Jökull - 01.01.2010, Blaðsíða 127

Holocene surface ruptures in the South Iceland Seismic Zone
ferent dips. The northern part of the fault dips to the
east, the southern part dips to the west (Hjaltadóttir,
2009). The tear in the surface ruptures is bridged by
a conjugate surface fault at Bitra (Clifton and Einars-
son, 2005), which, by the way, does not show up in
the hypocentral distribution of aftershocks.
The Réttarnes, Leirubakki, and Tjörvafit faults
Unusually clear fault segments are exposed in a flat
lava flow near the farm Leirubakki in the eastern
part of the seismic zone (Figure 7). Three parallel
fault segments can be distinguished, here named by
the nearest place names, Réttarnes, Leirubakki, and
Tjörvafit. The first two were discovered and described
by Einarsson and Eiríksson (1982a), the third one was
identified later. The Leirubakki fault was the subject
of a study by Bergerat et al. (2003). The distance be-
tween the parallel faults is 1–2 km.
The Réttarnes fault is exposed in a rather rough aa-
lava and is traceable primarily by its push-up hillocks
for about one kilometer. Faint trace of it can be fol-
lowed for longer distances. The Leirubakki fault is
beautifully exposed for about 1.5 km in a smooth
pahoehoe-lava. This section perhaps provides the best
examples of strike-slip structures anywhere. The fault
can be traced further in both directions, a total of
about 4 km, but it becomes obscured in the rough lava
to the north and the alluvial deposits to the south. The
Tjörvafit fault is mainly traceable by its push-ups and
pressure ridges. It is clearly traced for 1.5 km. Conju-
gate structures are prominent in its southern part.
It is not known with certainty when these faults
were last active. At least two of the large historical
earthquakes originated in this general area, in 1294
and 1732. Skálholts-annal states (Thoroddsen, 1899)
for an earthquake in the year 1294: "the earth was
fractured in the district Rangárvellir and Rangá river
was forced out of its bed and broke people’s houses".
The Rangá river crosses the seismic zone in this area
and only very few farms are close enough to the river
that such damage could be done. The most likely
farm is Svínhagi (Figure 7). The most probable fault
is therefore the Leirubakki fault although the other
two faults cannot be excluded. The 1732 earthquake
caused extensive damage in the Land district and the
damage zone as mapped by Björnsson (1978) is cen-
tered on the farms Leirubakki and Svínhagi. One of
the three faults shown in Figure 7 must be regarded as
likely source of this earthquake.
A somewhat contrary view on historical activity
is expressed by Bergerat et al. (2003). They argue
that the surface structures of the Leirubakki fault are
formed in one single earthquake. The evidence they
quote is mainly a comparison with the 1912 earth-
quake fault and the events of 2000. In both quoted
examples the earthquakes took place on an old fault
with pre-existing fault structures (Clifton and Einars-
son, 2005). It is therefore by no means certain that the
structures visible along the Leirubakki fault are due
to one single event. In fact, it appears more likely that
more than one event are responsible. Their magnitude
estimate of 7.1 must therefore be regarded as the abso-
lute upper limit. They furthermore argue (Bergerat et
al., 2003) that such a large earthquake would probably
have been recorded in the historical accounts, which
it is not. Therefore it must be pre-historic. This ar-
gumentation does not hold if the structures are due to
more than one event.
DISCUSSION
Conjugate faulting
In the beginning of the mapping project it was gener-
ally accepted that the South Iceland Seismic Zone was
a transform zone and thus a zone of horizontal shear
(e.g. Stefánsson, 1967, Ward, 1971). A major E-W
fault had not been identified in spite of a clear align-
ment of earthquake sources in that direction. A sys-
tem of conjugate faults was to be expected. It soon be-
came evident, however, that a large majority of recog-
nizable surface fracture systems had northerly trends
and that the conjugate trend of ENE was poorly repre-
sented (Einarsson and Eiríksson, 1982a,b). This was
consistent with the observation of Björnsson (1978)
and Einarsson and Björnsson (1979) that damage
zones of individual historical earthquakes were elon-
gated in a N-S direction. The data presented here con-
firm this finding. Fracture arrays indicating left-lateral
faulting on ENE-striking faults exist but they are an
order of magnitude less common in the SISZ than
fracture arrays showing right-lateral displacement on
JÖKULL No. 60 127