Jökull - 01.12.1984, Qupperneq 9

the magma before eruption. When this is done it
can be seen for example that FeO and Ti02 have
a tendency to increase with decreasing MgO. In
this case the variation observed is likely to be
real.
In each sample about ten grains were analyzed
and the average of these was then used to re-
present the sample in Table 1. In some of the
samples variation between individual grain analy-
ses was greater than in other but since this is so
close to the precision limit it is not, at this stage,
possible to decide whether it is random error of
the machine or real variation within individual
samples as suggested for the samples as a whole.
The conclusion therefore is that the glass phase
of the ash erupted has a fairly uniform chemical
composition but minor variation near the limit of
the analytical method may be present. Variation
in the amount of crystals is not reflected in the
composition of the accompanying glass.
The 1934 eruption. In the ten samples from this
eruption the glass phase was similarly analyzed
and with very similar results (Table 1). The
samples from the tephra fall of April lst 1934
show a very small variation which for all the
elements is within the precision limit of the
method and gives some indication of the possible
precision for samples analyzed during one
period.
The result is therefore that the composition is
similar as that of the 1983 ash and minor variation
may be present between the samples.
Analysis of the glass phase of ash from the 1934
eruption from a drill core in Bárdarbunga (Stein-
thórsson 1977) is also listed in Table 1. The
apparent difference between that and the other
samples is not significant as it was analyzed under
different conditions.
A number of whole rock analyses are available
from the 1934 tephra and other eruptions. One of
them, SAL 51 (Steinthórsson 1977), used the
same sample as 0018, but since that sample is
dominantly black tachylitic glass only a limited
number of grains could be analyzed. The similar-
ities of the glass and the whole rock support the
observation that the crystal content of the sam-
ples is small.
The 1922 eruption. The two samples from the
1922 eruption have an identical composition
although sample 0007 has significantly higher
crystal content. This composition is very similar
to that of the two later eruptions.
The 1903 eruption. This one sample has similar
composition as the other samples but appears to
be slightly less evolved.
In two of the samples, 0025 from the 1934
eruption and 0008 from the 1922 eruption, one
grain was found in each that had markedly less
evolved composition from the rest. The signifi-
cance of this is very uncertain since these could
easily be xenoglasses incorporated during the
eruption.
DISCUSSION
Detailed knowledge of earlier eruptions is
scarce but the behaviour of the 1983 eruption
seems to be similar in most respects. No jökul-
hlaup accompanied the eruption like most other
known eruptions but similar behaviour has been
noted previously. The pattern of small eruptions
at short time intervals, nine this century, seems to
be characteristic of Grímsvötn. The chemical
composition of the products of this eruption also
seems to be very similar to that of earlier erup-
tions, from this century at least.
The basalt type erupted in all the eruptions
discussed here is a quartz normative tholeiite
similar to that found to dominate other central
volcanoes. It is highly evolved and very different
from any basalt that is likely to be derived from
the mantle below Iceland. The presence of three
mineral phases and low crystal content suggests a
final evolution of the magma at relatively low
pressures (O’Hara 1982). Significant evolution
from a more primitive original liquid must there-
fore take place during ascent through the crust.
The frequent sampling of this magmatic system
is specially interesting when considered in rela-
tion to the very high geothermal energy output of
the area. This energy, estimated at 5000 MW,
must be of magmatic origin and it is estimated
that this corresponds to the cooling of 50 million
m3 of magma during the last 120 years at least
(Björnsson 1983). Estimates of the amount of the
magma involved, although uncertain, suggest
that 89% cools as intrusions, 8% deposit in the
lake and 3% form the tephra layers (Björnsson
1983). These observations put serious constraints
on possible evolution processes within the
magmatic system.
The most realistic model for the geothermal
system in Grímsvötn (Björnsson et al. 1982,
JÖKULL 34. ÁR 7