Radiocarbon AMS dating was carried out by Dr. K. VAN DER BORG, Utrecht (Netherlands). We express our thanks to Prof. Dr. I. MARIOLAKOS, Athens (Greece), for fruitful discussion, and to the Institute of Geology and Mineral Exploration at Athens for working permits. We wish to thank Dr. L. KOLONAS, Greek Ministry of Culture, and the Ephory of Patras for support during field work. J. LUTHER helped during field work and laboratory analyses. Thanks are also due to Dr. A.J. LONG, Durham (UK) and Mark BESONEN (U.S.A.), for polishing our English.
1The delta regions of the Mediterranean have witnessed the most extensive coastal changes in history. They are geological archives, storing large volumes of sediment which was mainly produced by terrestrial erosion and delivered to the coast by rivers. They also document the transition from environments dominated by natural processes to environments controlled by human activities. Deltas serve as excellent geo-archives for the effects of settlement phases on one hand, and of sea-level fluctuations, earthquakes and other natural phenomena on the other.
2The beginning of the evolution of nearly all Mediterranean deltas is the same: during the Last Glacial Maximum (LGM) around 20 000 years ago, sea level fell to its lowest point, ca. 120 m below that of today. From ca. 14000 BP onwards, sea level rose at an average speed of 1.5 cm/yr, presumably reaching its present position (or close to it) around 6000-5500 BP. In some Mediterranean areas, this furthest inland shoreline is associated with Neolithic coastal sites. Only after this date did the significant expansion of Mediterranean deltas occur. In the following 5500 or so years, many of the former marine embayments silted up, mainly due to fluvial sedimentation but also as a result of the input of sediment derived from coastal erosion of headlands and longshore drift.
3In this paper we document the mid- and late Holocene evolution of three deltas. Case study 1 deals with the evolution of the Küçük Menderes alluvial plain and delta in western Turkey, and pays particular attention to the development and decline of the famous ancient city of Ephesus. Two different scenarios will be presented. The first is exclusively based on literary sources, whereas the second is the result of geoarchaeological research based on systematic coring. Our second case study focuses on the progradation of the Karamenderes which seriously affected the history of ancient Troia. Both scenarios are mainly based on geomorpho-logical and geoarchaeological evidence. However, they differ considerably in the interpreta-tion of the data which, among others, leads to considerable discrepancies for the reconstruc-tion of the shoreline during the time of the so-called Troian War. Case study 3 deals with the Holocene progradation of the Acheloos delta in northwestern Greece. Selected sections are used to establish the first palaeo-geographical scenario for the region based on systematic corings. While case studies 1 and 2 are based on the results of geoscientific investigations lasting more than a decade, case study 3 is an ongoing project.
4The major purpose of geoarchaeological research in delta and floodplain areas is to detect the maximum extent of the Holocene transgression ca. 5000-6000 years ago and to decipher the shifts in the shoreline during the following regression and the associated delta progradation. In general, as well as for the three presented case studies, the main tool for palaeoenvironmental reconstruction is the interpretation of sediment cores taken by a Cobra vibration coring device, and which reach average depths of 10-20 m below ground surface. The general stratigraphy of these cores is as follows: on top of abraded bedrock lies the transgression deposit (often pebbles in sandy matrix), indicating the first transition of the shoreline during the early Holocene sea-level rise. This littoral facies grades upwards into a marine one as water depths increase. Depending on the sedimentation rate, this stratum may be several metres thick. The second transition of the shoreline occurs during the following regression – which is either a glacio-eustatic phenomenon (Kayan, 1995) or a sedimentological effect due to the by-passing of the delta. It is recorded either by another littoral facies (e.g. beach sand) or by a lagoonal layer (mostly silt and clay). The latter reflects a sand bar or barrier beach developing seaward of the cored site. The lagoonal facies may subsequently change into a freshwater lake deposit. The uppermost sedimentary units are deposits of the delta top and the river floodplain, and typically comprise river gravel, sand or fine grained alluvium.
