Jökull - 01.12.1999, Qupperneq 87

Figure 12. Fluvioglacial alluvial fan cover against the
background of the development of level VI. 1) area
of the field of research, 2) area of concentrated occur-
rence of kettles currently with a relief greatly trans-
formed by nival waters, 3) individual kettles, con-
siderably filled with deposits.— Yfirborð rannsókn-
arsvœðisins. Örvarnar benda á svæðið þar sem jök-
ulkerin eru þéttust (2) og stök, sandfyllt jökulker (3).
rims, till rims, and boulder rims whereas Thwaites
(1935) introduced the term kettle rim, underlining the
participation of debris, moraine and boulders, from
melted ice, in its structure.
The kettles under study are limited to a con-
fined region of the Höfðabrekkujökull forefield. This
part of Mýrdalssandur has many similar features as
the sediments from the 1918 jökulhlaup (Figure 11).
However, the flat surface of level VI within the ket-
tle region (Figure 12), lacks various forms which are
common on the surrounding plain, such as rims and
gravel, diamicton, moraine and boulder mounds. The
kettle on level VI contain genetically complex forms
which were not recognized by Maizels (1992) on level
II, defined by the authors as rimmed-normal kettle.
However, the rim structure is not visible on the surface
but occurs at a depth of 1 m and therefore one cannot
exclude that some of the kettles on level II could have
formed in a similar way.
The main ílow of the 1918 jökulhlaup lay west of
the Hafursey massif (Jónsson 1982; Larsen and As-
björnsson 1995; Tómasson 1996), creating the sand-
ur level VI, which originally extended between Haf-
ursey and Vatnsrásarhöfuð. A small fragment of
outwash level VI is still visible by Vatnsrásarhöfuð,
above the main channel of the present day river. Ac-
cording to the principle of the chronological devel-
opment of sandur plains, all levels lower than the
higest 1918 jökulhlaup level (II) should be younger.
We can identify level VI on aerial photographs from
1946 on which the kettle field (level Vla) is larger
than today with twice as many kettles. The geolog-
ical structure of level VI is typical of sandurs from
jökulhlaups (classified as type III sandur by Maizels
(1991)) and documents the decline of this catastrophic
flood. Floods after 1946 have thus erased a large part
of the kettle field. Our mapping, being more detailed
than Maizels's (1992), shows that the area labelled
“active sandur” on Figure 1A encompasses a part of
the 1918 jökulhlaup sediments as well. If the dead
ice which settled in the kettle field was transported
there during a jökulhlaup after 1918, then only the
June 25th, 1955 jökulhlaup can be taken into account
as the jökulhlaup on January 20th, 1956 was, accord-
ing to Rist (1967), significantly smaller. Although we
do not know the level of the Sandvatn river relative
to the 1918 plain prior to the 1955 jökulhlaup it is
unlikely that the river overflowed the kettle field, at
that time as level VI currently lies 5.5 m above the
surrounding plain (level VIII) and 8-9 m above the
river basin. The 1955 jökulhlaup was large enough
to overílow and destroy the bridge across the narrow
JÖKULL, No. 47 85