Submerged Sites Between Geology and Pseudoarchaeology

In 1986, a diving tour operator named Kihachiro Aratake was scouting hammerhead shark dive sites off the southern coast of Yonaguni Island when he came upon a series of step-like sandstone formations beneath the surface. [1] [2] [3] [41] Shortly afterwards, a group of scientists led by University of the Ryukyus marine geologist Masaaki Kimura visited the formations, and Kimura came to argue that they were man-made, identifying a pyramid, castles, roads, monuments and a stadium. [41] Boston University geologist Robert M. Schoch examined the same formation and judged it primarily natural, over 95 percent by his estimate, with the rock breaking along horizontal bedding planes and vertical joints. [8] The rock is Early Miocene sandstone and mudstone of the Yaeyama Group, believed to have been deposited about 20 million years ago, and neither Japan’s Agency for Cultural Affairs nor the government of Okinawa Prefecture recognises the features as important cultural artifacts. [41]

Illustration: Yonaguni Monument DSC02806. Photo: Vincent Lou from Shanghai, China / Wikimedia Commons. DSC02806. Source

Yonaguni Monument DSC02814. Photo: Vincent Lou from Shanghai, China / Wikimedia Commons. DSC02814. Source

Yonaguni Monument DSC02807. Photo: Vincent Lou from Shanghai, China / Wikimedia Commons. DSC02807. Source

Sevier River Formation (Miocene; Joe Lott Creek Canyon, Tushar Mountains, Utah, USA) 52. Photo: James St. John / Wikimedia Commons. Source

Illustration: Yonaguni Island ISS045. Photo: Image Science and Analysis Laboratory, NASA-Johnson Space Center. “The Gateway to Astronaut Photography of Earth.” / Wikimedia Commons. Yonaguni Island, located in Okinawa Prefecture, Japan. Source

Illustration: Coarse sandstone beds dipping southwards - geograph.org.uk - 1096728. Photo: Nic Howes / Wikimedia Commons. Source

In December 2000, India’s National Institute of Ocean Technology ran side-scan sonar surveys across the Gulf of Khambhat and detected geometric patterns that the investigators described as “unnatural features”. [11] When dredging later pulled up a piece of wood radiocarbon dated to approximately 7500 BC, prominent archaeologists rejected the date as significant because the artifacts had been recovered by dredging rather than controlled excavation, and the wood’s stratigraphic relationship to the sonar patterns could not be established. [4]

Gujarat Gulfs. Photo: Wikimedia Commons. Source

Map of Gulf of Cambay, A map by Petrus Bertius, Amsterdam, 1606. Photo: Petrus Bertius / Wikimedia Commons. A map of Gujarat by Petrus Bertius, Amsterdam. Source

Illustration: Britannica Dredge and Dredging 15. Photo: Encyclopædia Britannica (1911), public domain / Wikimedia Commons. Deep-sea Dredge, with Tangle Bar. Source

Yonaguni: The Sandstone Steps Beneath Japan

Dive operator Kihachiro Aratake was hunting for hammerhead shark observation sites off the south coast of Yonaguni Island in 1986 when he reported a submerged structure that has since become Japan’s most contested piece of bedrock. The main rectangular formation, known as the Yonaguni Monument, measures roughly 100 m long by 60 m wide and rises about 25 m from the seabed, with its top lying about 5 m (16 ft) below the surface. [5] The rock itself is unremarkable geologically: it is Lower Miocene Yaeyama Group sandstone and mudstone deposited approximately 20 million years ago, mostly connected to the underlying rock mass rather than assembled from free-standing blocks. The 20-million-year figure is the geological age of the substrate — not any claimed construction date — and similar step-like sandstone formations occur onshore on Yonaguni Island itself, at Sanninudai. [5] [41]

Illustration: Location Ryukyu Islands. Photo: Nanshu, Michiel1972 / Wikimedia Commons. Source

Masaaki Kimura, a marine geologist at the University of the Ryukyus, has spent more than 15 years measuring and mapping the formation, and he insists the site is a man-made complex of stepped monoliths. His catalog of features he interprets as artificial includes a stepped pyramid, a castle, a triumphal arch, five temples, a stadium, roads, canals, retaining walls, quarry marks, animal sculptures, and what he calls an ‘underwater sphinx’. Kimura’s chronology has shifted. He initially placed construction about 10,000 years ago, when the site would have stood above water [7]; in 2007 he described the remains of a city at least 5,000 years old that sank in an earthquake about 2,000 years ago [6], and in a report to the 21st Pacific Science Congress that year he revised his estimate to 2,000 to 3,000 years ago, suggesting that tectonic activity rather than post-glacial sea-level rise submerged the structure after its construction. [41]

The rock arch at the Yonaguni Monument through which dives usually begin, leading to the “Twin Megaliths”. Photo: Melkov / Wikimedia Commons, CC0. Source

Robert M. Schoch of Boston University, who has dived the site personally, reaches a different conclusion. He estimates the formation is over 95% natural, produced by well-defined bedding planes and orthogonal jointing in sandstone that fracture along straight lines. But here’s the part that troubles Kimura’s reading: geometrically similar stepped and rectangular fracturing appears above sea level along Yonaguni’s own northeast coast, in surface outcrops untouched by any ancient engineers. Schoch attributes the step geometry to natural bedding planes and joints in a region he calls earthquake-prone; Kimura and his supporters reject a natural origin. Schoch allows that prehistoric inhabitants may have ‘touched up’ portions and mimicked the natural step-like features in now-visible rock-cut tombs on the island, but he rejects the artificial-origin reading overall. One Japanese seismologist, Toru Ouchi of Kobe University, has sided with Kimura, dissenting from the natural-formation consensus — a reminder that the dispute is active rather than closed. [8] [6] [9]

