EFECTOS NATURACIÓN

7/30/2019 EFECTOS NATURACIN

1/14

Building and Environment 43 (2008) 480493

Temperature decreases in an urban canyon due to green walls and green

roofs in diverse climates

Eleftheria Alexandria, Phil Jonesb,

aMantzakou 2-6, 114 73 Athens, GreecebWelsh School of Architecture, Cardiff University, King Edward VII Avenue, Cardiff CF10 3NB, UK

Received 23 February 2006; received in revised form 24 July 2006; accepted 31 October 2006

Abstract

This paper discusses the thermal effect of covering the building envelope with vegetation on the microclimate in the built environment,

for various climates and urban canyon geometries. A two-dimensional, prognostic, micro scale model has been used, developed for the

purposes of this study. The climatic characteristics of nine cities, three urban canyon geometries, two canyon orientations and two wind

directions are examined. The thermal effect of green roofs and green walls on the built environment is examined in both inside the canyon

and at roof level. The effects of this temperature decrease on outdoors thermal comfort and energy savings are examined. Conclusions

are drawn on whether plants on the building envelope can be used to tackle the heat island effect, depending on all these parameters

taken into consideration.

r 2006 Elsevier Ltd. All rights reserved.

Keywords: Green roofs; Green walls; Urban canyon; Plants; Built environment

1. Introduction

Since the beginning of human existence man has clearly

intended to alter his microclimate, to a more human-

friendly one, protecting himself from extreme climatic

conditions. Even from the first evidence of Neolithic houses

and settlements, it is obvious that they were not sited in a

purely natural environment, but in a part of nature

transformed according to a human plan [1]. With the

evolution of human societies, settlements were trans-

formed, evolved into villages, towns or cities, developed

or faded away, according to the geographical, economic,

social and cultural transformations taking place through-out time. With the Industrial Revolution, urban spaces

expanded dramatically, much faster and with much more

significant changes than in their previous evolutionary

periods. The large areas modern cities occupy, their

structure, materials and the general lack of vegetation

cannot but have altered the climatic characteristics of

urban spaces.

These changes have a direct effect on the local climate ofurban spaces, especially the central parts of the city,

causing a significant rise of the urban temperature and

other alterations, known as the heat island effect. This may

cause serious local climatic unpleasant conditions and even

imperil human health, especially for cities in climates with a

distinctively hot season [2,3]. The moderation of extreme

heat in the local environment of such climates could mean

not only their sustainability, but also the potential of

occupying them without the morbidity and mortality risks

caused by excessive heat [4,5].

On prima facie evidence, the general lack of vegetation in

existing cities is one of the factors affecting the formationof raised urban temperatures. In most urban spaces,

appreciable amounts of vegetation exist mostly concen-

trated in parks or recreational spaces. Although parks

manage to lower temperatures within their vicinity [69],

they are incapable of thermally affecting the concentrated

built spaces where people live, work and spend most of

their urban lives. By placing vegetation within the built

space of the urban fabric, raised urban temperatures can

decrease within the human habitats themselves and not

only in the detached spaces of parks. Urban surfaces which

ARTICLE IN PRESS

www.elsevier.com/locate/buildenv

0360-1323/$- see front matter r 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.buildenv.2006.10.055

Corresponding author.
http://www.elsevier.com/locate/buildenvhttp://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.buildenv.2006.10.055http://localhost/var/www/apps/conversion/tmp/scratch_3/dx.doi.org/10.1016/j.buildenv.2006.10.055http://www.elsevier.com/locate/buildenv


2/14


3/14

specific humidity (in kg/kg), c1 is the isobaric specific heat

of component 1 (moisture) of the mixture (in J/kg K) and c2the isobaric specific heat of component 2 (air) of the

mixture (in J/kg K). According to Eckert and Drake [12],

the effect of thermal diffusion on heat transfer (fourth term

of Eq. (1)) can be neglected in normal engineering mass-

transfer processes. However, they point out [12] that itcontributes essentially when temperature gradients are

extremely large. In the description of the thermal ex-

changes in the built environment, where temperature

gradients in the boundary layer of surfaces exposed to

direct solar radiation are relatively large, the expression of

thermal diffusion is essential for the accurate description of

the phenomenon.

When air velocity is considered, eddy diffusion is much

stronger than molecular diffusion (conduction) in the air in

the atmosphere away from the boundary layer of the

surface. Despite the fact that molecular diffusion always

takes place in the air, it is omitted from both heat and mass

transfer in the air at these levels, as it is 104105 smaller

than eddy diffusion [13]. The effect of vapour gradients

onto temperature in the air nodes well above the ground is

expressed through eddy diffusion coefficients [14,15]. Heat

and mass transfer in the two-dimensional model of the

binary airwater vapour mixture thus becomes:

q

qt u

q

qx w

q

qz

T

q

qzKHz

qT

qz

q

qxKHx

qT

qx

,

(3)

q

qt u

q

qx w

q

qz q q

qzKEz

qq

qz q

qxKEx

qq

qx , (4)

where KHz and KEz are the eddy diffusion coefficient of

energy and water vapour, respectively, in the vertical axis

and KHx and KEx are the respective diffusion coefficients in

x-axis. The expression of these diffusion coefficients is

given by the MoninObukhov similarity theory [14,15].

The water vapour gradients are taken into consideration in

the calculation of the eddy diffusion coefficients of energy.

Regarding solid materials, they are considered as a

system, consisting of a capillary-porous building material

in the medium of wet air and in a region of positive

temperatures (no ice). The equations describing the one-

dimensional heat and mass transfer can be expressedby [11]

dT

dt ac

q2T

qz2

l

cc

qq

qt, (5)

dq

dt am

q2q

qz2, (6)

where ac is the building material thermal diffusion

coefficient (in m2/s), cc is the building material specific

heat capacity (in J/kg K), e is the evaporation number of

the building material, and am is the diffusion coefficient of

moisture in the building material (in m2/s).

Regarding plants, they are considered to be a layer

consisting of canopy leaves and the air among them.

Equations describing heat and mass transfer in the air are

the ones given by Eqs. (1) and (2), while heat transfer in the

leaf is given by

rcpdT

dt Fn C lE, (7)

where r is the density of the leaf tissue (in kg/m3), cp is

the specific heat capacity of the leaf tissue (in J/kg K), Tis the leaf surface temperature (in K), Fn is the net heat

gain from radiation (in W/m2), C is the net sensible heat

loss (in W/m2) and lE is the net latent heat loss (in W/m2).

