Skip to navigation – Site map
Paléorisques

Late Holocene tsunami imprint at the entrance of the Ambrakian gulf (NW Greece)

Impacts des tsunamis le long des côtes du golfe Ambracien (Grèce nord-occidentale)
Andreas Vött, Helmut Brückner, Matthias May, Franziska Lang and Svenja Brockmüller
p. 43-57

Abstracts

This paper presents geomorphological, sedimentological and geoarchaeological evidence of multiple tsunami impacts on the Aktio headland (NW Greece). Numerous vibracores revealed high energy event deposits most of which were accumulated on terrestrial sites on top of erosional unconformities. The tsunami sediments, up to 3 m thick, mostly consist of middle to coarse sand and include gravel and marine shell fragments. Both badly sorted and well sorted, laminated event deposits were observed. In several places, an intersecting palaeosol unit documents repeated tsunamigenic influence after a phase of weathering and soil formation. Underwater studies offshore Aktio headland showed that the Plaka palaeo beach ridges, including its beachrock base, out of beachrock was destroyed by tsunamigenic wave activity. Dislocated beachrock fragments with a maximum diameter of 60 cm were encountered on Aktio headland up to 2.5 km to the west of the in-situ beachrock unit in elevations up to 3.15 m a.s.l.
A series of 14C-AMS dates and diagnostic ceramic fragments found in vibracores allow to reconstruct several tsunami landfalls. Around 2870-2350 cal BC, a major event hit the entire headland. Parts of the study area were probably affected by further tsunamis around 1000 cal BC and 300 cal BC which are known to have hit the adjacent Lefkada coastal zone. A mega tsunami struck the Aktio headland around 840 cal AD. The minimum height of the tsunami surge is estimated at 6m. Further tsunami landfalls occurred during the last 700 or so years. The results document an extraordinarily high tsunami risk for the study area.

Top of page

Index terms

Geographical index :

Grèce, Golfe anbracique
Top of page

Author's notes

We thank K. Gaki-Papanastassiou, H. Maroukian, and D. Papanastassiou (Athens) for joint field surveys and intense discussion. Sincere thanks are due to L. Kolonas (Athens), M. Stravropoulou (Mesolongion), and C. Melisch (Berlin) for various support during field work. U. Ewelt and R. Grapmayer (Grünberg) carried out scuba diving at Skoupeloi Achilleos. Radiocarbon dating was accomplished by A. Scharf (Erlangen), P.M. Grootes (Kiel), and K. van der Borg (Utrecht). Work permits were issued by the Greek Institute of Geology and Mineral Exploration (Athens). We gratefully acknowledge funding of the research project by the German Research Foundation (DFG, Bonn, VO 938/2-1).

Full text

1 - Introduction

1The Ionian Sea and its coasts belong to the seismically most active regions of the Mediterranean and has been affected by many earthquakes and associated tsunamis. The high tsunami risk has decisively been revealed by the increasing number of tsunami studies which have been carried out during the last two decades. By these studies, a variety of event deposits have been identified and used to reconstruct tsunami landfall history during the Holocene (for instance Dominey-Howes, 2002; Dominey-Howeset al., 2000; Kelletat and Schellmann, 2002; Mastronuzzi and Sansò, 2004; Minouraet al., 2000; Morhangeet al., 2006; Reinhardtet al., 2006; Scheffers, 2006a; Stefatoset al., 2006).

2Geological records for multiple tsunami impacts on the coast near Lefkada were encountered during the course of interdisciplinary investigations on the Holocene evolution of coastal Akarnania in northwestern Greece (Vöttet al., 2006a, 2006b, 2006c). Different types of interrelated tsunami deposits were identified. They are part of a system of tsunami sediment traps which seem to be unique for the eastern Mediterranean (Vöttet al., 2006d, 2007).

3This paper gives evidence for tsunami landfall at Aktio headland at the entrance to the Ambrakian Gulf (NW Greece). The main objectives of our study are (i) to present geomorphological, sedimentological and geoarchaeological records of tsunamigenic activity, (ii) to establish a tsunami geochronology, and (iii) to reconstruct the flow direction and the flow intensity of palaeo tsunami waves in order to assess the present tsunami risk of the area.

Fig. 1 - Topographic overview of the coastal zone between Preveza and Lefkada and detailed map of the Aktio headland showing vibracoring sites and location of underwater studies.
Schéma topographique du secteur d’étude entre Preveza et Leucade. Croquis détaillé du cap Aktio.

Fig. 1 - Topographic overview of the coastal zone between Preveza and Lefkada and detailed map of the Aktio headland showing vibracoring sites and location of underwater studies.Schéma topographique du secteur d’étude entre Preveza et Leucade. Croquis détaillé du cap Aktio.

2 - Topography and geotectonic setting

4Aktio headland, almost 7 km long and 4 km wide, is located in the coastal zone between Preveza and Lefkada (Fig. 1). It is bordered by the Ionian Sea to the west, by the Ambrakian Gulf to the east, and by the Strait of Aktio/Preveza, only 600 m wide and 8 m deep, to the north. The headland is a low lying coastal area with maximum elevations of 6 m above present sea level (m a.s.l.) and has a slightly undulating surface. The Phoukias sand spit which extends into the Bay of Aghios Nikolaos and the Plaka are the most striking geomorphological features of the present-day coast. The Plaka represents the partly submerged ruins of a former strandline and is entirely made up of beachrock (von Seidlitz, 1927), up to 6-10 m thick. The formation of the Plaka beachrock started in the 5th millennium BC at the latest closing off the lagoons of Lefkada and Aghios Nikolaos from the open Ionian Sea (Vöttet al., 2006d). The shallow waters of the Lagoon of Saltini cover the central parts of the Aktio lowlands. In the north, the spit of Aktio bears the sparse archaeological remains of the ancient sanctuary of Apollo which was used by the Akarnanian people at least during Hellenistic times (Murray, 1982).

5According to Paschoset al. (1991) and IGME (1996), the Aktio headland is largely made up of Holocene coastal sediments and swamp deposits. The Preveza area and parts of the hills of Stoupas (Fig. 1) consist of Pleistocene sand units and Pliocene to Pleistocene marls, sandstones and conglomerates of a Flysch-like facies. In contrast, the Plaghia Peninsula, the hills between Aghios Nikolaos and Vonitsa as well as the hills east of Aghios Thomas near Preveza are composed of limestone, partly brecciated, and dolomite of Triassic to Cretaceous age. Parts of the underground are characterized by thick Triassic evaporitic units which have considerable influence on tectonodynamic (Galonopoulos and Ekonomides, 1973; Laigleet al., 2004) and coastal evolution (Underhill, 1988).

6The Ambrakian Gulf is a tectonic basin which has been formed by NE-SW directed crustal compression and NNW-SSE trending extension (Clews, 1989). The main triggering factors are pull-apart dynamics caused by the rapid southwest-ward movement of the Aegean microplate (Doutsos and Kokkalas, 2001). Moreover, northwestern Greece underwent a 40° clockwise rotation during 15-8 Ma and a 20° clockwise rotation after 4Ma (van Hinsbergenet al., 2005). Northern Akarnania even shows a 90° clockwise rotation since Oligo-Miocene times (Broadleyet al., 2004). Both rifting and rotation are responsible for the opening and spreading of the Ambrakian Gulf.

7The northernmost section of the Hellenic Arc lies not far offshore Aktio headland. As part of a multiple plate junction, the right-lateral strike slip Cefalonia transform fault (CF) and the Lefkada fault (LF) induce the high seismo-tectonic activity of the area (Cocardet al., 1999; Louvariet al., 1999). Rates of crustal motion south of the CF and LF reach up to 40mm/a whereas north of the CF, they are negligible (Kahleet al., 2000; Peteret al., 1998).

8The study area experienced a Mw = 6.2 earthquake shock on August 14, 2003, which was induced by the Lefkada fault (Karakostaset al., 2004) and triggered a small tsunami of 0.5 m south of Nidri on Lefkada Island (EERI, 2003). The Preveza-Lefkada coastal zone is strongly affected by high frequency and mid to high magnitude earthquakes (Papazachos and Papazachou, 1997) and is thus characterized by a high tsunami risk (Papazachos and Dimitriu, 1991). It is known to have been repeatedly hit by tsunami events. A regional tsunami catalogue, mainly based on historical sources analyzed by several authors, has been compiled by Vöttet al. (2006d).

3 - Materials and methods

9A series of 28 vibracores was drilled at Aktio headland in search of tsunamigenic deposits. Vibracorings were carried out by means of an Atlas Copco mk1 corer using core diameters of 6 cm and 5 cm. The maximum recovery depth in the study area was 18 m below ground surface (m b.s.). Vibracores were analyzed in the field using geomorphological and sedimentological methods. Colour, grain size, texture, macrofossil fragments and carbonate content were determined for each sediment layer. Macrofaunal remains were used to elucidate the sedimentary environment.

10Sediment samples were taken for geochemical analyses of parameters such as electrical conductivity, pH-value, loss on ignition, content of orthophosphate, as well as of the concentration of (earth-) alkaline and heavy metals. Geochemical values were helpful to detect facies changes (Vöttet al., 2003).

