Energy and Buildings 36 (2004) 403–413

Modelling microclimate in urban environments and assessing

its influence on the performance of surrounding buildings
Francisco Sánchez de la Flor∗ , Servando Alvarez Domı́nguez
E.S. Ingenieros, Grupo de Termotecnia, Avda. de los descubrimientos s/n, 41092 Seville, Spain

Abstract

The present paper, presents a computational model that assess the modification of the climatic variables in an urban context, and how
they affect to the thermal performance of urban buildings. There is a clear interaction between both systems, and they are able to modify
their mutual energy balances. Therefore, a coupling of the urban model and the building thermal performance, it is, absolutely, necessary if
we are interested in clarify the consequences of that interaction on heating and cooling requirements, and even on urban comfort. Finally,
and from the building point of view, it could be useful to develop a methodology to evaluate the impact of the microclimatic modifications
in the building energy demands. In fact, it treats of a double methodology, the first one characterizes the influence of the climatic variables
due to the urban context, upon the complete building, and the second upon the contribution to the demands by envelope components.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Heating and cooling requirements; Rural–urban layout; Urban heat island; Vegetation and mass of water

1. Introduction erful, disseminated and easy to use passive cooling in the

The interaction between building’s energy balance and
their surroundings is well-known. The envelope that sepa-
rates them is also the one that determines the magnitude of
that, upon permitting a greater or smaller heat transfer. This
does not depend only on the own building, but, on the con-
trary, is seen very influenced by the surroundings, and in
general, by the geometric and constructive characteristics of
the urban layout in which it is situated.
The situation, besides, is different in the rural environ-
ment and in the urban one, since all the energy exchanges
are modified by the so-called microclimatic changes due to
human activity. As a consequence, the needs of heating and
cooling of the same building, are not the same in both en-
vironments. On the other hand, the problem can be stud-
ied also from the point of view of the comfort in the open
spaces: streets, squares and courtyards of the cities, since
the presence of building modifies the outdoor conditions.
In urban areas, the impact of the urban heat island on en-
ergy requirements of buildings is tremendous especially on
cooling loads. During town planning, passive solutions and
integration of renewable energy sources have to be proposed
by urban designers. In this aspect, the more attractive, pow-
∗ Corresponding author. Tel.: +34-954-487-259; fax: +34-954-463-153.
E-mail address: fjs@tmt.us.es (F.S. de la Flor).
urban environment is vegetation. Vegetation for cooling pur-
poses can be considered as a renewable heat sink, in a sim-
ilar way that the sun is a renewable heat source for heating
purposes.
With the objective to analyse the effect of urban layout
on climate, and its modification, different models for each
mechanisms of heat transfer in the building envelope have
been developed. These models have been coupled among
themselves, and implemented in data processing tools, and
they have been validated through the use of experimental
results.
2. General aspects of the urban canyon model
In the framework of the GreenCode project, a program
focused on calculations of the urban context impact in the
building energy performance, was developed (GreenCanyon
[1]). The whole process is divided into several parts or ap-
proximation levels. The developed models for the first level
calculate the ambient conditions in the surrounding of the
building, like air temperature, air velocity, impinging so-
lar radiation, etc. These conditions are calculated assuming
an air temperature distribution at a rural environment, like
for instance an airport meteorological station, and then it is
modified by the influence of the intermediate area between
that rural site and the urban site under consideration.

