The Recent Eruption Beneath Vatnajokull, Iceland.
A semi-chronological account

Below is an account of the recent eruption beneath Vatnajokull, Iceland which began Sep 29, 1996 and ceased October 12th 1996. Click on thumbnail images for full size versions.

Pictures and news were provided by the following people during the course of the eruption - thanks for making the info available.
  • Nordic Volcanological Institute photos 1, 3, 4, 5, 10 and 11.
  • Daily News Iceland daily updates and the photo above.
  • University of Iceland, Science Institute photos 2, 6 and 7.
  • The Icelandic Metrological Office maps of earthquake locations during the eruption.
  • Air Charter Inc. photos 8 and 9.
  • Tromso Satellite Station satellite images 12 and 13.

  • Initial earthquakes

    The sequence of events leading to the eruption started with a magnitude 5.4 (Ms) earthquake at 10:48 GMT located on the northern rim of Bardarbunga, a large caldera on the northwestern edge of the glacier. Events of this magnitude have occurred before in this area though they didn't lead to an eruption. What was different about this event was the swarm or fairly large events which followed, several up to magnitude 3.5.

    These maps, provided by Vedurstofa Islands, show the growth of the swarm.
  • September 29, 10:48 - 12:00 Initial event and beginnings of swarm.
  • September 29, 12:00 - 24:00 Growth of swarm and beginnings of migration to the south.
  • September 30 Swarm spreads further south to the Grimsvotn caldera.

    At around 22:30 on September 29 the earthquakes in the south ceased and a low amplitude volcanic tremor started - note the difference in earthquake distribution between September 30 and October 1. This is the time at which it is believed the eruption started and lava reached the surface (beneath the ice cover).
  • October 1 Events cease in the south.

    Over the following few days the swarm moved to the north and ceased.
  • October 2 October 3 October 4

    Photos of the developing eruption

    The subsidence cauldron

    1 2 3

    On the morning of October 1st, after the beginning of the volcanic tremor the previous night, a plane went up over the glacier to see if there was any evidence of the eruption. They found several elongated subsidence cauldrons 1 to 2 km wide and 200 to 300 m deep, photo 1 (Oct 1). They delineated a 4 km long fissure which was erupting beneath, photo 2 (Oct 1). To give some idea of scale see photo 3 (Oct 1) which show a small light aircraft over the cracks in the ice as seen in photos 1 and 2.

    The eruption

    4 5

    The first signs of the eruption breaking through the glacier were observed on the morning of October 2nd. Once the eruption had melted though the glacier a column of steam could be observed rising from the breach. Later the eruption started ejecting ash up into the atmosphere turning the white steam black, photos 4 and 5 (Oct 2).

    6 7 8

    As the eruption continued the size of the crater in the glacier grew and the surrounding ice was covered in tephra. Photos 6 and 7 (Oct 3) show the increasing size of the crater. Photo 8 is a view looking along the length of the fissure and down into it.


    Photo 9 (Oct 3) shows the steam column rising 10 km up into the atmosphere.

    Pending Flooding

    The subglacial eruption melted an enormous volume of ice along the length of the fissure. The location of the fissure resulted in most of the melt waters flowing into the Grimsfjall Caldera Lake, Grimsvotn, just south of the eruption (see one of the earthquake event maps). This subglacial lake usually fills over a period of 5 to 10 years due to geothermal melting, and then empties in a catastrophic flood, a jokulhlaup. While the lake was low as the last flood was in 1995 it filled rapidly surpassing all previously measured levels by October 4th. Many predicted immediate flooding which did not occur, the level continues to rise without any evidence of flooding yet.

    Here are some typical figures. Flooding usually occurs when lake reaches around 1450 m and a typical drop in lake level on flooding is 70 m. As of October 18th the level of the lake had surpassed 1500 m thus we would expect the level to drop around 120 m almost twice the usual value.


    Photo 10 is an image of the ice above Grimsvotn. In the background part of the caldera rim protrudes through the ice. The foreground ice floats on the lake beneath, the threshold of which can be seen in the surface of the ice, as indicated in photo 10.

    Overview of the eruption site


    Photo 11 is an overview shot of the eruption site taken on October 3rd. It shows the eruptive fissure as indicated by the three lines, Grimsfjall behind Grimsvotn, and on the horizon Oraefajokull. The photo is taken from the north. At the southern end of the fissure there is evidence of the subglacial channel which runs from the fissure to the lake. This can be seen as subsidence in the ice due to the melting caused by the eruption heated waters flowing towards the lake.

    12 13

    The development of the fissure and the subglacial channel to Grimsvotn can be seen on satellite images. Images 12 and 13 are from ERS-2 SAR and show the same area of the glacier, north is up. Image 12 shows part of the glacier in September 1996. The dark circular region half way down the image on the left is Bardarbunga (see earthquake event maps above), and the dark diagonal ridge in the bottom right is the southern rim of Grimsfjall with the lake above it. Image 13, from October 6th 1996, shows and additional north-south feature between the two calderas with is the erupting fissure, and a meandering line connecting the fissure with Grimsvotn to the south, this is the depression above the subglacial channel.

    The eruption ceases


    The eruption came to an end on October 12th, 1996 after a gradual decline. Photo 14 was taken several days later and shows the new mountain within the fissure in the ice. The heat from the eruption will continue to melt the glacier adding more volume to Grimsvotn, the caldera lake. Scientists have stopped making predictions as to when the lake will empty resulting in a massive flood. Based on the height of the lake it should have flooded over a week ago, local farmers insist it won't flood until the next thaw however.

    The flood - Jokulhlaup

    15 16 17

    The flood started early morning November 5th as the volume of Skeidara began to increase with a strong smell of sulphur. During the course of the day the volume gradually increased flowing out to the south of the glacier across the flood plain. By mid afternoon one bridge had been destroyed and another, the 900 m Skeidara bridge, was badly damaged. The power and phone lines were destroyed. Icebergs broke off the snout of the glacier and were carried across the flood plain by the flow. These were later left spread across the plain, photos 15, 16 and 17, notice the helicopter, man and road respectively for scale.

    The flood culminated at 2230 November 5th. At the peak an estimated 45 000 cubic m/s flowed from Grimsvotn under the glacier and out onto the flood plain, Skeidararsandur. It was over by early moring November 8th having done an estimated 15 million US$ worth of damage.

    Back to Eruption Index Page
    This page is maintained by Richard M Allen.