11Detailed geomorphological surveys were conducted on land and under water by means of scuba diving along transects across the submerged Plaka beachrock remains. Underwater findings of dislocated beachrock blocks and rubble ridges were measured, photographed, and sampled. A high-resolution differential GPS system (Leica SR 530) was used to determine the position and elevation of vibracoring sites.

12The geochronological framework of the study is based on (i) 14C-AMS analyses of organic material and shells made out of marine carbonate as well as on (ii) the age determination of diagnostic ceramic fragments encountered in vibracore profiles and during field surveys. Concerning samples from marine environments, the palaeo-reservoir effect may differ from the modern one and has possibly been variable through time and space (Geyh, 2005). This is why we refrained from determining the modern reservoir effect and corrected all marine samples for a mean marine reservoir age of 402 years (ΔR = 0; Reimer and McCormac, 2002). Calibration was carried out using the Calib 5.0.2 software.

4 - Event deposits at Aktio headland

13We present sedimentological and geomorphological evidence of multiple tsunami impacts encountered in 15 selected vibracores from the Aktio headland and during underwater studies around Skoupeloi Achilleos (Figs. 2 to 6).

4.1 - Phoukias and the Phoukias sand spit transect A

14Vibracore transect A shows a general SSE-NNW direction (Fig. 1). The transect starts at the Phoukias sand spit and, in its greatest part, runs almost parallel to the modern coastline some 200-400 m inland. Facies distribution for transect A is illustrated in Fig. 2.

Fig. 2 - Facies distribution and tsunami deposits found in vibracores from the Phoukias sand spit and Phoukias (transect A).
Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Phoukias (transect A).

Fig. 2 - Facies distribution and tsunami deposits found in vibracores from the Phoukias sand spit and Phoukias (transect A).Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Phoukias (transect A).

15The base of vibracore ANI 2 is made up of clayey silt which was deposited in the sublittoral zone of a quiescent shallow marine to lagoonal environment. These deposits are abruptly covered by fine sand (11.11-10.17 m below sea level (= m b.s.l.)). The medium to high energy influence subsequently led to a complete change of the sedimentary environment and a thick package of clearly laminated silty to fine sandy deposits with abundant sea weed remains was deposited. The latter document upgrowing generations of sea weed mats. Although wave dynamics seem to have been restricted, conditions were much more energetic compared to the preceding quiescent phase. An intersecting layer of homogeneous fine sand (6.16-5.72 m b.s.l.) may indicate another temporary shift towards higher energy conditions.

16Quiescent sedimentary conditions at the base of ANI 14 were interrupted by the input of a thick sequence of badly sorted middle sand including marine shell debris and fragments of Echinoidea and Bryozoa (9.65-7.77 m b.s.l.). Afterwards, lagoonal conditions were re-established. Intersecting layers of shell debris embedded in a silty to sandy matrix and partly rich in organic material (7.08-4.87 m b.s.l.) document, once more, abrupt high energy interferences of the quiescent environment. In the uppermost parts of profiles ANI 2 and ANI 14, silt-dominated lagoonal units are abruptly topped by coarse to middle sand reflecting another sudden increase of energetic input to the system (ANI 2: 3.27-1.22m, ANI 14: 2.88-0.36 m b.s.l.). Hereafter, ANI 2 came under shallow marine influence and sandy littoral sediments were deposited. Meanwhile, at ANI 14, a soil was formed. Lagoonal conditions were then neither re-established at ANI 2 nor at ANI 14. However, a layer of gravel in a silty to sandy matrix, 6 cm thick, which was found at ANI 14 on top of the subrecent soil (0.33-0.39 m a.s.l.), documents short-term influence from the littoral zone.

17Vibracores AKT 1-3 largely differ from the stratigraphic sequences of cores ANI 2 and ANI 14. The base of profile AKT 1 consists of weathered silty to clayey sediments. They are covered by strongly weathered silty lagoonal and fine to middle sandy littoral deposits. The upper part of the sand unit shows a brownish palaeosol the top of which is clearly eroded and overlain by grey coloured and badly sorted sand with abundant fragments of marine mollusks and pieces of gravel (1.90-1.65 m b.s.l.). Subsequently, a layer of well sorted sand follows (1.65-0.58 m b.s.l.) which, in turn, is covered by another thick stratum of badly sorted middle sand including shell fragments and gravel (0.58 m b.s.l. – 0.84 m a.s.l.). This stratum is of a brownish colour and obviously underwent weathering after it had been deposited. The following stratum of well sorted sand (0.84-1.42 m a.s.l.) is covered by aeolian deposits. Vibracore AKT 1 clearly documents the high energy input of marine sediments into a terrestrial environment and the partial erosion of the underlying pre- to mid-Holocene palaeosol. The intersecting upper palaeosol documents that the site must have been affected twice by sudden wave activity from the seaside. The total thickness of the event deposits at AKT 1 is 3.32m.

18Vibracore profile AKT 2 is similar to AKT 1. The erosional unconformity at the base of the sandy event layer (1.59 m b.s.l. – 0.50 m a.s.l.), however, is formed on top of strongly weathered silty lagoonal deposits.

19At vibracoring site AKT 3, the older generation of event deposits (0.43 m b.s.l. – 0.40 m a.s.l.), including numerous shell fragments, is partly weathered. The younger generation is characterized by shell debris and badly sorted sand which was deposited well above sea level (0.40-0.83 m a.s.l.). The uppermost part of AKT 3 is made up of aeolian fine sand.

4.2 - Underwater geomorphology at Skoupeloi Achilleos

20Skoupeloi Achilleos belongs to the northern section of the partly submerged Plaka palaeo coastline. It is located some 2.3 km to the west of Phoukias and some 2.2 km to the north of the outermost tip of the Santa Maura beach ridge (Fig. 1). Underwater studies were conducted along a WNW-ESE running transect.

21Around 40 m east of the in-situ beachrock ruin of the Plaka, we found a ridge made out of isolated beachrock fragments at 2.5-3.2 m water depth. The rubble ridge is up to 0.9 m thick. The dislocated beachrock slabs along the transect show a maximum diameter of 1.0 m and a maximum volume of 0.3m3 (Fig. 3a). Close-by the transect, we encountered individual slabs up to 2 m long (Fig. 3d). The ridge structure is completely separated from the Plaka and its components lie on top of or are partly embedded in sandy deposits. Approaching the Plaka, numerous large beachrock blocks lie in extraordinarily disordered positions completely detached from the in-situ beachrock ruin at 2.9-4.6 m water depth. Some of the blocks were obviously broken into pieces when they hit the ground. Along the scuba dive transect, the largest dislocated blocks show a maximum volume of ≈ 3m3. However, not far off the transect, we measured blocks up to ≈ 25m3. Many blocks are piled up to block clusters. In some places, we found clear imbrication (Figs. 3e and 3f). Isolated beachrock blocks larger than 1m3 were encountered up to 15-20 m to the east of the Plaka.

22Both rubble ridge and dislocated blocks are densely overgrown by marine algae and other marine organisms. This documents that the deposits are not affected by recent wave activity even during storms. Bio-erosion and bio-construction features found at the bottom sides of slabs and blocks prove dislocation of the material from its original position.

23From a palaeogeographical point of view, the lagoonal deposits encountered at the base of vibracore ANI 2 (Fig. 2) prove that the formation of the Plaka beachrock had already started and induced quiescent hydrodynamic conditions east of it. The Plaka beachrock was part of a beach ridge system which separated the lagoonal Bay of Aghios Nikolaos from the Ionian Sea. Geomorphological underwater evidence found at Skoupeloi Achilleos clearly documents high energy impact to the former coastline. The impact destroyed the beach ridge and the Plaka beachrock throwing stones, slabs and blocks in an easterly direction. We assume that the beachrock had already been broken into pieces by previous earthquake influences along the Lefkada fault. The adjacent lagoon of Aghios Nikolaos, deprived of its natural barrier, subsequently experienced increasing sea water influence and wave activity. The findings at Skoupeloi Achilleos correspond well to dislocated beachrock units described from the Santa Maura beach ridge system and the southern part of the Plaka by Vött et al. (2006d, 2007).

Fig. 3 - Geomorphological underwater studies at Skoupeloi Achilleos to the west of the Phoukias sand spit.
Vues sous-marines de blocs démantelés de beach-rock, secteur de Skoupeloi Achilleos.

Fig. 3 - Geomorphological underwater studies at Skoupeloi Achilleos to the west of the Phoukias sand spit.Vues sous-marines de blocs démantelés de beach-rock, secteur de Skoupeloi Achilleos.

(a), (b), (d): Isolated rubble ridge made up of beachrock fragments. Beachrock stones and slabs originate from the partly submerged Plaka ruin. They were dislocated by tsunamigenic wave action and moved over a distance of 40m. Beachrock fragments are partly embedded in or lie on top of sandy deposits. (c), (e), (f): Beachrock slabs and blocks some 15-20m east of the in-situ Plaka beachrock. One slab sticks vertically between dislocated blocks. Some of the blocks show imbrication. All slabs and blocks are detached from the in-situ beachrock and partly embedded in a sandy substrate. Body length of snorkeler is around 1.90m. Photos taken by R. Grapmayer, 2006.

