Dependence of Outdoor Thermal Comfort on Street

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

ISSN 1435-618X Alle Rechte, insbesondere die Rechte der Vervielfältigung und Verbreitung sowie der Übersetzung vorbehalten. Eigenverlag des Meteorologischen Instituts der Albert-Ludwigs-Universität Freiburg 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: [email protected]

Dokumentation:

Ber. Meteor. Inst. Univ. Freiburg Nr. 15, 2005, 224 S.

Dissertation, angenommen von der Fakultät für Forst- und Umweltwissenschaften 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 Institute MIF, University of Freiburg, for supervising this work, for his support throughout the period of my stay in MIF and for giving me the valuable possibility to join several 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 participate in the experimental work conducted in south Algeria is gratefully acknowledged. 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 manuscript. 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

Page

TABLE OF CONTENTS 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 architectural 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 sensation across the street, i.e. street centre vs. street sides, which influence the frequentation of the street. A special emphasis is placed on a human bio-meteorological assessment of these microclimates by using the thermal index PET, Physiologically Equivalent Temperature. The investigation is carried out by using the three-dimensional numerical model ENVImet 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 orientation 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 pedestrian 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 conditions. 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 geometry 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 subjective 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 Klimabedingungen vom Straßendesign

Die vorliegende Untersuchung beschäftigt sich mit dem Beitrag, den das Straßendesign zur Ausbildung eines komfortablen Mikroklimas für Menschen im Straßenbereich leisten kann. Der Schwerpunkt liegt bei subtropischen Klimabedingungen, d.h. bei heißem und trockenem Klima. Die Arbeit wurde anwendungsorientiert durchgeführt. Sie versucht, quantitative Ergebnisse bereitzustellen, die aus der Sicht eines Stadtplaners leicht verstanden und interpretiert werden können. Verschiedene Geometrien von Straßenschluchten werden untersucht. Sie schließen ein variables H/W-Verhältnis (H: Höhe der Randbebauung, W: Straßenbreite), unterschiedliche Richtungen und spezifische Designausführungen (z.B. Überhänge oder Straßenbegleitgrün in Form von Bäumen) ein. Am Anfang werden symmetrische Straßenschluchten mit H/W = 0.5, 1, 2 und 4 analysiert, die in E-W, N-S, NE-SW und NW-SE verlaufen. Daran schließt sich die Untersuchung von asymmetrischen Straßenschluchten an, die eine größere bzw. kleinere Öffnung zum Himmel aufweisen. Hier werden zusätzlich die Effekte von planerischen Details, wie Galerien, Überhänge und Straßenbegleitgrün, berücksichtigt, da sie eine weitere Möglichkeit zur Verbesserung der thermischen Komfortbedingungen im Außenbereich darstellen. Bei den Ergebnissen erfolgt eine räumliche Differenzierung im thermischen Empfinden von Menschen quer zur Straße. Neben der Straßenmitte werden daher auch die Straßenränder betrachtet, da sie vorwiegend von Menschen in der Stadt frequentiert 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 repräsentiert, dient die Region Ghardaia (32.40 °N, 3.80 °E) in Algerien. Die meisten Simulationsberechnungen 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 kurzund langwelligen Strahlungsflüsse vor dem Hintergrund der Wärme, die dadurch eine stehende Person absorbiert, im Mittelpunkt. Zusätzlich ließen sich über die durchgeführten 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 Straßenschluchten. 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 Straßenschluchten möglich. Gegensätzliche Muster der Komfortbedingungen treten zwischen Straßenschluchten mit sehr hohem und niedrigem H/W-Verhältnis sowie mit unterschiedlicher Orientierung auf. Breite Straßenschluchten (H/W ≤ 0.5) weisen im Sommer 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 subtropischen Klima der Tageszeitraum mit extremer Wärmebelastung und die räumliche Verteilung 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 häufigen 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 parametrisiert werden. Stehende Menschen in Straßenschluchten absorbieren tagsüber hauptsächlich Strahlungswärme von bestrahlten Umgebungsflächen, während der Wärmegewinn aus der direkten Sonnenstrahlung von zweitrangiger Bedeutung ist. Daraus ergibt sich, dass nur über die Abschattung der direkten Sonnenstrahlung auf Umgebungsflächen die klimatischen Bedingungen in Straßenschluchten einen Zustand erreichen, 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 aufgefasst werden, da ihre räumliche und zeitliche Variabilität in Straßenschluchten deutlich 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 Straßenschluchten 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 Befragungen 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 deficit (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 progressively 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 exchanges 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 environments, 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 meetings (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 observed 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 canopy. 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 conflict 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 compactness and openness to the sky, respectively. Oke (1988) argues that a “zone of compatibility” which ensures a compromise between apparently conflicting objectives in the design of the street can be found. Swaid (1992), for instance, proposes as “intelligent buildings” some removable arrangements within the street in order to control shading according to seasons, which attests the conflicting task. Moreover, traditional and contemporary ar18

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 methods, 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 microclimatic changes within an urban street canyon has already been gathered, mainly from studies conducted in mid-latitude cities. All studies point out the prime importance of the aspect 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 irradiation 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 temperatures and consequently the amount of heat transferred to air as sensible flux and consecutively the air temperature (Nakamura and Oke 1988, Yoshida et al. 1990/91, Santamouris 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 focusing on the role of urban geometry. The number of methodological questions on the assessment 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 temperature 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 exposed locations and even up to 5 K in shaded parts due to the diffuse and reflect solar radiation components. However, although methodologically more accurate, these studies focused mainly on land use differences by considering various urban densities and vegetated 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 advantage of shading in the reduction of the radiant heat absorbed by a human body compared 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 methods, 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 between 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 (Nikolopoulou et al. 2001). Thus, an investigation which also considers the spatial microclimatic differences 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 sensation 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 microclimate 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 spatial 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 relation to street design is certainly the huge number of urban variables and processes involved. 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 factors (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. Arnfield 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 growing 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 arrangements could be analyzed and compared. Urban microclimate models vary substantially in many aspects: their physical basis, temporal 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 meteorological 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 simulation 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 thermodynamics). 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 parameters and calculates all required meteorological factors, namely air and surface temperatures, 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 discussed 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 temperature (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 creation 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 modelling 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 comparison 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 govern 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, Germany 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 international 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 surface 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 assumes that the interdependence between the urban geometry and the climate is well understood and translated in readily understandable guidelines for the designer, so that the climatic 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. Literature review

2.1. The microclimate of an urban street canyon Obviously, a representative urban canyon is quite impossible to find if all modifying parameters 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 vegetation 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 before 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 exchange 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 energy 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 canyon 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

27


negligible in comparison with the sunlit walls. Thus, it was suggested that turbulent heat 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 (represented 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 become narrower they become increasingly isolated in term of heat exchange from the overlaying 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 allwave 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 measuring 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’ frequency towards the floor and walls (Fig. 2.2). The street had its axis oriented E-W, an aspect 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 includes 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 temperature 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 temperature Ts and the adjacent Ta for directly irradiated urban facets. Near midday, the difference (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 aspect 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 between 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 increased with height within the canyon. This difference became insignificant in the nighttime (< 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 regimes 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 skimming 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, Santamouris 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 proportionally 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 circulation but thermal effects due to canyon facets irradiation may play a role (e.g. Nakamura 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 onevortex 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 perpendicular 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 contribute 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 promotes 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, Santamouris 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 orientations 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 exposure 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 oriented 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 obstructions. 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 orientation, while the shading patterns observed for N-S extend up to 30° from due north towards 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 sunshine 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 symmetrical 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 determining 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 envelope 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 density 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. Capeluto 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 exposed 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 emphasized (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 particular 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 advantages of urban vegetation are assessed either according to their effects on the meteorological factors (e.g. Ta, RH or v) or to the induced energy savings in the buildings as a result of less cooling and/or heating loads. Primarily, the vegetation possesses three main properties which affect the climate: shading, humidification (evapotranspiration) and windbreak (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 seasonal changes and growth of the vegetation (density, size, etc.). The main results of relevance for the present work are summarized below. For a single tree, the shading effect is easily estimated while the cooling by evapotranspiration 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 increase 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 insignificant for small vegetated areas (Shashua and Hoffman 2000). Hence, it has been suggested 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 intersections or in rows along the streets. The usefulness of the latter solution should not be underestimated 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 variability 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 vegetation can influence the microclimate negatively if not appropriately designed, e.g. by blocking 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 variable 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 being 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 temperature 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 nighttime 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 expresses satisfaction with the thermal environment. A more rational definition relates comfort 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 sensation of cold or heat (e.g. Houghton and Yaglou 1923, Missenard 1948). The thermal environment 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 responses 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, ASHRAE 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 (ASHRAE 2001a). These are often limited to the estimation of the combined effect of air temperature, 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 equilibIndex 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, clothing, 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 energy 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 expressions 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 diffusion 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 storage 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 temperatures 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 outdoors. The differences between the various existing models are attributable to the complementary parameterizations related to personal data required to solve eq. 2.3. The comfort equation proposed by Fanger (1970) is probably the most well-known application 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 temperature (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 people (≈ 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 insulation Icl. Solving that equation provides the Predicted Mean Vote PMV defined as the corresponding 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 obtained 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 radiation 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. Similarly 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 temperature (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 surface 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 limitations:

