Jökull - 01.01.2016, Side 13

The subglacial topography of Drangajökull ice cap, NW-Iceland
from 5 m to 20 m. The lower value is expected near
surveyed profiles in relatively gentle bed slopes under-
neath the ablation area, for glacier thickness <100 m
(Cgl well established from Figure 4). The higher val-
ues are expected in thick ice, relatively far from sur-
vey profiles and in steep slopes. The error of the bed
DEM (and thickness map) is likely to exceed 20 m in
some cases, particularly at locations where all criteria
to enhance errors are fulfilled.
A possible bias in Cgl over the whole ice cap also
results in errors for the integrated ice cap volume and
corresponding average thickness (Table 1). We do
however expect Cgl to vary by less than 2.5%. Other
errors are unlikely to contribute significantly to the in-
tegrated volumes. As explained in Data and Methods
we compensate for a possible bias caused by the 2D
migration with the choice of Cgl (Figure 4). Errors
caused by the interpolation are expected to be random
with close to zero mean and hence unlikely to result
in significant bias of the glacier thickness. Assuming
an uncertainty of 2.5% results in total ice volume of
15.4±0.4 km3 for Drangajökull in July 2011. If at-
tached ice and snow patches are included the volume
increases by less than 0.1 km3.
The ice and water divides of ice and river catch-
ments on the ice cap were delineated (Figures 8 and 9,
respectively). The details of outlining those are given
in Magnússon et al. (2012). The resulting area and
volumes of individual ice catchments, average and
maximum glacier thickness (relative to the glacier sur-
face in July 2011) are displayed in Table 1. Kaldalóns-
jökull ice catchment ranks in first place in terms of all
key dimensions, with a total volume of 4.6±0.1 km3
and mean thickness of 131±3 m. The maximum ob-
tained thickness of Kaldalónsjökull as well as Dranga-
jökull ice cap is 284±14 m. This thickness is obtained
at a surveyed profile with relatively gentle bed slopes,
hence we expect the uncertainty of this point to be
governed by uncertainty in Cgl (5%). Slightly higher
values between profiles crossing the thickest part of
the glacier cannot be excluded.
The water divides of the largest glaciated drainage
catchments are drawn assuming static water potential,
ϕ, with water pressure equal to the glacier overburden
pressure (see e.g. Paterson, 1994) resulting in:
Figure 9. The glaciated drainage basins of the 6 main
rivers draining from Drangajökull ice cap, assum-
ing full (black line) and no (red line) ice load. Ice
and snow patches attached to the ice cap are not in-
cluded. The largest branch of each labelled river is
shown with blue lines. – Vatnasvið 6 stærstu jökul-
áa Drangajökuls. Svartar línur afmarka vatnasvið
á jökli út frá æstæðu vatnsmætti þar sem gert er
ráð fyrir fullu ísfargi og rauðar línur vatnasvið fyrir
æstætt vatnsmætti án ísfargs. Stærstu kvíslar ánna eru
sýndar með bláum línum.
ϕ = ρw gzb+ρi gH
where zb is the bed elevation, H is the glacier thick-
ness, g = 9.82 m s−2 the acceleration due to gravity
and ρw = 1000 kg m−3 and ρi = 900 kg m−3 the den-
sity of water and ice, respectively. Water divides as-
suming no ice load were also drawn. This was done
because the ice of Drangajökull is at some locations
quite thin and therefore we inspect how the water di-
vides are likely to change if the long term thinning
of the ice cap continues (Figure 10, see Magnússon et
al., 2016 for more details). Pressure in water channels
beneath thin ice (few tens of meters) is likely often at
JÖKULL No. 66, 2016 13