4.3 - Paliokoulio and Koumaros - transect B

24Transect B trends in a SSE-NNW direction across the southern outlet of the Lagoon of Saltini (Fig. 1). Vibracores AKT 9-12 were drilled within a distance of 30 m along a line running from the top of a natural ridge-like topographic unit across its northern flank towards the southern shore of the Saltini channel. Vibracores AKT 23 and AKT 24 are located on the opposite shore of the lagoon.

25The lower part of profile AKT 9 is a gleyic palaeosol comprising fine sandy to predominantly silty deposits and is characterized by iron oxide spots down to 0.66 m b.s.l. (Fig. 4). On top of an erosional unconformity, we found a layer of badly sorted sand rich in fragments of a marine macrofauna and gravel (0.09-0.61 m a.s.l.) which also includes ceramic fragments. The subsequent stratum has a similar texture (0.30-0.61 m a.s.l.) and contains gravel and potsherds. However, it is strongly weathered and thus void of carbonate. The top of the core is made up of weathered aeolian fine sand.

26The facies pattern encountered at site AKT 10 is almost identical to the one described for AKT 9. However, the lower unweathered unit of coarse grained deposits (0.08-0.17 m a.s.l.) mainly consists of gravel and includes fragments of beachrock, up to 6 cm large. The erosional discordance is even clearer than at AKT 9. The upper coarse grained unit is almost 30 cm thick (0.17-0.46 m a.s.l.). We found ceramic fragments in these two units (Fig. 4). The top of profile AKT 10 is made up of weathered dune sand.

Fig. 4 - Facies distribution and tsunami deposits found in vibracores from Paliokoulio and Koumaros (transect B).
Chronostratigraphie des carottes et dépôts de tsunami des carottes des secteurs de Paliokoulio et de Koumaros (transect B).

Fig. 4 - Facies distribution and tsunami deposits found in vibracores from Paliokoulio and Koumaros (transect B).Chronostratigraphie des carottes et dépôts de tsunami des carottes des secteurs de Paliokoulio et de Koumaros (transect B).

27At AKT 11, the sandy layer (0.23-0.12 m b.s.l.) on top of the palaeosol is entirely weathered and void of fossil remains out of carbonate. It is covered by clayey silt which was deposited under limnic to brackish conditions and which, at its top, forms a dark palaeosol. The palaeosol, in turn, is abruptly covered by another sand stratum (0.16-0.29 m a.s.l.) which includes ceramic fragments.

28The basal palaeosol of vibracore AKT 12 shows spots of iron oxide down to 0.55 m b.s.l. The subsequent layer consists of coarse sand and gravel (0.54-0.43 m b.s.l.) with embedded remains of a marine macrofauna. Another layer of marine sand, laminated and comparatively well sorted, appears on top of intersecting lagoonal mud and is also rich in marine fossils (0.38-0.10 m b.s.l.). The upper part of the profile is made up of clayey to silty marsh deposits.

29Vibracores AKT 9-12 clearly document that sediments from marine environments were thrown onto near-coast terrestrial sites by high energy impact. The depositional event caused an erosional unconformity. The event deposits, according to their unsorted nature and the included fragments of marine fossils, seem to be mixed up and consist of material from different marine environments. According to their weathered appearance, most of the event layers were deposited above sea level at the time of deposition. Profiles AKT 11 and AKT 12 were affected twice by a sedimentary event.

30The base of vibracore AKT 23 is made up of cemented fine sand possibly identical to the aeolianite of Pleistocene age described by Paschoset al. (1991) from the Preveza area. It is covered by littoral sand and subsequent lagoonal mud. In its upper part, the mud, according to the decreasing content of marine macrofossils, reflects increasing influence by freshwater. The following limnic sediments were cut by an erosional unconformity and covered by mean sand (2.73-1.98 m b.s.l.) which, partly weathered (1.98-0.05 m b.s.l.), documents sudden and abrupt high energy influence at the site. On top of it, we found coarse sand (0.05 m b.s.l. – 0.20 m a.s.l.), including fragments of Cerastoderma glaucum and Pecten sp., which is suggested to corresponds to a younger event generation.

31The facies pattern encountered at AKT 24, drilled some 50 m inland, is almost identical to the one of AKT 23. However, the erosional unconformity was formed in a peat layer rich in gastropod fragments. Moreover, the overlying stratum of coarse sand (1.61-1.15 m b.s.l.) includes fragments of a marine macrofauna. Young generation event deposits were not found at AKT 24. In contrast, the top of the profile is made up of weathered aeolian sand.

32Vibracores AKT 23 and AKT 24 confirm the twofold landfall of strong wave events. The comparatively deep position of the coarse grained event deposits however indicate that the area (i) has been affected by coastal subsidence along NE-SW running faults (Paschoset al., 1991), (ii) represents a pre-existing topographic low and/or (iii) reflects a deep erosion channel. We found spots of iron oxide down to 1.20 m b.s.l. at AKT 23 and 1.60 m b.s.l at AKT 24 which indicates that subaerial weathering of the event deposits took place for a certain time period after deposition.

Fig. 5 - Facies distribution of vibracore AKT 14 from the northeastern shore of the Lagoon of Saltini.
Photographie des faciès de la carotte AKT 14 (transect B).

Fig. 5 - Facies distribution of vibracore AKT 14 from the northeastern shore of the Lagoon of Saltini.Photographie des faciès de la carotte AKT 14 (transect B).

The autochthonous limnic environment was abruptly affected by tsunami waves which left several shell debris layers. Gaps are due to rodding processes. The top/up direction for the core segments is to the left. Photo taken by M. May, 2006.

4.4 - Northeastern shore of the Lagoon of Saltini

33Several vibracores were drilled at the northeastern shore of the Lagoon of Saltini east of the road between Vonitsa and Preveza (Fig. 1). In this area, the lagoon has almost completely desiccated. Fig. 5 exemplarily illustrates vibracore AKT 14 (ground surface at 0.20 m a.s.l.) which was drilled some 30 m from the shore of the Ambrakian Gulf (Fig. 1). The lower part of the profile shows brownish grey, clayey to silty deposits (4.80-3.20 m b.s.l.) which – due to numerous spots of iron oxide – were accumulated in an ephemeral shallow lake. Several indeterminable ceramic fragments found at 3.37 m b.s.l. indicate a mid- to late Holocene age. The following deposits are of similar grain size and texture (3.20-1.44 m b.s.l.). Their grey colour, however, reflects a permanent limnic environment. An intersecting peat layer documents temporary semi-terrestrial conditions (2.43-2.30 m b.s.l.).

34Numerous fragments of marine shells, partly in the form of shell debris layers, abruptly appear in the subsequent layer (1.44-0.96 m b.s.l.) which consists of clayey silt and was accumulated on top of a slight erosional unconformity. The deposits also include wood remains and a lot of organic matter. However, subsequent brownish silty clay void of marine shell fragments (1.09-0.70 m b.s.l.) documents that the water body came again under prevailing freshwater influence and partly dried up. The uppermost part of the profile is made up of similar deposits including some fragments of Cerastoderma glaucum which indicate slight saltwater influence.

35Vibracore AKT 14 documents the sudden and temporary influence of saltwater on a near-coast limnic environment typical of an extreme marine event.

4.5 - Vasiliko - transect C

36Vibracores AKT 20-22 were drilled in the northern part of Aktio headland (Fig. 1). They are arranged in the W-E running transect C (Fig. 6) across an erosional terrace facing the low lying coastal plain.

Fig. 6 - Facies distribution and tsunami deposits found in vibracores from Vasiliko (transect C).
Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Vassiliko (transect C).

Fig. 6 - Facies distribution and tsunami deposits found in vibracores from Vasiliko (transect C).Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Vassiliko (transect C).

37The base of profile AKT 20 is made up of clay deposited in a limnic environment. An intersecting shell debris layer (1.81-1.61 m b.s.l.) indicates temporary influence from the seaside. The top of the limnic deposits is a clear erosional surface followed by a thick stratum of middle and coarse sand (0.92-0.12 m b.s.l.). The sand layer includes gravel and abundant marine shell fragments such as from Cerastoderma glaucum, Cerithium sp., Pecten sp., Tellina sp., and Echinoidea. It is covered by well sorted middle to fine sand (0.12 m b.s.l. – 0.22 m a.s.l.). The uppermost part of profile AKT 20 is a strongly weathered soil (0.22-0.74 m a.s.l.) out of the same substrate. The sandy units encountered at AKT 20 contain some indeterminable ceramic fragments.

38Basal lake deposits found at AKT 21 are abruptly covered by strongly weathered middle and coarse sand (1.75-1.20 m b.s.l.) including pieces of gravel and a chert artefact. Provided that the material is of marine origin, it may have been deposited during an extreme event. The subsequent thick stratum of well sorted brown to light brown fine sand probably represents weathered aeolian deposits. This dune palaeosol shows an erosional discordance on top of which strongly unsorted sand and gravels were deposited (0.10-0.65 m a.s.l.). The upper section of the core consists of deeply weathered and comparatively well sorted sand (0.65-1.50 m a.s.l.). Several indefinable ceramic fragments were found in the sand unit on top of the erosional surface.

