Diss Freidok Ali Toudert Fazia

Berichte des Meteorologischen Institutes
der Universität Freiburg

Nr. 15
Fazia Ali Toudert
Dependence of Outdoor Thermal Comfort

on Street Design
in Hot and Dry Climate
Freiburg, November 2005

Dekan: Prof. Dr. Enrst E. Hildebrand
Referent: Prof. Dr. Helmut Mayer
Korreferent: Prof. Dr. Gerd Jendritzky
Mündliche Prüfung: 23.11.2005
ISSN 1435-618X
Alle Rechte, insbesondere die Rechte der Vervielfältigung und Verbreitung so-
wie der Übersetzung vorbehalten.
Eigenverlag des Meteorologischen Instituts der Albert-Ludwigs-Universität Frei-

burg
Druck: Druckerei der Albert-Ludwigs-Universität Freiburg
Herausgeber: Prof. Dr. Helmut Mayer und PD Dr. Andreas Matzarakis

Meteorologisches Institut der Universität Freiburg
Werderring 10, D-79085 Freiburg
Tel.: 0049/761/203-3590; Fax: 0049/761/203-3586
e-mail: meteo@meteo.uni-freiburg.de
Dokumentation: Ber. Meteor. Inst. Univ. Freiburg Nr. 15, 2005, 224 S.
Dissertation, angenommen von der Fakultät für Forst- und Umweltwissenschaf-

ten der Albert-Ludwigs-Universität Freiburg
2
in memory of Nabila
and
to my parents
3
4
Acknowledgements
First, I am greatly indebted to Prof. Dr. Helmut Mayer, head of the Meteorological In-
stitute MIF, University of Freiburg, for supervising this work, for his support through-
out the period of my stay in MIF and for giving me the valuable possibility to join sev-
eral conferences.
I also wish to express my deep gratitude to Dr. Michael Bruse whose advice and support
while working with his model facilitated enormously my work and kept me going on
during this study.
I would also like to thank Prof Dr. Gerd Jendritzky (German Weather Service, DWD)
and Prof Dr. Wilhelm Kuttler (University of Duisburg/Essen) for their interest in my
study.
The invitation of Prof. Dr. Rafik Bensalem (School of Architecture of Algiers) to par-
ticipate in the experimental work conducted in south Algeria is gratefully acknowl-
edged. Thanks go also to Moussadek Djenane, Omar Douag and the association for the
protection of the environment of the city of Beni-Isguen, Algeria, for their assistance
during the measuring campaign.
I highly appreciated the permanent disposal for assistance of Dr. Florian Imbery and the
friendship and support of Pamela May and Carolin Vassigh. Thanks also to all other
MIF members who helped me in one way or another. Further thanks are addressed to
Nicky and Peter Lafferty, and Dr. Argwings Ranyimbo for proofing the final manu-
script.
The present work would not be possible without the financial support of the German
academic exchange service DAAD. Through the DAAD, I had the great opportunity for
an enriching experience which has enlarged my horizon.
Not least, I would especially like to thank my parents, close relations and friends for
their care and encouragements. Thanks for always being there when I need you.
I hope that you find it worthwhile and helpful. Enjoy the read!
Freiburg, July 2005
5
6
TABLE OF CONTENTS Page
Acknowledgments ...................………………………………………………...…… 5
Table of contents …...………………………………………………………….…… 7
Summary …………....…...………………………………………………………..… 11
Zusammenfassung ….………………………………………………………….…… 13
1. Introduction ...………………………………………………………………….. 17
1.1. Necessity of the present study ...…...………………………...………….... 17
1.2. Objectives of the present study ….……………………………………...… 21
1.3. Methodology …….……………………………………………...……….... 22
1.4. Structure of the thesis …………………………………………...………... 23
2. Literature review .……………………….……………………………………... 25

2.1. The microclimate of an urban street canyon ...…………..……………...… 26
2.1.1. Energy budget of an urban canyon ……………...………...........… 26
2.1.2. Thermal characteristics of an urban canyon …….………..………. 28
2.1.3. Wind flow in an urban canyon …..……………………………...… 31
2.1.4. Solar access outdoors ……………………………………...……… 33
2.1.5. Solar access indoors ……..……………………………...………… 36
2.1.6. Effects of the vegetation ...……………………………...………… 37
2.1.7. Further aspects …..……………………………………………...… 40
2.2. Outdoor thermal comfort ..………………………………………………... 41
2.2.1. The thermal comfort indices …….…….………………….………. 41
2.2.2. The human energy balance ……...……….……………………….. 43
2.2.3. The mean radiant temperature …...…………………………...…... 46
2.2.4. Methodological problems in assessing comfort outdoors …....…… 48
2.2.5. Effects of urban design on comfort outdoors ……………...……… 51
2.2.6. Conclusion …....…………………………………………………... 55
3. The numerical model ENVI-met 3.0 ……………………….…………………. 57

3.1. Numerical modelling of the urban microclimate ..…………………...…… 57
3.2. Relevance of ENVI-met to the present study ……………………......…… 58
3.3. General structure of ENVI-met 3.0 ……...………………………...……… 59
7
Table of contents
3.4. The atmospheric model ….………………………..………………………. 61

3.4.1. Mean air flow ……………………………………………..………. 61
3.4.2. Temperature and humidity .………...…………...………………… 62
3.4.3. Atmospheric turbulence ………………………………...………… 62
3.4.4. Radiation fluxes …………………………………………………... 64
3.4.5. The ground and building surfaces ……….……………………..…. 67
3.5. The soil model ………………………………………………………….… 69
3.6. The vegetation model ……………………………………………………... 70
3.7. The human-biometeorological dimension …………………………..…..... 71
3.8. Boundary conditions and course of a simulation …………………………. 72
3.9. Simulations with ENVI-met in the present work ..………………...……... 74
3.9.1. Site climate ……………………………………………...………... 74
3.9.2. Simulation conditions …...………………………………………... 75
3.9.3. Case studies ……..………………………………………………... 76
4. Results of the numerical simulations ………………………………………… 79

4.1. Symmetrical canyons oriented east-west ..……………………………..…. 79
4.1.1. Air temperature ….…………………………………………..……. 79
4.1.2. Radiation fluxes ……………………………………………..…..... 80
4.1.3. Thermal comfort analysis .……………………………..…………. 82
4.2. Symmetrical canyons oriented north-south ……...…………...………..…. 89
4.2.1. Air temperature ….……………………………………………..…. 89
4.2.2. Radiation fluxes ……………………………………………..……. 90
4.2.3. Thermal comfort analysis .………………………………………... 93
4.3. Comparison between E-W and N-S streets ……...…………………….…. 96
4.4. Intermediate orientations NE-SW and NW-SE ……………….………….. 103
4.5. Complex urban canyons …………………………………………………... 105
4.5.1. Air temperature ….………………………………………………... 105
4.5.2. Role of galleries …………………………………………………... 108
4.5.3. Role of the asymmetry and overhanging façades ………………… 113
4.5.4. Role of the vegetation …...………………………………………... 121
4.6. Role of the wind …...……………………………………………………… 125
4.7. Solar access in summer and winter …………..….………………………... 130
8
Table of contents
5. Field measurements in Freiburg, Germany …….........………………………. 137

5.1. Site and observations ……………………………………………………... 137
5.2. The microclimate in the canyon….…………………………….…………. 140
5.2.1. Air and surface temperatures ……………………………………... 140
5.2.2. Wind direction and wind speed ….……………………...………... 143
5.3. Thermal comfort analysis ..…………………………………...…………... 146
5.3.1. Short-wave radiation fluxes ……..………………………………... 146
5.3.2. Long-wave radiation fluxes ……..………………………………... 147
5.3.3. Heat gained by a standing person .………………………………... 149
5.3.4. Human thermal comfort ....………………………………………... 152
5.4. Comparison with ENVI-met simulation ...………………………………... 154
5.5. Discussion and conclusion ………………………………………………... 156
6. Field measurements in Beni-Isguen, Algeria ………….……………......….… 157

6.1. Site description …...…………………………………………...………….. 157
6.2. Measurements …………………………………………………………….. 162
6.3. The microclimate in the canyon …...………………………..……………. 162
6.3.1. Air temperature and air humidity …..………………...…..………. 162
6.3.2. Wind speed …...……………………………………...………….... 164
6.3.3. Surface temperatures …...…………………………………………. 165
6.3.4. Radiation fluxes …………………………………………………... 166
6.3.5. Mean radiant temperature ……………………………………….... 168
6.4. Thermal comfort analysis …….……………………..…...……………….. 170
6.5. Discussion and conclusion …...…………………………………………… 171
7. Discussion and conclusion …………………………………………..………… 175

7.1. Street microclimate ..………………………………………………..……… 175
7.2. Heat gained by a human body …….……………..…………...……………. 179
7.3. Street design and outdoor thermal comfort ………..………...………..…… 180
7.3.1. Aspect ratio and solar orientation .…………………..……………. 180
7.3.2. Asymmetry, galleries and overhanging façades .…….….…...…… 182
7.3.3. Vegetation ..…………………………………………..…………… 183
7.4. Recommendations and design examples…………………………………… 184
9
Table of contents
7.5. Limits and current development of ENVI-met …...………………………... 189

7.5.1. Boundary conditions …….……………………………………..…. 189
7.5.2. Heat storage in the building materials …….……..………...…..…. 190
7.5.3. Mean radiant temperature and comfort ….…………………..……. 191
7.6. Concluding remarks .……………………………………………..………… 192
References …..……………………………………………………………………… 194

List of figure captions ……...……………………….………………………………. 207
List of tables captions ...…………………..………………………………………… 213
List of symbols and abbreviations ...………………………………………………... 214
Appendix …...……………………………………………………………………….. 219
10
Summary
Summary
The present work addresses the contribution of street design toward the development of
a comfortable microclimate at street level for pedestrians. The work is design-oriented
and seeks to provide a quantitative knowledge readily interpretable from the perspective
of urban designers. Street geometries are investigated, including various aspect ratios,
i.e. height-to-width ratio H/W, solar orientations and a number of design details. First,
symmetrical urban canyons with H/W equal to 0.5, 1, 2 and 4 and for different solar
orientations (i.e. E-W, N-S, NE-SW and NW-SE) are studied. Secondly, asymmetrical
profiles with different openness to the sky are investigated together with the role of ar-
chitectural details such as galleries, horizontal overhangs on façades and rows of trees,
considered as possible ways to improve the outdoor thermal comfort further in the
summertime. Moreover, the analysis focuses on the local differences in the thermal sen-
sation across the street, i.e. street centre vs. street sides, which influence the frequenta-
tion of the street. A special emphasis is placed on a human bio-meteorological assess-
ment of these microclimates by using the thermal index PET, Physiologically Equiva-
lent Temperature.
The investigation is carried out by using the three-dimensional numerical model ENVI-
met 3.0, which simulates the microclimatic changes within urban environments in a
high spatial and temporal resolution. Model calculations are run for typical summer
conditions in Ghardaia, Algeria (32.40° N, 3.80° E), a subtropical region characterized
by a hot and dry climate. Additionally, short-term field measurements are carried out in
Freiburg, Germany, and in Ghardaia (Beni-Isguen), Algeria, during the summer 2003.
In the former site, the microclimate changes due to geometry and the effects of the street
irradiation patterns on the heat gained by a human body are dealt with in detail. In the
latter site, a quantitative evaluation of the thermal effectiveness of existing architectures
in a hot-dry climate is the focus.
The simulations show that the thermal comfort is difficult to reach passively in such an
extreme climate but improvements are possible by means of appropriate geometrical
forms. All investigated urban describers are found to influence the final thermal sensa-
11
Summary
tion. Contrasting patterns in the comfort situation are found between shallow and deep
urban streets as well as between the various orientations studied.
Wide streets (H/W ≤ 0.5) are highly uncomfortable for both orientations. Yet, N-S ori-
entation shows some advantage over E-W orientation, and this benefit increases as the
aspect ratio increases.
Explicitly, this is expressed by a shorter period of heat stress and lower PET maxima.
Moreover, heat stress can effectively be mitigated if galleries, trees or textured façades
are appropriately combined with the aspect ratio and solar orientation.
A comparison of all case studies reveals that the duration, the period of day of extreme
heat stress, as well as the spatial distribution of PET across the canyon depend strongly
on aspect ratio and on street orientation. This is crucial since this will directly influence
the design choices in relation to street usage, e.g. streets exclusively planned for pedes-
trian use or including motor traffic, and also the time of frequentation of urban spaces.
The simulations as well as the on-site measurements also confirmed the dominant role
of the radiation fluxes expressed by the mean radiant temperature Tmrt for summer con-
ditions. The human body absorbs energy from the irradiated surrounding surfaces and
from a direct exposure of his body. This fact points out the necessity of shading as a
main strategy for keeping the street area in comfort range. Air temperature and wind
speed are secondary factors with respect to comfort as these vary less with urban ge-
ometry changes in comparison to Tmrt. The issue of solar access indoors has been briefly
discussed as an additional criterion in designing the street by including winter needs and
draw attention on the double role of the street, i.e. as interface of urban and architectural
scales. Design recommendations are also outlined for designing a comfortable urban
street.
Methodologically, ENVI-met revealed to be a good tool for the prognosis of the urban
microclimate changes within urban areas, and also in the assessment of outdoor comfort
through a satisfactory estimation of the mean radiant temperature. A number of eventual
refinements of the model are mentioned to improve its accuracy.
The work also highlights the necessity of more on-site measurements and more subjec-
tive votes of people for validating the simulations results and in order to strengthen a
practice-oriented knowledge about comfort in urban areas.
12
Zusammenfassung
Zusammenfassung
Abhängigkeit des thermischen Komforts unter heißen und trockenen Klimabedin-

gungen vom Straßendesign
Die vorliegende Untersuchung beschäftigt sich mit dem Beitrag, den das Straßendesign
zur Ausbildung eines komfortablen Mikroklimas für Menschen im Straßenbereich leis-
ten kann. Der Schwerpunkt liegt bei subtropischen Klimabedingungen, d.h. bei heißem
und trockenem Klima. Die Arbeit wurde anwendungsorientiert durchgeführt. Sie ver-
sucht, quantitative Ergebnisse bereitzustellen, die aus der Sicht eines Stadtplaners leicht
verstanden und interpretiert werden können. Verschiedene Geometrien von Straßen-
schluchten werden untersucht. Sie schließen ein variables H/W-Verhältnis (H: Höhe der
Randbebauung, W: Straßenbreite), unterschiedliche Richtungen und spezifische De-
signausführungen (z.B. Überhänge oder Straßenbegleitgrün in Form von Bäumen) ein.
Am Anfang werden symmetrische Straßenschluchten mit H/W = 0.5, 1, 2 und 4 analy-
siert, die in E-W, N-S, NE-SW und NW-SE verlaufen. Daran schließt sich die Untersu-
chung von asymmetrischen Straßenschluchten an, die eine größere bzw. kleinere Öff-
nung zum Himmel aufweisen. Hier werden zusätzlich die Effekte von planerischen De-
tails, wie Galerien, Überhänge und Straßenbegleitgrün, berücksichtigt, da sie eine weite-
re Möglichkeit zur Verbesserung der thermischen Komfortbedingungen im Außenbe-
reich darstellen. Bei den Ergebnissen erfolgt eine räumliche Differenzierung im thermi-
schen Empfinden von Menschen quer zur Straße. Neben der Straßenmitte werden daher
auch die Straßenränder betrachtet, da sie vorwiegend von Menschen in der Stadt fre-
quentiert werden.
Zur human-biometeorologischen Bewertung der thermischen Komponente des Klimas
in Straßenschluchten wird von den modernen, thermophysiologisch relevanten Indizes
die physiologisch äquivalente Temperatur PET verwendet. Die zur Berechnung von
PET erforderlichen Variablen werden in hoher räumlicher und zeitlicher Auflösung über
die Anwendung des dreidimensionalen numerischen Modells ENVI-met, Version 3.0,
13
Zusammenfassung
ermittelt. Die Grundzüge von ENVI-met werden so weit beschrieben, so wie es für das
Verständnis der Simulationsergebnisse notwendig ist.
Als Untersuchungsgebiet, das im Sommer heiße und trockene Klimabedingungen reprä-
sentiert, dient die Region Ghardaia (32.40 °N, 3.80 °E) in Algerien. Die meisten Simu-
lationsberechnungen werden für den 1. August durchgeführt.
In Ergänzung zu den numerischen Simulationsuntersuchungen erfolgten im Sommer
2003 kurzzeitige Messungen der thermophysiologisch relevanten meteorologischen
Variablen in einer Straßenschlucht in Freiburg und auf einer Profilroute durch die Stadt
Beni-Isguen in der Region Ghardaia, Algerien. Bei der Messkampagne in Freiburg
(Deutschland) standen das Mikroklima und die räumlich/zeitliche Variabilität der kurz-
und langwelligen Strahlungsflüsse vor dem Hintergrund der Wärme, die dadurch eine
stehende Person absorbiert, im Mittelpunkt. Zusätzlich ließen sich über die durchge-
führten Messungen Simulationsresultate aus ENVI-met validieren. Die Messkampagne
in Beni-Isguen ermöglichte in Bezug auf die thermischen Komfortbedingungen eine
Bewertung der Freiräume in der bestehenden Anordnung von Gebäuden und Straßen-
schluchten.
Die Ergebnisse aus den Simulationsberechnungen zeigen, dass thermischer Komfort für
Menschen unter den gegebenen, extremen klimatischen Verhältnissen schwer erreichbar
ist. Verbesserungen sind jedoch über geeignete geometrische Anordnungen von Stra-
ßenschluchten möglich. Gegensätzliche Muster der Komfortbedingungen treten zwi-
schen Straßenschluchten mit sehr hohem und niedrigem H/W-Verhältnis sowie mit un-
terschiedlicher Orientierung auf. Breite Straßenschluchten (H/W ≤ 0.5) weisen im Som-
mer einen hohen thermischen Diskomfort auf. Dabei sind Straßenschluchten in N-S
Richtung etwas günstiger als solche in E-W Richtung zu beurteilen. Angezeigt durch
einen kürzeren Zeitraum während des Tages mit Wärmebelastung und niedrigere PET
Maxima nimmt dieser Vorteil der Straßenschluchten in N-S Richtung mit ansteigendem
H/W-Verhältnis zu. Es ließ sich auch quantifizieren, wie sich die Wärmebelastung, die
in Straßenschluchten auf Menschen wirkt, durch die Berücksichtigung von weiteren
Varianten des Straßendesigns, wie Galerien, Überhänge oder Straßenbegleitgrün, in
Kombination mit dem H/W-Verhältnis und der Richtung der Straßenschlucht reduzieren
lässt.
Auf der Grundlage der Ergebnisse aus allen Fallstudien zeigt sich, dass im subtropi-
schen Klima der Tageszeitraum mit extremer Wärmebelastung und die räumliche Ver-
teilung von PET in der Straßenschlucht am stärksten vom H/W-Verhältnis und von der
14
Zusammenfassung
Richtung der Straßenschlucht abhängen. Das hat Auswirkungen auf die Möglichkeiten
zur optimierten Nutzung von Straßenschluchten (z.B. ausschließliche Nutzung durch
Fußgänger oder einschließlich von Kraftfahrzeugverkehr) und auf die Zeiten der häufi-
gen Nutzung dieser urbanen Freiräume.
In Bezug auf den thermischen Komfort von Menschen im Sommer im Freien bestätigen
die Simulationsberechnungen und experimentellen Fallstudien die dominierende Rolle
der
Strahlungsflüsse, die durch die mittlere Strahlungstemperatur Tmrt der Umgebung para-
metrisiert werden. Stehende Menschen in Straßenschluchten absorbieren tagsüber
hauptsächlich Strahlungswärme von bestrahlten Umgebungsflächen, während der Wär-
megewinn aus der direkten Sonnenstrahlung von zweitrangiger Bedeutung ist. Daraus
ergibt sich, dass nur über die Abschattung der direkten Sonnenstrahlung auf Umge-
bungsflächen die klimatischen Bedingungen in Straßenschluchten einen Zustand errei-
chen, der im Sommer von Menschen als weniger thermisch belastend empfunden wird.
Lufttemperatur und Windgeschwindigkeit können in Bezug auf thermischen Komfort
im Sommer in den Subtropen als meteorologische Variable von zweiter Bedeutung auf-
gefasst werden, da ihre räumliche und zeitliche Variabilität in Straßenschluchten deut-
lich geringer als diejenige von Tmrt ist.
Ein zusätzliches Kriterium für das Straßendesign, auf das kurz eingegangen wird, ist die
Verfügbarkeit von Strahlung in Innenräumen im Winter. Unter Berücksichtigung der
Anforderungen durch Menschen im Sommer und im Winter an das Design von Straßen-
schluchten werden Empfehlungen an die Planung von Freiräumen in Straßenschluchten
gegeben.
Insgesamt hat sich das mikroskalige Modell ENVI-met als ein sehr gutes Werkzeug für
diese Untersuchung herausgestellt. Vor allem wird Tmrt in zufriedenstellender Weise
simuliert. Zweckmäßig wären allerdings weitere Validierungen von ENVI-met über
geeignete experimentelle Fallstudien. Resultate zu thermischen Indizes sollten über Be-
fragungen von Menschen über ihre Einschätzung der thermischen Bedingungen ergänzt
werden, weil dadurch die Möglichkeit besteht, thermische Indizes in abgestufter Form
zu klassifizieren.
15
16
1. Introduction
1.1. Necessity of the present study
The urban climate is a shared field to climatologists and designers. Each of them, however,
has dealt for a long time with this issue differently in a number of ways, including scale,
relevant variables and object of study (Mills 1999). Hence, and as noted by several authors,
the integration of the climate dimension in the design process is lacking as a consequence
of poor interdisciplinary work. Therefore, they increasingly emphasize the necessity of
translating the available knowledge on applicable design guidelines to overcome this defi-
cit (e.g. Bitan 1988, Oke 1988, Arnfield 1990a, Kuttler 1993, Mayer 1993, Golany 1996,
Mills 1999).
In fact, climatologists were more concerned with the causality of the urban climate, while
designers were more interested in the effects of environmental forces on buildings. The
urban climatology concentrated first on the urban heat island (UHI) and moved progres-
sively to micro-scales as the urban geometry was found to be decisive in the UHI, (e.g.
Barry and Chorley 1978, Landsberg 1981, Oke 1987, Escourrou 1991, Oke et al. 1991,
Kuttler 2004). The focus was then placed on understanding the surface-air energy ex-
changes and mass exchanges between the urban canopy and the overlaying boundary layer
(Mills 1997).
By contrast, designers focused initially on indoor climate of individual buildings, on design
strategies, and on the resulting energy needs for maintaining internal comfort. The interest
for these issues was exacerbated by the oil crisis of 1973, as thoroughly documented (e.g.
Olgay 1969, Givoni 1976, Markus and Morris 1980). A special attention was dedicated to
passive solar gains as a way to enhance the environmental efficiency of buildings. Next,
they attempted to apply that environmental approach to a larger scale, i.e. urban environ-
ments, and the challenge consisted mostly in managing the reduced potential of sun and
17
1. Introduction
wind energy due to the mutual obstructions between buildings and implied by the high
urban density. The concept of solar envelope, initiated by Knowles (1981), and which
manages the solar availability inside the buildings illustrates well this evolution.
Recently, one can notice that the environmental quality of urban open spaces has become a
central issue for both disciplines. This can be observed in the latest related scientific meet-
ings (e.g. PLEA, ICUC and AMS conferences) as well as in the practice-oriented literature
(e.g. Herzog 1996, Rogers 1997, Asimakopoulos et al. 2001, Littlefair et al. 2001, Hawkes
and Foster 2002, Steemers 2003, Thomas 2003). The topic of comfort in outdoor spaces
will certainly foster more collaboration between both fields, so that the disconnection ob-
served so far between the sophisticated but theoretical results of the urban climatology on
one hand, and the more empirical but design-oriented findings of urban design on the other
hand, can be overcome.
In this respect, the street appears as the interface of urban and architectural scales, as it
consists on “shared” active facets between the building envelope and the open urban can-
opy. Designing the street is, hence, a key issue in a global approach for an environmental
urban design (e.g. Oke 1988, Ali-Toudert and Bensalem 2001). Indeed, the shape of the
street canyon has been reported to influence both outdoor and indoor environments, i.e. the
potential for passive solar gains inside and outside the buildings, the permeability to wind
flow for internal and urban ventilation, the urban absorption versus reflectance of radiation,
as well as the potential for cooling of the whole urban system. By implication, the street
form affects the thermal sensation of people as well as the global energy consumption of
urban buildings.
The strategic importance of the street is also attributable to its function: the street network
of an urban entity has, from a design point of view, a structural role and accounts for the
main support for mobility, urban activity, social life, and even reflects cultural specificities
(e.g. Moughtin 2003).
Climatologically, the main difficulty faced by the designer in shaping a street is the con-
flict in the seasonal internal and external needs, i.e. the required protection from the sun in
the summer and the need for solar access in the winter. Theoretically, these imply com-
pactness and openness to the sky, respectively. Oke (1988) argues that a “zone of compati-
bility” which ensures a compromise between apparently conflicting objectives in the de-
sign of the street can be found. Swaid (1992), for instance, proposes as “intelligent build-
ings” some removable arrangements within the street in order to control shading according
to seasons, which attests the conflicting task. Moreover, traditional and contemporary ar-
18
1. Introduction
chitectures provide a number of attempts of street design according to climate (e.g. Roche
1970, Golany 1982, Herzog 1996, Krishan 1996, Asimakopoulos et al. 2001, Hawkes and
Foster 2002, Thomas 2003). However, quantitative information, based on scientific meth-
ods, about the optimal street design for regulating the climate comfort is still required.
For convenience, the urban canyon (UC) has been widely adopted in urban climatology as
the basic structural unit for describing a typical urban open space (e.g. Oke 1988, Arnfield
1990a, Swaid et al. 1993, Asimakopoulos et al. 2001, Arnfield 2003) namely filtered from
non-climatic relevant aspects. A great deal of information on the most important microcli-
matic changes within an urban street canyon has already been gathered, mainly from stud-
ies conducted in mid-latitude cities. All studies point out the prime importance of the as-
pect ratio or height-to-width ratio (H/W) and the street orientation, being the most relevant
urban parameters responsible of these changes. In fact, these two describers were found to
be decisive in the energy balance of an urban canyon (e.g. Nunez & Oke 1977, Todhunter
1990, Yoshida et al. 1990/91, Arnfield and Mills 1994), in a differentiated potential of ir-
radiation of canyon facets, i.e. floor and walls (e.g. Arnfield 1990a, Mills 1997, Bourbia
and Awbi 2004). Exposure versus shadow patterns affects strongly the canyon surface tem-
peratures and consequently the amount of heat transferred to air as sensible flux and con-
secutively the air temperature (Nakamura and Oke 1988, Yoshida et al. 1990/91, Santa-
mouris et al. 1999). The potential of wind flow at street level also depends on these factors
(e.g. Hussein and Lee 1980, de Paul and Shieh 1986, Nakamura and Oke 1988, Arnfield
and Mills 1994, Santmouris et al. 1999). The building materials of the canyon surfaces
were also found to be decisive in the diurnal heat storage rate of a street canyon (Oke 1976,
Arnfield et al. 1998) as well as in the nocturnal cooling rate (Arnfield 1990b, Mills 1997).
The potential of solar access inside the buildings and, by implication, the site layout and
urban density have also been directly related to street vertical profile and orientation
(Knowles 1981, Capeluto and Shaviv 2001, Krisl and Krainer 2001, Pereira et al. 2001).
In contrast to the large number of studies on street microclimate, studies dealing directly
with outdoor thermal comfort in urban environments are very few, in particular those fo-
cusing on the role of urban geometry. The number of methodological questions on the as-
sessment of human thermal sensation outdoors is also rising:
Many investigations extend indoor comfort methods to outdoors by considering only air
temperature, humidity and wind speed (e.g. Swaid et al. 1993, Coronel and Alvarez 2001,
Grundström et al. 2003). Sometimes, comfort is implicitly related to the only air tempera-
ture and is expressed as “cooling effect” (e.g. Coronel and Alvarez 2001, Shashua-Bar and
19
1. Introduction
Hoffman 2000). Though still used, this approach is inaccurate and valid only in areas
where the mean radiant temperature Tmrt is nearly equal to air temperature Ta and the wind
speed very weak (usually indoors). This is unrealistic outdoors and up-to-date methods of
human-biometeorology already emphasized the prime importance of radiation fluxes in the
human energy balance (Mayer and Höppe 1987, Jendritzky and Sievers 1989, Jendritzky et
al. 1990, Mayer 1993, 1998). Hence, Tmrt can be of more than 30 K higher than Ta in ex-
posed locations and even up to 5 K in shaded parts due to the diffuse and reflect solar ra-
diation components. However, although methodologically more accurate, these studies
focused mainly on land use differences by considering various urban densities and vege-
tated areas and were not explicitly related to street geometry.
One of the very few studies (Pearlmutter et al. 1999), which focused on the effects of street
geometry on radiation fluxes and on the heat gained by a human body, confirmed the ad-
vantage of shading in the reduction of the radiant heat absorbed by a human body com-
pared to a person standing in a fully exposed location. Yet, the actual thermal sensation has
not been clearly evaluated, and one can speculate that the thermal situation would hardly
be comfortable owing to the relatively large aspect ratio (H/W = 1) and the hot and dry
subtropical location considered.
Arnfield (1990a) and Bourbia and Awbi (2004) compared, by means of numerical meth-
ods, the potential of irradiation (or shading) in-canyon for a large number of aspect ratios
and various orientations. The results, given as monthly average values for streets of simple
symmetrical shapes, highlighted the large differences in solar access of canyon facets be-
tween all case studies. Extending such an investigation to human comfort, by considering a
greater diversity of street geometries, i.e. more realistic street forms, and considering the
most relevant times of frequentation by people of urban spaces on a daily basis is thus
highly advisable.
Moreover, a number of studies deal, for convenience, with only one or few points within
the canyon, accounting for representative of the whole area of the street (e.g. Swaid et al.
1993, Pearlmutter et al. 1999). As a result, the spatial microclimatic differences across the
street (centre and edges) are inhibited, yet known to be influencing the microclimate (e.g.
Nakamura and Oke 1988, Arnfield 1990a) as well as the human adaptive behaviour to
thermal stress, favoured by the presence of various microclimatic sub-spaces (Nikolopou-
lou et al. 2001). Thus, an investigation which also considers the spatial microclimatic dif-
ferences on the resulting human comfort is also advisable.
20
1. Introduction
1.2. Objectives of the present study
The present work is primarily motivated by the will to link the theoretical knowledge on
urban microclimate and the practical design process, as this was widely reported to be
lacking. This study seeks to contribute towards a deeper understanding of the thermal sen-
sation in urban open spaces. It addresses the contribution of street design, i.e. aspect ratio,
solar orientation, and further design details towards the development of a comfortable mi-
croclimate at street level for pedestrians in the summertime with a special emphasis on hot
and dry climate. The focus is put on the applicability of the results, i.e. expressed in form
of design guidelines. To do so, the thermal situation in-canyon is analysed with a high spa-
tial resolution in order to highlight the very local variability in the thermal sensation within
the area of the street. This assessment of comfort is also performed on a daily basis in order
to deal with the subjective dimension of the time of frequentation of people. In this work
the following aspects are investigated:
the role of the vertical urban geometry and solar orientation in creating a different
microclimate within the canyon at street level,
the effects of these microclimatic changes on the human thermal sensation outdoors,
the effects of street design details on outdoor thermal comfort, and
the combined effects of outdoor comfort in the summer and indoor solar access in the
winter on the final design of an urban street.
The following design strategies are analysed:

