Jökull - 01.01.2014, Side 35

Volume estimate of nine prehistoric Katla tephra layers
in this study. The actual uncertainty in layer volume
is considerably higher since the offshore part is un-
accounted for and same applies to fallout outside the
mapped area on land.
DISCUSSION
Volume difference between historical and prehis-
toric Katla tephra layers
Volume estimates of selected prehistoric Katla tephra
layers confirm that layers produced in late prehistoric
time are larger than tephra layers produced during his-
torical time. The average size of the eight prehistoric
layers is 0.9 km3 (range 0.2–2.7 km3) whereas the his-
torical average is 0.4 km3 (range 0.04–0.8 km3). Ad-
ditionally, the tephra layer frequency during the his-
torical time is lower (1.0 layer/100 years) than during
the prehistoric period discussed here (2.8 layers/100
years; Óladóttir et al., 2005). It is worth noting that
the prehistoric tephra layers discussed here are among
the larger layers produced during this time period,
their average thickness in the Atley outcrop (Figure 1;
Óladóttir et al., 2005; 2008) being 12.8 cm whereas
the average thickness of all 45 Katla layers preserved
in the same outcrop is 3.5 cm. However, the eight
historical layers used for comparison are also among
the larger eruptions of that time period justifying the
comparison.
The observed volume difference between the two
time periods could be related to changes in the magma
plumbing system under the Katla volcano. Compo-
sitional changes of Katla tephra dating back ∼8400
years have been interpreted as a result of a cyclically
changing magma transfer system under the volcano.
It changes from a simple magma transfer system (rep-
resented by a stable concentration of K2O) to a sill-
and-dike system (represented by irregular K2O con-
centration) that then evolves into an active magma
reservoir showing increasing K2O concentration with
time (Óladóttir et al., 2008). These cycles influence
both composition of erupted material and eruption
frequency which is highest during periods controlled
by sill and dike transfer system and lowest when the
transfer system is simple (Óladóttir et al., 2008). The
historical time represents a simple magma transfer
system and the part of the prehistoric time under con-
sideration an active magma chamber.
To what extent does the calculated volume repre-
sent the total erupted volume?
In terms of eruption numbers the tephra-producing
explosive eruptions appear to have been dominant in
Katla over long periods (e.g. Larsen, 2000; Óladóttir
et al., 2008). There are no constraints on lava pro-
duction associated with the prehistoric tephra form-
ing eruptions and there are no known lava flows from
the time period covered in this study. Furthermore,
the tephra production and its sulphur content strongly
indicates the existence of the ice cap during the pre-
historic period studied here (Óladóttir et al., 2007)
probably preventing effusive eruptions taking place
in the Katla central volcano. However, 5–10 small
lava-forming eruptions are known, probably most of
them older than 4000 years (Jóhannesson et al., 1990;
Larsen, 2000). Therefore, the tephra productivity
of the nine estimated eruptions, ranging from 0.2–
2.7 km3 (Table 5), is the best available indicator of
magma productivity from the Katla volcanic system
in prehistoric time.
Still, it is highly likely that part of the erupted ma-
terial was transported away by large glacial outbursts,
jökulhlaups, accompanying the subglacial eruptions
as has been the case for the historical eruptions (e.g.
Tómasson, 1996). The difficulty with estimating vol-
ume of the water transported part is to know what pro-
portion of the material is newly formed, related to the
tephra forming eruption, and what is older material
transported with the melt water produced during the
eruption. Additionally, the size of the glacier cover-
ing the Katla central volcano at each eruption time
is not known, although high sulphur concentrations
in the tephra glass suggest high water/magma ratios
during the Holocene, best explained by a permanent
glacier (Óladóttir et al., 2007). Hence, it is difficult
to speculate about the size of the associated jökul-
hlaups and thereby the amount of water-transported
material. During the K-1918 eruption the volume of
water transported volcanic debris has been estimated
0.7–1.6 km3 (Larsen and Ásbjörnsson, 1995; Tómas-
son, 1996; Larsen, 2000), whereas the tephra layer has
been estimated 0.7 km3 (Eggertsson, 1919).
JÖKULL No. 64, 2014 35