0378-7788/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.enbuild.2004.01.050
404 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413
A further approximation level is integrated by a set of dif-
ferent sub-models that are coupled forming the so-called Ur-
ban Canyon Model, as it is explained below. Each sub-model
treats the calculation of a different heat flux in an open urban
environment, this is: wall and ground conduction, short-wave
and long-wave radiation, convection, and finally water ponds
and trees or vegetated surfaces in the streets.
2.1. Urban modelling approach
The computational model calculates the air temperature
and surfaces temperatures of a defined street canyon by solv-
ing the air and surface thermal balance, respectively, for each
cell of a grid defined by the program user. Thus, the user
has to introduce the canyon size and the number of cells in
which to divide the whole volume, and on the other hand has
to set the façade and ground composition and thermal prop-
erties. Finally, the program can be run with a design day,
previously defined by setting the geographical site, and the
maximum and minimum air temperatures, or with a modi-
fied day using the first level of the program (Fig. 1).
A brief description of the theoretical basis of each
sub-model is given below.
Conduction through walls is modelled through the use of
an equivalent wall composed by a layer of massive material,
concrete, that accounts for the inertia of the wall and a layer
of insulation. This equivalent wall has the same global heat
transfer coefficient of the actual one, and the same transient
thermal behaviour. The advantage of an equivalent wall in
comparison with the actual one is that the coefficients of
its representative transfer function can be obtained using
correlations based on a little number of parameters, like
for instance the insulation layer position. This sub-model
considers three different positions of the insulation layer:
inside, outside, or centred. Then, the transient heat flux by
conduction through walls can be obtained for the time step
t, as follows [2]
qsconduction (t) = a0 Ts (t) + a1 Ts (t − 1) + a2 Ts (t − 2) − C
− d1 qsconduction (t − 1)
Fig. 1. Description of heat and mass transfer between the buildings.
where a0 , a1 , a2 , C and d1 , are the coefficients of the trans-
fer function, and can be obtained through correlations, as it
has been explained. The whole process to obtain these co-
efficients is based in a finite difference method.
The ground conduction is also modelled using a transient
one-dimensional approach on the conductive heat transfer
phenomena of the ground. As in the case of wall conduction,
an assumption has been made in order to obtain correlations
for the calculation of the coefficients of a transfer function.
This assumption consists in the consideration that the ground
is composed by two layers of different materials: one is bare
ground, and the other one is selected by the user among a
data base with many materials. The objective of the model
is to represent the thermal behaviour of the ground, with an
expression of the heat flux useful to make the coupling with
the other models.
The ground temperature can be represented with a high
accuracy into the addition of two periodical waves, with
yearly and daily periods, respectively [3]. This allows solv-
ing the problem using superposition. As the calculations are
realised only for a design day, the heat flux corresponding
to the effect of the yearly fluctuation can be considered as a
constant that can be calculated using the mean yearly fluc-
tuation of the ground surface temperature (MYF).
The second term is due to the daily fluctuation, and pro-
vides a time dependence calculation. Then, the ground tem-
perature can be calculated as the solution of the superposi-
tion of a daily solution in transient state and the yearly one
in ‘steady state’, as it has been shown in Fig. 2.
The depths, at which daily and yearly fluctuations are sup-
posed to be negligible, are marked in the plot with dashed
lines. In both graphs typical curves representing the mini-
mum, medium and maximum values have been also drawn.
The heat flux by conduction through the ground can be
calculated then using the following equation:
qsiconduction (t) = a0 Tsi (t) + a1 Tsi (t − 1) + a2 Tsi (t − 2) − C
− d1 qsiconduction (t − 1) + qannual
where a0 , a1 , a2 , C and d1 , are the coefficients of the transfer
function, and can be obtained through correlations, as it has
been explained.
The short-wave solar irradiance absorbed by the surfaces
of the urban canyon can be calculated multiplying the total
solar radiation incident on those surfaces, by the absorption
coefficient (α). On the other hand, the total solar radiation
(direct, diffuse and reflected) incident on each surface is
calculated by the improved model of Elnahas and Williams
[4] which is given by the equation:
I(t) = Idir (t)(1 − PSA(t)) + Idif (t)SVF


+ Idir+dif (t)(1 − αi )FiA
i
where PSA(t) is the partial shaded area of the surface; I(t),
mean hourly total solar radiation incident on the surface
(W/m2 ); Idir (t), hourly mean unobstructed direct solar radi-
 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413 405

Daily Fluctuation Yearly Fluctuation

Ground Temperature Distribution Ground Temperature Distribution
0 0

0.1 1

2
0.2
3
0.3
Depth (m)

4
Depth (m)

0.4 5

0.5 6

7
0.6
8
0.7
9

0.8 10
15 20 25 30 35 -25 -15 -5 5 15 25
Temperature (ºC)
+
Temperature (ºC)

Fig. 2. Superposition of daily solution in transient state and the yearly one in ‘steady state’.

ation (W/m2 ); Idif (t), hourly mean unobstructed diffuse so-
lar radiation on a horizontal plane; Idif+dir (t), hourly mean
direct and diffuse solar radiation; F, view factor for two sur-
faces; α, solar absorptivity of the surface; SVF, sky view
factor for the surface.
In this equation, the incoming diffuse radiation is assumed
to be isotropic and the first reflection is only included as
it is very difficult to model multiple reflections. The direct
radiation has to be previously calculated for the specific solar
altitude.
The air movement model use two kinds of models, in the
low part of the street it is used a zonal model and in the high
part the wind effects are taking into account by a specific
airflow model. Applying the zonal model to a street between
two buildings involves partitioning the volume into “cells”.
Then, equations for mass balance and heat balance are solved
for those cells. Additional equations describe conduction
through the walls, convection between outside wall surfaces
and cells, and long-wave and short-wave radiant interchange
among the outside wall surfaces. Each type of cell has a
different set of equations for the internal behaviour of the
cell and for the transfer with other cells.
The specific airflow model used the Nicholson proposed
equation [5] to calculate airflow velocity in the high part of
the street:
 ∗  
u z + d + z0
ū(z) = ln
k z0
A dynamic complete model of the thermal performance
of a water pond that is subjected to various atmospheric
conditions is used for ponds, fountains and mass of water.
This model is based on the application of a global thermal
balance, with the assumption of the absence of vertical gra-
dients of temperatures inside water mass. In the case of the
water pond of little depth (ca. 1 m), it can consider that the
entire pond water mass is at average temperature, therefore,
the temperature field is only function of time. In these con-
ditions, the distribution of the temperature field is calculated
from a global balance in the pond water per unit of pond
surface.
ρcp V dT
U = = qRC − qRL − qe − qc − qs + qa − qp
A dt
where U is the heat stored in pond per units of area and
time, qRC , solar heat gain; qRL , radiation from pond surface
to air; qc , heat flux due to the convection from the pond sur-
face; qe , heat flux due to evaporation from the pond surface;
qs , heat flux from sprays, qa , heat flux input from outside
and qp , the conduction heat loss to the earth surrounding the
pond.

