Skip to navigation – Site map

HomeIssues126Case studiesAn Early Bronze Age pile-dwelling...

Case studies

An Early Bronze Age pile-dwelling settlement of discovered in Alepu lagoon (municipality of Sozopol, department of Burgas), Bulgaria

Découverte d’un site palafitte de l’âge du bronze ancien dans la lagune d’Alepou (municipalité de Sozopol, département de Burgas), Bulgarie
Clément Flaux, Pauline Rouchet, Tzvetana Popova, Myriam Sternberg, Frédéric Guibal, Brigitte Talon, Alexandre Baralis, Krastina Panayotova, Christophe Morhange and Atila Vassiliev Riapov
p. 57-70


A new pile-dwelling settlement has been discovered during coring investigations on the shores of the Alepu lagoon (municipality of Sozopol, department of Burgas), on the western Black Sea coast, in Bulgaria. A multi-disciplinary methodology was applied to analyze the archaeological dataset, composed of wood piles, abundant charcoals and wood fragments, seeds, fish and shell remains, a few small bone fragments, some lithic fragments and potsherds. The piles were trimmed from oak trees and sunk into lagoonal muds, and currently lie 5.8 to 6.8 m below mean sea level. It highlights a wooden building at the edge of Alepu palaeo-lagoon. Charcoal remains confirm the use of oak tree as a dominant timber resource, consistent with pollen data for this period. Palaeo-botanic remains highlight gathering activities and the consumption of wild grapes, raspberries and figs. The herbaceous assemblage evokes deforestation activities. Exploitation of coastal resources is well attested by the great density of fish remains, dominated by anchovy (61%), highlighting possible preservation of fish products. Five radiocarbon dates constrain the age of the site to between 3350 and 3000 cal. BC.
The Alepu piles-dwelling settlement sheds new light on the very beginning of the Early Bronze Age in coastal Bulgaria. Adding fresh information to the local archaeological record, it completes the well-discussed issue of the protohistoric submerged settlements, revealing in turn the economic strategies of the societies at the end of the transitional period. Considerations about geomorphological settings of these sites underline the evolution of the regional settlement patterns, as well as the importance of lagoonal locations. Lagoon contexts not only offer abundant fish resources, as attested by our data from Alepu, but also good conditions for anchorage. These sites were later drowned, preserved and buried following the relative rise of the Black Sea level.

Top of page

Full text

The study tooks was supported by the French Archaeological Mission at Apollonia Pontica, funded by the French Foreign Office, the Museum of Louvre and the French National Agency (ANR – project “Pont-Euxin”). Additional funding for radiocarbon datings was obtained from the Artemis program. The authors would like to thank Ingrid Rossignol and Guillaume Soulet for their help during fieldwork. C. Morhange thanks LABEX OT-Med and A*Midex.

1The first observations carried out on the submerged settlements of the Bulgarian Black Sea coast date back to 1921, when such a site was discovered in the ancient ria of Varna (SHKORPIL and SHKORPIL, 1921; PEEV, 2004). In 1927, the dredging of the modern harbour of Sozopol provided new evidence for the existence of protohistoric settlements below current sea level (LAZAROV, 1974). However, it was not until the spread of the submarine research during the 1960’s that these types of sites were systematically studied, and today 18 submerged settlements are known. Among them, the identification of the site of Atiya’s harbour (1968) and Burgas (1971) preceded the discovery of the settlement of Kiten (1973), on the edge of the Urdoviza peninsula (LAZAROV, 1993). All of these sites are attributed to the final phase of the Chalcolithic period or to the Early Bronze Age (PEEV, 2004), at the transitional period attributed by the recent 14C investigations to 3800‑3200 cal. BC (BOJADJIEV, 2007).

2These sites were first described as lacustrine settlements (LAZAROV, 1993). However, the discovery of similar ceramics and wood-pile remains in Atiya bay suggested that most of them were located within marine bays in very close connection with the open sea (TODOROVA, 1995). Furthermore, the depth of the remains, which were found from 4 down to 9 m below the modern water level, led some scholars to theorize a dramatic change in sea level of about 7 m since the Chalcolithic. Numerous multi-disciplinary studies were undertaken from the 1990s onwards, looking to refine the palaeo-environmental context of these coastal settlements from a geo-archaeological perspective (e.g. FILIPOVA-MARINOVA et al., 2013). There is now a general consensus that relative sea-level changes have determined the history of these coastal sites (ANGELOVA and DRAGANOV, 2003; FILIPOVA‑MARINOVA et al., 2011). However, recent controversy about Black Sea-level changes during the Holocene, where a “gradual‑rise” model contrasts with an oscillating one (e.g. BRÜCKNER et al., 2010), means that the relationships between sea-level and pile-dwelling settlements remains an open question.

3Palaeo-geographic studies of known sites are considered to be an important task in assessing coastal areas where new archaeological sites could be discovered (ANGELOVA and DRAGANOV, 2003). By chance, during coring fieldwork undertaken in March 2012 in Alepu lagoon south of Sozopol (Fig. 1), a sequence drilled at the edge of the water body yielded a surprising sedimentary facies: three vertical wood piles, ca. one meter long, were retrieved from the core gouge, with part of their bark still preserved.

Fig. 1 ‑ Geomorphological and geological setting of the Sozopol Coast.

Fig. 1 ‑ Geomorphological and geological setting of the Sozopol Coast.

Location of site SOZ-7. Upper left inset: the Bulgarian coast, simplified geological units after DOGLIONI et al. (1996) and the location of Eneolithic and Early Bronze Age pile-dwelling sites along the Bulgarian Black Sea

After Draganov, 1995; Angelova and Draganov, 2003.

4Above the pile, an assemblage of charcoal, seeds, fish remains, bone fragments and some potsherds encouraged a multi-disciplinary investigation of the site, including analysis of the archaeological remains and their palaeo-environmental context. In spite of the narrow window provided by the coring investigation, the stratigraphy of the site is vertically very well preserved, constrained by a precise chronology of the site supported by seven radiocarbon dates. This study provides a new site to the series of prehistoric pile-settlements identified on the Bulgarian coast and raises questions about the interactions between relative sea level and the geography of ancient coastal settlements.

1 - Geographic setting

5The coast of Sozopol, south of the bay of Burgas, is characterized by an alternation of rocky promontories and bays. Coastal sediments from the Holocene Black Sea transgression have partially filled the embayments (BARALIS et al., 2011). Alepu bay is located south of Sozopol between two promontories – Sveta Agalina and Sveti Toma – the southern one separating the bay from the Ropotamo estuary (Fig. 1). A sand spit, about 3 km long and 100 to 200 m wide, formed between the two promontories, in response to north-south longshore currents. The spit encloses a lagoon at the foot of Medni Rid mountain chain. Alepu lagoon is connected to the sea by a narrow channel north of the spit. A linear dune field is formed in the adlittoral zone, about 5‑10 m high and 50‑150 m long. Along and just below the beach shoreface, a narrow strip of lithified sands has formed, 30‑40 m wide and ending with a subaquaeous cliff at a depth of 2-3 m below mean sea level. Deeper layers of this beach rock were formed at a lower mean sea level (GEORGIEV, 1989).