5To establish a chronostratigraphy, the most suitable materials in the cores are macrofloral remains (e.g. grape seed, olive pips, peat), articulated mollusc shells (although the reservoir effect is a pending problem) and, of course, characteristic datable ceramic fragments. The different environments of deposition (marine, littoral, lagoonal, lacustrine, fluvial) are defined by microfaunal analysis (e.g. ostracods; Handlet al., 1999) or by geochemical parameters (Vöttet al., 2002). The large number of corings and datings allows the development of palaeogeographical maps that depict changes in the shoreline through time. The reconstruction of sea-level fluctuations is challenging since there are seldom good sea-level indicators (i.e. possessing a defined age and height relationship to a former sea-level) in the sediment cores. The best are: (a) the transgressive facies at the base of the core; (b) the change from marine to terrestrial facies in the upper part of the core, which may be detected by sedimentological and faunal criteria; and (c) coastal peat. This information must then be supplemented by archaeological data, such as submerged ruins. The geoarchaeological approach of our studies is outlined in Fig. 1 (see also Brückner, 2003; Brückneret al., 2002).
Fig. 1 - The geoarchaeological approach
Source: H. Brückner (2003: Fig. 2), modified.
6A major aim of our geoarchaeological research in and around Ephesus was to reconstruct the delta growth of the Küçük Menderes (Cayster, in antiquity: Kaystros) in space and time, and thereby to detect the shifts in the shoreline as well as the palaeogeographical changes of the city’s environs (Photo 1). The following chapter is mainly based on Brückner (2004).
Photo 1 - The Roman harbour of Ephesus silted up due to the progradation of the Küçük Menderes (Kaystros) delta
The Arcadiane (street with erected columns) connects the theatre with the harbour basin («inner port» in Fig. 3). View from the theatre looking west with Bülbüldag to the left and Aegean Sea in the far background. (Photograph by H. Brückner, August 2001).
7The different scenarios for the progradation of the Küçük Menderes delta shown in Fig. 2 (I) are based on the interpretation of documentary sources as oppose to geologic-chronostratigraphic evidence. It is, therefore, not surprising that they are contradictory in many aspects.
8Eisma (1978) proposed a slow shoreline advance during the Archaic and Classical times (ca. 750-300 BC), during which the former island of Syrie was reached. This was followed by rapid delta growth in the Hellenistic period (5 km, ca. 300–100 BC). In Roman times the speed of progradation decelerated (2 km, ca. 100 BC–200 AD), and since ca. 700 AD the coastline has been more or less stable. This latter assumption is, however, wrongly based on the attribution of a pumice layer in the penultimate beach ridge to the Santorini eruption of 726 AD. According to Erinç’s (1978) interpretation of ancient writings, the island of Syrie and the site of the later Artemision were reached by the coastline in the 7th century BC, after which time there was a phase of rapid delta growth. In the 3rd century BC (Hellenistic time) he suggests that the delta front was already situated in the area of the so-called outer harbour. In the following four to five centuries, it prograded another 1.5-2.5 km westwards (cf. Fig. 2(I)).
9We learn from Strabo (16.1.24) that around the time of Christ’s birth an attempt was made to save the harbour of Ephesus by constructing a canal-like connection to the sea and by the erection of a dam against the floods of the river Kaystros. The «inner harbour» (Photo 1 and Fig. 2(I), no. 4) was still in use in Early Byzantine times (5th – 7th centuries AD) and was connected to the sea via a canal. Thereafter, an «outer harbour» was needed. This «panormos» is assumed to have been situated in the northernmost part of the Kenchrios valley (Fig. 2(I), no. 3). Philippson (1936: 30) assumes that its position was west of the Arvalia (= Kenchrios) valley and that it had already become the main harbour in Roman Imperial times. Meriç (1985) located the outer harbour in the area of Alaman Gölü (cf. Hess, 1989).
Fig. 2 - Scenarios of the progradation of the Küçük Menderes (Kaystros) delta in space and time
I (upper part): These scenarios are based on the interpretation of literary sources. They are quite controversial. II (lower part): Coastline positions are based on the association of geological, archaeological and historical data. Note that within the inner embayment the morphology was that of an irregular delta with distributary arms, typical of a low energy coast. This pattern has changed to higher wave energies since Late Byzantine times, after the delta had reached the outer parts of the embayment with a longer fetch and the influence of the longshore current. The white areas in the figure show regions where data are missing.
Sources for I: Brückner (1997a: Abb. 1), modified; based on: Eisma (1978: Fig. 2), Erinç (1978: Fig. 4), Meriç (1985: Abb. 1 & Tafel VI), Hess (1989: Abb. 2); supplemented by: Geological Map of Turkey, 1:500,000, Sheet Denizli, Ankara 1964; own research. Source for II: Kraftet al. (2003a: Fig. 8), modified.