39The base of vibracore AKT 22 shows weathered sand and gravel followed by a thick brownish red palaeosol. Another palaeosol comprising deeply weathered reddish brown silty to clayey sand (2.34-2.68 m a.s.l.) follows. It is covered by a thin sand stratum rich in carbonate and beachrock fragments (2.68-2.73 m a.s.l.) which is partly weathered (2.73-3.15 m a.s.l.). Above 2.19 m a.s.l., we found a lot of ceramic fragments and, on top of the terrain surface, abundant stones and blocks out of beachrock, associated with numerous large potsherds in different states of conservation and roundness. The beachrock fragments presumably originate from the submerged Plaka several kilometers to the west.

40Vibracores AKT 20-22 document the sudden input of marine deposits in both limnic and terrestrial environments. The presence of erosional unconformities as well as the composition, the grain size distribution, and the degree of sorting of the overlying sand and gravel units indicate that an extreme event hit the former shoreline. Vibracore transect C shows that the coast at that time lay to the west of AKT 20 (Figs. 1 and 6).

41We also conducted geomorphological surveys in the southern part of Vasiliko. Along the shore of the Ambrakian Gulf, some 0.9 km to the southeast of transect C, we found pebbles, up to 10-15 cm in diameter, with boreholes and living in-situ specimens of marine boring mussels. Some of the pebbles were covered by barnacles (Figs. 7a and 7b). Recent marine bio-erosion and bio-construction demonstrate that the pebbles are barely moved by modern wave activity. We thus suggest that the coarse beach material is allochthonous and was dumped by a sedimentary event. Colonization by marine organisms occurred post-depositionally.

42Further inland – some 0.6 km to the south of transect C, around 1 km east of the open sea and 2.5 km east of the nearest in-situ beachrock off cape Skilla – the terrain surface is strewn with large stones and blocks. These stones and blocks are also of allochthonous origin. We found (i) breccias including chert fragments and beach pebbles (Fig. 7c), (ii) calcareous sandstones, possibly sandy beachrock, with boreholes from marine boring organisms (Fig. 7d), and (iii) beachrock slabs, up to 60 cm in diameter, comprising gravel, with signs of bio-erosion (Figs. 7e and 7f). These findings support the idea of a high energy impact on the Aktio headland by a strong marine event approaching from a westerly direction.

Fig. 7 - Geomorphological indicators of a tsunami imprint found at Vasiliko.
Dépôts de tsunami, secteur de Vassiliko.

Fig. 7 - Geomorphological indicators of a tsunami imprint found at Vasiliko.Dépôts de tsunami, secteur de Vassiliko.

(a), (b): recent shore of the Ambrakian Gulf. Large pebbles colonized by boring mussels and barnacles indicate that the beach material is not moved by modern wave action. (c)-(f): Findings of large stones and blocks scattered on top of the terrain surface in the southern part of Vasiliko including a breccia with cemented beach gravel (c), a limestone block with boreholes from marine boring organisms (d), and dislocated beachrock slabs (e, f). Photos taken by A. Vött, 2006.

5 - Ages derived from ceramic findings and radiocarbon data

43Ceramic fragments were found in a number of vibracores. In most cases, age determination was difficult due to the small sizes of the fragments or because they were strongly weathered. Diagnostic ceramic fragments used for relative dating are depicted in Figs. 4 and 6 within the stratigraphic context.

44Radiocarbon dating of samples taken from event deposits is problematic as they may be reworked. 14C-dating thus results in a maximum age or a simple terminus ad or post quem for the wave impact. Provided that sufficient and appropriate material for dating was found, we preferred sampling the underlying and/or covering sedimentary units (sandwich dating). However, an erosional unconformity at the base of an event layer may represent a considerable hiatus. Moreover, datable material is missing in cases where event deposits were accumulated above sea level and weathering started immediately after deposition.

45Altogether, 14 samples from vibracores from Aktio headland were 14C-AMS analysed (Table 1). Samples AKT 10/4+ PR and AKT 12/3 PR yielded modern ages and were not  considered for further interpretation. It is assumed that the samples were parts of (sub-)recent plant roots.

Table 1. Radiocarbon dating results for selected samples from Actio headland.
Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Vassiliko (transect C).

Table 1. Radiocarbon dating results for selected samples from Actio headland.Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Vassiliko (transect C).

Note: a.s.l. – above sea level; b.s. – below ground surface; b.s.l. – below sea level; artic. spec. – articulated specimen; Mytilus gall. – Mytilus galloprovincialis; * – marine reservoir correction with 402 years of reservoir age; 1σ max; min cal BP/BC (AD) – calibrated ages, 1σ-range; “;” – there are several possible age intervals because of multiple intersections with the calibration curve; Lab. No. – laboratory number, University of Erlangen-Nürnberg (Erl), University of Kiel (Kia), University of Utrecht (UtC).

6 - Discussion

6.1 - Tsunami sediments versus storm or sea level high stand deposits

46Based on sedimentological characteristics of vibracores, on the geomorphological underwater findings at Skoupeloi Achilleos and on geoarchaeological records at Paliokoulio and Vasiliko, we assume that the high energy event deposits encountered at Aktio headland represent tsunamigenic sediments. This interpretation goes well with historical accounts and tsunami catalogues which document that the region has been repeatedly affected by tsunamis. Moreover, the data presented fit with the scenario of multiple tsunami imprints on the Lefkada coastal zone and the Lake of Voulkaria reported by Vöttet al. (2006d, 2007).

47At Aktio headland, tsunami deposits were encountered between several meters below and up to 3.15 m above present sea level (Fig. 6). In some cases, they were completely weathered, in other cases they were intersected by a palaeosol. The latter indicates that the headland experienced at least two different tsunami landfalls. The maximum thickness of tsunami deposits was found at AKT 1measuring 3.32 m (Fig. 2).

48Considerable parts of the tsunamigenic sediments at Aktio headland lie well above present sea level (Figs. 2, 4 and 6). The following arguments contradict their interpretation as storm or sea level high stand deposits.

49(i) Two different types of tsunami deposits were found. One type consists of poorly sorted middle and coarse sand and includes gravel as well as abundant shell fragments. It corresponds perfectly with deposits accumulated by the 1992 tsunami in Indonesia described by Shiet al. (1995). The second type is made up of well sorted sand which is, in part, clearly laminated and may also contain fragments of marine fossils. The first type represents bi- or even multi-modal deposits and documents turbulent transportation and more or less chaotic sediment accumulation possibly induced by runup or backwash vortices. Bimodal deposits are regarded, by some researchers, as typical of tsunami deposits (Scheffers and Kelletat, 2004; Scheffers, 2005, 2006b). The second type, due to its more homogeneous texture and lamination, is the result of a laminar water flow. Laminar flow dynamics occur both during runup and backwash of tsunami waters as reported, for instance, from the December 26, 2004 tsunami in southeast Asia (Richmond et al., 2006). Both types of event deposits found at Aktio headland therefore show characteristic features of tsunami events. The fact that they can be found in one and the same sedimentary sequences indicates that the site was affected by both turbulent and laminar flow probably during the same event. Storm deposits, however, produce more or less well sorted sand or gravel sheets fining landwards (Schäfer, 2005).

50(ii) Most of the event deposits encountered at Aktio headland lie on top of an erosional surface (Fig. 8). Vibracore transects A, B and C reveal that the elevations of erosional discordances are different and partly even lie above present sea level (AKT 9, AKT 10, AKT 21). Moreover, they seem to increase in a landward direction (Figs. 2, 4 and 6). Similar unconformities at the base of tsunami deposits are described by Shiet al. (1995) from Flores Island and by Lavigneet al. (2006) from Banda Aceh, both in Indonesia (see also Dawson and Shi, 2000). Storm-borne sand sheets which reach further inland, in most cases concordantly overlie the pre-existing topography (Schäfer, 2005).

51(iii) Event deposits at Aktio headland were either thrown onto terrestrial sites and lie on top of a palaeosol (Figs. 2, 4 and 6) or into shallow limnic or lagoonal water bodies (Figs. 4 and 6). Both types of environment do not show any signs of earlier, pre-event as well as later, post-event influence from the seaside. Onshore and near-shore environments are usually affected repeatedly by storm activity (Reineck and Singh, 1980; Einsele, 2000).

52(iv)Pirazzoliet al. (1994) radiocarbon dated emerged beachrock near Cape Gyrapetra and deduced a 0.5-1.0 m a.s.l. relative sea level high stand during the 4th and 3rd millennia BC. Considering the well-known mole of the Corinthians at the southern entrance to the Lagoon of Lefkada which has been submerged by around 2.5 m since the 5th century BC, the authors concluded that the Lefkada area has experienced tectonic up- and down movements during the Holocene. Similar yo-yo dynamics were already suggested by von Seidlitz (1927). However, according to recent findings of tsunamigenically uplifted and dislocated beachrock mega blocks along the northern coast of Lefkada (Vöttet al., 2006d, 2007), we assume that the beachrock unit sampled by Pirazzoliet al. (1994) and the beachrock to which von Seidlitz (1927) refers is not in-situ but probably relocated and uplifted by tsunamigenic influence. Moreover, it cannot be excluded that the tsunami events which hit the coast at different times during the Holocene went hand in hand with co-seismic displacements of the coastal zone.

