Jökull - 01.01.2009, Page 13

Langjökull, energy balance and degree-day models
Figure 8. Changes in ddf1 of DDM1 (a,b) and ddf2 of DDM2 (c,d), presented in relation to albedo. Black lines:
experimental monthly values of ddf -parameters obtained from the EBM and by assuming constant albedo val-
ues (shown at the ends of the lines). Note: time independency would yield horizontal straight lines. Thick grey
line: observed monthly values of ddf1 and ddf2, accompanied by the corresponding observed monthly albedo
values. – Næmni stuðla í reynslubundnum líkönum fyrir breytingum í endurkastsstuðli yfirborðs.
solar zenith angle, which agrees with Guðmundsson
et al. (2005) who gained better degree-day prediction
of seasonal change in runoff at the northeastern part
of Vatnajökull (Figure 1), by incorporating modelled
clear sky irradiance into the degree-day scaling (as in-
troduced by Hock, 1999).
The performance of degree-daymodels in response
to regional temperature changes
Both the energy balance and the degree-day models
(parameters from Table 3) were used to estimate melt
rate changes in response to climate change. As a
case study we considered a seasonally constant tem-
perature change !TS outside the ice cap (at S475).
The calculation was restricted to a fixed period cor-
responding to the ablation season 2001 at the two
AWSs sites. The eddy fluxes were calculated step-
wise ( !TS = 1!C ) for temperatures deviating -5 to
5 !C from the observed values of TS in 2001. The
corresponding variations of TG (in !C) were derived
by using a piecewise-linear regression of the data in
Figure 9.
The observed values ofQi, Io (from amelting sur-
face), r (close to that of saturation), and the winter
balance (bW ) for the year 2001 were selected as a ref-
erence. Changes in air temperature outside the glacier
marginwere expected to have an effect on the strength
of the down slope glacier winds (e. g. Björnsson et
al., 2005). For estimating this effect (!u in m s"1),
we approximate the relation between temperature and
wind speed observations at G1100 and G500, and
temperature at S475 during May-September 2001–
2005 by piecewise linear regression (Björnsson et al.,
2005) based on our glacio-meteorological data from 5
ablation seasons:
!u =
#
$
%
0.25!TS TG # 0
0 otherwise
(11)
JÖKULL No. 59 13