The main tephra sector formed November 1-3 trends to north-northeast as a result of strong
southerly winds. The sector is about 30 km wide near the north edge of Vatnajokull at a
distance of 50 km from the eruptive site. Tephra fell in inhabited areas in north and northeast
Iceland, but only in small quantities. The eruption plume was seen on satellite images and ash
drifted over large parts of the North Atlantic and reached Scandinavia. Air traffic was
disrupted; an area of 311 thousand square kilometres was closed for flights from the
beginning of the eruption until the morning of November 4. Farmers sheltered grazing animals
in North Iceland to prevent them from consuming soluble fluorine adhering to ash grains.
On November 4, winds changed to north-westerly direction and in the afternoon tephra had
been dispersed over the ice cap east of Grimsvotn.

After the onset of the eruption insignificant earthquake activity occurred at the eruptive site,
but continuous low-frequency tremor has been recorded during the eruption. The tremor
was steady for the initial 15-hours of the eruption. After that it was pulsating and declining.

The jokulhlaup, the glacial outburst flood that preceded the eruption by few days and
triggered the eruption (Sigmundsson et al., 2004), reached a maximum in the afternoon of
November 2. At that time the peak discharge in the rivers on Skeidararsandur was 3000-
4000 m3/s (based on information from the Icelandic Hydrological Service). Discharge
declined fast after the peak. No damage has occurred to roads or bridges. The total volume
of the jokulhlaup is about 0.5 km3.

The eruption follows a pattern similar to previous eruptions in 1983 and 1998, with probably
less than 0.1 km3 of magma erupted. These eruptions, together with the 1996 Gjalp eruption
north of Grimsvotn reveal much higher activity at Grimsvotn than during the middle part of last
century, and may indicate that Grimsvotn is entering into a new period of high volcanic activity
that may last for decades. Such a high activity period has been predicted on the basis of the
observed cyclic volcanic activity in the area in the preceding millennium (Larsen et al., 1998).
From NORVOL, Iceland, authors: Freysteinn Sigmundsson, Pall Einarsson, Magnus Tumi Gudmundsson, Thordis Hognadottir, Anette K. Mortensen, Institute of Earth Sciences,
University of Iceland, Reykjavik, Iceland.
Steinunn Jakobsdottir, Matthew Roberts, Kristin Vogfjord, Ragnar Stefansson,
Icelandic Meteorological Office, Reykjavik, Iceland.

Release of overburden pressure triggers an eruption of Grimsvotn volcano, Iceland.

The subglacial Grimsvotn volcano, Iceland, started erupting on November 1, 2004 around
22 GMT. An intense swarm of volcanic earthquakes that started about 3 hours earlier
changed at that time to continuous low frequency tremor, indicating onset of an eruption.
Weather conditions prohibited direct observations of the beginning of the eruption at this
remote volcano situated near the center of Europe's largest ice cap, Vatnajokull.

In the morning of November 2nd Photo: © Magnús Tumi Guðmundsson, www.jardvis.hi.is.

The eruption was preceded by both long-term and short-term precursors, and finally
triggered by release of overburden pressure associated with a glacial outburst flood
(jokulhlaup), originating from the Grimsvotn subglacial caldera lake, that preceded the
eruption. Accumulation of magma in a shallow magma chamber under the Grimsvotn caldera
has been ongoing since its last eruption in 1998 (Sturkell et al., 2003; Sigmundsson et al.,
2004). GPS measurements show uplift of 5-10 cm/year in the caldera center, and horizontal
displacements away from the caldera. Earthquake activity increased in middle of 2003, at
about the same time uplift exceeded its 1998 maximum. Pressure in the Grimsvotn magma
chamber is likely to have exceeded its pre-eruption level from 1998 at this time.Additional
uplift and expansion of the volcano since then suggested approaching failure of the volcano.

In the morning of November 4th Photo: © Fredrik Holm, www.norvol.hi.is

Earthquake activity increased further in late October, 2004. Geothermal heat sustains a lake
in the caldera that intermittently causes glacial outburst floods. On October 26 high frequency
seismic tremor indicated increased water flow from the caldera lake and suggested that a
glacial outburst flood was about to begin. On October 29 discharge increased in river
Skeidara. The outburst flood was caused by high water level in the Grimsvotn caldera lake
from ice melting by geothermal activity. The release in overburden pressure associated with
the outburst flood triggered the eruption. The drop in water level in the Grimsvotn caldera at
the onset of the eruption is uncertain, but is probably on the order of 10-20 meters,
corresponding to a pressure change of 0.1-0.2 MPa on the volcano surface. This modest
pressure change triggered the eruption because internal pressure in the Grimsvotn shallow
magma chamber was high after continuous inflow of magma to the volcano since 1998.

From NORVOL, Iceland, authors: Freysteinn Sigmundsson, Pall Einarsson, Magnus Tumi Gudmundsson, Thordis Hognadottir, Anette K. Mortensen, Institute of Earth Sciences,
University of Iceland, Reykjavik, Iceland.
Steinunn Jakobsdottir, Matthew Roberts, Kristin Vogfjord, Ragnar Stefansson,
Icelandic Meteorological Office, Reykjavik, Iceland.