Radiative heat exchanges between the canyon surfaces

have been described analytically, according to the radiative

heat transfer theory [16] and not with the use of a

combined convection and radiative heat transfer coeffi-

cient. Thus the radiative heat exchanges between surfaces

with different emissivities in closed enclosures is expressed

by

qii

XNj1

1

j 1

Fijqj H0i

XNj1

FijEbi Ebj for i 1; 2; . . . ; N, 8

where qi is the radiation emitted from the surface i(W/m2),

ei is the isurfaces emmisivity and ej is the emmisivity of the

rest of the surfaces, Fij is the view factor of surface i

towards surface j, H0i is any external radiation arriving at

surface i, and Ebi equals to sTi4, where s is the

StefanBoltzmann constant (5.67 108

W/m2

K4

) and Tiis the temperature of the ith surface.

Climatic characteristics, such as air temperature, relative

humidity and wind speed, are set as the boundary nodes of

the model, placed 10 m above the upper part of roofs.

These climatic characteristics, as well as solar radiation

derive from meteorological data from METEONORM

[17]. Solar radiation is input onto the surfaces, according to

their orientation, inclination and shading pattern. The

shading pattern, determined by the canyon geometry and

the geographic latitude, was calculated with the software

ECOTECT [18], where the same canyon geometries, as the

ones described below were input, for the different latitudes

and longitudes examined. Air velocities in the vicinity of

the canyon were calculated with the CFD code WinAir4

[19]. WinAir4 is an in-house code, which uses the fixed

viscosity models, with a variation on simple solution

scheme. The canyon geometries in the CFD model are

the same as for the heat and mass transfer model. As the

CFD model is three-dimensional, the length of the canyon

was 40 m and the rest of the canyon dimensions (height,

width) varied, according to the canyon geometry, as

discussed below. The CFD code mesh was also the same

as for the two-dimensional heat and mass transfer model,

as described in Fig. 1. The grid is not uniform; near the

building and road surfaces the grid is 0.30 m, while two

ARTICLE IN PRESS

E. Alexandri, P. Jones / Building and Environment 43 (2008) 480493482


4/14

nodes away from the surface, the grid varies according to

the canyon geometry, from 1 to 3 m for the canyon width,

and from 1 to 2 m for the canyon height. For the length, it

is constantly set at 2 m. At the boundary surface are set

10 m away from the roof of the buildings, 10 m away from

the windward surface and 100m away from the leeward

surface when the building height is 5 m, and 10, 20 and200 m, respectively, when the building height is 10 m. The

input climatic data of wind speed, air temperature and

relative humidity are input as the boundary conditions for

each hour of the diurnal profile. The heat gains on

buildings and the street are calculated according to the

solar radiation absorbed by the surface, which depends on

its orientation, and shading pattern, the latter having been

defined by [18]. The output air velocities from the CFD

code in the middle of the canyon are input at the respective

nodes of the two-dimensional heat and mass transfer

model.

Four types of vegetation covering the building envelope

are examined for each canyon geometry:

(a) a base case, where no green is placed in and around the

canyon, referred to as the no-green case,

(b) the green-roofs case, where both roofs are covered

with vegetation (ground-covering grasses)

(c) the green-walls case, where both walls inside the

canyon are covered with vegetation (ivies) and

(d) the green-all case, where both roofs and walls are

covered with vegetation.

Three types of canyon geometries are examined, accord-

ing to the wind flow developed in each:

(a) a canyon with height (H) 10m and width (W) 5 m ,

referred to as H10W5 canyon, where, according to

Santamouris [6], skimming flow is developed, with very

low air velocities, and sun shaded,

(b) a canyon with 5 m height and 10 m width, referred to as

H5W10 canyon, where wake interference flow is

developed, with bigger air velocities and more exposed

surfaces to direct solar radiation and

(c) the H5W15 canyon, with 5 m height and 15 m width,

where isolated roughness flow is developed, with much

larger air velocities, and greater exposure to solar

radiation.

The canyons are examined with two orientations:

(a) one where the canyons axis was parallel to the

EastWest axis (referred to as the EW canyon) and

(b) one where the canyons axis was parallel to the

NorthSouth axis (referred to as the NS canyon).

Two directions of wind flow are considered:

(a) one aligned to (referred as x) and

(b) one parallel to the canyons axis (referred as y).

Buildings are made of concrete, and the street is covered

with asphalt. A summary of the hydrothermal properties of

the materials and vegetation considered in the canyons is

made in Table 1. All these cases are examined for nine cities

in nine different types of climates, where cities and

evapotranspiring vegetation can be found. Based on

Koeppens climatic classification [20], the nine citiesstudied, and the climatic type in which they belong, are

summarised in Table 2. All cases are examined for a typical

day of their hottest month. Their climatic data have

derived from hourly data from METEONORM [17]. The

effect of vegetation on the urban texture of each city is

examined for its hottest month. For Athens, Hong Kong,

London, Montre al, Moscow and Riyadh, July is chosen as

the hottest month, while for Mumbai May is used, for

Beijing June, and for Braslia September. The typical day

ARTICLE IN PRESS

Table 1

Hydrothermal properties of plants, soil, building materials (concrete) andstreet materials

Characteristic Concrete Asphalt Soil Plants

Specific thermal capacity

(MJ/m3K)

1.60 2.00 1.15 2.60

Thermal conductivity (W/mK) 1.70 1.30

Vapour diffusivity (106 m2/s) 0.55 1.58

Ratio of vapour diffusion

coefficient to total moisture

diffusion coefficient

0.20 0.10

Emissivity 0.94 0.81 0.94 0.94

Albedo 0.23 0.10 0.23 0.30

Hydraulic conductivity

(10

4

m/s)

0.01

Moisture potential, when soil is

saturated (cm)

49.0

Maximum volumetric water

content (m3/m3)

0.492

Coefficient b 10.40

Convective heat resistance

(s/m)

200

Resistance expressing the plant

type (s/m)

100

Canopy extinction coefficient 1.4

Level of soil moisture below

which permanent wilting of the

plant occurs (m3/m3)

0.25

Table 2

Table of cities studied

City Climate Location

London, UK Temperate 51.32N, 0

Montreal, Canada Subarctic 45.31N, 73.34W

Moscow, Russia Continental cool summer 55.45N, 37.37E

Athens, Greece Mediterranean 37.59N, 23.43E

Beijing, China Steppe 39.48N, 116.23E

Riyadh, Saudi Arabia Desert 24.38N, 46.43E

Hong Kong, China Humid subtropical 22.16N, 114.12E

Mumbai, India Rain forest 18.54N, 72.5E

Braslia, Brazil Savanna 15.48S, 47.54W

E. Alexandri, P. Jones / Building and Environment 43 (2008) 480493 483


5/14

of the hottest month is a 24-h profile whose climatic data

are averaged for each hour of the month examined, to

make a diurnal typical climatic profile for the month

studied. The maximum, minimum, average and daytime

average quantities of temperature, relative humidity, wind

speed and solar radiation on a horizontal plane are

presented in Table 3. In Figs. 25 the diurnal profile ofair temperature, relative humidity, solar radiation on a

horizontal plane and wind speed are presented for the

typical day of the hottest month of all nine cities.

