Jökull - 01.07.2003, Síða 6

de Ruyter de Wildt et al.
Figure 2. Scatterplots of calculated against observed hourly means of global radiation (a) and of the incoming
longwave radiation (b). Only measurements from stations for which cloud observations are available are shown
(A1, U2, A4, A5, I6, R4 and R5). For calculation of the incoming longwave radiation data from the radio
soundings are used, which are available only twice a day. – Samband mældrar og reiknaðrar sólgeislunar (a)
og jarðgeislunar (b).
of the images. It displays few clouds and much bare
ice and shows that Breiðamerkurjökull and Skeiðarár-
jökull in the south have broad bands of low albedo and
are quite inhomogeneous. Tungnaárjökull and Köldu-
kvíslarjökull in the west are more homogeneous. The
same is true for Dyngjujökull in the north but this out-
let displays lower albedo values than the other ones
(below 0.05 in places). For each model evaluation site
we use the mean ice albedo that was determined for
that site from the satellite images.
In November 1996 a volcanic eruption took place
underneath Vatnajökull, which had a significant im-
pact upon the ice cap (Gudmundsson et al., 1997).
The eruption melted the ice, causing a large flood
or jökulhlaup south of Skeiðarárjökull. The eruption
also covered large parts of Vatnajökull with tephra.
From the AVHRR images we find that during the sum-
mer of 1997, the average albedo in the accumulation
area was about 0.08 lower than in the years prior to
and after 1997. In the model we therefore lower the
albedo of snow and firn by 0.08 during the melting
season of 1997. The eruption had no measurable ef-
fect upon the albedo of the ablation areas presumably
because there the albedo already was relatively low.
Longwave radiation
 stems from the lowest part of the atmospheric
boundary layer and from the upper-hemisphere slopes
that surround a measurement site. For an exact cal-
culation one needs to know the atmospheric profiles
of temperature and humidity, but generally the influ-
ence of the vertical profiles can be well described with
the temperature and water vapor pressure at screen-
height,  and e. We use a slightly altered ver-
sion of the parameterization given by Greuell et al.
(1997) which is based on Konzelmann et al. (1994)
and which calculates  as a function of , e and
cloudiness. However, the katabatic layer over Vatna-
jökull is generally not very thick (Oerlemans et al.,
1999) and  and e are not representative of that part
of the atmosphere that generates  . Near A5, for ex-
4 JÖKULL No. 52, 2003