh and finally mixed with 1% poly vinyl alcohol before mounting in Norland Optical Adhesive No 63. Counts of all palynomorphs were made, with the aim of obtaining a minimum of 200 pollen grains, excluding unknowns, aquatics and reworked taxa, which make up the pollen sum (PS). Additional quantitative counts were made of non-pollen palynomorphs (NPP), including fern spores, fungal bodies, algae, dinocysts and reworked forms. Most of the pollen types could be identified with reference to sources, including Erdtman (1971) and Moore and Webb (1978). The Russian pollen and spore floras of Kuprianova and Alyoshina (1972, 1978) and Bobrov et al. (1983) were also used. In situ dinocysts were identified with reference to Marret et al. (2004), Mudie et al. (2001), Popescu et al. (2009), Sorrel et al. (2006) and others.

CONISS cluster analysis (Grimm, 1987) was applied to identify pollen and non-pollen zones. Two separate CONISS analyses were made for each section. The first includes only the PS and the clusters obtained are labelled P1–P4 (P = Pollen). The second CONISS analyses were run using a separate percentage sum of the totals of the other groups (aquatics, algae etc.). These clusters are labelled N1–N4 (N = Non-Pollen). Assigned clusters are prefixed with the number of the core (e.g. 7-P1, 7-P2, 8-N2, 8-N3, etc.). Where counts are less than 50, individual occurrences are not shown on percentage diagrams and are excluded from the CONISS analysis.

Samples studied for ostracods and other calcareous microfossils were washed through a 125 µm sieve and the residue dried at 100°C. Individual specimens were picked from the residue and identifications made principally with the aid of Agalarova et al. (1961) and Meisch (2000).

Seismic, lithology and radiocarbon dating

Of the principal seismic facies, distinct channel features (Figure 5a) are visible throughout the delta, notable in the region of the D-08, D-09 and D-24 core localities (survey S9). Heterogeneous sets (Figure 5b) are also widespread and include individual beds of between 10 cm and 1 m thick, interpreted as small-scale channel/levee complexes, highstand delta front deposits and possibly terminal splays occurring, for example, close to the D-22 core locality (survey S3). More evenly bedded parallel sets (Figure 5c) mostly coincide with grey clays and are associated with highstand sedimentation, mainly in the lower delta and avandelta region, for example, at the D-07 core locality (survey S4). There are no direct tie points between the cores and the seismic lines, as the cores were taken on land and the seismic data are from the adjacent fluvial channels. Typical lithofacies include laminated dark grey clays, silty clays and fine sands, which most probably represent mainly avandelta and kultuk sediments, brown massive fine-grained sands indicative of fluvial channels and reddish brown silty clays and fine sands, which occur in the deeper parts of several boreholes and represent the sub-strata of the Holocene delta (Hoogendoorn et al., 2010). Work is currently ongoing to integrate the seismic, lithological and palynological datasets fully. Results of the radiocarbon dating (Table 1 and Figure 6) for the cores studied for palynology give a spread of ages from 9010 (core D-18) to 670 (core D-24) cal. BP.

OPEN IN VIEWER

Figure 5.

Examples of shallow seismic profiles. (a) channel sets, (b) heterogeneous sets and (c) parallel sets (after Hoogendoorn et al., 2010).

OPEN IN VIEWER

Table 1.

Accelerated mass spectrometry (AMS) radiocarbon dates from the Damchik cores in conventional ¹⁴C years and in calibrated years before present (cal. BP) using the IntCal09 calibration curve. Calibrated ages are not corrected for reservoir effect. U = Utrecht, B = Beta Analytic Inc. The 95% probability 2σ ranges are shown as well as the best-fit ‘intercept’ with the IntCal09 curve as calculated by Beta Analytic Inc. (rounded to the nearest 10 years).

