Navigation – Plan du site

Human activity and landscape change at Adjiyska Vodenitsa, central Bulgaria

Activités humaines et changements environnementaux à Adjiyska Vodenitsa, Bulgarie centrale
Richard C. Chiverrell et Zosia H. Archibald
p. 287-302

Résumés

La découverte d’un site urbanisé datant de l’Age du Fer, Adjiyska Vodenitsa (Thrace antique, Bulgarie centrale), à une distance de 500 m de la rivière Maritsa, pose la question du pourquoi de l’implantation d’un site d’occupation au milieu de terrains exposés aux inondations périodiques et à l’érosion fluviatile. Cette étude vise à présenter une interprétation cohérente de l’histoire des dynamiques environnementales qui puissent expliquer l’existence du site à cet endroit, les activités humaines révélées par les fouilles archéologiques, et éclaircir les relations existant entre les communautés humaines et leurs ressources naturelles. Les découvertes archéologiques (produits de commerce lourds et produits industriels) faites à Adjiyska Vodenitsa montrent que le site correspondait à un port fluvial. L’inscription dite de Pistiros fait référence aux échanges internationaux réalisés à partir du site et confirme l’importance supra-régionale des vestiges fouillés. Une analyse systématique des dynamiques environnementales nous permet d’expliquer les indices apparemment contradictoires offerts par les données archéologiques, tandis que les dates obtenues par le radiocarbone offrent des paramètres pour caractériser les phases principales de la géomorphologie locale à Adjiyska Vodenitsa. La ville fut fondée dans un lieu stratégique à la surface d’un cône alluvial construit par un affluent de la Maritsa ; depuis l’Antiquité, ce cône a été érodé par la migration de la Maritsa. Deux datations 14C nous permettent de situer les changements dynamiques majeurs entre 520-400 av. J.-C. et 1010-1150 ap. J.-C., date vraisemblablement de l’abandon de la ville.

Haut de page

Notes de la rédaction

Article soumis le 19 juin 2009, accepté le 9 novembre 2009

Texte intégral

RC and ZA are grateful to the Deans of the Faculties of Arts and Social and Environmental Studies for their support of this research with Interfaculty Research Development Funding. Two anonymous reviewers and Gilles Arnaud-Fassetta, Editor-in-Chief, are acknowledged for contributions in improving the clarity and structure of this paper.

Introduction

1The foci of this paper are the landscape-human interactions during the Classical and early Hellenistic periods (~600-100 BC) associated with the former river port located at Adjiyska Vodenitsa adjacent to the Maritsa River in southern central Bulgaria. Adjiyska Vodenitsa (discovered in 1988) was an inland urbanised settlement and represented an important commercial and cultural centre in western Thrace (Domaradzki, 1987, 1993 and 2002; fig. 1), and it contradicted the then assumed coastal distribution of urban sites and the supposed exclusively rural character of inland Thrace (see Archibald, 2004). Ancient Thrace extended from the Danube in the north to the Aegean in the south, and was bounded in the east by the Black and Marmara Seas, and the Vardar (Axios) and Velika Morava rivers in the west (fig. 1).

Fig. 1 – General location and setting of the archaeological site of Adjiyska Vodenitsa in the Maritsa basin.
Fig. 1 – Localisation du site archéologique d’Adjiyska Vodenitsa dans le bassin de la Marica.

Fig. 1 – General location and setting of the archaeological site of Adjiyska Vodenitsa in the Maritsa basin.Fig. 1 – Localisation du site archéologique d’Adjiyska Vodenitsa dans le bassin de la Marica.

1: study site; 2: published pollen sites; 3: ice extent at last glacial maximum; 4: rivers mentioned in the text; 5: other rivers; 6: limit of the Maritsa basin; 7: international boundaries.
1 : secteur d’étude ; 2 : sites sur lesquels les données polliniques ont été publiées ; 3 : extension des glaces au Dernier Maximum Glaciaire ; 4 : cours d’eau mentionnés dans le texte ; 5 : autres cours d’eau ; 6 : limite du bassin de la Marica ; 7 : frontières internationales.

2Rare contextual information is provided by a large inscription carved in fine Greek letters, on a huge block of granite. This find, though not in context, since it was reused at a Roman road station 2 km away from the river port, would not have been moved far and refers to emporia (markets) as well as to emporitai, those who inhabit an emporion. The inscription (Chankowski and Domaradzka, 1999; Domaradzka, 2002b) names the city as Pistiros, and the members of its socio-political community (emporitai), as well as citing an earlier decree of Kotys (probably Kotys I of Thrace, who ruled from 383-382 to 359 BC), and further emporia, including Belana and unnamed others. The inscription also names Thracians, Thasians, Maronitans, and people from ‘Apollonia’ (which city of this name is meant here has proven hard to specify), as associated with the emporion called Pistiros, and perhaps these references reflect trading routes (Domaradzka, 2002a). The absence of any significant evidence of human activity contemporary with the inscription in the vicinity of the road station, or of any other nearby setting in which the stone may originally have been displayed, has led to the conclusion that the stone must originally have been displayed at Adjiyska Vodenitsa (Chankowski et al., 2004-2005). Many artefacts, some built features, and other cultural markers point to direct connections between this site and some of the partners referred to in the inscription; notably coins of Thasos and Maroneia, the style of masonry fortifications, and preferences for certain distinctive name forms (Domaradzka and Domaradzki, 1999; Domaradzki, 2000; Archibald, 2002d; Domaradzki, 2002). Though the attribution of the archaeological site at Adjiyska Vodenitsa as the Pistiros referred to in the inscription appears robust, throughout this paper the location is referred to as Adjiyska Vodenitsa.

3The archaeological record at Adjiyska Vodenitsa shows traces of human activity emerging around the beginning of the fifth century BC with the visibility and quantity of material evidence of human activities increasing markedly through this period, advancing to a gradual and progressive decline in the second century BC. The site is located next to the Maritsa River, which rises in the Rila Mountains (2925 m), and is some 150 km from the nearest coast, the Aegean, and separated from it by the Rhodopes mountain range (240 km east-west extent; 2191 m elevation; fig. 1). The current river is a much managed system and has been constrained within extensive gravel embankments to prevent lateral migration of the channel and flood inundation of adjacent agricultural land. Inside these embankments the Maritsa River displays a complex multiple channel planform with numerous gravel islands (braid bars). As recently as the late 1890s the Maritsa River was believed to have been navigable upstream as far as Adjiyska Vodenitsa, and the archaeological evidence suggests a river harbour existed at the emporion (Bouzek, 1996; Archibald 2002c). Thus the Maritsa River was integral to the economy and wealth of Adjiyska Vodenitsa. Artefact evidence shows that the emporion was well connected and served by commercial trade routes; both overland across the Rhodopes range with the Aegean to the south and the Black Sea to the east, and also along the course of the Maritsa River (fig. 1; Archibald, 1998; Domaradzki, 2000). In contrast to the wealth of archaeological research undertaken around Adjiyska Vodenitsa (Bouzek et al., 1996, 2002 and 2007), change in the surrounding landscape has received little attention (Chankowski et al., 2004-2005; Tonkova and Gotsev, 2008). The Maritsa River appears to have regularly threatened the settlement, directly through periodic inundation, and latterly by erosion, which has had the affect of denuding the archaeological record (Domaradzki, 1996 and 2002).

4This paper presents a brief overview of the history of occupation and human activity at Adjiyska Vodenitsa underpinned by 20 years of archaeological investigation at the site, but focuses in particular on evidence for exploitation of the surrounding landscape, water-related commerce and the impacts of flooding. The paper presents the findings of preliminary investigations of the physical landscape that involved a programme of mapping (summer 2006); the fluvial geomorphology; and radiocarbon dating of key sections. The geomorphological analyses benefited from remote sensing (geophysical surveys) and field survey data collected by a Liverpool University field project between 1999 and 2004. The aim was to integrate the archaeological and geomorphological evidence, and to produce a better understanding of the evolution of the landscape surrounding the site and of the linkages between the Maritsa River and the former community at Adjiyska Vodenitsa.

Methodology

5The site at Adjiyska Vodenitsa has attracted the attention of archaeological teams from Bulgaria, the United Kingdom, France and the Czech Republic over the past two decades. Investigation has included a variety of remote sensing techniques (conducted principally by the Bulgarian and British teams), as well as extensive walk-over survey of the region (Bulgarian and French teams), but the bulk of the work carried out has focused on excavation in and around the eastern gateway of the settlement. One of the aims of the British archaeological project (1995; 1999-2006) has been to enhance the investigation of the environment and resource base of the historical community identified at Adjiyska Vodenitsa, so as to provide a broader context for interpreting subsistence and consumption practices represented in excavated features. A programme of remote sensing, using gradiometry and magnetic susceptibility, has revealed a wide range of anthropogenic features inside and particularly beyond the limits of the excavated area (Archibald, 2002b).

6Investigation of the fluvial development has focused on mapping of the geomorphology, description of sections, and radiocarbon dating of organic materials associated with the two elevations of Maritsa River fluvial terraces adjacent to Adjiyska Vodenitsa. Geomorphological mapping was initially based on interpretation of the NASA Shuttle Radar Topographic Mission (SRTM) digital elevation model (DEM), supplied by CGIAR-CSI (http://srtm.csi.cgiar.org/​). The DEM affords a relatively high resolution, spatial interval ~90×90 m and vertically accuracy <15 m. Interpretive geomorphological mapping was undertaken on an ARCGIS (ESRI™) platform digitising close-to horizontal surfaces (potential river terrace fragments) and linear depressions (potential palaeochannel courses). Mapping focused on the Maritsa basin downstream from Belovo (fig. 2). Further mapping (not shown) was undertaken for an alluvial basin of smaller dimensions upstream of Belovo centred on Kostenets (fig. 1). There, this mapping included an assessment of the agricultural hillslopes north of the Maritsa River near Kostenets (not shown). Owing to the often subtle and complex nature of the geomorphology, field survey was undertaken for these areas to ‘ground truth’ and refine the DEM-based mapping, and identify finer detail in the geomorphology (fig. 2).

