Jökull - 01.12.1983, Blaðsíða 107

Fig. 6. Dropstones in the basal part of rhythmic
bed VIII.
Mynd 6. Jakabomir steinar í syrpu VIII.
layer, and one from the consolidated part in section
C, were examined in thin sections. The basal part
contains at least 95% of unaltered glassy grains.
The shape is altemating from needle grains to
grains with concave outlines. The glassy grains in
the consolidated part show flowage pattern. Grains
of feldspar, iron ore, altered glass and a trace of
pyroxene are seen in both samples.
One sample from the basal part was analysed for
grain size by pipette using the method described by
Galehouse (1971). The part of the sample coarser
than 4 0 was drysieved. Grain size distribution is
shown in Fig. 8. Three unaltered grains were
analysed in an electron-microprobe at the Nordic
Volcanological Institute, The chemical com-
position is shown in Table 1; by comparison with
the alkali:silica diagram by Jakobsson (1979) the
tephra belongs to the tholeiitic series.
GENERAL DISCUSSION AND
CONCLUSIONS
The lithology of the Skagafjall deposits is charac-
teristic ofglacial lacustrine sediments.
The surface of glacier ice, where it abuts against a
hillslope, is often gently convex, owing to increased
melting at the ice margin because of heat reflected
from the bedrock slope. A longitudinal hollow may
form between the ice margin and the bedrock slope
(Embleton and King 1975, p. 338). Any meltwater
would collect in such hollows and could form a
marginal lake if the longitudinal gradient of the ice
margin is very small and meltwater adequate.
It is suggested that an extensive glacial lake was
formed in this way between the bedrock slope of
Skagafjall and the margin of the glacier in Dýra-
fjördur. The surface of the glaciers which occupied
Ingjaldssandur and Dýrafjördur must have been at
least 700 m above the present sea level when the
lake was in existence. The depth of the seawater in
the Dýrafjördur mouth is up to 50 m so the ice
thickness must have been at least 750 m when it
dammed up the Skagafjall lake. At the same time an
ice dammed lake was blocked up in Arnarneshvilft
by a glacier which occupied Gerdhamradalur. The
highest terrace in Arnarneshvilft and Oþoli indicate
that the maximum level of the lakes was at least 695
m a.p.s.l.
Late in summer 1981 I found boulder gravel
between 520-590 m elevation on the crest slope of
Saudanes in the northern side of Önundarfjördur.
This indicates that the ice thickness in Önundar-
fjördur mouth was at least 600 m when the boulder
gravel was formed.
The following physical environment of deposit-
ion in Skagafjall glacial lake is [Xistulated. Sediment
was transported by meltwater streams from the
glaciers and uplands. The location of the snowline
when the Skagafjall lake was formed remains a
fundamental problem because meltwater form-
ation, as a result of heat from the adjacent exposed
ground, has hardly been enough to form such an
extensive lake as the Skagafjall lake must have been.
Gravel and sand were carried out into the lake. Clay
became the dominant deposit in winter when the
lake was frozen at its surface and little or no melt-
water was entering. This sequence of events was
repeated every year producing one pair each year.
De Geer (1912) introduced the term “varve” todes-
cribe such annual couplets.
Kuenen (1951) stressed that turbidity currents are
JÖKULL 33. ÁR 105