The ‘steps’ identified on the structure are too big to be climbed as stairs, which is hard to reconcile with a practical stairway. [9] Kimura himself acknowledges that pottery and wood do not survive on the seafloor [6], and the site has yielded no in situ diagnostic archaeological finds. [32] Neither Japan’s Agency for Cultural Affairs nor the Okinawa prefectural government considers the area of cultural significance [10], and neither has carried out research or preservation work at the site. [6]

Yonaguni Monument DSC02829. Photo: Vincent Lou from Shanghai, China / Wikimedia Commons. DSC02829. Source

The combination of missing artifacts and dramatic visual geometry has made Yonaguni a fixture of pseudoarchaeological literature. Kimura himself hypothesizes the site may be a remnant of the mythical lost continent of Mu, and his structural identifications — castle, triumphal arch, stadium, temples — have been taken up by supporters of an artificial origin such as Graham Hancock without independent peer-reviewed corroboration. [41] The Mu hypothesis has no accepted historical backing; it derives from 19th-century mistranslations and occult literature, not from geology or archaeology, and mainstream encyclopedic sources explicitly classify Kimura’s man-made-construction reading as pseudoarchaeological. Kimura cites only one historical event in support of submergence — the April 1771 Yaeyama tsunami, with a reported run-up over 131 ft — but that event is documented, not prehistoric, and provides no evidence for artificial construction. Forty years of diving, documentaries, and scientific expeditions have failed to recover a single diagnostic artifact from the submerged structure itself — and that absence is precisely what has allowed the Yonaguni story to anchor Japan’s version of the global pseudoarchaeological Atlantis circuit. [6]

Gulf of Khambhat: Sonar Anomalies and a Dredged Log

In December 2000, a side-scan sonar survey by India’s National Institute of Ocean Technology (NIOT) operating in the Gulf of Khambhat detected a series of geometric patterns on the seafloor approximately 20 km west of Hazira, at water depths of 20–40 m — patterns the NIOT team described in its own publication as “very unusual in the marine domain”. [11] The official response was swift: on 19 May 2001, India’s Union Minister Murli Manohar Joshi announced that the ruins of an ancient civilization had been discovered, and a follow-up NIOT campaign in November 2001 used dredging rather than controlled excavation to recover material from the seafloor. [4] The dredge recovery yielded stone artefacts, potsherds, hearth pieces, human bones and teeth embedded in fluvial sands and silts, along with a carbonised wooden log radiocarbon-dated to approximately 9500 years BP (~7500 BC). [11]

Illustration: Gulf of Cambay (South part) 1896. Photo: The British Library / Wikimedia Commons. Gulf of Cambay (South part) 1896. Source

Gulf of Cambay (North part) 1896. Photo: The British Library / Wikimedia Commons. Gulf of Cambay (North part) 1896. Source

The November 2001 NIOT campaign used dredging rather than controlled excavation — and that single methodological choice became the central objection. Critics pointed out that dredging may collect errant artefacts along with those that belong to the site and makes analysing stratification virtually impossible, so no recovered object can be securely tied to the sonar anomalies. [4] The 7500 BC wood date, although striking, does not by itself imply a civilization: D. P. Agrawal, founder of carbon-14 testing facilities in India, argued that such wood is a common find, because 20,000 years ago the Arabian Sea stood about 100 m lower than today and the gradual sea-level rise since then submerged entire forests. [4] The lithic material itself remains contested: critics argue that many of the items identified as artefacts by the NIOT investigators are stones of natural origin. [4] NIOT’s own thin-section study (Kathiroli et al. 2003) described perforated calcareous sandstone samples from the site and attributed their holes to deliberate boring by a stone tool, reading the sandstone as an ancient land surface on which fluvial and human activities took place. [11]

The Gulf of Khambhat is not a forgiving place to date anything from. Artifacts at the site were recovered by dredging rather than by controlled archaeological excavation, so they cannot be definitively tied to the sonar features, and the reported site lies about 20 km from the Gujarat coast, spans 94 km and sits at a depth of 30-40 m. Follow-up surveys found two palaeochannels of old rivers in the middle of the gulf under 20-40 m of water. The area is very tectonically active, with several faults and periodic earthquakes, and that activity means there is not sufficient stratification to be sure the recovered artifacts can be associated with the site at all. [4]

Illustration: ONGCbridge. Photo: Urva222 / Wikimedia Commons. ONGC Bridge, Over Tapi River. Source

MODIS satellite view of sediment streaks across the Gulf of Khambhat, 10 January 2022. Photo: MODIS Land Rapid Response Team, NASA GSFC / Wikimedia Commons. Source

MODIS satellite view of the Gulf of Khambhat. Photo: MODIS Land Rapid Response Team, NASA GSFC / Wikimedia Commons. Source

Illustration: The Narmada inland in Madhya Pradesh; the river reaches the sea through an estuary near Bharuch in the Gulf of Khambhat. Photo: Abhirock123 / Wikimedia Commons. Source

The gulf’s coast has a well-attested archaeological baseline: the Harappan port of Lothal lies at the head of the gulf. [4] A terrace sequence at the mouth of the Sabarmati River, an area that hosted Harappan cultural centres during the Holocene, records the area from about 11,300 to 1,300 years BP, shifting from fluvial and marginal-marine to predominantly marine deposits and back to a mixed environment. [12] A synthesis of relative sea-level data from five coastal sites in the gulf points to a mid-Holocene high stand of about +1.2 m at Lothal to about +2.0 m at Diu (uncorrected for tectonic uplift), fluctuations that may have influenced the navigability and habitability of key Harappan settlements. [13] Within the Gulf of Khambhat itself, two palaeochannels of former river courses were identified at 20-40 m water depth, approximately 20 km offshore — real drowned landscape features, independent of any “city” interpretation. [4]

