Jökull - 01.01.2005, Qupperneq 136

Sverrir Guðmundsson et al.
Table 2: Coefficients of Eqs. (4-6) (DDM , TIM andEEB, respectively) optimized for the whole Brúarjökull
outlet (constant in time and space) by using the AWS data from 1996–2001. i and ii: Before and after exposure
of the previous year’s summer surface, respectively. – Stuðlar í reynslubundnum líkönum af leysingu jökuls.
ddfi ddfii MF ai aii TF bi bii
mm ◦C−1 mm ◦C−1 mm ◦C−1 mmW−1 m2 ◦C−1 mmW−1 m2 ◦C−1 Wm2 ◦C−1
DDM 5.7 6.6 - - - - - -
TIM - - 2 0.01 0.015 - - -
EEB - - - - - 13.9 0.1 0.24
when the runoff was mainly produced by the net radia-
tion alone. Better prediction is achieved with the TIM
model in Eq. (5) than the DDM in Eq. (4); unlike the
degree-day factor (ddf ) of the DDM model, the scal-
ing factor of the TIM model accounts for the reduced
solar zenith angle, and hence gives a better prediction
during the latter parts of the ablation season (Figure
13a,b). None of the three empirical models accounts
for wind-speed changes, but all predict runoff reason-
ably from July 28 to August 2, when intensive wind-
speeds, and hence high turbulent heat fluxes, were ob-
served on the glacier (Figures 3b,c,d,f and 12b). This
can be explained by low net radiation compensating
for the strong turbulent heat fluxes (Figures 12 and
3e).
CONCLUSION
Meteorological observations within the boundary
layer of Brúarjökull, along with mass balance mea-
surements at stakes, were successfully used to create
energy budget maps for the ablation season during
2004. The energy budget maps effectively estimate
glacial runoff, and are in good agreement with the ob-
served river discharge of Jökla, which drains the out-
let. The flood that occurred during the period August
3-6 was related to exceptionally intensive rain, but
glacial melting caused the flood that occurred from
August 9-14. The circumstances leading to the sec-
ond flood were i) five days with high turbulent heat
fluxes driven by strong southerly winds (July 28 to
August 1) and heat supplied by rain (August 1–6), all
speeding up the removal of snow in the ablation area
and lowering the albedo, and ii) exceptionally high air
temperatures and solar radiation along with abruptly
reduced albedo (August 9-14).
Seasonal variations in glacial runoff during 2004
can be calculated by three empirical ablation models,
which use air temperature observed away from the
glacier front as the only input. The best result was ob-
tained using an empirical energy balance model, espe-
cially during periods when melting was maintained by
the net radiation alone. The glacial peak-runoff dur-
ing the second flood was predicted acceptably by both
an empirical energy balance- and a temperature index
model that accounts for changes in the solar zenith
angle by incorporating theoretically calculated clear-
sky irradiance. A simple degree-day model, not ac-
counting for changes in the solar radiation, predicted
the second flood but overestimated the glacial peak-
runoff. The simplicity of the empirical models is a
great advantage. Air temperature away from glacier is
easier to assess than weather parameters on the glacier
needed for the energy balance calculations. How-
ever, the empirical models do not provide the same
detailed insight into the physical processes generating
the glacial runoff.
Acknowledgements
The work was supported by the National Power Com-
pany of Iceland, the University of Iceland Research
Fund and the Nordic project Climate, Water and
Energy. We are indebted to Einar Sveinbjörnsson
at the Icelandic Meteorological Office for supplying
data and assisting with interpretation of meteorolog-
ical data and optical satellite images, and Amy E.
136 JÖKULL No. 55