Jökull - 01.12.1984, Qupperneq 22

Fig. 7. Photograph of the eruption site taken on May 29, at 16 h, from an aeroplane. The eruption pit
is under the southern caldera rim. The thin ash fan is visible to the south and the apron of tephra and
ice on the plane to the northwest of the pit. According to the seismographs the eruption is in progress,
but only occasional explosion was seen to break the water surface in the pit.
Photo Páll Einarsson.
7. mynd. Gosstöðvarnar séðar úr lofti um kl. 16, 29. maí. Gígurinn er við suðurjaðar öskjunnar, beint
undir Vestari Svíahnúk, enfrá honum má sjá þunna öskudreif til suðurs og gjóskuslettu á ísbreiðunni til
norðvesturs. Samkvœmt óróa á skjálftamælum var gos í gangi þegar þessi mynd var tekin, en ekki sáust
þó nema stakar sprengingar rjúfa vatnsyfirborðið.
ociated with an inflating magma chamber. This
interpretation is strengthened by the fact that the
earthquakes ceased when the eruption began.
Similar cessation of seismic activity has followed
every deflation event of the Krafla volcano, and
reflects the lowered stress in the crustal rocks
around the magma chamber, when magma press-
ure decreases. Assuming that the epicenters in
Fig. 2 are correctly located, the magma chamber
is situated under the SE rim of the main caldera.
Whether this chamber is the same as that pre-
sumed to feed the geothermal system in Gríms-
vötn (Björnsson et al. 1982), a subdivision of it,
or an independent reservoir, we cannot say.
The earthquake swarm on May 28 probably
reflects the process of fracturing of the chamber
walls and subsequent propagation of magma
away from the reservoir. The propagation must
have been mostly vertical, since the epicentral
area coincides with that of^the inflation earthqu-
akes. Some lateral propagation must also have
taken place, however, if the distance between the
epicentral area and the eruption site seen in Fig.
3 is real. Earthquake swarms have preceded all
recent eruptions in Iceland. The time from the
beginning of the swarm to the outbreak of the
eruption is variable, from about 25 minutes in the
case of Hekla in 1980 (Grönvold et al. 1983) to 30
hours in Heimaey in 1973 (Einarsson and Björns-
son 1979). In all cases the earthquakes have
diminished or ceased when the magma reached
the surface. If this also holds for Grímsvötn, the
20 JÖKULL 34. ÁR