2 - Methods

6Coring was carried out in Alepu lagoon by means of a vibracorer (Cobra type TT ©), using core diameters of 8 and 6 cm. Because the presence of wood piles made the coring process challenging, a series of four cores was undertaken in order to get the complete stratigraphy of the site, including pre- and post-occupation sediments. This strategy also provided a higher volume of bulk sediments for the occupation layer. Wood piles were sampled during fieldwork. Because the pile diameter was greater than the gouge window, the timber was cut and sampled in 3 cm long pieces and then conserved in a cold room. Samples of bulk sediments 2 cm thick were taken along the whole core sequence, including the archaeological layer. Classic bio-sedimentological analyses, including grain size and the identification of bio-indicators, were performed in order to investigate deposition processes and environments. For comparison, modern analog sediment samples were also taken from the present Alepu coastline, along a profile from the top of the dune to the mid-tidal deposits.

7A total amount of 1.8 liters of bulk sediments were sampled from the archaeological layer, and wet sieved to separate and clean the gravels (> 2 mm), coarse sands (> 0.5 mm), medium sands (> 0.2 mm) and fine sands (> 0.063 mm) fractions. Charcoal and potsherds (> 2 mm), fish remains and seeds (> 0.5 mm) and ostracods (> 0.2 mm) were then picked for further examination.

8The chronological framework was established using seven radiocarbon data, with a special focus on the archaeological level (Tab. 1).

Tab. 1 ‑ Radiocarbon determinations and calibrations

Tab. 1 ‑ Radiocarbon determinations and calibrations

9Dated samples included wood fragments, charcoal and organic macro‑remains. One additional sample was taken from the sapwood of a pile, using the five youngest growth rings, with the exception of the first ring to avoid possible contamination from the preserved bark. All dates were calibrated using IntCal09 calibration curve (REIMER et al., 2009).

3 - Results

3.1 - Stratigraphy and bio-sedimentary analyses

10Cores SOZ-7, SOZ-7 bis, SOZ-7 ter(1) and SOZ-7 ter(2) were undertaken at one site, within a radius of 1 m, at the edge of the Alepu lagoon, behind the sand dunes. The surface was at the lagoon level during coring. Because the lagoon is connected to the Black Sea through a narrow channel, the depth in the cores is plotted in relation to present mean sea level. Figure 2 provides the synthetic stratigraphy of the site, based on cores SOZ-7, SOZ-7 bis, Soz-7 ter(1), Soz-7 ter(2), and figure 3 presents a bio-sedimentological analysis of the sequence. The retrieved sequence was 6.85 m long and comprises two well-defined sedimentary facies (units A and B; Fig. 2 and 3).

Fig. 2 ‑ Synthetic stratigraphy of the site, based on cores SOZ-7, SOZ-7 bis, Soz-7 ter(1), Soz-7 ter(2)

Fig. 2 ‑ Synthetic stratigraphy of the site, based on cores SOZ-7, SOZ-7 bis, Soz-7 ter(1), Soz-7 ter(2)

Fig. 3 ‑ Bio-sedimentology and chronology of the sequence

Fig. 3 ‑ Bio-sedimentology and chronology of the sequence

3.1.1 - Unit A: lagoonal muds

11Unit A comprises a dark grey mud unit, deposited between 6.85 and 4 m below the surface. The facies is largely dominated by silts and clays (90-98 %). The molluscan faunal density is low, a few individuals per 100 g of sediment aggregate, mainly represented by Bittium reticulatum and Rissoella opalina. These two species, that live respectively in subtidal sands and on hard substrates, have probably been reworked. The micro-fauna comprises the ostracod Cyprideis torosa, a euryhaline and eurythermal species that can be found in a variety of different coastal habitats. Here, this monospecific assemblage and the low-energy depositional context demonstrates that unit A characterizes a lagoonal environment, similar to the modern Alepu lagoon lying behind a wide and high sand spit. The absence of other lagoonal ostracods and macrofauna may suggest rapid changes in the ecological conditions of the lagoon, within which only the opportunist Cyprideis torosa was able to develop.

12In SOZ-7, a piece of wood, 27 cm long, was found from 6.51 to 6.78 m below the surface within the lagoonal muds of unit A (sub-unit A1 in Fig. 2). Above, between 5.9 and 5.75 m, an organic rich layer, with sharp upper and lower limits, containing wood fragments, charcoal, seeds, organic macro-remains, fish remains, a few shells, some bone fragments, and a few lithic and sherd fragments, were deposited within the dark mud matrix (sub-unit A2 in Fig. 2). In cores SOZ-7 bis, SOZ-7 ter(1) and SOZ-7 ter(2), three wood piles, ca. 1 m long, were cored and retrieved, respectively, at the following depth ranges: 6.01-6.83 m, 5.8-6.82 and 5.59-6.71 m. At the same depth range, core SOZ-7 recorded lagoonal muds, indicating that the wood piles were fixed within the lagoon bottom. Just above the wood pile, in each of the three cores, an organic rich layer identical to the one retrieved from core SOZ-7 (sub-unit A1), was deposited. It was 14, 19 and 16 cm in thickness, respectively in cores SOZ-7 bis, SOZ-7 ter(1) and SOZ-7 ter(2). The lower and upper limits of these layers were sharp. Detailed analysis of this archaeological assemblage, composed of wood piles and the organic layer is described below.

3.1.2 - Unit B: dune sands

13Unit B is a sand-rich unit (> 99%), 4 m thick. The color ranges between grey to light grey and to yellow from the base towards the top. The grain-size distributions were calculated in order to quantify the sorting of this sand deposition. 27 samples were analysed from unit B and compared with modern sands deposited at the Alepu coast (Fig. 3). All palaeo-sand samples display an identical grain-size distribution. Mean grain size is 0.41 ± 0.04 (2σ) mm and sorting index after FOLK and WARD (1957) ranges between 0.33 and 0.43, indicating well-sorted sediments. By comparison, cumulative frequency curves from surface samples taken across the present dune system of the Alepu coast are superposed on the palaeo-sands from SOZ-7, with a mean grain size and a sorting index range, respectively, of 0.41 ± 0.05 (2σ) and 0.3‑0.47. By contrast, the dune sands in erosion and mid-tidal sands displayed very less well‑sorted sediments, characterized by an increase of the coarse sand fraction (Fig. 3).

14There is a sharp increase in the faunal diversity, from 2‑3 species in unit A to 7 recurrent species in unit B. Species derive from various habitats including subtidal sands, upper clean sands and hard substrates (Fig. 3). This assemblage therefore attests to secondary deposition. Bittium reticulatum, a small gastropod a few millimeters long, largely dominates the faunal group (always > 75 %), probably because of sorting processes which led to the dominant deposition of medium sands in unit B.

15These results clearly emphasize that unit B of core SOZ‑7 was deposited under aeolian dynamics. Unit B is a buried part of the modern dune. The lagoonal unit A, characterized by decantation processes protected from the open sea, was deposited behind a coastal spit which isolated the lagoon from coastal hydrodynamics, like today. The lagoonal muds were then buried by landward growth or the migration of sand dunes.

3.2 - Archaeological remains (Fig. 4)

Fig. 4 ‑ Archaeological remains

Fig. 4 ‑ Archaeological remains

A- Wood pile taken from core SOZ-7 ter(1), with dark muds still covering the wood. The pile was cut into pieces to extract it from the gouge. The real dimensions are 1.02 m long and 4.5-6 cm in diameter. The bottom of the wood pile appeared to be bevelled cut. B- Proportion of seed remains (n=199) and illustrations (1- Rubus fruticosus L., 2- Polygonum aviculare L., 3 and 4- Vitis vinifera ssp. Sylvestris; bar scale 1 mm). C- Proportion of charcoal remains (core SOZ-7, n=40). D- Potsherd, frontal and lateral view. Scale bar = 5 mm; E- Proportion and illustrations of fish remains (142 were identified from a total of 1965 remains). D.1 and D.2: vertebras, usually < 2mm, dominate the best preserved remains; square = 1 mm2

3.2.1 - Wood piles

16Three wood piles, 1 m long, stuck into lagoonal muds, were sampled by coring. The piles have been trimmed from oak trees (Quercus sp.) 4.5 to 6 cm in diameter. In some places, the bark was still in situ, indicating that the wood was used directly after the felling of the trees. Given their small diameter, wood piles were issued from relatively young oaks, which were probably growing in conditions of strong inter-stem competition. These piles come from a pole stand which did not exceed 40 to 50 years.