Lab. no.	Core	Depth (m)	Material	δ13C	Conventional age		Calibrated ages (IntCal09)		Intercept age
					14C yr BP		2σ range (95%)
B-372497	D-07	7.00–20	Shells	−3.6	7350	±40	7720	7910	7820
B-372498	D-07	7.30–40	Shells	−1	7270	±40	7650	7820	7720
B-246233	D-08	2.80	Shells	−10	3240	±40	3560	3380	3450
U-14787	D-08	2.90	Shells	−9	3020	±60	3370	3000	3230
U-14788	D-09	2.60	Shells	−7.8	3451	±43	3835	3592	3700
U-14789	D-11	0.60	Shells	−11.5	3143	±33	3442	3274	3370
U-14790	D-11	1.00	Shells	−11.3	1511	±30	1508	1340	1390
U-14791	D-11	1.60	Plant	−27.7	1704	±29	1696	1539	1610
U-14792	D-11	2.60	Shells	−8	3212	±34	3478	3371	3420
U-14793	D-11	3.80	Shells	−7.6	3609	±39	4068	3834	3900
U-14794	D-11	5.40	Shells	−7.2	3902	±39	4425	4183	4360
U-14795	D-11	6.00	Shells	−6.6	4274	±37	4869	4825	4850
U-14796	D-11	6.70	Shells	−7.1	4895	±36	5697	5587	5600
U-14797	D-11	7.90	Shells	−5.8	4904	±42	5722	5587	5640
U-14798	D-11	9.00	Shells	−5.9	27900	±200	32,720	31,520	31,910
U-14799	D-12	3.60	Shells	−11.4	2501	±31	2735	2466	2620
U-14800	D-12	8.50	Wood	−29.1	863	±30	899	726	750
U-14801	D-15	5.90	Shells	2.3	41800	±900	46,400	44,100	45,230
B-246234	D-17	1.80	Nut	−9.4	1760	±40	1810	1560	1700
B-246235	D-17	2.70	Nut	−8.8	1890	±40	1920	1720	1850
B-246236	D-17	5.20	Shells	−9.8	3370	±40	3700	3480	3630
B-246237	D-17	5.40	Shells	−7.9	3430	±40	3830	3580	3690
B-246238	D-17	5.68	Shells	−9.1	3600	±40	4060	3830	3900
B-246239	D-17	5.70	Shells	−11.2	3930	±40	4510	4240	4420
B-246240	D-17	6.30	Shells	−8.5	4460	±40	5300	4890	5200
U-14802	D-18	0.80	Shells	−11.3	1624	±28	1563	1416	1530
U-14803	D-18	3.70	Shells	−13.5	3388	±32	3697	3564	3640
U-14804	D-18	4.70	Shells	−9.4	3264	±30	3564	3407	3470
U-14805	D-18	5.10	Shells	−5.5	3638	±49	4089	3836	3940
U-14806	D-18	6.10	Shells	−5.2	8072	±46	9088	8793	9010
B-246241	D-21	1.50	Shells	−10.6	1640	±40	1690	1410	1540
B-246242	D-21	2.00	Shells	−11.5	1770	±40	1820	1570	1700
U-14807	D-22	1.00	Shells	−10	3316	±34	3636	3461	3560
U-14808	D-22	1.60	Shells	−12.5	3547	±35	3923	3721	3840
U-14809	D-22	3.80	Shells	−10.5	3423	±33	3820	3586	3660
U-14810	D-22	4.20	Shells	−7.5	3827	±36	4407	4094	4190
U-14811	D-22	5.20	Shells	−8.7	4253	±36	4862	4728	4840
U-14812	D-22	7.00	Shells	−2.5	7287	±44	8182	8007	8120
U-14813	D-24	1.00	Shells	−11.3	1706	±51	1718	1522	1610
U-14814	D-24	2.20	Shells	−11.7	1889	±36	1918	1727	1850
U-14815	D-24	4.70	Plant	−27.8	886	±60	929	685	790
U-14816	D-24	5.70	Plant	−27.6	772	±40	739	665	690
U-14817	D-24	7.60	Plant	−27.1	1064	±91	1177	787	960
U-14818	D-24	8.95	Plant	−27.5	707	±35	688	567	670

OPEN IN VIEWER

Figure 6.

Age–depth plots of radiocarbon ages in calibrated years before present (cal. BP) for selected samples/cores. Data points for cores D-08, D-09, D-12 and D-15 are not shown for reasons of diagram clarity.

Surface samples

Palynological data from the nine surface samples are shown in Figure 7, with ostracod and other microfossil results displayed in Table 2. More or less equal proportions of arboreal pollen (AP) and non-arboreal pollen (NAP) occur across the study region, with highest NAP in the most distal locality (AVD3). AP is represented by abundant Salix and consistently common Pinus, Alnus, Betula, Quercus and several others. NAP components are made up predominantly by Poaceae and Amaranthaceae, which show very similar distributions and abundances, becoming slightly more frequent towards the north (e.g. D-15 core locality). Artemisia is also well represented throughout. Aquatic pollen is present in all samples with increased Sparganium to the north (D-15 core locality) and Potamogeton towards the south (AVD3). Reworked palynomorphs of various ages occur in every sample but are not especially common. Of the NPP, fungal bodies (mainly hyphae) are most frequent in the most landward sites (D-18 and D-15 core localities) that are barren of microfauna. Dinocysts occur in low numbers in all samples but have slightly higher abundances in the southern part of the study area (at and to the south of the D-09 core locality) in the lower delta and avandelta. These include good specimens of S. cruciformis, P. cruciformis, I. caspienense and C. rugosum. A minor peak in representation of I. caspienense occurs at the surface at the AVD2 locality, whereas a sample from 0.20 m below water level contains abundant Pediastrum. Sample AVD3 (from 1.50 m water depth) contains frequent Poaceae pollen, mixed aquatic taxa and low numbers of foraminiferal test linings. Ostracod diversity is richest in the more distal samples; AVD2 (ilmen) sample yielded abundant carapaces of Cypridopsis vidua together with a few specimens of Pseudocandona (mainly juveniles). Specimens of Cyprideis torosa, Bacuniella dorsoarcuata and Candona were rare and probably reworked. Sample AVD2 (0 m) contained a single specimen of Limnocythere inopinata and a juvenile Pseudocandona. Also present but probably reworked were C. torosa (common), Leptocythere quinquetuberculata and Caspiolla. Sample AVD2 (0.20 m) yielded single specimens of L. inopinata, C. vidua, C. torosa and rare Callistocythere, as well as several reworked specimens. Sample AVD3 also contained a single C. vidua and several undifferentiated gastropods. Charophyte gyrogonites were recorded in the AVD2 (0.20 m) and AVD3 samples.

OPEN IN VIEWER

Figure 7.

Palynological analysis chart for surface samples collected in the Damchik region. Data for AP and NAP (upper chart) are presented as a percentage of the pollen sum (PS). Data for aquatics and NPP (lower chart) are presented as a percentage of the PS + individual groups. Filled circles denote occurrences of less than 1%. For sample locations see Figure 4d.

OPEN IN VIEWER

Table 2.

Results of ostracod and calcareous microfossil analyses from surface and core samples.

