ENVIRONMENTALLY FRIENDLY CITIES, Proceedings of PLEA '98, Lisbon, Portugal, June 1998
Pages 163-166 © 1998 James & James Publishers Ltd

Street canyon geometry and microclimate:
Designing for urban comfort under arid conditions


The Center for Desert Architecture and Urban Planning
J. Blaustein Institute for Desert Research ~ Ben-Gurion University of the Negev
Sede-Boqer Campus 84990 Israel


While urban design guidelines have been developed for responding to climate in various regions, these recommendations are often based more on intuition or sporadic observation than on an integrated microclimatic analysis of thermal comfort conditions. Quantitative studies on desert environments are especially lacking, since most arid regions remain sparsely populated. In the present study, empirical data taken from extensive full-scale measurements in a number of low-rise urban street canyons in the arid Negev region of Israel are integrated with a simple numerical model representing the overall thermal energy exchange between a pedestrian and the street canyon environment. The integrated thermal index produced allows a comprehensive means for comparing geometric alternatives and generating guidelines which can aid in the design of urban spaces under climatically similar conditions. Analysis of overall energy balance suggests that in summer, overheating within the canyon is sensed primarily as a nocturnal phenomenon, and that during hours of substantial heat stress in a desert climate, compact urban spaces do in fact constitute potential "cool islands," mainly due to internal shading. In winter, a compact geometry was found to provide relatively warm conditions during most hours, with the key factor being protection from chilling by strong winds.


Climatic response in urban design is typically based on intuition or sporadic observation, rather than on systematic analysis of local conditions. An illustration of this issue can be seen in the planning of towns and cities in Israel's arid southern region, the Negev, where many neighborhoods were planned according to the "Garden City" model imported from Europe, with relatively large open spaces resulting from a broad dispersal of buildings throughout the landscape. In response to the perception of such an approach as being climatically inappropriate for the arid Negev, attempts have been made to create denser, more compact residential environments with the goal of improving microclimatic conditions [1]. The success of such attempts has yet to be systematically evaluated.

In theoretical terms, many of the advantages and drawbacks of compact planning in a desert environment have been documented. A dense urban fabric may provide solar shading of pedestrians within deep street canyons; on the other hand, such canyons may become relative "heat traps" due to multiple solar reflection and reduced albedo, diminished night sky radiation, and substantially restricted ventilation [2,3].

The relative impacts of such opposing thermal phenomena, and particularly their implications for pedestrian comfort in an arid setting, remain to be sufficiently addressed. Attempts at modelling have long been limited by the range of variables which can be accurately simulated, and when overall energy balance is considered, it is often done so for the general "urban terrain" or for the urban canyon as a system, rather than for a body located at a particular point within [4]. Finally, existing studies which do address the integrated thermal effect of urban spaces on human comfort have rarely focused on desert environments, since most arid regions have remained sparsely populated. Special attention to cities in such regions may be justified not only by their unique climatic characteristics, but by differences in their traditional urban morphologies (e.g. low-rise rather than high-rise street canyons) and by recent increases in arid zone urban development.

In the present study, empirical data taken from extensive full-scale measurements in a number of low-rise urban street canyons in the Negev are integrated with a numerical model representing the energy exchange between a pedestrian and the street canyon environment. Analysis of the monitored and calculated results reveals the relative nature of thermal stresses imposed by differences in canyon geometry on urban inhabitants of an arid region.


fig 1
Figure 1. Patio House neighborhood plan, showing points of observation and summer afternoon shading patterns.

The site investigated was a "patio house" neighborhood, located in the city of Dimona (elevation 600m) in the arid Negev Highlands of southern Israel. The neighborhood is composed of single-storey row houses and attached walled courtyards, with a grid of narrow (3m width) pedestrian paths ("street canyons") dividing the rows (Fig. 1). The pattern of streets and housing blocks is relatively compact and of a regular, well defined three-dimensional geometry, and street canyons are symmetrical with a height-width (H/W) ratio close to unity.

As is typical for many arid zones, the climate of the region is characterized by wide diurnal and seasonal thermal fluctuations. An average daily temperature range of 20-32oC in July is accompanied by low daytime relative humidity, intense solar radiation and strong late afternoon winds, predominantly from the northwest. In winter, while minimum daily temperatures occasionally reach freezing and winds are strong, clear skies and abundant solar radiation prevail in the daytime [5].

