Island secrets discovered beneath the waves

We are taking a look back at stories from Cosmos Magazine in print. In December 2020, Patrick Nunn looked at how geoscience can explain why some islands have partly or completely disappeared, sometimes abruptly.

The idea that an entire island can disappear seems unbelievable, anathemic to our understanding of the natural world, even offensive to rational thinking. Something substantial and upstanding shouldn’t vanish in less time than it would once have taken you to walk around its perimeter. Yet this has undeniably happened in many places, and many times.

Today you won’t find an island named Tolamp on any map. But if you travel among the smaller islands off the east coast of 2041 sq. km Malekula, the second-largest island in Vanuatu, in the southwest Pacific, and ask about it, chances are you’ll learn where it once was and how, several hundred years ago, it abruptly disappeared. Local people may tell you how two mischievous boys, left behind on Tolamp after all the adults travelled to the mainland, went to the other side of the island – where they had been forbidden to go – and dug a hole into which the ocean poured and eventually submerged the entire island. As with most such stories, the details are undoubtedly apocryphal – blame is a common way of rationalising such events – but the recollection of submergence is undoubtedly true.

Owing to their implausible details (boys digging holes), Western science has often dismissed such stories as mythical, but in the case of Tolamp the evidence that stories are essentially eyewitness accounts of sinking islands is compelling. Geologically, Vanuatu is young and dynamic, exactly the kind of place you might expect to encounter stories of vanished islands.

It’s no mystery where Tolamp is today: its summit lies under some 15 metres of ocean, about 5 kilometres from Malekula’s northeast coast. Mapping of the sea floor in the area shows that its disappearance was not without precedent: many other submerged islands project from the steep slopes undersea below.

 People were living in the islands of Papua New Guinea (PNG) more than 40,000 years ago. Most other island groups in the western Pacific were first occupied around 3000 years ago, with those in the eastern and southernmost Pacific quite a bit later; both Easter Island and the islands of New Zealand are thought to have been first settled around 700 years ago. So it’s unsurprising that Pacific Island stories exist recalling the disappearance of islands like Tolamp. Yet there is evidence that such oceanic islands have periodically collapsed and disappeared throughout the hundreds of millions of years the Pacific Basin has been in existence.

A great example is provided by the island of Oahu in Hawaii, which is formed from the remaining halves of two giant volcanoes, connected in the centre island by a low-lying saddle (where some of the tinned pineapple in your supermarket probably originated). The southwest half of the volcano that forms much of western Oahu slipped into the sea to create the undersea Waianae Slide more than three million years ago. A million years later a similar event in Oahu’s northeast formed the massive Nuuanu Slide – 5000 cubic kilometres of crumbled volcanic rock tumbling chaotically down the island’s steep underwater flanks, coming to rest hundreds of kilometres away from its coastline beneath hundreds of metres of ocean.

Along the spectrum from vanished Tolamp to half-vanished Oahu, there are numerous other such islands, investigations of which can reveal the reasons for their disappearance. Take the example of Ritter Island in PNG, about 100 kilometres off the east coast of the main island. This island volcano was first named Cone Island and sketched in 1834 by Thomas Jefferson Jacobs. Early in the morning of 13 March 1888, a relatively minor eruption of the Ritter volcano caused the island’s western flank to collapse abruptly, decapitating the conical landmass that Jacobs drew and leaving only a tiny portion of the original. Should another collapse of one of Ritter’s undersea flanks occur, the entire island is likely to disappear.

Science has developed a mistrust of information about phenomena like vanished islands obtained from what it regards as non-scientific sources. One reason for this is self-protection, a reluctance to muddy processes of scientific deduction with potentially unreliable data – a fine example of which is provided by the tale of Atlantis.

Around 350 BCE, the Greek philosopher Plato invented the story of Atlantis as a means of illustrating his views about how an ideal society might evolve – and how it might be undone, punished by the gods by being violently and abruptly submerged. Atlantis never existed, but Plato was not above extracting details of real events from the geologically active eastern Mediterranean where he lived to enliven his narrative, to enhance both its believability and memorability.

 One of the undoubted influences on the original story of Atlantis was the eruption of ancient Santorini (Thera) in the Aegean Sea. Like other island volcanoes nearing the end of their active lives, it underwent a catastrophic terminal eruption about the year 1600 BCE that saw the shallow magma chamber beneath the island empty explosively, leaving a massive void into which most of the old volcano slipped. People might understandably have thought the volcano was gone forever but, ominously, about 1400 years later, eruptions started within its undersea caldera, later resulting in the formation of two small active volcanic islands, Palea Kameni and Nea Kameni. One day these volcanoes will build another massive island volcano that will in turn explode and collapse, ending the cycle, and laying the foundations for a new one to begin.

