Iceland’s eruptions since 1900

In 1900, Iceland was a very different country. It wasn’t a country, for one thing. Although Iceland had been granted self-rule, it was still a part of Denmark and the local leaders were largely Danish-educated. There was still a Danish governor. The position (not the governor) was terminated in 1904. On Dec 1, 1918 the nation officially became independent, and the union with Denmark ended in 1944. In 1900, the country was depopulated: the climate had worsened, and many people had emigrated to Canada. Reykjavik was little more than a small fishing village, and fewer than 80,000 people lived in all of Iceland. (Nowadays there are some 350,000 Icelanders.) It always was a hard place to live. After the settlement era, around the year 1000, perhaps 30,000 people lived in Iceland. The numbers grew and declined with the climate. Even as late as 1700 there were only some 50,000 Icelanders. Laki reduced the population by 25%, in part by mortality and in part by emigration. This nation of fishermen (and women) always lived at the edge of its carrying capacity. Once sailing boats took over from rowing boats, more people could live from fishing and the population could increase, but in Iceland that was a remarkably late change.

Reykjavik, around 1900

And always there were volcanoes and there were eruptions. Demographics and politics has little influence on them. With so few people to watch, minor eruptions in isolated areas (central Iceland and Vatnajokul, especially) may have erupted in vain, and were missed. What was there to see?

There was a lot to see, and few to see it. Iceland is full of volcanoes, and many are considered active. Some (mainly Grimsvotn) erupt every decade, some (e.g. Torfajokull) only once every millennium. But many volcanoes are in regions where no one lives. That makes sense: frequent tephra falls, lava flows and glacial floods do not go well with keeping sheep. The land has changed; areas that were fertile 1000 years ago are now forbidding wastelands. This has to do with the changing glaciers, and the jokulhaups that changed in tandem. Perhaps a future post. The desolation near active volcanoes means that at the start of this post, the beginning of the last century, eruptions were often poorly observed – or perhaps missed entirely. An unusual flood in a river might be the only sign. In the 1920’s some seismographs appeared, and some of what had been hidden became visible. From 1934 Vatnajokull was frequently visited by explorers, and shortly afterwards overflights began. Interestingly, that dramatic improvement in monitoring coincided with a lull in volcanic activity!

But let’s go through the century (and more) of volcanic activity, from the hesitant beginning to the open-science era of VC.

The volcanic century started with activity at Vatnajokul, setting a pattern that was to continue for the next 120 years. This was at a time that scientific exploration of the area was only just taking off. A Danish expedition attempted to map the entire glacier in the period 1902-1904, but they somehow missed all the excitement, and the eruptions of that time are known from distant reports, not from Danish feet on the ground. There were two eruptions during their time, and both seem strangely unusual. Over time eruptions became better observed and exceptions became rarer although not absent. There were fires (fissures), explosions, and in some cases both. Many eruptions were repeat events, others were unique. Some lasted less than an hour, others a decade. And every eruption was a surprise. Let’s look.

1902 Bardarbunga

In December, fire was seen towards Vatnajkull, by people in northern Iceland. Simultaneously, a jökulhlaup occurred in two rivers: the Jökulsá á Fjöllum and Skjálfandafljót. They were strong enough to damage ferries. This requires that the eruption was from a rift on the north side of Bardarbunga, as these rivers are only fed together from there. A little ash was found later in the wool of sheep from southern Iceland, but it is not known whether that is related. This is one of only two confirmed/likely eruptions of Bardarbunga during the entire 20th century.

1903 Grimsvotn

Six months later an eruption occurred on the Grimsvotn system, during a jökulhlaup. We don’t know for certain where the eruption occurred. The eruption was described by Pastor Magnús Bjarnason, dated 11 June 1903:

[…]it was nonetheless one of the largest and most frightful of jökulhlaups as it occurred at the same time as an eruption in Skeiðarárjökull which often happens, to a greater or lesser extent, when there is an outburst in Skeiðará river. Flames arose north and west of the so-called Grænafjall mountain, which lies to the north of Eystrafjall, on Thursday 28 May, upon which the floodwaters increased as the glacier melted, searching for channels with such power that they burst open the front of the outlet glacier and surged forwards with blocks of ice that, according to the post bearer, were 60 feet high, and over a wide area there were others on top of them another 25– 35 feet thick.

The Eystrafjall is a small peak on the eastern edge of the Djupa area, due south of Lake Graenelon (a glacial lake that now no longer exists due to the melt of the glacier that dammed it). The eruption is often attributed to Thordarhyrna, a volcano that has not erupted since. But was it? Bjarnason writes that it was in Skeiðarárjökull, and Thordarhyrna is just to the west of that glacier. If we go due north of Eystrafjall, we find Haabunga and Grimsvotn, at the top of Skeiðarárjökull. Thordarhyrna is well to the west.

If we consider that jökulhlaups in the Skeiðará river come from Grimsvotn, that Grimsvotn is due north of Eystrafjall, that Grimsvotn indeed often erupts during a jökulhlaup, and that Bjarnason describes it as a normal eruption (‘often’), it seems unlikely that the eruption came from Thordarhyrna. This appears to have been a normal Grimsvotn eruption, rather than a rare event.

The eruption started explosively, with tephra fall northeast. A ship 30 km off the coast report ash fall. The eruption continued for 5 months, until October. That is unusual for Grimsvotn: its eruptions typically last for a week. Together with Bjarnason reporting ‘north and west’, this may suggest that after the summit explosion, a small rift opened up towards the southwest and that this continued the eruption. But the precise location is unclear, and it is unlikely that the rift extended as far as Thordarhyrna.

A contemporaneous document states that the 1902 and 1903 eruptions came from different mountains, Bardarbunga and Grimsvotn, although it has the order the other way around, something that is incompatible with the location of the jökulhlaups.

As an aside, the first reported name for Vatnajokull is Grímsvatnajökull (The Glacier of Grimur’s Lakes). It derived its name from Grímsvötn (The Lakes of Grimur). However, the name Grimsvotn was also used for Lake Graenelon, and it is not clear which of the lakes was the original name bearer.

1910 Bardarbunga / Hamarinn

After this long diversion, we get to the next contentious eruption. On June 18, ashfalls were reported in southern Iceland. Further ashfalls happened on 15 August and 21 September, and during the last two weeks of October. A flood came down the Skafta river. The location of this eruption is uncertain. A painting with an eruption column aligns it with the East Skafta Cauldron on the Loki ridge, east of Hamarinn. This is a geothermal area; it is not known to be eruptive, but appears to have erupted once since. The long duration suggests a fissure eruption; the most likely origin is Bardarbunga. The eruption was largely below the glacier, but broke the ice a few times.

1913 Lambafell

This eruption started on 25 April 1913. Initially a 5 km long fissure opened at Mundafell, 5 km east of Hekla. This was after several hours of earthquake activity which was felt as far as Reykjavik. This fissure lasted a day. A second fissure, also 5 km long, opened up at Lambafit, 15 km further northeast along the same direction as the previous fissure. This fissure was weaker but lasted much longer, until 18 May. The eruption was effusive and basaltic. Although it was near Hekla, this fissure is not part of the Hekla system. The first fissure produced a lava field of 3.8 km2, while the second fissure covered 6.3km2. The total erupted lava volume is estimated at 0.02 km3. An eruption had happened in 1878 on the same rift but in between the two fissures.

Katla 1918

Now it was time for a serious eruption. As if to celebrate Iceland’s full independence which began on Dec 1, Katla broke the ice. The eruption started on October 12; it lasted for 3 weeks. The eruption started in the southeastern side of the caldera, made short shrift of the 400-m thick icecap, went airborne, and deposited ash widely across Iceland. Even Akureyri on the north coast received an ash layer thick enough for footprints.The strongest explosions came early in the eruption, but minor explosions continued for weeks. Lahars came down the river valleys: the estimated 8 km3 of flood water carried house-sized glacier blocks with it. Of the 0.8 km3 of tephra, half was washed down with the flood and half was explosively ejected. (Some estimates of the volume are higher.) Even the ice of Vatnajokul was blackened by the ash, and was difficult to cross for a year after: sledges and skis do not cope with tephra. The south coast was reshaped by the flood sediment which extended the coast line by 2 km. This was a high VEI-4 eruption as is typical for major events in Iceland, VEI-5 being fairly rare here.

