Jökull - 01.12.1984, Qupperneq 15

Seismic Activity preceding and during the 1983
Volcanic Eruption in Grímsvötn, Iceland
PÁLL EINARSSON AND BRYNDÍS BRANDSDÓTTIR
Science Institute, University of Iceland
ABSTRACT
The eruption of Grímsvötn in May—June 1983
was preceded by increased earthquake activity.
The premonitory activity began in December 1982
— March 1983 and increased gradually in April
and May. The earthquakes originated under the
SE rim of the Grímsvötn caldera and are inter-
preted as being caused by brittle failure of the crust
above and around an ínflating magma chamber.
An intense earthquake swarm (Mmax = 4.0) occur-
red in the same area on May 28, presumably
related to the failure of the magma chamber walls
and subsequent migration of magma towards the
surface. The eruption is inferred to have begun
shortly after the swarm ceased or about the time,
when continuous volcanic tremor was first
recorded on the nearest seismograph between
ll:47h and 12:00h on May 28. The trernor was
most intense in the first 12 hours and then gra-
dually diminished until it disappeared early on
June 2. It came in bursts of several minutes dura-
tion, separated by longer periods ofmore uniform
background tremor. If the tremor amplitude is
taken as an indicator of extrusion rate, the erup-
tion was most vigorous during the first 26 hours.
The several km high eruption columns observed in
the last two days of the eruption thus probably
reflect the reduced water pressure on the magma,
when the volcanic orifice approached the surface
of the caldera lake, rather than high extrusion rate.
Earthquake activity was very low in Grímsvötn
during the eruption and remained so for three
months afterwards, probably indicating the
relaxed stress state around the deflated magma
chamber. The 1983 eruption demonstrates that
small eruptions can occur in Grímsvötn without
causing or being triggered by jökulhlaups, and
without being noticed from the inhabited low-
lands. Eruptions are, however, likely to be
detected by the seismograph network in SE-lce-
land. The absence of any abnormal seismic activ-
ity in Grímsvötn prior to 1983 and since at least
1971 indicates that no eruptions occurred during
this time, in particular not during the jökulhlaups
of 1972, 1976 and 1982.
INTRODUCTION
The eruption of May 28 — June 2, 1983 is the
first confirmed eruption in Grímsvötn after the
Iarge increase in seismograph coverage of Iceland
that began in the early seventies. The eruption
was preceded by a significant increase in seismic
activity for about three months and an intense
earthquake swarm on the day of the presumed
outbreak. Characteristic volcanic tremor was also
recorded during the eruption. In this paper we
document the seismic observations and discuss
them in context with the ”Grímsvötn problem",
i.e. the intriguing interplay between volcanic,
geothermal and glacial processes at work in
Grímsvötn. The direct observations of the erup-
tion are summarized by Grönvold and Jóhannes-
son (1984) who also present chemical analyses of
the eruptive products and discuss their con-
straints concerning the magma system. Geoth-
ermal and glaciological aspects are discussed by
11. Björnsson and H. Kristmannsdóttir (1984).
The Grímsvötn volcano with its 6—8 km wide
caldera is situated beneath the Vatnajökull ice
sheet, and is one of a group of central volcanoes
that appears to characterize the structure of this
part of the volcanic rift zone. A powerful geoth-
ermal area, presumably fed by a shallow magma
chamber (Björnsson et al. 1982), makes Gríms-
vötn unique among these subglacial volcanoes.
The geothermal heat in the caldera melts the
glacier ice from below and forms a caldera lake
dammed by the ice. The lake level rises steadily
until a critical level is reached. The water is then
released under the glacier, the lake is partly
drained and a burst of water up to 6—7 km3 in
total volume issues from the glacier, flooding the
JÖKULL 34. ÁR 13