Performance of Different Sub-Grid-Scale Surface Flux Parameterizations for Urban and Rural Areas - Meteorological and Air Quality Models for Urban Areas - page 20

is density of air, q is specific humidity, U(z 1 ) is the main flow in
x-direction at height z 1 ,

Where ρ

ψ q are the stability functions for momentum and
humidity (Dyer, 1974), L i is the Obukhov stability length for land use class i ,

ψ m and

κ
is the von Kárman constant (0.4) and l 21 the latent heat of vaporisation of water.
The sub-grid-scale surface fluxes are then averaged to a box averaged flux at the
blending height, which is the height, where the influence of the individual surface
characteristics is blended out.

3.2.2 Simulated Situations

The 3D version of the atmospheric MEsoscale TRAnsport and Stream (METRAS)
model (Schlünzen, 1990) is initialized with 1D background profiles assuming hori-
zontal homogeneity. The simulations are nested in European wide observational data
based on soundings and surface observations using a nudging approach (e.g. Dierer
et al., 2005). Table 3.1 summarizes the situations simulated. The selected situations
differ by the influence of surface fluxes. This is characterized by the locality index
( I lt ), a measure for the locality of the meteorological situation, which describes how
relevant local influences might be for the selected day. It is a measure for the strength
of the turbulent transport from the surface and is mainly a function of the friction
velocity and the free convection velocity.

Table 3.1 Situations simulated

Comment on
situation Date Grid sizes [km] I lt

21.07.1974 18, 4 0 very dry

04.03.2003 16, 8, 4 10 humid

04.01.2003 16, 8, 4 20 humid

10.03.2003 16, 8, 4 30 very dry

03.06.2003 16, 8, 4 40 very dry

3.2.3 Result Evaluation

The model performance is evaluated by calculating hit rates from routine data of
four DWD stations in the Berlin area (10381, 10382, 10384, and 10389).
The model results of air temperature, dew point, wind speed and direction are
interpolated into the locations of the DWD stations at 2 m height above the surface.
The hit rates are calculated following Eq. (3.3):

n i , with 1for

m

100
m

| difference ( measurement, model ) | < A

=

H (3.3)

0for | difference ( measurement, model ) |≥ A

i = 1