Urban Soil-Canopy-Atmosphere Exchanges at Submesoscales: Learning from Model Development, Evaluation, and Coupling with LES - Meteorological and Air Quality Models for Urban Areas - page 43

a) b)

800

800

600 600

400 400

200

200

0 0

-200 -200

000 0

0 600 1200 1800 2400

0000 0600 1200 1800 2400

Time (UTC)

Time (UTC)

c) d)

800

800

600 600

400

400

200 200

0 0

-200

-200

0000 0600 1200 1800 2400 0000 0600 1200 1800 2400

Time (UTC) Time (UTC)

Fig. 6.3 Evaluation of SM2U with Marseille city centre measurements. Mean diurnal cycle of
observed ( dots ) and simulated ( solid lines ) net radiation ( a ), turbulent latent heat flux ( b ), turbulent
sensible heat flux ( c ), and (storage) heat flux by conduction into wall/roof/soil materials, estimated
as the imbalance of the measured/simulated energy budgets ( d ). Net radiation and conduction
fluxes are positive downwards, while turbulent heat fluxes are positive upwards

scales (Berthier et al., 2006). For long time scales water removal from the deep soil
layer reservoir by infiltration in the draining network is important (Dupont et al.,
2006). Alternate simulations demonstrated the potentially large influence of surface
layout or management on the canopy microclimatology, especially in dry summer
conditions (Fig. 6.7): a small watering of a day ground may reduce the daytime
peak by several degrees, as well as the reduction of paved fraction. The difference
between the top and bottom curves in Fig. 6.7 is what may be observed in two simi-
lar housing estates when the common surfaces are either kept as natural as possible
or all paved and connected to the draining network.
Coupling the SEB model with an ABL model requires SM2U input parameters,
averaged over each computational grid mesh. These parameters include canopy mor-
phology descriptors and surface/material radiative and thermodynamic properties.
In a first step the urban land use and morphological parameters are extracted from
statistical analyses of urban databases by the DFmap pre-processor; in a second
step a semi-automatic classification is operated with the k-means method to identify
the urban districts, allowing to further identify the surface materials, hence their