Katla 1918

Katla would not erupt again in the 20th century, apart from very minor hydrothermal events which stayed well below the ice.

1921-1929 Askja

The minor rift activity in Vatnajokull now moved north, to Askja. A large eruption in 1875 had left a crater lake, called Öskjuvatn. Over the next 40 years it slowly grew as the crater continued to collapse. Now the area came back to life. Solfatara activity had been noted a few years before and this may have been a precursor of the first eruption in March 1921.

Four small lava flows formed in 1921, 1922, and 1923. They are called Bátshraun, Mývetningahraun, Kvíslahraun, Suðurbotnahrau; all flowed into the lake. In 1926, a lava flow formed a small island in the southern part of the lake; the island is now called Askjur. A much larger lava flow (0.22 km3 DRE) came from a 6-km long fissure on the southern side of the lake. The lava flow, called Thorvaldshraun (not the only lava flow in Iceland with this name!), it formed between 1924 and 1929, perhaps in multiple episodes. The lava flowed eastward from the rift, and is on average 5 meter thick. The earlier eruption were on the ring fault, but the NNE-SSW orientation of the new fissure instead followed the line of the rift zone along which Askja lies. The earlier flows had resolved the stress left from the crater formation, and now the large scale stress regime of the northern rift had taken over.

1922 Grimsvotn

After almost two decades, Grimsvotn was ready to try again. This was a violent eruption which distributed a thin layer of ash across northern and eastern Iceland. The ash fell for three days, from 5 to 7 Oct; the large majority of it was on Oct 6. The eruption happened at the end of a jökulhlaup. This is a fairly common event: the emptying of the lake can trigger an eruption that was already building. The jökulhlaup produced some 7 km3 of water. Whereas the previous Grimsvotn eruption (1903) was a relatively rare rift, this was a more typical summit eruption, brief, explosive, but with little or no lava involved.

Ashfalls of the 1903, 1918, 1922 and 1934 eruptions. And ashfall of 300 g/m2 is enough to leave footprints. Source: Larsen et al. 2014, Last millennium dispersal of air-fall tephra and ocean-rafted pumice towards the north Icelandic shelf and the Nordic seas.

1927 Esjufjöll (uncertain)

This eruption was not observed, but was inferred: In the beginning of September 1927, a jökulhlaup came down the Jökulsá á Breiðamerkursandi, towards the east coast, brining a smell of sulfur. This is the only jökulhlaup known from this river. It is however insufficient to define as a definite eruption, as solfataric sulfur can also accumulate in a glacial lake. The event is notable because one person died from the flood. If this was an eruption, this would be one of only two volcano casualties during the century.

1933 Grimsvotn

Very little is published on this event. It is reported to be north of Grimsvotn, perhaps Gjalp. No earthquakes were reported.

1934 Grimsvotn

This was the first well-studied eruption in Vatnajokull. It started, as often for Grimsvotn, with a jokulhaup. The Skeidará river began to slowly rise on March 22. The full jokulhaup came a week later, and peaked on March 31. The flow reached levels of 40,000 m3 per second, carrying ice blocks as big as houses and covering the entire flood plain. The flood abated on the next day. In the mean time, the eruption began on in the evening of March 30. On the next day the eruption cloud reached 13 km in height, and ashes spread east and northeast. The eruption quickly subsided, and fully ended in the middle of April. Subsequent expeditions found scoria and pumice around the eruption site, and two new craters in the southwestern part of the caldera. The easternmost crater was 500 m in size. The ice below the rim of the caldera had sunk by 200 meters since two Swedish students had rediscovered the caldera in 1919. In addition to the tephra, the eruption produces hyaloclastites below the ice, with a volume of 0.03 km3.

1938 Grimsvotn

This started out as a repeat event of 1934. In April, a jokulhaup of similar size to that of 1934 covered the entire floodplain, stretching 50 km wide. The eruption that followed remained subglacial and was not noted above the ice cap. The reason for this was found later: the eruption had taken place not underneath the fragile ice around the lake, but 10 km north, near Gjalp. Here, a 2km wide cauldron had formed, 150 meters deep in the centre. A rift-like depression in the ice pointed directly at Grimsvotn. The ice rift was 1 km wide and more than 100 meter deep. The ridge below has a volume of 0.4 km3, and this is assumed to be the total volume of the eruption. That puts it among the larger eruptions in Iceland of the 20th century. The eruption likely had started earlier, around the middle of March. Melt water which forms in the region around Gjalp flows into the caldera lake and this can trigger a jokulhaup. Caldera eruptions such as 1934, in contrast, tend to come after a jokulhaup.

After this eruption, Grimsvotn did not have another certified eruption for almost 50 years. In fact all of Vatnajokull went quiet.

Hekla 1947-1948

The union with Denmark ended on 17 June 1944, and now Iceland stood on its own (leaving the minor issue of the British occupation aside). The first eruption of the nation was a memorable one. Hekla had been quiet for some time (the repose time was 101 years). Now it came back with a vengeance, with its largest eruption in the 20th century.

The eruption began on 29 March. As usual with Hekla, it started highly explosive. The initial explosion lasted an hour, and reached a height of 30 km. Intermittent explosions continued for several months. Effusive lava flows followed. In total 0.74 km3 of lava was erupted; the flows covered 40 km2, on either side of the summit rift. The eruption lasted for 13 months, well into 1948. But the eruption did not end when it ended. CO2 continued to seep from the ground for several months, and collected on the west side of the mountain where both birds and sheep were killed. These were the first recorded volcanic casualties in Iceland during the 20th century, other than the possible 1927 eruption.

Hekla eruptions in the 20th century

1961 Askja

The eruption began on Oct 12 with intense solfatara activity, before progressing to a significant effusive eruption on Oct 26. It occurred on a fissure north of the caldera, and the lava flowed east.

The eruption ended in the first half of December (the precise date is not known. The tholeiite lava had a volume of 0.1 km3; the flow extended some 11 km. Early lava was a’a; later it became pahoehoe. Thorarinsson wrote an impression of the early phase of the eruption, obtained during an overflight: “The glowing river that belched forth through Oskjuop was a spectacle beyond description. The eruption column rose to 6000 m and its upper parts were lit by the full moon and flaming northern lights. A strong wind blew from the north and carried the column southwards. Now and then four fountains of brightly glowing lava were visible through the clouds; the easternmost was highest and its height was estimated at 500 m.”

1963 – 1967 Surtsey

Now 15 years of volcanic silence followed, apart from one suspected hydrothermal event under the icecap of Katla, in 1955. The silence broke in an unexpected place: 140 meters under the sea, off the southern coast. The activity started in mid November, and within a day broke the surface. The island of Surtsey began to form. The eruption continued intermittently until June 1967, along a 6-km which covers 14 km2. Two further islands formed during the eruption, but unlike Surtsey they survived only for weeks. The eruption produced about 0.7 km3 DRE. When the eruption ended, Surtsey contained two 140 m high tuff cones and a small pahoehoe lava field. The island was immediately proclaimed a nature reserve.