where d is the zero-plane displacement and depends on the

building height; z0 , roughness height; U∗ , frictional velocity;
k, Karman’s constant (0.38).
Coupling this velocity equation as a boundary condition
of the zonal model, it is possible to obtain all horizontal air
flow rates perpendicular to the street (Fig. 3). Fig. 3. Air movement perpendicular to the street axis.
406 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413
Finally, the presence of trees inside the canyon influences
the distribution of incoming solar radiation in the canyon
and the shading of the canyon surfaces, and thus for every
cell of the canyon it is necessary to investigate if there is
shading due to trees. Also, the calculation of view factors
between the cells of the grid is affected by the trees. In
that way trees are taken into account in the calculations
of the shaded area of the canyon (considering the tree’s
canopy as an obstacle) and in the calculations of the view
factors between the canyon elements. Trees are considered
to emit long-wave radiation with tree temperature equal to
air temperature, as it is explained in more details below (see
Fig. 10).
2.2. Validation using experimental results
Two experimental campaigns have been used to validate
the GreenCanyon program results. The first one, was real-
ized for a real urban layout, by the professor David Pearl-
mutter of Ben-Gurion University in Israel, during the 1998
summer [6]. The professor David Pearlmutter and Francisco
Sánchez de la Flor, 2 years latter, realized the second one,
using a scale model. In both cases, the measurements were
made in terms of air temperature and air velocity in several
heights in the longitudinal central plane of the street, sur-
faces temperatures in both walls and in the ground, and also
absorbed solar radiation.
Each measurement variable was used in the validation in
a different way. During a first stage, the absorbed solar ra-
diation by each surface was used to estimate, indirectly, the
surface albedo. In a second stage, the air velocity measured
above roof was used as an input data for a CFD program, in
order to calculate, with a high definition, the airflow pattern
inside the canyon. Finally, as a last stage of the comparison
procedure, using the outdoor air temperature and air velocity,
the surfaces albedos, and the geometrical and geographical
Fig. 4. Scheme and picture of the experiment site.
data of the canyons in the experiments, the program Green-
Canyon was run. The validation was made by comparing the
surfaces temperatures, and the absorbed solar radiation for
both walls and for the ground.
The first set of experiments used for the comparison were
carried out in two pedestrian streets of aspect ratio H/W close
to 1, of Dimona city in Negev desert (Israel). The following
figure shows a scheme and a picture of the experiment site
(Fig. 4).
Comparing these experimental results with those obtained
by GreenCanyon program, it can be concluded that they are
in good agreement, as it can be seen from the following fig-
ures, where the program results are represented with a line,
and the experimental results with dots. In these figures, the
air temperature has been measured at the central point of the
canyon, and is compared with the average values obtained
from the program. On the other hand, the surface tempera-
tures have been measured at a central point of the ground,
and both façades (façade 1 is facing southeast, and façade 2
is facing northwest). The deviations in surface temperatures
are due to different reasons like the precision of the measure-
ments devices, the goodness of the model input data, and
mainly to the fact that the measurements have been done at
a certain point, and the model results refer to average values
(Figs. 5 and 6).
The second group of experiments had as one of the ma-
jor advantages the possibility of doing a large sensitivity
analysis, since the urban layout, the distance and height of
the buildings, and their orientation could be changed easily.
The scale model was composed by concrete blocks of 15 cm
width × 15 cm height × 30 cm length. Additionally, a high
number of rows of blocks were placed in the upwind zone
of the experimental site, thanks to a very constant wind di-
rection (from northwest).
The experimental scheme and a picture of the experi-
mental site with measurement devices are shown in Fig. 7.
 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413 407

Average Air Temperature inside the Canyon Average Surface Temperature, Ground
40 55
Surface Temperature

35 50
Air Temperature

45
30 40
35
25 30
20 25
20
15 15
12:00 16:00 20:00 0:00 4:00 8:00 12:00 16:00 20:00 0:00 4:00 8:00

Fig. 5. Air and soil temperature.

Average Surface Temperature, Façade 1 Average Surface Temperature, Façade 2

35 35
Surface Temperature
Surface Temperature

30 30

25 25

20 20

15 15
12:00 16:00 20:00 0:00 4:00 8:00 12:00 16:00 20:00 0:00 4:00 8:00

Fig. 6. Temperatures of the façades 1 and 2 in the canyon with an aspect ratio of 1.

Fig. 7. Scheme and picture of the experimental site.