3. Discussion and analysis

3.1. Direct cooling effects

Air and surface temperatures lower significantly in all

climates examined, when walls and roofs are covered with

vegetation, as can be observed in Figs. 710. The heat

fluxes on the vegetated and on the non-vegetated surfacesare very different. As can be observed in Fig. 6 for a green

roof and for a concrete roof in Montre al, the 24-h profile of

ARTICLE IN PRESS

T

able3

M

aximum,minimum,averageandday-timeaveragevaluesofclimaticcharacteristics(totalsolarradiationonahorizontalplane,air

temperature,relativehumidity,windspeed)forthe24-hprofileof

th

etypicaldayofthehottestmonthofeach

city

C

ity

Totalsolarradiationonahorizontalplane

Airtem

perature

Relativehu

midity

Windspeed

Max

Min

Average

Daytime

average

Max

Min

Average

Daytime

average

Max

Min

Average

Day-time

average

Max

Min

Average

Daytime

average

A

thens,July

809.0

0.0

298.4

457.5

30.1

22.3

25.7

27.0

57.4

37.3

50.2

45.9

6.4

3.5

4.8

5.2

B

eijing,June

589.2

0.0

216.4

331.9

29.1

19.6

24.2

25.6

75.0

47.1

62.7

57.9

4.3

3.3

3.8

4.0

B

raslia,September

726.0

0.0

228.4

437.8

28.0

20.4

23.5

25.4

65.7

44.0

58.6

52.1

4.7

3.2

3.8

4.0

H

ongKong,July

704.3

0.0

235.8

417.2

33.1

26.7

29.5

30.9

85.1

62.9

77.1

70.9

4.9

3.0

3.8

4.2

L

ondon,

July

497.3

0.0

199.3

286.5

19.7

14.5

17.2

18.0

77.3

63.5

74.3

70.5

4.4

3.9

4.2

4.2

M

ontreal,July

692.2

0.0

261.9

376.5

25.9

17.1

21.2

22.3

77.7

49.7

66.1

62.1

5.5

3.3

4.3

4.6

M

oscow,

July

527.8

0.0

220.2

297.9

20.9

13.9

17.3

18.0

86.3

60.3

74.7

71.6

5.2

3.5

4.3

4.5

M

umbai,

May

846.9

0.0

283.8

502.1

33.5

25.3

28.9

30.8

79.1

53.1

69.5

62.4

4.6

2.3

3.3

3.8

R

iyadh,

July

850.0

0.0

284.2

466.9

42.8

31.2

36.5

38.3

45.8

23.5

36.2

32.4

5.9

3.8

4.8

5.2

0

5

10

15

20

25

30

35

40

45

1 9 11 13 15 17 19 21 23

Athens Beijing Brasilia

Hong Kong London Montreal

Moscow Mumbai Riyadh

Air Temperature

Time (Hours)

Temperature(C)

3 5 7

Fig. 3. Twenty four-hour profile of air temperature for the hottest month

of each city, which is input at the boundary nodes of the heat and mass

transfer model.

0

1 11 13 15 17 19 21 23

Athens

Beijing

Brasilia

Hong Kong

London

Montreal

Moscow

Mumbai

Riyadh

Total Solar Radiation on a Horizontal Plane

SolarRadiation(W/m2)

900

800

700

600

500

400

300

200

100

3 5 7 9

Time (Hours)

Fig. 2. Twenty four-hour profile of the total solar radiation on a

horizontal plane, for the hottest month of each city, which is input on the

unshaded, horizontal surfaces of the heat and mass transfer model.



6/14

the convective, conductive, evaporative and radiative heat

fluxes on the green and the concrete roofs differ

significantly. The convective heat flux density at the

external surface of the concrete roof is much larger than

the convective heat flux density in the upper part of the green

roofs canopy. For the concrete roof its 24-h profile ranges

from 345.1 to 128.6 W/m2, while for the green roofs upper

surface, it only ranges from 51.3 to 99.9W/m2. The

convective heat exchanges between the grass foliage and

the air are milder than those between the solid concreteroof and the air. The total radiative heat flux density (both

short and long wave radiation) on the external surface of

the concrete roof is also larger than that on the upper part

of the canopy layer. It ranges from 158.2 to 355.1 W/m2

on the concrete roof and from 38.8 to 229.5 W/m2 on the

green roof. Due to the redistribution of radiation within

the vegetated layer, the total radiative heat exchanges are

smaller on the vegetated surface, when compared with the

concrete roof. As can be observed in Eq. (7), the conductive

heat component is omitted in the relationship governing

heat transfer in plants as too small [1315,21,22], while it is

an important factor in the heat transfer of a concrete roof,

with a range from 444.5 to 154.5 W/m2 on the external

part of the roof. Nonetheless, the greatest differences are

observed at the evaporative heat fluxes, which range from

46.3 to 170.6 W/m2 for the concrete roof and from

593.2 to 26.4 W/m2 for the green roof. As the

evaporative heat transfer on the green roof acts constantly

as a heat sink and the radiative energy absorbed by the

green roof is smaller than that absorbed by the concrete

roof, the energy fluxes on a green surface can only offer

lower surface and air temperatures, when compared to

those produced by concrete surfaces.

Because of these energy distributions, canyon air

temperature lowers the most when both walls and roofsare covered with vegetation in all climates examined. This

can be explained by the fact that when roofs are covered

with vegetation, air masses enter the canyon much cooler,

from the vegetated roofs. On the other hand, when only

walls are covered with vegetation, air masses enter the

canyon heated by the plain roofs, which absorb the quite

ARTICLE IN PRESS

0

10

20

30

40

50

60

70

80

90

1 11 13 15 17 19 21 23

RelativeH

umidity(%)




100

Time (Hours)

Relative Humidity

3 5 7 9

Fig. 4. Twenty four-hour profile of relative humidity for the hottest

month of each city, which is input at the boundary nodes of the heat and

mass transfer model.

0

1

2

3

4

5

6

7

1 11 13 15 17 19 21 23




Wind Speed

WindSpeed(m/s)

3 5 7 9

Time (Hours)

Fig. 5. Twenty four-hour profile of wind speed for the hottest month of

each city, which is input at the boundary nodes of the heat and mass

transfer model.