10In order to solve the question of the siltation of the marine embayment around Ephesus from a geomorphological-geoarchaeological perspective, a transect of boreholes was drilled from the footslope of Panayırdag close to the Koressos gate via the former island of Syrie to Cevasirdag at the northern flank of the graben (Fig. 2; location of coring sites see Fig. 2(I)). The transition of the delta front in this area was of particular interest. For detailed information about profile stratigraphy see Brückner (2004).
11The summary view of the geological data leads to the scenario for the delta growth shown in Fig. 2(II) (more details in Brückner, 1997; Kraftet al., 2000, 2001, 2003a). When compared with previous reconstructions presented in Fig. 3(I), this new interpretation shows three major differences relating to (i) the chronology of shoreline displacements and (ii) the geomorphology of the delta.
12Delta growth in space and time is based on geoarchaeological evidence with radiocarbon and ceramic chronostratigraphy, cross-checked with written sources. The reliability of the latter was discussed with historians. In the context of this paper we want to emphasize the third difference concerning the geomorphology of the delta. From Neolithic to High Byzantine times the shoreline configuration was dominated by an irregular delta distributary pattern typical of a low energy coastline. A modern day analogy is the Sperchios delta in central Greece, especially its morphology in the inner part of the Gulf of Maliakos where wave energy is low and water depth shallow (Kraftet al., 2003a: 367 and Fig. 4). Also the Arachthos and Thyamis river deltas in western Greece show typical features of a bird’s foot delta (Besonen, 1997) due to their sheltered position within the Ambrakian Gulf or behind Corfu Island respectively. The morphodynamic pattern changed in Late Byzantine times when the delta front had reached the outer parts of the embayment and higher wave energy and longshore current became dominant factors. This resulted in the evolution of a barrier accretion plain.
13The consequences of the permanent shift in the shoreline – especially for the harbours, their repeated relocation, and the measures taken against their siltation – and the impact of continuous landscape changes on ancient societies are outlined in Kraftet al.(2000, 2001).
14Troia (Troy) is one of the most legendary archaeological sites due to Homer’s Iliad, the epos of the «Troian War», Heinrich Schliemann’s excavations between 1871 and 1890, and multidisciplinary research since 1988 led by Manfred Korfmann. The city rises over the flood plains of Karamenderes (ancient name: Scamandros or Scamander) and Dümrek Su (Simois) rivers (Photo 2). Geoarchaeological research has been carried out for many years by I. Kayan, based on earlier works by J.C. Kraft and O. Erol (Kraftet al., 1980; latest publications: Kayan, 2001; Kayanet al., 2003; Kraftet al., 2003a, b).
Photo 2 - Troia (HisarlIk) hill overlooking the Karamenderes (Scamander) alluvial and delta plain
This former marine and later swampy terrain has recently been transformed into intensively used arable land. In the background the entrance to Çanakkale Strait (Dardanelles or Hellespont; to the right) and the northern extension of Yeniköy ridge (Sigeum or Sigeion; to the left). View towards the northwest.
(Photograph by H. Brückner, August 2000).
15The fame of Troia was partly based on its strategic position: the city controlled navigation to and from the Black Sea via the Çanakkale Strait (Dardanelles, Hellespont). With frequent northern winds in the eastern Mediterranean, ships had to anchor in Besik Bay since they were unable to sail windward. It was then that Troia could collect taxes.
16Geological cross-sections published by Kraftet al. (1980; 2003a, b) and Kayanet al. (2003) show over 50 m of sedimentary strata that provide a depositional record between the Sigeum ridge to the west and the cuestas Hisarlık (with Troia) and Yenikumkale to the east. They reveal sediments of a marine embayment, coastal swamps, coastal barriers and lagoons as well as backswamps, floodplains, river channels and marshes. The Holocene transgression reached ca. 17 km inland up to the area immediately northwest of Pınarbası. During the following 6000 years, the former marine embayment silted up with alluviations and colluvial sediments resulting in the progradation of the Scamander and Simois river deltas. Of special interest are the questions: When did Troia lose direct access to the sea? Where were the harbours? What was the palaeogeography like during the time of the Troian War»?
17There are two scenarios for the siltation of the marine embayment. Both are based on the same geoarchaeological evidence (roughly 250 drilled cores in the floodplain), but they differ in interpretation.