Lothal Dockyard. Photo: Raveesh Vyas / Wikimedia Commons. Source

More than two decades after the initial NIOT survey, the core problem remains unchanged: no controlled excavation has been conducted at the site, and the academic literature continues to treat the anthropogenic origin of the larger “city-like” structures as unproven. The “lost civilization” interpretation has been amplified in popular media, including a 2002 BBC report and Graham Hancock’s writings, but mainstream archaeological opinion cites sonar misinterpretation, carbon-dating errors, and contextual problems with dredged artefacts. The submerged landscape is real. The 7500 BC wood is real. Whether either connects to a pre-Harappan urban settlement remains undemonstrated. [14] [15] [4] [16]

Bimini Road: Beachrock or Ancient Pavement?

On September 2, 1968, divers J. Manson Valentine, Robert Angove, and Jacques Mayol descended into roughly 5.5 meters of water off the northwest coast of North Bimini and reported a 0.8-kilometer linear arrangement of rectangular limestone blocks oriented northeast-southwest, with a hook at its southwest end. Within months, the feature was being called a “road” and linked to Atlantis. Yet the rocks themselves are not masonry. They are beachrock — a natural sedimentary rock formed in the intertidal zone when carbonate-supersaturated seawater cements sand, shell hash, and coral fragments beneath the beach surface. [17] The blocks are rounded and cushion-like, not tool-cut; their high degree of erosion makes it “highly implausible that any original surface features, including any tool marks and inscriptions, would have survived”. [18] This is the fundamental starting point: Bimini Road is made of natural beachrock, the kind of which forms routinely across the Bahamas and worldwide. [17]

Illustration: Bimini island. Photo: Original picture by NASA / Wikimedia Commons. Source

Illustration: Map of North and South Bimini (labels partly in German). Map: Joachim Greiner / Wikimedia Commons. Source

The radiocarbon evidence reinforces the natural origin. A core collected by Shinn in 1977 and dated at the University of Miami in 1978 returned whole-rock ages of 2,780±70, 3,500±80, and 3,350±90 BP, with shells at 3,510±70 BP and carbonate cement at 2,770±80 and 2,840±70 BP — the shells older than the cement that binds them, internally consistent with sediments accumulating first and being cemented later. Davaud and Strasser concluded that the beachrock layer formed beneath the surface of North Bimini Island and was only exposed by coastal erosion approximately 1,900–2,000 years ago. Gifford and Ball’s direct examination produced eight specific observations weighing against an anthropogenic interpretation: the three linear features are unconnected at the southwest end; no second course of blocks exists anywhere; insufficient blocks remain to have formed a destroyed second course; bedrock closely underlies the entire area; the blocks of the inner and middle features have always rested on a layer of loose sand, with no evidence of being cut into or founded on the underlying bedrock; no regular or symmetrical supports were observed beneath any block; no regular or repeated grooves interpretable as tool marks were present; and the seaward feature is not continuous enough to have served as a thoroughfare. A uranium-thorium date of 14,992±258 BP from underlying marine limestone, sometimes cited as evidence of great antiquity, is scientifically invalid for two independent reasons: the sample was partially recrystallized and therefore not a closed system, and at roughly 15,000 years ago sea level stood 95–100 meters below present, meaning the sampling locality was 90–95 meters above sea level when that carbonate supposedly accumulated. The geological verdict is not contested at the level of evidence — it is contested only in popular and fringe forums. [18]

The Atlantis identification faces a second, independent line of problems. Plato’s Timaeus and Critias place the destruction of Atlantis roughly 9,000 years before his own time (around 9,600 BCE) and locate it beyond the “Pillars of Hercules” in the Atlantic — the Strait of Gibraltar, not the Bahamas. The Cayce connection, meanwhile, rests on psychic readings: the American mystic Edgar Cayce predicted that part of Atlantis would rise again in 1968 or 1969, a prediction his followers linked to his readings placing Atlantean temples near Bimini, decades before the discovery. This prediction is the sole historical basis for the Atlantis reading; the radiocarbon dates place the cementation of the beachrock about 2,800 years ago and its exposure by coastal erosion about 1,900 to 2,000 years ago, some nine thousand years after Plato’s date. [18] The geographic mismatch and chronological mismatch are both fatal on their own; together they leave the Cayce prophecy as the only thread holding the identification together. [19] [20] [21]

The peer-reviewed methodology for distinguishing natural from anthropogenic submerged features, articulated by Galili and colleagues in 2024, requires geological context, sea-level considerations, associated archaeological finds (including coastal survey), broader archaeological context, and direct visual assessment by an underwater archaeologist; the authors applied it to two case studies in the Sicilian Channel. [32] [33] Applied here to Bimini Road, which Galili and colleagues do not discuss, the criteria are not met: Shinn found no artifacts and no wheel ruts [19], and Gifford and Ball found that the blocks of the inner and middle features rest on a layer of loose sand, with no evidence of being cut into or founded on the bedrock. [18] The Galili framework explicitly warns that remote sensing alone is insufficient and that direct visual assessment is required — a warning that the original popularizers of the Atlantis interpretation, working from a 1968 diver’s report and a psychic prediction, never heeded. [32]

