Jökull - 01.07.2003, Side 7

A calibrated mass balance model for Vatnajökull
Figure 3. NOAA AVHRR image of Vatnajökull, taken on August 23rd, 1998. It has been processed so that it
represents the surface albedo. The black pixels represent clouds. – Gervitunglamynd sem sýnir endurkaststuðul
sólgeislunar á Vatnajökli 23. ágúst 1998. Svörtu reitirnir stafa af skýjum.
ample, the temperature inversion typically occurs in
the lowest 20 to 30 m above the surface. This means
that  , even for this station which is situated in the
lower part of the ablation area, is only partly influ-
enced by the katabatic layer and is best described as a
function of meteorological variables in the free atmo-
sphere (Meesters and van den Broeke, 1996). When
we tune the parameterization to the measurements of
 (Appendix), we indeed find that  is best de-
scribed as a function of the temperature and the water
vapor pressure in the free atmosphere at the same alti-
tude as the measurement site ( and e). Apart
from this, when the mass balance model is forced
with external data,  and e need to be computed
from the external data, which introduces extra uncer-
tainty. When we compute  from the cloud obser-
vations and from  and e (measured by the ra-
dio soundings), we obtain a correlation coefficient be-
tween computed and observed  of 0.86. The resid-
ual standard deviation is 17.6 W/m (Figure 2b).
 depends on the surface temperature and on
the surface emissivity, which were not measured.
However, the emissivity of snow and glacier ice is
very close to 1 (Warren, 1982), while we assume the
surface temperature to be equal to  when  is be-
low the freezing point. Otherwise, the surface temper-
ature is set to 0ÆC.
JÖKULL No. 52, 2003 5