Jökull - 01.12.1991, Qupperneq 42

Although this apparent difference in eruption fre-
quency may, at least partly, be due to incomplete
records, it suggests that a more conservative estimate
of the total volume erupted should be made. If the
production rate prior to 1600 was, say, one fourth to
half that observed at present, the estimate of the total
volume erupted in historical times would be 3-5 km3.
The conclusions reached in this paper are consis-
tent with the interpretation of internal reflections
acquired with a multiband synthetic pulse radio-echo
sounder in the Grímsvötn area in 1991 (Bjömsson et
al., in preparation). The internal reflectors are inter-
preted as being tephra layers within the ice. Sounding
profiles in the eastern part of the Grímsvötn area
show four tephra layers above the depth of 200
metres. The two uppermost layers are considered to
be from the eruptions of 1934 and 1922, on the ice
divide to the east of Grímsvötn, the depth to these
layers is 130 m and 150 m, respectively. In the cen-
tral part of the Grímsvötn ice shelf the depth to the
layer of 1934 is reduced to 80 m. No tephra layers
were detected above the layer of 1934, supporting
our conclusion that no eruptions occurred within
Grímsvötn between 1934 and 1983.
Temporary existence of an ablation area in the
Grímsvötn depression would complicate the use of
tephra layers for estimating glacier mass balance
from ice cores and radio-echo soundings. Conditions
similar to those observed during the period from
1934 to the 1950's may have been frequent in
Grímsvötn in previous centuries. Such conditions
may also be expected in the neighbourhood of the
most active subglacial volcanoes in Iceland, like
Kverkfjöll and Bárðarbunga in Vatnajökull, and
Katla in Mýrdalsjökull. However, the size of ablation
areas of this type is probably always very limited. In
most parts of the accumulation areas of the ice caps
the dispersed ash layers only affect the mass balance
of one year as they are buried permanently by the
snowfall of the next winter (cfr. results of radio-echo
soundings discussed above). Such conditions have
existed, at least over the last few hundred years, south
of Bárðarbunga, NW-Vatnajökull, where a 415 m
long ice core was recovered in 1972 (Steinþórsson,
1977). No evidence was found for a discontinuity in
the tephrachronological record.
CONCLUSIONS
Firm evidence can only be found for three eruptions
within or near Grímsvötn in the period 1934-1991, i.e.
in 1934, 1938 and 1983. In addition, seismic data
alone indicate a small subglacial eruption in 1984.
Data on these events are summarized in Table 1. An
eruption at the end of the jökulhlaup of 1954 can be
discarded as no crater is situated at the proposed erup-
tion site. Further, aerial observations indicate that
eruptions did not occur during the jökulhlaups of 1941
and 1948. The exact location of the sinkhole forrned
in 1945 is not known and consequently the existence
or absence of a crater in the bedrock cannot be estab-
lished. However, we consider a steam explosion or
increased upwelling in a geothermal upflow zone the
most likely explanation for the sinkhole of 1945.
An ash cover that was observed in the northwest-
ern part of Grímsvötn in the period 1934-1960 dates
back to the eruption of 1934 but may contain ash dis-
persed in earlier eruptions. During this period (since
1934 and possibly longer) the ash cover was exposed
every year in late summer as the winter accumulation
was melted, and this part of the Grímsvötn depres-
sion was an ablation area. Moreover, the frequent
small jökulhlaups from Grímsvötn in the period
1938-1948 are considered to have been caused by
melting at the site of the eruption of 1938, north of
Grímsvötn. Consequently, both the ash layer and the
jökulhlaups in the 1940’s can be adequately
explained without assuming eruptions after 1938.
ACKNOWLEDGEMENTS
We would like to thank Haukur Jóhannesson and an
anonymous reviewer for constructive criticism of
the manuscript. Assistance from the Iceland Geodetic
Survey in locating air photos from Grímsvötn is
gratefully acknowledged as well as grants from the
Research Fund of the University of Iceland and the
Icelandic Road Authority. Members of the Iceland
Glaciological Society assisted in the fieldwork.
40 JÖKULL, No. 41, 1991