As can be seen from the scheme, three different aspect dots and triangles, depending on the measurement device
ratios (height to width) were checked: narrow streets (temperature sensor measurements are represented with dots,
(H/W = 3), medium streets (H/W = 1) and broad streets and infrared camera measurements with triangles). These
(H/W = 0.3). comparisons reveal in most of the cases a good agreement
Some of the results of this comparison are also shown be- between program results and experimental ones. The devia-
low (Figs. 8 and 9). In these figures the program results are tions are due to the same reasons explained in the previous
represented with a line, and the experimental results with comparison, and additionally, in this case a comparison be-

Average Air Temperature inside the Canyon Average Surface Temperature, Ground
45 60
Surface Temperature

55
Air Temperature

40
50
35 45
40
30 35
30
25
25
20 20
12:00 16:00 20:00 0:00 4:00 8:00 12:00 16:00 20:00 0:00 4:00 8:00

Fig. 8. Air and soil temperature with an aspect ratio of 1.

408 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413
Average Surface Temperature, Façade 1
50
Surface Temperature
45
40
35
30
25
20
12:00 16:00 20:00
Fig. 9. Temperatures of the façades 1 and 2 with an aspect ratio of 1.
tween both kinds of experimental results shows also devia-
tions of the same order.
3. Coupling with a building thermal simulation
program
The described models have been useful to estimate the
modification of the outdoor conditions in urban canyons,
provided by the buildings and their surroundings. When the
focus is on the building thermal behaviour, and the build-
ing energy requirements must be calculated, then a build-
ing thermal simulation program must be used. Furthermore,
it would be very interesting, and in many cases crucial, to
undertake this analysis in conjunction with an urban model
like the previous one.
The output of the Urban Canyon model has been used in a
building simulation model in order to derive the correlations
presented in the following sections.
3.1. Modifications introduced by the urban context
The presence of buildings modifies in a different degree
all the energy balance terms in an urban context [7]. In
order to understand the origin of these modifications, a brief
description for each energy balance terms in an urban context
is done in the following paragraphs.
3.1.1. Air temperature
Air temperatures in densely built urban are higher than
the temperatures of the surrounding rural country. This is a
known consequence of increasing urbanisation, urban traffic,
and the decreasing of vegetation and trees. The phenomenon
is known as ‘heat island’ and its main effect on urban envi-
ronment is to increase temperatures especially during sum-
mer period.
On the other hand, in many cases, this effect can be miti-
gated by using the ability of green areas and mass of water
(urban heat sinks) to modify the air temperature of an ur-
ban environment. This cooling effect takes place when hot
air is put into contact with a surface at a lower temperature.
The vegetation and the water are able to maintain a lower
temperature due to the evaporation that acts as a regulatory
mechanism. In order to analyse the effect of a cool surface
Average Surface Temperature, Façade 2
50
Surface Temperature
45
40
35
30
25
20
0:00 4:00 8:00 12:00 16:00 20:00 0:00 4:00 8:00
on the air temperature is necessary the use of a complex
program like a commercial CFD code, or for very simple
urban configurations, the use of simplified programs based
on correlations or on experimental results [1,8].
3.1.2. Wind pattern
The presence of buildings acts over the wind as a shelter
decreasing in general the wind speed, and always modifying
the pattern of airflow.
If a building is part of an urban area the flow pattern de-
pends on the geometry of the array, especially the height to
width ratio (aspect ratio). When the buildings are relatively
widely spaced (H/W < 0.3) their flow pattern appears al-
most the same as if they were isolated. At spacings closer
than these the main flow starts to skim over the buildings
tops and drives a lee vortex in the street or courtyard. This
vortex is reinforced by the downward deflection produced
by the windward face of the succeeding building [9].
Although in general we can expect that wind speeds within
the urban canopy are reduced in comparison with rural winds
at the same height, there are some situations in which the
opposite can be found. One of these situations occurs when
the faster moving upper air layers are either deflected down-
wards by relatively tall buildings or are channelled into ‘jets’
along streets oriented in the same direction as the flow. An-
other one occurs when regional winds are very light or calm.
In this situation, the horizontal temperature gradient can be
sufficient to induce a low-level breeze from the country into
the city in the same manner as a sea breeze.
In this case, like in the study of air temperature modifica-
tion, a CFD code is normally necessary. Another possibility
is to apply a zonal model to a the volume between buildings.
Then, the whole volume is divided into a little number of
cells, and equations for mass balance and heat balance have
to be written for each cell. Additional equations describe
conduction through the walls, convection between outside
wall surfaces and cells, and long-wave and short-wave radi-
ant interchange among the outside wall surfaces.
3.1.3. Short-wave and long-wave radiation
The most important radiative effects are a decrease in the
solar radiation receipt by areas in shadow, a local increase
in solar receipt by reflection from sunlit walls, and the re-
duction of net long-wave cooling from surfaces near the
 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413
building. This last heat lost reduction is due both to a reduc-
tion in losses by long-wave exchange with sky, caused by
the reduced sky view factor, and to an increase in gains by
long-wave exchange with the usually warm buildings. Also
in the immediate vicinity of a building, soil and air temper-
atures are often warmer than in an open area due to heat
losses from the building, and to the wind shelter provided
[10].
The reduction in energy received at the ground is greatest
at low solar elevations, i.e. in the early morning and late af-
ternoon hours. In winter and autumn the frequent low-level
inversions contribute to the accumulation of pollutants and
hence to the radiation loss. In spring the generally higher
wind velocities and, in summer, the greater convection con-
tribute to the dispersal of pollutants and thus the relatively
smaller radiation loss [11].
The presence of obstacles like trees complicates extremely
the calculation of the multiple reflections in the surround-
ing of a building. Furthermore, the long-wave exchange is
controlled by a huge number of different surfaces and, in
consequence, of view factors. All these reasons lead to the
development of a calculation method able to take into ac-
count the direct, diffuse and reflected radiation, and to treat
with many surfaces.
3.1.4. Water in street landscaping
Ponds and fountains can be effective air conditioning sys-
tems in open spaces because of their ability to keep water
temperatures lower than air temperature, and their low re-
flectivity. Ponds have a reflectivity of approximately 3% at
times of maximum solar radiation, and therefore, reflect lit-
tle solar radiation towards occupied zones. They absorb a
lot of solar radiation, up to 80% depending on the depth
of pond. All this solar radiation does not, however, produce
a significant increase of water temperature because of the
pond’s thermal inertia and evaporation at its surface. The
water pond inertia is directly proportional to water mass,
