Jökull - 01.12.1983, Síða 19

years. This may be due to lack of intrusions of
magma to shallow levels so the magma body has not
got a recharge. The solidification front of the magma
body sinks deeper and deeper and the heat ex-
change is slowly retarded. Second, peaks in the heat
flux may amount to a rise above the base flux by a
factor of two to three. They are presumably due to
intrusion of magma to the glacier bed.
It is interesting to note that at the same time as
the heat flux declines at Grímsvötn a comparable
increase in heat output has been observed beneath
the cauldrons 10 km northwest of Grímsvötn. We
can speculate if the recharge to Grímsvötn went
amiss into a magma body in the area beneath the
cauldrons.
The heat energy removed by volcanic eruptions
at Grímsvötn is an order ofmagnitude less than the
heat which is considered to be drawn from a deeper
sited magma body with the aid of hydrothermal
convection. On the average, the magma which soli-
difies beneath Grímsvötn equals the annual rate at
which magma has been erupted to the surface of
Iceland in historical times (Thorarinsson 1976).
A NOTE ADDED IN PROOFS
Since this paper was submitted new information
has become available about jökulhlaups and erup-
tions in the Grímsvötn area. This does not alter the
conclusions of the paper but is added here to
update the history of the caldera.
1. In February 1982 a jökulhlaup was discharged
from Grímsvötn. The total water discharged
was only 1.3 km3 (Sigurjón Rist, personal com-
munication 1982) and the accompanying Ta.ll of
the lake level was only 50 m. This does not
mean that the accumulation ofwater in Gríms-
vötn was abnormally low in the 5.3 years
between the jökulhlaups of 1976 and 1982, as
the lake level in 1982 had risen to the critical
level required to trigger ajökulhlaup. Sofar this
exceptional event has not been explained but it
emphasizes the caution taken in interpretating
the variations shown in Fig. 2.
2. On the 28th or 29th May 1938 an eruption
started in Grímsvötn and was active for five or
six days. The eruption was not observed from
outside the ice cap and the question arises how
many small eruptions in Grímsvötn could have
been overlooked in the past. A possible example
of this is provided by Haukur Jóhannesson (in
preparation, personal communication 1982,
1983) who claims that a small eruption occurred
in the Grímsvötn area in November-Desember in
1933. The eruption in March-April in 1934 may
have been the continuation of this event. Further,
we can speculate whether craters observed in
Grímsvötn in 1945 (Áskelsson 1959) and 1954
(Thorarinsson 1974, figs. 54 and 55, p. 188, 191)
were caused by volcanic eruptions rather than
steam explosions.
The eruption in May 1983 cannot have transport-
ed more than 1— 2X106 m3 of lava into Grímsvötn.
This is only 2—4% of the total annual inflow of
magma to the Grímsvötn area.
I do not expect this additional information to
significantly alter the calculations of the rates of
magma extrusion at the glacier surface, transport at
the glacier base and solification in the upper crust.
SKAFTÁRHLAUPS.
Fig. 4. Volume ofwater in jökulhlaups in the river
Skaftá (Rist 1974—80, personal communication
1983). Computed rate of melting by the subglacial
heat source beneath the caldrons NW ofGrímsvötn.
4. mynd. Rúmmál vatns í Skaftárhlaupum skv. mielingum
Sigurjóns Rists. Stöplaritið sýnir mat á ísbráðnun vegna
jarðhita milli hlaupa og afl varmagjafans.
JÖKULL 33. ÁR 17