		Ostracods																	Other microfossils
		Freshwater									Brackish		Caspian Sea
		Candona neglecta	Candona spp.	Cypridopsis vidua	Darwinula stevensoni	Iliocypris bradyi	Iliocypris spp.	Limnocythere inopinata	Limnocytherina sanctipatricii	Pseudocandona spp.	Cyprideis torosa	Cytherissa bogatschovi	Bacuniella dorsoarcuata	Callistocythere spp.	Caspiolla spp.	Leptocythere quinquetuberculata	Leptocythere spp.	Ostracods (undifferentiated)	Charophytes	Dreissena spp.	Bivalves (undifferentiated)	Gastropod operculum	Gastropods (undifferentiated)
Surface samples	Water depth (m)
D-15	0
D-18	0
D-09	0						P											P
D-07	0		R (?)																				P
D-07 (ilmen)	0		R (?)					P										R
AVD2	0							P		P	C (rw)				P (rw)	P (rw)				R (rw)	A (rw)
AVD2	0.20			P	P (rw)			P			P			R				A (rw)	P	P (rw)
AVD2 (ilmen)	0		R (rw)	A						R	R (rw)		P (rw)										R
AVD3	1.50			P															P			P	R
Core samples	Core depth (m)
D-17	1.80
D-17	2.70
D-17	5.20–5.30		R		C	R
D-17	5.40		P		P
D-17	5.68		R		P
D-17	5.70	R (?)	A		R	C			R					P
D-17	6.30–6.40		R		R
D-08	2.80–2.90					P (rw)					R (rw)	P (rw)				R	R
D-07	7.00–7.20		R	R	P			R			A							R		R			C
D-07	7.20–7.40		R		A			C		P	P							C	P	P			R
D-21	1.50–1.60				P						A			P			C		P
D-21	2.00					P					A						R

P: present; R: rare; C: common; A: abundant; (rw): reworked; (?): questionable specimens.

Core samples

Detailed descriptions of the palynological assemblages recorded in all of the cores studied are provided in the Supplementary Information section, available online, including complete pollen diagrams (Figures S1–S15, available online), lithologies and correlations (Figures S16–S19, available online). Results of ostracod and other microfossil analyses are shown in Table 2.

The majority of the core samples contain rich recoveries of palynomorphs, including mixed assemblages of pollen, spores, algae and dinocysts. The pollen spectra include several types of AP, including Pinus, Quercus, Ulmus, Tilia, Carpinus, Betula, Alnus and Salix. NAP also occurs frequently and includes Poaceae, Artemisia, other Asteraceae and Amaranthaceae. Aquatic pollen from submerged (e.g. Potamogeton) and emergent forms (e.g. Sparganium and Typha) is locally frequent. Algae include Pediastrum and Botryococcus and dinocysts (e.g. I. caspienense, S. cruciformis and P. cruciformis) occur in low but consistent numbers. P. cruciformis has not been previously documented from the Caspian Sea, being originally described from the Pleistocene of the Black Sea (Rochon et al., 2002). Also, as reported by Sorrel et al. (2006) from the Aral Sea, specimens similar to Romanodinium areolatum were occasionally observed, and these are included with the counts of S. cruciformis. Some of these also closely resemble Galeacysta etrusca, originally described from the late Miocene (Messinian) of the Mediterranean Sea (Corradini and Biffi, 1988) and also known from the early Pliocene of the Pannonian and Dacic Basins (Popescu et al., 2009; Plates 1–3). Fungal bodies are locally frequent. Reworked components include Neogene and Paleogene forms (mainly dinocysts) and Mesozoic and Paleozoic pollen, spores and microplankton.

OPEN IN VIEWER

Plate 1.

Photomicrographs of dinoflagellate cysts from the Damchik locality (surface samples and cores). The scale bar = 20 µm (all photos). (a–c) Spiniferites cruciformis (Wall et al., 1973), AVD2 (ilmen). (d–f) S. cruciformis (Wall et al., 1973), Core D-07 (6.90–7.00 m). (g–i) S. cruciformis (Wall et al., 1973), Core D-07 2.90–3.00 m. (j–l) S. cruciformis (Wall et al., 1973), extreme morphotype, similar to Galeacysta etrusca complex (see Popescu et al., 2009), Core D-24 (7.70–7.30 m).

OPEN IN VIEWER

Plate 2.

Photomicrographs of marine palynomorphs from the Damchik locality (surface samples and cores). The scale bar = 20 µm (all photos). (a–c) indeterminate granular dinocyst, D-07 locality, (surface sample). (d and h) indeterminate granular dinocyst, D-15 locality (surface sample). (e) microforaminiferal test lining, AVD3 locality, ‘avandelta’ mud (1.50 m water depth). (f and g) indeterminate granular dinocyst, D-15 locality (surface sample). (i) and (j) indeterminate granular dinocyst, AVD2 locality (0.20 m water depth). The ‘indeterminate granular cysts’ have varied ornamentation ranging from even ‘granules’ (e.g. (i and j)) to small ‘spines’ (e.g. (d and h)). They appear to have an apical archaeopyle, often with attached plates or operculum, similar to Batiacasphaera and the ‘baggy cysts’ of Traverse (1978). (k–m) Impagidinium caspienense (Marret et al., 2004) AVD2 locality (surface sample). (n–p) I. caspienense (Marret et al., 2004), morphotype similar to the freshwater form Gonyaulax apiculata, and also spiny forms of Spiniferites inaequalis (Wall et al., 1973) Core D-24 (0.40–0.60 m). (r–t) Caspidinium rugosum (Marret et al., 2004), small specimen, Core D-24 (3.10–3.50 m). (q and u) C. rugosum (Marret et al., 2004), fragment of typical, larger specimen, Core D-24 (1.70–2.00 m).

