Jökull - 01.01.2009, Side 12

Guðmundsson et al.
Figure 7. Observed and predicted (with DDM1 and DDM2) summer balance (b S) at locations of stakes at
Hagafellsjökull, from 2001–2005. The predicted b S is not corrected for snow that falls and melts within the
summer. – Reynslubundin líkön borin saman við sumarleysingu í mælistikum.
(errors in Table 3) are identical for ddf1 and ddf2 at
G1100 but substantially higher for ddf1 than ddf2 at
the lower G500. One explanation for the increased
ddf2snow downglacier could be the earlier timing of
the snow/ice transition at lower elevations, resulting
in higher incident solar radiation falling on a surface
with already reduced albedo, and hence higher ab-
sorbed solar radiation, during the period of parameter
optimisation. SPOT5 images show that the low albedo
for ice at G500 (" = 0.07), reflected as high ddf2ice
(Table 3), reaches only to elevation !100 m above
the station. The ice at 700–1000 m a. s. l. is much
cleaner (" $ 0.35), resulting in stable ddf2ice (Table
3). The ddf2 values in Table 3 are close to being the
same as previously found for northeastern Vatnajökull
and northern Hofsjökull ice caps, Iceland (Figure 1),
by using energy balance observations (Guðmundsson
et al., 2003) and mass balance observations on stakes
(Jóhannesson et al., 1995), respectively. The degree-
day parameter was found to be slightly higher for ice
on northern Hofsjökull than on southern Langjökull.
To investigate the seasonal sensitivity of the ddf -
parameters during unchanged surface conditions, we
used the observedweather parameters along with con-
stant albedo values of 0.1 to 0.9 for the period of
May through September at G1100 and G500, and op-
timised the ddf -parameters separately for each of the
months and each albedo value (black lines in Figure
8). The lower solar radiation generally resulted in
reduced degree-day factors during unchanged surface
conditions, indicating that the degree-day factors are
sensitive to seasonal changes in solar radiation; this is
particularly evident at the higher station G1100. The
high ddf -parameters optimised for May (black lines
in Figure 8) can be explained by the relatively strong
contribution of net radiation to melting during that
month. The impact of the abrupt transition from snow
to ice/firn depends on its timing during the summer.
For example, a drop in the albedo at G1100 from 0.5
to 0.3 in June-July would increase ddf1 from!17 mm
to 21 mm per !C, but the same decline in albedo dur-
ing July-August would result in unchanged ddf1 (Fig-
ure 8a). The parameter ddf2 has a lower value, varies
more slightly and is less sensitive to changes in the
weather parameters and to the timing of the snow-ice
transition than ddf1 (black lines in Figure 8); hence,
ddf2 comes nearer to depending solely on conditions
at the glacier surface. The typically strong contribu-
tion of net radiation in May affects ddf2 more than
ddf1, but occurrence of the strong winds and relatively
high temperatures in September affected ddf1, but not
ddf2 (thick gray lines in Figure 8a,c).
A justification for assuming time-independent
degree-day factors, varying only with surface condi-
tions (snow or ice), is that the reduced solar radiation
and increased heat fluxes as summer proceeds jointly
counteract the lowering of albedo, which explains the
stability with time in the monthly values for ddf1 and
ddf2 at both stations obtained by using the observed
albedo values (thick grey lines in Figure 8).
Our results indicate that the accuracy of degree-
day models can be improved by accounting for the
12 JÖKULL No. 59