Hekla 1970

On May 5, at 20:58 an earthquake swarm hit Hekla. That was as much warning as it was going to give: the explosion occurred 25 minutes later. It lasted several hours; the eruption column reached 16 km height. This was quite a damaging eruption, not so much because the explosion itself, but because of the amount of fluorine which was deposited over the grazing ground to the northwest. Several fissures opened and closed, as is usual with Hekla; the last fissure closed on July 5. The lava covered an area of 18 km2. The volume is reported as 0.21 km3

1973 Heimaey

Icelandic eruptions tend to be fairly safe. This is because they tend to be fairly predictable, not in terms of when they will erupt, but where. There are some 45 volcanoes-at-risk, and in addition the ‘dead zone’ and the northern volcanic zone can produce long-distance rifts. Icelandic people are by-and-large sensible, and no one lives there. But sometimes even Icelanders can be surprised. Heimaey was the biggest surprise of the century. It erupted suddenly, unexpectedly, at night, and almost in the middle of a town. The successful evacuation was greatly aided by luck. A 1-km shift, or just different weather could have caused disaster with perhaps significant casualties. The eruption lasted from 23 Jan to 28 June. It build a new cone (Eldfell), buried part of a town, and activated a 3.5 km rift (10 km if a small submarine event is added). There were two casualties: one horse, and one person who died indoors from the poisonous gasses.

1975 -1984 Krafla

After this excitement came a prototypical Icelandic rift eruption. The Krafla volcano is a low 25-km wide shield, with an oversized 9-km caldera. On Dec 20 it announced its presence with an earthquake followed by a small eruption within the caldera. The eruption was quickly quenched when a dike formed and diverted the magma supply. The dike stalled before it could reach the surface. Three weeks later, a M6.5 earthquake occurred near the Grimsey rift, perhaps aided by the filling of the rift. Now the stage was set. Over the next decade, 20 such events happened, each with magma intruding into the central chamber, followed by the formation of a dike either north or south, extending some 10 km. Initially, lava was rare. There were two small fissure flows in 1977, whilst the next flows did not appear until 1980. Each flow lasted no more than a day. But from mid-1980 the flows grew in size, became longer and lasted up to a week. There were two further eruptions in 1981 and one in 1984. Afterwards there was some activity around Grimsey for another 5 years, but no further eruptions. The longest eruption of the era had come to an end.

Maybe an overly relaxed attitude to eruptions?

The event was almost a carbon copy of the 1725-1729 eruption. The two lava fields are closely overlapping. The 1975-1984 eruption eventually covered and area of 36 km2, 1 km2 more than the flows 250 years earlier. The eruption produced about 0.35 km3 of lava. This is very little, compared to the duration of the activity. This is in part because of a low magma supply in this region, and in part because much of the intrusion went into filling the rift. A similar effect had been seen in the 1875 Askja eruption, where the eventual volume of the crater far exceeds at of the erupted lava. Much of Iceland’s growth occurs underground.

1980 Hekla

The 1970 eruption had given 25 minutes warning. But in 1980, Aug 17, there was only 23 minutes between the onset of earthquakes at 13:04 and the onset of the eruption at 13:27. The eruption lasted only 3 days, unusually short for Hekla, but it had the usual mixture of an explosive start lasting several hours, and an effusive follow-on. The eruption cloud peaked at 15 km. The volume of the eruption was 0.12 km3.

1981 Hekla

On April 8 Hekla resumed its previous eruption, albeit very restrained. The eruption was not observed (the main event was between 2 and 5 am), but the earthquake activity was noted and there was some tephra fall. The eruption cloud did not reach much above 5 km. Small lava flows came from the summit rift, amounting to 0.05 km3. Everything ended on April 16. This was such an unusual event (for Hekla) that it is often considered as a late activity of the 1980 eruption.

1983 Grimsvotn

After almost 50 years of withdrawal, Vatnajokull came back in action. But it did so somewhat hesitantly. A small fissure eruption started on May 28, with along the southern crater wall of Grimsvotn. It was preceded by 6 months of enhanced earthquake activity: Hekla can learn from this! The eruption lasted only a few days. It produced hyaloclastites with a volume of around 0.01 km3.

1984 Grimsvotn

The 1983 event was repeated a year later. After a few months of earthquake activity, a brief eruption occurred in August. It lasted less than an hour and was unobserved apart for some tremor, but the depressions left in the ice showed that this was indeed a mini-eruption. One wonders whether the 1983/84 Grimsvotn hick-ups were related to Krafla 1975 and Askja 1961, all small events on or near the rift zone. Much stronger activity on the rift was to come in 2014.

1991 Hekla

After the embarrassment of 1981, Hekla tried again a decade later. It caught observers by surprise. There was a small earthquake swarm at 16:30, but the eruption itself (17:05) was first noticed by people on a nearby farm. This initial explosive event lasted 2 hours, and gave rise to an eruption plume 12 km high. The effusive phase that followed focussed on a short rift where a cinder cone build up. The lava field extended to 23 km2. The volume of the eruption was around 0.2 km3. The eruption ended on 11 March.

1996 Grimsvotn/Gjálp

On 30 September Grimsvotn rejoined the world of the true volcanoes. It erupted on its northern rift, at Gjalp. The eruption took more than a day to break through the ice, 600 meter thick. By the time it had managed to do so the strongest eruption as over, and the cloud nowreached only a few km high. The plume came up through a 100 m high ice chimney. Tephra topped at 500 metres. But underneath the ice this was a substantial eruption, which led to the formation of 6 km long and 500 m high subglacial hyaloclastite ridge.

The unusual aspect of this eruption was that it started at Bardarbunga. On Sept 29 it suffered an M5.7 earthquake, followed by a swarm which then migrated to Gjalp. Was this in fact a Bardarbunga eruption? The composition of the ejecta did not indicate this, but neither were they that typical for Grimsvotn. The 4-km long fissure, though, pointed at Grimsvotn and left little doubt about the culprit. A later re-analysis of the data suggested that there may have been a minor eruption in Bardarbunga before the M5.7 earthquake, and that afterwards the earthquake swarm came from a collapse on the crater rim. This reduced the pressure on Gjalp, and allowed the rift there to open.

As is common, the Gjalp eruption caused the Grimsvotn lake to rise, break the ice, and escape as a major jokulhaup. A water volume of 3.5 km3 came down the river. This is much less than early in century. Possibly the thinning of the ice cap in recent decades doesn’t allow the lake to rise as far.

This was Grimsvotn’s 4th rift eruption outside of the caldera in the 20th century. The commonly quoted number is that 6 out of 7 of Grimsvotn’s eruptions take place in the central crater. This was not the case in the 20th century, up to this point. The 1996 eruption produced around 0.4 km3 DRE but little tephra.

1996 Bardarbunga

On November 6, there apparently was a brief eruption in Bardarbunga. It was picked up by seismographs, and lasted less than half an hour. The eruption formed two ice cauldrons in the most southeasterly part of the Bárðarbunga caldera. The eruption was a day after a jokulhaup from Grimsvotn, but that may be a coincidence.

1998 Grimsvotn

Now Grimsvotn came back into its stride with its frequent but fairly minor eruptions. The 18 December eruption began after a ew months of earthquakes, and a dike injection underneath the caldera. The eruption lasted ten days. It occurred on a fissure along the southern caldera wall. Less than 0.1 km3 of lava was ejected.

2000 Hekla

We leave out an uncertain mini eruption at Katla in June 1999, and move back to Hekla. On Feb 26 it gave a longer than usual warning that something was brewing: earthquakes started more than an hour before the eruption. The signal was recognized in time, and an advance warning went out on the radio 20 minutes before the explosion: a unicum for Hekla! The initial explosion reached 11 km height. It gave rise to some pyroclastic clouds, the first reported Icelandic pyroclastics of the century. The eruption lasted until March 5. The lava covered 17 km2, and the eruption volume was around 0.1 km3. This event continued the pattern that after the larger 1947 and 1970 eruptions, the eruptions became less strong and shorter. Hekla has not erupted since.

Hekla 2000.