7Extensive flooding in August 2005 damaged and removed parts of the modern flood protection embankments and exposed lateral sequences of fluvial sediment around Adjiyska Vodenitsa. These sections were described in detail and provided the materials used for radiocarbon dating. Interpretations of past fluvial environmental changes were based on established lithofacies and sediment landform models for fluviatile deposits (e.g., Miall, 1996), with a particular focus on switches between coarse clast-supported sediments indicative of active-channel bed or bar deposition, and intervening fine-grained silt/clays and silty/sand beds, typical of lower energy and overbank inundations processes. The radiocarbon ages were targeted on fine-grained alluvial deposits and soils, and used the humic acid fraction of the soil organic matter. Humic acids are the more mobile fraction in soils and often provide younger age estimates for palaeosols (Matthews, 1993), here we are interested in age for the cessation of soil formation and 14C measurements for the humic acid fraction in the top of the palaeosol provide the best estimate. The radiocarbon measurements were undertaken at the Scottish Universities Reactor Research Centre (SUERC) at East Kilbride. The radiocarbon results are quoted in accordance with the international standard known as the Trondheim convention (Stuiver and Kra, 1986), and the original radiocarbon age determinations are listed in tab. 1, with calibrated age ranges quoted at the 95% (2 σ) confidence interval in calendar years BC and AD calculated using CALIB 5.0.1 (Stuiver et al., 2005). Supporting information on the fluvial geomorphology and sediments of the Maritsa basin around Adjiyska Vodenitsa is available in published borehole records (Baltakov et al., 1996; Kenderova et al., 2002 and 2007), and in the gradiometry and magnetic susceptibility surveys undertaken as part of the archaeological investigations (Ovenden-Wilson, 1999 and 2001; Katevski, 2002).

Fig. 2 – Geomorphology of the upper North Thracian basin around Adjiyska Vodenitsa, also identifying the location of key archaeological sites [base map derived from NASA Shuttle Radar Topographic Mission (SRTM) digital elevation model (DEM), supplied by CGIAR-CSI (http://srtm.csi.cgiar.org/​) SRTM)].
Fig. 2 – Géomorphologie du bassin supérieur de la Marica dans les environs d’Adjiyska Vodenitsa et localisation des sites archéologiques [la carte a été réalisée à partir des données issues du modèle numérique de terrain (MNT) acquis lors de la mission STRM de la NASA, CGIAR-CSI (http ://srtm.csi.cgiar.org/)SRTM)].

Fig. 2 – Geomorphology of the upper North Thracian basin around Adjiyska Vodenitsa, also identifying the location of key archaeological sites [base map derived from NASA Shuttle Radar Topographic Mission (SRTM) digital elevation model (DEM), supplied by CGIAR-CSI (http://srtm.csi.cgiar.org/​) SRTM)].Fig. 2 – Géomorphologie du bassin supérieur de la Marica dans les environs d’Adjiyska Vodenitsa et localisation des sites archéologiques [la carte a été réalisée à partir des données issues du modèle numérique de terrain (MNT) acquis lors de la mission STRM de la NASA, CGIAR-CSI (http ://srtm.csi.cgiar.org/)SRTM)].

1: archaeological sites; 2: study site; 3: extent of fig. 3; 4: alluvial fan; 5: low Maritsa river terrace; 6: high Maritsa river terrace; 7: river bank; 8: slope lines; 9: palaeochannels; 10: sediment exposures; 11: road bridges; 12: contours at 10 m intervals; 13: contours at 100 m intervals.
1 : sites archéologiques ; 2 : secteur étudié ; 3 : délimitation de la fig. 3 ; 4 : cône alluvial ; 5 : basse terrasse de la Marica ; 6 : haute terrasse de la Marica ; 7 : levée de berge ; 8 : lignes de plus grande pente ; 9 : paléochenaux ; 10 : coupes stratigraphiques ; 11 : ponts routiers ; 12 : courbes de niveau (équidistance 10 m) ; 13 : courbes de niveau (équidistance 100 m).  

Tab. 1 – Radiocarbon ages secured for the geomorphology.
Tab. 1 – Datation des données géomorphologiques par le radiocarbone.

Lab code

Materials

Method

13C

14C years

2 sigma cal.
range

SUERC 57112

Humic acid

AMS

-25.3

2405±35 BP

750-395 BC

SUERC 57113

Humic acid

AMS

-24.9

985±35 BP

AD 990-1155

Geomorphological record

Regional geology and geographical setting

8Adjiyska Vodenitsa is located towards the head of a large alluvial basin drained by the Maritsa River (fig. 1), which descends from the Rila and western Sredna Gora mountains into an upper basin centred around Kostenets (altitude 550-525 m) and then passes through a 15 km gorged reach, before entering the extensive and broadly flat North Thracian lowlands (altitude 275-160 m) at Belovo (fig. 2). This extensive lower basin (area of 53 000 km2) drains eastwards past Plovdiv, while receiving flows from the Rhodopes Mountains (2191 m) (to the south) and the parallel Sredna Gora (1604 m) and Stara Planina (2376 m) ranges (to the north). At Edirne (across the Bulgarian border in Turkey), the Maritsa/Evros turns southwest and eventually after ~125 km enters the Aegean Sea (fig. 1; Okay and Okay, 2002). The archaeological site at Adjiyska Vodenitsa is located on the bajada formed by a series of coalescing alluvial fans towards the head of the North Thracian basin (fig. 2). Adjiyska Vodenitsa is on the distal end of an alluvial fan issuing south-eastwards from river systems rising in the foothills of the Sredna Gora Mountains on the north side of the Maritsa basin (fig. 2). On the southern flank of the basin another series of alluvial fans line the mountain front of the Rhodopes range. These north and south-wards extending alluvial fans constrain the dominant west to east flow and alluvial plain of the Maritsa River. The gradient of the main Maritsa floodplain is shallower than the flanking alluvial fans, but it has a relatively steep west to east gradient, 170 m in 5 km and these gradients reduce downstream towards Plovdiv, 17 m over 8 km.

9The terrain in south central Bulgaria reflects the cumulative impacts of tectonic and orogenic activity since the late Mesozoic. The end result has been the production of a series of raised horst blocks (e.g., the Pirin, Rila and Rhodopes massifs) separated by fault-bounded grabens (e.g., the Strouma, Mesta and Maritsa river systems; Zagorchev, 2007). Erosion of the higher region during uplift has led to the uncovering of Palaeozoic basement rock including granite and gneiss. The Maritsa River lies in the 25 km wide Maritsa Fault Zone (25-29 km wide), and is bounded in the south and north by faults, fronting the Rhodopes and Stara Planina ranges respectively (Zagorchev, 2007). The graben may have formed an extensive series of connected lake systems during Pliocene to early Pleistocene times, as lake sediments presumed to be of this age are present in boreholes at depths of ~70 m beneath Septemvri and elsewhere (Baltakov et al., 1996). The sequence of Neogene and Quaternary deposits in this basin rarely exceeds 400 m, and comprises a mixture of lacustrine, alluvial and alluvial fan deposits (Baltakov et al., 1996).

10During the Pleistocene, the Rila, Rhodopes and Stara Planina Mountains experienced repeated glaciations (Velčev, 1995). The most extensive ice-field in the Rila Mountains during the last glacial maximum is thought to have occupied the summit regions (2925 m) and extended down to 1400 m (fig. 1), but the Rhodopes and Stara Planina Mountains were largely unglaciated except for localised small cirque glaciers. Melting of glaciers following the end of the Pleistocene has left a classic alpine terrain of cirque basins and moraines in the Rila mountains. The cold and warm cycles of the Pleistocene have been identified as a principal driver of the staircases of river terraces seen in the major river systems of Bulgaria, including the Maritsa, Mesta and Strouma (fig. 1; Westaway, 2006; Zagorchev, 2007). In settings as tectonically active as the Balkan Peninsula, the sequence of terraces typically reflects the cumulative impacts of both climatically-driven changes in sediment supply and the hydrological regime, and the impact of tectonic movement. During the late Pliocene-Pleistocene, the Maritsa River cut downward c. 200 m through gneiss and marble to form the Momina Klissoura gorge, while the development of terraces was limited to the upstream basin around Kostenets (fig. 1). Downstream of the Momina Klissoura gorge around Adjiyska Vodenitsa a combination of tectonic subsidence and continuous sediment supply have produced a thick basin fill.

11Tectonic control of the Maritsa basin would have been important over Quaternary timescales (Burchfiel et al., 2008), but it is difficult to gauge the absolute vertical movement during the last 3000 years. Extended GPS measurements taken over the past decade are insufficiently precise to detect any real vertical movement. Nevertheless, given the longer term (Pleistocene) trend of tectonic activity, an extensional regime and subsidence of the Maritsa lowlands relative to the flanking mountain ranges is likely to have been an important contributor to change in the fluvial system over Holocene timescales (Zagorchev, 2007). The key drivers of environmental change in the fluvial system over late Holocene timescales are likely to have been the impacts of climate change, extreme events (floods) and anthropic landuse activities on the water and sediment discharge regime.

The fluvial geomorphology at Adjiyska Vodenitsa

12The Maritsa River currently occupies a northerly position on the alluvial plain adjacent to Adjiyska Vodenitsa, and this channel is one of three zones for which there is surface expression for former channels. Adjiyska Vodenitsa itself is currently perched on a bluff (254.7 m) fronting the tributary alluvial fan issuing from the foothills of Sredna Gora at Vetren. The palaeochannels of the Maritsa River are both evident in the field and on the SRTM digital elevation model (fig. 2). In mountain front river systems of this character, with a relatively steep gradient, seasonally high flows and high rates of sediment supply, it is typical for multiple channel systems to exist and for the system to experience episodic avulsion between channels driven by sediment delivery and the surface topography. The two other palaeochannels follow a southerly route and have trimmed the edges of the southern alluvial fans (fig. 2). The age of these palaeochannels is poorly constrained, though Roman and Medieval settlement evidence on the northern outskirts of the town of Septemvri is buried beneath a thick mantle of alluvium and silts (Domaradzki, 1996, fig. 1.18) and testifies to extensive sediment influx and aggradation in the southerly Maritsa palaeochannel during the last 2000 years.

13Our investigations have focused on the current river path, which has trimmed to bluff fronting the Vetren alluvial fan and eroded part of the archaeological record. The archaeological site is 12-14 m above the current Maritsa River and not affected by contemporary flooding, which is intriguing given evidence for flooding of the settlement in the archaeological stratigraphy. The low angle floodplain adjacent to the contemporary Maritsa River contains two distinct levels of fluvial terrace (fig. 2 inset), which at Adjiyska Vodenitsa are separated by approximately 4 m of relief. The higher river terrace is more extensive, covering much of the headward zone of this alluvial basin. The higher terrace (252.4 m) is closer in altitude but distant from Adjiyska Vodenitsa (254.7 m), whereas the lower fluvial terrace (~248 m) abuts against the steep bluff fronting the archaeological site. The bluff was most probably formed by the migration of the Maritsa River after the abandonment of the higher fluvial terrace.