18Using the present-day scenario of the so-called asmak (multiple stream channel estuarine) model of the Scamander delta, Kayanet al. (2003) propose a series of phases of delta growth shown in Fig. 3(I). According to their work the city was still situated by the sea during the settlement phases of the Maritime Troia Culture (Troia I – III, ending 2200 BC). Soon after that time, however, it lost this direct access, when the delta passed by the city. During the times of the Anatolian Troia Culture (Troia IV–V, 2200–1700 BC) and the Troian High Culture (Troia VI – Troia VIIa, 1700–1200 BC) the distance from the coastline continued to increase. Both rivers – the Scamander and Simois – contributed to the siltation.
Fig. 3- Scenarios of the progradation of the Karamenderes (Scamander) and Dümrek (Simois) deltas in space and time
I (upper part): according to Kayanet al. (2003). II (lower part): according to Kraftet al. (2003a). Both scenarios are based on geoarchaeological evidence from drill cores. They differ in the interpretation of the effects of coastal dynamics and in the chronology of the events. Scenario II emphasizes the transition from a bird’s foot delta in a quiescent embayment with low energy wave climate, to the present-day longshore-current dominated coast (cf. Fig. 2). The palaeogeographies around Troia (Troy, Ilion, Ilium) are shown for Neolithic – Late Bronze Age times, around 1200 BC, the time of the «Troian War» described in Homer’s Iliad, for the time of Strabo around 0 BC/AD and for the present situation. Noted are also Luce’s (2003) identification of the Greek camp and ship station based on Strabo’s (13.1.36) account that it was 20 stades from Troia (A), and (B) that the coast was 12 stades away from the city during his time (0 BC/AD) while it was half that distance at the time of the «Troian War» (ca. 1200 BC). The noted 14C dates are from molluscs at the transition from marine to brackish environments indicating the transition of the shoreline during delta progradation.
19The so-called «Troian War», probably Homer’s legendary condensation of several wars around Troia, is nowadays placed at the end of the 13th century BC (destruction of Troia VIIa). By then, according to this interpretation, the shoreline was at a considerable distance to the northwest of the city. Delta growth continued throughout the periods of Troia VIII (Greek Troia = Ilion, 334 – 85 BC) and Troia IX (Roman Troia = Ilium, 85 BC – 450 AD) and continued to the present day.
20An alternate interpretation is suggested by Kraftet al. (2003a) shown in Fig. 3(II). Here, as in the case of the Küçük Menderes river (Chapter 3), the siltation process is compared with that of the Sperchios delta in the Greek Gulf of Maliakos. In the quiescent low wave climate embayment, a delta forms with multiple bird’s foot distributaries and irregularly shaped bays. It is only around 0 BC/AD that the strong east-west current in the Dardanelles starts to take over coastal dynamics, eroding distributary arms and creating sand spits. The other difference to Kayan’s scenario is the use of dates reflecting the last occurrence of mollusks of a marine-brackish environment for establishing a chronology for shoreline displacements. These mollusks render the age of the shoreline transition at the cored site during regression, triggered either by eustatic sea-level fall or by delta growth.
21The latter scenario has two advantages: it fits better the coastal dynamics and it supports evidence from ancient literature. Strabo (13.1.31 and 13.1.36) mentions the distance between Troia and the so-called Achaian harbour during the «Troian War» as 20 stades (ca. 4 km). The best candidate for a natural harbour at that time is the present-day Kesik plain, a former marine embayment on the eastern slope of Sigeum ridge. Strabo mentions that by then the distance from the city to the sea was 6 stades (ca. 1.2 km), half the distance during his own times (cf. also Luce, 2003). It is interesting to note that the geological evidence – as interpreted by Kraft et al. (2003a, b) – correlates well with the relevant Homeric geography: «Nothing in our research negates the writings of Homer!» (Kraft et al., 2003a: 375).
Fig. 4 - Geomorphological map of the Acheloos alluvial plain
Source: A. Vött et al. (2003b), modified.
22The Acheloos alluvial plain lies in the westernmost part of the Greek mainland. Oiniadai, an ancient city in the plain’s southern part, is famous for its shipsheds of the 3rd century BC. Due to siltation processes caused by the progradation of the Acheloos delta, it lost its function as a seaport. Today, Oiniadai lies some 9 km distant from the open sea (Fig. 4). Within the framework of a regional study dealing with Akarnania’s coastal evolution, we carried out vibracorings and analysed sediments to decipher the evolution of the Acheloos delta with special regard to shoreline displacements. This chapter presents selected geomorphological-palaeogeographical results based on a transect along the modern Acheloos River.