the solar orientation, i.e. E-W, N-S, NE-SW, NW-SE,
height-to-width ratio: H/W
¾ simple geometries, i.e. symmetrical profiles, from shallow to deep streets
¾ complex geometries, that combine various design details, namely:
asymmetry of the street profile
galleries
horizontal overhangs
vegetation
A further objective of the work is related to the methodology used and is explained below.
21
1. Introduction
1.3. Methodology
One reason for the very limited number of field studies on outdoor thermal comfort in rela-
tion to street design is certainly the huge number of urban variables and processes in-
volved. This complexity makes it difficult performing comprehensive field measurements
and is probably the reason why most investigations concentrate on air temperature and
humidity, which are much easier to measure. Indeed, it is costly to record continuously and
for a large sample of streets all-wave radiation flux densities from the three dimensional
surroundings of a human body, in addition to the commonly measured meteorological fac-
tors (i.e. air temperature Ta, wind speed v, and vapour pressure VP).
In this respect, numerical modelling has a distinct advantage over comprehensive field
measurements and is, therefore, a powerful alternative for urban climate issues (e.g. Arn-
field 1990a, Mills 1997, Capeluto and Shaviv 2001, Kristl and Krainer 2003, Bourbia and
Awbi 2004, Asawa et al. 2004). In a recent review of the state of research development in
urban climatology during the last two decades, Arnfield (2003) draw attention to the grow-
ing popularity of numerical simulation, described as a methodology perfectly suited to
dealing with the complexities and non-linearities of urban climate systems.
Hence, the present research is mainly carried out by using a numerical methodology, so
that a series of geometries combined with various street orientations and other arrange-
ments could be analyzed and compared.
Urban microclimate models vary substantially in many aspects: their physical basis, tem-
poral and spatial resolution, input and output quantities, etc. (see 3.1). In our study, the
three-dimensional model ENVI-met, release 3.0 was chosen for the prognosis of all mete-
orological factors and comfort quantities within an urban area (Bruse and Fleer 1998,
Bruse 1999, Bruse 2004). The major advantage of ENVI-met is that it is one of the first
models that seeks to reproduce the major processes in the atmosphere including the simula-
tion of wind flow, turbulence, radiation fluxes, temperature and humidity, and this on a
well-founded physical basis (i.e. the fundamental laws of fluid dynamics and thermody-
namics). ENVI-met simulates the microclimatic dynamics within a daily cycle in complex
urban structures, i.e. buildings with various shapes and heights as well as vegetation. Its
high spatial and temporal resolution enables a fine understanding of the microclimate at
street level.
22
1. Introduction
According to our objectives, it is then possible by means of ENVI-met 3.0 to point out the
spatial variations of human thermal sensation at street level by differentiating between
street edges (sidewalks) and street centre. The model also requires relatively few input pa-
rameters and calculates all required meteorological factors, namely air and surface tem-
peratures, wind speed and direction, air humidity, short-wave and long-wave radiation
fluxes as well as the mean radiant temperature needed for comfort analyses.
Assessing comfort outdoors is not easy and methodological differences observed in the
related literature make any comparison with available results difficult, and this will be dis-
cussed in the next chapter. Basically, comfort can be assessed by means of comfort indices.
In this work, thermal comfort is expressed by means of the physiologically equivalent tem-
perature (PET), which is an up-to-date human-biometeorological thermal index. It is based
on the human energy balance and is taking into account the physiological capacities of a
human body to adjust to stressful microclimates (Höppe 1993, 1999).
A further objective of this work is to provide some additional information on the accuracy
of the relatively new model ENVI-met. This is not an easy task, and as already noticed by
Arnfield (2003), the validation of numerical models, unfortunately, lags behind their crea-
tion and when performed, is often weak, relying more on plausibility of outputs than on
direct comparison with process variables. According to the author, this is not surprising,
because the difficulty of measuring such variables is a prime reason why numerical model-
ling is so popular, and a closer collaboration between modellers and field climatologists is
encouraged to close the methodological gap.
Therefore, two short-term field measurements are also conducted to allow further compari-
son and discussion. On-site measurements have been carried out for:
Freiburg, Germany, mid-latitude, in temperate climate.
Ghardaia (Beni-Isguen), Algeria, subtropical, in hot-dry climate.
1.4. Structure of the thesis
Chapter 2 summarizes the most significant findings related to urban canyon microclimate
and human comfort outdoors. Chapter 3 recalls briefly the physical statements which gov-
ern the model ENVI-met, with a focus on those of particular relevance in the framework of
this work. The results of the numerical simulations are discussed comprehensively in
Chapter 4. In addition, field observations are presented in Chapter 5 for Freiburg, Ger-
many and in Chapter 6 for Ghardaia, Algeria. A general discussion on the relevance of
23
1. Introduction
street design on thermal comfort follows in Chapter 7. It includes a number of design rules
of thumb as well as an evaluation of the model ENVI-met as prognosis tool of the urban
microclimate and outdoor comfort.
Remark: Symbols used in this work correspond to those commonly used in the interna-
tional literature. Yet and for convenience of the reader, the nomenclature used to describe
ENVI-met is kept unchanged from the original source (Bruse 1999). Therefore, some
physical quantities are referred to with more than one symbol through this manuscript. All
are listed in p. 214.
24
2. Literature review
Microclimatology deals spatially with the layer of air directly above the earth surface in
which the effects of the surface (frictions, heating, cooling) are felt directly on time scales
of about one hour and in which significant fluxes of momentum, heat or matter are carried
by turbulent motions (Garrat 1992). This layer can extend up to 2 or 3 km above the sur-
face and is known as the atmospheric boundary layer ABL. Above the ABL, the air is
mainly influenced by macro-scale processes and reacts slowly to the changes near the
ground. Urban areas are typical examples of profound local climatic modifications and are
commonly known as the urban canopy layer UCL (Oke 1987). The UCL extends from the
ground surface up to building roof heights. The urban structure is an inhomogeneous and
rough “surface”, which increases the turbulent processes and leads to a high variability in
space and time of all meteorological quantities and so to a mosaic of microclimates taking
place in a quite limited area.
Designing comfortable urban spaces must take into account these modifications. This as-
sumes that the interdependence between the urban geometry and the climate is well under-
stood and translated in readily understandable guidelines for the designer, so that the cli-
matic dimension can be included within the design process, together with all other design
imperatives, i.e. economic, functional, socio-cultural and aesthetic.
The following material is a theoretical background, which reports on the most relevant
studies dedicated to the investigation of the microclimate of an urban street and to thermal
sensation of people in outdoor spaces.
25
2.1. The microclimate of an urban street canyon
Obviously, a representative urban canyon is quite impossible to find if all modifying pa-
rameters have to be considered: aspect ratio, orientation, construction materials, presence
of vegetation, etc. (Oke 1987). Nevertheless, the major characteristics of the microclimate
of an urban canyon were clarified by a large number of studies and are summarized below.
2.1.1. Energy budget of an urban canyon
The pioneering investigation of Nunez and Oke (1977) identified the basic knowledge on
the energy budget of an urban canyon. Further studies confirmed and completed these first
findings (e.g. Todhunter 1990, Yoshida et al. 1990/91, Mills and Arnfield 1993, Arnfield
and Mills 1994). The study dealt with a street located in Vancouver (49 °N), north-south
oriented, with an aspect ratio close to unity and nearly symmetric (H1/W = 0.86 and H2/W
= 1.15). The walls are made of concrete, white painted and windowless. Some sparse vege-
tation is also available. The energy balance of both walls and floor are given by:
*
Qwall = QH + ∆QS (2.1)
and
Q *floor = QH + QE + ∆QS (2.2)
where Q* is the net all-wave radiation, QH is the sensible heat flux, QE is the latent heat
flux, ∆QS is the energy stored in the walls. Advection is neglected and anthropogenic heat
included in ∆QS. All energy components were measured during three days, but the sensible
heat flux which was obtained as residual. The main results showed that the influence of the
canyon geometry on the radiation exchanges affected strongly the timing and magnitude of
the energy regime of the individual canyon surfaces and were very different from each
other. The orientation was also found to have an evident importance on the energy balance.
Fig. 2.1 illustrates the diurnal course of all fluxes for the floor, the east-facing wall and the
urban system (canyon-top). The east-facing wall is first irradiated in the morning and the
second peak in the afternoon corresponds to the diffuse radiation mainly reflected from the
opposite wall which experiences a maximum irradiation at that time. According to the N-S
orientation, the floor is exposed at midday and the west and east walls about 1.5 hours be-
fore and after solar noon. By day, about 60% of the radiant energy surplus was dissipated
as a sensible heat flux, 25-30% stored in the materials and 10% transferred to air as latent
26
heat. The diurnal course of the energy balance of the canyon system is relatively smooth
and symmetric, comparable to a horizontal surface, in spite of the different energy ex-
change schemes for each surface. This is due to the convection of the heat energy out of
the canyon whereas the rest is stored in the materials.
Fig. 2.1. Daily energy balance of urban facets of an urban canyon oriented N-S with H/W
≈ 1 for a sunny summer day in Vancouver, 49 °N (Nunez and Oke 1977)
In the night-time the net radiative deficit is almost entirely offset by the release of the en-
ergy stored (∆QS) in the canyon materials and turbulent exchange is minor. Moreover, the
wind direction and speed as well as the nature of the surrounding thermal environment may
have contributed in advective motions which, however, should not have exceeded 100
Wm-2. Finally, the authors suggest that with airflow directed at an angle in relation to can-
yon axis it appears as if transport by the mean flow may be important.
A similar field study was carried out by Yoshida et al. (1990/91) and confirmed many of
these results. The street investigated is located in Kyoto (35°N), oriented E-W, symmetric
and with H/W = 0.96. Surface and air temperatures, wind flow as well as energy fluxes
were measured continuously for several days under clear summer conditions. The energy
fluxes were reported as canyon-top totals and were compared to roof surface. The authors
showed that the energy budget into the canyon is about 1.5 times as much as that into the
roof surface. The daytime canyon-top sensible heat flux averages 40% (against 60% for the
Vancouver canyon) attributed to more shading involved by the orientation, the presence of
windows, as well as to weak in-canyon airflow. The heat transfer of the shaded walls was
negligible in comparison with the sunlit walls. Thus, it was suggested that turbulent heat
27
fluxes from the air to shaded surfaces may occur leading to a reduction in the canyon-top
sensible heat flux. In the night-time, there is no significant difference in the energy balance
between the roof surface and the canyon system; the net radiation and the substrate heat
fluxes were nearly equal.
A numerical simulation which compares the energy balance of an E-W urban canyon (rep-
resented at its top) and a parking lot (Sakakibara 1996) confirmed that the urban system
absorbs more heat in the daytime and releases more heat at night than a horizontal surface,
whereas another field study (Mills and Arnfield 1993) argued that as street canyons be-
come narrower they become increasingly isolated in term of heat exchange from the over-
laying atmosphere.
Most of the studies focused on symmetrical geometries and the microclimatic variations
attributable to urban canyon asymmetry and orientation have been rarely investigated (e.g.
Todhunter 1990, Mills and Arnfield 1993). A detailed numerical simulation of the energy
budget of six symmetric and asymmetric canyons with various orientations for summer
conditions was preformed by Todhunter (1990). Basically, it was found that net radiation,
net solar radiation and turbulent sensible heat are particularly sensitive to urban geometry
and the interurban system daily net short-wave variations K* are significant ranging from
25 to 85% and the variations in daily long-wave irradiances L* about 10 to 15%. Net all-
wave radiation Q* and sensible heat flux QH variations are significant between the different
geometries.
2.1.2. Thermal characteristics of an urban canyon
The temporal evolution and spatial distribution of air temperature within an urban canyon
were first thoroughly investigated by Nakamura and Oke (1988). A network of 63 measur-
ing points was set in a vertical cross-section of an urban canyon located in Kyoto, Japan,
during clear summer days. The grid network was arranged with increasing points’ fre-
quency towards the floor and walls (Fig. 2.2). The street had its axis oriented E-W, an as-
pect ratio of nearly 1 and was made of typical urban materials.
The results showed small differences between roof and canyon air temperature Ta of about
0.5 to 1 K, mainly due to the well-mixed turbulent air within the canyon and above it. The
roof air was systematically cooler by day and warmer by night.
28
Fig. 2.2. Isotherm distribution across an E-W canyon at selected daytime hours, also in-
cludes wind speed, wind direction and stability conditions at 1 m height (Nakamura and
Oke 1988).
Basically, Ta gradients within the canyon air were found to be small, except adjacent to the
irradiated urban facets, at which Ta was visibly higher than the mean value measured at the
canyon centre. The difference reached 2 to 3 K (Fig. 2.2).
Yoshida et al. (1990/91) and Eliasson (1993) confirmed the small roof-canyon air tempera-
ture vertical gradients, across the street as well as along it. No clear correlation could be
found between street geometry and Ta. In the Kyoto’ canyon, the south facing wall and
floor were the primary sites of solar absorption during the day, and their role as a source of
sensible heat for the canyon continued in the night-time.
Moreover, Nakamura and Oke (1988) observed large differences between the surface tem-
perature Ts and the adjacent Ta for directly irradiated urban facets. Near midday, the differ-
ence (Ts – Ta) exceeded 10 K. At night, the residual heat kept the surface by a few degrees
warmer. These differences are much smaller at the shaded part of the canyon, which only
received diffuse solar radiation. Air temperature may even be higher than Ts in the shade,
likely due to the warming of the whole air volume by turbulent sensible heat flux transfer
from sunlit surfaces and its mixing through vortex air circulation.
29
Fig. 2.3. Surface and air temperatures of urban canyon facets, for an E-W street of an as-
pect ratio H/W = 0.96 under sunny summer conditions for Kyoto, Japan, 35°N (Yoshida et
al. 1990/91)
These findings were confirmed by Yoshida et al. (1990/91) for a similar canyon (Fig. 2.3)
and by Santamouris et al. (1999) for a deeper street and a different orientation.
Santamouris et al. (1999) performed measurements under hot weather conditions in Athens
(30°N) in a pedestrian street oriented NNW/SSE with H/W = 2.47. This study highlighted
the vertical Ts and Ta distribution in deep profiles. Surface temperature differences be-
tween the opposite surfaces at different levels were high, with a maximum difference of 14
K to 19 K. By day, the simultaneous difference in Ts was lower at ground level and in-
creased with height within the canyon. This difference became insignificant in the night-
time (< 2 K). Furthermore, Ts stratification was found to be larger for the SW façade (0 to
10 K) than for NE façade (0 to 3 K) due to the different daily solar exposure. The higher
temperatures were recorded for the higher part in both cases.
In contrast to the larger Vancouver and Kyoto streets, the roof-canyon ∆Ta as well as the
vertical air stratification in the Athens’ street was larger. A maximum Ta difference ≈ 2 to
3 K vertically. Moreover, Ta close to the SW façade was higher than at the NE façade with
a mean difference of 3 K on average and reached 4.5 K. The (Ts –Ta) difference was about
13 K for the SW façade against only 5 K for the opposite façade mainly shaded. Unlike
these studies, Coronel and Alvarez (2001) and Grundström et al. (2003) reported on much
lower in-canyon Ta, i.e. 8 K for H/W = 5 and 10 K for H/W = 10 in comparison to free air
in old compact cities in subtropical locations, respectively.
30
2.1.3. Wind flow in an urban canyon
The wind flow within an urban canyon is a secondary circulation feature driven by the
above-roof dominant flow (Nakamura and Oke 1988, Santamouris et al. 1999), which is
strongly affected by the street orientation and geometry (H: height, L: length, W: width).
When the flow over arrays of buildings is approximately normal to street axis, three re-
gimes can take place depending on the aspect ratio (H/W) and building ratio (L/W). The
transition from one regime to another occurs at critical combinations of H/W (Hussein and
Lee 1980) and L/W (Hosker 1985) as shown in Fig. 2.4.
Fig. 2.4. (a) Wind flow regimes (Oke 1988) and (b) corresponding threshold lines dividing
flow into three regimes as function of canyon (H/W) and building (L/W) geometry (Hosker
1985)
The isolated roughness flow occurs between well spaced buildings, when the windward
and leeward flows do not interact, comparably to a wind movement around an isolated
obstacle. As the H/W increases, the wakes are disturbed leading to a wake interference
regime. With further increase of H/W, the street canyon becomes isolated from the above
circulating air and a stable circulatory vortex is established in-canyon, leading to a skim-
ming flow. The latter regime is the most common in urban contexts and has, therefore,
drawn the most attention.
The correlation of wind speed between in-canyon and above-roof wind flow is found to be
marked for high winds, resulting in a stable vortex circulation. This correlation is lost for
lower wind speeds, leading to much more scattering (Nakamura and Oke 1988, Santa-
mouris et al. 1999). A threshold wind speed of 1.5-2.0 ms-1 between both situations was
31
mentioned by McCormick (1971), de Paul and Shieh (1986) and Nakamura and Oke
(1988). Above this threshold, wind speed in the canyon was reported to increase propor-
tionally to free ambient air (de Paul and Shieh 1986, Yamarito and Wiegand 1986, Lee et
al. 1994). In case of light winds, the air canyon flow is not only a mechanically driven cir-
culation but thermal effects due to canyon facets irradiation may play a role (e.g. Naka-
mura and Oke 1988, Santamouris et al. 1999). Explicitly, the temperature cross-section
suggests the formation of one circulatory vortex mixing cool air from above the roof into
the canyon space and expelling warming air. Moreover, the differential heating of street
surfaces can shift the flow from one regime to another (Sini et al. 1996) and from a one-
vortex flow to a flow with several contra-rotative vortices (Sini et al. 1996, Kim and Baik
1999).
In a deep canyon, wind flow at street level relates to the prevailing wind speed above roof
level as well as to in-canyon thermal stratification (Santamouris et al. 1999). For perpen-
dicular winds, either one circulatory vortex occurs driven by the ambient air flow, or a
double vortex takes place, where the lower vortex is induced from the upper one and is in
the opposite direction (Santamouris et al. 1999, Baik et al. 2000). The speed of the single
vortex results from three specific mechanisms: the ambient air flow above the canyon, the
vertical stratification of air inside the canyon, and the mechanism of advection from the
buildings ends. When the wind above roof is predominant, the speed within the canyon
increases (under a threshold of 2 to 3 ms-1). A double vortex is almost always observed,
together with temperature stratification. Under these conditions, higher ambient winds con-
tribute to the transmission of more energy from the upper to the lower vortex and hence
increase its speed.
The wind direction within the canyon depends on the incidence of ambient air in respect to
street axis. When the flow at roof level is normal to the canyon axis, an opposite direction
prevails at street level (e.g. Hoydysh and Dabbert 1988, Nakamura and Oke 1988, Arnfield
and Mills 1994, Santamouris et al. 1999). A wind blowing parallel to street axis generates a
channelling of the mean wind (Wedding et al. 1977, Nakamura and Oke 1988, Santamouris
et al. 1999) with possible uplift along the canyon walls due to increased friction near the
surfaces (Nunez and Oke 1977). The canyon wind speed is then proportional to free wind
speed. When the flow above roof is at some angle of attack on the canyon, a spiral vortex
(or cork-screw type) is induced along the canyon (Wedding et al. 1971, Dabberdt et al.
1973, Nakamura and Oke 1988, Santamouris et al. 1999). This oblique incidence improves
32
the potential of urban ventilation and hence promotes the ventilation indoors in comparison
to a perpendicular incidence (Wiren 1985, 1987, Bensalem 1991).
The parallel component of the main wind determines the along canyon stretching of the
vortex and the transversal component drives the canyon vortex (Yamartino and Wiegand
1986). A simple reflection concept is proposed by Nakamura and Oke (1988) as a useful
first approximation, i.e. ddcanyon = 180° - ddroof for ddroof = 0°-180° and ddcanyon = 540° -
ddroof for ddroof = 180°-360°, where 0° corresponds to north direction. Wind speeds above
and within the canyon indicate an approximately linear relationship: ucanyon = 2/3 uroof ,
applicable for similar canyon dimensions (i.e. H/W ≈ 1) and horizontal wind speed (u) at
roof level up to 5 ms-1. Nakamura and Oke (1988) also suggest an angle of incidence
equals to reflection’s angle. However, the latter shows some evidence of being lower, due
to eventual entrainment along the canyon.
Furthermore, Chan et al. (2001) found that non-uniformly building heights provide better
ventilation and tall buildings do not necessarily promote blockage. A wider canyon pro-
motes better mixing of air and canyon geometry should be restricted to threshold value for
skimming flow and maximum relative canyon length ratio L/H should be kept at five.
End canyon effects play an important role on the air flow distribution in canyons. For L/W
≈ 20 finite canyon length begin to dominate over the vortex (Yamarito and Wiegand 1986).
Intermittent vortices are shed on the building corners, and these vortices are responsible for
the mechanism of advection from building corners to mid-block creating a convergence
zone in the mid-block region of lowest wind speeds (Hoydysh and Dabbert 1988, Santa-
mouris et al. 1999).
2.1.4. Solar access outdoors
Arnfield (1990a) investigated by means of a numerical method the solar access within
various canyons. The aim was to explore the dependence on aspect ratio and orientation of
the irradiances on canyon facets (walls and floor) and on a pedestrian model. The study
was conducted for aspect ratios ranging from 0.25 to 4 and for E-W and N-S street orienta-
tions for all latitudes and seasons. The monthly average irradiations revealed that the H/W
ratio first determines the quantity of energy received on the whole urban surfaces. The ex-
posure of urban surfaces to the sun in an urban profile decreases as the profile becomes
deep (Fig. 2.5).
33
Fig. 2.5. Monthly mean canyon irradiances simulated for June for E-W and N-S canyons
and various aspect ratios. The symbols +, x, ∗, , ∆, ο correspond to H/W = 0.25, 0.5, 1, 2, 3,
and 4 respectively (Arnfield 1990a)
However, the availability or distribution of this solar energy on the different surfaces of the
profile is unequal. Basically, the vertical surfaces (walls) are less irradiated than the streets
for a same canyon and the variation of H/W seems to affect the streets more than the walls.
Besides, one can differentiate between the urban borders (often pedestrian paths) and the
central part of the street in relation to the irradiations they receive.
The orientation appears to be more decisive for the exposure of the walls and the depth of
the profile more decisive for the exposure of the ground. The importance of the orientation
is also more significant in summer than in winter. The exposure of building’s walls ori-
ented N-S (i.e. E-W streets) allows an easier seasonal solar control because the walls are
protected in the summer and exposed in the winter. For the pedestrian, the irradiations are
nearly independent of the orientation and vary little during the year.
In the winter, the sun position is lower for higher latitudes and generates strong obstruc-
tions. Hence, the irradiances decrease for high latitudes. This is particularly noticeable for
the E-W orientation. The latitudes 20°- 40° show the largest contrasts in the exposure of
34
the street floor and walls in dependence to H/W, orientation and between the seasons. This
suggests that urban geometry is of prime importance in the solar control in the subtropics.
This has been highlighted by Bourbia and Awbi (2004) who did a similar investigation for
latitude 33°N. They used a shading factor SF, which is the complementary fraction of SAI
as suggested by Arnfield 1990a, and where SAI is the potentially available solar energy
received by a facet. The shading factor was calculated on a monthly basis for a variety of
street orientations with a step of 15° (Fig. 2.6).
Fig. 2.6. Mean monthly shading fraction SF for canyon, floor and walls in dependence with
aspect ratio H/W during summer and winter for latitude 33 °N (Bourbia and Awbi 2004).
For E-W oriented streets, the walls show large contrasts in SF between summer and winter.
Explicitly, SF is very small in the winter and less dependent on H/W, while SF is high in
the summer. Inversely, the floor is mostly shaded in the winter for H/W ≥ 1 and only partly
shaded in the summer even for high aspect ratios. For a N-S orientation, the differences are
less manifest between summer and winter for the floor and shading increases as the aspect
ratio increases, i.e. about 0.6 for H/W = 1 and about 0.8 for H/W = 4. The walls are more
shaded in the winter than in the summer as the aspect ratio increases. Fig. 2.6 also shows
that the situation observed for an E-W oriented street extends only to ± 15° from due orien-
tation, while the shading patterns observed for N-S extend up to 30° from due north to-
wards E and W.
35
2.1.5. Solar access indoors
Indoor solar gain is a challenging issue in urban context, where the main difficulty is the
reduction of the potential of solar irradiation due to the effects of sun obstruction from the
surroundings. The concept of solar envelope initiated by Knowles (1981, 2003) was a first
attempt to resolve this problem and shows how the urban forms are shaped according to
architectural purposes.
Basically, the solar envelope can be defined as the largest volume on a plot, which allows
solar access to all adjacent parcels to useful sunshine times. The shape and size of a solar
envelope depends on the sun path (determined by the latitude) and on the desirable sun-
shine duration. Moreover, the solar envelope depends on the plot form and its orientation,
and controls directly the geometry of the streets (Fig. 2.7): the resulting streets are symmet-
rical if the street axis is N-S oriented and rather asymmetrical if E-W or NE-SW oriented,
with the walls facing the sun being higher. Furthermore, the association of individual plots
is more profitable in E-W direction, so that each building could have a façade oriented to
the south. Yet, even though the solar envelope decides on the urban prospects for determin-
ing building heights and prospects, it does not deal with the outdoor microclimate.
Fig. 2.7. Three different building blocks orientations showing the effect of the solar enve-
lope on the shape and size of the urban streets geometries (Knowles 1981)
36
Pereira and Minache (1989) and Pereira et al. (2001) confirmed the usefulness of the
asymmetry of a street canyon in relation to orientation for optimising the internal solar
gains and justly questioned the utility of fixed regulations based on simple symmetrical
aspect ratios. Capeluto and Shaviv (2001) highlighted the importance of street orientation
in deciding on the plan density (by implication H/W). Basically, rows of buildings along
NE-SW, E-W and NW-SE streets reveal to allow (in this order) the highest urban plan den-
sity while preserving winter solar rights. Similarly, Kristl and Krainer (2001) showed that
by increasing building heights the orientation becomes more decisive in increasing the plan
density while keeping sufficient solar access.
This “envelope” concept has been even used to design self-shading buildings (e.g. Cape-
luto 2003). This results, for instance, in inclined façades which shapes reversed pyramids,
so as both sidewalks and façades are shaded during a required period in summer and ex-
posed to winter’s sun. The solar envelope shows, that by shaping the buildings, a street
may not necessarily be vertical and can promote or prevent solar access through design
details.
2.1.6. Effects of the vegetation
The vegetation is a modifying factor of the local climate. The use of the green as a strategy
to mitigate the urban heat island (UHI) and improve the microclimate has been widely em-
phasized (e.g. Escourrou 1991, McPherson et al. 1994a, Akbari et al. 1995, Avissar 1996,
Taha et al. 1997). A quantitative evaluation of the climatic role of the urban vegetation is
required since this is also planned for other tasks, e.g. acoustics, reduction of pollution,
aesthetics, social issues, etc. (e.g. Givoni 1997).
Studies on the effects of the vegetation on thermal outdoor comfort are very few, in par-
ticular those specifically addressed to urban streets (e.g. Shashua-Bar and Hoffmann 2000).
However, a number of studies on the climatic effects of urban vegetation is available and
provides much useful information for urban designers. Methodologically, the climatic ad-
vantages of urban vegetation are assessed either according to their effects on the meteoro-
logical factors (e.g. Ta, RH or v) or to the induced energy savings in the buildings as a re-
sult of less cooling and/or heating loads. Primarily, the vegetation possesses three main
properties which affect the climate: shading, humidification (evapotranspiration) and wind-
break (McPherson et al. 1994a). Indirectly, it also acts as a medium to trap water inside the
soil. Any use of vegetation for improving the microclimate has to exploit judiciously these
37
properties according to site comfort requirements (Moffat and schiler 1981). Moreover,
many related studies use numerical modelling, on one hand because field experiments are
costly, and on the other hand to allow comparisons of various scenarios that include sea-
sonal changes and growth of the vegetation (density, size, etc.). The main results of rele-
vance for the present work are summarized below.
For a single tree, the shading effect is easily estimated while the cooling by evapo-
transpiration is more difficult to assess. This is because the fresh air generated is rapidly
diffused in the air volume which traverses the tree crown (McPherson and Simpson 1995).
Evapotranspiration effects and wind speed reduction can be easily evaluated for aggregated
trees. For instance, the cooling effect in a residential area largely provided with trees can
experience lower air temperatures (up to 5 K) and 50% less wind speed (McPherson and
Simpson 1995).
Within an urban structure, the climatic effectiveness of the vegetation depends on the ratio
green area / built-up area, as well as on the size, location and own characteristics of the
plant (species, density, shape, size, volume, age, etc.). The benefits of the vegetation in-
crease with the increase of its proportions (Saito et al. 1990, Avissar 1996). The cooling
effects (e.g. Ta decrease) of large parks was found to extend in the surroundings at a radius
of several hundred metres ( e.g. Avissar 1996, Ca et al. 1998) and can even lead to breezes
(e.g. Geiger et al. 1995, Eliasson and Upmanis 2000). However, these effects become in-
significant for small vegetated areas (Shashua and Hoffman 2000). Hence, it has been sug-
gested that several smaller areas with sufficient intervals are more cooling effective than
one large green space (Honjo and Takakura 1990, McPherson 1992, McPherson et al.
1994b).
In densely built environments, trees can be located in places, parking areas, street intersec-
tions or in rows along the streets. The usefulness of the latter solution should not be under-
estimated as reported by McPherson (1994) and McPherson et al. (1994b). They found for
Chicago large economies gained from green cover, from which one-third consisting in
alignment of trees in urban streets. The vegetation leads to the most energy savings if
planned in residential areas where the energy needs are high, i.e. yards or streets, from
50% to 65% of energy savings.
McPherson and Simpson (1995) assessed various trees’ properties on energy savings. Tree
efficiency is found to be dependent on orientation: a tree located for shading a west facing
wall is as efficient as two identical trees on the east. On the south, the benefits are slightly
offset by the negative effects of obstruction in the winter. The shape and volume of the tree
38
can be more important than its density in relation to shading because the seasonal variabil-
ity of the crown diameter is greater (4 to 16 m) than that of its density (60 to 90%). Large
trees can shade more surface than narrow and dense trees and two trees are about five
times more effective than one tree.
Moreover, the imbrications of green and built-up areas can be conflicting and the vegeta-
tion can influence the microclimate negatively if not appropriately designed, e.g. by block-
ing solar access in the winter or reducing wind movement in the summer. The branches of
deciduous trees planted for summer shading may reduce up to 30-40% of the desirable
solar gains in the winter (McPherson 1992) and the permeability of trees to solar radiation
has to be evaluated for both seasons if planned in the vicinity of buildings (Cãnton et al.
1994). This points out the importance of the tree species, for instance evergreen trees
nearby buildings facing south should be avoided. The optimal location of the vegetation is,
hence, a crucial design criterion.
Shashua-Bar and Hoffman (2000) investigated the cooling effects of shade trees at small
urban green sites, courtyards and streets in a subtropical location. They found for several
planted urban streets that the cooling effect is about 1 K and up to 3 K at the hottest hour of
the day. The highest effects are registered at the centre of the canyon at mid-distance from
the edges but the cooling effect decreases noticeably when moving to street edges. The
vapour pressure VP recorded inside the planted areas was found to be insignificantly vari-
able from the nearby non-planted reference point. This was explained by the lack of soil
irrigation which led to a low evapotranspiration rate. Shading was estimated to be at 80%
responsible of the resulting cooling effect. However, shading was effective only locally
and fluctuations in the reduction of Ta were observed in the same street between different
measuring points (spaced by about 20 m) because of the non-uniform shading.
For hot climates, the best use of the vegetation should profit from its shading property to
mitigate the intense solar radiation in the summer as the overheating is mainly due to the
storage of heat by the sunlit surfaces. The evapotranspiration is often weak owing to the
lacking water in the soil unless irrigation is supplied. A sparser vegetation well mixed
within the urban structure to produce as much shadow as possible has to be preferred in hot
and dry climates (McPherson et al. 1994b). For cold climates using the vegetation as
screen against high winds is more appropriate and dense vegetation located at the urban
edges is advisable.
39
2.1.7. Further aspects
Further important aspects affecting the microclimate of an urban street are the nocturnal
cooling and the nature of canyon surfaces. The nocturnal cooling of the urban fabric was
directly related to the sky view factor and thus to the aspect ratio (e.g. Oke 1981, Arnfield
1990b, Eliasson 1993, Mills 1997). At night, the canyon balance consists of the deficit be-
ing offset by the release of energy stored in the canyon materials and the role of the floor
and the façades as a source of sensible heat for the canyon continues in the night-time (e.g.
Nunez and Oke 1977, Nakamura and Oke 1988, Arnfield and Mills 1994). Surface tem-
perature of the street remains at 0.5 - 1 °C lower than the façade temperatures by night, due
to a larger sky view of the horizontal surface (Santamouris et al. 1999). During the night-
time vertical stratification of the air temperature is low, i.e. less than 0.5 °C for each level,
with higher air temperatures measured at the ground level and decrease with the height. At
night the simultaneous differences in the surface temperatures are insignificant, max. 2 K
(Santamouris et al. 1999).
With respect to urban surfaces, Aseada et al. (1996) pointed out the importance of the
pavement materials in the resulting heat fluxes and air-ground interface on summer days.
They reported that an asphalt pavement emits an additional 150 Wm-2 infrared radiation
and 200 Wm-2 sensible transport compared to a bare soil surface. The water content in a
bare soil and thus the evaporation from it produces much lower surface temperatures. By
contrast, waterproof soils such as asphalt, increasing thickness of the covering material
increase the temperature and heat stored under the surfaces (Asaeda and Ca 1993). Urban
surfaces with high albedos typical of light colours reduce the storage in the materials (Doll
et al. 1985, Akbari et al. 1995, Taha 1997, Taha et al. 1997).
40
2.2. Outdoor thermal comfort
2.2.1. The thermal comfort indices
While comfort sensation indoors is well documented (e.g. Fanger 1970, Givoni 1976,
Brager and de Dear 1998, ASHRAE 2001a), assessing comfort outdoors is by far less well
understood. Basically, comfort assessment methods applied outdoors have been adjusted
from those originally conceived for indoors. The following material discusses the most
important questions related to this issue.
Several definitions of thermal comfort exist. ASHRAE (2001a) highlights its subjective
and psychological dimension by describing comfort as a condition of mind, which ex-
presses satisfaction with the thermal environment. A more rational definition relates com-
fort to energy gains and losses and describes the state of comfort as satisfied when the heat
flows to and from the human body are in equilibrium (Fanger 1970). This is achieved when
the body data, i.e. skin temperature, sweat rate and/or core temperature, are within a range
of comfort. These data are partly governed by the thermo-physiological regulations of a
human being.
Assessing the human thermal comfort is not a recent issue and is not obvious. People have
always been concerned by their well being and looked for methods to quantify their sensa-
tion of cold or heat (e.g. Houghton and Yaglou 1923, Missenard 1948). The thermal envi-
ronment and its impact on a human body cannot be described as a function of one single
factor (e.g. Ta) because the body does not possess individual sensors for each factor and
consequently feels the thermal environment as a whole. A thermal index is based on the
same idea: it combines several factors (e.g. Ta, RH, v, radiation fluxes, etc.) into a single
variable which sums up their simultaneous effects on the sensory and physiological re-
sponses of the body (Givoni 1976, ASHRAE 2001a).
A large number of thermal indices exist and this might be confusing at first, but in fact,
most of them share many common features and can be classified in two groups: empirical
or rational. These indices are well documented (e.g. Givoni 1976, Houghton 1985, ASH-
RAE 2001a) and some of them are exemplarily listed in Table 2.1.
The indices of the former group, generally developed earlier, are based on measurements
with subjects or on simplified relationships that do not necessarily follow theory (ASH-
RAE 2001a). These are often limited to the estimation of the combined effect of air tem-
perature, air humidity and air speed on people in sedentary activity (Givoni 1976).
41
Table 2.1. Selected thermal comfort indices for indoors and outdoors (Fanger 1970, Givoni
1976, and ASHRAE 2001a)
Index Definition
Empirical indices
ET set in Monograms and represent the instantaneous thermal sensation estimated experimen-
Effective Temp. tally as a combination of Ta, RH and v
RT comparable to ET but tested for a longer time to meet assumed thermal equilibrium
Resultant Temp.
HOP temperature of a uniform environment at a relative humidity RH = 100% in which a person
Humid Operative Temp. looses the same total amount of heat from skin as the actual environment (comparable to
ET* but RH equals 50% for HOP)
OP arithmetic average of Ta and Tmrt, that is including solar and infrared radiant fluxes
Operative Temp. weighted by exchange coefficients
WCI based on the rate of heat loss from exposed skin caused by wind and cold and is function of
Wind Chill Index Ta and v, suitable for winter conditions
Rational indices
ITS assumes that within the range of conditions where it is possible to maintain thermal equilib-
Index of Thermal Stress rium, sweat is secreted at sufficient rate to achieve evaporative cooling.
HSI ratio of the total evaporative heat loss Esk required to thermal equilibrium to the maximum
Heat Stress Index of evaporative heat loss Emax possible for the environment, for steady-state conditions
(Sskin=Score=0) and Tsk = 35°C constant
ET* temperature of a standard environment (RH = 50%, Ta = Tmrt, v < 0.15 ms-1) in which the
new Effective Temp. subject would experience the same sweating SW and Tsk as in the actual environment. It is
calculated for light activity and light clothing.
SET* similar to ET* but with clothing variable. Clothing is standardized for activity concerned.
Stand. Effective Temp.
OUT_SET* similar to SET* but adapted to outdoors by taking into account the solar radiation fluxes.
Out. Stand. Eff. Temp. Reference indoor conditions are: Tmrt = Ta ; RH = 50% ; v = 0.15 ms-1.
PMV and PT PMV expresses the variance on a scale from -3 to+3 from a balanced human heat budget
Predicted mean vote and PT the temperature of a standardized environment which achieves the same PMV as
Perceived Temp. the real environment. Clothing and activity are variables.
PET temperature at which in a typical indoor setting: Tmrt = Ta ; VP = 12h Pa ; v = 0.1 ms-1, the
Physiol. Equiv. Temp. heat balance of the human body (light activity, 0.9clo) is maintained with core and skin
temperature equal to those under actual conditions, unity °C.
Yet, these empirical indices ignore the decisive role of human physiology, activity, cloth-
ing, and other personal data (height, weight, age, sex). Rational indices are more recent,
promoted by the lately development of computing techniques, and rely on the human en-
ergy balance. Here, the heat transfer theory applies as rational starting point to describe the
various sensible and latent radiation flux exchanges, together with some empirical expres-
sions describing the effects of known physiological regulatory controls (ASHRAE 2001a).
42
2.2.2. The human energy balance
The energy exchanges between a person and the surrounding environment is illustrated in
Fig. 2.8 and expressed by the following heat energy balance equation:
M + W + Q * + QH + QL + QSW + QRE = S (2.3)
All terms of equation 2.8 are expressed in (W), where M is the metabolic rate (i.e. internal
energy production by oxidation of food), W the physical work output, Q* the net radiation
budget of the body, QH the convective heat flow (sensible), QL the latent heat flow for dif-
fusion of water vapour, QSW latent heat flow due to evaporation of sweat, QRE respiratory
heat flux (sum of heat flow for heating and humidifying the inspired air) and S is the stor-
age heat flow for heating (positive value) or cooling the body (negative value).
The detailed mathematical statements for each of these terms are thoroughly documented
(e.g. Fanger 1970, Gagge et al. 1971, Gagge et al. 1986, Höppe 1984, VDI 1998, ASHRAE
2001a). Basically, the body state influences many of these heat fluxes through body tem-
peratures and skin wetness. The meteorological factors also affect a number of individual
terms as follows:
QH = f (Ta, v); QRE = f (Ta, RH); QSW = f (RH, v; and Q* = f (Tmrt).
Fig. 2.8. The components of the human heat balance (Houghton 1985)
43
Equation (2.3) is the basis for all energy balance models for indoors as well as for out-
doors. The differences between the various existing models are attributable to the comple-
mentary parameterizations related to personal data required to solve eq. 2.3.
The comfort equation proposed by Fanger (1970) is probably the most well-known appli-
cation of the human energy balance. The Fanger’s equation applies indoors and assumes
comfort conditions, i.e. by setting the term S equal to zero. The unknown mean skin tem-
perature (Tsk) and sweat secretion (SW) are replaced assuming a linear relationship with
the activity (internal heat production). These interrelations were defined empirically in
indoor conditions on the basis of field surveys involving a large number of sedentary peo-
ple (≈ 1300). The Fanger’s equation in its full form gives all human-related terms as a
function of the internal heat production, together with Ta, Tmrt, VP, v and the clothing insu-
lation Icl. Solving that equation provides the Predicted Mean Vote PMV defined as the cor-
responding thermal index. PMV indicates comfort when lying around zero (-0.5 to +0.5).
The deviation from zero was referred to as thermal stress and varies on a seven-point scale
from -3 (cold stress) to +3 (heat stress).
In the two-node model developed by Gagge et al. (1971, 1986), the empirical relationships
used for determining Tsk and SW are not related to the internal heat production, but ob-
tained by including some thermoregulatory processes of a human being (i.e. conduction
through body tissue and convection through blood flow). In fact, this model represents the
body as two concentric cylinders, where the first cylinder states for the body core and the
outer cylinder for the skin layer. Each of them is governed by one equation. The thermal
index new effective temperature (ET*) and standard effective temperature (SET*) are the
main outputs of this model (Table 2.1).
So far, the issue of comfort dealt with indoor environment. While trying to extend these
tools to outdoor conditions, it was necessary to incorporate the solar and terrestrial radia-
tion fluxes. These are decisive because extremely variable spatially as well as temporarily.
The output PMV of the Klima-Michel Model and the perceived temperature PT (Jendritzky
et al. 1990, Staiger et al. 1997) are as an extension of PMV to outdoor conditions. Simi-
larly OUT_SET* is an extension of SET*(Pickup and de Dear 1999). In addition, PET
calculated by MEMI (Munich Energy Model for Individuals) was especially developed for
outdoor environments (Höppe 1993, 1999); see Table 2.1 for definitions’ comparison.
The MEMI model follows the same approach as the Gagge two-node model. The human
thermoregulations considered include the constriction or dilation of the peripheral blood
vessels, sweating, and the production of heat by shivering. It also differentiates between
44
the core of the body, the skin layer and the clothing layer, so that the heat flow for covered
and uncovered parts of the body are calculated separately. Hence, two further heat flux
equations were added for evaluating the clothing surface temperature (Tcl), skin tempera-
ture (Tsk) and the core temperature (Tc) needed to solve the heat balance equation. The first
additional equation (2.4) describes the heat flow (Fcs in Wm-2) from body core to skin sur-
face and the equation 2.5 gives the heat flow transfer (Fsc in Wm-2) from the skin surface
through clothing layer to clothing surface as follow:
FCS = Vb ⋅ ρ b ⋅ cb ⋅ (Tc −T sk ) (2.4)
FSC = (1 / lcl ) ⋅ (Tsk − Tcl ) (2.5)
where Vb (ls-1m-2) is the blood flow from body core to skin, ρb (kgl-1) is the blood density
and cb (WsK-1kg-1) its specific heat. lcl (Km2W-1) is the heat resistance of the clothing.
The system of equations (2.3), (2.4) and (2.5) along with some thermo-physiological
parameterizations derived from the two-node model (for details see Höppe 1984, 1993 and
ASHRAE 2001a) enables the calculation of all relevant heat flows, as well as actual body
temperatures and sweat rate. MEMI provides as output the Physiologically Equivalent
Temperature PET and is calculated for standardized conditions (Table 2.1).
Comparing these three indices, a number of remarks could be made about on their limita-
tions:
PMV and OUT_SET* set Icl and the activity as variables, which means that the human
adaptive behaviour is included, whereas these are kept invariable in PET, meaning that
only the thermal environment is assessed. The two former indices are, hence, more
suitable than PET if calculations are directly confronted with subjective votes obtained
from social surveys, which must take into account the actual personal data.
PET and OUT_SET* were tested for identical hot outdoor conditions (data from Chap-
ters 5 and 6) and the same Icl and metabolic rate. OUT_SET* provided systematically
lower values, following a linear relationship: OUT_SET* = 0.73 PET + 3.1, with a
very high correlation coefficient R = 0.9998. In fact, OUT_SET* is about 27 % lower
because OUT_SET* considers a relative humidity RH = 50 % in the reference indoor
situation which is changing with Ta. This interdependence inhibits partly the assess-
ment of thermal stress, whereas PET considers a vapour pressure of 12 hPa which is a
constant water content in the air independent from Ta. Hence, this makes PET more ac-
curate than OUT_SET.
45
Theoretically, PET and OUT_SET have the advantage on PMV in that it takes into
account the thermoregulations of a human body and are therefore more accurate for ex-
treme conditions (typically outdoors).
The reliability of these indices is also discussed in the following pages. In this study PET
was used.
2.2.3. The mean radiant temperature
A critical issue in assessing the human comfort outdoors is the need for the mean radiant
temperature (Tmrt), which sums up all short-wave and long-wave radiation fluxes absorbed
by a human body. Tmrt is the key variable in evaluating thermal sensation outdoors under
sunny conditions regardless of the comfort index used (e.g. Mayer and Höppe 1987, Jen-
dritzky et al. 1990, Mayer 1993, Spagnolo and De Dear 2003). Tmrt is, per definition, the
uniform temperature of an imaginary black enclosure in which an occupant would ex-
change the same amount of radiant heat as in the actual non-uniform enclosure (ASHRAE
2001b). However, its calculation in outdoor spaces is not evident, particularly in complex
urban environments. This, certainly, explains the usual focus on air temperature and air
humidity in comfort related studies as these are easier to measure.
Theoretically, Tmrt applicable outdoors is given by the following formula (Fanger 1970):
0.25
  n α n
αk 
Tmrt = 1 / σ B  ∑ Ei Fi + k ∑ D F + f I  (2.6)
  i =1 εp i i
ε p p 
 i =1
where the surroundings are divided into n isothermal surfaces, for each one Ei (Wm-2) is
the long-wave radiation component (Ei = σB εi Ti4), Di (Wm-2) is the diffuse and diffusely
reflected short-wave radiation component, Fi is the angle weighting factor, I (Wm-2) is the
direct solar radiation impinging normal to the surface, fp is the surface projection factor
which is a function of the sun height and the body posture, αk is the absorption coefficient
of the irradiated body surface for short-wave radiation (≈ 0.7), εp is the emissivity of the
human body (≈ 0.97), and σB is the Stefan-Boltzmann constant (σB = 5.67 . 10 –8 Wm-2K-4).
The calculation of the angle factor Fi is the most problematic aspect when dividing the
environment into several surfaces. A procedure for calculating the angle factors is given by
Fanger (1970) for simple shapes, but the task becomes much more complicated for com-
plex urban forms and simplifications are thus necessary.
Several calculation procedures for Tmrt do exist, depending on whether it is modelled or
measured. One method, for instance, is to divide the human surroundings in two hemi-
46
spheres upwards and downwards and with the weighting factor Fi set to 0.5 for each of the
two directions (e.g. Jendritzky et al. 1990, Pickup and de Dear 1999). Although easier to
use, this method is probably only reliable for unobstructed open spaces. Obstruction effects
may be added if fish-eye photography is used to replace Fi (Watson and Johnsson 1988,
Chalfoun 2001). Yet, all surface temperatures as well as direct and diffuse short-wave ra-
diation components are still required.
To avoid such difficulties, the most suitable method would be to use an integral radiation
instrument. Such an instrument exists for indoor purposes, i.e. a globe thermometer (e.g.
Givoni 1976, ASHRAE 2001b). The globe thermometer consists of a hollow sphere (usu-
ally 15 cm in diameter), a flat black paint coating and a thermometer bulb at its centre. The
temperature assumed by the globe at equilibrium results from a balance between THE heat
gained or lost by radiation and convection. Empirical formulas derivate Tmrt from the globe
temperature Tg, together with Ta and v (Givoni 1976, ASHRAE 2001b). Alternatively, a
comfort index can be directly calculated, namely the Wet Bulb Globe Temperature WBGT,
usually used for assessing comfort at working spaces (Givoni 1976, ISO 1989, ASHRAE
2001b).
This instrument gives a good approximation of Tmrt indoors, where the heat irradiated from
the surroundings is rather uniform. However, the globe thermometer is less suitable out-
doors for several reasons, including the non-homogeneity of the radiant environment in-
duced by the additional solar beam radiation. Moreover, because of its spherical shape, the
globe thermometer may be well approximated for a seated person, as it averages the ab-
sorbed radiation equally from all directions, but not for a standing one for which the lateral
fluxes are dominant. Tmrt, integrally obtained, assumes equal energy absorption from a hu-
man body in both long-wave and short-wave range, and the black colour overestimates the
absorption of short-wave radiation, unless it is replaced by a grey globe more suitable to
describe normal clothing (ASHRAE 2001b). Finally, the globe thermometer is not conven-
ient because it needs a relatively long time to reach equilibrium (15-20 minutes). Alterna-
tively, one can use a smaller and light-coloured sphere for faster response of the instrument
(ASHRAE 2001b). Despite these disadvantages, it has been implemented for outdoors is-
sues, e.g. for workspaces outdoors (wet globe bulb temperature, WGBT) or even in social
surveys (Nikolopoulou et al. 2001, RUROS 2004). To date, there is no reliable instrument
for integral measurement of Tmrt outdoors, even though some attempts have been made
(e.g. Brown and Gillespie 1986, Krys and Brown 1990).
47
Another measuring technique of Tmrt (°C) has been proposed by Höppe (1992) including
all radiation fluxes, angle factors, human shape, etc. The surrounding environment is di-
vided into six main directions (upwards, downwards and the four lateral orientations) and
expressed by:
[ (
Tmrt = S rad / ε p ⋅ σ B )]0.25 − 273.2 (2.7)
with
6
S rad = ∑W (α
i =1
i k ⋅ K i + α l ⋅ Li ) (2.8)
where the related angle factors are the percentage of the hemisphere taken up by each part
of the body in each direction and expressed as a fraction (Wi), the short-wave (Ki in Wm-2)
and long-wave (Li in Wm-2) heat fluxes are summed as the mean radiation flux density
(Srad in Wm-2). Wi equals 0.22 for lateral directions and 0.06 upwards and downwards for a
standing body assumed to be cylindrical. Pyranometers and pyrgeometers, arranged in the
six directions are required for the measurement of the short-wave and long-wave radiation
fluxes, respectively. This method is accurate but costly and time-consuming, making it
difficult to implement in extensive measuring campaigns. Hence, the lack of an easy and
reliable method for determining Tmrt accounts for the main difficulty in conducting com-
prehensive investigations on comfort outdoors. Modelling Tmrt also requires simplifica-
tions. Surface temperatures are here an additional limitation, and are only accurately de-
termined if substrate and wall heat storage is included. The method used in ENVI-met re-
lies on sky view factors, and is detailed in section 3.7.
2.2.4. Methodological problems in assessing comfort outdoors
The assessment of comfort faces a number of methodological problems, including Tmrt

measurement mentioned above, the lack of validation of available indices for use outdoors,
the difficult interpretation in respect to actual people’ sensation, as well as the missing link
to urban geometry effects which are important in relation to design. These issues are dis-
cussed briefly in the following paragraphs.
Thermal indices applied indoors were extended to outdoors with the assumption that the
theory of comfort is also valid outdoors (Spagnolo and de Dear 2003). However, although
the thermal comfort indices, based on the human energy balance, are from a physical and
thermo-physiological point of view well founded, they are still facing the critical problem
of interpretation. In other words, what is precisely the meaning of an index value on a
48
given scale? A PMV value of + 3 or a PET value of 48°C, for example, can at most be in-
terpreted as heat stress, but nothing about the actual degree of discomfort can be drawn
with confidence, unless comparison with social surveys is carried out. Indeed, the differ-
ences between internal and external spaces are numerous: in typical indoor conditions, Tmrt
is almost equal to Ta; the air movement is weak and the activity mostly sedentary, while
outdoor conditions may experience larger differences in Tmrt in space and time, much
higher wind speeds and a different level of environmental stimulation. Indeed, the rele-
vance of exclusively thermo-physiologically based methods has been recently questioned,
and more social surveys based on questionnaires were undertaken, seeking to validate
those indices against actual people’s votes (e.g. Nikolopoulou et al. 2001, Spagnolo and de
Dear 2003, 2004). Although the general conditions and methods employed vary greatly,
making any comparison difficult, some common findings can be drawn.
The Nikolopoulou et al. (2001) survey was undertaken in a number of recreational loca-
tions in Cambridge, UK, at different seasons. The measurements included all relevant me-
teorological factors and a collection of about 1000 subjective people’s votes along with the
observation of their behaviour. It was assumed conditions of free choice for people to sit
outdoors but without changing their route on the basis of comfort. The subjective data were
compared to calculated PMV values. The authors confirmed that thermal environment is of
prime importance in influencing people’s use of these spaces, and the subjective response
to microclimate is subconscious leading to seasonal patterns of frequentation of outdoor
spaces. However, they claimed that a purely physiological approach is inadequate in char-
acterizing comfort outdoors, as the psychological adaptation is also found to be of great
importance. Available choice, environmental stimulation, thermal history of the person,
memory effect of recent weather conditions and expectations were all found to be decisive.
In a similar field study, Spagnolo and de Dear (2003) focused on the causal linkage be-
tween biophysical environments and subjective states of thermal comfort. They discussed
whether the standards applied indoors are also reliable outdoors and seek to determine the
range of a neutral comfort zone outdoors. All relevant meteorological data for comfort
were recorded and compared to 1018 subjective votes of people. The main finding was that
the thermal neutrality in terms of human comfort was significantly higher than for indoors
with OUT_SET* equals 26.2 °C versus 24 °C indoors (given by SET*). This agrees with
other studies (de Freitas 1985, Potter and de Dear 1999), which argued that people would
prefer slightly warmer conditions, corresponding to a positive value on the ASHRAE
seven-point scale, rather than theoretical neutrality.
49
Furthermore, Spagnolo and de Dear (2003) indicated that indoor standard comfort limits
are not directly transferable to outdoor environment. People’s expectations outdoors are
much more variable over space and time since they perceive their lack of control. This sug-
gests a significant widening of the comfort zone for outdoors, and consequently less dis-
comfort than usually interpreted. OUT_SET* ranging between 23 °C and 28 °C was found
to correspond to the zone of comfort for Sydney, and this is far above the standards
adopted indoors (Spagnolo and de Dear 2003). Recently, the lack of control has also been
cited to explain the larger tolerance of people in naturally ventilated versus air conditioned
buildings (Brager and de Dear 1998, Fanger 2004) and seems to corroborate for outdoors
this thesis of increased tolerance in case of evident lack of control.
A further interesting point handled in this study was to compare the most used comfort
indices (PMV, PET, OUT_SET*, PT, OP, ET*) when confronted to people subjective as-
sessments. The comparison between all these indices shows substantial discrepancies in the
assessment of comfort. PET and OUT-SET* seem to provide closest results (e.g. tempera-
ture of neutrality of 24.1°C for OUT_SET* vs. 23.4 °C with PET) whereas larger differ-
ences are found for PT or PMV. However, these results depended on the climate type: the
climate of Sydney shows small amplitudes and moderate air temperatures. This does not
necessarily reveal how these indices would vary if used for other climate conditions, e.g.
hot-dry. For example, a calculation of PET and OUT_SET* with the same inputs for ex-
treme hot conditions showed that OUT_SET* provides lower values (by 27 % less) be-
cause of different humidity assumptions.
The Nagara et al. (1996) study agrees with both studies above: The results given on a
seven-point scale revealed that thermal sensitivity of the subjects was affected by the his-
tory of exposure. An uncomfortable hot thermal sensation is registered when people moved
from air-conditioned spaces to sunlit open spaces. This calls attention to the issue of the
relevance of stationary vs. instationary models in assessing comfort outdoors.
Indeed, the history of exposure can be dealt with statistically by means of instationary
models (e.g. IMEM or Gagge two-node model). These are able to assess the evolution of
the human thermal sensation during a period of time by including the heat stored in the
body. Non-steady models can give additional information compared to steady-state models
in providing temporal courses of thermo-physiological parameters, e.g. skin and core tem-
perature (Höppe 2002). This can be valuable if various thermal conditions take place in a
restricted area, i.e. combination of shade and sunlit areas and also the transition from in-
door to outdoor milieu.
50
By means of IMEM, Höppe (2002) showed that the time needed by a human body to adjust
to outdoor thermal conditions lasts longer in the winter than in the summer, i.e. about sev-
eral hours for cold outdoor conditions against half an hour for hot conditions. This sup-
poses that a stationary assessment of human comfort is a good approximation for summer
conditions whereas a non-steady approach is more suitable in the winter.
With respect to planning, studies directly focusing on the consequences of urban design
strategies on comfort are dramatically lacking. A few studies relying on human-
biometeorological methods were undertaken within urban structures and highlighted the
major dependences of individual factors (Ta, VP, v, Tmrt) on thermal sensation outdoors
(e.g. Mayer and Höppe 1987, Jendritzky and Sievers 1989, Mayer 1993, 1998): PMV and
PET increases with the increase of Tmrt and Ta. Under typical hot and sunny summer condi-
tions, Tmrt is of prime importance in the thermal sensation. Tmrt shows a linear relationship
with strong correlation (R2 > 0.93) with either PET or PMV. In addition, the interdepend-
ence of Tmrt and the global irradiation G was as important (R2 > 0.92), with a clear distinc-
tion between irradiated and shaded areas was observed. This points out the usefulness of
shading (by buildings or trees) in maintaining comfort. The dominant impact of Tmrt dimin-
ishes, however, for transitional seasons when the solar radiation is lower. Moreover, a sta-
tistical regression between Ta and PET revealed an exponential relationship, however, with
less good correlation than that observed for Tmrt, certainly because PET experiences con-
trasting values between exposed and shaded locations. Vapour pressure VP fluctuations
were found to have insignificant impact on PET. Increasing the wind speed v leads to a
decrease of PET, yet no strong relationship could be found.
2.2.5. Effects of urban design on comfort outdoors
A number of urban design experiences illustrate a real concern and consciousness in de-
signing with the climate, either by taking advantage from the potential of natural energy or
by protecting the living spaces from adverse climatic conditions. These can be verified
through history (Ali-Toudert 2000), in the traditional built heritage (e.g. Knowles 1981,
Ravérau 1981, Golany 1982, Krishan 1996) as well as in contemporary urban projects (e.g.
Herzog 1996, Asimakopoulos et al. 2001, Hawkes and Forster 2002, Thomas 2003). Few
examples are shown in Figs. 2.9 to 2.11(see also Fig. 6.2). Many of these arrangements
deal directly with the street geometry and confirm its structural role.
From these examples, one can observe several common features:
51
- The street as climate regulator is one aspect within a whole urban design methodology
(see also Ali-Toudert et al. 2005),
- comfort outdoors and comfort indoors are simultaneously considered, so that winter
and summer needs are satisfied,
- The aspect ratio and solar orientation are basic describers of a street microclimate, but
details in the design of the street are essential. These include the use of galleries, vege-
tation, “textured” or self-shading façades and the use of different building heights for a
better seasonal solar control.
These solutions are based on theory or long-term experience. Surprisingly, very few stud-
ies investigated their climatic efficiency quantitatively (e.g. Pearlmutter et al. 1999, Grund-
ström et al. 2003). Moreover, investigations based on actual scientific methods, which
prove the efficiency of commonly used street design concepts on outdoor thermal comfort,
are also lacking. Therefore, the current knowledge on the subject is still mainly qualitative.
Available studies are reported and discussed below.
Swaid et al. (1993) carried out one of the first investigations on outdoor thermal comfort
directly related to street design. They considered street canyons with an aspect ratio H/W
of 0.5 and 1, oriented N-S and E-W, and located in the Mediterranean climate in Tel Aviv
(32°N). The authors found that the comfort conditions are more sensitive to H/W than to
street orientation. Comfort is better for H/W = 0.5 than for 1. N-S streets are always closer
to comfort irrespective of the canyons’ aspect ratios all the day round. However, H/W = 1
is warmer in the night-time. E-W streets are uncomfortable between 14:00 and 18:00 LST
for H/W = 0.5 and all the day for H/W = 1. This was attributed to shading from the walls,
which, according to the sun course, is more efficient for a N-S orientation than for an E-W
orientation. Air temperature in an E-W street including a gallery is found to be lower than
in a street without gallery. The thermal situation is close to the comfort limit but it does not
reach it except between 17:00 and 20:00 LST. Nocturnal discomfort was reported, and was
attributed to weak winds.
The methodology used in this study was based on the only prediction of the urban air tem-
peratures by means of the empirical CTTC model, together with a simple linear relation-
ship for the prediction of the wind speed reduction within a canyon. Comfort was assessed
using the index ITS. The study assumes shade within the street (i.e. Tmrt = Ta), although
ITS, per definition, assesses the combined effect of metabolic rate, environmental condi-
tions including solar radiation, and clothing on physiological strain (Table 2.1).
52
Fig. 2.9. bedZED project showing an E-W asymmetrical street shape for ensuring solar
access, together with using galleries and vegetation for outdoor comfort (Thomas 2003)
Fig. 2.10. Solar control through self-shading façade in a hot-dry climate (Krishan 1996)
Fig. 2.11. Housing quarter of Linz-Pichling, Austria showing the link between urban and
architectural concepts in relation to climate (Herzog 1996)
53
This is due to the relatively wide streets considered (H/W = 0.5 and 1) and the subtropical
latitude of the site at which the sun position in summer is high (see Arnfield 1990a, Bour-
bia and Awbi 2004).
Hence, the thermal sensation in sunlit locations is not really investigated. This assumption
probably explains that the street H/W = 0.5 was found to be more comfortable than H/W =
1, since the wind speed is higher in the former case (caused by sea breezes in the after-
noon) and leads to more cooling. The shade provided by the buildings only seems to be
insufficient for ensuring the comfort required. The authors advised the use of additional
shading devices to reduce the heat stress, either by planting trees or by means of arcades on
the sidewalks.
Pearlmutter et al. (1999) performed the first investigation which focused on the radiation
fluxes within urban canyons along with their impact on a human body. Full-scale meas-
urements were conducted in the arid Negev region in two low-rise urban street canyons
(H/W = 1), with different orientations, at the centre of the street and on the roof.
The canyon is described as a potential “cool island” due to solar shading from direct, dif-
fuse and reflected radiation. A pedestrian gained less radiant heat in comparison to a per-
son standing on the unobstructed roof. However, the absolute dimensions of the street in
respect to human size (H = W = 3 m) are logically responsible for the large shading advan-
tages and this may differ in larger canyons. Moreover, no explicit information is provided
about the actual thermal sensation in the canyon. The orientation is found to be important
regarding the ventilation potential during the late afternoon and in the evening. The micro-
scale heating effect in the canyon is found to be a nocturnal phenomenon. In the winter, the
street provides relatively warmer conditions owing to the protection from the strong cold
winds.
Furthermore, Grundström et al. (2003) conducted comparative measurements in a Saharan
location between a traditional desert city and a nearby modern neighbourhood in streets
with H/W = 10 and 0.5, respectively. In both summer and winter, the minimum air tem-
perature is found to be 2 to 4 K lower in the modern quarter than in the traditional one. The
maximum air temperature is 10 K higher in the modern structure. In the summer, the street
in the modern site is extremely uncomfortable whereas the traditional one was in the com-
fort zone. However, only Ta and RH were taken into account in the comfort assessment.
The radiant fluxes were neglected by setting Tmrt equal to Ta. This is particularly question-
able for the wide street, which is most of the time sunlit. In the winter, none of the two
areas achieved comfort but the large canyon was better.
54
Similarly, Coronel and Alvarez (2001) studied the thermal properties in the summer of
confined urban spaces in Santa Cruz, Spain. They found that confining and reducing the
dimensions of a street is very important in the final thermal behaviour of these spaces,
which was compared to an oasis effect. Air temperature decrease of 8 K for narrow streets
(H/W = 5) in summer was recorded. This was explained by the reduced solar access, use of
white colours and to the weak anthropogenic heat generation. Thermal driven air move-
ment by night was also very important for the nocturnal cooling. Massive walls reduced
the night-day oscillations.
2.3. Conclusion
This chapter presented an overview on the available knowledge on the urban canyon mi-
croclimate, on outdoor thermal comfort methods and on the dependence of comfort upon
the urban structure. The existing studies provided some basic knowledge on the energetic,
thermal and wind flow characteristics of an urban canyon. The methods for assessing com-
fort outdoors are numerous, and basically relying on the same principles. These are either
energy balance based or rather focusing on adaptive behaviour of people. Both methods are
expected to be complementarily used in the next years. By contrast, the relationship be-
tween urban geometry and comfort is by far less well understood. Design concepts known
from the practice and collected throughout history for managing the climate dimension in
architecture and urban design is recognized. Yet, the quantitative assessment of these solu-
tions is lacking, or performed with weak methods. These statements are the main motiva-
tion of the following investigation, which aims at providing more knowledge on the de-
pendence of comfort on design choices.
55
56
3. The numerical model ENVI-met 3.0
This chapter presents a brief review on the state of research development regarding
modelling of urban microclimates, followed by a summary of the main features of the
model ENVI-met 3.0 used as the main investigation method in the present work.
3.1. Numerical modelling of the urban microclimate
The use of numerical methods for urban climate issues has a distinct advantage over
comprehensive field measurements. Their versatility in dealing with the manifold vari-
ables and atmospheric processes make them increasingly popular (Arnfield 2003).
Urban climate models can be first classified according to their scale, which can range
from kilometres to few centimetres. Usually, models developed for urban climate pur-
poses (UHI) use a large space resolution (e.g. Gross 1991, Masson 2000). These are
probably more suitable for urban planning issues (scale up to 1/5000) rather than for
urban design issues (∼ 1/500). The following review addresses the microclimatic nu-
merical models from the latter category, which are comparable in scale and to some
extent in task with ENVI-met.
Urban microclimate models vary substantially according to their physical basis and their
temporal and spatial resolution. At the microscale, three-dimensional (3D) wind flow
models are the most well founded (e.g. Eichorn 1989, Johnsson and Hunter 1995), while
those including all hydrological, thermal and energy processes are very few, inter-alia
because very time-consuming. Such models are often simplified by assuming several
parameterizations and limitations in order to save time and solve problems linked to
variables difficult to determine. Typically, these models use simplified turbulence
schemes (e.g. Mills 1993, Arnfield 2000). Urban canyon models are also typical exam-
ples: 2D rather than 3D, they focus on the energy fluxes prognosis and assume prede-
fined street configurations, with buildings of uniform shape and height, dry surfaces, no
57
vegetation (no latent heat) and no heat storage in the building fabric (e.g. Herbert et al.
1998). Alternatively, models which combine 3D flow modelling and 2D energy model-
ling are faster and more accurate (e.g. Arnfield et al. 1998). Other models are more em-
pirical and are based on equations derived from few available measured data, which
may make them context specific, e.g. Nunez and Oke (1980) or the CTTC model
(Swaid and Hoffman 1990, Shashua-Bar and Hoffman 2000). Moreover, many of these
models deal with the urban canyon volume as a whole, i.e. all calculations are made for
one point at street level, and spatial differences within a canyon are not considered. By
contrast, CAD-based models seek to reproduce with precision the 3D urban scene, as
these models are especially dedicated to designers (e.g. Teller and Azar 2001, Asawa et
al. 2004) and possibly assess the interdependence between indoors and outdoors in
terms of daylight and sunlight availability on the urban surfaces, e.g. SOLENE (Groleau
and Miguet 1998). The focus in this case is the calculation of the surface temperatures
and mean radiant temperatures. Yet, most of the weather data (wind speed, Ta, etc.) are
assumed to be known.
Furthermore, very few microclimate models assess the resulting thermal comfort in ad-
dition to the urban microclimate changes (Teller and Azar 2001, Asawa et al. 2004).
This is mainly due to the problematic determination of the radiation fluxes from the
surroundings of a human body in complex urban areas. The issue of modelling outdoor
thermal comfort is thus often dealt with using simplified and averaged methods, in
which many atmospheric processes are removed. These are then replaced by data set as
inputs by the user, which assumes their availability (e.g. daily data of v, Ta, RH). For
instance, thermal comfort in the model TOWNSCOPE (Teller and Azar 2001) is calcu-
lated on a daily basis, however, with Ta, v, RH, and Ts assumed as mean daily average
and held constant during simulation.
Finally, a decisive aspect in choosing a model is the output information. The outputs
may vary from only one variable prognosis, e.g. Ta (Swaid and Hoffman 1990), to a
detailed microclimate description, e.g. ENVI-met (Bruse 1999).
3.2. Relevance of ENVI-met to the present study
In this study, the three-dimensional model ENVI-met, version 3.0 was used (Bruse
1999). The major advantage of ENVI-met is that it is one of the first models that seeks
to reproduce the major processes in the atmosphere that affect the microclimate on a
58
well-founded physical basis (i.e. the fundamental laws of fluid dynamics and thermody-
namics). According to the objectives of the present work, ENVI-met presents several
advantages:
1. ENVI-met simulates the microclimatic dynamics within a daily cycle. The model is
in-stationary and non-hydrostatic and prognoses all exchange processes including
wind flow, turbulence, radiation fluxes, temperature and humidity.
2. A detailed representation of complex urban structures is possible, i.e. buildings with
various shapes and heights or design details like galleries and irregular geometrical
forms, particularly relevant for the present work. The vegetation is handled not only
as a porous obstacle to wind and solar radiation, but also by including the physio-
logical processes of evapotranspiration and photosynthesis. Various types of vegeta-
tion with specific properties can be used. The soil is also considered as a volume
composed of several layers and the ground can be of various types.
3. The high spatial resolution (up to 0.5 m horizontally) and the high temporal resolu-
tion (up to 10 s) allow a fine reading of the microclimatic changes, especially sen-
sible to urban geometry and pertinent for comfort issues.
4. The model requires a limited number of inputs and provides a large number of out-
puts.
5. The key variable for outdoor comfort, i.e. mean radiant temperature Tmrt, is also
calculated.
3.3. General structure of ENVI-met 3.0
Fig. 3.1 shows the construction scheme of ENVI-met, which is composed of a 3D core
model (including atmospheric, vegetation and soil sub-models) and 1D border model.
The task of the 3D model is to simulate all processes inside the actual model area. The
upper horizontal boundary and the vertical windward boundary act as interface of the
1D border model and the 3D core model. The 1D model extends the simulated area to
the height H = 2500 m (i.e. an average depth of a boundary layer) and transfers all start
values to the upper limits of the 3D volume needed for the actual simulation.
The core area to be simulated is a volume of the dimensions (X, Y, Z) plotted into n
grid modules. Z is determined by the maximum height Hmax of the urban elements
within the model (Z ≥ 2Hmax). Each module (∆x, ∆y, ∆z) can either be a part of a build-
ing, of vegetation, or of an open space (e.g. street) and possible oblique urban forms
59
have to be approximated in steps. At street level, the first grid is vertically subdivided
into five equal parts in order to record thoroughly the microclimate near the surface.
The soil model provides the system with the surface temperatures and humidity. The
soil model is 1D, except the grids of the ground surface which are connected in 3D for
ensuring homogeneity. The nesting grids consist on a “buffer zone”, which acts as an
offset of the actual edges of the model area in order to avoid numerical disturbances, i.e.
boundary effects. The nesting grids also ensure a representative 3D profile of the wind
at the windward boundary by adjusting the initial 1D wind profile. These grids get pro-
gressively larger as their distance from the core model increases and are composed of
two soils types. The nesting area extends at least to the double of the highest obstacles
in the model area (2Hmax) beyond the actual modelled area.
The equations that govern ENVI-met are too numerous to be presented thoroughly here.
Only, the main important are reported below, since the model is well documented
(Bruse and Fleer 1998, Bruse 1999) and is also regularly updated and a freeware (Bruse
2004).
at H=2500m, u , v , θ q inputs, constant