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 Chapters 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 assessment 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 accurate 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 extreme 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, Jendritzky 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 exchange 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):

Tmrt

  n α = 1 / σ B  ∑ Ei Fi + k  i =1 εp  

 αk  D F f I + ∑ i i ε p p  i =1 n

0.25

(2.6)

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 complex 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 radiation 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 (usually 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 outdoors for several reasons, including the non-homogeneity of the radiant environment induced 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 absorbed 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 human 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 convenient because it needs a relatively long time to reach equilibrium (15-20 minutes). Alternatively, 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 issues, 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 divided 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

k

⋅ K i + α l ⋅ Li )

(2.8)

i =1

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 comprehensive investigations on comfort outdoors. Modelling Tmrt also requires simplifications. Surface temperatures are here an additional limitation, and are only accurately determined if substrate and wall heat storage is included. The method used in ENVI-met relies 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 discussed 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 interpreted 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 differences 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 relevance 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 locations in Cambridge, UK, at different seasons. The measurements included all relevant meteorological 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 characterizing 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 between 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 suggests a significant widening of the comfort zone for outdoors, and consequently less discomfort 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 assessments. 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. temperature of neutrality of 24.1°C for OUT_SET* vs. 23.4 °C with PET) whereas larger differences 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 extreme hot conditions showed that OUT_SET* provides lower values (by 27 % less) because 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 history 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 temperature (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 indoor 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 several hours for cold outdoor conditions against half an hour for hot conditions. This supposes 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 humanbiometeorological 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 conditions, 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 interdependence of Tmrt and the global irradiation G was as important (R2 > 0.92), with a clear distinction 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 diminishes, however, for transitional seasons when the solar radiation is lower. Moreover, a statistical regression between Ta and PET revealed an exponential relationship, however, with less good correlation than that observed for Tmrt, certainly because PET experiences contrasting 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 designing 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

2. Literature review -

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, vegetation, “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 studies investigated their climatic efficiency quantitatively (e.g. Pearlmutter et al. 1999, Grundström 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 temperatures by means of the empirical CTTC model, together with a simple linear relationship 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 conditions 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, Bourbia 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 afternoon) 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 measurements 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, diffuse and reflected radiation. A pedestrian gained less radiant heat in comparison to a person 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 advantages 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 microscale 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 temperature 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 comfort 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 questionable 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 movement 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 microclimate, 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 comfort 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 between 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 solutions is lacking, or performed with weak methods. These statements are the main motivation of the following investigation, which aims at providing more knowledge on the dependence 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 variables 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 purposes (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 numerical 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 examples: 2D rather than 3D, they focus on the energy fluxes prognosis and assume predefined street configurations, with buildings of uniform shape and height, dry surfaces, no 57

3. The numerical model ENVI-met 3.0

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 modelling are faster and more accurate (e.g. Arnfield et al. 1998). Other models are more empirical 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 addition 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 calculated 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 thermodynamics). 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 physiological processes of evapotranspiration and photosynthesis. Various types of vegetation 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 resolution (up to 10 s) allow a fine reading of the microclimatic changes, especially sensible to urban geometry and pertinent for comfort issues. 4. The model requires a limited number of inputs and provides a large number of outputs. 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 building, 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 progressively 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 E, ε from 1D boundary model

H = 2500 m

I N F L O W

O U T F L O W

u , v , E, ε, θ , q

Z≈2Hmax u , v , E, ε, ∂ θ ,q / ∂x = 0

( )

∂(u,v,E, ε ) / ∂x = 0

z y

Hmax Nesting grids T0 and q0 provided by 3Dmodel

x

for z = 0, u = v = E = ε = 0

1D soil model 3 layers H = -1.75 m

Fig.3.1. General scheme 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. equations 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 Navier-Stokes equations. In ENVI-met, the non-hydrostatic incompressible form is used:

 ∂ 2u  ∂u ∂u ∂p ′ + ui =− + K m  2  + f (v − v g ) − S u ( x, y, z ) ∂t ∂xi ∂x  ∂xi 

(3.1a)

 ∂ 2v  ∂v ∂v ∂p ′ + ui =− + K m  2  + f (u − u g ) − S v ( x, y , z ) ∂t ∂xi ∂y  ∂x i 

(3.1b)

 ∂2w  θ (z ) ∂p′ ∂w ∂w − S z ( x, y , z ) + K m  2  + g =− + ui θ ref ( z ) ∂z ∂xi ∂t  ∂xi 

(3.1c)

∂u ∂v ∂w + + =0 ∂x ∂y ∂z

(3.2)

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 Navier-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):

S ui =

∂p′ = cd , f LAD(z ) ⋅W ⋅ ui ∂xi

(3.3)

61


where W = (u2+v2+z2)0.5 is the mean wind speed at a height z, LAD is the leaf area density 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:  ∂ 2θ ∂θ ∂θ + ui = K h  2 ∂t ∂xi  ∂xi

 1 ∂Rn ,lw  + + Qh  c p ρ ∂z

(3.4)

 ∂ 2q  ∂q ∂q + ui = K q  2  + Qq ∂t ∂xi  ∂xi 

(3.5)

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 longwave 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 determined 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 obstacles such as buildings or vegetation, which are typical in urban environments. The closure 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 advection 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 coefficients, namely the turbulent kinetic energy E and its dissipation ε. These equations are given by Mellor and Yamada (1975) as follows:  ∂2E  ∂E ∂E + ui = K E  2  + Pr + Th + QE − ε ∂t ∂xi  ∂xi 

(3.6)

 ∂ 2ε  ∂ε ∂ε ε ε ε2 + ui = K ε  2  + c1 Pr + c3 Th − c2 + Qε ∂t E E E ∂xi ∂ x  i 

(3.7)

where c1, c2 and c3 are standard values taken from Launder and Spalding (1974) but different values might be used for special conditions. The terms Pr and Th describe respectively the production and dissipation of turbulent energy due to wind shearing and thermal stratification (buoyancy production) and are given by:  ∂u ∂u j Pr = K m  i +  ∂x  j ∂xi Th =

g

θ ref ( z )

Kh

 ∂ui   ∂x  j

(3.8)

∂θ ∂z

(3.9)

θref(z) is the reference potential temperature at the inflow boundary and g is the acceleration 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 considered 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 diffusion coefficients Km, Kh and kq are calculated with the assumption of isotropy of the local turbulence: K m = cµ

E2

(3.12a)

ε

K h = K q = 1.35 K m

(3.12b)

63


KE = Kε =

Km

(3.12c)

Km

(3.12d)

σE 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 turbulence 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 decisive in the nature of turbulent exchange, i.e. free convection or molecular exchange, and can be expressed by the Bulk-Richardson number: Rib =

g ∆θ ⋅ ∆w θ (∆u )2

(3.13)

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 absorption 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 (Paltridge 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: ↓ Rlw (z)=

N

∑σ

BT

4

( n )[ε n ( l + ∆l ) − ε n ( l )]

(3.14)

n =1

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 integration of the radiation intensity of the sun I0 in the wavelength range of λ = 0.29 to λ = 4.0. R*sw =

∫

4.0

0.29

I 0 (λ ) exp{− α R (λ )m + α M (λ )m}dλ

(3.15a)

