Jökull - 01.01.2010, Page 11

Fifteen years of CGPS in Iceland
modate the right-lateral transform motion between
these two segments of the ridge, the Húsavík-Flatey
fault (HFF) and the Grímsey oblique rift (GOR) (Fig-
ure 1). A third zone, the Dalvík zone, is indicated
by seismicity about 30 km south of the HFF (Einars-
son, 1991, 2008) where earthquakes as large as Ms 7
have occurred, latest in 1963. These three zones com-
prise the Tjörnes Fracture Zone (TFZ). The last ma-
jor earthquake rupturing the on-shore part of the HFF,
where the fault runs through a town, occurred in 1872,
and another major earthquake may therefore possibly
be due on the fault system. It is thus important to es-
timate the slip rate, slip deficit and resulting moment
accumulation on the two fault systems to gain further
information about the seismic hazard in the region.
The observed GPS velocities in north Iceland show
a gradual increase as one crosses the TFZ (Figures 1
and 4). Preliminary modeling of the CGPS velocities
show that the distribution of motion between HFF and
GOR appears to be closer to 30/70 (Metzger and Jóns-
son, 2010) than to the 40/60 percent estimated earlier
(Geirsson et al., 2006) from a more sparse network.
Despite the low slip rate of the HFF, the accumulated
moment on the HFF since 1872 would still correspond
to an earthquake of a magnitude just below 7 (Metzger
and Jónsson, 2010).
Earthquakes
On June 17, 2000, an earthquake sequence started in
the South Iceland Seismic Zone with a magnitude 6.5
event. The event triggered earthquakes across south-
west Iceland, all the way west to the Reykjanes penin-
sula, with the largest triggered event being a magni-
tude 6.5 earthquake that occurred 81 hours later, lo-
cated 17 km west of the initial mainshock (Árnadóttir
et al., 2001, 2003, 2004; Geirsson et al., 2006). The
CGPS network was rather sparse at the time, but pro-
vided important constraints on the co-seismic offsets
from these events. The June 2000 sequence demon-
strated the need for automatic processing of the data
with low latency, which was initiated at IMO shortly
after the earthquakes. An earthquake doublet with
a composite Mw=6.3 struck the western part of the
South Iceland Seismic Zone on May 29, 2008. Within
a few seconds of the onset of the first mainshock, the
second event was triggered on a fault located 5 km
further west (Hreinsdóttir et al., 2009). A clear co-
seismic offset is observed in the time series at the
CGPS stations in the epicentral area, with a maximum
offset of about 20 cm (Figure 2).
The 29 May 2008 earthquake sequence is the first
event recorded by the high rate CGPS network in Ice-
land (Hreinsdóttir et al., 2009). The two mainshocks
occurred so close in space and time that the seismic
waves from the second rupture were embedded in the
coda from the first event, making precise location and
timing of the second event difficult using conventional
seismic data. Source models of the two main faults
were estimated from static offsets of the CGPS sta-
tions. Based on these source models and the high-
rate CGPS time series, Hreinsdóttir et al. (2009) con-
cluded that the second fault ruptured within 3 seconds
of the initial mainshock. Data from the CGPS net-
work therefore provided important constraints on lo-
cation, timing, and magnitudes of the earthquakes.
Post-seismic deformation was observed after both
the 2000 and 2008 earthquake sequences. During the
first months after the 2000 sequence, deformation was
most rapid in the epicentral area and dominated by
poroelastic rebound, as observed by InSAR (Jónsson
et al., 2003). A slower transient was observed from
annual campaign GPS measurements during the first
4–5 years following the main shocks (Árnadóttir et
al., 2005). Modeling of the GPS data indicates fairly
low viscosities, with 5–10×1018 Pa s in the lower
crust and 3×1018 Pa s in the upper mantle (Árnadóttir
et al., 2005). Subsequent analysis of InSAR data indi-
cate that the year-scale transient signal is more likely
due to visco-elastic relaxation in the lower crust and
upper mantle, rather than afterslip (Jónsson, 2008).
A ∼2 cm deformation transient was observed at the
CGPS station HVER in the week following the May
2008 earthquakes (Figure 2). The post-seismic defor-
mation following the May 2008 earthquakes appears
to be less pronounced than the signals observed after
the June 2000 earthquakes. This is perhaps not sur-
prising since the June 2000 earthquakes were signifi-
cantly larger events.
The main shocks in the June 2000 and May 2008
sequences released about half of the moment accumu-
lated by plate motion since the previous earthquake
JÖKULL No. 60 11