Mid-oceanic rifts should be in the middle of the ocean they formed. And often they are, but there are exceptions. The Reykjanes Rift, south of Iceland, is one of these. It is well known for its connection to Iceland, and events there are often discussed in this blog. Reykjanes has separated Greenland from Scotland, and as spreading rifts spread symmetrically, with both sides moving away at the same speed, a quick revisit to the geometry lessons of our youth tells us that both sides will have moved the same distance. Reykjanes should thus be mid-way between Greenland and Scotland, equidistant from both. But its location is nowhere near. It is 500 kilometer from Greenland but 1000 kilometer from Scotland. It seems to be fleeing from the scary Scots. What happened?
450 kilometer west of Scotland, out in the Atlantic Ocean, is a single rock. The name, Rockall, is Norse: Rok means spraying, stormy sea. (In old Saxon or Frisian it means smoke, showing how the word was transferred to a similar phenomenon with a very different origin between in-land people and sea farers). Indeed, Rockall sticks out of an oft wild sea, and the steep sides and the massive waves make landings rather hazardous. The rock has a size of 30 by 25 meter, peaking 17 meter above sea. There are a few other rocks, smaller ones, nearby, but Rockall is the only one permanently above the sea. (Although ‘permanently’ does not mean the same as ‘always’: nearby, a wave was recorded at 18.5 meter above average sea level which would have completely submerged it). Rockall thus qualifies to be called an island – one of the most isolated in the world. Of course, that made it desirable property. The UK claimed ownership in 1955, against competing interest from Ireland and The Faroe, perhaps as consolation for losing India. The ownership was confirmed by the UN only three years ago. It is now officially part of fearsome Scotland.
Isolated rocks in the middle of an ocean are a hazard. One 19th century captain tells the story of chasing a ship with very white top sails but dark lower sails, only to discover it was this rock: we discovered, on running close to this mysterious vessel, that we had been actually chasing a rock – not a ship of oak and iron, but a solid block of granite, growing as it were, out of the sea, at a greater distance from the main land than, I believe any other island or inlet or rock of the same diminutive size, is to be found in the world. What they had seen, of course, was the dark granite rock with a topping of bird droppings. Worse happened: one of the smaller rocks sank the SS Norge, a few years before the Titanic, which is among the worst peace-time disasters on the Atlantic Ocean.
Rockall, this little rock, isolated and unknown, holds the answer to our problem. Here is the reason for the misplacement of the Reykjanes ridge, and thereby, of Iceland.
The Rockall sunken continent
The presence of this single rock, so far from anywhere, shows that something is going on beneath the sea. And so it is. Peel away the ocean, and you will find that Rockall is the tip of a vast submerged plateau, 450 kilometer in length and around 500 to 1000 meters deep which in the Atlantic Ocean counts as shallow. The plateau is not oceanic: it is a sunken fragment of continent, about the size of the British Isles, of which only 800 m2 (!) is not submerged. This is the real Atlantis (by name – don’t take this literally!), a micro-continent – although judging by the sub-aerial area, ‘nano-continent’ might a better description. The plateau is separated from Ireland and Scotland by the steep-sided Rockall trough, 3 kilometer deep and 250 kilometer wide. One of the early predictions of continental drift was that the Rockall Plateau should be continental in origin: it was needed to fill in a gap between Europe and America when trying to close the Atlantic Ocean. This turned out to be correct, a validation of the theory.
The Rockall plateau consists of two main parts, the higher eastern bank, and the lower western bank which is more broken up. The two main parts are called the Rockall Bank (a certain uniformity in naming is apparent), and the Hatton Bank; the basin between the two is imaginatively called the Hatton-Rockall Basin. Some smaller components have more unexpected names: Lorien Knoll, Edoras Bank, Rohan Seamount, to name a few. Who knew Tolkien’s Middle Earth was an archipelago of Atlantis? This finally explains the mystery of the Ent-wives. The southern edge is the Charlie-Gibbs fracture zone which runs east-west into the Atlantic. If you think you now can guess the name of its discoverer, there might still be a surprise. ‘Charlie’ is the name of the nearest ocean weather station, and Josiah Willard Gibbs was the name of the vessel which first mapped the fracture zone. Gibbs was an American theoretical physicist, brilliant but working at a time when science was based in Europe, not America. If he had worked in Europe, he would certainly have won the first Nobel prize. His name is still found in physics and chemistry text books worldwide, but few people outside of these sciences will recognize the name. Half a name of an oceanic fracture zone is his consolation.