3.2.2 - Seeds

17A total of 199 seeds were sampled, among which 22 % come from fruits and 78 % from wild herbaceous plants (Fig. 4). Fruit seeds are dominated by wild grape (Vitis vinifera ssp. sylvestris), raspberry (Rubus fruticosus L.) and fig (Ficus carrica L.), consistent with other Neolithic and Bronze Age sites in Bulgaria (POPOVA, 2010). Weed remains are mainly represented by Polygonum aviculare L., Chenopodium album L., and Rumex acetosa L., and to a lesser extent by Euphorbia helioscopia L., Saponaria officinalis L., and Hypericum perforatum L. The assemblage is heterogeneous, taking into account the original habitat of the seeds recovered. For example, Euphorbia helioscopia L. and Hypericum perforatum L. develop upon dry land, rocky and sandy soils, while Saponaria officinalis L. is found on rich soils in meadow environments. The plant assemblage thus translates gathering activities and documents vegetation resources used at the site.

3.3 - Fish remains

181965 fish remains were picked from the archaeological layer, which totaled some 1.8 liters. By comparison with fish-remain densities from Mediterranean Gaul sites of the Iron Age (STERNBERG, 1995), the remains collected from Alepu are nearly a hundred times greater. Amongst the remains collected however, most are fragments impossible to determine, including scales and bones (spines, axonosts, lepidotrichs). The majority are badly damaged and with a size inferior to 2 mm. Only 142 bone fragments have been preserved well enough to allow a taxonomic determination (7 %). These are vertebraes and some parts of the mandibulatory apparatus, with some of the dental and pre-maxillary parts. However, given the preservation status of the remains, determining the species was not possible in most instances. At least six taxa are present: Engraulis encrasicolus (61 %), Mugilidae (5,5 %), Cyprinidae (4 %) , Atherina sp. (2,2 %), Trachurus sp. (1,7 %) and Gobius sp (1 %). The maximal size of Engraulis encrasicolus, Atherina sp. or Gobius sp. is estimated to be 7 cm, thus very small fishes. The other remains have been attributed to larger fish: Mugilidae, Cyprinidae, Trachurus sp. or other unidentified fish that measured more than 15 cm in size (35 % of the remains). All the identified taxa are euryhaline species, not linked to any specific biotope: they are as likely to populate lagoons, deltas, swamps or coasts. None are related exclusively to the preservation status of the remains of marine species.

3.3.1 - Charcoal

19Sixty pieces of centimetric-size and well-preserved charcoal pieces have been analyzed, taken from the rich-organic layer (A2) in core SOZ-7. Some of them show the remains of fungal hyphae, indicating that dead wood was burned. Seven taxa have been identified. The assemblage is dominated by oak tree (Quercus deciduous, Fagaceae; 40 pieces), followed by maple (Acer psp, Sapindaceae; 8 pieces), and hop hornbeam (Carpinus betulus, Betulaceae; five pieces). Minor taxa include two pieces of walnut (Juglans, Juglandaceae), and a single piece from spindle tree (Euonymus, Celastraceae), hawthorn (Crataegus, Rosaceae) and a probable plum tree (Prunus, Rosaceae).

3.3.2 - Pottery

20Several sherds were discovered in the occupation layer of the four corings. Unfortunately, most of them < 1 cm. This poor conservation strongly limits the scope of our observations. Nevertheless, all these fragments share the same feature with a dark brownish colour, more or less homogenous, a coarse clay well baked with many medium-size inclusions of quartz. Furthermore, the surface does not show any remains of polishing. One of the most common sherds belongs to a handmade storage vase (Fig. 4D).

3.4 - Chronology

21Seven radiocarbon datings were performed along the sedimentary sequence, from the base of the lagoon muds to the top of the occupation layer (Tab. 1 and Fig. 3). Two determinations pre-date the archaeological layer, dating back to ca. 3700 cal. BC. The sample Lyon-10877 corresponds to a charcoal discovered at the bottom of the pillar in section SOZ-7 ter (2) at a depth of 670-673 cm. It yielded a date of 4855 ± 30 BP (3634‑3484 cal BC), dating the stratigraphic layer used for the installation of the substructure of the house.

22Dating of the beginning of occupation includes a sample from a pile’s sapwood, and charcoal taken from the base of the organic-rich layer. Three samples were obtained at the top of the same layer to date the end of occupation. Despite the detailed sampling around and within the occupation layer, and although deeper samples provide older uncalibrated radiocarbon ages, calibrated age ranges overlap, because the calibration curve for this period presents a plateau. Taken as a whole, the dataset provides a chronology ranging between 3350 and 3000 cal. BC for the occupation layer (Tab. 1).

23However, it is possible that the occupation, included in this period, was shorter. Both the lower and upper limits of the archaeological layer were sharp (Fig. 3), indicating from a sedimentological point of view, that beginning and end of occupation were not progressive nor discontinuous, but rather rapid and definitive. Taking into account a sedimentation rate of ca. 2 mm year-1 measured in pre‑occupation lagoon muds, and an averaged 160 ± 20 mm thickness of the archaeological layer, this would suggest a maximum occupation time of 80 ± 10 years. Although this is a broad estimation, it means that the lake dwelling site would have been occupied by a few generations, but not three centuries. By comparison, the Early Bronze Age site in Sozopol was occupied for 150-200 years according to ANGELOVA and DRAGANOV (2003). In Kiten, a dendrochronological study has elucidated five stages of construction, ranging between 2778 ± 10 and 2715 ± 10 BC, while radiocarbon dates provide a chronology between 2850 and 2600 yr cal. BC (KUNIHOLM et al., 2007).

24On the basis of the 14C samples, Alepu can be attributed to the very beginning of the Early Bronze Age, according to the new 14C calibration of Y. Bojadjiev who places the end of the Eneolithic along the Struma river at around 3850-3800 BC and the spread of Sitagroi IV around 3200 BC (BOJADJIEV, 2007). It thus appears contemporary with the Dubene-Sarovka IIA, Ezero 13 and Dyadovo 10 in the Upper Plain of Thrace, all attributed to the EBA I (NIKOLOVA and GÖRSDORF, 2002). Alepu is therefore earlier than the other EBA sites discovered close to Sozopol (Atiya, Sozopol 1 and Kiten-Urdoviza) which are all representative of the second phase of the EBA. Sozopol 1 can be assigned to Ezero VI-V, with a possible first layer contemporary to Ezero VIII-IX, according to DRAGANOV (1995). Kiten-Urdoviza seems close to Ezero VIII-V according to ANGELOVA and DRAGANOV (2003).


