Jökull - 01.12.1999, Blaðsíða 37

ing and graben formation across Eyjafjarðaráll within
the Tjörnes Fracture Zone (Flóvenz and Gunnarsson,
1991), where seismic reílection data show a system of
large-scale listric faults down to 4 km depth in a sed-
imentary basin. The fault curvature is clearly visible
in sediments about 1 km below the ocean floor. A few
of these faults may extend down to 9-12 km depth, i.e.
into the basement.
Listric faulting is also suggested on land for the
model of normal faults in both the upper and lower
crusts in Iceland (see i.e. Guðmundsson and Báck-
ström, 1991; Guðmundsson, 1992, 1996; Forslund
and Guðmundsson, 1992). Based on analyses of
fault swarms as well as theoretical considerations, this
model suggests that “ all normal faults in Iceland initi-
ate either from columnar joints when the lava pile has
become tilted, and/or from large-scale and smaller-
scale tension fractures which in turn develop from
joints” (Guðmundsson and Backström, 1991). The
model of Guðmundsson also implies that the progres-
sive tilting of the lava pile in Iceland (1° for every
150-170 m depth) causes the initially vertical colum-
nar joints in the lavas and the resulting normal faults
to gradually decrease in dip with increasing crustal
depth. As consequence of tilting of the lava pile the
columnar joints form sets that dip at variable angle
towards or away from the rift axis, generating poten-
tial (curved) normal faults which also dip in both di-
rections. According to Guðmundsson’s model (1992,
1995) tension fractures commonly change into nor-
mal faults at a crustal depth of 500 m and normal
faults should mainly occur at 500-1500 m depth, be-
low which dykes are the most dominant extension
fractures.
Although Forslund and Guðmundsson (1992)
state that “ listric faults do not, however, appear to be
common in the exposed lava pile of Iceland ” and also
that “ where vertical exposures reach several hundred
meters, the faults appear to be planar”, they claim that
“ Some of the large faults, with throws of tens of me-
ters or more, may, however, penetrate the entire crust
and are likely to be listric ”. They support this conclu-
sion by data from a short profile in the Tertiary moun-
tain Akrafjall in soutwestern Iceland (located within
50 km from the Borgarfjörður area) where “ the dip
of the lavas increases by 5° over a vertical distance of
200 m in the hanging wall of a fault with 48 m throw,
indicating a curved fault plane ”. Based on vertical ze-
olite zonation, the Akrafjall lavas have not been buried
deeper than 300 to 600 m (Franzson, 1978; personnal
communication, 1998) and do thus not represent pro-
cesses at more than 1000 m depth. Furthermore, as
discussed below, the data presented here do not agree
with this model.
The six Borgarfjörður profiles presented in this
paper are particularly relevant to fault geometry with
respect to depth as they span a 1500 m depth-range
from one of the deepest on-land exposures in Iceland.
An average paleoburial depth of 1.3 to 1.5 km is in-
ferred from mesolite/scolecite boundary and stilbite,
which are commonly present below the Hreðavatn un-
conformity (profiles 1 to 3), in the old basement lavas
of the SRZ. Generally, these zeolites are to be found in
the uppermost 1 km (Kristmannsdóttir and Tómasson,
1978), or more precisely, from around 800 m depth
down to more than 1700 m, according to “Walker's
vertical zeolite zonation” (Walker, 1960). The pro-
files, also being located on each side of the Borgarnes
anticline, promote our understanding of the nature of
crustal tilting towards rift zones.
Listric faults
The bulk of the Borgarfjörður data shows that faults
dip from 50° to 90° on both sides of the Hreðavatn
unconformity (Figures 7a and 7b). On both sides of
the unconformity, only 9 faults (including two exam-
ples discussed below) dip as shallow as 50° to 65°.
These faults have a throw ranging from less than a
metre to 4.5 m. Apart from these few small-scale
faults which have very limited extent, no significant
difference in fault dip is detected in a vertical section
down to a paleo-depth of 1.5 km. Neither large-scale
listric fault nor deformation associated with it, such as
subsidence, vertical block rotation and roll over were
observed. If listric faults exist in the Borgarfjörður
region, they must thus be curved below 1.5 km depth.
Two examples of small-scale listric normal faults
(with throw less than 1 m) were observed on both
sides of the unconformity (Figures 6c and 6d). These
faults dip 70° in their upper parts but become curved
JÖKULL, No. 47 35