Jökull - 01.01.2019, Qupperneq 50

Historical accounts of pre-eruption seismicity in Iceland
for the several decades of seismic monitoring (e.g.,
Einarsson, 1991a; Einarsson and Brandsdóttir, 2000;
Jakobsdóttir, 2008; Jónsdóttir et al., 2007). Hekla,
on the other hand, is seismically very quiet between
eruptions (Soosalu and Einarsson, 2005). This differ-
ence in long-term behaviour may be explained by the
depth of the magma storage of the two volcanoes. The
caldera region of Katla is underlain by low-velocity,
high attenuation material at shallow level, as revealed
by undershooting and 3-D tomography (Gudmunds-
son et al., 1994; Jeddi et al., 2016). At Hekla, on the
other hand, attempts to find such volumes have been
unsuccessful (Soosalu and Einarsson, 2004) and the
pressure source responsible for inflation and deflation
of the volcano in connection with recent volcanic ac-
tivity have been shown to be at depths of 10-24 km
(Ófeigsson et al., 2011; Geirsson et al., 2012; Sturkell
et al., 2013).
The unusually short precursor time for Hekla is
difficult to understand in light of the large depth to
the magma pressure source. A speed of 40 km/hour
is required if magma propagating from a depth of 20
km reaches the surface in half an hour. Furthermore,
the first events recorded prior to the eruption in 2000
had the signature of very shallow earthquakes, even
though the depth could not be determined with cer-
tainty (Soosalu et al., 2005). A solution to this enigma
was proposed by Sturkell et al. (2013), who suggested
that the time between eruptions of Hekla was not long
enough for the magma in the conduit to solidify. The
magma feeding the initial phase of the next eruption
therefore came from the top of the conduit at shallow
depth, and not from the deep pressure source.
The high seismicity that seems to accompany
eruptions of the Hekla volcanic system when the erup-
tion site is outside the main edifice remains enigmatic.
These eruptions take place on fissures, are mainly ef-
fusive and produce basalt (Jakobsson, 1979) as op-
posed to the more andesitic and dacitic composition
of typical Hekla products. Perhaps the intense seismic
activity reflects the longer path of the magma through
the brittle part of the crust, whether by lateral propa-
gation from the central volcano or by vertical propa-
gation through the cooler crust surrounding the central
volcano.
Seismic activity preceding the two eruptions of
Öræfajökull volcano in SE Iceland in Historical time,
in 1362 and 1727, is not well documented. Both erup-
tions appear to have been preceded by seismic activ-
ity, however, sufficient to alarm the local population.
This understanding is important now because of the
recent re-awakening of this dangerous volcano, in-
crease in seismicity and new geothermal activity, that
is associated with inflation of the volcano (Geirsson et
al., 2018).
CONCLUSIONS
A systematic study of historical documents describing
the beginning of volcanic eruptions in Iceland con-
firms the general public knowledge that most erup-
tions are accompanied by earthquakes. These data
complement and add to the instrumental data for the
eruptions of the last half century that show that all
eruptions have a detectable short-term precursory ac-
tivity (Einarsson, 2018). This activity can be used to
issue short-term warnings of an impending eruption if
detected and recognized soon enough.
There are reliable descriptions of pre-eruption
seismicity for the Katla volcano for the seven sig-
nificant eruptions during 1625–1918. These are the
only data for this important and hazardous volcano.
No eruptions large enough to penetrate the ice cover
of the Katla caldera have occurred after sensitive in-
struments were installed to monitor the volcano. All
seven eruptions were preceded by felt earthquakes
and were accompanied by catastrophic floods, jökul-
hlaups, from the caldera. The precursor times, i.e.
time from the first felt earthquake until the eruption
was confirmed, range between one and nine hours.
The time from the first felt earthquake until the flood
wave reached the inhabited area, the jökulhlaup time,
was always longer than two hours and in one case
more than 12 hours.
Reliable descriptions exist of ten eruptions of
Hekla during the pre-instrumental period 1510-1947.
Eruptions of Hekla fall into two categories regard-
ing seismic behaviour. Eruption from the main edi-
fice tend to be explosive and are generally accompa-
nied by mild seismicity with short precursor times. In
some instances (1510, 1597, 1693) the first felt earth-
JÖKULL No. 69, 2019 49