4.1 - Significance of the bio-archaeological dataset

25None of the biological remains retrieved from the organic-rich layer were found below or above this peculiar strata (A2, Fig. 2), even though lagoonal muds would have provided good preservation conditions. This suggests that biological remains did not sediment there naturally. Trimmed and vertical wood piles, various original habitats of the biological remains, as well as a few potsherds further emphasize the anthropogenic origin of the organic-rich layer. Because this layer, which accumulated on the lagoon bottom, was found just above the wood piles in cores SOZ-7 bis, ter(1) and ter(2), it is likely that only the buried part of the wood piles, up to 1 m deep, was conserved, while submerged and aerial parts were destroyed, as confirmed by numerous wood fragments found in the archaeological layer. Abundant charcoal, millimetric to centimetric in size and angular in form, indicate proximity to their source and probably result from fire activities, such as cooking. The assemblage of mussels, fish and fruit remains strongly suggest food wastes. A few pottery fragments further emphasize cooking or storage activities. Taken as a whole, the archaeological dataset suggests a habitation site. The Alepu pile-dwelling settlement can thus be attributed to the chain of Eneolithic and Early Bronze Age woodwork structures identified along the Bulgarian coast (e.g. DRAGANOV, 1995). Although the aerial structure was not conserved, bio-archaeological remains deposited on the lagoon bottom, probably as waste, provide a clear picture of resource exploitation and use of the surrounding landscape during the EBA I (Fig. 4).

4.1.1 - Gathering resources

26Wood, fruit and wild herbs collected from various original habitats, including oak forest, meadows and coastal rocky and/or sandy dry lands, testify to the importance of gathering activity for inhabitants of the Alepu dwelling site. Wild grape pips are known from several archaeological sites in Bulgaria, such as Malak Preslavetz, Drinovo, Eleshnitsa, Kapitan Dimitrievo, and the Karanovo, Karanovo-II and -III and Orlovets cultures, dating back to Neolithic times (POPOVA, 2010). The first indications of grape cultivation and production of fermented beverage appears during the EBA (VALAMOTI, 2004), although at Alepu, only wild seeds were found. TONKOV et al. (2011) described the presence of Vitis in the form of liana within riverine forests, based on pollen data from the northern Bulgarian coast. Raspberry is also known from Neolithic sites in Greece and Bulgaria. Early fig remains identified in central Bulgaria (Galabovo site, Madretz) date back to the Bronze Age (POPOVA, 1995; POPOVA and BOZILOVA, 1998). The site of Alepu confirms that fig trees were already exploited during the EBA. The dominance of Chenopodium album L. and Poligonum sp. within the wild herbaceous assemblage (Fig. 5) may be related to deforestation activities, because these species generally rely on fields and human habitats. At Sozopol, agricultural imprints are clear although very local according to the pollen record (ANGELOVA, DRAGANOV, 2003).

Fig. 5 ‑ Compiled age-depth plot of coastal muds for the Bulgarian coast

Fig. 5 ‑ Compiled age-depth plot of coastal muds for the Bulgarian coast

1- Bozilova and Tonkov, 1998; Marinova, 2003; 2- Tonkov et al., 2011; 3- this study; 4- Bozilova and Beug, 1992. The age-depths of wood piles are also presented for Sozopol (after Filipova-Marinova et al., 2011 and Kuniholm et al.., 2007), Alepu (this study) and Urdoviza (Kiten; Angelova and Draganov, 2003 and Kuniholm et al., 2007). Relative sea-level envelope reconstructed for the Black Sea based on a glacio-hydro-isostatic model is also depicted after Lambeck et al., 2007. The model does not take into account local land-level movements.

27Pollen, archaeobotanical and osteological material from other EBA coastal sites also suggest the existence of modest livestock farming in addition to fishing, hunting and gathering (FILIPOVA‑MARINOVA, 2006).

4.1.2 - Fish resources

28Identified fish taxa all denote a marginal coastal environment with variable salinity. The presence of Cyprinidae amongst the fish remains further points to a low salinity range. This observation is supported by the monospecific ostracod assemblage, represented by the high density of Cyprideis torosa. In this sense, the fish resources may be provided by Alepu lagoon itself. For this reason, it is not possible to use the ichtyological dataset to resolve whether these remains are the result of a natural phenomenon (thanatocenosis) or human waste. Nonetheless, the absence of whole skeletons and clear anatomical connections would tend to suggest human waste deposition. Indeed, because these remains are associated with wood piles, charcoal, seeds, bones and ceramics, we strongly suggest that fish remains as well as mussel shells concur the exploitation of lagoonal/coastal resources. In this case, the dominant presence of anchovy (Engraulis encrasicolus) is particularly interesting because it suggests interest of the inhabitants for preservation (fish-salting or sauce production) and maybe the trading of fish products. At the EBA II dwelling site of Kiten (Fig. 1), DRAGANOV (1995) and ANGELOVA and DRAGANOV (2003) present hunting and fishing as the dominant economy of the site. Some fish remains were represented by dolphin and tuna, showing that fishing took place, at least partially, in open sea.

4.1.3 - Vegetation resources

29Wood piles from Alepu were trimmed from oak tree (Quercus sp.). Previous investigations at other pile-dwelling sites from the Bulgarian coast also attest to the use of oaks in Kiten (BOZILOVA and FILIPOVA‑MARINOVA, 1994), Durankulak (BOZILOVA and FILIPOVA‑MARINOVA, 1991) and Sozopol (ANGELOVA AND DRAGANOV, 2003; FILIPOVA‑MARINOVA et al., 2011). Charcoal assemblages from Alepu translate the dominant use of oak for fire activities. These data suggest the abundance of oak resources during the EBA. Although the analysed anthracological dataset is insufficient for robust palaeoecological interpretations, the assemblage nonetheless illustrates mesophilous climatic conditions (temperate continental climate). Numerous Holocene pollen studies have been carried out in Bulgaria during the past 30 years. Mid-Holocene pollen data include studies from cores taken from the Black Sea continental slope (FILIPOVA-MARINOVA et al., 2012), the bay of Kiten (BOZILOVA and FILIPOVA-MARINOVA, 1994), Sozopol harbour (FILIPOVA-MARINOVA et al., 2011), and in coastal lakes at Durankulak (BOZILOVA and FILIPOVA, 1991; Bozilova and Tonkov, 1998; MARINOVA, 2003; TONKOV et al., 2013), Bolata (TONKOV et al., 2011) and Arkutino (BOZILOVA and BEUG, 1992), and in the floodplain behind the Gerena coast, north-west of Sozopol (ROSSIGNOL, 2014; Fig. 1). All pollen assemblages studied from coastal sequences display the predominance of deciduous oaks and attest to a mixed oak forest, with Tilia, Ulmus, and Fraxinus, during the last 6000 years. In these sequences, anthropogenic impacts, as attested by the decrease of mixed oak forests together with the appearance of ruderals and cerealia-type taxa, have been evidenced since the Eneolithic on the northern coast of Bulgaria and since the EBA on the southern coast (FILIPOVA‑MARINOVA, 2006). However, the palynological record from the Black Sea continental slope, which gives a regional picture of the vegetation cover, display a homogeneous mixed oaks forest assemblage between 7000 and 600 years cal. BC (FILIPOVA‑MARINOVA et al., 2012). Thus, it appears that anthropogenic impacts are only recorded in pollen diagrams deriving from coastal sites, indicating the local expansion of agriculture and deforestation. The latter could be, at least partially, the result of oak exploitation for dwellings (DRAGANOV, 1995; ANGELOVA and DRAGANOV, 2003; (FILIPOVA‑MARINOVA et al., 2006).

