Jökull - 01.01.2005, Page 100

Heidi Soosalu and Páll Einarsson
HE at the flank of the volcano is taken as the end of
the eruption.
After the initial swarm of earthquakes on January
17, 1991, earthquake activity at Hekla and the Hekla-
Vatnafjöll area was modest during the eruption (Soos-
alu and Einarsson 2002). On January 19–February 21
fourteen events (ML 1.1–2.6) were observed in the
Hekla-Vatnafjöll area, nine of them at Hekla proper.
At the end of the eruption, on March 11, the analogue
station HE, on the flank of the volcano, recorded a
swarm of about thirty small events before noon. They
likely represent conduit collapse after the volcanic ac-
tivity had ceased. The depths of the Hekla events after
the onset day, about 8–12 km, were similar to typ-
ical events in the eastern part of the South Iceland
seismic zone. All the earthquakes recorded by the
SIL network during the eruption at Hekla were high-
frequency tectonic events with distinct S-phases, none
of them looked like low-frequency volcanic earth-
quakes (Chouet 1996). During later phases of the
2000 eruption, only one earthquake, with a size ofML
0.8 was detected on March 1. It does not have a well-
constrained location, due to large gaps between the
stations.
DISCUSSION
During non-eruptive times the few earthquakes which
occurred at Hekla do not have an apparent correlation
to Hekla as a volcano. Instead, the seismicity in the
area around Hekla and the Vatnafjöll volcano to the
south have the same characteristics. The earthquakes
cluster loosely along two N-S lineaments and occur
mainly at 8–13 km depth, similar to the distribution
of seismicity at the eastern end of the South Iceland
seismic zone. A magnitude 5.9 (Mw) earthquake oc-
curred in the SW part of Vatnafjöll in 1987. Its fault
plane solution showed right-lateral strike-slip faulting
on a N-S striking fault, i.e. characteristics of South
Iceland seismic zone earthquakes (Bjarnason and Ein-
arsson 1991). It was thus discovered that “bookshelf
faulting”, the seismicity pattern of the seismic zone,
continues to the east as far as western Vatnafjöll, some
10 km further east than the surface expression of the
seismic zone. A portion of the earthquakes in our
data set occurred on the same lineament as the Vatna-
fjöll earthquake with its fore- and aftershocks. An-
other, fuzzier, N-S lineation can be discerned further
east, through the central parts of Hekla and Vatnafjöll.
Fault plane solutions for five events in this area are
primarily of the strike-slip type (Soosalu and Einars-
son 1997). Thus, South Iceland seismic zone tecton-
ics extend well into the volcanic zone, according to
our observations, all the way to longitude 19◦40’W.
Depth estimates for Hekla earthquakes before the
onset of the eruptions (Kristín Vogfjörð and Sigurður
Th. Rögnvaldsson, unpubl. data; Soosalu et al. 2005)
point to a shallow origin for the first earthquakes. Al-
though it is likely that the initial earthquakes are re-
lated to stress changes caused by the intrudingmagma
reaching the surface, it is clear that they are not form-
ing a propagating front close to the tip of the intrusion.
We suggest that the lack of seismicity preceding
Hekla eruptions is evidence for a deep magma source.
The stress change related to a deep-seated, inflating
magma chamber is distributed over a wider area and
occurs aseismically until a dyke starts propagating.
We have studied seismic rays between SIL stations
and local earthquakes to look for signs of volumes
of magma (Soosalu and Einarsson 2004). We did
not find evidence for a substantial magma chamber
at Hekla in the volume we could cover, i.e. the depth
range of 4–14 km. This is in contrast with former geo-
physical studies which place a magma chamber un-
der Hekla at 5–9 km depth (Kjartansson and Grönvold
1983; Eysteinsson and Hermance 1985; Sigmundsson
et al. 1992; Linde et al. 1993; Tryggvason1994). New
interpretation of strain data by Sturkell et al. (2005a)
suggest a Hekla magma chamber at 11 km depth, with
a radius of 2 km, in line with our suggestion. Because
of scarce data we could not examine well the even-
tual existence of a molten volume in the uppermost
4 km under Hekla. However, Hekla lacks the typical
expression of a shallow magma chamber, such as per-
sistent microearthquake activity and geothermal sys-
tems, and it is thus considered unlikely.
Our method was restricted to volumes with di-
mensions larger than about 800 m (see Soosalu and
Einarsson 2004). If the Hekla magma chamber actu-
ally is located somewhere at 5–9 km, it must be too
small for us to detect. The amount of erupted ma-
100 JÖKULL No. 55