Odysseus Unbound - The Search for Homer’s Ithaca

Onshore Drilling Outcomes in Kefalonia

2011-2014 Geoscientific Research Report

Issued 16 September 2015

Thinia Valley Panorama Thinia valley panorama looking east from Paliki

Project Background

The objective of this phase of geoscientific research has been to test the proposition that the island of Ithaca may have been accurately described in Homer’s Odyssey as the furthest west of a group of four islands off the western coast of Greece, facing dusk, the open sea and being of low elevation.

In 2005 Cambridge University Press published “Odysseus Unbound: The Search for Homer’s Ithaca”, written by Robert Bittlestone with James Diggle, Professor of Classics at Cambridge University, and John Underhill, then Professor of Stratigraphy at Edinburgh University (now Chair of Exploration Geoscience at Heriot-Watt University). The book explored the descriptions provided by Homer in the Odyssey and also a later reference by the geographer Strabo that described a partially submerged isthmus at the narrowest part of the island of Kefalonia.

If this isthmus (now called the Thinia valley) was once submerged, then the western peninsula of Kefalonia (now called Paliki) would have been a free-standing island in its own right, thereby fitting Homer’s description accurately. In this context, it is worth noting that Mycenaean-era remains have been previously found on Kefalonia, including on the Paliki peninsula. (The time frame described by Homer in the Iliad and Odyssey is generally accepted to have fallen in the late-Mycenaean period.)

Research Sponsorship

After initial field studies and the successful drilling of one onshore borehole in 2006, research sponsorship for the project was announced in March 2007 from Fugro (provider of geotechnical, survey and geoscience services) under the direction of Professor Underhill.

Fugro Marine and Helicopter Survey
Fugro marine survey team deploying cables during acquisition of seismic data.
Fugro helicopter-mounted electromagnetic survey equipment flying over the northern exit of the Thinia valley.

Over the last few years, a number of land, sea and airborne geoscientific techniques have been used to investigate the Thinia isthmus and surrounding areas with the aim of determining the nature of the underlying geology of the Thinia valley. These have included: helicopter-mounted electromagnetic and LiDAR surveys, ground-based resistivity and seismic refraction surveys, gravity surveys, shallow marine seismic reflection surveys, sidescan sonar, sub-bottom profiling and bathymetry.

Fugro ground-based resistivity and marine seismic reflection survey
Fugro ground-based resistivity survey undertaking onshore gravity and field investigations.
Fugro marine seismic reflection survey being undertaken. The vessel is pulling a streamer array of sonic detectors.
Drilling rig on site Drilling rig on site in the Livadi Marsh area.

During 2010-11, Fugro, advised by Professor Underhill, undertook a programme of land-based shallow (less than 105 metres in depth) drilling and rock coring using a small mobile rig.

This drilling program was informed by the results of the earlier geophysical studies and geological field mapping. The resultant locations were selected on the basis of their strategic importance for the theory being tested. The equipment used enabled continuous rock-cores to be obtained at 17 drilling locations in Kefalonia, mainly in the Thinia valley. The continuous nature of the cores provides an extensive and unprecedented calibration of the subterranean geology in the Thinia valley. Almost 700 metres of core samples were collected and these were analyzed at Fugro’s geo-specialist labs in North Wales in 2011-12. The results provide a new-found basis for understanding the age, nature and post-depositional deformation history of the sediments that lie buried beneath the modern valley.

The multiple research surveys and core samples were analyzed by Dr Kirsten Hunter, PhD, University of Edinburgh, and the results are now available through her 2013 PhD thesis titled “Evaluating the geological, geomorphic and geophysical evidence for the re-location of Odysseus’ homeland, Ancient Ithaca”.


The main outcomes of the geoscientific studies and implications for the theory being tested are:


1) The geology of the Thinia valley is substantially more complex as a result of geo-tectonic movements than might initially have appeared to be the case.

Aerial view of Thinia Valley An oblique aerial view of the Thinia Valley from the SW. The image illustrates the low lying nature of the ground that separates the Paliki peninsula from the main body of the island of Kefalonia. If this valley were submerged, Paliki would have formed a western island.

