Jökull - 01.01.2009, Page 14

Guðmundsson et al.
Figure 9. Relationship of air temperature TG at G1100
(a) and G500 (b) during the summers of 2001–2006, to
temperature TS at S475, outside the glacier, presented
in each instance as a one-hour mean. Piecewise linear
regression between temperatures on and off the glacier
are shown as imprinted grey lines. The scatter values in
(b) were separated into southern regional winds (light
grey) and northern down slope glacier winds (grey dark)
when optimising Eq. 12. – Samband mælds hita yfir
bráðnandi jökulyfirborði við hita mældan utan jökuls.
!u =
#
$
%
0.54!TS TG # 0
0 otherwise
(12)
at G1100 and G500, respectively. Equations 11and 12
apply for a melting glacier surface, approximatedwith
TG # 0. The observed incoming long wave radiation
(Ii) of the ablation seasons 2001–2005 varies between
%ST 4G and %ST 4S (%S being the Stefan-Boltzmann
constant), indicating a too shallow boundary layer to
eliminate effects from the warm air above it. The
boundary layer was expected to become thicker with
increased TS and the changes of Ii taken as mid-
values between %ST 4G and %ST 4S .
Calculations of the energy fluxes, suggest that
the melting rates would be accelerated both by eddy
fluxes and net radiation, however, more by eddy fluxes
at the lower station (Figures 10a-b and 11a-b). The net
radiation would be affected more by increased long
wave radiation than albedo despite earlier exposure of
the summer surface.
Reasonable agreement was obtained between the
complete model of physical energy balance (Eq. 5)
and the degree-daymodels (Eqs. 9-10) at G1100when
"5!C & !TS & 5!C (Figure 10c-d), and this also
applied to G500 for "5!C & !TS & 2!C; keeping
the wind speed unchanged from the 2001 reference
values as well as wind speed changes proportional to
temperature (Figure 11c-d). The empirical models di-
verged increasingly from the physical model when as-
suming glacier winds to change proportionally to TS
(Figures 10d and 11d).
14 JÖKULL No. 59