Jökull - 01.12.1999, Síða 57

Figure 7. Deposition in the lee of a cluster of ice bergs. Note compass-clinometer for scale (centre).
— Setmyndun í hléi við þyrpingu ísjaka. Attaviti (í miðri mynd) er mælikvarði.
2 and 6b), declining downstream to 400-900 Wm“2
within 2 km (Figure 2). These values are similar to
those published by Baker and Costa (1987) for floods
in bedrock channels. Lower flood powers calculated
for cross-sections 2 and 3 fall within upper the range
of floods in alluvial channels (Baker and Costa, 1987).
DISCUSSION AND CONCLUSIONS
The primary geomorphological impact of the Novem-
ber 1996 jökulhlaup on the Gígjukvísl river chan-
nel was erosion of the narrow channel through the
moraines to create a much wider channel with a less
sinuous course (Figures 2-4 and 6). Downstream of
the moraines channel enlargement was accompanied
by net aggradation in the bed of the former channel
thalweg of up to 3 m (Figure 4c). Waning stage flow
and hence erosion was concentrated on the eastern
bank of the Gígjukvísl (Figure 2). The large bar on the
western margin of the channel is largely an erosional
feature on top of which large numbers of ice bergs are
grounded, often creating classic obstacle marks char-
acterised by proximal substrate erosion (Figures 2 and
3b) (Russell, 1993).
Peak jökulhlaup discharge reconstructed for the
Gígjukvísl channel using the slope-area technique
suggests a peak discharge of 19,500 m3s_1 ±5,000
m3s“ J . The errors on this value are high, due to un-
certainties regarding the amount and timing of chan-
nel scour and fill, thereby reducing the accuracy of
cross-sectional area estimates. Comparison of our
discharge figures for the Gígjukvísl river with peak
discharge values provided by Björnsson (1997) and
Snorrason et al. (1997), suggests that the Gígju-
JÖKULL, No. 47 55