Stratigraphy and geochronology

14Near Adjiyska Vodenitsa, on the north bank of the Maritsa River, ~500 m west of the excavated trenches, a 70 m outcrop of the toe of the Vetren alluvial fan, 12-14 m in height, exposes the fan deposits that underly the settlement (section 2, fig. 3). The sequence comprises a basal unit of laminated lacustrine muds (1 m of which is visible but the base is not seen), which in previous research were mapped as forming the extensive higher elevation terrace extending north from the Maritsa River to the Sredna Gora foothills around Vetren (Chankowski and Fouache, 2000). This interpretation seems now not to be the case. Whilst the North Thracian basin does contain lacustrine deposits, in a fill some 100 m in thickness in places, the upper 10 m of these exposures at Adjiyska Vodenitsa are dominated by two units of gravel, separated by fine-grained alluvium. The clast angularity (rounded to sub-rounded), the overall clast-supported character, and the broadly northwest to southeast fabric (clast imbrication) of these gravels attest to a fluvial origin, most probably lain down as alluvial fans extending southwards off the Sredna Gora. In situ Neolithic remains in the surface layers of these alluvial fans near Adjiyska Vodenitsa suggests that the underlying deposits pre-date 5000-6000 BC and may include deposits lain down during cold episodes of the Pleistocene. D. Katinčarova (2007) has published the contents of a pit, c. 1 m below the surface of the ancient east-west road, dated to Karanovo III-IV/9, i.e., 6th millennium BC, and loose finds of flaked tools and occasional ceramic evidence has been found re-deposited in the upper layers of the excavation trenches, indicating wider spatial distribution of the Neolithic habitation on the surface of the alluvial fan. Preliminary analyses of the clast lithologies support a Sredna Gora provenance, as granitoid and gneiss rock types characteristic of the Rila Mountains are absent (in contrast to the deposits of the Maritsa floodplain terrace).

15On the south bank of the Maritsa River, a 100 m-long section, cuts through two segments of the lower fluvial terrace, as well as the higher fluvial terrace (section 1, fig. 3). The stratigraphy demonstrated that both ends of the higher Maritsa terrace deposits were trimmed, showing erosional contacts with the deposits of the adjacent lower Maritsa terrace. The deposits of the higher terrace comprise two gravel units separated by a 0.3 m thick overbank alluvium. The lower terrace deposits comprise sequences of basal channel gravels overlain by finer grained organic-rich palaeochannel fills, which in turn have locally been buried by further channel gravels. The gravels comprise well mixed assemblages of rock types entirely derived from the Sredna Gora and Rila-Rhodopes upland massifs and are typical of the materials issuing from the Momina Klissoura gorge (Kenderova et al., 2007). These exposures provide an opportunity to secure the history of geomorphic development through the radiocarbon dating of the overbank alluvium within the deposits associated with the higher terrace, and the base of a palaeochannel associated with the lower terrace. Two radiocarbon ages (tab. 1) were obtained and constrain the most recent phase of alluviation associated with the higher terrace to after 2405±35 BP (750-395 BC), and the date of 985±35 BP (AD 990-1155) constrains abandonment of an early palaeochannel associated with the lower terrace. These dates provide terminus post quem and terminus ante quem constraints respectively for the incision to the lower fan terrace and the erosion of the site to after 750-395 BC and before AD 990-1155.

Fig. 3 – Stratigraphy and radiocarbon dating at two exposed sections near the archaeological site at Adjiyska Vodenitsa.
Fig. 3 – Stratigraphie et datation au carbone 14 de deux profils stratigraphiques à proximité du site archéologique à Adjiyska Vodenitsa.

Fig. 3 – Stratigraphy and radiocarbon dating at two exposed sections near the archaeological site at Adjiyska Vodenitsa.Fig. 3 – Stratigraphie et datation au carbone 14 de deux profils stratigraphiques à proximité du site archéologique à Adjiyska Vodenitsa.

1: silty clay; 2: rounded gravel; 3: sand; 4: woody peat; 5: peat; 6: soil layer 1; 7: soil layer 2; 8: angular gravel; 9. lake mud. Inset map shows the general location (outlined on fig. 2).
1 : argiles limoneuses ; 2 : graviers émoussés ; 3 : sables ; 4 : tourbe ligneuse ; 5 : tourbe ; 6 : sol 1 ; 7 : sol 2 ; 8 : graviers anguleux ; 9 : vases de marécage. Localisation générale sur la fig. 2.

16Remote sensing techniques (Geophysical prospecting) have revealed further information about the fluvial environment. Electrical imaging alongside targeted gradiometry on the lower terrace of the Maritsa River immediately adjacent to Adjiyska Vodenitsa (Ovenden-Wilson, 1999 and 2001; Archibald 2002b) shows a series of alluvial channels and longitudinal bars typical of a braided multiple channel system. The gradiometry data typically shows features ~ 1 m in depth with stronger responses interpreted as reflecting near surface gravel islands and softer responses a deeper gravel surface and possible palaeochannels. These interpretations are supported by borehole information for the low Maritsa terrace between Adjiyska Vodenitsa and the current river (Kenderova and Fitova, 2002). Away from what clearly are responses to archaeological remains close to Adjiyska Vodenitsa there are equivalent signals in gradiometry data obtained for the Vetren alluvial fan, on which the site has been built, and it is possible that a similar interpretation should be given to these anomalies (Ovenden-Wilson, 1999 and 2001; Katevski, 2002). The flows responsible for these channel environments extend off the Vetren alluvial fan rather than relating to the Maritsa River. There are no geophysical data for the higher Maritsa terrace to shed light on the character of the alluvial environment associated with those deposits. The limited exposures available for the higher terrace (section 1, fig. 3) and other exposures close to Septemvri show no real difference between the type of deposits associated with the higher and lower terraces. Both are dominated largely by gravel (60-70%) with the intervening fine-grained over bank deposits.

Archaeological record

The city at Adjiyska Vodenitsa

17The excavations at Adjiyska Vodenitsa (fig. 4) have revealed a substantial walled settlement covering a minimum of c. 2500 m2 (0.25 ha), although a substantial area, up to 97%, has been lost to erosion by the Maritsa River (Domaradzki, 1993 and 2002). The discovery of masonry blocks belonging to the northern and western continuation of the fortification wall between 200 m and 400 m west of the eastern gateway provides confirmation of the settlement’s extent in those directions (fig. 4; Bouzek et al., 2007). The area bounded by these northern and eastern limits, along with projection of the former dimensions for where the archaeological features have been removed by the river, suggests a substantial settlement occupying approximately 150,000 m2 (15 ha). Fieldwork included in the international programme of investigations covers a more extensive area including territory outside the eastern gateway, and encountered kilns and other industrial processing areas, as well as residential and other functional quarters, extending into the modern arable field systems to the north and east. The overall proportions and configuration of buildings and activities in the surviving area on the river’s northern bank are consistent with a site much larger than the 0.25 ha available for excavation. A fortification wall 2 m wide, with a tower whose inner dimensions are 6.29 m by 4.75 m (fig. 5), and a neighbouring building complex with dimensions of 18.2 m by 14.35 m are all structures on a scale appropriate for a significant urban unit, comparable to Olynthos in the Chalkidic peninsula (630,000 m2; Cahill, 2002) or the town of Thasos (380,000 m2; ‘la Porte du Silène’, Grandjean and Salviat, 2000).

Fig. 4 – Plan of the excavations at Adjiyska Vodenitsa, showing the eastern gateway, fortification wall and east-west road
Fig. 4 – Plan des fouilles à Adjiyska Vodenitsa, montrant la porte orientale, le mur de l'enceinte et la rue ouest-est

Fig. 4 – Plan of the excavations at Adjiyska Vodenitsa, showing the eastern gateway, fortification wall and east-west roadFig. 4 – Plan des fouilles à Adjiyska Vodenitsa, montrant la porte orientale, le mur de l'enceinte et la rue ouest-est

(after Bouzek et al., 2007; © Pistiros international excavation project).
(d’après Bouzek et al., 2007 ; © Equipe internationale Pistiros).

Fig. 5 – Excavated structures (A, C) showing the fortification wall (A, B), the eastern gateway and road (C), and house (No 1) and the road (D; © Pistiros international excavation project).
Fig. 5 – Structures découvertes lors des fouilles (A, C) montrant le mur de l'enceinte (A, B), la porte orientale et la route (C) et la maison (N° 1) et la rue (D ; © Equipe internationale Pistiros).

Fig. 5 – Excavated structures (A, C) showing the fortification wall (A, B), the eastern gateway and road (C), and house (No 1) and the road (D; © Pistiros international excavation project).Fig. 5 – Structures découvertes lors des fouilles (A, C) montrant le mur de l'enceinte (A, B), la porte orientale et la route (C) et la maison (N° 1) et la rue (D ; © Equipe internationale Pistiros).

18Leaving aside the prehistoric (Neolithic) pits at Adjiyska Vodenitsa, which partly underlay the main west-east roadway (Katinčarova, 2007); occasional finds suggest a measure of activity in the vicinity during the first half of the fifth century BC. The fortification wall, constructed in the third quarter of the fifth century BC (fig. 4), is an ambitious feature, made of two masonry faces of granite slabs, separated by a rubble fill and crowned with a superstructure of rubble and clay, and represents the most significant single investment of time and resources (fig. 5). An external tower at least two storeys high that crowned the eastern gateway was refashioned in the 370s BC and sections of the circuit wall were refurbished, perhaps more than once, in the course of the same century. The integration of a network of clay pipes and stone-lined drains with the fortification wall, via a number of purpose-built conduits, constructed in the base course of the circuit wall, shows that the interior space had been planned as a comprehensive exercise, along with the main street running NE-SW across the settlement (Domaradzki, 2002; Katinčarova, 2007). The international significance of the site at Adjiyska Vodenitsa, and the probable identification as the ancient Pistiros described in the inscription (Chankowski and Domaradzka, 1999), stems from the unusual level of economic activities signified and confirmed by the concentrations of high value pre-Roman coin finds in the vicinity (Domaradzki, 1987 and 1993). Several phases of rebuilding and refurbishment took place during the following century and a half. A major disturbance, which included the conflagration of the eastern gateway, took place c. 300-280 BC. This was evidently the first of a number of crises that led to the destruction of parts of the settlement, involving the concealment of a pot hoard, containing 434 Macedonian copper alloy drachms, 115 silver tetradrachms, and three gold staters, the latest of which series belong c. 280 BC (Bouzek and Musil, 2007).

