Jökull - 01.12.1985, Síða 40

Fig. 7. The NE
flank of Fella-
heiði, showing
the discon-
tinuity between
the lower
basalts and the
overlying col-
umnar basalt,
together with
the directions of
the channel sys-
tems. S shows
Sandvatn stóra,
F is Fjallsel, R is
Rangárhnjúkur.
— 7. mynd.
Loftmynd af
Fellaheiði.
otherwise dry, level floors rising to the N. The present
stream is a much more recent feature and its course is
largely determined by the channels, which appear to
end on the margin of the plateau, guided by the trend of
the flexure zone, (Fig. 2).
The landscape between the N end of Lagarfljót and
the Fellaheiði plateau is one of hill masses deeply
dissected by ice and channel systems. One of the higher
hiils, Hafrafell, 218 m, has channels right on its summit,
associated with large and small cavettos close by, which
often end in sharply pointed hollows. The channels are
extremely sharply cut also, and one, containing a sheep-
fold, has horizontal floor and almost vertical sides
marked with ice scratches and hollows. It is difficult to
visualise how such a cross-section could be cut by ice
alone, however, even when working in dykes less re-
sistant than the local basalt flows, and the most likely
agent could be water under thick ice, which could also
account for the cavettos. Some rather similar features of
channels and cavettos are also found on the top and
flanks of Heiðarendi, at 400 m the point of the plateau
overlooking Jökuldalur, but the main assemblage of
channels is around Rangárhnjúkur, 562 m. This pro-
minent point forms the N end of a thick band of
columnar basalt, which runs from the SW on top of the
plateau, (Figs. 1 and 7), at a marked angle to the strike
of the Lagarfljót flexure zone and showing up clearly as
a distinct feature on Landsat images. It appears that the
northerly trend of the moving ice was interrupted by
this resistant feature and diverted round its end, where
the cliff is steep and there are no channels. As soon as
the ice was able to resume its northerly trend, however,
the slopes are cut with lines of channels running up and
down over the spurs, and extending to the summit
point, just below which is a small channel with a pond in
it.
Here, again, it seems possible to suggest the com-
bined activity of ice and water and in this case where the
ice was apparently constrained against the resistant rock
face, internal pressure would be high and water be
38 JÖKULL 35. ÁR