Jökull - 01.07.2003, Blaðsíða 16

de Ruyter de Wildt et al.
The outlet glaciers in the south and southeast also re-
ceive more precipitation than the ablation areas in the
north and northwest, even though the latter are situ-
ated on higher grounds. Consequently, the Equilib-
rium Line Altitude (ELA) in the south and southeast
is lower than in the northwest (Figure 8). The ELA
ranges from about 1000 m at some locations in the
south and southeast to over 1400 m on Dyngjujökull
and Köldukvíslarjökull in the northwest. The accu-
mulation area ratio of Vatnajökull is found to be 64%.
Figure 9 shows the simulated mean specific mass
balance of the northwestern drainage basins of Vatna-
jökull as a function of time. The winter mass bal-
ance has slightly increased between 1965 and 2000,
whereas the summer mass balance has decreased. As
a result the annual  decreased from slightly posi-
tive in the 1970s to near-neutral in the 1990s (changes
in glacier surface not taken into account). In general,
fluctuations in winter and summer balance are more
or less equally large. Interestingly, the most extreme
positive and negative annual values both occur in the
early 1990s and are both due to summer conditions. In
1991 the summer was exceptionally sunny and warm,
whereas in 1992 the summer was cool with consider-
able snowfall in June and August (Björnsson et al.,
1998a). The model reproduces  over the years
1993 to 1999 fairly well (we estimate the uncertainty
in the measured  to be 0.25 m w.e.) The effect
of the albedo lowering due to the volcanic eruption
of November 1996 is adequately simulated: without
the additional lowering of the snow albedo during the
summer of 1997, the simulated mass balance would
have been considerably too high. Table 4 displays
correlation coefficients between observed and mod-
eled  for all drainage basins where has been
measured. In general the correlations are good, al-
though for the winter mass balance the correlations
are low for some drainage basins. This is obviously
caused by the distribution of precipitation over Vatna-
jökull, which varies from year to year, whereas we
use a fixed distribution of precipitation. Vatnajökull is
sufficiently high and large to act as a topographic bar-
rier, so the distribution of precipitation depends upon
the large-scale atmospheric circulation.
Figure 9. Observed (thick solid lines) and recon-
structed (dashed lines) mean specific mass balance
for the northwestern part of Vatnajökull (formed by
the drainage basins Tungnaárjökull, Köldukvíslarjök-
ull, Dyngjujökull and Brúarjökull). Shown are the
winter balance (crosses), the summer balance (open
circles) and the annual balance (solid squares). The
thin solid lines are smoothed curves. – Mæld (þykk
lína) og reiknuð (brotin lína) afkoma á norðvestur-
hluta Vatnajökuls; vetrarafkoma (krossar), sumaraf-
koma (opnir hringir), ársafkoma (svartir ferningar);
þunnar heildregnar línur sýna meðaltöl.
THE SENSITIVITY OF VATNAJÖKULL
TO CLIMATIC CHANGE
We compute the sensitivity of  to a change in a
variable  as (Oerlemans, 1996):
 Æ 
Æ 

 



(8)
where
 is the change in variable . We compute the
sensitivity to changes in the two most important atmo-
spheric variables, temperature (
 = 1K) and precip-
itation (
 = 10%). We perturb these variables over
the period 1965–1999 and then compute the average
change in  over this period. Table 5 shows  and
 for different parts of Vatnajökull. Most obviously,
14 JÖKULL No. 52, 2003