Jökull - 01.12.1957, Qupperneq 23
SIGURDUR THORARI NSSON:
The Jökulhlaup from the Katla Area in 1955 compared
with other Jökulhlaups in Iceland
The eruptions o£ the volcano Katla in Mýr-
dalsjökull are known to start always with jökul-
hlaups which at their maximum are really en-
ormous. It has been estimated that the discharge
of the jökulhlaup of Katla (Kötluhlaup) in
1918 at its maximum was between three and
four hundred thousand m3/sec. (Hannesson
1934). This figure is probably somewhat —
hut in my opinion not necessarily very
much — too high. It must also be taken into
consideration that the discharge of a Kötlu-
hlaup may consist to a great part o£ ice, not
only icebergs of all sizes, but also a more fine-
grained ice, that can make the hlaup look like
a sort of sludge. This ice should be subtracted
from the total volume of the hlaup when dis-
cussing the origin and the storage of the hlaup
water. However I regard it as likely, judging
from descriptions of Kötluhlaups, that their
discharge of water (the ice subtracted) may reach
an euphemeral maximum of at least 100 000
m3/sec, i.e. ab. the size of the river Amazon,
or more than twice as much as a Skeidarár-
hlaup at its maximum. We know from Rist’s
measurements (Rist 1955) that tlie maximum
discharge of the last Skeidarárhlaup, in July
1954, was 10 500 m3/sec, or at least 10 times less
than that of a normal Kötluhlaup and the total
quantity of water discharged by this Skeidarár-
hlaup was ab. 3.5 km3. Tlie Skeidarárhlaup in
March 1934 may be regarded as normal both
as regards the drainage discharge pattern and
the total quantity of water discharged. Its ap-
proximate maximum discharge I have estimated
at ab. 45 000 m3/sec, and the total quantity of
water discharged at ab. 7 km3 (Thorarinsson
1953). In my opinion most of the water dis-
charged by a Skeidarárhlaup is stored up sub-
glacially in the Grímsvötn depression between
the eruptions and dammed up by an ice barrier
until the water level has been raised so high
that the water can force its way under the
barrier. In the above mentioned paper I stress-
ed the close resemblance between the discharge
graphs of the jökulhlaups from Grímsvötn and
graphs of drainage of normal ice dammed lakes
such as Lake Grænalón, and I classified Gríms-
vötn as a type of ice-dammed lake. The seismic
soundings within the depression, carried out by
the French-Icelandic expedition in 1951 and
the Icelandic-French expedition 1955, indicate
that the clepression is much deeper and its ice
cover much tliicker than previously assumed
(Holtzscherer 1954, Thorarinsson 1953, 1955).
However, I am still of the opinion that the
drainage mechanism of the hlaups from Gríms-
vötn is essentially the same as at the hlaups
from the ice-dammed lakes. J. W. Glen reminds
of the well known fact that an empty hole in
a glacier will fill in when more than some 15 to
20 m deep, due to the rapid variation of strain
rate with shear stress, and he points out that
a waterfilled hole or lake will tend to enlarge
itself by the same mechanism if it is more than
150 to 200 m deep. This phenomenon he be-
lieves to be responsible for the emptying of
lakes such as Grænalón and Grímsvötn (Glen
1951). This is certainly a theory worth con-
sidering and the enlarging process could explain
a lot in the Grímsvötn area, but I must stress
the fact that ice dammed lakes much shallower
than 150 m, such as lake Gjánúpsvatn at Hof-
fellsjökull, show discharge graphs of the same
type as the deep lakes such as Grænalón, and
these shallow lakes are often completely empti-
ed subglacially by a hlaup, which is not in
favour of Glens hypothesis. Nor does his hypo-
thesis explain, why the ice dammed marginal
lakes sucli as Grænalón and the Vatnsdalur
lake, reflect the thinning of the damming
glacier (Skeidarárjökull and Heinabergsjökull
respectively) in such a manner as they have
21