19Archaeological material dating from Classical antiquity (~600-100 BC), has been found buried below layers of floodplain alluvium and alluvial gravel at a number of locations across the lower Maritsa terrace, extending from immediately adjacent to Adjiyska Vodenitsa (Domaradzki, 1996; Kenderova and Fitova, 2002) to the northern periphery of the town of Septemvri. As excavations of the urban layout have developed, it has become clear that some of the archaeological contexts contained significant amounts of fine-grained overbank alluvium (Domaradzki, 1996 and 2002). This was puzzling for the archaeologists, because either such alluvial materials were used during occupation to build up and structure parts of the site (Archibald, 2002b and 2007); or some at least of these deposits, particularly where artefacts were scarce, might correspond to periodic flooding. If the latter is the correct explanation it would imply that the river was closer in elevation to the site (it is currently some 12-14 m above the bed of the Maritsa River and out of reach from contemporary flooding). Given the flow regime and magnitude of tributary systems, the primary agent of this flooding must have been the Maritsa River, although such an interpretation renders it difficult to understand why the inhabitants would want to reside in such conditions (Domaradzki, 2002). Sand and silt was readily available for reconstruction, which was a periodic necessity in a heavily built up urban centre where available building materials had limited durability (packed clay around wooden frames on stone footings) and were liable to burn down. The presence of at least two significant destruction levels, the first c. 300-280 BC, when the eastern gateway collapsed in an intense blaze and was not rebuilt; the second at some time after 280 BC (Domaradzki, 1996; Kovacheva and Gigov, 1996; Kovacheva et al., 2002), make it all the more likely that mechanical filling of damaged areas of the site may account for some of the presence of alluvium on the site. However, overbank alluvium is also present in archaeological contexts not associated with these destruction layers.

Cultural and economic exchange in the region

20Though the exact dimensions of the of the settlement at Adjiyska Vodenitsa remain uncertain, the intensity and character of human activity in the excavated sector testifies to a level of interaction generated by exchange over two centuries or more of inter-regional commerce. The number of artefacts registered across the site manifestly exceeds the volume generated in a large rural unit of equivalent age on the Crimean coast, and by several orders of magnitude (over 13,700 ceramic finds over 300 m2 at Adjiyska Vodenitsa, compared with 637 ceramic finds in 1190 m2 in a monumental building at Panskoye, near Chersonesos). Adjiyska Vodenitsa has yielded the largest single concentration in the east Balkan peninsula of coins, imported tableware, wine and oil amphorae, as well as a wide range of manufactured objects, representing a range of styles that can be associated with workshops or working methods of regional and extra-regional origin. The value of this type of analysis is constrained in that comparative statistics from cognate urban sites around the eastern Mediterranean remain very unevenly distributed. However it is testimony to the level of trade undertaken at the inland emporion at Adjiyska Vodenitsa, even though other factors probably contribute to the differences with the Panskoye I record, for example 50 years (minimum) commercial exchange via the port city of Chersonesos compared to 80 years of more isolated inland exchange at Adjiyska Vodenitsa.

21Most of the detailed information accumulated so far about the character of activities at the site and in its hinterland is derived from excavated features. The distribution of activities within the area of the circuit wall changed over time, although it is clear that in all investigated areas, there was a high concentration of artefacts in each of the three chronological phases that have been identified. The earliest phase represents the planned organisation of the city as a fortified urban unit, on the site of a previous minor settlement and/or crossing point on the Maritsa River. Apart from the masonry circuit wall, architecture in this phase seems mainly to have consisted of structures made of organic materials, interlaced wattle stakes overlaid with thickly packed clay, and roofed with a timber frame and tiles, which were re-used in subsequent, more substantial buildings on stone footings. In the case of buildings on either side of the main east-west street, these structures had stone walls made of river pebbles packed with clay and masonry quoins, fronted by roofed colonnades supported on wooden columns fixed into stone bases. From outside the city, based on the construction materials and planform of the buildings, Adjiyska Vodenitsa would have resembled many urban sites around the Mediterranean. Behind the colonnaded frontage were shops, store rooms and courtyards. Side streets linked the main road to structures further away from the arterial routes, including workshops and residential units (Archibald, 2002b and 2007; Bouzek and Musil, 2007).

22Adjiyska Vodenitsa was a commercial centre. A wealth of metal finds, including regional and inter-regional coins, weights and measures, scales, and ceramic containers for bulk transportation, demonstrate the significance of trade. However, this was by no means the only kind of activity attested. Smithing is represented by the presence of specialised tools, ingots, moulds, and waste material. Textile manufacture is implied by the widespread presence of loom weights in a variety of designs. The kilns outside the city walls show that ceramics were locally produced as well as imported. Many elements of the built environment as well as portable features reflect the importance of cultural considerations. The low-level decorated baked clay altars, and associated clay and stone sculptures, are the most visible expression of these motives, which were probably iterated on the painted clay walls, of which we only have minor fragments. The generous provision of cooking and tableware, and frequent deposits in underground pits of food refuse, show that eating and drinking played an important part in the social life of the site.

23Anthropogenic engagement with the landscape in and around Adjiyska Vodenitsa is reflected in arable and pastoral activity and in the exploitation of local mineral resources, including clay (there are modern clay pits nearby; Chankowski and Gotzev, 2002), and granite (from Boshulya and perhaps another local source). The sources of copper, silver, iron ore, and other metals worked by smiths on the site, have not yet been located. Arable farming is reflected in the range of plant species identified from a range of local contexts at Adjiyska Vodenitsa (Popova, 1996 and 2002). In contrast to the rarity of organic finds from the most intensively used areas in the centre of the settlement, which may in any case have been swept clean at frequent intervals, a rich series of deposits was examined from the vicinity of a low clay altar and more particularly from a fire place (Popova, 2002). The numerically dominant cereal was wheat, particularly of the mild/compact variety (Triticum aestivo-compactum) followed by barley (Hordeum vulgare), millet (Panicum miliaceum), with some limited evidence of oats (Avena sativa) and rye (Secale cereale). Lentils, bitter vetch, and, occasionally, peas, made up the available legumes, together with fruits, of which the most important was the domesticated vine (Vitis vinifera) and the plum (Prunus sp.). The overall profile of arable cultivation differs in certain respects not only from similarly aged upland sites in the Rhodopes, such as the mountain sanctuary on Mount Babyak, but also from late Iron Age sites in the east Thracian plain, in the vicinity of the modern town of Svilengrad near the Bulgarian-Turkish border, where barley was the dominant cereal (Ignatov et al., 2006; Popova, 2006).

24Preliminary analysis of faunal remains from Adjiyska Vodenitsa shows that cattle represent the most prominent species of domesticated mammal (43-50%) in many contexts, closely followed by sheep and goats (30-69%), with pigs appearing regularly (11-17%; Stallibrass, 2007 and 2008). Dogs and equids (horses and perhaps donkeys or mules), are also represented, together with small numbers of wild species [red deer (Cervus elaphus), hare (Lepus), and a few bones of bear]. Bird bones consist mainly of fowl, probably domestic chickens, but also some large wild fowl, such as cranes and storks, which could have been caught close to the site, as they are today (Stallibrass, 2007). The water-side location of the settlement would have provided suitable pasturage for cattle and the surrounding alluvial plain offers more than adequate grazing opportunities for sheep. The agricultural regime reflected in these plant and animal remains implies extensive planned cultivation (perhaps including surplus production for export downriver) with an integrated pastoral strategy. How such a strategy may have operated in the vicinity of the urban centre, and how agricultural considerations on the one hand, and industrial needs (including mining and smelting, as well as ceramic production) on the other were adjusted, has not yet been determined.

Archaeological survey and the wider region

25The dramatic increase in population levels attested to in many parts of the Mediterranean region and the Near East during the second half of the first millennium BC is also seen in the Balkans (see e.g., Sbonias, 1999). Although intensive surveys, which form the principal method of systematic and large scale spatial demographic analysis, have not yet been applied at this location, a programme of site registration conducted in Bulgaria between 1992 and 1994 provides an alternative indicator of settlement configuration in this period. This survey (Domaradzki, 1996) covered an area of ~25 km2 in the upper reaches of the north Thracian basin, and provided the basis for a more detailed survey conducted by a Franco-Bulgarian team between 1999 and 2003 (Chankowski and Gotzev 2002; Chankowski et al., 2004-2005). These surveys (locations identified on fig. 2) show that in the immediately vicinity of Adjiyska Vodenitsa a small number of sites has yielded evidence of early prehistoric (Neolithic) activity, e.g., at the village of Lozanitsa (Chankowski et al., 2004-2005), and at Adjiyska Vodenitsa itself (Katinčarova, 2007). Late Bronze Age or Early Iron Age activity has been recorded at Izvora, south east of Belovo (Chankowski et al., 2004-2005), and near Semchinovo, in the foothills of the northern Rhodopes (Chankowski et al., 2004-2005). The settlement mound of Yunatsite (early Bronze Age to late Iron Age), on the eastern periphery of this area, is the most significant prehistoric settlement in the region. Thus activity prior to c. 500 BC has been registered at a number of other locations on the alluvial fan south and east of Vetren, and in the northern Rhodopes in both the foothills (below 400 m) as well as in the upland zone (e.g., Tsepina and Milevi Skali; Domaradzki, 1996; Tonkova and Gotzev, 2008). The apparent absence of prehistoric and Iron Age sites in a wide area north and south of the modern bed of the Maritsa River has been attributed to burial of settlement evidence beneath a thick mantle of alluvium and silts (Domaradzki, 1996, fig. 1.18). This is best demonstrated by the 2 m deposits of river-borne sediments overlying Roman and Medieval layers on the northern outskirts of the town of Septemvri (Domaradzki, 1996) and testifies to extensive sediment influx and aggradation in the southerly Maritsa palaeochannel during the last 2000 years. At many of the sites discussed above, occupation began or expanded during the Late Iron Age (the second half of the first millennium BC), but the precise chronology is hard to specify, in the absence of closely datable ceramic finds. Nevertheless, from the artefact record at these sites, it seems clear that many settlements in the region arose as a consequence of the foundation of a river port at Adjiyska Vodenitsa.