Both the Hatton bank and Rockall bank have a unique fauna of cold-water corals, now under treat from destructive fishing methods which involve scraping nets over the bottom to make sure nothing escapes (perhaps taking a clue from common financial management practices and therefore considering that a bank needs a bail-out). Parts of the bank are considered for conservation status, but although mandated by the UN and submitted to the European Commission, this protection has not yet been implemented. Out of sight is out of mind.
The rocks of the plateau have a large variety of ages. These have been measured from a number of drilling programs, which were mainly done by groups looking for hydrocarbons – oil. The base rocks are old, at 1.9 billion years. This is not quite as old as the Lewisian rocks of Scotland and the Hebrides, but is impressive none-the-less. The area is clearly a part of the ancient core of the northern continents. Rocks of this age run from Canada and Greenland to Scandinavia, and perhaps northern China. At times they were connected; at other times they were separated by deep oceans. They weren’t always in the deep north: over time, the different continental building blocks have drifted up from the south. It is hard to be sure where on Earth this was 1.9 billion years ago and where the travels have taken it since. At the speed of continental drift, the Rockall rocks may have circumnavigated the Earth twice over.
But the drilling mostly brings up much younger rocks. A thick layer of basalts dated to 50-60 million years ago has covered the entire region. Rockall Islet itself is part of that: it is made of a granite peculiar to Rockall, of this same date. Often Rockall Islet is listed as of volcanic origin. It may perhaps instead be called plutonic, but the underlying plateau definitely is volcanic. (If you don’t like plutonic, the Atlantic location suggests ‘neptunic’ as alternative.)
The Rockall plateau contains a number of separate volcanic complexes, one of which surrounds Rockall Islet. In the deep trough between the Rockall Plateau and the continental shelf are a number of seamounts: Rosemary seamount, Anton Dohrn seamount, and Hebrides terrace seamount. The first two are named after the vessels that discovered them. Rosemary seamount is possibly the tallest mountain of the British Isles, 2 kilometer from base to top, although far from the highest as the top is 300 meter below sea level. All three are volcanic in origin, and are dated (not very accurately) to between 40 and 70 million years ago. For the Hebrides terrace seamount, a large gravity anomaly is seen, which has been interpreted as due to a 17-kilometer thick body of solidified magma underneath the seamount. These were massive volcanoes.
All of this is part of the Atlantic large igneous province, immense basaltic eruptions, the remnants of which are found here, and on the Faroe, and on the east coast of Greenland, and on the west side of the British Isles. To have been affected by the same flood basalt event means that these distant regions were close together at the time. Therefore, this was erupted before the Atlantic Ocean had begun to spread. Greenland, Rockall, and Scotland were part of one continent; they once were located not at the coast, but in the middle of the most recent supercontinent, Eurasia. This continent was zipped apart by the opening of the North Atlantic Ocean, with the zip running between on one side the British Isles and Norway, and America/Greenland on the other. The massive eruptions came during this unzipping.
The map of the Atlantic Ocean shows the Reykjanes Ridge running southeast from Iceland. It is clearly much closer to Greenland than it is to the British Isles. The Rockall Plateau is visible in the Atlantic Ocean: the Reykjanes Ridge is rather better centred between it and Greenland. This already shows that the Reykjanes spreading centre cut the connection between the Rockall Plateau and Greenland, and pushed them apart. The Rockall plateau is therefore the true westernmost point of Europe, as well as the eastern boundary of the true Atlantic Ocean. We are standing, or rather swimming, on an ancient border. Once, from here it was short walk or swim to southeast Greenland, in the subtropical conditions of the time. Now, one side is a frozen wilderness, covered in glaciers, and the other side is deep under water. The break-up didn’t work out for either of them.
Forming the Atlantic Ocean
So what happened? And when did it happen? Why were both Greenland and the Rockall Plateau left as flotsam, continental debris in the Atlantic?
Breaking up a continent is hard to do – and it takes time. Not everything goes at once. Modern examples (with some liberty on what counts as ‘modern’) are found around Africa: it has been splitting for a long time, but has been amazingly resilient. India split off first, Madagascar next, and now Somalia seems on the way out. The most recent splitting began with the Afar hot spot, which arrived around 30 million years ago in the appropriately named Afar region, at that time within the African-Arabian continent. The hot spot created a triple junction, with one rift arm slowly extending to the south. But even now, after tens of millions of years, it has not yet managed to break Africa apart. When it does, Somalia will join the great trek into the Indian Ocean. Afar has done better on its western arm, managing to largely split Arabia off from Africa, but it has taken 30 million years to progress up the Red Sea and the process is still not complete – the last part, Suez, had to be dug by hand.