01 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

HeatFlux

Density(W/m2)

Conv, rf-gr

Evap, rf-gr

Rad, rf-gr

Cond, rf-con

Conv, rf-con

Evap, rf-con

Rad, rf-con

400

300

200

100

-100

-200

-300

-400

-500

-600

Time (Hours)

2 3 4 5 6 8

Fig. 6. Convective (Conv), evaporative (Evap), long and short-wave

radiative (Rad) and conductive (Cond) heat fluxes on a concrete roof (rf-

con) and on a green roof (rf-gr) in Montre al. -12

-10

-8

-6

-4

-2

0

4 12 16 20 24

TemperatureDecrease(C)

Athens

Beijing

Brasilia

HongKong

London

Montreal

Moscow

Mumbai

Riyadh

Time (Hours)

Decrease of canyon air temperature,

green-all case

8

Fig. 7. Air canyon temperature decrease in the EW, H5W10 canyon, with

parallel wind flow, when both roofs and walls are covered with vegetation,

for all climates examined.



7/14

great amounts of summer insolation. For this reason, in the

instance of Hong Kong, canyon air temperature decrease

reaches 8.4 1C maximum and 6.9 1C daytime average in the

green-all case (Fig. 7), while for the green-wall case these

numbers become only 3.9 and 2.5 1C, respectively (Fig. 8).

In general, the temperature decrease is quite significant

for both surface and air temperatures both inside thecanyon and at roof level. Regarding surface temperature

decrease of the south-oriented wall, it reaches from 18.7 1C

maximum and 14.3 1C daytime average for Riyadh to 9.8

and 5.6 1C, respectively, for Moscow (Fig. 9). Roof surface

temperatures lower even more, due to the greatest amounts

of solar radiation horizontal surfaces receive in summer;

the greatest day-time average temperature decrease is noted

for Riyadh (12.8 1C) and the greatest maximum for

Mumbai (26.1 1C), while the smallest decreases are noted

for Moscow and London (Fig. 10). Moscow reaches the

smallest daytime average surface temperature decrease

(9.1 1C), while London the smallest maximum (19.3 1C). In

the subject of air temperature decrease inside the canyon

for the green-all case, it reaches its peak for Riyadh

(11.3 1C maximum and 9.1 1C daytime average), while its

smallest decreases are noted in Moscow (3.6 and 3.0 1C,

respectively) (Fig. 7). For the green-wall case, air tempera-ture decrease reaches its maximum again for Riyadh

(5.1 1C maximum and 3.4 1C daytime average) and its

lowest decreases for Moscow (2.6 and 1.7 1C, respectively)

(Fig. 8).

3.2. Indirect radiative cooling effects

On prima facie evidence, the air inside a canyon with

vegetated walls is reduced due to the evapotranspirational

rate from plants and the lower surface temperatures of

vegetated surfaces. The latter are responsible not only for

lowering the air temperature but also for lowering surface

temperatures of surfaces not covered with vegetation.

As the radiative heat exchanges between the urban

canyon surfaces have been modelled analytically, a

decrease is observed in the asphalt surface temperature

when walls are covered with vegetation. In Fig. 11 the

decrease of the asphalt surface temperature is presented for

the H5W10 canyon for all the climates examined. As can be

observed, the greatest decreases occur for hot and with

high solar radiation Riyadh, with a maximum decrease of

2.0 1C and a daytime average 1.3 1C. The lowest surface

asphalt temperature decreases take place in much colder

and with lower insolation Moscow (maximum 0.9 1C,

daytime average 0.6 1C). As the air temperature near the

ARTICLE IN PRESS

-6

-5

-4

-3

-2

-1

0

4 12 16 20 24Athens

Beijing

Brasilia

HongKong

London

Montreal

Moscow

Mumbai

Riyadh

Decrease of canyon air temperature,

green-wall case

TemperatureDecrease(C) 8

Time (Hours)

Fig. 8. Air canyon temperature decrease in the EW, H5W10 canyon, with

parallel wind flow, when only walls are covered with vegetation, for all

climates examined.

-10

-5

0

4 12 16 20 24

TemperatureDecrease(C) Athens

Beijing

Brasilia

HongKong

London

Montreal

Moscow

Mumbai

Riyadh

Decrease of surface temperature of

south-oriented wall

Time (Hours)

-15

-20

8

Fig. 9. Surface temperature decrease of the south-oriented wall, when

covered with vegetation, in the EW, H5W10 canyon, for all climates

examined.

-5

5

4 12 16 20 24

TemperatureDecrease(C) Athens

BeijingBrasiliaHongKongLondonMontreal

MoscowMumbaiRiyadh

Time (Hours)

Decrease of roof surface temperature

-35

-25

-15

8

Fig. 10. Roof surface temperature decrease when covered with vegetation,

for all climates examined.

4 12 16 20 24


Athens

Beijing

Brasilia

HongKong

London

Montreal

Moscow

Mumbai

Riyadh

Decrease of asphalt surface temperature

0.0

-0.5

-1.0

-1.5

-2.0

-2.5

Time (Hours)

8

Fig. 11. Asphalt temperature decrease when walls are vegetated in the

H5W10 canyon for all the nine climates examined.



8/14

ground is primarily affected by the horizontal surface

temperature [14,15,21], the effect of vertical surface

temperatures is not as crucial as that of the horizontal

surface. This radiative cooling of the street asphalt by green

walls, reaching up to 2.0 1C in hot Riyadh has an

additional effect on lowering air temperatures, apart from

evapotranspirational and convective cooling effects.

3.3. Climatic characteristics

It can be said with certainty that, the hotter and drier a

climate is, the more important the effect of green walls and

green roofs on mitigating urban temperatures is (Figs. 7

and 8). As can be observed in Figs. 3 and 4 and Table 3,

Riyadh is the hottest and most arid of all the cases

examined, with urban temperatures reaching 42.81C

maximum and 31.2 1C minimum, with a daytime average

of 38.3 1C, while relative humidity spans from 45.8% to

23.5%, with a daytime average of only 32.4%. These

extreme climatic inputs, benefit the most from green walls

and green roofs and, as has been mentioned in Section 3.1,

the green-all case reaches temperature decreases of the

magnitude of 11.31C maximum and 9.1 1C daytime

average and the green-walls case 5.1 and 3.41C, respec-

tively. Much more humid Mumbai (Fig. 4) reaches smaller

decreases, of the magnitude of 6.6 1C daytime average and

8.0 1C maximum for the green-all case and 2.7 and 4.4 1C,

respectively, for the green-walls case. The colder climates

of London, Moscow and Montre al benefit the least,

reaching daytime average decreases from 1.7 to 2.11C

and maxima from 2.6 to 3.2 1C for the green-walls case and

from 3.0 to 3.8 1C and from 3.6 to 4.5 1C, respectively, forthe green-all case.

