Jökull - 01.01.2010, Page 129

Holocene surface ruptures in the South Iceland Seismic Zone
segment is (e.g. Angelier and Bergerat, 2002, Bergerat
and Angelier, 2003). Other, much less conspicuous
examples of conjugate faulting are found in the south-
ern part of the Selfoss township (Imsland et al., 1998
a, b), at the Litlu-Reykir fault zone, where it crosses
Highway #1, and in the fracture system active in the
1912 earthquake near the farm Haukadalur.
Length of faults
Since the largest seismogenic faults trend transversely
to the seismic zone, the width of the zone should give
an indication of the length of the faults. The zone
of mapped fault structures, as seen e.g. in Figure 1,
concides very well with the zone of background seis-
micity. This width is about 20 km, almost uniformly
along the zone. The lengths of the source faults of
recent large earthquakes in the zone are consistent
with this. Modeling of surface deformation fields of
the two Mw 6.5 earthquakes of June 2000 gives fault
lengths of 15 km (e.g. Pedersen et al., 2001, 2003)
and the two faults responsible for the double event
of May 2008 were 11 and 17 km long (Decriem et
al., 2010). The mapped surface ruptures of the 1912
earthquake (MS7.0) are 11 km long (Einarsson and
Eiríksson, 1982) and Bjarnason et al. (1993) suggest
that the original rupture may have been as long as
30 km. It has been argued that the length of the de-
struction zones of historical earthquakes in South Ice-
land as shown by e.g. Einarsson and Björnsson (1979)
and Björnsson and Einarsson (1981) may be taken
as a proxy for fault length (e.g. Guðmundsson, 1995,
2000, Angelier et al., 2008) and therefore may be as
long as 50 km. This is not so. Assuming a simple
model where earthquake intensity is a function of only
the distance to the nearest segment of the source fault,
it is easy to show that the fault length should be equal
to L-W, where L is the length and W the width of the
destruction zone. This would give a fault length of
about 30 km for the largest earthquakes with known
destruction zones, i.e. those of 1784 and 1912.
Strike of the faults
All the larger faults, that are reasonably known, strike
almost due N-S. This is clearly seen in the 1912
mapped fault traces, the 1896 second event fault traces
(Einarsson et al., 1981), the aftershock distribution
of the 1987 Vatnafjöll event (Bjarnason and Einars-
son, 1991) and the earthquakes of 2000 (Hjaltadóttir,
2009). Modeling of the deformation fields of the 2000
and 2008 earthquakes shows the same results (Ped-
ersen et al., 2001, 2003, Decriem et al., 2010). Yet
there is frequent reference in the literature to a strike
of NNE for the major faults (e.g. Gudmundsson, 1995,
Gudmundsson and Brynjólfsson, 1993, Bergerat and
Angelier, 2000). The reason for this misconception
appears to be the en echelon arrangement of the fault
structures. The first order en echelon is on the scale of
a kilometer. One may therefore see a kilometer long
strike-slip segment with a NNE strike (see e.g. Figure
7) but fail to see that it is a part of a larger structure
with a strike of N-S. Detailed studies of the hypocen-
tral distribution of the earthquake sequence of 2000
and modeling of the deformation field confirms that
the en echelon fracture arrays at the surface are un-
derlain by a continuous, near-vertical fault plane with
a northerly strike. This relationship is to be expected
where a strike-slip fault, initiated at depth, propagates
towards the free surface.
The Grímsnes volcanic system
A population of NE-SW striking fractures in Gríms-
nes, at the northern border of the SISZ is identified
as belonging to a fissure swarm associated with the
Grímsnes Volcanic System. The GVS is placed un-
conformably on top of older crust (Jakobsson, 1966)
and judging from the lack of long, continuous struc-
tures, its fissure swarm is immature. The fractures are
expressed as rows of sinkholes and depressions in the
Holocene surface. They are mostly extensional and
rarely exhibit a normal component. Their widths are
implied to be of the order of 1–2 m and lengths are
generally less than 1 km. En echelon arrangements
are hardly seen at all, and push-ups are not known
here. The fractures are fairly evenly distributed in
a 6×20 km swarm. Total dilatation in the Holocene
is estimated to be of the order of 10–20 m. Frac-
tures judged to belong to the GVS are not included
in Figure 1.
JÖKULL No. 60 129