Jökull - 01.07.2003, Qupperneq 18

de Ruyter de Wildt et al.
Table 5. Static sensitivities of the mean specific mass balance of various parts of Vatnajökull to changes in
free atmospheric temperature and precipitation. All quantities are given as the mean during the period 1965 to
1999. – Metin breyting í afkomu á ýmsum hlutum Vatnajökuls við einnar gráðu hlýnun ( ) og 10% aukningu
í úrkomu ( ).
Region Area   m  m
(km ) (m) (
) (w.e./10%)
Síðujökull 514 2.7 -0.76 0.31
Tungnaárjökull 491 2.0 -0.77 0.30
Köldukvíslarjökull 330 1.7 -0.50 0.28
Dyngjujökull 1239 2.0 -0.49 0.28
W-Brúarjökull 956 2.3 -0.62 0.29
E-Brúarjökull 780 2.5 -0.73 0.30
Eyjabakkajökull 121 2.4 -0.74 0.27
Fláajökull 173 3.3 -0.73 0.32
E-Breiðamerkurjökull 631 3.0 -0.69 0.32
W-Breiðamerkurjökull 384 3.1 -0.80 0.35
E-Skeiðarárjökull 1002 3.1 -0.61 0.34
W-Skeiðarárjökull 434 2.8 -0.75 0.31
Vatnajökull 8198 2.6 -0.65 0.31
 , which is somewhat lower than the val-
ues used by Oerlemans (1992) (7  ). Other
authors (e.g., Kuhn, 1989; Braithwaite, 1995) calcu-
late C from 2 m variables and obtain values between 6
and 31  . These values compare well with
the values that we find for  

  and
that are two to three times as high as

(i.e. 9–13  ). So when  is explicitly cal-
culated in the model, the bulk method can be used, but
when only  is available and if equation 9 is used,
a correct transfer coefficient equal to

should be applied.
Another difference with some other models (e.g.
Oerlemans, 1992) is the dependence of snow albedo
upon the age of surface snow. For Vatnajökull the
occurrence of summer snow fall in the accumulation
area is strongly related to temperature. Modeling the
albedo of snow as a function of snow age will in-
crease  with respect to a model with constant snow
albedo. When we use a constant mean value of 0.72
for the albedo of snow, we find that  is roughly
25% lower.
We studied the sensitivity of  of Vatnajökull
to external climatic changes with the model. Meteo-
rological stations in the vicinity of Vatnajökull are all
located south of the ice cap, so we could only force the
model with data from these weather stations. This in-
troduces errors in cloudiness and precipitation, which
are not well known over Vatnajökull. Especially the
distribution of precipitation may vary from year to
year, dependent on the predominant wind direction.
In spite of this, the mean specific mass balance over
seven years could be simulated reasonably well. The
fact that Vatnajökull shapes its own precipitation gra-
dients results in strongly varying sensitivities to tem-
perature changes over the ice cap. In the south and
southeast, where the climate is more maritime,  is
higher than in the drier northwest. The variation of
 over Vatnajökull stresses the importance of an op-
timization procedure as that described in this work.
Even for much smaller ice caps or glaciers, the pre-
cipitation can vary considerably irrespective of alti-
tude, and this strongly affects their sensitivity to cli-
mate change.
16 JÖKULL No. 52, 2003