Jökull - 01.07.2003, Side 10

de Ruyter de Wildt et al.
Table 1. Observed and simulated sensitivities of temperature and windspeed at screen-height. The sensitivities
were obtained by fitting a straight line to daily mean values. – Mæld (obs.) og reiknuð (cal.) aukning í lofthita
og vindhraða í veðurstöðvum við einnar gráðu hlýnun andrúmslofts.
Station Altitude (m)





(  Obs. Cal. Obs.
U3 165 0.32
0.05 0.34
0.00 0.30
0.08
A4 279 0.38
0.05 0.38
0.01 0.18
0.10
A5 381 0.41
0.05 0.43
0.01 0.12
0.08
I6 715 0.62
0.05 0.57
0.01 -0.04
0.09
R6 820 0.94
0.06 0.62
0.01 0.17
0.14
R4 830 0.70
0.07 0.62
0.01 0.21
0.11
U9 870 0.74
0.08 0.63
0.01 0.32
0.09
R2 1100 0.62
0.07 0.73
0.02 -0.20
0.13
R5 1140 0.75
0.06 0.75
0.02 0.05
0.12
U8 1210 0.78
0.08 0.78
0.02 0.19
0.12
U7 1530 0.99
0.08 0.91
0.02 -0.12
0.11
R3 1712 0.95
0.09 0.99
0.03 -0.02
0.15
is the most suitable. From this station we use daily
means of temperature, vapor pressure, pressure and
cloudiness from 1965 to 1999. Upon the temperature
we impose a daily cycle with an amplitude of 2.2 K,
which is the observed mean daily amplitude at the sta-
tions A1, U10 and R1. Precipitation in KIR was found
to be not particularly well correlated to the observed
winter mass balance of northern and western Vatna-
jökull. The precipitation measured in Fagurhólsmýri
proved to be more useful, which is why we force the
model with monthly values from this station.
Variables in the free atmosphere
The temperature in the free atmosphere influences  (see the Appendix) and the occurrence of solid precip-
itation. It is calculated from the temperature at KIR
(
) and the atmospheric lapse rate (), which
is obtained from the radiosonde data (-5.8 K/km).
The atmospheric pressure () is extrapolated from the
pressure measured at KIR (
) with the same expo-
nential decrease with altitude as was measured by the
radiosondes. Vapor pressure in the free atmosphere
(e) is calculated by assuming a constant relative
humidity with altitude. The radiosonde data showed
that this was, on average, the case.
Variables in the katabatic layer
When melting conditions prevail a cool katabatic sur-
face layer develops in which temperature and wind
speed deviate from their values in the free atmosphere.
All weather stations on the ice cap, except U7, R3 and
R6, display a preferred wind direction (Oerlemans et
al., 1999). This means that a large part of the ice
cap is subject to a more or less persistent katabatic
wind regime which shields the surface from fluctua-
tions in the atmosphere. This is clearly reflected in the
sensitivities of 2 m variables to external changes (Ta-
ble 1). The sensitivity of  to external temperature
changes,


, is smallest for the lowest sta-
tions and increases with altitude. It approaches 1 for
the stations where a katabatic layer was weak or ab-
sent (U7 and R3 high in the accumulation area and R6
on an open spot at the very glacier margin). Neglect-
ing the small sensitivities of the 2 m variables leads to
an over-estimation of the sensitivity of the turbulent
fluxes to external temperature changes (Greuell and
Böhm, 1998).
From a physical point of view, parameterizations
of the 2 m variables should be based upon the dynamic
equations that describe the turbulent exchange of mo-
8 JÖKULL No. 52, 2003