H = 2500 m E, ε from 1D boundary model
O
I U
N T
F F
L L
O O
W u , v , E, ε, θ , q W
Z≈2Hmax
u , v , E, ε,
( )
∂ θ ,q / ∂x = 0 ∂(u,v,E, ε ) / ∂x = 0 z
y
Hmax
x
Nesting grids 1D soil model

T0 and q0 for z = 0, u = v = E = ε = 0 3 layers
provided by 3Dmodel
H = -1.75 m
Fig.3.1. General schema of the ENVI-met model including the boundaries (symbols see
text)
60
3.4. The atmospheric model
The atmospheric model prognoses the evolution of the wind flow (speed and direction),
turbulence, temperature, humidity, short-wave and long-wave radiations fluxes. It is
based on the fundamental laws of dynamics and thermodynamics of fluids, i.e. equa-
tions of conservation of mass, momentum, heat and moisture (e.g. Garrat 1992).
3.4.1. Mean air flow
The spatial and temporal evolution of the turbulent wind flow is described by the Na-
vier-Stokes equations. In ENVI-met, the non-hydrostatic incompressible form is used:
∂u ∂u ∂p ′  ∂ 2u 
+ ui =− + K m  2  + f (v − v g ) − S u ( x, y, z ) (3.1a)
∂t ∂xi ∂x  ∂xi 
∂v ∂v ∂p ′  ∂ 2v 
+ ui =− + K m  2  + f (u − u g ) − S v ( x, y , z ) (3.1b)
∂t ∂xi ∂y  ∂x i 
∂w ∂w ∂p′  ∂2w  θ (z )
+ ui =− + K m  2  + g − S z ( x, y , z ) (3.1c)
∂t ∂xi ∂z  ∂xi  θ ref ( z )
∂u ∂v ∂w
+ + =0 (3.2)
∂x ∂y ∂z
where f (= 10-4 s-1) is the Coriolis parameter, p´ is the local pressure perturbation, Km the
exchange coefficient and θ the potential temperature at height z. The nomenclatures ui
and xi correspond to u, v, w and to x, y, z with i = 1, 2, 3, respectively. The modelled
area is relatively of small extent and the Boussinesq-Approximation (i.e. ρ = ρ O = 1.29
kgm-3) can be applied to remove the air density ρ from the original compressible Na-
vier-Stokes equations, analytically difficult to determine. The continuity equation (3.2)
is, hence, added to make the model mass conserving as well as a term for thermal forced
vertical motion in the w-equation (3.1c). Su, Sv and Sz are added as local source/sink
terms which describe the loss of wind speed due to drag forces induced by possible
vegetation elements and are given by (Yamada 1982, Liu et al. 1996):
∂p′
S ui = = cd , f LAD(z ) ⋅W ⋅ ui (3.3)
∂xi
61
where W = (u2+v2+z2)0.5 is the mean wind speed at a height z, LAD is the leaf area den-
sity and informs on the porosity of the plant. cd,f is the mechanical drag coefficient and
is usually set to 0.2.
3.4.2. Temperature and humidity
The distribution of the potential temperature θ and the specific humidity q inside the
atmosphere is given by the combined advection-diffusion equation with internal
source/sink terms:
∂θ ∂θ  ∂ 2θ  1 ∂Rn ,lw
+ ui = K h  2  + + Qh (3.4)
∂t ∂xi  ∂xi  c p ρ ∂z
∂q ∂q  ∂ 2q 
+ ui = K q  2  + Qq (3.5)
∂t ∂xi  ∂xi 
where Qh and Qq are used to link heat and vapour exchanges between the plant surface
and the surrounding air. These quantities are provided by the vegetation model. Kh and
Kq are the diffusion coefficients for heat and vapour. The vertical divergence of long-
wave radiation ∂Rn ,lw / ∂z accounts for cooling and heating effects of radiative fluxes.
3.4.3. Atmospheric turbulence
To solve the basic equations given above, modelling faces the necessity of determining
the turbulent processes. These can not be described numerically and have to be deter-
mined as approximated values from definable quantities (closure problem).
Several turbulence closure solutions exist (e.g. Garrat 1992). The closure of first-order
(also called K theory) bases on the exchange coefficients K and the gradient of each
quantity in all directions. This method is very convenient and widely used. However, it
is only reliable for homogenous terrain because it does not integrate the effects of obsta-
cles such as buildings or vegetation, which are typical in urban environments. The clo-
sure of second-order (or more) is too complex to be applicable in numerical modelling
because of the time processing it implies. A compromise solution between both is the E-
ε model (also called 1.5-order model). The E-ε model allows the simulation of advec-
tion processes as well as the incorporation of the influence of the horizontal non-
62
homogeneity. This makes it suitable for urban context and was, therefore, adopted in
ENVI-met (Bruse 1999).
In the E-ε model, two further variables are added to determine the exchange coeffi-
cients, namely the turbulent kinetic energy E and its dissipation ε. These equations are
given by Mellor and Yamada (1975) as follows:
∂E ∂E  ∂2E 
+ ui = K E  2  + Pr + Th + QE − ε (3.6)
∂t ∂xi  ∂xi 
∂ε ∂ε  ∂ 2ε  ε ε ε2
+ ui = K ε  2  + c1 Pr + c3 Th − c2 + Qε (3.7)
∂t ∂xi ∂x
 i  E E E
where c1, c2 and c3 are standard values taken from Launder and Spalding (1974) but dif-
ferent values might be used for special conditions. The terms Pr and Th describe respec-
tively the production and dissipation of turbulent energy due to wind shearing and ther-
mal stratification (buoyancy production) and are given by:
 ∂u ∂u j  ∂ui
Pr = K m  i +  (3.8)
 ∂x  ∂x
 j ∂xi  j
g ∂θ
Th = Kh (3.9)
θ ref ( z ) ∂z
θref(z) is the reference potential temperature at the inflow boundary and g is the accelera-
tion due to gravitation (= 9.81 ms-2). Th is usually neglected under stable conditions.
QE and Qε account for the additional turbulence produced by the vegetation as well as
the accelerated cascade of turbulent energy from large to small scales near plant foliage.
These are expressed according to Liu et al. (1996) and Wilson (1988) as follows:
QE = cd , f LAD ( z ) ⋅ W 3 − 4cd , f LAD ( z ) ⋅ W ⋅ E (3.10)
Qε = 1.5cd , f LAD ( z ) ⋅ W 3 − 6cd , f LAD ( z ) ⋅ W ⋅ ε (3.11)
where cd,f is a drag coefficient at the plant foliage and W the wind speed at the consid-
ered height and ε in the latter equation is found by the application of the Kolmogorov
relation (ε = 0.16 E3/2/l). Once the E-ε field is calculated, the exchange coefficients Km,
Kh and kq are calculated with the assumption of isotropy of the local turbulence:
E2
K m = cµ (3.12a)
ε
K h = K q = 1.35 K m (3.12b)
63
Km
KE = (3.12c)
σE
Km
Kε = (3.12d)
σε
with cµ = 0.09, σE = 1 and σε =1.3. Scaling functions are additionally used to adjust
theses diffusion coefficients to thermal stratification according to Sievers et al. (1987)
and Businger et al. (1971).
One disadvantage, however, of the E-ε model is that it tends to overestimate the turbu-
lence in higher atmosphere layers. Therefore, ENVI-met offers the option of using the
closure of first-order (gradient equations) in case of a homogenous model area.
The exchange coefficients between the ground or building surfaces and the air, i.e. first
grid point next to the surface, are not calculated by the E-ε model but with empirical
formulations based on the physical state of the air close to the surface and the surface
itself.
The wind field (wind shear) and the thermal forces (buoyancy production) are here de-
cisive in the nature of turbulent exchange, i.e. free convection or molecular exchange,
and can be expressed by the Bulk-Richardson number:
g ∆θ ⋅ ∆w
Rib = (3.13)
θ (∆u )2
where ∆w is the distance between the surface and the first grid of air next to it and u the
horizontal wind speed.
The turbulent fluxes of momentum, heat and vapour at surfaces are then calculated after
the Monin-Obukhov similarity law which states that the fluxes in the surface layer are
highly constant, and expressed as a function of the scaling quantities u*, θ* and q* ( e.g.
Stull 1988, Garrat 1992):
The momentum flux: (u*)2
The turbulent heat flux: u* · θ*
The turbulent moisture/water vapour flux: u* ·q*
3.4.4. Radiation fluxes
The atmospheric long-wave radiation depends on air temperature, as well as on absorp-

tion and emission coefficients for each single air layer. The actual absorption and emis-
64
sion rate of air depends on the water content but also on gases like carbon dioxide CO2
and ozone O3. Yet, only absorption due to water (i.e. VP) is taken into account (Pal-
tridge and Platt 1976, Gross 1991) because of the complex absorptive relationships as
well as the lacking information about the vertical distribution of carbon dioxide CO2
and ozone O3. Hence, the long-wave atmospheric radiation at a height z, if not modified
by vegetation, can be approximated after integration for n single layers (Paltridge and
Platt 1976) by:
N
↓
Rlw (z)= ∑σ
n =1
BT
4
( n )[ε n ( l + ∆l ) − ε n ( l )] (3.14)
where l is the water content in the layer between the height z and the lower layer n, εn is
the emissivity of a layer n and T is the absolute temperature.
The short-wave radiation fluxes at the model boundary R*sw is calculated with the inte-
gration of the radiation intensity of the sun I0 in the wavelength range of λ = 0.29 to λ =
4.0.
4.0
R*sw = ∫ I 0 (λ ) exp{− α R (λ )m + α M (λ )m}dλ (3.15a)
0.29
I0 is available from tables (Houghton 1977). The optical mass m is function of the solar
height h, the Rayleigh scattering ( i.e. αR = 0.00816 λ-4 ) and Mies scattering (αM = λ-
1.3
βtr). The absolute amount of direct short-wave radiation at the model boundary Rsw
0
,dir
is obtained after the deduction of the energy quantity absorbed Rsw,abs by the water con-
tained in the atmosphere after Liljequist (1979), namely:
,dir = R sw − R sw ,abs = R sw − (70 + 2.8VP2 m ⋅ m )

0 * *
R sw (3.16a)
0
The short-wave diffuse radiation Rsw ,dif for cloudless sky conditions depends on the di-
rect solar radiation flux and the sun height φ and is estimated after Brown and Isfält,
(1974):
(
R 0sw ,dif = f R 0sw ,dir , φ ) (3.16b)
0
For cloudy sky conditions, the direct solar radiation Rsw ,dir is reduced according to
Taesler and Anderson (1984).

The radiation fluxes are strongly modified in the model area by obstructing buildings
and vegetation. A number of coefficients (σ(…)) are introduced to include these effects
and range from one for undisturbed fluxes to zero for totally obstructed fluxes.
65
For a given grid point in the model area, if a building blocks the direct solar radiation
then σ sw,dir is set equal to zero. For all other radiation fluxes, the partial obstruction of a
solar radiation depends on the proportion of the sky “viewed” by a surface and is given
by the sky view factor σsvf :
1 360
σ svf = ∑ cos ω (π )
360 π =0
(3.17)
where ω is the vertical angle determined by an obstacle at the azimuth angle π.

Obstructions coefficients due to plants for direct and diffuse short-wave radiation
(σsw,dir, σsw,dif) as well as atmospheric and terrestrial long-wave radiation ( σ lw↓ and σ lw↑ )
are expressed as follows:

σ sw,dir ( z ) = exp(− F ⋅ LAI * ( z )) (3.18a)
σ sw,dif ( z ) = exp(− F ⋅ LAI (z, z p )) (3.18b)
σ lw↓ ( z ) = exp(− F ⋅ LAI (z, z p )) (3.18c)
σ lw↑ ( z ) = exp(− F ⋅ LAI (0, z )) (3.18d)

where F is the extinction coefficient and LAI is the 1D leaf are index which is obtained
by the vertical integration of LAD over the height of the plant. For short-wave solar ra-
diation LAI is replaced by LAI*, which is “3D” i.e. calculated with respect to the angle
of incidence from the incoming sun’s rays.
The direct short-wave radiation Rsw ,dir at any point z is then given by:
Rsw ,dir ( z ) = σ sw ,dir ( z )Rsw

0
,dir (3.19)
and the total diffuse radiation (including the diffusely reflected radiation) Rsw,dif is given
by:
Rsw,dif ( z ) = σ sw,dif (z )σ svf ( z )Rsw ,dif + (1 − σ svf ( z ))Rsw ,dir ⋅ a
0 0
(3.20)
Where the first term corresponds to the diffuse radiation, assumed isotropic scattered
and is weighted by the sky view factor. The second term is the diffusely reflected radia-
tion with a an averaged albedo for all walls and ground surfaces in the model area.
In the presence of vegetation, the long-wave radiation fluxes Rlw↓ ( z ) upwards and
Rlw↑ ( z ) downwards are then expressed by:
( )
Rlw↓ ( z ) = σ lw↓ ( z )Rlw↓ ,0 + 1 − σ lw↓ ( z ) ε f σ B T f +
4
(3.21)
66
( )
Rlw↑ ( z ) = σ lw↑ ( z )ε sσ BT04 + 1 − σ lw↑ ( z ) ε f σ B T f −
4
(3.22)
4 4
where T f + and T f − are the average foliage temperature of the overlying (+) and underly-
ing (-) vegetation layer; T0 is the ground temperature and Tw is the average surface tem-
perature of buildings walls; σB is the Stefan-Boltzman constant; εf, εs and εw are the
emissivities of foliage, ground surface and walls, respectively.
3.4.5. The ground and buildings surfaces
The ground surface temperature T0 is calculated by solving the energy balance of the
surface:
Rsw ,net + Rlw ,net − G0 − H 0 − LE0 = 0
(3.23)
where Rsw ,net is the net short-wave radiation received by the surface, Rlw ,net is the net
long-wave radiation, G is the soil heat flux, H0 and LE0 are the sensible and latent turbu-
lent heat flux, respectively. The calculation of Rlw ,net is complex and includes the effects
of buildings and vegetation. The long-wave radiation fluxes from the buildings are ap-
proximated in one quantity Rlw↔ which is calculated on the basis of an averaged surface
temperature of all walls Tw :
Rlw↔ (z ) = (1 − σ svf ( z ))ε wσ B T w

4
(3.24)
The turbulent fluxes H0 and LE0 are function of the calculated turbulent exchange coef-
ficients and the air temperature and humidity of both ground surface (z = 0) and the first
grid point vertically (z = 1). The molecular flux G and the turbulent and molecular en-
ergy fluxes at the ground surface are given by:
∂T T − Tk =−1
G0 = λs = λs (k = −1) 0 (3.25)
∂z 0.5∆z k =−1
 ∂T   0 T0 − θ k =1 
H 0 = ρc p .− K h0  = ρc P  K h  (3.26)
 ∂z z =0   0.5 ∆z k =1 
 ∂q   q − qk =1 
LE0 = ρ .L0 − K q0  = ρ ⋅ L0  K q0 0  (3.27a)
 ∂z   0.5 ∆z k =1 
with
67
L0 = (5.501 − 0.00237 (T0 − 273.13 ))10 6 (3.27b)
where λs is the soil heat conductivity, k = ± 1 corresponds to the first grid point over or
under the ground surface, K h0 and K q0 are the exchange coefficient for heat and vapour
between air and surface, calculated with respect to thermal stratification (Asaeda and Ca
1993).
The presence of buildings plays obviously an important role in the energy budget in an
urban area. The wall and roof temperatures are calculated by solving the energy balance
equation for each surface respectively. However, the heat storage in the building materi-
als is not taken into account.
Similar to the ground surface, the energy balance of a wall or roof surface is given by:
Rsw ,net + Rlww ,,rnet − H w ,r − Qw ,r = 0 (3.28)
where Hw and Qw,r are the turbulent sensible heat flux and the heat flux through the roof
or wall, respectively. The net short-wave radiation flux Rkw,net is given by:
{ }
Rsw ,net = cos β * ⋅ Rkw ,dir (z w ,r ) + Rkw,dif (z w ,r ) (1 − as )
(3.29)
where z w,r is the surface height. The albedo as is defined for the wall and for the roof. as
is calculated for natural materials and predefined in case of waterproof materials. β* is

the angle between the incident direct solar beam Rkw,dir and the normal to the surface
(Lambert cosine law).
For the net long-wave radiation, a differentiation is made between roof and wall sur-
faces, because of their different orientation (i.e. horizontal and vertical). Hence, for a
roof surface, the net long-wave radiation flux Rlwr ,net is given by:
Rlwr ,net = σ svf Rlw↓ ,0 + (1 − σ svf )ε wσ B T r − ε rσ B T r

4 4
(3.30)
The vertical wall surfaces receive additionally a part of radiant heat from the ground.
Here, some assumptions are made for the façades: Explicitly, it is set that one half of the
heat originates from the ground and one half from the sky in case of visible sky, and
one-third of the emitted heat is assumed coming from the ground and two-third from the
façades in case of sky obstruction.
The net long-wave radiation flux Rlww ,net for a wall surface is given by:
Rlww ,net =
( 4
) ( 4 4
)
σ svf 0.5ε sσ B T 0 + 0.5 Rlw↓ ,0 + (1 − σ svf ) 0.33ε sσ B T 0 + 0.67ε wσ B T w − ε wσ B T w
4
(3.31)
68
The sensible heat flux Hw from the wall surface to the air close to it (i.e. w and w+1
grids) is given by:
∂T Tw − Tw+1
H w = ρc p K hw = ρc p K hw (3.32)
∂xi w
∆w
It depends on the exchange coefficient at the wall and the gradient between the surface
temperature and the air temperature at the first grid point next to the wall surface. The
heat flux through the wall Qw,r can be directly calculated from the wall or roof surface
temperature and the internal air temperature with:
Qw ,r = U (Tw − Ta ,i ) (3.33)
where U is the heat transmittance of the building material. Standard U-values for vari-
ous materials are available in the literature (e.g. Koenigsberger et al. 1973, Markus and
Morris 1980).
3.5. The soil model
This sub-model calculates the temperature and humidity at the ground surface and un-
derground up to a depth of -1.75 m for the whole model area. The soil model takes into
account the hydrological and thermo-dynamical processes which vary according to the
individual soil properties. It consists of a one dimensional vertical profile with 14 nodes.
Beyond this level, the daily variations of soil temperature and humidity are assumed
negligible. For each grid point, a soil structure can be chosen. This is composed of three
layers (0- 20 cm, 20-45 cm, and 45-175 cm) and each layer corresponds to a soil type.
Numerous natural and artificial soil types with specific hydrological and thermo-
dynamical properties are available in a database (see Appendix A1 and A2) based on
Clapp and Hornberger (1978). More soils can be added (e.g. “as” and “s2” in Appendix
A1 and A2). A 1D calculation is used for non equidistant nest grids, except for the up-
per surface, at which a 3D calculation of the heat exchange is used to get homogenous
surface temperatures. For waterproof materials the heat conductivity is directly avail-
able in the databank describing the materials. No water content is calculated and no
source term is considered since no exchange occurs. For natural soil types the heat con-
ductivity and albedo are calculated.
69
3.6. The vegetation model
The reference to vegetation in the previous pages showed that the plants in ENVI-met
are more than physical obstacles against wind and radiation. They are biological bodies
which interact with the surrounding environment by exchanging heat and water vapour.
The vegetation is schematized as a 1D column with height zp and a root depth of -zr..
This “column” consists of ten equally distant nodes above the ground as well as in the
root part. To each layer correspond a leaf area density (LAD) and a root area density
(RAD). This scheme is used for all vegetation types ranging from small green cover like
grass to tall trees. Physiological parameters are also needed, such as the stomata-
resistance, the nature of the plant (deciduous or evergreen) and the albedo of the foliage.
The emissivity of the foliage is kept constant. All this information is summarized in a
database (Appendix B), which is expandable to more species.
The exchanges between the plant and its surrounding air consist on direct heat flux Jf,h,
evaporation flux Jf,evap and transpiration flux Jf,trans and are given by:
J f ,h = 1.1ra−1 (T f − Ta ) (3.34a)
J f ,evap = ra−1 ∆qδ c f w + ra−1 (1 − δ c )∆q (3.34b)
J f ,trans = δ c (ra + rs ) (1 − f w )∆q

−1
(3.34c)
where Tf is the leaf temperature and ra the aerodynamic resistance of the leaf. ∆q is the
leaf-to-air humidity deficit. δc is a factor set at 1 if evaporation and transpiration can
occur (∆q ≥ 0), otherwise δc is set to zero for only possible condensation. fw is the frac-
tion of wet leaves. rs is the stomatal resistance which depends on short-wave irradiance
input and available soil water, and is calculated after Deardoff (1978) or alternatively
after Jacobs (1994) which is a more dynamic description including the photosynthesis
process.
The energy balance of the leaf enables the calculation of the leaf temperature and is
expressed by:
Rsw,net + Rlw,net = c p ρJ f ,h + ρL(J f ,evap + J f ,trans ) (3.35)
where L is the latent heat of vaporization.

Net short-wave radiation flux is given by:
Rsw,net ( z ) = (F ⋅ Rsw,dir ( z ) + Rsw,dif (z ))(1 − a f − tr f ) (3.36)
70
where F describes the orientation of the leafs toward the sun ( = 0.5 for randomly ori-
ented leaves), af is the foliage albedo and trf a transmission factor set equal to 0.3
Net long-wave radiation flux is given by:
Rlw ,net (z ,T f ) = ε f Rlw↓ ( z ) + Rlw↔ ( z ) + ε f Rlw↑ ( z ) − 2ε f σ BT f4 − (1 − σ svf ( z ))σ BT f4 (3.37)
with εf is the emissivity of the leaf.

Source/sink terms expressed in the atmospheric model (equations 3.4 and 3.5) are then
calculated by:
Qh ( z ) = LAD( z )J f ,h (3.38a)
Qq ( z ) = LAD( z )J f ,evap + J f ,trans (3.38b)
The vegetation model is also coupled with the soil model since the water transpired by
the plant is supplied by the soil and hence has to be deduced from it.
3.7. The human-biometeorological dimension
A discussion on the importance of Tmrt for comfort issues and the difficulty related to its
determination was presented in section 2.2. In this respect, ENVI-met gives a good ap-
proximation of Tmrt at street level, which is expressed for each grid point (z) as follows
(Bruse 1999):
0.25
 1  
Tmrt =  Et ( z ) + α k (Dt ( z ) + I t ( z )) (3.39)
 
σ B  εp 
The surrounding environment consists of the building surfaces, the free atmosphere
(sky) and the ground surface. All radiation fluxes, i.e. direct irradiance It(z), diffuse and
diffusely reflected solar radiation Dt(z) as well as the total long-wave radiation fluxes
Et(z) from the atmosphere, ground and walls, are taken into account.
At street level, Et(z) is assumed to originate as 50 % from the upper hemisphere (sky
and buildings) and 50 % from the ground. This approximation is only valid at street
level for which the calculation of Tmrt is foreseen, because the influence of the radiation
of the ground decreases with increasing height; Et(z) is expressed by:
[ ]
Et ( z ) = 0.5 (1 − σ svf ( z ))Rlw↔ + σ svf ( z )Rlw↓ ,0 + 0.5ε sσ BT0
4
(3.40)
with the heat flux from the ground ( Es = ε sσ BT 40 ) is calculated for the actual surface
temperature (T0) at the grid point (z) in order to take into account the exposure versus
71
shadow situation. The downward radiation flux Rlw↓,0 coming from the visible part of the
sky is weighted by σsvf. The long-wave radiation emitted by the walls Rlw↔ is calculated
as an average by considering a mean value for the building surface temperatures Tw :

4
Rlw↔ ( z ) = (1 − σ svf ( z )) ε w σ B Tw (3.41)
The total diffuse radiation Dt(z), which comes partly from the sky and partly from the
walls as diffusely reflected solar radiation, is expressed by:
Dt ( z ) = σ svf ( z ) Rsw ,dif + ( 1 − σ svf ( z )) a Rsw ,dir
↓ ,0 ↓ ,0
(3.42)
where a is the mean albedo of the model area.

The irradiated part of a human body absorbs one part of the direct solar irradiance. This
is expressed by the projection factor (fp) which depends on the sun height φ and given
by:
↓
I t ( z ) = f p Rsw ,dir ( z ) (3.43a)
f p = 0.42 cos φ + 0.043 sin φ (3.43b)
3.8. Boundary conditions and course of a simulation
Fig. 3.1, shown previously, illustrates the following description. The equations used in
the boundary model are a 1D simplified form of those used in the 3D model with some
parameterisations when necessary. The vertical inflow profile up to 2500 m is calcu-
lated with the 1D model by applying a logarithmic law, based on the input values of the
horizontal wind (u, v) at 10 m height and on the roughness length z0.
The potential temperature (θstart) given as input parameter at a height of 2500 m is set to
the whole vertical profile assuming start conditions of neutrality. A vertical gradient
forms if the initial surface temperature differs from the initial air temperature. The sur-
face temperature is provided to the 1D model by the soil sub-model, and is calculated
on the basis of three input values of soil temperatures and soil humidity. The air humid-
ity profile is linear and is calculated by means of input values at 2500 m i.e. q 2500 m and
the relative humidity RH at 2 m. Turbulence quantities E and ε are constant at 2500 m

and are function of the local friction velocity u*. The surface temperature and humidity
are provided by the 3D model as mean values of the nesting area related values.
72
The initialization of the 1D model is run during a period of 8 hours with a time step of
∆t = 1s until the interactions between all start values reach a stationary state, i.e.
dK m dt 〈10 −3 m 2 ⋅ s −2 . The atmospheric equations are solved by integration of the variables
in the following order: u , v , θ , q , E and ε, and the exchange coefficients Km, Kh, and
Kq.
Start values at the inflow boundary of the 3D model are provided by the 1D boundary
model as a vertical profile. The transition from 1D to 3D schemes needs an adjustment
in a non-homogenous urban milieu. This is solved by the use of the 3D nesting area. On
the horizontal boundary, homogeneity is assumed. Wall and roof temperatures are cal-
culated at all physical boundaries in the model area. The wind speed components at
building grids are set following a no-slip condition i.e. u = v = w = 0. The wind field is
adjusted to the presence of the obstacles gradually during the initializing phase (dias-
trophic phase). At the ground surface (z = 0) and on the walls, E and ε are calculated as
a function of u* from the flow components tangential to the surface. It is assumed that
no gradient exists between the two last grids close to the outflow border.
The actual 3D simulation includes, in the following order, the calculations of soil pa-
rameters (T,η), surface quantities (T0 , q0, as), radiation update, the update of wind com-
ponents ( u , v , w ), pressure perturbation p´, turbulence quantities E, ε, Km , Kh , Kq , and
air temperature and humidity θ, q. The process is repeated once the 1D model is updated
again.
Numerically, all differential equations are approximated using the finite difference
method and solved forward-in-time. Time steps adopted vary depending on the quantity
to be calculated. The main time step is 10 minutes for the wind flow calculations.
Smaller time steps are used for E-ε system to obtain numerically stable solution (3 min-
utes). Solar radiation is usually updated in larger time-steps and can be set by the user.
To solve the advection-diffusion equation, dynamic pressure is removed from the equa-
tions of motion (equations 3.1 to 3.3) and auxiliary flow components are calculated,
these are then corrected by incorporating the dynamic pressure which has been sepa-
rately defined by means of the Poisson equation (Bruse 2004).
73
3.9. Simulations with ENVI-met 3.0 in the present work
3.9.1. Site Climate
The simulations are carried out for Ghardaia in the Mzab region, a location in the Sa-
hara of Algeria at 32.40° N, 3.80° E and 469 m above sea level. The Saharan climate is
characterized by mostly clear sky, which leads to a comparatively high solar irradiance
in the daytime and a high long-wave net radiation during the night. Hence, the summer
is hot and dry as well as long owing to the subtropical location of the region. Air tem-
perature Ta > 40 °C is not rare and the daily Ta amplitude is relatively large. The atmos-
pheric moisture content reaches only a low level (RH = 35%). In most places, wind
sweeps dust and sand for several months of the year. The winters are short and cold,
particularly at night (reaching freezing point). The rainfall is scarce but of high intensity
when it occurs (ONM 1985). The living conditions are very difficult as shown by typi-
cal old cities in this region (see chapter 6). Fig. 3.2 illustrates long-term measurements
in the region in August (1974-1985), plotted together with ENVI-met simulation for a
typical summer day (1st August).
Ghardaia, 32.40° N, 3.80° E, 1st August, bare soil

1200 42
G, ENVI-met
1000 Ta, ENVI-met 36 Ta (°C), VP (hPa)
800 S, ENVI-met 30
S, G ( Wm-2)
Ta,measured
600 24
400 18
VP, measured
200 12
VP, ENVI-met
0 6
7 9 11 13 15 17 19 21 23 1 3 5
time (LST)
Fig. 3.2. Average air temperature Ta and vapour pressure VP humidity in Ghardaia in
August (1974-1985, ONM 1985) plotted against ENVI-met simulation results for a bare
soil on the 1st August for Ta, VP, direct irradiance S and global irradiance G.
74
3.9.2. Simulation conditions
The main simulation conditions and building properties used for the case studies re-
ported in this work are listed in Table 3.1. The domain simulated is composed of two
long buildings separated by a street of a constant width of 8 m. The building height is
variable according to the aspect ratio H/W. Corresponding sky view factor SVF are
given in appendix C. The building length equals six times its height to meet the dimen-
sions of an urban canyon (see Fig. 2.4).
Table 3.1. Typical inputs’ configuration of a simulation as used in this study

% ---- Basic Configuration File for ENVI-met ---------------
% ---- MAIN-DATA Block -------------------------------------
% Symmetrical urban canyon of H/W = 2, E-W oriented, perp. wind
Name for Simulation (Text): = Ghardaia_EW_H/W=2
Input file Model Area = [INPUT]\Ghardaia_EW_H/W=2
File base name for Output (Text): = Ghardaia_EW_H/W=2
Output Directory: = [OUTPUT]\Ghardaia_EW_H/W=2
Start Simulation at Day (DD.MM.YYYY): = 01.08.2003
Start Simulation at Time (HH:MM:SS): = 06:00:00
Total Simulation Time in Hours: = 15.00
Save Model State each ? min = 60
Wind Speed in 10 m ab. Ground [m/s] = 5
Wind Direction (0:N..90:E..180:S..270:W..) = 0
Roughness Length z0 at Reference Point = 0.1
Initial Temperature Atmosphere [K] = 306
Specific Humidity in 2500 m [g Water/kg air] = 7.8
Relative Humidity in 2m [%] = 25
Data base Plants = Plants.dat
( -- Following: Optional data. The order of sections is free. --)
[POSITION]_______________________________Where the area is located on earth
Longitude (+:east -:west) in dec. deg: = 3.80
Latitude (+:northern -:southern) in dec.deg: = 32.40
Longitude Time Zone Definition: = 15.0
[SOILDATA] ______________________________________Settings for Soil
Initial Temperature Upper Layer (0-20 cm) [K]= 301
Initial Temperature Middle Layer (20-50 cm) [K]= 305
Initial Temperature Deep Layer (below 50 cm)[K]= 305
Relative Humidity Upper Layer (0-20 cm) = 30
Relative Humidity Middle Layer (20-50 cm) = 30
Relative Humidity Deep Layer (below 50 cm) = 30
[TIMING]_____________________________________Update & Save Intervals
Update Surface Data each ? sec = 60.0
Update Wind and Turbulence each ? sec = 900
Update Radiation and Shadows each ? sec = 600
Update Plant Data each ? sec = 600
[TURBULENCE]_________________________________Options Turbulence Model
Turbulence Closure ABL (0:diagn.,1:prognos.) = 1
Turbulence Closure 3D Model (0,1,2 ) = 1
Upper Boundary for e-epsilon (0:closed,1:open) = 0
[BUILDING]__________________________________Building properties
Inside Temperature [K] = 293
Heat Transmission Walls [W/m²K] = 1.7
Heat Transmission Roofs [W/m²K] = 2.2
Albedo Walls = 0.3
Albedo Roofs = 0.15
75
U-values for “standard” building materials were used (Koenigsberger et al. 1973, Mar-
kus and Morris 1980) with typical albedos (Oke 1987, VDI 1998) to allow some com-
parison with other climates eventually. Since no heat storage is included in the model,
the use of materials of high thermal capacity was not particularly relevant. The simula-
tion is started preferably at 6:00 LST at which most atmospheric processes are slow.
Input start values were not obvious to set and test simulations were often necessary. In
some cases, these are based only on theory: for instance, the ground surface temperature
was set by few degrees lower than Ta (Asaeda and Ca 1993) and a roughness length z0 =
0.1 m is chosen as a typical value for urban areas (Oke 1987). Wind speed is set equal
to 5 ms-1 at 10 m height and arranged perpendicular to street axis. A perpendicular inci-
dence of wind is used complementarily for few cases to allow some comparison
Simulations are run for daytime hours, on one hand because daylight hours represent the
period of day of usual frequentation of outdoor spaces, and on the other hand owing to
the weakness of the model in the prognosis of the night-time situation, i.e. overestima-
tion of air temperature (Fig. 3.2) and the lack of nocturnal heat release process as no
heat is stored in the building fabric.
3.9.3. Case studies
The simulations reported here were run according to the following plan (Fig. 3.3):
** symmetrical urban canyons with rectangular shape with H/W equal to 0.5, 1, 2, 3,
and 4 for East-West orientation (I-1 to I-5);
** symmetrical urban canyons with rectangular shape with H/W equal 0.5, 1, 2, 3,
and 4 for North-South orientation (I-1 to I-5);
** intermediate orientations NE-SW and NW-SE for H/W = 2, considered as an av-
erage profile between shallow and deep profiles (I-3);
** Complex urban canyons:
a. asymmetrical urban canyon with large openness to the sky (II-2),
b. urban canyons with H/W = 2 with galleries, oriented E-W, N-S, NE-SW and
NW-SE (II-1),
c. asymmetrical urban canyon with overhanging façade, including galleries and
with a smaller openness to the sky (II-3),
d. urban canyon with H/W = 2 oriented E-W with a row of trees on the north side
(II-4)
76
e. urban canyon with H/W = 1 oriented N-S with a central row of trees and two
lateral galleries (II-5).
** study of the impact of wind incidence on comfort by simulating for few selected
cases a parallel wind orientation to be compared to the perpendicular incidence
used as default condition (i.e. for I-1, I-3, II-4, II-5);
** study of the thermal comfort is conducted for summer conditions; some additional
simulations of solar access in the winter period are made, as this is a decisive issue
in the design of an urban street (I-1 to I-5).
In order to avoid redundancies, a few street geometries are selected from the large num-
ber of simulations undertaken. Explicitly, each of the irregular streets is concerned with
one or more design details simultaneously. Table 3.2 and Fig. 3.3 give an overview on
these geometries together with their actual dimensions.
The 3D grid resolution used for the simulated area is 1 m horizontally and 2 m verti-
cally. In ENVI-met, the first grid above the ground (i.e. on the z-axis) is subdivided into
five equal parts to enable a better understanding of the microclimate at pedestrian level.
All the results discussed below are given for the central part of the street, i.e. at mid-
block distance from the street ends, and calculated for a height of 1.2 m above the
ground. This height is representative for comfort calculations for a standing person.
Table 3.2. Dimensional characteristics of the investigated urban canyons

spatial resolution 1 m horizontally, 2 m vertically
street width W 8 m
building height H 4 m, 8 m, 16 m, and 32 m
building length L 6 x H (≈ urban canyon)
gallery 4 m high and 3 m width
canyon materials street: asphalt, gallery: pavement (appendix A)
Buildings: brick
Tree 6 m high, leafless base, dense and light dense
(see appendix B for LAD)
Tree row of 2 m width at the edge or 4 m width at the centre
(schemes at scale)
overhanging façade 1 m and 2 m width (schemes at scale)
wind perpendicular and parallel
77
Fig. 3.3. Geometry of the urban canyons selected for the simulations
H
W
I-1 (H/W = 0.5) I-2 (H/W = 1) I-3 (H/W = 2) I- 4 (H/W = 3) I-5 (H/W = 4)
II-1 II-2 II-3 II-4 II-5
N-S E-W NE-SW NW-SE
1.2 m
2m
street: 8 x 1 m
Section at which main analysis is made
(mid-block distance)
(not at scale)
Plan of urban canyon and vertical resolution used by ENVI-met at street level
78
4. Results of the numerical simulations
4.1. Symmetrical canyons oriented east-west
4.1.1. Air temperature
Fig. 4.1 shows the evolution of the air temperature Ta at the height 1.2 m during the
daytime for symmetrical urban canyons oriented E-W with aspect ratios H/W varying
from 0.5 to 4. Basically, Ta decreases with the increase of the aspect ratio. The shallow-
est canyon H/W = 0.5 is the warmest case study and has a comparable Ta evolution as
an unobstructed surface. The deepest canyon H/W = 4 is the coolest one.
Ghardaia, 32.40° N, 1st August, E-W orientation
40
Ta, H/W = 0.5
39
Ta, H/W = 1 unobstructed
38 surface
Ta, H/W = 2
37
Ta, H/W = 3
36 Ta, H/W = 4
Ta (°C)
35
34
33
32
31
30
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
time (LST)
Fig. 4.1. Diurnal variation of simulated air temperature Ta at 1.2 m within the canyon
for E-W oriented streets of aspect ratios H/W of 0.5, 1, 2, 3 and 4
79
Yet, the differences do not exceed 1 K between two successive canyons (i.e. H/W = 1
and 2, 2 and 3, etc.). The variances are quite small in the early morning up to 10:00 LST
between all streets. A maximum difference of less than 0.8 K is calculated. The differ-
ences become larger after 10:00 LST and reach their maximum around 14:00 LST, i.e.
∆Ta,max = 3 K. The highest air temperature equals 39.5 °C, recorded at 15:00 LST for
H/W = 0.5 while of only 36.7 °C for H/W = 4. After 17:00 LST, the canyons of H/W ≤
2 cool faster and have almost the same Ta. The deep profiles with H/W equal to 3 or 4
remain cooler than the other streets also in the evening, by approximately 1 K at 20:00
LST.
Fig. 4.2 compares the (a) direct, (b) diffuse, and (c) global radiation fluxes received at
street level for all symmetrical canyons oriented E-W. As expected, the role of the ratio
H/W is found to be decisive. The exposure to direct solar radiation diminishes as the
aspect ratio increases. As shown in Fig. 4.2a, the street H/W = 0.5 is the most exposed
to direct solar radiation (S), about 900 Wm-2 from 09:00 to 17:00 LST with two peaks at
10:00 and 16:00 LST of 950 Wm-2. Shading becomes effective only for very high aspect
ratios. The E-W orientation implies symmetry in the street irradiation in relation to
midday. The shading effect by increasing the aspect ratio is maximal around noon, and
the deepest profile H/W = 4 receives no beam irradiation during 2 hours (from 11:00 to
13:00 LST). As a first approximation, the street floor irradiance is about 200 Wm-2 less
between two successive aspect ratios. All streets are highly exposed at two times, i.e.
around 09:00 and 17:00 LST. This is due to the fact that an E-W orientation limits
strongly the effectiveness of the walls in shading the street level at these times as the
sun rays reach the street level laterally from the sides.
Fig. 4.2b shows a reverse trend for the diffuse irradiation D, which rises with the in-
crease of the aspect ratio. This is because sky view factor SVF becomes smaller as the
vertical surfaces become higher. This leads to an increase of the diffusely reflected irra-
diation from the façades more than the diffuse irradiation is decreased (see eq. 3.20).
Basically, the diffuse irradiance is less than 250 Wm-2 in all cases. The differences are
rather small, about 20 Wm-2 for each two successive streets, and the maximum differ-
ence is of less than 100 Wm-2 between the shallowest and the deepest profile.
80
st
Ghardaia, 32.40° N, 1 August, E-W orientation
1000
S_H/W = 0,5
900 S_H/W = 1
800 S_H/W = 2
S_H/W = 3
700
S_H/W = 4
600
S (W/m )
2
500
400
300
200
100
0
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.2a. The simulated direct solar radiation (S) at street level for E-W oriented streets
of aspect ratios H/W of 0.5, 1, 2, 3 and 4
st
250
D_H/W = 0,5
D_H/W = 1
200 D_H/W = 2
D_H/W = 3
D_H/W = 4
150
D (W/m )
2
100
50
0
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.2b. The simulated diffuse radiation (D) at street level for E-W oriented streets of
various aspect ratios of 0.5, 1, 2, 3 and 4
81
The global radiation G (Fig. 4.3c) is mostly affected by the direct solar component be-
cause of the subtropical location of Ghardaia and the typical clear sky. It is also worthy
of note that the global irradiance received at streets with H/W = 2 or less varies little (<
100 Wm-2), whereas the differences become substantially larger for H/W = 3 and 4.
Moreover, these figures also suggest that a deep street canyon exposed to direct solar
beam would receive much more irradiation than a horizontal surface because of an addi-
tional diffusely reflected irradiation as reported by others (e.g. Givoni 1997, Yoshida et
al. 1990/91).
st
1200
G_H/W = 0,5
G_H/W = 1
1000
G_H/W = 2
G_H/W = 3
800 G_H/W = 4
G (W/m )
2
600
400
200
0
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.2c. The simulated global radiation (G) at street level for E-W oriented streets of
4.1.3. Thermal comfort analysis
The isotherm representation shown on the right in Fig. 4.3 is privileged in the following
comfort analysis. This graph sums up the evolution in time and space of one parameter
(mainly PET in this work) across the street at mid-canyon distance as representative of
an urban canyon (i.e. 0 m ≤ x ≤ 8 m, y = L/2, and z = 1.2 m). This is preferred to a two-
dimensional representation of the whole urban canyon (fig. 4.3 plan on the left) for one
parameter at a specific time and a specific plane since the variability of most parameters
along the canyon is small.
82
Fig. 4.3. Example of an isotherm representation chosen for a detailed spatial and tempo-
ral illustration of the thermal comfort outdoors
This representation has the advantage of giving a complete picture of the diurnal evolu-
tion of the thermal situation at street level and points out the differences between the
sidewalks and the centre of a street canyon. Indeed, very local variations would be un-
derestimated if the investigation was limited to a single point within the street, as usu-
ally assumed (e.g. Pearlmutter at al. 1999, Swaid et al. 1993). This information is inter-
estingly twofold with respect to street design:
- First, the purpose of the street determines the area used by pedestrians. This may be
the whole area of the street or limited to peripheral sidewalks if motor traffic is also
planned. Hence, the differentiation between these sub-spaces is useful.
- Secondly, it allows determining whether and how long the street is totally or partly
comfortable, and consequently to assess pedestrian adjustments’ possibilities. This
is important since the frequentation of urban spaces is favoured if the period of com-
fort is long enough and if pedestrians are given the choice of moving to shaded sub-
spaces in order to adjust to a stressful climatic situation.
Fig. 4.4 compares the simulation results between (a) Ta and (b) Tmrt for the same street.
The figure shows a large difference in relation to Ta for the sunlit part of the street and
of about 6 to 10 K for the area in shade.
83
Most studies revealed that Ta shows a relatively uniform distribution in a street canyon
(e.g. Nakamura and Oke 1988, Yoshida et al. 1990/91) except for the air layer close to
urban canyon surfaces (ground or wall) which can be slightly warmer if irradiated (Na-
kamura and Oke 1988). This uniformity is well reproduced by the model but the warm-
ing of air close to the surfaces is underestimated, probably due to the 1 m grid resolution
used and the lack of heat storage in the materials. Moreover, the warming of canyon air
in comparison to roof air is known to be not significant for H/W ≈ 1 because the air
driven from roof level is well mixed inside the canyon (e.g. Nakamura and Oke 1988,
Yoshida et al. 1990/91). As already shown in Fig. 4.1, Ta differences in relation to as-
pect ratio is limited to a few degrees, which is also in good agreement with available
field studies dealing with comparable H/W ratios (e.g. Santamouris at al. 1999, Coronel
and Alvarez 2001). This confirms that Ta alone is not relevant in describing the human
comfort conditions in the summer, and contrasts with many studies which focused on Ta
as main indicator for thermal comfort in outdoor spaces (e.g. Swaid et al. 1993, Grund-
ström et al. 2003). The use of Tmrt appears to be playing a more decisive role as it influ-
ences strongly the human energy balance during sunny days (e.g. Mayer and Höppe
1987, Jendritzky and Sievers 1989, Jendritzky et al. 1990, Mayer 1993, 1998). Hence,
Tmrt is more relevant for comfort assessment.
(a) air temperature (Ta) (b) mean radiant temperature (Tmrt )