The structural geology of Thinia is more complicated than initially thought, with an underlying easterly-dipping extensional structure over-thrust by Hellenide compressional deformation. This in turn is faulted by NW-SE striking offsets, which post-date the extensional-compressional regime. There is indication of shear and contraction (thrust faulting) within the marls and limestone, suggesting strong post-depositional deformation. The steeply dipping deformed marl sediments are prone to frequent slope failure, and slumping is prevalent particularly on the northern coast and down-stream gullies.

In layman’s terms, this means that layers of rock of different periods have been thrust on top of one another and then deformed during geo-tectonic movements. This has led to the creation of steep geological dips, the creation of high angle slopes and driven significant movements in the Thinia valley through gravity-driven slope failure particularly at times of high rainfall and during co-seismic tectonic (earthquake) events. The tectonic structural deformation is additional to and distinct from the former proposed “sedimentary” mechanism involving only rock fall derived from the higher valley sidewall.

2) Near the northern end of the potential channel, inland of the village of Zola, evidence of a former marine beach was identified under valley fill at about sea-level.

Drilling at north channel exit near Zola Drilling at north channel exit near Zola.

Borehole C2 was positioned at the northern end of the Thinia valley, inland of the village of Zola. The core from C2 represented a tectonically-disturbed succession of Late to Early Pliocene (Zanclean to Piacenzian) marls and clays, alternating with Late Miocene (Tortonian) limestones and marls.

Beach deposits from the Pliocene epoch (5.3million – 2.6 million years ago) occurred at a depth of about 62 m beneath a thrusted contact. Although the original classification of the buried beach deposits as dating to the more recent Quaternary period was ruled out, what does remain clear is that these beach deposits confirm the original existence, at the northern end of the Thinia valley, of an ancient marine channel that has subsequently been buried under landslip material. This is another significant finding.

3) The sedimentary evidence shows that the Thinia valley was inundated more recently than previously suspected. The confirmed existence of an ancient Quaternary period channel (c. 400,000 years BP) is a significant new finding.

The shallow sedimentary coring survey for Thinia shows that the valley sediments are composed of a thick deposit of south-easterly-dipping Early / Middle Miocene – Early Pleistocene sediments deposited in a marginal to open marine neritic setting. These sediments occurred in stratigraphically conformable units emplaced onto one another by thrust faults, and post-depositional deformation has fractured and tilted these sediments steeply (40° and 60°). They were discovered at an elevated position of approximately 170 metres above sea level.

This finding confirms that a marine channel existed in the Thinia valley in the Quaternary period (2.6 - 0 million years ago). Although no sediments younger than the Gelasian (1.8 million years ago) were sampled, this does not rule out the continuation of this marine neritic environment into the Calabrian (1.8 – 0.8 million) and Ionian ages (0.8 - 0.1 million years ago) and potentially more recently still. Unusually, the sedimentary evidence was found in an elevated position, instead of buried at sea level, which (a) reflects the fact that unusual geotectonic deformation of the valley has occurred (see below), and (b) implies that younger marine neritic sedimentary evidence elevated to the surface could potentially have been bowed up and eroded before later layers formed above. This imbricated geology will be further examined below, in paragraph 5.

Consequently, we can say that what we have now is evidence for the original existence of a marine channel in the Thinia valley, that the channel is present at either end of the valley and rises up into the saddle of the valley where it has been subject to erosion; what we do not yet have evidence for is the date that this channel was obliterated, although we can say that it is some time after the Early Pleistocene.

4) The existing valley fill is not simply the result of landslip and slope collapse and consequently this hypothesis does not fully explain the absence of the marine channel today.

On the one hand, the lower slopes of the valley consist of easily-eroded soft marls conducive to severe undercutting of the limestone and conglomerate above and thus consequent slope collapse. Slumping of the marl is very evident at the posited northern exit (Agia Kiriaki Bay) and collapse of the limestone valley sides is evident throughout the Thinia valley, particularly at the southern end.

On the other hand, mapping of the surface and sub-surface geology of the Thinia valley shows that colluvial deposits are not as widespread as previously believed and that in situ bedrock would constrain the path of the channel in certain places. In particular, our new discovery of a previously unmapped thrust fault (the “Petrikata Thrust”) effectively restricts the potential channel route at the Petrikata Quarry to less than 50 m across. Borehole C1 penetrated the hanging wall of an easterly-dipping thrust fault, demonstrating that the high elevation Petrikata Quarry limestone outcrop is also part of a thrust. The presence of these thrusts means that even a vertiginous channel at this central location would encounter limestone bedrock long before reaching sea level (at about 140 m above sea level).