2004 Grimsvotn

A jokulhaup occurred on Oct 27, and it immediately led to warnings of an impending eruption, based on previous earthquake activity and the know relation between jokulhaps and Grimsvotn caldera eruptions. The eruption started on Nov 1, on a 1-km fissure along the southeastern caldera wall. It lasted for five days. The eruption quickly broke through 200-meter thick ice, and the column reached 12 km in height. This was a typical small Grimsvotn eruption.

2010 Eyjafjallajökull

This was one of the more unexpected eruption of this period. Eyjafjallajökull erupts perhaps once every 300 years. There had been signs. The first intrusion was as early as 1994. A further intrusion in 2009 and one in early 2010 brought the pressure to breaking point. It broke on March 20 on the east side of the volcano, at Fimmvörðuháls. The fissure dried up a few weeks later, on April 12. Two days later, at 1:15 on April 14 a weak summit eruption began. At 6:50 the ice cover broke and the real eruption followed. It was not a major event, but luck was lacking. The eruption column was 10 km high and very ashy — and the wind blew in the wrong direction. European air space closed, and the world’s era of volcanic fragility began. The eruption ended after 40 days, on May 22. The world never went back to its old self-confidence. This was the time that (apart from Indonesia which had learned long before), governments first began to listen to volcanologists.

2011 Grimsvotn

In the evening of May 21 Grimsvotn erupted once more. It was a large eruption, orginating from the same location at the southwestern corner of the lake as the 2004 eruption but much more significant. And unusually, this eruption was not closely associated with a jokulhaup. The eruption plume reached 17 km. Ash fell widely over southern Iceland. The eruption quickly slowed down and it ended by May 28. Some air space was briefly closed, but overall the wind was much more favourable than during 2010, and during the peak of the eruption the ash was blown away from the main transatlantic routes. The eruption was a high VEI 4.

How Iceland has changed. But not its volcanoes.

2014 – 2015 Holuhraun

What can I say? The time period we cover ended as it began, with a fissure eruption from Bardarbunga. But there the similarity ends. Holuhraun was the most voluminous eruption of this period, and by far the longest continuous eruption. It started in Aug 2014 and ended in Feb 2015. The magma was transported more than 40 km down-rift before breaking the surface. As the fissure erupted, the caldera collapsed in a seemingly never-ending sequence of earthquakes. The eruption volume is 1.2 km3 and the flows covered an area of 80 km2 in the unoccupied region between Vatnajokul and Askja. The eruption occurred in the same region as the 1797 rift eruption. In Iceland, eruptions repeat.

The next 120 years?

We have left out a number of possible but disputed eruptions. Even so, this is an impressive list. It shows the phases, the rising and waning of eruptions with the quiet era mid-century. It shows the number of rift eruptions. And it shows that not all eruptions are expected or predicted.

What is next? Which volcanoes will announce themselves in the next 120 years? We can make one easy prediction: Grimsvotn is likely to erupt within the next year, and some 10 times more during the period. The 2021 eruption will likely be a small eruption from the summit, as that is the most common. After that, three other regions are on the volcanic radar. Oraefajokull has had a significant intrusion in recent years. It is coming back to life and perhaps in another 10 or 20 years, after another 1 or more intrusions, will give way. If that happens, it may well be the largest Icelandic eruption of the period. The Reykjanes peninsula is showing increased activity, and seems ready to play its part again after 800 years of silence. It is uncomfortably close to towns, although Reykjavik itself is safe. Askja is having earthquake swarms, and seems a likely candidate for the next rift. Otherwise, we may expected 1 or 2 Bardarbunga eruptions and 1 to a few Hekla explosions. And we should expect something unexpected. After all, this is Iceland: independent, self-reliant, and ready to play on the world stage.

And how has Iceland fared? In 1900 it was fishing nation with some agriculture. In 2020, fishing remains important but for the economy, it has been overtaken by tourism. And the tourism uses those volcanoes. In the past, a badly placed eruption wiped out a quarter of the country. Now, volcanoes account for a quarter of its economy. How times have changed.

Albert, November 2020

106 thoughts on “Iceland’s eruptions since 1900

  1. Great post & an interesting list.

    You don’t mention Katla as a contender for a future eruption in the (geological speaking) not too distant future. Is there any reason why she is less likely to erupt than the others?

    • Katla is a massive volcano and probaly haves a quite robust supply, been over 400 ”large” VEI 4 holocene eruptions and some large VEI 5 and VEI 6 eruptions. Katla does larger but more infrequent eruptions than Grimsvötn. The 2011 Grimsvötn maybe the typical Katla Eruption ( Katla may produce 10 km3 every 1000 years ) so its quite productive
      But Grimsvötn is much more productive but smaller eruptions

    • I woud go for Grimsvötn next time as its Icelands most active system, but Hekla and Bardarbunga are likley too.

      Bardarbunga may sleep for now knowing it already had a massive drainout ( 1,4km3 ) are likley enough to keep her shut. Thjorsahraun must have keept Bardarbunga dormant for many many 100 s of years it became totaly drained out.
      Perhaps only recently, Bardarbunga have recovered from the 1400 s and 800 s big shows. Bardarbunga seems to have a quite prolific supply

      Hekla have increased her activity in the 1900 s and become more mafic than before ( the years before where infrequent ) she coud erupt at anytime, or stay dormant until 2180. The more Hekla stays dormant the more she will crystalize and evolve.
      In 2000 she was Basaltic Andesite as she flushed out earlier andesites and dacites

    • Katla does not sit directly on the plume center, so its not as frequently active as Vatnajökull is, but on a geological scale Katla is highly active once or twice every 100 years, faaar more active than Icelands western volcanoes near Reykjanes and Langjökull

    • Katla is likely to erupt over the next 120 years, perhaps more than once. The current repose time is the longest for many centuries, I think. But that says little – it could easily sleep for another 100 years if it feels like it.The reason I did not mention it specifically is that there is no sign of any build up, and it has been quiet for so long.

      • Katla is s semi alkaline basalt too. Alkalinity is a sign and measure of magma production, the more alkaline it is the smaller the melting production is in general. Vatnajökull is Thoelitic and haves higher melting rates in the mantle. Katlas magmas are product of smaller melt rates than Grimsvötn

        The Eldgja eruption is amazingly huge for being even a mildly alkaline basalt, always belived that Eldgja was Thoelitic basalt, but its not.

        Katla may become fullyThoelitic in the future as the east rift becomes dominant and the Hotspot stronger that will increase melt rates in the mantle under Katla

      • Seems to be no holocene shields from Katla, none acually lots of magma gets trapped at depth

        • Katla is not as close to the plate boundary as some of the other Icelandic volcanoes so her magmas would differ from Bardarbunga and Grimsvotn. But she does appear to be on a rift. Unfortunately, as her last large subaerial eruption was in 1918, we don’t know what the precursors in the run up to an eruption are; there are only the eye-witness accounts as the eruption started.

    • 2 magnitude 3+ just southeast of the caldera. Eruptions dont happen often in that area but its where the rift and summit connect, and theres been quite a lot of quakes around keanakako’i crater at about 2 km deep this past few days. Neither quake shows up very well on the seismometers so it might be tremor like the one that happened at the south edge of the caldera a few weeks ago.

      • The last 2 swarms around Keanakako’i are long-period earthquakes, the last one conincides with the start of the ongoing DI event, so they are probably related.

  2. Good article, Albert!

    The Vatnajokull glacier is shrinking at such a rate that it could be at least 80-90% gone before the 24th Century. Or maybe completely so. The Ok Glacier is not the only one likely to disappear within the next few decades, that’s for sure.

    The retreating of the ice does have me wondering if Iceland’s going to see more frequent eruptions over the next century or so, especially those underneath the Vatnajokull ice cap. That is, even if there’s not another pulse of magma from the Icelandic mantle plume. Less ice on top means less pressure on the magma below. Iceland did indeed have a major uptick in the frequency and intensity of volcanic activity during the Early Holocene, as the Icelandic ice cap was retreating greatly.

