Institute of Earth Sciences

The Gjálp eruption in Vatnajökull 30/9 - 13/10 1996

Resume

On September 29th, 1996 at 10:48 an earthquake of magnitude 5 on the Richter scale was detected within the Vatnajökull icecap in SE Iceland. This event was followed by an intense earthquake swarm with a large number of small events with intermittent larger quakes of magnitude 3-4 on the Richter scale. The activity continued at similar level until shortly before midnight on September 30th when volcanic tremor was recorded and earthquake shocks became less frequent. In the afternoon of September 29 a warning was issued to national and international aviation authorities that an explosive (phreatic) eruption might be expected within the near future and an ash cloud might interfere with the dense air traffic across the Atlantic.


In the morning of October 1st an over flight discovered a subsidence bowl in the glacier surface at a location where an eruption had occurred in 1938. More or less continuous over flights during the day observed continuous increase in the size and depth of the initial subsidence bowl as three additional bowls formed on a line N30E indicating intensive melting at the base of the glacier along a fissure 5-6 km long.
Simultaneously the ice cover on the 10 km diameter Grimsvötn caldera some 15 km to the south of the active subglacial fissure started to rise indicating that melt water from the eruption was flowing into the caldera depression and lifting its ice cover.

On October 2nd, in the early morning, an over flight observed that the eruption had broken through the ice. Rhythmic explosions resulted in black ash clouds rising to a height of 500 meters while the buoyant eruption column rose to 3000 meters before being deflected by a southerly wind at a velocity of 30-40 knots. Visibility and flying conditions were extremely poor and allowed observation for only a few minutes.

Overview of the eruption site

The eruption fissure is located between two volcanic centres, Grimsvötn and Bardabunga, with a direction parallel to the regional tectonic lineament. The subglacial topography directed meltwater from the erupting fissure toward the Grimsvötn caldera which was rapidly filling. By the evening of October 1st the ice cover above the subglacial lake had risen 10-15 meters. A rise of the water level by additional 35 meters would trigger a flood.
 

Jökulhlaup on Skeiðarársandur from the Grimsvötn caldera

Jökulhlaup started on the morning of November 5th quite abruptly. The growth was fast and many icebergs were floating in the flood water from the glacier reaching all the way to the sea. Grimsvötn tremour started at 21:30 on Monday November 4th, the jökulhlaup came out of the glacier 11 hrs later. Distance is about 50 km from Grimsvötn to the Skeidararsandur. The discharge growth was more sudden than was ever expected to be possible. While all was quiet at 0800 the flood had increase its discharge 80-100 times in less than 2 hours.

The estimated peak-flow was around 45.000 m3/s, that is about 10.000 to 15.000 m3/s higher than in the jökulhlaups of 1934 and 1938. Flood peak was reached at about 2300 hrs.  The flood channel from Grimsvötn caldera could be clearly seen. It formed a depression like structure on the surface of the glacier dotted with several holes. It is estimated that the Grimsvötn lake was totally emptied, since the glacier seal had been destroyed due to melting of the water. This had never occurred in Grimsvötn before. Such high melting can only be associated with lake water temperature greater or equal to 10 deg. celsius.

A small eruption started in Bardarbunga around 1300 hrs on November 6th. The eruption lasted for about 20 to 30 min. According to seismograms at the Meteological office, the eruption was initiated by some intrusive activity. The intrusive activity is based on recorded eruption tremor picked up by on of the seismometers. Eruption column reached about 4 km in to the air. Relation between pressure decrease due to the flooding have been suggested as the main cause of the eruption.