Jökull - 01.06.2000, Blaðsíða 18

Holocene eruptions within the Katla volcanic system,
S-Eldgjá segment where the 345 km

western lavas,
including the Álftaver lava field, emanated while all
the remaining subaerial segments contributed to the
435 km

eastern lavas, including the Meðalland and
Landbrot lava field. The lava fields were previously
thought to have a combined volume of 14-16 km


(Miller, 1989) but new field observations indicate that
it may exceed 18 km


(Thordarson et al., in press).
The lava fields are pahoehoe lavas, dotted with, and in
places dominated by, rootless vents (pseudocraters),
indicating emplacement over wet ground or shallow
lakes. The main lobe of the Álftaver lava extends
more than 50 km from the source to the south coast,
while several shorter lobes with distinct flow fronts
occur closer to the source. Activity on the subaerial
part apparently outlasted that on the subglacial part
and may have continued intermittently for a prolonged
period, as implied by the late flow lobes.
Jökulhlaups accompanied the Eldgjá eruption but
their extent and timing are only partly known. A
fan deposited by a major jökulhlaup emerging south
of Öldufell (or from below Öldufellsjökull), con-
formably overlain by bedded Eldgjá tephra, was most
likely emplaced during the early stages of the Eld-
gjá eruption. Other potential early flood deposits oc-
cur on Mælifellssandur and north of Álftaver. De-
bris fans that can be fitted into the eruption sequence
were formed during the later stages of the eruption by
floods apparently emerging near Öldufell and Sand-
fell. Such debris is found on the lower slopes of Atla-
ey where it fits into the tephra stratigraphy at the same
level as the Eldgjá tephra. There it overlies or in-
tercalates the airfall deposit and is also found sand-
wiched between lava lobes. Evidence of a jökulhlaup
from below Sólheimajökull in the early 10th century
is found on the vegetated slopes east of the outwash
plain. This is the last verified occurrence of a jökul-
hlaup leaving a discernible deposit of volcanic debris
in that area (Larsen and Dugmore, unpublished data).
The deposit either belongs to the Eldgjá event or the
previous Katla eruption (ca. 920).
The Eldgjá tephra and lavas have the chemical
characteristics of the Katla volcanic system, being a
transitional alkali basalt with high iron and titanium
content (Table 5 and Jakobsson, 1979). A tephra unit
erupted on the caldera segment during a late stage of
the explosive phase contains minor amounts of sili-
cic (SiO ca 64%) glass interspersed in the basaltic
tephra. This component resembles the silicic tephra
layers described in the previous section. A character-
istic feature of the basaltic tephra units erupted within
the caldera is the abundance of small glomerocrysts,
mostly plagioclase, in the glass. Contamination of
FeTi basalt by acid melt would result in crystalliza-
tion of plagioclase. Compositional variations along
the Eldgjá fissure as described by Miller (1989) in-
dicate contamination by tholeiite, which is most pro-
nounced at the northern part of the fissure, confirming
Jónasson’s (1974) observations.
Course of events in the Eldgjá fires
The course of events during the Eldgjá eruption is
not known in detail. The first material to appear was
tephra erupted on the topographically low fissure seg-
ments west of Öldufellsjökull and at S-Eldgjá. Jökul-
hlaups from below Öldufellsjökull seem to have ac-
companied this activity, and emanation of lava from
S-Eldgjá may have begun at the same time. Next to
appear was a batch of tephra erupted on the caldera
segment, seemingly in a single short burst, while ac-
tivity on the first-mentioned segments appears to have
been continous. Tephra units indicate that the re-
maining subaerial fissure segments became active in
a stepwise fashion, first C-Eldgjá, followed by Eld-
gjá proper and finally by N-Eldgjá. Lava production
probably began with the first appearance of tephra on
each segment. The duration of this phase, the progres-
sive stage, is not known. By analogy with the 1783-
85 Skaftá fires it may have lasted weeks (Thordarson
and Self, 1993). Ejection of airborne material from
the caldera segment resumed after the opening of the
C-Eldgjá segment. Activity at the caldera may have
ceased or been reduced as other segments set in. Al-
ternatively, activity on the caldera segment may have
been continuous with the greatest part of the material
emplaced as hyaloclastic flow(s) through the Kötlu-
jökull gap down to Kriki, only a small part of the ma-
terial becoming airborne until a later stage.
The duration of the Eldgjá event may have been
years. The well defined fronts of some of the late lava
lobes may be taken as an indication that the under-
JÖKULL No. 49 17