20:00 20:00
19:00 19:00
18:00 18:00
39°C
17:00 17:00
86°C
16:00
38°C 16:00
80°C
15:00 37°C 15:00
time (LST)
time (LST)
74°C
14:00 14:00
68°C
36°C
13:00 13:00
62°C
12:00 35°C 12:00
56°C
11:00 11:00 50°C
34°C
10:00 10:00 44°C
33°C
9:00 9:00 38°C
8:00 32°C 8:00 32°C
0m 1m 2m 3m 4m 5m 6m 7m 8m 0m 1m 2m 3m 4m 5m 6m 7m 8m
I----------------- street width ------------------I I----------------- street width ------------------I
Fig. 4.4. Comparison between air temperature Ta and mean radiant temperature Tmrt in
time and space for an E-W oriented street of an aspect ratio H/W = 2 at 1.2 m a.g.l.
84
Moreover, the wind speed v is a determinant parameter in the calculation of PET. In the
main simulations, the wind speed was found, as expected, to be strongly reduced in all
streets and averages a velocity of less than 0.3 ms-1 at pedestrian level at mid-canyon
distance. This is due to the perpendicular incidence of the wind in relation to street axis.
For convenience, the role of the wind incidence on comfort is discussed separately (see
section 4.5.6.). The air humidity (VP) equals 12 hPa as set initially and experiences no
change in all cases where no source for water exchange (and latent heat) is available
(waterproof surfaces, no trees or water plants).
In order to take into account Ta, VP and v together with the strongest effects of the solar
irradiation (i.e. Tmrt), the following comfort analysis is based on PET (see Chapter 2.2).
The following graphics (Figs. 4.5a to 4.5e) represent a detailed spatial and temporal
distribution of PET for E-W streets with H/W = 0.5, 1, 2, 3 and 4 respectively, on a
typical summer day (1st August) during the period time from 8:00 to 20:00 LST at
which thermal comfort is mostly required. Basically, PET values are high and range
between 38 °C and 66 °C. PET pattern in Fig. 4.5a shows extremely high values and
indicate clearly that for H/W = 0.5 the street is highly uncomfortable throughout the
day. The street area is almost fully exposed, and only about 1/8 of the street on the
north-facing part lies in shade from 11:00 to 15:00 LST with PET about 42 °C. PET
reaches a peak value of 66 °C between 16:00 and 17:00 LST. This is due to the intense
solar irradiation combined with the daily maxima of air temperatures (mean monthly
average of Ta = 39 °C in summer, ONM (1985)). The street with H/W = 1 (Fig. 4.5b)
provides almost no improvement in the extreme thermal situation, except minimally on
the north facing part where the area of lowest discomfort is somewhat larger at midday
hours. Both situations would exclude any leisure activity unless improvements of the
thermal quality based on further shading strategies are planed, e.g. higher aspect ratios,
other orientations, or further shading devices such as trees or galleries.
For H/W = 2 shown in Fig. 4.5c, about one half of the street remains exposed for the
major part of day in spite of the high walls, with PET higher than 60 °C and a peak
value of 66 °C occurring around 16:00 LST. This is due to the latitude of Ghardaia
where the sun’s height reaches 75° in the summer, making the walls only partly effec-
tive in shading the street from the sun rays impinging laterally. PET is high even for the
shaded part, with a minimum value of about 40 °C developing along the south sidewalk
during 6 hours from 10:00 to 16:00 LST. This area of shade extends to 40 % of the
street width at midday hours from 12:00 to 14:00 LST.
85
Ghardaia, 32.40° N, 3.80° E, 01 August
PET, H/W = 0.5, East-West PET, H/W = 1, East-West
20:00 20:00
19:00 19:00 70 °C
18:00 18:00 66 °C
17:00 17:00
62 °C
16:00 16:00
58 °C
15:00 15:00
54 °C
14:00 14:00
50 °C
86
time (LST)
13:00 13:00
46 °C
12:00 12:00
42 °C
11:00 11:00
10:00 38 °C
10:00
9:00 9:00 34 °C
8:00 8:00 30 °C
|----------------- street width --------------------| |----------------- street width --------------------|
Fig. 4.5a: Diurnal variation of PET street level for E-W oriented streets of an aspect ratio H/W = 0.5 (left)
Fig. 4.5 b: Diurnal variation of PET at street level for E-W oriented streets of an aspect ratio H/W = 1 (right)
PET, H/W = 2, East-West PET, H/W = 3, East-West

20:00 20:00
19:00 19:00 70 °C
18:00 18:00
66 °C
17:00 17:00
62 °C
16:00 16:00
58 °C
15:00 15:00
54 °C
14:00 14:00
50 °C
87
time (LST)
13:00 13:00
46 °C
12:00 12:00
11:00
42 °C
11:00
10:00 10:00 38 °C
9:00 9:00 34 °C
8:00 8:00 30 °C
|----------------- street width --------------------| |----------------- street width --------------------|
Fig. 4.5c: Diurnal variation of PET at street level for E-W oriented streets of an aspect ratio H/W = 2 (left)
Fig. 4.5 d: Diurnal variation of PET at street level for E-W oriented streets of an aspect ratio H/W = 3 (right)
The canyon H/W = 2 shows two opposite thermal situations on the two sides. This can
be seen as an alternative for pedestrians to move to nearby less stressful areas. Fig. 4.5d
shows that H/W = 3 brings minimal improvements compared to H/W = 2, namely a
somewhat larger area across the street with minimal PET. Yet, no improvement is found
around 8:00 and 17:00 LST in comparison to H/W = 2.
The deepest urban canyon with H/W = 4 (Fig 4.5e) shows spatially a larger area of low
PET values. These are about 40 °C and are close to Tmrt. This corresponds to the shaded
part of the street and lasts 6 hours in the south side and 1 to 4 hours on the opposite side.
The entire street is shaded between 12:00 and 13:00 LST. Increasing the aspect ratio has
in fact more impact on the maxima than on the minima. Indeed, the relatively lower air
temperatures found for deeper canyons (Fig. 4.1) seem to play a very limited role in
decreasing PET minima, which range in all cases between 38 °C and 42 °C.

PET, H/W = 4, East-West
20:00
19:00 70 °C
18:00 66 °C
17:00
62 °C
16:00
58 °C
15:00
time (LST)
54 °C
14:00
50 °C
13:00
46 °C
12:00
42 °C
11:00
10:00 38 °C
9:00 34 °C
8:00 30 °C
0m 1m 2m 3m 4m 5m 6m 7m 8m
|----------------- street width --------------------|
Fig. 4.5e. Diurnal variation of PET at street level for an E-W oriented street of an aspect
ratio H/W = 4
88
It is worthy of note that the deepest street with H/W = 4 remains highly uncomfortable
at 2 times of the day, namely in the early morning and in the late afternoon. The street is
overheated from 8:00 to 10:00 LST due to the exposure to the sun rays coming from the
east direction. Symmetrically the street experiences the highest thermal discomfort be-
tween 16:00 and 17:00 LST, due to the westerly sun exposure combined with Ta
maxima. During this period a maximum value of PET of 66 °C is calculated.
The street cools rapidly and PET reaches 34 °C at 20:00 LST. This scheme is to some
extent comparable to H/W = 2, but the deeper street cools somewhat faster due to the
shorter time of exposure of its surfaces which results in a lower amount of radiant heat
transfer.
This example shows that, despite the high aspect ratio, PET values are still above the
comfort level. This explains the typical design of very deep streets in traditional desert
cities located at the same latitudes (see Chapter 6). As well, it shows that the effective-
ness of the aspect ratio is rather limited in ensuring comfortable microclimates for an E-
W orientation for subtropical latitudes. The main reason is the lack of shading as sug-
gested by Arnfield 1990a and Bourbia and Awbi 2004 (see e.g. Fig. 2.5 and Fig. 2.6).
4.2. Symmetrical canyons oriented north-south
Fig. 4.6 shows the evolution of the air temperature Ta during the daytime for urban can-
yons oriented N-S with aspect ratios H/W = 0.5, 1, 2, 3 and 4. In the early morning
(07:00 to 09:00 LST), the differences in the air temperatures are very small between all
profiles. The maximum difference is calculated between the largest and deepest canyon
(1 K at 09:00 LST). The curve of Ta for H/W = 0.5 is similar to the one in unobstructed
milieu due to the high exposure of this street. The canyon H/W = 1 is about 1 K cooler
than H/W = 0.5 at the warmest period of the day. Deeper streets with H/W ≥ 2 show
lower amplitude, with a maximum value of about 1.5 K lower. H/W = 3 and 4 have
their maximum around 13:00, time at which the street area is directly exposed to solar
radiation. For H/W ≥ 2 Ta maximum reaches 37.2 °C and the streets are slightly cooler
before 8:00 LST and after 20:00 LST in comparison to wider canyons with H/W of 0.5
or 1.
89
Ghardaia, 32.40° N, 1st August, N-S orientation

40
Ta, H/W = 0.5
unobstructed
Ta, H/W = 1 surface
38 Ta, H/W = 2
Ta, H/W = 3
Ta, H/W = 4
36
Ta (°C)
34
32
30
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
time (LST)
Fig. 4.6. Diurnal variation of simulated air temperature Ta at 1.2 m within N-S oriented
streets with aspect ratios of 0.5, 1, 2, 3 and 4
Fig. 4.7 compares the average (a) direct, (b) diffuse and (c) global radiation fluxes for
all symmetrical streets oriented N-S received at street level. Fig. 4.7a shows symmetry
of exposure in respect to noontime as a consequence of the sun course combined with
the N-S orientation. Basically, all streets are noticeably less irradiated than E-W ori-
ented canyons of the same aspect ratio. Shading, resulting from increasing the aspect
ratio H/W, is also more significant for N-S orientation. The highest exposure to direct
solar radiation is registered at noon for all urban canyons. All streets are largely pro-
tected from the sun in the morning as well as in the afternoon. The only exception is the
street with H/W = 0.5 which receives the largest amount of solar irradiation during ten
hours. The deepest profile H/W = 4 is protected from the sun all the day except shortly
around noon. The intensities vary also strongly with increasing aspect ratio.
The diffuse radiation D (Fig. 4.7b) shows the same distribution in relation to geometry
as found for E-W streets (see Fig. 4.3). This is not surprising as the calculation of D by
the model depends only on the sky view factor and a mean albedo, both independent
from the orientation.
90

1200
S_H/W = 0,5
1000 S_H/W = 1
S_H/W = 2
800 S_H/W = 3
S_H/W = 4
S (W/m2)
600
400
200
0
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.7a. The simulated direct solar radiation (S) at street level for N-S oriented streets
with aspect ratios of 0.5, 1, 2, 3 and 4

250
D_H/W = 0,5
D_H/W = 1
200 D_H/W = 2
D_H/W = 3
D_H/W = 4
150
D (W/m2)
100
50
0
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.7b. The simulated diffuse radiation (D) at street level for NS oriented streets with
aspect ratios of 0.5, 1, 2, 3 and 4
91

1300
G_H/W = 0,5
1200
1100 G_H/W = 1
1000 G_H/W = 2
900 G_H/W = 3
800 G_H/W = 4
G (W/m )
2
700
600
500
400
300
200
100
0
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.7c. The simulated global radiation (G) at street level for NS oriented streets of
The resulting global radiation G as given in Fig. 4.7c shows to some extent a similar
pattern to S. Around noon; all streets are irradiated at almost the same intensity and a
variation of less than 100 Wm-2 is due to the diffuse component. The shallowest urban
canyon is by far the most exposed, while other canyons receive mostly a diffuse com-
ponent of irradiation. This is roughly of the same magnitude, namely by 200 Wm-2 or
less.
92
4.2.3. Thermal comfort analysis
The graphics given in Figs. 4.8a to 4.8e represent a detailed spatial and temporal distri-
bution of PET for N-S streets with H/W = 0.5, 1, 2, 3 and 4 respectively, for the 1st Au-
gust and the period of time between 8:00 and 20:00 LST.
Fig. 4.8a shows PET values lower than 40 °C between 8:00 and 11:00 LST on the west-
facing side, which is shaded in the morning. A PET value lower than 40 °C is calculated
on the east-facing side from 14:00 LST when more shade is produced by the buildings.
Otherwise, the street remains generally as uncomfortable as the E-W oriented street (i.e.
Fig. 4.5a). Values of PET exceeding 60 °C are calculated in the morning from 9:00 to
13:00 LST on the east-facing side of the street and move gradually to the opposite side
with maximum values occurring between 13:00 to 17:00 LST on the west-facing side.
Increasing the aspect ratio to unity (Fig. 4.8b) shows a noticeable amelioration in com-
parison to the E-W case. The period of extreme heat stress where PET exceeds 60 °C
lasts only 3 hours, i.e. from 10:00 to 13:00 LST on the east-facing side and from 12:00
to 16:00 LST on the north facing side. Between 8:00 and 9:00 LST the whole street ex-
periences an average PET of 34 °C.
Fig. 4.8c shows that higher aspect ratios, i.e. from H/W = 2, visibly accentuate the posi-
tive effect of the N-S orientation on outdoor thermal comfort. Peak PET values are
lower and the period of extreme heat stress (up to 58 °C) is substantially reduced, occur-
ring approximately during 2 hours around noontime. Until 11:00 LST, PET does not
exceed 38 °C. Similarly, from 14:00 LST the street becomes progressively shaded with
PET values equal to or less than 40 °C. On the west-facing side of the street, PET re-
mains below 40 °C until 11:00 LST and after 14:00 LST, at which time the street begins
to receive some shading on the opposite side. This situation extends to the whole space
of the street after 17:00 LST.
Increasing the aspect ratio H/W to 3 or 4 (Figs. 4.8d and 4.8e) allows only minimal fur-
ther improvement of the thermal situation compared to H/W = 2. The period of highest
discomfort in the whole street space lasts only one hour at midday with maximal PET
values of approximately 54 °C. No significant difference can be found between H/W =
3 and H/W = 4 but slightly lower maxima.
93
Ghardaia, 32.40° N, 3.80° E , 01 August
PET, H/W = 0.5, North-South PET, H/W = 1, North-South
20:00 20:00
19:00 19:00
18:00 18:00 70 °C
17:00 17:00 66 °C
16:00 16:00 62 °C
15:00 15:00 58 °C
14:00 14:00 54 °C
94
time (LST)
13:00 13:00 50 °C
12:00 12:00 46 °C
11:00 11:00 42 °C
10:00 10:00 38 °C
9:00 9:00 34 °C
8:00 8:00 30 °C
|------------------- street width -----------------------| |------------------- street width -----------------------|
Fig. 4.8a: Diurnal variation of PET at street level for N-S oriented streets of an aspect ratio H/W = 0.5 (left)
Fig. 4.8b: Diurnal variation of PET at street level for N-S oriented streets of an aspect ratio H/W = 1 (right)
PET, H/W = 2, North-South PET, H/W = 3, North-South
20:00 20:00
19:00 19:00 70 °C
18:00 18:00 66 °C
17:00 17:00
62 °C
16:00 16:00
58 °C
15:00 15:00
54 °C
14:00 14:00
50 °C
95
time (LST)
13:00 13:00
46 °C
12:00 12:00
42 °C
11:00 11:00
10:00 10:00 38 °C
9:00 9:00 34 °C
8:00 8:00 30 °C
|------------------- street width -----------------------| |-------------------- street width ----------------------|
Fig. 4.8c: Diurnal variation of PET at street level for N-S oriented streets of an aspect ratio H/W = 2 (left)
Fig. 4.8d: Diurnal variation of PET at street level for N-S oriented streets of an aspect ratio H/W = 3 (right)
Ghardaia, 32.40° N, 3.80°E, 01 August

PET, H/W = 4, North-South
20:00
19:00 70 °C
18:00 66 °C
17:00 62 °C
16:00 58 °C
15:00 56 °C
time (LST)
14:00 54 °C
13:00 50 °C
12:00 46 °C
11:00 42 °C
10:00 38 °C
9:00 34 °C
8:00 30 °C
0m 1m 2m 3m 4m 5m 6m 7m 8m
|--------------------- street width ----------------------|
Fig. 4.8e. Diurnal variation of PET at street level for N-S oriented streets of an aspect
ratio H/W = 4
4.3. Comparison between E-W and N-S streets
Fig. 4.9 shows the air temperature differences (∆Ta) between E-W and N-S streets for
each aspect ratio respectively. The differences are moderate in all cases and do not ex-
ceed 1.3 K. The air temperature is in fact more sensitive to increased aspect ratio than to
orientation. Basically, E-W streets are warmer than N-S streets, except around noon for
higher aspect ratios where E-W canyons become slightly cooler. In fact, the warming of
air in the canyon is directly related to solar exposure of canyon surfaces as this influ-
ences the amount of sensible heat transferred to air. Explicitly, this corresponds to the
morning hours (max. by 10:00 LST) and afternoon hours (max. by 17:00 LST) for E-W
streets and at midday hours for N-S streets. This relationship is made clear in fig. 4.9
which also compares the global radiation between both orientations for each aspect ra-
tio, and ∆G shows the same temporal trend as ∆Ta .
96
Ghardaia, Algeria, 32.40 °N, 1st August

1.6
1.2 H/W = 2
0.8
0.4 H/W = 1
∆Ta K
H/W = 0.5
0
H/W = 3
-0.4
H/W = 4
-0.8
-1.2
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
1000
800
600
400
∆ G (W/m )
200
2
-200
H/W = 0,5
-400
H/W = 1
-600 H/W = 2
H/W = 3
-800
H/W = 4
-1000
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.9. Differences in (a) air temperature (∆Ta) and (b) global radiation (∆G) between
E-W and N-S oriented streets for aspect ratios H/W of 0.5, 1, 2, 3, and 4; positive values
mean higher values for E-W cases
97
Figs. 4.10a to 4.10e represent PET differences (∆PET) between E-W and N-S orienta-
tions for each H/W ratio from 0.5 to 4, respectively. Primarily, the graphics show a lar-
ger area of positive than negative values, meaning that E-W oriented streets experience
more heat stress and for a longer time than N-S oriented streets. For wide profiles, i.e.
H/W ≤ 1, N-S streets reveal to be thermally better than E-W, even at midday hours
where the differences are small or spatially very limited. As the aspect ratio increases, a
mixed situation is observed where each orientation has some advantage at some time: E-
W streets are more stressful than N-S streets in the morning (from 8:00 to 11:00 LST)
and in the afternoon from 15:00 to 17:00 LST, and N-S streets are more stressful at
midday hours (12:00 to 13:00 LST). Moreover, E-W streets are up to 25 °C (on PET
scale) warmer than N-S streets in the morning and afternoons, while N-S streets are up
to 10 °C warmer at midday hours. This is mostly due to lower PET maxima in N-S
streets for a same aspect ratio.
Fig. 4.10a. ∆PET between an E-W and N-S oriented street for an aspect ratio of 0.5;
positive values mean higher PET values for E-W orientation
98
Figs. 4.10b to 4.10e. ∆PET between an E-W and N-S oriented street for an aspect ratio
of 1, 2, 3 and 4 respectively; positive values mean higher PET for E-W orientation
In order to explain the differences between both orientations, Table 4.1 lists the four
variables (i.e. Tmrt, Ta, VP, v) corresponding to the PET maxima recorded for the two
orientations. It appears that Tmrt is the first factor responsible in PET differences, fol-
lowed by Ta as a second modifying factor, while v and VP are not relevant in these dif-
ferences.
99
Table 4.1. Tmrt, Ta, v and VP corresponding to PET maxima for E-W versus N-S streets
for H/W varying from 0.5 to 4
orientation aspect ratio PET[max] Tmrt Ta v VP
-1
(°C) (°C) (°C) (ms ) (hPa)
E-W H/W = 0.5 62 - 68 75.6 – 83.2 35.2 – 39.6 0.03 – 0.14 12.3 -12.4
H/W = 1 62 - 66 75.9 – 81.3 33.6 – 38.8 0.03 – 0.13 12.3 -12.4
H/W = 2 62- 67 79.8- 86.3 37.9 - 38 0.16 -0.2 12.3 -12.4
H/W = 3 62 - 65 78.7 – 82.3 34.6 – 37.6 0.1 12.3 -12.4
H/W = 4 62 – 65.6 80.3 – 85.3 34.3 – 37.1 0.16 – 0.2 12.3 -12.4
N-S H/W = 0.5 62 – 65.4 75.5 – 82.1 35 – 39.5 0.01 – 0.09 12.3 -12.4
H/W = 1 60 – 65.7 72.9 – 79.6 34.3 – 38.5 0.04 – 0.13 12.3 -12.4
H/W = 2 58 - 62 73.3 – 78.4 35.5 – 37.1 0.23 – 0.24 12.3 -12.4
H/W = 3 58 – 61.2 71 – 75.6 35.6 – 37.3 0.1 12.3
H/W = 4 54 - 58 73.8 – 74.9 36.7 0.25 12.3
Hence, Tmrt is analyzed in detail for selected points below. Exemplarily, Fig. 4.11 com-
pares the individual radiant terms accounting in the calculation of Tmrt (see equations
3.39 to 3.43) for a central point between an E-W and a N-S canyon with H/W = 2. In
fact, this location i.e. grid n° 3 from north and west wall, respectively, is relevant be-
cause it experiences contrasting PET’s in relation to orientation (see Figs. 4.5c and
4.8c). Complementarily, Table 4.2a lists the actual values of each energy term absorbed
by a standing person for the most critical hours, i.e. 9:00-10:00 LST and 16:00-17:00
LST for E-W canyon and 12:00-13:00 LST for N-S canyon. Basically, the different time
and duration of sun exposure are the main explanations for the variations observed:
The amount of direct short-wave irradiance absorbed ( 0.7 f p Rsw
↓
,dir
) is maximal around
9:00 and 17:00 LST in an E-W case because the sun is low (φ ≈ 37°- 49°) and impinges
laterally on a longitudinal-shape body. The projection factor is thus high fp (0.30 < fp <
0.37). By contrast, when PET maximum occurs for N-S street (i.e. φ ≈ 72°-75°), the sun
position is maximal at midday and leads to minimal fp (≈ 0.14). The short-wave diffuse
↓
irradiance 0.7σ svf Rsw ,dif absorbed in a N-S street at noontime is slightly higher than in the
early morning and late afternoons (max. 10 Wm-2). This is proportional to G, and is
rather insignificant in the final differences. A maximal difference in the total SW irradi-
ance absorbed at peak hours between E-W and N-S equals 80 Wm-2. Moreover, Ta is
higher for an E-W street in the evening than a N-S street at noontime by up to 2 K and,
hence, contributed in rising PET in the afternoon in case of E-W orientation.
100
Ghardaia, 32.40 °N, 01 August

600 41
total LW absorbed
E-W
500 39
400 Ta 37
Radiation flux Wm-2
total SW absorbed LW from ground
Ta (°C)
300 35
SW direct
200 33
SW diffuse
LW from sky
100 31
LW from buildings
0 29
8 9 10 11 12 13 14 15 16 17 18 19 20
600 41
N-S
500 total LW absorbed 39
400 37
Radiation flux Wm-2
Ta
Ta (°C)
300 LW from ground 35
200 SW diffuse total SW absorbed 33

LW from sky
100 31
LW from buildings
SW direct
0 29
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.11. Individual short-wave (SW) and long-wave (LW) energy terms absorbed by a
standing person at the street centre in an E-W and N-S oriented street with H/W = 2
Table 4.2a. Individual radiant energy terms (Wm-2) absorbed by a standing person at the
most stressful hours for E-W vs. N-S canyon of H/W = 2 at street centre (SVF = 0.569)
Σ Σ Σ
Time SW_dir. SW_dif. SW LW_grd. LW_sky LW_bldgs LW (SW+LW) Tmrt Ta PET
LST Wm-2 Wm-2 Wm-2 Wm-2 Wm-2 Wm-2 Wm-2 Wm-2 °C °C °C
E-W orientation
9:00 230.3 120.5 350.8 267.5 125.9 86.8 480.2 831.0 77.7 33.1 58.9
10:00 210.7 133.2 343.9 294.0 129.6 86.9 510.5 854.4 80.0 34.4 61.2
16:00 219.2 130.0 349.2 335.6 139.2 87.0 561.8 911.0 85.9 38.0 67.9
17:00 230.2 115.0 345.2 322.4 138.3 86.9 547.6 892.8 84.2 38.0 66.9
N-S orientation
12:00 126.1 144.7 270.8 300.5 133.3 86.9 520.7 791.5 73.1 35.5 57.4
13:00 107.5 145.7 253.1 327.6 137.0 86.9 551.5 804.6 74.6 36.7 59.0
101
Inversely, Ta is the lowest in the E-W street in the morning which had an opposite effect
on PET and explains the 58.9 °C versus 66.87°C in the late afternoon. Yet, this influ-
ence is small. The absorbed atmospheric radiation (εp0.5σsvf Rlw↓ ) varies slightly
throughout the day and the final differences are moderate (max. 10 Wm-2) since
weighted by the same sky view factor (σsvf ) for both orientations.
Secondly, the longer the period of exposure of the ground surface to direct solar radia-
( )
tion the warmer it is and thus the greater heat it releases i.e. ε p 0.5 ε sσ BT0 . The upward
4
heat flux is larger for an E-W street in the afternoon and is lower in the morning with
differences reaching 70 Wm-2. A N-S street shows clearly lower values in comparison to
E-W, except at 13:00 LST after one hour of irradiation’s absorption. Finally, the radiant
heat gained from the buildings (εp0.5(1-σsvf) Rlw↔ ) is invariable between both orientations
and throughout the day as this is based on the assumption of an average wall tempera-
ture Tw for the whole area simulated.
Furthermore, if the above values are compared with their corresponding data for H/W =
4 (Table 4.2b), one can notice that PET differences between E-W and N-S oriented can-
yons are also valid. Yet, the absolute quantities show:
no change in the direct short-wave irradiance absorbed,
more diffuse radiation as the sky view factor decreases (by implication (1-σsvf) in-
creases),
less radiant heat from the ground as this is for a shorter time irradiated and cumulates
less heat, especially at 17.00 LST, and
less atmospheric radiation absorbed and more heat gained from the buildings as both
depend on the sky view factor.
Table 4.2b. Individual radiant energy terms (Wm-2) absorbed by a standing person at the
most stressful hours for E-W vs. N-S canyon of H/W = 4 at street centre (SVF = 0.375)
Σ Σ Σ
E-W orientation
9:00 230.3 142.6 372.9 266.6 85.4 124.3 476.2 849.1 78.4 33.2 60.7
10:00 210.7 156.1 366.8 297.8 87.2 124.4 509.4 876.2 81.0 34.3 62.9
16:00 219.2 152.7 371.9 308.7 91.8 124.5 524.9 896.9 83.3 37.0 65.4
17:00 230.2 136.6 366.8 308.5 91.4 124.4 524.4 891.3 83.0 37.1 65.3
N-S orientation
13:00 107.5 168.6 276.0 321.3 92.5 124.5 538.2 814.3 74.4 36.7 58.7
102
4.4. Intermediate orientations NE-SW and NW-SE
The precedent pages showed the contrasting thermal comfort situation between E-W
and N-S oriented streets. In the following examples, a number of complementary results
for intermediate orientations are presented, i.e. NE-SW and NW-SE.
Theoretically, these orientations allow more exposure of the façades to the sun in the
winter than a N-S street and at the same time offer an easier protection of the façades
from the sun in the Summer (Givoni 1976, see also Fig. 2.6). The hypothesis was then
to verify whether these orientations ensure comfortable outdoor comfort conditions and
hence, whether they can be a compromise between summer and winter comfort needs.
Wide canyons were found to be highly uncomfortable for both E-W and N-S orienta-
tions and it is expected that intermediate orientations bring no improvement. Higher
aspect ratios offer principally a better thermal situation at street level, but differences
between the two cardinal orientations are more manifest. Exemplarily, the comfort
situation in a street with H/W = 2 is investigated for the four orientations. Interestingly,
the comparison (Fig. 4.12) shows that some similarity in the PET patterns between a N-
S orientation and the intermediate orientations NE-SW and NW-SE, whereas an E-W
orientation is noticeably different and uncomfortable for a much longer period of time.
Yet, the duration of extreme discomfort for intermediate orientations is longer compared
to N-S. Extreme PET values are recorded during 4 hours for the NE-SW street, 3 hours
for a NW-SE street and only about 2 hours for a N-S oriented street. The extreme ther-
mal stress affects simultaneously about two-third of the street space in all three cases,
while one-third of the street width will always experience lower values of PET, mainly
due to shading. Moreover, the time of highest discomfort occurs at different times of the
day depending on the orientation. In fact, a NW-SE street is extremely uncomfortable
roughly between 10:00 and 13:00 LST. This situation is shifted to 12:00 to 13:00 LST
for N-S orientation and from 13:00 to 14:00 LST for a NE-SW street.
103
Ghardaia, 32.40° N, 01 August, H/W =2
(a) PET, H/W = 2, East-West (b) PET, H/W = 2, North-South

20:00 20:00
19:00 19:00
18:00 18:00
17:00 17:00
16:00 16:00
time (LST)
15:00 15:00
14:00 14:00
13:00 13:00
12:00 12:00
68°C
11:00 11:00
64°C
10:00 10:00
60°C
9:00 9:00
8:00
56°C
8:00
0m 1m 2m 3m 4m 5m 6m 7m 8m 0m 1m 2m 3m 4m 5m 6m 7m 8m 52°C
(c) PET, H/W = 2, NE-SW (d) PET, H/W = 2, NW-SE 48°C
20:00 20:00
44°C
19:00 19:00
40°C
18:00 18:00
36°C
17:00 17:00
32°C
16:00 16:00
28°C
time (LST)
15:00 15:00
14:00 14:00
13:00 13:00
12:00 12:00
11:00 11:00
10:00 10:00
9:00 9:00
8:00 8:00
|--------------- street width ---------------| |--------------- street width ---------------|
Fig. 4.12. Comparison of PET patterns according to street orientations E-W, N-S, NE-
SW and NW-SE, with an aspect ratio H/W = 2
104
4.5. Complex urban canyons
The case studies II-1 to II-5 (see Fig. 3.2) are analysed below, as well as the role of
wind incidence upon a street canyon on comfort.
Figs. 4.13a to 4.13c show a comparison of air temperatures Ta at street level for canyons
with irregular vertical profiles, overhanging façades and including galleries as well as
trees. Basically, E-W streets are the warmest with the largest differences occurring in
the afternoon around 16:00 LST (up to 1.5 K between E-W and N-S orientation). NE-
SW streets are also warmer than NW-SE and N-S canyons because of a longer exposure
to direct solar radiation. Yet, the differences are very small between the various geome-
tries for the same orientation.
Fig. 4.13a compares the air temperature Ta between an asymmetrical canyon with H1/W
= 1 and H2/W = 2 and a symmetrical canyon with H/W = 2 for the four orientations. All
include galleries. Asymmetrical canyons are slightly warmer during the day (up to 0.6
K) due to their larger exposure to sun radiation and show a trend to be cooler from
17:00 LST (≈ 0.3 K) when the streets become shaded. This attests for a potentially
faster cooling due to a larger openness to the sky ( σ svf : 0.462 versus 0.390). When the
street includes horizontal shading and is asymmetrical with H2/W = 2 and H1/W = 1.5
(Fig. 4.13b), it tends to warm more in the morning hours in comparison to regular can-
yons of H/W = 2 because of a larger exposure of the canyon surfaces leading to more
heat transfer to air. These differences are reduced in the late afternoon, but the E-W
streets remain the warmest, yet the irregular E-W canyon cools faster owing to its larger
sky view. Fig. 4.13c shows a selection of case studies including rows of trees with vari-
ous crown densities and two wind incidences. Air temperature observed in planted can-
yons is up to 1.5 K lower in comparison with unplanted streets with the same aspect
ratio, namely 37.3 °C against 38.8 °C. The differences are larger for the case study of
H/W = 1 where the row of trees is larger (4 m versus 2 m). The differences are, how-
ever, much smaller among planted canyons when changing the leaf area density (LAD)
from dense to light, as well as between a perpendicular and a parallel wind. A maximum
difference of 0.8 K is recorded between 11:00 and 18:00 LST, more likely because of
the different aspect ratio and orientation.
105

39
38
37
36
Ta (°C)
35
34 N-S, asym. H/W = 2 & 1 NE-SW, asym. H/W = 2 & 1
E-W, asym. H/W = 2 & 1 NW-SE, asym. H/W = 2 & 1

33
NS, H/W = 2 NE-SW, H/W = 2
32
EW, H/W = 2 NW-SE, H/W = 2
31
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.13a. Average air temperature Ta at street level (1.2 m a.g.l.) for asymmetrical
urban canyons and symmetrical canyons of H/W = 2

39
38
37
36
Ta (°C)
35
34
N-S, asym. & overhangs NE-SW, asym. & overhangs
33 E-W, asym. & overhangs NW-SE, asym. & overhangs
NS, H/W = 2 NE-SW, H/W = 2

32
EW, H/W = 2 NW-SE, H/W = 2
31
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.13b. Average air temperature Ta at street level (1.2 m a.g.l.) for asymmetrical
urban canyons with overhanging façades and symmetrical canyons of H/W = 2
106

39
38
37
36
35
Ta (°C)
34 E-W, dense tree, H/W = 2, perp. wind

E-W, dense tree, H/W = 2, paral. wind
33 E-W, light dense tree, H/W = 2, paral. wind
N-S, dense tree, H/W = 1, perp. wind
32 E-W, light dense tree, H/W = 1, paral. wind
E-W, H/W = 2
31 N-S, H/W = 1
E-W, H/W = 1
30
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.13c. Average air temperature Ta at street level (1.2 m a.g.l.) for urban canyons
with trees and similar canyons without trees

40
39
38
37
36
Ta (°C)
35
34 E-W, H/W = 0.5, perp. wind

E-W, H/W = 1, perp. wind
33 E-W, H/W = 2, perp. wind
32 E-W, H/W = 0.5, paral. wind
E-W, H/W = 1, paral. wind
31
E-W, H/W = 2, paral. wind
30
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.13d. Average air temperature Ta at street level (1.2 m a.g.l.) for selected urban
canyons for a perpendicular and parallel incidence of wind
107
Finally, a parallel wind incidence in respect to street axis leads to a slight increase in Ta
in wide canyons (Fig. 4.13d). This can be explained by a higher transfer of heat as sen-
sible flux (∼ 30 Wm-2) induced by a higher exchange coefficient. These effects become
negligible for a higher aspect ratio of H/W = 2 probably due to noticeably more shading
of the canyon surfaces.
4.5.2. Role of galleries
Using galleries as a shading device is usual and already known from the Greek portico
in the Antiquity (e.g. Lechner 1991). Colonnades are especially suitable in hot climate
and are of common use in traditional and contemporary architectures. However, this
issue has been rarely investigated from the point view of climate comfort (e.g. Swaid et
al. 1993, Littlefair et al. 2001). The following examples (Figs. 4.14a to 4.14d) present a
quantitative evaluation of the thermal situation within urban streets of H/W = 2 includ-
ing galleries and for various street orientations. The gallery is 4 m high and 3 m wide,
i.e. 2 and 3 grids respectively (II-1 in Fig. 3.2).
On the whole, the thermal situation in the area of the galleries is, as expected, better
than at irradiated locations within the street. The covered areas have minimal PET val-
ues, which range between 34 °C and 42 °C. However, these covered spaces also experi-
ence periods of high stress in form of an extension of the discomfort zone observed at
the sidewalks. This is due to an exposure of the standing body and the ground surface to
direct solar beam in spite of the relatively high aspect ratio. This depends on the orienta-
tion of the street combined with the dimensions of the gallery itself, i.e. height and
width (Littlefair et al. 2001).
With respect to orientation, Fig. 4.14a shows that the two galleries in an E-W street are
well protected and the extent of discomfort is very limited. The gallery on the north side
is only partially stressful before and after noontime and contrasts strongly with the ex-
treme PET values in the adjacent open area. This is attributable to the effectiveness of
horizontal shading in an E-W orientation (e.g. Lechner 1991). The gallery on the south
side is as expected shaded, except shortly around 17:00 LST because of lateral and
skimming sun’s rays. Similarly, the gallery SE in a NE-SW street remains in shade all
time with the lowest PET values, even when the street is highly uncomfortable (Fig.
4.14c). Galleries SW and E are at most stressful in 1/3 of their width during one hour.
108

(a) PET, H/W = 2 with galleries, E-W orientation
20:00
19:00 70 °C
18:00 66 °C
17:00
62 °C
16:00
58 °C
15:00
time (LST)
54 °C
14:00
50 °C
13:00
46 °C
12:00
42 °C
11:00
10:00 38 °C
9:00 34 °C
8:00 30 °C
0 m 1 m 2 m 3 m 4 m 5 m 6 m 7 m 8 m 9 m 10 m 11 m 12 m 13 m 14 m
|------ gal. N ------|------------------- street width -----------------------|------ gal. S ------|
(b) PET, H/W = 2 with galleries, N-S orientation

20:00
19:00 70 °C
18:00 66 °C
17:00
62 °C
16:00
58 °C
15:00
time (LST)
54 °C
14:00
50 °C
13:00
46 °C
12:00
42 °C
11:00
10:00 38 °C
9:00 34 °C
8:00 30 °C
|------ gal. W ------|------------------- street width -----------------------|------ gal. E ------|
Figs. 4.14a and 4.14b. PET distribution across symmetrical urban canyons including
galleries on both sides for (a) E-W and (b) N-S oriented streets (H/W = 2)
109

(c) PET, H/W = 2 with galleries, NE-SW orientation
20:00
19:00 70 °C
18:00 66 °C
17:00
62 °C
16:00
58 °C
15:00
time (LST)
54 °C
14:00
50 °C
13:00
46 °C
12:00
42 °C
11:00
10:00 38 °C
9:00 34 °C
8:00 30 °C
|---- gal. NW -----|------------------- street width -----------------------|----- gal. SE ------|
(d) PET, H/W = 2 with galleries, NW-SE orientation

20:00
19:00 70 °C
18:00 66 °C
17:00
62 °C
16:00
58 °C
15:00
time (LST)
54 °C
14:00
50 °C
13:00
46 °C
12:00
42 °C
11:00
10:00 38 °C
9:00 34 °C
8:00 30 °C
|----- gal. SW -----|------------------- street width -----------------------|------ gal. NE -----|
Fig. 4.14c and 4.14d. PET distribution across symmetrical urban canyons including gal-
leries on both sides for (c) NE-SW and (d) NW-SE oriented streets (H/W = 2)
110
In the other galleries (W, NE) extreme PET values are recorded in about 2/3 of the cov-
ered area for approximately 2 hours for H/W = 2. In fact, the aspect ratio of the gallery
in combination with the orientation and aspect ratio of the street are all decisive.
Hence, the discomfort observed in the examples mentioned is expected to decrease for
deeper streets and to increase for wider streets and/or higher galleries. This is further
discussed with the next examples (see Figs. 4.16 to 4.19). Moreover, the period of ex-
treme discomfort does not occur at the same time in comparison to the main street area,
especially for intermediate orientations (except for gal. N). Indeed, this period is
“shifted” to about one hour before or after the most critical time within the open street,
suggesting that an alternative for people to move into shade is available.
It is worthy of note, however, that PET minima under the galleries are not lower than
those recorded in shaded parts of the “open” street. PET maxima are also anomalously
higher in the gallery by up to 4 K. One explanation to this is the insignificant differ-
ences in the air temperature and wind speed found across the street at mid-block dis-
tance. Another reason is probably related to the way Tmrt is calculated by the model. For
more clarity, Fig. 4.15 compares radiation fluxes between a location in the gallery and
at the canyon centre of a N-S oriented street with H/W = 2. Table 4.3 lists the values of
individual energy terms accounting in Tmrt calculation for the most stressful hours.
Around 11:00 LST, when the gallery is irradiated, a standing person absorbs more direct
radiation than later at noontime when the street centre becomes irradiated. This is due to
a lower sun position which implies a higher fp (0.24 at 11:00 LST vs. 0.17 at 12:00
LST). The outgoing heat from the ground increases slightly in the gallery when the
ground surface becomes shortly irradiated. The ground surface at street centre heats
more and irradiates more because of a longer period of exposure and also because of the
asphalt material used, whereas the gallery’s floor is set as pavement (Appendix A).
Moreover, the gallery is reported to receive more diffuse radiation than the street centre,
i.e. up to 55 Wm-2. This is surprising and this overestimation is attributable to the lower
sky view factor of the gallery (0.1171 vs. 0.5692) which leads to an important increase
in the diffusely reflected radiation according to equation 3.42. For the same reason, the
covered area receives less radiant heat from the sky (27 Wm-2 against 133 Wm-2 on av-
erage) and more radiant heat from the walls, namely 87 Wm-2 against 178 Wm-2 (see
equations 3.41 and 3.42). For these reasons, it is expected that the mitigation of thermal
stress under the galleries is underestimated by the model.
111

