Iceland does it rather well. Over the 1100 years or so since the Viking settlement, it has erupted some 63 km3 of lava, the large majority of it basaltic. A similar volume of tephra was also produced. The East Volcanic Zone is the prime suspect for any eruption, and with good reason: it is responsible for 51 km3 of this lava. All four regular eruptors are here: Katla, Hekla, Grimsvotn and Bardarbunga. But this means that 12 km3 of lava is missing, or at least went elsewhere. Where is it?
Of the missing amount, 8 km3 is hiding in clear sight to the west of Langjökull, in the Western Volcanic Zone. It forms the Halmundarhraun lava field, and was erupted between 800 and 1000 AD, probably a little before Eldgja. A further 1 km3 is in the North Volcanic Zone; a region which is much better at tephra than lava.
That means we are still short of 3 km3 of lava, a not inconsiderable amount. It is enough to bury Reykjavik (a very spread-out city) under 10 meters of lava. And it is in fact found not far from Reykjavik. The missing lava is located on the Reykjanes Peninsula, in between Reykjavik and its airport, Keflavik. No other nation in the world has build its capital so close to so much young lava.
So why is this volcanic activity on Reykjanes Peninsula so little known? It turns out that this region goes for larger and occasional, rather than smaller and often. It has been a long time since the last event here. But things are stirring and after 8 centuries, this sleeping beauty may have met her prince. She seems to be waking up.
Iceland in faults
Iceland is riddled with faults. In this map, black lines indicate volcanic zones, and red are non-volcanic faults. The blue lines indicate volcanic flank zones, with some minor rifting, where stratovolcanoes have developed. The black arrows show the spreading direction.
The main volcanic zones are pure spreading centres, excellent for generating magma. The red faults are transform faults: these can provide pathways for some magma to reach the surface but they do not generate magma themselves. The pink lines are combinations of transform and spreading faults, also called trans-tensional faults. They have both sideways (lateral) motion and spreading (extension).
The Reykjanes Peninsula contains the only transtensional fault found on land in Iceland. At the tip of the peninsula, the fault bends to join the Reykjanes Ridge, where it becomes a spreading centre. On the peninsula, it is a bit of both.
Comparing the map to the volcanic eruptions shows that the majority of Iceland’s (on-land) lava is generated by the two main spreading centres, the EVZ and the WVZ. The Reykjanes Peninsula is the only place where the eruptions occur in a fault with a transform component. There is still some spreading: the spreading rate here is about 8 mm/yr, compared to 20 mm/yr for the EVZ (it varies a bit along this fault). The Reykjanes Peninsula therefore has 30% of the spreading of the EVZ, but has only 6% of its lava production. Allowing for the fact that the Peninsula is only half as long as the EVZ, it produces 2-3 times less lava per kilometer per centimeter of spreading. This shows the effect of the mantle heat underneath the Vatnajokul which increases the magma production there. Reykjanes lacks such a pronounced hot spot. On the other hand, it is much closer to the actual MAR where the Atlantic Ocean forms its new crust. It may perhaps borrow some magma from there.
But look closer. The eruptions may belong to the Reykjanes Fault System, they don’t use this fault for their progression. On the EVZ, eruptions form long fissures along the spreading fault. But on the peninsula the volcanic activity runs at an angle to the fault. They form a series of NE-SW segments, while the Reykjanes Fault System runs E-W. In fact these segments are approximately along the main direction seen in the EVZ and in the Reykjanes Ridge. It is the direction of the spreading rift. Dikes develop much more easily in this direction, as the crust is already being pulled apart. Magma is a lazy substance, and it will always follow the path of least resistance.
On the peninsula, eruptions may begin on the Reykjanes fault, but when they spread they take their directions from the spreading, and ignore the transform component.
The easternmost segment, Hengill, is considered part of the WVZ. The next ones are Brennisteinsfjall and Krisuvik, and these are the most productive of the segments. Fragadalsfjall seems to be a new development. Svartsengi and Reykjanes make up the end of the peninsula. There is a further system off-shore, invisible below the waves.
The most famous eruption on the peninsula dates from 1226 and occurred on the Reykjanes system, or a few kilometers out to sea. It was part of a longer sequence of eruptions which left the tip of the peninsula a volcanic waste land. Around this time, Icelandic sagas mention ‘fires in the sea’ at the Reykjanes Peninsula for 1223 and again 1225-1227 (which may or may not have been one continuous eruption). Within this series there was one particularly violent eruption: after this explosion in 1226, the southwesterly wind spread 0.1 km3 of tephra across the area, with a thickness of almost 1 cm as far as Reykjavik. The tephra layer from this explosion is found across the peninsula and is the most pronounced layer here over the past 5000 years.The sagas mention that it caused ‘darkness in the middle of the day’. The winter of 1226/1227 was called a ‘sand winter’. This word is normally used for the year after an eruption, and indicates the effect of the volcanic ash rather than the ash fall itself. It causes mortality in the farm animals, and famine. Fluorine may be to blame, but there are other possibilities. Ash and tephra are unhealthy for cattle when digested with the gras they cover. Sulfur emissions may also have killed vegetation, leaving insufficient fodder to last the winter. Farming and volcanoes make for an uneasy partnership.