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 = λ1.3

0 βtr). The absolute amount of direct short-wave radiation at the model boundary Rsw ,dir

is obtained after the deduction of the energy quantity absorbed Rsw,abs by the water contained in the atmosphere after Liljequist (1979), namely: 0 * * R sw ,dir = R sw − R sw ,abs = R sw − (70 + 2.8VP2 m ⋅ m )

(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 :

σ svf =

1 360 ∑ 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 radiation 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: 0 Rsw ,dir ( z ) = σ sw ,dir ( z )Rsw ,dir

(3.19)

and the total diffuse radiation (including the diffusely reflected radiation) Rsw,dif is given by: 0 0 Rsw,dif ( z ) = σ sw,dif (z )σ svf ( z )Rsw ,dif + (1 − σ svf ( z ))Rsw ,dir ⋅ a

(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 radiation 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

4

(3.22)

4

where T f + and T f − are the average foliage temperature of the overlying (+) and underlying (-) vegetation layer; T0 is the ground temperature and Tw is the average surface temperature 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 turbulent 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 approximated 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 coefficients 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 energy fluxes at the ground surface are given by: G0 = λs

T − Tk =−1 ∂T = λs (k = −1) 0 ∂z 0.5∆z k =−1

 ∂T H 0 = ρc p .− K h0 ∂z 

(3.25)

 0 T0 − θ k =1     = ρc P  K h z =0   0.5 ∆z k =1 

 q − qk =1  ∂q   LE0 = ρ .L0 − K q0  = ρ ⋅ L0  K q0 0  0.5 ∆z k =1  ∂z   

with

67

(3.26)

(3.27a)


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 materials 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 surfaces, 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 =

(

)

(

)

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

4

68

4

4

(3.31)


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: H w = ρc p K hw

∂T ∂xi

= ρc p K hw w

Tw − Tw+1 ∆w

(3.32)

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 various 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 all boundary surfaces in the whole model, namely the ground surface, wall and roof surfaces. The soil model takes into account the hydrological and thermo-dynamical processes and consists of ten nodes in a vertical profile with a depth of -1.75 m. 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 upper 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 available 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 conductivity 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 corresponds a leaf area density (LAD) and a root area density (RAD), respectively. 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

(3.34c)

−1

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 fraction 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 )

70

(3.36)


where F describes the orientation of the leafs toward the sun ( = 0.5 for randomly oriented 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 approximation of Tmrt at street level, which is expressed for each grid point (z) as follows (Bruse 1999): Tmrt

 1 = σ B

   Et ( z ) + α k (Dt ( z ) + I t ( z ))   εp  

0.25

(3.39)

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 : Rlw↔ ( z ) = (1 − σ svf ( z )) ε w σ B Tw

4

(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: ↓ ,0 ↓ ,0 Dt ( z ) = σ svf ( z ) Rsw ,dif + ( 1 − σ svf ( z )) a Rsw ,dir

(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 calculated 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 surface 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 humidity 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 calculated 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 (diastrophy 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 parameters (T,η), surface quantities (T0 , q0, as), radiation update, the update of wind components ( 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 minutes). 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 equations 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 separately 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 Sahara 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 temperature Ta > 40 °C is not rare and the daily Ta amplitude is relatively large. The atmospheric 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 typical 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 Ta, ENVI-met

800

36 30

S, ENVI-met Ta,measured

600 400

24 18

Ta (°C), VP (hPa)

S, G ( Wm-2)

1000

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 Ghadaia 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 reported 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 dimensions 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 Modell (0,1,2 ) = 1 Upper Boundary for e-epsilon (0:clsd.,1:op.) = 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

U-values for “standard” building materials were used (Koenigsberger et al. 1973, Markus and Morris 1980) with typical albedos (Oke 1987, VDI 1998) to allow some comparison with other climates eventually. Since no heat storage is included in the model,

75


the use of materials of high thermal capacity was not particularly relevant. The simulation 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 incidence 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. overestimation 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 average 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) e.

urban canyon with H/W = 1 oriented N-S with a central row of trees and two lateral galleries (II-5).

76


**

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 number 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 vertically. 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 midblock 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)

II-1

N-S

I-2 (H/W = 1)

II-2

I-3 (H/W = 2)

I- 4 (H/W = 3)

I-5 (H/W = 4)

II-4

II-5

II-3

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 shallowest 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 = 2

37 Ta (°C)

unobstructed surface

Ta, H/W = 1

38

Ta, H/W = 3

36

Ta, H/W = 4

35 34 33 32 31 30 7

8

9

10

11

12

13 14 15 time (LST)

16

17

18

19

20

21

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

4. Results of the numerical simulations

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 differences 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. 4.1.2. Radiation fluxes

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 increase 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 irradiation 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 difference 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 S_H/W = 2

800

S_H/W = 3

2

S (W/m )

700

S_H/W = 4

600 500 400 300 200 100 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

2

D (W/m )

D_H/W = 4

150

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 because 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 additional 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 G_H/W = 4

2

G (W/m )

800 600 400 200 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 various aspect ratios of 0.5, 1, 2, 3 and 4 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 twodimensional 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 temporal illustration of the thermal comfort outdoors This representation has the advantage of giving a complete picture of the diurnal evolution 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 underestimated if the investigation was limited to a single point within the street, as usually assumed (e.g. Pearlmutter at al. 1999, Swaid et al. 1993). This information is interestingly 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 comfort is long enough and if pedestrians are given the choice of moving to shaded subspaces 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 (Nakamura and Oke 1988). This uniformity is well reproduced by the model but the warming 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 aspect 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, Grundström et al. 2003). The use of Tmrt appears to be playing a more decisive role as it influences 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

15:00 14:00

16:00

37°C

15:00

36°C

11:00

74°C

14:00

68°C 62°C

12:00

56°C

34°C

11:00

50°C

10:00

44°C

9:00

38°C

8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

32°C

33°C

8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

80°C

35°C

10:00 9:00

86°C

13:00

13:00 12:00

time (LST)

time (LST)

16:00

38°C

32°C

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 effective 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

time (LST)

86

30 °C 8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m |----------------- street width --------------------|

Fig. 4.5 b: Diurnal variation of PET at street level for E-W oriented streets of an aspect ratio H/W = 1 (right)

Fig. 4.5a: Diurnal variation of PET street level for E-W oriented streets of an aspect ratio H/W = 0.5 (left)

|----------------- street width --------------------|

34 °C

9:00

9:00

38 °C

42 °C

46 °C

10:00

50 °C

10:00

14:00

54 °C

11:00

14:00

11:00

15:00

15:00

58 °C

12:00

16:00

16:00

62 °C

12:00

17:00

17:00

66 °C

70 °C

13:00

18:00

18:00

PET, H/W = 1, East-West

13:00

19:00

20:00

19:00

20:00

PET, H/W = 0.5, East-West

Ghardaia, 32.40° N, 3.80° E, 01 August


time (LST)

87 9:00


9:00


Fig. 4.5 d: Diurnal variation of PET at street level for E-W oriented streets of an aspect ratio H/W = 3 (right)

Fig. 4.5c: Diurnal variation of PET at street level for E-W oriented streets of an aspect ratio H/W = 2 (left)

10:00

14:00

10:00

14:00

15:00

11:00

15:00

11:00

16:00

16:00

12:00

17:00

17:00

12:00

18:00

18:00

13:00

19:00

19:00


13:00

20:00

20:00



30 °C

34 °C

38 °C

42 °C

46 °C

50 °C

54 °C

58 °C

62 °C

66 °C

70 °C



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 Ta. 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

time (LST)

16:00

58 °C

15:00

54 °C

14:00

50 °C

13:00

46 °C

12:00

42 °C

11:00 10:00

38 °C

9:00

34 °C


30 °C

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 between 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 effectiveness of the aspect ratio is rather limited in ensuring comfortable microclimates for an EW orientation for subtropical latitudes. The main reason is the lack of shading as suggested 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

4.2.1. Air temperature

Fig. 4.6 shows the evolution of the air temperature Ta during the daytime for urban canyons 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 surface

Ta, H/W = 1

38

Ta, H/W = 2

Ta (°C)

Ta, H/W = 3 Ta, H/W = 4

36

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 4.2.2. Radiation fluxes

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 oriented 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 protected 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

4. Results of the numerical simulations Ghardaia, 32.40° N, 1st August, N-S orientation 1200 S_H/W = 0,5 S_H/W = 1

1000

S_H/W = 2 S_H/W = 3

S (W/m2)

