Jökull - 01.01.2010, Síða 8

Geirsson et al.
are shown in Figure 3. The observed station veloc-
ities presented here are estimated after the time se-
ries have been corrected for offsets due to earthquakes
or equipment changes. Data from stations north of
Vatnajökull, affected by the Upptyppingar intrusive
episode (see below), are not used when estimating the
velocities shown in Figures 1 and 4, nor are data span-
ning the Eyjafjallajökull episode. Figure 1 shows the
average station velocities from the installation of the
stations until January 2010 with respect to a stable
North-American plate, and Figure 4 shows the same
velocities with respect to a fixed Eurasian plate. Euler
rotation poles from the MORVEL model (DeMets et
al., 2010) were used to transform the velocities from
ITRF2005 to stable North-America and Eurasia (Fig-
ures 1 and 4).
RESULTS
The plate boundary
The plate boundary in Iceland is composed of several
rift segments and transform zones connecting the rift
segments (Figure 1), as a result of interaction of the
Icelandic hotspot with the mid-Atlantic plate bound-
ary. The Mid-Atlantic ridge is spreading at a full
rate of a little less than 2 cm/yr in Iceland, accord-
ing to plate motion models based on either geologi-
cal and/or geodetic observations (e.g. DeMets et al.,
1994; Kreemer et al., 2003; DeMets et al., 2010). Re-
constructions of the plate spreading rates from mag-
netic anomalies indicate constant rates since about
6.5 to 7.5 Ma (Merkouriev and DeMets, 2008). The
spreading rate varies along the plate boundary (e.g.
DeMets et al., 2010) and the rate is slightly higher in
south Iceland than in north Iceland (Figure 1). The
observed CGPS velocities generally agree reasonably
well with plate motion model predictions (Figures 1
and 4) though formally many of the individual sites
do not agree with the MORVEL model. This dis-
crepancy is because many of the sites are located
within the plate boundary deformation zone and thus
are moving at intermediate rates, or the sites are af-
fected by other processes such as glacial rebound or
volcanic deformation. For example sites VMEY and
RHOF are likely still slightly affected by the proxim-
ity to the plate boundary (Figure 1), and the horizon-
tal velocity of site HOFN (higher than the MORVEL
prediction) is affected by the glacial rebound. Site
HEID, on the other hand, should be less affected by
glacial rebound and indeed its velocity agrees better
with the MORVEL model (Figure 4). The most com-
plete study to date of the plate boundary in Iceland is
published in Árnadóttir et al. (2009), where they use
the velocity field from the ISNET 1993 and 2004 na-
tionwide campaigns, along with available CGPS data.
Their results agree fairly well with predicted rates, ex-
cept the observed spreading rate is slightly higher in
the Northern Volcanic Zone. While the CGPS net-
work is still too sparse in many places to constrain the
over-all spreading of the plate boundary, it does pro-
vide important observations, especially where there
are transient changes in the site velocities and/or co-
seismic offsets.
The plate boundary on the Reykjanes peninsula
is highly oblique, where the on-land continuation of
the Reykjanes ridge connects to the Western Volcanic
Zone and the South Iceland Seismic Zone (SISZ).
Geodetic studies show that the plate spreading across
the peninsula is accommodated by left-lateral shear
(17–19 mm/yr) and a significant component of open-
ing (7–9 mm/yr) below a locking depth of 6–9 km
(Árnadóttir et al., 2006; Keiding et al., 2008). The
South Iceland Seismic Zone is an east-west trending
transform zone. The deformation in the brittle part of
the crust (above 10–15 km) is taken up by many paral-
lel north-south trending, right lateral strike slip faults,
a fault configuration that has been called "book-shelf"
faulting (Einarsson et al., 1981; Einarsson, 2008).
Several sequences of magnitude 6–7 earthquakes are
documented in the SISZ (Stefánsson and Halldórsson,
1988), with the latest sequence starting in the year
2000, as described below.
The Eastern and Western volcanic zones have
been observed to divide the plate spreading between
each other (LaFemina et al., 2005; Sigmundsson et
al., 1995). The CGPS stations in the CHIL subnet-
work can be used to constrain the interplay of the vol-
canic zones and the role of the Hreppar block (Figure
1). The CGPS stations indicate low spreading rates
across the WVZ and we observe a small increase in
8 JÖKULL No. 60