23Neumann & Partsch (1885: 350), referring to ancient writers such as Herodotus and Thukydides (both 5th century BC), were the first geomorphologists who pointed out the role of the former Echinades islands, the geographical arrangement of which had accelerated siltation of the former marine embayments. According to Philippson (1958) the ancient harbour of Oiniadai was connected to the sea via the Acheloos river. As the distance from Oiniadai to the river mouth given by Strabo (70 stadia = ca. 13 km) is still accepted as valid today, Philippson argued that the tremendous coastline shift was already completed by the time of Strabo’s account. Piper & Panagos (1981) found three systems of abandoned river channels south and east of the modern river. Apart from minor changes in the area of the actual river mouth, they concluded that for the last 2300 years no larger environmental changes had occured. Villas (1984) was the first to carry out corings in the alluvial plain. Her work suggested that deltaic sedimentation began approx. 5700 BP in a southeastern direction into the lagoon of Etoliko lying approx. 10 km east of the Lesini mountains (see Fig. 4) (Villas, 1984: 115, Fig. 33; cf. Stanley & Warne, 1994: Tab. 1). Later, the fluviodeltaic sedimentation concentrated more towards the southwest and west. Freitag (1994) and Fouache (1999: 71) support the idea that Oiniadai, with its shipsheds, was directly connected to the sea and that the river Acheloos was not directly responsible for the siltation of the Lesini swamp north of Oiniadai. Grove & Rackham (2001: 341) emphasize that the last phase of considerable deltaic sedimentation must have taken place before the time of the ancient writers, at the latest during the 5th century BC, due to an increase in sediment transport, or a temporary decrease of subsidence rates.
24The main objectives of our research were to delineate the Holocene progradation of the Acheloos based on geological-geomorphological evidence in order to shed light on the controversial questions discussed in literary sources.
25Fig. 5 shows a summary view of a coring transect alongside the Acheloos river composed of vibracores OIN 10, 12 and 13. At OIN 10, dark grey fine sand with abundant macrofaunal remains up to 11.32 m b.s.l. document littoral facies. It is followed by light grey silt of a sublittoral environment and bluish grey medium sand of a subsequent high energy sand bar system. Up to 5.40 m b.s.l., we found lagoonal deposits with numerous seashell fragments as well as articulated specimens, interrupted at 7.82 – 6.59 m b.s.l. by sediments of a coastal swamp. A sharp break marks the transition to sandy fluviodeltaic deposits of the prograding Acheloos delta. The topmost stratum consists of brown loamy to sandy levee and flood plain sediments.
26Vibracore profile OIN 12 shows grey sand up to 8.29 m b.s.l. with abundant seaweed at 8.40 m – 8.29 m b.s.l. which, together with macro- and microfauna, indicates a shallow marine environment. Subsequently follows reddish to greenish grey silt void of macrofauna which we attribute to a marshy or shallow marine depositional environment. As in OIN 10, there is a sharp transition to the overlying brownish grey, well sorted sands of the Acheloos delta front, which extend almost to present-day mean sea level. The upper part of the profile shows oxidized sand and loam typical of a fluvial depositional context.
27Profile OIN 13 is characterized by grey sand with large amounts of plant remains up to 11.69 m b.s.l., followed by a thick layer of partly well sorted grey coarse and medium sand. This layer was deposited by the passing-by Acheloos delta. The fluviodeltaic sediments give way to dark grey coastal swamp and clearly laminated marsh deposits (up to 1.66 m b.s.l.). The uppermost layer is made up of flood plain sediments which show features characteristic of freshwater marshes.
28At OIN 10 fluviodeltaic sediments dominate the upper part of the profile and provide the first indicators of a huge river discharging into the sea. The delta sediments clearly truncate a persistent lagoonal environment. At 5.72 m b.s.l. these sediments contained a ceramic fragment proving early human presence. At least for the first phase of settlement at Oiniadai, we can assume that the northern flank of the hill lay adjacent to a widespread lagoonal system (Vött et al., 2003a, b). OIN 12 reflects the evolution of a shallow marine environment in front of the delta to a delta top marsh. Marsh sedimentation was abruptly terminated by delta sand deposition. Profile OIN 13 also shows a regressive sequence initiated by sediments of the Acheloos river with a shallow marine, probably prodeltaic, environment suddenly affected by a system of rapid fluviodeltaic sedimentation.
