The strange story of Sandy Island, the remote Pacific outcrop that never existed

Chris Leadbeater
·12-min read
Sandy Island on a 1908 map
Sandy Island on a 1908 map

The RV Southern Surveyor was sailing through the sunlight of a warm summer day when Dr Maria Seton noticed that something was wrong.

November 22 2012 should have been an ordinary date in the diary for this doughty Australian research vessel, as it studied the lay-out of the seabed in the Coral Sea, some 900 miles east of mainland Queensland. But after a few minutes of confusion, the geologist on its deck realised that, somehow and somewhere, the ship had left the chart. Seton and her colleagues were travelling freely across a horizon-wide patch of open ocean. The trouble was, the map said otherwise. In its esteemed opinion, they had just washed ashore – onto an island the size of Manhattan.

Seton had just done something that went against the logic of exploration, and would overthrow the accepted “knowledge” of 136 years. She had “undiscovered” Sandy Island.

What had been marked – since as far back as 1876 – as a solid landmass, some 15 miles long by three miles wide, was nothing of the sort. Not even close. “We had a cached version of Google Earth for the area – as we had no internet,” Dr Seton would explain. “We saw that the island was depicted as a big black blob. This made us very suspicious”.

It did not take long to ascertain the reality of the situation. The Southern Surveyor was in the midst of a 25-day scientific expedition to examine the tectonic contours of this Australasian corner of the Pacific – and was able to see, at a glance, what lay beneath the location where Sandy Island was supposed to be. In short, water – and a lot of it. “We would find that the ocean floor didn’t ever get shallower than 1,300m [4,265ft] below the wave base,” Dr Seton added. “We sailed over it. We’re not sure how it got onto the maps. There must have been an error on one of the coastline data studies that has been propagated through scientific literature – our weather maps on board showed the island.”

“We all had a good giggle at Google when we sailed through the island,” another member of the team, Steve Micklethwaite, told The Sydney Morning Herald. “Then we started compiling information about the seafloor, which we will send to the relevant authorities.”

There was – in a very real sense of the familiar phrase – nothing to see here. And yet, there had been firm talk of a Sandy Island (or a Sable Island) in this remote position – 19.22°S 159.93°E – since the final quarter of the 19th century, and reference to it a full century earlier.

Captain James Cook at Botany Bay - Getty
Captain James Cook at Botany Bay - Getty

On September 14 and 15 1774, during his “Second Voyage” into Australasian waters, Captain James Cook charted a “Sandy I” at co-ordinates of between 19°S and 20°S, and 163°E and 164°E. This was not the vague ghost that would still be popping up on Google in September 2012 – Cook made few errors, and his sighting corresponds to the Grande Terre barrier reef which surrounds the French territory of New Caledonia, a further 300 miles to the south-east, on the outer edge of the Coral Sea. But the name – admittedly a common one for a random sliver of land in an enormous expanse of ocean – appears to have stuck. In 1876, the Velocity, a whaling vessel, would pass through the area – and would identity a “Sandy Island” in its path. A phantom was born.

It seems that the Velocity sailed out of the Tasmanian capital Hobart on April 12 1876, under the captaincy of one JW Robinson – returning to the same harbour almost a year later, on March 20 1877. A report in the city’s newspaper the Mercury suggests that it had a troubled voyage, running into heavy winds in the Coral Sea, and springing a leak that would force it to drop anchor for repairs in the uninhabited Chesterfield Islands – a series of small outposts scattered to the north-west of New Caledonia. From here, the vessel lumbered north to San Cristobal in the Solomon Islands. En route, it spotted what it took to be an unknown segment of land. It would report its “discovery” when it arrived home.

A rufous fantail on San Cristobal - Getty
A rufous fantail on San Cristobal - Getty

Within three years, the error had been accepted as confirmed fact. In 1879, Australia’s Admiralty Hydrographic Office (a forerunner of the modern-day Australian Hydrographic Service) published a directory of maps of Australian waters – a navigational aid for shipping in the region – which stated that “in 1876, the master of the whaler Velocity reported that, while cruising on the eastern side of the Chesterfield and Bampton reefs, he observed heavy breakers in the latitude 19.83°S, longitude 158.83°E. The master of Velocity also reported a line of sandy islets extending north and south along the meridian of 159.95°E, between the latitudes 19.12°S and 19.33°S.” The error was now in writing.

And it would stay that way. Returning to the University of Sydney, Dr Seton would look into how a non-existent island came to be inked onto global maps. Her research showed that Sandy Island had made it to the British edition of the admiralty map by 1908, before seeping into information banks worldwide. By 1982, it had arrived in the American military database, reproduced on a chart issued by the US Defense Mapping Agency (DMA). From there, it was a short hop to online services like Google. Seton’s curiosity took her to the World Vector Shoreline (WVS) database – a facility developed by the US military, which transformed old-fashioned hard-copy charts into digital maps – where the error had been copied across without question. “Inconsistencies in data exist [for] some of the least explored parts of our planet,” she would write in an article for Eos, the house journal of the American Geophysical Union (AGU). “[This is] a function of both human digitising errors, and errors in the original maps – from which the digitising took place.”

Phantom islands are surprisingly common - Getty
Phantom islands are surprisingly common - Getty

Google corrected its maps on November 26 2012, less than a week after the Southern Surveyor’s “undiscovery”; the National Geographic Society followed three days later. Sandy Island was no more. But not everyone had been so late to the realisation. France, for one, had long had its concerns. Well versed in the waters of the Coral Sea – New Caledonia is a French territory – it had deleted the island from its charts as early as 1974. Other maps had already labelled the island as “ED” – for “existence doubtful”. Picking up the trail after the Australian media had widely reported Dr Seton’s news, Shaun Higgins – a New Zealand librarian – delved into the archive at his place of employment, Auckland Museum. He unearthed another map from the Admiralty Hydrographic Office, this time dated to 1908, which, though it named the Velocity, identified Sandy Island using a dotted circle. “It has the same shape as on Google Earth, but it’s dotted,” he told the Sydney Morning Herald. “And it could have been dotted because it was an unidentified hazard”.

