Jökull - 01.12.1974, Blaðsíða 39

with the changes in the beliaviour of Gríms-
vötn in historical times. (Thorarinsson, 1974).
There is not reason to assume that the pro-
cess of the jökulhlaup in Grímsvötn is clifferent
from that of other ice-dammed lakes. Grímsvötn
is though, because of its great distance from
the glacier snout and the many years between
jökulhlaups, a rather complicated phenomenon.
A jökulhlaup from Grímsvötn can occur at
any time of the year, although during the last
half a century it has been almost twice as com-
mon during the first half of the year as the
second one. This indicates that the process of
jökulhlaups from Grímsvötn is unusually com-
plicated, as for other ice-dammed lakes there
is a strong tendency for jökulhlaups to occur
in late summer or fall. That is also in accord-
ance with a theory that leakage from the ice-
dammed lakes takes up the drainage system of
the glacier when meltwater decreases either
temporarily or seasonally.
Well in advance of a jökulhlaup from Gríms-
vötn a substantial rising of the glacier tongue
of Skeidarárjökull has been observed by farmer
R. Stefánsson in Skaftafell. This rising of the
glacier has been explained by the glacier float-
ing on water and thus indicating that all the
glacier is afloat all the way from Grímsvötn
down to the glacier snout. This bulge of the
glacier disappears in a few months after the
jökulhlaup. To me this rather indicates a rapid
advance of the glacier due to higher hydro-
static pressure at the base in accordance with
Weertman’s and Lliboutry’s theories on glacier
sliding. With real uplift of the glacier this
bulge should disappear at the end of the jökul-
hlaup.
Often a substantial decrease in the discharge
of the rivers on Skeidarársandur has been ob-
served before the jökulhlaup. In my opinion
this decrease is due to cut-off of the water
channels at the base of the glacier in connec-
tion with the surge of the glacier. This cut-off
could either be as Nye has described (1973), or
at pressure planes. In 1965 the well-marked
water from Grímsvötn was cut off before the
jökulhlaup as seen in Fig. 2.
The mineralogical composition of the sedi-
ment changed in a regular manner during the
1972 jökulhlaup. This is shown on Fig. 8. This
mdicates that in the beginning the material is
derived from moraine near the glacier snout,
but originates from farther and farther in and
closer to Grímsvötn as the jökulhlaup proceeds.
This would indicate that the limiting cross-
sectional area is initially near the glacier snout
but subsequently moves upward.
How can this be explained in view of the
referred theories above? Most likely there are
many factors at work here. One is that a sub-
glacial channel can be formed according to
Glen’s law of pressure difference in horizontal
and vertical directions. Such a channel can be
widened through melting due to the heat con-
tent of the water although flow through it is
very small initially. The heat content may be
greater during winter than summer because of
less mixing of 0° C surface water. In this way
a channel is wedged and melted until it reaches
tlie subglacial drainage system and a jökulhlaup
is started.
In the beginning of the jökulhlaup most of
the heat content of the water is absorbed in
melting of ice in the upper part of the channel,
out the widening of the lower part takes place
through melting by frictional heat. This would
satisfactorily describe the mineralogical com-
position of the sediment load as the high velo-
city in the tunnels in the lower reach of the
glacier tongue would immediately carry out the
melted moraine. On the other hand the sedi-
ment from the upper most of the channel woulcl
not be carried away until in the latter half of
the jökulhlaup when high velocities extend all
the way up.
It is not necessary that a Glen’s type mechanism
is at work in a jökulhlaup. It can just as well
be initiated by leakage as described by Mat-
hews (1973) and others. The leakage would initi-
ate a Liestöl type of mechanism when pressure
in the subglacial channel systems drops either
temporarily or seasonally. This can also play
an important role in connection with Gríms-
vötn especially in the case of a small jökul-
hlaup. A prerequisite for that would be an
unusually well developed subglacial channel
system.
In the jökulhlaup the water leaves the margin
through many outlets. The contacts between
the outlets can best be visualized through trans-
verse cavities at the bottom of the glaciers at
places where a sudden increase in bedrock slope
JÖKULL 24. ÁR 37