The 1996 Bardarbunga Eruption, Iceland

Saturday, November 23, 1996

Alan V. Morgan,
Department of Earth Sciences,
University of Waterloo,
Waterloo, ON
N2L 3G1

The eruption

Iceland has long been known as the "Land of Ice and Fire", and the term came to life again at the end of September when these two ancient elements again fought for mastery beneath Europe's largest glacier. Ever since the 1973 eruption of Eldfell on the island of Heimaey off the south coast of Iceland the residents of this northern country have carefully watched any signs of impending volcanic activity, but this was to be an eruption with quite different consequences.

In mid morning of September 29, 1996, a force 5 earthquake was detected beneath Vatnajokull Ice Sheet in southcentral Iceland. Over the course of the next 36 hours multiple earthquakes of about 3 to 4 Richter indicated that magma was rising under the northwestern portion of the ice sheet. Mindfull of the intensity of some previous Icelandic eruptions, particularly from Hekla and Askja, situated southwest and north of the new focus of activity, the Icelandic authorities issued warnings of a potential eruption to national and international aviation authorities. Both of these volcanoes have issued large ash eruptions in the past and many European to North America and Asia airline routes cross over Vatnajokull. Aircraft jet engines and fine rock particles disagree with each other and at least three commercial jet liners have come close to disaster when engines have failed whilst flying through volcanic ash clouds.

On October 1, a subsidence bowl was detected in the ice sheet and further observations located additional collapse in the ice surface above an old eruptive site, last active in 1938. The fissure which was being created beneath Vatnajokull was between 5 and 6 km in length. The following morning observers noted that the new eruption had broken through the ice surface, projecting tephra at least 500 m, whilst the eruption column rose over 3 km above the ice surface. In the same 24 hour period observations at Grimsvotn (a caldera within the ice sheet some 15 km to southeast) revealed that the ice level in this region was rising. The inflating ice surface was due to subglacial water from the new eruption flowing into the depression, and raising the ice level.

Icelandic geologists were concerned about this sequence of events, because volcanic warming in Grimsvotn periodically causes "jokulhlaups" or "glacier bursts" which emerge on the outwash "sandur" plains of south Iceland. The most recent, measured at a peak flow rate of 3,000m3/s, was in 1995. The new eruption reported initially at Bardabunga (64 degrees. 63"N; 17 degrees. 53"W) continued a pattern of activity which extends back over 1,000 years, with eruptions recorded in 871, 1477, 1862 and 1910. The area is characterised by a complex of fissures extending at least 100 km beneath the ice and an additional 50 km north of the ice sheet. The Icelandic authorities have pinpointed the eruption at 64 degrees. 30" 32' N; 17 degrees .22" 00' W.

By October 7, the activity had concentrated into continued melting of the ice sheet (up to 2 cubic km of ice was reported gone by October 10) with water flowing mainly south to the Grimsvotn caldera and earthquake activity around force 4 closer to the centre. The fissure had opened to a length of about 8 km and the eruption cloud was rising to 10 km, being clearly visible from Hofn, a small settlement on the south coast. Icelandic geologists were forecasting an imminent jokulhlaup at Skeidararsandur, however, as is often the case, Mother Nature succeeded in throwing another curve ball at the scientists! The eruption died down on October 13 and, although geologists knew that over two cubic km of water had filled the Grimsvotn caldera, no water had emerged from the glacier front some 50 km away. For three weeks the scientists waited and wondered, until early on the morning of November 5 observers noted that the Skeidara River was starting to smell strongly of sulphur and that the discharge and sediment load was increasing.

The jokulhlaup

At the time of writing the long expected glacier burst has finally taken place and with more disastrous effects than had been expected. The flow rates on the Skeidara; River increased exponentially during November 5. At 0800 discharge was about 70m3/s; by 0930 it was 3,000m3/s; by 1200 the smallest bridge across the Gigjakvils River had been destroyed with flow rates of 7000m3/s. This discharge was twice the rate of the 1995 jokulhlaup from Grimsvotn, and was starting to get close to the 9,000m3/s design limits for the main bridges across Skeidararsandur. By 1500 the discharge was estimated at 20,000m3/s, and by darkness (around 1700) estimated at 25,000m3/s. Discharge peaked at about 45,0003/s. To try to put this in perspective it is about the same as the volume of water passing through the mid-reaches of the Mississippi River!

It is a little frustrating not being able to close this article without the final results, but there has been little news out of Iceland in the last 24 hours. Certainly one bridge and as many as 7 kilometres of road have vanished. I suspect that there is some damage to the long causeway across the main part of Skeidarársandur, although the main Skeidarbru bridge is still standing, and there are reports that the fibre-optic communications line has also been destroyed. Preliminary estimates of damage are at least $15m (U.S.); more than the damage cost during the 1973 eruption of Eldfell on Heimaey. Once again the citizens of eastern Iceland are faced with a 15 hour journey to the west coast through northern Iceland, and it is uncertain when the south coast road will be repaired. In effect the island nation has had its road network returned to a pre-1974 situation - a reversal of 22 years of progress in a matter of days! More on this in the next issue of Wat On Earth.

  1. 2018 (1)
  2. 2016 (4)
  3. 2013 (4)
  4. 2012 (6)
  5. 2011 (11)
  6. 2010 (5)
  7. 2009 (4)
  8. 2007 (8)
  9. 2006 (10)
  10. 2005 (7)
  11. 2004 (10)
  12. 2003 (12)
  13. 2002 (15)
  14. 2001 (17)
  15. 2000 (21)
  16. 1999 (31)
  17. 1998 (22)
  18. 1997 (11)
  19. 1996 (28)
  20. 1995 (28)
  21. 1993 (7)
  22. 1992 (6)
  23. 1991 (4)
  24. 1990 (9)
  25. 1989 (9)