The Atlantic Ocean was similarly reticent, and tortuously slow in forming. The South Atlantic started with the splitting up of Gondwana, 130 million years ago. The North Atlantic was more of an afterthought, starting when Iberia separated from America about 110 million years ago. From here, a rift migrated north. The connection between Greenland and Scotland was the last to give way. It did so with a tremendous outpouring of basalt – as mentioned above.
But this process was far from straightforward, with many failed attempts and rifts which lost momentum and died. On both sides of the future ocean there were great complications. A brief spreading centre operated between Canada and southern Greenland, separating the two and forming the Labrador Sea. This separation may have started already 92 million years ago. It began, as so often, with continental stretching. A basin formed at this time where the crust thinned. The thinner crust could not support the weight and the basin sank, dropping to below sea level. The Ethiopian rift is currently in such a sinking phase, just about to let in the sea. Sea floor spreading started 80 million ago, with a spreading rate of about 1 cm per year (this is the ‘half rate’, the speed at which the ocean floor moves away from the centre. The sea widens at twice this rate.) About 60 million years ago a triple junction formed south of Greenland – at that time, the spreading rate in the Labrador sea briefly doubled. Subsequently, spreading in the Labrador Sea slowed, and the spreading centre became extinct by 40 million years ago. It never quite completed the job and Baffin Island remained in close contact with Greenland.
This stretching of the crust in the late Cretaceous also happened next door, on the other side of the non-existent Atlantic Ocean. Basins and rifts began to form in several locations, places where the crust extended and thinned, and the land dropped. Some of these were more successful than others. The Rockall trough was one of the more successful basins, growing northward from Iberia around 90 million years ago but never becoming a real ocean. One paper calls it ‘quasi-seafloor spreading’, whatever that means: the crust is thin but it is in fact continental, not seafloor. At this time, the area between Greenland and Europe may have been a bit like the western US, with a series of basins (although lacking the intervening mountain ranges) or it may have been a triple junction with the Iberia rift, Labrador and the Rockall trough forming the three arms. The Rockall trough developed into a sea arm and the Plateau was now separated from the British Isles, but it wasn’t an island and the Rockall trough did not develop into a spreading centre, or if it did so it was very briefly. Europe and Greenland were still connected.
Another failed attempts to complete the ocean was a rift extending from Baffin Island southeastward, to the Hebrides, 62 million years ago. It appears to have split there, with one branch going into the central North Sea, and one branch extending southward towards the Bristol channel, along the western side of Scotland and England. If this had succeeded, a part of the UK and most of Ireland would now have been on the other side of the Atlantic, an early brexit. But it became a failed rift. The attempt left a legacy of faults which occasionally shake up the UK, and caused the volcanic marvels of Skye and Mull, as a reminder of what could have been. The seamounts also developed during this time, 60 million years ago, mostly underwater but the peaks may have formed islands.
But finally a spreading centre did develop in a basin on the western side of the Rockall Plateau, fed from a new triple junction southwest of Greenland. Did the older Rockall Trough triple junction move there? Or was it a separate event? In any case, this time the northeastern arm did better. After a slow start, magma began to upwell in the deepening basin, and this eventually became the Reykjanes ridge. The basin now extended the North Atlantic Ocean. But the new spreading ridge never lined up with the older one further south: when the Labrador rift died, it cut the link between Reykjanes and the Iberian rift. Even nowadays you can see the large jump between the mid-Atlantic Ridge going north from the Azores and the Reykjanes Ridge going south from Iceland. This jump is a tribute to the adhesive strength of Eurasia, although in the end the resistance was futile.
The ocean floor itself can be used to date the final separation event. This method uses the magnetic reversals which have happened on not-very-regular times in the past. Every time the Earth’s magnetic field has a reversal, that reversal is frozen into the oceanic crust. As the liquid rock pours out of the spreading centre, during the cooling-down the iron takes on the direction of the magnetic field. As the lava solidifies, the iron is frozen in and is now unable to move: if in the future the magnetic poles reverse, the iron cannot respond. Thus, it maintains a memory of what the magnetic field was at the time the rock solidified. As the ocean widens, the oldest sea floor is at the edges and the youngest sea floor is at the spreading centre. Every reversal is engraved into the ocean floor at the distance from the spreading centre which corresponds to its age. The ocean floor shows stripes of normal and reversed magnetic field. In fact this striping was the final evidence that convinced even the most conservative scientist that continental drift was real. The stripes have been numbered: the one closest to the Rockall Plateau is called anomaly 24, or also ‘chron 24’. This is the oldest ocean floor known between Greenland and Rockall. Anomaly 25 is not seen: at that time, this part of the Atlantic Ocean had not yet begun to form.