Field measurements

Measurements were carried out in two perpendicular street canyons, whose axes are approximately east-west and north-south, respectively, with the neighborhood grid rotated 30o counter-clockwise from the cardinal directions (Fig. 1). Comparative measurements were taken above the roof of the single-storey row house building. Parameters monitored included air temperature, relative humidity, wind speed and direction, global radiation, net radiation, and radiant surface temperature. Continuous 24-hour monitoring was conducted during summer (June through August) and winter (December through February) seasons over a one year period.

Energy exchange modelling

In order to integrally characterize the thermal environment within each canyon, the overall energy balance of a theoretical body representing a pedestrian in the street was calculated, and compared with that of a similar body above the roof. This energy balance model encompasses both radiative and convective thermal exchanges, while evaporative exchanges, although significant to physiological comfort and overall energy balance, were excluded from the model on the assumption that neither absolute nor relative humidity would vary significantly between points of comparison within the study area, which is largely devoid of vegetation.

fig 2
Figure 2. Schematic diagram of energy exchanges between cylindrical body and street canyon environment.

The radiation balance was derived from measured radiation fluxes and surrounding surface temperatures, and the convective balance from air temperature and wind velocity. By calculating the overall balance of simultaneous radiative and convective fluxes, the effects of solar exposure, radiant heating, air temperature and wind were evaluated in one integrated index which expresses the total rate of thermal energy flux between a pedestrian and the urban environment.

Monteith [6] provides an equation used to calculate the instantaneous Net Radiation (Rn) at the surface of a vertical cylinder which simulates a person standing on the ground, with account taken of the energy exchanged between the two surfaces. Calculation of the energy balance between a cylindrical body and the urban street canyon space (Fig. 2) requires the following generalized version of this expression, to account for the additional effects of vertical surfaces (all fluxes in W/m2 of cylinder surface area):

Rn = (1-rs)(Gdir + Gdif + r hGh + rvGv) + Ld + Lh + Lv - esTs4 . . . [1]

in which:

Rn = net radiation (all wavelengths);
rs = albedo of skin;
Gdir and Gdif = direct and diffuse solar radiation incident on the cylindrical surface;
rhGh = solar radiation reflected from the horizontal surface;
rvGv = solar radiation reflected from vertical surfaces;
Ld = downward long-wave radiation emitted by atmosphere;
Lh = upward long-wave radiation emitted by horizontal surface;
Lv = long-wave radiation emitted from vertical surfaces;
e = emissivity of the body;
s = Stefan-Boltzman constant;
and Ts = average surface temperature of the body (oK).

It may be seen that the absorption of short-wave radiation is divided into direct, diffuse and reflected components, with each flux averaged over the body surface to account for angles of incidence and shading coefficients (for direct radiation on body and surfaces), and for View Factors (for exposure to diffuse and reflected radiation).

Absorption by the body of long-wave radiation emitted from the environment is similarly divided into several components representing the surrounding urban surfaces, each computed as a product of the relevant View Factor (calculated as an average for the surface using the cylinder's vertical midpoint) and flux density, which for built surfaces was based on their measured temperature. For calculation of the instantaneous long-wave radiation flux emitted by the body, a constant skin temperature of 34oC was taken to represent a pedestrian encountering street canyon conditions after emerging from the indoor environment. The sum of these long-wave fluxes, together with short-wave fluxes, yields an overall radiation balance between the cylindrical body and its urban environment.

The rate of convective heat transfer (C), or sensible flux, per unit area of the body in W/m2 is given as (C = hCDT) in which hC is a transfer coefficient (W/m2oC), and D T (oC) is the differential between the body’s surface temperature and that of the surrounding air (measured on-site). One empirical expression for hC developed from observation using a vertical cylinder of diameter 0.17m, is given by Mitchell [7] as (hC = 8.3V0.6)in which V is the wind velocity in m/s.

The results of the overall energy balance, in particular the radiative balance, are largely dependent on the size and position of the "human cylinder" representing a pedestrian in the street. By placing a cylinder of height 1.50 m and radius 0.17 m at the center line of the canyon floor, a configuration was chosen which represents “average” conditions.


Temperature variation

fig 3
Figure 3. Summer daily cycle of air and surface temperatures, in E-W street canyon and above roof.

