Jökull - 01.12.1984, Qupperneq 29

jökulhlaups from Grímsvötn. The first report of a
jökulhlaup dates back to A.D. 1332. From 1600
until 1934 there occurred about one jökulhlaup
per decade with an estimated discharge of 6-7
km3 of water and a maximum discharge rate of
approximately 40 000 m3/s (for example in 1903,
1913,1922,1934). But since 1934 there have been
two, even three, bursts per decade with corres-
pondingly smaller volumes (except for 1938),
1-3.5 km3, and maximum discharge rates of
5-8000 m3/s (1938,1939,1941,1945,1948,1954,
1960,1965,1972,1976,1982 and 1983; Thorarins-
son 1974, Rist 1955,1973,1976, Björnsson 1983).
In December 1983 the smallest jökulhlaup ever
was observed with a total volume of 0.5 km3
(Sigurjón Rist, pers. comm.).
The Grímsvötn lake owes its existence to the
geothermal area. Melting due to the geothermal
activity creates a depression in the surface of the
Vatnajökull ice cap. Ice and water are diverted
towards the depression from a 300 km2 drainage
basin (Fig.l). The meltwater accumulates in a 25
km2 subglacial lake, which is covered by a 200 m
thick floating ice shelf (Thorarinsson 1953b,
1974, Björnsson 1974). The lake is sealed and no
water drains out of it between the jökulhlaups.
Water accumulates in the lake and when the
water level has risen to a critical level, water is
forced out of the lake beneath a barrier east of
the lake (Fig.2). The subglacial waterways are
enlarged by frictional melting and the lake is
drained in a fortnight by a catastrophic flood.
Eventually, the ice overburden pressure is able to
close the water tunnels and the flood stops
abruptly before the lake is empty (Björnsson
1974, Nye 1976). The water-level change in the
larger jökulhlaups has been estimated to be
about 150 m, but is observed to be 80-100 m
during the last three decades (Fig.3). The fre-
quency and water volume are believed to depend
on the thickness of the glacier. This model is
considered to explain most of the jökulhlaups.
The jökulhlaup in December 1983, however,
occurred at a water level that was 20-30 m below
the present critical level for triggering jökul-
hlaups. Explanation of this will be discussed in
the paper.
The heat output of the subglacial geothermal
area has been estimated by using the Grímsvötn
lake as a natural calorimeter. A long term mass
balance of the drainage basin has been given by
Björnsson (1974). The average accumulation in
Fig. 2. The Grímsvötn area. Contour map of
the Icelandic Geodetic Survey from 1946 on
which we have marked the position of craters in
1934 and 1983, the depression north of Svarti-
bunki from 1938, and ice cauldrons. The
cauldron that formed during the jökulhlaup in
1983 is marked on the map. The cauldron north-
west of Grímsvötn drains to the river Skaftá.
2. mynd. Grímsvatnasvœðið. Kortið sýnir hœð-
arlínur frá 1946, en á það eru merktir sigkatlar og
gígarnir frá gosunum 1934 og 1983. Rennan
norðan við Svartabunka myndaðist 1938 og sést
enn árið 1946. Sigketillinn, sem myndaðist við
hlaupið í desember 1983 er merktur á kortið.
Sigketillinn norðvestan við Grímsvötn hleypir
vatni í Skaftá.
the form of ice is equivalent to 2200 mm /yr of
water and the surface ablation amounts to 500
mm /yr. A long-term steady-state model for the
drainage basin proves to be a valid approxima-
tion (Björnsson 1974). The water added to the
lake is 6.6T011 kg/yr. About 1.5T011 kg/yr are
melted at the surface of the glacier by
meteorological processes, but the difference,
about 5T011 kg/yr is melted by the geothermal
heat within the drainage basin. The heat flux
required to melt this ice is about 5000 MW (ther-
mal). In the present work we estimate the geoth-
ermal mass fraction in the lake and present new
estimates of the thermal power of the geothermal
system.
JÖKULL 34. ÁR 27