Jökull - 01.01.2019, Page 57
Sgattoni et al.
surface expression of subglacial geothermal activity
(Guðmundsson et al., 2007).
A zone with reduced P-wave velocities (with its
base at ∼3 km below the bedrock surface) and absent
S-waves was identified beneath the Katla caldera with
seismic undershooting and interpreted as evidence of
a magma chamber (Guðmundsson et al., 1994). This
is in agreement with an aeromagnetic survey (Jóns-
son and Kristjánsson, 2000) and a recent tomographic
study by Jeddi et al. (2016). A shallow magma
reservoir is consistent with geobarometric analyses by
Budd et al. (2016) that imply polybaric magma crys-
tallization pointing to simultaneous deep and shallow
magma storage. This is in contrast with previous geo-
chemical studies by Óladóttir et al. (2008) suggesting
the absence of shallow magma reservoirs in the cur-
rent plumbing system at Katla.
The Katla volcanic system has possibly been ac-
tive for several hundred thousand years (Jakobsson,
1979; Björnsson et al., 2000), producing FeTi-rich
alkali basalts and mildly alkali rhyolites, with very
subordinate intermediate rocks (Lacasse et al., 2007;
Óladóttir et al., 2008). Many outcrops along the
caldera rims and glacier margins are composed of rhy-
olitic lavas (Jóhannesson and Sæmundsson, 2009; La-
casse et al., 2007). The age of the caldera is unknown.
Seismicity
Despite its tectonic location outside the main rift
zones, persistent seismicity has been detected at Katla
since the first sensitive seismographs were installed
in Iceland in the 1960s (Einarsson and Brandsdóttir,
2000). Until the 2011 unrest, this seismicity has been
concentrated mostly within the caldera and imme-
diately to the west, at Goðabunga (Figure 1). The
caldera seismicity consists mostly of high-frequency
and hybrid events, probably associated with sub-
glacial geothermal activity and volcano-tectonic pro-
cesses (Sturkell et al., 2008). The Goðabunga cluster
consists mainly of low-frequency shallow events with
emergent P-waves, unclear S-waves and long low-
frequency coda. These events have a controversial in-
terpretation in the literature, either as a response to a
rising viscous cryptodome (Soosalu et al., 2006) or
due to glacial processes such as ice-fall events (Jóns-
dóttir et al., 2009). Recent discovery of a massive
landslide of about 1 km2 that has been active since at
least 1945 at the site of the proposed cryptodome (Sæ-
mundsson et al. 2020) calls for reinterpretation of the
possible sources of the Goðabunga seismicity. More-
over, volcano-tectonic microearthquakes are recorded
on the eastern flank of the volcano beneath the surface
at around 3.5 km depth, near the tip of Sandfellsjökull
glacier (Jeddi et al., 2017).
The Katla seismicity also shows a seasonal vari-
ation, particularly at Goðabunga, where the seismic
activity peaks in autumn. A less pronounced peak
of seismicity in the caldera occurs instead during the
summer (Jónsdóttir et al., 2007). This seasonal cor-
relation has been interpreted as a result of ice-load
change and resulting pore pressure variation at the
base of the glacier (Einarsson and Brandsdóttir, 2000)
or due to enhanced glacial motion during periods of
distributed subglacial water channels (Jónsdóttir et
al., 2009).
Holocene volcanism
The Holocene volcanic activity at Katla has been char-
acterized by three main eruption types. The most
frequent are phreatomagmatic explosive eruptions
due to magma-ice interaction below the glacier that
produced jökulhlaups and widespread tephra layers
(0.02–1.5 km3 volume; Thorarinsson, 1975; Larsen,
2000). At least 300 subglacial explosive eruptions are
known during the Holocene, with 20 in historic times,
i.e. since about 900 AD (Óladóttir et al., 2005). The
least common are effusive basaltic eruptions along the
fissure swarm in the ice-free part of the volcanic sys-
tem (8–10 in the Holocene). These include the two
largest eruptions of AD 934–40 Eldgjá Fires (19.6
km3; Thordarson et al., 2001) and ∼7.7 ka Hólmsá
fires (≥5 km3; Larsen, 2000). The third type of ac-
tivity consists of explosive silicic eruptions from the
central volcano that produced tephra fallout (<0.01–
0.27 km3; Larsen et al., 2001) and probably jökul-
hlaups (Larsen, 2000). It is not established whether
these events also generated effusive silicic products
that are exposed on the volcano today. At least 12 sili-
cic tephra layers are identified in the 1.7–6.6 ka time
interval between the Hólmsá and Eldgjá fires (Larsen,
2000). A minor silicic component was erupted dur-
ing the Eldgjá fires (Einarsson et al., 1980), but sili-
56 JÖKULL No. 69, 2019