Jökull - 01.12.1991, Qupperneq 28

Fig. 5. Cross-section of the eruption sites of 1934 and
1983 showing the mounds formed in the eruptions.
For location see Fig. 2. Þversnið gegnum
gosstöðvarnar frá 1934 og 1983. Kollarnir sem
mynduðust í gosunum eru merktir með ártali.
1983 can be estimated from the size of the mounds
formed and the tephra that was spread over the glac-
ier surface. It may, however, be an underestimate as
some of the erupted material may have been deposit-
ed on the lakefloor as lahar beds. Moreover, reflec-
tion seismics and magnetics indicate that parts of the
lake floor are covered with lava flows which must
have formed under water (Guðmundsson, 1989). The
volume of the mound of the main cater in 1934 is 15-
20-106 m3 while the volume of material deposited at
the other two sites was less, probably no greater than
5-106 m3. Þórarinsson (1974) estimated the total vol-
ume of tephra in 1934 as 10-20-106 m3. Therefore,
the minimum estimate of the volume produced in the
eruption of 1934 is 30-40-106 m3.
The hillock formed in 1983 has a volume of 6-
8-106 m3 and the amount of ash erupted was minute
compared to 1934 (Grönvold and Jóhannesson,
1984). Hence, the total volume produced by the
eruption of 1983 was not greater than 10-106 m3.
Caldera Opening in ice shelf
Fig. 6. A schematic cross-section showing the main
features of the eruptions of 1934 and 1983. A dashed
line shows the form of the ice shelf before the
eruption. Aðstœður við dæmigert gos innan
Grímsvatnaöskjunnar. Gosið bræðir gat í íshelluna og
upp hleðst gígur.
THE ERUPTIONIN1938
A large unexpected jökulhlaup occurred on May 23 -
June 6, 1938, only 4 years after the large jökulhlaup
in 1934 (Þórarinsson, 1974; Björnsson, 1988).
Usually the Grímsvötn jökulhlaups rise for 7-15 days
and then recede in 1-3 days (Bjömsson, 1988). The
1938 jökulhlaup reached a maximum in 3 days and
receded slowly for two weeks. The reason for this
was that a large area to the north of Grímsvötn sub-
sided and a large volume of water was drained into
Grímsvötn causing a sudden rise in the lake level,
sparking off the jökulhlaup (Bjömsson, 1988) (Figs.
1, 7 and 9a). In a reconnaissance flight on May 28,
1938, no tephra was seen on the surface of the glaci-
er within or near the subsided area. The depression
trended towards the north for 9 km from the north-
west corner of Grímsvötn. The area could be divided
into two parts. The southern part was linear while the
northern part, which was the wider of the two, was
elliptic with the long axis striking NNE. The border
area between the two was narrower and the subsi-
26 JÖKULL,No. 41, 1991