600
LW_total street
500
LW_total gallery
radiation flux Wm-2
400
300
SW_total gallery
200
100 SW_total street
0
8 9 10 11 12 13 14 15 16 17 18 19 20
450
gal. SW-dif gal. SW-dir gal. LW-ground gal. LW-sky
400 gal. LW-walls str. SW-dif str. SW-dir str. LW-ground
str. LW-sky str. LW-walls
350
300
radiation flux Wm-2
250
200
150
100
50
0
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Figs. 4.15. Individual short-wave (SW) and long-wave (LW) energy terms absorbed by
a standing person for a N-S oriented street with H/W = 2 for points within a gallery and
at the street centre
Table 4.3. Individual short-wave (SW) and long-wave (LW) energy terms (Wm-2) ab-
sorbed by a standing person at the most stressful hours in a gallery and at street centre
of for a N-S canyon of H/W = 2
Σ Σ Σ
in gallery
11:00 171.1 198.7 369.8 271.6 26.8 177.9 476.3 846.0 83.3 34.7 64.6
within street
12:00 126.1 144.7 270.8 302.6 133.3 86.8 522.6 793.4 73.5 36.0 58.3
13:00 107.5 145.7 253.1 330.6 137.0 86.8 554.5 807.6 75.1 37.4 60.0
112
In spite of these uncertainties, the model gives a good differentiation of Tmrt between
irradiated and shaded situations because ENVI-met takes into account accurately the
direct irradiation of the body and the ground surface, both decisive in these cases. How-
ever, a different parameterisation than the SVF seems to be necessary for a better esti-
mation of the various fluxes accounting in Tmrt in case of covered urban spaces. No in-
formation in the literature could be found for a comparison with measured data. Hence,
attention is drawn here on the relevance of more on-site measurements for assessing
comfort within galleries.
4.5.3. Role of the asymmetry and overhanging façades
The following examples (II-2 and II-3 in Fig. 3.2) introduce a design alternative which
is opposite to the previous ones. The street II-2 is asymmetric with a greater openness to
the sky in order to keep a higher potential of solar access in winter. The street II-3 is
more complex and combines between a relatively larger exposure of the walls in com-
parison to a symmetrical canyon with H/W = 2 but with an offset of the facades to pro-
mote more shade at the street level in summer. The relevance of asymmetrical street
geometries has been pointed out by the solar urban architecture for optimizing internal
solar gains (e.g. Knowles 1981, Djenane 1998, Ali-Toudert 2000, Littlefair et al. 2001,
Pereira et al. 2001, Thomas 2003). Enlarging the sky view implied by this asymmetry
also promotes a faster cooling at night (Oke 1988, Arnfield 1990a). Obviously, it is ex-
pected that this geometry leads to more solar exposure of the street in the summer. So,
galleries as a way to protect pedestrian spaces are added and simultaneously assessed.
The first example illustrated by Figs. 4.16a and 4.16b is an E-W oriented street. The
canyon has an aspect ratio H1/W = 1 on one side and H2/W = 2 on the other side.
Fig. 4.16b shows, as expected, that in the asymmetrical profile the thermal situation is
more stressful than in a corresponding regular street (i.e. H/W = 2). The warming of the
street reaches 20 K on the PET scale if compared to H/W = 2 for an additional 1/8 of the
street width on the south side (see Fig. 4.5c). Yet, no further effect on the north half part
is observed, which is equally uncomfortable. Also, no difference is found if compared
to H/W = 2 after 17.00 LST in the whole street area. If compared to the regular geome-
try of H/W = 1 (see Fig. 4.5b), the spatial and temporal evolution of PET is noticeably
similar.
113

PET, asymmetrical profile: H2/W = 2 and H1/W = 1, E-W
20:00
19:00 70 °C
18:00 66 °C
17:00
62 °C
16:00
58 °C
15:00
time (LST)
54 °C
14:00
50 °C
13:00
46 °C
12:00
42 °C
11:00
10:00
38 °C
9:00 34 °C
8:00 30 °C
|------ gal. N ------|------------------- street width ------------------------|------ gal. S ------|
Fig. 4.16a. PET distribution across an asymmetrical profile with H2/W = 2 and H1/W =
1 (case II-2) oriented E-W and including galleries
Fig. 4.16b. ∆PET between asymmetrical canyon (H2/W = 2, H1/W = 1) and symmetrical
canyons H/W = 2 (left) and H/W = 1 (right) for E-W orientation
114
However, some advantage for the asymmetrical street is noted with a better thermal
situation after 16:00 LST and in the early morning when the sun’s rays coming laterally
from the sides are blocked by the higher façades.
This comparison reveals that such an asymmetry offers an intermediate thermal situa-
tion between the two regular streets H/W = 2 and H/W = 1. It allows a shorter period of
time of discomfort than H/W = 1 in the afternoon, while keeping a higher plan density
with a relatively small disadvantage on comfort in comparison to H/W = 2. As well,
more solar caption in winter in ensured together with a faster heat release in summer.
As previously shown, air temperature vary little (Fig 4.13a). The wind speed is also
insignificantly variable. In contrast, the radiation fluxes summarized by Tmrt play the
main role in the PET differences observed. A comparison of each single radiation com-
ponent for these two canyons (i.e. I-3 and II-2 in Fig. 3.3) allows understanding to
which extent these are responsible in the differences observed in the thermal comfort.
The sky view factor obviously explains the differences in the diffuse radiation (diffuse
and diffusely reflected) received at street level, yet, these are insignificant. The asym-
metrical street has a larger sky view factor (0.1 larger, Appendix C) and leads on one
hand to more diffuse radiation (≤ 12 Wm-2) but on the other hand to less diffusely re-
flected radiation (≈ 25 Wm-2). The total diffuse radiation received at street level is,
therefore, less for an asymmetrical profile than in a symmetrical canyon. Hence, the
main reason to higher Tmrt is found to be, as expected, the greater exposure to direct
solar radiation (S) promoted by the larger openness to the sky, which increases the heat
released by the irradiated ground surface as well as the direct solar radiation absorbed
by a pedestrian.
The following graphics (Figs. 4.17 to 4.19) illustrate the thermal comfort situation for
the same asymmetrical geometry for N-S, NE-SW and NW-SE orientations, with the
highest wall facing E, S-E and S-W respectively. Complementary observations can be
summarized as follows:
- For a N-S orientation, the extreme discomfort period extends to the morning hours
for 2/3 of the street canyon in comparison with H/W = 2 (Fig. 4.17b). If compared to
H/W = 1, the street shows a substantial improvement in the thermal situation (up to
24 K lower) between 14:00 and 17:00 LST for 75 % of the street width. The inter-
mediate orientations show the same similar trends.
115
PET, asymmetrical profile: H2/W = 2 and H1/W = 1, N-S

20:00
19:00
66°C
18:00
62°C
17:00
58°C
16:00
15:00
54°C
time (LST)
14:00 50°C
13:00
46°C
12:00
42°C
11:00
38°C
10:00
9:00
34°C
8:00 30°C
0m 1m 2m 3m 4m 5m 6m 7m 8m 9 m 10 m 11 m 12 m 13 m 14 m
|------- gal. W -------|---------------------- street width --------------------------|-------- gal. E ------|
Fig. 4.17a. PET patterns across an asymmetrical profile with H2/W = 2 and H1/W = 1
oriented N-S and including galleries
Fig. 4.17b. ∆PET between asymmetrical canyon (H2/W = 2, H1/W = 1) and symmetrical
canyons H/W = 2 (left) and H/W = 1 (right) for N-S orientation
116
PET, asymmetrical profile: H2/W = 2 and H1/W = 1, NE-SW

20:00
19:00 66°C
18:00
62°C
17:00
58°C
16:00
15:00 54°C
time (LST)
14:00 50°C
13:00 46°C
12:00
42°C
11:00
38°C
10:00
9:00 34°C
8:00 30°C
|------ gal. NW ------|----------------------- street width -------------------------|------ gal. SE ------|
Fig. 4.18. PET patterns across an asymmetrical profile with H2/W = 2 and H1/W = 1
(case II-2) oriented NE-SW and including galleries
PET, asymmetrical profile: H2/W = 2 and H1/W = 1, NW-SE

20:00
19:00 66°C
18:00
62°C
17:00
58°C
16:00
54°C
15:00
time (LST)
14:00 50°C
13:00 46°C
12:00
42°C
11:00
38°C
10:00
34°C
9:00
8:00 30°C
|------- gal. SW ------|---------------------- street width -------------------------|------- gal. NE ------|
Fig. 4.19. PET patterns across an asymmetrical profile with H2/W = 2 and H1/W = 1
(case II-2) oriented NW-SE and including galleries
117
- Intermediate orientations show an appreciable amelioration in the thermal comfort

situation in summer, however the gallery NW of the NE-SW street still experiences
maximum PET values around 10:00 LST. PET is maximal for only two hours for
each point across the street, indicating that during the whole day an alternative is
available to walk in a comfortable part of the street.
- With regard to the areas within the galleries, these figures show clearly that the ef-
fectiveness of the galleries in mitigating the heat stress is reduced if the aspect ratio
decreases. Explicitly, the period of time of extreme discomfort within the galleries
becomes longer depending on the orientation. The E-W orientation is the less af-
fected by the aspect ratio and insignificant differences are observed in the comfort
situation of the gallery N (Figs. 4.14a and 4.16a).
For N-S, NE-SW and NW-SE orientations, the period of extreme discomfort is longer
(about 2 to 3 hours) due to the combination of relatively low sun position and lateral
incidence of direct solar beam. This suggests that the galleries are moderately effective
for wide street canyons (H /W ≤ 1) oriented NE-SW or NW-SE. By contrast, galleries
along an E-W orientation seem to be noticeably more effective even for wide street can-
yons.
The use of self-shading façades or horizontal shading devices on the walls is well-
known in traditional architectures in hot climates (e.g. Fig. 2.9). The advantages of
these strategies are known in respect to indoor climates as these supply protection from
undesirable solar radiation in the summer. This issue is addressed in the next examples
which combine the following design strategies: use of galleries, asymmetry and over-
hanging façades. The geometry used here is simplified owing to the limits induced by
the resolution of the model, but horizontal shading devices can also be balconies or in-
clined façades, etc. Complementary observations can be summarized as follow:
- The area and period of highest discomfort is noticeably lower for all 4 orientations
(Figs. 4.20a to 4.20d) if compared to a simple geometry of higher aspect ratio, i.e.
H/W = 2 (Figs. 4.14a to d). PET maxima are also basically lower than those re-
corded in Fig. 4.14, i.e. 62 °C against 58 °C.
118
Figs. 4.20a and 4.20b. PET patterns across an asymmetrical profile with overhanging
façades (H2/W = 2 and H1/W = 1.5) oriented E-W and N-S, respectively
119
Figs. 4.20c and 4.20d. PET patterns across an asymmetrical profile with overhanging
façades (H2/W = 2 and H1/W = 1.5) oriented NE-SW and NW-SE, respectively
120
- The overhangs on the façades are most efficient for a N-S and NW-SE streets and
less for a NE-SW street and E-W streets. The E-W oriented street remains the most
uncomfortable. Yet, offsetting the façades leads to a better protection of the street’s
sidewalks.
- The N-S oriented street is the most comfortable with a very restricted area of ex-
treme values, namely at street centre at noontime only with a full protection of the
galleries.
4.5.4. Role of the vegetation
The use of vegetation is a complementary strategy for mitigating heat stress at street
level (see 2.1.6). This solution is especially suited when the façades do not operate as an
efficient shading device for the street area, either because of a large aspect ratio or an
inappropriate street orientation or both. In the following case studies, the trees have a
total height of 6 m, including a leafless base of 2 m height and a dense crown (II-4 and
II-5 in Fig. 3.3).
Fig. 4.21 shows the PET patterns for an E-W oriented street of H/W = 2 including a
narrow row of trees on the north side. A similar case without trees was discussed with
Fig. 4.5c. Air temperature was found to play a secondary role in the final comfort situa-
tion as Ta decreased in the planted streets up to 1.5 K (Fig. 4.13c). VP showed almost no
change because of the lacking water in the soil and was insignificant in the differences
observed. This is also the case for v lying by 0.3 ms-1. By contrast, the radiation fluxes
are decisive and confirm that shading is the most effective climatic property of the vege-
tation in improving comfort (e.g. McPherson 1992, McPherson and Simpson 1995,
Shashua-Bar and Hoffmann 2000).
The use of trees leads to a decrease of PET up to 22 K directly under the tree crowns
because of less irradiation. The decrease in the received direct solar radiation (∆S) at 1.2
m a.g.l. is at least 200 Wm-2 and over 800 Wm-2 as shown in Fig. 4.22. In fact, the at-
tenuation of solar irradiation is function of an extinction coefficient and leaf area index
LAI (see equation 3.18). For the direct irradiation, LAI takes account the actual dis-
tance “traversed” by the sun’s rays for the integration of LAD, i.e. an optical length and
LAI is then expressed as LAI*. This optical length is increased when the sun’s rays are
nearly “parallel” to the row of trees and depends on the sun position together with the
orientation.
121

PET, H/W = 2 with a row of trees , E-W
20:00
19:00 70 °C
18:00 66 °C
17:00 62 °C
16:00
58 °C
15:00
time (LST)
54 °C
14:00
50 °C
13:00
46 °C
12:00
42 °C
11:00
10:00 38 °C
9:00 34 °C
8:00 30 °C
0m 1m 2m 3m 4m 5m 6m 7m 8m
|---------------- street width --------------------|
Fig. 4.21. PET patterns within a street oriented E-W with H/W = 2 and a row of trees on
the south-facing side (……. projection of trees’ area)
∆ S, H/W =2, with vs. without trees, E-W ∆ L-upwards, E-W, H/W =2, with vs. without trees
20:00 20:00
19:00 19:00
18:00 18:00
17:00 17:00
16:00 16:00
15:00 15:00
time LST
time LST
14:00 14:00
13:00 13:00
12:00 12:00
11:00 11:00
10:00 10:00
9:00 9:00
8:00 8:00
I-------------------- street width ----------------------I I------------------------- street width ---------------------------I
Fig. 4.22. Differences in (a) direct solar radiation (∆S) and (b) long-wave radiation (∆L-
upwards) emitted by the ground between streets with a row of trees vs. without trees
122
This occurs between 9:00 and 10:00 LST as well as 16:00 to 17:00 LST for an E-W
orientation and result in the greatest heat stress mitigation (Fig. 4.21). Another explana-
tion for the decrease of PET is the strongly reduced heat absorbed by the ground surface
(up to 200 Wm-2) under the vegetation and hence the heat emitted upwards and ab-
sorbed by a human body (Fig. 4.22). The graphics also show that the cooling effect is
effective mostly under the tree crowns and does not extend to the surroundings. This
agrees with the observation made by Shashua-Bar and Hoffmann (2000).
Fig. 4.23 gives the PET values for a N-S street with H/W = 1 including a large central
row of trees. For comparison, a similar case without trees is shown in Fig. 4.8b. In this
case PET was up to 24 K lower than in a street without trees. One can see that the best
screen effects of the vegetation occurs on in the central part of the vegetated area
whereas shortly less effective at the edges when the optical length is minimal and re-
sults in low LAI*, e.g. grid No 6 around 14:00 LST. This is explicit in Fig. 4.24 which
compares between the individual irradiance terms accounting for the energy gained by a
standing person, for a central grid point (x = 6 m) in a N-S street when planted or not.
Table 4.4 lists these values for the most critical daytime hours, i.e. from 11:00 to 14:00
LST.
PET, H/W = 1 with a central row of trees, N-S
20:00
19:00
62 °C
18:00
58 °C
17:00
16:00 54 °C
15:00
time (LST)
50 °C
14:00
46 °C
13:00
12:00 42 °C
11:00
38 °C
10:00
34 °C
9:00
8:00 30 °C
0m 1m 2m 3m 4m 5m 6m 7m 8m 9m 10m 11m 12m 13m 14m
|----- gal. W -----|--------------- street width ----------------------|---- gall. E -----|
Fig. 4.23. PET pattern within a street oriented N-S with H/W = 1 and a large central
row of trees (− − − limit of gallery, ------ projection of trees’ area)
123

600
LW_total, no trees
500
radiation flux (Wm )

-2 400
LW_total, with trees
300
SW_total, no trees
200
100 SW_total, with trees
0
8 9 10 11 12 13 14 15 16 17 18 19 20
400
SW-dif. SW-dir. LW-grd. LW-sky LW-bldgs.
350
SW-dif., tree SW-dir., tree LW-grd., tree LW-sky, tree LW-bldgs., tree
300
radiation flux (Wm-2)
250
200
150
100
50
0
8 9 10 11 12 13 14 15 16 17 18 19 20
time (LST)
Fig. 4.24. Individual short-wave (SW) and long-wave (LW) energy terms absorbed by a
standing person located in a N-S street with H/W = 1 without vs. with trees
Table 4.4. Individual short-wave SW and long-wave LW energy terms (Wm-2) absorbed
by a standing person at the most stressful hours in a N-S canyon of H/W = 1 with and
without trees for grid No 6
Σ Σ Σ
no trees
11:00 171.1 116.9 287.9 281.3 177.3 49.5 508.1 796.0 74.7 35.7 59.9
12:00 126.1 120.7 246.8 294.2 182.1 49.5 525.7 772.5 72.2 36.9 58.8
13:00 107.5 121.7 229.1 301.6 186.1 49.5 537.2 766.3 71.5 38.0 58.7
14:00 138.9 119.8 258.8 304.2 187.4 49.5 541.1 799.8 75.2 38.5 61.5
with trees
11:00 35.7 116.9 152.6 236.6 42.9 49.4 328.9 481.5 44.3 34.4 39.7
12:00 29.6 120.7 150.4 240.3 43.8 49.4 333.6 483.9 44.9 35.3 40.5
13:00 26.0 121.7 147.7 243.6 44.6 49.4 337.7 485.4 45.3 36.2 41.2
14:00 99.9 119.8 219.7 251.7 45.1 49.5 346.3 566.0 62.2 36.8 51.3
124
Basically, the human body absorbs up to 135 Wm-2 less short-wave irradiance mainly in
form of less direct solar radiation gain and up to 199 Wm-2 less long-wave radiation,
mostly in form of less downwards radiant heat from the free atmosphere (∼140 Wm-2)
and outgoing from the ground (50 Wm-2). The diffuse irradiation is kept unchanged by
the model to replace the direct radiation which would be converted into diffuse radiation
within the crowns and not considered in the calculations.
The previous cases showed that a N-S orientation allows more sun within the galleries
than an E-W orientation, in particular for large aspect ratios. Fig. 4.23 shows, in addi-
tion, that a central row of trees does not protect the galleries better, which still experi-
ence about 2 hours of highest discomfort on each gallery in the morning and in the late
afternoon, respectively. This suggests that planting on street edges would be preferable
in case of wider canyons for further protecting the sidewalks and galleries.
Furthermore, the extent to which a tree is an efficient strategy for mitigating the heat
stress depends on its density (LAD, LAI) and geometry (dimensions). Light density
trees normally allow less shading but more air circulation under the crown than a dense
crown tree. Test simulations (not shown here) were also made for light and dense crown
trees as well as for a parallel wind incidence (e.g. channelling in-canyon) in order to
assess whether promoting shading is more critical than allowing more ventilation or
inversely. PET results showed small differences between the two cases, with a minimal
advantage for a dense tree, for which PET values are about 2 to 4 K lower during one
hour than under light-dense trees.
4.6. Role of the wind
The issue of urban wind is complex and a detailed analysis of the wind flow mecha-
nisms is not possible within the framework of this study. Nevertheless, a number of ob-
servations of relevance for the comfort issue are summarized below. These deal mainly
with the effects of the incidence of above-roof wind in relation to street axis (perpen-
dicular or parallel) on the near ground wind speed in the canyon.
Unlike other meteorological factors (e.g. Ta, VP), the wind speed (v) shows large differ-
ences along the street, namely a strong contrast between street ends and street centre.
Exemplarily, Fig. 4.25 compares the wind speed between parallel and perpendicular
incidence for a wide canyon (H/W =0.5) and for a deep canyon (H/W = 2).
125
Ghardaia, 32.40° N, 3.80° N, 01August
(a)
v (m/s), H/W = 0.5, parallel wind v (m/s), H/W = 0.5, perpendicular wind
8m 8m
6m 6m
4m 4m
2m 2m
|------ street width ----|

0m 0m
0m 4m 8m 12 m 16 m 20 m 0m 4m 8m 12 m 16 m 20 m
|--------------------- street lenght ------------------------| |--------------------- street lenght ------------------------|
(b)
v (m/s), H/W = 2, parallel wind
8m
3.2 m/s
6m
126
4m 2.8 m/s
2m
2.4 m/s

0m
0m 6 m 12 m 18 m 24 m 30 m 36 m 42 m 48 m 54 m 60 m 66 m 72 m 78 m 84 m 90 m 2.0 m/s
v (m/s), H/W = 2, perpendicular wind
8m
1.6 m/s
6m 1.2 m/s
4m
0.8 m/s
2m

0m 0.4 m/s
at 1.2 m a.g.l. level for both perpendicular and parallel wind incidence on street axis
0 m 6 m 12 m 18 m 24 m 30 m 36 m 42 m 48 m 54 m 60 m 66 m 72 m 78 m 84 m 90 m
|--------------------------------------------------------- street lenght --------------------------------------------------------------| 0.0 m/s
Fig. 4.25. Mean wind velocity within urban canyons of (a) H/W= 2 and (b) H/W = 0.5,
1. Perpendicular flow: For wide canyons, i.e. H/W = 0.5 (Fig. 4.25a), v is strongly re-
duced at street level in the whole canyon area and is about 0.1 ms-1. In fact, the street
canyon is rather isolated from the ambient air from above roof (skimming flow) and
even the perturbation zone at the canyon corners is very limited, i.e. about 10 % of the
buildings length with at most 0.4 ms-1 for H/W = 0.5. This is likely attributable to the
small height of the walls which limit their role in deflecting the flow inside the canyon
and hence minimize the advection toward the mid-canyon zone.
With increasing aspect ratio (Fig. 4.25b) the areas on street corners become strongly
influenced by the main flow and experiences noticeably higher wind speeds than the
centre of the canyon, namely up to 1.5 ms-1 on the windward side. This is due to inter-
mittent vortices which are responsible for the mechanism of advection from building
corners to mid-block canyon creating a convergence zone in the mid region of lowest
wind speeds (Hoydysh and Dabbert 1988, Santamouris et al. 1999). The strength of this
advection at street level increases with the aspect ratio until a critical value as shown in
Fig. 4.26. Explicitly, near ground wind speed rises with H/W until a proportion of 3 ,
but decreases again for H/W = 4, at which the strong eddy circulation deviating the
wind flow from the upper corners through the canyon occurs only at the higher part of
the street (v > 1.2 ms-1 only from 0.5 H) and does not reach the street level, leading to
lower wind speeds.
The mid-canyon is characterized by a convergent flow from both sides and ranges be-
tween 0.3 and 0.1 ms-1 for all canyons regardless of their aspect ratio. The symmetry of
these patterns is slightly altered, likely because of thermal effects (Santamouris et al.
1999).
2. Parallel flow: The wind is channelled along the street and flows in the same direc-
tion. Because of friction forces near the surfaces, the wind speed decreases progres-
sively along the street with an uplift along the canyon walls until a critical point within
the canyon depending on H/W. The flow is then accelerated again, pushed by the down-
ward flow from roof level. Explicitly, with increasing H/W ratio the canalisation and
acceleration effects are stronger and the wind speed at the opposite corner experiences
higher wind speeds than at the entrance of the canyon, e.g. 3.4 ms-1 vs. 3.0 ms-1 for H/W
= 2 (Fig. 4.25b). The extent of the acceleration zone also increases, e.g. 20 % for H/W =
0.5 vs. 60 % for H/W = 2.
A sensitivity analysis for summer conditions with dominating mean radiant tempera-
tures and high air temperatures revealed that PET can decrease by 8 % for an increase in
127
v of 1 ms-1 and 11% for an increase of 3 ms-1. Figs. 4.27 and 4.28 compare PET for both
wind incidences at mid-distance of the canyon in 2 case studies: in a canyon of H/W = 2
oriented E-W and in the same street but including a row of trees. In Fig. 4.27, v at 1.2 m
height ranges between 0.1 and 0.3 ms-1 for a perpendicular wind incidence against 2.2
and 2.9 ms-1 in the course of the day. In case of a street with a row of trees (II-4, dense
with leafless base), the wind speed at street level is slightly lower than without trees for
both wind incidences, namely v ranges between 0.1 and 0.2 ms-1 for a perpendicular
wind and between 1.5 and 2.6 ms-1 for a parallel wind.
First, it can be seen that the effects of wind speed vary throughout the day and across
the street suggesting that the importance of v in mitigating the heat stress depends also
on the absolute values of Tmrt and Ta. The decrease in PET, due to wind speed, ranges
between 2 K and 12 K for lower values when the wind is parallel and hence stronger.
Fig. 4.26. Zones with different ventilation potential and depending on canyon dimen-
sions according to simulation results
128

(a) PET, H/W = 2, parallel wind, E-W (b) ∆ PET, parallel vs. perpendicular wind
20:00 20:00
19:00 19:00
18:00 18:00
17:00 17:00
16:00 16:00
15:00 15:00
time (LST)
time (LST)
14:00 14:00
13:00 13:00
12:00 12:00
11:00 11:00
10:00 10:00
9:00 9:00
8:00 8:00
|---------------- street width -------------------| |---------------- street width -------------------|
Fig. 4.27. (a) PET pattern for an E-W street of H/W = 2 for a parallel wind incidence,
(b) PET between parallel and perpendicular wind for the same canyon (see Fig. 4.5c)

PET, H/W = 2 with trees, parallel wind, E-W ∆ PET, parallel vs. perpendicular wind
20:00 20:00
19:00 19:00
18:00 18:00
17:00 17:00
54 °C
16:00 16:00
50 °C
15:00 15:00
time (LST)
time (LST)
14:00 46 °C 14:00
13:00 42 °C 13:00
12:00 12:00
38 °C
11:00 11:00
34 °C
10:00 10:00
9:00 30 °C 9:00
8:00 26 °C 8:00
|---------------- street width -------------------| |---------------- street width -------------------|
Fig. 4.28. (a) PET pattern for an E-W street of H/W = 2 including a row of trees (dense,
leafless base) for a parallel wind incidence, (b) ∆PET between parallel and perpendicu-
lar wind for the same canyon (see Fig. 4.21), negative values mean lower PET’s
129
The lowest impact is recorded on the shaded areas where Tmrt is minimal and close to
Ta, namely on the north side and after 18:00 LST. In contrast, PET is up to 12 K slowed
down at time periods where Tmrt is maximal combined with highest air temperatures
(16:00 to 17:00 LST). Statistically, a linear relationship was found with a determination
coefficient r2 = 0.9843, namely PET (paral. wind) = 0.648 PET (perp. wind) + 11.7.
Unlike Ta which shows almost no change along the street, the wind speed is suggested to
bring more variability in space in the thermal comfort according to the location within
the street.
This influence is much more perceptible for a parallel wind than a perpendicular wind
since the wind velocities are higher in the former case, e.g. for H/W = 2 for a parallel
wind v varies between 2.4 ms-1 and 3.4 ms-1 against a range of 0.1 ms-1 to 1.4 ms-1 for a
perpendicular incidence. Yet, thermal comfort close to the street corners is also charac-
terized by a greater potential of solar exposure which might reduce the advantages ob-
tained by a stronger wind flow. This calls attention on the necessity of assessing com-
fort in the particular case of street intersections.
4.7. Solar access in summer and winter
Street design affects not only the outdoor microclimate: urban geometry choices are
often also motivated by solar access purposes in indoor spaces in winter (see 2.1.4 and
2.1.5). Therefore, a comparison of the thermal comfort outdoors and solar access in-
doors turns out to be necessary for a complete evaluation of the climate efficiency of
any street design solution.
Increasing the aspect ratio obviously leads to less potential of solar irradiation of the
facades and wall orientation according to sun exposure is just as important. Walls facing
south are preferred for optimal solar gains in winter and easy solar control in the sum-
mer, east and west are alternatives but with a number of disadvantages in comparison to
the south, and the north almost receives no solar beam (e.g. Givoni 1976, Arnfield
1990a, Lechner 1991, Bourbia and Awbi 2004, etc.).
The precedent analysis showed that outdoor comfort in the summertime is efficiently
guaranteed by increasing the aspect ratio to a proportion of 4 or more in the subtropics,
especially for E-W orientations. This is, for instance, the typical solution adopted in old
desert cities (see chapter 6). Such a design assumes that winter sunlight and daylight
inside the buildings are ensured through internal courts or patios, but this has been ques-
130
tioned (e.g. Ouahrani 1993). In case of conventional sun-lighting and day-lighting

through external façades, a ratio H/W ≥ 4 will compromise the winter solar passive
gains because of the low sun position. Moreover, the dilution of pollutants in deep can-
yons can be strongly reduced if any source of pollution at street level exists (e.g. motor
traffic). Thus, this kind of street is more appropriate as pedestrian paths in housing areas
of a dense urban plan. If all these goals are considered, wider streets have to be pre-
ferred together with complementary shading devices as discussed previously.
The solar access index (SAI) has already been proposed as a useful indicator for design
purposes (Arnfield 1990a). SAI is defined as the actual direct solar radiation received by
an urban surface (ground floor or façades) reported to the maximum potentially avail-
able direct solar radiation on an unobstructed surface. Similarly, one can also use a
shading factor SF as index, which is the complementary fraction of SAI as used by oth-
ers (e.g. Kristl and Krainer 2001, Bourbia and Awbi 2004). These indices help to find
an appropriate geometry for both summer and winter needs at early design stages.
The following graphics show the variability of SAI in relation to aspect ratio (from 0.5
to 4.5) for E-W and N-S orientations for summertime and wintertime. Fig. 4.29 shows a
diurnal evolution and Fig. 4.30 a spatial evolution across the street. Obviously, very
different SAI patterns are found between E-W and N-S orientations. The seasonal solar
access potential at street level varies much more for E-W than for N-S oriented streets
(Fig. 4.29). The floor area of E-W oriented streets of H/W = 0.5 receives up to 50%
solar irradiation and at most 20% for H/W = 1, against 90 % in the summer for H/W ≤
1. For N-S streets, the exposure potential does not vary significantly throughout the
year, and the floor area of N-S streets is also irradiated at midday hours in the winter
even for deep canyons. This suggests that N-S orientation is more appropriate for pedes-
trian use as the outdoor comfort is more probable even for winter season when sun ex-
posure is desirable. By contrast, an E-W orientation is less appropriate since irradiated
in the summer and shaded in the winter, leading to discomfort in both cases.
Fig. 4.30 shows that winter sun exposure of the street area in an E-W oriented street is
differentiated between the north part which receives more solar energy than the south
part for canyons up to H/W = 1. For larger proportions, SAI equals 0.2 in the whole
area.
131
Fig. 4.29. Dependence of the solar access index SAI on the aspect ratio H/W at street
level for (a) summer conditions and (b) winter conditions
132
Fig. 4.30. Dependence of the SAI on the aspect ratio across the street space for E-W and
N-S oriented streets in winter
133
The summer exposure for the same orientation shows a particular case of SAI = 1.1
which corresponds to more irradiation of the canyon floor than a horizontal unob-
structed surface and attributable to additional diffusely reflected radiation from the ver-
tical surfaces. A N-S street shows the same exposure patterns on both sides of the street
with almost the same seasonal, e.g. about 0.4 and 0.5 for H/W = 1 in the summer and in
the winter respectively.
Fig. 4.31 compares between the solar access potential on the walls (by implication in-
doors) for both orientations in the winter. It appears clearly that the availability of solar
energy on the facades decreases very rapidly with the increase of H/W for E-W streets.
Moreover, the walls receive less solar energy in the morning (before 11:00 LST) and
afternoon hours (after 15:00 LST) with increasing aspect ratio. While the irradiation of
the façades is optimal for H/W = 0.5, it is only of 0.2 to 0.4 for H/W = 2 and only about
0.1 for deeper streets. For wide canyons the walls receive solar energy during all the
sunshine period.
Ghardaia, 32.40° N, 3.80° E, 01 January
(a) S.A.I., Walls, Winter, E-W (b) S.A.I., Walls, Winter, N-S
20:00 20:00
19:00 19:00
18:00 18:00
17:00 17:00
16:00 16:00
15:00 15:00
time (LST)
time (LST)
14:00 14:00
13:00 13:00
12:00 12:00
11:00 11:00
10:00 10:00
9:00 9:00
8:00 8:00
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
|-------------- aspect ratio (H/W) -------------| |-------------- aspect ratio (H/W) -------------|
Fig. 4.31. Dependence of solar access index on aspect ratio for (a) E-W and (b) N-S
oriented street
134
Hence, a street with H/W = 2 seems to be inappropriate for winter needs and only partly
efficient for summer comfort outdoors, suggesting that wider streets (e.g. I-1 or I-2 in
Fig. 3.3) or asymmetrical street (e.g. II-2 or II-3 in Fig. 3.3) have to be preferred if in-
ternal solar gains in winter are of prime importance. If the indoor issue is less relevant,
deeper canyons (e.g. I-4 and I-5 in Fig. 3.3) are more advisable for comfort for the hot
season.
The exposure of the walls for N-S orientation shows a different pattern: Increasing the
aspect ratio leads to a shorter time of exposure around noontime. Even deep canyons of
H/W > 3 receive maximal energy during one hour, and about one half (0.5) of the po-
tentially available energy for at least 2 hours. Façades of wider streets are exposed to the
sun much longer, e.g. up to 6 hours of full exposure for H/W = 0.5. Yet, this exposure
potential is also valid in the summer as can be deduced from Fig. 4.29 and suggests
necessary complementary solutions for summer shading. The use of asymmetrical can-
yons e.g. II-2 or II-3 with the high façade oriented to the east is a good alternative to
take advantage of more exposure in the winter and at the same time offer more shading
from west exposure in the winter.
These values must, however, be appreciated according to the actual values because
these indices do not reveal the effect of the incidence angle of the direct solar beam on
the façade according to its orientation on the amount of energy received by the surface.
135
136
5. Field measurements in Freiburg, Germany
5.1. Site and observations
Meteorological measurements were conducted in an urban canyon (Erbprinzenstraße)

in the downtown of Freiburg, a medium-sized city in the southern upper Rhine plain in
south-west Germany. Freiburg is located at 48 °N, 7° 50´ E and 280 m above sea level.
The canyon axis is oriented in east-west direction. The aim of these measurements was
to assess experimentally the effects of the canyon geometry and orientation on the street
microclimate, on the heat gained by a pedestrian and the resulting thermal comfort.
The canyon axis is oriented in east-west direction (Fig. 5.1). The street is flanked by
long buildings, which preserve the canyon alignment for at least 150 m, despite the
presence of a number of gaps in the building’s fronts. At the measuring site, the canyon
is symmetric with an aspect ratio H/W = 1 and a sky view factor SVF = 0.26 (Fig. 5.2).
The buildings are almost of equal height, typically of two or three stories with pitched
tile roofs. The street is made of asphalt and is 12 m wide.
Fig. 5.1. Plan view of the east-west canyon street in Freiburg with the location of the
permanent station and the measuring points MP1 to MP4
137
Fig. 5.2. Fish-eye photography of the canyon at the station location, Freiburg
The walls of the building are made of bricks and painted with light colours. Windows
constitute about 30% of the walls. A small park with tall trees is located in the vicinity
of the canyon on the west side. The east end of the canyon opens onto a small planted
place while the west ends onto a main north-south road. Some sparse vegetation along
the street is also noticeable.
The experimental work was conducted on 14 and 15 July 2003: two sunny and hot days.
Although the period of data collection was short, the prevailing conditions on these two
days where considered representative of typical summer in Freiburg. A vertical mast
fitted with temperature, wind and radiation sensors was installed at a distance of 1 m
from the northern wall. This location corresponds to the pedestrian sidewalk where
comfort is required. For this orientation E-W, this is also the most critical location in
relation to comfort as reported in chapter 4 for the subtropics (see also Ali-Toudert and
Mayer 2006). Air temperature, air humidity, wind speed and wind direction were con-
tinuously recorded at regular time intervals at two heights: 1.4 m and 3.1 m. The short-
wave and long-wave radiation flux densities were measured from the three-dimensional
surroundings i.e. upwards and downwards together with the four lateral directions (N,
E, S and W). All factors were recorded in form of 10-minute-averages (scan interval: 10
s) over a 30-hour-period. The instrumentation used in this study is listed in Table 5.1.
138
Table 5.1. Instrumentation used at the station within the street canyon in Freiburg
Item Unit Instrument height (a.g.l) number
-2
(1) Global radiation K Wm Pyranometer, CM21, CR11, Kipp 1.4 m 6 ( ↑, ↓, N, E, W, S )
& Zonen
-2
(2) Net radiation Q Wm Pyradiometer, Schenk 1.4 m 6( ↑, ↓, N, E, W, S )
-2
(3) Long-wave radiation Wm difference between all-wave and 1.4 m -
L short-wave radiation (Q – K)
(4) Air temperature Ta °C PT100, HMP Vaisala 1.4 m, 3.1 m 2
(5) Air humidity VP hPa PT100, Humicap, HMP Vaisala 1.4 m, 3.1 m 2
(vapour pressure)
(6) Wind speed v ms-1 Cup anemometer, Vector Instru- 1.4 m, 3.1 m 2
ments
(7) Wind direction dd ° Wind vane, Vector Instruments 3.1 m 1
Fig. 5.3. Set of radiation sensors for the measurement of the global radiation from the
3D surroundings within the urban canyon in Freiburg
Such extensive measurements of the radiation fluxes are required for an accurate calcu-
lation of the mean radiant temperature Tmrt (see equations 2.7 and 2.8), according to the
method proposed by Höppe (1992) and VDI (1998), see Fig. 5.3.
Other supplementary readings collected on 14 July included manually taken measure-
ments of air and surface temperatures on both sides of the street (MP1 to MP4, Fig.
5.1). This allows one to get a spatially differentiated picture of the street microclimate.
In addition, the data obtained in the street were compared to those provided by a perma-
nent urban climate station in order to clarify the microclimatic changes within the can-
139
yon. This “background station” is run by the Meteorological Institute, university of

Freiburg, and is located on the roof of a high-rise building at a height of 51 m above
ground level (MIF 2005). It is situated in the northern part of Freiburg at about 1500 m
far away from the investigated street. Air temperature and humidity sensors were placed
2 m above roof level (a.r.l) and the wind sensor at 10 m a.r.l. The data of the “back-
ground station” were compared to those obtained at street level in order to clarify the
microclimatic changes inside the canyon due to the obstructing effects of the buildings.
5.2. The microclimate in the canyon

5.2.1. Air and surface temperatures
Fig. 5.4 shows the daily course of the air temperature Ta as recorded by the fixed sta-
tions in the canyon and above-roof, together with supplementary readings measured
manually at the four additional points along the sidewalks.
Basically, Ta within the canyon varies between 18 °C and 35 °C, which is a much wider
range in comparison to the average monthly values for Freiburg (i.e. 18 °C to 25 °C).
This depicts the record-breaking heat-wave which affected Europe in summer 2003 dur-
ing which a maximum temperature of 40.2 °C was reached in Freiburg.
Freiburg, 14/15 July 2003

36
32
28 Ta_3.1m
Ta (°C)
Ta_1.4m
24 Ta_roof
Ta_MP1
Ta_MP2
20
Ta_MP3
Ta_MP4
16
8 12 16 20 0 4 8 12 16 20
time (LST)
Fig. 5.4. Daily variation of air temperature Ta in the urban canyon on a cloudless sunny
day in Freiburg
140
In the canyon, there was little difference in Ta measured at various points before 13:00
LST and after 18:00 LST, owing to the well mixed air inside the canyon. Compared to
Ta above roof level, one can also see that almost no difference is found during the period
between 8:00 and 13:00 LST, during which time the street is yet to warm up. In the af-
ternoon on 14 July, from 14:00 to 18:00 LST, Ta measured at the sunlit part of the
street, i.e. at the fixed station, MP3 and MP4, was a few degrees higher than those re-
corded on the opposite side of the street (MP1 and MP2) which were mostly in shade.
The urban canyon surfaces facing south are most of the daytime irradiated and experi-
ence high Ta values (Fig 5.5a, calculated from L with ε = 0.98), leading to increased
heat transfer to air as sensible heat flux.
By contrast, air temperatures at the north facing side show lower values due to the lim-
ited heat released by the adjacent surfaces, as these have noticeably lower temperatures
(Figs. 5.5a and 5.5b). Maximum Ta difference between both sides is reached at 17:00
LST and is of approximately 3 K for the same height (1.4 m). Moreover, Ta measured at
the fixed station (1.4 m a.g.l.) is 1.2 K higher than the one measured at 3.1 m height, as
a consequence of increased proximity to the sunlit and warm ground and walls. The
differences are negligible in the evening hours and do not exceed 0.5 K.
The results of Ta are in good agreement with previous studies conducted for streets with
almost the same characteristics: E-W orientation, H/W ≈ 1, and latitude 35 °N. Naka-
mura and Oke (1988) found that there was small difference between the temperatures of
the air in the canyon and that at the roof, except close to sunlit urban facets where the
heat transferred from the heated walls leads to warmer adjacent air. Yoshida et al.
(1990/91) confirmed the insignificant warming of the canyon air in comparison to free
ambient air and the homogeneity of Ta across and along a street canyon. They also re-
ported on large differences in the surface temperatures between sunlit and shaded sur-
faces. Surfaces in shade are noticeably cooler than irradiated surfaces, and the surface
temperatures can even be lower than Ta. Santamouris et al. (1999) report of almost simi-
lar findings for a deeper street oriented close to N-S, but with a slight thermal stratifica-
tion. As well, air temperature differences of up to 2 K were found between irradiated
and shaded sidewalks in various urban canyons under hot summer conditions (e.g.
Mayer and Höppe 1987, Nakamura and Oke 1988).
In the night, the street is cooler than the free air above roof and the difference reaches a
maximum ∆Ta of 3 K. This feature seems anomalous and could only be attributed to the
141
different microclimates within the city of Freiburg as reported by Nübler (1979) and
hence between the canyon station and the background station. Indeed, the urban canyon
under study is on the passage of a local circulation system from the Black Forest (so-
called Höllentäler) driven down a valley and leading to a noticeably cooler night situa-
tion, while the urban climate station at roof level is located in a typical heat island zone.
Freiburg, 14/15 July 2003

60
Tground Twall Ta_3.1 m Ta_1.4 m
55
50
45
Ta, Ts , Tw (°C)
40
35
30
25
20
15
8 12 16 20 0 4 8 12 16 20
time (LST)
60
50
40
Ts, Tw (°C)
30
20 Ts MP1 Tw MP1 Ts MP2 Tw MP2
Ts MP3 Tw MP3 Ts MP4 Tw MP4
10
10 12 14 16 18 20 22
time (LST)
Figs. 5.5a and 5.5b. Daily variation of (a) air temperature Ta, ground temperature Ts and
wall temperature Tw at the station on the north side of the street and (b) Ts and Tw at
points MP1, MP2 (southern side) as well as MP3 and MP4 (northern side) on 14 July
2003
142
5.2.2. Wind direction and wind speed
The relationship between wind flow above-roof and within the canyon, discussed be-
low, should be read bearing in mind the following uncertainties: The background station
is about 1500 m far away from the canyon studied and the data describing the above-
roof wind conditions were recorded at 61 m a.g.l. This is about four times the canyon
height and as a result the wind speed directly at roof level (13 m a.g.l.) is much lower.
Moreover, the discontinuity of the building fronts may have influenced the wind flow at
the typical low speeds recorded at street level.
The air flow in the canyon is known to be a secondary circulation feature driven by the
above-roof imposed flow (e.g. Nakamura and Oke 1988, Santamouris et al. 1999). In
this study, the correlation between the canyon wind speed with that above roof is found
to be more marked for high wind speeds, whereas the coupling between the upper and
secondary flow is lost for lower velocities, leading to much more scattering. The wind
direction in the canyon was found to depend on that above roof, more precisely on the
angle of incidence of the upper wind with respect to the canyon’s axis.
During the measuring period, the wind was either parallel or oblique. Almost no per-
pendicular incidence was recorded. Three distinct and temporarily consecutive episodes
could be observed, with remarkably different combinations of wind directions and
speeds, which have in turn influenced the wind flow characteristics within the canyon
(Figs. 5.6 and 5.7). When the wind above-roof is nearly parallel to the canyon axis (+/-
30°), the wind in the canyon flows in the same direction due to channelling. In this case,
it corresponded to a local wind called “Höllentäler” which blew from the east during the
night from 20:00 to 6:30 LST. On the first day, from 12:00 LST to 20:00 LST, the wind
was blowing at an angle of incidence with moderate velocity: from NW quadrant and
faster than 5 ms-1. This led to a wind inside the canyon flowing in the SE direction.
This flow scheme has been described as a spiral vortex induced along the canyon (e.g.
Wedding et al. 1977, Nakamura and Oke 1988, Santamouris et al. 1999). The simple
relationship suggested by Nakamura and Oke (1988) for an urban canyon with H/W ≈ 1
(see section 2.1.3, p. 33) seems to apply to the present case study as a first approxima-
tion.
143
Freiburg, 14./15.7.2003
360
270
wind direction (°)
180
90
at 51 m height
in-canyon at 3.1 m height
0
10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18
time (LST)
Fig. 5.6. Wind direction within the street canyon and above roof level (at 61 m a.g.l.)
12
v at 51 m height
v in-canyon at 3.1 m
10
v in-canyon at 1.4 m
8
wind speed (m/s)
0
10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18
time (LST)
Fig. 5.7. Wind speed within the canyon street and above roof level (at H = 61 m)
4 360
from E (70° to 125°) above 5 m/s
wind speed in-canyon (z = 3.1 m a.g.l.) ms-1
below 5 m/s
dd in-canyon (z = 3.1 m a.g.l.) (°)
from NW (270° to 335°)
3 other directions R2 = 0.6211 270
R2 = 0.4593
2 180
1 90
0 0
0 2 4 6 8 10 12 0 90 180 270 360
-1
wind speed above roof (z = 51 m a.g.l.) ms dd above roof (z = 51 m a.g.l.) (°)
(a) (b)
Fig. 5.8. Wind speed and wind direction dd outside the canyon plotted against inside
corresponding conditions
144
From 10:00 to 13:00 LST on 14 July and after 6:30 LST on 15 July, weak winds with
no dominant direction prevailed above roof level. This led to a large scattering in the
canyon. In this case, the wind flow in the canyon was not only a mechanically driven
circulation but thermal effects may have also played a role (e.g. Nakamura and Oke
1988, Sini at al. 1996, Santamouris et al. 1999) especially at the sunlit part of the street
where the wind was measured.
Neither the direction nor the speed of the winds, within and without the canyon, was
found to be clearly correlated as suggested by Nakamura and Oke (1988). It is quite
noticeable that low-speed winds in the canyon tend to north eastwards. The spacing lo-
cated near the station on the north side may have influenced the wind direction, so that
air flowed between the two buildings.
From Fig. 5.8a, we find that there is a linear relationship between the speed of winds in
the canyon and that of winds above the roof. Generally winds above the roof move
faster relative to those in canyon. A linear regression line fitted to the data for winds
blowing from E (70° to 125°) and from northwest NW (270 to 335) had a coefficient of
determination R2 = 0.62. This means that approximately 62% of the variation in the
speed of winds, from E and NW, in the canyon is accounted for by the movement of
winds at roof level. Similarly, considering winds from other directions, the coefficient
of determination was R2 = 0.49. The inference made here is that a greater proportion of
the variations in the speed of winds from other directions, in the canyon, cannot be ex-
plained by the dynamics of the winds above the roof. It was also observed that winds
from the east (E) were faster both in the canyon as well as at the roof level. However,
from our study we could not confirm the simple linear relationship suggested by Naka-
mura and Oke (1988) between the speed of wind in the canyon and at the roof level.
Some studies suggest the existence of a threshold above which a coupling between the
wind outside and inside the canyon may take place (e.g. de Paul and Shieh 1986, Na-
kamura and Oke 1988). In this study, a wind speed of 5 ms-1 (measured at 61 m) may be
considered as threshold as shown in Fig. 5.8b, above which correlation is found be-
tween inside and outside wind direction whereas much more dispersal is observed be-
low this limit. By invoking the power law of wind profile, the corresponding wind
speed directly above roof level (at 13 m a.g.l.) could be approximated to 2 ms-1, which
agrees with estimates from previous studies.
145
5.3. Thermal comfort analysis