    • The Early Holocene uptick in activity had the interesting effect of volcanoes starting to erupt large monogenetic shields all over Iceland. Large fissure eruptions like Thjorsa may have also been encouraged by the relieved pressure.

    • True. But it does make a lot of difference whether a 1-km ice sheet covering the entire nation melts, or a 200-meter sheet covering 10% of the country. I would expect melt but it will not be like the early holocene

  3. Thank you, Albert! An enjoyable summary of Iceland’s volcanic activity. Perhaps this will set the scene for the next event…

  4. I npticed it says Askja is a good candidate for the next rifting event but it is only 150 years since it last did so, and that was a major rift probably much bigger than 2014 even if the volume erupted was less. I dont really know how often that part of Iceland is supposed to rift but 150 years seems not long enough after something like that. Krafla fires was 270 years after the previous rift, that seems a good benchmark of interval to expect north of Askja, but maybe that is too long and it rifts more often closer to the hotspot.

    Next rift I think will be somewhere southwest of Vatnajokull probably in Veidivotn area because its been several centuries since the last eruption. Problem is eruptions there are from Bardarbunga and Holuhraun was also from Bardarbunga, I dont know if it has historical precedent on doing multiple large rift eruptions per cycle. Maybe one of the other volcanoes near Bardarbunga will get its time to shine. I know of Thorsjahraun but that was 9000 years ago, eruptions there more recently are not too different from Holuhraun, so big flow without big problem 🙂

          • 40 km3 apparently, but its not well studied and I wouldnt be surprised if it was more than one eruption. In any case it is a pahoehoe shield like Pu’u O’o, it was not really the sort of eruption that would be visually impressive at any given moment.

            Biggest eruption that was fast was probably Thorsjahraun, then Eldgja.

          • I said this in another post, it is about 10x the volume of Pu’u O’o, so maybe 350 years. It might have been a bit higher rate but in all it seems the volcano was probably active for at least a century easy. Its probably not the absolute biggest monogenetic shield though, theres a few all over the country that look to be of similar dimension. Its also never confirmed, only speculated, that the shields are actually formed in one continuous eruption.

            Its also not known for sure but also not unlikely similar volume eruptions have happened in Hawaii and in Africa. Average long term supply to Kilauea has seen it possibly get over 150 km3 in the past 1000 years and the filling and overflowing of the old Powers caldera took a semi-continuous 5 century eruptive event that should be of comparable volume to Theistareykjarbunga though probably determining this is very difficult now.
            I think the Tarso Tousside volcano is also mostly a pahoehoe shield with a strombolian volcano on top (separate eruption?) and that might be of similar volume too but not much is known apart from its age being mid Holocene. Erta Ale is also probably entirely Holocene.

      • Just south of Theistareykjarbungas summit there is a very impressive tube channel system!

        These are the biggest tubed channel system in holocene. The lava channels merge into semi tube pit chains downhill, that can be followed for 10 kilometers in two paths. The vent of this is larger than Baugur and channel system. And it seems to have flowed like that for a few years with medium to high eruptive rates. A rare example of a fast long lived channel trying to form a tube system. Its two chains of massive tube pits, many dozen s of them. The lava rivers of these keept flowing for certainly a few years. The tubed channels are as wide as 70 meters and becomes more tubed downhill. Overflows from these carpets that arera too. It woud have glowed like hell, two arms with numerous dozen glowing holes, and Big dome fountains at the summit

        This feature of Theistareykjarbunga is perhaps the only recent example on Earth. The other examples are on Mars, Venus and IO.

      • Theistareykjarbungas lava channel system on its northen slope probaly started with a massive flank eruption with 600 m tall fountains that later focus on one vent.

        That later dies down and it becomes a boiling bathtub with huge dome fountains and the channelized Aa rush downslope, the eruptive rates seems to have been many 100 s of cubic meters s second, with maybe 100 a second when settling into a long lived phase. The fast channels flowed for a very long time, because they started to tube, and perhaps they where huge tubes, that now collapsed into pit chains. These are massive channels – tubes capable of carrying very large volumes downhill. Its likley that the original fountain cones are long since buried there. The tubes are also deep and buried below other flows. A wonderful sight in the winter darkness it must have been

        Fast eruptions normaly does not form tubes, because that takes a long time but at this it seems to be the same process as the Big marsian or Venusian vents

        But the biggest tube channel system on Venus is 6800 kilometers long!

      • But most of the Theistareykjarbunga shield was much slower than that it was as slow as Puu Oo, with no open channels with pahoehoe and large placid calm lava lakes residing in its summit pits. Not a very visible eruption, it resembled perhaps Kupaianaha but much larger

        The channel system was a late faster event

  5. Nice overview, thanks Albert.
    6 years ago…, looks like ages!

    Some footage.
    Really spectacular vids in those days…, and thinking back at all those hours watching the mila cam… and the time lapses made by vc contributors.
    Hoping for real action soon.

      • Did Holuhraun have surges whenever there was a collapse quake at Bardarbunga?

        There also looks to be lots of pahoehoe along the channel just like in Hawaii, maybe it doesnt look as shiny in Iceland (chilled faster?) so is harder to see than it is in Hawaii, but its quite clear up close in that first vid. The channel also was really low so maybe that does go with the idea of pulses in effusion, though that was at the start of the eruption so maybe not so obvious yet.

        • I don’t remember any relation of the surges with the collapse quakes.
          I remember though we counted the surges and it was obvious that number was decreasing over time.

          In fact the fissure with many fountains in the beginning of the eruption had become the large lava bath tub (named Baugur later) with three main lava sources. The lava poured over one edge of the tub (growing walls created by the spattering of the fountains) into the river. The force of the sources was strong. The lava was surging spectacular in the tub. At the end of the eruption one could see clearly the fountaining was becoming less high and pulsing with a max once in 10 minutes or so. Apparently the pulsing was showing the pressure reached an equilibrium.

          • The thing that surprised me a bit from the post was that the official volumes of Holuhraun and Leilani are actually the same, 1.2 km3. I thought Holuhraun would have been at least bigger, given it lasted twice as long, and that video it looks like 3 fissure 8s going at the same time. Only thing I can think of is that Holuhraun was much less linear. To have these two such similar eruptions in the same decade too.

          • Holuhrauns volume is 1,4 km3 in most papers
            So just a bit larger than Leilani

          • Bardarbunga is probably spent for this round. 1477 eruption was apparently not hugely bigger than Holuhraun, 4 km3 DRE, 1862 eruption was under 1 km3, both of those werent followed up by other similar eruptions so probably we have to look elsewhere. Problem is Veidivotn area seems to rift every 400-500 years so its right about due. Isnt there a volcano southwest of Bardarbunga, that might be the one to watch out for instead,

            It looks right now that the next lava flows in Iceland will be either at Reykjanes or maybe Askja, possibly in Hekla area.
            Reykjanes it looking very interesting now though, non stop intrusion and swarms all year.

          • Katla is quiet.
            Grims earthquakes are ar low rate. Still and not making a lot of progress the past months.
            The Reykjanes unrest are adjustments of slip strike events earlier this year, in my opinion.
            Askja has deep quakes occasionally, but no bursts what we should be looking for.

            I grabbed a strip posted in ancient VC days

        • 1,5 km3 is also referenced for Holuhraun. Holuhraun had faster eruptive rates the first month to first weeks.

          I agree that Bardarbunga central volcano is pretty much drained out for now. Its struggling to recover, and sits on spreading that steals alot of incomming magma. There wont be a huge Bardarbunga effusive flow for a ”very long time” Bardarbunga is not a frequent erupter of large rift lava flows

          • Seems the most recent study of Holuhraun, in 2018, found the volume to be about 1.2 km3, so the two really were the same size.

      • When Holuhraun started I really thought we woud get a new 100 years long dynga shield, but it never tapped the large deeper resovairs required for that.