3.4. Roof versus canyon

Temperatures at roof level decrease more than inside the

canyon, when the building envelope is covered with

vegetation. This is because the roof, being more exposed

to the much larger amounts of summer solar radiation on

the horizontal plane, raises its temperatures even more

when plain, low albedo-building materials are exposed to

direct solar gains. However, the canyon, due to its

geometry, is generally more shaded, not reaching the peak

temperatures roof surfaces do. By covering the roof with a

vegetated medium, which regulates its temperature so as

not exceed some crucial levels, roof temperatures decrease

more than the temperature inside the canyon, when both

roofs and walls are covered with vegetation. For all the

nine climates examined, the maximum temperature de-

crease at the air layer 1 m above the roof reaches from

26.0 1C for Riyadh to 15.51C for London and daytime

average temperatures from 12.8 1C for Riyadh to 5.8 1C for

Moscow (Fig. 12). The air inside the canyon reaches lower

decreases; for the green-all case the air temperature

decrease reaches a maximum from 11.3 1C for Riyadh to

3.6 1C for Moscow and daytime average from 9.1 1C for

Riyadh to 3.0 1C for Moscow (Fig. 7). However, tempera-

tures at roof level start falling after 12:00, while for the

more stable conditions inside the canyon, temperatures due

to vegetation on walls start decreasing from early in the

morning, as can be observed by comparing Figs. 7 and 12.

3.5. Canyon orientation

Canyon orientation determines the shading pattern on

both the horizontal and the vertical parts of the canyon

geometry. It determines the amount of insolation received,

especially for the vertical planes, depending on their

orientation. During the summer months examined, the

amount of irradiation received on vertical planes is much

smaller than the horizontal one, for all orientations. Thus,

the orientation, despite the fact that it plays an importantrole in temperature distributions in and around the canyon,

it does not affect temperature decreases so significantly

when vegetation covers its vertical surfaces and roofs. The

magnitude of the effect strongly depends on the geographic

latitude. The examples of Hong Kong (22.16N) and Athens

(37.59N) are discussed below.

For all the climates examined, it has been observed that

the amount and geometry of vegetation is more important

than the canyons orientation. In the instance of Hong

Kong, solar radiation in all vertical orientations is not so

high, reaching a maximum of only 185 W/m2 for the

south orientation and 427 W/m2 for the west orientation

(Fig. 13). The green-walls case of the EW and NS oriented

H5W10 canyons result in 2.4 and 2.0 1C daytime average

temperature decrease, respectively, inside the canyon, with

an only 0.4 1C difference between the two orientations. For

the maximum, this difference becomes 0.7 1C (3.81C

maximum temperature decrease for the EW oriented

canyon and 3.11C for NS one). For the green-all case,

the differences between the two orientations become even

smaller, reaching 0.2 1C for the daytime average (tempera-

ture decrease being 6.8 1C for EW and 6.6 1C for NS), and

0.0 1C for the maximum (maximum temperature decrease

being 8.5 1C for both EW and NS orientation). It can be

observed in Fig. 14 that the amount and geometry of

ARTICLE IN PRESS

-30

-25

-20

-15

-10

-5

0

5

4 12 16 20 24

Temperatu

reDecrease(C)

AthensBeijingBrasiliaHongKongLondonMontreal

MoscowMumbaiRiyadh

-35

8

Decrease of air temperature 1m above the roof

Time (Hours)

Fig. 12. Air temperature decrease 1 m above the roof, for all climates

examined.



9/14

vegetation mostly affect the temperature decrease and not

so much the canyons orientation. The green-all case in the

EW-oriented canyon results in 5.4 1C higher for the

daytime average and 4.7 1C higher for the maximum

temperature decrease than the green-walls case. For the

NS orientation, these differences become 4.6 and 4.4 1C,

respectively.

However, in the example of Athens, the amount of

irradiation received in the east and west oriented vertical

planes is much larger than on the south and north

orientations and proportionally larger to those of Hong

Kongs. The maximum solar radiation received by the

south-oriented vertical plane is 374.0 W/m2, while for the

east-oriented plane reaches the magnitude of 616.7 W/m2

(Fig. 15). This has a direct effect on the way the canyon

orientation affects temperature decreases due to vegetated

surfaces. For the green-all case, the difference between the

temperature decrease of the air inside the canyon remains

small, of the magnitude of 0.1 1C for the daytime average

(temperature decrease being 5.6 1C for EW and 5.5 1C for

NS), and 0.2 1C for the maximum (temperature decrease

being 6.6 1C for EW and 6.8 1C for NS orientation). For

the green-walls case these differences become larger,

reaching 0.81C for the daytime average (temperature

decrease being 3.01C for EW and 2.2 1C for NS) and

1.2 1C for the maximum (temperature decrease being 4.5 1C

for EW and 3.3 1C for NS). Yet again, the difference

between the two amounts of vegetation (green-all and

green-walls) is more crucial than the difference between the

decreases of different orientations. The difference between

the temperature decrease of the green-all and the green-

walls cases of the EW oriented canyon reaches 2.6 1C for

the daytime average and 2.1 1C for the maximum. For the

NS orientation, these differences become larger, 3.2 and

3.4 1C, respectively (Fig. 16).

In general, it can be concluded that the orientation may

play a countable role in temperature decreases due to

vegetation, only when the amounts of solar radiation

received by the vertical planes differ significantly. Yet

again, concerning temperature decreases, the amount of

vegetation placed on buildings is more crucial than the

orientation of the canyon, with the green-all case, when

ARTICLE IN PRESS

0

100

200

300

400

500

600

700

800


Global rad, horiz.

Globalrad, south

Global rad, north

Global rad, east

Global rad, west

1.00

Time (Hours)

4.00 7.00 10.00 13.00 16.00 19.00 22.00

Fig. 13. Hong Kong global solar radiation on horizontal and vertical

planes of east, west, south and north orientation in July.

0

5

10

15

20

25

30

35

12 15 18 21

Temperature(C)

-9

-8

-7

-6

-5

-4

-3

-2

-1

0 Ta,can[no gr]EW-HgKg

Ta,can [no gr]NS-HgKg

Ta,can[gr a]EW-HgKg

Ta,can[gr a]NS-HgKg

Ta,can[gr w]EW-HgKg

Ta,can[gr w]NS-HgKg

DTa,can[gr a]EW-HgKg

DTa,can[gr a]NS-HgKg

DTa,can[gr w]EW-HgKg

DTa,can[gr w]NS-HgKg

Time (Hours)


Fig. 14. Hong Kong temperature distributions and decreases inside the

canyon for no-green [no-gr], green-all [gr-a], green-walls [gr-w], for EW

and NS oriented H5W10 canyon.

0

100

200

300

400

500

600

700

800

900

1 10 13 16 19 22


Global rad, horiz.

Global rad, south

Global rad, north

Global rad, east

Global rad, west

4 7

Time (Hours)

Fig. 15. Athens global solar radiation on horizontal and vertical planes of

east, west, south and north orientation in July.

