Jökull - 01.12.1974, Qupperneq 23

Consider the form of subglacial waterways in
this’area. Beneath the wide ice cover of Gríms-
vötn water moves slowly toward tunnels east of
the lake. The smooth flat roof of the lake’s
ice cover leads gradually into arched tunnels
at the west side of the subglacial ridge and to
more circular tunnels running down to the
Skeidarárjökull valley. The glacier bottom is
expected to be rather smooth on the west fac-
ing side of the subglacial ridge but to have a
high relief on the east sloping face. Tlie cutting
off of the jökulhlaups may be explained by rapid
plastic deformation of the arched subglacial
tunnels as the vaulted smooth roof of the ice
cover of Grímsvötn falls clown onto the rela-
tively flat west-facing side of the subglacial
ridge. The more circular tunnels on the east-
facing side of the ridge are then left open, but
they contain no water. Note that this is assum-
ecl to occur within the lake (nortli of Svíahnúk-
ur eystri on Grímsfjall).
DISCUSSION
A model of the jökulhlaups that have occurr-
ed during the last three decades has been de-
scribed. The accuracy of the model and whether
the model can explain the greater jökulhlaups
will now be discussed.
The significance of the topographical model
The model is based on maps which are not
very reliable north of Grímsvötn. It can, on the
other hand, be shown that water would accu-
mulate in Grímsvötn whatever the shape of the
bottom of the water, basin given the existing
glacier surface slope. Tlie dominating influence
of Bárdarbunga on the surface map makes
detailed knowledge of the bottom topography
irrelevant for understanding the process of
water accumulation in Grímsvötn.
The glacier surface map also reveals that the
jökulhlaups must escape down Skeidarárjökull.
The only part of the glacier bottom map that
is important for the present study is the sub-
glacial ridge east and south-east of Grímsvötn.
Fortunately, this part of the bottom map is
well known.
The present model does explain how the pre-
jökulhlaup water level in Grímsvötn could rise
150—200 m in the early part of this century.
Vatnajökull was thicker then than it is now; if
the glacier surface elevation was 1600 m at the
threshold, the water level in Grímsvötn might
have risen up to 1500 m.
Without the geothermal area the glacier sur-
face would either have a continuous slope from
Bárdarbunga (2000 m) towarcl Grímsfjall
(1719 m) or both mountains would slope down
to a stagnation point. In neither of these cases
could water accumulate in the Grímsvötn de-
pression and cause jökulhlaups by lifting the
glacier.
Comparison witli experience
The relationship between the frequency and
water volume of jökulhlaups from Grímsvötn
ancl the climatic conditions on Vatnajökull
has been discussed. Decreased surface ablation
and increased positive mass balance would cause
the glacier to thicken at the threshold ancl
would reduce the flow of water into the lake.
Both effects would increase the period between
jökulhlaups. Increased ablation and decreased
balance would have the opposite effect, in-
creasing the frequency and decreasing the water
volume of the jökulhlaups.
The only detailed observations of jökulhlaups
have been obtained since 1954 (Fig. 3). The
critical water level has increased by about 10 m
since 1960. The period of water accumulation
was 67 months before the jökulhlaup of 1960
and 1965, but was 78 months before the jökul-
hlaup of 1972. Balance measurements at the
threshold during this period are not available
but geodetic surveys on the line Grímsfjall—
Itverkfjöll (Fig. 8) in 1961 and 1975 showecl no
net change in ice depth along the profile. How-
ever, a slight thickening can be observed along
the first 5 km of the profile.
The increase of 11 months in the accumula-
tion period before the 1972 jökulhlaup may be
clue to the increasecl activity of jökulhlaups in
Skaftá since 1964. Around 1955 a change seems
to have taken place in the Grímsvötn water
basin causing some water to flow to Skaftá.
Such a change in drainage should delay the
jökulhlaups from Grímsvötn. According to
Björnsson (in press), the runoff to Skaftá was
about 0.4 km3 during accurnulation periods be-
fore the 1960 ancl 1965 jökulhlaups from Gríms-
vötn. However, the runoff was about 0.9 km3
JÖKULL 24. ÁR 21