        Instead it drained the magma in upper parts of Bardarbunga and drained itself dry in 5 months

        • The largest such eruptions act through an intermediate magma chamber. The second chamber feeds the eruption and the pressure drops below eruptable values after a few months. But the primary reservoir is feeding the intermediate one, and one the eruption stops the intermediate one inflates until it opens a new channel. Laki did that 5-10 times. Holuhraun did not have that recharge.

          • The 100 s of year long Icelandic shields then?
            How are they feed? Is it massive resovairs or constant supply from plume? Decompression melting perhaps
            Can a massive shield happen again after the Early holocene?

          • Bardarbunga I think is probably more likely for a shield, it has done one at least once from the looks of it. That area is pretty much at the hotspot so unlike the rest of the shields near Thingvellir and up north this area is probably still up to the task, though maybe not until the glacier melts, that is probably what happened the first time.

            I do wonder if that shield ( trolladyngja I think it is called) was terminated by a flood basalt, I dont think the area is very well dated outside of recent flows because they bury each other, but if big shields in Hawaii are terminated by flood basalts it seems if a shield forms above a rift in Iceland the same thing would happen.

          • The next rifting event of the volcano would be expected to kill the shield , and this could end in a large fissure eruption, so that is an interesting option.

          • Trolladyngja is 1400 meters tall, and its base is about 950 meters above sea level, most of the Veidivotn rift is between 500 and 700 meters above sea level. Trolladungja formed apparently about 7000 years ago but not well constrained, theres also a Veidivotn eruption 200 years later that is equally unconstrained, so it seems at least plausible, though I have no idea how big the latter eruption was. Thorsjahraun was 1500 years before Trolladyngja and seems pretty securely dated, so seems there isnt a connection there.

            Maybe the only problem with this is that Trolladyngja doesnt have a summit collapse.

        • There is many black Aa flows among Trölladyngjas grey shiney pahoehoe. So they coud be because of high rates overflows, or high rate flows from bursted lava tubes similar to Puu Oo tubes leaks.

          But Theistareykjarbunga seems to have been a more massive shield eruption. Even today there is a very large magma resovair under there. Last eruption was 2000 years ago, so in theory this Arctic Erta Ale coud still be alive. The larger the shields grow, the more developed and established their plumbing systems becomes

  6. Green star at Hellisheiðarvirkjun Iceland. To my amateurish eye and Alzheimer reduced memory this looks like a slow jumping progression of activity from the western tip of Reykjanes east along the ridge.

    • It’s not very far from a geothermal plant. OTH the earthquake swarm (small so far) seems to line up along a fault line.

    • Your memory has been impressive and I would not have guessed. And I too would expect that when the peninsula breaks, the main activity will initially be in the east.

    • None of their articles are well written, and seems mostly be made to make money and spread alarmism as well conspiracy theories. Nibriu cataclysm have often popped up on that site😂
      I never reads these articles from them.
      Mostly paranoia in Daily Express

      Edited to remove politics – admin

    • I like the reader comment on the article: “now I have something great to look forward to that will get me out of this lock-down situation”.

      It looks like the readers are also rather world-weary about the Express’s fascination with this fading volcano! And it’s just another writer flogging a book. I doubt Bryan Walsh will make much money out of this old chestnut!

        • Not I. I burned out a few weeks ago. Plotting growth curves can take it out of you. (intently waiting for the next NHC update and listening to the weather twits on TV making total arses of themselves with their mindless banter.) {I actually saw a Weather channel idiot put a TS graphic up on the screen that was rotating the wrong way one time… not to mention a frontal storm a few months ago where he pointed out a well defined rear flank downdraft and called that the tornado when you could clearly see the couplet on his Doppler about 2 miles in front of it.} → Sometimes it’s funny, usually it’s just pathetic.

          Do note that this particular one has a growth curve that shattered all of my preconceived notions from watching them.

          For reference, WINTER, “Bomb Cyclones” strengthen at about 1mb/hr. “Rapid Intensification” (an old term used in tropical systems) is about 1.7 to 2 mb/hr. This thing blew that out of the water. It’s intensification rate was in a nutshell… “obscene”.

          • I started noticing BS concerning the WC since I was a kid, but it kept my interest. I am sure you’ve seen the video of that guy faking struggling against modest winds.
            This year has set a high mark for numerous rapidly intensifying tropical cyclones, that could be surpassed next year if this la nina continues or strengthens. Iota had an deepened at rate of 5 millibars an hour if I am not mistaken, would be awesome if this happens with an extratropical system.

          • And it was not the first one to do so this year. Hurricanes seem to suffer rapid intensification more often. You can interpret that as an indication that their early formation is being hampered by something, and once they get going there is a lot of heat still available. The form late and then grow up very fast. The total storm ‘power’ this year has been high but not record breaking. The number of storms has been unheard off. The storm power has I think been held back by the high number of landfalls. The bottom line: if you keep on pumping heat into the oceans, the oceans will come back to bite you.

      • Just remember that it is making landfall on the same bit of coast that Eta used just two weeks ago. It is going to be catastrophic, and likely there will be many casualties from the already battered communities.

        • It is a very sparsely inhabited area, I believe. The biggest risk is probably further in-land, from rain.

          • True. Around 74 folk died in Eta mostly from land and mudslides. That will likely increase, as the ground must still be totally saturated. The NHC in one forecast have predicted up to 30″ of rain in some parts.

  7. Are there any volcanos in the impact zone that could produce lahars with all that rain?

    • It is predicted to head for the southern part of Central American Volcanic Arc at Honduras and Nicaragua. I don’t know how far inland the rain will reach but we are looking at the narrow bit the joins North and South America.

  8. Come on Chad, you could’ve done better than that. you didn’t really argue against me as much as you stated your opposite position. Be more aggressive this time, it won’t hurt my feelings.
    I could monitor 1000 volcanoes and still miss out on the lottery, I have got a decent deck with 5 volcanoes, 3 of which I have already mentioned and the other 2 will have articles


    The lava flows and vents of all the eruptions in this post. The vents under glaciers are mostly uncertain in location though, for obvious reasons.

    It seems pretty likely that except for Reykjanes probably erupting some point in the next decades there will be more eruptions in the areas as on this map, Hekla area and Vatnajokull, and maybe Katla eventually.

    • I must say I was quite surprised after making this. Maybe to be expected, Bardarbunga was the most productive volcano, Holuhraun was 1.5 km3 more or less, and the 1910 eruption seems to have been quite big, probably similar to Gjalp, about 0.4 km3. I estimate about 2 km3 of magma has erupted on the Bardarbunga system since 1900.

      The second place is what surprised me, it is Hekla, not Grimsvotn. Hekla erupted 1.45 km3 of magma since 1900, 1.47 if you add the 1913 eruption being in the general area. While apart from 1947-1948 most of the eruptions were smaller than the average, there were also 5 of them, it seems the productivity of hekla now is maybe over double what it was before 1104. For being not anywhere near the hotspot the area is very productive. Katla also has been totally silent at the same time as Hekla has erupted 5 times.

      Grimsvotn has erupted, since 1983, 0.75 km3 of magma, and 1.2 km3 of magma since 1900. That is quite decent, about the same as Etna. It is however, rather lower than the reputation it has on this site, and it is nothing even close to Hawaii. The ridge is a double edge sword, it keeps the supply high but also hinders the plume from erupting.

      • Yes, I have often said that when looking at the long-term history of Iceland, Bardarbunga is the real beast, not Grimsvotn.

        • Yes, it is quite obvious which one has erupted more lava when you look at the dozens of absolutely massive a’a flows that extend from the Bardarbunga rifts, and not just Veidivotn rift but also Trolladyngja rift and Holuhraun rift have done full scale flood basalts too.