Human activity and the fluvial system

26The circulation of ceramic products is one of the clearest surviving symptoms of a lively network of connections between communities in the east Balkan region, the Black Sea, and the Aegean in the second half of the first millennium BC (Archibald, 1998 and 2002a; Bouzek et al., 2007). Communications established in early prehistory, shown by finds of Bronze Age metalwork in the region around Adjiyska Vodenitsa, testify to long-established patterns of contact that suggest periodic long-distance connections. However, the integration of overland and riverine routes of communication across the Rhodopes with Adjiyska Vodenitsa seems to have been largely a development of the Late Iron Age (second half of the first millennium BC; Domaradzki et al., 2000). It is notable that the inhabitants of the island of Thasos in the north Aegean were obliged to change economic strategy when their resources in the Thracian mainland opposite the island were forcibly removed by the Athenians in 463 BC. One of the strategies developed by the islanders in response to these losses was the systematic development of the island’s viticulture, and the exportation of wine, which is reflected in widespread finds of Thasian wine amphorae along the rivers of the east Balkans, particularly along the banks of the Maritsa River (Archibald, 1998; Brunet, 2000; Bouzek et al., 2007).

27The evidence from shipwrecks as well as from recipient centres suggests that products from any one manufacturing centre rarely travelled alone. Mixed cargoes were the norm. This is also reflected in the range of wine jars at Adjiyska Vodenitsa, where other production centres, from the Black Sea coast (notably Herakleia Pontika), as well as from other north Aegean centres. Though the statistical evidence is a little unclear, the minimum number of individual specimens originating from Herakleia on the Black Sea appears to exceed the minimum numbers at some recipient centres close to the Pontic coast itself (Bouzek et al., 2007; Tsotzev, 2007). The current state of the statistical evidence from these various centres means that direct comparisons cannot be made, but the role of Adjiyska Vodenitsa as the principal entrepôt in the western part of the Thracian plain is not in dispute. The jars from Chios on the one hand and Herakleia on the Black Sea on the other are examples of the pioneering role of Adjiyska Vodenitsa as a transhipment centre for the surrounding region and beyond (Archibald, 1998; Domaradzki, 2000; Tsotzev, 2007). The commercial role of Thasians in this network of transportation is confirmed by finds of Thasian coins at Adjiyska Vodenitsa, although the overwhelming number of silver and copper alloy issues from an outside city comes from the Thracian Chersonese, with some neighbouring cities also being represented, notably Kardia, Lysimacheia and Parion in the Hellespontine Straits (Domaradzki, 1987; Archibald, 1998; Taneva, 2000). Other north Aegean centres, notably Maroneia, which is a prominent partner in the Pistiros inscription, but also Ainos and Kypsela, located in the lower estuary of the Maritsa River, are not without importance (Taneva, 2000). The commercial significance of cities in the region of the Hellespontine Straits is in part a reflection of chronological changes. This region became much more significant for commercial traffic in the second half of the fourth and early third century BC, as competition for dominance of the Straits increased.

28Many of the formal features of the river port at Adjiyska Vodenitsa reflect the range and variety of its inter-regional connections. The stone architecture is partly inspired by models in the north Aegean, notably Thasos, where granite is one of the types of building stone widely used in urban construction (alongside schists and marble), but the style and choice of features, including details of the fortification walls and drainage channels, are best paralleled in the most ambitious north Aegean coastal cities, such as Amphipolis and Abdera. This suggests that architects and masons were drawn from a wide pool. The cultural life of the settlement reflects its integration within the region of the western Rhodopes (in the style and shapes of handmade pottery), and more broadly within the expectations of Thracian communities in the central plain (the proliferation of decorated altars in domestic and non-domestic contexts; the typology of tools and equipment).

29The quantity and bulky nature of commodities traded in and out from Adjiyska Vodenitsa point to efficient routeways, and the Maritsa River was most likely integral to this commerce alongside overland routeways across the Rhodopes. The fluvial regime posed severe constraints over navigation. Although the mean annual flows recorded (1936-1988) at Plovdiv are ~51.6 m3/s,flow in different seasons varies between 2 and 258 m3/s suggesting that river access to Adjiyska Vodenitsa may have been seasonal, especially given that mean annual discharges of 7.82 m3/s were recorded closer to Adjiyska Vodenitsa at Belovo (fig. 2). These constraints apply even though water abstraction, regulation (reservoirs) and channel management is greater now than would have been the case 2000 years ago. Autumn/winter extreme flows, low flows during the summer months, and winter conditions in the Rhodopes Mountains may all have been barriers to transport of trade goods. The character of the river system also poses limitations over potential use. The contemporary Maritsa River around Adjiyska Vodenitsa is a gravel-bedded multiple channel system. In part this character reflects the managed condition, with the channel migrating in a reach between pronounced gravel embankments. However, geomorphological, geophysical and sedimentological evidence suggest that the Maritsa alluvial plain has displayed muliple channel (braided), gravel dominated and laterally-dynamic characteristics. Similar longitudinal bar-forms to those seen at present and braiding characterise both the higher and lower Maritsa terraces, thus the fluvial system appears to have been broadly similar during the last 2500 years. The persistence of a complex channel planform further limits the potential of the Maritsa River as a trade route, with water discharge perhaps spread across multiple channels compounding problems of navigation at low flow. Perhaps the transport and trade were limited to the seasons with advantageous flow conditions, or 2500 years ago there may have been more stable and dominant main channels even though the system appears to have operated as a braided river.

Drivers of change in the fluvial system

30Prior to the switch from the higher to lower Maritsa terraces the fluvial environment appears to have been gravel dominated and with evidence for a multichannel channel habit. Between 750-395 BC and AD 990-1155 the geomorphology of the Maritsa River at Adjiyska Vodenitsa changed, with lowering of base level (perhaps accomplished through lateral channel migration) followed by aggradation. This lower terrace phase of aggradation was associated with a dynamic migrating braided channel geometry that contributed substantial erosion of the archaeological remains. Though broadly coincident, from the available data it is not possible to discern whether the basal lowering and erosion took place and threatened the city whilst it was still occupied. This sequence of geomorphological changes could be related to a wide range of potential forcing or conditioning factors. Over Quaternary timescales the Plovdiv basin has been subsiding driven by a regional extensional tectonic regime (Zagorchev, 2007). Recent monitoring suggests this movement continues today, but lacks the sensitivity to detect the vertical displacement over the last 20 years (Burchfiel et al., 2008). However annual rates of subsidence of for example -2 mm/a in front of the Rhodopes Mountains would be sufficient to have created ~2-3 m of vertical accommodation space for sediment in the North Thracian basin, and tectonic activity of this nature has perhaps contributed to the changes seen in a fluvial system issuing from the active fault-zones fronting the Rhodopes and Sredna Gore Mountains. Notwithstanding any tectonic change, it is also possible to produce the geomorphology around Adjiyska Vodenitsa without external forcing, with relatively steep gradient settings like the headward zone of the North Thrace basin very dynamic fluvial systems. The Maritsa River has occupied at least three alternative flow paths on the higher terrace, and sediment accumulation encourage avulsion and occupation of alternate flow paths, particularly in the absence of large scale river management. Thus the phase of fluvial activity associated with the lower terrace and that eroded Adjiyska Vodenitsa might reflect a migration of the dominate flow path 750-395 BC and AD 990-1155 to occupy this northerly channel zone. Channel changes of this nature during the last 2500 years are further evidenced by the Roman and Medieval archaeology buried beneath alluvial gravel in the former middle Maritsa channel zone to the north of Septemvri (Kenderova and Fitova, 2002).

31Climate or anthropic landuse changes are further potential drivers of for the geomorphic changes seen. In terms of sediment supply and capacity, the phase of incision broadly reflects an increase in available power (or reduced sediment supply) and the subsequent aggradation excess sediment delivery. Increased water discharge might expected 600-400 BC, a time period which has been identified across Europe as one of cooling and wetter conditions (van Geel et al., 1996; Chambers et al., 2008). Similar increases in sediment delivery have been identified in river systems elsewhere in the Balkans and Northern Levant, and been attribute to a combination of increasing landuse pressure and climatic deterioration (e.g., Howard et al., 2004; Casana, 2008). The archaeological data from Adjiyska Vodenitsa and montane pollen sites in the wider region show evidence for reductions in woodland cover and increases in arable and pastoral activity testifying to more intensive human activities from around 500 BC. Greater pressure on the landscape would have lowered erosion thresholds encouraging erosion of hillslopes, perhaps gullying, for which there is undated evidence in the low foothills of the Sredna Gora, particularly around Kostenets (fig. 1). Reductions in riparian woodland flanking the river system may have encouraged reduced bank stability and a greater lateral channel change, also. Land pressure releasing new sediment sources upstream of the Momina Klissoura gorge would have contributed increased sediment availability and delivery to the North Thrace basin. At Adjiyska Vodenitsa there is evidence for extensive interaction with the surrounding landscape, which was exploited for arable crops (dominantly wheat, with barley, oats and rye) and for grazing pasture (cattle, sheep, goats and pig). Palaeoecological (pollen and plant macrofossil) data from the lowlands of the North Thracian basin are sadly lacking, with available pollen data from locations above 1400 m and of limited relevance for more the heavily populated lower altitude locations. That said, these studies (Bozilova et al., 1986; Filipovitch, 1995; Filipovitch and Lazarova, 2002; Lazarova, 2003) show evidence for increases in anthropogenic activity manifested in the expansion of arable (cereals) and pastoral indicators (open ground indicators including Poaceae, Plantago lanceolata) alongside changes in the composition of the forest (Abies decline and Fagus expansion in the Rila, Bozilova et al., 1986; Fagus expansion in the Sredna Gora, Filipova-Marinvoa, 1995) from 500 BC onwards.

32The wetter climate conditions 600-400 BC and greater land pressure from 500 BC are both conducive to pattern of geomorphological change seen at Adjiyska Vodenitsa, but discerning the relative importance of these two forcing or conditioning factors is at this stage not possible. Particularly so, given the role tectonic activity and internal dynamics in the fluvial system may have played in the geomorphic development. In summary this is a complex system that was utilised by the habitants of Adjiyska Vodenitsa, and the river almost certainly has had a significant role in the establishment, success, failure and abandonment of the city, as well as destroying much of the archaeological heritage. To better understand the regional context there is a need for palaeoecological data from lowland basins around Kostenets (fig. 1), and in the foothills of the Sredna Gora, and palaeochannels on the alluvial plain for the Maritsa River nearer Adjiyska Vodenitsa. Further geomorhological data are also needed focusing on the prehistory for the fluvial system (Maritsa) around Kostenets upstream of the Momina Klissoura gorge and the hillslope gully systems in the agricultural landscapes of the Sredna Gora foothills.