Although these findings might initially seem to imperil the marine channel hypothesis, in fact the newly discovered Petrikata Thrust may be one element of the more complex geotectonic deformation that has occurred in the Thinia valley and may lead to a different explanation (see below, paragraph 5).

5) At the southern end, there is new evidence for a large rotational, translational slump ending in a toe thrust that could have blocked the channel course and elevated the channel floor and marine fossils significantly above sea level.

In southern Thinia, above Agia Ioanni Bay, the Ainos Thrust is visibly displaced by around 150 m along a pronounced curvilinear valley, which extended around the top of a large portion of the thrust’s hanging wall to the village of Nifi. At the top of the mountainside, the valley forms a smooth, NW-dipping slip plane with a throw of around 560 m (dubbed the “Agia Ioanni Fault”). This displacement and scarp suggests that this valley forms a major listric (concave) normal fault. The “detached” portion of the Ainos Thrust hanging wall (a lens-shaped hill measuring around 1 km x 4 km) may represent a massive coherent slump. Independent movement of this block would explain the 8 m of lateral displacement of the Ainos Thrust plane, which occurred during the 1953 earthquake at Petrikata but nowhere else along the eastern side of Thinia. This fault then dives below the Gulf of Argostoli, so it is reasonable to assume that the displacement along this fault extends further westwards.

Rotational Slump Diagrams Schematic diagrams illustrating the processes and products of rotational slumping. The slip occurs along a rotational failure surface sited in an easy slip horizon, which in the case of the Thinia Valey is along steep westerly-dipping bedding planes. The central part of the rotation is translated and the toe of the slip is characterized by reverse (thrust) faulting.

In layman’s terms, this suggests the possibility that a massive segment of the Thinia valley was detached, as a result of geo-tectonic, ground-water and/or gravitational forces, from its bedrock position and moved as a coherent unit down the hillside to the bottom of the valley and then up the other side in a rotational slide to form what is called a toe-thrust at its base. It is plausible that this slope failure may have led to the displacement and elevation of an in-situ marine channel from sea level to an elevated position on the top of the toe-thrust. This could then cause marine sedimentary material to be raised to an elevated situation above its normal position. And, in fact, as described in paragraph 2 (above), marine neritic sedimentary deposits were discovered in the Thinia valley boreholes at an elevated position of approximately 170 metres above sea level. It goes almost without saying that marine deposits should not be found in this elevated position without recourse to some very unusual geo-tectonic or geological activity. It is also worth noting that, once elevated and exposed to the elements, the more recent sedimentary deposits would be the first to be removed as a result of erosion caused by the elements.

This last point could be one plausible explanation for the fact that examples of Emiliana huxleii (Ehux), a distinctive Quaternary nanno-fossil, were found in the initial Thinia valley borehole in 2006 and not identified in duplicate boreholes drilled in the same location in 2011. (Other plausible explanations could include wind-blown contamination of the original core sample from surface deposits or original mis-identification of Ehux in 2006 through confusion with a similar-looking Neogene species, Orbulina universa.)

The occurrence of a large extensional fault on the eastern side of the valley is a major new finding. The fault utilizes steeply east-dipping bedding planes, was formed as a result of gravitational (rotational) collapse and is young since it offsets all other structures. If linked to the compressional toe-thrust complex on the western side of the valley, it has the potential to create the mechanism by which the valley fill is deformed, uplifted and the marine channel bowed upwards. Such rotational rockslides are not uncommon in other areas where the geological circumstances are similar and they are often triggered by seismic activity or elevated groundwater levels. One recent example of this phenomenon occurred in the mountains of Nepal as a result of the Gorkha earthquake that struck the area on April 25 2015. In this case, the main co-seismic slip was directed east from the hypocenter, with a maximum displacement of >4m leading to slope failure with dimensions of about 120 by 80 km.1

6) An unusual feature in the Thinia valley is an elevated flat lakebed, named Lake Katachori, which was created by geo-tectonic movements and has subsequently dried out. In origin this feature is strikingly similar to Quake Lake in southwestern Montana, USA, which was formed as a result of seismic activity in 1959.

Former lakebed at Lake Katachori The flat plain that characterizes the Thinia Valley represents a former (palaeo) lake bed (Lake Katachori) that was instigated by the rotational slump having blocked upland drainage and ponded headwaters to form a lake.