Was this what had happened? Had the crew of the Velocity – worried about rough seas and a ship that was stubbornly taking on water, and more focused on their hunt for prey and pay than any novelties of geography – seen something in the Coral Sea; just not an island?

Higgins would also tell the paper that, although the original mistake was probably the whaling ship’s, he doubted that it was intentional. “They could have been further west, where there are reefs,” he continued. “The primary function of whalers was to look for whales. [But] I think any responsible captain, if they spot something, would mark it down – particularly at this time, in the Pacific, when whalers were often [the] first explorers.”

Whalers were among the first explorers - Getty
Whalers were among the first explorers - Getty

But if JW Robinson and his men did not see an island, then what was it? In her article for Eos, Dr Seton would present an answer to this question, suggesting that, while the crew did see solid matter in the water, it was only surface-deep, only temporary – and volcanic.

Thanks to the fractious condition of the seabed in the region, the South Pacific and the Coral Sea are known for cases of “pumice rafts”. These are tranches of ultra-light buoyant rock, created by the cooling of lava spewed upwards from active fissures in the ocean floor. The rapid loss of heat once in the water causes a proliferation of bubbles, which allows the stone to float – and, depending on the eruption, can create formations of a significant size.

In August 2012, the HMNZS Canterbury spotted an enormous pumice raft floating in the sea to the north-east of New Zealand. The naval vessel was sailing from Auckland to the outpost of Raoul Island, and had to take care when passing an obstacle that was jutting almost a metre above the surface. Satellite images would measure it as being 300 miles long and 30 miles wide – about 10,000 square miles in area. It was deemed to be the result of an eruption, probably on July 18-19 2012, in the Havre Seamount – a submarine mountain near New Zealand’s outlying Kermadec Islands (of which Raoul Island is the largest). The stones would disperse over the next three months, but patches of the raft would still be washing up on the Australian coast over a year later.

Such sightings can cause concern for passing mariners – in the 21st century, let alone the 19th. In August 2006, travellers on the yacht Maiken, journeying in Tongan waters, were astounded to sail into “a vast, many-miles-wide belt of densely packed pumice” after leaving the sanctuary of the Vava’u islands. Photos posted on the yacht’s blog show the stones lying flat and silt-like on the surface, giving the impression – to the untrained approaching eye – of a low-slung landmass. They would also encounter Home Reef – a volcano hidden just below the waves, north of Tonga and east of Fiji, which has a habit of breaching the surface. Its eruptions have repeatedly created an “ephemeral island” – a temporary landmass of loose new rock that can be a hazard for shipping, but is usually eroded back into invisibility by the ocean’s movements. The August 2006 eruption was not the first time Home Reef had appeared – it was also spotted in 1852, 1857 and 1984.

It is unlikely that the whalers of the Velocity stumbled across a concealed volcano in 1876 – as mentioned earlier, Maria Seton’s studies of the seabed below “Sandy Island” showed nothing but deep water. But it is not difficult to comprehend why mariners of that era, distracted by the perils of their main job, might have mis-identified an apparent landmass.

Nor is there any embarrassment to this. They would not be the last sailors to make such an error. Witness the unusual case of Dougherty Island – another outcrop that was thought to exist in the South Pacific, this time between Cape Horn and New Zealand. It was reported in 1841 by the eponymous Captain Dougherty, an English mariner aboard the whaling ship the James Stewart, who declared he had sighted it at 59.33°S, 120.33°W. He described it as a some six miles long, and covered in snow. His “discovery” would be corroborated in 1860 by a Captain Keates of the Louise, who “identified” the same place, albeit at slightly different coordinates (59.33°S, 120.3°W) – and again by a Captain Stannard of the Cingalese in 1886 (at 59.35°S, 119.12°W). No islet exists in this location.

A map showing the phantom island
A map showing the phantom island

The three men had probably all encountered the same thing – either a fog bank or a series of icebergs. Crucially, none of them tried to anchor off Dougherty Island, or they might have understood their error. Nonetheless, it could be found on maps until as late as 1934.

But then, such mistakes can happen to the best of seafarers. Step forward the USS San Francisco, a nuclear submarine (now retired from active service) in the employ of the United States Navy – which suffered a similar yet directly contrasting encounter on January 8 2005. Navigating in Pacific waters some 420 miles south-east of the US Territory of Guam, it collided with an unmapped seamount that was “marked” on one of its maps as nothing more than discoloured water. The incident was almost disastrous. The submarine was moving at full speed, at a depth of 160m (525ft), making for an impact so severe that one crewman died (of head trauma) and 98 were injured. The vessel would struggle to re-surface due to a rupture in its forward ballast tanks. It limped into Guam two days later. Images showed it in a dry dock, its nose heavily bandaged. It would still be there in the May – and ultimately, would require a transplant, taking on the bow of the USS Honolulu, which was due to be decommissioned. It would not be until April 2009 that it would fully reappear on duty, after an overhaul that cost some $79million (£58m).

The submarine’s captain, Commander Kevin Mooney, would bear the brunt of the responsibility. He was reassigned to a shore unit in the immediate aftermath, and later stripped of his command – even though the investigation found that the US Navy’s charts did not have the latest geographical data, because the area was not a priority for the DMA. The accident could have been considerably more serious – but it was a reminder that, even in the face of 21st-century technology, the sea is ready to play tricks on the eye.