4.2 - Geoarchaeology of EBA coastal settlements

30The new site discovered at Alepu is attributed to EBA I. It completes the previous listing of the contemporary coastal sites of Bulgaria (DRAGANOV, 1995; ANGELOVA and DRAGANOV, 2003; see locations in Fig. 1). The main function of wood piles was to raise the occupation floor above the lagoon’s high-water level. Because these settlements are now submerged sites, sea-level changes constitute the main environmental determinant driving the coastal occupation history (e.g. ANGELOVA and DRAGANOV, 2003; FILIPOVA-MARINOVA et al., 2011). However, these sites, especially for the Chalcolithic, comprise a specific category within the local settlement patterns.

4.2.1 - Evolution of the local settlement pattern

31Despite the partial recording of the settlements along the Bulgarian coast, which does not include several areas, like the foothills of the Medni Rid in the Sozopol region, some common features can be underlined. The local archaeological map around Alepu shows a clear diversity of the settlement pattern during the Chalcolithic. Some sites are located in river estuaries (Akladi Chayri), whilst others lie at the sea front (Sveti Toma), in sheltered bays (Sozopol), on a rocky peninsula (Budjaka) or at some distance from the coast, on a shallow ridge (Garmitsa). At the same time, the metallurgical activities support mining activities in the Medni Rid mountain range (LESHTAKOV, 2010). This area of the Western Black Sea coast resembles the Aegean coast of Thrace, where coastal occupation, especially of its lagoons (Diomeideia, Paradimi, Lafrouda), constitutes an important element within the very diversified local settlement patterns. The latter appear organized around complementary territories, providing the ground for the emergence of a local specialization thanks to the growth of regional exchange networks (BARALIS, 2007). However, the transition from the Chalcolithic to the Bronze Age modified this balance, provoking an important decrease in the number of sites and deep-seated changes in the economic strategies of the local populations. In the Sozopol region, the sites of Garmitsa, Budjaka and Sveti Toma were abandoned, although the frequentation of Akladi Chayri seems to decline. Alepu appears in this context, before the reoccupation of Sozopol and the spread of the Atiya and Urdoviza settlements.

4.2.2 - Settling the coast 6000 years ago

32Pile-settlement sites were not located on the sea front, but rather in protected areas, such as lakeshores behind coastal sand barriers (Durankulak, Shabla, Alepu), in river estuaries (Varna, Burgas, Ropotamo, Kiten) or in sheltered bays (Atiya, Sozopol). At Sozopol, it has been
demonstrated that the tombolo linking the palaeo-island of Skamni to the mainland, was already in place by the 4th millenium BC (MORHANGE 
et al., 2010; BARALIS et al., 2011). The sandy isthmus provided a sheltered bay on the western side of Sozopol, where the Eneolithic and EBA II settlements are located (Fig. 1), protected from the dominant NE and E swell. Moreover, analysis of diatom flora and pollen from the Eneolithic layer attest to freshwater to brackish conditions (ANGELOVA and DRAGANOV, 2003), indicating that the site was located at the outlet of a coastal river.

33In northern Bulgaria, the present geomorphology of Durankulak lake clearly shows that it is an ancient drowned valley (ria) later disconnected from the sea by a sandy barrier (POPOV and MICHEV, 1974). At the base of a core taken from Durankulak lake organic clay was found, dated to 4150 ± 160 cal. yr. BC, and whose pollen record was dominated by Chenopodiaceae (BOZILOVA and TONKOV, 1998). It respectively indicates a protected environment and the proximity of a saline soil, meaning that the area was a lagoon at this time, probably protected behind a sandy barrier which closed, at least partially, the estuary from the open sea. The same morphogenesis can be reconstructed for Shabla, Arkutino (at the estuary of Ropotamo river) and Bolata (ca. 15 km south of Shabla) lakes, in which earlier ages obtained in lake sediments were respectively radiocarbon dated to ca. 5700 (FILIPOVA, 1985), 5000 (BOZILOVA and BEUG, 1992) and 3500 (TONKOV et al., 2011) cal. yr BC. The site at Alepu is no exception because it was located in a lagoonal environment, protected from the open sea behind a sandy spit. The older lagoonal muds recorded at site SOZ-7 were dated to 3620 ± 80 cal. yr BC (Fig. 3).

34Such geomorphological configurations, i.e longshore drift obstruction at river mouths, forming coastal lakes or lagoons behind sand spits, is very frequent along the Bulgarian coast. Our review of older available radiocarbon dates constrains this environment to between 5700 to 3500 cal. yr BC (Fig. 5). By comparison, older sandy barriers in the Danube delta occured between ca. 5000 (CAROZZA et al., 2012) and 3200 (GIOSAN et al., 2006) cal. yr BC. At the world scale, STANLEY and WARNE (1994) demonstrated that deltaic sequences began to accumulate between 6000 and 5000 yr cal. BC, because of the deceleration of sea level rise at this time. Very abundant studies undertaken along the Mediterranean coast also show that fluvio-deltaic progradation has occurred since the mid-Holocene, in the context of sea-level stabilization (e.g. ANTHONY et al., 2014; STEWART and MORHANGE, 2009). In the Nile delta, STANLEY and WARNE (1993) argued that the deceleration in sea-level rise was a key factor in the development of the fertile deltaic plain and subsequent occupation by Predynastic agricultural communities. A similar scenario may be evoked for the Bulgarian coast, where the settlements had spread during the late Eneolithic ca. 4000 cal. yr BC, on the shores of semi-closed embayments, built up in response to the relative sea-level stabilization. DRAGANOV (1995) claimed that late Eneolithic coastal sites were then settled by Eneolithic continental populations which were undergoing pressure from incoming steppe tribes, interpreting these sites as shelter settlements for the last Eneolithic communities.

35In these coastal settings, migrants found abundant gathering, hunting and fishing resources. The density of lagoonal fish remains in Alepu highlight the strong fishing vocation of the site. Settlers also systematically selected the edge of the protected water body rather than the seashore.

4.2.3 - Occupation and relative sea level

36A rise in relative sea level was deduced at Sozopol based on an increase in euryhaline marine dinoflagellate cysts and diatom assemblages (OGNJANOVA-RUMENOVA, 2008) within the final Eneolithic archaeological layers. Also, traces of destruction by fire were visible on a large portion of the woodwork and some pottery (DRAGANOV, 1995). Re-occupation was possible in EBA II using wooden platforms. This occupation hiatus in Sozopol would correspond to the archaeological hiatus known as the ‘Transitional Period’, dated in Bulgaria from 3850‑3200 cal. BC (FILIPOVA-MARINOVA et al., 2011).

37The EBA II settlement would also have been disrupted by both fire destruction and a slow rise in the water table (DRAGANOV, 1995). In Kiten, the EBA II cultural layer contained an important quantity of well-conserved ceramics, because the site was rapidly sealed by shelly sandy muds, 45 cm thick, in response to a sea transgression which also contains disturbed archaeological material (ANGELOVA and DRAGANOV, 2003). In Durankulak, BOZILOVA and TONKOV (1998) also envisage the interruption of the occupation in connection with relative sea-level rise during the EBA.