and therefore, proportional to its depth. With increasing wa-
ter pond inertia, the water temperature decreases. The daily
range of water temperature (maximum to minimum differ-
ence) is reduced and there is a phase shift between air and
water temperatures.
When the pond is in shadow, the incoming solar radiation
is reduced, with a reduction in water temperature. This tem-
perature reduction increases with increased shading of the
pond.
Additionally to an energy balance of the water pond, an
analysis of the cooling effect provided by sprays is necessary
in fountains. A single water drop moving through the air
experiences the following processes:
• Heat flows from the air to the drop (if the air is hotter than
the drop).
• Water evaporates from the drop to the surrounding air.
The hotter the drop is, the more water will be evaporated.
• The drop slows down as it moves through the air.
409
The first two processes affect the temperature of the drop
in different ways. Heat transfer will warm it up but evapo-
ration will cool it down. As a result of these two opposite
tendencies, an equilibrium drop temperature is reached (the
wet bulb temperature of the air). Once the drop has reached
the wet bulb temperature, the extra energy needed to evapo-
rate more water has to come from the surrounding air. This
means the surrounding air is cooled.
3.1.5. Trees in street landscaping
The complexity of a vegetated surface as a system of sinks
and sources of heat, mass and momentum is such that an ex-
act description of its physical behaviour is almost impossi-
ble. Two problems are faced to describe a correct behaviour
of a vegetated system. The first one is linked to the com-
plexity and the heterogeneity of the foliage and the second
one, to the turbulent nature of the air flow within and above
the canopy. These reasons force us to use a simplified model
based on reasonable assumptions [12,13].
The main effects of trees on urban climate are shading,
wind shielding and cooling effect by absorbing solar radia-
tion which is then dissipated by evapotranspiration and sen-
sible exchange with the environment.
Using the global energy balance of a leaf, its surface
temperature can be calculated. The difference between the
leaf temperature and the air temperature depends on the
value of the diffusive resistance and the gap between the net
long-wave and solar radiation flux, and the latent heat flux.
In this, if the net radiation flux is almost dissipated by the
evaporation process, the leaf temperature will be very close
to the ambient temperature. Furthermore, evapotranspiration
estimation in urban environment is very complex for trees
because lots of parameters have to be taken into account
(available water, stomata closure, shade and sunlight, wind
movement,. . . ) and it should be easy in a first level to let
the temperature of trees at a value close to air temperature.
In that way, the cooling effect is included via the absorp-
tion of the solar radiation by the foliage of the trees without
an increase of the foliage temperature, i.e. no sensible heat
transfer with the outside air.
The following three figures represent the thermograph im-
ages of a canopy tree when the outdoor air temperature was
28 ◦ C. In these figures, from left to right, the white dots
represent the isotherms of 27, 28 and 29 ◦ C, respectively
(Fig. 10).
3.2. Characterization of the climate influence on the total
building energy demand
The analysis of the exterior conditions over the cooling
and heating demands of a giving building could be studied
in an easy and clear way using the concept of climatic sever-
ity index (CSI, as defined below). Previously, it is necessary
to express the relation between the CSI and the building de-
mands, and between the CSI and the known climatic vari-
ables.
410 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413
Fig. 10. Thermograph images of a canopy tree in 28 ◦ C air temperature. Isotherms of 27, 28 and 29 ◦ C.
Two winter climatic conditions that in principle are
“different”, can be considered “identical” when the heating
demand of a certain building is the same in both climates.
Then, it can be said that both climatic conditions have the
same “CSI of winter”. The same definition can be applied
to the cooling demand, and in that case it is called “CSI of
summer”.
Furthermore, it can be said that a climatic condition is
“x” times more severe than other if the energy demand of
a certain building is “x” times higher in the first case than
in the second one. These CSIs are not depending on the
building for which they have been calculated, as it has been
demonstrated by the Thermal Energy Group of Seville [14].
Using the CSI concept, the influence of a climatic variable
change over the cooling and heating demands of a certain
building can be estimated, assuming that this change affects
in the same way the whole building. A typical modification
like this is the variation of the external air temperature, since
it modifies the degree days. Then an increase of the DD can
be expressed as an increase of the CSI [15].
n  n 2
CSI = a × DD + b + c × DD2 + d +e
N N
where DD is the mean value of the summer degree days
(June, July, August and September) with a base temperature
of 20 ◦ C; n/N, ratio between the actual insolation hours and
the maximum for each month.
a b c d e building characteristics.
1.090E−02 1.023E+00 −1.638E−05 −5.977E−01 −3.370E−01
3.3. Characterization of the climate influence on the
contribution to the building energy demand by components
The CSI index can be used to relate heating and cooling
energy demand of a building with ambient air temperature
(as ambient air temperature is assumed to be in the same
value around the building; irrespective of orientation; and
thus, any change affects in the same way the energy demand
of the building).
When the climatic change affects in a different way to dif-
ferent parts of a building, because it is, for instance, a func-
tion of the orientation; the previous methodology can not
be applied. That methodology could be applied only to the
whole building, and not to part of the building. In this case it
is necessary to calculate the effect of those climatic modifi-
cations on the building energy demand by components, due
to each roof, exterior wall and window. A typical climatic
variable that affects in a different way to each component
of a building is the solar radiation.
Thus, the first stage has to be to obtain the correlations be-
tween the gains or losses through those components (roofs,
exterior walls and windows) and the known climatic vari-
ables.
As in the case of characterization of the climate influ-
ence on the total building energy demand, for this other
case, it is possible to obtain the increase of each contribu-
tion as functions of the climatic variables and their incre-
ments.
4. Influence on the building energy demand
4.1. Air temperature
The effect that a modification of the urban air temperature
has over the heating and cooling energy demands of a certain
building depends on the climatic conditions, and on the own
Using the methodology introduced for the evaluation of
the climate influence on the total building demand, it is
possible to obtain the variation of the CSI as a function of
the air temperature increment.
As an example, let us assume that a certain urban context
is responsible of an average air temperature decrease of 10◦
per h by day during the cooling period (base temperature for
the DD calculation is 20 ◦ C). Then, as a direct consequence,
it can be calculated that DD = −12.7. Assuming also that
the building is placed in Seville (Spain) where DD = 175.61,
and CSI = 1.51, then the CSI modification is CSI =
−0.07, and in consequence CSImodified = 1.44.
 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413 411

