Environmental Engineering Reference
In-Depth Information
developments and challenges of urban regional modeling is
discussed in Masson (2006) and Martilli (2007).
In order to resolve the air flow around individual buildings
computational fluid dynamics (CFD) models such as Reynolds-
averagedNavier-Stokes and Large-Eddy Simulationmodels with
spatial resolutions of
∼
1 m are applied to areas of interest within
cities. Thosemodels can resolve urban aerodynamic features such
as street level flow. Increasingly output fromregional atmospheric
models is used to provide initial and boundary conditions to the
CFD models while on the other hand CFD model output on can
be aggregated and transferred back to the regional model (Chen
et al
., 2010).
The main physical approaches to represent urban areas
in regional atmospheric models are: roughness approach,
single-layer urban canopy approaches, and multi-layer urban
canopy approaches.
are the Town Energy Balance Scheme (TEB; Masson, 2000), the
Noah Urban Canopy Model (Noah UCM; Kusaka and Kimura,
2004) and models by Mills (1997) and Oleson
et al
. (2008).
21.2.3
Multilayer urban canopy
approaches
The multilayer UCM is currently the most complex among
the urban approaches used in regional atmospheric modeling.
Examples of multilayer UCMs include the Building Parameteri-
zation Scheme (BEP) by Martilli
et al
. (2002) and the scheme by
Dupont, Otte and Ching (2004). In a multilayer UCM, exchanges
with the atmosphere occur at multiple vertical levels within the
urban canopy by directly modifying the prognostic differential
equations of the regional atmospheric model to include addi-
tional terms such as urban drag force, heating, turbulent kinetic
energy production and dissipation terms. At each level within
the urban canopy the urban surface energy balance is solved for
roofs and walls and, at the bottom of the model, for roads. Hence,
within urban canopy profiles of air temperature, humidity and
wind speed can be predicted by the regional model and therefore
the environmental conditions were humans live. However, the
models are difficult to couple with regional atmospheric models
and significantly increase their computational time. The coupling
of UCMs with regional atmospheric models requires the prepro-
cessing of LULC and urban parameters; to enable land surface
models to average calculated energy and momentum fluxes and
surface temperatures fromnatural and urban fractions of cover of
amodel grid cell; and in case ofmultilayerUCMs themodification
of the prognostic equations of the atmospheric model.
Single-layer UCMs make use of the geometry of a generic
two-dimensional street canyon as an abstraction of the true
heterogeneous urban geometry. The geometry is enhanced in the
multilayer UCMs by considering subgrid scale features such as
street canyons of several street directions and vertical changes
of building density. However, for both approaches there are
limitations particularly when investigating micrometeorological
or neighborhood-scale characteristics of meteorological variables
where site-specific details may become important. This applies
for example when studying mitigation strategies for the urban
heat island.
The demand on model input parameters in terms of urban
geometry is also increased. Table 21.1 gives an overview of input
parameters for the different urbanmodel types (fromChing
et al
.,
2009). Burian and Ching (2009) present a detailed derivation of
UCPs for the example of Houston. In the next section we will
discuss how those parameters can be derived from remotely
sensed data.
21.2.1
Roughness approach
A relatively simple urban parameterization to account for the
influence of urban areas on the atmosphere is the ''roughness
approach'', in which the urban surface is treated physically like
a soil surface, but with an adjusted increased roughness, heat
capacity and conductivity and with modified albedo, emissiv-
ity and water availability for evapotranspiration. The roughness
approach was successfully enhanced by Taha (1999) by means of
including an empirical Objective Hysteresis Model by Grim-
mond, Cleugh and Oke (1991) thereby accounting for the
increased heat storage in cities.
The advantage of the roughness approach is the relatively
low demand on input parameters and simplicity of coupling
the approach with regional atmospheric models. Usually, remote
sensing-derived LULC classes with certain associated physical
characteristics are used as model input. The roughness approach,
however, does not resolve vertically the effects of buildings
on the urban canopy air, which is important for some model
applications, usually related to air quality and human comfort
(Masson 2006).
21.2.2
Single-layer urban canopy
approaches
The single-layer UCMs incorporate urban geometry into the
surface-energy balance and wind-shear calculations by using the
geometry of a generic street canyon that is characterized by road
width, building width and building heights. The surface energy
balance and amultilayer heat conductivity equation are solved for
roof, wall, and road surfaces. Within-urban canopy profiles of air
temperature, humidity andwind speed can be diagnosed fromthe
predicted variables above roof level and empirical profile func-
tions. The demand on input parameters is significantly larger than
for the roughness approach andmust include parameters defining
the average urban geometry, together with physical characteristics
of roof, road, and wall materials such as heat conductivity, heat
capacity, albedo and emissivity. The application of single-layer
UCMs with regional atmospheric models increases their compu-
tational needs anddemands by coupling the urban surface param-
eters to the atmospheric models. Examples of single layer UCMs
21.3
Remotely sensed data
as input for regional
atmospheric models
Remotely sensed data is an important tool for obtaining spatially
extensive, detailed and temporally repeatable data that records
the physical character of the urban land surface that is a result
Search WWH ::
Custom Search