Jökull - 01.01.2004, Side 21

Glaciological application of InSAR topography data of W-Vatnajökull
Some critical points exist, in the static potential
field, where we obtained ambiguous results for the
water divides. For example, the model can not pre-
dict with any certainty whether water from the area
north of the star in Figure 2 should drain to the rivers
Hverfisfljót or Djúpá. It is also uncertain whether
water from the areas north of regions marked by di-
monds drains into Skaftá or elsewhere. The area
north of the southernmost diamond could belong to
the Hverfisfljót river basin and the northern area the
Sylgja/Bryðja river basin. The two areas north of the
diamonds in Figure 2 will be referred to later as the
uncertainty areas for the Skaftá and Hverfisfljót river
basins. The water divides around Sylgjujökull are
also uncertain since the outermost part of this outlet
glacier has not been mapped in detail by radio echo
sounding. For example, a ridge clearly observed in
the surface DEM of Sylgjujökull is missing in the
basement DEM. Due to this ridge, the river basin of
Skaftá might be substantially larger at the expense of
Sylgja/Bryðja and even Kaldakvísl river basins (Fig-
ure 2).
Estimates of summer balance within the water
divides of Djúpá and Hverfisfljót (Björnsson et al.,
1998, 2002) were compared with river discharge mea-
surements during June to September (Figure 3), pro-
vided by the National Energy Authority. The glacier
meltwater contribution during the summer was cal-
culated by subtracting the direct runoff, attributed to
precipitation, from the measured river discharge, as-
suming that the volume of ground water draining each
river basin equals the measured runoff from that basin.
The direct runoff for each river was crudely estimated
by scaling the precipitation measured at Kirkjubæj-
arklaustur (www.vedur.is/vedurfar/yfirlit/medaltalstoflur/
Stod
−
772
−
Kirkjubajarklaustur. ManMedal.txt, 2003) with
the area of the river basin below 1100 m. The pre-
cipitation above 1100 m is hard to estimate based on
the Kirkjubæjarklaustur data, as it lies significantly
further away from the station. It is also uncertain to
which river the area above 1100 m is connected (Fig-
ure 2). Excluding this region causes an underestimate
in the direct runoff for the river draining it, leading
to an overestimate in the glacier contribution for this
river.
Discharge values and summer balance measure-
ments agree fairly well, especially since 1999, assum-
ing that the area west of Grímsvötn is connected to
Hverfisfljót. The difference in 1996 to 1998 could be
due to higher negative balance for Síðujökull as a re-
sult of its surge in 1994 (Björnsson et al., 2003). The
summer balance is not measured specifically for Síðu-
jökull but estimated from direct measurements from
Tungnaárjökull and Breiðamerkurjökull and in some
years from Skeiðarárjökull as well.
The average ratio between the glacier runoff, dur-
ing the summer, in Hverfisfljót and in Djúpá and sum-
mer balance is around 3 assuming the area west of
Grímsvötn drains to Hverfisfljót and 0.9 if it drains
to Djúpá, while the corresponding average ratio esti-
mated from the measured river discharge and the pre-
cipitation is 2.2 ± 0.7 assuming an error of up to 50%
in the precipitation contribution. The results indicate
that most or all of the water from the area directly
west of Grímsvötn drains into Hverfisfljót. The possi-
bility of a minor amount draining into Djúpá should,
however, not be rejected.
The derived water divides for both branches of the
river Skaftá are also of interest. According to the static
potential model, the river basin of the eastern branch
extends into the Skaftá cauldrons (Figure 2). How-
ever, discharge from the Skaftá cauldrons during re-
cent jökulhlaups all drained into the western branch
(Bjarni Kristinsson, pers. comm., 2003). This dis-
crepancy could be caused by several effects; The wa-
ter pressure during jökulhlaup may be totally different
from normal conditions drasticly changing long-term
water divides and flow paths. One might argue that
water in jökulhlaups tends to drain in pre-existing tun-
nels. This might be true in many cases but the shape
of the hydrographs for jökulhlaups in Skaftá suggests
that their water does not pass through a single tunnel
but instead spreads out as distributed flow (Björnsson,
1992). A simple model may apply to the more regu-
lar jökulhlaups in Skeiðará only whereas distributed
flow is acquired to explain larger events such as the
>3 km3 jökulhlaup in 1996 (Björnsson, 2002) which
emerged into all rivers draining Skeiðarárjökull, al-
though initially bursting into Skeiðará. This model
may also oversimplify the water divides between in-
JÖKULL No. 54 21