Cooling Demand Saving Contribution to the Cooling Demand

Due to the Decrease of Air Temperature Wall: U=1 W/m ² ºC ; Window: U=5.7 W /m² ºC and SC=1
7 180
Cooling Demand Saving

Absolute Decrease (kWh/m2)

6 160
5 140
(k Wh/m2)

Burgos 120
4 10
Madrid 100
3 50
Sevilla 80
2 90
60
1 40

0 20
0 5 10 15 20 0
0 20 40 60 80 100
Mean Decrease of Air Temperature (ºC)
Percentage of Original Incident Radiation
Fig. 11. Absolute saving of cooling demand.
Fig. 13. Absolute decrease of contribution to the cooling demand.

This result can be interpreted in two different ways.

Firstly, following the CSI definition, it can be said that the
building performance in this modified urban context, will
be the same that the corresponding building performance
in a non-modified urban context placed in another location
with CSI = 1.44. This could be, more or less, the value of
Jaén (Spain). Secondly, taking into account that the CSI is
a dimensionless energy demand, these values can be used
to obtain the real values, and then the energy saving due to
the climatic modification, that for this example is 6%.
A sensitivity analysis of the savings of cooling demands
that can be achieved due to the decrease of mean air tem-
perature depending on this last value, and on the location, is
shown in the figures below. In these figures, it can be seen
that the savings are proportional to the mean decrease of
outdoor air temperature, and it is also bigger for cold cities,
this is, it is inversely proportional to the CSI (Seville: 1.51,
Madrid: 1, Burgos: 0.43) (Figs. 11 and 12).
4.2. Solar radiation
The effect of the modification of the solar radiation due
to the urban context, on the building energy demand, has
to be calculated for each external component (roofs, walls
and windows), since this modification affects them in a dif-
ferent way. Therefore, it is necessary to use the method-
ology introduced above, but considering the climate influ-
ence on the contribution to the building energy demand by
components.
As an example, let us assume that the impinging solar ra-
diation over the south façade of a certain building has been
decreased in a 10% of the original value, due to urban con-
text modifications like new surfaces providing shades, low
albedo materials, etc. If for instance, the façade is com-
posed by a wall of thermal global transmission coefficient
U = 1 W/m2 ◦ C, and of a window with U = 5.7 W/m2 ◦ C
and shading coefficient SC = 1; then, the saving in terms
of contribution of this façade to the building energy demand
can be estimated.
The following figures show the decrease of this con-
tribution assuming for a building located in Seville, and
taking into account three different percentage of glazing
in the façade (10, 50 and 90%). The first figure repre-
sents the absolute values and the second one the relative
ones. In both figures, the x-axis refers to the percentage
of the original incident solar radiation, this is, a value
of 100% means that there is not any reduction (Figs. 13
and 14).
As can be observed, the decrease variation is proportional
to the original impinging solar radiation, and if it were com-
pared for other locations, almost the same results in absolute
values will be obtained, since the variation of the total so-
lar radiation is very little for the latitude range considered
(Seville: 37.4, Madrid: 40.41, Burgos: 42.3). Another im-
portant conclusion is that this saving is higher with higher
percentage of glazing in the façade.

