Jökull - 01.12.2003, Blaðsíða 35

The 1783–1785 Laki-Grímsvötn eruptions
rubble (Keszthelyi et al., 2000), which covers about
one half of the Laki lava, indicate that it was initially
formed by break up of coherent and stationary pahoe-
hoe crust as a result of surging within the underly-
ing lava. This process exposes the incandescent flow
interior and at this stage the rubble grows in thick-
ness; first by autobrecciation of incandescent lava and
then by the moving lava as it piles the rubble up into
concentric pressure ridges oriented perpendicular to
the flow direction (e.g. Figure 7f). However, the base
of these rubbly lavas is typified by smooth pahoehoe
surface, showing that they have a hybrid morpholog-
ical character that falls between the two end-member
basalt lava flow types, a’a and pahoehoe. However the
emplacement mechanism of these lavas differs signif-
icantly from that of a’a and pahoehoe and therefore it
is considered to be a distinctive flow type, named rub-
bly pahoehoe by Keszthelyi and Thordarson (2000).
All of the flow structures mentioned above are in-
dicative of endogenous growth, suggesting that insu-
lated lava transport and growth by inflation played an
important role during the emplacement of the Laki
lava (e.g. Thordarson and Self, 1993; Keszthelyi et
al., 2000). It is, therefore, reasonable to assume that
the lava, including the surges, was transported from
the vents to the active flow fronts within preferred in-
ternal pathways (i.e., master lava tubes). Such path-
ways were presumably established early on during
emplacement of the lava in the Skaftá River gorge,
and later in the Hverfisfljót River gorge, and subse-
quently lengthened as the lava flow field grew in size.
The lobed architecture of the Laki flow field is also
consistent with this view, because compound flows
are one of the characteristic features of endogenous
flow emplacement (e.g., Walker, 1991; Hon et al.,
1994; Mattox et al., 1993). Each lobe is produced
as a breakout from the main pathways or pre-existing
lobes at the active flow front or onto the lava surface
(Figure 7f), resulting in incremental lengthening and
thickening of the flow field.
ACONCEPTUALMODELONTHE PROGRESS
OF THE LAKI ERUPTION
The above analysis of the contemporary accounts
clearly shows that the Laki eruption featured distinc-
tive eruption episodes reflecting periodic increases in
the magma discharge (Figure 6). Each episode be-
gan with an earthquake swarm of increasing inten-
sity that was followed by vigorous explosive activ-
ity at the fissures and sudden increases in outflow of
lava from the fissures. The seismic swarms gener-
ally lasted for several days to a week, with the ex-
ception being the first which lasted for 3–4 weeks.
Each earthquake swarm was followed by a short-lived
subplinian or phreatomagmatic explosive phase lead-
ing into a longer-lasting phase of lava fountaining and
effusive activity. A surge of lava emerged from ei-
ther the Skaftá River or Hverfisfljót River gorge 3–5
days after the beginning of each explosive phase. The
events described above define an eruptive episode and
all in all Laki event featured 10 such episodes (Figure
6). The Laki cone-row is composed of at least 10 en
echelon fissures, trending N47–48◦E (Figure 3). Us-
ing tephra stratigraphy, coupled with descriptions on
the location and timing of explosive activity and lava
surges, each of the ten eruption episodes is linked to
the opening of a new fissure segment and the stepwise
propagation of the locus of activity from the south-
west to the northeast (Thordarson and Self, 1993). In
this context it is worth noting that there is a percep-
tible increase in the frequency of eruptions at Gríms-
vötn during fall 1783 and that the volcano was active
through to May 1785, adding four eruption episodes
to the 1783–1785 activity on the Grímsvötn volcanic
system (Table 2). It is conceivable that these trends
are linked. The obvious synchronisation in the ac-
tivity at the Laki fissures and the Grímsvötn volcano
indicates that both eruptions resulted from the same
volcano- tectonic event. However, a range of evidence
show that the magma erupted at the Laki fissures is
derived from a large deep-seated reservoir located at
the crust-mantle boundary, whereas the eruptions at
Grímsvötn occurred from a shallow crustal magma
chamber (Gudmundsson, 1987; Sigmarsson et al.,
1991; Thordarson and Self, 1993). Thus, the trends
mentioned abovemay indicate a gradual restoration in
the flow of magma from the deep-seated reservoir to
the shallow magma chamber beneath the Grímsvötn
volcano.
JÖKULL No. 53, 2003 33