29Fully marine sediments could only be detected at OIN 10 where they are free of any direct deltaic influence. At OIN 12 and 13 fully marine deposits have to be assumed below maximum coring depth. This means that water depth of an early marine embayment increased towards the southwest. OIN 10 is the only transect site which shows sand bar sediments causing the formation of a lagoon. Based on these data, we assume that during that time at OIN 12 and 13 shallow marine conditions must have prevailed. At the same time the delta front itself is supposed to have been east or even northeast of OIN 10. At OIN 10 a later phase is characterized by a coastal swamp, followed by a second lagoon. Marsh and/or deltaic sand deposition dominate OIN 12 and 13 at comparable depths. This seems to support the idea that the lagoon of Oiniadai reached the area from the west – north of Kounovina – and did not extend to the area around OIN 12 and 13.
30According to the sharp boundaries between the sedimentary units, the following phase is dominated by an abrupt appearance of the Acheloos delta front. Comparing the bases of delta sediments at OIN 10 (5.40 m b.s.l.), OIN 12 (6.16 m b.s.l.) and OIN 13 (11.69 m b.s.l.), there is a clear sedimentation gradient to the southwest which documents increasing water depths. This corresponds to the main flow direction of the modern Acheloos. Further, fluviodeltaic sediments show the highest thickness at OIN 13 and the lowest at OIN 10. This is due to rapid delta progradation which caused sedimentation to last longer downstream. It seems as if deltaic sand sedimentation persisted at OIN 12 when subaerial flood channel sediments were accumulating at OIN 10.
31It is possible to determine the beginning of fluviodeltaic sedimentation at OIN 10 using the 14C AMS technique. The ages are corrected for a reservoir effect of 402 years. An articulated specimen of Dosina exoleta within lagoonal sediments yielded an age of 3482–3372 cal BC (5032±48 BP, 1 sigma error range). A ceramic fragment from 7.90 m below surface (5.72 m b.s.l.) found next to the dated mollusc indicates a very early human occupation of the area. As shown above, the delta front prograded further southwest and reached OIN 12 some 4100 years later. A fragment of Dosina exoleta from the base of the fluviodeltaic sand deposits yielded an age of 681–721 cal AD (1687±26 BP, 1 sigma error range). Since numerous findings of marine macrofaunal remains were made within this sandy layer, it seems unlikely that it belonged to a younger river channel which had simply eroded older marsh deposits. Consequently, the Acheloos river delta prograded to the southwest with an average rate of approx. 1.5 m per year. At OIN 13, fluviodeltaic influence stopped at 1469–1626 cal AD – an age yielded by a plant remain found within coastal swamp deposits overlying the deltaic sequence (365±37 14C years, 1 sigma error range).
FIg. 5 - Holocene delta evolution shown in facies profiles along the modern Acheloos River
Source: Own research.
32Our results show clear discrepancies when compared to scenarios described in literature. Villas’ (1984: Fig. 33-35) scenarios are based on the interpretation of sedimentary data from seven cores related to the eustatic sea level curve which O. Erol had developed for the Anatolian coast. Apart from the problematic transferability, it remains unclear if reservoir effects were taken into account when processing radiocarbon data. Unfortunately, Villas’ 14C ages (Villas 1984: 199) do not allow a calibration according to modern standards which makes direct comparison difficult. However, our results show that delta progradation reached the eastern flank of Trikardo at OIN 10 at least 1000 14C years earlier than suggested by Villas. This indicates that the shifting of deltaic sedimentation to the southwest probably took place earlier than thought before. Also, for the area around OIN 12 we found deltaic influence starting earlier than assumed by Villas. Nevertheless, the two scenarios show good general accordance. At OIN 13– according to our data – delta progradation stopped at about 1550 cal AD; for this time Villas describes the area to be near the present shoreline. To summarize, our data clearly indicate that delta progradation to the southwest started earlier and was faster than proposed by Villas (1984).
