Over the past weeks, the Fagradalsfjall has settled into in almost predictable routine. There are regular cycles of eruptions and interruptions. During the interruption, the tremor goes quiet. Nothing is shaking or moving on the drum plots. Over several hours, there is a slow build-up of the tremor. Lava begins to return to the crater in this phase. It reaches a peak when the crater is full and is vigorously bubbling. The crater overflows and the lava moves down the slopes, towards Meradalir. After 10 to 16 hours, the tremor very suddenly stops. The drump lot goes flat, and within minutes the lava disappears. A quiet, lava-free period begins; the cone is dark and empty. It stays likes this for perhaps 10 hours, the build-up starts and the cycle repeats. The build-up is slow but the end is almost instantaneous. In one case we saw that the end was triggered by a small wall collapse. The complete cycle takes about 1.5 days, or at least has done for the past week or so.
There are many drone videos of the eruption. Here is one that shows the phase of high activity. The vigorous bubbling occurs on one side of the crater, and it has build up a higher slope on this side.
During one of those quiet periods, there was a surprise. This was perhaps not unexpected. As one commenter wrote “This volcano keeps doing something different every few hours. Wouldn’t shock me if it started belching out wildflowers for a bit.” To be precise, a big bang. It was brief, and it was after midnight while nothing much was going on. The flash was still noted by the ever-present army of watchers, who see with blinding sight. After the flash, a gap was seen in the rim of the crater. The next day, lava flowed through this gap but it was quickly repaired by the eruption.
Here is the recording of the flash.
The flash came from the same area of the cone where the bubbling occurred during the active phase. It was not at the location of the gap and it seems this gap may have formed before the explosion, not during it.
Explosions are not uncommon during a phase where an eruption tapers off. The lava retreats below a blockage, either rubble or a thin surface of solidified lava no longer broken up by the movement of bubbles. Gas can collect below the blockage, and once the pressure become too much, cause an explosion. We have seen this happen in Agung. One explosion in the night is not yet a pattern, but it is a warning. Do not assume the cone is safe even during its sleepy time.
Gentle into that good night
Is it possible that the eruption is waning? Recent activity episodes have not been as vigorous. The lava does not get as far in Meradalir and instead spends more time building up the shield on the old valley-with-no-name which now is almost a hill-with-no-name. Where there used to be a lava river running down the slope, now there are many braided channels. Carl noted that the low-frequency tremor is beginning to lag behind the high frequency, and that the build-up phase takes longer. It is appearing as an eruption in decline. The coastal road may yet be safe.
The latest report can be found at http://jardvis.hi.is/eldgos_i_fagradalsfjalli, with data taken on August 8. They find that the flow rate is a bit lower, 9.3 m3/s averaged over 12 days. The rate has fluctuated over the past month. The scientists write ‘There are strong indications that the flow was lower in the first half of July, 7-9 m3/s, but then came a peak that lasted for 8-10 days, where the flow could have reached 17 -18 m3/s on average.’ It is too early to call an end, but the eruption is having difficulty in keeping going. Was the explosion its Rage, rage against the dying of the light?
In spite of its difficulties, it is already an impressive eruption. The volume has reached 0.12 km3; the area that is covered by lava is now 4.4 km2. The area has increased little in recent weeks, because it is locked in by the walls of Meradalir. Will it manage to break out?
The last few days have had fairly clear weather, a notable difference to the dense fogs of July. This has allowed some useful satellite images, not entirely cloud-free but giving us a complete overview of the lava fields. This satellite image is from August 7, taken with the ESA Sentinel. It was during a vigorous lava flow into Meradalir; the red flows dominate the image. How far did these flows get?
It is easy to recognize the hot lava. In contrast, the cold lava is black, not the easiest colour for images, and not easy to distinguish from the surrounding burned vegetation. We can see that the flows went into Meradalir, but to see whether they expanded the flow field requires a better comparison.
The current full extent of the lava flow can be found at http://www.viewsoftheworld.net/?p=5783 together with detailed contour levels. Taking that map and overlaying it with the satellite image shows where the recent flows reached the walls. This did happen but only at a few places. The flow is pushing the edges in the northern most lobe and it may still be expanding there. The only other place was on the eastern side of the northern lobe but there it is pushing against a hill – it can go up but not out. It is possible there is some further lava movement underneath the surface which is hidden from our view. But mostly it seems that the lava likes to stay in the upper reaches of Meradalir.
In the western part of Meradalir, the elevation of the lava now reaches 150, in places 160, meters. There is a 20 meter drop in level around the narrow part of the valley. In the north-south-oriented valley to the east, the lava has reached an elevation of 130 meter. Northward it still has to inflate by 20 meters to escape: Meradalir there is enclosed by the 150 meter contour. To the east and south there are some escape routes at 140-145 meters elevation. The lava is still some ways short and it won’t escape the valley until the lava flows reach these borders again.
Instead, much of the lava seems to be building up the shield in front of the cone, and also thicken the slope into Meradalir and the upper valley. It is not as mobile as it used to, either because of lower flow rates or because of higher viscosity. It is getting old, but still going. Old age should burn and rave.