5.3.1. Short-wave radiation fluxes
The aspect ratio (H/W = 1) and the street orientation (E-W) together with the day of
year and latitude are responsible for the solar exposure patterns prevailing within the
canyon at street level. Fig. 5.9 shows the simulated short-wave radiation fluxes (imping-
ing on a normal surface) for the street area investigated, and including the actual build-
ing heights with a spatial resolution of 2 m.
The north wall is almost permanently irradiated as well as about one half of the canyon
floor, whereas the south wall and opposite half part of the street surface are mostly
shaded. The buildings block the sun rays at the northern side before 7:00 LST and after
18:00 LST. During these periods, the opposite facets are shortly irradiated as the sun
crosses over the street. These patterns help to understand the following results relating
to human heat gain within the canyon.
Fig. 5.9. Temporal and spatial distribution of short-wave radiation in Wm-2 (normal to
surface) across the street, simulated by ENVI-met 3.0 (---- location of the measuring
station)
146
Fig. 5.10 illustrates the short-wave radiation fluxes (K) from the six directions on a per-
son standing near the north wall. The importance of the orientation and the location
within the canyon is evident. The irradiation from above (K↓) recorded in the canyon
has a daily course comparable to a location with a free horizon, i.e. the reference city
station (K↓roof). The only exceptions are before 9:00 LST and after 17:00 LST where
the façades obstruct the direct solar beam. In the morning, the solar radiation comes
from the south-east quadrant. Hence, K↓ is already by 11:00 LST at 800 Wm-2 and the
radiation fluxes coming from the east K(E→) and form the south K(S→) are also rela-
tively high, about 500 Wm-2 and 300 Wm-2, respectively. The maximum values of K↓
and K(S→) are recorded around 13:00 LST as the sun reaches its highest position (63°)
with the sun rays coming exactly from the south. Both street and north wall are then
irradiated at maximal angle of incidence.
The radiation fluxes registered parallel to the street axis (K(W→) and K(E→)) are
symmetrically reversed as the sun moves from the south-east quadrant to the south-west
quadrant. Explicitly, K(W→) and K(E→) undergo the same pattern in the morning and
afternoon respectively. A sharp increase in the energy received from both directions is
then recorded as the sun rays become parallel to the street axis, i.e. around 9:00 LST for
K(E→) and around 17:00 LST for K(W→) with a maximum value of 650 Wm-2. In con-
trast, values not exceeding 100 Wm-2 are measured before 13:00 LST for K(W→) and
after 13:00 LST for K(E→). These correspond to diffuse and diffusely reflected solar
radiation components. Similarly, the short-wave radiation upwards K↑ as well as from
the sunlit wall K(N→) do not exceed 120 Wm-2 and correspond to reflected irradiation
from the street and wall, that have a mean albedo of 0.15 and 0.13 respectively.
5.3.2. Long-wave radiation fluxes
The solar exposure patterns influence the long-wave radiation fluxes (L) as shown in
Fig. 5.11. The asphalt road is mostly irradiated during the daytime and constitutes the
highest source of long-wave irradiance within the canyon at street level (L↑).
The peak value occurs at 15:00 LST and exceeds 600 Wm-2, whereas the lowest value is
recorded at 6:00 LST and equals 376 Wm-2. The radiant heat from the south-facing wall
L(N→) is also substantial as it is mostly irradiated. L(N→) shows a comparable tempo-
ral evolution as L↑, though of less magnitude.
147
Erbprinzenstraße, Freiburg, 14/15 July 2003

1000
K↓roof
800 K↓
600
K (Wm-2)
K(W→) K(E→)
K(E→)
K(S→)
400
200 K(N→) K↑
0
8 12 16 20 24 4 8 12 16 20
time (LST)
Fig. 5.10. Short-wave radiation fluxes (K) received from the six directions surrounding
a standing person located at the north side of an E-W oriented street with H/W = 1

700
L↑
600 L(W→) L(N→)

L (Wm-2)
500
L↓
L(S→)
L(E→)
400
300
8 12 16 20 24 4 8 12 16 20
time (LST)
Fig. 5.11. Long-wave radiation fluxes (L) received from the 6 directions surrounding a
standing person located at the north side of an E-W oriented street with H/W = 1
148
Indeed, being a vertical surface, it receives less short-wave radiation than the horizontal
ground surface. Moreover, the asphalt pavement heats much more than the light col-
oured brick walls. After 17:00 LST, the rate of heat release is slowed down as the can-
yon surfaces become shaded and thus cooler.
The radiation flux densities from east L(E→) and west L(W→) are composed of the
heat released from the ground surface, the walls and the atmosphere. Yet, the influence
of the north wall may have been dominant as the measurement was conducted close to
it. In the morning, the amount of radiant heat from both directions is almost equal to
that emitted by the north wall until 13:00 LST and shows approximately the same in-
creasing trend. However, the east side releases slightly more heat in the early afternoon
while for the west side it happens in the late afternoon, due to the sun exposure patterns
previously mentioned.
Because of the location of the radiation sensor close to the north wall, the heat emitted
by the atmosphere together with the street surface and the south wall constitute a large
part of L(S→). At 11:00 LST, it is clearly lower than the other fluxes. The reason to this
is the relatively low long-wave atmospheric radiation L↓ combined with the small
amount of heat released by the opposite shaded wall. There is a rapid increase in the
afternoon. The peak value is recorded at 15:00 LST as a result of the accumulated heat
stored in the ground along with the late additional exposure of the whole street surface
and the south façade. The long-wave irradiation slump from 18:00 LST when all canyon
facets become shaded. The slight increase at 8:00 LST of L(S→) can be explained by
the short exposure time of the southern side of the street in the early morning.
In the night-time, the street surface and the north wall remain the main sources of heat
and show an almost equal nocturnal cooling rate. The influence of these two surfaces is
also perceptible in east and west fluxes and to less extent in the south direction. Maxi-
mal differences reach 100 Wm-2, for example between L(S→) and L(N→), as the south
part of the street is cooler in the daytime.
5.3.3. Heat gained by a standing person
In order to better understand the impact of the radiative environment described above on
a human body, the actual short-wave and long-wave irradiances absorbed by a standing
person in each direction are reported in Fig. 5.12 and Fig. 5.13. These fluxes take into
149
account the human absorption coefficients and the human shape as expressed by the
equations 2.7 and 2.8.
The total absorbed short-wave irradiation (Fig. 5.12) ranges from 9:00 to 17:00 LST
between 160 Wm-2 and 200 Wm-2. The peak values were recorded at 11:00 LST and
15:00 LST when the sun irradiates the human body laterally from the south-east or
south-west directions. Because the aspect ratio H/W = 1, direct solar radiation before
and after these two time points is blocked. This explains the sharp increase or decrease
of energy gained at both times, respectively. At midday, the short-wave irradiance ab-
sorbed is somewhat lower because of the smaller body surface irradiated as the sun is
high. Basically, the human body absorbs less than 20 Wm-2 from the wall which corre-
sponds to the diffusely reflected solar radiation, whereas he absorbs up to a maximum
of 70 Wm-2 from the opposite side facing the sun in the early afternoon. The absorption
from east and west sides shows similar patterns, but with the situation prevailing in the
morning at the east side being symmetrically reflected on the west side in the afternoon.
The highest values of global radiation reach 100 Wm-2 at 9.00 LST for the east and
17:00 LST for the west, respectively.
The individual absorbed long-wave irradiance in each direction as well as the total
amount is shown in Fig. 5.13. The evolution of the long-wave radiation fluxes is much
smoother than the short-waves. Owing to the cylinder-like shape of a standing person,
the long-wave irradiance absorbed from the lateral sides is much higher than those di-
rected upwards and downwards. These vary from 80 to 130 Wm-2 for the lateral direc-
tions as opposed to between 20 and 40 Wm-2 in the vertical direction. Notably, the dif-
ferences in absorption from the lateral directions: east, west, north and south, do not
exceed 20 Wm-2. This means that the radiant environment is relatively homogenous
vertically, in spite of the complex and variable shade patterns within the canyon. This is
due to the fact that the radiant heat received from each direction originates from all sur-
faces (walls and ground) and from the sky simultaneously. The largest variance, how-
ever, is observed for the south facing side which shows absorbs the lowest amount at
10:00 LST as the associated surfaces are still cool and the highest value at 15:00 LST,
when the canyon surfaces have in the meantime stored a lot of heat and receive addi-
tional energy from the irradiated opposite part of the street canyon.
150

200
Ktotal
160
Kabs (Wm-2)
120
K(W→) K(E→)
80 K(S→)
K(N→)
40
K↓
K↑
0
8 12 16 20 24 4 8 12 16 20
time (LST)
Fig. 5.12. Actual short-wave radiation (Kabs) absorbed by a standing person at the south
facing side of an E-W oriented with an aspect ratio H/W = 1

160 560
Ltotal
120 L(W→ L(N→) 420 (Wm )

Labs (Wm-2)
-2
L(S→)
L(E→)
80 280
Ltotal
40 L↑ 140
L↓
0 0
8 12 16 20 24 4 8 12 16 20
time (LST)
Fig. 5.13. Actual long-wave radiation (Labs) absorbed by a standing person at the south
facing side of an E-W oriented with an aspect ratio H/W = 1
151
5.3.4. Human thermal comfort
The mean radiant temperature Tmrt which is a function of the absorbed short-wave and
long-wave fluxes is plotted in Fig. 5.14 together with the comfort index PET. The
maximum Tmrt registered during the two days of study was about 66 °C. This value is
reached at about 15:00 LST. The minimum was 20 °C and was observed at 5:00 LST.
This curve is easily understood in light of the previous two graphs. The exposure to
short-wave irradiation is high from 9:00 LST to 17:00 LST while that for long-wave
irradiation gets progressively high throughout the day, with a maximum occurring at
15:00 LST. After 17:00 LST Tmrt decreases drastically because the short-wave irradia-
tion becomes negligible and hence results to a reduction of the surface temperatures. As
a consequence, less heat is radiated. In the night, Tmrt values remains between 20 °C and
30 °C. This is attributable to the surplus heat released by the surfaces.
The thermal comfort index (PET) shows, as expected, the same pattern as Tmrt (Fig.
5.14). During the daytime, the extremely high temperatures and the low-speed winds
accentuate the PET. The maximum PET value of 48 °C is registered at 16:00 LST. Dur-
ing the night, the relatively low air temperature and high-speed winds reduce the PET
values further: a minimum value of about 15 °C was noted.
70
60
Tmrt
50
Tmrt, PET ( °C)
40 PET
30
20
10
0
8 12 16 20 24 4 8 12 16 20
time (LST)
Fig. 5.14. Daily evolution of the mean radiant temperature Tmrt and the physiologically
equivalent temperature PET at the south facing side of an E-W oriented with an aspect
ratio H/W = 1
152
A particularly interesting feature is discernable from Table 5.2: a standing person ab-
sorbs at least about two-third of the total energy gained as long-wave irradiance from
the surrounding built environment. This highlights the importance of shading in reduc-
ing the heat absorbed by a standing person: firstly, because it prevents a direct exposure
of the person and secondly because it keeps the nearby surfaces cooler.
In Fig. 5.15, the radiant heat (L) gained by a standing person is plotted against the heat
released by the ground and the north wall. There is an almost perfect linear relationship
between the heat released and that absorbed. More absorption is from the north wall
because the body is considered in standing posture (vertical). The importance of the
ground has already been reported by Watson and Johnson (1988) and it is probably
more sensible for design proposes if these findings could be extended to other locations
across the street (less close to a wall).
Table 5.2: Percentage of short-wave radiation (SW) and Long-wave radiation (LW)
absorbed by a standing person at the south facing side of an E-W oriented with an as-
pect ratio H/W = 1 in Freiburg in summer
Time LST 11 12 13 14 15 16 17 18 19 20 21 to 5 6 7 8 9 10
SW (%) 29 16 24 26 27 27 24 3 2 1 0 2 4 16 28 29
LW (%) 71 74 76 74 73 74 76 97 98 99 100 98 96 84 72 71
560
y = 0.9387x - 3.8101
2
R = 0.99
L absorbed by a human body (Wm )
-2
520
y = 0.699x + 96.011
R2 = 0.9652
480
440
400
from north wall
from ground
360
400 440 480 520 560 600 640 680
-2
L emitted by a nearby surface (Wm )
Fig. 5.15. Long-wave radiation (L) absorbed by a standing person versus the radiant
heat emitted by the ground and nearby north wall
153
5.4. Comparison with ENVI-met simulation
The above results were compared with simulated results obtained with ENVI-met. A
horizontal resolution of 2 m x 2 m was chosen for the model area. For this comparison,
a new version of the model was used, and an adjusting factor (= 0.84) could be set for
the global radiation G to fit with the measured data of G provided by the urban climate
station.
Fig. 5.16 compares between simulated and measured Tmrt. Considering the complexity
of the three dimensional environment, ENVI-met is found to represent well the trends of
Tmrt with its two contrasting periods. However, simulated values of Tmrt are overesti-
mated in the morning (8.00 and 12:00 LST) and underestimated in the night-time in
comparison to field data. A full explanation of these differences is difficult owing to the
different methods used on-site and by the model for dividing the 3D radiative surround-
ings (see equations. 2.7 and 2.8 vs. 3.39 to 3.43, respectively). Each individual radiation
terms involved in the calculation of Tmrt could not be compared. Nevertheless, a com-
parison of the total short-wave and total long-wave irradiances absorbed by a human
body for both measurements and simulations is made (Fig. 5.17a and Fig. 5.17b).
Fig. 5.17a shows that the simulated short-wave irradiance absorbed by a standing body
is correctly reproduced if compared to measured data. The only discrepancies occur at
the critical times between 17:00 and 18:00 LST and between 7:00 and 8:00 LST.
Freiburg,14./15. 7. 2003
70
Tmrt measured
60
50 Tmrt simulated
40
Tmrt °C
30
20
10
0
10 12 14 16 18 20 22 24 2 4 6 8 10 12
time (LST)
Fig. 5.16. The mean radiant temperature Tmrt simulated by ENVI-met plotted against
measured Tmrt
154
These periods correspond to the times at which the sun crosses to or over the street. The
differences can be, therefore, attributed to the largest time step used for updating the
radiation fluxes within the model (set at 15 minutes). Moreover, the vertical resolution
used in ENVI-met (i.e. 2 m) may have been inaccurate for representing the actual com-
plex building heights, and hence influenced the final amounts of short-wave radiation
received. However, it is worthy of note that the projection factor formula used for esti-
mating Tmrt is expressed by fp = 0.3345 -0.00272φ (VDI 1998) and differs from the one
given in eq. 3.43b and used for all simulations in chapter 4. Indeed, the latter empirical
equation was found to overestimate the short-wave radiation flux absorbed by a stand-
ing person, especially when the sun is relatively low, when compared to measured data.
600
Labs measured
500
Labs simulated
400
-2
Labs, Kabs Wm
300
Kabs simulated
200
Kabs measured
100
0
11 13 15 17 19 21 23 1 3 5 7 9 11
(a) time (LST)
700
Lupwards measured
600
-2
Lupwards Wm
Lupwards simulated
500
400
11 13 15 17 19 21 23 1 3 5 7 9 11
time (LST)
(b)
Fig. 5.17. (a) Simulated individual short-wave (SW) and long-wave (LW) energy terms
absorbed by a standing person and (b) simulated long-wave irradiance emitted by the
ground, plotted against measured data in Freiburg
155
The long-wave irradiance simulated shows much more discrepancy with the field data
(Fig. 4.17b) and certainly played the main role in the differences observed in Tmrt,
which reached 50 Wm-2. This difference is, however, acceptable owing to the fourth
power law which links the radiant heat and the surface temperature. ENVI-met esti-
mated lower values mostly because of lower heat irradiated by the ground surface as
this is set to 50%. This is partly due to the rather arbitrary ground properties used in the
absence of accurate information on the actual ground: asphalt thermal properties, soil
humidity and soil temperature, as well as the assumption for constant temperature at the
boundary depth of 1.75 cm. Additionally, the heat released from the walls may have
been underestimated by the model due to the mean value assumed, especially because of
the location of the station close to the wall. In the night-time the simulated long-wave
radiant flux is lower as a consequence of the lower daytime ground heating and the
lacking heat release from the façades as no storage in the building materials is included.
5.5. Discussion and conclusion
In this work, the findings of an in-situ study conducted in an east-west oriented urban
canyon, with an aspect ratio 1, located in the mid-latitude city of Freiburg (Germany)
during hot summer days, is presented. The microclimate of the street and its impact on
human comfort were investigated.
The changes in the basic climatic variables (Ta, v and dd) in the canyon in comparison
to an unobstructed roof level location are found to be in good agreement with previous
studies: a small increase of Ta in the canyon adjacent to irradiated surfaces and a good
correlation in wind speed and direction between canyon and roof air. The daily dynam-
ics of canyon facets irradiances and their impacts on the heat gained by a pedestrian
depend strongly on street geometry and orientation. Thermal stress is mostly attribut-
able to solar exposure. A standing body is found to absorb up to 730 Wm-2 with more
than 70% of energy in form of long-wave irradiance against less than 30 % of short-
wave irradiance in the daytime. Shading of the pedestrian as well as the surrounding
surfaces is, hence, the first strategy in mitigating heat stress in summer under hot condi-
tions. More field investigations are needed to verify the generality of these results for
other locations and climatic conditions. Another example is presented in chapter 6.
156
6. Field measurements in Beni-Isguen, Algeria
This chapter reports on a field study conducted for the same location and season as as-
sumed for the main simulations (chapter 4). It seeks to provide some comparison with
the numerical results. This field campaign is part of an ongoing research initiated by the
School of Architecture of Algiers, Algeria. The site is located within a typical desert
city and the focus was to provide some knowledge on the effectiveness of traditional
design forms in relation to outdoor comfort.
6.1. Site description
The region of Ghardaia in the Saharan Mzab valley includes five compact cities, i.e. El-
Atteuf, Bounoura, Melika, Beni-Isguen and Ghardaia, each of them possessing an oasis.
Fig. 6.1 shows the city and oasis of Beni-Isguen.
Fig. 6.1. The old city of Beni-Isguen and its oasis in the Mzab valley, Algeria
157
These cities, built in the 10th century and designated world culture heritage buildings by
UNESCO since 1982, provide in addition to an architectural authenticity (Donnadieu et
al. 1977, Ravérau 1981) a climatic-conscious design developed over centuries of build-
ing experience.
It is commonly claimed that this type of architecture is perfectly adapted to suit the sur-
rounding climatic environment. Yet, this has not been proven and the current knowl-
edge on this issue is still mainly qualitative. The climatic effectiveness of traditional
solutions has been questioned as these also reflect cultural specificities (e.g. Givoni
1997). The positive climatic effects of numerous traditional solutions may have been
overestimated: Givoni (1997) and Meier et al. (2004) argue that the excessive thermal
inertia of such architecture in hot-dry climates prevents the nocturnal cooling of the
houses and leads to discomfort indoors at night. Also, Ouahrani (1993) found that light-
ing during the day is insufficient in the typical inward-looking houses because of the
small size of the courtyard which is the only source of natural light. Thus, more investi-
gation, based on scientific methods, is required to further quantitatively evaluate com-
mon design concepts and establish the veracity of this common belief of climatic adap-
tation. Furthermore, available studies undertaken in such built environments focused on
the architectural dimension i.e. indoor climate (e.g. Ouahrani 1993, Krishan 1996,
Potchter and Tepper 2002, Meier et al. 2004,), whereas very few published studies are
available to date which deal with the urban design level i.e. outdoor spaces (e.g. Grund-
ström et al. 2003).
The living conditions for people are very difficult in hot-dry climates. However, they
can be improved by using an appropriate housing design. A number of strategies have
been frequently reported and advised in the literature (e.g. Koenigsberger et al. 1973,
Golany 1982, Golany 1996, Givoni 1997). These include fabric compactness, the high
inertia of the construction, shading, night ventilation and evaporative cooling. In the
winter season, provision for sunshine is recommended with heat storage capacity. The
Mzab cities typically illustrate these recommendations. The old settlements in the Mzab
valley form a system where environmental concepts can be stated at the 3 consecutive
design scales: (1) the location in the valley, (2) the urban fabric and (3) the architecture
of the house (Figs. 6.1. and 6.2). See Ali-Toudert et al. 2005 for details on this issue.
According to the objectives of this study, measuring points were selected in 8 locations
with various orientations and aspect ratios in the old city of Beni-Isguen, Mzab valley,
158
Algeria (Table 6.1). The city is built on a hill slope facing east and follows the topogra-
phy of the site. The measuring points were arranged along a downward measuring route
(Fig. 6.3) with the starting point (point 1) being the highest (525 m a.s.l.) and the last
point (point 8) at the market place (484 m a.s.l.).
Fig. 6.2. A bird view on the a typical compact urban fabric of Beni-Isguen in the Mzab
valley, Algeria (Roche 1970)
Table 6.1. Geometry and material properties at the eight measuring points in the old city
of Beni-Isguen, Mzab valley, Algeria
site street width aspect ratio H/W SVF orientation ground ground material
W(m) (angle from N) albedo
1 2.5 H1/W = 1.5; H2/W = 0.6 0.45 NE-SW; 45° 0.20 concrete
2 1.4 H1/W = 7.5; H2/W = 4.7 0.11 N-S; 166° 0.15 stone & concrete
3 2.1 H1/W = 3.5; H2/W = 3.8 0.13 NEE-SWW; 63° 0.15 concrete
4 1.5 H1/W = 1.4; covered 0.03 NW-SE; 130° 0.15 concrete
5 2.1 H1/W = 4.6; H2/W = 3.8 0.16 NE-SW; 50° 0.20 stone & concrete
6 2.4 H1/W = 3.1; H2/W = 3.5 0.14 NW-SE; 122° 0.15 concrete
7 1.7 H1/W = 4.3; H2/W = 4.3 0.09 NW-SE; 125° 0.15 concrete
8 market place H1/W = 0.1; H2/W = 0.1 0.67 -- 0.30 stone
159
The city can be divided into 2 parts: the upper part is composed of small houses of ir-
regular shapes (points 1 to 4) and the lower part is almost flat with more regular streets
and houses (points 5 to 8). The urban structure is compact with very narrow streets of
various orientations and high aspect ratios. The H/W ratios of the selected streets vary
between 7.5 and 0.6. The building materials are heavy, mostly made of stone. The walls
are thick and heavy, covered with a layer of gypsum and painted with light colours
(rose, blue or ochre). To get a better impression of the site conditions, Fig. 6.4 shows
some photos of selected measuring sites as well as some fish-eye photos. Fig. 6.2, in
addition shows the upper part of the city and the measuring point 3 is located in the
main street on the left.
Fig. 6.3. Route and all measuring points within different street geometries in the ver-
nacular city of Beni-Isguen, Mzab valley, Algeria
160
Fig. 6.4. Photographs and fish-eye photographs of selected measuring sites within the
city of Beni-Isguen, Mzab valley, Algeria
161
6.2. Measurements
As the climatic conditions in the Sahara are relatively homogeneous in summer, the
measurements of the necessary meteorological variables were restricted to a few days.
Therefore, the meteorological measurements of this study were carried out on 2 days
only (24 and 26 June 2003), which had typical summer conditions, i.e. hot, sunny and
cloudless.
The meteorological measurements were performed consecutively starting at point 1
from 6:00 to 24:00 LST and lasting 15 min on average at each site. For each measuring
point, the time interval between 2 measurements was about 3 h. Point 8 was considered
as a reference site as it is an unobstructed location.
Air temperature (Ta), air humidity (VP) and wind velocity (v) were measured at 1.2 m
a.g.l. In addition, the mean radiant temperature Tmrt had to be determined precisely be-
cause it is an important variable in the human energy balance. The use of a globe ther-
mometer for measuring Tmrt is common but this method was dismissed in this investiga-
tion because of its inaccuracy outdoors as previously discussed (see section 2.2.3). Tmrt
was determined according to Höppe (1992) and VDI (1998), i.e. expressed by equations
2.8a and 2.8b. All short-wave and long-wave radiation fluxes from these 6 directions
were recorded by means of a pyranometer and an infrared thermometer. The long-wave
atmospheric radiation was calculated after Oke (1987) as a function of the measured air
temperature and air humidity. The sky view factor SVF was determined for all locations
by means of a camera with fish-eye lens. The albedo of the ground has been estimated
separately before starting the main measurements as the ratio of short-wave reflected
and short-wave global irradiance around noon for each street.
6.3. The microclimate in the canyon

6.3.1. Air temperature and air humidity
The following results show exemplarily the data obtained on 26 June 2003. Fig. 6.5
shows the air temperature Ta for all measuring sites. The highest value of Ta was re-
corded around 15:00 LST at the sunlit point 1. The diurnal course of Ta showed very
small differences between the various urban streets in the morning until 11:00 LST.
162
Ghardaia, 26 June 2003

45
pt. 1
pt. 2
pt. 3
40 pt. 4
pt. 5
pt. 6
Ta (°C)
pt. 7
35 pt. 8
30
25
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
13
pt. 1
pt. 2
pt. 3
12 pt. 4
pt. 5
pt. 6
VP (hPa)
pt. 7
11 pt. 8
10
9
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
6
pt. 1
pt. 2
5 pt. 3
pt. 4
pt. 5
4
pt. 6
v (m s-1)
pt. 7
3 pt. 8
0
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
Figs. 6.5 to 6.7. Air temperature Ta, vapour pressure VP and wind speed v, at all meas-
uring points during a typical summer day (26 June 2003) within the city of Beni-Isguen,
Algeria
163
With the increased turbulent transfer of heat induced by the irradiated surfaces (Naka-
mura and Oke 1988), the disparity in Ta became larger between non-shaded and shaded
streets. A peak difference ∆Ta = 2 K was reached between 15:00 and 16:00 LST.
The measuring points 4 and 5 showed a tendency to be slightly cooler than the others
because of their lower exposure to direct solar beam. In fact, point 4 is a covered path-
way and point 5 is a deep canyon oriented close to N–S, which allows a longer time of
protection from direct solar radiation as revealed by the simulations. Point 2, which was
the deepest street investigated, was not as cool as expected. This is likely due to its loca-
tion at the city boundary, which leads to a stronger air mass exchange with the adjacent
largely exposed areas. After 21:00 LST, when Ta averages 32.5°C, almost no difference
(∆Ta) was found between all investigated urban streets. The market place (point 8),
however, cooled faster from 22:00 LST and became 1.5 K cooler at midnight in com-
parison to the other enclosed measuring points. The SVF at the market place was high
(0.67, see Table 6.1) and allows a rapid dissipation of released heat. The urban streets
have low SVF values and therefore the heat release from the canyon materials is trapped
in the canyon air volume.
The vapour pressure (VP) was low and corresponds to the typical water content in this
location (Fig. 6.6). It reached values around 12hPa in the morning until noontime and
was 10 hPa during the night. A systematic influence of the specific site conditions could
not be detected. It should be mentioned here that many kitchens are located on the street
sides, which, as local sources of heat and humidity, might have influenced these results.
6.3.2. Wind speed
The values of the wind velocity (v) are only indicative and can not give a comprehen-
sive analysis of the air flow in the streets of Beni-Isguen (Fig. 6.7). Nevertheless, some
observations are worth mentioning.
Table 6.2 lists the wind speed recorded on the 2 days of measurements. The wind speed
measured at the unobstructed measuring point 8 at the market place was temporally
more than 5 ms–1 while in the urban streets (v) was reduced up to 4.6 ms–1. Although
unexpected, this indicates that ventilation at street level does exist despite the high
compactness of the urban structure. The measuring point 4 is worth mentioning: even
though covered and enclosed, it turns out to be a ‘windy’ site being the best ventilated
of all streets investigated. This may be (1) because of its location at the limit between
164
the high and down part of the city, which means that the buildings downwards do not
obstruct the incoming wind flow, and (2) because the street canyon faces the main wind
direction, the east.
Table 6.2. Mean wind velocity (ms–1) measured at all measuring sites on (a) 23 June and
(b) 26 June 2003 in Beni-Isguen, Algeria (32.40°N, 3.80°E)
Time from 6 LST 9 LST 12 LST 15 LST 18 LST 21 LST 24 LST

(a) (b) (a) (b) (a) (b) (a) (b) (a) (b) (a) (b) (a) (b)
Point 1 0.0 0.8 0.3 1.1 1.8 0.5 1.0 2.8 0.8 2.0 1.3 0.9 0.1 0.2
2 0.3 0.8 0.5 0.4 0.1 0.6 1.2 1.9 0.1 0.9 0.3 0.1 0.1 0.3
3 0.3 0.4 0.8 0.5 1.2 1.8 1.2 1.7 0.7 0.5 0.4 0.4 0.1 0.3
4 0.6 0.5 1.5 1.8 1.7 2.7 4.6 3.5 1.5 2.3 1.2 1.4 0.8 0.6
5 0.5 0.3 0.8 0.9 2.2 1.9 1.4 2.9 1.6 1.1 0.8 0.2 0.5 0.1
6 0.9 0.6 1.7 1.3 1.7 1.7 2.1 2.0 1.2 0.3 1.0 0.2 0.7 0.2
7 0.7 0.8 0.5 0.6 1.1 1.5 2.0 0.9 0.6 0.4 0.4 0.7 0.7 0.2
8 0.4 0.1 3.4 2.9 2.3 5.6 1.2 2.3 1.5 2.5 0.5 1.5 0.7 0.1
6.3.3. Surface temperatures
The surface temperatures of the ground (Ts) and of the 2 walls (Tw) are given in Figs.
6.8 and 6.9. They inform on the exposure versus shading of the canyon facets (Naka-
mura and Oke 1988, Yoshida et al. 1990/91). In the morning, street surface tempera-
tures showed relatively small differences between all urban canyons. Points 8 and 1
showed maximum values around 50 to 55°C in the afternoon, whereas point 4 experi-
enced Ts of about 34°C at the same time in the afternoon, i.e. 3 to 4 K lower than Ta.
Fig. 8 shows that the street floor can be irradiated even for deep canyons, i.e. at points
5, 6 and 7 at midday hours (12.00 to 13:00 LST). Yet, Ts values were below 46°C be-
cause of their short duration of exposure.
The temperatures of the canyon wall surfaces showed generally small differences be-
tween both sides when shaded. In this case, Tw values are below the corresponding air
temperature for each measuring point as observed by others (e.g. Yoshida et al.
1990/91). For subtropical latitude, high aspect ratios combined with a high sun position
result in a good protection of the façades in comparison to the ground surface as re-
ported by the numerical studies of Arnfield (1990a) and Bourbia and Awbi (2004). The
165
measuring point 4 showed almost the same Tw values as Ts. At point 1, ∆Tw between
both walls was much larger (up to 11 K) because of its large SVF which allows a longer
period of solar irradiation of the south-east facing wall. In the evening all surfaces were
warmer than the air by few degrees except at point 8 where Ta and Ts were almost equal.
65
pt. 1
pt. 2
pt. 3
55 pt. 4
pt. 5
pt. 6
pt. 7
pt. 8
(°C)
45
Ts
35
25
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
50
pt. 1
pt. 2
45 pt. 3
pt. 4
pt. 5
40 pt. 6
Tw (°C)
pt. 7
35
30
25
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
Figs. 6.8 and 6.9. Surface temperature Ts and wall temperature Tw at all measuring
points during a typical summer day (26 June 2003) within the city of Beni-Isguen, Alge-
ria
The heat gained by a human body consists of short-wave irradiance (Kabs) due to the
exposure to direct and diffuse solar radiation and to a long-wave irradiance (Labs) ab-
sorbed from heat emitting surrounding surfaces. For a better understanding of the role
166
of both components on the total energy absorbed by a standing person, the 2 quantities
are represented separately in Figs. 6.10 and 6.11.
At the subtropical location of Beni-Isguen, the sky is clear and the global radiation in
the summer is dominated by the direct solar radiation, while the diffuse radiation is very
small. Hence, Kabs depends strongly on the sun course, canyon geometry, and orienta-
tion.
400
pt. 1
pt. 2
pt. 3
pt. 4
300
pt. 5
pt. 6
Kabs (W m-2)
pt. 7
pt. 8
200
100
0
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
580
pt. 1
pt. 2
pt. 3
pt. 4
540
pt. 5
pt. 6
Labs (W m -2)
pt. 7
pt. 8
500
460
420
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
Figs. 6.10 and 6.11. Short-wave (Kabs) and long-wave (Labs) radiation fluxes absorbed
by a standing person at all measuring points during a typical summer day (26 June
2003) within the city of Beni-Isguen, Algeria
167
The unobstructed market place (SVF: 0.67) showed the highest Kabs values (Fig. 6.10),
namely 215 W m–2 in the morning (8:00 LST) and a maximum of 260 W m–2 around
11:00 LST. The high Kabs value recorded at 8:00 LST and around 17:00 LST at the
market place is due to the relatively low sun height (~ 25° to 35°), which increases the
amount of energy absorbed laterally by a standing person. At the measuring point 1,
which also had a relatively high SVF value (0.45), Kabs reached 257 W m–2 at 15:00
LST.
At all other measuring points, Kabs is clearly lower and did not exceed 60 Wm–2, except
at noontime when the sun is at its highest position (∼ 75°). At this time, the measuring
points 2, 3 and 5 were irradiated despite their high aspect ratios. These results suggest
that a deep geometry has the advantage of shortening the period of direct exposure to
sun regardless of the orientation.
In contrast to Kabs the pattern of the daily course of the long-wave radiation fluxes (Labs)
absorbed by a standing person (Fig. 6.11) is different, and Labs was generally higher and
reached values between 460 and 550 W m–2. No systematic Labs difference could be
found between the various street canyons, probably because of the high thermal admit-
tance of the building materials, which clearly levelled the daily surface temperatures in
contrast to the larger fluctuations of Ta. However, the streets at measuring points 2 and 4
clearly released less heat than streets at other measuring points in the afternoon hours as
these had lower surfaces temperatures. Moreover, the daily amplitudes of Labs were rela-
tively low, also being a logical consequence of the high thermal inertia of the urban
canyon materials. The market place showed lower Labs values in the early morning and
during the night because the larger SVF leads to a faster nocturnal cooling.
6.3.5. Mean radiant temperature
As expected, the mean radiant temperature Tmrt (Fig. 6.12) was noticeably lower within
the urban streets than in an unobstructed location (e.g. point 1 vs. point 8). The differ-
ence between sheltered and exposed measuring points reached 36 K at the hottest time
of the day (e.g. between point 1 and point 2 around 15:00 LST). The market place (point
8) experienced the highest Tmrt values ranging between 60 and 75°C from 8:00 to 17:00
LST. The high Tmrt values in the morning were due to the lateral irradiation of the stand-
ing person when the sun is still relatively low, as could be seen in Fig. 6.10. In more
detail, Tmrt differences between the different urban streets are clearly higher than ∆Ta.
168
The lowest Tmrt values were calculated for the measuring point 4, corresponding to the
lowest Kabs und Labs values. This is not surprising since point 4 is a covered pathway
(SVF: 0.03) and is not directly influenced by solar radiation. At this location, Tmrt
showed a ‘flat’ diurnal course with values varying between 32 and 37°C. Measuring
point 2 is as protected as point 4, except at midday, at which time Tmrt reaches 55°C.
100
pt. 1
pt. 2
pt. 3
80 pt. 4
pt. 5
pt. 6
pt. 7
Tmrt (°C)
60 pt. 8
40
20
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
50
pt. 1
pt. 2
40 pt. 3
pt. 4
pt. 5
30
pt. 6
Tmrt - Ta (K)
pt. 7
20 pt. 8
10
-10
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
Figs. 6.12 and 6.13. Mean radiant temperature Tmrt and the difference (Tmrt-Ta) at all
measuring points during a typical summer day (26 June 2003) within the city of Beni-
Isguen, Algeria
169
This is due to the N–S orientation of this urban street canyon which prevents shadowing
even though the canyon is very deep. The role of the orientation can also be seen for
point 6 and point 7. Their NW–SE orientation leads to a lower amount of energy being
gained by a human body, particularly in the early afternoon, in contrast to higher energy
gain for measuring points 1, 3 and 5 (nearly NE–SW oriented). After 18:00 LST, negli-
gible differences were registered between all urban streets and Tmrt averaged 35°C. In-
deed, the solar radiation intensity is less and the low sun position promotes shade at
street level. Yet, the deep geometry has partly inhibited the influence of the orientation,
as well as the discontinuous measurements.
The diurnal fluctuations of Tmrt as well as Tmrt maxima are mainly attributable to Kabs
(Fig. 6.10), whereas Tmrt minima depend on Labs (Fig. 6.11). The latter rely on the ther-
mal admittance of the building materials and on SVF. Compared to low admittances
(usually light materials), high admittances reduce the heat emitted from the surfaces in
the daytime, which extends to the night-time period.
In the daytime, Tmrt was for the most part higher than Ta (Fig. 6.13). By night Tmrt was
approximately equal to Ta. Tmrt was a few degrees higher in the shade in the urban can-
yons and up to 14 K higher in irradiated situations. Points 2 and 4 had even lower val-
ues (1.5 to 2 K) shortly around 15:00 LST because of their confined aspects. At the
market place, the maximum difference (Tmrt – Ta) reached 38 K around 11:00 LST.
6.4. Thermal comfort analysis
As previously shown, PET schemes are basically influenced by Tmrt in summer under
sunny conditions. Therefore, the patterns of the diurnal courses of PET (Fig. 6.14) and
Tmrt (Fig. 6.12) are similar. Regressions analyses between PET and Tmrt lead to coeffi-
cients of determination of R2 = 0.900 for a linear relationship and R2 = 0.939 for a loga-
rithmic relationship. This is not surprising as Ta, VP and v vary comparatively much
less. The most uncomfortable locations were those exposed to the sun. PET peak values,
occurred in the afternoon and ranged from 53 to 55°C at point 8 (market place, SVF:
0.67) and 54°C at the measuring point 1 (SVF: 0.45). These high values were mitigated
by high wind speed (i.e. 5.6 m s–1 at point 8 and 2.8 ms–1 at point 1). By contrast, the
lowest PET value (37°C) was determined for the measuring point 4 (SVF: 0.03). In the
early morning and late afternoon, PET did not reveal clear differences between the sites
within street canyons, i.e. PET ≈ 30°C before 8:00 LST and ≈ 36°C after 18:00 LST.
170
60
pt. 1
pt. 2
pt. 3
50 pt. 4
pt. 5
pt. 6
PET (°C)
pt. 7
40 pt. 8
30
20
6:00 9:00 12:00 15:00 18:00 21:00 0:00 3:00
time (LST)
Fig. 6.14. Physiologically equivalent temperature PET at all measuring points during a
typical summer day (26 June 2003) within the city of Beni-Isguen, Algeria
PET differences between the various streets were more pronounced around noon at
points 1, 2, 3 and 5, indicating a higher level of heat stress (∼ 45 to 51°C). The measur-
ing points 6 and 7 showed a slightly lower level of heat stress 1 h later (< 45°C). PET
values decreased very slowly in the night-time and roughly equalled the air temperature
at midnight.
6.5. Discussion and conclusion
On-site meteorological measurements were carried out in an old desert city. For the first
time, a human-biometeorological based method was applied in vernacular desert cities
with the goal of investigating the effectiveness of traditional design solutions in ensur-
ing comfortable outdoor conditions. This experimental work provides, in spite of the
limited data sample, some quantitative information and suggests a number of potential
future investigations. The results confirm many of the findings gathered with numerical
modelling:
Quantitatively, the results show a high thermal discomfort in a non-shaded location at
subtropical latitude of 32.40°N under summer conditions, with Tmrt and PET reaching a
maximum of 74 and 55°C, respectively. In the absence of shading, heat stress is experi-
171
enced in the morning hours and lasts for a large part of the day. Yet, PET maxima re-
corded on-site are lower than those found by simulation for two main reasons. First, the
wind speed on the market place (point 8) and at point 1 was high, while the wind was
calm in the canyons (e.g. 5.6 ms-1 at point 8 versus 0.2 ms-1 in simulated stretched can-
yons). In the absence of strong wind, which is also common for that region, PET would
be about 8 K higher for the same radiation context, i.e. about 63 °C instead of 55 °C.
Secondly, the global radiation calculated by the model was higher in comparison to the
one measured. A maximum global radiation G of 1000 Wm-2 was measured at the open
space on the market place whereas ENVI-met calculated 1100 Wm-2 at a horizontal sur-
face. In fact, this difference is likely due to more absorption in the real case and no ad-
justing factor for solar radiation was used in the present simulations. Within street can-
yons, the reflections from the walls lead to further increase of G (e.g. G ≈ 1250 Wm-2
for irradiated locations in streets with H/W = 2).
This field study confirms that shading achieved by means of high aspect ratios reduce
substantially the thermal discomfort of people at street level. High aspect ratios were
found to be an effective strategy in shortening the duration of exposure to solar energy
and mostly affected the amount of absorbed short-wave irradiance. The very high aspect
ratios investigated have partly inhibited the influence of the orientation; however, it was
observed that the N–S orientation is the most comfortable except shortly around noon,
and a NE–SW oriented street is more stressful than a NW–SE one. This is in good
agreement results obtained by modelling. Moreover, covered streets experience the low-
est PET values as the heat emitted from these surfaces is noticeably lower in compari-
son to other canyons and the sheltered site is almost not influenced by the daily course
of solar radiation. This corroborates the usefulness of using galleries as pedestrian
pathways.
The heat gained by a standing person was also found to be high even in the early morn-
ing or late afternoons because of the low position of the sun and agrees with simulation
results. This energy gain depends on (i) the exposure of the body itself and (ii) on the
exposure of the surrounding urban surfaces. As for the Freiburg study, the distinct rep-
resentation of the absorbed short-wave and long-wave irradiances (accounting in Tmrt
calculation) revealed that the long-wave irradiance is a significant source of heat load
and the absolute values largely exceed the short-wave irradiance absorbed. Hence, shad-
ing the surrounding urban surfaces is as crucial as shading the person in mitigating the
heat stress.
172
Moreover, these measurements give some information on the role of the building mate-
rials. Thick and heavy materials with high thermal capacities help to decrease the long-
wave radiant heat during the day and minimize the differences between the streets of
different geometries and orientations. However, when high thermal capacity is com-
bined with high aspects ratios, the heat released from the canyon surfaces in the night-
time is slowed down and delayed the nocturnal cooling of the urban fabric. Although,
night-time outdoor comfort is of small relevance in comparison to day-time, the cooling
of the houses would last longer and would extend the period of night-time discomfort
indoors as reported by Meier et al. (2004). This suggests the necessity of further inves-
tigation on this issue.
Contrarily to common opinion, air temperature was found to be moderately lower in the
urban canyons in comparison to a free location (∆Tmax = 2 K). No clear correlation
could be found between the aspect ratio and the air temperature. This contrasts with the
higher air temperature differences reported by Coronel & Alvarez (2001) and Grund-
ström et al. (2003), which might be due to the larger cities investigated. Furthermore, Ta
as a conservative quantity reacts little to urban geometry and can therefore be used only
as a secondary indicator for comfort outdoors. Indeed, the reason why Ta is still often
used as the main comfort indicator is probably that any decrease of Ta is almost always
associated with increased shading and hence lower irradiances.
The present study is based on an energy-model approach which assesses comfort by
means of comfort indices. Yet, all subjective aspects that may affect the actual thermal
sensation of people and revealed by social surveys to be important (e.g. Nikolopoulou et
al. 2001, Spagnolo & de Dear 2003) are not dealt with in the present work. Social sur-
veys will bring more knowledge on the reliability of these indices and refine their scal-
ing. Such work is particularly lacking in such severe climates, where people‘s subjec-
tive perception of the climate may play an important role in their sensation of comfort.
Vernacular architectures provide valuable knowledge on climate-conscious design, and
this study draws attention on issues still needing further investigation, for instance:
To compare old and new typologies in the Mzab valley. This is particularly relevant
for the region where the new settlements contrast strongly with the old. These new
typologies have noticeably larger urban plan densities and open spaces and make
more use of vegetation. Larger open spaces also imply a different use, as these are
more appropriate for social activity and can include traffic.
173
This study shows some evidence for existing urban ventilation in the city’s streets in
spite of the high density of urban fabric and points out another field for further in-
vestigation. Moreover, this work shows that more continuous measurements are
needed for establishing the dependence between air temperature and the urban struc-
ture.
Ali-Toudert et al. (2005) described the strong interdependence between the various
scales in these urban structures regarding climate adaptation. An exhaustive assess-
ment of the effectiveness of these design concepts should, therefore, deal simultane-
ously with the indoor and outdoor climates, e.g. at a neighbourhood scale, and in-
clude summer and winter issues, i.e. internal and external thermal and visual com-
fort, ventilation, etc. Such extensive studies are lacking.
174
7. Discussion and conclusion
The present study dealt with outdoor thermal comfort in dependence on street design
under summertime conditions in hot and dry climate. The street properties investigated
included the aspect ratio H/W, street orientation and a number of design details, i.e. gal-
leries, rows of trees, and overhanging façades. The methodology was mainly based on
numerical modelling using the microscale model ENVI-met 3.0. The simulations results
were supplemented by two short-term on-site measurements under typical hot summers:
i) in the mid-latitude location of Freiburg, Germany, and ii) the subtropical location of
Ghardaia, Algeria. A human-biometeorological method for the assessment of the ther-
mal comfort was used, and expressed by the energy-based index PET.
The following material discusses the main results obtained from the simulations along
with the findings of the field measurements. Design recommendations are outlined,
supported by few design examples to illustrate their applicability. Finally, some remarks
are made on the methodology employed, namely on the accuracy of the model ENVI-
met and on eventual refinements.
7.1. Street microclimate
Using ENVI-met, Ta is found to decrease moderately with the increase of the aspect
ratio. This is mostly perceptible after 10:00 LST and is less than 1 K at the hottest hours
of the day between two “consecutive” street aspect ratios (e.g. H/W = 2 and 1, 3 and 2,
etc.). H/W = 4 can be up to 3 K cooler than H/W = 0.5 around 15:00 LST. This is in
good agreement with experimental studies, which revealed only a weak warming of
175
canyon air in relatively large streets (e.g. Nakamura and Oke 1988, Yoshida et al.
1990/91, Pearlmutter et al. 1999) and an increasing thermal stratification as the canyons
become deep (Santamouris et al. 1999, Coronel and Alvarez 2001). Moreover, small
differences in Ta were found when changing the orientation for the same aspect ratio,
but ∆Ta increases between E-W and N-S oriented street with increasing H/W. The
warming rate of the canyon air reflects the irradiances patterns of the canyon facets and
Ta maxima in deep canyons are reached at different times of day according to the orien-
tation, in particular as the aspect ratio increases. Explicitly, Ta,max occurs in the early
afternoon for N-S canyons and in the late afternoon for E-W canyons. For H/W < 2, E-
W streets are all the day warmer than N-S streets with a maximum ∆Ta reaching 1.2 K.
For higher aspect ratios (H/W ≥ 2), E-W streets are the warmest around 10:00 LST and
17:00 LST when compared to N-S streets, but are shortly cooler at midday hours. If the
same aspect ratio is considered, E-W streets are warmer because of a longer period of
time of solar exposure, followed by the NE-SW orientation which is irradiated in the
morning hours. N-S streets and NW-SE are found to be cooler.
Experimentally, the “Freiburg canyon” showed that Ta is relatively uniform across the
street, except close to the permanently sunlit surfaces on the north side where Ta was up
to 2 K higher, which agrees with previous findings (e.g. Nakamura and Oke 1988).
Moreover, in the Saharan city of Beni-Isguen, where the climate conditions are extreme
and the urban fabric very dense, air at street level in various very deep urban canyons
(H/W up to 7.5) showed at most 3 K higher temperatures in comparison to a fully unob-
structed site. This contrasts with the larger gradients reported by Grundström et al.
(2003) and Coronel and Alvarez (2001) and calls attention on the necessity of more
field measurements to clarify the dependence of Ta on the urban geometry.
ENVI-met reports on relatively uniform air temperature across the canyons and under-
estimates the additional warming of air adjacent to irradiated surfaces noted in-situ. This
is due to lacking storage in the building materials and also to the spatial resolution used.
Certainly for the same reasons, Ta within the galleries varies little from Ta calculated in
canyon centre. Yet, in-situ measurements within galleries are not available to allow any
comparison. The simulations also revealed that using geometrical irregularities in the
vertical profile has minimal impact on Ta in comparison to a simple geometry. Yet, a
larger openness to sky of the canyon (i.e. larger sky view factor) shows an evidence to
warm more in the daytime and cool faster in the evening.
176
By contrast, the model results report that more effects on Ta are obtained if vegetation is
used. Explicitly, Ta is found to decrease up to 1.5 K in comparison to a similar canyon
shape without trees. This is attributable to less warming up of the ground surface
through shading, together with moderate water content within a natural soil (water con-
tent 30%) against waterproof canyon floors used in unplanted streets. Shashua-Bar and
Hoffman (2000) report on a similar magnitude of air cooling in planted streets for com-
parable summer conditions.
Santamouris et al. (1999) and Nakamura and Oke (1988) reported that no clear correla-
tion could be found between street geometry and Ta in their field studies. The present
parametrical study showed a relationship between Ta, the aspect ratio, the use of vegeta-
tion, and the orientation. Increasing H/W or planting trees are the most efficient strate-
gies for decreasing Ta, even though these changes are limited. More field measurements
in various street canyons are still required to validate the results obtained numerically.
Furthermore, Ta, being a conservative quantity, is confirmed to be not suitable as main
comfort indicator and can at most deal as a secondary indicator. This contrasts with the
few published numerical studies (including comfort assessment), for example those
based on the CTTC model, which report on much larger air temperature variations due
to street geometry and suggest the dominant role of Ta in evaluating comfort (e.g. Swaid
et al. 1993, Grundström et al. 2003). This discrepancy is likely attributable to the em-
pirical assumptions of the CTTC model and calls attention to the methodological prob-
lems in assessing outdoor comfort, which make any comparison vulnerable.
Indeed, the analysis of comfort by means of PET showed that Ta has a secondary effect
on comfort, far behind the radiations fluxes. The correlation between Ta and PET is very
weak. The decrease of Ta due to urban geometry alone is noticeably less efficient in
mitigating heat stress. For instance, a sensitivity analysis, made with the measured data
recorded in the Freiburg campaign, estimates a PET increase of about 0.75 K for each 1
K increase of Ta: ∆PET = 3/4 ∆Ta. In fact, the positive effects of lower Ta in mitigating
the heat stress is overestimated, certainly because any decrease in Ta (even by a few
degrees) is often combined with a shading situation and hence lower amounts of radiant
energy absorbed by a person. This might explain the wide use of Ta as comfort indica-
tor. On the contrary, in a sunlit location during hot and clear summer day, the present
study confirmed that thermal indices like PET are strongly dominated by Tmrt as previ-
ously reported by others (e.g. Mayer and Höppe 1987, Jendritzky et al. 1990).
177
Vapour pressure, as expected, does not react to geometry changes for all dry configura-
tions tested and insignificantly for those with vegetation, mainly because of the simu-
lated typical lack of water in the soil in hot-dry climates (water content 30 % set at the 3
soil layers) which limits the evapotranspiration. It is unlikely that a moderate increase in
air humidity will improve sensitively the thermal sensation under very hot summer con-
ditions.
The wind speed is found to play an important role in the final comfort sensation and the
influence of the wind speed differs depending on whether the location is irradiated or
not, i.e. high or low Tmrt values (close to Ta). Explicitly, increasing the wind speed im-
proves the comfort sensation: for a wind speed of about 2 to 2.5 ms-1, PET can decrease
up to 12 K in case of extreme discomfort against 2 to 4 K for locations with low Tmrt
values. The wind speed is found to depend strongly on the wind incidence in relation to
street axis. Two main patterns are identified, depending on whether the wind is perpen-
dicular or parallel to the canyon axis. In each case, the mid-canyon and the street ends
experience different ventilation zones and lead to different comfort situations.
For a perpendicular incidence, the street ends experience relatively high wind speeds (>
2 ms-1 for 5 ms-1 at H = 10 m) due to intermittent vortices shed on the building corners.
These vortices are responsible for the mechanism of advection from building corners to
canyon centre, creating a convergence zone in the mid-block region of lowest wind
speeds (Hoydysh and Dabbert 1988, Santamouris et al. 1999). The intensity of these
eddies and their extent inside the street depend on the aspect ratio. Increasing the aspect
ratio strengthens the air flow up to a certain proportion (H/W ≈ 3). Above this threshold,
the strong eddy circulation takes place in the upper part of the street and does not reach
the street level. Hence, lower air velocities are felt by pedestrians. In the central part of
the canyon, the wind speed is reduced in all cases and reaches at most 0.3 ms-1 (for 5
ms-1 at z = 10 m) in all case studies, including wide streets (H/W = 0.5) and complex
geometries. In fact, this is due to the relatively important length of the canyon investi-
gated. This agrees with available knowledge on urban canyon wind flow (e.g. Naka-
mura and Oke 1988, de Paul and Shieh 1986, Santamouris et al. 1999). A parallel inci-
dence of wind, leads to much higher wind speeds due to channelling within the canyon
(e.g. for H/W = 2, 3.4 ms-1 > v > 2.4 ms-1 ). The wind speed decreases then progres-
sively along the street up to a critical distance and is accelerated again. The magnitudes
of channelling and acceleration increase with higher aspect ratios.
178
7.2. Heat gained by a human body
The simulation results showed that the radiation fluxes expressed by Tmrt are visibly
more sensitive to street geometry and orientation than air temperature or wind speed.
This dependence is evident temporarily (evolution throughout the day) as well as spa-
tially (i.e. centre and edges of the street) and points out the great potential of climate
control through design for meeting comfort purposes.
The individual energy components of Tmrt were dealt with in detail in the measurements
made in Freiburg in order to clarify the most critical components that affect the thermal
sensation of a standing person in relation to its surroundings. In addition, a similar
analysis was conducted for selected points from the simulated case studies. The meth-
ods used for calculating the heat gained by a pedestrian differed between simulations
and measurements, making difficult a comparison of each individual component. Never-
theless, a number of common results could be drawn. The heat gained by a standing
person outdoors depends strongly on the exposure to direct solar radiation of:
↓
- the body itself, given by the projection factor and solar intensity (fp, Rsw ,dir ), and
- the surrounding surfaces, which provide additional radiant heat to the body.
The maximal amount of heat gain estimated for a pedestrian is recorded for irradiated
locations, especially in the early morning and late afternoon, because of the high value
of the projection factor fp as the sun position is relatively low, leading to a large amount
of direct solar radiation absorbed by a standing person. This thermal stress is amplified
when the air temperature reaches its maximum, typically in the afternoon.
Moreover, the standing person absorbs more energy as long-wave irradiance than short-
wave irradiance: as a first approximation (from the measurements in Freiburg) about 70
% against 30% in an E-W street with H/W = 1. Consequently, shading is an efficient
strategy because it keeps the surfaces cooler as well as because of shading the person
itself.
A strong correlation was found between the total long-wave irradiance absorbed by a
standing person and the long-wave irradiation emitted by nearby sunlit surfaces and
confirms that the ground surface is particularly important as suggested by Watson and
Johnson (1988). The long-wave heat absorbed laterally is found to be relatively ho-
mogenous if compared to the complex irradiation patterns in-canyon from all directions.
179
This supports the choice made in ENVI-met to set the ground radiant heat to 50 % and
to decide on an average value for the radiant heat from the walls in the calculation of
Tmrt. This also draws attention on the importance of the thermal properties of the urban
canyon materials as already suggested by the measurements in Ghardaia. Explicitly,
high thermal admittances (heat capacity) decrease the surface temperatures and thus the
emitted radiant heat. Nocturnal cooling can, however, be slow down in case of deep
canyons because of a limited sky view factor. Possible evaporation from the ground is
also advisable for decreasing the radiant heat from the ground as observed in the planted
streets investigated.
7.3. Street design and outdoor thermal comfort