 People lived on Santorini before it exploded and mostly disappeared. Ancient Akrotiri, the Minoan-era forerunner of modern Akrotiri, was built around 3700 years ago. It featured several multi-storey dwellings facing paved streets with sub-surface drainage; the town prospered from maritime trade and commerce. Although Akrotiri was destroyed in the 1600 BCE eruption, the absence of human remains suggests its people evidently read the signs of impending disaster sufficiently clearly to evacuate the island before this event. But their stories were preserved: some buried in dubious places like Plato’s account of Atlantis; others in oral traditions and written histories; yet others in paintings and in folklore. All of this material, complemented by scientific research, allow the nature and sequence of this catastrophe to be reconstructed.

This leads us to speculate about vanished islands that exist in traditional stories but of which science is completely ignorant. In some Indigenous Australian stories, the founding ancestors of particular groups, especially in the continent’s northeast, are said to have come from a now-disappeared island named Baralku. When such stories were first written down a hundred years or so ago, it was easy to dismiss them as fantastic. But perhaps it’s time to treat such stories more seriously, to regard them not as myth or legend, but as memories.

In Yolngu stories, Baralku is said to have been the place to the east of Arnhem Land (where lies the modern Gulf of Carpentaria) from which came the Djanggawul, spiritual beings responsible for creating the original landscape of Australia and introducing all the plants that originally grew there. Most of us treat such tales as fictional creation stories, but consider that during the last ice age, about 70,000 to 12,000 years ago, there was a land bridge connecting northeast Australia with the islands of PNG. So there was, indeed, land east of Arnhem Land. Baralku may have been the name of one of the last pieces of that land to be submerged, as sea level rose in the aftermath of the ice age.

The idea that the Djanggawul of Baralku created the Australian landscape and its flora could well recall the time, perhaps about 10,000 years ago, when the people of Baralku were forced from their island home by rising sea level and took their environmental knowledge and skills, honed by their ancestors on the now-submerged land bridge, into the rest of Australia. Far-fetched? Not when you consider that linguists believe that the most widespread group of Aboriginal Australian languages (Pama-Nyungan), once spoken across 90% of the continent, originated on this land bridge and were dispersed as the bridge progressively drowned. LANDS OF LEGEND Long ago, in the low equatorial Pacific islands of Kiribati, more than 4000 kilometres northeast of Australia, people were accustomed to seeing ghost islands, named abaia anti. These islands might one day be present – vegetated and habitable piles of sand and gravel on familiar coral-reef platforms – but the next day be gone without trace. Later these islands might apparently re-appear elsewhere. The world views of the people of Kiribati once held there to be parallel worlds to the visible material one; like people, islands also often moved between these worlds.

It’s plausible to suppose that ideas about such ghost islands and even the parallel worlds through which they were said to pass derive from observations of superficial islands – sand cays. These often appear on reef flats in the region, typically after storm waves carry reefal debris from undersea slopes and dump it onto shallow submerged reef surfaces. Similar events can also remove such islands – powerful waves can sweep over reef flats and wash away their superficial sediment cover.

Sand cays are really the only types of islands that can disappear as a result of extreme-wave impact – or even progressive sea-level rise (see “Islands lost to rising seas”, page 90). The tsunamis that followed the 1755 Lisbon Earthquake shredded islands along the Algarve coast of Portugal just as the 15-metre high tsunamis generated by the 2004 Indian Ocean Earthquake washed away islands in the Nicobar group, killing around 7000 residents – about onefifth of the islands’ inhabitants. But this event did not significantly damage other low Indian Ocean islands, notably those in the Maldives, because the comparative steepness of their undersea flanks slowed the tsunamis and did not allow them to build as high as was the case on gentler slopes.

A likely disappeared island that lies at the nexus of science and memory is Thompson Island in the South Atlantic Ocean. First located on 13 December 1825 by George Norris, captain of the whaler Sprightly, just over 70 kilometres northeast of Bouvet Island (now the Norwegian dependency of Bouvetøya), Thompson was seen again in 1893 by Joseph Fuller, captain of the Francis Allyn. It has not been seen since. Several searches were made for Thompson Island, including one by the well-outfitted 1927 Norvegia-Ekspeditionen, which made the first landing on nearby Bouvet, but to no avail.

Thompson Island was important to British colonial aspirations in the early 20th century, and in the 1920s the British Foreign Secretary stated in the House of Commons that “there did not appear to be any ground for questioning the existence of Thompson Island”, a view shared by nautical and geological experts. How did such a landmass come to disappear sometime between 1893, when Fuller sighted it, and 1898 when the Deutsche Tiefsee-Expedition was unable to re-locate it?

In weather records of the time from Chile and New Zealand, which have similar latitudes to Thompson Island, there is a period of cooling (1896–1907) that does not show up in records elsewhere. There’s no obvious explanation for this, yet we know that volcanic eruptions have caused similarly enduring episodes. It’s possible that Thompson Island was an active volcano, like we know Bouvet Island to be, which blew itself up some time in 1896, causing its own disappearance. An event unnoticed and unrecorded by anyone, and a tantalising mystery.