53Recent sea level studies along the Akarnanian coast, including the Palairos area (Fig. 1), revealed that during the Holocene relative sea level has never been higher than today (Vött, 2007). Further research is required to check if the relative sea level evolution between Preveza and Lefkada shows a different pattern. However, the fact that tsunami deposits at Aktio headland bear signs of relic subaerial weathering – such as spots of iron oxide – down to a minimum depth of 1 m b.s.l. indicates that the relative sea level must have been considerably lower at or shortly after the time of their deposition and that the area has subsequently experienced a considerable sea level rise. Weathering of the event deposits and soil formation seem to have begun immediately after deposition and the sites have never been affected again by sediment input from the seaside.

54(v) It is impossible to explain the Aktio headland event deposits by abrupt co-seismic submergence even if it occurred in the order of several meters. Firstly, co-seismic subsidence is not known to cause strong water currents towards the coast. Secondly, it is not able to activate large sediment masses and transport them in a landward direction. The elevation of the event layers above present sea level would furthermore require subsequent coastal uplift.

55(vi) Both subaerial (Vött et al., 2006d) and underwater findings of dislocated mega blocks along the Plaka prove tsunamigenic impact on the coast. Hydrodynamic calculations show that a block of 6m3 requires storm waves higher than 30 m to be transported. The same block may be dislocated by tsunami surge, only 11 m high (Nott, 2003; Bartel and Kelletat, 2003).

56We conclude that Aktio headland experienced at least two strong tsunami landfalls approaching from the west. Tsunami waves caused strong erosion at the western coast of the headland as well as in its northern part and left thick sequences of coarse grained deposits. Geological findings from the eastern part of the headland around Vasiliko and the Lagoon of Saltini, however, prove that the Aktio headland was entirely inundated by tsunami waters which then flowed into the Ambrakian Gulf. Furthermore, vibracore profiles of transect B (Fig. 4) show that the Lagoon of Saltini has not been created by longshore sand transport and related beach accretion but seems to be the result of strong tsunamigenic erosion in a SSW-NNE direction. The resulting basin was probably submerged due to ongoing relative sea level rise by tectonic subsidence along active faults.

57At vibracoring site AKT 22, tsunami deposits were found up to 3.15 m a.s.l. (Fig. 6). Considering the large blocks found at Vasiliko, it is assumed that the tsunami surge reached at least 5 m above present sea level. If we take into account that the relative sea level at the time of deposition of the tsunami sediments was at least 1 m lower than today, this results in a minimum height of inundation of 6m.

Fig. 8 - Examples of erosional unconformities due to tsunamigenic impact encountered in vibracores from Phoukias (AKT 1), Paliokoulio (AKT 10), Koumaros (AKT 24), and Vasiliko (AKT 20). Photos taken by M. May, 2006.
Photographies des carottes AKT 1, AKT 10, AKT 24 et AKT 20.

Fig. 8 - Examples of erosional unconformities due to tsunamigenic impact encountered in vibracores from Phoukias (AKT 1), Paliokoulio (AKT 10), Koumaros (AKT 24), and Vasiliko (AKT 20). Photos taken by M. May, 2006.Photographies des carottes AKT 1, AKT 10, AKT 24 et AKT 20.

6.2 - Geochronology of tsunami landfalls at Aktio headland

58Aktio headland plays a key role in understanding the Holocene evolution of the Ambrakian Gulf. Previous studies conducted in the area concentrated on the analysis of underwater sediments by means of seismic reflection profiles (Pouloset al., 1995; Kapsimaliset al., 2005) or on the reconstruction of Holocene palaeoenvironments and the progradation history of the Louros and Arachthos River deltas based on sediment cores (Jing and Rapp, 2003; Pouloset al., 2005). Tsunamigenic deposits were not identified within the framework of these studies. However, Piperet al. (1982, 1988) found isolated sandy deposits at the sea bottom in the Strait of Aktioand Tziavos (1997) reports on a sequence of “recent terrigenous silty sand”, 11-28 m thick, which he encountered in underwater cores from the same area. Considering that, to the west of Preveza, the longshore drift is limited due to prevailing rocky coasts and that there is no significant fluvial input of sand, it is difficult to explain these deposits merely by outflow dynamics in the Strait of Aktio. We suggest that they represent tsunamigenic sediments, partly reworked, which probably correlate to the event deposits encountered at the Aktio headland.

59The geochronology of Holocene tsunami landfalls presented in this paper is based on ten radiocarbon dates. At least two distinctive clusters of 14C-AMS ages were recognized, one around 2870-2350 cal BC and another around 840 cal AD (Table 1).

6.2.1 - The mid-Holocene environment and the 2870-2350 cal BC tsunami event

60An articulated specimen of Mytilus galloprovincialis taken from lagoonal deposits at the base of vibracore ANI 14 shows an age of 5373-5280 cal BC (ANI 14/25 M, 11.16 m b.s.l.). This documents that the formation of the Plaka beach ridge and beachrock west of Phoukias, which induced quiescent hydrodynamic conditions in the Bay of Aghios Nikolaos around ANI 14, had already started in the 6th millennium BC, i.e. some 550 years earlier than further south around the Santa Maura beach ridge (Vöttet al., 2006d).

61The oldest dated tsunami which affected the Lagoon of Aghios Nikolaos and its shores occurred in the 3rd millennium BC. Indeterminable plant remains found below the base of the tsunami deposits at vibracore AKT 2 date to 2872-2679 cal BC (AKT 2/9 PR, 1.60 m b.s.l.). As vibracores AKT 1-3 show almost identical stratigraphies (Fig. 2) and are drilled within a short distance (Fig. 1), the event layers encountered at AKT 1 and AKT 3 seem to be of the same age. The underlying deposits experienced strong weathering before they were affected by tsunamigenic influence. They presumably date to Eemian/Tyrrhenian times. At AKT 2 and AKT 3, seawater influenced sediments of Eemian/Tyrrhenian age are in a slightly higher position than those found at AKT 23 and AKT 24 (Fig. 4) which indicates post-sedimentary subsidence of the Koumaros area relative to Paliokoulio and Phoukias (Figs. 1 and 2) possibly along a SW-NE running fault system (Paschoset al., 1991). However, vibracores AKT 1-3 prove that around 2872-2679 cal BC, a tsunami impact eroded palaeosols and covered them by thick event layers (Section 4.1).

62A wood fragment from the very base of the tsunami deposits at AKT 14 at the northeastern shore of the Lagoon of Saltini yielded an age of 2858-2623 cal BC (AKT 14/4+ PR, 1.44 m b.s.l.), which is almost identical to the age found for the event layer at site AKT 2. This suggests that, around that time, tsunamigenic inundation affected the entire Aktio headland between Koumaros and the shore of the Ambrakian Gulf. Unidentified plant remains from vibracore AKT 20 at Vasiliko – possibly roots from plants growing on a higher surface – were dated to 2476-2346 cal BC (AKT 20/10 PR, 1.17 m b.s.l.). The sample was taken out of slightly weathered limnic sediments shortly below an erosional unconformity and a subsequent event layer (Section 4.5). It seems that the erosional surfaces at AKT 20 and AKT 21 were produced by the same tsunami impact which influenced AKT 1-3 and AKT 14 in the 3rd millennium BC. Further datings will have to clarify the difference in age. Sedimentological and geochronological data from Phoukias, Vasiliko, and the northeastern shore of the Lagoon of Saltini thus indicate a major tsunami landfall at 2870-2350 cal BC which caused considerable erosion along the western flank of the Aktio headland.

6.2.2 - Intermediary events

63An articulated specimen of Tellina planata taken from the lower part of the badly sorted tsunami sediments of vibracore AKT 20 revealed an age of 1326-1209 cal BC (AKT 20/7 M, 0.74 m b.s.l.). This is a mere terminus ad or post quem which documents that the area was also affected by a younger extreme event, possibly by the one which hit the Bay of Aghios Nikolaos around 1000 cal BC and which brought marine sediments into the Lake Voulkaria freshwater environment (Vöttet al., 2006d).

64Further ages were obtained for samples from the Phoukias sand spit. A sequence of shell debris layers (7.08-4.87 m b.s.l.) encountered in the middle section of vibracore ANI 14 is probably of tsunamigenic origin. Radiocarbon sample ANI 14/11+ PR (4.88 m b.s.l.) yielded an age of 143-230 cal AD and post-dates tsunami-borne influences at the site. Vöttet al. (2006d) detected tsunamigenic impact on the nearby Lefkada coastal zone around 300 cal BC documented by tsunamigenic washover fans east of the Santa Maura beach ridge. One of the mentioned shell debris layers of core ANI 14 therefore may be attributed to the 300 cal BC or even to an unknown younger event.

65There are no usable radiocarbon dates to estimate the age of tsunami sediments at Paliokoulio. However, the youngest ceramic fragments found in vibracores AKT 9 and AKT 10 date to ancient times (7th century BC to 4th century AD) resulting in a terminus ad or post quem for the event. At site AKT 11, potsherds associated with the older generation of tsunami deposits yield a Hellenistic to Roman age (1st century BC to 4th century AD) as another terminus ad or post quem. Deposits of the younger tsunami generation are related to fragments dating to Classical-Hellenistic times (5th to 1st centuries BC). Obviously, these fragments were reworked by tsunami wave action. We assume that Paliokoulio was affected by the 300 cal BC event detected by Vöttet al. (2006d) in the area east of Santa Maura (Fig. 1). Younger tsunami deposits are probably associated to the 840 cal AD event.