Conclusions

33The archaeological record shows a strong association between human activity at Adjiyska Vodenitsa and the surrounding landscape. The inhabitants of the city used the landscape for agriculture, cereal cropping and for pasture. These activities appear to coincide with and perhaps contributed to driving changes in the geomorphic regime. Adjiyska Vodenitsa was situated on a high tributary alluvial fan surface, to some extent buffered from direct impact from the Maritsa River, away from perhaps all but the most extreme flows, but sufficiently close to benefit from the trading routes offered by riverine transport. Perhaps during and certainly after human activity at the site, these circumstances changed, with lateral channel migration and erosion of the alluvial fan deposits threatening the site. The timing of these changes in river behaviour is not as yet sufficiently well secured to assess any putative causal association between geomorphic change and abandonment of the site. Artefact evidence shows that the emporion was well connected with the Aegean to the south and the Black Sea to the east, with the Maritsa River an important conduit for commerce and trade. The geomorphology identifies challenges the inhabitants of Adjiyska Vodenitsa would have faced in using the river, particularly the nature of the river system, dynamic, changing and multiple channel, and with flow regimes varying through the seasons from little or no flow in summer months, to high magnitude floods in snow-melt and winter months. These factors can be combined to suggest that unless the geomorphic regime was radically different, e.g., a single channel more stable system, trade and river use may have been strongly seasonal. At present there is insufficient data on the precise nature of the channel system 2500-2000 years ago to further this question, save that the fluvial regime appears to become more dynamic during the period 750-395 BC and AD 990-1155.

34Assessment of the character and past behaviour of the Maritsa River at Adjiyska Vodenitsa is only part of the story. Catchments operate more or less as a sediment cascade, with materials transmitted between sediment sinks of varying durations, and the location in the mountain front zone at the head of the North Thracian basin renders the fluvial system sensitive to delivery of materials through the Momina Klissoura gorge from further upstream. Around Kostenets the rural low-upland agricultural landscape displays evidence for hillslope erosion (gullying), though for the most part stabilised owing to the declines in the intensity of agriculture in the post-communist era. Soils developed over alluvial fan surfaces in these gullies are immature suggesting recent activity (the last ~200-500 years). Securing a history for gully inception and development (sensu Chiverrell et al., 2007 and 2008) would enable a more holistic understanding of late Holocene sediment delivery and the fluvial regime of the upper Maritsa basin. There are equivalent gaps in understanding in the regional vegetation history and, in order to explore linkages between land cover and geomorphic response, these types of data are critical. Thus securing low altitude palaeoecology sites in or adjacent to these agricultural landscapes is an important next phase to research in this region. There is a clear need for more detailed investigation of the wider Maritsa floodplain to better constrain the histories of sediment accumulation and geomorphic development in the current channel and for the two major palaeochannel zones. An integrated palaeoecological, geomorphological and geoarchaeological approach has proven and will be critical underpinning future efforts to generate a better understanding of the site, fluvial system and surrounding landscape.

Haut de page

Bibliographie

Archibald Z.H. (1998)The Odrysian Kingdom of Thrace. Orpheus Unmasked. Oxford Monographs, Oxford, 370 p.

Archibald Z.H. (2002a) – Attic Figured Pottery from Adjiyska Vodenitsa (Adžijska Vodenica), Vetren 1989-95. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros 2: Excavations and Studies. Charles University Press, Prague, 131-148.

Archibald Z.H. (2002b) – A River Port and emporion in Central Bulgaria: An Interim Report on the British Project at Vetren. Annual of the British School at Athens 97, 309-351.

Archibald Z.H. (2002c) – The Odrysian river port near Vetren, Bulgaria, and the Pistiros inscription. Talanta 32-33, 253-275.

Archibald Z.H. (2004)–Inland Thrace. In Herman Hansen M., Heine Nielsen T. (Ed.): An Inventory of Archaic and Classical Poleis. Oxford University Press, Oxford, 885-899.

Archibald Z.H. (2007) – Excavations by the British team (1999-2005) in the northern and southern sectors of AV1, Vetren – Pistiros. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros 3: Excavations and Studies. Charles University Press, Prague, 81-110.

Baltakov G., Kenderova R., Fitova E. (1996) – Geomorphological and palaeogeographical environmental condition of the emporion. In Bouzek J., Domaradzki M., Archibald Z.H. (Ed.): Pistiros I: Excavations and Studies. Charles University Press, Prague, 181-185.

Bouzek J. (1996) – Pistiros as a river harbour: Sea and river transport in antiquity. In Bouzek J., Domaradzki M., Archibald, Z.H. (Ed.): Pistiros I: Excavations and Studies. Charles University Press, Prague, 221-222.

Bouzek J., Domaradzki M., Archibald Z.H. (Ed.) (1996)Pistiros I : Excavations and Studies. Charles University Press, Prague, 240 p.

Bouzek J., Domaradzka L., Archibald Z.H. (Ed.) (2002)Pistiros II : Excavations and Studies. Charles University Press, Prague, 348 p.

Bouzek J., Domaradzka L., Archibald Z.H. (Ed.) (2007)Pistiros III : Excavations and Studies. Charles University Press, Prague, 286 p.

Bouzek J., Musil J. (2007) – Preliminary report of the Czech mission: the tripartite house south of the main east-west street (Southern House). In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros III : Excavations and Studies. Charles University Press, Prague, 62-80.

Bouzek J., Rückl Š., Titz P., Tsotschev C. (2007) – Trade amphorae. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros III : Excavations and Studies. Charles University Press, Prague, 133-186.

Bozilova E., Tonkov S., Pavlova D. (1986) – Pollen and plant macrofossil analyses of the Lake Sucho Ezero in the south Rila mountains. Annual of Sofia University 80, 48-57.

Brunet M. (2000) – Thasos : Structure du territoire et fonctionnement économique de la cite. In Domaradzki D., Domaradzka L., Bouzek .J, Rostropowicz J. (Ed.) : Pistiros et Thasos : structures économiques dans la péninsule balkanique dans le 1er millénaire av. J.-C. Opole University Press, Opole, 183-190.

Burchfiel B.C., Nakov R., Dumurdzanov N., Papanikolaou D., Tzankov T., Serafimovski T., King R.W., Kotzev V., Todosov A., Nurce B.(2008) – Evolution and dynamics of the Cenozoic tectonics of the South Balkan extensional system. Geosphere 4, 919-938.

Cahill N. (2002)Household and City organization at Olynthus. Yale University Press, New Haven, 383 p.

Casana J. (2008) – Mediterranean valleys revisited: Linking soil erosion, land use and climate variability in the Northern Levant. Geomorphology 101, 429-442.

Chambers F.M., Mauquoy D., Brain S.A., Blaauw M., Daniell J.R.G. (2008) – Globally synchronous climate change 2800 years ago: Proxy data from peat in South America. Earth and Planetary Science Letters 253, 439-444.

Chankowski V., Domaradzka L. (1999) – Réedition de l’inscription de Pistiros et problèmes d’interprétation. Bulletin de correspondance hellénique, 123, 247-258.

Chankowski V., Fouache E. (2000) – Pistiros (Bulgarie). Bulletin de correspondance hellénique, 124, 643-654.

Chankowski V., Gotzev A. (2002) – Pistiros et son territoire : les premiers résultats de la prospection franco-bulgare. In Bouzek J., Domaradzka L., Archibald Z.H., (Eds): Pistiros II: Excavations and Studies. Charles University Press, Prague, 271-282.

Chankowski V., Gotzev A., Nehrizov G. (2004-2005) – Pistiros. Bulletin de correspondance hellénique, 128-129, 1227-1253.

Chiverrell R.C., Harvey A.M., Foster G.C., (2007) – Hillslope gullying in the Solway Firth-Morecambe Bay region, Britain: responses to human impact and/or climatic deterioration? Geomorphology 84, 317-343.

Chiverrell R.C., Harvey A.M., Hunter S.Y., Millington J., Richardson N.J. (2008) – Late Holocene environmental change in the Howgill Fells, Northwest England. Geomorphology 100, 41-69.

Domaradzka L. (2002a) – Catalogue of graffiti discovered during the excavations at Pistiros-Vetren 1988-1998. Part one. Graffiti on fine imported pottery. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros II: Excavations and Studies. Charles University Press, Prague, 209-228.

Domaradzka L. (2002b) – Addendum ad Pistiros vol. I: The Pistiros – Vetren inscription. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros II: Excavations and Studies. Charles University Press, Prague, 339-342.

Domaradzka L., Domaradzki M. (1999) – Population structure of Pistiros. Ancient Macedonia, Sixth International Symposium, I, 383-392.

Domaradzki M. (1987) – Les données numismatiques et les études de la culture thrace du second Age du Fer (in Bulgarian with French summary). Numizmatika 21, 4-18.

Domaradzki M. (1993) – Pistiros : centre commercial et politique dans la vallée de Maritza (Thrace). Archeologia (Warsaw) 44, 35-57.

Domaradzki M. (1996) – Interim report on archaeological investigations at Vetren-Pistiros, 1988-94. In Bouzek J., Domaradzki M., Archibald Z.H. (Ed.): Pistiros I: Excavations and Studies. Charles University Press, Prague, 13-34.

Domaradzki M. (2000) – Problèmes des emporia grecques. In Domaradzki M., Domaradzka L., Bouzek J., Rostropowicz, J. (Ed.) : Pistiros et Thasos : structures économiques dans la péninsule balkanique dans le 1er millénaire av. J.-C. Opole University Press, Opole, 29-38.

Domaradzki M. (2002) – Interim report on fieldwork at Vetren-Pistiros, 1995-98. In Bouzek J., Domaradzka L., Archibald Z.H., (Ed.): Pistiros II: Excavations and Studies. Charles University Press, Prague, 11-29.

Domaradzki M., Domaradzka L., Bouzek J., Rostropowicz, J. (Ed.) (2000)Pistiros et Thasos : structures économiques dans la péninsule balkanique dans le 1er millénaire av. J.-C. Opole University Press, Opole, 274 p.

