Jökull - 01.01.2019, Side 85

Nikkola et al.
ing system beneath the volcano, and near-direct verti-
cal magma transport from mantle depths (Sigmunds-
son et al., 2010; Sigmarsson et al., 2011; Keiding and
Sigmarsson, 2012; Tarasewicz et al., 2012a,b; Laeger
et al., 2017).
Regardless of the advances above, our understand-
ing of crustal storage and evolution of magmas in
SEVZ is inadequate, with our best inferences relying
on the 2010 Eyjafjallajökull eruption alone. Analysis
of minerals with pressure- and temperature-sensitive
compositions offers a way forward, as it allows us
to make inferences on the conditions of magma dif-
ferentiation in the SEVZ crust. In this paper, we
present major element data on olivine, clinopyrox-
ene, spinel, and melt inclusions from the most prim-
itive Eyjafjallajökull volcanic units: the Brattaskjól
and Hvammsmúli ankaramites. We show that these
two ankaramites host primitive olivine and clinopy-
roxene macrocrysts as well as chromian spinel that is
the most Cr- and Ti-rich reported from Iceland. On
the basis of thermobarometric calculations and diffu-
sion modelling, we suggest that these crystals were
derived from agitated and disaggregated mid-crustal
(10.7±5 km) wehrlitic or plagioclase wehrlitic crystal
mushes and that they ascended from these depths in a
carrier magma within a few weeks only.
EYJAFJALLAJÖKULL VOLCANIC
SYSTEM
Eyjafjallajökull (Figure 1) is a glacier-covered, mod-
erately active volcano that has acted as a locus of
magmatism for over 0.78 Ma (Kristjánsson et al.,
1988). Typical of off-rift volcanic systems in Ice-
land, Eyjafjallajökull magmas are enriched in K2O
and Na2O compared to axial rift magmas (Jakobsson,
1972; Hémond et al., 1993) and show trace element
and isotopic signatures of "enriched" mantle source
(e.g., elevated 206Pb/204Pb in comparison to MORB)
(Peate et al., 2010). These characteristics are most of-
ten attributed to low-degree melting of incompatible
trace element rich and mineralogically distinct mantle
source compared to the depleted source of common
MORB (Chauvel and Hémond, 2000; Kokfelt et al.,
2006).
Observations of the Eyjafjallajökull 2010 eruption
build a strong case for a trans-crustal magma plumb-
ing system beneath the volcano, with defined magma
intrusions in the brittle upper crust (<10–12 km depth,
Hjaltadóttir et al. 2009) underlain by poorly con-
strained magma storage zones in ductile crust and up-
per mantle (Sigmundsson et al., 2010; Sigmarsson et
al., 2011; Keiding and Sigmarsson, 2012; Tarasewicz
et al., 2012a, b). Although Eyjafjallajökull has a shal-
low silicic magma intrusion at 5 km depth (Sigmars-
son et al., 2011), and intrusions of magma into the
shallow crust (e.g., 5 km beneath the eastern flank
of the volcano; Sigmundsson et al., 2010) predated
the eruption, the Eyjafjallajökull 2010 eruption was
fed by magma from considerable depths. During the
first weeks of the eruption, seismicity was focused
on the brittle crust (<10 km), suggesting a mid-crustal
magma source. However, later a downward propaga-
tion of earthquakes down to ∼30 km below the sur-
face was detected, with distinct seismic clusters at
depths of ∼19 km and ∼25 km, potentially reflecting
depressurization of two or more 1–10 km3 intrusions
at these depths (Tarasewicz et al., 2012b, a). This sug-
gests that the eruption tapped magma from the mantle
(from depths greater than ∼22 km; Brandsdóttir and
Menke, 2008). Exhaustion of the mid-crustal reser-
voirs and deep tapping of magmas supports views of
SEVZ as an embryonic rift segment (e.g., Mattsson
and Oskarsson, 2005), where crustal magma storage
zones are still small and maybe ephemeral (Sigmars-
son, 1996) in comparison to active rift zones with hot-
ter crust (Flóvenz and Saemundsson, 1993) and pre-
sumably larger crustal magma reservoirs.
SAMPLES
The volcanic units sampled for this study, Bratta-
skjól and Hvammsmúli (Figure 1), are the most prim-
itive volcanic units of the Eyjafjallajökull volcano
described so far (Loughlin, 1995). Both are partly
eroded outcrops on the southern slope of Eyjafjalla-
jökull and seemingly subaerial in character (Lough-
lin, 1995), although a shallow intrusive origin for
Hvammsmúli has also been suggested (Steinthórs-
son, 1964). We follow the established practise (e.g.,
Steinthórsson, 1964; Loughlin, 1995) and refer to
84 JÖKULL No. 69, 2019