So when was anomaly 24? From fossil and geological records elsewhere, we know it was near the start of the Eocene, with an age of 52-54 million year. The same rocks have also been found in East Greenland, where they have been dated using radioactive elements, and so this date is rather well established. This is the time the Reykjanes Ridge began its long career as mid-ocean spreading centre.
Further north, the Aegir ridge has about the same age as the Reykjanes ridge: it was creating ocean floor by 50 million years ago. But it never connected to Reykjanes and its formation was a separate event. Whilst the Reykjanes ridge slowly extended north into East Greenland, the Aegir ridge extended southward, staying east of Greenland. They passed each other at a distance: for a while there was a small range of latitudes where there were two parallel spreading centres. In between was what became the Jan Mayen plateau: it was split off from Greenland about 33 million years ago. Neither ridge completed the job. The growth of the Reykjanes ridge came to a halt at the latitude of Jan Mayen, and the Aegir ridge did so just north of the modern latitude of Iceland. Finally, 24 million years ago, the Aegir ridge went extinct and the Kolbeinsey ridge formed. This did connect to the Reykjanes, and now the North Atlantic Ocean was finally complete. Jan Mayen found itself afloat mid-Atlantic, a victim of the competing rifts, stranded on the European plate.
There is an intriguing possibility here. While the two rifts extended parallel to each other, there was a possible path around it, going north of the Reykjanes ridge, south over Jan Mayen, and east again south of the Aegir ridge. This was a bridge between Greenland and Europe, with some suggesting it was dry land as late as 30 million years ago. And because the Labrador sea had stopped extending north, there was also a path from Greenland via Baffin Island or Ellesmere to Canada, completing the bridge. A land bridge, or perhaps shallows with a chain of islands between America and Europe, could have allowed animals to spread between the two. There is some evidence for this but it is disputed. The earliest primates, 56 million years ago, spread almost instantaneously to the Americas, Europe, and Asia, showing easy land connections still existed at that time. The similarity in fauna shows that there was further interaction between America and Europe at a later time, perhaps indeed as late as 30 million years ago, but it is not clear how this occurred. Apart from this occasional bridge around 55 degrees north, a more northerly connection has also been considered, and of course the back route via the Bering Straits remained possible. If a land bridge still existed, it was intermittent and not easy to traverse.
From satellite images, the Atlantic Ocean seems straightforward. But it isn’t. The many false starts left debris and scars, some as obvious as Greenland, some as secret as Rockall. Everywhere, the Earth shows its history.
The Faroe problem
In the North Atlantic, the Faroe Islands are sitting on their own plateau, further north than Rockall. It is sometimes considered as part of the purported Icelandic hotspot track, the end of a ridge running from Greenland to the Faroe. In this interpretation, Faroe formed as material from Iceland was carried east by the moving oceanic plate (no model has ever predicted that the hot spot itself spent any time on the European plate east of Iceland). The map of the North Atlantic shows that this ridge is not continuous, but has a gap between the Faroe and Iceland. It would imply that there was a period of time where the hotspot was not active, and that Faroe came from an earlier version of Iceland. However, the map also suggests an easier explanation, namely that the Faroe is a fragment of Europe, similar to the Rockall Plateau.
In the traditional model, a mantle plume triggered the break up between Greenland and Europe, starting from the east coast of Greenland. But the way the Atlantic Ocean formed does not fully support this: the separation should have started at the point where the plume arrived, but in fact this region, close to where Iceland is now, was the last place where the continents separated. The best indications for a hot spot comes from the massive volcanic eruptions which left basalt around the North Atlantic. But the lateness of the split at this point is a problem for this model. One or more triple points formed south of Greenland: were these a separate hot spot, or perhaps the real one? What caused the Rockall trough? It is all far from clear.
This single rock sticking out of the Atlantic Ocean, unknown and isolated, standing up against waves higher than itself, is a memorial to the birth of an ocean. Few people will ever see it; tradition has it that more people have walked on the Moon than on Rockall (I don’t think this quite correct, though). It guards its memories well. But the rock shows that Europe and America still remain closer together than appears from a surface map, and that the separation took a long time to complete. Under the waves lies the record of a common heritage. The Scotts weren’t so scary after all.
Albert, May 2017
As I write this, my home city is still recovering from a suicide bombing targeting young people. The perpetrators were trying to push cultures apart by acts of evil. Geology shows how much countries have in common, and it teaches not separation but similarity. Even oceans do not divide forever. I would like to dedicate this post to the memory of the 22. We all stand together.