OPEN IN VIEWER

Plate 3.

Photomicrographs of dinoflagellate cysts from the Damchik locality (surface samples and cores). The scale bar = 20 µm (all photos). (a–c) Pterocysta cruciformis (Rochon et al., 2002), AVD2 locality (surface sample). (d–f) P. cruciformis (Rochon et al., 2002), Core D-21 (3.50–3.60 m). (g–i) Spiniferites cruciformis (Wall et al., 1973), end-member morphotypes; (g) from Core D-24 (7.70–7.30 m) shows some similarity with Thalassiphora balcanica (Baltes, 1971) and (h and i) from Core D-24 (5.90–6.30 m and 7.00–7.30 m) with Galeacysta etrusca complex (see Popescu et al., 2009).

Representative summary diagrams are shown in Figure 8a–e for pollen and Figure 9a–e for aquatics and NPP. Cores D-07 (Figure 8a) and D-17 (Figure 8b) show good examples of typical pollen distributions, with frequent NAP in the lower sections of the cores and increased AP in the upper parts. Core D-18 (Figure 8c) shows a similar pattern but is almost barren of pollen in the lowermost section, whereas core D-22 (Figure 8d) has a more complete pollen record. The assemblages in core D-24 (Figure 8e) show little significant change within the studied section. Changes in aquatic pollen and NPP assemblages, including reworked elements, are visible in cores D-07 (Figure 9a) and D-09 (Figure 9b). Core D-17 (Figure 9c) shows the presence of frequent freshwater algae (Pediastrum) together with increased numbers of dinocysts, reworked forms, fungal hyphae and aquatic pollen in the upper sections. Core D-18 (Figure 9d) shows similar overall trends, with the additional presence of frequent fungal spores and localized Pediastrum in the lowermost section (where pollen was virtually absent). The aquatic and NPP record from core D-24 (Figure 9e) again shows little overall change throughout.

OPEN IN VIEWER

Figure 8.

Simplified percentage pollen diagrams for selected cores. Only selected taxa are shown. Data are presented as percentage of the PS. QUTC = the sum of Quercus, Ulmus, Tilia and Carpinus (after Bolikhovskaya, 1990). Radiocarbon ages in parentheses are considered unreliable. (a) Core D-07, (b) Core D-17, (c) Core D-18, (d) Core D-22 and (e) Core D-24. U: unassigned. For core locations see Figure 4d.

OPEN IN VIEWER

Figure 9.

Simplified percentage diagrams for selected cores showing aquatics, dinocysts and other NPP. Only selected taxa are shown. Data are presented as percentage of the PS + individual groups. Dinocyst occurrences are shown with a 10× exaggeration curve. Radiocarbon dates in parentheses are considered unreliable. (a) Core D-07, (b) Core D-09, (c) Core D-17, (d) Core D-18 and (e) Core D-24. U: unassigned. For core locations see Figure 4d.

Radiocarbon ages

The age–depth profiles (Figure 6) show a good deal of consistency between most of the radiocarbon ages from the cored sections. The ages tend to fall in clusters, which might be an indication that gaps in sedimentation have occurred, at least in some parts of the delta. These clusters are 9010–7720 cal. BP, 5640–3230 cal. BP and 1850–670 cal. BP and correspond with the main phases of active delta deposition during the Holocene, as originally shown by Hoogendoorn et al. (2010). Outliers from these main clusters do occur due to reworking and/or sediment inversion. Sedimentation rates will differ between individual cores due to variations in lithology, but have an average of between 1.00 and 1.5 mm/yr, similar to the rate of 1.3 mm/yr estimated by Overeem et al. (2003). The most useful ¹⁴C ages are from cores D-07, D-17, D-18, D-22 and also D-11 (not yet studied for palynology). The age–depth plot for core D-24 shows clear evidence for sediment inversion and reworking, which is consistent with the regional setting and palynological and lithological data.

Surface samples

Pollen data from the surface samples (Figure 7) are broadly comparable with the local and regional vegetation in that proportions of AP and NAP are fairly equally represented. High proportions of Salix pollen are due to the sample locations, most of which are on the river-margin levees where willow forest commonly grows. Sample AVD3 was taken away from the willow forest and has the lowest percentages of Salix pollen, as might be expected as Salix is not a large pollen-producer and the pollen is dispersed both by wind and insects. The remaining AP content is made up of mixed tree taxa such as Alnus, Betula, Carpinus, Quercus, Tilia and Ulmus. Of these, only Quercus and Ulmus grow in the present-day delta (Losev et al. 1998) in the mixed forest areas (see Figure 4c). The pollen from Alnus, Betula, Carpinus and Tilia is, therefore, presumably sourced from outside of the delta region. These are widely dispersed types and also occur in pollen rain and sea-bed samples in the north Caspian region (Vronsky, 1976). Pinus accounts for less than 10% of the AP component in the Damchik surface samples and could be wind-blown or water-transported from within the Volga catchment. NAP components, in particular Poaceae, Artemisia and Amaranthaceae, generally occur at frequencies of around 20% or less of the PS in the surface samples, and these are also likely to be sourced from herbaceous associations within the delta itself, or wind-blown from the adjacent steppe and semi-desert areas. The highest proportions of Poaceae (c. 40% of the PS) occur in AVD3, and much of this is probably derived from reedbeds, where Phragmites is abundant. Pollen components within the surface samples are therefore thought to be representing pollen sourced mainly from the local area (i.e. the delta) and the adjacent steppe and semi-desert vegetation, but with part of the AP component resulting from long-distance transport. Aquatic pollen and spores also reflect the local delta vegetation in that Sparganium (emergent) is common in the most northerly (drier) locality (D-15), whereas Potamogeton and Salvinia (submerged/floating) are most frequent in the most distal locality (AVD3), where a single occurrence of Nelumbo pollen was also recorded. The parent plant, although very frequent in the lower delta, is insect pollinated and therefore under-represented in pollen profiles.

