Jökull - 01.12.1974, Qupperneq 38

limiting factor until the reservoir is almost
empty, then the supply of water limits the flow.
Liestöl’s theory does not explain the beginning
of the jökulhlaup. The buoyancy or Glen’s
theory could therefore be valid for the beginn-
ing of a jökulhlaup while Liestöl’s theory ex-
plains very well the subsequent process. Still
the two first named theories are too specific for
explaining the beginning of all jökulhlaups.
In search for a more general theory one has
to pay attention to the permeability of glaciers.
Some authors state that glacier ice is imperme-
able (Lliboutry, 1971), but others predict some
permeability (Nye and Frank, 1973). Röthlis-
berger (1972) and Shreve (1972) have in theo-
retical papers carried out further theories of
the behaviour of water in glaciers. Shreve starts
from intergranular openings and then discusses
larger channels, while Röthlisberger discusses
larger channels ancl their behaviour. They both
take into account the heat generated by the
flow of water, plastic movement of the ice, and
a tendency for larger channels to win over the
smaller ones. These theories predict a network
of channels under a glacier similar to usual
river systems.
Theories of glaciers sliding (Lliboutry, 1968;
Weertman, 1964) assume melting of water at
the base of glaciers due both to geothermal
heat and frictional heat from the sliding. They
also agree on the importance of this for sliding
velocity, but disagree on the importance of the
hydrostatic pressure of the water film. Llibou-
try considers it important, but Weertman only
considers the presence and thickness as import-
ant. Such a layer cannot, because of the sub-
glacial channel system, reach appreciable thick-
ness. It is probably much less than 1 mm. A
thick water layer would cause extremely fast
sliding of the glacier.
A very important observational work has
been done at Summit Lake, British Columbia,
described by Gilbert 1971, Matthews 1973, ancl
Fisher 1973. These studies prove that a sub-
stantial leakage took place from the lake be-
fore ancl between the jökulhlaups and also that
Liestöl’s theory can explain the hydrograph
from observed heat content of the lake and
calculated heat from friction. The initiation of
the jökulhlaup cannot be stated with certainty,
but a drop in the ground water table in the
ice coulcl be the explanation. This was also
theoretically concludecl by Röthlisberger 1972.
This drop in the ground water table in the
autumn when the subglacial drainage system is
well developed could cause the leakage water
to take over the drainage system and extend it
by melting into the lake.
THE MECHANISM IN JÖKULHLAUPS
FROM GRÍMSVÖTN
Nielsen (1937) was among the first to discuss
the origin of the water discharge in Gríms-
vatnahlaup. He concludecl that it mainly ori-
ginated from melting of ice in volcanic out-
breaks and that it escaped immediately with
approximately the same speed as the melting.
Thorarinsson (1953) pointed out that a vol-
canic outbreak is not a precondition for a jökul-
hlaup. On the contrary the decreased pressure
due to the release of water could rather be the
cause of the volcanic outbreak. He states that
the source of the water is melting by geothermal
heat from the high temperature thermal area
underlying Grímsvötn. Surface melting of ice
is also a part of the process.
Today everybody agrees with Thorarinsson
in this matter. A model of Grímsvötn would
be that under it flow out 3 m3/sec of 220°C
water with dissolved solids, about 1500 mg/1.
This melts 9 m3/sec of ice containing dissolved
solids, about 30 mg/1. Surface melting amounts
to 4 m3/sec. This mixture has dissolved solids
of about 325 mg/1. All this fits to the observa-
tion. The only snag is, that this is approximately
twice the volume of stored water in Grímsvötn.
My suggestion in this matter is that a sub-
stantial part of the water does leak out over
the 5 years period between jökulhlaups.
The mechanism of escape of the water from
Grimsvötn was first discussed by Thorarinsson
(1953) were he assumes that the bouyancy the-
ory is valid for it. Glen in 1954 explains it with
his theory of jökulhlaup mechanism. Thorarins-
son has later (1965) declared that in the light
of the present knowledge of the surface- and
bedrock topography the bouyancy theory seems
doubtful for the mechanism of water escape
from Grímsvötn. Yet he cloes not exclude it as
impossible and still thinks it in best accordance
36 JÖKULL 24. ÁR