7.3.1. Aspect ratio and solar orientation
The results obtained with ENVI-met on the effects of street design on comfort during
summer in a hot-dry climate can be summarized as follow:
Thermal comfort is very difficult to reach passively in hot-dry climates at subtropical
latitude and summer conditions (e.g. in Ghardaia: Ta ≈ 40°C, RH ≈ 35 %). In effect,
Arnfield (1990a) and Bourbia and Awbi (2004) suggested the substantial irradiation of
the street surface for the subtropics (∼ 20°N to 40°N) even for deep geometries. PET
maxima reached 68 °C and PET minima were in all cases by few degrees higher than Ta
(up to 4 K). Nevertheless, an improvement is possible by means of appropriate design
since both aspect ratio H/W and solar orientation were found to affect strongly the out-
door thermal comfort at street level.
Wide streets, e.g. H/W ≤ 0.5, are highly uncomfortable during the largest part of
daytime for both orientations. These are largely irradiated and have high air tem-
peratures (almost equal to that above an unobstructed surface). Yet, N-S streets have
a small advantage over E-W streets as the thermal conditions at their edges along the
walls are less stressful. Hence, for shallow canyons, implementing shading strategies
at street level (galleries, trees, etc.) is the only way to improve substantially the com-
fort situation.
Increasing the aspect ratio ameliorates the thermal comfort for both E-W and N-S
orientations. However, the mitigation of heat stress is by far more effective for N-S
streets than for E-W streets. Indeed, E-W streets are for a longer period of time and
for a larger area of the street uncomfortable compared to N-S streets of the same as-
180
pect ratio. The contrast in the thermal sensation between the two orientations be-
comes more manifest for deeper canyons. By increasing the aspect ratio, PET
maxima decrease noticeably for N-S streets (about 58 °C) whereas still by 66 °C for
E-W canyons. The highest discomfort period occurs in the late afternoon for E-W
streets as the sun rays, impinging laterally, irradiate a large surface of the pedestrian.
This is combined with maximal daily Ta and large amounts of heat released by the
ground due to a permanent irradiation. In contrast, N-S streets are uncomfortable at
midday hours but have lower maxima. This is attributable to less solar irradiance ab-
sorbed as fp is minimal. This is combined with lower Ta and lower heat released by
the shortly irradiated ground.
Moreover, PET patterns showed two different distributions in respect to street ori-
entation, i.e. spatial or across canyon and temporal or diurnal evolution. Explicitly,
E-W streets show simultaneously two zones, on south and north side, with contrast-
ing thermal stress, while N-S canyons experience globally the same comfort condi-
tions on the whole area across the street. It is difficult to keep an E-W street under
comfort conditions because the effectiveness of building walls in intercepting sun’s
rays received laterally from the sides is limited, unless a very high aspect ratio is
chosen, e.g. H/W = 4 or higher.
Intermediate orientations NE-SW and NW-SE show some similarity in the temporal
and spatial evolution of the thermal situation within a N-S oriented canyon for the
same aspect ratio (e.g. H/W = 2). Yet, the discomfort period lasts about 2 hours
longer in the former cases. By contrast, these orientations experience a noticeably
shorter period of time of discomfort than E-W streets, with the street being always
partly in shade. This suggests that intermediate orientations are potentially good al-
ternatives for combining summer and winter needs in relation to solar energy, i.e.
shading and solar access, respectively, as these orientations offer a better summer
comfort than E-W orientation and a larger potential of internal solar gains than a N-
S orientation.
Experimentally, on-site measurements in Ghardaia confirmed the very high thermal
stress for unobstructed locations. Shading through high aspect ratios was proven to
be of prime importance in any improvement of thermal sensation. Street orientation
showed some evidence to play a role although its relevance has been partially inhib-
ited by the systematic deep geometries studied. A preference for N-S streets could
be observed, as these are only overheated around noontime. Maximum PET values
181
obtained on-site in Ghardaia were lower than those calculated by ENVI-met. The
reasons to this are twofold:
o The wind speeds were much higher on-site at the market place in comparison to
the very weak wind modelled at mid-block distance in-canyon
o The radiation fluxes calculated by the model are higher, as no adjusting factor
was used to match the measured data (a systematic difference in G was ob-
served, up to 100 Wm-2 for Gmax). This difference rises with increasing aspect ra-
tios because of more diffusely reflected radiation.
o The absorbed direct solar radiation flux by a standing person is overestimated
because of high values of fp, as calculated by equation 3.43b.
7.3.2. Asymmetry, galleries and overhanging façades
Using galleries revealed to be beneficial for mitigating thermal stress. This is due to the
reduced direct solar radiation received by a human body and to less long-wave irradia-
tion received from the surrounding surfaces, in particular the ground. No perceptible
reduction of air temperature was found under the galleries, and this disagrees with the
numerical results of Swaid et al. (1993). Discomfort can shortly extend under galleries
when the sidewalks in the “open” street area already experience extreme thermal stress.
This is a consequence of direct irradiation of the pedestrian and the ground surface. This
is more marked for wide canyons and depends on street orientation and gallery’s height
and width. The galleries of an E-W street are the best protected as well as a SE gallery
in a NE-SW oriented street. For all other orientations, the galleries are for 1 to 3 hours
highly stressful for H/W = 2. This intrusion can be reduced for higher aspect ratios or
with galleries of low height.
The asymmetrical profile (H2 > H1) revealed a clear similarity in the comfort situation
with the symmetrical profile H1/W owing to the almost same exposure to the sun. Yet, it
offers better comfort conditions in the early morning and late afternoon, suggesting that
it could be preferred if solar access inside the buildings in winter has also to be satisfied
while keeping a higher plan density. An evidence of faster nocturnal cooling is also ob-
served in case of asymmetry.
The design of overhanging façades as horizontal shading devices (e.g. balconies) help to
increase the area and duration of shade within the street and reduce further the heat
stress. E-W streets appear to be the best protected in comparison to N-S orientations and
182
also to NE-SW and NW-SE orientations. This solution is advisable if combined with an
asymmetrical profile: on one hand, there is more shading at street level in the summer,
and on the other hand more internal solar access is ensured in the winter. Moreover,
these “self-shading” facades reduce the overheating of indoor spaces by less warming of
their surfaces and hence less heat conduction towards indoors.
According to the large surface temperature differences between shaded and sunlit sur-
face observed in the field measurements, it is believed that the positive effects on com-
fort by using galleries and further horizontal shading devices on façades may be greater
than assessed by the model since ENVI-met uses an average wall temperature for the
whole model area. This certainly overestimated the heat released by shaded surfaces,
which are considered as the main emitting surroundings (SVF is very low).
7.3.3. Vegetation
The use of a row of trees is found to improve the thermal comfort situation within the
canyon. On one hand, Ta is moderately reduced and on the other hand the direct solar
radiation is also strongly decreased. Shading is the main property of the vegetation that
leads to heat stress mitigation, rather than evapotranspiration given the lack of water in
the soil. Thermal cooling effects (i.e. lower PET) are directly perceptible in the shade of
the tree crowns. However, almost not extent of these advantages could be found in the
surrounding space, as found experimentally by Shashua-Bar and Hoffmann (2000).
Trees are often considered as a costly strategy and their climatic usefulness has to be
maximized, e.g. preferably set at locations where discomfort is high and lasts a long
period of time. This has been made clear by the PET representation used in the present
study, i.e. across the street and throughout the day. Hence, it appears that a distinction
has to be made between wide profiles (H/W ≤ 0.5) and deeper profiles (H/W > 0.5) in
respect to tree planting. For wide streets, the vegetation is, basically, a good strategy for
both orientations since almost the entire street area is uncomfortable during all the day.
Depending on the use of the street, a row of trees (central or on the sidewalks) may be
planned on the pedestrian areas. For deeper urban streets, vegetation seems to be more
relevant for E-W than for N-S orientation because of the much longer period of discom-
fort in the former case. For N-S streets with an aspect ratio higher than 1, the time of
discomfort is limited to a short period around noon and may not necessitate planting.
183
The optimal location of trees within the street canyon also depends on street orientation
and on aspect ratio. For E-W orientation, the highest discomfort period occurs on the
north side during a large part of the day suggesting the use of trees at this location.
7.4. Recommendations and design examples
The discussion above revealed the interdependence of all investigated design aspects on
the resulting comfort within a street area. This study also showed that the thermal situa-
tion could be properly observed by investigating the whole space of the street versus
one central point within a street, and including simple and complex canyon shapes.
Manifold design possibilities are, hence, possible for controlling the microclimate. From
a climatic point of view, shading is the key strategy for promoting comfort in hot-dry
climate because it leads to:
a reduction of the direct solar radiation absorbed by a standing person
a reduction of the heat released by the surroundings, in particular the ground.
a decrease of the air temperature as a second effect.
Several design possibilities based on promoting shading can be suggested: i) a judicious
combination of aspect ratios and orientation ii) by arranging galleries, planting trees,
greening the façades or by using other shading devices for the irradiated wall and
ground surfaces.
The examples below seek to illustrate realistic situations and support the following dis-
cussion. Although, the study was mostly completed for a subtropical location with ex-
treme hot-dry climate for summer conditions, it is believed that the design recommenda-
tions discussed here can be more efficient for transitional seasons and also applicable to
less extreme climates such as mid-latitudes with typical hot summers as shown for
Freiburg. The Mediterranean basin, for instance, experiences to a large extent similar
irradiation potentials in the hot season (see Arnfield 1990a). Obviously, some adjust-
ments related to sun course geometry (zenith and azimuth angles) accounting for lati-
tude differences have to be considered (Arnfield 1990a, Mills 1997).
Designing a street is primarily conditioned by:

1. Street utility: structural role of the street in the whole urban plan, implying scale (i.e.
absolute dimensions: width and height), activity, and usage (pedestrian streets or in-
cluding motor traffic). This has a direct impact on the period of time at which com-
184
fort is essential (frequentation time by people) and also the area of the street where
comfort is at most required (whole area, edges, etc.).
2. Building usage: domestic (housing) or non-domestic (e.g. office or educational
buildings). Domestic buildings are concerned with comfort the day round and re-
quire passive solar gains. South, south-east or east exposures of the façades are the
most suitable. In non-domestic buildings, comfort is mostly crucial during the day-
time where day-lighting is the main concern. The potential of natural light is almost
equal for all solar orientations and is much more sensitive to sky view, i.e. aspect ra-
tio.
Usually, the street orientation is chosen first and the aspect ratio is set according to the
orientation. The street orientation, if not already determined by non-climatic arguments
(site constraints, surrounding built environment, etc.) should take into account the needs
for solar energy inside the buildings and as much as possible pay attention to the domi-
nant wind directions for promoting ventilation or protecting from cold winds (Ali-
Toudert and Bensalem 2001). An E-W orientation is well known to be preferable if so-
lar gains have to be maximized. Intermediate orientations are less optimal but still pro-
vide a good potential of sunlight and daylight. N-S orientation is appropriate for day-
light issues but requires a good protection of the façades from the sun in the summer.
As a first rule of thumb, the urban canyon can be divided into two parts: i) street level
and ii) building part (Fig 7.1), which refer to outdoor and indoor issues, respectively.
The street area is in turn subdivided into 3 sub-spaces: the central part, the edges and
possible extensions in the building basement in form of galleries.
β2
α2 β1
α1
indoor
climate
indoor
climate
1. centre
street
climate 2. sidewalk
3 2 1 2 3 3. gallery
Fig. 7.1. Scheme on the subdivision of a street canyon volume according to climatic
design needs
185
Given that the passive solar gains are needed only in the upper building part, then, the
“effective” aspect ratios for the façades (expressed by α1 and α2) are less restrictive
than the absolute aspect ratios, applicable for the street area (β1 and β2). Secondly, it is
advisable to offer a diversity of arrangements at street level in order to increase the
probability for a sustained frequentation of an outdoor space. The previous results put
forward the necessity of differentiating between wide and deep streets, say H/W ≤ 1
versus H/W > 1 as a first approximation.
Wide streets allow a good solar access in the winter but are highly uncomfortable in the
summer at street level. Detail arrangements are thus required. Deep streets are better
protected in the summer but do not support winter issues.
Fig 7.2 shows a typical wide street flanked by 2 rows of trees on both sides, either ori-
ented N-S. Large and high trees act for shading on the lateral sides while the central part
is foreseen for motor traffic and kept untreated. In case of E-W orientation, adding gal-
leries is advisable. Deciduous species with sufficient distance from the north wall avoid
the obstruction of desirable solar gains in the winter.
Fig 7.3 shows another example of a wide street canyon of H/W = 0.6. The street is E-W
oriented and allows optimal internal solar gains on the south façade. In this case, the
largest part of the street would be highly uncomfortable if no shading strategies are
planned (see Fig. 4.5a).
The street area is divided in sub-spaces consisting of pedestrian areas and motor traffic
areas. Pedestrian areas are placed on the southern half part of the street, arranged under
galleries or protected by vegetation. Trees are planned so that they maximize shade
through their large crowns and high size. At the same time they are at some distance
from the south facade to prevent overshadowing in the winter. Trees are preferably de-
ciduous in order to save solar access indoors in winter and for people sitting outdoors in
the winter.
Traffic is also located on the north side on the potentially most uncomfortable location.
On the south side a deciduous tree can be added to prevent from overheat in the early
morning and late afternoon in the summer season if required by the activity taking place
at that part of the canyon. The relatively large aspect ratio promotes a rapid nocturnal
cooling.
186
Fig. 7.2. Example of wide canyon combining motor traffic and pedestrian areas pro-
tected by deciduous trees
Fig. 7.3. Example of wide street canyon oriented E-W, combining comfortable pedes-
trian zones and motor traffic
Fig 7.4 shows an example of higher aspect ratio with an asymmetrical geometry espe-
cially advisable when high plan density is required. The walls have to face the sun, i.e.
street axis preferably oriented E-W, NE-SW or NW-SE. The flanking buildings have
different heights. The wall facing the sun predominates with a large openness to sky of
the building part promoted by the lower height of the opposite building.
187
Fig. 7.4. Example of an asymmetrical canyon combining summer comfort at street

level, winter solar access and high density
At street level, horizontal overhangs, e.g. in form of galleries and trees are planned to
keep the whole space comfortable since exclusively foreseen for pedestrian use. More-
over, a special attention has to be given to the surfaces themselves. Ground pavements
should be preferably of light colour, porous and/or of thin layer materials to keep lower
surface temperatures. Pavements mixed to green surfaces for promoting evaporation
from underground are also advisable (Asaeda and Ca 1993, Asaeda et al. 1996, Ca et al.
1998), especially in latitudes where summers are not dry like in Freiburg. Building ma-
terials also play a role: High thermal capacity and high albedos (e.g. white colours)
help to reduce the surface temperatures further and thus the heat released.
Horizontally, the street canyon can also be differentiated into street corners and street
centre with respect to ventilation aspects (see Fig. 4.26). The incidence of the wind
upon the street axis is decisive and a parallel direction promotes more air movement
than a perpendicular one. Increasing the building heights leads to a larger zone of inter-
ference at street corners. For stretched canyons, the wind speed in the central part is
isolated from ambient wind, suggesting that, for better ventilation, canyons of limited
length are more suitable than long stretched canyons.
An urban structure composed of small size buildings and possibly staggered in a
checker-board pattern, may be preferred, as this promotes a much uniform wind flow
and eliminate stagnant air zones (e.g. Koenigsberger et al. 1973, Asimakopoulos et al.
188
2001). Since ventilation is strongly reduced even for wide canyons in case of perpen-
dicular flow, it is suggested that buildings of medium height with moderate aspect ratios
provide better ventilation than low-rise buildings with small aspect ratios, due to
stronger convergent flow from the sides and higher wind speeds. An oblique incidence
of wind as showed in the case of Freiburg offers good ventilation potential at street level
as reported by previous studies (e.g. Wiren 1985, 1987, Bensalem 1991).
7.5. Limits and current development of ENVI-met
Methodologically, ENVI-met 3.0 revealed to be a good tool for the prognosis of micro-
climatic modifications due to urban environments and for assessing the thermal comfort
of pedestrians. Indeed, the model has a well-founded physical basis and offers many
advantages in comparison to many other available urban microclimate models. Yet, and
as for many models, very few validation studies are available to date (Arnfield 2003). In
the following material, some observations on the performance and accuracy of ENVI-
met, made in the framework of this study, are reported. This information helps to under-
stand the limits of the present investigation, and gives an overview on eventual refine-
ments of the model.
7.5.1. Boundary conditions
Initial conditions, i.e. inputs for the start time of simulation, are given at the boundary
height of 2500 m and serve, using the 1D model, to re-create the large surrounding envi-
ronment for the main simulation volume assessed by the 3D-model. These inputs in-
clude the potential temperature (θ), air humidity (VP) and wind speed (v, z = 10 m
a.g.l.) and are kept constant at 2500 m during the simulation. For the soil, temperature
and humidity values at 1.75 m depth are also kept constant. Test simulations are usually
necessary for the determination of the appropriate start values for a specific climate
situation. These unchanged boundary conditions have the following consequences:
1. Air temperature: Test simulations showed that ENVI-met reproduces well the daily
cycle of air temperature for a mid-latitude location like Freiburg. However, for a conti-
nental and subtropical region like Ghardaia, it showed a trend to underestimate the typi-
cal large daily amplitudes where Tmax - Tmin ≈ 12 K in summer. If start values are appro-
priately chosen for the daytime, the simulated data are then up to 4 K higher by night.
189
One direct consequence on this study was the limitation of the simulation period to day-
time. This is sensible for comfort purposes; however no assessment of the nocturnal
cooling was possible.
2. Wind: non-stationary assessment of the wind speed over the day is possible with
ENVI-met, which means that the daily thermal effects are included. Yet, the model as-
sumes no change in the geostrophic wind at 2500 m. This makes it difficult simulating
locations where the wind is strongly influenced by local climate: A coastal location, for
example, is strongly influenced by marine breezes which results in sensible changes in
the wind speed and direction and affect even air temperature daily cycle. Test simula-
tions made for the coastal location of Algiers (36°N) showed these patterns.
The ongoing development of ENVI-met considers the possibility of forcing the model
with external measured data at the boundary of 2500 m height, so that a progressive
adjustment of the 3D-model data becomes possible. This feature is yet at an early stage
of experimentation and will eventually apply to air temperature, wind speed and wind
direction (Bruse 2004).
3. Radiation fluxes: In the calculation of incoming short-wave radiation, the absorption
in the atmosphere takes into account the water vapour but no attenuation by other gases
(e.g. CO2 or O3) or aerosols is included. This is usual but leads to overestimated values,
in particular in urban areas. ENVI-met gives in its recent version the possibility to ad-
just the incoming solar radiation amount by means of a factor varying from 0.5 to 1.5
(i.e. 50% to 150%). This function was not used in this study and consequently, the solar
radiation was overestimated if compared to the field measurements gathered in Ghar-
daia. Heat stress calculated is, hence, also slightly overestimated.
7.5.2. Heat storage in the building materials
The energy balance of the wall and roof surfaces in ENVI-met takes into account a con-
ductive heat flux, exclusively depending on heat transmittance (U-value) and the gradi-
ent between internal and external wall surface temperatures. No heat storage within the
materials is taken into account. In real cases, each material has thermal properties which
define its ability to accept or release heat and is expressed for a surface by thermal ad-
mittance µ = (k C )
0.5
where k and C are the thermal conductivity and the heat capacity,
respectively. A review of available models shows that the storage in the materials is
190
often neglected (e.g. Todhunter 1989, Sievers and Zdunkowski 1986, Mills 1993) or
only considered together with simplifications of other atmospheric processes such as the
wind prognosis (e.g. Arnfield et al. 1998, Arnfield and Grimmond 1998, Groleau 1998,
Arnfield 2000). The lack of heat storage in ENVI-met has a double effect on the simula-
tion results: First, the instantaneous surface temperature is overestimated, and secondly,
no heat can be released after sunset since not stored. Hence, the long-wave radiation
emitted by the walls is overestimated in the daytime and underestimated by night. This
was a second argument for limiting the study to diurnal conditions. Moreover, whether
the asymmetry of a street canyon is significant in improving the nocturnal cooling could
not be assessed. Moreover, the air temperature was found to be almost uniform in the
canyon air volume with insignificant warming up of air close to irradiated surfaces. This
disagrees with field study results, e.g. Nakamura and Oke 1988 and the results obtained
in Freiburg (Chapter 5). This can be partly attributable to the lacking storage of heat.
The potential effects of wall thermal properties in reducing the amount of long-wave
radiation absorbed by a person from the surroundings could not be assessed.
Taking into account this feature would require an additional sub-model involving each
grid point on each façade and roof. This is not foreseen in the next update versions,
among others because of the high processing time implied by the size and resolution of
the model area (x, y, z : 250 x 250 x 100 grids. The last release of ENVI-met 3.0 (Oct.
2004) uses a coefficient of 0.5 for reducing the impinging solar radiation on façades in
order to counterbalance the effects of lacking storage on the wall surface temperatures.
7.5.3. Mean radiant temperature and comfort
The present study confirmed the prime importance of the mean radiant temperature on
outdoor comfort. Yet, the discussion in section 2.2.3 underlined the difficulty of esti-
mating Tmrt using both measurements or modelling. A number of observations can be
made on Tmrt prognosis method used by ENVI-met by comparison to the measurements
gathered in Freiburg:
In spite of the different methods used for calculating Tmrt, simulated results provided a
satisfactory agreement with measured data given the complexity of the urban environ-
ment when the projection factor fp is appropriately set, e.g. as in chapter 5, section 5.4.
This fp as given by eq. 2.43b (former version of the model) has lead to abnormally high
values of Tmrt in chapter 4 and hence overestimates the heat stress.
191
Positive aspects in calculating Tmrt by ENVI-met are:

the relative importance accorded to radiant heat from the ground by setting it to 50%
of the whole long-wave radiant energy which has been verified experimentally.
the use of the sky view factor for differentiating between the radiation components
the precise modelling of the short-wave solar radiation.
The calculation of the long-wave radiation for each grid point at a given time is esti-
mated using an average quantity Tw which is a mean surface temperature for all walls
within the model, indifferently whether the person stands nearby a sunlit or shaded wall.
This simplification shows to be sensible for the canyon of H/W = 1 in comparison to
measurements as the lateral radiant environment was found to be relatively homogenous
and because of the more decisive role of the ground.
The relevance of the sky view factor as weighting factor could not be verified for higher
aspect ratios (no measured data available). Yet, the simulations results revealed that Tmrt
in largely shaded locations like galleries is overestimated because the fractions of long-
wave energy gained from walls vs. sky, as well as the short-wave diffuse radiation are
systematically affected by SVF.
Finally, adding further outputs in ENVI-met can be worthwhile: the irradiation of build-
ing surfaces in order to integrate the solar access is, in our opinion, useful to enlarge the
model to indoor climate issues for a comprehensive analysis. As well, a display of the
street level area (i.e. in plan XY) of the sunshine duration for each grid point is helpful
as first working information for a design practitioner, because it highlights the timely
longest irradiated areas versus shaded, and so potentially the most critical areas needing
corrective measures through design in summer. These will be added in next releases of
the model.
7.6. Concluding remarks
The present study was particularly motivated by the will to link between the theoretical
knowledge on urban microclimate and the practical design process, as this was widely
reported to be lacking. Surprisingly, the literature review also reports on a dramatic lack
of studies, either experimental or numerical, which deal directly with the effects of ur-
ban geometry on human comfort and thus make this transition difficult. This underlines
192
the relevance of the present study on one hand, but also draws attention to the difficulty
of comparison with similar studies on the other hand. Further work is hence needed:
A number of interesting questions arise in the course of this study:

1. Field measurements are required for validating numerical results, from which
the effects of the urban parameters investigated in this study, i.e. aspect ratio, orien-
tation, vegetation, galleries and self-shading façades. This is particularly sensible for
a better estimation of the energy gained by a human body (and consequently Tmrt),
which can improve modelling parameterisations.
2. Extend the study to night-time situation by investigating the effects of urban
geometry on the nocturnal cooling (of the street and buildings).
3. Human comfort is a multifaceted issue, which combines physical, physiological
and psychological dimensions. An overview of available studies pointed out im-
portant methodological differences in assessing comfort. It is still difficult to inter-
pret the actual human thermal sensation from the currently used thermal indices.
Complementing energy-based methods with adaptive methods (social survey’s) is
necessary as a next research stage for a better understanding of human comfort
(Brager and De Dear 1998, Spagnolo and de Dear 2003) and eventually setting a
universally applicable tool for comfort evaluation.
4. More connection between architectural and urban scale is highly advisable, since
urban buildings are in the practice primarily conceived to cope with indoor comfort.
Developing microscale numerical tools which assess simultaneously the effects of
urban geometry on outdoor and indoor climate (i.e. energy efficiency of buildings)
is a promising alternative (Mills 1999).
193
References
References
Akbari H., Rosenfeld A.H., Taha H. 1995: Cool construction materials offer energy
saving and help reduce smog. ASTM standardization news 23, 11: 32 -37.
Ali-Toudert F. 2000: Intégration de la dimension climatique en urbanisme. Mémoire de

Magister, EPAU, Alger.
Ali-Toudert F., Bensalem R. 2001: A methodology for a climatic urban design. Proc.
18th Int. Conf. on PLEA, Florianopolis, Brazil: 469-473.
Ali-Toudert F., Mayer H. 2006: Numerical study on the effects of aspect ratio and solar
orientation on outdoor thermal comfort in hot and dry climate. Building and Envi-
ronment 41: 94-108.
Ali-Toudert F., Djenane M., Bensalem R., Mayer H. 2005: Outdoor thermal comfort in
the old desert city of Beni-Isguen, Algeria. Climate Research 28: 243-256.
Ali-Toudert F., Mayer H., 2005: Thermal comfort in an east-west oriented street canyon
in Freiburg (Germany) under hot summer conditions. Theor. Appl. Climatol. (in
press).
Arnfield J. 1990a: Street design and urban canyon solar access. Energy and Buildings
14: 117-131.
Arnfield J. 1990b: Canyon geometry, the urban fabric and nocturnal cooling: a simula-
tion approach. Phys. Geogr. 11: 220-239.
Arnfield J., Mills G. 1994: An analysis of the circulation characteristics and energy
budget of a dry, asymmetric, east-west urban canyon. II. Energy budget. Int. J.
Climatol. 14: 239-261.
Arnfield J., Herbert, J.M., Johnson, GT. 1998: A numerical simulation investigation of
urban canyon energy budget variations. Proc. 13th Int. Conf. on Biometeorol. and
Aerobiol., Albuquerque, New Mexico, AMS: 2-5.
Arnfield J., Grimmond C.S.B. 1998: An urban canyon energy budget model and its ap-
plication to urban storage heat flux modelling, Energy and Buildings 27, 61-68.
Arnfield J. 2000: A simple model of urban canyon energy budget and its validation.
Phys. Geogr. 21: 305-326.
194
References
Arnfield J. 2003: Two decades of urban climate research: A review of turbulence, ex-
changes of energy and water, and the urban heat island. Int. J. Climatol. 23: 1-26.
Asaeda T., Ca V.T. 1993: The subsurface transport of heat and moisture and its effects
on the environment: a numerical model. Boundary−Layer Meteorol. 65: 159-179.
Asaeda T., Ca V.T., Wake A. 1996: Heat storage of pavement and its effect on the
lower atmosphere. Atmos. Envir. 30: 413-427.
Asawa T., Hoyano A., Nakaohkubo K. 2004: Thermal design tool for outdoor space
based on numerical simulation system using 3D-CAD. Proc. 21th Int. Conf. on
PLEA, Eindhoven. Netherlands. Vol. 2: 1013-1018.
ASHRAE 2001a: Chapter 8 – Comfort. In: Handbook of Fundamentals. American Soci-

ety for heating Refrigerating and Air Conditioning, Atlanta: 8.1-8.29.
ASHRAE 2001b: Chapter 13 – Measurements and instruments. In: Handbook of Fun-

damentals. American Society for heating Refrigerating and Air Conditioning, At-
lanta: 13.26 –13.27.
Asimakopoulos D. N., Assimakopoulos V. D., Chrisomallidou N., Klitsikas N., D.

Mangold, Michel P., Santamouris M., Tsangrassoulis A. 2001: Energy and climate
in the urban built environment. James & James. London.
Avissar R. 1996: Potential effects of vegetation on urban thermal environment. Atmos.

Envir. 30: 437-448.
Baik J-J., Park R-S., Chun H-J., Kim J-J. 2000: A laboratory model of urban street can-
yon flows. J. Appl. Meteorol. 39: 1592-1600.
Bärring L., Mattsson J.O., Lindqvist S. 1985: Canyon geometry, street temperatures and
urban heat island in Malmö, Sweden. J. Climatol. 5: 433-444.
Barry G.R., Chorley R.J. 1978: Atmosphere, weather and climate. Methuen & Co Ltd.
London. Chapter 7: Urban and forest climates: 322-352.
Bensalem R. 1991: Wind driven natural ventilation in courtyard and atrium-type build-
ings. PhD thesis. University of Sheffield. Department of Building Science. UK.
Bitan A. 1988: The methodology of applied climatology in planning and building. En-
ergy and Buildings 11: 1-10.
Bitan A. 1992: The high climatic quality city of the future. Atmos. Envir. 26B: 313-329.
195
References
Bourbia F., Awbi H.B. 2004: Building cluster and shading in urban canyon for hot-dry
climate. Part 2: Shading simulations. Renewable Energy 29: 291-301.
Brager G.S., de Dear R.J. 1998: Thermal adaptation in the built environment: a litera-
ture review. Energy and Buildings 27: 83-96.
Brazel A.J., Arnfield A.J., Greenland D.E., Willmott C. J. 1991: Physical and boundary
layer climatology. Phys. Geogr. 12: 189-206.
Brown R.D., Gillespie T. 1986: Estimating outdoor thermal comfort using a cylindrical
radiation thermometer and an energy budget model. Int. J. Biometeorol. 30: 43-52.
Brown G., Isfält E. 1974: Solar irradiation and shading devices. Report R19. Nat. Swed-
ish Council for Building Research, Stockholm.
Bruse M., Fleer H. 1998: Simulating surface-plant-air interactions inside urban envi-
ronments with a three dimensional numerical model. Envir. Model. Software 13:
373-384.
Bruse M. 1999: Die Auswirkungen kleinskaliger Umweltgestaltung auf das Mikrokli-

ma. Entwicklung des prognostischen numerischen Modells ENVI-met zur Simula-
tion der Wind-, Temperatur-, und Feuchtverteilung in städtischen Strukturen. PhD
Thesis, Univ. Bochum, Germany.
Bruse M. 2004: ENVI-met website. http://www.envi-met.com.
Bussinger J.A., Wyngaard J.C., Izumi Y., Bradley E.F. 1971: Flux-Profile relationships
in the atmospheric surface layer J. Atmos. Sci. 28: 181-189.
Ca V.T., Aseada T., Ashie Y., 1998: Utilization of porous pavement for the improve-
ment of summer urban climate. Proc. 2nd Japanese-German Meeting “Klimaana-
lyse für die Stadtplanung”. Research Centre for Urban Safety and Security. Kobe
University, Sp. Rep. No. 1: 169-177.
Canton M.A., Cortegoso J.L., De Rosa C. 1994: Solar permeability of urban trees in
cities of western Argentina. Energy and Buildings 20: 219-230.
Capeluto I.G. 2003: Energy performance of the self-shading building envelope. Energy
and Buildings 35: 327-336.
Capeluto I.G., Shaviv E. 2001: On the use of solar volume for determining the urban
fabric. Solar Energy 70: 275-280.
196
References
Chalfoun N.V. 2001: Thermal comfort assessment of outdoor spaces using MRT© and
fish-eye lens photography of architectural scale models: a case study of the “arts
oasis” plaza at the university of Arizona, USA. Proc. 18th Int. Conf. on PLEA,
Florianopolis, Brazil: 1021-1025.
Chan A.T., So E.S.P., Samad S.C. 2001: Strategic guidelines for street canyon geometry
to achieve sustainable street air quality. Atmos. Envir. 35: 5681-5691.
Clapp R.B., Hornberger G. 1978: Empirical equations for some soil hydraulic proper-
ties. Water Resource Research, 14: 601-604.
Coronel J.F., Alvarez S. 2001: Experimental work and analysis of confined urban
spaces. Solar Energy 70: 263-273.
Dabberdt W.F., Ludwig F.L., Johnsson W.B. 1973: Validation and applications of an
urban diffusion model for vehicle emissions. Atmos. Envir.7: 603-618.
Deardoff R.W. 1978: Efficient prediction of ground surface temperature and moisture
with inclusion of a layer of vegetation. J. Geophys. Research, 83: 1189-1903.
De Freitas C.R. 1985: Assessment of human bioclimate based on thermal response. Int.
J. Biometeorol. 29: 97-119.
De Paul F.T., Shieh C.M. 1986: Measurements of wind velocity in a street canyon. At-
mos. Envir. 20: 455-459.
Djenane M. 1998: Participation de la forme urbaine au contrôle de l’irradiation solaire.