6.2.3 - The 840 cal AD and (sub-)recent events

66The lower section of vibracore ANI 2 documents tsunamigenic interference of a lagoonal environment by the abrupt input of fine sand including marine shells and plant remains (11.11‑10.17 m b.s.l., Section 4.1, Fig. 2). The tsunami impact induced a severe facies change from quiescent lagoonal to higher energy shallow marine conditions characterized by gradually upgrowing mats of sea weed. Sea weed remains from shortly above the event layer post-date the tsunami to 849-966 cal AD (ANI 2/16++ PR, 10.00 m b.s.l.). At almost the same time, the lagoonal environment at site ANI 14 experienced the sudden input of a thick package of well sorted middle to coarse sand. Sea weed remains from the uppermost part of the lagoonal unit pre-date the impact to 741-838 cal AD (ANI 14/7+ PR, 2.89 m b.s.l.). It is concluded that a major tsunami affected the Phoukias sand spit area between 741-838 cal AD and 849-966 cal AD and induced a general change of sedimentary conditions in the area east of the former Plaka strandline. The Plaka beach ridge system which, up till then, closed off the Aghios Nikolaos lagoon from the open Ionian Sea obviously sustained severe damage by the tsunami impact around 840 cal AD. It is concluded that previous tsunami events only partly affected the palaeo beach along the Plaka. Based on sedimentological and geochronological data found for ANI 2, the 840 cal AD event, however, seemed to have completely flushed westwards the loose Plaka beach material and destroyed the solid beachrock base of the system. Rubble ridges and dislocated mega blocks studied at Skoupeloi Achilleos (Fig. 3) as well as dislocated beachrock material encountered along the southern part of the Plaka (Vöttet al., 2006d) are most probably due to the same tsunami event.

67Undetermined plant remains from the base of weathered tsunami deposits found at AKT 23 yielded an age of 726‑867 cal AD (AKT 23/6+ PR, 1.52 m b.s.l.). In fact, this is a mere maximum age or terminus ad or post quem for the tsunami as the event may have reworked older material. However, sedimentological and geochronological evidence of the 840 cal AD impact from the nearby Phoukias sand spit suggests that the thick tsunami deposits at AKT 23 and AKT 24 were accumulated by the same event.

68Ceramic findings in vibracore AKT 22 document that the coast around Vasiliko was hit by a tsunami during or after Byzantine times (7th to 15th centuries AD, Fig. 6). This event may possibly also correspond to the 840 cal AD impact. Multiple impacts of Vasiliko by tsunami waves at 2870-2350 cal BC, around 1000 cal BC and probably at 840 cal AD possibly explains why there are almost no remains left of the ancient sanctuary of Apollo (Murray, 1982). Collignon (1886) reports on multiple destruction of the sanctuary by unknown invaders. Based on our results, we instead suggest that these destructions are, at least partly, due to repeated tsunamigenic influences.

69At Phoukias, samples from vibracore ANI 2 indicate that younger tsunami events occurred shortly after 1325-1405 cal AD (ANI 2/12+ PR, 6.25 m b.s.l.) and around 1688-1926 cal AD (ANI 2/7+ PR, 6.25 m b.s.l., Fig. 2). The tsunamigenic gravel layer encountered in the uppermost part of vibracore ANI 14 (0.39-0.50 m a.s.l.) seems to have been deposited during one of these younger events. The age of the younger generation of tsunami deposits found at sites AKT 1-3 remains unclear. The degree of weathering and the thickness of the overlying dune sands, however, indicate that the event layers are possibly related to the tsunami landfall at 840 cal AD.

7 - Conclusions

70Detailed geomorphological, sedimentological and geochemical analyses of vibracores, geoarchaeological studies and geomorphological underwater surveys give evidence of strong tsunamigenic impact on Aktio headland. Based on 14C‑AMS dates and archaeological age determination of ceramic fragments our results reveal multiple tsunamis which hit the coast between Preveza and the Bay of Aghios Nikolaos during the late Holocene. The following conclusions can be made.

71(i) Around 2870-2350 cal BC, a major tsunami hit and inundated the entire Aktio headland between Phoukias and Vasiliko. The tsunami water flow seem to have eroded a SW-NE running wide channel-like basin east of Koumaros which, due to subsequent relative sea level rise, was filled by the lagoonal waters of the Lagoon of Saltini. Strong erosional unconformities induced by the 2870-2350 cal BC event were found both in the southern- and in the northernmost parts of the headland at Phoukias and Vasiliko.

72(ii) Parts of the Aktio headland were affected by tsunamis which struck the coast northeast of Lefkada and which were identified in a former study (Vöttet al., 2006d, 2007). The northernmost part of the Aktio headland around Vasiliko was possibly hit by the tsunami which intruded into the Lake Voulkaria around 1000 cal BC. The areas around Phoukias and Paliokoulio were probably influenced by a tsunami around 300cal BC which produced large washover fans northeast of Lefkada.

73(iii) A strong tsunami landfall occurred around 840 cal AD and hit the entire Aktio headland. At Koumaros, the tsunami caused considerable erosion and left thick layers of event deposits. Sedimentological and geochronological data indicate that the 840 cal AD event destroyed the former Plaka beach ridge and its beachrock base and thus exposed the former lagoonal environment of the Bay of Aghios Nikolaos to open sea wave dynamics.It is further assumed that it was the 840 cal AD tsunami which (a) formed underwater rubble ridges and dislocated mega blocks at Skoupeloi Achilleos and all along the Plaka and (b) scattered beachrock blocks and material from the littoral zone on top of the terrain surface at Vasiliko up to 3.15 m a.s.l. We estimate the distance of transport of this material at minimum 2.5 km and the minimum elevation of the tsunami surge at 6m.

74(iv) Further tsunami impacts have occurred during the past 700 or so years but further field data and alternative dating techniques are needed to reconstruct and precisely date these events.

75(v) Multiple tsunamigenic influences during the late Holocene show that Aktio headland is exposed to a high tsunami risk. The adjacent NATO airport and nearby centres of mass tourism indicate the high vulnerability of the area. The sensitivity of the Preveza-Lefkada coastal zone to tsunami events is due to (a) the funnel-like shape of the shoreline which amplifies tsunami waves approaching the coast and (b) the availability of excellent tsunami sediment traps.

Top of page

Bibliography

Bartel P. and Kelletat D., (2003), Erster Nachweis holozäner Tsunamis im westlichen Mittelmeergebiet (Mallorca, Spanien) mit einem Vergleich von Tsunami- und Sturmwellenwirkung auf Festgesteinsküsten, in Daschkeit A. and Sterr H., (eds.), Aktuelle Ergebnisse der Küstenforschung. 20. AMK-Tagung Kiel, 30.05.-01.06.2003. Berichte, Forschungs- und Technologiezentrum Westküste, 28, Büsum, p. 93-107.

Broadley L., Platzman E., Platt J., Matthews S., (2004), Palaeomagnetism and tectonic evolution of the Ionian thrust belt, NW Greece, in Chatzipetros, A. and Pavlides, S., (eds.), Proceedings of the 5th International Symposium on Eastern Mediterranean Geology, 14-20 April 2004, Thessaloniki. Extended Abstracts Volume plus CD-ROM, Volume II, Thessaloniki, p. 973-976.

Clews J.-E., (1989), Structural controls on basin evolution: Neogene to Quaternary of the Ionian zone, western Greece, Journal of the Geological Society, 146, London, p. 447-457.

Cocard M., Kahle H.-G., Peter Y., Geiger A., Veis G., Felekis S., Paradissis D., Billiris H., (1999), New constraints on the rapid crustal motion of the Aegean region: recent results inferred from GPS measurements (1993-1998) across the West Hellenic Arc, Greece, Earth and Planetary Science Letters, 172, p. 39-47.

Collignon M., (1886), Torses archaiques en marbre provenant d’Actium, Gazette Archéologique, 11, p. 235-243.

Dawson A.-G., Shi S., (2000), Tsunami deposits, Pure Applied Geophysics 157, p. 875-897.

Dominey-Howes D.-T.-M., (2002), Documentary and Geological Records of Tsunamis in the Aegean Sea Region of Greece and their Potential Value to Risk Assessment and Disaster Management, Natural Hazards, 25, p. 195-224.

Dominey-Howes D.-T.-M., Cundy A., Croudace I., (2000), High energy marine flood deposits on Astypalaea Island, Greece: possible evidence for the AD 1956 southern Aegean tsunami, Ma­rine Geology, 163, p. 303-315.

Doutsos T., Kokkalas S., (2001), Stress and deformation patterns in the Aegean region, Journal of Structural Geology, 23, p. 455-472.

Earthquake Engineering Research Institute, (EERI, 2003), Preliminary observations on the August 14, 2003, Lefkada Island (Western Greece) earthquake, EERI Speical Earthquake Report, 11 p., http://www.eeri.org/lfe/pdf/greece_lefkada_eeri_preliminary_rpt.pdf [access: february 23, 2007].