Dinocyst records provide some new information about the geographic ranges and perhaps environmental tolerances of several taxa. I. caspienense, C. rugosum, S. cruciformis and P. cruciformis all occur in low numbers in samples taken from the partially flooded areas of the lower delta (to the south of core D-09 locality), suggesting that this part of the delta may be influenced by occasional brackish incursions, or that the dinocysts observed are tolerant of extremely low salinity to fully freshwater conditions. An unknown form (a granular cyst with an apical archaeopyle; Plate 2) occurs further north still (at the core D-15 locality) and may, therefore, be a freshwater cyst. Rare occurrences of L. machaerophorum were noted in the mid-part of the study area (core D-07 and D-09 localities), where salinity is almost zero. This species is tolerant of a wide variety of salinities (e.g. Mertens et al., 2012), but it is unlikely to occur naturally in fully freshwater conditions, and these occurrences may be due to reworking, brackish incursions or wind-blown water surges. A peak representation of Pediastrum (at AVD2 0.20 m) suggests a likely ilmen origin for the sample. Fresh-looking foraminiferal test linings occur in the most distal sample (AVD3) with a probable reworked occurrence in the AVD2 (ilmen) sample, and suggest slightly increasing salinity in the avandelta region.

The overall species composition of pollen and dinocysts from the surface samples is similar to that obtained from many of the core samples studied, in particular those deposited since around 4000 BP, suggesting that the pollen and dinocyst spectra obtained, both in core and at the surface, are giving accurate representations of the delta vegetation and deltaic environments. The findings are comparable with previous modern pollen studies from the north Caspian region, such as Vronsky (1976) and Bolikhovskaya and Kasimov (2008). There is no clear evidence that a large proportion of the pollen in the surface samples, except probably Pinus and low numbers of several other AP taxa, for example, Carpinus, have been carried over very large distances by wind or water. In general, the results are comparable with the findings of Cambon et al. (1997) in the Rhône delta in that the pollen in the surface samples includes mixed wind dispersed (atmospheric) and fluvially transported components, depending on the sample locations. Low numbers of reworked forms do occur in the Volga delta surface samples, which must be sourced from the Russian hinterland and transported by water.

The Damchik surface samples also contained varying proportions of in situ and reworked ostracods. Autochthonous specimens were fresh in appearance, often preserved as carapaces and some retained soft body parts. Abraded specimens were interpreted as allochthonous, either transported penecontemporaneously from adjacent environments or reworked from older Quaternary deposits. Pseudocandona, Limnocythere, Candona, Darwinula and Iliocypris are all freshwater taxa and found at various locations within the studied area. C. vidua, which occurs in several samples and is abundant in the AVD2 (ilmen) sample, typically inhabits well-oxygenated, permanent water bodies rich in vegetation (Meisch, 2000), whereas charophyte gyrogonites indicate proximity to shallow freshwater or brackish pools at the AVD2 (0.20 m) and AVD3 localities. The presence of C. torosa and Callistocythere in AVD2 (0.20 m) also suggests a degree of brackish influence. Taxa such as Leptocythere, Caspiolla, Bacuniella and Loxoconcha (present in several of the AVD2 and AVD3 samples) are all typical Caspian Sea taxa with brackish-marine affinities and are most probably reworked. Samples from the most landward locations, close to cores D-15 and D-18, were devoid of any microfossil content. The ostracod assemblages in the surface samples are comparable with the palynological observations, and indicate mainly freshwater conditions in the delta region with minor brackish influences in the most distal lower delta and avandelta localities.

Core samples

The CONISS zonations based on the pollen and NPP records from the eight studied cores indicate that at least four principal P zones (P1, P2, P3 and P4) and N zones (N1, N2, N3 and N4) can be recognized. P2 can be further sub-divided into sub-zones P2a and P2b, and P4 can be sub-divided into sub-zones P4a and P4b. Sub-zones N1a and N1b are also recognized. As the P and N zonations are based on separate CONISS sums, it is to be expected that the zone boundaries are not always coincident. The P zones give an indication of changes in the land vegetation, whereas the N zones illustrate environmental changes in the delta and aquatic realm. In most instances, however, the P and N zones occur in a consistent manner, and the principal boundaries usually coincide with lithological changes and are further constrained by radiocarbon ages. Additional information from ostracod and calcareous microfossil studies is included within the N zone descriptions.