800

S_H/W = 4

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

Ghardaia, 32.40° N, 1st August, N-S orientation

250 D_H/W = 0,5 D_H/W = 1

200

D_H/W = 2

D (W/m2)

D_H/W = 3 D_H/W = 4

150

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

4. Results of the numerical simulations Ghardaia, 32.40° N, 1st August, N-S orientation

2

G (W/m )

1300 1200

G_H/W = 0,5

1100

G_H/W = 1

1000

G_H/W = 2

900

G_H/W = 3

800

G_H/W = 4

700 600 500 400 300 200 100 0 8

9

10

11

12

13 14 time (LST)

15

16

17

18

19

20

Fig. 4.7c. The simulated global radiation (G) at street level for NS oriented streets of various aspect ratios of 0.5, 1, 2, 3 and 4 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 component 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 distribution of PET for N-S streets with H/W = 0.5, 1, 2, 3 and 4 respectively, for the 1st August 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 westfacing 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 comparison 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 experiences an average PET of 34 °C. Fig. 4.8c shows that higher aspect ratios, i.e. from H/W = 2, visibly accentuate the positive 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, occurring 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 remains 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 further 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

time (LST)

94

34 °C 30 °C

9:00 8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

9:00

8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

Fig. 4.8b: Diurnal variation of PET at street level for N-S oriented streets of an aspect ratio H/W = 1 (right)

Fig. 4.8a: Diurnal variation of PET at street level for N-S oriented streets of an aspect ratio H/W = 0.5 (left)

|------------------- street width -----------------------|

38 °C

10:00

10:00

|------------------- street width -----------------------|

42 °C

11:00

11:00

14:00

14:00

46 °C

15:00

15:00

12:00

54 °C

16:00

16:00

12:00

58 °C

17:00

17:00

50 °C

62 °C

18:00

18:00

13:00

66 °C

19:00

19:00

13:00

70 °C

20:00

PET, H/W = 1, North-South

20:00

PET, H/W = 0.5, North-South

Ghardaia, 32.40° N, 3.80° E , 01 August


time (LST)

95

8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

Fig. 4.8d: Diurnal variation of PET at street level for N-S oriented streets of an aspect ratio H/W = 3 (right)

Fig. 4.8c: Diurnal variation of PET at street level for N-S oriented streets of an aspect ratio H/W = 2 (left)

|-------------------- street width ----------------------|

9:00

9:00

|------------------- street width -----------------------|

10:00

10:00

14:00

11:00

14:00

11:00

15:00

15:00

12:00

16:00

16:00

12:00

17:00

17:00

13:00

18:00

18:00

13:00

66 °C

19:00

19:00

30 °C

34 °C

38 °C

42 °C

46 °C

50 °C

54 °C

58 °C

62 °C

70 °C

20:00


20:00





Ghardaia, 32.40° N, 3.80°E, 01 August PET, H/W = 4, North-South

time (LST)

20:00 19:00

70 °C

18:00

66 °C

17:00

62 °C

16:00

58 °C

15:00

56 °C

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 0m 1m 2m 3m 4m 5m 6m 7m 8m

30 °C

|--------------------- 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 exceed 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 influences 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 ratio, and ∆G shows the same temporal trend as ∆Ta .

96


Ghardaia, Algeria, 32.40 °N, 1st August

1.6 H/W = 2

1.2

∆Ta K

0.8 H/W = 1

0.4

H/W = 0.5

0

H/W = 3

-0.4

H/W = 4

-0.8 -1.2 8

9

10

11

12

13 14 time (LST)

15

16

17

18

19

20

1000 800 600

2 ∆ G (W/m )

400 200 0 -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 orientations for each H/W ratio from 0.5 to 4, respectively. Primarily, the graphics show a larger 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: EW 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, followed by Ta as a second modifying factor, while v and VP are not relevant in these differences.

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 E-W

N-S

aspect ratio

PET[max]

Tmrt

Ta

v

VP -1

(°C)

(°C)

(°C)

(ms )

(hPa)

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

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 compares 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 because 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: ↓ ) is maximal around The amount of direct short-wave irradiance absorbed ( 0.7 f p Rsw ,dir

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 irradiance 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

4. Results of the numerical simulations Ghardaia, 32.40 °N, 01 August 600

41

E-W

total LW absorbed

39

400

Ta

total SW absorbed

37

LW from ground

300

35 SW direct

200

Ta (°C)

Radiation flux Wm-2

500

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

39

total LW absorbed

400

37

Ta LW from ground

300

200

35

33

total SW absorbed

SW diffuse

Ta (°C)

Radiation flux Wm-2

500

LW from sky

100

31 LW from buildings

SW direct

0

29 8

9

10

11

12

13 14 15 time (LST)

16

17

18

19

20

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) Time LST 9:00 10:00 16:00 17:00

SW_dir. SW_dif. Wm-2 Wm-2 E-W orientation 230.3 120.5 210.7 133.2 219.2 130.0 230.2 115.0 N-S orientation

Σ LW_grd SW . Wm-2 Wm-2

LW_sky Wm-2

Σ Σ LW_bldgs LW (SW+LW) Wm-2 Wm-2 Wm-2

Tmrt °C

Ta °C

PET °C

350.8 343.9 349.2 345.2

267.5 294.0 335.6 322.4

125.9 129.6 139.2 138.3

86.8 86.9 87.0 86.9

480.2 510.5 561.8 547.6

831.0 854.4 911.0 892.8

77.7 80.0 85.9 84.2

33.1 34.4 38.0 38.0

58.9 61.2 67.9 66.9

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 influence 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 temperature 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 canyons 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) increases),

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) Time LST

SW_dir. SW_dif. Wm-2 Wm-2 E-W orientation 9:00 230.3 142.6 10:00 210.7 156.1 16:00 219.2 152.7 17:00 230.2 136.6 N-S orientation 13:00 107.5 168.6

Σ Σ Σ SW LW_grd. LW_sky LW_bldgs LW (SW+LW) Wm-2 Wm-2 Wm-2 Wm-2 Wm-2 Wm-2

Tmrt °C

Ta °C

372.9 366.8 371.9 366.8

266.6 297.8 308.7 308.5

85.4 87.2 91.8 91.4

124.3 124.4 124.5 124.4

476.2 509.4 524.9 524.4

849.1 876.2 896.9 891.3

78.4 81.0 83.3 83.0

33.2 34.3 37.0 37.1

276.0

321.3

92.5

124.5

538.2

814.3

74.4

36.7 58.7

102

PET °C 60.7 62.9 65.4 65.3


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 orientations 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 NS 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 thermal 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 (b) PET, H/W = 2, North-South

time (LST)

(a) PET, H/W = 2, East-West 20:00

20:00

19:00

19:00

18:00

18:00

17:00

17:00

16:00

16:00

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

0m 1m 2m 3m 4m 5m 6m 7m 8m

64°C 60°C 56°C

0m 1m 2m 3m 4m 5m 6m 7m 8m

52°C

(d) PET, H/W = 2, NW-SE

48°C

(c) PET, H/W = 2, NE-SW

time (LST)

68°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

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

28°C

0m 1m 2m 3m 4m 5m 6m 7m 8m

0m 1m 2m 3m 4m 5m 6m 7m 8m

|--------------- street width ---------------|

|--------------- street width ---------------|

Fig. 4.12. Comparison of PET patterns according to street orientations E-W, N-S, NESW 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. 4.5.1. Air temperature

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). NESW 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 geometries 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 canyons 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 various crown densities and two wind incidences. Air temperature observed in planted canyons 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, however, 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


Ghardaia, 32.40° N, 3.80°E, 01 August 39 38 37

Ta (°C)

36 35 34

N-S, asym. H/W = 2 & 1

NE-SW, asym. H/W = 2 & 1

33

E-W, asym. H/W = 2 & 1

NW-SE, asym. H/W = 2 & 1

NS, H/W = 2

NE-SW, H/W = 2

EW, H/W = 2

NW-SE, H/W = 2

32 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 Ghardaia, 32.40° N, 3.80°E, 01 August 39 38 37

Ta (°C)

36 35 34 33 32

N-S, asym. & overhangs

NE-SW, asym. & overhangs

E-W, asym. & overhangs

NW-SE, asym. & overhangs

NS, H/W = 2

NE-SW, H/W = 2

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


Ghardaia, 32.40° N, 3.80°E, 01 August 39 38 37

Ta (°C)