For enjoyment, here is another recent satellite image. This was taken by Planet Labs, also on Aug 7 but during the quiet phase. The high resolution image can be found at https://postimg.cc/w1JWNddj It shows more detail, including the shape of the cone
Finally, the 3d overview always adds to the information:
Rhymes and Reasons
So why is the output currently cyclic? Why is it variable? Is this an old eruption needing its daily naps?
Let’s start with volcano basics, VC 101. The magma flow is determined by a combination of factors (ok, VC 102). One is the magma itself: the overpressure it is subject to, and the buoyancy. The other is the conduit: the capacity of the conduit is set by its width and by the viscosity of the magma, perhaps reduced further in places by bottle necks.
Originally the flow was fed by the magma dike that was emplaced in February and March. The conduit connected this dike to the surface, and it limited the flow rate to 5-7 m3/s. The increase in the flow rate came in April when a new vent opened, with (apparently) a wider conduit. The dike probably contained around 0.1-0.5 km3, within range of what has erupted so far. Whether the eruption is purely living of this magma emplaced in the dike or that it has direct access to magma from deeper down is unclear. If it is the former, then we may now be in the winding-down phase where the pressure is decreasing and the viscosity may be increasing. If the latter, then it could continue as we are for longer. Of course, if it did stop the pressure in the dike could still force another opening elsewhere to give it another (shorter) lease of life. And volcanoes have been known to push out a block viscous magma to restart the faster flow. The problem with predicting the future is that there are too many possibilities. There is a butterfly effect for the weather – perhaps there is an earthworm effect for volcanoes.
Local GPS stations, e.g. Krysuvik and Svartsengi, indicate a small amount of deflation, with the stations moving a bit closer (1-2cm) to the eruption site since May, and since mid July also showing downward motion of some 2 cm. Svartengi is almost back to the elevation it had before March. Krysuvik still retains more than half of its elevation gain. This tells us that the deeper magma (perhaps 5 km or more) may have lost some volume but the shallow dike is still there. Of course it may have partly solidified.
The rapid fluctuations in flow rate suggest that the magma flow is no longer limited purely by the conduit but that maintaining pressure is becoming a problem. The most likely cause for the sudden interruptions is that something is quenching the degassing. Gas bubbles make the magma buoyant, and if they suddenly dissolve back into the liquid the magma column will collapse. The fact the cycle is fairly regular indicates it is not due to random effects, but it is running out of gas. Once it is running low, and small disturbance which adds weight to the column (rubble falling down) can stop the eruption. Now it takes time for gas to be replenished from below, carried up by fresh magma. Bubbles slowly reform, the column rises and the eruption restarts.
Where does this happen? The degassing is mainly in the lava pond inside the cone. If the pond has a size of 100 by 50 meters, it would need to be around 100 meter deep to provide the amount of lava erupted in one cycle. This is a rather rough number but let’s assume that in the quiescent phase, lava withdraws to that depth. That means that without the aid of the gas bubbles the magma column would reach to 100 meter below the rim of the cone. This is the height reached by the pressure in the dike or conduit, and it is not sufficient to drive an eruption. The cone has grown much too tall for the eruption. The eruption is at risk of ending itself by over-ambition. It may explain the parasitic vent, by the way, as the capillary action of the rubble of the sides of the cone can aid the rising of the lava.
The cone is currently not increasing in height. The region around the cone is growing, and is doing so rapidly. You can see this by staring at the above images taken on Aug 10 and Aug 12, or easier from this animation close-up from the Meradalir camera. The shield to the left is notably growing: you can see how the two hills immediate to left left of the cone are disappearing behind the shield. (One is now only barely visible.) Much of the lava is currently going into building the shield and not into the valley. But the cone itself is not changing (and neither is the region to the right). It indicates that the cone has reached its maximum height: the magma has insufficient pressure to go higher.
This suggests a way the eruption could re-invigorate: get rid of the cone. A new outlet at the base of the cone would change the state of play and could restart a phase of continuous, perhaps even vigorous (if not too viscous) flow. Could the cone collapse? So far it has avoided that – not even the explosion did much damage. But this volcano is unpredictable. Creating a new eruption site is harder as it has to break a new conduit through the solid rock. That will only happen if the eruption stops completely and the current conduit blocks. The pressure inside will build up, and eventually a new fissure may form. Or if the energy is gone, Fagradalsfjall may fall silent again, for 10,000 years or more.
The drum plot show the phase of low activity, the build-up and the high activity. You also see intermittent brief bursts of activity, lasting perhaps 20 seconds. These puzzled us for a while. They are seen at any phase, but only during day light – were they solar-induced? They are on a 2-hour schedule (notice the repeating colours), starting around 10am. We now think these are helicopters flying over the seismograph.
Early on in the eruption, it was found that the magma came from a depth of 14-16 km, which is almost exactly where the crust-mantle boundary is underneath Reykjanes. Crystals in the magma also showed that the magma had spend some additional time at a depth of 0.5-2 km, in the shallow dike. This was measured in the lava that was erupted during the first two days. The composition changed a bit in April, perhaps as the shallow dike became exhausted. The larger dike that was emplaced at 4-6 km depth during the rifting and shaking phase in February and March may have become the source of the magma, or it may be sourced directly from deeper down. Or both.