Comparisons with the Volga–Akhtuba floodplain region

The palynological records obtained from the Damchik cores can be broadly correlated to the palynological vegetation and climate scheme (Bolikhovskaya, 1990; Bolikhovskaya and Kasimov, 2010) for the lower Volga at the Solenoe Zajmishche locality in the Volga–Akhtuba floodplain region, to the north (Figure 10). The period 11,500–10,350 cal. BP at Solenoe Zajmishche was characterized by frequent Pinus, Picea and Abies pollen, suggesting forest and forest-steppe vegetation and a relatively cool and humid climate, and conditions at Damchik were probably broadly similar. Increased numbers of Amaranthaceae pollen, and subsequently tree taxa such as Carpinus, Tilia and Ulmus in the period 10,350–8900 cal. BP suggest that steppe and forest-steppe were present at the Solenoe Zajmishche locality, and the climate was predominantly dry. Incised valleys and dune fields (Baer hills) were present at Damchik during the Mangyshlak lowstand at around the same time. Maxima of Amaranthaceae pollen suggest cold and dry conditions at Solenoe Zajmishche between 8400 and 8240 cal. BP, and a similar palynological event occurs in Damchik in core D-22 (8.30–7.30 m, sub-zone P2a) occurring just below a horizon with a radiocarbon age of 8120 cal. BP. A comparable event in core D-07 (7.40–7.00 m) produced slightly younger ¹⁴C ages of 7720 and 7820 cal. BP. It is possible, although unproven, that the Amaranthaceae event at Solenoe Zajmishche and in core D-22 at Damchik may be a response to the so-called ‘8200 BP cold event’ which affected much of Europe (Magny et al., 2003; Thomas et al., 2007) and has also been observed in the western Black Sea region (Filipova-Marinova et al., 2013). Pollen evidence from the other cores (e.g. D-07) suggests that dry, but not necessarily cold, conditions continued until at least 7720 cal. BP.

OPEN IN VIEWER

Figure 10.

Holocene stages, vegetation and climate in the lower Volga region (after Bolikhovskaya, 1990; Bolikhovskaya and Kasimov, 2010; Richards and Bolikhovskaya, 2010) and correlation to delta phases and pollen (P) and non-pollen (N) zones in the present study.

Warm climatic events during the mid-Holocene period 6970–4040 cal. BP are evident at Solenoe Zajmishche where increased numbers of AP, including Carpinus, Ulmus, Fagus, Tilia, Betula and various conifers indicate that forest-steppe and mixed forest were widespread, whereas NAP elements remained dominant at Damchik (sub-zone P2b), suggesting that steppe vegetation was prevalent close to the delta region. Increased numbers of NAP, including Amaranthaceae, Artemisia and Ephedra occurred at Solenoe Zajmishche in the period 4040–3770 cal. BP which is thought to reflect the expansion of dry steppe and semi-desert vegetation in response to cold and dry climatic conditions (Bolikhovskaya, 1990; Bolikhovskaya and Kasimov, 2010). An equivalent event is not clearly seen in the Damchik pollen records but may be marked by a condensed section or gap in sedimentation: the inferred timing is also close to the P2/P3 zone boundary. The period from 3770 cal. BP to the present-day at Solenoe Zajmishche shows fluctuating but gradually falling proportions of AP suggesting changes in vegetation from forest-steppe to steppe, then from dry steppe to semi-desert, indicating a trend towards generally drier climates (Bolikhovskaya, 1990; Bolikhovskaya and Kasimov, 2010). This is broadly comparable with Damchik in that increased AP (in zone P3) indicates the presence of forest or woodland communities in the delta region. Frequencies of NAP then show further increases (in sub-zone P4a), indicating steppe or semi-desert vegetation and presumably drier climatic conditions leading up to the Derbent lowstand period. Subsequently, aquatic pollen records indicate landward expansion of avandelta vegetation (in zone N4). These events probably occurred at a time when the Solenoe Zajmishche locality was subject to alternating warm/humid and cold/dry climatic conditions, with steppe and dry steppe vegetation present.

Ages for the phases of delta deposition observed at Damchik are based on radiocarbon dates from the individual sections studied, although some gaps in the radiocarbon record remain. A more complete picture can be obtained from the inferred correlation with the regional vegetation and climate scheme of Bolikhovskaya (1990) and Bolikhovskaya and Kasimov (2010). Using this correlation, Delta Phase 1 is likely to represent the period of approximately 11,500–8900 cal. BP, Delta Phase 2 approximately 8900–3770 cal. BP, Delta Phase 3 approximately 3770–2080 cal. BP and Delta Phase 4 approximately 2080 cal. BP to the present-day (Figure 10).

Caspian Sea level

Conventional (i.e. uncalibrated) radiocarbon dates, lithological and palynological data from the present study have been used by Kroonenberg et al. (2008) and Hoogendoorn et al. (2010) to produce a new partial sea level reconstruction, which also incorporates lowstand data from the Kura delta (Hoogendoorn et al., 2005) and highstand data from Dagestan (Kroonenberg et al., 2007). This is generally in agreement with other published studies (e.g. Rychagov, 1997), in that a major lowstand episode (the Mangyshlak event) occurred in the early Holocene before c. 9000 BP. Subsequently, Caspian Sea levels were variable in the early to mid-Holocene, but with a rising trend (at least in the northern Caspian Sea), reaching a peak at around 2600 BP (Kroonenberg et al., 2007), coincident with the beginning of the humid Subatlantic period in Western Europe (van Geel and Renssen, 1998). The palynological results from the Damchik cores (this study) are in broad agreement with these findings, although it is clear that there are regional differences during the Holocene and these remain only partially explained.