36 35 E-W, dense tree, H/W = 2, perp. wind E-W, dense tree, H/W = 2, paral. wind E-W, light dense tree, H/W = 2, paral. wind N-S, dense tree, H/W = 1, perp. wind E-W, light dense tree, H/W = 1, paral. wind E-W, H/W = 2 N-S, H/W = 1 E-W, H/W = 1

34 33 32 31 30 8

9

10

11

12

13

14 15 time (LST)

16

17

18

19

20

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 Ghardaia, 32.40° N, 3.80°E, 01 August 40 39 38

Ta (°C)

37 36 35 E-W, H/W = 0.5, perp. wind

34

E-W, H/W = 1, perp. wind

33

E-W, H/W = 2, perp. wind E-W, H/W = 0.5, paral. wind

32

E-W, H/W = 1, paral. wind

31

E-W, H/W = 2, paral. wind

30 8

9

10

11

12

13

14 15 time (LST)

16

17

18

19

20

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 sensible 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 including 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 values, which range between 34 °C and 42 °C. However, these covered spaces also experience 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 orientation 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 extreme 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


Ghardaia, 32.40° N, 3.80° E, 01 August (a) PET, H/W = 2 with galleries, E-W orientation 20:00 19:00

70 °C

18:00

66 °C

17:00

62 °C

time (LST)

16:00

58 °C

15:00

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 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

30 °C

|------ 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

time (LST)

16:00

58 °C

15:00

54 °C

14:00

50 °C 13:00

46 °C

12:00

42 °C

11:00 10:00

38 °C

9:00

34 °C


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


Ghardaia, 32.40° N, 3.80° E, 01 August (c) PET, H/W = 2 with galleries, NE-SW orientation 20:00 19:00

70 °C

18:00

66 °C

17:00

62 °C

time (LST)

16:00

58 °C

15:00

54 °C

14:00

50 °C 13:00

46 °C

12:00

42 °C

11:00 10:00

38 °C

9:00

34 °C


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

time (LST)

16:00

58 °C

15:00

54 °C

14:00

50 °C

13:00

46 °C

12:00

42 °C

11:00 10:00

38 °C

9:00

34 °C


30 °C

|----- gal. SW -----|------------------- street width -----------------------|------ gal. NE -----|

Fig. 4.14c and 4.14d. PET distribution across symmetrical urban canyons including galleries 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 covered 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 extreme 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 differences in the air temperature and wind speed found across the street at mid-block distance. 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 average) 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

4. Results of the numerical simulations Ghardaia, 32.40 °N, 01 August 600 LW_total street

500 radiation flux Wm-2

LW_total gallery

400 300 SW_total gallery

200 100

SW_total street

0 8

9

10

11

12

13

14

15

16

17

18

19

20

450 400

radiation flux Wm-2

350

gal. SW-dif

gal. SW-dir

gal. LW-ground

gal. LW-sky

gal. LW-walls

str. SW-dif

str. SW-dir

str. LW-ground

str. LW-sky

str. LW-walls

300 250 200 150 100 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) 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 Σ SW_dir. SW_dif. SW LW_grd. Wm-2 Wm-2 Wm-2 Wm-2 in gallery 11:00 171.1 198.7 369.8 271.6 within street 12:00 126.1 144.7 270.8 302.6 Time LST

13:00

107.5

145.7

253.1

330.6

Σ Σ LW_sky LW_bldgs LW (SW+LW) Tmrt Wm-2 Wm-2 Wm-2 Wm-2 °C

Ta °C

PET °C

26.8

177.9

476.3

846.0

83.3 34.7 64.6

133.3

86.8

522.6

793.4

73.5 36.0 58.3

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. However, a different parameterisation than the SVF seems to be necessary for a better estimation of the various fluxes accounting in Tmrt in case of covered urban spaces. No information 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 comparison to a symmetrical canyon with H/W = 2 but with an offset of the facades to promote 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 expected 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 geometry of H/W = 1 (see Fig. 4.5b), the spatial and temporal evolution of PET is noticeably similar.

113

4. Results of the numerical simulations Ghardaia, 32.40° N, 3.80° E, 01 August 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

time (LST)

16:00

58 °C

15:00

54 °C

14:00

50 °C 13:00

46 °C

12:00

42 °C

11:00 10:00

38 °C

9:00

34 °C


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 Ghardaia, 32.40° N, 3.80° E, 01 August

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 situation 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 component 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 asymmetrical 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 reflected 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 intermediate orientations show the same similar trends.

115

4. Results of the numerical simulations Ghardaia, 32.40° N, 3.80° E, 01 August 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

time (LST)

16:00 15:00

54°C

14:00

50°C

13:00

46°C

12:00

42°C

11:00

38°C

10:00

34°C

9:00 8:00 0m

30°C 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 Ghardaia, 32.40° N, 3.80° E, 01 August

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

4. Results of the numerical simulations Ghardaia, 32.40° N, 3.80° E, 01 August 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

time (LST)

16:00 15:00

54°C

14:00

50°C

13:00

46°C

12:00

42°C

11:00 38°C

10:00

34°C

9:00

30°C

8:00 0m

1m

2m

3m

4m

5m

6m

7m

8m

9 m 10 m 11 m 12 m 13 m 14 m

|------ 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 Ghardaia, 32.40° N, 3.80° E, 01 August

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

time (LST)

16:00 15:00

54°C

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 0m

1m

2m

3m

4m

5m

6m

7m

8m

9 m 10 m 11 m 12 m 13 m 14 m

|------- 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

4. Results of the numerical simulations -

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 effectiveness 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 affected 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 canyons. The use of self-shading façades or horizontal shading devices on the walls is wellknown 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 overhanging 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 inclined 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 recorded 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

4. Results of the numerical simulations -

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 situation 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 vegetation 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 attenuation 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

4. Results of the numerical simulations Ghardaia, 32.40° N, 3.80° E, 01 August

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

time (LST)

16:00

58 °C

15:00

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 0m 1m 2m 3m 4m 5m 6m 7m 8m

30 °C

|---------------- 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) Ghardaia, 32.40° N, 3.80° E, 01 August ∆ L-upwards, E-W, H/W =2, with vs. without trees 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

∆ S, H/W =2, with vs. without trees, E-W 20:00

14:00 13:00

14:00 13:00

12:00

12:00

11:00

11:00

10:00

10:00

9:00

9:00

8:00 0m

1m

2m

3m

4m

5m

6m

7m

8:00 0m

8m

I-------------------- street width ----------------------I

1m

2m

3m

4m

5m

6m

7m

8m

I------------------------- street width ---------------------------I

Fig. 4.22. Differences in (a) direct solar radiation (∆S) and (b) long-wave radiation (∆Lupwards) 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 explanation 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 absorbed 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 results 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. Ghardaia, 32.40° N, 3.80° E, 01 August PET, H/W = 1 with a central row of trees, N-S 20:00 19:00

62 °C

18:00 58 °C

17:00

time (LST)

16:00

54 °C

15:00

50 °C

14:00 46 °C

13:00 12:00

42 °C

11:00

38 °C

10:00 34 °C

9:00 8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m 9m 10m 11m 12m 13m 14m

30 °C

|----- 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

4. Results of the numerical simulations Ghardaia, 32.40 °N, 01 August 600 LW_total, no trees

-2

radiation flux (Wm )

500 400

LW_total, with trees

300 SW_total, no trees

200 SW_total, with trees

100 0 400

8

350

9

10

11

12

13

14

15

16

17

18

19

20

SW-dif.

SW-dir.

LW-grd.

LW-sky

LW-bldgs.