The deeper methodological pattern is more telling than any single test. The “New Atlantis” framework associated with Cayce, and later popularized by writers like Graham Hancock, has been characterized in academic critique as a form of historical creationism — selecting evidence that confirms the desired conclusion (a single prophecy, the linear shape of an eroding beachrock exposure, Plato’s literary description of a fictional island) while ignoring the geological dating, the field observations of trained geologists, the absence of any cultural material, and the basic geography of Plato’s own account. When a prophecy made decades before the feature was discovered is cited as confirmation, while a 1978 radiocarbon study showing 2,800-year-old beachrock cement is dismissed, the asymmetry is the methodology. Bimini Road is a textbook case of how a feature with a documented natural origin — exposed by ordinary coastal erosion of an intertidal cementation surface — gets recruited into a lost-civilization narrative built on prior belief. The geological evidence resolves the question; the framework simply refuses to update. [22] [19] [23] [21]

Doggerland: The Benchmark of Submerged-Landscape Science

The trawler Colinda’s recovery of a lump of peat containing a barbed antler harpoon in 1931 was the first tangible archaeological evidence of human occupation of what is now the southern North Sea bed. [24] [25] The red-deer antler harpoon came up in September 1931 from the Leman and Ower Banks, from a depth of 19–20 fathoms, inside a mass of the peat known as moorlog. [49] No sample of the peat around the harpoon was kept, but pollen from peat that the same trawler later dredged elsewhere on the Leman and Ower Banks indicated mixed woodland, evidence that much of the North Sea had once been dry land. [49] When the harpoon itself was radiocarbon-dated in 1988 (OxA-1950, 11,740 ± 150 BP, c. 11,950–11,300 cal BC), the result replaced an earlier date of c. 6500 BC for the peat and gave the first direct date for Upper Palaeolithic settlement on this part of the North Sea Plain. [49]

The scale of Doggerland research expanded dramatically when the University of Birmingham’s North Sea Palaeolandscapes Project, with Vincent Gaffney and Ken Thomson among its researchers, used 3D seismic data originally acquired by the petroleum industry, more than 22,000 km² of English surveys in the southern North Sea, to map the Mesolithic landscape. [40] That footprint grew further under the ERC-funded “Europe’s Lost Frontiers” project (grant 670518, 2015–2020), which integrated seismic survey, vibrocoring, marine magnetometry and sedaDNA analysis across the Southern River palaeo-channel system. Sediment core ELF019, recovered under that programme, documented the environmental transition from freshwater silts through terrestrial organic deposits to marine clay, with the freshwater-to-marine transition radiocarbon-dated to 10,243–10,199 cal BP, anchoring the drowning sequence to the Greenlandian stage of the Lower Holocene. [26] [27]

The Storegga tsunami debate exemplifies how Doggerland research has moved beyond assertion into multi-proxy resolution. Around 8150 BP, the tsunami generated by the Storegga submarine landslide off Norway struck North-west Europe, and its size has led many to assume a devastating impact on contemporaneous Mesolithic communities, including the final inundation of Doggerland. [29] Weninger et al. (2008, Documenta Praehistorica 35, 1–24) argued for a precise calendric date and framed the tsunami as the catastrophic final flooding of Doggerland [38], while Hill et al. (2017), in a scenario Walker et al. call Dogger Bank survival, estimated that a maximal inundation would have covered around 35 per cent of the exposed landmass. [29] The 2020 Gaffney et al. multi-proxy characterisation of vibrocore ELF001A from the Outer Dowsing Deep provided the first direct physical evidence of Storegga deposits in the southern North Sea, but the authors concluded that the tsunami’s impact was “temporary and likely localized and mitigated by the dense woodland and topography of the area,” with clear physical remnants restricted to inland basins and incised river valleys; its sedimentary DNA recorded rodents and larger animals such as bear, boar and cloven-hoofed ruminants in a shady, closed woodland. [28] Walker et al. (2020, Antiquity 94, 1409–1425), working from the same core, presented this first evidence of the tsunami from the southern North Sea as a reason to rethink traditional notions of a catastrophically destructive event. [29] The disagreement has since become partly a methodological question: Waddington and Wicks (2017) argue that the tsunami may have had a taphonomic effect that reduced the number of surviving or detectable sites. [37]

What distinguishes Doggerland from the other candidates in this comparison is the steady accumulation of peer-reviewed evidence rather than a single headline find. Brown Bank, roughly 85 km off the Dutch coast, has been surveyed with simultaneous high-resolution (0.5–1 m) and ultra-high-resolution (10–20 cm) seismic, vibrocoring and dredging to recover Mesolithic artefacts within their preserved landscape context. [47] The Wessex Archaeology campaign for Vattenfall’s Norfolk Boreas offshore wind farm delivered what the team described as the most complete and extensive set of sediment cores ever extracted from the southern North Sea, covering approximately 85 km² of mapped peat deposits and providing a near-continuous environmental record of roughly 3,500 years. [30]

Doggerland is the only one of the four sites under consideration where the submerged-landscape literature expands year on year with new primary data. The Yonaguni Monument, Gulf of Khambhat and Bimini Road each remain frozen at their original disputed interpretations, with no independent corroboration of an anthropogenic origin and with mainstream sources still classifying them as natural or unverified. Doggerland, by contrast, is being actively re-dated, re-modelled and re-sampled: a 2025 Nature paper, for example, reconstructed global sea-level rise in the Early Holocene from North Sea peats. [31] The unresolved question is whether the Storegga tsunami’s social signature on Mesolithic communities can ever be extracted from the offshore record — a taphonomic problem, not a research-design problem, and the clearest indicator that Doggerland has moved from speculative survey into operational science. [31]

