Jökull - 01.12.1990, Qupperneq 160

son, 1988, p. 103). Further, a 10-15 km long row of
mounts extends NE from the southeastern comer of
Grímsvötn. The peaks, some of which exceed 1300
m in elevation, probably represent eruption sites on a
fissure swarm. This feature, however, does not reach
to the Kverkfjöll mountain massif. The topographic
data therefore do not indicate a structural connection
between the Grímsvötn and the Kverkfjöll volcanic
systems.
The broad highland on the western and southern
border of the Grímsvötn caldera continues to the SW
to Háabunga (c.1600 m), and the peaks Þórðarhyma
(1660 m), Geirvörtur, Hágöngur, and Pálsfjall (1332
m, Fig. 4). Farther to the SW the Rauðhólar fissures
emerge from the glacier. The subglacial topography
may indicate that Háabunga, Þórðarhyrna, Pálsfjall
and the Rauðhólar and Laki fissure swarms all belong
to the same volcanic system. Þórðarhyrna is probably
a separate volcanic centre within this system.
Most of the earthquakes at Grímsvötn shown in
Fig. 5 belong to the active period in 1983 - 1984,
when one and possibly two eruptions occurred within
the caldera (see below). The epicentres were mostly
located slightly SE of the caldera. Seismically the
Grímsvötn volcano does not seem to be particularly
active, at least when it is not in an eruptive state. A few
seismic events have been located near Þórðarhyma,
indicating a separate volcanic centre in the area.
The Grímsvötn volcano is notorious for its be-
haviourwithrespect tojökulhlaups (Björnsson, 1988).
The extensive geothermal activity within the caldera
continuously melts the ice and produces a subglacial
lake. When the lake reaches a critical level determined
by the glacier thickness in the caldera breach, water
is drained out of the lake in a large jökulhlaup. The
lake level drops, the glacier blocks the water flow,
and the cycle starts again. Thus jökulhlaups occur
at semi-regular intervals regardless of eruptive activ-
ity. Eruptions on the lake bottom increase the melting
rate of the floating ice shelf but do not affect the lake
level directly because of the lag in the flow of ice into
the lake from the surrounding glacier. They do not
necessarily lead to immediate jökulhlaups, assuming
that the eruption does not disturb the normal trigger-
ing mechanism and that the volume of the volcanic
materials erupted into the lake is small compared to
the lake volume. Eruptions in the areas surrounding
the lake can, on the other hand, lead to an instanta-
neous rise of the lake level, which may in turn result
in a sudden and unexpected jökulhlaup. An erup-
tion or increase in geothermal activity near the caldera
breach may also alter conditions for the draining of the
lake, and thus affect the pattem of jökulhlaups. This
appears to have happened in 1983. Following the
eruption in the spring of that year, new cauldrons were
formed in the ice near the caldera breach. The next
jökulhlaup, later in the year, was released at a lower
water level than known before. Jökulhlaups from the
caldera lake dump into the rivers Skeiðará and Súla. It
is worth pointing out that the large calderas of Gríms-
vötn, Bárðarbunga and Kverkfjöll are practically the
only areas of Vatnajökull where a substantial eruption
can occur without producing a jökulhlaup.
Most eruptions in Grímsvötn have occurred at the
same time as jökulhlaups. Þórarinsson (1953) pointed
out that several eruptions have been observed at the
end of jökulhlaups when the lake level has fallen, typ-
ically by the order of 100 m. He suggested that the
eruptions might be triggered by the pressure release
above the volcano. Hence, the eruptions were trig-
gered by the jökulhlaups. Alternatively, the eruptions
might have been going on during the jökulhlaup but
with the plumes first becoming visible from afar when
the crater emerged as the lake level sank.
Eruptions on the external sides of the Grímsvötn
volcano, i.e. outside the caldera, may lead to floods in
various rivers. Eruptions on the S flank would drain
to Súla and Skeiðará whereas meltwater produced on
the W flank may cause floods in Skaftá. Eruptions on
the N flank and on the ridge NE of the volcano will
release water into Jökulsá á Fjöllum.
Eruptions on the E flank of Háabunga and
Þórðarhyrna would release water to Súla. Núpsvötn,
Djúpá and Hverfisfljót may receive meltwater from
eruptions in the Þórðarhyrna-Geirvörtur-Hágöngur-
Pálsfjall area. The ice cauldron 1 km E of Pálsfjall in-
dicates subglacial geothermal activity in that area. The
drainage of meltwater from beneath the ice cauldron
has not been studied but watercourses are predicted to
run towards Hverfisfljót. No jökulhlaups that can be
156 JÖKULL, No. 40, 1990