0

5

10

15

20

25

30

35

40

12 15 18 21

-8

-7

-6

-5

-4

-3

-2

-1

0 Ta,can[no gr]Ath-EW

Ta,can[no gr]Ath-NS

Ta,can[gr a]Ath-EW

Ta,can[gr a]Ath-NS

Ta,can[gr w]Ath-EW

Ta,can[gr w]Ath-NS

DTa,can[gr a]Ath-EW

DTa,can[gr a]Ath-NS

DTa,can[gr w]Ath-EW

DTa,can[gr w]Ath-NS

Temperature(C)


Time (Hours)

Fig. 16. Athens temperature distributions and decreases inside the canyon

for no-green [no-gr], green-all [gr-a], green-walls [gr-w], for EW- and NS-

oriented H5W10 canyon.



10/14

both roofs and walls are covered with vegetation, leading

to much larger temperature decreases.

3.6. Canyon geometry

For all geometries and cases examined, it can be

concluded that the wider a canyon is the smaller the effectof green roofs and green walls on its temperature decreases.

For wider canyons, temperatures inside the canyon are

dominated by the proportionally larger street surface and

the fact that it is more exposed to direct solar radiation. In

Riyadh, temperature decreases in the wide H5W15 canyon

are of the magnitude of 1.2 1C for the daytime average and

1.7 1C for the maximum of the green-wall case and 7.3 and

9.3 1C, respectively, for the green-all case. For the narrower

H10W5 canyon these decreases reach 6.3 1C daytime

average and 9.1 1C maximum for the green-walls case and

8.9 and 12.3 1C, respectively, for the green-all case (Figs. 17

and 18). It can be observed in Fig. 18 that the wider a

canyon is the smaller the effect of green walls is on its

temperatures. Nevertheless, when the combination of green

roofs and green walls is implemented (green-all case),

temperature decreases rise significantly. In contrast, for the

narrow H10W5 canyon, whose walls are proportionally

more dominant than the street, the green-walls case has a

significant effect on lowering urban temperatures. The

combination of both green roofs and green walls does not

lead to such significant further decreases, as is the case in

the wider H5W10 and H5W15 canyons. The differences

between the temperature decreases of the green-all and the

green-walls case reaches an average of 5.7 and 6.1 1C,

respectively, for the H5W10 and the H5W15 canyons,while for the H10W5 canyon it is only 2.6 1C. For the

maximum, the discrepancy between the two wider and the

narrower canyon is even larger, reaching 6.2, 7.6 and

3.2 1C, respectively.

3.7. Wind direction

Wind direction affects temperature decreases inside the

canyon even less than orientation. Although it is a

significant factor for temperature distributions, for the

decreases due to vegetation it is the vegetation itself that

plays the most important role. The differences betweentemperature decreases in the same canyons for different

wind directions are insignificant. In Fig. 19 temperature

distributions and temperature decreases for the EW-

oriented H5W10 canyon in Mumbai are presented, for

both parallel (y) and perpendicular (x) to the canyons

axis wind directions. It can be observed, that for the low air

velocities inside the canyon, temperature distributions are

not so different for the two wind directions as they were for

the canyon orientations (Figs. 14 and 16). Temperature

differences between the temperature decreases of the two

wind directions become quite insignificant, reaching a

0.1 1C difference for both green-walls and green-all cases. It

can also be observed that for both wind directions the most

important factor for temperature decreases is the amount

of vegetation, with the green-all case reaching temperature

decreases 3.7 1C higher than the green-walls case.

For the wider H5W15 canyon, with its much larger air

velocities, temperature decreases are similar for the two

wind directions (Fig. 20). The differences between the

temperature decreases of the two wind directions reach a

maximum of 0.3 1C, with 0.2 1C daytime average for the

ARTICLE IN PRESS

-5

0

7 11 15 19


DTa [gr a]H5W10

DTa [gr a]H10W5

DTa [gr a]H5W15

Air temperature decrease for different canyon

geometries in Riyadh for the green-all case

Time (Hours)

-15

-10

Fig. 17. Air temperature decrease during the day in the H5W10, H10W5

and H5W15 canyon, for the green-all case, Riyadh.

-8

-6

-4

-2

0

7 11 15 19


DTa[gr w]H5W10

DTa[gr w]H10W5

DTa[gr w]H5W15

Time (Hours)

-10

Air temperature decrease for different canyon

geometries in Riyadh for the green-wall case

Fig. 18. Air temperature decrease during the day in the H5W10, H10W5

and H5W15 canyon, for the green-wall case, Riyadh.

0

5

10

15

20

25

30

35

40

12 15 18 21

Temperature(C)

-9

-8

-7

-6

-5

-4

-3

-2

-1

0 Ta,can[no gr]Mumb-EW-x

Ta,can[no gr]-Mumb-EW-y

Ta,can[gr a]Mumb-EW-x

Ta,can[gr a]-Mumb-EW-y

Ta,can[gr w]Mumb-EW-x

Ta,can[gr w]-Mumb-EW-y

DTa,can[gr a]Mumb-EW-x

DTa,can[gr a]-Mumb-EW-y

DTa,can[gr w]Mumb-EW-x

DTa,can[gr w]-Mumb-EW-y

TemperatureDecrease(

C)

Time (Hours)

Fig. 19. Mumbai temperature distributions and decreases inside the

canyon for no-green [no-gr], green-all [gr-a], green-walls [gr-w], for a

parallel (y) and perpendicular (x) to the canyons axis wind direction in

the EW-oriented H5W10 canyon.



11/14

green-wall case and 0.4 and 0.3 1C, respectively, for the

green-all case. Again, the amount of vegetation has a

stronger effect than the wind flow direction, even in the

wider canyon.

It can thus be concluded, that for the generally low air

velocities inside the urban canyons [6], the effect of wind

direction is not so strong on temperature decreases due to

vegetated roofs and walls, as is the amount and geometry

of vegetation itself.

4. Thermal comfort

In order to assess the thermal comfort improvements in

outdoors spaces when walls and roofs are covered with

vegetation, the physiological equivalent temperature (PET)

is used, its expression deriving from Ref. [23] and its

relationship with thermal sense from Ref. [24]. The results

for the EW-oriented H5W10 green-all and no-green cases

are presented here, for Moscow, Athens and Riyadh in

Figs. 2123. Emphasis is given on thermal comfort, not

only inside the canyon (symbolised with EW in the graphs),

but also at the roof level (symbolised with rf).

It can be observed in Fig. 21, that for the much milder

summer of Moscow, the greening of the building envelope

does not lead to such major improvements of the outdoors

thermal comfort. PET ranging from slightly warm and

comfortable levels on the roof and inside the canyon for

the no-green case, lowers to cooler levels, from comfortable

to slightly cool, during daytime, when roofs and walls are

covered with vegetation (green-all case). Although moving

from slightly warm to comfortable might not be so

spectacular, it could prove to be beneficial for the thermal

comfort and well being of populations used to cooler

climatic conditions.