Contribution to the Cooling Demand

Cooling Demand Saving Wall: U=1 W/m² ºC ; Window: U=5.7 W/m² ºC and SC=1
Due to the Decrease of Air Temperature
100% 100%
Cooling Demand Saving

80% 80%
Relative Decrease

Burgos 60% 10
60%
Madrid 50
40% 40% 90
Sevilla
20% 20%

0% 0%
0 5 10 15 20 0 20 40 60 80 100
Mean Decrease of Air Temperature (ºC) Percentage of Original Incident Radiation

Fig. 12. Relative saving of cooling demand. Fig. 14. Relative decrease of contribution to the cooling demand.
412 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413

4.3. Surface temperature Contribution to the Cooling Demand

Wall: U=1 W/m² ºC ; Window: U=5.7W/m² ºC and SC=1
1.5
The effect of a surface temperature modification is more
Absolute Decrease( kWh/m2)

1
complicated than the previous ones. On one hand, if the sur- 0.5
face belongs to the building under consideration, its temper- 0 10
ature modifies the heat flux only by conduction; but if this -0.5 50
surface does not belong to the building (e.g. opposite side -1 90
buildings, street surface), then it is able to modify both: the -1.5

heat flux by long-wave radiation, and the outdoor air tem- -2

-2.5
perature. -30 -20 -10 0 10 20 30 40
The first effect is due to the modification of the surface Surface Temp. Increaseover Air Temp.
albedo, and it has been studied included in the solar radiation
Fig. 15. Absolute decrease of the cooling demand.
study. The modification of the outdoor air temperature has
been also studied, so the only effect that has to be study is
Contribution to the Cooling Demand
that due to the long-wave radiation modification. Wall: U=1 W/m² ºC ; Window: U=5.7W/m² ºC and SC=1
In this case, it is possible to use the analogy between 2%
the surface temperature and the air temperature, using the
1%
Relative Decrease

concept of equivalent temperature defined as follows:

0% 10
hc ta + hr tr
teq = 50
hcr -1% 90

-2%
where hc is the outdoor convective transfer coefficient;
hr , outdoor radiant transfer coefficient; hcr , outdoor -3%
-30 -20 -10 0 10 20 30 40
convective-radiant transfer coefficient hcr = hc + hr ; ta ,
Surface Temp. Increase over Air Temp.
outdoor air temperature; tr , outdoor radiant temperature.
With this equation, and using the concepts of radiant tem- Fig. 16. Relative decrease of the cooling demand.
perature, and radiant heat transfer coefficient, it is possible
to express the relation between them and the surface tem-
perature, and finally with the air temperature. function of the increment of surface temperature. As can be
seen from these figures, the savings are higher for higher
tr = FFsky × tsky + (1 − FFsky )tsuf window to wall ratio, but in all cases they are very low
(Figs. 15 and 16).
hr = 4σ(tr + 273.15)3