Filipova-Marinova M. (1995) – The late Quaternary history of the genus Fagus in Bulgaria. In Bozilova E., Tonkov S. (Ed.): Advances in Holocene palaeoecology in Bulgaria. Pensoft, Sofia, 84-95.

Filipovitch L. (1995) – Palynological data on the formation of recent vegetation in Dospat Mountain in the western Rhopodes. Phytologica Balcanica 1, 5-11.

Filipovitch L., Lazarova M. (2002) – Late and post-glacial vegetation dynamics in the Western Rhodopes (Bulgaria) based on pollen analyses and radiocarbon dating. Academie Bulgarian Sciences 55-7, 55-60.

Grandjean Y., Salviat F. (2000)Guide de Thasos. École française d’Athènes, Athens, 330 p.

Howard A.J., Macklin M.G., Bailey D.W., Mills S., Andreescu R. (2004) – Late-glacial and Holocene river development in the Teleorman Valley on the southern Romanian Plain. Journal of Quaternary Science 19-3, 271-280

Ignatov V., Kuncheva-Ruseva T., Velkov K., Popova T. (2006) – Archeologicheski razkopki v mestnostta « Shihanov Bryag » do Harmanli (in Bulgarian). In Nikolov V., Nehrizov G., Tsvetkova J. (Ed.): Spasitelni Archeologicheski Razkopki po traseto na Zhelezopatnata liniya Plovdiv-Svilengrad prez 2004. Bulgarska Akademiya na Naukite, Archeologicheski Institut s Muzei, Veliko Turnovo, 335-397.

Katevski I. (2002) – New data from the geophysical surveys on emporion Pistiros. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros II: Excavations and Studies. Charles University Press, Prague, 299-302.

Katinčarova D. (2007) – Survey along the main street – from the Eastern Gate of the Emporion Pistiros inwards and in the sector of the Eastern city wall In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros 3: Excavations and Studies. Charles University Press, Prague, 35-61.

Kenderova R., Fitova E. (2002) – Development of the left banks of the Maritsa river on the territory of Pistiros. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros II: Excavations and Studies. Charles University Press, Prague, 321-327.

Kenderova R., Baltakov, G., Fitova E. (2007) – Palaeohydrological evolution of the Maritsa Plain in the area of the emporion Pistiros during the last 2500 years (Upper Thacian Lowland, Bulgaria) In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros 3: Excavations and Studies. Charles University Press, Prague, 272-279.

Kovacheva M., Gigov V. (1996) – Emporion Pistiros and archaeomagnetic studies. In Bouzek J., Domaradzki M., Archibald, Z.H. (Ed.): Pistiros I: Excavations and Studies. Charles University Press, Prague, 187-195.

Kovacheva M., Jordanova N., Gigov V. (2002) – Supplementary archeomagnetic investigations on structures of burnt clay from emporion Pistiros. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros II: Excavations and Studies. Charles University Press, Prague, 307-319.

Lazarova M. (2003) – Late Holocene vegetation history of the central Rhodopes mountains, southern Bulgaria. In Tonkov S. (Ed.): Aspects of palynology and palaeoecology. Pensoft, Sofia, 245-256.

Matthews J.A. (1993) - Radiocarbon dating of arctic-alpine palaeosols and the reconstruction of Holocene palaeoenvironmental change. In Chambers F.M. (Ed.): Climate Change and Human Impact on the Landscape. Chapman and Hall, London, 83–100.

Miall A.D. (1996)The geology of fluvial deposits: sedimentary facies, fasin analysis, and petroleum geology. Springer, Berlin, 582 p.

Okay N., Okay A. (2002) – Tectonically induced quaternary drainage diversion in the northeastern Aegean. Journal of the Geological Society 159, 393-399.

Ovenden-Wilson S.M. (1999) – Vetren-Pistiros 1999, Bulgaria. Unpublished Geophysical Survey Report 99/88, 11 p.

Ovenden-Wilson S.M. (2001) – Pistiros 2001, Bulgaria. Unpublished Geophysical Survey Report 2001/46, 8 p.

Popova T. (1996) – The archaeobotanic samples from the clay altar. In Bouzek J., Domaradzki M., Archibald, Z.H. (Ed.): Pistiros I: Excavations and Studies. Charles University Press, Prague, 173-174.

Popova T. (2002) – Studies of carbonised vegetation remains from Pistiros. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros II: Excavations and Studies. Charles University Press, Prague, 289-297.

Popova T. (2006) – Rastitelni ostanki ot yamno svetilishte ot zhelyasnata epocha pri Svilengrad’, (Bulgarian) In Nikolov V., Nehrizov G., Tsvetkova J. (Ed.): Spasitelni Archeologicheski Razkopki po traseto na Zhelezopatnata liniya Plovdiv-Svilengrad prez 2004. Bulgarska Akademiya na Naukite, Archeologicheski Institut s Muzei, Veliko Turnovo, 518-520.

Sbonias K. (1999) – Investigating the interface between regional survey, historical demography and paleodemography. In Bintliff J., Sbonias K. (Ed.): Reconstructing Past Population Trends in Mediterranean Europe. The Archaeology of Mediterranean Landscapes. Oxbow Monographs, Oxford, 219-234.

Stallibrass S. (2007) – Research questions and some preliminary results for the animal bones from the excavations at Adjiyska Vodenitsa, Vetren, Bulgaria. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros 3: Excavations and Studies. Charles University Press, Prague, 264-271.

Stallibrass, S. (2008) - Body parts, placements and people in an Iron Age town in Bulgaria.In Campana D., Choyke A., Crabtree P., deFrance S., Lev-Tov J. (Eds): Anthropological approaches to Zooarchaeology: Colonialism, Complexity and Animal Transformations. Oxford, Oxbow Books, in press.

Stuiver M., Kra R.S. (Ed.) (1986) – Calibration issue, Proceedings of the 12th International 14C conference. Radiocarbon 28, 805-1030.

Stuiver M., Reimer P.J., Reimer R. (2005) - CALIB Radiocarbon Calibration Programme Version 5.0.2html. Belfast, http://calib.qub.ac.uk/calib/.

Taneva V. (2000) – Les monnaies de Pistiros. In Domaradzki M., Domaradzka L., Bouzek J., Rostropowicz, J. (Ed.) : Pistiros et Thasos : structures économiques dans la péninsule balkanique dans le 1er millénaire av. J.-C. Opole University Press, Opole, 47-53.

Tonkova M., Gotsev A. (Ed.) (2008)Trakiyskoto Svetilishte pri Babyak i negovata Archeologicheska Sreda (The Thracian sanctuary at Babyak and its Archaeological Setting: in Bulgarian). Municipality of Belitsa, Bulgaria, unpublished report submitted to the Phare initiative of the European Union, 256 p.

Tsotchev C. (2007) – Black Sea Transport amphorae. In Bouzek J., Domaradzka L., Archibald Z.H. (Ed.): Pistiros 3 : Excavations and Studies. Charles University Press, Prague, 187-189.

Van Geel B., Buurman J., Waterbolk H.T. (1996) – Archaeological and palaeoecological indications of an abrupt climate change in The Netherlands, and evidence for climatological teleconnections around 2650 BP. Journal of Quaternary Science 11, 451-460.

Velčev A. (1995) – The Pleistocene glaciations in the Bulgarian mountains (in Bulgarian with English summary). Annual of Sofia University, Faculty of Geology and Geography 87, 53-65.

Westaway R. (2006) – Late Cenozoic extension in southwest Bulgaria: a synthesis. In Robertson A.H.F., Mountrakis, D. (Ed.): Tectonic Development of the Eastern Mediterranean Region. Geological Society of London Special Publication 260, 557-590.

Zagorchev I. (2007) – Late Cenozoic development of the Strouma and Mesta fluviolacustrine systems, SW Bulgaria and northern Greece. Quaternary Science Reviews 26, 2783-2800.

Haut de page

Annexe

   

Version française abrégée

Cet article essaie de faire la part des interactions entre les activités humaines et les dynamiques environnementales pour expliquer la manière dont le port fluvial d’Adjiyska Vodenitsa (Bulgarie centrale ; fig. 1), centre commercial et culturel de la Thrace occidentale ancienne, localisé au bord de la rivière Maritsa, fonctionnait avant sa fossilisation partielle. Cette cité a joué un rôle important dans un environnement essentiellement rural qui était celui de la Thrace continentale entre le début du Vè siècle et la fin du IIè siècle av. J.-C. L’inscription dite de Pistiros, gravée en grec sur un grand bloc de granit, et l’important matériel (pièces de monnaie, amphores à vin, tessons de céramiques, etc.) livré par les fouilles archéologiques démontrent l’importance des échanges commerciaux entre le site, le monde Egéen et les régions de la Mer Noire. Le site archéologique occupe une position de carrefour et un point de rupture de charge entre une route continentale nord-sud et fluviale ouest-est.

Le site d’Adjiyska Vodenitsa est étudié par les archéologues depuis 1988. L’étude régionale s’est appuyée sur la télédétection (Bulgares, Britanniques et Français), des prospections de surface régionales (Bulgares et Français), mais le cœur de l’étude archéologique est constitué par des fouilles stratigraphiques réalisées sur et autour de la porte orientale de la ville (Tchèques, Britanniques et Bulgares). Les études géomorphologiques initiées par G. Baltakov et al. (1996) et V. Chankowski et E. Fouache (2000) se sont centrées sur la cartographie des lits fluviaux et des systèmes de terrasses et de cônes de déjection associés, sur la description de coupes et sur des datations 14C réalisées sur des échantillons de matière organique prélevés dans les deux niveaux de terrasse alluviale de la Maritsa, adjacents au site d’Adjiyska Vodenitsa. La cartographie géomorphologique repose sur une interprétation des données SRTM, un modèle numérique de terrain (MNT) et des observations de terrain. D’autres cartes non présentées ici ont été réalisées dans un petit bassin fluvial situé en amont de Belovo, à proximité de Kostenets (fig. 1). Les inondations d’août 2005 ont mis à jour des coupes dans des sédiments fluviatiles à proximité d’Adjiyska Vodenitsa, lesquelles ont été décrites et échantillonnées pour des datations radiométriques. Les interprétations des paléoenvironnements fluviaux reposent sur une chronostratigraphie des paléofaciès reconstitués à partir des informations fournies par des carottages réalisés antérieurement (Chankowski et Fouache 2000), des analyses granulométriques et des mesures radiométriques et de susceptibilité magnétique.