Einsele G., (2000), Sedimentary basins. Evolution, facies, and sediment budget, 2nd edition, Springer, Berlin, Heidelberg, 792 p.

Galanopoulos A. and Ekonomides A., (1973), Rapid diapir growth, a triggering agent of a 6 1/2magnitude earthquake on October 29, 1966, in Acarnania (Greece), Annali di Geofisica, XXVI, Rome, p. 605-612.

Geyh M.-A., (2005), Handbuch der physikalischen und chemischen Altersbestimmung, Darmstadt, 211 p.

Institute of Geology and Mineral Exploration, (IGME, 1996), Geological map of Greece, 1:50 000. Vonitsa Sheet, Athens.

Jing Z., Rapp G. (Rip), (2003), The coastal evolution of the Ambracian embayment and its relationship to archaeological settings, in Wiseman J. and Zachos K., (eds.), Landscape Archaeology in Southern Epirus, Greece I, The American School of Classical Studies at Athens, Hesperia Supplement, 12, Athens, p. 157-198.

Kahle H., Cocard M., Peter Y. Geiger A. Reilinger R. Barka A., Veis G., (2000), GPS-derived strain rate field with the boundary zones of the Eurasian, African and Arabian Plates, Journal of Geophysical Research, 105, B10, p. 23353-23370.

Kapsimalis V., Pavlakis P., Poulos S.E., Alexandri S., Tziavos C., Sioulas A., Filippas D., Lykousis V., (2005), Internal structure and evolution of the Late Quaternary sequence in a shallow embayment: The Amvrakikos Gulf, NW Greece, Marine Geology, 222-223, p. 399-418.

Karakostas V.-G., Papadimitriou E.-E., Papazachos C.-B., (2004), Properties of the 2003 Lefkada, Ionian Islands, Greece, earthquake seismic sequence and seismicity triggering, Bulletin of the Seismological Society of America, 94, 5, p. 1976-1981.

Kelletat D., Schellmann G., (2002), Tsunamis on Cyprus: Field Evidences and 14C Dating Results, Zeitschrift für Geomorphologie N.F., Suppl. Vol., 137, p. 19-34.

Laigle M., Sachpazi M., Hirn A., (2004), Variation of seismic coupling with slab detachment and upper plate structure along the western Hellenic subduction zone, Tectonophysics, 391, p. 85-95.

Lavigne F., Paris R., Wassmer P., Gomez C., Brunstein D., Grancher D., Vautier F., Sartohadi J., Setiawan A., Gunawan Syahnan T., Waluyo Fachrizal B., Mardiatno D., Widagdo A., Cahyadi R., Lespinase N., Mahieu L., (2006), Learning from a major disaster (Banda Aceh, December 26th, 2004): A methodology to calibrate simulation codes for tsunami inundation models, in Scheffers A. and Kelletat D., (eds.), Tsunamis, hurricanes and neotectonics as driving mechanisms in coastal evolution, Zeitschrift für Geormorphologie N.F., Suppl. Vol., 146, p. 253-265.

Louvari E., Kiratzi A.-A., Papazachos B.-C., (1999), The Cephalonia Transform Fault and its extension to western Lefkada Island (Greece), Tectonophysics, 308, p. 223-236.

Mastronuzzi G., Sansò P., (2004), Large boulder accumulations by extreme waves along the Adriatic coast of southern Apulia (Italy), Quaternary International, 120, p. 173-184.

Minoura K., Imamura F., Kuran U., Nakamura T., Papadopoulos G.-A., Takahashi T., Yalciner A.-C., (2000), Discovery of Minoan tsunami deposits, Geology 28, 1, p. 59-62.

Morhange Ch., Marriner N., Pirazzoli P.-A., (2006), Evidence of late-Holocene tsunami events in Lebanon, in Scheffers A. and Kelletat D., (eds.), Tsunamis, hurricanes and neotectonics as driving mechanisms in coastal evolution, Zeitschrift für Geomorphologie N.F., Suppl. Vol., 146, p. 81-95.

Murray W.-M., (1982), The coastal sites of western Akarnania: a topographical-historical survey, Ph.D. thesis, University of Pennsylvania, 557 p.

Nott J., (2003), Tsunami or Storm Waves? – Determining the Origin of a Spectacular Field of Wave Emplaced Boulders Using Numerical Storm Surge and Wave Models and Hydrodynamic Trans­port Equations, Journal of Coastal Research, 19, 2, p. 348-356.

Papazachos B.-C., Papazachou C., (1997), The earthquakes of Greece, Ziti Editions, Thessaloniki, 304 p.

Papazachos B.-C., Dimitriu P.-P., (1991), Tsunamis in and near Greece and their relation to the earthquake focal mechanism, in Bernard E.N., (ed.), Tsunami hazard. A practical guide for tsunami hazard reduction, Natural Hazards, 4, 2+3, p. 161-170.

Paschos P., Papadakis J., Rondoyanni-Tsiambaou Th., (1991), Neotectonic study of Preveza-Aktio area, Institute of Geology and Mineral Exploration, Athens, 23 p.

Peter Y., Kahle H.-G., Cocard M., Veis G., Felekis S., Paradissis D., (1998), Establishment of a continuous GPS network across the Kephalonia Fault Zone, Ionian Islands, Greece, Tectonophysics, 294, p. 253-260.

Piper D.-J.-W., Kontopoulos N., Panagos A.-G., (1988), Deltaic sedimentation and stratigraphic sequences in post-orogenic basins, western Greece, Sedimentary Geology, 55, p. 283-294.

Piper D.-J.-W., Panagos A.-G., Kontopoulos N., (1982), Some observations on surficial sediments and physical oceanography of the Gulf of Amvrakia, Thalassographica, 5, p. 63-80.

Pirazzoli P.-A., Stiros S.-C., Laborel J., Laborel-Deguen F., Arnold M., Papageorgiou S., Morhange Ch., (1994), Late Holocene shoreline changes related to palaeoseismic events in the Ionian Islands, Greece, The Holocene, 4, 4, p. 397-405.

Poulos S.-E., Lykousis V., Collins M.-B., (1995), Late Quaternary evolution of Amvrakikos Gulf, Western Greece, Geomarine Letters, 15, 1, p. 9-16.

Poulos S.-E., Kapsimalis V., Tziavos C., Pavlakis P., Leivaditis G., Collins M., (2005), Sea-level stands and Holocene geomorphological evolution of the northern deltaic margin of Amvrakikos Gulf (western Greece), Zeitschrift für Geomorphologie N.F., Suppl. Vol., 137, p. 125-145.

Reimer P.-J., McCormac F.-G., (2002), Marine radiocarbon reservoir corrections for the Mediterranean and Aegean Seas, Radiocarbon, 44, p. 159-166.

Reineck H.-E., Singh I.-B., (1980), Depositional sedimentary environments. With reference to terrigenous clastics, 2nd edition. Springer, Berlin, Heidelberg, New York, 551 p.

Reinhardt E.-G., Goodman B.-N., Boyce J.-I., Lopez G., van Henstum P., Rink W.-J., Mart Y., Raban A., (2006), The tsunami of 13 December A.D. 115 and the destruction of Herod the Great’s harbor at Caesarea Maritima, Israel, Geology, 34, 12, p. 1061-1064.

Richmond B., Jaffe B.-E., Gelfenbaum G., Morton R.-A., (2006), Geologic impacts of the 2004 Indian Ocean tsunami on Indonesia, Sri Lanka, and the Maldives, in Scheffers A. and Kelletat D., (eds.), Tsunamis, hurricanes and neotectonics as driving mechanisms in coastal evolution, Zeitschrift für Geomorphologie N.F., Suppl. Vol., 146, p. 235-251.

Schäfer A., (2005), Klastische Sedimente. Fazies und Sequenzstratigraphie, Elsevier, München, 414 p.

Scheffers A., Kelletat D., (2004), Bimodal tsunami deposits – a neglected feature in palaeo-tsunami research, in Schernewski G. and Dolch T., (eds.), Geographie der Meere und Küsten. 22. AMK-Jahrestagung, 28.-30.04.2004, Coastline Reports, 1, p. 67-75.

Scheffers A., (2005), Coastal Response to extreme wave events - hurricanes and tsunami on Bon­aire, Essener Geographische Arbeiten, 37, Essen, p. 7-96.

Scheffers A., (2006a), Coastal transformation during and after a sudden Neotectonic uplift in western Crete (Greece), in Scheffers A. and Kelletat D., (eds.), Tsunamis, hurricanes and neotectonics as driving mechanisms in coastal evolution, Zeitschrift für Geomorphologie N.F., Suppl. Vol., 146, p. 97-124.

Scheffers A., (2006b), Argumente und Methoden zur Unterscheidung von Sturm- und Tsunamischutt und das Problem der Datierung von Paläo-Tsunami, Die Erde, 136, 4, p. 413-428.

Shi S., Dawson A., Smith D.-E., (1995), Coastal sedimentation associated with the December 12th, 1992 tsunami in Flores, Indonesia, Pageoph, 144, 3-4, p. 525-536.

Stefatos A., Charalambakis M., Papatheodorou G., Ferentinos G., (2006), Tsunamigenic sources in an active European half-graben (Gulf of Corinth, Central Greece), Marine Geology, 232, p. 35-47.