SW-dif., tree

SW-dir., tree

LW-grd., tree

LW-sky, tree

LW-bldgs., tree

radiation flux (Wm-2)

300 250 200 150 100 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 Time LST

SW_dir. Wm-2

SW_dif. Wm-2

Σ SW Wm-2

LW_grd. Wm-2

LW_sky Wm-2

LW_bldgs Wm-2

Σ LW Wm-2

Σ (SW+LW) Wm-2

Tmrt °C

11:00 12:00 13:00 14:00

no trees 171.1 126.1 107.5 138.9 with trees

116.9 120.7 121.7 119.8

287.9 246.8 229.1 258.8

281.3 294.2 301.6 304.2

177.3 182.1 186.1 187.4

49.5 49.5 49.5 49.5

508.1 525.7 537.2 541.1

796.0 772.5 766.3 799.8

74.7 72.2 71.5 75.2

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

Ta °C

PET °C

35.7 59.9 36.9 58.8 38.0 58.7 38.5 61.5


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 addition, that a central row of trees does not protect the galleries better, which still experience 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 mechanisms is not possible within the framework of this study. Nevertheless, a number of observations 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 (perpendicular 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 differences 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

126

(a)

(b)

|------ street width ----|



v (m/s), H/W = 0.5, parallel wind

v (m/s), H/W = 0.5, perpendicular wind

v (m/s), H/W = 2, perpendicular wind

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

v (m/s), H/W = 2, parallel wind

0m 0m 4m 8m 12 m 16 m 20 m |--------------------- street lenght ------------------------|

2m

4m

6m

8m

0m 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 --------------------------------------------------------------|

2m

4m

6m

8m

0m 0m

2m

4m

6m

8m

0m 0m 4m 8m 12 m 16 m 20 m |--------------------- street lenght ------------------------|

2m

4m

6m

8m |------ street width ----|

Ghardaia, 32.40° N, 3.80° N, 01August

at 1.2 m a.g.l. level for both perpendicular and parallel wind incidence on street axis 0.0 m/s

0.4 m/s

0.8 m/s

1.2 m/s

1.6 m/s

2.0 m/s

2.4 m/s

2.8 m/s

3.2 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 intermittent 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 between 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 progressively 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 downward 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 temperatures 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 dimensions according to simulation results

128


Ghardaia, 32.40° N, 3.80° E, 01 August (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)

(a) PET, H/W = 2, parallel wind, E-W

14:00 13:00

14:00 13:00

12:00

12:00

11:00

11:00

10:00

10:00

9:00

9:00

8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m |---------------- 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)

Ghardaia, 32.40° N, 3.80° E, 01 August 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

50 °C

15:00 14:00

46 °C

13:00

42 °C

12:00

38 °C

11:00

time (LST)

time (LST)

16:00

15:00 14:00 13:00 12:00 11:00

34 °C

10:00

30 °C

9:00 8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m

26 °C

|---------------- street width -------------------|

10:00 9:00 8:00 0m 1m 2m 3m 4m 5m 6m 7m 8m |---------------- 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 perpendicular 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 characterized by a greater potential of solar exposure which might reduce the advantages obtained by a stronger wind flow. This calls attention on the necessity of assessing comfort 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 indoors 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 summer, 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 ques130


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 canyons 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 preferred 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 available 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 others (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 pedestrian use as the outdoor comfort is more probable even for winter season when sun exposure 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 unobstructed surface and attributable to additional diffusely reflected radiation from the vertical 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 indoors) 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 (b) S.A.I., Walls, Winter, N-S 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)

(a) S.A.I., Walls, Winter, E-W 20:00

14:00 13:00

14:00 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

|-------------- aspect ratio (H/W) -------------|

0.5

1.0

1.5

2.0

2.5

3.0

3.5

|-------------- 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

4.0


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 internal 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 potentially 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 canyons 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

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

5. Field measurements in Freiburg, Germany

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 continuously recorded at regular time intervals at two heights: 1.4 m and 3.1 m. The shortwave 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 (1) Global radiation K

Unit -2

Wm

Instrument

height (a.g.l) number

Pyranometer, CM21, CR11, Kipp 1.4 m

6 ( ↑, ↓, N, E, W, S )

& Zonen (2) Net radiation Q

-2

Pyradiometer, Schenk

-2

difference between all-wave and 1.4 m

Wm

(3) Long-wave radiation Wm

1.4 m

6( ↑, ↓, N, E, W, S ) -

short-wave radiation (Q – K)

L (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

ms-1

Cup anemometer, Vector Instru- 1.4 m, 3.1 m

2

(vapour pressure) (6) Wind speed v

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 calculation 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 measurements 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 permanent 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 “background 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 stations 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 during which a maximum temperature of 40.2 °C was reached in Freiburg. Freiburg, 14/15 July 2003 36

Ta (°C)

32

28

Ta_3.1m Ta_1.4m Ta_roof

24

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 afternoon 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 recorded 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 experience 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 limited 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. Nakamura 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 reported on large differences in the surface temperatures between sunlit and shaded surfaces. 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 similar findings for a deeper street oriented close to N-S, but with a slight thermal stratification. 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 (socalled Höllentäler) driven down a valley and leading to a noticeably cooler night situation, while the urban climate station at roof level is located in a typical heat island zone. Freiburg, 14/15 July 2003 60 Tground

55

Twall

Ta_3.1 m

Ta_1.4 m

Ta, Ts , Tw (°C)

50 45 40 35 30 25 20 15 8

12

16

20

0

4

8

12

16

20

time (LST) 60

Ts, Tw (°C)

50

40

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 below, 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 aboveroof 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 perpendicular 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 approximation.

143

5. Field measurements in Freiburg, Germany Freiburg, 14./15.7.2003 360

wind direction (°)

270

180

90 at 51 m height in-canyon at 3.1 m height

0 10

12

14

16

18

20

22

0 2 4 time (LST)

6

8

10

12

14

16

18

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

wind speed (m/s)

8

6

4

2

0 10

12

14

16

18

20

22

0 2 4 time (LST)

6

8

10

12

14

16

18

Fig. 5.7. Wind speed within the canyon street and above roof level (at H = 61 m) 360 from E (70° to 125°)

above 5 m/s

dd in-canyon (z = 3.1 m a.g.l.) (°)

wind speed in-canyon (z = 3.1 m a.g.l.) ms-1

4

from NW (270° to 335°)

3

R2 = 0.6211

other directions

R2 = 0.4593

2

1

0

180

90

0 0

(a)

below 5 m/s

270

2

4

6

8

10

12 -1

wind speed above roof (z = 51 m a.g.l.) ms

0

(b)

90

180

270

360

dd above roof (z = 51 m a.g.l.) (°)

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 located 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 explained 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 Nakamura 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, Nakamura 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 between inside and outside wind direction whereas much more dispersal is observed below 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 (impinging on a normal surface) for the street area investigated, and including the actual building 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 person 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 relatively 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 contrast, 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 temporal evolution as L↑, though of less magnitude.

147


Erbprinzenstraße, Freiburg, 14/15 July 2003

1000

K↓roof K↓

K (Wm-2)

800

600 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

Erbprinzenstraße, Freiburg, 14/15 July 2003 700 L↑

L (Wm-2)

600

L(W→)

L(N→)

500 L↓ L(S→) L(E→)

400

300 8

12

16

20

24 4 time (LST)

8

12

16

20

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 coloured brick walls. After 17:00 LST, the rate of heat release is slowed down as the canyon 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 increasing 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. Maximal 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 absorbed 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 corresponds 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 directed upwards and downwards. These vary from 80 to 130 Wm-2 for the lateral directions as opposed to between 20 and 40 Wm-2 in the vertical direction. Notably, the differences 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 surfaces (walls and ground) and from the sky simultaneously. The largest variance, however, 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 additional energy from the irradiated opposite part of the street canyon.

150

5. Field measurements in Freiburg, Germany Erbprinzenstraße, Freiburg, 14/15 July 2003 200 Ktotal

Kabs (Wm-2)

160

120 K(W→)

K(E→)

80

K(S→) K(N→)

40

K↓ K↑

0 8

12

16

20

24 4 time (LST)

8

12

16

20

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

Erbprinzenstraße, Freiburg, 14/15 July 2003

160

560 Ltotal

-2

(Wm )

420

L(N→)

L(W→

L(S→)

L(E→)

80

280

Ltotal

Labs (Wm-2)

120

L↑

40

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 irradiation 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. During 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. Erbprinzenstraße, Freiburg, 14/15 July 2003 70 60 Tmrt

Tmrt, PET ( °C)

50 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 absorbs 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 reducing 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 aspect 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

-2

L absorbed by a human body (Wm )

y = 0.9387x - 3.8101 2 R = 0.99

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 overestimated 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 surroundings (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 comparison 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 Tmrt simulated

Tmrt °C

50 40 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 complex 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 estimating 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 standing person, especially when the sun is relatively low, when compared to measured data.