L’analyse de la géomorphologie fluviale montre que la plaine d’inondation de la Maritsa comprend deux niveaux de terrasse alluviale (fig. 2) : une haute terrasse, pratiquement située à l’altitude sur laquelle est fondée le site archéologique, et une basse terrasse qui « attaque » le site en falaise vive. En coupe, on constate que le site archéologique est construit sur un cône alluvial latéral constitué de matériaux grossiers se raccordant latéralement à la haute terrasse alluviale de la Maritsa (fig. 3). Les résultats des datations 14C (tab. 1) réalisées sur un sol alluvial fossilisé entre deux lentilles graveleuses sur la haute terrasse et sur un paléochenal de la basse terrasse montrent de grands changements de dynamique. La phase d’alluvionnement la plus récente, à la surface de la haute terrasse, est datée 2405±35 BP (750-395 av. J.-C.), et l’abandon d’un paléochenal ancien au niveau de la basse terrasse est daté d’avant 985±35 BP (990-1155 ap. J.-C.). A partir de cette dernière date, le déplacement latéral du chenal de la Maritsa vers le nord a érodé la berge qui supportait le site archéologique, contribuant à le détruire partiellement. En croisant ces résultats avec ceux des fouilles (fig. 4), il est dès lors possible de reconstituer l’extension minimale du site, soit environ 2500 m2 (2,5 ha), dont 97 % ont été détruits par le sapement latéral de la rivière Maritsa. La présence de dépôts d’inondation au sein des couches d’occupation témoigne d’un niveau plus élevé d’écoulement dans l’Antiquité.

La structure urbaine de la ville est comparable à celle de la cité d’Olynthos en Chalcidique (Grèce) ou de la ville de Thasos (Grèce). La ville est fondée vers 450-425 av. J.-C. (fig. 5). On observe plusieurs phases de réaménagement mais la destruction de la porte orientale, aux alentours de 300 av. J.-C., marque le début d’une période de trouble, dont témoigne un trésor monétaire daté de 280 av. J.-C., qui va aboutir à la destruction et à l’abandon de la cité. La richesse monétaire préromaine du site est un bon indicateur de l’importance de cette cité comme place commerciale d’importance régionale tandis que le plan urbain est caractéristique de fonctions commerciales d’ampleur exceptionnelle. L’exploitation des ressources naturelles locales était importante comme en témoigne le proche terroir agricole, dédié à l’agriculture et aux activités pastorales, les nombreuses espèces cultivées identifiées dans les fouilles, mais aussi l’exploitation du granite et de l’argile. Les ressources minières régionales comme le cuivre, l’argent et le fer, étaient travaillées également en abondance sur le site. La Maritsa n’était pas navigable toute l’année et n’était pas franchissable en crue, ce qui nous amène à titre d’hypothèse à envisager que le transport des marchandises ne se faisait que durant certaines périodes seulement.

Antérieurement à l’emboîtement de la basse terrasse de la Maritsa dans la haute terrasse, l’écoulement de la Maritsa correspond à un style fluvial en tresses. Entre 750-395 av. J.-C. et 990-1155 ap. J.-C., on assiste à un abaissement du niveau de base, progressivement acquis lors des migrations latérales, suivi d’une phase d’aggradation. Cette aggradation correspond aux dépôts des alluvions qui composent aujourd’hui la basse terrasse. Ils sont associés à un style fluvial à méandres libres, style qui a contribué au sapement et à l’érosion du site archéologique. Il n’est pas possible de déterminer si la cité a été confrontée à ce sapement durant son occupation. L’évolution dynamique de la Maritsa est la conséquence de plusieurs facteurs dont la combinaison est difficile à établir : tectonique, aggradation alluviale, variabilité climatique et conséquences de la gestion par l’Homme de son environnement, notamment par le pâturage et l’agriculture. La période d’incision qui conduit à l’édification d’une basse terrasse en contrebas de la plus haute peut refléter aussi bien une augmentation des débits qu’une réduction des apports sédimentaires, tandis que l’aggradation traduit un excès sédimentaire. Il est sûr que les conditions plus humides attestées entre 600-400 av. J.-C. et le développement continu de la mise en valeur agricole à partir de 500 av. J.-C. ont joué un rôle, mais il est difficile d’affecter à l’un ou à l’autre de ces facteurs un rôle prépondérant. Cependant, on peut affirmer que la rivière Maritsa a joué un grand rôle et qu’elle a été responsable de la destruction de la cité postérieurement à son abandon. En revanche, sur la période d’occupation de la cité, seules des crues temporaires et d’extension limitée ont affecté la cité.

Haut de page

Table des illustrations

Titre Fig. 1 – General location and setting of the archaeological site of Adjiyska Vodenitsa in the Maritsa basin.Fig. 1 – Localisation du site archéologique d’Adjiyska Vodenitsa dans le bassin de la Marica.
Légende 1: study site; 2: published pollen sites; 3: ice extent at last glacial maximum; 4: rivers mentioned in the text; 5: other rivers; 6: limit of the Maritsa basin; 7: international boundaries.1 : secteur d’étude ; 2 : sites sur lesquels les données polliniques ont été publiées ; 3 : extension des glaces au Dernier Maximum Glaciaire ; 4 : cours d’eau mentionnés dans le texte ; 5 : autres cours d’eau ; 6 : limite du bassin de la Marica ; 7 : frontières internationales.
URL http://geomorphologie.revues.org/docannexe/image/7747/img-1.jpg
Fichier image/jpeg, 584k
Titre Fig. 2 – Geomorphology of the upper North Thracian basin around Adjiyska Vodenitsa, also identifying the location of key archaeological sites [base map derived from NASA Shuttle Radar Topographic Mission (SRTM) digital elevation model (DEM), supplied by CGIAR-CSI (http://srtm.csi.cgiar.org/​) SRTM)].Fig. 2 – Géomorphologie du bassin supérieur de la Marica dans les environs d’Adjiyska Vodenitsa et localisation des sites archéologiques [la carte a été réalisée à partir des données issues du modèle numérique de terrain (MNT) acquis lors de la mission STRM de la NASA, CGIAR-CSI (http ://srtm.csi.cgiar.org/)SRTM)].
Légende 1: archaeological sites; 2: study site; 3: extent of fig. 3; 4: alluvial fan; 5: low Maritsa river terrace; 6: high Maritsa river terrace; 7: river bank; 8: slope lines; 9: palaeochannels; 10: sediment exposures; 11: road bridges; 12: contours at 10 m intervals; 13: contours at 100 m intervals.1 : sites archéologiques ; 2 : secteur étudié ; 3 : délimitation de la fig. 3 ; 4 : cône alluvial ; 5 : basse terrasse de la Marica ; 6 : haute terrasse de la Marica ; 7 : levée de berge ; 8 : lignes de plus grande pente ; 9 : paléochenaux ; 10 : coupes stratigraphiques ; 11 : ponts routiers ; 12 : courbes de niveau (équidistance 10 m) ; 13 : courbes de niveau (équidistance 100 m).  
URL http://geomorphologie.revues.org/docannexe/image/7747/img-2.jpg
Fichier image/jpeg, 696k
Titre Fig. 3 – Stratigraphy and radiocarbon dating at two exposed sections near the archaeological site at Adjiyska Vodenitsa.Fig. 3 – Stratigraphie et datation au carbone 14 de deux profils stratigraphiques à proximité du site archéologique à Adjiyska Vodenitsa.
Légende 1: silty clay; 2: rounded gravel; 3: sand; 4: woody peat; 5: peat; 6: soil layer 1; 7: soil layer 2; 8: angular gravel; 9. lake mud. Inset map shows the general location (outlined on fig. 2).1 : argiles limoneuses ; 2 : graviers émoussés ; 3 : sables ; 4 : tourbe ligneuse ; 5 : tourbe ; 6 : sol 1 ; 7 : sol 2 ; 8 : graviers anguleux ; 9 : vases de marécage. Localisation générale sur la fig. 2.
URL http://geomorphologie.revues.org/docannexe/image/7747/img-3.jpg
Fichier image/jpeg, 288k
Titre Fig. 4 – Plan of the excavations at Adjiyska Vodenitsa, showing the eastern gateway, fortification wall and east-west roadFig. 4 – Plan des fouilles à Adjiyska Vodenitsa, montrant la porte orientale, le mur de l'enceinte et la rue ouest-est
Crédits (after Bouzek et al., 2007; © Pistiros international excavation project).(d’après Bouzek et al., 2007 ; © Equipe internationale Pistiros).
URL http://geomorphologie.revues.org/docannexe/image/7747/img-4.jpg
Fichier image/jpeg, 268k
Titre Fig. 5 – Excavated structures (A, C) showing the fortification wall (A, B), the eastern gateway and road (C), and house (No 1) and the road (D; © Pistiros international excavation project).Fig. 5 – Structures découvertes lors des fouilles (A, C) montrant le mur de l'enceinte (A, B), la porte orientale et la route (C) et la maison (N° 1) et la rue (D ; © Equipe internationale Pistiros).
URL http://geomorphologie.revues.org/docannexe/image/7747/img-5.jpg
Fichier image/jpeg, 475k
Haut de page

Pour citer cet article

Référence papier

Richard C. Chiverrell et Zosia H. Archibald, « Human activity and landscape change at Adjiyska Vodenitsa, central Bulgaria », Géomorphologie : relief, processus, environnement, vol. 15 - n° 4 | 2009, 287-302.

Référence électronique

Richard C. Chiverrell et Zosia H. Archibald, « Human activity and landscape change at Adjiyska Vodenitsa, central Bulgaria », Géomorphologie : relief, processus, environnement [En ligne], vol. 15 - n° 4 | 2009, mis en ligne le 01 janvier 2012, consulté le 20 octobre 2017. URL : http://geomorphologie.revues.org/7747 ; DOI : 10.4000/geomorphologie.7747

Haut de page

Auteurs

Richard C. Chiverrell

Department of Geography, University of Liverpool, Roxby Building, Liverpool, L69 7ZT, UK (rchiv@liv.ac.uk)

Zosia H. Archibald

School of Archaeology, Classics, and Egyptology (SACE), University of Liverpool, 12-14 Abercromby Square Liverpool L69 7WZ (Z.Archibald@liverpool.ac.uk)

Haut de page

Droits d’auteur

© Groupe français de géomorphologie

Haut de page
  • Logo CNRS - Institut des sciences humaines et sociales
  • Logo Groupe français de géomorphologie
  • Revues.org