Connecting the Dots

The Yonaguni Monument occupies a sand-and-mudstone outcrop of Lower Miocene Yaeyama Group rock that weathers and erodes along its bedding planes and joints and lies below the Arakawabana cliff at the island’s southern tip. [41] Yet marine geologist Masaaki Kimura of the University of the Ryukyus has spent fifteen-plus years mapping what he calls stepped monoliths [6], while Robert M. Schoch of Boston University concludes the formation is more than 95 percent natural, with parts possibly touched up by ancient islanders. [8] The peer-reviewed Galili et al. methodology for distinguishing anthropogenic from natural submerged landscapes requires geological context, sea-level considerations, associated archaeological finds, broader archaeological context, and direct visual assessment by an underwater archaeologist rather than remote sensing alone — when applied to two Sicilian Channel cases, both were reclassified as natural features. [32] [33] Yonaguni fails those tests on multiple counts: no definitive artifacts (Kimura himself acknowledges that pottery and wood do not last on the bottom of the ocean), no published tool-mark photographs, and the steps are spaced several feet apart rather than at the roughly eight inches that a human stairway would require. [6] [42] Archaeologist Richard J. Pearson considers Kimura’s scenario of construction followed by tectonic submergence unlikely. [41]

The dating evidence collapses under the same logic that makes Doggerland science possible. Van der Plicht et al. (2016) demonstrate that Mesolithic Doggerland inhabitants had significant freshwater-fish dietary components, producing freshwater reservoir effects on radiocarbon dates that preclude precise calibration to absolute ages. [48] A piece of wood dredged from the Gulf of Khambhat was radiocarbon dated to approximately 7500 BC, but prominent archaeologists reject this date because the wood’s stratigraphic relationship to the sonar features cannot be established — artefacts there were recovered by dredging rather than controlled archaeological excavation. [4] The absence of organics at Yonaguni is consistent with Kimura’s own admission that pottery and wood do not survive on the seafloor [6] — but it provides no positive evidence for human activity, only for the absence of preservation. At Bimini, radiocarbon dating of beachrock samples collected by E. A. Shinn in 1977 and processed at the University of Miami yielded whole-rock dates of 2780±70, 3500±80, and 3350±90 BP, with cement dates of 2770±80 and 2840±70 BP [18] — internally consistent with shells being older than the cement that bound them. Davaud and Strasser concluded the beachrock layer composing Bimini Road originally formed beneath North Bimini Island and was exposed by coastal erosion approximately 1,900 to 2,000 years ago, far younger than Plato’s 9,600 BCE date for Atlantis. [18] [21]

Doggerland’s transformation from obscure dredge-finds into a mapped landscape required deliberate reframing by Bryony Coles’s formative 1998 paper “Doggerland: a Speculative Survey” and the mobilization by Vincent Gaffney’s Birmingham team of more than 22,000 km² of petroleum-industry three-dimensional seismic data into a research programme. [40] That same evidentiary chain — geological context, sea-level data, organic preservation, broader regional archaeology — is what the Galili et al. criteria operationalize and what the advocate-driven Yonaguni, Gulf of Khambhat, and Bimini claims consistently bypass. Edgar Cayce’s prediction that part of Atlantis would rise again in 1968 or 1969, which his followers linked to his readings placing Atlantean temples near Bimini, preceded the September 2, 1968 discovery of Bimini Road by J. Manson Valentine, and Graham Hancock’s Netflix series Ancient Apocalypse frames the feature as a mysterious underwater megalithic structure that, if man-made, must date to the Ice Age. The persistence of that interpretation despite Shinn’s beachrock demonstration (1978, 2004), Gifford and Ball’s no-tool-marks finding, and the radiocarbon sequence inconsistent with Plato’s date fits the posture Hancock himself describes as that of “an attorney defending a client”, as Fagan and Hale point out. [21] It is the same media ecosystem that promotes Yonaguni as a sunken Mu and the Gulf of Khambhat as a lost Hindu civilization, and that even extends, in Jean Deruelle’s largely dismissed hypothesis, to mapping Plato’s “great plain” onto Doggerland itself. Pseudoarchaeology is identified by professional methodologists precisely as the rejection of accepted data-gathering and analytical methods; the petroleum-seismic and AI-assisted “Goldilocks zones” approach now used to target Doggerland geo-archaeological potential exemplifies the systematic methodology that separates submerged-landscape science from the advocate’s case. [34] [35] [36]

The Other Side

On Yonaguni, Boston University geologist Robert M. Schoch — who has personally dived the Monument — argues the structure is the product of natural bedding planes and joints in a sandstone/mudstone sequence, with similar step-like formations onshore, and he judges it over 95 % natural, allowing that parts may have been touched up by ancient islanders. [8] At Bimini, geologist Eugene Shinn found no artifacts and no wheel ruts at the site [19], and scientists who have evaluated the structures found them to be natural beach-rock formations. [39]

For the Gulf of Khambhat (Cambay), the structural setting itself complicates any reading: the gulf was formed by a major rift, it is very tectonically active today, with several faults and periodic earthquakes, and several archaeologists therefore state that the site is not in a secure enough context to be reliably dated. [4] NIOT’s own report (Kathiroli et al. 2003) describes the sonar patterns as “very unusual in the marine domain” and reads them as archaeological; its date comes from a dredged piece of carbonized wood, dated by two laboratories to a calibrated age of about 9,545 years BP. [11] Critics reject the date because dredged material cannot be tied stratigraphically to the features. [4] Even on Doggerland — where the submerged landscape itself is uncontested — Gaffney and colleagues read the Storegga-era core through sedimentary ancient DNA and other proxies and judge the tsunami’s impact on its woodland temporary and likely localized [28], and Walker et al. suggest that traditional notions of a catastrophically destructive event may need rethinking. [29]