Differences in air temperature (Fig. 3) proved to be relatively minor: at head-height, air within the canyons was warmer than above-roof air by up to 3oC during afternoon hours, with the highest temperatures observed in the E-W street. Thermal stratification of air within the street canyon (expressed as a consistent vertical gradient by which air temperature decreased with height) was in fact less intense than that above the roof surface - an effect attributed, presumably, to heating from adjacent walls - which, in conjunction with impeded ventilation, results in a "pocket" of relatively warm air within the street canyon.

While radiant surface temperatures measured on vertical walls expectedly varied by orientation and height of measurement, the scale of these variations was consistently small in comparison with the difference between temperatures of vertical and horizontal (either floor or roof) surfaces, with the latter reaching peaks of 50oC, or higher by some 10oC than the average temperature of canyon walls and floor.

Relative wind speed attenuation

fig 4
Figure 4. Wind attenuation vs. wind direction; regression lines for various heights of measurement.

A pattern of selective wind speed attenuation, with canyon winds abated in general and to a greater extent in the street perpendicular to wind flow, was observed universally in all monitoring periods and for all measurement heights within the street canyon.

Figure 4 shows a fairly clear relation, for measurements at 1.50 m above the canyon floor, between wind speed attenuation, (expressed by the diminution factor P, or the ratio between canyon and "free" above-roof wind speeds) and relative wind direction, as expressed by the angle of attack between the average wind direction and the canyon axis. When the attack angle is near-parallel, canyon wind speed is attenuated to an average of 2/3 (P = 0.66) of the free flow. Perpendicular winds are more severely attenuated to about 1/3 of free flow.

The intensity of wind attenuation with attack angle was also seen to vary with depth below the canyon walls (measurement points varying from a height of +350 down to a height of +050): at near-perpendicular attack angles, the diminution factor (P) expectedly decreases with depth - but at near-parallel angles, the relation is reversed. The practical interpretation of these patterns is that when the prevailing wind is at a sharp angle to the street, an increasingly compact street canyon geometry (high H/W ratio) will most certainly reduce the rate of air flow, even in a low-rise street canyon. When the wind is directed along the street's length, however, air flow may be stronger in a deep canyon than it is in a shallower one.

Pedestrian Energy Exchange

fig 5
Figure 5. Energy exchange differential (body in canyon vs. above-roof) by street orientation and season.

Calculated man-environment energy exchanges in W/m2 are presented here in terms of "relative exchange," or as the difference between the energy balance of a body in the street canyon and that of a reference (above-roof) body. Thus, positive values indicate an increased gain (or reduced loss) for the body in the canyon, and negative values indicate a reduced gain (or increased loss).

Summation of the overall relative energy exchange by radiation and convection between body and street canyon (Fig. 5) shows that during most summer daytime hours, a pedestrian in the street canyon absorbs less thermal energy from the environment than one in the open. Thus while previously discussed observations of elevated air temperature and impaired wind flow within the street canyon suggest it to be a relatively overheated environment, the combination of all relevant parameters paints a somewhat different picture.

Calculation of individual energy exchange components indicates that this summer daytime difference (which in late afternoon peaks at -140 W/m2 for the E-W street and -200 W/m2 for the N-S street) is primarily due to shading of the body in the canyon from solar radiation (both direct and diffuse), which compensates for a reduction in its heat loss by convection (due to higher air temperatures and lower wind speeds).

In winter, a relative reduction in energy loss from the body to its street canyon surroundings (positive differential) is seen during most hours, primarily due to protection from cold winds (reduction in sensible heat loss). While overall energy loss during summer night hours was typically reduced in both canyons by about 100 W/m2, the reduction on winter nights reached over twice that during a strong gust.

Discussion and Conclusions

The above evidence suggests that like the urban "heat island" in general, the micro-scale heating effect within an urban canyon such as those analyzed here is primarily a nocturnal phenomenon: during daylight hours, the compact canyon is in fact a potential "cool island." This finding, however, is based on the thermal condition of a body in the street rather than that of the air surrounding it, and may be attributed to several factors:

1) Shading - As illustrated in Fig. 1, the street canyon’s closely spaced flanking walls (particularly in the N-S street) provide both its surfaces and its with considerable shade during all but midday hours in the summer. However, the reduced absorption of short-wave radiation by a pedestrian in the street is not exclusively a function of shading from direct sun. A pedestrian standing at the midpoint of either street canyon is exposed not only to less direct radiation (due to shading of the body), but also to less diffuse radiation (due to a restricted Sky View Factor) and to less radiation reflected from the horizontal ground surface (due to both shading of the street and a restricted View Factor) throughout summer daylight hours. Together, these considerably offset the absorption by a body in the street of radiation reflected from vertical wall surfaces. The above relationships hold for winter as well as summer, though they of course vary in scale according to street orientation, and in their significance for thermal comfort.

