Jökull - 01.01.2010, Page 114

Hjartardóttir et al.
the stress field would be governed by the differential
movements caused by the deglaciation. The Kerling-
ar fault displays characteristics which may indicate
that the fault was formed without the involvement of
magma; it is long (∼30 km) and continuous, as op-
posed to the sinuous and discontinuous normal faults
characteristic of dike-induced rifts. This indicates that
the stress field causing the formation of the Kerlingar
fault extended over a large area. We consider it likely
that numerous other faults at the boundary between
the NVZ and the EFB formed in a similar manner,
and that magma intruded some of them, either verti-
cally from the mantle, or horizontally from the Kverk-
fjöll central volcano. This process could have formed
the distinct arcuate pattern of hyaloclastite ridges seen
at the boundary between the NVZ and the EFB, in
the continuation of the Kverkfjöll fissure swarm (Fig-
ure 1).
Differential movements, as suggested here, would
have been confined to this area and not applicable
to other flank zones of the Icelandic rift. In par-
ticular, this does not apply to the southernmost part
of the NVZ-EFB boundary, as no marked difference
in crustal thickness exists there (e.g. Allen et al.
2002; Brandsdóttir and Menke 2008; Darbyshire et
al. 1998). Variations in crustal thickness along the
margin of the NVZ could also explain why the dis-
tinct arcuate pattern of hyaloclastite ridges seen at the
NVZ-EFB boundary does not extend all the way to
the Kverkfjöll central volcano (Figure 1).
It has been proposed that stress changes during
deglaciations have caused the formation of faults in
continental settings, such as in Fennoscandia (e.g.
Muir Wood 1989). These are thrust faults, consistent
with horizontal compression in the old, continental
crust (Sykes and Sbar 1974; Wu et al. 1999). How-
ever, the reaction of the NVZ-EFB crust to deloading
should be different. Crustal extension is expected in
the much younger crust (Árnadóttir et al. 2009; Sykes
and Sbar 1974).
In this paper, we have presented an overview on
the processes that may cause differential movements
at the NVZ-EFB boundary. For a more detailed study,
modeling of these deloading effects (i.e. with Finite
Element Modelling) is preferable.
CONCLUSIONS
1. The Kerlingar fault is a 30 km long normal fault
located on the eastern boundary of the Northern Vol-
canic Rift Zone, Iceland. It is a gently curved NNW-
oriented feature. The fault is located within and par-
allel with hyaloclastite ridges which form an arcuate
pattern along the boundary of the NVZ and the EFB.
2. The fault has some notable features: a. It is un-
usually long and continuous, compared with fractures
and normal faults within the NVZ. b. It has a throw
down to the east, although it is located at the eastern
boundary of the NVZ. c. It is not parallel with the fis-
sure swarms in the NVZ at this latitude, although it
is parallel with hyaloclastite ridges at the NVZ-EFB
boundary, as well as with several other faults at the
same boundary.
3. Although no earthquakes have been instrumen-
tally detected in the area, the sharpness and continu-
ity of the fault indicate that it has been active in the
Holocene. The fault has most likely been active in
many earthquakes but assuming it ruptured in only
one event, its magnitude would have been close to
Mw=6.7.
4. The offset of fault segments we observed in the
field was in the range 2–9 m. The higher number
might be an overestimate because of erosion due to
snow melting.
5. We found a sharp offset in a moraine formation
in a part of the Kerlingar fault, which shows that the
fault has been active since the Pleistocene glacier dis-
appeared from the area.
6. Considering three possible explanations for the
formation of the Kerlingar fault, we conclude that
the most likely process is differential movements due
to deglaciation, isostatic rebound and variations in
crustal buoancy. Lower viscosity of the lower crust
and uppermost mantle induces faster rebound within
the NVZ and buoyancy generates higher uplift of the
NVZ than the adjacent EFB, explaining why the Kerl-
ingar fault is situated at the NVZ-EFB boundary, why
it is parallel with the boundary, and why it is so long
and continuous. Other faults on the NVZ-EFB bound-
ary may be formed in a similar manner. Magma may
have intruded some of them, forming the arcuate pat-
tern of hyaloclastite ridges at the NVZ-EFB boundary.
114 JÖKULL No. 60