600

Labs measured

500 Labs simulated

Labs, Kabs Wm

-2

400 300 Kabs simulated

200 Kabs measured

100 0 11

13

15

17

19

(a)

21

23

1

3

5

7

9

11

23 1 time (LST)

3

5

7

9

11

time (LST)

700

Lupwards Wm

-2

Lupwards measured

600

Lupwards simulated

500

400 11

(b)

13

15

17

19

21

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 estimated 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 dynamics 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 attributable 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 shortwave 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 conditions. 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 assumed 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. ElAtteuf, 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

6. Field measurements in Beni-Isguen, Algeria

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 building experience. It is commonly claimed that this type of architecture is perfectly adapted to suit the surrounding climatic environment. Yet, this has not been proven and the current knowledge 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 lighting 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 investigation, based on scientific methods, is required to further quantitatively evaluate common design concepts and establish the veracity of this common belief of climatic adaptation. 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. Grundström 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 topography 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 W(m)

aspect ratio H/W

SVF

orientation (angle from N)

ground albedo

ground material

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 irregular 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. start 2

1 3 4

5

8 market place

6

7

route 0

50

1

100 m

3

4

5

6

7

Fig. 6.3. Route and all measuring points within different street geometries in the vernacular city of Beni-Isguen, Mzab valley, Algeria

160

6. Field measurements in Beni-Isguen, Algeria pt. 1

pt. 2

pt. 4

pt. 6

pt. 7

pt. 8

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 because it is an important variable in the human energy balance. The use of a globe thermometer for measuring Tmrt is common but this method was dismissed in this investigation 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 recorded 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

6. Field measurements in Beni-Isguen, Algeria Ghardaia, 26 June 2003 45 pt. 1 pt. 2 pt. 3

40

pt. 4 pt. 5

Ta (°C)

pt. 6 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

VP (hPa)

pt. 5 pt. 6 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

v (m s-1)

4

pt. 6 pt. 7

3

pt. 8

2

1

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 measuring 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 (Nakamura 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 pathway 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 location 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 comparison 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 comprehensive 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 (Nakamura and Oke 1988, Yoshida et al. 1990/91). In the morning, street surface temperatures 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 experienced 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 because of their short duration of exposure. The temperatures of the canyon wall surfaces showed generally small differences between 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 reported 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. Ghardaia, 26 June 2003 65 pt. 1 pt. 2 pt. 3 pt. 4 pt. 5 pt. 6 pt. 7 pt. 8

45

Ts

(°C)

55

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

Tw (°C)

pt. 5

40

pt. 6 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, Algeria 6.3.4.

Radiation fluxes

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) absorbed 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 orientation. Ghardaia, 26 June 2003 400 pt. 1 pt. 2 pt. 3 pt. 4 pt. 5 pt. 6 pt. 7 pt. 8

Kabs (W m-2)

300

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 pt. 5 pt. 6 pt. 7 pt. 8

Labs (W m -2)

540

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 admittance 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 relatively 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 difference 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 standing 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. Ghardaia, 26 June 2003 100 pt. 1 pt. 2 pt. 3 pt. 4

80

pt. 5

Tmrt (°C)

pt. 6 pt. 7 pt. 8

60

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

Tmrt - Ta (K)

30

pt. 6 pt. 7

20

pt. 8

10

0

-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 BeniIsguen, 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, negligible differences were registered between all urban streets and Tmrt averaged 35°C. Indeed, 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 thermal 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 canyons and up to 14 K higher in irradiated situations. Points 2 and 4 had even lower values (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 coefficients of determination of R2 = 0.900 for a linear relationship and R2 = 0.939 for a logarithmic 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

6. Field measurements in Beni-Isguen, Algeria Ghardaia, 26 June 2003 60 pt. 1 pt. 2 pt. 3 pt. 4

50

pt. 5

PET (°C)

pt. 6 pt. 7 pt. 8

40

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 measuring 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.


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 ensuring 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 experi171


enced in the morning hours and lasts for a large part of the day. Yet, PET maxima recorded 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 canyons). 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 surface. In fact, this difference is likely due to more absorption in the real case and no adjusting factor for solar radiation was used in the present simulations. Within street canyons, 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 lowest PET values as the heat emitted from these surfaces is noticeably lower in comparison 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 morning 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 representation 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, shading 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 materials. Thick and heavy materials with high thermal capacities help to decrease the longwave radiant heat during the day and minimize the differences between the streets of different geometries and orientations. However, when high thermal capacity is combined with high aspects ratios, the heat released from the canyon surfaces in the nighttime 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 investigation 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 Grundström 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 surveys will bring more knowledge on the reliability of these indices and refine their scaling. Such work is particularly lacking in such severe climates, where people‘s subjective 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 investigation. Moreover, this work shows that more continuous measurements are needed for establishing the dependence between air temperature and the urban structure.

Ali-Toudert et al (2005) described the strong interdependence between the various scales in these urban structures regarding climate adaptation. An exhaustive assessment of the effectiveness of these design concepts should, therefore, deal simultaneously with the indoor and outdoor climates, e.g. at a neighbourhood scale, and include summer and winter issues, i.e. internal and external thermal and visual comfort, ventilation, etc. Such extensive studies are lacking.

174

7.


The present study dealt with outdoor thermal comfort in dependence to 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. galleries, 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 bio-meteorological method for the assessment of the thermal 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 ENVImet 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

7. Discussion and conclusion

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 orientation. 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, EW 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 NWSE 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 unobstructed 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 underestimates 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 content 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 comparable summer conditions. Santamouris et al. (1999) and Nakamura and Oke (1988) reported that no clear correlation 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 vegetation, and the orientation. Increasing H/W or planting trees are the most efficient strategies 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 empirical assumptions of the CTTC model and calls attention to the methodological problems 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 indicator. 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 previously 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 configurations tested and insignificantly for those with vegetation, mainly because of the simulated 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 conditions. 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 improves 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 perpendicular 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 investigated. This agrees with available knowledge on urban canyon wind flow (e.g. Nakamura and oke 1988, de Paul and Shieh 1986, Santamouris et al. 1999). A parallel incidence 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 progressively 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 simulations 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 spatially (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 methods used for calculating the heat gained by a pedestrian differed between simulations and measurements, making difficult a comparison of each individual component. Nevertheless, 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 shortwave 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 homogenous 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 outdoor 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 temperatures (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 comfort 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 as180


pect ratio. The contrast in the thermal sensation between the two orientations becomes 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 absorbed 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 orientation, 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 contrasting thermal stress, while N-S canyons experience globally the same comfort conditions 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 alternatives 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 NS 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 inhibited 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 observed, up to 100 Wm-2 for Gmax). This difference rises with increasing aspect ratios 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 eq. 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 irradiation 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 observed 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 surface observed in the field measurements, it is believed that the positive effects on comfort 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 moderatly 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 discomfort 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 situation 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 discussion. Although, the study was mostly completed for a subtropical location with extreme hot-dry climate for summer conditions, it is believed that the design recommendations 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 adjustments related to sun course geometry (zenith and azimuth angles) accounting for latitude 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 including 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 require 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 daytime 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 ratio. 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 dominant wind directions for promoting ventilation or protecting from cold winds (AliToudert and Bensalem 2001). An E-W orientation is well known to be preferable if solar gains have to be maximized. Intermediate orientations are less optimal but still provide a good potential of sunlight and daylight. N-S orientation is appropriate for daylight 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 oriented 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 galleries 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 deciduous 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 protected by deciduous trees