For much hotter Athens (Fig. 22) and Riyadh (Fig. 23),

the improvements of outdoors thermal comfort are more

dramatic. For both climates, the bare concrete roof reaches

the very hot level in the afternoon. When covered with

vegetation, the sensation warm is reached only for 4 h in

Athens and 5 in Riyadh. Most of the daytime, the exposed

to direct solar radiation roof reaches the slightly warm

and comfortable zone for both cities. For inside the

ARTICLE IN PRESS

0

510

15

20

25

30

35

40

45

12 15 18 21

-8

-7-6

-5

-4

-3

-2

-1

0

1 Ta,can[no gr]MumbH5W15-EW-x

Ta,can[no gr]-MumbH5W15-EW-y

Ta,can[gr a]MumbH5W15-EW-x

Ta,can[gr a]-MumbH5W15-EW-y

Ta,can[gr w]MumbH5W15-EW-x

Ta,can[gr w]-MumbH5W15-EW-y

DTa,can[gr a]MumbH5W15-EW-x

DTa,can[gr a]-MumbH5W15-EW-yTemperatureDecrease(C)

T

emperature(C)

Time (Hours)

Fig. 20. Mumbai temperature distributions and decreases inside the

canyon for no-green [no-gr], green-all [gr-a], green-walls [gr-w], for a

parallel (y) and perpendicular (x) to the canyons axis wind direction in

the EW-oriented H5W15 canyon.

Fig. 21. PET for the EW-oriented H5W10 canyon, for the no-green [no

gr] and green-all [gr-a] cases, inside the canyon (EW) and on the roof (rf),for Moscow.


gr] and green-all [gr-a] cases, inside the canyon (EW) and on the roof (rf),

for Athens.


gr] and green-all [gr-a] cases, inside the canyon (EW) and on the roof (rf),

for Riyadh.



12/14

canyon, the thermal sensation improves even more, from

hot, in both cases to slightly warm and comforta-

ble, reaching even slightly cool for both cases in the

early morning and late evening hours and even cool for

Athens, in the early morning hours.

In general, green roofs and walls can improve outdoors

thermal conditions not only at street level, but also at rooflevel, turning these empty urban spaces into potentially

usable ones, in the form of superterrestrial gardens. By

covering roofs and walls with vegetation, thermal comfort

in the built environment can improve significantly, not only

for hot climates, but for cooler ones, in which populations

are acclimatised to lower temperatures.

5. Energy savings from green walls and roofs

Apart from creating outdoor conditions, which are more

human-friendly, from a thermal point of view, green

roofs and green walls can also prove beneficial for indoorthermal conditions. In addition to the fact that they add a

further insulation layer to the buildings fabric, they can

decrease cooling load demands inside the building quite

significantly due to the microclimatic modifications dis-

cussed in this paper.

In a simplified steady-state analysis, without taking into

consideration internal thermal gains, heat gains/losses (qE)

from the buildings fabric with an average U-value U, an

indoors temperature Tin and an outdoors temperature Toutare given by the relationship:

qE UTout Tin. (9)

For the no-green base case [no gr], the cooling load for

the non-vegetated canyon is given by the relationship:

qEno gr U Tno gr Tin

, (10)

where T[no gr] is the averaged air temperature inside the

canyon when no vegetation is placed either on walls or on

roofs. For the green-all case, with an average air

temperature inside the canyon T[gr a], heat gains are1:

qEgr a U Tgr a Tin

. (11)

Thus, the decrease in the cooling load, when both walls

and roofs are covered with green is given by

DqEgr a qno gr qgr a

qno gr) ,

DqEgr a Tno gr Tgr a

Tno gr Tin. 12

For T[no gr]6Tin, T[no gr]4Tin and T[gr a]4Tin.

Similarly, for the green-walls case, if T[gr w] is the

average air temperature inside the canyon with the green

walls, the cooling load decrease becomes:

DqEgr w Tno gr Tgr w

Tno gr Tin. (13)

For T[no gr]6Tin, T[no gr]4Tin and T[gr w]4Tin.

Considering an indoor limit temperature for cooling of

23 1C for all climates studied, the cooling load decreases

due to green-all and green-walls cases are given as a

daytime average in Fig. 24 and for an hourly basis in

Fig. 25.

As can be observed in Fig. 24, the largest cooling load

decreases in all climates examined, occur for the green-all

case. For the geometries examined for Braslia and Hong

Kong, the cooling load decreases for the green-all casereach 100%; no cooling load is needed after covering roofs

and walls with vegetation, while in both cities cooling load

is needed in the afternoon and early evening hours for the

no-green case (Fig. 25c and d). London and Moscow are

not affected, regarding cooling loads, as no cooling load is

needed for the typical day examined, even before vegeta-

tion was placed around the canyon. Riyadh experiences a

quite high cooling load decrease, of the magnitude of 90%,

as does Montre al (85%) for the green-all case, lowering

their total hours of cooling demand from 12 to 5 and from

8 to 4, respectively (Fig. 25e and g). For the green-all case,

Mumbai reaches a 72% decrease, lowering its cooling

energy demand from 11 h to 6 (Fig. 25f), while for Athens

and Beijing the decrease is 66% and 64%, respectively,

lowering their energy demand by 4 and 3 h, respectively

(Fig. 25a and b).

For the green-walls cases, cooling load decreases are less

dramatic. The largest one is noted for Braslia (68%), with

6 h decrease in cooling demand (Fig. 25d). It is followed by

a 66% and 2 h decrease for Hong Kong (Fig. 25e), 52%

and 2 h for Montre al (Fig. 25e), 43% and 2 h for Athens

(Fig. 25a), 37% and 2 h for Beijing (Fig. 25b), 37% and 3 h

for Riyadh (Fig. 25(g) and 35% and 3 h for Mumbai

(Fig. 25f). It can be noted that the differences between the

green-all and green-walls cooling loads are smaller for

ARTICLE IN PRESS

Fig. 24. Average cooling load decreases (%), with a 231C indoors

temperature, for the green-all and green-walls cases of all the climates

examined.

1Despite the fact that the U-value is altered, when vegetation is placed

on the buildings fabric, leading to further cooling load decreases, this is

not taken into consideration here, as the aim is to directly compare

between the effects of the microclimatic alterations on the buildings

cooling load, without it being affected by alterations to the fabric.



13/14

humid climates (of the magnitude of 3237%) and greater

for arid climates (53% for Riyadh), due to the different

humidity concentrations in the two climatic groups.

In general, green roofs and green walls cool the

microclimate around them, which can lead to quite

important energy savings for cooling, depending on the

climatic type, the amount and position of vegetation on the

building. In cases where little cooling load is needed,

cooling demand can be reduced to zero by covering

building surfaces with vegetation. In other cases, energy

savings can also be significant, varying from 90% to 35%.2

In addition to the energy savings themselves, this could

lead to successful applications of further passive cooling

techniques, especially ones employing ventilation, which

are not easy to implement in the extremely hot urban

conditions, in cases of large heat island densities.