38Late Quaternary sea-level changes in the Black sea have been discussed by a very abundant scientific literature (YANKO-HOMBACH et al., 2007a). Two ongoing debates have animated the scientific community: one concerns a catastrophic or a smooth reconnection of the Black Sea to the Mediterranean, the second focuses on mid-to-late Holocene Black Sea-level changes, arguing either for a smooth gradual rise, or an oscillating rise, the latter with amplitudes of several meters, during the mid-to-late Holocene. The main criticism regarding the latter hypothesis is based on the widely accepted consensus that the Black Sea and the Mediterranean Sea have been connected since at least the last 7500 years (YANKO‑HOMBACH et al., 2007b), an age recently revised to ca. 9000 cal. yr BP (SOULET et al., 2011). Because the two water bodies have been in equilibrium since this time, oscillations of several meters in the Black Sea do not fit the stabilization of eustatic sea level since 6000 years described in the Mediterranean (POROTOV, 2007; BRÜCKNER et al., 2010; FOUACHE et al., 2012). Rather, given local tectonic movements, compaction processes, the heterogeneity of sea-level markers used, reworking processes, reservoir age, and others local factors, BRÜCKNER et al. (2010) concluded that it is impossible to reconstruct a single relative sea-level curve for the entire Black Sea. Such a conclusion implies that only local relative sea-level curves should be produced to better understand the history of coastal settlements.

39Wood piles were discovered in Kiten at 6.4 m below mean sea level (msl), but their installation occurred at a base level of 6.5 m. Those in Ropotamo estuary and Sozopol harbour were found, respectively, at 5.5 and 5-6 m depth below msl. Wood piles from Alepu lagoon were recorded at a similar depth range, between 5.8 and 6.8 m below msl, but installed at a base level of 5.8 m. By contrast, piles from Varna lakes (Fig. 1), were found at only 3-4 m below msl, and archaeological layers at Durankulak were found on land and within a shallow marsh (DRAGANOV, 1995). As previously stated by ANGELOVA and DRAGANOV (2003), this south-north zonation could be related to tectonic movements of the coast. Indeed, deeper submerged sites are located within the Burgas subsiding basin, delimitated to the north by the Emine normal fault, which would have been active in the Quaternary (DOGLIONI et al., 1996; Fig. 1). However, the relationship between the upper-pile level or archaeological layers with regards to the corresponding sea-level is not clearly defined. In the case of Alepu, the upper plateform was not conserved and the only sea-level marker is the lagoonal bottom, which comprises an imprecise lower boundary of the sea-level envelope. Age-depths of coastal lake mud bottoms for the Bulgarian coast are compiled in figure 5. There is no north-south zonation within the scatter plot. Assuming that mud samples were deposited close to the shore or under shallow water in lagoonal conditions, the scatter plot broadly depicts a maximum coastal lake water column of ca. 3 m which encompasses the predicted sea-level envelope (Fig. 5). As coastal lake bottoms are not significantly lower than contemporaneous predicted sea level, mid-to-late Holocene tectonic movement of the Bulgarian coast cannot be deciphered from the available palaeo-geographic dataset itself. Relative sea-level rise is nevertheless attested at the Bulgarian coast since 6000 years (Fig. 5). Taken into account the higher and lower vertical boundaries of the coastal muds, deposited close to msl or below shallow waters during the Eneolithic to the EBA (Fig. 5), a broad mean relative sea-level rise range can be estimated at 0.4‑1.2 mm. yr-1. By contrast, modern coastal subsidence rates of 7.5 ± 0.4 and 3.6 ± 0.4 mm. yr-1 were respectively measured in Varna and Burgas between 1928 and 1998, using tide-gauge data (PASHOVA, 2002). Although coastal settlements were submerged and buried below coastal sediments, a mean rate for the last 6000 years cannot, however, be compared with the displacement or abandonment of local settlements. In Kiten, five main dendrochronological phases of wood constructions intercalated by periods of wood remplacement during 65-year period were elucidated. This result is important because it shows that maintenance activities were continuous (KUNIHOLM et al., 2007) and that the community could have adapted to slow relative sea-level rise.


40Oak piles sunk into lagoonal muds highlight a wood habitat at the edge of Alepu palaeo-lagoon, which further documents Bulgarian coastal settlements during the very Early Bronze Age. Alepu sheds new light on the economic strategies of an EBA I settlement in a coastal location. Although archaeological materials were recovered from coring only, the high density of charcoal, seeds and fish remains allow us to probe the gathering activities and exploitation of lagoonal resources. The high density of fish remains found in the sediments suggests that the lagoon of Alepu was an attractive environment for settlers. This situation explains that all the EBA coastal sites in this sector of the Western Black Sea area share common features, namely a location at the edge of a protected lagoonal waterbody, rather than at the sea front, and, from a vertical point of view, at sea level, rather than upon an elevated rocky promonotory. A natural lagoonal harbour, providing anchorage facilities and a connection to the sea, may have been another attractive factor. Alepu provides new data on submerged ancient settlements in the Black Sea. Their position along the Bulgarian coast attests to a mean relative sea-level rise broadly estimated at 0.4‑1.2 mm.yr-1, during the last 6000 years.

Top of page


ANGELOVA H., DRAGANOV V., (2003), Archaeological excavations of submerged late Eneolithic and Early Bronze Age settlement, in Kiten and Sozopol (South Bulgarian Black Sea coast), Thracia Pontica, 6, 2, p. 9-22.

ANTHONY E. J., MARRINER N., MORHANGE C., (2014), Human influence and the changing geomorphology of Mediterranean deltas and coasts over the last 6000 years: from progradation to destruction phase ?, Earth-Science Reviews, 139, p. 336-361.

BARALIS A., (2007), Essai de monographie régionale. Habitat et réseaux d’occupation spatiale en Thrace égéenne (fin du Mésolithique-période classique), PhD Thesis, Université d’Aix-Marseille I, 770 p.

BARALIS A., DEVILLERS B., MARRINER N. et al., (2011), Coastal geoarchaeology of Apollonia Pontica (Bulgaria), Méditerranée, 117, p. 103‑109.

BOJADIJEV J., (2007), Absolute Chronology of the Neolithic and Eneolithic Cultures in the Valley of Struma, in TODOROVA H., STEFANOVICH M., Ivanov G. (eds), The Struma / Strymon River Valley in Prehistory, Sofia, p. 309-316.

BOZILOVA E., BEUG H.-J., (1992), On the Holocene history of vegetation in SE Bulgaria (lake Arkutino, Ropotamo region), Vegetation History and Archaeobotany, 1, p. 19-32.

BOZILOVA E., FILIPOVA M. (1991), Palynological and palaeoethnobotanical evidence about the human impact on the vegetation along the Bulgarian Black Sea coast from the Neolithic till the Greek colonisation, in LAZAROV M. et al. (eds), Thracia Pontica IV: Les agglomérations côtières de la Thrace avant la colonisation grecque. Les sites submergés – méthodes de recherches, Actes du IVe symposium international, Sozopol, 6-12 octobre 1988, Sofia, p. 87-96.

BOZILOVA E., FILIPOVA-MARINOVA M., (1994), Palaeological conditions in the area of the prehistoric settlement of Urdoviza near Kiten, in LAZAROV M., ANGELOVA H. (eds), Thracia Pontica V: les ports dans la vie de la Thrace ancienne, Actes du Ve symposium international, Sozopol, 7-12 octobre 1991, Varna, p. 39-50.

BOZILOVA E., TONKOV S., (1998), Towards the vegetation and settlement history of the southern Dobrudza coastal region, north-eastern Bulgaria: a pollen diagram from lake Durankulak, Vegetation History and Archaeobotany, 7, p. 141-148.

BRÜCKNER H., KELTERBAUM D., MARUNCHAK O. et al., (2010), The Holocene sea level story since 7500 BP, lessons from the Eastern Mediterranean, the Black and the Azov Seas, Quaternary International, 225, p. 160-179.