2) Radiant Heating - In terms of a body's net-radiative balance, relationships during daytime hours are similar to those for short-wave absorption; that is, the body in the canyon absorbs less total radiation of all wavelengths than that above the roof. This is in contrast to an observed increase in measured Net Radiation for the street canyon as an overall urban surface, when the measurement is taken at the imaginary upper plane of the canyon volume and therefore does not account for shading of a body within the space.

The limited extent of radiant heat gain by a pedestrian in the canyon during daytime hours may be partially explained by the restricted exposure of a pedestrian to the horizontal ground surface - which of all surfaces is the greatest source of radiant heat, as it a) intercepts the highest flux density of solar radiation, and b) is often of a dark color, thus maintaining by far the highest surface temperature during the daytime. It is only at night that this relationship changes, and the importance of restricted radiant heat loss to the sky becomes dominant.

3) Ventilation - During most daytime hours, ventilation is not a dominant factor, since its effectiveness in an arid region like the Negev Highlands is in any case limited by high air temperatures and low wind speeds. Ventilation is crucial, however, during the late afternoon and evening hours, when winds are consistently strong and cool, and during this period the importance of orientation is most evident: a space such as the north-south street in the Patio House neighborhood can provide substantial convective cooling by wind from the north-northwest, in addition to the previously mentioned effect of solar shading.

4) Thermal inertia - A further explanation is that most urban surfaces have a high thermal inertia. Although net radiative income during the day is higher in a compact street (as an overall system), the impact of such heating is sensed primarily at night, when the built elements release their stored heat.

Given the desert atmosphere's rapid cooling during summer evening hours, the nocturnal nature of overheating which was observed in a compact street canyon poses considerably less of a burden than it would in more humid areas, where the need for cooling in the summer is spread over a greater part of the daily cycle. In winter, the night time reduction in heat loss is a distinct thermal advantage, and the negative effect of daytime shading is relatively minor - particularly in a N-S street.

Comfort in a desert environment - whether indoors or outside - depends to a great extent on the stabilization of thermal extremes, between day and night and between summer and winter. Given proper attention to geometry and detail, the ‘compact’ urban street canyon appears to exhibit the potential for such thermal moderation.


  1. Y. Gradus and E. Stern (1985) From preconceived to responsive planning: Cases of settlement design in arid environments, in Desert Development: Man and Technology in Sparselands, Ed. Y. Gradus, D. Reidel Publishing Co., Dordrecht, etc.
  2. T.R. Oke (1988) "Street design and urban canopy layer climate," Energy and Buildings 11, pp. 103-113.
  3. Y. Nakamura and T.R. Oke (1988) Wind, temperature and stability conditions in an east-west oriented urban canyon. Atmospheric Environment 22, 2691-2700.
  4. A.Yoshida, K. Tominaga and S.Watatani (1991) Field measurements on energy balance of an urban canyon in the summer season. Energy and Buildings 15-16, 417-423.
  5. A. Bitan and S. Rubin (1991) Climatic Atlas of Israel for Physical Planning and Design, Tel Aviv University, Israel Meteorological Service and Ministry of Energy and Infrastructure.
  6. J.L. Monteith (1973) Principles of Environmental Physics, Edward Arnold Publishers, London, pp.74-77.
  7. D. Mitchell (1974) Convective heat transfer from man and other animals, in Heat Loss from Animals and Man: Assessment and Control, Ed. J.L. Monteith and L.E.Mount, Butterworths, London, pp. 59-76.

ACKNOWLEDGEMENTS - The author wishes to acknowledge the invaluable contribution made to this paper by Dr. Pedro Berliner of the Blaustein Institute for Desert Research. Thanks also to Prof. Arieh Bitan of Tel-Aviv University and Prof. Yehuda Gradus of Ben-Gurion University for their support of the original research, and to members of the Center for Desert Architecture and Urban Planning for their assistance throughout.