What We Actually Know

The case for significant human presence beneath the waves rests on its strongest evidence yet. Doggerland, the now-submerged landscape of the southern North Sea, has long been known to archaeologists as an area of Mesolithic occupation, and Bryony Coles’s formative 1998 paper “Doggerland: a Speculative Survey” rallied interest in its marine prehistory. [40] The first direct archaeological evidence, the Colinda harpoon, was dredged up in September 1931 in a lump of peat from the Leman and Ower Banks: a barbed red-deer antler point later radiocarbon-dated (OxA-1950, 11,740 ± 150 BP, c. 11,950–11,300 cal BC) to the Upper Palaeolithic. [49] In the Gulf of Khambhat, NIOT side-scan sonar surveys in December 2000 identified what was described as an area of regularly spaced geometric structures [4], and the Yonaguni Monument off Japan presents measurable stepped formations at depths of about 5 to 25 m. [5]

The well-documented baseline of all four sites is more straightforward than the popular literature suggests. The Yonaguni Monument is composed of early Miocene Yaeyama Group sandstones and mudstones deposited about 20 million years ago and was discovered in 1986 by diving tour operator Kihachiro Aratake while scouting hammerhead shark dive sites. Marine geologist Masaaki Kimura of the University of the Ryukyus believes he can identify a pyramid, castles, roads, monuments and a stadium. Robert M. Schoch of Boston University concludes the formation is primarily natural — he estimates more than 95% natural [8] — with apparent steps arising where the rock breaks along horizontal bedding planes and vertical joints. Neither Japan’s Agency for Cultural Affairs nor the government of Okinawa Prefecture recognises the site as cultural property, and neither body has carried out research or preservation work there. [6] No definitive human artifacts have been recovered from the submerged structures. Gulf of Khambhat artifacts were recovered by dredging rather than controlled archaeological excavation, and a piece of wood radiocarbon dated to approximately 7500 BC cannot be stratigraphically tied to the sonar features. [4] [41] [42]

The plausible-but-uncertain tier maps the actual mechanics of the debate. Two palaeochannels of former rivers were identified within the Gulf of Khambhat under 20-40 m of water, at a distance of about 20 km from the present day coast [4]. Seismologist Toru Ouchi of Kobe University has sided with Kimura, stating that the Yonaguni structures were not caused by earthquakes. [6] Gifford and Ball directly examined the Bimini Road blocks [18] and found no evidence for a second course of masonry, no tool marks, and no symmetrical supports, noting that the three linear features are unconnected at the southwest end. Archaeologist Richard J. Pearson considers Kimura’s scenario of construction followed by tectonic submergence unlikely. [41] Natural geological processes — including columnar jointing from cooling, differential erosion of layered strata, and coastal erosion — can produce rock formations with geometric features that resemble human construction. Geochemical analysis of sediment core ELF019 from the southern North Sea documents a transition from freshwater silts through terrestrial-vegetation-dominated deposits to saline marine clay. [26]

The speculative tier is where interpretation runs ahead of evidence. Whether the Yonaguni Monument is primarily natural or a man-made stepped construction remains unresolved, and Kimura’s claimed age has shifted from at least 10,000 years ago to 2,000–3,000 years ago [41], invoking tectonic subsidence rather than Holocene sea-level rise as the submergence mechanism. Whether the Gulf of Khambhat sonar features represent anthropogenic structures (a “lost city”) or natural geofacts is similarly unresolved; proponents claim sonar scans revealed an area spanning 94 km at a depth of 30-40 m with dredged pottery, a tooth and bones, and a piece of wood dated to about 9,500 years ago [4], while a low-tier YouTube source claims that by ~2003–2004 geologists identified the features as natural iron-oxide concretions and silty sandstone formations, a reclassification not attested in any peer-reviewed publication cited here. Whether Bimini Road is a natural beachrock formation or anthropogenic construction remains contested, and whether a continuous second course of stones lies beneath the exposed pavement (claimed by Greg Little and others) lacks peer-reviewed documentation [18]. Fringe interpretations connect Yonaguni to the mythical Pacific continent of Mu and to an undefined pre-Ice-Age civilization; French engineer Jean Deruelle’s identification of Doggerland as Plato’s Atlantis fails on both chronological and geographical grounds. [10] [43] [44] [45] [46]

References

[1] The Yonaguni Ruins in Japan | KCP Japanese Language School — https://www.kcpinternational.com/2018/10/the-yonaguni-ruins/ (accessed 2026-08-15)

[2] Dive Into Japan’s Mysterious Yonaguni Monument — https://www.scubadiving.com/dive-japans-mysterious-yonaguni-monument (accessed 2026-08-15)

[3] The Yonaguni ruins… an underwater mystery - Diveplanit — https://www.diveplanit.com/japan/okinawa/yonaguni-monument/ (accessed 2026-08-15)

[4] Marine archaeology in the Gulf of Khambhat — https://en.wikipedia.org/wiki/Marine_archaeology_in_the_Gulf_of_Khambhat (accessed 2026-08-15) [Reputable]

[5] The Deep End of History: What is the Yonaguni Monument? — https://www.discoveryuk.com/mysteries/the-deep-end-of-history-what-is-the-yonaguni-monument/ (accessed 2026-08-15)

[6] Japan’s Ancient Underwater “Pyramid” Mystifies Scholars — https://www.nationalgeographic.com/history/article/yonaguni-jima-japan-underwater-city (accessed 2026-08-15) [Reputable]

[7] Ancient megalithic construction beneath the sea off Ryukyu islands in Japan submerged by post glacial sea-level change — https://www.researchgate.net/publication/4129815_Ancient_megalithic_construction_beneath_the_sea_off_Ryukyu_islands_in_Japan_submerged_by_post_glacial_sea-level_change (accessed 2026-08-15)

[8] Robert M. Schoch: Research Highlights Yonaguni, Japan — https://www.robertschoch.com/yonaguni.html (accessed 2026-08-15)