33Piper & Panagos (1981: 120) emphasize that there was «rather little change» in the Acheloos delta «in the last 2300 years except for some progradation of the Acheloos to the southwest». Considering that deltaic sedimentation at OIN 12 did not take place until about 700 cal AD, we state that even within the last 1300 years considerable coastal changes did take place. Moreover, delta progradation initiated other landscape changes – for instance the silting up of lagoonal systems as shown by Vött et al. (2003a, 2003b). According to Fouache (1999) delta progradation did not stop until the 19th century when increasing human impact led to a sedimentary crisis. Modern analysis of coastal sedimentary units and radiocarbon dates clearly reject the assumptions of Philippson (1958) and Grove & Rackham (2001) who argued that delta progradation had already stopped at about 1 cal BC or during the 5th century BC respectively.
34From a historical point of view, the delta front of the Acheloos river may be located somewhere between the coring sites OIN 10 and 12 for the time when the people of Oiniadai occupied Trikardo and the shipsheds were used. At that time, Trikardo was surrounded by a large lagoonal system which guaranteed the connection of the shipsheds to the sea (Vött et al., 2003b). When the delta front reached OIN 12, Oiniadai had already been abandoned for several centuries. Nevertheless, it has to be taken into account that, initially, the Acheloos delta might have prograded southwards and reached the Ionian Sea between the islands of Skoupas and Koutsilares and afterwards turned further west towards OIN 12 and the modern river mouth (Photo 3). Furthermore, we cannot definitely exclude that the dated fluvio-deltaic layer found at OIN 12 is made up of reworked deltaic sediments of an older age.
35The presented results allow us to outline some new aspects about the progradation of the Acheloos river in a southwestern direction. In order to generate a detailed and consistent scenario for the whole delta region, more corings and radiocarbon dates are needed. A convincing figurative presentation of palaeogeographical changes during the Holocene needs a maximum of input data. Therefore, further investigations will have to focus on the question whether sediments of earlier periods of delta progradation in southeastern and southern directions – as assumed by Piper & Panagos (1981) and Villas (1984) – can be found.
Photo 3 - View of the Acheloos River near the modern river mouth north of the former island of Koutsilaris
Kefalonia, one of the Ionian Islands, is visible in the background (at a distance of ca. 37 km). (Photograph by A. Vött, August 2001).
36The evolution of the Küçük Menderes delta plain was studied with a geoarchaeological focus. Our results show that the ancient site of Ephesus was not severely affected by siltation until Archaic times. Sedimentary data indicate that the famous Artemision was erected at a former shoreline as indicated in ancient accounts. Within the last 2500 or so years the former embayment gradually silted up and forced the inhabitants of Ephesus to repeatedly relocate their harbour. There are indications of an anthropogenically related acceleration of delta growth due to landscape degradation and soil erosion in Roman times. As for Troia, the Karamenderes and Simois rivers initiated considerable coastal changes affecting the direct environs of the city as early as in the 3rd millennium BC. The former bay, once the reason for Troia’s strategic and commercial importance, was rapidly infilled with sediments. The city lost its direct connection to the sea at the latest during the middle of the 2nd millennium BC .
37From a geomorphological-palaeogeographical point of view, the first two case studies show the progradation of deltaic environments in former quiescent embayments. Distributary channels prograded into the sea far beyond the coastlines, forming typical bird’s foot deltas. The detection of this kind of sedimentary pattern within the strata of the mentioned coastal plains is a new contribution to palaeogeographical research. This approach is closely related to the study of recent deltaic sedimentation in the quiescent Gulf of Maliakos (Greece) – a good example for actualism in geomorphology.
38In contrast to the examples from Turkey, the growth of the Acheloos delta took place under completely different conditions. Bound to active distributary arms of the Acheloos, winds predominantly from the southwest caused the evolution of extended barrier beach-lagoon systems attached to the rocky islands of the Echinades. The river channels were of decisive importance to coastal changes. Here, we present results related to the southwestern distributary system and its progradation history. The former island of Trikardo, where ancient Oiniadai is situated, was first affected by fluviodeltaic sedimentation and the subsequent formation of large lagoonal systems in the 4th millennium BC. Further delta growth was faster than suggested by previous investigations and literary data. Research on several other distributary-like former river channels and their palaeogeographical significance is currently in progress. It will result in a complete scenario of Holocene coastal changes in this area.
39It is important to note that the evolution of the three deltas was not synchronous but closely connected to different phases of settlement, specific tectonic patterns (see Fouache & Dalongeville, 2003 for other Mediterranean regions) and, of course, sediment supply.