CAROZZA J.-M., MICU C., MIHAIL F. et al., (2012), Landscape change and archaeological settlements in the lower Danube valley and delta from early Neolithic to Chalcolithic time: A review, Quaternary International, 261, p. 21‑31.

DOGLIONI C., BUSATTA C., BOLIS G. et al., (1996), Structural evolution of the eastern Balkans (Bulgaria), Marine and Petroleum Geology, 13, 2, p. 225-251.

DRAGANOV V., (1995), Submerged coastal settlements from the final Eneolithic and the Early Bronze Age in the sea around Sozopol and Urdoviza Bay near Kiten, in BAILEY D.W., PANAYOTOV I. (eds), Prehistoric Bulgaria, Prehistoric Press, Madison, p. 225-242.

FILIPOVA M., (1985), Palaeoecological investigations of Lake Shabla-Ezeretz in North Eastern Bulgaria, Ecologia Mediterranea, 11, p. 148-158.

FILIPOVA-MARINOVA M., (2006), Données palynologiques sur la dynamique de la végétation et les changements climatiques au cours du Quaternaire, Bulgarie, mer Noire, PhD, Université de Sofia, 80 p. (in Bulgarian).

FILIPOVA-MARINOVA M., GIOSAN L., ANGELOVA H. et al., (2011), Palaeoecology of submerged prehistoric settlements in Sozopol harbour, Bulgaria, in BENJAMIN J., BONSALL C., PICKARD C., FISCHER A. (eds), An offprint from Submerged Prehistory, Oxbow Books, p. 230-244.

FILIPOVA-MARINOVA M., PAVLOV D., COOLEN M. et al., (2012), First high-resolution marinopalynological stratigraphy of Late Quaternary sediments from the central part of the Bulgarian Black Sea area, Quaternary International, 293, p. 170-183.

FILIPOVA-MARINOVA M., PAVLOV D., VERGIEV S. et al., (2013), Palaeoecology and geoarchaeology of Varna lake, northeastern Bulgaria, Comptes rendus de l’Académie bulgare des Sciences, 66, 3, p. 377-392.

FOLK R.-L., WARD, W.-C., (1957), Brazos river bar: a study in the significance of grain size, Journal of Sedimentary Petrology, 27, 1, p. 3-26.

FOUACHE É., KELTERBAUM D., BRÜCKNER H. et al., (2012), The late Holocene evolution of the Black Sea – a critical view on the so-called Phanagorian regression, Quaternary International, 266, p. 162-174.

GEORGIEV V. M., (1989), Recent beach rock formation in the Alepou Bay, Bulgarian Black Sea Coast, Comptes rendus de l’Académie bulgare des Sciences, 42, 3, p. 73-76.

GIOSAN L., DONNELLY J. P., CONSTANTINESCU S. et al., (2006), Young Danube delta documents stable Black Sea level since the middle Holocene: morphodynamic, paleogeographic, and archaeological implications, Geology, 34, 9, p. 757-760.

KUNIHOLM P.I., NEWTON M.W., KROMER B., (2007), Dendrochronology of submerged Bulgarian sites, in YANKO‑HOMBACH et al. (eds)The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement. Springer, Dordrecht, The Netherlands, p. 483-488.

LAMBECK K., SIVAN D., PURCELL A., (2007), Timing of the last Mediterranean sea – Black sea connection from isostatic models and regional sea-level data, in Yanko-Hombach et al. (eds), The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement, Springer, Dordrecht, The Netherlands, p. 797-808.

LAZAROV M., (1974), Localités prégrecques sur le littoral de la mer Noire au sud du Balkan, Thracia, 3, p. 107-113.

LAZAROV M., (1993), Les sites submergés le long du Pont ouest dans le contexte de l’histoire pontique et méditerranéenne, Pontica, 26, p. 7-18.

LESHTAKOV P., (2010), Two prehistoric sites on the southern Bulgarian Black Sea coast, in CHOLAKOV I., CHUKALEV K. (eds.), Archaeology in Bulgaria, 2007-2009, American Journal of Archaeology, 114, p. 734-736.

MARINOVA E., (2003), The new pollen core Lake Durankulak-3: a contribution to vegetation history and human impact in Northeastern Bulgaria, in S. TONKOV (Ed), Aspects of Palynology and Palaeoecology, Festschrift in honor of Elissaveta Bozilova, Sofia – Moscow, p. 279-288.

MORHANGE C., DEVILLERS B., MARRINER N., (2010), Études géomorphologiques des rivages d’Apollonia, in Apollonia du Pont (Sozopol), la nécropole de Kalfata (Ve-IIIe s. av. J.-C.), in HERMARY A., PANAYOTOVA K., BARALIS A. et al. (eds.), Apollonia du Pont (Sozopol), La nécropole de Kalfata (Ve-IIIe s. av. J.-C.). Fouilles franco-bulgares (2002-2004), éd. Errance, Aix-en-Provence / Paris, p. 23-30.

NIKOLOVA L., GÖRSDORF J., (2002), New radiocarbon dates from the Balkans (Dubene-Sarovka) : approach to the Early Bronze absolute chronology in Upper Thrace, Radiocarbon, 44, p. 531-540.

OGNJANOVA-RUMENOVA N., (2008), Palaeoenvironment and archaeology: the use of diatom analysis in archaeology, in KOSTOV R. I., GAYDARSKA B., GUROVA M. (eds), Geoarchaeology and Archaeomineralogy, Proceedings of the International Conference, 29-30 october 2008, Sofia, p. 291-294.

PASHOVA L. G., (2002), Investigation of sea level variations at two tide gauges in Bulgaria, Vistas for Geodesy in the New Millennium International Association of Geodesy Symposia, 125, p. 475-480.

PEEV P., (2004), Submerged prehistoric settlements along the Western Black sea coast: the problem of situation, in DOBRZAŃSKA H., JEREM E., KALICKI T. (eds), The Geoarchaeology of River Valleys, Budapest, p. 161-169.

POPOV V., MICHEV K., (1974), Géomorphologie de la côte et du plateau continental bulgares de la mer Noire, Bulgarian Academy of Sciences, Institute of Geography, Sofia, 267 p.

POPOVA T., (1995), Plants remains from Bulgarian Prehistory (7000-2000 B.C.), in BAILY D., PANAJOTOV I. (eds), Prehistory of Bulgaria, Monographs in World Archaeology, 22, Madison, Wisconsin, p. 193-207.

POPOVA T., (2010), Plant environment of man between 6000 and 2000 B.C. in Bulgaria, Oxford, BAR International Series, 2064, 107 p.

POPOVA T., BOZILOVA E., (1998), Palaeocological and palaeobotanical Data from the Bronze Age in Bulgaria, in STEFANOVITCH M., TODOROVA H., HAUPTMANN H. (eds), J. H. Gaul, In memoriam, Sofia, p. 391-399.

POROTOV A., (2007), Relative sea-level changes and submersion of archaeological sites along the northern shoreline of the Black sea, Méditerranée, 108, p. 29-36.

REIMER P. J., BAILLIE M. G. L., BARD E. et al., (2009), IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP, Radiocarbon, 51, 4, p. 1111–1150.

ROSSIGNOL I., (2014), Reconstruction paléoenvironnementale de deux sites portuaires antiques (Orgamè et Apollonia du Pont) sur le littoral occidental de la mer Noire, PhD Thesis, Université Paul Sabatier Toulouse III, 305 p.

SHKORPIL H., SHKORPIL K., (1921), Наколни постройки в Eзерото, Известия на варненското археологическо дружество, 7, p. 79.