          Veidivotn area rifts every few centuries, its been a good 500 years since the last big rift there so probably a good spot for the next one this century. Bardarbunga drained out during Holuhraun so its probably going to stay small for a while but Hamarinn is still inflating and primed, and it is next to Veidivotn… 🙂

      • Grimsvotn is a very frequent eruptor, enough so that it dominates all statistics on Iceland. but it does not erupt very much. The 1903 eruption may have been larger. Its lava output is completely dominated by Laki.


        • Yes, I suspect that Grimsvotn is still recovering from the events in the 1780s. Like in your analysis posts it was more complicated than just a bigger Holuhraun and magma was supplied from depth in large part, but there was definitely a drain in total, probably at least the southern caldera is from then. I read the old article from Carl too, he says there was also eruptions in 1783 at the southeast edge of the ice and the other central volcanoes, a second fissure system, if that is real then this was a very big event and probably not likely to repeat again for millennia.

        • Grimsvötn did 0,7km3 in 2011. And another 0,7 km3 in 1996, but 1996 was n”Not a ”competely fresh” basalt so maybe 1996 does not count?

          2004 and 1997 did both around 0,1 km3 with 2004 being sligthly larger than 1998

          • Again, 2011 0.8 km3 is for the tephra which is much less dense than solid basalt, the DRE is 0.2 km3. Gjalp is 0.4 km3 DRE on Global Volcanism Program. All the other eruptions in the past 40 years are under 0.1 km3 DRE. As I said before a few times I think this site has overhyped Grimsvotn and no-one really bothered to factcheck.

          • It is, but its not fed by only 1 volcano, all of the central volcanoes erupted in 1783-1785. If it was only Grimsvotn feeding the event then it would be in a very deep sleep right now, and for millennia to come.

            Skafta Fires was an extraordinary event, the 3rd biggest eruption in Iceland in the Holocene. Many shields are similar but I dont really count shields this way because leaking for 100+ years vs flooding 1000 km2 of land under lava is not the same thing. Think Pu’u O’o vs fissure 8.
            If you add up just Thjorsahraun, Eldgja and Laki you get nearly 60 km3 of magma to fuel these events, the entire Holocene output of Iceland is estimated at about 250 km3 DRE, so 3 eruptions are responsible for about a quarter of the total. That is how colossal these eruptions are, only 3 years out of 10,000…

          • Basically what I am saying is to expect another Laki this century is very unrealistic, these eruptions probably require their volcano to be an actual mountain and then create the caldera, Katla Bardarbunga and Grimsvotn are all sizable calderas, Bardarbunga is the most recovered but it still isnt capable to do anything close to Thjorsahraun today.

            The only volcano in Iceland today which is both productive and not a caldera is Hekla, but it isnt really old or big enough to have a 20 km3 flood basalt, its probably maxed out at about 1.5 km3 which is a pretty standard large flow in Iceland.

          • Think Laki may have been more a consequence of a rifting event rather than “just” being fed from Grimsvotn, or, indeed, Mt Laki.

          • That is the question. There was an intermediate magma chamber: we know that from the crystals. But it received a high influx from the deeper reservoir shortly before and during the eruption. The biggest mystery is the missing crater. Grimsvotn isn’t large enough. Either it has recovered, or it was distributed over a larger area

          • That seems to be the idea that was in Carls article, the fissures were fed out of the mantle, but that was before Holuhraun and Leilani. There really isnt very much that is known about most of the subglacial volcanoes, really we dont know if they are erupting unless the icecap breaks even today. Given that the subglacial eruptions during the Skafta Fires were not overly large or unusual and there was a massive lava flood at the same time it is entirely likely that the calderas of both Thordarhyna and Grimsvotn formed that year and no one would have noticed in all the action.

            I do remember that before the eruption of Oraefajokull in 1362 the outwash plain south of Vatnajokull was a habitable valley, but now it is a wasteland prone to lahars from eruptions. It doesnt really make sense for it to be so different even with the Medieval warm period until you consider that maybe Grimsvotn didnt have a caldera back then. No caldera means no lake to collect water and cause frequent flooding. The only eruption in the Holocene big enough to create the caldera was in 1783, and notably there wasnt a flood in 1783 either, and reports of glow in the glacier which suggest open vents.

          • Laki is also just a mountain, its not a vent I dont think, its like all the other hills in the area just the 1783 rift went through it.
            There is also a lava flow called Botnahraun which is mostly buried under the 1783 lava and is on GVP refered to as Laki, probably because of confusion more than anything. I think this is the flow that in Carls article he called the Laki fires.

          • There was evidence of a mantle source in Holuhraun, in addition to the feed from Bardarbunga.

            Kilauea is fed solely from a hotspot; it is not at a plate boundary, so its rifts are a slightly different mechanism.

          • I don’t think there was. That was an early claim based on incomplete analysis. To my knowledge, it was not repeated afterwards and the close relation between the eruption at Holuhraun and the collapse at Bardarbunga left little doubt about the source of the magma. Of course, eventually all magma has a deep origin. The question is about how long it spends in the magma chamber of the host volcano.

          • There was a subaerial eruption of Grimsvotn during the Laki eruption. There was also a significant volume of tephra. We do know from Holuhraun and others that fissure eruptions on their own do not tend to form a lot of tephra. Either the tephra came from Grimsvotn or Mt Laki or there was a significant body of water in the area (not seen any accounts of a pre-eruption lake) – or all three.

          • There was definitely a significant body of *ice* in the area. Perhaps the northern tip of the fissure was under the southern margin of Vatnajokull. If the ice was thin enough there, the eruption could punch through and do an Eyjafjallajokull.

          • There was a 2 year long explosive eruption at Grimsvötn (at its summit) that could be related to a caldera collapse, being thus a big basalt blast. Do we know the exact volume of the caldera? It seems quite big.

            Another interesting fact is that the fissures opened episodically backwards towards Grimsvötn, this suggests a series of dyke intrusions, perhaps 10, each reaching a smaller distance from the central volcano, presumably Grismvötn. This behaviour has been observed in the Krafla Fires and in the Manda Hararo rifting of 2005 in Afar. See how the dykes grow smaller distances from their respective central volcanoes, which are Krafla and the Ado’ale complex:

            Hawaiian volcanoes also do a similar thing after their rifts get opened up by flank slip, where dykes migrate back towards the summits over a few decades. Mauna Loa did that after the M 8 earthquake of 1868 where each following southwest rift zone eruption opened a bit further uprift than the previous one, as magma closed back the rift.

            My opinion is that whatever fed the Skaftar fires must have been under Vatnajökull and that the shallow chamber of Grimsvötn must have gotten involved even if it was backed by other deeper pressure sources. Magma could have originated from anywhere that was in a pressure balance with the summit of Grimsvötn.

          • The backwards opening is normal, and understandable. Magma flowing in a dike will cool on the way. Once too cool, it will solidify and stop flowing. The cooling rate depends on the flow rate, and this depends on the pressure. As the pressure drops, the flow rate reduces and the magma will solidify earlier. Therefore the eruption will stop, and resume closer to the source. This can happen several times in a large eruption. The situation is more complex than this, of course, but the basics is correct.

          • Oops that image is incomplete, here it can be seen better:

          • The colored bars represent dyke intrusions in those diagrams, and AVC stands for Ado’ale volcanic complex

          • I am with Merlot here. Maybe we think too much central volcanoes as inbetween station of the magma in (at least) some of the events.

            We like to imagine a (or more) chamber(s) in which magma sits maybe a bit to much? What is happening at depth of 10 to 25 km?

            The rift doesn’t affect the upper part of the crust only. In some hot parts, like the dead zone seems to, not much signals are picked up. That hot parts, within EVZ and NVZ are remains of non dyke eruptions and wouldn’t have to be sourced by central volcanoes only.
            Question is then if there would be a missing crater or some evidence remaining at the crusts surface.

          • North end of Lakagigar craters definitely goes under Vatnajokull, theres a hyaloclastite ridge which lines up that is melting out of the glacier. Probably trying to say with certainty anything about the exact location of the subglacial vents is impossible, so its entirely likely the observed eruptions were extensions of the fissure. That also would make sense with most of the activity being after the lava flood.
            Thordarhyna probably also erupted in the Skafta Fires.