[9] Exploration Mysteries: Yonaguni Monument - Explorersweb » — https://explorersweb.com/exploration-mysteries-yonaguni-monument/ (accessed 2026-08-15)

[10] A Scientist Says a Baffling Seafloor Structure Marks Japan’s Atlantis — https://www.popularmechanics.com/science/environment/a71748136/japan-atlantis-yonaguni-monument/ (accessed 2026-08-15)

[11] New Archaeological sites in the Gulf of Cambay, India — https://www.episodes.org/journal/download_pdf.php?doi=10.18814/epiiugs/2003/v26i1/003 (accessed 2026-08-15)

[12] Chintan Vedpathak, Archana Das, Siddharth P. Prizomwala, and Nisarg Makwana (2024). Holocene Palaeoenvironmental change at the mouth of Sabarmati River, Gulf of Khambhat, Western India. Quaternary Environments and Humans. DOI 10.1016/j.qeh.2024.100016 - https://www.sciencedirect.com/science/article/pii/S2950236524000148 (accessed 2026-08-15) [Academic]

[13] Chintan Vedpathak, Archana Das, Sunny Tiwari, Nisarg Makwana, and Siddharth P. Prizomwala (2026). Holocene relative sea-level history of the Gulf of Khambhat, Western India: Revisiting regional records with standardized SLIP protocols. Quaternary International. DOI 10.1016/j.quaint.2026.110163 - https://ui.adsabs.harvard.edu/abs/2026QuInt.75910163V/abstract (accessed 2026-08-15) [Academic]

[14] Surprising underwater discovery in India may lead to a lost civilization — https://www.futura-sciences.com/en/surprising-underwater-discovery-in-india-may-lead-to-a-lost-civilization_22388/ (accessed 2026-08-15)

[15] Meetings With the Ancient Teachers of Mankind | Graham Hancock | Megalithomania 2006 (Audio) [70:26-72:22] — MegalithomaniaUK (YouTube) — https://youtu.be/6uNVfcOFaPM?t=4226 (accessed 2026-08-15)

[16] Lost Hindu Civilization They Don’t Want You To Know About [22:26-24:25] — Universe Inside You (YouTube) — https://youtu.be/TWMV0zlUbZE?t=1346 (accessed 2026-08-15)

[17] The Mysteries of the Bimini Road: Ancient, Underwater… and Artificial? — https://www.historicmysteries.com/unexplained-mysteries/bimini-road/32771/ (accessed 2026-08-15)

[18] Bimini Road — https://www.crystalinks.com/biminiroad.html (accessed 2026-08-15)

[19] Bimini Road: Natural Rock Formation Or Path To Atlantis? — https://allthatsinteresting.com/bimini-road (accessed 2026-08-15)

[20] Some Believe The Bimini Road To Be The Remains Of A Wall From Atlantis — https://virginianreview.com/238608/ (accessed 2026-08-15)

[21] The New Atlantis and the Dangers of Pseudohistory — https://www.hallofmaat.com/lostciv/the-new-atlantis-and-the-dangers-of-pseudohistory/ (accessed 2026-08-15)

[22] Underwater “Roads” — https://www.nuscimagazine.com/underwater-roads-a29e175fd78b (accessed 2026-08-15)

[23] Bimini Road in Bailey’s Town — https://en.advisor.travel/poi/Bimini-Road-15197 (accessed 2026-08-15)

[24] Doggerland: Lost ‘Atlantis’ of the North Sea gives up its ancient secrets — https://www.theguardian.com/science/2021/aug/01/doggerland-lost-atlantis-of-the-north-sea-gives-up-its-ancient-secrets (accessed 2026-08-15) [Reputable]

[25] The Last Humans To Roam Doggerland [2:01-3:48] — Stefan Milo (YouTube) — https://youtu.be/kfvnR8ZHi-8?t=121 (accessed 2026-08-15)

[26] Mohammed Bensharada, Alex Finlay, Ben Stern, Richard Telford, and Vincent Gaffney (2026). A Multiple-Proxy Geochemical Investigation of a Shallow Core from Doggerland: Implications for Palaeolandscape and Paleoenvironmental Reconstruction. Humans. DOI: 10.3390/humans6010005 (accessed 2026-08-15)

[27] Mohammed Bensharada, Alexander J. Finlay, Ben Stern, Richard Telford, and Vincent Gaffney (2025). A Multiple-Proxy Geochemical Investigation of a Shallow Core from Doggerland, Implications for Palaeolandscape Reconstruction. Preprints.org. DOI: 10.20944/preprints202510.0855.v1 (accessed 2026-08-15)

[28] Vincent Gaffney, Simon Fitch, Martin Bates, Roselyn Ware, Tim Kinnaird, Benjamin Gearey, Tom Hill, Richard Telford, Cathy Batt, Ben Stern, John E. Whittaker, Sarah J. Davies, Mohammed Ben Sharada, Rosie Everett, Becky Cribdon, Logan Kistler, Samuel Harris, Kevin Kearney, James Walker, Merle Muru, Derek A. Hamilton, Matthew Law, Richard Bates, and Robin G. Allaby (2020). Multi-proxy evidence for the impact of the Storegga Slide Tsunami on the early Holocene landscapes of the southern North Sea. bioRxiv (Cold Spring Harbor Laboratory). DOI: 10.1101/2020.02.24.962605 (accessed 2026-08-16) [Academic]

[29] A great wave: the Storegga tsunami and the end of Doggerland? — https://www.semanticscholar.org/paper/A-great-wave%3A-the-Storegga-tsunami-and-the-end-of-Walker-Gaffney/0b29610c99c17ceee55921d93d85d0ca4784c9a9 (accessed 2026-08-15) [Academic]