The bulk composition is shown in the figure earlier in the post. MgO started out at around 8.8% but increased after a few weeks to 9.6%. These values are high for Reykjanes: typical values are 7-8%. TiO2 is measured at around 1%, while for other Reykjanes lava fields it is around 1.5-2%. SiO2 has not been reported but for Reykjanes is normally around 49%. The magma clearly is not a common type for the region.
The bulk composition was last measured in June. It may not be easy to get access to fresh lava at the moment! (Note that this is NOT a call for help.)
Abundances of other elements can help in tracing the origin. These were measured in the first few days of the eruption, and to my knowledge not since. The most important result is from lead. This element has four common isotopes, which form in part from radioactive decay. That is a very slow process, but the mantle is a very slow beast. Different convection currents n the mantle can end up with material of different ages and therefore somewhat different isotope ratios. The different isotopes are chemically identical, so that the ratio of the original material that provided the melt is kept in the magma.
Here is the result. The lead isotopes follow the sequence of the other lavas, but with an interesting detail. The blue squares are the Krysuvik and Svartsengi lavas from 800 years ago. The cluster of open squares to the top right are other lavas from this period, the Reykjanes fires. The open squares on the bottom left are older lavas from the end of the ice age when there was a spike in activity driven by decompression melting. The top right is called ‘enriched’, for obvious reasons, and the bottom left ‘depleted’. The current eruption is less enriched than any measured lava of the previous period, but it is not as depleted as that of the post-ice-age period.
This sequence is thought to result from mixing of different magmas. How many different magmas there are is disputed, but it seems at least three are needed: the decompression melt at the ice age (long since run out), an enriched magma and a mildly depleted magma. The current eruption could be an almost pure example of the latter.
Notice the red squares? They are not on the peninsula but are from two recent lava flows (20th century) on the Reykjanes Ridge, in the Atlantic Ocean beyond Iceland. They are similar to those of Fagradalsfjall.
This is worth exploring further. The lead isotopes have been measured below the sea along much of the north Atlantic rift – the MAR. This has shown a marked variation. While on-land a large range is seen, the MAR shows distinct regions can be seen, each with their own value. In particular, the Reykjanes Ridge shows a particular range with little scatter, with differs from that seen elsewhere. The Fagradalsfjall ratios are identical to those of the Reykjanes Ridge.
There are other elements in the lava. The rare earth elements in particular show notably low enrichment in the current eruption. On the Reykjanes Peninsula, high enrichment of rare earth elements is seen mainly in the east and the lowest level of enrichment are on the western tip. The off-shore lava is even less enriched. The current eruption is again close to the off-shore lavas.
As mentioned, recent Reykjanes magmas (i.e. the fires 800-1200 years ago) are thought to be formed by mixing two (or more) components: an enriched one and a less-enriched one. Fagradalsfjall is an example of the latter, and it seems to relate to the spreading ridge, which can cause melt at 20-50 km depth. This ridge magma would have collected at the crust-mantle boundary, 16 km deep.
There are other sources of magma in Iceland. The plume underneath Vatnajokull brings up a deeper melt, and this may spread to the peninsula. Magma pockets from previous eruptions may still be stored underneath the Reykjanes Peninsula, slowing changing composition as the magma evolves. At one time in the past there was also decompression melt, affecting shallower regions in the mantle and perhaps the upper crust. This last component is now gone: magma left behind from this will have solidified. The mixing of the other components cause the spread in lava properties on the peninsula.
Fagradalsfjall is in a funny place for an eruption. There are four volcanic centres on the Reykjanes Peninsula, but this eruption is in between two of them. There were eruptions north of here after the ice age, but Fagradalsfjall itself may not have seen an eruption since 35,000 years. The magma that had collected at the top of the mantle came up to feed the new eruption. But it did not find other magma pockets or stores: This was not a volcanic area. It gave us almost pure (primitive, as it is called, to the horror of any archeologist) mantle magma. What we see here is in effect a mid-oceanic rift eruption – on-land. It is not identical to one: the magma formed under an additional few kilometers of crust which may have provided extra weight and insulation, increasing both the temperature and the pressure. In particular, the high MgO and low TiO2 of Fagradalsfjall are not typical for the Reykjanes Ridge, or for other mid-oceanic spreading ridges. But it is close, and the lead isotopes and rare earth elements suggest the material shares its source with the MAR.
We know little about mid-oceanic-ridge eruptions. The most recent case on the Reyjanes Ridge was in 1970, at Eldeyjarbodi, 55–60 km off-shore, where lava was found to have formed on the sea floor. The eruption itself was not observed. This same region also erupted in March 1830, causing a large plume that was seen from Reykjavik. That eruption lasted with intermittent activity for a year. Perhaps it gives us an idea what to expect from Fagradalsfjall. Or perhaps not. But it is an interesting idea that we are seeing a real-life MAR eruption, the first ever tourist-friendly one. There is much to learn.
Albert, August 2021