Jökull - 01.06.2000, Page 8

Holocene eruptions within the Katla volcanic system,
Table 2. Chemical composition of basaltic tephra from intracaldera eruptions of the Katla system. – Efnasamsetning glers
úr basísku Kötlugjóskulagi, K-1625.
K tephra n SiO TiO Al O FeO MnO MgO CaO Na O K O P O Total
K 1625 7 46,28 4,56 12,62 14,75 0,23 4,89 9,97 2,72 0,71 0,72 97,44
0,46 0,22 0,33 0,30 0,03 0,08 0,24 0,12 0,05 0,12 0,69
result of rapid escape of meltwater from the eruption
site or high initial mass eruption rates.
The vent area in several historical Katla eruptions
is known either from sightings (1755, 1860, 1918) or
from direct observation of the site (1823, 1918). In the
latter case, the vent area was in the southeastern part
of the caldera, near its margin. The location of ice
cauldrons formed in 1955, as shown by Rist (1967),
does not coincide with the location of the 1918 and
1823 eruption sites, which lie some 3 km farther south
as described by G. Sveinsson (1919) and Austmann
(S.t.s. Ísl. IV, 1907-15).
Katla eruptions apparently occur on short fissures,
but their length and orientation is difficult to assess
from the scant descriptions. If the 1955 incident was
caused by an eruption, a NNW-SSE trending 1-1.5 km
long fissure is implied by the cauldrons. Its orienta-
tion is sub-parallel to the caldera rim as defined by
Björnsson et al. (1993 and this volume). The ice de-
pression over the 1918 vent was 0.8-1 km wide from
north to south but the E-W length could not be de-
termined (G. Sveinsson, 1919). Another source esti-
mated the length of the depression to about one Dan-
ish mile (7-8 km) whereof a 0.5-0.8 km long chasm
near its NW termination was thought to be the vent
area (P. Sveinsson, 1919). The depression may have
been partly modified by meltwater channels, in which
case the original fissure was shorter than implied by
the depression.
All documented Katla jökulhlaups since the late
12th century have escaped from the caldera along
the Kötlujökull pass onto Mýrdalssandur, with the
exception of a minor “jökulhlaup” from below Sól-
heimajökull during the 1860 eruption (Hákonarson,
1860). Accounts of volcanogenic jökulhlaups onto
Skóga- and Sólheimasandur in the 13th and 14th cen-
turies are not supported by field data (Dugmore, 1987;
Larsen and Dugmore, unpubl. data), as the last jökul-
hlaup to leave detectable deposits in sections around
these sandur plains occurred in the early 10th cen-
tury. Prehistorical jökulhlaups also escaped through
the Entujökull pass (Sigurðsson, 1988). The historical
jökulhlaups have emerged in several outbursts onto
Mýrdalssandur, those of the first day usually being
most voluminous. The first 15 km of their route lies
below, within or on top of the Kötlujökull glacier, and
they may emerge out from under its snout or break out
well above the ice margin. The locations of the out-
lets determine the routes across Mýrdalssandur, some
of which are shown in Figure 4.
Jökulhlaups accompanying Katla eruptions are a
mixture of meltwater, ice and volcanic debris, mostly
in the ash and lapilli range (Einarsson, 1975; Sveins-
son, 1994). They have been defined as debris flows
(Jónsson, 1982; Maizels, 1993), water floods (Karls-
son, 1994; Tómasson, 1996) and as alternating be-
tween mud flows and water flows (Björnsson, 1993)
on their 35-40 km route from the eruption site to
the coast. The velocity of the leading edge of the
1918 jökulhlaup on Mýrdalssandur plain was close to
20 km/h or 6 m/s according to eyewitness accounts
(Jóhannsson, 1919) while sediment structures indi-
cate a velocity of up to 15 m/s in channels (Maizels,
1993). The maximum discharge for the 1918 jökul-
hlaup has been estimated at 100.000 to 300.000 m

/s
and the volume of meltwater has been estimated to
be as much as 8 km

(Tómasson, 1996). Estimates
of water-transported volcanic debris vary between 0.7
and 1.6 km

respectively (Larsen and Ásbjörnsson,
1995; Tómasson, 1996).
Lightning is common in the clouds of the Katla
eruptions. Lightning frequency appears to correlate
to the intensity of the eruption, but occurrences are
also reported when activity is low. When frequency is
highest, lightning occurs at intervals of a few seconds.
Most appear to be intracloud discharges but strikes be-
JÖKULL No. 49 7