Jökull - 01.12.1966, Qupperneq 16

soon seep out under the glacier and play no
further role in the happenings. The cooled
fragmented lava would, on the other hand, be-
come a good insulator between the ice ceiling
and the gfowing lava. If instead of crumbling,
the lava is turned into pillows the result would
not, be much different. Each pillow would get
a colcl crust and let the heat out slowly ancl
the mass as a whole would be a good insulator.
Thus after the initial contact between ice
and lava, further lava flow will soon be insulat-
ed from direct contact with the ice and now
the weight of the ice cap comes into the pic-
t.ure. This immediately decreases the flow by
counter-pressure and if the lava is effectively
prevented from any contact with the ice this
slowing down will ease the quenching process
by any meltwater that comes from above. The
wall between lava and ice becomes thicker and
a final sealing of the fissure takes place if the
ice sheet is 150 m or more. The easiest way
for the lava would be sideways at the base of
the ice, and this is in keeping with what the
two cases previously mentioned show: the ridges
are low with a flat surface and the breadth is
5—10 times larger than the height.
Now, even if the glacier were fractured the
stress will spreacl laterally upwards. The ice
load then corresponds to the weight of a wedge
of about 90° opening angle. The thrust of the
lava in the 2 m wide dyke can then be counter-
balanced by a 100 m thick ice cap, or, in the
case of a point eruption, by a 75 m thick ice.
Thus for an ice-cap 400 or 500 m thick the
blocking is fully effective whether the ice is
solid or crossed by a wide mesh of fractures.
Finally, if a hydraulic press develops or if
steam pressure becomes effective, slight lifting
will let the lava or steam sideways out under
the glacier.
Sofar we have been able to see, the latent
heat of the lava would be very ineffectively
used for such a process as melting a wav up
through a thick ice sheet. But we shall also
face such a case. The lava is then supposed to
find ways to the top o£ the pile of brecciated
material above the volcanic fissure, and in the
end to penetrate to the surface. The gap so
formed might be some 100 m wide at the base
and in spite of good insulation between ice
wall and the lava column, due to cooled
crumbled material, we shall suppose that be-
cause of slumping of fractured ice the gap
wiclens upwards at an angle of 90°, being one
km wide at the surface of a 500 m thick ice.
At this point the eruption will in the beginn-
ing be explosive and a laver of tuff and bombs
will settle on the surface of the ice. As this
material was cooled in the air before settling
it has no effect on the ice, as was well demon-
strated by a layer of bombs falling on snow
around Hekla in 1947. On the contrary the
layer forms an effective insulation if the erup-
tion should eventually become effusive; the
lava will just spread over the ice as if this was
a surface of sand.
When in the end the eruption stops, there
lias been formed an extremely unstable struc-
ture supported by ice, and it seems most likely
that as the ice flow must be assumecl to con-
tinue, this structure will very largely be carried
away by the ice. In any case the remnant left
at the end of glaciation would have little re-
semblance of the original structure and in special
it would not be a regular table-mountain with
dominant horizontal stratification.
Compare with this picture what the sub-
glacial extrusion hvpothesis of the table-moun-
tains demands, Fig. 4. Through an ice-cap 500
m thick it is expected that melting forms a
hole some 5 km in diameter, i.e. 10 times the
thickness o£ the ice sheet. The melting is,
further, expected to be so neatly regulated that
fragmental material can settle simultaneouslv
in the entire horizontal section of the gap to
form a regular succession of horizontal layers.
All this could be done by a Maxwellian demon
who directed the latent energy of the lava into
the desired channels. But instead of such an
agency we are left with the second law of
thermodynamics which prohibits an arbitrary
prescription of the flow of heat in a natural
process. We must inquire into the probability
of a process ancl the forming of the wide gap
seems a very unlikely use of the lava heat.
This paper is nearly exclusively concerned
with the physical aspects of a sub-glacial erup-
tion and we think that such a consideration
clarifies what will reasonably happen in most
cases. Field observations must not be summarily
explained be reference to physically most un-
likely prcesses but must take the normal geo-
logical agencies, such as faulting and erosion,
fully into account.
170 JÖKULL