On the southwest flank of the Thinia valley in an area called Katachori, there is clear evidence that a major rock fall has originated from the eastern slopes of the valley, travelled across the valley floor and has come to rest (on-lap) upon the upwards-sloping western hillside, infilling the valley itself in the process. The tail end of the rock fall has covered some pre-existing walls which end abruptly at the debris and their continuity can be identified underneath it. Up-thrust combined with rock fall and valley fill set up a barrier for drainage that caused an ancient lake to form subsequently above it. The lakebed has now silted up and forms a low agricultural plain in the centre of the valley.

The formation of Lake Katachori appears to have followed a very similar process to the one that created Quake Lake in southwestern Montana, USA, in 1959. In the case of the latter, an earthquake with a magnitude of approximately 7.4 on the Richter scale triggered a landslide of an estimated 80 million tons of rock and earth that dammed the flow of the Madison River below the Hebgen dam and thus created a lake that today is almost 10 km long and about 0.4 km wide at its widest point.

Quake Lake, Montana Quake Lake, Montana: the rotational slide blocked outflow from the rivers that drained the valley to set up a lake behind it. A similar mechanism is envisaged for the formation of Lake Katachori in Thinia. Photos © Blox Images, Chicago.2

While confirming a freshwater environment, the six cores drilled for Lake Katachori detected only a thin veneer (0.65 to 6.40 m) of lake-fill, suggesting this was a very shallow feature. These deposits were sited on steep, easterly-dipping Plio-Pleistocene sediments uplifted to ~170 m above sea level. The occurrence of freshwater algae with uppermost Pleistocene (Gelasian) sediments suggests that departure from a marine depositional setting occurred in Thinia sometime after Early Pleistocene (1.80 million years ago), while the rock-fall’s overlap of pre-existing walls, which can probably be dated to the modern era, potentially implies that the rock-fall and creation of the lake occurred within geologically recent times.

7) In the Livadi Marsh, evidence was found of ancient marine sediments and beach deposits indicating the potential for an ancient harbour to have existed at the foot of Kastelli Hill.

Livadi Harbour & Gulf from Kastelli
Livadi harbour site from Kastelli Hill.
Livadi harbour site beside Kastelli Hill.

The oldest marine sediments and beach deposits (Early Pliocene) in the Livadi Marsh occurred all the way to the base of Kastelli Hill, demonstrating that the marsh area could have formed a deep natural harbour between the Pliocene and Middle Pleistocene periods (5.3 million – 0.8 million years ago).

A record of the more recent Late Holocene flooding history came from the radiocarbon dated samples taken from borehole C6c.

Core from Livadi harbour bore hole Cores from Livadi harbour borehole: the cored section shows breccias and conglomerates found at the unconformity at the base of the Holocene section.

The deposition of thick recent sediment in C6c indicates that the marsh area was partially flooded after the Last Glacial Maximum, which occurred between 26,500 and 19,000–20,000 years ago. The calibrated radiocarbon samples yielded the following results, calculated in term of years before present (BP) and then re-stated as years BC:

5.0 m depth: 3362 ± 34 BP = 1,412 BC ± 34
6.3 m depth: 4471 ± 31 BP = 2,521 BC ± 31
7.2 m depth: 5206 ± 67 BP = 3,256 BC ± 67

1,412 BC would be considered to be the late period of the Mycenaean age, approximately 100-200 years prior to the generally accepted time of the Trojan War in the 13th century BC.

Further research is required to determine the appropriate adjustment factor to take account of tectonic uplift and eustatic sea-level change on Kefalonia subsequent to these dates. It is not advisable to use macro global measures as Kefalonia sits over abutting tectonic plates that create a unique geo-tectonic environment.

At a minimum, it is possible to state that a natural harbour once existed at the foot of Kastelli Hill. It is not yet possible to say whether it could be characterized as a deep harbour (πολυβeνθής λιμήν), as the harbour of Ithaca is described, but it is a realistic possibility.


The geological research sponsored by Fugro and directed by Professor John Underhill has been completed safely, successfully and with no harm to the environment. The results of the geoscientific campaign unveiled a complex geo-tectonic and geological landscape beneath the Thinia valley. It has refuted the simple side-wall collapse and in-fill hypothesis and has raised another possibility, that of a massive rotational slump, leading to a toe-thrust and elevation of an ancient marine channel.