For the period 11,500–8900 cal. BP (Delta Phase 1, zones P1 and N1), palynological data from Damchik are mostly consistent with relatively low Caspian Sea levels, with the dominant assemblages indicative of erosional and sub-aerial conditions, as would have occurred during the Mangyshlak lowstand. Following the relative lowstand, a small rise in Caspian Sea level, perhaps in association with wind-blown surges, would have easily flooded over the delta front areas, and probably occurred at the onset of Delta Phase 2 (zone P2 and N2) at around 8900 cal. BP, suggested the presence of low numbers of brackish dinocysts, for example in core D-07 (Figure 9a) and core D-18 (Figure 9d). A cold and dry period (sub-zone P2a) then occurred, possibly related to the ‘8200 year cold event’, which may have coincided with a fall in Caspian Sea level interpreted by Varushchenko et al. (1987) at around the same time. Pollen composition in the remainder of Delta Phase 2 is dominated by NAP with freshwater algae (Pediastrum) occurring frequently and dinocysts rarely; these are consistent with generally high Caspian Sea levels and kultuk-ilmen deposition. The high levels of NAP are likely to have been sourced from steppe and reedbeds. Dinocyst evidence from the central and southern Caspian Sea regions also suggests that early to mid-Holocene Caspian Sea levels were mainly high, with highest levels being reached prior to c. 4000 cal. BP, coinciding with maximum abundances of the dinocysts P. psilata and I. caspienense, caused by increased glacial melt-water from the Pamir Mountains via the Uzboy and Amu-Darya river systems (Leroy et al., 2007, 2013a). The end of Delta Phase 2 at Damchik occurs at more or less the same time as a cold and dry episode interpreted at around 4000 cal. BP at Solenoe Zajmishche (Bolikhovskaya and Kasimov, 2010). This event is not discernible in the Damchik pollen records, but it may be represented as a hiatus or break in sedimentation (between zones P2/N2 and P3/N3).

The onset of Delta Phase 3, at around 3770 cal. BP, is represented in the palynological records by increased AP, increased reworking and more frequent dinocysts (especially I. caspienense). It is likely that this was a period with variable Caspian Sea levels; warm and humid conditions in the catchment (Bolikhovskaya and Kasimov, 2010) would have enabled significant Volga runoff, leading initially to a Caspian Sea rise, followed by a period of fall, with an outbuilding of the delta allowing colonization of the delta plain by arboreal species. There is no positive evidence that the ‘2600 BP highstand’ (Kroonenberg et al., 2007) is represented in the delta palynological records; it could be represented by one of the peaks in aquatic (avandelta) taxa (e.g. in core D-08 at 2.10 m), but there is no specific age information to confirm this.

The period from 2080 cal. BP to the present-day (Delta Phase 4) is also likely to have been a period of varying Caspian Sea levels, with an initial rise indicated by increased aquatic pollen and freshwater algae in several cores after c. 2080 cal. BP, followed by increased NAP and fungal bodies associated with the Derbent lowstand event, which is well documented from around the northern Caspian region. Caspian Sea level is thought to have dropped to at least 3 m (Rychagov, 1997; Rychagov et al., 2010) or perhaps 20 m or more (Kroonenberg et al., 2008) below its present level at this time. This event is clearly evident in the palynological records from core D-24 where an incised channel succession is dated at between 1180 and 570 cal. BP, more or less time-equivalent to the ‘Medieval Warm Period’ (Kroonenberg et al., 2007, 2008). A major increase in aquatic pollen is evident in all of the studied cores towards the top of zone N4, indicating landward expansion of the avandelta associations once again. This is likely to have occurred within the last few hundred years and probably represents recent rises in Caspian Sea level coincident with the ‘Little Ice Age’ (Badukova and Kalashnikov, 2010; Kroonenberg et al., 2007) and also observed by Leroy et al. (2011) and Naderi Beni et al. (2013b) in the southern Caspian region.

Comparison with the Mio-Pliocene palaeo-Volga delta

The modern Volga delta is a partial analogue for the Mio-Pliocene palaeo-Volga delta (Abdullayev et al., 2012; Kroonenberg et al., 2005), which contains large oil and gas reserves in the central and south Caspian basins, more than 500 km to the south of the present-day delta, in sediments of the Productive Series (Reynolds et al., 1998). The palaeo-Volga was much larger than the present-day delta and deposited mainly clastic sediments up to c. 6000 m or more in thickness over a period of about 2.6 Myr (Abdullayev et al., 2012) in the latest Miocene and Pliocene. Depositional models and generalized stratigraphy for the Productive Series are presented by Abdullayev et al. (2012), Abreu and Nummedal (2007), Baganz et al. (2012), Green et al. (2009), Hinds et al. (2004) and others. During the late Miocene (Pontian), the Caspian Sea was connected to the Para-Tethyan oceans (Reynolds et al., 1998). Subsequently, between 6 and c. 5.3 Myr, a major base-level fall of approximately 1500 m (Abdullayev et al., 2012) resulted in the formation of the palaeo-Volga canyon. This is incised more than 500 m below the present ground level and can be traced almost 2000 km northwards from the mid-Caspian region into the Russian platform (Kroonenberg et al., 2005). The isolated south Caspian basin was then rapidly re-filled with fluvio-lacustrine sediments, initially with Lower Productive Series deposits during the latest Miocene to Pliocene, followed by the Upper Productive Series during the Pliocene. The latter contains the Fasila (Pereriv) and Balakhany units which are the principal oil and gas reservoirs (Abdullayev et al., 2012; Baganz et al., 2012; Green et al., 2009).

