Jökull - 01.01.2019, Blaðsíða 68

The 2011 unrest at Katla volcano
aerial activity, because the flanks are (currently) to a
large degree ice-free. This deeply influences what sort
of hazards can be expected at Katla in case of vol-
canic reactivation. In light of this, further studies are
needed to better constrain the volcanic history and en-
vironment of eruptions and intrusions on the southern
flank, preferably combined with radiometric dating.
The potential involvement of magma during the
unrest episodes is not supported by clear signs of sur-
face deformation in the recent activity of Katla. This
might be explained by very shallow processes respon-
sible for the persistent seismicity, sometimes evolving
into real seismic crises. If, for example, a shallow
magma chamber exists, as postulated by Guðmunds-
son et al. (1994) and supported by more recent studies
(Budd et al., 2016; Jeddi et al., 2016), and the unrest
episodes are related to small batches of magma reach-
ing or approaching the surface, the associated defor-
mation signal might be negligible and hidden beneath
the glacier. In this scenario, it is important to note that
the 1955, 1999 and 2011 unrest episodes occurred in
different parts of the caldera, indicating that a large
portion of the volcanic system has been activated re-
cently.
Although no evidence of large-scale inflation is
currently observed at Katla, the possibility that the
volcano is preparing for an eruption cannot be dis-
carded. It is not unlikely that the edifice is already
in an inflated state that was reached before the defor-
mation measurements started. At Krafla, for exam-
ple, inflation of the edifice immediately followed the
1975–1984 rifting episode, and then since 1989 the
volcano has been slowly deflating, but the pressure of
the magma system is currently on the same order as
before the rifting episode (Buck et al., 2006). The sce-
nario at Katla may be similar and rapid inflation might
have occurred after the 1918 eruption. Accordingly,
the persistent seismicity may reflect the processes tak-
ing place in a pressurized system persistently close to
failure. The 1955, 1999 and 2011 episodes might be
signs of this. If this was the case, these unrest episodes
would represent important warnings, because they oc-
curred suddenly with no precursory activity and could
have triggered a larger eruption, or could do so in the
future.
CONCLUSIONS
The 2011 unrest episode at Katla was character-
ized by a sudden jökulhlaup, tremor bursts originating
at the source of the water flood (where ice cauldrons
deepened), increased earthquake activity within the
caldera and the appearance of a new source of persis-
tent earthquakes on the southern flank. This marked
a clear change of the long-term seismicity pattern, as
outlined by the analysis of the 1998–2015 seismic cat-
alogue, that likely highlights a modification of the vol-
canic system, possibly involving magma movements.
The 2011 unrest, similarly to the 1955 and 1999
episodes, was likely associated with hydrothermal
processes or magma movement, possibly a small sub-
glacial eruption. All unrest episodes were apparently
accompanied by only small deformation fields. In the
absence of evidence for recent large-scale recharging
of a magma chamber beneath Katla, it may be spec-
ulated whether Katla is already in an inflated state,
having inflated immediately following the most recent
large eruption in 1918. This may be the cause of the
persistent seismicity at Katla.
During a pilot study of the geology of the southern
flank in the source area of the new seismicity we have
found evidence that flank eruptions occurred during
the pre-Holocene (to recent) history of the volcano.
We have identified two sources in particular, one cor-
responding to the Gvendarfell ridge, and the other
one to Gæsavatn lake, documenting subglacial to surt-
seyan, basaltic to rhyolitic volcanic activity. This
highlights new scenarios for Katla’s eruptive history
and future activity and suggests that magmatic intru-
sions on its flanks should not be neglected as a poten-
tial seismic source.
Acknowledgements
The authors would like to thank the Icelandic Meteo-
rological Office for access to catalogue data from the
permanent seismic network. We thank Hugh Tuffen,
whose valuable suggestions helped improve the anal-
ysis of the geological data. This work was funded by
the University of Bologna, University of Iceland and
Uppsala University, as a part of a joint PhD project.
We thank the sheriff in Vík for logistic support. The
Generic Mapping Tool (Wessel et al., 2013) was used
to produce some of the graphics.
JÖKULL No. 69, 2019 67