Tziavos C., (1997), Paleogeographic evolution of the Amvrakikos Gulf, Western Greece, in Marinos P.-G., Koukis G.-C., Tsiambaos G.-C. and Stournaras G.-C., (eds.), Engineering Geology and the Environment, Proceedings of the International Symposium on Engineering Geology and the Environment, 23-27 June 1997, Athens, p. 425-430.

Underhill J., (1988), Triassic evaporites and Plio-Quaternary diapirism in western Greece, Journal of the Geological Society, 145, London, p. 269-282.

van Hinsbergen D.-J.-J., Langereis C.-G., Meulenkamp J.-E., (2005), Revision of the timing, magnitude and distribution of Neogene rotations in the western Aegean region, Tectonophysics, 396, p. 1-34.

Vött A., (2007), Relative seal level changes and regional tectonic evolution of seven coastal areas in NW Greece since the mid-Holocene, Quaternary Science Reviews (in press).

Vött A., Brückner H., Handl M., (2003), Holocene environmental changes in coastal Akarnania (northwestern Greece), in Daschkeit A. and Sterr H., (eds.), Aktuelle Ergebnisse der Küstenforschung. 20. AMK-Tagung, Kiel, 30.05.-01.06.2002, Berichte Forschungs- und Technologiezentrum Westküste Christian-Albrechts-Universität zu Kiel, 28, Büsum, p. 117-132.

Vött A., Brückner H., Handl M., Schriever A., (2006a), Holocene palaeogeographies and the geoarchaeological setting of the Mytikas coastal plain (Akarnania, NW Greece), Zeitschrift für Geomorphologie N.F., Suppl. Vol., 142, p. 85-108.

Vött A., Brückner H., Handl M., Schriever A., (2006b), Holocene palaeo­geographies of the Astakos coastal plain (Akarnania, NW Greece), Palaeogeography, Palaeoclimatology, Palaeoecology, 239, p. 126-146.

Vött A., Brückner H., Schriever A., Luther J., Handl M., van der Borg K., (2006), Holocene palaeogeographies of the Palairos coastal plain (Akarnania, NW Greece) and their geoarchaeological implications, Geoarchaeology 21, 7, p. 649-664.

Vött A., May M., Brückner H., Brockmüller S., (2006), Sedimentary evidence of late Holocene tsunami events near Lefkada Island (NW Greece), in Scheffers A. and Kelletat D., (eds.), Tsunamis, hurricanes and neotectonics as driving mechanisms in coastal evolution, Zeitschrift für Geomorphologie N.F., Suppl. Vol., 146, p. 139-172.

Vött A., Brückner H., Brockmüller S., May M., Fountoulis I., Gaki-Papanastassiou K., Herd R., Lang F., Maroukian H., Papanastassiou D., Sakellariou D., (2007), Tsunami impacts on the Lefkada coastal zone during the past millennia and their palaeogeographical implications, in Papadatou-Giannopoulou H., (ed.), Proceedings of the International Conference Honouring Wilhelm Dörpfeld, August 6-9, 2006, Lefkada (in press).

von Seidlitz W., (1927), Geologische Untersuchung der Inselnatur von Leukas, in Dörpfeld W., (ed.), Alt-Ithaka. Ein Beitrag zur Homer-Frage, 2 volumes, München, p. 352-372.

Top of page

List of illustrations

Title Fig. 1 - Topographic overview of the coastal zone between Preveza and Lefkada and detailed map of the Aktio headland showing vibracoring sites and location of underwater studies.Schéma topographique du secteur d’étude entre Preveza et Leucade. Croquis détaillé du cap Aktio.
URL http://mediterranee.revues.org/docannexe/image/164/img-1.png
File image/png, 112k
Title Fig. 2 - Facies distribution and tsunami deposits found in vibracores from the Phoukias sand spit and Phoukias (transect A).Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Phoukias (transect A).
URL http://mediterranee.revues.org/docannexe/image/164/img-2.png
File image/png, 52k
Title Fig. 3 - Geomorphological underwater studies at Skoupeloi Achilleos to the west of the Phoukias sand spit.Vues sous-marines de blocs démantelés de beach-rock, secteur de Skoupeloi Achilleos.
Caption (a), (b), (d): Isolated rubble ridge made up of beachrock fragments. Beachrock stones and slabs originate from the partly submerged Plaka ruin. They were dislocated by tsunamigenic wave action and moved over a distance of 40m. Beachrock fragments are partly embedded in or lie on top of sandy deposits. (c), (e), (f): Beachrock slabs and blocks some 15-20m east of the in-situ Plaka beachrock. One slab sticks vertically between dislocated blocks. Some of the blocks show imbrication. All slabs and blocks are detached from the in-situ beachrock and partly embedded in a sandy substrate. Body length of snorkeler is around 1.90m. Photos taken by R. Grapmayer, 2006.
URL http://mediterranee.revues.org/docannexe/image/164/img-3.jpg
File image/jpeg, 32k
Title Fig. 4 - Facies distribution and tsunami deposits found in vibracores from Paliokoulio and Koumaros (transect B).Chronostratigraphie des carottes et dépôts de tsunami des carottes des secteurs de Paliokoulio et de Koumaros (transect B).
URL http://mediterranee.revues.org/docannexe/image/164/img-4.png
File image/png, 37k
Title Fig. 5 - Facies distribution of vibracore AKT 14 from the northeastern shore of the Lagoon of Saltini.Photographie des faciès de la carotte AKT 14 (transect B).
Caption The autochthonous limnic environment was abruptly affected by tsunami waves which left several shell debris layers. Gaps are due to rodding processes. The top/up direction for the core segments is to the left. Photo taken by M. May, 2006.
URL http://mediterranee.revues.org/docannexe/image/164/img-5.jpg
File image/jpeg, 32k
Title Fig. 6 - Facies distribution and tsunami deposits found in vibracores from Vasiliko (transect C).Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Vassiliko (transect C).
URL http://mediterranee.revues.org/docannexe/image/164/img-6.png
File image/png, 39k
Title Fig. 7 - Geomorphological indicators of a tsunami imprint found at Vasiliko.Dépôts de tsunami, secteur de Vassiliko.
Caption (a), (b): recent shore of the Ambrakian Gulf. Large pebbles colonized by boring mussels and barnacles indicate that the beach material is not moved by modern wave action. (c)-(f): Findings of large stones and blocks scattered on top of the terrain surface in the southern part of Vasiliko including a breccia with cemented beach gravel (c), a limestone block with boreholes from marine boring organisms (d), and dislocated beachrock slabs (e, f). Photos taken by A. Vött, 2006.
URL http://mediterranee.revues.org/docannexe/image/164/img-7.jpg
File image/jpeg, 84k
Title Table 1. Radiocarbon dating results for selected samples from Actio headland.Chronostratigraphie des carottes et dépôts de tsunami des carottes du secteur de Vassiliko (transect C).
Caption Note: a.s.l. – above sea level; b.s. – below ground surface; b.s.l. – below sea level; artic. spec. – articulated specimen; Mytilus gall. – Mytilus galloprovincialis; * – marine reservoir correction with 402 years of reservoir age; 1σ max; min cal BP/BC (AD) – calibrated ages, 1σ-range; “;” – there are several possible age intervals because of multiple intersections with the calibration curve; Lab. No. – laboratory number, University of Erlangen-Nürnberg (Erl), University of Kiel (Kia), University of Utrecht (UtC).
URL http://mediterranee.revues.org/docannexe/image/164/img-8.png
File image/png, 78k
Title Fig. 8 - Examples of erosional unconformities due to tsunamigenic impact encountered in vibracores from Phoukias (AKT 1), Paliokoulio (AKT 10), Koumaros (AKT 24), and Vasiliko (AKT 20). Photos taken by M. May, 2006.Photographies des carottes AKT 1, AKT 10, AKT 24 et AKT 20.
URL http://mediterranee.revues.org/docannexe/image/164/img-9.jpg
File image/jpeg, 78k
Top of page

References

Electronic reference

Andreas Vött, Helmut Brückner, Matthias May, Franziska Lang and Svenja Brockmüller, « Late Holocene tsunami imprint at the entrance of the Ambrakian gulf (NW Greece) », Méditerranée [Online], 108 | 2007, Online since 01 January 2009, connection on 24 September 2017. URL : http://mediterranee.revues.org/164 ; DOI : 10.4000/mediterranee.164

Top of page

About the authors

Andreas Vött

Faculty of Geography - Philipps-Universität Marburg - Deutschhausstr. 10 - D-35032 Marburg/Lahn, Germany

By this author

Helmut Brückner

Faculty of Geography - Philipps-Universität Marburg - Deutschhausstr. 10 - D-35032 Marburg/Lahn, Germany

By this author

Matthias May

Faculty of Geography - Philipps-Universität Marburg - Deutschhausstr. 10 - D-35032 Marburg/Lahn, Germany

Franziska Lang

Department of Classical Archaeology - Technische Universität Darmstadt - El-Lissitzky-Str. 1 - D-64287 Darmstadt, Germany

Svenja Brockmüller

Faculty of Geography - Philipps-Universität Marburg - Deutschhausstr. 10 - D-35032 Marburg/Lahn, Germany

Top of page

Copyright

Tous droits réservés

Top of page