SOULET G., MÉNOT G., LERICOLAIS G., BARD E., (2011), A revised calendar age for the last reconnection of the Black Sea to the global ocean, Quaternary Science Reviews, 30, p. 1019-1026.

STANLEY D.-J., WARNE A., (1993), Sea level and initiation of Predynastic culture in the Nile delta, Nature, 263, p. 435-438.

STANLEY D.-J., WARNE A., (1994), Worldwide initiation of Holocene marine deltas by deceleration of sea-level rise, Science, 265, p. 228-231.

STERNBERG M., (1995), La pêche à Lattes dans l’Antiquité à travers l’analyse de l’ichtyofaune, Lattara, 8, 152 p.

STEWART I. S., MORHANGE C., (2009). Coastal geomorphology and sea level changes, in WOODWARD J. (ed.), The Physical Geography of the Mediterranean, Oxford University Press, p. 385-414.

TODOROVA H., (1995), The Neolithic, eneolithic and Transitional period in Bulgarian prehistory, in BAILEY D. W., PANAYOTOV I. (eds), Prehistoric Bulgaria, Prehistoric Press, Madison, p. 79-97.

TONKOV S., BEUG H.-S., BOZILOVA E. et al., (2011), Palaeoecological studies at the Kaliakra area, northeastern Bulgarian Black Sea coast: 6000 years of natural and anthropogenic change, Vegetation History and Archaeobotany, 20, p. 29-40.

TONKOV S., MARINOVA E., FILIPOVA-MARINOVA M. et al., (2013), Holocene palaeoecology and human environmental interactions at the coastal Black Sea Lake Durankulak, northeastern Bulgaria, Quaternary International, 328-329, p. 277–286.

VALAMOTI S., (2004), Η διατροφή στη Βόρεια ελλάδα κατά την προϊστορική περίοδο, με έμφαση στα φυσικά συστατικά της τροφής, Αρχαιολογικό έργο στη Μακεδονία και Θράκη, 18, p. 417-430.

YANKO-HOMBACH V., GILBERT A. S., PANIN N. et al. (eds), (2007a), The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement, Springer, Dordrecht, The Netherlands, 971 p.

YANKO-HOMBACH V., GILBERT A. S., DOLUKHANOV P. M., (2007b), Controversy over the great flood hypothesis in the Black Sea in light of geological, paleontological, and archaeological evidence, Quaternary International, 167-168, p. 91-113.

Top of page

List of illustrations

Title Fig. 1 ‑ Geomorphological and geological setting of the Sozopol Coast.
Caption Location of site SOZ-7. Upper left inset: the Bulgarian coast, simplified geological units after DOGLIONI et al. (1996) and the location of Eneolithic and Early Bronze Age pile-dwelling sites along the Bulgarian Black Sea
File image/jpeg, 240k
Title Tab. 1 ‑ Radiocarbon determinations and calibrations
File image/jpeg, 124k
Title Fig. 2 ‑ Synthetic stratigraphy of the site, based on cores SOZ-7, SOZ-7 bis, Soz-7 ter(1), Soz-7 ter(2)
File image/jpeg, 424k
Title Fig. 3 ‑ Bio-sedimentology and chronology of the sequence
File image/jpeg, 164k
Title Fig. 4 ‑ Archaeological remains
Caption A- Wood pile taken from core SOZ-7 ter(1), with dark muds still covering the wood. The pile was cut into pieces to extract it from the gouge. The real dimensions are 1.02 m long and 4.5-6 cm in diameter. The bottom of the wood pile appeared to be bevelled cut. B- Proportion of seed remains (n=199) and illustrations (1- Rubus fruticosus L., 2- Polygonum aviculare L., 3 and 4- Vitis vinifera ssp. Sylvestris; bar scale 1 mm). C- Proportion of charcoal remains (core SOZ-7, n=40). D- Potsherd, frontal and lateral view. Scale bar = 5 mm; E- Proportion and illustrations of fish remains (142 were identified from a total of 1965 remains). D.1 and D.2: vertebras, usually < 2mm, dominate the best preserved remains; square = 1 mm2
File image/jpeg, 348k
Title Fig. 5 ‑ Compiled age-depth plot of coastal muds for the Bulgarian coast
Caption 1- Bozilova and Tonkov, 1998; Marinova, 2003; 2- Tonkov et al., 2011; 3- this study; 4- Bozilova and Beug, 1992. The age-depths of wood piles are also presented for Sozopol (after Filipova-Marinova et al., 2011 and Kuniholm et al.., 2007), Alepu (this study) and Urdoviza (Kiten; Angelova and Draganov, 2003 and Kuniholm et al., 2007). Relative sea-level envelope reconstructed for the Black Sea based on a glacio-hydro-isostatic model is also depicted after Lambeck et al., 2007. The model does not take into account local land-level movements.
File image/jpeg, 127k
Top of page


Bibliographical reference

Clément Flaux, Pauline Rouchet, Tzvetana Popova, Myriam Sternberg, Frédéric Guibal, Brigitte Talon, Alexandre Baralis, Krastina Panayotova, Christophe Morhange and Atila Vassiliev Riapov, An Early Bronze Age pile-dwelling settlement of discovered in Alepu lagoon (municipality of Sozopol, department of Burgas), BulgariaMéditerranée, 126 | 2016, 57-70.

Electronic reference

Clément Flaux, Pauline Rouchet, Tzvetana Popova, Myriam Sternberg, Frédéric Guibal, Brigitte Talon, Alexandre Baralis, Krastina Panayotova, Christophe Morhange and Atila Vassiliev Riapov, An Early Bronze Age pile-dwelling settlement of discovered in Alepu lagoon (municipality of Sozopol, department of Burgas), BulgariaMéditerranée [Online], 126 | 2016, Online since 01 January 2018, connection on 13 June 2024. URL:; DOI:

Top of page

About the authors

Clément Flaux

CNRS, EcoLab, (Laboratoire d’écologie fonctionnelle et environnement), Toulouse, France

By this author

Pauline Rouchet

Aix Marseille Univ, CEREGE, Europôle méditerranéen de l’Arbois, Aix-en-Provence, France

Tzvetana Popova

National Institute of Archaeology and Museum, Bulgarian Academy of Sciences, Sofia, Bulgaria

Myriam Sternberg

Aix Marseille Univ, CNRS, ministère de la Culture et de la Communication, Inrap, Centre Camille Jullian, Maison méditerranéenne des sciences de l’homme, Aix-en-Provence, France

Frédéric Guibal

Aix Marseille Univ, CNRS, IRD, UAPV, IMBE, Europôle Méditerranéen de l’Arbois, Aix-en-Provence, France

Brigitte Talon

Aix Marseille Univ, CNRS, IRD, UAPV, IMBE, Europôle Méditerranéen de l’Arbois ,Aix-en-Provence, France

Alexandre Baralis

Département des Antiquités grecques, étrusques et romaines, musée du Louvre, Paris, France

By this author

Krastina Panayotova

National Institute of Archaeology and Museum, Bulgarian Academy of Sciences, Sofia, Bulgaria

Christophe Morhange

Aix Marseille Univ, CEREGE, Europôle méditerranéen de l’Arbois, Aix-en-Provence, France

By this author

Atila Vassiliev Riapov

Aix Marseille Univ, CNRS, ministère de la Culture et de la Communication, Inrap, Centre Camille Jullian, Maison méditerranéenne des sciences de l’homme, Aix-en-Provence, France

Top of page


The text and other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search