Palynological analyses of the Upper Productive Series in core samples from the offshore Azeri–Chirag–Gunashli (ACG) field showed that most of the mud-prone samples studied were of lacustrine origin (Reynolds and Lowe, 2004) and the associated reservoirs are interpreted as fluvial, fluvio-lacustrine and deltaic (Abdullayev et al., 2012; Bowman et al., 2008a, 2008b). Three principal ‘palynofacies types’ can be recognized: ‘Delta Front’, ‘Caspian Lake’ and ‘Lake Margin’ (Figure 11). ‘Delta Front’ associations are characterized by frequent Pinus and Pediastrum, and represent phases when catchment climate was relatively humid, and palaeo-Volga discharge relatively high (Figure 11a). The Pediastrum algae may be sourced from the catchment, ilmen and from the shallow, largely fresh waters at the delta front. ‘Caspian Lake’ associations (Figure 11b) contain increased pollen from mixed and swamp forest, locally common Amaranthaceae and consistent numbers of freshwater to brackish dinocysts, most often S. cruciformis and C. rugosum, and show some similarities with the kultuk in the modern delta. ‘Lake Margin’ associations (Figure 11c) are characterized by localized increases in fungal hyphae or spores, together with increased NAP and, in some instances, dinocysts. These ‘Lake Margin’ assemblages relate to periods of minimal runoff, with sub-aerial exposure of lake-margin muds and expansion of halophytic and xerophytic plant associations within the delta plain regions. The three palynofacies types described do not adequately account for all the palynological and depositional variations observed in the offshore area, and they are most directly applicable to the Fasila and lower Balakhany units. These ‘palynofacies types’ have sedimentary significance in that they broadly mirror the features and processes observed in the modern Volga delta, for example, by Belevich (1956, 1960, 1963), Rusakov (1990) and Kroonenberg et al. (1997; Figure 12). ‘Delta Front’ palynofacies in the Mio-Pliocene delta are associated with amalgamated to isolated channel sands, minor siltstone and mudstone which are the main hydrocarbon reservoirs in the Productive Series. They were deposited in shallow water, not more than tens of metres according to Reynolds and Lowe (2004), at the delta front (equivalent to the lower delta and avandelta in the present-day delta). ‘Caspian Lake’ palynofacies in the Mio-Pliocene delta are associated with finely laminated mudstones, siltstones and minor sandstone, which typically have low reservoir quality. They were deposited at the distal delta front to pro-delta (equivalent to the distal avandelta in the present-day delta); ‘Lake Margin’ palynofacies extended over very large areas of the basin during the latest Miocene and Pliocene in response to extremely low river runoff (Richards, unpublished). They vary in thickness by only a few millimetres to tens of metres, consisting mainly of fine-grained sediments with little or no reservoir quality. There is no exact analogue for the ‘Lake Margin’ palynofacies in the modern delta, although similar palynomorph assemblages (with frequent fungal bodies) occur in several of the Damchik cores associated with the early Holocene Mangyshlak and late Holocene Derbent lowstand events. True deltaic influences are diminished in the upper part of the Upper Productive Series in the Sabunchy and Surakhany units; outcrops are predominantly of fluvial origin (Vincent et al., 2010) with ‘playa lake’ sediments and thick evaporites present in the offshore area (Abdullayev et al., 2012). Potential analogues for these Upper Productive Series units include Lake Eyre, Australia, and some of the isolated lakes in central Asia, for example, the Karakiya depression, Kazakhstan (Richards, unpublished).

OPEN IN VIEWER

Figure 11.

Simplified depositional/palynofacies models for the Mio-Pliocene palaeo-Volga delta in the south Caspian basin. (a) ‘Delta Front’, (b) ‘Caspian Lake’ and (c) ‘Lake Margin’. Reproduced with permission from BP Exploration (Caspian Sea) Ltd.

OPEN IN VIEWER

Figure 12.

Schematic diagram of typical sediment distribution in the Mio-Pliocene palaeo-Volga delta at the ACG locality in the south Caspian basin showing analogies with sedimentary features in the modern delta and ‘palynofacies types’. The inset panels are schematic cross-sections based on typical well sections within the Fasila and Balakhany units of the Upper Productive Series. High net-to-gross sediments have high sand content, low net-to-gross sediments have low sand content. Reproduced with permission from BP Exploration (Caspian Sea) Ltd.

Both the Mio-Pliocene and modern deltas were deposited in closed lacustrine basins, not directly affected by marine influences, with sedimentation controlled by climate, fluvial discharge and base level, and both show evidence for significant variations in Caspian Sea level. The Holocene delta was largely aggradational during rising levels and progradational when levels were falling (Overeem et al., 2003), whereas the Mio-Pliocene delta was predominantly aggradational (Abdullayev et al., 2012) and shows more extremes of deposition as it represents the infilling of an underfilled lake (Bohacs, 2012; Green et al., 2009). Sediments deposited during the early Holocene Mangyshlak lowstand in the modern delta are characterized by largely reworked palynological assemblages, and similar reworking patterns can be seen in several intervals within Productive Series outcrops in Azerbaijan (Richards, unpublished). The modern delta supports largely cool temperate and semi-arid plant associations, whereas the Mio-Pliocene delta shows good pollen evidence for the presence of warm temperate and swamp associations in some parts of the stratigraphic record (Richards, unpublished). In both cases, cooler climates are indicated by a southward expansion of the boreal forest belt, with higher frequencies of coniferous pollen (mainly Pinus) recorded. River margin trees during the Mio-Pliocene included frequent Taxodiaceae, which became extinct in the Caspian region towards the end of the early Pleistocene according to Filippova (1997), although the possibility of relict populations persisting more recently is considered by Sorrel et al. (2007). The niche occupied by Taxodiaceae in the Mio-Pliocene delta is filled by Salix in the modern delta. Steppe, desert and halophytic associations are probably represented by more or less the same pollen types (e.g. Poaceae, Asteraceae, Artemisia and Amaranthaceae) during both the Mio-Pliocene and Holocene, and are indicative of periods of dry climate. Dinocysts (e.g. S. cruciformis and C. rugosum) and algae (e.g. Pediastrum) occur in both the Mio-Pliocene and modern deltas, and are useful for estimating relative brackish and freshwater influences. Fungal bodies are also present in both deltas, and are thought to be broadly indicative of Caspian Sea relative lowstands.