6. Conclusions

From this quantitative research, it has been shown that

there is an important potential of lowering urban

temperatures when the building envelope is covered with

vegetation. Air temperature decreases at roof level can

reach up to 26.0 1C maximum and 12.81C day-time

average (Riyadh), while inside the canyon decreases reach

up to 11.3 1C maximum and 9.1 1C daytime average, again

for hot and arid Riyadh. It can be concluded that the

ARTICLE IN PRESS

Fig. 25. Cooling load decreases (%) for (a) Athens, (b) Beijing, (c) Hong Kong, (d) Braslia, (e) Montre al, (f) Mumbai and (g) Riyadh for green-all and

green-walls cases.

2These percentages can become even greater, when a higher than 23 1C

limit temperature for cooling is considered. In general, inhabitants of hot

climates are accustomed to higher temperatures (in the instance of Greek

regulations, the limit temperature for cooling is set to 26 1C).



14/14

hotter and drier a climate is, the greater the effect of

vegetation on urban temperatures. However, it has been

pointed out that also humid climates can benefit from

green surfaces, especially when both walls and roofs are

covered with vegetation, reaching up to 8.4 1C maximum

temperature decrease for humid Hong Kong. Temperature

decrease due to vegetation is primarily affected by thevegetation itself (amount and geometry), more than the

canyon orientation in hot periods. In general, the larger

amounts of solar radiation a surface receives, the larger its

temperature decreases are when it is covered with vegeta-

tion. For the low air velocities inside the canyon, the wind

direction does not have any significant effect on tempera-

ture decreases due to vegetation.

Regarding the urban geometry, the wider a canyon is,

the weaker the effect green roofs and green walls have on

temperature decrease. For all climates examined, green

walls have a stronger effect than green roofs inside the

canyon. Nonetheless, green roofs have a greater effect at

roof level and, consequently, at the urban scale. The

combination of both green roofs and green walls leads to

the highest mitigation of temperatures inside the canyon. If

applied to only one unit block, green roofs and green walls

can create a small area of mitigated temperatures to the

urban heat island effect, as has been shown in this

microclimatic study. If applied to the whole city scale,

they could mitigate raised urban temperatures, and,

especially for hot climates, bring temperatures down to

more human-friendly levels and achieve energy saving

for cooling buildings from 32% to 100%.

Acknowledgements

This research has been funded by the State Scholarship

Foundation of Greece (IKY) from 2001 to 2003. The

authors are extremely grateful to Panagiotis Doussis for his

guidance and contribution to computer modelling.

References

[1] Benevolo L. The history of the city. London: Scholar Press; 1980.

[2] Koppe C, Kovats S, Jendritzky G, Menne B. Health and global

environmental change; heat-waves: risks and responses, series no. 2,

energy, environment and sustainable development. Copenhagen:

World Health Organization; 2004.[3] White KS, et al. Technical summary, climate change 2001: impacts,

adaptation and vulnerability. In: McCarthy JJ, Canziani OF, Leary

NA, Dokken DJ, White KS, editors. Climate change 2001: impacts,

adaptation and vulnerability. Cambridge: United Nations Inter-

governmental Panel on Climate Change (IPCC) Cambridge Uni-

versity Press; 2001. p. 1973.

[4] Daz J, Ballester F, Lopez-Velez R. Impacts on human health. In:

Moreno Rodrguez JM, editor. The preliminary assessment of the

impacts in Spain due to effects of climate change, project ECCE.

Madrid: Ministry of the Environment; 2005.

[5] Davies RE, Knappenberger PC, Novicoff WM, Michaels PJ. Decadalchanges in summer mortality in US cities. International Journal of

Biometeorology 2003;47:16675.

[6] Santamouris M, editor. Energy and climate in the urban built

environment. London: James & James; 2001.

[7] Dimoudi A, Nikolopoulou M. Vegetation in the urban environment:

microclimatic analysis and benefits, PRECis project, the European

Commission, Directorate General XII, Joule III, contract JOR3-

CT97-0192. Centre for Renewable Energy Sources, Pikermion,

2000.

[8] Bruse M, Fleer H. Simulating surfaceplantair interactions inside

urban environments with a three dimensional numerical model.

Environmental Modelling and Software 1998;13:37384.

[9] Giridharam R, Ganasan S, Lau SSY. Daytime urban heat island in

high-rise and high-density residential developments in Hong Kong.

Energy and Buildings 2004;36:52534.[10] Versteeg HK, Malalasekera W. An introduction to computational

fluid dynamics. Harlow, London: Prentice-Hall; 1995.

[11] Luikov AV. Heat and mass transfer in capillary-porous bodies.

Oxford: Pergamon Press; 1961.

[12] Eckert ERG, Drake RM. Heat and mass transfer. New York:

McGraw-Hill Book Company Inc.; 1959.

[13] Noilhan J. Les facteurs physiques du microclimat au voisinage dun

batiment. Etude bibliographique. Centre Scientifique et Tecnique du

Batiment, Etablissement de Nantes, 1979.

[14] Jacobson MZ. Fundamentals of atmospheric modeling. Cambridge:

Cambridge University Press; 1999.

[15] Pielke RA. Mesoscale meteorological modeling. San Diego: Aca-

demic Press; 2002.

[16] Modest MF. Radiative heat transfer. New York: McGraw-Hill

International Editions; 1993.[17] Remund J, Lang R, Kunz S, Meteonorm, version 3.0, Meteotest,

Bern, 1997.

[18] Marsh A. Ecotect, version 5.20(b), square one research PTY LTD,

2003.

[19] Welsh School of Architecture, Cardiff University, WinAir4. Cardiff,

2003.

[20] Critchfield HJ. General climatology. London: Prentice-Hall Inter-

natioal Inc.; 1974.

[21] Geiger R. The climate near the ground. Cambridge, Massachusetts:

Harvard University Press; 1965.

[22] Monteith JL, Unsworth M. Principles of environmental physics.

London, New York: Edward Arnold; 1990.

[23] Matzarakis A, Mayer H, Rutz F. Radiation and thermal comfort.

Sixth Hellenic conference in meteorology, climatology and atmo-

spheric physics, Ioannina 2; 2002. p. 73944.[24] Kuttler W. Stadklima, Teil 2: pha nomene und wirkungen. In: Mo ller

D. (Ed.), Beitragsserie klimaa nderung und klimaschutz, UWSF,

vol. 16(4); 2004. p. 26374.

ARTICLE IN PRESS


EFECTOS NATURACIÓN

Documents

Transcript of EFECTOS NATURACIÓN