Jökull - 01.12.1966, Blaðsíða 13

TRAUSTI EINARSSON:
Physical Aspects of Sub-glacial Eruptions
For many years there has been mucli refer-
ence in the literature to the supposed para-
mount role of sub-glacial eruptions in the pro-
duction of the great masses of fragmental,
largely glassy basaltic rocks in Iceland. There-
by the physical conditions in such eruptions
have not been given much attention, and it
has been tentatively assumed that a sub-glacial
eruption is a simple and obvious process: the
heat of the extruded Iava goes to melt the over-
lying ice, until eventually a wide gap through
the ice, often many kilometres in diameter, has
been formed, even by an ice thickness of 400—
500 m or more. Table-mountains in Iceland
are claimed to have been formed in this way
(cf. Kjartansson 1966), the topmost lava cap
being due to sub-aerial extrusion after the pile
of pillow lavas, breccias and tuffs had penetrat-
ed the ice-cap and formed a dry surface.
This theory of the table-mountains can in
a number of instances be refuted, as it is found
(Einarsson 1962) that these mountain are made
up of flat-lying layers of fluviatile sediments
alternating with lavas up to middle height. In
some cases the lavas form groups of alternating
magnetic polarity and thus cover a wide range
in time. In these cases of quite typical table-
mountains we are obviously concerned with
isolated parts of much more extended strata —
either due to erosion or, which in other cases
can be demonstrated, due to local uplift.
On the other hand there exist in Iceland
several low and flat ridges that by field evid-
ence are very likely to have been formed under
an ice-cap, and several isolated hills may also
belong to this category. As examples of ridges
may be mentioned Draugahlíðar near Jósefs-
dalur and Dráttarhlíð at the river Sog.
In a country of dominant fissure eruptions
the elongated ridge is a likely result of sub-
glacial eruptions. For whereas in a sub-aerial
fissure eruption the lava extrusion soon be-
comes concentrated at individual points along
the fissure, the resistance of the ice-cover
against extrusion raises the lava pressure in
the fissure, the lava penetrates into all avail-
able parts of the fissure and finally meets the
ice all along the fissure.
To understand how, after such an initial
phase, a linear or point eruption under an
ice-cap will proceed further, it must be realized
that two phvsical factors now become import-
ant: 1) the melting of ice and quenching of
lava by meltwater, 2) the pressure of the over-
lying ice-cap.
If the extrusion of the lava were slow and
if the heat of the extruded lava were wholly
used for melting of ice, it is easily found that
the volume of ice melted would be 6—7 times
that of the lava; that thereby a 5 km wide hole
through a 500 m thick ice-cap would be a
natural outcome is another matter which we
shall leave for the moment. But the point to
be made is that in fissure eruptions the lava
output is usually great in the beginning. In
that case, and probably generally, there would
be insufficient time to create space for the lava
bv melting and draining of the meltwater, and
the pressure of the covering ice would becorne
an important factor. We are therefore concern-
ed with the capabilitv of the lava to lift the
ice cover, and we shall see that this may be
remarkably small.
From what was said above about the initial
stage of a sub-glacial eruption, and taking into
account the two examples of supposed sub-
glacial ridges, we envisage a ridge, a few tens
of metres thick and a few hundred metres
Fig. 1. Schematic section through a sub-glacial-
ly formed volcanic ridge, a few tens of metres
high.
1. mynd. Skýringarmynd af hrygg, nokkrum
tuga metra háum, sem myndazt hefur við gos
undir jökli.
JÖKULL 167