It has been suggested that the Karlinn rockstack comes from this eruption, but this stack seems a bit too close to the shore and the explosion probably occurred a bit further out. There is an underwater ridge some 50 meters deep which seems suitable.
The soil layers show that a few years before the 1226 eruption there had been an effusive eruption on land, which became explosive when it reached the sea. The 1226 ash fell on top of the lava sand from this earlier eruption. This was a double hit: the vegetation would not come back for a long time afterwards.
These were not isolated eruptions. The peninsula contains a number of lava flows which were erupted within a 400 year period, between about 940 and 1340. The series began in the east, in the Brennisteinsfjöll volcanic system, with a series of eruptions, a few decades apart and ending before 1200. After 1150 eruptions began further west, in Krisuvik, also lasting until about 1200. Now the Svartsengi and Reykjanes systems joined in with a series of eruptions, from 1211 to 1240. The 1226 eruption was part of this group. Then it all ended, and the area has remained quiet ever since.
This behaviour of progressive eruptions turned out to be typical. There had been a similar series of eruption 1000 years earlier – and the same another 1000 years before that, and again before that one. In this region, eruptions move from the east to the west over several hundred years, terminate, followed by 700 years of volcanic sleep before it restarts.
(Hengill is not part of this sequence. It erupts rather infrequently: the last time was 2000 years ago. Its eruptions are connected to the WVZ, not the Reykjanes Fault System.)
The sequence may have gone on ever since the ice melted, but the oldest series have not left a record. Their flows were overwritten by later lava. It seems that the earliest eruptions, shortly after the ice went, were larger than they are now. As the land rebounded from the ice age 14000 yars ago, it may have generated more magma. Large shields formed: the largest one covers over 20 km2. Over time, the eruptions became smaller. But they did not fade quietly into the volcanic night: The thickest tephra layer dates from 6100 BP, up to 1.5 meters thick.
Why this progressive, sequential pattern? That is guesswork. My personal guess is as follows. The volcanic systems have formed a series of echelon (parallel) faults, which are being pulled apart by the rifting component. Over 500 years the spread adds up to some 4 meters, meaning that each fault has to counter about 1 meter of movement. That seems to be the maximum stress here that can be accommodated. When the stress exceeds this, the easternmost of these faults snaps open. Several fissure eruptions follow. Now the faults give way one after another, with Reykjanes as the last one.
Towards the end of the series, eruptions can also happen beyond the end of the peninsula where the volcanic systems extend for some 15 kilometers beyond the land. It ends before Eldey: Eldey is part of another system, offset from the Reykjanes Peninsula volcanoes, with its own eruptions. Here the bend in the fault gives a larger spreading component.
Magma in motion
As the rift opens, magma begins to come up. This seems confined to a fairly narrow band, perhaps 4-5 kilometers wide. Initially, it appears that a sill forms, pushing up the land. Eventually a dike forms, going either upwards or sideways. Eruptions can happen along the segments, showing magma traveling sideways. But they tend to happen on land, and not too far from the Reykjanes fault. However, the dikes can continue for much further. They have been traced 50 km to the northeast, well beyond Reykjavik and even beyond Hvalfjordur. But the dikes do not come to surface beyond a few kilometers from the Reykjanes Fault System. If magma takes this route, it stays underground.
But even underground magma can have an effect on the surface. The peninsula has permeable rock, where sea water can get in. Once it hits the magma sills, even after 700 years, it gets heated and moves up. These are the geothermal centres: there are four on the peninsula, each located along a segment close to the fault. The Blue Lagoon is one of these, although it should be noted that it is completely artificial: the lake contains the waste water of the local geothermal plant. And the distant dikes still generate some geothermal heat as well.
We are now some 700 years into the years of quiet. Over the past decades there have been episodes of unrest, perhaps starting with an inflation event at Hengil in the 1990’s. There have been several earthquakes swarms on the peninsula. The current inflation shows that magma is now accumulating, something that had not been seen here before. Is this another step in the re-activation? Will it slow down and stop, as IMO considers most likely? Or will the magma prince find a weakness, wake the sleeping beauty and come to the surface? Or will a dike develop taking the magma southwest? Time will tell.
But in any case, these are only the initial stirrings. If history repeats itself, the real eruptions won’t begin here. It will be further east, south of Reykjavik, with several eruptions occuring decades apart. It will take two to three centuries before the tip of the peninsula erupts. But will history really be condemned to repeat itself? Predictability is not normally a strong point of volcanoes. They may still surprise us.