[30] The secrets of Doggerland - Wessex Archaeology — https://www.wessexarch.co.uk/news/secrets-doggerland (accessed 2026-08-15)

[31] Marc P. Hijma et al. (2025). Global sea-level rise in the early Holocene revealed from North Sea peats. Nature 639, 652-657. DOI 10.1038/s41586-025-08769-7 - https://www.nature.com/articles/s41586-025-08769-7 (accessed 2026-08-15) [Academic]

[32] Ehud Galili, Liora Kolska Horwitz, Ilaria Patania, Amir Bar, and Isaac Ogloblin Ramirez (2024). Identifying Anthropogenic Versus Natural Submerged Prehistoric Landscapes: Two Case Studies from the Sicilian Channel. Journal of Marine Science and Engineering 12(11), 1981. DOI 10.3390/jmse12111981 - https://www.mdpi.com/2077-1312/12/11/1981 (accessed 2026-08-15) [Academic]

[33] Identifying Anthropogenic Versus Natural Submerged Prehistoric Landscapes: Two Case Studies from the Sicilian Channel (preprint) — https://www.preprints.org/manuscript/202410.0500 (accessed 2026-08-15)

[34] Edgar Cayce: Tales From the “Sleeping Prophet” — https://www.ripleys.com/stories/edgar-cayce-tales-from-the-sleeping-prophe (accessed 2026-08-15)

[35] Bimini Road: The Underwater Stones Linked to Atlantis - Apple Podcasts — https://podcasts.apple.com/us/podcast/bimini-road-the-underwater-stones-linked-to-atlantis/id1820661946?i=1000778711521&l=zh-Hant-TW (accessed 2026-08-15)

[36] Bimini Road: The Underwater Stones Linked to Atlantis — https://play.headliner.app/podcast/a1b0f97ac16b4976b709bb10b55382e5/episode/8095a13a-1841-4ef3-80e1-03875cdcc2e7 (accessed 2026-08-15)

[37] Astrid J. Nyland, James Walker, and Graeme Warren (2021). Evidence of the Storegga Tsunami 8200 BP? An Archaeological Review of Impact After a Large-Scale Marine Event in Mesolithic Northern Europe. Frontiers in Earth Science 9, 767460. DOI 10.3389/feart.2021.767460 - https://www.frontiersin.org/journals/earth-science/articles/10.3389/feart.2021.767460/full (accessed 2026-08-15) [Academic]

[38] Bernhard Weninger, Rick Schulting, Marcel Bradtmöller, Lee Clare, Mark Collard, Kevan Edinborough, Johanna Hilpert, Olaf Jöris, Marcel Niekus, Eelco J. Rohling, and Bernd Wagner (2008). The catastrophic final flooding of Doggerland by the Storegga Slide tsunami. Documenta Praehistorica. DOI: 10.4312/dp.35.1 (accessed 2026-08-15) [Academic]

[39] Was Atlantis Real? 6 Theories About the Lost Civilization — https://www.history.com/articles/top-6-theories-about-atlantis (accessed 2026-08-15)

[40] Internet Archaeol. 22. Fitch. In Sight of Doggerland — https://intarch.ac.uk/journal/issue22/fitch_index.html (accessed 2026-08-15) [Academic]

[41] Yonaguni Monument — https://en.wikipedia.org/wiki/Yonaguni_Monument (accessed 2026-08-15) [Reputable]

[42] Yonaguni: Monumental Ruins or Natural Geology? — https://ahotcupofjoe.net/2017/10/yonaguni-monumental-ruins-natural-geology/ (accessed 2026-08-15)

[43] In search of the lost continent of Mu, Asia’s answer to Atlantis — https://www.thetimes.com/world/asia/article/japan-yonaguni-sea-bed-ruin-ctnjcqlrl (accessed 2026-08-15)

[44] Forbidden Archaeological Evidence of Atlantis: Forgotten Civilization [8:56-10:47] — Universe Inside You (YouTube) — https://youtu.be/g2h8O0Fni8g?t=536 (accessed 2026-08-15)

[45] What You MUST Know about Lost Ancient Civilizations [4:32-6:14] — Bright Insight (YouTube) — https://youtu.be/0aP8_V39yxY?t=272 (accessed 2026-08-15)

[46] Peru’s Megaliths, Atlantis Myths & Moon Truths | Ancient Mystery Podcast with Aaron Hynes [3:15-5:07] — Nikkiana Jones (YouTube) — https://youtu.be/GxR-A9uhi_E?t=195 (accessed 2026-08-15)

[47] Tine Missiaen, Simon Fitch, Rachel Harding, Merle Muru, Andy Fraser, et al. (2021). Targeting the Mesolithic: Interdisciplinary approaches to archaeological prospection in the Brown Bank area, southern North Sea. Quaternary International 584, 141-151, DOI 10.1016/j.quaint.2020.05.004 - https://www.vliz.be/imisdocs/publications/53/371153.pdf (accessed 2026-10-04) [Academic]

[48] Johannes van der Plicht, L. W. S. W. Amkreutz, M. J. L. Th. Niekus, J. H. M. Peeters, and B. I. Smit (2016). Surf’n Turf in Doggerland: dating, stable isotopes and diet of Mesolithic human remains from the southern North Sea. Journal of Archaeological Science: Reports, DOI 10.1016/j.jasrep.2016.09.008 - https://research.rug.nl/en/publications/83c2183f-9928-4931-a59b-8c613919dd68 (accessed 2026-10-04) [Academic]

[49] Upper Palaeolithic antler harpoon, Leman and Ower Banks, Doggerland (North Sea, poorly located), NHER 11171 — Norfolk Heritage Explorer — https://www.heritage.norfolk.gov.uk/record-details?MNF11171 (accessed 2026-10-20)