Jökull - 01.01.2004, Side 23

Glaciological application of InSAR topography data of W-Vatnajökull
dividual rivers under normal conditions. Furthermore,
unknown, small-scale basal features may exist, such
as a canyon controlling the flow at some crucial point.
The derived water divide between the two branches of
Skaftá is most sensitive to failure in the static model
or due to unknown basin features at locations shown
in Figure 4.
To analyse the uncertain water divides in more
detail, the summer balance for both river basins of
Skaftá in 1996–2002 (Björnsson et al., 1998, 2002)
was calculated for various divides between the Skaftá
branches (Figures 2 and 4). The results are only dis-
played as ratios (Figure 5) as continuous measure-
ments for the two branches do not exist. The derived
ratios support the existence of the suggested break
point (star in Figure 4) at which the water from the
area south of the Skaftá cauldrons drains over to the
western branch of Skaftá.
The total discharge in Skaftá estimated from sum-
mer balance (Figure 5) is slightly less than discharge
surveyed by the National Energy Authority at Sveins-
tindur (Figure 3) during the summer months, exclud-
ing jökulhlaups (Zóphóníasson, 2002). The amount of
summer balance that percolates into groundwater and
reappears below Sveinstindur has been estimated 15–
20 m3/s, during the whole year (Snorrason and Sig-
urðsson, 2002). This means that roughly 50% of the
estimated summer balance should be measured as sur-
face water in the river at Sveinstindur, implying that
the difference in flow measured at Sveinstindur and
calculated water flow from summer balance is on av-
erage around 45–70 m3/s over the summer months,
which can hardly be explained with summer precip-
itation alone. A rough estimate of the summer pre-
cipitation derived by multiplying the precipitation at
Kirkjubæjarklaustur with the area of the sub-glacial
river basin of Skaftá gives around 20 m3/s on average
for the years 1996–2002.
The location of water divides is also sensitive to
glacier surface changes. That applies to the critical
points (Figure 2) and the break point on the derived
water divides between the Skaftá branches. Improved
maps of the basal topography around these points are
required to erase uncertainities in the determination
of water divides and river basins. However, even
very dense radio echo sounding profiles would not
reveal narrow canyons. Moreover, the static model
of the basal water potential does not display temporal
changes in the drainage system.
The location of the water divides for Tungnaá
shown in Figure 2 is quite different from older esti-
mates (Figure 6) based on the DEM in 1981 (Björns-
son et al., 1992b). The surge of Tungnaárjökull in
1994–1995 (Björnsson et al., 2003) decreased the
size of the Tungnaá river basin, from 149 km2 to
95 km2. The average estimated glacier summer bal-
ance of Tungnaá in 1996–2002, using the new divides,
is thus only 73% of the older measurements. The
difference for the years 1996–2002, which on aver-
age corresponds to a 13.5 m3/s average summer dis-
charge from June to September, and varies between
10 and 15 m3/s, currently drains into Skaftá. This
agrees with discrete measurements of the discharge
in Tungnaá close to the margin, which have given less
water runoff than expected from summer balance es-
timates with the old divides (Hannes H. Haraldsson,
pers. comm., 2003). The derived size of the glacier
part of the river basin of various rivers is shown in
Table 1 in Appendix. The given values for Djúpá,
Hverfisfljót and the Skaftá branches are adjusted to
discharge measurements.
Pathways of main sub-glacial watercourses
The predicted basal water potential may be used to
map the location of the main sub-glacial watercourses.
Even though the low discharge condition is violated in
these watercourses during the summer one may argue
that they start forming in the spring while there is still
little water at the base of the glacier and that the water
maintains a tunnel system during the summer. Figure
7 shows the estimated main watercourses of Tungnaá,
Skaftá, Hverfisfljót, Brunná and Djúpá. Sylgja is not
included since only a part of the Sylgja river basin
has been surveyed with radio echo sounding. The ac-
tual position of the main runoff outlets for Skaftá and
Hverfisfljót seen in polarimetric SAR (PolSAR) im-
ages, acquired by EMI at the same time as the InSAR
data, is shown as well. The river outlets for Djúpá and
Tungnaá could not be identified from the PolSAR im-
ages partly because these rivers don’t appear to have
any large outlets but also because they flow substan-
JÖKULL No. 54 23