The critical question then becomes one of timing: could this rotational slump have occurred recently (in the last ~3,100 years) and therefore be responsible for in-filling and displacement of a Mycenaean age marine channel?

Other promising findings include the discovery of ancient beach deposits buried under land-slip infill at the northern end of the Thinia valley and evidence of an ancient marine harbour that reached the foot of Kastelli Hill and that contained radiocarbon-datable samples contemporaneous with the time of Odysseus. An anomalous late-period lakebed, now dried out, is another intriguing finding.

It is also worth reiterating that Mycenaean-era sites have been previously identified on both the western peninsula (called Paliki) and the remaining part of the island of Kefalonia, confirming that this land is of considerable historical and archaeological interest dating to the period in question.

Next Steps

The proposed next steps in the geological research are to conduct a field-based mapping assessment of the southern exit to investigate whether there is evidence for a tilted and elevated channel rising up the eastern sides of the Gulf of Argostoli.

In addition, we will seek to undertake a marine-based shallow coring at the northern end of the Gulf of Argostoli in order to establish a date profile for the in-fill of a former marine channel identified on the seismic data and further investigate the nature of its sedimentary in-fill. This will provide a baseline from which it should be possible to describe how and when this valley has evolved and been in-filled. It is hoped that this research, postponed from 2012, can be accomplished in 2016, but the timing will depend on funding and local and national regulatory approvals and permits being obtained.

Other preliminary conceptual work will be developed in relation to Kastelli Hill, Livadi Harbour and other sites on the Paliki peninsula in an effort to define future parameters for possible sub-surface investigation. Next steps could involve discussion with local and national authorities on the possibility of non-intrusive sub-soil scanning of promising sites identified in the course of the geological exploration.


This acknowledges the very significant contribution of Dr Kirsten Hunter to the understanding of the underlying geology of the Thinia valley and adjacent sites on Kefalonia through the work described in her PhD thesis titled “Evaluating the geological, geomorphic and geophysical evidence for the re-location of Odysseus’ homeland, Ancient Ithaca”. A substantial number of short excerpts from Dr Hunter’s thesis have been included in this summary.

The project research team would also like to acknowledge the enormous contributions made by the geo-scientific company Fugro in sponsoring and carrying out the extensive phases of research, under the direction of Professor Underhill, that have led to these findings.

The views expressed in this update, however, represent the opinions of the OU project team, and may or may not coincide with Dr Hunter’s and/or Fugro’s views in every instance.

Contact information

For geological inquiries:
Professor John Underhill
Past-President - European Association of Geoscientists & Engineers (EAGE)
Chair of Stratigraphy & Fellow of the Royal Society of Edinburgh (FRSE)
Professor of Exploration Geoscience
Centre for Exploration Geoscience,
School of Energy, Geoscience, Infrastructure & Society,
Heriot-Watt University, Edinburgh Campus,
Riccarton, Edinburgh, Scotland, EH14 4AS;
Email: J.R.Underhill@hw.ac.uk

For Homeric contextual inquiries:
Professor James Diggle
Emeritus Professor of Greek and Latin at Cambridge, & Fellow of Queens’ College, Cambridge
Email: jd10000@cam.ac.uk

For publisher’s inquiries
Cambridge University Press

Cambridge office:
Dr Michael Sharp, msharp@cambridge.org, +44 (0) 1223 325733
Classics Editor, Humanities & Social Sciences

New York office:
Melissanne Scheld, mscheld@cambridge.org, +1 212.337.5988

For general inquiries


Anne Rouse (formerly Anne Stephenson), Metapraxis Ltd.
anne.rouse@metapraxis.com or + 44 (0) 20 8541 2743

John Crawshaw, OU Project Coordinator
Email: odysseus.unbound.info@gmail.com


  1. Source: The April 25, 2015 M 7.8 Gorkha Earthquake and its Aftershocks, Earthquake Educational Slides, Created & Compiled by Gavin Hayes, U.S. Geological Survey, National Earthquake Information Center.
  2. Image credits:
    Unless otherwise specified, images shown are the copyright of the Odysseus Unbound project.
    Black & white view of Quake Lake: Blox Images, Chicago
    Colour image of Quake Lake: Blox Images, Chicago