Jökull - 01.01.2016, Page 2

Eyjólfur Magnússon et al.
entists realised that by using radar with frequency in
the range of a few MHz, temperate ice could be pen-
etrated as well (Watts and England, 1975; Ferrari et
al., 1976). Icelandic scientists became pioneers in the
development of instruments to survey the thickness of
temperate glaciers (Björnsson, 1977; Sverrisson et al.,
1980). Since 1980, the Icelandic ice caps have been
systematically surveyed with RES (e.g. Björnsson et
al., 1987; Björnsson and Einarsson, 1990; Björns-
son et al., 2000; Björnsson, 2009; Magnússon et al.,
2012).
The density of RES-measurements is often sparse,
particularly for large ice caps and ice sheets (e.g.
Bamber et al., 2013; Fretwell et al., 2013). The sur-
veys are usually on profiles with sufficient resolution
along profile but with considerable distance between
the profiles. The profile density is improving rapidly
due to better survey technology (both instruments and
logistics) and various ongoing missions such as Ice-
Bridge (e.g. Li et al., 2013). The under-sampling
of topographic features with RES surveys is how-
ever still a problem for most glaciers. The existing
bedrock DEMs of all the large ice caps in Iceland are
based on RES profiles, commonly with ∼1 km profile
separation, often with significant gaps in inaccessible
crevassed areas. In an attempt to construct realistic
bedrock DEMs from the under-sampled data set, the
DEMs have been created from manually drawn con-
tour lines using the a prior knowledge of likely land-
scape in volcanic regions eroded by glaciers. This
method avoids many artefacts created by directly ap-
plying interpolation algorithms. The under-sampling
is however the largest source of elevation error in the
bedrock DEMs.
In this study we present the first subglacial topog-
raphy and glacier thickness maps of Drangajökull ice
cap. The survey and data post-processing were im-
plemented to extract more bedrock details than previ-
ously attempted for Icelandic ice caps. Typical sepa-
ration between RES profiles was 200–300 m revealing
the landscape beneath the glacier in unusually high
detail. We describe the data processing, the construc-
tion of the bedrock DEM and associated uncertainties.
We extract key dimensions of the ice cap and indi-
vidual ice catchments, including mean and maximum
thickness as well as the net ice cap volume. The net
volume of Drangajökull and how it has evolved since
1946 is derived using the new bedrock map and avail-
able surface DEMs (Magnússon et al., 2016). The
water catchments of rivers, draining meltwater from
the ice cap, are delineated. We also discuss some
unique features of Drangajökull ice cap, which the
bedrock mapping either reveals or helps explain.
STUDY AREA
Drangajökull ice cap, the fifth largest ice cap in Ice-
land, covered 144 km2 in 2011 (Magnússon et al.,
2016). It is located in NW-Iceland (Figure 1), where
the snow line altitude is lowest in Iceland; almost the
entire ice cap is currently at altitude below 900 m
a.s.l. The frontal locations of three outlet glaciers
of Drangajökull, Kaldalónsjökull, Leirufjarðarjökull
and Reykjarfjarðarjökull have been monitored an-
nually since 1931 (Sigurðsson, 1998; http://sporda-
kost.jorfi.is/). These records have shown retreat in-
terrupted with periods of advances during surges; the
three outlets are all surge type glaciers (Björnsson et
al., 2003; Brynjólfsson et al., 2015). Despite this
early effort, which is ongoing (Einarsson and Sigurðs-
son, 2015), Drangajökull was until recently among the
least studied ice caps in Iceland. This has changed in
recent years, starting with annual monitoring of the
mass balance at selected profiles on Leirufjarðarjökull
and SW-Drangajökull in 2005 (Sigurðsson, 2006) and
RES glacier thickness measurements at a few sites
(point surveys) (Magnússon et al., 2004).
The surface was mapped in high resolution with
laser scanning (Lidar) from an airplane 20 July 2011
(Jóhannesson et al., 2013). Various studies have been
carried out in this area to construct the past glaciation
history of Drangajökull using landforms in the periph-
ery of the ice cap and lake sediment cores. These
studies have focused on advances during surges in the
past 300 years (Brynjólfsson et al., 2015), glacial ex-
tent during the Holocene (Schomacker et al., 2016;
Harning et al., 2016a; 2016b) and the Little Ice Age
(LIA) maximum (Brynjólfsson et al., 2014; Harning
et al., 2016a). The most recent study on the LIA
maximum estimates a size almost twice the current
glaciated area (Harning et al., 2016a). A study of
2 JÖKULL No. 66, 2016