Jökull - 01.12.1974, Blaðsíða 24

during the period before the 1972 jökulhlaup.
The difference of 0.5 km3 is roughly equal to
the water that accumulates in Grímsvötn dur-
ing an 11 months period.
Many jökulhlaups have been accompanied by
volcanic eruptions in the mountains surround-
ing Grímsvötn, e. g. in 1922 and 1934. Thorar-
insson (1953) observed that these eruptions do
not. c.ause the jökulhlaups, and he suggested
that the eruptions might be triggered by the
pressure drop at the volcano as the water drains
out of Grímsvötn. Volcanic eruptions inside the
water basin north of Grímsvötn could trigger
jökulhlaups from Grímsvötn by adding extra
meltwater to the lake, and this is what might
have caused the jökulhlaup in 1938. Volcanic
eruptions inside Grímsvötn, on the other hand,
can hardly trigger unexpected jökulhlaups as
they could only melt the ice cover on the lake
and could not raise the water level. (This as-
sumes, of course, that the volume of the vol-
canic materials introduced into Grímsvötn is
negligible compared to the volume of the lake.
The average volcanic eruption in Iceland pro-
duces about 0.2 km3 of volcanic materials com-
pared with the lake’s volume of, say, 3 km3).
An eruption at the subglacial ridge east of the
lake, causing melting of ice at the threshold,
could cause a jökulhlaup by breaking the seal.
It might also change the release mechanism by
increasing the height of the subglacial ridge
and by reducing the glacier thickness. If 13 vol-
ume units of ice were melted for one volume
unit of volcanic material introduced, the glacier
surface would drop more than the subglacial
ridge would rise; although the water level in
Grímsvötn would have to rise higher than at
present to give a certain value of the hydraulic
head relative to the height of the new ridge, the
glacier which would have to be lifted would be
much thinner than the present one. It may
take some decades for the glacier at the seal to
attain a height appropriate to the climate
conditions. After that a higher water level in
Grímsvötn would be required to lift the glacier
and the jökulhlaups would be greater in vol-
ume.
The mass balance model predicts water vol-
umes for the jökulhlaups which are in a reason-
able agreement with the estimated runoff in
the rivers on Skeidarársandur. This model
should, however, be subject to further tests.
Detailed data on the mass balance ancl the'ice
velocity should be collected in the water basin.
The bottom topography of the lake should be
mapped. This map, together with data on the
rate of rise of the water level (Fig. 3) in the
lake, would give the best estimate of the filling
rate of the lake.
The present study does not confirm Malm-
berg’s (1974) estimates of the jökulhlaup in
1972. His estimate of water volume was only
i/3 to i/, of Rist’s (1973) estimate.
In the present theory no attempt is made to
link the reported observations (Thorarinsson,
1974; Stefánsson, 1973) of a rise of the lower
part of Skeidarárjökull to the explanation of
the jökulhlaup mechanism. On the whole, they
seem to be somewhat contradictionary and do
not appear to be an essential part of the re-
lease mechanism. They should, however, be
thoroughly examined.
Plastic deformation as a trigger mechanism.
Glen (1954) advanced the hypothesis that
water might escape from Grímsvötn by causing
plastic deformation of the ice. Consider a gen-
eral model of water in a vertical hole in the
glacier. The difference between the water pres-
sure and the ice overburden pressure far from
the hole is
(^) Pw§ Pi fí li
in which lw is the depth of the water in the
hole and 1; is the length of the hole, pw and p,
are the density of water and ice, respectively,
and g is the acceleration of gravity. Glen’s
(1954) model of a water-filled hole is the special
case for which lw = lj. For a lw of 200 m he
obtainecl a Ap of 2 bars, which induces a shear
stress of 1 bar on planes at 45° to the hori-
zontal. Deformation of the ice at the bottom of
the hole would then allow water to penetrate
the ice. However, such a model does not de-
scribe the situation in Grímsvötn.
If a water-filled hole were to reach the glacier
bottom one would expect lifting of the glacier.
Also, the hole may be only partly filled with
water. A water depth for which the water pres-
sure at the bed matches the ice overburden
pressure could cause lifting of the glacier. Since
lifting of the glacier can occur at lower water
22 JÖKULL 24. ÁR