Jökull - 31.12.2001, Qupperneq 92

Steinunn S. Jakobsdóttir et al.
but it was possible to collect and save very valuable
data for later research. Several swarms were recorded
in SW-Iceland during the first half of 1991 while
the system was still not running in a fully automatic
mode, including 2 earthquakes larger than magnitude
4. On January 30 an earthquake of M/= 4.6 occurred
in the WVZ near Skjaldbreiður south of Langjökull
and on June 19 an earthquake of M/= 4.0 was recorded
in Ölfus (Figure 2), the largest event in a swarm last-
ing from June 17 to June 22. The next swarm activ-
ity was at Kleifarvatn on November 20, 1992 starting
with an M/= 4.0 earthquake preceded with a few earth-
quakes in the Kleifarvatn area during the night before.
On December 27, 1992 an M/= 4.3 earthquake oc-
curred in the Hengill area as the second earthquake in
a small swarm. In 1993 there was one swarm recorded
near Krýsuvík on Reykjanes Peninsula with 3 earth-
quakes reaching magnitudes larger than 3.
Except for the first half of 1991, these 3 first
years of the SIL system were rather quiet with some
5-20 events detected per day in the SISZ. The first
half of 1991 had swarm activity every month with
over 100 and up to 500 earthquakes per day, but af-
ter June 20 the only days with more than 100 earth-
quakes detected are mentioned above and the number
of events per day never reached 200. At the end of
year 1993, after a magnitude ~4 earthquake in Eyja-
fjörður, the SIL system was expanded with 6 stations
in NE-Iceland, which improved the coverage for all
of Iceland. Few additional earthquakes were detected
in other parts of Iceland during the first three years,
but they are not very well located as their epicenters
are well outside the network. They are therefore not
included here.
EARTHQUAKE EPICENTERS 1994-2000
The map in Figure 3 (page 92) shows epicenters of
selected earthquakes with magnitudes greater than or
equal to M/=1 during 1994-2000 as determined by
the SIL network. During this period some 145,000
earthquakes were recorded. The majority of the events
have their origin in the Hengill-Ölfus area, at the in-
tersection between the SISZ, the Reykjanes Penin-
sula and the WVZ, where about 85,000 earthquakes
were generated in an episode of crustal deformation
and high seismic activity during 1994-1998. The
Hengill-Ölfus swarm activity culminated with events
of M/~5 in June and November 1998 (Rögnvaldsson
et al, 1996; Sigmundsson et al., 1997; Feigl et al.,
2000). A large fraction of the seismicity occurred in
the TFZ, offshore northern Iceland. The largest earth-
quake there, Ms=5.3 occurred in February 1994, just
after the SIL network was installed in NE-Iceland.
The event was accompanied by high seismicity in the
period 1994-1997.
Seismicity in connection with eruptions has also
been observed during this period. Two eruptions in
Vatnajökull, Gjálp in October 1996 (Einarsson et al.,
1997) and Grímsvötn in December 1998 were pre-
ceded by increased seismic activity and confirmed by
pilots that were asked to survey the region. A waming
was issued to the National Civil Defence of Iceland
prior to both these eruptions. The Gjálp eruption was
preceded by an Ms=5.4 earthquake in Bárðarbunga
two days earlier. In May 1994 an Ms= 5.2 earthquake
occurred in Bárðarbunga. The largest event prior to
the Grímsvötn eruption was M/=3.8.
The Hekla eruptions in January 1991 (Gudmunds-
son et al., 1992; Soosalu and Einarsson 2002) and
February 2000 (Stefánsson et al., 2000a) had seis-
mic precursors that were first noticed by the ana-
log highland network operated by the Science Insti-
tute, University of Iceland (Einarsson and Björns-
son, 1987). Again, a warning was issued to the Na-
tional Civil Defence of Iceland prior to both these
eruptions. The increased seismicity activated the
alert system of the SIL network (Böðvarsson et al.,
1996) sorne 50 minutes before the surface rupture.
When strainmeters started to show an anomaly, in the
same way as known from the Hekla eruption in 1991
some 20 minutes before the onset, a prediction was
made with predicted time accuracy within 5 minutes
(http://hraun.vedur.is/ja/heklufrettir.html).
Volcanic events in Eyjafjallajökull and Mýrdals-
jökull have also been monitored. Eruptions have not
been observed directly in these cases, but geodetic
measurements confirm uplift at Eyjafjallajökull as-
sociated with the seismic activity (Sturkell et al, in
press) and a minor magmatic episode in Mýrdalsjök-
ull resulted in the formation of a new cauldron and a
90 JÖKULLNo. 51