where σ is the Stefan–Boltzmann constant σ = 5.67E −

5. Conclusion
8 W/m2 K4 ; tsuf , mean surface temperature; tsky , sky tem-
perature; FFsky , form factor with the sky.
The complexity of the urban context, and the particular-
Therefore, an increment of surface temperature, tsuf ,
ities of each case compare to a rural case, are the reasons
gives an increment of radiant temperature, tr = (1 −
that explain the difficulty of obtain general results, and con-
FFsky ) tsuf , that can be expressed in terms of increment of
clusions. Each case has to be analyse separately, and for it,
equivalent temperature, teq = (hr / hcr ) tr . The same in-
all the design alternatives have to be compare in terms of
crement can be obtained modifying the air temperature in-
energy requirements.
stead of the surface temperature, and then it can be con-
In this paper, an urban canyon model is presented, show-
cluded that, ta ≡ (hr / hc ) tr .
Assuming hc ∼ = 20 W/m2 ◦ C, and hr ∼
ing the main characteristics of the integrated sub-models.
= 5 W/m2 ◦ C, then
ta ∼
This model has been validated, and the results obtained from
= (1/4) tr , and finally if the sky form factor is equals
it can be used in order to modify the inputs for a building
to 0.5
    thermal simulation program.
hr 1 Energy consumption of buildings are related to solar
ta = (1 − FFsky ) tsup ∼= 0.5 tsup
hc 4 loads, wind flow patterns and external air temperature. So,
  improvements on urban microclimate should have direct
1
= tsup and indirect consequences on energy savings. Therefore,
8
programs and studies like the one presented above, are
Thus, in this case can be used the same methodology that completely necessary to assess the influence of the urban
in the case of modification of air temperature. context on building energy demand.
For the previous example, the absolute and relative de- Using an urban canyon model in conjunction of a build-
crease of the total cooling demand can be represented as a ing thermal simulation program, is possible to obtain cor-
 F.S. de la Flor, S.A. Domı́nguez / Energy and Buildings 36 (2004) 403–413
relations like those presented above, where the influence
of each climatic modification on the building requirements
can be seen separately. Thus, the presented methodology
has been useful to obtain quantitative results for the in-
fluence of the outdoor air temperature decrements, the
reduction of solar radiation, and the variation of surface
temperatures.
Acknowledgements
The experimental work presented in this paper has
been possible to the grant that one of the authors (Fran-
cisco Sánchez de la Flor) obtained for the stay during
the summer of 2000 at the Jacob Blaustein Institute for
Desert Research in Israel support by the “Improving Hu-
man Potential–Access to Research Infrastructure”. Other
experimental results used in this paper were obtained by
the professor David Pearlmutter of the Jacob Blaustein
Institute.
References
[1] GreenCode Project, Reglementary Frame For Renewable Energy Use
In Urban Site Through Vegetation Plantings And Strategic Surfacing,
ALTENER Project, Contract No. XVII/4.1030/Z/98–172.
[2] D.G. Stephenson, G.P. Mitalas, Calculation of heat conduction trans-
fer functions for multi-layer slabs, ASHRAE Transactions 77 (Part
II) (1971) 112–126.
[3] J.L. Molina, Simulación Energética de Edificios Mediante un Modelo
de Respuesta en Frecuencia, Thesis Doctoral, Universidad de Sevilla,
Sevilla, 1987.
413
[4] M.M. Elnahas, T.J. Williamson, An improvement of the CTTC model
for predicting urban air temperatures, Energy and Buildings 25 (1997)
41–49.
[5] S.E. Nicholson, A pollution model for street-level air, Atmospheric
Environment 9 (1975) 19–31.
[6] D. Pearlmutter, Street Canyon Geometry and Microclimate: Design-
ing for Urban Comfort Under Arid Conditions, PLEA’98, 1998,
pp. 163–166.
[7] F. Sánchez, S. Alvarez, Assessing the influence of the urban context
on building energy demand, EPIC 2002 AIVC, in: G. Guarracico
(Ed.), Proceedings of the Third European Conference on Energy
Performance and Indoor Climate of Buildings and Twenty third
Conference of the Air Infiltration Ventilation Centre, Lyon, France,
October 2002. Energy Efficient and Healthy Buildings in Sustainable
Cities, 2002.
[8] URBACOOL Project, Energy efficient cooling systems and tech-
niques for urban buildings, European Commission DG XVII For
Energy, SAVE II Programme, Contract No. XVII/4.1031/Z/98–224.
[9] F. Sánchez, Estudios Numéricos de la Capa Lı́mite Urbana Orientados
a Mejorar el Microclima Alrededor de Edificios, Proyecto fin de
carrera, Universidad de Sevilla, ETSII, 1998.
[10] T.R. Oke, Boundary Layer Climates, Routledge, London, UK, 1987,
ISBN 0-415-04319-0.
[11] H. Landsberg, The urban climate, in: International Geophysics Series,
vol. 28, Academic Press, 1981.
[12] J. Kondo, T. Watanabe, Studies of the bulk transfer coefficients over
a vegetated surface with a multilayer energy budget model, Journal
of Atmospheric Sciences 39 (23) (1992) 2183–2199.
[13] M. Bruse, H. Fleer, Simulating surface–plant–air interactions inside
urban environments with a three dimensional numerical model, En-
vironmental Modelling and Software 13 (3–4) (1998) 373–384.
[14] Ministerio de Fomento e IDAE, Fundamentos Técnicos de la Califi-
cación Energética de Viviendas, Centro de Publicaciones Secretarı́a
General Técnica Ministerio de Fomento, 1999.
[15] F. Sánchez, Modificaciones Microclimáticas Inducidas por el Entorno
del Edificio y su Influencia sobre las Demandas Energéticas de
Acondicionamiento, Thesis Doctoral, Universidad de Sevilla, ETSII,
2003.