Jökull - 01.01.2010, Page 13

Fifteen years of CGPS in Iceland
edifice subsides slightly relative to sites away from
Hekla, which is in a general agreement with the In-
SAR observations.
Eyjafjallajökull and Katla
The 2010 eruption at Eyjafjallajökull caught wide-
spread media attention because of the severe effects
of the ash-plume on air travel. The eruption is but a
part of a long chain of events leading up to the erup-
tion. The most recent eruption of Eyjafjallajökull be-
fore 2010 occurred in 1821–1823, and the volcano
experienced two inflation episodes in 1994 and 1999
(Pedersen and Sigmundsson, 2004, 2006; Sturkell et
al., 2010). The CGPS site THEY, south of Eyjafjalla-
jökull, was originally installed to monitor the 1999
intrusion, but the episode ceased before the site was
installed in May 2000.
In July 1999 a jökulhlaup emerged from the south-
ern part of the neighboring Mýrdalsjökull ice cap,
that covers the Katla volcano. Katla erupted last in
1918, and is historically known for more frequent and
more violent eruptions than Eyjafjallajökull. Follow-
ing the 1999 jökulhlaup, cauldrons at a few places un-
der the icecap were observed to deepen, indicating in-
creased heat flow from the volcano (Gudmundsson et
al., 2007). Episodic GPS measurements at nunataks
showed steady inflation of Katla between 1999 and
2004, and were explained by inflation of a magma
chamber at depths of 2–5 km (Sturkell et al., 2008).
From 2004 onwards seismic activity has been at lower
levels. The station SOHO south of Katla moves out-
ward from the caldera at a rate of ∼6 mm/yr in excess
of plate movements (Figure 1). This rate decreased
slightly at the 2004 transition (Sturkell et al., 2010),
but the site continues to move outward from the Katla
caldera.
In May 2009 intrusive activity resumed under
Eyjafjallajökull, which had remained quiet since the
1999 intrusive episode. Seismic activity increased
and subtle surface deformation was observed at the
CGPS station THEY (Figure 6). The activity contin-
ued until mid-August 2009, when both deformation
and seismicity halted. During this intrusive episode,
THEY moved southwards by about 15 mm, which is
about 10 times less than displacements of a nearby
episodic GPS site during the 1999 intrusion (Sturkell
et al., 2003a). Towards the end of the 2009 episode,
two new semi-continuous stations were set up in the
area, and four more were installed in the spring of
2010 (Sigmundsson et al., 2010).
In December 2009 seismic activity and deforma-
tion resumed, in a more intense manner compared to
the summer’s activity. In early March 2010 a change
in the surface deformation and seismic activity indi-
cated the formation of an ESE-striking dike east of the
summit (Sigmundsson et al., 2010). On 20 March the
magma broke its way to the surface at the eastern flank
of the volcano, mid-way between the Eyjafjallajökull
and Katla volcanoes. The flank eruption was hawaiian
style and confined to a small area. The eruption did
not seem to relieve much of the pressure previously
built up by the intrusions, as the observed co-eruptive
deformation was subtle (Figure 6). On April 12th,
the flank eruption ceased. On April 14th 2010 a new
phase of the activity started with the summit erup-
tion. This phase was much more explosive, and the
co-eruptive deformation rates were larger than during
the flank eruption (Figure 6), indicating that magma
pressure was released more efficiently than during the
summit eruption.
Grímsvötn eruption 2004 and subsequent inflation
Grímsvötn is Iceland’s most frequently erupting vol-
cano in recent times (Thordarson and Larsen, 2007).
It erupted last in 2004 and before that in 1998. It
is of great interest to try to cast light on the magma
movements of the volcano and try to forecast when
the next eruption could occur. The Grímsvötn caldera
is in the interior of the Vatnajökull ice-cap and is
mostly covered by ice with only one nunatak suitable
for a GPS site, located on the SE-rim of the caldera.
Episodic GPS measurements were occasionally made
at the nunatak, capturing the deformation associated
with the 1998 eruption (Sturkell et al., 2003b) and
subsequent inflation (Sturkell et al., 2006).
By 2004, episodic GPS measurements and seismic
observations indicated that the volcano had reached
its pre-1998 eruption state. In June 2004, a CGPS sta-
tion (GFUM) was installed on the nunatak on a tem-
porary monument 500 meters from the episodic site.
Due to the intense atmospheric icing conditions on the
JÖKULL No. 60 13