Jökull - 01.01.2021, Qupperneq 66

Magnússon et al.
riod 1755–2016 was close to the one required to main-
tain the present elevation at these ice divide, while it
was above this value in the period 1755–1918, result-
ing in net elevation gain and below it in 1918–2016
resulting in net lowering.
The good agreement between the depth of the
1755 layer from the static modelled velocity field and
the depth of the deep tephra layer, which independent
of the model result is quite likely to be formed by the
1755 eruption of Katla, raises the question: How old
is the ice below the tephra layer, which we mark as
1755? With the same method, we would approximate
the elevation of ice made from snow falling on the
glacier in 1200 AD to be somewhat less than 100 m
from the glacier bed. However, our model does not
include basal melting, so older ice will not be consid-
ered. Given the good agreement between the mod-
elled and observed depth of tephra layers, as well
as recent work estimating basal melting of Icelandic
glacier (Jóhannesson et al., 2020) we consider it un-
likely that the annual basal melt rate beneath these ice
divides (Figure 8b and 8d) is higher than 0.1 m a−1
(90 m of ice since 1200 AD); at present no clear caul-
drons formed by subglacial geothermal activity are
at this location. Our age estimate should, however,
be considered with caution due to the assumption of
fixed glacier surface; the glacier is expected to have
been substantially thinner for a good part of the pe-
riod since 1200 AD (e.g. Björnsson, 2017). With that
in mind we still consider it likely that the age of the
ice near the bed at the ice divides is 600–800 years
or even higher. Applying time-dependent models for
Mýrdalsjökull with varying geometry and mass bal-
ance should provide further constraints on this. Mod-
els aimed to simulate the development of Mýrdalsjök-
ull and its mass balance in the past centuries should
ideally use the constraint given by the depth, at the ice
divides, down to the 1918 tephra layer as well what we
identify as the 1755 tephra layer.
CONCLUSIONS
The comprehensive RES-surveys presented here have
resulted in a revised bedrock DEM of the Katla
caldera, which is unprecedented in terms of details
for an ice covered volcano, particularly in areas where
dense survey profiles allowed for 3D migration of the
RES-data. It has also resulted in unique maps, show-
ing the topography of two tephra layers, buried in ice,
that have been shaped by subglacial geothermal activ-
ity and high surface mass balance rates, since 1918
for the younger tephra layer and since 1755 for the
older layer, according to our dating. In addition to pre-
senting these new results and describing in detail how
they were obtained, we have listed and discussed var-
ious interesting features revealed in the bedrock and
tephra layer topography. Further studies are required
in many cases to understand some of these features,
including various topographic structures beneath the
glacier, or whether some observed shapes in the tephra
layers are caused by geothermal activity, ice dynam-
ics or spatially variable surface mass balance. The
presented data sets will serve as vital input in various
new studies, aiming to better understand subglacial
geothermal activity, ice cauldrons, ice dynamics, sub-
glacial hydrology and jökulhlaups, as well as the link
between all these processes. Such studies have al-
ready started (e.g. Jarosch et al., 2020). Furthermore,
the detailed bedrock DEM and the topographic map-
ping of the dated isochrones within the glacier may
serve as key data for studying the development and
mass balance of Mýrdalsjökull in the past centuries.
Addendum: The bedrock topography at the esti-
mated location of the 1918 eruption
Results from a recent study (Gudmundsson et al.,
2021) have led to relocation of the 1918 eruption to
a site outside the span of the RES-data acquired in
2012–2019. It was therefore decided to carry out more
RES-surveying in this part of the Katla caldera in May
2021. This included both a ∼1 km2 area, surveyed
with dense RES-profiles allowing for 3D migration,
as well as 15 km of 2D migrated profiles (Figure 12a).
The survey was carried out after the acceptance of this
paper for publication in Jökull but despite this the new
RES-data has been incorporated into the text, figures
and tables of this paper, without changing any of the
main findings presented above.
The bedrock map of the area specifically mea-
sured in 2021 (Figure 12b) is also shown as 3D rep-
resentation in Figure 10b and on the cover image of
64 JÖKULL No. 71, 2021