Référence particulière au rôle de la rue dans les régions chaudes et sèches. Mé-
moire de magister, Centre universitaire Mohamed Khider, Biskra, Algérie.
Doll D., Ching J.K.S., Kaneshiro J. 1985: Parameterization of subsurface heating for
soil and concrete using net radiation data. Boundary−Layer Meteorol. 32: 351-
372.
Donnadieu C., Donnadieu P., Didillon J.M. 1977: Habiter le désert, maisons Mozabites.
Mardaga. Paris.
Eichorn J. 1989: Entwicklung und Anwendung eines dreidimensionalen mikroskaligen

Stadtklima-Modells. Dissertation. Univ. Mainz.
Eliasson I. 1993: Urban climate related to street geometry. PhD thesis. University of
Gothenburg. Dept. Phy. Geogr. GUNI rapport 33.
197
References
Eliasson I., Upmanis H. 2000: Nocturnal airflow from urban parks-implications for city
ventilation. Theor. Appl. Climatol. 66: 95-107.
Escourrou G. 1991: Le climat et la ville. Géographie d'aujourd'hui. Nathan. Paris.
Fanger P.O. 1970: Thermal comfort. Danish Technical Press. Copenhagen.
Gagge A.P., Burton A.C., Bazett H.D. 1971: A practical system of units for the descrip-
tion of heat exchange of man with his environment. Science 94: 428-430.
Gagge A.P., Fobelets A.P. Berglund L.G. 1986: A standard predictive index of human
response to the thermal environment. ASHRAE Transactions 92: 709-731.
Garrat J.R. 1992: The atmospheric boundary layer. Cambridge University Press. Cam-
bridge.
Geiger R., Aron R., Todhunter P. 1995: The climate near the ground. Vieweg. Wies-
baden.
Givoni B. 1976: Man, Climate and Architecture. Van Nostrand Reinhold. New York.
Givoni B. 1997: Climate considerations in building and urban design. Van Nostrand
Reinhold. New York.
Golany G. 1982: Design for arid regions. Van Nostrand Reinhold, New York.
Golany G. 1996: Urban design morphology and thermal performance. Atmos. Envir. 30:
455-465.
Groleau D., Miguet F. 1998: Solène et la simulation des éclairements directs et diffus
des projets architecturaux et urbains. IBPSA France '98. Sophia Antipolis: 61-65.
Gross G. 1991: Anwendungsmöglichkeiten mesoskaliger Simulationsmodelle darge-

stellt am Beispiel Darmstadt. Teil 1 Wind und Temperaturfelder. Meteorol. Rund-
schau 43: 267-274.
Grundström K., Johansson E., Mraisi M., Ouahrani, D. 2003 : Climat et urbanisme – la
relation entre confort thermique et la forme du cadre bâti. Report 8. Housing De-
velopment and Management. Lund University.
Hawkes D., Foster W. 2002: Energy efficient buildings, Architecture, Engineering, and
Environment. Norton. New York.
Heisler G.M. 1986: Energy savings with trees. J of Arboriculture 12: 113-124.
198
References
Herbert J.M., Johnson G.T., Arnfield J. 1998: Modelling the thermal climate in city
canyons. Envir. Model. Software 13: 267-277.
Herzog T. 1996: Solar energy in Architecture and urban planning. Prestel. Munich.
Honjo T., Takakura T. 1990/91: Simulation of thermal effects of urban green areas on
their surrounding areas. Energy and Buildings 15-16: 457-463.
Höppe P. 1992: Ein neues Verfahren zur Bestimmung der mittleren Strahlungstempera-
tur im Freien. Wetter und Leben 44: 147-151.
Höppe P. 1984: Die Energiebilanz des Menschen. Dissertation. Wissenschaftlicher Mit-

teilung Nr. 49. Universität München.
Höppe P. 1993: Heat balance modelling. Experientia 49: 741-746.
Höppe P. 1999: The physiological equivalent temperature - a universal index for the
biometeorological assessment of the thermal environment. Int. J. Biometeorol. 43:
71-75.
Höppe P. 2002: Different aspects of assessing indoor and outdoor thermal comfort. En-
ergy and Buildings 34: 661 - 665.
Hosker R.P.J. 1985: Flow around isolated structures and building clusters: a review.
ASHRAE Transactions 91: 1671-1692.
Houghton J.T. 1977: The physics of the atmosphere. Cambridge University press. New
York.
Houghton D.D. 1985: Handbook of applied Meteorology. John Wiley and Sons. New
York.
Houghton F.C., Yaglou C.P. 1923: Determination of the comfort Zone. ASHVE Re-
search report No. 673. ASHVE Transactions 29: 361.
Hoydysh W., Dabbert W.F. 1988: Kinematics and Dispersion characteristics of flows in
asymmetric street canyons. Atmos. Envir. 22: 2677-2689.
Hussain M., Lee B.E. 1980: An investigation of wind forces on the 3D roughness ele-
ments in a simulated atmospheric boundary layer flow. Part II- Flow over large ar-
rays of identical roughness elements and the effect of frontal and side aspect ratio
variations. Department of Building Sciences. Univ. of Sheffield, UK.
ISO 1989: 7243 Hot environments − estimation of the heat stress on working man,
199
References
based on WBGT index (wet bulb globe temperature). Int. Stand. Org.
Jacobs C.M.J. 1994: Direct Impact of atmospheric CO2 enrichment in regional transpi-
ration. PhD Thesis. Wageningen Agricultural University. Netherlands.
Jendritzky G., Nübler W. 1981: A model analysing the urban thermal environment in
physiologically significant terms: Arch. Met. Geoph. Biokl., Ser. B 29: 313-326.
Jendritzky G. Sievers U. 1989: Human-biometeorological approaches with respect to

urban planning. Proc. 11th ISB-Congress. SPB Academic publishing. The Hague,
Netherlands: 25-39.
Jendritzky G., Menz G., Schirmer H., Schmidt-Kessen W., 1990: Methodik zur räumli-
chen Bewertung der thermischen Komponente im Bioklima des Menschen. Fort-
geschriebenes Klima-Michel-Modell. Beiträge der Akademie für Raumforschung
und Landesplannung. Hannover, Band 114.
Johnsson G. T., Hunter L.J. 1995: A numerical study of dispersion of passive scalars in
city canyons: Boundary−Layer Meteorol. 75: 235-262.
Knowles R.L. 1981: Sun, Rhythm and Form. MIT press. London.
Knowles R.L. 2003: The solar envelope: its meaning for energy and buildings. Energy
Koenigsberger O.H., Ingersoll T.G., Mahyew A., Szokolay S.V. 1973: Manual of tropi-
cal housing and building. Part 1: climatic design. Longman. London.
Krishan A. 1996: The habitat of two deserts in India: hot-dry desert of Jaisailmer (Ra-
jasthan) and the cold-dry high altitude mountainous desert of leh (Ladakh). Energy
Kristl Z., Krainer A. 2001: Energy evaluation of urban structure and dimensioning of
building site using ISO-Shadow method. Solar Energy 70: 23-34.
Krys S.A.; Brown R.D. 1990: Radiation absorbed by a vertical cylinder in complex out-
door environments under clear sky conditions. Int. J. Biometeorol. 34: 69-75.
Kuttler W. 1993: Planungsorientierte Stadtklimatologie. Aufgaben, Methoden und Fall-

beispiele. Geogr. Rundschau 45. H. 2.
Kuttler W. 2004: Stadtklima. Teil 1:Grundzüge und Ursachen. Beitagsserie: Kilamände-

rung und Klimaschutz. Umweltchem Ökotox 1-12.
200
References
Landsberg H. E. 1981: The urban climate. Int. Geophys. Series Vol. 28. New York.
Launder B., Spalding D.B. 1974: The numerical computation of turbulent flow. Com-
puter Methods in Applied Mechanics and Engineering 3: 269-289.
Lechner N. 1991 Heating cooling, Lighting. Design methods for Architects. John Wiley
& Sons. New York.
Lee I.Y., Shannon J.D., Park H.M. 1994: Evaluation of parameterisations for pollutant
transport and Dispersion in an urban street canyon using a three-dimensional dy-
namic flow model. Proc. 87th Annual Meeting and Exhibition, Cincinnati, OHIO:
19-24.
Liljequist G. H. 1979: Meteorologie. Dept. of Meteorology. Univ. of Uppsala. Sweden.
Littlefair P.J., Santamouris M., Alvarez S., Dupagne A., Hall D., Teller J., Coronel J.F.,
Papanikolaou N. 2001: Environmental site layout planning: solar access, micro-
climate and passive cooling in urban areas. CRC. London.
Liu J., Chen J.M., Black T.A., Novak M.D. 1996: E-ε modelling of turbulent air flow
downwind of a model forest edge. Boundary−Layer Meteorol. 77: 21-44.
Markus T.A., Morris E.N. 1980: Buildings, Climate and Energy. Pitman. London.
Masson V. 2000: A physically-based scheme for the urban energy budget in atmos-
pheric models. Boundary−Layer Meteorol. 94: 357-397.
Mayer H., Höppe, P. 1987: Thermal comfort of man in different urban environments.
Theor. Appl. Climatol. 38: 43-49.
Mayer M. 1993: Urban bioclimatology. Experientia 49: 957-963.
Mayer H. 1998: Human-biometeorological assessment of urban microclimates accord-

ing to the German VDI-guideline 3787 part II. Prepr. 2nd Urban Envir. Symp. Al-
buquerque. New Mexico. AMS: 136-139.
McCormick R. A. 1971: Air pollution in the locality of buildings. Phil. Trans. R. Soc.
London. Ser. A. 269: 515-526.
McPherson E.G. 1992: Shading urban heat islands in U.S. desert cities. Wetter und Le-
ben 44: 107-123.
201
References
McPherson E.G. 1994: Energy-saving potential of trees in Chicago. Chicago's urban

forest ecosystem: Results of the Chicago urban forest climate project. USDA for-
est service, General technical report NE-186: 95-113.
McPherson E.G., Rowntree R.A, Wagar J.A. 1994a: Energy efficient landscapes. Urban
Forest Landscapes: integrating multidisciplinary perspectives. University of
Washington Press. Seattle: 150-160.
McPherson E.G., Nowak D.J., Rowntree R.A. 1994b: Chicago's urban forest ecosystem:
Results of the Chicago urban forest climate project. USDA forest service. General
Technical report NE-186.
McPherson E.G., Simpson J.R. 1995: Shade trees as a demand - side resource, Home
energy 2: 11-17.
Meier I.A., Roaf S.C., Gileard I., Runsheng T., Stavi, I., Mackenzie-Bennett J. 2004:
The vernacular and the Environment towards a comprehensive Research method-
ology. Proc. 21st Int. Conf. on PLEA, Vol. 2: 719-724.
Mellor G.L., Yamada T. 1975: A simulation of the Wangara atmospheric boundary

layer data. J. Atmos. Sci. 32: 2309-2329.
Mills G. 1993: Simulation of the energy budget of an urban canyon I. Model structure
and sensitivity test. Atmos. Envir. 27B: 157-170.
Mills G.M., Arnfield J. 1993: Simulation of the energy budget of an urban canyon - II.
Comparison of model results with measurements. Atmos. Envir. 27B, 171-181
Mills G. 1997: An Urban Canopy-Layer Climate Model. Theor. Appl. Climatol. 57:
229-244.
Mills G. 1999: Urban climatology and urban design. ICB-ICUC; Sydney, Australia:
541-544.
Missenard H. 1948: Equivalence thermiques des ambiances, équivalences de passage,

équivalences de séjour. Chaleur et Industrie. Jul-Aug.
Moughtin C. 2003: Urban design, street and square. Architectural Press. Amsterdam.
Nagara K., Shimoda Y., Mizuno M. 1996: Evaluation of the thermal environment in an
outdoor pedestrian space. Atmos. Envir. 30: 497-505.
Nakamura, Y., Oke T. 1988: Wind, temperature and stability conditions in an east-west
202
References
oriented urban canyon. Atmos. Envir. 22: 2691-2700.
Nikolopoulou M., Baker N., Steemers K. 2001: Thermal comfort in outdoor urban
spaces: understanding the human parameter. Solar Energy 70: 227-235.
Nübler W. 1979: Konfiguration und Genese der Wärmeinsel der Stadt Freiburg. Frei-
burger Geogr. Hefte, Heft 16.
Nunez M., Oke, T. R. 1977: The energy balance of an urban canyon. J. Appl. Meteorol.
16: 11-19.
Oke, T.R, 1976: The distinction between canopy and boundary - layer urban heat island.
Atmosphere 14: 268-277.
Oke T. 1981: Canyon geometry and the nocturnal urban heat island: comparison of
scale model and field observation. J. Climatol. 1: 237-254.
Oke T.R. 1987: Boundary Layer Climates. Methuen. London.
Oke T. 1988: Street design and urban canopy layer climate. Energy and Buildings 11:
103-113.
Oke T., Johnson G.T., Steyn D.G., Watson I.D., 1991: Simulation of surface urban heat
islands under “ideal” conditions at night. Part 2: Diagnosis of causation. Bound-
ary− Layer Meteorol. 56: 339-358.
Olgay V. 1969: Design with the climate: Bioclimatic approach to architectural regional-
ism. Princeton University Press. New Jersey.
ONM 1985 : Atlas climatologique national. Station Ghardaïa. Partie 1: Recueil de don-
nées. Office National de la Météorologie. Alger.
Ouahrani D. 1993: Light and housing in the desert: Case study of Ghardaia, Algeria.
Lighting Research Technol. 25: 1-11.
Paltridge G.W., Platt C.M.R. 1976: Radiative processes in Meteorology and Climatol-
ogy. Elsevier. New York.
Pearlmutter D., Bitan, A., Berliner P. 1999: Microclimatic analysis of “compact” urban
canyons in an arid zone. Atmos. Envir. 33: 4143-4150.
Pereira F.O.R., Minache J.A.C. 1989: Insolation in the built environment: Criteria for its
normalisation and regulation. Proc. 2nd Europ. Conf. on Architecture. Paris.
France: 36-38.
203
References
Pereira F.O.R., Silva C.A.N., Turkienikz B. 2001: A methodology for sunlight urban
planning. A computer-based solar and sky vault obstruction analysis. Solar Energy
70: 217-226.
Pickup J., de Dear R. 1999: An outdoor thermal comfort index (OUT-SET*) –Part 1-
The model and its assumptions. Proc. 15th Int. Congr. Biometeorol. & Int. Conf.
Urban Climatol. Sydney. Australia: 279-283.
Potchter O., Tepper Y. 2002: The climatic behaviour of the internal courtyard in ex-
tremely hot and dry zones. Proc. 19th Int. Conf. on PLEA. Toulouse, France: 469-
473
Potter J., de Dear R. 2000: Field study to calibrate an outdoor thermal comfort index.
Biometeorlogy at the turn of the millennium. Proc. ICB-ICUC’99. Sydney. Aus-
tralia: 315-319.
Ravérau A. 1981: Le M’zab une leçon d’architecture. Sindbad. Paris.
Roche, M. 1970: Le M’zab, architectures ibadites en Algérie. Arthaud. Strasbourg.
Rogers R. 1997: Cities for a small planet. London. Faber and Faber. London.
RUROS 2004: Rediscovering the urban realm and open spaces. Designing open spaces
in the urban environment: a bioclimatic approach. CRES.
http://alpha.cres.gr/ruros/
Saito I., Ishihara O., Katayama T. 1990: Study of the effect of green areas on the ther-
mal environment in an urban area. Energy and Buildings 15-16: 443-446.
Sakakibara, Y. 1996: A numerical study of the effect of urban geometry upon the sur-
face energy budget. Atmos. Envir. 30: 487-496.
Santamouris M., Asimakopolous D. 1996: Passive cooling of Buildings. James &

James. London.
Santamouris, M., Papanikolaou N., Koronakis I., Livada I., Asimakopoulos D. 1999:
Thermal and air flow characteristics in a deep pedestrian canyon under hot
weather conditions. Atmos. Envir. 33: 4503-4521.
Shashua-Bar L., Hoffman M.E. 2000: Vegetation as a climatic component in the design
of an urban street. Energy and Buildings 31: 221-235.
Sievers U., Mayer I., Zdunkowski W.G. 1987: Numerische Simulation des urbanen
204
References
Klimas mit einem zweidimensionalen Modell. Teil 1: Die Modellgleichungen und

deren numerische Behandlung. Meteorol. Rundschau 40: 40-52.
Simpson J.R., McPherson E.G. 1998: Simulation of tree shade impacts on residential
energy use for space conditioning in Sacramento. Atmos. Envir. 32: 69-74.
Spagnolo J., de Dear R. 2003: A field study of thermal comfort in outdoor and semi-
outdoor environments in subtropical Sydney Australia. Building and Environment
38: 721-738.
Staiger H., Bucher G., Jendritzky G. 1997: Gefühlte Temperatur: Die physiologischge-
rechte Bewertung von Wärmebelastung und Kältestress beim Aufenthalt im Freien
mit der Maßzahl Grad Celsius. Annal. Meteorol. 33: 100-107.
Steemers K. 2003: Cities, energy and comfort: a PLEA 2000 review. Energy and Build-
ings 35: 1-2.
Stull R.B. 1988: An introduction into Boundary-Layer Meteorol. Kluwer Academic

Press. Utrecht.
Swaid H. 1992: Intelligent urban forms (IUF): a new climate-concerned urban planning
strategy. Theor. Appl. Climatol. 46: 179-191.
Swaid H., Hoffman M.E. 1990: Prediction of urban air temperature variations using the
analytical CTTC model. Energy and Buildings 14: 313-324.
Swaid H., Bar-El M., Hoffman M. E. 1993: A bioclimatic design methodology for ur-
ban outdoor spaces. Theor. Appl. Climatol. 48: 49-61.
Taha H. 1997: Urban climates and heat islands: albedo, evapotranspiration, and anthro-
pogenic heat. Energy and Buildings 25: 99-103.
Taha M.R., Douglas S., Haney J. 1997: Mesoscale meteorological and air of quality
impacts of increased urban albedo and vegetation. Energy and Buildings 25: 169-
177.
Taesler R., Anderson C. 1984: A method for solar radiation computing using routine
meteorological observations. Energy and Buildings, 7: 341-352.
Teller J., Azar S. 2001: TOWNSCOPE II - A computer system to support solar access
decision-making. Solar Energy 70: 187-200.
Thomas R. 2003: Sustainable urban design, an Environmental approach. Spon. London.
205
References
Todhunter P.E. 1990: Microclimatic Variations Attributable to Urban Canyon Asymme-

try and Orientation. Phys. Geogr. 11: 131-141.
VDI 1998: Methods for the human-biometeorological assessment of climate and air
quality for urban and regional planning. Part 1: Climate. VDI Guideline 3787, Part
2. Verein Deutscher Ingenieure.
Watson I.D., Johnsson G.T. 1988: Estimating person view-factors from fish eye lens
photographs. Int. J. Biometeorol. 32: 123-128.
Wedding J.B., Lombardi D.J., Cermak J.E. 1977: A wind tunnel study of gaseous pol-
lutants in city street canyons. J. Air Pollut. Contr. Ass. 27: 557-566.
Wilson J. D. 1988: A second-order closure model for flow through vegetation. Bound-
ary-Layer Meteorol. 42: 371-392.
Wiren B.G. 1985: Effect of surrounding buildings on wind pressure distribution and
ventilation heat losses for single family houses. Part 1:1 ½ - Storey detached
houses. Int. Swedish inst. for building Research. Gavle. Sweden. Report N° M
85:19.
Wiren B.G. 1987: Effect of surrounding buildings on wind pressure distribution and
ventilation heat losses for single family houses. Part 2:2 ½ - Storey detached
houses. The international Swedish institute for building Research. Gavle. Sweden.
Yamada T. 1982: A numerical study of turbulent airflow in and above a forest canopy.
J. Meteorol. Soc. Japan 60: 439-454.
Yamartino R. J., Wiegand G. 1986: Development and evaluation of simple models for
the flow, turbulence and pollution concentration fields within an urban street can-
yon. Atmos. Envir. 20: 2137-2156.
Yoshida A., Tominaga, K., Watani S. 1990/91: Field measurements on energy balance
of an urban canyon in the summer season. Energy and Buildings 15-16: 417-423
206
List of figure captions page
Fig. 2.1. Daily energy balance of urban facets of an urban canyon oriented N-S with 27
H/W ≈ 1 for a sunny summer day in Vancouver, 49 °N
Fig. 2.2. Isotherm distribution across an E-W canyon at selected daytime hours, also 29
includes wind speed, wind direction and stability conditions at 1 m height
Fig. 2.3. Surface and air temperatures of urban canyon facets, for an E-W street of an 30
aspect ratio H/W = 0.96 under sunny summer conditions for Kyoto, Japan,
35°N
Fig. 2.4. (a) Wind flow regimes and (b) corresponding threshold lines dividing flow 31
into three regimes as function of canyon (H/W) and building (L/W) geometry
Fig. 2.5. Monthly mean canyon irradiances simulated for June for E-W and N-S can- 34
yons and various aspect ratios. The symbols +, x, ∗, , ∆, ο correspond to
H/W = 0.25, 0.5, 1, 2, 3, and 4 respectively
Fig. 2.6. Mean monthly shading fraction SF for canyon, floor and walls in depend- 35
ence with aspect ratio H/W during summer and winter for latitude 33 °N
Fig. 2.7. Three different building blocks orientations showing the effect of the solar 36
envelope on the shape and size of the urban streets geometries
Fig. 2.8. The components of the human heat balance 43
Fig. 2.9. bedZED project showing an E-W asymmetrical street shape for ensuring 53
solar access, together with using galleries and vegetation for outdoor comfort
Fig. 2.10. Solar control through self-shading façade in a hot-dry climate 53
Fig. 2.11. Housing quarter of Linz-Pichling, Austria showing the link between urban 53
and architectural concepts in relation to climate
Fig. 3.1. General scheme of the ENVI-met model including the boundaries 60
Fig. 3.2. Average air temperature Ta and vapour pressure VP humidity in Ghadaia in 74
August (1974-1985, ONM 1985) plotted against ENVI-met simulation results
for a bare soil on the 1st August for Ta, VP, direct irradiance S and global ir-
radiance G.
Fig. 3.3. Geometry of the urban canyons selected for the simulations 78
Fig. 4.1. Diurnal variation of simulated air temperature Ta at 1.2 m within the canyon 79
for E-W oriented streets of aspect ratios H/W of 0.5, 1, 2, 3 and 4
Fig. 4.2a. The simulated direct solar radiation (S) at street level for E-W oriented 81
streets of aspect ratios H/W of 0.5, 1, 2, 3 and 4
207
List of figure and table captions
Fig. 4.2b. The simulated diffuse radiation (D) at street level for E-W oriented streets 81
of various aspect ratios of 0.5, 1, 2, 3 and 4
Fig. 4.2c. The simulated global radiation (G) at street level for E-W oriented streets 82
of various aspect ratios of 0.5, 1, 2, 3 and 4
Fig. 4.3. Example of an isotherm representation chosen for a detailed spatial and 83
temporal illustration of the thermal comfort outdoors
Fig. 4.4. Comparison between air temperature Ta and mean radiant temperature Tmrt 84
in time and space for an E-W oriented street of an aspect ratio H/W = 2 at 1.2
m a.g.l.
Fig. 4.5a & 4.5b. Diurnal variation of PET at street level for an E-W oriented street 86
with an aspect ratio H/W of 0.5 and 1
Fig. 4.5c & 4.5d. Diurnal variation of PET at street level for an E-W oriented street 87
with an aspect ratio H/W of 2 and 3
Fig. 4.5e. Diurnal variation of PET at street level for an E-W oriented street with an 88
aspect ratio H/W of 4
Fig. 4.6. Diurnal variation of simulated air temperature Ta at 1.2 m within N-S ori- 90
ented streets with aspect ratios of 0.5, 1, 2, 3 and 4
Fig. 4.7a. The simulated direct solar radiation (S) at street level for N-S oriented 91
streets with aspect ratios of 0.5, 1, 2, 3 and 4
Fig. 4.7b. The simulated diffuse radiation (D) at street level for NS oriented streets 91
with aspect ratios of 0.5, 1, 2, 3 and 4
Fig. 4.7c. The simulated global radiation (G) at street level for N-S oriented streets of 92
Fig. 4.8a to 4.8b. Diurnal variation of PET at street level for N-S oriented streets 94
with an aspect ratio H/W of 0.5 and 1
Fig. 4.8c to 4.8d. Diurnal variation of PET at street level for N-S oriented streets 95
with an aspect ratio H/W of 2 and 3
Fig. 4.8e. Diurnal variation of PET at street level for N-S oriented streets with an 96
aspect ratio H/W of 4
Fig. 4.9. Differences in (a) air temperature (∆Ta) and (b) global radiation (∆G) be- 97
tween E-W and N-S oriented streets for aspect ratios H/W of 0.5, 1, 2, 3, and
4; positive values mean higher values for E-W cases
Fig. 4.10a. ∆PET between an E-W and N-S oriented street for an aspect ratio of 0.5; 98
208
positive values mean higher PET values for E-W orientation

Figs. 4.10b to 4.10e. ∆PET between an E-W and N-S oriented street for an aspect 99
ratio of 1, 2, 3 and 4 respectively; positive values mean higher PET values for
E-W orientation
Fig. 4.11. Individual short-wave (SW) and long-wave (LW) energy terms absorbed 101
by a standing person at the street centre in an E-W and N-S street with H/W =
2
Fig. 4.12. Comparison of PET patterns according to street orientations E-W, N-S, 104
NE-SW and NW-SE, with an aspect ratio H/W = 2
Fig. 4.13a. Average air temperature Ta at street level (1.2 m a.g.l.) for asymmetrical 106
urban canyons and symmetrical canyons of H/W = 2
Fig. 4.13b. Average air temperature Ta at street level (1.2 m a.g.l.) for asymmetrical 106
urban canyons with overhanging façades and symmetrical canyons of H/W =
2
Fig. 4.13c. Average air temperature Ta at street level (1.2 m a.g.l.) for urban canyons 107
with trees and similar canyons without trees
Fig. 4.13d. Average air temperature Ta at street level (1.2 m a.g.l.) for selected urban 107
canyons for a perpendicular and parallel incidence of wind
Figs. 4.14a and 4.14b. PET distribution across symmetrical urban canyons including 109
galleries on both sides for (a) E-W and (b) N-S oriented streets (H/W = 2)
Fig. 4.14c and 4.14d. PET distribution across symmetrical urban canyons including 110
galleries on both sides for (c) NE-SW and (d) NW-SE oriented streets (H/W
= 2)
Figs. 4.15. Individual short-wave (SW) and long-wave (LW) energy terms absorbed 112
by a standing person for a N-S oriented street with H/W = 2 for points within
a gallery and at the street centre
Fig. 4.16a. PET distribution across an asymmetrical profile with H2/W = 2 and H1/W 114
= 1 (case II-2) oriented E-W and including galleries
Fig. 4.16b. ∆PET between asymmetrical canyon (H2/W = 2, H1/W = 1) and symmet- 114
rical canyons H/W = 2 (left) and H/W = 1 (right) for E-W orientation
Fig. 4.17a. PET patterns across an asymmetrical profile with H2/W = 2 and H1/W = 116
1 oriented N-S and including galleries
Fig. 4.17b. ∆PET between asymmetrical canyon (H2/W = 2, H1/W = 1) and symmet- 116
209
rical canyons H/W = 2 (left) and H/W = 1 (right) for N-S orientation
Fig. 4.18. PET patterns across an asymmetrical profile with H2/W = 2 and H1/W = 1 117
(case II-2) oriented NE-SW and including galleries
Fig. 4.19. PET patterns across an asymmetrical profile with H2/W = 2 and H1/W = 1 117
(case II-2) oriented NW-SE and including galleries
Figs. 4.20a and 4.20b. PET patterns across an asymmetrical profile with overhanging 119
façades (H2/W = 2 and H1/W = 1.5) oriented E-W and N-S, respectively
Figs. 4.20c and 4.20d. PET patterns across an asymmetrical profile with overhanging 120
façades (H2/W = 2 and H1/W = 1.5) oriented NE-SW and NW-SE, respec-
tively
Fig. 4.21. PET patterns within a street oriented E-W with H/W = 2 and a row of 122
trees on the south-facing side ( ……. projection of trees’ area)
Fig. 4.22. Differences in (a) direct solar radiations and (b) long-wave radiations emit- 122
ted by the ground surface between a street with vs. without a row of trees
Fig. 4.23. PET pattern within a street oriented N-S with H/W = 1 and a large central 123
row of trees (− − − limit of gallery, ------ projection of trees’ area)
Fig. 4.24. Individual short-wave (SW) and long-wave (LW) energy terms absorbed 124
by a standing person located in a N-S street with H/W = 1 without vs. with
trees
Fig. 4.25. Mean wind velocity within urban canyons of (a) H/W= 2 and (b) H/W = 126
0.5, at 1.2 m a.g.l. level for both perpendicular and parallel wind incidence on
street axis
Fig. 4.26. Zones with different ventilation potential and depending on canyon dimen- 128
sions according to simulation results
Fig. 4.27. (a) PET pattern for an E-W street of H/W = 2 for a parallel wind incidence, 129
(b) ∆PET between parallel and perpendicular wind for the same canyon
Fig. 4.28. (a) PET pattern for an E-W street of H/W = 2 including a row of trees 129
(dense, leafless base) for a parallel wind incidence, (b) ∆PET between paral-
lel and perpendicular wind for the same canyon, negative values mean lower
PET
Fig. 4.29. Dependence of the solar access index SAI on the aspect ratio H/W at street 132
level for (a) summer conditions and (b) winter conditions
Fig. 4.30. Dependence of the SAI on the aspect ratio across the street space for E-W 133
210
and N-S oriented streets in winter

Fig. 4.31. Dependence of solar access index on aspect ratio for (a) E-W and (b) N-S 134
oriented street
Fig. 5.1. Plan view of the east-west canyon street in Freiburg with the location of the 137
permanent station and the measuring points MP1 to MP4
Fig. 5.2. Fish-eye photography of the canyon at the station location, Freiburg 138
Fig. 5.3. Set of radiation sensors for the measurement of the global radiation from 139
the 3D surroundings within the urban canyon in Freiburg
Fig. 5.4. Daily variation of air temperature Ta in the urban canyon on a cloudless 140
sunny day in Freiburg
Figs. 5.5a and 5.5b. Daily variation of (a) air temperature Ta, ground temperature Ts 142
and wall temperature Tw at the station on the north side of the street and (b) Ts
and Tw at points MP1, MP2 (southern side) as well as MP3 and MP4 (north-
ern side) on 15 July 2003
Fig. 5.6. Wind direction within the street canyon and above roof level (at 61 m a.g.l.) 144
Fig. 5.7. Wind speed within the canyon street and above roof level (at H = 61 m) 144
Fig. 5.8. Wind speed and wind direction dd outside the canyon plotted against inside 144
corresponding conditions
Fig. 5.9. Temporal and spatial distribution of short-wave radiation in Wm-2 (normal 146
to surface) across the street, simulated by ENVI-met 3.0 (---- location of the
measuring station)
Fig. 5.10. Short-wave radiation fluxes (K) received from the 6 directions surrounding 148
a standing person located at the south-facing side of an E-W oriented street,
H/W = 1
Fig. 5.11. Long-wave radiation fluxes (L) received from the 6 directions surrounding 148
a standing person located at the south-facing side of an E-W oriented street
with H/W = 1
Fig. 5.12. Actual short-wave radiation (Kabs) absorbed by a standing person at the 151
south facing side of an E-W oriented with an aspect ratio H/W = 1
Fig. 5.13. Actual long-wave radiation (Labs) absorbed by a standing person at the 151
south facing side of an E-W oriented with an aspect ratio H/W = 1
Fig. 5.14. Daily evolution of the mean radiant temperature Tmrt and the physiologi- 152
cally equivalent temperature PET at the south facing side of an E-W oriented
211
with an aspect ratio H/W = 1

Fig. 5.15. Long-wave radiation (L) absorbed by a standing person versus the radiant 153
heat emitted by the ground and nearby north wall
Fig. 5.16. The mean radiant temperature Tmrt simulated by ENVI-met plotted against 154
measured Tmrt
Fig. 5.17. (a) Simulated individual short-wave (SW) and long-wave (LW) energy 155
terms absorbed by a standing person and (b) the simulated long-wave irradi-
ance plotted against measured data in Freiburg
Fig. 6.1. The old city of Beni-Isguen and its oasis in the Mzab valley, Algeria 157
Fig. 6.2. A bird view on the a typical compact urban fabric of Beni-Isguen in the 159
Mzab valley, Algeria
Fig. 6.3. Route and all measuring points within different street geometries in the ver- 160
nacular city of Beni-Isguen, Mzab valley, Algeria
Fig. 6.4. Photographs and fish-eye photographs of selected measuring sites within the 161
city of Beni-Isguen, Mzab valley, Algeria
Figs. 6.5 to 6.7. Air temperature Ta, vapour pressure VP and wind speed v, at all 163
measuring points during a typical summer day (26 June 2003) within the city
of Beni-Isguen, Algeria
Figs. 6.8 and 6.9. Surface temperature Ts and wall temperature Tw at all measuring 166
points during a typical summer day (26 June 2003) within the city of Beni-
Isguen, Algeria
Figs. 6.10 and 6.11. Short-wave (Kabs) and long-wave (Labs) radiation fluxes ab- 167
sorbed by a standing person at all measuring points during a typical summer
day (26 June 2003) within the city of Beni-Isguen, Algeria
Figs. 6.12 and 6.13. Mean radiant temperature Tmrt and the difference (Tmrt-Ta) at all 169
measuring points during a typical summer day (26 June 2003) within the city
of Beni-Isguen, Algeria
Fig. 6.14. Physiologically equivalent temperature PET at all measuring points during 170
a typical summer day (26 June 2003) within the city of Beni-Isguen, Algeria
Fig. 7.1. Scheme on the subdivision of a street canyon volume according to climatic 185
design needs
Fig. 7.2. Example of wide canyon combining motor traffic and pedestrian areas pro- 187
tected by deciduous trees
212
Fig. 7.3. Example of wide street canyon oriented E-W, combining comfortable pe- 187
destrian zones and motor traffic
Fig. 7.4. example of an asymmetrical canyon combining summer comfort at street 188
level, winter solar access and high density
List of table captions page
Table 2.1: Selected thermal comfort indices for indoors and outdoors 42
Table 3.1. Example of a typical inputs’ configuration of a simulation; Data as 75
used in this study
Table 3.2. Dimensional characteristics of the investigated urban canyons 77
Table 4.1. Tmrt, Ta, v and VP corresponding to PET maxima for E-W versus N-S 100
streets for H/W varying from 0.5 to 4
Table 4.2a. Individual radiant energy terms (Wm-2) absorbed by a standing person 101
at the most stressful hours for E-W versus N-S canyon of H/W = 2 at
street centre (SVF = 0.569)
Table 4.2b. Individual radiant energy terms (Wm-2) absorbed by a standing person 101
at the most stressful hours for E-W versus N-S canyon of H/W = 4 at
street centre (SVF = 0.375)
Table 4.3. Individual short-wave (SW) and long-wave (LW) energy terms (Wm-2) 112
absorbed by a standing person at the most stressful hours in a gallery
and at street centre of for a N-S canyon of H/W = 2
Table 4.4. Individual short-wave SW and long-wave LW energy terms (Wm-2) 124
absorbed by a standing person at the most stressful hours in a N-S can-
yon of H/W = 1 with and without trees for grid No 6
Table 5.1. Instrumentation used at the station within the street canyon in Freiburg 139
Table 5.2: Percentage of short-wave radiation (SW) and Long-wave radiation 153
(LW) absorbed by a standing person at the south facing side of an E-W
oriented with an aspect ratio H/W = 1 in Freiburg in summer
Table 6.1. Geometry and material properties at the eight measuring points in the 159
old city of Beni-Isguen, Mzab valley, Algeria
Table 6.2. Mean wind velocity (m s–1) measured at all measuring sites on (a) 23 165
June and (b) 26 June 2003 in Beni-Isguen, Algeria (32.40°N, 3.80°E)
213
List of symbols
List of symbols
Symbol Quantity Unit

af albedo of leaf surface -
as albedo of ground surface -
aw albedo of wall surface -
a averaged albedo for walls and ground surfaces in the model area -
-
cµ , σE, σε constants for the turbulence model: cµ = 0.09, σe = 1 ,σε =1.3
c1, c2, c3 standard values for calibrating ε-equation, available from lit. -
cb specific heat WsK-1kg-1
cd,f drag coefficient at the plant foliage (= 0.2) -
cp specific heat of air at constant pressure (=1847) Jkg-1K-1
D diffuse solar radiation Wm-2
Di diffuse and diffusely reflected short-wave radiation in direction i Wm-2
Dt total short-wave diffuse radiation flux absorbed by a human body Wm-2
E turbulent kinetic energy m2s-2
Ei long-wave radiation component in direction i Wm-2
Et total long-wave radiation flux absorbed by a human body Wm-2
F extinction coefficient -
F orientation of leafs toward the sun (= 0.5 for random orientation) -
Fi angle weighting factor in direction i -
FCS heat flow from body core to skin surface Wm-2
FSC heat flow from skin surface to clothing surface Wm-2
Feff effective area of body for energy exchange with envir. (≈ 0.75) -
f coriolis parameter (=10-4 ) s-1
fp surface projection factor -
fw fraction of wet leaves -
g acceleration due to gravity ( = 9.81) ms-2
G(0) substrate heat (0: ground) Wm-2
G global radiation Wm-2
H Height of the building m
H(o,w,r) turbulent sensible heat (for ground o, wall w or roof r surface) Wm-2
I direct solar radiation impinging normal to the surface Wm-2
i, j, k cartesian coordinates in grid points -
I0 solar constant Wm-2
It total short-wave direct irradiance absorbed by a human body Wm-2
214
List of symbols
Jf, evap evaporative heat flux between plant and surroundings Wm-2
Jf,trans transpiration heat flux between plant and surroundings Wm-2
Jf,h direct heat flux between plant and surroundings Wm-2
K* Total short-wave radiation Wm-2
KE, Kε diffusion coefficients for local turbulence (production & dissipation) m2s-1
K m(0 ,w ) exchange coefficient for momentum at ground 0 or wall surface w m2s-1
K h(0 ,w ) exchange coefficient for heat at ground 0 or wall surface w m2s-1
K q(0 ,w ) exchange coefficient for vapour at ground 0 or wall surface w m2s-1
Kabs short-wave radiation flux absorbed by a human body Wm-2

Labs long-wave-wave radiation flux absorbed by a human body Wm-2
L length of the building m
L latent heat of vaporization Jkg-1
L* Total long-wave radiation Wm-2
l water content in a layer of air g
LAD leaf area density m m-3
2
LAI leaf area index m3m-3

LAI* “3D”leaf area index including angle of incidence of sun rays m3m-3
lcl heat resistance of the clothing Km2W-1
LE(0) turbulent latent heat density (0: ground) Wm-2
M metabolic rate W
m optical mass -
p´ local pressure perturbation Pa
PET physiologically equivalent temperature °C
PMV predicted mean vote -
Pr production of turbulence energy due to wind shearing -
q specific humidity (q0 at the surface) kgkg-1
∆q leaf-to-air humidity deficit kgkg-1
Qε additional turbulence dissipated by vegetation s-1
Q(w,r) heat flux through wall or roof Wm-2
Q* net all-wave radiation Wm-2
Q* net radiation budget of the body W
QE latent heat flux density Wm-2
QE additional turbulence produced by vegetation s-1
QH sensible heat flux density Wm-2
QH convective heat flux from body W
215
List of symbols
Qh sink/source terms due to heat ms-2

QL latent heat flow for diffusion of water from the body W
Qq sink/source terms due to vapour ms-2
Qr heat flux through roof Wm-2
QRE respiratory heat flux W
QSW latent heat flow due to evaporation of sweat W
Qw heat flux through wall Wm-2
ra aerodynamic resistance of the leaf sm-1
Rlw( g,net
,w ,r ) net long-wave radiation (g,w,r: for ground, wall or roof surface) Wm-2
R*sw short-wave radiation flux at the model boundary Wm-2

(0 ) direct short-wave radiation flux density(0: at model boundary) Wm-2
Rsw ,dir
(0 ) diffuse short-wave radiation flux density(0: at model boundary) Wm-2

Rsw , dif
Rsw,abs absorbed short-wave radiation flux by water in the atmosphere Wm-2
Rlw↓,(0 ) atmospheric long-wave radiation flux density (0: at model boundary) Wm-2
Rlw↑ long-wave radiation flux density from ground Wm-2
Rlw↔ long-wave radiation flux density from walls Wm-2
Rn ,lw divergence of long-wave radiation flux density Wm-2
Rsw, net net short-wave radiation density Wm-2

RAD root area density m2m-3
RAI root area index m3m-3
RH relative humidity %
Ri Richardson number 1
Rib bulk-Richardson number 1
rs stomatal resistance at leaf surface sm-1
S Direct solar radiation Wm-2
S storage heat flow in body W
Srad mean radiation flux density absorbed by a human body Wm-2
Su , Sv, Sz sink/source terms due to wind drag ms-2
SAI Solar access index -
SVF sky view factor -
t time s
T absolute temperature K
Ta air temperature K or °C
Ta,i air temperature inside the buildings K
216
List of symbols
Tc core temperature °C
Tcl clothing temperature °C
Tf(+,-) leaf temperature (+ overlying side, - underlying side of the leaf) K
Tmrt mean radiant temperature K or °C
To ground surface temperature K or °C
Tr roof temperature K
Ts ground surface temperature K or °C
Tsk skin temperature °C
Tw wall temperature K
Th dissipation of turbulence energy due thermal stratification -
trf transmission factor (= 0.3) -
u horizontal wind speed m s-1
U heat transmittance Wm-2K-1
u* friction velocity ms-1
u, v, w wind speed in the x, y, and z directions (ug, vg, wg, geostr. wind) ms-1
ui, xi i.e. u, v, w and to x, y, z with i = 1,2,3 (einstein summation) -
v horizontal wind speed m s-1
Vb blood flow from body core to skin ls-1m-2
VP vapour pressure hPa
W Width of the street m
W physical work output (activity) W
2 2 2 0.5
W mean wind speed at height z (w = u + v + w ) ms-1
Wi angle weighting factor in direction i -
x, y, z cartesian coordinates m
X, Y, Z horizontal and vertical dimensions of the core model m
∆x, ∆y, ∆z grid resolution of the model in the 3 directions m
zo roughness length m
zp vegetation height m
zr root depth m
∆Qs energy stored in buildings Wm-2
αk body absorption coefficient for short-wave radiation (≈ 0.7) -
αl body absorption coefficient for long-wave radiation (≈ 0.97) -
αR Mies scattering (αm = λ-1.3βtr) -
αR, Rayleigh scattering (αr = 0.00816 λ-4 ) -
β* angle of incident direct beam / normal to surface °
θ potential temperature K
217
List of symbols
θref average temperature over all grids at height z K

σB stefan-boltzmann constant (=5.664 10-8) Wm-2K-4
σ lw↓ modification factor for downwards long-wave radiation -
σ lw↑ modification factor for upwards long-wave radiation -
σsvf sky view factor -

σsw,dif modification factor for diffuse short-wave radiation -
σsw,dir modification factor for direct short-wave radiation -
ε dissipation of turbulence -
εf emissivity of foliage -
εn emissivity of layer n -
εp emissivity of the human body (≈ 0.97) -
εs, εw emissivity of ground and wall surface -
ρ air density (ρ0 =1.29) kgm-3
µ thermal admittance Jm-2s-0.5K0.5
ρb blood density kgl-1
δc factor depending on evaporation and transpiration probability -
λs heat conductivity Wm-1K-1
λ wavelength range -
φ sun position °
ω vertical angle of an obstacle °
π azimuth angle °
218
Appendix
Appendix A1: Database of soil types in ENVI-met 3.0. These individual soils are com-
bined to form a “soil profile” as shown in Appendix A2.
ID ηs ηfc ηwilt ψpot Kη,s ρi c i B λ Name
Natural soils
0 0.451 0.24 0.155 -0.478 7 1.212 5.39 0 Default soil (loam)
sd 0.395 0.135 0.0068 -0.121 176 1.463 4.05 0 Sand
ls 0.41 0.15 0.075 -0.09 156.3 1.404 4.38 0 Loamy sand
sl 0.435 0.195 0.114 -0.218 34.1 1.32 4.9 0 Sandy loam
sl 0.485 0.255 0.179 -0.786 7.2 1.271 5.3 0 Silt loam
le 0.451 0.24 0.155 -0.478 7 1.212 5.39 0 Loam
ts 0.42 0.255 0.175 -0.299 6.3 1.175 7.12 0 Sandy clay loam
tl 0.477 0.322 0.218 -0.356 1.7 1.317 7.75 0 Silty clay loam
lt 0.476 0.325 0.25 -0.63 2.5 1.225 8.52 0 Clay loam
st 0.426 0.31 0.219 -0.153 2.2 1.175 10.4 0 Sandy clay
ts 0.492 0.37 0.283 -0.49 1 1.15 10.4 0 Silty clay
to 0.482 0.367 0.286 -0.405 1.3 1.089 11.4 0 Clay
tf 0.863 0.5 0.395 -0.356 8 0.836 7.75 0 Peat
Artificial soils
zb 0 0 0 0 0 2.083 0 1.63 Cement concrete
mb 0 0 0 0 0 1.75 0 2.33 Mineral concrete
ak 0 0 0 0 0 2.214 0 1.16 Asphalt (with gravel)
ab 0 0 0 0 0 2.251 0 0.9 Asphalt (with basaltl)
as 0 0 0 0 0 1.94 0 0.75 Asphalt (Oke 1987)
gr 0 0 0 0 0 2.345 0 4.61 Granite
ba 0 0 0 0 0 2.386 0 1.73 Basalt
ww 0 0 0 0 0 0 0 0 Water
Key of symbols of Appendix A1
ηs saturation water content m3m-3

ηfc field capacity m3m-3
ηwilt permanent wilting point m3m-3
ψpot Matrix potential at saturated water content m
Kη,s hydraulic conductivity at saturated water content 10-6.ms-1
ρici constant (clapp and horrnberger 1978) 106.Jm-3K-1
λ soil conductivity Wm-1K-1
219
Appendix
Appendix A2. Database of multilayered soils profiles in ENVI-met. Individual soil

types are given in Appendix A1.
Appendix B: Database of various vegetation types in ENVI-met
220
Appendix
Appendix C. Sky view factors for each grid point and all investigated aspect ratios
SVF/H/W H/W = 0.5 H/W = 1 H/W = 2 H/W = 3 H/W = 4 ASYM. OVERH.
gal.1 0.227 0.117 0.117 0.117
gal.2 0.227 0.117 0.117 0.117
gal.3 0.430 0.284 0.343 0.117
pt. 1 0.774 0.629 0.475 0.393 0.342 0.552 0.117
pt. 2 0.852 0.702 0.528 0.432 0.371 0.608 0.249
pt. 3 0.899 0.756 0.569 0.461 0.391 0.654 0.416
pt. 4 0.907 0.764 0.576 0.466 0.395 0.665 0.452
pt. 5 0.901 0.760 0.574 0.465 0.394 0.667 0.444
pt. 6 0.887 0.748 0.566 0.460 0.390 0.659 0.398
pt. 7 0.824 0.686 0.522 0.429 0.370 0.604 0.242
pt. 8 0.736 0.611 0.469 0.391 0.341 0.532 0.155
gal. 4 0.413 0.279 0.351 0.155
gal. 5 0.227 0.117 0.227 0.155
gal. 6 0.227 0.117 0.224 0.155
SVF street 0.848 0.707 0.535 0.437 0.374 0.618 0.309
SVF with gal. 0.848 0.541 0.390 0.437 0.374 0.462 0.240
221
Berichte des Meteorologischen Institutes der Universität Freiburg
Nr. 1: Fritsch, J.: Energiebilanz und Verdunstung eines bewaldeten Hanges.

Juni 1998.
Nr.2: Gwehenberger, J.: Schadenpotential über den Ausbreitungspfad Atmo-

sphäre bei Unfällen mit Tankfahrzeugen zum Transport von Benzin,
Diesel, Heizöl oder Flüssiggas. August 1998.
Nr. 3: Thiel, S.: Einfluß von Bewölkung auf die UV-Strahlung an der Erdober-
fläche und ihre ökologische Bedeutung. August 1999.
Nr. 4: Iziomon, M.G.: Characteristic variability, vertical profile and modelling of

surface radiation budget in the southern Upper Rhine valley region. Juli
2000.
Nr. 5: Mayer, H. (Hrsg.): Festschrift „Prof. Dr. Albrecht Kessler zum 70. Ge-
burtstag“. Oktober 2000.
Nr. 6: Matzarakis, A.: Die thermische Komponente des Stadtklimas. Juli 2001.
Nr. 7: Kirchgäßner, A.: Phänoklimatologie von Buchenwäldern im Südwesten

der Schwäbischen Alb. Dezember 2001
Nr. 8: Haggagy, M.E.-N.A.: A sodar-based investigation of the atmospheric

boundary layer. September 2003
Nr. 9: Rost, J.: Vergleichende Analyse der Energiebilanz zweier Untersu-

chungsflächen der Landnutzungen “Grasland“ und „Wald“ in der südli-
chen Oberrheinebene. Januar 2004
Nr. 10: Peck, A.K.: Hydrometeorologische und mikroklimatische Kennzeichen

von Buchenwäldern. Juni 2004
Nr. 11: Schindler, D.: Characteristics of the atmospheric boundary layer over a
Scots pine forest. Juni 2004
Nr. 12: Matzarakis, A., de Freitas, C.R., Scott, D. (eds.): Advances in Tourism
Climatology. November 2004
223
Nr. 13: Dostal, P.: Klimarekonstruktion der Regio TriRhena mit Hilfe von direk-
ten und indirekten Daten vor der Instrumentenbeobachtung. Dezember
2004
Nr. 14: Imbery, F.: Langjährige Variabilität der aerodynamischen Oberflächen-

rauhigkeit und Energieflüsse eines Kiefernwaldes in der südlichen O-
berrheinebene (Hartheim). Januar 2005
Nr. 15: Ali Toudert, F.: Dependence of outdoor thermal comfort on street de-
sign in hot and dry climate. November 2005
224

Diss Freidok Ali Toudert Fazia

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Diss Freidok Ali Toudert Fazia

Hochgeladen von

Copyright:

Verfügbare Formate

Berichte des Meteorologischen Institutes

der Universität Freiburg

Fazia Ali Toudert

Dependence of Outdoor Thermal Comfort

Freiburg, November 2005

Eigenverlag des Meteorologischen Instituts der Albert-Ludwigs-Universität Frei-

Druck: Druckerei der Albert-Ludwigs-Universität Freiburg

Herausgeber: Prof. Dr. Helmut Mayer und PD Dr. Andreas Matzarakis

Dissertation, angenommen von der Fakultät für Forst- und Umweltwissenschaf-

Freiburg, July 2005

2. Literature review .……………………….……………………………………... 25

3. The numerical model ENVI-met 3.0 ……………………….…………………. 57

3.4. The atmospheric model ….………………………..………………………. 61

4. Results of the numerical simulations ………………………………………… 79

5. Field measurements in Freiburg, Germany …….........………………………. 137

6. Field measurements in Beni-Isguen, Algeria ………….……………......….… 157

7. Discussion and conclusion …………………………………………..………… 175

7.5. Limits and current development of ENVI-met …...………………………... 189

References …..……………………………………………………………………… 194

Abhängigkeit des thermischen Komforts unter heißen und trockenen Klimabedin-

1.1. Necessity of the present study

1.2. Objectives of the present study

The following design strategies are analysed:

 Ghardaia (Beni-Isguen), Algeria, subtropical, in hot-dry climate.

1.4. Structure of the thesis

2.1. The microclimate of an urban street canyon

2.1.1. Energy budget of an urban canyon

2.1.2. Thermal characteristics of an urban canyon

2.1.3. Wind flow in an urban canyon

2.1.4. Solar access outdoors

2.1.5. Solar access indoors

2.1.6. Effects of the vegetation

2.1.7. Further aspects

2.2. Outdoor thermal comfort

2.2.1. The thermal comfort indices

2.2.2. The human energy balance

FSC = (1 / lcl ) ⋅ (Tsk − Tcl ) (2.5)

2.2.3. The mean radiant temperature

2.2.4. Methodological problems in assessing comfort outdoors

The assessment of comfort faces a number of methodological problems, including Tmrt

2.2.5. Effects of urban design on comfort outdoors

3.1. Numerical modelling of the urban microclimate

3.2. Relevance of ENVI-met to the present study

3.3. General structure of ENVI-met 3.0

at H=2500m, u , v , θ q inputs, constant

Nesting grids 1D soil model

3.4. The atmospheric model

3.4.1. Mean air flow

3.4.2. Temperature and humidity

3.4.3. Atmospheric turbulence

Qε = 1.5cd , f LAD ( z ) ⋅ W 3 − 6cd , f LAD ( z ) ⋅ W ⋅ ε (3.11)

3.4.4. Radiation fluxes

The atmospheric long-wave radiation depends on air temperature, as well as on absorp-

,dir = R sw − R sw ,abs = R sw − (70 + 2.8VP2 m ⋅ m )

Taesler and Anderson (1984).

where ω is the vertical angle determined by an obstacle at the azimuth angle π.

are expressed as follows:

σ sw,dif ( z ) = exp(− F ⋅ LAI (z, z p )) (3.18b)

σ lw↓ ( z ) = exp(− F ⋅ LAI (z, z p )) (3.18c)

σ lw↑ ( z ) = exp(− F ⋅ LAI (0, z )) (3.18d)

Rsw ,dir ( z ) = σ sw ,dir ( z )Rsw

Rlw↑ ( z ) downwards are then expressed by:

3.4.5. The ground and buildings surfaces

temperature of all walls Tw :

Ghardaia (Beni-Isguen), Algeria, subtropical, in hot-dry climate.