Jökull - 01.12.1984, Qupperneq 10

Björnsson 1982, 1983) assumes that heat extrac-
tion from the magma takes place by penetration
of water into cooling magma, most likely in high
level intrusions. The apparent recent shift in
geothermal activity from Grímsvötn towards the
west would best be explained by a shift in intrus-
ive activity (Björnsson 1983). For a magmatic
system with a 10 km2 areal extent the solidifica-
tion front would progress at 5 m/yr.
Further complications arise with attempts to
relate Grímsvötn to the surrounding area and
other eruptions. Comparison with ice free volca-
nic areas and satellite photographs suggest that
the Grímsvötn volcano is a part of an elongated
volcanic system consisting of fissures and hyaloc-
lastite ridges similar to other such systems in
Iceland (Jakobsson 1979). Suggestions about the
extent of the Grímsvötn system vary. The erup-
tion of the Laki craters in 1783—84 may have
been caused by lateral magma flow from reser-
voirs below Grímsvötn (Saemundsson 1978,
Sigurdsson and Sparks 1978) which would extend
the system about 70 kilometres to the south-west
(Figure 1). This suggestion is supported by
Jakobsson (1979) who includes Thórdarhyrna
volcano in this system. Larsen (1982) suggested
on the basis of chemical similarities that Gríms-
vötn and Kverkfjöll, a central volcano at the
northern margin of the ice sheet belong to the
same volcanic system while Saemundsson (1978)
and Jóhannesson (1984) show them as three sepa-
rate systems, Grímsvötn, Kverkfjöll and Thór-
darhyrna (Figure 1). Recently a number of sam-
ples from the Lakagígar eruption of 1783—1784
were analyzed in a similar way as the Grímsvötn
samples (Grönvold 1984). The result of the glass
analyses show a very homogeneous chemical
composition and an average of these is included
in Table 1 for comparison. The chemical com-
position of the glass phase is identical with that of
Grímsvötn which strongly supports the sugges-
tion of a single magmatic system (Jakobsson
1979).
The volume of the lava from the Laki craters,
about 12 km3, corresponds to a 200 year heat
supply of the Grímsvötn geothermal area. No
significant effects were noted in the frequency of
eruptions and jökulhlaups from Grímsvötn after
the Lakagígar eruption (Thórarinsson 1974). This
must either mean a very quick recovery of a high
level magma system or that the behaviour of the
volcanic system, which may include Grímsvötn
and the fissures swarm to the south, is more
complicated. One complication would be that the
Lakagígar eruption was an addition to the “norm-
al“ magmatic activity of this volcanic system
rather than an eruption simply brought about by
a rifting event.
The chemical composition of the magmatic
intrusions that maintain the geothermal system is
not inevitably the same as that of the relatively
minor eruption products. It has even been sug-
gested that a more primitive magma rises below
Grímsvötn where it partly crystallizes to feed the
geothermal system but then the liquid convects
down again (Tryggvason 1982) moving the space
problem to a deeper level. Another possibility
could be a density trap (Stolper and Walker
1980). The more primitive magma would then be
denser and trapped as intrusions while a more
evolved magma escapes to the surface in erup-
tions. The apparent homogeneity of the ash,
however, seems to argue against both these possi-
bilities as neither would result in the strict control
of the chemical composition observed.
The observations of this eruption, considered
in the context of the general behaviour and
character of the area, show a magmatic system
capable of producing continuously large volumes
of evolved magmatic liquids of uniform chemical
composition. The evolution processes of this
magma seem to take place at deeper levels than
the heat extraction that maintains the geothermal
system, and the buffering capacity that maintains
this evolved composition is not exceeded by large
eruptions like that of the Laki craters 1783-84.
8 JÖKULL 34. ÁR