T. Klein et al., Simulation of the katabatic flow near the Greenland ice margin using a high-resolution nonhydrostatic model, METEOROL Z, 10(4), 2001, pp. 331-339
Simulations of the katabatic wind system in the stable boundary layer near
the margin of the Greenland ice sheet are presented using the high-resoluti
on (2.8 km) non-hydrostatic Lokal-Modell (LM) of the Deutscher Wetterdienst
(DWD). The LM is nested into numerical forecasts provided by the hydrostat
ic Norwegian Limited Area Model (NORLAM). The area of Kangerlussuaq in West
Greenland was selected as the LM domain, since comprehensive measurements
of the katabatic wind structure are available for that area (aircraft and A
WS). One focus of the LM simulations is to study processes in the transitio
n region between the ice sheet and the tundra, which are not captured by pr
evious model studies because of the micro-scale complexity of the topograph
y.
The first simulation is performed for an idealized case using a wintertime
atmosphere at rest as initial conditions. For this situation lacking synopt
ic forcing, a well-developed katabatic wind system is simulated over the in
land ice, while a system of fjord winds is simulated in the tundra area. In
the second simulation study a realistic case of strong katabatic winds (ex
ceeding 20 m s(-1)) is investigated, and AWS data and vertical profiles fro
m aircraft measurements are used for the model validation. For this real ca
se, the simulation results are in good agreement with instrumented aircraft
and automatic weather station observations. While NORLAM fails to reproduc
e the wind field in the tundra region, the LM simulations show again a patt
ern of fjord winds and a complex wind field near the ice margin, which agre
es with the observed atmospheric state. Because the intensity and the three
-dimensional structure of the katabatic winds depend on the specific synopt
ic environment, the success of the LM simulations is also dependent on the
quality of the initial and boundary conditions, i.e. the forecasts of synop
tic and mesoscale processes by NORLAM.