Jökull - 01.01.2013, Page 108

F. Fuchs et al.
ally contribute to scattering of the epicenter loca-
tions, since P-wave picks were not available for all
stations. The respective locations are shown as red
circles in Figure 2. The observed seismicity clus-
ters beneath the southern and north-eastern flanks of
Snæfellsjökull volcano, respectively with some addi-
tional events ranging towards the lower lands in the
North-West. Red circles in Figure 2 denote epicen-
ters located by phase picks on OLA, GLA and STA
recordings. Additionally, few seismic events showed
clear arrivals on both GLA and STA but could not
be identified at OLA due to large background noise.
The respective phase arrivals on GLA and STA allow
for two possible epicenter locations that are shown as
grey circles (shaded and solid) in Figure 2. However,
the potential epicenter locations towards the south
(marked as grey shaded) fall outside the volcanic com-
plex, whereas the alternative locations (grey solid) are
within the cluster at the north-eastern flank of the vol-
cano and appear more plausible. We therefore suggest
that these events belong to the same seismic source
volume and are located within the north-eastern flank
of Snæfellsjökull as well.
Figure 3. Cross section along the Snæfellsjökull vol-
cano as marked by the dashed line in Figure 2, show-
ing the depth of the seismic events as a function of lat-
itude. Uncertainties in the determination of the depth
are indicated by error bars and range from ±3 to ±5
km. – Snið gegnum Snæfellsjökul eftir strikalínum á
2. mynd. Sniðið sýnir dýpi skjálfta sem fall af norðlæ-
gri breidd. Óvissan í dýptarútreikningum er ±3 til ±5
km.
The dashed line in Figure 2 represents the trace
of the profile shown in Figure 3. Although hypocen-
ter depths are not well-constrained (the nominal depth
error ranges from 3 km to 5 km deep) and weak
P-waves complicated phase picking (see below) we
observe that most of the seismic activity originates
from depths of 9 to 13 km. Still, the poorly con-
strained hypocenter locations do not allow us to iden-
tify significant changes in depth. Note that includ-
ing only three seismometers with the given geome-
try and unclear P-arrivals render the resulting depths
highly uncertain and thus the hypocentral depth might
be much shallower than stated here (Tarasewicz et
al., 2011). The actual seismic wave velocities under-
neath the Snæfellsnes peninsula are likely to deviate
from the 1D SIL velocity model used for this study.
We therefore checked how hypocentral depths are af-
fected by different P-wave velocities.Yang and Shen
(2005) report a low velocity anomaly in the Snæfells-
nes area while in the same region Allen et al. (2002)
observe slightly higher wave velocities in the upper
crust. Thus, we reduced and enhanced P-wave ve-
locities by 0.25 km/s at all depths of the 1D model.
Reduced wave velocities result in hypocenters raised
to shallower depths by about 600–1000 m, while in-
creased velocities would place them about 200–1000
m deeper. Thus, the effect is of the same order as the
uncertainties on the depth.
Figure 4 shows the waveform and the spectrogram
of the strongest seismic event (Ml 1.1) recorded on
July 18th 2011, located beneath the Snæfellsjökull
volcano. The picks for phase arrivals of the P- and
S-wave are marked by solid red lines and are identi-
cal with the theoretically calculated arrival times for
the given hypocenter. Figure 4a shows waveforms
bandpass filtered at 4–12 Hz where the signal-to-noise
ratio is best, while for the same reason the spectro-
grams shown in Figure 4b were generated from f > 3
Hz highpassed data. The signal is dominated by fre-
quencies between 2–12 Hz and is most prominent at
5–6 Hz. In general, waveforms of all recorded seis-
mic events are comparable and show a similar low-
frequency content. Except for the Ml 1.1 event the
magnitude range of the measured seismicity is be-
tween Ml -0.5 and 0.7. Generally, S-wave arrivals
108 JÖKULL No. 63, 2013