Fig. 7.3. Example of wide street canyon oriented E-W, combining comfortable pedestrian zones and motor traffic Fig 7.4 shows an example of higher aspect ratio with an asymmetrical geometry especially 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. Moreover, 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 materials 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 interference 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 perpendicular 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 microclimatic 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 ENVImet, made in the framework of this study, are reported. This information helps to understand the limits of the present investigation, and gives an overview on eventual refinements 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 environment for the main simulation volume assessed by the 3D-model. These inputs include 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 continental and subtropical region like Ghardaia, it showed a trend to underestimate the typical large daily amplitudes where Tmax - Tmin ≈ 12 K in summer. If start values are appropriately 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 daytime. 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 assumes 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 simulations 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 adjust 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 Ghardaia. 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 conductive heat flux, exclusively depending on heat transmittance (U-value) and the gradient 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 admittance 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 simulation 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 estimating 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 environment 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 ENVImet 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 estimated 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 longwave 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 building 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 urban 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, orientation, 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 both physical, physiological and psychological dimensions. An overview of available studies pointed out important methodological differences in assessing comfort. It is still difficult to interpret 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 Environment 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 simulation 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 application 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, exchanges 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 Society for heating Refrigerating and Air Conditioning, Atlanta: 8.1-8.29. ASHRAE 2001b: Chapter 13 – Measurements and instruments. In: Handbook of Fundamentals. American Society for heating Refrigerating and Air Conditioning, Atlanta: 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 canyon 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 buildings. PhD thesis. University of Sheffield. Department of Building Science. UK. Bitan A. 1988: The methodology of applied climatology in planning and building. Energy 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 literature 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. Swedish Council for Building Research, Stockholm. Bruse M., Fleer H. 1998: Simulating surface-plant-air interactions inside urban environments with a three dimensional numerical model. Envir. Model. Software 13: 373-384. Bruse M. 1999: Die Auswirkungen kleinskaliger Umweltgestaltung auf das Mikroklima. Entwicklung des prognostischen numerischen Modells ENVI-met zur Simulation 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 improvement of summer urban climate. Proc. 2nd Japanese-German Meeting “Klimaanalyse 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 properties. 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. Atmos. 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. Mémoire 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: 351372. 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 description 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. Cambridge. Geiger R., Aron R., Todhunter P. 1995: The climate near the ground. Vieweg. Wiesbaden. 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 dargestellt am Beispiel Darmstadt. Teil 1 Wind und Temperaturfelder. Meteorol. Rundschau 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 Development 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 Strahlungstemperatur im Freien. Wetter und Leben 44: 147-151. Höppe P. 1984: Die Energiebilanz des Menschen. Dissertation. Wissenschaftlicher Mitteilung 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. Energy 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 Research 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 elements in a simulated atmospheric boundary layer flow. Part II- Flow over large arrays 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 transpiration. 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 räumlichen Bewertung der thermischen Komponente im Bioklima des Menschen. Fortgeschriebenes 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 and Buildings 35: 15-25. Koenigsberger O.H., Ingersoll T.G., Mahyew A., Szokolay S.V. 1973: Manual of tropical housing and building. Part 1: climatic design. Longman. London. Krishan A. 1996: The habitat of two deserts in India: hot-dry desert of Jaisailmer (Rajasthan) and the cold-dry high altitude mountainous desert of leh (Ladakh). Energy and Buildings 23: 217-229. 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 outdoor environments under clear sky conditions. Int. J. Biometeorol. 34: 69-75. Kuttler W. 1993: Planungsorientierte Stadtklimatologie. Aufgaben, Methoden und Fallbeispiele. Geogr. Rundschau 45. H. 2. Kuttler W. 2004: Stadtklima. Teil 1:Grundzüge und Ursachen. Beitagsserie: Kilamänderung 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. Computer 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 dynamic 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, microclimate 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 atmospheric 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 according to the German VDI-guideline 3787 part II. Prepr. 2nd Urban Envir. Symp. Albuquerque. 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 Leben 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 forest 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 methodology. 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. Freiburger 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. Boundary− Layer Meteorol. 56: 339-358. Olgay V. 1969: Design with the climate: Bioclimatic approach to architectural regionalism. Princeton University Press. New Jersey. ONM 1985 : Atlas climatologique national. Station Ghardaïa. Partie 1: Recueil de données. 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 Climatology. 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 1The 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 extremely hot and dry zones. Proc. 19th Int. Conf. on PLEA. Toulouse, France: 469473 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. Australia: 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 thermal 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 surface 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 semioutdoor environments in subtropical Sydney Australia. Building and Environment 38: 721-738. Staiger H., Bucher G., Jendritzky G. 1997: Gefühlte Temperatur: Die physiologischgerechte 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 Buildings 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 urban outdoor spaces. Theor. Appl. Climatol. 48: 49-61. Taha H. 1997: Urban climates and heat islands: albedo, evapotranspiration, and anthropogenic 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: 169177. 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 Asymmetry 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 pollutants 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. Boundary-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 canyon. 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

page

List of figure captions

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 irradiance 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 streets of aspect ratios H/W of 0.5, 1, 2, 3 and 4 207

81

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

various aspect ratios of 0.5, 1, 2, 3 and 4 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;

208

98


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, respectively 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 parallel 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 (northern 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 irradiance 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 BeniIsguen, 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

page

List of table captions

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 canyon 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 June and (b) 26 June 2003 in Beni-Isguen, Algeria (32.40°N, 3.80°E)

213

165

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.

WsK-1kg-1

cb

specific heat

cd,f

drag coefficient at the plant foliage (= 0.2)

cp

specific heat of air at constant pressure (=1847)

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

substrate heat (0: ground)

Wm-2

G

global radiation

Wm-2

H

Height of the building

G(0)

H(o,w,r) I i, j, k

Jkg-1K-1

m

turbulent sensible heat (for ground o, wall w or roof r surface)

Wm-2

direct solar radiation impinging normal to the surface

Wm-2

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

L

latent heat of vaporization

Jkg-1

L*

Total long-wave radiation

Wm-2

m

water content in a layer of air

g

LAD

leaf area density

2

m m-3

LAI

leaf area index

m3m-3

LAI*

“3D”leaf area index including angle of incidence of sun rays

m3m-3

l

lcl LE(0)

Km2W-1

heat resistance of the clothing 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

heat flux through wall or roof

Wm-2

Q*

net all-wave radiation

Wm-2

Q*

net radiation budget of the body

QE

latent heat flux density

QE

additional turbulence produced by vegetation

Q(w,r)

QH

sensible heat flux density

QH

convective heat flux from body

W Wm-2 s-1 Wm-2 W

215

List of symbols

ms-2

Qh

sink/source terms due to heat

QL

latent heat flow for diffusion of water from the body

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

net long-wave radiation (g,w,r: for ground, wall or roof surface)

Wm-2

short-wave radiation flux at the model boundary

Wm-2

(0 ) Rsw ,dir

direct short-wave radiation flux density(0: at model boundary)

Wm-2

(0 ) Rsw , dif

diffuse short-wave radiation flux density(0: at model boundary)

Wm-2

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 ,r ) Rlw( g,net

R*sw

Srad Su , Sv, Sz

W

W

mean radiation flux density absorbed by a human body

Wm-2

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

Ta,i

air temperature inside the buildings

K or °C

216

K

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

Ts

ground surface temperature

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

U

heat transmittance

u*

friction velocity

ms-1

wind speed in the x, y, and z directions (ug, vg, wg, geostr. wind)

ms-1

u, v, w ui, xi

K K or °C

m s-1 Wm-2K-1

i.e. u, v, w and to x, y, z with i = 1,2,3 (einstein summation)

-

horizontal wind speed

m s-1

Vb

blood flow from body core to skin

ls-1m-2

VP

vapour pressure

W

Width of the street

W

physical work output (activity)

v

hPa m W 2

2

2 0.5

ms-1

W

mean wind speed at height z (w = u + v + w )

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

Wm-2

energy stored in buildings

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

σB

stefan-boltzmann constant (=5.664 10-8)

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

-

K Wm-2K-4

kgm-3

ρ

air density (ρ0 =1.29)

µ

thermal admittance

ρb

blood density

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

°

Jm-2s-0.5K0.5 kgl-1

218

-

Appendix

Appendix A1: Database of soil types in ENVI-met 3.0. These individual soils are combined to form a “soil profile” as shown in Appendix A2. ID

ηs

ηfc

ηwilt

ψpot

ρi c i

Kη,s

λ

B

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

gal.1

0.227

gal.2 gal.3

H/W = 3

H/W = 4

ASYM.

OVERH.

0.117

0.117

0.117

0.227

0.117

0.117

0.117

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 Atmosphäre 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 Erdoberfläche 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. Geburtstag“. 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 Untersuchungsflächen der Landnutzungen “Grasland“ und „Wald“ in der südlichen 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 direkten und indirekten Daten vor der Instrumentenbeobachtung. Dezember 2004 Nr. 14: Imbery, F.: Langjährige Variabilität der aerodynamischen Oberflächenrauhigkeit und Energieflüsse eines Kiefernwaldes in der südlichen Oberrheinebene (Hartheim). Januar 2005 Nr. 15: Ali Toudert, F.: Dependence of outdoor thermal comfort on street design in hot and dry climate. November 2005

224