            As far as future large lava flows go in Iceland, Grimsvotn system is a no go for me, its very infrequent. My two picks are Hamarinn and Hekla, Hamarinn is still primed and is next to Veidivotn, its pretty much the same as Bardarbunga before Holuhraun but in the opposite direction, I would not bet money on this scenario happening this century though.
            Hekla is seemingly not in the habit of doing eruptions every 10 years now, so its next eruption could be quite big. The combined 20th century volume is larger than any single eruption in the preceeding millennium, so it is possible the supply has increased and a only slightly longer repose will deal a significant effusive eruption along with the usual VEI 3-4 opening. Theres also the basalt vents, not all of them were small like 1913, some are comparable to Holuhraun.

          • From the maps of the lava flows for Laki’s 1783 eruption, the lava came mainly from the subaerial fissures, not from beneath the ice-cap.

          • If the Laki eruption had started under the ice-cap, there would be eye-witness accounts of a jökulhlaup(s). Not heard of any.

          • No it ended under the ice, the fissures went uprift towards the glacier over time and no doubt eruptions happened under it. Eruptions under the ice would be like Gjalp, the big flows did come from the vents outside the glacier.

            Again it is completely ambiguous to where the subglacial eruptions were. The lack of floods is what makes me think the caldera wasnt there, it is the lake overflowing which causes floods now, back then it probably didnt exist and the caldera it occupies was created in the Skafta fires. Before 1362 Oraefajokull eruption the area was a habitable valley, and theres no way to explain why that is the case unless you assume floods were very rare or even absent for millennia before. The area was wiped out by Oraefajokull but the floods dont come from there, they come from Grimsvotn. There would be no flood to the south without Grimsvotn lake, the floods in 1783 would have gone down the Skafta river which was full of several km3 of new and still hot lava.

          • Tensile stress being relieved by the intrusions probably plays a role too in the backward opening. The first intrusion always travels the farthest away because it is easier to snap open after centuries of slow spreading. The successive intrusions fill the rift, which makes it more compressed until intrusions cannot travel much distance from the volcano. Eventually rifting ends.

            But in either of both mechanisms it points to the main feeder of Laki being located under the ice of Vatnajökull, in the direction of Grimsvötn. It may not be the same chamber that feeds the summit eruptions of Grimsvotn, but it must be part of its system.

          • The narrowing of the flow channel is a major part it, indeed. In a dike the magma pressure helps keep the dike open. As you say, this is easier during a rifting event when everything is quite relaxed, although that is true mainly at shallow depth. As the magma pressure drops, the dike narrows again and this increases the cooling rate. Magma can’t flow as far in a narrow channel.

            Laki was certainly fed from the direction of Grimsvotn.

            Eldgja however had a late eruption at the far end of the rift, almost at Vatnajokull. That appears to have been caused by heating of pre-existing, cooled magma. That magma probably originally came from Bardarbunga or relatives. The different volcanic centres interact in this area.

          • So what happened to the lava and the melt-water for the subglacial part of the fissure – there is a lot of ice-cap between the edge of the ice-cap and Grímsvötn, itself. Both the lava and the melt-water must have gone somewhere.

          • The lava would be hyaloclastite, so no flows, just like at Gjalp. The water would have gone in the Skafta valley and had to flow over the colossal lava delta, which could have evaporated it entirely, even if not the flood would have been relatively minor compared to what had just happened, it could have been ignored.

      • Africa haves lava lakes despite being rift + hotspot. Thats because Africa spreads 10 times slower than Iceland. And more magma can come up in Africa. But Iceland remains more productive as Grimsvötn than most single African volcanoes. Iceland is also oceanic geology.
        Nyiramuragira benefits from souch a slow rifting, that does not lateraly steal its magma as what happens in Iceland. Nyiramuragira is becomming a very productive rift volcano.

        If Iceland Hotspot emerged in the ultrafast EPR that spreads 18 times faster than Iceland Atlantic, its unlikley that Iceland woud even reached the surface, it woud form a boulge at the ridge, but no Island. Eruptions there are long but short lived fissure eruptions with more than 10% ending up below surface. EPR must be magma stealing as heck, and thats why NO central volcanoes have formed over the ridge axis swell

        • I means perhaps less than 10% of an intrusion ends up on the surface in the super ultrafast EPR

  10. We have a seismic swarm at Kilauea, the pressure of the Halema’uma’u reservoir has done a small but very quick pressure jump after the last DI event. This rapid change in pressure has started a ground cracking swarm. I expect the earthquakes will migrate south-eastward through the East Rift Conduit and also perhaps through the Southwest Rift Conduit though this one is usually more quiet. This could happen because the pressure increase may propagate as a wave through the molten cores of the rift zones.

    Other than the earthquakes it is unlikely anything else will come out this, it is the routine stuff at Kilauea although a little bit more dramatic than usual.

    • You are right about it being likely nothing happens but this shows at least there is the possibility of a small eruption suddenly happening under the lake, it doesnt take much to make a decent explosion and especially so unexpected while the alert is on lowest setting is a scenario we have seen a few times… Probably if anything these swarms will only get more frequent going forward until there is an open hole again, at which point eruptions resume with enthusiasm.

  11. Given that my location is surrounded by both volcanoes and mega/giga-fires plus a prolonged period of drought (save for a couple of “blip” years in 2010 and 2017, we’ve been in/out of severe drought since the turn of the century) it’s only natural to be curious if similar weather/climate linkages exists between volcanic eruptions and Pyro-Cb’s + massive amounts of lower elevation ash and debris as generated by mega-fires.
    Here are two excellent (recent) papers that go into detail on the atmospheric/stratospheric impacts of the Australian wildfires of 2019-20 and the Northwest-America fires of July-Sept, 2018. Of note is the comparisons of the wildfires’ injection height, rate and composition of the various effluent that got thrust into the stratosphere as compared to a moderate-sized volcanic eruption (with the VEI 4 Kasatochi eruption in 2008 used for reference).
    But, there is another potential climate impact I had not considered; which is a mega-fire like in Australia produced a self-sustaining anti-cyclonic vortex over a 1000 km diameter that perpetuated in the stratosphere (>35km) at detectable levels for 13 weeks. ATTM, I have not seen any such research into the impacts of this current Summer/Fall’s massive wildfire outbreaks over much of the western N.A. continent (which included two “giga fires”). As I write, fires are still burning in many areas (especially in southern California and the desert SW), with total acreage burned approaching 4,500,000 acres in California alone (the previous high was a mere ~1,900,000 acres burned in 2018). While it’s going to take some time for re-analysis to reveal if/how much impact this fire season has/is having on our weather (including a record-setting NA Hurricane season) I suspect that it is rather significant.
    While admittedly we’re not talking about a global impact on the same scale as Pinatubo for any single fire, but then again, we’re now seeing the Arctic tundra (i.e. Siberia) burning at an unprecedented rate, plus here in the western U.S. we are just now exiting our steadily lengthening “fireseason” that literally dwarfs any previous fireseason. In Colorado early in October, Denver went from 101F to heavy snow in less than 48hrs as massive, record-setting fires riding near-hurricane force winds descended on many towns/cities including Boulder and Estes Park. In total, virtually every state west of the Rockies experienced multiple fires that eclipsed (by a wide margin) the largest previously documented fireseason.
    Anyway, I could go on and on about this emerging metric and the real and theoretical implications of the changing composition of the both the troposphere as well as the stratosphere from the clearly emerging threat of wildfires on a global scale…..but this is a volcano blog after all!

  12. Green star on Katla. Specialist remarks “This morning at 11:08 a magnitude 3.1 earthquake occurred on the eastern part of the Katla caldera. No aftershocks nor tremor has been detected since.”

Leave a Reply