Jökull - 01.12.1982, Qupperneq 70

Fig. 7. Hypothetical cross section of a Katla flow. (a) Fine grained base of the flow, mainly pumice (cf.
Fig. 8). (b) Coarse-grained central section consisting of clast in all dimensions and pumice. (c) Top
section containing mixture of fresh volcanic material, clasts, soil fragments picked up by the flow,
fragmental ice and iceblocks. (d) Big iceblocks.
Mynd 7. Þverskurður af Kötluhlaufii. Hugmynd. (a) Neðsta, fínkornótta lag hlaupsins, aðallega vikur (sbr. mynd 8).
(b) Grófari miðkafli hlaupmassans, sem samanstendur af grjóti í öllum stærðum ásamt vikri og sandi. (c). Efsta lag
hlauþmassans, sem samanstendur af bl'óndu af ferskum gosefnum, grjóti,jarðvegstorfum, sem hlauþið hefur rifið með sér,
íshrafli og jöklum. (d) Stórir jökuljakar.
stroyed by a debris flow about 1480±75 C14 years
ago or about AD 470. The material left on the
sandur plain is mainly pumice.
CONCLUSIONS
1) The initial — and the main— phase of the Katla
ílows is to be regarded as a volcanoglacial debris
flow.
2) The Katla flows are considered to contain 80%
or rnore solid material.
3) As the material consists of up to 95% fresh vol-
canic glass, it is concluded that thc first pulse of
the flow is derived directly from the volcanic
vent.
4) The first phase of the flow rushes down on top of
the glacier. The second phase advances partly
intraglacially (through shear planes and crev-
asses in the ice) and partly subglacially.
5) The density of' the flow is supposed to be 2. 5
g/cm3 or more.
6) The iceblocks are carried on top of the flow, and
big boulders are carried passively in laminar
flow.
7) Between Hafursey and Vestureggjar the 1918
Katla flow has, by Professor 4’horbjörn Karls-
son, been calculated to amount to 1. 5 mifl.
mVsek. Consequently the total volume of the
flow must have been considerably more than 2
mill. mVsek.
8) I he flows are supposed to give rise to turbidity
currents in the ocean, cf. Malmberg (1978) and
Walker (1972).
ACKNOWLEDGEM ENT
The author is highly indebted to G. Ó. Fridleifs-
son and I. B. Fridleifsson for discussions in conn-
ection with this paper and for correcting the English
66 JÖKULL 32. ÁR